基于BP神经网络和MIV算法的注塑件工艺参数优化研究  

Research on Optimization of Injection Molding Process Parameters Based on BP Neural Network and MIV Algorithm

作　　者：张鲁滨 黄海松[1] 姚立国[1] Zhang Lubin;Huang Haisong;Yao Liguo(Key Laboratory of Advanced Manufacturing Technology,Ministry of Education,Guizhou University,Guiyang 550025,China)

机构地区：[1]贵州大学现代制造技术教育部重点实验室

出　　处：《塑料科技》2018年第12期94-99,共6页Plastics Science and Technology

基　　金：贵州省科技重大专项计划(黔科合重大专项[2017]3004号);贵州工业攻关重点项目(黔科合GZ字[2015]3009);贵州工业攻关重点项目(黔科合GZ字[2015]3034);贵州省教育厅项目(黔教合协同创新字[2015]02)

摘　　要：针对注塑件翘曲变形问题,以某塑料叶轮为研究对象,首先设计了正交试验对叶轮进行翘曲分析,通过正交试验获得的相关数据,建立了基于BP神经网络的注塑件翘曲量预测模型。在预测模型的基础上,通过采用平均影响值(MIV)算法对模型的输入参数进行筛选后,再进行仿真模拟。结果表明:经MIV算法优化后的塑件翘曲量预测模型具有较高的预测精度,模型预测的相对误差由原来的13%减小到7%,对实际注塑加工生产具有重要意义。Aiming at the problem of warpage deformation of injection molded parts, taking a certain plastic impeller as the research object, the orthogonal experiment was designed to analyze the warpage of the impeller. Based on the relevant data obtained from the orthogonal experiment,the prediction model of warpage of injection molded parts was established based on BP neural network. On the basis of the prediction model, the input parameters of the model were filtered by using the mean impact value (MIV) algorithm, and then the simulation was carried out. The results show that: the prediction model optimized by MIV algorithm has higher prediction precision. The relative error of the model prediction is reduced from 13% to 7%, which is of great signifi cance to the actual injection molding production.

关 键 词：注塑成型 参数优化 正交试验 BP神经网络预测模型 MIV算法 

分 类 号：TQ327.1[化学工程—合成树脂塑料工业]