Artificial neural network modelling of weld bead characteristics during GMAW of nitrogen strengthened austenitic stainless steel

Nitrogen strengthened austenitic stainless steels is gaining its popularity in replacing the SS 304 component by cheaper AISI 201 grade in many industrial application including repair and maintenance due to its outstanding combination of mechanical property and corrosion resistance. Gas metal arc welding (GMAW) is preferred in fabrication industry due to its mechanization and high productivity. In this work, welding experiments of GMAW of AISI 201Gr stainless steel plate was carried out using response surface methodology (RSM) of 27 experimental runs. Wire feed rate (WFR), voltage (V), nozzle to plate distance (NTD) and welding speed (S) are considered as weld parameters; each of three different levels. The weld bead characteristics such as penetration (P), width (W), reinforcement (R), weld penetration shape factor (WPSF), weld reinforcement form factor (WRFF) are considered. The modelling of weld characteristics is carried out by a popular soft computing approach i.e., Artificial Neural Network (ANN). The developed ANN models have been validated for its adequacy and it shows less percentage of error. The proposed models will be useful for online implementation of ANN model in fabrication industry.