Modeling of EDM parameter with novel W/O nano-emulsion

• Published in: Journal of the Brazilian Society of Mechanical Sciences and Engineering 2024-09, Vol.46 (9), Article 584
• Main Authors: Jain, Nikhil, Jain, Jinesh Kumar, Unune, Deepak Rajendra
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Dielectric properties; Electric discharge machining; Emulsions; Engineering; Error analysis; Impact analysis; Material removal rate (machining); Mathematical models; Mechanical Engineering; Microhardness; Nanoemulsions; Parameter identification; Polydispersity; Response surface methodology; Surface roughness; Surfactants; Synthesis; Technical Paper
• ISSN: 1678-5878; 1806-3691
• DOI: 10.1007/s40430-024-05160-x

This study investigates the potential of a novel water-in-oil nano-emulsion (W/O) synthesized from non-edible refined jatropha oil as a dielectric fluid to enhance the electric discharge machining (EDM) process. W/O nano-emulsion was synthesized in two different surfactant concentrations, namely 10% and 15%. Influence of different input parameters, including current, pulse-on time, pulse-off time, and the concentration of the dielectric fluid, was explored on the key output responses, namely the material removal rate (MRR) and surface roughness. After evaluating its dielectric properties and observing a smaller particle size (46.33 nm) with a polydispersity index of 0.103, the 15% surfactant concentration was selected for EDM testing in comparison with conventional EDM oil. A mathematical model was developed using a face-centered cubic design model of the response surface methodology, and the results were subsequently compared with those obtained from an artificial neural network (ANN) model. The quadratic mathematical model for both MRR and surface roughness closely aligned with the experimental data, with current identified as the most influential parameter affecting both MRR and surface roughness. Pulse-off time, on the other hand, exhibited minimal impact on EDM performance. Furthermore, error analysis was conducted to compare the accuracy of the RSM and ANN models, with the ANN model demonstrating the least error. Notably, the machined surface exhibited fewer cracks and pores in the presence of the proposed (W/O) emulsion and displayed improved micro-hardness compared to conventional EDM oil.