Real-time data monitoring of steel cladding on mild steel and spatter recognition in WAAM using machine learning

In the present study, wire arc additive manufacturing (WAAM) cladding parametric characteristics and spatter recognition were analyzed using a machine learning process. Additionally, data acquisition software was employed to record the variations in WAAM parameters during cladding, and machine learning was used for spatter detection. The cladding parameters, such as voltage, were higher than the prefix voltage; the current decreased as the voltage increased at a similar wire feed rate; and the power increased with higher voltage at the same wire feed rate. It was observed that ceramic MMCs, SiC, could absorb more heat while increasing the dilution across the clad width in the melt pool due to higher temperature stability at elevated temperatures. Spatters are unpredictable but can be minimized by using appropriate cladding parameters when steel is clad on a mild steel substrate. Furthermore, it was observed that spatters increased from mild steel, AISI304 stainless steel, to SiC-AISI304 stainless steel cladding on a mild steel substrate. Finally, it was also found that the heat-affected zone is lighter in SiC-AISI304 stainless steel cladding due to the higher amount of heat absorbed by SiC.