Ocean Engineering 134 (2017) 35–49

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Passive flow control in the near wake of a circular cylinder using attached MARK
permeable and inclined short plates
⁎
Gokturk M. Ozkana, , Erhan Firatb, Huseyin Akillia
a
Faculty of Engineering and Architecture, Department of Mechanical Engineering, Çukurova University, 01330 Adana, Turkey
b
Faculty of Engineering, Department of Mechanical Engineering, Munzur University, Tunceli, Turkey

A R T I C L E I N F O A BS T RAC T

Keywords: The main objective of this study is to assess the effectiveness of attached permeable plates in suppressing
Cylinder nominally two-dimensional vortex shedding from a fixed cylinder. The permeable plate along the entire span of
Drag reduction the cylinder was made of a chrome-nickel screen and was attached normal to the cylinder surface. The main
Flow control parameters of the study are porosity of the permeable plate, β, and the plate angle of the cylinder-plate
Flow visualization
arrangement, θ, with respect to the freestream. Experiments were performed using Particle Image Velocimetry
Permeable plate
(PIV) and Flow Visualization Techniques. Additionally, drag force measurements were carried out in a wind
Porosity
VIV tunnel to analyze the effects of permeable plates on drag coefficient of the circular cylinder. The results revealed
that the use of a permeable plate successfully suppresses the vortex shedding downstream of the circular
cylinder by reducing the velocity fluctuations in the wake, elongating the vortex formation region further
downstream and attenuating the vortex shedding frequency. The porosity values of β=0.4 and 0.5 with plate
angles in the range of 35°≤θ≤90° were found to be effective on suppressing the vortex shedding. Furthermore,
both the drag coefficients of the cylinder-plate arrangement and the < TKE > max were reduced almost for all
porosity ratios for θ < 30°.

1. Introduction many others) due to its tempting simplicity in setting up such an

The shedding of vortices generated from a bluff body has been of
great interest for many researchers since it continuously results in
serious problems in a wide range of engineering applications. The
physical mechanism of vortex shedding generates vortex-induced
vibrations (VIV) and high unsteady loadings which could damage the
structural integrity of even huge constructions. The vortical formation
behind bluff bodies also directly affects the amount of drag and lift
forces exerted on the body. Some examples for the engineering
applications concerning the vortex shedding are risers in offshore
engineering, high-rise buildings, bridge piers, chimney stacks, cooling
towers. In recent years investigations have focused on applications in
offshore engineering since oil risers (made of steel pipes) break like
thin wires due to large amplitude VIVs caused by the deep ocean
currents. Ocean engineers pay special attention to this problem
because riser systems cost one-third of an entire offshore production
unit. Due to its great importance, bluff body wakes and vortex shedding
regimes have been attractive research areas. Considering only the
circular cylinder wake as a bluff body have been preferred (Williamson,
1996; Gabbai and Benaroya, 2005; Bearman, 2011 may be cited among
arrangement in experimental or computational laboratory.
Various flow control techniques to eliminate and/or reduce the
adverse effects of vortex shedding were reviewed extensively by Choi
et al. (2008). They introduced recent major achievements in bluff- body
based flow controls such as 3D forcing, active feedback control, control
based on local and global instability, and control with a synthetic jet.
They also classified the controls as boundary-layer controls and direct-
wake modifications and discussed important features associated with
these control methods. More recently, the review of Rashidi et al.
(2016) exhibited a different perspective on the control methods
including their advantages and limitations. They classified the existing
flow control methods into nine different categories and gave useful
conclusions and recommendations. With a broader perspective, a
general classification of the methods of vortex shedding control is
known as active and passive techniques. Some geometrical modifica-
tions without energy consumption is valid for passive control methods,
whereas an energy input is required to control the vortical structure in
an active manner. Generally, introducing a passive control method is
easier to implement and less costly since no energy input is needed.
Zdravkovich (1981) classified the passive control techniques in three

⁎
Corresponding author.
E-mail address: gmozkan@cu.edu.tr (G.M. Ozkan).

http://dx.doi.org/10.1016/j.oceaneng.2017.02.014
Received 15 June 2016; Received in revised form 12 October 2016; Accepted 14 February 2017
Available online 20 February 2017
0029-8018/ © 2017 Elsevier Ltd. All rights reserved.
G.M. Ozkan et al.
ways: impacting separated shear layers or separation lines with
protrusion devices such as helical strakes, wires, studs or spheres;
affecting the shear layer emanating from both sides of a bluff body with
devices such as shrouds; and preventing the interaction of the
entrainment layers with some devices such as splitter plate and base
bleed. A selective review of various aerodynamic passive control
measures which could be adopted to reduce vortex-induced vibrations
(most general type of flow-induced vibration) is presented by Kumar
et al. (2008). They introduced recent patents on this topic and
concluding notes on the current and future developments expected in
the passive control of VIV. Throughout the passive control methods,
splitter plate has attracted the attention of most researchers starting
with Roshko (1955) who indicated that the use of a splitter plate with a
length of 5D (D is the cylinder diameter) at Re=1.45×104 results in a
complete suppression of vortex shedding, consequently decrease in
drag force due to increase in base pressure. According to the results of
Roshko (1954, 1955), many researchers started to use splitter plates as
flow control devices for suppressing the vortex shedding (Bearman,
1965; Gerrard, 1966; Apelt et al., 1973, Apelt and West, 1975; Unal
and Rockwell, 1987; Nakamura, 1996; Anderson and Szewczyk, 1997;
Ozono, 1999, Hwang and Yang, 2007; Akilli et al., 2008, Shukla et al.,
2009; Gu et al., 2012; Gozmen et al., 2013; Assi et al., 2014; Huera-
Huarte, 2014; Barman and Bhattacharyya, 2015; Zhang and Weiping,
2016). The study of Bearman (1965) investigated the flow behind a
two-dimensional model with a blunt trailing edge and fitted with
splitter plates at Reynolds numbers between 140,000 and 256,000. He
measured the base pressure and shedding frequency and showed that
the existence of a splitter plate increases the base pressure significantly
(thereby reducing the pressure drag), even for very short plate length.
He also found that a splitter plate length of 2D completely suppresses
the vortex shedding for the given Reynolds numbers. A lower
Reynolds-number-study was performed by Gerrard (1966) who studied
the mechanism of vortex generation behind a cylinder at ReD=2×104.
He showed that the Strouhal number is decreased when the length of
the splitter plate was smaller than cylinder diameter, D but it increased
for D < L < 2D. Akilli et al. (2008) experimentally investigated the
passive control of vortex shedding behind a circular cylinder by splitter
plates of various lengths attached on the cylinder base in shallow water
flow at Re=6300. They found that the flow characteristics in the wake
region of circular cylinder sharply change up to the splitter plate length
of L/D=1.0 and large-scale vortex shedding behind cylinder has
completely disappeared for L/D > 1.2. Assi et al. (2009) presented a
variation of control methods using plates; free-to-rotate splitter plate,
double splitter plate, parallel plates and parallel plates with gap. Within
these control methods, they achieved a drag reduction of 38% by using
parallel plates. They also reported that a free-to-rotate splitter plate
also suppresses VIV, however a mean transverse force was generated
which should be eliminated. Cardell (1993) brought a new approach to
splitter plate studies by examining the permeable wake splitter plate
spanning in the wake center plane. He defined the permeability by
means of solidity (which is opposite of the term porosity) and derived it
using the pressure drop across the permeable plates which are made
from woven screens available in the market (β=0.35, 0.51, 0.65, 0.71,
0.78, 0.81). He used infinitely long permeable plates to examine their
effects on wake structures for various solidities and used flow visualiza-
tion followed with the hot wire, pressure and drag measurements for a
Reynolds number range of 2500≤Re≤18,000. He concluded that the
use of a low permeable splitter plate separates the large scale vortical
structures in the near wake of the cylinder whereas the wake structure
for a high permeable splitter plate preserves its basic form similar to
the no-splitter plate case. According to the discussions of Cardell
(1993) it can be said that the use of permeable materials brings out new
alternatives to current techniques on flow control since the control
elements such as a woven metal screen are readily available in the
market. However, Cardell (1993) used very long splitter plates in the
experiments, 16D in length, which is quite difficult to adopt for various
36
Ocean Engineering 134 (2017) 35–49
engineering applications. In addition, the splitter plates oriented only
at 0° with respect to the oncoming flow i.e. the effects of the plate angle
on the flow characteristics were not investigated. These shortcomings
contribute to the rationale for the present study. A recent study of Assi
and Bearman (2015) showed the flow-induced vibration (FIV) response
of elastically-mounted circular cylinders fitted with non-rotating, three
different geometries of splitter plates: short solid (0.5 diameter in
length), long solid (1 diameter in length) and slotted (1 diameter in
length and 70% solidity). It was said that the splitter plates seemed to
keep the drag below that observed during the VIV of plain cylinder,
even though vibrating with much larger transverse displacements in
the Re range of 3000–7000. They also revealed that the cylinders fitted
with splitter plates show a continuous increase in y-displacement
compared with the plain cylinder as Re increased, and that the
response curves of the cylinders with solid splitter plates are steeper
than those of the slotted one. Their study clearly showed that when
compared with the impermeable (solid) plate, permeability (porosity)
offers some advantages from the point of view of reducing the
amplitude of vibration.
It can be noted that permeable (porous) material was not only used
as an attached splitter plate but also used as a detached plate to
suppress the vortex shedding behind a circular cylinder in shallow
water flow (Negretti et al., 2005), as a cover to control the flow noise
generated by a set of cylinders that consist of a rigid core cylinder
covered by different porous material in cross-flow (Geyer and Sarradj,
2016), as a thin outer shell (shroud) to suppress the vortex shedding
from the inner solid cylinder in cross-flow (Bruneau and Mortavazi,
2006; Bhattacharyya and Singh, 2009; Ozkan et al., 2012, 2013, 2015;
Gozmen and Akilli, 2014), as a cover to enhance the heat transfer from
the heated cylinder in cross-flow (Al-Salem et al., 2011), a perforated
cylinder to attenuate large-scale vortex formation in the wake of it in
shallow water flow (Pinar et al., 2015; Durhasan et al., 2016).
In this study, the unsteady vortical flow structure downstream of a
circular cylinder was controlled using a permeable plate by varying
plate angles and porosity ratios. The permeable plate is made of a
woven screen which is commercially available and easy to implement.
Among the wide-range research including splitter plate applications,
the combination of permeability and splitter plate was considered
within this study for a large parameter interval.
2. Experimental details
2.1. Flow facilities
During the course of this study, a close-circuit water channel
(located in the Mechanical Engineering Department of Çukurova
University) and an open circuit low-speed wind tunnel (located in the
Mechanical Engineering Department of Niğde University) were used.
The water channel with a test section of 750 mm×1000 mm×8000 mm
(height, width and length, respectively) was used for obtaining
instantaneous, two-dimensional, whole field velocity measurements
using the Particle Image Velocimetry (PIV) technique, and for captur-
ing the dye streaklines in the near wake of the models by dye injection.
These experiments were performed considering deep water flow
conditions (Fig. 1). The height of the water level and freestream
velocity were kept constant as hw=600 mm and U∞=95.2 mm/s,
respectively. The Reynolds number, based on D (diameter of the
cylinder), was about 5000. The freestream turbulence intensity (Tux)
was less than 2%. The walls of the test section were made of Plexiglas
material to allow optical access to the desired part of the flow in
channel. A 15 kW centrifugal pump was used to drive the water
through the close-circuit channel. The impeller speed of the pump
can be controlled by a frequency converter.
Force measurements were performed in the suction-type, open-
circuit, subsonic wind tunnel with a test section of 570 mm high,
570 mm wide and 1000 mm long (Fig. 2). The walls of test section (also
G.M. Ozkan et al.
Fig. 1. Schematic illustration of the cylinder with attached permeable plate installed in
the recirculating water channel with positions of the PIV camera and the laser sheet.
made of Plexiglas material) were expanded with a divergence angle of
0.3° from inlet to outlet so as to ensure constant static pressure along
the whole test section. The freestream turbulence intensity (Tux) in the
test section was less than 1% for the studied Reynolds number
(ReD=5000). A Pitot-static tube in connection with an electronic micro
manometer was installed into the test section to determine the
freestream velocity.
2.2. Models
Hollow cylinders with a wall thickness of 3 mm and diameter (D) of
50 mm were used in the experiments. These cylinders are made of
Plexiglas material because of its good surface quality and transparency.
Permeable splitter plates are made of woven chrome-nickel wire
meshes (screens) with various square openings. To produce the
permeable plates, meshes with various opening area cut to a length
that is equal to D and a length that is equal to 600 mm. These
Fig. 2. A sketch of the wind tunnel test section with force measurement system. Flow is from left to right.
37
Ocean Engineering 134 (2017) 35–49
Fig. 3. Presentation of the cylinder with permeable plate, details of the woven screen
with definition of porosity, β.
nominally planar plates were first heated, then properly attached
normal to the cylinder surface without any distortion. That is, no
adhesive was used to stick them together. The main parameters of this
study are the porosity/permeability (β=0, 0.4, 0.5, 0.6, 0.7, 0.8) and the
plate angle (θ=0°, 15°, 30°, 35°, 40°, 45°, 60°, 75°, 90°) with respect to
the wake centerline (Fig. 3).
Models considered in the present work were oriented vertically both
in water channel and wind tunnel. Models were immersed into the
water and fixed. Immersed end of the models were in contact with the
bottom of the channel. That is, models were bounded in the top by free-
surface and in the bottom by the channel floor. Models inserted into the
water channel were produced having a blockage ratio (BR=wc/D where
wc is width of the channel) of 5% and an aspect ratio (AR=hc/D where
hc is height of the channel) of 12. On the other hand, two-dimensional
models spanned the height of the wind tunnel test section; therefore,
no end plate was employed. The gap between the models and the
tunnel's walls were about 4–6 mm. Thus, models inserted into the wind
tunnel were produced having a blockage area of 8.8% and an aspect
ratio of 11.4.
2.3. Flow visualization experiments
The flow visualization experiments were carried out first to gain a
preliminary understanding of the flow structure. By qualitative flow
visualization via the injection of the fluorescent dye into the flow,
vortex formation length, development of free shear layers, vortex
shedding frequency, etc. can be approximately evaluated beside
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

comparing it with quantitative PIV measurements.

The fluorescent dye was comprised of water ( > 99.5% w/w) and
small amount of Rhodamine B powder (Sigma-Aldrich Co.). The
density of the dye was very close to that of the water i.e. dye is
neutrally buoyant. Dye was flowed by gravity from the dye reservoir
that is located 1000 mm above the free-surface of the water channel.
The injection rate was controlled by a roller clamp on the tubing. Main
flow was inevitably disturbed by the presence of the ejecting device. To
minimize this, attached two thin needles (ejecting device) aligned with
the main flow were located 8D upstream of the model. The behavior of
the nominally two-dimensional, unsteady wake in the plane of a
continuous laser light sheet (Spectra-Physics Lasers, Model 177-
G0232) was recorded by a high definition camcorder (Sony, Model
HDR-SR1) having a capturing frequency of 24 Hz. The laser light sheet
was oriented at mid-span of the immersed part of models.

2.4. Velocity measurements

Standard (2D2C) Particle Image Velocimetry (PIV) system was used

to obtain whole field velocity measurements (Dantec Dynamics A/S).
The field of view was illuminated by double-pulsed Nd: YAG laser that
produces 120 mJ pulses at 532 nm and a repetition rate of 15 Hz (New
Wave Research Inc., Model Solo 120XT). The water was seeded with
nominally 10 µm diameter silver coated hollow glass sphere particles
having a density of 1400 kg/m3. The laser light sheet that illuminated
the mid-depth of the channel flow was parallel to the bottom wall of the
channel. A charge coupled device (CCD) camera (Dantec Dynamics A/
S, Model FlowSense 2M) equipped with a lens (Nikon, Model AF
Micro-Nikkor 60 mm f/2.8D) was used to capture the frames for
velocity evaluation.
For configuration, acquisition, analysis and post-processing of
acquired PIV data, Dynamics Studio v3.2 software was used. For each
case, 1000 images with a pixel depth of 8 bit were acquired in double-
frame mode with a time delay of 1.75 ms between laser pulses.
Adaptive correlation technique with 2 refinement steps, 2 pass per
Fig. 4. The results of plain cylinder: Time averaged streamline topology, < Ψ >
step, 32 pixel×32 pixel interrogation windows and 50% overlap was superposed with the vorticity contours, < ω > , Reynolds Shear Stress, < u´v´/U∞2 >
used to process the double-frame images prior to 2D velocity recon- and Turbulent Kinetic Energy, < TKE > contours.
struction so as to spatially resolve the gradients in the flow. The
resulting instantaneous velocity maps covered an area of with the rotary unit. The uncertainties in the θ and < CD > were found
219 mm×164 mm and contained 7326 (99×74) vectors. by applying the general uncertainty analysis method presented by Kline
Spurious velocity vectors (less than 3%) were detected using the and Mcclintock (1953). The uncertainty in the θ was estimated to be
local median-filter technique. Thereafter, they were replaced by a new approximately ± 0.5°. The total uncertainties in time-averaged drag
vector that was calculated by employing bilinear least-squares techni- coefficients of the models were estimated to be less than 8%. Also, force
que between surrounding vectors. During the analysis, the instanta- measurements were performed from θ=0–45° by 5° step, and then, 15°
neous or time-averaged velocity (1000 instantaneous vector maps were step up to 90°.
used for time-averaging) data were also used to calculate other flow
features, such as, vortex shedding frequency, vorticity, time-averaged
recirculation length, topology of streamlines, etc. 3. Results and discussion

2.5. Force measurements 3.1. Results for the plain cylinder

Force measurements were made using a six axis force/torque (F/T)
measurement system (ATI Industrial Automation, Model Gamma/SI-
32-2.5). ATI DAQ F/T software with previously loaded transducer
calibration data was used to set the sampling rate (500 Hz), number of
samples (20,000) beside acquiring the F/T data. The force measure-
ments were averaged over two runs. The time-averaged drag coeffi-
cients of the models were defined as
2∙<FD >
<CD > =
ρ∙U∞2 ∙AP (1)
where FD is the drag force acting on the model, ρ is the density of the
air, AP is the frontal (projection) area of the model. A rotary unit (isel
Germany AG, ZD30) was used to precisely (0.042°) adjust the plate
angle, θ, of the models. Dantec Dynamics application software
(MiniCTA) and traverse controller unit were used to communicate
Fig. 4 presents the results obtained for the plain cylinder where the
magnitudes of time-averaged vorticity contours and streamwise velo-
city contours are presented by legends shown on the right side of the
figure. Here, the results are in good agreement with the previous works,
i.e. a recirculation region appears in the near wake of the cylinder with
formations of two well- defined vortices, whose respective focal points
are denoted by F1 and F2 which are almost symmetrical (slight
asymmetry might be due to the measurement uncertainties) with
respect to the centerline and generated at 0.8D (in good agreement
with the validated results of Aljure et al. (2015)) away from the base of
the cylinder. The fluid flow gains zero velocity at a position 1.4D away
from the cylinder base which is called as the saddle point designated by
S1. This corresponds to the vortex formation length lf, being consistent
with the results of Norberg (1998) and Aljure et al. (2015). The
vorticity contours are almost symmetrical and get closer to each other

38
G.M. Ozkan et al.
in the wake region which is an indication of interaction of the vortices
and consequent vortex shedding. Reynolds Shear Stresses evaluated
from PIV data are presented in the second row of Fig. 4 which were
calculated by the multiplication of velocity fluctuations in streamwise
(u´) and spanwise (v´) directions. Taking the average of the multi-
plications gives the time- averaged Reynolds Shear Stresses,<u´v´ >.
The values thus obtained are divided by the depth-averaged free stream
velocity squared U∞2 in order to make better comparisons. Here, the
continuous and dashed lines present the positive and negative
Reynolds Shear Stresses, respectively. The minimum of positive and
maximum of negative values of RSS contours are, +0.01 and −001,
respectively while the increment was taken as 0.01. It is clearly seen by
the results that small- scale positive and negative clusters of Reynolds
Shear Stress exist just behind the cylinder base in the near wake. In
accordance with the large-scale vortices generated by the unsteady
vortex shedding, the large- scale clusters of Reynolds Shear Stress
contours are formed approximately at 1.4D away from the cylinder
base which also correspond to the saddle point of the cylinder. The
maximum Reynolds Shear Stress values for the positive and negative
clusters are evaluated as 0.098 and −0.096, respectively. The last row of
Fig. 4 shows the distribution of time- averaged Turbulent Kinetic
energy, < TKE > where the minimum and incremental values were
taken as 0.1. Assuming isotropic flow, the third fluctuating velocity
component can be estimated by a 2D approximation (Sheng et al.,
2000) where the third component is taken to be equal to half of < u´u´
> + < v´v´ >, and then < TKE > can be calculated with the expression,
< TKE >=3/4( < u´u´ > + < v´v´ > ). As seen by the third row of Fig. 1,
the concentrations of < TKE > gradually increase in downstream
direction by approaching a maximum value of 0.33 which has been
made dimensionless by dividing it with the square of the depth-
averaged free-stream velocity (U∞)2. The < TKE > concentrations lose
their magnitude in further downstream locations since the unsteady
vortex shedding weakens with increasing streamwise locations. It
should be pointed out that the location of the maximum < TKE > is
measured to be 1.4D away from the cylinder base which is the same
location as evaluated for the saddle point and maximum Reynolds
Shear Stress in first and second rows of Fig. 4, respectively.
3.2. PIV results of the controlled cases
Since the study has been carried out for a wide range of parameters,
it would be better to get started by observing the effects of all
parameters on a single chart rather than analyzing them separately.
Therefore the peak magnitudes of < TKE > for all cases as well as the
plain cylinder case are presented by two dimensional curves which are
given in Fig. 5. For controlling the unsteady structure of the flow
downstream of bluff bodies it is necessary to reduce the intensity of
Turbulent Kinetic Energy, especially in the near wake. According to this
Fig. 5. The effect of plate angle, θ on the peak values of turbulent kinetic energy,
< TKE > max for various porosities.
39
Ocean Engineering 134 (2017) 35–49
physical reality, the lower magnitude Turbulent Kinetic Energy peaks
of the controlled cases in Fig. 5 would give the effective intervals of the
current study. First of all, the maximum value of < TKE > generated
downstream of the plain cylinder is measured to be 0.33 which is
shown by the dashed line on the graph. Attaching a solid plate which is
known as splitter plate in the literature, β=0.0 shows a remarkable
decrease in the peak value of TKE compared with the plain cylinder. It
is well known from the previous works that use of a splitter plate
downstream of the cylinder is an effective way of flow control since
interactions of shear layers can be prevented or moved farther away
from the cylinder wake, consequently resulting in decreased fluctua-
tions in the near wake (Gozmen et al., 2013). For the current study, a
reduction of 49% in the peak value of <TKE > is achieved for the
splitter plate ( β =0.0, θ =0°). Better results are obtained increasing the
plate angle to θ =15°, resulting in a reduction of 55% of the maximum
value. Further increase in the plate angle results in a more complex
flow structure due to the increased cross sectional area among the cross
stream flow. Therefore, the trend of solid plate for 30°≤θ≤60° shows
that the peak magnitudes are lower than that of the plain cylinder case.
Increasing the plate angle increases the peak magnitudes of < TKE > ,
exceeding the < TKE > value related to the uncontrolled case. A slight
reduction in the peak magnitude of < TKE > exists for the solid plate
angle of θ=90° when compared with the plain cylinder.
As being the innovative side of the present research, using a
permeable plate instead of the solid one is expected to bring a more
effective flow control, especially with higher plate angles, since the free
stream flow now can enter through the permeable material (screen).
According to Fig. 5, use of the permeable plates having porosities of
β=0.4 and β=0.5 is found to be remarkably effective on reducing the
maximum < TKE > values in the wake of a cylinder, especially for
the plate angles of 45°≤θ≤90°. For β=0.4 and θ=45°, a reduction in the
< TKE > is obtained as 74% which is the highest reduction being the
best case of decreasing < TKE > . However, the peak magnitudes of
< TKE > are still lower than those of the plain cylinder for 0°≤θ≤40° as
seen by the Fig. 5. The higher porosities of β=0.6 and β=0.7 are also
effective on reducing the maximums. However, a slight increase in the
trend also exists with respect to the lower porosities. It should be noted
that for β=0.8 a drastic increase in the peak magnitude of < TKE >
occurs for 30°≤θ≤40° which could be due to the complete interaction of
lower shear layer with the high-porosity permeable plate. This condi-
tion is also valid for the solid plate and it is probably due to the
complex flow structure occurring downstream of solid plate- cylinder
arrangement. Fig. 5 shows that the porosity of β=0.8 results in higher
peak magnitudes of < TKE > with respect to the lower porosities. The
peak magnitudes are still lower than that of the plain cylinder except
for the plate angle of θ=35° in which the maximum value of < TKE > is
6.4% higher than that of the plain cylinder. Finally, it can be concluded
that the intervals of 0.4≤β≤0.6 and 45°≤θ≤90° seem to be the most
effective cases in order to decrease the velocity fluctuations in the wake
of the cylinder. Dramatic reduction in the peak magnitudes of < TKE >
in the near wake are observed.
According to Feng and Wang (2010), the < TKE > production can
be related to the dynamics of the vortices. Therefore, it is requisite to
analyze the vortical structures in the wake of the cylinder especially for
the cases with have reduced < TKE > values compared to the plain
cylinder. In order to study the vortex shedding mechanism, the
instantaneous dye pictures and vorticity contours evaluated by PIV
experiments are analyzed and presented in Fig. 6a and b. To eliminate
crowding of the figures, two porosity values are selected considering
their effectiveness on reducing the maximum < TKE > . The continuous
and dashed lines of vorticity contours present the positive (counter
clockwise) and negative (clock wise) vortices, respectively. In Fig. 6a,
the results are presented for the plain cylinder and the controlled case
of β=0.5 which was found to be an effective porosity on reducing
< TKE > . Instantaneous picture of the plain cylinder shows the roll-up
of the upper and lower shear layers into the wake region which results
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 6. a. Instantaneous dye visualization pictures (left) and vorticity contours (right) for the plain cylinder and cylinder with permeable plate of β=0.5. b. Instantaneous dye
visualization pictures (left) and vorticity contours (right) for the plain cylinder and cylinder with permeable plate of β=0.7.

40
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 6. (continued)

41
G.M. Ozkan et al.
in alternating vortex shedding and consequent Karman Vortex Street
downstream of the cylinder. It can be said that the instantaneous
vorticity contours evaluated from PIV data are in good agreement with
the dye picture of plain cylinder where the interaction of upper and
lower shear layers and consequent vortex shedding are clearly seen.
From a physical point of view, it is expected that attaching a permeable
plate along the midspan of a circular cylinder and giving a certain angle
to the plate would directly affect the shear layer (i.e. lower shear layer
of the cylinder). The latter rows of Fig. 6a presents the instantaneous
data for the permeable plate of β=0.5 for specified plate angles. For the
case of θ=0°, the upper and lower shear layers interact with each other
at trailing edge of the permeable plate. This means that the vortices
formed by the shear layers cannot roll-up through the meshes because
of the low porosity value. Therefore, well-organized Karman Vortex
Street is formed in farther downstream regions compared with the
plain cylinder. Increasing the plate angle to θ=15° results in an
increase in the distance of first contact point of upper and lower shear
layers. The formation of vortex shedding and Karman Vortex Street are
clearly seen on both dye picture and vorticity contours. For the plate
angle of θ=15°, the plate was not long enough to interfere with the
generated shear layer emanating from the lower shoulder of the
cylinder, but increasing the plate angle to a value of θ=30° led to a
direct interaction as seen on the fourth row of Fig. 6a. Similar with the
results of Assi et al. (2009), the plate deflects the separating shear layer
from the cylinder by attaching to the tip of the plate which may stabilize
the cylinder wake. Here, the upper shear layer rolls-up into the wake
region whereas the lower shear layer cannot entrain into the wake since
it is partially blocked by the permeable plate, i.e. some part of the fluid
flows through the meshes and some of the fluid passes round the
permeable plate. It should be pointed out that the amounts of these two
different flows depend on the porosity and the plate angle. With
increasing porosity larger amount of fluid flows through the meshes
due to increase in open area of the permeable plate. Higher amount of
flow is blocked by the permeable plate with increasing plate angle since
it occupies a wider area in cross stream flow. Therefore, deterioration
of the lower shear layer is more dominant on the flow structure for
θ≥30°. The fluid flow through the meshes of permeable plate results in
a uniform momentum injection into the wake region whose effective-
ness changes by the plate angle. Increasing the plate angle to θ=45°
and θ=60° results in a complete suppression of vortex shedding since
the formation of a lower shear layer is significantly blocked by the
permeable plate. This results in the absence of a vortex interaction
between upper and lower shear layers. There exists a weaker shear
layer on the lower side of the cylinder (compared with the upper shear
layer) generated by the flow passing through the permeable plate. This
low-magnitude lower shear layer is insufficient to entrain into the wake
and interact with the upper shear layer. Hence, the upper shear layer
takes the dominant role on the flow structure. It should also be noted
that the jet-alike flow through the meshes of permeable plate is
observed especially for θ=60° which injects momentum into the wake
by generating negative vortices at the rear surface of the plate.
Additionally, the positive vortices are formed at the trailing edge of
the plate, but they cannot entrain into the cylinder wake since they are
blocked by the negative vortices formed by the jets emanating through
the meshes near the trailing edge of the plate. This interaction of
positive and negative vortices might cause a small-scale vortex shed-
ding downstream of the trailing edge of the permeable plate. It can be
concluded that the vortex shedding is eliminated and consequent
formation of Karman Vortex Street is prevented for θ=45° and θ=60°
of β=0.5 in accordance with the TKE results. Furthermore, instead of
large scale Karman Vortex Street, small-scale Kelvin- Helmholtz
vortices are generated along the upper shear layer for all plate angles
and along the trailing edge of the permeable plate for θ≥45°. It should
be noted that even though the latter plate angles are not provided in
42
Ocean Engineering 134 (2017) 35–49
Fig. 6a, based on TKE results and physical interpretations, it is definite
that the vortex suppression mechanism works for the plate angles of
θ=35°, 40°, 75°, 90°.
Fig. 6b includes the instantaneous dye visualization pictures and
vorticity contours for β=0.7. Compared with the β=0.5 presented in
Fig. 6a, one can see that the plate angles of 0°≤θ≤30° show a similar
behavior both qualitatively and quantitatively. As mentioned earlier,
the amount of fluid passing through the meshes depends mostly on the
porosity ratio. Hence, increased porosity would result in higher amount
of momentum injection through the permeable plate. Accordingly, the
recirculation in the wake region formed by the fluid flowing through the
meshes, as well as the upper vorticity layer are shown by the dye
pictures of θ=45° and 60° with the red arrows. The instantaneous
vorticity contours for the plate angles of θ=45° and 60° are consistent
with the qualitative results. Compared with β=0.5, the magnitude of
lower shear layer generated by the fluid passing through the meshes
significantly increases and rolls-up into the wake region. As a result,
interaction of upper and lower vorticity layers exist for β=0.7 resulting
in consequent vortex shedding and Karman Vortex Street downstream
of the cylinder-permeable plate arrangement. However it should also
be pointed out that the vortex street formed by the cases of β=0.7 for
plate angles of θ=45° and 60° covers a smaller region in its width.
Hence, the magnitudes are lower compared to the smaller plate angles
and the plain cylinder. This could be an indication of attenuation of
vortex shedding even though it cannot be suppressed for β=0.7.
The main function of the permeable plate on flow control is to
prevent or attenuate the velocity fluctuations caused by interaction of
shear layers in the near wake. Reducing the wake width is another
indication of flow control downstream of a circular cylinder. The best
way of observing the shear layer interactions and wake width variations
is to analyze the vorticity contours and streamlines, respectively.
Fig. 7a presents the time-averaged streamline topology, < Ψ > super-
posed with the vorticity contours, < ω > for the porosities of β=0.4 and
0.5 and the plate angles of 35°≤θ≤90° which was found to be the most
effective range on reducing the fluctuations. It should be noted that the
reduction of velocity fluctuations can be related to the disappearance of
vortex shedding in the near wake of the cylinder. It can easily be seen
that both porosity values have similar effects on the flow structure of
cylinder- permeable plate arrangements for θ=35°. The previous cases
are not presented to avoid crowding of the figures. It was emphasized
by Fig. 6a that a new shear layer is generated at the trailing edge of the
permeable plate since the sizes of the meshes are narrower for β=0.4
and 0.5 compared with higher porosity values. Therefore, the recircu-
lation region of the cylinder and the corresponding location of the
generated focal and saddle points (shown by the plain cylinder results
in Fig. 4) are significantly affected by the existence of permeable plate.
The two main vortex have an asymmetrical shape and the vortex
formation region elongates in downstream direction due to the
difficulty of the upper shear layer to interact with the lower one and
to enter into the cylinder wake. Furthermore, the width of the wake
slightly increases compared with the plain cylinder as a result of
increased cross-sectional area of the cylinder- permeable plate ar-
rangement. As a general trend, both porosity values qualitatively show
similar behaviors on the flow structure. This is valid almost for all plate
angles. Increasing the plate angle up to θ=60° results in an interesting
and complex flow structure, i.e. the focal points get closer to the
cylinder base while the wake width gets narrower compared with the
latter porosities. The maximum reduction in the wake width exists for
the plate angle of θ=60° which was found to be the best-controlled case
according to the results obtained so far in terms of velocity fluctuation
(Fig. 5) and vortex shedding (Fig. 6). It should be noted at this point
that two independent recirculation regions exist for the plate angle of
θ=60° for β=0.4 porosity. One of them is formed just behind the rear
stagnation point of the cylinder with two focal points, and the other is
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 7. a. Time averaged streamline topology, < Ψ > superposed with the vorticity contours, < ω > for 35°≤θ≤90° plate angles of β=0.4 and β=0.5. b. Time averaged streamline
topology, < Ψ > superposed with the vorticity contours, < ω > for 35°≤θ≤90° plate angles of β=0.6 and 0.7.

43
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 7. (continued)

at farther downstream regions approximately at 4D away from the trailing edge. This situation might also be valid for larger plate angles
cylinder base. This is due to the small mesh sizes of the permeable plate which is probably out of plane of the current flow field. As mentioned
where the trailing edge of the cylinder is solid enough to generate a earlier in Fig. 6a, a small- magnitude negative vorticity layer is formed
shear layer and consequently a new wake exists in farther regions of the by jet-alike flows through the meshes of the screen, especially for plate

44
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

angles of 40°≤θ≤60°. This small- magnitude negative shear layer play

an important role on flow control as well, i.e. the amount of fluid
entrainment from the positive shear layer into the wake region is
limited or prevented by the existence of this negative shear layer.
However, the formation of a new recirculation region as specified by
θ=60° of β=0.4 case is due to the interaction of shear layers on trailing
edge of the permeable plate. For plate angles of θ=75° and 90°, the
recirculation region grows and focal points move away from each other
in both streamwise and transverse directions of the flow. Besides, since
the lower vorticity layer is partially blocked by the permeable plate, the
upper vorticity layer has a dominant role on the flow structure.
Consequently, the upper vorticity rolls up into the wake region
resulting in increased < TKE > max as shown in Fig. 5 for θ≥75°.
As emphasized before, increasing the porosity would result in larger
Fig. 8. Dimensionless vortex formation lengths as a function of plate angles for
mesh sizes. Hence the larger amount of fluid emanates through the 0.6≤β≤0.8.
meshes into the wake region, especially for larger plate angles.
Therefore, it is important to see the effect of greater porosities on flow identical for the plate angles of θ=0° and 15°. For θ≥30°, the values
control by means of shear layer interactions and streamlines. The time- of β=0.8 remain close to the plain cylinder line. However the vortex
averaged streamline topology, < Ψ > superposed with the vorticity formation region is elongated by β=0.6 and 0.7 with increasing plate
contours, < ω > for β=0.6 and 0.7 cases for the plate angles of angle. The largest vortex formation length is obtained as 3.17D for the
35°≤θ≤90° are shown in Fig. 7b. Similar to the results of lower plate angle of θ=90° of β=0.6. It should be pointed out that although
porosities, it is seen that the shear layers elongate in downstream the porosity of β=0.8 was not found to be effective on suppressing the
direction where the upper shear layer is longer than that of the lower vortex shedding, it can be said that this high porosity screen can be
shear layer for all plate angles of β=0.6. The recirculation region for used for the reduction of vortex- induced vibrations due to its effect on
β=0.6 slightly grows up by increasing plate angle while generating two altering the vortex formation length of the cylinder. It can be concluded
well- defined foci and a saddle point for all plate angles presented in that the permeable plates having larger mesh sizes, i.e. higher porosity
Fig. 7b. Formation of these focal and saddle points are the main ratios, can be used with the aid of reducing the vortex-induced
differences when compared with the lower porosity results presented vibrations rather than a complete suppression of vortex shedding
previously in Fig. 7a where there existed higher number of saddle and downstream of the cylinder. Next section of the manuscript deals with
focal points. The elimination of additional focal and saddle points is the complete elimination or reduction of vortex shedding frequencies of
due to the increased permeability of β=0.6 and 0.7. Although the flow the cylinder.
structures are similar for both porosity values, increasing the porosity
to β=0.7 dramatically increases the vorticity magnitudes as seen by the
3.3. Spectral analysis
different vorticity contour legends. Almost 50% increase in the vorticity
magnitude takes place due to the increased open area of the permeable
In this section, the Fast Fourier Transformation (FFT) technique
plates, i.e. sizes of the meshes. This dramatic increase in the vorticity
was used to evaluate the frequency of vortex shedding for various cases.
magnitudes from those evaluated for β=0.4 and 0.5 porosities indicates
Determination of the vortex shedding frequency gives us the frequency
that β=0.6 should be a transition porosity. Therefore, it can be
of vibration on the body due to fluid-structure interaction and hence,
concluded that the porosity of β=0.6 is the transition point for the
the magnitude of vortex induced vibrations (VIV). Therefore it is
plate from low permeable to high permeable, hence the transition from
necessary investigate the effect of permeable plates on reducing or
effective to ineffective flow control of the applied method. This
eliminating the vortex shedding frequency downstream of the cylinder.
conclusion may be supported by the < TKE > max curves presented
Fig. 9 presents the plain cylinder having a dominant frequency of
in Fig. 5 where the β=0.6 curve is in between the β=0.4, 0.5 and β=0.7,
f=0.38 corresponding to a Strouhal number of St=0.2 (St=fD/U∞)
0.8 curves which have corresponded to the effective and ineffective
which is in good agreement with the results of Blevins (1990). A clear
porosities, respectively, according to the results presented so far. The
and distinct peak of frequency shown in Fig. 9 is an indication of the
results of β=0.7 clearly show that the existence of permeable plate is
existence of a dominant vortex structure owing to vortex shedding and
not effective on prevention of the shear layer interactions since the
consequent vibrations of the vortices. Therefore, it is essential to
generated shear layers get closer to each other although their length is
reduce or eliminate the dominant vortex shedding frequency occurring
elongated. Furthermore, the lower shear layer which is formed by the
flow passing through the plate has higher vorticity magnitudes
compared with the lower porosities. It can be concluded that the
interaction of the shear layers is unavoidable in the case of β=0.7,
hence the shedding of the vortices would start at this porosity ratio.
The vortex formation length, evaluated by measuring the distance
between saddle point and the cylinder base is an important factor since
the increase in the vortex formation length would be effective on vortex
shedding control. Vortex formation lengths for the porosities of β=0.4
and 0.5 cannot be calculated since there exists a complex flow structure
for almost all plate angles including more than one focal and saddle
points as can be seen in Fig. 7a. Therefore, the vortex formation lengths
were obtained for 0.6≤β≤0.8 and presented as a function of the plate
angle in Fig. 8. Here, the dashed line represents the vortex formation
length of the plain cylinder which was evaluated as 1.2D away from the
cylinder base. First of all, it is obvious from the figure that all cases are
effective on increasing the vortex formation length compared to the Fig. 9. Spectra of streamwise velocity fluctuations for the plain cylinder at streamwise
plain cylinder. The values obtained for all porosities are almost location of x/d=1.6 and spanwise location of y/d=2.

45
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 10. a. Spectra of streamwise velocity fluctuations for the permeable plate of β=0.5 and β=0.6 at the plate angle of θ=60°. b. Spectra of streamwise velocity fluctuations for the
permeable plate of β=0.7 and β=0.8 at the plate angle of θ=60°.

downstream of the cylinder to eliminate vortex-induced vibrations.
The effects of θ=60° plate angle on suppression of vortex shedding
frequency are presented in Fig. 10a and b for the porosities of β=0.5–
0.6 and β=0.7–0.8, respectively. The spectra of streamwise fluctuations
are evaluated on various locations of field of view in order to find out
the possible vortex shedding frequency. The points are shown on the
figures as 1, 2, and 3 on the upper shear layer; 4, 5 on the lower shear
layer; and 6, 7 on the shear layer formed by the trailing edge of the
permeable plate. The points 1, 2, 3, 4, 5, 6 and 7 correspond to the
coordinates of (x/D, y/D)=(1.6, 2.0), (3.2, 2.0), (4.0, 2.0), (1.6, 1.0),

46
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 10. (continued)

(1.4, 1.0), (1.6, 0.2) and (2.4, 0.2), respectively. It was already found
that the permeable plate is effective on reducing the < TKE >
magnitudes for 0.4≤β≤0.6 and 45°≤θ≤90° and the porosities of
β=0.4 and β=0.5 showed a similar trend for both quantitative and
qualitative means. In addition to these, β=0.0 was not found to be
effective on reducing the < TKE > max for 45°≤θ≤90° since it increases
the wake width due to increased cross section of the cylinder- plate
arrangement. Therefore, FFT is applied by disregarding β=0.0 and 0.4
cases and presented for θ=60° to eliminate the crowding of figures.
Fig. 10a presents the spectra of streamwise velocity fluctuations for the
permeable plate of β=0.5 and β=0.6 at the plate angle of θ=60°. The
results clearly reveal that both porosity values effectively control the

47
G.M. Ozkan et al. Ocean Engineering 134 (2017) 35–49

Fig. 11. The effect of plate angle and porosity on drag coefficient of the cylinder, < CD > .

vortex shedding since a dominant frequency does not emerge for all
points. These FFT results indicate that the use of a permeable plate Table 1
Comparison of the drag coefficient of plain cylinder with available literature results.
having the porosity range 0.4≤β≤0.6 of θ=60° plate angle completely
suppresses the vortex shedding in the near wake of a circular cylinder, Researcher Re Facility Technique < CD >
as well as the vortex-induced vibrations. Fig. 10b shows the spectra of
streamwise velocity fluctuations for higher porosity values, i.e. β=0.7 Current study 5000 Wind Force balance 1.07
tunnel
and 0.8. For the results of β=0.7 there exists a dominant frequency for
Khalak and 5100 Water Force balance 0.98 or
the points 4 and 5 having a value of f=0.31. This frequency corresponds Williamson (1996) channel 1.07
to a Strouhal number of St=0.16 and it is lower than that of the plain Aljure (2015) 5000 – Direct numerical 1.05
cylinder by 20%. Another peak was evaluated at point 1 as f=0.29 with simulation
a lower amplitude. For the other points of β=0.7, no dominancy was Najafi et al. (2016) 5000 Water Integral momentum 1.02 or
channel theorem 1.2
obtained. The last and highest porosity value β=0.8 was ineffective by
both PIV and dye visualization results and it was not considered useful
for flow control. The FFT results of the highest porosity are in which a reduction of 27% is achieved. The drag coefficient increases
consistent with those comments made earlier; the fluctuations of slightly by increasing the plate angle. The general trend of all porosities
streamwise velocities along the upper shear layer (namely the points is similar for θ < 30° where the < CD > values are below the plain
1, 2 and 3) of the cylinder have a frequency value of f=0.32 with cylinder line. Among all permeable plate cases, the largest reduction in
decreasing amplitude in increasing downstream direction. This fre- < CD > is evaluated for β=0.5 porosity having a plate angle of θ=15°
quency value is also evaluated along the lower shear layer of the where a reduction of drag is calculated as 25% compared with the plain
cylinder at points 4 and 5 for a Strouhal number of St=0.17, still 15% cylinder. For 0° < θ < 30°, the values of CD are quite close to each other
lower than that of the plain cylinder. Furthermore, according to the which is found to be the effective interval on reducing the drag
highest open area of β=0.8, almost all incoming fluid passes through coefficient. Increasing the plate angle further to θ=30°, the calculated
the meshes. Therefore, there is no frequency value determined for the < CD > values suddenly jump over the drag of plain cylinder except for
points 6 and 7 which correspond to the trailing edge of the permeable β=0.5 porosity. This increase in < CD > is due to the increased cross-
plate. According to these results one can easily say that although β=0.7 sectional area of the cylinder- permeable plate arrangement where a
and 0.8 are not effective on completely eliminating the vortex shedding, contact between the trailing edge of the plate with the shear layer of the
they are effective on attenuating the shedding frequency. Thus, they cylinder occurs at this plate angle. Thus, total cross sectional area
can be used for detuning the vortex shedding and induced vibrations. increases by increasing plate angle as demonstrated by Fig. 11, the drag
forces grow up slightly for 30°≤θ≤90°. This condition is well-presented
by the results of solid plate where a dramatic increase in < CD > is
3.4. Drag force measurements
obvious by increasing plate angle θ. The < CD> of θ=90° for the solid
plate (β=0.0) is almost 4 times greater than that of a plain cylinder. The
By the results presented in the preceding sections, the effect of
increasing trend is also valid for permeable plate cases, however the
porosity and plate angle on the flow structures were examined. The
growth rates are slightly lower with respect to the solid plate cases. It
influence of the current control technique on drag force is another issue
can be said that the general behavior of increase in the drag force by
regarding the aerodynamic applications in engineering. Fig. 11 pre-
varying plate angle is inconsistent with the definition of porosity. For
sents the calculated drag coefficients of the cylinder with attached
θ > 30° the largest < CD > is evaluated for the smallest porosity value
solid- permeable plates for various porosities and plate angles. The
β=0.4 and the < CD > values slightly decrease by increasing porosity.
dashed line corresponding to a value of < CD > =1.07 shows the drag
This is because of the increase in open area for greater porosities, and
coefficient of the plain cylinder as being the reference for better
thus the lower sized cross- sectional areas for cylinder- permeable plate
comparisons. This value is similar to that found by other researchers
arrangements. It can be concluded by < CD > results that the attached
presented in Table 1.
plates are effective on reducing the drag force of a circular cylinder for
The results obtained for plate angles of θ < 30° reveal that the drag
plate angles of θ < 30° except for β=0.5 which was found to be the most
coefficient decreases almost for all attached plates. It should be noted
effective plate on reducing the drag force for 0°≤θ≤35°.
that the solid plate is the most effective one for θ=0° (splitter plate), in

48
G.M. Ozkan et al.
4. Conclusions
In this study, a passive control method was applied in order to
control the flow downstream of a circular cylinder using attached
permeable plates along its span. The parameters of the study were the
porosity β of the permeable plate and the plate angle θ. In order to
compare with the permeable plate, the effect of a solid plate (β=0.0-
known as splitter plate in literature) was also investigated. The results
revealed that for some effective intervals, the existence of a permeable
plate has two impacts on the flow structure: a) it directly preventing a
vortex roll-up from lower side of the cylinder, b) it injects high amount
of momentum through its meshes which directly prevents the roll-up of
the upper shear layer. Consequently the lower vortex is not formed and
upper vortex is not rolled into the wake region and the vortex shedding
is successfully prevented. According to this physical reality, the
fluctuations significantly decrease in the wake of the cylinder for
porosity ratios of 0.4≤β≤0.6 and permeable plate angles of
45°≤θ≤90°. This might be useful for elimination of the vortex shedding
and consequently Karman Vortex Street as well as Vortex- Induced
Vibrations. A maximum reduction in the peak values of Turbulent
Kinetic Energy in the wake region is evaluated as 74% (compared with
the plain cylinder) for the porosity ratio and plate angles of β=0.4 and
θ=45°, respectively, which is an indication of effective suppression of
the vortex shedding. The porosity ratio of β=0.6 was found to be the
transition point from effective to ineffective on the flow control. On the
other hand, drag force measurements showed that the use of attached
plates for θ < 30° would slightly decrease the drag coefficient compared
with the plain cylinder. Compared with the plain cylinder, 25%
reduction in drag was achieved by the permeable plate of β=0.5 at
θ=15° and 27% reduction in drag was obtained for the splitter plate
case, i.e. β=0.0 at θ=0°.
Within these effective intervals, this study brings out a new
approach on controlling the unsteady flow structure around bluff
bodies and reduction of drag coefficient. Based on these observations
the use of a permeable material could be suggested for engineering
applications concerning the vortex shedding and consequent vortex-
induced vibrations for deep water flow conditions. If a complete
suppression of vortex shedding and induced vibration is needed for a
specific application, permeable plates of 0.4≤β≤0.6 for θ≥30° are
suggested, although they will give larger drag forces exerted on the
body. If a slight reduction in vortex shedding frequency is sufficient and
drag reduction is preferential, then use of higher porosity permeable
plates will solve the problem for plate angles of θ≤30°.
References
Akilli, H., Karakus, C., Akar, A., Sahin, B., Tumen, N.F., 2008. Control of vortex shedding
of circular cylinder in shallow water flow using an attached splitter plate. J. Fluids
Eng.-T ASME 130, 041401.
Aljure, D.E., Rodriguez, I., Lehmkuhl, Perez-Segarra, Oliva, A., 2015. Influence of
rotation on the flow over a cylinder at Re=5000. Int. J. Heat Fluid Flow 55, 76–90.
Al-Salem, K., Oztop, H.F., Kiwan, S., 2011. Effects of porosity and thickness of porous
sheets on heat transfer enhancement in a cross flow over heated cylinder. Int
Commun. Heat Mass 38 (9), 1279–1282.
Anderson, E., Szewczyk, A., 1997. Effects of a splitter plate on the near wake of a circular
cylinder in 2 and 3-dimensional flow configurations. Exp. Fluids 23, 161–174.
Apelt, C.J., West, G.S., 1975. The effects of wake splitter plates on bluff body flow in the
range 104 < Re < 5×104: part 2. J. Fluid Mech. 71, 145–160.
Apelt, C.J., West, G.S., Szewczyk, A.A., 1973. The effects of wake splitter plates on the
flow past a circular cylinder in the range 104 < Re < 5×104. J. Fluid Mech. 61,
187–198.
Assi, G.R.S., Bearman, P.W., 2015. Transverse galloping of circular cylinders fitted with
solid and slotted splitter plates. J. Fluids Struct. 54, 263–280.
Assi, G.R.S., Bearman, P.W., Kitney, N., 2009. Low drag solutions for suppressing vortex-
induced vibration of circular cylinders. J. Fluids Struct. 25, 666–675.
Assi, G.R.S., Bearman, P.W., Tognarelli, M.A., 2014. On the stability of a free-to-rotate
short-tail fairing and a splitter plate as suppressors of vortex-induced vibration.
Ocean Eng. 92, 234–244.
Barman, B., Bhattacharyya, S., 2015. Control of vortex shedding and drag reduction
49
Ocean Engineering 134 (2017) 35–49
through dual splitter plates attached to a square cylinder. J. Mar. Sci. Appl. 14,
138–145.
Bearman, P.W., 1965. Investigation of flow behind a two-dimensional model with blunt
trailing edge and fitted with splitter plates. J. Fluid Mech. 21, 241–255.
Bearman, P.W., 2011. Circular cylinder wakes and vortex-induced vibrations. J. Fluids
Struct. 27, 648–658.
Bhattacharyya, S., Singh, A.K., 2009. Reduction in drag and vortex shedding frequency
through porous sheath around a circular cylinder. Int. J. Numer. Methods Fluids 65,
683–698.
Blevins, R.D., 1990. Flow-Induced Vibration. Van Nostrand Reinhold Company, New
York.
Bruneau, C.-H., Mortavazi, I., 2006. Control of vortex shedding around a pipe section
using a porous sheath. Int. J. Offshore Polar Eng., 16.
Cardell, G.S., 1993. Flow past a Circular Cylinder with a Permeable Wake Splitter Plate
(Ph.D. Thesis). California Institute of Technology, Pasadena, California, USA.
Choi, H., Jeon, W., Kim, J., 2008. Control of flow over a bluff body. Annu. Rev. Fluid
Mech. 40, 113–139.
Durhasan, T., Aksoy, M.M., Pinar, E., Ozkan, G.M., Akilli, H., Sahin, B., 2016. Vortex
street suppression of a circular cylinder using perforated semi-circular fairing in
shallow water. Exp. Therm. Fluid Sci. 79, 101–110.
Feng, L.H., Wang, J.J., 2010. Circular cylinder vortex-synchronization control with a
synthetic jet positioned at the rear stagnation point. J. Fluid Mech. 662, 232–259.
Gabbai, R.D., Benaroya, H., 2005. An overview of modelling and experiment of vortex-
induced vibration of circular cylinders. J. Sound Vib. 282, 575–616.
Gerard, 1966. The mechanics of formation region of vortices behind bluff bodies. J. Fluid
Mech. 25, 401–413.
Geyer, T.F., Sarradj, E., 2016. Circular cylinders with soft porous cover for flow noise
reduction. Exp. Fluids 57, 30.
Gozmen, B., Akilli, H., Sahin, B., 2013. Passive control of circular cylinder wake in
shallow flow. Measurement 46, 1125–1136.
Gu, F., Wang, J.S., Qiao, X.Q., Huang, Z., 2012. Pressure distribution, fluctuating forces
and vortex shedding behavior of circular cylinder with rotatable splitter plates. J.
Fluid Struct. 28, 263–278.
Huera-Huarte, F.J., 2014. On splitter plate coverage for suppression of vortex-induced
vibrations of flexible cylinders. Appl. Ocean Res. 48, 244–249.
Hwang, J.-Y., Yang, K.-S., 2007. Drag reduction on a circular cylinder using dual
detached splitter plates. J. Wind Eng. Ind. Aerodyn. 95, 551–564.
Khalak, A., Williamson, C.H.K., 1996. Dynamics of hydroelastic cylinder with very low
mass and damping. J. Fluid Struct. 10, 455–472.
Kline, S., Mcclintock, F.A., 1953. Describing uncertainties in single-sample experiments.
Mech. Eng. 75, 3–8.
Kumar, R.A., Sohn, C.-H., Bangalore, H.L., 2008. Passive Control of Vortex-Induced
Vibrations: An Overview 1. Recent Patents on Mechanical Engineering, Korea, 1–11.
Najafi, L., Firat, E., Akilli, H., 2016. Time-averaged near-wake of a yawed cylinder. Ocean
Eng. 113, 335–349.
Nakamura, Y., 1996. Vortex shedding from bluff bodies with splitter plates. J. Fluids
Struct. 10, 147–158.
Negretti, M.E., Socolofsky, S.A., Rummel, A.C., Gerhard, H.J., 2005. Stabilization of
cylinder wakes in shallow water flows by means of roughness elements: an
experimental study. Exp. Fluids 38, 403–414.
Norberg, C., 1998. LDV-Measurements in the near Wake of a Circular Cylinder. Advances
in Understanding of Bluff Body Wakes and Vortex-Induced Vibration, Washington
DC.
Ozkan, G.M., Oruc, V., Akilli, H., Sahin, B., 2012. Flow around a cylinder surrounded by
a permeable cylinder in shallow water. Exp. Fluids 53, 1751–1763.
Ozkan, Akilli, H., 2015. Vortex shedding control using a permeable plate located around
the separation point of a circular cylinder, ASME International Mechanical
Engineering Congress and Exposition, (IMECE), 4A-2015.
OZKAN, G.M., AKILLI, H., SAHIN, B., 2013. Effect of high porosity screen on the near
wake of a circular cylinder, In: Proceedings of the EPJ Web of Conferences, 45,
01071.
Ozono, S., 1999. Flow control of vortex shedding by a short splitter plate asymmetrically
arranged downstream of a cylinder. Phys. Fluids 11, 2928–2934.
Pinar, E., Ozkan, G.M., Durhasan, T., Aksoy, M.M., Akilli, H., Sahin, B., 2015. Flow
structure around perforated cylinders in shallow water. J. Fluids Struct. 55, 52–63.
Rashidi, S., Hayatdavoodi, M., Esfahani, J.A., 2016. Vortex shedding suppression and
wake control: a review. Ocean Eng. 126, 57–80.
Roshko, A., 1954. On the drag and shedding frequency of two dimensional bluff bodies.
Natl. Advis. Comm. Aeronaut. (NACA).
Roshko, A., 1955. On the wake and drag of bluff bodies. J. Aeronaut. Sci. 22, 124–132.
Sheng, J., Meng, H., Fox, R.O., 2000. A large eddy PIV method for turbulence dissipation
rate estimation. Chem. Eng. Sci. 55, 4423–4434.
Shukla, S., Govardhan, R.N., Arakeri, J.H., 2009. Flow over a cylinder with a hinged-
splitter plate. J. Fluids Struct. 25, 713–720.
Unal, M.F., Rockwell, D., 1988. On vortex formation from a cylinder, Part 2: control by
splitter-plate interference. J. Fluid Mech. 190 (513), 529.
Williamson, C.H.K., 1996. Vortex dynamics in the cylinder wake. Annu. Rev. Fluid Mech.
28, 477–539.
Zdravkovich, M.M., 1981. Review and classification of various aerodynamic and
hydrodynamic means for suppressing vortex shedding. J. Wind Eng. Ind. Aerodyn. 7,
145–189.
Zhang, H., Weiping, S., 2016. Numerical simulation of flow over a circular cylinder with a
splitter plate near a moving wall. Ocean Eng. 122, 162–171.