Determination of wind load on high-rise buildings by applying Computational Fluid Dynamics

The interest of Computational Wind Engineering to estimate wind loads for the structural design of high-rise buildings has significantly increased with easier access to ever-growing high-performance computing. To ensure accurate comparisons between experimental and digital wind tunnels, the approach of replicating an experimental wind tunnel with Computational Fluid Dynamics (CFD) and performing Large Eddy Simulation (LES) was adopted.

In this study, the Precursor Database method was applied in the CFD simulations, where the approaching wind flow is created in an auxiliary domain, and the target building is composed in a test section domain. 13 precursor domains were tested to find an optimal model that was workable in size while ensuring high accuracy for reflecting the atmospheric boundary layer.

The standard tall CAARC building was investigated for five wind angles of attack with both experimental and digital wind tunnel tests. The façade pressure coefficients, peak predicted structural response base moments and appertaining Floor-by-Floor loadings were evaluated. The results of the two approaches were in good agreements, and the mean average differences were found to 10.5 %, 13.0 % and 12.8 % for the base peak Mx, My and Mz moments.

Based on the gained knowledge, guidelines to standard CFD setups of the inflow and test section domain were suggested. To test the viability on other shapes, the suggested guidelines were employed on the CAARC building with chamfered and rounded corner modifications for seven wind angles of attack. The simulations were executed as blind tests to demonstrate a transparent study, where no optimisation of the mesh or numerical schemes were performed.

Encouraging agreements were found, however with slightly decreased accuracy, with average deviations of 14.92 %, 14.08 % and 22.07 % for the base peak Mx, My and Mz moments of the chamfered corner high-rise building, and 12.56 %, 11.58 % and 15.51 % for Mx, My and Mz for the rounded corner high-rise building.

The results showed that further research is necessary for wind directions perpendicular to the windward face. Lastly, the suggested guidelines were employed on a single high-rise building, and on a large real project in Ramboll that was compared to experimental wind tunnel tests from an external wind engineering consultant.

Despite the required extra research, the suggested guidelines are believed to be useable by computational wind engineers in the future, in order to secure that wind loading CFD simulations have a certain standard and accuracy. Moreover, CFD is a useful tool that allows us to observe the invisible wind flow from accurate numerical simulations so that we, in the end, can learn and develop as engineers.