Estimation techniques for common cause failure events
Common cause failure probability estimation techniques, including ..beta..-factor basic parameter, binomial failure rate, multiple Greek, and C-factor estimators, are evaluated and compared using simulation data that captures the real world problem of sparse data from different plants. The effects on the estimators' performances from underlying factors such as common cause shock rates, lethal shock rates, probability of failing given a shock, independent failure rates, and system operational time are discussed. Worst case results are reported, and it is seen that for extremely small common cause failure probabilities the binomial failure rate estimators are best. However, these estimators can underestimate the true probabilities when the failured deviate from the binomial failure rate mode. The ..beta..-factor technique is shown to be conservative, and in some cases to overestimate the true probability by several orders of magnitude. When there are observed failutes for each failure event, the basic parameter technique is best and is easily calculated. This estimator is investigated in detail and is used to develop an estimator for the probability of K or more units failing due to a common cause. Uncertainty limits for this probability are also developed.
- Research Organization:
- Los Alamos National Lab., NM (USA)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 6984678
- Report Number(s):
- NUREG/CR-5044; LA-11179-MS; ON: TI88011214
- Country of Publication:
- United States
- Language:
- English
Similar Records
Computational problems with the binomial failure rate model and incomplete common cause failure reliability data
Computational problems with the binomial failure rate model and incomplete common cause failure reliability data
Related Subjects
22 GENERAL STUDIES OF NUCLEAR REACTORS
FAILURES
CALCULATION METHODS
PROBABILISTIC ESTIMATION
FAILURE MODE ANALYSIS
REACTOR COMPONENTS
SIMULATION
SYSTEM FAILURE ANALYSIS
SYSTEMS ANALYSIS
990230* - Mathematics & Mathematical Models- (1987-1989)
220200 - Nuclear Reactor Technology- Components & Accessories