Boundary conditions for fluid

From: Bruce Boghosian (bmb@Think.COM)
Date: Thu May 27 1993 - 22:57:23 UTC


   Date: 27 May 1993 15:41:58 -0400
   From: butterwo@cs.unc.edu (Jeff Butterworth)

   In article <1993May27.071230.22564@crs4.it> ralf@grappa.crs4.it (Ralph Santos) writes:
   >> I am programming a fluid flow model using a hexagonal mesh.  Its a 2-D
   >> channel flow with a solid boundary at the 'top', a solid boundary at the
   >> 'bottom' and the flow from 'left' to 'right'. The solid boundaries are non-
   >> slip.
   >>
   >> I would appreciate hints on how to do the left and right boundary conditions.
   >>

The two methods mentioned so far are (1) using periodic boundary
conditions, and then driving the system with an artificial gravity, and
(2) introducing particles with different probabilities at the left and
right boundaries.  Indeed, both methods can be effective.

With regard to the first, it has been noted that the artificial gravity
can be a volumetric force, set up by introducing occasional
non-momentum-conserving collisions at random sites throughout the
medium.  The driving force per unit volume is then equal to the amount
of momentum introduced per unit volume per unit time.  Note that there
has to be some obstacle when you do this -- you have to have drag past
some object -- otherwise you have a source for momentum and no sink, so
the momentum keeps ramping up...  Also note that when taking
quantitative measurements you sometimes have to be careful that the wake
doesn't wrap around and distort the incoming flow.

With regard to the second method, constant velocity boundary conditions
can be set up by sampling the incoming particles at the enterance and
exit from the Fermi-Dirac distribution that is the equilibrium for the
lattice gas (assuming that you are using a lattice gas that satisfies
semi-detailed balance).  Outgoing particles at either the enterance or
exit are just allowed to leave the system.  You have to ensure that the
incoming flux is equal to the outgoing flux in order to achieve a steady
state.  For details on the Fermi-Dirac equilibrium, see one of the
introductory texts on lattice gases that have been quoted on this list
in the recent past.  I think that you can implement constant pressure
boundary conditions in this way too, though I have never tried it.

Note that the two methods correspond to different mathematical
formulations of the fluid: The first method solves the Navier-Stokes
equations with an external force term and periodic boundary conditions,
while the second method solves it without the external force term but
with constant velocity boundary conditions.

Finally, note that either method introduces a probabilistic component to
the lattice gas rule.

--
Bruce M. Boghosian             | Internet: bmb@think.com
Thinking Machines Corporation  | Bitnet: bmb%think.com@mitvma.bitnet
245 First Street               | Phone: (617) 234-2140 (work)
Cambridge, MA 02142-1264  USA  | FAX: (617)-234-4444


This archive was generated by hypermail 2.1.7 : Tue Oct 14 2003 - 21:44:15 UTC