Sailboat keel section choice/options?

So much agreement on here, lol. FYI while the Santana 35 keel is a 63A012, the Soverel 33 is a 0012.

 

I've heard a lot of experienced people caution against small nose radius foils. I also know a designer with an aerodynamics background, Richard Roake, who tried quite a radical foil on a Mount Gay 30, with unimpressive results. Most keels are stalled immediately after a tack, and the flow on the leeward side must reattach as the boat accelerates (made more difficult when the boat is pitching in waves). This generally makes blunter noses better in sloppy seas and tactical situations.

One recommendation that seems to be bourne out by tests is that the keel thickness should be 10% or slightly less when possible, making me think that reducing thickness generally might be better that going to a small nose radius foil in many cases.

The 15% versions seem to work for ballast bulbs, where low drag is sought rather than maximum lift/drag ratio. A fine nose foil might also be good on an older style one-design that has a whopping keel area, and therefor a smaller typical angle of incidence. But if you want to reduce drag on a modern boat I think you're often better off with less keel area (by reducing chord length) and a more conservative foil.

 

Well, I see that my use of "nose radius" was inappropriate, the E386's is bigger (0.0198) than the 0012's (0.0172), altough I would describe the former as more "pointed"

For my 9 meter cruising boat the minimal keel volume is too important for 10% or less thickness (that gives too much lateral area, more than 2 meter draft, which is already quite a lot here), this could be solved with a 100% lead keel or a bigger bulb which sticks out from the leading edge (good way to drag all what floats along).

Do you know what's the code for symmetrical Eppler, Wortmann? (I need something better than NACA 00 for the tip, and Xfoil is hungry )

 

Short steep choppy and sloppy seas are di rigour around here when the wind pipes up. I have noticed on several occasions (often after a tack or helm screwup like pinching into a big wave) where the boat will suddenly drop from around 5.5-6 kts to 4-4.5 kts, and it mysteriously takes f o r e v e r to get it back up to top speed. It seems no amount of bearing off and easing the rig and sheets etc. can get it easily rolling again. I have little doubt this is costing valuable time and distance vs the fleet.

My keel is a trapezoid fin, 4.3' chord root, 2.2' c tip, 4.7' span, sweepback angle 22.5 deg. An 0009 section would probably work best, but I don't feel like planing off lead ballast weight, so if the 12% fits I'll fair the first 30% of chord into the current max thickness, maybe leave the hollow in the aft section maybe not?, cut down the trailing edge thickness (currently about 3/16 or 1/4 inch) to around 1/16 and add a 30 degree taper since I've noticed some keel hum at higher speeds, and square off the back of the tip (it is slightly rounded upwards into the trailing edge at the aft corner).

 

Chickadee, Tspeer, others:
Chickadee's reference to E386 foil is interesting.... I'd like to hear more about it...

 

Nose radius is another one of those things that I think people pay great attention to, but is misleading. For the NACA 4-digit sections, the formua for the thickness distribution didn't really do a good job of describing the leading edge itself, so they put in a small circular arc at the leading edge.

But when you design a section by a process called conformal mapping, where the flow about a circular cylinder is mathematically warped into the flow about a teardrop shape, the whole leading edge shape is completely defined. So you don't want to mess with the leading edge by putting in an arbitrary leading edge radius. First of all, the designed leading edge shape isn't circular. And where the leading edge radius meets the rest of the section shape you'll have a discontinuity in the curvature that, while it looks fair, will cause a perturbation in the pressure distribution that will damage the foil's performance.

A modern section is like a mirror. The pressure influence from the whole section is focused by the shape on the leading edge. When you mess with the leading edge shape, you're destroying that focus. Now, if it wasn't especially well focused to begin with, like the NACA 4-digit and 5-digit sections, that doesn't matter much. But when you start messing with a NACA 6-digit, or a Wortmann, Eppler, Selig, Roncz, etc. design, then you are not likely to improve the performance by making an arbitrary geometric change, especially at the leading edge.

The general trend is, the narrower the angle of attack range of the section is, the smaller the leading edge radius tends to be. But you can have a highly tuned high lift section that has a surprisingly small leading edge radius. For symmetrical sections, like a keel, thickness is all you have to work with, so those sections tend to follow the rule better than cambered sections.

I definitely wouldn't put a finer leading edge radius on a NACA 6-series section. Those already suffer from a tendency toward a nasty leading edge stall, and a sharper leading edge will only make it worse. You're not going to reduce the drag by making the leading edge sharper - that's determined by the length of laminar flow toward the "shoulders" of the section. A leading edge pressure spike from a too-sharp leading edge can trigger laminar separation and transition, and increase the drag by eliminating laminar flow on the suction side.

Most airfoil codes assume incompressible flow to compute the pressure distributions, then account for compressibility by correcting the incompressible results. And even those codes that do truly compute the characteristics using compressible flow, you can enter 0 or a very low Mach number. Below, say, Mach 0.5 or so, there's not much effect of compressibility at all, assuming that there aren't some really high local velocities. So compressibility isn't much of an issue when applying airfoil data to hydrofoils.

However, laminar to turbulent transition seems to happen sooner in water than in air. In XFOIL, this can be modeled by changing the value of ncrit - the level of exponential growth at which a disturbuance in the laminar boundary layer is assumed to have become so large that the boundary layer becomes turbulent. XFOIL's default value is 9, meaning a disturbance has to grow by a factor of 8,000 before it transitions to turbulent. A value more like ncrit = 3, meaning the disturbance is amplified by a factor of 20, seems to fit the experimental data better for foils in water. I'm not sure why this is, whether the disturbances in the water tend to be higher to begin with (wave action, perhaps?), or whether it's due to disolved air, solids and critters in the water. Of course, the disturbance level in the NACA wind tunnels used to produce much of the classical data was pretty high, too. So those data may be more representative of what happens in water than data taken in a modern low-noise low Reynolds number wind tunnel.

 

Tom,

Could you elaborate on 'insert a small wedge in the coordinates'? I am bit familiar with XFOIL (certainly no expert) & Profili and understand the rationale for the thicker, sharp T.E.

Regards, RW

 

XFOIL has a function (TGAP) in it's geometric design (GDES) menu that does this. However, you can do it easily in a spreadsheet.

Here's an example. I took the NACA 63-012 coordinates and modified them in Excel using this formula:

Ymod/c = Y/c + sign(Y/c)*Yte/c*X/c*X/c

Yte is the thickness I wanted to add to each side of the trailing edge. I arbitrarily picked Yte/c = 0.0025 to make the trailing edge half a percent thick. The wedge added by the second term is parabolic, so there's no change in slope and no change in thickness near the leading edge. The same amount gets added to both sides of the section, and the change is smoothly blended in along the entire length.

The plots below show the modifications to the geometry, including a blowup of the trailing edge, and the XFOIL predicted performance. The minimum drag is raised by a couple of counts, and the maximum lift is raised by 8%. Considering how much more robust the thicker trailing edge is, I'd say the modification is quite a reasonable tradeoff.

 

Thanks much Tom. That helps a lot.

RW

 

Did you read the article on keel and rudder design in pbo?
http://www.proboat-digital.com/proboat/200506/?=pg1