Control surface issue

[Shakil Salim]

I am working on a Prandtl-inspired flying-wing model in OpenVSP 3.47.0 and I am having trouble with VSPAERO control surface deflections.

I created two elevon/control surfaces on a symmetric wing. In VSPAERO → Control Grouping, both surfaces appear in the Pitch and Roll groups. For the Pitch group, I set the gains as:

• WingGeom_Surf0_Elevon_suef1: +1.00
• WingGeom_Surf1_Elevon_suef1: -1.00

Then in VSPAERO → Overview → Control Group Angles, I set:

• Pitch = +10 deg
• Roll = 0 deg

I clicked Prepare Solver, then Launch Solver, and exported the polar file.

The problem is that the Pitch +10 polar file is exactly identical to the neutral Pitch 0 / Roll 0 polar file. I also tried another debugging run with Elevator/Pitch = 7 deg and Roll = 0 deg. The AoA setup changed correctly, so VSPAERO is running a new case, but the aerodynamic coefficients are still unchanged compared with neutral.

Example neutral and pitch-deflected values at AoA = 0 deg are identical:

• CLtot ≈ 0.700636
• CDtot ≈ 0.029938
• CMytot ≈ 0.417135
• L/D ≈ 23.4027

This makes me think the solver is using the new run settings, but the control surface deflection is not actually being applied to the VSPAERO geometry.

My VSPAERO setup:

• OpenVSP version: 3.47.0
• Analysis: VSPAERO
• Wing-only test first, no fuselage
• Mach: 0.001
• Re: 1e7
• Beta: 0 deg
• AoA test range: -1 to +1 deg
• Control group angle tested: Pitch +10 deg and Pitch +7 deg
• Roll = 0 deg

Questions:

1. Is there a specific setting needed to make the OpenVSP elevon subsurfaces actually deflect in VSPAERO?
2. Should the deflected control surfaces be visibly rotated in Preview VSPAERO Geometry after clicking Prepare Solver?
3. Could this be caused by an old control surface grouping/tag issue or by the way the elevon subsurfaces were created?
4. What files should I check, such as .vspgeom, .csf, .taglist, or other VSPAERO control surface files, to confirm the deflection is being passed to the solver?

I attached the .vsp3 file 

Thank you for any help.

[Rob McDonald]

I vaguely remember this bug - there is probably some chatter about it in the group.

I believe VSPAERO currently has a bug where it only applies the last control surface from a list -- it doesn't apply them cumulatively.

So, when you have pitch +10 and Roll 0, it only applies Roll 0.

If you delete the Roll and just analyze the pitch, it should work fine.

I'll remind the VSPAERO dev about this.

Also, you are using both the thick and thin geometry -- don't do that.  Put each component (in your case only a wing) in one and only Thick or Thin.

Also, you're running 80 AOA points.  There really can't be any good reason to run that many.

The deflected geometry is not visualized in OpenVSP -- the deflection only happens in VSPAERO.  You can see the deflected state in the Viewer.

Rob

[Shakil Salim]

Thank you Rob! 

[Shakil Salim]

Thank you. But why vspaero is stopped working?

--
You received this message because you are subscribed to the Google Groups "OpenVSP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openvsp+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/openvsp/59718dfe-83df-4a82-8f65-4b183887f94dn%40googlegroups.com.

[Rob McDonald]

I don't know, but the first problem is that you have the Thin Set set to Shown and the Thick Set to All.  Don't do that.

If you are going to use Shown / All -- the the other Set needs to be None.  If you are going to use both thin and thick sets, then you need to use Sets properly -- put the desired components into different Sets and use those.

You can't have the same component in both Thin and Thick Sets at the same time.

Then, in your Geom browser, there are a bunch of 'MeshGeom_NGon' items.  Those are created on-the-fly by VSPAERO.  They get automatically deleted when you run VSPAERO a second time, but only if you don't quit VSP before the next run.  They are essentially empty and I should get rid of them, but having this many means you 1) OpenVSP 2) Run VSPAERO 3) Save 4) Quit 5) Repeat.  You can safely delete all of them.

Rob

[Shakil Salim]

Thank You Rob. I am working on a Prandtl-D3C style flying-wing model in OpenVSP/VSPAERO. The clean wing case is behaving reasonably well and gives expected Prandtl-D type trends: bell-shaped span loading, low tip loading, and peak L/D in the expected design-alpha region.

Now I am trying to add a small center fuselage/payload pod and compare the new results against the clean wing. I am keeping the same reference values for comparison:

Sref = 40.5 ft²
Bref = 25 ft
Cref = 1.97 ft
Xref = 2.125 ft
Yref = 0
Zref = -0.196 ft

The fuselage is centered at Y = 0 and is modeled as an ellipse/point body. I set the wing/elevons as the Thin Set and the fuselage as the Thick Set. However, after adding the fuselage, the results are acting strangely. In some runs, the L/D becomes higher than the clean wing, sometimes above 30, which does not seem physically reasonable because I expected the fuselage to add parasite/interference drag and reduce L/D. In an earlier setup, I also saw nonzero side force/rolling moment in the neutral symmetric case, and I suspect I may have had the Thick Set incorrectly set to “All” at one point.

My current question is:

1. For a VSPAERO case with a flying wing plus fuselage, should the setup be:
   • Thin Set = wing/elevons only
   • Thick Set = fuselage only?
2. Should the fuselage ever be included in the Thin Set, or should it only be in the Thick Set?
3. Does VSPAERO underpredict fuselage parasite drag or wing-body interference drag for thick bodies? If so, is it normal for a raw VSPAERO fuselage result to show unrealistically high L/D?
4. What is the recommended way to compare a clean wing vs wing + fuselage case? Should I keep the same Sref/Bref/Cref and moment reference for comparison, then apply a separate parasite-drag correction for the fuselage?
5. Are there specific fuselage modeling or meshing practices I should follow to avoid strange results, such as negative drag, nonzero side force at beta = 0, or artificial L/D improvement?

Any guidance on the correct VSPAERO setup for wing + fuselage analysis would be very helpful. I  share the .vsp3, .polar, and .lod files if needed.

Thank you,
Shakil

[Rob McDonald]

Putting the wing in the Thin Set and the body in the Thick Set is a very reasonable approach -- that is probably how I would do it.

I would certainly leave Sref, cref, bref the same for the comparisons with the fuselage on/off.  Those are reference quantities -- they should relate to the reference wing.

For now, I would probably not rely on the VSPAERO parasite drag results -- particularly for the fuselage.  You can dig into the files to see how the contributions break down a bit, but they are based on extremely simple models with limited applicability.

Both the wing and body parasite drag models for VSPAERO are undergoing a major overhaul -- so I'm hoping in the future that my answer will be different.

For now, I would probably ignore the CDo and CDt (and therefore the L/D) results -- just use VSPAERO for the inviscid results and then in a separate tool add in a parasite drag contribution.

Rob