Surface Roughness Prediction of mechanical parts in ultrasonic vibration-assisted finishing machining

Abstract

Surface roughness is an important quality factor of mechanical parts. When finishing machining with the support of ultrasonic vibration, the surface roughness of the part depends on some parameters including frequency f (kHz), voltage V(V) supplied to the piezo vibration head, and the rotation speed n (rpm) of the machine tool spindle. In this study, the authors conducted an experimental design and proposed two regression models to evaluate the effectiveness of predicting the surface roughness of mechanical parts when finishing machining with the support of ultrasonic vibration with different sets of factor parameters. Those are linear regression using the Minitab software tool and non-linear regression using the Random Forest (RF) algorithm. The results show that the linear regression model using Minitab produces false prediction results of approximately 20,1%, and is 7,5% lower than the results obtained from the non-linear regression model using the Random Forest algorithm.