摘要
为解决注塑制件成型过程翘曲变形问题,采用Moldflow软件对自动化设备电子元器件外壳注塑过程进行模流分析,但是模拟分析需要的样本数量较多,整个模拟过程缓慢。为了解决这一问题,采用拉丁超立方抽样方法对制件进行随机取样,建立RBF神经网络代理模型。通过模拟退火算法对代理模型进行全局寻优,对制件模具温度、熔体温度、保压压力以及冷却时间进行多目标优化,以制件的翘曲变形量为响应目标,获得最佳的工艺参数组合。结果表明:代理模型R2为0.92098,模拟值与预测值基本一致,误差为0.84%。通过模拟退火算法优化后,最佳的成型工艺参数保压压力为59 MPa,冷却时间为18 s,模具温度为50℃,熔体温度为240℃,此时制件翘曲量最小为0.5385 mm,通过该方法为改善制件翘曲变形提供参考。
In order to solve the problem of warpage deformation in the injection molding process,Moldflow software was used to analyze the injection molding process of the shell of electronic components of automation equipment.However,the simulation analysis requires a large number of samples,and the whole simulation process is slow.In order to solve this problem,the Latin hypercube sampling method is used to randomly sample the parts,and the RBF neural network surrogate model is established.The global optimization of the surrogate model was carried out by simulated annealing algorithm.The multi-objective optimization of mold temperature,melt temperature,packing pressure and cooling time was carried out.The warpage deformation of the part was taken as the response target to obtain the best combination of process parameters.The results show that the R2 of the surrogate model is 0.92098,the simulated value is basically consistent with the predicted value,and the error is 0.84%.After optimization by simulated annealing algorithm,the optimal molding process parameters are as follows:holding pressure is 59 MPa,cooling time is 18 s,mold temperature is 50℃,melt temperature is 240℃,and the minimum warpage of the part is 0.5385 mm.This method provides a reference for improving the warpage of the part.
作者
谭波
TAN Bo(Sichuan Vocational College of Chemical Technology,Luzhou 646300,China)
出处
《塑料科技》
CAS
北大核心
2023年第6期75-79,共5页
Plastics Science and Technology
