Integration of GA and neuro-fuzzy approaches for the predictive analysis of gas-assisted EDM responses

This research work discusses the application of three intelligent prediction models, based on artificial neural network (ANN) with back-propagation algorithm, adaptive neuro-fuzzy inference system (ANFIS) and hybrid ANFIS and genetic algorithm (ANFIS-GA). These techniques are used for prediction and comparison of machining aspects such as material removal rate (MRR) and surface roughness during gas-assisted electrical discharge machining of D3 die steel. In the present work, helium-assisted EDM with perforated tool has been performed. In this work, parameters considered for machining are discharge current, pulse on time, duty cycle, tool rotation and discharge gas pressure. The suggested approach is based on up-gradation of ANFIS with GA. The GA algorithm is applied to improve the precision of the ANFIS model. The soft computing models were trained, tested and validated with experimental data. Mean square error (MSE), mean absolute error (MAE), root-mean-square error and correlation coefficient (R2), were used to measure the efficacy of models predicting abilities developed through ANN, ANFIS and hybrid ANFIS-GA approaches. The experiment and anticipated measure of MRR and SR of the process, acquired by ANN, ANFIS and hybrid ANFIS-GA, was found to be in good agreement. The prediction potential of proposed models was tested using new set of data for the training and testing process. The ANFIS-GA technique provides more accurate prediction of the responses in comparison with the ANN and the ANFIS. In general, the inference of this work discloses that the hybrid algorithm like ANFIS-GA is an efficient and effective approach for precise prediction of EDM process responses.