Surface Roughness Prediction of AISI 304 Steel in Nanofluid Assisted Turning Using Machine Learning Technique

Abstract

Machining is a complex process which uses cutting tool for finshing the workpiece material. A sequence of machining tests costs a lot of expense and effort to complete. It's critical to avoid time-consuming runs and put technology first. Surface roughness (Ra) has been used to signal quality of product in the turning process as part of an automated monitoring system deployed in-process. This research uses machine learning models to estimate surface roughness while machining AISI 304 stainless steel rods. The key elements impacting surface quality are the input variables of turning, namely feed rate, depth of cut, and spindle speed. Four machine learning (ML)-based algorithms were used to predict surface roughness in this study: Gradient Boosting Regression (GBR), Decision Tree Regression (DTR), Extreme Gradient Boosting Regression (XGB), and Random Forest (RF) of Surface Roughness (Ra). The baseline models' predictive ability was measured using error measures such as Root Mean Square Error (RMSE), mean squared error (MSE), and coefficient of determination (R2). Overall, the XGB and GBR models appear to have the most accuracy in predicting surface roughness (Ra).