[Nek5000-users] Internal boundary condition

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

Dear Mohsen,

Years ago we did support internal walls, but that's no longer the case with the current workflow.

The issue is that the gridpoints across the interface are identified as being one and the same if
they are physically adjacent.  It would be possible to hack into your vertex numbering to make
these different (i.e., after they are generated by genmap), or to make a geometry in which they
are distinct by having a finite-thickness wall, but either way requires some application specific
coding in your usr file or modifications to your map file.

Paul

________________________________
From: nek5000-users-bounces at lists.mcs.anl.gov [nek5000-users-bounces at lists.mcs.anl.gov] on behalf of nek5000-users at lists.mcs.anl.gov [nek5000-users at lists.mcs.anl.gov]
Sent: Monday, April 21, 2014 6:12 PM
To: Nek 5000
Subject: [Nek5000-users] Internal boundary condition

Dear all,
I have a question regarding the internal boundary condition in the code.
My problem consists of a fan-shaped nozzle, which parts of its walls are located within the flow domain (a cylindrical tank with inflow/outflow boundary conditions). The question is how to set the boundary conditions for such internal zero-thickness walls?

I thought the walls could be defined as internal faces by creating a side-set (in the Cubit) and setting the velocity and the temperature fields of these faces to 'W  ' and 'E  ', respectively, similar to the solid-fluid interface  in the conjugate heat transfer example. Nevertheless, there is a fluid at both sides of the interface in this case.
Below is part of the boundary conditions in .rea file.
.....
.....
1 0     ! 1 fluid set, 0 other/solid sets
10     ! fluid set is 10
1 ! block 10 is material 1
10       ! no. bc sets (side-sets); bc set id,
100  1    W  , (external walls)
100  2    I  , (external walls)
120  1    W  , (internal walls corresponding to the nozzle)
***  *    *  * (internal walls corresponding to the nozzle)
200  1    v  , (inflow)
200  2    t  , (inflow)
300  1    O  , (outflow)
300  2    O  , (outflow)
400  1    SYM,
400  2    SYM,
.....
.....

When I run this case, warning messages appear during assigning the sidesets to the element faces, saying that: "face shared by 2 Hexes (say) 2331 5227". It seems that this warning is only because I defined some faces as "internal", however, no "other solid/fluids" has been defined in the input file.
Nevertheless, the issue with such definition is that, although, the velocity field is zero at the internal walls, there seems to be a pressure gradient across the wall, which is not physically consistent with internal walls, and drives a flow through the walls.

It would be appreciated if anyone could help me on this problem.

Thanks,
Mohsen

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20140422/860915d8/attachment.html>