IMPACT LABORATORY IMPROVING MANUFACTURING PRODUCTIVITY WITH ADVANCED COLLABORATION TECHNOLOGY A Backward Mapping Approach for Complex System Design

Parametric design for complex engineering systems at the early stage involves identifying feasible candidates for detailed stages that satisfy multiple performance requirements at different levels. Conventional methods, using system analysis through “trial-and-error” iterations to pursue optimal solutions, are not robust and inflexible for further design and performance changes. In this paper, several methods are studied and a Backward Mapping approach is presented to deal with this problem. The Backward Mapping Methodology can help design engineers quickly identify robust design alternatives through set-to-set mappings from performance space to design space in complex system designs. In this approach, comprehensible and efficient surrogate models are used to increase design knowledge and expedite the design process. The use of design regions with desirable levels of confidence not only enhances the robustness of design under a high degree of uncertainty but also offers more alternatives for better decision making at the early and later design stage. The main framework of the approach is demonstrated by using an example of vehicle system design at the early stage. The preliminary results show the effectiveness of this approach and its potential applications in different engineering design scenarios.