Improved Precision in Grain Size Prediction by Thermal Analysis and Statistical Analysis in Aluminium Castings

摘要

Thermal analysis has been used for decades for melt control before casting aluminum alloys. However, obtaining a good grain refinement in a standard cup does not ensure that the grain refinement is correct in real parts which may solidify at very different cooling rates. For this study, the effect of cooling rate on AlSi7Mg alloy with different metal qualities in terms of grain refinement was tested. A mathematical model has been built for prediction of grain size in aluminium castings, knowing the following parameters: grain refining potential of the melt, evaluated by using thermal analysis (Thermolan~R-Al system) and cooling rate of the selected zone in the casting, evaluated in terms of DAS by metallographic measurement or simulation tools. The model was first developed by the correlation between nucleation potential in term of grain size measured on the standard thermal analysis cup with those obtained in cylindrical test parts with various diameters cast in sand moulds. It was found possible to set up a formula for the grain size of the cylindrical test parts as a function of their modulus and of the nucleation potential evaluated on the standard thermal cups. A correlation between modulus and DAS (metallographic and simulation prediction) is established also for the cylinders. The new model has been validated in steering knuckles cast with melts having different grain refining potential. Good correlations, with R2 factors ranging from 0.85 to 0.91 were obtained when comparing measured grains sizes with the corresponding calculated values in cylinder casting and steering knuckles. Thus a simple thermal analysis of the melt prior to casting allows to predict if the desired grain size will be obtained at different locations of the part or if corrective actions must be taken in order to improve the grain refinement potential of the melt.