CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   non-positive volumes (

john March 2, 2001 17:53

non-positive volumes
I keep getting error checking my mesh.what is a non-positive volume anyway?!! john

John C. Chien March 2, 2001 18:25

Re: non-positive volumes
(1). It simply says that the volume computed, based on the conventional definition, is negative. (2). Since the negative volume is not acceptable, the mesh generated is not useful. (3). In other words, the mesh generated is no good for the solver. (4). Based on my experience in using the commercial codes, this failure is very common. I normally had to try at least 30 to 50 times in order to get an acceptable mesh. (5). In case I could not get any solution, I normally send the file to the vendor to get it fixed. (6). It is hard for someone to create a general mesh generation code without knowing your problem in the first place. (using a commercial cfd code is a double guessing game, the vendor of the code is guessing the user's problem, and the user of the code is guessing the vendor's code capability. there is no answer to this problem. this normally does not happen when you write your own mesh generation code for your own problem.) (7). My comment is: you are likely to get lost in Los Angeles downtown, if you are using the map of L.A. to get around. You will have to use Yahoo Internet map and get the detailed instruction to go from one place to another. You also need to zoom in and out the Internet map several times to get the correct route. Somehow, the cfd mesh generation is not like Internet map yet. So,...

john March 2, 2001 19:34

Re: non-positive volumes
Thankx John for your suggestions.The problem is I don't think the volumes are negative .I checked the mesh in Gambit and I also checked the volume there ,I couldn't see any problem with them.Though when I change the mesh from hex to tet it works fine!!any reason for that? I'm just too frustrated with Gambit. john

John C. Chien March 3, 2001 03:33

Re: non-positive volumes
(1). I had similar problems with other mesh generation codes. (2). In 3-D, sometimes it is hard to "see" the bad cells on the screen in the mesh generation code. (3). The other possibility is that the solver had a different requirement. (4). Sometimes it is possible to locate the bad cells and check the angles, volumes, or skewness,etc.. It is a good idea to check with the support engineer. (5). I am currently not using any of the commercial codes. I had enough problems with these commercial codes already.

Max Nutini March 5, 2001 03:35

Re: non-positive volumes
John, try to translate the whole domain of a very small, negligible quantity, starting from the y direction and see what happens. It sometimes happen to me to face the problem you're having, and this is a method that usually works. Bye Max

Flav March 5, 2001 10:27

Re: non-positive volumes
I suggest you try this, if you haven't done it yet : in the text-user interface,


Then check your mesh again, the problem should be fixed.

Althea March 5, 2001 12:47

Re: non-positive volumes
Hello John,

Is your geometry modelled as axisymmetric? If so you will need to be very careful that all the x and y values are positive and the z values are zero. Sometimes zero values of -0.000000 are too zero so have to be corrected to make them +0.000000

I have had problems with negative volumes when using axisymmetric geometries.

Good luck. I hope you resolve the problem.


john March 5, 2001 16:02

Re: non-positive volumes
Thanx all,actually I tried to translate the geometry or /grid/modify-zones/repair-face-handedness and also my geometry is not axisymmetric!!!I guess that makes it really hard to solve ,eh? john

Shyam Kishor March 5, 2001 19:04

Re: non-positive volumes
Please contact your local FLUENT Support Engineer for help.

Jack Keays March 12, 2001 08:17

Re: non-positive volumes

I have experienced similar problems with 3D meshes. I have checked them in Gambit and also in Fluent and they are fine. Then when I scale the mesh I get negative volumes. It seems that FLuent has scaling problems, so perhaps this is due to round-off error. I think it might be that my mesh size was too small. Correcting face-handedness doesn't always work I'm afraid. Good luck.

All times are GMT -4. The time now is 12:55.