On the use of quantitative judgment in risk assessment: a Bayesian approach

Thesis/Dissertation · Thu Jan 01 04:00:00 EST 1981

OSTI ID:6670363

Mosleh, A

The development of a methodology for the use of expert quantitative judgement with particular emphasis on the problems of interest in risk assessments, is the subject of this work. A general methodology in the context of Bayesian theory of probability is presented and several models for using the opinions of one or more experts are proposed. These models account for both the experts' opinions and the information about the experts themselves. It is shown that existing aggregation techniques are special cases of the proposed methodology. A method is presented for the estimation of distributions for cases where the opinion of experts is expressed in the form of point or interval estimates or some percentiles of a distribution. The methods are then applied to estimate fragility curves for a nuclear reactor component from several expert's estimates for the percentiles. Finally, the methodology is used in the analysis of component failure data. Futhermore, the performance of different aggregation techniques in several examples with real data is studied and some conclusions are drawn.

• Research Organization: California Univ., Los Angeles (USA)
• OSTI ID: 6670363
• Country of Publication: United States
• Language: English

Related Subjects

22 GENERAL STUDIES OF NUCLEAR REACTORS
220200* -- Nuclear Reactor Technology-- Components & Accessories
EVALUATION
FAILURES
PROBABILITY
REACTOR COMPONENTS
RISK ASSESSMENT
USES