Kriging模型与优化算法在汽车乘员约束系统仿真优化中的应用研究

Occupant Restraint System Simulation Design Optimization Based on Kriging Model and Optimization Method

汽车乘员约束系统的优化涉及到安全带、安全气囊等部件的众多参数.首先采用一阶线性响应表面模型、正交实验设计方法进行参数筛选.然后构造Kriging代理模型来代替计算机仿真,并结合两种不同的优化算法来进行乘员约束系统参数的优化设计.结果表明:该方法对汽车乘员约束系统参数优化具有明显效果,可行方向法在该优化中优于多目标遗传算法.

Vehicle occupant restraint system optimization involves many parameters of seat belts, airbags and other parts. A new optimization method was proposed, in which linear response surface approximation and orthogonal experimental design methods were first used to choose the parameters, then the Kriging surrogate model was constructed to substitute computer simulation, and two optimization methods were combined to optimize occupant restraint system parameters. The result has shown that this method has obvious effect on occupant restraint system parameters optimization, and CONMIN is better than NSGA-II.