基于BP神经网络遗传算法的汽车塑件CAE优化分析

Optimization Analysis of Automobile Plastic Part CAE Based on BP Neural Network and Genetic Algorithm

以汽车内饰中立柱本体注射成型为例,基于Moldflow中CAE分析基础上,对塑件注塑所需的成型工艺参数进行了仿真,并分析了塑件翘曲成因,给出了翘曲改善优化目标。结合注塑工艺规律,借助于Tugachi正交试验法、BP神经网络遗传算法、Matlab数值分析对塑件注射成型工艺参数协同进行优化,并对优化结果进行了CAE比对验证。结果表明:神经网络预测推荐的工艺参数能有效将翘曲结果控制在质量误差范围内,提出的优化设计方法能有效降低模具试模成本,改善塑件成型质量。

As an example by studying on the injection molding of the automotive interior column body, based on analysis of Moldflow CAE,the molding parameters required for injection molding products were simulated, and the causes of product warpage deviation were analyzed, which gave warpage optimization goal. The plastic injection molding process parameters were optimized with the help of the Tugachi orthogonal experiment method, BP neural network genetic algorithm and Matlab numerical analysis, and the optimization results were verified on CAE. The results show that the neural network prediction of the recommended parameters can effectively control the warpage results in quality error range. The proposed optimization design method can effectively reduce the mold cost and improve the molding quality of plastic parts.