Scholarlink

<p>In this paper, Back Propagation (BP) neural network, BP neural network improved by Genetic Algorithm (GA-BP), and BP neural network enhanced by Particle Swarm Optimization (PSO-BP) were employed to predict the conductivity of Cu-Fe-P alloy, aiming to optimize the hot rolling process parameters. The data set was filtered using Mutual Information method and Pearson Correlation Coefficient method to identify input features (rolling force, rolling temperature, rolling speed, roll gap) highly correlated with conductivity. The results indicate that BP, GA-BP, and PSO-BP neural network models can effectively predict the conductivity of copper alloy under different rolling process parameters. Among them, the GA-BP neural network exhibits the highest prediction accuracy (MSE = 0.040, <i>R</i><sup>2</sup> = 0.93), leading to the optimization of hot rolling process parameters. Employing Electron Backscatter Diffraction (EBSD) technique to analyze the microstructure of samples before and after optimizing the hot rolling process, the results indicate that the microstructure of the samples after optimization contains more low-angle grain boundaries (LAGBs) and larger average grain size.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimization of Hot Rolling Process and Conductivity Prediction for Cu-Fe-P Alloy Based on GA-BP and PSO-BP Neural Networks

  • 期刊论文

  • 发表日期：

    2026年1月26日

查看全文
  • Yong Hu,
  • Shuyan Liu,
  • Shangyuan Jiang,
  • Qingyang Duan,
  • Shaohua Wang,
  • Zhijie Yan,
  • Xin Song,
  • Xin Liu

摘要

In this paper, Back Propagation (BP) neural network, BP neural network improved by Genetic Algorithm (GA-BP), and BP neural network enhanced by Particle Swarm Optimization (PSO-BP) were employed to predict the conductivity of Cu-Fe-P alloy, aiming to optimize the hot rolling process parameters. The data set was filtered using Mutual Information method and Pearson Correlation Coefficient method to identify input features (rolling force, rolling temperature, rolling speed, roll gap) highly correlated with conductivity. The results indicate that BP, GA-BP, and PSO-BP neural network models can effectively predict the conductivity of copper alloy under different rolling process parameters. Among them, the GA-BP neural network exhibits the highest prediction accuracy (MSE = 0.040, R2 = 0.93), leading to the optimization of hot rolling process parameters. Employing Electron Backscatter Diffraction (EBSD) technique to analyze the microstructure of samples before and after optimizing the hot rolling process, the results indicate that the microstructure of the samples after optimization contains more low-angle grain boundaries (LAGBs) and larger average grain size.

Graphical Abstract