Modeling the damage and recovery of interdependent critical infrastructure systems from natural hazards

Understanding the impact of disasters to civil infrastructure network will guide strategic pre-disaster hazard mitigation and post-disaster recovery planning of a community. The facilities in civil infrastructure network depend on each other to exchange product, information or services. When disaster happens, these dependencies would aggravate the initial damage and lead to cascading failures. Thus, understanding the dependencies among infrastructure facilities is essential in modeling the damage and recovery of a community under disruptive events. This study builds upon the Dynamic Inoperability Input-output Model to assess the recovery of the civil infrastructure facilities by considering the facility-level dependencies. The recovery of the integrated network was assessed utilizing the characteristic parameters from graph theory. This methodology was demonstrated with a hypothetical infrastructure network, consisting of power, water and telecommunication systems under a hurricane. The sensitivity of the network recovery to the dependency measures was investigated, which showed that the recovery time is sensitive to the level of dependency between different systems. The proposed methodology was compared with a conventional model with system-level interdependencies and showed that considering facility-level dependencies would yield more refined result. The model was then applied to simulate the power system restoration of Galveston City, Texas under Hurricane Ike.