Modeling and TLBO algorithm based optimization of corrosion rate of friction stir welded dissimilar Mg–Al–Zn alloys

Magnesium alloys are used to manufacture lightweight vehicles to improve fuel economy and reduce carbon emissions. Implementing the idea of friction stir welding for joining increases the utility of these alloys in the automotive sector. In this study, dissimilar Mg alloys AZ31 and AZ91 were welded using the Box-Behnken experimental design criterion. The modeling of corrosion rate has been done using Response Surface Methodology (RSM), while optimization has been done using the Teaching-Learning-Based Optimization (TLBO) algorithm. The microstructural investigation suggest variation in grain size with process parameters and the affinity of magnesium towards oxygen form MgO in welded region which generates microcracks. Microstructure of corroded samples included some corroded circular rings, pits and MgO as white layers. The minimum corrosion rate 0.220 mm/year is obtained at 700 rpm of Rotational Speed (RotS), 30 mm/min of Welding Speed (WeldS) and 18 mm of Shoulder Diameter (ShD) suggesting improved homogeneity due to dynamic recrystallization. The most influencing factor for corrosion is tool rotational speed. The TLBO algorithm suggests a 5.90% improvement at 1000 rpm of RotS, 50 mm/min of WeldS and 19 mm of ShD.