CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   3d-pipe intersection in same plane (http://www.cfd-online.com/Forums/fluent/29406-3d-pipe-intersection-same-plane.html)

Pravir Kumar Rai February 25, 2002 19:18

3d-pipe intersection in same plane
 
I have drawn 3 cylinders(of diff.dimension) in same plane and united them using a sphere.But when i try to SMOOTH/SWAP it in FLUENT ,the error message says non positive volume exists.

Kindly tell me where i am wrong?

Erwin February 26, 2002 11:49

Re: 3d-pipe intersection in same plane
 
I have gotten negative volumes in the past by meshing a volume with tets and then have Gambit smooth it. I solved that by either 1) not smoothing in Gambit or 2) use different meshing intervals and play with the settings a bit. Optimizing the face meshes before meshing the volume can do wonders.

After reading the mesh into Fluent, run a Grid Check and confirm if the mesh contains these negative elements already from the beginning. Or check the mesh in Gambit to find these negative elements.

Also, face swapping inside Fluent does not work for hex or hybrid meshes. You can only use the smoothing option.


Pravir February 26, 2002 13:43

Re: 3d-pipe intersection in same plane
 
how do i prevent myself from not smoothing in Gambit because i dont do it myself as an operation. If i do not swap then my iterations are not starting.It says 'Divergence detected'.

Erwin February 26, 2002 14:56

Re: 3d-pipe intersection in same plane
 
Is the negative volume created in Gambit? If so, try changing the face mesh very slightly, e.g. by changing the interval 0.01 or by moving some nodes. Play with it. Then remesh the volume and check if negative elements keep appearing. If have found that stubborness is essential when using Gambit.

Divergence can be created by lots of things: bad initialization, forgetting to scale the grid, etc. Start slowly, for example with adiabatic incompressible flow and gradually turn on more features. Start with 1st order discretization and low URF values. Limit pressure and temperature values to realistic numbers.



All times are GMT -4. The time now is 02:47.