基于BP神经网络的连接器注塑工艺参数多目标优化

Multi-objective Optimization of Connector Injection Molding Process Parameters Based on BP Neural Network

为提高连接器注塑工件质量,获得最优的连接器注塑工艺参数,设计一种基于BP神经网络的连接器注塑工艺参数多目标优化方法。利用方差分析法获得对试验结果有显著影响的参数,进而得到更全面的信息。建立相应工艺参数优化模型,并添加多层级的结构;建立BP神经网络集预测模型,映射工艺参数与质量指标的非线性关系,利用预补偿法最终实现工艺参数优化及误差补偿。在相同的测试环境之下,对比于传统优化补偿测试组,新型的优化补偿组所得出的翘曲平均值较低。测试结果证明:新型优化补偿组处理效果更佳,具有一定的应用价值。

In order to raise the quality of connector injection molding workpieces for the optimal injection molding process parameters,a multi-objective optimization method of injection molding process parameters based on BP neural network is designed.Variance analysis is conducted to abtain the parameters which significantly affects experimental results for more comprehensive information.The corresponding process parameter optimization model is built,with a multi-level structure being added.Prediction model of BP neural network set is established,the nonlinear relationship between process parameters and quality index is mapped and pre-compensation is used to achieve process parameter optimization and error compensation.Under the same test condition,compared with the traditional optimized compensation test group,the improved one has a lower average warpage with better processing effect and certain application significance.