ABSTRACT
A continuous dynamic monitoring system was installed at Pedro e Inês footbridge, located in Combra, Portugal, in 2007 by the CONSTRUCT/Vibest research group from the Faculty of Engineering of the University of Porto (FEUP). Such a system was further developed for research purpose by implementing an automated operational modal analysis procedure, in order to track the footbridge’s long term dynamic properties and detect early structural modifications. Modal parameters are identified by automated Covariance based Stochastic Subspace Identification (SSI-COV) technique. Their variations are tracked and the influence of temperature is analyzed. With the final purpose of detecting potential structural modifications at an early stage, the Multiple Linear Regression (MLR) method is employed to effectively eliminate temperature effects. Novelty analysis on the residual errors of the MLR model is established to provide a statistical structural health index. During five years from 2007 to 2012, a clear deviation of the proposed health indicator suggests a slight long term modification of the footbridge.
