Study On Damage Identification Method For Shear Frame Structure Under Non-stationary Ambient Excitation

Civil structures will suffer from neglect and overloading, such as earthquakes, typhoons and blasting, leading to accelerated deterioration and even catastrophe. Early identification of damage for a structure plays an important role to avoid further risks, both in life-safety and property of the users. Therefore, vibration-based structural damage identification has been increasingly employed as an invaluable technique in the field of civil engineering during the last few decades. Unfortunately, due to the significant differences among each type of structures, impossible application of artificial excitation and tremendous complexity of ambient excitations, it is very hard to utilize a common method to detect structural damages for all types of civil engineering structures. So, it is important to develop the pertinent damage identification method according to the characteristic of each type structure. As a typical type of engineering structures, the shear frame is widely used for houses, schools, offices and hospitals all round the world. Consequently, developing a suitable damage identification method for the shear frame is very important to structure assessment after an emergency.Focusing on the widely used shear frame structures, the dissertation proposes a new damage detection method using the ambient excitation, such as microseism. The main contents and results are described in detail as following:(1) A novel damage localization and quantitation method for the shear frameA new damage localization and quantitation method is proposed using two arbitrary modal charaters, which includes natural frequencies and mode shape ratios, under the health condition of the structure and an arbitrary modal charater under the damaged condition. Hence, the proposed method is convenient for application. The rate of story stiffness degradation(RSSD), which has a clear physical significance, is used as the damage index and can be used for damage localization and quantitation concurrently. The relationship between the modal charater and the rate of story stiffness degradation is derived based on the principle of structure dynamics. And then the equations are established using RSSD and are solved by the restricted least-square algorithm. Finally, the RSSD is employed as the damage index and the damage location and severity are indentified concurrently utilizing RSSD.(2) A new mode frequencies and mode shape ratios identification method under non-stationary ambient excitationFocusing on the non-stationarity characteristic of ambient excitations, a new mode frequencies and shape ratios identification method is proposed. The traditional random decrement method is expanded and the free vibration response curves of structures can be extracted from the structural dynamic response under non-stationary excitations. And then the mode frequencies and mode shape ratios can be identified using the singular value noise reduction algorithm and eigen-system realization algorithm.(3) A novel uncertainties reduction method for damage detectionThe RSSD damage index will fluctuate by the uncertainties of the ambient excitations, environmental factors, noise and calculation errors. Based on optimization algorithm and interval theory, a new uncertainties reduction method is developed in the article. Firstly, the interval models of the damage index are established. And then, a new damage localization and quantitation method based on the interval models of the damage index is proposed. The proposed method avoids the deficiency of the existing method based on probability theory, which requires lots of test data to establish the probability model of the damage index, and can be used for damage localization and quantitation with small test data.(4) Experimental verficationA plane steel frame structure is used to verify the proposed method. The damage occurred at columns from one single floor or two floors are considered in the test. And the uncertainties are come from the different seismic waves, measurement noise and calculation error. The result shows that the proposed method can be used to extract the mode natural frequencies and mode shape ratios from the non-stationary structural dynamic response and reduce the influence of the uncertain factors. The proposed damage localization and quantitation method for the shear frame under non-stationary seismic load is verified, while there are uncertainty in the process of damage detection.