Thermo-Physical Interpretation and Predictive Optimization of Material Removal Rate in Die-Sinking Electrical Discharge Machining of Hardened SKD61 Steel Using an Using a Response Surface Methodology–Particle Swarm Optimization Framework

That Nguyen Huu1, Chien Vu Ngoc1,
1 Faculty of Mechanical Engineering, Nha Trang University, Khanh Hoa, Vietnam

Main Article Content

Abstract

Electrical discharge machining (EDM) is indispensable for producing complex cavities in hardened tool steels, yet reliable productivity prediction remains challenging owing to the strongly nonlinear dependence of material removal rate (MRR) on discharge parameters. This study investigates MRR in die-sinking EDM of hardened SKD61 steel (50 HRC) using a cylindrical copper electrode. A Taguchi L27 orthogonal array was employed to characterize the effects of discharge current (5–11 A), gap voltage (50–70 V), and pulse-on time (Tₒₙ; 3–7 machine-scale units) on MRR. A full second-order response surface model (RSM) was fitted by least-squares regression, achieving a coefficient of determination R² = 0.995 (Adj-R² = 0.993) and confirming model adequacy by lack-of-fit analysis and residual diagnostics. Regression analysis revealed that discharge current is the dominant factor; its significant quadratic term and the negative current×pulse-on-time interaction term are consistent with the thermal-erosion mechanism of EDM, specifically progressive plasma-channel widening and debris-induced reduction in spark efficiency at prolonged high-energy discharge conditions. The validated RSM was integrated with particle swarm optimization (PSO; 40 particles, 50 iterations) to maximize MRR within the experimental domain. The optimal setting—I = 11 A, V = 50 V, Tₒₙ = 3—yielded a predicted MRR of 38.33 mm³/min, a 1.68 % improvement over the best experimental result. The RSM–PSO framework is computationally efficient, physically interpretable, and directly transferable to multi-response EDM optimization problems.

Article Details