Numerical Simulation On Solidification Microstructures Using Cellular Automaton Method

Solidification process and microstructures formation during solidification have attracted researchers from different fields for a long time because they have not only profound background of engineering application, but also significant theoretical value. With the development of computer technology, numerical simulation on solidification has made great progress. In the last decade, based on the success of macro simulation on temperature field, microstructure simulation has become the major focus.On the basis of previous researches, a solute diffusion controlled model for dendritic growth is developed and relative numerical method is provided. Cellular automaton method is applied to transform the sharp change at the solid-liquid interface to gradual change of solid fraction in interfacial cells, which avoids direct front tracking while sharp interface hypothesis is still kept. The effects of constitutional undercooling, curvature undercooling, anisotropy of interfacial energy and interface perturbation are considered in this model.The model is applied to simulate steady growth behavior in undercooled melt and the simulated results are in agreement with the prediction of theoretical model for tip growth. Branching mechanism and competition growth of side branches are simulated by imposing perturbation at the interface and the effects of perturbation wavelength and amplitude on the degree of side branch are studied. The results are in agreement with marginal stability theory.The model is extended to simulate constrained growth in the directional solidification. By imposing different combinations of temperature gradient and solidification rate, the typical interface morphology including planar, cellular and dendritic are successfully simulated as well as the branching and adjustment of primary arm. The variation of primary arm spacing with the temperature gradient and solidification rate are simulated respectively and agree with the prediction of theoretical model.Combined with the nucleation model based on normal distribution, the dendritic growth model is directly applied to simulate grain structures, which avoids the artificial geometry hypothesis in traditional grain structures simulation. In this way, simulation of grain structures and dendritic growth is integrated with cellular automaton method, which greatly improves the capability and the modeling scale of this method.Equiaxed grain structure normally solidified in metal mold is simulated with coupling of heat transfer calculation. By imposing different cooling rates, temperature gradients and nucleation parameters, the columnar-to-equiaxed transition is simulated and the effects of processing parameters on the competition growth between columnar and equiaxed grains are analyzed. The results are in agreement with practical process and theoretical analysis.