Machine Learning-Based Predictive Modelling of EDM and EAM-V Processes for Performance Analysis

The promotion of Electrical Discharge Machining (EDM) and vibration aided Electric Arc Machining (EAM-V) processes is characterized in the study in terms of their capability for precision manufacture, mainly drawing any performance comparisons from a machine learning approach. The present machine learning study aims to predict some important metrics of machining utility, such as Material Removal Rate (MRR), Tool Wear Rate (TWR), and Surface Roughness (SR), against process parameters like current, pulse-on/off time, etc. Some advanced models like Gradient Boosting and Random Forest are used to analyse the efficacy and effectiveness of EDM and EAM-V, comparing the respective influences these parameters have on honing outcomes. The study describes an elaborate methodology: data collection, preprocessing, feature scaling, and application of multiple regression algorithms for machining performance forecasting. The experimental data for model training and testing were partitioned into 80% and 20%, respectively. The results revealed that Gradient Boosting (GB) performed better than Random Forest (RF) for all parameters. In GB, the R² values of MRR, TWR, and SR were higher; hence, its degree of accuracy was superior in comparison with RF. For instance, an R² value of 0.970, 0.994, and 0.999 was achieved by GB for MRR, TWR, and SR, respectively, thus proving its better predictive ability. Moreover, according to average predicted values, EAM-V performs better for MRR; EDM, comparatively, from TWR and SR, is more suitable for precision applications. The performance validation of GB through RMSE and MAE also confirms its efficacious predictions.