Tandem wing for high efficiency? Case Proteus

Discussion in 'Aircraft Design / Aerodynamics / New Technology' started by karoliina.t.salminen, Aug 13, 2011.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Sep 14, 2011 #141

    Rick McWilliams

    Rick McWilliams

    Rick McWilliams

    Well-Known Member

    Joined:
    Feb 27, 2011
    Messages:
    274
    Likes Received:
    30
    Location:
    Santa Paula CA
    The canard stall in a 3LS configuration greatly increases the static margin and adds a pitch down moment. A large elevator may be able to over-power this effect. The elevator stick force will rise for sure. Many years ago I designed a 3LS airplane with a Y tail and pusher propeller. The canard flap extended with wing flaps to produce no trim change.

    The deep stall problem is usually associated with aircraft with very large fuselage, and a highly swept inboard main wing. A T tail may just place the horizontal stabilizer deep within the wing wake. The DC9 super 80 did a spectacular demonstration of this during flight tests.

    Should we design a little lift bump in the wing of a conventional airplane to be cancelled by the negative lift of the tail?
     
    Last edited: Sep 14, 2011
  2. Sep 14, 2011 #142

    bmcj

    bmcj

    bmcj

    Well-Known Member HBA Supporter

    Joined:
    Apr 10, 2007
    Messages:
    12,974
    Likes Received:
    4,919
    Location:
    Fresno, California
    Why is stalling the canard and wing any different than stalling the wing on a conventional aircraft? Are you referring strictly to a T-tail config where there may more blanketed area for the tail to fall in?
     
  3. Sep 14, 2011 #143

    Denis

    Denis

    Denis

    Well-Known Member

    Joined:
    Feb 24, 2009
    Messages:
    304
    Likes Received:
    4
    Location:
    Lviv, Ukraine
    The simultaneous stalling of the wing and the canard results in a pronounced kink (a "spoon") in the pitch moment curve. This kink starts when the stall cowers more than half of the canard area and its Clmax is reached. The canard Cl then decreases with further increase in the AoA, so the nose-down moment previously produced by the canard disappears. Nothing prevents the main wing from further increase in its AoA deep into the stall region. If the elevator just barely overcomes the canard diving moiment at its Clmax point, the further nose-up rotation develops as an avalanche. It may be possible to limit the elevator travel to prevent entering the canard Clmax point at stick pulled to the stop, but the vertical gust or inertial rotation after sudden elevator deflection may still do the same.

    I've mentioned the pitch lock particularly considering the P-180 plane with the Tee tail. The above described kink in the pitch moment accelerates the rotation movement and makes this phenomenon more probable in real flight conditions as compared with the conventional T-tailed planes with back-swept wings.

    By the way, the back swept wing features very the same pitch kink or spoon, but the wing and canard pair does it much stronger than the swept wings commonly used in jets.

    A 3LS plane with low-mounted horisontal tailplane will not lock in pitch, but it still will experience noncontrollable pitch increase beyond the canard stall, and this exteme rotation will lead to the flat spin, especially if the stall occurs at full power. The longer nose in front of the main wing will be here a spin-stimulating factor.

    A small canard in a 3LS configuration possibly will not generate much kink in the pitch moment, but such canard may be useless for bringing the flight CG range enough forward.

    On the constructive interference in a conventional layout:

    This effect differs from simple modification of the lift distribution. As either negative or positive lift of the conventional tail normally is within few percents of the wing lift, and the tailplane span is much less than the wing span, it does not appreciably influence the whole aircraft span efficiency. Here the primary source of a constructive interference is the downwash velocity field itself. The tail with a negative lift redirects some portion of the downwash up again, thus reducing the energy dissipated in the wake. If one looks at the Prandtl formula for the induced drag of a biplane, it is evident that if one of the two lifts is negative, the interference member is negative too and subtracts from the total induced drag.
     
  4. Sep 15, 2011 #144

    Rick McWilliams

    Rick McWilliams

    Rick McWilliams

    Well-Known Member

    Joined:
    Feb 27, 2011
    Messages:
    274
    Likes Received:
    30
    Location:
    Santa Paula CA
    It is easy to design a 3LS airplane for canard stall ahead of main wing stall. It takes a very large elevator to go beyond canard stall.

    You must have some special design criteria that I do not know.
     
  5. Sep 15, 2011 #145

    Denis

    Denis

    Denis

    Well-Known Member

    Joined:
    Feb 24, 2009
    Messages:
    304
    Likes Received:
    4
    Location:
    Lviv, Ukraine
    As I've described, the elevator easely overrides the canard in Piaggio P-180 and an additional measure was taken to limit the angle of attack, the ventral fins. The dimensions for the canrad and the tail stabilized werechosen to meet the following design goals:

    1. The entire useful CG range should be shifted forward sufficiently to place the whole cabin ahead of the wing spar box, the latter penetrates the fuselage in the middle position.
    2. The trim drag at cruise should be minimized.
    3. The Clmax should be high enough.

    The resulting elevator has quite conventional dimensions, just like those in the existing T-tailed planes. One can note that the 3LS configuration had no positive impact on the P-180 performance, rather the opposite is true.
     
  6. Sep 15, 2011 #146

    Denis

    Denis

    Denis

    Well-Known Member

    Joined:
    Feb 24, 2009
    Messages:
    304
    Likes Received:
    4
    Location:
    Lviv, Ukraine
    SmartCockpit - Airline training guides, Aviation, Operations, Safety

    This is the link to the detailed description of the P-180 Avanti.

    Vs0=97KIAS, Clmax based on this indicated stall speed and total area of the main wing and the canard, reveals that in the landing configuration Clmax=1.93

    The corrected stall speed is most likely even higher, due to typical behavior of the pitot error with the AoA increase, at list it is expected to be the same.

    Certainly such Clmax is too low against what one could expect from so elaborate set of flaps.

    Moreover, one can expect this or even higher (up to 2.0) value for a clean wing operting in similar conditions on a conventional well-designed twin turboprop.

    If one accepts the takeoff anf landing distances of the Avanti (and they are definitely acceptable for the typical missions of this plane), one can not only amputate tah **** canard, but considerably shrink the wing area and tailplanes. The total wetted surface could be reduced by about 10%. This is not a small figure for such airframe.

    Here is avery interesting comparative analysis:

    http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20100021005_2010021539.pdf

    Note that the total are of the wing and tail surfaces of this proposed 20 seat regional airliner is just equal to the total area of all three surfaces of the P-180, the same is true for overall length. Look at P.334 for analysis of airframe configurations.
     
  7. Sep 15, 2011 #147

    Rick McWilliams

    Rick McWilliams

    Rick McWilliams

    Well-Known Member

    Joined:
    Feb 27, 2011
    Messages:
    274
    Likes Received:
    30
    Location:
    Santa Paula CA
    What similar twin turboprops have better cruise perfomance than the Avanti?

    Many years ago I did a trade study of 20 different aircraft configurations. I used CLmax / Cdcruise as the figure of merit. Conventional airplanes were essentially tied with tandem and three wing configuration. The better three wing airplanes had small canards, large main wing and a small tail. All of the configurations benefitted from long length. Structure weight was not included. I consider the 3 wing option as a way to move the aerodynamic center and center of gravity where the variable loads are located.
     
  8. Sep 15, 2011 #148

    Denis

    Denis

    Denis

    Well-Known Member

    Joined:
    Feb 24, 2009
    Messages:
    304
    Likes Received:
    4
    Location:
    Lviv, Ukraine
    It is a pity there are no other small turboprops built to the vary same design requirements as the Avanti. P-180 is also the youngest such development in chronological order. indeed the most important factor here is the superior altitude performance of the PT6A-66 engine. The equivelent drag area of P-180 probably may me even greater and span efficienncy not better than those of Turbo Commander AC-1000. Also the P-180 airframe is designed for higher dynamic pressure and critical M number from the beginning, while the AC1000 is a descendant of a piston Twin Commander. But the comparison of the aerodynamical characteristics of these two planes can be surprising and not in favor of P-180.

    The balanced configuration analysis of a small twin with considerations for structural weight and range to payload relationship shows the benefits of a relatively thick and short fuselage against the narrow and long one as soon as the cabin dimensions are practical for the comfortable seating of the specified number of passengers. The above NASA report clearly reveals how such shape reduces the drag of the fuselage. Note that the wing of such plane has a relatively minor contribution into the equvalent drag area. The conventional wing and tail layout also proves its superiority over everything other, by the best balance between the structural weight and total induced drag.
     
  9. Sep 15, 2011 #149

    Rick McWilliams

    Rick McWilliams

    Rick McWilliams

    Well-Known Member

    Joined:
    Feb 27, 2011
    Messages:
    274
    Likes Received:
    30
    Location:
    Santa Paula CA
    Fuselage and engine nacalle drag is usually much greater than the wings. The problem with fuselages is people.

    I think that there is an interesting trade-off between advanced flap systems and increased wing area. A good laminar flow section can have a cruise drag Cd of 0.004. An advanced flap on a high lift section can have a maximum landing Cl of 3.2. The compromises are complicated.
     
  10. Sep 15, 2011 #150

    Denis

    Denis

    Denis

    Well-Known Member

    Joined:
    Feb 24, 2009
    Messages:
    304
    Likes Received:
    4
    Location:
    Lviv, Ukraine
    Thge starting point is the maximal L/D Cl. the Clmax should be at least 1.69 x this Cl, corresponding to minimal safety margin above sall speed, 1.3Vs1.
    From other hand, the maximal reasonable cruise speed should not exceed 1.41 x V for max L/D, at this speed the L/D is sitll 80% of the maximal value, and at higher speeds it falls quickly. So the minimal profile drag should be ideally maintained down to the 0.5Cl for maximal L/D. I tis important taht the maximal L/D Cl gradually increases as we shrink the wing chord while maintaining the constant span, so the whole optiomal drag polar of the airfoil should move towards higher Cl values.Assume clean Clmax=2.0, which actually possible at reasonably high Re, one obtains the low drag bucket Cl range from 1.0 down to 0.5 there are turbulent airfoils, which can provide Cd<0.006 at this Cl range, and this wing will be so small in respect to the fuselage wetted area, that the profile drag will contribute to just about 1/4 of the plane Cd0. If we imagine some airfoil, which will have Cd=0.004 within the same Cl range, the resulting Cd0 will be reduced by 8% and the fast cruise airespeed will increase les strhan by 3%, such small increment will be at the edge of test accuracy.
    But it is difiicult to imagine such high-lift airfoil with so low Cdmin. Moct likely the Clmax will considerably reduce with tsuch improvement in Cdmin and we'll need to increase wing area. Therefore the equivalent drag area will not change altogether, but the larger wing and correspondingly larger tail will result in increased airframe weight. Finally, here we discussed the hypothetical airplane witl maximal L/D=20.8, fast cruise L/D=16.7 and Vcrmax/Vs0=2.45 at sea level.
     
  11. Sep 16, 2011 #151

    autoreply

    autoreply

    autoreply

    Moderator

    Joined:
    Jul 8, 2009
    Messages:
    10,732
    Likes Received:
    2,542
    Location:
    Rotterdam, Netherlands
    Indeed. Doing the math for a (much slower) design, I concluded that advanced slotted flaps or plain flaps were on par. While the slotted flaps achieved a Cl of 2.7 for the whole wing, while the plain flaps didn't come further as 1.8, the plain flaps still had lower drag for a given stallspeed, since they could be laminar and thus had lower drag, even with it's almost 50% larger area. Not to mention the lower cost and complexity. Only real drawback is the bumpier ride in turbulence and the lower Va and the like.

    @ Denis, a highly swept wing, preferably forward, would be a very interesting option for a pusher turboprop.
     
  12. Sep 19, 2011 #152

    Himat

    Himat

    Himat

    Well-Known Member

    Joined:
    May 5, 2011
    Messages:
    2,811
    Likes Received:
    654
    Location:
    Norway
    Tread drifted away, but interesting.

    Anyway, I do find it difficult to follow as to me some asumtions are left out.
    Why L/D=20.8, fast cruise L/D=16,7 and Vcrmax/Vs0=2,45 at sea level?

    I'm not convinced that a larger wing(area) do have a higher weight than a small. I do think Barnaby Weinfan found/argued that the Dyke Delta was a more efficient transport vehicle than a modern composite airplane due to lower structural weight. The composite airplane had a much better L/D, but the Dyke Delta outperformed it on a better useful load/structural weight.
     
  13. Sep 20, 2011 #153

    karoliina.t.salminen

    karoliina.t.salminen

    karoliina.t.salminen

    Well-Known Member

    Joined:
    Oct 7, 2008
    Messages:
    407
    Likes Received:
    80
    Location:
    Finland
    When the aircraft design comes from aerodynamic speculation to the structural design, it seems to become apparent that some compromises have to be made, otherwise the penalties from the structural weight will outweight the aerodynamic advantages from a better aerodynamic design. Bottom line is that additional lift causes additional induced drag. More weight means needing more lift and the planes becomes more draggy. Best compromise on optimized design is both less draggy design and also has lower structural weight and better payload fraction. That's a challenging dilemma which I need to start thinking about next.
     
  14. Sep 20, 2011 #154

    Autodidact

    Autodidact

    Autodidact

    Well-Known Member

    Joined:
    Oct 21, 2009
    Messages:
    4,513
    Likes Received:
    799
    Location:
    Oklahoma
    It would be nice to have an algorithm that would calculate the structural weight and performance of a design that had a fixed engine and payload requirement as it was "morphed" digitally from long thin small wings to low AR large area wings, i.e., from say, a Diamond to a Facetmobile type craft or anything else in between. It should be possible using equations like the Rockwell equations for weights that Raymer provides, I wouldn't want to be the one who writes that program, though!
     
  15. Sep 20, 2011 #155

    autoreply

    autoreply

    autoreply

    Moderator

    Joined:
    Jul 8, 2009
    Messages:
    10,732
    Likes Received:
    2,542
    Location:
    Rotterdam, Netherlands
    If you're satisfied with only calculating spar weight it's pretty straightforward to write. For an aspect ratio of 10, I get a spar weight for example that's just over 1% of MTOW, while an aspect ratio of 8 "saves" me 0.06% of MTOW.
    Most of the other weight is per area of skin. Problem though is that for large cross sections, like on a Facetmobile you need additional stiffeners to keep the skin from buckling, which in the end might make a wing with a slightly higher aspect ratio lighter. I ran into a similar problem for my design, when my aspect ratio was at or below 7.5, the skin would buckle, forcing me to a thicker foam core, completely correcting the pound of so I might have saved on the lighter spar.

    If you have an Excel-sheet or program that does the math, it's fairly easy to compare a few different wings with each other.
     
  16. Sep 20, 2011 #156

    karoliina.t.salminen

    karoliina.t.salminen

    karoliina.t.salminen

    Well-Known Member

    Joined:
    Oct 7, 2008
    Messages:
    407
    Likes Received:
    80
    Location:
    Finland
    I may write that program actually because I think I will need it myself on next phase.
     
  17. Sep 20, 2011 #157

    karoliina.t.salminen

    karoliina.t.salminen

    karoliina.t.salminen

    Well-Known Member

    Joined:
    Oct 7, 2008
    Messages:
    407
    Likes Received:
    80
    Location:
    Finland
    This is very interesting insight. For example for Diamond DA40 that would mean that the spar would be weighting 11.5 kg (which is quite nothing compared to the weight of the rest of the aircraft). This makes me think that is it so that the spar weight exaggerated in the usual thinking when the weight is somewhere else.

    Can I then interpret this this way: if AR 10 spar weight 1% of MTOW, the AR 20 spar will weight 4% of MTOW if all else being equal? So by doubling the length, is the rule of thumb that the spar weight becomes four times bigger?

    Auroreply; finding the buckling points of your skin - do you use some FEM software? If yes, what? Can you calculate that by hand? What books should I get to write software (myself, because I a software engineer, I like to solve things by writing software because it is easy for me) that does that? I have Mac and Linux machines and no Windows, unable to run Windows software.
     
  18. Sep 20, 2011 #158

    autoreply

    autoreply

    autoreply

    Moderator

    Joined:
    Jul 8, 2009
    Messages:
    10,732
    Likes Received:
    2,542
    Location:
    Rotterdam, Netherlands
    Yes, I think so. Maybe in the old aluminium days, an stubby aspect ratio was beneficial, but in composites there's really not that much of a weightwise penalty for longer wings. Of the 5 best-selling composite ships, none has an aspect ratio of under 9 if I recall correctly.
    Yeah, sort of. But the problem isn't the spar(weight), which is really only a minor issue. Even with an aspect ratio of 57 (!) your spar weight is probably less then 10% of MTOW. The real issue are all the tiny details which become exponentially more problematic with a higher AR, while the decrease in drag exponentially decreases.

    Things like control surface clearance, aileron/flap bending, push/pull tube bending, flutter, twist under load and so on are virtually irrelevant if you have a stubby wing, but when you go to sailplane-like AR's, those problems explode in magnitude. At the same time, the possible gain decreases, going from AR=10 to AR=18 gives you really only a minimal increase in climb or cruise.
    Honestly, the difference between AR=9 and AR=11 is minimal. Just pick one that feels good and let other factors (for me that was maximum chord length for the mold covers..) determine the final AR, don't "over-optimize" it.
    I basically made a cross-section, analytically determined all the forces and converted that to Java where all the variables are processed, but admittedly this only works if you have a constant cross section shape (not dimensions). Judging the highest stress by analytically locating it, I used that as a cut-off for "too-weak" outcomes. Picking a lot of things cleverly (skins, spar caps and shear webs are discrete numbers for example) you get pretty good results with minimal (but still considerable) work.
     
  19. Sep 20, 2011 #159

    Autodidact

    Autodidact

    Autodidact

    Well-Known Member

    Joined:
    Oct 21, 2009
    Messages:
    4,513
    Likes Received:
    799
    Location:
    Oklahoma
    Here is an excellent book - not many formulae, but lots of insight, for example the reason (one of them) that spoilers were used for roll control on the B-52 was that the wings were so flexible and spoilers impart less of a torsional moment helping to avoid aileron reversal: Amazon.com: Airplane Stability and Control: A History of the Technologies that Made Aviation Possible (Cambridge Aerospace Series) (9780521021289): Malcolm J. Abzug, E. Eugene Larrabee: Books

    What would be cool is if you could have all of the analysis tools parametric to a drawing, then you could stretch the drawing this way and that and watch the numbers change as you do, it would probably take a lot of money both for software and hardware.
     
  20. Sep 20, 2011 #160

    Himat

    Himat

    Himat

    Well-Known Member

    Joined:
    May 5, 2011
    Messages:
    2,811
    Likes Received:
    654
    Location:
    Norway
    Yes, it would be very nice to have such an algorithm in a computer program.
    One major difficulty I can think of is how to handle transitions between different fabrication metodes and materials as the AR changes.
    Building a wing with an AR of 20 composite is probably the way to do it.
    Going down to AR of 10 there is a choise between composite, wood, and metal, with none much lighter than the other.
    Down at AR close to 2 or 3, tube and fabric might give the lightes structure.
    Then, how to handle this in computer code?
     

Share This Page



arrow_white