Setting up higher fidelity analysis

[Sebastian Leblic]

Hi everyone,

I've recently begun using SUAVE, having a beginners-level background in all things python, linux, etc. but a good understanding of cfd and aerodynamics analysis.

So far, I have installed and ran tutorials with not too much of an issue. I ensured integration with OpenVSP, SU2, and GMSH are all working through us of the BWB tutorial and removing the given history file so as to force the program to work through the entire meshing and cfd process.

My goal at the moment is to take the Concorde tutorial and work with it to get higher fidelity results. I'm hoping this process should help me get a basic understanding/aptitude to produce results of various fidelity for any aircraft I am looking to analyze.

At the moment, I have copied the Concorde tutorial script and have made slight changes (ex. calling SU2_Euler_Super instead of Supersonic_Zero) to try and initiate a higher fidelity analysis. The problem I am currently trying to figure out is that the base_premesh.vsp3 file that is generated is not being read properly and so the necessary base files are not created. The error looks like this:

 

I was hoping the geometry settings given by the concorde tutorial to describe the aircraft would be enough; when comparing to the BWB tutorial, it specifies meshing parameters for specific sections of the wing but I am of the understanding that it shouldn't be necessary to specify meshing parameters to run. I understand it shouldn't produce an optimal mesh on its own, but I am just trying ensure the process works in general when I set up my own analysis.

An aside from the above problem: I am wondering how to determine what would be the best methods to look at for a higher fidelity supersonic analysis as when I look into the doxgyen notes on the methods, for example SU2_Euler_Super, they indicate the assumption of being within the subsonic regime. Nowhere have I confirmed what methods would be best to assess the supersonic regime. Any indicators as to where I can get better information on analysis methods would be much appreciated!

I hope this is clear enough - please let me know if more clarification is needed from my end. Thanks in advance for the help, I look forward to hearing back soon!

Sebastian

[ebot...@gmail.com]

Hi Sebastian,

Yeah, it looks like there's a couple of small bugs, when we updated the aero analyses the SU2_Euler_Super didn't get updated. I'll pull that into the develop branch early this week. However, I didn't get any of the same errors as you got.

Also you're right the defaults for the Mach numbers are subsonic (someone must have gotten a little too copy and paste happy) but that shouldn't change the results. The methods that are being called are indeed tuned to supersonic performance. I'll fix the defaults, although I suggest you set those manually anyways.

I was able to get the Concorde running by dropping the Concorde vehicle into the BWB tutorial though. A couple of small changes were required including getting rid of mesh setup (as you mentioned) and renaming the turbojet. If you don't rename the turbojet it will try to mesh a solid cylinder, it's better to just not include the turbojet.

One thing I would caution about this, it was not intended to run the full flight profile of the vehicle from takeoff to landing with CFD. It's more intended to run CFD at cruise. Doing the full flight profile would be prohibitively expensive and I wouldn't trust automatically generated meshes at higher angles of attack. So the surrogate models look like they assume just supersonic (no mixed super/sub/trans). We can update that if you really want, since we do it all the time with Fidelity Zero.

Since you have some CFD experience, my suggestion long term would be to export to VSP, mesh manually, run SU2 manually, and then use the CFD analyses like we do with the BWB tutorial with an aero deck. CFD is so mesh dependent that an experienced user can do far better than these scripts can do.

Let us know if you have anymore questions.

-Emilio

[Sebastian Leblic]

Hi Emilio,

Thank you for getting back to me.

I will be looking to use the BWB template as a starting point and see how that goes. This is just my initial stab at SUAVE, so in the near future I hope to move away from using just tutorials as templates and become proficient at starting from scratch. Anyways, thanks for looking into this and for your feedback - much appreciated.

I definitely understand that SUAVE is not expected to generate the highest quality meshing/CFD process and if higher/highest accuracy is sought, then it is best to do the CFD externally.

Just wanted to say thanks for all the effort you guys put into keeping on top of things/problems people are facing - your timely responses don't go unnoticed.

Sebastian