I compute the motion of threee
 

I compute the motion of threee bubbles with interFoam. The cGamma is 1.5 for 3D case. From the following jpg figure, we can find some dirty fragments. These fragments may be introduced by the algorithm embedded in interFoam? I have not got idea to solve this problem. Can anyone give me any advie?

time=0.1s and time=.35s:

http://www.cfd-online.com/OpenFOAM_D...ages/1/965.jpg

http://www.cfd-online.com/OpenFOAM_D...ages/1/966.jpg


Thanks.

The result for the single mode
 

The result for the single mode RT instability is as follows:

http://www.cfd-online.com/OpenFOAM_D...ages/1/968.jpg

The interface is beautiful, yet.

It is not clear that those fra
 

It is not clear that those fragments are incorrect, they may be a natural consequence of a break-up process caused by the up-draft in the wake of the middle bubble but they may also be numerical as you suggest. What schemes are you using for the terms of the U an gamma equations?

The scheme list: ddtSchemes
 

The scheme list:
ddtSchemes
{
// Default scheme
default Euler;
}

// Gradient discretisation schemes
gradSchemes
{
// Default gradient scheme
default Gauss linear;
grad(U) Gauss linear;
grad(gamma) Gauss linear;
}

// Convection discretisation schemes
divSchemes
{
div(rho*phi,U) Gauss upwind;
div(phi,gamma) Gauss Gamma201 0.2;
div(phirb,gamma) Gauss Gamma201 1;
}

// Laplacian discretisation schemes
laplacianSchemes
{
// Default scheme
default Gauss linear corrected;
}

// Interpolation schemes
interpolationSchemes
{
// Default scheme
default linear;
}

// Surface normal gradient schemes
snGradSchemes
{
// Default scheme
default corrected;
}

// Calculation of flux
fluxRequired
{
// Create storage for flux for all solved variables?
default no;
pd;
pcorr;
gamma;
}

I will try to do some tests with different schemes.

Thanks.

I don't think your choice of s
 

I don't think your choice of schemes are appropriate for thise case, in particular upwind on U will cause a lot of unnecessary damping and given that your case is low Re you could probably use linear of if that gives trouble Gamma2V 1. However, this will not affect the fragmentation of the interface, to see if this is real or numerical try with

div(phi,gamma) Gauss Gamma01 1;
div(phirb,gamma) Gauss Gamma01 1;