Comparison of methods to determine variations in unstrained unit cell parameter across welds

Many alloys undergo complex changes in local chemistry in the vicinity of weldments due to the thermal excursion during welding. The resulting changes in solute concentrations can lead to significant local variations in the unstrained unit cell parameter which, if not accounted for, can lead to serious error when determining residual stress by diffraction methods. Age-hardening aluminium alloys are particularly susceptible to such effects. The present paper compares three methods (plane stress assumption, sin(2)psi method, and comb correction method) for evaluating the stress-free unit cell parameter variation for friction stir welds in AA7449-W51 plates of two different thicknesses. All three methods gave comparable results for thin (5 mm) sheet, but for the thicker (12.2 mm) plate the results calculated on the basis of the plane stress assumption diverged from the other two, largely because in this case the other methods indicate there to be a significant triaxiality of stress. In the example cases, hardness and unstrained unit cell parameter variations were found to be strongly correlated across the welds. The advantages and disadvantages of the three methods are compared.