HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Feb 28, 2013, 07:30 PM
Jim C Patrick
jcpatrick's Avatar
Shenandoah County
Joined Jan 2008
819 Posts
Swifts and swallows have scythe-like wings; whippoorwills, goatsuckers and nighthawks are night-flyers with wings like that too. All are land birds.

Barn swallows use the full wing to attain speed, then partly close them and use the tip section to spin, weave, zip, and turn. They catch flying insects (gnats and mosquitos) but also fly just for the fun of it.
jcpatrick is offline Find More Posts by jcpatrick
Reply With Quote
Sign up now
to remove ads between posts
Old Feb 28, 2013, 08:54 PM
Registered User
The Willamette Valley, Oregon
Joined Dec 2008
1,118 Posts
Quote:
Originally Posted by jcpatrick View Post
Swifts and swallows have scythe-like wings; whippoorwills, goatsuckers and nighthawks are night-flyers with wings like that too. All are land birds.

Barn swallows use the full wing to attain speed, then partly close them and use the tip section to spin, weave, zip, and turn. They catch flying insects (gnats and mosquitos) but also fly just for the fun of it.
Absolutely. And generally speaking, not soaring birds. Though swifts and swallows like thermals if they are full of bugs. Interesting.

Note the extreme dihedral of nighthawks, typically. Possibly not in the midst of maneuvering?

Swifts strike me as flying with slight anhedral. Swallows, I'm not so sure. Hey, they'll be here soon!

Steve
aeronaut999 is offline Find More Posts by aeronaut999
Reply With Quote
Old Mar 01, 2013, 12:29 PM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
Below is my attempt at a Vortex-Lattice Seagull. Unfortunately, the seagulls don't appear to have standardized their soaring configuration, so I tried to base the geometry on some of the more exaggerated anhedral-dihedral pictures I could find.

Not too surprisingly (given the significant outboard anhedral), this configuration has negative apparent dihedral effect (beak to the left wants to roll to the right). With no twist, this planform shows a span efficiency factor of just over 1.0
ShoeDLG is offline Find More Posts by ShoeDLG
Reply With Quote
Old Mar 01, 2013, 01:54 PM
A witty saying proves nothing.
kcaldwel's Avatar
Joined Jan 2007
3,565 Posts
Quote:
Originally Posted by ShoeDLG View Post
With no twist, this planform shows a span efficiency factor of just over 1.0
I wonder if these effects show up with VLM? The analysis I've seen uses higher order panel methods. That seems to be required to really get decent results for planform and tip shapes. Force-free wakes also have to be modelled.

Is your VLM like the one in XFLR5, and almost any planform has a span efficiency over 1?

What happens to the span efficiency if you use a similar dihedral angles, instead of anhedral, in the outer panels?

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Mar 01, 2013, 02:11 PM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
Kevin,

I'm calculating drag by summing the streamwise force component on each element. I suspect I could do better with Trefftz plane integration, but I haven't implemented that yet. Although the wake isn't relaxed (force free), it is streamwise (drag free). I have seen cases where the span efficiencies are suspect (like 1.005 for a planar rectangular wing), but they are not routinely above 1.0. I can "reflect" the anhedral into dihedral and see what happens to the span efficiency. My guess is it won't change much.
ShoeDLG is offline Find More Posts by ShoeDLG
Reply With Quote
Old Mar 01, 2013, 05:40 PM
A witty saying proves nothing.
kcaldwel's Avatar
Joined Jan 2007
3,565 Posts
I don't think you will see the 3D wake effects using VLM.

This is quite a good paper that looks at the errors due to 3D wakes that are not captured in most modelling:

http://ntrs.nasa.gov/archive/nasa/ca...1996034550.pdf

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Mar 02, 2013, 01:45 AM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
Kevin,

I think it depends what you're looking for, and what you're calling 3D wake effects. A VLM will predict that you can achieve a reduction in induced drag with a winglet. This reduction is essentially a 3D wake effect. In the same way, a VLM is capable of capturing changes in induced drag for configurations with significant dihedral/anhedral.

A VLM obviously has errors associated with it that you can only address by going to higher order panel methods and implementing wake relaxation schemes. That said, there's nothing in Steve's paper that suggests a VLM is insensitive to the 3D wake effects associated with non-planar geometries. Will they under-predict or over-predict certain features? Absolutely, but to the extent that they accurately characterize the distribution of vorticity in the wake, they can offer valid insight into the induced drag.

For example, the drag results in the work you referenced here:
Quote:
Originally Posted by kcaldwel View Post
are based entirely on a Vortex Lattice Method using a streamwise (unrelaxed) wake. VLM's do have significant limits, but that isn't to say they can't offer useful results if you stay within those limits.
ShoeDLG is offline Find More Posts by ShoeDLG
Last edited by ShoeDLG; Mar 02, 2013 at 02:03 AM.
Reply With Quote
Old Mar 02, 2013, 08:51 AM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
Kevin,

This discussion motivated me to implement a Trefftz plane drag analysis. The plot below shows the span efficiency factor as a function of AOA for three configurations:

- A "Dihedral/Anhedral" configuration (shown in the images in post #18)
- A "Dihedral-Only" configuration with the same planform, but all anhedral "reflected" into dihedral (shown beak-on in the second image below)
- A "Planar" configuration with the same planform, but no dihedral or anhedral

I think the span efficiency factors are plausible given the non-planar shape of the wake. I'll have to play around with this some more. Interestingly, the dihedral-only configuration is perhaps a little exaggerated, but not too different from what I would expect a Turkey Vulture to look like.

I don't have the code set up to easily compare lift distributions, but the third image shows the dihedral/anhedral lift distribution (clc) at 20 deg AOA compared to the dihedral-only at the same AOA. They're not too different, but the dihedral-only carries slightly more lift outboard.
ShoeDLG is offline Find More Posts by ShoeDLG
Last edited by ShoeDLG; Mar 02, 2013 at 09:41 AM.
Reply With Quote
Old Mar 02, 2013, 04:04 PM
A witty saying proves nothing.
kcaldwel's Avatar
Joined Jan 2007
3,565 Posts
I still think VLM is missing many of the effects of planform shape, both 2D and 3D. "The Split-Wing analysed by Smith did show an increase of e of some 11% in experiment and free force vortex sheet CFD (which translates to 11% reduction of induced drag), about 6% more than calulated using the linear Prandtl-Munk model."

The optimum planform for a planar wing in the Smith paper, and a couple of others using high-order panel analysis, does not result in an elliptical lift distribution. The span efficiencies are better than elliptical for some odd looking planform shapes and lift distributions.

VLM also neglects viscous effects, which at the Re of a seagull wing are going to be very important. I don't have access to the paper referred to in this quote:

"There is some evidence that the optimal winglet dihedral is strongly influenced by viscous effects. For example, Gerontakos and Lee40 varied wingtip dihedral in an experimental investigation and found that the winglet-down case produced lower induced drag than the corresponding winglet-up case; however, they noted that there was an order of magnitude discrepancy between the induced drag predicted by lifting-line theory and the experimental results obtained using the Maskell wake survey method."

http://homepages.rpi.edu/~hickej2/pa...-2008-5807.pdf

I think the simpler analysis tools like lifting line and VLM are missing a lot of the finer effects of planform and cambered spans, as well as the viscous effects of low Re. This could lead designs in the wrong direction.

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Mar 04, 2013, 09:25 AM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
I think it's useful to keep in mind just how challenging it is to accurately compute the span efficiency factor, e.

- An airplane's maximum Lift to Drag ratio is only half of its Lift to Induced Drag ratio at L/D_max.

- For an airplane with a modest maximum L/D of 20:1, the Induced Drag is only 1/40th of the Lift at L/D_max.

A 1% change in span efficiency near L/D_max for a 20:1 airplane corresponds to a change in Induced Drag of only about 0.00025 times the Lift force. In other words, in order to resolve a 1% change in span efficiency, you have to be able to resolve the drag to within 3/10000 of the Lift force.

I think its amazing that VLMs have proven to be accurate as they have.
ShoeDLG is offline Find More Posts by ShoeDLG
Reply With Quote
Old Mar 11, 2013, 02:37 PM
A witty saying proves nothing.
kcaldwel's Avatar
Joined Jan 2007
3,565 Posts
I'm not too concerned about the absolute accuracy of the wing efficiency factor, although within 5% would be nice for a change. It seems VLM doesn't sort planform shape efficiencies properly relative to one another.

Does your VLM rank the planform shapes from this wind tunnel test correctly? XFLR5 (VLM, LL, or 3D panel) doesn't. The differences in the wind tunnel testing seem large enough.

http://ntrs.nasa.gov/archive/nasa/ca...1994012101.pdf

I hope your VLM does, because it would be nice to have a tool that indicated the right direction to go on planform shapes for at least planar wings. I see people optimizing planform shape for elliptical lift distributions to 1mm accuracy in XFLR5, when it is missing 1/4 sweep and trailing edge shape effects completely.

Kevin
kcaldwel is offline Find More Posts by kcaldwel
Reply With Quote
Old Mar 11, 2013, 04:49 PM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
Kevin,

My VLM doesn't sort them the same way. My code showed all three elliptical planforms in a virtual tie with respect to span efficiency factor (all within 0.1% of each other). My code sorted them by the amount of "tip shear" (with the rectangular wing last): C - B - A - D

According to the paper, the elliptical planforms are within 4% of each other and sort: B - A - C - D.

The paper uses a very indirect technique for measuring e:

e = (pi * AR * dC_D/dC_L^2)^-1

This technique is extremely prone to capturing lift-dependent profile drag effects as induced drag effects. The only reliable way to sort the span efficiencies out experimentally would be to do wake surveys (there's no way we would have found a span efficiency near 1.1 for the split-tip using this technique). I think this technique penalizes planform C for 3D boundary layer effects unrelated to the span efficiency.

I'm not suggesting my code is appropriate for studying the difference between these planforms (it's not), but I also don't think the paper does a credible job of calculating span efficiency based on induced drag (I suspect the "actual" span efficiencies are well inside of 4% of each other for the AOA's considered) . Bottom line: I think the best you could hope a Vortex Lattice code would tell you is that the induced drag differences between these planforms is small. I think my code is remarkably reliable in that respect.

The numbers in parentheses in the images below are for the experimental results
ShoeDLG is offline Find More Posts by ShoeDLG
Last edited by ShoeDLG; Mar 11, 2013 at 04:55 PM.
Reply With Quote
Old Mar 11, 2013, 05:15 PM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
Just for yucks...

I don't show e continuing to increase with tip shear.
ShoeDLG is offline Find More Posts by ShoeDLG
Reply With Quote
Old Mar 11, 2013, 06:48 PM
Grad student in aeronautics
United States, GA, Atlanta
Joined Oct 2010
440 Posts
Shoe, do you use straight wakes? If so, are they body-fixed or straight in the freestream direction?
DPATE is offline Find More Posts by DPATE
Reply With Quote
Old Mar 12, 2013, 12:50 AM
Registered User
ShoeDLG's Avatar
Germany, BW, Stuttgart
Joined Mar 2012
757 Posts
DPATE,

Straight wakes in the freestream direction (drag free). I've pondered implementing a wake relaxation scheme, but that seems overkill given the inherent limitations in the method. Wake interaction with tail surfaces is always a concern for VLMs, and it would be interesting to see if relaxing the wake makes that better or worse. The code is designed to run on an iPad, so relaxing the wake won't be quick.
ShoeDLG is offline Find More Posts by ShoeDLG
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Just a thought about a hotwire cutter... e-flyer Foamies (Kits) 3 Jul 16, 2003 07:30 AM
Just a good R/C Story I thought I'd share.. Tram Electric Plane Talk 30 May 14, 2003 01:00 PM
Just a technical thought... Hopperfly22 Electric Ducted Fan Jet Talk 1 Sep 13, 2002 08:26 AM
Just when you thought it was safe to go back to the park... Martin Hunter Parkflyers 13 Jul 19, 2002 07:18 PM