Preventing Torsion

[DHKpeWI]

I am designing a bent C channel spanning 14'-0". The channel is loaded by hat channels on the bottom flange. See attached drawing. Since I am not loading the C channel through the shear center the loading will create torsion in the C. I think I can resist the torsion by welding the hat channel to the C. What do you think?

Thanks

DHK

 

[DHKpeWI]

The C channel is 10 inches deep and the hat channel is 3 inches deep.

 

[JedClampett]

Are the hat channels supported at their other end?

 

[DHKpeWI]

The other end of the hat channel also bears on a C channel, same of the same size.

 

[JedClampett]

In this case I wouldn't worry about the torsion as long as your hat channel is adequately designed.

 

[GregTirevold]

Open sections like channels are really, really poor at resisting torsion. The calculated polar moment of inertia is deceptive with these shapes since the shear forces are no longer uniform and standard Tc/J formulas no longer give valid results.

-----------------
An example from Design of Weldments by Blodgett:

A pair of C6x8.2 connected with cross members like ]_[ subject to 1000in-lb or torque have an angle of twist of something like 2 degrees.

Put those same two channels back to back like this ][ and subject them to the same torque and your angle of twist drops to like 0.8 degrees.

Put the same two angles toe to toe like this [] and subject them to the same torque and the angle of twist drops to like 0.02 degrees.
-----------------

So the basic idea is you want to approximate a "box" shape as much as possible with your torsion members. If you can't do that then you need to attach your left and right side channels with diagonal cross bracing, not straight bracing that resembles a bookshelf. The reason being that diagonal braces transform the twisting of the sidewalls into compressive and tensile loads much like beam in bending or members of a truss.

In your picture the top flange of your 10" channel is basically unrestrained and welding the bottom 3" of the web to a straight cross brace isn't going to improve the situation much. If you added another hat channel to the underside of the top flange you would get an approximate "box" section and stiffen the frame quite a bit as well.

 

[DHKpeWI]

Thanks guys.

Unfortunately I cannot install a hat channel at the top flange of the C channel.

 

[DHKpeWI]

I need to use the C shape for a variety of reasons. I intended to weld the hat channel all around to the C channel. This will definately provide some torsional resistance. The maximum torsion I need to restrain is less than 2,000 lb*inches.

 

[graybeach]

Can you weld stiffeners to the web and flanges of the 10" channel, and make a moment connection between the stiffener and the hat channels? If you do that the torsional rotation of the 10" channel will be limited to the bending rotation of the hat channels, and you should not have a problem.

 

[csd72]

eccentric load does not always result in torsion particularly if the ends of the beam are not torsionally restrained.

Normally what results is a small rotation of the beam to a point where biaxial bending resists the force. In your case the fact that the load is hanging from the bottom of the beam would tend to resist the rotation and minimise the twisting.

 

[JedClampett]

I guess I disagree with most of you. In my opinion, as long as the hat section can span the distance, the torsion on the channel can be ignored. It's no different than any other steel framing. There is some torsion that is somewhat of a secondary effect due to connection eccentricity. The open members twist negligibly, but that is allowable in pinned framing.
As long as you're not cantilevering off the open section, you're OK.

 

[csd72]

agreed, most of the time channels are designed with no consideration for shear centre.

I was just trying to give a scientific explanation.

 

[Lion06]

This shear center idea always gets me. I asked the question here once before about a channel loaded through it's geometric center. The channel is welded along both flanges and the web at both ends such that the connection weld mimics the section. Now the channel itself should see torsion, but the connection won't. Now, if you load the channel at the shear center, but keep the connection the same then you have no torsion in the section, but you do in the connection. How do you get twist at the connection with no twist on the section?

 

[paddingtongreen]

SEIT, if you weld the flanges, you prevent them from rotating about the vertical axis, one wants to go one way and the other the opposite way.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[hokie66]

I would turn the channel around and connect to the web. That way, your load is closer to the shear centre. If you don't want to do that, use a closed section.

 

[prex]

StructuralEIT, the welded ends of the channel will have a shear and (coinciding) torsion center very close to the one of the channel, so also at the ends a load through the shear center will produce no net torsion.
What others didn't perhaps say clearly, in stating that there is no need to account for torsion of C beams, is that this is true because the torsional deflection is generally resisted by cross members (of course not true for a lifting beam, as an example).
So my answer to DHKpeWI will be: if you don't weld your cross members (or hat channels as you call them, don't personally know the meaning of this nomenclature), you are at a risk of having the cross beam not well supported on the lower flange, because of the torsional deflection: with high or limit loads, the failure mechanism here could be with the cross beams losing their support and falling down.
If on the contrary you weld the cross beam to the main ones, then you can forget about torsion and shear center.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[Lion06]

prex-

Are you saying that a C-shaped weld has a "shear center" in a similar location to that of the Channel? I'm not sure if I believe that. If that is true then the C-shaped weld tables for AISC are all wrong, no? If that were true then the eccentricity would need to be measured to the shear center and not to the centroid of the weld group, but the tables all measure the eccentricity to the centroid of the weld group.

 

[paddingtongreen]

StructuralEIT, I rejected the idea when I first heard of it but I made a model from stiff paper, I cut it out in the curved shape of a deflected flat bar. Then I bent the top and the bottom as flanges and the thing became clear, the top and the bottom were forced in opposite horizontal directions, twisting the beam. The twist though, is relative to the ends.

A side effect is that the flanges rotate horizontally about the vertical axis at the ends, a warping distortion, as in torsional analysis, so welding changes the end conditions.

This tendency for the flanges to distort in opposite directions is a function of the bending moment, the instantaneous curvature is a function of the instantaneous bending moment, so there is a difference in the effects of point and uniform loads.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[BAretired]

The hat section can be welded to the channel to resist torsion, but the web of the channel may not be adequate to carry the torsional moment up to the top flange. This problem can be remedied if the web is stiffened at each hat section.

Turning the channel around as suggested by hokie is the best solution, but I assume there is some reason why this cannot be done.

I do not agree with Jed that the torsional moment is a secondary effect. It is a primary effect and a moment connection is necessary for stability if the channel is not capable of accepting the torsional moment. A secondary effect is one where torsional moment is not needed for stability such as the torsion in a spandrel beam caused by joists framing into one side.

BA

 

[slickdeals]

It is pretty amazing that there are still a few structural engineering topics where people seem to have different approaches and methodologies (makes it fun I suppose). The other one that instantly comes to mind is that unbraced length issues related to inflection point. Wish there was closure to these issues........

We are Virginia Tech
Go HOKIES