Problem with CDtot values

[Steve]

Hi everyone,

I'm an aerospace engineering student currently running aerodynamic simulations on a research aircraft using VSPAERO, and I'm running into an issue I can't seem to resolve.

I'm getting negative CDtot values across my components, and I'm not sure whether this is a numerical artifact or something wrong with my setup. I recently increased the Num U tessellation parameter on all components, which may or may not be related.

Has anyone experienced this before? Could it be a mesh conditioning issue — e.g., a bad U/W aspect ratio causing errors in the pressure integration? Any tips on what to check or how to fix it would be greatly appreciated.

Thanks in advance!

[Rob McDonald]

Please post your model -- this post does not have anything for anyone to use to be able to help you.  More information is needed.

Rob

[Steve]

Sorry my bad. 

https://drive.google.com/file/d/1lmed51_V11JH1i9G0rGTpWBgeBOlKEDG/view?usp=drive_link

[Steve]

[Rob McDonald]

There are lots of places to improve your model.

The aft end of the fuselage actually self-intersects and inverts.  That causes a hot-spot in the solution that is likely a problem.

Your vertical tail pops out of the top of your horizontal tail -- you probably don't want this, adjust the model so this doesn't happen.

Overall, your surface tessellation is far too fine - and in non-sensical places.  All this is doing is making everything take a lot longer during testing and development.

Likewise, you have VSPAERO set up to run 20 points.  Start with one, get things working the way you want.  Then, maybe go to three, see how it spans the area you're interested in.  Then, if you actually have a good reason, add more.  You probably don't need more.

Next, your fuselage skinning is not well done, it creates shapes that you probably don't want.  Watch the Ground School Skinning video to get some quality tips on how to use skinning.

Rob

[Steve]

Thank you for the tips. I was able to get the total drag coefficient to a more realistic value. However, I now have another issue: when calculating the parasite drag, the wetted area that OpenVSP computes is significantly larger than the actual geometric surface area.  I also compared my Cl values of the wing to the NACA report and found a certain misalignment.  I've tried troubleshooting it but haven't been able to resolve the issue.

[Rob McDonald]

Why do you think the wetted area numbers are wrong?  When I look at Parasite Drag and also CompGeom, everything looks reasonable.  What numbers are you seeing -- what do you expect -- what makes you think there is a problem?

You say you're comparing to "the NACA report" -- there are thousands of NACA reports.  We can't possibly have any idea what comparison you're making, what you expect to see, or even what you are seeing.

You need to put yourself in the position of the people who could possibly help you -- you must provide us with enough information to be able to help.

Rob

[Steve]

When I ran the parasite drag calculation, I got some questionable results. Either my calculation method is off, or I made a mistake somewhere: the calculated wetted area (Swet​) turned out to be significantly larger than the actual geometric surface of my aircraft. Additionally, the component drag coefficients (CD0​​) I obtained seem far too high based on my judgment.

I have already watched several tutorials on this calculation, but I still couldn't figure out how to get reasonable results. 

For the verification of my simulation model, I compared my results against the data presented in NACA Report No. 460, page 35 (https://ntrs.nasa.gov/api/citations/19930091108/downloads/19930091108.pdf). In that report, the lift coefficient (CL​) versus angle of attack (α) diagram shows a CL​ of approximately 0.4 at an angle of attack of around 0 degrees.

However, when I ran my own simulation, I obtained a CL​ of about 0.55 – and this was for the full aircraft model, not just the wing. I don't understand how it is physically possible that the lift coefficient of my entire aircraft configuration could be higher than that of the isolated wing from the NACA report. This seems counterintuitive, since the wing is the primary lift-generating component, and adding a fuselage and tail surfaces typically adds drag and may slightly reduce or at best maintain overall lift efficiency – but certainly shouldn't increase it beyond the clean wing's performance.

[Rob McDonald]

There have been times when the parasite drag tool behaves poorly and gives seemingly random results...  Does it give the same answer if you use it repeatedly?

It may also be that you had a MeshGeom from a prior analysis in the geometry Set that was throwing things off.  Your model is using either the 'All' or 'Shown' Set.  When some operations run, they add a geometry to the browser.  Then, if you later run an analysis with 'All', your geometry will effectively be duplicated with a perfectly overlapping geometry that can mess things up.  You'll want to delete those Geoms when you run.  Better yet, learn how Set works and always use an intentional Set when doing analyses...

If you still get those unreasonable results, please document the exact steps you take from opening a file and every click to duplicate the issue.  Also, tell us what platform (Windows / Mac / Linux) you are running on, etc.

It appears you're comparing to a NACA 6412 airfoil, but your settings are slightly off.  It appears you moved the sliders until the text name read 6412.  There is a lot of rounding that goes into making that name, so if you want to do careful comparisons, you'll want to edit the fields and type in exact values to represent a particular airfoil.  In this case, it only changes the design lift coefficient by about 3%, so it isn't the primary problem, but it is something you'll want to track if you're being careful about your work.

There are a lot of things that are different between the test in the NACA report and your analysis.  We need to be able to exactly duplicate whatever you are testing in order to be able to help you.  A first step towards that is to save your file with exactly the analysis settings (say in VSPAERO) that you are running -- that way we can just open the file, hit run, and see the same thing you see.

Rob

[Steve]

I somehow still get these unreasonable results, although I have used it repeatedly.
I'm on a Windows 10 system.

So, I first open the .vsp3 file. Then I open the CompGeom window. There, I have Set_0 selected as the Normal Set, and Degen Set is set to None. I also selected Subsurfs. Then I execute it. Although I get a MeshGeom, the window stays empty where normally the Wetted Area, etc., are listed.

Then I do a Degen Geom; the Set is Set_0, and I execute it. It says: "Wrote 34 components and 0 blank geoms" (I saw these steps in a video: https://www.youtube.com/watch?v=MZywiGpBlME&t=920s).

Then, when I launch the Parasite Drag analysis, I set the Geometry Set to Set_0 and the Length Unit to meters. As the Lam. Cf Eqn., I use Blasius, and for the Turb. Cf Eqn., I use Schlichting Compressible. For the reference area, I take the values from the model and set the ref. wing to 3_Flugel.

The atmosphere is the US Standard Atmosphere 1976 with Vinf = 252 km/h, Alt = 27000 ft, dTemp = 0, and Temp = 234.7 K.
The FF equation for the Rumpf is Hoerner Streamlined Body, and for the Hohenleitwerk, Seitenleitwerk, and Flügel, it is also Hoerner.

For the Airfoil, I now have selected the right parameters.

Then, in the VSPAero analysis, I use Set_0 as the Thick Set and None for the Thin Set (because I read that it is better to use the 3D geometries).

For the Wake, I have: Num It = 30, Num Nodes = 64, Relaxation = 1, Freeze It = 10. For Alpha, I have 0 to 12 with 3 steps, Beta = 0, Mach = 0.23, and ReCref = 6.6e+06 – this is one difference from the NACA report, where they had 3.09e+06. For the reference length, I took the values from the model.

In the Advanced section, I have Stallmode set to on and CLMax2D set to 1.7 (because this was the highest value in the report). Clo2D is 0 – I didn't find anything that could really help me figure out what this actually is.

For the Convergence Factors, I have: forward mat = 0.1, adjoint mat = 1, non-linear = 0.1, core size = 1, freeze it = 10000, far away = 5.
For the Advanced Flow Conditions, I have Vinf = 70 and rho = 0.002377.
In the other tabs, I didn't change anything.

I hope that I have given you all the necessary information and the required files.