Göktun, 1975 PDF

Calhoun: The NPS Institutional Archive
DSpace Repository
Theses and Dissertations 1. Thesis and Dissertation Collection, all items
1975-12
The drag and lift characteristics of a cylinder

placed near a plane surface
Geöktun, Selâahattin
Monterey, California. Naval Postgraduate School
http://hdl.handle.net/10945/20826
Downloaded from NPS Archive: Calhoun

THE DRAG AND LIFT CHARATERISTICS OF A
CYLINDER PLACED NEAR A PLANE SURFACE
Selahattin Goktun
L POSTGRADUATE 80 OL
Monterey, California
T
THE DRAG AND LIFT CHARACTERISTICS
OF A
CYLINDER PLACED NEAR A PLANE SURFACE
by
Selahattin Goktun
December 1975
Th esis Advisor: T. Cooper
Approved for public release; distribution unlimited.
T 1 7nnoc
UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE (Whan Data I Entered)
REPORT DOCUMENTATION PAGE READ INSTRUCTIONS

BEFORE COMPLETING FORM
1. REPORT NUMBER 2. GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER
4. TITLE (and Subtitle) 5. TYPE OF REPORT A PERIOD COVERED

Master's Thesis;
The Drag and Lift Characteristics of a December 1975
Cylinder Placed Near a Plane Surface
€. PERFORMING ORG. REPORT NUMBER
7. AUTHORf*,) • - CONTRACT OR GRANT NbM8ER(-»J

Selahattin Goktun
9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT. PROJECT, TASK
AREA A WORK UNIT NUMBERS
Naval Postgraduate School
Monterey, California 93940
11. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE
Naval Postgraduate School December 1975
Monterey, California 93940 13. NUMBER OF PAGES
114
14. MONITORING AGENCY NAME & ADDRESS,'!/ dtlferent Irom Controlling OiUce) IS. SECURITY CLASS, (ot thla riport)
Unclassified
Ma. DECLASSIFl CATION/' DOWN GRADING
SCHEDULE
16. DISTRIBUTION ST ATEMEN T (o( (fits Roport)
Approved for public release; distribution unlimited.
17. DISTRIBUTION STATEMENT (ol <h» abstract entered In Block 30, It dlttarent from Report)
18. SUPPLEMENTARY NOTES
19. KEY WORDS (Continue on reverse alda It naceaeary and Identity by block number)
Drag Coefficient
Lift Coefficient
Cylinder, Right Circular
20. ABSTRACT (Continue on reverse tlda It nacattaiy and Identity by block number)
Surface pressure, drag and lift coefficients have been
experimentally determined for a right circular cylinder
located near a plane surface and placed in a cross flow of air.
Parametric studies were carried out for Reynolds number
varying from 90,000 to 250,000, three plate lengths and a
variety of cylinder to plate spacings The variation of the .
drag coefficient as a function of gap size was found to

exhibit an interesting and unexpected trend. The drag was a
DD
(Page
,
F
JA r73
1)
1473 EOITION OF NOV 65
1
S/N 0102-014- 6601 |

IS OBSOLETE UNCLASSIFIED
SECURITY CLASSIFICATION OF THIS PAGE (*hen Data Kntaratt)
UNCLASSIFIED
fliCURlTY CLASSIFICATION OF THIS P»rJ Emii.n Hrla Enf.r.cf
(20. ABSTRACT Continued)
minimum when the cylinder was resting on the plate and was
a maximum at a gap size of approximately one cylinder radius.
Flow visualization studies together with detailed measurements
of the vortex shedding frequency in the cylinder wake indicate
that the plate interferes with the formation of the vortex
street in the cylinder wake when it is located within a
cylinder radius of the cylinder. This interference disturbs
the cylinder base pressure which in turn influences the
magnitude of the drag coefficient.
DD Form 1473
, Jan 73
1 UNCLASSIFIED
S/N 0102-014-G601 SECURITY CLASSIFICATION OF THIS P A G E fT** •" Dmtm Enfrmd)
The Drag and Lift Characteristic
of a
Cylinder Placed Near a Plane Surface
by
Selahattin Goktun
Lieutenant, Turlcish Navy
B.S.M.E., Naval Postgraduate School, 1975
Submitted in partial fulfillment of the

requirements for the degree of
MASTER OF SCIENCE IN MECHANICAL ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOL

December 1975
senook
MONTEREY
ABSTRACT
Surface pressure, drag and lift coefficients have been
experimentally determined for a right circular cylinder
located near a plane surface and placed in a cross flow of
air. Parametric studies were carried out for Reynolds
numbers varying from 90,000 to 250,000, three plate lengths
and a variety of cylinder to plate spacings. The variation
of the drag coefficient as a function of gap size was found
to exhibit an interesting and unexpected trend. The drag
was a minimum when the cylinder was resting on the plate and
was a maximum at a gap size of approximately one cylinder
radius. Flow visualization studies together with detailed
measurements of the vortex shedding frequency in the cylinder

wake indicate that the plate interferes with the formation of
the vortex street in the cylinder wake when it is located
within a cylinder radius of the cylinder. This interference
disturbs the cylinder base pressure which in turn influences
the magnitude of the drag coefficient.

TABLE OF CONTENTS
I. INTRODUCTION 12
II. BACKGROUND 16
III. EXPERIMENTAL APPARATUS 18
A. WIND TUNNEL 18
B. PRESSURE COEFFICIENT EXPERIMENT 20
C. FLOW VISUALIZATION EXPERIMENT 22
D. VORTEX FREQUENCY EXPERIMENT 25

'
IV. EXPERIMENTAL PROCEDURE 28
A. PRESSURE COEFFICIENT EXPERIMENT 29
B. FLOW VISUALIZATION EXPERIMENT 30
C. VORTEX FREQUENCY EXPERIMENT 30
V. RESULTS AND DISCUSSION 33
A. PRESSURE EXPERIMENT AND CYLINDER

SURFACE PRESSURE DISTRIBUTIONS 33
B. PRESSURE DRAG AND LIFT COEFFICIENTS — 51
C. FLOW VISUALIZATION RESULTS 56
D. VORTEX FREQUENCY RESULTS 66
VI. SUMMARY 69
APPENDIX A. PRESSURE DATA AND DATA REDUCTION 70
APPENDIX B. CALCULATION OF PRESSURE DRAG AND

LIFT COEFFICIENTS 74
APPENDIX C. UNCERTAINTY ANALYSIS 75
APPENDIX D. POLAR PLOTS OF PRESSURE COEFFICIENT

VERSUS ANGULAR LOCATION 80
LIST OF REFERENCES 113
INITIAL DISTRIBUTION LIST 114

LIST OF ILLUSTRATIONS
1. Schematic diagram of plate-cylinder arrangement 14
2. Schematic diagram of the wind tunnel 19
3. Photograph of the cylinder used to collect

pressure data 21
4. Photograph of the plate and cylinder arrangement

in the wind tunnel 23
5. Photograph of the plate and cylinder arrangment

in the water channel 26
6. Sketch of a typical subcritical flow pattern

on the surface of a cylinder 36
7. Sketch of a typical critical flow pattern

8. Sketch of a typical supercritical flow pattern

9. Pressure coefficient and local Nusselt number

on the surface of a free cylinder for
Re = 90,000 37

on the surface of a free cylinder for
Re = 153,000 38
11. Pressure coefficient on the surface of a free

cylinder for Re = 250,000 39

on the surface of a cylinder placed near a
plane surface for: Re = 153,000, L/D =4.0,
d/r = 0.0 40

plane surface for: Re = 153,000, L/D = 4.0,
d/r =0.25 41

d/r = 0.50 42
d/r = 1.0 43

d/r =2.0 44

d/r = 3.0 45

d/r = 4.0 46

plane surface for: Re = 153,000, L/D - 4.0,
d/r = 5.33 47
20. Pressure drag and lift coefficients variation

as a function of plate length and gap size for
a Reynolds number of 153,000 54
21. Pressure drag and lift coefficients variation

as a function of gap size and Reynolds numbers
for L/D =4.0 55
22. Photograph of the flow pattern at d/r =0.0 58
23. Photograph of the trapped vortices at d/r =0.0 — 59
24. Photograph of the flow pattern at d/r = 0.2 5 61
25. Photograph of the flow pattern at d/r = 0.5 62

d/r = 0.0 81

d/r = 0.25 82
d/r = 0.5 83

d/r = 1.0 84

d/r = 2.0 85

d/r = 3.0 86

d/r = 4.0 87

plane surface for: Re = 9 0,000, L/D = 4.0,
d/r = 5.33 88
36. Pressure coefficient on the surface of a

cylinder placed near a plane surface for:
Re = 250,000, L/D = 4.0, d/r = 0.0 89

Re = 250,000, L/D = 4.0, d/r = 0.25 90

Re = 250,000, L/D =4.0, d/r =0.5 91

Re = 250,000, L/D = 4.0, d/r = 1.0 92

Re = 250,000, L/D = 4.0, d/r = 2.0 93

Re = 250,000, L/D = 4.0, d/r = 3.0 94
Re = 250,000, L/D = 4.0, d/r = 4.0 95

Re = 250,000, L/D = 4.0, d/r = 5.33 96

d/r =0.0 97

d/r = 0.25 98

d/r =0.5 99

d/r = 1.0 100

d/r = 2.0 101

d/r = 3.0 102

d/r = 4.0 103

d/r = 5.33 104

d/r = 0.0 105
d/r = 0.25 106

d/r = 0.5 107

d/r =1.0 108

- d/r = 2.0 109

plane surface for: Re = 153,000, L/D =8.0,
d/r =3.0 110

d/r =4.0 111

d/r = 5.33 112
10
ACKNOWLEDGMENT
I would like to express my sincere thanks and
appreciation to my thesis advisor, Professor Thomas E.
Cooper, whose professionalism, knowledge, and encouragement
were inspirational to me throughout this entire endeavor.
Also, I would like to extend my appreciation to Professor
Paul F. Pucci, who provided me with many helpful suggestions
regarding the finalization of this work.
11
I. INTRODUCTION
In this study, the flow field, drag and lift charac-
teristics of a cylinder placed in close proximity to a plane
surface in a cross flow of air were investigated. Previous
work of this nature has concentrated primarily on free

cylinders.
This project is a follow-on study to an investigation
originated by McComas [1] and continued by Gnerlich [2].
In both of these previous studies, primary emphasis was
devoted to obtaining the heat transfer characteristics of
a cylinder placed in proximity to a plane surface. Although

surface pressure data were obtained in these studies, these
data were not analyzed in detail. The present investigation
has focused on obtaining drag and lift information on the
cylinder as a function of the size of the gap between the
cylinder and plate. A range of Reynolds numbers varying
from 90,000 to 250,000 was investigated and the length of
plate ahead of the cylinder was varied from 2.0 to 8.0
cylinder diameters. Also, the flow field surrounding the
cylinder was studied in detail with the aid of a flow
visualization technique and with a hotwire anemometer.
The present experiments were conducted using the
Mechanical Engineering Department wind tunnel located in
Building 500 and the water channel located in Building 234
on the Naval Postgraduate School campus. The wind tunnel,
12
.
described in detail in Section III, has a speed range of
15 to 250 ft/sec, and is equipped with a traversing mechanism
and hotwire anemometer which can be used to obtain velocity
information throughout the test section. The hotwire was
also used to measure the vortex shedding frequency behind
the cylinder.
Experiments were conducted in the wind tunnel at
uncorrected free stream Reynolds numbers of 90,000, 153,000,
250,000 for plate to cylinder spacings of d/r = 0.0, 0.25,
0.5, 1.0, 2.0, 3.0, 4.0, 5.33, and 9.33 (no plate case)
where d = gap size and r = cylinder radius (See Figure 1)
For these studies the plate length ahead of the cylinder was
four cylinder diameters (L/D = 4.0) . Experiments were also
performed at a free stream Reynolds number of 153,000 with
the same d/r ratios but with plate leading edge lengths of
2.0 and 8.0 cylinder diameters (L/D =2.0 and 8.0).
In calculating and plotting the pressure drag coefficient
as a function of gap spacing an interesting and unexpected
trend was revealed. Over the range of Reynolds numbers
investigated, it was found that the minimum pressure drag
coefficient occurred when the plate was attached to the
cylinder. Further, these curves also indicated a maximum
drag coefficient at d/r = 1.0. The lift coefficient was
found to be a maximum when the plate and cylinder were in
contact and continuously decreased as the gap size was
increased. In order to provide a qualitative explanation
13
p
c
<u
B
Q)
Cr>
C
td
n
5-1
c
•H
iH
>i
O
I
Q)
-P
(d
rH
cu
I
n
-rH
u
•H
-p
rd
e
O
CO
U
H
En
14
for the existence of a maximum in the pressure drag
coefficient curve at d/r = 1.0 , a flow visualization
experiment was conducted in the water channel. In this
water channel (See Section III for detailed description)
a 14 inch long plexiglass flat plate and 1.5 inch diameter
hollow plexiglass cylinder were used. The system was
scaled for geometric similitude with the experimental
apparatus in the wind tunnel. The cylinder was equipped
with three dye ports and the space between the cylinder
plate was varied using a lifting device which was attached
to the plate. Due to limitations in the water pump, the
maximum Reynolds number which could be attained with this
apparatus was approximately 10,000.
The flow visualization revealed very clearly that in the
range _< d/r < 1 the plate was interfering with the
formation of the vortex street that forms in the wake of a
free cylinder. The surface pressure data indicated that
this interference produced a significant effect on the
cylinder base pressure and an attendant effect on the
cylinder drag. Hotwire anemometer measurements of the vortex
shedding frequency were also made in the cylinder wake and
tend to substantiate this claim.
15
II. BACKGROUND
The present investigation of the flow field around a
cylinder placed in proximity to a plane surface was conducted
concurrently with an investigation by Kosemen [3] of the heat
transfer characteristics of a heated cylinder located near a
plane surface. Information obtained from the flow experi-
ments proved useful in explaining the heat transfer trends
observed by Kosemen. Likewise, Kosemen' s measurements
provided insight into a better understanding of the flow
field.
Data relating to the drag and heat transfer characteris-
tics of the cylinder-plate geometry under investigation were
not found in the literature. Such information would be of
obvious value in the design of heated piping systems located
near walls and in heat exchanger design.
While numerous investigations have been conducted upon

free cylinders having continuous surfaces and diverse shapes,
few studies have been made of the flow patterns associated
with a plate-cylinder system. Roshko [4] conducted an
experiment using a splitter plate located at various locations
behind a circular cylinder. He concluded that vortex dynamics
were indeed important and that any interference with their
formation would have a strong influence on the base pressure
and drag of the cylinder. By placing a splitter plate along
the centerline of the wake, it was found that the normal
16
.
periodic vortex formation was inhibited and the base pressure
was found to be increased considerably. Roshko's investiga-
tion was very informative and provided insight into our
investigation.
Another study which provided insight into our problem
was an investigation by Spivak [5] of the characteristics of
vortex formation in the wake behind a pair of cylinders
located at various gap spacings in a cross flow. Spivak
found:
a) When two cylinders located in cross flow are separated
by a gap just smaller than a cylinder diameter, an instability
in the wake vortex pattern occurs.
b) At larger gap sizes the cylinders behave like independent
bodies
c) At spacings less than one-half diameter a low gap vortex
shedding frequency is observed. Between one-half and one
diameter a high gap vortex shedding frequency is present
decreasing to the independent cylinder value at a critical
spacing of one cylinder diameter.
d) At very small gap sizes individual von Karman vortex
streets are not detached from each cylinder; instead, a
coupled vortex street is formed by the overall system along
with certain internal vortices that depend upon the gap.
The internal vortices are generated in vortex sheets
bordering a jet that emanates from the orifice formed by the
adjacent sides of the two cylinders. Spivak 's investigation
was very informative and provided significant insights into
our investigation.
17
III. EXPERIMENTAL APPARATUS
The wind tunnel, water channel, test cylinders and
plates and their mounting are discussed in this section.
A. WIND TUNNEL
All data were collected in a subsonic, open circuit,
wind tunnel manufactured by the Kenny Engineering Corporation
of Monrovia, California. This tunnel was powered by a 75 HP,
1750 rpm variable-speed motor driving a 45 inch diameter
adjustable pitch, vaneaxial fan (blower) capable of supplying
air at 10,000 to 70,000 cubic feet per minute. The tunnel
was equipped with six graded-mesh screens to control turbu-
lence and a 24-inch manometer tube tapped into the stilling
chamber and test section for velocity measurement. The
tunnel was designed to operate at velocities ranging from
approximately 15 ft/sec to 250 ft/sec. The turbulence
intensity in the clear tunnel was measured as 0.2%.
The wind tunnel consists of the inlet section, test
section (28" high x 20" wide x 6' long) , diffuser and blower.
The blower pulls air through the inlet past the test section.
After passing the test section, the air enters the diffuser
and then the blower. The blower then expels the air to the
atmosphere. A schematic of the wind tunnel with pertinent
dimensions is shown in Figure 2.
18
E
v
w
W
c
•H
0)
-p
m
o
e
(d
td
•H
o
•H
+>
<a
e
o
Xi
o
CO
CM
,Q
e .
H
w
o
o p
u
H
c
o
ffi
W
-p
E
c h
M U.
19
.
B. PRESSURE COEFFICIENT EXPERIMENT
In order to obtain static surface pressure measurements
on a cylinder placed in this particular flow situation, a
3 inch diameter, 20 inch long, smooth, right circular
cylinder constructed of aluminum was used. This cylinder
was fitted with four 0.02 5 inch surface pressure taps located
at the cylinder mid-point and spaced at 90-degree intervals
around its circumference . The circumference of the cylinder
was scribed in 5 degree increments. Reference lines were
scribed on the test section wall at 90 degree intervals so
that the matching of the scribe marks on the cylinder with
these lines provided a means of accurately establishing the
angular location of the pressure taps. The pressure cylinder
is shown in Figure 3. The surface pressure taps were
connected to a bank of 8 water filled manometer tubes by

means of plastic tubing. Two tubes of the manometer bank
were referenced against atmospheric pressure. The wind
tunnel stilling chamber static pressure and test section
(free stream) static pressure were measured on the same
manometer bank through pre-existing connections on the wind
tunnel. The experiments were performed to find the variation
of the surface pressure around the circumference of the
cylinder as a function of gap size, Reynolds number and plate
length
The plates used in these experiments were manufactured
from 1/2 inch plexiglass. These plates were 40 inches,
46 inches, and 58 inches long, respectively. To minimize
20
m
4J
ro
<L>
i-
ZJ
10
CO
<D
i-
Q.
+->
U
a;
o
u
o
+->
o
a>
co
u
s-
>>
o
a>
.c
+->
<+-
o
-C
«
S-
C7>
O
o
ro
i-
en
21
the influence of the downstream edge on the flow patterns
around the cylinder, the trailing edge in all studies was
maintained at a length of 34 inches. This provided leading
edge lengths of 6 inches, 12 inches, and 24 inches, respec-
tively. The width of the plates was 20 inches and they
were beveled on the under side of the leading edge to a

12 degree angle.
The plates and cylinder were placed in the tunnel
horizontally in such a way that the leading edge of the
plates extended 2, 4, and 8 cylinder diameters ahead of the
cylinder centerline. The plates were connected by means of
two iron angles to a threaded rod extending through the
tunnel floor which was capable of moving the plates vertically
and a rear brace which locked in position by means of a split
nut. A typical plate and cylinder arrangement is shown in

Figure 4. It was possible to locate the plate at any position
relative to the cylinder from attached to a maximum possible
distance of 8 inches away.
C. FLOW VISUALIZATION EXPERIMENT
The apparatus used for flow visualization consisted of a
converging wooden water channel 15 feet in length and
13 inches deep with a plexiglass test section 13 inches long,
12 inches wide, and 13 inches deep. A hollow, 1.5-inch
diameter plexiglass cylinder was attached between the test
section walls with its longitudinal axis located eight inches
from the test section floor. In order to preserve geometric
22
T3
C
-H
Q)
x:
+j
c
-H
4->
C!
Q)
£
0)
tn
C
5-1
ffl
5-1
O
T3
C
-H
H
>i
U
•O
(0
cu
+J
td
.h
Qj
CU
£
-P
m
o
xi
a,
id
n
tr>
O
-P
o
D
O
H
III
23
.
similitude, a 1/2 inch thick plexiglass flat plate was
constructed to the scale of the ratio of the water channel
cylinder diameter to the wind tunnel cylinder diameter
(0.5:1). Thus, this plate was 14 inches long and 12 inches
wide. The leading edge was beveled to a 12-degree angle as
were the larger flat plates used in the wind tunnel. This
flat plate was placed above the cylinder and was connected
to an apparatus which raised or lowered the plate and also
moved the plate to downstream and upstream directions
horizontally. By moving the plate up and down and also
horizontally, the desired gap sizes and L/D ratios were
obtained.
The cylinder was fitted with three dye ports which were
connected by polyethylene tubing to three separate dye
reservoirs with individual valving located approximately
three feet higher than the cylinder. This additional height
insured that a constant head was maintained on the injection
system, and the individual valving was necessary to allow
each port to be operated separately in order that the flow
field at various sections of the cylinder could be observed
independently
Water was pumped into the channel by means of an irriga-
tion pump directly coupled to a Westinghouse Type SK
dynamometer running as a motor. With the motor running at
1200 rpm, the pump output was approximately 1000 gallons per
minute. Because of the limitation in the pump capacity, the
highest Reynolds number which could be obtained in this
24
,
channel was approximately 10,000. The depth of water in
the channel was controlled by a tilting weir at the end of
the channel. For this particular experiment, the water
depth was maintained at approximately 12 inches. This
allowed the entire apparatus to be immersed to a depth
sufficient to minimize free surface effects.
The cylinder and plate arrangement in the test section
of the water channel is shown in Figure 5.
D. VORTEX FREQUENCY EXPERIMENT
The apparatus used to measure the vortex shedding
frequency behind the cylinder in the wind tunnel consisted
of the hotwire anemometer, TYPE 555 DUAL BEAM OSCILLOSCOPE
with amplifier TYPE 1A1, KROHON-HETE MODEL 3100 (R) Filter,
and FLUKE 19 50A Digital Counter.
The hotwire was attached to a tranversing mechanism by
means of a narrow depth cantilever. The transversing
mechanism allowed the hotwire to be moved to the desired

locations during the experiments. The output of the hotwire
anemometer was displayed on the oscilloscope. The Model
3100 (R) filter is a solid state variable electronic band-pass
filter which allows both high and low cut-off frequencies to
be independently and continuously adjusted over a frequency
range from 10 Hz to 1 MHz. The pass-band gain is unity (0 dB)
with an attenuation rate of 24 dB per octave outside the
pass-band, and a maximum attenuation of 80 dB . Using this
filter the turbulent high frequency noise was cut-off at an
25
C
C
id
X!
o
u
Q)
•P
(d
-P
C
•H
-P
C
I
<D
tn
C
<d
rd
<u
c
•H
.H
>i
U
T3
C
(d
-p
rd
Ch
Q)
X!
-P
m
O
Xi
&
<d
u
o>
o
-p
o
XI
CM
in
8
D
U
H
26
output of less than 100 microvolts. This allowed the low
shedding frequency to be obtained easily. The vortex
shedding frequency was read in Hertz from the FLUKE 19 50A
Digital Counter.
27
,
IV. EXPERIMENTAL PROCEDURE
In order to ensure that significant and accurate data
would be obtained, it was necessary to compare our initial
results with previous work. Since Gnerlich [2] performed
his experiment at a Reynolds number of 153,000 and L/D = 4.0,
it was decided to check the apparatus at that Reynolds number.
For this reason, an air stream speed of 98 ft/sec was used
for comparative purposes. Experiments were then carried out
using Reynolds numbers of 90,000 and 250,000 at L/D =4.0 .
To examine the influence of the location of the plate leading
edge, additional runs were also made at a Reynolds number of
153,000 and L/D =2.0 and L/D =8.0 .
In the present work the Reynolds number is defined in the
conventional fashion
U -D
Re = JSL-
where
Re = Reynolds number,
U = Free stream velocity at test section entrance, ft/sec

(uncorrected for blockage)
D = Cylinder diameter, ft,

2
v = Kinematic viscosity of air, ft /sec.
28
A. PRESSURE COEFFICIENT EXPERIMENT
Pressure data were obtained at Reynolds numbers of
90,000, 153,000, and 250,000. After considerable checks
with the hotwire anemometer, it was decied to use the wall
manometer to set the tunnel speeds of 98 ft/sec and 160 ft/sec
for Reynolds numbers of 153,000 and 250,000. To accurately
set a tunnel speed of 57.6 ft/sec, which yielded a Reynolds
number 90,000, the hotwire anemometer was used.
Initially, all leads were checked for tightness and the
manometer bank was leveled. The pressure taps were initially
located at , 90, 180, and 270 degrees. The wind tunnel was
started, and the velocity was increased to a predetermined
value. The manometer readings were recorded for each pressure
tap on the cylinder, the stilling chamber pressure, and the
test section pressure. The cylinder was then rotated in
five-degree increments, with data being recorded at each
increment, until the entire circumference of the cylinder
had been traversed. This procedure was repeated for all
desired Reynolds numbers, d/r , and L/D spacings. For
all experiments, data were then reduced to obtain pressure
coefficient around the circumference of the cylinder. Also
polar plots of the pressure coefficient versus angular
location were obtained. The polar plots are shown in Figures
12 through 19 in Section V and Figures 28 through 59 in
Appendix D. The pressure data and data reduction scheme used
to obtain the pressure coefficient can be found in Appendix A.
29
The scheme used to calculate the pressure drag coefficient
and lift coefficient from the pressure coefficient is

described in Appendix B.
B. FLOW VISUALIZATION EXPERIMENT
Before starting the motor, the sump was completely filled
with water. The motor was then started and brought up to a
speed of 1200 rpm. The tilting weir was adjusted such that
a height of water was obtained which completely submerged
the plate and cylinder. The flow was allowed to steady out
and the polyethylene tubes were checked for tightness. The
dye was then injected through the top and bottom ports, which
were 120-degrees apart from each other, in a trial and error
fashion until the flow could be visualized clearly. This
procedure was repeated for every desired gap spacing and
plate leading edge length. As previously described, the gap
spacing (d/r) and plate leading edge length (L/D) were 0.0,
0.25, 0.5, 1.0, 2.0, 3.0, 4.0, 5.33 and 2.0, 4.0 and 8.0,
respectively. These quantities were obtained according to a
scale which was attached to the water channel test section
wall. Because of the limitation in the pump capacity, only
one run was made with the highest Reynolds number of 10,000
which could be obtained in this experiment.
C. VORTEX FREQUENCY EXPERIMENT
In order to explain the trends which were observed in the
drag curve, a vortex shedding frequency experiment was
conducted in the wind tunnel at a Reynolds number of 153,000.
30
To start the experiment, the speed of the wind tunnel was
set to 98 ft/sec (2.2 in-H 2 0) . The frequency measurements
were made with a standard hotwire anemometer. Readings were
taken at the cylinder mid-span at various locations in the
wake behind the cylinder. Because of the limitation on the
hotwire anemometer's transversing mechanism, measurements
were restricted to a maximum downstream distance of
2.5 cylinder diameters.
The output of the hotwire anemometer was connected to a
TYPE 1A1 amplifier which removed the mean velocity (D.C.
level) and amplified the signal to a useful level. This
amplifier was connected internally to an oscilloscope display
which provided a visible signal for confirming the readings

from the counter. Also, this amplifier was connected
internally to a filter described in detail in Section III
which was used to remove the noise and turbulence. This
filter was connected to a TYPE 1A1 amplifier which amplified
the signal to a useful level and passed it to the display.
The second output of the last filter was connected to a
precision counter which yielded the vortex shedding frequency.
Using the above electronic circuit and hotwire anemometer,
the vortex shedding frequencies were obtained at d/r = 0.0,
0.25, 0.5, 1.0, 2.0, 3.0, 4.0, 5.33 and L/D = 4.0 . The
error of the frequency measurements was estimated to be
two percent.
31
In exploring the flow in the cylinder wake, two different
frequencies were observed for the same air speed when the gap
spacing was small. These frequencies were observed over the
top of the cylinder and between the cylinder and plate in the
flow field. Results of the vortex shedding frequency
measurements are shown in Section V.
32
.
V. RESULTS AND DISCUSSION
The results shown and discussed in this section are
the cylinder surface pressure distributions, pressure drag
and lift coefficients, qualitative observations from the
flow visualization experiment, and the measurements of the
vortex shedding frequency in the wake of the cylinder.
A. PRESSURE EXPERIMENT AND CYLINDER SURFACE

PRESSURE DISTRIBUTIONS
Raw data for three typical pressure runs and a sample
calculation of the pressure coefficient are contained in
Appendix A. This section presents a graphical presentation
of the pressure coefficient versus angular location on polar
coordinates for:
(a) Reynolds numbers of 90,000, 153,000, 250,000 with no

plate, Figures 9 through 11.
(b) Reynolds number of 153,000 with plate, L/D =4.0

and d/r ranging from 0.0 to 5.33, Figures 12
through 19
Polar plots for Reynolds numbers of 90,000 and 250,000
with plate L/D = 4.0 and d/r ranging from 0.0 to 5.33
are shown in Appendix D, Figures 28 through 43. Also
Appendix D includes the polar plots for Reynolds number
153,000 with plate L/D =2.0 and 8.0, and d/r ranging
from 0.0 to 5.33, Figures 44 through 59.
All polar plots include Nusselt number versus angular
location curves which were taken from Kosemen's [3] data.
Again this study was a parallel study with Kosemen who used
33
.
the same Reynolds number range (except 250,000) and the same
size cylinder to investigate the heat transfer characteristics
of a heated cylinder. Because Kosemen's data indicate
precise separation points and the flow characteristics on
the top and bottom half of the cylinder, his data were found
very informative and were plotted with the curves of pressure
coefficients versus angular location.
The trend in the plots and the magnitude of the pressure
coefficient for the no-plate case were found to agree well
with published values in the literature [6] (within 2%)
See Figures 9 through 11.
Before proceeding with an examination of the polar plots
which are shown in Figures 12 through 19, a brief description
of the classification of flow past a cylinder will be given.
This description will help to explain our observations about
the flow phenomena.
Flow is characterized by the nature of the separation of
the boundary layer from the cylinder surface. Using the
definitions employed by Achenbach [6]:
1. When the boundary layer separates laminarly, this
state of flow is called subcritical flow. In subcritical
flow, a laminar boundary layer develops at the forward
stagnation point of the cylinder and grows to a maximum
thickness at an angular location between 80 and 85 degrees.
At 80 to 85 degrees, the kinetic energy of the moving fluid

is no longer sufficient to overcome the adverse pressure
gradient present on the cylinder's surface, and the laminar
34
boundary layer separates. Upon separation, a wake forms in
the rear portion of the cylinder. A typical subcritical flow

pattern is shown in Figure 6.
2. When a separation bubble, followed by a turbulent

reattachment, occurs, this state of flow is called critical
flow. In critical flow, the laminar boundary layer again
grows from the forward stagnation point to the 80 to 85
degree point where laminar separation again occurs. But,
unlike the subcritical flow case, a transition of the
separated laminar layer to a turbulent state occurs which
results in the reattachment of a turbulent boundary layer.
This reattached turbulent boundary layer ultimately separates
at a position further downstream. The region between the
laminar separation point and the point of reattachment of
the turbulent boundary layer is called a "separation bubble."
A typical critical flow pattern is shown in Figure 7.
3. When an immediate transition from the laminar to the
turbulent boundary layer occurs without separation, this
state of flow is called supercritical flow. A typical
supercritical flow pattern is shown in Figure 8.
In examining the polar plots which are shown in Figures
12 through 19 for significant trends or unusual phenomena,
two items of interest were immediately apparent: one was
the change from critical to subcritical flow, the other was
a shift in the location of the stagnation point and the
laminar separation points as the plate was moved away from
the cylinder.
35
Laminar
Boundary
Layer
FIGURE 6. Sketch of a typical subcritical flow pattern

on the surface of a cylinder
Laminar _,
Reattachment
Separation Point
Turbulent
Boundary Layer
Separation Point
FIGURE 7. Sketch of a typical critical flow pattern

Turbulent
Boundary Layer
Transition
Region
JTurbulent
Separation Point
FIGURE 8. Sketch of a typical supercritical flow pattern

36
p
Nu
FIGURE 12. Pressure coefficient and local Nusselt number

plane surface for:
Re = 153,000, L/D = 4.0, d/r =0.0
.40
Nu

on the surface of a free cylinder for Re = 90,000
37
C *
P
Nu
LLIIIIIllJl-U

on the surface of a free cylinder for Re = 153,000
38
260
FIGURE 11. Pressure coefficient on the surface of a free

cylinder for Re = 250,000
39
Nu

plane surface for:
Re = 153,000 , L/D = 4.0 , d/r = 0.25
41
C x
P
Nu

plane surface for:
Re = 153,000 , L/D = 4.0 , d/r =0.5
42
p
Nu
260

plane surface for:
Re = 153,000 , L/D = 4.0 ,d/r = 1.0
43
p
Nu
FIGURE 16. Pressure coefficient and local Nusselt

number on the surface of a cylinder placed
near a plane surface for:
Re = 153,000 , L/D = 4.0 , d/r =2.0
44
C x
P
Nu

plane surface for:
Re = 153,000 , L/D = 4.0 , d/r =3.0
45
p
Nu

plane surface for:
Re = 153,000 , L/D =4.0 , d/r =4.0
46
Cp x
Nu

plane surface for:
Re = 153,000 , L/D = 4.0 , d/r =5.33
47
Kosemen' s [3] data were used to determine the d/r ratio
associated with a change from critical to subcritical flow.
The data obtained are shown in Table I. The results for a
Reynolds number of 16,000 which was used by Kosemen in his
investigation were included in this table for additional
information. This table shows the d/r ratios where the
flow undergoes a critical to subcritical transition for various
Reynolds numbers and L/D ratios. For the data collection,
liquid crystal thermography technique employed by Kosemen
were used. In short, this technique allows visual determina-
tion of both qualitative and quantitative heat transfer and
fluid flow information to be obtained on heated objects
placed in forced convection environments. This technique
employs cholesteric liquid crystals as the temperature-
sensing agent. The liquid crystals indicate temperature by
exhibiting brilliant changes in color over discrete,
reproducible temperature ranges. As a result, the liquid
crystal thermographic technique afforded the opportunity to
visually observe the effects of flow separation, the
separation bubble region, the turbulent boundary layer, and
the turbulent wake on the surface of a heated cylinder.
During the experiment, uniform heat flux was applied to
the cylinder surface by means of an electrical resistance
paper (Temsheet) . The paper was coated with S-4 3 liquid
crystal which indicates in three distinct color bands the
following event temperatures:
48
L/D = 2.0 L/D = 4.0 L/D = 8.0
Re
TOP BOTTOM TOP BOTTOM TOP BOTTOM
At "d/r" At "d/r" At "d/r" At "d/r" At "d/r" At "d/r"
16,000 0.20 1.00 0.20 0.90 0.20 0.80
90,000 0.15 0.90 0.15 0.90 0.15 0.80
153,000 0.15 1.20 0.10 1.20 0.10 1.20
TABLE I. Transition from critical to subcritical flow

as a function of Reynolds number, d/r and
L/D ratios
49
:
COLOR TEMPERATURE
Red 43.2° C
Green 43.9° C
Blue 44.6° C
In the wind tunnel, the electrodes on the inner surface
of the Temsheet cylinder were attached to a power supply.
The power supply was energized and voltage was adjusted to
bring the surface temperature of the cylinder to a temperature
of approximately 50° C. This caused the S-43 liquid crystal
on the surface of the cylinder to pass through its event
temperature ranges. Having pre-heated the cylinder, flow
was established in the wind tunnel. Immediately after flow
was initiated, the cylinder started to cool. To counter this,
the power supplied to the cylinder was increased to maintain
the hottest point on the cylinder at about 43.2° C. This
corresponded to the red point on the liquid crystal. The
plate was then attached to the cylinder (d/r = 0.0).
Adjusting the voltage to obtain the event temperature ranges
on the surface of the cylinder, the following observations
were made
Due to the fact that all of the air approaching the
cylinder is forced to go "up and over", the flow was found
critical (See Figure 7) over the top of the cylinder with
the plate attached. Also a separation bubble which manifested
itself as a local cool spot (liquid crystal does not display
color) bounded by two local hot spots indicated the presence

of critical flow.
50
As soon as a space existed between the cylinder and
plate (d/r = 0.25), the separation bubble suddenly was lost.
This was the change of flow from critical to subcritical
(See Figure 6) according to Reynolds numbers studied in this
investigation. By alternately moving the plate and adjusting
the voltage, the transition points over the top of the
cylinder for various Reynolds numbers and L/D ratios were
found to be within the range . 1 _< d/r <_ . 2 . See Table I.
Using the same procedure of carefully adjusting spacing
and voltage the transition from critical to subcritical flow
on the bottom of the cylinder was found in the interval
0.8 < d/r < 1.2 for various Reynolds numbers and L/D ratios
See Table I. One can easily see from Table I that the transi-
tion points which occurred at various d/r ratios are
approximately independent of L/D ratios but dependent on
Reynolds numbers. When the Reynolds number was increased,
the d/r ratios for transition on the top of the cylinder
decreased slightly. On the bottom of the cylinder the trend
was reversed.
B. PRESSURE DRAG AND LIFT COEFFICIENTS
In order to gain further information from the pressure
distribution around the cylinder, the pressure drag (C, ) and
the lift coefficients (C T ) were computed and are shown in
Table II. Appendix B illustrates the method which was used
to calculate the pressure drag and lift coefficients.
For the Reynolds numbers range used in this investigation
Achenbach [6] states that the friction forces are nearly
51
Re L/t> d/r Re L/D d/r
'dp °L 'dp °L
90,000 NO PLATE 1.17 0.00 153,000 4.0 2.0 1.43 0.05

90,000 4.0 0.0 0.85 1.30 153,000 4.0 3.0 1.35 0.03
90,000 4.0 0.25 1.17 0.40 153,000 4.0 4.0 1.34 0.01
90,000 4.0 0.50 1.12 0.26 153,000 4.0 5.33 1.29 0.01
90,000 4.0 1.0 1.41 0.12 153,000 8.0 0.0 0.60 1.20
90,000 4.0 2.0 1.39 0.05 153,000 8.0 0.25 1.12 0.30
90,000 4.0 3.0 1.35 0.02 153,000 8.0 0.50 1.08 0.22
90,000 4.0 4.0 1.32 0.02 153,000 8.0 1.0 1.44 0.11
90,000 4.0 5.33 1.32 ,
0.01 153,000 8.0 2.0 1.41 0.05
153,000 NO PLATE 1.23 0.00 153,000 8.0 3.0 1.38 0.03
153,000 2.0 o.o 0.75 1.58 153,000 8.0 4.0 1.32 0.01
153,000 2.0 0.25 1.21 0.42 153,000 8.0 5.33 1.32 0.00
153,000 2.0 0.50 1.11 0.28 250,000 NO PIATE 1.10 0.00
153,000 2.0 1.0 1.46 0.11 250,000 4.0 0.0 0.61 1.39
153,000 2.0 2.0 1.43 0.04 250,000 4.0 0.25 1.15 0.42
153,000 2.0 3.0 1.36 0.04 250,000 4.0 0.50 1.08 0.27
153,000 2.0 4.0 1.33 0.02 250,000 4.0 1.0 1.38 0.13
153,000 2.0 5.33 1.33 0.02 250,000 4.0 2.0 1.32 0.06
153,000 4.0 0.0 0.64 1.44 250,000 4.0 3.0 1.23 0.02
153,000 4.0 0.25 1.16 0.41 250,000 4.0 4.0 1.20 0.01
153,000 4.0 0.50 1.09 0.27 250,000 4.0 5.33 1.16 0.00
153,000 4.0 1.0 1.46 0.12
TABLE II. Pressure drag and lift coefficients as a function

of gap spaces for various Reynolds numbers
52
negligible in calculating the drag coefficient. Therefore,
the skin friction drag was neglected compared to the pressure
(form) drag which constitutes 98 percent or more of the total

drag according to Achenbach.
The plots of the pressure drag and lift coefficients as
a function of gap spacing for various Reynolds numbers and
L/D ranges are shown in Figures 20 and 21. Figure 20 shows
the effect of L/D ratio on C, and CL and Figure 21
shows the effect of Reynolds numbers on C. and C, . It
is interesting to note the unexpected shapes and trends of
the pressure drag versus d/r ratio curves. For all
Reynolds numbers and L/D ratios, the pressure drag coeffi-
cient increases to a local maximum value, then decreases to
a local minimum then increases to a maximum at d/r =1.0 ,
and finally gradually decreases to a nearly constant value
as d/r approaches relative infinity. It is important to
emphasize that the trends of the pressure drag coefficient
versus d/r ratio curves are the same for all Reynolds
numbers and L/D ratios investigated. This implies there
is no effect on the trends of the curves due to changing
the L/D ratio or Reynolds numbers.
A question that immediately comes to mind when examining

the drag curves, Figure 20 and Figure 21, concerns the
maximum value of the drag coefficient that occurs at

d/r =1.0 . The flow visualization technique described in
Section IV and hotwire anemometer measurements of the vortex
shedding frequency (Section IV) were used to provide both a
53
o
"o
>i
a • o -P
C
•H
o O o 4-> L°.
• • • M
(N «<3< 00
o
II II II
D
P a Q
\ \ \
^ a J H
00
d S
54
>i
• O o
•H
o
o o
o o
o
<tf
-M
oMb
o fe
o •*
M
o n
in
o
LD
d)
G\
i-H <N
L>
II II II
& s a H
55
qualitative and quantitative explanation of this phenomena.
The results of this aspect of the study are shown in Part C
and Part D of this section. In short, it was found that at
d/r = 1.0 the vortex shedding frequency of the flow field
reached a maximum value compared to its values at d/r = 0.0,
0.25, 0.5, 2.0, 3.0, 4,t)and 5.33. This caused the base pressure
to exhibit a minimum value and as a result of this, the
pressure drag coefficient reached its maximum value at
d/r =1.0.
In examining the lift coefficient as a function of gap
spacing the following observations can be made:
a) The maximum lift coefficient occurs when the plate
is attached to the cylinder.
b) The lift coefficient suddenly drops to a small value
at d/r = 0.2 5 and then gradually decreases to an ultimate
value near zero at d/r = 5.33. See Figures 20 and 21.
With no plate in the tunnel the lift coefficient was
found to be zero as expected and the pressure drag coeffi-
cients were found to agree with literature values [6]. See
Table II.
C. FLOW VISUALIZATION EXPERIMENT
In order to obtain qualitative answers to the questions
which arised from the pressure (form) drag characteristics

of the flow field, a flow visualization experiment was
performed in the water channel (described in detail in
Section III) . Because of the limitation in the pump capacity,
56
:
the highest Reynolds number which could be obtained in this
experiment was approximately 10,000. This was well below
the Reynolds numbers range examined in the pressure experi-
ment. However, the behavior observed agreed well with the
proposed theory of vortex formation.
During the experiment the following observations were
made
1. At d/r =0.0
Referring to Figure 22, when the gap spacing is zero
there is no flow between the cylinder and plate. The flow
appeared to be relatively stagnant in the downstream cavity
formed by the cylinder and plate. No "trapped" vortices
were observed. Indistinct vortices were observed at approxi-
mately four diameters downstream. These vortices most
probably detached from the outer side of the cylinder. At
the upstream cavity a series of small, circular, three-
dimensional vortices were clearly evident. Also a large
turbulent vortex was observed (See Figure 23) . According to
McComas [1] , this large turbulent vortex occurs within the
separated region where the flow separates from the flat
plate and reattaches to the cylinder at a point just aft of
the forward stagnation point. Due to the stagnant region
behind the cylinder, the base pressure is found to exhibit
a maximum value and thus the drag on the cylinder is
decreased considerably from the free cylinder value. On the
other hand, the lift force reaches its maximum value due to
the plate effect. See Table II for the minimum value of the
57
FIGURE 22 . Photograph and sketch of the flow pattern
at d/r =0.0
58
mm?*
FIGURE 23. Photograph and sketch of the trapped vortices

at d/r =0.0
59
pressure drag coefficient and maximum value of the lift
coefficient at all Reynolds numbers and L/D ranges for
d/r =0.0.
2. At d/r = 0.2 5
plate a stable jet flow was observed between the plate and
cylinder in the immediate vicinity of the gap. See Figure 24.
This jet widened as it moved downstream and mixed with the
flow in the wake. Vortices were not detached immediately
from the top and bottom of the cylinder probably due to the
effect of the jet flow. A qualitative comparison with the
d/r = 0.0 case from astrength viewpoint, indicated that
distinguishable vortices were observed at approximately four
diameters downstream for d/r = 0.25 while only very
indistinct vortices were observed for d/r = 0.0. Probably
due to these vortices, the base pressure exhibited a more
negative value compared to the d/r = 0.0 value. As a
result of this, the drag increased to a local maximum value
as indicated in Figure 20 and Figure 21 for all Reynolds
numbers and L/D ranges.
3. At d/r =0.5
As the gap further increased in size, a jet flow was
still observed between the plate and cylinder. See Figure 25.
In this state vortices again were observed at four diameters
downstream. Compared with the d/r = 0.2 5 case, the vortices
did not appear as strong. Due to this decrease in vortex
60
' *.. A**5-'
FIGURE 24. Photograph and sketch of the flow pattern

at d/r =0.25
61
mi
//*

at d/r =0.5
62
strength, the drag decreased to a minimum value which one
can see from Figures 20 and 21 for all Reynolds numbers and
L/D ratio ranges which were used in this investigation.
4. At d/r =1.0
This was the location where more distinct three-
dimensional vortices were observed during the flow visualiza-
tion experiment. The boundary layer appeared to separate
from both the top and bottom surface of the cylinder in the
form of alternating three-dimensional vortices and a very
strong vortex shedding frequency was observed. Also, a
rather regular, violent vortex street formed at about
one-half diameter downstream. See Figure 26. Due to the
formation of this strong vortex street very close to the
cylinder, the base pressure decreased to its minimum value
and the wake width became much greater than the cylinder
diameter. As a result of this, the pressure (form) drag
reached its maximum value as illustrated in Figures 20
and 21.
As the gap space was further increased to d/r = 2.0,
3.0, 4.0, 5.33 it was observed that the strength of the vortex
street and shedding frequency also appeared to decrease.
Therefore, the drag on the cylinder gradually decreased
to its "free" cylinder value. See Figure 27.
63
N.,
// •A
\\
i \
<F=

at d/r =1.0
64
f-&>*• >
<3=

at d/r =5.33
65
D. VORTEX FREQUENCY EXPERIMENT
In order to obtain quantitative answers to the questions
arising from the pressure drag curve characteristics, an

experiment was performed to obtain the vortex shedding
frequency behind the cylinder (described in detail in
Section III) . Because of limitations in the hotwire
anemometer's transversing mechanism, the vortex shedding
frequency was measured from 1 to 2.5 cylinder diameters
behind the cylinder. In spite of no limitations in the
vertical direction, the best results were obtained at
1.5 cylinder diameters from the center of the cylinder.
During the experiment the following results were obtained,
1. At d/r =0.0
No frequency signal was observed on the display of
the oscilloscope and the shedding frequency read from the
counter was approximately zero. This was found to agree
with the flow visualization results at d/r = 0.0.
2. At d/r = 0.2 5
plate, two different frequencies were obtained at the top and
bottom of the cylinder. The top frequency was approximately
three times greater than the bottom frequency. This implied
that there were no strong vortices in the gap between the
cylinder and plate. This result also agrees with the flow
visualization observation at d/r = 0.25.
66
.
3. At d/r =0.5
At this gap size, two frequencies again appeared.
The top frequency was about the same as the frequency which
was found at d/r = 0.2 5 but the bottom frequency was less
than the d/r = 0.25 case.
4. At d/r =1.0
The maximum frequencies observed during the experi-
ment were at this gap size. This was similar to the
qualitative results observed in the flow visualization
experiment and also coincides with the location at which
the maximum drag coefficient was observed.
As the gap space was further increased to d/r = 2.0,
3.0, 4.0, and 5.33 it was found that the shedding frequency
gradually decreased to its "free" cylinder value. The
results, including the Strouhal number, are shown in
Table III. Also, the results were found to agree qualita-
tively with the results which were obtained by Roshko [4]
and Spivak [5]
67
TOP
LOCATION
d/r
n (Hz) S
X Y Z
0.0 7.5 4.5 0.0 0.0 0.0
0.25 7.5 4.5 0.0 73 0.186
0.50 7.5 4.5 0.0 73 0.186
1.0 3.0 4.5 0.0 88 0.224
2.0 3.0 4.5 0.0 79 0.201
3.0 3.0 4.5 0.0 79 0.201
4.0 3.0 4.5 0.0 78 0.198
5.33 3.0 4.5 0.0 78 0.198
BOTTOM
0.0 7.5 4.5 0.0 0.0 0.00
0.25 7.5 -0.88 0.0 23 0.058
0.50 7.5 -1.24 0.0 20 0.050
1.0 3.0 -2.0 0.0 81 0.206
2.0 3.0 -3.5 0.0 79 0.201
3.0 3.0 -3.9 0.0 79 0.201
4.0 3.0 -4.5 0.0 78 0.198
5.33 3.0 -4.5 0.0 78 0.198
TABLE III. The vortex shedding frequency, n, and Strouhal

number, S = nD/U s as a function of gap size at
,
a Reynolds number of 153,000 with L/D =4.0
X = Distance measured downstream from the center of the

cylinder (inch)
Y = Vertical distance from the center of the cylinder (inch)
68
VI. SUMMARY
The experimental investigation described in this thesis
was conducted to gain an understanding of the fluid flow
surrounding a right circular cylinder placed near a plane
surface. Surface pressure coefficients were measured
around the cylinder circumference for various plate to
cylinder spacings, plate lengths and Reynolds numbers.
Drag and lift coefficients were obtained from these data
by numerical integration.
An unexpected trend which was observed in the plots of
drag versus cylinder to plate separation prompted a flow
visualization experiment and hotwire anemometer measure-
ments of the vortex shedding frequency in the cylinder wake.
Based on these observations and measurements, it was
concluded that the structure of the vortex pattern that
develops in the wake of the cylinder exerts a controlling
influence on the shape of the drag curve. For a range of
plate to cylinder spacings on the order of one cylinder
radius, it was found that the plate interferes with the
formation of the vortex street in the wake of the cylinder.
This interference significantly affected the cylinder base
pressure which, in turn, influenced the drag on the cylinder
69
APPENDIX A
PRESSURE DATA AND DATA REDUCTION
The values of the pressure coefficient were determined
from the formula
P - P P - P
= e s _ e s
c
'
2
P ipU
2 K s
P
ps
-P
s
where
C = Pressure coefficient
P
p = Density of air, lb/ft 3
U = Air stream velocity, ft/sec

s
P = Pressure on the cylinder surface at angular
location 6, in-H 2
P = Air stream pressure in the wind tunnel test section,

S
in-H Q
2
P = Pressure in the wind tunnel stilling

3 chamber, ' in-HÔ
ps 2
A sample calculation for data obtained at Re = 250,000,
d/r = 0.2 5 and L/D =4.0 is shown below
9 = 235° P n = 14.10 P = 8.00 P = 1.90

6 s ps
14.10 - 8.00
p 1.90 - 8.00
C = -1.00
70
3
Air Speed- 0.
J£ fa . /£ Q Date i /ofujr-^3- iBlS
Time i Z2-40 ?M
Air Temp - £ / 'f
Re tf0j 00
- 0-25" L/D = 4-0
e Po Pp3 Pe Cp e Ps
- ...
Pps Pe Cp
o 3- 6fi ^-"75- 0, . Q 0. # = 8 1 9,0 3.6 8 2.15 4. oO -o.3%9
5 j
5- OJ? c 70$ ! 8 5 v- ^0
I I
3- /^ 0. S2 6 13 /- £tf
1 5 .
2 a _£ 0. 3 07 13 5 4 £ O \
.20 a/35 200 7. 6

2 5" 3-7£ -0.10? 2 05 4- 60
30 3-^5 -0-29O 2 1
lu &o
35" ^2 - a ^53 J? f
5" 4- Go
4 (j.39 -0.14,3 i2 4. ^^
45 H.5? -0.361 £25 ^. £
-/. 01 s
1
50 k-68 2.3 y. ^^ x]/
d
1
55 ! tf.1-% -/./ %z ? <r 4- -1.0 1

£0 I
r 1 f7£ -l.)SZ 2 40 1 4- Co 2 -1.0 1
6S |
i
8
li.6 -/.015 ,2 *5 4. (,2. -\.o\
10 4.60 -0.389 2 CO 6. & 1 -1.0 1
IS 1
\.se -0.918 cr t=~
4'. £ 1 -1. ceo
RO 1
!
\4-S9 -G.318 -2 6 4-^/ -/. 00 c
P5" I •
\/ f 5S -o.3?8 ^ 65 \ / \/ r-^ 9 -{.OIS
30 3.*ff £.15 ^.60 -as S3 37 3.£g 2.1S 4-X2 -1-230
4.6o -0.383 ^7 5 ^75 -I. ISO
100 4-60 210 I
^.55 -O.S35\
1
05" 4.6o 29 5 4.3 6? L# ^^^j
I I 4. 60 23 ^. t7^ -0.3441
115 4. 6C 23 2-7^ -0. 036\
\20 4. 6 300 2-eo 0. C? (-
\ZS 4. 6 30S 2. jo O.301

130 ^60 31 3-2 7 c.^40
1 ?>S 4- 60 3i 5 3./' a ss 1
1 40 l{. fo 32 3- £5 C C 1 T .
r~
1 hs 4. (0 <
_\2 0. %o6
1 50 4. to 33 2- #3 c. S\t{
55 4. h £. 16 0.961
^
I
1 Af) 4- 3 4 z-75 1.000

1 ^5
1
j
£ 62 34 5"
. 2riS i. COO
1 IO j
l/. fl
•-* ^-
S.?7 C-31?
1 7 5 V »
1 4» £ W 35 5 V y . z.ez 0.32 4'
71
rivivoounij 3
VAiA
Air Speed- 2.2 In- M2 Date I J^y-^9- /5 7 5"

- 38 ft /sec Time t fl -2<? ?/M
Air Temp . 7#°F Re \SZ;OoO
0.25 L/D = 4.0
e Pa Pps Pe Cp e Ps Pps Pe Cp
o 4-40 2-00 |^.5"0 0. 1S\ 1 SO 4-4o 2- CO £-7 5 -c.-j~3
5 ^.75 0- £07 ! S5" £.JO -/• ceo
1 1 3 • -2 5" 0-47,9 13 £> Po
1 5 3-62- )35 ._
6. go
20 ^.20 a o £3 200 £. £o i
2 £T 4- *o -a /^ 205 O' n O
50 |
5"-^0 — - 33 3 2 1 s .?n
35 1
5-'7P -0.575" 2. ! S 6* kO I
4 £. 25 -0. ?7C .22 6- So i
4 5" £. ^6? -0.3 s8 225 l£-8o v

50 7-05" -1.10*1 2 30 U. g.5 _/. C2C
55 |
'
7-25" -/./«? 2 0,5- 1

£. b5|-/.(9i o
60 i 7.25 -».i*7- 2 40 6>. 25 -/. ozo
65" 1
£-00 -/. 04/ 2 ^5 <?,S2 -A 0/0
N 5"
70 •'-
-1.020 Z C £. 3£ -/. 0/ O
15 i
1 | \6. *0 -/. ccc 1 k. g2 1-/. o/O

90 i
£. g -/- OQo 2 60 r i
|^. go I-/- CC70
f?5" V/ V./ 6- 8 o -/- (90 Z 65 V \/ ^. g5 I-/-020
30 4-4o .2-00 £ Oo -/. GOO 210 iô ^-OO 7.55" \-l.2Z3
3 5" 6. 84 -/- 0/^ 215 1
7.£o !_/. /^ j
100 £.35- -/.02O 2%0 £. 62 \-C.3Zi\

1 05 £.#0 -( .COO 2. £5 |
£. CC \-C63l j
1 1 £ 77 ~o.9e3 23 J s*.ô i-o-î

1 10 6.1-9- -a $93 22 S U.rfO -C 0$2\
)20 & £o -/. 5£0 300 J u. 20 C OS
125 6- Bo 3 OS 2. + C 0.231
!30 6, Qo 3 1 3-3 C. 4 5&
5"
135 6.eo 3i Z.SS' O.6o4
AD 6,po 52 2.6^ \C.1Z3
T
}
1 MS 6. $0 Z25~ 1
2.25 o. ?s-f
! 50 6- r O \/ 33 2.20 0.316
I 55 6. 7^ • - 0.37 3 i
U-/0 C.J5f
1 60 6. is 3 4 2. OS 0.373
1
-5" 6.^5 34 5 2. OS C.3?3
1 1O 6. 75" 3 5 2. 10 0.35*6
! 175 V
f
J îT , >
' 35 5"
\,
>
_V 2.2 0.31 6
72
rnr,oautt£ DATA
Air Speed- 5.85 ,n../-/z O

Date I August- 31- 11^5
m f 60 -ft- /sec
Time t 9.3c An
Air Temp - 66 °F
Re 250,000
*h o.zs L/D -4.0
e Pa Pp3 Pe Cp
- ,
Ps Ppa Pe Cp
. .. .
o 2-0 \.3o o.OO 0. 2)9 1 50 g. co / 5o CO

/y . -
0.983
5 3.1<o 5.63 I
! %S 13-So -C.J6-
f 4-^7 5.5/3 13 /5 oo -Cj'rZ
I 5 c.£P (7-34? \3 5 13.33 -CSS 1
20 7-^5" 0. 121 5.0 13-33 -0.3 9}

2 5 g-70 -aiiM 205 13.23 -0.391
3 3. So -0.3/1 2 ! \k-oo - C.J S3
35 ll. 3$ -0.5-54 J2 f
5" oo -C.J 83
14-
4 J2.6C -0.1 u 22 14-05 -0.331

4 5" 14- CO -0.3 €3 £2 5" I -CO -C-3P3
1
!
50 14*90 -1.131 J? 3 14-05 -c.es 1
55 /5. 7 O -I.JL6Z I 35 \4- 10 -/. Coo

6 /5-.7S- -I.Z 7 2 40 \\4.l? -I.00B
6S i/^-^5 -/. 03 o 2 ^5 j
/yÔ -i.eit
10 \H.30 -I.C32 5. 5 1
l^ 20 -I.CU
75 1 \M-30 I-/. 03.2 -8
^3" /5/5 -1-ooS
RO 1/4. 30 -I.CZ1 2 6 Ik -05 - 5. £5- \
05" -Low }u. 3 -! 032

i i
!
l4-4o .2.65 1 V/
30 9.00 I.SO /4.<3<? -1.032 -8 7 £.00 1.90 15- tO -1-262
\
3 5"
/f lo -1. ooo £75 j
\5> 1 O \-f.U3
100 It 10 _/. CCO Z9>0 )3- 60 \-0.3l 9,
1 05 /*. CO -0.313 2. ?5 II.35 \-0-6 5? .
! 1 j/4. 55 -0.9% 3 J2 3 1 CIO -0.344

I
15" \l4-oo -0.993 iô i
g.so \-o.ori \
\20 \ik-oo -0.393 300 t.3$ O.ICI

\%5 14.00 -0.3C3 3 05 6.15 0.303
r
!30 In- 05 -0.33 1
3 1
5. ;0 5.-55
1 ?>5 /f 0<7 -0.383 5i 5"
4- 3o 5.606
1 40 '5. 55 -5.55/ 52 0.704
1 4.5 /4- 05" -CJi/ .
325" 5-. co C. $ ij
1 50 /*. CO -0-3 9>3 33 : r 0. C3li
1 55 /y. C C -0.3 93 5- CO C.3S3

1 A0 /v. c -0.393 3 4 1. %9 I. c c l
! r5" 14- co -0.0 S3 34 5 1.3 1.ceo

! 1O iz.3 o - C- 9 6 7 3 5 2,.Q8 0.970
75 V 13.95 -5.57 5 35 5"
03
1
.., \/ -
1 V 2. 7 3 C 3 .
73
APPENDIX B
CALCULATION OF PRESSURE DRAG AND LIFT COEFFICIENTS
The pressure drag coefficient and lift coefficient were
calculated from the following formulas.
2tt
c - c (e) cos e dl
dp 7 ! p
o
2tt
c '
ij Z. p
where
C-, = Pressure drag coefficient
C XJ
T
= Lift coefficient
C = Pressure coefficient
P
d0 = Angular increment (5-degree)
Pressure drag and lift coefficients were obtained by
numerically integrating the above formulas using the
computer (IBM 360) and Simpson's Rule of Quadrature with
a step size of five degrees. The results are shown in
Table II in Section V.
74
APPENDIX C
UNCERTAINTY ANALYSIS
Since uncertainties existed in our measured variables,
an uncertainty analysis was performed in order to ascertain
the degree of uncertainty in the experimental results. The
method of Kline and McClintock [7] was used.
Uncertainties existed in the following variables:
Wall manometer Z ±0.05 inch (20:1)
Manometer bank H ±0.05 inch (20:1)
The uncertainties associated with the properties of air were
considered negligible.
REYNOLDS NUMBER, Re :
The Reynolds number was calculated from the equation
U -D
Re = -1—
Therefore,
Re
Re
/W 1
(*)"*(*)
The value of U is directly proportional to the square
root of the wall manometer reading Z; hence
75
—U_U
§. 05
= U,D —
o)
Z
7
Again, since the uncertainty in the kinematic viscosity of
air and in the cylinder diameter are negligible.
^=0 5^
Re
u *
D
Z
Therefore at various Reynolds number ranges the uncertainties
were calculated as follows
a) At a Reynolds number of 90,000
Z = 0.76 in-H u> = ± 0.05 in-H 2

2
-=£= = 0.5(0.05/0.76)
Re
-#= = 0.0328 or w_ = 2968 ~ 3000

Re Re
Re = 90,000 ± 3000 (20:1)
b) At a Reynolds number of 153,000
Z = 2.20 u> = ± 0.05 in-H

2
w Re
-—
Re
= 0.5(0.05/2.20)
76
^Re
-^- = 0.011 or o) = 1738 z 2000
Re
Re = 153,000 ± 2000 (20:1)
c) At a Reynolds number of 250,000
Z = 5.85 in-H a> = 0.05 in-H

2 2
^Re
—^ = 0.5(0.05/5.85)
0)
Re =
Re
0.004 or ^Re
a)
= 1000
Re = 250,000 ± 1000 (20:1)
PRESSURE COEFFICIENT, C :
The pressure coefficient was calculated from the equation
P -P AP n
= _6 s 1
r ..
p
r P -P AP n
ps s 2
Therefore,
2 u ^
"C_ //"ap^ / ap 2
C
£ _
AP AP
p V \ 1 / V 2
77
Since both pressure differences were measured on the same
manometer bank they shared the same uncertainty which is

shown on page 75. Therefore, for the various Reynolds
numbers the uncertainties were calculated for the pressure
coefficient as follows.
a) At a Reynolds number of 90,000 and an angular

location of 270 degrees,
d/r = 0.25 L/D = 4.0 P. = 4.82 P„ = 3.68 P = 2.75

s ps
W
C
-£ = 0.069 or io„ = 0.08
P P
C = -1.23 ± 0.08 (20:1)
b) At a Reynolds number of 153,000 and an angular

location of 270 degrees,
d/r = 0.25 L/D = 4.0 P Q = 7.35

o
P^ = 4.40
s
P_
ps
= 2.00
c
'2- = 0.026 or oo^ = 0.03
P P
C = -1.23 ± 0.03 (20:1)

P
c) At a Reynolds number of 250,000 and an angular

location of 270 degrees
d/r = 0.25 L/D = 4.0 P Q = 15.7 P^ = 8.00 P^ o = 1.90

u s ps
78
C
-^ =0.01 or a) = 0.01
P P
C = -1.26 ± 0.01 (20:1)
PRESSURE
- DRAG AND LIFT COEFFICIENTS, C-. , CT
'
dp '
L
Since the pressure drag and lift coefficients were obtained
integrating the pressure coefficient/ it seems reasonable
to assume that the drag and lift coefficients exhibit the
same uncertainty as the pressure coefficient. Therefore,
a) At a Reynolds number of 90,000, d/r = 0.25, and

L/D =4.0
C, = 1.17 ± 0.08
dp
(20:1)
CT = 0.40 ± 0.08
Li
b) At a Reynolds number of 153,000, d/r = 0.25, and

L/D =4.0
C,, = 1.16 ± 0.03

dp
(20:1)
C
L
=0.41 ±0.03
c) At a Reynolds number of 250,000, d/r = 0.25, and

L/D =4.0
C
dp = 1.15 ± 0.01
(20:1)
CT = 0.42 ± 0.01
Li
79
APPENDIX D
POLAR PLOTS OF PRESSURE COEFFICIENT VERSUS
ANGULAR LOCATION
The polar plots for Reynolds numbers of 90,000 and 250,000
with plate L/D = 4.0 and d/r ranging from 0.0 to 5.33
are shown in Figures 28 through 43. Also this Appendix
includes the polar plots for Reynolds number of 153,000
with plate L/D = 2.0, 8.0 and d/r ranging from 0.0 to
5.33. See Figures 44 through 59.
80
p
Nu
FIGURE 28. Pressure coefficient and local Nusselt number on

the surface of a cylinder placed near a plane
surface for:
Re = 90,000, L/D =4.0, d/r =0.0
81
p
Nu
"260
ou

plane surface for:
Re = 90,000 , L/D = 4.0 d/r =0.25
,
82
C x
P
Nu

plane surface for:
Re = 90 000
, , L/D = 4.0 , d/r =0.5
83
p
Nu
ou 260

plane surface for:
Re = 90,000 , L/D = 4.0 ,d/r =1.0
84
C x
P
Nu
0*2 260

plane surface for:
Re = 90,000 , L/D = 4.0 d/r =2.0
,
85
Nu
0^? *6
°

on the surface of a cylinder near a plane
surface for:
Re = 90,000 , L/D =4.0 d/r =3.0
,
86
C x
P
Nu

plane surface for:
Re = 9 0,000, L/D = 4.0 , d/r =4.0
87
p
Nu
760

plane surface for:
Re = 90,000 , L/D = 4.0 ,d/r =5.33
88
ou
FIGURE 36. Pressure coefficient on the surface of a

cylinder placed near a plane surface for
Re = 250,000, L/D =4.0, d/r =0.0
89
Re = 250,000 , L/D = 4.0 , d/r = 0.25
90
~"J^ ' '
UJ I Ml I I
ill
/60

Re = 250,000 L/D = 4.0
, d/r =0.5 ,
91
^T"-^' ' 1 1 1 1 1 1 1
1 1 1

Re = 2 50,000, L/D = 4.0 d/r =1.0 ,
92
Re = 2 50,000, L/D = 4.0 , d/r =2.0
93
LU I I I I I I UXU 60
Oil *

Re = 250,000 , L/D = 4.0 d/r =3.0 ,
94
OU 260

Re = 250,000 , L/D = 4.0 ,d/r =4.0
95
Re = 250,000 , L/D = 4.0 , d/r = 5.33
96
p
Nu

plane surface for:
Re = 153,000 , L/D = 2.0 , d/r =0.0
97
C x
P
Nu
ocz 260

plane surface for:
Re = 153,000 , L/D = 2.0 d/r = 0.25
,
98
p
Nu

plane surface for:
Re = 153,000 , L/D = 2.0 , d/r =0.5
99
C x
P
Nu
ou W
FIGURE 47. Pressure coefficient and local Nusselt nuinber

plane surface for:
Re = 153,000 , L/D = 2.0 , d/r =1.0
100
p
Nu
260

plane surface for:
Re = 153,000 , L/D =2.0 d/r =2.0
,
101
p
Nu
OU 1*°

plane surface for:
Re = 153,00 , L/D = 2.0 d/r =3.0
,
102
Nu
260-
ou

plane surface for:
Re = 153,000 , L/D = 2.0 d/r =4.0
,
103
p
Nu
•oae 760-
ou

plane surface for:
Re = 153,000 , L/D = 2.0 d/r =5.33
,
104
p
Nu

plane surface for:
Re = 153,0 00, L/D = 8.0 , d/r =0.0
105
p
Nu

plane surface for:
Re = 153,000 , L/D =8.0 , d/r =0.25
106
x
s
Nu

plane surface for:
Re = 153,0-0 , L/D = 8.0 , d/r =0.5
107
p
Nu
OCZ 260

plane surface for:
Re = 153,000 , L/D = 8.0 , d/r = 1.0
108
p
Nu
OU 26<>

plane surface for:
Re = 153,000 , L/D = 8.0 d/r =2.0
,
109
p
Nu
OLZ

plane surface for:
Re = 15 3,000, L/D = 8.0 , d/r =3.0
110
C x
P
Nu
OU 26°

plane surface for:
Re = 15 3
, 00 , L/D = 8.0 d/r =4.0
,
111
p
Nu

plane surface for:
Re = 153,000 , L/D = 8.0 , d/r =5.33
112
LIST OF REFERENCES
1. McComas, J. P., Experimental Investigation of Ground

Effects on a Heated Cylinder in Crossflow M.S.M.E.
,
Thesis, Naval Postgraduate School, 1974.
2. Gnerlich, C. H., The Convective Heat Transfer Behavior

of a Heated Cylinder Located Near a Plane Surface ,
M.S.M.E. Thesis, Naval Postgraduate School, 1975.
3. Kosemen, Ender, Thermal Study of a Heated Cylinder Near

a Plane Surface, M.S.M.E. Thesis, Naval Postgraduate
School, 1975.
4. Roshko, Anatol, On the Drag and Shedding Frequency of

Two Dimensional Bluff Bodies Nat. Adv. Comm. Aero.,
,
Washington, Tech. Note 3169, 1954.
5. Spivack, Hermann N., "Vortex Frequency and Flow Pattern

in the Wake of Two Parallel Cylinders at Varied Spacing
Normal to an Air Stream," Jour Aero Sci
. , vol. 13, no. 6,
June 1946, pp. 289-301.
6. Achenbach, E., "Distribution of Local Pressure and Skin

Friction Around a Circular Cylinder in Cross-Flow up to
Re = 5xl0 6 ," J. Fluid Mech., (1968), vol. 34, part 4,
pp. 625-639.
7. Kline, S. J. and McClintock, F. A., "Describing Uncertain-

ties in Single-Sample Experiments," Mechanical Engineering ,
vol. 75, pp. 3-8, January 1953.
113
INITIAL DISTRIBUTION LIST
No. Copies
1. Defense Documentation Center 2

Cameron Station
Alexandria, VA 22 314
2. Library, Code 0212 2

3. Department Chairman, Code 59 1

Department of Mechanical Engineering
4. T. E. Cooper, Code 59Cg 5

Department of Mechanical Engineering
5. C. F. Markarian 1
Code 4061
Naval Weapons Center
China Lake, California 93555
6. LT Selahattin Goktun, Turkish Navy 1

Harmanlik Sokak No:
Kartal-Istanbul
Turkey
7. Istanbul Teknik Universitesi 1

Makina Fakultesi
Taskisla, Istanbul
Turkey
114
*»«0
Thesis
G3155
053^+
c.l Goktun
The drag and lift
characteristics of a
a
cylinder placed near
plane surface.
/
thesG534
The drag and lift characteristics of a c
3 2768 002 13060 1

DUDLEY KNOX LIBRARY

Göktun, 1975 PDF

Uploaded by

Copyright:

Available Formats

Göktun, 1975 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Göktun, 1975 PDF

Uploaded by

Copyright:

Available Formats

Calhoun: The NPS Institutional Archive

Theses and Dissertations 1. Thesis and Dissertation Collection, all items

The drag and lift characteristics of a cylinder

Downloaded from NPS Archive: Calhoun

REPORT DOCUMENTATION PAGE READ INSTRUCTIONS

4. TITLE (and Subtitle) 5. TYPE OF REPORT A PERIOD COVERED

7. AUTHORf*,) • - CONTRACT OR GRANT NbM8ER(-»J

16. DISTRIBUTION ST ATEMEN T (o( (fits Roport)

Approved for public release; distribution unlimited.

18. SUPPLEMENTARY NOTES

drag coefficient as a function of gap size was found to

S/N 0102-014- 6601 |

(20. ABSTRACT Continued)

Submitted in partial fulfillment of the

MASTER OF SCIENCE IN MECHANICAL ENGINEERING

NAVAL POSTGRADUATE SCHOOL

Surface pressure, drag and lift coefficients have been

experimentally determined for a right circular cylinder

located near a plane surface and placed in a cross flow of

air. Parametric studies were carried out for Reynolds

numbers varying from 90,000 to 250,000, three plate lengths

and a variety of cylinder to plate spacings. The variation

of the drag coefficient as a function of gap size was found

to exhibit an interesting and unexpected trend. The drag

was a maximum at a gap size of approximately one cylinder

radius. Flow visualization studies together with detailed

measurements of the vortex shedding frequency in the cylinder

the vortex street in the cylinder wake when it is located

within a cylinder radius of the cylinder. This interference

disturbs the cylinder base pressure which in turn influences

the magnitude of the drag coefficient.

III. EXPERIMENTAL APPARATUS 18

B. PRESSURE COEFFICIENT EXPERIMENT 20

C. FLOW VISUALIZATION EXPERIMENT 22

D. VORTEX FREQUENCY EXPERIMENT 25

IV. EXPERIMENTAL PROCEDURE 28

A. PRESSURE COEFFICIENT EXPERIMENT 29

B. FLOW VISUALIZATION EXPERIMENT 30

C. VORTEX FREQUENCY EXPERIMENT 30

V. RESULTS AND DISCUSSION 33

A. PRESSURE EXPERIMENT AND CYLINDER

B. PRESSURE DRAG AND LIFT COEFFICIENTS — 51

C. FLOW VISUALIZATION RESULTS 56

D. VORTEX FREQUENCY RESULTS 66

APPENDIX A. PRESSURE DATA AND DATA REDUCTION 70

APPENDIX B. CALCULATION OF PRESSURE DRAG AND

APPENDIX C. UNCERTAINTY ANALYSIS 75

APPENDIX D. POLAR PLOTS OF PRESSURE COEFFICIENT

LIST OF REFERENCES 113

INITIAL DISTRIBUTION LIST 114

1. Schematic diagram of plate-cylinder arrangement 14

2. Schematic diagram of the wind tunnel 19

3. Photograph of the cylinder used to collect

4. Photograph of the plate and cylinder arrangement

5. Photograph of the plate and cylinder arrangment

6. Sketch of a typical subcritical flow pattern

7. Sketch of a typical critical flow pattern

8. Sketch of a typical supercritical flow pattern

9. Pressure coefficient and local Nusselt number