Need a book? Engineering books recommendations...

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

RE: Wind load design for Photovoltaic panel installations

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Karen, and others

When you say: "Use the same formula for roof uplift that you would use on
the roof surface for the roof member design.", are you also suggesting using
the same pressure coefficients as were used for the roof, or adjusting them
to suit.

Most existing houses don't have good solar orientation or appropriate roof
slopes, so if the solar panels are to be any good, then they have to be
located to face the sun and tilted at an appropriate angle. That generally
means they should not be mounted flat and hard against the existing roof
slope, and therefore pressure coefficients for the existing roof are not

In Holmes "wind loading of structures" he indicates the critical parameters
relevant to wind loading of solar panels. There are many: two or which are
the angle of the panel relative to the house roof and the stand-off spacing
between roof and panel.

Increasing stand-off apparently reduces uplift, but increases wind load
parallel to roof, and a steeper pitch than the house roof increases wind
loading. He doesn't give any pressure coefficients.

However the wind tunnel tests for free roofs are basically done for flat
plates and therefore should give similar results to using the formula Jordan
was suggesting. None of which accounts for interference between fluid flow
along roof slope and obstruction of panel. It is basically a judgment call
if no research papers readily available: comparing flows over monoslopes and
gables, to assess the influence of the roof shape. The wind can approach
from any direction, and oblique angles produce the worst effects. Consider
wind flowing from one side of gable and in behind the solar panel. Look at
drag coefficients on exposed framing, frictional drag across cladding
profiles, pressure coefficients for free edge of a free wall. All this needs
to be taken into consideration when attempting to make a conservative

ASCE7-05 Fig 6-21 does give some pressure coefficients for roof top mounted
equipment. But the table seems more appropriate for box shaped type
equipment, such as water tanks and air-conditioning units.

Your approach seems reasonable if future research proves the pressure
coefficients adopted are conservative. If not then risk based probabilistic
design can allow assessment of what the consequences of underestimating the
required resistance are. That is work backwards from the resistance
installed, and calculate the maximum reference pressure qz that can be
applied using more appropriative pressure coefficients, then calculate the
risk of that being exceeded for the desired exposure period (life expectancy
of the installation and house: noting if the house exists already then part
of its life has expired already.).

In Australia (AS1170.2 & BCA) for temporary structures (importance level 1)
the acceptable annual probability of exceedence is (1/100) for normal
structures (importance level 2) it is (1/500), this is for ultimate strength
limit state. Inference therefore is that older buildings and secondary
structures around residential properties are less important than normal
structures (new house), but more important than temporary structures, and
looking for an importance level between 1 and 2, the closer to 2 the better.

Importance level 1 represents the minimum acceptable hazard to life, it is
not permitted to remain, it is only permitted a short exposure period and
has to be removed. Other structures have short economic lives and are
expected to be replaced after a few years: verandahs for example: this year
timber and flat, 5 or 10 years from now steel and gable. The structures may
be designed for 50 years, but with poor maintenance cease to be
aesthetically desirable and get replaced: or simply temporary until the
owners can afford something larger. The built environment is full of
buildings having differing life expectancies for a given risk due to the
codes they were designed and built to.

A code may be enacted by law, but it is still none the less only a
guideline, and the subjective opinion of the various political parties
contributing to its writing: building industry associations, insurance
associations, government departments. It mostly boils down to what is
economically viable.

The real wind pressure a building needs to resist is unknown, no matter what
magnitude of load we choose it could be exceeded. Therefore even if the
future shows our best guess to be in error, it still cannot provide an exact
load to prevent other future events.

There are multiple variables involved in getting the wind load and we can
push any of these up or down as need be to get a more conservative estimate
of loading: there is more than one way to get the same load. A more thorough
and detailed investigation can push the values to be less conservative, and
yet show over all the resultant is more conservative than a more simplistic
assessment suggested. For example push qz up for design, but for more
critical evaluation reduce it to the minimum can get.

My point is that design is always dependent on uncertainty, and needs to
make use of simple practical rules for decision making, at law however the
codes have to permit assessment using the most detail possible. If not then
there would be little to no innovation. Just because a code allows something
doesn't mean it should be adopted for day to day design, it still pays to be
conservative, however the economic constraints of individual projects need
to be considered, and that can push the codes to their limits of

All buildings and products are real world experiments. Until they have been
built we don't really know anything about them, and even if they are
repetitive products, we still don't know anything about what the future
holds. Designers are always attempting to forecast the future, based on very
limited information. Only the real building in its real environment, under
real operating conditions can give all the information required: but not
allowed to build it to find out. Therefore judgement based on codes to get
something real which can be tested. Unfortunately far too many people assume
a structure is fit-for-function and sufficient-for-purpose if it complies
with the code: so there is little in service monitoring until failure
occurs, and after failure can only monitor something different, that which
hasn't failed yet.

Any case wind is an uncertain load, it becomes an hazard to life when it
starts up-rooting trees, which from the Beaufort wind scale chart starts at
about 55 mph [25 m/s] to the Australian code the minimum design wind speed
is 30m/s [67 mph] for which the reference pressure is qz=0.55 kPa. Our
bureau of meteorology issues a severe wind warning when the average speed
exceeds 63km/h [17.5 m/s, or 39mph], and gusts exceed 90km/h [25 m/s, 55
mph], at these speeds are expected to take shelter. These speeds are
exceeded every year, therefore expect to take shelter in our houses. The
regional limit state basic wind speed is 45 m/s [162 km/h, 101 mph], if the
weather news reports this speed you do not want to be sheltering in your
house: it is ultimate strength speed, the yield strength of materials will
be exceeded where permitted, the structure will be permanently deformed and
cease to be serviceable. Post disaster facilities will be designed for an
annual probability of exceedence of (1/2000) not the (1/500) used for

ASCE7 simplifies by using 50 year mean return period speeds, and applying
load factors, and importance factors. It also adjusts the pressure at the
site rather than adjusting the basic wind speed to account for terrain and
height, topography, shielding and other factors to get a site/building
specific design wind speed. With a site wind speed, the speed can be
monitored at the site. Weather reports give speeds at airports and weather
stations, those speeds are equivalent to the regional, they can be adjusted
to the speed expected at a site. It is therefore possible to develop an
awareness of how reliable or conservative the statistical models in the code
are. It's a simple matter to rearrange the formula for qz in ASCE7 to get a
site wind speed.

With improved awareness of speeds experienced locally, it is then possible
to make better judgement of how much reserve is in estimate of qz and in
estimate of pressure coefficients. Pushing qz up is the easiest way to get
conservatism, when uncertain about pressure coefficients, ignore possible
reductions and increase the height (use maximum height of building rather
than average, and don't interpolate anything where permitted).

My essential point is that we can only make a judgement. If we are above the
code minimum then the hazard to life is taken care off, at the default
importance levels and associated design wind speeds all kinds of rubbish in
the built environment is going to pose a hazard to life, trees crashing
through house roofs, cars rolling, utility poles dragged over by falling
trees, older house tearing apart: it is time to seek shelter in something
designed to be serviceable at these extreme conditions. Life safety requires
attention to a lot more issues than simply increasing the resistance of the
structure. Any thing above the absolute minimum permitted by the code is
about economic loss, loss of amenity and economic recovery after an extreme

Therefore if under estimate the wind load, due to under estimating the
pressure coefficients, then the principal issue is an earlier loss of
structure and cost of replacement, rather than an hazard to life. That turns
the problem into an economic model, including the designer, owner, and their
associated insurance companies. If guess a high pressure coefficient now
then have a high up front cost, and small annual insurance payments. If
guess a low pressure coefficient, then have low up front cost but higher
annual insurance payments.

And roughly speaking if the problem is directly within the scope of the code
then there is a 5% probability that the load will be exceeded for a given
exposure period. Or alternatively an expectation that will loose 5% of
buildings, and therefore need at least 5% of the capital value of the
building stock to recover after a disaster event. Need more because of older
buildings and consequential damage. Someone, somewhere has to have the funds
for recovery. Disaster level events however are another issue, here
concerned with localised event to one of more sites. If we haven't made an
error of judgement then our insurance shouldn't have to pay, it is the house
owners insurance which pays, or the house owner.

If not happy about the potential risk, then load the fees accordingly. Low
estimate of wind loading then high design fee to cover your potential payout
for losses and damages, high estimate of wind loading then low design fee
but owner has to pay more for construction. That is assessing the effects on
cost of varying your estimate of the pressure coefficient up or down: like
the authors of the code who don't have a code to tell them what to put in
the code. Failure is inevitable and the cost of replacement has to be
covered, or loss accepted. House roof we want, solar panel at present we can
do without: therefore can accept loss of solar panel, and cover cost of
replacing roof. Of course failure is also a waste of materials and fuel, so
if really want to be green, owner should accept a high estimate of wind

As for life we risk it every time we step into a car and travel by road. To
be cynical the medical profession needs a resource to harvest organs, if we
fixed the road hazards (blind spots etc ...) where would the doctors turn
to? But I reiterate that low loads don't cause a hazard to live, and high
resistances don't remove the hazard. To protect life we need to be more
ingenious than simply picking big numbers.

Sorry for all that! My thoughts escape me.

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: 
******* ****** ****** ****** ******* ****** ****** ********