CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   My 3D interFoam test (http://www.cfd-online.com/Forums/openfoam-solving/60548-my-3d-interfoam-test.html)

zou_mo July 18, 2005 22:23

I use interFoam to test the 3D
 
I use interFoam to test the 3D Rayleigh-Taylor problem with a initial random interface. Top and bottom patches are set to wall. The other patches are symmetry planes. The density-ratio is 200:1.

In my result interface, besides the bubbles and spikes, I find many other little slices which are seperated from the main interface. Is the result resonable?

zou_mo July 18, 2005 23:21

This is a picture of the resul
 
This is a picture of the result: http://www.cfd-online.com/OpenFOAM_D...your_image.gif.

Thanks.

henry July 19, 2005 04:21

What combination of differenci
 
What combination of differencing schemes have you tried and what values for the interface compression factor cGamma?

What are the viscosities and surface tension?

zou_mo July 19, 2005 09:04

The differencing schemes,and t
 
The differencing schemes,and the interface compression factor cGamma are the default value. The nu is 0.0111, the density is 4.99 and 998, sigma is 0.005.

When the initial inerface is a simple cos shape, the result interface is well captured without small slices.

Thanks

henry July 19, 2005 09:20

I am unhappy with the concept
 
I am unhappy with the concept of "default value", we did not intend for the settings that FoamX has when it starts up to be considered reliable "default values" they are simply valid values which it needs to start. In the next version of FoamX I am going to add functionality so that "null" values are given for everything at start-up and require all settings to be specified by the user before the case can be saved and run.

It really is the users responsibility to choose all physical properties and operatiing parameters for CFD runs, defaults simply aren't acceptable because all cases are different and may require different choices depending on a wide range of considerations including stability, reliablility, accuracy, mesh quailty, case complexity, computational cost, users requirements for the solution etc. etc.


All times are GMT -4. The time now is 11:45.