期刊
JOURNAL OF ALLOYS AND COMPOUNDS
卷 695, 期 -, 页码 2389-2399出版社
ELSEVIER SCIENCE SA
DOI: 10.1016/j.jallcom.2016.11.131
关键词
Constitutive equation; Hot deformation; Neural networks; Dynamic recrystallization
资金
- Doctoral Research Fund of Henan Polytechnic University [64998007]
- Key Research Project of Higher Education Institution of Henan [16A430037]
Hot deformation behavior of M50NiL steel is studied via hot compression tests on the Gleeble-3500 thermal-mechanical simulator. The strain-compensated Arrhenius constitutive equation and the modified Fields-Backofen (FB) equation are developed using true stress-strain data. In order to incorporate the influence of strain on materials constants, back-propagation artificial neural networks are used to approximate materials constants. The developed Arrhenius constitutive model can predict accurately flow stress across the whole of hot working domain, describe very well the characteristics of true stress-strain curves at the stage of dynamic recovery and recrystallization, and can remove striking fluctuation caused by polynomial fitting of materials constants in predicted stress-strain curves. The FB equation can describe the work hardening, strain-rate sensitivity and temperature softening effects during hot deformation of M50NiL steel; however it cannot very well describe the real variation law with strain of flow stress at the stage of dynamic recrystallization. (C) 2016 Elsevier B.V. All rights
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据