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.

new geometrie in the moving mesh appl. mode

Michael Rembe Certified Consultant

Please login with a confirmed email address before reporting spam

I calculate a transient moving mesh in a axi-symmetrical model. By aid of postprocessing -> subdomain integration I evaluate the volume of my geometry:
t=0 I get V1
t=t_stop I get V2

After creating a new geometry from the deformed mesh I check the volume:
for t=0 I get V3 and V3 > V2 !!

I expected V3=V2!

Is there anybody who gets the same results and who knows the reason to prevent this difference?


********** Note! *****************
The volume "increases" by updating the model after creating the new geometry from the deformed mesh.
*********************************

Thank you and best regards
Michael

4 Replies Last Post Jun 26, 2009, 11:06 a.m. EDT
Mina Sierou COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 18, 2009, 1:30 p.m. EDT
Hi Michael,

How large are the differences you see? If there are pretty small, it is possible that this is just a post-processing/meshing issue? (the volume is calculated based on whatever mesh you have before and after remeshing -- a finer mesh that would represent the geometry a bit different could give a slightly different answer)?

Mina Sierou, COMSOL Inc
Hi Michael, How large are the differences you see? If there are pretty small, it is possible that this is just a post-processing/meshing issue? (the volume is calculated based on whatever mesh you have before and after remeshing -- a finer mesh that would represent the geometry a bit different could give a slightly different answer)? Mina Sierou, COMSOL Inc

Michael Rembe Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 19, 2009, 3:19 a.m. EDT
Dear Mina,

the differences are very high, about 30-40%! The reason is - I tested it using a small model - that I start a new time stepping after recreating the new geometry and remeshing.
- evaluate V0(t=0) (step 1)
- simulating ALE t=0,....,1
- evaluate V1(t=1) (step 1)
- creating new geometry
- remeshing
- evaluate V1(t=1) (step 2) it agrees with V1 from step 1
- evaluate V0(t=0) (step 2) it doesn't agree with V0 from step 1 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Because I don't restart the model but I start a new solver sequence after remeshing and new time stepping, the volumes differ. But there must be a way to prevent this case!

Best regards
Michael
Dear Mina, the differences are very high, about 30-40%! The reason is - I tested it using a small model - that I start a new time stepping after recreating the new geometry and remeshing. - evaluate V0(t=0) (step 1) - simulating ALE t=0,....,1 - evaluate V1(t=1) (step 1) - creating new geometry - remeshing - evaluate V1(t=1) (step 2) it agrees with V1 from step 1 - evaluate V0(t=0) (step 2) it doesn't agree with V0 from step 1 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Because I don't restart the model but I start a new solver sequence after remeshing and new time stepping, the volumes differ. But there must be a way to prevent this case! Best regards Michael

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 26, 2009, 9:54 a.m. EDT
Hi, there is probably some conceptual mistake in the model. If you submit the model to this thread or support@comsol.com I am sure the problem will be resolved by somebody.

Niklas
Hi, there is probably some conceptual mistake in the model. If you submit the model to this thread or support@comsol.com I am sure the problem will be resolved by somebody. Niklas

Michael Rembe Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 26, 2009, 11:06 a.m. EDT
Hello Niklas,

the case is supported by Ms. Bannach and Mr. Geis from Comsol Germany (Göttingen). There are some activities in the "background", thats why I dont send the model file again. Hope the reasons will be found next week.

Today we set up mass transport for the k-eps turbulence model. It should be included in the next version of comsol! Now the model of a large cavern runs perfectly.

Best regards
Michael Rembe
Hello Niklas, the case is supported by Ms. Bannach and Mr. Geis from Comsol Germany (Göttingen). There are some activities in the "background", thats why I dont send the model file again. Hope the reasons will be found next week. Today we set up mass transport for the k-eps turbulence model. It should be included in the next version of comsol! Now the model of a large cavern runs perfectly. Best regards Michael Rembe

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.