A strategy for resolving evolutionary performance coupling at the early stages of complex engineering design

Abstract

In the performance-driven complex engineering design process, at an early design stage, some required design parameters or equations cannot be determined precisely, which can have significant impact on the chain of design decisions and have to be validated with the outcome of design results from the later stages requiring redesign effort to resolve inconsistencies. So it is imminent to evaluate the extent of, and systematically manage, the couplings of design models and develop an appropriate resolving strategy in order to obtain more accurate design results in the shortest time period. This paper presents the research work carried out via the analyses of performance evolution and potential performance coupling at the early stage of complex engineering design. Four strategies based on performance model transformation for resolving evolutionary performance coupling are studied (i.e. decoupling, coupling, first-decoupling-then-coupling and first-coupling-then-decoupling). A selection method for resolving performance coupling based on the synthetic analysis of sensitivity of uncertainty propagation, solvability of coupled models, coupling strength and performance interface and availability of design information is proposed. To demonstrate the related concepts and method, the solving process of a complex design problem related to a suspension system design for tracked vehicle is given.

Keywords:

• descriptive models of the design process
• evolutionary design
• process modelling

Acknowledgement

This research work is funded by the National Science Foundation of China, Grant number 50705083. The authors are very grateful to the anonymous reviewers for their insightful and valuable comments on this paper.

Notes

To preserve confidentiality of the product, detailed specification variable values have been omitted.