|Title:||Damage detection and reliability assessment of building structures using stochastic approaches|
|Subject:||Hong Kong Polytechnic University -- Dissertations|
Structural health monitoring
|Department:||Department of Civil and Structural Engineering|
|Pages:||1 v. (various pagings) : ill. (some col.) ; 30 cm.|
|Abstract:||Civil structures begin to deteriorate once they are built due to harsh environment such as corrosion, earthquake, and typhoon. Vibration-based structural damage detection methods have thus attracted considerable attention for assessment of functionality and safety of civil structures. Nevertheless, the damage detection of civil structures still remains as a challenging task. One of the main obstacles is that the current damage detection methods are either insensitive to local structural damage or sensitive to measurement noise. In addition, a significant amount of uncertainties are inherently and inescapably associated with structural damage detection of civil structures, which limits the successful application of most deterministic damage detection methods. In this regard, this thesis focuses on the establishment of a framework in which novel stochastic approaches in consideration of the uncertainties involved in measurements, structures and external excitations are proposed to detect damage of building structures and assess their reliability effectively. This thesis first proposes a novel structural damage detection method using a new damage index based on the statistical moments of dynamic responses of a building structure under random excitation. The principle of the statistical moment-based damage detection (SMBDD) method is first put forward in the frequency domain through a single-degree-of-freedom (SDOF) system under white Gaussian noise excitation. The sensitivity of statistical moment to structural damage is discussed for various types of structural responses and different orders of statistical moment. The formulae for statistical moment-based damage detection are derived. The effect of measurement noise on damage detection is ascertained. The new damage index and the proposed SMBDD method are then extended to multi-degree-of-freedom (MDOF) systems with resort to the nonlinear least squares method. As numerical studies, the proposed method is applied to both single and multi-story shear buildings in consideration of measurement noise. Numerical results show that the fourth-order statistical moment of story drifts is a more sensitive indicator to structural stiffness reduction than the natural frequencies, the second order moment of story drift, and the fourth-order moments of velocity and acceleration responses of the shear building. The fourth-order statistical moment of story drifts can be used to satisfactorily identify both locations and severities of various damage scenarios of the shear building. Furthermore, a significant advantage of the SMBDD method lies in that it is not only sensitive to local structural damage but also insensitive to measurement noise. After that, the SMBDD method is advanced in the following three aspects in the frequency domain for its practical application: (1) the type and location of external excitations, (2) the type of building structures, and (3) the number of structural responses measured. The equations of the SMBDD method are accordingly extended to be more general for any type of building structures under any type of random excitation as long as it complies with the Gaussian distribution. The generalized SMBDD method is also extended from the necessity of complete measurements of all DOFs to the proper selection of measurements of incomplete DOFs. Extensive numerical examples are presented to demonstrate the feasibility and effectiveness of the generalized SMBDD method. A MDOF shear building structure under colored noise excitations at different locations is investigated. Various damage scenarios of a high-rise building and a frame structure using selected measurement responses are investigated. The effect of measurement noise on the quality of identified results is also investigated for all the damage scenarios concerned by numerically contaminating the external excitations and the measured responses with white Gaussian random noises. Numerical analysis results show that the damage locations and severities of all the concerned various damage scenarios can be identified satisfactorily even though the structural responses used are incomplete and the measurement noise has a high noise-to-signal ratio of 15%. Furthermore, the feasibility and effectiveness of the generalized SMBDD method are explored in the time domain for the building structures under non-Gaussian excitations. The algorithm of the generalized SMBDD method in the time domain is developed. Various damage scenarios of different damage locations and damage severities of shear buildings, high-rise buildings and frame structures are numerically investigated. Numerical results demonstrate that the generalized SMBDD method is feasible and effective for building structures under either Gaussian or non-Gaussian excitations in the time domain. Even with the measurement noise intensity as high as 15%, the structural damage locations in various damage scenarios with incomplete measurements can be identified satisfactorily no matter whether the external excitation is of Gaussian distribution or not. Furthermore, the identified damage severities are exactly equal to the real values when measurement noise is not considered. Otherwise, the quality of the identified damage severities in the time domain is similar with that in the frequency domain when measurement noise is considered. Nevertheless, the requirement of proper optimization methods in model updating is required for the generalized SMBDD method in the time domain.|
Before the generalized SMBDD method can be applied to real building structures, experimental investigations are necessary. This thesis therefore presents an experimental investigation on the generalized SMBDD method through shaking table tests. Three three-story shear building models of different lumped masses are designed and manufactured. The column width of each building model is reduced from both sides of the columns at designated stories to simulate building damage. Different damage severities are simulated by different reductions of column widths. A total of eight damage scenarios of different single or multi damage locations and different damage severities are constructed in the building models step by step. The undamaged and damaged shear building models are subjected to ground motions generated by the shaking table. Two band-limited white noise time histories and two colored noise time histories are randomly simulated and inputted as scheduled into the control system of the shaking table to generate the expected ground excitation to the building models. The ground motion and the displacement responses of each building model at each floor are recorded. The recorded ground motion and building responses as well as identified structural damping ratios are then used to identify damage locations and severities using the generalized SMBDD method in both the frequency domain and the time domain. To provide a basis for the assessment of the proposed damage detection method, the theoretical value of damage severity for a given damage scenario is also computed using the flexibility method and the principle of virtual work. The identified damage locations and severities are compared with the theoretical values. The comparison is found satisfactory in both the frequency domain and the time domain. The effectiveness and feasibility of the generalized SMBDD method are therefore demonstrated by the shaking table tests. The research work on structural damage detection then makes further progress by proposing a new stochastic damage detection method based on the generalized SMBDD method in consideration of random parameters or uncertainties which are inescapable for civil structures. Although many researchers have studied statistical structural damage identification to consider the uncertainties involved in civil structures in the last three decades, most of the statistical approaches still stay in primitive forms. Generally speaking, they have one or more of the following problems: computational complex, inherent limitation to identifying presence of damage only, ineffectiveness in dealing with uncertain parameters which are not normally distributed, and prohibitiveness for most practical applications. The new stochastic damage detection method proposed in this thesis can not only locate structural damage but also identify damage severities without the extensive computational efforts, and it can also handle both Gaussian and non-Gaussian random parameters. The algorithm of the stochastic damage detection method is first presented in the thesis. New damage indices are proposed to identify damage locations and damage severities. The numerical investigation is then conducted to demonstrate the proposed method through a shear building structure. The first damping ratio of the shear building is selected as a random parameter with a lognormal distribution. Three damage scenarios including single and multi damage scenarios are explored. Numerical analysis results show that the proposed method in consideration of uncertainties or random parameters can identify both damage locations and damage severities of the building structures of random parameters satisfactorily. The last piece of work described in this thesis is concerned with reliability assessment of instrumented building structures. Although the deployment of structural health monitoring systems has now attained some degree of maturity, the application of measured response data of a building structure for evaluating structural reliability is still in its infancy. In most of previous investigations, structural system identification and structural reliability assessment are treated separately. When uncertainties or random parameters are taken into account, the stiffness parameters of all the elements in a building structure identified are random parameters coupled with each other. Under this circumstance, it is prohibitive to evaluate structural reliability by the current reliability analysis methods for civil structures. This study therefore presents two integrated methods to evaluate structural component reliability and structural system reliability, respectively. The integrated methods accept the measurement responses as input and produce as output the reliability of the concerned instrumented structures. Structural system identification is embedded in the procedure of the reliability analysis in the proposed methods. Numerical investigation is conducted on the aforementioned stochastic shear building structure with three damage scenarios under the EL Centro excitation. The computation results manifest that the undamaged shear building has higher reliability than the damaged shear buildings of various damage scenarios. The larger damage severity, the lower is the reliability of the building. In addition, the values of structural system reliability are always no larger than those of structural component reliability. The structural reliability of stochastic building structures integrated with system identification can be effectively evaluated by the proposed methods using limited measurement responses. In summary, the research described in this thesis involves the development and application of the SMBDD method, the stochastic damage detection method, and the reliability assessment methods for stochastic structures. Extensive numerical and experimental investigations have demonstrated that the SMBDD method is not only sensitive to local structural damage but also insensitive to measurement noise, and this method can be used in either the frequency domain or the time domain. The stochastic damage detection method enables the damage identification of building structures of uncertainties or random parameters. By using the proposed reliability assessment methods, the reliability of a stochastic structure without explicit damage identification can be obtained by using limited measurement responses.
Files in This Item:
|b23517050.pdf||For PolyU Staff & Students||3.35 MB||Adobe PDF||View/Open|
|12095.pdf||For All Users (Non-printable)||3.37 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: