L.
Hill
ABSTRACT
In this paper a summary is given of the published model experiments on the
wind pressure on cylinders and of some previously unpublished experiments made
at the National Bureau of Standards on models, on a large cylinder in natural
This information is
winds, and on the power plant chimney in natural winds.
collected and analyzed for the purpose of estimating the wind pressure on chimneys and other cylindrical structures at known wind speeds.
The new experiments are in large measure determinations of the local distribution of pressure over the entire surface of the cylinder in the model experiments,
and over a suitably chosen section in the experiments in natural winds. In addition, the overturning moment on a 10 by 30 foot cjdinder in the natural wind was
directly
measured.
CONTENTS
Page
I.
II.
III.
Introduction
Summary
General procedure
3.
Results
Discussion and comparison with other published data
4.
IV.
of
679
feet high
(a)
(6)
(c)
3.
653
655
660
660
661
664
666
677
677
677
Power-plant chimney
V. Resume"
I.
679
683
686
689
692
INTRODUCTION
Bureau
654
[Vol. s
ments described in the present paper were undertaken for the purpose
of obtaining information as to the distribution of wind pressure over
chimneys and other structures of substantially cylindrical form.
Because of its great complexity, we have suggested that the windpressure problem be subdivided into several parts, namely, (1) the
determination of the wind pressure in a uniform and steady wind,
(2) the determination of the frequency of occurrence of winds of
specified mean speeds as given by the Weather Bureau records, (3) the
determination of the maximum gust speeds in winds of specified mean
speeds, and (4) the determination of the stresses produced in the
The division of
several structural members by specified loadings.
the problem into wind loads and wind stresses is well recognized, but
the separation of the commonly adopted wind loads into the several
components is not often made. Too often a figure of 20, 25, or 30
is first adopted and then some method of computation is
found which will result in the value of the pressure originally adopted
and thus in a measure justify its adoption. The loads are often said
to correspond to a speed of 100 miles per hour, a figure which seems
to be regarded as a limiting value irrespective of exposure.
The problem of choosing a wind speed to be used for design purposes is essentially a problem of probability, and unfortunately there
are not sufncient data available to make an exact mathematical
study possible. The period of observation for which records are available is not sufficiently extensive to give reliable information regarding the frequency of very high winds. The only records available
in sufficient quantity to be useful are those of the Weather Bureau
on the mean speeds over 5-minute periods. In many of the records,
especially for stations in cities, the exposure and elevation of the
anemometer have been occasionally changed, and the records are not
comparable for the entire period of years. If we examine the records
for a period of years at some station for which the elevation was not
changed (for example, for Washington, D. C, for the years 1911
lbs. /ft.
somewhat as follows.
The mean speed
exceeds 40 miles per hour about 1 day in 85, 45 miles per hour about
1 day in 195, 50 miles per hour about 1 day in 850, and 55 miles per
hour about 1 day in 1,950. The maximum observed over a period of
58 years was 68 miles per hour, so that the mean speed exceeds 70
miles per hour, say 1 day in 25,000.
The engineer must decide how
we
1927),
find relations
much
These
design.
The data
While approximately
correct,
for
use in
Dry den'
Wind
Hill
Table
1.
True speed
(miles
per hour)
Indicated
speed, old
4-cup
Indicated
speed, new
3-cup
standard
standard
5
11
5
10
15
10
15
20
25
35
40
45
50
55
Before Jan.
1,
1928,
theU.
S.
True speed
(miles
per hour)
23
30
37
20
25
31
00
65
70
75
80
85
44
50
57
64
71
36
41
47
52
57
90
95
100
105
110
17
30
655
Pressure on Chimneys
Indicated
speed, old
4-cup
Indicated
speed, new
3-cup
standard
standard
78
85
91
105
112
63
68
73
79
84
89
118
125
132
138
145
95
100
105
111
116
98
"Weather Bureau used the 4-cup instrument; after that date, the 3-cup.
We
New
whom
We
!II.
Bureau
656
[Vol.
In the
first
place -
is
in
many
cases
speed and the size of model, so that from a single value of it for any
given shape of body the pressure at the corresponding point on a
similar body of any size at any wind speed can be readily computed
with the aid of a table of velocity pressures. Inmany other cases
the variation of
TTT
"-.
pressures are
all
Table
is
small.
The
Velocity
pressure
speed
(miles per
(lbs./ft-
hour)
True wind
speed
(miles per
hour)
Velocity
pressure
(lbs./ft.2)
70
75
80
10.81
12.53
14.39
16.37
40
1.600
2.302
3.133
4.092
85
90
95
100
18.48
20.72
23.08
25. 58
45
50
55
60
5.179
6.394
7.737
9.208
105
28.20
30.95
33.83
36.83
15
20
25
30
35
is
wind speeds
0.064
.256
.575
1.023
10
Note.Velocity "pressure
units.
2.
True wind
the density
W
a pure number independent of the units used so long as the
2
ratio is
65
no
115
120
that at 15 C.
The term static pressure is u?cd to indicate tho pressure which would be measured by a pressure gauge
moving with the air and, therefore, "static" with respect to the air. In actual practice the measurement
the wind direction.
Is made by means of holes in the side of a closed tube, the axis of which is parallel to
The form of the tube is such that the air flows smoothly past the holes.
2
B. S.
Excluding
Sci.
Paper No.
importance only
in special cases.
Drvntn
Wind
Hill
657
Pressure on Chimneys
by
we have =
I- dA
q,
sin
ff
6.
found convenient to
It is
Again
^ = fp ds sin 0,
where ds dl
is
the area
dF
it is
6
7
s
9
for
The terms, resistance coefficient, drag coefficient, shape coefficient, are also used.
Alcune esperienze di idrodinamica, 2, No. 4; Rome 1912.
Comptes Rendus, 155, p. 1597; Dec. 30, 1912.
Comptes Rendus, 15G, p. 109; Jan. 13, 1913.
See for example the following summaries: Technical Report No. 185, National Advisory Committee
Aeronautics; Reports and Memoranda No. 190, Brit. Adv. Comm. Aero.; Annual Report, Brit. Adv.
Comm.
w
und
Motorluftschiffahrt,
5, p. 140; 1914.
Bureau
658
[Vol. 6
much
(low
also
smaller above the critical region than below it, and that smooth
around the sphere continued well past the major section. It was
found that the flow, and hence the force, was most sensitive to
m,
11
For a further development of tix>so Ideas,
Phil. Trans. Roy. Boo. (London). 174, Pt.
p.t )3S; iss:i.
see Technical Report 842 of the National Advisory Committee for Aeronautics.
Zeitschrift fur Flugtechnlk und Motorluftschlfitahrt, 5, p. 140; 1914.
" Reports and Memoranda, No. 191. itrit. Adv. Comm, Aero. Annual Report, 7, p. 30; 1915-16.
w Tiavaua Laboratolre Aerodynamlque Eiffel (B. Chiron, Paris), p, 60, L915-1918.
gebnlsse der Aerodynamlschen versuchsanstalt, QOttlngen, 1923 (R. Oldenbourg), 2, p. 28.
,
r Aeronautics.
chnical Report No. 231, National Advisory Committee
Reports and Memoranda No. L02, Brit. Adv. Comm. Aero. Annual Report,
f(
5, p. 47; 1918-14.
Drijdni
Hill
Wind
[Table
3.
Effect
059
Pressure on Chimneys
Length-diameter ratio
Force coefficient
1
0.
03
0.
69
0.
3
75
20
0.92
10
5
0.
74
0.
83
40
00
1.
I.
20
3.0
\
\\
\\
\\
2.0
C^>__v
lIj
^SL
~^ N
'
1.0
\
\
\
4.0
3.0
a.o
1.0
5.0
6.0
Lo 9 I0 VL/V
.
Figure
1.-
= kinematic
viscosity
=0.000161 for
(ft.
infinite length
/sec).
air.
Relf.
Wiesclsborger.
660
Bureau
[V01.5
been studied by Karman 19 and Relf. 20 For approximate computations at Reynolds numbers below the critical value the product of
frequency (cycles per second) by diameter of the cylinder (in feet)
may be taken as 0.29 times the wind speed in miles per hour. At the
critical Reynolds number the frequency increases more rapidly with
the speed, and finally the flow loses its periodic character; it continues to vary with time, but in an erratic manner.
It is obvious
that conditions of resonance in wires or flagpoles in high winds are
to be avoided, and it seems probable that the erratic fluctuation of
the flow may set up severe oscillations in chimneys of small stiffness.
Model experiments fail to give values of force coefficients applicable to full-scale chimneys, because of the occurrence of the critical
region.
Furthermore, the information available as to the effect of
the length-diameter ratio and as to the details of the pressure distribution at Reynolds numbers above the critical value is quite limited.
Our first experiments were intended to supply further information
on the pressure distribution above the critical value. They were
carried out in the 10-foot wind tunnel of the Bureau of Standards
and are described in the next section.
III.
APPARATUS
was approximately
w Physikalisohe Zeitschrift,
13, p. 49;
H)12.
J21.
Reports
and Memoranda
Research
Dryden
Wind
Hill
661
Pressure on Chimneys
over the greater part of the length were less than 0.015 inch. The
top of the cylinder was closed by a plate % 6 inch thick, The total
length from the bottom of the
Sect.
mounting to the top of the top
I
mounting plates.
The wind stream was supplied
by the 10-foot wind tunnel which
is
described in Scientific Paper
No. 523. The method of measuring wind speed and the pressure
gauge used for measuring pressures
at the station on the cylinders were
In
described in the same paper.
the present tests speeds up to about
66 miles per hour were obtained.
The effect of the ground on winch
the chimney stands was simulated
by a platform placed under the
..
ii
I
3
,.
..
,.
it
..
10
>.
12.
its
the platform.
2.
m
The
GENERAL PROCEDURE
,-,
Bureau
662
[Vol. 5
small hole in a static plate located 7% feet upstream from the model or,
on some occasions, to a small hole in the downstream quadrant of a
short cy Under placed with axis normal to the wind. 21 When the
wind speed was steady, an observer read the 12 tubes of the multiple
gauge, which gave the differences in pressure between the holes at the
surface of the model and the reference pressure.
Readings were
taken for speeds of approximately 40, 60, 80, and 97 feet per second.
Figure
3.
1, 2, 8,
sections.
The
w This was dune in order to keep the differentia] pressure within the range of the multiple manometer.
Dryden]
Hill
Wind
Pressure on Chimneys
663
Figure
4.
7,
and 8
For explanation, see legend
of Figure 3
Bureau
6G4
[Vol.
RESULTS
The pressure distribution for the four air speeds are given in part in
The distributions of
the polar diagrams of Figures 3 to 8, inclusive.
average force per unit projected area along the length are given in
Figures 9 and 10 and the average force per unit projected area for the
whole cylinder in Figure 11. The quantities plotted are absolul
coefficients, the values of which do not depend on the system
units, provided a self-consistent system is used throughout.
<
Figure
5.
and
at sections 9,
10,
'
Far explanation,
The
8-mch cylinder
Distribution of pressure on
see Figure 8
(V a at any section
CD
for the
whole cylinder
& Bin
"
si
ds
\\
|
sin o da dl
force
irhj
J I,
Drydcn]
Wind
Hill
where
665
Pressure on Chimneys
w = width
I
dsdl
=
=
= the
p p = the
line.
/!****"'
'
1
'v
^XN
^
\
Figure
6.
and
For explanation,
/ /
/ /
at
/,
Bureau
666
[Vol.5
It will
be seen that
wind speed,
Figure
7.
7,
and 8
For explanat ion, see legend
4.
DISCUSSION
of
Figure 3
All of the speeds used for the 8-inch cylinder are equal to or
greater than the critical value. Because of this fact the computation
For, owing to
of average pressures is subject to a small error.
chances in the barometric pressure and the temperature of the air,
the kinematic viscosity and, hence, the Reynolds number for a given
The total range in values of the
air speed changes from day to day.
Dry den]
/Jill
Wind
Pressure on Chimneys
607
Reynolds number at a given speed was about 3 per cent. The effect
on the value of the average pressure is probably not very great.
A careful examination of the diagram shows a lack of symmetry
with respect to the wind direction. By reading off the angles at
which p p s vanishes we may estimate the nature and amount of
the asymmetry. Figure 12 shows the average angle of asymmetry or
angle through which the diagram for each section must be rotated to
bring the points of zero pressure difference to symmetrical positions.
~H\/\
1
y(Z "V\\ \
A[J/
^^y\
~
Figure
8.
/I
./\.
/
/
For explanation,
and 12
see legend of Figure 3
Bureau
[Vol. 5
'""
i-
<N
>(j
a -
-o'
t.
-." ?"
J,
~o-
-_
Ifl
o
o
-*-
^, -*
_
'"
-- --
~"
hi
<r>
|
5
-5
>
t-
>
i(
Js
\N
Vs
^?
^==
.3
.2.
10
30
2.0
40
50
Figure
9.
section
60
80
ft./sec.
ft./sec.
97 ft./sec.
-0
\Jr
s
Y \
hi
\s
-.-- -
^;
-o-
.3
>N
*7
"
--'
V *^
s.
"*
"B.<x*
--
r "
rfl
=*
-
--
"
a.
>
_ a
30
10
40
Fiqure
10.
60 ft./sec.
80 ft./sec.
97 ft./sec.
Dry den'
Hill
Wind
1.0
Pressure on Chimneys
669
\
\
\\
0.8
\
v
0.6
>
\
0.4
vv
*.-*
""-V-
~f
.*--
0.2.
1*10'
1*10'
3x10'
4x10'
5xi0"
6xl0
7x!0
VL/v
Figure
11.
Cd = force
V= wind
=diameter
(ft.), v
at large
Reynolds numbers
whole cylinder).
= kinematic
viscosity
(ft. 2 /sec.)
= 0.000161
for air.
Eiffel.
Wieselberger.
and Hill, 8-inch cylinder.
Dryden
Dryden and
o
x
'
<
u
z
40
30
Distance of Section from Top ~ Inches
2.0
Figure
12.
^.n^Ze through which the pressure diagrams for the various
sections must be rotated in order to give symmetrical positions for the
points of zero pressure difference
670
of Research
[Vol. 5
of images. In other words, the tests actually made were on a cylinder in the presence of other cylinders placed in positions corresponding to the images of the first cylinder in the walls of the fictitious
small tunnel. The smallest projected area was about 1.8 per cent
and the next was about 35 per cent of the area of the tunnel cross
section, so that it is not possible to estimate very closely where the
effect of the tunnel walls begins to exceed a predetermined small value.
The values of the ratio which we have employed are in line with values
used in other laboratories. At Gottingen the projected area was
about 1 1 per cent and at the Eiffel Laboratory about 9 per cent of the
area of the tunnel cross section. The ratio of cylinder diameter to
tunnel diameter was 0.105 for the 12.6-inch cylinder, 0.067 for the
8-inch cylinder, 0.134 at Gottingen, and 0.100 at the Eiffel Laboratory. Taylor states that a ratio of 0.125 is evidently too large.
recognize, therefore, that the conditions of the experiment do
not correspond to ideal conditions as fully as one might desire, especially as regards the ratio of the projected area of the cylinder to
the area of the tunnel cross section. The conditions do, however,
represent the nearest approach to ideal conditions in this respect
that can be obtained at large Reynolds numbers in atmospheric
wind tunnels of the sizes at present available at the Bureau of
We
Standards.
The observed type of distribution of pressure is well known from
previous measurements, differences being matters of detail as to relaIn every case except for sections 1 to 4,
tive numerical values.
which are in the uppermost 5 per cent of the length, the maximum
pressure increase on the front is equal to the velocity pressure within
2 per cent, hence within the experimental error. The point of maximum pressure is within 1 of the point farthest upstream and the
variations within this limit which have been previously discussed are
at least partly due to a slight spiral motion in the air stream. The
pressure falls off to either side and reaches the static pressure at a
zonal angle of from 31 to 35. The pressure then becomes less than
the static pressure by increasing amounts up to an angle in the
neighborhood of 65 to 75. At low Reynolds numbers, the pressure
decrease drops slightly from its maximum value near 70 to a nearly
constant value extending from 85 or 90 to 180. At high Reynolds
numbers, the pressure decrease varies slightly between 70 and 100
or 100 and then drops rather sharply to a nearly constant value
extending from about 135 to 180.
The chief differences observed over the range of Reynolds number
from 168,000 to 642,000 (most of the change occurring between
252,000 and 257,000) may be summarized as follows:
Reynolds No.
KIS.OOO
'l.
:'..
Maximum
Mean
642,000
35
31
1.2
2.4
.7
.4
den
niu
671
These changes, especially the last two, are well shown in Figures
and 5. The excessive development of the winglike protuberances extending across the direction of the air stream are characteristic of Reynolds numbers above the critical value.
Table 4 gives all of the published information which we have been
able to find on the general characteristics of the pressure distribution
around circular cylinders. As it seems probable that a cylinder
spanning an open ah- stream does not act as a cylinder of infinite
length (owing to the feeding of air into the low pressure regions at
its ends), the results of Hemke (serials 15 and 16) and of Toussaint
and Weisenburger (serial 17) have been considered applicable to
cylinders having lengths equal to the diameter of the air stream, and
3, 4,
672
Bureau
* i
co
P.
P.
.rH 03
[Vol.
9rfa
ood
OS OS OS OS Cs
OS OS OS I- CO co
OS
ii| ill
,_,
f^illllil
"22
3 go
8 3 8
s a
aj
23
nlT
-o
In
,2 >>
S.t?
B(S S*s
_< IC ^h >0
rH
SSSS8
HHlOrtlO
rH
88888 8
O
8
lO
IO
a
i
--'loggs
3
k
o
J^ifi
oO
O
o
o5
o)
ooooo<
j3xi
"tf<
<N CM CM
O CM CM
SSooSS
i-H
^,2
.
tsc-o
hi
C*
ss
a o
!
flonacici
a d -2.2b
o b
o
h
<
as a>
(u
o n
.
ro
<i>
n pq
?;
(-i
^;
>-;
ill
Si
ffi
s-i
< pq "1 -^ Pw
*"
x>
oo
oo
r-
^ CO CO CO CO
CO
CO CO CO
t~ >C OP 00 50
rH
tN OS 00 <N t^ cq
CO (N W CO OS CM
O
**
Tf
t^
o w
o<i
03
C.I
GO
03
R?3
.an
03
h+ -'
pi.t.'
*<
00
CO "0
Pi 03
III!
g>
c3
K Ph OQ c_
.
3s
o 3O
o
x:
h/1
a!?
o
^f^^
CO
WCOTf
^
O
888888
!l1
OS
ift
CS
iJ5
-<*<
OS
HMM'tiOtOt^OOOlOH
CC to CO CO CO
(NCMOJCNCM
SS2
?>>>>>>!>>>>
p X3 .o x: Xi
si
7-
JS
3 O
"IO
03
So
io >o o
g
f: CM CM tN
io io
o CN
rH
rH
CO
CM
(M
CO
CS
>
CO COCO
<NCMCM
0000
>> r*.^
T "T
>>
>>
Xi
hQ
,>
X2X5X!
rH
r-J
OOXXhOO
o
CM
!>>
bbfc
CNCN
OSOrHCMCO'**
'
Dry den
Hill
Wind
Pressure on Chimneys
671
42
8fe
3 P.
4J
03
id
w>
,Q
^ o
8
d
la
31
O^CN
o
C3
3-
3
> O
"3-3
ass!
a <u
d a
3
!*
8
CO
'O
O.S
si
OOaOCNCS
-* tT
-^i
<tf<
CO
CO CO CO
aw a
<j
I- o
Co
OT
o U *-
HHi
So 3
P-T3 E?
w
I
CI Cj
iO
CO 00 00 00
- CI
Iffl
CI CI CI <N
l> tg
C3 a)
J,r
3 I
I i
L^a
1 8|
.9
1^ t
co
CO
o
lO
CM
lO >0 >C
d
<N CI <M
,-,,-,00
l-O
,_,
OiOlOlO
00 CO 00 00
'-C r-l r-l
^H
>0 CI
CN
OO^hiOQ-*
rtOCOXOO
OlCH
hWh
OO
r-H
.S
0(00 OOOO
HHrtH
OOOO
I
.-2
Oil"
to
3^
CD
3 -?>
CO
I~-
co
id
co
Coo
OOOO OOOO
o
- O
\Z
3S'~
3"to"
CI SO CI
rt
CO
11a3
fc
su;sj
C.3
.s
lOiOHOO'
Sco'C
35 " 3
&./-- <u
c^5 -
"j
+s'o
CI CI
,-<
CO CO CO CO
hONMOO
CO CO IO 25 00 CO
ma
00 00
"
cu
S C-
** t 1 - CO
OCOMC1
*>*-<
-<r
O ri CI ro
*-r
0000
8383
O
'.
00
'
"O lO iC >0
oodS
CO
CScsi-CO
>
*-i
MONh
O
CO
O -< CI-*
CI CI CI C1C) tc
c c c
t*
ifl
to
ococo
rH CO t
U3COO
iC o >C "0
IO
-*< "9<
'-C
oo
'.
t>< *<
-<*l
O
I
-}<
-*
o io ooo
CN-3"OCO
* "0 1- oo
HMWOo
CO -* CT> Ol
- CI oo
co o a> ^h
co a - eo
*owo)
ic *o o o
io o *0
& * *r >n
Cf 'C
CI
O CI CI
o CO
"i 00
i~-"
-<
co
CI
-*<
-f
"<
'
"j i'j -
CI CO**
OOOOO
oo9
OOOO
OJ
CO CO CO
'C US
O 0> OS 0>
ifl
U3
tfO
*r s>
fe
C
3
a s
<u
/'
r/
"-
'
<M CO "0
esj
-*<
# co -r m'
hOCIO
i-l
a^Ta
aS
S
ce
>* -J*
OO OOOO OQOQ
88 SSS8 S8S3
'OOO
:888
;338
o> 5v 35 *
!gg<
cococo
-'
go o> i
CI CI i-i 00
i
r-H
3 "5;
S-5
SS
13
O co O CO
o^
erf
oJ os
ss*a
S
t
COCO
>>>>
'
.Q.O.
88
00 00 00 00
CO 00 00 00 SO 00
00 00 00 00
00 CI CI CI CI CI
O N
^
i-h
COOt-
I-
h OlO-N
r<!iri:i
00
OQO
8 88!
CO
000000
CO-fi
ci
CO
S S
C-3^3
gi|Og
CO CO CO CO
CI CI CI CI
o i ci co
" C-.i"
674
of Research
[Vol.6
50
</>
-jail
CC
o
u
lj
_j
o
l.f
30
'
%*
K^X Ar+
o
z
.0
3.0
Z.0
1.0
4.0
6.0
5.0
Lo 9 , VL/v
Figure
13.
on an
p= pressure on
p= static
F=wind
cylinder.
pressure.
speed
=diameter
(ft./sec).
(ft.).
y=kinematic viscosity
(ft.2/sec.)
=0.000161 for
air.
4.
ratio.
Dry den
Wind
Hill
Pressure on Chimneys
675
cylinders. Below the critical region, Fage finds 1.02 for the cylinder
of infinite length as compared to 0.44 for a cylinder of length-diameter
ratio of 9.
There are, however, certain discrepancies, as, for example (serials
13 and 18), between the results of Taylor and Fage at Reynolds
3.0
o
o o
^o
-1.0
L0
1.0
3.0
6.0
5.0
4.0
Loa. l0 VL/v
Figure
14.
Effect of
p,= static
pressure.
pressure.
q= velocity
V=wind
for air.
4.
diameter
ratios.
-2..0
c 1-
-to
~o
V?
"S""o
"**&
"0
*-
1.0
1.0
3.0
4.0
6.0
5.0
Lo$ 10 VL|v
Figure
15.
Effect of
an
infinite cylinder
p,=mean
V=wind
speed
L= diameter
(ft ./sec).
(ft.).
i>=kinematic viscosity
(ft. Vsec)
=0.000161 for
air.
4.
effect of
676
Bureau
[Vol.5
Table
5.
Height
of section
above
plat-
Reynolds
number
of
168,000, ex-
expressed as
fraction of
average Cd
for cylinder
Reynolds
numbers
335,000 to
642,000, ex-
pressed as
fraction of
average Cd
cylinder
99
95 -.
90..80
0.96
1.12
1.11
1.07
1.14
1.33
1.27
1.04
70
60
50
1.01
.94
.94
.96
.84
.84
.84
.91
.98
.98
.98
.98
1.00
1.04
1.04
1.00
40. ..
30
20
10--.
for
Figure 11, showing the coefficients for the entire cylinder, indicates
a difference in turbulence in the several wind tunnels.
Taylor's
pressure distribution curves indicate a drop at least as early as do
Eiffel's.
In the Gottingen results and in ours the force and pressure distribution, respectively, are still changing at the highest Reynolds
number. The Reynolds number corresponding to a chimney 20 feet
in diameter in a wind of 100 miles per hour is 9,250,000.
Extrapolation to such a value can not safely be made from model experiments
limited to 642,000.
den
fm
Wind
5.
Pressure on Chimneys
677
We
678
Bureau
[Vol.5
There
50
^^_
t
p
U
-I
30
z
10
LENGTH
Figure
16.
Angle
diameter ratios
is
at
12.
16
2L0
DIAM.
which the pressure on cylinders of various lengthto the static pressure: Reynolds numbers below
equal
the critical
p = pressure on cylinder.
p,=static pressure.
is only 4 and the total change between a length-diameter
ratio of 9 and infinity is only 4, it appears that there is some hope of
fixing the ang]e within 2.
This method of attack has the advantage of securing a value of the
speed of the wind striking the chimney at the level where the measurements are made. It also happens, as will be explained later, that the
force coefficient can be obtained from a pressure distribution curve
plotted to an arbitrary scale and with an arbitrary reference pressure.
The method has the disadvantage that the results of the measurements on the large structure do not stand alone but are interpreted
on the basis of a particular kind of extrapolation of data for models, an
extrapolation which is, however, far safer than a direct extrapolation
of force measurements.
wish again to point out that this method of treatment of the
wind-pressure observations implies the division of the wind-pressure
region
We
B. S. Journal of Research,
RP22I
Figure
17.
The 10
30
en
fna
Wind
Pressure on Chimneys
G79
(a)
10
10
FEET IN DIAMETER
As the Bureau of Standards was to erect a new power plant chimney on which a more or less permanent installation could be made, it
was considered advisable to try the new method on a somewhat
smaller scale and under conditions such that the natural wind measurements could be made to stand alone. For this purpose an experimental stack 10 feet in diameter and 30 feet high was erected on the
roof of the West Building of the Bureau of Standards.
(Fig. 17.)
The stack was constructed of a comparatively light timber framework
consisting of rings, separated by compression members, to which a
The structure was located over a
sheet metal covering was nailed.
skylight opening, and guy wires were carried inside the structure
At an elevation of about 20 feet
to the roof beams of the building.
above the roof, 24 pressure holes were spaced equally around the
circumference. From each hole a separate pipe (three-eighths-inch
diameter) was led down the interior of the stack to the observation
room in the attic of the building. As first installed, each pressure
hole was formed by extending the pipe itself through the sheet metal
by a distance of about one-sixteenth-inch and holding the pipe
It was found that the presence of these
in position by lock washers.
small projections produced large suction effects and consequent
Wind-tunnel tests showed that the size of the opening
errors.
Depression of the hole
(within the limits used) had no large effect.
below the surface gave only small effects, but any projection gave
large errors.
The pressure holes were then reconstructed by soldering
a tube on the inside of the sheet metal covering and drilling from the
All burrs
outside into the tube with a one-thirty second-inch drill.
around the hole were carefully removed.
The 24 pipes were brought together at the observing station with
a valve in each line so arranged that all 24 valves could be opened or
The valves
closed simultaneously by a common operating lever.
were of the plug type so that no volume changes occurred on closing
them. Each line after passing through its valve led to one tube of
a multiple manometer similar to the one which was used for the windtunnel tests and which was described in Scientific Paper No. 523.
680
Bureau
[Vo.
The other side of the manometer was left open to the air in the observation room. An installation similar to the one used for the experimental stack is shown in Figure 24.
About 10 feet above the top of the stack on the west side a Pi totstatic tube was mounted on a wind vane.
A sketch of the tube is
shown in Figure 18. The
pressures from the Pitot and
static tubes were led through
oil seals to pipes running
down to the observation
room. The differences between the Pitot and static
pressures and the pressure in
the observation room were
measured on the manometer
used to determine the pressure distribution, replacing
two stations on the lee side
of the stack where the pressure is nearly the same for a
number of stations.
The experimental procedure was as follows: On a
day when the wind reached
speeds greater than 25 miles
per hour, the valves were
opened and the observer,
watching the fluctuations,
closed the valves when the
Dryden'
Wind
Bill
Pressure on Chimneys
681
maximum
pressure. 24
The value
Table
6.
29, 1929
No.
2,
Run
1]
Pressure
dif-
ference with
maximum
pressure as
base pressure
sure at hole to
mean
pres-
sure on the
rear.
Base
pressure
is
maximum
pressure
Lbs./ft*
Lbs./ft.*
1
2
3
4
5
6
-13.2
-12.1
-3.3
-1.7
-2.7
-4.34
-3.98
-1.08
-.56
-.89
-7.81
-7.45
-4.03
-4.36
1.83
1.75
1.07
.95
1.02
-2.0
-.5
-.66
-.16
-4.13
.97
9.3
3.05
-2.3
-.76
-4.23
.99
-2.8
-2.5
-2.8
-4.3
-.92
-.82
-.92
1 03
1 01
-1.41
-4.39
-4.29
-4.39
-4.88
-12.3
-9.1
-2.4
-4.04
-2.99
-.79
5.1
9.5
1.68
3.12
-7.51
-6.46
-4.26
-1.79
-.35
1.76
1.52
1.00
.42
.08
-.25
-1.63
-3.86
-6.36
.06
.38
tube
7,
static
8,
Pitot tube
.__..._.
9
10
-4. 55
0)
11
12
13
14
15
1.03
1. 15
0)
16
17
18
19
20
21
22
23
24
Tube
3.22
1.84
-.39
-2.89
.91
1.49
Estimated
Mean
9.8
5.6
-1.2
-8.8
maximum
pressure on rear
distribution.
Velocity pressure from Pitot, 3.21 lbs./ft.a (35.4 miles per hour).
Velocity pressure from distribution, using static from tube, -3.63 lbs./ft. 2 (37.6 miles per hour).
Angular separation of points at which the pressure equals static of tube, 69.
The curve
of Figure
around the cylinder, with the maximum pressure, p m *%, as the reference pressure and the difference between the maximum pressure
2Wx and the pressure on the rear p r as the unit of pressure. The
next step is to determine the average force per unit length exerted
24
fore,
The
475430-
-12
The
682
Bureau
upon
and
[Vol.5
wind, namely
dw
sin edi
J Pr-P
Vr
/\
/
\\
.b
s^
'
/
/
330
\
\
--
360
ANGLE
Figure
300"
/
/
170
\
\
or
19.
p= pressure
at station.
stack.
\
+
/
/
/
-
L/
\\
\ _J
.*
.1
Ficiure 20.
"^
ELEMENT loWlNO
p= pressure
at station.
where 6 is the angle of the surface element to the wind, ds is an element of the circumference and dw is an element of the projection of
the surface element on a plane normal to the wind; sin 6ds = dw, and
it is
The
20.
obvious that
Pm&x _ Pm&x
Pr
Pmax
Pm&x
P
Pr
fm
en
Wind
683
Pressure on Chimneys
average wind force per unit area at this section is obtained in terms
Vm&x~VT as unit. For the example used, a value of 0.22 is obtained.
Attention is again called to the fact that this value is arrived at, or
can be arrived at, from a plot of the pressure distribution on any
arbitrary scale and with any arbitrary reference pressure to begin
with; that is, from readings on a manometer of any slope (so long
as all tubes have the same slope) and with its reservoir connected
If, as in the experiments described, the
to any reference pressure.
velocity pressure is measured on the same manometer, we may determine at once the ratio of p m ax~Pr to the velocity pressure (1.17 in
the example) and hence obtain the true force coefficient (1.17X0.22
= 0.26) without any knowledge of the slope of the manometer. Or,
if from Table 4 we estimate the probable value of the zonal angle at
which the pressure is equal to the static pressure as 31 for a cylinder
of infinite length at Reynolds numbers above the critical region, and
if we estimate from Figure 16 that the correction to a length-diameter
ratio of 6 is 6, obtaining a probable value of 37, we may proceed as
From the estimated value of the zonal angle (37) the
follows.
angular width of the distribution curve at such an ordinate that the
pressure equals the static pressure is estimated to be 74. From
Figure 19, this angular width corresponds to
of
^max
But
at this ordinate,
pressure
a.
p=p
s,
Pr
= 0.95
Thus
Pm&x
Pr
or
^-^
\i
only.
We
postpone all uncertainties until the last step, namely, until the
determination of the ratio {p m ^~p T )lg_ or its equivalent, the determination of the zonal angle at which the pressure equals the static
pressure.
(b)
10
30
FEET HIGH
Bureau
684
When
points.
by the weight
of Standards
no wind
Journal of Research
is
of the stack.
h,
We
,
The
R$E
I
F=^FNW 2+FNE
"-"*
and
by
its
azimuthal angle,
a,
from
the meridian
a = 315 + tan- 1
^NW$
^NW5
J-
Diagram illustrating
wind force on the expcrimental stack was computed from
the
measured reactions at four
FlGTJRB
how
21.
The
force coeflicient
CD is
given by
the
Cd =
points of support
F=wind
force.
Fnr= component
of
v.
ind force
in
direction.
Fs w= component
of
wind
force
in
NE-SW
NW-SE
direction.
ft height of center of stack above the
the four supports.
2a = distance between
and
and SE supports.
Rub "Upward reaction on cylinder
support.
/?sB=reaction at SE support.
support, and
.Karat
at
NE
SW
Rnw
SW
plane of
or
NW
at
NE
where
is
stack (300
F
Ao
ft.
pressure.
The reactions were measured by
the use of pressure capsules of
the sylphon type (metal bellows).
NW support. The
Dryden
Wind
Hill
685
Pressure on Chimneys
Table
7.
[/i
Typical
= 14.33 feet,
set
of observations on overturning
a=3.98
feet, effective
area of sylphon = 97
NW
Sylphon
13.6
7.1
-fi.5
lbs./in.2
Reading
do..
do..
Net pressure
-630
pound
Reactions
Hence
(-291-107)3.98
= 1t1
111 pounds
Vne =
aaoo
14.33
,
i NW~
(572
+ 630)3.98 =
14.33
334 pounds
moment
in. 2 ]
NE
12.8
13.9
1.1
107
SE
14.9
20.8
5.9
572
sw
15.0
12.0
-3.0
-291
Bureau
686
From Table
q ro
3.63X300
6,
0.32.
q
o:
= 315 + tan"
^
111
-T
334
lbs. /ft.
[Voi.s
2
.
Hence
CD
SUMMARY, STACK
10
30
FEET HIGH
moment and
mum
'
Dryderi]
Hill
J
Wind
from
dis-
10
coco
O
*
IN
<
<*> ^J*
t>-
CO
MNtttO
CO
CO CO
<*<
HtOffl
<*!
CO CO
b-CM-*
*<
CO
CO
if
Tfl
Tr*
CO
-CMO
r-l
""i1
pressure
sylphons
687
Pressure on Chimneys
2 e -^
d
to
coef-
>
tribution
Velocity
referred
Force
from
ficient
15 14
a
o
dis-
from
iO N
CM CM CN CO
CM CO CO CO
CO
pressure
distribution
OQ CON t~CM CM CM CM
co
to
coeffi-
tribution
velocity
column
Column
referred
from
cient
force
I,
sS
X
on
pres-
Force
OCSKCO
.215 .217 .211
0.254
to
distri- bution
coefficient
S3
if t
0 CM lO
CM CM CN CM
IS S3 O
rear
referred
from
sure
to
CO
i-j.fcm
l CM CM CM
cmSjcocm
^1 OS io
CO CM <N CM
<
**
c.S2
rear
maxi-
pressure
pressure
Average
mum
referred
>-
In
on
of
to
at
distri-
bution
to
iO
tj<
IO CJ
CMCMCMr-i
CO
".1
4
mum
from
equal
NSHffl
11
88
tN
maxi-
ordinate
Width
tStiCON
II
OlrHHCD
r-00
NOSCSH
CBcONN
column
curve
Ph
a
a
Tfioco
Tf CO t- IO
CM
coacoo
,
1
3a
If
'
'
"o
D ^3
T-l
.CO
IO CM l> IQ
'
HUM
i~
--.
i~-
CMCMCNCM
Illlil'iffa
c-.
CMCOCMCN
-# O
"*
OJ CM
0O CM -#
00 00
CM
CM
-<
33
fl
-rf
fl
* T-i 1-J
ci^CMi-H
COCOCO
gsl|J
5^ Eg
Q.
*^CM
CM*
t~-"
'
COO-^t-i
CMCNCMCM
i!SS
COCOCMCM
t> ob oo
CM CM CM CM
COCOCMCM
COt^CO
CM
*<
CO -*
oo
*
t- 00 O
CM CO
CO i* 00
OC0J5 10
->
CM CM CO
ONwO)
COCMCOCN
IO 00 ON00N
^
CO CD CO
r
nNHCO
IO
i-i r-
'
-st<
"O 0
Ortr^O
00
00
CO
CNCMCMCM
<M CO CO CM
CO
>*>
CMCOCOCO
-<*<
00
ON^H
00 **
CO
"O
l>.
CM
CO CM
<
rH
1-5 i-i
T-i i-J
i-lt^CM-*l
IOH1CJO
Tf< Tf<
COCOCOCO
C CO
CO* CM'
CM
COCO^CO
r-i
lO 00
CONMU)
if i
oo' co'
cm
r>i <-
'
r-J I-i
N .-
GO OO
CO 6 O O
CMCOCOCM
CMCMCNCO
CM CM IO CO
NOOOCO
CM
CM CM
1-1
'
r-J
rH
i-H
'
'
if CO CM CO
rH
i-H
rH i-H
' 00 ' lO
CO CM CM CM
^j
ClONHO
o COCMCO
^CN
t-~'
cm'
od
r-^
-**<
t~~
m"
CM CM CO CM
oo' co'
00
-
il
00
oo' oo'
if
IO
<*<
CO CM CM
CM CM CM
MCOON
IOCM00
CO CO
O
CM
>o 00
IO
CO CM CO
CM
HrtriH
COCOCOCO
COCOCOCO
06
if CO CM CM
iCMCMCO
COCM1MCM
CO CM -* CO
COCO
l^t~
co" cm' cm' cm'
cm"
O0
io co
CMCMCNCO
'<
1*?
CM CM CM CM
E-j'3
i^i>-
f*Hla-isa^>>a5b5"
s.3^1
00 i-i
co o
CM CO
CO*
CO
i^ t^ o io
.0 Tt< l**
mo5
COCO
I
I
cm'
cm
cm' CO*
cm'
>coco
C0
a "oO3
11*1
0O1-H111-1
Tji
CO
CM'
cooon
co co
I
1-1 CM
t^t^i-inri
f-^coco
-*CNC<0S
CO -t -^
1
^c
CO CM t^ CO
>*"
CO CM CM
CM CM CM
CO CO CM CM
CO CM CO CM
r--
CM CM CM
CMoooot--
r-5
'
CM
co' CO*
CM
|3Si3S.:
"0
CM CM CM CO
10
CO 00
CM -H CO
1-H
i-i CM'
CM
CO*
CO*
OiHHH
1--
iCMCM
CNrir-ir-i
688
Bureau
CO COO
"^
CO CO CO
from
dis-
CN CO
COO
CO CO CO CO
CO CO CO CO
00 OS CO CO
CO COCO CO
tHNCO'*
Tti
CO coco
[Vol.
>o -h 10 00
CO CO CO CO
C<.h-COCN
COCO CN CO
OS HH
COO
'
pressure
sylphons
to
coef-
fl
tribution
velocity
referred
Force
from
ficient
15 14
OS CO OS CO
to
coeffi-
CO 00 10 iO
CNCOCNCN CN
CN CN CN
>* CN CO hh
CN CO CN CN
00 CO lO o>
10 t-
CN
OO)
CO
CCOS'ON COM
CNO
dis-
from
CO
pressure
-H
distribution
tribution
velocity
column
Column
referred
cient
from
force
X
O CN t 0
ta
00
ONOH
0)000
CN CN CN CN
CN CN CN CN
CD
CN
r-l
0>Si
CO 1
CN CN
rear
distri- bution
coefficient
00 OS 1 rH
r-lCOrHO os -j; o co
CNCNCNCN CNCNCNCN rHCNCNrH
on
pres-
Force
referred
from
co co cot-
HCOOH
O
CN CN CN CN
sure
coo
CN
to
HONCl)
CO
co co co tCOCO 00 CN
rear
maxi-
t- 3) t- CD
CN T-l rH CN
00 CO 00
r-H
CNCN rHCN
S2clS
OCOrHlO
CN00OO
00 iO t- CO t- CD CO rH
MrHlfllO
t- t- "3
**<
rti
iQ
rti
OHON
00
r-H
CN >C
Scn'Scn'
CO 00 CO t^
mNoe
CNCOrHCO CO
T*l
-*
*<
Tf
->tfl
CNO
J
10
COCO
H- lO
co CO
CN
O CO OS CN "OOOOO
i-i
lO
hji
r-H
1 CO 00
CO hH
COUOCOTf
mum
referred
on
to
at
to
COiCNrH
i- t- t- r-
Width
t- f- t- CO
cot- t- CO
OS 1-H OS r-H
CO t- COt~
100
t-t-coco t-
or.
CO cot-
o'o
OCN
mum
from
equal
OS CN Oi 00
cor- co co
e*
maxi-
ordinate
distri- bution
NNr-lN
*9
pressure
pressure
of
CN t- t^ t^
CNCN rHCN
'
OO
Average
t- 00 os co
(N CO CM <N
-*
73
column
curve
fl
COOCNrH
So""
CO HH t- CN
OOCNt-rH t-t-COCN OS
171
CN CO CN Tf
OS "0 t-
'
rprH
17
1
'
'
"SJ.I
qOs 3
OClOiOO
direc-
tion from
syl-
N)
Wind
direc-
tion from
distri- bution
t- t- 1-H
CN (M CO CO
phons
Wind
Hinwoi
OONON
CNCNCOCN
CN CD 00 CO
h*
OO
C; 00
*
CO 00
-r"
CNCNCNCO CNCNCNCN
IfllOOlH
CO CO CO CO
CS O
OS
O OS
CO
OS OS
<-<
OS
o>
O) CN
O CO lOCHN
CO CO CN CO
>C CN
CO CO CN
CD
OONO
ooaco OMHOJ
CNCOCNCO COCOCNCN CO
t- 00 CO 00
hJ<
r-H Tfl
OS JOgS t-
CNCOCNCN
O O Hf COCNT+-CN
OC'O HNNrO
CO CO CN CO SgScN'
"HH
r-l
CNCOCNCN CNCNCOCO
from
OffiOM
O rH 0 rnoos*o
00
of
rH
umn
umn
r-l
S2g
-1
I-l T-l
O COOS O
col-
colRatio
to
COONM
Wind speed
from
Pitot, column
0,0' co cn
tj<
rf
00 t^ OS CN
OOOJH
rH
rH
r-o
t-!
0 00 *C CO
CO CO Tt< CO
^COCNCNCN
OhNO
t- cn
CO CN IH CO
CNCOCO0
CNCO
CO
t- OS CO 00
rH
tH
rH
Tf CO 1-00
Tt<
rH t- OS CN
CN t-- CO
CN Tf t- 00
OONH
rH
m 00 "OO OOOOO
O 2S
COO
00
OOOH
HHHH rid
tCO
Tj<
I-
>*
CN t- OS
CO CO CO CO
CN OS CC CO
r- 10 coco
CD r-. t-O
*H
h.
Wind speed
distri-
bution
from
column
CCC0t->O
p. 30.0
cs
00 CO Oil-:
CO CO CO CO
CO CO CO -<h
CO CN os' CN
CO -w CO -<*i
tO-*H t-CN
>o cs >o
O-HtCTji
CO lO CO 1^
CO CO OS
t- lO OS
M.
t- CO CN 1-H
-g;CO COCO OS
jjj
2-col-
presVeloc-
ity
sure
from
Pitot, column
umn
HOOIOH
CN t- CO CO
<*<
CO CO "5 CO
H^
CO CO HH
OONH ICHOH
CO 0: CO CO
Tp ti CO CN
CO Tf lO Tt<
CO
Tfl Tt<
T*
CO CN CO CO
r-H
O CO
CO
OS 00
00 OS 00 CN
005
* -^
CNCNCOTJ<
nr CO CO CO
ocs*
NHCOrJI t-H<00
Tfi
CC CO CO
~3
a*;
r- cn co h*
-colpresVeloc-
ity
sure
distri-
bution
from
1
pres-
Static
pres-
sure
>*
column
umn
OS
Lbs
sure
to
pres-
CO
i^r-tOOO
JS
attic
r-l
pres-
sure
CO t^ CO 00
CN lO CO O
CONOO
00 t-
CO CO
co co h< co
CO
hJ<
CO
<*
ot^Tf tr-HO
rHOOO rHrH OO
1
00 rH
HrH HH
1
rH rH
1
O rH
1
00 00 <e
in
attic
referred
I-*
CO CO t- t-
COHlflO
CO OOO OS
ONhm<
CO
">
"5 CO
-3
in
referred
_g CN CN
e*
OO'O'"*
HHO-
3.2s
1
55 80
4. 5.
Lbs.jfW
4.11
prespres-
Maxi-
sure
*c3
in
attic
referred
mum
sure
to
^cn-SSg*
l- go cn rTf CO -<I< CO
"OOOM
NO0H^< HfOSf
00 CO 00 o
i**~^
^
CO CO-* CO
CO CO H^ CO
coTrcOT*.
OS 00 CN CN
t- COI- 1
CN C-J CO 'j'
8"
OS
CO
OS
^3
sure
to
a"
8"
S?a
\i
ft
I*
Il-H-O
CO
CD
00
CNO
* rH OS
CO CO CO CO
CN
h<
OS O.
1-H
>
CO
10 CO
j$f "]
Wind
Pressure on Chimneys
689
On
Bureau
11.8 feet,
690
Bureau
[Vol.6
The dark
is
h'U^H
PLANE OF
_EBSSURE HOLES
Figure
ZL
23.
Sketch of power plant chimney
showing principal dimensions
runs from each opening to the interior of the chimney, thence around
the inside of the chimney and down to the point of assembly, where
The openings
all are brought outside and down under the ladder.
are simply the squared off ends of the three-eighth inch pipes.
Precaution was taken to have no projections, the installation being
supervised by one of the authors. The area around the pipes was
pointed up. Notwithstanding these precautions the surface of the
chimney is not a smooth surface, even around the holes, but in view
of our experience with the experimental stack and with a model of
B.
S.
Figure
24.
The
kn
n'ifl
Wind
Pressure on Chimneys
69
to be very great.
Besides the ladder running down the south side of the chimney
pipes, there is another projection on the north side, a
lightning conductor leading from the lightning rods at the top of the
chimney to the ground. The stack is slightly tapered as indicated
in Figure 23, and there are certain projections in connection with the
ornamentation at the top. The chimney is not, therefore, exactly
a circular cylinder, its surface is not smooth and, in view of the
surrounding topography (fig. 22), a determination of its effective
length-diameter ratio is somewhat difficult. We may perhaps estimate the effective length-diameter ratio as roughly 15 to 20, considering the chimney as equivalent to a 100-foot stack standing on an
From Table 4 and Figure 16 we would then
infinite plane surface.
estimate the zonal angle at which the pressure equals the static
pressure as 31.
An attempt was made to secure independent checks of the wind
speed by means of a Pitot-static tube (the same used on the experimental stack and shown in fig. 18) mounted on a tower on the Northwest Building. (See fig. 22 ) The Pitot-static tube was approximately
130 feet horizontally from the chimney and some 20 feet below the
The gauge was installed in the
elevation of the pressure holes.
Northwest Building, and an electrically operated shut-off valve was
mechanism on the pressure distribution
connected to the trigger
apparatus. Repeated series of observations have shown only a
chance relation between the speed at the chimney and the speed at
the wind gauge. One speed may be zero when the other is a maximum, or the two may occasionally vary together. In the case of the
experimental stack where the wind gauge was only 20 feet from the
pressure holes, the correlation was good. At six times this distance,
there is no relation between the two speeds.
The observing station
is very similar to that described for the experimental stack.
It is
shown in Figure 24. The details of the mechanism do not show very
clearly since one sees the 24 operating levers end on.
The details
of operation and of computation were the same as described for the
experimental stack. In general, however, the results have been computed in the form of averages of groups of points. A typica] plot of
a group of observations is shown in Figure 25.
The result of some nine groups of observations gives an average
value of 0.37 0.02 for the ratio of the average force per unit area
of the projection to the pressure difference between the lee side of the
chimney, p T and the maximum pressure p m&x
Assuming the zonal
angle at which the pressure is equal to the static pressure to be 31,
- >r to the velocity pressure is found to be 1.82, and
the ratio of
hence the true shape coefficient 1.82X0.37 = 0.67, a value corresponding to a pressure of 17.1 lbs. /ft. 2 at a true wind speed of 100
miles per hour.
This value is considerably higher than the value
obtained on the short cy Under, a fact to be attributed mainly to the
greater length-diameter ratio, but probably to some extent to the
greater roughness of the actual chimney.
and the 24
692
Bureau
[Vol.5
RESUME
V.
/s
/
\\
i
i
\\
)
.
/
J
fj
\%
H
/r
/M
\ t
-A-
170
360
330
300
90
60
30
ANGLE
Figure
25.
of
150
110
180
110
140"
170'
/\
\
\
1
.
\\
i
,7
V\
\
V
\
\\
Vy
/
\\
//
V 7
YX y
^v >y
/
/
30"
60
ANCLE
Figuue
26.
or
90
110
180
test,
experimental
= pressure on chimney.
Dryden']
mil
Wind
093
Pressure on Chimneys
on the lee side and the maximum pressure. In Figure 27, the same
curves are given in terms of the velocity pressure, known in the cases
of the 12.6-inch cylinder and the 10 by 30 foot stack, but for the
power plant chimney obtained on the assumption of a zonal angle of
31 for the point at which the pressure equals the static pressure.
The distributions are of the same general nature but very considerably
different in detail.
diameter and
its
f VV
/
f
~^
'
k-
\
i
\
\
//
',
"\
Vy /
J
/
"
y/
V\
/
/
nt
ANC.LL
Figure
27.
p pressure
or
SURFACE.
ELEMENT t.WIND
test,
experimental
on chimney.
p.=static pressure, q= velocity pressure. For the power-plant chimney (curve C), p,
to be equal to the value of p at the zonal angle 31.
A, l2.G-inch cylinder in wind tunnel, distance from top=1.4 diameters.
B, 10 by 30 foot experimental stack, distance from top= 1.0 diameter.
C, power-plant chimney, distance from top=3.0 diameters.
is
assumed
Washington, March
26, 1930.