Cascading failure analysis of multistate loading dependent systems with application in an overloading piping network

Abstract

Many production and safeguard systems consisting of multiple components are susceptible to the cascading failures, where one possibility is that the failure of a component leads to more workloads of other components. Such loading dependence can result in failure propagation, make the systems more vulnerable and maintenance decision-makings more difficult. In this study, we develop a model for analyzing the propagation process of failures in loading dependent systems considering overloading states and degradation of components. The multinomial distribution is applied to characterize the probabilities of total numbers of failed- and overloading components, and probability distributions of different stop scenarios of cascading process are derived. A practical case in piping network is investigated to illustrate the analysis procedure, and to compare the effectiveness of the proposed model with those of the existing methods. Numerical analyses are conducted for evaluating the factors influencing the probability distributions of total number of failed- and overloading components, as well as the occurrence frequencies of different stop scenarios. It is expected that design and maintenance of loading dependent systems can be optimized with the support of this new cascading analysis approach.