Epic racking malfunction

[littleharbor]

Came across these photos today. I think someone's got some major explainin' to do here.
. Solar racking failure..jpgSolar racking failure 2.jpgSolar Racking failure 3.jpg

[SunEagle]

Wow.

I'll huff and I'll puff and blow the whole house down!

[bcroe]

I'm adding to my list of Ground Mount Advantages, less likely to be disturbed by high wind. Bruce Roe

[SunEagle]

I can't be sure but it looks like that panel system was just sitting on the roof using very small supports that might be considered as "weights" but certainly not enough to anchor the whole thing.

I wonder if there were any tie downs at the corners?

[DanS26]

Good deal on some slightly used panels on eBay next week.........

[foo1bar]

I'm adding to my list of Ground Mount Advantages, less likely to be disturbed by high wind. Bruce Roe

I'd say ground mount is more likely to be disturbed.
Ground Mount has no structure shielding it, while a typical roof-mount does have a structure on one side. (So wind blowing at the panels pushes them into the roof, wind blowing away from the panels hits the roof/structure first)

In this case it looks like they did a ballasted mount and didn't use enough ballast.
You could just as easily do a ground mount and not use a good enough mount (not put in sufficient footing, or skimp on the metalwork holding up the racks or do other things that make the mount insufficient)

[inetdog]

FWIW, wind speed at roof level is often significantly greater than within 6-10 feet of the ground.

[dennis461]

I can't be sure but it looks like that panel system was just sitting on the roof using very small supports that might be considered as "weights" but certainly not enough to anchor the whole thing.

I wonder if there were any tie downs at the corners?

Those are 'sticky feet' glued to the roof, instructions were not followed (Attorney for supplier),
Glue did not dry (Attorney for installer)
Wind was above design limit (Attorney for AE firm)

Could we have a clue (not glue) from the OP, what country, what photo source?

[peakbagger]

I used to have to deal with industrial roofs in a high wind zone. They usually fail at the edges, usually the upwind side. Air blowing against the flat wall of the building has to go somewhere so it goes up over the roof and creates a low pressure zone above the roof. It will lift the roof membrane and decking right up and once the air get under it the whole roof zips off. The insurance company would always make us test how much tension the fasteners would withstand in the underlying decking near the sides and corners of the building. We generally needed tighter screw spacing on the edges due to this uplift potential. The other issue we would run into would be fasteners rotting out due to water getting under the membrane, either due to roof leaks or damp conditions on the inside. Indoor swimming pools with their associated chemicals can really raise heck.

I would guess that someone didn't design the ballasting to have more weight on the edges or the design wind speed was exceed. There always can be tornado or unusual straightline wind event that can ruin anything, that's what insurance is for.

[solar pete]

Epic fail alright, I wonder who installed it

[foo1bar]

It doesn't look like it was glued to the roof to me.
It looks to me like concrete pads on top of some isolating thing (strips of something like cork?) to separate it from the roof.
If it were glued, I would expect to see some parts of the roof where it was pulled up.

It also looks like the array on the standing seam roof in the background wasn't damaged.

[dennis461]

It doesn't look like it was glued to the roof to me.
It looks to me like concrete pads on top of some isolating thing (strips of something like cork?) to separate it from the roof.
If it were glued, I would expect to see some parts of the roof where it was pulled up.

It also looks like the array on the standing seam roof in the background wasn't damaged.

http://www.bigfootsupport.com/en/big-foot-solar-support.html

[littleharbor]

Those are 'sticky feet' glued to the roof, instructions were not followed (Attorney for supplier),
Glue did not dry (Attorney for installer)
Wind was above design limit (Attorney for AE firm)

Could we have a clue (not glue) from the OP, what country, what photo source?

There was an assumption in the thread that it was the Netherlands. The images were taken from here with no real info. https://imgur.com/a/RnkqC

[J.P.M.]

There sure doesn't look to be any roof penetrations visible to me. If glue/adhesive was used, and it held, I'd think some of what looks like the top surface of the roof would have come with it, at least in one/two places. I also see no adhesive on the bottom of what look like plinths.

I'm also pretty sure those supports (plinths) are way too small to act as effective ballast for any, however slight, wind loadings. They do look to be designed to distribute what would otherwise be a point load (note that the pins/supports in the center of the plinths appear to be flush with the bottom surface of the plinths.), although given the slenderness (thinness) of the plinths, I'm not sure how long it might take the bottom side of the plinths to spall under any recurring wind induced vibrational tensile stress, but I'm no expert in concrete design.

It may also be that support beams for the rows were affixed to the roof at the ends of the rows somehow and what we're seeing are intermediate supports that were designed as not anchors, but designed for dead load only (again, note that the pins in the center of the plinths appear to be flush with the bottom of the plinths), although I wouldn't have done it that way. It may be that the area has a design wind speed of zero, but I seriously doubt it.

It's interesting, to me at least, that at least some of the array portions affixed to a metal roof appear to have held.

As for wind velocities and characteristics as f(elevation above some datum, site conditions,other), and mostly going along with Peakbagger on this one, something that's sometimes called bluff characteristics or bluff mechanics are one of many things that can impose varying wind regimes on upwind, or leading sections of structures that can easily have a +z, or positive upward vector component and thus "lift" all along leading or upwind "edges" of structures. That effect and all the others can be present on ground mounted structures and also those mounted atop other structures - anywhere really, depending on site conditions and characteristics. It may be a decent assumption to begin an analysis by thinking that wind speed does vary as the cube of elevation above datum, but that needs verification at the site. See ASCE 7-05, method 2 for some particulars about wind design, and a decent text on wind design such as "Wind Effects on Structures", 2D ed., Wiley, 1986, ISBN #0-4781-86613-X, among others, for some background if curious. I sometimes did the preliminary wind design for distillation, cooling towers and other such structures and equipment as part of the mechanical pressure vessel design which mechanical design I stamped/sealed, and then handed the wind/seismic design over to a structural engineer competent in such things for review/mod./stamping for that portion of the work.

Without a lot more information, and eyeballs on the site, I'd be reluctant to form a hanging party just yet, but it sure looks like someone was having their way with the pooch on that design.

In any case, my guess and hope is that this pictured site is not in the U.S. If so, U.S.engineering competence and oversight is further in the toilet than I thought.

There's also a perhaps small but still non zero probability that, without further information, the whole thing is a hoax.

Take what you want of the above. Scrap the rest.

[littleharbor]

From the "Making Lemonade Dept." Looks like the mid and end clamps passed the test with flying colors