We have performed Large-Eddy Simulation of turbulent boundary layer over simplified urban like roughness. Totally six types of arrays are adopted with several roughness densities and roughness aspect ratios. The turbulent flow fields at a pedestrian level are investigated to identify probability characteristics of unsteady flows. Conclusions are summarized as follows. First, the probability density functions of wind speeds at a pedestrian level show long-tail shape with positive skewness for all arrays and well explained by dimensionless probability variable defined by the mean and standard deviation of wind speeds. On the other hands, the discrepancies of the probability between arrays become considerable for events with wind speeds larger than triple of the standard deviations. Second, such events with large wind speed rarely happen; the cumulative probability shows approximately 1%. Lastly, linear relationships are confirmed between the mean and percentile wind speeds. These findings will help the understanding of the unsteadiness nature of turbulent flow in urban boundary layer especially when evaluating very rare events such as gusts or extreme low winds.
Computational fluid dynamics (CFD) analysis using large-eddy simulation (LES) and standard k-ε model (k-ε) was conducted for flow within and above urban canyon under various conditions of thermal stratification. The results of LES and k-ε were compared with the wind tunnel experiment to investigate the prediction accuracy of CFD analysis. k-ε failed to reproduce the spatial distribution of the mean wind velocity vector and mean temperature within the urban canyon compared with LES in unstable case. k-ε underestimated the Reynolds stress <u1’u3’> and turbulent heat flux <u3’T’> within the urban canyon in unstable case. It is thought that the prediction accuracy of k-ε decreases because k-ε underestimates the exchange of momentum and heat within and above the urban canyon compared with LES and the wind tunnel experiment.
We have performed Large-Eddy Simulation of flow and pressure fields over two types of urban-like block arrays in order to investigate the relationships between instantaneous building wall pressure and surrounding flow fields. The adopted block arrays are: lattice-type square and staggered pattern. The analysis on the distributions of statistics for wall pressure shows that the large values of the statistics are observed along with the top and side edges of the wall for the square array, whereas they are shown on an upper half region of the wall for the staggered array. These large values are due to air flow introduction from both above and side of the target block for square array; in contrast, strong flows introduced into canopy area cause large pressures of the upper half of the wall for staggered array.