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 The information provided may be out of date.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.

Is the problem positive definite? Floquet periodic boundary conditions or PML using

Please login with a confirmed email address before reporting spam

Hello, I am doing some elastic wave simulations using solid mechanics module and piezoelectric module. And sometimes the DOF of my models are quite large so that I have to choose the solver manually to save memory and time. I have read the Comsol Multiphysics Reference Guide and realized that the choice of the best solver is depend on the type of the equation.
Here is my situation: I use Floquet-type periodic BC in eigenfrequency study and PML in frequency domain analysis very often, and because my lack of mathematical knowledge I don't know whether the problem is positive definite or not under these kind of boundary settings. (They are not ordinary boundaries that appear in my math textbooks) So could anyone please help me? Thanks!

2 Replies Last Post 6 févr. 2013, 07:59 UTC−5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 6 févr. 2013, 04:13 UTC−5
I believe that it's positive definite in eigenfrequency study with Floquet periodic boundary conditions since I always got real eigenfrequencies (can i draw such a conclusion?)

But in frequency domain with PML, models with large DOF, I tried several iterative solvers and preconditioners, still couldn't find a way to converge.
I believe that it's positive definite in eigenfrequency study with Floquet periodic boundary conditions since I always got real eigenfrequencies (can i draw such a conclusion?) But in frequency domain with PML, models with large DOF, I tried several iterative solvers and preconditioners, still couldn't find a way to converge.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 6 févr. 2013, 07:59 UTC−5
Hello.

You can reduce DOF by changing the order of the shape function (by default it is 2) but it will affect on the accuracy of the solution. I frequently use Generalized Alpha cause I think it"s more stable.
Hello. You can reduce DOF by changing the order of the shape function (by default it is 2) but it will affect on the accuracy of the solution. I frequently use Generalized Alpha cause I think it"s more stable.

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.