[christoph_heinrich]

Hi everybody,

at the moment I am trying to implement a stationary Stokes solver in Matlab for a simple geometry. I am using the FVM with midpoint rule and CDS. In the pressure field I get oscillations. How can I circumvent this problem on a collocated grid? Any idea?

Thanks in advance!

Kind regards
Christoph

[henrik]

Dear Christoph,

you need to do Rhie & Chow interpolation to avoid pressure-velocity decoupling on collocated grids.

http://www.tfd.chalmers.se/~hani/kurser/OS_CFD_2007/rhiechow.pdf

Reference to original paper:
http://powerlab.fsb.hr/ped/kturbo/Op...jeJasakPhD.pdf

Henrik

[victorconan]

Originally Posted by henrik

Dear Christoph,

you need to do Rhie & Chow interpolation to avoid pressure-velocity decoupling on collocated grids.

http://www.tfd.chalmers.se/~hani/kurser/OS_CFD_2007/rhiechow.pdf

Reference to original paper:
http://powerlab.fsb.hr/ped/kturbo/Op...jeJasakPhD.pdf

Henrik

Hi, Henrik
I suppose the Rhie-Chow interpolation mentioned by Fabian Peng Karrholm does not exist in Openfoam at all. According to Rhie-Chow interpolation, we should have interpolation for U and p on surface respectively. I have felt puzzled for a long time about how to make it in openfoam. I tried this:
volScalarField Vmesh=mesh.V()
phi=fvc::interpolate(U) & mesh.Sf()
- pEqn.flux()
+ fvc::snGrad(p) * fvc::interpolate(Vmesh);
I suppose these three above items are what Rhie-Chow interpolation is, but the compiling of it failed. I don't know why and how to correct it. Could you do me a favor?