A Study on Experimental Design Methods for the Automatic GMA Welding Process

Abstract

Generally, the welding parameters directly affect the weld forming and the joint performance. Because many parameters are involved in the automatic arc welding process, it is not realistic to use traditional experimental methods, such as full factorial design. Therefore, it is important to find out the good experimental design method to determine the welding parameters for optimal joint quality with a minimal number of experiments. Therefore, this study is aimed at investigating the effect of DOE (Design of Experiment) methods on bead width of mild steel parts welded by the automatic GMA (Gas Metal Arc) welding process. In this work, Taguchi method was used for studying the effect of the welding parameters on optimization of bead width, while Box-Behnken method was utilized to develop a mathematical model relating the bead width to welding parameters such as welding voltage, arc current, welding speed and CTWD (Contact Tip to Work Distance). The S/N (Signal-to-Noise) ratio and the ANOVA (Analysis of Variance) were employed to find the optimal bead width. Confirmation tests were carried out to validate the effectiveness of the Taguchi method. The experimental results show that welding current mainly affected the bead width. The predicted bead width of 3.12mm was in good agreement with the confirmation tests. With the regression coefficient analysis in the Box-Behnken design, a relationship between bead width and four significant welding parameters was obtained. A second-order model has also been established between the welding parameters and the bead width as welding quality. The developed model is adequate to navigate the design space.