CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ANSYS Meshing] CutCell Mesh & Inflation (http://www.cfd-online.com/Forums/ansys-meshing/104657-cutcell-mesh-inflation.html)

CaglarCoskun July 12, 2012 09:26

CutCell Mesh & Inflation
 
Hello
I am trying to do a simulate the natural convection in a cavity, having a complex geometry. I used the new Cut Cell method to mesh the geometry. I need to use inflation layers to resolve the thermal boundary layers better.
When I add the inflation layers, they create stair-stepping at various locations, and results a bad quality mesh. This causes convergence problems in Fluent.
I decreased the stair stepping problem by adjusting advanced inflation settings, but couldn't prevent it.
Any ideas?

Far July 12, 2012 15:22

cant u try ICEM?

CaglarCoskun July 12, 2012 15:33

I don't know how to use icem. Should I try? Is it better at handling complex geometries?

Far July 12, 2012 15:36

Rule of thumb: For simple geometries use Ansys meshing and for complex cases use ICEM CFD.

ICEM CFD is very good at handling complex geometries. You should try it and you can also get very good help at this forum from experts including The Simon .

PSYMN July 13, 2012 08:56

Cutcel is really in a separate category within ANSYS Meshing, it should be able to handle your complex geometry (it was designed for automotive underhood which is as complicated as it gets).

Are you using version 14? There was a lot of work on Cutcel prism between 13.0 and 14.0 and I noticed it got a lot better at handling these things.

Maybe post a picture so we can see what you mean. Typically, the mesher only allows prism stair stepping when you set your physics to CFD => CFX, so check that setting also.

CaglarCoskun July 14, 2012 08:44

Far and PSYMN, thanks for replies!
Here are some pictures of my mesh. In fact, in most parts of the geometry, there is no stair stepping. Instead of that, layers are compressed a lot.
http://postimage.org/gallery/123kqsxo/
My min. mesh size is 4mm, and number of inflation layers is 5.
PSYMN, it is not possible to change physics or solver preference if cutcell is active, as seen on pictures.

PSYMN July 16, 2012 16:52

I asked around in development for you and was told that they had a name for this (they called it the "corner shrink bug") and that it was sorted out back in May (therefore fixed in R14.5).

In the mean time,
Quote:

reduce gapfactor to 0.5...
Keeping Height over base below 0.5 is also recommended.

CaglarCoskun July 16, 2012 17:12

Quote:

Originally Posted by PSYMN (Post 371728)
reduce gapfactor to 0.5...

ah.. the gap factor is 2 in my mesh :) is it better to reduce it? I made it 2 because, in tutorials, it is written that increasing this value makes the inflation layers more compressed, thus reducing the stair stepping.

The other inflation parameters are as follows:
max. height over base: 0.5
max. angle: 180
fillet ratio: 0.5
fix first layer: no
inflation type: smooth transition

What can you suggest about these settings?

Thanks a lot for your help...

PSYMN July 16, 2012 17:32

Give it a try and tell us if it works...

I am not so expert at this (yet) that I can diagnose it remotely.

nurjs October 22, 2016 12:15

Hello Simon,
I need your help in meshing my geometry. I using different types of meshes tetrahedron with inflation, very fine, take alot of time to convegence.
cut cell but i fail to using inflation with it, ICEM good determinet and angle but it give me un expected results for Nusslet number.
So i feel i west my time without gooe results
please can you help me

Far October 22, 2016 12:32

why you dont try Hexa mesh?


All times are GMT -4. The time now is 00:53.