Multi-response Mathematical Modeling for Prediction of Weld Bead Geometry of AA6061-T6 Using Response Surface Methodology

Abstract

In the present paper, multi-response mathematical model is established for prediction of weld bead geometry in cold metal transfer (CMT), MIG pulse synergic (MIG P), and MIG manual (MIG M) welding of AA6061-T6 using ER4043 (AlSi5%) as a filler material. Central composite face-centered design under response surface methodology is employed to develop the design matrix for conducting the experiments. The developed model is employed in finding the optimal process parameters for good weld bead aesthetics. Current (I) and welding speed (S) are opted as input process parameters for response output such as penetration, dilution, and heat input. This model is proficient to forecast the main effects and interactive effects of two factors of the opted welding process parameters. Results show that higher current values with low welding speeds result in deeper penetration, high amount of dilution with higher heat input, and vice versa. With lower heat input, CMT has high dilution and penetration with respect to MIG pulse synergic and standard MIG welding. Repeatability of CMT process is much higher as compared to the other two processes. The optimal process parameters are 92.518 A and 7.50 mm/s for CMT, 109.418 A and 10.873 mm/s for MIG P, and 110.847 A and 11.527 mm/s for MIG M with 61.11%, 68.80%, and 72.6% desirability, respectively. Predicted output values generated from regression model equation obtained from welding process parameters are very close and sometimes overlaid on actual output that obviously demonstrates the suitability of the second-order regression equations. A good amount of penetration and dilution with low heat input is required for better joint efficiency.