Hello Essence,

I used a RNG k-e model. This choice is in accord with the reference paper, I need to simulate that model for having a comparison with my results.

I tried what you suggested and, besides that, I also lowered the order of the numerical methods (from 2nd to 1st order upwind).

Nevertheless, there was no big improvement in the residuals:

I brought the URFs gradually down to 0.1.

Is it possible that there might be a problem with the BCs?