Wettability studies of formulated SMAW electrode coating fluxes with regression analysis and neural network approach

The present work aims to formulate and investigate the SMAW electrode coating fluxes for pipeline steels. The wettability of the flux on steel has a significant influence on weld properties. Twenty-seven flux compositions are produced by the extreme vertices design approach. The contact angle, spread area, work of adhesion, and floatation coefficient at 1573 K are investigated for the SiO2−CaO−CaF2−BaO−Al2O3−TiO2 fluxes. The corrected optical basicity and surface tension of the fluxes are estimated. The XRD, FTIR, and UV–vis analysis techniques are employed to investigate the phases present in the flux and slag. The structural analysis of the melt is done by examining the powder of quenched slag. The corrected optical basicity is estimated, which is a global measure of oxygen species in the melt. Regression models are developed to study the effect of flux composition on the wettability properties. An artificial neural network method is used, and the MSE and MAPE values for the predicted result have been successfully lowered.