12ft skiff dinghy centreboard discussion

Hi, I thought this might be a better place for this topic than a thread I posted to yesterday.
[ http://www.boatdesign.net/forums/sailboats/best-foil-shape-rn-250-000-750-000-a-1577.html ]

The Centreboard is for a 12ft skiff (see link in sig. for more details!)
Construction would likely be NC machined moulds with a carbon skin and foam core; this gives so much design flexibility that the problem really is about the theoretically best shape. In particular I'm interested in what the best section to use would be, considering there should be many more choices now than the 1960's

I've looked for information fairly extensively on this topic but information is relatively sparse.

I've been looking around and have done a couple of quick calcs to guide this.

Problem: At the moment the foil I'm using is nominally a NACA 0012 but with a LE rad about half of what it should be. I feel that it's stalling out of tacks and when accelerating off the start line, and limits our ability to pinch due to excessive leeway, esp. in light winds.

After looking through some information, lift-aoa plots and a few simple calcs, the theory also points to this. Current section probably stalls out around Cl=0.8 or less

Existing board:
Chord is ~300mm, span ~1350mm, area is 0.29m2, actual section shape probably resembles a NACA 0012-34
The planform is a highly tapered (reflecting a bending moment optimised planform a-la Munk)
I feel the area and section thickness is about right.

Here's some numbers I've estimated:

Side force is 718N in most conditions, so the Cl varies wildly depending on windspeed (i.e. boat speed) and whether we're going upwind or downwind.
In stronger winds, the Centre of Effort is lower in the rig due to more twist and shorter masts - the side force would increase to say 870N.
Drag at High Lift is not too important, so long as the section doesn't stall.. but low drag at low lift is important

V(kts) Cl
4 1.2 (accel. out of tacks, start etc.)
8 0.3 (lighter upwind conditions)
10 0.2 (heavier upwind conditions)
11 0.18 (fastest common upwind speed, with higher side force due to smaller rig)
17 0.06 (offwind)
17 0.045 (off wind, one only on trapeze)

These values are only relative to planform area, I haven't adjusted them for effect of the finite aspect ratio. It seems the section shape needs to accommodate slightly higher CL values than current to achieve the low speed targets. I'm guessing the current section probably maxes out at Cl ~ 0.8 or less.

So the sweet spot for upwind efficiency would be in a Cl range of 0.3 - 0.18, with Re centred on 1 - 1.5 E6 ish.

Reynolds number: Offwind Re. up to 3.3 E6; foil tip Re could be as low as 200,000 when tacking, etc. assuming a tapered planform.
Low drag at higher Re and low Cl will be helpful.
We can lift the board a little in stronger winds to reduce surface area and limit 'tripping' due to a deep board (as much a seakeeping issue over outright speed)

I'm wondering if there's a section that has not so much a drag 'bucket' like the 6-series foils, but a gentler low drag 'cup' - a forgiving foil that's not going to stall early, and where the transition point from laminar to turbulent will gradually move forward along the chord as the board loads up (yet maintain a reasonably robust laminar BL at low Cl, for offwind speed)

I took a look at
http://airfoiltools.com/compare/index
to compare some foils.

It seems that the Eppler 836 stalls relatively early at around Cl = 0.8, as would our existing foil.
The Wortmann sections that Tom Speer suggested won't fit at 15% t/c, and I'm not entirely convinced.

Looking at the actual NACA 0012, it seems to be reasonably suitable.. but I'd really like something more modern with, hopefully, better boundary layer control and less drag. It's been quite hard finding any decent information out about Eppler or Wortmann sections although there is a mass of information about NACA 4 digit and 6-series section. Especially any notes on their intended use or special properties.
I've heard Eppler sections are good for low Re applications and Wortmann are centred on medium Re applications.

Anyway, I went searching for a few foils and came up with the following list of reasonably similar foils:

NACA 0012
Eppler EA 6-1-012
Wortmann FX-76-120
Wortmann FX-79-L-120
12% Joukowski

The FX-79-L-120 looks promising at a hunch, but I've no further information on this section.

Now before everyone says 'use Xfoil', I'd prefer to use a section shape that someone with actual expertise has come up with. If anyone has made foils using Eppler or Wortmann sections (like above) I'd like to hear about it. I'm happy to have a crack at Xfoiling a pre-existing section, out of interest.

Can anyone shed some light on these sections? Are they any good and what might be better??

Thanks,
Ben

PS to Eric Sponberg: I noticed a (possible) error in the article 'Keel and Rudder Design'. You suggest narrowing a keel of fixed depth increases its aspect ratio. While this is true, the drag may not reduce dramatically in the way 'figure 6' suggests. Reduced surface area will be a benefit, however the induced drag will remain the same for a given load at a fixed span, no?
Also max. surface roughness for foils I guess would be 0.004" or 0.1016mm on p 15

 

Have you looked at Selig's SD8020?

 

No - thanks for the tip! I hadn't seen that one but I've just looked it up. It looks promising, the lift and drag polars look good... I've just looked at the quick 'comparison' on airfoil tools website between the SD8020, FX 79-L-100, NACA 0010 (and NACA 0012). Due to the scale of the plots it's hard to see if there's much drag difference at low angles of attack.

What can you tell me about the Selig 8020 foil? It looks very close to the NACA 0010 foil.

I've just looked around the net and found a few notes.
'SoarTech 8' at http://m-selig.ae.illinois.edu/uiuc_lsat/Airfoils-at-Low-Speeds.pdf looks like the original paper.
They concentrate on reynolds numbers that are very low and there's lots of talk about separation bubbles.
I guess one priority is highest lift at low reynolds number. Now I've got the idea of separation bubbles etc., foils that perform well at low Re are in fact more important than I would first expect... because at higher speed most foils will give enough lift (assuming the area is enough to stay within acceptable leeway limits) so from there it's just a matter of getting the drag down.
It's interesting to look at the Cl / Cd polars and how it changes from say Re = 100,000 and Re = 500,000 due to the presence of separation bubbles.

Is there anything with a T/c of 12%? 10% is ok (could even do a gybing board) but the laminate will be that much thicker.

 

The relationship between induced drag, span and area was discussed at length in http://www.boatdesign.net/forums/hydrodynamics-aerodynamics/myth-aspect-ratio-36836.html

 

Thanks that'a a pretty epic thread,
along with the referenced thread http://http://www.boatdesign.net/forums/sailboats/tandem-keel-1058.html
it will take me a while..

My first thought is that I want the highest effective span possible, ie ditch the highly tapered planform, optimised for bending moment, in favour of an elliptical planform (or one with taper ratio ~0.4). While there's no hard limit on draft there are practical limitations, ie launching and the ability to sail further into shallow sections of the course, and durability when running aground etc.

Another point which reading (just a small section) of this has highlighted, is that induced drag is less important as speed increases (because Cl is decreasing rapidly). So skin friction & profile drag become more important with speed. So if you take the tapered (Munk) planform - keep it optimised by bending moment - but now add weight for a skin friction drag component. I'm pretty sure this has been solved for a fixed airspeed and loading - the resulting planform is now a compromise between an ellipse and the munk planform (I don't remember who showed this)

Now if you consider that during a race, and across many days of a regatta, your boatspeed will vary greatly.
This means the planform will also be a compromise between efficient lift at lower speeds but also low drag at low lift. I think this would skew the optimal planform even more heavily in favour of reduced skin friction / surface area. So in the end, an elliptical planform may be close to the best even when taking into account overturning moment

 

A NACA with a pointy nose will give you all the disavantages of a laminar section with out giving the advantage of a low drag bucket.

Try an accurate NACA 0012. They perform very well at Re=0.5 to 2M+

I did that with my windsurfer fin using a NACA 0009. I was able to get the board on plane faster and point higher than with the laminar foil section fin sections that are common. Reaching and running seemed unaffected.

I can also Reccomend the SD8020 but have not used it on fins. It is designed to function best at the lower end of your reynolds number range though. I have used it on radio controll glider tails.

 

Pay attention to interference drag and separation at the fin-hull juncture. Trouble there could completely wipe out any advantage of a special planform.
Also design for effective use of your variable area option with retracting and extending the fin.

Putting these pragmatic aspects in the mix cooks up a planform with untapered chord 2/3 of the fin span and then a 0.5 taper to the tip in the last 1/3. (more taper invites tip stall issues at low reynolds numbers)

This would allow you to retract the fin to 1/3 of the draft and less than 1/3 of the wetted area while maintaining a good hull-fin fit.

BTW a planform with the outer 1/3 tapered can be very close to the elliptical planform in terms of induced drag. Especially at these lower Re.

 

Thanks for the input John.
The more I think about it the more variables come to mind! There are many other considerations and some background I haven't written down yet, so at this stage I'm trying to patch up the holes in my understanding of the system.

Solutions for this class of dinghy can be surprising due to their uncompromising nature, so it pays to think of all aspects.

I'm not too worried about the hull - board junction at this point, the hull is essentially a flat plate where the foil exits. I'd consider a small fillet or ramp at the leading edge but it's probably not worth the effort. I think the section is narrow enough that the horseshoe vortex around the leading edge would be small enough to neglect. The current practical concern of the junction is the gaps in the bog from when I fitted the board Over about 15 knots it sucks air out the bottom of the case which aerates the planing surface and rudder. I've never noticed it cause trouble apart from my inattention from theorising about bubbles when I should be watching where I'm steering!

In practice the variable area option is probably only good for maybe 3-4 inches, as the top of the board impedes tacking & gybing. But I may be able to do something clever with cutouts etc.
Having a parallel section is a good idea though, it also allows the board to be cut shorter and still fit well, if needed!
Another option along those lines, is possibly to reduce the chord and section thickness slightly, and use the extra 'slop' available in the case to have a gybing board in lighter weather. Would need to be careful to maintain enough lift during tacking though.
Another possible problem with raising a lot of board is whether the boat would become too twitchy (less roll damping). I once sailed another boat, with smaller foils, in heavy weather, and it was wild... hardly a controlled comparison though.

I thought a taper ratio of 0.4 is roughly equivalent to an ellipse (ie tip chord is 0.4 of root chord)? If so I would have thought a parallel section to 2/3 then the last 1/3 only tapered to half would make for higher induced drag than necessary...? How does the Re affect the foil loading / induced drag?

My thoughts exactly about the section shape btw.. all the disadvantages of a laminar flow section combined with the disadvantages of a 4 digit section!
What makes the current foil even worse is the high taper ratio (tip is 95mm root is 320mm) AND the section shape slims from 12% at the root to 9% at the tip! The tip will already be heavily loaded.. then the sharp leading edge...

I'm interested to hear more about tip stall though.. As mixing up some bog and re-profiling (back to the original 0012) is a lot easier than constructing new moulds. But bog can't fix the planform.

Keep in mind though, that offwind I can still get RE as high as 3 x 10^6. Any ideas on how the drag of an SD8020 would compare to a NACA 0010 or 0012, at low AoA?

Also, does anyone know how a Wortmann FX79-L-100 (or FX79-L-120) would stack up? Would they do better at high Re to offset their worse performance at low Re (?!)

 

One quick and dirty way to look at a planform is to just lay it over an elliptical wing of the same area and span. Roughly, where Chord_ellipse/Chord_wing is greatest you can expect the stall to occur. Also look with understanding that lower Re areas are also more prone to stall. A perfectly elliptical wing will usually stall near the tip because of Re number effects.

http://www.betsybyars.com/guy/soaring_symposia/71-wvu.html

As for the airfoil comparison just plug them into Xfoil and take a look or Google search for wind tunnel results. NACA 824 has 4 and 5 digit NACA wind tunnel tests at 3 million. It might be a good idea to estimate the percent difference in total drag that you will see going from one airfoil to another. It could turn out that a section that prevents 1 sec of stalled fin has more benefit than using a section that might give 100 sec of imaginary laminar fin flow.

 

Can't it be easier?!

So I've looked more and more... the more I find out the more annoying it gets! Every other paper condradicts the last paper etc etc.

Currently thinking
NACA 0012
FX-76-120
Selig 8020 (scaled to 12% for strength and a slightly higher Cl..?)

A paper on the wortmann sections at low Re here:
http://aerosociety.com/Assets/Docs/About_us/HPAG/Papers/HP_aerofoils.pdf
Interesting because it talks about performance in the range of Re that I'm motsly interested in.

Seems testing at low reynolds numbers is usually done in laminar flow tunnels..

The Wortmann FX-76-120 is very close to the NACA 0012.. Slightly more front loaded I guess. The design has this range of reynolds number in mind. These sections deal with the laminar flow problems of low reynolds numbers better and Have high CL's around 1.2 ish in the range I'm interested in. High CL's at low speed are good as it reduces the surface area (chord) needed and therefore skin friction at higher speed.

It's hard to find info on the drag of these foils at higher RE and low CL's (ie for offwind performance). But as far as I can tell, the FX-76-120 and S 8020 are reasonably low drag.
It seems that a laminar flow section (ie NACA 63 012) may not provide a large benefit offwind anyway, as laminar flow apparently occurs depending on the mood of the water etc etc. Any possibility of a separation bubble (even at low Cl's) would destroy the advantage of the laminar flow section. The risk of leading edge stall / laminar separation upwind would never seem worth it. (This may not be the case for larger boats)

As far as planform shape goes.. yes tip stall is not great. Seems the elliptical planform isn't actually the go, John as you say we may want a more squat ellipse. Or a truncated ellipse. Or a trapezoidal planform with a 'low drag' wingtip (whatever that is!)
A pointed, elliptical wing-end may only end up with prematurely separated flow.. ?
Taking it to the extreme you could have a different section for the tip, to suit the smaller reynold number. But this seems a waste of time as the range of Reynolds no's is already large, so a conservative section will be the best for the whole foil anyway.

Tempted to have a slightly raked, straight leading edge. This would make for more accurate surface generation and also more accurate fairing / polishing of the leading edge of the foil. Keep any curvature in the trailing edge; not straight as is the latest trend.
The rake of the leading edge could be say 5-6 degrees (giving a slightly raked quarter - chord). Again, keeping a straight leading edge may give slightly less drag than curved. A curved LE profile ends up quite raked at say 75% of the span, which could disturb the boundary layer more.. creates cross flow instability leading to premature transition to turbulent boundary layer.

A few papers on 505 boards:

https://docs.google.com/file/d/0B3Zekt2OTzO4VHp0dk5PckhtOFE/edit?pli=1
https://docs.google.com/file/d/0B3Zekt2OTzO4SEVKc0VfQU5HMjA/edit?pli=1


This paper is kind of interesting but I don't understand why they changed the profile from tip - root as well as a few other points:

http://www.google.com.au/url?sa=t&r...W-HwKzZm44aSjUMgFLriIDw&bvm=bv.66111022,d.dGI

 

Note not everything frm a 505 is relevant to a 12'...
https://docs.google.com/file/d/0B3Zekt2OTzO4VjY1NUQwQ1RiaXM/edit?pli=1

This paper seems impressive at first but on looking further has some serious flaws when it comes to designing a foil that can be fast around a course. It never defines the 'rectangular' planform or the 'trapezoidal' planform, it's pointless quantifying an 'improvement' if you control foil is not told. The trapezoid looks highly tapered and is not raked, so the tip may drag unnecessarily and may develop lift more slowly than a slightly raked board.
The final planform has the root chord at the max limit, so the result is peculiar only to that centreboard case.

They have assumed that L/D is a singular target, optimised for a single sideforce. The plot of 'L/D improvement' stops at low angles of attack. Total drag is never mentioned. The planform shows a small improvement in L/D at high angles of attack, but there may be more improvement here using a thicker section. Also it doesn't show max lift at high angles of attack, which is vital when starting, tacking, and rounding the leeward mark. Also there is no allowance given to no - load conditions seen on other boats when running.
There is no attempt to explain why the planform may be better apart from 'it looks like a bird'.

 

In the windsurfer fin experiments I concluded that the advantages of using a template to put over the leading edge was nessesary if the foil section was going to be accurate enough. I went with a rectangular planform so I only had to make one set of templates (one that fit over the nose of the airfoil, and one that was shaped to check full-chord top and bottom, NACA 0009) The reason for such a crude planform being ok was that, at the Cl I expected, the additional drag was minimal compared to what would have happened if the foil section was wrong.

The most important thing here is to be aware that there comes a time to kill all the engineers and actually build something. Anything you build will work better than anything not built......

Make this rectangle board and try it with the mindset that you will be making another one soon. You will learn a lot and have a better idea of what to do on the fancy schmancy one...

 

I quite agree - I was about to post something very similar. I would at least keep the LE straight for simplicity of construction. And the possibility to use a template/templates on the LE. Even rectangle is not bad.

I wouldn't also expect miracles from your new CB - like a 3xgold-medal-winner friend of mine used to say: "90% of boatspeed is the driver".

 

Thanks - yes I'm keen to keep a straight leading edge for those reasons, even if the planform tapers, and keep the foil relatively simple.
It seems a correctly proportioned, trapezoid shape foil with (say) an accurate NACA 0012 section and smooth tip would give 90 - 95% of what's practically possible.
Unfortunately it takes much much more reading to understand that something so basic is almost as good as it gets.

I'm going to try a look at Javafoil and have a crack at comparing different operating scenarios and aspect ratios etc.

I also found this thread informative, which I hadn't seen before:
http://www.boatdesign.net/forums/sailboats/rudder-daggerboard-profiles-937.html

Part of the problem about building again is the centreboard I have should be fine if you believe the hype.. one of the other boats is considering changing back to his old board for the same reasons I've noticed.
So it's not worth building something unless I actually do it right, or about right, and understand why.. then I can go changing things!

There's a huge variety of shapes and sizes of board in the class and they all work.. from 15% laminar sections to 10% and also a variety of planform shapes and areas. Choosing a 'baseline' board is not that obvious.

 

Exactly my point. Maybe one reason for great theoretical profiles & planforms not delivering the performance difference you would expect lies in the nature of the sea itself, and it's turbulence level. Below a RANS simulation of a keelboat underwater hull with 3 different turbulence intensity (Ti) levels - the laminarity starts to vanish at Ti 0,5% and at ti 1% it's practically all turbulent. The first one, Ti 0,1%, is very mooth indeed, it's the turbulence value they strive after in low turbulence windtunnels, for low Re airfoil testing.

Boats passing in the water, the waves they generate and wind generated waves all create eddies & turbulence, which combined to the relatively slow speed at which boats are moving through the water, easily create a Ti of 2%-5%, have I understood. The turbulence from wind waves penetrate at least the height of the wave below the sea surface measured from the wave bottom, so at least the hull bottom and upper part of the keel & rudder are affected. The lower tips again experience higher motions due to pitching & rolling of the boat passing in waves.