Validation of Welding Residual Stress Model Using Results From a Pressurizer Surge Nozzle Mockup

Abstract

Although numerical welding simulation is now commonly used in the nuclear industry to predict residual stresses in reactor vessels and associated piping components, there are currently no universally accepted guidelines for performing such analysis. Moreover, due to the complexity of the calculations and varying analytical procedures among analysts, there remains a need to validate predictions of residual stress against benchmark studies. As part of an industry initiative to manage the degradation of dissimilar metal welds in pressurized water reactor piping that are susceptible to primary water stress corrosion cracking, the U.S Nuclear Regulatory Commission embarked on a multi-phased program to validate welding residual stress models. The aim of Phase II of this program is to obtain measured residual stresses from a pressurizer surge nozzle dissimilar metal weld mockup for use in comparisons with numerically predicted stresses. This paper presents results of finite element analysis for various stages during the fabrication of a 14–inch pressurizer surge nozzle mockup, including an Alloy 82 dissimilar metal weld between a stainless steel safe end and carbon steel nozzle, an inside surface weld repair (back weld) and fill-in weld (weld build-up), and a stainless steel “field” weld attaching a section of straight pipe to the safe end. The NRC validation program was structured to allow participants to first calculate results using their own material properties, and then tune their welding simulations to thermocouple data. This was followed by reanalysis using NRC-supplied material properties. The program was conducted as a round robin analysis among an international group of participants and formatted as a blind validation project wherein results were submitted to the NRC prior to receipt of thermocouple and material property data. Results were obtained for both kinematic and isotropic hardening rules to study the effect of these two extreme measures of material characterization on the development of residual stress. Predicted stresses are then compared to measured stress data obtained by the deep-hole drilling technique at multiple locations through the thickness of the weld. The NRC residual stress model validation project serves as a valuable contribution to the understanding of how residual stresses are developed in dissimilar metal welds. The correlation of calculated residual stresses with measured data from a relevant mockup also serves to increase confidence in predicting crack growth in these primary pressure boundary welds by removing much of the uncertainty previously associated with residual stress input to crack growth analysis.