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 The information provided may be out of date.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.

Discretization schemes of Comsol stationary solver

Please login with a confirmed email address before reporting spam

Hi,
I would like to ask you how can I see what are the discretization schemes that a Comsol solver uses.
This information is of great importance for any numerical study but Comsol doesn't show this !
Thank you in advance.
PS: references are welcome
Mehrez

4 Replies Last Post 4 gen 2017, 12:00 GMT-5
Eric Favre COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 ott 2014, 11:31 GMT-4
Hello Mehrez,

your question is interesting and pretty large.
You mention "discretization schemes for a solver". That is actually probably a bit too wide.

A solver might be time dependent, stationary, eigenvalue. It can be as well an optimization solver, AWE, plug flow, time-discrete (this list is found in the documentation). And the problem can be non-linear, linear, parametric, and sometimes combinations of those! You might want to distribute your problem in a cluster (or on your own machine), make use of the adaptive mesh, create your own segregated algorithm, use domain decomposition, etc...

For instance, a time-dependent Navier-Stokes will make use of the time-dependent solver, that calls the non-linear solver, that calls the linear solver. At the linear solver level, you have several options : direct or iterative. Now you solve! For each of those choices, you have several options. Most of the time, you don't need to change the defaults since this should be taken care automatically by COMSOL. Sometimes you do, and you can easily.

Most of the general concepts that are related to Solvers are introduced in details in some classical references. Textbooks I personally recommend are those two :

- Multiphysics Modeling With Finite Element Methods, William B. J. Zimmerman
www.comsol.com/books/mmwfem
This one, although not updated with latest version, gives crystal clear explanations about finite elements in the context of COMSOL.

And the classical Numerical Recipies :
www.nr.com/

References given in COMSOL documentation are usually a bit more advanced. You can find them at the end of a chapter in COMSOL Reference Manual : "References for the Solution Operation Nodes and Solvers".
If you are looking for a basic understanding of the methods and how it works in their principles, more than a detailed explanation that needs a deeper understanding, you should find your way in the 2 references above.

I hope this helps,
good luck,
Eric Favre,
COMSOL France
Hello Mehrez, your question is interesting and pretty large. You mention "discretization schemes for a solver". That is actually probably a bit too wide. A solver might be time dependent, stationary, eigenvalue. It can be as well an optimization solver, AWE, plug flow, time-discrete (this list is found in the documentation). And the problem can be non-linear, linear, parametric, and sometimes combinations of those! You might want to distribute your problem in a cluster (or on your own machine), make use of the adaptive mesh, create your own segregated algorithm, use domain decomposition, etc... For instance, a time-dependent Navier-Stokes will make use of the time-dependent solver, that calls the non-linear solver, that calls the linear solver. At the linear solver level, you have several options : direct or iterative. Now you solve! For each of those choices, you have several options. Most of the time, you don't need to change the defaults since this should be taken care automatically by COMSOL. Sometimes you do, and you can easily. Most of the general concepts that are related to Solvers are introduced in details in some classical references. Textbooks I personally recommend are those two : - Multiphysics Modeling With Finite Element Methods, William B. J. Zimmerman http://www.comsol.com/books/mmwfem This one, although not updated with latest version, gives crystal clear explanations about finite elements in the context of COMSOL. And the classical Numerical Recipies : http://www.nr.com/ References given in COMSOL documentation are usually a bit more advanced. You can find them at the end of a chapter in COMSOL Reference Manual : "References for the Solution Operation Nodes and Solvers". If you are looking for a basic understanding of the methods and how it works in their principles, more than a detailed explanation that needs a deeper understanding, you should find your way in the 2 references above. I hope this helps, good luck, Eric Favre, COMSOL France

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 3 gen 2017, 22:22 GMT-5
Hello,When iam using the optimization solver to solve a topology optimization problem of eigenfrequency with MMA argorithm,i find i can not chose a study step of eigenfrequency ?and this question is really hindering my optimization.
best wishes to you
Hello,When iam using the optimization solver to solve a topology optimization problem of eigenfrequency with MMA argorithm,i find i can not chose a study step of eigenfrequency ?and this question is really hindering my optimization. best wishes to you

Eric Favre COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 4 gen 2017, 07:41 GMT-5
Hello,

Topology optimization with eigenfrequency seems a bit odd -I would recommend you to address your question to the technical support of COMSOL with more details.

Thanks,
Eric Favre
COMSOL France
Hello, Topology optimization with eigenfrequency seems a bit odd -I would recommend you to address your question to the technical support of COMSOL with more details. Thanks, Eric Favre COMSOL France

Eric Favre COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 4 gen 2017, 12:00 GMT-5
Correction : Topology optimization with eigenfrequency is perfectly valid.
Addressing your question to technical support might still be the best option!
Best regards,
Eric Favre
Correction : Topology optimization with eigenfrequency is perfectly valid. Addressing your question to technical support might still be the best option! Best regards, Eric Favre

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.