code_saturne User's Forum

Hello,
I want some advice on how to get fluctuating (u', v', w') and RMS quantities in Saturne.
I need those to calculate different turbulence quantities such as turbulent kinetic energy and Reynolds stress.
I primarily use the Saturne GUI and Saturne version 1.3.3.
Any suggestion / help will be really helpful.
Thanks and Best Regards,
Debashis

Hello,
Are you running a DNS or using a turbulence model.
If you are running a DNS, you will need to use the usproj.F user subroutine (possibly also using time value averages, as described in the user documentation) to compute the quantitie you need.
If you are using a turbulence model, the is no way you may compute those quantities, as the computed velocity is the average velocity, and the turbulent quantities (k, epsilon, Rij, ... depending on the model) are already computed by the code.
Best regards,
  Yvan Fournier

I understand.
But can usproj.F calculate u', v' and w' for any case ?
RMS = root mean square quantities.
Can any subroutine calculate that ?
Suppose RMS of u' or v' ?
Is there any subroutine that calculates that ?
Regards,
Debashis

Sorry, I forgot to mention that I am using the k-epsilon turbulence model.
 
Previously Yvan Fournier wrote:

Hello,
Are you running a DNS or using a turbulence model.
If you are running a DNS, you will need to use the usproj.F user subroutine (possibly also using time value averages, as described in the user documentation) to compute the quantitie you need.
If you are using a turbulence model, the is no way you may compute those quantities, as the computed velocity is the average velocity, and the turbulent quantities (k, epsilon, Rij, ... depending on the model) are already computed by the code.
Best regards,
  Yvan Fournier

 

Hello,
If you are using the k-epsilon model, you already have k (turbulent kinetic energy) en epsilon (turbulent dissipation) available as main variables, which you can post-process (note that specialists recommend the linear production variant of the k-eps over the legacy version).
You will not be able to compute those values using usproj.F and just the velocity, as the k-epsilon adds a turbulent viscosity (much higher than the laminar viscosity), which amounts to computing a "mean" velocity with a lower equivalent Reynolds number. This is how the k-epsilon model works, and is not specific to Code_Saturne.
If you want more advanced post-processing than those available in the code (k, epsilon, y+ at the boundary, ...) you will need to have a good (i.e. developper-level) understanding of the turbulence models and how they are implemented in the code first.
Best regards,
  Yvan

Thank you. But suppose I run LES or DES type turbulence calcultaion, there will be a modeled turbulent kinetic energy (k) and also a resolved turbulent kinetic energy [1/2*(u' 2 +v' 2 +w' 2 )]. Also there will be Reynolds stress component such as u'v' etc. If I run LES or DES then how can I calculate these quantities ? Kindly let me know. I am fine with k and epsilon and the time averaged modeled quantities. It is this resolved turbulent quantities and the Reynolds stress quantities, that I am interested in.
Also, I have another question.
I have kind of downloaded everything related to Saturne version 2.0 and about to install.
 
However, in the directory ncs-2.0.1/doc/theory I can see a lot of .tex files (Latex files).
 
I can also see two files Makefile.in and Makefile.am.
 
I can pretty much guess that these two files are related to the Latex documents.
 
Could you please advice me how to use these two Makefiles and convert the tex files to pdf files.
 
I would really appreciate your help in this regard.
 
Best Regards,
 
Debashis
 

Hello,

For the documentation, in a build directory, run

make pdf
make install-pdf

I believe the installer script does this automatically, but it will work only if you have the necessary packages installed (pdflatex and fig2dev).

You will have Reynolds-stress components automatically using Rij-epsilon, but if you want them using LES, you'll need to do your own statistics in usproj.F.

Best regards,

  Yvan