ABSTRACT
The supervised machine learning (SML) algorithms have been implemented for predicting the rate of material removal (RMR) and the surface roughness (Ra) during dry turning of graphene-based aluminum metal matrix composite with polycrystalline diamond (PCD) carbide-insert tool. The composite specimen are synthesized by reinforcing graphene particles with 0 (pure aluminum), 0.5, 1, 2 and 3 wt.% in pure aluminum matrix via stir casting process. Experiments have been conducted according to the full factorial design considering the speed, feed, depth of cut, and wt.% of graphene as the controlled parameters. This experimentally acquired dataset has been utilized to find the best performing regression model using the back-propagation artificial neural network (ANN) and Supervised Regression Methodologies (Random Forest Regression [RFR], Decision Tree Regression [DTR] and Light Gradient Boost Regression [LGBR]) for accurate prediction of RMR and Ra, based on current data as input. The surface quality and RMR are found to be improving with wt.% of graphene reinforcement. Also, the RFR model is found to provide comprehensive accuracy in terms of all the aspects and therefore, can be endorsed as a suitable model for prediction of performance measures during the current dry turning operation.
