CFD Online URL
[Sponsors]
Home > Forums > Main CFD Forum

Code Saturne, Thoughts after finishing the First Tutorial

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   April 22, 2009, 00:01
Red face Code Saturne, Thoughts after finishing the First Tutorial
  #1
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
For some time, I have heard good opinions about Code Saturne, an open source, and free CFD solver (you can have a very first contact here

2007.rmll.info/IMG/pdf/RMLL07Sc_CodeSaturne_EDF.pdf

I tried to install it on my open suse partition, but the result was not positive, it seems that open suse has memory management problems with hdf5 & med libraries (needed to import and export geometry and post processing data with Salome ).
Recently I added a Debian partition to my computer and decided to have another go with the installation. Every thing went OK except med (Still have problems installing the right med, but nothing with memory management), any way, except med, the rest is running fine. This week end, I printed the first tutorial and rolled up my sleeves.

First, I would like to ask the readers that have negative opinions about Open source programmes to set aside their negative waves till they try this one by themselves.
The programme has a nice graphical user interface with a file tree so well structured that does not allow you to forget any step. It reminds me of The old Fluent text menues, that led you through the solution procedure a step by step so here they rightly deserve an A+ grade.
A very nice GUI indeed.
Upwind and central discretizations are available
A bunch of Turbulence models spanning from the old Prandtle mixing length model to the latest k-omega. Another A+ here.
The Energy equation (Both in temperature and enthalpy) are available, so another A+
A bunch of physical models from simple pipe cases to nuclear reactors and combustion models, here they deserve a mark of excellence above A+.
In short, I invite my fellow readers of this forum to try it and sure they will change their minds about open source codes, period period period.
Very few things annoys me about this programme, one such a thing is the way they treat gravity forces, my first impression was this programme is good for outer space applications but I am wrong, I just need to go through their theory manual.
Another negative point is how the different libraries needed to compile the programme are scattered along different sites, yes, they indicate links to these sites, but what about the correct version number of each library, my problem now is to find the correct med library that can be compiled on my Debian partition and my 64 bits Open Suse partition.
We can easily forget about these last two remarks, and Enjoy a well written programme with an excellent GUI.
By the way, the programme can export the results in Ensight format, so our good friend ParaView can read them.
Enjoy your CFD computations.

If after trying this Open Source programme you find I am wrong, Please put your opinion here.
I hope one of these days we convice Johnas and Peter to have a dedicated forum for such an excellent Free Open source programme.

Good Luck to all, and thank you for patiently reading this post.
Ahmed is offline   Reply With Quote

Old   April 23, 2009, 20:16
Default
  #2
New Member
 
david.notay
Join Date: Mar 2009
Posts: 20
Rep Power: 7
gonski is on a distinguished road
I want to use code_saturne, but some its documents are written in French which I cannot read. The use of an Open source code should give me an hope that I can learn/ know most knowledge underlying the code. This helps me make further development. Otherwise, I would like to use commercial packages which are generally more friendly in use although they are black boxes. the unreadable documents absolutely put me in the position that I feel I am using a commerical package, not an open source.

Cheers,
Gonski
gonski is offline   Reply With Quote

Old   April 24, 2009, 03:06
Default
  #3
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
Quote:
Originally Posted by gonski View Post
I want to use code_saturne, but some its documents are written in French which I cannot read. The use of an Open source code should give me an hope that I can learn/ know most knowledge underlying the code. This helps me make further development. Otherwise, I would like to use commercial packages which are generally more friendly in use although they are black boxes. the unreadable documents absolutely put me in the position that I feel I am using a commerical package, not an open source.

Cheers,
Gonski
I did not say I have years of experience using the programme, but if you google for Code_saturne, you will find so many papers and articles by so many people in the industry and academics that have used the programme.
So what is the value of what is said by someone who has never used the programme, (by your own word).
Just for the record, if you search the doc folders in your installation, you will find the Theory manual, the user guide manual and the Tutorials manual all in English (this is at least what I have here on my computer) what else do you need?
Ahmed is offline   Reply With Quote

Old   April 24, 2009, 03:21
Default
  #4
New Member
 
david.notay
Join Date: Mar 2009
Posts: 20
Rep Power: 7
gonski is on a distinguished road
are you sure? I just check the code, finding that its theory manual, notes in the code are in french. This is the same to those couple of months ago. I did hear that the developing group are trying to do some translations, but the progress goes very slowly. They may have other important things to do.
Cheers,
Gonski
gonski is offline   Reply With Quote

Old   April 24, 2009, 11:24
Default
  #5
New Member
 
Join Date: Mar 2009
Posts: 18
Rep Power: 7
highnelly is on a distinguished road
Thanks for that Ahmed,

This looks really interesting I will give it a shot. Hoave you any ideas on an OS mesher? Have EDF developed something in that department other than Salome, which as far as I know cannot generate prism layers. Engrid looks good but I think it needs a good surface mesh to begin with. How do you overcome the meshing issue?
highnelly is offline   Reply With Quote

Old   April 25, 2009, 00:52
Default
  #6
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
Quote:
Originally Posted by gonski View Post
are you sure? I just check the code, finding that its theory manual, notes in the code are in french. This is the same to those couple of months ago. I did hear that the developing group are trying to do some translations, but the progress goes very slowly. They may have other important things to do.
Cheers,
Gonski
You have a problem with your installation. Check the variable language in the setup file (saturne/Install) if it is set to FR then your installation will be in French (menus, documents) and you need to change it to EN so you wll have English installation.
Ahmed is offline   Reply With Quote

Old   April 25, 2009, 01:06
Default
  #7
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
Quote:
Originally Posted by highnelly View Post
Thanks for that Ahmed,

This looks really interesting I will give it a shot. Hoave you any ideas on an OS mesher? Have EDF developed something in that department other than Salome, which as far as I know cannot generate prism layers. Engrid looks good but I think it needs a good surface mesh to begin with. How do you overcome the meshing issue?
I have not looked into this question yet, I was planning on using Salome and that is why I am looking for a hdf5/med files that can be installed on my Installation. Engrid may be a good solution, but I do not know, installing on Linux is not a simple straight forward procedure as you know. Let us wait and see
I have installed the thermal SYRTHES code so heat conduction through solid walls can be accounted for during a CFD simulation and of course radiation boundary conditions can be included
Good Luck
Ahmed is offline   Reply With Quote

Old   April 25, 2009, 06:26
Default
  #8
New Member
 
Join Date: Mar 2009
Posts: 18
Rep Power: 7
highnelly is on a distinguished road
Dear Ahmed,

I am trying to install code_saturne on Ubuntu 8.10 but I am running into the following error during my compilation:
Code:
make[1]: Leaving directory `/home/tom/code_saturne/Noyau/ncs-1.4.0/build/Linux_x86_64/BASE'
make: *** [privcompil] Error 2
   Directory: /home/tom/code_saturne/Noyau/ncs-1.4.0/build/Linux_x86_64/BASE

   Error : make lib_install MOD=BASE failed

Thank you for choosing Code_Saturne
On the CAELinux site the fix was to " to change the h_addr structure item in h_addr_list[0] at line 313 in order to compile Code_Saturne" s
I am not yet experienced enough with linux to do this but I was wondering if you ran into the same difficulty. I am presuming your Debian sys is Ubuntu?

Anyway, thanks alot if you can help; I am eager to try this out
highnelly is offline   Reply With Quote

Old   April 25, 2009, 07:50
Default
  #9
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
Quote:
Originally Posted by highnelly View Post
Dear Ahmed,

I am trying to install code_saturne on Ubuntu 8.10 but I am running into the following error during my compilation:
Code:
make[1]: Leaving directory `/home/tom/code_saturne/Noyau/ncs-1.4.0/build/Linux_x86_64/BASE'
make: *** [privcompil] Error 2
   Directory: /home/tom/code_saturne/Noyau/ncs-1.4.0/build/Linux_x86_64/BASE

   Error : make lib_install MOD=BASE failed

Thank you for choosing Code_Saturne
On the CAELinux site the fix was to " to change the h_addr structure item in h_addr_list[0] at line 313 in order to compile Code_Saturne" s
I am not yet experienced enough with linux to do this but I was wondering if you ran into the same difficulty. I am presuming your Debian sys is Ubuntu?

Anyway, thanks alot if you can help; I am eager to try this out

Unless someone from the code_saturne support team is reading your message and gives you another advice, here is what I think, you ought to do:
rename the installation folder to just saturne not code_saturne.

/home/tom/saturne

The following two lines are from the README.txt

When installing Code_Saturne, a specific directory must be prepared, referred to here as INSTALLPATH. It is usually /home/saturne.

Good luck
Ahmed is offline   Reply With Quote

Old   April 25, 2009, 20:22
Default
  #10
New Member
 
david.notay
Join Date: Mar 2009
Posts: 20
Rep Power: 7
gonski is on a distinguished road
Quote:
Originally Posted by Ahmed View Post
You have a problem with your installation. Check the variable language in the setup file (saturne/Install) if it is set to FR then your installation will be in French (menus, documents) and you need to change it to EN so you wll have English installation.
I want to study the code, then develop something regarding granular flow based on the code. so theory manual and notes in the code is the most important thing to me. Otherwise, it is difficult to continue my work.

Cheers,
Gonksi
gonski is offline   Reply With Quote

Old   April 26, 2009, 10:05
Default
  #11
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
Quote:
Originally Posted by gonski View Post
I want to study the code, then develop something regarding granular flow based on the code. so theory manual and notes in the code is the most important thing to me. Otherwise, it is difficult to continue my work.

Cheers,
Gonksi
Gonk

I do not work for EDF or the Code_Saturne developers, No relation at all.
So, I do not care whether you use the programme or not, I do not care if you want to study the programme, develop the programme or whatever,
in simple language, I do not care about you or your work.
if you have read the pdf file (in my first post) and you understand the English language then you know where to send your silly questions.

I am a member of this forum, just want to exchange opinions, ideas and experiences with other members of the forum but not with a child like you
Ahmed is offline   Reply With Quote

Old   April 26, 2009, 20:15
Default
  #12
New Member
 
david.notay
Join Date: Mar 2009
Posts: 20
Rep Power: 7
gonski is on a distinguished road
Quote:
Originally Posted by Ahmed View Post
Gonk

I do not work for EDF or the Code_Saturne developers, No relation at all.
So, I do not care whether you use the programme or not, I do not care if you want to study the programme, develop the programme or whatever,
in simple language, I do not care about you or your work.
if you have read the pdf file (in my first post) and you understand the English language then you know where to send your silly questions.

I am a member of this forum, just want to exchange opinions, ideas and experiences with other members of the forum but not with a child like you
Hi,Ahmed
Please note that I never ask any help from you or suggest you anythings. I admire the guys who are working on open sources, but not you. Therefore, I try to suggest here EDF developers (some may visit this forum) to do something regrading translations so as to expand the user group who in turn help to develop the code. I do not understand why you are so angry at my points.
gonski is offline   Reply With Quote

Old   May 3, 2009, 15:38
Default
  #13
New Member
 
David Monfort
Join Date: May 2009
Posts: 9
Rep Power: 7
dmonfort is on a distinguished road
Hello Ahmed,

As a developer of Code_Saturne, many thanks for having tested the code and provided such feedback. All these A+ are amazing!

I'll try to give you more information on the code, but don't hesitate to ask for more.

Quote:
Originally Posted by Ahmed View Post
one such a thing is the way they treat gravity forces, my first impression was this programme is good for outer space applications but I am wrong, I just need to go through their theory manual.
Can you explain what you intended by this? If needed I can give you inner details on how gravity (or body) forces are treated.

Quote:
Originally Posted by Ahmed View Post
Another negative point is how the different libraries needed to compile the programme are scattered along different sites, yes, they indicate links to these sites, but what about the correct version number of each library, my problem now is to find the correct med library that can be compiled on my Debian partition and my 64 bits Open Suse partition.

Indeed you got a point. Nonetheless, all the library you mention are optional. You only need the ones provided on Code_Saturne website to run the code. If you need specific features, then you have to install them and the latest version of each of them should work:
  • libxml2 for reading XML files created by the Graphical User Interface
  • Tcl/Tk/Tix or PyQT 4 for the GUI (depending on Code_Saturne version)
  • MED/HDF5 for MED files support (mesh format exported by Salome)
  • CGNS for CGNS files support (standard format in CFD)
  • MPI distribution (say OpenMPI for example) and Metis (higly recommended for better performance) for parallel runs.
  • BLAS library for better performance (about 20% faster)
Essentially, only the two first items are important for a user-friendly experience but these libraries should be installed with your package manager. The next two items are for different mesh formats support and must be installed "by hand" because they are rarely available as packages (though MED is available on Debian as of version 2.3.1). And the two last ones are very specific, and again should be installed with your packager manager (except for Metis).
A quick note on installation, you need to install library and dev packages for some libraries in order to compile Code_Saturne (e.g. libxml2 and libxml2-dev)


Quote:
Originally Posted by Ahmed View Post
I hope one of these days we convice Johnas and Peter to have a dedicated forum for such an excellent Free Open source programme.
Concerning the forum, we are about to set up one on Code_Saturne website. It took us longer than what we had foreseen but it should be available in a couple of weeks now (just a few fixes needed left). So stay tuned

Again thanks for the feedback!

Regards,
David
dmonfort is offline   Reply With Quote

Old   May 3, 2009, 15:54
Default
  #14
New Member
 
David Monfort
Join Date: May 2009
Posts: 9
Rep Power: 7
dmonfort is on a distinguished road
Again speeking as one of Code_Saturne developers

Quote:
Originally Posted by gonski View Post
I want to use code_saturne, but some its documents are written in French which I cannot read.
That's true and we hope to have the theory manual fully translated by the end of this year.

Quote:
Originally Posted by gonski View Post
The unreadable documents absolutely put me in the position that I feel I am using a commerical package, not an open source.
I clearly understand your concerns. But you have to understand that the code was released open source only two years ago, whereas it has been developped for about 10 years (without counting the knowledge on older codes...) and of course everything was in French at the moment. So we need to translate everything and it is quite a long process, and unfortunately not always the highest priority one

We should have a release condidate of the next version (we validate a version every two years) before the end of this year, expectedly with a proper documentation in English, so stay tuned
Meanwhile, if you have any question, you can contact the support at saturne-support@edf.fr.

Regards,
David
dmonfort is offline   Reply With Quote

Old   May 3, 2009, 16:02
Default
  #15
New Member
 
David Monfort
Join Date: May 2009
Posts: 9
Rep Power: 7
dmonfort is on a distinguished road
Dear Highnelly,

The fix below is not needed anymore (at least for 1.3.3 and 1.4.0 version), so your issue does not come from here.
Quote:
Originally Posted by highnelly View Post
On the CAELinux site the fix was to " to change the h_addr structure item in h_addr_list[0] at line 313 in order to compile Code_Saturne"
Could you post your setup file as well as the installation output if there is more information on the error? i'll try to help you (or you can contact the support).

Regards,
David
dmonfort is offline   Reply With Quote

Old   May 5, 2009, 14:03
Wink Code Saturne - Reply to David
  #16
Senior Member
 
Ahmed
Join Date: Mar 2009
Location: NY
Posts: 241
Rep Power: 8
Ahmed is on a distinguished road
David Monfort wrote
"Can you explain what you intended by this? If needed I can give you inner details on how gravity (or body) forces are treated."

When doing analysis for devices that will be operating in the outer space, the gravity vector is set equal zero, the same as the default values in the body forces panel of Code-saturne and on the contrary of what is usually done in commercial programmes I used before. So when I got to this panel it just refreshed my memory and brought back some old experiences
As I said, I need to dive into the theory manual.

You see, when you read a comment made by someone with a different experience than yours, just read them with open mind, these comments normally reflects their own experiences and preferences, no offense intended.
Ahmed is offline   Reply With Quote

Old   May 5, 2009, 14:45
Default
  #17
New Member
 
David Monfort
Join Date: May 2009
Posts: 9
Rep Power: 7
dmonfort is on a distinguished road
Quote:
Originally Posted by Ahmed View Post
the gravity vector is set equal zero, the same as the default values in the body forces panel of Code_saturne
You are right, the default options are with zero gravity. The reason is that the code was originally designed for confined flows (most of our in-house studies are still confined flows), and if the density is assumed to be constant setting the gravity to a non-zero value is useless. And we do not provide the user with a standard law for the density variation, but if you specify classical Boussinesq approximation for the density law, you will get the classical source term depending on the gravity in the Navier-Stokes equations.

Quote:
As I said, I need to dive into the theory manual.
I hope you can read French because more than half of it isn't translated yet but I hope it will change in the future!

Quote:
no offense intended.
No poblem
dmonfort is offline   Reply With Quote

Old   March 26, 2010, 01:37
Default Code_saturne
  #18
Member
 
MSarkar
Join Date: Dec 2009
Posts: 99
Rep Power: 7
msarkar is on a distinguished road
Hi
I want to use code_saturne and I have it installed on my machine but I am little confused how to start. I was trying to follow user manual and tutorial but I could not do it. I set up the environment variable properly. Then how do I approach to do any tutorial? By following tutorial case1, I was trying to create study directory T_JUNCTION CAS1 but I could not do it. It says command not found. Could anyone please guide me how to start a tutorial? What will be working directory? Do I need create any new directory in my Home folder and copy tutorials there and start it? I really do not have any idea, please help me out. My machine has Ubuntu linux system and code_saturne was installed with CAE linux by root. I am an user only.

Regards
Msarkar
msarkar is offline   Reply With Quote

Old   March 28, 2010, 13:26
Default
  #19
New Member
 
David Monfort
Join Date: May 2009
Posts: 9
Rep Power: 7
dmonfort is on a distinguished road
Hi,

IIRC, CAELinux provides Code_Saturne 1.4 in its distro, all my explanation will be based on this hyptothesis.

First thing to do is to correctly setup your environment (if not already done by CAELinux):
Code:
. /path/to/saturne/Noyau/ncs-1.4.0/bin/cs_profile
Once this is done, then you can launch the tutorial documentation by typing
Code:
cs_info tutorial
and follow the different steps.

First step will be to create a new study:
Code:
cs_create -study TUTORIAL
and to launch the graphical interface
Code:
cd TUTORIAL/CASE1/DATA && ./SaturneGUI
Latest point, I would advise you to go for the newest version "2.0-rc1" that you can find here.

Hope this will help,
David
dmonfort is offline   Reply With Quote

Old   March 29, 2010, 04:01
Default
  #20
Member
 
MSarkar
Join Date: Dec 2009
Posts: 99
Rep Power: 7
msarkar is on a distinguished road
Hello David,

Your hypothesis is correct. CAE linux has 1.4.b version of Code-saturne. Thanks a lot for your help! With your help I was able to run the first tutorial, Simple_Junction case but I could not figure out how to visualize the results yet. If you can guide me for this, I would appreciate your help.

Code_Saturne 2.0.rcl:- actually I downloaded this version from EDF website but I could not install it. install_saturne python script is not working for my sytem (ubuntu linux). I was trying to install by following the option II (Manual Install) explained on readme file but it was giving some error. I will try it again and let you know the error messages..

Thanks
M. Sarkar
msarkar is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
own Code vs. commercial code Bernhard Mueck Main CFD Forum 10 February 16, 2000 11:07
State of the art in CFD technology Juan Carlos GARCIA SALAS Main CFD Forum 39 November 1, 1999 15:34
Use of an hydrodynamic model integrated in the vertical by the PHOENICS code Hafsia Zouhaier Main CFD Forum 2 March 15, 1999 09:23
public CFD Code development Heinz Wilkening Main CFD Forum 38 March 5, 1999 12:44
What kind of Cmmercial CFD code you feel well? Lans Main CFD Forum 13 October 27, 1998 11:20


All times are GMT -4. The time now is 03:10.