Modeling of Wear Behavior of Al–Si/Al2O3 Metal Matrix Composites

Abstract

Artificial neural network (ANN) approach, analysis of variance (ANOVA), and multiple regression model were developed to predict the wear rate for the aluminum (Al)-silicon (Si) alloy. These methods were based on weight fractions of alumina (Al₂O₃), load, and sliding distance as inputs. Metal matrix composites (MMCs) were prepared using stir casting method. The Al–Si alloy was reinforced with the addition of 0, 10, 15, 20, and 25 wt % of Al₂O₃ particles. The ANN model was utilized to predict the wear rates of the composites. Experimental results indicated that the increase of both load and sliding distance increases the wear rate. However, the increase of weight fractions of alumina (Al₂O₃) decreases the wear rate. Both ANN and ANOVA revealed that the sliding distance has the major influence on the wear rate in comparison with the factor of alumina weight fraction. However, the applied load has a relatively low influence on the wear rate of Al–Si/Al₂O₃ composite. A multiple regression approach suggested in this study reveals the correlation between the weight fractions of Al₂O₃, load, and sliding distance and the wear rate.