[quarkz]

Hi,

I'm simulating a 3d double cone in high speed flow using rhoCentralFoam. Initially I used a coarse grid (0.4M cells) and although the speed is slow, it runs fine with 4 procs. Using more procs doesn't seem to help much.

I then changed to a finer grid with 4M cells. It diverges after a very short while - about 1 min. I tried to reduce the initial time step to as low as 1e-12, with a low CFL of 0.1 but it still diverges.

So I'm wondering what should I change now?

For BC, most faces are more or less standard (fixed value inlet, symmetry).

For the outlet, I'm using inletOutlet and waveTransmissive for U and p. I wonder if I need to change.

For the fvSolution, I'm using:

solvers
{
"(rho|rhoU|rhoE)"
{
solver diagonal;
}

U
{
/*solver smoothSolver;
smoother GaussSeidel;
nSweeps 2;
tolerance 1e-9;
relTol 0.01;*/
solver PCG;
preconditioner FDIC;
tolerance 1e-09;
relTol 0.01;
}

while for the fvSchemes, I'm using:

fluxScheme Kurganov;

ddtSchemes
{
default Euler;
}

gradSchemes
{
default Gauss linear;
}

divSchemes
{
default none;
div(tauMC) Gauss linear;
}

laplacianSchemes
{
default Gauss linear corrected;
}

interpolationSchemes
{
default linear;//limitedLinear 1.0; //
reconstruct(rho) upwind;//vanLeer; MUSCL; Gamma
reconstruct(U) upwind;//vanLeerV; MUSCL; GammaV
reconstruct(T) upwind;//vanLeer; MUSCL; Gamma
}

snGradSchemes
{
default corrected;
}

Being a new user to OpenFOAM, I hope someone can give some advice.

Thanks!