[marine]

hello dear foamers !

I'm simulating flow around a ship on OF1.6 with a mesh exported from StarCCM+.

I first tried a simpleFoam simulation with first order schemes and k-epsilon for modeling turbulence with BCS as follows :

k and epsilon = fixedValue at the inlet and zeroGradient at the outlet, wallfunctions on the hull.

My case didn't work as the continuity equation exploded after 30 iterations.

I think my mesh is correct because it worked on Fluent and STARCCM+ and the checkMesh only reported 4 skewFaces and few non orthogonal cells.

So I tried to modify the BCs for turbulence and I have now :

inlet : k=zeroGradient and epsilon=fixedValue
outlet : k=inletOutlet and epsilon=zeroGradient
(still wallfunctions on the hull)

I think it worked, I obtained coherent velocity and pressure fields (distribution and values) and forces of the same order as the ones I obtained on Fluent and StarCCM+.

My questions are : does someone know if this set-up is reliable or is not ? Does someone ever use this kind of set-up? because on the forum I only found cases where people fixed k and epsilon at the inlet and put zeroGradient for both at the outlet but it seems strange to me to impose values at the inlet where the flow is not "very" turbulent.

One more, what would be the best between wallfunction or fixedValue on the hull? (Re=111.6)

(I know the k-epsilon model isn't the best for this type of simulation but I want to compare with an old simulation using this model)

thank you in advance for your replies and advices !

Marine

[Mo-ITB]

hey marine,

i am working on a similar problem at the moment, ahmed body aerodynamics. have you tried to make a polymesh in ccm+? this solved most of my problems, as it has much better quality.

bye,
moritz

[marine]

yes my mesh is a polymesh, but I have a low density of cells in the volume (0.4).
My mesh doesn't seem to be the problem because with the turbulent BCs I choose for the second simulation it worked perfectly.

[marine]

To add precision : my case works with first order schemes, when I try with a second order scheme for convection (div(phi,U)=linear) it explodes...I'm know looking for a good 2nd order scheme If anyone have an advice?

thanks !

Marine