Construction of hot deformation processing maps for 9Cr-1Mo steel through conventional and ANN approach

Abstract

The behavior of hot deformed flow curves in a 9Cr-1Mo steel have been analysed using constitutive and artificial neural network (ANN) approaches. At each temperature (850 ° C-1100 ° C) various strain rates (0.001 s−1–10 s−1) of deformation behavior has been analysed. The corrections of flow stress due to the adiabatic temperature rise during hot deformation have been made through artificial neural network (ANN) method and compared it with the linear interpolation (conventional) method. The temperature corrected (conventional) flow stress has been compared and analyzed with the predicted flow stress data obtained by established constitutive equation. Additionally, the experimental flow stresses obtained by thermomechanical simulator have also been compared and analyzed with predicted flow stresses obtained by ANN method. Further, the predicted and corrected flow stress curves obtained by ANN method have been compared and analyzed. The processing maps at 0.6 true strain based on flow stress data corrected through both ANN and conventional methods have been compared. Upon analyzing the safe and unsafe domains of the maps with the corresponding microstructures, the ANN approach depicted the best precise results over that of the conventional method. Variation in strain hardening rate and correlated it with EBSD analysis have also been carried out to analyze the variation in hardness for some of the deformed specimens.