Walking Column and Resolving Forces 1

[Struct_Matt]

Walking columns and associated slab diaphragm forces: We have a condition on one face of a building that requires walking columns back from their position at the upper levels (residential units) to their position in the garage below.

That said, we are using the floor to floor resistance to resolve the eccentricity induced by the walking column. The questions really boil down to these (large concentrated) forces at each affected diaphragm level and how people have gone about handling them.
1. Need to develop and detail the mechanism that transfers the forces into the slabs. There are pretty straight forward ways to handle this with added reinforcement, but I think it becomes a little more delicate when introducing these forces into a PT slab system.
2. Need to properly distribute these forces into the diaphragm so that they can then be resolved back into the buildings shear walls.

Any words of wisdom (details of reinforcement, etc) from engineers who have actually implemented this kind of thing in one of their own building designs would be appreciated.
Please see below for a sketch of the circumstance along with a screen shot from the RISA FEM model used to determine the resultant slab forces needed to resolve the eccentricity of the column walk. All forces from the RISA model are in kips.

 

[JLNJ]

Interesting problem. I have not designed a building with walking columns like this.

Does the flexibility of the diaphragm and the shear walls play into the column shear load resolution (as shown in the RISA reactions) or is the diaphragm so stiff it has no effect?

 

[Trenno]

Can you form a strut/tie diaphragm truss that is 5' deep that spans between the two cores?

Draw the truss out such that the resolved walking column forces occur at a truss node.

 

[bookowski]

I've always seen/done this by a simple resolution of force P x e force by dividing the overall walk height, So T=C=Pe/d. If you're going to try and model it as you've done and not do the above simple Pe/d I'd recommend removing those horizontal reaction points and replacing with springs (good luck guessing at spring stiffness but may not be so important as long as relative to each other they are reasonable). The weird back and forth reactions you are getting are a result of the reactions you've modeled, they don't appear correct.

For large walks i've seen vertical shear checks but yours doesnt look big.

These seems like there should be 20 different checks but I've always seen them done with little in depth check. This looks like a small walk so I wouldnt worry much. Just a couple of things:
- you're obviously inducing tension into your slab at one end of it, this should be accounted for in terms of cracking/deflection etc.
- You want to drag the tension force back far enough to get it into the core (probably obvious point)
- I've seen recommendations to limit the tension steel stress to some relatively low value, so rather than check the tension as .9AstFy for example maybe you do .5AstFy etc to try and limit strains/elongation/cracking

 

[Struct_Matt]

bookowski, Thanks for responding to the thread here, appreciate your thoughts. The hand calc version of the resolving couple is something we looked at quickly but it appears (based on FEM) that by lengthening the walking transfer zone to multiple stories it does bring the forces down. For reference, Pu = 1840 kips, e = 5ft, d = approx 10 ft/story. So by the simple form, T = C = 920 kips with a single story transfer. Would like to maintain 8" typical PT slabs (with some special detailing) throughout the transfer zone. With the 3 story transfer (RISA output linked above) the highest forces are 347.5 kips (C) and 247.4 kips (T). These forces seem more manageable with the 8" PT slab system (but still the crux of our thread here). A 2-story version of the FEM transfer zone indicated 406.9 kips (C), 160.7 kips (T) and 250.7 kips (T) at the 3 diaphragms bounding the transfer zone. We are hoping to keep the thread focused on the diaphragm force transfers and distribution.

 

[Celt83]

does adding support at each node along the bottom of the column mesh you modeled change the floor loading by any appreciable amount, as you have it modeled now it has a bit more eccentricity as the support seems to only be at the far base node?

Open Source Structural Applications:

https://github.com/buddyd16/Structural-Engineering

 

[Struct_Matt]

Celt83 - The other nodes at the base of the column are vertical supports, they are just not selected.

 

[Agent666]

Thinking outside the box, at one level can the wall extend to the next column line to turn it into a deep transfer element between columns. Guessing architecturally it would be a bit of a stretch but worth considering as a possibility as it eliminates the horizontal loads on the diaphragm.

 

[WARose]

I'm not sure if I follow the supports (as shown). It's showing a single lateral restraint (per level). As far as the diaphragm goes, under lateral load, I guess that makes sense (to some degree).....but for a complete analysis, I would think you'd have to model all this together. The slab will attempt to restrain the column from rotating (from a variety of loads).....and that would induce some high slab shears and moments. It may also (in some circumstances) increase moments/shears in the "walking" column.

For joint design, I'd see ACI 352.

 

[bookowski]

When I say pe/d the d is the total height, so in your case 3 stories or 30ft. so T/C would be 306kips.

You can't model a restraint at each floor, this doesn't make sense. You could model a spring at each floor, but not restraint.

We usually specify steel that is in addition to the flexural slab steel, typically larger diameter bars and specified at mid-depth and going back/far as you need to form load path back into the core and resolve it.

 

[formi]

Interesting topic (walking columns) but it's hard to find any publications/guides on how to design it correctly.
WSP uses this concept very often and I managed to find their 'Technical Reference Manual' on walking column design. They do suggest modelling it to derive secondary moments in columns above and below the "walk".

In terms of height of walking column, I heard that the rule of thumb is: height of wall shall be 3 (or more) times the distance that you are walking...

Any thoughts on this?

 

[bhiggins]

I've never handled this situation in practice but this is my interpretation. If there was no diaphragm, then all of the columns will have to resist 100 percent of the moments induced by the Pe effect. Your diaphragms will provide restraint against Pe moments, but the diaphragms are not infinitely rigid thus will only partially reduce these internal column moments. The columns closest to your shear walls will have the most restraint and will transfer more loads to the diaphragm vs. the columns farther away from shearwalls due to diaphragm displacement. It seems like you will need to make a 3d model with your appropriate diaphragm and shear wall stiffness to accurately get diaphragm forces and internal column moments. The more flexible the diaphragm and shear walls are, the more internal column moments you will have.