Menu

Discussion Forum

New Discussion
Filtrer

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 incompressible Navier Stokes, using Flow in porous media (Brinkman Equation)

Posted 25 juin 2010, 04:37 UTC−4 12 Replies

Pablo Briones
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi guys,

I have a problem when a use the incompressible Navier Stokes with the activation of Flow in porous media (Brinkman Equation).

I get the next error:

Error:
Failed to find consistent initial values.
Matrix_has_zero_on_diagonal

What can I do?, I have tried everithing but nothing happen.

Regards

Pablo
12 Replies Last Post 27 oct. 2010, 00:44 UTC−4
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2010, 05:29 UTC−4
Hi Pablo,

This is because of the solver, as the first guess. Switch to a Direct solver if possible to see if the problem is the solver or you have missed some physics. Also, since you enabled another physic, your stiffness matrix changes character and without fine tuning of pre-conditioners for iterative solvers (venka) it will give that error. I'm sure more experienced people can give your more detailed answer.
Hi Pablo, This is because of the solver, as the first guess. Switch to a Direct solver if possible to see if the problem is the solver or you have missed some physics. Also, since you enabled another physic, your stiffness matrix changes character and without fine tuning of pre-conditioners for iterative solvers (venka) it will give that error. I'm sure more experienced people can give your more detailed answer.
Pablo Briones
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2010, 05:34 UTC−4
Hi, Thank for your help

What you mean with "Direct solver"? Is a option or something in the program?



Hi, Thank for your help What you mean with "Direct solver"? Is a option or something in the program?
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2010, 05:35 UTC−4
Iterative solvers vs. direct solvers. In the solver box, there are several direct solvers (UMFPACK, PARDISO) and some iterative ones. Direct solvers are usually much easier to use than iterative ones.
Iterative solvers vs. direct solvers. In the solver box, there are several direct solvers (UMFPACK, PARDISO) and some iterative ones. Direct solvers are usually much easier to use than iterative ones.
Pablo Briones
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2010, 05:51 UTC−4
Hi,

Thank a lot i didn't know this part.
I will try with this now and I will tell you the results

regards

Pablo
Hi, Thank a lot i didn't know this part. I will try with this now and I will tell you the results regards Pablo
Pablo Briones
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2010, 06:13 UTC−4
Hi,

now is running with the "Flow in porous media", thanks a lot for your help

Other question: There are a lot of different between the direct solve and the iteration solve? or is a goos solution too?

regards
Hi, now is running with the "Flow in porous media", thanks a lot for your help Other question: There are a lot of different between the direct solve and the iteration solve? or is a goos solution too? regards
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 juin 2010, 07:04 UTC−4
It depends on your degree of freedom (DOF), complexity of your physics, 2D-3D, and plenty more. I have used Direct solvers for 700k DOF, but usually it works the best for <300k. They have less tuning and easier to setup. But for special cases e.g. 3D direct solvers are problematic; both speed and memory usage.
It depends on your degree of freedom (DOF), complexity of your physics, 2D-3D, and plenty more. I have used Direct solvers for 700k DOF, but usually it works the best for
NGUYEN Quang Thanh
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 juin 2010, 07:02 UTC−4
Hi,
I would like to know what solver and what preconditionner that you used for your problem.
Regard
Hi, I would like to know what solver and what preconditionner that you used for your problem. Regard
Pablo Briones
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 juin 2010, 09:12 UTC−4
Hi,

I was using BigStab, but was not working, after i have change to PARDISO and is work.
But i have also the problem that when i try to mix with the Convection and diffusion equation (using the ouput of the navir stokes u,v,w ) only work with small values of the Initial Velocity in the Navier and this value is like 10000 time more small that the real value that i nedd.

What you mean with precondition. which values you need?

Hi, I was using BigStab, but was not working, after i have change to PARDISO and is work. But i have also the problem that when i try to mix with the Convection and diffusion equation (using the ouput of the navir stokes u,v,w ) only work with small values of the Initial Velocity in the Navier and this value is like 10000 time more small that the real value that i nedd. What you mean with precondition. which values you need?
NGUYEN Quang Thanh
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 juin 2010, 10:02 UTC−4
Thanks for your answer.
we have preconditioner when we use indirects solvers like GMRES or FGMRES
Thanks for your answer. we have preconditioner when we use indirects solvers like GMRES or FGMRES
Pablo Briones
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 29 juin 2010, 10:22 UTC−4
ok, so i don't have in this case.

You have some ideo why is not working with big values in the Initial Velocity and only for small values?

regards
ok, so i don't have in this case. You have some ideo why is not working with big values in the Initial Velocity and only for small values? regards
mohammad shamsul
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 oct. 2010, 19:31 UTC−4
Which solver do you think suitable incase of solving porous media problem?
Which solver do you think suitable incase of solving porous media problem?
Xife
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 oct. 2010, 00:44 UTC−4
what about the default setting of solver provided by COMSOL40a? is there a downside to using these defaults? it's just unclear what rule of thumb must be applied per problem base. thanks for your suggestion.

--
merci
what about the default setting of solver provided by COMSOL40a? is there a downside to using these defaults? it's just unclear what rule of thumb must be applied per problem base. thanks for your suggestion. -- merci

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