A Bayesian Network model to assess seismic risk of reinforced concrete girder bridges

Fragility curves are required for the probabilistic evaluation of performance of an earthquakedamaged transportation network. Recently, fragility curves of several hundreds of bridges have been obtained via 3D inelastic response history analysis, and then used to calibrate a Bayesian network (BN) model of seismic fragility. This paper investigates the sensitivity of network-level results to the use of such a BN-based surrogate fragility model. Bridge damages were evaluated with both the exact FEM-based and the approximate BN-based fragilities, and network flow analysis was carried out on the damaged network for thousands of seismic events in several Monte Carlo simulations. Even though the BN-based model does not perform in an equally good manner over all bridges, overall more accurate results at the network level were observed, indicating that network performance does not in general depend on any single bad match between reference and approximate fragility curves.