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 with periodic boundary condition

Jean-Philippe Regnard

Please login with a confirmed email address before reporting spam

Hi,
i have a problem that i don't really understand.
I made some simulation of a square impedance tube to compute sound absorption coefficient.
The mode works very well.
when i apply a periodic condition for the lateral faces of the tube, the model failed (Failed to find a solution for the initial parameter). The relative error (1e+004) is greater than the relative tolerance).
It can be possible to solve this problem by increasing the mesh resolution bu it work only for some frequencies and the problem get back.
Does someone can explain to me what's wrong?
Thanks a lot.

Jean-Philippe

6 Replies Last Post 23 feb 2017, 20:47 GMT-5
Jean-Philippe Regnard

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 24 ago 2015, 12:49 GMT-4
Hi,
an important thing to specify is that my model is a 3D model.
Hi, an important thing to specify is that my model is a 3D model.

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: 9 years ago 24 ago 2015, 13:55 GMT-4
Hi

I suspect that when you add a periodic condition, you still need to restrain one point, or some other boundary to fully constrain the solution, and that could be the issue, that your solution is modulo a constant and that sometimes the numeric make COMSOL to manage to +/- converge to one possible solution, but then by changing the mesh or something else it does not converge anymore.

Always check that you have enough BC for your equations, acoustics are second order equations in time and space so you need at least two independent BC's, and reasonable initial conditions

--
Good luck
Ivar
Hi I suspect that when you add a periodic condition, you still need to restrain one point, or some other boundary to fully constrain the solution, and that could be the issue, that your solution is modulo a constant and that sometimes the numeric make COMSOL to manage to +/- converge to one possible solution, but then by changing the mesh or something else it does not converge anymore. Always check that you have enough BC for your equations, acoustics are second order equations in time and space so you need at least two independent BC's, and reasonable initial conditions -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 21 feb 2017, 16:09 GMT-5

Hi

I suspect that when you add a periodic condition, you still need to restrain one point, or some other boundary to fully constrain the solution, and that could be the issue, that your solution is modulo a constant and that sometimes the numeric make COMSOL to manage to +/- converge to one possible solution, but then by changing the mesh or something else it does not converge anymore.

Always check that you have enough BC for your equations, acoustics are second order equations in time and space so you need at least two independent BC's, and reasonable initial conditions

--
Good luck
Ivar


Dear Ivar Kjelberg,

I also got the problem about the period boundary condition. I am dealing with a simple geometry and when I set the inflow and outflow boundary as the Period Flow Condition, it always tells me "Failed to find a solution. Maximum number of Newton iterations reached. There was an error message from the linear solver. The relative residual (0.041) is greater than the relative tolerance. Returned solution is not converged." But it actually is accuracy to meet the requirements, so where can I set it to avoid the error.

Thank you!

Jiaqing

[QUOTE] Hi I suspect that when you add a periodic condition, you still need to restrain one point, or some other boundary to fully constrain the solution, and that could be the issue, that your solution is modulo a constant and that sometimes the numeric make COMSOL to manage to +/- converge to one possible solution, but then by changing the mesh or something else it does not converge anymore. Always check that you have enough BC for your equations, acoustics are second order equations in time and space so you need at least two independent BC's, and reasonable initial conditions -- Good luck Ivar [/QUOTE] Dear Ivar Kjelberg, I also got the problem about the period boundary condition. I am dealing with a simple geometry and when I set the inflow and outflow boundary as the Period Flow Condition, it always tells me "Failed to find a solution. Maximum number of Newton iterations reached. There was an error message from the linear solver. The relative residual (0.041) is greater than the relative tolerance. Returned solution is not converged." But it actually is accuracy to meet the requirements, so where can I set it to avoid the error. Thank you! Jiaqing

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 22 feb 2017, 08:14 GMT-5
Hello Jiaqing,
It's hard to say for sure what is wrong with your set-up without seeing your mph file, but it's quite possible that Ivar's comment above applies directly to your model as well. Indeed, a Navier Stokes problem does not have a unique solution unless a reference pressure is defined somewhere. See e.g. www.comsol.com/community/forums/general/thread/117651
Best regards,
Jeff
Hello Jiaqing, It's hard to say for sure what is wrong with your set-up without seeing your mph file, but it's quite possible that Ivar's comment above applies directly to your model as well. Indeed, a Navier Stokes problem does not have a unique solution unless a reference pressure is defined somewhere. See e.g. https://www.comsol.com/community/forums/general/thread/117651 Best regards, Jeff

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 22 feb 2017, 15:23 GMT-5
Hi Jeff,

Thank you so much for the answering. I attach three pictures which you can have a closer look at the problem I am encountering. Picture 1 shows the model and periodic flow boundary, and Picture 2 shows the Solver Configuration where you can see, I just adopt the default settings and the relative tolerance is 0.001. Picture 3 shows the Converge plot, as you can see the Error is really tiny. But I do not know why the calculation is not stopped when the Error is below 0.001.

I am trying to upload the mph file, so you can see it directly.
Hi Jeff, Thank you so much for the answering. I attach three pictures which you can have a closer look at the problem I am encountering. Picture 1 shows the model and periodic flow boundary, and Picture 2 shows the Solver Configuration where you can see, I just adopt the default settings and the relative tolerance is 0.001. Picture 3 shows the Converge plot, as you can see the Error is really tiny. But I do not know why the calculation is not stopped when the Error is below 0.001. I am trying to upload the mph file, so you can see it directly.


Please login with a confirmed email address before reporting spam

Posted: 8 years ago 23 feb 2017, 20:47 GMT-5
Dear Jeff,

I solve it according to your comments, thank you so much.
I have not got your meaning until I have a try.

Thank you again, it really helps.

Jiaqing
Dear Jeff, I solve it according to your comments, thank you so much. I have not got your meaning until I have a try. Thank you again, it really helps. Jiaqing

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.