ARTICLE IN PRESS

Journal of Wind Engineering

and Industrial Aerodynamics 96 (2008) 1015–1028
www.elsevier.com/locate/jweia

Wind loads on free-standing canopy roofs:

Part 1 local wind pressures
Yasushi Uematsua,, Theodore Stathopoulosb, Eri Iizumic
a
New Industry Creation Hatchery Center, Tohoku University, Aoba-ku, Sendai 980-8579, Japan
b
Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, Canada H3G 1M8
c
Department of Architecture and Building Science, Tohoku University, Aoba-ku, Sendai 980-8579, Japan
Available online 26 July 2007

Abstract

Wind loads on free-standing canopy roofs have been studied in a wind-tunnel. Three types of roof
geometries, i.e. gable, troughed and mono-sloped roofs, with roof pitches between 01 and 151, were
tested. Wind pressures were measured simultaneously at many points both on the upper and lower
surfaces of the roof model for various wind directions. The paper describes the characteristics of the
local wind pressures with special attention to the mechanism and estimation of the peak net
pressure–difference coefficients. Based on the results for the most critical positive and negative peak
pressure coefficients irrespective of wind direction, the wind force coefficients for the design of
cladding and its immediately supporting structure are proposed. The proposed values are compared
with the previous wind-tunnel results as well as with the specifications of the Australian/New Zealand
(AS/NZ) Standard (2002).
r 2007 Elsevier Ltd. All rights reserved.

Keywords: Free-standing canopy roof; Wind pressure; Wind-tunnel experiment; Cladding; Design wind load;
Codification

1. Introduction

Free-standing canopy roofs are used for many structures, such as agricultural facilities
(barns, etc.), bus and railway stations, carports, and modern, lightweight, tension
membrane structures. Since the roofs are supported by columns and no walls, wind action

Corresponding author. Tel./fax: +81 22 795 7875.

E-mail address: yu@venus.str.archi.tohoku.ac.jp (Y. Uematsu).

0167-6105/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jweia.2007.06.047
 ARTICLE IN PRESS
1016 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

is directly exerted both on the upper and lower surfaces. Therefore, these roofs seem
more vulnerable to wind actions than those of enclosed buildings. In practice, such
roofs often experience damage during windstorms. Since wind flow around free-standing
canopy roofs is rather complicated, the wind forces on the roofs are influenced by
many factors, such as roof shape, roof pitch, obstruction under the roof and wind
direction. Although several researchers have made extensive studies on this subject, the
number of the studies is rather limited, compared with that of enclosed buildings. This is
probably due to difficulties in model making and pressure measurement in the wind-tunnel
experiments.
The primary purpose of the present study is to investigate the characteristics of wind
forces on free-standing canopy roofs in detail based on a series of wind-tunnel experiments
and thereby to discuss the design wind force coefficients for this type of structures. The
present paper and its companion paper describe the experimental results and discussion.
This paper (Part 1) focuses on the local net pressures, or the pressure difference between
the upper and lower surfaces, and the wind force coefficients for the design of cladding and
its immediately supporting structure. The results on the overall wind forces and moments
are described in a companion paper (Part 2).

2. Survey of previous studies

Gumley (1984) made an extensive parametric study investigating the effects of roof
shape, roof pitch, aspect ratio of the roof, eaves height, wind direction, and internal
stacking arrangement on canopy roof wind forces. He measured the mean and peak
pressures spatially averaged over various areas of the roof. Subsequently, full-scale
measurements of wind pressures on two agricultural canopy roof structures were reported
by Robertson et al. (1985). Based on the results, they proposed a set of wind force
coefficients for designing such structures. Another important set of experiments was
undertaken by Letchford and Ginger (1992) and Ginger and Letchford (1994). They
measured the mean and peak pressures at points and spatially averaged over several areas
of the roof. Based on a comparison of their results with the Australian wind loading code
(1989 version), they pointed out that the code provisions might underestimate the wind
loads significantly in some cases. Recently, Altman (2001) has made extensive
measurements of overall forces and moments acting on mono-sloped and duo-pitched
canopy roofs using a high-frequency force balance developed at Clemson University, USA.
Based on the experimental results, he proposed a set of wind force coefficients to be used
for designing the main wind force resisting systems.
Uematsu and Stathopoulos (2003) reviewed the previous studies, including the above-
mentioned ones, and compared the results with each other. Regarding the characteristics
of the wind forces on canopy roofs, they detected the following features:

(1) The roof pitch b affects the wind forces most significantly. However, its effect becomes
minute when b4301.
(2) There exists a significant disparity between the results of previous experiments for low
roof pitches, such as bo151, for example. The results for medium roof pitches,
151obo301, agree relatively well with each other.
(3) Details of the model, such as roof thickness and roof supporting system, affect the
results significantly.
 ARTICLE IN PRESS
Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028 1017

(4) The effect of roof aspect ratio (ranging from 1 to 4) on the wind force coefficients is
small.
(5) The experimental data both for mono-sloped and troughed roofs are rather limited.
The rarity of data is even more distinct for local pressure coefficients.

Furthermore, a survey of the pitch of canopy roofs constructed in Japan was made.
Troughed roofs with b ¼ 31 or 111 and mono-sloped roofs with b ¼ 111 are commonly
used for the Japan Railway (JR) stations. Regarding gable-roofed agricultural facilities,
the roof pitch of approximately 151 is often used for barns and that of approximately 181 is
common for livestock housings.
Based on the above discussion, this paper focuses on relatively low roof pitches.

3. Notation and coordinate system

The notation and coordinate system used in the present study are schematically illustrated in
Fig. 1; the definition for troughed roofs is similar to that for gable roofs. The pressure
coefficient Cp is defined in terms of the velocity pressure qH at the mean roof height H. The net
pressure coefficient Cf, positive downward, is defined as Cf ¼ CpUCpL, with CpU and CpL
being the pressure coefficients on the upper and lower surfaces, respectively.

4. Wind-tunnel experiments

4.1. Wind simulation

The experiments were carried out in the boundary layer wind-tunnel of the Centre
for Building Studies, Concordia University. The working section of the tunnel is
12.2 m long and 1.8 m wide. The height of the working section is approximately 1.6 m.

Mono-sloped roof b*

= 0° b
b H

= 0° x l
b*/2 b*/2

Gable roof

b
= 0° H
= 90°

Fig. 1. Notation and coordinate system. (a) Cross section; (b) roof plan.
 ARTICLE IN PRESS
1018 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

An open-country exposure was simulated in the wind-tunnel. The mean velocity profile
of the flow has a power-law exponent of a ¼ 0.14. The roughness length z0 is
approximately 0.01 cm. Assuming a boundary layer scale of 1:100, the full-scale value of
z0 is approximately 1 cm, which is within the range of full-scale measurements for flat
open-country exposure. The longitudinal integral scale Lx of turbulence is approximately
0.4 m at a height of z ¼ 5 cm. The turbulence intensity Iu at z ¼ 5 cm is 0.14. The mean
wind speed at z ¼ 5 cm is approximately 9.8 m/s, which results in a velocity scale of
approximately 1:2.8 for a strong wind event.

4.2. Models

Fig. 2 shows a canopy roof model on the turntable of the wind-tunnel of Concordia
University. When making the model, special care was taken in decreasing the roof
thickness and column width to avoid the distortion of the flow around the roof. The roof
consists of two identical ‘basic’ models (mono-sloped roof) 7.5 cm wide and 15 cm long.
The geometric scale of the models is assumed 1:100. The roof model is made of two
galvanized steel sheets 0.3 mm thick and consists of a sandwich structure; the thickness of
the roof is only 2 mm. Twelve pressure taps of 0.4 mm diameter are drilled on each side of
the basic model. Tapping locations are identical on both the upper and lower surfaces so
that the net pressure difference can be obtained. Two types of tap arrangements are used;
one is for the overall wind force measurements (regularly arranged over the whole area)—
see Fig. 3a—and the other is for the local pressure measurements (densely arranged in the
edge and corner regions)—see Fig. 3b. The roof pitch b can be changed from 01 to 151. The
column length can also be changed so that the mean roof height is kept constant, i.e.
H ¼ 5 cm. The basic model is supported by six 5-mm wide square hollow columns. The
columns are placed inside the eaves to reduce interference at the separation lines.
Combining the two basic models in different ways, three kinds of roof geometries with an
aspect ratio of l/b ¼ 1.0 are tested; that is, gable, troughed and mono-sloped roofs.
Therefore, each wind-tunnel model has 24 pressure taps both on the upper and lower
surfaces (48 taps in total) for each tap arrangement.

Fig. 2. Wind-tunnel model (gable roof, b ¼ 151).

 ARTICLE IN PRESS
Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028 1019

12.5

12.5
25

25

21

25

21
4

4
15

20
20
37.5

20

150
45

150
50
37.5

36
15

4
75

75

Fig. 3. Pressure tap arrangements for the basic model. (a) Arrangement 1; (b) arrangement 2.

4.3. Pressure measurements

Wind pressures at 48 points are measured by a ZOC33 Electronic Pressure Scanning

Module (ZOC33/64Px) and sampled in parallel at a rate of 250 Hz on each channel for a
period of 16.4 s (10 min in full scale) in each run. The tubing system consists of a 4.8 cm
long stainless steel tube of 0.4 mm ID and a 30 cm long PVC tube of 1.5 mm ID with a
brass restrictor. The connections between the stainless steel tube and the PVC tube are all
installed inside the hollow columns. The tubing effects are numerically compensated by
using the gain and phase-shift characteristics of the pressure measuring system. The mean,
standard deviation, maximum and minimum pressure difference coefficients are computed.
In the present paper, the ensemble average of the results obtained from nine consecutive
runs is used. The wind direction y is varied from 01 to 3451 at a step of 151; in some cases, it
is from 01 to 901 or from 01 to 1801 considering the symmetry of the tap arrangement.

5. Results and discussion

5.1. Mean wind pressures

Fig. 4 shows a comparison between the present experiment and a full-scale measurement
by Robertson et al. (1985) for the distribution of mean pressure difference coefficient along
 ARTICLE IN PRESS
1020 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

ridge

1.5
Present
1.0
Robertson et al. (1985)
0.5
Cf (mean)

0.0

-0.5

-1.0

-1.5
0.0 0.2 0.4 0.6 0.8 1.0
x/b*

Fig. 4. Comparison for the mean pressure difference coefficients along the centerline (y ¼ 01).

Table 1
Geometry of the test models

Roof configuration Present work Robertson et al. (1985)

b 151 151
l/b 1.04 (15 cm/14.4 cm) 1.61 (14.5 m/9.0 m)
H/b 0.345 (5.0 cm/14.4 cm) 0.731 (6.58 m/9.0 m)

the centerline of a gable roof when y ¼ 01. The geometry of the test model is summarized
in Table 1. The roof pitch is 151 in both cases, but the l/b and H/b ratios are somewhat
different from each other. It is seen that the agreement between these two results is fairy
good. This indicates the validity of the present experiment regarding the mean flow field
around the roof. Furthermore, the experimental data show that the roof pitch is a
parameter dominating the mean flow field.

5.2. Peak wind pressures

Shown in Fig. 5 are computerized contours of the most critical positive and negative
peak Cf values (Cfmax and Cfmin) irrespective of wind direction for the three types of roofs
with b ¼ 151. Considering the symmetry of the roof, only one quarter or one half of the
roof area is depicted. Very large negative peak values are induced in the leeward ridge
corner for the gable roof and in the windward eave corner for the troughed and mono-
sloped roofs; in practice, such large values occur when y ¼ 30–451, probably due to the
generation of conical vortices. The most critical values, both positive and negative,
generally increase in magnitude with an increase in b. When the roof pitch is the same,
larger peak values are induced in mono-sloped roofs than in gable and troughed roofs.
Fig. 6 shows the largest positive and negative Cf values irrespective of wind direction
over the whole roof area for gable roofs, plotted as a function of b; in practice, these values
are induced in the edge and corner regions. In the figure, the results of Gumley (1984) are
 ARTICLE IN PRESS
Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028 1021

Positive Negative
(eaves) (eaves)

Positive Negative
(higher eaves) (higher eaves)

(ridge) (ridge)
C.L C.L

Positive Negative
(eaves) (eaves)

(ridge) (ridge) (lower eaves) (lower eaves)

C.L C.L

Fig. 5. Most critical positive and negative pressure difference coefficients (b ¼ 151). (a) Gable roof; (b) troughed
roof; (c) mono-sloped roof.

6
Gable roof
4

2 Present (max)
Cfmax, Cfmin

Present (min)
0
Gumley (1984) (max)
-2 Gumley (1984) (min)

-4

-6
0 10 20 30 40
Roof pitch (deg)

Fig. 6. The largest positive and negative Cf values irrespective of wind direction over the whole roof area.

also plotted for the purpose of comparison. The present results are consistent with
Gumley’s results, except for Cfmax in the range of bp101. Similar features were observed
for mono-sloped roofs. The difference between these two results for Cfmax in the range of
bp101 is not clear, because details of Gumley’s results are not given in his paper.
 ARTICLE IN PRESS
1022 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

5.3. Correlation of pressures on the upper and lower roof surfaces

Fig. 7 shows a phase-plane representation of the pressure coefficients CpU and CpL on
the upper and lower roof surfaces of a mono-sloped roof with b ¼ 151 at a pressure tap
where the largest negative pressure difference is observed. The overall correlation between
CpU and CpL is not high; the coefficient of correlation is approximately 0.6. However,
regarding the largest values, the correlation is relatively high; i.e. large negative and
positive pressures occur almost simultaneously on the upper and lower roof surfaces,
which results in the largest pressure differences. This is also the case for the other roof
configurations. Such a feature implies that the maximum and minimum pressure difference
coefficients Cfmax and Cfmin are approximated by the difference between the extreme
pressure coefficients on both surfaces:
C nfmax ¼ C pUmax  C pLmin ; C nfmin ¼ C pUmin  C pLmax . (1)
A comparison between the actual and estimated values of Cfmax and Cfmin on a mono-
sloped roof is shown in Fig. 8; the results for all pressure taps and wind directions are
plotted in the figure. The solid lines represent the linear regression lines for C nfmax and
C nfmin ; the slope is approximately 1.0 and the magnitude of the intercept is 0.3–0.4. Quite
similar results were observed for all the cases tested. Therefore, Eq. (1) can be used for
estimating Cfmax and Cfmin reasonably. For larger values of Cfmax and |Cfmin|, which are
induced in the edge and corner regions of the roof, the prediction error (overestimate) is
less than approximately 10%.

5.4. Estimation of the maximum and minimum peak pressure difference coefficients

In Fig. 9 the gust effect factor Gf, defined as the ratio of the peak to the mean pressure
difference coefficients (i.e. Cfpeak/Cfmean), is plotted against |Cfpeak| for mono-sloped roofs.
Note that Cfpeak represents both Cfmax and Cfmin for simplicity. The data for various taps
and wind directions are plotted in the figure. When the values of |Cfpeak| are small, the data
scatter over a wide range. However, as the value of |Cfpeak| increases, the value of Gf
collapses into a narrow range around Gf ¼ 2.2 (dashed line). Based on the quasi-steady

2.0
higher eaves
tap
1.5
CpL (bottom)

lower eaves
1.0

0.5

0.0
-8 -6 -4 -2 0
CpU (top)

Fig. 7. Phase-plane representation of CpU and CpL (mono-sloped roof, b ¼ 151, y ¼ 301).
 ARTICLE IN PRESS
Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028 1023

6
y = 0.97 x + 0.37
4

Cf*max, Cf*min 0
-10 -5 0 5
-2

-4
y = 1.02 x - 0.33
-6 Cf*max
-8 Cf*min
-10
Cfmax, Cfmin

Fig. 8. Relation between the actual and estimated values of Cfmax and Cfmin (mono-sloped roof, b ¼ 151).

20
beta: 5 deg

15 beta: 10 deg
beta: 15 deg
Gf

10

5
Gf = 2.2

0
2.0 3.0 4.0 5.0 6.0 7.0 8.0
|Cfpeak|

Fig. 9. Gust effect factor for Cf (mono-sloped roof).

assumption, Cfpeak may be given by

C fpeak  G f C fmean  G 2v C fmean ¼ ð1 þ gv I uH Þ2 C fmean , (2)
where Gv, gv and IuH represent the gust factor, peak factor and turbulence intensity of the
flow at the mean roof height H, respectively. Substituting IuH ¼ 0.14 and Gf ¼ 2.2 into the
equation, we obtain gv ¼ 3.5, which is often used for evaluating the gust wind speed.
Furthermore, the standard deviation C 0f of the pressure difference coefficient may be
given by
C 0f  2I uH jC fmean j. (3)
This relation is valid only when the values of |Cfmean| are relatively large. Denoting the
peak factor of pressure difference by gf, Cfpeak is represented by the following equation:
C fpeak ¼ C fmean þ sgnðC fmean Þgf C 0f  ð1 þ 2gf I uH ÞC fmean . (4)
 ARTICLE IN PRESS
1024 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

From Eqs. (2) and (4), we obtain

gf  gv þ 0:5 g2v I uH . (5)
Substitution of gv ¼ 3.5 and IuH ¼ 0.14 into Eq. (5) gives gfE4.4. Such an interpretation is
validated by the experimental results of the present as well as previous studies (for
example, Stathopoulos, 1983). Fig. 10 shows the peak factor gf plotted against |Cfpeak| for
mono-sloped roofs. For larger peak values, such as |Cfpeak| 4 4, for example, the value of
gf collapses into a narrow range around gf ¼ 4.4 (dashed line).
From the above discussion, it may be concluded that the largest peak pressures induced
in the edge and corner regions of the roof are mainly induced by large-scale turbulence in
the approach flow. These peak pressure difference coefficients can be evaluated by the gust
effect factor approach. Note that this is not the case for the internal region where the
values of |Cfmean| are relatively small.

5.5. Wind force coefficients for the design of cladding and its immediately supporting
structure

Based on such results as shown in Fig. 5, the peak wind force coefficients for the design
of cladding and its immediately supporting structure is discussed; a tributary area of 1 m2 is
considered in the present paper. The roof is divided into several zones, such as edge, corner
and internal zones, and the most critical positive and negative peak values are provided for
each zone. The results are summarized in Figs. 11–13. The definition of the zones depends
on the sign of the wind force coefficient as well as on the roof shape.
The proposed values are compared with the experimental results by Gumley (1984) and
with the specifications of the Australian/New Zealand (AS/NZ) Standard (Standards
Australia, 2002). Sample results are shown in Figs. 14 and 15 for gable and mono-sloped
roofs, respectively. For calculating the peak wind force coefficients based on the AS/NZ
Standard, it is assumed that the area reduction factor is 1.0. Furthermore, the gust factor is
assumed to be 1.55, based on the ‘ratio of probable maximum speed averaged over t sec to
hourly mean speed for hurricanes’ given in ASCE 7-95 (American Society of Civil
Engineers, 1996). The experimental results for carports with curved mono-sloped roofs are

6
Peak factor

gf = 4.4
4

beta = 5 deg
2 beta = 10 deg
beta = 15 deg
0
0 2 4 6 8
|Cf(peak)|

Fig. 10. Peak factor for Cf (mono-sloped roof).

 ARTICLE IN PRESS
Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028 1025

Fig. 11. Design wind force coefficients for gable roofs. (a) Positive wind force coefficient; (b) negative wind force
coefficient.

Fig. 12. Design wind force coefficients for troughed roofs. (a) Positive wind force coefficient; (b) negative wind
force coefficient.

also plotted in Fig. 15 (see Uematsu, 2003). Note that the roof of the carport is not planar
but curved. Therefore, the roof pitch is defined by the slope of the straight line connecting
the higher and lower eaves. Regarding the mono-sloped roofs, the present results are
consistent with the specification of the AS/NZ Standard, which is hereafter referred to as
the AS/NZ specification. They are also in relatively good agreement with the experimental
data for carports. Some difference is seen for the negative values in zone B (higher eaves
 ARTICLE IN PRESS
1026 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

Fig. 13. Design wind force coefficients for mono-sloped roofs. (a) Positive wind force coefficient; (b) negative
wind force coefficient.

5 0
Present
4 -1
AS/NZ (2002) Gumley (1984)
-2
3
Cf

-3
Cf

2
AS/NZ (2002) -4
Present
1 -5
Gumley (1984)
0 -6
0 10 20 30 0 10 20 30
Roof pitch (deg) Roof pitch (deg)

Fig. 14. Design wind force coefficients as a function of roof pitch (gable roof). (a) Eaves corner (positive);
(b) ridge corner (negative).

corner). This difference is probably due to the effect of curvature of the roof on the local
flow around the eaves. A significant difference is seen between the present results and the
AS/NZ specification for gable roofs; the specification is rather small in magnitude. The
results of Gumley (1984) for the positive values are generally large compared with the
present results as well as with the AS/NZ specification. Regarding the troughed roofs, not
 ARTICLE IN PRESS
Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028 1027

10 0
-1
8
-2
6
Cf

-3

Cf
AS/NZ (2002)
4
Present -4 Present
Gumley (1984) Gumley (1984) AS/NZ (2002)
2 -5
Uematsu (2003) Uematsu (2003)
0 -6
0 10 20 30 0 10 20 30
Roof pitch (deg) Roof pitch (deg)

Fig. 15. Design wind force coefficients as a function of roof pitch (mono-sloped roof). (a) Lower eaves corner
(positive); (b) higher eaves corner (negative).

shown in the present paper, the agreement between the present results and the AS/NZ
specification is relatively good. There exists a significant difference between the present and
Gumley’s results. A full comparison is presented in Iizumi (2005).

6. Concluding remarks

The wind pressures acting on free-standing canopy roofs have been investigated based
on a series of wind-tunnel experiments. The roof pitch investigated in the present study is
limited to a range from 01 to 151. Large wind forces are induced in the edge and corner
regions in oblique winds. The magnitude is larger for mono-sloped roofs than for gable
and troughed roofs. The results indicate that the maximum and minimum wind force
coefficients on these regions are approximated by the difference between the extreme
pressure coefficients on both surfaces with an overestimate less than approximately 10%.
Furthermore, they can be evaluated by a gust effect factor approach, in which the gust
effect factor is estimated from the gust factor of the approach flow. This feature implies
that we can evaluate the design wind force coefficients from the experimental data for the
mean wind pressure coefficients on the upper and lower surfaces. Furthermore, based on
the results for the most critical positive and negative peak pressure difference coefficients,
the peak wind force coefficients for the design of cladding and its immediately supporting
structure are proposed. The roof is divided into several zones and the design wind force
coefficient for each zone is provided as a function of roof pitch. Comparisons with the AS/
NZ Standard (2002) show generally good agreement for mono-sloped roofs, and some
difference for gable and troughed roofs.

Acknowledgment

This study was made during the first author’s appointment at the Department of
Building, Civil and Environmental Engineering, Concordia University, Montreal, Quebec,
Canada, as a visiting professor, supported by the Ministry of Education, Culture, Sports,
Science and Technology, Japan, from May 2002 to February 2003. The authors are much
indebted to Mr. Kai Wang, graduate student of Concordia University, for his assistance
with the experiments.
 ARTICLE IN PRESS
1028 Y. Uematsu et al. / J. Wind Eng. Ind. Aerodyn. 96 (2008) 1015–1028

References

Altman, D.R., 2001. Wind uplift forces on roof canopies, M.Sc. Thesis, Clemson University.
American Society of Civil Engineers, 1996. Minimum design loads for buildings and other structures,
ANSI/ASCE 7-95.
Ginger, J.D., Letchford, C.W., 1994. Wind loads on planar canopy roofs, part 2: fluctuating pressure distributions
and correlations. J. Wind Eng. Ind. Aerodyn. 51, 353–370.
Gumley, S.J., 1984. A parametric study of extreme pressures for the static design of canopy structures. J. Wind
Eng. Ind. Aerodyn. 16, 43–56.
Iizumi, E., 2005. Design wind loads for free-standing canopy roofs, M. Eng. Thesis, Tohoku University.
Letchford, C.W., Ginger, J.D., 1992. Wind loads on planar canopy roofs, Part 1: mean pressure distributions.
J. Wind Eng. Ind. Aerodyn. 45, 25–45.
Robertson, A.P., Hoxey, R., Moran, P., 1985. A full-scale study of wind loads on agricultural ridged canopy roof
structures and proposals for design. J. Wind Eng. Ind. Aerodyn. 21, 167–205.
Standards Australia, 2002. Australian/New Zealand Standard, AS/NZS 1170.2.
Stathopoulos, T., 1983. Fluctuating wind pressures on low building roofs. J. Struct. Eng. ASCE 109 (1), 226–271.
Uematsu, Y., 2003. Wind tunnel study of design wind force coefficients for carports. Japan Exterior Industrial
Association (in Japanese).
Uematsu, Y., Stathopoulos, T., 2003. Wind loads on free-standing canopy roofs: a review. J. Wind Eng. JAWE
95, 245–256.