Risk-Informed Safety Margin Characterization

Abstract

The concept of safety margins has served as a fundamental principle in the design and operation of commercial nuclear power plants (NPPs). Defined as the minimum distance between a system's "loading" and its "capacity", plant design and operation is predicated on ensuring an adequate safety margin for safety-significant parameters (e.g., fuel cladding temperature, containment pressure, etc.) is provided over the spectrum of anticipated plant operating, transient and accident conditions. To meet the anticipated challenges associated with extending the operational lifetimes of the current fleet of operating NPPs, the United States Department of Energy (USDOE), the Idaho National Laboratory (INL) and the Electric Power Research Institute (EPRI) have developed a collaboration to conduct coordinated research to identify and address the technological challenges and opportunities that likely would affect the safe and economic operation of the existing NPP fleet over the postulated long-term time horizons. In this paper we describe a framework for developing and implementing a Risk-Informed Safety Margin Characterization (RISMC) approach to evaluate and manage changes in plant safety margins over long time horizons.

Estimating a failure probability using a combination of variance-reduction techniques

“…Then feed the X s, j into the function w s , and its output (L ′ s ,C ′ s ) has joint distribution H ′ s . For example, this is exactly the setting of the RISMC studies in Sherry, Gabor, and Hess (2013) and Dube et al (2014), where w s represents a detailed nuclear-specific computer code modeling the progression of a hypothesized accident in a NPP; see Hess et al (2009) for a survey of widely used codes in the nuclear industry. In general, a code run involves numerically solving a system of differential equations, which can require enormous computational effort, making variance reduction crucial to reduce the number of runs needed to achieve an acceptable precision.…”

Quantile Estimation Via a Combination of Conditional Monte Carlo and Randomized Quasi-Monte Carlo

“…Nuclear engineers employ detailed computer codes (Hess et al 2009) to generate an observation of (L,C), which often entails numerically solving systems of differential equations. We express this through two deterministic functions b L : (0,…”

A Strategic Approach to Employ Risk-Informed Methods to Enable Margin Recovery of Nuclear Power Plants Operating Margins

“…Compared to LB-LOCA events, Plant PRAs have indicated that the Small Break LOCA event also can challenge the ability to maintain adequate core cooling. These events typically have a larger impact on CDF and LERF than the LB-LOCA event due to (1) a much higher initiating event frequency, and (2) fewer alternative mechanisms for providing core cooling than for the LB-LOCA case. Small Break LOCA transients are characterized by a gradual top-down draining of the reactor coolant system, with low flow rates in the core relative to those occurring at steady-state or for Large Break LOCA transients.…”

“…The basic concept associated with margins is shown schematically in Figure 1 [2]. In its most simplistic form, the margin is represented as a distance between the load that an SSC experiences (L) and the capacity (C) that the SSC is capable of withstanding.…”