Process Modeling and Analysis in Practice: The Ferrari Experience

Article written with I. Bozza, U. Cugini, M. Lamperini and G Stefani. Published in June 1997 on 30th ISATA, Conference on Mechatronics and Automotive Electronics, Firenze, Italy – June 16-19 1997. ISBN 094771987

Abstract

The need to remain competitive in a fickle, ever growing market, forces the automotive companies to internally quantify the costs and benefits coming from new production criteria and/or technologies and to quickly design their practical introduction. This leads to a continuous evaluation, rationalization, re-arrangement and re-engineering of the design and production processes: process modeling and analysis techniques can provide the necessary support to these activities. In this area, Ferrari promoted a pilot project focused on the design of a meaningful automobile component. First, a functional model of the component design process, AS it IS currently performed, has been produced by means of iterated interviews of the Ferrari design experts. The model, composed of a structured set of IDEF0 diagrams and associated glossary and textual annotations, immediately revealed its effectiveness on documenting the process, thus providing a consistent base for the following analysis and improvement phases. A set of possible enhancements has been individuated and, step by step, designed so that the model of a new design process, as it has TO BE performed in the future, has been produced. Finally, in order to put this new process into practice, a comparative analysis of the two models has been conducted and a set of organizational issues have been highlighted and submitted to the Ferrari management.

 Introduction

Companies from the automotive industry are facing an increasingly competitive market. Even a company working in the top range market such as Ferrari S.p.A., has to internally quantify the costs and benefits coming from new production criteria and strategic technologies and, when they can provide real advantages, quickly design their practical introduction. This implies a continuous evaluation, rationalization, re-arrangement and re-engineering of the design and production processes: process modeling and analysis techniques can provide the necessary support to these activities. In this area, Ferrari S.p.A. and the University of Parma jointly promoted a pilot project aimed at experimenting the use of consolidated process modeling and analysis methodologies in an industrial problem. 

The Ferrari technical management was looking for something able to provide a complete and precise map of the activities, resources, data and control flows involved in the design and production processes of the vehicle. The project involved designers, production engineers and managers from Ferrari and analysts from the University of Parma, and allowed to evaluate:

1. the kind of support that process modeling and analysis can provide to the management;
2. the time and the resources required to produce a reliable and trustworthy model of an industrial problem and to analyze it;
3. the level of collaboration that can be obtained from designers and production engineers, usually uninterested or without time to invest in this kind of activities.

In order to perform the modeling and analysis activities, the choice of an appropriate methodology was required. On the base of the previous experiences gained at the University of Parma, the IDEF0 methodology was chosen because it had proven to be effective, it includes a graphical language generally appreciated by people involved in the process modeling and analysis activities and it is supported by a considerable number of commercial software tools.

 The choice of the study case

The first step of the project consisted in the choice of the study case among the components of a typical Ferrari sport car.

The following criteria were used for the choice of the study case:

• Its design process shall be representative of the one of others components of a typical vehicle; this ensures a general validity to the results of the experience.
• It shall be a component designed within the company and produced directly by Ferrari or under the company direct control; this allows a straight and simple interaction with designers, production engineers and other experts involved in the design and production processes.
• It shall be a component critical enough to receive from and to impose to the others vehicle components heterogeneous constraints, such as dimensional and positional constraints, characteristics, etc.; this allows to model and analyze the interdependencies between the component and the vehicle design processes.
• Its design process shall receive partially or fully defined constraints from the surrounding parts, at different stages; this ensures that the selected component is representative also from the “concurrent engineering” point of view.

After a short examination, the Ferrari design staff suggested to focus on the design and production processes of the fuel tank. As in any other vehicle, the fuel tank is designed to hold the petrol for the engine; in a typical Ferrari configuration, it is placed within the motor space and behind the passengers compartment.

The tank design is subject to many functional and mechanical constraints and must satisfy a number of specifications. In particular, in the recent years, the adoption of stricter international regulations and the need for extra space in the cockpit, forced the company to rethink this part and its design process.

The main constraints for the fuel tank, as described by the Ferrari design staff, are:

• It shall hold enough petrol to ensure 300350 Km of economy to the vehicle.
• It shall fit in the motor space, according to all the functional, structural and dimensional requirements.
• It shall satisfy -in the basic model or in some special versions- all the national and regional regulations for the automotive industry, e.g., CEE 91/441 and USA 40CFR par 86. This constraint applies for each nation in which the vehicle will be registered and exported.
• It shall be easily accessible in order to reduce the vehicle maintenance costs.
• It shall adopt some kind of protection from the thermal radiency of the engine in order to ensure that the fuel temperature is always lower than 60°-70° C.
• It shall satisfy the fuel evaporation limits, as defined by national laws and regulations.
• It shall support strong accelerations and decelerations, typical of a sport car, without compromising the functions of the evaporation and fuel transport subsystems.
• It shall ensure safety in crash conditions and, in particular, it shall be protected from lateral collisions.
• It shall be produced with a technology convenient for the small series, as usual for this category of vehicles.

The following fuel tank specifications were extracted from the technical sheets of the latest Ferrari model:

• Shape: one big tank behind the cockpit placed in central position or two lateral tanks with some external protection.
• Capacity: 120 liters.
• Production technology: sheet metal bending.

In the past, as described by the Ferrari designers, this component was usually designed at the end of the vehicle design process, when all the others components were almost completely defined. This approach has been recently abandoned: the need to fulfill stronger requirements on the cockpit internal space and on the fuel evaporation control, reduced the degrees of freedom of the fuel tank design. Therefore, the tank design has been anticipated in order to allow the evaluation of different solutions and, eventually, the imposition of conditions on the chassis and on the surroundings parts. In particular, the relevant dimensions of the fuel thank and the presence of two big canisters for the elimination of the petrol vapor impose the preliminary evaluation of the volumes in the motor space.

 The “As-Is” modeling phase

The fuel tank and its design and production processes change continuously for a number of reasons. The first reason is that its functional requirements are becoming stronger and stronger, as previously highlighted; others reasons are the company organizational changes and the growing and evolving experience of the engineers staff. The tentative to extrapolate a general design process from the history of the recent Ferrari models was discarded because it was noticed that the processes differed significantly. Therefore, it was decided to focus on the design process of the fuel tank of a specific Ferrari model: the last one completed was chosen.

As a preliminary step, the analysts organized a meeting with the technical management in order to identify the Ferrari people involved and their professional roles.

The activities started with interviewing the designers and the production engineers about the main design activities, information flows, etc. From the interview minutes, a first high level tentative model of the process was produced. Then, the first approach consisted in showing the preliminary IDEF0 diagrams to the interviewed people in order to obtain the feedbacks necessary to refine the model: the original plan was to iterate the discussion-refinement process until a consolidated and approved model was produced. This tentative generated however some problems. Unavoidably, each person, designer or production engineer, describes the process from his personal point of view with differences, not so obvious, with the descriptions obtained from other personal points of view: it is in charge to the analysts to reconcile the different descriptions in order to converge to a single model. As the model reached deeper levels of details, the analysts had to assume a super partes position and show to each expert only the set of IDEF0 diagrams of his competence.

In this phase of intensive interaction with the people, the IDEF0 graphical formalism was effective; its simplicity and descriptive power were appreciated by everyone. Usually designers and production engineers don’t have time to spend in reading long textual documents and prefer to analyze and discuss immediate and synthetic graphical representations, such as the IDEF0 one.

It was noticed that the representation immediately influenced the way the designers and production engineers see the process. For example, when the modelization highlighted an inconsistent procedure, with some unnecessary serialization of activities, the people unconsciously updated their mental model of the process and asked for a corresponding change in the IDEF0 representation. While this is the first benefit of such kind of activity, the analysts tried to control it in order to get a real “As-Is” model of the process.

The result of this activity was a complete and readable IDEF0 model describing the design and production processes of the fuel tank of a typical Ferrari model. The IDEF0 model was integrated with a glossary defining all the terms used, according the definitions provided by the Ferrari people. Besides, the list of process improvement suggested by the designers and production engineers during the interviews was collected. The final document released didn’t have any correspondent or equivalent among the Ferrari internal documentation and constituted an effective base for the analysis of the current process and for the definition of future evolution and scenarios.

 The “To-Be” modeling phase

At the beginning of the second phase, the analysts discussed the documents released in the “As-Is” modeling phase with the technical management in order to define some preliminary rationalization and re-engineering hypothesis. A few enhancements of the process were suggested directly by the design and production experts during the interviews and the discussion of diagrams, while many others came directly from the IDEF0 diagrams analysis. An IDEF0 model is a functional description of a process with a clear hierarchical structure that facilitates the analysis of the process, the individuation of critical activities and the design of the reorganization and optimization initiatives. For example, when an IDEF0 diagram doesn’t show a critical dependence among two or more boxes representing activities, it becomes evident that they can be reorganized in order to be performed concurrently, according to the “concurrent engineering” principles.

On the base of the analysis results, the technical management defined the main hypothesis for the future evolution of the design and production processes, such as the adoption of new design methodologies and the acquisition and use of new software tools. In particular, the main target of the management was to reorganize the design and production processes in order to anticipate all the decisions and the verification on the component. This verification, as described in the “As-Is” model, were previously performed only at the end of the design phase, directly on the physical mock-up. In order to reach such objective, a preliminary design of the fuel tank should be done in the firsts design stages of the vehicle. Besides, many physical characteristics of the fuel tank should be simulated with computer programs but without abandoning the traditional verification on the physical mock-up. Starting from these inputs and under the direct control of the technical management, the analysts reorganized the previous IDEF0 model in order to obtain the “To-Be” model. In this phase, as in the previous one, the analysts had to iterate the discussion and diagram update activities many times.

In the transition from the “As-is” to the “To-Be” model, the analysts neglected some inessential levels of detail and focused themselves on the main structure of the design and production processes. At the end of the work, they released a new, complete IDEF0 model describing a future scenario, as described and designed by the technical management and including the designers and production engineers suggestions. The technical management released the “To-Be” model to the other Ferrari departments, in order to illustrate the advantages of the planned rationalization and reorganization processes and obtain the resources and the collaboration required to put it into practice.

 Conclusions

The paper describes a pilot project jointly promoted by Ferrari S.p.A. and the University of Parma, in order to experiment the use of process modeling and analysis methodologies in the Ferrari framework. The project involved designers, production engineers and managers from Ferrari and analysts from the University of Parma. The project, focused on the design and production processes of the fuel tank, required about six months and produced two distinct IDEF0 models. The first model, called “As-Is” model, describes the current (at the time of the analysis) design and production processes; the second model, called “Tobe” model, was derived from the first one, taking into accounts the reorganization and rationalization plans of the technical management. The project demonstrated the practicability and the significant advantages of process modeling and analysis on an industrial problem. The experience highlighted some non-obvious problems in the application of the IDEF0 methodology, due mainly to the knowledge elicitation. Finally, the project show that the modeling and analysis activities, if coordinated by external analysts, are able to provide immediate and medium term benefits: the “To-Be” model constitutes an effective base for the future process reorganizations and rationalizations initiatives.

Acknowledgements

We wish to tank the designers, the production engineers and the managers from Ferrari S.p.A. involved in this project, for their active collaboration and the support provided.

References

[1] J. M. Lamperini. Analisi e sperimentazione di metodologie per la modellazione di processo nell’ambito della progettazione di serbatoi carburante per autovetture sportive. Science Doctor Thesis, Università degli Studi di Parma, Parma Italy, 1996.

[2] U. Cugini, G. Bocchi, F. Folini and M. Galloni. Rapid prototyping and parametric CAD systems. In: Proceedings of 27ISATA International Symposium on Automotive Technology and Automation, Dedicated Conference on Rapid Prototyping for the Automotive Industries and Laser Applications for the Transportation Industries, (Aachen Germany, October 31–November 4 1994), pp. 51–57, ISATA, Automotive Automation Limited, Croydon UK, 1994.

[3] Integration definition for function modeling (IDEF0). Draft Federal Information Processing Standards Publication 183 (FIPSPUB 183). FIPS, USA, 1993.[4] François B. Vernadat. Enterprise Modeling and Integration: principles and applications. Chapman & Hall, London UK, 1996.