Abstract
Correlated component failures, recognised to be a major impediment in the power of redundancy to improve system reliability, need to be factored into the reliability assessment of fault-tolerant systems. In this paper, we present a methodology to derive an approximate analytical expression for mean reliability of on-demand systems with correlated failures. Applying the methodology, we derive expressions for the mean reliability of common redundant structures, namely, k-out-of-n, series-parallel and bridge systems. We illustrate the potential of these expressions to systematically assess the influence of pairwise component correlations on system reliability. An examination of the error between the system reliability estimates obtained from the analytical expressions to those obtained from simulations indicates that our method accurately captures the trends in system reliability and can hence provide a computationally faster alternative to simulation for system reliability analysis.