Acta Metallurgica Sinica (English Letters) ›› 2025, Vol. 38 ›› Issue (8): 1275-1292.DOI: 10.1007/s40195-025-01885-3

Previous Articles     Next Articles

Hot Deformation Behavior of CoNiV Medium-Entropy Alloy: Constitutive Model, Convolutional Neural Network, Hot Processing Map, and Microstructure Evolution

Biao Zhang1, Yuntian Du2, Huishuang Jia2, Yuanyi Zhou1, Liguang Wang1, Minghe Zhang1(), Yunli Feng1, Weimin Gao1, Ning Xu3   

  1. 1Key Laboratory of Ministry of Education for Modern Metallurgy Technology, North China University of Science and Technology, Tangshan 063210, China
    2College of Science, North China University of Science and Technology, Tangshan 063210, China
    3Branch of Iron and Steel Research Institute, Ansteel Beijing Research Institute Co., Ltd., Beijing 102209, China
  • Received:2024-12-20 Revised:2025-02-10 Accepted:2025-02-26 Online:2025-06-17 Published:2025-06-17
  • Contact: Minghe Zhang

Abstract:

This study systematically investigates the hot deformation behavior and microstructural evolution of CoNiV medium-entropy alloy (MEA) in the temperature range of 950-1100 °C and strain rates of 0.001-1 s−1. The Arrhenius model and machine learning model were developed to forecast flow stresses at various conditions. The predictive capability of both models was assessed using the coefficients of determination (R2), average absolute relative error (AARE), and root mean square error (RMSE). The findings show that the osprey optimization algorithm convolutional neural network (OOA-CNN) model outperforms the Arrhenius model, achieving a high R2 value of 0.99959 and lower AARE and RMSE values. The flow stress that the OOA-CNN model predicted was used to generate power dissipation maps and instability maps under different strains. Finally, combining the processing map and microstructure characterization, the ideal processing domain was identified as 1100 °C at strain rates of 0.01-0.1 s−1. This study provided key insights into optimizing the hot working process of CoNiV MEA.

Key words: Hot deformation, Arrhenius model, Machine learning, CoNiV MEA, Hot processing map