Stochastic Sampling for Efficient Seismic Risk Assessment of Transportation Network

Abstract

For accurate seismic risk assessment of transportation network under probabilistic seismic hazard, the uncertainty in the seismic hazard, the damage states of links/bridges in the network, and network performance need to be quantified. Stochastic simulation is well suited for this task. However, it typically requires large number of model evaluations, which entails significant computational effort, especially for large network. To address the above challenges, an efficient stochastic sampling-based approach is proposed. It relies on generating one set of samples for earthquake magnitude and carrying out analysis for the corresponding set of networks. This set of evaluations are used for seismic risk assessment under different risk measures, different probabilistic seismic hazards (e.g., with or without considering spatial correlation), and also for risk-based importance ranking of all bridges/links in the network for risk mitigation purpose. No additional evaluation of the network model is needed. The proposed approach is applied to seismic risk assessment and mitigation of the transportation network of Los Angeles and Orange countries. The impact of spatial correlation in seismic hazard on the seismic risk assessment and mitigation is investigated.