Numerical Wind Tunnel Simulations Of Wind Effects On Super High-rise Buildings And Long-span Roof Structures

Wind effects on buildings and structures are mainly determined by means of wind tunnel testing, full-scale measurement or computational fluid dynamics (CFD) predications. Compared to wind tunnel testing and full-scale measurement, CFD simulation has many advantages such as:(1) it can provide much more detailed, visualized, and comprehensive information on flow fields and pressure distributions on building surfaces, even on relatively small areas with large pressure gradient variations. In such a case, it is hard or impossible for the surface pressures to be captured by wind tunnel testing and full-scale measurement.(2) There are a number of limitations involved in wind tunnel testing, like the difficulty of fulfilling model tests under high Reynolds number flow conditions. However, for CFD simulations, it is feasible to conduct full-scale size simulations with high Reynolds number flows by adopting recently developed CFD techniques. Such drawbacks can thus be overcome accordingly.(3) CFD simulation is able to reproduce realistic wind flow fields in the atmospheric boundary layer to ensure obtaining credible results of wind effects on structures. Furthermore, it is convenient to analyze the effects of different turbulent flow fields on wind pressure coefficients, wind forces and wind-induced responses of structures.(4) It is convenient to conduct parameter analysis, mechanism investigation of wind flow phenomenon and wind effects on structures.(5) There is no need to make expensive models or purchasing costly sensors. Hence, it has the merit of low costs.Steady/unsteady Reynolds averaged Navier-Stokes (RANS) methods used to be adopted to study wind effects on buildings and structures. But, in recent years, Large-Eddy Simulation (LES) approach has been identified as a useful tool for investigation of wind loads and wind-induced response of civil structures. Although encouraging results have been achieved by LES, most of the concerned buildings in previous studies with LES are too idealized or simplified, such as with rectangle or circular shapes. Meanwhile, most commonly utilized sub-grid scale (SGS) models are merely suitable under the conditions of relatively high-order numerical discritization and high quality structural or hexahedral grids. This study adopts novel key techniques of LES to predict wind effects on super-tall buildings and long-span roof structures with irregular shapes. Besides, the distributions of wind-driven snows on a long-span roof are investigated by means of CFD simulations. In addition, this study estimated the noise levels and distributions around wind turbines installed in a super-tall building, the Pearl River Tower. The major contents involved in this dissertation are as follows: (1) Basic theoretical knowledge of CFD simulations of wind effects on structures is briefly introduced first. The LES approach is highlighted as an efficient CFD method for its ability of reproducing complex turbulent flows around and wind-induced loads on buildings and structures. Novel key techniques for LES are introduced, which include a new inflow turbulence generator called the discretizing and synthesizing random flow generation (DSRFG) approach and a new one-equation dynamic SGS model. These techniques play an important role in accurate estimations of wind effects on civil structures by using LES approach and form the theoretical foundation of the research works presented in this dissertation.(2) The wind effects on the Shenzhen new railway station is first studied using the new LES techniques based on the parallel computation on the software platform of Fluent6.3. The predicted distributions of both mean and fluctuating pressures on the station roof are found to agree with those from wind tunnel tests. The numerical simulation approach is proofed to be an efficient tool for structural engineers to assess the wind effects on long-span roof structures at the design stage. Specially, the numerical simulation can provide much more detailed information of pressures on small areas with large pressure gradient variations, such as joints of the roof undersurface and the uprights where such information is hard or impossible to be obtained by wind tunnel testing due to insufficient number of pressure taps in a scaled building model. Furthermore, wind flow characteristics around this station are investigated in detail. It is found that the adopted numerical approach is of great efficiency and effectiveness. It can give finer depiction of wind fields around obstacles and provide reliable pressure distributions on the station building. In addition, the mechanisms for the pressure distributions on the station roof are discussed by analyzing the correlations between the wind pressures on the building surfaces and the wind fields around the building.(3) The wind effects on the Shenzhen Universiade Stadium are studied with consideration of all the major mountains and structures (in full-scale sizes) within a range of10km. Wind fields around the stadium, wind environment at pedestrian level, the wind-induced equivalent static loads on the roof surfaces and interference effects on the pressure distributions from surrounding obstacles are investigated, respectively. The numerical results have been successfully utilized in the design of the stadium.(4) The508m high Taipei101Tower and the660m high Shenzhen Ping-An International Finance Centre are considered as objects for investigation of wind effects on the two super-tall buildings by the LES approach. Based on the computational results, wind flow fields around the buildings, wind pressure distributions, the wind-induced responses and the equivalent static wind loads on these two buildings are presented and discussed. The effects of different turbulent flow fields on the wind pressure coefficients, wind forces and wind-induced responses are discussed. The obtained results are compared with those from wind tunnel tests and full-scale measurements. It is found that the numerical simulation approach as adopted in this dissertation is an effective tool for structural engineers to assess the wind effects on super-tall buildings at the design stage.(5) The phenomenon of wind-driven snow on the Jilin New Railway Station is investigated based on the Euler-Euler two-phase flow theory. UDF programs are compiled to realize parallel computation on the Fluent software platform under Linux system. The distributions of wind-driven snow loads on the station roof and the snow loads under the different incident wind directions are presented and discussed. The distributions of the averaged snow loads on the station roof and their variations with respect to the approaching wind directions are demonstrated. These results provide meaningful references for the design of this station and other similar roof structures in cold climate regions.(6) Pearl River Tower is regarded as a global origination "High Performance" Sustainable tall building installed with wind turbines in four tunnels inside the building for power generation, which may produce notable noises when wind blows through the tunnels. It is thus necessary to predict and evaluate the noise levels and distributions around the Pearl River Tower at the design stage. Broadband noise source model based on the SST turbulence model is used in the numerical simulation of the aerodynamic-induced noise distributions around the Pearl River Tower. The influence of the wind turbines on the noise distributions is discussed and the noise levels inside and outside the tall building are evaluated.(7) To the best knowledge of the author, it is the first time in wind engineering and structural engineering to adopt LES for predictions of wind loads and wind-induced vibrations of large-scale complex structures under extremely high Reynolds numbers (108) conditions. The numerical results are found to be in good agreement with those obtained from wind tunnel tests and full-scale measurements.