Abstract
•A reliability approximation method for heterogeneous cold standby systems•The method is fast, accurate and based on Lyapunov central limit theorem•Exponential, Weibull, normal, and mixed component time-to-failure distributions•Accuracy and efficiency are verified and compared with convolution methods
Cold standby sparing is a widely-used fault-tolerant technique where spare units are unpowered and non-operational before being activated to replace a malfunctioned on-line unit. The dynamic failure rate behavior of cold standby units poses unique challenges to the reliability analysis of cold standby systems (CSSs). The existing reliability analysis methods developed for CSSs have various limitations, such as being applicable to only the exponential component time-to-failure distribution, CSSs with identical components, and small-scale systems due to high computational complexity. In this paper, we advance the state of the art by proposing a fast and accurate reliability approximation method based on the Lyapunov central limit theorem for heterogeneous 1-out-of-n cold standby systems with non-identical components. The Lyapunov's conditions are proved for CSSs with exponential, Weibull, normal, and mixed component time-to-failure distributions. The accuracy and efficiency of the proposed method are verified and compared with the existing methods through comprehensive case studies on CSSs with different sizes and different types of distributions. The results show that the proposed method can estimate the reliability of large-scale CSSs efficiently and accurately.