The Study On Strain Modal Parameters Identification For Structures Under Ambient Excitation

In this dissertation, the investigation is mainly focused on structural strain modal parameters identification, meanwhile, the application of the strain mode type index in structural damage location identification has been discussed by the author. Combining with theoretical derivation, finite element numerical simulation and experimental verification, three contents are expounded in the dissertation as follows:(1) Based on relationship between strain and displacement, like the Natural Excitation Technology and the Eigensystem Realization Algorithm using displacement response data, the Natural Excitation Technology and the Eigensystem Realization Algorithm for strain response data are established respectively. What's more, the NExT-ERA method based on strain response data for the identification of structural strain modal parameters is proposed, and the corresponding program is written based on MATLAB platform.(2) The identification accuracy of the NExT-ERA method based on strain response data is verified by the experiments of a steel cantilever beam. The experimental investigation shows that the natural excitation technique which using on the strain response signals is able to gain the approximate strain impulse response signals for structure. using the approximate strain impulse response signals from the natural excitation technique as input signals, the Eigensystem Realization Algorithm based on strain response data can validly identify the structural strain modal parameters. The difference between the identified structural inherent frequencies and that identified by Peak Picking method (PP) is within 3%, meanwhile, the identified strain mode shapes are comparable to the accuracy of that identified by PP. During the experiment, the identified results of damping ratio is unstable and its fluctuation range is considerable large, and this may be cause by the difference between the Viscous Proportional Damping theory and actual damping. Therefore, an improved stabilization diagram is presented to obtain the order of the system and distinguish the true strain modes quickly and accurately. The improvement to classic stabilization diagram is introducing the power spectral density curve of strain response signal and eliminating the damping ratio indicators as well as keeping the frequency indicators.(3) With the simulation of structural damage by elemental stiffness reduction factor, this paper simulates the transient response of a damaged steel cantilever beam under Guass white noise excitation by mean of finite element analysis software and gain the strain response signals of the cantilever beam, and then identify the first four strain mode shape of the damaged cantilever beam based on the NExT-ERA method which use the strain response signals as input signals. Furthermore, it is discussed and investigated in the paper that the effect of damage location identification by mean of the strain mode type index such as strain mode shapes and strain mode deviation. From the investigation result, it shows that the strain mode shapes and strain mode deviation are able to identify the damage location effectively, and the strain mode deviation can identify the damage location of a small damaged structure better. For small damaged structure, if the SNR of the acquired signals is lower, then the impact to the identification result by noise is bigger, moreover, the effect to the result of different order strain mode deviation is vary by the noise. Generally, the strain mode deviation, with good anti-interference ability of noise, has a strong ability on structural damage location identification. Besides, it is able to identify the damage location well by mean of taking the identification results of different order strain mode deviation into consideration synthetically.