Multi-variable statistical models for predicting bead geometry in gas metal arc welding

Statistical models were developed to study the effect of gas metal arc welding process parameters (i.e., wire feed speed, voltage, travel speed, and shielding gas chemistry) on the resultant weld bead width, penetration, and reinforcement height, using a factorial design of experiment. Analysis of variance (ANOVA) indicated that the weld width depended on voltage, travel speed, gas type, and the interactions between these factors. The weld penetration depended only on wire feed speed and gas type, as well as the two-way interactions of wire feed speed with travel speed and gas type. Reinforcement height depended on travel speed, wire feed speed, and their two-way interactions with gas type. Residual analysis revealed that all assumptions inherent in the regression analysis were satisfied over the range of welding parameters considered in this study. The predictive power of the statistical models was validated using intermediate process parameter values in the experimental design, and it was found that predicted values were mainly in agreement with the measured values for a 95% prediction interval.