Electrical Discharge Machining (EDM) is a complex process due to the involvement of numerous machining indicators. The key input machining indicators affect machining efficiency, tool material, dielectric fluid, and electrical parameters. The conductive silicon carbide and aluminium powder, having a size of 50 microns is used for the improvement of the machining efficiency of EDM. The machining is carried out on AISI D3 material, which is difficult to machine by conventional techniques. The machining performance of EDM is identified by measuring the Surface Roughness (SR). Moreover, research explores the optimization of surface roughness prediction for silicon carbide and aluminium powder using Particle Swarm Optimization (PSO) combined with machine learning algorithms. The performance of the Random Forest (RF), Decision Tree (DT), and Gradient Boosting (GB) is evaluated, trained and tested on respective datasets, with hyperparameters optimized using PSO to minimize prediction errors. For the Aluminium Powder dataset, the Decision Tree model demonstrated superior performance with the lowest Mean Squared Error (MSE) of 0.0603, indicating its ability to accurately capture the dataset’s feature relationships. Conversely, the Gradient Boosting model excelled on the Silicon Carbide dataset, achieving the best MSE of 0.1021, attributed to its robust ensemble learning approach. The significant differences in model performance across the two datasets underscore the importance of dataset-specific optimization and model selection. The findings of the research work highlight the effectiveness of PSO in enhancing model accuracy and emphasize the necessity of modified strategies for different materials in predictive analytics. Moreover, it showed that aluminium powder results are better as compared to silicon carbide powders on Surface roughness analysis.
Huu-Phan et al. (Fri,) studied this question.