Confidence Limits for System Reliability,
Abstract
The statistical problem of constructing confidence limits for system reliability from pass-fail data on the components is discussed. Usually such data are obtained from fixed sample size tests on the components. Exact and approximate methods for dealing with series systems are reviewed. A new approximate method is proposed and its performance evaluated in a number of special cases. An alternative approach that avoids some of the problems inherent in the above methods is to use variable sample size testing of the components. A sequential experimental design that is applicable to any coherent structure is given. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1976
- Accession Number
- ADA047533
Entities
People
- P. F. Preston