Go to the table of contents.

Previous ...



6 DISCUSSION


6.1 Initial research motivations

The motivation for this research has its roots in a lack of a systematic development approach that could support the whole product development process from early phases until completion as well as supporting different engineering disciplines – especially mechanical engineering, electrical engineering, and software engineering. Complementing this problem area were weaknesses identified in the area of product descriptions. The state of affairs in the product description systems based on a parts paradigm when the research work was initiated was their lack of traceability of the design sources responsible for the product variety represented. Lack of traceability in this context is the fact that a part was manually assigned a usage statement on the product level. There was little or no support for connecting the specification and design of the part to the assigned usage – except by asking the engineer responsible and the person in charge of product variant definition. The parts-based paradigm furthermore implied that early phases of product development could either not be supported or had to be tweaked into a behavior where the parts-based paradigm could be artificially forced upon the development process. An additional issue, that has not been in focus in this research, is the labor resources needed to keep these types of product descriptions up to date in a business operation where changes are frequent and many.

The guidance and control of product development provided another set of motivations. A project-based approach supported by a stage-gate process model was used to drive and control development programs. In the automotive industry the main thrust of development is refinement design rather than new product development. This implies a high degree of repetition in the work tasks to be performed during development. Conceptually, this relates more to a manufacturing process control paradigm than to a project-based paradigm. However, given that the analogy is not straightforward, there seem to be potential benefits of adopting a manufacturing process control approach for product development as well. A flow based control paradigm for product development of this kind seemed to be promising and worthwhile to investigate further. Although a stage-gate process model can be a step toward a production process oriented approach to product development, the combination with a project oriented control paradigm still includes potential problem areas. In a project oriented paradigm there is an almost inevitable set-up scenario associated with the start of a project. Potentially, these set-up efforts can be avoided in a more production process oriented approach. In addition, there is an inherent tendency in a project management team to take decisions that are more optimized towards the project objectives with the risk of being sub-optimal for the larger range of products that will be based on the same product platform. In a more production process oriented approach many of these decisions are more or less built into the respective operations that constitute the process.

The research work presented in this thesis is grounded in the organizational setting of an automotive company. The company has been subjected to substantial changes in its environment as well as in its internal affairs before and during the time this research was conducted. The work presented spans over a time when the company has been transformed from one of the smallest independent and globally acting vehicle developer and manufacturer to the current state of affairs, where the company has been fully integrated into the functional organizational structure of the largest automotive company in the world.


6.2 The conception of the configurable component concept

The early thoughts about the weaknesses in the traditional approach to the core product description system came several years before this research work was formally initiated. In a team effort, these weaknesses were analyzed and approaches to deal with them discussed. The initial approach was referred to as "smart parts". Gradually, this concept was refined into the configurable component concept described in section 5.4. The concept was furthermore implemented in the commercial PDM system iMAN in order to provide support for an ongoing vehicle development program (see section 5.6.1). It was realized that the contemporary available commercial PDM systems would not be able to implement all aspects of the conceived concept. However, a commercial PDM system (in this case iMAN from Unigraphics) provided a reasonable starting point and capability to be operationally used. With the operational experiences of using the configurable component concept came the understanding of the necessity to utilize a systematic approach to the identification and definition of appropriate components. One effect of this experience and understanding was the initiation of the research work reported in this thesis.


6.3 Experiences from the implementation

The configurable component concept described above was implemented as a new core product description system to support a full-scale vehicle development program and its subsequent role in releasing information for manufacturing operations. Paper C is primarily focused on describing this implementation and some of the experiences and conclusions drawn. The implementation of this new core product description system was discontinued after a couple of years in operational use owing to a requirement to utilize globally common systems within the organization. Another change of product description system was therefore conducted. This switch was a switch back to a third-generation system (see part c in Figure 55), but a system slightly different from the previously abandoned system. An interesting observation here is that this has actually meant that the organization has conducted two successive full-blown changes of the core product description system within approximately a five year period.

A benefit of the configurable component (CC) concept when implemented as a replacement of a core product description system was the ability to allow definition of components similar to the parts used in the replaced system. Even though the intention of the CC concept is to adopt a systems oriented approach and to utilize a systematic approach (like the presented function-means methodology) in the design and definition task, there is nothing to prevent a user from utilizing a physical parts paradigm in order to remain close to the previously available product description approach. This capability substantially reduces the changeover efforts and risks and allows for the organization to gradually make a transition from a physical parts paradigm toward a more systems oriented approach that more fully utilizes the modeling capabilities available in the CC concept.


6.4 Expected effects on product development

Modularization is the conscious goal driven decomposition and grouping of design solutions in order to provide building blocks (modules) suitable for selection into several products (product configurations, product variants, derivative products). Note that these building blocks can be created on various levels of abstraction. The most concrete level of abstraction is when the building blocks are composed of a set of physical parts. A more abstract level of building blocks would be functional building blocks that can be combined into higher-level functions by selection of a configuration of the functional building blocks defined. In between, there is a potential to create building blocks based on parameterized design solutions that are a couple of steps further into the development process than the functional building blocks, but also some steps away from making the final decisions on the physical realization of the design solutions. This mid-abstraction level is the target of the concept of configurable components.

Modularization is regarded to be a key strategy in order to facilitate a high level of re-use without compromising too much on providing distinctive design solutions and products that can be tailored to meet individual customer needs. Creating product variants through a selection of pre-defined physical components has been used extensively for a long period of time. The purpose of introducing configurable components as the main carrier of design definitions and descriptions in the core product description system of a company is to move this capability of creating product variants earlier into the development process and to a higher level of abstraction. By moving the capabilities earlier and to a higher level of abstraction, more parameters are left open that can be tailored to meet customer needs for individualized product variants, but also to respond more effectively and efficiently to requests for development of new product derivatives based on an existing design solution platform.

Ullman (1997) has presented a well-known illustration of the design paradox (see Figure 77). As an adaptation based on the research presented here additional curves have been added. Overlaid in this Figure 77 using dashed lines is an illustration of the expected effects on utilization of the proposed concept of configurable components. When a new design problem is begun, very little is known about the solution, especially if the problem is a new one for the designer. As work on the project progresses, the designer’s knowledge about the technologies involved and the alternative solutions increases. Throughout the solution process, knowledge about the problem and its potential solution is gained, and conversely, design freedom is lost. This traditional wisdom about product development is illustrated in Figure 77 by the solid lines. The dashed lines in Figure 77 intend to illustrate the expected effects of successful application of modularization, parameterization, and re-use of design solutions.

Figure 77: The design process paradox (adapted from Ullman, 1997).

The re-use of design solutions leads to a starting point with less than 100 % design freedom, but also with more pre-knowledge about the design problem and the design solutions. The modularization and parameterization of the design solutions lead to a situation where progress in the product development can be made with less reduction in design freedom. Please note that the time scale is graded in percent from the initiation of a development task to its completion (probably defined by the time the first products are shipped to customers). If the time had been presented in absolute measures the successful application of the proposed concepts would be expected to have the potential to drastically improve responsiveness and shorten the time to providing new derivative products to the market.



Continue ...

Go to table of contents.