Evaluation of wind loads on high-rise buildings at various angles of attack by wall-modeled large-eddy simulation

Abstract

This study is to validate the accuracy of large-eddy simulation (LES) of wind pressures and loads on high-rise buildings at various angles of attack based on approaching flow simulation framework with synthetized inflow turbulence. The wake flow field and pressure distribution on building surfaces are quite different at different angles of attack due to sensitivity of flow phenomena (separation, vortex, wake effects, etc.). To better capture such complex flow features around the building, local octree mesh refinement is proposed as an efficient meshing strategy to improve the simulation of wind loads statistics with acceptable additional computational cost. Its accuracy is demonstrated by comparing the simulation results from progressive levels of mesh refinement with experimental data. It is shown that octree mesh refinement can improve the prediction of skewness and kurtosis of wind pressures, which are important for the prediction of peak pressures. Moment-based Hermite model approach is introduced for better assessment of probability distribution of peak pressures from short duration of simulation samples.