C h a p t e r 1.

F l o w a r o u n d a cylinder i n
steady current

1.1 R e g i m e s o f f l o w a r o u n d a s m o o t h , circular cylinder

The non-dimensional quantities describing the flow around a smooth circu­

lar cylinder depend on the cylinder Reynolds number

Re = — (1.1)
u

in which D is the diameter of the cylinder, U is the flow velocity, and u is the
kinematic viscosity. The flow undergoes tremendous changes as the Reynolds
number is increased from zero. The flow regimes experienced with increasing Re
are summarized in Fig. 1.1. Fig. 1.2, on the other hand, gives the definition
sketch regarding the two different flow regions referred to in Fig. 1.1, namely the
wake and the boundary layer. While the wake extends over a distance which is
comparable with the cylinder diameter, D, the boundary layer extends over a very
small thickness, 6. which is normally small compared with D. The boundary layer
thickness, in the case of laminai' boundary layer, for example, is (Schlichting, 1979)

LRD1051166014
2 Chapter 1: Flow around a cylinder in steady current

No separation.
Creeping flow Re < 5

A fíxed pair of
symmetric vortices 5 < Re < 4 0

Laminar
vortex 4 0 < Re < 2 0 0
street

Transition
to turbulence 200 < Re < 300
in the wake

Wake completely turbulent. 300 < Re < 3x10®

A:Lamlnar boundary layer
separation Subcrl tical

A:Lamlnar boundary
layer separation
S x 1 0 % R e < 3 . 5 x lo''
BiTurbulent boundaiy
layer separatlon;but Critical (Lower transition)
boundary layer laminar

B: Turbulent boundary 3.5 X 10® < Re < 1.5 x 10®

layer separation;the
boundary layer partly
laminar partly turbulent Supercritical

0 ; Boundary layer com­ 1.5xlO®< Re < 4x10®

pletely turbulent a t
one side Upper transition

4xl0®< Re
C: Boundary layer comple­
tely turbulent a t Transcritical
two sides

Figure 1.1 Regimes of flow around a smooth, circular cylinder in steady

current.
 Regimes of flow around a smooth, circular cylinder 3

(1.2)

a n d it is seen t h a t 6/D << 1 for Re larger t h a n O(IOO), say.

Wake region
Bo
laj
'Separatie

Incc
flow

Figure 1.2 Definition sketch.

Now, returning t o Fig. 1.1, for very small values of Re n o separation occurs.
T h e separation first appears when Re becomes 5 (Figs. 1.1a).
For t h e range of t h e Reynolds number 5 < Re < 40, a fixed pair of vortices
forms in t h e wake of t h e cylinder (Fig. 1.1 b). T h e length of this vortex formation
increases with Re (Batchelor, 1967).
W h e n t h e Reynolds number is f u r t h e r increased, t h e wake becomes unsta­
ble, which would eventually give b i r t h t o t h e phenomenon called vortex shedding
in which vortices are shed alternately a t either side of t h e cylinder a t a certain
frequency. Consequently, t h e wake h a s a n appearance of a vortex street (see Fig.
1.3d-f).
For t h e range of t h e Reynolds number 40 < Re < 200 t h e vortex street is
laminar (Fig. 1.1c). T h e shedding is essentially two-dimensional, i.e., it does not
vary in t h e spanwise direction (Williamson, 1989).
W i t h a f u r t h e r increase in Re, however, transition t o turbulence occurs in
t h e wake region (Fig. l . l d ) . T h e region of transition t o turbulence moves towards
t h e cylinder, as Re is increased in t h e range 200 < Re < 300 (Bloor, 1964). Bloor
(1964) reports t h a t a t Re = 400, t h e vortices, once formed, are turbulent. Obser­
vations show t h a t t h e two-dimensional feature of t h e vortex shedding observed in
t h e range 40 < Re < 200 becomes distinctly three-dimensional in this range (Ger-
r a r d , 1978 a n d Williamson, 1988); t h e vortices are shed in cells in t h e spanwise
direction. (It may b e noted t h a t this feature of vortex shedding prevails for all
t h e other Reynolds number regimes Re > 300. This topic will b e studied in some
detail in Section 1.2.2 in t h e context of correlation length).
For Re > 300, t h e wake is completely turbulent. T h e boundary layer over
t h e cylinder surface remains laminar, however, for increasing Re over a very wide
Chapter 1: Flovj around a cylinder in steady current

a) R e = 3 2

b) R e = 5 5

c) Re = 6 5

d) R e = 7 3

e) R e = 1 0 2

f) R e = 1 6 1

Figure 1.3 Appearance of vortex shedding behind a circular cylinder m a

stream of oil (from Homann, 1936) with increasing Re.
 Regimes of flow around a smooth, circular cylinder 5

range of Re, namely 300 < Re < 3 x 10®. T h i s regime is known a s t h e subcritical
flow regime (Fig. l . l e ) .
W i t h a f u r t h e r increase i n Re, transition t o turbulence occurs i n t h e bound­
ary layer itself. T h e transition first takes place a t t h e point where t h e boundary
layer separates, a n d t h e n t h e region of transition t o turbulence moves upstream
over t h e cylinder surface towards t h e stagnation point as Re is increased (Figs,
l.lf-l.li). '
I n t h e narrow Re b a n d 3 x 10® < Re < 3.5 x 10® (Fig. l . l f ) t h e boundary
layer becomes turbulent a t t h e separation point, b u t this occurs only a t one side
of t h e cylinder. So t h e boundary layer separation is turbulent a t one side of t h e
cylinder a n d laminar a t t h e other side. T h i s flow regime is called t h e critical (or
t h e lower transition) flow regime. T h e flow asymmetry causes a non-zero m e a n
lift o n t h e cylinder, as seen f r o m Fig. 1.4.

Figure 1.4 Non-zero mean lift in the critical-flow regime (3 X 10® < Re <
3.5 X 10®). Schewe (1983).

T h e side a t which t h e separation is turbulent switches f r o m one side t o t h e

other occasionally (Schewe, 1983). Therefore, t h e lift changes direction, as t h e
one-sided transition t o turbulence changes side, shifting f r o m one side t o t h e other
(Schewe, 1983).
T h e next Reynolds number regime is t h e so-called supercritical flow regime
where 3.5 x 10® < Re < 1.5 x 10® (Fig. l . l g ) . I n this regime, t h e boundary
layer separation is turbulent o n b o t h sides of t h e cylinder. However, transition
t o turbulence i n t h e boundary layer h a s n o t been completed yet; t h e region of
transition t o turbulence is located somewhere between t h e stagnation point a n d
t h e separation point.
T h e b o u n d a r y layer on one side becomes fully turbulent when Re reaches t h e
value of a b o u t 1.5 x 10®. So, in this flow regime, t h e b o u n d a r y layer is completely
turbulent o n one side of t h e cylinder a n d partly laminar a n d partly turbulent on
6 Chapter 1: Flow around a cylinder in steady current

t h e other side. T h i s type of flow regime, called t h e upper-transition flow regime,

prevails over t h e range of Re, 1.5 x 10® < Re < 4.5 X 10® (Fig. l . l h ) .
Finally, when Re is increased so t h a t Re > 4.5 x 10®, t h e b o u n d a r y layer
over t h e cylinder surface is virtually turbulent everywhere. T h i s flow regime is
called t h e transcritical flow regime.
Regarding t h e terminology in relation t o t h e described flow regimes a n d also
t h e ranges of Re i n which they occur, there seems t o b e n o general consensus among
various authors (Farell, 1981). T h e preceding classification a n d t h e description
are mainly based o n Roshko's (1961) a n d Schewe's (1983) works. Roshko's work
covered t h e Reynolds number range f r o m 10® t o 10^, which revealed t h e existence
of t h e upper transition a n d t h e transcritical regimes, while Schewe's work, covering
t h e range 2.3 x 10^ < Re < 7.1 x 10®, clarified f u r t h e r details of t h e flow regimes
from t h e lower transition t o t h e transcritical flow regimes.

1.2 V o r t e x s h e d d i n g

T h e most important feature of t h e flow regimes described i n t h e previ­

ous section is t h e vortex-shedding phenomenon, which is common t o all t h e flow
regimes for Re > 40 (Fig. 1.1). For these values of Re, t h e b o u n d a r y layer over
t h e cylinder surface will separate due t o t h e adverse pressure gradient imposed by
t h e divergent geometry of t h e flow environment a t t h e rear side of t h e cylinder.
As a result of this, a shear layer is formed, a s sketched i n Fig. 1.5.
As seen f r o m Fig. 1.6, t h e boundary layer formed along t h e cylinder contains
a significant amount of vorticity. This vorticity is fed into t h e shear layer formed
downstream of t h e separation point a n d causes t h e shear layer t o roll u p into a
vortex with a sign identical t o t h a t of t h e incoming vorticity. (Vortex A in Fig.
1.5).
Likewise, a vortex, rotating in t h e opposite direction, is formed a t t h e other
side of t h e cylinder (Vortex B).

Mechanism o f vortex shedding

It has been mentioned in t h e previous section t h a t t h e pair formed by
these two vortices is actually unstable when exposed t o t h e smaU disturbances
for Reynolds numbers Re > 40. Consequently, one vortex will grow larger t h a n
t h e other if Re > 4 0 . Further development of t h e events leading t o vortex shedding
has been described by Gerrard (1966) in t h e following way.
T h e larger vortex (Vortex A in Fig. 1.7a) presumably becomes strong
enough t o draw t h e opposing vortex (Vortex B) across t h e wake, as sketched in
Fig. 1.7a. T h e vorticity in Vortex A is in t h e clockwise direction (Fig. 1.5b), while
t h a t in Vortex B is in t h e anti-clockwise direction. T h e approach of vorticity of
 Vortex shedding 1

a)

Stagnation
point

S h e a r layer
Vorticity

Boundary
layer

D e t a i l e d p i c t u r e of
flow n e a r s e p a r a t i o n

Figure 1.5 The shear layer. The shear layers on both sides roll up to form
the lee-wake vortices, Vortices A and B.

Boundary
layer

Figure 1.6 Distribution of velocity and vorticity in the boundary layer, a;

is the vorticity, namely o) =
8 Chapter 1: Flow around a cylinder in steady current

t h e opposite sign will t h e n cut off f u r t h e r supply of vorticity t o Vortex A f r o m its

boundary layer. T h i s is t h e instant where Vortex A is shed. Being a free vortex,
Vortex A is t h e n convected downstream b y t h e flow.
Following t h e shedding of Vortex A, a new vortex will b e formed a t t h e
same side of t h e cylinder, namely Vortex C (Fig. 1.7b). Vortex B will now play
t h e same role as Vortex A, namely it will grow i n size a n d strength so t h a t it will
draw Vortex C across t h e wake (Fig. 1.7b). T h i s will lead t o t h e shedding of
Vortex B. T h i s process will continue each t i m e a new vortex is shed a t one side
of t h e cylinder where t h e shedding will continue t o occur i n a n alternate m a n n e r
between t h e sides of t h e cylinder.

b)

Figure 1.7 (a): Prior to shedding of Vortex A, Vortex B is being drawn

across the wake, (b): Prior t o shedding of Vortex B, Vortex C
is being drawn across the wake.

T h e sequence of photographs given i n Fig. 1.8 illustrates t h e time develop­

ment of t h e process during t h e course of shedding process.
O n e implication of t h e foregoing discussion is t h a t t h e vortex shedding oc­
curs only when t h e two shear layers interact with each other. If this interaction
is inhibited in one way or another, for example by p u t t i n g a splitter plate a t t h e
downstream side of t h e cylinder between t h e two shear layers, t h e shedding would
b e prevented, a n d therefore n o vortex shedding would occur in this case. Also, a s
another example, if t h e cylinder is placed close t o a wall, t h e wall-side shear layer
will not develop as strongly a s t h e opposing shear layer; this will presumably lead
t o a weak interaction between t h e shear layers, or t o practically n o interaction if
t h e cyHnder is placed very close t o t h e wall. I n such situations, t h e vortex shed-
 Voriex shedding S

F i g u r e 1 . 8 T i m e d e v e l o p m e n t o f v o r t e x $ h e d d i n g dxiring a p p r o x i m a t e l y
t w o - t h i r d of t h e s h e d d i n g p e r i o d . H e = 7 x 1 0 ^ .
10 Chapter 1: Flow around a cylinder in steady current

ding is suppressed. T h e effect of close proximity of a wall o n t h e vortex shedding

will be examined i n some detail later i n t h e next section.

1.2.1 Vortex-shedding frequency

T h e vortex-shedding frequency, when normalized with t h e flow velocity U

and t h e cylinder diameter D, can on dimensional grounds b e seen t o b e a function
of t h e Reynolds number:
St = St{Re) (1.3)

in which
SÇJí = —
îv^ A\ (1.4)

and is t h e vortex-shedding frequency. T h e normalized vortex-shedding fre­

quency, namely 5i,"is called t h e StrouhaJ number. Fig. 1.9 illustrates how t h e
Strouhal number varies with R e , while Fig. 1.10 gives t h e power spectra corre­
sponding t o Schewe's (1983) d a t a shown i n Fig. 1.9.

St
0.4

0.3

0.2

Re
0.0
10 10'

V Subcrltical ł S uper- ł Transcritlcal

Laminar Transition critical
vortex to t u r b u l e n c e
street in t h e w a k e Critical, Upper
or lower Transition
transition

Figure 1.9 Strouhal number for a smooth circular cylinder. Experimental

data from: Solid curve: Williamson (1989). Dashed curve:
Roshko (1961). Dots: Schewe (1983).
 Vortex shedding 11

Subcrltlcal

4.0 Re = 1 . 3 X 10®

2.0

Supercritical

0 . 0 0 8 •• Re = 7 . 2 X 10®

0.008

Re = 1.9 X 10®

0.02 Upper transition

R e = 3 . 7 X 10®

0.02
B e g i n n i n g of t r a n s c r i t i c a l

Re = 5 . 9 X 10®

0.075

Transcritical

0.2 0.4 0.6 Re = 7 . 1 X 10®

fD/U

Figure 1.10 Power spectra of the lift oscillations corresponding to Schewe's

data in Fig. 1.9 (Schewe, 1983).
12 Chapter 1: Flow around a cylinder in steady current

T h e vortex shedding first appears a t Re = 40. From Fig. 1.9, t h e shedding

frequency St is approximately 0.1 at this Re. It t h e n gradually increases as Re
is increased a n d a t t a i n s a value of about 0.2 at Re = 300, t h e lower end of t h e
subcritical flow regime. From this Re number onwards throughout t h e subcritical
range St remains practically constant (namely, a t t h e value of 0.2).
T h e narrow-band spectrum w i t h t h e sharply defined dominant frequency
in Fig. 1.10a indicates t h a t vortex shedding i n t h e subcritical range occurs i n a
well-defined, regular fashion.
As seen f r o m Fig. 1.9, t h e Strouhal frequency experiences a sudden j u m p
a t Äe = 3 — 3.5 X 10^, namely i n t h e critical Re number range, where St increases
f r o m 0.2 t o a value of about 0.45. T h i s high value of St is maintained over a
r a t h e r large p a r t of t h e supercritical Re range, subsequently it decreases slightly
with increasing Reynolds number.
T h e large increase in St in t h e supercritical-flow range is explained as fol­
lows: in t h e supercritical flow regime, t h e b o u n d a r y layer on b o t h sides of t h e cylin­
der is turbulent a t t h e separation points. This resultsi n a delay i n t h e boundary-
layer separation where t h e separation points move downstream, a s sketched i n Fig.
1.11. This means t h a t t h e vortices (now being closer t o each other) would interact
a t a faster r a t e t h a n i n t h e subcritical flow regime, which would obviously lead t o
higher values of t h e Strouhal number.

Laminar separation Turbulent separation

i n subcritical regime in supercritical regime

Figure 1.11 Sketch showing positions of separation points at different sepa­

ration regimes.

T h e power spectrum (Fig. 1.10b) a t Re = 7.2 x 10®, a Reynolds number

which is representative for t h e supercritical range, indicates t h a t i n this Re range,
too, t h e shedding occurs in a well-defined, orderly fashion, since t h e power spec­
t r u m appears t o b e a narrow-band s p e c t r u m with a sharply defined, dominant
peak. T h e fact t h a t t h e magnitude of t h e spectrum itself is extremely small (cf.
Figs. 1.10a a n d 1.10b) indicates, however, t h a t t h e shed vortices a r e not as strong
as they a r e in t h e subcritical flow regime. A n immediate consequence of this,
as will b e shown later, is t h a t t h e lift force induced by t h e vortex shedding is
relatively weak in this Re range.
 Vortex shedding 13

T h e Strouhal number experiences yet another discontinmty when Re reaches

t h e value of about 1.5 x 10®. A t this Reynolds number, transition t o turbulence in
one of t h e b o u n d a r y layers h a s been completed (Fig. l . l h ) . So, t h e b o u n d a r y layer
at one side of t h e cylinder is completely turbulent a n d t h a t a t t h e other side of
t h e cylinder is p a r t l y laminar a n d partly turbulent, a n asymmetric situation with
regard t o t h e formation of t h e lee-wake vortices. T h i s situation prevails over t h e
whole u p p e r transition region (Fig. l . l h ) . Now, t h e asymmetry in t h e formation of
t h e lee-wake vortices inhibits t h e interaction of these vortices partially, resulting
in a n irregular, disorderly vortex shedding. This can b e seen clearly f r o m t h e
broad-band spectra in Figs. 1.10c a n d d.
T h e regular vortex shedding is re-established, however, (see t h e narrow­
b a n d power spectra in Fig. l.lOe a n d f ) , when Re is increased t o values larger
t h a n approximately 4.5 x 10®, namely t h e transcritical flow regime where t h e
Strouhal number takes t h e value of 0.25 — 0.30 (Fig. 1.9).

Effect o f surface roughness

For rough cylinders t h e normalized shedding frequency, namely t h e Strouhal

number, should b e a function of b o t h Re a n d t h e relative roughness

St = St{Re, ks/D) (1.5)

in which ks is t h e Nikuradse's equivalent s and roughness of t h e cylinder surface.

St
Smooth

-3
9 x 10

0
lo" 2 5 lO'"" 2 8 10® 2 5
,7
10
Re

Figure 1.12 Effect of surface roughness on vortex-shedding frequency.

Strouhal number against Reynolds number. Circular cylinder.
Achenbach and Heinecke (1981).
u Chapter 1: Flow around a cylinder in steady current

Fig. 1.12 illustrates t h e effect of t h e relative roughness o n t h e Strouhal num­

ber where t h e experimentally obtained St values for various values of kg/D are
plotted against Re (Achenbach a n d Heinecke, 1981). Clearly, t h e effect is signifi­
cant. From t h e figure, it is apparent t h a t , for rough cylinders with fcs/D > 3 x 1 0 " ' ,
t h e critical ( t h e lower trajasition), t h e supercritical a n d t h e u p p e r transition flow
regimes merge into one narrow region in t h e St-Re plane, a n d t h e flow regime
switches directly t o transcritical over this narrow Re range, a n d this occurs a t very
low values of Re number. ( T h e figure indicates for example t h a t , a t Re 0.3 X 10^
for fcj/D = 30 X 1 0 " ' a n d a t Re = 1.5 x 10^ for k^/D = 3 x 1 0 " ' ) . T h i s result is
in fact anticipated, a s it is well known t h a t trajisition t o turbulence occurs much
earlier (i.e., a t much smaller values of Reynolds number) over rough walls.

E x a m p l e 1.1: N i k u r a d s e ' s e q u i v a l e n t s a n d r o u g h n e s s

I n practice th ere exists a n extremely wide variety of surface roughnesses,

f r o m small protrusions existing i n t h e texture of t h e surface itself t o extremely large
roughnesses i n t h e f o r m of mari ne growth such a s mussels a n d acorn barnacles,
etc..
Therefore, normally it is n o t a n easy task t o relate t h e roughness of t h e
surface t o some typical scale of t h e roughness elements, partly because t h e elements
are quite unevenly distributed. ( O n a loose s and bed, for example, t h e roughness
is measured t o b e 2-3 times t h e grain diameter). To tackle this problem, t h e
concept "Nikuradse's equivalent s and roughness" h a s been introduced. T h e idea
is t o relate a n y kind of roughness t o t h e Nikuradse roughness so t h a t comparison
can b e m a d e o n t h e same basis. Very systematic a n d careful measurements o n
rough pipes were carried out by Nikuradse (1933), who used circular pipes. Sand
with known grain size was glued o n t h e pipe wall inside t h e pipe. By measuring
t h e flow resistance a n d velocity profiles, Nikuradse obtained t h e following velocity
distribution law

^ = 5 . 7 5 1 o g i o f + 8.5 (1.6)
Uf Kg
which can b e p u t in t h e following f o r m

f =l l n f ï (1.7)
Uf K k,
i n which u is t h e streamwise velocity, Uf is t h e wall shear-stress velocity, k is t h e
K a r m a n constant ( = 0.4), y is t h e distance f r o m t h e wall a n d fc, is t h e height of
t h e s a n d roughness t h a t Nikuradse used in his experiments ( a detailed account of
 Vortex shedding 15

t h e subject is given b y Schlichting (1979)). T o judge a b o u t t h e roughness of a

particular surface, t h e usual practice is first t o measure t h e velocity distribution
above t h e surface i n consideration a n d then, based o n this measured velocity dis­
tribution u{y), t o determine kg, t h e Nikuradse's equivalent sand roughness of t h e
surface, f r o m Eq. 1.7.

Effect o f cross-sectional shape

Fig. 1.13 shows t h e Strouhal-number d a t a compiled by Blevins (1977) for
various non-circular cross sections, while Fig. 1.14 presents t h e Strouhal numbers
for a variety of profile shapes compiled by A S C E Task Committee (1961). Modi,
Wiland, Dikshit a n d Yokomizo (1992) give a detailed account of flow a n d vortex
shedding a r o u n d elliptic cross-section cylinders.

0.30
St
0.25 0.5

i }--5--j
0.20

0.15

0.10
4 5
10 2 3
10' 10 10
Re

Figure 1.13 Effect of cross-sectional shape on vortex-shedding frequency.

Strouhal number against Reynolds number. Blevins (1977).

As f a r as t h e large Reynolds numbers are concerned (iîe>10®), t h e vor­

tex formation process is relatively uninfluenced by t h e Reynolds number for t h e
cross sections with fixed separation points such as rectangular cylinders. So, t h e
Strouhal number m a y not undergo large changes with increasing Re for such cross-
sectional shapes, in contrast t o w h a t occurs in t h e case of circular cylinders.

Effect o f incoming turbulence

Quite often, t h e approach flow is turbulent. For example, a cylinder placed

on t h e sea b o t t o m would feel t h e approach-flow turbulence which is generated
within t h e b o t t o m boundary layer. T h e turbulence i n t h e approach flow is also
a n influencing factor with regard t o t h e vortex shedding. T h e effect of turbulence
Chapter 1: Flow around a cylinder in steady current

Profile D i m e n s i o n s Value Profile D i m e n s i o n s Value

(mm) of S t (mm) of S t

t=2.0 , t=1.0

T
50
A l I i
25

1 -50-
1^5|
T 1-—50—^
\-J

t=0.5 t=1.0
J _
l _ L
25 H E , ,12^5
12^1 I T
25
I h—50—H

t=1.0 t=1.0

25
T 50

-50
/ 1 — 5 0 ^

t=1.5 t== 1 . 0

J _
25
1
H 25
/
t=1.0 t=1.0

_ L
25 25

t 5 0 — I |-25--|-25-|»25H

Figure 1.14 Effect of cross-sectional shape on Strouhal number. Strouhal

numbers for profile shapes. A S C E Task Committee (1961).
 Vortex shedding 17

Profile D i m e n s i o n s Value Profile D i m e n s i o n s Value

(mm) of S t (mm) of S t

t=i.o t=i.o
•12.5
1 ^ 1 I 25
T 1 — 5 0 ^
1-2 5> 25-1

t=1.5 t=1.0

50

-50- - 1 0 0 •

Cylinder t=1.0
1 1 8 0 0 <Re< 1 9 1 0 0
_1_
25

T ' 50
25

Figure 1.14 (continued.)

on t h e vortex shedding h a s been studied by various authors, for example by Che­

u n g a n d Melbourne (1983), Kwok (1986) a n d Norberg a n d Sunden (1987) among
others. Fig. 1.15 presents t h e Strouhal number d a t a obtained by Cheung a n d
Melbourne for various levels of turbulence in their experimental tunnel. Here,
is t h e turbulence intensity defined by

in which V ÎÎ is t h e root-mean-square value of t h e velocity fluctuations a n d « is

t h e m e a n value of t h e velocity.
T h e variation of St with t h e Reynolds number changes considerably with
t h e level of turbulence i n t h e approach flow. T h e effect of turbulence is rather
similar t o t h a t of cylinder roughness. T h e critical, t h e supercritical, a n d t h e upper
transition flow regimes seem t o merge into one transitional region.
18 Chapter 1: Flow around a cylinder in steady current

0.3

0.2

Mt. Isa s t a c k full scale d a t a for I 7.8%

St 0.20 at Re

St 0.15 at Re

Re

10 5 10
6

Figure 1.15 Effect of turbulence in the approach flow on vortex-shedding

frequency. Strouhal numbers as a function of Reynolds number
for different turbulence intensities, ƒ„ is the level of turbulence
(Eq. 1.8). Cheung and Melbourne (1983).

It appears f r o m t h e figure t h a t t h e lower end of this transition range shifts

towards t h e smaller a n d smaller Reynolds numbers with t h e increased level of
turbulence. T h i s is obviously d u e t o t h e earlier transition t o turbulence in t h e
cylinder boundary layer with increasing incoming turbulence intensity.

E f f e c t o f s h e a r i n t h e i n c o m i n g flow
T h e shear in t h e approach flow is also a n influencing factor in t h e vortex
shedding process. T h e shear could b e present in t h e approach flow in two ways:
it could b e present in t h e spanwise direction along t h e length of t h e cylinder (Fig.
1.16a), or in t h e cross-flow direction (Fig. 1.16b). T h e characteristics of shear flow
around bluff bodies including t h e non-circular cross-sections have been reviewed
by Griffin (1985a a n d b). In t h e case when t h e shear is present i n t h e spanwise
direction (Fig. 1.16a), t h e vortex shedding takes place i n spanwise cells, with a
 Vortex shedding 19

frequency constant over each cell. Fig. 1.17 clearly shows this; it is seen t h a t t h e
shedding occurs i n four cells, each with a different frequency. W h e n t h e Strouhal
number is based on t h e local velocity (the dashed lines in t h e figure), t h e d a t a are
grouped a r o u n d t h e Strouhal number of about 0.25.

a) b )

Figure 1.16 Two kinds of shear in the approach flow, a: Shear is in the
spanwise direction, b: Shear is in the cross-flow direction.

Regarding t h e length of cellular structures, research shows t h a t t h e length

of cells is correlated with t h e degree of t h e shear. T h e general t r e n d is t h a t t h e
cell length decreases with increasing shear (Griffin, 1985a).
W h e n t h e shear takes place in t h e cross-stream direction (the conditions in
t h e spanwise direction being uniform), t h e shedding is only slightly influenced for
small a n d moderate values of t h e shear steepness á which is defined by

D du
(1.9)
Uc dy
For large values of s, however, the shedding is influenced somewhat substantially
(Kiya, Tamura and Arie, 1980). Fig. 1.18 shows the Strouhal number plotted
against the Reynolds number for three different values of s. A s is seen for s = 0.2,
the Strouhal number is increased substantially relative to the uniform-flow case
( . = 0),
20 Chapter 1: Flow around a cylinder in steady current

1.0
z
' 0.8

0.6
/
• /

0.4
/ - U.2
+
St

0 . 6 0 . 8 1.0 1.

0.32-
0.28-

0.24-
0.20-

—1 1 r •z / D
10 14 1Í

Figure 1.17 Effect of shear in t h e approach flow on vortex-shedding fre­

quency. Shear in t h e spanwise direction. Circles: Strouhal num­
ber based on t h e centre-line velocity Uc- Dashed lines: Strouhal
number based on t h e local velocity, î7iocal- -ñe = 2.8 X lO"^. T h e
shear steepness: s — 0.025. Maull a n d Young (1973).

s t = i p

0.24
0.24 - 0.24
0.20
0.20 0.20
0.16
0.16 0.16
0.12
0.12 0.12
2 3 3
10 10 10 10 10 Re 10

Figure 1.18 Effect of shear in t h e approach flow frequency. Shear in cross-

flow direction. T h e Strouhal number against t h e Reynolds num­
ber for three different values of t h e shear steepness s. Hatched
band: Uniform-flow results. Circles: Shear-flow results. Kiya et
al. (1980).
 Vortex shedding 21

Effect of wall proximity

T h i s topic is of direct relevance with regard t o pipelines. W h e n a pipeline is
placed on a n erodible sea b e d , scour m a y occur below t h e pipe d u e t o flow action.
This m a y lead t o suspended spans of t h e pipeline where t h e pipe is suspended
above t h e bed with a small gap, usually in t h e range f r o m 0 ( 0 . I D ) t o 0 ( 1 D ) .
Therefore it is important t o know w h a t kind of changes t a k e place i n t h e flow
around a n d in t h e forces o n such a pipe.

Stagnation

Stagnation

777777777777777777777

Figure 1.19 Flow around a) a free cylinder, b) a near-wall cylinder. S =

separation points.

W h e n a cylinder is placed near a wall, a number of changes occur in t h e

flow a r o u n d t h e cylinder. These changes a r e summarized as follows:
1) Vortex shedding is suppressed for t h e gap-ratio values smaller t h a n about
e/D = 0.3, as will b e seen later i n t h e section. Here, e is t h e g a p between t h e
cylinder a n d t h e wall.
2) T h e stagnation point moves t o a lower angular position as sketched in
Fig. 1.19. T h i s can b e seen clearly f r o m t h e pressure measurements of Fig. 2.20a
a n d Fig. 2.20b where t h e m e a n pressure distributions around t h e cylinder are
given for th r e e different values of t h e g a p ratio. While t h e stagnation point is
located at about > (/ = 0° when e / D = 1, it moves t o t h e angular position of about
4) = —40° when t h e g a p ratio is reduced t o e/D = 0.1.
3) Also, t h e angular position of t h e separation points changes. T h e sepa­
ration point a t t h e free-stream side of t h e cylinder moves u p s t r e a m a n d t h a t a t
t h e wall side moves downstream, as shown in t h e sketch given in Fig. 1.19. T h e
22 Chapter 1: Flow around a cylinder in steady current

] Free-Stream side
140 separation point

120 -

100 -
//////////////

80

60 -
Cfl o
-e-

(b)
1 1 1

140

120 _
Wall s i d e
100 o separation point

80 -

60
1 1 1 I 1 1 1 1 1 1 1 1 1 1 k.
0 1 2 e/D

Figure 1.20 Angle of separation as a function of the gap ratio, (a): At the
free-stream side of the cylinder and (b): At the wall side of the
cylinder. Re = 6 X 10'. Jensen and Sumer (1986).

separation angle measured for a cylinder with Äe = 6 x 10' is shown in Fig. 1.20;
the figure indicates t h a t for example for e/D = 0.1 the separation angle at the
free-stream side is ^ = 80°, while it is <^ = —110° at the wall side for the same gap
ratio.
4) Finally, the suction is larger on the free-stream side of the cylinder than
on the wall-side of the cylinder, as is clearly seen in Fig. 2.20b and c. When the
cylinder is placed away from the wall, however (Fig. 2.20a) this effect disappears
and the symmetry is restored.
 Vortex shedding 23

At A At B

- 2 . 0 -2.0

= 2
-3.0 -3.0

-4.0
e
0 . 8 1.6 2 . 4 0 0.8 1.6 2.4
Log. f r e q u e n c y (Hz) Log. f r e q u e n c y (Hz)

C
i)
TS •2.0 -2.0

b) 0 . 3
-3.0 -3.0

tifl - 4 . 0 -4.0
o
J 0 0 . 8 1.6 2 . 4 0 0 . 8 1.6 2 . 4
Log. f r e q u e n c y (Hz) Log. f r e q u e n c y (Hz)

-- - 2 . 4 -2.4
c) 0.2
-3.2 -3.2

^ -4.0 -4.0
• • '
0 0 . 8 1.6 2 . 4 0 0 . 8 1.6 2 . 4
Log. f r e q u e n c y (Hz) Log. f r e q u e n c y (Hz)

d) -2.4

-3.2

-4.0

0 0 . 8 1.6 2 . 4
Log. f r e q u e n c y (Hz)

Figure 1.21 Effect of wall proximity on vortex shedding. Power spectra

of t h e hot-wire signal received from t h e wake. Bearman and
Zdravkovich (1978).
24 Chapter 1: Flow around a cylinder in steady current

Vortex shedding may b e suppressed for a cyhnder which is placed close t o a

wall. Fig. 1.21 presents power spectra of the hot-wire signals received from b o t h
sides of the wake of a cylinder placed at different distances from a wall (Bearman
and Zdravkovich, 1978). As is clearly seen, regular vortex shedding, identified by
the sharply defined, dominant peaks in the power spectra, persists only for values
of the gap-to-diameter ratio e/D down t o about 0.3. This result, recognized first
by Bearman and Zdravkovich, was later confirmed by the measurements of Grass,
Raven, Stuart and Bray (1984). T h e photographs shown in Fig, 1.22 demonstrate
the supression of vortex shedding for gap ratios e/D below 0.3.
T h e suppression of vortex shedding is linked with the asymmetry in the
development of the vortices on the two sides of the cylinder. The free-stream-side
vortex grows larger and stronger t h a n the wall-side vortex. Therefore the interac­
tion of the two vortices is largely inhibited (or, for small e / D , totally inhibited),
resulting in partial or complete suppression of the regular vortex shedding.
Regarding the effect of wall proximity on the vortex-shedding frequency for
the range oí e/D where the vortex shedding exists, measurements show t h a t the
shedding frequency tends t o increase (yet slightly) with decreasing gap ratio. In
Fig. 1.23 are plotted t h e results of two studies, namely Grass et al. (1984) and
Raven, Stuart, Bray and Littlejohns (1985). Grass et al.'s experiments were done
in a laboratory channel with b o t h smooth and rough beds. The surface of the test
cylinder was smooth. Their results collapse onto a common curve when plotted
in the normalized form presented in the figure where 5io is the Strouhal number
for a wall-free cylinder. The d a t a points of Raven et al.'s study, on the other
hand, were obtained in a n experimental program conducted in the Severn Estuary
(UK) where a full-scale pipeline (50.8 cm in diameter with a surface roughness
of kjD = 8.5 X 10"^) was used. In both studies, St is defined by the veloc­
ity at the top of the cylinder. There are other d a t a available such as Bearman
and Zdravkovich (1978) and Angrilli, Bergamaschi and Cossalter (1982). While
Bearman and Zdravkovich's measurements indicate t h a t the shedding frequency
practically does not change over t h e range 0.3 < e/D < 3, Angrilli et al.'s mea­
surements show t h a t there is a systematic (yet, slight) increase in the shedding
frequency with decreasing gap ratio in their measurement range 0.5 < e/D < 6
(they report a 10% increase in the shedding frequency at e/D = 0.5).
It is apparent from the existing d a t a t h a t the vortex-shedding frequency is
insensitive t o the gap ratio, although there seems t o b e a tendency t h a t it increases
slightly with decreasing gap ratio. This shght increase in the Strouhal frequency
may be attributed t o the fact t h a t the presence of the wall causes the wall-side
vortex t o be formed closer t o the free-stream-side vortex. As a result of this, the
two vortices interact at a faster rate, leading t o a higher St frequency.
 Vortex shedding

d ) = 0 . 0 5

1.22 Effectof wall proximityo n vortex shedding. F l o w in t h e wake

o f a n e a r - w a l l c y l i n d e r . S h e d d i n g i s a p p a r e t k t f o r e/D = 0.4
a n d 0 . 3 b u t s u p p r e s s e d f o r c / Z ? = 0 . 2 a n d 0 . 0 5 . Re = 7 x 10'.
26 Chapter 1: Flow around a cylinder in steady current

Vortex s h e d d i n g

Figure 1.23 Effect of wall proximity on vortex shedding frequency. Nor­

malized Strouhal number as a function of gap ratio. St^ is the
Strouhal number for wall-free cylinder. Circles; Raven et al.
(1985). Solid curve: Grass et al. (1984).

Jensen, Sumer, Jensen a n d Fredsøe (1990) investigated t h e flow around a

pipeline (placed initially on a f l a t bed) a t five characteristic stages of t h e scour
process which take place underneath t h e pipeline. Each stage was characterized in
t h e experiments by a special, frozen scoured bed profile, which was a n exact copy
of t h e measured bed profile of a n actual scour test. T h e investigated scour profiles
a n d t h e corresponding m e a n flow field a r e shown i n Fig. 1.24. It was observed
t h a t n o vortex shedding occurred for t h e first two stages, namely stages I a n d II,
while vortex shedding did occur for stages III - V. Fig. 1.25 depicts t h e shedding
frequency corresponding t o t h e different stages.
T h e variation of t h e Strouhal number, which goes f r o m a£ high a value as
0.36 for Stage III t o a n equilibrium value of 0.17 in Stage V , can b e explained by
t h e geometry of t h e downstream scour profile as follows.
For profiles III a n d IV, t h e steep slope of t h e u p s t r e a m p a r t of t h e d u n e
behind t h e cylinder forces t h e shear layer originating f r o m t h e lower edge of t h e
cylinder t o b e n d upwards, t h u s causing t h e associated lower vortex t o interact
with t h e u p p e r one prematurely, leading t o a premat ure vortex shedding. T h e
result of this is a higher vortex shedding frequency a n d a very narrow formation
region. T h e flow visualization s t u d y carried out in t h e s a m e experiments (Jensen
et al., 1990) confirmed t h e existence of this narrow region.
 Vortex shedding

y(cm)
6
4 20 cm/s
2 I
0

-* - 'ÍS
y7Z^777777777777?7777777777777777

/ ' : III
~Z777777777777777777P7777777

. r r IV

yyyyyyy yyyyyyyyyyyyyyyyyyyy

-4 -3 -2 -1 0 1 2 3 4 5 6 7 8 x/D

Figure 1.24 Vector plot of the mean velocities, S = the approximate posi­
tion of the stagnation point. Jensen et al. (1990).

time
(min)

Figure 1.25 Time development of Strouhal number during the scour process
below a pipeline. Jensen et al. (1990).
28 ChapteT 1: Flow around a cylinder in steady current

1.2.2 Correlation length

As h a s been mentioned in Section 1.1, vortex shedding in t h e turbulent

wake regime (i.e. Ä e > 2 0 0 ) occurs in cells along t h e length of t h e cyUnder.
These spanwise cell structures are visualized in Fig. 1.26 which shows t h e
time evolution of t h e shedding process in p l a n view.
T h e cells are quite clear f r o m t h e photographs i n Fig. 1.26. Shedding does
not occur uniformly along t h e length of t h e cylinder, b u t rather in cells (designated
by A, B a n d C in Fig. 1.26). It can also b e recognized f r o m t h e pictures in Fig.
1.26 t h a t t h e cells along t h e length of t h e cylinder are o u t of phase. Consequently,
t h e maximum resultant force acting o n t h e cylinder over its total length may b e
smaller t h a n t h e force acting o n t h e cylinder over t h e length of a single cell.
T h e average length of t h e cells m a y b e termed t h e correlation length. T h e
precise determination of t h e correlation length requires experimental determina­
tion of t h e spanwise variation of t h e correlation coefficient of some unsteady quan­
tity related t o vortex shedding, such as fluctuating surface pressure, or a fluctuat­
ing velocity j u s t outside t h e shear layer at separation.
T h e correlation coefficient is defined by

p'(C) p'(C + ^)
(1.10)

in which Ç is t h e spanwise distance, z is t h e spanwise separation between two

measurement points, a n d p' is t h e fluctuating p a r t of t h e unsteady quantity in
consideration. T h e overbar denotes t h e time averaging. T h e correlation length L,
on t h e other h a n d , is defined b y t h e integral

(1.11)

Fig. 1.27 gives a typical example of t h e correlation coefficient obtained in a

wind tunnel with a cylinder 7.6 cm i n diameter a n d 91.4 c m in length with large
streamlined end plates (Novak a n d Tanaka, 1977). T h e Reynolds number was
1.9 X 10^. T h e measured quantity was t h e surface pressure a t a n angle 60° t o t h e
main stream direction. T h e correlation length corresponding t o t h e correlation
coefficient, given i n Fig. 1.27, on t h e other h a n d is found t o b e L/D = 3 f r o m Eq.
1.11.
Figure 1.26 Photographs, illustrating the time evolution of spanwise cell structure. Cylinder smooth. Re = 6x 10®
30 Chapter 1: Flow around a cylinder in steady current

For a smooth cyhnder, t h e correlation length changes with t h e Reynolds

number. Table 1.1 presents t h e correlation-length d a t a compiled by King (1977).

Table 1.1 Correlation lengths and Reynolds numbers of smooth cylin­

ders.

Reynolds number Correlation Source

length

40 < ü e < 150 (15-20)D Gerlach and Dodge (1970)

150 <Re< 10^ (2-3)D Gerlach and Dodge (1970)
1.1 X 10^ <Re< 4.5 X 10^ (3-6)D El-Baroudi (1960)
> 10^ 0.5D Gerlach and Dodge (1970)
2 X 10® 1.56D Humphreys (1960)

T h e table shows t h a t t h e correlation length is (15-20)D for40 < i î e < 150

b u t experiences a sudden d r o p t o (2-3)D a t Re = 150. T h e latter Re number is
quite close t o t h e Reynolds number (see Fig. l . l d ) , a t which t h e laminar vortex
shedding regime disappears. Regarding t h e finite (although large) values of t h e
correlation length i n t h e range 40 < Äe < 150, t h e correlation length i n this flow
regime should theoretically b e infinite, since t h e vortex regime i n this range is
actually two-dimensional. However, purely two-dimensional shedding cannot b e
achieved in practice due t o t h e existing end conditions. A shght divergence f r o m
t h e purely two-dimensional shedding, in t h e f o r m of t h e so-called oblique shedding
(see for example WiUiamson, 1989), may result in finite correlation lengths.
Other factors also affect t h e correlation. T h e correlation increases consider­
ably when t h e cyhnder is oscillated in t h e cross-flow direction. Fig. 1.28 presents
t h e correlation coefficient d a t a obtained b y Novak a n d Tanaka (1977) for several
values of t h e double-amphtude-diameter ratio 2A/D where A is t h e amplitude of
cross-flow vibrations of t h e cyhnder. T h e figure shows t h a t t h e correlation coef­
ficient increases tremendously with t h e amplitude of oscillations. Similar results
were obtained b y Toebes (1969) who measured t h e correlation coefficient of fluctu­
ating velocity in t h e wake region near t h e cylinder. Fig. 1.29 presents t h e variation
of t h e correlation length as a function of t h e amplitude-to-diameter ratio (curve a
in Fig. 1.29). Clearly, t h e correlation length increases extensively with increasing
t h e amplitude of oscillations.
 Vortex shedding SI

0.8
0.6
0.4
0.2
•-z/D
0 2 4 6 8 10

Figure 1.27 Correlation coefficient of surface pressure fluctuations as func­

tion of the spanwise separation distance z. Cylinder smooth.
Re = 1.9 X 10^. Pressure transducers are located at 60° to
the main stream direction. Novak and Tanaka (1977).

1.0
2A/D = 0.20
0.8-

0.6- 0.15

0.4- 0.10

0.2-
0.05

0.0
0 2 4 10

Figure 1.28 Effect of cross-flow vibration of cylinder on correlation co­

efficient of surface pressure fluctuations. Cylinder smooth.
Re = 1.9 X 10^. Pressure transducers are located at 60° to
the main stream direction. A is the amplitude of the cross-flow
vibrations of cylinder. Novak and Tanaka (1977).

Turbulence in the approaching flow is also a significant factor for the corre­
lation length, as is seen from Fig. 1.29. The turbulence in the tests presented in
this figure was generated by a coarse grid in the experimental tunnel used in Novak
and TanaJia's (1977) study. T h e figure indicates t h a t t h e presence of turbulence
S2 Chapter 1: Flow around a cylinder in steady current

in the approaching flow generally reduces the correlation length. It is interesting

t o note t h a t with 2 A / D = 0.2, while the correlation length increases from about 3
diameters t o 43 diameters for a turbulence-free, smooth flow, the increase is not so
dramatic when some turbulence is introduced into the flow; t h e correlation length
increases t o only about 10 diameters in this latter situation.

L
D
40 Flow:

30 a: Smooth

20 b: Turbulent

10

0 2A/D
0 0.1 0.2

Figure 1.29 Correlation length. Cylinder smooth. Re = 1.9 X lO''. Pres­

sure transducers are located at 60° to the main stream direc­
tion. A is the amplitude of cross-flow vibrations of the cylinder.
Turbulence in the tunnel was generated by a coarse grid, and
its intensity, = 11%. Novak and Tanaka (1977).

The subject has been most recently studied by Szepessy and Bearman
(1992). These authors studied the effect of the aspect ratio (namely the cylin­
der length-to-diameter ratio) on vortex shedding by using moveable end plates.
They found t h a t the vortex-induced lift showed a maximum for a n aspect ratio of
1, where the lift could be almost twice the value for very large aspect ratios. This
increase of the lift amplitude was found t o be accompanied by enhanced spanwise
correlation of the flow.
Finally, it may be noted t h a t Ribeiro (1992) gives a comprehensive review
of the literature on oscillating lift on circular cylinders in cross-flow.
 References SS

REFERENCES

Achenbach, E . a n d Heinecke E . (1981): O n vortex shedding f r o m smooth a n d

rough cyhnders i n t h e range of Reynolds numbers 6 x 10® t o 5 x 10®. J .
Fluid Mech., 109:239-251.

Angrilh, F., Bergamaschi, S. a n d Cossalter, V. (1982): Investigation of wall-

induced modifications t o vortex shedding f r o m a circular cylinder. Trans,
of t h e ASME, J . Fluids Engrg., 104:518-522.

A S C E Task Committee on W i n d Forces (1961): W i n d forces o n structures. Trans.

ASCE, 126:1124-1198.

Batchelor, G.K. (1967): An Introduction t o Fluid Dynamics. Cambridge Univer­

sity Press.

Bearman, P.W. a n d Zdravkovich, M.M. (1978): Flow around a circular cylinder

near a plane boundary. J . Fluid Mech., 89(l):33-48.

Blevins, R.D. (1977): Flow-induced Vibrations. Van Nostrand.

Bloor, M.S. (1964): T h e transition t o turbulence in t h e wake of a circular cylinder.

J . Fluid Mech., 19:290-304.

Cheung, J.C.K, a n d Melbourne, W.H. (1983): Turbulence effects on some aerody­

namic parameters of a circular cylinder a t supercritical Reynolds numbers.
J . of W i n d Engineering a n d Industrial Aerodynamics, 14:399-410.

El-Baroudi, M.Y. (1960): Measurement of Two-Point Correlations of Velocity

near a Circular Cylinder Shedding a K a r m a n Vortex Street. University of
Toronto, UTIAS, TN31.

Farell, C. (1981): Flow around fixed circular cylinders: Fluctuating loads. Proc. of
ASCE, Engineering Mech. Division, 107:EM3:565-588. Also see t h e closure
of t h e paper. Journal of Engineering Mechanics, ASCE, 109:1153-1156,
1983.

Gerlach, C.R. a n d Dodge, F . T . (1970): An engineering approach t o t u b e flow-

induced vibrations. Proc. Conf. o n Flow-Induced Vibrations in Reactor
System Components, Argonne National Laboratory, pp. 205-225.

Gerrard, J . H . (1966): T h e mechanics of t h e formation region of vortices behind

bluff bodies. J . Fluid Mech., 25:401-413.
Si Chapter 1: Flow around a cylinder in steady current

Gerrard, J.H. (1978): The wakes of cylindrical bluff bodies at low Reynolds
number. Phil. Transactions of the Royal Soc. London, Series A,
288(A1354):351-382.

Grass, A.J., Raven, P.W.J., Stuart, R.J. and Bray, J.A. (1984): T h e influence
of boundary layer velocity gradients and bed proximity on vortex shedding
from free spanning pipelines. Trans. ASME, J . of Energy Res. Technology,
106:70-78.

Griffin, O.M. (1985a): Vortex shedding from bluff bodies in a shear flow; A Re­
view. Trans. ASME, J . Fluids Eng., 107:298-306.

Griffin, O.M. (1985b): The effect of current shear on vortex shedding. Proc. Int.
Symp. on Separated Flow Around Marine Structures. T h e Norwegian Inst,
of Technology, Trondheim, Norway, J u n e 26-28, 1985, pp. 91-110.

Homann, F . (1936): Einfluss grosser Zähigkeit bei Strömung u m Zylinder.

Forschung auf dem Gebiete des Ingenieurwesen, 7(1):1-10.

Humphreys, J.S. (1960): O n a circular cylinder in a steady wind at transition

Reynolds numbers. J . Fluid Mech., 9:603-612.

Jensen, B.L. and Sumer, B.M. (1986): Boundary layer over a cylinder placed near
a wall. Progress Report No. 64, Inst, of Hydrodynamics and Hydraulic
Engineering, ISVA, Techn. Univ. Denmark, pp. 31-39.

Jensen, B.L., Sumer, B.M., Jensen, H.R. and Fredsøe, J . (1990): Flow around and
forces o n a pipeline near a scoured bed in steady current. Trans, of t h e
ASME, J . of Offshore Mech. and Arctic Engrg., 112:206-213.

King, R. (1977): A review of vortex shedding research and its application. Ocean
Engineering, 4:141-171.

Kiya, M., Tamura, H. and Arie, M. (1980): Vortex shedding from a circular cylin­
der in moderate-Reynolds-number shear flow. J . Fluid Mech., 141:721-735.

Kwok, K.C.S. (1986): Turbulence effect on flow around circular cylinder. J . En­
gineering Mechanics, ASCE, 112(11):1181-1197.

Maull, D.J. and Young, R.A. (1973): Vortex shedding from bluff bodies in a shear
flow. J . Fluid Mech., 60:401-409.

Modi, V.J., Wiland, E., Dikshit, A.K. and Yokomizo, T . (1992): O n the fluid
dynamics of elhptic cyhnders. Proc. 2nd Int. Offshore and Polar Engrg.
Conf., San Francisco, CA, 14-19 J u n e 1992, 111:595-614.
 References 35

Nikuradse, J . (1933): Strömungsgesetze in rauhen Rohren. Forsch. Arb.Ing.-Wes.

No. 361.

Norberg, C. and Sunden, B. (1987): Turbulence and Reynolds number effects on

the flow and fluid forces on a single cylinder in cross flow. Jour. Fluids and
Structures, 1:337-357.

Novak, M. and Tanaka, H. (1977): Pressure correlations on a vibrating cylin­

der. Proc. 4th Int. Conf. on Wind Effects on Buildings and Structures,
Heathrow, U.K., Ed. by K.J. Eaton. Cambridge Univ. Press, pp. 227-232.

Raven, P.W.J., Stuart, R.J., Bray, J.A. and Littlejohns, P.S. (1985): Full-scale
dynamic testing of submarine pipeline spans. 17th Annual Offshore Tech­
nology Conference, Houston, Texas, May 6-9., paper No, 5005, 3:395-404.

Ribeiro, J.L.D. (1992): Fluctuating lift and its spanwise correlation on a circular
cylinder in a smooth and in a turbulent flow: a critical review. Jour, of
Wind Engrg. and Indust. Aerodynamics, 40:179-198.

Roshko, A. (1961): Experiments on the flow past a circular cylinder at very high
Reynolds number. J . Fluid Mech., 10:345-356.

Schewe, G. (1983): On the force fluctuations acting on a circular cylinder in cross-

flow from subcritical u p t o transcritical Reynolds numbers. J . Fluid Mech.,
133:265-285.

Schlichting, G. (1979): Boundary Layer Theory. 7.ed. McGraw-Hill Book Com­

pany.

Szepessy, S. and Bearman, P.W. (1992): Aspect ratio and end plate effects on
vortex shedding from a circular cylinder. J . Fluid Mech., 234:191-217.

Toebes, G.H. (1969): T h e unsteady flow and wake near a n oscillating cylinder.
Trans. ASME J . Basic Eng., 91:493-502.

Williamson, C.H.K. (1988): T h e existence of two stages in the transition t o three-

dimensionality of a cylinder wake. Phys, Fluids, 31(11):3165-3168.

Williamson, C.H.K. (1989): Oblique and parallel modes of vortex shedding in

the wake of a circular cylinder a t low Reynolds number. J . Fluid Mech.,
206:579-627.