CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)

Hi everybody,

I want to apply this BC to my U:

Code:

`du/dx=-dv/dy , dv/dx=0`
which the u is the x-component and v is the y-component of U.

any body knows how can I define this BC?
any hint will valuable for me.

Thank you.

 immortality December 16, 2012 15:01

what shape does your boundary have?is it aligned with x or y axis or any other?
Can't you simplify the formula on the boundary?

Quote:
 Originally Posted by immortality (Post 397806) what shape does your boundary have?is it aligned with x or y axis or any other? Can't you simplify the formula on the boundary?
It's not a matter that my bc is horizontal or vertical, I want to know how can I apply such a bc.

 ata December 18, 2012 07:55

Hi
Is your problem 2-D? If yes first BC says that div(V)=0 so I think this may be a problematic BC. For the second one I think you can use OF BCs.

1 Attachment(s)
Quote:
 Originally Posted by ata (Post 398058) Hi Is your problem 2-D? If yes first BC says that div(V)=0 so I think this may be a problematic BC. For the second one I think you can use OF BCs.
ِDear ata,

These conditions are for only one boundary. I want to apply this two conditions for my boundary simultaneously.
yes, I's 2D.
the complete form of my bc is attached.

thank you.

 joenick December 18, 2012 09:40

Hello,
without knowing the context of your problem, a good answer will be difficult. Which process do you calculate? What are P, theta, in and out in the attached figure?

Have you tried to reformulate the BC using the PDE that is being solved?

Quote:
 Originally Posted by joenick (Post 398068) Hello, without knowing the context of your problem, a good answer will be difficult. Which process do you calculate? What are P, theta, in and out in the attached figure? Have you tried to reformulate the BC using the PDE that is being solved?
Thanks Joe,

Assume a cavity that its right wall is open and the conditions I refered above are the conditions at this open side. my solver is buoyantBoussinesqSimpleFoam and as I mentioned above the pressure is P=0 and the theta is the temperature.

What is your meaning of 'reformulation of BC using the PDE that is being solved'?

Thank you

 joenick December 18, 2012 13:12

Hello,
I'm afraid I cannot help, because I don't know this solver.

What I mean by reformulating the BC is to make use of the differential equations for U in your solver to simplify your BC. Just a thought.

If you explain why you choose this BC instead of showing only the equation, somebody might be able to help. What are you trying to solve conceptually, not mathematically?

dv/dx=0 seems to make assumptions about the normal direction of the boundary and can't be applicable in general unless you have some anisotropic forces.

thank you Joe,

As I mentioned in my last post, I want to model an open cavity. this kind of cavities have an open side that have special conditions that I want to model it. I tried to replace it with other boundary conditions and even with a huge space with special BCs that might make my desired BC. but all of them failed. :(
I will try if can I include this bc in my solver or not!!

 wyldckat December 24, 2012 09:10

Greetings to all,

Since adambarfi sent me an email the other day about this and although I'm not much of help here, here's what I know on the subject:
• Look into the source code of all of the other boundary conditions for clues on how to do this.
• For simulating open areas, AFAIK there are a few standard boundary conditions already in OF:
• Symmetry patch - probably not exactly what you're looking for, but it's possibly the closest you'll find in OF that will respect such a boundary condition.
• Atmospheric patches: "atmBoundaryLayerInletEpsilon", "atmBoundaryLayerInletVelocity" and "pressureInletOutletVelocity". For more ideas, run:
Code:

`find \$FOAM_TUTORIALS -name U | xargs grep -isl atm`
and check the tutorials listed by that command line.
Best regards,
Bruno

thank you Bruno,

I simplified my problem so that my problematic boundary has just one condition:

Code:

`dv/dx=0`
which the v is y-component of U.

any body knows how can I apply this condition?

Thanks

 All times are GMT -4. The time now is 08:01.