Machine Learning Approach of Bio Silica Suspended Dielectric in EDM of Ti6Al4V for Biomedical Application

Abstract

A machine learning technique, the artificial neural network is used to predict the output responses of Ti6Al4V in electrical discharge machining, using nano bio silica infused vegetable oil methyl ester dielectric fluid and optimizing the operating parameters of electrical discharge machining by using the Jaya algorithm. The input parameters of the experiments are peak current, pulse on time, discharge voltage, and duty cycle, and the measured output responses are the material removal rate, tool wear rate, and surface roughness. The optimal architecture of the artificial neural network model is recognized as 4-10-10-3. The correlation coefficient of the artificial neural network prediction is 0.9626, and the least mean absolute percentage error of the material removal rate, tool wear rate, and surface roughness are 0.8129, 0.3337, and 1.2595%, respectively. The artificial neural network predicted accurate results and the Jaya algorithm optimized the operating parameters of Ti6Al4V in electrical discharge machining.