Modal analysis using substructuring strategy

Hello guys,

I'm Quentin, a french master of science student. I have just started an internship in a company that produce a small wind turbin called "Wind Tree" (in you guys want more information about the product, here is the webpage : URL). My job is to produce a modal analysis of the structure using sub structuring strategy. My choice of finite element software turned to ASTER, cause it's free and the compagny does not have the founds to get some licence elsewhere.

I must confess that it's only my second day I read things about using ASTER. Cause I worked hard on my bibliographic study before launching any calculation^^.

My final project is using beams as substructures and then get the whole structure modal analysis. But for now, I want to learn about ASTER using a simple example. During my bibliography's study, I worked on a simple case (embedded beam cut into two substructures and only in traction) which I have solved "manualy" with MATLAB (you can watch out for the linked image).
My first goal is to reproduce this simple case with ASTER to help me understand how to write and organize my .comm.

What I did not get :
1. How must I structure my CAD models -> one CAD per substructure ? or one CAD for the whole structure and define the super elements after that ?
2. For now I'm turning for a solution without using SALOMEMECA (SolidWorks + GMSH + ASTER + Paraview). Is that a good idea ?
3. I did not find any tutorial about sub structuring technics in ASTER. Any one knows where I can find an exemple ?

I thank you all for your future answers. If you guys need more information, I think I will haunt this forum quite a bit in the next days, so i would be able to answer fast.

"May ASTER be with you" (mASTER YODA)

See you all !

Hello,

quite interesting project!

Here some training materials:
Modal Analysis
Dynamic Substructuring

And some Code_Aster validation cases that might be a good starting point:
Harmonic Response of A Sub-Structured Beam
Calculation of plate modal substructuring with Ritz basis

Good luck!

Best,
Richard

Hi Richard,

Thank you for your fast answer ! I took a rapid look at the papers you have linked. They look really interesting.
Today I get an appointment with an ASTER developer. So I will work on the paper you gave me as soon as I can.

Are sub structuring technics, methods you already used ? In that case I may ask you some real technical question when I start my code.

Hello Quentin,

I have no personal experience in using sub-structuring techniques with Code_Aster,
so i might not be the best person to ask for specific advice. But with general problems or error messages
I will try to help you if I have some spare time.

Best,
Richard

Hello Richard and the others,

I begin to understand right how ASTER works. My meetings with the ASTER developper made me learn a lot about using sub structuring methods, and dynamics study in general.

1. My biggest problem now is to export a 3D/shell design into a beam structure. Do you guys know a good way to get the neutral fiber of a 3D design.

2. The other thing I'm worried of is how to get a node at a precise coordinate on the beam (mesh). Because interfaces will need a node at the right coordinate (my beams have multiples interfaces that are not at the edges of the beams). Maybe by creating a group of node from geometry where I want a node ?
I want something solid/stable cause my next job on the model will be optimizing the structure mass using python scripts.

3. Last question, you may need to go see how the "arbre à vent" looks like, but basically the leafs are the wind turbines. I don't need to model the rigidity of them. What I'm looking for, is to represent the inertia/mass of such leafs. Does modeling the leafs with a little beam with infinite rigidity, and the right mass and mass center seems to be a great idea ? The Aster developper told me that it's kinda complex to create single mass point.


See you all,
Quentin

Hello Quentin,
I'll try to answer your questions...

Gavroche wrote: 1. My biggest problem now is to export a 3D/shell design into a beam structure. Do you guys know a good way to get the neutral fiber of a 3D design.

I think this has been discussed already in the forum and to my knowledge there exists no open source or free CAD tool that allows to create midsurface or neutral fiber models out of 3D geometries. whereas almost any commercial CAD tool does.

Gavroche wrote: 2. The other thing I'm worried of is how to get a node at a precise coordinate on the beam (mesh). Because interfaces will need a node at the right coordinate (my beams have multiples interfaces that are not at the edges of the beams). Maybe by creating a group of node from geometry where I want a node ?
I want something solid/stable cause my next job on the model will be optimizing the structure mass using python scripts.

This is possible with Salome. You can explicitly define that you want to have a node at a specific point. But this question would be better answered in the Salome forum.

Gavroche wrote: 3. Last question, you may need to go see how the "arbre à vent" looks like, but basically the leafs are the wind turbines. I don't need to model the rigidity of them. What I'm looking for, is to represent the inertia/mass of such leafs. Does modeling the leafs with a little beam with infinite rigidity, and the right mass and mass center seems to be a great idea ? The Aster developper told me that it's kinda complex to create single mass point.

I Don't think that it would be very complex to add point masses that represent the mass and inertias of your "leaves". You can create POI1 elements , at the center of mass of the leaves, assign the correct mass and inertias to it and connect those POI1 elements with LIAISON_MAIL or other linear relations to your tree.

Best regards,
Richard