A structural health monitoring framework for civil structures

A health monitoring framework that detects and localizes damage in civil structures is proposed in this dissertation. Dynamic measurements of the systems as they vibrate under the influence of ambient and service loads are used to characterize the structural condition at any given time. Localization of damage is achieved by comparing characterizations in the pre and post damage states from a statistical point of view. Damage is localized regardless of the environmental conditions that affect the structures' dynamic behavior. Full automation of the data analysis is pursued. The methodology is tested with simulated responses of a finite element model that represents a typical highway bridge. Several uncertainties and modeling errors typically found in real applications were considered in this numerical example. The main advantages and limitations of the proposed framework are explained. The research emphasizes realistic applicability to civil structures in which only a few low vibrational modes can be measured with relatively high accuracy, and noise in sensors is unavoidable. However, the proposed framework as a whole (or in part) is likely to perform well in other engineering fields.