[TurBoris]

Hi,

When I set the outlet BC as mass flow rate or exit corrected mass flow rate for the unsteady compressor simulations, the flowfield cannot converge to the quasi-steady state eventhough the simulation runs for sufficient wheel rotation time. Mass flow rate at each sliding interface shows some oscillations and the average of this deviates from the set BC value. FFT of probes shows different frequency values than expected ones. However, when I change it to pressure outlet the simulation easliy converges to quasi-steady state within a few wheel rotation time. The outlet is extended enough and the mesh is gradually coarsened in order to avoid artificial reflections. The simulations are done away from stall boundary (on aerodynamically stable point).

I'd like to see your comments regarding this issue.

Thank you.

[Opaque]

Do you understand the difference between Dirichlet conditions, and non-Dirichlet conditions when applied to a differential equation?

The mass flow boundary condition is non-Dirichlet; therefore, it is adjusted as the solution is converging to satisfy the user-specified value. That means, there is an additional non-linearity that must be carefully tamed/controlled to obtain a fully converged solution.

Either it requires additional linearization, under-relaxation, more non-linear coefficient loops (in case of transient), or a smaller timestep