Buckling analysis - Local mode

[t.sautierr]

Good evening everyone,

I'm currently studying a tower structure and begin a global analysis by calculating the critical alpha coefficient.
Have you a trick to avoid local modes, and to highlight the global modes more easily? like in a modal analysis using a coherent matrix or removing the density of vibrating bars.

[t.sautierr]

Bonsoir à tous, 

je suis entrain d'étudir une structure de tour et de commencer une analyse globale par calcul des coefficient alpha critique.

Est ce que vous avez une astuce pour éviter les modes locaux, et faire ressortir les modes globaux plus facilement, un peux comme dans une analyse modale en utilisant un matrice cohérente ou en supprimer la densité des barres qui vibrent.

Merci beaucoup.

[HoshangMustafa]

Hi

Please follow this link

http://forums.autodesk.com/t5/robot-structural-analysis/global-buckling-shape/td-p/5001086

[t.sautierr]

Thanks, I will analyse it.

[Artur.Kosakowski]

Hi Thomas,

I'm afraid there is no easy way to filter out the global modes. What you may try is to work with the Buckling Analysis Results table and use these steps:

1. Display results for a single mode

2. Find the column with 1.0/-1.0 value

3. Sort this column

When the most of the values in this column are not close to zeros you may assume this is a global mode (top on the atatched picture) whereas when they are close to or zeros it is probably the local one ( bottom on the picture).

.

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[Artur.Kosakowski]

Please check:

http://forums.autodesk.com/t5/robot-structural-analysis/buckling-analysis-local-mode/td-p/6380910

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[t.sautierr]

I checked it Arthur but It didn't helped. I must confessed I have not understood entirely the effect to set a dephasing : for example if I set it to 1 for the buckinlg analysis, It will force Robot to search eigen value from 1 and above?

[t.sautierr]

I will continue to speak on the english forum, in order not to double the discussion and I will set here the solution when we arrived to it 🙂

[t.sautierr]

It a good method, we used it last night, but unfortunately the analysis (around 50 modes) contained only local modes.

This means there are no trick to reduce the number of modes to calculate, one must maybe calculate 500 and after only sort the eigen vector to find what is relevant.

[Artur.Kosakowski · Accepted solution]

I can't think of any other option 

[Artur.Kosakowski]

I must confessed I have not understood entirely the effect to set a dephasing : for example if I set it to 1 for the buckinlg analysis, It will force Robot to search eigen value from 1 and above?

No, it will search for values closed to the specified "target".

E.g. for 10 modes and no shift you have:

and when you reduce number of modes to be calculated and set the shift to e.g. 60 you will have:

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[t.sautierr]

That's brilliant !

So if I want to keep global modes with alphacr around 8 :

I first do an analysis without shift to see the distribution of the critical coefficient and numebring

Let's say 100 modes with the first and only  global at 50th position, value 8 

Then I can shift to 8 and with calcultions for 10 modes, searched around 8 -> it will "filter" the local modes because of the shift!

The best would be to be able to make a multi-shift in order to be able to focus precisely around values of global modes (same idea than with a modal analysis where you can filter the interessting modes taken into account in seismic, after a first analysis of all the modes)

[Artur.Kosakowski]

Few buckling analysis with different shifts? 

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[t.sautierr · Accepted solution]

Re-Brilliant!