为在设计过程中有效减少轿车乘员约束系统的设计参数,并快速优化轿车乘员约束系统的安全保护性能,采用一阶线性响应表面模型和D-最优试验设计方法进行约束系统设计变量的筛选。对筛选后的变量应用Kriging代理模型和均匀试验设计方法进行设计参数的仿真优化,将优化结果与直接使用碰撞仿真软件MADYMO计算得到的结果进行对比.研究表明,一阶线性响应表面模型和D-最优试验设计方法能有效筛选出重要的并对约束系统性能具有决定作用的设计变量,而Kriging代理模型和均匀试验设计方法能在选定的设计变量的基础上,快速取得较好的优化设计结果.
In order to reduce the design parameters of vehicle occupant restraint system during the design process, and quickly optimize the system's performance, a new approach was developed for the optimization of vehicle occupant restraint system. At first, some design variables were chosen by using the linear response surface approximation and the optimal experimental design method; then Kriging surrogate model was constructed; and finally, the optimization design was conducted by using the even test method. By comparing the optimization result with the values of MADYMO simulation, it was found that the linear response surface approximation and the optimal experimental design method are effective in choosing important design parameters, and Kriging surrogate model with even test design method can give the expected optimization result quickly.