A quantitative concurrent engineering design method using virtual protyping-based global optimization and its application in transportation fuel cells.

摘要

Concurrent engineering and virtual prototyping are two emerging techniques that are bringing considerable economical benefits to the manufacturing industry. This work proposes the use of virtual prototyping to produce quantitative measures of product lifecycle performances to facilitate the implementation of concurrent engineering. A multiobjective, virtual prototyping-based global optimization problem is formulated to close the open loop of present virtual prototyping methods and to allow concurrent engineering design to be carried out systematically and automatically.; Virtual prototyping-based design optimization faces several technical challenges. First, virtual prototyping is usually computationally intensive; relations between design variables and product life-cycle performances are often implicit. Secondly, the optimization problem usually consists of multi-modal design (objective and constraint) functions. The complexity and multi-modal nature of the optimization problem preclude the direct use of conventional local and global optimization methods. In this work, a new and efficient search method for virtual prototyping-based global design optimization is introduced. The method, called Adaptive Response Surface Method (ARSM), carries out systematic “design experiments” through virtual prototyping to build second-order regression models to approximate the design functions. Through an iterative process, the regression models are improved and the global design optimum is obtained. The ARSM search scheme requires only a modest number of design function evaluations, making virtual prototyping-based global design optimization feasible.; The proposed quantitative concurrent design method is then applied to the components, stack and system design of a transportation fuel cell. The approach led to an optimized multi-functional component, a reduction of the system cost, and an improvement of the system performance. The approach can be applied to the concurrent design and design optimization of other complex mechanical components, assemblies and systems.