A damage identification method using wavelet package and fractal for cable-stayed bridge

In this paper, a new structural damage identification method using correlation dimension (Dr) of wavelet package of vibration detection signals to monitoring structural damage is proposed based on the significant ability of wavelet in signal local analysis with that of fractal in nonlinear processing. Firstly, the vibration detection signal is decomposed into each frequency band according to wavelet packet analysis, and then Dr of each wavelet packages are calculated, lastly, cross entropy correlation dimensions between damaged and undamaged signal are served as an index for damage identification. A numeric example shows that the method can resist noise disturbs and detect the structural damage effectively.