Investigation of GMAW Parameters and Plate Thickness Effects on Weld Bead Geometry and Development of Prediction Model Using RSM and ANN

Abstract

The influence of gas metal arc welding (GMAW) parameters such as current, voltage, electrode extension, and welding speed on weld bead geometry (bead width, bead height, penetration) has been investigated using Response Surface Methodology (RSM). Through a designed experiment matrix generated by RSM, comprehensive experimentation has been conducted. The effects and effectiveness of these factors have been analyzed across three workpieces of varying thicknesses for quality control testing. For each geometric or physical characteristic mentioned above, a corresponding regression model has been developed. Furthermore, the impact of workpiece thickness on physical characteristics has been assessed using analysis of variance (ANOVA) techniques to construct regression models with varying thickness variables. Subsequently, an Artificial Neural Network (ANN) has been employed to predict the aforementioned physical characteristics or quality control metrics. Finally, the output of the network has been compared and analyzed against the output of the regression models and actual data.

Keywords: Weld bead geometry, RSM, ANN.