Modelling Staircases in Softwares for Structural Analysis 2

[notpasible]

Greetings everyone!

I am fairly new to the industry and have little knowledge about the usual practice in designing structures. I am confused about how staircases should considered in structural engineering softwares. Should they be modelled as inclined shells, as a flat horizontal membrane, or by just simply leaving it as an opening (apply loads on adjacent beams)? Is it a case to case basis, and if so, when should we model staircases for each type and why?

I am looking forward to your recommendations, thank you.

 

[tmac7285]

It would depend on what you are trying to get from you analysis and how you plan to detail the element in question. Do you have a particular problem in mind? Can you ask a more experienced engineer in your office about it and what common practice is?

If you are considering a lateral stability system, we should be modelling what elements contribute to building stiffness in general, which is particularly important for EQ design. Practically however, in the case of stairs and ramps, these can be detailed to have sliding joints which would mean that they don't act as a load path to resist lateral forces. The Australian concrete code (AS 3600) for instance has tightened the provisions in the code meaning that either stairs are separated such that they don't contribute to building stiffness or they are modelled. This came out of the findings of the 2011 Christchurch earthquake where they had a number of stairs fail in buildings, though it's a fairly innocuous line in the code. It appears the intent is for us here in Aus to more closely consider our detailing, though current practice doesn't seem to have caught up yet, but I expect it will with time.

For vertical load, these should be quick and simple calculations for the majority of cases and software/spreadsheet will be able to get you the answer quickly.

 

[MegaStructures]

I second detailing sliding joints and then designing stairs in a separate model or by hand. If sliding joints aren’t detailed the stairs should be included in the stiffness model to ensure they aren’t picking up more lateral load than they can handle

“Any idiot can build a bridge that stands, but it takes an engineer to build a bridge that barely stands.”

 

[hardbutmild]

Will the stairs pick up any lateral load if the whole "stair area" is enclosed in a RC core?

 

[tmac7285]

hardbutmild,

I'm not sure I understand what you mean by the whole "stair area' being enclosed in an RC core. Do you mean that the stair is just internal to core? Why would this change anything?

If you have a diagonal element, such as a stair or ramp, that is rigidly connected to a diaphragm at the top and bottom (i.e. no slip type joint) then these diagonal elements will act in brace like manner. This is because the floor to floor drift cannot develop as per a core in flexure because the stair/ramp will attract axial loads as this becomes the stiffest load path for the lateral force to travel along. Adding diagonal elements to your model will drastically change the model's behaviour and you would need to detail the diagonal elements accordingly. As MegaStructures emphasises, if you don't put them in the model, but they are there in reality, they may well attract loads that they can't handle and fail. This is essentially one of the findings from the Christchurch earthquake.

 

[hardbutmild]

Do you mean that the stair is just internal to core?

Yes, I was thinking of that.
I understand the idea behind what you're talking about and I generally agree.
I was just wondering if the extremely stiff wall structure would change that, the idea being that walls will probably be very stiff on their own. I might be very much wrong, I'm bringing it up to try and understand the problem better.

It seems logical to me that walls should change the influence of the stairs since they can limit the drift and with that, stair activation. Look at the picture below. Is the stair activation the same for both case a and case b? I wouldn't expect them to be, but I might be way off.

I'm asking this because if I got the case b) or if I got the structure from the first picture, I'd simply ignore the stairs in lateral model, but in the case a) I'd definitely model the stairs.

 

[Trenno]

I think what you're trying to say is that stair cases enclosed within a stair core are analogues to a rigid diaphragm, ie as the walls are so stiff, then there will be minimal differential membrane strain and thus minimal diaphragm forces would develop?

 

[r13]

I think the force on the stair depends on the relative displacement of the floor and the adjacent landing. In a stiff core, the relative displacement should be quite small.

 

[CrabbyT]

I tend to model stairways and their support structures separately. That is, I model the stairway in one model and determine the reaction loads at the ends of the stringers. Then, I apply those reaction loads to the support structure in a separate model.

By doing this, it prevents the stairways from contributing stiffness to its supporting structure.

 

[structSU10]

Crabby - that is all well and good, but ASCE 7-16 has expanded seismic requirements such that if it is an egress stair if you perform your analysis as you indicate you must provide a slip connection for the stair to the floor.

 

[SocklessJ]

A while back I brought up this issue to my colleagues and got a lot of strange looks. The typical practice was to leave stairs out of the structural model, because they will suck up load and fail the stringers in compression.

I replied that in reality the structure doesn't care how you've modelled it, and now ASCE 7-16 is pretty explicit that the relative displacements need to be accommodated with sliding joints, or the stairs need to be in the model and withstand the overstrength load combos. Rather than come up with a sliding joint detail, I just put the stairs in the model. I typically work with industrial structures (braced frames) in low-seismic areas, so the drifts are relatively small. Weak-axis bending of the supporting beams helps accommodate drift, and we use beefy C10x15.3 and C12X20 shapes for stringers.

I was wondering if anyone could post a sliding stair stringer detail they've used with success. I was disappointed that the new AISC Design Guide 34 doesn't include any example details.

 

[structSU10]

do you run the check against a drift load combo using the drift amplification factor (0.7*3)? or apply overstrength to the forces?

 

[SocklessJ]

I just applied overstrength (omega) to the forces.

 

[CrabbyT]

struSU10 - thanks for the heads up, I had no idea. We haven't adopted ASCE 7-16 here yet, and I suspect this provision will go mostly overlooked in low seismic areas.

I can see why this would be a concern in a seismic event, but it would be nice to see some justification. Did a stairway or a structure fail because the stairs were too rigid and didn't slide? I mostly mean that rhetorically, but I's love to know the actual answer if you have it.

Often, the supporting structure is constructed well before the stairs are. I've wondered about this for a long time, but how is it possible to account for that in a structural model based on the tools we have today? That is, if I'm adding a set of stairs to a building that is already loaded, and if the building is deflecting under whatever loads exist before the stairs are implemented, we wouldn't expect the existing stresses to redistribute to the stairs. It's almost like designing shoring, or analyzing a beam encased in concrete.

 

[hardbutmild]

I can see why this would be a concern in a seismic event, but it would be nice to see some justification. Did a stairway or a structure fail because the stairs were too rigid and didn't slide?

As someone mentioned before, it happened in Christchurch (I believe in 2011, or maybe 2009). I stumbled upon a report regarding specifically stair collapse after that earthquake somewhere on this board not long ago (it's a few pages long, the stairs part).

how is it possible to account for that in a structural model based on the tools we have today? That is, if I'm adding a set of stairs to a building that is already loaded, and if the building is deflecting under whatever loads exist before the stairs are implemented, we wouldn't expect the existing stresses to redistribute to the stairs.

I wouldn't expect that stress to matter much, if we're still talking about seismic. Surely an earthquake won't come before the stairs were built and any other horizontal loads (wind) probably won't cause the structure to behave inelastically. So during the time the stairs are built, no (significant) horizontal loads exist.
But even if that's not the case, there is a possibility to model a phased construction process, although to my knowledge it's mostly used for bridges and I doubt it would (usually) be necessary for stairs.

Yes, I guess you put it better than I could.
That would also comply with what r13 said.
I wonder what some of the NZ guys say about it.

 

[r13]

I think there is a simple solution, no matter there is core wall or not. Set up monitoring joints in your model at strategic locations, then find the absolute displacement, and design the member for the required strength to resist the force due to displacement (for small displacement), or design the sliding device to accommodate the displacement. For the former, some engineering judgement is required