A predictive mesoscale model for continuous dynamic recrystallization

Abstract

Thermomechanical processing of titanium alloys often requires complex routes to achieve the desired final microstructure. Recent advances in modeling and simulation tools have facilitated the optimization of these processing routes. However, existing models often fail to accurately predict microstructural changes at large deformations. In this study, we refine the physical principles of an existing mean-field model and propose a calibration method that uses experimental results under isothermal conditions, accounting for the actual local deformation within the workpiece. This new approach improves the predictability of microstructural changes due to continuous dynamic recrystallization during torsion and compression experiments. Additionally, we integrate the model into the commercial FEM-based DEFORM™ 2D software to predict the local microstructure evolution within hot torsion specimens thermomechanically treated by resistive heating. Validation using non-isothermal deformation tests demonstrates that the model provides realistic simulations at high strain rates, where adiabatic heat modifies temperature, flow stress and microstructure. This study demonstrates the intrinsic correlation between microstructure, flow behavior, and workpiece geometry, considering the impact of deformation history in thermomechanical processes.