How to control parameters like skewness etc. ?

Hi

I've close packed sphere geometry and I'm trying to generate really fine mesh for cfd calculation. I played a lot with the setting of the NETGEN 1D 2D 3D algorithm, but as you can see in the CheckMesh results the mesh is not well generated. I want to reduce max skewness below to 0.7 and max non-orthogonality below to 65 but my attempts have failed. Also attempts takes too long (I use Optimization steps 20). I tried that reduce growth rate to 0.08 and min mesh size too low but these are not helped me much. I also reduced the max mesh size a lot and reached 70 million cells, but the mesh was of poor quality too. Any suggestions are appreciated.

Mesh stats

    points:           6026479

    faces:            62568029

    internal faces:   57657431

    cells:            30056365

Checking geometry...

    Overall domain bounding box (-30.5 -30.5 -30.5) (30.5 30.5 30.5)

    Mesh has 3 geometric (non-empty/wedge) directions (1 1 1)

    Mesh has 3 solution (non-empty) directions (1 1 1)

    Boundary openness (4.08417e-16 -1.17489e-18 9.45926e-17) OK.

    Max cell openness = 2.27192e-15 OK.

    Max aspect ratio = 163.522 OK.

    Minimum face area = 1.6149e-07. Maximum face area = 0.382216.  Face area magnitudes OK.

    Min volume = 8.02769e-11. Max volume = 0.05664.  Total volume = 71563.7.  Cell volumes OK.

    Mesh non-orthogonality Max: 88.8249 average: 26.1101

   *Number of severely non-orthogonal (> 70 degrees) faces: 227858.

    Non-orthogonality check OK.

  <<Writing 227858 non-orthogonal faces to set nonOrthoFaces

    Face pyramids OK.

    Max skewness = 3.48082 OK.

    Coupled point location match (average 0) OK.

Mesh OK.

good luck, 

on would take time of it

Hi Said,

Can we have a look at the geometry first? Or an HDF file (or a TUI script) would be even better. Hard to say anything before that. It may be likely, that you need to do some tricks with the geometry first (e.g., split it into more elementary parts) in order to achieve a better mesh quality.

Cheers,

Roman

Hi Roman,

Thanks for your reply. After lot of tries I did this (How to fuse a lot of spheres and cylinders — SALOME Platform (salome-platform.org)) for fixing lot of meshing issue due to spheres contact points. The neck or bridge approximation is a good approach for this kind of complex geometries. But now I'm strugguling with fusing them together. Have you any suggestion for this?

Python script is attached.

Kind regards,
Said.

Said,

Do you stil have this problem ?

You might try another meshing algorithm that I interfaced with Salome. It is based on a variational rather than a Delaunay or a frontal approach so you might have different results. The new mesher is also multithreaded. You can find it at:

https://github.com/MoiseRousseau/Variational-Tet-Mesh

Also, can you send your salome sutdy with your geometry and possibility with the mesh if not too heavy (for example, https://wetransfer.com/) ? I created a small piece of software that allow optimizing mesh orthogonality, but I need further tests to get it works and increase its performance. I would like to test it on your mesh and share it with the community if it works.

Have a nice day,

Moise

Hi Moise,

Unfortunately I delete all mesh files because I changed the method that I generate the mesh. As you said with NETGEN these geometry can't be meshed properly. Now I'm generate just 2D mesh with SALOME and export as STL then I generate 3D mesh with OpenFOAM utility named snappyHexMesh. If you want previous Script files ofcourse I can share everything with you. 

But If I can solve the Fusing error and fuse the cylinders with spheres succefully, I will generate mesh with SALOME because I need grouping them by their height. With snappyHexMesh and a STL file this is not possible.

In short I'll generate the mesh when I'm done with this error. After that I'm gonna send you the files that you want.

Kind regards,
Said.

saidcatalbas@gmail.com