sail aerodynamics

Thank you for this very interesting discussion about the respective role of the front & rear part of the sail.

-It sparks a candid question regarding the 2D "Design philosophy" for a sail:

In an airplane environment, I guess the front part of the wing ,will be designed in order to maximize the length of the laminar rooftop, among other requirements.

-But for a sail, subject to high level of turbulence in the airflow, which translate into almost fixed transition @5 or 10% of LE;

The design philosophy can't target a longer laminar BL ?

So, in this environment I wonder what should be the "Design Philosophy"

Managing the thickness of the turbulent BL ? Avoiding separation ?

Happy week-end

Erwan

 

Just a small note: the extension of the laminar airflow is not a general requirement in aircraft design either. For example, think of STOL airplanes, which require a reliable and controllable airfoil behavior at high lift coefficients.

For a classic sail with a naked mast, the transition actually occurs at the mast, in most cases followed by the separation. Check the pics in this post: http://www.boatdesign.net/forums/hydrodynamics-aerodynamics/sail-aerodynamics-457-41.html#post745718 .

For the rest of your questions, and especially when it comes to wing sails, prof. Drela and Mr. Speer are the definitive authorities for them on this forum. I'll sit and learn together with you.

 

I should have precised the following:
Design philosophy is, implicitly for teardropmast + full-battened sail
in its simpliest form (uni-rig) to be more or less similar to a single wing.

 

Sorry meant to say it under-rotates.

It is bad aerodynamically because it makes a teardrop worse so you are stuck with more or less round, and because even with round it moves the luff groove around towards the windward side more rather than around towards the leeward side more. The bend characteristics also change but that can be accommodated because the under-rotation is predictable.

 

Perhaps it is less of a problem now as the boom does not bend as much as it used to. I think they changed the dimensions of it. It was originally based on the dimensions of a wooden boom that was relatively tall and narrow, and was wedged into the mast to act as a vang. When they went to a proper boom vang and an aluminum boom it stayed relatively narrow and it bent sideways quite a bit. I think this may have changes when they went to a carbon spar. My brother Fred was the Finn sailor. I was a girly 170# laser sailor back then, but I did follow the Finn development with him and others in the late 70s and 80s before they went to carbon.

Here is a photo from before the switch. I think it shows some sideways bend.
If the wind was stronger and he was hiking harder it would bend some more.

 

Under-rotating a teardrop mast can be a good thing, compared to rotating it to the point where the lee side forms a smooth contour. It's common for there to be a large separation bubble behind the mast on the windward side. When the mast is under rotated, you can get a modest separation bubble on both sides. The two smaller bubbles can have less drag than the one large one.

One good strategy for trimming the mast rotation is to put one or more small wind vanes very close to the mast so the fin of the vane just clears the mast. The purpose of the vane(s) is to see where the stagnation point is, and to rotate the mast so the stagnation point is just to windward of the leading edge. This minimizes the leading edge suction peak that forms if the mast is under- or over-rotated. The high velocity in the leading edge suction peak causes a lot of skin friction that contributes to profile drag.

 

Yes, but as a first approximation, the drag penalty of a severe bubble scales as the bubble thickness times the cube of the velocity over the bubble. When the sail is at a high CL, this V^3 factor will be many times larger on the lee bubble than on the windward bubble. So it's prudent to rotate the mast so as to prevent a lee-side bubble first and foremost.

A lee-side bubble will also thicken the entire downstream boundary layer and thus reduce resistance to separation near the leech, which will decrease the sail's CLmax capability. In contrast, the thickening of the windward-side boundary layer has no adverse effect on CLmax. This is yet another reason to make a smooth lee-side contour the highest priority.

 

The only possible benefit to under-rotating a mast like this is in the case where some separation off the mast is inevitable. In that case, you want to produce a minimum pressure point a short distance behind the mast, so that the lee-side boundary layer which separates off the mast can reattach ASAP before it's hit with the main pressure recovery.

 

Aircraft aerodynamics vs Sailing aerodynamics

I am trying to remember, but I seem to recall it was not an entire article, but more like a news clipping. I'll try going back to my old hard drive storage unit and see if I can find it.

My real question in bringing this up was to bring up the question of differences of interpretations as to what the aircraft theorist might say was the ideal configuration for the best lift-to-drag airfoil compared to what might actually occur with sails on our sailing vessels.

ie: if the aircraft designer was designing the rig on Hall's J-90, he/she would call for the uni-rig in accordance with his theories. But what actually occurs on the sailing vessel in real practice? Perhaps not ALL aircraft aero is directly compatible with sailing aero

I saw a similar thing expressed by Arvel Gentry at one time in the past, but can't find it right now.

What is the aerodynamic explanation for the uni-rig not being the better of the 2 for going to windward?

 

Depends on wingsails dimensions and shape, on the airfoils used, at what windspeed was it tested or raced etc.

A wingsail is not better than a cloth sail by a mere definition. But it can be much better in some conditions if properly designed. A classic low-AR sail can be pretty stall-resistant, while wingsails are generally not. This gets more true as winds get lighter. So if the design of the wing sail was wrong for the racing conditions, it is not so difficult to imagine a situation in which it can be outperformed by a classic sail, especially when matched to low-speed hulls. By the way, when I say "low-AR" and "high-AR" I use common aeronautical wings as reference.

A hypothetical vector diagram which shows such possibility is attached below. Low-AR sail represents a classic sail (main+jib), high-AR sail represents a hypothetical badly-designed wingsail (exaggerated for clarity). Hopefully the diagram is clear enough.

Of course, this is just a conjecture which shows one of many possible cases. More input info would be necessary for an in-depth technical analysis.

Cheers

P.S.
There is a typo in the graph - the axis should read "Lift" and "Drag" instead of CL and CD, respectively.

 

Since you mention Hall's J-90, I'm assuming that we're comparing a conventional una-rig with a sloop rig (not a wing-sail).

At the risk of oversimplifying things, here goes :

I'm attaching the graph from Marchaj's book where he shows a typical drag polar, rotated to show the thrust & sideforce components for various apparent wind angles. (repeating the attachment from post #585, earlier in this thread).

For the 3 curves with the largest apparent-wind angles, the angle of attack for maximum thrust is very close to the angle with maximum CL. Even for the case with the smallest apparent-wind angle (30Deg), the max-thrust point is much closer to the CLmax point than it is to the max CL/CD point. Some boats may sail to windward at smaller angles than this, but by far the majority do not.

So, if we can agree that the una-rig probably won't be able to generate as large a value of CLmax as the sloop rig, then it would be very common for the sloop-rigged boat to be faster - even going to windward.

 

Everything I've seen the "aircraft theorist" write here actually matches perfectly with what actually occurs in our sailing vessels. The same can't be said of most of the amateurs who decry the rigs that work and spuriously claim that they are only used because of rating rules or tradition.

 

Tom/Mark,

Thanks for the continuing education opportunity.

I can not seem to get a real "feel" for the finer points of wing design (I tried playing around with the free tools for wing design as noted in http://www.boatdesign.net/forums/hydrodynamics-aerodynamics/what-significance-wing-thickness-45383-2.html). However, some of the recent discussion (front part vs rear part & under rotate vs smooth lee) really "hit home" with the results I noted when I was playing around.

For those that do not want to bother with the above link, I was studying sections for "reversible camber" three section rigid wing (two hinges, no slots). I wanted a fat front section (for structural reasons) and the mid and rear sections were going to be flat surfaces (sample pics attached). I used a "plug and chug" approach with XFLR5 and Java Foil to see what worked and what did not work.

When I was looking for min drag as a priority (think upwind work with plenty of wind) I found that slight under rotation of my fat front part with no extra angle on the flap (flap angle at the same as the mid portion) worked well. Although the front portion was slightly under rotated, there was no Lee separation due to the low AOAs I was focused on.

When I was looking for good to high lift without terrible drag (reaching / down wind sailing) having the "front part" rotation set for a smooth lee allowed good L/D with transition from good lift to max lift available by going with more flap (more angle on the very rear most portion).

 

P Flados,

Beware of XFLR5, I've never be able to display the Boundary Layer Shape function, nor the Cf function.
So I don't know how to manage TE separation.

Did you get it with XFLR5?

If no, To get it you ve to go back to XFOIL.

Cheers

EK

 

You can ask Andrè or other users of XFLR5 for help, whenever you encounter a problem. This is the XLR5 forum: https://sourceforge.net/p/xflr5/discussion/679396/
and they usually reply very quickly.
By the way, the new version (6.11) of XFLR5 was released right yesterday.
Cheers