The wind-induced dynamic load on high-rise structures by the use of CFD

Abstract

During the last decades, the development of high-rise increased quickly. The Eurocode is frequently used to make wind calculations for high-rise structures. To make those calculations, the dynamic wind load is simplified to a conservative equivalent static load. An alternative to calculate wind loads is a physical wind tunnel test. Wind tunnel tests are expensive. An alternative for these methods is the use of Computational Fluid Dynamics(CFD). CFD makes it possible to simulate the wind. At this time, CFD is still very user-dependent and therefore it is stated in the Eurocode that CFD simulations cannot be used as a design tool. This research aims to examine the impact of different parameters in evaluating the dynamic loading on a high-rise structure by the use of CFD. The results of the CFD simulations are compared with a spectral approach. The Navier Stokes (NS) equations are the basics for CFD. Different models are used to solve the NS equations. The costs of CFD increase exponentially with demanding more accuracy. An extension of the RANS model is used in this research: the Unsteady-RANS(U-RANS) model. Where the RANS model only computes the steady-state solution in a simulation, the U-RANS captures the slow turbulences as well. For the objective of a dynamic wind load on a high-rise structure the slow turbulences are of more importance than the faster turbulences. Especially the frequencies close to the natural frequencies of the structure are important. The CFD simulations are performed using Star CCM+ by Engineering Company ABT BV. A reference structure is used to validate the CFD simulation procedure with existing available wind tunnel data of the Tokyo Polytechnic University. The wind tunnel data contains dimensionless pressure coefficients on several locations on the model. The results show a lot of noise, which is visible when the time domain's pressure coefficients are translated into the frequency domain. A dominant peak is observed at a frequency of 10 Hz, but other frequencies are triggered as well. The CFD simulation for the same reference block show less fluctuating results than the wind tunnel data: the results are only fluctuating on one frequency of 12 Hz. Afterwards, a case study is used to compare CFD results with a spectral approach. A high-rise structure of 40 x 40 x 200 meters is used for this study. The results of the spectral approach are compared with existing literature using the structural factor cscd. The results coincide. Several CFD simulations are performed on the same model. The absolute values of the peak velocity pressure are of the same order of magnitude, with a difference of 10 to 25 percent. However, the values of the standard deviation differ more substantially. The CFD simulation stores the average pressure over an area of 10 x 10 meters, which balances out some fluctuations. Furthermore, it can be concluded that the CFD simulations underpredict the absolute values of the pressure consistently compared to the spectral approach. It can be concluded that the CFD simulations require exact defined input values.