Investigation Of Structural Damage Detection Based On Hilbert-Huang Transformation

The structural damage detection is the key of the structural health monitoring techniques. Damage detection methods based on the structural dynamics are involving new methods research on signal testing and data processing techniques, not only including the deep research on the theory of structural dynamics and experiment skills as well. The Hilbert-Huang Transform (HHT) is suited to be used in the nonlinear and unstable signal processing for it offers a approach to describe the signal's instantaneous characters. The application of HHT in the oceanography, seismology, bioengineering and structure health monitoring has already proved the advantage of this method. But the HHT still needs further study for it is only proposed recently. The main work in this paper includes three parts:(1) Empirical Mode Decomposition (EMD) method is the basic item of HHT. Based on the Mirror Extending Method and General Extrema Extending Method, a new method named Symmetrical Extrema Extending Method is proposed to eliminate the end effect of the Empirical Mode Decomposition (EMD) method. Analytical results show that the algorithm of this method is simple to implementation with good elimination of end effect of EMD, and can maintain the character of original signal when reconstructed.(2) Correlation function amplitude vector (CorV) is used to identify the structural damage. The influence of the exciting frequency bandwidth in this method is studied through analyzing the damage detecting effect while the structural system is excited by signals with different frequency bandwidths.(3) Frequency of intrinsic mode function (Fimf) and Amplitude of intrinsic mode function (Aimf) are proposed on the base of EMD and HHT. These two concepts are also used in the damage detection, and both simulation and experimental results show that they can work effectively.