A MONTE CARLO METHOD FOR DETERMINING LOWER CONFIDENCE LIMITS FOR SYSTEM RELIABILITY ON THE BASIS OF SAMPLE COMPONENT DATA,
Abstract
A Monte Carlo method is presented for the determination of the lower confidence limit for the estimated reliability of a complex network of components (e.g., a guided missile, adaption kit, or other weapon system) when the only data available for estimating system reliability are sample failure data for each component. This method can be easily adapted to the particular circuitry of a specific system and can be utilized even in the case where sample sizes vary sharply from component to component. A method of solution is indicated for the general case in which the reliability of each component in the system is dependent upon the functioning of other components in the system and where the sample data are in either binomial or variable form. A detailed solution is shown for the case of component independence where the sample data for each component are in binomial form, e.g., 1 failure out of 100, 2 failures out of 400, etc. A numerical solution for a simple circuit is also shown and the result compared to that obtained from an analytical solution. The agreement between the two is shown to be good. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1960
- Accession Number
- AD0627799
Entities
People
- Donald S. Orkand
Organizations
- Picatinny Arsenal