Structural Vibration Signal Processing And System Development Based On Analytic Mode Decomposition

Structural dynamic response analysis and feature extraction is one of the hot research issues in the area of structural health monitoring data analysis.Especially,the instantaneous characteristics of the structure vibration signal,is the critical index for structural parameter identification and damage detection,and plays significant roles for structural safety evaluation.In order to decompose the structural vibration signal and obtain the time-frequency characteristics of the vibration response,an extended discrete analytical mode decomposition is proposed.Based on the decomposition of extended discrete analytic mode,the optimization method of cut-off frequency selection is also proposed,and the cutoff frequency of the traditional Fourier spectrum is replaced with the self-regression power spectrum.In order to provide a convenient,fast and practical application of civil engineering structure vibration signal processing tool,this thesis develops the structure vibration signal processing software named as Structural dynamic response analysis(SDRA)based on the commercial software MATLAB.Finally,the modal parameter identification of Ganjiang river highway bridge is identified based on developed SDRA software and vibration signal processing method.The main research works are as follows: 1.Based on analytical model decomposition,the extended discrete analytic mode suitable decomposition method is proposed..The four steps decomposition method is also developed to extract the low frequency component from the vibration signal.2.In order to optimize the cut-off frequency,the self-regression power spectrum method instead of the Fourier spectrum method is developed to obtain the cutoff frequency.Numerical examples are used to verify the accuracy and validity of the self-regression power spectrum in the cut-off frequency selection.3.Based on the MATLAB platform,the SDRA software is also developed.The software interface,the preparation of menu key,and the specific function of the software,are introduced in the thesis.The developed software provides the Fourier spectral,and self-regression power spectral analysis.The cutoff frequency can then be obtained from the signal spectrum.The instantaneous amplitude and frequency of the decomposed signal can also be obtained based on the software with Hilbert transform.The time-frequency spectrum can then be plotted by the software.4.Finally,the Ganjiang highway bridge is taken as an engineering example.The vibration accelerations of the bridge due to ambient vibration are measured,and the cutoff frequency of signal is obtained by using the self-regression power spectrum.Then,the first two modal responses of the bridge is obtained by analytic mode decomposition,and the first two modes of the bridge are finally identified.