Surface integrity and multi-objective optimization of Ti-6Al-4V alloy by EDM using a WC/Cu powder metallurgy electrode: an integrated RSM–ANN approach

Abstract

This study examined the electro-discharge machining (EDM) performance of Ti-6Al-4V alloy utilizing a WC/Cu powder metallurgy electrode using hybrid response surface methodology (RSM) and artificial neural network (ANN) techniques. The effects of peak current ( I p ), pulse-on time ( T on ), pulse-off time ( T off ), and discharge voltage ( V g ) on material removal rate (MRR) and surface roughness (Ra) were analyzed using Box–Behnken design. Regression models developed using RSM demonstrated excellent prediction accuracy with the coefficient of determination ( R 2 ) values of 99.89% for MRR and 99.88% for Ra. ANOVA results confirmed that I p and T on were the most significant parameters influencing machining performance. MRR increased from 13.5 to 30 mm 3 /min as I p increased from 5 to 15 A, while Ra increased from 2.4 to 4.3 μm owing to the increased spark energy and crater formation. The ANN model with a 4-10-2 architecture demonstrated outstanding prediction performance with an overall regression coefficient ( R ) of 0.99054 and a minimal validation mean squared error of 0.36696, confirming strong agreement between predicted and experimental data. Multi-response optimization using the desirability function approach identified optimal machining conditions at I p = 14.0909 A, T on = 54.0404 μs, T off = 15.6566 μs, and V g = 80 V, yielding a predicted MRR of 23.942 mm 3 /min and an Ra value of 3.7608 μm with a composite desirability of 0.9782. SEM and EDS analyses confirmed enhanced surface integrity with shallow craters, fewer microcracks, and Cu and W element deposition on the machined surface. The outcomes demonstrated that the hybrid RSM–ANN technique effectively optimized EDM performance and improved machining characteristics of Ti-6Al-4V alloy using a WC/Cu PM electrode.