sail aerodynamics

Discussion in 'Hydrodynamics and Aerodynamics' started by Guest, Mar 21, 2002.

  1. daiquiri
    Joined: May 2004
    Posts: 5,371
    Likes: 259, Points: 93, Legacy Rep: 3380
    Location: Italy (Garda Lake) and Croatia (Istria)

    daiquiri Engineering and Design

    And how would he measure the sail shape too, without which the BL data is useless?
     
  2. Erwan
    Joined: Oct 2005
    Posts: 462
    Likes: 29, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    QUOTE:
    Mark D. and Tom S. have done a great job teaching fluid dynamics here and elsewhere* on the web.

    *You should add :and at the MIT for Mr Drela and at Boeing for Mr Speer.

    I would love to "measure" the impact of local separation on lift, drag , of an A-Cat rig, compared to a single asymetric wingsail with similar lift and no separation.

    The good new: it starts like a 2D problem.

    Any idea?

    Cheers
     
  3. Mikko Brummer
    Joined: May 2006
    Posts: 574
    Likes: 83, Points: 28, Legacy Rep: 258
    Location: Finland

    Mikko Brummer Senior Member

    A paper about pressure measurement in fullscale, windtunnel and comparison to CFD (there's many more). Mostly, I would see the use of pressure measurement in CFD validation, but even there the problem of how to measure the apparent wind (necessary "inflow" for the CFD comparison), is rather overwhelming.
     
  4. powerabout
    Joined: Nov 2007
    Posts: 2,944
    Likes: 67, Points: 48, Legacy Rep: 719
    Location: Melbourne/Singapore/Italy

    powerabout Senior Member

    lucky for us several billionaires in the AC also want to know
     
  5. Mikko Brummer
    Joined: May 2006
    Posts: 574
    Likes: 83, Points: 28, Legacy Rep: 258
    Location: Finland

    Mikko Brummer Senior Member

    Not much separation on a properly trimmed A-cat sail on the leeward side, at least... some on the edges around the foot & the head, but that's it, all leech most of the time attached. On the windward side there may be a large bubble (vortex) especially downwind. But as the illustrations from a study about 10 years ago show, you will not learn much about a 2D-study - the separation will be highly 3-dimensional.
     

    Attached Files:

  6. markdrela
    Joined: Jun 2004
    Posts: 307
    Likes: 30, Points: 28, Legacy Rep: 324
    Location: MIT Aero & Astro

    markdrela Senior Member

    The simplest way is with an integrating rake (IR) mounted on the trailing edge or just a little behind it. On a soft sail it could be attached to the rear end of a batten perhaps. A single differential pressure measurement between the IR pressure, p_IR and the freestream total pressure pt_inf (e.g. sampled at the top of the mast) provides a very good estimate of the profile drag contribution at that spanwise location. You'd also need the freestream static to get the wind dynamic pressure. This has been used on sailplanes in flight to estimate the profile drag, to detect early transition, optimize flap settings, etc.

    The actual formula for the local profile drag coefficient is

    cd = [(pt_inf - p_IR) / (pt_inf - p_inf)] * (h_rake / chord) / (1 + H)

    where H is the local shape parameter which is the only thing that needs to be estimated. If all you need is relative cd measurements, e.g. for sail trim evaluation, then assuming H is fixed is very reasonable.

    One could use it to manually find the best sail trim. If we're at one point of sail, most of the stuff in the cd formula is fixed, in which case we have

    cd ~ pt_inf - p_IR

    so this pressure signal could be used to assist in setting the best sail trim. This is especially true in strong wind when the sail is depowered, and the lee side telltales are always attached and hence don't tell you much.

    For a jib+mainsail combo it gets more complicated, since what counts is the combined cd from the two sheets. In theory this could be measured with one tall rake which sees both wakes. Or one could have one rake on each trailing edge, and the two p_IR pressures would then need to be combined with appropriate plumbing. And if one really wants to get picky, there should be a number of rakes across the span all plumbed together. I don't know how practical or impractical this setup would be. It might be useful for sail-trim training at least.
     
  7. Mikko Brummer
    Joined: May 2006
    Posts: 574
    Likes: 83, Points: 28, Legacy Rep: 258
    Location: Finland

    Mikko Brummer Senior Member

    Interesting - could you point somewhere with information about the integrating rake?

    But this would be just the profile drag, right? You'd need to measure the induced drag, too, for total drag? I recall this could be done with vanes outside the boundary layer (?). And max L/D is a desired trim condition only for heavier winds, in lighter air max. drive (thrust) is what you you are after.
     
  8. markdrela
    Joined: Jun 2004
    Posts: 307
    Likes: 30, Points: 28, Legacy Rep: 324
    Location: MIT Aero & Astro

    markdrela Senior Member

    Try searching for "integrating wake rake".

    Here's one, but I don't know if it's available to everyone.
    http://journals.sfu.ca/ts/index.php/ts/article/download/289/271
    Their rake has a huge number of tubes, which is overkill. About 10-15 tubes is usually sufficient, especially for relative measurements.
     
  9. Erwan
    Joined: Oct 2005
    Posts: 462
    Likes: 29, Points: 28, Legacy Rep: 97
    Location: France

    Erwan Senior Member

    Thank you very much Mikko, for posting your A-Cat research,

    So the main separation to address is in a low velocity environment and probably accounts for little in the global section drag ??

    Is it correct??

    Can you disclose the freestream velocity assumption for this A-Cat simulation ?

    I guess it is between 1m/s and 7 m/s

    Always fascinating to run the simulations you post on a regular basis

    Regards

    EK
     
  10. Skyak
    Joined: Jul 2012
    Posts: 1,461
    Likes: 147, Points: 63, Legacy Rep: 152
    Location: United States

    Skyak Senior Member

    Real time drag among other things. This all started with a video of AC teams using MEM pressure sensor arrays, the soft sail guys want the same capabilities, I say soft sails need shape and boundary layer measurement.

    This is just my best idea to give soft sails what the AC wing guys have -closed loop design optimization linked to theory (CFD, or even just a stack of 2D layers).

    The experts are here. I deffer to them.
     
  11. daiquiri
    Joined: May 2004
    Posts: 5,371
    Likes: 259, Points: 93, Legacy Rep: 3380
    Location: Italy (Garda Lake) and Croatia (Istria)

    daiquiri Engineering and Design

    Correct.
    And the drive force depends on the heading relative to the true wind.
    Then there is the influence of the sail shape, which is an independent, both manually and automatically (elasticity) adjustable variable. How about the mast tuning? CoG position?
    There are so many independent and semi-independent variables in case of cloth (or membrane) sails, that the info coming from just the the boundary-layer measurement is grossly insufficient (and I dare to say - almost insignificant) for any practical use on board, IMO.

    A much more usable approach might be a mapping of the rig and hull characteristics for ALL combinations of independent geometric parameters, by performing a comprehensive CFD and/or wind-tunnel analysis, and feeding it into the on-board computer. It would then give the optimum overall boat trim (sail+hull) for a given:
    - heading
    - apparent wind direction
    - apparent wind speed
    - boat weight
    - boat speed

    But hold on for a moment and just think about the amount of input data required. The amount of required man-hours and computing/wind tunnel time for analysing all the possible combinations of the above independent parameters is simply frightening.
     
  12. Skyak
    Joined: Jul 2012
    Posts: 1,461
    Likes: 147, Points: 63, Legacy Rep: 152
    Location: United States

    Skyak Senior Member

    Mark, thanks for straightening this out.

    A wake rake on the end of a batten being the starting point, I was thinking how I could get the info cheaper and with less to go wrong. Could I sweep a single sensor through the wake and use position to calculate? For example a rotary setup would give info over a greater span. An oscillating setup would be simpler. Is there a percentage of cord this sensor would need to cover?

    Do I need pressure info all the time or is it possible that after the sail is 'characterized' that a microphone could be used for the real time, all the time, indication of boundary layer thickness and turbulence? these sensors are so good, looking into the future I can see race rules to limit their use.
     
  13. Skyak
    Joined: Jul 2012
    Posts: 1,461
    Likes: 147, Points: 63, Legacy Rep: 152
    Location: United States

    Skyak Senior Member

    It's not just profile drag.

    You don't need to measure induced drag if you can calculate it reliably from your sensor data.

    Max L/D is not the only boundary of the performance envelope, but it is an important one, if not the most important one. Knowing the boundary layer layer thickness also helps determine max force and it's direction. Knowing it in real time would improve the dynamic trim -how to play the sail.

    The other sensor I am proposing is sail shape. To judge feasibility I need to know what kind of accuracy is needed to link CFD to real sails. Can you tell me about what has worked and what didn't? Have you measured sails in 3D as flown? If so have you taken the real data into CFD?
     
  14. Skyak
    Joined: Jul 2012
    Posts: 1,461
    Likes: 147, Points: 63, Legacy Rep: 152
    Location: United States

    Skyak Senior Member

    FEA is a process that allows engineers to calculate the behavior of complex systems they don't know about by breaking up a system into many small pieces that behave in a predictable manor and capturing the complexity in the shape of their interaction. Small bits of a known material for example. But there is often a a smaller grid of higher order elements that give equally accurate results. Higher order elements can be more accurate when you admit that all those little bits might deviate from reality by a small amount that adds up. This error in knowledge of the environment you are trying to predict can be addressed by closing the loop with sensor data -that is what I am proposing here for sails.
     

  15. daiquiri
    Joined: May 2004
    Posts: 5,371
    Likes: 259, Points: 93, Legacy Rep: 3380
    Location: Italy (Garda Lake) and Croatia (Istria)

    daiquiri Engineering and Design

    Then your idea makes sense. Validation of CFD data and improvement of physical models used for the prediction of fluid behavior.
    I had understood instead that the idea was to give the sailor on board the possibility to evaluate the sail drag - and just couldn't grasp the reason for doing it, since such info would be pretty much useless to him.
     
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.