RAPID RISK ASSESSMENT AND DECISION SUPPORT FOR URBAN INFRASTRUCTURE NETWORKS BY MATRIX-BASED SYSTEM RELIABILITY METHOD

Junho Song,M. ASCE,Seung-Yong Ok,Liang Chang,Student M. ASCE

2008-01-01

Abstract:Risk assessment of urban infrastructure networks under natural and man-made hazards is often performed by repeated computational simulations based on random samples of hazard intensities and corresponding component status. This sampling-based approach may prevent near-real-time application of probabilistic methods to hazard mitigation of infrastructure networks and risk-informed decision support. This paper proposes a non-sampling-based, multi-scale network reliability analysis approach based on the recently proposed matrix-based system reliability method. This approach accounts for the spatial correlation of hazard intensities and makes use of the disjoint cut sets or link sets identified by a state-of-the-art network analysis algorithm. The paper demonstrates the proposed approach through a case study of a Memphis Light, Gas and Water natural gas network under seismic hazard. As an attempt to identify opportunities and challenges in near-real-time application of probabilistic methods to hazard mitigation of urban infrastructure networks, this paper proposes a non-sampling-based, multi-scale network reliability analysis methodology based on the matrix-based system reliability (MSR) method (Song and Kang, 2007). The proposed method is demonstrated through a case study of a gas network under earthquake hazard scenarios. The probabilities that service nodes are disconnected from the source nodes are computed by a multi-scale system reliability analysis approach that can account for spatial correlation of seismic demands. For efficient network connectivity analysis, disjoint cut sets or link sets are identified by a state-of-the-art network analysis algorithm. We examine the accuracy and efficiency of the proposed method by comparison with Monte Carlo simulations. The paper also introduces future research topics to further