Investigation on Mechanical Behavior of Friction Stir Welded Nylon-6 Using Temperature Signatures

Abstract

The present investigation focused on parametric optimization for improving joint efficiency of friction stir welded nylon-6 sheets using particle swarm optimization algorithm using a response surface method-based regression model. Initially, parametric effects on weld quality characteristics such as weld bead profile, bead shape and its microstructure along with micro-hardness variation and stress–elongation behavior of the butt weld have been studied considering cylindrical, square and triangular pin contours. The thermal cycles along advancing and retreating side of tool rotation have also been acquired during welding which was further used in the response surface models for the improvement in weld quality prediction capability. The joint efficiency was found to be maximum (49.68%) at intermediate value of each parameter using square pin. The material scooping action and undercut defects along with non-uniform grain morphology with micro-cavities were primary reasons for weld failure at weld interface. The response surface-based regression model of joint strength of the weld was significant, whereas weld hardness and percent elongation were found to be insignificant. However, weld peak temperature with associated cooling rate-based regression models was found to be highly adequate as per substantial improvement for the prediction of each weld quality-based features. Finally, response surface-based regression model of joint strength was further used for parametric optimization using response surface method as well as evolutionary particle swarm optimization tools. The particle swarm optimization was found to be more precise with better optimization capability than response surface methodology.