Modelling and measurement of residual stresses in autogenously welded stainless steel plates: Part 1 – fabrication and modelling

Abstract

Fabrication details and finite element derived predictions of residual stress are presented for two types of 316 L stainless steel test specimens of simple geometry, both containing an autogenous (no filler material) weld bead. The specimens were conceived in order to validate analytical weld residual stress predictions, which are only available for components of simple geometry and which do not account for phase transformations. The numerical modelling presented was undertaken by four different participants as part of a small ‘round robin’. The simplicity of the specimens permitted conclusions to be drawn concerning the relative importance of various modelling parameters. It is demonstrated that the assumptions made regarding the material hardening behaviour have the most significant effect on the predicted residual stress fields. Maximum differences in residual stress magnitudes of approximately 250 MPa were seen, but in general the variations were around 100 MPa. This underlines the need for caution when considering numerical residual stress predictions and underlines the need to validate predictions with measurements.