Predictive Process Modeling of Tool Wear in Friction Stir Welding of Metal Matrix Composites

Abstract

Metal Matrix Composites (MMCs) are strong, lightweight materials consisting of a metal matrix (usually an Aluminum alloy) reinforced with ceramic particles or fibers. Because of their high strength to weight ratio, temperature resistance, and hardness, these materials are excellent candidates for use in aerospace and defense applications. Melting of the matrix alloy during fusion welding is accompanied by reactions between the molten alloy and the reinforcement material, resulting in the formation of deleterious phases linked to degradation in joint strength. These conglomerates are absent in friction stir welded MMC joints since the process occurs below the melting point of the workpiece material. FSW of MMCs is complicated by rapid and severe wear of the tool pin, a consequence of contact between the tool and the comparatively harder reinforcement material. Harder tool materials which would be resistant to abrasion from these inclusions are often too brittle to withstand the stresses associated with FSW. The dissertation presents several separate but related studies on tool wear in FSW of MMCs. Multiple regression techniques are used to construct a predictive model for tool wear based on empirical data. A phenomenological model, based on the existing rotating plug model for material flow, is developed to explain the mechanics of the wear process in FSW. Effects of reinforcement particle size, inclusion percentage, and the hardness of the tool relative to the reinforcement on wear are assessed. The feasibility of sensing tool wear in-process by monitoring changes in spindle torque is also explored.