Steel testing frame - sway-buckling NOT possible? 4

[Ingenuity]

I have designed and detailed a 6 foot tall steel testing frame that shall be used for static testing of post-tensioning strand & wire tendon assemblies. Simple set up of multi-wire or multi-strand vertical tendons, concentric within a steel frame.

I have an immediate need for a 200,000 lb.f test load, but I may design the frame for a possible 400,000 lb.f capacity with modifications to the beam spans/sizes.

A hydraulic center-hole ram on the top beams provides statically applied incrementing loads to the tendon assembly, and reacts upon the top and bottom beams, with load passing into the 4 x vertical columns.

I checked the flexural buckling of the 4 vertical legs - they are 3" XX STRONG pipes, with W4x13 top and bottom horizontal beams, so a nominal axial capacity of 200 kips each leg to AISC 360-16. I need to factor in some accidental eccentricity and check for beam/column actions.

BUT...since the vertically-applied post-tensioning tendon is the source of the applied load, it is NOT possible for a sway buckling mode - similar to a vertical self-centering system. Do you agree?

I need a reality check before I start welding up the frame. Thanks.

 

[bugbus]

I agree it is self-centering so the normal sway type buckling would be unlikely.

But given how torsionally flexible the whole frame is without any sort of diagonal bracing between the 4 legs, you may well get a torsional-compression type buckling where the points of load application don't necessarily have to move relative to each other (although they would rotate relative to one another). So the two ends of each tube may well 'sway' relative to each other even though the points of load application remain aligned.

 

[human909]

I ran a rough model out of curiosity. A linear elastic buckling analysis. Torsional buckling is narrowly edged out by regular Euler buckling in this circumstance. (A few dimensions etc have been estimated.) Either way with a base load of 1000kN the Euler values of 34 and 37 are well beyond yielding, buckling at 34,000kN or 7,643.5kips.

I adding a slight torsional load but it didn't make much difference to the buckling behaviour.

I'll have a quick look at plastic buckling if the model behaves.
UPDATE: Plastic buckling drops a little. But not much. From my quick analysis buckling isn't a worry here with these thick stocky pipes.

 

[Ingenuity]

bugbus and human909: Many thanks for your contributions. It is much appreciated.

bugbus - nice call on the compression-torsional mode.

human909 - is that STRAND analysis and output, or other?

Those "thick stocky pipes" sure are a SOB to lift and cut and weld...

Thanks again to you both for helping out an 'AU brother'.

 

[human909]

human909 - is that STRAND analysis and output, or other?

Inventor Nastran-InCad

Thanks again to you both for helping out an 'AU brother'.

So what was with those damn imperial units you are using!

 

[Ingenuity]

Inventor Nastran-InCad

Wow. Nice. I first used NASTRAN back in 1984 (in AU at Uni) it was on main frame - looks like it has come a loooong way since those days for pre and post-processing.

So what was with those damn imperial units you are using!

I have lived in the US for near 30 years, and whilst I love the metric system, those damn yanks hate it...so I am bi-lingual when it comes to systems of units.