Exploring the impact of applying activated fluxes in TIG welding of AISI 304 stainless steel

The present work analyses how multi-component activated fluxes affect weld-bead geometry during Tungsten Inert Gas (TIG) welding. TIG welding of thick materials usually requires multiple passes, thereby reducing productivity and limiting its industrial application. To address this issue, a variant of the TIG welding process, Activated Flux TIG (A-TIG) welding, was introduced. A slurry of flux is made by mixing acetone and applied to the faying surface using a flat paintbrush. The Exp. set uses a combination of SiO2, ZrO2, CaCO3 and Fe2O3 flux mixture for 8 mm-thick AISI 304 SS flat through autogenous bead-on-plate welding. In this experiment set, a design of experiment based on a two-factor, three-level full factorial methodology has been considered, where Heat Input is varied from 1.664 kJ/mm to 2.021 kJ/mm and the content of CaCO3 in the flux mixture is varied from 2–3 g in wt.%. The best depth of Penetration of 6.7 mm with an aspect ratio of 0.51 is achieved at a flux ratio of (SiO2:ZrO2:CaCO3:Fe2O3 = 2:2:3:2) and a heat input of 2.021 kJ/mm. Artificial Neural Network (ANN) modeling is done to optimize the neural network based on the performance of Levenberg-Marquardt (LM), Bayesian Regularization (BR) and Scaled Conjugate Gradient (SCG) neural network algorithms, with varying their number of nodes in hidden layers 4, 6, 8, 10 and 12, respectively.