EDM processing of LPBF AlSi10Mg lattice structures with data-driven modeling and multi-objective optimization

The study investigates Electrical Discharge Machining (EDM) of LPBF-fabricated AlSi10Mg lattice structures using an additively manufactured copper (AM-Cu) electrode. An L18 orthogonal array was employed to evaluate eight process parameters, including polarity, pulse-on/off time, current, sensitivity, spark gap, dielectric pressure and lifting time. Increasing current from 2 to 4 A enhanced the material removal rate (MRR) from 0.052 to 0.225 g/min, while pulse-on time also significantly influenced MRR. The ANN–NSGA-II hybrid model achieved (R2>0.96) for MRR, tool wear rate (TWR) and surface roughness (SR). Optimization yielded an MRR of 0.228 g/min and SR of 4.65 µm, representing a 41% MRR improvement. TWR ranged from 0.0008 to 0.00245 g/min, while debris exhibited resolidified globular morphology. Porosity varied from 19.758% to 41.552%. Negative polarity improved MRR but increased SR, whereas positive polarity provided better surface quality, demonstrating the potential of machine-learning-assisted EDM optimization for lattice post-processing.