Você está na página 1de 6

Fluctuating wind loads for cladding design

David J. Henderson1 and John D. Ginger1


1
Cyclone Testing Station
School of Engineering and Physical Sciences
James Cook University, 4811, Australia
david.henderson@jcu.edu.au

ABSTRACT
Peak cladding pressures derived from the corner and general roof areas for different
building geometries representing typical low rise buildings have been determined using
wind tunnel time series data. These pressures are significantly larger than design
pressures derived from AS/NZS1170.2:2002.

Moving point load testing has demonstrated the elongated tributary area of typical
pierced fixed cladding types. The influence coefficients developed from the point load
testing were used in conjunction with an analysis of time series data from a high density
tapped wind tunnel model to reveal the relationship of the conical vortices and the
elongated tributary area for the pierced fixed cladding. The tributary area and
appropriate relationship with the relevant pressure taps were used to show the validity
of generating simulated cyclonic wind pressure traces using wind tunnel pressure time
series data. From analyzing the generated traces it has been determined that the
Australian wind loading design standard’s pressures are un-conservative for cladding
design for a range of building configurations.

INTRODUCTION
Windward roof edges are subjected to intense fluctuating external pressures during
windstorms. Loads on these areas can be greatly increased when combined with large
positive internal pressures resulting from a breached windward wall, giving a large net
uplift load. Thus the roof envelope and fixings generally experiences the highest wind
loads of the building’s components, and are the components most susceptible to failure.
Industrial sheds typically have open interior floor plans that are subjected to the “same”
internal pressure. The application of internal pressures on roof cladding in a house
would be dependent on location of openings (man holes, vents) in ceiling and eaves
linings in relation to openings in walls associated with the division of rooms and wall
openings (e.g., Kopp et al. 2008).

For this paper, the building configuration used, is shown in Fig. 1. The building
dimensions relate to a scaled wind tunnel model test contained within the United States’
National Institute of Standards and Technology (NIST) building data base, as described
by Ho et al. (2005). The model building had equivalent full-scale plan dimensions of
38.1 m × 24.4 m, a gable end roof of slope 1:12 (4.8°) and an eaves height of 4.9 m and
was tested in the Boundary Layer Wind Tunnel II at the University of Western Ontario
(UWO). The building is subsequently referred to in the text as Building 5-5 representing
5 m wall height with a 5° roof slope. Areas A, B, C and D represent cladding fastener
tributary areas of 0.152 m × 0.9 m on the roof and are shown in Fig. 1, along with wall
point pressure locations WL and WG.
D. Henderson, J. Ginger

The building is similar in size to the average of 26 sheds surveyed by Leitch et al.
(2007). The study showed that the roof cladding design in six of the 26 sheds did not
account for possible large internal pressure as per design standard requirements or used
an incorrect wind load modifier such as for terrain or local pressure factor.

38.1 m

4.9 m 0.15m
θ = 0o WL
WG
R2
1.8 m R1
1.0m
24.4 m

Fig. 1: Diagram of Building 5-5 including location of pressure taps

WIND LOADS FROM AS/NZS 1170.2


Wind loading codes and standards such as AS/NZS1170.2:2002 (Standards-Australia
2002), specify ultimate limit state design wind speeds based on return periods of 500 to
1000 yrs for design of typical buildings. In Region C, this design gust wind speed (V500)
is 69.3 m/s at a 10m elevation in open country terrain for a typical low-rise building.

Design loads of cladding and fixings on buildings are typically calculated from
pressures derived using nominal shape factors or pressure coefficients, provided in
AS/NZS 1170.2:2002. The peak (ppeak) external pressures are calculated from Equation
(1), where ρ is the density of air, and Cfig is the aerodynamic shape factor. For external
and internal pressures, the aerodynamic shape factors are Cfig = Cp,e(Ka × Kc × Kl × Kp)
and Cfig = Cp,i × Kc, respectively. The relevant external and internal pressure
coefficients, Cp,e, and Cp,i, are obtained from Section 5, and Appendix C in AS/NZS
1170.2:2002, and Ka, Kc, Kl and Kp are factors for area-averaging, load combination,
local-pressure effects, and cladding permeability. The dynamic response factor Cdyn is
taken equal to 1.0 for these types of (i.e., “static”) structures, and Vh is the 3sec peak
design gust wind speed at mid-roof height.
ppeak = 0.5 × ρ × Vh2 × Cfig × Cdyn (1)

AS/NZS1170.2:2002 gives an external pressure coefficient Cp,e of -0.9 for edge regions
of a low pitch roof, and internal pressure coefficients Cp,i of 0 and 0.7 for the nominally
sealed and dominant windward wall opening cases. A local pressure factor Kl of 2.0
applies for the design of cladding and fixings, located in an area < a2/4 and within a
distance a/2 from the windward roof edge. The dimension ‘a’ is the minimum of 0.2 ×

2
D. Henderson, J. Ginger

building depth, 0.2 × breadth or its height. A Kl of 3.0 has been proposed to apply in an
area < a2/4 only in the roof corners bounded by distance a, for low pitch roofs in the
revised draft standard of AS/NZS1170.2.

For Building 5-5, shown in Fig. 1, with a regional ultimate limit states (ULS) design
wind speed of 69.3 m/s situated in open terrain, design external pressures given in Table
1, were determined using AS/NZS1170.2:2002.
Table 1: Design external pressures for Building 5-5 using AS/NZS1170.2
Cp,e Kl Design Pressure (external)
Area A -0.9 2 -4.68 kPa
Area B -0.9 2 -4.68 kPa
Area C -0.3 1 -0.78 kPa
Area A -0.9 3 -7.02 kPa (draft standard)
Wall +0.7 1 +1.82 kPa

WIND LOADS FROM WIND TUNNEL DATA


Pressure time series measurements from the NIST database have been used to derive
pressure traces for roof cladding areas for the analysis of fatigue wind loading on
cladding. The building configuration, Building 5-5, shown in Fig. 1, was tested at a
length scale of 1/100 in an atmospheric boundary layer. External point pressure
measurement locations (taps), for example R1 and R2, were combined to give area
averaged pressures representative of the load on the cladding fastener, tributary area A.
Two taps were also used for the area B with one tap located approximately 1 m from the
ridge used for areas C and D. Taps WL and WG were used for the wall pressures.

The building 5-5 time series pressure data for the taps were analysed for approach winds
over a 180° range in 5° intervals. Measured mean and peak pressures are defined in
terms of a pressure coefficient: C p = p ( 12 ρU 2 ) ; where p is the pressure and 12 ρU 2 is
the mean dynamic pressure at mid-roof height. Pressure acting towards the surface is
defined as positive.

Area averaged minimum peak and mean external pressure coefficients acting on areas
A, B and C, and maximum peak and mean for the wall, WL, for wind directions θ = 0°
to 180° are shown in Fig. 2. The large suction pressures were generated on area A (taps
adjacent to the gable end) by the formation of conical vortices for θ = 30° to 60°. The
peak suction pressures acting on Area B, from the wind normal to the long wall (~90°)
were approximately 60% of the peaks acting on A from the quartering wind. Positive
peak pressure coefficients greater than 2.0 were experienced on the wall area WL, for θ
= 45° to 110°.

Comparison with AS/NZS 1170.2


The NIST wind tunnel time series database contained data on building models for both
open and suburban exposures. The exposures were defined as having roughness length
of 0.03 m and 0.3 m, respectively. Gust factors (Û3sec/Ū) of 1.4 and 1.7 were used for
open and suburban terrain, respectively. For example, the suburban building wind
tunnel data trace was derived by modifying the 70 m/s design wind speed at 10 m high

3
D. Henderson, J. Ginger

by a 0.89 Mz,cat from AS1170.2:2002, resulting in the design wind speed of 62 m/s.
Applying the 1.7 gust factor gave a mean speed of 36 m/s at 10 m height.

Fig. 2: Peak and mean Cp for Areas A, B and C and Wall WL for Building 5-5

Table 2 gives a summary of the peak external and net pressures derived from the time
series data for Building 5-5 in both open and suburban terrain. The table also provides
the ratios of the wind tunnel derived pressures divided by the AS1170.2:2002 design
pressures, including the Kl of 3.

Table 2: Building 5-5 external and net pressures for open and suburban terrain
Open terrain Suburban terrain
Area Pressure Ratio of Wind tunnel Pressure Ratio of Wind tunnel
(kPa) over Design pressure (kPa) over Design pressure
A ext -7.70 1.65 (1.10 if Kl of 3) -5.95 1.75 (1.16 if Kl of 3)
A net -8.61 1.35 (0.97 if Kl of 3) -6.49 1.38 (1.01 if Kl of 3)
B ext -5.34 1.14 -3.31 0.97
B net -6.50 1.00 -4.43 0.94
C ext -1.53 1.96 -1.14 1.99
C net -3.96 1.52 -3.36 1.78
Wall +3.35 1.47 +3.03 1.83

The 5° pitch roof had the largest magnitude suctions at roof area A. The peak pressures
from the open terrain were greater than those from the suburban terrain, as would be
expected. However, AS1170.2:2002 typically underestimated these peak cladding
pressures. For the highly loaded gable end roof area A, the design pressures
underestimated the wind tunnel results by more than 60%. The inclusion of a Kl of 3 in
the proposed draft wind loading standard reduces the underestimation to approximately
15%. The Kl of 3 is proposed for only the corner regions of low pitch roofs. Henderson

4
D. Henderson, J. Ginger

(2010) showed that AS1170.2:2002 also underestimated the cladding pressures for a
medium pitch (14°) gable roof with plan dimensions as per Fig. 1.

TRIBUTARY AREA FOR A CLADDING SCREW


Influence coefficients for fastener reactions were determined from a series of point load
tests, as described by Henderson (2010), on a triple span cladding specimen. The
influence coefficient was defined as the ratio of fastener reaction at screw divided by the
specific load that has been applied at various locations across the cladding specimen.
The point load testing to derive influence coefficients showed that the pierced fixed
cladding tributary area is elongated along the crest or rib of the cladding with only about
5% of the reaction at a fastener resulting from the load applied in an adjacent screwed
crest.

Wind tunnel data of a low pitch low-rise building, with a very high density of pressure
taps, was used to explore the effects of area averaging pressures over an elongated
rectangle representative of the area of cladding being held down by a screw. The full
scale dimensions of the building were 46 m × 30.5 m × 10 m high, with a roof slope of
0.25 on 12. The model was tested at UWO, with details of the model study described by
Morrison and Kopp (2010).

Fig. 3 shows the contours of minimum pressure at the roof corner for 0°, 30° and 45°
wind angles. The tributary area for the cladding fixing in from the gable end and centred
on the first purlin in from the eaves is denoted by the dark rectangle. The pressure
gradient from the conical vortices is evident for the quartering winds. Area averaging
for 1, 2, 9 or 51 taps, of the pressure for each time step throughout the traces was
analysed, with examples given in Table 3. There was no significant variation from the
point pressure measurement to an area averaged result, highlighting the effect of conical
vortices on the elongated tributary area.
Table 3: Comparison of area averaging method with direction
Dir Number of taps Method Cp % of one tap Time stamp
30° 1 Point -9.44 100 12389
30° 2 Area average -9.49 101 12389
30° 9 Area average -9.65 102 12389
30° 51 Influence 3 rows -9.81 104 12389
45° 1 Point -8.41 100 13405
45° 2 Area average -8.84 105 13406
45° 9 Area average -8.74 104 13405
45° 51 Influence 3 rows -8.68 103 13405

CONCLUSIONS
Wind tunnel test data produce spatially and temporally varying pressures on the roof of
a building. These pressures generate highly fluctuating reactions in the cladding
fasteners. Influence coefficient tests on cladding shows that the fastener reaction is
primarily affected by pressures along its elongated tributary area.

The analysis of wind tunnel test time series for a range of taps over a cladding fastener
tributary area, in conjunction with the point load influence testing showed for pierced

5
D. Henderson, J. Ginger

fixed cladding, the small elongated cladding tributary area could be satisfactorily
represented by one or two pressure taps centred about the screw location within the
elongated area, for both peak pressure and load cycles.

Peak pressures for cladding loads determined from two building configuration’s wind
tunnel studies typically exceeded AS/NZS1170.2:2002.

Fig. 3: Contours of minimum peak pressure for roof corner

REFERENCES
Henderson, D. (2010). "Response of pierced fixed metal roof cladding to fluctuating wind
loads," James Cook University, Townsville, Australia, Submitted April 2010.
Ho, T. C. E., Surry, D., Morish, D., and Kopp, G. A. (2005). "The UWO contribution to the
NIST aerodynamic database for wind loads on low buildings: Part 1. Archiving format
and basic aerodynamic data." J. Wind Eng Ind. Aerodyn, 93, 30.
Kopp, G. A., Oh, J. H., and Inculet, D. R. (2008). "Wind-induced internal pressures in houses."
Journal of Structural Engineering-ASCE, 134(7), 1129-1138.
Leitch, C. J., Ginger, J. D., and Henderson, D. J. (2007). "Vulnerability of metal-clad sheds in
cyclonic regions." 12 Int. Conf. Wind Engineering, Cairns, Australia.
Morrison, M., and Kopp, G. (2010). "Analysis of Wind-Induced Clip Loads on Standing Seam
Metal Roofs." Journal of Structural Engineering, 136(3).
Standards-Australia. (2002). "AS/NZS1170.2:2002 Structural design actions Part 2: Wind
actions." AS/NZS1170.2:2002, Standards Australia, Sydney NSW, Australia.

ACKOWLEDGEMENTS
The authors gratefully acknowledge the support of an Australian Research Council Linkage
Grant and Stramit Building Products as the industry partner. They also gratefully acknowledge
the support provided by the Queensland Government Department of Public Works.

Você também pode gostar