Using a back propagation neutral network based modeling and heuristic algorithms based optimization technique in activated gas tungsten arc welding process

Abstract

In this study, a modeling method based on an artificial neural networks model combined with a back propagation algorithm (BPNN) and an optimization procedure based on heuristic algorithms (particle swarm optimization (PSO) and simulated annealing (SA) algorithms) have been proposed for modeling and optimization of activated gas tungsten arc welding (A-GTAW) process in order to tackle the poor penetration drawback occurs during GTAW process. in this study effect of the most important process adjusting variables including welding current (C), welding speed (S)) and percentage of activating fluxes (TiO2 and SiO2) combination (F) on the most important quality characteristics (weld bead width (WBW), depth of penetration (DOP), and consequently aspect ratio (ASR)) in welding of AISI316L austenite stainless steel parts have been investigated. Box-behnken and central composite designs (BBD and CCD) based on response surface methodology (RSM) in design of experiments (DOE) method have been employed to gather the required data for modeling and optimization purposes. Then, BPNN has been used to determine the relations between A-GTAW process input variables and output responses. To determine the proper BPNN model architecture (the proper hidden layers’ number and their corresponding neurons/nodes in each layer) PSO algorithm has been used. Next, PSO and SA algorithms have been used to optimize the proposed BPNN model in such a way that desired AR, minimum WBW, and maximum DOP achieved. Finally, confirmation experimental tests have been conducted to evaluate the proposed procedure performance. Based on the results, the proposed method is efficient in modeling and optimization (less than 7% error) of A-GTAW process.