Structural Health Monitoring of an In-Service Highway Bridge with Uncertainties

In recent years, structural health monitoring (SHM) has become increasingly important for the assessment and maintenance of aging infrastructure. However, some challenges remain in applying SHM in the field. For example, high cost of wired sensors limits the SHM application to small number of structures within a network. Wireless smart sensors (WSSs), on the other hand, provide a versatile, cost-effective method of monitoring dynamic conditions of structures in near real-time. While the wireless sensor technology has been available for nearly a decade, there have been limited numbers of full-scale applications due to its intrinsic limitations such as power supply issue. In this dissertation, based on the power consumption of the Imote2 wireless sensor platform, a vibration-based piezoelectric energy harvester has been developed and validated. The feasibility of structural monitoring by means of a WSS system has been validated using a lab-scale truss bridge and an in-service highway bridge in Meriden, Connecticut. The results from the WSS system have been compared with the results from traditional wired sensors as well as finite element model to show the efficiency of SHM using WSSs. One of the major objectives of SHM is to detect damage in the monitoring structures based on changes in their mass, damping, and stiffness. These parameters are also affected by operational and environmental variability, which challenge the reliability of damage detection. In this study, a full year monitoring results from a long-term SHM system on the Meriden Bridge is presented. Especially, the variations in the identified modal frequencies have been observed, and these changes have been shown to be strongly correlated with temperature measurements. The ultimate goal of this research is to detect structural damage in presence of uncertainties. In order to achieve this goal, damage detection procedure based on machine learning algorithms has been developed. Two machine learning algorithms have been employed: a linear algorithm based on the Mahalanobis Squared Distance (MSD), and a nonlinear algorithm based on the Auto-Associative Neural Network (AANN). The MSD and AANN algorithms have been tested on the lab-scale truss bridge as well as the one-year monitoring data measured from the Meriden Bridge