Reliability analysis based on dynamic Bayesian networks: A case study of an unmanned surface vessel

Abstract

Unmanned surface vessels (USVs) have been widely used in various activities to date. Reliability is particularly significant for USVs because of the increasing complexity of USV tasks and their characteristic that cannot be repaired during task execution. This study aims to analyse the reliability of USVs' overall system for providing an optimisation reference for their reliability improvement in the design phase. The method for modelling and analysing the reliability of complex systems based on Dynamic Bayesian Networks (DBNs) is studied. The reliability of a USV for seafloor topography survey is analysed. The reliability model of the USV overall system is established by mapping the Dynamic Fault Trees (DFTs) into DBNs considering the system logical structure, redundant configuration and dynamic behaviour. Reliability analysis, including reliability evaluation, importance measure and sensitivity analysis, are conducted on the basis of DBN inference. The overall reliability parameters of the USV are obtained to determine the critical subsystem and components. This research may provide a reference for the reliability design and optimisation of USVs.