Post-hazard system reliability analysis of complex interacting networks based on the Logical Expansion of Recursive Decomposition Algorithm

This paper proposes to solve sophisticated reliability problems of complex systems using the Logical Expansion of the Recursive Decomposition Algorithm (LE-RDA), a non-simulation-based network reliability analysis method. These problems include the reliability analysis of k-out-of-n system event definitions and importance assessment of the interaction effect between interdependent networks in terms of overall system reliability. These problems are defined by a complex mixture of events with correlation caused by shared components. This study also proposes a new reliability-based approach to evaluate the contribution of network component groups which are defined by their functions, to system events e.g. disconnection of water to an area. These new approaches are applied to a bench mark electric power grid and water network in Memphis/Shelby County to demonstrate the capabilities and effectiveness of the new approaches. The significance of interdependency effect is addressed and the network component groups' priorities are identified according to given earthquake magnitudes and interdependency level for decision support in hazard mitigation planning and disaster prevention management.