Investigation Of Modal Parameter Identification Based On HHT And SSI For Engineering Structures Under Ambinent Excitation

Modal parameter identification for civil engineering structures play an important role in several research and application fields, such as structural health monitoring, structural damage assessment, active control (fuzzy control), etc. It is difficult to excite the civil engineering structures by the artificial resources and to measure the excitation signal because of huge dimension and complex configuration. So it is not easy to use traditional methods, such as Fourier Transform, with known input and output to identify modal parameters for these structures. But, in contrast, the ambient excitation methods just require process the output data to identify modal parameters in time domain or frequency domain. Because of its reliability and feasibility, it has been paid more attention by both researchers and professionals.In this paper, two identification methods which are suitable for operational modal parameter identification are investigated, namely, Hilbert-Huang Transform (HHT) and Stochastic Subspace Identification (SSI). And sufficient effort has been focused on the primary problem of modal parameter identification and computer programming.HHT is a new signal processing technology superior to the traditional one (such as Fourier analysis), and has been proved to be ideal for non-linear and non-stationary data. This paper investigates the basic theory of HHT Method for signal processing. And two simulated signals are taken as examples to illustrate signal preprocessing and the advantage of analyzing signal in time-frequency domain using HHT method. Furthermore, based on the basic theory of modal parameter identification, a new identification method is proposed by combining HHT and Natural Excitation Technique (NExT). Meanwhile, as a promising output-only time domain identification method, stochastic subspace identification is also investigated. SSI use spatial projection technique to eliminate the noise effect. Thus it has better anti-noise capacity. And this method is also more effective for systems with closely-spaced modes. Finally, in order to be used for modal parameter identification under ambient excitation, the approaches are proposed and the corresponding programs are coded using Matlab based on the above-mentioned rationale, respectively.In order to verify the feasibility and practicality of the proposed approaches and the corresponding programs, the original signals from the field ambient vibration test of an arch bridge and the shaking table test of a 12 storeys RC frame model arc processed and modal parameters are identified using the above two approaches, respectively. Identification results and the comparison with the results from traditional fast Fourier transform and finite element analysis indicate that these two approaches and the corresponding programs can identify modal parameters rapidly and efficiently for civil engineering structures.