System identification of large-scale bridge model using digital image correlation from monochrome and color cameras

Infrastructure systems and key components such as bridges are usually considered for structural health monitoring especially with the increasing deterioration and rising need for repair or replacements. Moreover, in cases of extreme loading, e.g. earthquake, the dynamic response of bridges and their condition after the event need to be rapidly evaluated. This is to make decisions with respect to opening the bridge for traffic, for instance, if no immediate repair is needed. Using modern technologies and non-contact methods such as wireless sensing or vision- based tracking, e.g. target-tracking Digital Image Correlation (DIC), can provide to the means to monitor our infrastructure for cases of extreme events and advance the current state-of-practice in SHM. Vision-based monitoring systems still face several challenges such as the high cost of high-speed or monochrome cameras used for monitoring. Therefore, it is beneficial to assess the possibility of using commercial or lower-end cameras for SHM purpose. In this study, an application of DIC for monitoring the behavior and seismic response of a large-scale bridge model behavior was applied using commercial DSLR cameras and high-speed monochrome cameras. The bridge model was a two-span concrete bridge that was tested on multiple shake tables under white noise and incremental seismic excitations at the University of Nevada, Reno. The objectives of the tests were to: (1) compare the monitoring performance between commercial and high-end monochrome cameras; (2) determine the modal properties of the bridge using system identification from DIC measurements and then compare it to the results from conventional contact sensors, e.g. accelerometers. The results show that feasible hardware such as commercial DSLR cameras can provide reasonable means to monitor and capture modal properties of bridges.