A model based on free-growth control of grain initiation can quantitatively predict the grain size in aluminium alloys inoculated with commercial refiner as a function of refiner addition level, solute content in the melt and cooling rate. This model is applied to predict the performance of refiners with hypothetical size distributions of nucleant particles. The distributions are log-normal, like those in commercial refiners, but the mean particle diameter and the relative spread of the distribution are varied to map the performance of the refiner in giving minimum grain size, maximum efficiency, or minimum sensitivity of grain size to cooling rate. The optimal nucleant particle size distribution is a compromise between these aims and depends on the alloy and processing conditions. This mapping can assist in choosing between available refiners or in developing improved refiners. Existing commercial refiners are highly effective, but their nucleant particle size distribution leaves room for improvement.
Read full abstract