[magnushaese]

Dear formers,

i have an issue with the zeroGradient BC at flow outlet combined with system rotation for a 2-D (xy) incompressible steady state problem. The whole domain is about to be rotated among the z-axis.

1. Whats working so far:
Plane-turbulent-channel flow setup, good results without system rotation using either SRFSimpleFoam or a modified simpleFoam with Coriolis force in the momentum equation. Both solvers work fine with arbitrary two-equation-turbulence-models (e.g. kEpsilon, kOmegaSST) and give the same results.

BCs velocity/relative velocity:
inlet = fixed value
outlet = zeroGradient
BCs pressure:
inlet = zeroGradient
outlet = fixed value 0

2. The issue:
Starting from a fully converged non rotation state, as soon as the system is rotated among the the z-axis in positive or negative direction the outflow shows unphysical behavior, not holding the condition of zeroGradient for the relative velocity (and absolute velocity plus turbulence quantities, whereas the pressure is fine). The result is the same for the modified simpleFoam and SRFSimpleFoam.

(Picture see attachment)
Top case, positive rotation among z-axis.
Bot case, negative though.

3. My comment:
The issue seems to be independent of the solver and turbulence model (since the used turb-models dont even represent rotational dependency anyways). Furthermore, the picture illustrates the impact of the rotational direction which shows that there might be a problem with the understanding of absolute and relative velocity at the outlet BC.
As far as i understand the "type" of a BC is independent of the basic flow quantities such as U,p,k,omega what ever, it just "zeros" the normal gradient of the quantity it is applied to, right? In my understanding this should not be the problem in the case of "Urel" though.

I am tankful for every advice!

Best regards,

Magnus

[magnushaese]

solved: new boundary condition with the pressure gradient balances the Coriolis force.

grad(P) = -2 Omega ^ U

[damon707]

Hello magnushaese,

Did you use SRFSimpleFoam for this simulation?

Best,
George