A FORM-based analysis of lifeline networks using a multivariate seismic intensity model

Governmental organizations, private firms, and others in seismically active regions are interested in how reliable lifeline networks will be in the event of an earthquake. Assessing risk in these networks is more complicated than assessing risk for individual sites using traditional Probabilistic Seis - mic Hazard Analysis (PSHA) because of interdependencies among different system links, as well as cor - relations among ground motion intensities across a region. The focus of this paper is three-fold: (1) to construct a multivariate probability distribution of seismic intensities at many locations in an example network, by analyzing simulated earthquake data, (2) to develop a framework based on the First Order Reliability Method (FORM) to calculate network flow capacity after a disaster, and (3) to illustrate the importance of more accurately describing joint distributions of ground motion intensities when com - puting lifeline network reliability. This proposed approach provides probability distributions of network flow capacity given an earthquake, and also quantifies the importance of each lifeline component, which would allow system administrators to identify the most critical links in their lifeline systems and prioritize network upgrades. The example results indicate that neglecting the correlation of the ground motions can result in gross overestimation or underestimation of the probability of achieving a given system flow.