XFOIL for hydrofoils

hi guys,

i'm trying to find better sections for windsurfer fins [racing].

is there anything i have to watch out for when using XFOIL for analysing hydrofoil sections?

so far i have been using a GUI for XFOIL called XFLR5 [ http://mapage.noos.fr/xflr5/xflr5.htm ] but i'm sure i will have to learn to drive the original some time as XFLR5 does not have the inverse design routines.

i have read a bit in Epplers book about his hydro sections and he refers quite often to the CPmin envelope. could anyone please explain why that is so important for hydrofoils [cavitation?] and what are the critical values to look out for?

any help is greatly appreciated
thnx
boogie

 

Cavitation will result in flow separation, extra drag, and possibly stall. See http://www.tspeer.com/Hydrofoils/Subcav/Subcav1.pdf

However, for a sailboard, I think the bigger problem is ventilation. See http://www.tspeer.com/Hydrofoils/Subcav/Subcav3.pdf

The other thing to watch out for is transition occurs earlier in water than in air, because of all the critters and particulate matter in the water. This can be simulated by setting Ncrit to 3 or even lower.

 

I was wondering about the ncrit-setting, so last week I mailed a leading yacht design company, who I know use Xfoil, to ask about their recommendations. I have read that ncrit = 1 might be a good value. Their answer was that they had done some research and were - understandably - not willing to reveal the results, but they indicated that ncrit = 1 was "a bit" to extreme (how much is "a bit"?). But they also mentioned two papers by Vijgen et al, that might come in handy:

Vijgen, P.M., van Dam, C.P. & Obara, C.J.:
- Turbulence and Heating Effects on the Design and Performance of Low-Drag Keels
- Turbulence Effects on the Design and Performance of Low-Drag Keels

I also contacted an english aero-/hydrodynamicist. He said that the above papers (I haven't read them yet) siad that the "critters and particulate matter" were of such a small scale, that they didn't really matter.
My own guess is that if you are only doing comparative studies (to see what foilsection is the best), the ncrit-value doesn't matter that much.
Tom: Am I right about this?

Søren Flening

 

Computers arent the answer to everything.

Ok - so you got a program and are trying to work out what will work best (fastest) while giving you the control you want.

My suggestion to you is to analyse nature (fish dolphins sharks etc) to see what works best for various situations...

If Sharks and Dolphins were the fastest fish in our oceans, then tuna's, marlins etc wouldn't exist, they'd all a been et by now!

So what fins do the survivors who can outrun a Shark Dolphin etc use.

Does a sailfish's fin stay "up" all the time?

Does a tuna's fin stay erevct while swiming at WOT (Wide open throttle)?

Answers no - they fold away into recesses in the outside profile of the fishes skin...

You want the fastes sail board fin - you want one that basically folds away to almost nothing until it's needed to either hold course or alter course, maybe it's vector thrust actuated...while running ahead of the wind there is no fin drag when heading across wind etc the fin deploys to varying degrees depending on side slip, when maneouvering again the side slip deploys the fin...

Thats what the fish do...fold them fins away for, adrenalin powered flight response and use em extended for tight turns etc when rounding up baitfish...

When you can emulate that with your sailboard fin, so that it works when needed and isn;t there when not needed - THEN the 'puter might be some help in solving your problem....particulate matter in the water or not.

Emulate nature, analyse some of those fins on the oceans fastest fish.

Maybe the fin should have some kinda "beak" out front like a marlin or sailfish?...or maybe it should retract like a tuna...thats fo you to work out!

That what you should be analysing with your puter - no some theoretical scientists ncrit value best guess IMHO...

The grand geometrician of the universe designed some of the best designs you'll ever see and he didn't write it up as a science paper - he dressed it up as dinner so you couldn't help but notice and learn!

Downunder we call it CDF.

Thats colloquial slang for 'common sense!'....

Play with the puter when your own minds already plumb worn out first...

Use the two analysis tools God gave you in your forehead (eyes) to work out what nature does - then use the grey matter god gave you to find a way to emulate it in an engineering sense is my suggestion.

The puters just a tool to help - once the eyes and brain have learned the lessons nature has to teach.

Least thats my 2c...

Cheers!

 

Trouty: You are absolutely right when you say that it would be beneficial to hide the fin (more or less) when it is not needed, and I would advice Boogie to do as you say. But: He specifically asked about sections!
Ironically, one of the authors of the papers I mentioned (C.P. van Dam) once did a (scientific!) study of fishtails. It was reported in NATURE (vol. 325, 29 january 1987, pp 435-437) under the title "Efficiency Characteristics of Crescent-Shaped Wings and Caudal Fins" (caudal fins being the tail fins of fish and aquatic mammals). His studies concluded, that "... backward curvature of a wing improves induced efficiency to a value greater than that of the flat untwisted wing of elliptical shape considered optimal in classical wing theory."
But this has nothing to do with section shape. A folding fin/centreboard won't help you much if the sections are bad, so why not use the 'puter to assist you in finding/designing the best section? One could even cut up some fish tails and fins, measure them and enter them in the 'puter. It would be interesting to see whether they are much better than the man-made shapes...

Søren Flening

 

thnx for the quick response Tom,

i have been visiting your site quite a few times before, but i never got around to actually read all the scientific articles you got on your site.
excellent site!
the subcav1 explains pretty well the function of the CPmin envelope, now i just have to digest it.
i will try with different settings for NCrit in XFoil and have a look what changes actually happen on the plots for the sections i have been analysing so far.
as soon as i find the time i'll have a closer look at your lifting line analysis spreadsheet too.
you are right mostly it's ventilation, but this article and the low speeds the cavitation is happening in them, made me think:
http://www.fluidlab.naoe.t.u-tokyo.ac.jp/Research/CavPictures/



for trouty and others:
i'm not at all theoretical about my development, rather the opposite. and the final proof is always the test on the water or rather the race course.
my fins helped guys to win 3 worldchampionships, so there must be something right. finding other sections is fine tuning in this case, i'm not trying to reinvent the wheel.
i'll try to attach a picture of one of my fins below. you will see that they actually have a fish fin like shape and one of my favourite pasttimes is to study nature and what they have come up with.
there are quite significant differences in the application of fins for fish and for windsurfers though and it would be interesting to see the shapes that mother nature would produce could it use carbon fiber for their bones and skins...
despite my fins being made from carbon fiber they still flex up to 150mm in the tip, so they are not quite as rigid as you might think.

i would love to have a very close look at some tuna fins, but would not like to kill the animal for that.
i don't have the budget either to get a large number of CNC machined moulds made just to trial and error with shapes. i rather try to gather as much information as possible and try to get more dots in the picture [which is still pretty blurred]. the final decision what shape to go for is a mix of gut feeling [computed by the brain] and info from my dumb helper, the computer.

you might not be aware of the steep development happening for course racing windsurfers, but the fins we are talking about here are 700mm long, the board is 1m wide but only 2.5m long, weighs about 8kg, no centreboard, and sails up to 12.5m^2. these boards plane in 6-7kn of wind and go around a windward-leeward course at a speed most sit-down sailors only dream of.

thnxfor all the comments
cheers
boogie

 

Hi Boogie

I grew up in a small coastal fishing villiage named Lancelin on the West Coast of Australia just north of the Capital city Perth.

Today Lancelin is world famous as wind surfing capital of Australia - with big fooshore race every year for wind surfers. (Lancelin Classic).

Wjhen i was a kid growing up there, windsurfers had not been invented - indeed surfcats were a new thing way back then...

Interesting fin you use - much different to what I'd envisioned.

I would think - a connection between sail angle (relative to board direction) and tilt on fin up into recess in board might be your answer...

When sail boom angle approaches 90 degrees to board, fin is fully extraced as you run downwind, and as sail angle decreases, fin extends for control proportional to sail boom angle relative to board...

Least thats how I see it...

Could you also borrow a leaf from ben Lexcens Book - and use a winged keel to get less slip at steep heel angles? i.e. have two fins, on a upside down "V" shape, so that wyhichever way the board heels at least one fin is verticle for minimum slip?

Preferable - would be to have your retractable single fin - heel to verticle whether board heels to port or starboard - ie again, some connection between sail mast angle relative to board......

Any side slip you can convert to forward motion must give you better speed, and keeping that fin virticle all the time might mean it could be shorter in length with less drag because it always remains verticle.

Just an idea.

Cheers!

 

hey trouty,

here are a few more dots for your picture... ;-)

in light winds we sail twice the wind speed. this is only possible because we sail the apparent wind. we never sail downwind in a straight line, but tack downwind on a broad reach. even on this course we never sheet out more than a few degrees.
there is no need to optimise the board for reaching condition as we only race up and down races.
the flex of the fins is actually built into the fins by purpose, not because we can't make them stiffer. especially in light winds softer fins [despite the loss in projected sideways area and force vectors being off angle] actually go better because they trim the board in a way that frees them up and helps to get and keep planing.
there are quite large loads on those fins and you are going fast, so making a retractable fin that works under load with no hands available to adjust it is quite a challenge.
windsurf course racing is a game of high average speeds not peak speeds.

cheers
boogie

 

Thanks for the kind words!

I don't have any test data to say what Ncrit should be. I based my recommendation on a talk that Laurie Smith and One World's ex-Boeing hydrodynamicist, Winfried Feifel, gave to SNAME here in Seattle a few years ago. Feifel was describing how one didn't get the laminar flow one would expect in water when compared to air and impact that had on their bulb design. There have also been some posts on the subject by Dr. Drela on the Yahoo XFOIL mailing list - you may be able to find them in the archives.

 

hi tom,

i understand that a lot of the parameters [like NCrit] are rather educated guesses, but at least it's a hint in the right direction.
the fins i'm working on are being used in such a dynamic environment that a lot of the theoretical analysis is probably flawed from the first keystroke anyway.
on the other hand i recognise quite a few of the characterisics of the different sections in real world sailing/testing. i guess it's an advantage that i sail competetively myself and don't have to rely solely on external feedback.

the Re numbers on the fins are quite low. i'm working with a range of 500k to 2 million for the different speeds and chord lengths. with the most time spent around 1 million.
despite the speeds being quite high [14-16kn upwind, 20-25kn downwind with peaks as high as 28kn. GPS measured] the chord lengths are quite small at around 125mm at the base and ~60mm close to the tip and have to stand quite high shock loads in choppy conditions.

do you think this is still a range where low reynolds number sections would work better? is excessive laminar flow and laminar separation still a problem at these Re? or is this offset in a way by the rather more turbulent freestream?

lot's of questions, i know. it would be great if you or someone else had the odd comment though.

cheers
boogie

 

In general, you can go up from the design Reynolds number range, but you shouldn't go down. 500K is a range where XFOIL works particularly well. It can handle the laminar separation bubble, and there's not the massive separation you might expect as you approach the fully-laminar regime.

The key to handling the laminar separation buble is to use rounded pressure distributions in the transition zone. That way the separation bubble is short and moves steadily forward as the angle of attack increases. Then you get the best of both worlds - lots of laminar flow for low drag at low angles of attack, and more turbulent boundary layer for resistance to stall at high angles of attack.

What you want to avoid is a transition point that suddenly jumps to the leading edge - that's a sign of forming a leading edge pressure peak. The danger with a sharp pressure peak is the increase in pressure on the downstream side can be steeper tnan the pressure recovery in the separation bubble after transition occurs. Then the bubble never reattaches and you get leading edge stall. So as long as you control the separation bubble, there's no such thing as too much laminar flow!

In XFOIL, I find it useful to design the leading edge using MDES. You can change the design lift or angle of attack, letting you flip back and forth between high and low so as to see if you're forming a pressure peak at one condition or the other. If you go to an angle of attack a little beyond your target you can clip the worst of the peak there, and when you come back down to your design range you'll have a nicely rounded peak. This is likely to add a bunch of thickness, though, so you have to go over to GDES to deflate the section a bit. This will undo some of your nice work in MDES, but by alternating back and forth you can home in on a good design.

If Ncrit really is lower for hydrofoils, this leads to earlier transition in the bubble and less succeptibility to laminar separation than for airfoils. You might be able to use that to your advantage by allowing your leading edge to be a bit more peaky than you'd otherwise be able to get away with. Since you're exerimenting on the scale of board fins, it's not that expensive to take a risk and see what works!

Cheers,