A hybrid learning approach for modelling the fabrication of super duplex stainless steel thin joints using GTAW process

The Gas Tungsten Arc Welding (GTAW) approach is employed in this research study to join the thin sheet (1.6 mm) of Super Duplex Stainless Steel (SDSS). Process parameters such as welding current (65-85 A), voltage (12-20 V), and welding speed (250-450 mm/min) are used to compute heat input. To anticipate all potential experiments and outcomes, an artificial neural network is deployed. The Taguchi technique and grey relational analysis (GRA) is employed to optimize the welding conditions. To identify the best suitable parameters, the Taguchi L25 orthogonal array technique is employed. The parameters 1-3-4-5 have been established by GRA. The welding current, voltage, welding speed, and gas flow rate were discovered to be 65 A, 16 V, 400 mm/s, and 12 L/min, respectively. After determining the optimal conditions, the butt joint of SDSS material is fabricated using the GTAW technique. The welded joint's microstructure, mechanical characteristics (tensile test, three-point bending, and hardness), and corrosion rate has been investigated. Based on the acquired experimental data, artificial intelligence tools have evolved and the capability of evolved tools has been analyzed. The outcomes of the analysis prove that the model is competent enough to predict the desired process variables.