Papers by Keyword: Cellular Automaton Method

Abstract: A cellular automaton (CA) model has been developed for simulating the microstructure evolution and dendrite morphology of cast magnesium alloys. The growth kinetics of dendrite tips is determined by the difference between local equilibrium composition and local actual composition obtained by solving the solute transport equation. Two sets of meshes, a hexagonal mesh and an orthogonal mesh, are used in the model to perform the simulation. The hexagonal mesh is used to perform CA calculation to reflect the texture of Mg alloy dendrites, and the orthogonal mesh is used to solve the solute transport equations. The model was applied to simulate single dendrite evolution and columnar dendrites growth of AZ91D Mg alloy, as well as multi-grain growth of Mg-10Gd-2Y-0.5Zr (wt%) Mg alloy. Permanent mold step-shaped castings of the two Mg alloys were poured and metallographic examinations were carried out for validating the present model. The simulation results agree well with metallographic results. The model can be applied to simulate the microstructure evolution and dendrite morphology of magnesium alloys.

Abstract: For the purpose of the prediction of casting structures, heterogeneous nucleation rate in the undercooled melt of solififying Al-Si alloys were evaluated by comparing experimentally observed macrostructures of solidified ingots with numerically simulated ones. Molten alloys were unidirectionally solidified in an adiabatic mold from a steel chill block located at the bottom of the mold. In the experiment, columnar to equiaxed transition (CET) was observed. A numerical simulation for grain structure formation of the sample ingots was carried out using a cellular automaton (CA) method, and heterogeneous nucleation rate in the solidifying alloys were evaluated by producing the similar structures to experimental ones. An attempt was made to predict the grain structure of conventionally cast ingots using the evaluated heterogeneous nucleation rate. However, the simulation could not predict the structure of ingot with low superheat due to crystal multiplication near the mold wall. The crystal multiplication mechanism, so-called "Big Bang mechanism", was introduced into the simulation and the simulation could predict the grain macrostructure composed of columnar and equiaxed crystals that were similar to experimentally observed one.