Optimal sensor placement for structural parameter identification of bridges with modeling uncertainties

Abstract

The uncertainties in the numerical models used for the optimal sensor placement (OSP) studies of civil infrastructures, specifically bridges, considerably affect the results. These effects can be more prominent if the modeling uncertainties are of a kind that significantly alters the mode shapes of the structure, such as the boundary conditions of the model. Yet, these effects on the results of OSP analysis remain unexplored, and there are no available methodologies to address all types of model uncertainties in OSP for civil infrastructures. This research presents a new framework to determine the optimal sensor locations to identify the modal properties of bridges under severe modeling uncertainties and its application on a railway bridge. The framework includes finite element model generation and a sensitivity study to select the most influential parameters that change the dynamic response of the bridge. The selected parameters are used in Monte Carlo simulations, and the results enable quantifying the relative amount of information presented at the candidate sensor locations. This information is combined with the spatial position of the candidate sensor locations, and a hierarchical clustering algorithm is used to obtain the optimal sensor locations and the number of sensors. In addition, the OSP analysis is carried out using the Effective Independence method to contribute to the state-of-the-art literature that investigates the uncertainties in OSP for civil infrastructures and uses this method.