A novel approach to assess the seismic performance of deteriorated bridge structures by employing UAV-based damage detection

Abstract

All civil infrastructure, including bridges, deteriorates over time. Unmanned aerial vehicle (UAV) based visual inspection of bridges has been proposed to assess the condition of bridges. However, existing methods cannot determine the seismic performance of bridges based on the results of UAV-based visual inspection. In this study, a novel approach is proposed to assess the seismic performance of deteriorated bridges with the results of UAV-based damage detection. The proposed approach consists of two phases: (i) the damage detection phase using a UAV and (ii) the seismic performance assessment phase. The images obtained from UAV survey are used to conduct condition assessment for the bridge, based on a previously developed region-based convolutional neural network (R-CNN), and the damage grade is assigned. Note that here damage includes both seismic damage and deterioration. Subsequently, the finite element (FE) model of the intact bridge is updated to correspond to the assigned damage index. To demonstrate the proposed approach, an in-service prestressed concrete box-girder bridge is investigated. In particular, the seismic response of the deteriorated bridge is assessed based on a comparison with the intact bridge responses; focus is placed on the maximum moment and maximum displacement at the pier and the girder. Predictions indicate that the seismic responses of the deteriorated in-service bridge are 10% poorer than those of the intact bridge. These results demonstrate the potential for the UAV-based approach for evaluating the seismic performance of deteriorated bridges.