Predictive Modelling on Machining Performance of ECDM Using Artificial Neural Network and Particle Swarm Optimization

Abstract

The electrochemical discharge machining (ECDM) process is developing into a potentially useful method of performing machining in conductive or non-conductive materials. The experiment was conducted utilizing full factorial design (2 4 ) with the 16 run test. The present study implemented artificial neural network (ANN) and particle swarm optimization (PSO) techniques to predict the Material Removal Rate (MRR) and surface roughness (SR), and the linear regression analysis were utilized for this purpose. The input parameters were adjusted with the objective of maximizing the MRR and minimizing the SR. The four selected process parameters are voltage, gap width, electrode type, and type of electrolyte, with each parameter has two levels, copper and brass were chosen as the electrodes in this particular technique. Linear regression (LR) modelling was calculated for comparison between the ANN model and LR. Artificial neural network (ANN) models have been developed to accurately predict SR and MRR, achieving a prediction accuracy of over 90%. The impacts of these findings have great significance for the future of machining tungsten carbide utilizing ECDM. Based on the results of R 2 , the modelling with ANN is closer to the actual value than the prediction using a linear regression model, the value of R 2 ANN greater than LR so that the prediction model of ANN is still the best model.