Development of displacement monitoring and prediction system for short- and mid-span bridges

Abstract

Various technologies for structural health monitoring (SHM) have been developed and applied to civil infrastructure including bridges. In Korea, monitoring systems are mainly installed on long-span bridges and are operated for health monitoring purposes. However, the construction and maintenance of these SHM systems require significant economic costs, and for this reason, they are applied to few short- and mid-span bridges. In this study, a low-cost prototype system of displacement monitoring and prediction for such bridges is presented. The system consists of three components: 1) vision-based displacement measurement; 2) probabilistic displacement prediction; and 3) a desktop application based on MATLAB App Designer. First, in the system, the vertical and horizontal displacements of a bridge are measured using a dual-camera system and a digital image correlation technique, and the data are transmitted to a remote file transfer protocol (FTP) server using wireless internet. Based on the collected data, in the FTP server, the future displacements are probabilistically predicted using the Gaussian process regression, which provides 99% confidence intervals as well as the predictive mean of the displacements. In addition, a desktop application is developed using MATLAB App Designer, enabling users to check the measurement and prediction results remotely. The proposed system was applied to an actual bridge in the UNIST campus, and it has been operating successfully for several months, showing reasonable measurement and prediction results. Although the system still needs to be further developed and tested in several aspects, it is expected to become the basis for a technology that can easily monitor short- and mid-span bridges at relatively low cost.