Ambient vibration re-testing and operational modal analysis of the Humber Bridge

• Published in: Engineering structures 2010-08, Vol.32 (8), p.2003-2018
• Main Authors: Brownjohn, J.M.W., Magalhaes, Filipe, Caetano, Elsa, Cunha, Alvaro
• Format: Article
• Language: English
• Subjects: Applied sciences; Bridges; Buildings. Public works; Exact sciences and technology; Operational modal analysis; Stresses. Safety; Structural analysis. Stresses; Structural identification; Suspension bridge; Suspension bridges. Stayed girder bridges. Bascule bridges. Swing bridges
• ISSN: 0141-0296; 1873-7323
• DOI: 10.1016/j.engstruct.2010.02.034

An ambient vibration survey of the Humber Bridge was carried out in July 2008 by a combined team from the UK, Portugal and Hong Kong. The exercise had several purposes that included the evaluation of the current technology for instrumentation and system identification and the generation of an experimental dataset of modal properties to be used for validation and updating of finite element models for scenario simulation and structural health monitoring. The exercise was conducted as part of a project aimed at developing online diagnosis capabilities for three landmark European suspension bridges. Ten stand-alone tri-axial acceleration recorders were deployed at locations along all three spans and in all four pylons during five days of consecutive one-hour recordings. Time series segments from the recorders were merged, and several operational modal analysis techniques were used to analyse these data and assemble modal models representing the global behaviour of the bridge in all three dimensions for all components of the structure. The paper describes the equipment and procedures used for the exercise, compares the operational modal analysis (OMA) technology used for system identification and presents modal parameters for key vibration modes of the complete structure. The results obtained using three techniques, natural excitation technique/eigensystem realisation algorithm, stochastic subspace identification and poly-Least Squares Frequency Domain method, are compared among themselves and with those obtained from a 1985 test of the bridge, showing few significant modal parameter changes over 23 years in cases where direct comparison is possible. The measurement system and the much more sophisticated OMA technology used in the present test show clear advantages necessary due to the compressed timescales compared to the earlier exercise. Even so, the parameter estimates exhibit significant variability between different methods and variations of the same method, while also varying in time and having inherent variability.