Steel Column outside of Bldg Slab

Good question. Sorry this post is absolutely no help, but could you calculate the buckling force at each point then check the one foot extention as a cantilever from the floor slab to resist the buckling force?

Your concern for lateral stability of the column is a legitimate one. I would take special care to study the lateral bracing requirements of the column in the direction parallel to the exterior wall.

What it sounds like is a column that depends upon the minor axis bending of the horizontal girder to cantilever out from the building 1 foot.

Joseph Yura with the University of Texas, Austin has worked with AISC over the years to develop design concepts for bracing (stability) of structures. You must ensure that you have BOTH enough strength and enough stiffness to adequately serve as a brace.

For your upper columns, I have the gut feeling that you should be OK with a fairly good sized girder. Near the bottom, there may at least be some concern over the stiffness provided by the girder in the weak axis.

See if you can find some of Yura's recommendations. He puts on a good stability seminar with Todd Helwig, from the Univ. of Houston and the one I attended a few years ago was very informative.

If you refer to the latest AISC LRFD spec, section C3, you will find the strength and stiffness parameters that, I believe, are based on Yura's work.

Can you plate the girders to develop some lateral restraint using a small percentage of the axial load as a measure of the force required to restrain (usually 1% to 5%, depending on the mood and circumstances) The moment about a vertical axis must then be transferred to the girder and then to the floor plate.

Hope it's in a warm climate... thermal transfer from the column to the girder may require the use of styrofoam shims <G>.

I've only been involved with one project with exterior supports... these were clad and insulation was provided around the columns.

The problem is the y-y axis buckling and effective length.
How about providing short knee braces (or stay) from the 2 unsupported sides of the column at 45degrees, back to the floor beam / floor slab.......a good architect could make a feature out of these.
The alternative would be to provide a longitudinal horizontal tubular tie at each floor level, outside the building.

suqlainUK:
agreed... but the plating of the girder would be used to provide the necessary restraint. Similarly the webs of the column would have to be stiffened at the connection. It will not be as stiff as a 'rigid' strut (the plating going from the toes of the girder to 'box' it in) Dealing with flexural stiffness as opposed to axial stiffness, but it is workable. It's not a very tall building so the equivalent forces to maintain stability should be relatively small. The plating is likely less obtrusive, than diagonal struts,to the architect.

Let me throw my 2 cents in...

Here's what I would do: (Assuming 14&quot; Col. & 24&quot; Girder)

Detail an end plate connection (3/4&quot; PL) at the girder/column connection. Make the plate about the same width as the &quot;T&quot; dimension of my column.

Provide 3/8&quot; stiffener plates at each side of the girder located @ 7.5&quot; & 16.5&quot; from the top of steel (Sym. about nuetral axis of girder). Provide a 1/2&quot; stiffener plates on the far side of the column located at the girder's stiffener plates.

Bolting to the column:

CL
|
|3&quot;|1.75&quot;|1.75&quot;|3&quot;|
-- o o | o o ---
3&quot; |
-- o | o 4.5&quot;
3&quot; ======|======---- Stiffener Pls
-- o | o
3&quot; |
-- o o | o o ---Nuetral Axis (Sym. About)

I would take 4% of the axial load,apply it at the column, and verify my girder could take the weak axis bending for the one foot moment arm. Unless you've got a really flimsy girder, this should be a verify calculation. I would not check fbx/Fbx + fby/Fby <=1 because of the location of this connection.

To back this up:
A beam is considered to have its compressive flange braced if a framing beam's connection is in the compressive area. (I don't remember the exact parameters, however, for the most part, the 3&quot; from top of steel to the first bolt satisfies this condition.) The framing beam must also be able to handle a certain percentage of the compressive flange force. I believe Soloman & Johnson say 2% but I typically double it to be conservative.

Apply this concept to your condition, this connection has now braced both flanges, and will not allow your column to rotate.

FYI:
This will not be an inexpensive connection. When checking the shop drawings, I'd circle this connection, and tell the fabricator to &quot;mill this end&quot; to provide consistency. A good fabricator would do this anyway, but just to be safe...

BTW:
Unless the steel is fabricated in an air conditioned environment, (Not going to happen), erected in an air conditioned environment (aww, now come on), I would not even CONSIDER thermal stresses induced by a one foot section of exterior steel. The building is going to feel more thermal stress once the skin is put on and the air conditioning turned on than than that small section could ever aspire to produce.

However, on exterior steel with long spans (i.e. pipe bridges, kiln walways, etc.), you do need to take it into account.

ChipB:
I wasn't thinking of thermal stresses... just cold transfer! I'm originally from Winnipeg... and the interior icicles may not be decorative <G>...

dik,
Well there you go! I'm in a warm climate!
Chip

tw,

You could analyze the entire column by substituting an equivalent spring for for each girder to consider the restraint of the girders. The spring constant to use is the lateral load at the free end of the girder that produce one inch displacement. The springs act in a direction perpendicular to the girders. In the other direction, as you said, the column is braced.

This method is similar to the analysis of a guyed tower, where the tower is supported by springs at each guy level.

regards

AEF