CFD Online Discussion Forums - RANS: difference between R and uPrime2Mean

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

- OpenFOAM Post-Processing (https://www.cfd-online.com/Forums/openfoam-post-processing/)

- - RANS: difference between R and uPrime2Mean (https://www.cfd-online.com/Forums/openfoam-post-processing/124674-rans-difference-between-r-uprime2mean.html)

RANS: difference between R and uPrime2Mean

Dear All,

I am running a RANS (simpleFoam) in a channel flow.

As far as the post-processing is concerned, I am trying to figure out the difference between R and uPrime2Mean.

Can anyone help?

Thanks,
Samuele

Hi Samuele
I think uprime2mean is the rms (root mean square) data that is also available from experiments, but what is R?
If you explain it maybe I could help you.
However I got another doubt.

Is there any way to calculate rms values (velocity, mixture fraction, etc) in RAS simulations in openFoam?

I posted this question long times ago, unfortunately with no reply.:(

I would be very glad if get any hint on this issue.:)

Regards
Bobi

Greetings to all

A hint for guys who want to have rms values in RAS simulations:

Your flow solver must be unsteady i.e. pimple or piso. Since this data processing needs to move in real times.

Regards
Bobi

Quote:

Originally Posted by samiam1000 (Post 456177)

I am trying to figure out the difference between R and uPrime2Mean.

R is the modelled Reynolds stress tensor, which you would typically obtain with Reynold Stress Models. UPrime2Mean follows from an averaging function object and denotes < U_i U_j > - < U_i><U_j>, where U is just the velocity of your simulation. So this is basically the "resolved" part of your Reynolds stress tensor.

Dear Bernhard,

is UPrime2Mean the residual-stress tensor $\tau_{ij}^R$ that results from the filtered momentum equation??? If yes, does it include the isotropic residual stress?

I've seen from the code that postChannel utility is supposed to deliver different values like: txx, txy,tyy, txy, eps, prod, vorticity, enstrophy and helicity.
postChannel.C file lists the header files: readFields.H, calculateFields.H and collapse.H
but the content of caculateFields.H is commented out! Why is that code inactive?

It computes txx, txy etc. . Are these values square root values of statistical moments?
What's the meaning of Txx? How can this field be extracted?

Could you maybe explain your answer more detailed please:confused:

Best regards,
Aylalisa

Hi aylalisa,

Did you get an answer to your question?

I have simulated the Channel395 case with a refined mesh but I don't know how get the contribution of the SGS terms using SGS TKE (as I am using the one eq model) - so that I can compare with the DNS data.

Thanks

Hi cfd!

I didn't get the answer.

According to my understanding (I am not sure if I am right) you can, once your system has reached the statistically stationary state, temporally average the Reynolds-Stress tensor that includes the fluctuations $<u^{''}u^{''}>,<v^{''}v^{''}>,<w^{''}w^{''}>,<u^{''}v^{''}>,<u^{''}w^{''}>,<v^{''}w^{''}>$ with help of function objects as described in the Forum:

Quote:

fieldAverage1
{
type fieldAverage;
functionObjectLibs ( "libfieldFunctionObjects.so" );
enabled true;
outputControl outputTime;

fields
(
U
{
mean on;
prime2Mean on;
base time;
}
);
}

prime2Mean : $u_{ij}^{''}$ (one-point, one-time correlations), since its extracted from temporally averaged velocity field U you have $<u_{ij}^{''}>$ (for graphs for DNS?)

Or use utility postChannel. Assuming your system is in the statistically stationary state and you have defined mesh periodicity in x and z directions the fluctuations become self-similar. Then postChannel collapses data in x and z direction, the remaining dependency is the y-direction. You find this data prepared for plotting in the new folder 'graphs'.

sgs kinetic energy - see answer of Sylv

I'd like to create plots like $u_{rms}^+$ and

over y and

over

.
Therefore I want to collapse time averaged fields e.g. prime2Mean in x and z. postChannel collapses different fields for each time step but not time averaged fields. How can I solve that?

The other question concerns

. How can I compute

for a cell that is not located on the wall?

Aylalisa

Hi Aylalisa,

Thanks for your response.

I think my message was not that clear - so please take a look at my two detailed questions on this thread.

Thank you very much again.

Hi again,

By the way I am already aware of the SGS TKE message from Sylv to your question. However, I am still not clear how to get the SGS stresses from SGS TKE. Maybe it is really stupid to ask or maybe not. I certainly need to get comfortable with these stresses - first in OpenFOAM (i.e how they are calculated) and then the physical picture. I have a background in RANS and all this is a little new.

Overall, at this point my aim is to reproduce figure 5.4 (on page 168) from Eugene's thesis.

Thanks