Experimental investigations on hybrid manufacturing: WEDM of WAAM-fabricated stainless steel components using ANFIS modelling

Wire Arc Additive Manufacturing (WAAM) enables the fabrication of large, near-net-shape stainless steel components, but the resulting surfaces require precision post-processing to meet industrial standards. In this study, Wire Electrical Discharge Machining (WEDM) was applied as a finishing process for WAAM-fabricated SS316L components, and a hybrid optimization–prediction framework was developed using Taguchi design, Grey Relational Analysis (GRA), and Adaptive Neuro-Fuzzy Inference System (ANFIS) modeling. In total, there were 27 experimental runs conducted at different pulse-on, pulse-off, and current conditions. The results showed that pulse-on time (Ton) was the dominant influencing factor in the case of material removal rate (MRR), dimensional deviation (DD), and GD&T errors, while pulse-off time (Toff) was significantly regulated to surface roughness (SR) and geometric stability. The experimental analysis revealed that pulse-on time (Ton) was the most influential parameter governing material removal, dimensional accuracy, and geometric errors, whereas pulse-off time (Toff) played a key role in controlling surface finish and geometric stability. This emphasizes the critical importance of discharge control for achieving high-quality post-processing of WAAM components. For multi-response optimization, GRA provided a composite performance index that was used to train the ANFIS model. The predictive outcomes exhibited excellent agreement with experiments, confirmed by very low error metrics (MAPE = 2.19%, RMSE = 0.027, MAE = 0.022) and a strong correlation (R² = 0.9985). Overall, the WAAM–WEDM hybrid framework not only improves surface quality and dimensional consistency but also establishes a scalable, intelligent manufacturing pathway with strong potential for aerospace, biomedical, and energy applications.