Prediction of cutting forces in MQL turning of AISI 304 Steel using machine learning algorithm

Abstract

Cutting force play a significant role in enhancing the machining performance as it affects the cutting tool life, surface finish generated and also the energy consumed in obtaining the final product. The machining cost is reduced considerably by effectively minimizing the cutting forces. Minimum quantity lubrication (MQL) is a technique by which cutting fluid is employed in the machining zone in the form of mist, thereby reducing the wastage of cutting fluid and improving the machinability of the process. In this paper, AISI 304 steel is machined using carbide tool in alumina nanoparticle enriched lubrication environment. The calculation of average cutting force is done by varying the input parameters namely cutting speed, feed rate, depth of cut and nanoparticle concentration respectively. The design of experiment is made using response surface methodology (RSM) and further analysis of variance is performed. Furthermore three machine learning based models namely linear regression (LR), random forest (RF) and support vector machine (SVM) are used for predicting the cutting force and comparing the experimental value with that of the predicted value. For accessing the performance of the predicted values, three different error metrics were used namely, coefficient of determination (R2), mean absolute percentage error (MAPE) and mean square error (MSE) respectively. The predicted values obtained by linear regression model for cutting forces are more accurate as compared to other models.