Dynamic substructural condition assessment

Abstract

Vibration measurements, such as dynamic acceleration response data from civil infrastructures, are usually used for structural condition assessment with system identification techniques. Substructural condition assessment approaches are receiving increasing attentions in recent years since they have the advantages of reducing the number of unknown system parameters to be identified and system degrees-of-freedom (DOFs) involved in the computation. Measurements at the interface DOFs are normally required and treated as input excitations to the target substructure in many existing substructural identification approaches. However, it may not be possible to measure all the responses at the interface DOFs. On the other hand, the interface forces may be identified as well as the system stiffness parameters in the substructural condition assessment. This dissertation proposes a dynamic substructural condition assessment approach without information of responses and forces at the interface DOFs. Dynamic response reconstruction techniques in both the frequency and wavelet domains are developed. The relationship between two sets of time-domain response vectors is formulated based on the frequency response function in the frequency domain or unit impulse response function in the wavelet domain. Only the finite element model of the intact target substructure and measured acceleration data from the target substructure in the damaged state are required in the identification. A dynamic response sensitivity-based method is used for the damage identification and the adaptive Tikhonov regularization technique is adopted to improve the identification results when large noise effect is included in the measurements. Local damage is identified as a change in the elemental stiffness factor. Numerical and experimental studies are conducted to validate the effectiveness and accuracy of the proposed substructural damage identification approach. The local damage in the target substructure can be identified efficiently with the measurement noise and initial model errors in the finite element model. Another development in this thesis is to detect the local damage using measured acceleration responses from the target structure subject to moving vehicular loads which serve as excitations to the structure. The dynamic response reconstruction in wavelet domain is developed for the scenario when a structure or a target substructure is subject to moving vehicular loads. The transmissibility matrix between two sets of time-domain response vectors is formulated using the unit impulse response functions when the moving loads are at different locations. Measured acceleration responses from the structure or the target substructure in the damaged state are used for the damage identification. A three-dimensional box-section girder subject to a two-axle three-dimensional moving vehicle is taken as an example to validate the proposed approach for damage identification. The simulated damage can be effectively identified with noise effect included in the measurements.