Menu

Discussion Forum

New Discussion
Filter

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Eigenfrequency Analysis in Fluid-Structure Interaction

Posted Jun 25, 2010, 7:39 a.m. EDT Studies & Solvers 8 Replies

Drake Mith
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

I'm trying to calculate the resonance frequencies of a fluid filled hollow shell and the results I get are NaN for each frequency, but it works when I use the Pressure-Acoustics (acpr) physics mode (I get frequencies and realistic mode shapes) on some arbitrary shape.

Here is my setup process (using Acoustics Module in version 4.0)<check attached file for further clarification>:

Load physics and the liquid water material. Load the Linear Elastic Material Model (from the physics tree "branch")

Draw two concentric spheres, set the outer one as solid elastics model, leave the inner as acoustics pressure.

From the Union.

Acoustic-Structure boundary is automatically assigned to their interface (8 surfaces)

Set material properties (Young's Modulus, Poisson Ratio, Density and get three green checks) manually in the solid outer sphere.

Do a normal Free mesh and then compute 18 or so eigenfrequencies.

Everything else was left to default

This results in all NaNs.



From my understanding, the computer now knows everything it needs to calculate the frequencies: forces and accelerations from material properties; geometrical relationships in the functions; and the mesh. It knows what part is fluid, what part is solid so I don't understand why its not working. It might be something in the solver

Attachments:

8 Replies Last Post Mar 7, 2011, 8:22 a.m. EST
COMSOL Moderator

Hello Drake Mith

Your Discussion has gone 30 days without a reply. If you still need help with COMSOL and have an on-subscription license, please visit our Support Center for help.

If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.

Bill Crowther
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Nov 2, 2010, 5:17 p.m. EDT
I'm having exactly the same problem... Did you ever resolve this issue?
I'm having exactly the same problem... Did you ever resolve this issue?
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Nov 4, 2010, 8:55 a.m. EDT
Hi

What puzzles me is that I see aluminium also for the inner part, I though it was hollow (= air ?)

--
Good luck
Ivar
Hi What puzzles me is that I see aluminium also for the inner part, I though it was hollow (= air ?) -- Good luck Ivar
Vladimir Bogachuk
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 26, 2011, 4:07 p.m. EST
hello,
Have you ever figure this out. because i have this problem I have a structure in a fluid and i need to obtain the natural frequencies of the structure with the viscous effects from the fluid. Thank you
hello, Have you ever figure this out. because i have this problem I have a structure in a fluid and i need to obtain the natural frequencies of the structure with the viscous effects from the fluid. Thank you
Drake Mith
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 27, 2011, 5:58 a.m. EST
Are you using 4.1? Apparrently eigenvalue calculations for fluid-structure interation wasn't supported until 4.1. That's how I did it to work recently, though I am still getting the "Relative Error is greater than relative residual" bug. I have been using a freq-domain sweep though.

Ivar, I did forget to set the inner domain.

--

----------------------------------------------------------------------------------------------------------------
Arts, crafts and sciences uplift the world of being and are conducive to its exaltation.
~Baha'u'llah
Are you using 4.1? Apparrently eigenvalue calculations for fluid-structure interation wasn't supported until 4.1. That's how I did it to work recently, though I am still getting the "Relative Error is greater than relative residual" bug. I have been using a freq-domain sweep though. Ivar, I did forget to set the inner domain. -- ---------------------------------------------------------------------------------------------------------------- Arts, crafts and sciences uplift the world of being and are conducive to its exaltation. ~Baha'u'llah
Vladimir Bogachuk
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 27, 2011, 4:25 p.m. EST
Hello,
Acoustic-Structure Interaction doesn't take into account the viscous damping effects on the structure from the fluid. I have simple problem that i made for a sanity check. Where you have fluid surrounding structure and when you get Eigenfrequency of the structure it should go up because of the stiffening effect of the fluid but instead its going down and you get different modes. Thank you

Vlad
Hello, Acoustic-Structure Interaction doesn't take into account the viscous damping effects on the structure from the fluid. I have simple problem that i made for a sanity check. Where you have fluid surrounding structure and when you get Eigenfrequency of the structure it should go up because of the stiffening effect of the fluid but instead its going down and you get different modes. Thank you Vlad
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 28, 2011, 1:31 a.m. EST
Hi

well I do not have the acoustic module (one of the three modules have not managed to get my boss to buy as I do not receive enough projects requests in these domains ;(

So, I would need to wait to I get a temporary demo license, that will perhaps arrive for one of the v4.2 workshops coming up soon, or at next Comsol Conference. But Vladimir, if this is so, send a request to "support" there should absolutely be some FSI set-up allowing such obvious (but not trivial) couplings in COMSOL at least to come

--
Good luck
Ivar
Hi well I do not have the acoustic module (one of the three modules have not managed to get my boss to buy as I do not receive enough projects requests in these domains ;( So, I would need to wait to I get a temporary demo license, that will perhaps arrive for one of the v4.2 workshops coming up soon, or at next Comsol Conference. But Vladimir, if this is so, send a request to "support" there should absolutely be some FSI set-up allowing such obvious (but not trivial) couplings in COMSOL at least to come -- Good luck Ivar
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 28, 2011, 4:38 p.m. EST
Hi

well I'm not COMSOL just to have that said, so I cannot comment for them.

But what you always can do i to add some damping on the boundaries via the equations:

if you look at the egenfrequency analysis equations of COMSOL you will see that external forces terms not related to iomega are dropped, so if you add a Cv*i*lambda term you will add some viscuous damping, up to you to scale Cv correctly.

The same way if you add a -k*i*lambda^2 term you will add a "k" spring (and change the eigenfrequency)

for a frequency swepp replace "lambda" by "2*pi*freq", and to avoid "orange unit errors", add two parameters
lambda = 0[1/s]
freq = 0[1/s]

But you can also add some ordinary BC damping (and look at the equations)

You are not linking the physics this way, but you do add one extra physics in, onto the other, via the equations like that, and you remain fully in Structural

I have attached a simple v4.1.0.154 2D model, play with the boundary loads (enable/disable for the different solvers)

--
Good luck
Ivar
Hi well I'm not COMSOL just to have that said, so I cannot comment for them. But what you always can do i to add some damping on the boundaries via the equations: if you look at the egenfrequency analysis equations of COMSOL you will see that external forces terms not related to iomega are dropped, so if you add a Cv*i*lambda term you will add some viscuous damping, up to you to scale Cv correctly. The same way if you add a -k*i*lambda^2 term you will add a "k" spring (and change the eigenfrequency) for a frequency swepp replace "lambda" by "2*pi*freq", and to avoid "orange unit errors", add two parameters lambda = 0[1/s] freq = 0[1/s] But you can also add some ordinary BC damping (and look at the equations) You are not linking the physics this way, but you do add one extra physics in, onto the other, via the equations like that, and you remain fully in Structural I have attached a simple v4.1.0.154 2D model, play with the boundary loads (enable/disable for the different solvers) -- Good luck Ivar

Attachments:

Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Mar 7, 2011, 8:22 a.m. EST
Hi all.
Check this paper from the Milan conference 2009 about eigenfrequency FSI. The idea is to linearize Navier Stokes.
www.comsol.com/papers/6747/

--
Niklas
Hi all. Check this paper from the Milan conference 2009 about eigenfrequency FSI. The idea is to linearize Navier Stokes. http://www.comsol.com/papers/6747/ -- Niklas

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.

Suggested Content