A comparison of three modeling approaches for the prediction of microporosity in aluminum-silicon alloys

Three computational models are presented for simulating porosity formation and growth due to hydrogen evolution in 7 wt% silicon aluminum alloy (Al7Si). The first model calculates the diffusion-limited growth of an average pore in one dimension and assumes that pore growth occurs under conditions of equiaxed grain formation. The second model uses a combined continuum-stochastic approach which determines the competitive, diffusion-limited growth of a set of stochastically-nucleated pores, assuming columnar grain growth in two dimensions. The third model applies a rule-based cellular automata technique, simulating porosity and grain growth in three dimensions. Fundamental thermodynamic and kinetic equations for each of the three models are given with their limiting assumptions. The model predictions are compared to experimental in situ radiographic observations of porosity growth during the solidification of Al7Si alloy. Further comparison of the models evaluates their computational speed, accuracy and relevance.