Risk informed validation framework for external flooding scenario

Abstract

Safety of nuclear plants against external flooding has gained significant attention following the accident at Fukushima Daiichi nuclear power station. In United States, Oyster Creek nuclear plant was safely shutdown when high storm surge during hurricane Sandy caused a potential flooding threat. Subsequently, the nuclear energy industry experienced a significant activity in Probabilistic Risk Assessment (PRA) for external flooding. Increasingly, methods of computational fluid dynamics including advanced simulation codes are being considered to evaluate the sequence of events during different scenarios of flooding at a plant. One of the key limitations in the use of advanced codes for external flooding is related to a lack of credibility of such simulations. The motivation of this study is to develop a formal validation approach that provides a basis to quantify credibility of risk assessments that are based on advanced simulation codes. In this paper, we take a simple synthetic example to evaluate the applicability of the proposed framework to validation of flooding PRA scenario.

Feature Extraction for Uncertainty Quantification and Reduction of Noise in Condition Monitoring Using Deep Learning

“…The three levels of degradation can therefore be characterized as [20%, 30%] for minor, [45%, 55%] for moderate, and [70%, 80%] for severe degradation. Then, Latin hypercube simulation (LHS) is used to generate random severity values for each degradation classification and these values are employed in the finite element model of the simple-piping system to collect a sensor response training database for the ANN [23,24,25,26,27,28].…”

Predictive Capability Maturity Quantification Using Bayesian Network

“…In this study, the maturity level of data applicability is Linyu Lin; VVUQ-20-1017 characterized by a R/S/U grading system. The R/S/U is firstly developed by N. Dinh's works in 2013 [38] and has been used in [39] to evaluate the quality of validation data. The R/S/U system categorizes evidence of data applicability into three sub-attributes: [R]elevance, [S]caling, and [U]ncertainty, and each of them is designed according to their intrinsic properties.…”

Risk-Informed Decision-Making in Verification and Validation Under Uncertainty

“…As experienced in Fukushima-Daiichi nuclear disaster, it is necessary to improve our understanding of the accident sequence to assist decision-makers appropriately. In this context, studies have shown that significant uncertainties exist in the predictive capability of advanced simulation codes to evaluate the risk assessments [1,2,3,4,5]. The range of applicability of a simulation code is assessed by the physics and the models it contains.…”

“…Verification, Validation, and Uncertainty Quantification (VVUQ) are ongoing active research areas that help to assess the adequacy of simulation models in predicting the risk over the range of an intended application and the uncertainties associated with those predictions. One of the major challenges in validation is related to the lack of relevant experimental data and lack of confidence in the applicability of simulation models [1,6,7]. This leads to excessive reliance on expert judgment and conservative assumptions in the nuclear plant safety assessment.…”

“…The motivation of this study is to develop a robust risk-informed framework for the validation of advanced simulation codes with a probabilistic criterion for an adequate level of validation. Some of the recent frameworks that are based on a PRA approach for validation, particularly in the context of external hazards, are proposed by Kwag et al [1] and Bodda et al [2,11]. Kwag et al [1] and Bodda et al [2] illustrate the application of the risk-informed methodology for validation using a seismic and flooding case study, respectively.…”