汽车B柱加强板热冲压工艺的遗传算法多目标优化

Multi-objective optimization on hot stamping process for vehicle B-pillar reinforced plate based on genetic algorithm

为了提高汽车B柱加强板热冲压件的厚度均匀性,提出了非支配排序遗传算法的工艺参数多目标优化方法。分析了B柱加强板的热冲压过程,选择板料加热温度、冲压速度、摩擦系数和模具间隙作为优化参数。以减小最大减薄率和最大增厚率作为优化目标,建立多目标优化模型。使用最优拉丁超立方抽样法在优化空间内设计实验点,并依据Autoform有限元软件得到实验值。基于BP神经网络拟合工艺参数与质量参数之间的关系,并依据预测均方根误差验证了拟合的精确性。使用非支配排序遗传算法搜索多个优化目标的Pareto前沿解。经生产验证,与厂家工艺相比,优化后的热冲压件最大减薄率减小了22.25%,最大增厚率减小了7.34%,说明优化后冲压件的质量得到了提高;且优化后最大减薄率和最大增厚率的标准差均减小,说明优化工艺的生产稳定性更好。

In order to improve the thickness uniformity of hot stamping part for B-pillar reinforced plate,a multi-objective optimization method of process parameters based on the non-dominated sorting genetic algorithm was proposed,and the hot stamping process of B-pillar reinforced plate was analyzed.Then,taking plate heating temperature,stamping speed,friction factor and die clearance as optimization parameters and the reduction of the maximum thinning rate and the maximum thickening rate as the optimization goals,the multi-objective optimization model was built.In the optimized space,the experiment points were designed by the optimal Latin hypercube sampling method,and the experimental values were obtained by finite element software Autoform.Furthermore,the relationship between process parameters and quality parameters was fitted by BP neutral network,the fitting accuracy was verified based on the predicted root mean square error,and Pareto front solutions of multiple optimization objectives were searched by the non-dominated sorting genetic algorithm.After production verification,compared with the thoriginal process used by manufacturer,the maximum thinning rate and the maximum thickening rate of the optimized hot stamping part are decreased by 22.25%and 7.34%respectively,indicating that the quality of stamping part is improved after optimization,and after optimization,the standard deviations of the maximum thinning rate and the maximum thickening rate are both reduced,indicating that the production stability of the optimized process is better.