Estimating washout from Airfoil Transitions?

[SoarNeck]

Hi all,

How would I go about estimating the effective washout that would
result from a transitional airfoil series? Programs like LiftRoll
don't account for this possibility.

I guess I could work from polars to determine the AoA where one
airfoil generates the same lift as the other, but I was curious if
there were any rules of thumb that would be less time consuming.

Thanks,
Adam

[Ollie]

Because the slope of the coefficient of lift versus angle of attack polar is the same for the various airfoils, all you have to do is campare their zero lift angles of attack (ZLAOA).

For example, if the mean line camber decreases as the airfoils transition from root to tip, the root airfoil will have a more negative ZLAOA then the tip airfoil and the wing will have washout equal to the difference in the ZLAOA of the airfoils.

As you know, this type of washout is known as aerodynamic washout to distinguish it from geometric washout. The total effective washout is the sum of the aerodynamic and geometric washouts.

To find which part of the wing stalls first, it is also necessary to compare the maximum lift coefficients of the root and tip airfoils as the wing's AOA increases. Some very effective wing designs employ a reverse transition with the root airfoil having less camber than the tip airfoil and compensating geometric washout to gain tip stall margin because, in this particular case, the tip airfoil stalled at a higher lift coefficient than the root airfoil.

[Sparky Paul]

Some years back Dave Jones did an excellent article in QFI on how to do this. Essentially what Ollie detailed.

[SoarNeck]

Thanks Ollie, very helpful, that's exactly what I was looking for.

Paul - I seem to remember reading that article, I'll have to go digging. Thanks for jogging my memory.

[Ollie]

Before I read the Dave Jones article I had been designing wings with the SD7037 airfoil, tripple taper and no washout. The wings were very efficient up to stall but the SD7037 has a very abrupt stall and was very unforgiving of my piloting errors.

After reading the last of three articles by Jones on the family of SA7035, -36, -38 and SD7037, I used the details of one of Jones wing designs. It incorporated the SA7036 at the root, SD7037 at the middle of the semispan and SA7038 at the tip with geometric washout that compensated for the aerodynamic washin. The result was a wing that had an efficient lift distribution at all angles of attack and a soft stall and quick recovery despite the abrupt stall characteristics of the individual airfoils. This was a big step forward in my understanding of wing design and for my plane's handling characteristics.

My understanding and appreciation of great wing design took another step forward when I studied Dr. Mark Drela's AG 24,-5,-6 and-7 airfoils and his composite Bubble Dancer wing taper and twist. The the plane that I built, based on Dr. Drela's wing design, was substantially better in efficiency and handling than anything I have flown in over 30 years of R/C sailplanes.

[SoarNeck]

I do remember that series, because I remember the transitional Selig 70XX wing.

Have you had a chance to review the 120/130" Aegea yet, Ollie? A fellow club member and I were discussing the airfoil choice, and he eventually decided to choose the MH32 for his model instead of the Drela series because he thought it would launch better (given that the MH32 has a better max Cl value, and at launch the airfoils are operating at max Cl).

I'm tempted to try building an Aegea, but the airfoil issue has been bothering me. The truth of the matter is that I'd really like to try molding a wing, and so I want to be really sure of what I'm doing before I commit to anything.

[Ollie]

There is no sense in comparing the maximum coefficient of lift of the MH32 in its unflapped state when you are going to launch with 15 degrees of flap and maybe 10 degrees of down aileron too. Has you friend used XFoil to analyze the flapped performance of the MH32? If he hasn't I wouldn't pay much attention to his reasoning.

In general, an airfoil or series of transitional airfoils have to be integrated with the wing's planform, sweep and twist for best results. Comparisons of performance should be made on the basis of the whole aircraft, not just the airfoil in isolation.

I have faith that Dr. Drela's wing design will be far superior to what any of us would design, myself most of all.

The Aegea wing is rather complex structurally. For that reason and for that reason only, I think a simpler wing, with a thicker airfoil and a stressed skin instead of an imbedded spar, would be a better choice for your first vacuum bagged wing construction. I think it is better to learn to walk before you try to learn to run.

[SoarNeck]

Hi Ollie,

I'm not sure what his conditions were for evaluating the max CL condition, but he's designed airfoils before so I'd imagine he took flap settings into account.

Actually, I was being literal when I said "molded" - I really meant I wanted to start hollow molding models. I've built 10 or 15 bagged models before, so am quite familiar with the techniques there. I understand the structural design behind the Aegea (even already have a fuselage mold done), I'm just trying to push myself to the point where I also understand the whys behind the aerodynamic design as well. Give me a couple of years :)

My first efforts at pulling it all together (though I think I'm a couple of revs beyond what I uploaded to the website by now):

http://www.soarcalgary.com/Extras/Articles/Sailplane%20Design%20Spreadsheet/index_html

I'm looking to have a repeatable form that allows me to play with structure without throwing aerodynamic changes to the mix...and to learn something new. I'd like to take a couple of years and mold a series of models that would be competitive in F3J, with the intention of going out for the Cdn team in 2005.

I may just stick with an F5D mold for now since the scale is much more manageable, but it's good to have long-term goals. The aerodynamic concepts are equally valid as well.

[Ollie]

SoarNeck,
Sorry I misinterpreted your question.
Do you plan to build a wing plug and make molds from it or do you plan to have the molds CNC machined and hand polish them?

[SoarNeck]

I might split the difference and get a plug cut by computer, it depends on how many favours I can call in from various people ;) I don't have the funds to get a negative mold cut, but I've done enough molds from positives to have a good feel for how that work.

After your discussion about Xfoil I took the opportunity to order the new version of Profili that is out, and man alive that program is amazing! I hadn't used it since back when Profili was a glorified airfoil database, but now it does so much that I'm shocked he isn't charging $100 a licence for it. Even then I'd feel like I was getting a deal!

I'd been putting off learning Xfoil due to the rather steep learning curve that's involved with the program, but happily now it looks like I won't have to.

Now we just need Dr. Drela to release one of his pressure distribution -> airfoil generators, and we'll have all the design tools we ever need.

Oh the schemes that I'm dreaming of now...

Cheers,
Adam

[Ollie]

Adam,
I think the Lift Roll spreadsheet uses the latice panel method and will deal with the evaluation of flap deflection by generating a lift distribution. I don't think it will show any significant difference in lift distribution for similar flap deflections associated with different airfoils. The line drag and flap drag are so high compared with the drag associated with the rest of the airfoil that such differences are completely swamped out during towing.

[SoarNeck]

Hi Ollie,

I'd definitely be using Xfoil (Profili) rather than LiftRoll to do anything with flap deflections. I played with it a bit last night, and am very impressed. In the last two days I've relearned a lot of what I forgot over the years regarding aerodynamics. Thanks for your help.

If you get a minute, could you check out my other thread in this section regarding local Cl in a turn?

Thnx
Adam

[durone]

Now we just need Dr. Drela to release one of his pressure distribution -> airfoil generators, and we'll have all the design tools we ever need.

With the new Professional version of Profili ( see www.profili2.com (http://www.profili2.com) )
you can also recompute an airfoil from a modified speed distribuction (the MDES menu of XFoil).

Please write me info@profili2.com if you need more info about

[AntonL]

if the mean line camber decreases as the airfoils transition from root to tip, the root airfoil will have a more negative ZLAOA then the tip airfoil and the wing will have washout equal to the difference in the ZLAOA of the airfoils.

To find which part of the wing stalls first, it is also necessary to compare the maximum lift coefficients of the root and tip airfoils as the wing's AOA increases.


Leaving aside geometric washout for a moment, I have tried to test this in Profili by running polars on 3 versions of the same foil with various changes in max camber (either increasing or decreasing camber from root to tip) and it seems that the foil that has the most negative ZLAOA is not necessarily the foil that has the highest alpha at max Cl. I thought that the general rule of thumb was, with most foils, that if you reduced camber from root to tip, this would have the effect of shifting the Cl vs alpha curve to the right as you progress towards the tip (i.e., aerodynamic washout) and that this would result in somewhat higher alpha at max Cl for the tip compared with the root. Well, it seems that when you do this it doesn't (with the foils tested so far) necessarily increase the alpha at max Cl seen at the tip. Could I be doing something wrong here with these tests? Is decreasing max camber from root to tip not the right way to produce aerodynamic washout?

Anton

[mharms]

(Pardon my basic questions, I'm not quite digesting all of this great thread)

I'm investigating how I might help reduce tip stalling through aerodynamic washout (rather than geometric washout). I'm not sure I'm understanding things correctly. I am pretty new to this. Specifically, I'm looking at all this for swept micro flying wing design (2 inch root chord!). I don't know if the super low Re has any implications for tip stalling, so perhaps that can be ignored for the moment.

I've been using Profili to help visualize what people are proposing to help reduce tip stalling. Please consider the Cl vs alpha plot I've attached:

I figured that I should simulate the tip at a lower Re than the root to reflect the lower tip speed in a turn (valid?). The black curve represents the root foil in a turn and the red and blue curves represent some possible tip airfoils. The grey curve is just the root foil at the same Re as the tips for reference.

For preventing tip stall, would I want the tip airfoil to maintain Cl at a higher AoA than the root (example: red tip curve #1 vs black root curve at 13 degrees alpha)? Seems to me, this is only relevant at really high AoA.

Or, do I want the tip airfoil to have a more positve AoA at the Cl=0 point (example: blue tip curve #2 vs. black root curve, which shows about 1 degree alpha difference)?

Or is the blue tip curve the opposite of what I want? For a given (mid-range) AoA, wouldn't I want the tip to have higher lift? Now I'm confused.

Maybe I need a basic walk-through on what happens on the Cl vs alpha plot during a turn or a tip stall. For example: I'm not clear on the AoA of the tip vs the root in a turn. Perhaps I also need to learn how to use the spanwise lift calculators I see mentioned.

Thanks for your patience,
Mark

[Sparky Paul]

Those curves look too good to be real. Those Res are in the insect range, where it's more swimming than flying.
Looking at the stall peaks, there's about a 4 degree range between root stall and tip stall. It wouldn't be a bad idea to incorporate something like that on the wing.
This real low stuff is in the mud for most wind tunnel information, so you may have to experiment yourself, with the data as a guide, but not an achievable situation.

[mharms]

Sparky Paul,

Yeah, I know calcs down at this low of Re may not reflect reality too well, but I was mainly trying to understand the curves as they relate to tip stalling.

Looking at the stall peaks, there's about a 4 degree range between root stall and tip stall.

So when you say there is a 4 degree range between root and tip stall, how are you seeing that on the curves? If you just mean the AoA delta between the Cl peaks, then I see maybe 2 degrees delta between red and grey curves. Like I said, I'm new at this.

BTW, the only way I got these curves to look this good is by the use of turbulators in the sim, one on the top, one on the bottom. Without the turbulators, the curves look horrible.

Those Res are in the insect range, where it's more swimming than flying.

I'm using 30 to 40kph as an estimated velocity, with a 5cm chord, so that works out to an Re around 30K to 40K. I'd say I'm in the really large insect range (probably more like small bird). For sure not the kind of insect I'd want to splat on my car windshield!

Mark

[Sparky Paul]

When your red section is beginning to stall, the black section has 5 or degrees left.. that's the washout effect. 5 degrees is a lot.
Some more doodling with thickness and turbulators could probably get a lower amount of washout, you don't need much.

[mharms]

Sparky Paul,

Sounds like I had things backwards. If I understand you correctly, then (to prevent tip stalls) it would be best for the red curve to be the root and the black curve to be the tip?

I had been thinking it was only the peak points that mattered. I saw the black peak being at lower alpha than the red peak and thought that meant the black curve was stalling first (and, hence, should be used as the root). I had not considered the reduction of slope of the red curve as important, but I guess it is just the start of a less abrupt stall.

So whichever curve starts reducing slope first (at lowest alpha) is the one I should chose for the root (in order to prevent tip stalls)?

Any comments on the blue curve (compared to the black)? I thought I had read that offsetting the Cl=0 point was a way to get aerodynamic washout?

Mark

[Sparky Paul]

Mark, washout is having less angle at the tip than the root. As the curve shows, when the red line is breaking at stall, the black line shows there's still some angle left before it stalls when the red line is stallling.
With the red section at the root, the black section at the tip, the plane will stall properly.
Too much washout washout makes the tip fly inefficently though, and can even create torsional moments that twist the tip down with increasing airspeed.
I don't think your plane will have this problem, but it can a consideration for higher performing items.