Network reliability analysis of complex systems using a non-simulation-based method

Youngsuk Kim, Won-Hee Kang

    Research output: Contribution to journalArticlepeer-review

    72 Citations (Scopus)

    Abstract

    Civil infrastructures such as transportation, water supply, sewers, telecommunications, and electrical and gas networks often establish highly complex networks, due to their multiple source and distribution nodes, complex topology, and functional interdependence between network components. To understand the reliability of such complex network system under catastrophic events such as earthquakes and to provide proper emergency management actions under such situation, efficient and accurate reliability analysis methods are necessary. In this paper, a non-simulation-based network reliability analysis method is developed based on the Recursive Decomposition Algorithm (RDA) for risk assessment of generic networks whose operation is defined by the connections of multiple initial and terminal node pairs. The proposed method has two separate decomposition processes for two logical functions, intersection and union, and combinations of these processes are used for the decomposition of any general system event with multiple node pairs. The proposed method is illustrated through numerical network examples with a variety of system definitions, and is applied to a benchmark gas transmission pipe network in Memphis TN to estimate the seismic performance and functional degradation of the network under a set of earthquake scenarios.
    Original languageEnglish
    Pages (from-to)80-88
    Number of pages9
    JournalReliability Engineering and System Safety
    Volume110
    DOIs
    Publication statusPublished - 2013

    Keywords

    • algorithm
    • binary
    • civil infrastructure
    • earthquake hazards
    • lifeline network
    • non-simulation-based method
    • recursive decomposition
    • system reliability

    Fingerprint

    Dive into the research topics of 'Network reliability analysis of complex systems using a non-simulation-based method'. Together they form a unique fingerprint.

    Cite this