Buoyancy Treatment in OpenFoam

AMK53 · October 29, 2017, 10:35

From all the standard solvers that I found on OpenFoam, it seems that the buoyancy term is always treated by modifying the pressure:

$-\nabla P + \rho g = -\nabla P_{rgh} - (g \cdot x) \nabla \rho$

where

$P_{rgh} = P - \rho g \cdot x$

In OpenFoam format, this is usually written as

Code:

solve        (
            UEqn
         ==
            fvc::reconstruct
            (
                (
                  - ghf*fvc::snGrad(rho)
                  - fvc::snGrad(p_rgh)
                )*mesh.magSf()
            )
        );

Is there a particular numerical benefit for treating the buoyancy term in this fashion instead of just treating it as a source term $\rho g$ ? It seems to me this way is more complicated in the sense that you have to track two pressures (P and P_rgh), and for certain problems, creating an initial condition for P_rgh is not straightforward.

eric · October 30, 2017, 04:31

Yes, using the modified pressure often makes the problem easier to handle numerically, and it's easier to specify boundary conditions.

AMK53 · October 30, 2017, 22:51

Quote:

Originally Posted by eric

Yes, using the modified pressure often makes the problem easier to handle numerically, and it's easier to specify boundary conditions.

Do you mind elaborating on how it makes it easier to handle numerically?

eric · October 31, 2017, 01:23

See this post on the issue tracker from Henry Weller, where he briefly explains it:
https://bugs.openfoam.org/view.php?id=1170

October 29, 2017, 10:35	Buoyancy Treatment in OpenFoam	#1
AMK53 New Member Join Date: Oct 2017 Posts: 12 Rep Power: 8	From all the standard solvers that I found on OpenFoam, it seems that the buoyancy term is always treated by modifying the pressure: $-\nabla P + \rho g = -\nabla P_{rgh} - (g \cdot x) \nabla \rho$ where $P_{rgh} = P - \rho g \cdot x$ In OpenFoam format, this is usually written as Code: solve ( UEqn == fvc::reconstruct ( ( - ghffvc::snGrad(rho) - fvc::snGrad(p_rgh) )mesh.magSf() ) ); Is there a particular numerical benefit for treating the buoyancy term in this fashion instead of just treating it as a source term $\rho g$ ? It seems to me this way is more complicated in the sense that you have to track two pressures (P and P_rgh), and for certain problems, creating an initial condition for P_rgh is not straightforward.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
OpenFOAM Training, London, Chicago, Munich, Houston 2016-2017	cfd.direct	OpenFOAM Announcements from Other Sources	0	September 14, 2016 03:19
OpenFOAM wall function treatment	kimotbwb	OpenFOAM Programming & Development	7	March 6, 2015 12:52
Suggestion for a new sub-forum at OpenFOAM's Forum	wyldckat	Site Help, Feedback & Discussions	20	October 28, 2014 09:04
OpenFOAM 2.x simpleFoam and kOmegaSST convergence and wall treatment	gillimaniac	OpenFOAM Running, Solving & CFD	6	August 1, 2013 01:36
Adventure of fisrst openfoam installation on Ubuntu 710	jussi	OpenFOAM Installation	0	April 24, 2008 14:25

October 30, 2017, 04:31		#2
eric Member Knut Erik T. Giljarhus Join Date: Mar 2009 Location: Norway Posts: 35 Rep Power: 22	Yes, using the modified pressure often makes the problem easier to handle numerically, and it's easier to specify boundary conditions.

October 31, 2017, 01:23		#4
eric Member Knut Erik T. Giljarhus Join Date: Mar 2009 Location: Norway Posts: 35 Rep Power: 22	See this post on the issue tracker from Henry Weller, where he briefly explains it: https://bugs.openfoam.org/view.php?id=1170