Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

RE: Wind Loads on Roof Mounted Solar Panels

[Subject Prev][Subject Next][Thread Prev][Thread Next]

The link appears to be for the EuroCode wind actions rather than the article
you are querying. It doesn't appear to have any references to solar in it,
nor have any tables in section 3.

Others: My usual waffle:

Free roofs or canopies are typically tested and net coefficients given.
AS1170.2 indicates that for an enclosed building the internal pressure
coefficient is equal to that which is generated on the surface where the
opening is located, holmes indicates in his book that this has been
confirmed by testing. I think it is Fig6-6 in ASCE7-05 which in effect
summarises most of the coefficents in AS1170.2 for enclosed buildings. 

If have an opening on the windward wall then the internal pressure
coefficient will equal that from the windward wall Cpe=Cpi=+0.8. If have
opening on windward edge of roof then Cpe=Cpi=-0.9 etc...

The commentary to AS1170.2 and Holmes book provides methods for balancing
the effects of various openings: in effect it balances the volumetric flow
in against the flow out: which reduces to the proportions of areas of

So in the main the pressure coefficents on a free roof are comparable to the
combination of internal pressure coefficients and external coefficents for
an enclosed building. But not all that verifiable if using the simplified
table given in ASCE7-05: with opening building from memory at +0.55. Which
potentially under estimates the internal pressure coefficient.

Also the blockage for a free roof is a volumetric blockage, like a large
caravan under a carport: not an adjacent fence. By messing with internal and
external pressure coefficients, and openings can reduce the pressures on a
carport for example when building officials are concerned about adjacent
fences. A lot of trouble to go to for a verandah, and low fee, but a lot of
them are built without approval due to industry which doesn't know
difference between pergola and canopy. A pergola, open structure, here
typically does not require approval: the covered structure does. Build
pergola, put roof covering on at later date get notice to pull it down, or
get a structural assessment. Not being into wasting materials, I waste my

Anycase the point is the coefficients on free roofs should be comparable to
those on enclosed buildings with openings, and often difficult to determine
whether have a canopy with walls or an enclosed building with large
openings. Especially rural buildings, and poolside shelters which have
entire walls missing, so typically all buildings here, this office, are
assessed on the basis of comparing enclosed buildings and free roof
coefficents. We typically adopt low internal pressure coefficents based on a
probabilitic argument presented by Kitipornchi, and also because high
internal pressure coefficents which reduce the resultant on a member to zero
is not conservative.

Also consider that a free roof is little more than a flat plate in an
airstream so should be comparable against the pressure drag coefficients for
a flat plate: and aerodynamics of flat plates is where Holmes book on wind
loading of structures starts.

As Thor Matteson indicated, adding the solar panel doesn't necessarily add
more wind load. The load which was on the on the roof is now in effect on
the panel, but now distributed to the roof differently: by point loads of
the support frame.

In past posts I explained that the pressure coefficient for a parapet can be
derived from the windward edge pressure coefficent and the windward wall
pressure coefficent: ASCE7-05 however prescibes a value. As a roof becomes
more vertical the wind ward pressure coefficent of the roof slope become
more positive and approaches that of a wall.

So can play around with the pressure coefficients. Also most of the
coefficients were obtained from simple wind tunnel tests on shapes, not real
buildings. So best to ignore what the code exactly says, and treat as
expermental data to be interpretted, relative to the problem at hand.

Also from a risk viewpoint. Consider a submarine commander:

Dear commander from a financial viewpoint, we cannot afford to provide
instrumentation and feedback control systems. But rest assured we have
designed the submarine on the basis there is only a 5% probability that you
are likely to operate the sub at a depth at which it will be crushed.

The commander is most likley to reject operating the submarine. Yet that is
how we go about designing buildings. How does the occupant of a a building
know that the wind load or for that matter earthquake load is less than the
design load, and it is relatively safe to take shelter in the building, and
when do they need to evacuate the building? For ultimately the building can
experience a load in excess of the design load and when it does, the
building itself becomes a hazard to life. If the building is ductile and
starts to deform then maybe the occuptants can take warning from and exit:
but where to? Where is the low hazard clear space, where are the purpose
made shelters?

One of the benefits of the performance based Building Code of Australia
(BCA), is that the performance criteria are relatively general and
non-specific, and only required to be met "to the degree necessary". To put
another way the wind loading code AS1170.2 is a deemed-to-satisfy provision,
it does not have to be complied with. Whilst it is typically a lot easier to
to comply with the deemed-to-satisfy provisions, it is also less onerus to
make judgement calls and deviate from those provisions.

To quote the objectives of the structural provisions:

a) safeguard people from injury caused by structural failure; and
b) safeguard people from loss of amenity caused by structural behaviour; and
c) protect other property from physical damage caused by structural failure;
d) safeguard people from injury that may be caused by failure of, or impact
with glazing.

Quoting the basic performance requirement:

a) A building or structure, to the degree necessary, must -
i) remain stable and not collapse; and
ii) prevent progressive collapse; and
iii) minimise local damage and loss of amenity through excessive
deformation, vibration or degradation; and
iv) avoid causing damage to other properties,

by resisting the actions to which it may be reasonably subjected.

The fundamental basis of approval is adequate evidence-of-suitability. To
the BCA the judgement of a suitably qualified person is adequate

The point is that the manufacturers, installers, and other suppliers of the
solar panels should really invest in the testing. The problem here in SA, is
that just about anyone can get started in the building industry, selling
manufactured structural building products: with out any real investment in
RD&D. The suppliers rely solely on the local government building departments
to accept of reject approval for a given project. There is a high level of
variablity in the competence and experience of the regulators. The result is
a great deal of inconsistency between what gets approved or rejected. The
suppliers consequently complain that: "that council approved why won't this
one.". Further more if explain that what they have been getting approval for
is based on deficient assessment, they don't really care. They do one thing
in one council area and another thing in another council area.

Fit-for-function doesn't necessarily involve complying with the
deemed-to-satisfy loads, and not complying with those loads doesn't
necessarily pose an hazard to life. It basically comes down to a matter of
economics. Using AS1170.2 there are probabilitic models, so that different
regional basic wind speeds can be calculated for different return periods.
That is can also play around with risk, life expectancy, and return period
relative to a given structure. Terrain category, shielding, pressure
coefficients and life expectancy of structure are all basically judgement
calls. The design wind load is little more than an educated guess: and what
ever it is there is always the possibility that it can be exceeded.

Here the Bureau of meteorology (BoM) issues severe weather warnings when the
average wind speed exceeds 63km/h, and instantaneous gusts exceed 90km/h. At
these speeds need to take shelter, these speeds are exceeded every year, and
wind damage occurs to trees, which fall over onto power lines or otherwise
crash through house roofs or onto cars. The minimum ultimate strength design
wind speed we can use is 30m/s, from the beaufort windspeed chart that is
defined as a hurricane: we don't have hurricanes in SA. From the BoM website
the maximum speed I can find recorded for SA is 167km/h, the regional design
wind speed is 162km/h which roughly has a 5% probability of being exceeded
for a 25 year life expectancy. Certain industrial activities are hampered by
wind speeds less than 20km/h. From the Beaufort chart the natural
environment is becoming something of an hazard at around 52km/h. The wind
codes are also not based on sustained wind loading but short term wind load.
A review of storm damage in newspapers indicates wind damage to houses
occurring at speeds of around 110km/h on a regular basis: mostly older
houses losing roof cladding/tiles, or trees crashing through roof.

Trees large enough to cause damage to house, are typically in the
significant tree category and are protected by law, and not allowed to be
cut down. When a tree is likely to crush the house, the exact magnitude of
the direct action of wind on the structure is of minor concern. First really
need to protect the house from the tree: and in the main no one is going to
do that.

My general view is I don't want to be walking or driving down the street and
get hit by flying debris from a failed structure. I expect a certain amount
of warning: like if branches are snapping of trees then I am in an hazardous
environment and need to take care. From perspective of general design, I do
not expect to design systems to accommodate peoples irreponsible stupidity,
only to design systems to accommodate peoples reasonable expected behaviour
under expected conditions. Houses are not storm shelters, on the other hand
we do not expect to rebuild them every year, but a certain amount of annual
maintenance is to be expected. If the same problem arises every year, then
likley to take steps to alleviate it. It is a cost benefit issue: annual
repair and maintenance versus a 5 year schedule or a 10 year schedule etc...

So my view is if small scale residential structures can remain serviceable
at wind speeds of 110km/h or more then the structure is fit-for-function,
and does not pose an undue hazard to life: and likely to remain serviceable
for many years without need for significant maintenance. Survival of the
structure at anything above 110km/h is an added bonus economically. At
speeds above 110km/h purpose made shelters should be provided, and the load
does not need to extend to the whole house. The shelter itself still has
limited usefulness and the design load can be exceeded. The design of the
house changes to retaining debris on site: the building is permanently
deformed and damaged and no longer fit for habitation but still otherwise
anchored to the site.

The estimated design pressure is a combination of estimated wind speed and
estimated pressure coefficient. A given design pressure can be determined by
pushing wind speed up and pushing pressure coefficient down, or vice versa.
We can also throw magnification factors in for good measure. The ultimate
design decision is whether the design pressure adopted is reasonable. The
parametric description of how that pressure was derived just provide some
environmental factors which can be observed on the ground and which provide
a more intuitive means of making decisions.

I am not saying that I design for 110km/h. Most structures built without
approval, have resistance fairly close to the resistance required for the
deemed-to-satisfy wind speed of 162km/h, but not quite good enough. Our code
doesn't mention permissible over stress for wind load: therefore have to
find some other basis of acceptance to avoid potentially unnecessary
disassembly to strengthen an existing structure: most of which have been
around for 6 to 15 years before discovered as built with out approval.

To me civil/structural engineers seemed highly focused on the hazard to
life, and a role of protecting life and providing safety: but only from the
perspective of what a code says. Yet they don't achieve such, in the main
their designs are the hazard to life. People a top a multistorey building
reporting looking out off the window as hurricane passes by: its crazy. All
very nice unless the hurricane exceeds the design event, then they won't be
able to get out off the building fast enough and they have no where to
retreat to anyway. Increasingly in all disciplines, design seems to be based
on resisting loads which should be avoiding. As I have mentioned before it
is the mode of failure and behaviour at failure which is the important issue
not resistance to the load. More effort should be put into failure modes
effects critically analysis (FMECA).

>From  design-of-experiment theory, typically have too many variables to take
one at a time, so pick a high and low value for a given parameter and test.
Ignoring the test, pick a high estimate of pressure coefficient but adopt a
low wind speed which expect to have a reasonable chance of being exceeded
every year: make sure structure can resist without permanent deformation.
For the low estimate of the pressure coefficient adopt a high wind speed and
allow permanent deformation but avoid fracture.

Preferably determine the collapse load and then determine the pressure
coefficient and high wind speed from that: and assess whether you find that
acceptable. Know the required design wind speed but don't know the pressure
coefficient, what coefficient is required to produce collapse of practical
support system? Does that coefficient seem reasonable?

And finally at what wind pressures are the solar panels themselves going to
be ripped apart?

The manufacturers in the main are not going to conduct testing until the
buyers question the performance of the systems. So the main issue is
defending the design decision made. As I have indicated in the main the
magnitude of the design loads in the codes are not about hazard to life but
economic loss. The hazard to life is not reduced by increasing resistance.
The solar panels themselves are probably a greater hazard to life than the
support structure: just try getting the manufacturer to change the panel
design. Their design decisions, and concepts of fitness-for-function are
totality different than those which flow into building codes.

Probably not much help: but its one perspective.

Conrad Harrison
B.Tech (mfg & mech), MIIE, gradTIEAust
South Australia

******* ****** ******* ******** ******* ******* ******* ***
*   Read list FAQ at:
*   This email was sent to you via Structural Engineers 
*   Association of Southern California (SEAOSC) server. To 
*   subscribe (no fee) or UnSubscribe, please go to:
*   Questions to seaint-ad(--nospam--at) Remember, any email you 
*   send to the list is public domain and may be re-posted 
*   without your permission. Make sure you visit our web 
*   site at: 
******* ****** ****** ****** ******* ****** ****** ********