Reliability analysis of dynamic reliability block diagram based on dynamic uncertain causality graph

Abstract

A dynamic reliability block diagram (DRBD) is often used in various industrial fields because of its ability to represent the dynamic behaviours of systems, despite its difficulty dealing with common cause failures and its quantitative result being hard to obtain due to the lack of a simple and practical solution method. Although some techniques, such as Petri net, Markov chain, and dynamic Bayesian network (DBN), have been proposed for analysing DRBDs, they are usually complicated and impractical for various reasons, such as state explosion in a Markov chain and a DBN and a complicated conversion process for a Petri net. To alleviate these problems, a common cause failure block is added to the DRBD, and a new solution method based on a dynamic uncertain causality graph (DUCG) is presented in this paper. Based on the conversion rule, the DRBD is mapped into a DUCG that needs fewer parameters and can be easily analysed according to the DUCG's reasoning algorithm. Finally, considering a subsea blowout preventer (BOP) control system as an example, a more accurate and concise model of the BOP control system is built.