Mechanical properties distributed in friction stir weld determined by spherical indentation testing

摘要

For characterizing the mechanical properties of different zones in a Friction Stir Weld of a high strength aluminum alloy, local stress-strain curves were determined from an array of spherical indents on the cross-section of the joint. The mapping of force-depth curves to stress-strain curves was carried out using neural network based analysis software which is commercially available. A slotted indentation test specimen has been introduced which promotes the relaxation of residual stress before indentation testing. In this way, systematic error in the neural network application can be avoided. The investigation demonstrates that spherical indentation tests combined with neural network analysis is capable of quantifying the local material property variation across butt welded joints. The results are comparable to much more expensive and time consuming experiments on micro-tensile specimens.