基于BP-SSA神经网络模型的中框塑件翘曲变形优化  

Optimization of the Warping Deformation of Mid-frame Plastic Components Based on BP-SSA Neural Network Mode

作　　者：杨明 侯健超 刘巨保[1] 姚建锋 王帅 廉成林 YANG Ming;HOU Jianchao;LIU Jubao;YAO Jianfeng;WANG Shuai;LIAN Chenglin(School of Mechanical Science and Engineering,Northeast Petroleum University,Daqing 163318,China;Hanking Plastic Manufactory(Shenzhen)Co.,Ltd.,Shenzhen 518117,China)

机构地区：[1]东北石油大学机械科学与工程学院,黑龙江大庆163318 [2]弘丰塑胶制品(深圳)有限公司,广东深圳518117

出　　处：《塑料工业》2024年第7期97-104,共8页China Plastics Industry

摘　　要：为减小挂壁风机中框在注塑成型中的翘曲变形,运用Moldflow软件对中框进行仿真,以模具温度、熔体温度、注射时间、冷却时间、保压时间和保压压力6个工艺参数进行正交试验设计,基于试验结果建立工艺参数与翘曲变形量之间的反向传播(BP)神经网络模型,利用麻雀搜索算法(SSA)对模型进行全局参数寻优。结果表明,当模具温度为80℃、熔体温度为250℃、保压压力为82 MPa、保压时间为20 s、注射时间为3 s、冷却时间为26.25 s时翘曲变形量最小,预测翘曲变形量为2.483 mm。利用Moldflow软件对寻优得到的工艺参数进行验证,结果显示仿真计算的翘曲变形量为2.449 mm,与优化算法结果误差为1.3%,并将寻优获得的参数进行试模验证,试模翘曲结果与优化算法结果误差为2.6%,证明了优化算法的准确性,且注塑件外观无飞边缩痕等缺陷,装配效果符合预期要求。研究结果表明BP神经网络结合麻雀搜索算法优化工艺参数的技术方法具有可行性。To mitigate the warping deformation of the mid-frame in the injection molding process of the wall-mounted fan,Moldflow software was used to simulate the mid-frame.An orthogonal experimental design was employed for six process parameters(mold temperature,melt temperature,filling time,solidification time,holding time,and holding pressure).Based on the experimental results,the back-propagation(BP)neural network model between process parameters and warping deformation was established.The model was optimized using the sparrow search algorithm(SSA)for global parameter optimization.The results indicate that when the mold temperature is 80℃,the melt temperature is 250℃,the holding pressure is 82 MPa,the holding time is 20 s,the injection time is 3 s,and the cooling time is 26.25 s,the warping deformation is minimized and its predicted value is 2.483 mm.Validation of the optimized process parameters using Moldflow software reveals a simulated warping deformation of 2.449 mm,resulting in a discrepancy of 1.3%between the predicted and actual values.The parameters obtained from the optimizations were verified by the test mold.The error between the result of the test mold warping and the result of the optimization algorithm is 2.6%,which proves the accuracy of the optimization algorithm.The appearances of the injection parts have no defects such as flash shrinkage,and the assembly effects meet the expected requirements.The research findings suggest the feasibility of utilizing the BP neural network in conjunction with the sparrow search algorithm to optimize process parameters.

关 键 词：注塑成型 翘曲变形量 模流分析 参数优化 反向传播神经网络 麻雀搜索算法 

分 类 号：TQ320.66[化学工程—合成树脂塑料工业]