Fluid Flow Analysis of A Shark-Inspired Microstructure

J. Fluid Mech. (2014), vol. 756, pp. 5–29.
c Cambridge University Press 2014 5

doi:10.1017/jfm.2014.447
Fluid flow analysis of a shark-inspired

microstructure
Samuel Martin1 and Bharat Bhushan1, †

1 Nanoprobe Laboratory for Bio & Nanotechnology and Biomimetics (NLBB),
The Ohio State University, 201 West 19th Avenue, Columbus, OH 43210-1142, USA
(Received 6 March 2014; revised 5 June 2014; accepted 25 July 2014;

first published online 1 September 2014)
The scales of fast-swimming sharks contain riblet structures with microgrooves,

aligned in the direction of fluid flow, that result in water moving efficiently over
the surface. In previous studies, these riblet structures have shown a drag reduction
of up to 10 % when compared with a smooth, flat surface. These studies have
suggested two prevalent drag-reduction mechanisms which involve the effect of
vortices and turbulence fluctuations. To further explore relevant mechanisms and
study the effect of riblet geometry and flow properties on drag, vortices and
turbulence fluctuations, various shark-skin-inspired riblet structures were created using
computational models in which velocity, viscosity, spacing, height and thickness
parameters were independently modified. A relevant mechanism of drag reduction is
discussed to relate riblet parameters and flow properties to drag change and vortex
size. Modelling information will lead to a better understanding of riblets and allow
for optimum drag-reducing designs for applications in marine, medical and industrial
fields.
Key words: drag reduction, flow control
https://doi.org/10.1017/jfm.2014.447 Published online by Cambridge University Press
1. Introduction
Nature has produced structures and materials that are efficient and multipurpose
using commonly occurring materials in order to survive. From these resourceful
designs, researchers have created bioinspired designs with the goals of replicating
and expanding their functionality in order to solve practical problems (Bhushan 2009,
2012). There are many aspects of nature that can serve as an inspiration for design.
For example, superhydrophobicity can be found from the leaves of Lotus flowers
(Bhushan 2009), reversible adhesion from the toes of a gecko (Autumn et al. 2000;
Gao et al. 2005; Autumn et al. 2006; Bhushan 2010), antireflectivity from the eyes
of moths (Genzer & Efimenko 2006) and self-cleaning from rice leaves and butterfly
wings (Bixler & Bhushan 2012b).
A model surface inspired by nature for drag reduction is provided by shark skin.
Fast-swimming sharks can quickly move through water with a low energy input. As
shown in figure 1, sharks are covered in scales called dermal denticles (little skin
teeth) and have riblets with microscopic grooves aligned parallel to fluid flow. The
† Email address for correspondence: bhushan.2@osu.edu

6 S. Martin and B. Bhushan
Actual shark skin (Mako, Isurus oxyrinchus)
Riblet
Dermal denticle
F IGURE 1. (Colour online) Scanning electron microscopy (SEM) micrographs of the skin
of a fast-swimming shark from living nature shown at two magnifications (adapted from
Bixler & Bhushan 2012b). These images highlight the scales (dermal denticles) and riblets
which are responsible for drag reduction. Arrows indicate fluid flow direction.
scales are generally 0.2–0.5 mm in size with the grooves spaced 30–100 µm apart
(Reif 1985; Bechert, Bruse & Hage 2000a,b; Bhushan 2012). Shark-skin-inspired
riblets have been studied by various researchers both experimentally (Walsh 1982;
Rohr et al. 1992; Bechert, Bruse & Hage 1997b; Lee & Lee 2001; Jung & Bhushan
2009; Dean & Bhushan 2010; Bixler & Bhushan 2012a, 2013) and computationally
(Choi, Moin & Kim 1993; Chu & Karniadakis 1993; Goldstein, Handler & Sirovich
1995). In these studies, various riblet geometries have been examined including blade,
scalloped and sawtooth designs, and riblet configurations such as continuous and
segmented.
Experiments to study the drag benefits of riblets have been carried out using the
so-called Berlin oil channel which contained two test plates, one with riblets and one
without, inserted into a flow channel (Bechert et al. 1992). Force measurements on
these plates were compared, and the percentage drag reduction on the riblet surface
was reported using 1τ /τ0 where 1τ is the difference in shear stress between the riblet
and smooth surfaces and τ0 is the shear stress for the smooth surface. The surface area
used to calculate the shear stress was based on the smooth surface area. The riblet
spacing from this experiment was non-dimensionalized using
s+ = svτ /ν. (1.1)
Non-dimensional values are denoted by a superscript +. In this equation, the

distance between microgrooves is the riblet spacing where s is the dimensional riblet
spacing, vτ = (τ0 /ρ)1/2 where vτ is the wall shear-stress velocity, ν is the kinematic
viscosity and ρ is the density. Instead of riblet spacing, García-Mayoral & Jiménez
(2011) used the riblet cross-sectional area for non-dimensionalization.
Figure 2 shows the drag-reduction data for the blade, scalloped and sawtooth riblets
in oil flow. The data shows a drag reduction of ∼10 % for blade, ∼6.5 % for scalloped
and ∼5 % for sawtooth riblets (Bechert et al. 1997b). Although blade riblet geometries
provide the greatest degree of drag reduction, due to the weak structural design of
blade riblets, commercial application would be difficult; therefore, other more durable
designs such as sawtooth or scalloped geometries may be desirable.
Researchers have not universally agreed upon the mechanism of drag reduction. Two
main mechanisms for drag reduction that have been proposed include controlling the
Fluid flow analysis of a shark-inspired microstructure 7
4 Sawtooth
Scalloped
2 Higher
Blade
drag
0
Lower
–2
drag
–4
–6
–8
–10
5 10 15 20 25 30 35
F IGURE 2. Drag-reduction curves for three continuous riblet geometries for various s+
spacing (adapted from Bechert et al. 1997b). Blade riblets provide the greatest drag
reduction followed by scalloped and sawtooth riblets. Drag reduction was calculated by
measuring the force on two test plates, one with riblets and one without, and reported as
1τ /τ0 where 1τ is the difference in shear stress between the riblet and smooth surfaces
and τ0 is the shear stress for the smooth surface. For this study, optimal spacing is at
s+ ≈ 16 where s+ is a non-dimensional spacing parameter defined by s+ = svτ /ν where s
is the dimensional riblet spacing, vτ = (τ0 /ρ)1/2 where vτ is the wall shear-stress velocity
and ν is the kinematic viscosity. Riblet geometries are commonly non-dimensionalized in
order to compare their drag change.
vortices in turbulent flow and reducing the turbulence fluctuations in the cross-flow
direction (table 1).
In the first mechanism stated in table 1, riblets control the streamwise vortices
that form in turbulent flow. These vortices rotate along the axis of the mean velocity
direction generating ejecting and sweeping motions in which high-velocity fluid mixes
with the low-velocity fluid near the surface. Consequently, this motion creates zones
of high shear stress and drag. By choosing an appropriately sized riblet spacing based
on the vortex size inherent to the flow, these vortices can be lifted away from the
surface, decreasing the area exposed to high shear stress and therefore reduce drag
(Kim, Moin & Moser 1987; Choi et al. 1993; Lee & Lee 2001).
In the second mechanism stated in table 1, riblets decrease the fluctuating velocity
component in the cross-flow direction. In turbulent flow, the flow is unsteady, and
the fluid velocity oscillates around the mean flow velocity. By channelling the flow
using riblets, this oscillation is reduced; therefore, the momentum transfer close to
the surface and overall drag is reduced. An effective protrusion height parameter has
been suggested to measure the riblet geometry effectiveness in reducing drag. This
parameter is the difference between the streamwise and cross-flow protrusion heights
shown in table 1 (Bechert et al. 1986; Bechert & Bartenwerfer 1989). The effective
protrusion height parameter (1h) increases with increased spacing (s) because 1h is
a constant fraction of the riblet spacing. With a larger spacing, the turbulent cross-
flow and momentum exchange near the wall decreases and therefore the cross-flow
protrusion height decreases. With riblets with small spacing only interacting in the
viscous regime, a linear relationship is expected due to the law of the wall, which
is the relationship between the velocity at the wall and the distance from the wall in
non-dimensional units (Spalding 1961) and so the initial slope of the drag reduction
curve is proportional to (1h/s)s+ (Bechert et al. 1997b).
Image Observations Questions

Vortex Riblets with appropriately sized spacing reduce viscous Previous studies varied velocity or
drag by lifting and pinning the streamwise vortices away non-dimensionalized the riblet parameters, which
from the surface and decreasing the mixing of high- and does not fully explain the effect of individually
low-speed fluid. varying riblet parameters such as spacing (s+ ;
However, with a small spacing, vortices contact many peak-to-peak distance), height (h+ ;
riblet peaks leading to many high-shear-stress zones; with peak-to-valley distance), or thickness (t+ ; peak
a large spacing, vortices can fall into the riblet valleys, width) on drag and vortex size.
exposing the surface to high-speed fluid
Two riblets
Flow visualization of a vortex (Choi et al. 1993).
interacting with a riblet surface using
smoke in air from atomized oil Previous computational work has suggested average vortex
(adapted from Lee & Lee, 2001). diameters of ∼30 wall units (s+ ) and located
at ∼20 wall units from the surface (Kim et al. 1987) and
experimental work has shown average vortex diameters of
30–40 wall units (Lee & Lee 2001).
Mean velocity profiles Riblets channel the flow in the mean x-direction by Previous computational work (Choi et al. 1993;
Cross- reducing the fluctuating turbulent cross-flow velocity Goldstein et al. 1995) and experimental work
Streamwise flow component in the z-direction. This obstruction reduces the (Weiss 1993; Suzuki & Kasagi 1994) have
turbulent momentum transfer close to the surface leading shown decreased cross-flow fluctuations;
hpc to a drag reduction. The effective protrusion height however, various riblet parameters have not
S. Martin and B. Bhushan
y parameter (1h) is a measure of how effectively the riblets been modified to see the change in turbulent
h ps x z reduce drag and is the difference between the streamwise cross-flow velocity component and compared
(hps ) and cross-flow protrusion heights (hps ) (Bechert et al. with the overall drag change.
Apparent flow origins
1986; Bechert & Bartenwerfer 1989).
Schematic of the mean velocity
profiles for flow in the streamwise and At low s+ , the drag reduction was found to be
cross-flow dirctions proportional to (1h/s)s+ (Bechert et al. 1997b).
(adapted from Bechert et al., 1997b)
TABLE 1. Two ideas for the riblet drag-reduction mechanism. The first idea shows an image of a lifted vortex created by atomizing oil
burned in air to produce smoke. With appropriate sized riblet spacing, vortices are lifted away from the surface decreasing the ejecting and
sweeping motions. The second idea shows an image of the effective protrusion height parameter which is a calculation to show the efficiency
of the riblet geometry in channelling the fluid and reducing the cross-flow turbulence fluctuations.
Researchers do not fully understand the mechanism of drag reduction and have
primarily suggested these mechanisms as possible explanations. Furthermore, from the
definition of s+ , the drag reduction should be the same for constant non-dimensional
parameters even if the spacing, velocity or viscosity terms are changed. This concept
has not been explored, and limited drag data can be compared to answer whether this
definition holds. In addition, it is not understood how the effect of riblet geometry
and flow properties impacts on drag, vortex size and turbulence fluctuations. The size
of vortices has been measured for limited geometries and Reynolds numbers, and
turbulence fluctuations have not been compared with drag change for various riblet
geometry.
By investigating and answering these concerns, riblets can become more widely
implemented in commercial applications as a method of drag reduction. For example,
riblet technology has applications in competitive swimsuits; air, water and oil pipe
flow; and transportation. Currently, riblet technology has been employed in competitive
swimwear such as the Speedo Fastskinr swimsuit to reportedly decrease drag by up
to 4 % (Krieger 2004). Applying riblets to the inside of pipes would decrease drag
(Bechert, Bruse & Hage 1997a; Weiss 1997) and reduce fouling in oil and water
flow (Bixler & Bhushan 2012a,b). In oil pipeline flow, sludge accumulation causes
high drag and requires higher pumping energy. In ships, biofouling in which aquatic
growth accumulates on the hull also increases drag. Riblets could be applied to the
hull of ships to reduce fouling and create a drag reduction. To investigate reducing
skin friction drag on aircraft using riblet structures, studies at the National Aeronautics
and Space Administration’s Langley Research Center (Walsh & Anders 1989) and
the German Aerospace Center (Bechert et al. 1997a) have been carried out. Riblets
by covering approximately 70 % of the hull of an aircraft can provide a 3 % drag
reduction resulting in desirable fuel savings (Bechert et al. 1997a).
In this paper, a computational fluid dynamic (CFD) model of blade riblet structures
was created in order to study drag, vortex formations and turbulence fluctuations.
These insights will help determine which of the two proposed riblet mechanisms is
relevant and help to develop optimal riblet designs. In addition, the definition of riblet
spacing can be evaluated to determine whether it is a valid non-dimensional parameter
in determining drag. To answer these questions, two fluid parameters (velocity and
viscosity) and three riblet parameters (spacing, height and thickness) for blade riblets
were independently changed.
2. Modelling
In this section, the CFD program and turbulence model will be presented first. Next,
a validation model of the flow around a square cylinder will be shown. Lastly, the
shark-skin inspired riblet models are presented along with the methods to analyse the
flow results.
2.1. CFD program

r
ANSYS FLUENT 14.0 was used as the CFD software, and the large eddy simulation
(LES) model with the Smagorinsky–Lilly subgrid scale was used to analyse the flow.
LES resolves the large-scale flow eddies and models the small-scale flow eddies which
results in greater accuracy over a Reynolds-averaged Navier–Stokes (RANS) model.
y
(a) Rectangular channel
Flow inlet Flow outlet
V
D Square cylinder
14D x
4.5D D 15D
(b)
F IGURE 3. Side view of the flow in a rectangular channel with a square cylinder for
LES validation showing geometry (a) and mesh (b). Computational mesh of the geometry
is 220 × 150 × 20.
2.2. Validation model: flow around a square cylinder

In order to validate the performance of this model, turbulent flow over a square
cylinder with edge length D = 0.04 m was performed. This simulation is a

well-documented experimental (Vickery 1966; Lee 1975; Cheng, Lu & Chen 1992;
McLean & Gartshore 1992; Lyn & Rodi 1994; Lyn et al. 1995) and computational
test case for the LES model (Murakami & Mochida 1995; Rodi 1997; Rodi et al.
1997; Ochoa & Fueyo 2004). The dimensions of the domain were normalized by
this edge length as shown in figure 3. The x, y and z axes denote the streamwise,
wall-normal and spanwise or cross-flow directions, respectively, and V denotes the
average velocity in the streamwise direction.
The inlet boundary condition was specified as a uniform velocity of 0.535 m s−1
with no added perturbations which corresponded to a Reynolds number of 21 400
in order to compare with the published data. The top and bottom surfaces were
prescribed as no-slip wall boundary conditions and the left and right surfaces
were prescribed as periodic boundary conditions. The computational mesh was
220 × 150 × 20 along the x, y, and z directions, respectively. The mesh was refined
in the region near the square cylinder as well as behind the cylinder to resolve the
flow more accurately.
Reference C̄L CLrms C̄D CDrms St

Numerical data – from Ochoa & Fueyo (2004)
Verstappen and Veldman 0.005 1.45 2.09 0.178 0.133
Porquie et al.
— −0.02 1.01 2.2 0.14 0.13
— −0.04 1.12 2.3 0.14 0.13
— −0.05 1.02 2.23 0.13 0.13
Murakami & Mochida (1995) −0.05 1.39 2.05 0.12 0.131
Wang and Vanka 0.04 1.29 2.03 0.18 0.13
Nozawa and Tamura 0.0093 1.39 2.62 0.23 0.131
Kawashima and Kawamura
— 0.01 1.26 2.72 0.28 0.16
— 0.009 1.38 2.78 0.28 0.161
Ochoa & Fueyo (2004) 0.03 1.4 2.01 0.22 0.139
Experimental data
Vickery (1966) — 0.68–1.32 2.05 — —
Lee (1975) — — 2.05 0.16–0.23 —
Cheng et al. (1992) — 0.1–0.6 1.9–2.1 0.1–0.2 —
McLean & Gartshore (1992) — 0.7–1.4 1.9–2.1 0.1–0.2 —
Lyn & Rodi (1994), Lyn et al. (1995) — — 2.1 — 0.132
This work (D = 0.04 m,
V = 0.535 m s−1 , Re = 21 400) −0.04 1.41 2.28 0.16 0.130
TABLE 2. Comparison among time-averaged square cylinder data.
A time step of 2.5 × 10−3 s was used for 10 000 time steps. This time step
was chosen to resolve the flow using the Courant condition which allows for the
calculations to capture the time scales at the cell residence time. The data was
averaged over the last 7000 time steps which was approximately 11 times the
residence time. From published experimental and numerical results, the mean lift
coefficient, C̄L , lift oscillation amplitude, CLrms , mean drag coefficient, C̄D , drag
oscillation amplitude, CDrms , and Strouhal number, St, were compared with the results
of the test case. It can be seen in table 2 that the results for this work agree and so
this model was validated for use in modelling riblet structures.
2.3. Shark-skin inspired riblet models

The cases that were analysed are presented in table 3. The variables that were
independently changed included two fluid parameters (velocity V and viscosity ν)
and three riblet parameters (spacing s, height h and thickness t). In the table, the
dimensional values are presented on the left, and the resulting non-dimensional
values are presented on the right. The values were non-dimensionalized using the
wall shear-stress velocity. The non-dimensional riblet spacing parameter has been
given in (1.1). Similarly, the height (h) and thickness (t) were non-dimensionalized
as h+ = hvτ /ν and t+ = tvτ /ν, respectively.
In order to run a computer model, the dimensional parameters V, ν, s, h and t
needed to be chosen. Because the Reynolds number and non-dimensional riblet
parameters s+ , h+ and t+ are the variables controlling the flow phenomenon, linearity
between V and vτ must be shown to satisfy (1.1). Using the wall shear-stress velocity
12
s (mm) h (mm) t (µm) ν/Vl (µm) s+ h+ t+ Number of riblets
Varying Vl Varying ν
ν = 10−6 m2 s−1 Vl = 0.42 m s−1
Baseline 1.00 0.50 20.0 2.38 2.38 16.2 8.1 0.32 20
Effect of V 4.00 10.3 5.1 0.21 20
2.38 16.2 8.1 0.32 20
1.39 24.5 12.3 0.49 20
0.98 32.4 16.2 0.65 20
0.76 46.8 23.4 0.94 20
Effect of ν 4.00 10.7 5.4 0.21 20
2.38 16.2 8.1 0.32 20
1.39 27.4 13.7 0.55 20
0.77 46.7 23.3 0.93 20
Effect of s 0.50 8.4 8.4 0.34 40
1.00 16.2 8.1 0.32 20
1.50 25.3 8.4 0.34 13
2.50 41.1 8.2 0.33 8
Effect of h 0.25 16.2 4.1 0.32 20
S. Martin and B. Bhushan
0.50 16.2 8.1 0.32 20

0.75 16.6 12.5 0.33 20
0.75 0.22 15.0 1.20 22.9 6.7 0.46 27
0.75 0.32 15.0 1.20 23.2 9.9 0.46 27
0.75 0.42 15.0 1.20 23.2 13.0 0.46 27
Effect of t 5.0 16.8 8.4 0.08 20
10.0 17.0 8.5 0.17 20
20.0 16.2 8.1 0.32 20
30.0 16.6 8.3 0.50 20
TABLE 3. Test cases for the simulations of turbulent flow over riblets.
0.05
0.04
0.03
0.02
0.01
0 0.5 1.0 1.5
F IGURE 4. Wall shear-stress velocity compared with velocity showing a linear relationship
which allows for changing velocity as a means to modify the wall shear-stress velocity.
and velocity from the test cases in table 3, figure 4 shows wall shear-stress velocity
compared with velocity. Because the relationship was approximately linear, changing
the velocity will linearly change the wall shear-stress velocity, and (1.1) could be
used to non-dimensionalize the riblet parameters. Therefore, the data trends presented
in this paper are independent of the dimensions selected for the models.
For the test cases with varying velocity or viscosity, either the viscosity or the
velocity was held constant and was listed in the table as ν/Vl . For the effect of
velocity cases, the viscosity was held constant at ν = 10−6 m2 s−1 ; for the effect of
viscosity cases, the velocity was held constant at Vl = 0.42 m s−1 . The baseline values
shown in bold numbers in the table were chosen for blank cells.
The values of the dimensional parameters were chosen to fulfil several requirements.
First, they were chosen over a range, such that when they were non-dimensionalized,
the cases would cover a wide non-dimensionalized range for the parameter. Second,
the values were selected so that the non-dimensional values could be compared
with the other cases. By having similar non-dimensional values that were created by
modifying different dimensional parameters, the definition of non-dimensionalizing
could be explored. Lastly, the dimensional values were chosen so that s+ of 15–20
would be in the centre of the parameter’s ranges. In other studies, this range has
been shown for optimal drag reduction. By centring the data on this value, it would
allow for test cases on either side of this range to be investigated.
Two sets of riblet height test cases were created to further explore the relationship
between riblet height and drag. In the first set, only the height was modified so the
other parameters were kept the same as the baseline case. In the second set, all of the
parameters were modified to increase the Reynolds number, centre s+ near 23 and h+
near 10.
2.3.1. Model set-up

Figure 5 shows that the models were created with the riblet surface on the bottom
wall and the smooth surface on the top wall. These walls had the no-slip boundary
condition. With both surfaces in the model, the drag change for the riblet surface
(a) Smooth top

surface
Riblet
surface
Flow V
t
h
s
W L
y
x
z
(b)
F IGURE 5. (a) Computational domain of the riblet geometry for the baseline case showing
geometry and the riblet dimensions. Twenty riblets are on the bottom surface and a smooth
surface is on the top surface for drag comparison. (b) Computational mesh near riblets.
can be directly compared with the smooth surface. The x, y, and z axes denote the
streamwise, wall-normal and spanwise or cross-flow directions, respectively.
Along the streamwise direction, a constant instantaneous volume flux boundary

condition was applied,
2
Z
Q= V dAc = Ac Vl (2.1)
Ac 3
where Ac = HW is the cross-sectional area, V is the average velocity in the streamwise
direction and Vl is the centreline velocity of a laminar parabolic profile with the same
volume flux. The Reynolds number was calculated with δ as the channel half-width
(δ = H/2) using
Re = Vl δ/ν. (2.2)
The inlet and outlet surfaces and the left and right surfaces were prescribed as
periodic so that wall effects would not interfere with the results. The solution was
initialized with a laminar velocity profile and afterwards turbulence fluctuations were
added. The solution was advanced using a computational time step of 1TVl /δ = 0.05
where 1T is the dimensional time step until natural turbulence was developed by
monitoring wall-shear stresses for quasi-periodic behaviour. After a statistically steady
flow field was reached, the simulations were continued for additional time to time
average the results.
The computational domain dimensions were chosen to be large enough to include
the expected scales of the flow structures and to be greater than the minimal flow
unit (100 wall units in the spanwise direction and 300 wall units in the streamwise
direction where a wall unit is non-dimensionalized as in (1.1) using a length scale
such as the spanwise or streamwise distance, respectively) as determined by Jimenez
& Moin (1991). For the models, the height H and width W remained constant at
0.02 m which corresponded to 206 wall units in the smallest case. The length L
changed for the cases to keep a grid spacing at x+ ≈ 35 and with 16 nodes this
corresponds to approximately 560 wall units in length. In the spanwise direction, the
z+ grid spacing varied between 0.4 and 2.0 with the resolution z+ ≈ 0.7 for most cases.
In the wall-normal direction, there were 128 grid points with the first grid point on
the smooth surface baseline case at y+ = 0.16. Furthermore, it was kept below y+ = 1
for all cases as required by the LES model, and the mesh was biased towards the
surfaces to capture the flow field near the walls more accurately. The grid resolution
was kept the same on the smooth and riblet surfaces for drag comparison purposes.
In a study by Choi et al. (1993), the number of grid points in the streamwise and
spanwise direction was increased from 16 × 129 × 32 to 32 × 129 × 64 to investigate
the influence of grid resolution on turbulence statistics. Doubling the grid points in the
streamwise and spanwise directions changed the maximum turbulence intensities and
Reynolds shear stresses by less than 2 %. The grid resolution in the streamwise and
spanwise direction was approximately the same as Choi et al. (1993) and finer than
Kim et al. (1987). Also, the grid resolution was much finer than the recommended
requirements of x+ = 40 and z+ = 20 suggested by FLUENT (2014).
To determine the error bounds using the grid resolution and time averaging length
for these models, the drag on two smooth walls in a plane channel flow were
compared by time averaging over 500 non-dimensional time steps (T + ) where
T + = T Vl /δ. A solution with enough time averaging should yield identical drag
on the surfaces. After averaging the drag over T + = 500, the drag on the plane walls
differed by ±2 % of each other. This percentage was in agreement with the drag
variations in Walsh (1982) and Choi et al. (1993).
2.3.2. Drag, vortex and turbulence fluctuation measurement

The instantaneous skin-friction drag for the smooth surface, Ds , and riblet surface,
Dr , were calculated with
Z
∂u

Ds = µ dAs 

∂n 
ZAs (2.3)
∂u
Dr = µ dAr 


Ar ∂n
where n is the coordinate normal to the surface, and As and Ar are the surface areas
for the smooth and riblet surfaces, respectively. Figure 6 shows that the instantaneous
drag for the riblet and smooth surfaces was quasi-periodic. To calculate the time-
averaged drag, the drag on each surface was averaged over a time period of at least
T + = 500. The percentage drag change was calculated by comparing the time-averaged
drag for these two surfaces as (D̄r − D̄s )/D̄s .
At time intervals of T + = 50, four equally spaced y–z planes with vorticity contours
and tangential velocity vectors were plotted. These images were used to measure the
5
Riblet surface Baseline case
Smooth surface
4
0 100 200 300 400 500
F IGURE 6. (Colour online) Time history of drag on the riblet and smooth surfaces. Due
to turbulence, the drag on each surface changes with time. The simulation was run for at
least 500 non-dimensionalized time units (T + ) to average the drag on each surface. This
figure shows an 8 % drag reduction on the riblet surface.
height, width and vortex centre distance from the wall for vortices in the buffer layer
(5–30 non-dimensionalized wall units from the wall) on the riblet and smooth surfaces,
as illustrated in figure 7. For each case over a time of T + = 500, this resulted in 44
images to analyse. The vortex size data taken from these images was averaged for
each case and will be presented to explain a possible riblet mechanism.
The definition of a vortex by Robinson (1990) states that ‘A vortex exists when
instantaneous streamlines mapped onto a plane normal to the vortex core exhibit
a roughly circular or spiral pattern, when viewed from a reference frame moving
with the center of the vortex core’. This definition was used to measure the
quasi-streamwise vortices located in the buffer layer (Robinson 1990) with the vortex
boundary defined as the edges at which the spinning motion of the fluid shown by the
velocity vectors began to break down. Vorticity plots were not used by themselves to
determine vortices because strong vorticity is not entirely indicative of actual vortices
especially near the walls (Robinson, Kline & Spalart 1989). For example, at the riblet
peaks in figure 7, strong vorticity was seen, but the velocity vectors do not suggest
a vortex at the peaks.
To measure the turbulence fluctuation in the cross-flow direction, points in the
model were created. These points were located at the centre of the riblet valleys (half
the riblet height and half the riblet spacing) and at the corresponding locations on the
smooth surface. The root-mean-square (r.m.s.) of the cross-flow turbulence fluctuation
was recorded at these points. The reduction in cross-flow turbulence was defined as
(w0s − w0r )/w0s , where w0s and w0r are the r.m.s. cross-flow turbulence fluctuations for the
smooth and riblet surfaces, respectively. This data was used to view the relationship
between riblet geometry and the reduction in turbulence fluctuations as explained in
the second riblet mechanism.
3. Results and discussion

In this section, the drag, vortex size and turbulence fluctuations data will be
presented to evaluate the performance of these riblets. This data was used to compare
(a) W
(b)
Width
Height
Distance
from wall
–100
100
Streamwise
vorticity (s–1)
F IGURE 7. Contour plot of streamwise vorticity and tangential velocity vectors on an y–z
plane showing the effects of turbulence. The height, width and distance from a wall for
an example vortex are presented. The parameters used in the image are s = 1.00 mm,
h = 0.50 mm, t = 20.0 µm, ν/Vl = 2.38 µm and Re = 4180.
the two possible mechanisms of drag reduction and present how riblet parameters
affect drag change and vortex size.
3.1. Drag change

In figure 8, the drag change for the various models was plotted with negative values
denoting a drag-reduction benefit. In figure 8(a) with varying velocity and viscosity,
both curves have similar trends with a maximum drag reduction at s+ = 18–20 of 9 %
and 12 %. Independently, the velocity and viscosity terms yielded approximately the
same drag reductions. Because of this similarity, the non-dimensional equation was
validated. The non-dimensional parameters will give the same drag change regardless
of which parameters in the non-dimensional definition are modified.
(a) 15 (b) 15
Varying V
10 Varying 10
Drag change (%)
5 5
0 0
–5 –5
Baseline
–10 –10
Baseline
–15 –15
0 10 20 30 40 50 0 10 20 30 40 50
(c) 15 (d ) 15
10 10
Drag change (%)
5 5
0 0
–5 –5
Baseline
–10 –10
Baseline
–15 –15
0 5 10 15 0 0.1 0.2 0.3 0.4 0.5 0.6
F IGURE 8. Drag change for the various models plotted against their non-dimensionalized
parameter. The data used in the analyses is for: (a) varying V or ν s = 1.00 mm, h =
0.50 mm, t = 20.0 µm, Re = 2500–13 000; (b) varying s h = 0.50 mm, t = 20.0 µm,
ν/Vl = 2.38 µm, Re = 4180; (c) varying h downward-pointing triangle s = 1.00 mm,
t = 20.0 µm, ν/Vl = 2.38 µm, Re = 4180; asterisk s = 0.75 mm, t = 15.0 µm, ν/Vl =
1.20 µm, Re = 8360; (d) varying t s = 1.00 mm, h = 0.50 mm, ν/Vl = 2.38 µm, Re = 4180.
In the models with varying velocity, viscosity or spacing (a,b), the curves agree to a
first order. At higher s values and constant h values, the h/s ratio becomes small leading
to a slightly different curve (b). From the models with varying velocity or viscosity (a),
optimal spacing was found to be at s+ ≈ 18–20. From the models with varying spacing
(b), optimal spacing was found to be at s+ ≈ 23. From the models with varying height (c),
optimal height was found to be at h+ ≈ 8–10. The s+ and h+ terms affect drag significantly,
whereas this range of t+ did not affect drag notably as shown in (d).
In figure 8(b) with varying spacing, the drag reduction minimum of 11 % was
located at s+ = 23. Even though the spacing was modified similarly to the models
with varying velocity or viscosity, the spacing curve did not match the velocity or
viscosity curve. This discrepancy was due to the riblet height and thickness not being
affected by modifying the spacing term. Changing velocity or viscosity resulted in
changing non-dimensional spacing, height and thickness; however, changing spacing
only resulted in changing non-dimensional spacing. In the models with varying
spacing, the height remained constant, and a greater non-dimensional riblet spacing
was required for the riblets to reach a drag increasing state. Due to the differences
between the velocity and viscosity curves and the spacing curve, it was apparent that
both riblet spacing and height played a large role in determining drag change.
Furthermore, in figure 8(c) with varying height and constant spacing, there was a
dramatic drag change between the baseline case and the taller riblet case for s+ ≈ 16.
In the baseline case, there was a drag reduction of nearly 10 % with h+ = 8.1, but
in the taller riblet case, there was a drag increase of 12 % with h+ = 12.5. Similarly
with s+ ≈ 23, there was a drag reduction of 8.2 % at h+ = 9.9 with less of a drag
reduction on either side of this case. A height in the region of h+ = 8–10 optimally
reduced drag for both cases even with different Reynolds numbers and riblet spacings;
however, increasing outside of this range quickly led to a drag increasing state and
decreasing outside this range led to less efficient state.
When the riblet thickness was modified, little drag change occurred (figure 8d).
The thicknesses ranged from t+ = 0.08–0.50, and the drag reduction was found to
be roughly constant at approximately 9 %. There were slight oscillations around this
value, but because there was not a strong positive or negative relationship, the trend
was reported as constant over this range of thicknesses. At much larger thicknesses,
the riblet valley width would decrease and the surface area at the riblet peaks would
increase. It is expected that these changes would generate a weaker drag reduction,
and eventually a drag increase.
3.2. Vortex height and width

Using the y–z planes with vorticity contours and tangential velocity vectors, the height
and width of the streamwise vortices were measured and averaged. In figure 9, the
average vortex height and width on the smooth and riblet surface were plotted. As
the velocity was increased in figure 9(a), the height and width of the vortices became
larger on both surfaces. In figure 9(b–d) for varying spacing, height and thickness,
the vortex size was fairly constant. In these figures, the smooth surface vortex width
averaged 37 wall units; the riblet surface vortex width averaged 32 wall units; the
smooth surface vortex height averaged 23 wall units; and the riblet surface vortex
height averaged 19 wall units.
The size of the vortices was related to the Reynolds number based on the channel
height. In the models with varying velocity or viscosity, the Reynolds number was
changing, and so the vortex height and width changed. Because the varying spacing,
height and thickness cases had the same Reynolds number of 4180, the vortex size
data was approximately the same for each of these cases.

On both surfaces, the vortex width was larger than the vortex height. For the vortex
height and width, the smooth surface had larger sizes than the riblet surface. Because
the sides of the model used periodic conditions, the vortices could grow in the width
direction; however, the riblet and smooth surfaces impeded the growth in the height
direction. The vortex sizes on the riblet surface were smaller than the smooth surface
because the riblets protruded sharply into the flow obstructing vortex rotation and
growth.
The vortex sizes from this research are in good agreement with the vortex sizes
suggested by Kim et al. (1987) and Lee & Lee (2001) for comparable Reynolds
numbers. In Kim et al. (1987), it was suggested that the streamwise vortices were 30
wall units in diameter at a Reynolds number of 3300. The work by Lee and Lee was
done at Reynolds numbers 2340 and 4950 and showed vortex diameters of 30–40
wall units.
The drag change and vortex size data were compared to determine when the drag
had a lowest value. The optimal drag reduction was located at s+ = 18–20 and s+ = 23
from figure 8(a,b), respectively. By comparing these s+ values with the riblet surface
(a) (b)
60 60
Smooth surface vortex width
Riblet surface vortex width
50 Smooth surface vortex height 50
Riblet surface vortex height
40 40
30 30
20 20
10 10
0 10 20 30 40 50 0 10 20 30 40 50
(c) (d)
60 60
50 50
40 40
30 30
20 20
10 10
0 5 10 15 0 0.2 0.4 0.6
F IGURE 9. Vortex height and width on the riblet and smooth surfaces for the various
models. The data used in the analyses is for: (a) varying V s = 1.00 mm, h = 0.50 mm,
t = 20.0 µm, Re = 2500–13 000; (b) varying s h = 0.50 mm, t = 20.0 µm, ν/Vl = 2.38 µm,
Re = 4180; (c) varying h s = 1.00 mm, t = 20.0 µm, ν/Vl = 2.38 µm, Re = 4180;
(d) varying t s = 1.00 mm, h = 0.50 mm, ν/Vl = 2.38 µm, Re = 4180. With increasing
velocity (a), Reynolds number based on channel height and vortex size increases. In the
models with varying s, h or t (b)–(d), vortex size remains approximately constant. Average
vortices have a greater width than height, and vortices on the smooth surface are larger
than on the riblet surface. Riblets pin the vortex keeping them from growing and becoming
the same size as the vortices on the smooth surface.
vortex width on figure 9(a,b), it was found that the average riblet surface vortex width
was approximately 32 non-dimensional units. This corresponded to a vortex-width-to-
spacing ratio of approximately 1.5. As the spacing and riblet surface vortex width
approached the same value, the drag became larger and eventually became a drag-
increasing case. It is thought that this ratio is effective in lifting away the vortices
from the surface and minimizing the exchange of high-velocity and low-velocity fluid.
As the size of the vortices and spacing approached the same value, there was a greater
possibility of a vortex fitting within a riblet valley and increasing drag.
The average vortex centre distance from the wall was plotted for the models with
varying spacing in figure 10. As s+ increased, the vortices were lifted less on the riblet
30
)
Riblet surface
Vortex center distance from wall (

Smooth surface
25
20
15
0 10 20 30 40 50
F IGURE 10. Vortex centre distance from the wall on the riblet and smooth surfaces for
the models with varying spacing. The data used in the analyses is for h = 0.50 mm, t =
20.0 µm, ν/Vl = 2.38 µm and Re = 4180. As s+ increases, vortex centre distance from
the wall decreases on the riblet surface and stays approximately constant on the smooth
surface.
surface and stayed at y+ ≈ 20 on the smooth surface. At small spacing, riblets lift up
the vortices, but have a much larger surface area. At large spacing, riblets do not lift
up the vortices. When the spacing was s+ ≈ 20, vortices were lifted from the riblet
wall without dramatically increasing the surface area and a greater drag reduction was
observed.
3.3. Turbulence fluctuations

The reduction in cross-flow turbulence fluctuations was plotted in figure 11. In
the models with varying s+ (figure 11a,b), the largest reduction was at small
non-dimensional spacing. As the spacing increased, more cross-flow turbulence was
observed in the riblet valleys. At larger spacings, the fluid was further from the no-slip
condition at the wall and was more affected by vortices in the flow; therefore, for
larger riblet spacings, it was easier for the fluid to travel in the cross-flow direction.
In figure 11(c) with the models with varying h+ , the taller riblets heights created
lower cross-flow turbulence fluctuations. As the riblets became taller, the velocity at
the riblet centre was less affected by the fluid above the riblets. The flow in tall
riblets was less perturbed and resulted in more streamwise flow. In figure 11(d) with
the models with varying t+ , modifying the riblet thickness minimally affected the
turbulence fluctuations. Because the valley spacing was nearly the same in these cases,
there was little difference in the turbulence fluctuations.
The reduction in turbulence fluctuations and the drag change trends did not match.
According to the second possible mechanism, these plots were expected to match
with a maximum drag reduction corresponding to a maximum reduction in turbulence
fluctuations. For example, in the models with varying s+ , there was not an inflection in
the reduction in turbulence fluctuations. In the models with varying riblet height, there
was the same reduction in turbulence fluctuations for the baseline case and taller riblet
case; however, one of these cases was a drag reduction while the other was a drag
(a) (b)
100 100
Varying V
80 Baseline Varying 80 Baseline
60 60
40 40
20 20
0 10 20 30 40 50 0 10 20 30 40 50
(c) (d)
100 100
Baseline
80 80 Baseline
60 60
40 40
20 20
0 5 10 15 0 0.2 0.4 0.6
F IGURE 11. Cross-flow turbulence fluctuations for the various models. The data used
in the analyses is for: (a) varying V or ν s = 1.00 mm, h = 0.50 mm, t = 20.0 µm,
Re = 2500–13 000; (b) varying s h = 0.50 mm, t = 20.0 µm, ν/Vl = 2.38 µm, Re = 4180;
(c) varying h downward-pointing triangle s = 1.00 mm, t = 20.0 µm, ν/Vl = 2.38 µm,
Re = 4180; asterisk s = 0.75 mm, t = 15.0 µm, ν/Vl = 1.20 µm, Re = 8360; (d) varying
t s = 1.00 mm, h = 0.50 mm, ν/Vl = 2.38 µm, Re = 4180. A maximum in the turbulence
fluctuations reduction should appear at the optimal drag reduction; however, the figures do
not match this trend. Because this data does not match the drag change data, this method
is not believed to be the drag reduction mechanism.
increase. Because the drag change and reduction in cross-flow turbulence fluctuations
did not match, this idea was not thought to be the drag reduction mechanism in riblets.
The vortices and cross-flow turbulence fluctuations mechanisms may not be
completely independent; however, from the simulations, the correlation between
the two mechanisms was poor. Vortices generate ejecting and bursting motions which
mixes high-velocity fluid with low-velocity fluid which increases drag. The presence
of vortices near the wall could increase the cross-flow turbulence fluctuations. Vortices
closer to the wall would have a larger effect on increasing cross-flow turbulence
fluctuations than vortices further from the wall. It is thought that lifting vortices from
the riblet surface is the primary mechanism rather than the channelling of flow in the
streamwise direction as in the second mechanism.
Baseline (s, h, t, )
Smooth surface Riblet surface
Effect of
At higher At lower
Re, , , Re, , ,
vortex height/width vortex height/width
vortices are lifted vortices fall down
Effect of s
At lower s At higher s
Contact area and Contact area and
Vortex height/width constant Vortex height/width constant
Vortices are lifted Vortices fall down
Effect of h
At lower h At higher h
Protrusion height Protrusion height
Vortex height/width constant Vortex height/width constant
Effect of t
At lower t At higher t
F IGURE 12. (Colour online) Conceptual models of various parameters effects on vortices
(flow into the page, relative scale with average vortex size data). Drag reduction was
obtained when vortices are lifted from the surface. Optimal riblet spacing was found to
be when the vortex width was 1.5 times greater than the spacing with a riblet height
h+ ≈ 8–10 which is near the lower boundary of the buffer layer. Minimal changes in drag
and vortex height or width were found for tested values of t+ . At large values of t+ , the
riblet spacing would be affected and increase drag.
3.4. Effect of riblet geometry on vortices and flow properties

Figure 12 shows a conceptual model of the effect of modifying the flow properties
and riblet geometry on vortices. In the baseline case with Reynolds number based
on channel height equal to 4180, the vortex height and width for the riblet and
smooth surfaces is given, as shown in figure 9. The sizes of the vortices for the other
parameters are relative to this baseline case. By changing the velocity or viscosity
term, the Reynolds number was affected which changes the vortex size. At high
ν/Vl and therefore low Reynolds number, a small spacing and vortex size occurred;
however, the spacing size decreased faster than the vortex size and so the vortices
were lifted and a drag reduction case was exhibited. In the opposite case with low
ν/Vl and therefore high Reynolds number, the spacing increased faster than the vortex
size and so the vortices fell down and a drag increase case was exhibited, as shown
in figure 9(a).
In the models with changing spacing, a smaller spacing was better than having
a large spacing. Even though a small spacing had more surface area, it was more
important to lift away the vortices from the surface. When the vortices were lifted
from the surface, only the small contact area from the riblet peaks was exposed to
high shear stress. In the models with varying spacing or height, the vortex size did
not change because the Reynolds number based on channel height was constant, as
shown in figure 9(b,c). Even though the riblet spacing and vortex size were the same
between these cases, the models with varying height exhibited drastically different
drag changes, as shown in figure 8(c). At low heights, the riblets lifted away the
vortices from the surface and made it slightly more difficult for the high- and low-
velocity fluids to mix. The riblets also do not protrude very far into the higher-velocity
flow. In contrast, the taller riblets protrude into the higher-velocity flow and increase
the surface area exposed to high drag.
The maximum drag reduction for each parameter’s curve was approximately 10 %,
and this drag reduction was similar to other optimal riblet drag reduction studies.
According to Pollard (1998), this threshold in riblet drag reduction is due to the
passive drag reduction nature of riblets. Even though riblets lift the vortices from the
wall, there will still be some exchange of high- and low-velocity fluid in the valleys.
The riblet mechanism lifts the majority of the turbulence from the wall, but does not
significantly affect the turbulence production mechanism.
3.5. Summary of riblet mechanism

Figure 13 shows a summary figure with vortex sizes and drag change for varying
riblet spacing. The data supports the first mechanism involving lifting of vortices by
the riblet peaks, appropriately sized riblets lift the vortices from the surface decreasing
the mixing of high-velocity and low-velocity fluid which leads to a drag reduction.
The model with s+ = 25.3 had a drag reduction of 11.7 %, whereas the model with
s+ = 41.2 had a drag reduction of 4.2 %. In both cases, the vortices are approximately
the same size on the smooth and riblet surfaces. The parabolic curve fit for the
drag change has a weaker drag reduction for small and large riblet spacings. At a
small riblet spacing, the combination of greater surface area and vortices contacting
many riblet peaks leads to a weaker drag reduction. At a large riblet spacing, the
vortices fall within the riblet valley increasing drag which again leads to a weaker
drag reduction. Eventually with even larger riblet spacings, a drag increase will be
expected. A vortex width that was approximately 1.5 times the riblet spacing yielded
an optimal drag reduction.
Based on the present study, optimal riblet spacing was in the range of s+ ≈ 18–25
and will depend on other riblet parameters. These values were on the same
order to the maximum drag reduction spacing of s+ = 16 (Bechert et al. 1997b)
(a)
(b)
Smooth surface Riblet surface
(c) 15
10
Drag change (%)
–5
–10
–15
0 10 20 30 40 50
F IGURE 13. Vortices on the smooth and riblet surface with varying spacing along with the
corresponding drag change for h = 0.50 mm, t = 20.0 µm, ν/Vl = 2.38 µm and Re = 4180.
Streamwise vorticity and tangential velocity vectors plotted to show vortices on smooth
and riblet surfaces for (a) s+ = 25.3 and (b) s+ = 41.2. (c) Drag change for varying
riblet spacing. Vortices are lifted from the surface with a smaller riblet spacing, but with
a larger riblet spacing, the vortices fit within the riblet valley leading to greater drag and
a weaker drag reduction. Increasing the riblet spacing even more will eventually lead to
a drag increase.
and s+ = 17.7 (El-Samni, Chun & Yoon 2007) for blade riblets. The riblet height
should be calculated to stay in the range of h+ = 8–10. In this range, the drag
reduction was maximized; however, increasing the height and extending further into
the channel quickly led to a drag increasing case. Because riblet thickness did not
play a significant role at the tested thicknesses, the width of the riblet should be
chosen to be durable enough for the application.
4. Conclusions
The scales of fast-swimming sharks have been shown to provide a drag-reduction
benefit over a smooth plate (Bechert et al. 2000b; Jung & Bhushan 2009). These
scales have riblet structures with microgrooves that efficiently control the flow
of water and have shown drag reductions of up to 10 %. In prior studies, two
prevalent drag-reduction mechanisms have been presented. In the first mechanism,
appropriately sized riblets lift vortices from the surface decreasing eject and sweep
motions associated with high shear stress and drag. In the second mechanism, riblets
channel the flow and reduce the cross-flow turbulence fluctuations. To study these
mechanisms and the effect of riblet geometry and flow properties on drag, vortices
and turbulence fluctuations, computational models were created.
In this research, the data agreed with the first mechanism and disagreed with the
second mechanism. When the vortices were approximately 1.5 times as large as
the riblet spacing, optimal drag reduction was observed. The largest drag reductions
occurred when the riblet spacing was s+ ≈ 18–25 and h+ ≈ 8–10 which was on the
same order to s+ = 16 and h+ = 8 from Bechert et al. (1997b) and s+ = 18 and
h+ = 9 from El-Samni et al. (2007). As the vortices approached and became smaller
than the spacing size, the riblet surface became a drag increasing case. In the second
mechanism, the cross-flow turbulence fluctuations were measured and compared with
the drag change. Because the cross-flow turbulence fluctuations and drag change
trends did not match, it showed that this mechanism did not adequately explain the
riblet drag-reduction mechanism.
The average vortex size is believed to be dependent on Reynolds number based
on channel height with smaller vortices on the riblet surface than the smooth surface
due to riblets obstructing vortex rotation and growth. For both surfaces as Reynolds
number increased, vortex size increased. In addition to choosing appropriate riblet
spacing, the height needed to be h+ = 8–10 for optimal drag reduction. Increasing the
height quickly led to a drag increasing case due to the riblets protruding further into
the higher-velocity flow. The thickness parameter had little effect on the drag change
at the tested values due to the much larger surface area in the riblet valleys.
By independently changing the velocity, viscosity, spacing, height and thickness
parameters, it was observed that the non-dimensional definition was valid. Comparable
drag changes were observed for similar non-dimensional spacings that were created by
modifying the velocity, viscosity or spacing variables; therefore, the non-dimensional
parameters can be used to determine drag change in an application.

A better understanding of the riblet mechanism of controlling the streamwise
vortices and the effect of riblet geometry and flow properties on riblet drag reduction
will expand riblet usage. Optimal drag reduction is believed to occur when the vortex
width is 1.5 times the riblet spacing and riblet height of h+ = 8–10. With streamwise
vortex diameters of approximately 30 non-dimensional units, this leads to s+ = 20
and keeping the riblet height h+ = 8–10. These riblet design principles can be applied
in various commercial applications such as marine, medical and industrial fields as a
method to reduce drag.
Acknowledgements
The authors would like to thank the Ohio State University for their university
fellowship support to the first author, Dr G. Bixler and Professor S. Mazumder for
their insightful discussion, and B. Woodall for Ohio Supercomputer Center (OSC)
support. This work was supported in part by an allocation of computing time from
the OSC.
Nomenclature
s Dimensional riblet spacing (peak-to-peak distance)

h Dimensional riblet height (peak-to-valley distance)
t Dimensional riblet thickness (peak width)
s+ Non-dimensional riblet spacing (peak-to-peak distance)
h+ Non-dimensional riblet height (peak-to-valley distance)
t+ Non-dimensional riblet thickness (peak width)
vτ Wall shear-stress velocity
ν Kinematic viscosity
1τ Difference in shear stress between the riblet and smooth surfaces
τ0 Shear stress for the smooth surface
ρ Fluid density
V Average velocity in the streamwise direction
Vl Centreline velocity of a laminar parabolic profile
Q Instantaneous volume flux
Ac Cross-sectional area of the riblet model
H Height of the riblet model
W Width of the riblet model
L Length of the riblet model
δ Channel half-width of the riblet model
Re Reynolds number based on channel height
1T Dimensional time step
T+ Non-dimensional time
1h Effective protrusion height parameter
hps Streamwise protrusion height
hpc Cross-flow protrusion height
D Edge length of the square cylinder
C̄L Mean lift coefficient
CLrms Lift oscillation amplitude

C̄D Mean drag coefficient
CDrms Drag oscillation amplitude
St Strouhal number
Ds Instantaneous skin-friction drag for the smooth surface
Dr Instantaneous skin-friction drag for the riblet surface
w0s Root-mean-square cross-flow turbulence fluctuation on the smooth surface
w0r Root-mean-square cross-flow turbulence fluctuation on the riblet surface
REFERENCES
AUTUMN , K., D ITTMORE , A., S ANTOS , D., S PENKO , M. & C UTKOSKY, M. 2006 Frictional adhesion:
a new angle on gecko attachment. J. Expl Biol. 209, 3569–3579.
AUTUMN , K., L IANG , Y. A., H SIEH , S. T., Z ESCH , W., C HAN , W. P., K ENNY, T. W., F EARING ,
R. & F ULL , R. J. 2000 Adhesive force of a single gecko foot-hair. Nature 405, 681–685.
B ECHERT, D. & B ARTENWERFER , M. 1989 The viscous flow on surfaces with longitudinal ribs.
J. Fluid Mech. 206, 105–129.
B ECHERT, D., BARTENWERFER , M., H OPPE , G. & R EIF , W. E. 1986 Drag reduction mechanisms
derived from shark skin. In ICAS, Congress, 15th, London, England, September 7–12, 1986,
Proceedings. Vol. 2 (A86-48976 24-01), pp. 1044–1068. American Institute of Aeronautics
and Astronautics, Inc.
B ECHERT, D., B RUSE , M., H AGE , W. & M EYER , R. 1997a Biological Surfaces and Their
Technological Application Laboratory and Flight Experiments on Drag Reduction and Separation
Control, Paper #AIAA-1997-1960, presented at AIAA 28th Fluid Dynamics Conference,
Snowmass Village, CO, AIAA, New York.
B ECHERT, D., B RUSE , M. & H AGE , W. 1997b Experiments on drag-reducing surfaces and their
optimization with an adjustable geometry. J. Fluid Mech. 338, 59–87.
B ECHERT, D., B RUSE , M. & H AGE , W. 2000a Experiments with three-dimensional riblets as an
idealized model of shark skin. Exp. Fluids 28, 403–412.
B ECHERT, D., B RUSE , M. & H AGE , W. 2000b Fluid mechanics of biological surfaces and their
technological application. Naturwissenschaften 87, 157–171.
B ECHERT, D., H OPPE , G., VAN DER H OEVEN , J. T. & M AKRIS , R. 1992 The Berlin oil channel
for drag reduction research. Exp. Fluids 12, 251–260.
B HUSHAN , B. 2009 Biomimetics: lessons from nature – an overview. Phil. Trans. R. Soc. Lond. A
367, 1445–1486.
B HUSHAN , B. 2010 Springer Handbook of Nanotechnology, 3rd edn. Springer.
B HUSHAN , B. 2012 Biomimetics: Bioinspired Hierarchical-Structured Surfaces for Green Science and
Technology. Springer.
B IXLER , G. D. & B HUSHAN , B. 2012a Biofouling: lessons from nature. Phil. Trans. R. Soc. Lond.
A 370, 2381–2417.
B IXLER , G. D. & B HUSHAN , B. 2012b Bioinspired rice leaf and butterfly wing surface structures
combining shark skin and lotus effects. Soft Matt. 8, 11271–11284.
B IXLER , G. D. & B HUSHAN , B. 2013 Shark skin inspired low-drag microstructured surfaces in
closed channel flow. J. Colloid Interface Sci. 393, 384–396.
C HENG , C., L U , P. & C HEN , R. 1992 Wind loads on square cylinder in homogeneous turbulent
flows. J. Wind Engng Ind. Aerodyn. 41, 739–749.
C HOI , H., M OIN , P. & K IM , J. 1993 Direct numerical simulation of turbulent flow over riblets.
J. Fluid Mech. 255, 503–539.
C HU , D. C. & K ARNIADAKIS , G. E. 1993 A direct numerical simulation of laminar and turbulent
flow over riblet-mounted surfaces. J. Fluid Mech. 250, 1–42.
D EAN , B. & B HUSHAN , B. 2010 Shark-skin surfaces for fluid-drag reduction in turbulent flow: a
review. Phil. Trans. R. Soc. Lond. A 368, 4775–4806.
E L -S AMNI , O. A., C HUN , H. H. & Y OON , H. S. 2007 Drag reduction of turbulent flow over thin
rectangular riblets. Intl J. Engng Sci. 45, 436–454.
F LUENT 2014 ANSYSr FLUENT, Release 14.0, Help System, FLUENT User’s Guide, ANSYS, Inc.
G AO , H., WANG , X., YAO , H., G ORB , S. & A RZT, E. 2005 Mechanics of hierarchical adhesion
structures of geckos. Mech. Mater. 37, 275–285.
G ARCÍA -M AYORAL , R. & J IMÉNEZ , J. 2011 Drag reduction by riblets. Phil. Trans. R. Soc. Lond.
A 369, 1412–1427.
G ENZER , J. & E FIMENKO , K. 2006 Recent developments in superhydrophobic surfaces and their
relevance to marine fouling: a review. Biofouling 22, 339–360.
G OLDSTEIN , D., H ANDLER , R. & S IROVICH , L. 1995 Direct numerical simulation of turbulent flow
over a modelled riblet-covered surface. J. Fluid Mech. 302, 333–376.
J IMENEZ , J. & M OIN , P. 1991 The minimal flow unit in near-wall turbulence. J. Fluid Mech. 225,
213–240.
J UNG , Y. C. & B HUSHAN , B. 2009 Biomimetic structures for fluid drag reduction in laminar and
turbulent flows. J. Phys.: Condens. Matter 22, 035104.
K IM , J., M OIN , P. & M OSER , R. 1987 Turbulence statistics in fully developed channel flow at low
Reynolds number. J. Fluid Mech. 177, 133–166.
K RIEGER , K. 2004 Do pool sharks swim faster? Science 305, 636–637.
L EE , B. 1975 The effect of turbulence on the surface pressure field of a square prism. J. Fluid
Mech. 69, 263–282.
L EE , S.-J. & L EE , S.-H. 2001 Flow field analysis of a turbulent boundary layer over a riblet surface.
Exp. Fluids 30, 153–166.
LYN , D. A., E INAV, S., RODI , W. & PARK , J. 1995 A laser-Doppler velocimetry study of ensemble-
averaged characteristics of the turbulent near wake of a square cylinder. J. Fluid Mech. 304,
285–319.
LYN , D. & RODI , W. 1994 The flapping shear layer formed by flow separation from the forward
corner of a square cylinder. J. Fluid Mech. 267, 353–376.
M C L EAN , I. & G ARTSHORE , I. 1992 Spanwise correlations of pressure on a rigid square section
cylinder. J. Wind Engng Ind. Aerodyn. 41, 797–808.
M URAKAMI , S. & M OCHIDA , A. 1995 On turbulent vortex shedding flow past 2D square cylinder
predicted by CFD. J. Wind Engng Ind. Aerodyn. 54, 191–211.
O CHOA , J. & F UEYO , N. 2004 Large-Eddy Simulation of the flow past a square cylinder. In
PHOENICS 10th International User Conference Melbourne, Australia, CHAM Limited, London.
P OLLARD , A. 1998 Passive and active control of near-wall turbulence. Prog. Aerosp. Sci. 33, 689–708.
R EIF , W. E. 1985 Squamation and Ecology of Sharks. E. Schweizerbartsche Verlagsbuchhandlung.
ROBINSON , S. K. 1990 A review of vortex structures and associated coherent motions in turbulent
boundary layers. In Structure of Turbulence and Drag Reduction, pp. 23–50. Springer.
ROBINSON , S., K LINE , S. J. & S PALART, P. 1989 A review of quasi-coherent structures in a
numerically simulated turbulent boundary layer. NASA STI/Recon Tech. Rep. N 90: 13723.
RODI , W. 1997 Comparison of LES and RANS calculations of the flow around bluff bodies. J. Wind
Engng Ind. Aerodyn. 69, 55–75.
RODI , W., F ERZIGER , J., B REUER , M. & P OURQUIE , M. 1997 Status of large eddy simulation:
results of a workshop. Trans. ASME J. Fluids Engng 119, 248–262.
ROHR , J., A NDERSEN , G., R EIDY, L. & H ENDRICKS , E. 1992 A comparison of the drag-reducing
benefits of riblets in internal and external flows. Exp. Fluids 13, 361–368.
S PALDING , D. B. 1961 A single formula for the law of the wall. Trans. ASME J. Appl. Mech. 28,
455–458.
S UZUKI , Y. & K ASAGI , N. 1994 Turbulent drag reduction mechanism above a riblet surface. AIAA
J. 32, 1781–1790.
V ICKERY, B. 1966 Fluctuating lift and drag on a long cylinder of square cross-section in a smooth
and in a turbulent stream. J. Fluid Mech. 25, 481–494.

WALSH , M. J. 1982 Turbulent Boundary Layer Drag Reduction using Riblets, Paper #AIAA-1982-0169,
presented at AIAA 20th Aerospace Sciences Meeting, Orlando FL, AIAA, New York.
WALSH , M. J. & A NDERS , J. B. J R 1989 Riblet/LEBU research at NASA Langley. Appl. Sci. Res.
46, 255–262.
W EISS , M. H. 1993 Drag reduction with riblets in pipe flow. PhD Diss., University of Calgary,
Alberta, Canada.
W EISS , M. 1997 Implementation of drag reduction techniques in natural gas pipelines. In 10th
European Drag Reduction Working Meeting, Berlin, Germany, 19–21 March.

Fluid Flow Analysis of A Shark-Inspired Microstructure

Uploaded by

Copyright:

Available Formats

Fluid Flow Analysis of A Shark-Inspired Microstructure

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fluid Flow Analysis of A Shark-Inspired Microstructure

Uploaded by

Copyright:

Available Formats

J. Fluid Mech. (2014), vol. 756, pp. 5–29.

c Cambridge University Press 2014 5

Fluid flow analysis of a shark-inspired

Samuel Martin1 and Bharat Bhushan1, †

(Received 6 March 2014; revised 5 June 2014; accepted 25 July 2014;

The scales of fast-swimming sharks contain riblet structures with microgrooves,

† Email address for correspondence: bhushan.2@osu.edu

Actual shark skin (Mako, Isurus oxyrinchus)

spacing from this experiment was non-dimensionalized using

s+ = svτ /ν. (1.1)

Non-dimensional values are denoted by a superscript +. In this equation, the

Image Observations Questions

2.1. CFD program

2.2. Validation model: flow around a square cylinder

cylinder with edge length D = 0.04 m was performed. This simulation is a

Reference C̄L CLrms C̄D CDrms St

2.3. Shark-skin inspired riblet models

s (mm) h (mm) t (µm) ν/Vl (µm) s+ h+ t+ Number of riblets

0.50 16.2 8.1 0.32 20

0 0.5 1.0 1.5

2.3.1. Model set-up

(a) Smooth top

Along the streamwise direction, a constant instantaneous volume flux boundary

2.3.2. Drag, vortex and turbulence fluctuation measurement

0 100 200 300 400 500

3. Results and discussion

3.1. Drag change

3.2. Vortex height and width

data was approximately the same for each of these cases.

Vortex center distance from wall (

3.3. Turbulence fluctuations

0 5 10 15 0 0.2 0.4 0.6

3.4. Effect of riblet geometry on vortices and flow properties

3.5. Summary of riblet mechanism

Smooth surface Riblet surface

parameters can be used to determine drag change in an application.

s Dimensional riblet spacing (peak-to-peak distance)

CLrms Lift oscillation amplitude

and in a turbulent stream. J. Fluid Mech. 25, 481–494.

You might also like