NOUFERROUS METALLURGICAL EQUIPMENT

Aluminum alloys are widely used in aerospace and transportation fields due to their high strength and low weight. In this paper, with aluminum alloy components as the input vector and tensile strength as the target variable, four different machine learning algorithm models, RF, ET, Bagging and Adaboost, are established. The results show that RF model has the best prediction performance, R=0.89, MAE=40.33. The content of Ti plays a positive role in predicting the tensile strength of aluminum alloys. The higher the content of Ti, the greater the tensile strength. The prediction of tensile strength of aluminum alloy with Mg and Cu content is not obvious. The content of Zn element, Ce element and Y element plays a negative role in predicting the tensile strength of aluminum alloy. The element content is great, that the tensile strength value is small. The importance of the characteristics is Ti＞Mg＞Cu＞Zn＞Ce＞Y.