Predictive modelling and optimization of WEDM parameter for Mg-Li alloy using ANN integrated CRITIC-WASPAS approach.

This work intended to improve the precision and machining efficiency of Magnesium alloy (Mg-Li-Sr) using Wire electrical discharge machining (WEDM). Mg-Li-Sr alloy is prepared through inert gas assisted stir casting route. Taguchi approach is used for experimental design for WEDM parameter such as pulse OFF time, pulse ON time, wire feed rate, servo voltage and current. L27 orthogonal array is considered to understand the influence of control parameter such as Kerf Width (KW), Roughness of the surface (Ra), Material Removal Rate (MRR). Integration of the CRITIC (Criteria Importance Through Intercriteria Correlation) -WASPAS (Weighted Aggregated Sum Product Assessment) multi-objective optimization method with Artificial Neural Network (ANN) modelling with different network structure for prediction and optimization is a novel approach that significantly improves prediction accuracy and machining outcomes. The developed ANN model with better R value of 99.9 % has better ability for prediction while correlated with formulated conventional regression equation. The error percentages identified through confirmation tests for regression and ANN models are Ra - 8.5 % and 3.4 %, MRR - 5.9 % and 2.8 %, KW - 6.7 % and 2.2 % respectively. Optimal output response attained by CRITIC-WASPAS approach yields surface roughness of 4.62 μm, material removal rate of 0.073 g/min and kerf width of 0.388 μm.