Multi-objective Optimization of Welding Parameters in Submerged Arc Welding of API X65 Steel Plates

Abstract

Submerged arc welding (SAW) is one of the main welding processes with high deposition rate and high welding quality. This welding method is extensively used in welding large-diameter gas transmission pipelines and high-pressure vessels. In welding of such structures, the selection process parameters has great influence on the weld bead geometry and consequently affects the weld quality. Based on Fuzzy logic and NSGA-II (Non-dominated Sorting Genetic Algorithm-II) algorithm, a new approach was proposed for weld bead geometry prediction and for process parameters optimization. First, different welding parameters including welding voltage, current and speed were set to perform SAW under different conditions on API X65 steel plates. Next, the designed Fuzzy model was used for predicting the weld bead geometry and modeling of the process. The obtained mean percentage error of penetration depth, weld bead width and height from the proposed Fuzzy model was 6. 06%, 6. 40% and 5. 82%, respectively. The process parameters were then optimized to achieve the desired values of convexity and penetration indexes simultaneously using NSGA-ll algorithm. As a result, a set of optimum vectors (each vector contains current, voltage and speed within their selected experimental domains) was presented for desirable values of convexity and penetration indexes in the ranges of (0. 106, 0. 168) and (0. 354, 0. 561) respectively, which was more applicable in real conditions.