high cfl number and discretization scheme
I have a general question concerning the accuracy of unsteady RANS calculations when using a high CFL number. Is it worth to use any kind of central differencing scheme with large CFL numbers? Visually I imagine that using cds the flow field is 'accurately' calculated at one time step, but the flow gets 'bad' resolved for the next time step!? This should actually matter especially for flows with no enforced direction!? Or do I understood this completely wrong?
|All times are GMT -4. The time now is 19:05.|