A New Three-Dimensional Tensor Method for Seismic Damage Detection in Concrete Piers of Bridges Using Response Signals

In scientific research works, time domain and frequency domain methods have been widely used to detect damage and its location. By making attention to the capabilities of square time-frequency representations in this study a new algorithm based on the representations was suggested for the first time to identify system and detect seismic damage in bridge piers. To evaluate the proposed algorithm, the analytical model of PEER-W180 Bridge was used. According to the method, before an earthquake event, some sensors were installed on top of the piers and vibrated the bridge by exciting force and recorded the bridge response signals. After the earthquake event which caused damages in the bridge piers, structural response signals of the piers were then recorded. Therefore the signals were processed using Reduced Interference Distribution, which is a time-frequency representation, and dynamic characteristics were extracted. Then, by using the suggested threedimensional tensor method, the damage and its location were detected. The results show the ability of the proposed algorithm for system identification and seismic damage detection in bridge pier. The algorithm is an output-only method. Furthermore, there is no need to create and update an analytical model of an existing bridge. According to the results, the damaged pier was detected correctly using the proposed algorithm. In addition, maximum error in the probability of damage in the piers was calculated equal to % 6.1. Therefore, the results indicate the ability of the algorithm in detection the damage and its location.