[OpenVSP] Airfoil changes

[Klarisse Ng]

Hello All, 

I am currently doing some research on the effects of airfoil changes on changes in aerodynamic coefficients in VSPAero using the B737 aircraft.

I have increased the T/C ratio such that my 5-digit airfoil changes from 22112 to 22122 to 22132 and so on. However, I noticed that some values in my graphs are overlapping and do not show much change. Here are some of my graphs for reference. 

I tried changing the cell size, but that only changed the values drastically.

(Green value is where cell size is changed)

Is this expected, and if not, what are some possible issues?

Thank you.

[Rob McDonald]

Most of your images are not coming through.

Are you representing the wings in thick or thin mode?

Rob

On Sun, Nov 30, 2025 at 7:49 PM 'CHEW ZHI XUAN, ERIN HCI' via OpenVSP <ope...@googlegroups.com> wrote:

Hello All, 

I am currently doing some research on the effects of airfoil changes on changes in aerodynamic coefficients in VSPAero using the same settings. 

I have increased the T/C ratio such that my 5-digit airfoil changes from 22112 to 22122 to 22132 and so on. However, I noticed that some values in my graphs are overlapping and do not show much change. Here are some of my graphs for reference. 

I tried changing the cell size, but that only changed the values drastically.

\

(Green value is where cell size is changed)

Is this expected, and if not, what are the possible issues?

Thank you.

--
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/180025fd-9497-48ca-a9fe-2da87b933210n%40googlegroups.com.

[Klarisse Ng]

Apologies, here are the images. The wings are in thin mode.

Here is the photo of the wing:

 

Graphs:

Graph after changing cell size (For green line, 22152)

Thank you.

To view this discussion visit https://groups.google.com/d/msgid/openvsp/CAEppYpGcsM_T0OrVcuDDXASOfpnayqMEjiCZ%2Bj_gfHVZ9z-w7Q%40mail.gmail.com.

[Rob McDonald]

When using thin surface representation, I would expect very little difference between cases where only the thickness of the wing is varying.

When using the thin surface representation, VSPAERO currently only 'knows' about the mean camber surface as represented in the mesh sent to VSPAERO.  At present, VSPAERO does not use any thickness information when analyzing a wing in thin surface mode.

When OpenVSP creates the thin surface mesh, the thickness has some influence (so we don't get exactly the same results).  Mesh points are arranged on the thick surface (before clustering) based on equal arc length.  Then, corresponding top and bottom surface points are averaged to find the mean surface.

Consequently, a thicker version of the same airfoil will end up with a different mesh -- it will effectively have more clustering at the LE.  That different mesh will result in a slightly different solution.

In particular, it looks like you've considered the wing with a t/c=1.0 -- which seems to blow up in some situations.  That should not be surprising, as that is an absurd thickness for a NACA 5-Digit airfoil.

Rob

[Klarisse Ng]

Thank you for the clarification. Could we check if we should expect a similar trend for the thick mode? We tried running the wings in thick mode as well but ended up with the same problem we faced with thin mode (I.e. graphs overlap, little difference in results)

Thank you! 

To view this discussion visit https://groups.google.com/d/msgid/openvsp/f53bfa94-28dd-413d-aa4c-f536623e01b8n%40googlegroups.com.

[Rob McDonald]

The results will also be very similar in thick mode.

VSPAERO is inherently an inviscid / incompressible code.  There is a very simple compressibility correction and a simple viscous drag model.

The simple viscous drag model does not currently take thickness into account.

I would expect the wing's thickness to make a very slight change to the lift curve slope -- but essentially no change in the load distribution (i.e. the induced drag at a given cl).

Larger thickness will increase the surface velocities on the upper/lower surface, but these effects largely cancel when you integrate to delta Cp and to loads (and VSPAERO does not calculate loads this way anyway).  However, the point is, unless you are monitoring the velocity at a certain point on the surface, you probably won't see a big change due to thickness.

This is why the thin wing representation is a good approximation -- the thickness doesn't matter that much if you aren't worrying about local surface velocities.

Rob

[Brandon Litherland]

Have you considered replacing the CD0 term from VSPAERO with results from the parasite drag tool?  I'd expect the thickness to have a greater effect on the parasite results which you could then use to obtain different estimates for L/D and the drag polars.  Are you also considering the stall model with this wing?

[Rob McDonald]

Is this question directed at me, or at the OP?

If me -- then yes, in the next ~6 months, I expect VSPAERO will have a different approach that will be sensitive to the airfoil thickness.

Rob

[Brandon Litherland]

Apologies, Rob. That was meant for the OP. I was suggesting that they take a different approach to capturing different wing performance parameters than relying solely on VSPAERO output.

OP, you will find other posts in the Group discussing drag buildup approaches including separating out the lift-dependent and non-lift-dependent parts of VSPAERO's CD0 result.