[Nek5000-users] Question about method & error

[nek5000-users at lists.mcs.anl.gov]

Hi Frank,

Actually, the only thing that would be specified at the interface
is the surface tension sigma, which would generate a jump in the 
stress across the interface.  The fluid and surface would dynamically
accommodate to these stresses - and would come to a steady state
geometry if that's the what the dynamics dictates.   The temperature
would still satisfy the heat equation, and of course you could have
different rho-Cp and k in each region.

The code will generate the correct jump in pressure according to the
surface tension.   (We can check this with the case of a spherical
drop.)

Of course, w/o Lee around we'll have to sort all this out to make
certain it does what we expect... but I'm fairly confident that it
will.

Paul

> Paul,
> 
> Great to hear that.  Question; an internal interface has the meaning of
> an internal boundary condition at which velocity and temperature, but
> not pressure, boundary conditions can be set? 
> 
> Since the normal velocity at the interface is set to zero and the
> interface completely separates the two fluids, the absolute value of the
> pressure in the two fluids is independent in the model I have in mind.
> To put in other words, no pressure gradients across the interface exist
> in the model.  But I guess that will always be the case in the spectral
> element method, as long as the interface is not inside an element.
> 
> Cheers,
> Frank