[Felix_Sp]

Hello everyone,

so I am struggling with understanding the implementation of the URANS turbulence model in REEF3D. In the source code, I only found one line of code (and some printing changes) that is changed in the files ikomega.cpp and ikepsilon.cpp (I suppose there is probably more). For the - - model this line of code is:

if(p->T10==21)
LOOP
a->eddyv(i,j,k) = MIN(a->eddyv(i,j,k), p->DXM*p->cmu*pow((kin(i,j,k)>(1.0e-20)?(kin(i,j,k)):(1.0e20)),0.5));

I suppose this is just ensuring the eddy viscosity to be in between a certain range of values to ensure stability.

From some forum posts (e.g. RANS and URANS. ) and the book of Ferzinger (https://doi.org/10.1007/978-3-662-46544-8) I basically came to the understanding, that in URANS we basically insert the time derivative and ensemble-average the flow for a certain time step (see the picture, which is taken from the book of Ferzinger, 2020; left side is supposed to be RANS and right side is supposed to be URANS).

The problem, that I have with understanding is, that we assign the time derivative for the momentum equation in N 40 in the ctrl.txt. Is this time derivative just not used, when we apply a RANS - turbulence model, e.g. T 10 1? And how is the ensemble averaging applied?

Since I have not found any papers on the implementation on the website of REEF3D, I would be happy to be pointed to a source of information or an explanation on the implementation.

Thanks for your great work and help in advance!

[valgrinda]

Of the different turbulence models, URANS is not very strictly defined. In literature, many different versions exist and URANS fills the whole region between RANS und high-resolution LES. In REEF3D, a very simple implementation is chosen, basically capping the eddy viscosity so that a larger part of the turbulence is actually resolved, see here for the reference: https://www.sciencedirect.com/scienc...42727X03001322. For RANS and URANS, the Reynolds-averaging is not performed explicitly, but happens through the (relatively) larger cells, the (relatively) larger time step and through the fact that the eddy-viscosity suppresses many of the higher-frequency turbulent fluctuations.

[Felix_Sp]

Thank you for the elaboration and the source!