Learning Forum

[JMA337]

Hi,

I am calculating the anistropic reynolds stress tensor. I believe I have a missconception regarding the definition of the turbulent kinetic energy in Fluent for pseudo-steady and unsteady RANS linear eddy visocsity models. 

Let's take the k-w model. The field created by Fluent called "turbulent kinetic energy" correpsonds to the modelled turbulent kientic energy which is obtained through the Boussinesq hypothesis. Right? 

But the turbulent kinetic energy can also be calculated as k = 0.5 ( u_rmse^2 + v_rmse^2 + w_rmse^2), assuming the sampling statistics have been activated. Does this turbulent kinetic energy then represent the fraction of the turbulent spectrum and unsteady phenomena which is actually resolved?

Therefore, is really the turbulent kinetic energy the compund of both? ie: k = k_modeled + k_resolved? So, k = k_"tke field in 2-eqn LEVM RANS" + k_"obtained from the velocity rmse"

Thank you

[SRP]

Hi,

The "turbulent kinetic energy" field in Fluent for RANS models (including k-ω) is the modeled TKE, representing all turbulence effects on the mean flow as per the Boussinesq hypothesis.

The RMS-based TKE from resolved fluctuations is not physically meaningful in RANS and should not be combined with the modeled TKE.

For hybrid or scale-resolving simulations, both modeled and resolved TKE are relevant and their sum represents the total TKE, but this does not apply to standard RANS.

[JMA337]

Hi,

Thank you for your prompt response. 

Ok. I undestand for steady RANS. But with unsteady RANS, depending on the spatial and temporal resolutions, and depending on separation of scales, there are some scales of turbuelnce (the largest) which are being resolved. Therefore in unsteady RANS, shouldn't we also combine the modeled TKE and the RMS-based TKE?

Thank you, 

Best wishes