[Thomashoffmann]

Hi.
I'm simulation a pipe flow with a non newtonian fluid. After about 2700 iterations my residual starts to oscillate (see attached).

Do you know if this is something that I should be worried about? And why it happends?


Thanks

Thomas

[oj.bulmer]

Perhaps, there is a need for a smaller under relaxation factor.

[Thomashoffmann]

ok, I just tried to run it again with a under relaxation of 0.8 and this is what I got - see attached..
It starts oscillating again at momentum residual=3E-4 - the same value as before... I think my result is accurate at this value, I just dont like that it is doing this oscillation thing...
Maybe I just shouldn't worry?

[oj.bulmer]

Typically such small oscillations are indicative of numerical noise amd all other artifacts, which exist even after the solution has converged. To be honest, though I suggested the URFs, or any other change for that matter to make the residual more smoother, there won't be a significant change in the results beyond a point.

I would plot the parameters of interest on important locations, along with the volumetric mass imbalance, to make sure the solution is converged, rather than worrying about residuals. It may be possible that the resiruals are gradually decreasing but monitors remain flatter, indicating convergence. The measure of convergence is the result you are interested in, not residual!

Regards
OJ

[Thomashoffmann]

I'll try studying the results..
Thank you for the helpful answer!