摘要
基于BP神经网络,以挤压速度、挤压道次和挤压方式为输入层参数,以抗拉强度为输出层参数,构建了BP神经网络模型用于分析ECAP强变形对Cu-3Cr合金性能影响,并进行了试验验证以及金相组织和SEM分析。结果表明,BP神经网络输出的抗拉强度预测值与试验值之间的相对误差均小于2%,平均预测误差为1.4%,模型的预测精度高、实用性强。从改善合金抗拉强度出发,Cu-3Cr合金的最佳ECAP工艺参数:挤压速度为5mm/s、挤压道次为4次、挤压方式为每次挤压后旋转180°再进入下一道次。
Based on BP neural network technology,the parameters of extrusion speed,extrusion passes and extrusion method were designed for the input layer,and the tensile strength was designed as the output layer parameter,thus BP neural network model was established for analyzing effects of ECAP strong deformation on properties of Cu-3Cr alloy.Meanwhile,experimental verification of metallurgical microstructure ovservation and SEM was carried out.The results show that the relative error between the predicted and experimental values of the tensile strength between BP neural network outputs and experimental value is less than 2% with the average prediction error of 1.4%,showing a higher prediction accuracy of the model.From the view of improving tensile strength of the alloy,the optimized ECAP process parameters for Cu-3Cr alloy were presented as follows：extrusion speed of 5mm/s,the extrusion pass of 4,and the extrusion method with rotating 180°and then entering the next pass after one squeeze.
出处
《特种铸造及有色合金》
CAS
CSCD
北大核心
2015年第3期240-243,共4页
Special Casting & Nonferrous Alloys
基金
浙江省教育厅科研基金资助项目(FG2014115)
浙江省教育厅基金资助项目(JG2013332)
