s a developer of yacht design and analysis softwai-e, I have i long awaited a chance to iLss% comment on what I consider /;/ %to be an incoiTect keel design technique-the cuiTent practice of using a very short keel root chord length on fm keels. (For those of you who don't deal often with "foil" hngo, the "Iceel root chord" is the section of the keel closest to the hull. For definitions of this and other terms, please see the glossary.) X TjTDically, a short root chord is com­ bined with an "elliptical" keel shape, which allows the use of longer chord lengths lower, near the middle of the keel draft.--- - ..-. .^ - ;' The logie of the current elhpt)cal keel design methodology appears hard to arfrue with, because the short root chord is intended to reduce the keel/huli interference drag (see glossary), and the longer chord lengths near the middle of the keel span can hold large volumes of lead, thereby significantly lowering the vertical center of gi-avity of the keel. The lower center of gravity means improved stabihty, and increased speed. The only obvious design problem is the thm root section, which makes attaching the keel to the hull very difficult. Using the "ellip­ tical" planform shape to achieve a lower center of gi-avity is a good concept, but I don't beheve it is necessary to incur the difficult mechanical problems of a very small root chord to reduce drag. My chance to prove that the small root chord is an unnecessary evil came when I was asked to design a new keel for the J/35 My Fair Lady. The new keel was part of an overall plan to make her more competitive in the 19S8 SORC IMS class. In order to understand how the design of tliis J/35 keel evolved, let's ni-st examine the problems I believe are m- herent with the current ellipticaUieels. Besides the obvious atiachmeni problems the short root chord creates, there are two main areas where perfor­ mance may suffer: 1. reduction of the end-plate effect (see glossal^) of the keel against the hull; 2. the placement of the longer chord lengths near mid-draft of the keel causes the center of effort to move towards the tip of the keel, especially when the elliptical shape is combined with a large leading-edge sweep-back angle. David Vacant"! explains the innovative keel Now let's explore these concepts ir he designed for the SQRC-winning J/35 David Vacanti is in his lUh year as a My Fair Lady, and makes a case for a long Pj-incipal Engineer at Boeing, anc keel mot chord. works exclusively on advanced researcl projects in the Boeing High Technolog] Center. He also owns Vacanti Yach Design, a small softioare clovelopnien fvnn that provides advanced yach design and analysis software for IB1\ PCs. performance. and showed clearly that large sweep- more detail. In order to understand my back angles on high-aspect ratio keels fu-st point about a reduction in end-plate Another factor that exacerbates the design problems of the short root chord are detrimental to performance.^ effect, let's take an extreme case where It is important to note that signifi­ we shorten the root chord until it no is the sweep angle of the leading edge. Towing tank di-ag tests, conducted at cant amounts of drag are generated by longer attaches to the hull, and we have keels with large sweep-back angles some magical way to carry this keel (gi-eater than 25 degi-ees). Minimum along under the hull as we sail. As we Typical Elliptical Keel, vortex drag for any given keel sweep shorten the root chord, we diminish the angle is achieved by selecting the proper end-plate effect to the point where the taper ratio. (See Taper Ratio gi-aph.) keel is removed from the hull, and we However, the taper ratios for large have an open end generating additional sweep angles requii-e an extremely short vortex drag and reduced lift. Now, in­ keel tip that approaches a point, while stead of having only the keel tip open smaller sweep-back angles allow longer and generating vortex di-ag (remember keel tip lengths. Since the extreme taper that wings were added to Australia IFa ratio required of large sweep-back keel to reduce vortex drag), the open angles is never achieved in practice, root chord is also creating vortex drag. these keels exhibit high levels of vortex When the keel is attached to the hull, drag. Even if the proper taper ratio is more lift is generated because there is achieved in highly swept keels, the only one open end of the keel for the minimum vortex drag levels of these water flow to escape around. However, keels is always higher than that of a there is some undesirable interference minimally swept keel. di-ag at the keel/hull joint. Not only does a large sweep-back My second concern with the cm'rent angle cripple the performance of high- eUiptical keel, designs and their short Delft in the Netheriands, showed that aspect ratios keels; it also helps move root chords is with the keel's center of the keel center of effort still lower on the effort. If you have sailed an unballasted sweeping back a keel's leading edge 45 degi-ees or more is beneficial in reduc­ keel span by encoui-aging water to flow racing dinghy with a centerboard, you down the span of the keel instead of are keenly aware that heel angle can be ing the drag of a low-aspect ratio keel along the desired root-chord direction. considerably reduced when sailing up­ with long chord lengths, while a much Obviously, sweep-back angle simply wind in heavy air by pulling up the smaller sweep-back angle is required for compounds the problem of reduced centerboard part way. A sailboat hter- high-aspect ratio keels with shorter stability described above. ally trips herself on her keel when chord lengths.'These experimental find­ • Wliile the mathematics and scientific resisting the sideforces of the sails when ings are also supported by research papers are impressive, there is nothing hard on the wind. This occurs because work done for the keel of Stars & Stripes more convincing than a winning record the keel side force or lift acts at a center '87, which stated in summary, "The in a major saihng event. The SORC- of effort location below the center of primary planform parameter which can winning keels carried by Abracadabra buoyancy, opposite to the direction of affect the level of viscous drag is the in 1986 and Splint in 1987 sported the sail forces, which act through the leading-edge sweep angle. Large sweeps longish root chords, 15-degi-ee sweep- sails' center of effort located high above can have an adverse effect on the back angles, and a special high cross- the center of buoyancy. The result is a laminar boundaiy layer...Optimization section NASA foU shape after they were large rolling moment to leeward. Mov­ of foil shape and planform geometry for optimized by Bernard Nivelt with my ing the keel area such that it concen­ low viscous di-ag can produce an efficient analysis software. trates lifting forces farther down thé keel capable of overall sailing perfor­ keel span, like that shown in elliptical mance gains." ^ In 1985 I pubhshed a At this point we can summarize a keels, tends to lower the keel center of reference work on keel performance formula for low performance by speci­ effort and is detrimental to stabihty and parameters using computer analysis. fying an elliptical keel planform, com-

Glossary winglets on 12-Meters and shoal-draft tip, the water leaving the tip is forced Chord: All lengths mecjGurod from the cruising keels. It's a means to force to spin. The spinning motion is called leading edge to the trailing edge of the water flow to remain over the a vortex, much like the swirl of water a keel, rudder or v/ing are referred to keel surface end prevent its escape seen in a bathtub drain. The spinning as chord lengths or chords. around an open lip. waterflow represents energy that was Interference Drag: Drag occurs Root Chord: The end of a keel or dissipated by the keel without gener­ whenever the energy present in a rudder that attaches to the hull Is ating lift. Any time we expend energy fluid flow is dissipated In some referred to as the "root" end. When without developing useful work, like wasted form without creating lift. we refer to the length of this end, we lift, we account for it as drag-In this Whenever two objects of different call it a "root chord." case vortex drag, sizes or orientations are placed in End-Plate Effect: Even the Wright Planform: The shape of the keel close proximit/ to one another, the Brothers understood that if an open when it is viewed from abeam of the fluid flow between them must adopt wingtip could be sealed with an "end hull, allowing the viewer to see the a speed and direction approximately plate" such as a flat piece of light leading and trailing edge shapes the average of the two Independent metal, they could prevent the natural flows. This cannot be done without Span: The height of the keel, escape of air from the wing bottom the slower flow speeding up (absorb­ measured from the root to the tip to the top side, and thus prevent a ing energy) and the faster flow loss of lift and generated drag. The Vortex: When the forward motion of slowing down (giving up energy). The hull of a boot always acts as an end the boat is combined with the natural result of this flow Interference and plate at the upper end of the keel; escape of water from the high- adjustment is energy loss that does the best known case of creating an pressure (leeward) side to the low- not result In lift. pressure (windward) side at the keel end plate at the tip is the use of SAILING WORL 40 Tip Chord Lenglh sition more gi-adual in this region as well. ing lilt over the wing tips and the critical Taper Ratio Root Chord Lenglh . These techniques are very common in aileron control surfaces (see drawing). ' , 1.5: .,..,„,,,. aircraft design. However, the swept-forward design B: Based on the Delft towing tank causes inboai-d spanwise flow wliich piles test results, we can use a large sweep- up at the wing/fuselage junction, thereby back angle only in the vicinity of the long increasing interference drag. This prob­ root chord to reduce interference drag. lem was solved by inserting a short wing The long root chord will provide the section near the fuselage which is swept :;0.5 necessary area to achieve high aft, before the remaining wing span is mechanical strength. swept forward. The aft-swept section Typical £ ilpllcal Keel : 0.1 C: Use a small leading-edge sweep next to the airframe directs the inboard angle below the root-chord area, chosen flow away from the fuselage, thereby -45 --30 -15 O -15 -30 +45 tóO reducing the interference' drag. _^ Sv.oe;) fo-ncfo . •. ;,/ 'Sw.eep Back:^- ' to match the taper ratio of the keel to achieve minimum vortex drag, and max­ A positive fallout of this design is a The points on this cun/e define the imum keel efficiency, beneficial vortex that forms at the toper ratio needed for minimum drag D: Use a laminar flow Oow drag), leading-edge joint in the wing between for a given svyieep-back angle. A high cross-section area foil shape to hold the different sweep angles. This vortex typical elliptical keel swept back becomes an active "fence" that limits the 45 degrees requires a tip chord more lead without long chord lengths. one tenth as long as the root chord to This foil shape reduces drag and can be outward (down in a keel) flow, and adds achieve minimum drag - practically used to lower the center of gravity energy into the local tlow to help main­ Impossible to achieve. without moving the center of effort. The tain laminar (low drag) conditions. center of effort will not shift because the A foil shape recently designed by bined with a short root chord and a required lead volume can be achieved NASA for low-speed general aviation au-- sweep-back angle of more than 30 with a shorter chord length, thus craft provides a large cross-sectional degrees. This design technique sig­ avoiding the need for long chord lengths area,, laminar flow characteristics, nificantly lowers the keel center of ef­ at mid span. resistance to stall, and high-lift xoeffi- fort, and negates the lowered center of E: Make the upper keel section an cients. The foil was originally called the gi-avity obtained when the longer chords empty sump, allowing room for a reason­ GA-W for General Aviation-Whitcomb, are placed at mid-span, thereby reduc­ able bilge and keeping lead low in the re­ after Richard Whitcomb, the father of ing stability and sail di-ive. The large maining keel span. the winglet at NASA. The foil series has sweep-back angle produces higher vor­ So far I have mentioned two con­ since become known simply as the LS tex di-ag due to reduced keel efficiency, cepts that need more explanation. The or Low Speed series. As originally increases viscous (friction) di-ag,' and pro­ first idea is using a large sweep-back designed, this foil shape was cambered duces less keel sideforce per unit of keel angle near the hull to reduce in­ and concave in the after sections. I have area. If the designer is not careful in terference di-ag, and the second is the modified the foil shape using a foil minimizing his choice of mid-span chord need for a special foil shape. analysis program to allow its use as a lengths, the aspect ratio may also be symmetric foil in keel design. The foil reduced, further degrading potential shape has a large radius nose section keel perfoi-mance. n excellent example of using that allows it to achieve high angles of attack and resist stall. It is also capable Having dealt the elliptical keel myth sweep angle to reduce in­ of generating what is known as leading- what I think is a considerable blow, terference drag is found on edge suction, which in effect is a force where can we turn for a design solution the advanced X-29 fighter in the direction of travel. This concept to the problems of reducing keel/hull icurrently under develop­ is exactly analagous to the forward di-iv- interference drag, while improving ment. The X-29 uses highly swept- ing force created by a sail all along its stability and keel efficiency? Our forward wings for remarkable maneu­ highly-cambered leading edge. criticism of the short root-chord design verability, because the forward sweep contains the answers. If we assemble a causes the inboard end of the wing to These design factors were combined list of desired characteristics, we can stall first in tight turns, thus maintain­ (continued on page 82) create a design approach. We need a good seal of the keel to the hull, and we -29 need a relatively long chord to allow suf­ ficient area to place keel bolts. We need a small leading-edge sweep angle for the keel, but we also have a conflicting need to sweep back the long root chords to reduce interference drag. We also need to lower the center of gi-avity to help improve stability.. Here's the design approach: A: Rather than shortening chord lengths to reduce interference drag, a more pragmatic approach is to use a relatively large radius fillet along the keel as it joins the hull. This makes the different flow speeds and directions along the hull/keel joint occur more The dual leading-edge wing-sweep on the new X-29 fighter minimizes interference gradually. It is also advisable to use a drag at the wing/fuselage junction. The author took a similar approach with his new fairing at the keel leading edge where J/35 keel by using a large sweep-back angle at the root chord and then a very it joins the hull, to help make the tran­ small sweep-back angle for the rest of the keel. »

JUNE 1988 41 A WINNING KEEL (continued from page H) REPRINTS into the keel that played a role in win­ ning the 1988 SORC for the J/35 My Fair Lady. Several people who saw the Sailing World articles keel before it raced in the SORC were are available to readers concemed about the long root chord, and about the "kinlt" in the leading edga The and advertisers at concept of double leading-edge angles or reasonable prices in "cranked" wings has been used exten­ sively, as shown in the case of the X-29. quantities of 500 or It is possible, however, to overdo the more. All reprints cranked wing concept and generate a are printed on 80 lb. low-performance, non-linear lift-vs.- The inflatable ten7 cloth beach pillow is leeway characteristic. perfect for the beach, the boat, or the car. coated stock. Research work conducted by the It's lightweight, portable, washable, and it Dutch hydrodynamicist Joop Sloof in­ comes in eight great colors: pink, hotpink, dicated that it was desu-able to reduce yellow, lavender, lightblue, royal blue and the amount of lift generated by the keel kelly green. For more Information, near the root chord.'' This was to be ac- To order, call or send check or money order for contact: compUshed by reducing the chord 1 pillow- $9.99 2-4 pillows-$8.49 each 4-10 pillows" 'STgg each plus $1.50 shipping & han­ lengths near the root chord and increas­ dling ("plus 2.50 shipping & handling) to— ing them near the keel tip. Mr. Sloof m- F.l.T.S. Distributors, P.O. Box 29, Hawthorne, MA dicates that this inverted keel planfonn 01937 tel. 617-777-8419 n.j,»»«ui*.. Production Manager with a longer tip chord than root chord Sailing World requires a swept-foi-wai-d leading edga 111 East Avenue for optimum performance." It is impor­ tant to note that by shortening the root- Norwalk, CT 06851 . chord lengths the intent is to reduce the Iblephone 203-853-9921 strength of the keel wake near the water sui-face. This is important because the keel wake near the surface generates F.l.T.S. DISTRIBUTORS wave di'ag. However, the invei-ted taper P.O. Box 29, Halhome MA 01937 Td 617.7T7-8419 ratio is only beneficial at hull speed and has diminishing value at lower speeds. Mr. Sloof s research makes no mention of how to offset the effects of lowering the center of effort, and the increased heeling moment that will result- and of coui'se the same mechanical problems of The Lightning. It hasn't forgotten the the short root chord will occur. His essence of sailing. The boat ia jimost too arguments for reducing wave drag are quick, too handsome, too sophisticated to vahd, but I'm concerned that some be a 50-year old, but those years of refine­ ment are what made her that way. Not just designers are using a short root chord a class, not just a boat, but a sailing legend. without following thi'ough on his entire Be a part of it. Contact the International concept. Lightning Class Association for a free The final concern many sailors may brochure and list ot builders and sailmakers. have is the problem of catching kelp with The International Lightning a nearly vertical section v.'hen using Design: Spaikman & Stephens minimal sweep angles. My only response Lengtti overall: 19 teet (5.8 w) to this concern is that I have seen many Beam: 6 It., 6 in. (2 mj highly swept-back keels do a more than Displacement: 700 Ib. (318 kg} rigged adequate job of kelp harvesting, and Sail area: M + J= I77sq.lt. when they were not fanning, their self- S = 300 sp.lt. hmiting pei-formance was robbing their Construction: Fiberglass or wood owners of well-deserved first-place Flotation: Tanks or loam honors. * Crew (racing): 3 Fleets chanered: Approx. 500 worldwide 1. "Experimental Analysis of Five Keel/Hull Com­ binations," J. GeiTitsma and J.A. Keuning. Ttie The International Lightning Seventh Chesapeake Sailing Yacht Symposium, 1985. 2. "Keel Design for Low Viscous Drag," Clifford J. Obara and CP. van Dam. The Eigiith Chesapeake Sailing Yacht Symposium 1987. ' V 3. "Keel Parameters and Performance," David C. Vacanti, SAIL magazine, August 1985. International Lightning Class Association 4. "On Wings and Keels U" J.W. Sloof. Ancient In­ 808 High Street, Worthington, Ohio 43085 terface XV, AIAA Symposium on the Aero- Telephone 614-885-0475 •Hydi-odynamics of Sailing Vol. 31, September 1985. Use Reader Service #369 For More Info.

SAILING WORLD 82