See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318135445

POD Analysis in the Wake of Wall-Mounted Cylinders

Conference Paper · June 2017

DOI: 10.2514/6.2017-3129

CITATION READS
1 161

4 authors, including:

Henrique Fanini Leite Leandra Abreu

Instituto Tecnologico de Aeronautica Instituto de Aeronáutica e Espaço
16 PUBLICATIONS   6 CITATIONS    6 PUBLICATIONS   13 CITATIONS   

SEE PROFILE SEE PROFILE

Ana Cristina Avelar

Instituto de Aeronáutica e Espaço
53 PUBLICATIONS   95 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Shock wave View project

All content following this page was uploaded by Henrique Fanini Leite on 04 July 2017.

The user has requested enhancement of the downloaded file.

 POD Analysis in the Wake of Wall-Mounted Cylinders

Henrique Fanini Leite 1 and Leandra Isabel de Abreu2

Instituto Tecnológico de Aeronáutica (ITA), São José dos Campos, 12228-900, SP, Brazil

Ana Cristina Avelar 3

Instituto de Aeronáutica e Espaço (IAE), São José dos Campos, 12228-904, Brazil

and

André V. G. Cavalieri 4
Instituto Tecnológico de Aeronáutica (ITA), São José dos Campos, 12228-900, Brazil

The flow patterns over a finite square cylinder of aspect ratio H/d = 3 were analyzed
experimentally. The measurements were carried out in a subsonic wind tunnel using the technique of
Time Resolved Particle Image Velocimetry (TR-PIV). The near wake flow structures and vortex
shedding characteristics were investigated using instant velocity maps and proper orthogonal
decomposition (POD). Tests were performed at a speed of 20m/s, resulting in Re = 65300, with the
cylinder facing the flow at 90°. For the wind tunnel tests, the cylinders were fixed on flat plate,
creating a boundary layer which interacted with the cylinder wake. The 2D PIV measurements were
conducted at three horizontal planes (z/H = 0.3, z/H = 0.5, z/H = 1) and the symmetry x-z plane. Due
to the complexity of the phenomena, the flow was characterized both in terms of average behavior
and time-resolved velocity fields. Both symmetrical and anti-symmetrical vortices structures occur in
the cylinder wake, which can be identified based on the coefficients of the first four POD modes. The
results indicated that the alternating Karman vortex structures are dominant, described by the two
first POD modes. However, this structure is sometimes suppressed, leading to periods of symmetrical
vortex shedding.

I. Introduction

T HE flow around cylinders has been a subject of investigation since the beginning of fluid mechanics. Starting
from D’Alembert’s Paradox, this type of flow has eluded scientists due to its simple geometry and complex flow
structure, which presents many patterns, depending on Reynolds number, ranging from quasi-potential flow for very
low Reynolds, to complex vortex-shedding mechanisms at higher Reynolds values. The wake behind 2-D cylinders
has been extensively investigated, both experimentally and through numerical simulations and, while some
questions remain, most of its aspects are well understood, at least at the empirical level1. However, in most practical
applications cylinder structures have one free end and the other end attached to a surface, often with aspect ratios
(H/d, defined as the ratio between height and diameter) lower than 10, and therefore cannot be considered 2-D.
Examples of such applications are the prediction of drag and buffeting on airplane antennas, vortex induced
vibrations in columns and buildings, landing gear noise generation and pollutant dispersing in chimneys.
Despite an increasingly large body of research on wall mounted cylinder wakes, there are still many points of
contend regarding the characterization of this type of flow2. Due to the influence of the free end and the wall, there
are significant differences between the wake of finite cylinders and two dimensional cylinders 2-5. Pioneering studies
have shown the type of wake to depend strongly on the cylinder’s aspect ratio. Quasi-periodical, Karmann-like
vortex shedding predominates up to a critical value (H/d)cr, below which it is suppressed by the tip induced

1
PhD Candidate, Instituto Tecnológico de Aeronáutica, leite05@hotmail.com.
2
PhD Candidate, Instituto Tecnológico de Aeronáutica, leandraabreu13@gmail.com.
3
Senior Engineer Researcher, Instituto Tecnológico de Aeronáutica, anacristina.avelar@gmail.com.
4
Associate Professor, Instituto Tecnológico de Aeronáutica, andre@ita.br.
1
American Institute of Aeronautics and Astronautics
downwash3,4,6. While some authors describe the existence of an arch-type symmetrical vortex shedding pattern4-6,
which is favored by lower aspect ratios and the upwash and downwash induced by the base and tip streamwise
vortices, others have only identified the anti-symmetric pattern1,3,7. For circular cross section cylinders, values for
(H/d)cr range from 2.5 to 6, depending on boundary layer thickness3,4,8. For a square cross section, (H/d)cr ~ 24.
Moreover, both symmetrical and anti-symmetrical types of shedding were observed at aspect ratios well above and
below critical, with spanwise variations in type predominance within the same cylinder5,9,10.
The vortex emission Strouhal number follows a power law of (H/d)4, while also correlating with boundary layer
thickness. The Strouhal number peak in the spectrum tends to spread at lower aspect ratios, something which was
also observed in regular cylinders11. For (H/d) below a critical value, the power law slope changed, which has been
attributed to a different vortex shedding behaviour 4. In Ref. 5, the shedding frequencies were measured throughout a
(H/d) = 7 cylinder span and it was found that the peak occurs at the same frequency at all locations, even 10
diameters downstream. The authors suggested that this result implied that the tip and base vortices may separate
from the cylinder at the same frequency as spanwise vortices, and each of the three types of vortices may be part of
the same entity of the organized structure. For (H/d) > 7, however, spanwise shedding frequency may vary in a
cellular fashion12-14. In Ref. 15, the authors have found a secondary frequency peak associated with the counter
rotating vortices formed at the free end, which is rapidly suppressed downstream.
In terms of streamwise vortices, two main structures can be identified. Two streamwise vortices form at the free
end, the so-called tip vortices, which arise from the coupling of the tip induced downwash and the upwash along the
side wall3,5,16. In addition to the tip vortices, a second pair of streamwise vortices has been detected originating at the
tip surface, where oil visualization indicated a swirling pattern17. Ref. 6, however, has not found evidence of this
second pair in cylinders of (H/d) = 1. When the only dominant streamwise vortices are formed at the tip, the wake is
called dipole-like. In cylinders with (H/d) above a certain value, which depends strongly on boundary layer
thickness, there is another pair of streamwise base vortices originating at the junction between the cylinder and the
wall7,18,19. These wakes are called quadrupole-like. Ref. 2 has provided a different hypothesis, according to which
the streamwise vorticity is the result of the upstream bending of the spanwise vortex structure, which seems to
contradict other authors3,6, who detected tip vortices even without organized spanwise vortex emission. The
spanwise vortex structures are significantly influenced by the shear layer thickness. In Ref. 7, the wakes in a (H/d) =
8 square cylinder with natural and tripped boundary layers were investigated. With an increase in boundary layer
thickness, the streamwise vortex arrangement changed from dipole-like to quadrupole-like. Correspondingly, the
spanwise shedding structure changed from a half-loop to a full-loop one. Ref. 19 investigated three boundary layer
thickness in a (H/d) = 5 square cylinder and has shown that the probability of anti-symmetric vortex shedding at
(z/d) = 1 was reduced from 84% to 34.5% with an increase in boundary layer thickness from 0.07d to 0.245d. At
(z/d) = 4, the trend was opposite, with the probability of anti-symmetric shedding going from 19.5% to 46.5%.
These results indicate a strong association between induced upwash or downwash and spanwise vortex shedding.
There is no consensus on how the streamwise vortices interact with the spanwise vortex sheet. Based on phase-
averaged data on a series of PIV instant velocity fields of the wake of a (H/d) = 4 square cylinder, Ref. 2 has
proposed the existence of connector strands, which are formed by the upstream bending of vertical vortices due to
the free end shear layer and that connect to the anteceding structure shed on the opposite side of the wake. Other
authors have presented similar results, with dipole wakes forming half-loop structures and quadrupole wakes
forming full-loop structures7, numerical simulations in cylinders of (H/d) = 25 and (H/d) = 50, have resulted in
similar structures near the tip of the cylinders20. These models do not take into account symmetrical shedding, since
the phase averaging technique did not indicate any dominant symmetrical shedding structure2,7. Ref. 5 has proposed
a model where spanwise vortices from both sides of the cylinder are connected with each other near the free end,
forming an arch-type vortex. This arch-type vortex structure has two legs that join each other near the cylinder free
end, that is, they form a full loop structure similar to the one proposed by Ref. 7. Under the influence of the free-end
downwash flow and the boundary layer over the wall, both the upper and the lower parts of the arch-type structure
are inclined upstream, accounting for the tip and base vortices. In addition, they postulate the existence of a
horseshoe vortex near the base due to observations from other authors5. The legs can be either symmetrically
arranged or staggered, with no significant distinction between the structures5. While the model proposed by Ref. 2
ignores symmetrical vortex shedding, the model proposed by Ref. 5 does not explain how unified shedding
structures can lead to different proportions of symmetrical and anti-symmetrical shedding throughout the cylinder
span. Other authors propose models where the streamwise vortices are not connected to the spanwise vortex
street3,16,21.
This work aims to investigate experimentally the flow patterns on a (H/d) = 3 square cylinder by means of
instant-velocity fields visualization and POD analysis, with a particular interest in understanding the nature of the
symmetrical vortex shedding and its relationship with other flow structures. It provides a different approach to flow
2
American Institute of Aeronautics and Astronautics
analysis, in which one or more POD modes are associated with a particular flow structure, and the value of the
coefficients of such modes is used as a guide to select snapshots where the structure they are associated with is
predominant. Thus, it is possible to identify both symmetrical and anti-symmetrical vortex shedding, as well as
periods of mixed predominance. Their dynamics and inter-relationship could also be investigated. Moreover, by
reconstructing the flow using only a few selected POD modes, it is possible to elucidate the many flow dynamics
behind observed behavior. The square cylinder has the advantage of a fixed separation point, which simplifies the
problem. The PIV measurements were conducted at three horizontal planes (z/H = 0.3, z/H = 0.5 and z/H = 1) as
well as the symmetry x-z plane.

II. Material and Methods

A. Experimental Measurements
The tests were carried out at the TA-3 wind tunnel, which is a closed-circuit wind tunnel with an open test
section. The maximum speed through its empty test section is around 38m/s and the turbulence level is 2%. The
open test section facilitated both optical and physical access to the model, allowing for greater flexibility in
experiment configuration. Four different cylinder models were tested, two with circular cross section, 40mm
diameter and aspect ratios of 3 and 6, and two with square cross section, 40mm side and aspect ratios of 3 and 6.

Figure 1. Cylinder models in the test section.

Here, the analysis of flow features in the wake of the cylinders will be conducted just for the mean velocity of
20m/s for the square cross section cylinder and aspect ratio AR = 3. The Reynolds number for the experiment was
Re = 65,300.

Horizontal and vertical 2D velocity fields were measured

using time-resolved PIV. In the PIV measurements, the light
sheet was generated by a Litron pulsed Nd:YLF Laser with
10mJ of energy per cavity, at 1kHz. The whole flow was
seeded with poly-ethylene glycol smoke using a high volume
liquid seeding generator of Nova Instruments. A 105mm
Nikkor lens with f#5.6 was used on a SpeedSense 9020 CMOS
camera, 1kHz, with full resolution of 1152 x 896 pixels. A
light guide arm from Dantec Dynamics was used to conduct the
laser beam, and, consequently, the light sheet to the region of
interest. DynamicStudio software V3.2, also from Dantec, was Figure 2. Schematics the x-z vertical PIV
used for data acquisition and processing. plane and the z/h = 0.5 horizontal plane
A total of 663 images at a frequency of 500Hz were taken
for each condition. This was the maximum allowed by the
camera internal memory. Velocity fields were determined by adaptive correlation, a method with higher
computational cost, but which is able to provide better results. It iteratively optimizes the size and shape of each

3
American Institute of Aeronautics and Astronautics
interrogation area to better adapt to local flow gradients and seeding densities, thus optimizing correlation conditions
at each interrogation window22. Window interrogation size was 32x32 pixels and 50% of overlap was considered.
The PIV resolution was 2,84mm, or 0.142d for the x-y horizontal planes and 3.01mm, of 0.150d for the x-z vertical
plane.
A total of 4 different planes were measured: three horizontal and one vertical. The horizontal planes were at z/h
= 0.3, z/h = 0.5 and z/h = 1. The analysis of different horizontal planes allows for investigating the influence of the
tip-induced downwash in the wake. Figure 2 illustrates the x-z vertical and the x-y z/h = 0.5 horizontal plane.

B. Proper Orthogonal Decomposition

The proper orthogonal decomposition (POD) is a quantitative method that has been increasingly applied to
instantaneous whole-filed measurement techniques, such as PIV, in order to identify coherent structures. The POD
technique provides a basis for the modal decomposition of a set of data into an ensemble of functions, called POD
modes. The first modes concentrate the most energy, and represent the largest and most coherent structures of the
flow. Therefore, if the flow dynamics presents only a few predominant structures, the data can often be represented
using just some of the first modes. This method has been successfully used in describing various types of flows,
such as mixing layers, jets, channel flows and cylinder wakes. The use of POD method in the PIV data has
previously been reported by Ref. 10 and Refs. 23-25.
The snapshot POD method was the selected approach to analyze the experimental data. In this method, each
snapshot corresponds to an instantaneous PIV data of the flow field. The present snapshot POD follows the
procedure outlined by Ref. 24.
The first step is to calculate the mean velocity field and subtract it from each snapshot, so that the analysis is
done only to the fluctuation parts of the velocity components , where u and v denote the fluctuation velocity
in streamwise and lateral directions, the index n runs through the N snapshots, and i runs through the M positions of
velocity vectors in a given snapshot.
All fluctuating velocity components from N snapshots are arranged in a matrix U:

Thus, the autocovariance matrix Cis created:

A set of N eigenvalues and corresponding set of eigenvectors which satisfy:

The eigenvalues are ordered by decreasing values as .The normalized POD modes are
constructed from the projection of the eigenvectors :

The discrete 2-norm is defined as:

4
American Institute of Aeronautics and Astronautics
 Each snapshot can be expanded in a series of POD modes with expansion coefficients for each . The
coefficients also called POD coefficients, are determined by projecting the fluctuating part of the velocity field
onto the POD modes:

Where . Then, the expansion of the fluctuation part of a snapshot n is obtained from:

The decreasing ordering of the eigenvalues and eigenvectors ensures that the most important modes in terms of
energy are the first modes. This usually means that the first modes will be associated with large scale flow
structures. If a flow has dominant flow structures, these are therefore reflected in the first POD modes and hence a
given snapshot can often be reconstructed satisfactorily using only the first few modes. More details on the POD can
be found in Ref. 26. The snapshot POD was conducted using MATLAB.

III. Results and Discussion

A. POD Results

All time-resolved velocity fields were analyzed by Proper Orthogonal Decomposition as a way of identifying
coherent structures within the flow. To investigate the dependence of the number of snapshots, the analysis has been
repeated for 200 snapshots. The results were qualitatively the same of those based on 663 snapshots, but some
random variations appeared, like slightly asymmetrical flow structures. The precision of the POD based on 663
snapshots was found to be satisfactory for the present analysis. The boundary layer momentum thickness was 0.05d.

1. Bottom Horizontal Plane z/h = 0.3

The eigenvalues ( ) and its cumulative sum representing the energy of each POD mode are shown in Figure 3.
POD modes 1 and 2 are significantly stronger than the others. They together represent about 53% of the energy in
velocity fluctuations whereas modes 3 and 4 together represent only about 7%.

Figure 3. POD eigenvalues (left) and cumulative energy from each POD mode.

5
American Institute of Aeronautics and Astronautics
 Figure 4 provides a spectral analysis of the first four POD modes coefficients. Both modes 1 and 2 present a
sharp peak at 50.6 Hz, corresponding to Sr = 0.10. This frequency is typical of the Von-Karmann vortex shedding in
3-D square cylinders, and matches both experimental and numerical data for cylinders of similar aspect ratios 2,27.
Mode 3 presents a
peak near at 0.76Hz,
but due to the small
amount of snapshots,
this frequency has a
low degree of
certainty. Mode four
does not present a
clear peak at any
frequency.
Figure 6 shows
the first four POD
modes of u and v for
a 2D square cylinder
wake in a horizontal
plane at z/h = 0.3.
Only the flow behind
the cylinder was Figure 4. Spectral Analysis of the modes coefficients.
presented. The
cylinder itself is
immediately to the left of the image, it was not included because of the presence of spurious vectors in the shade it
cast, which contaminated the POD analysis.
Figure 5 provides the mean flow computed with velocity fields that presented only a) negative or b) positive
values of the mode 3 coefficients. Although this does not allow us to associate this mode with any particular
vorticial structure, it becomes clear that positive values for mode 3 correlate with a stronger reverse flow and a
larger recirculating region. In it, the cylinder, in grey is seen from above, and the black area represents the shade it
cast.
The contour structures of Mode 1 and 2 are highly correlated. Particularly, Mode 2 can be interpreted as a
streamwise shift of Mode 1, with the phase difference about 1/4 shedding circle. For Modes 1 and 2, the contours
with positive and negative
values present an alternating
pattern, the contours of are
anti-symmetrical about the
wake centerline, which
inherently resembles the
alternating Von Karman
vortex structure. The vortices
are convected downstream
and the first two POD modes
describe two phases of the
convective motion. The
circular distribution of the a1
and a2 coefficients, shown in
Figure 7, suggests a cyclic
variation of POD modes 1
and 2, which is exactly what Figure 5. Average flow computed with velocity fields presenting a) negative
is expected if two POD and b) positive mode 3 coefficients.
modes describe different
phases of a tender process creating convecting vortices. The a1 and a2 coefficients change in a fashion similar to
cosine and sine functions, respectively, in a time-resolved study of the flow. The random variations from a simple
circle in Figure 7 accounts for the influence of turbulence in terms of both local velocity fluctuations and
irregularities in the creation of the vortices in the wake. Another factor that supports the idea that modes 1 and 2

6
American Institute of Aeronautics and Astronautics
correspond to a Von-Karmann shedding
pattern is the oscillatory nature of its
coefficients, as presented in Figure 9. The
alternation between negative and positive
values corresponds to the emission of
clockwise and anti-clockwise vortices.
The pattern of Modes 3 and 4 are different
from those of the first two modes, as shown in
Figure 6. The contours of are symmetrical
about the wake centerline. The overall
contribution of modes 3 and 4 to the total
fluctuation energy is limited. This correlates
well with previous reports, which indicate a
predominance of the Von-Karmann type
vortex shedding in square cylinders of aspect
ratio above h/L = 2.4
Based on the energy distribution for the
POD modes, around 60% of the fluctuation
energy is described by the first four modes.
Figure 8 provides examples of snapshots and
their reconstructions using just the four first
modes, showing that all the main flow
structures are accurately reconstructed, while
small scale, random fluctuations are
eliminated. It is interesting to observe how
different combinations of just four modes can
lead to a great variety of velocity fields. This,
in turn, indicates that even small variations in
the relative weight of each mode can lead to
important changes in flow behavior. This
observation is crucial to elucidate the many
flow structures detected by instant velocity
field visualization.
One must bear in mind, however, that the
analysis of POD modes energetic
contributions takes into account the whole
time length of the experiment. As shown in
Figure 9, the instant values of the POD Figure 6. The first four POD modes for the wake at z/h = 0.3
coefficients vary considerably, indicating
important time-dependent variations in flow 1.5
behavior, as will be discussed in section B.
The instant values of each POD coefficient 1
can be used as a guide to find the vortex
structures they relate to. While it is clear that
0.5
modes 1 and 2 relate to a Karman-like vortex
structure, the nature of the two other most
a2

energetic modes can only be determined by a 0

deeper analysis, which includes all the

measured planes, as well as instant velocity -0.5
fields. An important characteristic of those
modes is that they present lower frequency -1
oscillations and, therefore, are probably not
directly associated with the Von-Karmann
-1.5
vortex sheet. Another interesting feature is -1.5 -1 -0.5 0 0.5 1 1.5
how the peak in the mode 3 coefficients a1
corresponds to lower values for modes 1 and Figure 7. Distribution of the first and second POD modes.
7
American Institute of Aeronautics and Astronautics
2, a pattern that is observed in all three horizontal planes.

Figure 8. Snapshot (left) and reconstruction (right) using the first four POD modes.

Figure 9. POD coefficients for each of the 663 snapshots at z/h = 0.3

8
American Institute of Aeronautics and Astronautics
 2. Middle Horizontal Plane z/h = 0.5

Results for z/h = 0.5 were remarkably similar to the z/h = 0.3 plane. Figure 10 presents the eigenvalues ( ) and
its cumulative sum representing the energy of each POD mode. In comparison with the z/h = 0.3 plane, the energy
of modes 1 and 2 is lower, together representing around 40% of all fluctuations energy. Conversely, mode 3 now

Figure 10. POD eigenvalues (left) and cumulative energy from each POD mode, z/h = 0.5.

represents nearly 8% of fluctuations energy and mode 4 around 4 %. This trend is important, as it suggests that
those modes, especially mode 3, suffer some influence from the tip induced downwash.
The mode shapes are the same observed at z/h = 0.3, and will not be presented in here for brevity. The similarity
between the modes of the two planes clearly indicate a region of coherent Karman vortex shedding, although there is
not enough data to determine the full extent of this region.
The fact that modes 3 and 4 have also remained unchanged is evidence that they represent organized
fluctuations, and not random disturbances. The results for the upper horizontal plane, presented in the next section
provide significant insight into mode 3.

3. Upper Horizontal Plane (z/h = 1)

The POD modes for the z/h = 1 plane were the same ones detected at the other horizontal planes. However, the
energy level associated with each one has changed considerably.
Figure 11 illustrates the modes for this plane. Mode 1 of the z/h = 1 plane is equivalent to the Mode 3 of the
other two horizontal planes, whereas the Mode 3 is the former Mode 1, albeit slightly modified due to the influence
of the main flow fluctuations and possibly the tip induced downwash. Mode 2 is similar to the ones detected on the
previous planes, but instead of a vortex, there is a source and a sink point, each at one side of the wake. From
previous results, it was already clear that this mode was associated with von-Karman vortex shedding. In this
particular case, it represents the general flow undulation induced by the Karmann vortex sheet. No Karmann-like
shedding is expected at the cylinder tip2,3,5, which is confirmed by the fact that the mode typically associated with
vortex shedding in the other planes, Mode 3, has a lower energy level than mode 2, associated with the vortex
induced undulation. Thus, both modes represent not local vortex shedding, but the influence of the shedding process
that is happening at lower planes in the tip wake.
9
American Institute of Aeronautics and Astronautics
 This is reflected by their lower energy levels.
Figure 12 provides the cumulative energy for each
POD mode. It can be seen that modes 2 and 3 together
represent only about 20% of total fluctuations energy,
compared to 53% for the equivalent modes at the z/h =
0.3 plane.
The very opposite happens with mode 1, which had
a much lower energy level at lower planes, and now
represents about 30% of total fluctuations energy. This
indicates that this mode is associated with the tip
induced downwash. More precisely, it is associated
with variations in the strength of the reverse flow
within the wake, as shown in figure 5 for the z/h = 0.3
plane, induced by variations in the tip-induced
downwash. It is likely that this mode is the horizontal
plane component of the downwash variations
represented by the mode 1 of the x-z plane, which will
be described in the following sections. It presents a
low-frequency oscillating pattern, which can be
detected both by analyzing the mode 1 coefficients for
the z/h = 1 horizontal plane and the POD results for
the vertical plane. Ref 10 detected the same shift in
mode importance throughout the span, observing a
remarkable similarity between the symmetrical mode 1
for z/d = 6 and mode 3 for z/d = 3.5 in a H/d = 7
square cylinder. However, he did not associate this
mode with fluctuations in the reverse flow strength or
suggested any correlation with the tip induced
downwash.
Moreover, a strong expression of this mode shape
in the z/h = 0.3 and z/h = 0.5 planes is related to a
suppression of the anti-symmetrical modes, and instant
velocity field visualization indicate that it also
corresponds to periods of symmetrical vortex
emission. When considering the results presented in
Figure 5, which indicates that this mode is associated
with a stronger reverse flow and larger recirculation
region, our findings seem to contradict the
observations of Ref. 4 which states that wakes with a
predominance of symmetrical vortices were much
smaller than its anti-symmetrical counterparts.
However, this statement was made when comparing
cylinders with different aspect ratios, while our
observations apply to different wake structures formed
Figure 11. The first four POD modes for the wake at z/H = 1
by the same cylinder.
The cumulative energy for the first four modes is
about 50%, close to the results for the z/h = 0.5 plane and lower than the 60% of the z/h = 0.3 plane, indicating a
slightly larger role for less organized velocity fluctuations within the wake.

10
American Institute of Aeronautics and Astronautics
 Figure 12. POD eigenvalues (left) and cumulative energy from each POD mode at z/h = 1.

The POD mode 4 was the same detected on all other horizontal planes. Again, this indicates that this mode
represents an organized structure, although with a much lower energy level. In contrast with the other 3 modes, its
energy has not been significantly affected by the plane height. Be it z/h = 0.3, z/h = 0.5 or z/h = 1, it has remained
with around 4% of total fluctuations energy, which can be verified in Figure 12. This indicates the tip induced
downwash has no influence on the type of fluctuations it represents, a behavior coherent with shear layer vortices
resulting from the lateral flow detachment. The same is true for the base vortex, since if this phenomenon had a role
in mode 4 fluctuations, its energy would also be expected to change as z/h increased. Therefore, the effect of the tip
induced downwash and base induced upwash is more in suppressing the Karman vortex sheet than in fostering
symmetrical vortex formation, which retains a constant share of fluctuations energy throughout the whole span.

4. The vertical Plane

Figure 13. POD eigenvalues (left) and cumulative energy from each POD mode at the x-z plane.
11
American Institute of Aeronautics and Astronautics
 Compared to the three horizontal
planes, vertical plane fluctuations are
remarkably less organized. Figure 13
presents the POD eigenvalues and their
cumulative energy. The first four
modes represent only 30% of total
fluctuations energy, compared to 50 -
60% for the horizontal planes. Mode 1
has about 15% of the energy, mode 2
slightly more than 6% and the
remaining two about 4% each.
Despite their low relative energy
levels, all four vertical modes display
clear peaks at the spectrum, presented
in figure 14. Mode 1 has a very low
frequency peak at Sr = 0.0045
(2.26Hz), but since there was a limited
number of samples, there is a high
degree of uncertainty to this value. Figure 14. Frequency spectrum of the x-z plane first four POD modes
Mode 2 has a peak at the same Sr =
0.0045 (2.26Hz) and a secondary one
spread between Sr = 0.20-0.22 (110-
100Hz). This is the same frequency of
mode 3, whereas mode 4 has very weak
peak at Sr = 0.056 (27.9 Hz). Since all
four first modes are associated with
frequency peaks at the spectrum, even
if weak ones, they must be coherent to
a certain degree. Interestingly, the
frequency associated with modes 2 and
3 is double the Karmann vortex
shedding frequency, which might
indicate a connection between those
modes and the spanwise vortex
emission. Mode 2 presents a peak at
2.26 Hz because mode 1 clearly
suppresses it, as can be seen in figure
15. When mode 1 coefficients are
positive, mode 2, and to a certain extent Figure 15. POD mode coefficients for the x-z simmatry plane
mode 3, has coefficients near zero,
while for negative mode 1 coefficients mode 2
presents a high frequency oscillatory pattern.
In figure 17 the mode shapes are represented.
The cylinder was not included in the image, but it is
immediately to the left, in a vertical position. Mode
1 is associated with an up-down motion in the
vertical wake. This low frequency fluctuation was
observed visually during the experiments. Regions
with positive mode 1 coefficients are associated
with longer recirculation regions, a weaker
downwash and stronger reverse flow. It is
associated with mode 1 of the z/H = 1.0 horizontal
plane and mode 3 of the z/H = 0.3 and z/H = 0.5
planes, which are related with variations in the Figure 16. Reconstruction of the flow using only the
reverse flow. These fluctuations are possibly related mean flow and modes 2 and 3.

12
American Institute of Aeronautics and Astronautics
 to instabilities in the incoming boundary layer, since
there were no tripping mechanisms and small
variations in boundary layer thickness can have
strong effects into the wake, maybe triggering a
transition from a dipole to a quadrupole type7,19. In
light of that, it is important to understand to what
structures the other modes are related, as well as
how they interact with mode 1.
By reconstructing the velocity fields using just
the mean flow and modes two and three allows us
insight into the structures they are related to. This
was done in figure 16, and a second, clockwise
vorticial structure becomes clearly visible. The
vortex forms at mid-span and propagates diagonally,
moving downstream and in direction of the plate.
This type of vortex shedding occurs frequently
during the periods when mode 1 is not predominant.
The same procedure was done using only mode
4, which is presented in figure 18. Thus, mode 4 is
visibly associated with an anti-clockwise vortex
which forms near the wall and quickly propagates
downstream.
The vorticial structures presented in figure 16
and figure 18 often pair with each other, but they
have also been observed independently. These
probably are not the mushroom-like longitudinal ribs
described by Ref. 28 and observed by Ref. 5 in a
(H/d) = 7 square cylinder, since the clockwise and
anti-clockwise vortices have different shedding
frequencies.
The emission frequency presented by modes 2
and 3 suggest an association with the spanwise anti-
symmetric vortex shedding. If those vortices are the
projection of the “connector strands” created by the
upstream bending of the spanwise vortices legs, one
would expect a vortex in the x-z plane for each
spanwise vortex shed, all of them clockwise due to
vortex bending. Therefore, the projection of the
connector strands in the x-z plane would appear as a

Figure 17. First 4 POD modes at the x-z simmetry plane.

series of clockwise vorticial structures shed with double the
frequency of the spanwise vortices, exactly as observed. The
initial position of the vortices, near mid-span, is close to the
results presented for a (H/d) = 4 cylinder2. The downwards
movement of this vortex is also coherent with a connector
strand structure, since it connects the bottom of one principal
core to the top of the principal core of the downstream
opposing side shed structure.
Figure 18. Flow reconstruction using the mean flow The simplest explanation for the apparent suppression of
and mode 4. modes 2 and 3 by the mode 1 is that the formation region of
the vortices they represent is convected downstream, outside
the image plane, but the shedding mechanism itself remains.
13
American Institute of Aeronautics and Astronautics
Image 19 presents the average
flow taken from negative and
positive mode 1 coefficient
values. That is, to create the
mean flow depicted in Figure
19a, only velocity fields with
negative mode 1 coefficients
were computed and vice versa
for Figure 19b. Positive values
are associated with a longer
recirculation region, stronger
backflow, but a weaker
downwash, while negative
values are associated with a
shorter reverse flow region and
stronger downwash. Therefore,
Image 19. Flow averaging for a) negative and b) positive mode 1 coefficients
this mode is probably
correlated with the
symmetrical mode 3 of the z/H = 0.3 and z/H = 0.5 horizontal planes. Periods of strong expression of mode 3 in
those planes are, in turn, associated with symmetrical vortex shedding. Thus, the absence of structures possibly
associated with the “connector strands” might be another evidence for a different, symmetrical shedding mechanism,
or at least the suppression of the anti-symmetrical shedding pattern.
Ref. 19 has shown that, in a H/d = 5 square cylinder, thicker boundary layers favored the formation of base
vortices, which, in turn, greatly increased the proportion of symmetric vortices shed at z/h = 0.2, while also
marginally increasing this proportion for z/h = 0.5. Unfortunately, while there is strong support for the linking of the
x-z mode 1 and z/H = 0.3 and z/H = 0.5 mode 3 (and mode 1 at z/H = 1), there is no information regarding
instantaneous boundary layer thickness, which makes it impossible to confirm if these two types of flow behavior
are caused by variations in the boundary layer.
In comparison, the vorticial structure associated with mode 4 is not completelly supressed by a strong expression
of mode 1. It represents an anti-clockwise vortex that forms near the wall and quickly propagates downstream.
Although there were much less vorticial structures associated with mode 4 during periods of positive mode 1
coefficients, they were observed nevertheless. Its shedding frequency is apparently not related to any of the other
flow structures, and there is no evidence of interaction of this vortices with any other, with the exception that in flow
visualization sometimes this structure is observed preceding the clockwise vortices associated with modes 2 and 3,
although both structures were also observed being shed independently. Despite a weak peak at Sr = 0.056 in the
spectrum, flow visualization revealed a highly variable shedding pattern, with periods of regular emission and
periods of subduing. The anti-clockwise sense of rotation of these structures is coherent with a second connector
strand in a “full-loop” vorticial structure7. However, full-loop structures were related with quadrupole streamwise
vortex distributions, which, if present in our case, would be associated with positive mode 1 coefficients due to the
reduced downwash, whereas most vortex shedding associated with mode 4 was observed during periods of negative
mode 1 expression. Moreover, there is a significant frequency mismatch between the spanwise vortex sheet and
vortex emission associated with mode 4. Therefore, the emission pattern associated with mode 4 appears to be
independent, arising from a roll up of the reverse flow, maybe with the insertion of lateral flow in the wake.
Considering that a reverse flow approaching the cylinder upstream is qualitatively similar to the main flow
approaching the cylinder downstream, the formation of the vorticial structures associated with mode four might be
qualitatively similar with the formation of the horseshoe vortex. The presence of the cylinder causes an adverse
pressure gradient, which causes the incoming flow to separate and roll-up, forming a vortex, which is then
convected downstream.

B. Time-averaged and Instant Velocity Fields

Time-averaged velocity fields for the AR3 square cylinder are presented in Fig. 20. Streamlines are only to
facilitate interpretation, and do not provide information regarding velocity magnitude. Color maps represent
vorticity (“S”), in (1/s). Scales vary for each image. The gray area represents the cylinder, while the black one
represents shade regions that could not be reached by the laser plane. At the bottom right, the 20m/s reference vector
is represented.
14
American Institute of Aeronautics and Astronautics
 Of the vortices represented in Fig. 20, only the vertical one, Fig. 20a, represents a stationary structure. Another
important information is that horizontal vortex core positions do not change considerably from the middle to the
bottom plane. This suggests that the influence of the free end induced downwash is equally felt at both positions, or
that the horizontal flow mechanisms predominate over it. At the top plane, no vortex structures can be noticed,
which suggests that the main flow predominates. As noted in a number of publications, no coherent spanwise vortex
emission is expected at the tip2,3,5. Overall, the mean velocity fields suggest the existence of a region dominated by
the Karman vortex sheet, which engulfs both the z/H = 0.3 and the z/H = 0.5 horizontal planes, and a region near the
tip where there is no coherent vortex shedding. In addition, a separation bubble could be detected at the tip in the x-z
vertical plane, which is induced by the cylinder’s sharp edge.
The horseshoe vortex could not be detected, since they would induce vertical velocity variations at the edges of
the cylinder, undetectable by the 2D PIV at horizontal planes, and far from the x-z symmetry plane shown in Fig.
20. Despite being present in most attempts at modelling the cylinder wake3,5, it is a small scale29 standalone feature
that does not interact with the vortex shedding mechanisms or the streamwise vorticial structures5.
Figure 21 presents instant horizontal velocity fields at z/h = 0.3. These were taken at cautiously selected

(a) (b)

(c) (d)

Figure 20. Time-averaged velocity fields and vorticity maps for: (a) vertical x-z plane;(b) bottom
horizontal plane (z/H=0.3); (c) tip horizontal plane (z/H=1) and (d) middle horizontal plane (z/H=0.5).

moments, corresponding to the peaks in certain values of the POD mode coefficients. Figure 21b corresponds to
image number 553, it exhibits a predominance of modes 1 and 2, which represent Karmann vortex shedding. In it,
this structure is clearly predominant. Figure 21a corresponds to image 301, which corresponds to a peak in the value
15
American Institute of Aeronautics and Astronautics
of POD mode 3. This sort of behavior was observed both at z/H = 0.3 and z/H = 0.5. Periods of strong mode 3
expression correlate with symmetrical vortex shedding, indicating a clear association between them.

(a) (b)

Figure 21. Instant Velocity fields for an AR3 square section cylinder, z/h =0.3 left: (a) symmetric vortex
shedding, right: (b) anti-symmetric vortex shedding.

By combining the information obtained through POD analysis of the four planes with instant flow visualization,
the following comments can be made:

1. The relative importance of the symmetrical mode 3 increases with z/H. While for the two lower horizontal
planes this mode concentrates much less energy than fluctuations related with the Von Karmann vortex
sheet, for z/H = 1 it becomes the most energetic POD mode (and thus becomes mode 1 at this plane). This
indicates strong influence of the tip induced flow in this mode.
2. Flow averaging using only snapshots with positive mode 3 coefficients indicate that this mode is associated
with a larger reverse flow region.
3. Positive values for the horizontal mode 3 also correlate with the suppression of the von-Karman vortex
emission.
4. At the x-z symmetry plane, periods of weak POD mode 1 expression correlate with vorticial structures
related with the “connector strands” structure, which is associated with anti-symmetrical vortex shedding.
Conversely, a strong mode 1 expression suppresses those structures.
5. The other important effect of the vertical plane mode 1 is a stronger reverse flow, which is also observed for
mode 3, supporting a connection between them.
6. Due to the absence of any structures related with anti-symmetrical vortex shedding during periods of strong
horizontal mode 3 expression and of the vertical mode 1, the predominance of these modes is associated
with the suppression of the Von Karmann vortex sheet and symmetrical vortex shedding.
7. Finally, the relative energy of the horizontal mode 4 was constant throughout the span. This mode is also
symmetrical, and probably is associated with the symmetrical vortex formation mechanism, while mode 3
has more to do with the suppression of the more organized von Karmann vortex shedding.

The results presented in Figure 21 indicate that the attempt to determine a critical aspect ratio – dependent on
boundary layer thickness and incoming turbulence intensity - after which vortex shedding behavior turns from
symmetrical to anti-symmetrical is not an accurate way of describing the flow. Although the symmetrical modes
were shown to contain less than 7% of flow energy, they still play an important part in flow behavior, even
predominating over other modes at times. In reality, both symmetrical and anti-symmetrical mechanisms co-exist,
each predominating at different moments. It remains unclear what makes each mode predominant, or if this
16
American Institute of Aeronautics and Astronautics
phenomenon is part of a greater pattern. Unfortunately, not enough data was collected to investigate if there is a
defined low-frequency pattern for the predominance of each mode.
In addition to those clear vortex shedding patterns, at times complicated patterns emerged, corresponding to
images where the three first POD modes coefficients were of comparable magnitude. Figure 22 exemplifies this:

(a) (b)

Figure 22. Instant Velocity fields for an AR3 square section cylinder, z/h =0.3 left: (a) a Karmann-like
structure with multiple vortex cores; right: (b) strong symmetrical vortices.

Figure 22a, corresponding to image 544 of the series, was taken at a point where the coefficients of all three
modes were negative and had a relatively low magnitude. It is possible to identify a Karmann vortex-shedding
pattern, but the upstream vortex has two cores, both rotating clockwise (which means there is some sort of saddle
point between them, or a very small anti-clockwise rotating region), whereas a third, secondary anti-clockwise
vortex structure is present at the bottom, slightly upstream from the two other cores. Downstream, there is a well-
defined vortex, which rotates anti-clockwise and forms a pair with the two-cored vortex structure. The low
magnitude of the first POD coefficients means that higher order modes contribute to a certain extent to flow
behavior, resulting in a higher amount of vortex cores and higher overall flow complexity. Figure 22b corresponds
to image 484. In it, the three first POD coefficients were of a similar magnitude to the ones in figure 22a, but this
time the third POD mode coefficient was positive. Figure 22b, however, does not present many vortex cores.
Instead, two big symmetrical vortices dominate the flow. An analysis of previous images shows that this is a
snapshot of a mixed type vortex-shedding pattern, in which the alternating Karman vortices are created from the
merger of small vortices.
Figure 23 shows eight consecutive snapshots of the flow immediately before the one presented in Figure 22b.
This is an upper view of the cylinder. The gray area represents the cylinder, seen from above, and the black region
represents the area the laser beam could not reach due to the shadow cast by the cylinder. It is possible to identify a
Karmann type of vortex shedding, but with some peculiarities. The upper four images describe the formation
process of a clockwise rotating vortex. This vortex is clearly formed from the merger of many smaller vortices. In
the first image, there are three vortices similar to those seen in Figure 21a, which merge in the second frame. In the
third fame, another vortex is formed at the upper side of the cylinder, while at the lower side there is already an anti-
clockwise rotating vortex. In frame four, the new upper-side vortex is merging with the big clockwise vortex,
whereas a series of complex vortex structures are formed at the lower side. Unfortunately, the vortices at the bottom
are not clear, but, judging by the last four snapshots, the same process happens at the lower side.
The process described in figure 23 is a combination of anti-symmetrical and symmetrical vortex shedding
patterns, with strong predominance of the former. This result indicates that the anti-symmetrical vortex sheet is
formed by the merger of many symmetrical vortices which are constantly being formed by the shear layer created by
the lateral flow detachment. This process was described by Ref. 28 for circular cylinders of a much larger aspect
ratio. The shear layer vortices formation is associated with the horizontal POD mode 4, which represents 4% of total
17
American Institute of Aeronautics and Astronautics
fluctuations energy in all horizontal modes. During periods of strong mode 3 expression, the formation of the Von-
Karmann vortex sheet is suppressed, but the symmetrical shear layer vortices remain, thus corresponding to the
periods of “symmetrical vortex shedding”. In reality, there is no coherent symmetrical shedding mechanism, but
only periods where the coherent anti-symmetrical shedding is suppressed, allowing for the visualization of
symmetrical vortices which haven’t merged into big anti-symmetrical ones. In regions near the tip and the wall,
down and upwash flows increase the likelihood of disruption of the anti-symmetrical vortex sheet, thus increasing
the amount of symmetrical vortices, while the energy associated with the shear layer vortices remains constant
throughout the span.

Figure 23. The vortex shedding pattern in a region of mixed POD mode predominance.

The mechanism which leads to the suppression of the anti-symmetrical vortex shedding remains unclear. POD
mode analysis has shown that it is associated with the tip induced flow. Periods of suppression are related to weaker
downwash and stronger reverse flows. These effects must be felt differently throughout the span to account for the
different symmetrical and anti-symmetrical shedding proportions19.
Total data acquisition time was slightly more than one second, and a series of different flow conditions could be
visualized, indicating either that small-scale variations in flow conditions can trigger strong variations in shedding
patterns or that a lower frequency pattern governs which type of flow behavior predominates. To elucidate the
questions that arise from our findings, bigger data sets acquired during longer periods would be of great value.

IV. Concluding Remarks

The flow patterns over a square cylinder with aspect ratio AR =3 were analyzed in three horizontal planes,
z/H = 0.3, z/H = 0.5 and z/H = 1, and at the x-z symmetry vertical plane. Following a POD analysis, time-averaged
and instant velocity fields were used to identify flow structures of interest. In the two lower horizontal planes, the
POD modes 1 and 2 were shown to be consistent with Von-Karman vortex structures, and contain more than half of
the wake’s energy. Mode 2 shows a flow structure displaced about a quarter of the “vortex width” compared to
mode 1. Modes 3 and 4, although representing a much lower share of the flow’s energy, were shown to play an
important part in determining flow behavior. At the tip horizontal plane, the structure associated with mode 3 on the
lower planes displayed a much larger share of the total fluctuations energy. This structure is related with variations
in the tip induced flow. A strong expression of this mode correlated with symmetrical vortex shedding, which was
detected by instant-velocity fields visualization, but it does not correspond to an organized symmetrical shedding
structure, but rather to the suppression of the anti-symmetrical shedding mechanism. Mode 4 presented a constant
share of the fluctuations energy throughout the span and is related to shear vortices arising from the lateral flow
detachment. The analysis of sequential instant-velocity fields revealed that the anti-symmetrical vortices are formed
by the merger of many shear layer vortices, and that the symmetrical vortex emission corresponds to periods were

18
American Institute of Aeronautics and Astronautics
the formation of anti-symmetrical vortices is suppressed, allowing the symmetrical shear vortices to propagate
downstream. Instances of complex vortex shedding, with many small-scale vortices were also identified. In the x-z
vertical plane, evidence of the “connector strands” was detected, but those structures disappeared when mode 1
coefficients were positive. A secondary anti-clockwise vorticial structure was also observed, but this structure seems
unrelated to the other observed mechanisms.

Acknowledgments
The Authors thank the Conselho Nacional de Pesquisa e Desenvolvimento(CNPq) for the financial support
through the grants: 402233/2013-1, 308829/2015-8 – DT and the process 381094/2016-2.

References
1 Williamson, C. H. K. “Vortex dynamics in the cylinder wake.” Annu. Rev. Fluid Mech., Vol. 28, 1996 pp. 477–539.
2 Bourgeois, J. A., Sattari, P. and Matinuzzi, R. J., “Alternating half-loop shedding in the turbulent wake of a finite surface-

mounted square cylinder with a thin boundary layer”. Physics of Fluids, Vol. 23, Sep, 2011, pp. 1-16. Paper no. 095101-1.
3Kawamura, T., Hiwada, M.,Hibino, T.,Mabuchi, I., Kumada, M., 1984. “Flow around a finite circular cylinder on a flat

plate”. Bulletin of JSME, Vol. 27, No. 232, 1984, pp. 2142 – 2151.
4 Sakamoto, H., Arie, M. “Vortex shedding from a rectangular prism and a circular cylinder placed vertically in a turbulent

boundary layer”, Journal of Fluid Mechanics, Vol. 126, 1983, pp. 147–165.
5 Wang, H., Zhou, Y. “The finite-length square cylinder near wake”, J. Fluid Mech., Vol. 638, Nov., 2009, pp. 453-491.
6 Pattenden, R. J., Turnok, S. R., Zhang, X. “Measurements of the flow over a low aspect ratio cylinder mounted on a ground

plane”. Experiments in Fluids, Vol. 39, No. 1, 2005 pp. 10-21.

7 Hosseni, Z., Bourgeois, J. A. and Martinuzzi, R. J., “Wall-mounted finite cylinder wake structure modification due to

boundary layer-wake interaction: Half-loop and full-loop coherent structure topologies,” Proceedings of The Seventh
International Colloquium on Bluff Body Aerodynamics and Applications (BBAA7), Vol 1, International Association for Wind
Engineering, Shanghai, 2012.
8 Okamoto, T., Yagita, M. “The experimental investigation on the flow past a circular cylinder of finite length placed normal

to the plane surface,” Bull. JSME, Vol. 16, 1973, pp.805–814.

9 Leite, H. F., Avelar, A. C., De Abreu, L. I., Schuch, D. “Pod And Spectral Analysis Of A Wall-Mounted Square Cylinder

Wake” Proceedings of EPTT, Vol. 1, ABCM, Rio de Janeiro, 2016.

10 Wang, H. “POD analysis of the finite-length square cylinder wake,” Proceedings of The Seventh International Colloquium

on Bluff Body Aerodynamics and Applications (BBAA7), Vol 1, International Association for Wind Engineering, Shanghai, 2012.
11 Okamoto, T. & Sunabashiri, Y. “Vortex shedding from a circular cylinder of finite length placed on a ground plane,” J.

Fluids Engng, Vol. 114, 1992, pp. 512–521.

12 Adaramola, M.S., Akinlade, O.G. Sumner, D., Bergstrom, D.J. and Schenstead, A.J., “Turbulent wake of a finite circular

cylinder of small aspect ratio,” Journal of Fluids and Structures, Vol 22, 2006, pp. 919-928.
13 Sakamoto, H., Oiwake, S. “Fluctuating forces on a rectangular prism and a circular cylinder placed vertically in a turbulent

boundary layer,” J. Fluids Engng, Vol. 106, 1984, pp. 160–166.

14 Uematsu, Y., Yamada, M., Ishii, K. “Some Effects Of Free-Stream Turbulence On The Flow Past A Cantilevered Circular

Cylinder,” Journal of Wind Engineering and Industrial Aerodynamics, Vol. 33, 1990, pp. 43-52.
15 Park, C. W., Lee, S. J. “Flow structure around a finite circular cylinder embedded in various atmospheric boundary layers,”

Fluid Dynamics Research, Vol. 30, 2002, pp. 197–215.

16 Etzold, F. and Fiedler, H., “The near-wake structure of a cantilevered cylinder in a crossflow,” Z. Flugwiss, Vol. 24, 1976,

pp. 77–82.
17 Roh, S.C. and Park, S., “Vortical Flow Over The Free End Surface of a Finite Circular Cylinder Mounted on a Flat Plate,”

Experiments in Fluids, Vol. 34, 2003, pp. 63-67.

18 Tanaka, S. and Murata, S. “An investigation of the wake structure and aerodynamic characteristics of a finite circular

cylinder,” JSME Intl J. Ser. B: Fluids Therm. Engng, Vol. 42, 1999, pp. 178–187.
19 Wang, H. F., Zhou, Y., Chan, C. K. & Lam, K. S. “Effect of initial conditions on interaction between a boundary layer and

a wall-mounted finite-length-cylinder wake,” Phys. Fluids, Vol. 18, No. 6, 2006, 065106.
20 Levold, P., “Viscous flow around finite length circular cylinder”. Master of Science Degree thesis, Norwegian

University of Science and Technology, Trondhein, Norway, 2012.

21 Wang, H. F., Zhou, Y. & Chan, C. K. “Flow around a finite length square prism,” Proceedings of the Fourth European and

African Conference on Wind Engineering Institute of Theoretical and Applied Mechanics, Academy of Sciences of the Czech
Republic, Prague, 2005.
22 Wang, G. “PIV experiment and CFD simulation of flow around cylinder”. Master of Science Degree thesis, University of

Texas, Austin, Texas, USA, 2015.

23 Graftieaux, L., Michard, M. and Grosjean, N., “Combining PIV, POD and vortex identification algorithms for the study of

unsteady turbulent swirling flows,” Meas. Sci. Technol., Vol 12, 2001, pp. 1422–1429.

19
American Institute of Aeronautics and Astronautics
 24 Meyer, K. E., Pedersen, J. M. and Ozcan, O., 2007.“Turbulent jetin crossflow analyzed with Proper Orthogonal

Decomposition.” J. Fluid Mech, Vol. 583, 2007, pp. 199-227.

25 Pedersen, J. M. and Meyer, K. E., “POD-analysis of flow structures in a scale model of a ventilated room,” Exps. Fluids,

Vol. 33, No. 6, 2002, pp. 940–949.

26 Holmes, P., Lumley, J. L. and Berkooz, G. Turbulence, coherent structures, dynamical systems and symmetry Cambridge

University Press, Cambridge, United Kingdom, 1998.

27 Wang, Y. Q., Jackson, P., Sui, J., “Simulation of flow around a surface-mounted square-section cylinder of aspect ratio

four,” Proceedings of the 20th Annual Conference of the CFD Society of Canada, Ottawa, Canada, 2011.
28 Wu, J., Sheridan, J., Soria J. and Hourigan, K., “Shear Layer Vortices and Longitudinal Vortices in the Near Wake of a

Circular Cylinder,” Experimental Thermal and Fluid Science, Vol. 12, 1996, pp.169-174.
29 Ho, T. Lin, C. and Dey, S., “Characteristics of steady horseshoe vortex system near junction of square cylinder and base

plate,” J. Engng Mech, Vol. 134, 2008, pp. 184–197.

20
American Institute of Aeronautics and Astronautics

View publication stats