Why aren't sailboat hulls designed as airfoils?

Discussion in 'Sailboats' started by kerinin, Jun 12, 2009.

  1. grob
    Joined: Oct 2002
    Posts: 216
    Likes: 5, Points: 18, Legacy Rep: 53
    Location: Cotswolds Waterpark, UK

    grob www.windknife.com

    While what C 249 says is correct a big part of the 10% perfromance difference between a H16 and a F16 is the weight of the boats a H16 weighs around 145kg and a F16 105Kg. The hull design is not the dominant factor here the weight is.

    Gareth
     
  2. Ad Hoc
    Joined: Oct 2008
    Posts: 6,609
    Likes: 617, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    grob
    "..What did I learn from this experience? Boat design is about compromise,.."

    Absolutely 100%. It takes a lot to fully appreciate this, but unless one has gone through the various stages and actually designed and built a boat, not a model, this becomes very apparent very quickly.

    Having a deep draft sadly doesn't help too much too. The residuary resistance increases with B/T ratio. However, as with everything in design, as noted, it also not just about compromise but also the caveats in the region one wishes to design to as well. Since what may be true at lower speeds may not be applicable at higher speeds etc.

    For example, the biggest influencing factor is the length displacement ratio.

    Take the classic series 64 hull forms produced in the 1950s, since these are applicable to any other slender catamaran. These clearly showed the influence of B/T and Cb. When B/T increased, so did the residuary. Increasing the Cb also increases the residuary; any authoritative work will show this.

    Having more volume below the waterline is dependent upon several factors and when successful can reduce pitching to a degree. Primarily the amount of added mass, owing to the chnage in shape, as a %'age of the total added mass needs to be high. But the waterplane area needs to be reduced in said location too. If the waterplane area is low, far fwd, ie a shape like a slice of cake, this has the effect of reducing the radius of gyration.

    Changing the added mass and reducing the radius of gyration has a significant impact on the pitching moments and period.
     
  3. C 249
    Joined: Jun 2006
    Posts: 25
    Likes: 1, Points: 0, Legacy Rep: 16
    Location: Sydney Australia

    C 249 Junior Member

    Sure, that's true. Interestingly though, the H16 is still a formidable performer downwind (especially in light airs) which is another time that the extra weight should be a handicap. That tends to make me think that the main difference in upwind performance is the other bits of the 4.9 design. Then again, I ain't no cat guru.
     
  4. kerinin
    Joined: May 2009
    Posts: 31
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Austin

    kerinin Junior Member

    First, I just want to say thanks for everyone's input. This has been a very interesting weekend and I really appreciate have a bunch of more knowledgeable and experienced people to bounce ideas off of.

    So I ran the two basic hull shapes we've been discussing through a quick CFD analysis, and it turns out I was indeed mistaken about the whole wing-pontoon idea. As suggested the aspect ratio of the pontoon was so low that the L/D ratio predicted for the airfoil in 2D analysis wasn't close to the 3D shape's performance and the surface drag ended up predominating. I've attached a few images showing the analysis results.

    I used two hull shapes, one which I've been developing for a few weeks based loosely on the Blade design and another based on an elliptical wing section using a modified NACA 1707 airfoil (I reduced the leading edge radius). Both hulls have about the same volume and I sized the daggerboard on the typical hull to produce a fairly similar amount of sideforce. These simulations are fully submerged in water (I don't have software capable of handling the air/water boundary), so the results aren't exactly perfect but it doesn't really matter.

    [​IMG]
    [​IMG]
    [​IMG]

    I was hoping the flow lines would show up - but it looks like it didn't export any of the CFD data.

    The basic result is that the wing-pontoon produced 635lbf of sideforce with 311lbf of drag. The 'typical' hull produced 466lbf of sideforce with 165lbf of drag. I didn't keep refining the models to make the sideforces match as it was clear that the L/D ratio of the wing-pontoon was much higher than that of the 'typical' hull. These simulations are in fluid moving 20mph down the pontoon and 1mph across.

    After I finished all that stuff, I started looking into symmetric foils, thinking that I could optimize the waterline cross-section of the original design using the wingfoil software I've been using, and I came across an interesting bit of information (which you guys probably already know). It turns out that if the trailing edge of the airfoil doesn't come to a sharp point (which would make a rudder attachment slightly difficult), the optimal shape of the airfoil seems to be something approaching an ellipse. I haven't spent enough time tweaking the airfoils and comparing them to actual ellipses, but to my naked eye that's what it looks like. After seeing the elliptical design I linked to earlier I thought I'd try running the CFD on a similar design - those images should be down there too. I basically took a sphere and stretched it to make the shape you see. The max depth is at 40% (60% fore, 40% aft) and the vertical section is slightly elongated. Comparing the drag of this hull to the one I was originally using (without the daggerboards), there's a reduction from 130lbf for the original design to 120lbf for the elliptical one.

    I'm currently playing with modifying the vertical section to be more 'egg' shaped; increasing buoyancy below the waterline and tapering the top half of the shape. Obviously, any observations/suggestions would be appreciated.

    Thanks again!

    [​IMG]
    [​IMG]
    [​IMG]
     
  5. Guest625101138

    Guest625101138 Previous Member

    I am interested to see the CFD data (field plots).

    You can do screen shots if you can view it. Place the screen shot into powerpoint and 'save as' a png file.

    Also when you include pictures place them as attachments using the little paperclip icon. It makes for easier scrolling when viewing and you get thumbnails that can be sized up.

    Rick W
     
  6. Ad Hoc
    Joined: Oct 2008
    Posts: 6,609
    Likes: 617, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    kerinin

    the basic difference between 2D flow of an ideal fluid and 3D flow is that in 2D flow, the flow follows the form and closes behind the the body, no matter how blunt/bluff the body/shape is, ie no separation. Or put another way, the potential energy, converted from kinetic energy is reconverted back to kinetic energy. in real 3D flow, this does not happen, separation and viscosity all play a part.

    In your symmetrical section, when the flow velocity is not high and angles of attack are not always known, such as a rudder or keel, having a blunt nose, acts to delay cavitation and/or eddies on the 'lee' side. I can't remember too much about my trailing edge theory from memory, especially regrading trailing tip vortices, but you should check affects of differences between leading and trailing edge shapes too.

    Ask Daiquiri, he is far more knowledgeable on foil theory than I am, he I'm sure will be able to explain far better than I can. It's been a while since i went into deep/heavy theory of foil sections.
     
  7. kerinin
    Joined: May 2009
    Posts: 31
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Austin

    kerinin Junior Member

    Thanks for the attachment tim Rick, I kept looking for it but didn't see it. I'll post the CFD images as soon as I can.

    Ad Hoc, that pretty much matches my memory of airfol dynamics (although I think the 2D simulations take boundary layer separation into effect, otherwise they wouldn't be very useful). My guess is that the trailing edge behavior is being caused by the fact that abruptly terminating the airfoil leads to a very low pressure cavity behind the tail, creating massive amounts of drag on the 'butt' of the foil.

    I suspect that this is less of a problem with boats, since the water surface allows the cavity formed at the wake to reduce the pressure on that portion. I wonder what the implications on wave resistance are of having a blunt stern rather than a finely tapered one?
     
  8. Ad Hoc
    Joined: Oct 2008
    Posts: 6,609
    Likes: 617, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    kerinin
    The 2d simulations have caveats to ensure that the flow is within set parameters to assume separation has already occurred, or not, depending upon the Reynolds number and the amount of actual data that supplements the algorithms for the sensitivity analysis by the program. Since to model the separate and its effects is well beyond basic 2D flow, but with some experimental data "joining the dots" in-between can give guidance. Otherwise very powerful full 3D Navier-stokes potential flow stuff is required, but even then, no guarantee that it is correct since it requires validation..and where is that from??

    The implications of blunt/taper sterns affects the design parameters. Such as would you have a tapered stern on a planing boat, or would you have a deep blunt stern on a canal boat etc etc.

    So, the implications are simple, in essence, there is a change in drag/resistance. However, as with everything in design, once the mechanism is known, for the given parameters, what one does with it, is part of the compromise of design. This is not to say one ignores it, quite the contrary, one acknowledges the phenomenon/fact and designs around it or uses it to enhance the design.
     
  9. kerinin
    Joined: May 2009
    Posts: 31
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Austin

    kerinin Junior Member

    Thanks for the explainer on the 2d foils. I've been playing with XFoil to get an understanding of the implications of different profiles - the separation issue might explain why some of my profiles refuse to evaluate.

    As for the Navier-Stokes stuff, I'm fortunate enough to have access to a full version of Solidworks through my job, so I've been running those analysis on the full 3D hull forms. Unfortunately Solidworks doesn't handle free surface wave effects, so I'm considering trying to get OpenFOAM to compile on my linux partition. Does anyone have experience with that software? I get the impression it's not particularly user-friendly...
     
  10. Ad Hoc
    Joined: Oct 2008
    Posts: 6,609
    Likes: 617, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Kerinin

    Even with full blown 3D nav-stokes software you need to ensure you're getting "believable" results. Best to start with simple basic shapes and validate the results against known theory and experimental work, then bit by bit change parameters to gauge their effects and see if the results are maintained as per theory. If so, then you can begin to investigate with a bit more confidence.

    CFD is no different to FEA. FEA has nice colour plots, but what does it mean, is it real??..without validation of simple models and slowly progressing into more and more complex models with as much validation against known established data....it is a case of GIGO.
     
  11. Ad Hoc
    Joined: Oct 2008
    Posts: 6,609
    Likes: 617, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    kerinin

    Ive never heard of Xfoil before, i don't use CFD, so had a quick look. Confirms my thoughts above.

    "..In the low Reynolds number regime the results are usually not very accurate if a laminar separation bubble or larger separated flow regions occur. This is a result of the integral boundary layer method, which simply cannot model separation (this would require some sort of coupling between boundary layer analysis and the calculation of the external flow). The code has a option to perform a displacement iteration in order to take the displacement effects of the boundary layer into account, but there is no direct interaction....Transition prediction is performed by testing the boundary layer parameters against a set of empirically derived transition relations, which work quite well for attached flow in a wide range of Reynolds numbers.."

    So, nothing surprising there really. It is a 'cheap bit' of software (free), so one shouldn't expect it to model 3D flow perfectly...but all programs like this are good for sensitivity analysis, espeically in the absence of validation.
     
  12. Guest625101138

    Guest625101138 Previous Member

    The wave drag is a major factor in the total drag unless you have long slender hulls.

    You may find this thread illuminating on different methods:
    http://www.boatdesign.net/forums/design-software/michell-v-cfd-efd-round-2007-a-20497.html

    Leo no longer has the time to support his software for free but there are quite a few users including the US Navy. He has been working on more complex versions for selected clients that allow for sinkage but if you are seeking the lowest drag hull, and not length constrained, this is not a significan consideration.

    The really valuable part of what Leo has achieved is the speed of analysis. He has a hull optimising routine bundled with the software that does about 1000 iterations per second so in a few minutes you rapidly home in on the optimum hull for the set conditions. It is very powerful and surprisingly accurate. I have verified its drag output on slender hulls to within 2%. He even adjusted the allowable range for the water viscosity so I could align with the cold water testing we were doing at one stage

    Rick W
     
  13. water addict
    Joined: Jun 2004
    Posts: 317
    Likes: 6, Points: 0, Legacy Rep: 73
    Location: maryland

    water addict Naval Architect

    back to your original questions:

    If the entire hull was a foil shape, you would have:
    -large wetted surface for given displacement, so more friction drag
    -larger wave-making drag than a well-designed canoe body hull form
    -poor lift characteristics from low-aspect foil shape

    I'm not sure I understand the stern shape question. Are you referring to the fact that sterns are not pointed like the back of a foil?
     
  14. gggGuest
    Joined: Feb 2005
    Posts: 828
    Likes: 24, Points: 18, Legacy Rep: 76
    Location: UK

    gggGuest ...

    In the early days of applying the reults of tank testing etc to yacht design there was an excellent example of this. It was a 12m yacht in the US defender trials called, IIRC, Mariner. The story goes that after yet another miserable day on the water Ted Turner (who was sailing the boat, not videoing it!) tuned to the designer, Britton Chance, and said "you know Britt, even a turd is pointed at both ends"
     

  15. yipster
    Joined: Oct 2002
    Posts: 3,486
    Likes: 95, Points: 58, Legacy Rep: 1148
    Location: netherlands

    yipster designer

    cut sterns above a certain speedrange are to prefer to cano sterns
    reasoning that at speed water coming clean from the bottom reduces friction drag so giving a better compromise
     
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.