26
0263–8762/04/$30.00+0.00 # 2004 Institution of Chemical Engineers www.ingentaselect.com =titles=02638762.htm Trans IChemE, Part A, April 2004 Chemical Engineering Research and Design, 82(A4): 527–552 INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC Case Study R. MADHUSUDANA RAO 1 , R. RENGASWAMY 1, *, A. K. SURESH 2 and K. S. BALARAMAN 3 1 Department of Chemical Engineering, Clarkson University, Potsdam, New York, USA 2 Department of Chemical Engineering, Indian Institute of Technology, Bombay, India 3 Research and Development, Chennai Petroleum Corporation Limited, Chennai, India P rocess modelling and simulation have emerged as important tools for detailed study and analysis of chemical processes. In activities such as design, optimization and control of processes, realistic process models, which incorporate physics and chemistry of the process in adequate detail, are becoming almost indispensable. Simulation studies also provide guidance in the development of new processes and can reduce both time and capital investment. A dif culty with process models is that they are based on the state of knowledge and simulation objectives de ned at the time of their formulation. In addition, it is not easy to modify process models to incorporate new knowledge as it becomes available and as new needs arise. There is a need, therefore, to use advanced modelling and simulation strategies such that re nements and additional capabilities can be incorporated in the model without dispropor- tionate additional effort. This work presents the framework of one such multipurpose process simulator, MPROSIM, an object-oriented process modelling and simulation environment. Though considerable literature is available on process modelling from a subjective or theoretical viewpoint, very little has been published on application of these ideas on complex industrial-scale processes. This being the focus of the paper, a case study of an object-oriented model for automatic generation of a uid catalytic cracking unit (FCCU) reactor=regenerator is presented. The utility of the framework is illustrated by demonstrating how the model for FCCU could be ne-tuned both structurally and parametrically to represent the behaviour under changing process operating conditions. Keywords: FCCU; FCC model; object-oriented model; process modelling; MPROSIM. INTRODUCTION There is a greater need today than ever before to design and operate chemical processes with a high degree of under- standing. This need arises from several factors: an incessant demand for higher quality yields and better products; an increasingly competitive environment that forces plants to be operated in an optimal manner; variable raw material quality; stringent environmental and safety regulations; an increased level of automation arising out of the above factors, and so on. This need for a greater level of under- standing has to be seen in context of the fact that chemical processes are usually complex to understand and operate. The complexity arises at different levels: at the level of physicochemical phenomena involved; at the equipment level; and nally at the level of the plant where topological factors resulting from recycles and other connectivities impose additional demands. All these factors mean that mathematical models of increasing complexity have to be used in order to design and operate processes pro tably. Good models can also guide the development of new processes and can substantially reduce both time and capital requirements in process development. The utility of a process model strongly depends on its predictive capabilities. The predictions should be reliable over wide ranges of feed composition and process conditions. The availability of such a model, in conjunction with a simulation tool, can then facilitate better insight into plant behaviour through simula- tions instead of the more expensive and time consuming route involving actual experimentation. The advantages resulting from efforts invested in the activities of process modelling and simulation have to be seen against the fact that any mathematical model is necessarily based on a set of assumptions. Typically, the assumptions are governed by several factors: objectives the model is meant to serve at the time of its formulation; 527 *Correspondence to: Professor R. Rengaswamy, Department of Chemical Engineering, Clarkson University, Potsdam, NY 13699-5705, USA. E-mail: [email protected]

Industrial Experience with Object-Oriented Modelling FCC Case Study

Embed Size (px)

Citation preview

Page 1: Industrial Experience with Object-Oriented Modelling FCC Case Study

0263ndash876204$3000+000 2004 Institution of Chemical Engineers

wwwingentaselectcom=titles=02638762htm Trans IChemE Part A April 2004Chemical Engineering Research and Design 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTEDMODELLING

FCC Case Study

R MADHUSUDANA RAO1 R RENGASWAMY1 A K SURESH2 and K S BALARAMAN3

1Department of Chemical Engineering Clarkson University Potsdam New York USA2Department of Chemical Engineering Indian Institute of Technology Bombay India3Research and Development Chennai Petroleum Corporation Limited Chennai India

P rocess modelling and simulation have emerged as important tools for detailed study andanalysis of chemical processes In activities such as design optimization and control ofprocesses realistic process models which incorporate physics and chemistry of the

process in adequate detail are becoming almost indispensable Simulation studies also provideguidance in the development of new processes and can reduce both time and capitalinvestment A dif culty with process models is that they are based on the state of knowledgeand simulation objectives de ned at the time of their formulation In addition it is not easy tomodify process models to incorporate new knowledge as it becomes available and as new needsarise There is a need therefore to use advanced modelling and simulation strategies such thatre nements and additional capabilities can be incorporated in the model without dispropor-tionate additional effort This work presents the framework of one such multipurpose processsimulator MPROSIM an object-oriented process modelling and simulation environmentThough considerable literature is available on process modelling from a subjective ortheoretical viewpoint very little has been published on application of these ideas on complexindustrial-scale processes This being the focus of the paper a case study of an object-orientedmodel for automatic generation of a uid catalytic cracking unit (FCCU) reactor=regenerator ispresented The utility of the framework is illustrated by demonstrating how the model forFCCU could be ne-tuned both structurally and parametrically to represent the behaviour underchanging process operating conditions

Keywords FCCU FCC model object-oriented model process modelling MPROSIM

INTRODUCTION

There is a greater need today than ever before to design andoperate chemical processes with a high degree of under-standing This need arises from several factors an incessantdemand for higher quality yields and better products anincreasingly competitive environment that forces plants tobe operated in an optimal manner variable raw materialquality stringent environmental and safety regulations anincreased level of automation arising out of the abovefactors and so on This need for a greater level of under-standing has to be seen in context of the fact that chemicalprocesses are usually complex to understand and operateThe complexity arises at different levels at the level ofphysicochemical phenomena involved at the equipmentlevel and nally at the level of the plant where topological

factors resulting from recycles and other connectivitiesimpose additional demands All these factors mean thatmathematical models of increasing complexity have to beused in order to design and operate processes pro tablyGood models can also guide the development of newprocesses and can substantially reduce both time and capitalrequirements in process development The utility of aprocess model strongly depends on its predictivecapabilitiesThe predictions should be reliable over wide ranges of feedcomposition and process conditions The availability of sucha model in conjunction with a simulation tool can thenfacilitate better insight into plant behaviour through simula-tions instead of the more expensive and time consumingroute involving actual experimentation

The advantages resulting from efforts invested in theactivities of process modelling and simulation have to beseen against the fact that any mathematical model isnecessarily based on a set of assumptions Typically theassumptions are governed by several factors objectivesthe model is meant to serve at the time of its formulation

527

Correspondence to Professor R Rengaswamy Department of ChemicalEngineering Clarkson University Potsdam NY 13699-5705 USAE-mail raghuclarksonedu

the need to keep the model tractable consistent with theseobjectives state of knowledge about the processes andequipment at the time of formulation of model and so onAny or all of these are subject to substantial changes duringthe life of a plant It would be therefore desirable to choose aframework for modelling and simulation that allows the exibility of incorporating changes in parts of the modelwithout having to undertake the entire modelling exerciseanew The present work concerns the development ofa process modelling and simulation environment thatattempts to address such concerns The framework calledMPROSIM is based on the concepts of object-orientedmodelling and simulation

In the next section we rst describe the idea and conceptof MPROSIM Some of the recent contributions in processmodelling and methodology of process modelling and hie-rarchy of modelling objects in MPROSIM are also descri-bed To demonstrate the utility and merits of MPROSIM wepresent a case study that concerns the modelling andsimulation of FCCU which in many ways typi es thetypes of challenges in process modelling and simulationthat MPROSIM seeks to address In MPROSIM FCCU ismodelled as a owsheet using the fundamental chemicalprocess units The object-oriented design and representationof the model as a owsheet are also described Variousstudies conducted with the model such as reactor tuningregenerator tuning parametric studies on integrated reactor=regenerator model and yield optimization studies are alsopresented We conclude this work with some comments onthe utility of an object-oriented framework such as MPRO-SIM and extension of the framework to form a singleenvironment wherein various process engineering activitiescan be conducted

MPROSIMETHMULTIPURPOSEPROCESS SIMULATOR

Concept

Typically in a chemical production unit process modelsare used of ine instead of online Some of the reasons forthis may have to do with issues of reliability of the processmodel inadequate knowledge of the model input para-meters and presence of noise in the data collected fromplant Model reliability is established through extensivevalidation after the model has been tuned by estimating themodel parameters Hence steady-state simulation datareconciliation (for error detection in the plant data) andparameter estimation are some of the process engineeringactivities that are very closely related Typically these studieswould be carried out in different environments and wouldalso often be based on different models The main aim ofMPROSIM is to develop an integrated framework whereinvarious process engineering activities can be carried outwithin the same environment and with the same model

Review of Process Modelling

Traditionally modelling involves identifying the keyphenomena occurring in the system that in uence thefeatures of interest developing mathematical equations todescribe the phenomena and solving these equations most

often using numerical techniqueswith the help of computersModel development as an activity is expensive requiring asit does highly skilled manpower with a good amount ofprocess knowledge Considerable work is also involved invalidating the models Because of the high costs associatedwith model development and validation it is desirable thatthe models retain their relevance and usefulness over areasonably long period of time

In recent years a variety of process modelling andsimulation tools have been developed As for the latterthere are several tools some of which have becomecommercially successful eg ASPEN PLUS PRO II etcThese tools are widely used in the petroleum industry assteady state simulators Some of the reasons for theirextensive use are uid data handling capabilities physicalproperty and component data base and an easy to usegraphical user interface (GUI) for unit operations and owsheet simulations Extending these tools to modelmore speci c or complex processes is dif cult Some ofthe typical examples that can be cited are mineral and solidsprocesses that involve particulate systems membraneprocesses polymer reaction systems and multiphase reac-tors To illustrate this consider a particulate process whereit would be necessary to describe the properties ofthe dispersed phase (say solids) such as particle sizeparticle shape and=or porosity as a part of the streaminput General purpose simulators allow only for the de ni-tion of a discretized particle size distribution by mass foreach solids stream (Gruhn et al 1997) Though the particlesize distribution may be easily added as a user de nedvariable for the process stream the user must also supply thecode to handle the additional data at each unit model ieeven at a simple splitter (Toebermann et al 2000) Also theparameters for a process unit (eg size reduction equip-ment) depend on the type of apparatus process-relevantdesign operating variables (gap size) and materials beingprocessed (Gruhn et al 1997) Apart from the extensibilityconstraints one of the main drawbacks of these simulationtools is that the models and problem solver are tightlycoupled in software modules ie FORTRAN subroutines(Nilsson 1993)

Because of the dif culties in adaptability and extensibilityof process simulation tools process modelling tools basedon equation-oriented approach have gained importanceExamples of commercial simulators are SPEEDUP andgProms They support the implementation of unit modelsand their incorporation in a model library ie the user canspecify all model equations Modelling of a process unitwould require a profound knowledge not only of chemicalengineering but also of such diverse areas as modelling andsimulation numerical mathematics and computer scienceHence extending the standard models and further develop-ment of speci c unit models can be undertaken only by asmall group of modelling experts Successful attempts havebeen made on extending SPEEDUP and gProms for model-ling speci c units of mineral processing problems (Bartonand Perkins 1988) and crystallization processes (Pantelidesand Oh 1996) However the large effort for the set-up andevaluation of the equation system as well as the expert userknowledge is a disadvantage (Toebermann et al 2000)Equation-oriented languages sometimes result in redundantmodelling as they do not assist the user in developingmodels using engineering concepts and thus reuse of even

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

528 R MADHUSUDANA RAO et al

well-designed and validated models is sometimes compro-mised (Marquardt 1996)

While both the above approaches have their advantagesand disadvantages the experience of early researchers inthese two approaches towards process modelling hastriggered considerable interest over the last decade Effortshave been directed towards addressing issues like modelformulation reusability of process models extensionsand adaptability to changing conditions As a result of theefforts of various researchers some of the developmentsin advanced process modelling tools and environmentsare ASCEND (Piela et al 1991) gPROMS (Barton andPantelides 1994) MODELLA (Stephanopoulos et al1990) and OMOLA (Nilsson 1993) One of the commonfeatures of all modelling tools is multi-level modularizationThe idea of modularization has been inherited from theconcepts of object-orientation which in turn has developedas a branch of computer science and engineering The mainidea of object-oriented design is to break a large andcomplex system into smaller fragments or modules so asto attain abstraction at each modular level The object-oriented approach directly supports reusability of commonfunctions in various parts of the program Further it alsoallows extensibility of the whole system by the implementa-tion of modules The reader is referred to the originalliterature cited above for a detailed summary on the devel-opment of each modelling tool State-of-the-art reviews ondevelopments in process modelling are also available(Marquardt 1991 1996 Biegler 1989 Boston et al1993 Pantelides and Barton 1993)

Some of the recent advances in the area of processmodelling have been towards a systematic approach to thedevelopment of process models (process modelling metho-dology) and formal representation of process model equa-tions (Marquardt 1994 1996 Bogusch and Marquardt1995 1997 Lohmann and Marquardt 1996) Marquardtand his group have laid the foundations for systematicprocess model development and at the same time haveproposed an object-oriented methodology for the task ofcomputer-aided process modelling Extending the ideas ofsystem theoretic approach and object-orientation a processunit model can be decomposed into smaller modules basedon its structure (for example reactor wall catalyst heatmaterial etc for a tubular reactor) and behaviour (conser-vation constitutive equations etc) Similarly the behaviouralobjects can further be decomposed into more fundamentalentities (for example holdup transport transfer etc for aspecies conservation equation) Following this methodologyMarquardt and his group have proposed a structure forhierarchy of elementary process quantities for developinga unit model A real unit model is de ned by aggregating aset of these canonical=fundamental modelling objects

The main aim of adopting such a modelling methodologyis to address the following two issues (1) at the fundamentallevel any chemical process be it petroleum petrochemicalmetallurgical or biological is the same ie conservationof species components or particles and conservation=conversion of energy Hence with a well-structured andgeneric modelling tool it should be possible (ideally) tode ne any model by aggregating the elementary modellingobjects and (2) to be able to develop a unit model from afundamental approach and also reuse and extend the exist-ing library of models to t the changing requirements

However in spite of all the developments in the area ofobject-oriented structured process modelling over the lastdecade there still remain several important issues that needto be addressed They can be classi ed into two categories

Generic issues Some of the issues aremdash the amount of effort and expert knowledge that would

be required in various elds of science and engineer-ing for developing such a fundamental and genericmodelling tool

mdash the capabilities and conduciveness of the existingmodelling and software tools for encoding and repre-senting the vast amount of diversi ed scienti c andengineering knowledge in a structured format

mdash the extent of granularity that one should build tomodel any given process system One of the practicalissues in this context is that a given hierarchy ofprocess model structuring however ne it may be inits granularity might still be not complete Forexample in mineral processing the liberation statein gravel and sand industry processes the fractionaldensity and in environmental processes like soil-washing the fractional contamination must be con-sidered along with the particle size distribution tocharacterize a process stream (Toebermann et al2000) Similarly processes from different elds ofengineering and technology may require additionalinput and parameter speci cations for both processstreams and process models

mdash uniqueness of a decomposition strategy A chemicalprocess can be represented by an aggregation ofseveral possible combinations of fundamentalmodelling objects of varying degrees of complexi-ties Thus the decomposition strategy for de ningthe model objects gives rise to many degrees offreedom The problem of selection of a particularstrategy for process representation can be adif cult task when a large scale industrial processis considered

mdash extension and adaptability of a set of fundamentalmodelling objects A generic process modellingframework must be capable and exible to be ableto extend and incorporate additional information=knowledge

Implementation issuesmdash Though some general guidelines for structured

process modelling have been stated (Marquardt1992 1996) the utility of the formalism can onlybe shown by the implementation and evaluation ofmodelling tools along with experimentation

While there is considerable literature on structuredprocess modelling from a subjective or theoretical view-point very little has been published on application ofthese ideas to complex real-life modelling problems Tothe authorsrsquo best knowledge there has been no directevaluation of the formalism or the guidelines for struc-tured process modelling through an implementation on alarge scale industrial process This is the main focus ofthe paper and FCCU is used as a case study The projectwas undertaken in joint collaboration with a re neryChennai Petroleum Corporation Limited (CPCL) IndiaThough there are few commercially available simulation

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 529

tools for FCCU like FCC Plus it is necessary to reiteratethat the concepts of object-oriented modelling and resultspresented in this contribution are not to undertake anevaluation of the simulation tools The modelling andsimulation of FCCU is chosen to demonstrate the meritsof MPROSIM and at the same time present an evaluationof structured object-oriented process modelling guidelinesThe reasons for the choice of FCCU are manifold

FCCU is a critical unit of a re nery and is thus importantin its own rightFCCU is a large system with complexities at severallevels heterogeneous operations solids handling hydro-dynamic complexities in the riser and uid-bed regen-erators complex kinetics resulting from the complexity offeedstock topological complexities due to recycles andinterconnected ows etc It is therefore a good test caseto test the implementation and evaluation of the ideas ofmodularization and structural decompositionVarious con gurations of reactor and regenerator are inoperation in the re neries Adaptability of the modellingframework to different con gurations can therefore betestedInputs to the unit often change in terms of both quality(composition of feed) and quantity (product demands inthe market) Often a new or improved catalyst becomesavailable This is precisely the situation where a simula-tion tool can prove its utility since ne tuning of theoperating conditions in response to such changes is oftenneededVarious assumptions have been made in literature formodelling the physics and chemistry of the crackingprocess As a result several approaches are available formodelling different sections of FCCU There is thus anopportunity to test the exibility of the framework tomodel the system using different approaches

In the next section we brie y describe the object-orientedmodelling framework developed in MPROSIM The frame-work is based on the principles of systematic processmodelling as enumerated in many of the works discussedabove

Modell ing in MPROSIM

MPROSIM is a framework for multipurpose processmodelling and simulation The simulator is based on theequation-oriented approach It has been developed usingthe concepts of object-oriented programming (OOP) Themodelling environment and the graphical user interface(GUI) are developed using JAVA Even though some ofthe simulator framework features are under development andseveral program module implementations are rudimentarythe overall software architecture of MPROSIM is illus-trated in Figure 1 The key idea that is being explored inMPROSIM is to make every activity of process engineeringa class or an object It is also designed to make all theimportant and essential features of process engineeringcompletely accessible from the GUI Hence the overallarchitecture of MPROSIM is at allowing the user tocarry out various process engineering activities with thesame model built using the GUI

Object-oriented modelling is a modern approach tohandle complexities The method involves modularization

and characterization of data into an abstraction called classA class is like a blueprint for an application or part of anapplication An object which is an instance or an incarna-tion of a class has both substance and behaviour bycomprising data and functions that perform operations onthis data A well abstracted class can be tested and imple-mented independently Abstraction helps in the reuse of aclass=object in various parts of the program Functionscommon to many classes can reside in one class and canbe implemented in other parts of the program with the helpof inheritance The common functionalities can be adaptedto meet speci c requirements with the help of polymorph-ism To protect parts of the program from unintentionalchanges or side effects variables and elements of an objectcan be encapsulated The ow of control in the program canbe threaded with the help of multithreading Multithreadingprovides a way for an application to handle many differenttasks at the same time (Niemeyer and Knudsen 2000) Eventhough JAVA supports only single inheritance it allowsmultiple implementation of interfaces All these featuresmake JAVA a conducive language to design the GUI and themodelling framework

The methodology of process modelling in MPROSIM hasbeen designed and developed based on the above conceptsof object-oriented programming and modelling The hier-archy of process modelling in MPROSIM as of present isdepicted in Figure 2 The rst step is a topological frag-mentation where a process is broken into model fragmentsor unit modules based on structure of the process Therepresentation of the original process is achieved by theaggregation of these model fragments with the help ofstreams in the form of a owsheet Each of the modelfragments resulting from the rst step of decomposition ismodular and well abstracted The model fragments can betested and implemented independent of other fragmentsthus ensuring their reuse At the second step a modelfragment or unit module is further decomposed into mole-cular fragments based on its internal structure and beha-viour A model fragment can be instantiated by a collectionof the molecular fragments

In the present development stage of MPROSIM thehierarchial modelling structure lies at the molecular fragmentlevel Further decomposition of the molecular fragmentsinto atomic fragments and issues relating to its testing andimplementation are some of the developments presentlyunder progress For the FCCU case study further

Figure 1 Software architecture of MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

530 R MADHUSUDANA RAO et al

decomposition of molecular fragments could not be under-taken due to limitations such as project schedule datacollection etc Though the modelling hierarchy builtcurrently is neither suitable nor extensible to model anygiven process system it is appropriate to address variousissues relating to the modelling of an industrial-scale FCCUIn addition some of the model fragments resulting from thedecomposition are generic in nature and can be extended tomodel other speci c units=processes as highlighted insection on Discussion and Clari cations below

For the implementation of the FCCU model a node classis de ned as the molecular fragment which contains all thetypical model equations such as mass balance energybalance etc Every unit module or model fragment inheritsthe node class and can modify or add additional informationto the inherited model equations to re ect its own behaviourThe FCCU model is realized in the form of a owsheet byan aggregation of unit modules by connecting them with the

help of streams A detailed description of the object-orientedmethodology for FCCU modelling in MPROSIM is given ina subsequent section As a precursor a brief review of theliterature on FCCU modelling is presented in the nextsection

Review of FCCU Modelling

Over the years many models have been proposed forFCCU These models have been based on different sets ofassumptions with respect to the kinetics of cracking reac-tions and hydrodynamics of the equipment involved such asriser and regenerator Some models concern themselves onlywith the regenerator (Ford et al 1976 Errazu et al 1979de Lasa et al 1981 Guigon and Large 1984 Krishna andParkin 1985 Lee et al 1989a) Some have only reactor orcracking models (Weekman and Nace 1970 Paraskos et al1976 Jacob et al 1976 Shah et al 1977 Lee et al

Figure 2 Hierarchy of process modelling in MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 531

1989b Larocca et al 1990 Takatsuka et al 1987) Therealso exist integrated models coupling both the regeneratorand reactor (Kumar et al 1995 Lee and Kugelman1973 McGreavy and Isles-Smith 1986 Bozicevic andLukec 1987 Arandes and de Lasa 1992 McFarlane et al1993 Arbel et al 1995 Arandes et al 2000) Table 1presents a list of literature which typify the approaches thathave been considered for FCCU modelling

The variety of approaches present in the literature can beanalysed with reference to Figures 3 and 4 The reactor riseris often modelled as being in plug ow (Arbel et al 1995)with uniform temperatures across a cross section and atemperature gradient along the height of the riser Otherassumptions such as quasi steady state no slip adiabaticoperation uniform temperature of the two phases at anypoint and constant heat capacities of oil vapour and catalystare usually made in the analysis On the other handMcFarlane et al (1993) assume isothermal conditions (ieCSTR type) in the riser though the bottom of the riserinvolves a non-isothermal zone due to a nite mixing time

Apart from this riser models differ mainly in their con-siderations of reaction kinetics Weekman and Nace (1970)developed a three lump model which was used by manyauthors Lee and Groves (1985) used the three lump model intheir integrated model Other models based on the Weekman-Nace three-lump model include those of McGreavy and Isles-Smith (1986) Kraemer and de Lasa (1988) Arandes andde Lasa (1992) Other authors expanded the three lumpsystem into more lumps Jacob et al (1976) developed aten lump system A four lump model was developed by Leeet al (1989b) Bozicevic and Lukec (1987) published a velump model Takatsuka et al (1987) developed a Weekmantype six lump model Kraemer et al (1990) extended theirmodel to eight lumps Two papers appeared in 1995 using tenlump kinetics with an integrated model Kumar et al (1995)used an isothermal reactor riser and Arbel et al (1995)modelled the riser assuming plug ow Arandes et al(2000) have considered the ten lump kinetics for dynamicand steady-state model of FCCU Plug ow conditions wereassumed in the riser and for gas ow in the regenerator

Among the models proposed for the regenerator mostfocus on the dense bed that is characterized by bubblesrising through an emulsion phase The earliest models were

single phase simple contacting models with plug owmixed ow dispersion and tanks in series The earlyresearchers Arthur (1951) Rowe and Partridge (1965) andWeisz and Goodwin (1966) simply divided the bed intoregions known as dense phase and dilute phase Later onmore complete models were developed the grid-effectmodel by Behie and Kehoe (1973) and Errazu et al(1979) the two-region model by de Lasa and Grace (1979)and de Lasa et al (1981) and the bubbling-bed model byKunii and Levenspiel (1968 1990)

A summary of the assumptions made by various research-ers for modelling the physics and chemistry of the processesin FCCU is illustrated in Figure 3 In the next section theFCCU case study is presented The case study describes indetail the process modelling methodology adopted for asystem like FCCU and its implementation in MPROSIMModel tuning and comparison of model predictions to theplant data are also presented The case study is concludedwith a discussion on some merits of developing FCCUmodel in an object-oriented framework like MPROSIM

CASE STUDY MODELLING OF FCCU

Process Description

The schematic of a current generation FCCU with stand-pipes and slide valves is represented in Figure 4 The reactorconsists of a vertical section called the riser The preheated

Table 1 A brief summary of literature on FCCU modelling

Author Year Model type Kinetic model

Weekman and Nace 1970 C 3-Lump modelJacob et al 1976 C 10-Lump modelFord et al 1976 R NAErrazu et al 1979 R NALee and Groves 1985 I 3-Lump modelKunii and Levenspiel 1990 R NAMcFarlane et al 1993 I NoneArbel et al 1995 I 10-Lump modelKumar et al 1995 I 10-Lump modelSriramulu et al 1996 R NAArandes et al 2000 I 10-Lump model

C Cracking model R Regenerator model I Integrated model

Figure 3 Summary of various assumptions for modelling FCCU

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

532 R MADHUSUDANA RAO et al

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 2: Industrial Experience with Object-Oriented Modelling FCC Case Study

the need to keep the model tractable consistent with theseobjectives state of knowledge about the processes andequipment at the time of formulation of model and so onAny or all of these are subject to substantial changes duringthe life of a plant It would be therefore desirable to choose aframework for modelling and simulation that allows the exibility of incorporating changes in parts of the modelwithout having to undertake the entire modelling exerciseanew The present work concerns the development ofa process modelling and simulation environment thatattempts to address such concerns The framework calledMPROSIM is based on the concepts of object-orientedmodelling and simulation

In the next section we rst describe the idea and conceptof MPROSIM Some of the recent contributions in processmodelling and methodology of process modelling and hie-rarchy of modelling objects in MPROSIM are also descri-bed To demonstrate the utility and merits of MPROSIM wepresent a case study that concerns the modelling andsimulation of FCCU which in many ways typi es thetypes of challenges in process modelling and simulationthat MPROSIM seeks to address In MPROSIM FCCU ismodelled as a owsheet using the fundamental chemicalprocess units The object-oriented design and representationof the model as a owsheet are also described Variousstudies conducted with the model such as reactor tuningregenerator tuning parametric studies on integrated reactor=regenerator model and yield optimization studies are alsopresented We conclude this work with some comments onthe utility of an object-oriented framework such as MPRO-SIM and extension of the framework to form a singleenvironment wherein various process engineering activitiescan be conducted

MPROSIMETHMULTIPURPOSEPROCESS SIMULATOR

Concept

Typically in a chemical production unit process modelsare used of ine instead of online Some of the reasons forthis may have to do with issues of reliability of the processmodel inadequate knowledge of the model input para-meters and presence of noise in the data collected fromplant Model reliability is established through extensivevalidation after the model has been tuned by estimating themodel parameters Hence steady-state simulation datareconciliation (for error detection in the plant data) andparameter estimation are some of the process engineeringactivities that are very closely related Typically these studieswould be carried out in different environments and wouldalso often be based on different models The main aim ofMPROSIM is to develop an integrated framework whereinvarious process engineering activities can be carried outwithin the same environment and with the same model

Review of Process Modelling

Traditionally modelling involves identifying the keyphenomena occurring in the system that in uence thefeatures of interest developing mathematical equations todescribe the phenomena and solving these equations most

often using numerical techniqueswith the help of computersModel development as an activity is expensive requiring asit does highly skilled manpower with a good amount ofprocess knowledge Considerable work is also involved invalidating the models Because of the high costs associatedwith model development and validation it is desirable thatthe models retain their relevance and usefulness over areasonably long period of time

In recent years a variety of process modelling andsimulation tools have been developed As for the latterthere are several tools some of which have becomecommercially successful eg ASPEN PLUS PRO II etcThese tools are widely used in the petroleum industry assteady state simulators Some of the reasons for theirextensive use are uid data handling capabilities physicalproperty and component data base and an easy to usegraphical user interface (GUI) for unit operations and owsheet simulations Extending these tools to modelmore speci c or complex processes is dif cult Some ofthe typical examples that can be cited are mineral and solidsprocesses that involve particulate systems membraneprocesses polymer reaction systems and multiphase reac-tors To illustrate this consider a particulate process whereit would be necessary to describe the properties ofthe dispersed phase (say solids) such as particle sizeparticle shape and=or porosity as a part of the streaminput General purpose simulators allow only for the de ni-tion of a discretized particle size distribution by mass foreach solids stream (Gruhn et al 1997) Though the particlesize distribution may be easily added as a user de nedvariable for the process stream the user must also supply thecode to handle the additional data at each unit model ieeven at a simple splitter (Toebermann et al 2000) Also theparameters for a process unit (eg size reduction equip-ment) depend on the type of apparatus process-relevantdesign operating variables (gap size) and materials beingprocessed (Gruhn et al 1997) Apart from the extensibilityconstraints one of the main drawbacks of these simulationtools is that the models and problem solver are tightlycoupled in software modules ie FORTRAN subroutines(Nilsson 1993)

Because of the dif culties in adaptability and extensibilityof process simulation tools process modelling tools basedon equation-oriented approach have gained importanceExamples of commercial simulators are SPEEDUP andgProms They support the implementation of unit modelsand their incorporation in a model library ie the user canspecify all model equations Modelling of a process unitwould require a profound knowledge not only of chemicalengineering but also of such diverse areas as modelling andsimulation numerical mathematics and computer scienceHence extending the standard models and further develop-ment of speci c unit models can be undertaken only by asmall group of modelling experts Successful attempts havebeen made on extending SPEEDUP and gProms for model-ling speci c units of mineral processing problems (Bartonand Perkins 1988) and crystallization processes (Pantelidesand Oh 1996) However the large effort for the set-up andevaluation of the equation system as well as the expert userknowledge is a disadvantage (Toebermann et al 2000)Equation-oriented languages sometimes result in redundantmodelling as they do not assist the user in developingmodels using engineering concepts and thus reuse of even

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

528 R MADHUSUDANA RAO et al

well-designed and validated models is sometimes compro-mised (Marquardt 1996)

While both the above approaches have their advantagesand disadvantages the experience of early researchers inthese two approaches towards process modelling hastriggered considerable interest over the last decade Effortshave been directed towards addressing issues like modelformulation reusability of process models extensionsand adaptability to changing conditions As a result of theefforts of various researchers some of the developmentsin advanced process modelling tools and environmentsare ASCEND (Piela et al 1991) gPROMS (Barton andPantelides 1994) MODELLA (Stephanopoulos et al1990) and OMOLA (Nilsson 1993) One of the commonfeatures of all modelling tools is multi-level modularizationThe idea of modularization has been inherited from theconcepts of object-orientation which in turn has developedas a branch of computer science and engineering The mainidea of object-oriented design is to break a large andcomplex system into smaller fragments or modules so asto attain abstraction at each modular level The object-oriented approach directly supports reusability of commonfunctions in various parts of the program Further it alsoallows extensibility of the whole system by the implementa-tion of modules The reader is referred to the originalliterature cited above for a detailed summary on the devel-opment of each modelling tool State-of-the-art reviews ondevelopments in process modelling are also available(Marquardt 1991 1996 Biegler 1989 Boston et al1993 Pantelides and Barton 1993)

Some of the recent advances in the area of processmodelling have been towards a systematic approach to thedevelopment of process models (process modelling metho-dology) and formal representation of process model equa-tions (Marquardt 1994 1996 Bogusch and Marquardt1995 1997 Lohmann and Marquardt 1996) Marquardtand his group have laid the foundations for systematicprocess model development and at the same time haveproposed an object-oriented methodology for the task ofcomputer-aided process modelling Extending the ideas ofsystem theoretic approach and object-orientation a processunit model can be decomposed into smaller modules basedon its structure (for example reactor wall catalyst heatmaterial etc for a tubular reactor) and behaviour (conser-vation constitutive equations etc) Similarly the behaviouralobjects can further be decomposed into more fundamentalentities (for example holdup transport transfer etc for aspecies conservation equation) Following this methodologyMarquardt and his group have proposed a structure forhierarchy of elementary process quantities for developinga unit model A real unit model is de ned by aggregating aset of these canonical=fundamental modelling objects

The main aim of adopting such a modelling methodologyis to address the following two issues (1) at the fundamentallevel any chemical process be it petroleum petrochemicalmetallurgical or biological is the same ie conservationof species components or particles and conservation=conversion of energy Hence with a well-structured andgeneric modelling tool it should be possible (ideally) tode ne any model by aggregating the elementary modellingobjects and (2) to be able to develop a unit model from afundamental approach and also reuse and extend the exist-ing library of models to t the changing requirements

However in spite of all the developments in the area ofobject-oriented structured process modelling over the lastdecade there still remain several important issues that needto be addressed They can be classi ed into two categories

Generic issues Some of the issues aremdash the amount of effort and expert knowledge that would

be required in various elds of science and engineer-ing for developing such a fundamental and genericmodelling tool

mdash the capabilities and conduciveness of the existingmodelling and software tools for encoding and repre-senting the vast amount of diversi ed scienti c andengineering knowledge in a structured format

mdash the extent of granularity that one should build tomodel any given process system One of the practicalissues in this context is that a given hierarchy ofprocess model structuring however ne it may be inits granularity might still be not complete Forexample in mineral processing the liberation statein gravel and sand industry processes the fractionaldensity and in environmental processes like soil-washing the fractional contamination must be con-sidered along with the particle size distribution tocharacterize a process stream (Toebermann et al2000) Similarly processes from different elds ofengineering and technology may require additionalinput and parameter speci cations for both processstreams and process models

mdash uniqueness of a decomposition strategy A chemicalprocess can be represented by an aggregation ofseveral possible combinations of fundamentalmodelling objects of varying degrees of complexi-ties Thus the decomposition strategy for de ningthe model objects gives rise to many degrees offreedom The problem of selection of a particularstrategy for process representation can be adif cult task when a large scale industrial processis considered

mdash extension and adaptability of a set of fundamentalmodelling objects A generic process modellingframework must be capable and exible to be ableto extend and incorporate additional information=knowledge

Implementation issuesmdash Though some general guidelines for structured

process modelling have been stated (Marquardt1992 1996) the utility of the formalism can onlybe shown by the implementation and evaluation ofmodelling tools along with experimentation

While there is considerable literature on structuredprocess modelling from a subjective or theoretical view-point very little has been published on application ofthese ideas to complex real-life modelling problems Tothe authorsrsquo best knowledge there has been no directevaluation of the formalism or the guidelines for struc-tured process modelling through an implementation on alarge scale industrial process This is the main focus ofthe paper and FCCU is used as a case study The projectwas undertaken in joint collaboration with a re neryChennai Petroleum Corporation Limited (CPCL) IndiaThough there are few commercially available simulation

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 529

tools for FCCU like FCC Plus it is necessary to reiteratethat the concepts of object-oriented modelling and resultspresented in this contribution are not to undertake anevaluation of the simulation tools The modelling andsimulation of FCCU is chosen to demonstrate the meritsof MPROSIM and at the same time present an evaluationof structured object-oriented process modelling guidelinesThe reasons for the choice of FCCU are manifold

FCCU is a critical unit of a re nery and is thus importantin its own rightFCCU is a large system with complexities at severallevels heterogeneous operations solids handling hydro-dynamic complexities in the riser and uid-bed regen-erators complex kinetics resulting from the complexity offeedstock topological complexities due to recycles andinterconnected ows etc It is therefore a good test caseto test the implementation and evaluation of the ideas ofmodularization and structural decompositionVarious con gurations of reactor and regenerator are inoperation in the re neries Adaptability of the modellingframework to different con gurations can therefore betestedInputs to the unit often change in terms of both quality(composition of feed) and quantity (product demands inthe market) Often a new or improved catalyst becomesavailable This is precisely the situation where a simula-tion tool can prove its utility since ne tuning of theoperating conditions in response to such changes is oftenneededVarious assumptions have been made in literature formodelling the physics and chemistry of the crackingprocess As a result several approaches are available formodelling different sections of FCCU There is thus anopportunity to test the exibility of the framework tomodel the system using different approaches

In the next section we brie y describe the object-orientedmodelling framework developed in MPROSIM The frame-work is based on the principles of systematic processmodelling as enumerated in many of the works discussedabove

Modell ing in MPROSIM

MPROSIM is a framework for multipurpose processmodelling and simulation The simulator is based on theequation-oriented approach It has been developed usingthe concepts of object-oriented programming (OOP) Themodelling environment and the graphical user interface(GUI) are developed using JAVA Even though some ofthe simulator framework features are under development andseveral program module implementations are rudimentarythe overall software architecture of MPROSIM is illus-trated in Figure 1 The key idea that is being explored inMPROSIM is to make every activity of process engineeringa class or an object It is also designed to make all theimportant and essential features of process engineeringcompletely accessible from the GUI Hence the overallarchitecture of MPROSIM is at allowing the user tocarry out various process engineering activities with thesame model built using the GUI

Object-oriented modelling is a modern approach tohandle complexities The method involves modularization

and characterization of data into an abstraction called classA class is like a blueprint for an application or part of anapplication An object which is an instance or an incarna-tion of a class has both substance and behaviour bycomprising data and functions that perform operations onthis data A well abstracted class can be tested and imple-mented independently Abstraction helps in the reuse of aclass=object in various parts of the program Functionscommon to many classes can reside in one class and canbe implemented in other parts of the program with the helpof inheritance The common functionalities can be adaptedto meet speci c requirements with the help of polymorph-ism To protect parts of the program from unintentionalchanges or side effects variables and elements of an objectcan be encapsulated The ow of control in the program canbe threaded with the help of multithreading Multithreadingprovides a way for an application to handle many differenttasks at the same time (Niemeyer and Knudsen 2000) Eventhough JAVA supports only single inheritance it allowsmultiple implementation of interfaces All these featuresmake JAVA a conducive language to design the GUI and themodelling framework

The methodology of process modelling in MPROSIM hasbeen designed and developed based on the above conceptsof object-oriented programming and modelling The hier-archy of process modelling in MPROSIM as of present isdepicted in Figure 2 The rst step is a topological frag-mentation where a process is broken into model fragmentsor unit modules based on structure of the process Therepresentation of the original process is achieved by theaggregation of these model fragments with the help ofstreams in the form of a owsheet Each of the modelfragments resulting from the rst step of decomposition ismodular and well abstracted The model fragments can betested and implemented independent of other fragmentsthus ensuring their reuse At the second step a modelfragment or unit module is further decomposed into mole-cular fragments based on its internal structure and beha-viour A model fragment can be instantiated by a collectionof the molecular fragments

In the present development stage of MPROSIM thehierarchial modelling structure lies at the molecular fragmentlevel Further decomposition of the molecular fragmentsinto atomic fragments and issues relating to its testing andimplementation are some of the developments presentlyunder progress For the FCCU case study further

Figure 1 Software architecture of MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

530 R MADHUSUDANA RAO et al

decomposition of molecular fragments could not be under-taken due to limitations such as project schedule datacollection etc Though the modelling hierarchy builtcurrently is neither suitable nor extensible to model anygiven process system it is appropriate to address variousissues relating to the modelling of an industrial-scale FCCUIn addition some of the model fragments resulting from thedecomposition are generic in nature and can be extended tomodel other speci c units=processes as highlighted insection on Discussion and Clari cations below

For the implementation of the FCCU model a node classis de ned as the molecular fragment which contains all thetypical model equations such as mass balance energybalance etc Every unit module or model fragment inheritsthe node class and can modify or add additional informationto the inherited model equations to re ect its own behaviourThe FCCU model is realized in the form of a owsheet byan aggregation of unit modules by connecting them with the

help of streams A detailed description of the object-orientedmethodology for FCCU modelling in MPROSIM is given ina subsequent section As a precursor a brief review of theliterature on FCCU modelling is presented in the nextsection

Review of FCCU Modelling

Over the years many models have been proposed forFCCU These models have been based on different sets ofassumptions with respect to the kinetics of cracking reac-tions and hydrodynamics of the equipment involved such asriser and regenerator Some models concern themselves onlywith the regenerator (Ford et al 1976 Errazu et al 1979de Lasa et al 1981 Guigon and Large 1984 Krishna andParkin 1985 Lee et al 1989a) Some have only reactor orcracking models (Weekman and Nace 1970 Paraskos et al1976 Jacob et al 1976 Shah et al 1977 Lee et al

Figure 2 Hierarchy of process modelling in MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 531

1989b Larocca et al 1990 Takatsuka et al 1987) Therealso exist integrated models coupling both the regeneratorand reactor (Kumar et al 1995 Lee and Kugelman1973 McGreavy and Isles-Smith 1986 Bozicevic andLukec 1987 Arandes and de Lasa 1992 McFarlane et al1993 Arbel et al 1995 Arandes et al 2000) Table 1presents a list of literature which typify the approaches thathave been considered for FCCU modelling

The variety of approaches present in the literature can beanalysed with reference to Figures 3 and 4 The reactor riseris often modelled as being in plug ow (Arbel et al 1995)with uniform temperatures across a cross section and atemperature gradient along the height of the riser Otherassumptions such as quasi steady state no slip adiabaticoperation uniform temperature of the two phases at anypoint and constant heat capacities of oil vapour and catalystare usually made in the analysis On the other handMcFarlane et al (1993) assume isothermal conditions (ieCSTR type) in the riser though the bottom of the riserinvolves a non-isothermal zone due to a nite mixing time

Apart from this riser models differ mainly in their con-siderations of reaction kinetics Weekman and Nace (1970)developed a three lump model which was used by manyauthors Lee and Groves (1985) used the three lump model intheir integrated model Other models based on the Weekman-Nace three-lump model include those of McGreavy and Isles-Smith (1986) Kraemer and de Lasa (1988) Arandes andde Lasa (1992) Other authors expanded the three lumpsystem into more lumps Jacob et al (1976) developed aten lump system A four lump model was developed by Leeet al (1989b) Bozicevic and Lukec (1987) published a velump model Takatsuka et al (1987) developed a Weekmantype six lump model Kraemer et al (1990) extended theirmodel to eight lumps Two papers appeared in 1995 using tenlump kinetics with an integrated model Kumar et al (1995)used an isothermal reactor riser and Arbel et al (1995)modelled the riser assuming plug ow Arandes et al(2000) have considered the ten lump kinetics for dynamicand steady-state model of FCCU Plug ow conditions wereassumed in the riser and for gas ow in the regenerator

Among the models proposed for the regenerator mostfocus on the dense bed that is characterized by bubblesrising through an emulsion phase The earliest models were

single phase simple contacting models with plug owmixed ow dispersion and tanks in series The earlyresearchers Arthur (1951) Rowe and Partridge (1965) andWeisz and Goodwin (1966) simply divided the bed intoregions known as dense phase and dilute phase Later onmore complete models were developed the grid-effectmodel by Behie and Kehoe (1973) and Errazu et al(1979) the two-region model by de Lasa and Grace (1979)and de Lasa et al (1981) and the bubbling-bed model byKunii and Levenspiel (1968 1990)

A summary of the assumptions made by various research-ers for modelling the physics and chemistry of the processesin FCCU is illustrated in Figure 3 In the next section theFCCU case study is presented The case study describes indetail the process modelling methodology adopted for asystem like FCCU and its implementation in MPROSIMModel tuning and comparison of model predictions to theplant data are also presented The case study is concludedwith a discussion on some merits of developing FCCUmodel in an object-oriented framework like MPROSIM

CASE STUDY MODELLING OF FCCU

Process Description

The schematic of a current generation FCCU with stand-pipes and slide valves is represented in Figure 4 The reactorconsists of a vertical section called the riser The preheated

Table 1 A brief summary of literature on FCCU modelling

Author Year Model type Kinetic model

Weekman and Nace 1970 C 3-Lump modelJacob et al 1976 C 10-Lump modelFord et al 1976 R NAErrazu et al 1979 R NALee and Groves 1985 I 3-Lump modelKunii and Levenspiel 1990 R NAMcFarlane et al 1993 I NoneArbel et al 1995 I 10-Lump modelKumar et al 1995 I 10-Lump modelSriramulu et al 1996 R NAArandes et al 2000 I 10-Lump model

C Cracking model R Regenerator model I Integrated model

Figure 3 Summary of various assumptions for modelling FCCU

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

532 R MADHUSUDANA RAO et al

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 3: Industrial Experience with Object-Oriented Modelling FCC Case Study

well-designed and validated models is sometimes compro-mised (Marquardt 1996)

While both the above approaches have their advantagesand disadvantages the experience of early researchers inthese two approaches towards process modelling hastriggered considerable interest over the last decade Effortshave been directed towards addressing issues like modelformulation reusability of process models extensionsand adaptability to changing conditions As a result of theefforts of various researchers some of the developmentsin advanced process modelling tools and environmentsare ASCEND (Piela et al 1991) gPROMS (Barton andPantelides 1994) MODELLA (Stephanopoulos et al1990) and OMOLA (Nilsson 1993) One of the commonfeatures of all modelling tools is multi-level modularizationThe idea of modularization has been inherited from theconcepts of object-orientation which in turn has developedas a branch of computer science and engineering The mainidea of object-oriented design is to break a large andcomplex system into smaller fragments or modules so asto attain abstraction at each modular level The object-oriented approach directly supports reusability of commonfunctions in various parts of the program Further it alsoallows extensibility of the whole system by the implementa-tion of modules The reader is referred to the originalliterature cited above for a detailed summary on the devel-opment of each modelling tool State-of-the-art reviews ondevelopments in process modelling are also available(Marquardt 1991 1996 Biegler 1989 Boston et al1993 Pantelides and Barton 1993)

Some of the recent advances in the area of processmodelling have been towards a systematic approach to thedevelopment of process models (process modelling metho-dology) and formal representation of process model equa-tions (Marquardt 1994 1996 Bogusch and Marquardt1995 1997 Lohmann and Marquardt 1996) Marquardtand his group have laid the foundations for systematicprocess model development and at the same time haveproposed an object-oriented methodology for the task ofcomputer-aided process modelling Extending the ideas ofsystem theoretic approach and object-orientation a processunit model can be decomposed into smaller modules basedon its structure (for example reactor wall catalyst heatmaterial etc for a tubular reactor) and behaviour (conser-vation constitutive equations etc) Similarly the behaviouralobjects can further be decomposed into more fundamentalentities (for example holdup transport transfer etc for aspecies conservation equation) Following this methodologyMarquardt and his group have proposed a structure forhierarchy of elementary process quantities for developinga unit model A real unit model is de ned by aggregating aset of these canonical=fundamental modelling objects

The main aim of adopting such a modelling methodologyis to address the following two issues (1) at the fundamentallevel any chemical process be it petroleum petrochemicalmetallurgical or biological is the same ie conservationof species components or particles and conservation=conversion of energy Hence with a well-structured andgeneric modelling tool it should be possible (ideally) tode ne any model by aggregating the elementary modellingobjects and (2) to be able to develop a unit model from afundamental approach and also reuse and extend the exist-ing library of models to t the changing requirements

However in spite of all the developments in the area ofobject-oriented structured process modelling over the lastdecade there still remain several important issues that needto be addressed They can be classi ed into two categories

Generic issues Some of the issues aremdash the amount of effort and expert knowledge that would

be required in various elds of science and engineer-ing for developing such a fundamental and genericmodelling tool

mdash the capabilities and conduciveness of the existingmodelling and software tools for encoding and repre-senting the vast amount of diversi ed scienti c andengineering knowledge in a structured format

mdash the extent of granularity that one should build tomodel any given process system One of the practicalissues in this context is that a given hierarchy ofprocess model structuring however ne it may be inits granularity might still be not complete Forexample in mineral processing the liberation statein gravel and sand industry processes the fractionaldensity and in environmental processes like soil-washing the fractional contamination must be con-sidered along with the particle size distribution tocharacterize a process stream (Toebermann et al2000) Similarly processes from different elds ofengineering and technology may require additionalinput and parameter speci cations for both processstreams and process models

mdash uniqueness of a decomposition strategy A chemicalprocess can be represented by an aggregation ofseveral possible combinations of fundamentalmodelling objects of varying degrees of complexi-ties Thus the decomposition strategy for de ningthe model objects gives rise to many degrees offreedom The problem of selection of a particularstrategy for process representation can be adif cult task when a large scale industrial processis considered

mdash extension and adaptability of a set of fundamentalmodelling objects A generic process modellingframework must be capable and exible to be ableto extend and incorporate additional information=knowledge

Implementation issuesmdash Though some general guidelines for structured

process modelling have been stated (Marquardt1992 1996) the utility of the formalism can onlybe shown by the implementation and evaluation ofmodelling tools along with experimentation

While there is considerable literature on structuredprocess modelling from a subjective or theoretical view-point very little has been published on application ofthese ideas to complex real-life modelling problems Tothe authorsrsquo best knowledge there has been no directevaluation of the formalism or the guidelines for struc-tured process modelling through an implementation on alarge scale industrial process This is the main focus ofthe paper and FCCU is used as a case study The projectwas undertaken in joint collaboration with a re neryChennai Petroleum Corporation Limited (CPCL) IndiaThough there are few commercially available simulation

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 529

tools for FCCU like FCC Plus it is necessary to reiteratethat the concepts of object-oriented modelling and resultspresented in this contribution are not to undertake anevaluation of the simulation tools The modelling andsimulation of FCCU is chosen to demonstrate the meritsof MPROSIM and at the same time present an evaluationof structured object-oriented process modelling guidelinesThe reasons for the choice of FCCU are manifold

FCCU is a critical unit of a re nery and is thus importantin its own rightFCCU is a large system with complexities at severallevels heterogeneous operations solids handling hydro-dynamic complexities in the riser and uid-bed regen-erators complex kinetics resulting from the complexity offeedstock topological complexities due to recycles andinterconnected ows etc It is therefore a good test caseto test the implementation and evaluation of the ideas ofmodularization and structural decompositionVarious con gurations of reactor and regenerator are inoperation in the re neries Adaptability of the modellingframework to different con gurations can therefore betestedInputs to the unit often change in terms of both quality(composition of feed) and quantity (product demands inthe market) Often a new or improved catalyst becomesavailable This is precisely the situation where a simula-tion tool can prove its utility since ne tuning of theoperating conditions in response to such changes is oftenneededVarious assumptions have been made in literature formodelling the physics and chemistry of the crackingprocess As a result several approaches are available formodelling different sections of FCCU There is thus anopportunity to test the exibility of the framework tomodel the system using different approaches

In the next section we brie y describe the object-orientedmodelling framework developed in MPROSIM The frame-work is based on the principles of systematic processmodelling as enumerated in many of the works discussedabove

Modell ing in MPROSIM

MPROSIM is a framework for multipurpose processmodelling and simulation The simulator is based on theequation-oriented approach It has been developed usingthe concepts of object-oriented programming (OOP) Themodelling environment and the graphical user interface(GUI) are developed using JAVA Even though some ofthe simulator framework features are under development andseveral program module implementations are rudimentarythe overall software architecture of MPROSIM is illus-trated in Figure 1 The key idea that is being explored inMPROSIM is to make every activity of process engineeringa class or an object It is also designed to make all theimportant and essential features of process engineeringcompletely accessible from the GUI Hence the overallarchitecture of MPROSIM is at allowing the user tocarry out various process engineering activities with thesame model built using the GUI

Object-oriented modelling is a modern approach tohandle complexities The method involves modularization

and characterization of data into an abstraction called classA class is like a blueprint for an application or part of anapplication An object which is an instance or an incarna-tion of a class has both substance and behaviour bycomprising data and functions that perform operations onthis data A well abstracted class can be tested and imple-mented independently Abstraction helps in the reuse of aclass=object in various parts of the program Functionscommon to many classes can reside in one class and canbe implemented in other parts of the program with the helpof inheritance The common functionalities can be adaptedto meet speci c requirements with the help of polymorph-ism To protect parts of the program from unintentionalchanges or side effects variables and elements of an objectcan be encapsulated The ow of control in the program canbe threaded with the help of multithreading Multithreadingprovides a way for an application to handle many differenttasks at the same time (Niemeyer and Knudsen 2000) Eventhough JAVA supports only single inheritance it allowsmultiple implementation of interfaces All these featuresmake JAVA a conducive language to design the GUI and themodelling framework

The methodology of process modelling in MPROSIM hasbeen designed and developed based on the above conceptsof object-oriented programming and modelling The hier-archy of process modelling in MPROSIM as of present isdepicted in Figure 2 The rst step is a topological frag-mentation where a process is broken into model fragmentsor unit modules based on structure of the process Therepresentation of the original process is achieved by theaggregation of these model fragments with the help ofstreams in the form of a owsheet Each of the modelfragments resulting from the rst step of decomposition ismodular and well abstracted The model fragments can betested and implemented independent of other fragmentsthus ensuring their reuse At the second step a modelfragment or unit module is further decomposed into mole-cular fragments based on its internal structure and beha-viour A model fragment can be instantiated by a collectionof the molecular fragments

In the present development stage of MPROSIM thehierarchial modelling structure lies at the molecular fragmentlevel Further decomposition of the molecular fragmentsinto atomic fragments and issues relating to its testing andimplementation are some of the developments presentlyunder progress For the FCCU case study further

Figure 1 Software architecture of MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

530 R MADHUSUDANA RAO et al

decomposition of molecular fragments could not be under-taken due to limitations such as project schedule datacollection etc Though the modelling hierarchy builtcurrently is neither suitable nor extensible to model anygiven process system it is appropriate to address variousissues relating to the modelling of an industrial-scale FCCUIn addition some of the model fragments resulting from thedecomposition are generic in nature and can be extended tomodel other speci c units=processes as highlighted insection on Discussion and Clari cations below

For the implementation of the FCCU model a node classis de ned as the molecular fragment which contains all thetypical model equations such as mass balance energybalance etc Every unit module or model fragment inheritsthe node class and can modify or add additional informationto the inherited model equations to re ect its own behaviourThe FCCU model is realized in the form of a owsheet byan aggregation of unit modules by connecting them with the

help of streams A detailed description of the object-orientedmethodology for FCCU modelling in MPROSIM is given ina subsequent section As a precursor a brief review of theliterature on FCCU modelling is presented in the nextsection

Review of FCCU Modelling

Over the years many models have been proposed forFCCU These models have been based on different sets ofassumptions with respect to the kinetics of cracking reac-tions and hydrodynamics of the equipment involved such asriser and regenerator Some models concern themselves onlywith the regenerator (Ford et al 1976 Errazu et al 1979de Lasa et al 1981 Guigon and Large 1984 Krishna andParkin 1985 Lee et al 1989a) Some have only reactor orcracking models (Weekman and Nace 1970 Paraskos et al1976 Jacob et al 1976 Shah et al 1977 Lee et al

Figure 2 Hierarchy of process modelling in MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 531

1989b Larocca et al 1990 Takatsuka et al 1987) Therealso exist integrated models coupling both the regeneratorand reactor (Kumar et al 1995 Lee and Kugelman1973 McGreavy and Isles-Smith 1986 Bozicevic andLukec 1987 Arandes and de Lasa 1992 McFarlane et al1993 Arbel et al 1995 Arandes et al 2000) Table 1presents a list of literature which typify the approaches thathave been considered for FCCU modelling

The variety of approaches present in the literature can beanalysed with reference to Figures 3 and 4 The reactor riseris often modelled as being in plug ow (Arbel et al 1995)with uniform temperatures across a cross section and atemperature gradient along the height of the riser Otherassumptions such as quasi steady state no slip adiabaticoperation uniform temperature of the two phases at anypoint and constant heat capacities of oil vapour and catalystare usually made in the analysis On the other handMcFarlane et al (1993) assume isothermal conditions (ieCSTR type) in the riser though the bottom of the riserinvolves a non-isothermal zone due to a nite mixing time

Apart from this riser models differ mainly in their con-siderations of reaction kinetics Weekman and Nace (1970)developed a three lump model which was used by manyauthors Lee and Groves (1985) used the three lump model intheir integrated model Other models based on the Weekman-Nace three-lump model include those of McGreavy and Isles-Smith (1986) Kraemer and de Lasa (1988) Arandes andde Lasa (1992) Other authors expanded the three lumpsystem into more lumps Jacob et al (1976) developed aten lump system A four lump model was developed by Leeet al (1989b) Bozicevic and Lukec (1987) published a velump model Takatsuka et al (1987) developed a Weekmantype six lump model Kraemer et al (1990) extended theirmodel to eight lumps Two papers appeared in 1995 using tenlump kinetics with an integrated model Kumar et al (1995)used an isothermal reactor riser and Arbel et al (1995)modelled the riser assuming plug ow Arandes et al(2000) have considered the ten lump kinetics for dynamicand steady-state model of FCCU Plug ow conditions wereassumed in the riser and for gas ow in the regenerator

Among the models proposed for the regenerator mostfocus on the dense bed that is characterized by bubblesrising through an emulsion phase The earliest models were

single phase simple contacting models with plug owmixed ow dispersion and tanks in series The earlyresearchers Arthur (1951) Rowe and Partridge (1965) andWeisz and Goodwin (1966) simply divided the bed intoregions known as dense phase and dilute phase Later onmore complete models were developed the grid-effectmodel by Behie and Kehoe (1973) and Errazu et al(1979) the two-region model by de Lasa and Grace (1979)and de Lasa et al (1981) and the bubbling-bed model byKunii and Levenspiel (1968 1990)

A summary of the assumptions made by various research-ers for modelling the physics and chemistry of the processesin FCCU is illustrated in Figure 3 In the next section theFCCU case study is presented The case study describes indetail the process modelling methodology adopted for asystem like FCCU and its implementation in MPROSIMModel tuning and comparison of model predictions to theplant data are also presented The case study is concludedwith a discussion on some merits of developing FCCUmodel in an object-oriented framework like MPROSIM

CASE STUDY MODELLING OF FCCU

Process Description

The schematic of a current generation FCCU with stand-pipes and slide valves is represented in Figure 4 The reactorconsists of a vertical section called the riser The preheated

Table 1 A brief summary of literature on FCCU modelling

Author Year Model type Kinetic model

Weekman and Nace 1970 C 3-Lump modelJacob et al 1976 C 10-Lump modelFord et al 1976 R NAErrazu et al 1979 R NALee and Groves 1985 I 3-Lump modelKunii and Levenspiel 1990 R NAMcFarlane et al 1993 I NoneArbel et al 1995 I 10-Lump modelKumar et al 1995 I 10-Lump modelSriramulu et al 1996 R NAArandes et al 2000 I 10-Lump model

C Cracking model R Regenerator model I Integrated model

Figure 3 Summary of various assumptions for modelling FCCU

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

532 R MADHUSUDANA RAO et al

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 4: Industrial Experience with Object-Oriented Modelling FCC Case Study

tools for FCCU like FCC Plus it is necessary to reiteratethat the concepts of object-oriented modelling and resultspresented in this contribution are not to undertake anevaluation of the simulation tools The modelling andsimulation of FCCU is chosen to demonstrate the meritsof MPROSIM and at the same time present an evaluationof structured object-oriented process modelling guidelinesThe reasons for the choice of FCCU are manifold

FCCU is a critical unit of a re nery and is thus importantin its own rightFCCU is a large system with complexities at severallevels heterogeneous operations solids handling hydro-dynamic complexities in the riser and uid-bed regen-erators complex kinetics resulting from the complexity offeedstock topological complexities due to recycles andinterconnected ows etc It is therefore a good test caseto test the implementation and evaluation of the ideas ofmodularization and structural decompositionVarious con gurations of reactor and regenerator are inoperation in the re neries Adaptability of the modellingframework to different con gurations can therefore betestedInputs to the unit often change in terms of both quality(composition of feed) and quantity (product demands inthe market) Often a new or improved catalyst becomesavailable This is precisely the situation where a simula-tion tool can prove its utility since ne tuning of theoperating conditions in response to such changes is oftenneededVarious assumptions have been made in literature formodelling the physics and chemistry of the crackingprocess As a result several approaches are available formodelling different sections of FCCU There is thus anopportunity to test the exibility of the framework tomodel the system using different approaches

In the next section we brie y describe the object-orientedmodelling framework developed in MPROSIM The frame-work is based on the principles of systematic processmodelling as enumerated in many of the works discussedabove

Modell ing in MPROSIM

MPROSIM is a framework for multipurpose processmodelling and simulation The simulator is based on theequation-oriented approach It has been developed usingthe concepts of object-oriented programming (OOP) Themodelling environment and the graphical user interface(GUI) are developed using JAVA Even though some ofthe simulator framework features are under development andseveral program module implementations are rudimentarythe overall software architecture of MPROSIM is illus-trated in Figure 1 The key idea that is being explored inMPROSIM is to make every activity of process engineeringa class or an object It is also designed to make all theimportant and essential features of process engineeringcompletely accessible from the GUI Hence the overallarchitecture of MPROSIM is at allowing the user tocarry out various process engineering activities with thesame model built using the GUI

Object-oriented modelling is a modern approach tohandle complexities The method involves modularization

and characterization of data into an abstraction called classA class is like a blueprint for an application or part of anapplication An object which is an instance or an incarna-tion of a class has both substance and behaviour bycomprising data and functions that perform operations onthis data A well abstracted class can be tested and imple-mented independently Abstraction helps in the reuse of aclass=object in various parts of the program Functionscommon to many classes can reside in one class and canbe implemented in other parts of the program with the helpof inheritance The common functionalities can be adaptedto meet speci c requirements with the help of polymorph-ism To protect parts of the program from unintentionalchanges or side effects variables and elements of an objectcan be encapsulated The ow of control in the program canbe threaded with the help of multithreading Multithreadingprovides a way for an application to handle many differenttasks at the same time (Niemeyer and Knudsen 2000) Eventhough JAVA supports only single inheritance it allowsmultiple implementation of interfaces All these featuresmake JAVA a conducive language to design the GUI and themodelling framework

The methodology of process modelling in MPROSIM hasbeen designed and developed based on the above conceptsof object-oriented programming and modelling The hier-archy of process modelling in MPROSIM as of present isdepicted in Figure 2 The rst step is a topological frag-mentation where a process is broken into model fragmentsor unit modules based on structure of the process Therepresentation of the original process is achieved by theaggregation of these model fragments with the help ofstreams in the form of a owsheet Each of the modelfragments resulting from the rst step of decomposition ismodular and well abstracted The model fragments can betested and implemented independent of other fragmentsthus ensuring their reuse At the second step a modelfragment or unit module is further decomposed into mole-cular fragments based on its internal structure and beha-viour A model fragment can be instantiated by a collectionof the molecular fragments

In the present development stage of MPROSIM thehierarchial modelling structure lies at the molecular fragmentlevel Further decomposition of the molecular fragmentsinto atomic fragments and issues relating to its testing andimplementation are some of the developments presentlyunder progress For the FCCU case study further

Figure 1 Software architecture of MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

530 R MADHUSUDANA RAO et al

decomposition of molecular fragments could not be under-taken due to limitations such as project schedule datacollection etc Though the modelling hierarchy builtcurrently is neither suitable nor extensible to model anygiven process system it is appropriate to address variousissues relating to the modelling of an industrial-scale FCCUIn addition some of the model fragments resulting from thedecomposition are generic in nature and can be extended tomodel other speci c units=processes as highlighted insection on Discussion and Clari cations below

For the implementation of the FCCU model a node classis de ned as the molecular fragment which contains all thetypical model equations such as mass balance energybalance etc Every unit module or model fragment inheritsthe node class and can modify or add additional informationto the inherited model equations to re ect its own behaviourThe FCCU model is realized in the form of a owsheet byan aggregation of unit modules by connecting them with the

help of streams A detailed description of the object-orientedmethodology for FCCU modelling in MPROSIM is given ina subsequent section As a precursor a brief review of theliterature on FCCU modelling is presented in the nextsection

Review of FCCU Modelling

Over the years many models have been proposed forFCCU These models have been based on different sets ofassumptions with respect to the kinetics of cracking reac-tions and hydrodynamics of the equipment involved such asriser and regenerator Some models concern themselves onlywith the regenerator (Ford et al 1976 Errazu et al 1979de Lasa et al 1981 Guigon and Large 1984 Krishna andParkin 1985 Lee et al 1989a) Some have only reactor orcracking models (Weekman and Nace 1970 Paraskos et al1976 Jacob et al 1976 Shah et al 1977 Lee et al

Figure 2 Hierarchy of process modelling in MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 531

1989b Larocca et al 1990 Takatsuka et al 1987) Therealso exist integrated models coupling both the regeneratorand reactor (Kumar et al 1995 Lee and Kugelman1973 McGreavy and Isles-Smith 1986 Bozicevic andLukec 1987 Arandes and de Lasa 1992 McFarlane et al1993 Arbel et al 1995 Arandes et al 2000) Table 1presents a list of literature which typify the approaches thathave been considered for FCCU modelling

The variety of approaches present in the literature can beanalysed with reference to Figures 3 and 4 The reactor riseris often modelled as being in plug ow (Arbel et al 1995)with uniform temperatures across a cross section and atemperature gradient along the height of the riser Otherassumptions such as quasi steady state no slip adiabaticoperation uniform temperature of the two phases at anypoint and constant heat capacities of oil vapour and catalystare usually made in the analysis On the other handMcFarlane et al (1993) assume isothermal conditions (ieCSTR type) in the riser though the bottom of the riserinvolves a non-isothermal zone due to a nite mixing time

Apart from this riser models differ mainly in their con-siderations of reaction kinetics Weekman and Nace (1970)developed a three lump model which was used by manyauthors Lee and Groves (1985) used the three lump model intheir integrated model Other models based on the Weekman-Nace three-lump model include those of McGreavy and Isles-Smith (1986) Kraemer and de Lasa (1988) Arandes andde Lasa (1992) Other authors expanded the three lumpsystem into more lumps Jacob et al (1976) developed aten lump system A four lump model was developed by Leeet al (1989b) Bozicevic and Lukec (1987) published a velump model Takatsuka et al (1987) developed a Weekmantype six lump model Kraemer et al (1990) extended theirmodel to eight lumps Two papers appeared in 1995 using tenlump kinetics with an integrated model Kumar et al (1995)used an isothermal reactor riser and Arbel et al (1995)modelled the riser assuming plug ow Arandes et al(2000) have considered the ten lump kinetics for dynamicand steady-state model of FCCU Plug ow conditions wereassumed in the riser and for gas ow in the regenerator

Among the models proposed for the regenerator mostfocus on the dense bed that is characterized by bubblesrising through an emulsion phase The earliest models were

single phase simple contacting models with plug owmixed ow dispersion and tanks in series The earlyresearchers Arthur (1951) Rowe and Partridge (1965) andWeisz and Goodwin (1966) simply divided the bed intoregions known as dense phase and dilute phase Later onmore complete models were developed the grid-effectmodel by Behie and Kehoe (1973) and Errazu et al(1979) the two-region model by de Lasa and Grace (1979)and de Lasa et al (1981) and the bubbling-bed model byKunii and Levenspiel (1968 1990)

A summary of the assumptions made by various research-ers for modelling the physics and chemistry of the processesin FCCU is illustrated in Figure 3 In the next section theFCCU case study is presented The case study describes indetail the process modelling methodology adopted for asystem like FCCU and its implementation in MPROSIMModel tuning and comparison of model predictions to theplant data are also presented The case study is concludedwith a discussion on some merits of developing FCCUmodel in an object-oriented framework like MPROSIM

CASE STUDY MODELLING OF FCCU

Process Description

The schematic of a current generation FCCU with stand-pipes and slide valves is represented in Figure 4 The reactorconsists of a vertical section called the riser The preheated

Table 1 A brief summary of literature on FCCU modelling

Author Year Model type Kinetic model

Weekman and Nace 1970 C 3-Lump modelJacob et al 1976 C 10-Lump modelFord et al 1976 R NAErrazu et al 1979 R NALee and Groves 1985 I 3-Lump modelKunii and Levenspiel 1990 R NAMcFarlane et al 1993 I NoneArbel et al 1995 I 10-Lump modelKumar et al 1995 I 10-Lump modelSriramulu et al 1996 R NAArandes et al 2000 I 10-Lump model

C Cracking model R Regenerator model I Integrated model

Figure 3 Summary of various assumptions for modelling FCCU

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

532 R MADHUSUDANA RAO et al

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 5: Industrial Experience with Object-Oriented Modelling FCC Case Study

decomposition of molecular fragments could not be under-taken due to limitations such as project schedule datacollection etc Though the modelling hierarchy builtcurrently is neither suitable nor extensible to model anygiven process system it is appropriate to address variousissues relating to the modelling of an industrial-scale FCCUIn addition some of the model fragments resulting from thedecomposition are generic in nature and can be extended tomodel other speci c units=processes as highlighted insection on Discussion and Clari cations below

For the implementation of the FCCU model a node classis de ned as the molecular fragment which contains all thetypical model equations such as mass balance energybalance etc Every unit module or model fragment inheritsthe node class and can modify or add additional informationto the inherited model equations to re ect its own behaviourThe FCCU model is realized in the form of a owsheet byan aggregation of unit modules by connecting them with the

help of streams A detailed description of the object-orientedmethodology for FCCU modelling in MPROSIM is given ina subsequent section As a precursor a brief review of theliterature on FCCU modelling is presented in the nextsection

Review of FCCU Modelling

Over the years many models have been proposed forFCCU These models have been based on different sets ofassumptions with respect to the kinetics of cracking reac-tions and hydrodynamics of the equipment involved such asriser and regenerator Some models concern themselves onlywith the regenerator (Ford et al 1976 Errazu et al 1979de Lasa et al 1981 Guigon and Large 1984 Krishna andParkin 1985 Lee et al 1989a) Some have only reactor orcracking models (Weekman and Nace 1970 Paraskos et al1976 Jacob et al 1976 Shah et al 1977 Lee et al

Figure 2 Hierarchy of process modelling in MPROSIM

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 531

1989b Larocca et al 1990 Takatsuka et al 1987) Therealso exist integrated models coupling both the regeneratorand reactor (Kumar et al 1995 Lee and Kugelman1973 McGreavy and Isles-Smith 1986 Bozicevic andLukec 1987 Arandes and de Lasa 1992 McFarlane et al1993 Arbel et al 1995 Arandes et al 2000) Table 1presents a list of literature which typify the approaches thathave been considered for FCCU modelling

The variety of approaches present in the literature can beanalysed with reference to Figures 3 and 4 The reactor riseris often modelled as being in plug ow (Arbel et al 1995)with uniform temperatures across a cross section and atemperature gradient along the height of the riser Otherassumptions such as quasi steady state no slip adiabaticoperation uniform temperature of the two phases at anypoint and constant heat capacities of oil vapour and catalystare usually made in the analysis On the other handMcFarlane et al (1993) assume isothermal conditions (ieCSTR type) in the riser though the bottom of the riserinvolves a non-isothermal zone due to a nite mixing time

Apart from this riser models differ mainly in their con-siderations of reaction kinetics Weekman and Nace (1970)developed a three lump model which was used by manyauthors Lee and Groves (1985) used the three lump model intheir integrated model Other models based on the Weekman-Nace three-lump model include those of McGreavy and Isles-Smith (1986) Kraemer and de Lasa (1988) Arandes andde Lasa (1992) Other authors expanded the three lumpsystem into more lumps Jacob et al (1976) developed aten lump system A four lump model was developed by Leeet al (1989b) Bozicevic and Lukec (1987) published a velump model Takatsuka et al (1987) developed a Weekmantype six lump model Kraemer et al (1990) extended theirmodel to eight lumps Two papers appeared in 1995 using tenlump kinetics with an integrated model Kumar et al (1995)used an isothermal reactor riser and Arbel et al (1995)modelled the riser assuming plug ow Arandes et al(2000) have considered the ten lump kinetics for dynamicand steady-state model of FCCU Plug ow conditions wereassumed in the riser and for gas ow in the regenerator

Among the models proposed for the regenerator mostfocus on the dense bed that is characterized by bubblesrising through an emulsion phase The earliest models were

single phase simple contacting models with plug owmixed ow dispersion and tanks in series The earlyresearchers Arthur (1951) Rowe and Partridge (1965) andWeisz and Goodwin (1966) simply divided the bed intoregions known as dense phase and dilute phase Later onmore complete models were developed the grid-effectmodel by Behie and Kehoe (1973) and Errazu et al(1979) the two-region model by de Lasa and Grace (1979)and de Lasa et al (1981) and the bubbling-bed model byKunii and Levenspiel (1968 1990)

A summary of the assumptions made by various research-ers for modelling the physics and chemistry of the processesin FCCU is illustrated in Figure 3 In the next section theFCCU case study is presented The case study describes indetail the process modelling methodology adopted for asystem like FCCU and its implementation in MPROSIMModel tuning and comparison of model predictions to theplant data are also presented The case study is concludedwith a discussion on some merits of developing FCCUmodel in an object-oriented framework like MPROSIM

CASE STUDY MODELLING OF FCCU

Process Description

The schematic of a current generation FCCU with stand-pipes and slide valves is represented in Figure 4 The reactorconsists of a vertical section called the riser The preheated

Table 1 A brief summary of literature on FCCU modelling

Author Year Model type Kinetic model

Weekman and Nace 1970 C 3-Lump modelJacob et al 1976 C 10-Lump modelFord et al 1976 R NAErrazu et al 1979 R NALee and Groves 1985 I 3-Lump modelKunii and Levenspiel 1990 R NAMcFarlane et al 1993 I NoneArbel et al 1995 I 10-Lump modelKumar et al 1995 I 10-Lump modelSriramulu et al 1996 R NAArandes et al 2000 I 10-Lump model

C Cracking model R Regenerator model I Integrated model

Figure 3 Summary of various assumptions for modelling FCCU

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

532 R MADHUSUDANA RAO et al

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 6: Industrial Experience with Object-Oriented Modelling FCC Case Study

1989b Larocca et al 1990 Takatsuka et al 1987) Therealso exist integrated models coupling both the regeneratorand reactor (Kumar et al 1995 Lee and Kugelman1973 McGreavy and Isles-Smith 1986 Bozicevic andLukec 1987 Arandes and de Lasa 1992 McFarlane et al1993 Arbel et al 1995 Arandes et al 2000) Table 1presents a list of literature which typify the approaches thathave been considered for FCCU modelling

The variety of approaches present in the literature can beanalysed with reference to Figures 3 and 4 The reactor riseris often modelled as being in plug ow (Arbel et al 1995)with uniform temperatures across a cross section and atemperature gradient along the height of the riser Otherassumptions such as quasi steady state no slip adiabaticoperation uniform temperature of the two phases at anypoint and constant heat capacities of oil vapour and catalystare usually made in the analysis On the other handMcFarlane et al (1993) assume isothermal conditions (ieCSTR type) in the riser though the bottom of the riserinvolves a non-isothermal zone due to a nite mixing time

Apart from this riser models differ mainly in their con-siderations of reaction kinetics Weekman and Nace (1970)developed a three lump model which was used by manyauthors Lee and Groves (1985) used the three lump model intheir integrated model Other models based on the Weekman-Nace three-lump model include those of McGreavy and Isles-Smith (1986) Kraemer and de Lasa (1988) Arandes andde Lasa (1992) Other authors expanded the three lumpsystem into more lumps Jacob et al (1976) developed aten lump system A four lump model was developed by Leeet al (1989b) Bozicevic and Lukec (1987) published a velump model Takatsuka et al (1987) developed a Weekmantype six lump model Kraemer et al (1990) extended theirmodel to eight lumps Two papers appeared in 1995 using tenlump kinetics with an integrated model Kumar et al (1995)used an isothermal reactor riser and Arbel et al (1995)modelled the riser assuming plug ow Arandes et al(2000) have considered the ten lump kinetics for dynamicand steady-state model of FCCU Plug ow conditions wereassumed in the riser and for gas ow in the regenerator

Among the models proposed for the regenerator mostfocus on the dense bed that is characterized by bubblesrising through an emulsion phase The earliest models were

single phase simple contacting models with plug owmixed ow dispersion and tanks in series The earlyresearchers Arthur (1951) Rowe and Partridge (1965) andWeisz and Goodwin (1966) simply divided the bed intoregions known as dense phase and dilute phase Later onmore complete models were developed the grid-effectmodel by Behie and Kehoe (1973) and Errazu et al(1979) the two-region model by de Lasa and Grace (1979)and de Lasa et al (1981) and the bubbling-bed model byKunii and Levenspiel (1968 1990)

A summary of the assumptions made by various research-ers for modelling the physics and chemistry of the processesin FCCU is illustrated in Figure 3 In the next section theFCCU case study is presented The case study describes indetail the process modelling methodology adopted for asystem like FCCU and its implementation in MPROSIMModel tuning and comparison of model predictions to theplant data are also presented The case study is concludedwith a discussion on some merits of developing FCCUmodel in an object-oriented framework like MPROSIM

CASE STUDY MODELLING OF FCCU

Process Description

The schematic of a current generation FCCU with stand-pipes and slide valves is represented in Figure 4 The reactorconsists of a vertical section called the riser The preheated

Table 1 A brief summary of literature on FCCU modelling

Author Year Model type Kinetic model

Weekman and Nace 1970 C 3-Lump modelJacob et al 1976 C 10-Lump modelFord et al 1976 R NAErrazu et al 1979 R NALee and Groves 1985 I 3-Lump modelKunii and Levenspiel 1990 R NAMcFarlane et al 1993 I NoneArbel et al 1995 I 10-Lump modelKumar et al 1995 I 10-Lump modelSriramulu et al 1996 R NAArandes et al 2000 I 10-Lump model

C Cracking model R Regenerator model I Integrated model

Figure 3 Summary of various assumptions for modelling FCCU

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

532 R MADHUSUDANA RAO et al

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 7: Industrial Experience with Object-Oriented Modelling FCC Case Study

feed is brought in contact with hot regenerated catalyst at thebottom of the riser Feed ashes at the bottom of the riser andis vapourized As the catalystndashvapour mixture rises in nearplug ow cracking reactions take place within the riser Cokeis deposited on the catalyst surface during cracking reactionsHeat required for endothermic cracking reactions is suppliedby the hot catalyst The bulk of the catalyst is separated fromproduct vapour in the riser termination device (RTD) at thetop of the riser and falls into the stripper section Catalyst nes that are entrained along with the product vapour areseparated in the reactor cyclones and returned to the stripperThe product vapour is fed to the main fractionator whereit is separated into various components such as light gases(C1ndashC4) liqui ed petroleum gas (LPG) gasoline cycle oil inthe diesel boiling range and heavy bottoms

In the stripper section steam is used to strip off hydro-carbons trapped within the bulk of the catalyst and catalystpores Stripped catalyst is regenerated in the regeneratorsection by burning coke in a uidized bed using hot airCoke combustion reactions occurring in the dense bedproduce CO CO2 and H2O Combustion reactions occurfurther in the dilute phase due to entrainment of catalystparticles from the dense bed by the ow of air The entrainedcatalyst is separated from stack gases in regenerator cyclones

and returned to the dense bed The heat due to combustionreactions raises the temperature of the regenerated catalystwhich is recycled to the reactor through the stand pipe Asshown in Figure 4 spent and regenerated catalyst circulationis controlled by slide valves in most of the modern units

Kinetic Lumping Scheme

The reactions that occur in the riser when hot regeneratedcatalyst comes into contact with the feed are described byconsidering a lumping scheme The lumps on the feed sideare characterized based on the boiling point range of gas oilfeed and its chemical composition mainly in terms ofparaf ns naphthenes and aromatics The lumps on theproduct side are characterized by the boiling point rangeof main fractionator side draws The various lumps consid-ered for building the reaction kinetics are listed in Table 2The lumps on the feed and product side are chosen based onthe theoretical basis provided by Jacob et al (1976) and theexperimental support from CPCL for feed and productcharacterization

The network of reactions using the 11 lumps given inTable 2 are shown in Figure 5 along with their boiling pointranges The salient points of the reaction network are

Figure 4 Schematic of a FCCU reactor=regenerator system

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 533

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 8: Industrial Experience with Object-Oriented Modelling FCC Case Study

The feed is mainly characterized as consisting of heavyand light fractions based on the boiling point range370 CndashFBP and IBPndash370 C respectively Each of theheavy and light fractions are further divided into fourlumpsmdashparaf ns naphthenes aromatics and side chainsThe product consists of three lumpsmdashgasoline (C5ndashIBP)light gases (C1ndashC4) and cokeIt is assumed that there are no cross reactions betweenheavy and light fractions For example heavy naphthenes(Nh) crack to give only light naphthenes (Nl) and not lightparaf ns (Pl) One exception to this assumption is thecracking of heavy aromatic substituent groups (Sh) goingto light aromatic bare rings (Al)Bare aromatic rings are mainly responsible for cokeformation Hence (Ah) and (Al) are shown to formonly cokeFigure 5 shows the reactions mainly from Ph and Pl andother typical reactions Apart from the reactions shown inthe gure naphthenes (Nh and Nl) and substituent groupson aromatics (Sh and Sl) will have reactions similar to thatindicated for paraf ns (Ph and Pl)

Based on the above set of assumptions the total numberof cracking reactions in the network is 27

As described in the review on FCCU modelling literatureresearchers have considered modelling the reaction kineticsin the riser based on various assumptions and parametersie number of lumps the feed is characterized into ratelaws governing the reaction kinetics kinetic parameters andvarious empirical catalyst deactivation functions Some ofthe recent literature proposes the use of structure-orientedlumps for modelling the reaction kinetics of complex feed-stocks (Quann and Jaffe 1992 1996) In the event of somany factors in uencing reaction kinetics in the riser it hasbeen modelled as a class=object with all the in uencingfactors featuring as attributes of reaction class Thus themodel is not restricted to a single lumping strategy The usercan choose any number of lumps build a reaction networkbetween various lumps input the various kinetic parametersde ne any type of rate laws governing the formation=depletion of various lumps and also include different typesof catalyst deactivation functions Modelling the reactionkinetics as a class=object does not restrict the model toa particular lumping strategy Moreover it imparts theimportant characteristicmdashease of extension and adaptabilityto model other complicated reaction mechanisms which isone of the desired essential features of a structured processmodelling

Object-Oriented Modell ing of FCCU

The object-oriented model for FCCU is developedfollowing the object-oriented and process modelling metho-dology described in section on Modelling in MPROSIMabove Catalytic reactions in the riser regenerator ui-dized bed hydrodynamics and coke combustion areconsidered in the model Detailed momentum calculationsat various sections of the unit have also been taken into

Figure 5 11-lump reaction kinetic scheme

Table 2 Kinetic lumping scheme feed and product lumps

Feed lumps BP range (macrC) Product BP range (macrC)

Ph 370ndashFBP Gasoline C5ndashIBPNh 370ndashFBP Light gases C1ndashC4

Ah 370ndashFBP CokeSh 370ndashFBPPl IBPndash370Nl IBPndash370Al IBPndash370Sl IBPndash370

IBP Initial boiling point of feed FBP Final boiling point of feedP Paraf ns N Naphthenes A Aromatic bare rings S Substituent groupson aromatics Subscripts l ndash Light fraction h ndash Heavy fraction

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

534 R MADHUSUDANA RAO et al

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 9: Industrial Experience with Object-Oriented Modelling FCC Case Study

consideration Based on the physicochemical phenomenaoccurring in various sections of the reactor=regeneratorand geometry of the unit a structural decomposition wasperformed The unit is divided into smaller fragments ofsystems and subsystems The decomposition is progres-sively done until each model fragment represents afundamental chemical unit operation The hierarchy ofmodel fragments as a result of decomposition performedon the commercial unit (see Figure 4) is illustrated inFigure 6 Different sections of FCCU which contribute tothe pressure drop in the system are also considered in thedecomposition and model hierarchy A model fragment oran aggregation of some model fragments represents apart or the whole of FCCU For example the riser bottomis emulated by an adiabatic mixer model wherein catalystand liquid feed combine and produce a single streamconsisting of catalyst and feed vapour Following thismethodology various sections of reactor and regeneratorare decomposed to fundamental chemical engineering unitoperation models as described below

ReactorThe reactor can be divided into various sections which

can be modelled using the following units

Mixer The mixer model emulates mixing of hot regene-rated catalyst and liquid feed at the bottom of the riser Itis modelled considering adiabatic conditions In additionthe mixer model represents other mixing processes withinthe reactor and regenerator for example mixing of steamfrom stripper bed and product vapour separated from theriser termination device

Riser The riser can be modelled either as a single con-tinuous stirred tank reactor (CSTR) representing com-plete back mixing with uniform temperature (McFarlaneet al 1993) or as a plug ow reactor with temperaturegradient along the riser height (Arbel et al 1995) Theabove two assumptions represent two extremes of model-ling the riser The actual conditions within the riser can be

assumed to represent a non-ideal mixing zone This is dueto the presence of slip between vapour and catalystparticles axial dispersion of catalyst due to turbulenceand temperature difference between the riser inlet andoutlet in actual conditions Hence the riser is modelled asa composite object characterized by series of CSTRs Slipratio catalyst circulation space velocity and void fractionare important factors which determine hydrodynamics ofthe riser The chemistry of reactions is accounted for by alumped kinetic scheme as described earlierRiser Termination Device (RTD) The RTD is modelledas a hydrodynamic unit which contributes to pressuredrop due to splitting of the vapour-catalyst stream intotwo (a lsquoTEErsquo junction) and a sudden expansion of thevapour through an opening into the reactor vesselSplitter The separation of catalyst from vapour at theRTD is modelled using a splitter which divides the inletstream into two and has a splitting ef ciency for eachcomponent It is modelled with the assumptions of nopressure loss and the three streams one inlet and twooutlet are in thermal equilibriumCyclone Separation of the catalyst nes from the reactorvapours is modelled using a cyclone Separation ef -ciency for each component and pressure drop are takeninto considerationStripper The stripper is modelled to represent stripping ofhydrocarbons It is modelled assuming a counter currentoperation with mass transfer The amount of unstrippedhydrocarbons in the stripper is a function of the steam tocatalyst ratio and is given by an empirical function (Arbelet al 1995)

A schematic of the reactor con guration in an object-oriented framework using the units described above is illu-strated in Figure 7 The riser is represented as a compositeobject in the object-oriented framework The number ofCSTRs for modelling the riser is a model input parameterThe user can specify any number of CSTRs instead ofphysically assembling that number of units on the owsheetAs a result the riser can be con gured to represent differentFCC riser models reported in literature

The riser termination device at the top of the riser (seeFigure 4) is modelled as a combination of the following unitsthe RTD with stream splitting into two halves along withpressure drop and a splitter which accounts for separation ofthe catalyst from the product vapour This methodology ofmodelling the pressure drop as a separate model fragment wasadopted in order to retain the generic nature of a component=stream splittingunit This results in a well de ned and abstractsplitter unit that can be reused The stripper is modelled intwo sections This is to account for recycling of the entrainedcatalyst through the dip leg of the reactor cyclones into thestripper bedThe stripperbed Stripper0 represents the stripperfrom the top of the bed to the dip leg exit inside the bed Thelower portion of the stripper bed from the exit of the dip leg tothe spent catalyst exit is represented by Stripper1

RegeneratorThe regenerator is mainly divided into two regions dense

bed and dilute region It also consists of hydrodynamic unitssuch as air distributor and cyclones for catalyst separationfrom ue gases Cyclones are modelled similarly to thosedescribed in the reactor sectionFigure 6 Structural decomposition of FCC reactor=regenerator

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 535

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 10: Industrial Experience with Object-Oriented Modelling FCC Case Study

Air distributor The air distributor is modelled as apressure drop unit for inlet air The pressure drop in theunit dependson the type of distributorFor a pipe grid typethe pressure drop depends on the number of open holes inthe grid the diameter of each hole and air inlet pressureDense region This region is characterized by a bed ofsolid catalyst particles and the air which ows through thebed for catalyst regeneration The dense bed is furtherdivided into two sectionsmdash Emulsion phase This phase mainly consists of solid

catalyst particles which are assumed to be completelymixed with the ow rate of air corresponding tominimum uidization velocity Hence this phase ismodelled as a single CSTR with uniform temperatureand gas composition

mdash Bubble phase The portion of air with ow rateexceeding minimum uidization velocity is assumedto ow in the form of bubbles These bubbles are con-sidered to be rising through the dense bed in near plug ow fashion The effect of expansion of bubbles dueto pressure gradient along the height of the dense bedcoalescence of bubbles and increase in volumetric owrate of bubbles due to pressure gradient are not con-sidered The near plug ow of the bubble phase ismodelled as a series of CSTRs The bubble phase isassumed to be completely free of the catalyst particles

Dilute region This region consists of combustion gasesfrom the top of the dense bed and the entrained catalystparticles Due to super cial velocity of the air and burstingaction of the bubbles at the top of the dense bed catalystparticles get entrained along with combustion gases andform a dilute phase above the dense bed In this regionboth gas and solid catalyst particles are assumed to be

moving in near plug ow In actual conditions there existssome amount of back mixing due to heavier particlesfalling off before they reach the cyclones Hencethis phase is also modelled as a series of CSTRs It isassumed that all the entrained catalyst particles return tothe dense bed through the cyclone dip legs

Various reactions occurring as a result of coke combustionin different regions of the uidized bed are considered inboth the dense bed and dilute phase Overall conversion ofcoke depends on ow rate of air its pressure and temperatureamount of catalyst in the bed catalyst circulation rate andbed temperature The combustion reactions are also stro-ngly in uenced by characteristics of the uidized bed Adetailed hydrodynamic calculation of the uidization processis considered to determine various parameters such as beddensity void fraction etc The reader is referred to theappendix for a detailed account of all the hydrodynamiccalculations and combustion reactions that were taken intoconsideration

The schematic of different regions considered for model-ling the regenerator and its owsheet representation inan object-oriented framework is shown in Figure 8 Theregenerator is a complex unit characterized by hydrodyna-mics of the uidized bed and coke combustion reactionsIn addition inlet air ow rate and its conditions stronglyin uence the physical and chemical phenomena thus givingrise to a coupled transfer and exchange of mass momentumand energy between various regenerator model fragmentsHence dense bed and dilute phase are modelled as a singlecomposite object similar to the riser The number of CSTRsassociated with both the bubble phase and regenerator dilutephase are part of input parameters to the composite object

Figure 7 Con guration of the FCC reactor system in object-orientedframework

Figure 8 Con guration of a FCC regenerator system in object-orientedframework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

536 R MADHUSUDANA RAO et al

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 11: Industrial Experience with Object-Oriented Modelling FCC Case Study

The complete owsheet schematic of the integratedFCC reactor=regenerator is shown in Figure 9 The repre-sentation of FCCU as a owsheet in MPROSIM is illu-strated in Figure 10 The gure also shows the inputtemplate for the riser In the next section we present theresults of simulation studies conducted using the FCCUmodel

Model Tuning

The FCCU model has been developed as a process owsheet in MPROSIM Various units are provided in thelibrary with which a user can select the units and connectthem to represent various con gurations of FCCU TheFCCU model has several input parameters In addition tothose associated with input streams ie ow rates compo-sition temperature pressure etc there are various unitparameters Many of these unit parameters are essentiallythe assumptions that are part of a FCCU model The usercan specify these parameters and thereby make a set ofassumptions for the model Simulation studies can becarried out by providing necessary inputs in order tocheck whether the model assumptions that the user hasmade are valid for the respective con guration of the

commercial unit If the results are not satisfactory the usercan change the modelling assumptions by changing theparameters Thus the model can be ne tuned to representthe con guration of the commercial unit

The model was tuned with industrial data provided byCPCL to represent their FCCU con guration Tuning themodel involved carrying out simulation studies by adjustingvarious unit parameters and comparing model predictionswith that of the plant data

Reactor tuningObject-oriented owsheet representation of the reactor in

MPROSIM is shown in Figure 11 Inputs to the reactor owsheet consisted of feed ow rates composition of feedin terms of various lumps regenerated catalyst and steam forthe stripper section Inputs to the reactor owsheet areshown in Table 3 One of the important tuning parametersfor the reactor owsheet is number of CSTRs for the riserStudies were conducted using the reactor model by varyingnumber of CSTRs The model predictions were comparedwith process conditions and product yield of the commercialunit All other parameters and inputs such as catalystcirculation rate feed conditions etc are kept constantResults of the simulation studies are given in Figures 12ndash16

Figure 9 FCC reactor=regenerator model in an object-oriented framework

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 537

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 12: Industrial Experience with Object-Oriented Modelling FCC Case Study

The trends of variables shown in the above gures provethat the assumption of approximating the reactor riser by a nite number of CSTRs-in-series is appropriate A smallnumber of CSTRs would mean the riser is close to idealmixing conditions At the same time a large number ofCSTRs would indicate a proximity to near plug owcondition It is evident from the above gures that beyonda certain number of CSTRs (say 10) change in the modelprediction is not signi cant Once the model is tuned withrespect to the number of CSTRs for the riser an appropriatenumber can be selected for carrying out further simulationstudies on the reactor and integrated reactor=regenerator owsheet We have chosen 10 CSTRs to represent CPCLrsquosreactor riser for conducting further simulation and optimiza-tion studies

Regenerator tuningThe regenerator modelled as a owsheet in MPROSIM is

illustrated in Figure 17 The gure also shows the input temp-late for the regenerator composite object Typical parametersused for tuning the regenerator model are (1) number ofCSTRs for bubble phase (2) number of CSTRs for dilutephase and (3) entrained catalyst as weight percent of

Figure 10 FCC reactor=regenerator model in MPROSIM

Table 3 Reactor inputs

Parameter (units) Value=source

Feed conditionsTotal feed ow rate (Kg hiexcl1) 10048240Composition (Wt fraction)

Ph 02866Nh 01061Ah 00836Sh 01905Pl 01151Nl 00426Al 00535Sl 01220

Feed temperature (macrC) 3510Feed pressure (Kg cmiexcl2) 274

CatalystRegenerated catalyst ow rate (Kg hiexcl1) 5940000Coke on regenerated catalyst (Wt percent) 025Temperature (macrC) 6600

SteamFlow rate (T hiexcl1) 150Temperature (macrC) 2200Pressure (Kg cmiexcl2) 280

Other inputsKinetic parameters Arandes et al 2000Catalyst deactivation Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

538 R MADHUSUDANA RAO et al

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 13: Industrial Experience with Object-Oriented Modelling FCC Case Study

catalyst in the regenerator Inputs to the regenerator ow-sheet are shown in Table 4 Tuning of the regenerator owsheet is done by varying the number of CSTRs ofboth bubble phase and dilute region for a xed value ofcatalyst entrainment After each simulation model predic-tions are checked with process data from the commercial

unit The combinations of number of CSTRs and catalystentrainment that match the plant data are chosen as repre-sentative values Regenerator model tuning results are givenin Table 5 and parameters chosen for CPCLrsquos regeneratorcon guration are highlighted

Parametric=Optimization Studies

The parameters chosen from tuning of the reactorand regenerator are used in the integrated FCC reactor=regenerator owsheet to conduct further simulation studiesA comparison of results of model prediction and corre-sponding process data obtained from the commercial unitare shown in Table 6 Some of the data corresponding tomodel predictions could not be derived from the commercialunit due to reasons of instrumentation and lack of samplingpoints

A process model can also be used for various optimiza-tion studies In the case of the FCCU model yield optimiza-tion studies were carried out with the model The effect offeed composition and temperature on gasoline yield isshown in Figure 18 Three different feed compositionsconsidered for this study are listed in Table 7 The behaviourshown in the plots is typically expected of gasoline Refer-ring to Figure 5 it can be seen that gasoline not only formsfrom heavy and light fractions of the feed but also under-

Figure 11 Reactor owsheet in MPROSIM

Figure 12 Wt fraction of feed lumps (heavy) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 539

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 14: Industrial Experience with Object-Oriented Modelling FCC Case Study

goes a cracking reaction to produce light components andcoke Due to the secondary cracking nature of gasoline itsyield exhibits a maximum as a function of feed temperatureIt is evident from Figure 18 that not only does the gasolineyield change with change in feed composition but theoptimum temperature also changes This result is in goodagreement with procedures typically followed in the opera-tion of the commercial unit Stable operation of FCCU ischallenged by frequent changes in feed composition due tochange of crude oil input to the re nery To maximize theyield from FCCU operating parameters such as catalystcirculation rate and feed inlet temperature are regularlymonitored and manipulated whenever feed compositionchanges

The next section brie y summarizes the authorsrsquo experi-ences in the implementation of object-oriented frameworkfor modelling a large scale industrial process such as FCCUThe section also highlights and clari es some of the lessonslearnt during model development stage and implementationperiod and in hindsight vis-a-vis the practical issues listedin the section on Review of Process Modelling above

Discussion and Clari cations

To begin with the FCCU model development projectwas undertaken in an academic institution with the supportof CPCL The scope of the project was to develop a steady-state model for CPCLrsquos FCC reactor=regenerator con -guration The model was proposed to be developed in anobject-oriented framework to support the re neryrsquos researchand development team in carrying out off-line simulationstudies for trouble-shooting due to frequent changes in feedcomposition catalyst evaluation and independent studies ofinternal components of the FCCU such as cyclone stripperetc From a detailed review of the literature on FCCUmodelling the model development was focussed towardsdeveloping a general framework for modelling variousFCCU con gurations proposed in the literature Thoughthe outcome of this development effort was more orientedtowards FCCU modelling and addressing some of there nery-based modelling issues there are some generalprinciples that can be culled from this effort towardsunderstanding the object-oriented modelling of chemicalprocesses Based on the modelling effort the following

Figure 13 Wt fraction of feed lumps (light) vs no of CSTRs

Figure 14 Wt fraction of product lumps vs no of CSTRs

Figure 15 Total riser yield (Wt fraction) vs no of CSTRs

Figure 16 Riser outlet temperature ( C) vs no of CSTRs

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

540 R MADHUSUDANA RAO et al

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 15: Industrial Experience with Object-Oriented Modelling FCC Case Study

comments could be made regarding generic and implemen-tation issues laid down in section on Review of ProcessModelling above

the development team consisted of chemical engineersfrom academia and industry speci cally from a re nerySo the knowledge domain of the group of developers isrestricted to a few areas of chemical engineering Typi-cally this would be the scenario in any model develop-ment activity undertaken in collaboration with anindustrial partner As a result the model developmentactivity is in uenced by the knowledge of the modellingexperts In addition several practical issues have to betaken into consideration as highlighted in the following

discussion Even if a prototype of a generic framework isdeveloped independent of any industrial input it wouldneed to be critically evaluated Implementation of variousrepresentative large scale processes from diverse engineer-ing elds such as petroleum petrochemical biologicalsolids-based etc have to be considered for the evaluationThe amount of time and expertise needed would always bean important limiting factorthe development team would have to judiciously choosea decomposition strategy to model the process in anobject-oriented framework Some of the choices availablefor developing an object-oriented model for FCCU areenumerated in Table 8 The choice of the authors for themodelling of FCCU in this contribution is alsohighlighted The selection of a particular option is notonly in uenced by the modelling experts from academiabut also by the industrial partner In addition the durationand time-frame of the project have to be taken intoconsiderationthe granularity to which the given process is decomposedand design of the framework directly depend on thechoice made from the options listed in Table 8 Typicallyat each level of decomposition some assumptions may beintroduced into some or all the model fragments Essen-tially a higher degree of fragmentation would imply alarger number of assumptions in the model Theseassumptions are veri ed as a result of experimentation

Figure 17 Regenerator owsheet in MPROSIM

Table 4 Regenerator inputs

Parameter (units) Value=source

AirAir ow rate (Kn m3hiexcl1) 430Temperature (macrC) 2370Pressure (Kg cmiexcl2) 260

CatalystSpent catalyst (Kg hiexcl1) 5940000Temperature (macrC) 5005Coke on spent catalyst (Wt percent) 0943

Other inputsFluidization parameters Arbel et al 1995

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 541

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 16: Industrial Experience with Object-Oriented Modelling FCC Case Study

Table 5 Regenerator tuning results

No CSTRs Tbubble (macrC) Te (macrC) Tdilute (macrC) Tdelta (macrC)

BP DP Model Plant Model Plant Model Plant Model Plant

I Wt fraction of regenerated catalyst entrained ˆ0051 1 74916 mdash 66436 65300 83534 78000 17098 127002 1 60187 mdash 67152 65300 100359 78000 33207 127003 1 62429 mdash 67744 65300 102464 78000 34720 127004 1 67846 mdash 67440 65300 104646 78000 37206 127005 1 70722 mdash 67843 65300 102234 78000 34391 12700

1 2 74868 mdash 66911 65300 85731 78000 18820 127002 2 60465 mdash 67484 65300 106190 78000 38706 127003 2 62765 mdash 68094 65300 108555 78000 40461 127004 2 67756 mdash 68262 65300 109152 78000 40890 127005 2 69273 mdash 68773 65300 104238 78000 35465 12700

1 3 74841 mdash 67526 65300 86157 78000 18632 127002 3 60466 mdash 67484 65300 106195 78000 38710 127003 3 62767 mdash 68096 65300 108601 78000 40505 127004 3 63947 mdash 68274 65300 107029 78000 38754 127005 3 69255 mdash 69608 65300 103496 78000 33888 12700

I Wt fraction of regenerated catalyst entrained ˆ0101 1 74870 mdash 66468 65300 80935 78000 14468 127002 1 61251 mdash 68404 65300 94029 78000 25625 127003 1 63560 mdash 68886 65300 93864 78000 24978 127004 1 64800 mdash 69025 65300 92798 78000 23773 127005 1 70932 mdash 67835 65300 94158 78000 26323 12700

1 2 74872 mdash 67148 65300 84613 78000 17465 127002 2 61784 mdash 69011 65300 99685 78000 30674 127003 2 64010 mdash 69310 65300 97738 78000 28428 127004 2 65242 mdash 69389 65300 96103 78000 26714 127005 2 66139 mdash 68375 65300 97180 78000 28805 12700

1 3 74847 mdash 67661 65300 85039 78000 17378 127002 3 61797 mdash 69026 65300 99839 78000 30813 127003 3 63978 mdash 69283 65300 97603 78000 28321 127004 3 65199 mdash 69357 65300 95956 78000 26599 127005 3 69147 mdash 69575 65300 94262 78000 24687 12700

II Wt fraction of regenerated catalyst entrained ˆ 0141 1 74930 mdash 66361 65300 78328 78000 11967 127002 1 61754 mdash 68976 65300 89515 78000 20539 127003 1 64107 mdash 69395 65300 89118 78000 19723 127004 1 65385 mdash 69496 65300 88186 78000 18690 127005 1 69301 mdash 68062 65300 86184 78000 18122 12700

1 2 74884 mdash 66867 65300 83172 78000 16305 127002 2 62437 mdash 69728 65300 94579 78000 24851 127003 2 64636 mdash 69861 65300 92224 78000 22363 127004 2 65946 mdash 69905 65300 90873 78000 20968 127005 2 65847 mdash 68470 65300 91187 78000 22717 12700

1 3 74877 mdash 67653 65300 84069 78000 16417 127002 3 62422 mdash 69713 65300 94530 78000 24816 127003 3 64582 mdash 69819 65300 92047 78000 22228 127004 3 65874 mdash 69861 65300 90697 78000 20836 127005 3 68258 mdash 69282 65300 89303 78000 20022 12700

II Wt fraction of regenerated catalyst entrained ˆ 0151 1 74914 mdash 66348 65300 77596 78000 11248 127002 1 61848 mdash 69081 65300 88548 78000 19467 127003 1 64213 mdash 69491 65300 88159 78000 18668 127004 1 67849 mdash 67610 65300 88512 78000 20902 127005 1 71880 mdash 68235 65300 85143 78000 16909 12700

1 2 74900 mdash 66978 65300 82972 78000 15994 127002 2 62552 mdash 69851 65300 93403 78000 23552 127003 2 64765 mdash 69968 65300 91136 78000 21168 127004 2 68675 mdash 68483 65300 89508 78000 21025 127005 2 66654 mdash 68930 65300 88526 78000 19597 12700

1 3 74813 mdash 67784 65300 83906 78000 16122 127002 3 62532 mdash 69831 65300 93329 78000 23497 127003 3 64705 mdash 69923 65300 90952 78000 21029 127004 3 66016 mdash 69958 65300 89664 78000 19706 127005 3 71185 mdash 69633 65300 88474 78000 18841 12700

BP Bubble phase DP dilute phase Parameters chosen for CPCLrsquos regenerator con guration

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

542 R MADHUSUDANA RAO et al

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 17: Industrial Experience with Object-Oriented Modelling FCC Case Study

and=or by matching the model predictions with theindustrial data However the quantity and quality ofdata that can be derived from a running industrial plantare certainly limited Some of the reasons for this areinstrumentation for the process is xed and no additionalprocess data can be extracted test runs that an industrialpartner is willing to conduct on the commercial unitinaccuracies of the process data and the availability ofsophisticated instrumentation for carrying out lab-scaleexperimental studies Hence hierarchy of the granularstructure designed for a process model is limited to alarge extent by the above practical issues To cite anexample the stripper section was modelled as an idealcounter-current mass and energy exchange operationthough in the actual design of the unit the operation isnot so Since no speci c tests could be conducted on thestripper of the commercial unit and due to unavailabilityof any lab-scale experimental setup the assumptionscould not be validated Hence an empirical function formass exchange ie stripping of hydrocarbons was takenfrom the literature (Arbel et al 1995) without any changeof function parametersuniqueness of decomposition decomposition of theFCCU model proposed earlier is not unique The decom-position strategy adopted resulted in more than onedegree of freedom in terms of the number of CSTRs forreactor riser regenerator bubble and dilute phases If aPFR model had been chosen for the same regions thedegrees of freedom would have been xed and alsounique to some extent But the choice of CSTR model

fragment allows the model to be exible and it can becon gured to represent many commercial FCC reactor=regenerator unitsas described earlier many FCCU models have beenproposed in the literature One of the important aspects ofthis model development has been to conceptualize a frame-work in which many FCCU models can be con gured Thehierarchy of structural decomposition illustrated in Figure 6is general enough to represent many FCCU modelsproposed in the literature An aggregation of the modelfragments along with appropriate model assumptions wouldresult in a particular con guration The following twoexamples are highlighted to illustrate the idea

mdash if the riser is assumed to be a single CSTRregenerator bubble and dilute phases are modelledby a large number of CSTRs and empirical kineticsare assumed for overall reactor conversion then themodel would represent that proposed by McFarlaneet al (1993) in its steady-state form

mdash on the other hand if the reactor riser regeneratorbubble and dilute phases are approximated by largenumber of CSTRs and a 10-lump kinetic model isused to describe the catalytic reactions then themodel would be that proposed by Arbel et al(1995)

similarly by considering different combination of CSTRsfor the bubble and dilute phases various models proposedfor the regenerator can be realized The object-orientedmodelling of FCCU as a owsheet consisting of variousparts of the process as model fragments has resulted inthe conceptualization of a general framework which can beeasily extended and adapted to model various con gura-tions of FCC reactor=regenerator However the issuesconcerning the reuse extension and adaptability of eachof the model fragments need to be critically evaluated Thesuccess and failure issues concerning the modelling of riserand regenerator composite objects are evaluated below

Table 6 Comparison of simulated and plant data

Process variable Model prediction Plant data

ReactorRiser bottom temperature (macrC) 55018 mdashRiser outlet temperature (macrC) 50178 49850Reactor yield 693 680

CatalystCatalyst circulation rate (T miniexcl1) 971 99Spent catalyst temperature (macrC) 5004 46950Regenerated catalyst temperature (macrC) 66361 6530Coke on spent catalyst (Wt) 0943 mdashCoke on regenerated catalyst (Wt) 045 025

RegeneratorBubble phase outlet temperature (macrC) 7493 mdashEmulsion phase outlet temperature (macrC) 66361 6530Dilute phase outlet temperature (macrC) 78328 7800

Figure 18 Optimization of gasoline yield

Table 7 Feed composition inputs for gasoline optimization

Wt fraction

Component Feed 1 Feed 2 Feed 3

Ph 02866 01791 01592Nh 01061 00707 00707Ah 00836 00557 00557Sh 01905 01190 01058Pl 01151 01957 02072Nl 00426 00823 00852Al 00535 00900 00963Sl 01220 02073 02195

Table 8 Degrees of freedom for modelling FCCU

Option Description Evaluation

1 FCC reactor=regenerator A monolithic block2 Reactor regenerator Two model blocks3 Reactor kinetics reactor

hydrodynamics regenerator kineticsregenerator hydrodynamics

Four model blocks

4 Decomposition as given in Figure 6 General framework5 Decomposition based on Figure 2 Fundamental approach

Option chosen in this contribution

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 543

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 18: Industrial Experience with Object-Oriented Modelling FCC Case Study

mdash the riser is modelled as a composite object It canfunction as a single CSTR or as a non-ideally mixedtubular reactor by a series of CSTRs It can also beapproximated as a PFR by specifying a large numberof CSTRs ideally in nite The riser is one of themodel fragments which truly re ects the behaviour ofa composite objectmdashone that can be completelyreused and can be easily extended and adapted

mdash the regenerator composite object is the model frag-ment for a uidized bed As highlighted earlier theregenerator uidized bed is a complex and coupledprocess In its present state the regenerator compositeobject can be used to represent a single stage regen-erator It is also possible that a two-stage regeneratorcan be modelled using a stack of two compositeobjects If the regenerator composite object is furtherdecomposed resulting in more fundamental modelfragments then it is possible to model many indus-trial-scale uidized bed processes The issuesconcerning this problem are currently being studied

As indicated in the issues listed above the developmentand implementation of a formal structure guidelines andprinciples of object-oriented modelling on a large-scaleindustrial process are strongly in uenced not only by theexpertise of the development team as a whole but also byvarious factors which comprise the needs and requirementsof the industry on a commercial scale Many of the exten-sion and adaptability issues related to MPROSIM as asimulation and modelling tool are currently being evaluatedFor this purpose we are undertaking a detailed study ofobject-oriented modelling in other processes such as fuelcells biochemical and biological processes etc

UTILITY AND MERITS

The representation of the FCCU model as a owsheet hasseveral advantages It is possible to carry out simulationstudies with an integrated reactor=regenerator owsheetwith individual reactor and regenerator owsheets andstudy the performance of individual units such as reactoror regenerator cyclones stripper air distributor etc Severalstudies on sensitivity of reactor yield and other importantprocess parameters to input conditions can be carried outwith the model Typically carrying out sensitivity para-metric and optimization studies involves a lot of effort Butwith the above model developed in an object-orientedframework studies like these can be conducted easily

The representation of modelling assumptions in terms ofunit parameters is essentially an extension of the object-oriented concept of encapsulation as described in the sectionon Modelling in MPROSIM above Some of the assump-tions that are encapsulated as unit parameters in the FCCUmodel are

the number of CSTRs for modelling the reactor riserregenerator bubble and dilute phasescatalyst deactivation function for catalytic cracking in theriser andsplitting ef ciencies (either direct numerical values or interms of empirical functional forms) for all components inunits such as splitter cyclone and stripper

One of the important problems that units in a re nerytypically FCCU have to face is that the feed composition

changes regularly With evolution of technologies and needsof particular markets in terms of product requirements andspeci cations it becomes necessary for the unit con gura-tion also to change with time Hence it is an inevitableexercise for any process model to be regularly ne-tuned torepresent the changing conditions in the plant In such ascenario a process model developed in an object-orientedframework can provide better support to the process engi-neer In the case of FCCU various reactor and regeneratorcon gurations are in operation in the re neries Some of thevariations in structural con gurations of the reactor andregenerator are

various types of disengagement section for separation ofproduct vapour from catalyst at the top of the riserdifferent designs for the stripper sectionsingle-stage and two-stage regeneratorssingle-stage and two-stage cyclone separation for catalyst nes

In addition to the structural variations based on processdesign operating conditions of the unit also vary widelySome of the reasons for this could be the spectrum ofproducts the unit is commissioned to produce frequentchanges in the quality of feed and uctuations in themarket demand affecting the quantity of throughput Suchvariations in the con gurations and operating conditions canbe modelled very easily and quickly in MPROSIM Achange in catalyst circulation rate and=or quantity of feedto the riser will change the hydrodynamics inside the riserSimilarly air ow rate and catalyst circulation rate willaffect hydrodynamics of the catalyst bed in the regeneratorThe model can be ne tuned by simply changing the numberof CSTRs to re ect the new process operating conditionsThe key factor responsible for exibility of the FCCUmodel in MPROSIM is its modularizationand representationin the form of a owsheet Unlike the traditional modelswhich might be dif cult to con gure to represent variouscon gurations of FCCU the object-oriented model can beeasily con gured to represent various con gurations

CONCLUSIONS

An evaluation of the principles of structured processmodelling was undertaken in this work This was done byhighlighting the critical issues and problems encounteredwhile implementing the guidelines for structured object-oriented process modelling As a case study for the processmodel an industrial-scale FCCU was chosen The FCCUmodel served as an appropriate example to bring forth theimportant object-oriented modelling issues such as granu-larity adaptability and extensibility

The model fragments proposed for the FCCU model aregeneral enough to be able to represent various modelsproposed in the literature and other con gurations ofFCCU However issues relating to the reusability andextensibility of these model fragments to other chemicalprocess systems are still open and work is under progress inthis regard As for the main aim of MPROSIM and develop-ment of a general purpose modelling tool it is proposed tostudy processes and systems from various other science andengineering elds to address issues such as generality ofprocess models and their implementational details

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

544 R MADHUSUDANA RAO et al

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 19: Industrial Experience with Object-Oriented Modelling FCC Case Study

APPENDIX FCCU MODEL EQUATIONS

The model equations for the various sections of the reactorand regenerator are summarized below Typically each unitcomprises the mass energy and momentum balances Thereactor riser incorporates the 11-lump kinetic model andpressure drop equations based on the pneumatic transportof the solids On the regenerator side the detailed hydro-dynamic equations are written characteristic of a uidizedbed taking into consideration the coke combustion kinetics

Reactor

MixerThe mixer is modelled under adiabatic conditionswith thetypical mass and energy balance equations The mixerdoes not contribute any pressure drop The equationspertaining to the pressure drop due to mixing ow acrossnozzles and so on are modelled as separate hydrody-namic units For the case of riser bottom the pressuredrop in the liquid feed across the distributor is modelledas a hydrodynamic unitRiserThe riser is modelled as a composite object The compo-site object contains mass and energy balance equationsincorporating reaction kinetics In addition detailedmomentum balance equations are taken into considera-tion The model equations for the riser are typi ed byCSTR model equations The effect of coking and catalystdeactivation are taken into consideration by includingappropriate terms in the model equations (McFarlaneet al 1993 Arbel et al 1995)mdash Mass balance The mass balance equation for a

component j (either a reactant or product) withinany CSTR inside the riser for the 11-lump kineticmodel can be written as

Fjout ˆ Fjin iexcl rvFcat

Foilf(t)

pound 11 Dagger KdyAh

X11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

Foil

(1)

mdash Energy balance A CSTR heat balance gives thetemperature pro le in the riser based on the numberof CSTRs considered to model the riser The energybalance equation for any CSTR in the riser compositeobject can be written as

FjinCpfvTin Dagger FcatCpcatTin

ˆ FjoutCpfvTout Dagger FcatCpcatTout

Dagger rvFcat

Foilf(t)

11 Dagger KdyAh

poundX11

kˆ1

nkKk0e(iexclEk=RT ) Fkout

FoilDHrk (2)

mdash Pressure drop The pressure drop in the riser ismodelled similarly to the pneumatic transport ofsolid particles in the vertical direction The followingterms are considered as the contributing factors to the

pressure drop within the riser (Jones et al 1967Yang 1973 1974)

acceleration of solid particles

DPr18F2

cat

(36 105)2p2rpartgd4riser

(3)

static head of the catalyst-vapour mixture

DPr2 ˆ Fcat Dagger Foil

(Fcat=rpart) Dagger (Foil=rv)

Aacute hriser

104 (4)

where rv feed vapour density at the bottom of theriser is calculated by the equation

PmixerMfeed

RTmixer(5)

friction between uid media and wall

DPr3 ˆ 2driser

hriser

sup3 acute0079Re025

sup3 acuterv

Foil=rv

(p=4)d2riser

sup3 acute2

(6)

assuming turbulent conditions The Reynolds number(Re) is given by the following equation

RedriserVvapourrv

mv(7)

The feed oil vapour velocity (Vvapour) inside the risercan be calculated by the following equation

VvapourFoil=rv

(p=4)d2riser

(8)

The viscosity of the vapour at the mixer outletconditions is given by

mv10319 10 6(Tmixer 27315)04896

10 (34737 102)=(Tmixer 27315)

(9)

acceleration of the uid media

DPr410

2 104

rvVvapour

g(10)

Riser Termination Device (RTD)The inlet stream of the RTD is assumed to be split intotwo equal halves taking into consideration the geometryof the disengaging section at the top of the riser Furtherall the three streams are assumed to be in thermalequilibrium The pressure drop in the RTD is modelledby considering the following terms (Bird et al 1960)mdash a lsquoTEErsquo junction which is approximated by two

90 bends

DPrtd1 ˆ rmix

2 pound gV 2

rtd23

b21

iexcl 10

Aacute

pound 10iexcl4 (11)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 545

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 20: Industrial Experience with Object-Oriented Modelling FCC Case Study

where b1 rmix and Vrtd are given by the followingequations

b1driser

darm(12)

rmixFcat Foil

(Fcat=rpart) (Foil=rv)(13)

Vrtd

(Fcat=rpart) (Foil=rv)

(p=4)d2riser

(14)

mdash sudden expansion of the catalyst and product vapourmixture into the reactor vessel

DPrtd2 ˆ rmix

gV 2

armb2(1 iexcl b2) pound 10iexcl4 (15)

where b2 is the area ratio of the opening in the sidearm of the RTD and the annular area in the reactorportion It is given by the following equation

b2 ˆ pwhw

p=4(d2reac iexcl d2

riser)(16)

and Varm is given by the equation

Varm

(Fcat=rpart) (Foil=rv)

(p=4)d2arm

(17)

CyclonesThe cyclone separators used for the recovery of catalyst nes are modelled for the pressure drop The totalpressure drop in the cyclone is a sum of the followingindividual pressure drops (Perry and Green 1997)mdash inlet contraction

DPc1 ˆ 05rgmix(V 2in iexcl V 2

vessel Dagger KV 2in) (18)

mdash particle acceleration

DPc2 ˆ LVin(Vpin iexcl Vpvessel) (19)

For small particles we can consider

Vpin Vin (20)

and

Vpvessel Vvessel (21)

mdash barrel friction

DPc3 ˆ2f rgmixV

2inpDcNs

din(22)

The friction factor f is based on the Reynolds numbercalculated at the inlet area The Reynolds number iscalculated by the following equation

Re ˆdinVinrgmix

mgmix(23)

and din is calculated as

din4(inlet area)

inlet perimeter(24)

and mgmix the viscosity of the vapour-steam mixturecan be calculated by the empirical equation

mgmix10319 10 6T04896

10 (34737 102)=T(25)

mdash gas ow reversal

DPc4 ˆrgmixV 2

in

2(26)

mdash exit contraction

DPc5 ˆ 05rgmix(V2exit iexcl V 2

c Dagger KV 2exit) (27)

SplitterThe splitter is modelled to separate the inlet streaminto two streams with an ef ciency of splitting foreach component present in the inlet The modelequations consist of the mass balance for each of thecomponents

Fjin ˆX

streams

Fjout j ˆ 1 2 11 (28)

and

Fjout ˆ ZjFjin j ˆ 1 2 11 (29)

where Zj is the ef ciency of splitting for the componentj The splitter is modelled such that the inlet andoutlet streams are in thermal and hydrodynamicequilibriumStripperThe stripper is modelled as a counter current operationwith steam stripping the hydrocarbons from the voids inthe stripper bed as well as within the catalyst pores Theef ciency of stripping depends on the steam to catalyst ow rate and is assumed to be a linear function of thisratio (Arbel et al 1995) Since the stripping is a countercurrent operation with steam injected at the bottom andcatalyst owing down the pressure of the streams whichlie on the same side of the stripper unit are considered tobe equal and the two outlet streams are assumed to be inthermal equilibrium

Regenerator

Air distributorAssuming a uniform distribution of air without anychannelling effects the following equation models thepressure drop for the pipe grid type air distributor (CPCL1999)

DPad ˆ PairVair

3960Tair(30)

Dense bedAs described earlier the regenerator dense bed is dividedinto two phases emulsion phase and bubble phase Theemulsion phase consists of the catalyst particles and air ow equivalent to the minimum uidization velocity The

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

546 R MADHUSUDANA RAO et al

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 21: Industrial Experience with Object-Oriented Modelling FCC Case Study

bubble phase consists of the air ow rate which is inexcess over the minimum uidization requirement Theair entering from the distributor stream U0 (see Figure 8)is divided into two streams (U07 Umf) corresponding tothe bubble phase ow and Umf equivalent to theminimum uidization velocity (Kunii and Levenspiel1968 1990)mdash Splitter 1

The ow in stream Umf will correspond to mini-mum uidization velocity which is given by theequation

Umf ˆ90 pound 10iexcl4 middotdd18

p b(rpart iexcl rair)gc0934

r0066air m087

air

(31)

where rair can be calculated from the ideal gasequation written at the air distributor exit conditionsand the viscosity of air can be calculated from theempirical correlation

mair ˆ (0016931 Dagger 49863 pound 10iexcl5Tair iexcl 31481

pound 10iexcl8T2air Dagger 12682 pound 10iexcl11T3

air) pound 10iexcl3

g

(32)

The average particle size can be calculated given theparticle size distribution (PSD) of the equilibriumcatalyst using the following equation

ddp1P

i (xi=dpi)

10 6 (33)

A balance across the Splitter 1 (see Figure 8) will givethe equation for calculating the air ow rate in theform of bubbles

Fairin Fairbubblep4

Umf d2reg (34)

mdash Emulsion phaseThe emulsion phase is assumed to be a completelymixed zone One of the attributes of the CSTR thatshould be known is the volume occupied by the CSTRHence before the balance equations for the individualcombustion components are presented detailed hydro-dynamic calculations for the uidized bed are listedbelow

The fraction of the total dense bed occupied by thebubbles is given by the equation

dUo Umf

Ub(35)

where the bubble rise velocity (Ub) is given by

Ub Uo Umf Ubr (36)

Ubr Uo Umf 0711(gdb)

p(37)

where db the effective bubble diameter is calculatedfrom the following equation

db 0667q0375o (38)

Therefore the volume occupied by the bubble phaseis given by

Vb dZbedAreg (39)

Zbed min ZcycWreg rdilAregZcyc

Areg(rdense rdil)

( )

(40)

rdil max00(aUo b) (41)

rdense (1 edense)rpart (42)

edenseUo cUo d

(43)

and the volume occupied by the emulsion phase is

Ve ˆ ZbedAreg iexcl Vb ˆ ZbedAreg(1 iexcl d) (44)

The emulsion phase is considered as a single CSTRand the bubble phase is approximated by CSTRs-in-series see Figure 8 The following combustionreactions are considered in the regenerator (Arbelet al 1995 Weisz and Goodwin 1966)

r1 C 05O2 CO

r2 C O2 CO2

r3 CO 05O2 CO2(Heterogeneous)

r3h CO 05O2 CO2(Homogeneous)

r4 H2 05O2 H2O

It is assumed that the complete hydrogen in the cokeis burnt in the emulsion CSTR itself and hence noreaction kinetics are taken into consideration for thehydrogen balance equation (Arbel et al 1995) Allthe other reactions (r1 r2 r3c and r3h) occur in theemulsion phase Writing the balance equation foreach of the components in the emulsion phase

Mass BalanceBalance equation for catalyst

X

streams

Fcatin ˆX

streams

Fcatout (45)

Balance equation for carbon

Coke is assumed to be of the formula CHnTherefore the balance can be written as

Fcokein

12 n12 r1 r2 Fcokeout (46)

The product stream from the emulsion CSTRdoes not contain any hydrogen in the coke as allthe hydrogen is converted to H2O in the emulsionphase itself The reaction rates for the carbon inthe rst two reactions described above r1 and r2

are given by

r1 ˆ Ve(1 iexcl edense)rpartK1Fcokeout

FcatoutPO2

(47)

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 547

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 22: Industrial Experience with Object-Oriented Modelling FCC Case Study

where the rate constant K1 and the partial pres-sure of O2 PO2

are de ned by the followingequations

K1 ˆ 26 pound 1011 exp(iexcl12926=Te)2512 exp(iexcl3420=Te) Dagger 1

(48)

PO2ˆ

(FO2 out=32)

FtotaloutPreg (49)

Ftotalout is total sum of the molar ow rates of thegaseous components in the emulsion phaseoutlet stream and is given by

Ftotalout

FO2

32FCO

28

FCO2

44

FN2

28

FH2O

18 out(50)

r2 Ve(1 edense)rpartK2Fcokeout

FcatoutPO2

(51)

where the rate constant K2 is de ned by thefollowing equation

K2 ˆ 1035 pound 108 exp(iexcl9506=Te)2512 exp(iexcl3420=Te) Dagger 1

(52)

Balance equation for CO2

The total CO2 produced in the emulsion phase iscontributed by three different reactions r2 r3cr3h taking place simultaneously Therefore abalance on CO2 gives us

FCO2 out4412

r24428

r3c r3h

XNb

k 1

MkCO2

(53)

where the last term on the right hand side of theabove equation represents the mass transfer ofCO2 between emulsion phase and bubble phaseCSTRs The reaction terms r3c and r3h arecombined and given by

r3 Ve K3 PO2PCO (54)

K3 xpt(1 edense)rpartK3c

edenseK3h (55)

K3c K3c0 expEk3c

RTe(56)

K3h K3h0 expEk3h

RTe(57)

The mass transfer between bubble and emulsionphase CSTRs is given by

M kCO2

Kbe

(FkCO2 out=44)

Fairbubble

(FeCO2out=44)

Fairemulsion

( )

dZbedAreg

2(58)

Balance equation for CO

The balance equation for CO is similar to thatwritten above for CO2

FCOout ˆ 2812

r1 iexcl r3 iexclXNb

kˆ1

M kCO (59)

The reaction terms have already been describedabove Writing the mass transfer equations

M kCO Kbe

(FkCOout=28)

Fairbubble

(FeCOout=28)

Fairemulsion

( )

dZbedAreg

2(60)

Balance equation for O2

The O2 supply to the emulsion phase takes placefrom both the ow stream as well as the masstransfer lines from the bubble phase Writing abalance would give

FO2out ˆ FO2in iexcl 12

cent 3212

r1 iexcl 3212

r2 iexcl 12

cent 3228

r3

iexclFcokein

12 Dagger npound n pound 1

4pound 32

DaggerXNb

kˆ1

MkO2

(61)

FO2in ˆ 21100

pound

Fairemulsion (Pair

Dagger 1033)=1033

82057 pound 10iexcl3

(Tair Dagger 27315)

pound 32

(62)

MkO2

ˆ Kbe

(FkO2out=32)

Fairbubbleiexcl

(FeO2 out

=32)

Fairemulsion

( )

pounddZbedAreg

2(63)

Balance equation for N2

The balance equations are similar to that writtenfor O2 with one exception being that there will

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

548 R MADHUSUDANA RAO et al

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 23: Industrial Experience with Object-Oriented Modelling FCC Case Study

be no reaction terms

FN2 out FN2in

XNb

k 1

M kN2

(64)

FN2 in79100

Fairemulsion (Pair

1033)=1033

82057 10 3

(Tair 27315)

28

(65)

MkN2

Kbe

(FkN2out=28)

Fairbubble

(FeN2 out=28)

Fairemulsion

( )

dZbedAreg

2(66)

Balance equation for H2O

The balance equation for H2O can be written as

Fcokein

12 nn

12

18

FH2Oout

XNb

k 1

M kH2O (67)

MkH2O Kbe

(FkH2Oout=18)

Fairbubble

(FeH2Oout=18)

Fairemulsion

( )

dZbedAreg

2(68)

Energy Balance The energy balance equation forthe emulsion phase can be written as

Qinput ˆ Qoutput Dagger Qgen (69)

Qinput ˆ Qair Dagger Qent Dagger Qsc Dagger Qcoke

Dagger QmtO2

Dagger QmtN2

(70)

The last two terms of input heat are due to the masstransfer of O2 and N2 from the bubble phaseCSTRs to the emulsion phase CSTR The heatinput due to air (Qair) can be calculated by

Qair MairCpair(Te Tref ) (71)

where the Mair is the ow rate of air in mol s 1 andis given by

Mair((Pad 1033)=1033)Fairemulsion 106

82057 Te

(72)

Pad is the pressure at the exit of the air distributorand is calculated by

Pad ˆ Pair iexcl DPad (73)

The heat input due to the entrained catalyst fallinginto the dense bed is given by

Qent Fentcat Cpcat(Tent Tref ) (74)

The heat input from the spent catalyst entering theregenerator dense bed is given by

Qsc FcatCpcat(Tsc Tref ) (75)

Heat input due to coke is calculated as

Qcoke FcokeCpcoke(Tsc Tref ) (76)

The heat inputs due to the mass transfer from bubblephase CSTRs for O2 and N2 are given by

QmtO2

XNb

k 1

MkO2

CpO2(Tbubblek Tref ) (77)

QmtN2

XNb

k 1

MkN2

CpN2(Tbubblek Tref ) (78)

Similarly the followingequationsgive the expressionsfor the heat output from the emulsion phase

Qoutput ˆ Qcat Dagger QO2Dagger QN2

Dagger QCO

Dagger QCO2Dagger QH2O Dagger Qmt

CO

Dagger QmtCO2

Dagger QmtH2O (79)

Qcat ˆ FcatCpcat(Te iexcl Tref ) (80)

QO2ˆ FO2

CpO2(Te iexcl Tref ) (81)

QN2ˆ FN2

CpN2(Te iexcl Tref ) (82)

QCO ˆ FCOCpCO(Te iexcl Tref ) (83)

QCO2ˆ FCO2

CpCO2(Te iexcl Tref ) (84)

QH2O ˆ FH2OCpH2O(Te iexcl Tref ) (85)

QmtCO ˆ

XNb

kˆ1

M kCOCpCO(Te iexcl Tref ) (86)

QmtCO2

ˆXNb

kˆ1

MkCO2

CpCO2(Te iexcl Tref ) (87)

QmtH2O ˆ

XNb

kˆ1

M kH2OCpH2O(Te iexcl Tref ) (88)

The heat generation in the emulsion phase is due tothe coke combustion reactions which are exother-mic in nature Hence the total heat generated canbe written as

Qgen r1DHTef CO r2DHTe

f2CO2r3DHTe

f3CO2

r4DHTef H2O (89)

The heats of formation at the temperature condi-tions of the emulsion phase CSTR are given by

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 549

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 24: Industrial Experience with Object-Oriented Modelling FCC Case Study

(Perry and Green 1997)

DHTefCO 763168 86604Te 8456

10 3T2e

458920Te

(90)

DHTef2CO2

416050082 158928Te

00236T 2e

8130804Te

(91)

DHTef3CO2

2961244 22836Te 10274

10 2T2e

7409644Te

(92)

DHTefH2O 101193858 63665Te

1387 10 2T2e

542305545Te

(93)

The speci c heats of the gaseous componentswhich are required for calculating the enthalpy ofthe streams can be written as functions of tempera-ture (Smith and Van Ness 1987)

CpCO 02357 4286 10 5T (94)

CpCO20235 623 10 5T

444318T2 (95)

CpH2O 0457 833 10 6T

744 10 8T2 (96)

Cpcoke 022275 1454

10 4T649444

T2 (97)

CpN20232 357 10 5T (98)

CpO202584 80625

10 6T5865625

T2(99)

mdash Bubble PhaseThe bubble phase is considered to be free of catalystparticles Hence the only reaction that takes placewithin each of the bubble phase CSTRs is the COcombustion reaction in the gas phase ie the homo-geneous reaction (r3h) The rate equation for eachCSTR can be accordingly written as

rCO ˆ V kb K3hPO2

PCO (100)

where the partial pressures have to be calculated asdescribed earlier but based on the outlet conditions ofthe corresponding bubble phase CSTR The balanceequations for other components are similar to thatwritten for the emulsion CSTR

Dilute phaseThe dilutephase is modelledas a series of CSTRs The massmomentumand energy balanceequationsfor the CSTRs aresimilar to those written for the emulsion phase Sincehydrogen in coke is assumed to be completely combusted

in the emulsion bed the corresponding mass and energybalance equations are not taken into considerationRegenerator CyclonesThe regenerator cyclone equations are similar to thereactor cyclone model equations To model a two-stagecyclone separator two single-stage cyclones are connectedin series and model equations are written accordingly

NOMENCLATURE

a b c d empirical constants in the regeneratorhydrodynamic equations

Areg cross sectional area of the regenerator m2

Cpair heat capacity of air kcal mol 1 C 1

Cpfv heat capacity of the oil vapour kcal kg 1 C 1

Cpcat heat capacity of the catalyst kcal kg 1 C 1

Cpcoke heat capacity of the coke on catalyst kcal kg 1 C 1

Cpi speci c heat of component i kcal kg 1K 1

darm inside diameter of the side arm of the RTD mdb effective bubble diameter mdp average particle size mdpi particle size mmdreac internal diameter of the reactor mdreg internal diameter of the regenerator mdriser inside diameter of the riser mDc cyclone barrel diameter mEk activation energy for cracking reaction of lump k kcal mol 1

Ek3c activation energy for the heterogeneous COcombustion reaction 1388855kcal kmol 1

Ek3h activation energy for the homogeneous COcombustion reaction 35555kcal kmol 1

Fairbubble ow rate of air through the bubble phase m3s 1

Fairemulsion ow rate of air through the emulsion phase m3s 1

Fcat catalyst circulation rate kgh 1

Fentcat entrained catalyst in regenerator cyclones kg h 1

Fjin inlet ow rate of component j kg h 1

Fjout outlet ow rate of component j kg h 1

Fejout outlet ow rate of component j from emulsion

phase CSTR kg h 1

Fkjout outlet ow rate of component j from bubble phase

CSTR k kg h 1

Foil oil feed ow rate kg h 1

g acceleration due to gravity 981m s 2

hriser total height of the riser mhw height of the opening in the RTD arm mK empirical proportionality constant for cyclone

pressure dropKbe mass transfer coef cient between emulsion and

bubble phase CSTRs kg m 3h 1

K1 rate constant for carbon combustion reaction toCO cm2kg 1h 1

K2 rate constant for carbon combustion reaction toCO2 cm2kg 1h 1

K3 rate constant for CO combustion reaction toCO2 kg CO m 3h 1(kg cm 2) 2

K3c rate constant for heterogeneous CO combustionreaction kg CO (kg cat) 1h 1 (kg cm 2) 2

K3c0 intrinsic rate constant for heterogeneous CO combustionreaction 3060kgCO (kg cat) 1h 1 (kgcm 2) 2

K3h rate constant for homogeneous CO combustionreaction kg CO m 3h 1 (kg cm 2) 2

K3h0 intrinsic rate constant for homogeneous CO combustionreaction 133 1016kgCO m 3h 1 (kgcm 2) 2

Kd aromatic adsorption coef cient (Wt frac of aromatics) 1

Kk0 intrinsic rate constant for cracking of component km3 (kg cat) 1h 1

L loading of the entrained catalyst kg m 3

Mfeed molecular weight of the oil feedM k

i mass transfer of component i between bubble phaseCSTR k and emulsion phase CSTR kg h 1

Nb number of bubble phase CSTRsNs number of spirals made by the solids inside the barrelPair blower discharge pressure psiPmixer pressure at the riser bottom kg cm 2

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

550 R MADHUSUDANA RAO et al

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 25: Industrial Experience with Object-Oriented Modelling FCC Case Study

PCO partial pressure of CO in the emulsion phase kg cm 2

PO2partial pressure of O2 in the emulsion phase kg cm 2

Preg pressure in the regenerator dilute phase kgcm 2

qo volumetric ow rate of air per hole in the air grid atthe operating temperature and pressure m3s 1

R universal gas constant 8314 Pam3mol 1K 11987kcal kmol 1K 1 82057atm cm3mol 1K 1

RCSV regenerated catalyst slide valveRe Reynolds numberSCSV spent catalyst slide valveT temperature of interest of any stream or unit KTair temperature of air at the inlet KTbubblek outlet temperature of the bubble phase CSTR k KTdelta temperature difference between emulsion and dilute

phase KTdilute dilute phase outlet or regenerator cyclone inlet

temperature KTe temperature of the emulsion phase CSTR KTent temperature of the entrained catalyst KTin riser CSTR inlet oil vapour temperature CTmixer temperature at the riser bottom CTout riser CSTR outlet oil vapour temperature CTref reference temperature for enthalpy calculations 27315 KTsc temperature of the spent catalyst KUo super cial velocity of inlet air m s 1

Umf minimum uidization velocity of air m s 1

Ub velocity of the rise of bubbles m s 1

Ubr velocity of the rise of bubbles with respect toemulsion solids m s 1

Vair velocity of air through jets ft s 1

Varm velocity of the vapour-catalyst mixture in the side armof RTD m s 1

Vkb volume of bubble phase CSTR k m3

Vb volume occupied by the bubble phase m3

Vc velocity of the vapour-steam mixture in cyclonebarrel m s 1

Ve volume occupied by the emulsion phase m3

Vexit exit velocity of the vapour-steam mixture from thecyclone m s 1

Vin velocity of the vapour-steam mixture at the inlet of thecyclone m s 1

Vpin actual particle velocity at inlet of the cyclone m s 1

Vpvessel super cial velocity of the particles inside the cyclone m s 1

Vrtd velocity of the vapour-catalyst mixture at the inlet tothe RTD m s 1

Vvapour velocity of the vapour inside the riser m s 1

Vvessel velocity of the vapour-steam mixture inside thecyclone m s 1

w width of the opening in the RTD arm mWreg catalyst loading in the regenerator bed kgxi fraction of particles in size range ixpt relative catalytic CO combustion rateyAh mass fraction of heavy aromatic rings within the riser CSTRZbed height of the uidized bed of catalyst mZcyc height of the cyclone inlet from the air distributor m

Greek symbolsd fraction of the dense bed occupied by the bubble phaseedense void fraction in the dense bedZj ef ciency of split for the component jnk stoichiometric coef cient for component k ( 1)mgmix viscosity of the vapour-steam mixture kg m 1s 1

mair viscosity of air at the exit of air distributor kg m 1s 1

mv viscosity of the oil vapour at the riser bottomconditions kg m 1s 1

f(t) catalyst deactivation functionrair density of air at the exit of air distributor kg m 3

rgmix mixture density of catalyst and steam kg m 3

rdense density of the dense bed medium kg m 3

rdil density of the dilute phase medium kg m 3

rpart particle density of the catalyst kg m 3

rmix mixture density of catalyst and vapour kg m 3

rv oil vapour density at the bottom of the riser kg m 3

t catalyst residence time within each riser CSTR sDHT

f i heat of formation of component i attemperature T kcal kg 1

DHrk heat of cracking of component k kcal kg 1

DPad pressure drop across the distributor psiDPri pressure drop terms inside the riser kg cm 2

DPrtd1 pressure drop term inside the RTD kg cm 2

DPrtd2 pressure drop term inside the RTD kg cm 2

DPci pressure drop terms inside the cyclone kg cm 2

References

Arandes JM Azkoiti MJ Bilboa J and de Lasa HI 2000 Modellingfcc units under steady and unsteady state conditionsCan J Chem Eng 78111ndash123

Arandes JM and de Lasa HI 1992 Simulation and multiplicity of steadystates in uidized FCCUs Chem Eng Sci 47 2535ndash2540

Arbel A Huang Z Rinard IH Shinnar R and Sapre AV 1995Dynamic and control of uidized catalytic crackers l modeling of thecurrent generation of FCCs Ind Eng Chem Res 34 1128ndash1243

Arthur JR 1951 Reactions between carbon and oxygen Trans FaradaySoc 47 164ndash178

Barton PI and Pantelides CC 1994 The modeling of combineddiscrete=continuous processes AIChE J 40 966ndash979

Barton GW and Perkins JD 1988 Experiences with speedup in mineralprocessing industries Chemical Engineering Research Design 66408ndash418

Behie LA and Kehoe P 1973 The grid region in a uidized bed reactorAIChE J 19(5) 1070ndash1072

Biegler LT 1989 Chemical process simulation Chem Eng Progr Oct50ndash61

Bird R Stewart W and Lightfoot E 1960 Transport Phenomena (JohnWiley amp Sons Inc New York USA)

Bogusch R and Marquardt W 1995 A formal representation of processmodel equations Comput Chem Eng 19(Suppl) S211ndashS216

Bogusch R and Marquardt W 1997 A formal representation of processmodel equations Comput Chem Eng 21(10) 1105ndash1115

Boston JF Britt HI and Tayyabkhan MT 1993 Software tacklingtougher tasks Chem Eng Progr Nov 38ndash49

Bozicevic J and Lukec D 1987 Dynamic mathematical model for a uidcatalytic cracking process Trans Inst Meas Control 9(1) 8ndash12

CPCL 1999 FCC Operating Manualde Lasa HI Errazu AF Barreiro E and Solioz S 1981 Analysis of

uidized bed catalytic cracking regenerator models in an industrial scaleunit Can J Chem Eng 59 549ndash553

de Lasa HI and Grace JR 1979 The in uence of a freeboard region in a uidized bed catalytic cracking regenerator AIChE J 25 984

Errazu AF de Lasa HI and Sarti F 1979 A uidized bed catalyticcracking regenerator modelmdashgrid effects Can J Chem Eng 57 191ndash197

Ford WD Reineman RC Vasalos IA and Fahrig RJ 1976 Modelingcatalytic cracking regenerators NPRA Annual Meeting

Gruhn G Rosenkranz J Werther J and Toebermann JC 1997 Devel-opment of an object-oriented simulation system for complex solidsprocesses Comput Chem Eng 21(Suppl) S187ndashS192

Guigon P and Large JF 1984 Application of the Kunii-Levenspiel modelto a multi-stage baf ed catalytic cracking regenerator Chem Eng J 28131ndash138

Jacob SM Gross B Voltz SE and Weekman VM 1976 A lumpingand reaction scheme for catalytic cracking AIChE J 22(4) 701ndash713

Jones JH Braun WG Daubert TE and Allendorf HD 1967 Estima-tion of pressure drop for vertical pneumatic transport of solids AIChE J13 608

Kraemer DW and de Lasa HI 1988 Catalytic cracking of hydrocarbonsin a riser simulator Ind Eng Chem Res 27(11) 2002ndash2008

Kraemer DW Sedran U and de Lasa HI 1990 Catalyticcracking kinetics in a novel riser simulator Chem Eng Sci 45(8)2447ndash2452

Krishna AS and Parkin ES 1985 Modeling the regenerator in commer-cial uid catalytic cracking units Chem Eng Prog 81(4) 57ndash62

Kumar S Chadha A Gupta R and Sharma R 1995A process simulatorfor an integrated FCC-regenerator system Ind Eng Chem Res 343737ndash3748

Kunii D and Levenspiel O 1968 Bubbling bed model for kinetic processin uidized beds gas-solid mass and heat transfer and catalytic reactionsInd Eng Chem Proc Des Dev 7(4) 481ndash492

Kunii D and Levenspiel O 1990 Fluidized reactor models l for bubblingbeds of ne intermediate and large particles 2 for the lean phase Freeboard and fast uidization Ind Eng Chem Res 29 1226ndash1234

Larocca M Ng S and Lasa H 1990 Fast catalytic cracking of heavy gasoils modeling coke deactivation Ind Eng Chem Res 29 171ndash180

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

INDUSTRIAL EXPERIENCE WITH OBJECT-ORIENTED MODELLING FCC CASE STUDY 551

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al

Page 26: Industrial Experience with Object-Oriented Modelling FCC Case Study

Lee L Chen Y Huang T and Pan W 1989b Four-lump kinetic modelfor uid catalytic cracking process Can J Chem Eng 67 615ndash619

Lee E and Groves F 1985 Mathematical model of the uidized bedcatalytic cracking plant Trans Soc Comp Sim 2(3) 219ndash236

Lee W and Kugelman AM 1973 Number of steady-state operating pointsand local stability of open-loop uid catalytic cracker Ind Eng Chemprocess Des Dev 12(2) 197ndash204

Lee L Yu S Cheng C and Pan W 1989a Fluidized-bed catalystcracking regenerator modeling and analysis Chem Eng J 40 71ndash82

Lohmann B and Marquardt W 1996 On the systematization of processmodel develoment Comput Chem Eng 20(Suppl) S213ndashS218

Marquardt W 1991 Dynamic process simulation mdash recent progress andfuture challenges Chemical Process Control IV 131

Marquardt W 1992 An object-oriented representation of structuredprocess models Comput Chem Eng 16(Suppl) S329ndashS336

Marquardt W 1994 Computer-aided generation of chemical engineeringprocess models Int Chem Eng 34(1) 28ndash46

Marquardt W 1996 Trends in computer-aided process modeling ComputChem Eng 20(6=7) 591ndash609

McFarlane RC Reineman RC Bartee JF and Georgakis C 1993Dynamic simulation for model IV FCCU Comput Chem Eng 17(3)275ndash300

McGreavy C and Isles-Smith PC 1986 Modeling of uid catalyticcracker Trans Inst Meas Control 8(3) 130ndash136

Niemeyer P and Knudsen J 2000 Learning JAVA (OrsquoReilly amp AssociatesInc California USA)

Nilsson B 1993 Object-oriented modeling of chemical processesDoctoral Dissertation Lund Institute of Technology Sweden

Pantelides CC and Barton PI 1993 Equation-oriented dynamic simula-tion mdash current status and future perspectives Comput Chem Eng17(Suppl) S263ndashS285

Pantelides CC and Oh M 1996 Process modelling tools and theirapplications to particulate processes Powder Tech 87 13ndash20

Paraskos JA Shah YT McKinney JD and Carr NL 1976 Akinematic model for catalytic cracking in a transfer line reactor IndEng Chem Process Des Dev 15(1) 165ndash169

Perry RH and Green DW (eds) 1997 Perryrsquos Chemical EngineersrsquoHandbook (McGraw-Hill New York USA)

Piela PC Epperly TG Westerberg KM and Westerberg AW 1991Ascend an object-oriented environment for modeling and analysis themodeling language Comput Chem Eng 15 53ndash72

Quann RJ and Jaffe SB 1992 Structure-oriented lumping describingthe chemistry of complex hydrocarbon mixtures Ind Eng Chem Res 312483ndash2497

Quann RJ and Jaffe SB 1996 Building useful models of complexreaction systems in petroleum re ning Chem Eng Sci 51(10)1615ndash1653

Rowe PN and Partridge BA 1965 An x-ray study of bubbles in uidizedbeds Trans Inst Chem Eng 43 157

Shah YT Huling GT Paraskos JA and McKinney JD 1977 Akinematic model for an adiabatic transfer line catalytic cracking reactorInd Eng Chem Process Des Dev 16(1) 89ndash94

Smith JM and Van Ness HC 1987 Introduction to Chemical Engineer-ing Thermodynamics (McGraw-Hill New York USA)

Sriramulu S Sane S Agarwal PK and Mathews T 1996 Mathematicalmodelling of uidized bed combustion 1 Combustion of carbon inbubbling beds Fuel 75 1351ndash1362

Stephanopoulos G Henning G and Leone H 1990 MODELLA Alanguage for process engineering Part I and II Comput Chem Eng 14813ndash869

Takatsuka T Sato S Morimoto Y and Hashimoto H 1987 A reactionmodel for uidized-bed catalytic cracking of residual oil Int Chem Eng27 107ndash116

Toebermann JC Rosenkranz J Werther J and Gruhn G 2000 Block-oriented process simulation of solids processes Comput Chem Eng 231773ndash1782

Weekman VW and Nace DM 1970 Kinetics of catalytic crackingselectivity in xed moving and uid bed reactors AIChE J 16 397ndash404

Weisz PB and Goodwin RB 1966 Combusion of carbonaceous depositswithin porous catalyst particles II Intrinsic burning rate J Catal 6227ndash236

Yang WC 1973 Estimating the solid particle velocity in vertical pneumaticconveying lines Ind Eng Chem Fund 12(3) 349ndash352

Yang WC 1974 Correlations for solid friction factors in vertical andhorizontal pneumatic conveying AIChE J 20(3) 605ndash607

ACKNOWLEDGEMENTS

The authors wish to acknowledge the support provided by Invensys IndiaPrivate Limited (IIPL) Foxboro Division Chennai India throughout theduration of the project We wish to specially acknowledge and thankB Jayaram (Manager Advanced ApplicationsGroup) for extendinghis supportand cooperation in various model development activities during the project

The manuscript was received 14 February 2003 and accepted forpublication after revision 3 September 2003

Trans IChemE Part A Chemical Engineering Research and Design 2004 82(A4) 527ndash552

552 R MADHUSUDANA RAO et al