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 The information provided may be out of date.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.

Problem using extrusion coupling variable.

Please login with a confirmed email address before reporting spam

My model consists of two geometry. Geo1 is 1D while geo2 is 2D. The type of mesh for geo2 is square mapped mesh.(Since geo2 is square.)

The physics of geo1 is diffusion in z direction(1D diffusion in the packed bed reactor) while the physics of geometry 2 is diffusion in r direction(1D diffusion in the spherical particle+reaction). The gas molecules travel along the packed bed reactor and diffuse into the catalyst particle.

I try to couple Geo1 and Geo2 using extrusion coupling variable. The extrusion variable for geo1 is subdomain extrusion variable while the extrusion variable for geo2 is boundary extrusion variable(It's the first derivative term.) at one side of geo2. In this case, I think I can use linear transformation.

The obtained results were compared with analytical solution.(To test the accuracy of the model) The results from COMSOL were close to analytical solution.(About 10% error) So I increased the numbers of meshes and decreased the tolerance values in solver setting(I hoped the results would be better.) but the results remained the same as before(Not significantly change).

Do you have any suggestions on this?

3 Replies Last Post Mar 16, 2011, 5:26 p.m. EDT
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 Mar 15, 2011, 11:53 a.m. EDT
Hi

not really,
you have done the #1 mesh sensitivity test ;) then remains (if you trust the analytical model better than the COMSOL results ;) to find any underlying hypothesis that one might have missed.

Unfortunately, I have problems undrstanding and visualising your extrusion coupling discussion, and further I'm not the most familiar/experienced with the chemistry modules. But there are certainly others out here that might give further hints

--
Good luck
Ivar
Hi not really, you have done the #1 mesh sensitivity test ;) then remains (if you trust the analytical model better than the COMSOL results ;) to find any underlying hypothesis that one might have missed. Unfortunately, I have problems undrstanding and visualising your extrusion coupling discussion, and further I'm not the most familiar/experienced with the chemistry modules. But there are certainly others out here that might give further hints -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Mar 16, 2011, 6:45 a.m. EDT
Dear Ivar

For geo1, I use subdomain extrusion variable to send the concentration values(cb) in geo1 to one boundary of geo2.

In the same way, I use boundary extrusion variable to evaluate the first derivative of concentration(d cp/dy) at the same boundary of geo2 and send it to geo1.

Nattapong Pongboutr





Dear Ivar For geo1, I use subdomain extrusion variable to send the concentration values(cb) in geo1 to one boundary of geo2. In the same way, I use boundary extrusion variable to evaluate the first derivative of concentration(d cp/dy) at the same boundary of geo2 and send it to geo1. Nattapong Pongboutr

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 Mar 16, 2011, 5:26 p.m. EDT
Hi

10% error is not that much, but if it's not improving with the mesh density it might come from some singularity or edge effects.
If you are evaluating a derivative on an edge it might be influenced either by next (sub)domain if exist, or some isolation or specific BC on the edge. Have you checked how the derivative is behaving close to the edge, or try to extrude from a line just before the edge, but inside the sub-domain ? OR to take some average over an area close to the edge ?

You might also have some numerical oscillations on the model too, specially close to edges

This is just a guess, as I do not have much details of your model, and as said before, it''s not my most experienced domain ;)

--
Good luck
Ivar
Hi 10% error is not that much, but if it's not improving with the mesh density it might come from some singularity or edge effects. If you are evaluating a derivative on an edge it might be influenced either by next (sub)domain if exist, or some isolation or specific BC on the edge. Have you checked how the derivative is behaving close to the edge, or try to extrude from a line just before the edge, but inside the sub-domain ? OR to take some average over an area close to the edge ? You might also have some numerical oscillations on the model too, specially close to edges This is just a guess, as I do not have much details of your model, and as said before, it''s not my most experienced domain ;) -- Good luck Ivar

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.