高洪晨,薛松,肖梦曦,等.基于正交试验与BP神经网络-遗传算法的空气净化器外壳注塑工艺参数优化[J].精密成形工程,2023,15(8):204-210.
GAO Hong-chen,XUE Song,XIAO Meng-xi,et al.Optimization of Air Purifier Shell Injection Molding Process Parameters Based on Orthogonal Experiment and BP Neural Network-Genetic Algorithm[J].Journal of Netshape Forming Engineering,2023,15(8):204-210. |
| 基于正交试验与BP神经网络-遗传算法的空气净化器外壳注塑工艺参数优化 |
| Optimization of Air Purifier Shell Injection Molding Process Parameters Based on Orthogonal Experiment and BP Neural Network-Genetic Algorithm |
| 投稿时间：2023-01-16 |
| DOI：10.3969/j.issn.1674-6457.2023.08.023 |
| 中文关键词: 注塑 优化 BP神经网络 遗传算法 翘曲变形 |
| 英文关键词: injection optimization BP neural network genetic algorithm warpage deformation |
| 基金项目:西南科技大学博士基金(13zx7153) |
|
| 摘要点击次数: 1018 |
| 全文下载次数: 490 |
| 中文摘要: |
| 目的 以某空气净化器外壳为研究对象，进行注塑工艺参数优化，从而提高塑料制品的成型质量。方法 设置4个影响塑料制品成型质量的因素:熔体温度、模具温度、保压压力、保压时间，以最大翘曲值作为衡量塑料制品成型质量的指标，通过Moldflow模流分析软件，基于正交试验及极差分析探究各因素的影响主次顺序;使用BP神经网络表征工艺参数与翘曲变形的非线性映射关系;采用遗传算法寻优获得最佳注塑工艺参数组合与翘曲变形量，并将所得参数组合用于实际生产指导。结果 经极差分析，保压压力对塑料制品质量的影响最为显著，其次分别为模具温度、保压时间、熔体温度。经BP神经网络预测与遗传算法寻优，发现当熔体温度为229.5 ℃、模具温度为77.9 ℃、保压压力为84.4 MPa、保压时间为6.5 s时，可以使注塑件达到最优质量，预测的最小翘曲值为2.94 mm，此工艺参数组合下的仿真计算翘曲值为2.91 mm，二者吻合程度较高。将优化后的工艺参数组合用于实际生产指导，获得了质量良好的注塑产品。结论 所提出的方法对产品注塑的成型及优化有良好的工程应用价值。 |
| 英文摘要: |
| The work aims to optimize the injection molding process parameters with an air purifier shell as the research object to improve the molding quality of plastic products. Four factors of great effects on the quality of injection molding were set. The maximum warpage was taken as a quality indicator of plastic products. With the help of Moldflow software and based on orthogonal experiment and range analysis, the ranking of significance effects of the four factors was obtained; the BP neural network was used to characterize the nonlinear mapping relationship between process parameters and warpage deformation. In addition, the genetic algorithm was used to find out the optimal parameter combination and warpage deformation. Finally, the resulting combination of parameters was used to guide production. The results of the range analysis showed that the packing pressure had the most significant effect on the quality of plastic products, followed by mold temperature, packing time and melt temperature. And the prediction of BP neural network and optimization of genetic algorithm showed that when the melt temperature was 229.5 ℃, the mold temperature was 77.9 ℃, the packing pressure was 84.4 MPa and the packing time was 6.5 s, the plastic part could reach the best quality. At this time, the predicted minimum warpage was 2.94 mm and the simulated minimum warpage was 2.91 mm, which were in good agreement. Finally, the optimized process parameter combination was used to guide production and plastic products of good quality were obtained. In conclusion, the proposed method has good engineering application value for optimization of injection products. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
渝公网安备 50010702501719号