Buckling analysis issue

[MicPrz]

Hello, 

I am running buckling analysis for my model. Simple steel building - storage hall.

I want to check buckling length for columns and got some strange results.

I am getting very high buckling length for member despite diagram for buckling mode (1) shows buckling moreless equal to nominal length.

Is it something that should correspond? Am I doing something wrong?

Would appreciate resonse 🙂 thanks

[Romanich]

Hi @MicPrz ,

Have you checked these topics?

https://forums.autodesk.com/t5/robot-structural-analysis-forum/results-buckling/td-p/3406121

https://forums.autodesk.com/t5/robot-structural-analysis-forum/results-buckling/td-p/3406121

[MicPrz]

Hi,

yes. Dividing columnt into parts is not affecting my results much.

One more thing. I am wondering... when I receive the results for buckling analysis and some modes correspond to a global stability loss eg Mode 1,2 - is it a right approach to use the calculated buckling lengths for compression members and than use them in CODE CHECK?

[Krzysztof_Wasik · Accepted solution]

Hi @MicPrz 

While, buckling analysis, Robot calculates buckling length for all members in buckling load case. Calculated buckling length is reasonable only for members which are subjected to buckling (green bar in example). For other elements calculated lengths are not correct (red bars in example).

Reason

Buckling length is calculated from Euler formula for all members in the structure. 

L is calculated from the following

Forces distribution, regarding critical force detected for members subjected to buckling is taken into account. But it may happen that only one element or part of the structure is subjected to buckling.

For members subjected to buckling for which critical force is reached (Fx similar to Pcr - green bar see below), calculated buckling length is reasonable. For members not subjected to buckling (Fx<Pcr), loads are smaller than critical force (or even tension can occur - right column in the example),  so using Euler formula, results in bigger (not reasonalbe) or N/A (members subjected to tension) buckling length.

Generally to obtain reasonable buckling length for member, we have to choose buckling mode in which this member is subjected to buckling. In other cases (modes) calculated length should be ignored.

I see that you check buckling length for mode 1 but regarding buckling modes it might be correct to regard modes 10-14 or even higher modes for which your bar is subjected to buckling. 

[MicPrz]

@Krzysztof_Wasik as always post on point

I think now I got it 🙂

really appreciate your help

[MicPrz · Accepted solution]

@Krzysztof_Wasik  Mayby one more thing.

If I am not getting any local mode for a column I want to investigate, would it be a right approach to add new simple case eg. and load this column directly in a node and than perform bucking analysis for this case?

Than for sure I will get buckling data corresponding for local column buckling.

[Krzysztof_Wasik]

Hi @MicPrz 

You can load member with axial force in separate (buckling test) load case. It is a good approach to obtain buckling length for member local buckling mode, when end supports stiffness should be regarded.

[MicPrz]

@Krzysztof_Wasik I have already did so - i mean after your first post you have explained me a lot I have created separate load cases for buckling checks to be sure that my assumptions in Code Check are +/- allright.

Thanks once again 😉