sail aerodynamics

On problem with the Finn mast is that it over-rotates due to the mid-sheeted boom bending to windward.

Correction - meant to say it under-rotated. Thanks

 

Is that bad because of its effects on the aerodynamics, or the mast bend?

 

Because besides lift, which was so far treated and excellently explained by prof. Drela through the results of the Thin Airfoil Theory (TAT), there is also the drag. Drag is the other side of the medal, very important for the performance of the sail when beating upwind. That's where means for the boundary-layer control and reduction of drag become necessary (such like streamlining of the mast, of the luff sleeve etc.), and they are beyond the limits of the TAT (which doesn't treat the airfoil drag).

 

Thin Airfoil Theory can include the effects of thickness on surface velocity by the use of a source singularity distribution.

If the flow remains attached then boundary layer theory can be used with the surface velocity predictions from inviscid calculations (including those from Thin Airfoil Theory)to calculate drag. Various criteria based on boundary layer theory can also be used to estimate if separation will occur, but good models of separated flow using invsicid theory are difficult at best.

 

No. We are talking about sails here, not wing airfoils or hydrofoils.

TAT cannot be used in a reliable manner for the estimation of the drag of the generic sail+mast combination, except for the descriptive (or qualitative) discussion. The flow immediately behind the mast is generally separated, with a bubble of turbulent recirculating flow which may reattach at low angles of attack but will not for higher, yet practically important, angles of attack. There is a range of sailing points where the sail is operating in near-stall condition, with a large separation on the leeward side of the sail. You won't analyse that with TAT, unless you just need a qualitative result and not a numerically valid one.
As an example of a typical 2-D flow over a sail, see the attached CFD visualizations, taken from this thesis work: http://oa.upm.es/14885/1/PATRICIA_IZAGUIRRE_ALZA_A.pdf

The main problem is how to mathematically model the boundary layer dynamics in these conditions, in such way to obtain quantitatively acceptable results (comparable to wind-tunnel measurements) in terms of lift, drag and moments. In practice, I have yet to see a 2-D method which gives reliable results for a wide range of regular and smooth airfoil shapes and Reynolds numbers, let alone for a discontinuous geometry like mast+sail.

The exception to the above might be masts shaped to a streamlined section (teardrop or wing masts) which can be treated with methods you describe, but with a grain of salt and a lots of experience, knowledge of aerodynamics and test data for the validation of results.

Back during my university studies, I remember my professor of aerodynamics (prof. Sergio De Ponte, of Politecnico di Milano Aerospace Department) warn us:
"Bear in mind that the Thin Airfoil Theory is an excellent tool for design and analysis of all types of airfoils, except the thin ones."

Only several years later I had fully understand the reasons behind his words, and their implications.

 

I don't think we disagree. Perhaps you overlooked where I said "If the flow remains attached ....." and then went on to say "good models of separated flow using invsicid theory are difficult at best". I agree that Thin Airfoil Theory (or any other inviscid theory) plus boundary layer theory usually can not accurately predict the drag of a typical mast and sail.

 

I agree that we fundamentally do not disagree.

But I felt a need to underline the fact that the conditions of attached airflow are more an exception than a common thing in classic sailing. At most points of sail the flow will be detached over the big part of, or the entire, leeward surface. At some points of sail the drag is an important or paramount component of the drive force.
So the range of sailing conditions in which the sail profile can be analyzed with the TAT is actually pretty narrow.

However, things are different for the modern high-speed hydrofoiling machines (which is the battlefield of prof. Drela, if I am not wrong). They do see a mainly attached flow at their principal sailing points, which makes them more suitable for analysis with tools commonly used in aeronautics. That's where the TAT becomes useful again.

Cheers

 

Nobody said that the front airfoil parts are unimportant. The rear parts do indeed control the amount of lift, but this assumes/requires that the front parts are at least present. Another way to say it: the front parts carry most of the lift, but how much lift they carry is mostly determined by the shape of the rear parts.

And all this assumes attached (mostly inviscid) flow, and does not consider separation and stall. The shape of the front parts certainly affects the maximum attainable lift, which is something quite distinct from the actual lift at any given incidence angle.

 

1 - The mast was rotating with the wing, like it would in real life.

2 - It was a long time ago... but I found the lab report in the shelf. The 0,5 m chord foil was run at 15 m/s and 25 m/s - at 25 m/s, Rn was 8x10E5, which corresponds to about 5 m/s in the real world scale. 15 m/s Rn was 5x10E5, and between these the Re-effect was minimal. A separate Re test was done, from 2x10E5 to 1.2x10E6. There was a slight difference, Cl decreasing from 1,27 to 1,22, and Cd likewise decreasing from about 0,03 to 0,027.

The tunnel was low turbulence, nominally 0,1% but I believe in reality probably 0,2%, as the screens were not exactly clean. That's actually a bad thing, as turbulence in natural wind close to sea level is much more.

3 - At least I think it worked better than no scheme at all. I haven't really done much comparison against modern codes, as my last Macintosh that would run the old system 6 died in about 2000. It was based on first predicting the separation point by a method described in the AMO Smith high lift paper, then grocely adjusting the Cp-curves on both sides to allow for the separation. Not scientifically sound at all, but served its purpose. I can probably find the Pascal-code somewhere, if you're interested. I don't have my notes here so that I could elaborate.

Look at the very end of our Macsail page http://cdn2.hubspot.net/hub/209338/news/95_11_MacSail/MacSail.html. Sailpowercalc and SailTrimSim are also based on macsail calculations.

 

I would rather say it underrotates, if anything, but not a problem. We have also measured very accurately, that the Finn mast does not twist at all.

 

Can the Airplane Aerodynamist Explain This?

I couldn't resist bring this old posting up again,...ran across this afternoon, and it haunted me again. Certainly a monoplane wing with no supporting wires etc should produce the best lift with the least drag .... so a free standing una-rig sail rig.....:?:

 

And as the bit you quoted says, the boat with the free-standing una rig "would be improved upwind with a No.1 jib. Generally, we could not point as high as the others here (Block Island) and therefore had difficulty holding lanes.”

The post seems to show, once again, that a wingmast is not particularly fast on keelboats, and that una rigs are slower than sloop rigs on keelboats. Mark and others have pointed out the reasons pretty clearly over time.

With the greatest respect to Eric, there are plenty of people who are pretty free-thinking in sailing. That is evidenced by the huge diversity of rigs in the development classes, and time and time and time again we see that the rigs that are in use in development classes are the ones that suit the characteristics of that type best. The free-standing una-rig sail is used in the world's two most popular dinghy classes. It's not as if people are blind to it - it's just that most of us know when and where it works.

To assume that the free-standing wing-mast rig is fastest but that people shy from it because of conservatism assumes that people like the Moth sailors, especially, are stupid or stuck in the mud. They are not. Like the people from many other development classes they have tried wings and/or free-standing rigs time and time again, and Amac's latest design uses the freestanding rig for simplicity. So we know for a fact that guys like Amac are very aware of the freestanding rig (and wing-type leading edges) and yet they normally use stays and pole masts. Why? Because they are better in that application. Same as sloop rigs in racing keelboats.

 

Think of a stereo. The sound is produced by the speakers, but the amount of sound is controlled by the volume knob on the amplifier. With an airfoil, the greatest pressure differences may be at the leading edge, but they are controlled by the conditions at the trailing edge. Yes, you can modify the sound somewhat by changing the speakers, but the real influence is in the amplifier. Likewise, you can modify the lift somewhat by changing the leading edge, but the trailing edge is where the action is with regard to determining how much lift is produced.

The whole point of a sail is to redirect the wind and thereby create a force that results from the change in momentum of the air. (Working on the direction part of momentum and trying to make as small an impact to the speed part of momentum as possible.) So it make sense that where the leech is pointed is the direction the wind will be going as it leaves the sail's influence. The shape of everything forward of that is dedicated to ensuring that the flow actually will get turned in the direction of the leech. If it just blows by the sail without really being turned all you get is a turbulent wake with a lot of drag and not much lift.

 

I will back that statement up. A GF can be replaced by a "proper" designed section and will result in a marginally better L/D ratio (depending on the size of the GF in question). A "proper" design section in the case of operating near maximum lift would have a TE hook or kick as some people call it which has a similar effect to the GF.

GF are popular in motorsport because of regulations as well as being a cheap and easy way to adjust the amount of lift on existing profiles.

 

I would like to read the whole article first, if you can provide a link.
And then - what exactly is the question?