CFL criterion in transient run
Hi Glenn, For an Implicit scheme Courant number has no significance, isn't it? For an explicit scheme it should be less than 1. The transient run in CFX uses an implicit scheme. Then why is there a CFL criteria for timestep adaptation in transient run? Should I choose 1, 5, 10 or ..?
|
Re: CFL criterion in transient run
Hi,
"For an Implicit scheme Courant number has no significance, isn't it?" - it has significance but it is not a fundamental stability parameter as it is for an explicit scheme. "For an explicit scheme it should be less than 1." - correct, but this is not relevant to CFX of course. "The transient run in CFX uses an implicit scheme. Then why is there a CFL criteria for timestep adaptation in transient run?" - because a) it is a measure of the size of the timestep relative to the mesh and therefore may be useful for some simulations, and b) some physics options may be limited by courant number, for instance some free surface flows. "Should I choose 1, 5, 10 or ..?" - Problem dependant. Do a sensitivity check and find what is required for your simulation. Regards, Glenn |
Re: CFL criterion in transient run
Remember that not all physics is implemented implicit. Not frequently used forces like virtual mass force or surface tension force are implemented in an explicit manner. Very small timesteps are required to statisfy your criterion and to obtain a converged solution.
Gert-Jan |
All times are GMT -4. The time now is 04:00. |