CFD Online Discussion Forums - [snappyHexMesh] nCellsBetweenLevels

Hello,

I'm experiencing a problem with the parameter nCellsBetweenLevels.
I was first using nCellsBetweenLevels 2 for a given mesh. I needed to increase it to nCellsBetweenLevels 3 and I obtained a mesh with less cells than in the first mesh.

The first thing I checked is the maximum number of cells allowed in sHM. There is no problem on that side, the limit is much higher than the mesh size.

I then looked into the log from sHM. The castellation step is split in 4 stages:
1. Feature edges refinement
2. Surface refinement
3. Cell removal inside the body
4. Shell refinement

I compared the number of cells at the end of each stage and the difference appears at the beginning of the "Shell refinement" stage where the number of cells selected for internal refinement is lower in the second mesh than in the first mesh (while it should be the opposite!).

Has anyone already experienced something similar? Do you have an idea what this is due to?

I appreciate your help!
Eloïse

Mesh 1 (with nCellsBetweenLevels 2):

Code:

Shell refinement iteration 0

----------------------------



Marked for refinement due to refinement shells    : 65967 cells.

Determined cells to refine in = 4.65 s

Selected for internal refinement : 983327 cells (out of 6561575)

Edge intersection testing:

    Number of edges             : 44510076

    Number of edges to retest   : 39665421

    Number of intersected edges : 3887457

Refined mesh in = 9.16 s

After refinement shell refinement iteration 0 : cells:13444864  faces:44510076  points:17659672

Mesh 2 (with nCellsBetweenLevels 3):

Code:

Shell refinement iteration 0

----------------------------



Marked for refinement due to refinement shells    : 65967 cells.

Determined cells to refine in = 3.74 s

Selected for internal refinement : 74967 cells (out of 6561575)

Edge intersection testing:

    Number of edges             : 24200235

    Number of edges to retest   : 2389770

    Number of intersected edges : 2975683

Refined mesh in = 4.2 s

After refinement shell refinement iteration 0 : cells:7086344  faces:24200235  points:10165443