Description
This research concerns Structural Health Monitoring (SHM), with the use of an advanced modal analysis. Using accelerometric measurements, the classical approaches are the study of correlation between damage and the frequencies, mode shapes or damping ratios of the structure, and the detection of cracks from modal curvatures. Moreover, without experimental reference solution, identification of local defects or non-linearities is challenging. This paper shows the potentialities of the proper use of the continuous wavelet transform (CWT) for an analysis of the dynamic responses of railway bridges under the passage of high-speed train. We analyzed the theoretical response; Firstly, the signal must be partitioned properly, as it is comprised of distinct time intervals containing specific modal information (especially bridge and bridge + train). An improper interval definition could lead to unexpected edge effects in the case of a CWT analysis, and if the Fourier transform is used, excessive leaking. Secondly, frequencies originating from the train excitation have to be clearly identified and separated from the natural frequencies of the whole system. A CWT-based procedure is then proposed to process the global signal with emphasis on the choice the wavelet function. An application using experimental accelerometric data collected during the passage of train on a real existing bridge is given. The dynamic response of the “bridge plus train” system is found to have a very distinct non-stationary shape dependent on the finite length of the train and the periodic spacing of its axles. It illustrates CWT capability to detect and characterize nonlinear behaviors.
