Accelerated Phase-Field Simulations for Static and Dynamic Recrystallization

Qi Zhang,Gang Fang

2023-01-01

Lecture Notes in Mechanical Engineering

Abstract:Due to rapid advances in computational power, various numerical methods have become feasible for simulating complex microstructure evolution including recrystallization, among which Monte Carlo (MC), cellular automaton (CA), and phase-field (PF) methods are the most commonly used in the mesoscale. Here, a PF model was established to predict the microstructure evolution during both dynamic recrystallization (DRX) and static recrystallization (SRX), which bear certain similarities in the nucleation and grain growth processes. Traditional algorithms solving the PF model usually assign each grain a unique phase-field variable to avoid grain coalescence, making the simulations computationally expensive especially when a large number of grains are present. Therefore, an efficient grain remapping algorithm was incorporated into the present PF model to accelerate the simulations, which allowed multiple grains to share the same variable and thus greatly cut down the number of variables required. Microstructure evolution during SRX and DRX was simulated, and it suggested that the algorithm yielded faithful results of the original model, and significantly reduced the computational costs.