Menu

Discussion Forum

New Discussion
Filtrer

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 The information provided may be out of date.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.

Velocity Oscillation near No-Slip Boundary

Posted 25 nov. 2015, 08:56 UTC−5 Fluid & Heat Version 4.4 4 Replies

Leo Mechelle
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Dear All,

I found a problem in my 2D comsol simulation when I am solving N-S equations in cylindrical coordinate. I use the creeping flow and neglect inertial. And I use P1+P1 discretization.
My solving geometry is shown in the attachment(First graph). The yellow line indicates the real boundaries of the system. Near the inner sphere boundary, I apply non-slip boundary condition. However, there is great body force just near the surface and it causes great oscillation of velocity right near the surface(second graph).
Therefore, I wonder is that possible to smooth this oscillation by giving multiply the body force by a decay facter(like step function) near the spherical surface just in the shaded region in first graph. However, I am not sure how to write this kind of decay function in 2D cylindrical coordinate near sphere surface. Could you please give me some suggestions?

Or any other suggestions to solve this oscillation?

Many Thanks!

Best,
Mechelle

Attachments:

4 Replies Last Post 1 déc. 2015, 06:31 UTC−5
Leo Mechelle
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 25 nov. 2015, 21:09 UTC−5
I have tried to reduce the mesh size near boundary using "boundary layers", number of mesh layers is 40 and stretch facter is 1.1. I think the mesh is fine enough.
I have tried to reduce the mesh size near boundary using "boundary layers", number of mesh layers is 40 and stretch facter is 1.1. I think the mesh is fine enough.
Leo Mechelle
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 30 nov. 2015, 22:39 UTC−5
Dear All,

I also wonder if the velocity is oscillating near boundary, could I still believe the viscous stress or pressure result obtained from creeping flow module?
Many Thanks!

Best Regards,
Mechelle
Dear All, I also wonder if the velocity is oscillating near boundary, could I still believe the viscous stress or pressure result obtained from creeping flow module? Many Thanks! Best Regards, Mechelle
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: 9 years ago 1 déc. 2015, 01:44 UTC−5
Hi

you might have two issues:
1) BC induced singularities, of numerical or equation induced form, that makes the solution oscillate
2) the default COMSOL plot is doing smoothing and interpolations, check the plot "Quality" tab, and change the settings there.

As a general rule, when one have such unexpected oscillations or "non physical values" such as negative temperatures or concentration for the diffusion equation, then one should always check the results at the mesh node level, by turning off all plot refinements and other "improvements".

Depending on your model, you might have a local singularity, probably of numerical form, in such cases you can do whatever you want with the mesh, it will only even grow with finer mesh. Check the COMSOL blog there are several explanations given there

--
Good luck
Ivar
Hi you might have two issues: 1) BC induced singularities, of numerical or equation induced form, that makes the solution oscillate 2) the default COMSOL plot is doing smoothing and interpolations, check the plot "Quality" tab, and change the settings there. As a general rule, when one have such unexpected oscillations or "non physical values" such as negative temperatures or concentration for the diffusion equation, then one should always check the results at the mesh node level, by turning off all plot refinements and other "improvements". Depending on your model, you might have a local singularity, probably of numerical form, in such cases you can do whatever you want with the mesh, it will only even grow with finer mesh. Check the COMSOL blog there are several explanations given there -- Good luck Ivar
Leo Mechelle
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 1 déc. 2015, 06:31 UTC−5
Dear Ivar,

Thank you so much for your detailed reply!

I guess my problem is due to the first issue you raised. I just found that if the body force is reduced near boundary, the oscillation of velocity will decrease a lot(in the attachment).
However, the large body force is the condition that I must deal with since they can exist in nature. In nature, if a large force is applied near no-slip boundary, the velocity will surely be smooth.
It seems a dilemma...

Best,
Mechelle
Dear Ivar, Thank you so much for your detailed reply! I guess my problem is due to the first issue you raised. I just found that if the body force is reduced near boundary, the oscillation of velocity will decrease a lot(in the attachment). However, the large body force is the condition that I must deal with since they can exist in nature. In nature, if a large force is applied near no-slip boundary, the velocity will surely be smooth. It seems a dilemma... Best, Mechelle

Attachments:

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