Title: Performance characteristic prediction of WEDM process using response surface methodology and artificial neural network
Authors: P.C. Padhi; S.S. Mahapatra; S.N. Yadav; D.K. Tripathy
Siksha 'O' Anusandhan University (SOAU), Bhubaneswar 751030, India
National Institute of Technology Rourkela, 769008, India
Central Institute of Plastics Engineering and Technology (CIPET), C.N.I. Complex, Patia, Bhubaneswar, 751024, India
Kalinga Institutes of Industrial Technology, Bhubaneswar, 751024, India
Abstract: In the present study, empirical relations have been reported for estimation of performance characteristics when EN-31 steel is machined by wire electrical discharge machining (WEDM) process using response surface methodology (RSM). The experimental plan was based on the face centred central composite design (FCCCD). In order to study the effects of the WEDM parameters on performance characteristics, second order polynomial models are developed. Cutting parameters such as pulse-on-time, pulse-off-time, wire tension, spark gap set voltage and servo feed are considered as inputs to the model variables whereas cutting rate, surface roughness and dimensional deviation as outputs. Further, analysis of variance (ANOVA) is used to analyse the influence of process parameters and their interaction on responses. Artificial neural network (ANN) model based on Levenberg-Marquardt (L-M) algorithm is employed to predict the performance characteristics.
Keywords: central composite design; CCD; artificial neural networks; ANNs; analysis of variance; ANOVA; Levenberg-Marquardt algorithm; cutting rate; dimensional deviation; performance prediction; wire EDM; WEDM; response surface methodology; RSM; electrical discharge machining; electrio-discharge machining; steel machining; pulse-on-time; pulse-off-time; wire tension; spark gap set voltage; servo feed; surface roughness; surface quality.
Int. J. of Industrial and Systems Engineering, 2014 Vol.18, No.4, pp.433 - 453
Available online: 31 Oct 2014