Simulation of a ground-mounted prism in ABL flow using LES: on overview of error metrics and distribution

Abstract

The accuracy of wind loading predictions using Large Eddy Simulation (LES) is usually influenced by numerous model parameters, which can influence the obtained results. The validation of numerical simulations with traditional Wind Tunnel Test (WTT) is still an important task, necessary to increase our a priori knowledge of possible inaccuracies and set up mitigation strategies. In this study, LES is used to simulate the wind fields around an isolated model high-rise building, under seven wind attack angles and validated with WTT results. The influence of various settings and parameters on the model performance is studied. For the angle of attack showing higher inaccuracy, different mesh refinement strategies and turbulence models are tested. Results indicate that LES can accurately predict the mean and local maxima of the pressure coefficients for both perpendicular and skew wind attack angles, as well as reproduce global forces and their envelopes with very good accuracy. Conversely, pronounced errors are found in the prediction of the pressure coefficient standard deviation and the local minima. The highest deviations between LES and WTT are found close to the leading edge in correspondence of flow separations which are observed in WTT and not in LES for skew flows. The addition of boundary layer cells and the use of different subgrid models have very limited effectiveness in modifying the obtained results for the analysed case.