search

idaholab / moose

Multiphysics Object Oriented Simulation Environment
https://www.mooseframework.org
GNU Lesser General Public License v2.1
1.76k stars 1.05k forks source link

3D Porous media Incompressible Finite Volume #18410

Closed liangyu-psi closed 2 years ago

liangyu-psi commented 3 years ago

Reason

This new feature is nice. But it seems only suitable for 2D calculation.

Design

It can be developed to 3D version.

Impact

Stronger demand for 3D calculations. It can be used to wider objects.

GiudGiud commented 3 years ago

It can do 3D already.

liangyu-psi commented 3 years ago

@GiudGiud Hi, Giud! Thanks for your information. From the equations, it seems that the constant physical properties are used. Could it be updated to the version whose physical properties changing with temperature?

GiudGiud commented 3 years ago

We will in the near future. For now this capability is only in Pronghorn You can request access to this application here: https://inl.gov/ncrc/

liangyu-psi commented 3 years ago

@GiudGiud That's great! May I have one more question? When I use a simple cubit 3D mesh. This module works. When I apply a unstructured mesh. Error happens. And the error is : matrix A is singular! Does it mean this module just suitable for structured mesh? Maybe there something wrong with my unstructured case. Hope to receive your professional reply.

GiudGiud commented 3 years ago

Is it a tetrahedral mesh? If so you need to turn two_term_boundary_expansion off for every variable.

This module works better with hex meshes as we don't have skewness correction

liangyu-psi commented 3 years ago

Yes, it's tetrahedral mesh. In other words, unstructured grids can be used. But, the convergence is not guaranteed in this situation? How about Polyhedral mesh? Should I use the same method?

GiudGiud commented 3 years ago

We can't do general polyhedral. The best mesh for this module right now is hexahedrons. Note that you can make an unstructured hex mesh. It does not need to be regular/structured.

tet meshes have issues with the two term boundary expansion.

liangyu-psi commented 3 years ago

@GiudGiud Thanks for your tips. I will try an unstructured hexahedral mesh.

GiudGiud commented 2 years ago

Closing this. Support for tetrahedral mesh is much better now, even with two terms boundary expansion. Still not support for general polyhedral, and recommandation is still that hexahedral is best.