Sample report of
a structural investigation - Metal
Building Roof LeaksIntroduction:
At the request of the owner, Thomas R.
Price P.E., of Structural Solutions LLC,
a licensed professional engineer,
inspected the Facility in the attempt to
determine the cause of weather-tightness
problems associated with the building
roof.
This
report is a follow-up to the inspection.
The inspection took place between noon
and 3:00 PM.
Weather was clear and sunny, with a
temperature of approximately 85 to 90 deg
F.
Present on the site was the owner, a
representative of the contractor, and Mr.
Price.
The building in question is a metal
building manufactured by a major metal
building manufacturer. It is 440’
wide and 400’ long. It has rigid
frames at 40’ spacings, with steel
bar joists at 5’ o.c. The roof is a
clip attached
vertical rib standing seam roof system,
with 18” wide panels. The ridge of
the building is at the center of the
440’ width. The roof panels are made
non-continuous by an approximate 6”
vertical step at 110’ from each
eave. (See Figure 1). Each of the four
110’ panel runs is fixed by screws
at its low end. The upper end of the
panel run is free, presumably to allow
for thermal movement due to ambient
temperature variations. The roof slope is
1/4” on 12”.
History:
The roof was installed, and complaints of
roof leaks began shortly thereafter.
Structural Solutions was shown various
correspondence between the owner, the
pre-enginnered metal building
manufacturer, the contractor, and others,
in regard to attempts to repair the
problem. It was immediately assumed that
the roof expansion joint was the source
of the leaks. We concur with this
conclusion.
Various repairs have been attempted,
however none have sealed the leaks
satisfactorily. The major repair attempts
involved the application of a sprayed on
rubber sealant covering the entire
expansion joint, and the installation of
a second metal flashing, (See Figure 2).
Inspection:
Prior to inspecting the site, we viewed a
videotape of a leak test where water was
being sprayed on the expansion joint. A
hole was made in the lower panel cap, and
water flowed freely through it. This was
at a point on the interior side of the
closure which presumably should be a
water tight area. From the video, it was
evident that water was flowing through at
least one other location as well. (This
was evidenced by rdips occurring down the
lower panel, about 6”).
At the site, the building occupant
provided us with locations of the leaks
during a recent rain. Each of the leak
locations was at or near a primary frame
line.
Our inspection began by walking along the
roof steps, visually inspecting the area
for any obvious problems. It was evident
that a lot of activity has occurred in
the area surrounding the expansion
joints. There were regions of poor panel
modularity, where the vertical ribs of
the upper roof panel were not closed
tight. There was caulk on many of the
fasteners, and various other components
of the expansion joint detail. The layer
of sprayed on rubber sealant had been cut
around numerous fasteners with a razor.
Of particular concern was many fasteners
whose neoprene washers had split. Some
fasteners which were removed, revealed
slotted holes in the panel. A clear,
apparently silicone based caulk was used
in various areas
Some screw heads were “tipping”
in various areas of the expansion joint.
We then selected a location where leaking
had occurred in the past, and removed an
approximate 10’ section of the
second metal flashing (See Figure 2) in
order to get a good look at the original
installation. We were able to lift the
upper roof slightly, so as to view the
seal between the flashing and upper roof.
Inside the building, we utilized a
moveable scaffold to view various leak
areas from the inside. The presence of
leaks was evidenced by staining and by
saturated insulation adjacent to the
expansion joint. We were able to view a
standing seam roof clip, and verify that
it was a slotted clip. The slot is
presumed to have the purpose of allowing
thermal movement of the roof panels.
Conclusions:
Of primary concern on a roof of this size
is the allowance for thermal expansion
and contraction of the roof panels. If a
continuous steel roof sheet is
110’-0” long at 50 deg F, it
will be 7/16” shorter at 0deg F, and
7/16” longer at 100deg F, assuming
it is unrestrained against longitudinal
movement. If it is restrained against
movement by screws fastening the sheet
into a rigid lateral support, large
forces will develop in the panel, causing
the screw holes to tear and elongate, and
tend to bend the screw, which causes the
screw head to tip. When this happens,
water can leak through. Since it is very
difficult to keep thermal stresses from
developing in a metal roof, it is
strongly advisable to utilize details
where, if an exterior fastener does leak,
it will rdip outside the building
envelope. This can be accomplished with Z
shaped end dams and eave plates. (See
Figure 3).
The sliding clips utilized in the roof
attachment are slotted to allow for
thermal movement. All sliding clips will
cause some amount of resistance to
thermal movement. It is difficult to tell
how much resistance these clips are
causing, but based on the locations of
past leaks (at frame lines), and logic,
it is quite possible that the clips are
causing excessive resistance to thermal
movement at the frame lines, which causes
screw leaks, and since the screws at the
low end of the upper roof panels
penetrate the building envelope, the
moisture enters the building.
Based on current information, it is
unclear how the slotted clips can slide,
since the screw anchoring the clip to the
joist clamps the mating surfaces around
the slot. They will slide, but only when
a large amount of force is applied. It is
the experience of this office that a
sliding standing seam clip should resist
no more
than 2 lbs. of lateral force before
static friction is broken, and the clip
slides. This way, even if there are 20 or
more clips attaching a roof seam all
along the slope, the total resistance
(which is equal to the force on the
screws anchoring the low end of the
panel) does not become so great as to
cause screw and screw hole deformation,
which will leak.
If the roof clips are not sliding freely,
leak problems will tend to be much more
pronounced at the frame lines. The bar
joists can move laterally in between
frames, as they are pushed by the
non-sliding clip. However, at the
frames, the joists are essentially
immovable, because the rigid joist seats
are welded to the frame.
The second flashing which was installed
at the expansion joint is screwed to the
upper roof sheet, and the lower roof
sheet. Combined with the first (original)
flashing, there is significant additional
rigidity created. This will tend to
inhibit thermal movement all along the
joint. It is desirable to remove this
second flashing, especially since it has
apparently not solved the leak problems.
Figure 4 shows a possible remedy for the
roof leaks, without requiring the removal
or cutting of roof sheets. Note, however,
that installation of an eave plate (See
Figure 3) is the better way to prevent
leaking screws at the low end of the
upper roof panel. Eave plates may be
difficult or impossible to install
without removing or cutting roof sheets.
However, the idea should be explored with
the manufacturer and erector. Figure 4 is
depending on roofing tape to seal screws,
which may be effective, but will also
cause a need for periodic inspection and
maintenance. It must be noted that the
roofing tape will cause a small amount of
water pooling just upslope from the tape.
This is not a desirable scenario, but the
water shouldn’t enter through the
screws if the panels are well cleaned,
and the roofing tape is installed with
care.
The clear, apparently silicone based
caulk which was prevalent in many
locations at the expansion joint should
not be used in the future. It did not
adhere to the metal components of the
roof. Caulking individual screws
generally will not provide a good seal.
Over time, the caulk will lose bond with
the panel due to thermal strains. A good
quality roofing tape is a better material
to use. Care must be taken in tourquing
down screws with rubber or neoprene
washers. If they are too tight, the
washer splits and leaks. The washer may
not split immediately. Tighten to
manufacturers recommendations and no
more.
The triangularization of the vertical
ribs is apparently caused by poor panel
modularity during erection. These should
be caulked, and covered with roofing
tape. (See Figure 4).
Structural Solutions LLC and its agents
are acting strictly as consultants in
this matter. Structural Solutions LLC
does not warrant that any action taken
using the information in this report will
resolve troubles associated with this
building. Furthermore, Structural
Solutions LLC does not recommend that any
action be taken which would void any
warranties of the product suppliers and
sub contractors involved.
If the primary reason for the roof leaks
is due to excessive clip resistance,
it’s quite possible no amount of
work at the roof step will completely
solve the leak problems. At this point
two items that should be considered by
the parties involved are:
1. Removal of all anchorage screws at the
low end of the upper roof. This would
create a situation where the entire upper
roof is free to float, theoretically
preventing thermal forces from occurring.
(Assuming the clips are
capable of sliding). This step should
only be done at the recommendation of the
manufacturer.
2. Consider the effects of the
insulation. If the insulation is packed
too tightly between the roof sheets, it
may be causing thermal stresses to build
up similar to, or in conjunction with the
clips not sliding. Removal of thermal
blocks at the roof panels adjacent to the
building’s primary frames may
relieve this problem
Summary:
It is our opinion that this roof has
problems beyond which can be solved by
caulk, or other quick fixes. The roof
system design by the pre-engineered metal
building manufacturer does not adequately
allow for thermal movement of the panels.
This is causing screw hole slotting,
screw tipping, and the subsequent leaks.
The details used may be effective for
cold-formed purlin roofs, but not for bar
joist roofs. Bar joists are very stiff at
their supports, whereas cold-formed
purlins are much more flexible.
The manufacturer details show fasteners
which penetrate the building envelope,
and roof panel clips that have excessive
resistance to sliding. Using these
details on a bar joist roof with 110 ft.
panel runs is inadvisable.
Of additional concern is the quality of
installation. Over tightened screws were
noted in some locations. This must be
avoided, as split rubber washers will
leak, and if the screw is penetrating the
building envelope, water will enter the
building. Fastener manufacturers should
be consulted for an appropriate screw and
installation procedure.
Installation of the roof panels should
have stopped at the first sign of panel
rib triangularization. It was
inappropriate to continue the
installation after it was apparent that
something was wrong with the panel
modularity.
