Probabilistic seismic assessment of reinforced concrete bridges using simulated records

Abstract

A prominent challenge in performance-based earthquake engineering is to select a suitable set of ground motions records with which to perform probabilistic seismic assessment of structures. The most common approach for engineering purposes is to employ actual recordings of worldwide events, given that large earthquakes do not occur frequently hence regional recordings of such events are usually not widely available. To address this not that uncommon issue, regionally simulated ground motions using a stochastic finite-fault method have been proposed as an alternative to real records. This study aims to explore the use of simulated records through a stochastic finite-fault method in probabilistic seismic assessment frameworks for reinforced concrete bridges, when compared to using real records. Direct seismic losses for a case-study existing bridge and a bridge portfolio are estimated and compared, as the reference risk metric. Finally, the similarities between seismic demands, obtained using both real and simulated record sets, are quantified and discussed via statistical hypothesis testing, resulting fragility curves and expected annual losses. The results show how simulated records can be a promising alternative to real records, becoming particularly useful in the absence of available recorded ground motions with specific seismogenic features.