hard wing vs soft wing?

Discussion in 'Hydrodynamics and Aerodynamics' started by Slingshot, Apr 11, 2020.

  1. Doug Halsey
    Joined: Feb 2007
    Posts: 640
    Likes: 212, Points: 53, Legacy Rep: 160
    Location: California, USA

    Doug Halsey Senior Member

    Agreed. In particular, whenever the apparent wind angle is greater than 90°, sail drag acts to increase the aerodynamic thrust, and as long as the additional hydrodynamic drag due to sideforce isn't too much, that will result in more boat speed.
     
  2. Erwan
    Joined: Oct 2005
    Posts: 460
    Likes: 28, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    Thanks for your participation Inquisitor,
    afaik JAVAFOIL can address 2 elements slotted wing, while, XFOIL addresses only single wing section with hinged flap.
    But JAVAFOIL outcomes are probably much more optimistic.

    A Cl=4.2, even for a 2D Cl, it looks a bit "too optimistic"

    And Cl=3.2 for hinged flap????

    I need to revisit Abbott & Van Doenhoff.

    Cheers

    EK
     
  3. Inquisitor
    Joined: Nov 2005
    Posts: 276
    Likes: 11, Points: 18, Legacy Rep: 155
    Location: North Carolina Mountains

    Inquisitor BIG ENGINES: Silos today... Barn Door tomorrow!

    Good God... I wouldn't want to curse someone with that one! Its probably my decade long absence. Tonight, I'll dumpster dive the hard disk and see if my recollection is the one that's optimistic.

    VBR
     
  4. tspeer
    Joined: Feb 2002
    Posts: 2,319
    Likes: 303, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    You need to take maximum lift predictions from Javafoil with a bit of skepticism. There area stall mechanisms that Javafoil cannot calculate. One big one is wake bursting. This occurs when the wake from the forward element passes through the region of increasing pressure generated over the flap. The low-speed flow in the wake gets pushed backwards by the pressure, leading to separation in the middle of the the flowfield. What's left is the flow through the slot that adheres to the flap as a wall jet.

    The attached figure shows the streamline-fitted grid calculated by MSES around an AC72 wingsail section with the flap at 30 degrees and an angle of attack of 5 degrees. The wake has burst, leading to flow separation. The flow above the wake from the main element is no longer following the flap. But the flow through the slot is attached to the flap. Telltales on the wing would all be laying flat, yet the wing is stalled!

    The calculations in Javafoil only concern the boundaries of the flowfield, but for this kind of stall the entire flowfield has to be discetized.

    In my studies with MSES, I find it's pretty hard to get to get the maximum Cl to approach 3. Cl for the example shown is 2.5.
     

    Attached Files:

    Last edited: Sep 18, 2020
    Eric Lundy, Inquisitor and Erwan like this.
  5. tspeer
    Joined: Feb 2002
    Posts: 2,319
    Likes: 303, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    Laminar flow can be important to high lift. A.M.O. Smith talks about it in the attached paper. BTW, this is probably the single best thing I've ever read on the subject of high lift airfoils. However, he doesn't cover topics like wake bursting, and his design philosophies are mostly aimed at high Reynolds number airfoils. But the principles he describes are still applicable to other situations.

    As a practical matter, it's best not to design a section that depends on laminar flow to avoid premature separation. The risk of not having laminar flow when you need it is too great. When the AC72s were sailing in the fog on San Francisco Bay, the wingsails were covered with water droplets. I doubt there was any significant laminar flow to be had in those conditions.

    Supercritical airfoils and laminar flow airfoils are two quite different things. Supercritical airfoils are sections that are designed to operate at transonic speeds, where the local velocity can exceed the speed of sound over part of the section. This is not directly related to the boundary layer characteristics.

    Laminar flow airfoils are designed to maintain substantial runs of laminar flow. This is all about the boundary layer characteristics.

    Supercritical airfoils and laminar flow airfoils can often look much the same, because both often employ "rooftop" design pressure distributions in which the pressure is constant over for forward part of the suction side. In the case of supercritical sections, the design pressure distribution is driven to keep the maximum velocity just below Mach 1, or to minimize how much it goes past Mach 1 in order to not form a strong shock wave when the flow has to decelerate. In the case of laminar flow sections, like the NACA 6-series, an adverse pressure gradient is destabilizing to boundary layer, so a constant pressure region is often used to delay the amplification of disturbances that can lead to transition to a turbulent boundary layer. So you get similar shapes for very different reasons.

    The angle of attack range for laminar flow can be much greater than a few degrees. I've attached MSES computed polar plots for the AC72 section shown in my previous post. These were computed with the assumption of natural transition. The different colored lines are for different flap deflections, ranging from 10 degrees to 35 degrees. The plot on the far right shows the location of the transition location on the upper and lower surfaces. In the middle of the operating range, say 1.0<Cl<1.5, there is laminar flow over roughly a quarter of the chord on both the suction and pressure sides. As the angle of attack increases, the transition point moves forward on the suction side and aft on the pressure side. When a leading edge suction peak forms on the main element, transition can move abruptly to near the leading edge, which causes the sudden increase in drag. The blue 30 degree flap lines are a good example of this.

    The sudden shift in transition location due to forming a leading edge suction peak is what causes the distinctive "drag bucket" associated with laminar flow section drag polars. A section can be designed to have a wide, shallow drag bucket, or a deep, narrow one.

    Camber can be used to shift the drag bucket to a different range of angles of attack or lift. The attached plot shows an example of this. Consider the section operating at Cl=1.5. With 20 degree flap deflection (green line), this operating point is outside the drag bucket and there is excessive profile drag. Increasing the flap deflection to 25 degrees (cyan line) and decreasing the angle of attack to get back to Cl=1.5 puts the operating point just inside the drag bucket. This produces the best performance, but is risky because any additional angle of attack will result in a substantial increase in drag. Increasing the flap deflection still more to 30 degrees (blue line) and reducing the angle of attack further shifts the drag bucket to a higher lift range. The Cl=1.5 is now comfortably inside the drag bucket a the cost of just a few counts of profile drag.

    The last figure shows the MSES computed polar data for same airfoil and flap deflections, but this time the boundary layer was artificially tripped at 2.5% of the chord from the stagnation point - essentially a fully turbulent boundary layer. That's why the transition locations plotted on the far right are all near the leading edge. The drag polars are all smooth curves because there is no rapid change in the transition location. The drag is higher in this figure compared to the previous one, showing the benefits of laminar flow even for a comparatively high-lift section, and over a wide range of angles of attack.

    The design philosophy for this section was to aim for a balanced design that took advantage of laminar flow but didn't depend on achieving it. There isn't much difference to the maximum lift with or without laminar flow. The profile drag could have been reduced by pushing for more laminar flow. Or it could have been optimized for higher maximum lift by sacrificing some laminar flow in the middle of the operating range.

    I agree that it's not enough to design for laminar flow. The wing has to be built and maintained to support a laminar boundary layer. However, if we're talking about high performance craft that sail faster than the speed of the wind, drag is always detrimental to performance. They never sail for long near dead downwind where the rig acts more like a parachute than a wing.
     

    Attached Files:

    Last edited: Sep 18, 2020
    Inquisitor, Doug Halsey and Erwan like this.
  6. Inquisitor
    Joined: Nov 2005
    Posts: 276
    Likes: 11, Points: 18, Legacy Rep: 155
    Location: North Carolina Mountains

    Inquisitor BIG ENGINES: Silos today... Barn Door tomorrow!

    • et al - My referenced high Cl values - I've found and gone back through my notes from several years ago. I did recollect incorrectly! They were not final solutions. Just numbers on the way to solutions. I must have registered those high numbers in passing, during a multi-day analysis, before a final solution and evaluation that threw them out. The "high" registered in my memory far more than the let down at the end. More later.
    • Tom - Thank you for correcting me in a very constructive way. As usual, lots of great content and information to ponder. I need to stay in my lanes better and/or caveat my comments when in gray areas that are outside my expertise. I have several questions already, but I want to digest your posts and do some research to gather more.
    VBR
     
  7. Doug Halsey
    Joined: Feb 2007
    Posts: 640
    Likes: 212, Points: 53, Legacy Rep: 160
    Location: California, USA

    Doug Halsey Senior Member

    I'm relieved to see the magic words "Reynolds number" finally appearing in this thread. I feel like any discussion of section characteristics and laminar versus turbulent flow is incomplete or even misleading, unless it spells out what conditions it's referring to.

    In this vein, I think it's worth mentioning that the NACA laminar-flow airfoils referred to were designed with high Reynolds number applications in mind, where the goal was to increase the extent of laminar flow. In many sailing applications however, Reynolds numbers are small enough that laminar flow is unavoidable, and designs that result in more gentle laminar separations are better. In these situations, the old classic NACA 4-digit sections are actually better than the later laminar flow ones.

    I'm also happy to see the favorable reference to A.M.O. Smith's paper, not just because it's a great paper, but also for purely selfish reasons. (My work on jet-flapped airfoils is included toward the end of the paper, and as a result, I had to sit through at least 3 dry runs while AMO was preparing his presentation.)
     
    DogCavalry and BlueBell like this.
  8. tspeer
    Joined: Feb 2002
    Posts: 2,319
    Likes: 303, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    Yeah, at low Reynolds numbers the problem is getting rid of laminar flow gracefully rather than trying to maintain it. But for sailing craft, we're usually talking moderate to high Reynolds numbers except in very light winds.

    Before the AC match in San Francisco, the OTUSA wingsails were wet-sanded to give a smooth surface and promote laminar flow. It didn't exactly call for skilled labor, so a lot of the sanding was done by engineers and office staff. Even spouses got to participate in the sanding. They got a big charge out of being able to participate in working on the boats instead of just being spectators.
     
    Erwan, jehardiman and BlueBell like this.
  9. jehardiman
    Joined: Aug 2004
    Posts: 3,762
    Likes: 1,152, Points: 113, Legacy Rep: 2040
    Location: Port Orchard, Washington, USA

    jehardiman Senior Member

    So true.
    Again so true. When America3 (cubed, Mighty Mary, USA 43) had a problem with the welding of its keel, a NA on the design team (a Webbie female) called her college roommate (a Webbie female) who worked with me, to talk to my wife (a UofM Metallurgical & Materials Engineer) to take a look at the problem. She got a trip, a hat, and a couple of bottles of bubbly out of the deal, but the team was so excited that it was all through the distaff side.
     
  10. DCockey
    Joined: Oct 2009
    Posts: 5,229
    Likes: 634, Points: 113, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    An aside, not directly relevant to this thread's topic: Through 1980 or so the NACA 64A006 airfoil sections was commonly used in studies of transonic flutter and related topics.
     
  11. Inquisitor
    Joined: Nov 2005
    Posts: 276
    Likes: 11, Points: 18, Legacy Rep: 155
    Location: North Carolina Mountains

    Inquisitor BIG ENGINES: Silos today... Barn Door tomorrow!

    I re-read through the JavaFoil documentation and see where he only claims accuracy below stall and freely states that anything after that is questionable. I've tried to model Tom's example as close as I can in JavaFoil. This is what I have so far. I didn't expect to match the Cl = 2.5 exactly, but I thought I could get close and also see the optimism that all have mentioned with JavaFoil. Using a Reynolds Number of 1E6 (6' chord at 16 knots), I got a maximum Cl = 2.34. Unfortunately, it was at AOA = 0 degrees. By an AOA = 5, it had dropped down to Cl = 1.86. My questions are then:
    1. Do you see optimism in these results? For me (non-Aerodynamicist) to recognize something being optimistic, I would expect to see a peak. In this case, near the 2.5 (which is there) a dip then some wild ride back up to say 4.2 :confused: At that point, I would clearly rationalize that anything after the dip is numerical garbage due to invalid modeling.
    2. Say I have this trade study program that:
      1. creates JavaFoil designs like this
      2. does the JavaFoil analysis
      3. grabs the results and pulls out the key values (max Cl, Cd at max Cl, min Cd... etc)
      4. uses some criteria to get a "score"
      5. then uses Nelder-Mead techniques to find local maximums of this score.
    Are there some JavaFoil results available or some other technique that I can programmatically determine when something strays into optimism? (Peaks, derivatives, etc)

    [​IMG]
    [​IMG]
    [​IMG]
     
    Eric Lundy and Will Gilmore like this.
  12. Erwan
    Joined: Oct 2005
    Posts: 460
    Likes: 28, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    Slingshot,

    There is a "concept" or CFD phenomenon, called : "Dissipation" I suggest you to investigate.

    Unfortunatly, I cannot help you as I am far from understanding it in depth, in fact I don't make difference between separation & dissipation (maybe dissipation is the consequence of separation?) and don't know if it is relevant regarding your intial question about surface roughness between a stitched fabric and smooth Mylar.

    The only point I understand about "Dissipation" is that it is proportional to V^3, while standard section Drag equation uses the Drag coefficient with V^2.

    ie: For a classic rig I don't know if that the hinge between the teardrop mast and the full batten sail can create some "Dissipation" or just separation.

    Looking for answers on this topic, I remember a workpaper about Turbulent BL, mentionning the "sub-layer" which accounts for more or less 10% of the BL thickness.

    AFAIR it was stated that any surface roughness or protrusion which extends beyond the sub-layer, will create extra mixing and therefore will increase BL thickness and drag. (to be checked).
    (But probably it is a totally different issue from dissipation)

    As I never hear anything about dissipation, my conclusion was that it is probably a marginal issue ??

    I feel confident CFD gurus will make it clear for all of us.:)

    Cheers & Fair Winds

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

    Erwan Senior Member

    Congratulations Inquisitor for your homeworks with JAVAFOIL.
     
  14. Inquisitor
    Joined: Nov 2005
    Posts: 276
    Likes: 11, Points: 18, Legacy Rep: 155
    Location: North Carolina Mountains

    Inquisitor BIG ENGINES: Silos today... Barn Door tomorrow!

    Thank you Erwan I will definitely explore Dissipation. I was not the O.P. and I hope SlingShot (a) got his question answered and (b) doesn't mind that I was party to hijacking his thread. :oops:

    I'm not an Aerodynamicist! However, over the years, I have spent a butt-load of time with JavaFoil, typically to build model planes (which is JavaFoil's creators passion also). In those uses, I was going after higher speed and/or extreme efficiency... not maximum lift in the trans and post stall realms for multi-element wings. For the realms I was using it, JavaFoil seems to be regarded highly and is accurate.

    Point being... I'm going to try to answer your question with what I think I know... :rolleyes:

    JavaFoil has options to simulate finish (smooth, fabric, NACA standard and "bug and dirt"). I believe these are merely fudge factors applied to the equations to make them mimic reality. But the point being, they do predict very well. Also remember that before WWII most all planes were covered in fabric... NACA - US Government funded (read-military usage) would have wanted that nailed down tight.

    The following two images might answer your questions. To summarize them: This is a plane-Jane foil, not a laminar design. This is at a Reynold's number of 1E6. (6' chord, 16 knots, warm day).

    Maximum lift reduction = 0%
    Drag at maximum lift increased = 15%

    [​IMG]
    [​IMG]
     

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

    Erwan Senior Member

    Thank You Inquisitor,

    As you seems very proficient with JAVAFOIL, do you know if NCrit options exist or natural turbulences of the atmosphere is among parameters you can modify ?

    Cheers

    EK
     
Loading...
Similar Threads
  1. W9GFO
    Replies:
    23
    Views:
    3,571
  2. petethai
    Replies:
    5
    Views:
    2,796
  3. Maarten88
    Replies:
    5
    Views:
    999
  4. lunatic
    Replies:
    7
    Views:
    2,152
  5. Forecaddie
    Replies:
    9
    Views:
    1,631
  6. Forecaddie
    Replies:
    14
    Views:
    1,952
  7. lunatic
    Replies:
    15
    Views:
    2,508
  8. dustman
    Replies:
    25
    Views:
    3,832
  9. Jhomer
    Replies:
    4
    Views:
    2,210
  10. mfly
    Replies:
    2
    Views:
    2,048
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.