Dear Dante,


I typically use the turb_outflow boundary condition that is in navier5.f (see below). It will guarantee no inflow.


Sometimes you need to set "uin" to get the correct scale if your inlet and outlet areas are disparate in one way or another, but I've rarely found a case where this does not work.  Other groups use some form of convective outflow treatment.  There were comments on this just last week.


hth,


Paul




c     . Set div U > 0 in elements with 'O  ' bc.

c

c     . rq is nominally the ratio of Qout/Qin and is typically 1.5

c

c     . uin is normally zero, unless your flow is zero everywhere

c

c     . d and m1 are work arrays of size (lx1,ly1,lz1,lelt), assumed persistant

c

c     This routine may or may not work with multiple outlets --- it has

c     not been tested for this case.

c

c

c     TYPICAL USAGE -- ADD THESE LINES TO userchk() in your .usr file:

c                      (uncommented)

c

c     common /myoutflow/ d(lx1,ly1,lz1,lelt),m1(lx1*ly1*lz1,lelt)

c     real m1

c     rq  = 2.

c     uin = 0.

c     call turb_outflow(d,m1,rq,uin)

c


From: Nek5000-users <nek5000-users-bounces@lists.mcs.anl.gov> on behalf of nek5000-users@lists.mcs.anl.gov <nek5000-users@lists.mcs.anl.gov>
Sent: Tuesday, March 6, 2018 1:33:29 AM
To: nek5000-users@lists.mcs.anl.gov
Subject: [Nek5000-users] Outflow boundary conditions
 
Dear all,

are outflow advective boundary conditions available in nek5000? 

Or is there any work around in order to impose outflow boundary conditions
(for the velocity and the passive scalars) that can grantee that there is no flow coming 
back from the outflow boundary?

Thanks,
Dante