EUROPEAN CONGRESS AND EXHIBITION ON ADVANCED MATERIALS AND PROCESSES - EUROMAT 2021

A physical-based model is implemented as a subroutine and used in FE simulations to predict the microstructure evolution during hot deformation and annealing of Ti alloys. During hot deformation, the material undergoes dynamic recovery of alpha and beta phases by forming new low angle grain boundaries, followed by continuous dynamic recrystallization. The microstructure evolves via static recovery or static recrystallization by strain strain-induced boundary migration during annealing. The model assumes a microstructure composed of three distinct populations of dislocations named mobile, immobile, and wall dislocations. Constitutive equations correlate the flow stress with the microstructure evolution and the flow softening in the α+β field is considered a result of the change in load partitioning. The grain sizes are related to the high angle grain boundary density. A subgrain is surrounded by low and high angle grain boundaries and is the representative microstructure entity. During deformation in the α+β domain, an initial α-lamellar structure suffers dynamic globularisation due to the formation of new boundaries within the α-platelet, and the model also predicts the evolution of this phenomena. During annealing, the static recrystallization kinetics is temperature and stored energy-dependent, and two phenomena of nucleation and growth describe the microstructure transformations. For validation of the model, the FE simulations were compared with EBSD maps measured after thermomechanical treatments. The results show that the mesoscale model predicts the non-linear microstructure evolution in complex-shaped objects during hot deformation up to large strains. Higher overall strain rates lead to higher wall dislocation production at the beginning of the deformation, smaller subgrain and grain sizes, and higher torque/stress values reached in the steady-state. The formation of subgrains, the progressive increase in boundary misorientation, and the transformation of LAGBs into new HAGBs via CDRX are well described in the proposed mesoscale model.