Flexural torsional buckling of column 2

I think so. If the shear connection were a shear tab, I'd be doubtful. But, with double angles, I would say that's a torsion restrain for the column.

I'm assuming these columns are part of a moment frame? Because a 40" deep WF column is quite large. Is that the case?

If you had floor beams framing into both flanges of the column with clip angles, I'd say yeah, probably braced. If the 'north' flange of the column is just hanging out by itself with no slab/edge angle surrounding it I might look at adding some sort of plan bracing to the flanges off of the top of the 'horizontal' beam. You said it's at a floor but also mentioned roof deck. That would also make a difference in my mind.

What magnitude of moment and axial is in the column at this location? What's your floor to floor height?

That said, with a column that big I'd like to have a more definitive answer than "yeah, probably braced", or "it depends", and get some input from your EOR.

Why doubtful with shear tab? Just trying to understand. what if the thickness of shear tab is designed for torsional moment from shear like extended shear tab.

Yeah, with a competent diaphragm in place, I would consider the column torsionally braced. You'll set up the resting couple shown below I would think. It might be a bit more questionable if the upper connection had slotted holes for fitup but, even then:

1) It still takes a bit to initiate sliding in a slotted bolt hole.

2) The column would only twist so far before engaging the slotted holes.

3) To a degree, both connections probably couple column torsion to weak axis beam flexure.

I like the setup a bit more if the deck is running from left to right.

@ dold

Why should both the flanges be braced? There is will be no twist as the columns is locked in both directions.

Thanks Kootk. That was my impression too. As the column is prevented from twisting in both directions, there is not requirement for both flanges of the column to be attached.

dold said:

Why should both the flanges be braced?

Click to expand...

Without question, direct bracing of both flanges is the most robust setup in my opinion. That is, after all, where the lion's share of the compressive load that you'll be bracing for resides if this is a moment frame column. Bracing the flanges indirectly by bracing the web, introduces additional flexibility into your bracing scheme, particularly in the absence of beam continuity stiffeners in the column.

dold was correct in his assumption about this being a moment frame column, right? Are these beams at an intermediate level that is not part of the moment frame or something? I was under the impression that this is a roof level which would normally have beams participating in the moment frame.

BAGW said:

Why should both the flanges be braced? There is will be no twist as the columns is locked in both directions.

Click to expand...

Well, that's what we're discussing: is the column actually locked in both directions, and what constitutes "locked". If this were at W10-W14 column (LTB/FTB not a problem anyway, probably), maybe even up to W24..., bracing through the web is probably workable. With a W40, the brace (perp beam in the web) is 20" away from the flange that wants to buckle, which my gut says is kindof a long way to bend a web out of plane and still be able to say that is a stiff enough 'brace'.

It depends on a lot of things that you haven't told us. Column size, is it part of a moment frame, internal forces, reason the column is so big, etc.

Just food for thought: in a special moment frame, both column flanges (at the moment connection itself) must be braced either directly (members or braces attached directly to both flanges) or indirectly (braced by a member perpendicular to the web, through the web, as is the case in our discussion here). Clearly this is not a moment connection, but the premise is similar.

I don't know that you'd have a Flexural-torsional buckling problem in the first place. Flexural-torsional buckling is for single symmetric shapes IIRC. (Channels and so forth.) Not "I" shapes.

Yes, this is a moment frame column. The beam framing that shown is at the floor level with a concrete deck. The beams are not part of moment frames. The beams are just a simply supported beams with a double angle clip connection. The flnage forces in the column are pretty high (close to plastic capacity of the clumn)

My concern is Lateral-torsional buckling per F2.2 which applies to doubly symmetric member as well

in my estimation, this would go a long ways towards convincing torsional bracing without adding a too much cost.

I like the end plate option. Otherwise, torsional bracing is iffy IMO.

Also, I agree with WARose. This is a Torsional Buckling mode, not a Flexural-Torsional Buckling mode.

My bad. I agree with WARose and 271828. There is no flexural-torsion as there is no lateral displacement. Its more of torsion problem.

I've played around with this a bit before using FEA and non linear buckling analysis. You really don't need much to stop the buckling but you do need something. Even a shear tab is sufficient in most cases even without a diaphragm. It is surprising how little stiffness you need.

Of course the moment in the column also matters. The last detail by Kootk would give me the warm fuzzy feeling that this is not something I need to worry about.