Converged?
 

This might be a moot point, but I ran a simulation overnight with the double precision solver, and I'm seeing something strange (at least to me) in the residual plots.

For reference, I'm modeling a turbulent multiphase shear flow with total energy heat transfer between two fluid layers. And I set the convergence criteria to E-5.

All of the momentum, heat transfer, and turbulence residuals have dropped below E-7 RMS. The momentum imbalance(%) is below 4e-005, and the mass fraction imbalance(%) is below 0.3. The strange thing though, is that one of the RMS Mass fraction residuals is oscillating between E-4 and E-5, but never drops below E-5 (Hence why the solver didn't reach convergence and stop). But, the mass fraction residual of the other phase is below E-7.

What I'm wondering is whether I should just accept this as a good result, or if I should revisit my mesh, and try to refine it.

It just seems strange that all the other residuals would converge except that one.

Re: Converged?
 

Hi,

Look at the important parameters of your flow and judge from that whether you are accurate enough. This could be outlet temperature, location of a separation, free surface level - whatever is important for your application. If it is not changing then you should be OK with your results as they are.

Glenn Horrocks

Re: Converged?
 

Apart from the fact that multiphase seems inherently difficult to converge, it might be an idea to plot the residual so that at least you know what the potential problem area is.