A Probabilistic Model for Stress Corrosion Cracking of Stainless Steel SNF Interim Storage Canisters

In the U.S., spent nuclear fuel (SNF) is stored in dry storage cask systems for long-term interim storage. The systems commonly used consist of welded stainless steel containers enclosed in ventilated cement or steel overpacks. These containers may be required to perform their waste isolation function for many decades, and failure by chloride-induced stress corrosion cracking (SCC) due to deliquescence of deposited salts is a major concern, particularly for near-marine storage sites.This paper presents a probabilistic performance assessment model to evaluate the probability of canister failure (through-wall penetration) by SCC. The model first assesses whether environmental conditions for SCC—the presence of an aqueous film—are present at canister weld locations (where tensile stresses are likely to occur) on the canister surface. Geometry-specific storage system thermal models and weather data sets representative of U.S. SNF storage sites are implemented to evaluate location-specific canister surface temperature and relative humidity (RH). As the canister cools and aqueous conditions become possible, the occurrence of corrosion is evaluated. Corrosion is modeled as a two-step process: first, pitting is initiated, and the extent and depth of pitting is a function of the chloride surface load and the environmental conditions (temperature and RH). Second, as corrosion penetration increases, the pit eventually transitions to a SCC crack, with crack initiation becomingmore likely with increasing pit depth.Once pits convert to cracks, a crack growth model is implemented. The SCC growth model includes rate dependencies on both temperature and crack tip stress intensity factor, and crack growth only occurs in time steps when aqueous conditions are predicted. The model suggests that SCC is likely to occur over potential SNF interim storage intervals; however, this result is based on many modeling assumptions. Sensitivity analyses provide information on the model assumptions and parameter values that have the greatest impact on predicted storage canister performance, and provide guidance for further research to reduce uncertainties.