Optimization of WEDM parameters for stir -cast Al6082–TiB ₂ –Gr–Mg hybrid composite using taguchi and machine learning approaches

Abstract This study investigates the optimization and prediction of Wire Electrical Discharge Machining (WEDM) performance for stir-cast Al6082–TiB 2 –Gr–Mg hybrid metal matrix composites. Pulse-on time (TON), pulse-off time (TOFF), wire feed rate (WF), and TiB 2 content were analyzed to maximize material removal rate (MRR) and minimize surface roughness (SR) using a Taguchi L27 orthogonal array, Analysis of Variance (ANOVA), and Grey Relational Analysis (GRA). Experimental results reveal that TON is the most influential parameter affecting both MRR and SR, contributing 48.48% and 45.34%, respectively. The optimal machining condition (TON = 15 μs, TOFF = 5 μs, WF = 9 m min −1 , TiB 2 = 9 wt%) yielded a maximum MRR of 28.812 mm 3 min −1 , while the minimum SR of 1.32 μm was obtained at 3 wt% TiB 2 . To enhance predictive capability, Linear Regression and Random Forest models were developed and evaluated using 5-fold cross-validation. Linear Regression exhibited better generalization (R 2 = 0.729 for MRR and 0.585 for SR), indicating predominantly linear parameter–response relationships. The novelty of this work lies in integrating Taguchi optimization, GRA, and machine-learning-based prediction within a unified framework. The findings provide practical guidelines for precision machining of aluminium hybrid composites in aerospace and automotive applications.