CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Non positive volume exist - How to find where ??!! (http://www.cfd-online.com/Forums/fluent/41794-non-positive-volume-exist-how-find-where.html)

 Cyril July 20, 2006 05:02

Non positive volume exist - How to find where ??!!

I've just finished my mesh, and fluent tells me that "non positive volume exists".

Houw can I do to find where, and correct it ???

Many thanks !

 Zafer Zeren July 20, 2006 07:21

Re: Non positive volume exist - How to find where

When I got the same problem, I always make smoothing and then read the mesh file into fluent again. I use ICEM CFD. That can work.

zafer

 Cyril July 20, 2006 07:59

Re: Non positive volume exist - How to find where

How do you smoth ??

(the part of my geometry which causes the problem is a single parallelogram, meshed in tri...! there is no curve!)

 Zafer Zeren July 20, 2006 08:28

Re: Non positive volume exist - How to find where

I use ICEM CFD to generate mesh. Smoothing is under "edit mesh" menu. I have structured + unstructured hybrid mesh around a bluff body. You can there smooth you mesh.

 Cyril July 20, 2006 08:34

Grrrr !

The problem, is that I'm running a 2D case... I don't understand why I've got a volume (negative or positive...)

I'don't have icem cfd...

 Jason July 20, 2006 09:07

Re: Grrrr !

Is it axisymmetric? If so, then you're problem is that you're below the X-Axis. In Fluent 2Daxi model, you must use the X-Axis as your axis of rotation, and your geometry must all lie in the +Y direction (above the X Axis).

Hope this helps, and good luck, Jason

 Cyril July 20, 2006 09:27

Re: Grrrr !

Thanks, but the face involved is far from the X axis. when I export the face alone (no mesh) everything's ok.

And when I mesh it -> I get a neg volume...

 Cyril July 20, 2006 10:35

find out !

"Gambit error, not found in gambit, but detected by fluent" the hotline said !

 Tayfun July 21, 2006 03:24

Re: find out !

I had the same problem in FLUENT, here is the way I solve it:

In FLUENT, you type(!) the following;

1. g 2. mz 3. rfh

This converts the negative volumes to positive for 2D case.

NOTE: "g" means "grid", "mz" means "modify zones" and "rfh" means "right face handedness"

 mayur October 30, 2009 09:50

Dear Tayfun,

I tried to follow the steps of Grid, modify zones, rfh. It says the negative volume have been converted but after checking the grid, it still says negative volume exists. Can you please help me in this.

 herntan November 4, 2009 01:17

I think maybe u can try to round/fillet all those edges.
sharp edges can give many problem while meshing.

Anyhow, just a suggestion.

 -mAx- November 4, 2009 07:46

If you have negative volume, the best way is to correct your mesh.
If you work with Gambit go back over there, and see where are those cells
Examine Mesh / Volume ... set the upper limit to 0
It will show you all the negative cells
Then correct your topology for avoiding this issue

 mayur November 4, 2009 08:01

Actually when I create the mesh, it shows positive volume. but as soon as I start the iteration in Fluent, it shows negative volume detected. I even checked the mesh parameters. The Grid info before iteration shows, Xmin = 0, Xmax=0.084 and Ymin=0, Ymax=0.09. After starting iterations, when I check the grid values, the values of Xmin and Ymin are changed. I even tried translation of the coordinates in Fluent, but it doesnt help:confused:.

 -mAx- November 4, 2009 08:16

*moving mesh?
*2d-axi?

 mayur November 4, 2009 08:22

Ya.. I am trying a moving mesh. Actually I am trying to compress fluid using fluent. I am using a simple rectangular geometry. The top wall, the side walls which are deforming and the bottom wall is the moving wall. I have drawn the geometry in a positive XY co-ordinate system.

 -mAx- November 4, 2009 08:34

ok, then your initial grid is ok, and your issue remains in the MDM-set up.
you have to set the moving wall as rigid body (here you have to give the motion with a profile or a UDF)
The side walls (adjacent to moving wall) as deforming
and the last one as stationnary.
Pay attention to meshing options panel, bad values will generate bad mesh

 mayur November 4, 2009 08:38

Thanks a lot max.

I will try these steps.
If I have any trouble, I will ask you again :)

 -mAx- November 4, 2009 08:44

 mayur November 4, 2009 08:45

Max,

I have one more query, I am defining the zones as you explained. While,defining the meshing options, we have to describe the cell height. Is there any way to find the optimum value as per the geometry. My current geometry is 84mm in length and 90 mm in height.