Wing sail questions

Discussion in 'Hydrodynamics and Aerodynamics' started by dustman, Jan 3, 2023.

  1. dustman
    Joined: Jun 2019
    Posts: 305
    Likes: 37, Points: 28
    Location: Tucson, AZ

    dustman Senior Member

    What's the optimum naca profile(chord to width) for a wing sail? I'm thinking 20-25%?

    A thicker foil would stall at a greater angle of attack than a thinner foil, would have a greater range of useable angles of attack, would be more forgiving?

    I thinner foil would allow you to sail closer to the wind?

    The thinner foil would have a better lift to drag ratio?

    How far along the chord should the mast be to balance the wing sail? Same rules as for balanced rudders? 15-20%?

    Are there any real disadvantages of a wing sail from an aerodynamics perspective?

    Since the wing sail is a more efficient lift device can you get away with a little smaller sail or a lower aspect ratio for similar performance?

    How much more efficient is the true foil than the standard sail, really?

    I noticed a lot of soft wing sails have a lot of creases, does this significantly affect their aerodynamics, and by using stiffer sail material could you minimize creasing?

    Is it more efficient to use a wing sail like a lift device or drag device downwind?
     
  2. dustman
    Joined: Jun 2019
    Posts: 305
    Likes: 37, Points: 28
    Location: Tucson, AZ

    dustman Senior Member

    Oh yes, and are wing sails better for light air?
     
  3. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    Is it a christmas wish list ? Santa Claus is already gone, you know ;)

    Here is me playing with Xfoil, around a NACA63012, turning it into a NACA63025, doing some calculations, assuming the flow has the viscosity of the air. Reynolds number is choosen where the turbulence becomes instable in real life. Top image is the boundary layer and pressure contour around the "thin" profile. Bottom image gives the same information, but for the "thick" profile. In the middle, the comparison between the two profiles.

    upload_2023-1-3_11-42-33.png

    How far along the chord should the mast be to balance the wing sail? Same rules as for balanced rudders? 15-20%?
    Really, it's depends on the plan view of the sailplane, ie of the surfaces of the wing sail and of the mast, together with the position of their geometrical center, as it is the case for balanced rudders.
    Are there any real disadvantages of a wing sail from an aerodynamics perspective?
    The wing sail is to have more stiffness than a traditionnal sail. So its control during gusts may be unforgivable.
    Since the wing sail is a more efficient lift device can you get away with a little smaller sail or a lower aspect ratio for similar performance?
    Yes, providing the fact that you do not decrease the aspect ratio nor the surface to a point where there would be no more advantages. Calculations are to be made.
    How much more efficient is the true foil than the standard sail, really?
    + Is it more efficient to use a wing sail like a lift device or drag device downwind?

    Answers are very particular to a given boat (stability,weight,surfaces ...). For instance, the efficient of a wingsail, compared with a standard sail, for a boat that sails faster than the true wind will not be the same as the one for an heavy boat stucked at 6kts.
    I noticed a lot of soft wing sails have a lot of creases, does this significantly affect their aerodynamics, and by using stiffer sail material could you minimize creasing?
    This is just an opinion, but even on a classical sail, creases effect seems to be very small, if these creases appear on a rightfully trimmed sail, of course. In very very very light wind, rigging pre-tension and sails tensions can be put to the minimum (for the mast not to fall on our nose), and surprinsgly, even with huge creases on the sails, your lift will be higher than the lift obtained by a sail with no crease at all. Of course, again, you must notice from that example that the wind conditions and the boat characteristics and speed must be taken into account to get a pertinent answer, in general.
     
  4. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    At null incidence
    upload_2023-1-3_13-16-42.png
     

    Attached Files:

  5. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    at 30° incidence

    upload_2023-1-3_13-31-24.png
     
  6. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    at 43° incidence

    upload_2023-1-3_17-11-9.png
     
  7. dustman
    Joined: Jun 2019
    Posts: 305
    Likes: 37, Points: 28
    Location: Tucson, AZ

    dustman Senior Member

    These simulations are like a late christmas gift, thank you!

    The reason I am asking all these question is because I thought of a way to build a wing sail pretty easily, significantly easier than the cambered quasi-junk rig I was set on before. Still would use junk sheeting and reefing.

    Anyway, I have a few observations/questions regarding what you posted.

    What wind speed would the reynolds number in this simulation be representative of?

    What do the terms Cp, Cm, and Ncr represent on the graphs? What do the lines represent?

    Am I seeing this right? The the thicker foil has greater lift than the thinner foil at 18 degrees, and a much better l/d ratio, and much lower drag? I would expect the thicker one to create more drag at lower angles of attack.

    It seems the thinner foil has greater lift at greater angles of attack than the thick one? That is the opposite of what I would expect.

    The thicker foil seems to maintain a better l/d ratio at all angles of attack presented, except for zero.

    It would have a square plan(for ease of construction) with high aspect ratio. Is the 15-20% range going to give me decent balance while still being controllable?
    That doesn't sound good. I imagine you meant to say unforgiving, rather than unforgivable? You mean to say that the wing sail won't twist off as much in gust?
    How much more efficient is the wing sail? 10%, 20%? I know, it depends... But if one had to make a generalization?
    It would be on a 24' ultralight catamaran with very high length/displacement. SA/disp of about 24.
    In certain circumstances the flow will stay attached longer with some creases? In my design, the only creases I can see happening are from sail twist, or from messing up the alignment of the sail components when building it.
     
  8. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    Then I should review the Reynolds used for the calculations. Please see the following comment

    The corresponding Apparent Wind Speed (AWA) is a little more than 10kts, assuming the E of the mainsail is around 10' (0.42 x LH).
    In the case of your ultralight catamaran, a Reynolds number higher than 2.0E6 will be a good design point, for you to attain "MACH1.0" @tws;TWA = 8.5kts;43degree
    TWS True Wind Speed
    TWA True Wind Angle

    Cp is the pressure coefficient. Pressure coefficient - Wikipedia https://en.wikipedia.org/wiki/Pressure_coefficient. Xhen the flow does not stall, the Cp on the extrado is negative, meaning that a depression occurs. On the intrado, the Cp is mainly positive. Flow tends to stall where the sign of the Cp changes.
    Cm is the moment coefficient. Aerodynamic Lift, Drag and Moment Coefficients https://aerotoolbox.com/lift-drag-moment-coefficient/. This moment is generated by the forces resulting from the pressure distribution on the airfoil, and is calculated relative to the quater-cord. A negative Cm gives a profile that has a "pitch down" tendency, which is what we expect from a stable airfoil, because it naturally compensate the loss of lift, due to variations of the incoming flow. But we don't want that pressure center to move too move aft too much when the angle of attack increase, to avoid any flutter phenomenon.
    Ncr is the critical exponent governing the transition of the flow regime from laminar to turbulent. What does the NCrit parameter indicate in a CFD analysis? https://aviation.stackexchange.com/questions/66175/what-does-the-ncrit-parameter-indicate-in-a-cfd-analysis. Below 9, the simulated flows tends to be more laminar, which is a situation that is not encounter in experiment, nor in our real life (on the water).

    The point calculated at 18degree is highly suspicious. Nevertheless, for the calculated angle, the thicker profile has indeed a better l/d ratio than the thinner. Indeed, the thicker one create probably more drag at low angles of attack (below 15° approximatively), but for higher angles of attack, the leading curvature of the thicker is more tolerant than the one of the thinner. It maintains the flow attached in the region where the NACA63 has been designed. It is no surprise, in fact, because the NACA63 has been designed for an optimum l/d at angle of attack below 15°. The scaling factor, applied on the profile, happens to be profitable for the NACA63 at higher angle of attack (again, do not take into account the point at 18degree for now, I will re-do calculations).

    Pushing the "thickening" of the profile would lead to something comparable with a circle, with has no lift at all. A usefull profile has to be designed for precise conditions. In that case, an adequate curvature on the thicker profile ( increasing the leading curvature, working on transition near Bmax...) would increase its lift coefficient.

    Yes. See comment above/

    Yes, I think so.

    My bad...Sorry. Yes, I meant unforgiving. Yes, the planform of the wing sails should take into account of a lower twist compared to a traditionnal sail. As a consequence, the apparent wind angle @Mast top can be significantly higher too, so good care should be taken for the structural design.

    A quick answer, after looking at the results and anticipating a true behaviour, I would say that the efficiency, in your case, could be more than 30%, beam reach. I'm working on more precise numbers, with to the global characteristics of the boat that you've given.

    In my example, my guess is that the creases act like stiffeners, maintaining a curvature to the sails, even when hardly no air is pushing on it. In very light wind, it's an advantage, compare with a flat sail without crease. As the wind speed increase, the crease would likely be the cause of an additional parasite drag. But remember that, below the total minimum drag, the main contribution is given by the drag induced by the lift of the sailplane itself. At the minimum drag of the sailplane, the parasite drag is equal to the induced drag. Minimizing the parasite drag due to the crease will certainly higher the speed where the drag of the sailplane is minimum, but if, as I suspect, the contribution of the crease to the parasite drag is small, then we are talking about increasing the total minimum drag speed of perhaps less than 0.5kts... In term of performance, this is nothing. Again, this reasoning should be considered under the hypothetis that the creases we are talking about are on a correctly trimmed sailplane.


    In the next post, I will review the calculations of the NACA63, adding some point to the curves, and working at a greater Reynolds number. Then, I will be able to put this in my VPP, to quantify how much efficient is this rough wing sail, compared to a more traditional one, varying the wind angle and the wind speed.
    As for now, I hope that these comments will give you the information you need. Cheers
     
  9. dustman
    Joined: Jun 2019
    Posts: 305
    Likes: 37, Points: 28
    Location: Tucson, AZ

    dustman Senior Member

    Wow, thanks for the thorough reply. Very helpful, though I still have a some questions.
    Sorry, not sure what E around 10'(.42 x LH) means?
    How do you determine what reynolds number to design to, and would this determine the appropriate aspect ratio?

    Do you think the rectangular sail planform is going to have a significant affect on its performance compared to an approximation of an elliptical planform?

    Here is all I can think to tell you about the plan for the catamaran as it is now:

    Length/beam: 24x16'

    Displacement: ~1000lbs

    Displacement/length: 16.5(Edit: it's actually 27.9)(assuming you calculate a catamaran by taking half the weight and the full waterline length)

    Sail area: 128ft2, divided between two sails in biplane configuration. Each sail would measure 4x16', for an aspect ratio of 4:1.

    Center of effort of sails: ~12' from waterline

    Sail area/displacement: 20.5 (I know this is different than the 24 i stated before. What i arrived at after doing more stability calculations)

    Keel and rudder area of 4.5ft2 total, which is 3.75% of sail area. Keel area 3ft2 divided by two keels at 1.5ft2 apiece. Rudder area 1.5ft2 divided between two rudders at .75ft2 apiece. Considering an aspect ratio between 2 and 4, leaning towards 2 for smaller draft.

    The stability formula i got off the Wharram website. rm formula.png According to this the sails should be reefed around 20mph wind speed.

    Wetted surface area: ~70ft2

    Looking forward to seeing your results. Thanks again for taking the time help me understand this.
     
    Last edited: Jan 10, 2023
  10. Erwan
    Joined: Oct 2005
    Posts: 460
    Likes: 28, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    A Wing sail!!!! That is an interesting project.
    For XFOIL, I would have the following suggestion:
    Above 6 knts true wind speed, the Boundary Layer of the wind above the water is turbulent, so your boundary layer on the sail/wingsail is not likely to be laminar, that is why it is reasonable to assume it will be turbulent.
    With XFOIL , you can assume a turbulent BL (boundary layer) choosing the position of the transition point at ie: 0.01 (1% from Leading edge) On this forum I noticed that for some of his XFOIL simulations, Tom Speer used to put the transition point at -1%
    to assume a fully turbulent boundary layer.
    It is common knowledge among CFD gurus that in case of a turbulent BL, the main driver of section/profil drag is the section thickness.
    That is why I am afraid a 18% relative thickness wing section is likely to be very draggy, unless you use it in a 2/3 elements wing sail like C-Cat rigs. All what is above mentioned is about section drag, but induced drag is probably more important to minimize.

    Considering and computing your maximum righting moment windward should be a good starting point.
    Then you choose your preferred true wind speed conditions and you can start to design your sail plan.

    For Induced Drag an elliptical lift distribution would be the least draggy sail plan for ONE apparent wind speed, the apparent wind speed where you can use your maximum righting moment without being overpowered.
    As soon as you have to de-power a little bit, the elliptical lift distribution is not the least induced drag solution any more.

    Cheers & fair wind

    Erwan
     
  11. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    I find generalization hard to make, considering the great variety of profiles, and the fact that we can adjust their geometry to a given purpose. Nevertheless, it is true that minimizing the wetted surface of a profile is to be considered, in the design loop, resulting in minimizing the relative thickness, in general, with the constraints of their structural strenght.

    I've re-do the calculations @18deg incidence, following your remarks, with a modification of the initial transition point on both the extrado and the intrado. Here are the results, to be compared with the initial calculations done without this presciption, in my first post.

    upload_2023-1-11_9-21-16.png

    upload_2023-1-11_9-21-33.png

    We observe that
    - I didn't change the name of the airfoil on the second image, but it is obvious that this result has been obtained on the thicker profile. I should have named it a "NACA63025" to be consistent with the previous calculations.
    - The lift coefficient of the NACA63012 is effectively increased, when the drag coefficient remains almost the same. As a result, the lift/drag ratio of the profile is increased, by a factor 1.4 (+40%)
    - The lift coefficient of the NACA 63025 is decreased by a factor of 0.82 (-18%). Its drag coefficient is increased by a factor of 1.14 (+14%). As a result, the lift/drag ratio of the profile is decreased by a factor of 0.72 (-28%)
    - The lift coefficient of the NACA63025 is 7% higher than the same coefficient for the NACA63012. The drag coefficient of the NACA63025 is 60% smaller than the same coefficient for the NACA63012.

    I conclude that the turbulence effectively increases the drag of the thicker profile, while increasing the lift of the thinner profile, as Erwann said. But the drag of the thicker profile remains significatively smaller than the drag of the thinner profile, resulting in the fact that the thicker profile, namely the NACA63025, could be more effective than the NACA63012, at this incidence, and with this flow conditions. Although I find that the drag calculated for the NACA63025 is too little, and couter-intuitive, this is the second set of calculations that shows this tendencie. It may be consistent, again, with the fact that the NACA63012 has been optimized for working below 12° or 15° of incidence. I foresee that, in this range, the NACA63012 could be more efficient than the thicker NACA63025. It is common that the performances of a modified and thicker version of a known airfoil are very often less than te performances of the original version. But we see here, that, by chance, outside the range 0° to 15°, a thicker version could indeed be more effective.

    This may demonstrate that the thickness is not the main parameter that governs the performance of a profile, but the repartition of the curvature along the profile is to be considered also, especially at 3 points : At the leading edge, at the Bmax, and at the trailing edge, or at the inversion point, when inversion in present. The curvature of the thicker profile may be, indeed, more prolific that initialy thought, to be used on a wing sails. Any CFD guy will play with these numbers to adapt a given profile to the working conditions and the requirement of the foil that is being made out of this profile.

    It will come in a second time. With the newly given informations by Dustman, I will be able to draw the planforms, in order to work on this induced drag. My approach is a bit different, since I've turned all the classical formulations (elliptic flow theory, panel code, VPP) upside down, or inside out if you prefer. Given a TWS, TWA, I am able to define a planform achieving a given boat speed. This is very handy when designing record breaking boats, because the only thing that you know at the beginning, in that cases, are the conditions that you have to fullfill in order to beat the numbers. This approach spares design loops.

    All that calculations will be done in concurrent engineering, for comparison with a traditionnal sailplane, seeking a trade-off between the two, limiting the aspect-ratio of the wing sails and its design complexity , to respect the wish of Dustman.
     
  12. Erwan
    Joined: Oct 2005
    Posts: 460
    Likes: 28, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    2 remarks:
    -I am afraid Ncrit 9 is not appropriate, should use NCrit=1 in order to consider the natural turbulence of wind.
    - Alpha=18° probably not relevant, unless you are studying the stall conditions of different wing sections.
    if you want to compare L/D for different wing sections, it should be investigated in the linear portion of the polar (ie: between 0 and 10° or 12° maximum angle of incidence.)
     
  13. Erwan
    Joined: Oct 2005
    Posts: 460
    Likes: 28, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    Your XFOIL simulation testifies on how separation is more important on a thin wing section compared to the thicker one (the yellow lines), but again studying wing sections at 18° does not bring information on the relative performance of these wing sections.
     
  14. Alan Cattelliot
    Joined: Jul 2021
    Posts: 506
    Likes: 212, Points: 43
    Location: La Rochelle (Fr)

    Alan Cattelliot Senior Member

    Exactly why thicker profiles are more tolerant than thinner profiles, in general.

    As you may have already noticed, I'm afraid this is worth than that, because I've also made some initial calculations at 30° and 43° of incidence. As previously said, I'm doing this small exercice in concurrent engineering, which means that I want make a comparison with the performances of a standard sail set, say upwind. The only validated model of catamaran main and jib set I have, provides lift coefficient and drag coefficient, only for a range a AWA between 9° and 47°. How often do you sail at awa between 0° to 12° (taking into account the leeway angle and a maximum angle of rotation of the mast(s), of course) ? I know that the wind gradient and the twist helps to decrease the aoa, but is it really what you are pointing out, here ?

    The choosen range, between 0 and 43°, is choosen in order to observe the drop in the polars that is awaited at the extreme of the sailing range upwind. And yes, I want to know how the flow is attached, or detached, in this whole range. If there is any opportunities to catch on the global polar curve of the boat, I want to be able to evaluate it, and quickly estimate what modifications are to be made in order to catch these opportunities , If choosen so.

    Again, for comparison purpose, I set the Ncrt equal to 9, resulting in flow simulations that may be compared with traditionnal windtunnel results. I will deal the problem of flow stability and its influence on the polar curves with finer DNS, when the big numbers will be in place, and with an overall geometry of the boat modelized in my tools. As you already said, the induced drag, but also the parasite drag, must be taken into account to define the plan view of the sailplane.
    Finer calculations for finer geometry. That's the rule.
     
    Last edited: Jan 11, 2023

  15. Erwan
    Joined: Oct 2005
    Posts: 460
    Likes: 28, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    You might find interesting informations reading a the following Tom Speer's work paper:
    "Optimum Plan Form"
    Planforms (tspeer.com)
     
Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.