CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ICEM] Volume mesh smoothing problem -DLR F6 (

Far April 22, 2012 11:47

Volume mesh smoothing problem -DLR F6
Dear All

Right now I am practicing the prism-tetra part of ICEM. I am practicing ICEM 14 Workshop tutorials and geometry is DLR F6. I am sucessful in creating all steps e.g. shell meshing, volume mesh, prism, density box and smoothing tet-4 elements.

But I am stuck in volume meshing step for quad-4, penta-4, tri-3 and as per insturction I have applied smoothing twice but nothing happens. Any one can suggest me, where I am getting it wrong.

This is what I get:

This is what, should be achieved:

energy382 April 23, 2012 04:12

I can't see the attached pics!?

siw April 23, 2012 05:22

The way I meshed this airfame (well the Common Research Model, but there very similar) was to first set-up the surface, volume and density region sizings based in the Octree criteria. Then I deleted the volume elements and smoothed the surface triangles until they would not improve any further (but I never used the laplace smoother as I found it made things worse). From that surface mesh I used the Delaunay (with TGlib/Use AF option switched on) to re-fill the volume. Again, I smoothed the volume tetras until no further improvment (no laplace again). Next I generated the prism layers to keep layering until it reached a smooth volume transition to the nearby tetras. The make a few floating prism layers and spliting/re-distributing does suit for this job I found. Then I smoothed it (sometimes freezing the penta elements). I did end up with pyramid elements due to the prism layers.

When smoothing the final mesh I freeze the pentas and aim to get a qualilty > 0.3. When I un-freezed the pentas I only smoothed up to 0.1 or it may have been lower.

I used ICEM v 13. As this was over 1 year ago so options my be different to v 14. I made posts with pictures asking for help with his mesh and PSYMN gave a loads of useful tips.

I biggest issue I found making a hybrid mesh was I could not get a fine resolution on the blunt trailing edges without getting a very high element total so I under resolved it with only 1 or 2 elements on the trailing edge height. Hexa does not have this issue.

Far April 23, 2012 06:27

Thanks for help. I shall try this method for this model as well.

But at the moment, I am following the tutorial and I assume that I should get what is depicted in tutorial.

I am going to attach the files with shell meshing and volume meshing with/without smoothing.

Thankyou again for your helpful hints and consideration.

Far April 23, 2012 06:28


Originally Posted by energy382 (Post 356329)
I can't see the attached pics!?

Try again, may be at that time "imageshack server" was down.

Far April 23, 2012 06:31

ICEM 13 and ICEM 14
This may be due to that fact that I am practicing ICEM 14 tutorials on ICEM 13 ;)

All times are GMT -4. The time now is 22:08.