Post-yield characterisation of welds based on instrumented hardness tester data

The strength of welded joints has a vital effect on the structural response of a welded structure under severe loading conditions. The thermal nature of the weld process causes considerable changes in the microstructure across the weld regions, which leads to different material properties in these regions, namely, the parent metal, the weld fusion zone, and weld heat -affected zone. The reliability of predicting the real behaviour of the welded structure using finite element analysis depends on the accurate determination of the material properties across these regions which are essential input data in the model. Whilst conventional tensile testing is incapable of providing these properties on such a small scale, the instrumented micro/nano-hardness test lends itself for such a task. In this non-destructive technique a small indenter, which is spherical in this study, is pressed onto the surface of the material and then unloaded while load and depth of indentation are continuously measured. Several approaches have evolved to analyse the measured load indentation data to extract various material mechanical properties conesponding to each indent location. The primary objective sought from such an analysis is the determination of the contact area, which is then used to obtain material properties. While the existing analytical approaches assume the contact edges sink-in, soft metals, however, tend to pile-up resulting in possible large errors in the delived results. The available connection formulae attempt to predict pile-up based on a prior knowledge of the strain hardening exponent of the test material. In addition, they assume a constant ratio of the pile-up lip height to indentation depth, which they defined as a function of the strain hardening exponent. Thus they do not predict pile-up based solely on the indentation data. This study shows, based on detailed finite element modelling, that pile-up starts with a negative value (sink-in) in the early stages of an indentation experiment and then develops and builds up until it stabilises at a certain level. An empilical relation is proposed to predict the extent of pile-up of an unknown material based on the residual to total indention depth ratio. This leads to a more accurate estimate of the contact area and thus the derived stress-strain curve. In addition, a characterisation model for an indenter of imperfect geometry is also proposed in this study which identifies the characteristic strain for an imperfect sphere. An algorithm is suggested which incorporates the proposed characterisation analysis in a complex iteration technique in the frame compliance calibration routine to process the raw acquired data. The proposed characterisation technique has been verified on experimental data of test materials, and it has then been applied to indentation data on butt-welded steel spanning the distinct weld regions.