Continuity Residual Does Not Converge
 

Hi all,

I'm solving a 3D compressible flow in a channel with perforated plates. I modeled the actual holes in the plates, rather than using the porous media approach.

I'm using a steady state solution, coupled scheme with quasi-transient approach.

After about 2500 iterations all my surface monitors have reached a steady state, the mass flow rate is perfectly balanced between the inlet and outlet, and all my residuals have converged, except for the continuity.

The continuity residual is reducing but at a very very slow pace, and at this stage it is ~6e-m.

Can I consider the solution converged? Should I continue the run (maybe beyond 10000 iterations)? In case the residual does not go below 6e-3, what should I do to obtain convergence?

Thanks!

Have you checked the mass imbalance?

OJ

Yes I did, the mass in/out is perfectly balanced, within 0.1% error.

If solution is mesh independent, all the monitors are flat, mass imbalances are small enough to your liking, and there is a monotonicity in your residuals, then I don't see a reason you should be worried about. I have observed that even within 2-eqn models, Realizable models reduces continuity to much smaller values than RNG. This must just be a numerical artifact!

OJ

Thanks a lot for the advice!