Addressing uncertainty in predictive estimates of risk

C. A. Bonebrake, P. Ramuhalli, W. J. Ivans, G. A. Coles, E. H. Hirt

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Probabilistic risk assessment (PRA) provides a static representation of risk associated with operations and maintenance (O&M) of nuclear power plants. Generic aging models can be used to predict the associated risk over time based on the assumed aging characteristics of specific components. These methods do not take into account the current condition of the components, and are susceptible to error in probabilistic estimates. Enhanced risk monitors (ERMs) use component condition (based on condition monitoring data) and time-dependent failure probabilities from prognostic health management (PHM) systems to calculate the risk associated with continued operation using potentially degraded components. ERMs are likely to be of value with advanced reactors, given the relative lack of operational data and the potential need to inform design choices and O&M actions to optimize plant performance, economics, and safety. Using equipment condition assessment methods to gather real-time conditions of advanced reactor components provides more accurate data to input into risk assessments within the ERM framework. This time-dependent condition data can be used as inputs for aging models in order to forecast the probabilistic risk associated with O&M. Such data is subject to uncertainties from measurements and historical operational conditions, along with uncertainties in aging models for components, resulting in uncertainty in estimates of component condition and predicted risk. Periodic updates with real-time measurements may be used to mitigate some uncertainty, which will need to be quantified. This paper describes the basic methodologies for incorporating periodic equipment condition monitoring data into an aging model to provide a forecasted risk assessment for prototypical advanced reactor components. This ERM methodology will include methods for propagating uncertainty over time within a dynamic predictive risk assessment that accounts for multiple interconnected components and failure events. Results of applying the ERM for a simplified advanced reactor PRA model will be presented.

Original languageEnglish
Title of host publication9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015
PublisherAmerican Nuclear Society
Pages1230-1238
Number of pages9
ISBN (Electronic)9781510808096
StatePublished - 2015
Externally publishedYes
Event9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015 - Charlotte, United States
Duration: Feb 22 2015Feb 26 2015

Publication series

Name9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015
Volume2

Conference

Conference9th International Topical Meeting on Nuclear Plant Instrumentation, Control, and Human-Machine Interface Technologies, NPIC and HMIT 2015
Country/TerritoryUnited States
CityCharlotte
Period02/22/1502/26/15

Funding

FundersFunder number
U.S. Department of EnergyDE-AC05-76RL01830

    Keywords

    • Enhanced risk monitors
    • Equipment condition assessment
    • Operations and maintenance
    • Probabilistic risk assessment
    • Prognostic health management

    Fingerprint

    Dive into the research topics of 'Addressing uncertainty in predictive estimates of risk'. Together they form a unique fingerprint.

    Cite this