Professional Documents
Culture Documents
2012 Bes Patterson Mehta PDF
2012 Bes Patterson Mehta PDF
AND
UNIVERSITY
OF
TEXAS
AT
ARLINGTON
SYMPOSIUM
ON
PAT T E R S O N
AND
M E H TA
125
ABSTRACT
It is generally assumed that wind causes only uplift pressures on a low-slope roofan
assumption widely reflected in various publications of the roofing industry. This assumption
is not entirely correct because wind can also cause downward pressures on a low-slope roof.
In most situations, the downward wind pressure on low-slope roofs is so small that it
can be ignored, but in some situations the downward wind pressure can be substantial.
This paper highlights situations where downward wind pressures, in combination with rain
load, must be examined by designers and forensic experts for reasons of life safety and the
prevention/reduction of property loss on buildings with low-slope roofs.
SPEAKER
STEPHEN PATTERSON, RRC, PE ROOF TECHNICAL SERVICES, INC.
STEPHEN L. PATTERSON is a licensed engineer and registered roof consultant who specializes in roofing and waterproofing. He has more than 40 years of experience in all
phases of the roofing industry. Patterson founded ROOFTECH in 1983 and has consulted
on more than 30 roof collapses in his career. He coauthored Roof Design and Practice
(Prentice Hall, 2001), Roof Drainage (RCI Foundation, 2003), and Wind Pressures on Low
Sloped Roofs (RCI Foundation, 2005).
OF
TEXAS
AT
ARLINGTON
126
PAT T E R S O N
AND
M E H TA
SYMPOSIUM
ON
ON
PAT T E R S O N
AND
M E H TA
127
Figure 3A
Pressurization of a
building interior in a
partially enclosed
building occurs when the
dominant opening(s) is
(are) located on the
windward wall.
Figure 3B
Depressurization of a
building interior in a
partially enclosed
building occurs when the
dominant opening(s) is
(are) located on the
leeward wall.
Thus, the exterior and interior uplift pressures add to give the resultant uplift pressure on a roof.
The leakage effect is large where a building has dominant opening(s) on its windward wall to produce a ballooning effect
from within (Figure 3A). In such buildings,
referred to as partially enclosed buildings,
the interior pressure is high. Most buildings
do not belong to the partially enclosed classification but to the enclosure classification,
referred to as an enclosed building. In such
buildings, the interior pressure is much
smaller than in a partially enclosed building.
Note that the interior pressures within a
building can cause either uplift pressure or
downward pressure on the roof (Figure 3B).
Both uplift and downward pressures must
be considered.
128
PAT T E R S O N
AND
M E H TA
SYMPOSIUM
ON
2. Downward Wind
Pressures on Low
Slope Roofs
If the downwind
dimension of the
building is relatively
large compared to its
height, the separated
airflow descends and
reattaches to the roof
and continues forward
parallel to the roof
(Figure 4). Apart from
roof height, the length
of the separated flow
region is also a function of the roughness
of the upwind terrain.
A rougher upwind terrain yields a shorter
length of the separated
flow region, increasing
the length of the reattached flow region
because it creates a
more turbulent flow on
the roof.1
Within the reattached flow region, wind
uplift pressures are much smaller than
those in the separated flow region, and this
region also experiences downward pressures. ASCE 7-10 Standard gives downward wind pressure values on low-slope
roofs (slope 7) for buildings with roof
height 60 ft. (Note that the downward
pressures can also be present in the separated-flow region, but their magnitude and
occurrence frequency are relatively small.)2,3
While the downward wind pressures in
zone 1 of a low-slope roof are quite small,
they can be high in the vicinity of a parapet
(Figure BA) or in the vicinity of an obstruction that creates a stepped roof, such as
where a tall block meets a flat-roof podium
(Figure 5B)4 or in multilevel flat roofs.5 ASCE
7-10 Standard requires that downward
pressures be considered in zones 2 and 3 of
a low-slope roof if (1) roof height 60 ft. and
(2) it is provided with a 3-ft.-high or higher
parapet.
Another factor contributing to the
downward pressure on a low-slope roof is
the depressurization of the buildings interior as shown in Figure 3B. Both enclosed
and partially enclosed buildings are subjected to this phenomenon; the latter category experiences a much higher downward
wind pressure.
3. Rain Load
The rain load on a low-slope roof
depends on the depth of the accumulated
rainwater and must be determined for each
individual roof, based on roof geometry,
drainage design, and the design rainfall
intensity for the location.
As per ASCE 7-10 Standard,6 an (interior) roof drain or an (exterior) scupper must
have a certain minimum hydraulic head to the maximum rain load on a typical lowachieve its full (design) drainage capacity. slope roof, we will assume a hydraulic head
For instance, a 4-in.-diameter roof drain of 3.5 in. over the drainage elements. We
requires a hydraulic head of at least 2.5 in., will further assume that the primary
and a 6-in.-diameter drain requires a head drainage system is nonfunctional (blocked),
of at least 3.5 in. to be fully effective. The and the overflow drainage is raised 2 in.
hydraulic head required for scuppers is above the primary drains so that the total
generally larger.
depth of accumulated water over a drain is
In order to obtain a rough estimate of at least 5.5 in. (Figure 6). Because the roof
is sloped and the drains are generally located at the lowest point on the roof, the actual rain load on the roof varies and is generally concentrated over the drains.
In Figure 6, the maximum rain load
occurs immediately over the drains and
equals (5.5 x 5.2) = 28.6 psf, where 5.2
implies that 1-in. depth of water over a
1-sq.-ft. area weighs 5.2 lb. (density of
water = 62.5 pcf). Note, however, that the
rain load of 28.6 psf may be exceeded for
short time intervals because the roof
drainage system is typically designed for a
100-year, 1-hour rainfall intensity. The
rainfall intensity during a shorter period
can exceed the 1-hour design rainfall intensity. For instance, the rainfall intensity for
the 100-year, 15-minute rainfall is generally two times the 100-year, 1-hour rainfall
intensity.7 During these short periods, the
roof is subjected to a higher load.
Figure 6 Depth of rainwater accumulation on a low-slope roof considering that
the primary drainage system is blocked.
SYMPOSIUM
ON
B U I L D I N G E N V E L O P E T E C H N O L O G Y O C T O B E R 2 0 1 2
PAT T E R S O N
AND
M E H TA
129
Figure 8 Detail section at eave region of the warehouse building shown in Figure
7. Average rain load on roof deck in the eave region = 0.5(5.5 + 2.25)5.2 = 20.15
psf 20 psf. Rain load on joist 1 = 4.25(5.2) 5 = 110.5 #/ft.
130
AT T E R S O N A N D
M E H T A
SYMPOSIUM
ON
Example 1
(Enclosed building)
Example 2
(Enclosed building)
Example 3
(Partially enclosed building)
Downward design
load on deck
Downward design
load on joist 1
170 #/ft.
265 #/ft.
an increase of 56%
over Example 1
316 #/ft.
an increase of 86%
over Example 1
Table 1 Downward loads for the roof structure of building in Figure 7, obtained fr
om Examples 1, 2, and 3.
two photographs in Figure 9 show the area
of the building that collapsed.
There were several issues involved in
this collapse, including a complicated
drainage system and unusual geometry of
the roof structure in the area of collapse
both of which made the evaluation of the
collapse difficult. Determining the amount
of water accumulation in the corner was difficult, and several models were used to estimate the rain load in the area of collapse.
The rain load occurring as a result of the
water accumulation from these models was
significant but not quite enough to collapse
the structureuntil the additional positive
downward pressure from the wind load was
taken into consideration.
Without question, this was an intense
storm, but this is the type of storm that
commonly occurs in North Texas and in
many other locations around the country.
While this was an intense thunderstorm,
the wind speeds and rainfall rates were
SYMPOSIUM
ON
PAT T E R S O N
AND
M E H TA
131
6. Conclusions
This paper has highlighted the situations where downward wind pressures
must be considered along with the rain load
for the design of a low-slope roof or in forensic investigation of a roof collapse. As discussed, these situations can occur in parapeted roofs, stepped (multilevel) low-slope
roofs, in high-wind locations, and in
enclosed or partially enclosed buildings.
In such situations, the downward wind
pressure on the roof is also accompanied by
the wind forcing the accumulated water
toward the parapet (or the wall adjoining
the roof), which justifies the simple addition
of the two fluid pressures (air and water) on
the roof.
While this paper has focused on rainload and wind-load combinations, similar
considerations should apply to snow-load
and wind-load combinations.
Finally, it should be stated that this
paper has only focused on the downward
wind pressures on roofs. Wind uplift pressures are important; and, as stated in the
introduction, they are often more critical
than downward wind pressures on low-slope
roofs. The ASCE 7 Standard, publications by
the Roof Consultant Institute Foundation,9
and other literature10 provide guidance to
the designers in this respect.
132
PAT T E R S O N
AND
M E H T A
REFERENCES
1. John D. Holmes, Wind Loading of
Structures, Second Edition (2007),
John Wiley and Sons, p. 180.
2. Ibid., p. 226.
3. American Society of Civil Engineers
(ASCE), Minimum Design Loads for
Buildings and Other Structures,
2002, p. 286.
4. American Society of Civil Engineers:
Minimum Design Loads for Buildings
and Other Structures, 2010, p. 336
and p. 339.
5. Jinxin Cao, Akhito Yoshida, Yukio
Tamura, Wind Pressures on MultiLevel Flat Roof of Medium-Rise
Buildings, Journal of Wind Engi
neering and Industrial Aerodynam
ics, Volume 103, April 2012, pp. 115.
6. ASCE 7, 2010, p. 452.
7. S. Patterson and M. Mehta, Roof
Drainage, Roof Consultant Institute
Foundation (RCIF) Monograph No.
02.03, 2003, p. 73.
8. Kishor Mehta and M. Delahay,
Guide to the Use of the Wind Load
Provisions of ASCE 702, ASCE
Press, 2004.
SYMPOSIUM
ON
ACKNOWLEDGMENTS
The authors gratefully acknowledge the
reviews and recommendations provided by
the following experts during the preparation
of this paper:
Arvindam G. Chowdhary, PhD, director,
Wind Engineering Research Laboratory, International Hurricane Research Center, Florida International
University, Miami, L.
Erik L. Nelson, PhD, PE, Nelson Architectural Engineers, Inc., Plano, TX.
Mahendra Raval, PE, Raval Engineering
Granby, CO.