Eutectoid growth, as the important reaction mechanism of the carbon steel heat treatment, is the basis to control the microstructure and performance. At present, most studies have focused on lamellar growth, and did not consider the nucleation process. Mainly due to the nucleation theory is inconclusive, a lot of research can support their own theory in a certain range. Based on the existing nucleation theory, this paper proposes a cooperative nucleation model to simulate the nucleation process of eutectoid growth. In order to ensure that the nucleation process is more suitable to the theoretical results, different correction methods were used to amend the model respectively. The results of numerical simulation show that when the model is unmodified, the lateral growth of single phase is faster than that of longitudinal growth, so the morphology is oval. Then, the effects of diffusion correction, mobility correction and ledges nucleation mechanism correction on the morphology of nucleation and the nucleation rate were studied respectively. It was found that the introduction of boundary diffusion and the nucleation mechanism of the ledges could lead to a more realistic pearlite.
The main work of this paper focuses on the simulation of binary alloy solidification using the phase field model and adaptive octree grids. Ni-Cu binary alloy is used as an example in this paper to do research on the numerical simulation of isothermal solidification of binary alloy. Firstly, the WBM model, numerical issues and adaptive octree grids have been explained. Secondary, the numerical simulation results of three dimensional morphology of the equiaxed grain and concentration variations are given, taking the efficiency advantage of the adaptive octree grids. The microsegregation of binary alloy has been analysed emphatically. Then, numerical simulation results of the influence of thermo-physical parameters on the growth of the equiaxed grain are also given. At last, a simulation experiment of large scale and long-time has been carried out. It is found that increases of initial temperature and initial concentration will make grain grow along certain directions and adaptive octree grids can effectively be used in simulations of microstructure.