Wind Characteristics Of Guangzhou Tower At High Altitude And Prediction Of Wind Pressures

Modern structures are constructed in higher direction, and many of them arebeyond450m. For these high and soft super-tall building, its inherent frequency isclose to the predominant frequency of strong winds and the structural response isfurther intensified under wind load. So the wind load has become the main designload of these super-tall buildings. But we are not clear on wind characteristics andwind pressure of the construction surface at the height above450m, there may appearsome new questions which we don’t have to focus on generally high buildings.Therefore, the major objective of this research study is to investigate the windcondition information and wind pressures through installing anemometer at the top ofGuangzhou Tower (464m) and wind pressure sensors at its mast (500m). The maincontent includes:(1)The paper simply introduces the development status of wind characteristicsand wind pressure and summarises the existence problem and expounds thesignificance of research study.(2)The wind field characteristics of the two measured typhoons are analyzedrespectively according to Stationary Wind Speed Model (SWM) and Non-stationaryWind Speed Model (NSWM). The study showes that the average wind speed obtainedusing the two models is basically the same. The turbulence intensity and integrallength calculated by non-stationary wind speed model are less than the values of thestationary wind speed model, but the gust factor is greater than the stationary windspeed model’s. Another, fluctuating wind speed obtained using the non-stationarywind speed model is better meeting the normal distribution.(3)Power spectrum for along wind velocity is converted to that for fluctuatingwind pressure according to the method proposed by Tamura etc and the convertedpower spectrum is compared with measured spectrum. It can be found that measuredspectrum agrees well with the Karman spectrum. The relative coefficient andcoherence function in horizontal and vertical direction are analyzed respectively andthe coherence curve is fitted based on the least squares procedure for single parameter.The results show that horizontal correlation index coefficient is less than the valuesuggested by Davenport, but the vertical correlation index coefficient agrees wellwith Davenport’value.(4) The time series model and the detailed modeling steps are introduced. In addition, wavelet decomposition is applied and a new method named wavelet-timeseries model is proposed. From the prediction cases, it can be found that wavelet-timeseries model significantly improves the wind speed prediction accuracy.(5)The wind pressure field is decomposed and reconstructed using the ProperOrthogonal Decomposition (POD) technique and the reconstruction effect is good justusing several first intrinsic modes.Finally, according to the measurement windpressure data of the known points, the POD technique is used to predict the timeseries of wind pressure where the pressure tap is not distributed and it can obtainsatisfactory results.