[praveen]

I have modified rhoCentralFoam so that one can use any flux function. Currently, Lax-Friedrichs and kinetic (kfvs) are available. You can get code here

http://sourceforge.net/projects/gfoam/develop

using svn or get latest tarball from here

http://gfoam.svn.sourceforge.net/viewvc/gfoam/?view=tar

There are three examples included. The obliqueshock and forwardstep cases run fine but there is some problem with the naca case. The case runs but the solution has strange behaviour near airfoil surface. If any of you try this and have some solution to this problem, please let me know.

[Solarberiden]

Originally Posted by praveen

I have modified rhoCentralFoam so that one can use any flux function. Currently, Lax-Friedrichs and kinetic (kfvs) are available. You can get code here

http://sourceforge.net/projects/gfoam/develop

using svn or get latest tarball from here

http://gfoam.svn.sourceforge.net/viewvc/gfoam/?view=tar

There are three examples included. The obliqueshock and forwardstep cases run fine but there is some problem with the naca case. The case runs but the solution has strange behaviour near airfoil surface. If any of you try this and have some solution to this problem, please let me know.

Hi, praveen:
I have reproduced the odd near airfoil surface you mention in the last post, I have tried to solve but without luck, do you have any new idea about how to fix the problem?

[praveen]

Based on advice of Luca Gasparini who is one of the original developers of rhoCentralFoam, I used Gamma reconstruction scheme. That improved the solution a lot. It appears that most reconstruction schemes in openfoam do not perform well on triangles/tetrahedra.

interpolationSchemes
	{
	    default          linear;
	    reconstruct(rho) Gamma  1.0;
	    reconstruct(U)   GammaV 1.0;
	    reconstruct(T)   Gamma  1.0;
	}

If this works for you too or you find a better solution, please post about it.

[Solarberiden]

alleviation.0007.jpg

alleviation.0004.jpg

I'm sorry but the situation is only improved a little bit and there should be some more efficient way to improve the result. My test result have been uploaded in the attachments, please check them out.
The 0004(at 0.02 second) is better than 0007(at 0.035 second). Along the iteration, the flaw has been appeared evidently.

Thanks Praveen.

[praveen]

I get similar results, not fully satisfactory. It might improve with finer grid. It works fine on quad/hex meshes. I dont know enough about reconstruction schemes in openfoam to say what is wrong.

[Solarberiden]

Have you ever tried out some polyhedral mesh on such test case whether the polyhedral mesh would alleviate the reconstruction issues?

[schwermetall]

Hi Foamers,
has one of you been successful in solving these pressure fluctuations? I have a similar problem and have been able to minimize these fluctuations using cellLimited grad schemes.

My supervisor thinks that this is a so called checker-board instability (I hope I'm not telling something you already know). What I can say for sure is that I have very high fluctuations using a linear scheme and pretty low ones using a cellLimited scheme.

See http://www.cfd-online.com/Forums/ope...eds-1-a-2.html

Would be nice, if you could comment on that, maybe one of you solved the problem.

regards

[ehsan]

Dear All

Could I kindly ask you whether it is possible to solve forward step test with turbulence model and rhoCentralFoam?

Regards