On the Resistance of Screens

K. E. G. WIEGHARDT
{Formerly Admiralty Research Laboratory, Teddington,
now at Hamburg University)

Summary: The resistance of various screens in a tube is reduced to the drag of

a single piece of wire by assuming that the average velocity in the screen itself rather
than the free stream velocity is the significant factor. Almost all the available tests
on screens or gauzes are brought into line with a single experimental function similar
to the drag of a cylinder plotted against Reynolds number.

Notation
d wire diameter
I width of the square mesh
P porosity=(area of holes)/(total area)
t*l» 4*2 velocity before and after the screen
static pressure before and after the screen
P density of fluid
v kinematic viscosity of fluid
k resistant coefficient = _
^2« l2
B2
c reduced resistance coefficient '•
i-/r
Co
drag coefficient of an infinite cylinder
1. Normal Flow Through Screens
The resistance of screens or gauzes in a tube has been measured several times
since 1932(1'7). Also some theories were established, two of which12,4) lead to
formulae deriving the resistance only from the geometrical dimensions. Considering
the complexity of the general problem of aerodynamic resistance, they are likely to
be only rule-of-thumb formulae, even if they cover the tests in a certain range, and
so a less ambitious approach might be preferable.
Attention is focused only on screens made of round wires forming a square
mesh, as these are defined geometrically alone by the wire diameter d and the
porosity
o_ area of holes / d \2 ,
P U)
~~ total area ~ V ~ 7 / • * • •
where / denotes the width of the square mesh.
Originally received March 1952; revised August 1952.
[The Aeronautical Quarter!}', Volume IV, February 1953]
186

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871
 RESISTANCE OF S C R E E N S

The resistance coefficient k is defined as the pressure drop through the screen
divided by the pitot head in the straight tube, i.e.
j, P1-P2 •
k=
-F~2 (2)

The proposal has been made'1' to reduce the resistance of the screen to that
of a single wire, i.e. an infinite cylinder. As a fraction of (1 - /?) of the whole area
is covered by pieces of wire with a drag coefficient c, the resistance of the screen
is expected to be roughly equal to the sum of the drag of all these wire pieces, or
* = (l-/8)c (3)
In the same way as c depends on the Reynolds number uYdh, k/(l - /3) should
be a function of uxd\v alone for all screens.
The experimental results, since they give various curves for various values of
/3, fail to substantiate this. The reason appears to be that the characteristic velocity
for the flow round the wires is not the one in front of the screen ux, but the higher
average velocity through the screen, ujfi. This idea has also been used by
G. I. Taylor(5).
Hence, for computing the drag it may be assumed that the single wires are in
a field of flow with free stream velocity uJ/3. This may be a bold assumption
because the actual velocity in front of the wires is u^, on the other hand, it is the
velocity (outsid'e the boundary layer) on the rear part of the cylinder which
influences the wake and therefore the drag. With this assumption the resistance per
unit area becomes

B2
so that c— -.—-ok. (5)

Now c should depend only on u,d/(Bv). That this is so is shown in Figs. 1 and 2,
where all the available tests are plotted on logarithmic scales. Although k varies
between 035 and 19, all the measurements lie along a single curve, particularly the
very accurate tests from Ref. 6 where, incidentally, an empirical formulae

=£ # (6)
is given for high Reynolds numbers.
Of this whole series from Ref. 6 only the points marked (O) for a single screen
are out of line, but in this case the mesh was rectangular with sides in the ratio 1:125,
187

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871
 K. E. G. W I E G H A R D T

40 60 60 100

Fig. 1.
The reduced resistance of gauzes, from Refs. 1, 3, 5, 6.

while in all other cases the mesh was square; the discrepancy presumably indicates
the influence of the geometry of the system. Other points which do not fit well are
the highest values of Ref. 1 (marked x). No explanation can be offered here, and
it might be useful to make some new tests at high Reynolds numbers, e.g. in a water
tunnel. In a water tunnel with a large working section it would be possible to
achieve even turbulent flow round the wires (critical Reynolds number u,d/(/3v)
about 3 x 105).
Further, it is seen from Fig. 1 that the difference between the average c and the
drag coefficient cD of a single cylinder in infinite flow is unexpectedly small, in view
of the crudeness of the whole reduction. In particular no attempt has been made to
take account of the fact that the wires composing a mesh are not straight, nor of the
influence of the points at which wires cross one another.

2. Comparison with Previous Theories

Another theoretical relationship for k has been derived in Ref. 5. Apparently
the main interest lay in the resistance of perforated sheets. Hence the flow through
such a sheet or a screen was reduced to that through a single hole as at the vena
contracta. This gives
(7)

with C=contraction coefficient and y=fraction of the pressure loss which is regained
when the stream again becomes uniform behind the sheet. Theoretically this
approach is as good as the previous one; but when considerations are restricted to
188

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871
 RESISTANCE OF SCREENS

Fig. 2.
The reduced resistance of gauzes, from Ref. 7.

screens made of round wire it seems more practical to take as basis the drag of a
cylinder than to find out how C and y depend on B and the Reynolds number.
Furthermore, with most screens, when /? is not too small, the mesh length or the
distance between the wires is several times the wire diameter. Therefore the com-
parison with the flow around a cylinder in an infinite fluid seems more natural than
that with the jet through a single hole. For screens, it was found that for high
Reynolds numbers
, 0-70
-1. (8)
8>
Assuming an average value of w^/v of 200, this relation gives the curve "T.D : Th."
in Fig. 1. (For other values of w^/v the same curve has to be moved horizontally.)
From a paper of Eckert and Pfliiger(4) a graph of k against B for uxd/'v = 200
is also plotted in Fig. 1, giving the curves "E.P : Exp." and "E.P : Th." Their
theoretical formulae is

-(¥) (9)

In Fig. 1 this curve " E.P : Th." does not fit the tests at all whereas in their
plotting the agreement is reasonable, though not very good; a striking example of
how much depends on the art of presentation.
Glauert, Hirst and Hartshorn(2) derived the formula
_ 2/
* = <*-*>(*-I) *
^B2 3,
(10)

which gives the curve "G.H.H:Th." in Fig. 1 again assuming M,d/v=200.

189

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871
 K. E. G. WIEOHARDT

u
u
l d:IN
FT/SEC
I
in water
H(2)
¥• £ 0 01. _o
K>- 0-9-
10— 1-02
_*4
08. 0-4
2_ 1-8 0-02 _
-0-001 07.
-0-6
.1-6 _0-8
.1-4
100- .1-2 J$2 0-64— I
•8"04 _ .- 1-5
100-- r l - 0 , -"
_fo "0.8 0-06 J >fe
05 2

3 — OOI
002

r-0.04
006
^~ .£L
-^
-07
crwf 04.
J
S
5

20. <>080„ 1. 0 2- 0-3 8

— 10-
_* I-10
0-2
40. - . 0 - 4 K,41
.0-6 0-2__20
60. .08
80-
100 _ - 2
.4
ISO .6

Fig. 3.
Nomogram for the pressure drop in a gauze in air flow.

Figure 1 suggests another formula. For M^/GSV) between 60 and 600 all the
tests at various Reynolds numbers and porosities can be approximated by the broken
line

c = 6-5
mr • (ID

On the other hand, the American tests(7) in Fig. 2 give lower values,

(12)

No explanation for this difference is obvious. Admittedly the test arrangements

were not quite the same. At the National Physical Laboratory the screens were put
into a tube, while at the National Bureau of Standards the screens were fixed at the
end of a square duct discharging into free air. Still, the flow through the screen
should have been the same. As there might have been a difference in the surface
roughness of the wires forming the gauzes, it would have been useful to exchange
the gauzes of the two test stations. Incidentally, the dotted line K in Fig. 2
corresponds to a silk cloth manufactured for bolting flour.
190

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871
 RESISTANCE OF SCREENS

Summarising, an average of all the tests may be assumed from Figs. 1 and 2
to be

f
or k^6(l-P)P ( ^ ) * • • • (14)

This formula has been used in the simple nomogram in Fig. 3 for calculating the
pressure drop of a gauze in air flow.
For example, if d=-fa in., /3=%, «1 = 20 ft./sec.

^v 3 5 0 ,

and the pressure drop in the gauze is

P , 1-/3 x c
Pi-Pi= ^2 "«ii xx - o£,
r"

=01x2x0-78=016 in. water.

3. Critical Reynolds Number for Eddy Production

Schubauer, Spangenberg and Klebanoff<7) also observed by a hot wire the eddy
shedding immediately after the screen and gave a graph for the critical Reynolds
number, for the beginning of eddy shedding or of the initial production of turbulence,
against /3. Obviously a straight line is a good approximation. Yet this gives a
constant critical Reynolds number if it is referred again to the velocity MJ//? in
the screen:

15
(#L-*- < >
A Reynolds number of the same order is the smallest at which a vortex street
is formed past a single cylinder in infinite fluid as is to be seen from F. Homann's
photographs of the wake(8).

ACKNOWLEDGMENT
The author acknowledges the permission of the Admiralty to publish the paper.
191

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871
 K. E. G. W I E O H A R D T

REFERENCES

1. FLACHSBART, O. (1932). The Resistance of Silk Gauzes, Round Wire—and Metal—Strips,

Grids with Square Mesh. Ergebnisse der Aerodynamischen Versuchsanstalt zu Gottingen,
IV Lieferung. Munchen, p. 112, 1932.
2. GLAUERT, H., HIRST, D. M. and HARTSHORN, A. S. (1932). Induced Flow through a Partially
Choked Pipe. R. & M. 1469, 1932.
3. COLLAR, A. R. (1939). The Effect of a Gauze on the Velocity Distribution in a Uniform
Duct. R. & M. 1867, 1939.
4. ECKERT, B. and PFLUGER, F. (1941). The Resistance Coefficient of Commercial Round Wire
Grids. Luftfahrtforschung 18, p. 142, 1941. N.A.C.A. T.M. 1003, 1942.
5. TAYLOR, G. I. and DAVIES, R. M. (1944). The Aerodynamics of Porous Sheets. R. & M.
2237, 1944.
6. SIMMONS, L. F. G. and COWDREY, C. F. (1949). Measurements of the Aerodynamic Forces
Acting on Porous Screens. With an Appendix by Sir Geoffrey Taylor. R. & M. 2276, 1949.
7. SCHUBAUER, G. B., SPANOENBERG, W. G. and KLEBANOFF, P. S. (1950). Aerodynamic
Characteristics of Damping Screens. N.A.C.A. T.N. 2001, 1950.
8. GOLDSTEIN, S. (Editor) (1938). Fluid Dynamics. Oxford University Press. Vol. II, p. 552.

192

Downloaded from https://www.cambridge.org/core. Purdue University Libraries, on 06 Jan 2018 at 15:27:26, subject to the Cambridge Core terms of use
, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0001925900000871