Re: [Throughflow-main] Your streamline curvature throughflow code at sourceforge

[Oliver B. <oli...@we...>]

Dear Sujith Joseph,

what do you mean, with translating the code from German to English, as 
the code is written in C++? Do you mean the comments, yes this can can 
be done easily, or do you mean the naming of functions and methods? The 
latter is a little bit tricky, as you have to change the naming at 
different locations.

The code already run under windows. For that you'll need a dll of a 
library called nurbs++, as the skv depends on that library. This library 
implements some really useful features in handling NURBS curves. 
Unfortunately this library is not maintained since one decade. We looked 
for other C/C++ libraries which implements something similar but didn't 
find anything suitable. So this is the main problem for windows. I think 
this library and the skv need to be build with the same toolchain, 
otherwise you'll probably have some problems with missing symbols. So 
that means either MinGW (which I preferred) or Visual Studio. For the 
latter case, you'll could transfer the GNU toolchain of the skv to the 
cmake build system, so we can have multiple OS supported easily. This 
should be not a big deal. But I don't want to have different versions 
for different operating systems. This is not maintainable.

To your questions:

1.) Yes definitely, I wrote the very first version of that code from 
scratch in my 6th semester :-)

2.) Yes I think so. The documentation should definitely improved, but I 
have not the time at the moment. So please go on with that.

Yes we can discuss problems of the program (but please not C++ basics) 
and only at the mailing list.

For the pre-processing we have developed a bladedesigner 
(http://sourceforge.net/projects/bladedesigner/). With this piece of 
software (which has a GUI) you can generate an input cgns file for the 
skv. The output cgns file from the skv can be converted to a vtk file 
format with the distributed python script: scm2vtmb.py So it can be 
viewed in paraview. So the skv is just the fluid solver and I don't 
think it will need a GUI.

A streamline curvature method works only good, if you have good loss 
correlations. Up until now, no loss correlations were implemented. We 
have done an more or less extensive literature research of loss 
correlations. So it would be an option for your students, to implement 
some of these correlations int the skv. I would be really interested in 
that!

If you want to start development, I suggest to open new branch(es) at 
the git repository. So you'll have to be become familiar with the git 
version control system first. These branches does not affecting the 
master branch during development, but can be merged at a certain time, 
when the implemented features are tested and ready. Then I can give you 
also write access to the git repo at sf.net.

So I hope I could answer the important questions. Further questions can 
be discussed at the mailing list.

Best regards,
Oliver Borm

Am 16.02.2013 06:04, schrieb Sujith Joseph:
> Dear Professor Oliver Borm,
>
>             My name is Sujith Joseph and I am an assistant professor  
> in the department of Mechanical Engineering at the Mar Baselios 
> Institute Of Technology And Science
> ( http://www.mbits.edu.in/ ) at Kothamangalam, Ernakulam district,  
> Kerala state, India.
>
> I am a project guide for a group of 4 students
>
> 1. Kevin Bon Francis
> 2. Basil Joy
> 3. Jorbins John
> 4. Elson Varghese
>
> in our final year (8th semester) Bachelor of Technology program. The 
> Bachelor of Technology is the usual 4 year undergraduate engineering 
> degree in India.
> I have suggested writing an axial compressor streamline curvature 
> through flow code as the final year project for these students.
>
> Our initial research gave us several reports on the implementation of 
> such codes in Fortran. Some of these reports were for centrifugal 
> compressors and had
> Fortran 4 code. Other reports were for throughflow codes for axial 
> flow compressors, written in Fortran but the source code was not 
> available.
> Our research finally led to your open source project - streamline 
> curvature method hosted at sourceforge:
>
> http://sourceforge.net/projects/throughflow/
>
> We saw that the code was in German and we would like to translate that 
> code to English. The aim that  I have set for the project group is to 
> study the theory
> and understand the source code and produce a working version of your 
> code for Windows in English. The code is to be fully documented. The 
> final aim
> is to develop  knowledge of streamline curvature through flow theory 
> and to understand how such a program is implemented in source code.
>
>
> We also plan to  translate the documentation (pdf file) you have 
> provided with the code to English.
> I have advised the students that trying to write a streamline 
> curvature code from code would be too ambitious since
>
> 1. They have available only 2-3 months for completing this project 
> before they have to present their project report.
> 2. Just understanding the code and trying to convert the code requires
>     1. Knowledge of compressor aerodynamics and the theory behind the 
> streamline curvature method
>     2. Knowledge of tensors which they have not learnt until now
>     2. Learning the C plus plus programming language
>     3. Programming numerical methods
>     4. If the program is displaying graphical output, then probably 
> the group should know opengl or Direct X etc.
>
> The students are willing to work hard. But before we start on this 
> project I would like your advice on the following questions
>
> 1. Is this a project that can be attempted by undergraduate students?
>
> The students have taken Thermodynamics, Fluid Mechanics, Hydraulic 
> machines (where they have worked with turbine and pump velocity 
> traingles) and Gas dynamics
> (Isentropic flow, normal shock, fanno and rayleigh flow and the basic 
> theory of jet and rocket propulsion)
>
> 2. Is the time frame (2-3 months )for translation feasible?
>
> 3. I assumed that you developed this code on Linux. If we attempt a 
> conversion to windows are there any programming libraries specific to 
> Linux
> which we might find difficult to get on the windows platform?
>
> and most importantly
>
> Will you and the group (your colleagues and student) who have 
> developed this code have some time to help us in our doubts about the 
> theory and code.
>
> We translated several files of the source code using google translate. 
> The comments in the code give us a good idea of what is being done in 
> that file,
> but it will make more sense only after we study the through flow code 
> theory. With some common sense and help from you we believe we can 
> translate the code.
> Apart from google translate we may be able to get the help of catholic 
> priest at a nearby seminary who know german and may be willing to help 
> translate and
> teach German to the degree required for this project.
>
>
> I myself worked for a period of close to 4 years at the French company 
> Safran aerospace (formerly Snecma motors) in Bangalore, India.
> As you know Snecma makes the  CFM 56 and other jet engines. I was in 
> their software engineering division and I worked on Snecma's stream 
> line curvature
> throughflow code called Goux. Snecma were porting this to C++ and were 
> making a QT based user interface for the program.
>
> I was working on the user interface programming with QT with Mr. Jean 
> Francois Rios, who was maintaining this program at Villaroche, Paris, 
> France.
> I have looked through the GOUX code, but could not understand much 
> because I did not know the theory. But from that time I have wanted to 
> attempt to write
> a streamline curvature code, for my own understanding.
>
> If we can port this code to windows and English, as a later project  I 
> plan to put a QT based interface to your code.
>
> Kindly gives us your thoughts in our planned project. We are very 
> grateful to you for taking time to help us in this matter.
>
>
> With best wishes and regards,
>
>
> Sujith Joseph
>
> Assistant Professor
> Department of Mechanical Engineering
> Mar Baselios Institute of Technology And Science
> Nellimatton P. O. Kothamangalam, Ernakulam district
> Kerala,
> India 686693
>
> Phone :           +91 0485 - 2837801 ( 30 Lines) / 0485 3258584
> Personal mobile : +91 9567222194
> Fax : 0485 - 2837830