[visakhmg]

Hello everyone,


I am running a free convection (Rayleigh-Benard) simulation using buoyantPimpleFoam. Does a laminar model with sufficient grid & time resolution to resolve the Kolmogorov scales and boundary layers make it a DNS? The flow field is expected to be turbulent at the Rayleigh number I solve for. As far as I know, Direct Numerical Simulation means solving the original NS (+ energy) equations without any modeling terms for the turbulence, which is what laminar solver does.

[mkraposhin]

Hi.

Theoretically - yes, but in practice - no. There are a lot of other things, which introduce errors - pressure-velocity coupling algorithm, BC's , etc.

In other words, when you use mesh with sufficiently fine resolution, the solution could tend to wrong due to approximation errors, which are not induced by discretization of equations terms.