A method for estimating common cause failure probability and model parameters : the inverse stress-strength interference (ISSI) technique
Name
EL_TR_1984_010.pdf
Size
10.45 MB
Format
Adobe PDF
Checksum (MD5)
64796f0a7e7dc3b98ee07fd775a53df1
Author(s)
Guey, Ching Ning
Alternative Title
The inverse stress-strength interference (ISSI) technique, A method for estimating common cause failure probability.
Common cause failure probability and model parameters, A method for estimating.
Date Issued
1984
Publisher
Cambridge, Mass : Massachusetts Institute of Technology, Energy Laboratory and Department of Nuclear Engineering, 1984
Series/Report no.
Energy Laboratory report (Massachusetts Institute of Technology. Energy Laboratory) no. MIT-EL 84-010.
Abstract
In this study, an alternative for the analysis of common cause failures (CCFs) is investigated. The method studied consists of using the Licensee Event Report (LER) data to get single component failure probability and using stress and strength parameters to evaluate multiple component failure probabilities. Since an inversion of stress-strength interference (SSI) theory is involved, the approach is called the inverse stress-strength interference (ISSI) technique.
The ISSI approach is applied to standby systems in commercial nuclear power plants. At a component level, major pumps and valves are studied. Comparisons with other CCF analysis methods indicate that the medians based on the ISSI method are slightly higher because of the inclusion of potential failure causes. Applications to multiple-train systems show that the ISSI method agrees well with the beta factor method. In all cases studied, it appears that uncertainty intervals associated with the ISSI are smaller than other methods.
This study suggests that the ISSI method is a promising. alternative to estimate CCF probabilities. The method will be particularly valuable when: (1) Component-specific and system specific values are needed. (2) Failure data are scarce. (3) Level of redundancy is high. (4) Uncertainty needs to be quantified.
Persistent DSpace Link
