Lessons from Wind Damage Reports

Technical investigation reports
on wind damage to roofi ng and
cladding can provide a wealth
of useful information relevant
to designers, researchers,
manufacturers, contractors,
and building owners. It is common to fi nd
multiple failures within roof systems, and
often the challenge is to identify which link
in the assembly broke fi rst. Forensic investigations
demand a methodical approach,
piecing together the available evidence of
the initial point of failure, which is often
hidden from view.
Several case studies of investigations of
wind-damaged roofs in the UK and Ireland
are presented here, exploring alternative
causation theories and developing options
for strengthening and repair.
WIND ATTACHMENT DESIGN
Over the past 25 years, there have been
three different standards used for calculating
design wind pressures on buildings in
the UK, each with its own factors, assumptions,
and methodology. The changes in
codes have led to confusion within the roofing
and cladding industry and, on occasion,
mistakes and oversight. The changes have
had a negative effect, and it is not unusual
during a wind damage investigation to
fi nd that no party has actually calculated
the design wind pressures on the roof and
checked the attachment strength.
The standard now used throughout
Europe is EN 1991, Part 1.4, which was
fi rst introduced in 1999. This comes in
two parts: the basic standard that is common
across all 29 member countries of
the European Union (EU), and a separate
National Annex, which is specifi c for use in
individual countries. Each member country
has developed its own variations, often
introducing signifi cant differences and the
requirement for local knowledge and experience.
The designer has to use the two documents
together and cross-reference both at
the same time, which adds complexity.
Wind suction forces acting on the upper
weathering skin of a roof are transferred
from layer to layer down through the roof
construction into the structural framework.
Each fastener transfers the upward load
to the next layer down. A useful analogy is
that of a chain anchoring the upper weathering
skin down to the support structure.
If one link in the chain were to break, then
potentially the outer sheets could become
detached from the roof (Figure 1).
Calculations should be prepared to estimate
the design wind suction load acting on
each link in the chain, and then compared
with the characteristic strength of the fastener
to determine their factors of safety
using the equation in Equation 1.
In the UK, the minimum acceptable factors
of safety used for checking the attach-
S e p t e m b e r 2 0 1 6 I n t e r f a c e • 2 5
Figure 1 – The chain analogy for the attachment strength of a multi-layer roof system.
where:
w = design wind pressure
f = characteristic strength
γ = factor of safety
Equation 1.
w < fγ
Editor’s Note: This is a shortened version of a paper presented at the
31st RCI International Convention and Trade Show in March 2016 in Orlando, FL.
ment strength of profiled metal and singleply
membrane roof fasteners are 2.0 for
pullout from steel or aluminum, 3.0 for pullout
from timber, and 4.0 for pullout from
masonry/concrete. In the UK, a permissible
strength approach is currently used in metal
roofing standards. This is likely to change
in the near future to a limit state approach
using partial factors of safety.
CASE STUDY A: STANDING SEAM ROOF
On New Year’s Day, 2005, a strong
gale blew across southern Ireland from a
southwesterly direction, resulting in an
extensive area of lightweight aluminum
standing-seam roofing becoming detached
from the windward verge of an aquatic center
and adjacent gym (Figure 2) and causing
consequential impact damage to roof cladding
and skylights downwind. The structures
were evacuated safely without injury
to members of the public or staff.
Roberts Consulting was contacted to
investigate. Instructions were to examine
the evidence relating to the wind damage
and to identify causation. The instructions
were received three months after the wind
event, such that on arrival on site, much of
the original roof construction and original
damage had been disturbed. Consequently,
the color photographs taken immediately
after the storm became a vital record.
The site inspection confirmed the as-built
arrangement of the roof assembly and the
details of the fasteners used.
The wind damage photographs showed
that there had been a detachment between
the steel top hat rail and the aluminum
saddle at 900-mm (3-ft.) centers (Figures 3
2 6 • I n t e r f a c e S e p t e m b e r 2 0 1 6
Figure 2 – Detachment of aluminum standing seam roofing from windward verge of the
Competition Pool.
Figure 3 – Close-up of wind
damage zone showing
detached top hat rails and
distorted saddles.
Figure 4 – Competition Pool roof assembly.
Location Dublin, Ireland
Building use Aquatic center
Altitude, site exposure, topography +74 m (240 ft.), severe, rural
Roof area 5,700 m2 (61,000 sf)
Roof slope Barrel vault 62 m wide, slope up to 20°
Roof type Aluminum standing seam
Roof sub structure Halters fixed to top hat rails, to saddles,
through liner into purlins
Basic wind speed (hourly mean) 23 m/s (52 mph)
Recorded peak speed (hourly mean) 14 m/s (31 mph)
Design wind suction pressure in -2.0 kN/m2 ↑ (42 psf)
damage zone
Extent of detachment 300 m2 (3,200 sf) of standing seam
60 m (200 ft.) of parapet capping
Estimated financial loss £10 million
Table 1 – Basic data for Case A.
and 4). In addition, there
was upward distortion in
the wide top plate of the
saddle. Calculations were
prepared to check the
strength at each of the
connections, and the conclusions
are summarized
in Table 2.
The calculations found
that the local bending
stress in the crown of the
aluminum saddle was
excessive, and that the factor
of safety against the
top hat rail to saddle fi xing
pulling out was 1.1,
signifi cantly less than the
recommended minimum of
2.0. This is the same weakest
link as observed in the
photos of wind damage.
The adjacent verge
cappings also became detached
in the storm, and their means of
attachment was closely examined. The aluminium
capping had been held in place with
rivets that had pulled through. The spacing
of the rivet holes through the supporting
cladding rail was measured, and the distances
were found to be greater than the
recommended 450 mm (18 in.). This was a
further weakness in the roof assembly.
S e p t e m b e r 2 0 1 6 I n t e r f a c e • 2 7
Element Material Design Area Fastener Load / Fastener Factor Sati sfactory?
wind loaded number fastener strength of
pressure m2 (sf) kN (lbf) kN (lbf) safety
kN/m2 (psf)
Standing 0.9-mm
seam Aluminium -2.0 ↑ – – – -5.0 kN/m2 2.5 􀁎
(42) (105 psf) Ordinance
Halter Aluminum
-2.0 ↑ 0.64 2 0.64 3.35 5.2 􀁎
(42) (6.9) (145) (753)
Top hat 1.25-mm
rail Galvanized
Steel -2.0 ↑ 1.89 2 1.89 2.1 1.1 X
(42) (20) (420) (472)
Saddle 2.0-mm
Aluminium -2.0 ↑ 1.89 4 0.95 18 19 􀁎
(42) (20) (210) (4047)
Purlin 10-mm
Steel Flange
Table 2 – Summary of factors of safety for the Competition Pool roof assembly.
American Hydrotech, Inc.
303 East Ohio | Chicago, IL 60611 | 800.877.6125 | www.hydrotechusa.com
© 2016 Garden Roof is a registered trademark of American Hydrotech, Inc.
American Hydrotech introduces the Garden Roof® Planning
Guide mobile app, a first-of-its-kind digital brochure that
helps design professionals take a vegetated roof from initial
concept to completion.
Packed with photography, technical information and videos,
design professionals can explore assembly options and
components, growing media and vegetation, and learn
about topics such as design considerations, economic
and sustainable benefits, installation and maintenance,
and much more.
20 years of vegetated roof experience…
brought to life in one app.
Download your copy today at
hydrotechusa.com/GRPG
7.50W X 5H.indd 1 1/12/16 5:06 PM
The original roofi ng contractor undertook
to replace the area of detached roofi ng,
both to make good the downwind isolated
impact damage, and to strengthen the
top hat rail to saddle and purlin connection
by installing additional long screw
fi xings through the top hat rail directly into
the steel purlin below. This repair scheme
had a number of disadvantages, including
puncturing the vapor control layer. Those
advising the building owners considered the
future condensation risks to be acceptable.
CASE STUDY B: SINGLE PLY
MEMBRANE
On St. Jude’s Day—September 28,
2013—a fast-moving, vigorous Atlantic
depression brought very strong winds and
heavy rain to southeast England, with winds
gusting up to 36 m/s (80 mph). One modern
building that suffered wind damage was
a hotel in Chelmsford, to the northeast of
London. Lengths of roof edging and singleply
membrane roofi ng became detached
from the western side of the second fl oor
roof and peeled back, resulting in
debris falling to ground level (Figure
5). This led to the closure of the public
highway immediately to the east
of the building for a period of several
days (Table 3).
An independent inspection
of the roof was commissioned two
months after the storm to examine
the evidence of damage, to identify
causation, and to advise on remedial
work to remaining roofs. The roof
system comprised a TPO single-ply
membrane that was adhered to a
tissue-faced mineral fi ber insulation
board, which in turn was screw fi xed
through the vapor control layer into
a galvanized steel deck.
At the time of the inspection,
temporary remedial works had been
completed to enable the hotel to
reopen. Much of the debris had been
removed from site by the repair contractor
and, fortunately, had been
kept for examination in his local yard. The
samples were closely examined to reveal an
inadequate thickness and spacing of the
adhesive bonding. Samples of the mineral
wool insulation showed that the top tissue
facing readily detached from the fi brous
core and was not the specifi ed insulation
board with a “single-ply adhered facing.”
The wrong product had been supplied,
which had not been identifi ed by the roofi ng
contractor, the supplier, or the other surveyors
initially inspecting the wind damage.
2 8 • I n t e r f a c e S e p t e m b e r 2 0 1 6
Figure 5 – Second-fl oor roof membrane rolled back from western edge.
ISSUE SUBJECT SUBMISSION DEADLINE
December 2016 Metal walls and roofs September 15, 2016
January 2017 Miscellaneous (bldg. envel.) October 14, 2016
February 2017 Codes and standards November 15, 2016
March 2017 Trends in wall & roof design December 15, 2016
April 2017 Air and vapor control January 13, 2017
May/June 2017 Conventi on review February 15, 2017
Publish in Interface
Interface journal is seeking submissions for the following issues. Optimum article size is 2000
to 3000 words, containing fi ve to ten graphics. Articles may serve commercial interests but
should not promote specifi c products. Articles on subjects that do not fi t any given theme may be
submitted at any time.
Submit articles or questions to Executive Editor Kristen Ammerman at 800-828-1902
or kammerman@rci-online.org.
The wind damage photos also showed
that the roof edge had become detached.
There were no record drawings of the initial
construction, and so the as-built assembly
was drawn up based on site measurements,
the opening-up examination, and
inspection of the debris in the contractor’s
yard. A short parapet had been constructed
using two channel sections with an internal
stud and no fi xings joining the channels
together. Under wind uplift pressure, the
capping and support could lift upwards.
From this, it was possible to determine the
probable sequence of detachment along the
western edge of the central roof in which a
Z-section fl ashing rotated upwards, increasing
the wind uplift pressure acting on the
underside of the fl ashing and entering the
upstand detail (Figures 6 and 7). This, in
turn, caused the perimeter channel and
studs to lift upwards, causing the single-ply
membrane to peel back readily with the lack
of adhesive restraint.
Within a month of the initial inspection,
the second-fl oor roof was fully replaced with
a new mechanically attached PVC single-ply
membrane system. The third- and fourthfl
oor roofs were investigated and found to
be of a similar construction, with areas of
debonded single-ply membrane adjacent to
the western edge. It was recommended that
a new mechanically fi xed PVC single-ply
membrane should be overlaid over the existing
membrane, with a new secure perimeter
detail developed.
There were delays in carrying out this
work, and over a six-month period, the
extent of the delaminated zone increased,
ultimately resulting in extensive ruckling
or wrinkling. This evidence of further progressive
damage persuaded the parties to
mechanically fasten and overlay the thirdand
fourth-fl oor roofs, with work satisfactorily
completed in the summer of 2014.
LESSONS LEARNED FROM WIND
DAMAGE INVESTIGATIONS
Safety Comes First
Our fi rst responsibility is to the safety of
ourselves and of those around us, including
members of the public. Roof consultants
inspecting wind-damaged roofs should be
experienced at working at heights and
in wearing appropriate personal protective
equipment. Particular care is required when
working close to unprotected roof edges,
often requiring the provision of a mobile
platform or fi xed scaffolding.
Immediately after a storm in which elements
of a roof have come loose, a cordon
should usually be set up at ground level
to keep people away from high-risk areas
where further pieces could fall and cause
injury. This is particularly important in
urban areas and around assembly buildings
where people gather, such as schools
and hospitals. Usually, these precautionary
measures have been undertaken in advance
of the arrival of the roof consultant.
Adopt a Scientifi c Approach
It is important to inspect the wind damage
as soon after the incident as possible,
before the areas are modifi ed, such as by
the removal of loose elements or local emergency
repairs. This often requires travel
W. R. MEADOWS
Waterproofing Solutions
W . R . M E A D O W S
H Y D R A L A S T I C 8 3 6
is a cold-applied, solvent-free,
single-component waterproofing
compound for horizontal surfaces,
s u c h a s p o d i um d e c k s a n d
planters. HYDRALASTIC 836
cures to a tough, flexible, elastic
membrane that i s appl i cable
for use on concrete sur faces,
including green concrete. The
product has a low VOC content
and meets all North American
VOC regulations. If your need
i s a f lexible, VOC- compl iant ,
wa t e r p r o o f i n g memb r a n e ,
choose HYDRALASTIC 836
from W. R. MEADOWS.
H8YDR3ALA6STIC
© W. R. MEADOWS, INC. 2016
For a solution based on your needs, visit
wrmeadows.com or call 1-800-342-5976.
WR_Meadows_Hydralastic_RCI_7-26_Layout 1 7/26/1
S e p t e m b e r 2 0 1 6 I n t e r f a c e • 3 1
Locati on Chelmsford, England
Building use Hotel
Alti tude, site exposure, topography +25 m (82 ft .), sheltered, urban
Roof area 2,000 m2 (21,500 sf)
Roof slope Flat
Roof type Single-ply membrane, adhered
Roof sub structure Mineral wool thermal insulati on, screwfi
xed through vapor control layer into
steel deck
Basic wind speed (hourly mean) 22 m/s (49 mph)
Recorded peak speed (hourly mean) 18 m/s (42 mph)
Design wind sucti on pressure in -1.8 kN/m2 ↑ (38 psf)
damage zone
Extent of detachment 200 m2 (2,150 sf) of single-ply membrane
roofi ng and roof edge
Esti mated fi nancial loss £1 million
Table 3 – Basic data for Case B.
at short notice, and the consultant
should be prepared.
On first sight, the pattern of
damage is often confusing, with multiple
failures and impact damage
caused by flying debris. The challenge
is to identify the first link
in the roof assembly that failed.
A methodical approach should be
adopted, following the good practice
set out in the “Guide to Surveys
and Inspections of Buildings” published
in the UK by the Institution
of Structural Engineers. Practical
difficulties on site often mean that it
is not always possible to gain access
to all roof areas on the same day, and
an adaptable approach is required to gather
as much information in what is often only a
short period of time at roof level.
Careful records of observations and
sketches with site measurements are
important, examining both the zone of
detachment and any loose debris. Color
photographs that are dated should include
general views of the wind-damaged building,
the surrounding topography as seen from
roof level, and close-ups of the roof assembly—
both damaged and undamaged. Where
there is some movement in the roof, video
recording can be of assistance. Collected
samples of damaged elements and fasteners
that can be removed should be properly
labelled. They can be of great assistance for
close examination and reference at the time
of report writing, and later for presenting
in meetings to assist in explaining likely
modes of failure.
Eyewitness Information Can Be Useful
Reliable eyewitness evidence can be
invaluable in determining the probable
sequence of events. This should include informal
discussions with security staff, maintenance
workers, neighbors, and roof repair
contractors, keeping a note of the names
of the eyewitnesses. Closed-circuit security
video records may also be of assistance.
Examine Weather Data
The meteorological data from the nearest
recording station should include the
hourly wind speeds and directions for the
period leading up to and including the time
of the reported damage. The wind speeds
should include the hourly mean, as well as
the maximum ten-minute and three-second
gust wind speeds. Occasionally, the air temperature
and rainfall data are also of assistance
in assessing the strength of the components;
for example, some artificial slates
absorb water and lose flexural strength.
Commission Testing
Basic static screw and nail pullout
strengths can be determined using screw
jack equipment, either on site or on a lab
bench. A first estimate of the degree of
movement can be measured by pulling a
flexible sheet by hand and measuring the
applied load with a spring balance. Other
3 2 • I n t e r f a c e S e p t e m b e r 2 0 1 6
Figure 6 – Examination of
upstand discovered top
channel not fixed to
internal stud.
Figure 7 – Probable
sequence of detachment
of western half of
second-floor roof.
test methods include the FM Global wind
uplift field test, and laboratory tests of fullscale
roof assemblies using compressed air
bags to apply an upward loading that allows
for changes in the shape of the roof cladding.
Prepare Calculations
To check the structural adequacy of
mechanically fixed roof systems, calculations
should be prepared to estimate the applied
wind loads on the different links in the chain,
transferring the upward wind loads down
into the supporting structure and to assess
their factors of safety. The summary given in
Table 2 of the first case study is a good way
to highlight the weakest link.
Within bonded multi-layer roof systems,
the wind pressure acting on each layer is
not the same. The layer with the greatest
upward pressure, or the critical layer, is
the first air-impermeable barrier from the
underside of the roof. The critical layer
could be a concrete roof deck, a vapor control
layer laid over perforated metal decking,
or a waterproofing layer above an unsealed
metal deck and fibrous insulation. This
should be considered by the roof consultant
in assessing the applied wind loading on the
roof system.
Undertake Desk Study
Any project drawings or specifications
should be examined, together with copies of
original manufacturers’ literature relating
to the claimed performance of the roofing
system. This, in turn, requires access to
a reference library with historical product
literature. Ordinance survey maps are particularly
useful in the UK for identifying
the topography and exposure of the land
around the site, which may give rise to
unusual wind features. Enquiries into similar
damage reports elsewhere—both nationally
and internationally—may provide useful
background information.
Prepare Report
The report should bring together the
information collected and present the facts
and discussion in a logical order with
concise conclusions and clear recommendations.
The use of photographs and line
drawings is particularly helpful to the lay
reader in illustrating where the failures
probably occurred. Recognizing the commercially
sensitive nature of some wind
damage investigations, there is a need for
matters to be dealt with on a strictly confidential
basis.
S e p t e m b e r 2 0 1 6 I n t e r f a c e • 3 3
You. Us. The project.
800 255 4255 PROSOCO.COM
When instructed, the options for repairing
or replacing the wind-damaged roofs
should be identifi ed, summarizing the
advantages and disadvantages of each.
From experience, there are occasions when
the roof consultant cannot be confi dent
of the precise reasons why the roof failed.
On such occasions, there is a need for
further examination, testing, and analysis
to improve the understanding of the performance
of the roof when subjected to strong
winds.
CONCLUSIONS
A methodical approach should be adopted
in gathering and recording the evidence
of wind damage, in undertaking the
desk study, and then producing the factual
report. It is recognized that these time-consuming
tasks often need to be completed
promptly to enable repairs to be started and
the building brought back into use.
There is a need to learn from experience
and to avoid repeating mistakes. This
recurred in Ireland in February 2014, when
the roof of another aquatic center in County
Wexford blew off, suffering the same mode
of failure as the aquatic center in Dublin a
decade before. The lessons from previous
investigations hadn’t been shared within
the roofi ng community.
It is hoped that by sharing feedback
from wind damage reports in an independent
and constructive way, we can
improve our common understanding of failure
mechanisms, enabling us to design and
build more reliable building envelopes that
are better able to withstand the extremes
of changing weather patterns in whatever
country we practice.
REFERENCES
A. Baskaran and T.L. Smith. 2005.
“Guide for the Wind Design of
Mechanically Attached Flexible
Membrane Roofs.” SIGDERS/
National Research Council of
Canada. Ottawa.
BS EN 1991-1-4: 2004. “Eurocode 1:
Actions on Structures – Part 1-4:
General Actions- Wind Actions” and
National Annex to BS EN. 1991-1-4.
January 2011. BSI. London.
CIB Publication 405: 2015. “Improving
Roof Reliability: Final Report of the
Reliable Roofi ng Task Group.” CIB.
Amsterdam.
K. Roberts. 1996. “Short-Term Roof
Inspections: Guidance Notes for the
Adoption of a Safe System of Work.”
RCI Interface. RCI, Inc. Raleigh,
North Carolina.
K. Roberts. 2005. “Wind Damage to
Lightweight Roofing Systems.”
Forensic Engineering: Diagnosing
Failures and Solving Problems.
Institution of Civil Engineers.
London.
T.L. Smith. 2013. “Evaluating Wind-
Damaged Low-Slope Roof Assemblies:
A Preliminary Protocol.” 12th
Americas Conference on Wind
Engineering. Seattle, Washington.
3 4 • I n t e r f a c e S e p t e m b e r 2 0 1 6
Keith Roberts is
principal of Roberts
Consulting, an
independent fi rm
of consulting engineers.
He is a
chartered civil and
structural engineer
who has investigated
wind damage
to more than
50 roofs in the UK
and Ireland over
the past 25 years, identifying the causes of
the failure and working with the parties to
agree on practical forms of repair or replacement.
He is chairman of the CIB International
Roofi ng Committee that examined roof reliability
and is a member of RCI.
Keith Roberts
Looking to stand out from the estimated
2000 other roofi ng companies scrambling to get
business in Colorado, one company has started
a promotion called “Get a Roof, Get a Gun.”
Every completed roof replacement and exterior
repair from a hail or windstorm receives an
AR-15 rifl e.
WeatherProof Roofi ng & Exteriors’ website
says the company is offering the promotion
through October 1 “because it symbolizes, at
its core, who we are: True American Patriots…
for patriots like us, its true representation is
the high cost of freedom we enjoy…a well-armed
population is the best defense for protecting your life, liberty, and happiness.”
The company is partnering with a local gun club to run the appropriate background checks. In lieu of the AR-15, customers
may choose a handgun or a $500 gift certifi cate to Cabela’s.
Colorado State Rep. Rhonda Fields said, “It is very troublesome to me that someone would want to increase business by
offering an incentive to get a new roof by getting a gun.”
— Fox31 Denver
“Get a Roof, Get a Gun” Promo Offered by Roofi ng Contractor