Length-Scale Similarity of Turbulent Organized Structures over Surfaces with Different Roughness Types

Abstract

We examine the similarity of turbulent organized structures over smooth and very rough wall flows. Turbulent flow fields in horizontal cross-sections were measured using particle image velocimetry, and the characteristics of turbulent organized structures over four types of surfaces were investigated. Measurements were conducted at several measurement heights across the internal boundary layer. The length and width of turbulence structures were quantified using a two-point correlation method. We selected two thresholds of two-point correlation coefficients to consider both large-scale and small-scale structures; the validity of these choices was examined through the analyses using proper orthogonal decomposition. For large-scale structures, the length and aspect ratios (streamwise length/spanwise width) of structures were highly correlated with the velocity gradient for each measurement height and boundary-layer thickness. This relationship was also examined in the results of previous studies, and the scaling of the aspect ratio with the non-dimensional velocity gradient again showed the importance of the velocity gradient, with slight differences found between smooth and rough surfaces. In contrast, the small-scale structures exhibited weak dependency on the velocity gradient and boundary-layer thickness. Instantaneous snapshots of turbulent organized structures at the same shear level also displayed differences in small-scale structures, but the structures of the organized motions resembled each other, as in the results of the two-point correlation method.

Effect of upstream flow regime on street canyon flow mean turbulence statistics

“…Nevertheless, the present results agree with those of Volino et al [35] who found that the length scales of the 2D roughness case were significantly higher than the 3D case throughout the height of the boundary layer. Conversely, the present results and those of Volino et al [35] are contradicted by Takimoto et al [34] whose results show that 3D configurations result in larger L u than those of 2D configurations with equal AR. This discrepancy may be a result of the simulation method leading to a smaller boundary layer to building height ratio as no spires were used by Takimoto et al [34] to produce turbulence and it is clear that L u tapers off to very small values with increasing height in their work.…”

“…Conversely, the present results and those of Volino et al [35] are contradicted by Takimoto et al [34] whose results show that 3D configurations result in larger L u than those of 2D configurations with equal AR. This discrepancy may be a result of the simulation method leading to a smaller boundary layer to building height ratio as no spires were used by Takimoto et al [34] to produce turbulence and it is clear that L u tapers off to very small values with increasing height in their work. All results seem to approach a similar value as z/h approaches unity except the 2D configuration of Volino et al [35] with a high AR.…”

“…The influence of varying the roughness geometry on the turbulence integral length scales was also investigated. Volino et al [35] and Lee et al [16,17] found that this change of geometry has a strong influence, with larger length scales above the roughness for 2D cases than 3D cases [35], but Takimoto et al [34] show no consistent variation of integral length scales between the 2D and 3D cases. Of the studies presented here, Volino et al [35] makes the most definitive conclusions concerning integral length scales.…”

Dispersion of a Passive Scalar Fluctuating Plume in a Turbulent Boundary Layer. Part I: Velocity and Concentration Measurements

“…Therefore the estimate of the Eulerian integral length scales is generally calculated as the distance at which the correlation function falls below a threshold value. For example Bewley et al (2012) assumed a value of 0.5 whereas Takimoto et al (2013) adopted 0.4. Note that this method may be problematic when computing scales associated to the ρ uu functions, since the extent of the iso-correlation lines corresponding to the threshold value may not be fully captured by the PIV field (see Fig.…”

Street-Level Ventilation in Hypothetical Urban Areas

“…The near-surface velocity gradient ∂ u /∂z| z/δ=0 over different rough surfaces shows a similar behavior, i.e., velocity increases more sharply with increasing aerodynamic resistance, signifying the elevated drag in the near-surface region (Figure 2). The length scales of flows are highly correlated with the velocity gradient at different elevation [54].…”