Can I support a pergola on roof without anchorage? 1

[milkshakelake]

I've been tasked to design this pergola to be on a 12th story rooftop. The wind loads are not bad; it's 98 MPH per ASCE 7-05 and 117 MPH per ASCE 7-10+. Negligible seismic loads. The owners want to do it without anchorage, which means using weights for uplift + sliding.

Is it permissible to do this without anchorage?

ASCE 7-10 says that anchorage is needed. Using 13.4 of ASCE 7-10 for guidance, it says that

Nonstructural components and their supports shall be attached (or anchored) to the structure in accordance with the requirements of this section and the attachment shall satisfy the requirements for the parent material as set forth elsewhere in this standard. Component attachments shall be bolted, welded, or otherwise positively fastened without consideration of frictional resistance produced by the effects of gravity.

This is taken from seismic part of the code, but it makes sense for wind.

However, I'm inclined to disagree with the code. I'm wondering if this is any different from a rooftop table or umbrella. I've seen trees in unanchored planters go on roofs. I'm not sure at what point something I consider to be merely a live load becomes a nonstructural component. I'm thinking that this is a matter of engineering judgment, but if there's code guidance, it would be useful.

 

[Alistair_Heaton]

There are codes for solar panels on flat roofs unanchored.

 

[milkshakelake]

@Alistair_Heaton Good point. A solar panel would have more uplift than this, so by that logic, this is okay to be unanchored.

 

[phamENG]

First, please don't use the code to cite building code...that's for posting segments of programming code. It's a nightmare to read text in there. Just use the quote box.

I don't think it makes sense to apply that provision to wind load - in a seismic event, you don't necessarily know which way things are going. You could have a combined lateral and vertical movement. So on the upward thrust of the earthquake, your dead load (and resulting normal force for and frictional resistance) is reduced. (At least this is my understanding of seismic activity...I'm a Floridian transplanted to the mid-Atlantic, so I'm far from an expert.)

A couple things to consider:
1) A tree, if overcome by wind load, will just fall over. It lacks the ability to take flight and the center of gravity is low enough that it'll just tip over rather than slide. So it makes a mess, maybe damages the roof membrane or decking, but nothing significant.
2) Your pergola is a little different. Unlikely that it'll get picked up, but it can certainly slide. Depending on the railing - is it a parapet wall or a railing? How high is it? Could it break through it or flip over it? What if somebody decides it's not quite enough and they put a sunshade over top? Now it could be picked up. Now you have a pergola falling 12 stories onto the street below. Bad news.

Now when I said it doesn't make sense to apply it to wind load, I just meant the direct justification. It always makes sense to provide positive anchorage.

If they insist on not providing positive anchorage (which is understandable for a tenant), then don't skimp on the uplift, overturning, and sliding resistance. I've been on a a 40th (+/-, maybe the 38th?) floor terrace in Manhattan...utterly miserable. I don't know how anyone can stand it. Calm day on the ground and outside the city, but the wind was howling up there. Probably not as bad on the 12th, though.

 

[milkshakelake]

@phamENG Thanks for letting me know about the formatting. Fixed it.

Good points there. It's true that it's fundamentally different than a tree; it could feasibly be picked up, especially if someone throws a shade over it. There is a 3'6" fully grouted CMU parapet so in an insane hurricane scenario, it's unlikely but it could tip over the edge. I think if I design it correctly and use friction while in uplift, it should be okay. Will overdesign it a bit and discuss w/ client what edge cases to consider in the design.

 

[JStructsteel]

Now, is the excess weight going to require the beams below be strengthened?

 

[milkshakelake]

@JStructsteel That's outside of my scope; someone else is checking that.

 

[Alistair_Heaton]

That's usually the problem with solar and ballast mounting. It also effects your design.

Your live load gets added in.

Maybe raise it and use the floor to be a load spreader and contain the ballast so it's not point loads or can be seen.

 

[skeletron]

Just as food for thought:
a) 10ft x 10ft is the limit for "accessory structures" which do not require permitting/engineering in my area...that would probably be my threshold to start
b) Without anchorage, I would try to rationalize to note somewhere in the plans and specifications that this is a temporary structure. Maybe a re-inspection period after ___ months/years or even a plan drawing showing acceptable regions to place the pergola within. I wouldn't design it for "temporary" loads though (ie. still designing for my q1/50 standard wind load, 2% in 50yr POE for seismic)
c) If there are issues with calling it "temporary" I would use anchorage
d) In telecom design (antenna ballast mounts on rooftops) there was big contention over the code change that clarified the definition from "positive connection" to also mean "do not rely on friction alone". I think that is how it is now worded in the Parts & Portions section of the code.
e) 12th floor rooftop sounds like there is a bit more "consequence to failure" than say a low-rise building. Although, the load is the load...

I would tend to want anchorage. I also understand the comparison to having BBQ's or park benches, trees, and benches on the rooftop without the same concern.

 

[dhengr]

Milkshakelake:
In any case, the ballast or counter weights have to be tied down or immovable. They can’t be conc. blocks, or some such, which some joker can move over there to make a little sitting bench.

 

[Tomfh]

To me it's akin to rooftop furniture. If you don't have to anchor an umbrella, then why this? I would make sure the weights cannot be removed.

 

[HTURKAK]

Yes... As long as it is proved with calculation that it is safe against sliding , overturning ..

You did not mention the surface finish ..probably ballast ? You may consider somehow large , heavy base plates over the waterproofing..









Tim was so learned that he could name a
horse in nine languages: so ignorant that he bought a cow to ride on.
(BENJAMIN FRANKLIN )

 

[Alistair_Heaton]

Stick a limitation on it that its not to be used for a hottub...

 

[milkshakelake]

@Alistair Heaton Yeah, I was thinking of doing something like raising it and hiding the ballast under it. In case the point load isn't a problem, I had a few other ideas like filling the columns with concrete, using concrete sonotube type stuff around the columns, or adding weight to the top.

@skeletron I think I might be able to get away with calling it "temporary" because that's why it's not anchored. If the tenant/owner changes, they might want to remove it. It's actually a lot less trouble to just anchor it. The code in my area doesn't have the 10' x 10' provision, but it's good to have that context...it makes things a bit less arbitrary. And I'm not sure if 12th floor stuff has more consequence. If it falls over, it has a lot more velocity and momentum than from a lower height. But at that point, the danger to the public is almost the same as if it fell off a 4 story building.

@dhengr Agreed. I was a teenager once, and I'd cause chaos for no reason except to cause it.

@HTURKAK 100% agreed. The surface finish is concrete slab. I'd throw welded/formed concrete ballast over the base plate and waterproofing over it, or just have the base plate be heavy.

 

[BAretired]

To build anything on a roof without anchorage is a foolish and dangerous idea. Forget it!!

 

[JStephen]

It just seems logical to me that anything on a roof would be securely attached. Part of that comes from seeing videos of some AC units blowing off a roof in a hurricane.
However, I can't find any such requirement in ASCE 7 or IBC. Googling the question, I did come up with this item:
"LOAD TRANSFER TO THE STRUCTURE
"Once the appropriate (wind) force is determined, a design can be completed to resist the forces or transfer the load to the building structure. For wind loading, this can be accomplished by ballasting or by mechanically anchoring the equipment to the structure. (Seismic loading must also be considered, which, under certain circumstances, can prohibit the use of ballasting.)
"Ballasted systems can be a very attractive option to address wind loading..."

Article:

https://roofingmagazine.com/wind-loading-rooftop-equipment/2/

I don't know that the article is authoritative on the question, but at least it shows that nailing everything down is not the universal practice.

 

[KootK]

Where it's palatable, a strategy that I like is ballast with a tether as a backup. For your pagoda, that may have undesirable implications for aesthetics and trip hazard.

 

[milkshakelake]

@BAretired I think it could work if designed properly.

@JStephen Thanks! Good article and research. I appreciate it.

@KootK Agreed. I think I can put some kind of hidden cables under the roofing. Not sure if the owner and/or architect will buy it, but it's a good idea.

 

[CrabbyT]

They make ballasted handrails, not sure why you couldn't apply the same principle to a purgola. Seems like as long as DL + WL results in an upward reaction, and as long as you have some way or resisting horizontal force so it doesn't slide off the roof, then you should be OK from a statics standpoint. Not sure why you'd take on that sort of liability though. Seems a penny wise and a pound foolish if you ask me.

If you don't have to anchor an umbrella, then why this?

Umbrellas definitely should be anchored
though. I don't care what the code doesn't say about it. Umbrellas kill a lot of people.

https://youtu.be/JEMvwr-vQUI

According to this source, 2 million people die worldwide each year from loosely secured canopies.

https://www.rmdlaw.com/personal-inj... 2018, a,to loosely secured outdoor canopies.

 

[BAretired]

There is no good reason not to anchor the bloody thing. It goes against all common sense. I have witnessed a similar structure blowing away on our golf course in a strong wind because it was not anchored down. Fortunately it did not hit anyone, but it could have.