Page 1 of 1

High derive values

Posted: Wed Feb 20, 2013 11:27 pm
by Pisolino
Hi all,
i'm running steady state MRF simulations (so 2 volume coupled) with saturne, some of them don't reach convergence and diverge easily.

My approach is :

Prepare mesh -> run few iterations with k-eps (i'm using air inside an oven) with scalable wall function -> check y+

if it's too big ( >300) -> remesh with finer elements at the wall where y+ is high -> check again with few iterations and if necessary rmesh again

if in the first iterations (10-15) y+ seems to be ok but then velocity increase a lot until divergence -> relaxation 0.5.

Often this doesn't help a lot... i would like to set relaxation = 0.01 but then pressure relax = 0.99 (1-0.01) so it's not helpful.

And again the derive values are very high (> e+6 ).

What can i do as strategy? How can go further in analising my mesh/simulation?

I'm quite a newbie of cfd so i would like all of your suggestions !!
Thank you

Andrea

Re: High derive values

Posted: Mon Feb 25, 2013 10:45 pm
by Jacques Fontaine
Hello,

Did you try the steady algorithm IDTVAR = 2 (time step variable in space and in time) ?

Regards,
JF

Re: High derive values

Posted: Tue Feb 26, 2013 5:57 pm
by Pisolino
hi, i'm using CS 3.0 beta.
So i can't find anymore this option, maybe i can use SIMPLEC algorithm instead of SIMPLE. is this what you mean?

Re: High derive values

Posted: Mon Mar 04, 2013 11:35 pm
by Jacques Fontaine
Hello,

Yes,

Steady + SIMPLEC

Regards,
JF