Role of probabilistic analysis in integrity assessments of reactor pressure vessels exposed to pressurized thermal-shock conditions

In recent years, probabilistic analyses have been playing an increasing role in safety assessments for nuclear power plants (NPPs) in the USA. An important example is the treatment of integrity analyses for the primary reactor pressure vessel (RPV) when it is exposed to pressurized thermal-shock (PTS) events. The U.S. Nuclear Regulatory Commission’s (NRC) approach to ensuring PTS integrity over the RPV lifetime has been to establish limits that ensure that the likelihood of through-wall-crack (TWC) development is kept below a prescribed very low level. The key elements for the probabilistic fracture mechanics (PFM) analysis package include thermal-hydraulic analysis input, neutron exposure models, embrittlement models for RPV steels, representations of fracture properties data, flaw characterization models, PFM models, and computational computer codes.The state of knowledge that existed when the NRC initially formulated the PTS rule in the 1980s necessitated that the rule and its technical bases use conservative treatment of parameters and models. Advancements made in the ensuing 20 years have motivated the NRC to undertake a comprehensive update of these technical bases to support potential revisions to the rule. The current status of the key technical elements has now been established, and an updated PFM methodology is being drafted and used for preliminary analyses of three U.S. PWRs. The frequencies of vessel failure predicted by these analyses have been compared with those calculated under the early PTS methodology and are encouraging that a revision to the PTS rule may be forthcoming.Because the PFM calculations utilize a series of deterministic LEFM fracture analyses, it is important to know the degree to which the LEFM methods are applicable to the analysis of thick-wall vessels. U.S. efforts to establish this applicability have been extensive and included a significant number of fracture experiments using large-scale vessels to validate analysis methods and demonstrate margins within design codes. This paper includes a summary of results from several of those key experiments.