Seismic restraint of rooftop elements without physical anchorage to structure

[atrizzy]

I'm debating if it's even possible to seismically restrain something to a building roof without anchoring it to the roof structure (and penetrating the waterproofing, thereby angering some of my team members).

It's tail chasing to add weight and rely on friction, as my lateral % of mass is larger than the coefficient of friction I could get. Not to mention that it's the building that moves under the element, and not an applied load (as the code design simplification simplifies it to).

Any ideas? My specific example consists of a rooftop fan and light framed rain screen.

 

[skeletron]

You can't rely on friction to resist seismic weight for rooftop elements. This is a huge debate in CSA S37-18 since most rooftop antennas were pre-eng designed in the 2000s and very light structures with ballast.

You need to attach it to the structure in someway so it doesn't move or get punted off the building.

 

[HTURKAK]

Mr atrizzy ; you did not provide necessary info. regarding the roof detail. Let me guess; this is walkable flat roof with some walking paths and gravel. The membrane insulation under the gravel. The roof top is connected to the exhaust shaft with some length of horizontal duct.

If this is the case, just provide a pc slab free supported on the waterproofing. The wt of slab shall resist the wind and seismic OT moments .

In case of a large eq, the EF ( at most ) will change location just a few cm.

 

[atrizzy]

HTURKAK, actually the roof is just exposed membrane. I don't understand in your view how a precast ballast could resist the seismic forces if (for explanatory purposes), my Veq = 50%Wd, and my friction sliding resistance is 0.4Wd.

Care to explain?

 

[KootK]

The friction interface between the unit and the roof is both the source of the lateral load on the unit and the resistance to it. So, at worst, you have a balanced condition. Still, as you say, adding mass doesn't improve things numerically beyond that. Does the equipment need to be operational after an earthquake or just not slide off of the roof? If it's the latter, maybe you can tie it off to an elaevator overrun or something. Roof penetrations in the pacific NW are unfortunate.

 

[HTURKAK]

Mr atrizzy , I just estimated the walkable roof construction detail = Ballast to protect the membrane 2-3 in+ Membrane WP+ Heat insulation 2-4 in + vapor barrier + structural deck or slab . Now you clarified that the roof is just exposed membrane.

..and you do not want to punch the membrane to anchor the EF supporting legs..

I said

(PRECAST SLAB)[/b] slab free supported on the waterproofing. The wt of slab shall resist the wind and seismic OT moments .

In case of a large eq, the EF ( at most ) will change location just a few cm.]

I just proposed = The E.F. will be fixed on precast ftg slab with post installed anchors and the precast slab will sit directly free on waterproof membrane without fixing , so without punching of membrane.

With this set up, you will increase the friction coeff. (between concrete surface and roof membrane) around to TAN 30 degr. 0.57 . Moreover, the EF +PC ftg set up will attract less seismic forces ..

If my explanation is not clear, pls provide the roof detail and EF details . to get more detailed responds.

 

[JedClampett]

I usually try to sell this as, "it was time for a new roof anyhow."
How old is the membrane? Is it sound? It's not like a new roof doesn't have value.

 

[bones206]

If you're adding stuff on a commercial membrane roof, adding penetrations is pretty much expected and just part of the business... not sure why people would get overly upset about that. Just make sure you specify appropriate penetration details and compatible materials. Try to use a roofing contractor who knows what they're doing (as opposed to the HVAC guy or rain screen guy MacGyvering their way through it).

 

[skeletron]

What about using a glue/adhesive/epoxy?

Of course it would be have to be weather-proof (equally challenging problem); there may be potential flaws if the object was ever repositioned or removed. However, it may allow you to use friction outside of the bounds of NBCC (or similar code) which specifies that you can't use "friction due to gravity loads...to provide resistance to seismic forces." Instead, the friction force would be argued to be caused by the adhesive bonds.