Influence of wing parameters on surrogate model

[Дмитрий Тарабащук]

Dear Suave team,

I run BWB.py tutorial with base_data.txt . And I have a few questions:

1) Why do we need wing parameters (I think we do not need these parameters since we already have data in txt file - CL, CD )? I wanted delete part of code with wing parameters, but code doesn't run.

2) When I change wing parameters, for instance quarter_chord  I see that drag coefficient changes. Why it happens?

3) Can I run code with surrogate data  without wing parameters? If yes, can you explain how

kind regards,

Dmitrii

[Emilio]

Hi Dmitrrii,

1) If you're CFD is inviscid then the viscous part of the drag needs to be built up. If you have viscous CFD with all corrections you can set your analyses to skip the relevant methods. Regardless the vehicle level reference area must be set to get dimensional forces.

2) See 1), it's changing drag parameters as part of the full build up. It's possible the default right now is using an oswald correction as well. All of this can be disabled.

3) You can see a list of calculations here. Disable what you don't need in your script (you don't need to edit any code). Use SUAVE.Methods.skip

It will look like this:

aerodynamics = SUAVE.Analyses.SU2_Euler()

aerodynamics.compute.drag.parasite = SUAVE.Methods.skip

...

I think you'll still need the total drag method.

-Emilio

[Дмитрий Тарабащук]

Emilio, Thank you a lot for your brilliant advices!

1) Could you clarify about vehicle reference area?
I use concorde tutorial as reference for my project with surrogate aerodynamic data. in code I removed all geometry parameters,  except from  basic parameters (code doesn't work without vehicle.reference_area) and  instead of manually filled data for Wing, fuselage, etc. I read VSP file . Also I find out that changing: "# basic parameters vehicle.reference_area"  my drag coefficients in mission are changed too, is that vehicle level reference area that you pointed out?

So what I did, I randomly picked the value  vehicle.reference_area and my results from  mission plots (AoA, M, Cd, CL) matched with my input data in base_data.txt. Am I did right?

2)  Where is placed  oswald correction? What does it?

3) I have CFD RANS data and I filled them into base_data.txt, so I wanna only interpolated my CFD data without others drag calculation in SUAVE. am I right that I should skip all drags? But to be honest when I skip one of the drags I have errors...

and why do you think that I need total drag?

Sorry for so many question. Emilio, you are my savior.

kind regards,

Dmitrii

понедельник, 29 августа 2022 г. в 21:18:54 UTC+2, Emilio:

[Дмитрий Тарабащук]

P.S.

I found out that SUAVE extrapolate only lift coefficient and drag is calculated inside SUAVE. 

So when I replace drag coefficients in my base_data.txt file from real number to  zero for each mach number and AoA plots with drag will not change.

How can I replace calculation of CD with only my data? I don't need calculation of different drag components , I need only extrapolation of Cd from my base_data.txt as it is done with lift coefficient.

Thank you for your patience :)

среда, 31 августа 2022 г. в 23:07:47 UTC+2, Дмитрий Тарабащук:

[Дмитрий Тарабащук]

As I understood root of my problem in SU2_Euler.py and SU2_inviscid.py.  SU2_inviscid extrapolates only Lift coefficient, then in SU2_Euler Cd calculates by means fidelity zero. I tried many times replace calculation of drag/ turn of  Fidelity_Zero drags with drags from base.txt file, but I always face with errors.

Dear Emilio, is it possible fix this problem?

kind regrards,

Dmitrii

четверг, 1 сентября 2022 г. в 23:43:09 UTC+2, Дмитрий Тарабащук:

[Emilio]

Hi Dmitrii,
1) A vehicle level reference area is used to dimensionalize the lift and drag coefficients into a force. Changing this will of course change the results of the mission. Given a fixed weight and fligt conditions a lower wing area will require a larger lift coefficient and consequently drag coefficient. Book keeping the area value is essential.

2) You won't need the oswald efficiency if you're inputting the drag values.

3) I think you should keep the untrimmed, trim, and total drag methods. The untrimmed sums the sources and the trim adds an extra markup for elevator deflection that's likely not accounted for in your CFD. Then total puts it all together. Keep the rest turned off as you've done.

About extrapolation, yes SUAVE will extrapolate if there is no data provided. This can cause mission convergence issues.

[Дмитрий Тарабащук]

Emilio,

1) I do not understand why in surrogate method is written only lift.inviscid why we do not have drag.inviscid? In base_data.txt drags are included too

2) am I right, that in SU2_Euler we do not use drag coefficients from base_data.txt file? We refer to Fidelity_Zero method and all  drags   are calculated by this methods. So we do not use extrapolation of Drag from SU2_inviscid?

As I see, for instance in SU2_Euler we have lift.inviscid where is located surrogate data of lift but there is no the same for drags, I tried add compute.drag.inviscid = SU2_inviscid(), but every time face with errors.

3) I kept  only the untrimmed, trim, and total drag methods, but faced with error 

Sorry for silly question. I'm stuck on this issue

Kind regards,

Dmitrii

вторник, 6 сентября 2022 г. в 20:11:13 UTC+2, Emilio: