Discussion ClosedThis 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.

coupled solver

Please login with a confirmed email address before reporting spam

Hi everyone

I am using comsol multiphysics 3.3a. I have balance equations for mass, momentum and energy and I have used non-conservative conduction convection equation along with non-newtonian flow equations to simulate them. I am trying to simulate time dependent behaviour.
The constitutive relationships I am using have gradient terms. I have tried solving them but i have seen ripples in the density profile. I guess comsol has got only segregated kind of solver with all equations solved sequentially.
Is there any option of coupled solver as we get in fluent?

regards

Nadeem

6 Replies Last Post 2009年7月29日 GMT+2 15:45
Jim Freels mechanical side of nuclear engineering, multiphysics analysis, COMSOL specialist

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 2009年7月29日 GMT+2 04:54
COMSOL has several types of solvers to choose from. If you choose a segregated solver, then that is what will occur. However, if you choose a stationary solution type, then pick PARDISO or UMFPACK solvers, this is a fully coupled direct solver, which is the best you can get for full coupling. However, if your problem is too large to fit in the computer memory, such as the case for most 3D problems, you will need to choose one of the iterative solvers to reduce memory. You have a choice, and it is not a fixed set of options.
COMSOL has several types of solvers to choose from. If you choose a segregated solver, then that is what will occur. However, if you choose a stationary solution type, then pick PARDISO or UMFPACK solvers, this is a fully coupled direct solver, which is the best you can get for full coupling. However, if your problem is too large to fit in the computer memory, such as the case for most 3D problems, you will need to choose one of the iterative solvers to reduce memory. You have a choice, and it is not a fixed set of options.

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 2009年7月29日 GMT+2 10:50
thanks James for the help.

I am solving the time dependant or unsteady case, for steady case I have got good results. Can we have coupled solver for time dependant simulations?

Nadeem
thanks James for the help. I am solving the time dependant or unsteady case, for steady case I have got good results. Can we have coupled solver for time dependant simulations? Nadeem

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 2009年7月29日 GMT+2 14:02
if both models (time/steady state) are the same, try to use the solution of the steady state as the initial values for the time dependent computation. This can help convergence.

Cheers
if both models (time/steady state) are the same, try to use the solution of the steady state as the initial values for the time dependent computation. This can help convergence. Cheers

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 2009年7月29日 GMT+2 14:14
thnks guys for ur help.

But the time dependant model is different from the steady state one. Also I am interested in the evolution of the solution with the time. The solvers seems to solve the equations in segregated form not in couple form. I can't find any option in comsol where i can used coupled solver. Also is there any option of implicit solver for time-dependant simulations? If i want to use dual time stepping is it avaiable in comsol?

cheers

Nadeem
thnks guys for ur help. But the time dependant model is different from the steady state one. Also I am interested in the evolution of the solution with the time. The solvers seems to solve the equations in segregated form not in couple form. I can't find any option in comsol where i can used coupled solver. Also is there any option of implicit solver for time-dependant simulations? If i want to use dual time stepping is it avaiable in comsol? cheers Nadeem

Jim Freels mechanical side of nuclear engineering, multiphysics analysis, COMSOL specialist

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 2009年7月29日 GMT+2 14:55
The set up for the solver type is the same for steady-state as is the time-dependent or transient. For the transient case, there is an additional menu and parameters to control the time stepping. You can use the fully-coupled direct solver in the transient case as well.

Perhaps you should follow some of the tutorials or read in the manual if it is not intuitive from the GUI.
The set up for the solver type is the same for steady-state as is the time-dependent or transient. For the transient case, there is an additional menu and parameters to control the time stepping. You can use the fully-coupled direct solver in the transient case as well. Perhaps you should follow some of the tutorials or read in the manual if it is not intuitive from the GUI.

Please login with a confirmed email address before reporting spam

Posted: 2 decades ago 2009年7月29日 GMT+2 15:45
James i have looked at the documentation but i am not sure if the direct solver uses dual time stepping? can you suggest the solver and additional options for the coupled solver. I think I have not been able to explain the segregated and coupled solvers. For example Fluent solves all the equations together in coupled solver where as in segregated solver it solves momentum first and then updates the rest of the field variables in a sequence of steps.
Also i cant see any implicit solver options in the comsol. I can't find any option which can give me implicit and explicit solver options.

Looking for a detail reply.

Nadeem
James i have looked at the documentation but i am not sure if the direct solver uses dual time stepping? can you suggest the solver and additional options for the coupled solver. I think I have not been able to explain the segregated and coupled solvers. For example Fluent solves all the equations together in coupled solver where as in segregated solver it solves momentum first and then updates the rest of the field variables in a sequence of steps. Also i cant see any implicit solver options in the comsol. I can't find any option which can give me implicit and explicit solver options. Looking for a detail reply. Nadeem

Note that while COMSOL employees may participate in the discussion forum, COMSOL®software users who are on-subscription should submit their questions via theSupport Centerfor a more comprehensive response from the Technical Support team.

Baidu
map