Comparative investigation of Artificial Neural Network (ANN) and Response Surface Methodology (RSM) expectation in EDM parameters

Abstract

In this research work, a correlation of various modeling approaches Artificial Neural Network (ANN) and Response Surface Methodology (RSM) is being studied. The models were created dependent on three-level models of investigations in Design Expert directed on EN-8 work material. Peak Current, Pulse on time, Di-electric Pressure and Tool diameter as the Electric Discharge Machining (EDM) parameters are considered as input and Material Removal Rate (MRR) are studied as the output. A back propagation neural network system was created to build the process model. The execution of the created ANN models and the RSM models are studied in contrast with the optimization. The investigation of this examination demonstrates that the ANN-RSM approach is fit for foreseeing the processed coefficient of assurance (R2), Root Mean-Square Error (RMSE) and Percentage of Average Absolute Deviation (AAD) for RSM model. In contrast with that of R2, RMSE, and Percentage of AAD for ANN is demonstrated with the high proportion of RSM model over the ANN model.