OSMOSE%Lua–A+Unified+Approach+to+Energy+ Systems ... · A variable is the smallest data structure handled by OSMOSE. A variable is deﬁned by 1 Name or tagname that litteraly identiﬁes

EPFL-‐SCI-‐STI-‐FM (IPESE) June 2015 ‹#›

OSMOSE-‐Lua – A Unified Approach to Energy Systems Integration with Life Cycle

AssessmentMin-‐Jung Yooa, Lindsay Lessardb, c, Maziar Kermanib, François Maréchalb

a EPFL, Laboratory for Computer-‐aided Design and Production (LICP), Lausanne, Switzerland b EPFL, Industrial Process & Energy Systems Engineering Group (IPESE), Lausanne, Switzerland

c Quantis, EPFL Innovation Park, Apt. D, 1015 Lausanne, Switzerland

PSE2015-‐ESCAPE25, Copenhagen, Denmark June 3, 2015


The system vision of the energy usage


Socio - ecologic - economic environment





Products &Energy services products

products

Demand

Price

t

t




Raw materials

Resources

Price

Availability

t

t


products

Demand

Price

t

t




Conversion systems

Raw materials

Resources

Price

Availability

t

t


products

Demand

Price

t

t




Conversion systems

Raw materials

Resources

Price

Availability

t

t


products

Demand

Price

t

t

Waste Emissions



Energy system design

Conversion systems

Raw materials

Resources

Price

Availability

t

t


products

Demand

Price

t

t

Waste Emissions


Problem definition

• Characterise and localise the technologies to be used to supply the energy services and the products needed over the lifetime of the system’s element at the time they are needed with the resources that are available • Decisions to be taken • Technology : choice, size and location • The operating strategy • The services that are provided • The resources that are used

• Criteria • Minimum cost = OPEX + APEX • Maximum profit = NPV (i,lifetime,cost) • Minimum environmental impact = LCIA • Max renewable energy integration

• Constraints • Mass & Energy Balances • Context specs • Bounds

Yu,l, Su,l

fu,l,t(p) 8p, u, lfu!s,l,t(p) 8p, u, l, sfr!u,l,t(p) 8p, u, l, r


The energy systems engineering methodology

Solutions

Energy services Resources Context & Constraints

Superstructure

System Boundaries

System performances indicators•Economic•Thermodynamic• Life cycle environmental impact

Results analysis•Exergy analysis•Composite curves•Sensitivity analysis•Multi-criteria

Technology options

Models

250

300

350

400

450

500

550

600

0 5000 10000 15000 20000 25000 30000

T(K

)

Q(kW)

Cold composite curve

Hot composite curveHeat & Mass integration

Decision variables

Solving method

Thermo-economic Pareto

Multi-objectiveOptimization

Out=F(In,P)


Energy systems integration : multi-‐scale problem

6

EnA / EnMS

Common element is the energy planning

• Energy consumption profile determination

• Energy performance evaluation

• Energy baseline generation

• Identification and evaluation of improvement options

• Implementation and monitoring

Step 1

Step 2

• Systematic approach

• Methodologies

• Tools

Context Literature review Problematic 1st year research Research plan Conclusion

Step 3

Details

Scale

Technologies

Processes

Industrial Clusters

Districts

Regions

* Model the interactions * Identify the synergies * Implement the interactions => Network or Grid

* Multi scale problem * Systematic framework

Materials


• Model of the interconnectivity (mass, heat, energy) • Model of the emissions (Equipment, Emissions) • Model of the cost (size => cost, maintenance)

The energy technology building block

-‐ Equipment sizing model -‐ Cost estimation

Heat transfer requirement

Heat transfer

Thermo-‐chemical conversion Model

Material streamsProduct streams

ElectricityHeat transfer requirement

Water streamsWater streams

Waste streamsWaste streams

ElectricityUnit parameters Decision Variables

Life cycle emissions

Life cycle of equipment -‐ Production -‐ Dismantling

MaintenanceInvestment cost

LCA models

LCA models


Energy technology approach

• Advantages • Expert developed model • Flowsheeting model and tool • Level of detail

• Independent of the energy systems integration • Model data base can be developed

• Difficulties • Remote access of models • Use of surrogate models

• Consistency of the models in the energy systems • from black box to white box

• Model data base • Confidence for reusing the models • Documentation • Certification of the model quality


Energy Technologies Encapsulation

Unit documentation DescriptionPurposeLimitationAuthors

Unit model report Summary of resultsImportant dataErrorsValidity

Unit execution handler Input dataCalculation engines

Existing flowsheeting softwareMimetic models

Output dataIntegration/interconnectivityCostingLCA => Ecoinvent LCI

Model ontologyModel encapsulation

Common interface of a module for system integration

Flowsheeting tools •BELSIM-VALI•gPROMS•ASPEN plus•HYSYS•Matlab•Simulink•(CITYSIM)•MODELICA•Others possible

•CAPE-OPEN ?•PROSIM

Unit connectivity Layers connections

Layer ontologyData base

ModelData base

KnowledgeData base


Ontology of a value in OSMOSE LUA

A variable is the smallest data structure handled by OSMOSE. A variable is defined by1 Name or tagname that litteraly identifies the variable 2 ID that is tagging the variable 3 Parent indicates the module to which the variable belongs 4 Value that defines one instance of the variable => Matrix

A priori value of the variable will exist for any time in the time sequence, for any period and for any scenario in which the system is calculated.

5 MinimumValue and a MaximumValue that defines the bounds for the value of the variables 6 DistributionType and DistributionParameters that defines the uncertainty of the value

7 DefaultValue that defines the value of the variable when it is created. this is also the recommended order of magnitude of the value recommended for the calculations

8 PhysicalUnit that defines the physical unit of the value of the variable 9 DefaultPhysicalUnits that defines the physical unit with which the calculation have to be made, It means that a

preprocessing phase will have to convert the value from the PhysicalUnits to the DefaultPhysicalUnits as well as a post-processing phase that is going to convert from the DefaultPhysicalUnits to the PhysicalUnits .

10 Status that identifies if the variable has a fixed value or a value that will be calculated. The status should identify as well if the value is calculated for each time or if it is constant or if it requires recalculation for each time. Status is equivalent to ISVIT in OSMOSE.i x : value is unique for the scenario (i.e. 1 value/scenario) ii x0 : value is period dependent (i.e. 1 value/period and / period)The value may calculated during

preprocessing phase using the model (i.e. constant that is period dependent).iii x00 : the value is time dependent (i.e. 1 value/time and period and scenario).

where x = 1 : for constant to be fixed by the user or the optimizer

2 : for constant fixed by the user or the optimizer otherwise use the default value -‐1 : for values calculated by the model

11 Dependence Matrix a dependence matrix that identifies the dependence of the value with respect to the other values of the problem is stored. Each of the dependence includes the mention if the dependence is linear, in which case the value of the dependence is stored in the field as a couple (ID,numerical value), or it can be stored as a reference to a non linear dependence that could be a surrogate model or the access to the model that calculates the dependence.

mu,l(t), Yi,u,l(t), Xi,u,l(t),⇧i,u,l(t),KPIk, m⇤u,l, Y

⇤i,u,l, X

⇤i,u,l,⇧

⇤i,u,l


Superstructure definitions

Luc Girardin page 25/77

Travail de diplôme : Méthodologie de conception optimale d'un réseau hydrogène d'un site industriel.

7.2 DÉFINITION DE LA SUPERSTRUCTURE

Le réseau d'hydrogène est conçu comme une superstructure qui contient toutes les sources et puits desunités productrices/consommatrices d'hydrogène, et qui incorpore toutes les connections possibles entre cessources et ces puits (Figure 7.2). Toutes les connections de la superstructure ne sont pas admissibles, cequi revient à introduire dans le modèle des contraintes sur l'établissement des nouvelles connexions.

Le point de fonctionnement de la structure est défini par l'ensemble des données requises pour chaqueconsommateur, chaque producteur, et chaque unité.

Chaque source de la structure est caractérisée par :

! la pureté de l'hydrogène produit Xp

! la pureté en impuretés produites x p , j

! un débit d'hydrogène disponible délivré "mp

! le niveau de pression total du flux délivré P p .

En contrepartie, chaque puits est défini par :

! la pureté requise du flux d'entrée Xc

! les limites de composition en impuretés du mélange admis xc , jmax

! un besoin de débit d'hydrogène "mc

! la pression totale requise du mélange à l'entrée Pc .

Le nombre des puits et sources, regroupé au sein d'une même unité, est égal au nombre de niveaux distinctsde pureté des flux d'hydrogène admis, respectivement rejetés. Dans un premier temps, on peut admettre quechaque unité possède au maximum 1 consommateur et 2 producteurs. Un nombre plus élevé de ceséléments par unité correspond à un affinement du modèle, et entraîne une augmentation de la densité deconnexions au sein de la superstructure

On a vu que les unités regroupent, dans une même structure, des puits et des sources qu'on appelleconsommateurs et producteurs.

Elles sont définies par :

! l'identité des puits et sources qui la compose,

! leurs coordonnées (X,Y) géo-référencées,

! un identifiant, qui permet de les regrouper en sous-systèmes pour l'analyse dans les diagrammes dePincement (cf. Chapitre "Représentation graphiques").

25/77 18/02/05

Figure 7.1 : Schéma de la superstructure complexe

250

300

350

400

450

500

550

600

0 5000 10000 15000 20000 25000 30000

T(K

)

Q(kW)


Hot composite curve

• Explicit • In flowsheeting software

• Automatic • Based on the connectivity description • Restricted matches • Routing

• Implicit • Heat cascades

• Combined • Mass and Energy


Geograhical informationAt a given location : combination of energy technologies


Geograhical information

Location i, (xi, yi, zi)

Location 1, (x1, y1, z1)


Geneva

At a given location : combination of energy technologies



• Coordinates attributed to locations • GIS data base => POSTGRE SQL • GIS Level of details

• Roads, district heating network, electricity grid • GIS Tools : Routing, Layers, Representations • GIS Data

• Buildings • Resources • Topology




Geneva







•Link between GIS data and Models •Link with unit models parameters •Access to GIS tools in the workflow




Geneva







•Link between GIS data and Models •Link with unit models parameters •Access to GIS tools in the workflow




Geneva

T TTT -20

0

20

40

60

80

0 50 100 150 200 250 300 350 400

T(C

)Q(kW)

AirWaste Water

Heating Hot water

recoveryRecoverable

Heat requirement



Energy system design : problem definition

Given a set of energy conversion technologies : Where to locate the energy conversion technologies ? How to connect the buildings ? How to operate the energy conversion technologies ?

1 Symbols

Roman lettersAn Annuities of a given investment [-]Ai,j,p Surface of the pipe of network p between nodes i and j [m2]B Arbitrarily big valueCaw Investment costs for air/water heat pump(s) [CHF]Cboiler Annual investment costs of the boiler(s) [CHF/year]Cfix Fix part of investment costs for a given device [CHF]Cgas Total gas costs [CHF]cgas Gas costs [CHF/kWh]Cgrid Total grid costs [CHF]cgrid Grid costs [CHF/kWh]C inv Investment costs of a given device [CHF]C inv

an Annual investment costs of a given device [CHF]Cpipes Total costs for the piping [CHF]Cprop Fix part of investment costs for a given device [CHF/kW]Cref Investment costs of a device chosen as reference [CHF]Cww Investment costs for water/water heat pump(s) [CHF]CO2

gas Total CO2 emissions for one year due to the combustion of gas [kg]co2

gas CO2 emissions due to the combustion of gas per kWh [kg/kWh]CO2

grid Total CO2 emissions for one year due to the consumption of electricity from the grid [kg]co2

grid CO2 emissions due to the consumption of electricity from the grid [kg/kWh]COP aw

t,k Coe⇤cient of performance for the air/water heat pump during period t at node k [-]COPww

t,k Coe⇤cient of performance for a water/water heat pump during period t at node k [-]cp Isobaric specific heat [kJ/(K· kg)]Disti,j Distance between nodes i and j [m]Ht Number of hours in period t [hour]Econs

t,k Electricity consumption during period t at node k [kW]Eexp

t,k Eletricity exported to the grid during period t by device e [kW]Egrid

t,k Electricity bought from the grid during period t [kW]Eaw

t,k Electricity consumed by the air/water heat pump during period t at node k [kW]Eww

t,k Electricity consumed by the water/water heat pump during period t at node k [kW]Eloss

t,k Electricity losses during period t [kW]Epump

t Pumping power for the network during period t [kW]Etech

t,k,e Electricity produced or consumed during period t at node k by device e [kW]Fm Maintenance factor [-]Gast Overall gas consumption during period t [kg]Lmin

e Minimum allowable part-load of device e [-]

Mbuildt,k

Mass flow of water from the network, circulating during period t through thebuilding at node k to heat up the building [kg/s]

Mpipet,i,j,p

Mass flow of water flowing during period t, in network p, from node i to nodej [kg/s]

Mmaxt,p,i,j

Maximum mass flow of water flowing in network p, from node i to node j, overall periods of time [kg/s]

M techt,k

Mass flow from the network being heated up by the device(s) at node k duringperiod t [kW]

MT buildt,k

Mass flow from the network flowing through a device at node k during periodt, to be re-heated, times its temperature [(kg/s)K]

2

MT pipet,i,j,p

Mass flow flowing during period t from node i to node j in the network p, timesits temperature [(kg/s)K]

N Expected lifetime for a given investment [year]Ploss Pressure losses in the pipes [Pa/m]Qaw

t,k Heat delivered by the air/water heat pump to the consumer at node k during period t [kW]Qboiler

t,k Heat delivered by the boiler at node k during period t [kW]Qcons

t,k Heat consumption at node k during period t [kW]Qnet

t,k Heat delivered by the network to the consumer at node k during period t [kW]Qnet ww

t,k Heat delivered by the network to the water/water heat pump at node k during period t [kW]Qtech

t,k,e Heat produced by device e located at node k during period k [kW]Qww

t,k Heat delivered by an water/water heat pump to the consumer at node k during period t [kW]r Interest rate [-]S Size of a given device [kW]Saw

k Nominal size of the air/water heat pump located at node k [kW]Sboiler

k Nominal size of the boiler at node k [kW]Snom

e Nominal power of a device [kW]Sref Size of a device chosen as reference [kW]Sww

k Nominal power of the water/water heat pump located at node k [kW]T atm Atmospheric temperature [K]T cold Temperature of the heat source for heat pumps [K]T cons

t,k Temperature at which the heat is required by the consumer at node k during period t [K]T hot Temperature of the heat sink for heat pumps [K]T net Design temperature of the network [K]v Velocity of the water through the pipes [m/s]X = 1 if a device exists, 0 otherwiseXaw

k = 1 if there is an air/water heat pump at node k, 0 otherwiseXboiler

k = 1 if there is a boiler at node k, 0 otherwiseXgas

e = 1 if device e needs gas to operate, 0 otherwiseXnode

k = 1 if a device e is implemented at node k, 0 otherwiseXtech

k,e = 1 if a device can be implemented at node k, 0 otherwiseXww

k = 1 if there is an water/water heat pump at node k, 0 otherwiseY i,j,p = 1 if a connection exists between i and j for network p , 0 otherwise

Greek letters�Theat Pinch at the heat-exchangers [K]�Tnet ww Temperature di⇥erence of the water in the network when it serves as heat source for water/water heat pump(s) [K]�boiler Thermal e⌅ciency of the boiler(s) [-]�ele Electric e⌅ciency of device e [-]�grid E⌅ciency of the grid [-]�the Thermal e⌅ciency of device e [-]⇥ Exergetic e⌅ciency [-]⇤ Density [kg/m3]

Indicesk nodesi, j connection from node i to node jt timee technologies

3

Networks superstructure

1 Symbols

Roman lettersAn Annuities for a given investment [-]Ai,j,p Area of the pipe between nodes i and j of network p [m2]B Arbitrarily big valueCaw Investment costs for air/water heat pump(s) [CHF]Cboiler Annual investment costs for the boiler(s) [CHF/year]Cfix Fix part of the investment costs for a given device [CHF]Cgas Total annual natural gas costs [CHF/year]cgas Natural gas costs [CHF/kWh]Cgrid Total annual grid costs [CHF/year]cgrid Grid costs [CHF/kWh]C inv Investment costs of a given device [CHF]C inv

an Annual investment costs of a given device [CHF/year]Cpipes Total annual costs for the piping [CHF/year]Cprop Fix part of the investment costs for a given device [CHF/kW]Cref Investment costs of a device chosen as reference [CHF]Cww Investment costs for water/water heat pump(s) [CHF]CO2

gas Total annual CO2 emissions due to the combustion of natural gas [kg/year]co2

gas CO2 emissions due to the combustion of natural gas [kg/kWh]CO2

grid Total annual CO2 emissions due to the consumption of electricity from the grid [kg/year]co2

grid CO2 emissions due to the consumption of electricity from the grid [kg/kWh]COPhp Coe⇤cient of performance of the central heat pumpCOP aw




t,e Eletricity exported to the grid during period t by device e [kW]Egrid

t,k Electricity bought from the grid during period t by node k [kW]Eaw



t Electricity losses during period t [kW]Epump


t,k,e Electricity produced or consumed during period t at node k by device e [kW]Fs Scaling factor [-]Fm Maintenance factor [-]Gast Natural gas consumption during period t [kg]Lmin

e Minimum allowable part-load of device e [-]Mbuild

t,k Water circulating during period t through the building at node k, to heat it up [kg/s]Mpipe

t,i,j,p Water flowing during period t, from node i to node j, in network p [kg/s]Mmax

i,j,p Maximum flow of water between nodes i and j, in network p, over all periods [kg/s]M tech

t,k Water being heated up by the device(s) during period t, at node k [kg/s]MT tech

t,k Water flowing through a device during period t at node k to be re-heated, times its temperature [(kg/s)K]

2

Q1..T

E1..T

Q1..T

E1..T

Q1..T

E1..T

Q1..T

E1..T

Q1..T

E1..TQ1..T

E1..T

Q1..T

E1..T

Q1..T

E1..T

Cinv =ne�

e=1

nk�

k=1

ae + ae � Se,k

Investment

Cgas =ne�

e=1

nk�

k=1

np�

t=1

HtQe,k,t

�the

1 Symbols























2

1 Symbols























2

Maintenance

Gas Grid

Electricity

Electricity

Emissions

1 Symbols























2

1 Symbols























2

Industry

Powerplant

Wastetreat.

Building plant.

. [5] Francois Marechal, Celine Weber, and Daniel Favrat. Multi-Objective Design and Optimisation of Urban Energy Systems, pages 39–81. Number ISBN: 978-3-527-31694-6. Wiley, 2008.

Water

Fuel

Water

FuelIndustrial plant.


Project data base

Predefined data

Project data base

OSMOSE LUA

Saved user dataEnergy technologies

Problems

ModulesDefault values

ThermodynamicSizing

Costing Life Cycle Inventory

EcoInvent

Work flow

ModelsSolving methods

in locations

GIS data base

GUI

Results data base

Layer Ontology

Superstructure

Yi,u,l(p(t)) = Fu(Xi,u,l(p(t)),⇧i,u,l(p(t)))


MILP process integration model @location k

Water: mii, Tin, hin

return : moo, Tout, hout

Water: mio, Tin, hin

return : moi, Tin, hin

Technologies w, @location k period s and time t(s)Demand @location k, period s and time t(s)

Storage tank

2

Tst,2

HEX h1(t)

heat demand

Storage tank

1

Tst,1

...

Tst,...

Storage tank

l

Tst,l

...

Tst,...

Storage tank

nl

Tst,nl

HEX h2(t) HEX h

..(t) HEX h

l(t) HEX h

nl-1(t)

heat demand heat demand heat demand heat demand

heat excess

HEX c1(t)

heat excess

HEX c2(t)

heat excess

HEX c..(t)

heat excess

HEX cl(t)

heat excess

HEX cnl-1

(t)

Units for heat integration

HL HL HL HL HL HL

HEX: Heat exchangers

HL: Heat losses

Storage system

S. Fazlollahi, G. Becker and F. Maréchal. Multi-Objectives, Mul.-Period Optimization of district energy systems: II- Daily thermal storage, in Computers & Chemical Engineering, 2013a

Buildings inertia

Heat exchange by heat cascade model

Electricity balance

Existence : w in location k


• Mass balance

• Mass/heat distribution

Layers models / automatic superstructure

nuX

u=1

k+u,l,t · fu,l,t(p) =nlX

j=1

eu,l,j,t(p) 8t(p) 8p8l

nuX

u=1

k�u,l,t · fu,l,t(p) =nlX

j=1

eu,j,l,t(p) 8t(p) 8p8l

nuX

u=1

nlX

l=1

k+u,l,t · fu,l,t(p) = 0 8t(p) 8p


Process integration models• MILP (Mixed Integer Linear Programming) models

• Integer => Selection/Usage • Continuous => Level of usage in t(period) • Parametrised (T)

• Discretisation • Decomposition

• Applications • Utility integration (Papoulias et al. , 1983, Marechal, et al., 1991)

• Combined Heat and Power + Steam network (Marechal et al. , 1997) • Refrigeration Cycle (Marechal et al., 2000) • ORC systems (Bendig, 2015)

• Multi-period (Marechal et al., 2003) • Multi period + Multi time (Weber et al., 2008) • Multi period + Multi time + Storage (Fazlollahi et al., 2012) • Multi period + Multi time + Multi-location+Storage (Fazlollahi et al., 2014) • Multi period + Multi time + Multi-location+Storage+Seasonal Storage (Rager et al., 2015) • Bio-refineries (Ensinas, 2015) • Industrial ecology (Gerber, 2013) • Restricted Matches (Becker et al., 2012) • Heat and Mass (Kermani et al., 2014) • Heat load distribution (Marechal et al., 1989) • Heat load Distribution multi-period (Mian et al. , 2014) • Heat load Distribution site scale (Pouransari et al., 2014) • Hydrogen Network (Girardin et al., 2006)

• MINLP (Mixed Integer Non Linear Programming) models • Heat Exchanger Network Design (Grossmann et al., Mian et al, 2014) • Heat Exchanger Network Design Multi-period (Mian et al., 2014)


Access to solvers

• MILP/MINLP solvers • Mathematical Programming Language • GLPK (MILP-Open source) • AMPL (MILP and MINLP - Proprietary)

• Solvers • LPSOLVE (Open Source) • Gurobi • CPLEX • BONMIN • Baron • NLPSOL

• Evolutionary algorithms • MOO (EPFL) • Dakota

• Parallelisation • MPI => grid computing


Computational platform -‐ OSMOSEevolutionary, multi-‐objective optimisation algorithm

(MINLP master problem)

Decision variables

performances

economic model

LCA models

LCI database (ecoinvent)

componentsunit processes elementary flows

energy-‐ and material-‐ flow models superstructure

detailed models: flowsheeting software

average models: LCI database

energy & mass integration (MILP slave sub-‐problem)

process integration software

supply chain synthesis

state variables

multi-‐time approach

Decision variables

background processes

process units

impact assessmentstate variables


OSMOSE : Computer Aided Platform



GIS data base Industrial ecologyUrban systems

GUI : Spreadsheets, Matlab

Data Structuring

Technology models data base Energy conversionSharing knowledge




•CAPE-OPEN ?•PROSIM•UNISIM ?

Energy technology data base •Data/models interfaces•Simulation•Process integration interface•Costing/LCIA performances•Reporting/documentation•Certified dev procedure

Modeling tools integration



Data Structuring








MILP/MINLP models Heat/mass integrationSub systems analysisSuperstructureHEN synthesis models

Optimal control models MILP/ AMPL or GLPKMulti-period problems

Sizing/costing data base LCIA database (ECOINVENT)

Process integration



Data Structuring








MILP/MINLP models Heat/mass integrationSub systems analysisSuperstructureHEN synthesis models

Optimal control models MILP/ AMPL or GLPKMulti-period problems

Sizing/costing data base LCIA database (ECOINVENT)

Process integration

Grid computing

Multi-objective optimisation Evolutionary - HybridProblem decompositionUncertainty

Decision support



Data Structuring



Results analysis

• Graphical representations • Composite curves • Sankey diagrams

• PDF reports • Sensitivity analysis • Pareto curves => data base of solutions • Decision variables values

• Uncertainty analysis => most probable optimal solutions

250

300

350

400

450

500

550

600

0 5000 10000 15000 20000 25000 30000

T(K

)

Q(kW)


Hot composite curve

64 66 68 70 72 74 76 78 80 82 84 86 88600

700

800

900

1000

1100

1200

1300

SNG efficiency equivalent εchem [%]

Spec

ific

inve

stm

ent

cost

cG

R [U

SD/k

W]

pressurisation

pressurisation, gas turbine integration,steam drying

& hot gas cleaning

hot gas cleaning & steam dryingwith heat cogeneration

453 K

513 K

5 %

35 %

1073 K

1173 K

573 K

873 K

1 bar

50 bar

573 K

673 K

573 K

673 K

0

1

40 bar

100 bar

623 K

823 K

323 K

523 K

SNG Outputminimal

maximal

min

max

min

max

min

max

Td,in Φw,out Tg Tg,p

pm Tm,in Tm,out yRankine

ps,p Ts,s Ts,u2

0 20 40 60 80 100 1200

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

SNG proccess configuration number

Pro

babi

lity

to b

e in

top

10 [−

]

Prod. costRes. profitabilityBM Break even

4.3. COMPARISON WITH CONVENTIONAL LCA 45

-500 %

-400 %

-300 %

-200 %

-100 %

0 %

100 %

200 %

scale-independent,conventional LCIA

(average lab/pilot tech.)

without cogenerationIntegrated industrial base case scenario (8 MWth)

with cogeneration

(a) Overall impacts

0

5

10

15

20

25

Remaining processes

Rape methyl ester

NOx emissions

Solid waste

Charcoal

Olivine

Infrastructure

Wood chips production

Electricity cons./prod.

Avoided NG extraction

Avoided CO2 emissions

UBP/

MJ w

ood

scale-independent,conventional LCIA

(average lab/pilot tech.)

harm

ful

bene

ficia

l

harm

ful

bene

ficia

l

harm

ful

bene

ficia

l

without cogenerationIntegrated industrial base case scenario (8 MWth)

with cogeneration

(b) Detailed contribution

Figure 4.3: Comparison of LCIA results obtained by conventional LCA and the proposedmethodology for Ecoscarcity06, single score, with detailed process contributions (Gerberet al. (2011a))

The differences between the conventional LCA and the base case scenario without and witha Rankine cycle are mainly due to the improved conversion efficiency obtained by applyingprocess design method, which has a direct influence on the quantity of SNG produced perunit of biomass. In the base case scenarios, the efficiency is considerably higher becausethe energy integration is optimized in the process design. As a direct consequence, the


Case study

PF(E)3 project: Calculation PlatForm for Energy Efficiency and Environmental optimisation

• Started in January 2013, duration of 30 months • Tool for design and decision making based on the notion of cost/benefit • Use of OSMOSE platform from EPFL’s IPESE group

EPFL, IFP Energies Nouvelles, IDEEL, EDF R&D, Armines CEP, …

Context of the project


Case study

A city of 100,000 habitants

Resource Price Units Uncertainty

Biomass (wood) 0.05 €/kWh ±30%

Biomass import 0.10 €/kWh ±30%

Natural gas 0.078 €/kWh ±30%

Electricity (grid) 0.20 €/kWh ±30%

Oil 1.05 €/kg ±40%

Diesel 1.80 €/kg ±40%

Petrol 1.88 €/kg ±40%

DME 0.65 €/lit ±50%

FT diesel 0.125 €/kWh ±50%

MeOH 0.323 €/lit ±50%

Services Value Units

Mobility 36.1 pkm/s

Electricity 22,000 kW

Hot water 24,210 kW

Space heating 117,623 kW

Wastewater 0.95 m3/s

Municipal waste 2.23 kg/s

Municipal organic waste 0.51 kg/s

Industry Size Units

Sawmill 100,000 m3 woodin/y

Greenhouse 12,000 tonnes tomatoes produced/y

Laundry service 20,000 tonnes dirty clothes washed/y

DME plant 50 MWthBM

FT plant 100 MWthBM

MeOH plant 50 MWthBM

Input data


Case study Models


Case study

Urban system

Models


Case study

Urban system

Biogas Diesel Electricity Heat

Mobility Natural gas (biogenic) Natural gas (fossil) Oil

Organic waste Petrol Solid waste Industry main product(s)

Wastewater Water Wood biomassUnit U

Required 1 Alternative 2 Alternative 3

Product 4 Product 5

Energy services

Waste treatment

Space heating

Domestic hot water

Electricity Transport

Municipal solid waste

Wastewater

Municipal organic waste

Models


Case study

Urban system






Product 4 Product 5

Energy services

Waste treatment

Space heating

Domestic hot water



Wastewater


Industries

Sawmill

Greenhouse

Industrial laundry

Biorefineries

Models


Case study

Urban system






Product 4 Product 5

Energy services

Waste treatment

Space heating

Domestic hot water



Wastewater


Industries

Sawmill

Greenhouse

Industrial laundry

Biorefineries

Available resourcesIndigenous resources

Imported resources

Water

Biomass (wood)

Natural gas

Diesel

Electricity import

Oil

Petrol

Imported biomass (wood)

Models


Case study

Urban system






Product 4 Product 5

Energy services

Waste treatment

Space heating

Domestic hot water



Wastewater


Conversion technologies

Boilers

Cars

Engines

Turbines

Dryers

Gasifiers

Biogas purification

Biometha-‐nation

Incineration

Wastewater treatment

Synthetic natural gas production

Industries

Sawmill

Greenhouse

Industrial laundry

Biorefineries

Available resourcesIndigenous resources

Imported resources

Water

Biomass (wood)

Natural gas

Diesel

Electricity import

Oil

Petrol

Imported biomass (wood)

Models


Case study User-‐interface


Business as usual (BAU)

• Space heating1 – Heat recovery from MSWI – 57% fuel oil, 18% natural gas, 14%

biomass (wood), 11% electricity

• Hot water – Assumed same as space heating

• Transport2 – 24% diesel, 76% petrol

• Industries – Biorefineries self-‐sufficient – Others ! heat from natural gas

• Electricity – From MSWI – From grid (UCTE)

Case study Assumptions


Business as usual (BAU)

• Space heating1 – Heat recovery from MSWI – 57% fuel oil, 18% natural gas, 14%

biomass (wood), 11% electricity

• Hot water – Assumed same as space heating

• Transport2 – 24% diesel, 76% petrol

• Industries – Biorefineries self-‐sufficient – Others ! heat from natural gas

• Electricity – From MSWI – From grid (UCTE)

Optimisation constraints

• Space heating and hot water • Possible heat recovery from MSWI,

WWTP, industries (refineries, laundry service), organic waste (OW) biogas

• Transport • 56% diesel / petrol • 44% non fossil fuel driven

• Industries • Heat recovery when possible

• Electricity • From MSWI, WWTP, OW biogas • Remainder from grid (UCTE)

Case study Assumptions


Case study Result of optimization

Monte Carlo simulation • Prices of resources are subjected to uncertainty • Finding the most resilient solutions

Probability of being the best solution in 800 Monte Carlo simulations



Monte Carlo simulation • Prices of resources are subjected to uncertainty • Finding the probability of being among the top 10 best solutions



Monte Carlo simulation • Prices of resources are subjected to uncertainty • Finding the probability of being among the top 10 best solutions


Case study

Total environmental impacts = 492,000 tons of CO2-eq/yr Total operating cost = 226.8 million CHF/yr

Result of optimization (one solution in the set)

The thicknesses represent the size of the connection in “kW” (except for connections to mobility which are scaled up to be visible) Each colour represents one type of connection, e.g. electricity, heat, natural gas, wood, mobility, …


Case study Climate change (tons CO2-‐eq/yr)

Electricity (grid)

Petrol

Diesel

Fuel oil

Natural gas

Biomass (import) for biorefineries

Biomass (import) for SNG producpon

Biomass (import) for sawmill

Biomass (indigenous) for sawmill

Climate change (tons CO2-‐eq/y)0 62'500 125'000 187'500 250'000

Oppmised systemBusiness as usual

Impo

rted

resources

Total

0 200'000 400'000 600'000 800'000


Case study Climate change (tons CO2-‐eq/yr)

Electricity (grid)

Petrol

Diesel

Fuel oil

Natural gas

Biomass (import) for biorefineries

Biomass (import) for SNG producpon

Biomass (import) for sawmill

Biomass (indigenous) for sawmill

Climate change (tons CO2-‐eq/y)0 62'500 125'000 187'500 250'000

Oppmised systemBusiness as usual

Impo

rted

resources

Total

0 200'000 400'000 600'000 800'000


Conclusions and future work

• Data management platform • Data bases • Technology Models => Ontology • Connectivity => Ontology • Geographical information system => SQL • Life Cycle Inventory => SQL • Costing

• Tools integration and handlers • Flowsheeting tools • Routing algorithms • Data analysis and processing • Clustering • Statistical analysis

• MILP problem formulation • Problem decomposition & methods • MILP/MINLP solvers


Conclusions

• Workflow organisation => list of tasks • Aggregation/disaggregation • Super-structure definition & models • Optimisation problems • Multi-objective • MI(N)LP

• Parallel computing • Decision support & uncertainty

• Applications • Process design • Process integration & HEN design • Site scale integration • Industrial Symbiosis development • Urban Energy Systems development • Industrial ecology • Region-City scale development • Autonomous systems (Ships, cars)


Thank you for your attention!


Appendices


A1. Life Cycle Impact Assessment (LCIA) in Osmose-‐Lua

Source: Gerber, L., Gassner, M., and Maréchal, F. (2011). Systematic integration of LCA in process systems design: Application to combined fuel and electricity production from lignocellulosic biomass. Computers & Chemical Engineering 35, 1265–1280.

Documents

OSMOSE%Lua–A+Unified+Approach+to+Energy+ Systems ... · A variable is the smallest data structure handled by OSMOSE. A variable is deﬁned by 1 Name or tagname that litteraly identiﬁes