Generation of regression models and multi-response optimization of friction stir welding technique parameters during the fabrication of AZ80A Mg alloy joints

Abstract

Conventional methodologies employed for the selection of weld process parameters for fabricating sound quality weldments have been found to consume more time and are often unreliable. Therefore, in this paper, an analysis was made to develop a quadratic regression model and empirical relationships by employing response surface methodology between various input parameters of the friction stir welding (FSW) technique including rotational speed of tool, axial force, and traversing speed of tool and five responses of output including percentage of elongation, yield strength, tensile strength, grain size, and microhardness. The discrepancies between the anticipated values and the genuine experimental outcomes are within ±1%, which reveals that the established mathematical quadratic regression model was a good fit to the actual experimental results. The experimental analysis also determined the elite combination of input parameters of the FSW technique for the output parameters.