Refining problem
 

Hi, I am Davide and I've just started to study fluent on my own. I am trying to solve a fully developed turbulent flow in a valve, but I am having some problems with the convergence of the continuity equation. I choose the k-omega SST model, but the continuity residual doesn't fall under 9 e-3 (the other residual are in the order of e-4, e-5). So I tryed to refine my mesh but.. all the residuals rised! I've done 4 refinements so i pass from 200000 elements to 2000000 but the problem still remain. I know the continuity residual is too high, but the mass imbalance shows a value of 1e-5 of 50 kg/s (water) so, first, I considered the problem converged. But how can i solve the refining problem?? Do you know what kind of error may it be?

I would love to have any suggestions because I do not know how to proceed (and this is my thesis project).
Thanks

Davide

Anyone could help me? As I write this is my thesis project so it's quite important..

Hi Davide,
I would try to switch to unsteady simulation if you're running steady; maybe you have some unsteadiness in your flow down to the valve, so you can't reach a full converged solution with the steady solver.

Daniele

Thank you for your replay ghost82. I've tried your way but I've obtained the same results. Moreover, after the solution with the transient solver finished, I have the "unfulfilled" symbol beside the solution box (in the workbench) and the "attention required" symbol beside the setup box.. but I can open the post-processor

Continuity residuals are by default scaled by the max residual in the first 5 iterations. For good initializations, this scaling may result in the continuity residual bottoming out at a higher value. The continuity residual alone can't tell you everything about the state of convergence of your simulation. If your other metrics of converstion, such as net mass imbalance, are good, then don't fret over the continuity residual.

You can confirm this by starting with a different initializaion. It may converge to the same solution, but have a different value for the continuity residual.

Thank you very much Kokemoor!!!