响应面和多目标遗传算法结合的副车架优化

Optimization for Sub-Frame Combining Response Surface with Muti-Objective Genetic Algorithm

为实现副车架设计过程中,质量和第一阶模态频率同时达到最优,在模态分析和三种工况副车架强度分析的基础上,首先应用Hyperworks进行了副车架参数化,建立了11个厚度尺寸变量。然后应用试验设计分析方法对尺寸变量进行筛选,去掉了3个对质量、最大应力和第一阶模态频率影响都不显著的因子,将基于移动最小二乘法构建响应面近似模型引入到副车架优化设计的复杂系统中。最后,基于副车架近似模型利用多目标遗传算法进行多目标优化,获得了副车架质量和第一阶模态频率的Pareto最优解。研究结果表明:通过获得的Pareto最优解的边界,可以指导副车架优化设计,将大幅缩减产品开发周期、降低产品开发成本。

In order to achieve the optimal mass and the first modal frequency in the process of the sub-frame’s design,it based on the modal and the strength analysis of three working conditions.Firstly,the sub-frame is parameterized by Hyperworks and 11 thickness dimension variables are established.Secondly,design of experiments(DOE)is explored to screen the size variables of sub-frame,and then the 3 size variables which are not significant factors for mass,maximum stress and the first-order frequency of sub-frame are removed,a response surface approximate model based on moving least squares method(MLSM)is introduced to the complex systems of sub-frame optimization design.Finally,multi-objective optimization is computed by multi-objective genetic algorithm(MOGA),using a approximate sub-frame model,to obtain the Pareto optimal solution of the mass and the sub-frame’s first modal.The results demonstrate that pareto optimal front has potential for guiding the sub-frame optimization design,reduction of the product development cycle and the decrease of product development’s cost significantly.