A new probabilistic evaluation model for weld residual stress

Abstract

Weld residual stress (WRS) is one of the most important factors in the structural integrity assessment of piping welds, and it is considered a driving force for crack growth. It is characterized by large uncertainty. For more rational assessment, it is important to consider the uncertainty of WRS for evaluating crack growth behavior in probabilistic fracture mechanics (PFM) analysis. In existing PFM analysis codes, WRS uncertainty is set by statistically processing the multiple WRS data obtained from finite element analyses. This process depends on the individual performing PFM analysis, which may lead to uncertainties derived from the PFM analyst. In this study, we developed a new WRS evaluation model based on Fourier transformation, and the model was incorporated into the PFM analysis code PASCAL-SP, which has been developed by Japan Atomic Energy Agency. Through improvements to the code, WRS uncertainty can be considered automatically and appropriately by inputting multiple WRS data directly as input for PFM analysis.