基于蝙蝠算法优化BP-GA模型的汽车前保险杠注塑成型优化

Optimization of injection molding for front bumper of an automobile based on BP-GA model with bat optimization

为降低某车型前保险杠注塑成型中产生的翘曲变形,基于数值模拟结果,将神经网络-遗传算法寻优模型与蝙蝠算法结合,确定了BP神经网络的初始权值和阈值。将BA-BP模型代入遗传算法中求解最佳工艺参数。由极差分析可知,影响翘曲变形最显著的因素为保压时间和模具温度。基于极差分析设计补充实验,训练BA-BP神经网络并作为遗传算法的适应度值进行迭代寻优。结果表明:BA-BP模型的相关系数R2可达0.99以上,平均绝对误差为1.05%,能较精准地预测翘曲量,为遗传算法提供可靠的适应度值。优化后的工艺参数为:模具温度50℃,熔体温度280℃,注射时间5 s,保压压力60 MPa,保压时间60 s,冷却时间30 s,翘曲量降低至6.8 mm,与原工艺相比降低了60.92%,满足生产要求。

To reduce the warpage caused by the injection molding of the front bumper of a certain car model,based on the numerical simulation results,the neural network-genetic algorithm optimization model was combined with the bat algorithm to determine the initial weights and thresholds of BP neural network.The BA-BP model was substituted into the genetic algorithm to solve the optimal process parameters.It can be seen from the range analysis that the most significant factors affecting the warpage are the holding time and the mold temperature.Supplementary experiments were designed based on range analysis,BA-BP neural network was trained and used as the fitness value of genetic algorithm for iterative optimization.The results show that the correlation coefficient R2 of BA-BP model can reach more than 0.99,and the mean absolute error is 1.05%,which can accurately predict the warpage amount and provide a reliable fitness value for the genetic algorithm.The optimized process parameters are mold temperature of 50℃,melt temperature of 280℃,injection time of 5 s,holding pressure of 60 MPa,holding time of 60 s and cooling time of 30 s.The warpage amount is reduced to 6.8 mm,which is reduced by 60.92%compared with that of the original process,and meets the requirement of production.