Evolutionary power spectral density of recorded typhoons at Sutong Bridge using harmonic wavelets

Abstract

The rapid development of structural health monitoring system (SHMS) over last decades makes the accurate monitoring of structural responses and surrounding environments possible. The obtained wind speed data during recorded typhoons typically show significant non-stationary features. Run-test method was employed to verify the stationarity of the recorded wind speed data. To investigate the non-stationary characteristics of recorded typhoons, the evolutionary power spectral densities (EPSDs) of four typhoons at Sutong Bridge site were estimated based on the wavelet transform (WT). A numerical experiment was conducted to identify the proper wavelet bases in the estimation of the EPSDs for non-stationary turbulent wind flows. Specifically, recorded data of typhoons Fung-wong, Meari, Damrey, and Matmo from the SHMS on Sutong Bridge were utilized to calculate the EPSDs using the generalized harmonic wavelet (GHW) and the filtered harmonic wavelet (FHW). The EPSDs estimated by GHW and FHW were comparatively investigated. The normalized Kaimal EPSDs were also compared with the normalized recorded EPSDs of the four typhoons. Results show that the HW-based EPSD estimations for the recorded typhoon data are accurate and reliable. The EPSDs estimated by FHW show the optimum time resolution in comparison with GHW and Morlet wavelet. The energy of the recorded wind speed data mainly concentrates in low-frequency band. The Kaimal EPSDs may be not appropriate to depict the evolutionary features of recorded data during typhoons Fung-wong and Matmo in this study. This observation may provide reference values for wind effect analysis on the long-span bridges with low natural frequencies.