CFD Online Discussion Forums - K and Omega Profiles in Fluent

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)

- FLUENT (https://www.cfd-online.com/Forums/fluent/)

- - K and Omega Profiles in Fluent (https://www.cfd-online.com/Forums/fluent/159387-k-omega-profiles-fluent.html)

K and Omega Profiles in Fluent

Hi all,

I am having a small issue with matching experimental data in Fluent 15.0 using standard k- $\omega$ model. From the experimental data obtained I have an x-velocity profile and corresponding RMS profile for the inlet conditions.

In Fluent there are various way to specify the turbulence (via hydraulic diamter etc), however, I want to use the actual RMS profile.

The method I have used thus far is to assume isotropic turbulence and determine the turbulent kinetic energy profile from:

$k = \frac{3}{2}\overline{u_{x}^{'}}$ ,

where $\overline{u_{x}^{'}}$ is the RMS profile in the x-direction (length).

$\omega$ specified in the Fluent manual is given by

$\omega = \frac{\epsilon}{k \beta^{*}}$ ,

where $\beta^{*}$ = 0.09, and

is specified above.

I have found an approximation to $\epsilon$ is various papers, listed as

$\epsilon \approx 15 \nu \overline{\frac{\left ( u_{x}^{'} \right )^{2}}{U_{x}}}$ .

Based on a wind tunnel velocity of approx. $4 ms^{-1}$ and turbulence intensity

of approx. 1.1%, this equates to average values of $\epsilon = 2.0 \times 10^{-9}$ and $\omega = 0.0001$ .

Is this the correct approach/equations given the experimental data?

Many thanks for any input,

Hob

Quote:

Originally Posted by hob (Post 563933)

Hi hob,

Did you ever come to a conclusion about this?

Cheers,

Quote:

Originally Posted by kingjewel1 (Post 573200)

Hi hob,

Did you ever come to a conclusion about this?

Cheers,

In the end I switched to the transitional sst model, the low turbulence value meant the BL growth was very thick for the k-w model, the trans-sst is better for low turbulence/ transitioning models.

I also used hydraulic diameter, although varying the dissipation rate seemed to make little difference compared to variance between turbulence models,

Regards,

Hob