Free-standing post stability checks

[trainguy]

Guys,

I am currently checking a free standing built-up section (2-I beams side by side connected with battens).
_
I_I

This built-up section supports loads in exactly the same way as a Hang-man structure (upside-down L rigidly fixed at the base) with wind loads thrown in for good measure...

FYI, its supporting catenary cables over railway track. Because of our recent (1998) ice-storm, the cables are laden with much ice, making them heavier and increasing the wind loads.

To get an appreciation of the magnitudes involved:
Cf = 200kN
Mfx = 750 kNm at the base.

I am planning to use Kx = Ky = 2.0 for axial buckling calcs (free at top, and fixed at bottom).

What approach would you take for the strong axis moment resistance, given lateral-torsional buckling? I am fully conversant in Canadian S16.1 for code bending strength of laterally unsupported members, but I'm not sure how to handle the completely free top.

Any ideas or design guides?

thx in advance

tg

 

[BAretired]

If you have wind loads, then the both legs of the L have biaxial bending. The vertical member also has torsion.

I believe you have already considered the free top by using K = 2.0.

BA

 

[trainguy]

BA,

I'll make the questions more specific:

What length of laterally unsupported compression flange would you use in calculating Mu for the vertical member? Twice the actual length? More than 2L? I'm using Mu as used in S16.1 Clause 13.6.

What values of J and Cw would you use for this built-up member? I'm leaning toward using Cw=0, and J=sum(.333*L*t^3)

Thanks

tg.

 

[BAretired]

tg,

Yes, 2L for length of unsupported compression flange as in Appendix F. Also, if Cw = 0 for hollow sections, it is appropriate for your built up section.

If your battens are continuous, you are being too conservative in your choice of J. A closed shape such as an HSS has a much larger J value than the sum of the individual plates. A method of calculating J can be found in Blodgett "Design of Welded Steel Structures" Section 2.10 for closed shapes. Also, see this link:

http://en.wikipedia.org/wiki/Torsion_constant

If the channels are connected with discontinuous battens, then it is not a closed shape and you might be correct in using the sum of the individual J values for the channels. Using a closed shape makes a huge difference in the value of J.

BA

 

[trainguy]

BA,

This is a great help - thanks. I may run a quick FEA of the assembly to explicitly determine the GJ term (Applied Torsion * height / theta).

I could then use this in the Mu equation.

Any thoughts on such an approach?

tg

 

[BAretired]

If your built up shape is a closed shape, I would calculate GJ by hand methods. I don't know enough about FEA to comment on your approach, but I would be interested in hearing the result, particularly if your J value differs significantly from the hand calc.

BA

 

[trainguy]

Actually, the built up shape is a hybrid (open - closed), because of the discontinuous battens.

I could conservatively disregard the battens and use the sum of the separate J's (for each I beam), but I think we could save weight with a slightly more rigorous assessment of J.

If I go ahead, I would first confirm the FEA approach by considering a continuous closed shape and comparing with a hand calc.

As I write this - I realize it's quickly turning into a research project...

tg