Savannah River Site generic data base development

Abstract

This report describes the results of a project to improve the generic component failure database for the Savannah River Site (SRS). Additionally, guidelines were developed further for more advanced applications of database values. A representative list of components and failure modes for SRS risk models was generated by reviewing existing safety analyses and component failure data bases and from suggestions from SRS safety analysts. Then sources of data or failure rate estimates were identified and reviewed for applicability. A major source of information was the Nuclear Computerized Library for Assessing Reactor Reliability, or NUCLARR. This source includes an extensive collection of failure data and failure rate estimates for commercial nuclear power plants. A recent Idaho National Engineering Laboratory report on failure data from the Idaho Chemical Processing Plant was also reviewed. From these and other recent sources, failure data and failure rate estimates were collected for the components and failure modes of interest. For each component failure mode, this information was aggregated to obtain a recommended generic failure rate distribution (mean and error factor based on a lognormal distribution). Results are presented in a table in this report. A major difference between generic database and previous efforts is that this effort estimates failure rates based on actual data (failure events) rather than on existing failure rate estimates. This effort was successful in that over 75% of the results are now based on actual data. Also included is a section on guidelines for more advanced applications of failure rate data. This report describes the results of a project to improve the generic component failure database for the Savannah River site (SRS). Additionally, guidelines were developed further for more advanced applications of database values.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 23 citations exist upstream.