Bracing ability for a tension bracing member?

[leeStruct]

Hi, ladies and gentlemen.

Please see attached file, 24 pieces of 20' W10x45 Wide Flange column are formed into a circle ring, every column is subject to a pure vertical load and both its ends are pinned. At its middle location, two 2x2x1/4" angle is welded to its flange to form a bracing member, the angle bracing member is 60" long. The question is what is the effective length of this W10x45 column? is it still 20' or it can taken as 10'?

The 1st argument is that based on AISC 13th, page 16.1-240 through page 16.1-242, when calculating the GA and GB value the bracing member is considered as compression member, only its compression properties are taken into calculation for G value, to compare with the W10x45 Wide Flange, the 2x2x1/4" angle compression capacity is negligeable, so its bracing function is negligeable. therefore the column effective length is still 20'.

The 2nd argument is that even if consider the tension function of the 2x2x1/4" angle as its bracing function, but because all the column can be deformed in same direction, therefore these angle still can not be considered as having any bracing function. therefore, the effective length of the column has to be considered as 20'.

What is your thoughts on this scenario?

Thanks a lot for your input.

 

[Lion06]

I would use 20' as the unbraced length, but this is a good discussion.

If the angle is attached to just one of the flanges as shown, I wouldn't consider it braced even if the angles were braced back to something else. I believe this creates a torsional buckling failure.

I'm not sure what you mean in "argument 1". G, as far as I know, has nothing to do with the compressive stiffness. It's all about the relative flexural stifnesses.

I agree with your "second argument".

 

[slickdeals]

If the angles were truly a tension ring, I think they would they prevent the column buckling in the strong axis. I think in order to be effective the attachment should be between the angle and the column web and should transfer the hoop stress through the web of the column.

If the column axis were rotated 90 degrees, then would the column design be more effective considering the weak axis is being braced by the tension ring?

This is indeed an interesting discussion.

 

[slickdeals]

Could a buckling analysis be done for such a system in SAP2000 or RAM/RISA to determine the behavior?

 

[JAE]

Unless the angles are tied in with some X brace somewhere I don't see them doing anything to brace the columns.

I've used the description of one column trying brace another column, while the other column tries to brace the first one - this is like two drunks leaning against each other - highly unstable!

In this case you have 20 drunks all depending on each other to brace their mid-heights. I can visualize a buckled shape where all 20 columns buckle circumferentially together. Not a good situation.

I'd use 20 feet. If you add some X braces between the angle ends and the column bases you might have a case for a brace.

Also - I agree with SEIT's comment that the brace needs to engage the whole column and not just one flange.

 

[westheimer1234]

i actually find that strut useless without brace bay..

 

[Galambos]

I believe there is another mode of failure that you might not be considering, that is along the lines of a global torsional buckling failure. Im pretty sure this is a concern in circular arena type structures.

I *think* you could find it in my namesake's book, Guide to Stability Design Criteria for Metal Structures, by Galambos.

 

[BAretired]

The structure you show is unstable and will collapse unless you provide diagonal bracing around the periphery.

BA

 

[dhengr]

I agree with Galambos & BA. That’s a house of cards waiting to fall over, in a circular domino fashion.

And, what we really need then, is someone experience in statistical and probabilistic considerations as applied to 20 drunks. In JAE’s apt description; will they tend to fall over all at once, or is there a greater probability that they will support each other and stay standing. I think Murphy’s Law comes into play here too.

 

[BAretired]

Attached is a suggestion for bracing. If you do this, the effective length of each column is 20' for the major axis and 10' for the minor axis.

BA

 

[KootK]

Lee: you get full marks for sketch quality.

If you want to avoid the cross bracing for architectural reasons, there is another way.

If the braces were stiffer and connected to the columns in such a manner that they could be counted on to transfer moment (both axes), you'd be back in business.

You would have to calc out the bracing stiffness of this arrangement rigorously however. It wouldn't be a slam dunk like an axial bracing system.

 

[vandede427]

I agree with the global torsional buckling failure comment.

maybe provide 8 braced bays; one at each 45 degree radian (or something similar)

 

[asixth]

Lee

What was the function of the 2" angle prior to you questioning whether it could be used to restrain the W-beam from buckling.

My opinion is 3 braced bays at 120 degrees.

 

[ToadJones]

designed a few hopper/silo style supports very similar to this in power plants to support silos and the like.
most of them were octagonal with a strut mid height and x-bracing above and below, all the way around.
My thinking was with just a few braced bays, the direction of lateral loading may render the bracing useless. Of course there was no architect to argue about the looks.
In one case, the lateral loading was so severe that a large "kicker" or buttress style brace was run down to the ground from one side of the octagon to its own foundation.