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 The information provided may be out of date.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.

3D simulations and meshing problems

Please login with a confirmed email address before reporting spam

Hello,

I am doing some 3D simulations. I have basically some very small structures on a bigger substrate. I have to use the ignore faces operation to remove the common faces between the substrate and those small structures. And This make the whole geometry as just one domain so I won’t be able to refine the mesh just on those small stretchers, is there any way to solve this problem please?

I do appreciate any suggestions

5 Replies Last Post 30 avr. 2013, 04:04 UTC−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 29 avr. 2013, 14:33 UTC−4
Hi
why remove the internal boundaries ?

in my view no reason, even it would make life easier for you to mesh with some internal boundaries. Even if it is the same material, just select all boundaries that apply, adding a few internal bondaries, even in the same material often mkes meshing far easier, and does not increas specially the sparsity of your matrices not the complexity of you model

--
Good luck
Ivar
Hi why remove the internal boundaries ? in my view no reason, even it would make life easier for you to mesh with some internal boundaries. Even if it is the same material, just select all boundaries that apply, adding a few internal bondaries, even in the same material often mkes meshing far easier, and does not increas specially the sparsity of your matrices not the complexity of you model -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 avr. 2013, 16:31 UTC−4
Thank you so so much Ivar for your reply.
So does this mean that those internal boundaries won’t have any effect on the simulations results?

And I have two more questions that I would appreciate any responses.
As I mentioned before I have some small structures on a substrate, actually those small stretchers are some pore s in the substrate. Only the material of the bottom of those pores differs from the rest of the materials anywhere else. And everything will happen at the bottom of the pores. I must refine mesh on the bottom of these pores to approach the convergence. Is there any way to choose only a boundary on a domain for refining the mesh or COMSOL is just happy by refining a selected domain?

Second question: Does the adaptive mesh just work for the whole geometry?

Thank you so much
Thank you so so much Ivar for your reply. So does this mean that those internal boundaries won’t have any effect on the simulations results? And I have two more questions that I would appreciate any responses. As I mentioned before I have some small structures on a substrate, actually those small stretchers are some pore s in the substrate. Only the material of the bottom of those pores differs from the rest of the materials anywhere else. And everything will happen at the bottom of the pores. I must refine mesh on the bottom of these pores to approach the convergence. Is there any way to choose only a boundary on a domain for refining the mesh or COMSOL is just happy by refining a selected domain? Second question: Does the adaptive mesh just work for the whole geometry? Thank you so much

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 avr. 2013, 18:00 UTC−4
Dear Ivar Kjelberg,

I am modelling a gas-liquid bubbly flow. While whenever I tried to compute my model, it always said that ' Undefined value found' and 'NaN or Inf found when solving linear system using SOR'. Could you please give me some suggestions to cover this problem? Thank you very much.

Best wishes,
DD Zhang
Dear Ivar Kjelberg, I am modelling a gas-liquid bubbly flow. While whenever I tried to compute my model, it always said that ' Undefined value found' and 'NaN or Inf found when solving linear system using SOR'. Could you please give me some suggestions to cover this problem? Thank you very much. Best wishes, DD Zhang

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 avr. 2013, 01:14 UTC−4
Hi

often that comes from insufficient BC conditions. Check your dependent variables, check the highest derivatbve of each depednent variable in your equations, and then count the BC, are you sure you have defined enough to give an unique solution, or are two "aligned" or not independent so you need one more.

Often the "p" pressure value is forgotten, some BC must define the p value somewhere, often its the outflow p=p, but i fou have no outflow, then adding a p=p at one point in the model will give aou a better gauge point. (in structural its best to restrain your model too, and not having "floating around" in space

--
Good luck
Ivar
Hi often that comes from insufficient BC conditions. Check your dependent variables, check the highest derivatbve of each depednent variable in your equations, and then count the BC, are you sure you have defined enough to give an unique solution, or are two "aligned" or not independent so you need one more. Often the "p" pressure value is forgotten, some BC must define the p value somewhere, often its the outflow p=p, but i fou have no outflow, then adding a p=p at one point in the model will give aou a better gauge point. (in structural its best to restrain your model too, and not having "floating around" in space -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30 avr. 2013, 04:04 UTC−4
Dear Ivar Kjelberg,

I am very grateful for your important suggestions. I checked momentum equation again (turbulent bubbly flow) and found liquid velocity has a second order derivative. Does it mean I need to give one more BC for liquid velocity? But in my model, liquid is stagnant in the tank and only gas passes through the tank. I can assume liquid velocity at inlet is zero and liquid pressure at outlet is P. But how could I add one more BC for liquid velocity?

I also assumed mass transfer and additional volume force in my tank is zero just to simplify it at beginning. I think I've already list all the BC, but it didn't work still. Do you know what I should do? The attachment is my model in the case you'd like to check it. (Only the second one is my model, I am sorry the first one is not relative and I just mixed them up)

Thank you very much with my best wishes.

Best regard,

Dd. Zhang
Dear Ivar Kjelberg, I am very grateful for your important suggestions. I checked momentum equation again (turbulent bubbly flow) and found liquid velocity has a second order derivative. Does it mean I need to give one more BC for liquid velocity? But in my model, liquid is stagnant in the tank and only gas passes through the tank. I can assume liquid velocity at inlet is zero and liquid pressure at outlet is P. But how could I add one more BC for liquid velocity? I also assumed mass transfer and additional volume force in my tank is zero just to simplify it at beginning. I think I've already list all the BC, but it didn't work still. Do you know what I should do? The attachment is my model in the case you'd like to check it. (Only the second one is my model, I am sorry the first one is not relative and I just mixed them up) Thank you very much with my best wishes. Best regard, Dd. Zhang

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.