基于神经网络和遗传算法的薄壳件注塑成型工艺参数优化

Optimization of Process Parameters for Thin Plastic Shell injection molding Based on Neural Network and Genetic Algorithm

建立了基于神经网络和遗传算法并结合正交试验的薄壳件注塑成型工艺参数优化系统。正交试验法用来设计神经网络的训练样本,人工神经网络有效的创建了翘曲预测模型;遗传算法完成了对影响薄壳塑件翘曲变形的工艺参数(模具温度、注射温度、注射压力、保压时间、保压压力和冷却时间等)的优化,并计算出了它们的优化值。按该参数进行试验,效果良好,可以有效地减小薄壳塑件翘曲变形,其试验数值与计算数值基本相符,说明所提出的方法是可行的。

This paper establishes an optimization system for thin plastic shell injection molding based on artificial neural network combined with orthogonal test and genetic algorithm .The sample for neural network model is designed by using orthogonal experimental method. A predictive model for warpage is created using artificial neural network. Through genetic algorithm, the injection molding parameters that affect thin-wall mould warpage, namely the mold temperature, melt temperature, filling pressure, packing pressure, packing pressure time and cooling time, etc,are optimized.and their optimum values are obtained as well. The experimental effect was excellent with the obtained parameters and the thin-wall mould warpage can be efficiently decreased, Which prove the optimizing method adopted is feasible.