Investigation of Residual Stress In Friction Stir Welded Aluminum Alloys

This study explored mechanical and metallurgical property development in friction stir welded 6xxx and 7xxx aluminum alloys through empirical observation and numerical modeling with an emphasis on residual stresses induced by processing. Thermal conditions and weld parameters are varied and resulting weld properties are correlated to deduce trends. In one study on thin plate 6.35 mm AA7050, thermal boundary conditions are manipulated within a weld process parameter envelope by welding underwater and compared to welding in ambient air. In a thick plate 32 mm AA7050 investigation, through-thickness property variations are observed due to non-uniform temperature conditions in the weld region. In a thin plate AA6019 study, post-weld heat treatment variation is coupled with a broad weld parameter envelope to characterize the material metallurgical and mechanical response. Simulation work was performed to aid in understanding how thermal steady state conditions and residual stresses develop during welding. This was applied for a basic set of weld conditions, then to a novel application of friction stir welding with a stationary shoulder tool. Results showed that convective thermal boundary conditions have a large effect on weld temperature and residual stress in AA7050 welds, both of which are reduced by welding underwater compared to in ambient air. It was also shown that there are strong property variations through-thickness in the thick plate welds. AA6019 weld properties, including hardness, tensile behavior, grain size, and residual stress were shown to be highly dependent on weld parameters and post weld heat treatment conditions. Simulation work showed that steady state conditions may be characterized well by a thermal contour area approach. Steady state conditions depended on welding speed and other factors. Residual stress modeling was able to reasonably resolve stresses observed in experimental data. The cut compliance technique was utilized throughout this work to obtain residual stress information. The repeatability demonstrated in results and modeling provide evidence of the technique\u27s reliability