The Nature of the Aircraft Component Failure Process: A Working Note
Abstract
The physics of component failures is normally assumed to follow a Poisson process. However, many studies have shown that the component demands in the U.S. Air Force supply system have a variance-to-mean ratio (VMR) much higher than 1.0, the VMR of a Poisson process. This apparent contradiction is resolved by modeling component failures as a Poisson process whose demand rate is not fixed, but rather is itself a stochastic process, wandering over time as a result of various causes such as weather, flying intensity, reliability growth, and, presumably, other unknown factors. Component and program data for the F-16 and A-10 aircraft show that demand over short time periods is Poisson. This is even more apparent when demands per flying hours are used instead of demands per day. However, when demands are aggregated over longer periods of time or more flying hours, the VMR increases. A gamma-Poisson model for future demands as a function of past observations fits the measured data. We call this process 'planetary Poisson' to distinguish it from the general class of nonstationary Poisson models.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1988
- Accession Number
- ADA197979
Entities
People
- Craig C. Sherbrooke
- F. M. Slay
Organizations
- LMI