Torsion and Flange Warping Restraint
Torsion and Flange Warping Restraint
(OP)
So, I thought I had a grasp on this -- but recent questions from a colleague have me wondering again.
The base of the question is when do warping normal and shear stresses develop in an open cross-section beam with torsion? For the sake of simplicity, let's consider only simply supported beams.
AISC Design Guide 9 goes to great lengths (p. 3, 11, 12, 14, etc) to express that when a member is allowed to warp freely, warping stresses do not develop. I've taken this "warping freely" primarily to be in the absence of fixed-style flange end connections, flange stiffeners or added webs at the flange tips, kickers, etc. (See WillisV response in http://www.eng-tips.com/viewthread.cfm?qid=340161)
But then, in the design examples and torsional functions, members that are flexurally and torsionally pinned (presumably a simple shear tab connection) are shown to develop Θ" and therefore warping normal stresses along the length of the members.
The best explanation I can think of is that the rate of change of twist itself is restraining each cross-sectional slice of a beam with respect to the adjacent slices. But if this is the case, and I'd imagine it would be the case in nearly every structural application, why would AISC even discuss a mythical "freely warping" beam?
What am I missing here? Thanks!
The base of the question is when do warping normal and shear stresses develop in an open cross-section beam with torsion? For the sake of simplicity, let's consider only simply supported beams.
AISC Design Guide 9 goes to great lengths (p. 3, 11, 12, 14, etc) to express that when a member is allowed to warp freely, warping stresses do not develop. I've taken this "warping freely" primarily to be in the absence of fixed-style flange end connections, flange stiffeners or added webs at the flange tips, kickers, etc. (See WillisV response in http://www.eng-tips.com/viewthread.cfm?qid=340161)
But then, in the design examples and torsional functions, members that are flexurally and torsionally pinned (presumably a simple shear tab connection) are shown to develop Θ" and therefore warping normal stresses along the length of the members.
The best explanation I can think of is that the rate of change of twist itself is restraining each cross-sectional slice of a beam with respect to the adjacent slices. But if this is the case, and I'd imagine it would be the case in nearly every structural application, why would AISC even discuss a mythical "freely warping" beam?
What am I missing here? Thanks!






RE: Torsion and Flange Warping Restraint
Whenever these types of questions come up, I always like to point people towards figure 4.4 and section 4.1.4 of the design guide. This give you a nice physical analogy for how warping stresses work. And, using this analogy you can come up with really quick approximate calculations rather than the crazy complex ones described in the rest of the design guide. I also use it to calculate warping stresses for situations that just aren't covered in the design guide.... Like situations that are somewhere between warping restrained and warping unrestrained.
RE: Torsion and Flange Warping Restraint
Still, the question remains -- do warping stresses always apply for I-shape members? And if so, why discuss free or restrained warping at all?
RE: Torsion and Flange Warping Restraint
Think of a moment connection with warping restraint at the end. This could be a pull pen moment connection with continuity plates in the column. This connection will develop large warping normal stresses at the location of maximum bending stresses.
Alternatively think of a moment connection where the flanges are free to warp. Maybe a case without continuity plates and where the column flange is relatively thin and can't provide much differential restraint to those beam flanges. This warping "unrestrained" moment connection will develop warping stresses in the member more towards the mid-point of the member. That's a location where the bending / flexural stresses are maybe 50% of what they would be at the connection location.
There is a tendency in our profession to ignore warping stress for simply supported beams. Typical example would be a clip angle connection. It provides torsional restraint, but not warping restraint at the end of the beam. There will be warping stresses along the length of that beam. Though many engineers ignore it. Not saying that's right or wrong. Just that I don't see many folks considering this.
To me, it's an engineering judgment call depending on how significant the torsion is. I personally prefer to use the flange bending analogy as a quick check to see how significant it is compared to the flexural stresses. Then I decide whether to include or ignore it.
RE: Torsion and Flange Warping Restraint
But after some more research, I found the SCI guide for Design of steel beams in torsion, which confirmed that the flange restraint can come from external sources (connection details) or develop internally (cross sections restrain each other when torsion varies along the length of the member). Also had a handful of other useful charts. I recommend it.
Thanks for your help Josh!
RE: Torsion and Flange Warping Restraint
Almost, but not quite, always. See the sketch below for a highly contrived example of a beam that is able to warp freely (relatively speaking).
For the most part because a discussion of torsional theory would be incomplete without it. To a lesser extent because there are a few, semi-practical cases of unrestrained warping torsion.
Additional commentary
- if a beam is torsionally fixed (encasement, full pen continuity plates, etc) at any point, it will not be able to warp freely and will develop normal and shear stresses as result of warping.
- if a beam is torsionally pinned (conventional shear connections etc) at two or more locations, it will not be able to warp freely and will develop normal and shear stresses as a result of warping. This is the part that throws people. The beam in question will be warping restrained between torsional supports in the same sense that a simply supported beam, symmetric in all respects, is effectively fixed for strong axis bending at mid-span (zero rotation).
- Beams that are able to warp freely are pretty rare. That's why one gets the impression that it's a case hardly worth discussing. Fee warping is generally precluded by one or more of the following:
1) Code specified limitations on where beam rotational restraint must be provided (LTB stuff).
2) Established practices that form the basis of sound detailing.
It's worth noting that the example shown in the sketch below would have a very high effective length factor with regard to lateral torsional buckling capacity. That is a direct consequence of the fact that the detail permits free warping or something close to it. I would expect that to be the case for virtually all open section beams that are allowed to warp freely.
I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.
RE: Torsion and Flange Warping Restraint
I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.
RE: Torsion and Flange Warping Restraint
RE: Torsion and Flange Warping Restraint