Torsion in Nonprestressed Concrete One-way Joists Systems 1

[AKM30]

Do nonprestressed concrete one-way joists beams that meet the requirements of ACI 318-14 Section 9.8 need to be designed for torsion? My structural model indicates that the members are over the threshold torsion, but the model does not recognize the members to be part of a one-way joist system. If the joists are designed for torsion, the result is lots of closely spaced stirrups due to the minimum spacing requirement of ph/8, per section 9.7.6.3.3.

 

[JAE]

Unless the structure depends on the joist torsional stiffness for stability I would ignore it.
Of the billions and billions of joists I've designed ( ) I've never considered torsion.

First - they tend to be fairly thin members. They also have small relative torsional stiffness relative to the surrounding beams and girders and finally it is difficult to fit in torsional stirrups in the thin webs.



Check out Eng-Tips Forum's Policies here:
faq731-376

 

[MIStructE_IRE]

Forget about the model for a minute and go back to free body diagrams. Where is the torsion coming from?

 

[rapt]

Most codes say if you are going to ignore compatability torsion, you should still provide minimum torsion reinforcement.

Part of the problem is that torsion stresses will increase your shear reinforcement requirements, even if you ignore torsion in design, as on one face torsion + shear are additive.

 

[hokie66]

Torsion is shear, but I don't know how you provide stirrups in one-way joists that would be effective in resisting torsion. Agree with JAE...It isn't done.

 

[AKM30]

MIStructue IRE, Good point. I am not sure where the torsion on the joists is coming from in my structural model. A free body diagram would not indicate any torsional forces.

 

[KootK]

I am not sure where the torsion on the joists is coming from in my structural model.

1) It's pretty hard for a CIP system, one way or other wise, to not behave a bit like a two way slab in the real world and in a 3D model. And two-way slabs have twist/torsion. So your ribs are going to pick up some twist/torsion. It makes sense to me that you would see some of that.

2) For these kinds of systems, there's always an interplay between flexural stiffness and torsional stiffness. I believe that the stiffer you make your system flexurally, the less torsion your ribs will pick up. Which is not to say that there's anything wrong with your current design. I agree with JAE's take on this although I haven't yet come up with a good argument for why the torsion should be ignored other than a) we always do that and b) nothing ever seems to come of it.

3) Well, maybe I do have one plausible argument for ignoring the torsion. The torsion in the stems would have to be transmitted to those stems via flexure in the topping slab. I suspect the topping slab would crack and shed stiffness long before you picked up an appreciable torsion in your ribs.

Out of curiosity:

1) What's the joist span?

2) What's the joist width?

3) What's the total depth of the system.

4) What's the width of the floor panel perpendicular to the span?

5) What is the spacing of the ribs?

6) What support conditions have you modelled around the perimeter of the floor panel? CIP girders?

 

[JAE]

KootK - Your reason number 3) for sure. There are, however, typically orthogonal distribution ribs in joists of some span distance so there is more stiffness there than just the slab...which for most typical pan joists at 3 ft. o.c. max. is a 4 1/2" thick slab with wire mesh.

But to start analyzing one-way joists for some inherent torsion, that is very difficult to model accurately due to varying and uncertain, partially cracked elements, is a waste of time in my book....so I also agree with your point 2) - we always do that and nothing ever seems to come of it.

Check out Eng-Tips Forum's Policies here:
faq731-376

 

[KootK]

There are, however, typically orthogonal distribution ribs in joists of some span distance so there is more stiffness there than just the slab

That is a fine point which I had failed to consider.

 

[AKM30]

@ KootK:

1) What's the joist span?

3-span continuous beam w/ exterior spans = 17'-3" and interior span = 35'. The interior span has bridging perpendicular to the joists at the 1/3 points.

2) What's the joist width?

7"

3) What's the total depth of the system.

18" (6"Slab + 12" rib)

4) What's the width of the floor panel perpendicular to the span?

18'-10" between column centerlines = girder span

5) What is the spacing of the ribs?

26" o.c.

6) What support conditions have you modelled around the perimeter of the floor panel? CIP girders?

The joists are supported by CIP Girders (30" wide x 18" deep). At column lines parallel to the joists, the joist rib width is increased to form a larger CIP beam (16" wide x 18" deep) which also serves to make the rib spacing geometry work out.

 

[KootK]

Thanks AMK. That all sounds pretty reasonable.

18'-10" between column centerlines = girder span

So here, you aspect ratio on the panel is nearing 2:1 with the short dimension perpendicular to your joists. That is indeed a situation where I might expect some unintended two way behavior that might induce torsion in the stems. There are beams, girders on these column lines, yes? If not, I rescind my comment.

 

[AKM30]

Correct KootK, there are beams and girders on the column lines.

 

[TLHS]

I agree with JAE, but you should also adjust your model so that loads are not being transmitted via torsion. What is currently being transferred as torsion may need to be carried somewhere else, so unless you can qualitatively say that the loads are not significant, you should include those forces elsewhere by releasing the torsional restraint on the joists.

 

[AKM30]

TLHS, this seems like a good idea but I am not sure how to apply this in my RISA 3-D model. It seems RISA limits the release of torsion to one end of the member. Do you have any suggestions on how to release the torsional restraint on the joints?

 

[jayrod12]

If you release it at one end, that should be sufficient. If you release it at both ends you'd end up with an instability where the member theoretically could spin on its axis.

 

[AKM30]

@ jayrod12, If I release the torsional restraint at one end, torsion still develops in the members, it just goes to zero at that end. Perhaps the torsion is generated through top slab (modeled as 6" concrete plate elements) or through the bridging in the interior span? Even if I remove these elements from the model, RISA still indicates torsional forces in the joists.

 

[jayrod12]

Then that means that something is providing rotational restraint along the length. Otherwise the joist should just rotate to relieve the torsion.

Care to post your model?