Abstract (EN):
The Laboratory of Vibrations and Structural Monitoring (ViBest) of FEUP has implemented, since 2007, two different continuous dynamic monitoring systems with remote control through Internet in two lively Portuguese footbridges: the Pedro e Ines footbridge, over Mondego river in Coimbra, and the stress-ribbon footbridge of FEUP Campus. This paper describes an investigation, developed on the basis of the monitoring data from those two bridges, with the purpose of demonstrating the feasibility of detecting damage using vibration data under normal operational conditions. This study shows that environmental and operational factors (e.g. temperature and pedestrian traffic) can induce linear or non-linear effects on the natural frequencies, which can mask subtle early damage. Taking into account linear relations between natural frequencies estimates of different modes, the Principal Component Analysis (PCA) is applied to remove those effects. Subsequently, Novelty Analysis on the residual errors of PCA builds a statistical damage indicator for long term structural health monitoring. The efficiency of the described damage detection methodology is evidenced by simulating some realistic damage scenarios based on previously tuned finite element models and observing the clear deviation of the damage indicator.
Language:
English
Type (Professor's evaluation):
Scientific
No. of pages:
8