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]
I can't argue your points - they are well stated and while I would agree
with you, it is a moot point as the client found another engineer who has
taken a simplified approach to the problem that may not be compliant with
either of our views but will get the contractor's three projects permitted
over the counter. 

>From my perspective, I had to view the practical solution and in the process
understood that there were other potential problems that needed to be
considered. Once I brought up my concerns about the capability of the
existing wood truss roof to carry the additional loads from the panels that
they were not designed for, the client decided their goal was to sell
photovoltaic panels and would therefore find an engineer willing to stick
his neck out and simplify the code back to the 97 UBC four line analysis
type of design.

This does not work for me. I won't put my stamp on anything that I believe
may be a risk unless I consider the alternatives. From what you have
written, the most accurate way to address the problem would essentially work
best if the industry establishes tables of wind values for various
conditions and provide the local engineer with the design loads in order to
allow the engineer only to address the ability of the roof framing
(especially if a metal side plate manufactured wood truss) to support the
reactions at the connections of the panel framework. This puts me in
agreement with your comments about the solar panel manufacturer's
responsibility to address the reserve capacity of the roof (if one exists). 

The bottom line is that by the time the project came to me it had been
submitted to the building department and had received a list of corrections.
The client was not willing to take the time for me to come up to speed or to
consider the research that I would have suggested to assure that the roof
could carry the load - especially where the roof is flat and there may be a
ponding problem due to the original design based only on the original
materials used without consideration for what technology may bring in the

Thank you for your response - I do agree with you but when it comes down to
a company with a product looking for a professional willing to stick their
potential liability neck out for a fee - they came to the wrong engineer.

Possibly Bob Garner can carry this further and I certainly would recommend
that you allow him to use your response since it takes the issue back to the
manufacturer's design stage to be more creative when addressing a retrofit
system on a tract style home or one constructed to the IRC prescriptive

Thanks again,

-----Original Message-----
From: Conrad Harrison [mailto:sch.tectonic(--nospam--at)] 
Sent: Wednesday, August 13, 2008 11:55 PM
To: seaint(--nospam--at)
Subject: RE: Wind load design for Photovoltaic panel installations


Jordan's approach may be the more difficult to sell to plan checkers, but it
is potentially the more correct approach.

In Wind loading of Structures, by John Holmes, the starting point for the
whole subject is with aerodynamics and fluid flow. He also provides some
guidance for design of roof mounted solar hot water systems, and antennae.

Using AS1170.2 we have drag coefficients for various structural sections,
and also frictional drag requirements over cladding. But still solar panels
and other roof mounted structures are outside the scope of the code: but
then so are most real things.

Using the code some judgement needs to be made regarding the interference of
airflow over the roof and the obstruction of the solar panel. Free monoslope
roofs give some guidance, but gap between panel and roof would have some
form of venturi effect. Then there will be turbulence where the airflow hits
and leaves the solar panel obstructing airflow over the roof surface. The
direction of the wind pressure could be uncertain.

Consider another example, a skillion roof lean-to against the wall to a
higher level gable roof. Does the wall experience wind ward positive
pressure, or more a continuation of the roof suction? What height above the
skillion roof experiences suction, and when does the interference between
the wall and roof cancel out, and the windward wall pressure dominate? 

Using the wind loading code is highly dependent on making judgement calls of
worst case conditions for real world buildings which seldom have the shapes
used in the wind tunnel testing.

It is also a matter of how easy it is for you to apply your own design
factor to magnify design loads to allow for uncertainties, and how well the
results are then accepted by those who have to pay for the implementation,
and how practical it is to implement.

It is outside the scope of the code, it can therefore only be a judgement
call: unless some manufacturer willing to do wind tunnel testing, in which
case they will have the test results and no other manufacturer will.

Here we don't have the same liability issues as you have in the US, so here
all parties are generally more willing to take the risk and make and accept
judgements. If risks are explained then losses are also more acceptable.
Insisting on great expense to strengthen something, then experiencing loads
greater than design load and still loosing is good cause for owners to be
unhappy. A balance needs to be made between the expense now, and the cost of
replacement. The hazard to life, is not addressed by the magnitude of the
load, but by the mode of failure and circumstances. When a severe weather
warning is issued then expect hazard to life: not the least of which is
trees being uprooted and crashing through house roofs or bringing cables and
utility poles down. Direct action of wind on the roof is minor hazard to
life: unless in hurricane or tornado territory.

Also it is not the 50 year mean return period wind speed need to worry
about, but the ultimate limit state wind speed with a mean return period of
500 years: and roughly 5% probability of being exceeded for a 25 year life
expectancy for a building: this is the wind which will start to tear the
building apart. Combining this with the 5 percentile resistance, the
probability of the load exceeded and the system below strength is very low:
more likely to be over strength and under loaded if everything worked out
right. None the less failure is inevitable. The question is will it happen
in our life times, will the buildings last long enough that we can move
forward, how often are we hampered from moving forward because of need to
rebuild: both as a society and as an individual?

My point is that generally things are built before regulations are imposed.
Regulations are imposed because of perceived hazard or a few failures, and
initial regulations tend to be prescriptions based on what is known to be
working versus what is known to be failing.

So when there is no existing specific guidance, any rational justification
is better than none. Engineering doesn't make it safe; it simply reduces the
risk of failure compared to having no engineering. With no engineering best
guess of the risk of failure is 50%: either it will or it won't. With
engineering hope to do better than that.

So one option is to start with the resistance of what the installers prefer
to do, then work that back to the maximum pressure coefficient which can be
applied, then compare that against the code. If lucky will get a coefficient
greater than anything in the code.

Using AS1170.2 I can go one step further, and maintain risk at 5% and adjust
life expectancy of the structure, on condition do not drop design wind speed
below 30m/s. (we vary site wind speed with height, then calculate pressure
compare ASCE7 which varies pressure with height)

I think some solar hot water systems only have a life of about 10 years, so
why design the support structure for a longer life, if it needs to be
modified for new systems? And who knows how long a solar cell remains
operational before pollution damages it. Of course the support structure may
be there for as long as the house, with solar panels replaced occasionally.

Further if solar panels are to be mounted on roof tops, then the
manufacturers should consider designing the panels and support structure to
suit the potential reserve capacity in the variety of existing roofs rather
than imposing wide scale assessment of existing dwellings. That is the panel
manufacturers should do the wind loading research and publish the
recommendations for support structure.

Conventional light timber framing represents the minimum resistance
available, if consider that is over stressed in the first instance then,
zero reserve resistance for additional loading. Therefore need an
independent support structure reaching to roof level if that is impractical,
unsightly and unacceptable: then the alternative is to accept lower design
load. A lower design load reduces life expectancy if risk level is

And by all accounts from most members of the list conventional timber
framing in the US is considered over stressed when checked against the
structural codes. Therefore if resistance calculated by codes is considered
accurate, then the design load needs to be lower for it to be acceptable,
which once again for a given risk means the life expectancy of conventional
framing is lower than engineered construction. That lower life expectancy is
acceptable to the community and authors of the codes; otherwise conventional
framing would be rejected by the regulators.

All structures designed to a code of practice are at risk of failure, it is
unhelpful to think in terms of providing safety: we are dealing with
uncertainty; there may be no safety at all. Also lawyers don't care what
codes say, but what the designer should have been expected to know.

So sometimes it is necessary to take the difficult path and educate the
building officials, or assist them by getting calculations independently
reviewed before submitting for checking. Another perspective is if the
building officials ask the questions and cause the problems for owners and
builders alike, then they better be capable of understanding the solution.
The person in the building department, who posed the problem, usually can,
and is also likely to accept rational judgement when the codes have no
specific requirements. Also when all the problems and practicalities are
pointed out, building departments are amenable, and introduce some
guidelines for acceptable practice, so as not to cripple local business and
the building products they are supplying.
Whilst most of the work I do comes from manufacturer/builders I don't much
like such businesses which rely on external consultants. I believe they
should employ engineers on staff and properly design and engineer their
products, and otherwise properly assess individual projects and provide
proper documentation for building approval. Rather than wait until the
building officials issue them with 4 pages of questions, then turn up at
some consultant's office crying for help: and a time constraint to avoid
paying the approval fees again. If the product is appropriately designed
then the manufacturer would be one step in front of the regulators.

Given that the owners and contractors will want a 10 minute solution and low
fee, and supplying at maybe a rate of one or more per day.

Really need to assess how quickly can get a conservative answer, and what
risks you are exposed to.

The simple answer is to try pushing some questions back, to both the
manufacturers and the building departments. See what practical compromise
can be reached.

(Oh! the probability stuff is in the commentary to ASCE7, and also the book
by Holmes it is also in AS1170.2 but the probability models are for
Australian conditions.)

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

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