Determination of time-varying means for non-stationary typhoon winds

The determination of time-varying means (TVMs) for nonstationary typhoon winds is of significance in the investigation of wind characteristics and the prediction of structural responses for super high-rise buildings. Recently, time-frequency analysis methods, especially Ensemble Empirical Mode Decomposition (EEMD) and discrete wavelet transform (DWT), have been widely adopted to extract TVMs of typhoon winds. A novel time-frequency method called analytical mode decomposition (AMD) was recommended here to determine the TVMs of typhoon winds. EEMD, DWT, and AMD were all used to extract the TVMs and obtain the turbulent wind characteristics of a typhoon wind. A comparison and further discussion were conducted on the results, effectively validating the great potential of AMD in determining TVMs. Nevertheless, the selection of the appropriate cut-off frequency for TVMs poses major challenges to their widespread application. An undated approach considering power spectrum entropy (PSE) was proposed to determine the cut-off frequency. Additionally, the effectiveness of this PSE-based approach was successfully verified by a numerical test with four deterministic types of TVMs. The study successfully utilized the AMD, EEMD, DWT, and PSE-based approach to extract the TVMs of field wind data and investigate the turbulent wind characteristics atop the Shanghai Tower during the occurrence of typhoon Lekima.