A New Approach to an Integrated CAD Method for Surface Ship Design

m K. T. Tan and T. R Bligh

ABSTRACT The objective of this paper is to introduce a new design approach of an integrated CAD method for surface ship design. This new approach is called the Concurrent Inte- grated CAD (CICAD) System. Instead of following the traditional spiral approach, the proposed work establishes a concurrent computer-aided design procedure. As the design process proceeds, all analyses would be carried out concurrently on the progressing design, to provide an overview of its performance. The paper starts with a brief discussion about the challenges encountered by the present shipbuilding industry. This is to identlfy the needs of this industry in the foresee- able future. The authors point out the problems of the ship design process, and investigate the impacts and trends of applying CAD methods, in order to justify the new approach. Other innovations of the ship design process, like Simulation Based Design (SBD), Deci- sion-based Design (DBD), etc. are also discussed. The paper illustrates the conceptual and function models of CICAD. It demonstrates how the new approach is being developed, with yacht design as an example; and it estimates the significance of this concurrent approach. Further work is underway to refine the new model prior to the system development.

A New Approach to an Integrated CAD Method for Surface Ship Design

INTRODUCTION

t is a fact that the downsizing of the U.S. Navy budget has led to research to achieve improvements in affordability and operational effectiveness of Navy ship designs. With the contraction in defence spending, many U.S. shipbuilders are planning to enter the com-

mercial market, which is highly competitive (Bennett and Lamb, 1996). This means that the conventional naval ship design and construction process is be- coming less applicable. A sirmlar trend is occurring to British shipbuilders such as Yarrow Shipbuilders Limited, Vickers Shipbuildmg & Engineering Limited and Vosper Thomycroft Limited, where general defence budget constraints have inevitably led to reductions in available UK orders (Walker and Mc- Cluskey, 1996; W&MF, 1994). When ship construction is down, most preliminary designs are cancelled. Thus no shipbuilder can afford a large expenditure of time and money in preparing preliminary designs (Lyon and Mistree, 1985). The time available pre-contract has to be capitalised on, and as much progress made as possible towards a truly advanced design (Firth, 1995). Time spent getting the design right at an early stage will not only be rewarded financially, but will make the designer a leader in his field. For shipbuilders to survive, they must be able to develop new products in shorter times, at considerably less cost, and at globally accepted quahty levels. Since the capability to design and develop new products over short periods of time is of increasing importance, it is clear that a step-change in design efficiency is necessary so that shipbudders can increase both the quantity and quahty of the preliminary designs in a specified time frame. The preliminary designs make a significant contribution to the success or otherwise of the vessel (Duffy and MacCallum, 1989).

To date, many attempts have been made to improve the efficiency of shipyards. Most shipbuilders have committed themselves to an electronic design process based on solid models, and some have used the Integrated Product and Process Development (IPPD) principles to design and construct world-class warships (Keane and Tibbitts, 1996). A few contemporary concepts, such as Concurrent Engineering and Total System Approach, have been introduced to the shipbuilding industry. Many of these attempts are focused on improvements in the process of ship acquisition, where cost and time are being cut down by changing or re-engineering the way ships were designed and built. On the other hand, some seek advanced computer technologies to improve the affordability of naval ship designs. For instance, simulating the design models in a virtual environment to eliminate hard prototypes (Boudreaux, 19951, and replacing military standard computing equipment with commercial ones, i. e. distributed computing (Geary and Masters, 1995).

That these attempts would improve the shipyards’ efficiency is beyond ques- tion. We are interested in whether the preliminary design stage could be improved further. Even though the Navy procurement methodology vanes from time to time, it is crucial for a shipbuilder to be able to respond effectively to a customer’s need, and produce the best possible p re l ina ry designs, w i t h the

NAVAL ENGINEERS JOURNAL January 1998 35

A New Approach to an Integrated CAD Mefhod for Surface Ship Design

constraints of cost, schedule and performance. Today, computer technologies have been used extensively to model and analyse preliminary designs; these include In- tegrated Computer-aided Design systems that embrace Electronic Product Definition (EPD). However, the principle of existing systems captures the features of the Ship Design Spiral concept, which is sequential, and therefore suffers from a number of problems. (The Ship Design Spi- ral concept was introduced by Evans (1959), and enables ship design problems to be solved systematically.) Due to its very nature, the analyses of preliminary hull forms are performed in a sequential manner, which usually begins with hydrostatics, goes on to stability and ends with performance analyses. This is driven by the fact that each analysis requires certain information from the preceding ones. For example, a weight estimation ought to be conducted, in order to determine the position of the centre of gravity, then stability studies can be carried out by using this value to obtain the righting arm, =. The authors believe this is one of the sigmficant obstacles to producing a better design in a shorter time, as it is unable to provide the designers with an overview of the design, especially at an early stage. Therefore, this paper introduces a new design approach that eliminates the inefficiency of the Ship Design Spiral. It is developed by establishing a concurrent computer-aided design procedure. This new approach is called the Concurrent Integrated CAD (CICAD) System.

PROBLEMS OF THE SHIP DESIGN PROCESS The design and building of ships was wholly a craft until the middle of the eighteenth century, before science af- fected ship design appreciably (Rawson and Tupper, 1966). For the next two centuries, the designer started with a number of assumptions and worked through the rules, which only imperfectly modelled the real situation, to see if the design satisfied the requirements, specified before- hand (Evans, 1959). This design process was an iterative, ‘trial and error’ procedure, where the final result had to satisfy the requirements (Larsson and Eliasson, 1994). In 1959, Evans made a significant contribution to Visuahsing and modelliig the process of ship design, which is now known as the ‘Ship Design Spiral’ (Mistree et al., 1990; Miller, 1965). The purpose of this technique was to assist designers in organising the thought process, so as to enable ship design problems to be solved more effectively (Evans, 1959). While some refinements have been made over time, the principle remains unchanged. Nowadays, this spiral model has become a widely accepted approach for ship designs (Hassan and Thoben, 1992; Larsson and Eliasson, 1994). A design spiral for surface cargo ship designs, which was suggested by Evans, is shown in Fig- ure 1. This spiral consists of fourteen stages, where each stage corresponds to a design task. The design process

Gencrrl

Sectional area and waterline characteristics

Freehoard Flosdahle length Siahility

F I G U R E 1. The design spiral for surface cargo ship designs [Evans, 19591.

keeps on iterating all these stages until it converges towards the centre, the final solution. Different groupings and sequences are possible, depending on the background data available (Miller, 1965).

The major characteristic of the Ship Design Spiral concept is that the design process is sequential and iterative. A designer must satisfy a limited set of criteria at each design stage; however, having satisfied those, the designer goes on to the next stage with little idea of how good the design actually is. Therefore, the design is continuously altered by means of consecutive iterations untd it reasonably fulfils all the requirements. The iterative processes may be conceived as moving in a spiral fashion to a balanced conclusion with all features compatible (Miller, 1965). As each design stage depends on the output of the preceding one, the process is arranged in a sequential manner. This leads to a design philosophy that is based on a straightforward sequential and iterative procedure (Kroemker and Thoben, 1996). Moreover, for certain design stages, internal iterations are required, and this makes the design process even more time consuming. For example, in the hydrostatics and stability analyses, internal iterations are required to find the proper sinkage and trim when the hull heels at large angles (Larsson and Eliasson, 1994).

Surprisingly, it is not easy for practising designers and naval architects to justify the effectiveness of the Ship Design Spiral approach, because by and large, it has been exercised for decades, and has become the routine procedure, especially for preliminary designs. Moreover, the majority of naval architects’ design activities are adaptive in nature, where new designs are usually derived from basis ships (Birmingham and Smith, 1997). Thus, accu- mulated data from previous designs is used extensively for short-cut design procedures (Miller, 1965). In this case, there is no immutable order (Trower, 1992). The design decisions are made largely based on type ship data: however, Andrews considers such approach could not pro-

36 January 1998 NAVAL ENGINEERS JOURNAL

A New Approach to an Integrated CAD Method lor Surface Ship Design

vide the scope for creative synthesis of new solutions (Andrew, 1986). Experienced naval architects may occa- sionally skip stages, as they are able to make a good guess, based on a rule of thumb (Condylis, 1997; Simpson, 1997). Part of the art of the naval architect lies in making his guess realistic so that the final solution can be achieved more rapidly (Rawson and Tupper, 1966). These designers, however, have to comply with the design spiral, when subtle modification to the design takes place, or when they begin a new design from first principles.

Despite the general acceptance of the spiral model, a few academics and operations researchers began to ques- tion its effectiveness. A number of general surveys were taken, and the results show that this method, by its very nature, may obstruct the exploration of optimum design (Kroemker, 1997, Kroemker and Thoben, 1996; Milgram, 1993; Mistree et al., 1990, Archer and Marshall, 1988; Lyon and Mistree, 1985). Due to this sequential, iterative, tedious and time-consuming design process, most of the designers are still unable to achieve near optimal designs. This is because the sequential process requires a great deal of design time and thus designers have limited re- sources to explore many potential designs (Kroemker and Thoben, 1996). Moreover, the spiral approach is unable to provide the designers with an overview of the designs, especially at an early stage, because the design information emerges at each stage as they slowly work through the spiral. Consequentlx they are incapable of recognising the effects of modifications on the design quickly Thus it is unlikely to get the design right first the. Another draw- back to this is that designers may adversely alter other design characteristics while concentrating on a particular design aspect (Condylis, 1997; Simpson, 1997). AU of these arguments strongly suggest that the spiral design process appears to be a sigrufcant obstacle to produce a better product in a shorter time. As a matter of fact, the longer the design time, the less competitive are designers in bidding the contract (Keane and Tibbitts, 1996). The authors, therefore, advocate the development of a new approach to the preliminary design stage.

EXISTING CAD SYSTEMS FOR SHIP DESIGN The advent of computer technology, with its unprece- dented processing speed and memory capacity, led to the development of Computer-aided Design and Manufacturing (CADICAM) programs. The shipbuilding industry, like others, began to use this new technology from the late 1960s (Hays and McNatt, 1994). An overview of the evolution of shlpbddmg CAD/CAM systems is shown in Table 1.

Table 1 shows that these computer-based tools were transformed from big and expensive computing hardware to affordable ones. The advances in software engineering contributed to the interactive processes and graphical user interfaces. Besides, computer architecture was enhanced

to represent a better portrayal of the real situation. All of these positive changes have resulted in general acceptance of CAD/CAM systems in the shipbuilding industa and are gradually including medium and small shipyards.

Today, CAD/CAM is a key technology for ship design and construction (Hays and McNatt, 1994). Typical CAD programs for ship design are built around a geometry generation capability which allows the designer to create, on the computer screen, a mathematically defined hull surface. Furthermore, a number of distinct programs al- low the post-processing of these surfaces in the areas of analyses and predictions, such as hydrostatics and stability calculation, resistant and powering, structural finite element analysis, Velocity Prediction Programs (VPP), Computational Fluid Dynamics (CFD) etc. (Lee et al., 1992; Biran and Kantorowitz, 1986). In addition to that, a number of complex programs, such as outfitting, piping, electrical cable ways, ventilation and arrangement designs, are incorporated to lay down the foundation for CAM programs (NSRP 0476, 1997). These programs play an important role in avoiding design inconsistencies, and en- hancing construction effectiveness. Compared with the pre-computer era, these computer-based tools undoubt- edly offer a number of advantages, which include replacing time-consuming laborious tasks, improving design accu- racy, making changes easier, and increasing production effectiveness (Ross, 1995). They link up the engineering and architectural aspects of ship designs.

The application of these tools has had a great impact on ship designs. First, many of the design tasks, to which the designer once devoted a large portion of time, are carried out at great speed by the computer; therefore, at an earlier stage more information about the design is available (Excell, 1997). In other words, the designers may now be able to assess their conceptual ideas, and detect errors in good time. Moreover, as the laborious tasks have been removed, it opens up greater opportunities for intel- lectual exploration, and eases the process of making changes. Second, as today’s powerful hardware and software can effectively model and fair a hullform precisely, the manual lofling process can be e h a t e d (Hays and McNatt, 1994; Larsson and Eliasson, 1994). This not only saves time, but also ensures that the real design is built, since the manual lofting process inevitably changes the hull form slightly Third, by transferring the hull image into a visual program, it is possible to view a perspective 3-D plot of the hull. Compared with the manual approach, where only three standard views are employed, this vis- uahsation is very important, because hulls that look good in lines plan may look quite ugly in reality (Larsson and Eliasson, 1994). Moreover, the representation of hull forms in 3-D solid models permits the usage of simulation and virtual environments (Boudreaux, 1995). Fourth, computer networking allows a multi-user environment, which means design information can be shared amongst designers, either through a workstation or the Internet.


A New Approach lo an Integrated CAD Method for Surface Ship Design

The evolution of shipbuilding CAD/CAM systems [adapted from NSRP 0479, 19971. Periods Subjects

Year 1972-1 978

Year 1979-1 986

Year 1987-1996

Hardware Processes Data End users

Main frames Batch processes Individual files Biq shipyards

Mid i/M i n i computers Interactive processes independent databases Biq and medium shipyards

PCs, Local Area Network Interactive graphic processes integrated data base Biq, medium and small shipyards.

This forms a huge database that may provide new ideas or opportunities for designers, as well as ease the selec- tion of components such as engines, deck equipment and fittings (Brooker, 1997). Fifth, the data generated during the design process can be tailored in format and content so that they can support the ship production process (Ross, 1995). From an integrated CAD/CAM feedback program, potentially costly mistakes can be ironed out before manufacture. Thus, this results in a production- oriented design, which may reduce the cost and time of manufacturing (Excell, 1997). Sixth, the digitisation of design information, instead of traditional engineering drawings, promotes interdisciphary co-operation. Working to- gether, naval architects, marine engineers, electrical engineers and interior designers, can transfer and super- impose drawings, which reduces error, and avoids mis- understandings (Mulligan, 1997).

THE TREND TOWARDS INTEGRATED SYSTEMS Many believe that reusing data in other computerised tasks will maximise the benefits of computerisation, and this is called Integration (Hays and McNatt, 1994). Thus, they regard the integrated ship design program as a com- pelling concept, and one whose time has come (Ross, 1995; Hays and McNatt, 1994; Lee, et al., 1992). This notion is driven by economic aspects. As it involves an enormous sum of money to computerise any given task in the design and manufacturing process, it is ddficult to justify the computerisation, if that task is taken as a stand- alone item. This is because the reduction in time and cost required to do that single task may not recompense the expense and effort of computerisation. However, if the entire process were integrated by computerising each of the tasks, then the task that cannot be justified on its own often still has big payoffs, because it allows greater efficiency elsewhere in the process. This becomes obvious when the design database is able to be manipulated to derive the manufacturing tool paths. As a result, there is a defirute trend to integrate the computer-aided ship design programs (Ross, 1995), and ultimately to embrace computer-aided manufacturing programs (NSRP 0476, 1997). In this case, the ship design process may be enhanced through individual programs sharing their results

with each other, preferably from a common database, which may be extended to computer-aided manufacturing (CAM). In addition, the emergence of Electronic Product Definition (EPD) had enabled distributed organisations and their suppliers to define and build products in a collaborative environment.

A few state-of-the-art integrated ship design programs have been developed, the representative ones include HULLTECH, AutoSHIP System, FORAN, HICADEC, IMSA, TRIBON, NAPA and NAVSEA CAD-2. In general, these programs can be categorised into two groups, where it is either integrated among modules, or by means of a product model (Ross, 1995). Although some consider that an integrated system must incorporate a common database and use the same language and the same access procedures for all modules (Meizoso et al., 19941, not many achieve ths; integration among the modules of the ship design process is the most common one that does. It means that the various programs of design tasks are designed to communicate data with one another to at least some extent. The HULLTECH, AutoSHIP and IMSA are examples of this group. The rest, which are called product models, share an integrated database of various ship design programs, which means there is no need for data conversion among the modules. This is a more advanced level of integration: however, it suffers from inflexibility of future growth (Ross, 1995); recent developments incorporate object-oriented computing techniques which exhibit the concepts of abstraction, encapsulation and inheritance. This results in simpler structures, which are easier to maintain (Wu, 1994). The majority of the above programs support conceptual, preliminary and detail designs. Con- ceptual and preliminary designs involve a suite of ship design programs, such as hull form geometry, hydrostatics, stability, longitudinal strength, resistance and powering. Although they are integrated in terms of reusing or sharing data, they still capture the features of the Ship Design Spiral concept. Detail designs include outfitting, piping, electrical cable ways, ventilation and compartment design programs; these programs usually utilise 3-D product model databases, and relate to computer-aided manufacturing tools (NSRP 0476, 1997). The development trends of these programs include increasing user friendli- ness, featuring open architecture, expandmg the scope of the program, developing document management, and as-

3a January 1998 NAVAL ENGINEERS JOURNAL

A New Approach to an Inlegrated CAD Method for Surface Ship Design

signing data ownership (NSRP 0479, 1997). All of these endeavours are aiming towards a fully integrated CAD/ CAM system. The Design Production Integration Panel (SP-4) of the NSRP (National Shipbuilding Research Pro- gram), in its evaluation report submitted to the U. S. Navy, concluded that an integrated CAD/CAM is necessary for United States shipyards to become competitive with overseas yards (NSRP 0479, 1997). Moreover, the Marine Division of SENER Ingenieria y Sistemas (Sener- mar) also suggested that for a shipyard to survive, it is absolutely necessary to adopt an integrated CADICAM system (Meizoso et al., 1994).

This trend towards full integration seems to be a rev- olutionary concept for the shipbuilding industry, as it may reduce the lead-time and cost, as well as improve the quality of the products in general. Computervision Cor- poration anticipates that there should be savings of around 25% on shipbuilding costs (Pullin, 19971, and Senermar estimates the savings in elapsed time for ship production would be between 10-15% (Meizoso et al., 1994). Hence, there cannot be serious opposition to the notion that an integrated system will enhance the shipbuilding industry However, it would be interesting to find to what extend this system resolves the problems of the ship design process, and whether the remaining problems are of great significance.

The above integrated CAD approach, with its powerful computing capability, may minimise the problems of the ship design process, in terms of removing laborious tasks, making change easy and promoting a collaborative design environment. During the conceptual and preliminary design stages, however, designers literally experience the same design process as they did before the computer era, except that now they are provided with 3-D images and high-speed calculations. In principle, they still adhere to the sequential nature of the Ship Design Spiral. Thus, this approach suffers from the consequences of the spiral approach, and therefore, still does not resolve the problems of the ship design process.

The significance of these problems is evidenced by various observations of designers. For a given specification, designers will require approximately the same number of iterations to converge to an acceptable preliminary design, no matter whether the programs are integrated or not. This is because in either case, the designers lack an overview of the design as the design is adjusted on a trial and error basis at each design stage which might well adversely alter other design characteristics; this is very an- noying (Condylis, 1997; Simpson, 1997). In fact, th~s problem would be avoided if they were provided with the global view of the design throughout the design process. Com- pared with pre-integrated computer-based tools, the existing integrated ones are unable to assist designers to attain a solution much more rapidly, because the rate of converging towards a solution remains constant. From a design perspective, these shipbuilding CAD/CAM sys-

tems contribute little to developing a better design over a short period of time.

A BRIEF REVIEW OF OTHER APPROACHES TO SHIP DESIGN Besides the above mainstream development, a number of ship design process innovations are underway A brief review of these approaches would provide a better under- standing of the whole issue. First of all, the discussion would be insufficient without mentioning the introduction of Concurrent Engineering (CE) in the shipbuilding industry, which purports to be a solution to shorten delivery time, reduce ship prices, and improve shipbuilding quality (Bennett and Lamb, 1996). Extensive investigation had been carried out by the Industrial Engineering Panel (SP- 8) of the NSRP (National Shipbuilding Research Program) to produce a user’s guide and primer for CE applications to the U. S. shipbuilding industry, and subsequently to im- plement CE at a specific shipyard. Overall, CE is a philosophy, not a technology It uses a parallel rather than a sequential process for the different functional parts of the product development. This is accomplished through the use of cross-functional teams (Bennett and Lamb, 1996). CE is very much a team management issue, where most of the obstacles are cultural and organisational (Ranky, 1994). As the navy procurement methodology vanes from time to time, it is not easy to establish generic strategies for CE application. The application of CE itself does not make sense unless the shipbuilders have a backlog of ships to build (Bennett and Lamb, 1996). Many shipyards, therefore, are reluctance to fully participate in this subject, and this is ample evidence that there may be a long way to go if CE is to become a core application in the shipbuilding industry (Bennett and Lamb, 1996).

A research project, which was funded by the European Commission w i t h the BRITE-EURAM (Basic Research in Industrial Technologies for Europe/European Research on Advanced Material) program, proposed a structured method for an improvement of the ship pre-design process (Kroemker and Thoben, 1996). This is to help the Euro- pean shipbuilding industry to develop adequate strategies in order to attain a successful market position when sub- jected to Far Eastern aggressive competition (Hassan and Thoben, 1992). Interviews with engineers and designers from nine European shipyards revealed that the current pre-design practice is inefficient due to the lack of applying a clearly defined methodology (Kroemker and Thoben, 1996). This structured method, which is named the ArgoShip, reorganised the most common components of various ship designs, in a systematic fashion with a modular structure and hierarchical order. ArgoShip extends support for generating design history, in order to monitor the progress of a particular design path. It is currently tested and vahdated in an industrial environment (Kro-


A Mew Approach lo an lnfegraled CAD Method for Surface Ship Design

emker and Thoben, 1996). Meanwhile, a similar attempt was made by Chou and Benjamin to develop a systematic procedure in the form of an interactive Decision Support System incorporating artificial intelligence (AI) techniques (Chou and Benjamin, 1992). It provides a comprehensive range of facilities throughout the ship pre-design process.

Instead of using the structured method, some under- took a different approach, i. e. the Object-Oriented method, to improve the ship design process. The object- oriented computing techniques were proposed because of their ability to represent the features of a ship in a natural and intuitive manner (Carnduff and Gray, 1992). This method offers some genuine advantages such as flexibility, reusability and extensibility; however, it does not actually tackle the problems of the Ship Design Spiral.

The emergence of a design methodology that is based on simulations and virtual environment has led to the development of Simulation Based Design (SBD) for naval ships. The concept of SBD focuses on the premise that the ship is designed, constructed, tested, and maintained in a computer using an integrated product and process development process before full commitment to full scale design and construction (Boudreaux, 1995). This is to function as an enabler for Concurrent Engineering and Integrated Product and Process Development. Like others, the development of SBD is also stimulated by the downturn in defence budgets of the U. S. Navy, where significant changes in the ship design process are essential to enable a continued, affordable ship program for the future. By utilising 3-D solid models from a single database, SBD will enable designers to place human models in the virtual environment to test interaction and operability of the product models (Boudreaux, 1995). Consequently, the properties and behaviour of a ship, which are associated with real world analogies, would be analysed. To imple- ment SBD, the ship design process must be changed, and commercial software and hardware suppliers as well as other technology developers must adopt standards for product models to be displayed in virtual environment (Boudreaux, 1995).

As multiple-objective optimisation techniques have been improved over the years, some operations researchers began to look into the possibility of solving ship design problems by using optimisation models. Farrokh Mistree contributed to this field by applying non-linear goal pro- gramming. He, and other associates, developed a Deci- sion-based Design (DBD) approach for the preluninary design of ships (Lyon and Mistree, 1985; Mistree, e t al., 1990, Mistree, e t al., 1993). This approach, which em- ploys optimisation techniques and is assisted by computers, is capable of “finding an optimal solution” by spec- ifying naval operational requirements, desirable characteristics and system constraints. The multicriteria ship design optimisation model has been identified as a global optimisation problem and is usually solved using global optimisation techniques (Ray, e t al., 1995). These

techniques are largely based on type ship data, and they, therefore, might not provide the scope for creative synthesis of novel solutions.

All of the above attempts imply that the existing ship design processes are no longer competitive in the present market. Therefore, it is clear that they need to be improved. The downturn in defence budgets after the Cold War, and the growing competitiveness in the global shipbuilding market, have motivated a number of researchers, many of whom intend to improve the affordability of ship design and construction. The above brief review reveals that much effort has been concentrated on the improvements of the design process itself, rather than other tech- nical aspects. This is because it may have direct impact on the competitiveness of a shipyard, in which it appears to be the most effective way to attain a successful market position. To a certain extent, this has justified the importance of an effective design process in the shipbuilding industry

TOWARDS A NEW DESIGN APPROACH General As the existing approach does not resolve the problems of the ship design process, there is a need to establish a new approach to an integrated CAD system. The new approach must be able to provide the right information to the designers at the right time, in order to increase the possi- bilities of getting the design right the first time. This is to assist the designers to converge towards a solution quickly In other words, as the design process goes on, all analyses will be executed concurrently to provide an overview of the progressing design. By doing so, the authors are convinced that the ability of designers to design better products over shorter periods of time would be enhanced.

This approach is called the Concurrent Integrated CAD (CICAD) System. It will be modelled and analysed by utilising I-CAM Definition (IDEF) techniques, developed by the US Air Force to perform modelling activities in support of enterprise integration (KBSI, 1995). For the analysis of requirements, the IDEF technique, which is a structured approach, will be employed to capture a static view of the functions performed by designers. This is to establish the function model, and to explore the possibility of executing the design process concurrently Besides, it helps focus on a specific aspect within the system. Subsequently, the IDEF4 techque, which is an object-oriented approach, will be used in the system development process in order to incorporate desirable life cycle qualities such as modu- larity, maintainability and reusability A knowledge-based system will also be developed to support the generation of assumptions for concurrent design analyses.

Conceptual Model The realisation of CICAD is shown in the conceptual model

40 January 1998 NAVAL E N G I N E E R S J O U R N A L

A New Approach lo an lnlegraled CAD Melhod for Sutface Ship Design

(see Figure 2). This model illustrates the information flow within CICAD. It consists of four modules, which are Modelling, Analyses, Knowledge-based and Evaluation Modules.

After the design requirements have been developed, the user may begin to model a preliminary hull form by using the Modelling Module. Meanwhile, the Knowledge-based Module might prompt the user for additional information to generate preliminary assumptions. As the hull modelling process goes on, the Analyses Module would carry out analyses programs concurrently, based on the preliminary assumptions generated from the Knowledge-based Mod- ule. The results of the analyses would be made available to the user in real time, and therefore, reflect the performance of the progressing design. In this case, the user would be provided with an overview of the design at all times. From time to time, the user may alter the preliminary assumptions or requirements in the Knowledge- based Module. Design alternatives are stored in the Eval- uation Module, where they would be compared with the design requirements to determine the final design.

The Modelling Module consists of a geometry generation program that allows the user to create a mathematically defined hull surface. The Analyses Module integrates a number of analysis programs concurrently, and will be capable of analysing the progressing design at specified intervals. The Knowledge-based Module will generate preliminary assumptions to support the analyses of hull forms, and will arrange the design requirements for the Evaluation Module. The Evaluation Module will assess design alternatives by using a scaling procedure. This paper deals with the functional modelling of the first two modules, in order to estimate the significance of this concurrent approach. It also outlines how the Knowledge- based Module will play its role in assisting the Analyses Module.

Function Model The Modelling and Analyses Modules of CICAD are developed by remodelling the sequential ship design process in a concurrent manner. Since each task is highly depen-

Find k l g ”

F I G U R E 2. Conceptual Model of CICAD

NAVAL ENGINEERS JOURNAL January 1998

dent upon the output of the preceding ones, it is crucial, at this point, to examine to what extend the process can be overlapped, and to determine which parts of the process require the support of the Knowledge-based Module. This section establishes the function model, and demonstrates how the design process could be overlapped, with a sailing yacht design as an example. The yacht design process is selected because it captures the basic tenets of the Ship Design Spiral, and has a moderate complexity

To begin with, each design stage was modelled in a top- down fashion by adopting the IDEFO technique. IDEFO is a modelling technique that minimises the need for elabo- rate descriptive text, and its graphical presentation provides a clear representation of a complex aspect of organ- isation. The model is supported by the inputs, controls, outputs and mechanisms (ICOMs). As this t echque is designated to perform the process planning, it must be modified in order to represent the actual stages of the Ship Design Spiral. Two crucial modifications were made. First, a closed loop structure was incorporated to account for the nature of the iterations in the Ship Design Spiral. Second, the controls and mechanisms were omitted, because they do not play an important role in determining the interdependency and they might, otherwise, compli- cate the entire model.

The overview of the design model is shown in Figure 3. The top level is known as the ‘environment’, and is given a node number LO (Level 0), this is to identify the inputs and outputs that cross the boundary of the model. To clearly identify the subject and extent of the model, a ‘context’ diagram node number L n, (Level n,) is estab- lished. One level down is the ‘viewpoint’, node number L n,m, (Level n,m,); this is used to establish each aspect of the ‘context’.

The core diagram, LO (see Figure 4) illustrates the generic description of sailing yacht design stages, which comprise of requirement development (Ll), conceptual design (LZ), preliminary design (L3) and detail design (L4). The design requirements are developed in the first place; those requirements will then inspire the designers to generate many sketches in the conceptual design stage, where these initial solutions will be evaluated intuitively, in order to select suitable combinations. If the clients are satisfied with the combination of these potentially most promising solutions, then the preliminary design will be carried out to embody the ideas in a design model. The design is continuously adjusted to converge into a best possible solution. This is the stage where the Ship Design Spiral concept is employed extensively The result of the prelirmnary design will be used to bid the contract; therefore, it appears to be the most crucial stage of the design process. If he does succeed in receiving the contact, then the detail design wdl be conducted. This involves the refinement of the preliminary design, and converts it into manufacturing instructions. For the requirement development, conceptual design and preliminary design stages,

41


F I G U R E 3. Overview of the design model

there are three closed loops respectively The first two loops are intended to determine the viability of each corresponding stage, while the third loop decides the acquisition of the design contract.

The development of the new approach is focused on the preliminary design stage, and its schematic diagram, L3 is shown in Figure 5 (see Appenduc A for diagrams L1 and L2). The greatest chance of arriving at an optimum solution occurs during the preliminary design stage; however, this is the design stage that suffers most from the problems of the Ship Design Spiral. This is because the conceptual designs do not have much sigmficant meaning in term of performance, as they are usually in the form of sketches, which are approximate and not measurable. Therefore, to achieve a best possible design, it requires the repetitive and subtle modifications that occur throughout the preliminary design stage.

Diagram L3 (see Figure 5) represents a typical function model of CICAD. It’s information flow indicates that the existing process of sailing yacht design is inevitable sequential, and it’s order has been arranged in a fairly effective way This can be ascertained by conducting a network analysis, which is known as the design structure matrix

(DSM). The DSM was first proposed by D. Y Steward in 1981, as a graphical technique to express design procedure information in a matrix form (Smith and Eppinger, 1994). The design tasks are arranged in a square matrix in which each row, and its corresponding column, are identified with one of the tasks. Along each row, the marks indicate from which other tasks the given task requires input. Diagonal elements do not convey any meaning, since a task cannot depend upon its own output (Smith and Eppinger, 1994). The DSM for diagram L3 can therefore be constructed as follows (see Figure 6):

Figure 6 shows that the preliminary design process is an entirely sub-diagonal matrix, which means all the tasks can be completed sequentially without having to make any guess or assumption throughout the design process (Smith and Eppinger, 1994); thus, it appears to be a reasonably effective order. To a certain extend, thls has justified the spiral concept of Evans, where he was attempt- ing to assist designers in organising the thought process, so as to enable ship design problems to be solved more efficiently (Evans, 1959). The authors are convinced of Evans’ model in respect of organising the thought process, but do enthuse over a better design approach.

To remodel the preliminary design process, tasks C, D and F (see Figure 5) have been expanded into Level n,m,, and are L33, L34 and L36 respectively (see Appendix A). This is to provide an insight into elementary processes, in order to investigate the interdependency of each design task at a lower level. The diagrams L33, L34 and L36 show that certain design tasks do not actually require the output of the preceding one, but they do require some input from elsewhere. Moreover, a small number of tasks may even be carried out at once after the conceptual design. These two findings hmt at the possibility of remodelling the whole design process.

Assuming that the duration of task A is 6; task B is 2; task C is 2; task D is 3; task E is 2; task F is 3 and task G is 2 units of time (Tan and Bligh, 19971, and the intervals

I IN

F I G U R E 4. Diagram Level 0, the sailing yacht design stages

42 January 1998 NAVAL E N G I N E E R S J O U R N A L

A New Approach to an Integrated CAD Method lor Surface Ship Design

F I G U R E 5. Diagram L3, the preliminary design

to produce each output within a specified task are spread uniformly, a p r e h a r y reformation can be constructed as follows (see Figure 7):

The new model shows that the design time in this example has been reduced by up to 40%. Moreover, the right information becomes available to the designers at an average of 30% earlier than in the spiral approach.

A

B

C

D

E F

G

A B C D E F G

Q X Q x x o X X X Q X X 0 X x x o x x Q

F I G U R E 6. The DSM for diagram L3

NAVAL ENGINEERS JOURNAL January 1998

Preliminary Assumptions Design time is defined as the sum of computational and interactive times. Computational time is the time required by the computer to perform a specified task. Interactive time is the time where a program waits for input from the user; it consists of two groups, i.e. independent and de- pendent ones. The former is the time for the task that requires no input from other computing tasks, and the latter is the opposite (see Figure 8).

The Knowledge-based Module generates preliminary assumptions to support these interactive tasks. There- fore, the analyses would be carried out concurrently, by utilising the preliminary assumptions.

FUTURE WORK The function model needs to be investigated further prior to system development. It requires refinement to ensure the acquisition of an accurate description of the problem situation. Based on the function model, a knowledge- based system will be developed to support the generation of preliminary assumptions. Subsequently, the system development process will employ the IDEF4 technique to incorporate desirable life cycle qualities such as modular- ity, maintainability and reusability Finally, system verifi-

43


Ship Design Spiral Approach

Design 7 Tasks I

Remodelling 0 1 Ship Design Spirvl

[-I F l

I E l i €2 I I

I A l j A 2 I c I0 12.x1 Design

Timc I F I G U R E 7. Remodelling of preliminary design process

cation and validation will be carried out to determine whether the system is functioning properly.

CONCLUSION This paper identifies the need for a step-change in ship design efficiency This is crucial to ensure a continued, affordable ship program for the future. The paper deduces that existing integrated CAD systems might obstruct the exploration of optimum designs, and they do not profoundly resolve the problems of the ship design process. There- fore, a new design approach, the CICAD, which aims to overcome the sequential nature of the Ship Design Spiral, is proposed. CICAD resolves the problems by integrating

1 DESIGN TIME I

F I G U R E 8. The components of design time

44

the design process concurrently; this supply the right information to the designers at a much earlier time, which provides a better global view of the design. Thus, the authors are convinced that the capability of designers to design better products within a shorter period of time would be enhanced. 4-

References Andrews, D., 1986, An Integrated Approach to Ship Synthesis,

The Naval Architect, April 1986, RINA London, pp. 73-102. Archer, D. J., and G. Marshall, 1988, Preliminary Ship Design:

A Rational Approach using Microcomputers, Marine and Off- shore Computer Application, Computational Mechanics Publi- cations, Southampton, pp. 143-172.

Bennett, J. G., and T. Lamb, 1996, Concurrent Engineering: Application and Implementation for U. S. Shipbuilding, Journal of Ship Production, Vol. 12, No. 2, pp. 107-125.

Besant, C. B., and C. W K. Lui, 1986, Computer-Aided Design and Manufacture, Ellis Horwood Limited, UK.

Biran, A., and E. Kantorowitz, 1986, Ship Design System Inte- grated Around a Relational DataBase, International Confer- ence CADMO 86, pp. 85-94.

Birmingham, R. W., and T A. G. Smith, 1997, Interpreting De- sign Intent to Facilitate Automatic Hull Form Generation, Sixth International marine Design Conference, June 1997, Newcas- tle, UK.

Boudreaux, J. S., 1995, Naval Ships and Simulation Based De- sign, SNAME Transactions, Vol. 103, pp. 111-129.

Brooker, D., 1997, A Craft Industry Welcomes The Computer, CAD Desk, London, Vol. 7/2, pp. 50.

Carnduff, T. W., and W. A. Grax 1992, Object-Oriented Com- puting Techniques in Ship Design, IFIP Transactions B-Ap- plications in Technology, Elsevier Science Publishers, pp. 299- 312.

Chou, Y., and C. 0. Benjamin, 1992, An AI-based Decision

January 1998 NAVAL E N G I N E E R S J O U R N A L

A New Approach to an lnfegrafed CAD Method for Surface Ship Design

Support System for Naval Ship Design, Naval Engineers Jour- nal, May 1992, ASNE, pp. 156-165.

Colquhoun, G. J. , and R. W. Baines, 1991, A Generic IDEFO Model of Process Planning, International Journal of Production Research, Vol. 29, No. 11, pp. 2239-2252

Condylis, T., 1997, Industrial collaboration: C & S Yacht Design, Southampton, UK.

Delchambre, A., 1996, CAD Method for Industrial Assembly: Concurrent Design of Products, Equipment and Control Sys- tems, Wiley, New York.

Duffy, A. H. B., and K. J. MacCallum, 1989, Computer Repre- sentation of Numerical Expertise for Preliminary Ship Design, Marine Technology, Vol. 26, No. 4, Oct. 1989, pp. 289-302.

Evans, J. H., 1959, Basic Design Concepts, Naval Engineers Journal, November 1959, ASNE, pp. 671-678.

Excell, J. , 1997, Has CAD killed the designer?, Design Engi- neering, London, Vol. March 1997, pp. 25-26.

Firth, G. R., 1995, CAD/CAM Design and Manufacture, CA- DAP’ 95, RINA.

Geary, J. W , and M. W Masters, 1995, Investigating New Com- puting Technologies for Shipboard Combat Systems, Naval En- gineers Journal, Vol. 107, No. 3, ASNE, pp. 253-265.

Hassan, N. E. A., and K. Thoben, 1992, Practical Design of Ships and Mobile Units, Elsevier Applied Science, London,

Hays, B. N., and T R. McNatt, 1994, The Integrated, Com- puter-based Ship Design Environment, Marine Engineering: Design and Operation of Ships and Offshore Structures, Com- putational Mechanics Publications, Southampton, pp. 75-84.

KBSI, 1995, Infomtion Integration for Concurrent Engineering (IICE) IDEF4 Object-oriented Design Method Report, Knowl- edge Based Systems Inc., US.

Keane, R. G., and B. E Tibbitts, 1996, A Revolution in Warship Design: Navy-Industry Integrated Product Teams, Journal of Ship Production, Vol. 12, No. 4, pp. 254-268.

Kroemker, M., 1997, Personal communication: Bremen Institute of Industrial Technology and Applied Work Science (BIBA), Germany, mk@biba. uni-bremen. de.

Kroemkec M., and K. D. Thoben, 1996, Re-engineering the Ship Pre-design Process, Computers In Industry, Elsevier Applied Science, London, Vol. 31, pp. 143-153.

Larsson, L., and R. E. Eliasson, 1994, Principles of Yacht De- sign, Adlard Coles Nautical, London.

Lee, K. Y., S. W Suh, and S. Han, 1992, Practical Design of Ships and Mobile Units, Elsevier Applied Science, London,

Lyon, T M., and E Mistree, 1985, A Computer-Based Method for the Preliminary Design of Ships, Journal of Ship Research, Vol. 29, No. 4, pp. 251-269.

Meizoso, M., E. Martinez-Abarca, and L. Garcia, 1994, Updat- ing of Shipbuilding CAD/CAM Systems, Marine Engineering: Design and Operation of Ships and Offshore Structures, Com- putational Mechanics Publications, Southampton, pp. 35-44.

Milgram, J. H., 1993, Naval Architecture Technology Used in Winning the 1992 America’s Cup Match, SNAME Transac- tions, Vol. 101, pp. 399-436.

Miller, R. T, 1965, A Ship Design Process, Marine Technology, October 1965, SNAME, pp 339-359.

Mistree, E, W E Smith, and B. A. Bras, 1993, A Decision- based Approach to Concurrent Design, Concurrent Engineer- ing: Contemporary Issues and Modem Design Tools, Chapman & Hall, London, pp. 127-158.

Mistree, E, W. E Smith, B. A. Bras, J. K. Allen, andD. Muster, 1990, Decision-based Design: A Contemporary Paradigm for

VOI. 11, pp. 2.1284-2.1297.

Vol. 11, pp. 2.1269-2.1283.

Ship Design, SNAME Transactions, Vol. 98, pp. 565-597. Mulligan, H., 1997, Computed Aided Drawing, Engineers Jour-

nal, Ireland, pp. 25-29. NSRP 0476, 1997, Evaluation of Shipbuilding CAD/CAM System

(Phase I), The Design Production Integration Panel (SP-4), NSRP (National Shipbuilding Research Program), US.

NSRP 0479, 1997, Evaluation of Shipbuilding CAD/CAM System (Phase II), The Design Production Integration Panel (SP-4), NSRP (National Shipbuilding Research Program), US.

Pullin, J., 1997 Shipshape and Computer Fashion, CIM Focus, The Engineer, London, Vol. Feb. 1997, pp. 15-16.

Ranky, P. G., 1994, Concurrent/Simultaneous Engineering: Methods, Tools and Case Studies, CIMware Limited, Eng- land, pp. 23.

Rawson, K. J., and E. C. Tupper, 1966, Basic Ship Theory, Longman, London.

Ray, T, R. F! Gokam, and 0. I? Sha, 1995, A Global Optimisation Model for Ship Design, Computers in Industry, Vol. 26, Elsev- ier Science B. i!, pp. 175-192.

Ross, J. M., 1995, Integrated Ship Design and Its Role in En- hancing Ship Production, Journal of Ship Production, Vol. 11, No. 1, pp. 56-62.

Simpson, A., 1997, Industrial collaboration: Andrew Simpson & Associates, Poole, UK.

Smith, R. l?, and S. D. Eppinger, 1994, A Predictive Model of Sequential Iteration in Engineering Design, M. I.Ts Schools of Engineering and Management, Massachusetts.

Tan, K. T., and ‘I E Bligh, 1997, A Concurrent Integrated CAD (CICAD) Method for Sailing Yacht and Ship Design, Proceed- ings of the qth International Conference on Computer Inte- grated Manufacturing, Springer-Verlag Singapore Pte. Ltd.,

Trower, G., 1992, Yacht and Small Craft Design, Helmsman Books, UK.

W & MF, 1994, Building Ships with Precision: A Visit to Vosper Thomycroft, Welding & Metal Fabrication, June 1994, pp. 252- 256.

Walker, A., and J. McCluskey, 1996, Restructuring the Manufac- turing Process of Modem Warships, GEC Journal of Research,

Wu, B., 1994, Manufacturing Systems Design and Analysis,

PP. 873-882.

Vol. 13, NO. 2, pp. 123-131.

Chapman & Hall, London. ~~

Kok Thong G Y i s a doctoral student at the Cambridge University Engineering Department, United Kingdom. He received a Bachelor of Engineering Degree in Marine Technology from Universiti Teknologi Malaysia. He has industrial experience with a few sh$building companies such as Brooke Dockyard and Sabah Shipyard in Malaysia, and A Lumen Weflt in Gemny. I n 1995, he was awarded the Best Academic Award from Malay- sia joint branch of Rwal Institution of Naval Architects and Institute of Marine Engineers. Thomas Percival Bligh is the Director ofstudies in Engi- neering at Gonville & Caius College, University of Cambridge. He received his B Sc., M Sc. and Ph.D. from the University of Witwatersrand. I n 1972, he went to the University of Minnesota as an Assistant Profeesol: Then, he had been appointed Associ- ate Pro&ssar in the Mechanical Engineering Department of Massachusetts Institute of Ethnology. I n 1986 he joined Cambridge University Engineering Department. His research interests are in the design and p e $ m n c e prediction of multi- hulls, and in applying expert systems and artificial intelligence to CAD.


P New Approach lo an lnfegrated CAD Method for Surface Ship Design

APPENDIX A

Markef Client

TY Pe Objective

Location Rating Rule Class Crew Information

Manufacturing Metho Time Scale Cost Estimation

Diagram L1 Requirement Development

Displacementhngth

Diagram L2 Conceptual Design

46 January 1998 N A V A L ENGINEERS J O U R N A L


- BM,GM \

KG \ ~ Heeling

Moment

Bonjean Curve Displacement, V

4

’ Intact Stability

2

r I b

b

L b

VCB. LCB

’ Damage Stability

3

’ ’

Change of V per ft Trim Moment to Alter Tridinch C”,. C” Wetted Surface

c

I

Diagram L33 Hydrostatics

Bonjean rn Lines Curve Plan 7 h Cross Curve

Cross Curve 1

Static Stability Dynamic Stability

Damage GZ Curve Floodable Length

L

Diagram L34 Stability

NAVAL E N G I N E E R S J O U R N A L January 1998 47

A New Aooroach to an Inlewated CAD Method lor Surface Ship Design

tines Plan RM Curve

Propeller

Sail Area & COE

Location

1 t- Velocity Prediction

Diagram L36 Velocity Prediction

48 January 1998 NAVAL ENGINEERS JOURNAL

Documents

A New Approach to an Integrated CAD Method for Surface Ship Design