Egress stairs in seismic regions. ASCE Wants to have their cake and eat it too.

[driftLimiter]

I've been wrestling with a particular code provision regarding egress stairs for a while now.

The in Ch 13, ASCE 7 gives two options for anchoring stairs to the main building:

Provide a slip connection.
Or
Rigidly tie in the stairs.

They also provide the required anchorage forces in Ch 13 (Fp).

The way I see it if you provide a slip connection, then you aren't able to resist the required Fp loads until all of the slip is taken up.
If rigid tie in, it says that the stair framing must be included in the MSFRS, in practice I have never seen this done.

Most of the time the EOR says anchor per chapter 13 and provide slip connection capable of 0.02 x Hsx displacement.

So the only feasible option I see is to provide a slip connection with a keeper system capable of resisting Fp.

Does anyone else have an opinion or references that can help clarify this?!

 

[Greenalleycat]

This is a big lesson learned from the 2010/11 earthquakes in Christchurch
Google things like "stairs fell out christchurch earthquakes 2010" and you'll find a bunch of different articles on it

https://www.stuff.co.nz/the-press/6478239/Damaged-stairs-not-inspected

https://canterbury.royalcommission.govt.nz/documents-by-key/20120223.3284/$file/BUI.COL764.0162.pdf

Here are a couple of links to get you started on the lessons we learned

https://www.building.govt.nz/buildi...ress-stairs-earthquake-checks-needed-for-some

https://www.sesoc.org.nz/design-res...-systems-following-the-canterbury-earthquakes

Download V11 then look at Page 48 and beyond

Basically, there were some very evident issues with stairs in the earthquakes

1) They are very axially rigid members so they behaved as struts between levels (same with carpark ramps) and caused unintended building behaviour
2) As a result of (1) they ended up buckling etc and many staircases fell out - people were literally abseiling out of high rise windows with fire hoses and all sorts as there were no stairs
3) Even stairs that were designed to slip previously had poor detailing - not enough deflection allowance, and often had small slots at the base that filled up with debris or were filled up by maintenance staff with grout etc, so they acted rigidly anyway...leading to (1) and (2)

Stair design has now massively changed in NZ

Key differences
i) Stairs are almost always fixed-sliding detailed, typically fixed at the top (or at least pin-sliding, either works) with a sliding detail at the base
ii) The sliding detail is now typically a bearing strip that the base of the stair lands on, rather than a pocket - this prevents (3) above
iii) The detailing at joints is significantly increased to provide hogging moment capacity at the upper knee in case of axial load causing the knee to pop

I don't know what specifically 'Fp' is as I can't access ASCE standards, but the general principle is completely legit
I would just design them with a bearing strip at the base of your landing strip and give it ~200mm of seating - then it probably won't fall off

 

[lexpatrie]

For those not in the know.... it's rapelling outside the U.S. (yeah I've seen it before but didn't recall immediately)

Abseil

 

[271828]

For the option of providing a slip connection, I thought the idea would be to resist Fp at the non-slip end. Is that not the case?

 

[Greenalleycat]

I see now Lex's post was to translate to American, not to nitpick on the specifics of abseiling a building vs a slope...
(to be fair, Lex's post was edited as I replied)

 

[Deker]

The way I see it if you provide a slip connection, then you aren't able to resist the required Fp loads until all of the slip is taken up.
If rigid tie in, it says that the stair framing must be included in the MSFRS, in practice I have never seen this done.

The stairs need to be rigidly tied somewhere to keep them from sliding off the floor. It's common to provide a rigid connection at the top and sliding connection at the bottom. If rigidly tied top and bottom, the stairs need to be designed for the forces that arise from deformation compatibility.

 

[driftLimiter]

Okay this is making more sense.

So if a single stringer spans from floor to floor we can rigidly connect it to one of the floors (design for the full anchorage load that is tributary), and have a slip connection at the other floor capable of accommodating the story drift.

For some reason I was stuck thinking that the slip needs to happen at each connection effectively leaving the stairs free for lateral loads.

 

[Struct123ure]

I have a follow up question:
Do precast mid landings and (in some cases) precast floor landings need to be rigidly connected to the core wall or placed to allow sliding deflection?
I agree if the stairs are rigidly attached to the landings and the landings are rigidly attached to the core walls "They are very axially rigid members so they behaved as struts between levels (same with carpark ramps) and caused unintended building behaviour".

What if we have the stairs as sliding (top of stair rigid and bottom sliding)and the mid landings/landings as rigidly connected? Best practice would dictate to have it modeled as part of the LFRS, but has anyone modeled to see whether rigidly connecting the landings significantly stiffens the core?

 

[TLHS]

It depends on the system. Detail so that you don't have significant stiffness contribution from the stairs unless you're explicitly analyzing for it, basically. You can integrate landings, or whatever, but then you need to make sure you have slip in the stairs attaching to the landing somewhere. You could also potentially look at the core and prove that it's significantly stiffer than the stairs that are attached to it, so maybe you don't need to do anything.

A set of wood stairs attached directly to the concrete elevator core is going to be different than a set of steel stairs out at the edge of the floor plate, is going to be different than a set of steel stairs in a portal frame pre-engineered building.

 

[lexpatrie]

No worries Alleycat. I have a lot of internal blockage when it comes to terminology, so I have to mentally translate a lot of stuff to the "correct" (for me) term.

Alright, third edit to a trivially short post, but here's my substantive answer to the question.

If you want the stair to slip, you need the separation required between the main building and the egress stair (I suppose this gap may become a bit of an issue, but you can put a plate over it and attach on only one side, and it's not ADA, from the sound of it, (it's a stair, after all), so you'd structurally isolate the egress stair as a free-standing structure "inside" the building. That way the two "buildings" (the stair and the actual building) don't hammer each other during (is the term "pounding?" I forget) an earthquake but people can still use the stair.

Alternately, it's attached to handle whatever force gets delivered to it, or you somehow detail it so it's able to impart load on the main structure, (sort of like a non-structural element, parapet, sprinkler system, etc), yet not accept the large force when the structure tries to impact the stair. (That seems to be what the New Zealand situation is getting at).

 

[driftLimiter]

@lex if you have a full gap from the main building to the stairs then the stairs needs to be self-supportive for lateral loads (Ch 15)
I was moreso talking about the case where you anchor the stairs to the structure via Ch 13 loading.

In order to accomplish sufficient anchorage and ensure the braces don't participate in lateral resistance, you would anchor one end to the structure rigidly, and at the other end have a slip connection.
This in my view satisfies Ch 13 requirements.

 

[Greenalleycat]

I would detail for slip all day long, it's so easy
If you don't, it's a double whammy of having to detail your stair to take a potentially huge axial compression load, and also to design your building for the impact of these random tension-compression struts that are probably in awkward places

Here are two stairs that I've designed that show options - 1 is pin based and sliding at the top, the other is fixed top and sliding at the base

P.S - bonus points if someone can figure out what I very nearly fucked up in the first Detail.... it's not related to sliding