[Nek5000-users] Set gradient and/or pressure at outlet

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

Dear Juan Pablo,


If you are doing a Boussinesq approximation with an open boundary you probably need to subtract off the hydrostatic mode from the pressure, given that pressure in that case is far from uniform in the vertical.


Please see:


http://www.mcs.anl.gov/~fischer/users.pdf


eqs. (13) and (17) for an example of how to do this.


The key is to have the "pressure" variable be close to uniform at outflow.


Paul


________________________________
From: Nek5000-users <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: Tuesday, December 26, 2017 10:20:29 PM
To: nek5000-users at lists.mcs.anl.gov
Subject: [Nek5000-users] Set gradient and/or pressure at outlet

Dear Nek users, I've been struggling to program an adecuate outlet BC. I simulate a vertical isothermal surface in an open domain, forming a free convection boundary layer. When there are vortex leaving the domain, the code diverges. I've tried turb_outflow subroutine, stabilized outflow BC (Dong 2014), convective BC (sommerfield radiation condition, thanks to Phillip and Eric, from this comunity), etc. But I can't achieve to make the code run propperly, and I can't go on with my research while I have this problem.

There is a paper, from Desrayaud (2013), "Benchmark solutions for natural convection flows in vertical channels submitted to different open boundary conditions".

I want to try these BC, which can be stated as: (x: horizontal and y:vertical.)

At the top face (outlet)

Ux=0
dUy/dy=0

If Uy > 0 (fluid going out)
dTEMP/dy=0
pressure=0

If Uy < 0 (fluid going in because there is backflow)
TEMP=0
pressure= -0.5Uy^2 or 0   (these are two options in the paper)



As far as I know, in Pn-Pn-2 formulation there are not pressure BC, but it is possible to set it using Pn-Pn. Am I right? If so, how can I set BC for pressure in Pn-Pn?

And for the velocity gradient and temperature gradient equal cero at the top face, how can it be done? Would it be enough to set dirichlet BC and replicate the last row of values to the last one?

Any help is welcomed, thanks in advance,

Juan Pablo.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20171227/27224a34/attachment.html>