[pro_]

https://www.eccomasproceedia.org/con...ress-2016/1903

As seen in the paper above, in particular Section 4.1.2, a flutter boundary was predicted for the Isogai Case A Wing Section. Within that section there has been mention of 'damping extracted from the dyanmic response' but no further. My questions are:

1. How did the authors of the mentioned paper go about in 'extracting the damping (coefficients)?
2. Does SU2 provide some form of python wrapper that would both simulate the aerodynamics, solve the Aeroelastic EOM AND calculate a damping coefficient?
3. (I believe this is more of a Aeroelastic/Flutter theory question) I have seen methods such as the Matrix Pencil Method https://doi.org/10.2514/1.J058663 used in order to estimate the damping coefficients froma pitch-plunge time history. Based on the pitch-plunge time history output of SU2, what type of methods would you suggest to obtain the flutter boundary? Would a flutter calculation in the frequency domain be possible?
4. Im using SU2 7.0.3 on Windows, Is there a way for me to use the python wrappers, I cant find any guidelines in the Build on Windows in the SU2 website concerning python wrappers? (Sorry I had to include this, although off-topic)

[hlk]

Thank you for your question.
I was an author on this publication, and although I was not the author responsible for that particular calculation, I believe that they use the classic mass-spring-damper equation, fitting the dynamic response to an equation of that form and selecting the damping coefficient from there.



For more information on the equation I referred to, see: https://faculty.washington.edu/seatt...reading-3b.pdf.



The paper you link involves multifidelity simulations and machine learning, and while I applaud you for going into that depth, if you are asking simply how to find the flutter boundary from a single time history, and trying to learn aeroelastic theory, those methods are not necessary.


You may be able to achieve what you by modifying on of the python scripts, which will require installing python with the correct modules (for example, the anaconda distribution should include everything you need), and setting the appropriate environment variables as described in the installation instructions. There are several example python scripts included with the code, beyond those described in the tutorials, and while I don't think any of them do exactly what you describe you may be able to find something you can adapt to what you need. For example, you will find the script used for FSI simulations like some of the other examples in that paper that involved flexing cantilevers.



For the python wrappers, you will need to compile the code as described here:
https://su2code.github.io/docs/Python-Wrapper-Build/. It is not clear from your question if you are referring to this specific functionality - and it looks like this may have been taken out for v7 and may no longer be supported. On the other hand, version 7 has improved the documentation for compiling the code on windows.



Since it has been a while since I studied flutter, the best I can do off the top of my head beyond the most basic equation I discussed at the top is to refer you to aeroelasticity textbooks for more method/theory recommendations. And yes, this can involve the frequency domain.

Originally Posted by pro_

https://www.eccomasproceedia.org/con...ress-2016/1903

As seen in the paper above, in particular Section 4.1.2, a flutter boundary was predicted for the Isogai Case A Wing Section. Within that section there has been mention of 'damping extracted from the dyanmic response' but no further. My questions are:

1. How did the authors of the mentioned paper go about in 'extracting the damping (coefficients)?
2. Does SU2 provide some form of python wrapper that would both simulate the aerodynamics, solve the Aeroelastic EOM AND calculate a damping coefficient?
3. (I believe this is more of a Aeroelastic/Flutter theory question) I have seen methods such as the Matrix Pencil Method https://doi.org/10.2514/1.J058663 used in order to estimate the damping coefficients froma pitch-plunge time history. Based on the pitch-plunge time history output of SU2, what type of methods would you suggest to obtain the flutter boundary? Would a flutter calculation in the frequency domain be possible?
4. Im using SU2 7.0.3 on Windows, Is there a way for me to use the python wrappers, I cant find any guidelines in the Build on Windows in the SU2 website concerning python wrappers? (Sorry I had to include this, although off-topic)

[pro_]

Originally Posted by hlk

Thank you for your question.
I was an author on this publication, and although I was not the author responsible for that particular calculation, I believe that they use the classic mass-spring-damper equation, fitting the dynamic response to an equation of that form and selecting the damping coefficient from there.



For more information on the equation I referred to, see: https://faculty.washington.edu/seatt...reading-3b.pdf.



The paper you link involves multifidelity simulations and machine learning, and while I applaud you for going into that depth, if you are asking simply how to find the flutter boundary from a single time history, and trying to learn aeroelastic theory, those methods are not necessary.


You may be able to achieve what you by modifying on of the python scripts, which will require installing python with the correct modules (for example, the anaconda distribution should include everything you need), and setting the appropriate environment variables as described in the installation instructions. There are several example python scripts included with the code, beyond those described in the tutorials, and while I don't think any of them do exactly what you describe you may be able to find something you can adapt to what you need. For example, you will find the script used for FSI simulations like some of the other examples in that paper that involved flexing cantilevers.



For the python wrappers, you will need to compile the code as described here:
https://su2code.github.io/docs/Python-Wrapper-Build/. It is not clear from your question if you are referring to this specific functionality - and it looks like this may have been taken out for v7 and may no longer be supported. On the other hand, version 7 has improved the documentation for compiling the code on windows.



Since it has been a while since I studied flutter, the best I can do off the top of my head beyond the most basic equation I discussed at the top is to refer you to aeroelasticity textbooks for more method/theory recommendations. And yes, this can involve the frequency domain.

Thanks for the reply, the basis of my topic is basically applying a regression model to a Mach vs Flutter Speed Index vs Damping Coefficient surface which would allow me to get a contour of the flutter boundary. I have mentioned the use of a Matrix Pencil Method to estimate damping from the pitch/plunge signals, and for the public, here is the link to a python file that applies the Matrix Pencil Method to the latter signals. Although it is only a part of a larger tool for aeroelastic analysis.

https://github.com/gjkennedy/funtofe.../matrix_pencil

Within that repository there is also some link to pySU2, is this noticed by the SU2 developers?
I have posted another thread: How do set a steady solution as an initial solution to an unsteady simulation?
Wherein, there are some issues in the steady state initialization for the aeroelastic analysis. (A reply from pcg). Maybe this can be looked into?

[hlk]

I'm sorry, I didn't fully understand what you are trying to do.


Yes, the developers are aware of pySU2, and both gjkennedy and pcg are quite involved with SU2 themselves. If I recall correctly, the pySU2 work is what is referred to as the 'python wrapper' in the SU2 documentation - where sub-functions of the C++ code are exposed to python tools. The link I shared in my last post should take you to the instructions for compiling that - but be aware that they may only be compatible with a past version of SU2.



As a point of clarity, sometimes folks can be confused that this is a completely different tool than the python scripts, which technically speaking are a wrapper around the SU2 code, but are distinct from the swig-based pySU2 wrapper. These scripts are the scripts stored in SU2_PY (or in bin/ after compilation or if downloading the already-compiled code) and used in several of the tutorials.

Effectively, there are two established ways (that I know of - the magic of open source means that while I wasn't looking someone could have generated a new and better one) to wrap SU2 with python - one of them requires recompilation and is referred to by pySU2 or the "python wrapper", and the other does NOT require recompilation and is usually referred to as the "python scripts" despite technically being a wrapper also.