Magnetically impelled arc butt (MIAB) welding is a pressure welding process used for joining of pipes and tubes with an external magnetic field affecting arc rotation along the tube circumference. Past studies based on experiments and finite element analysis (FEA) suggested that the weld quality can be significantly improved by proper adjustment of its input parameters such as current, voltage, flux density, etc. Experiments are useful for understanding the physics behind the process, but the quantification and optimization of the MIAB process needs a thorough numerical investigation. Therefore, the present work presents a combined experimental and numerical study for investigating the arc speed indicator of the T11 grade chromium alloy tubes of the MIAB process. Firstly, experiments on the MIAB process are conducted, and the arc speed is measured based on the four inputs (welding current, welding voltage, magnetic coil current, and voltage). This is followed by introduction of a numerical approach of multi-gene genetic programming (MGGP) to formulate a functional relationship between arc speed and the four design variables. The statistical error metrics based on the square of correlation coefficient, relative error (%), and root mean square were used to validate the model performance against the experimental data. The proposed model can be utilized to select the best combination of input variables to optimize the arc speed, which has a vital role in governing the weld quality. In addition, 2-D analysis conducted reveals that the welding current influence the arc speed of the clay the most followed by coil current and coil voltage.
