critical coefficient for buckling (is my structure ok?)

[kkengtip]

Hello All,

I am using non-linear P-Delta buckling analysis for designing a curved structure with stepping section sizes (CHS244.5*16 and CHS 244.5*8). The FA-package, I am using, is Robot Millennium V.21.

Right now, the lowest critical load coefficient from the model is about 5.

Does that mean my structure can take 5 times the current load I applied to it?

If the above question is true, can I say my structure is safe base on a critical load coefficient greater than 1.

Currently, I am not very confident with the critical coefficient, so I take the effective length from buckling analysis and check the bars using BS5950. Is this necessary providing I have got a critical coefficient greater 5? Also, if I use the effective length method, I often get some bars with extreme large effective lengths (e.g.1000m) but tiny forces (less 3KN axial; 2KN.m moment, I believe the forces are correct), which cause the bar fails the BS5950 buckling check, and then I have to increase the bar size though it doesn’t carry much load.

I have been struggling with this problem for a week now and still couldn’t find a confident answer.

Hope you guys can give me a hand Please.



Thanking you in advance,

KK

 

[crisb]

The load coefficients from an Eigenvalue buckling analysis are not comparable to real life load factors. The buckling curves used for the back calculation based on effective length include allowances for initial imperfections and residual stresses. Your method sounds like the conventional way of the using Eigenvalue buckling results. If you have unrealistic effective lengths it may be that you are not choosing the most appropriate mode for that member. If you are in any doubt about the results I recommend carrying out a nonlinear run with some initial imperfections included (if your program will do this).