Predicting capacity model and seismic fragility estimation for RC bridge based on artificial neural network

Abstract

The uncertainty of a structural capacity plays an important role in regional fragility, risk, and resilient estimation. This paper proposes an artificial neural network (ANN)-based predicting capacity model to consider the uncertainty of the seismic performance of the reinforced concrete (RC) bridge columns. The seismic fragility of three typical RC bridges is studied. A database that includes 78 testing results is used to train, validate, and test the ANN model. The capacity measures of RC bridges are predicted using the constructed ANN model. Case studies are conducted for three typical simply supported T-beam bridges, and nonlinear time history analysis is performed to obtain the structural responses of the RC bridges. Seismic fragility models are established and a comparison is performed to study the discrepancies of damage probability for RC bridges with and without considering the uncertainty of the capacity model. The main results of the paper are as follows: (1) the ANN-based predicting capacity model can provide an acceptable capacity measure for seismic fragility estimation and (2) the damage probability of RC bridges is related to the uncertainty of seismic performance that needs to be considered for structural fragility estimation.