(1. 西北工业大学 材料科学与工程学院,西安 710072;
2. 河南科技大学 材料科学与工程学院, 洛阳 471003)
摘 要: 利用神经网络对Cu-Cr-Zr合金变形量、 时效温度和时间与硬度和电导率样本集进行训练和学习,采用改进的BP网络算法—Levenberg-Marquardt算法, 建立了形变热处理工艺BP神经网络模型, 得出了具有较高综合性能的最佳工艺参数: 在80%变形量, 450~480℃, 2~5h形变热处理条件下, 硬度和电导率分别可达HV150~157和74%~77%(IACS)。
关键字: 铜合金; 形变热处理; 神经网络; Levenberg-Marquardt算法
alloy by artificial neural network
(1. College of Materials Science and Engineering,
Northwestern Polytechnical University, Xi′an 710072, China;
2. College of Materials Science and Engineering,
Henan University of Science and Technology, Luoyang 471003, China)
Abstract:By using a supervised artificial neural network(ANN), Levenberg-Marquardt algorithm, the non-linear relationship between parameters of thermomechanical treatment processes and properties such as hardness and conductivity of Cu-Cr-Zr alloy was analyzed. A basic repository on the domain knowledge of thermomechanical treatment processes is established via sufficient data mining by the network. The results show that the ANN system is effective and successful for predicting and analyzing the properties of Cu-Cr-Zr alloy.The optimum process parameters are obtained for the thermomechanical treatment. By aging at 450~480℃ for 2~5h after 80% deformation, the hardness and conductivity reach HV150~157 and 74%~77%(IACS) respectively.
Key words: copper alloy; thermomechanical treatment; artificial neural network; Levenberg-Marquard algorithm