Coupling Acoustic-Structure

Please login with a confirmed email address before reporting spam

Hi,

I have modeled in solid mechanics a structure excited by a load and i would like to calculate the radiation in the air domain of a point located in the far field. I have used pressure acoustic boundary element to reduce the computational domain and coupled with acoustic-structure interaction. I only considering frequency domain. The problem is that the model is too computational expensive. I have tried to use axisymmetric symmetry but still the model is not able to converge. the structure is a wind turbine tower by the way.

I saw that i could actually split the problem in two part: first study the vibration of the structure in solid mechanics and afterward use the displacement from that as input in normal displacement boundary condition using boundary element method. But i do not get it.
Any advice or specific instruction.


5 Replies Last Post 15 janv. 2025, 10:02 UTC−5
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 weeks ago 9 janv. 2025, 04:26 UTC−5

Can you be more specific about what the problem is?

Do you have a problem with accessing the solution from the first study or is it a problem in writing the boundary condition for the acoustic problem?

If you create a small example, would you be able to do the same thing for FE based pressure acoustics?

-------------------
Henrik Sönnerlind
COMSOL
Can you be more specific about what the problem is? Do you have a problem with accessing the solution from the first study or is it a problem in writing the boundary condition for the acoustic problem? If you create a small example, would you be able to do the same thing for FE based pressure acoustics?

Acculution ApS Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 2 weeks ago 9 janv. 2025, 09:08 UTC−5

You may want to look into a segregated solver setup a la https://doc.comsol.com/5.4/doc/com.comsol.help.aco/aco_ug_acousticstructure.06.33.html

But if you cannot get convergence with 2Daxi, there may be some underlying problem that needs handling before you concern yourself with solver optimization. Get the model going with a rough mesh and build confidence from there.

-------------------
René Christensen, PhD
Acculution ApS
www.acculution.com
info@acculution.com
You may want to look into a segregated solver setup a la https://doc.comsol.com/5.4/doc/com.comsol.help.aco/aco_ug_acousticstructure.06.33.html But if you cannot get convergence with 2Daxi, there may be some underlying problem that needs handling before you concern yourself with solver optimization. Get the model going with a rough mesh and build confidence from there.

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 weeks ago 9 janv. 2025, 11:17 UTC−5

Filipo,

Can you describe what the real world (3D) problem looks like (geometry, loads, etc)? I have trouble figuring out how a 2D axisymmetry assumption is possible.

Jeff

-------------------
Jeff Hiller
Filipo, Can you describe what the real world (3D) problem looks like (geometry, loads, etc)? I have trouble figuring out how a 2D axisymmetry assumption is possible. Jeff

Please login with a confirmed email address before reporting spam

Posted: 1 week ago 13 janv. 2025, 04:33 UTC−5

I attach model to clarify. I have tried to use segregated solver but still as suggest from Acculution ApS but still is too computationally expensive. I do not know how to set the boundary condition from the solid mechanics study and use them in pressure acoustic boundary element. I had a look at some tutorial but still is not working.

I attach model to clarify. I have tried to use segregated solver but still as suggest from Acculution ApS but still is too computationally expensive. I do not know how to set the boundary condition from the solid mechanics study and use them in pressure acoustic boundary element. I had a look at some tutorial but still is not working.


Mark Cops COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 7 days ago 15 janv. 2025, 10:02 UTC−5

I noticed that the model is using a Fully Coupled, Direct Solver which is quite expensive for BEM in 3D. In the model attached here, I right clicked Solver Configurations->Reset Solver to Default. I then enabled the second suggestion for iterative solver (GMRES with Direct Precon.) as this showed better convergence after a test run.

The last change I made was under Pressure Acoustics, Boundary Elements->Quadrature to enable Thin gap handling. This is needed for accuracy for BEM when narrow regions exist in the model, such as the thin walled cylinder.

The first model attached here solved in ~20 minutes with about 20 GB RAM for me, which is not bad. The screenshot shows the acoustic pressure from the vibrating structure.

Out of curiosity, I also setup the model with a segregated solver. Step 1 solves for the structural DOFS, while Step 2 is the acoustics. In terms of run time, this was similar however the log reported about 15 GB RAM this time.

Hope this helps, -Mark

I noticed that the model is using a Fully Coupled, Direct Solver which is quite expensive for BEM in 3D. In the model attached here, I right clicked Solver Configurations->Reset Solver to Default. I then enabled the second suggestion for iterative solver (GMRES with Direct Precon.) as this showed better convergence after a test run. The last change I made was under Pressure Acoustics, Boundary Elements->Quadrature to enable Thin gap handling. This is needed for accuracy for BEM when narrow regions exist in the model, such as the thin walled cylinder. The first model attached here solved in ~20 minutes with about 20 GB RAM for me, which is not bad. The screenshot shows the acoustic pressure from the vibrating structure. Out of curiosity, I also setup the model with a segregated solver. Step 1 solves for the structural DOFS, while Step 2 is the acoustics. In terms of run time, this was similar however the log reported about 15 GB RAM this time. Hope this helps, -Mark

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.