The use of large scale potts pseudo-atomic model for 3D grain growth simulation

A three-dimensional generalization of a Potts pseudo-atomic model based on a stochastic process has been implemented on a CRAY-2 computer to study large networks during long simulations. The three-dimensional network was then sampled with two-dimensional cuts, which were analyzed and compared with reality. For many parameters a satisfactory agreement between simulation and reality was observed but discrepancies still exist. In particular, the presented results show that the simulated system is somewhat anisotropic. Moreover, it appears that, given the transition mechanism, the energy function presents several local minima in which the system stabilizes.