Vibration-based system identification and damage detection of civil engineering structures

[Truncated abstract] Civil engineering structures inevitably subject to the adverse effects of environment erosion, material aging, fatigue, long-term effect of loadings, and natural disasters etc., which result in an accumulation of damage in structures. The structural deterioration has become a worldwide concern as it is a hidden danger which may cause sudden break or collapse of the structure. Structural Health Monitoring (SHM) has been attracting enormous research efforts in structural engineering because it targets at monitoring structural conditions to prevent catastrophic failure, and to provide quantitative data for engineers and infrastructure owners to design reliable and economical asset management plans. Many SHM systems have been installed on large-scale structures worldwide, however, a great deal of research is still needed for developing more reliable and applicable real-time SHM systems. A critical issue needs to be well addressed is the development of a vibration-based system identification method and a damage detection technique that are applicable to the real-time SHM under operational conditions. The research carried out in this thesis focuses on developing the vibration-based system identification technique and damage detection method that are applicable to the SHM of structures under ambient excitations. It consists of: (1) applicability of the widely used output-only system identification methods including time domain methods, frequency domain methods and time-frequency domain methods is comparatively studied. Variability of modal parameter identification produced by using various methods is quantified through analysing the ambient vibration response of an example beam indicating that on average a 2% to 3% error is generated in natural frequency identification, which should be accounted in damage detection...