摘要
为了提高车门静态刚度,采用基于HDMR理论的多参数解耦优化方法与Morph技术结合,对一款处于概念设计阶段车门的主截面形状进行优化。这种建模方法的特点在于通过少量的样本点识别形状变量之间的耦合关系,进而将高维问题进行分解,在保证精度的前提下提高计算效率。采用Kriging插值方法对Cut-HDMR的各项进行构造,得到车门各刚度下的考察点位移和质量的近似模型,进而利用非支配排序多目标遗传算法(NSGA-Ⅱ)对近似模型进行优化,在车门静态刚度均达到设计要求的前提下合理进行减重,使垂直刚度和扭转刚度得到了很大的提高。
In order to improve the static stiffness of vehicle door, this paper used HDMR（high dimensional model representation） based multi-parameters decoupling theory combined with Morph technology to optimize main cross section shapes of a vehicle door in the conceptual design stage. It compared with other popular metamodel techniques, this method could identify the coupling relationships among shape variables by a few sample points, and then decomposed the high-dimensional problem into a series of low-dimensional sub-problems, so it could improve the efficiency remarkably without loss of accuracy. It adopted Kriging model to construct the terms of Cut-HDMR and then abtained the displacement approximate models of investigated points under five stiffness conditions. Based on the Kriging-HDMR models, then it used NSGA-Ⅱ（non-dominated sorting based genetic algorithm） to optimize the approximate model. The results show that it lightens the weight of the vehicle door and improves both the vertical stiffness and the torsional stiffness.
出处
《计算机应用研究》
CSCD
北大核心
2014年第1期157-161,共5页
Application Research of Computers
关键词
车门刚度
多参数解耦
多目标遗传算法
形状优化
vehicle door＇s stiffness multi-parameters decoupling multi-objective genetic algorithm shape optimization
作者简介
苏晅（1987-），男，硕士研究生，主要研究方向为汽车结构与CAE（suxuan259@126．com）；
王琥（1975-），男，副教授，硕导，主要研究方向为数值计算方法、工程优化问题．
