Microstructure Simulation Of Magnesium Alloy

Magnesium is the lightest structural metal with high specific strength and stiffness. In transportation area, the application of magnesium alloy achieves the demand for saving weight which makes it as a kind of metal material with most bright future. To a great extent, performance of magnesium alloy castings depends on the microstructure formed in the solidification. Fine grains can dramatically improve its formability and increase its comprehensive properties. Normally, by adding grain refiner or enhance the cooling conditions, fine grain structures can be obtain. However, the relationship between processing parameters and grain size is very complex. Numerical simulation technology is an important and effective means to predict the structure evolution during the solidification process with low cost, then to optimize the processing parameters.A probabilistic method-cellular automaton method with square cells, von Neumann neighborhood configuration for microstructure simulation of magnesium alloys under uniform temperature field was developed. During the solidification process, nucleation was controlled by the Gaussian distribution model; the crystallographic orientation was taken randomly; the growth kinetics of dendrite tips were obtained by solving KGT model and the relationship between undercooling and growth velocity of dendritic tips was established with a third order polynomial;"Corner Growth Algorithm" was proposed and applied to determine the capture and state transition of the cell. By using this new method, the growth of a single grain was simulated. The grain always retain its original preferential growth orientation and will not bias to the CA lattice which shows its ability in avoiding artificial anisotropy, and it is very easy to realize growth of random crystallographic orientation. The simualtion of multiple grains was executed. The result shows that this model can reveal the growth competition occuring among the columnar grains and the columnar to equiaxed transition. The good agreement of both experimental and simulated result shows the efficiency of this method.The assumption of uniform temperature field limits the application of the model. So a fully coupled finite difference-cellular automaton model is developed for the simulation of magnesium alloys. The model has two sets of meshes—macro finite difference meshes and micro CA lattices. The macro finite difference meshes are used to calculate the temperature field and the micro cells are used to model the microstructure evolution. The latent heat released by the cells is feed back to macro nodes. A fully new mechanism is introduced to control the liberation of latent heat during the calculation. The latent heat starts to liberate when the temperature drops below virtual liquidus which is defined as the nucleation temperature of a cell. The nucleation parameters were experimentally determined and the boundary conditions were adjusted. Based on these parameters, the solidification process of the entire casting is simulated. Similar with the actual cooling process, the predicted cooling curves display distinct slope change at a certain temperature below liquidus. Although detailed morphology of the dendrite is not considered in the present, the simulated grain densities agree well with the experimental investigations.At last, the effects of heterogeneous particles on the micro structure are simulated. The more the amount of the heterogeneous particles and the smaller the nucleation undercooling, the finer is the grain size. The growth restriction caused by the heterogeneous particles and the solute segregation will enhance grain refinement.