Flaps curving instead of being straight when extended spanwise with fixed length

[Jasper]

Dear Openvsp Team,

I am trying to make full span maneuver flaps on my aircraft wing leading edge, however, when i try to extend it spanwise beyond a certain point, the lines start wiggling and curving. I have tested what can cause this on testing wings and what my observations so far are; when the wing has very low aspect ratio, such that there is a negative trailing edge sweep, the lines indicating the flaps become curved as they are made longer (spanwise). This does not occur on regular wings/conventional wings, where there is similar trailing edge sweep as leading edge sweep (as far as my testing goes). Furthermore, this only occurs when I try to have constant chordwise length of the flap. If I adjust it to L/C, then it is no problem again.

Would you know why this happens and how to fix it?  I am working in OpenVSP 3.49.0. on windows 11, and have attached images of the problem and my working file to the message.

Kind regards,
Jasper

[Brandon Litherland]

You're seeing the underlying curve of the UW surface coming into play.  You're tapering the wing while also significantly changing the L/C of the control surface which causes the underlying surface curve to wiggle across the UW.  Note that, visually, the curve is there but if you were to run this through VSPAERO with deflected control surfaces, the panels would be affected mostly correctly.  I can take a look but I'd first try running your simulations and see if anything looks weird.

[Brandon Litherland]

Quick note: VSPAERO does an intersection of the CS onto the thin surface and then bends the panels based on that subsurface section.  You can see it if you do a Thin set CompGeom. You'll have to determine if this is good enough for your analyses.

[Rob McDonald]

Brandon describes the situation correctly.

The subsurfaces are computed in survade UW (or UV) space.  Show the wireframe and you will see lines of constant U & W.  U is spanwise and is easy to deal with.  The tricky part on this one is W.

When doing control surfaces, OpenVSP solves for the W to give the desired control surface length -- it does this at four points along the control surface (endpoints and two intermediate).  It then fits a cubic curve (in UW coordinates) along each edge of the control surface.

Under many circumstances, this approximation is not noticeable.  However, in situations like yours (where the L/C of the control surface is dramatically different from start to end), it becomes noticable.

The easiest solution is to break the control surface into two (or more) segments in the spanwise direction.  This will effectively make the approximation better.

Rob

[Jasper]

Dear Brandon and Rob,

Thank you for the quick responses, that makes it clear. I have now adjusted the HLD's by putting two segments next to each other and it works well.

thanks again,
kind regards,
Jasper