Abaqus mesh analysis

[Kenci555]

Hi everyone.

I am modelling a steel frame in Abaqus. The frame is modelled as a 3D-problem built by shell elements.
In the step-module i have chosen "buckle" as the type and eigensolver as Lanczos with 6-7 eigenvalues.
I should remind that i am a new user in Abaqus, few months of experience.
However, when i want to mesh the structure i am allways using the same types of elements. However, when i change the number of elements, that means changing the number in "approximate global size", it never seems to convergence. The eigenvalue changes alot if I for instances put 0.2 or 0.1 as the value of "approximate global size". It simply never converge to a stable value of the eigenvalue.

Is there something wrong with the model?
The way i mesh the structure?
Or is it simply impossible to a convergence analysis for this analysis?

Best regards
Kenan

 

[IceBreakerSours]

I am not sure if I understand. What has mesh convergence got to do with eigenvalues? Could you point me to a section in this documentation that talks about this?

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[Kenci555]

Thanks for the answer IceBreakerSours.
Thats the thing that i dont understand.. because this anlaysis is giving me some eigenvalues..
Now, if I change the mesh.. For instances using, lets say 1000 elements in one analysis and 6000 elements in the second analysis, the resulting eigenvalues have a huge difference between them.. Also the deformed structure behaves strange... i have posted a picture of the structure deformed with the first positive eigenvalue in one analysis where i use a certain of elements.. and a picture of another analysis were i use even more elements (finer mesh).. as you can see from the result.. it is behaving very strange.. you have any idea of what is could cause this?

Picture of first analysis

Picture of second analysis with a fine mesh

 

[Kenci555]

 

[Kenci555]

Link

 

[Kenci555]

First analysis
Link

Second analysis with a FINE mesh
Link

 

[IceBreakerSours]

I had suspected it but now I know that you do not know what you are doing. And that is fine, as long as you wish to learn.

So, you must go back to undergrad structural dynamics and differential equations texts to understand the concept of eigenvalues/eigenvectors. While you are reading up on it, visualize a rigid bar, a linked-chain, and a rope hanging from a ceiling with a pin joint at the ceiling.

Once you feel good about the concepts, open a book on practical finite element analysis and study the section on mesh convergence.

If you still have questions, feel free to continue the thread.

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083