CFL number

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
Post Reply
Fabio89

CFL number

Post by Fabio89 »

During my first simulations I had HIGH CFL numbers, about 20000, my velocities into the nozzle were over 10^17 (m/s) and when I tried to show particles (displacement) and their path (trajectory) I had still particles and no trajectoy.
Than I handled my mesh and now I have 0.78<CFL<16, velocities=0.24 - 0.31 (m/s) and at last but not least I can show my particles.
What is the meaning of CFL number in CS code?

Now I'm setting "Time Step Option: UNIFORM and COSTANT", but I know I can change it in "VARIABLE in TIME" and "VARIABLE in TIME and SPACE", handling some parameters: CFL max number, Fourier max number, Min time step, Max time step.
What do they mean? and what happen if I set CFL max = 1?

Thank you for your attention

Best Regards,

Fabio
CAVT
Posts: 9
Joined: Wed Mar 14, 2012 9:10 pm

Re: CFL number

Post by CAVT »

It's the same as the Courant number, roughly it means how many cells a particle will go by in one timestep. It is reccomended that CFL<=1 to avoid divergence, but I think CS admits locally a value up to 20. So I suppose you're on a safe road.
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: CFL number

Post by Yvan Fournier »

Hello,

I'll add that the "variable in time and space" option imples a local time step, automatically determined to adapt to a target local time step (with min an max defined by the user).
This option usually accelerates convergence quite a bit, but is meaningful only for steady flows, as an unsteady flow with local time step is not interpretable relative to a given "global" instant.

And in any case, robustness of a calculation depends on many factors, notably mesh quality, but the CFL should remain below 20 (though choosing a very small time step to avoid a large CFL in only a few cells, leading to a CFL << 1 elsewhere, is not always a good option either).

Best regards,

Yvan
Fabio89

Re: CFL number

Post by Fabio89 »

Ok perfect,

Thaks for your help CAVT and Yvan, always very appreciated.
I can be sure my output gives good values to fluid velocity and particles velocity
Just an other question about Time Step: VARIABLE IN TIME and COSTANT IN SPACE, do you think setting it with:

time step 0.1
number of iteration 600 (in order to simulate one minute)

max CFL 10.0 (in my case max CFL is 16)
max Fourier 10.0 (default)
min time step 0.01
max time step 1
time step max variation 0.1

is a good idea?
I'm working with an ari flux and particles into a nozzle of 1 mm diameter.

Best Regards,

Fabio
Post Reply