Tandem Keel...?

Discussion in 'Sailboats' started by Lew Morris, Nov 6, 2002.

  1. Lew Morris
    Joined: Jun 2001
    Posts: 124
    Likes: 4, Points: 18, Legacy Rep: 92
    Location: Pismo Beach, Ca

    Lew Morris Industrial Designer

    The November 2002 issue of Sailing magazine (Perry on Design) show an option "tandem" keel offered on their 26i model.

    This tandem keel is basically two keels (in line) connected by a common bulb (for lack of a better term).

    I would think that the turbulance created by the forward keel would really screw up the water for the aft keel....

    Can anyone enlighten me of the concept behing this design?
  2. MDV
    Joined: Feb 2002
    Posts: 31
    Likes: 1, Points: 0, Legacy Rep: 60
    Location: Singapore

    MDV Junior Member

    Given enough seperation, the leeway of the yacht should give smooth flow on both foils. As far as I understand it the concept is trying to obtain two high aspect ratio sections which collectively have less drag for the same amount of lift as a single section. This was the driving factor with the twin struts fixing the front wing on a F1 car, although in that case the driving factor was minimal drag for a given strength.

    Check out the following site for further information.



  3. Lew Morris
    Joined: Jun 2001
    Posts: 124
    Likes: 4, Points: 18, Legacy Rep: 92
    Location: Pismo Beach, Ca

    Lew Morris Industrial Designer

    Thanks Michael,

    I should proof-read my posts... the boat in question is the ETAP 26i.


    Unfortunately there is no further information regarding the benefits of this keel configuration on their website.

    As a reference, the side elevation drawing provided in the article shows the forward and aft keels to be of the same width. The distance between the trailing edge of the forward keel, and the leading edge of the aft keel, is roughly 2/3 of the width of either keel.... the lower surface of the bulb ( in the side elevation)appears to be nearly flat. A view looking aft would have been helpful.

    Thanks for your comments!
  4. origamiboats
    Joined: Nov 2001
    Posts: 20
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: BC, Canada

    origamiboats Junior Member

    I can't offer much information, but I can offer a lead in terms of some keywords that might help your search for information. Warwick Collins is the designer of the tandem keel of cast iron used on the Hill's Bendford Badger, a junk-rigged boat featured in their book "Voyaging on a Small Income". A web search on this name may send you in the right direction.


  5. origamiboats
    Joined: Nov 2001
    Posts: 20
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: BC, Canada

    origamiboats Junior Member

    I found a litte more on the Collins tandem keel at the website page for the Sadler 34 at http://www.sadlerandstarlight.co.uk/docs/sadler34.htm#tandem

    A quote from that site:
    Warwick Collins spent more than three years in developing his tandem keel, with tests on tank models and on full-sized boats.
    Claims made for the Tandem Keel:

    1. A lower centre of gravity

    2. Better directional stability and resistance to broaching

    3. Better damping of pitch and roll

    4. Smarter tacking

    5. Improved weatherliness in rough water

    There is now plenty of evidence from owners and skippers that where a standard keel is replaced by a tandem keel the boat's general behaviour is much improved, and that speed made good to weather is either the same or better.

    Attached Files:

  6. Lew Morris
    Joined: Jun 2001
    Posts: 124
    Likes: 4, Points: 18, Legacy Rep: 92
    Location: Pismo Beach, Ca

    Lew Morris Industrial Designer

    thanks for the link.

    it would appear that the "bulb" in this image, in addition to having a tandem keel, is of a "winged" configuration.

    interesting stuff...
  7. Steve Gray
    Joined: Oct 2002
    Posts: 22
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Weymouth, UK

    Steve Gray Junior Member

    Hello, Lew

    A few years ago there was a move towards putting a slot just behind the leading edge of the fin on high-speed sailboards to help combat 'spin-out', which is the tendency of the board's tail to swing off to leeward when water turbulence or poor rider stance caused the fin to come unstuck at speed (usually a problem at about 15+Kts).

    The concept was that the slot acted like the slat at the LE of an aircraft's wing (or the primary feathers on a bird's wing (or, of course, a jib in front of a main)) to assist the fluid (water or air) to stick smoothly to the higher-aspect foil surface rather than turbulate--it effectively allows the foil to maintain a better grip on the water. The keel configuration you show will obviously provide these benefits, and the bulb wing will also act as a 'tip fence' to stop water flow being diverted downwards (look at the tips of a Jumbo 400), thereby retaining even more traction (which shouldn't be confused with drag).

    I reckon that the list of quoted claims can all be put down to good traction of the keel. Incidentally, on sailboards, the fashion died when it was realized that no amount (or absence) of plastic would compensate for pilot error ('heavy feet'); now we just use a longer, very high-aspect blade.

    Hope this helps...
  8. tspeer
    Joined: Feb 2002
    Posts: 2,304
    Likes: 269, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    Splitting the area into two surfaces of the same span does not increase the aspect ratio. Aspect ratio is the span^2 / total area. If you substitute the keel dimensions into the formulae for drag to get the total drag (in pounds or newtons), as opposed to the drag coefficient, I think you'll see what I mean.

    Most of the reasons given for a twin keel don't make sense to me. On a beat, the lift from the keel is dictated by the sail rig, so getting more lift from a slotted section only makes sense when the boat is tacking or if the area of the keel is reduced.

    The induced drag due to lift depends on the total lift distribution over the span of the keel. If the lift distributions of the single and tandem keels are similar in shape and the depth of the two keels is the same, then the induced drag will be the same. This is true regardless of the area of the keels or how that area is distributed in the streamwise direction (this is known as the Munk stagger theorem). Likewise, the effect of winglets on the keel's induced drag should be largely independent of their location relative to the keel in the streamwise direction, so the same winglets should be comparably effective for either keel.

    The tandem keel can have less skin friction drag if the total area is less than the single keel. However, the tandem keel's I've seen pictured don't seem to have markedly less area. So the profile drag should be comparable for both designs.

    With comparable profile drag, induced drag, and lift coefficients, that pretty much rules out all the conventional explanations as to what makes tandem keels go. But I think there are some other factors that haven't been mentioned.

    For keel sections of comparable thickness ratio, the thickness of the tandem keel is half that of the single keel. The intersection drag at the boat and bulb junctions depends on thickness squared, so it's possible the junction drag is halved in the tandem design. The tandem keel's leading edge can go right to the nose of the bulb, with the attachment line of the keel ending at the stagnation point on the bulb. This would eliminate the necklace vortex that would form at the junction of a blade attached to the middle of the bulb, and reduce junction drag even more. Of course, the aft surface on the tandem keel is still subject to forming a necklace vortex at its junction.

    The pressure disturbance of a narrow fin keel is concentrated in one location, and there's a wave drag penalty for this. From a wave drag point of view, it's as though you wrapped a bump around the hull at the keel location, much like the equivalent body of revolution used to analyze transonic wave drag in air. The tandem keel may well provide a better distribution of the equivalent cross sectional area and reduce the wave drag over the single keel.

    The two surfaces of the tandem keel will provide more yaw damping than a single blade. Yaw damping varies with the square of the distance between the surface and the center of gravity, so damping provided by the rudder is much more dominant than either keel. but if you didn't have a rudder, the increased damping would be significant.

    The real gains may be structural rather than hydrodynamic. The twin keel will be thinner, and thus weaker than a comparable single keel with respect to sideways bending moments. However, it should be much stiffer in torsion when supporting a long, heavy bulb because the torsion loads are reacted by differential bending in the two surfaces. The tandem keel is also attached over several more floors, which would be especially valuable in a grounding situation.

    Whether or not there's a net gain for a tandem keel probably depends on the details of its implementation. However, the most effective way to improve the efficiency of any keel going to weather is simply to make it deeper. There's no substitute for span.

    A good figure of merit for comparing two keels is the boat's wetted aspect ratio. This is the depth-squared divided by the total wetted area, including hull, keel, bulb, wings, etc. Chances are the boat with the highest wetted aspect ratio will be the more closewinded one.
  9. Guest

    Guest Guest

    I don't understand your statement about the aspect ratio. If I have a keel of span X and area Y and I then slit the keel into 2 keels each of the same span (x) but each having an area of Y/2, wouldn't the apsect ratio of each keel be x^2/(y/2) as opposed to the original single keel of X^2/Y. Is there a reason the two keel would not be considered seperately and treated as two seperate surfaces? I would think that the increase in aspect ratio would be the best argument for the twin keel. If this argument is invalid that really puts a substancial dent in the argument for twin keels.
  10. tspeer
    Joined: Feb 2002
    Posts: 2,304
    Likes: 269, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    It's a common misconception that induced drag depends on aspect ratio. It doesn't. Induced drag depends on span and is independent of aspect ratio if you fix the span. Here's why:

    Let's assume a parabolic drag polar, typical of wings and keels when wave drag is not a factor. The drag coefficient can be written as

    AR = b^2/S

    CD = CDo + CL^2/(pi * AR * e)

    AR = aspect ratio
    b = span (depth of keel)
    CD = total drag coefficient
    CDo = parasite drag coefficient
    CL = lift coefficient
    e = efficiency factor (approx. 1.0)
    pi = 3.142...
    S = reference area

    This makes it look like the induced drag (second term in the equation above) depends on the aspect ratio. But this is a nondimensional drag coefficient - it's been divided by the really big influences; the velocity, fluid density, and reference area.

    Now take a look at what the drag really is, in pounds or newtons:

    D = total drag
    L = lift
    rho = fluid density

    D = CD * 1/2 * rho * V^2 * S
    L = CL * 1/2 * rho * V^2 * S

    D = CDo*1/2*rho*V^2*S
    + [L/(1/2*rho*V^2*S)]^2 / (pi*b^2/S*e) * (1/2*rho*V^2*S)

    D = CDo*1/2*rho*V^2*S
    + L^2 / (1/2*rho*V^2*pi*b^2*e)

    Aspect ratio has totally disappeared! The parasite drag depends on the reference area, and if you picked the wetted area as the reference area, CDo would be close to the skin friction coefficient.

    The induced drag depends on span-squared, not aspect ratio, and e, which depends on the planform shape. Winglets get into the act by increasing e, and improving b when the boat heels. If you keep the span, b, the same and increase the chord, the drag goes up because you're adding wetted area to the parasite drag. The aspect ratio is going down, too, but the induced drag isn't changing (assuming e stays constant). Finally, the induced drag goes down with speed if lift is held constant. So if you have a short span (like a boardless catamaran) you can make up with speed what you're lacking in depth. This is all completely opposite of what you'd expect from looking at the drag coefficient instead of the actual drag.

    Instead of thinking of aspect ratio as being a "skinny-ness" factor, think of it as the nondimensional form of span (squared). When nondimensionalizing an equation, anywhere you have a length^2 component, you divide it by reference area. So b^2 becomes b^2/S and you have aspect ratio. Long skinny glider wings have low drag because they have a lot of span for their area - this is what aspect ratio is really saying. Or because they have small area for their span, which would be another valid way of looking at it.

    If you look at it in this light, then the nondimensional equation starts to make more sense. The induced drag is inversely proportional to the nondimensional span-squared (aspect ratio), just like it is in the dimensional equation. For a fixed lift, the lift coefficient drops with speed, and this accounts for the reduction in induced drag with speed. So the two approaches are consistent.

    Now let's look at a tandem keel. For comparsion, I'll assume the skin friction is the same (it would actually increase somewhat because the chord Reynolds number will be lower for the tandem), and total area and span will be the same. We're just going to take a single keel and split it into two tandem surfaces.

    Drag of the single keel is

    D = CDo*1/2*rho*V^2*S
    + L^2 / (1/2*rho*V^2*pi*b^2*e)

    Now distribute the lift between two equal sized surfaces and assume that the surfaces are independent:

    L1 = L/2
    S1 = S/2
    D1 = CDo*1/2*rho*V^2*S1
    + L1^2 / (1/2*rho*V^2*pi*b^2*e)

    D1 = CDo*1/2*rho*V^2*S/2
    + (L/2)^2 / (1/2*rho*V^2*pi*b^2*e)

    D1 = (CDo*1/2*rho*V^2*S) / 2
    + L^2 / (1/2*rho*V^2*pi*b^2*e) / 4

    Dtandem = 2 * D1
    Dtandem = CDo*1/2*rho*V^2*S
    + L^2 / (1/2*rho*V^2*pi*b^2*e) / 2

    Wow - the induced drag has been cut in half! The trouble is, the surfaces aren't independent. The wake from the forward surface affects the aft surface and, surprisingly, vice versa. The real expression for the induced drag of two surfaces is (ref 1):

    Di = L1^2 / (e1*b1^2)
    + 2*(sigma/e3)*(L1/b1)*(L2/b2)
    + L2^2 / (e2*b2^2)

    For two surfaces of equal span and no separation in the crossflow direction, (sigma/e3) is approximately 1.0. And the interference between the two surfaces exactly fills in the difference between the tandem wing and single wing.

    Figure 9 of Ref. 4 (http://aero.stanford.edu/Reports/MultOp/multop9.gif) shows the relative drag of two surfaces of various sizes. For equal span, the total drag is greater than that of a single surface, and two equally sized surfaces is about as bad a choice as you can make.

    As you say, I think this does put quite a dent into the arguement for twin keels.

    However, this isn't the whole picture, either. A twin keel with a rudder is actually a 3-surface configuration. And with the right choice of parameters, it is possible to get a net savings in the trimmed drag (http://aero.stanford.edu/Reports/MultOp/multop13.gif). But it's not a slam-dunk.

    1. Rokhasaz, K, and Selberg, B. P., "Comparison of vortex lattice and Prandtl-Munk Results for Optimized Three-Surface Aircraft", AIAA-86-2695, Oct., 1986.

    2. Kendall, Eric R., "The Theoretical Minimum Induced Drag of Three-Surface Airplanes in Trim", AIAA Journal of Aircraft, Vol. 22, No. 10, October 1985, pp. 847 - 854.

    3. Feistel, T. W., "Interdependence of Paramters Important to the Design of Subsonic Canard-Configured Aircraft", SAE 850865, SAE Aerospace Technology Conference & Exposition SP-757 Advanced Aerospace Aerodynamics, Oct. 1988.

    4. Kroo, Ilan, "Design and Analysis of Optimally-Loaded Lifting Systems", AIAA 84-2507, Oct. 1984. http://aero.stanford.edu/Reports/MultOp/multop.html

    An interactive java applet that will calculate the trimmed drag of tandem wings can be found at
    http://www.desktopaero.com/appliedaero/configuration/canardcalc.html. You will have to adjust the static margin (sm) to control the loading on the two surfaces. e = 1 means the drag is the same as a single elliptically loaded wing of the same span.
  11. Steve Gray
    Joined: Oct 2002
    Posts: 22
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Weymouth, UK

    Steve Gray Junior Member

    Tom, you certainly know how to deliver a nuke to make a point. I consider myself fortunate that I can get a mental grasp on the basic idea--it must be great to actually understand the workings as well. The detailed math intimidates me a bit, so I tend to rely on analogies to get the picture.

    I accept that the slot between the keel sections doesn't equate to a higher aspect ratio, and that high aspect is defined by a contiguous longitudinal section so that the whole leading edge and chord work as a whole; I guess that part of what your last evidence shows is that, in the case of the tandem keel and aspect ratio, one plus one doesn't equal two.

    So, do you have any comment/instruction concerning the view I posted earlier about the effect of the slot in maintaining a good flow over the foils? Your detail responses seem to be focussed a lot on 'drag', but isn't this significantly affected by any turbulence caused by poor adhesion of the fluid's progress over the surface? And is there a relationship between induced drag and the 'grip/traction' that I mentioned (which is something that you can feel very acutely when moving at 20+kt on a sailboard--those long narrow fins make such a positive difference)?

    Sorry, Tom, you're obviously being pumped for free lessons here.

  12. tspeer
    Joined: Feb 2002
    Posts: 2,304
    Likes: 269, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    No, what I was saying is the aspect ratio is the span-squared divided by the total reference area. The reference area is whatever you want to define it as - it's typically the planform area for a keel. It doesn't matter if the area is divided up into two narrower surfaces or is one surface.

    Actually, the best parameter for comparing two keels would be the wetted aspect ratio, which is the depth^2/total wetted area. The wetted area would include both sides of the keel, the bulb, wings, and the bottom of the boat.

    What strikes me most about the keel in the picture is that it's SHORT. I'd bet a single fin twice its depth would beat the pants off it going to windward. But obviously, there were other considerations driving the design of that keel than windward performance.

    By "poor adhesion" I take it you are talking about separated flow, such as when the keel stalls. I think a clean fin keel with a well designed section would perform as well. The section design would take into account the speed range and size of the keel, as well as how heavily it was loaded. The section would also have to have enough thickness for the required strength and ballast.

    A slotted section can have a higher maximum lift than a single element. But it's not so much that there's a slot there, but rather the way the two sections are designed to benefit from their mutual interference. It has to do with altering the pressure distribution on each piece to control the development of the boundary layer along each surface. It's hard to get the right benefit from two symmetrical sections that are lined up in the same plane like they are with a tandem keel.

    I'm not sure what you mean by this. I'm not a board sailor, and I don't know what the conditions are on the fin of a sailboard.
  13. Guest

    Guest Guest


    Just one more question. How far would the keels have to be staggered in order for them to be considered as individual surfaces.
  14. tspeer
    Joined: Feb 2002
    Posts: 2,304
    Likes: 269, Points: 83, Legacy Rep: 1673
    Location: Port Gamble, Washington, USA

    tspeer Senior Member

    The factor sigma in the formulae above is factor that handles that. See http://www.desktopaero.com/appliedaero/configuration/multiplesurfaces.html for some typical values. For the tandem keel, the interference is cut in half when the surfaces are roughly 20% of the span apart in the crossflow direction.

    BTW, this whole discussion also applies to two-masted rigs. Cat schooners are generally not as efficient as a single rig, so why wouldn't people expect the same thing to apply to keels?

  15. Guest

    Guest Guest

    Tandem keel

    Many thanks so far for the insightful comments on tandem keels. I have been searching the net for some information, but found little except Sadler and Etap yachts home pages.

    The Etap product brochure gives a technical drawing about the tandem keel, showing a really big wing/ bulb at the bottom, which brings the center of gravity down.

    From all material I have read and understood on this keel,
    the claims are:
    -reduced draft
    -same range of stability as deep keel
    -same pointing ability a deep keel

    The principle involves, that

    - the center of gravity is kept low although the draft is reduced

    -the bulb/ wing is really big and works as a fence/ endplate (dont know the right expression in English), thereby creating an almost 2D- flow with little vortex around the keel "tip"/ endplate.

    -the "window"/ slot between the keels is claimed to be carefully designed based on towing tank experiments. The slot widens towards the upper side of the keel. This is supposed to encourage the fluid passing between the tandem keels to have an upward component, thus counteracting the downward flow on the pressure side and the associated tip vortex.

    If all this was true, an ingenius way of reducing draft without penalty was found. For the cruising sailor this would mean a breakthrough, avoiding all the problems associated with lifting keels.

    In spite of all the insightful comments of tspeer, a comparison of the collins keel with a double wing configuartion or a shooner rig apears unjust, as there is much more control over the flow with the short distance of the foils and the big endplate - similar to the controlled flow in a turbine.

    But where are the hard facts? Does anybody know about reports/ proceedings about the Collins keel? Are there test reports with the same hull design, but different keel configurations?
    And who is or was Mr. Warwick Collins., if not present in the net.?
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.