Lower confidence interval bounds for coherent systems with cyclic components

Abstract

Three lower confidence interval estimation procedures for system reliability of coherent systems with cyclic components are developed and their accuracy measured using Monte Carlo techniques. The procedures use either the Poisson approximation to the Binomial distribution, the lower Binomial confidence limit procedure, or a modified procedure using the Poisson approximation to the Binomial distribution to obtain an equation for the lower confidence limit. The accuracy of the interval estimators were evaluated using standard computer simulation methods for series, parallel, series-parallel, and Wheatstone Bridge systems. The method determined to be most accurate can be combined with similar procedures for components that have continuous failure times and applied to yield a lower confidence interval procedures for the reliability of coherent systems with cyclic and continuously operating components.