Predicting the Various Responses in EDM of Ti6Al4V Using Deep Neural Network

This research aims to develop modeling and prediction for the electrical discharge machining (EDM) process of Ti6Al4V with a graphite electrode. The three models have been developed for the prediction of material removal rate (MRR), electrode wear ratio (EWR), and surface roughness (SR) using two artificial neural network models (ANN_sigmoid and ANN_Relu) and a deep neural network model (DNN). Herein, the four inputs of the network were the discharge current, duty factor, pulse on time, and voltage. Taguchi’s L27 (34) orthogonal array was used to designed and conducted the experiments. Comparing the values predicted with the experimental data indicated that all neural network models provided accurate results. Moreover, this study showed that the DNN model gave the best performance and can successfully predict the MRR, EWR, and SR of the stochastic and complex EDM process.