The Role of the Poisson Distribution in Approximating System Reliability of k-Out-of-n Structures.

摘要

The determination of the exact reliability of a complex system of components (or subsystems) is typically a formidable theoretical and computational problem. Often one resorts merely to bounds on system reliability. The present paper delineates, for k-out-of-n structures, a simple and easily implemented approach based on the Poisson distribution. The components are allowed to be statistically dependent and to have unequal reliabilities. Special attention is given to the cases of independent and Markov dependent components, and some numerical illustration is provided. Comparisons are made to other reliability bounds in the literature. Reliability bounds for k-out-of-n structures have wide application, including series systems, parallel systems, monitoring systems, and systems with spares. The results are also applicable in the context of availability. (Author)