A full-scale Monte Carlo Potts model and real time conversion under non-uniform temperature distribution

Abstract

Prediction of microstructure evolution is important to design structural materials, and grain growth is the most fundamental but important one which determines mechanical properties. The Monte Carlo Potts is a widespread simulation model for grain growth with high computational efficiency but has not been considered suitable for the quantitative prediction of practical processes since the physical meaning of variables is not fully understood yet. Efforts have been made to assign realistic physical meaning to variables. The existing equation for converting simulation time to real time involves a strong temperature dependence, which causes a problem under non-uniform temperature distribution, as in additive manufacturing. This problem could be solved by assigning a temperature dependence to the simulation result, which was possible by using realistic temperature and energy values. We found that the simulated activation energy for grain growth could be equated to the experimental value by adjusting the activation energy for grain boundary migration within a reasonable range in the simulation. The modified model made the time conversion equation completely temperature-independent and was free from the above-mentioned problem in time conversion. The Monte Carlo Potts simulation can now more reasonably predict grain growth in various materials under various temperature conditions.