[Indu Kant Deo]

Dear SU2 users,

I am new to CFD and SU2. I will love the help in understanding the difference between Nu_Tilde, Eddy_Viscosity and Laminar_Viscosity definition in the SU2.

From my basic CFD knowledge I know that MU is molecular dynamic viscosity, NU is molecular kinematic viscosity which equals to MU divided by the density, MU_T is turbulent dynamics viscosity and NU_T or NU_Tilde is turbulent kinematic viscosity which equals MU_T divided by density.

Eddy viscosity and laminar viscosity can be dynamic or kinematic. I would appreciate the help regarding the definition of Nu_tilde, Eddy viscosity and Laminar Viscosity in the SU2 CFD solution output for SA model.

Thank you very much.
Indu Kant Deo

[deenriqu]

Im not the expert on this, but I have a partial answer.

Eddy viscosity is the viscosity of the turbulent eddies in the flow. Laminar viscosity is a property of the fluid (intramolecular forces).

I think this reference can help https://turbmodels.larc.nasa.gov/spalart.html

[TKatt]

Hello Mr Kant ,

The given resource of the previous person is generally regarded as a very trustworthy resource which I know many SU2 models are directly implemented from (as they also list corrections/ mistakes from the original papers). https://turbmodels.larc.nasa.gov/spalart.html#sa

I think you are wrong with

NU_T or NU_Tilde is turbulent kinematic viscosity

because NU_Tilde is a somewhat artificial quantity that is solved in a transport equation (SA is a 1-eq model). The resemblance from NU_Tilde to a real kinematic viscosity wanted I guess. BUT the turbulent dynamic viscosity MU_T(urbulent) is computed using NU_Tilde with some additional factor besides the Density rho (See the first two equations under the link above). The switch between dynamic MU and kinematic NU for laminar or turbulent viscosities is the same for both and just multiplying or dividing by rho.


Eddy viscosity is then MU_T(urbulent) which is just added to the laminar viscosity in the Navier-Stokes equation. That approach is often called Boussinesq assumption. Another thing you might want to look at is just some derivations of the reynolds averaged navier stokes equations (RANS) to complete a good idea of whats going on.


Hope that helps, best, Tobi