Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

Fluid Structure Interaction problem

Please login with a confirmed email address before reporting spam

Dear all,

I am trying to 2D model a solid body (circle) rotating in simple shear flow in a channel and I am using the fluid structure interaction module to try and solve it. I have followed the standard tutorial for deformation of a narrow vertical obstacle by horizontal channel flow and made changes to suit my needs. I have specified the following conditions in the fsi module.

- upper and lower walls have been specified as sliding walls moving opposite to each other so that the net velocity in channel is 0 an the solid body simply rotates without translation.
- mean inlet and outlet velocities have been specified as 0
- the body has been specified to be a linear elastic material model
-the boundaries of the body have been specified in the fluid solid interface boundary

I am currently using default values in the "initial values" and "prescribed mesh displacement" sections of the fsi module. Again, I am not sure if that is quite right.

The mesh is a fine free triangular mesh over the entire geometry. When I run the model over a chosen time dependent range I am continuously getting the following error

"Failed to evaluate variable Jacobian.
- Variable: mat.nu
- Geometry: 1
- Domain: 2"


I would greatly appreciate any help I can get. I have been working on this problem for the past few months without getting anywhere.

Thank you very much.

5 Replies Last Post May 28, 2014, 10:03 a.m. EDT
Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 10 years ago Apr 27, 2014, 11:24 p.m. EDT
It seems like Poisson’s ratio is not defined.
It seems like Poisson’s ratio is not defined.

Please login with a confirmed email address before reporting spam

Posted: 10 years ago Apr 30, 2014, 4:48 p.m. EDT
Dear Nagi,

Thanks for replying. It assumed that becasue I had specified the material it would automatically get those values. I made that change, however, I am still running into some error --- " Failed to find consistent initial values Segregated group 2 " etc.

Was wondering if you'd have any insights as to whats going wrong now. I am attaching the mph file for reference.

Once again, thanks a lot for getting back on this ....
Dear Nagi, Thanks for replying. It assumed that becasue I had specified the material it would automatically get those values. I made that change, however, I am still running into some error --- " Failed to find consistent initial values Segregated group 2 " etc. Was wondering if you'd have any insights as to whats going wrong now. I am attaching the mph file for reference. Once again, thanks a lot for getting back on this ....


Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 10 years ago May 2, 2014, 6:23 p.m. EDT
Hi Seemit,

I only managed to take a quick look at your model and I have two comments. It is safer to start prescribed velocities in a time dependent analysis with zero so that the first time step solution is close to the initial conditions. Also it’s not suitable to put velocity boundary conditions at both inlet and outlet. Switch one to pressure.

Nagi Elabbasi
Veryst Engineering
Hi Seemit, I only managed to take a quick look at your model and I have two comments. It is safer to start prescribed velocities in a time dependent analysis with zero so that the first time step solution is close to the initial conditions. Also it’s not suitable to put velocity boundary conditions at both inlet and outlet. Switch one to pressure. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 10 years ago May 24, 2014, 7:58 p.m. EDT
Dear Nagi,

Thanks a lot for your insights. I was actually able to solve the 2d problem without the use of FSI by solving two separate problems with and without rotation and simply adding them up as it is creeping flow conditions. Now I am trying to solve the problem in 3d but i am running into one small problem. Is there any way to rotate a solid sphere at a given angular velocity in comsol ??? I am trying to use the rotating machinery module but once again I am not quite sure if its the right way, I do feel I could solve this without the module, just not getting the right BC on the surface of the spheres.

Once again, very grateful for your replies. They have helped a lot in getting to solve this problem.

Thanks,
Seemit
Dear Nagi, Thanks a lot for your insights. I was actually able to solve the 2d problem without the use of FSI by solving two separate problems with and without rotation and simply adding them up as it is creeping flow conditions. Now I am trying to solve the problem in 3d but i am running into one small problem. Is there any way to rotate a solid sphere at a given angular velocity in comsol ??? I am trying to use the rotating machinery module but once again I am not quite sure if its the right way, I do feel I could solve this without the module, just not getting the right BC on the surface of the spheres. Once again, very grateful for your replies. They have helped a lot in getting to solve this problem. Thanks, Seemit

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 10 years ago May 28, 2014, 10:03 a.m. EDT
Hi Seemit, glad I could help.

Yes you can rotate a solid sphere at a prescribed angular velocity as you described. You do not need the Multibody Dynamics Module for this simple kind of rigid body motion. You can do it using a Rigid Connector. Constrain all or part of the sphere using the connector and prescribe a rotation to its center, sometime like omega*t to get a constant angular velocity omega. We used this approach in this peristaltic pump animation: www.youtube.com/watch?v=Gj0MGjyhFCc.

Nagi Elabbasi
Veryst Engineering
Hi Seemit, glad I could help. Yes you can rotate a solid sphere at a prescribed angular velocity as you described. You do not need the Multibody Dynamics Module for this simple kind of rigid body motion. You can do it using a Rigid Connector. Constrain all or part of the sphere using the connector and prescribe a rotation to its center, sometime like omega*t to get a constant angular velocity omega. We used this approach in this peristaltic pump animation: http://www.youtube.com/watch?v=Gj0MGjyhFCc. Nagi Elabbasi Veryst Engineering

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.