Modelling for SAW Square Butt Joints by Using ANFIS to Predict the Weldment Characteristics of Joint

Abstract

Weldment characteristics like bead width, bead height, penetration and weldment cross-sectional areas are the factors that contribute the weld quality. Submerged arc welding (SAW) is mostly used to join the heavy structures around the world because of its high deposition rate and good weld quality. The present investigation deals with single pass double side full penetration SAW butt joints with process parameters for minimizing the welding distortions. The plates were without having any edge preparation and welded from both the sides using full factorial design of experiments with process parameters such that full penetration was achieved. The shrinkage forces from both the sides of the joints acted in opposite direction to minimize distortions. The process was modelled using adaptive neuro – fuzzy inference system (ANFIS) technique to predict the effects of process parameters on weldment characteristics and angular distortion of SAW butt joint. The ANFIS model has been developed on the basis of full factorial experimental data. The validation of ANFIS model has been also done for a number of test cases.