Blended Winglet in Version 3.4

[Vijay Matheswaran]

Hello, 

 I've been trying to design a very simple winglet to test a VLM code that I've been writing. However, I notice that the wing blend functionality has been removed for version 3.4.0. Is there a way to go about designing a blended winglet in version 3.4?

Regards, 

Vijay

[Rob McDonald]

Depending on the severity of the blend, you might be able to construct
a blend with a fuselage or stack component.

There is no blended wing support right now. I've done some work on
making it work well and have a roadmap for how to get there, but it is
going to take some non-trivial development work in the underlying
geometry library.

Rob

> --
> 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.
> For more options, visit https://groups.google.com/d/optout.

[Vijay Matheswaran]

Rob, 

 Thank you. Could you perhaps share the roadmap you mentioned? 

Vijay

[Rob McDonald]

On Fri, Dec 4, 2015 at 9:54 AM, Vijay Matheswaran
<vijaymat...@gmail.com> wrote:
> Rob,
> Thank you. Could you perhaps share the roadmap you mentioned?

Roadmap may have been too strong a word (:

I was never happy with the wing blending in v2. As with some other
features, rather than bring forward capability that was not up to
standard, we've set blended wings on the back burner until they can be
done right.

After fuselage skinning was introduced, some users asked that I extend
this functionality to wings to create blended wings. I suspected it
wasn't going to be that simple, but I did some experimenting on my
own.

Originally, skinning control allowed you to control 'angle',
'strength', and 'curve' along four spines of a fuselage. The 'angle'
parameter is defined perpendicular to the surface. This works pretty
well for fuselage and stack. However, when you think about a swept
wing, angle would work for the LE and TE, but the top/bottom surface
need to be skewed parallel to the surface.

Sure enough, my experiments showed that simply porting skinning to
wings was insufficient.

So, sometime later, I extended skinning control to also allow 'slew'.
This is an angle control parallel to the surface. This has succeeded
in making fuselage and stack components more flexible and is a great
feature to have added.

I updated my blended wing experiments to utilize slew (why slew was
created after all), and while it was a dramatic improvement from
before, it still was unsatisfactory.

Skinning works by taking a set of boundary conditions and solving a
linear system of equations to place the intermediate Bezier control
points. In the typical case, the root/tip airfoils are 3D curve
position boundary conditions. You can imagine that the 'spanwise'
derivative of the wing has a value all the way around the airfoil --
this is another 3D curve boundary condition. When you specify
curvature, you add in an additional 3D curve boundary condition.
Skinning allows you to specify any combination of those conditions at
root and tip. It then does the magic to solve for a surface that
obeys all that information.

Skinning however has no means to control, constrain, or restrict what
happens between the boundary conditions. In the case of skinning
wings, the top/bottom surfaces near the trailing edge had a tendency
to get quite warped -- often intersecting one another. They certainly
did not maintain the integrity of something that looks like an
airfoil.

That is where the code stands today.

The plan going forward is to use a totally different mathematical
approach to generate the surface in-between two airfoils. We believe
we understand the math and theory enough that we know this will work
and we know that it is a mathematically well posed problem. However,
this bit of code hasn't been started, is non trivial, and will need to
be implemented in a pretty tricky place.

The new algorithm goes something like this...

-Start with root and tip airfoils for a section positioned in 3D in
their final location using the current wing code.
-Construct the straight-line wing surface as is done today (perhaps
revisit this in some subtle ways later).
-Extract a number of intermediate curves from root to tip across that
section. These airfoils will transition in thickness and camber while
maintaining the integrity of the airfoil shapes.
-Construct 3D blended leading and trailing edge curves from
user-specified parameters. This will be done in a way very similar to
skinning -- but just curves, not the entire surface.
-Intersect the intermediate curve planes with the LE and TE curves --
this will find the new corresponding LE and TE points for each
intermediate curve.
-Transform the intermediate curves (scale, skew, translate, rotate)
such that their LE and TE points coincide with the intersection points
found in the previous step.
-Solve for the surface that goes through the root/tip and all
intermediate curves.

This last step is really where the magic happens. I can't explain it
without an assumption of a pretty deep understanding of Bezier curves
and surfaces (and probably a white board). So, I'll stop here for
now. But that bit is the tricky bit.

So, my 'roadmap' has each of those algorithmic steps as milestones,
but I don't have a schedule or deadline associated with each of them.
In some cases they depend on other members of the development team who
of course have a wide variety of demands on their time.

That is the plan, no promises on when it will be complete.

Rob

[Karén Melikov]

Rob,

are there any plans of maybe returning the capabilities of the dihedral menu from v.2 for making blended winglets? 

Thanks.

пятница, 4 декабря 2015 г., 12:09:40 UTC-5 пользователь Rob McDonald написал:

[Rob McDonald]

Our team is working on blended wings.

It won't be like v2's blended dihedral -- and it won't be like the
v2's BWB component. In a really good way, it should be awesome!.

We've actually made two tries at it internally. First, just applying
Fuselage's skinning to wings. Then, adding in Slew control and using
it appropriately. Unfortunately, the airfoil behavior between
specified wing stations was unacceptable.

We have a plan that I believe will work, but we aren't there yet.

Rob

[Karén Melikov]

Sounds exciting. I'm looking forward to it.

пятница, 6 мая 2016 г., 18:12:38 UTC-4 пользователь Rob McDonald написал: