Full-Scale Application of Stochastic DLV Method for Highway Bridge Health Monitoring

Structural health monitoring (SHM) has drawn significant attention lately for public safety as well as efficient infrastructure asset management. Over three decades, numerous SHM algorithms have been developed for bridge health monitoring. Research on full-scale damage detection on in-service bridges under operational loads is challenging and limited. In this paper, a practical application of a flexibility-based damage detection algorithm, the stochastic damage locating vector (SDLV) method, is developed and implemented for full-scale health monitoring of an in-service highway bridge. The test bed is a curved box-girder bridge in Connecticut. Based on long-term acceleration and temperature data measured for 5 years, statistical modal analysis has been conducted. During the measurement period, a permanent tilt was reported possibly by expansion joint misalignment due to permanent tilts of the pier. Potential damage locations are determined using the SDLV method with a sparse array of sensors before and after the permanent tilt event.