Is support condition same for both flexure and torsion?

[faromic80]

I'm analyzing a channel supporting a uniform load. i've assumed a C8x11.5 spanning 11.5'. the channel supports a metal deck. the channel is the end support and the span to the adjacent beam is 4'. I'm designing the channel as a simply supported member in flexure. The channel will be connected to the in-place columns with L3x3x1/4 angles welded to both the channel and column web. My question is, is the end support condition the same for torsion. I'm thinking, the angle and welds carry only shear and are flexible enough to model a ss end condition. When putting torsion on the three sided weld between the angle column web, isn't the connection fixed? This is a pretty big difference since the theta prime factor is much larger for simply suppored spans. Also if the channel is unbraced in the 11.5' span, only pure torsion would occur, correct? Another question I have regarding the tables in the back of AISC design guide 9 - the torsion at the end of the fixed span is 0 and the torsion at the end of a simply supported span has a value. Shouldn't this be the opposite? Why does a fixed condition have 0 torsion at the ends (See case 7 for theta prime).

 

[Lion06]

The difference between torsionally fixed and pinned is the ability to warp, not the ability to support a torsional reaction. It has to be able to support a torsional reaction or it will be unstable. From everything I've read, it is extremely difficult to get a torsionally fixed condition.
Also, when designing your channel, be sure to use the eccentricty to the shear center and not to the centroid of the web.

 

[DaveAtkins]

I would ignore torsion for your situation. The deck which spans to another member 4' away will prevent the channel from rotating a significant amount, and will load the channel close to its shear center, anyway.

With a channel, the support is generally a single angle (not two angles like with a wide flange). The angle should not be welded to the column on three sides, but only along the bottom and vertical leg. Leaving the top of the angle "unwelded" allows the angle to flex, which means the assumption of a simple support is correct.

DaveAtkins

 

[asixth]

If the element has limited torsional stiffness (J), it will not attract torsion. C and I sections are relativly flexible when it comes to torison. If you release the torsional fixities at both ends of the beam in a 3D analysis will create a singularity.

 

[faromic80]

Dave,
What do you mean exactly by "close to the shear center". Is e0 (distance from face to shear center considered close? If the deck/slab bears on the entire flange, the distance to the shear center would be half the flange width plus e0, right? Is this considered negligible?

You said not to weld the top of the angle to the column to approximate a simple shear connection and allow some flexibility. Is this also the case for attaching other members designed as simply supported to columns/beams. I've never seen this done. From what I remember, I always saw 3 sided welds for WF members connected to a supporting member. Correct me if I'm wrong?

asixth: Are you saying that if J is a low value then the member won't attract torsion? If the member has a low J, doesn't that mean it can't resist torsion as well as another member might be able to and yield/fail faster than a member with a higher J might?

 

[Lion06]

what asixth is referring to is what is known as compatibility torsion. If the torsion is there only as a result of member deformation, then it is true that a member with low J won't attract much. If, however, it is equilibrium torsion, then it doesn't matter how torsionally stiff the member is (or isn't), it must support that torsion.

About the 3-sided weld on the angle. You should weld 3 sides of the angle to the supported member, but not to the supporting member. That connection relies on the flexibility of the angle to act as a simple shear connection. If you lock up the angle by welding all 3 sides, then you don't have a simple shear connection anymore. That being said, I would not leave the top of the angle completely unwelded. I would return the weld on the top for a length of 2*weld leg. The max allowable per AISC is 4*weld leg.

 

[DaveAtkins]

I agree with asixth and StructuralEIT. Regarding the shear center comment, maybe I am wrong--maybe the distance is not negligible. But I dare say most engineers would ignore torsion in this case--again, because the channel is torsionally flexible.

DaveAtkins

 

[faromic80]

By torsionally flexible, does that mean the polar moment of inertia is low? So that means that if the channel/member was torsionally stiff or relatively stiff (higher J), torsion has to be accounted for? What constitutes a non-flexible channel/member?
sorry if some of these questions are repeated, I'm just trying to grasp the concept.

 

[Lion06]

The torsion always has to be accounted for. What asixth was getting at is that if the torsion is a result of member deformation (and not of equilibrium) that a torsionally un-stiff member will not attract much torsion.
I would explain it this way. If you have a single beam framing to a spandrel that happens to have a J of 0 (just for demonstration purposes), the spandrel will provide no torsional restraint and the beam will be free to have a simple end rotation.
If the beam is framing to a very torsionally stiff member (say a HSS 16x16), then that spandrel will provide end restraint for the beam (meaning the beam will have an end moment) that must be resisted through torsion in the spandrel.