Prioritizing failure events in fault tree analysis using interval-valued probability estimates

Antti Toppila*, Ahti Salo

*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

Probabilistic Safety Assessment (PSA) of nuclear power plants and other potentially hazardous facilities is often based on fault tree analysis. In conventional fault tree analysis, the probabilities of failure events are usually modeled with crisp estimates. However, because these estimates may be uncertain, we describe probability estimates with confidence intervals within which the 'true' values of the components' probabilities are assumed to lie. With interval probabilities, the risk importance measures that are commonly used for prioritizing components become set-valued so that they may lack a natural ordering. Therefore, in order to prioritize failure events, we define a dominance relation between events so that one event dominates another if its risk importance measure is at least as high as that of another event for all probabilities within these confidence intervals and, moreover, strictly higher for some. For computing the dominance relations, we use the rare event approximation and approximate three widely used risk importance measures, namely Fussell-Vesely, Birnbaum and Critical Importance using the minimal cut sets of the fault tree. We present a Branch and Bound-algorithm based on multilinear programming for checking the dominance relations when the probabilities are constrained to lie within their respective confidence intervals. The approach is illustrated by providing numerical results for a fault tree of a safety system at a fictive nuclear power plant. We also show how the results can be used to target improvements or repair/maintenance activities of the components.

Original languageEnglish
Title of host publication11th International Probabilistic Safety Assessment and Management Conference and the Annual European Safety and Reliability Conference 2012, PSAM11 ESREL 2012
Pages2538-2547
Number of pages10
Volume3
Publication statusPublished - 2012
MoE publication typeA4 Article in a conference publication
EventProbabilistic Safety Assessment & Management Conference and European Safety and Reliability Conference
- Helsinki, Finland
Duration: 25 Jun 201229 Jun 2012

Conference

ConferenceProbabilistic Safety Assessment & Management Conference and European Safety and Reliability Conference
Abbreviated titlePSAM ESREL
CountryFinland
CityHelsinki
Period25/06/201229/06/2012

Keywords

  • Fault tree analysis
  • Imprecise probability
  • Probabilistic risk analysis

Fingerprint Dive into the research topics of 'Prioritizing failure events in fault tree analysis using interval-valued probability estimates'. Together they form a unique fingerprint.

  • Cite this

    Toppila, A., & Salo, A. (2012). Prioritizing failure events in fault tree analysis using interval-valued probability estimates. In 11th International Probabilistic Safety Assessment and Management Conference and the Annual European Safety and Reliability Conference 2012, PSAM11 ESREL 2012 (Vol. 3, pp. 2538-2547)