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.

What the desirable discretization methods with higher Rayleigh number ?

Please login with a confirmed email address before reporting spam

Hello all I am modelling natural convection in a closed enclosure 2D geometry. I employed two physics Laminar Flow+Heat Transfer in fluids. My simulation is well when Ra<1E6 but as soon as I make Ra = 1E8 the solution take a long time and iteration is finished without converage? What do you recommend for best discretization methods for Laminar flow and heat transfer may let me get the solution converage? knowing that I am running unsteady case with a time step of 0.001 s?

Many thanks


1 Reply Last Post 5 oct. 2021, 18:34 UTC−4
Jim Freels mechanical side of nuclear engineering, multiphysics analysis, COMSOL specialist

Please login with a confirmed email address before reporting spam

Posted: 3 years ago 5 oct. 2021, 18:34 UTC−4

Recommend fully coupled for any natural convection problem due to strong coupling between temperature and flow. As far as discretization method, using linear is OK, but will required more mesh detail. Using any higher order discretization (quadratic, cubic, etc.), will give you more accuracy for less mesh, but also puts higher demands overall due to the increased bandwidh and coupling in the Jacobian matrix. I would recommend doing the best you can with linear elements to start, then bump up the discretization order with the same mesh until your results are mesh independent and you are satisfied with the results.

-------------------
James D. Freels, Ph.D., P.E.
Recommend fully coupled for *any* natural convection problem due to strong coupling between temperature and flow. As far as discretization method, using linear is OK, but will required more mesh detail. Using any higher order discretization (quadratic, cubic, etc.), will give you more accuracy for less mesh, but also puts higher demands overall due to the increased bandwidh and coupling in the Jacobian matrix. I would recommend doing the best you can with linear elements to start, then bump up the discretization order with the same mesh until your results are mesh independent and you are satisfied with the results.

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.