Machine Learning Hot Deformation Behavior of Nb Micro-alloyed Steels and Its Extrapolation to Dynamic Recrystallization Kinetics

In the present paper, a novel method was developed to predict the dynamic recrystallization (DRX) behavior during hot rolling based on machine learning in combination with physical models of niobium (Nb) micro-alloyed steels. It has been demonstrated that the comprehensive modeling could accurately predict the DRX behavior under different compositional and processing conditions, which not only exhibited higher precisions but also was in better agreement with physical metallurgical principles than traditional methods.