Impact of momentum sources on pedestrian-level peak wind predictions of an urban-like array using large-eddy simulations and statistical models

The effect of urban geometries on the peak wind speed at the pedestrian level was investigated using simplified urban-like arrays in wind tunnel experiments. To scrutinize the turbulent flow's spatial and temporal characteristics, large-eddy simulations (LESs) were adopted to simulate suitable experimental conditions using an external force accelerating the flow with the periodic boundary condition. Although previous studies have revealed that conventional LESs use a constant pressure gradient as a momentum source, driving the flow differs from those in developing boundary layers in experiments, and the effect of the momentum source on peak wind speeds at the pedestrian level remains unknown. Therefore, this study used a series of LESs based on the three driving methods to investigate the impact of the momentum provision on the relevant statistics and peak values. The turbulent statistics showed good agreement among the cases driven by the momentum sources regardless of the shape of the profiles of the momentum source. Peak wind speeds, quantified by percentiles, were estimated using statistical models based on the Weibull distribution. Overall, the results showed a good agreement between the LESs and statistical model estimations when higher-order moments were adopted as the estimation parameters.