基于Moldflow和BP神经网络的薄壳注塑件预测  

Prediction of Thin-shell Injection Molded Parts Based on Moldflow and BP Neural Network

作　　者：修辉平[1] 徐敏[1] 刘晓红[1] 

机构地区：[1]九江职业技术学院机械工程学院,江西九江332007

出　　处：《塑料科技》2017年第5期69-72,共4页Plastics Science and Technology

摘　　要：基于Moldflow平台,通过正交试验研究了一模两腔薄壳注塑件成型工艺参数对翘曲量的影响。以流动速率、保压压力、保压时间、熔体温度和模具温度五个工艺参数为输入,注塑件翘曲量参数作为输出,在Matlab神经网络工具箱中建立三层BP神经网络;利用Moldflow正交试验所得工艺参数与翘曲量的数据作为训练样本对神经网络进行训练,得到注塑工艺参数与注塑件翘曲量之间的非线性映射关系,并验证该神经网络的泛化能力。结果表明,该神经网络能较好地预测薄壳注塑件翘曲量。Based on the Moldflow platform,the influence of forming parameters on the warpage of the injection mold with two cavities was studied by orthogonal test.The BP neural network with three layers which takes flow rate,packing pressure,packing time,melt temperature and mold temperature as input parameters and the injection molding warpage as output paramers was established in Matlab neural network toolbox.Training the neural network with data of process parameters and warpage obtained through orthogonal experiment in Moldflow,get the nonlinear mapping relationship between process parameters of injection molding and warpage,and verify the generalization ability of the neural network.The results show that the neural network can predict the warpage of thin-shell plastic parts.

关 键 词：MOLDFLOW 正交试验 薄壳注塑件 神经网络 翘曲 

分 类 号：TQ320.662[化学工程—合成树脂塑料工业]