基于BP神经网络的双色塑件翘曲变形量预测
|
摘要
以某遥控器前壳双色塑件注塑成型为例,以该塑件在注塑成型过程中的翘曲变形量为研究目标,提出了一种结合AMI数值模拟、正交试验和BP神经网络的双色塑件翘曲变形量快速、准确的预测方法。首先建立了基于AMI数值模拟的CAE模流分析模型,并对注塑成型工艺参数及翘曲变形量进行数值模拟分析;之后结合正交试验设计法使AMI软件数值模拟结果在指定的工艺参数范围内实现了离散分布;最后以正交试验数据为基础建立BP神经网络预测模型,通过Matlab训练网络使其满足误差精度要求,从而达到准确预测新工艺参数下翘曲变形量的目的。结果表明:训练出的BP神经网络模型具有很高的预测精度,能够满足对该双色塑件翘曲变形量准确、快速的预测要求。
Taking the injection molding of a remote double color front shell as an example and focusing on the study of warpage deformation the injection molding process of the plastic. A fast and accurate prediction method for the warpage deformation of two-colored plastics was proposed by combining AMI numerical simulation, orthogonal test and BP neural network. Firstly, the CAE flow analysis model was established based on AMI numerical simulation and the injection molding process parameters and warpage deformation were analyzed by numerical simulation. Then the discrete distribution of AMI numerical simulation results was realized within the range of specified process parameters combining with the orthogonal test method. Finally, A BP neural network prediction model was established based on the orthogonal test result data, and trained using by Matlab software to meet the error precision requirements,thus to achieve the purpose of accurately predicted warp deformation under the new process parameters combination. The results show that the BP neural network model has high prediction accuracy and can meet the accurate and rapid prediction requirements of the remote double color front shell warp deformation.
Taking the injection molding of a remote double color front shell as an example and focusing on the study of warpage deformation the injection molding process of the plastic. A fast and accurate prediction method for the warpage deformation of two-colored plastics was proposed by combining AMI numerical simulation, orthogonal test and BP neural network. Firstly, the CAE flow analysis model was established based on AMI numerical simulation and the injection molding process parameters and warpage deformation were analyzed by numerical simulation. Then the discrete distribution of AMI numerical simulation results was realized within the range of specified process parameters combining with the orthogonal test method. Finally, A BP neural network prediction model was established based on the orthogonal test result data, and trained using by Matlab software to meet the error precision requirements,thus to achieve the purpose of accurately predicted warp deformation under the new process parameters combination. The results show that the BP neural network model has high prediction accuracy and can meet the accurate and rapid prediction requirements of the remote double color front shell warp deformation.
引文
[1]冯刚,田雅萍,张朝阁.双色注射成型技术研究进展[J].工程塑料应用,2014,42(8):111-115.
[2]黄海龙,李晓春.基于AMI与正交试验的餐盘翘曲变形优化[J].塑料,2016,45(4):92-94.
[3]胡开元,王雷刚,黄瑶.注塑工艺参数对双色注射成型影响分析[J].现代塑料加工应用,2009,21(6):37-40.
[4]Zhu Chundong,Li Futao,Zhu Hongyan.Optimizing the injection molding process parameters based on orthogonal test and moldflow[J].China Plastics Industry,2009(6):30-32.
[5]陆广华.基于BP神经网络的注塑成型翘曲分析及工艺优化[J].塑料工业,2017,45(2):58-60.
[6]李瑞娟,黄力.基于BP神经网络的汽车内饰面板注塑成型工艺参数优化[J].塑料,2016,45(3):81-85,35.
[7]黄海龙,周亮.基于正交试验的遥控器双色前壳翘曲变形优化[J].轻工机械,2017,35(3):79-82.
[8]Guan Zhijie,Lu Jun.Optimization design of technological parameters for surface quality of injection molding product based on moldflow and orthogonal experiment[J].Engineering Plastics Application,2010,38(8):36-38.
[9]Zhao Guo,Xin Yong.Optimization of parameters in injection molding based on multi-target comprehensive evaluation[J].Engineering Plastics Application,2008,36(1):31-34.
[10]张成军.实验设计与数据处理[M].北京:化学工业出版社,2011:8-51.
[11]崔如坤.薄壁配光镜的翘曲变形控制及工艺优化[D].吉林:吉林大学,2016:45-51.
[12]周浩文.基于人工神经网络的注塑成型模拟及工艺参数优化的研究[D].广州:广东工业大学,2013:34-60.
[2]黄海龙,李晓春.基于AMI与正交试验的餐盘翘曲变形优化[J].塑料,2016,45(4):92-94.
[3]胡开元,王雷刚,黄瑶.注塑工艺参数对双色注射成型影响分析[J].现代塑料加工应用,2009,21(6):37-40.
[4]Zhu Chundong,Li Futao,Zhu Hongyan.Optimizing the injection molding process parameters based on orthogonal test and moldflow[J].China Plastics Industry,2009(6):30-32.
[5]陆广华.基于BP神经网络的注塑成型翘曲分析及工艺优化[J].塑料工业,2017,45(2):58-60.
[6]李瑞娟,黄力.基于BP神经网络的汽车内饰面板注塑成型工艺参数优化[J].塑料,2016,45(3):81-85,35.
[7]黄海龙,周亮.基于正交试验的遥控器双色前壳翘曲变形优化[J].轻工机械,2017,35(3):79-82.
[8]Guan Zhijie,Lu Jun.Optimization design of technological parameters for surface quality of injection molding product based on moldflow and orthogonal experiment[J].Engineering Plastics Application,2010,38(8):36-38.
[9]Zhao Guo,Xin Yong.Optimization of parameters in injection molding based on multi-target comprehensive evaluation[J].Engineering Plastics Application,2008,36(1):31-34.
[10]张成军.实验设计与数据处理[M].北京:化学工业出版社,2011:8-51.
[11]崔如坤.薄壁配光镜的翘曲变形控制及工艺优化[D].吉林:吉林大学,2016:45-51.
[12]周浩文.基于人工神经网络的注塑成型模拟及工艺参数优化的研究[D].广州:广东工业大学,2013:34-60.