Comprehensive functional resilience assessment methodology for bridge networks using data-driven fragility models

This paper proposes a probabilistic seismic resilience assessment methodology for bridge networks subjected to spatially correlated earthquakes. The proposed method integrates the effects of influencing factors (e.g., regional seismic hazards, bridge fragilities, and traffic flow) and their uncertainties in the seismic performance assessment of bridge networks. However, there are two major challenges in applying such a methodology that need to be addressed: rapid seismic damage assessment of regional bridges and comprehensive resilience quantification of bridge networks. Therefore, a data-driven fragility scheme based on an artificial neural network is developed to predict the damage states of regional bridges under earthquakes, which can support the simulation of the functionalities of bridges and roadways by Monte Carlo simulation. Furthermore, a multi-dimensional functional resilience vector is explored to quantify the ability of bridge networks to maintain the holistic system function, important subsystem function, and emergency response function. Finally, the proposed methodology is illustrated on the Sioux Falls bridge network, which shows the quantitative effects of regional bridge fragilities, traffic flow, and epicenter location on the functional resilience of bridge networks.