[ewlks]

Hey, I've been having some success with cfMesh for 3D problems, and have recently started exploring the 2D cartesian meshing workflow to answer some specific questions about the numerics I'm using with a less expensive simulation.

Two features that I've made use of with 3D meshes don't seem to work in the 2D workflow, namely:

• layer optimization which smooths the boundary layer cell normals. This doesn't seem to run at all in the cartesian2DMesh workflow.
• refinementThickness does result in greater refinement near the specified patch, but the transition between refinement layers is far more abrupt than in the cartesianMesh case.

I don't think its an issue with the meshDict (spelling, etc.) because with the same meshDict and a suitable geometry the cartesianMesh routine does generate these features.


Does anyone have any ideas? Is the cartesian2DMesh routine simply not as feature rich as the more general cartesianMesh case? Should I just give up, and generate a thin 3D mesh and then extrude a surface à la snappyHexMesh?



Thanks in advance for your help!

[einstein_zee]

Originally Posted by ewlks

Hey, I've been having some success with cfMesh for 3D problems, and have recently started exploring the 2D cartesian meshing workflow to answer some specific questions about the numerics I'm using with a less expensive simulation.

Two features that I've made use of with 3D meshes don't seem to work in the 2D workflow, namely:

• layer optimization which smooths the boundary layer cell normals. This doesn't seem to run at all in the cartesian2DMesh workflow.
• refinementThickness does result in greater refinement near the specified patch, but the transition between refinement layers is far more abrupt than in the cartesianMesh case.

I don't think its an issue with the meshDict (spelling, etc.) because with the same meshDict and a suitable geometry the cartesianMesh routine does generate these features.


Does anyone have any ideas? Is the cartesian2DMesh routine simply not as feature rich as the more general cartesianMesh case? Should I just give up, and generate a thin 3D mesh and then extrude a surface à la snappyHexMesh?



Thanks in advance for your help!

Hi there,

I personally never made much experiment with cartesian2DMesh but one thing I remember and just wanted to notify you (only to make sure things are done in proper way) is that it has a restriction which only works when your geometry lies in x-y plane.
As a suggestion, if you want to check the difference you may try to draw a simple sketch in 2D and mesh it with a sample meshDict (using cartesian2DMesh), and then for the second case just extrude the 2D sketch in the third direction and mesh it with the same meshDict (this time using cartesianMesh). In that case you can really say that things are working or not.

hope it helps...

[quarkz]

Hi,

I would like to add some pointers. You can download a tutorial guide from wolf dynamics:

http://www.wolfdynamics.com/training...cfmesh2017.pdf

Do take note that after using surfaceGenerateBoundingBox, you will need to edit the generated stl file and remove the part starting from this line:

solid zMin ...
...

Else, you will not be able to mesh and cartesian2DMesh will complain of errors.

Hope this helps.