CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   specifying interface boundary condition (http://www.cfd-online.com/Forums/cfx/88512-specifying-interface-boundary-condition.html)

specifying interface boundary condition

Need help !!

I need to specify the following interface boundary condition :

(viscosity1)*d(velocity1)/d(x)=(viscosity2)*d(velocity2)/d(x)

where 1 and 2 represents two domains containing different fluids
the interface is planer

 ghorrocks May 19, 2011 07:44

Is the x direction prependicular to the surface? Then isn't this simply the shear stress on one side of the interface is the same as the other?

I cannot think of an easy way to implement this with two domains. You could do it as a multiphase free surface model, but freeze the volume fraction field using expert parameters. This would match the shear stresses across the interface.

yes, the x-direction is perpendicular to the interface
Actually, I am new to this software, please explain in detail

 ghorrocks May 19, 2011 18:48

Set up a multiphase simulation with one fluid as one phase and the other fluid as the other phase. Use a single domain. Set an initial conidtion on the volume fraction for the location of the two fluids. Use the expert parameter "solve volume fraction" set to false so the solver does not update the volume fraction field.

But I suspect the interface will actually move, so you could probably leave the volume fraction on as default and let the solver sort it out.

Do the CFX tutorials on free surface flow to understand setting it up.

Multiphase models take 2-3 days to solve just momentum equations. And here we have to solve momentum, mass tranfer and reaction equations too. So, I am trying it in single phase

Actually I am using single phase model with two domains. And the boundary condition is to given at the interface of the two domain.

 ghorrocks May 31, 2011 08:18

That's why I say to freeze the volume fraction equation and the run time will not be too much longer than a single phase model.