[MissCFD]

Hello everybody,

Is it fair to think that eddy viscosity allow to represent convective momentum transfer engendered by a turbulent flow by diffusive momentum transfer like a laminar flow ?

Thanks

[ghorrocks]

That is exactly how it works. The assumption is that turbulence causes additional viscosity which simply adds to the molecular viscosity. Therefore turbulence simply increases the viscosity (which is the diffusive momentum transfer).

[MissCFD]

Thank you

An other litte question about it, how can i know if my turbulent viscosity ratio is representative or not of my inlet flow ? Or how can i know if i must change my turbulence conditions (Intensity or especially turbulent length scale) on inlet ?

For example, i found a turbulent viscosity ratio of 200 in the inlet of my flow and I calculated a Reynolds number of 10^6.

Thank

[ghorrocks]

How do you know whether your turbulent viscosity is accurate? By whether your simulation is accurate or not.

When must change my turbulence conditions? .... the corollary - if your simulation is inaccurate then you need to thing harder about your simulation, and that include the inlet boundary conditions.

[MissCFD]

I'm not sure to understand very well this notion finally because in my case, a circular pipeline where I have imposed a turbulent velocity profil at inlet, I observe after an elbow of 90° and a convergent, an eddy viscosity ratio of zero near wall and many near center like 3000. Does it mean necessarily that the flow is more turbulent at the center of pipeline ? Can i say that turbulent stress-strain doesn't exist near wall and that boundary layer is only laminar ?

But, why eddy viscosity ratio is zero near wall, there is inevitably velocity gradients, no ?

Thank you for your help

[ghorrocks]

The turbulence at a wall is zero because the no slip boundary condition means there is no flow. In a pipe as you move away from the wall the turbulence (in general) increases to a maximum in the middle of the flow. So this is consistent with what you saw.

Have a read of a turbulence modelling textbook (like Turbulence modelling for CFD by Wilcox) to get some background on near-wall modelling. There are several layers in the standard turbulence profile and they all have interesting characteristics.

To answer your direct questions:

Does it mean necessarily that the flow is more turbulent at the center of pipeline ?

This is usually the case, yes.

Can i say that turbulent stress-strain doesn't exist near wall and that boundary layer is only laminar ?

This is a massive over simplification of the true situation. Refer to a turbulence modelling textbook for more details.

But, why eddy viscosity ratio is zero near wall, there is inevitably velocity gradients, no ?

Eddy viscosity is zero at the wall as the flow is zero at the wall. And yes, there is a velocity gradient at the wall (that is where the wall shear comes from). This is the integration to the wall approach.