Groth, Katrina M., Denman, Matthew R., Cardoni, Jeffrey N., Wheeler, Timothy A. Proof of Principle Framework for Developing Dynamic Risk-Informed Severe Accident Management Guidelines (Technical Report) Sandia National Laboratories Albuquerque, NM, (SAND2013-8324), 2013.

BibTeX

@techreport{SAND2013-8324,
title = {Proof of Principle Framework for Developing Dynamic Risk-Informed Severe Accident Management Guidelines},
author = {Katrina M Groth and Matthew R Denman and Jeffrey N Cardoni and Timothy A Wheeler},
year = {2013},
date = {2013-09-01},
number = {SAND2013-8324},
address = {Albuquerque, NM},
institution = {Sandia National Laboratories},
abstract = {Severe accident management guidance (SAMGs) provides support for collecting information and assessing the state of a nuclear power plant during severe accidents. However, SAMGs developers cannot anticipate every possible accident scenario. Sandia National Laboratories conducted a proof-of-principle study to demonstrate the capability to use Discrete Dynamic Event Trees (DDETs) to probabilistically explore dynamic uncertainties in a plant simulation models (MELCOR) to develop comprehensive, risk-informed Severe Accident Management Guidelines (SAMGs). That work was documented in April 2012 (SAND2013-3352) as part of this body of work under the Department of Energy Advanced Small Modular Reactor research plan.. This report documents the application of the DDET capability developed in the first half of fiscal year 2013 to the development of a proof-of-principal risk management tool designed to enhanced operator diagnosis of accident conditions and to lead to risk-informed SAMGs that are informed by DDET methods developed under this work package.

DDET analyses can be combined into a decision support system based on a Bayesian Network (BN) framework. The BN model synthesizes the information from the extensive set of DDET and MELCOR runs into a single framework to be used, in real-time, for supporting diagnosis and response planning during off-normal conditions.

The proof-of-principle study developed a framework for integrating results from DDET/ MELCOR simulations and component reliability information into a BN. A model was created for diagnosing an example problem involving the status of Emergency Core Cooling System (ECCS) valves in a generic iPWR design. The concept extends to advanced small modular reactors as well.},
keywords = {Artificial intelligence, Bayesian Networks, Decision support systems, Dynamic PRA, nuclear power, Probabilistic risk assessment (PRA), risk-informed procedures, uncertainty},
pubstate = {published},
tppubtype = {techreport}
}

Abstract

Severe accident management guidance (SAMGs) provides support for collecting information and assessing the state of a nuclear power plant during severe accidents. However, SAMGs developers cannot anticipate every possible accident scenario. Sandia National Laboratories conducted a proof-of-principle study to demonstrate the capability to use Discrete Dynamic Event Trees (DDETs) to probabilistically explore dynamic uncertainties in a plant simulation models (MELCOR) to develop comprehensive, risk-informed Severe Accident Management Guidelines (SAMGs). That work was documented in April 2012 (SAND2013-3352) as part of this body of work under the Department of Energy Advanced Small Modular Reactor research plan.. This report documents the application of the DDET capability developed in the first half of fiscal year 2013 to the development of a proof-of-principal risk management tool designed to enhanced operator diagnosis of accident conditions and to lead to risk-informed SAMGs that are informed by DDET methods developed under this work package.

DDET analyses can be combined into a decision support system based on a Bayesian Network (BN) framework. The BN model synthesizes the information from the extensive set of DDET and MELCOR runs into a single framework to be used, in real-time, for supporting diagnosis and response planning during off-normal conditions.

The proof-of-principle study developed a framework for integrating results from DDET/ MELCOR simulations and component reliability information into a BN. A model was created for diagnosing an example problem involving the status of Emergency Core Cooling System (ECCS) valves in a generic iPWR design. The concept extends to advanced small modular reactors as well.