Continuous dynamic monitoring of a lively footbridge for serviceability assessment and damage detection

• Published in: Mechanical systems and signal processing 2012-11, Vol.33, p.38-55
• Main Authors: Hu, Wei-Hua, Moutinho, Carlos, Caetano, Elsa, Magalhães, Filipe, Cunha, Álvaro
• Format: Article
• Language: English
• Subjects: Applied sciences; Automated stochastic subspace identification; Bridges; Buildings. Public works; Damage detection; Environmental/operational effects; Exact sciences and technology; Footbridge; Fundamental areas of phenomenology (including applications); Measurement and testing methods; Measurements. Technique of testing; Physics; Principal Components Analysis; Serviceability; Solid mechanics; Structural and continuum mechanics; Vibration, mechanical wave, dynamic stability (aeroelasticity, vibration control...)
• ISSN: 0888-3270; 1096-1216
• DOI: 10.1016/j.ymssp.2012.05.012

This paper aims at analyzing the feasibility of applying a vibration based damage detection approach, based on Principal Components Analysis (PCA), to eliminate environmental effects using the large amount of high quality data continuously collected by the dynamic monitoring system of Pedro e Inês footbridge since 2007. Few works describe real data, regularly collected along several years by reliable continuous dynamic monitoring systems in bridge structures. One main contribution is to show a large difference between making academic research based on numerical simulations or limited experimental samples, and making validity tests of innovative procedures using large high quality databases collected in real structures. The monitoring system, installed with the only initial objective of checking the efficiency of vibration control devices used to mitigate lateral and vertical vibrations, was therefore further developed for research purposes by implementing LabVIEW based automated signal processing and output-only modal identification routines, that enabled the analysis of the correlation of modal estimates with the temperature and the vibration level, as well as the automatic tracking of modal parameters along several years. With the final purpose of detecting potential structural damage at an early stage, the Principal Components Analysis (PCA) was employed to effectively eliminate temperature effects, whereas Novelty Analysis on the residual errors of the PCA model was used to provide a statistical indication of damage. The efficiency of this vibration based damage detection approach was verified using 3 years of measurements at Pedro e Inês footbridge under operational conditions and simulating several realistic damage scenarios affecting the boundary conditions. It is demonstrated that such a dynamic monitoring system, apart from providing relevant instantaneous dynamic information, working as an alert system associated to the verification of vibration serviceability limits, can also serve as an effective tool for long term bridge health monitoring. ► A continuous dynamic monitoring system was implemented in a slender footbridge. ► It was used for serviceability assessment and health monitoring under operational conditions. ► During three years, acceleration levels generally satisfied comfort limits. ► Natural frequencies automatically identified were correlated with environmental factors. ► PCA and Novelty Detection were used to remove environmental ef