3D dam breaking in a box

koWUalsky · February 25, 2010, 03:07

Hi, everyone,
I am just a beginner on using OF. I made a "3D dam breaking in a box" for practice.

Here is the animations I made,
http://sites.google.com/site/kowualsky/3dKEdamBreak.wmv
http://sites.google.com/site/kowuals...EdamBreak2.wmv

It used the k-epsilon turbulence model.

I am wondering that in VOF, the freesurface should be at alpha=0.5, right? But in this case i made, the iso-surface of alpha=0.5 is very strange. It seems alpha=0.9 or 0.99 (as i shown in the animations) is better to represent the free surface. Why?

And, the result may be beautiful. It showed the bubble and splash. They will go back to freesurface. Basically, the flow field is symmetric against x=z. But I am also wondering that when bubbles rise back to the freesurface, they should produce jet, right?

Best regards,
Wu

ata · February 26, 2010, 06:30

Dear koWUalsky
Hello
Because of my low speed internet connection I can't download your films but it is true, alfa=1 is interface.
Best regards

koWUalsky · February 26, 2010, 16:43

Dear ata,
Thanks for your reply. Now, the data is not avarible for me. But I will show some pics after back to my lab.

Maybe my setting is something wrong.

For choosing alpha=0.5, it got reasons. I had a little experience in Fluent. We wanted to verify its VOF by computing the submerged foil. Compared with experient (from Dr. Duncan) and level-set method (from Dr. Young and Dr. Stern), the alpha=0.5 matched well. So, it should be alpha=0.5.

Plus, I thought that because of no clear interface for defining the free surface in VOF, it is a distribution, the free surface should be located beween air and water. So, the alpha=0.5 makes sense. The alpha=1.0 means it is totally water.

Recently, I also try to learn the simulation of wave elevation surrounding the ship. I download and refer the OF tutorial made by Dr. Paterson's lab. They also suggest the alpha=0.5 as the free surface.

Actually, I always think about definition of the free surface in VOF. Theoretically, the interface for being the free surface should satisfy the free surface kinematic and dynamic condition. But in VOF, it didn't consider this part explicitly. The level-set method solves this part, but VOF doesn't.(However, the single-phase level set method doesn't consider the air flow field, but VOF does.) We might or should find the streamline which satisfies the free surface conditions. it may not be easy. Then, it would be the free surface. So, I am thinking that if the alpha=0.5 is a streamline, it should be the free surface. It is a streamline and the free surface, right?

Best regards,
Wu

ata · February 27, 2010, 00:40

Dear Ping-Chen Wu
Hello. I hope you are well.
As you know because of numerical diffusion, in VOF method there is not an strict interface between fluids especially in long times. If numerical method hasn't numerical diffusion alfa=0.5 and alfa=1 are almost same but in common cases that numerical diffusion exist it seems that you are right and it is better to select alfa=0.5 as interface.
For your last question as I see, if there is no normal velocity to the free surface, free surface is stream line, else, it in not.
All the best

Ata

February 25, 2010, 03:07	3D dam breaking in a box	#1
koWUalsky New Member Ping-Chen Wu Join Date: Feb 2010 Posts: 4 Rep Power: 16	Hi, everyone, I am just a beginner on using OF. I made a "3D dam breaking in a box" for practice. Here is the animations I made, http://sites.google.com/site/kowualsky/3dKEdamBreak.wmv http://sites.google.com/site/kowuals...EdamBreak2.wmv It used the k-epsilon turbulence model. I am wondering that in VOF, the freesurface should be at alpha=0.5, right? But in this case i made, the iso-surface of alpha=0.5 is very strange. It seems alpha=0.9 or 0.99 (as i shown in the animations) is better to represent the free surface. Why? And, the result may be beautiful. It showed the bubble and splash. They will go back to freesurface. Basically, the flow field is symmetric against x=z. But I am also wondering that when bubbles rise back to the freesurface, they should produce jet, right? Best regards, Wu Last edited by koWUalsky; February 25, 2010 at 12:48.

February 26, 2010, 06:30	3D dam breaking in a box	#2
ata Senior Member ata kamyabi Join Date: Aug 2009 Location: Kerman Posts: 323 Rep Power: 18	Dear koWUalsky Hello Because of my low speed internet connection I can't download your films but it is true, alfa=1 is interface. Best regards

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
CFD Animations of breaking waves	Dommermuth	Main CFD Forum	0	June 17, 2009 12:47
[OpenFOAM] Visualization of breaking of a dam tutorial	lasb	ParaView	2	July 13, 2007 09:15
Immersol Simulation of a Heated Box	Dong	Phoenics	0	March 2, 2006 22:20
Dam Breaking Problem	B Godderidge	Siemens	0	May 19, 2005 05:35
Diverge in 2d dam breaking case	Jason	FLUENT	0	November 1, 2002 04:12

February 26, 2010, 16:43		#3
koWUalsky New Member Ping-Chen Wu Join Date: Feb 2010 Posts: 4 Rep Power: 16	Dear ata, Thanks for your reply. Now, the data is not avarible for me. But I will show some pics after back to my lab. Maybe my setting is something wrong. For choosing alpha=0.5, it got reasons. I had a little experience in Fluent. We wanted to verify its VOF by computing the submerged foil. Compared with experient (from Dr. Duncan) and level-set method (from Dr. Young and Dr. Stern), the alpha=0.5 matched well. So, it should be alpha=0.5. Plus, I thought that because of no clear interface for defining the free surface in VOF, it is a distribution, the free surface should be located beween air and water. So, the alpha=0.5 makes sense. The alpha=1.0 means it is totally water. Recently, I also try to learn the simulation of wave elevation surrounding the ship. I download and refer the OF tutorial made by Dr. Paterson's lab. They also suggest the alpha=0.5 as the free surface. Actually, I always think about definition of the free surface in VOF. Theoretically, the interface for being the free surface should satisfy the free surface kinematic and dynamic condition. But in VOF, it didn't consider this part explicitly. The level-set method solves this part, but VOF doesn't.(However, the single-phase level set method doesn't consider the air flow field, but VOF does.) We might or should find the streamline which satisfies the free surface conditions. it may not be easy. Then, it would be the free surface. So, I am thinking that if the alpha=0.5 is a streamline, it should be the free surface. It is a streamline and the free surface, right? Best regards, Wu

February 27, 2010, 00:40		#4
ata Senior Member ata kamyabi Join Date: Aug 2009 Location: Kerman Posts: 323 Rep Power: 18	Dear Ping-Chen Wu Hello. I hope you are well. As you know because of numerical diffusion, in VOF method there is not an strict interface between fluids especially in long times. If numerical method hasn't numerical diffusion alfa=0.5 and alfa=1 are almost same but in common cases that numerical diffusion exist it seems that you are right and it is better to select alfa=0.5 as interface. For your last question as I see, if there is no normal velocity to the free surface, free surface is stream line, else, it in not. All the best Ata