Data Mining Methods for Prediction of Multi-Component Al-Si Alloy Properties Based on Cooling Curves

Abstract

The paper concerns the mechanical properties of hypoeutectic Al-Si alloy (silumin) with the addition of Cr, Mo, V and W. Changes in microstructure under the impact of these elements result in a change in the mechanical properties. Crystallization of Al-Si alloys determines grain size reduction, which causes a significant increase in their strength properties. Crystallization subjected to modifications through the influence of alloying additives can be described by the cooling curve run. Statistical relationships between the characteristic values of cooling curves and mechanical properties are investigated with data mining techniques of regression, especially regression trees. Such knowledge could provide an ability of a property prediction on the basis of cooling curves in terms of the benefits of a short time of the curve registration.