A computational model for the prediction of porosity due to dissolved hydrogen in binary aluminum-silicon alloys has been developed. The model combines the cellular automata technique for the simulation of the growth of the solid phase, the finite-difference technique for the simulation of diffusion of the dissolved species, and a quasi-equilibrium model for the growth pf individual bubbles. The growth of the solid and gas phases is initiated by a stochastic nucleation model, depending upon the undercooling (for the solid) or the supersaturation ratio (for the gas). The results agree favorably with experiments. The low supersaturation values needed to simulate the experimental results are consistent with a nucleation mechanism of gas pockets entrained within the melt.
展开▼
