Document Type



Doctor of Philosophy


Civil Engineering

First Adviser

Pakzad, Shamim

Other advisers/committee members

Wilson, John; Ricles, James; Frangopol, Dan; Cheng, Liang


Different elements of structural health monitoring (SHM) can benefit from the application of wireless sensor Networks (WSNs), as advanced sensing systems. While WSNs can significantly enhance the SHM by facilitating deployment of scalable and dense monitoring systems, challenges in the power consumption and data communication, and concerns regarding the possible impacts of their associated quality on the results have restricted their broad application. This research contributes in addressing the limitation associated with the prohibitive data communication delay and power consumption by introducing a novel time- and energy-efficient distributed algorithm for modal identification, and also addressing the concerns regarding the possible effects of their sensing quality by development of quality assessment approaches for modal identification and damage detection practices. The onboard processing techniques attempt to reduce the communication and power consumption by pushing the computation into the network. Efforts in developing onboard processing algorithms are restricted by the topology and algorithms, and their efficiency is not high enough to alleviate the challenge. A novel approach for modal identification of structural systems in a distributed scheme is developed which assigns the entire computational task of modal identification to remote nodes and limits the communication to transmission of only system's parameters. The algorithm is based on estimation-updating steps at remote nodes and iterations by passing the results through the network for convergence of estimation. The algorithm is first developed for input-output scenarios and then is further expanded to address output-only systems as well. Development of approaches such as Cumulative System Formation for providing initial estimates of the system (as starting point of iteration) and also a novel AR-ARX approach for expediting the convergence also further enhanced the developed algorithm. Experiments and implementations have proved the functionality and performance of the algorithm. While the focus of the research is on development of algorithms for enhancing the application of wireless sensors in modal identification, other aspects of data-driven SHM such as damage detection, and performance evaluation through field-testing of real-life structures are also studied. A framework for damage detection algorithm including accuracy indicators and statistical approaches for change point detection is developed and validated through implementation on different experimental models. Moreover, the state of the art in structural monitoring and vibration evaluation is presented in two field deployments.