Effect of assumed inflow turbulence intensities and lengthscales generated via spectral representation on LES of tall building wind response

Turbulence integral lengthscale and lateral and vertical turbulence intensities are sometimes not measured in wind tunnel experiments. The resulting uncertainty in inflow conditions cascades into uncertainties in the surface pressure, drag and lift coefficients, and the wind response of tall buildings. This paper utilizes Large Eddy Simulation (LES) with a divergence-free turbulence generator to investigate a set of five inflow conditions generated using different assumptions related to building height and turbulence integral lengthscale, and the ratio between three components of turbulence intensities. The effects of the inflow assumptions on the downstream incident flow, on the surface pressure coefficient of a tall rectangular building of 1:4 width to height ratio, on drag and lift moment coefficients are all evaluated. Furthermore, an example steel-frame-tube structure is considered to investigate the effects of the inflow turbulence on its background, resonant and peak displacement response using modal analysis with gust peak factors. It is observed that distinct turbulence inflow conditions tend to converge towards similar statistical profiles downstream. However, within this envelope of convergence, a range of potential values persists, indicating significant remaining uncertainty. The inflow turbulence integral lengthscale is found to have an effect up to 13% on the peak along-wind response and 4% on the across-wind response. The lateral and vertical turbulence intensity is found to have effects of up to 13% and 24% on the two peak responses, respectively. These results highlight the importance of considering a range of inflow assumptions when conducting LES to reproduce and interpret WT results.