Optimization of Process Variables for Prediction of Penetration Depth of HSLA Steel Welds Using Response Surface Methodology

Abstract

The statistical model is created for predicting penetration depth in an alternating current-based additional axial magnetic field controlled shielded metal arc welding of ASTM A 516 Gr.70 steel. The design for the trials is developed using the Placket-Burman design and response surface methodology. The created model determines the optimum process variables for getting excellent penetration depth. The input variables (current, magnetic field density, and magnetic frequency) are chosen for a response like penetration depth. This model can predict the main effects and the interacting effects of three process variables. The findings reveal that a higher current value with a low magnetic field density leads to deeper penetration and vice versa. Furthermore, a greater penetration depth is achieved at lower magnetic field density and higher magnetic frequency. With a desirability of 98.8%, the optimum process variables are 110 A, 0 mT, and 60 Hz. The predicted response values produced from the regression equation based upon process variables are extremely similar to the observed output, demonstrating the usefulness of second-order regression equations. For improved joint efficiency, a high level of penetration is needed.

Influence of Welding Parameters on Strength of Metal Inert Gas Welded Mild Steel Joints

“…Different methods of experimental design have been used in various studies, such as orthogonal array based on Taguchi [3], [12], [20], [21], factorial design [22], central composite design [22]- [24], response surface methodology [23], [25], [26] and D-optimal design [27]. One of the drawbacks of full factorial design is that it requires a large number of trials, which can be costly and time consuming [3].…”