基于神经网络的工艺参数对翘曲变形和收缩的影响研究

摘要

以拉伸和冲击试样(无缺口和有缺口)三种塑料零件的注射成型为例,以翘曲变形和收缩为评价指标,采用Taguchi方法及极差和方差分析方法,优化了模具温度、熔体温度、注塑压力、注塑时间、保压压力、保压时间和冷却时间,获得了最优的工艺参数组合.建立了神经网络模型,利用神经网络的预测功能,预测出变动单个工艺参数下的翘曲变形量和收缩率,研究了单个工艺参数对翘曲变形和收缩的影响,以指导生产实践.%Taking tensile specimen,impact specimen (no gap or without gap) for example,with warpage and shrinkage as the optimization objective,the mold temperature,melt temperature,injection pressure,injection time,packing pressure,packing time and cooling time were optimized by using the Taguchi,range analysis and variance analysis methods.The optimal injection molding process parameters were determined.A neural network model was developed and then the warpage and shrinkage under changing a single process parameter were predicted by the predicted function of neural network.The influence of process parameters on warpage and shrinkage were studied,which were useful for the practical production.