Enhancing WEDM performance on Mg/FeCoCrNiMn HEA composites through ANN and entropy integrated COCOSO optimization

Abstract

The aim of this experimental work is to find the ideal wire electric discharge machining (WEDM) parameter combination for processing a novel FeCoCrNiMn High Entropy Alloy (HEA)-reinforced magnesium composite. This composite is developed with varying weights of FeCoCrNiMn at 5%, 10%, and 15% through powder metallurgy. Experiments are performed to examine the effects of HEA and wire-EDM variables on surface roughness (Ra) and kerf width (KW) using Taguchi’s L27 orthogonal array. The hybrid ENTROPY-COCOSO (Combined Compromise Solution) methodology is used for multiple objective optimizations after the Taguchi method for optimization. The most significant constraints on Ra and KW are found to be pulse ON time and current. Wider kerfs and rougher surfaces are the result of longer pulse ON times and higher current. The ideal input parameters recommended by ENTROPY-COCOSO for minimal Ra and KW are 2 A of current, 20 µs of pulse ON time, 25 µs of pulse OFF time, and 4 mm/min of wire feed rate. To predict outcomes, both linear regression models and artificial neural networks (ANNs) are used, and the results are compared with experimental data. The results are validated by the fact that ANN predictions closely match experimental data with minimal deviation.