Residual stress effects on crack driving force in multipass welds

In a structural integrity assessment of structures containing weld defects, it is often necessary to consider the effects of residual stress on crack driving forces. Most commonly used assessment procedures suggest using a yield magnitude residual stress for welded joints which have not gone through any post-weld stress relief. The assumed residual stress is then treated as an externally applied stress. This procedure involves some unrealistic simplifications. For instance, the transverse residual stress in the long butt weld of an unrestrained plate is often significantly less than yield magnitude. Moreover, when a crack is initiated in a region with weld residual stresses, a redistribution of the residual stresses occurs. It is, therefore, expected that the amount of crack driving force exerted by the residual stresses could be different from that of externally applied stress of the same magnitude. To quantify the effects of residual stress on crack driving forces, we analyzed one of the most commonly encountered structure geometries: butt-welded plates with externally applied load transverse to the weld line. The analysis consists of two main steps: (1) thermal elasto-plastic analysis to determine the residual stresses, and (2) fracture mechanics analysis to determine the crack driving force from a combination of welding residual stresses and externally applied stresses. In the residual stress analysis, two butt-welded plates of different strengths were simulated using the real welding parameters. The welding consumables were the same for both plates. In one plate, the weld strength overmatched the base plate strength by 26%. In the other plate, the weld strength undermatched the base plate strength by 20%. Good agreement was obtained between the computed residual stresses and the experimentally measured residual stresses. In the follow-on analysis, the residual stress fields were imposed on models with built-in cracks. The residual stresses were allowed to reach equilibrium in the cracked body. External load was then applied to create a model-I loading condition. The crack driving force in terms of CTOD was computed from the crack tip deformation. For comparison, the same cracked body without the initial residual stress was also subjected to an external load. This elastic-plastic analysis extends the loading from small scale yielding to large plasticity. The comparison of CTOD vs. applied load relations showed that the residual stress contribution to the crack driving force is very small regardless the magnitude of the externally applied load. For the plate/crack configuration analyzed, assuming yield magnitude residual stress in fracture assessment would significantly overestimate the crack driving force. The overestimation is most severe in welds containing long (deep) cracks.