J . Fluid Mech. (1968), vol. 34, part 4, p p .

625-639 625
Printed i n Cheat Britain

Distribution of local pressure and skin friction around

a circular cylinder in cross-flow up to Re = 5X106
By E. ACHENBACH
Institut fur Reaktorbauelemente, KFA-Julich, Germany

(Received 15 March 1968)

I n a large range of Reynolds numbers, 6 x lo4 < Re < 5 x 106, the flow around
single cylinders with smooth surfaces has been investigated. The high values of
the Reynolds numbers were obtained in a test channel which could be pressurized
up to 40 bar of static pressure. New experiments were performed to measure the
local pressure and skin friction distribution around the cylinder. From these
results the total drag, the pressure drag and the friction drag were calculated.
By means of the skin friction distribution the position of the separation points,
separation bubbles or transition points can be localized. These data allow one to
define three states of the flow: the subcritical flow, where the boundary layer
separates laminarly; the critical flow, in which a separation bubble, followed by
a turbulent reattachment, occurs; and the supercritical flow, where an immedi-
ate transition from the laminar to the turbulent boundary layer is observed at
a critical distance from the stagnation point. According to the total drag
coefficient the values found in this paper connect the subcritical region repre-
sented by the measurements of Wieselsberger (1923) and Fage & Warsap (1930)
with the supercritical range in which Roshko (1961) carried out his experiments.

1. Introduction
Investigations on tube banks in cross-flow have shown that, beginning a t a
critical Reynolds number, an improvement of heat transfer in connexion with
the increase of the pressure-drop coefficient is to be observed. This change is
rather important so it must be taken into account when designing heat
exchangers for extremely high Reynolds numbers. I n order to get a deeper
insight into these phenomena, we started investigations on heat transfer and
flow mechanism. I n this paper those preliminary experiments are communi-
cated which have been made with a single cylinder in cross-flow with a smooth
surface. I n a region of 6 x lo4 < Re < 5 x lo6 the distribution of the local pres-
sure and skin friction has been measured around the cylinder. With these values
available, the total drag, the pressure drag and the friction drag could be
calculated. Finally we found the position of the separation point as a function
of the Reynolds number.

40 Fluid Mech. 34

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 626 E. Achenbach

2. Experimental arrangement
The investigations were performed in two test rigs which were available in
the ' Institut fur Reaktorbauelemente, Kernforschungsanlage Julich '. The
fluid was air. The atmospherical tests were carried out in the low-pressure wind
tunnel (ND channel). The tests carried out up to 40 bar static pressure were
conducted in the pressurized wind tunnel (HD channel). The rectangular test
section (500 x 900 mm) could be mounted alternatively in the ND channel or in
the HD channel. Table 1 gives some technical data. A detailed description of the
test rigs has been published by Grosse & Scholz (1965, pp. 150-8).

Tempera- Velo-
Pressure ture city,, Mach TZL
Channel (bar) ("C) (n1is) number (%)
1
Re1na.T

ND 1 60 35 3 x 105 < 0.1 0.7

HD 1-40 60 15 4.7 x 106 < 0.05
TABLE1. Experimental data of the test channels

The test cylinder was a brass tube with a length of L = 500mm and a
diameter cl = 150mm. The maximum deviation of the tube diameter was
0.15 mm. For control another cylinder 1000 mm in length was used. The
surfaces of the tubes were polished. The total height of the roughness was 2pm,
measured in the axiaI direction. Before starting the experiments the flow condi-
tions were calibrated. The velocity profile in the test section was nearly plane
within the range of 1 yo of the mean value.
The turbulence level
TU = J-,
C".')
u
r
n

where u' is the velocity fluctuation and Urn,the undisturbed velocity, was
measured with a hot wire. The obtained value was Tu = 0-7 yo.The tempera-
ture of the fluid was controlled at a value of nearly 60°C. The deviation of the
blower rotation was kept in the range of f 0.2 yo.

3. Measurement techniques
The velocity at the entrance of the test section U, was calculated from the
mass flow, measured by means of a calibrated Venturi nozzle. The accuracy was
about 1 %. The calculated velocity was checked by a Prandtl tube. The agree-
ment was very good. The static pressure around the cylinder was measured by
means of a pressure hole (1mm of diameter). It was found nearly a t the half length
of the cylinder. At the same generator the skin friction probe, described later on,
was installed. The pressure hole and the skin friction probe were 10 mm apart.
For control another pressure hole was drilled at a distance of 30 mm (detailed
construction see figure 1).
The cylinder could be turned around its longitudinal axis from # = 0"
(stagnation point) to q5 = 360". Thus the static pressure and the skin friction

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 Measurements on flow past a circular cylinder 627
could be measured simultaneously. At the ND channel the pressure was mea-
sured by a Betz manometer. At the HD channel a set of five electronic pressure
transducers of various ranges was used. The accuracy was 0.5% of the full
deflexion.
For the dimensionless presentation of the local static pressure the term

was used. p = p ( 4 ) is the static pressure at the peripherical angle 4 of the

cylinder and p m the static pressure of the infinite flow.

Skin friction probe

Pressure control
-
Y

FIGURE
1. Test cylinder (scheme).

The local shear stresses at the wall were picked up by means of a skin friction
probe (figure 2). The application of this type of probe was communicated by
Konstantinov & Dragnysh (1955). I n Rechenberg’s (1962, 1963) experiments,
carried out in turbulent boundary layers, the reliability of the probe was con-
firmed. The principle is based on the measurement of the pressure difference at
an edge which transversely projects only some hundredths of a millimetre into
the boundary layer. The quantity of the pressure difference Ap only depends on
the wall-shear stressesr,,. The advantage of this probe, compared with other ones,
is the change of the sign in the reading out, when the flow changes direction.
This fact is important for measurements around the total circumference of the
cylinder and for localizing a recirculation. Above all it is possible to work in
extremely thin boundary layers, because the edge can be ground off as it
becomes necessary.
The procedure of the calibration must be regarded as a disadvantage.
Because of the very small size of the probe the surroundings of it must not be
changed. This means that the probe has to be calibrated while being mounted.
I n our case the calibration curve was obtained in the following way: the local
40-2

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 628 E . Achenbach
skin friction was calculated in the surroundings of the stagnation point from the
distribution of the pressure around the cylinder. The procedure used was that
of the so-called 'Blasius Reihe' (see Schlichting 1955). The values of To aud Ap
have been brought into the following relationship :

\
\

u,
Test cylinder 150

FIGURE
2. Skin friction probe, dimensions in mm.

This equation is obtained by a dimensional analysis, as it is done by Rechenberg

(1962) for a test body in a boundary layer. 7 is the viscosity of the fluid. The
term h has the dimension of a length. We ought to introduce h as the height of
the edge. I n our case, however, the calibration curve must be available for only
one height of the probe edge. Therefore an exact measurement of h was not
necessary. The calibration curve was picked up for both directions of the flow and
for the pressure range of 1 to 40 bar. There is only one line for the entire range,
if one takes the calculated values of the skin friction in the region of nearly
0" < 9 < 45O.
The small pressure differences obtained from the skin friction probe could
be measured by means of a fused quartz precision pressure gauge (Texas Instru-

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 Measurements on flow past a circular cylinder 629
ments, Houston, Texas, U.S.A.).This manometer based on the deflexion of a
Bourdon tube even works up to large values of static pressure with a high
resolution.
For the presentation of the experimental skin friction the dimensionless term

was used, which results in this form from the boundary-layer calculation.

4. Results
4.1. Pressure and skin friction distribution
The distribution of the local pressure and skin friction was measured in steps of
q5 = 5" around the circumference of q5 = 360" of the cylinder. When necessary
the step width was diminished.

g
h

0
9
5 -1 0
0 0
0
-2 0
0 0
<)o,oO
-3
0 30 60 90 120 150 180 210 240 270 300 330 360
1

I
u
F4 - 1

3. Circular cylinder: skin friction and pressure distribution. Re = 105.

FIGURE

I n the four figures 3-6 the skin friction and the corresponding pressure distri-
butions are plotted versus the peripheric angle q5 of the cylinder.?
Figure 3 shows the case of the subcritical flow a t Re = lo5. The boundary
layer separates laminarly at q5 = 78" (q5 = 282") before reaching the main cross-
t Details of the numerical measurements are available on request from the editorial
office.

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 630 E . Achenbach
section. The separation is indicated by the vanishing of the skin friction. If
T~ = 0, the velocity gradient at the wall

This is the condition for the separation of the boundary layer from the wall.
Figure 4 shows the behaviour of the flow at Re = 2.6 x 105. The flow is just
before the transition into the critical region which begins at Re = 3 x 105. The
boundary layer still separates laminarly at an angle of $ = 94' ($ = 266').
7

-2

-3
0 30 60 90 120 150 180 210 240 270 300 330 3%
1

n
N8
L3
- 0

.
Q
2
u
n
8
R -1
I
u
R

-2

FIGURE
4. Circular cylinder: skin friction and pressure distribution. Re = 2.6 x 106.

The minimum of the pressure has changed with respect to the magnitude and
the position compared with the flow at Re = lo5.The dimensionless pressure in
the back of the cylinder rises; this means a decrease of the drag coefficient.
Figure 5 shows the typical distribution of the skin friction which occurs in the
critical region. At q5 = 105" ($ = 255") there is no final separation, but a so-
called separation bubble. This means that there is a region between laminar
separation and turbulent reattachment in which the wall shear stresses theo-
retically vanish. Downstream an intensive rise of the skin friction follows,
showing values which are mostly greater than those of the laminar maximum.
It may be concluded that the boundary layer is now turbulent. At an angle of
$ = 147" ($ = 220') it separates finally.
Also the pressure distribution confirms that the flow follows the shape of the
wall up to $ = 147' ( Q = 220'). The measured distribution comes close to the
values of t,he potential theory in a large range of angle.

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 Measwements on $ow past n circular cylinder 631
This region of Reynolds number characterized by the described behaviour
may be called the critical state of the flow. I n our measurements it covers the
range 3 x lo5 < Re < 1.5 x los. Re = 3 x lo5 is called the critical Reynolds
number. The critical flowis extremely sensitive. Both turbulence level and surface
roughness have a significant influence on the flow. This can be recognized by
comparing the drag coefficient curves of several authors (figure 9). The critical
Reynolds number varies from 2 x lo5 to 5 x lo5, depending on the special flow
conditions.

Skin friction

FIGURE5. Circular cylinder: skin friction and pressure distribution. Re = 8-5 x 105.

I n the region of Reynolds number greater than 1-5x lo6 the phenomenon of
the separation bubble is no longer observed. The boundary layer changes with-
out any intermediate state from the laminar to the turbulent flow having
reached a critical distance from the stagnation point. The supercritical state of
the flow is developed. The position of the transition point is indicated by a small
rise and a following retarded drop of the skin friction distribution. The transi-
tion point shifts in the direction of the stagnation point if the Reynolds number
is increased. In figure 6 it can be localized near $ = 65" ($ = 295"). Down-
stream the boundary layer is turbulent. Due to the higher level of energy the
turbulent boundary layer is able to run against a pressure gradient from about
q5 = 85" to q5 = 115" (from $ = 275" to $ = 245O).
I n order to demonstrate the dependence of the pressure and skin friction
distribution upon the Reynolds number, the results just discussed are combined
in one diagram (figure 7 ) . I n the frontal part of the cylinder the normalized

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 632 E . Achenbach
pressure and skin friction distributions are nearly independent of the Reynolds
number. Important changes, however, are found in the rear as was to be
expected. I n particular, the pressure distribution in the back of the cylinder
causes the variation of the drag coefficient presented in figure 9.
The distribution of the static pressure in the subcritical range and in the lower
range of the critical flow state has been measured by many authors in former
times. However, the flow conditions, above all the turbulence level and the

(I"

0 30 60 90 120 150 180 210 240 2iO 300 330 360

6. Circular cylinder: skin friction and pressure distribution. R e = 3.6 x 106.

FIGURE

surface roughness, differ in nearly all cases. For the rest, these data are not
usually communicated. Therefore, one does not hope to find a satisfying agree-
ment in a comparison of the results. Nevertheless we compared, in figure 8, our
own results with those of Giedt (1951) and Fage & Falkner (1931); for they too
measured the local skin friction up to Re = 2.1 x lo5. At Re = lo5 the results
agree rather well except our own pressure distribution in the region of qi = 70".
The comparison at Re = 2.1 x 105 shows that Giedt as well as Fage & Falkner
measured the beginning of the separation bubble, while in our case a pro-
nounced laminar separation occurs. This fact seems to be due to the turbulence
level. Giedt and Fage & Falkner used a Stanton tube for the determination of
the skin friction. Therefore they were not able to measure in the region of the
recirculation.
4.2. Position of the separation point
The .r,,-curves crossing the zero line indicate the geometrical position of the
separation. In figure 10 the position of the separation points is plotted as a
function of the Reynolds number. The diagram allows one to distinguish three

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 iMeasurements o n flow past a circular cylinder 633
flow rkgimes: the subcritical, the critical and the supercritical. The sudden rise
from the subcritical to the critical flow at Re = 3 x lo5 is obvious. The separa-
tion point shifts from q5 = 95" to q5 = 140". The vanishing of the separation
bubble at Re = 1.5 x lo6indicates that the supercritical state of flow is reached.
The separation point moves in the direction of the frontal stagnation point. The

4
0 30 60 90 '% 150 180 210 240 270 300 330 360

G 1
?.
h

El8 0
9
G?
v

-1

-2

-- 3

-4

-5
1

-
s8
-
h

.
5-1
Y

8
R
I
R
u

-2

-3
FIGURE 7. Circular cylinder: comparison of the skin friction and pressure distribution at
various Reynolds numbers. . . .,Re = 1.0 x 106; -.-, Re = 2.6 x 106; --, Re = 8.5 x
106; - x -, Re = 3.6 x lo6.

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 634 E. Achenbach
separation occurs now in the range of 115" < 4 < 120". I n the subcritical
region the minimum of the separation angle at R e = 1.5 x lo5 is remarkable.
Here the lowest values of 4 = 72" have been measured. This minimum indicates
the beginning of the decreasing of the total drag coefficientC,.

- 4 O

80

FIUURE 8. Circular cylinder : skin friction and pressure distribution, comparison with other
authors. Re = lo5: -, own values; - x -, Giedt (1951); . . . .,Fage & Falkner (1931).
Re = 2.1 x 106: --, own values; -0-, Giedt (1951); -.-, Fage & Falher (1931).

4.3. Total drag coeficient

By means of the measured pressure and skin friction distribution the total drag
coefficient C, can be calculated by an integration around the tube

c -- ( & pD) U & d L '

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 Measurements o n $ow past a circular cylinder 635
I n figure 9 the results are plotted for two ratios of length to diameter L l d . It
must be expected that there is an effect of the span-diameter ratio. The scat-
tering of the results, however, does not allow one t o outline the systematic
influence. The recent investigations of Morsbach (1967) show that, in the sub-

FIGURE 9. Circular cylinder: total drag coefficient as a function of the Reynolds number.
-, Wieselsberger (1923); --, Fage & Warsap (1930); -.-, Roshko (1961); A ,
L / d = 3.33 ND channel; A, L / d = 3.33 HD channel; 0, L / d = 6.66 HD channel.

3
Subcriticdi critical -.k supercritical
I I
4; 150

140

130

120

110

100

90 Separation point .

80

70
104 2 5 105 2 106 2 5 .lo7
-Re
FIQ~RE 10. Circular cylinder :position of the separation point as a function of the Reynolds
number. A , L l d = 3.33 ND channel; A, L / d = 3.33 HD channel; 0,L / d = 6.66
HD channel.

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 636 E. Achelzbach
critical state of flow, there is no influence of the span-diameter ratio. In the
critical flow, however, an effect was observed. For small values of Lld
(Lld < 3.0) the influence was considerable, while in the case of L l d = 5 the
flow at the middle section of the cylinder was no longer influenced by the walls.
Concerning the investigations of Wieselsberger (1923) and Roshko (1961), as
well as those of Page & Warsap (1930))the turbulence level has, unfortunately,
not been communicated. An increasing turbulence level effects a decrease of the
critical Reynolds number. Therefore it may be supposed that Wieselsberger
measured a t a lower, Page & Warsap at a higher turbulence level than we did
(Tu = 0.7 yo).This assumption implies that the surfaces of the cylinders were
polished in all cases.

30

20

15

-
&?
u
10

.
Q
- 05
G
u” 30

02

FIGURE
01
104 2 4 6 lo5 2 ’i 6 1IJh

11. Circular cylinder: rate of the friction coefficient with respect to the total drag
coefficient. A, L / d = 3.33 ND channel; A, L / d = 3.33 HD channel; 0,L / d = 6.66
2 4

- 6

Kr
10’

HD channel.

Finally it must be noticed that no correction for the blockage ratio d / E ,

where H is the channel width, was made. This correction was omitted, in order
to avoid the measurements being subjected to a theory, which may not be
correct.
4.4. Friction forces
The total drag of a cylinder is composed additively of pressure forces and
friction forces. Thus the drag coefficient can be written

Figure 11 presents the percentage of friction referred to the total drag as a

function of the Reynolds number. I n the subcritical region our own values
nearly join those of Thom (1929) obtained in a semi-theoretic way. The results
published by Schiller & Linke (1933) are of the same order. It is suggested that
Thom’s values are a little too large, because he made an addition of 4 % for the
rear, while a reduction would be obvious with respect to the observed recircula-

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 Measurements on flow past a circular cylinder 637
tion. The results of Schiller & Linke (1933) may be regarded to be of moderate
accuracy because of the method used. Their results are the difference of two
large terms of nearly the same order. The rate of friction was obtained by a
comparison of the directly measured forces with the integration of the local
pressure around the cylinder. I n our case, however, the experimental skin
friction has been integrated.
By means of figure 11 we find that the friction forces are nearly unimportant
with respect to the quantity of the total drag. Nevertheless they influence the
flow causally by effecting the separation of the boundary layer.

5. Discussion of the results

For the measurements of the pressure and the skin friction, as well as for the
calculations, the flow around the cylinder was regarded to be of the steady type.
This is not true exactly because of the alternating eddy separation, which causes
an oscillation of the boundary layer up to the stagnation point. Therefore the
measured local values are temporal mean values. The averaging process is
unknown, because it depends on the co-operation of the probe, mains and
measuring instruments. However, an eventual systematic error is so small that
there is no discrepancy between the mean pressure distribution and the corre-
sponding mean skin friction distribution. The application of the skin friction
probe leads to the question, whether the edge, which can be regarded as a
discrete roughness element, influences the flow. According to the measurements
of Page & Preston (1941) the transition from laminar to turbulent boundary
layer in a flow without pressure gradient
- occurs at a critical dimensionless
roughness height y* = 20;
y* = (7ih4+hP
7
I n our own atmospherical experiments this value y* was always lower than 4.5.
Concerning the investigations in the HD channel, the largest value y* = 40 was
reached a t the highest Reynolds number in the position of the maximum skin
friction; this means that a premature transition may be caused by the probe.
This fact seemed to be confirmed when applying Kraemer's (1961) criterion :

uhp < 900,

Recrit = __
7
because the obtained maximum critical Reynolds number was Recrlt = 1800.
I n order to get a decision from the experiment, we measured the pressure
distribution for control by means of an additional pressure hole, drilled at a
distance of 30 mm from the skin friction probe on the same generator. Within
the accuracy no deviation of the pressure distribution could be observed. From
this result it was concluded that the probe had not yet influenced the flow. It
is suggested that the reason for this discrepancy is the intensive pressure
gradient.
The pressure and the skin friction were measured around the circumference
of the cylinder from g5 = 0" to g5 = 360". Thus it is noticed that sometimes

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 638 E . Achenbach
there is an unsymmetrical flow. The asymmetry has been observed independent
of different length to diameter ratios. The flow around the cylinder is sensitive
in such a way that the smallest disturbances of the flow a t the entrance of the
test section are suggested to cause the asymmetry.
At high Reynolds numbers the boundary-layer changes from the laminar to
the turbulent state. The skin friction probe was calibrated in the laminar
boundary layer only. Therefore we cannot be sure that the values of the turbu-
lent boundary layer are reliable. I n this sense the investigations of Bradshaw &
Gregory (1959) are interesting. The authors analyse different methods of
measuring local skin friction. One of the methods used by them in their experi-
ments is that of the Stanton tube, which, concerning the principle, comes
closest t o our procedure. They demonstrate, for a flow without, respectively
with very small pressure gradient, that a Stanton tube calibrated in a laminar
boundary layer and working in a turbulent flow yields values of wall shear
stresses which are too low. From these results it may be concluded that our
turbulent skin friction may be too small. Unfortunately the results of Bradshaw
& Gregory cannot be evaluated for our case, because the influence of the
considerable pressure gradient is unknown.

The author is indebted to Dr H. Grosse, director of the ‘Institut fur Reaktor-

bauelemente der Kernforschungsanlage Julich ’ who made this investigation
possible. He is also grateful to all people whose co-operation he was dependent
on, especially to Mr H. Gillessen, Mr F. Hoffmanns, Mr H. Reger, Mr W.
Schmidt and Mr G. Turk for their assistance in preparing and performing the
experiments.

REFERENCES
BRADSHAW, P. & GREGORY, N. 1959 The determination of local turbulent skin friction
from observationsin the viscous sub-layer. Aero. Res. Counc. Lond. R. & M . no. 3202.
FACE, A. & FALKNER, V. M. 1931 Further experiments on the flow around a circular
cylinder. Aero. Res. Counc. Lond. R. & M . no. 1369.
FAGE,A. & PRESTON, J. H. 1941. On transition from laminar to turbulent flow in the
boundary layer. Proc. Roy. SOC. A 178, 201-27.
FACE, A. & WARSAP,J. H. 1930 The effect of turbulence and surface roughness on the
drag of a circular cylinder. Aero. Res. Counc. Lond. R. & M . no. 1283.
GIEDT,W. H. 1951 Effects of turbulence level of incident airstream on local heat transfer
and skin friction on a cylinder. J . Aero. Sci. 8, 725-30.
GROSSE, H. & SCHOLZ, F. 1965 Der Hochdruck-Gaskanal. Kerntechnik, Heft 4. Munchen:
Thiemig.
KONSTANTINOV, N. I. & DRAONYSH, G. L. 1955 The measurement of friction stress on a
surface. D.S.I.R. RTS 1499 (1960).
KRAEMER, K. 1961 Uber die Wirkung von Stolperdriihtenauf den Grenzschichtumschlag.
Z . Plugwisa, 9, 20-7.
MORSBACH,M. 1967 n e r die Bedingungen fiir eine WirbelstraBenbildunghinter Kreis-
zylindern. Dissertation T. H. Aachen (1967).
RECRENBERG, I. 1962 Zur Messung der turbulenten Wandreibung mit dem Prestonrohr.
Jahrbuch der Wissenschaftlichen Ctesellschaft fiir Luft- und Raum-fahrt e.V. (WGLR),
151-9.

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120
 Measurements on jow past a circular cylinder 639
RECHENBERG, I. 1963 Messung der turbulenten Wandschubspannung. 2. Flugwiss, 11,
429-38.
ROSHKO,A. 1961 Experiments of the flow past a circular cylinder a t very high Reynolds
number. J. Fluid Mech. 10, 345-56.
SCHLLLER,L. & L I ~ W. , 1933 Druck- und Reibungswiderstand eines Zylinders bei
Reynolds’ schen Zahlen 5000 bis 40000.2. Plugtech. Motorluftschiff. 24, Heft 7 , 193-8.
SCHLICHTING,H. 1955 Boundary layer theory. Oxford : Pergamon.
THOM,A. 1929 An investigation of fluid flow in two dimensions. Aero. Res. Counc. Lond.
R. & M . no. 1194.
WIESELSBEROER, C. 1923 Versuche uber den Widerstand gerundeter und kantiger
Korper. Ergebnisse A V A Gottingen 11 Lieferzcng (1923).

Downloaded from https:/www.cambridge.org/core. Université Grenobles Alpes, on 27 Jan 2017 at 16:16:30, subject to the Cambridge Core terms of use, available at
https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0022112068002120