基于数值模拟的B柱加强板成形工艺优化

Numerical-simulation-based process optimization on B-pillar reinforcement panel forming

在对使用拼焊板拉深成形的汽车B柱加强板进行数值模拟的基础上,采用BP神经网络建立工艺参数与成形质量之间的非线性映射关系,通过多目标遗传算法NSGA-II获得最优成形工艺参数.研究结果表明:神经网络结合多目标遗传算法可以获得最优成形工艺参数,可较好地解决B柱加强板的成形问题.

Based on the numerical simulation on car B-pillar reinforcement panel forming from the tailored welded blanks,a nonlinear mapping relation between process parameters and forming qualities is first established via BP neural network. By applying the multi-objective genetic algorithm, i. e. NSOA-II, for optimal parameters, it is found from results that the integration of neural network and multi-objective genetic algorithm can obtain the optimal parameters to resolve the B-pillar reinforcement panel forming problems.