Mathematical models for operational availability.

Abstract

The point availability of a system of components each of which is subject to random failures and has random restoration times is determined. Each component is assumed to have a fixed number of spares such that when all spares are exhausted no restoration can take place. Exact expressions are obtained for the availability of any series/parallel system composed of independently operating components. Furthermore, expressions are obtained for the availability of series systems under five operating scenarios for which components are not independent. Order relationships on system availability for the various operating scenarios are determined. An algorithm which optimally allocates spare parts over the components in a system is presented. The algorithm offers computational savings over solutions derived from dynamic programming.