Supervised Machine Learning Approach for Modeling Hot Deformation Behavior of Medium Carbon Steel

Metal forming process parameters selection highly depends on the consistent and realistic characterization of material behavior under the combined effects of strain, strain rate, and temperature on the material flow stress. Hot deformation tensile tests are performed for AISI 1045 steel at deformation temperatures and strain rates ranges from 650 to 950 °C and 0.05 to 1.0 s⁻¹, respectively. The received flow curves indicate that flow stress increases with a decrease in deformation temperature and an increase in strain rate. In this study, it is investigated the supervised machine learning techniques such as support vector regression, single decision tree, and random forest regression (RFR) models to characterize material-flow behavior during hot deformation. Overall, the proposed RFR model results are in good agreement with the experimental observations. Besides, the proposed model's predictability is assessed using graphical and numerical validations. The numerical quantification confirms that the RFR models perform significantly better with a higher coefficient of determination (R ²), 0.9983, and low prediction error, 1.021%. Furthermore, it is revealed through the comparison with previous findings, that the proposed machine learning models can precisely calculate flow stress better than conventional models.