CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Mesh distortion close to cylind wall-GAMBIT 2.2.30 (

Mirek April 13, 2006 18:38

Mesh distortion close to cylind wall-GAMBIT 2.2.30
Hello everybody I have a problem with Gambit 2.2.30 when I try to perform extra fine mesh (1e-5) close to the wall of cylinder (circle 2-D case). I built geometry using square and circle placed in the center then I subtract the circle from square (it's a part of discrete area). I divided such a face into four pieces (faces).I tried to mesh this small face using submap (the opposite edges have the same number of mesh elements) in two ways but failed: 1. Create boundary layer and then perform mesh face but the mesh and the boundary layer don't meet each other i. e. the mesh is abnormally skewed close to the wall. There aren't straight radial lines but zigzaged irregular lines! 2. Create extra fine mesh close to the wall by the aid of first length scheme (=1e-5). In this manner I'm obtaining better result but very close to the wall radial lines are similar to waved lines and try of smoothing cause worse effect.

I want to mention that I don't remember that I had such problems when I working on previous versions of Gambit. Boundary layers were working very well. Very strange situation take place when you try to perform boundary layer on whole not dividededge edge of circle - farther lines at vertex are very distorted. It appeares something like long triangle with top diverted in normal direction to the wall (when you create about 15-20 rows). Why? I don't understand.

Some details: a) square side = 22mm b) circle diameter = 12mm c) mesh spacing on circle edge = 0.2mm d) number of mesh elements on square edge results from (c) e) max growth factor b/a=1.2 f) quadrilateral mesh elements

Could enyone meet such a strange problem? Thanks in advance for any kind of help.

Mirek April 15, 2006 17:22

Re: Mesh distortion close to cylind wall-GAMBIT 2.
I helped myself simply changing defaults settings - Edit\Edit defaults\Mesh face ...

All times are GMT -4. The time now is 21:37.