I think that the smallest (Kolmogarov) length scale in turbulent flow can be related to Re. no. by:

N^3 = Re ^ (9/4)

(please correct me if I'm wrong!)

Is there an equivalent for a smallest temporal scale, i.e. something that would indicate the smallest time step required for DNS?

Can one consider a DNS solution to be the same as a transient laminar solution with appropriately (small) spatial and temporal step and grid sizes? If not, why not!!

(I can almost hear the moans and groans of the DNS gurus out there

going some way to answer my own question....

...I suppose the whole issue of inflow/outflow BCs has to be readressed. I doubt one can assume a steady state flow BC.

Hi,

Another issue that need to be assessed is the numerical methods! That's if your numerical methods are too dissipative or dispersive, they will impact your solution in a non-physical way (so this is not DNS).

Best of luck, Julien