Sie sind auf Seite 1von 8

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

A conceptualisation of design context to explain design trade-offs in the automotive industry


Roxana Belecheanu 1, Johann c.k.h. Riedel1, Kulwant S Pawar 1
1

University of Nottingham, University Park, Nottingham, NG7 2RD, UK, {Roxana.Belecheanu,Johann.Riedel, Kulwant.Pawar}@nottingham.ac.uk

Abstract was undertaken. This paper presents a case study with a mass-market car manufacturer, aiming to improve the understanding of how design trade-offs are made in the context of a design organisation. The investigation covered designers, supervisors and managers from body, powertrain and doors engineering departments, in small and medium size car programs. Through triangulation of data sources and data collection methods, a large and detailed set of design trade-off examples was collected as part of the empirical evidence. The case study revealed how unformalised practices of decision making in trade-off situations override the guidelines for decision making of the company. Based on these findings, a framework for conceptualising the context of design decisions in trade-off situations was developed. A method for how to apply this framework to understanding the rationale of design trade-offs in different situations is also provided, along with an illustration of its use on two real life trade-off examples. Keywords Decision making, design trade-offs, design context, automotive industry.

1 Introduction
Design trade-offs are a useful and fruitful means to study decision making. A design trade-off is a non-trivial type of a decision situation, when the choice is not obvious because no solution can satisfy all the conflicting objectives. In these situations, an improvement in one performance attribute of the design can only be achieved by damaging another performance attribute. For example, a cost/performance trade-off in which the two conflicting design objectives are minimise cost and maximise performance, has several solutions: a 20% improvement in performance at the same cost, or a 20% reduction in cost at the same level of performance, or a 10% improvement in performance at 10% cost reduction. Therefore, trade-offs are potentially more challenging than simple decisions and can illustrate a decision making rationale which is not self-evident. In car design, making design trade-offs is particularly complex and challenging, due to the existence of a large number of interacting parameters and the fact that interactions between these parameters are often incompletely defined. In this case, customer satisfaction is delivered by a complex set of performance characteristics (e.g. weight, style, ergonomics, efficiency), under strict legislative requirements, safety and reliability concerns, while balancing subjective product attributes with objectively measured ones. Making a design trade-off is a very involved and iterative process and is not guaranteed to yield a desirable result (Tate and Nordlund, 1996). As many design problems are complex and as often there is no numerical analysis possible to support the decision making process, personal preferences of the design engineers and constraints derived from the product development process are the basis for the decisions they make (Ullman, 1997). Conflicts in preferences, resource requirements, etc. play an important role in the result of the design and analysing these conflicts can provide insight into how certain design decisions are made (Rajan, 1997).

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

2 Existing research on design trade-offs


Design trade-offs have been addressed mostly in prescriptive design research, in the context of design evaluation (Scott and Antonsson, 1996; Thurston and Nogal, 2001) and design optimisation (Sen and Yang, 1998). These are methods for decision support, originating in the decision theory. The focus here is on determining the optimal design solution, rather than analysing how decisions are made in order to reach an optimal design solution. These methods require assigning precise values to weights and preferences for the decision criteria, which is not always possible in the case of design problems of the complexity encountered in real life (e.g. car design). For this reason, they are thought to address design trade-offs only in a rationalistic way, which resemble a scientific ideal of replicability and not how engineering decisions are made in practice (Grudin, 1996). Therefore, they offer limited applicability to understanding how design trade-offs are made in practice. On the other hand descriptive design research, which investigates decision making in practice, pays insufficient attention to design trade-offs. Studies focusing on the decision maker (e.g. Lindemann, 1999; Hansen and Andreasen, 2000) seek to understand design as a cognitive and social activity; however the setting here is usually experimental and little research is done in the natural, real-life environment of a design organisation. Also, case studies on capturing design decisions rationale in context are limited in number and do not address the particular case of design trade-offs. Moreover, while Design Rationale researchers (e.g. Grudin, 1996; Sharrock and Anderson, 1996) acknowledge that tools for capturing design rationale must have a mechanism for showing the impact of design context issues on design decisions, they are not able to show how such a tool would be applicable to d ifferent project contexts. The reason is that a conceptualisation of the context of the design project does not exist. Similarly existing research on design in context (Hales, 1993; Maffin, 1996; Birmingham et al., 1997) fails to provide a coherent picture of design context. Firstly, there is no coherent understanding about what design context means. Secondly, its impact on design decisions is addressed only at a general, broad level, without showing the influences or constraints imposed on design decisions in trade-off situations. In summary, the review of literature identified two main knowledge gaps: 1. The need for an overarching framework for conceptualising the design context, as a linking mechanism between the context factors which impact on design trade-offs. 2. The need for a systematic way to approach and understand the design context and its impact on design trade-offs, due to the complexity of the design context

3 Research approach
An in-depth case study was carried out with a volume car manufacturer (na med here AutoVM1). The company was chosen to represent a typical car manufacturer, i.e. a mass-market car maker. It was thus aimed to target typical design decisions taken in the development of mass-market car models and to exclude exceptional decisions that can occur in specialist designs like luxury or sport cars. The investigation involved semi-structured interviews, qualitative and quantitative questionnaires, as well as non-participant observation. The respondents were designers, supervisors and managers from small and medium size car programs of family saloon type. They were selected from body, powertrain and doors. The data collection process lasted for one year, in which a relationships with key contacts in the company was initially established, then two series of semi-structured interviews were carried out, and then a qualitative questionnaire was
1

AutoVM (Automotive Volume Manufacturer) is a fictitious name, used in this research for confidentiality reasons.

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

used to collect trade-off examples in a systematic way and with a high level of detail. The questionnaire was first piloted and filled in through interviews and then it was posted. By using multiple sources of evidence and a triangulation of data collection methods, reliability of data was ensured. A large set of detailed examples of design trade-offs was collected and provided evidence for the impact of several design context factors on design trade-offs. The study exploratory, since a pre-defined agenda did not exist at the outset. This allowed the researcher to identify gradually the important variables of the design context of AutoVM, in a manner in which data collection and interpretation were intertwined and iterative.

4 Case study findings


The analysis of the empirical data showed the decision making practices of designers, supervisors and design managers in AutoVM differ from the prescribed form of decision making (e.g. guidelines and methods for making trade-offs) offered by the company. Thus, the following decision making patterns were found: The relative importance of decision criteria in trade-off situations: 1. changes over time, within the period of a program, 2. varies with the position of the decision maker within the organisational hierarchy 3. varies depending on the product part where the decision is made. Time changes the importance that people give to decision criteria: While reducing variable cost and improving performance are equally important in the early program stages, performance and quality are prioritised in the late program stages, at the expense of product cost and development cost. Development cost suffers towards the end of the program mainly because meeting production deadlines is crucial: the company must start selling cars, because revenues from selling can counteract the increase in development cost. Exceptionally, saving product cost in the late stages can be a reason for not meeting the deadlines, only if the savings are significant enough to account for the loss of market revenues (e.g. millions of dollars). It was thus found that cost is one of the targets with most variable importance throughout the program.

1.

2. A design trade-off is solved differently by different decision makers at different hierarchical levels: design engineers tend to prioritise functionality and performance, design supervisors tend to treat equally performance, cost and time, design managers tend to prioritise cost, time, customer and competitiveness.

3. Thirdly, it was also observed that the product component or product system where the trade-off is made affects the trade-off. For example, weight is often more important than performance in trade-offs on car body, while weight is usually perceived as less important than performance in trade-offs on doors. These findings reflect implicit, unformalised practices of decision making which override company guidelines and which are simple principles to enable the researcher and practitioner to approach a particular design trade-off in a particular design context. It is therefore necessary that, in order to correctly capture the decision making rationale, a conceptualisation framework for design trade-offs should accounts for: 1. when the trade-off is made 2. 3. who makes the trade-off the product part(s) where the trade -off is made.

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

A typical example to illustrate these patterns is the weight-performance-cost trade-off (Figure 4.1):
Hierarchical position

Design manager Design supervisor Design engineer

Managers tend to prioritise cost Supervisors prioritise almost equally cost, performance and weight Engineers tend to prioritise performance Program Stage

Early stages

Late stages

Weight loses against performance Weight is more important than for doors, due to impact on fuel economy Weight is more important than for powertrain, due to its impact on car weight Product System Body

Reducing weight Performance is critical, while improving performance while weight and cost and cost are sought lose importance Doors Powertrain

Figure 4.1: The weight-performance-cost trade-off positioned in the design context space

The variability of design targets relative importance with time is as follows:


Importance of target performance

cost weight Program Stage (time)

Figure 4.2 Symbolic representation of the variance of importance of targets with time

in the early stages, there is time to research means of lowering the weight at the same time as lowering the cost and improving the performance; in the late stages, performance must be delivered at all costs, hence it becomes more important than cost and weight.

otherwise, the tendency is to allow a slight increase in weight, if performance is improved significantly. Conflicts between the relative importance of the traded-off variables given at different hierarchical levels make decisions go through several iterations or be reversed. The decision making process taking into account these iterations was mapped on three hierarchical levels (Figure 4.3):

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

Design Group A Are internal targets met? Targets of own group (internal) and of other groups (external) Assign values to design variables, simulate and test iterations Decide what variables to change in order to meet targets Engineer level No Are external targets met? No Yes Are cost, time, customer affected? Design Group B

Horizontal (cross-functional)

Yes

OK!

Cross-functional trade-off: Negotiate with Design Group B or decide changes of own designs/targets

No

Supervisor level Yes

Cross-functional trade-off: Decide which targets to meet or ask for more iterations Manager level Vertical (hierarchical)

Figure 4.3: Negotiation process for design trade-offs

5 A conceptualisation framework of design context


As the design literature review had identified a need for a design context conceptualisation, these three dimensions were seen as a useful and meaningful start to form this framework of the design context in AutoVM. A conceptualisation framework of design context was thus constructed, having at the centre a 3-dimensional space Figure 5.1): 1. Program stage (the time), 2. 3. Hierarchical position of the designer in the organisation, Product system/ component.
Subjectivity

Engineering complexity

Hierarchical position of decision maker

?
Time pressure Program Stage

Product System

Carry over Customer impact

Data availability

Figure 5.1: A conceptualisation of the design context

Any design trade-off situation can be understood by mapping it firstly on this 3-dimensional space. The three dimensions must then be supplemented by a representation and analysis of other contextual factors which explain, modify or reinforce the initial mapping (explanatory variables). Some of the following variables investigated within this case study and their impact on design trade-offs are outlined in Table 5.1 below:

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

Context variable Product complexity

Impact on a design trade -off in AutoVM

1. The complexity of a product in terms of the number of functionally interconnected parts determines the number and difficulty of trade-offs inherent in design. 2. Component standardisation makes cost/performance trade-offs easier, by reducing the cost. 3. The number of components is a concern for AutoVMs designers due to its impact on car weight. 4. Sometimes, the investment needed or the performance risks associated with reducing product complexity can outweigh the cost and time benefits initially sought. Reuse of 1. Functional constraints imposed by reused parts make achieving new design targets components more difficult. (Carry 2. The number of conflicts between old and new components and hence the number of over) cross-functional trade-offs are likely to increase. 3. In AutoVM, the engineering difficulties and the added development cost associated with the use of carry over parts are less significant than the product cost and the time saved. Subjectivity of qualitative product targets Time pressure and data availability 1. Subjectivity interferes in determinin g the relative importance of different product attributes 2. Qualitative product attributes which are not objectively quantified can lead to misinterpretation and/or errors in design trade-offs. 3. The relative importance of design attributes can be underestimated or overestimated 1. Design trade-offs made under time pressure are likely to be made with incomplete or inaccurate data. 2. As result of time pressure or insufficient data, design trade-offs are made under risk (the risk of higher manufacturing cost, manufacturing problems, quality problems, etc.). Table 5.1: Some design context variables with impact on design trade-offs

The design context of each trade-off is different and depends on the situation. Therefore, to produce a description of the context of a trade-off, a pick and mix approach must be used: only the mix of the relevant variables must be investigated, to see their impact on the decision.

6 Example
The following example is a trade-off between NVH (Noise, Vibration, Harshness), durability and time. The situation was generated by NVH problems at the right engine mount, problems which are visible to the customer and which also affect the durability of the part. Designers were confronted with the choice between: 1. To fix the NVH problems completely, which would have meant a risk to meeting the design freeze deadline (here called Engineering Sign-Off, or ESO) in terms of testing the durability. This would make the customer happy, but it would endanger safety. To fix the NVH problem partially, which would give time to test durability completely, thus meeting safety requirements. Not to fix the NVH problem - with no risk to durability, but with negative impact on customer.

2. 3.

Because of the need to solve a customer related problem in short time, a compromise had to be reached. The final decision was to implement only some of the design changes and solve the problem partially. The changes which minimised the risk of negative impact on the customer were implemented, while a secondary series of design changes was left for after the ESO. Durability testing was carried out only to allow for a minor risk to overrun the deadline.

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

In order to understand the rationale of this deci ion, the steps below must be followed: s 1. Firstly, the NVH-durability-time trade-off situation must be mapped on the 3dimensional space, to understand how the three dimensions impact upon that trade-off (Figure 6.1). This generates the following assumptions in terms of the importance of the 3 design targets: Program Stage: Confirmation Prototype (i.e. a late program stage) The trade-off situation occurs 10 weeks before ESO (i.e. at a late program stage). Production tooling has started and no design changes are allowed after ESO. Meeting ESO with all the targets tested and all design functional problems solved is crucial. Hence, customer, performance and meeting the deadline are more important than cost. Hierarchical Position of Decision Maker: Design Manager The decision was taken by the NVH Manager and Vehicle Engineering Manager. Cost, time and customer are expected to be the main concerns of these decision makers.
Hierarchical position of decision maker

Design manager level Trade-off (NVH, durability)

10 weeks before ESO

Engineering Sign-Off (ESO)

Program Stage (time)

Powertrain system

Confirmation Prototype

Pilot Production

Product System

Figure 6.1: Mapping the NVH/durability trade-off in the 3-dimensional space

Product System: the Powertrain System

The right engine mount is part of the powertrain system - which is highly complex, with highly specialised engineering and with a strong interaction with the suppliers. It is expected that cross-functional negotiation is needed due to interactions between engine mount and body, hence several decision iterations are likely. Furthermore, the product attributes involved in the trade-off determine the following patterns in decision making: Cost (variable cost and tooling cost) increases due to design changes, but the general trend on cost is to lower its importance at this stage. Weight increases due to adding material for fixing NVH problems. However, the general trend on weight at this stage is to allow higher weight if performance gains. 2. Secondly, other explanatory variables from the design context applicable to this situation must be identified and investigated. In this case, these were: Customer visibility:

NVH is noticed by the customer and is of average importance to the customer. Durability is a measure of safety, legally crucial, but less noticeable by the customer.

th The Proceedings of the 9 International Conference of Concurrent Enterprising, Espoo, Finland, 16-18 June 2003

Engineering complexity: Design changes needed to solve NVH affect stiffness, hence the decision must consider the impact on durability, and determining this impact takes time. Hence the decision needs to be made through several iterations (distributed in time) and not completed at one moment. Data availability: Durability test data was missing for the decision and would have only been available after ESO. Hence, making a decision required taking risks.

Time to implement the change: Not enough time to solve all NVH issues before ESO, not enough time to test the durability impact before ESO. The trade-off was solved through a compromise, by taking a risk in terms of durability, when substituting missing data with engineering judgement. The decision maker took this risk, believing that there would be opportunities to fix potential problems even after the deadline.

7 Conclusions
This paper presented a conceptualisation framework for design context to explain decision making in trade-off situations. The framework shows how to systematically investigate design context, thus helping designers to analyse the possible constraints imposed by design context on design trade-offs, depending on when, where and by whom the trade-off is made. This framework is useful to designers because it brings simplicity and clarity to the complex landscape of the design context. It is also practically important because it makes explicit a phenomenon (i.e. the variance in the importance of decision criteria) of which decision makers might not be aware or do not reflect upon, when making trade-offs.
References Birmingham, R., Cleland, G., Driver, R. and Maffin, D. (1997). Understanding engineering design: context, theory and practice, First Edition, Prentice Hall, London. Grudin, J. (1996). Evaluating Opportunities for Design Capture, In: Design Rationale: Concepts, Techniques and Use, Moran, T.P. and Caroll, J.P. (Eds.), Lawrence Erlbaum Associates, Hillsdale, New Jersey. Hales, C. (1993). Managing engineering design, Longman Scientific & Technical, Harlow, England. Hansen, C.T. and Andreasen, M.M. (2000). Basic thinking patterns of decision-making in engineering design. In: Proceedings of the First International Workshop on Multi-Criteria Evaluation (MCE), Neukirchen, 1-8. Lindemann, U. (1999). A model of design processes of individual designers, In: Proceedings of the 12th International Conference on Engineering Design, Munich, August 24-26. Maffin, D. (1998). Engineering Design Models: Context, Theory and Practice, Journal of Engineering Design, 9(4), 315-348. Rajan, V.N. (1997). Decision-Based Design. Position Paper, http://dbd.eng.buffalo.edu/papers/rajan.html, (accessed January 2003). Scott, M.J. and Antonsson, E.K. (1996). Formalisms for negotiation in engineering design, Proceedings of the 1996 ASME Design engineering Technical Conference and Computers in Engineering Conference (96DETC/DTM-1525), August 18-22, Irvine, California. Sen, P. and Yang, J.-B. (1998). Multiple criteria decision support in engineering design, Springer Verlag. Sharrock, W. and Anderson, R. (1996). Organisational Innovation for Design Capture, In: Design Rationale: Concepts, Techniques and Use, Moran, T.P. and Caroll, J.P. (Eds.), Lawrence Erlbaum Associates, Hillsdale, New Jersey. Tate, D. and Nordlund, M. (1996). A design process roadmap as a general tool for structuring and supporting design activities, In: Proceedings of the Second World Conference on Integrated Design and Process Technology (IDPT), 3, Society for Design and Process Science, Austin, Texas, December 1-4, 97-104. Thurston, D.L. and Nogal, A. (2001) Meta-level strategies for reformulation of evaluation function during iterative design, Journal of Engineering Design, 12(2), 93-115. Ullman, D.G. (1997). Decision Based Design. Position paper, The ASME Design Technical Conference (DTC), Sacramento, California, September, http://dbd.eng.buffalo.edu/papers/ullman.html (accessed January 2003).

Das könnte Ihnen auch gefallen