Sailboat keel section choice/options?

I spent a fair amount of time carefully crafting some naca 0012 templates for my fin keel, as that seemed like the best compromise section for forgiveness, max L/D at 3-8 deg of leeway, etc. - basically taken off the vacanti site and others that discuss the theory. Boat is a Dehler DB1, 33'.

After rough cutting them and checking them against the keel (which is stripped down to the lead), it appears that the keel may be more like a 63 section with max draft around 40-50% or so. The nose is narrower than the 12 section, to fit the 12 it would need some additional thickness over the first 30% or so. Also there appears to be a bit of a hollow towards the trailing edge, perhaps over the 70-95% area. It was fairly consistently off at each of the 4 chord lengths I made (top, 1/3 down, 2/3 down, bottom).

So I have a couple options, I could make up some 63 series templates with the draft aft more to see if they fit better, or I could shape the foil to a 12 section. My understanding is that the 63 series may have less theoretical drag especially dead downwind, it may be twitchier to drive efficiently with the finer nose, it has less lift overall and especially at higher AOA, it may be faster at higher speeds than the 12 section.

So, question is, what sort of characteristics of boats would lend themselves better to having the 12 section, vs. say a 63 section? Where are the very general tradeoffs or trends that would make it more or less worth it. I'm thinking things like 12 may help windward performance or acceleration out of tacks, but may be slower on downwind or reaching than 63. Boats with high aspect ratio sails and lots of sail carrying ability may benefit from additional lift 12 can provide versus the drag involved, underpowered boats probably not. Either or may be better in light air, choppy water, etc. for whatever reason, or a wash. The original designer didn't use a 12 section with good reason, or 12 sections are old news, and you're daft for even thinking of changing it - figure out what section the lead currently approximates the closest and fair it to that...

Whatever it happens to be, suggestions/ideas welcome.

 

I think the 63 series is a good section for a keel fin. The 65 series also works. These sections have realitively low attack angles but don't forget that you only see, as you said, 4-8 degrees of leeway at the keel. The 00 series is a good all round foil and can withstand higher angles of attack. I generally think the 00 series is better suited for rudder sections as these foils see higher angles of attack and the 00 series give a good compromise between stall angle, lift and drag. That is my cents on the subject. Drag kills speed so use the section that will operate in your leeway range that provides the least amount of drag.

Cheers

 

Burnsy,

A 63 section is a very likely suspect, and besides, it's a lot easier to cut templates than sanding fairing compound.

Yoke.

 

The 63 and 65 series are NACA Laminar flow sections, they are not the only laminar flow sections, nor are they the best. Check out Vortmann's work with laminar flow.

Typically, Laminar sections are fantastic to 2 or 3 degrees but much above that and you stop getting the high L/D ratio that made it so good. Instead, the turbulent sections (possibly with a turbulator) actually have better L/D ratios. It depends on what you're asking the foil to do. Is this an upwind/general boat, or a superfast downwind boat?

Tim B.

 

Tim B....

That was the impression I was under, from things I've read such as vacantisw site, hence my question. The turbulent sections are reported to be less drag in the 3 to 8+ degree leeway range... and appear more forgiving to drive to especially if not in perfectly smooth condition. It is probably a 63-012 section, although I might do better with a 0009 section I don't want to cut into the keel, so I was thinking of fairing it out as close to a 0012 section as practical. I've not seen Vortmann's work, I'll look for it. I was also thinking about doing a small end plate since it is a trapezoid fin keel and I'd rather not increase draft, but I've heard this may be very difficult to make worth the effort, never mind possible rating hit. Turbulator? You'll have to explain what you're thinking on that.

This is definitely an upwind/general boat in 99% of conditions we see - the 7/8 fractional kite does not have the horsepower ratio I'd like downwind, especially against the masthead boats and in medium/light air conditions (Great Lakes). My closest competition, a masthead Farr 33, I'd say with proper boat prep we'd be about even speed upwind, I'm a little faster close reaching right now as is, he walks away downwind in under 15ish kts true. My buoy race goal is to get to the weather mark first, if I can do that I stand a good chance the rest of the way - I'm one of the smaller and phrf rated slowest boat in my fleet. My other use for the boat is mid-distance races, 40 nm up to Mac race length ~250 nm.

Boat is very similar to a Sov 33, with slightly less horsepower (mine has longer boom, but smaller jib and only 7/8 vs 15/16, 505 vs 535 sq ft upwind) and slightly heavier (mine 7275 lbs, 3500 in keel. sov 5800, with 2800 in keel.)

 

On page 323 of Abbott & Doenhoff is a 0012-64 foil, which has a nose radius like a 0012 but has its maximum at 40% aft. One of my former employers, Jim Taylor, used a 0010-64 foil (page 318) for keels regularly.

If you think the aft part is too fat you could use the offsets from the foil above from 30% forward and the offsets from the 65A012 (page 370), or from the 65-012 if you want to keep a hollow (page 363), from 55% aft, fairing between with a spline. At least my gut says you could. The more mathematical types might frown on this.

 

Spline fitting between two aerofoils (or even averaging) is perfectly acceptable, but analyse the resulting section afterwards. By averaging the sections you will not average the CL,CD graphs.

For upwind sailing, I would definitely recommend a turbulent section. You might get away with using a laminar section at the root, and turbulent section at the tip. This may help as the angle of attack varies down the keel.

Tim B.

 

Thanks for the input guys.

I don't have abbott & doenhoff but I'll see if I can find a copy. Averaging a couple foils (like the 0012-64) was sort of my line of thinking... start with a 0012 section toward the front, fair it into wherever the max thickness amount and % back of the lead keel currently is, and then fair it back to the trailing edge. I figured this would not result in a averaging of CL,CD, but with the 0012 section it should be fairly close to a true 0012 at the end... maybe with the max draft ending up around 35% though, or with a slight flatter spot between 30 and 35%. This seems to be the best compromise between work involved, reusing what I have to start with, and developing on top of it.

As for laminar root vs. turbulent tip... I did notice when I held up my rough 0012 templates that the tip seemed to be fatter and fit differently than the other 3 going up toward the root (they all fit similarly). Maybe this was an original attempt to get more lead in the bottom and a more turbulent section for larger AOA as you suggested.

I'm not going to get this thing perfect regardless (I'm no nav arc, no analysis tools etc. just workin on my boat), but I'd like to avoid the major no-no's. I realize the first 1/3 or so of the chord is most important to get right. So, couple more questions:

1) Is having a flat spot of any sort to be avoided at all costs? I.E. should I fair it so there is a single particular widest point whether it be at 30% or 35% or 40% and then fair or average the curves so that it works... or should I use an exact 0012 foil at the front with max at 30% and then let the rest of the foil fall in where it best fits to the 0012 template... even if there is a little flatter spot because the max thickness of the 63-012 is (probably) aft more and the same amount as the 0012.

2) The current keel has a hollow toward the trailing edge, what is the effect of a hollow? Is it desireable, undesirable, or doesn't matter? Fill it or not worth it?

3) What is a turbulator and how/why would you build one in? Haven't heard of this being put on a boat keel...

4) What is a spline vs averaging? Spline = more abrupt but still faired transition area between two different sections? Averaging between 2 sections seems fairly straightforward.

 

I have averaged two foils, as you're suggesting. In a perfect world you'd average on perpendiculars normal to the surface, but I think averaging offsets suffices. You need a formula to average the two nose radii together, though. I have the formula at home and I'll post it here tomorrow.

 

From my experience over the years with somewhat similar boats (X-3/4 ton, X-99, Albin Express and J/24's) I would say that using NACA 4-digit sections (like 0012) on the keel is never fast in any condition even if calculations tend to suggest that they would be superior upwind, especially in difficult conditions. Moderate laminar sections like the 63 and 65 series or even better custom designs adapted for the actual load and Rn-range simply works much better. The 4-digit sections are suitable for rudders though.

 

The laminar flow sections require smooth surfaces to achieve their low-drag properties. With bottom paint, etc., you may not be seeing much laminar flow at all. The NACA "standard roughness" curves are probably too rough, but give some idea as to what you might be seeing.

I think you ought to take the offsets for the keel you have and plug them into XFOIL. Then see what characteristics it has and where you might want to improve it. XFOIL has a capability (QDES) specifically intended for changing one part of a section while leaving the rest of it alone.

 

Plane profiles with large hollow parts are called auto-stable in french, but I'm not sure how this works in turbulent water flows. That effect would be great for cruise, something like a longer keel.

The 63-66- NACA series have hollow trailing edges, too (smaller, 90 to 100%). Would you keep them for a keel or shorten the last 15-20% to have flat surfaces?

 

I don't know exactly where this came from - possibly my own active imagination. Nevertheless here's the formula I've used to average nose radii when averaging together two foils:

P=First nose radius
Q=Second nose radius
N=Nose radius of averaged foil

N = P{[(Q/P)^0.5+1]/2]^2}

Let's test it by averaging together a 0008 and a 0012 and see if we get the proper nose radius for a 0010

0008: 0.70*c
0012: 1.58*c

N = 0.70c{[(1.58/0.70)^0.5+1]/2]^2} = 1.10c

Right answer.

 

I would keep them for a keel. The only modification I'd make is to insert a small wedge in the coordinates to give the trailing edge a thicker, sharp-edged square shape. This will make for a stronger trailing edge and won't have a measurable effect on performance.

A lot of people seem to make a big deal of the hollows. They are simply there to give the desired pressure distribution. The NACA 6-series sections have a design pressure distribution that is flat from the leading edge back to the second digit in the designation (30% chord for 63-, 60% chord for 66- sections, etc.). From there it descends in a straight line to the trailing edge. That's it. To get that straight-line pressure recovery requires a slightly hollow shape.

The effect of turbulence and roughness on the keel is to cause the boundary layer to transition from laminar to turbulent flow closer to the leading edge. If it doesn't transition by the time it gets to the end of the flat "rooftop" section of the pressure distribution, the increasing pressure of the pressure recovery region will quickly cause the laminar flow to separate, whereupon it will transition to turbulent flow and reattach, forming a laminar separation bubble. The NACA 6-series section characteristics are not very sensitive to just where the transition point is.

No averaging is required. The leading edge radius for NACA 4-digit sections is defined to be:

rt = 1.1019 t^2, where t is the maximum thickness ratio.

The thickness distribution of NACA 4-digit and 5-digit sections is given by the formula:

yt = t/0.20 * (0.29690*sqrt(x) - 0.12600*x - 0.351060*x^2 + 0.28430*x^3 - 0.10150*x^4)

Plug in the value for t and splice the yt(x) curve with the rt circle and you have the exact coordinates of any NACA 00XX section.

There's nothing particularly special about either the leading edge radius or this formula. It's simply the definition of the geometry they selected for their wind tunnel tests. The NACA arrived at this formua by curve-fitting the thickness distributions of a number of sections (like the Clark-Y, Eiffel and RAF sections) that all seemed to perform pretty well. Like the hull forms of the Delft series, they were simply creating a family of geometric shapes for the purposes of performing a series of systematic experiments.

It wasn't until the 1-series and the later 6-series sections that they used an inverse method, specifying the aerodynamics first and deriving the shape that gave them the desired characteristics. That's why we don't have a formula for the 6-series sections. You'd have to plug in the parameters for the pressure distribution and rerun the program. Which is eminently practical today, but amounted to a heroic calculation in the 1940's when a "computer" was a person (usually a woman) armed with a hand-cranked calculator.

 

Thanks.

Perhaps it's the surface/volume ratio and increased friction that worries some people, altough the gain would be far less than 1% in the NACA 6. (No computation here, and I haven't found any way to automatically enter the values of my custom foils splines from autocad into Xfoil to estimate the loss)

Another "gut feeling": I'd use smaller nose radii and slender (than a 65!) leading part in water than in air - in clear contradiction with my last justification!

Nothing scientific here, but I wonder why the change in fluids compressibility doesn't affect the profiles used in air or water?