Modal Analysis + Mass Partiticipation Factors

[Clarabella]

Interesting, we are studying the collapse of a 70 meters multilayer dome similar to the one you are investigating. I believe the nonlinear model that we developed is better than that shown in section 4 of the paper. We used fiber sections, distributed plasticity, force-based elements, and a corotational approach for nonlinear geometry.
We are able to capture very well the local and global instability of the dome. I think that this approach would work for this problem very well.

[akshaysomkuwar123]

Dear Clarabella

Thank you for the suggestions I tried after changing displacement based beam column element to forced based beam column element still I am getting strange eigenvectors above mode 150.
Please let me know any suggestions that will help mw in resolving the issue.

Thank you,
Akshay S

[Clarabella]

Sorry I was thinking of the nonlinear model described in section 4. Please post your last model.

[akshaysomkuwar123]

Dear Clarabella

I have shared my model in previous threads. I am sharing again please find the model attached in the post.
Please let me know any valuable suggestions.

Thank you

Regards
Akshay S

[marafini.f]

Dear akshaysomkuwar123,

we made a little modification to your model, to solve the problem of the high eigenvalues.
Here is an updated file.

K8-4ele-1mem-massstudy1.rar

(37.02 KiB) Downloaded 397 times

We introduced the mass and rotary inertia indicated in the article you provided.

In addition, we noticed the following:
- your model does not have the same geometry like the one in the article, therefore I am not sure how you can compare the results;
- the balls in the article are not concentrated masses at the nodes, in the plot, it is evident that they are modelled as spheres, which means that the stiffness they provide together with the beam sections, is much different than the one you modelled.

We added the mpco recorder in your file so you can open it in STKO Post-Processor to visualize the results.
Why don't you try to do the whole model in the Pre-Processor?

Hope this helps.
Francesca

[akshaysomkuwar123]

Dear Francesca

Thank you for the help.

-The model you discussed is the test dome which is in section " 3.4 Finite element model ",
-The model I am analyzing is available in section "4.0 Finite element model of prototype single-layer reticulated dome" Hence you found out difference in the geometry.

Still I am getting problem of high eigenvalues.

I have started developing whole model in Pre-processor.

Thank you

Regards,
Akshay S

[marafini.f]

Dear Akshay S,

I am sorry, I had not understood you were referring to that second structure.
You are right, it is not evident whether they use or not rotational masses, but I repeat the higher eigenvalues are there because those modes after a certain value engage rotational mechanisms.
If you add even a small percentage of rotary inertia, like in the file I attach, here

K8-4ele-1mem-massstudy1_suggestion.rar

(37.13 KiB) Downloaded 360 times

you will not have that initial error:

missing operand at _@_
in expression "...0000000000000000000 * _@_"

and obtain results until the 200th mode.

Why don't you plot the Total effective mass ratio graph and see if you get consistent results?
Unfortunately, I do not have the time available to study the paper in detail, just a few ideas, did you check if the element formulation of Abaqus B31 beam used is the same as the displBeamColumn you used in OpenSees? It says at some point that it uses 8 integration points along the elements, did you check if you are using the same number of integration points?

If you have further questions on this model we would be happy to support you with the modeling or the post-processing in STKO, let us know how the results comparison goes.

Francesca

[akshaysomkuwar123]

Dear Francesca,

As you suggested rotational masses might be taking part in higher modes so I tried playing with it but I am getting the same problem of very large eigenvectors, may be apart from this some other issue might be causing the problem.

However I am sharing the results of comparison, I am getting the same pattern but percentage error in the frequencies that I am getting is like 30%.

Do let me know if you have any suggestion to improve my model.

Thank you for all the suggestions you have given to me

Regards,
Akshay S

[Horace Horsecollar]

Please send us the .scd model
Thank you

[akshaysomkuwar123]

Thank you very much for all the suggestion, I was able to solve the problem. It occurred due to some unconstrained/unconnected node.