Optimizing submerged arc welding using response surface methodology, regression analysis, and genetic algorithm

• Main Authors: Ajitanshu Vedrtnam, Gyanendra Singh, Ankit Kumar
• Format: Article
• Language: English
• Published: KeAi Communications Co., Ltd. 2018-06-01
• Series: Defence Technology
• Online Access: http://www.sciencedirect.com/science/article/pii/S2214914717302441

The weld quality is significantly affected by the weld parameters (arc voltage, welding current, nozzle to plate distance and welding speed) in the submerged arc welding (SAW). Bead-on-plate welds were performed on stainless steel plates by automated SAW machine. The experimental data were collected in accordance with the response surface methodology (RSM). In addition to RSM, the regression analysis was performed to set up input–output relationships in the SAW process. It was found that weld parameters define the geometry of weld bead and determine the mechanical properties of the joint. The influence of the input variables on weld bead geometry is represented as graphs. It was found that an increment in voltage increases the bead width but decreases the bead height, whereas the current increment result-in an increment in bead height and no change in bead width. The bead width and height decrease with the increment in the welding speed. With an increment in the nozzle-to-plate distance, bead width decrease, but bead height increases. The value of bead hardness increases with the increment in current but the increment in voltage and travel speed does not have a significant influence on the bead hardness. The predictions from the mathematical model developed and the corresponding experimental results are having a fair agreement. Further, the genetic algorithm (GA) is also used for predicting the weld bead geometry. Keywords: SAW, Bead hardness, RSM, Regression, Optimization