关键词: 微型泵, 有限元, 粗糙集, 遗传算法, 回归分析

Abstract: To deal with imprecision in single parameter optimized design for the micro-pumps, a multidisciplinary approach integrated with Finite Element Method (FEM), Rough Set Theory (RST), Multi Regression Analysis (MRA) and Genetic Algorithm (GA) was proposed. With this approach, the FEM was used to analyze the dynamic properties of the virtual product applications with coupled-energy-domain and information data could be obtained; The RST and the GA were used to mine the data from achieved information, and key attributes affecting the properties of the micro-pumps could be defined. Based on achieved data and key attributes, the general optimized design model was constructed by the MRA method. The prototype software of optimized design synthesis of the micro-pumps was developed based on these approaches. And the optimized design synthesis of a thermal actuated micro-pump was conducted by making use of the prototype software. Efficiency of the thermal actuated micro-pump has been significantly improved. The research results indicated that the method and the prototype software were feasible and practical and could be applied to the optimized design synthesis of other micro-components.

Key words: micro-pump, finite element method, rough set theory, genetic algorithm, regression analysis