Modeling the Electrical Parameters in EDM Process of Ti6Al4V Alloy Using Neural Network Method

Abstract

Electrical discharge machining (EDM) is a very complex and stochastic process. Thus, it is difficult to predict or to estimate its output characteristics accurately by mathematical models. Therefore, the non-conventional techniques for modeling become more effective. In this research, the Artificial Neural Network (ANN) was applied as an effective tool for modeling and predicting the surface roughness (SR), Material Removal Rate (MRR) and Tool Wear Rate (TWR). The EDM performance of Cu compact electrode have been investigated with peak current (Ip), pulse duration (ton) and duty factor (η) as the input variables. A set of experimental data was obtained with different levels. The experiments were planned and implemented using Central Composites Design (CCD) of Response Surface Methodology (RSM) with three input factors at five levels. The neural network model was built by using MATLAB. The results indicate that even with the complexity of developing a model and predicting the results in EDM process, the neural network technique is found to be adequate in predicting the SR, MRR and TWR. Predictive neural network models are found to be capable to give high accuracy.