Stochastic analysis of human-machine systems

Abstract

This study presents a stochastic analysis of three-state device systems subjected to critical human errors and common-cause failures. The systems analysed incorporate elements of commonly used redundant configurations such as parallel, series, k-out-of-n and standby. For the systems that are subjected to constant failure and repair rates, the Markov State Space technique is used to perform the reliability and availability analysis. Generalized and special-case expressions for the system reliability, the mean time to failure, the steady state availability and the time-dependent availability for each of these systems are developed. For the parallel system with non-constant failure and repair rates, the Device of Stages approach is used to formulate a modified Markov model with the addition of dummy states. The impact of human error, common-cause failure, failed system repair policy and elements of redundant configurations on the values of the system reliability, the mean time to failure, the time-dependent availability and the steady state availability is demonstrated by means of plots. (Abstract shortened by UMI.)