Context Dependent CAD Systems Evaluation: the FIAT Experience

Article published in 1997 in the Proceedings of the 30th ISATA

Dr I Bozza, Prof U Cugini, Dr F Folini,
Università degli Studi di Parma, Italy
Ing V Romagnoli, FIAT Auto Mains, Italy

Abstract

Today market competition imposes to the manufacturing companies to take advantage from CAD/CAM technological innovations as soon as they become market available and sufficiently mature, provided that costs and benefits have been properly evaluated. Thus, continuous evaluation activities tailored on the specific design and organizational contexts seem to belong to the future of each medium/large size manufacturing company.
The paper describes a comprehensive methodology for the context oriented evaluation of CAD technologies and systems, successfully adopted by FIAT. Some activities are preliminary to the first benchmarking phase: 1) Identification of an emerging CAD technology. 2) Set up a scenario showing the application of the identified CAD technology within the company. 3) Identification of the CAD functionalities to be evaluated. 4) Definition of the benchmark design sequence. 5) Set up the evaluation method. The benchmarking phase is performed on the market available systems implementing the identified CAD technology and consists in going through the design sequence and contextually scoring the effectiveness, efficiency and usability of the used CAD functionalities: the indicative technical evaluation produced is then used for screening the most promising systems. The second pilot project phase consists in introducing the screened systems in the company and evaluating how they face a real design problem in the real design context: the definitive technical evaluation produced is delivered to the company strategic department.

 Introduction

Today market competition imposes to manufacturing companies to deliver innovative, high quality products, shortening the time to market; this led the companies to reengineer their internal processes on the base of innovative design and production principles and technologies. In this perspective, innovations coming from CAD/CAM area are of great interest.

CAD systems have definitively abandoned their connotation of mere geometric modellers, provide functionalities for modeling design rules and procedures and are supplied or integrated with specific modules for the structural analysis, kinematics, simulation, etc. The trend is toward systems able to support the product life cycle, from the conceptual design to the manufacturing phase, coordinating and integrating modeling activities with functional and technological verifications and managing the heterogeneous data constituting the product. The CAD system is hence doomed to provide wider support to product modeling and to pierce deeply into the company processes and organization; the obvious implication is that it isn’t -and won’t become- a plug and play system and its adoption has to be carefully evaluated and previously designed within the company specific framework. Let’s consider, for example, one of the recent innovations in the CAD area: the parametric/associative CAD technology. The adoption of the parametric/associative CAD technology allows to reduce dramatically the time usually spent in modifying a part; the drawback, usually neglected, is that the designers have to “think parametric”, i.e. they have to design the part having its main features clearly in mind and following a predefined, well organized process. But, is that the working way of the designers of my company? How can they start to think parametric and how much does it cost in terms of process reorganization, design environment set up, etc.? And, of course, does this technology really serve the design requirements of my company? [1] These questions should be carefully evaluated, because the parametric/associative CAD technology, as many other emerging technologies, really provides great opportunities for shortening the time to market but has also strong implications on the company processes and organization [3].

The CAD market offer and trend should be also taken into due consideration: CAD technologies are implemented by a variety of systems having different capabilities and costs. Unfortunately, the ratio capabilities/cost is not a true indicator of the effectiveness of a system within a company. First of all, the systems capabilities should not be evaluated on brossures or demos, but rather on the company specific design problems. Moreover, the technical evaluations are not the only factor to take into consideration: the support provided by the vendor, the vendor representativeness on the CAD market (typically the number of installations) and its trend, etc. [2] are other crucial factors.

So, the aim of increasing the company overall performance by operating enhancements on the its CAD area presents two different yet interrelated sides: which technology? which system? The paper presents a comprehensive methodology enabling a medium/large size company to technically evaluate CAD technologies and systems through a first benchmark and a following pilot project. The methodology lends itself to a twofold use: on one side, it supports the company in the choice of its own CAD technology and system; on the other side, it provides tools for continuously monitoring the CAD technologies and market. At the end, the paper illustrates how FIAT adopted such methodology in a recent evaluation of the parametric/associative CAD technology and systems. 

 The Benchmark

The benchmark is organized in two phases: a preparation and an execution phase. The preparation phase consists in identifying an emerging CAD technology and producing the benchmark execution guidelines, tailored both on the company design processes and on the identified CAD technology. The execution phase consists in benchmarking the market available CAD systems implementing the identified technology in order to screen the most promising ones.

Preparation

The preparation is a five-steps phase, as detailed in the further.

Step 1: identify the emerging CAD technology

This step consists in analyzing jointly the CAD technologies state of the art and the market current offer in order to identify the emerging CAD technologies, understand their applicability, foundations and main features and their short/medium term availability on the CAD market (how many vendors already supply that technology? how many vendors plan to supply that technology and which is their timetable? etc.). This investigation will highlight the technology most appealing for the manufacturing field the company belongs to: it should be a settled technology, possibly supplied by representative CAD vendors.

This step should be performed by CAD specialists, either internal or external consultants.

Step 2: set up a scenario showing the application of the identified CAD technology within the company

This crucial step is aimed at sketching a scenario showing the application of the identified CAD technology within the company design processes. The work starts from a suitable model of the product design process and proceeds re-elaborating it on the base of the CAD technology basic principles identified in the previous step.

Both the starting and resulting model of the product design process shall clearly highlight: the design activities, eventually grouped/split into activities of higher/lower level of detail, the product and information flows and, for each activity, the criteria, rules and constraints used and the skill of the executor/executors. The IDEF0 technique [4, 5] can be of great support to this modeling activity: it allows to represent all the above information both graphically, through diagrams, and textually, through annotations, glossaries and explanation documents. Moreover, the simplicity of the IDEF0 graphical language, its capability of structuring the model into levels of details and the availability of a considerable set of commercial software tools supporting the technique, allow to build manageable models, suitable to be analyzed, re-elaborated and enriched even by non professionals, i.e. analysts as well as designers, CAD specialists, etc.

The starting model shall provide a functional description of the product design process abstract from the technologies currently available in the company: the modeling activities should not be detailed in terms of specific CAD functions; the data conversion and exchange activities should not be included; etc. If the starting model of the product design process is not available it shall be created; if, on the contrary, a model of the product design process already exists, it shall be analyzed and eventually re-arranged in order to meet the above requirement. 

The joint analysis of such model of the product design process and of the CAD technology basic principles should lead to identify design activities that can be automated, low-cost functional and technological verifications that can be performed, new methods for the product data sharing that can be applied, etc. Such information can be easily represented by re-arranging, detailing and enriching the starting model of the product design process. The resulting model will describe the support the identified CAD technology is expected to provide to the company design processes.

For complex products assembling technologically heterogeneous parts (e.g. machined parts, sheet metal parts, etc., as in the automotive industry), this step and the following ones shall be performed separately on parts representative of the different technological domains, provided that the connections among the various design processes are properly modeled. 

This step should be performed jointly by the company designers, at different levels of responsibility, CAD specialists and process analysts, either internal or external consultants.

Step 3: define the set of the CAD functionalities to be evaluated

This step is aimed at defining the set of the CAD functionalities to be evaluated. The followings are base CAD functionalities:

  1. base modeling, e.g. functionalities to handle curves, surfaces, solids, mixed dimensions models, fillets, etc.;
  2. interaction modalities and techniques, e.g. presentation techniques, interaction paradigm, uniformity and consistency of the user interface, etc.;
  3. product data management, e.g. functionalities to manage non-geometric data, to handle assemblies, etc.;
  4. customizability, e.g. functionalities to record and execute design sequences, to manage libraries of parts, etc.;
  5. integration with other tools (FEA, kinematics, documentation, etc.), either supplied or external.

Other functionalities are specific of the identified CAD technology.

This step should be performed by CAD specialists, either internal or external consultants.

Step 4: define the benchmark design sequence

This step is aimed at defining a meaningful design sequence to be used as guideline for the following execution phase. The work starts from the model of the design process produced in step two and proceeds simulating one of its possible executions, on given input data and parameters. The simulation report is a sequence of design operations whose free parameters (dimensions, offset values, radiuses, etc.) have been set; the simulation output is a product.

The simulation provides also the base for establishing the evaluation points of the CAD functionalities identified in step three: analyzing each design operation, its execution and effect on the product, it is possible to identify and record the set of the CAD functionalities involved, both directly and indirectly, in the execution of that operation.

In order to limit the amount of work yet ensuring the significance of the result, the design sequence and the resulting product can be simplified, provided that:

  • the simplified design sequence preserves the base structure of the real sequences;
  • the resulting simplified product preserves the main functional and technological features of the real products;
  • the simplified design sequence contains at least one non trivial occurrence of each design pattern typical of the real sequences (loop backs, modifications, functional and technological verifications, etc.);
  • the simplified design sequence provides at least one evaluation point for each CAD functionality identified in step three.

Typical simplifications can be: limit the number of the repeated modification operations by considering them only once, perhaps in an advanced stage of the design; limit the number of the similar product features by considering only the most recurring and/or most complex ones; etc.

This step should be performed jointly by the company designers, at different levels of responsibility, CAD specialists and process analysts, either internal or external consultants.

Step 5: set up the method for the technical evaluation of the systems implementing the identified CAD technology

The following three-steps method allows to technically evaluate the support provided by a system implementing the identified CAD technology.

Let: S a system implementing the identified CAD technology; F=fnn=1N the set of the CAD functionalities to be evaluated and R=1,L
the evaluation range; DS=omm=1M the design sequence composed of M design operations; Fm=fkmk=1Km, FmÍF the set of the CAD functionalities involved in the design operation
om, omÎDS.

The evaluation of the support provided by the system S proceeds as follows:

1. Evaluation of a CAD functionality on the system S A CAD functionality is evaluated within the different design operations involving it. The evaluation of the CAD functionality fkm, fkmÎFm, involved in the design operation om consists in executing o1,…,om on the system S and producing the score sfkm, sfkmÎR, measuring the effectiveness, the efficiency and the usability of fkm within the design operation om executed on S.

2. Evaluation of the support provided by the system S to a design operation Let:

WFm=wfkmk=1Km, k=1Kmwfkm=1the set of weights associated to the CAD functionalities involved in the design operation om and measuring their relative importance within om.

The support provided by the system S to the design operation om is evaluated as follows:

sm=k=1Kmwfkmsfkm

3. Evaluation of the support provided by the system S to the design sequence Let:

WDS=wmm=1M, m=1Mwm=1the set of weights associated to the design operations of the sequence DS and measuring their relative importance within DS.

The support provided by the system S to the design sequence DS is evaluated as follows:

SDS=m=1Mwmsm

The adoption of this technical evaluation method requires to establish in advance the sets of valuesR, WFm and WDS. 

This step should be performed jointly by the company designers, at different levels of responsibility, and CAD specialists, either internal or external consultants.

Execution

The execution phase is aimed at producing an indicative technical evaluation of the market available systems implementing the identified CAD technology. For each system, the work consists in executing the design sequence and step by step evaluating the system performance using the above method. In order to take the full advantage from this work, the system vendor is required to collaborate with a selected group of designers and CAD specialists in charge of the evaluation: the vendor, able to use the system at its best, directly execute the design sequence while the company designers and the CAD specialists carefully observe and evaluate. 

For complex products, the use of a simplified yet significant design sequence allows to shorten the work to two/three working days per system, thus reducing the costs of this phase.

The indicative technical evaluations produced are then used for screening the two/three most promising systems.

 The Pilot Project

The pilot project is aimed at producing a definitive technical evaluation of the systems screened in the benchmark. The work consists in introducing the systems in the company and letting them face a real design problem in the real design context: the data to handle, in terms of amount and complexity, are the real ones and the systems users are a group of designers previously trained. The evaluation is performed by the designers working on the different systems: they periodically meet, report on their development status and on the problems and opportunities they experienced, discuss and finally agree on a common evaluation of the systems.

 The FIAT Experience

The present section illustrates how FIAT recently adopted the described methodology for evaluating the parametric/associative CAD technology and systems. 

The choice of investigating the parametric/associative CAD technology was quite obvious considering that FIAT estimated that the a considerable portion of its total design time was devoted to modification operations and that the specific target of the parametric/associative CAD technology is to automate, at least partially, this kind of operation. 

The first benchmark phase was promoted within the body design area and focused on the hatchback door design because its functional and technological features and its design process had been considered to be representative of the body parts design area. 

The work took as base the FIAT internal documentation detailing the hatchback door dimensioning rules and design criteria. This documentation was abridged with the functional model of the FIAT hatchback door design process, produced in accordance with the IDEF0 methodology: a set of meetings between a group of process analysts and FIAT body designers was organized and the elicited information was step by step arranged into an IDEF0 model complete with annotations, glossaries and explanation documents.

The IDEF0 model was then easily re-elaborated on the base of the parametric/associative CAD technology basic principles and a new IDEF0 model was produced: fig. 1 shows the backbone of the hatchback door design process (IDEF0 node A0) and fig. 2 details the props housings design process (IDEF0 node A222).

Fig. 1 The backbone of the hatchback door design process

In parallel, the set of the CAD functionalities to be evaluated was identified. In addition to the base CAD functionalities listed previously, the set contained functionalities specific of the parametric/associative CAD technology, e.g. functionalities to impose constraints on 2D/3D geometries, functionalities to handle under-constrained parametric models, functionalities to edit sets of constraints, etc.

Fig. 2 The props housings design process

A reduced execution of the hatchback door design process was simulated on real input data and parameters and a simplified design sequence, consisting of eighty-five design operations, was produced; fig. 3 describes the design subsequence produced by simulating the execution of the props housings design process (see fig. 2).

Fig. 3 The props housings design sequence

During the simulation, the set of the evaluation points of the CAD functionalities was established. For example, it was established that the design operation 26 Props choice and positioning (see fig. 2) provided an evaluation point for the functionalities to manage libraries of parts, e.g. open and query a library of parts, instantiate a part from a library, etc.

Finally, the sets of values R, WFm and WDS were set: the evaluation range of the CAD functionalities was set to R=1,3 and WFm and WDS were calculated on the base of FIAT internal statistics.

The benchmark was executed on a set of high-range parametric/associative CAD systems; it took two working days per system and involved the systems vendors, CAD specialists and experienced designers. 

The pilot project proceeded on the systems screened in the benchmark and extended to all the FIAT design areas, i.e. part design areas as well as equipment design areas.

Acknowledgments

The authors were involved in the FIAT benchmark and would like to thank the people from ELASIS, FIAT, Università di Parma and Università di Napoli for their cooperation.

References

[1] Gary Magoon. CAD/CAM connection: beyond the technology hype. Computer Graphics Word, December 1995, pp. 19-20.

[2] Caren D. Potter. CAD/CAM special report: CAD contracts won and lost. Computer Graphics Word, June 1996.

[3] Caren D. Potter. Special CAD/CAM supplement: best practices in solid modeling. Computer Graphics Word, November 1995.

[4] François B. Vernadat. Enterprise Modeling and Integration: principles and applications. Chapman & Hall, London UK, 1996.

[5] Integration definition for function modeling (IDEF0). Draft Federal Information Processing Standard Publication 183 (FIPAPUB183). FIPA, USA, 1993.