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 The information provided may be out of date.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.

Solving pure advection problem

Please login with a confirmed email address before reporting spam

Hi,

I use COMSOL 4.2a to solve colloid transport in a pipe using convection-diffusion equation. The diffusion is anisotropic with diffusion coefficient in axial direction zero. But COMSOL is showing error:

Undefined value found.
- Detail: Undefined value found in the equation residual vector. There are 15004 degrees of freedom giving NaN/Inf in the vector for the variable mod1.c

Is this because that peclet number is infinite in the axial direction?
But COMSOL 3.5a could solve such problems with zero diffusion without showing any error.

Can anybody help in this regard?

Regards,
Seetha

4 Replies Last Post 30.05.2012, 12:29 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30.05.2012, 09:33 GMT-4
Hi

interesting, there are obviously some differences between the versions.
OK but have you tried to replace "0" in the axial direction by a "small" value, just to check it's really the anisotropy that is giving you problems and nothing else ?

--
Good luck
Ivar
Hi interesting, there are obviously some differences between the versions. OK but have you tried to replace "0" in the axial direction by a "small" value, just to check it's really the anisotropy that is giving you problems and nothing else ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30.05.2012, 11:24 GMT-4
Hi

Its the problem only with diffusion, I think.
I tried using very small value of diffusion coefficient like 1E-9 in axial direction. But in that case also, it is showing the same error. May be because then also peclet number is so large. Galerkin FEM is unstable for peclet greater than 1.
COMSOL 4.2a is able to solve the problem if I give diffusion coefficient in the axial direction as 1E-3 times that in transverse direction, but that is a big value for my problem.

But 3.5a is not showing any error and working properly.

How can I resolve this issue in 4.2a?

Regards,
Seetha
Hi Its the problem only with diffusion, I think. I tried using very small value of diffusion coefficient like 1E-9 in axial direction. But in that case also, it is showing the same error. May be because then also peclet number is so large. Galerkin FEM is unstable for peclet greater than 1. COMSOL 4.2a is able to solve the problem if I give diffusion coefficient in the axial direction as 1E-3 times that in transverse direction, but that is a big value for my problem. But 3.5a is not showing any error and working properly. How can I resolve this issue in 4.2a? Regards, Seetha

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30.05.2012, 11:59 GMT-4
Please check the discretization and other artificial diffusion setting for variable 'conc' (convection-diffusion equation). I had similar issues, where solution for convection-diffusion equations were not matching for 3.5a (quad discretization) and 4.2 (linear discretization) version. If you make sure your discretization and other artificial diffusiuon settings are same, it should give same results. The documentation also contains more about artificial diffusion and other related settings.

Good luck.
Please check the discretization and other artificial diffusion setting for variable 'conc' (convection-diffusion equation). I had similar issues, where solution for convection-diffusion equations were not matching for 3.5a (quad discretization) and 4.2 (linear discretization) version. If you make sure your discretization and other artificial diffusiuon settings are same, it should give same results. The documentation also contains more about artificial diffusion and other related settings. Good luck.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30.05.2012, 12:29 GMT-4
Hi

Thanks for your reply.
I used the same settings of 3.5a.
In 3.5a, there was no need to add artificial diffusion, as the solution was non-oscillatory.

But 4.2a is not solving at all with zero or very small diffusion coefficient. Also I am already using very fine mesh, so if peclet needs to be less than one (if I use very small diffusion coefficient), then the mesh size should be very very fine one. But I am not sure, whether the problem is due to very high peclet or something else.

Regards,
Seetha
Hi Thanks for your reply. I used the same settings of 3.5a. In 3.5a, there was no need to add artificial diffusion, as the solution was non-oscillatory. But 4.2a is not solving at all with zero or very small diffusion coefficient. Also I am already using very fine mesh, so if peclet needs to be less than one (if I use very small diffusion coefficient), then the mesh size should be very very fine one. But I am not sure, whether the problem is due to very high peclet or something else. Regards, Seetha

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.