Coupled crystal plasticity finite element and cellular automata model for simulating dynamic recrystallization in AA3104 aluminium alloy during hot rolling

A three-dimensional coupled crystal plasticity finite element (CPFE)–cellular automata (CA) framework is developed to simulate dynamic recrystallization (DRX) in aluminium alloy AA3104 under hot deformation. The model explicitly accounts for spatial variations in misorientation and dislocation density and directly incorporates real microstructural data as input. Unlike some existing approaches, no a priori assumptions are made regarding the nucleation rate, nucleation sites, or initial dislocation density. DRX nucleation is governed by local misorientation, dislocation density mismatch, and critical dislocation density criterion. The challenge of length- and time-scale coupling is addressed through an automatic time-increment control implemented within a UMAT subroutine. The approach introduces an element-wise evaluation of the DRX time step and uses it to regulate the finite element time increment, ensuring accurate tracking of DRX evolution while avoiding excessively small global time increments and maintaining computational efficiency. The numerical framework is applied to hot rolling conditions of AA3104 and successfully predicts flow stress evolution, texture development, and DRX evolution. Simulation results show good agreement with experimental observations and Avrami-based kinetic models, demonstrating the capability of the proposed approach to capture the mechanisms governing DRX in aluminium alloys.