Prediction of surface roughness using semi-empirical and regression models in machining of metal matrix composites using abrasive waterjet

Prediction of surface roughness has always been a challenging task in the abrasive waterjet machining (AWJM) of metal matrix composites (MMCs), since it involves quite a number of process and material variables. The present study explores the prediction of surface roughness (Ra) through development of models, namely regression and semi-empirical models for AWJM of unreinforced Al 6063 alloy, Al 6063 reinforced with 5%, 10%, and 15% boron carbide (B4C) particles. Accordingly, AWJM trials were performed using Box-Behnken experimental design by keeping the process variables viz. mesh size (ms), mass flow rate (ma), water pressure (p), and traverse speed (u) each at three levels. Response surface methodology (RSM) was applied to formulate a regression model using the data generated from experiments. The results analyzed through analysis of variance (ANOVA) revealed that ms, water pressure, and traverse speed render a significant contribution in the generation of Ra. The experimental investigations showed that the combination of process parameters such as ms of # 120, mass flow rate of 340 g/min, water pressure of 275 MPa, and traverse speed of 60 mm/min have resulted in lower Ra. By utilizing both the experimental data and materials properties data, a semi-empirical model was developed through dimensional analysis by applying Buckingham’s π-theorem. Predictions of the developed models were in good correlation with the data obtained through experiments under corresponding conditions. The study undertaken here proved that the semi-empirical models could efficiently predict the Ra in AWJM of MMCs.