【摘要】 为了对轿车车门密封条结构参数进行优化设计,采用遗传算法和神经网络相结合的策略,首先利用神经网络建立密封条结构设计参数与压缩负荷、应力等的非线性全局映射关系,获得求解结构优化问题所需的目标函数,然后用遗传算法进行优胜劣汰的寻优搜索运算,求出最优解。优化结果表明,椭圆形结构在壁厚为1.5mm、高度为20mm时,压缩负荷和应力能达到目标函数要求。压缩负荷和应力的优化结果与理论计算值的误差分别为7.4%、9%,因此,利用神经网络和遗传算法进行结构参数优化的方法是可行的。更多还原

【Abstract】 Because of the sectional shape is irregular and the wall thickness is uneven,a new method was used to optimize structural parameters of the seal,which adopted the strategy of combining Genetic(Algorithm)(GA) and Neural Network(NN).Using learning algorithm of NN,nonlinear global mapping function can be found between structural parameters and mechanical properties,such as compression load and stress.Objective function of GA can be obtained synchronously.Optimum solution can be solved using search operation of GA.The results show that the values of the compression load and stress can be satisfied to the objective function of GA when the shape of seal is ellipse and the wall thickness is 1.5mm,the height is 20mm.The errors between optimization values and theoretic results of the compression load and stress are 7.4% and 9% respectively.This method is effective to optimize structural parameters.更多还原