accuracy of weak solutions

Please login with a confirmed email address before reporting spam

Finite elements give weak solutions, that is, the solutions to a PDE are only correct in its integral form. Is it possible that in comsol, the solution differs a lot from the analytic one locally while it's only accurate in its integral form (globally)?


3 Replies Last Post 29 sep. 2021 17:00 GMT−4
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 months ago 27 sep. 2021 13:30 GMT−4

Hello Carl,

The finite element method is one of several numerical techniques used by COMSOL in arriving at an approximate solution to equations. You are correct that if employed improperly, any numerical method can yield an approximate solution that is further away from the exact solution than the user may presume. Theorists and practitioners have long studied how to quantify and minimize the error. At a minimum, performing a mesh refinement study is recommended. You will find many works on this topic in any serious textbook on the finite element method. There is a short introduction to the topic in the COMSOL Multiphysics Cyclopedia.

Best regards,

Jeff

-------------------
Jeff Hiller
Hello Carl, The finite element method is one of several numerical techniques used by COMSOL in arriving at an approximate solution to equations. You are correct that if employed improperly, any numerical method can yield an approximate solution that is further away from the exact solution than the user may presume. Theorists and practitioners have long studied how to quantify and minimize the error. At a minimum, performing a mesh refinement study is recommended. You will find many works on this topic in any serious textbook on the finite element method. There is a short introduction to the topic in the [COMSOL Multiphysics Cyclopedia](https://www.comsol.com/multiphysics/finite-element-method?parent=physics-pdes-numerical-042-62). Best regards, Jeff

Please login with a confirmed email address before reporting spam

Posted: 2 months ago 28 sep. 2021 22:54 GMT−4

Hi Jeff, thanks so much for the info but what if mesh refinement doesn't help?

Hi Jeff, thanks so much for the info but what if mesh refinement doesn't help?

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 months ago 29 sep. 2021 17:00 GMT−4
Updated: 2 months ago 29 sep. 2021 17:04 GMT−4

Hi Carl,

I tried to hint in my original message that inappropriate meshing is only one of various possible causes for the numerical (hence approximate) solution returned by an FEA solver to deviate from the exact (but usually unknown) solution of a PDE. Solver settings could be another (think for instance of what happens if you set convergence tolerances too loosely: you're then basically asking the software to content itself with a relatively inaccurate solution). And of course there are other possible causes, some of which are specific to particular physics, such as the presence of singularities in the analytical solution, etc, etc. There are no shortages of textbooks on the FEA method; Zienkiewicz, Hughes, Bathe, to name just a few of the classics.

Best,

Jeff

-------------------
Jeff Hiller
Hi Carl, I tried to hint in my original message that inappropriate meshing is only one of various possible causes for the numerical (hence approximate) solution returned by an FEA solver to deviate from the exact (but usually unknown) solution of a PDE. Solver settings could be another (think for instance of what happens if you set convergence tolerances too loosely: you're then basically asking the software to content itself with a relatively inaccurate solution). And of course there are other possible causes, some of which are specific to particular physics, such as the presence of [singularities](https://www.comsol.com/blogs/singularities-in-finite-element-models-dealing-with-red-spots/) in the analytical solution, etc, etc. There are no shortages of textbooks on the FEA method; Zienkiewicz, Hughes, Bathe, to name just a few of the classics. Best, Jeff

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

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.