A solution to the problem of the mesh anisotropy in cellular automata simulations of grain growth

Abstract

Cellular automata modeling is a powerful tool used for simulating complex grain growth phenomena. However, a computational mesh may give rise to artificial anisotropy, which is a highly undesirable calculational problem. To eliminate this drawback of the approach, we have introduced two new corrections into a two-dimensional cellular automata algorithm for grain growth. The two-dimensional cellular automata model built in the framework of the approach developed by Rappaz and Gandin is based on a combination of the cellular automata and finite difference methods. The simulation results obtained for the cases of single grain growth and evolution of polycrystalline structure during solidification of alloys have demonstrated that the proposed corrections enable the mesh anisotropy problem to be solved.