Sensitivity of LES predictions of wind loading on a high-rise building to the inflow boundary condition

Abstract

Large eddy simulations (LES) can provide a powerful tool for cladding design, but their accuracy strongly depends on the correct representation of the incoming atmospheric boundary layer (ABL). The objective of this work is to validate LES of a wind tunnel experiment of wind loading on a high-rise building, focusing on accounting for sensitivity to the incoming ABL. We define three uncertain parameters, i.e. the roughness length of the terrain, the turbulence kinetic energy, and the integral time-scale, using the available experimental data. Subsequently we perform 27 LES using different combinations of these parameters. The results indicate that (1) accurately quantifying the turbulence statistics at the building location, rather than at the inflow, is essential when analyzing LES of wind loading, and (2) correctly accounting for inflow uncertainty when performing validation against experiments could drastically improve confidence in the predictions. By accounting for this uncertainty, we predict intervals that encompass the experimental data for the rms pressure coefficients. The magnitude of the local peak pressure coefficients is generally under-predicted by 12%; however, the predicted intervals for design pressure coefficients on cladding panels of different sizes located at the upper-windward corner of the building fully encompass the experimental result.