Additional roof loading from solar array

[OHEcoEng]

Hi, all. I am new to these forums and I hope that I can invoke a little interest in the alternative energy field.

I have been a practicing PE for 10 years now in the wastewater/environmental field. I recently made the jump to solar and wind design. I certify electrical diagrams for these systems, and will be designing basic foundations for pole and ground-mounted systems.

Along with this comes structural evaluations of existing roofs, and I'm a bit rusty after being out of school for 15 years. There have been a lot of changes and advances in doing this.

I have been studying ASCE 7-05 for about five days now, and feel rather comfortable with the dead, roof live, and snow loading for our area (Ohio), but I am having trouble getting a grip on the appropriate analysis for wind loading.

This is a 1960's/70's era warehouse, flat roof (1.5 deg slope), built up membrane, no ballast. The roof drains into gutters on the downslope side, so ponding from drain blockage is not a concern. Joists are 28LH06, 6' OC, 50' span.

Following are my parameters:

Roof live load: 20 PSF
Existing dead load: 16.1 PSF (roof, decking, mechanical allowance, insulation, joists, sprinklers/lighting)
Snow load: 25 PSF

Proposed additional PV load (modules, racking, ballast): 8.2 PSF

Wind load data: Occ Cat II
Imp fact., I: 1
Basic wind speed: 80 MPH
Exposure: B
GCpi: +/- 0.18

Using Method 2 for MWFRS, I find a worst case force on the original structure of -15.32 PSF in zone 2E (I didn't find any downforces).

The solar modules are at a 10 deg. tilt...here is where I'm struggling. Do I evaluate this as C&C, or as another MWFRS, and then add my "open structure" forces to the existing building uplift for worst case (downforce on array subtracted from to internal pressure uplift)?

Since the solar modules are not mechanically attached to the roof, the uplift does not transfer to the decking, only additional downforce.

I realize, too, that there are some potential snow drifting issues in the vicinity of the array.

Since this is a standard evaluation that I will be performing on a repetitive basis, I'd like to get this right.

Building is 30 ft tall, 300 ft x 100 ft (load bearing masonry wall at the 50 ft length; utilizing two spans across 100 ft).

I am using LRFD for this evaluation, and have a factored total safe uniform load of 469 PLF. I have calculated 308 PLF for the existing structure, with a maximum uplift of 60.1 PLF.

 

[ishvaaag]

In my view one of the intents of the codes is to be technically accurate. Hence you only need to add wind forces as really induced bt the surmised wind of each case. Except for construction issues, I would look the as built situation for forces, i.e. the building with the panels. Construction sequence may oblige to consider alternative dispositions and/or separately the building from the panel outfits, but this should be rare except on bad construction practices, I think.

You may also support your decision on some CFD analysis if not for the forces (it is said the current state of the art can't cover nor be substitute for code specs) at least yes to see whether and where the building with panels is getting bigger wind actions and becoming worse on the addition of the panels, that, since wind is assumed to be from the whole rosetta, should be almost anywhere.

I also see that you must be using laying and integral panels to the ballast, because in other case the weight would be inadequate for the wind speed.

 

[vmirat]

I'm currently designing a project to add solar water heat panels on the top of an existing municipal pool roof. This is a gable roof with 10 degree slope on both sides. Building is 98' parallel to ridge and 118' perpendicular. Eave height is 15 feet. The solar panels are 10' long by 4' high at an angle of 69 degrees to horizontal. Panels are installed in rows 7 feet apart, perpendicular to the ridge.

I am using Figure 6-18A to determine the load from wind on the panel which transfers to the roof joists. Although this is for MWFRS, I believe it more accurately represents the actual wind pressures on the solar panel. Figure 6-19A is for C&C but it is for a total roof area where there are localized pressure areas due to the size of a roof.

The more complicated question is the issue of snow drift. I've tried to find as much info as I could off the internet since ASCE 7 doesn't really cover this situation. I found a document from a Canadian source which indicates that the drifts will form in a somewhat parabolic shape (in a convex shape) in front of the face of the panel. However, this document doesn't address what happens if the wind blows from the other side of the panel. I've decided to approximate a drift form which is triangular from both directions. I've attached a powerpoint file showing what I'm doing.

There is a paper by Michael O'Rourke which discusses snow drifting and solar panels but it costs money to download.

 

[BAretired]

Roof live load: 20 PSF
Existing dead load: 16.1 PSF (roof, decking, mechanical allowance, insulation, joists, sprinklers/lighting)
Snow load: 25 PSF

If the snow load is 25 psf, why would you consider live load?

The solar modules are at a 10 deg. tilt...here is where I'm struggling. Do I evaluate this as C&C, or as another MWFRS, and then add my "open structure" forces to the existing building uplift for worst case (downforce on array subtracted from to internal pressure uplift)?

For design of the roof, I would think MWFRS. For the panels themselves, C&C would seem appropriate.

Since the solar modules are not mechanically attached to the roof, the uplift does not transfer to the decking, only additional downforce.

I would think the solar modules must be attached to the roof to prevent them from flying away.

BA

 

[OHEcoEng]

Thanks, vmirat. At this point, I don't know if I'm being way too conservative or what.

I looked at your project, and I can't tell what you have on your existing roof. I am assuming that the 52DLH15 joists are capable of supporting the additional load, but your decking is not?

I am not too comfortable with my review, since factored loads came out to right around what the maximum loading is. I didn't even look at maximum moments and shear.

It looks like this stuff is great discussion material, and if you (or anyone else) can point me to existing threads - or get involved int eh discussion - I would truly appreciate it. I know that this type of evaluation isn't anything new/rocket science, but it is something that I'm going to have to repeat on a regular basis.

I am now going into another situation with modules being mounted on a monoslope (1.4 degrees) roof with standing seam and 28K10 joists, 5' OC and 49 ft. span...seems to have a boatload of extra capacity. It will use s5 clamps and clicksys rails.

 

[vmirat]

OHEcoEng,

Actually, I'm still in the process of determining if the existing structure can support the additional dead load of the panels and additional live load of the drifting. I've pretty much decided that the corrugated roof deck won't be capable of carrying the concentrated load of the solar panel supports, hence the need for new purlins to transfer the loads to the OWSJ's.

Because I couldn't find a whole lot on this subject, I'm being pretty conservative with my analysis. I'd like to get that paper by Michael O'Rourke but I work for the federal government and they're too cheap to buy it for me

 

[OHEcoEng]

BA:

ASCE 7-05 details load combinations in Chapter 2. Using 2.3.2 equations does in fact exclude Lr when considering S. I was just providing parameters.

This is a ballasted system (hence the 8.2 PSF dead load). I am not designing the building, just evaluation it's capacity to support the additional loading...

 

[pwp2010]

I am a mfgr of racking to mount Solar panels. The uplift from the 90 mph building code requires a considerable amount of ballast blocks on a ballast mounted system. The big advantage of the ballasted racking system is the elimination of the need for mechanical anchors. My standard rack has a tilt of 10 deg. I would like to come up with a standard formula to calculate the weight of ballast required for a variable array. This formula would have to consider all relevevant variables. I can create the formula and the computer program if I can get help in identifying the correct elements of the formula.

 

[msquared48]

I can see a ballasted system working for flat roofs, but have a real problem with sloping roofs without some form of mechanical anchorage.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[OHEcoEng]

I'm still trying to work through this one...

And, yes, msquared, ballasted systems are for flat roofs only, though the more I look into this, the more it appears they aren't suited for many of those, either.

 

[PSlem]

Good example of roof calcs from a thread on pg. 4 called win loading, am I doing this right?

http://www.solarabcs.org/wind/

 

[OHEcoEng]

That is a good article, but not applicable in this case (ballasted with an angle off of the roof).

 

[ishvaaag]

http://heron.tudelft.nl/52-3/2.pdf

Another article on the matter.

 

[SteelPE]

Maybe I have done this wrong in the past... but in this instance for, rooftop equipment, I use section 6.5.15.1 of ASCE 7-05 to get my wind loads on the panels.

 

[IRstuff]

Not particularly relevant to this discussion, but there is a company that makes solar PV panels in the same form factor as a conventional "Eaglelite" tile, but in shingle length. This allows you to simply remove the existing roofing tiles and replace them with shingle strips of PV panels. Near as I can tell, the dead load would actually be somewhat lower than that of the original tiles replaced.

TTFN

FAQ731-376