Scholarlink

<p>This article reports on the findings obtained from the hot deformation behaviour of AISI 304 stainless steel. Uniaxial compression tests were done using the Gleeble<sup>®</sup> thermal mechanical equipment. The test conditions were a deformation temperature range of 950–1050℃ and a strain rate range of 0.1–15&#xa0;s⁻¹. The study analysed the metal flow pattern and compared the prediction accuracy of the Arrhenius, strain-compensation, and physical and Artificial Neural Networks (ANN) models using statistical parameters: the correlation coefficient <i>R</i> and average absolute relative error <i>AARE</i>. The results show that flow stress increases with a decrease in the deformation temperature and an increase in strain rate, and vice versa. The predicted data obtained using the ANN model accurately tracks the experimental data throughout the entire loading condition range. However, the constitutive model analyses show a marked deviation from experimental data. The statistical parameters <i>R</i> and <i>AARE</i> analysis were: the <i>R</i>-values 0.994 (Arrhenius), 0.994 (strain compensated), 0.980 (Physical model) and 0.998 (ANN), and the <i>AARE</i>-values 15.05% (Arrhenius), 17.32% (strain compensated), 4.78% (Physical model) and 1.96% (ANN). The statistical analyses indicate that the trained ANN model exhibited the highest prediction accuracy for predicting the flow stress behaviour of the AISI 304 stainless steel.</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

Hot workability behaviour of AISI304 stainless steel: constitutive and ANN modelling

  • 期刊论文

  • 发表日期：

    2025年10月30日

查看全文
  • Japheth Obiko,
  • Brendon Mxolisi,
  • Malatji Nicholus

摘要

This article reports on the findings obtained from the hot deformation behaviour of AISI 304 stainless steel. Uniaxial compression tests were done using the Gleeble® thermal mechanical equipment. The test conditions were a deformation temperature range of 950–1050℃ and a strain rate range of 0.1–15 s⁻¹. The study analysed the metal flow pattern and compared the prediction accuracy of the Arrhenius, strain-compensation, and physical and Artificial Neural Networks (ANN) models using statistical parameters: the correlation coefficient R and average absolute relative error AARE. The results show that flow stress increases with a decrease in the deformation temperature and an increase in strain rate, and vice versa. The predicted data obtained using the ANN model accurately tracks the experimental data throughout the entire loading condition range. However, the constitutive model analyses show a marked deviation from experimental data. The statistical parameters R and AARE analysis were: the R-values 0.994 (Arrhenius), 0.994 (strain compensated), 0.980 (Physical model) and 0.998 (ANN), and the AARE-values 15.05% (Arrhenius), 17.32% (strain compensated), 4.78% (Physical model) and 1.96% (ANN). The statistical analyses indicate that the trained ANN model exhibited the highest prediction accuracy for predicting the flow stress behaviour of the AISI 304 stainless steel.

Graphical Abstract