Implementation of condition-dependent probabilistic risk assessment using surveillance data on passive components

Abstract

A great deal of surveillance data are collected for a nuclear power plant that reflect the changing condition of the plant as it ages. Although surveillance data are used to determine failure probabilities of active components for the plant's probabilistic risk assessment (PRA) and to indicate the need for maintenance activities, they are not used in a structured manner to characterize the evolving risk of the plant. The present study explores the feasibility of using a condition-dependent probabilistic risk assessment (PRA) framework that takes a first principles approach to modeling the progression of degradation mechanisms to characterize evolving risk, periodically adapting the model to account for surveillance results. A case study is described involving a potential containment bypass accident sequence due to the progression of flow-accelerated corrosion in secondary system piping and stress corrosion cracking of steam generator tubes. In this sequence, a steam line break accompanied by failure to close of a main steam isolation valve results in depressurization of the steam generator and induces the rupture of one or more faulted steam generator tubes. The case study indicates that a condition-dependent PRA framework might be capable of providing early identification of degradation mechanisms important to plant risk.

Condition-based probabilistic safety assessment of a spontaneous steam generator tube rupture accident scenario

“…Although various events of inappropriate tubes plugging have been recorded (see the NRC's LERSearch system, https://lersearch.inl.gov/LERSearchCriteria.aspx), in this work it is assumed that crack detection is perfect, neglecting any possibility of non-detection of cracks. The reason for this is that as the flaw depth approaches the 100% through-wall crack depth, the probability of non-detection drops to zero as shown in (Kupperman et al, 2009) and (R. Lewandowski et. al.…”

“…The tube cracking process can be divided into onset, formation and propagation of cracks inside the tube well. The crack onset (i.e., the generation of microcracks inside the tube bundle) is modelled relying on the actual data collected in the Zion Plant (see (R. Lewandowski et. al.…”

“…1995), the SG tube is plugged if the crack depth exceeds 40% of the nominal tube wall thickness t. Being t=1.27 mm, the plugging limit for the crack depth results to be equal to 0.51 mm. Assuming a maximum depth-to-length ratio equal to 1/3 (R. Lewandowski et. al.…”

“…For the SGTR accident scenario, the novelty of CB-PSA with respect to LPSA is introduced in terms of the capability of updating, at each refueling cycle, the SGTR frequency according to the operative conditions that the NPP is expected to deal with during the following cycle. Other authors have proposed approaches for PSA updating using condition monitoring data (R. Lewandowski et. al.…”

Dynamic business continuity assessment using condition monitoring data

“…It is noteworthy that, due to the lack of real data, the condition monitoring data employed in the case study is generated from a known physical model. For illustrative purposes, the evolution of the tube crack growth process is assumed to follow the Paris-Erdogan model, which has been applied to model SGTR in [52,55],…”