The horizontal position (2G) of welding is a common in assembling the blocks or vertical pipes of large welded structures such as ships and plants. However, the studies of this welding position have not been focused by researchers because of its difficulties in understanding the behavior of weld pool. In this study, to investigate the shape of back bead as a weld quality and then to select the optimal condition of root-pass TIG welding of horizontal position, an experimental design and
the RSM(response surface method) has been employed. The RSM is a powerful method for searching an optimal condition of factors producing the maximum outputs by implementing experiments. It is based on statistics, especially on regression model and analysis such as F-test. And the regression model is used to build a desirability function that provides the objective function for the optimization.
In this study, three parameters are employed as the input variables, which include base current, peak current and welding speed. And the back bead width is selected as an output variable repesenting the weld quality, the target value of back bead width is 5.4 mm.
By implementing the experiments according to Box-Behnken experimental design, a 2nd regression model for the back bead width was made and then the validation of the model was confirmed by using the F-test. And the desirability function was built through the nominal-the-best formula for the appropriate back bead width. Finally an optimal condition of welding was selected by using the RSM, which results base current of 0.8678, peak current of 0.9204, and welding speed of 0.3776 in coded values. For verifying, a test welding with the optimal condition was executed and the result shows the back bead width of 5.38 mm which is well matched to the target value.