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.

Comsol can resolve a problem heat transfer 3D?

Posted Sep 8, 2011, 5:22 a.m. EDT 7 Replies

Thien An Nguyen Van
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hello everyone,

I use Comsol to model a coupling problem thermo-mechanics. However, when I tested a simple problem of heat transfer in 3D (with a internal heat source and adiabatic case), the results received are not good. Because they depend on the mesh.

When I tested with 3 elements in the height, there were 3 layers of temperature distribution in this part. Or, 4 elements in the height, there were also 4 layers of temperature.

You can see these results accompanied in this message.
Do you have any opinion to improve the solver of Comsol?

Thank you very much for yours discussions.
7 Replies Last Post Sep 9, 2011, 11:24 a.m. EDT
Lechoslaw Krolikowski
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 8, 2011, 10:35 a.m. EDT
Hi,
The results depend on the grid when the grid is too coarse. You should perform a mesh convergence study.
Regards.
Hi, The results depend on the grid when the grid is too coarse. You should perform a mesh convergence study. Regards.
Thien An Nguyen Van
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2011, 7:20 a.m. EDT
Thanks for your discussion.
I try this test in Abaqus. The temperature field does not depend on the mesh. I'll give you this result.
Regards
Thanks for your discussion. I try this test in Abaqus. The temperature field does not depend on the mesh. I'll give you this result. Regards

Attachments:

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2011, 8:50 a.m. EDT
I suspect that none of your models (COMSOL or Abaqus) represent an adiabatic condition. The analytical solution in the adiabatic case (i.e. heat flux=0 on all boundaries) with a uniform distributed heat source is a uniformly-increasing temperature.
I suspect that none of your models (COMSOL or Abaqus) represent an adiabatic condition. The analytical solution in the adiabatic case (i.e. heat flux=0 on all boundaries) with a uniform distributed heat source is a uniformly-increasing temperature.
Thien An Nguyen Van
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2011, 9:15 a.m. EDT
None of my examples is adiabatic condition. These examples take account a convective flux with the environement : q=h. (Text-T).
The result of temperature in abaqus is continuous. But in the Comsol, the distribution of temperature is depend on layers of mesh as in the picture.
Do you have any opinion to improve this result?
Regards
None of my examples is adiabatic condition. These examples take account a convective flux with the environement : q=h. (Text-T). The result of temperature in abaqus is continuous. But in the Comsol, the distribution of temperature is depend on layers of mesh as in the picture. Do you have any opinion to improve this result? Regards
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2011, 9:22 a.m. EDT
Well, your original posting said you were trying to model an adiabatic condition...
Anyways, if you want help figuring out what's wrong with your model set-up, please post the COMSOL mph file.
Well, your original posting said you were trying to model an adiabatic condition... Anyways, if you want help figuring out what's wrong with your model set-up, please post the COMSOL mph file.
Thien An Nguyen Van
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2011, 9:37 a.m. EDT
Thank you for your discussion.
I have posted this model.
Regards
Thank you for your discussion. I have posted this model. Regards
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Sep 9, 2011, 11:24 a.m. EDT
I had a quick look at your model file and it looks like given the values of the material properties and time span you selected you have extremely thin thermal boundary layers that simply cannot be resolved by these meshes. You'd need a mesh with much finer elements along the boundaries to arrive at a solution that's reasonably close to the analytical solution.
I had a quick look at your model file and it looks like given the values of the material properties and time span you selected you have extremely thin thermal boundary layers that simply cannot be resolved by these meshes. You'd need a mesh with much finer elements along the boundaries to arrive at a solution that's reasonably close to the analytical solution.

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