Recurrent neural network estimation of material removal rate in electrical discharge machining of AISI D2 tool steel

Abstract

An Elman network is used for the prediction of material removal rate (MRR) in electrical discharge machining (EDM). An Elman network is a dynamic recurrent neural network that can be used to model non-linear dynamic systems. Training of the models is performed with data from series of EDM experiments on AISI D2 tool steel from finishing, semi-finish to roughing operations. The machining parameters such as discharge current, pulse duration, duty cycle, and voltage were used as model input variables during the development of predictive models. The developed model is validated with a new set of experimental data that was not used for the training step. The mean percentage error of the model is found to be less than 6 per cent, which shows that the proposed model can satisfactorily predict the MRR in EDM.