Vous êtes sur la page 1sur 16

ARTICLE IN PRESS

Journal of Wind Engineering


and Industrial Aerodynamics 91 (2003) 10071022

Experimental study of the wind forces on


rectangular latticed communication towers
with antennas
Ce! lio F. Carril Jr.a,*, Nicholas Isyumovb,
Reyolando M.L.R.F. Brasilc
!
Laboratorio
de Estruturas e Materiais Estruturais, Departamento de Engenharia de Estruturas e
Funda@oes,
* Escola Polit!ecnica da Universidade de Sao
* Paulo, Caixa Postal 61548,
Sao
* Paulo CEP 05424-970, Brazil
b
Boundary Layer Wind Tunnel Laboratory, The University of Western Ontario, London, Ontario,
Canada N6A 5B9
c
Departamento de Engenharia de Estruturas e Funda@oes,
* Escola Polit!ecnica da Universidade de Sao
* Paulo,
* Paulo CEP 05424-970, Brazil
Caixa Postal 61548, Sao
a

Abstract
With todays expanding communication systems, a large number of lattice towers to
support cellular antennas are being constructed in Brazil. Due to the lightweight of these
structures, wind forces are the primary concern in the design. An experimental investigation
on the subject was carried out at the Boundary Layer Wind Tunnel Laboratory, University of
Western Ontario (UWO), Canada. Three section models were designed and constructed based
on existing latticed towers built in Brazil. The wind incidence angle; the tower solidity; the
shielding effect; the inuence of the wind turbulence on the drag coefcient were analyzed.
Measurements were made of the mean and RMS drag and crosswind forces. The results were
compared with some existing codes and standards including the Canadian (NBCC, 1995),
American (ASCE 7-95, 1995), Australian/New Zealand (AS/NZS 1170.2-2002), Australian
(AS 3995-1994), British (BS8100, 1986), Eurocode 1 (European Committee for Standardization, 1995) and Brazilian (NBR 6123, 1988). It is a common approach to consider the wind
forces on antennas independent of the lattice tower, without considering the effects of their
presence on the computation of the wind forces. The question arises whether this is a good
approach or not. These effects can be described by introducing an interference factor. This
factor depends, among other things, on the tower solidity. Two models with different
solidity were tested for wind incidence angle of 0 degrees and antenna dishes simulated

*Corresponding author. Tel.: +55-11-38185705; fax: +55-11-38185181.


E-mail address: cfcarril@usp.br (C.F. Carril Jr.).
0167-6105/03/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0167-6105(03)00049-7

ARTICLE IN PRESS
1008

C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

with disks made of Styrofoam attached to the windward face. The results were compared
with ESDU.
r 2003 Elsevier Ltd. All rights reserved.
Keywords: Lattice tower; Wind tunnel tests; Microwave antennas

1. Introduction
The use of freestanding latticed steel towers to support cellular and microwave
antennas in Brazil has been intensive in the last few years with the expanding of the
telecommunication systems. Due to the lightweight of these structures, wind forces
are the primary concern in the design. Also, in Brazil there are no codes specifying
how to consider the wind loads from microwave antenna dishes on lattice towers.
This paper presents an experimental investigation on the wind action on a
designed freestanding lattice tower made of angle members based on existing towers
used for telecommunication in Brazil. This work was carried out at the Boundary
Layer Wind Tunnel Laboratory of the University of Western Ontario.
For the present work, a lattice tower was designed based on existing towers for
telecommunication in Brazil. The tower geometry is presented in Fig. 1. It is
composed by 17 sections of 5.9 m. It is 100.3 m tall and can be used for smaller
heights by just subtracting some of the lower 5.9 m sections.
On such structure, overall drag coefcients are used to calculate the wind forces.
Most of the codes present these drag coefcients as functions of the tower solidity.
The tower is separated in sections and for each section the force coefcients are
determined. The crosswind forces are negligible compared to drag forces. For square
towers most of the codes specify only drag coefcients at 0 and 45 of wind
incidence angle (the largest force coefcient).
Some important parameters, which inuence the wind loading, and which are
contained in codes of practice related to lattice towers, are examined. These are:
effect of solidity on overall forces; shielding effect; wind incidence angle; inuence of
turbulence. This is a long list of variables and it was a major challenge to select a
meaningful combination of these for this study.
Another subject of this work is the interference of antenna dishes on the wind
forces of lattice towers. It is a common approach to consider the wind forces on
antennas independent of the lattice tower, without considering the effects of their
presence on the computation of the wind forces. The question arises whether this is a
good approach or not. To describe the inuence of the antenna dishes an interference
factor is introduced. This factor depends, among other things, on the tower solidity.
This investigation does not intend to solve the problem of the interference factor
entirely due to the various parameters involved such as the position of the antenna,
type of tower, wind incidence angle, number of antennas and tower solidity. It is
intended to examine just some of these important parameters. The tests were done
varying the number of the antennas and the tower solidity. The wind incidence angle

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1009

1
2

17.7000

3
4
5
6
7
8
82.6 m

9
10
11
12
13
14
15
16
17

5.9 5.9 5.9 5.9 5.9 5.9 5.9 5.9 5,9 5.9 5.9 5.9 5.9 5.9 5.9 5.9 5.9

1.8 m

Horizontal bracing

9.5

Fig. 1. Lateral view of the tower.

and the position of the antennas were xed. The results of the interference factor
were compared with [1].
Holmes et al. [2] studied the interference factor of microwave antenna dishes
attached to lattice towers with different wind incidence angles, and found values
greater than unity for some wind directions. In their experiments Holmes et al. [2]
tested only one and two antennas at the same tower cross-section. The present study
is considering only the interference factor of antenna dishes attached to the
windward tower face with 0 wind incidence angle, In this case the interference factor
is always less than unity.

2. Wind tunnel model


2.1. Section model
A section model was designed and constructed based on the tower described. The
model was built on a scale of 1:40 and represents part of the tower (Fig. 1) at 40 m
height approximately. It is 1 m long and 0.102 m wide (Fig. 2). Two solidities were
tested. For the lowest model solidity, 0.162, the main bars were 4.6 mm thick and the
secondary bars constituted by diagonals and horizontal bars, 2 mm thick. For the

ARTICLE IN PRESS
1010

C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

Fig. 2. Model 1 (0.102  0.102  1.035 m).

Fig. 3. Model 2 (0.102  0.204  1.035 m).

highest solidity, 0.267, pieces made of styrene were added to the model bars using
double-sided tape. The tests were conducted for Reynolds number for the main bars
of 5000 and 6000 for the lowest solidity, 7000 and 11000 for the highest solidity.
Despite the fact that the actual tower was designed for angle members, the secondary
model bars were designed with square members with the same external dimensions.
The thickness of main angle members was not properly scaled.
Force balances were mounted at each end of the model. A rig was prepared to
simulate two-dimensional ow, Fig. 5. To study the shielding effects with the
distance between frames, two other models were built with the ratio between distance
and width of s=B 2 and 3 besides the rst of s=B 1 (Figs. 3 and 4). All
experiments were carried out in smooth ow (exposure 1) and turbulent ow
(exposure 2) generated by a grid placed upstream, except the tests with the models
with the highest solidity carried out only in turbulent ow. The model wind spectra

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1011

Fig. 4. Model 3 (0.102  0.304  1.035 m).

Fig. 5. View of the model, rig and grid.

ts better the wind spectra of ESDU 74031 [3] with zo 0:3 m and H 40 m, Fig. 6.
Only the highest frequency was simulated. The intensity of turbulence generated was
6.8% (Figs. 5 and 6).
2.2. Antenna model
Some disks made of Styrofoam (Fig. 7) were built at the same scale of the model to
simulate shrouded antenna dishes attached to the tower. The tests were done using 1,
2, 4, and 6 disks attached to the model by a double-sided tape in turbulent and
smooth ow. Only the wind perpendicular to one plane of the tower and to the disk
was tested.

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1012

RMS = 0.196 volts


Test Speed at H: 9.0 m/s

0.01

f S(f )/V

o Measured data
____ ESDU 74031
0.001

ESDU 74031
z o = 0.3 m
z = 50 m
Length scale 1:40
Velocity Scale 1:4

0.0001
0.0001

0.001

0.01

0.1

Wave Number (f/V)


Fig. 6. Wind spectrum generated by the grid and spectra from [3].

60 mm

22.5 mm
Fig. 7. Model of the shrouded microwave antenna dish (antenna disk). Scale 1:40.

Fig. 8. Position of the antenna disks on the model.

A small wind tunnel, with 0.5  0.5 m test section, was used for the determination
of the wind forces on the disks alone. The forces in one disk were measured
separately using a small sensitive balance. A grid was used to generate turbulence. To
minimize the disturbance caused by the load cell in the ow, the tests were conducted
with the disk connected and disconnected to the load cell, but xed to the tunnel wall
with a rod of diameter 9.5 mm. The blockage effect was not considered because the
frontal disk area represents only 1.1% of the tunnel test section area. The
interference of the load cell on the drag force was found negligible.
All tests of the antenna attached to the section model were conducted in smooth
and turbulent ow. The mean forces were taken for two different wind speeds. The
measurements were made with and without the antennas. The disks were distributed
symmetrically along the front frame of the tower (Fig. 8). For each test, the antennas
were xed in different positions: for one antenna, position 1 was used; for two

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1013

antennas, position 2 was used; for 4 antennas, positions 2 and 3 were used; and for 6
antennas, positions 2, 3 and 4 were used.

3. Results
Mean and RMS drag and crosswind forces were measured at 15 intervals for the
full 180 azimuth range. The results are presented in tables and graphs. Sign
convention used is presented in Fig. 9.
The mathematical denition of these coefcients is indicated in sequence. All have
been rendered non-dimensional using the dynamic pressure at the model height
q 1=2rV 2 ; where r=air density and V is the mean hourly velocity at the
reference height. Other factors used in these denitions are nominal cross-sectional
dimensions B; as dened in Fig. 10, H of the section model (B 0:102 m and
H 1:022 m) and tower solidity f dened as the ratio of the effective area of the
tower to the area limited by the external bars. All coefcients vary with wind
incidence angle. The models tested are specied in Table 1.
CD

mean along wind force


;
qBHf

CL

mean cross wind force


:
qBHf

3.1. Drag and crosswind forces


3.1.1. Model 1
Fig. 11 shows the results for drag and crosswind force coefcients for smooth
(exposure 1) and turbulent (exposure 2) ow.

CL

CD

Wind
B
B

Fig. 9. Model drag and crosswind force sign convention.

ARTICLE IN PRESS
1014

C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

s
B

Wind
face 1

face 2

Fig. 10. Parameters B and s of the model cross-section.

Table 1
Model specication
Model

BH (m2)

B (m)

s (m)

s=B

1
2
3
4
5

0.104
0.104
0.104
0.104
0.104

0.162
0.162
0.162
0.267
0.267

0.102
0.102
0.102
0.102
0.102

0.102
0.204
0.306
0.102
0.204

1
2
3
1
2

C D exposure 1
C L exposure 1

C Dexposure 2
C L exposure 2

Force coefficients

3
2
1
0
-1
0

15

30

45

60

75

90

Angle
Fig. 11. Model 1drag and crosswind force coefcientsexposure 1Re 6800 and exposure 2
(Re 70; 000; VD 6500; V (m/s), D (m)).

3.1.2. Models 2, 3 and 5


These models were tested only for 0 of wind incidence for a comparison of the
shielding factor related to the ratio s=B of each model. The results of the force
coefcients are presented in Tables 2 and 3.

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1015

Table 2
Models 2 and 3drag and crosswind coefcients for exposures 1 and 2
Model 2

Model 3

Exp.

CL

RMS

CD

RMS

Re

Exp.

CL

RMS

CD

RMS

Re

1
1
2
2
2

0.01
0
0
0.01
0

0.016
0.017
0.031
0.012
0.011

2.91
3.15
3.00
2.94
3.17

0.047
0.020
0.124
0.024
0.022

3976
6875
4012
5019
6414

1
1
1
2
2
2

0.04
0.04
0.05
0.04
0.01
0.05

0.020
0.016
0.016
0.042
0.043
0.049

3.04
3.04
3.05
3.19
3.18
3.11

0.025
0.029
0.022
0.127
0.127
0.126

3960
4817
6875
4009
5010
6424

Table 3
Model 5drag and crosswind coefcients for exposure 2
Exposure

CL

RMS

CD

RMS

Re

2
2
2

0.01
0.01
0.01

0.013
0.013
0.013

2.81
2.77
2.74

0.116
0.109
0.116

4031
5062
6485

Force Coefficients

3
2.5
2
CL

1.5

CD

1
0.5
0
-0.5
0

15

30

45

60

75

90

Angle
Fig. 12. Model 4drag, CD ; and crosswind, CL ; coefcients for exposure 2 (Re 10800)f 0:267:

3.1.3. Model 4
Fig. 12 shows the force coefcient results for drag and crosswind force for wind
incidence angle varying from 0 to 90 for turbulent ow.
3.1.4. Comparisons with codes
Fig. 13 shows a comparison between codes and test data. Table 5 shows the force
coefcients obtained from different codes compared to the test average. The test data

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1016

4
NBR6123, 1988

3.5

ASCE 7-95
EUROCODE 1

3
CD

BS8100, 1986
2.5

NBCC, 1995
Exposure 1

Exposure 2
AS 3995-1994

1.5
1
0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Fig. 13. Drag coefcients from codes and tests.

Table 4
Drag coefcient: experimental data
Exposure 1
Re 6000
f
0.162

0
2.86
2.80
2.86

Exposure 2
Re 6800

45
3.35
3.31

0
2.85
2.78
2.84

Re 3900


45
3.35
3.37

0
2.80
2.75

Re 6800


45
3.35

0
2.84
2.77

Re 6400


45
3.40

Exposure 1
f

Re 6000

0.277

0
2.85
2.91

Re 3900


45
3.34

0
2.90
2.98

45
3.41

Exposure 2
Re 6800

Re 3900

Re 6800

Re 10800

Re 6900

2.55
2.59

2.55
2.59

2.83

2.84

lead to drag coefcients slightly smaller than predicted by codes. Larger differences
between codes were expected as stated generally by Georgiou [4,5] about the
inconsistency of the data between codes and experiments. The data collected, within
the range of solidity studied, showed that all codes are on the safe side except [6] for
incidence angle of 45 , which is almost the same from the average test results. For
solidity less than 0.2, the Canadian Code presented the mean drag forces in the range
of 10% higher than the average of the data from other codes.
In Table 4 and Fig. 13 the test data presented are from different experiments using
different setups. For each degree the model was xed manually. The force
coefcients that have the same setup are presented at the same table line with the
same incidence angle but different Reynolds numbers (Tables 4 and 5).

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1017

Table 5
Drag coefcient from codes and test average
f 0:162

NBR6123, 1988 [12]


ASCE 7-95 [8]
EUROCODE 1, 1995 [6]
AS 3395-1994 [10]
NBCC, 1995 [7]
BS8100, 1986 [11]
Test average

f 0:277

0

45

0

45

3.09
3.15
2.98
3.07
3.34
3.09
2.84

3.58
3.53
3.34
3.47
4.15
3.55
3.36

2.59
2.67
2.51
2.57
2.73
2.63
2.57

3.01
2.99
2.94
2.97
3.41
3.17
2.84

Table 6
Shielding factor: codes and tests
Tests: Cfn =C1

Codes Kx

NBR 6123, 1988 [12]


AS/NZS 1170.2-2002 [9]
NBCC, 1995 [7]
Smooth
TEST smooth
Turbulent

s=B 1

s=B 2

s=B 3

s=B 1

s=B 2

s=B 3

0.162
0.277
0.162
0.277
0.162
0.277
0.162
0.277
0.277

0.884
0.731
0.876
0.723
0.878
0.687

0.902
0.785
0.938
0.878
0.919
0.762

0.93
0.823
0.969
0.919
0.929
0.785

1
1
1
1
1
1
1
1
1

1.01
1.031
1.033
1.013
1.022
1.045
1.067
1.047
1.09

1.024
1.053
1.05
1.042
1.027
1.058
1.071
1.09

A conclusion may be reached that the mean force coefcients have no practical
variability with turbulence and model Reynolds number. Also it is noticed that most
of the code data are obtained from tests with smooth ow, which is reasonable.
3.2. Shielding factor
To study the variation of the shielding factor with tower frame spacing, the drag
force was measured for three different models with the ratio s=B 1; 2 and 3.
Dening:
CDn 1 Kx n

;
3
CD1 1 Kx 1
where n is the ratio s=B and Kx is the shielding factor.
The test results and code data are presented in Table 6 and Fig. 14. The tests
presented higher coefcients compared to codes showing that it might have a
contribution of the lateral members to the nal load. This is not considered by the
codes of practice when determining the shielding effect. The differences in wind force
from s=B 1; 2 or 3 are so small that for practical cases it has no meaning. In most

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1018

1.09

C Dn /C D1

1.08
1.07

Test - 0.162 - exp. 1

1.06

Test - 0.277 - exp 1

1.05

Test - 0.277 - exp. 2

1.04

AS 1170.2 - 0.162
AS 1170.2 - 0.277

1.03

NBCC - 0.162

1.02

NBCC - 0.277

1.01

NBR6123 - 0.162

1
1

1.5

NBR6123 - 0.277

2.5

s/B
Fig. 14. Shielding effect. Comparison from experimental data and codes of practice for tower solidities of
0.162 and 0.267.

of the codes, the shielding factor in tower design is used only to determine the forces
that acts on each tower face and most of the towers are square or triangular with the
shielding factor of s=B 1:
3.3. Interference factor
While the shielding factor is applied to towers members, the interference factor fa
is applied to the antennatower interaction. It is dened as
fa

CDantenna on the tower


;
CDseparate antenna

where CDantenna on the tower is the drag coefcient of the antenna when it is attached to
the section model, and CDseparate antenna is the drag coefcient of a separate antenna.
Fig. 15 shows the incremental drag coefcient derived from Eq. (5).
DCD

FDtower

and antenna

qBHf

 FDtower

where FDtower and antenna is the drag force measured for the antenna disks attached to
the tower section model. FDtower is the drag force measured for the tower section
model without antenna disk.
The interference factor, obtained from Eq. (4), is presented in Table 7. Fig. 16
presents a comparison of the interference factor with an empirical expression given
by [1]
fa expkCD f2 ;

where k is 1.2 for square tower, CD is the tower drag coefcient and f is the solidity.

ARTICLE IN PRESS

CD

C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

Re=3900 - smooth
Re=5900 - smooth
Re=3900 - Turbulent
Re=5900 - Turbulent
Re=6900 - Turbulent
Re=10800 - Turbulent

= 0.162
= 0.277

1019

Antennas
Fig. 15. Incremental drag coefcient DCD: Re Dmodel V =n (Table 1).

Table 7
Interference factor CDdisk
Antennas

on the tower =CDdisk separate

f 0:162

f 0:277

Smooth ow

Turbulent ow

Turbulent ow

Re 3900

Re 5900

Re 3900

Re 5900

Re 6900

Re 10; 800

0.908
0.832
0.838
0.832

0.898
0.839
0.778
0.793

0.915
0.827
0.851
0.847

0.691
0.701
0.754
0.749

0.660
0.583
0.531
0.468

0.369
0.355
0.386
0.484

ESDU

Test data for solidity=0.162

test data for solidity=0.277

1
0.8
0.6

fa

1
2
4
6

Re Dmodel V=n

0.4
0.2
0
0

0.2

0.4

0.6

0.8

Tower solidity without antennas


Fig. 16. Interference factor from experiments and ESDU [1].

ARTICLE IN PRESS
1020

C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

Table 8
Drag coefcients on a separate antenna disk derived from the small wind tunnel test, Re Dantenna V=n
Smooth ow
Re
CD

24,192
1.13

Turbulent ow
44,405
1.12

24,342
1.02

Turbulent ow
42,542
0.9

24,342
1.03

44,405
0.86

This expression does not take into account the study of [2], which considers the
variation of the interference factor with the wind incidence angle. Eq. (7), taken from
[2], is the modied empirical form of Eq. (6).
fa expkCD f2 1 b b cos 2y  yd  90 ;

where b is an adjustable parameter; yd is the angle of the normal to the dish


antenna relative to the tower; and y is the wind incidence angle relative to the
tower.
It is noted that the Australian Standards AS 3995-1994 and AS/NZS 1070.2:2002
use Eq. (7) for interference factor with b 0:5; k 1:2 for square towers and k 1:8
for triangular towers.
Table 8 presents the drag force on a separate antenna disk taken from an
experiment in the small wind tunnel with similar conditions of turbulence, 9.7%.
Some change on the ow pattern between the antenna Reynolds number of 24,000
and 44,000 was found. The experiment was repeated conrming the results as it is
seen in Table 8 (for turbulent ow the drag coefcient changes from 1.02 to 0.9 or
1.030.86). The explanation to this could be the combination of the antenna
geometry, roughness and wind turbulence. The antenna disk tested is not a at disk,
Fig. 7. The ow must be reattaching due to turbulence and disk roughness when the
speed rises. More experiments are needed to check this ow pattern. Therefore, the
interference factor, Eq. (4), was determined using CDseparate antenna of 1.13 for smooth
ow and 1.02 for turbulent ow.
From Fig. 15, it is observed that there are two tendency lines, which depends on
tower solidity. The slopes indicate that the interference is higher for higher tower
solidity. It indicates also that, for the test conditions, the interference factor does not
depend much on the number of antennas. This may not be true for higher number of
antennas and different relative positions of the antennas on the tower. More
experiments are needed.
Depending on tower solidity, the consideration of wind forces from one antenna
disk separately is not a good approach. For 0 wind incidence angle, it seems that for
tower solidity of 0.2 or less, the consideration of the wind load on the antenna
separately is a good approach, but for higher solidity the designer should use the
interference factor taken from [1]. More research in this subject is needed because the
aerodynamic interference depends on other factors not studied here like the antenna
position on the tower and wind incidence angle. Also in this study the antenna
models are simulating only the front part of a shrouded antenna dish.

ARTICLE IN PRESS
C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

1021

4. Conclusions
There were no practical differences between the mean drag coefcients taken from
turbulent and smooth ow.
The data from the mean drag coefcients showed good agreement with those
provided by codes within the range of solidity studied. Only the Canadian code
presented disparity from other codes and from experiment for lower solidities.
Experimental results indicate some differences between measured shielding factors
and those predicted by codes. This is attributed to lateral members, which increase
actual wind forces, but which usually are not considered in a code based analysis.
However the differences between forces are very small and it is not meaningful for
practical purposes, within the range of spacing ratio s=B studied.
The shielding effect of the antenna rises with tower solidity. The authors suggest
use of interference factor of 1.0 for sections of lattice towers that has solidity of 0.2
or less. For higher solidities the authors suggest to use the curve taken from [1].
However more tests are needed as interference factor depends also on antenna
position and wind incidence angle.

Acknowledgements
We acknowledge the funding support given by FAPESPFunda@*ao de Amparo a"
Pesquisa do Estado de S*ao Paulo and the funding support given by CAPES,
Funda@*ao Coordena@*ao de Aperfei@oamento de Pessoal de N!vel Superior, Brazil
that made this work possible.
We also acknowledge the contributions by various members of the technical staff
of the Boundary Layer Wind Tunnel Laboratory of the University of Western
Ontario to carry out the experimental phases of the study.

References
[1] ESDU Item 81028, Engineering Science Data Unit, Lattice Structures. Part 1: Mean Fluid Forces on
Tower-like Space Frames, London, October 1990.
[2] J.D. Holmes, R.W. Banks, G. Roberts, Drag and aerodynamic interference on Microwave dish
antennas and their supporting towers, J. Wind Eng. Ind. Aerodyn. 50 (1993) 263270.
[3] ESDU Item 74031, Engineering Science Data Unit Characteristics of atmospheric turbulence near
ground, Part II single point data for strong winds (neutral atmosphere) London, March 1975.
[4] P.N. Georgiou, A study of the wind loads on building frames, Thesis (Master degree), Faculty of
Engineering Science, London, Canada, University of Western Ontario, 1979.
[5] P.N. Georgiou, B.J. Vickery, Wind loads on building frames, In: J.E. Cermak, (Ed.), Proceedings of
the Fifth International Conference, Vol. 1, Fort Collins, Colorado, USA, July 1979, Pergamon,
Oxford, 1980, pp. 421433.
[6] European Committee for Standardization, Eurocode 1: Basis of design and actions on structures
part 24: Actions on Structureswind actions, CEN, 1995.
[7] National Building Code of Canada, NBCCLive Loads Due to Wind, Canadian Commission on
Buildings and Fire Codes, National Research Council of Canada, 1995.

ARTICLE IN PRESS
1022

C.F. Carril Jr. et al. / J. Wind Eng. Ind. Aerodyn. 91 (2003) 10071022

[8] American Society of Civil Engineers, ASCE 7-95, Minimum Design Loads for Buildings and
Structure, New York, 1995.
[9] Standards Australia/Standard New Zealand AS/NZS 1170.22002, Structural design actions, Part 2:
Wind actions, Sydney, 2002.
[10] Standards Association of Australia AS 39951994, Design of steel lattice towers and masts, Sydney,
1994.
[11] British Standard BS 8100, Lattice towers and mast, Part 1, Code of Practice for Loading, London,
1986.
*
[12] Associa@*ao Brasileira de Normas T!ecnicas, NBR-6123 For@as devidas ao vento em edica@oes,
Rio de Janeiro, 1988.

Vous aimerez peut-être aussi