Bulletin of the American Physical Society
68th Annual Meeting of the APS Division of Fluid Dynamics
Volume 60, Number 21
Sunday–Tuesday, November 22–24, 2015; Boston, Massachusetts
Session R22: Turbulent Boundary Layers: Roughness 
Hide Abstracts 
Chair: Christina Vanderwel, University of Southampton Room: 210 
Tuesday, November 24, 2015 12:50PM  1:03PM 
R22.00001: Characteristics of secondary flows in roughwall turbulent boundary layers Christina Vanderwel, Bharathram Ganapathisubramani Largescale secondary motions consisting of counterrotating vortices and low and highmomentum pathways can form in boundary layers that develop over rough surfaces. We experimentally investigated the sensitivity of these secondary motions to spanwise arrangement of the roughness by studying the flow over streamwisealigned rows of elevated roughness with systematicallyvaried spacing. The roughness is created with LEGO blocks mounted along the floor of the wind tunnel and StereoPIV is used to measure the velocity field in a crossplane. Results show that the secondary flows are strongest when the spanwise spacing of the surface topology is comparable with the boundary layer thickness. We discuss how these results are relevant to flows over arbitrary topologies and how these secondary motions influence the Reynolds stress distribution in the boundary layer. [Preview Abstract] 
Tuesday, November 24, 2015 1:03PM  1:16PM 
R22.00002: The effect of transitionallyrough surfaces on nearwall turbulence Nabil AbderrahamanElena, Ricardo GarcĂaMayoral We present results of DNSs of channel flow with rough walls in the transitionallyrough regime, for $k^+<=15$. Through flow visualization and statistical analysis, we show that the resulting fluctuations can be separated into two components: one due to the overlying nearwall turbulence, and one due to the presence of the roughness. The latter is essentially the phaseaveraged fluctuation that is observed also for laminar flows, but intensely modulated in amplitude by the overlying turbulence. The above decomposition of the fluctuations can be used to develop predictive models for the onset of roughness effects. [Preview Abstract] 
Tuesday, November 24, 2015 1:16PM  1:29PM 
R22.00003: Direct numerical simulations of the dense regime of roughness Michael MacDonald, Leon Chan, Daniel Chung, Nicholas Hutchins, Andrew Ooi We investigate the sparse and dense regimes of roughness using Direct Numerical Simulations (DNS) of turbulent flow over threedimensional sinusoidal surfaces in the transitionally rough regime. The sparse regime is known to lead to an increase in the Hama roughness function, $\Delta U^+$, as the roughness density increases, while the dense regime is associated with a decrease in $\Delta U^+$ as density increases. In this parametric study, the wavelength of the sinusoidal roughness elements is varied while the roughness height is fixed. The minimalspan channel is used, as the high cost of the grid would otherwise make the dense roughness simulations unattainable. It was found that the dense regime began at solidity values (frontal area divided by wallparallel projected area) greater than 0.15, in agreement with the literature. An analysis of the mean momentum balance above the roughness reveals that the decrease in $\Delta U^+$ in the dense regime is due to a reduction in the Reynolds shear stress. This reduction is located just above the roughness crest in the nearwall region, and the difference in the energy spectra of streamwise velocity between smooth and dense roughness clearly demonstrates that this is at long streamwise length scales. [Preview Abstract] 
Tuesday, November 24, 2015 1:29PM  1:42PM 
R22.00004: High Reynolds number roughwall turbulent boundary layers Dougal Squire, Caleb MorrillWinter, Michael Schultz, Nicholas Hutchins, Joseph Klewicki, Ivan Marusic In his review of turbulent flows over roughwalls, Jimenez (2004) concludes that there are gaps in the current database of relevant experiments. The author calls for measurements in which $\delta/k$ and $k^+$ are both largelow blockage, fullyrough flowand where $\delta/k$ is large and $k^+$ is smalllow blockage, transitionallyrough flowto help clarify ongoing questions regarding the physics of roughwallbounded flows. The present contribution details results from a large set of measurements carried out above sandpaper in the Melbourne Wind Tunnel. The campaign spans 45 roughwall measurements using single and multiplewire hotwire anemometry sensors and particle image velocimetry. A floating element drag balance is employed to obtain the roughwall skin friction force. The data span $20 < k_s^+ < 160$ and $30 < \delta/k_s < 200$ across a friction Reynolds number range of $2800 < Re_\tau < 30000$, targeting areas in the parameter space identified by Jimenez (2004) as being sparsely populated by preexisting data. Smoothwall data are also obtained across a similar Reynolds number range to enable comparison of smooth and roughwall structural features. Generally, the data indicate similarity in the outerlayer of smooth and fullyrough wallbounded flows. [Preview Abstract] 
Tuesday, November 24, 2015 1:42PM  1:55PM 
R22.00005: Coupling between roughness and freestream acceleration in turbulent boundary layers Junlin Yuan, Ugo Piomelli To explain various roughwall flow responses to different types of freestream conditions previously observed, we carried out a direct numerical simulation of a spatially developing turbulent boundary layer with freestream acceleration. Unlike the equilibrium (selfsimilar) accelerating scenario, where a strong acceleration leads to complete laminarization and lower friction, in the present nonequilibrium case the friction coefficient increases with acceleration, due to the faster nearwall acceleration than that of the freestream. At the same time, roughness reduces the nearwall time scale of the turbulence, preventing the acceleration from linearly stretching the nearwall eddies and freezing the turbulence intensity as in the smooth case. In addition, acceleration leads to similar decrease of meanvelocity logarithmic slope on rough and smooth walls; this allows a clear definition of the roughness function in a local sense. Interestingly, this roughness function correlates with the roughness Reynolds number in the same way as in selfsimilar or nonaccelerating flows. This study may also help develop benchmark cases for evaluating roughwall treatments for industrial turbulence models. [Preview Abstract] 

R22.00006: ABSTRACT WITHDRAWN 
Tuesday, November 24, 2015 2:08PM  2:21PM 
R22.00007: Turbulent boundary layer over 2D and 3D largescale wavy walls Leonardo P. Chamorro, Ali M. Hamed, Luciano Castillo In this work, an experimental investigation of the developing and developed flow over two and threedimensional largescale wavy walls was performed using highresolution planar particle image velocimetry in a refractiveindexmatching flume. The 2D wall is described by a sinusoidal wave in the streamwise direction with amplitude to wavelength ratio a/$\lambda $x $=$ 0.05. The 3D wall is defined with an additional wave superimposed on the 2D wall in the spanwise direction with a/$\lambda $y $=$ 0.1. The flow was characterized at Reynolds numbers of 4000 and 40000, based on the bulk velocity and the flume half height. Instantaneous velocity fields and timeaveraged turbulence quantities reveal strong coupling between largescale topography and the turbulence dynamics near the wall. Turbulence statistics show the presence of a wellstructured shear layer that enhances the turbulence for the 2D wavy wall, whereas the 3D wall exhibits different flow dynamics and significantly lower turbulence levels, particularly for \textless u'v'\textgreater which shows about 30{\%} reduction. The likelihood of recirculation bubbles, levels and spatial distribution of turbulence, and the rate of the turbulent kinetic energy production are shown to be severely affected when a single spanwise mode is superimposed on the 2D wall. POD analysis was also performed to further understand distinctive features of the flow structures due to surface topography. [Preview Abstract] 
Tuesday, November 24, 2015 2:21PM  2:34PM 
R22.00008: Micro PIV measurements of turbulent flow over 2D structured roughness Joel Hartenberger, Marc Perlin We investigate the turbulent boundary layer over surfaces with 2D spanwise square and triangular protrusions having nominal heights of 100  300 microns for Reynolds numbers ranging from Re$_{\tau }\approx $ 1500 through Re$_{\tau }\approx $ 4500 using a high speed, high magnification imaging system. Micro PIV analysis gives finely resolved velocity fields of the flow (on the order of 10 microns between vectors) enabling a detailed look at the inner region as well as the flow in the immediate vicinity of the roughness elements. Additionally, planar PIV with lower resolution is performed to capture the remainder of the boundary layer to the freestream flow. Varying the streamwise distance between individual roughness elements from one to ten times the nominal heights allows investigation of ktype and dtype roughness in both the transitionally rough and fully rough regimes. Preliminary results show a shift in the mean velocity profile similar to the results of previous studies. Turbulent statistics will be presented also. [Preview Abstract] 
Tuesday, November 24, 2015 2:34PM  2:47PM 
R22.00009: Turbulent shearflow over fractal arrays of surfacemounted cubes Daniel J. Wise, Wernher Brevis The turbulent shearflow over a bottomwall fully covered by periodic multiscale arrangements of obstacles is examined via Particle Image Velocimetry (PIV), Volumetric 3D Velocimetry (V3V) and Acoustic Doppler Velocimetry (ADV) measurements. Three obstacle patterns are utilised, all based on different numbers of iterations of the Sierpinski carpet fractal. In each case 2D/3D velocity fields of the flow formed within the porous channels, namely the flow beneath the mean obstacle height, are presented and analysed with respect to standard statistics such as the mean, rms velocity profiles, and the Reynolds stresses. Pointwise measurements within the obstacle arrays reveal that the presence of the obstacles, and in particular their injection of energy at the associated wavelengths, has unexpected effects on the slope of the energy spectra within the turbulent porous flow. The region dominated by these spectral characteristics is defined. It is also shown that this behaviour is not observed in the outer flow. [Preview Abstract] 
Tuesday, November 24, 2015 2:47PM  3:00PM 
R22.00010: DNS study of amplitude modulation statistics of turbulent channel flows over rough walls Sicong Wu, Kenneth Christensen, Carlos Pantano DNS of long turbulent channel flows over rough walls at friction Reynolds number up to 400 are considered. The walls are hexagonally packed with hemispheres with roughness height h/k=10 and 20 and average spacing between hemispheres from 2 to 4 times the roughness height. A conforming grid approach (unstructured) using spectral finite elements is used to fully resolve the flow with up to 1.6 billion grid points. Analysis of twopoint correlation and Hilbert transform applied to the spectrally filtered fluctuating velocity signals is used to study the interaction between largescale and smallscale turbulence structure. The talk focuses on the effect of roughness parameters on the amplitude modulation accuracy, both for the wallparallel and wallnormal velocities. [Preview Abstract] 
Tuesday, November 24, 2015 3:00PM  3:13PM 
R22.00011: Amplitude modulation of streamwise velocity fluctuations in the roughness sublayer: evidence from largeeddy simulations Ankit Awasthi, William Anderson Largescale motions in the logarithmic region of turbulent boundary layers amplitude modulate the viscous sublayer (Marusic et al., 2010: Science; Mathis et al., 2009: J. Fluid Mech.). This finding has promising implications for largeeddy simulation of wallbounded turbulence at high Reynolds number (wherein the turbulence integral length exhibits linear proportionality with wallnormal elevation). Existing amplitude modulation studies have addressed smooth wall flows, though high Reynolds number rough wall flows are ubiquitous. Under such conditions, roughnessscale vortices ablate the viscous sublayer and result in the roughness sublayer. The roughness sublayer depth scales with aggregate element height, k, and is typically 2k $\sim$ 3k. Above this, Townsend's Hypothesis dictates that the logarithmic layer is unaffected by the roughness sublayer. Here, we present largeeddy simulation results of turbulent channel flow over rough walls. We follow the decoupling procedure of Mathis et al., 2009: J. Fluid Mech., and present evidence that outerlayer dynamics amplitude modulate the roughness sublayer. Below the roughness element height, we report enormous sensitivity to element proximity. Above the elements, but within the roughness sublayer, topography dependence rapidly declines. [Preview Abstract] 
Tuesday, November 24, 2015 3:13PM  3:26PM 
R22.00012: A phenomenological model for the roughness function in turbulent boundary layers with macroscale roughness elements. Jasim Sadique, Xiang Yang, Charles Meneveau, Rajat Mittal ~There has been extensive work done in the past to predict the roughness function associated with rough wall boundary layers and to connect it to the roughness topology. Correlations have been obtained from experiments for a variety of cases and attempts have also been made to use physics based models to obtain the roughness function. In this talk we present a way to derive an explicit formula that connects the rough wall boundary layer parameters to the roughness geometry and arrangement. We assume a twolayer model for the velocity: a loglaw in the outer layer and an exponential profile in the canopy layer, and make use of the concept of `mutual sheltering'. The analysis focuses mainly on rectangular prism shaped roughness elements with different arrangements such as aligned, staggered, rotated, and inclined to the flow, and also with a distribution of heights. It is found that the derived formula, which is simple to apply, matches the results from a variety of largeeddy simulations and experiments. It is also shown that the disparate cases collapse onto a single curve using a parameter depending only on the geometry. This formula gives a quick and accurate way to predict the roughness function from the surface geometry, and can also be extended to other types of surfaces. [Preview Abstract] 
Follow Us 
Engage
Become an APS Member 
My APS
Renew Membership 
Information for 
About APSThe American Physical Society (APS) is a nonprofit membership organization working to advance the knowledge of physics. 
© 2021 American Physical Society
 All rights reserved  Terms of Use
 Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 207403844
(301) 2093200
Editorial Office
1 Research Road, Ridge, NY 119612701
(631) 5914000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 200452001
(202) 6628700