The Eighth Asia-Pacific Conference on Wind Engineering,

December 10–14, 2013, Chennai, India

WIND LOADS ON CURVED ROOFS: EFFECT OF SIDE WALLS

A. Abraham1, Nagesh R. Iyer2, G. M. Samuel Knight3, K. Muthumani4 and N. Lakshmanan5
1
Senior Scientist, CSIR-Structural Engineering Research Centre, Chennai, TN, India, abraham@serc.res.in
2
Director, CSIR-Structural Engineering Research Centre, Chennai, TN, India, director@serc.res.in
3
Professor, Structural Engineering Division, Anna University, Guindy, Chennai, TN, India, gmsk@annauniv.edu
4
Chief Scientist, CSIR-Structural Engineering Research Centre, Chennai, TN, India, kmm@serc.res.in
4
Former Director, CSIR-Structural Engineering Research Centre, Chennai, TN, India, nlakshmanan@live.com

ABSTRACT

Wind tunnel pressure measurement studies on a rigid model (scale 1:300) of structure with curved roof has been
carried out under open and suburban terrain conditions. The external pressures exerted over the roof due to the
presence of side walls have been investigated for the following test cases: (i) roof springing from ground level
i.e., side wall height, h = 0.0d and (ii) roof at elevated levels i.e., h = 0.15d and 0.29d, where ‘d’ is span of the
roof. The model has been instrumented with point pressure taps on the roof and area averaged pressure taps on
the gable end wall and sidewalls. In this paper, results pertaining to mean and standard deviation of pressure
~
coefficients, ( Cpe and C pe ), at edge and mid arch locations on the roof for typical angles of wind incidence i.e.,
wind blowing (i) parallel (T = 0q), (ii) perpendicular (T = 90q), and (iii) diagonal w.r.t. normal to gable end (T =
45q) have been presented. Based on the study, it is observed that the centre half and windward region of the roof
of the model with h = 0.29d is subjected to high suction than that for model with h = 0, which is due to the
presence of sidewall. Further, evaluated values of Cpe for wind blowing parallel and perpendicular to the axis of
the arch have been compared with values given in codes of practice and the evaluated values are observed to be
between two values reported in AS/NZS and ASCE/SEI codes of practice for T = 0q and reasonable agreement at
windward quarter for T = 90q.

Keywords: Pressure measurement, Curved roof, Open and Sub-urban terrain, External pressure
coefficients

Introduction

Curved roof structures are being used for industrial structures especially in power
sector, aviation sector, etc, due to their unobstructed clear span, which are essential for the
functionality purposes. These structures are springing either from ground or at an elevated
level. Wind loads are one of the important load criteria in design of such structures. Currently
available information of wind loads on curved roofs is scanty. Various codes of practice viz.
IS: 875 (Part 3): 1987 (1989), AS/NZS: 1170.2: 2002 (2005), ASCE/SEI 7-05 (2006), AIJ
(2004) and User’s Guide-NBC: 1995 (2002) specify the external pressure coefficients on
curved roof structures for wind blowing parallel and perpendicular to the axis of the arch and
while few codes of practice viz. GBJ 9-87 (1994) and EN1991-1-4 (2004) specify only for
wind blowing in perpendicular to the axis of the arch. The importance of wind effects on
curved roofs in boundary layer flows by using Boundary Layer Wind Tunnel (BLWT) have
been reported by researchers viz. Ginger (2004), Blackmore and Tsokri (2006), Kasperski
(2008) and Natalini et al. (2013). Application of Proper Orthogonal Decomposition (POD)
technique on structures with large span roof to predict wind loads and/or its effects was
reported by Li et al. (2006) and Yang et al. (2013). Analytical prediction of load effects using
wind tunnel pressure data applied on structures with curved roofs was reported by Ding and

Proc. of the 8th Asia-Pacific Conference on Wind Engineering – Nagesh R. Iyer, Prem Krishna, S. Selvi Rajan and P. Harikrishna (eds)
Copyright c 2013 APCWE-VIII. All rights reserved. Published by Research Publishing, Singapore. ISBN: 978-981-07-8011-1
doi:10.3850/978-981-07-8012-8 117 786
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

Tamura (2013). From the literature review, it is observed that (a) the rise ‘f’ to span ‘d’ ratio
‘f/d’ has strong influence on the aerodynamic loads on curved roofs, and (b) the other two
geometric parameters i.e., side wall height ‘h’ to span ratio ‘h/d’ and building length ‘l’ to
span ratio ‘l/d’ do not have significant on wind induced loads. Since the actual load
distribution on curved roof structures are so complex, wind tunnel tests are required for
evaluating aerodynamic behavior and wind induced pressures on such structures. In this
paper, an attempt has been made to understand the distribution of pressures exerted over the
roof, which is springing from ground and at elevated levels. This paper presents results on
~
Cpe and C pe obtained at edge and mid arch locations based on wind tunnel studies conducted
on rigid models of curved roofs for T = 0q, 45q and 90q. Further, a comparison between
evaluated values of Cpe and values given in various codes of practice for T = 0q and 90q have
been made.

Wind Tunnel Experimental Programme

Model Fabrication and Instrumentation

A rigid model (scale 1:300) of a structure with curved roof has been fabricated using
acrylic sheet of uniform thick of 3 mm. The roof of the model has a rise, ‘f’ of 0.107 m and
span, ‘d’ of 0.367 m. The roof has been investigated for three cases i.e., with and without side
walls height, ‘h’ i.e., h = 0.15d & 0.29d and h = 0.0d. The roof has been instrumented with
165 point pressure taps, along with 5-port manifolds to pneumatically averaged pressure taps
of 7 nos., at the gable (windward) end wall. The roof without sidewall hereafter referred as
Model1 (M1). Depending upon the sidewall height, 5-port manifolds to pneumatically
averaged pressure taps of 13 nos. and 35 nos. on the side walls of models, hereafter, referred
as Model2 (M2) and Model3 (M3) have been provided. These are in addition to pressure taps
provided in model M1. The relative locations of the pressure taps along the arch direction
(i.e., x/d) and along the length direction (y/l) of the model are shown in Fig. 1.
y = l/2
(a) f (b)
y = l/3
1
h
l
Roof l/2

d/2 E y = l/12 y/l

C y/l 0 x/d 1
A x/d D A B
0 B Point pressure taps C D Side wall
d 1
l/2 Area ave. pressure taps
E
0q T Windward
Wind Wind Not to scale
gable wall 0q T
Fig. 1 Schematic diagram of curved roof structure and relative positions of pressure taps. (a) Model
M3 (typical) and (b) Typical instrumentation scheme for model M3 (in plan)

Wind Tunnel Tests and Data Processing

The wind tunnel tests have been conducted under suburban terrain conditions for three
different test cases on models with f/d = 0.29 and l/d = 1.23, as given below:
Test case no. Sidewall height to span ratio ‘h/d’ Sidewall height to rise ratio ‘h/f’ Model code
1 0.0 0.0 M1
2 0.15 0.5 M2
3 0.29 1.0 M3

787
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

For a given test case, simultaneous pressures have been collected for various angles of
wind incidence ranging from T = 0q to 360q in steps of every 30q, and also at angles 45q,
135q, 225q and 315q. Photographic views of models M1, M2 and M3 tested in BLWT under
suburban terrain condition for typical angles of wind incidence are shown in Fig. 2. Pressures
have been acquired for a duration of 15 sec (by considering velocity scale ratio as 1:1, it
corresponds to a sampling duration of 1¼ h in full scale) with sampling frequency of 625 Hz
in case of PSI scanners (2 nos., each 32 channels) and with sampling frequency of 500 Hz in
case of scanivalve pressure scanners (2 nos., each 64 channels + 1 no. of 32 channels).
Pressure data have been acquired for mean wind velocities of about 9 m/s, 10 m/s and 11 m/s,
measured at the crown of the roof for models M1, M2 and M3. These mean wind velocities
are considered as reference velocities, ‘ Vref ’, and all the pressure coefficients are based on
corresponding reference dynamic pressure, ‘ q ref ’, corresponding to the above reference wind
velocities. Pressures have been measured on the roof from edge section y = l/12 (near
windward gable end for wind blowing normal to gable end i.e., y/l = 0 and x/d = 0) to mid
section y = l/2 at an interval of y = l/24 in length direction. Three data runs have been taken
for every angle of wind incidence in each test case.

M1, T = 90q M2, T = 45q

M3, T = 0q

Fig. 2 Photo views of models M1, M2 and M3 in BLWT under suburban terrain condition

Simulation of Terrain Characteristics

The curved roof models are considered to be located in open and suburban terrain
conditions. Accordingly, profile of mean velocity, profile of turbulence intensity and
turbulence spectrum of longitudinal wind velocity corresponding to the above terrains have
been simulated in the wind tunnel by using a trip board followed by boards of wooden
roughness elements, as vortex generators. The results of the simulation in terms of profiles of
mean velocity and turbulence intensity, and turbulence spectrum with Karman spectrum given
in literature [Simiu and Scanlan (1996)] are shown in Fig. 3. The power law coefficient, ‘D’
of the mean velocity profile has been experimentally found to be equal to 0.165 for open and
0.215 for sub-urban terrain, respectively. The measured turbulence intensities at crown of the
models of M1, M2 and M3 are about 0.16, 0.15 & 0.14 for open and 0.25, 0.23 & 0.21 for
sub-urban terrain, respectively.

788
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

120 120
sub-ur sub-ur
open open
90 sub-ur 90
Height (cm)

Height (cm)
open sub-ur

n*S(n) / Variance
open
60 60

30 30
(c)
(a) (b)
0 0
0 10 20 30 0 10 20 30 40
Mean velocity (m/s) Turbulence intensity (%)
n*Lu / U
Fig. 3 Simulated wind characteristics (a) Profile of mean velocity (b) Profile of
turbulence intensity and (c) Turbulence spectrum of longitudinal wind velocity

Results and Discussions

All pressure data have been analysed by using the in-house developed script in
~
MATLAB and the values of C pe and C pe are obtained, as given below:
~
Pe ~ Pe
C Pe 2
C Pe 2
(1)
1 2 U Vref 1 2 U Vref
~
It can be noted that the trends of distributions of C pe and C pe obtained for the models M1, M2
and M3 for various angles of wind incidence are found to be similar but varying in
magnitudes, which could be due to the overall dimension of the model and oncoming flow
~
conditions (mean velocity and turbulence intensity). Hence, distributions of C pe and C pe for
model M1 (typical) at different arch locations (from edge ‘l/12’ to mid ‘l/2’, where l is the
length of the curved roof) for two important angles of wind incidence 0q and 90q under open
and suburban terrain condition have been plotted and are shown in Fig. 4. The following
observations were made:

x For T = 0q, suction values of C pe obtained at edge location (y = 0.083l) are found to be
maximum (due to flow separation), gradually reducing in the direction of wind and
when the flow reaches at mid location (y = l/2) it is found that suction values of C pe
are minimum (due to flow reattachment), for both terrain conditions.
x For T = 90q, small portion of the roof (x/d # 0.8 to 1.0, windward region) is subjected
to +ve pressures and centre region of the roof (x/d # 0.3 to 0.7) is subjected to high
suction pressures, due to flow separation and the leeward region (x/d # 0 to 0.3) is
subjected to low –ve suctions, due to pressure recovery. The above phenomenon
occurred at all the locations under both terrain conditions.
x The values of C pe obtained at mid location found to be higher for both terrain
conditions when compared to edge location, due to edge effects of the model with
respect to flow.
x The values of C pe (in suction zone) are found to be higher under open terrain than
those under suburban terrain condition, due to less oncoming turbulence level.
~
x For T = 0q, values of C pe obtained at edge location are found to be higher and
gradually reducing in the direction of wind and when the flow reaches at mid location
it is found that values of C pe are relatively minimum, for both terrain conditions.
~
x The values of C pe obtained under suburban terrain found to be higher than those
values obtained under open terrain condition.

789
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

0.8 0.083l 0.60 0.083l

l/8 l/8
T = 0q open T = 0q open
0.4 l/6 l/6
0.45

Std. dev. of Cp
0.208l 0.208l
Mean Cp

0.0 l/4 l/4

0.292l 0.30 0.292l
-0.4 l/3 l/3
0.375l 0.375l
0.15
-0.8 0.417l 0.417l
0.458l 0.454l
-1.2 l/2 0.00 l/2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d

0.8 0.083l 0.60 0.083l

T = 90q open l/8 T = 90q open l/8
0.4 l/6 l/6
0.45

Std. dev. of Cp
0.208l 0.208l
Mean Cp

0.0 l/4 l/4

0.292l 0.30 0.292l
-0.4 l/3 l/3
0.375l 0.375l
0.15
-0.8 0.417l 0.417l
0.454l
0.458l
-1.2 0.00 l/2
l/2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d
0.8 0.083l 0.60 0.083l
l/8 l/8
T = 0q suburban
0.4 l/6 l/6
0.45
0.208l
Std. dev. of Cp

0.208l
Mean Cp

0.0 l/4 l/4

0.292l 0.30 0.292l
-0.4 l/3 l/3
0.375l 0.375l
0.15
-0.8 0.417l 0.417l
0.458l T = 0q suburban 0.454l
-1.2 l/2 0.00 l/2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d

0.8 0.083l 0.60 0.083l

T = 90q suburban l/8 T = 90q suburban l/8
0.4 l/6 l/6
0.45
Std. dev. of Cp

0.208l 0.208l
Mean Cp

0.0 l/4 l/4

0.292l 0.30 0.292l
-0.4 l/3 l/3
0.375l 0.375l
0.15
-0.8 0.417l 0.417l
0.458l 0.454l
-1.2 l/2 0.00 l/2
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d
~
Fig. 4 Variation of Cpe and C pe at different arch locations on the model M1

~
Effect of Sidewalls on Distributions of Cpe and C pe
To understand the effect of sidewalls on pressures exerted over curved roof,
~
distributions of C pe and C pe for model M1 (without sidewall case, i.e., h = 0.0d) and models
M2 and M3 (with sidewall cases, i.e., h = 0.15d and 0.29d) at two different arch locations

790
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

(edge ‘0.083l’ and mid ‘l/2’) for T = 0q, 45q and 90q under open and suburban terrain
condition have been plotted and are shown in Fig. 5. The following observations are made:
x At both locations, the suction values of C pe obtained on the roof of model M3
(highest sidewall height) are observed to be higher than those values obtained for
models M1 and M2, for the selected angles of wind incidence. These values are
found to be higher under open terrain than under suburban terrain.
x The pressure values of C pe obtained on the roof of model M1 (which has no
sidewall) are observed to be higher than those for models M2 and M3, for T = 45q
and 90q. These values are found to be higher under suburban terrain than under open
terrain condition.
~
x The values of C pe are observed to be significantly affected due to the presence of
sidewalls for individual angles of wind incidence. For example, at edge location, the
~
values of C pe obtained on the roof of model M1 are observed to be higher than those
values obtained for models M1 and M2 for T = 0q.

0.8 0.60
T = 0q O: Open & S: Suburban M1_O
0.4 M1_O
0.45
Std. dev. of Cp

M1_S M1_S
Mean Cp

0.0 M2_O M2_O

0.30
M2_S
-0.4 M2_S
M3_O
0.15 M3_O
-0.8 M3_S T = 0q
M3_S
-1.2 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d
0.8 0.60
T = 45q M1_O
T = 45q
0.4 M1_O
0.45
Std. dev. of Cp

M1_S M1_S
Mean Cp

0.0
M2_O M2_O
0.30
-0.4 M2_S M2_S
M3_O 0.15 M3_O
-0.8
M3_S M3_S
-1.2 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d

0.8 0.60

T = 90q M1_O T = 90q M1_O

0.4
0.45
Std. dev. of Cp

M1_S M1_S
Mean Cp

0.0
M2_O M2_O
0.30
-0.4 M2_S M2_S
M3_O 0.15 M3_O
-0.8
M3_S M3_S
-1.2 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d
~
Fig. 5a Comparison on distributions of Cpe and C pe at edge location

791
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

0.8 0.60
T = 0q O: Open & S: Suburban M1_O T = 0q M1_O
0.4
0.45

Std. dev. of Cp
M1_S M1_S
Mean Cp

0.0
M2_O M2_O
0.30
-0.4 M2_S M2_S
M3_O 0.15 M3_O
-0.8
M3_S M3_S
-1.2 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d
0.8 0.60
T = 45q T = 45q
0.4 M1_O M1_O
0.45

Std. dev. of Cp
M1_S M1_S
Mean Cp

0.0
M2_O M2_O
0.30
-0.4 M2_S M2_S
M3_O 0.15 M3_O
-0.8
M3_S M3_S
-1.2 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d

0.8 0.60
0.5 T = 90q M1_O T = 90q M1_O
0.2 M1_S 0.45
Std. dev. of Cp

M1_S
Mean Cp

-0.1 M2_O
M2_O
0.30
-0.4 M2_S
M2_S
-0.7 M3_O
0.15 M3_O
-1.0 M3_S
M3_S
-1.3 0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/d x/d
~
Fig. 5b Comparison on distributions of Cpe and C pe at mid location

~
x In general, the C pe obtained on the roof of models M1, M2 and M3 under suburban terrain
are observed to be higher than those values obtained under open terrain condition.

Comparison Between Evaluated C pe and Values Given in Literature

A comparison have been made between evaluated values of C pe and the values
recommended in various codes of practice for three different regions at the roof for T = 90q
and is given in Table 1. Further, comparison on the values of C pe for T = 0q reported
elsewhere (Abraham et al., 2013) is also given in Table 2, for readers benefit. The following
observations are made:
x From Table 1, for T = 90q, the centre half of the roof is subjected to high suction, as
observed in many codes of practice and the values of Cpe for model M3 are
relatively higher than that for model M1 under both terrain conditions, is due to the
presence of side wall.

792
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

Table 1: Comparison of Cpe at three different regions as recommended in codes (for T = 90q)
Regions/ Windward quarter (d/4) Leeward
Centre half (d/2)
Codes of practice Model M1 Model M3 quarter (d/4)
IS: 875 (1989) +0.39 -0.34 -0.99 -0.40
-0.65, 0.0 (h/d=0) -0.31, 0.0 (h/d=0)
AS/NZS (2005) +0.16, 0.0 -0.24, 0.0
-0.85, 0.0 (h/d=0.29) -0.51, 0.0 (h/d=0.29)
ASCE/SEI (2006) +0.41 -0.36, +0.14 -0.99 -0.50
GBJ (1994) +0.32 +0.18 -0.80 -0.50
-1.19 (h/d=0)
AIJ (2004) -0.12 -0.44 -0.50
-1.38 (h/d=0.29)
EN (2004) +0.44 - -1.00 -0.40
-0.55 (h/d=0) +0.06 (h/d=0)
Suburban +0.42 +0.16
Present -0.76 (h/d=0.29) -0.07 (h/d=0.29)
study -0.70 (h/d=0) -0.05 (h/d=0)
Open +0.39 +0.06
-0.90 (h/d=0.29) -0.14 (h/d=0.29)

Table 2: Comparison on values of Cpe (for T = 0q)

Locations on the roof Remarks
Regions/
Codes of practice edge mid
IS : 875 (1989) -0.7 -0.5 -
f(have./ l); two values given to take into
AS/NZS (2005) -0.9, -0.4 -0.2, 0.2
account for turbulence effect
ASCE/SEI (2006) -0.9, -0.18 -0.3, -0.18 f(have./d)
f(f/d, h/d and l/d)
NBC (2002) -0.8* -0.3*
*for f/d = 0.17, h/d = 0.08 and l/d = 1
AIJ (2004) -1.105 -0.605 f(f/d and h/d)
-0.90 (h/d = 0) 0.02 (h/d = 0)
Open
Present -1.05 (h/d = 0.29) -0.20 (h/d = 0.29)
f(f/d, h/d and l/d)
study -0.73 (h/d = 0) 0.07 (h/d = 0)
Suburban
-0.88 (h/d = 0.29) -0.10 (h/d = 0.29)
x From Table 2, for T = 0q, the edge location of the roof is subjected to high suction,
as observed in many codes of practice and the values of Cpe for model M3 are
relatively higher than that for model M1 under both terrain conditions, is due to the
presence of side wall.

Summary and Conclusions

Wind tunnel pressure measurement studies on models of structure with curved roof
with and without sidewalls have been carried out for open and suburban terrain condition. The
tested models had sidewall to span ratios of 0.15 & 0.29 and 0.0, respectively. The effect on
pressures exerted over curved roof with and without sidewalls have been investigated from
the evaluated distributions of mean and standard deviation of pressure coefficients at edge and
mid locations of the roof for typical angles of wind incidence i.e., wind blowing parallel,
diagonal and perpendicular to the axis of the arch and are reported in this paper. Based on the
study, it is observed that the suction values of mean pressure coefficients obtained on the roof
of model with sidewall to span ratio of 0.29 are observed to be higher than those values
obtained for models with sidewall to span ratios of 0.0 and 0.15, which indicates the pressures
were affected due to presence of sidewalls. A comparison between evaluated values of mean
pressure coefficients and the values given in various codes of practice have been made and
found the evaluated values of mean pressure coefficients at windward quarter and centre half
showed reasonable agreement for wind blowing perpendicular to the axis of the arch and at
edge location for wind blowing parallel to the axis of the arch.

793
 Proc. of the 8th Asia-Pacific Conference on Wind Engineering (APCWE-VIII)

Further, the distributions of mean and standard deviation of pressure coefficients for
diagonal angle of wind incidence are found to be unique when compared with distributions of
pressure coefficients for angles of wind incidence parallel and perpendicular to the axis and
hence, in addition to above angles of wind incidence, pressure coefficients corresponding to
diagonal angle of wind incidence also need to be considered for design of industrial buildings
with curved roofs in evaluating and verifying against worst load effects.

Acknowledgements

The support rendered by the staff of Wind Engineering Laboratory is gratefully

acknowledged. This paper is being published with kind permission of Director, CSIR-
Structural Engineering Research Centre, Chennai.

References
AIJ (2004), AIJ recommendations for loads on buildings, Architectural Institute of Japan, Japan.
Abraham, A., Nagesh R. Iyer, Muthumani, K., Samuel Knight, G.M., and Lakshmanan, N. (2013), “Wind tunnel
pressure measurement studies on a model of curved roof: Influence of wind angles on mean suction
pressures,” National Conference on Wind Tunnel Testing, Thiruvananthapuram, Kerala, August 22-23, 2013.
AS/NZS: 1170.2: 2002 (2005), Australian/New Zealand Standard, Structural design actions, Part 2: Wind
actions, Standards Australia & Standards New Zealand.

ASCE/SEI 7-05 (2006), Minimum design loads for buildings and other structures, American Society of Civil
Engineers Standard.
Blackmore, P. A., and Tsokri, E. (2006), “Wind loads on curved roofs”, Journal of Wind Engineering and
Industrial Aerodynamics, 94, 833-844.
Ding, Z., and Tamura, Y. (2013), “Contributions of wind-induced overall and local behaviors for internal forces
in cladding support components of large-span roof structure,” Journal of Wind Engineering and Industrial
Aerodynamics, 115, 162-172.
EN1991-1-4 (2004), Eurocode 1: Actions on Structures-Part 1-4: General actions-Wind actions.
GBJ 9-87 (1994), National Standard of the People’s Republic of China, Load code for the design of building
structures, New World Press, Beijing.
Ginger, J. D. (2004), “Fluctuating wind loads across gable-end buildings with planar and curved roofs”, Wind
and Structures, 6 & 7, 359-372.
IS: 875 (Part 3)-1987 (1989), Indian Standard: Code of Practice for Design Loads (Other than Earthquake) for
Buildings and Structures, Part 3: Wind Loads, Bureau of Indian Standards, New Delhi, India.
Kasperski, M. (2008), “Design wind loads for arched roofs,” Journal of Wind and Engineering, 5, 17-30.
Li, Y.Q., Tamura, Y., Yoshida, A., Katsumura, A., and Cho, K. (2006), “Wind loading and its effects on single-
layer reticulated cylindrical shells,” Journal of Wind Engineering and Industrial Aerodynamics, 94, 949-973.
Natalini, M. B., Morel, C., and Natalini, B. (2013), “Mean loads on vaulted canopy roofs,” Journal of Wind
Engineering and Industrial Aerodynamics, 119, 102-113.
User’s Guide-NBC: 1995 (2002), Structural Commentaries (Part 4), NRC-CNRC, Canadian Commission on
Building and Fire Codes, National Research Council of Canada.
Yang, Q., Chen, B., Wu, Y., and Tamura, Y. (2013), “Wind-induced response and equivalent static wind load of
long-span roof structures by combined ritz-proper orthogonal decomposition method,” Journal of Structural
Engineering, 139, 997-1008.

794