Timber Detail for Torsion Due to Eccentricity

[drlt5]

Hi all,

Some of you may be interested in this problem or may even have already encountered this and had a solution. The question is about a clever and efficient detail to avoid unexpected torsion and splitting on a timber beam. The timber beam in question is taking an unequal loading with an eccentricity. You can see the condition on the attached basic sketch and the model to understand the geometry.
Left hand side is CLT slab and on the right hand side timber rib panels are framed into this hardwood laminated beam known as Baubuche. The unequal loading at the Baubuche beam is causing torsion and the same unequal loading is eccentrically introduced to the Baubuche beam. I thought about having a tension strap at the top that is screwed into CLT, Baubuche and ribs so that everything would be tied together but more I think about it actually the tension tie does not do much when the Baubuche beam rotates creates a compressive stress at the top of the ribs, and tension at the bottom and would do the same on the left hand side and would eventually split in my opinion.
Have you guys had this kind of an issue and maybe even already thought about it and came up with an efficient way to solve this ?

 

[XR250]

I agree, a tension strap would not be all that effective. I assume the beam can't be installed under the ribs?

 

[drlt5]

@XR250 Correct. That is not possible.

 

[phuduhudu]

My first thing would be to check if you can get enough moment at the junction between the column and the beam to counter the torsional moment by taking width/6 of the column x the load. If that doesn't work you could put a strap on the tension side of the column to the side of the beam to get more capacity. Then the other thing to look at is how much moment capacity you can get at the rib to beam connection. You might be able to get enough to take out the half beam width eccentricity

 

[KootK]

Wood members in torsion are terrible things, truly. Get any mid-section longitudinal checking or cracking and your torsional stiffness and strength drop drastically. To my knowledge, we don't even have codified methods for checking torsion in the US other than resorting to the Timber Construction Manual. And I suspect that this is precisely why. Admittedly, it's been some time since I've looked into this.

I would expect glulam to perform better than sawn timber. How much better I could not say.

That said, I feel that the best approach with something like this is:

1) As much as possible, keep your CLT to beam connections pinned. If you moment connect the CLT to the beam in torsion, then you force a bunch of torsion into the beam as the beam is forced to twist to match the CLT end rotations (assuming beam supports that are not torsionally pinned).

2) Use only compression connections to the CLT, per the sketch below, to restrain beam twist arising from eccentric loading or LTB tendency.

This is a common strategy used in the design of precast hollow core planks supported by inverted tee girders. Similar principles.

 

[jayrod12]

I can't say I've ever known anyone to take this into account directly. But yet this unequal spans, or different framing schemes on each side of flush beams, is extremely common. Often due to the physical dimensions of the beams, they're inherently more stable for torsion. The one time I actually needed to check this was when I was hanging a folding partition off the side of a large glulam beam. The top of the beam was restrained by CLT slabs, so I checked the bottom of the beam as a wind girt spanning horizontally between columns. I checked the beam to column connections for the lateral load from the torsion. In my case, the columns were extended through the beams to the top of the CLT, so the column to beam joint wasn't as unstable as only a single level where the column would stop at the underside of beam.

Most column to beam connections do have a bit of rotational stability as well.

Kootk, no sketch attached.

 

[KootK]

The eccentricity associated with supported load delivery usually isn't much. As such, one can often make the argument that the "torsion" is really just eccentric shear. The danger in moment connecting the slab elements, I feel, is in inadvertently creating much greater eccentricity and, thereby, creating a situation where a true reversal of shear across the cross section is required.