[Ulixes]

Hi everyone,

I am simulating a 2D drop motion with VOF model in Fluent. I am quite surprised that the residuals of x- and y-velocity are two orders of magnitude higher than the one of the continuity.

I have set the absolute convergence criteria at 10^(-6) for all of the three monitored residuals (x-velocity, y-velocity and continuity). From the beginning the x- & y-velocity residuals are higher than the continuity ones. Here comes my first question: is that normal ? Should it not be the opposite ?

Than my second issue is the following. Yesterday I launched the simulation (transient) and the convergence occured within 20 iterations per time step. This morning I saw that convergence was not reached anymore even after 200 iterations. Continuity residuals was at 10^(-8) but x- and y-velocity were only at 3.10^(-6). Is that ok or is there a problem somewhere ?

Thank you for reading and sorry for the terrible english.
Here attached a screen shot of the residuals.

[LuckyTran]

Continuity residuals are scaled differently (by the worst residual in the first 5 iterations). If you initial guess is very bad, you'll have terrible residuals in the early iterations, which will make later continuity residuals look very good.

If your solution is converged, you can reset your residuals and you'll find that the continuity residual is stuck at 1 whereas velocity, etc. will be at the same values as before.

Basically, the continuity residuals are dependent on your initial guess and they can be higher or lower than the others. For transient simulations, it is okay. But for steady calculations, don't stare at residuals.

[Ulixes]

Thanks LuckyTran for responding.

Basically my initial guess is very good since I perform steady state simulations that I interpolate then to begin the transient VOF computation.

I was surprised that in these simulation it is the x- your y-velocity residuals that limit the convergence and not the continuity. Whereas in my previous simulations it was the opposite.