sail aerodynamics

[markdrela]

Originally Posted by tspeer

With regard to the mast, I don't think there's much reduction in the drag of the mast due to a change in dynamic pressure.

I dunno about this. I'd think the jib's effect on the mast drag would be pretty big. The contribution of a compact bluff body to the total drag scales as the cube of the velocity at the body. So if the jib induces a modest 10% reduction in airspeed at the mast, the drag of the mast goes down by a factor of 1-0.9^3 = 0.73 , or a 27% reduction. Seems quite significant.

[brian eiland]

Interesting formulation of questions by Skippy, and as always, an interesting reply from Tom Speer. Regrettably I can't enter the discussion right now as I am leaving for the Miami Show. I hope to find a client that's willing to spend a little time on my predominately headsailed arrangement.

[Skippy]

Thanks Brian. And thanks for the corrections Tom. It's nice to know I didn't get all of 'em wrong.

1. Slat effect: In the vicinity of the mast, the velocities due to circulation around the jib run counter to the velocities around the main
This is true only on the leeward side of the main. Are we talking about negative pressure peaks? That sounds like simple interference: The reduced net circulation is another way of saying the pressure is moderated in the slot. I can see that reducing separation off the leeward side of the mast. Wouldn't that reduce it's drag? I thought that would be a primary benefit. Otherwise, how do you take advantage of the reduced pressure peaks? Sail in higher winds?

2. Circulation effect.
I guess this sounds to me like an "upwash" effect. The upwash of the main increases the speed and improves the direction of the airflow experienced by the jib. It affects the trailing portion of the jib more than it does the forward areas, so it's kind of like lowering a trailing-edge flap on a foil.

5. Fresh-boundary-layer effect. Lower Reynolds numbers are not more efficient.
Oops. I think what I had in mind was that the shorter chord would dump the air sooner. How about this:

Multiple elements can be smaller than an equivalent single element, and therefore have a shorter chord. This dumps the air before or soon after the boundary layer turns turbulent.
This effect applies to each element individually. It's not a product of their aerodynamic interaction, via the slot or otherwise. It's allowed by the mechanical fact that the two elements are attached to the same vessel, both helping to propel it.

[tspeer]

Originally Posted by Skippy

1. Slat effect: In the vicinity of the mast, the velocities due to circulation around the jib run counter to the velocities around the main
This is true only on the leeward side of the main. Are we talking about negative pressure peaks? That sounds like simple interference: The reduced net circulation is another way of saying the pressure is moderated in the slot.

Yes, negative pressure peaks. Yes, it is interference. The point is to use that interference in a way that is meaningful to improving the performance of the whole system.

I can see that reducing separation off the leeward side of the mast. Wouldn't that reduce it's drag? I thought that would be a primary benefit. Otherwise, how do you take advantage of the reduced pressure peaks? Sail in higher winds?

Yes, of course reduced separation means reduced drag. I suspect one of the benefits of overlapping headsails is to establish a favorable pressure gradient aft of the mast to reattach the flow separating from the mast.

One way to take advantage of reduced pressure peaks is to increase the angle of attack and go to higher lift coefficients in the same or lighter winds.

2. Circulation effect.
I guess this sounds to me like an "upwash" effect. The upwash of the main increases the speed and improves the direction of the airflow experienced by the jib. It affects the trailing portion of the jib more than it does the forward areas, so it's kind of like lowering a trailing-edge flap on a foil.

Exactly. The main does indeed make the jib act like it's got a flap deflected. This is why people find that jibs are more powerful than mainsails of the same area.

5. Fresh-boundary-layer effect. Lower Reynolds numbers are not more efficient.
Oops. I think what I had in mind was that the shorter chord would dump the air sooner. How about this:

Multiple elements can be smaller than an equivalent single element, and therefore have a shorter chord. This dumps the air before or soon after the boundary layer turns turbulent.
This effect applies to each element individually. It's not a product of their aerodynamic interaction, via the slot or otherwise. It's allowed by the mechanical fact that the two elements are attached to the same vessel, both helping to propel it.

Basically. Since each element starts off with laminar flow, it might be possible to achieve a greater proportion of laminar flow than for a single element.

However, what makes this difficult to realize in practice is the job of the after elements is to achieve the deceleration of the flow from the high velocity of the forward element to somewhat below freestream velocity at the final trailing edge. These elements tend to have adverse pressure gradients over nearly all their suction surface, and a turbulent boundary layer can tolerate a much more aggressive deceleration than a laminar boundary layer. The figure above is a good example. The forward element can maintain laminar flow over half the total chord, but the flap will see laminar separation almost immediately and transition to turbulent just behind the slot. So there's turbulent flow over pretty much all of the aft half of the total chord.

I think the more important aspect of the fresh boundary layer is it can stand a steeper initial deceleration. The Stratford distribution has a very sudden deceleration followed by a gradual tapering off. Each time you refresh the boundary layer, you get a new abrupt deceleration segment.

[Skippy]

Thanks a lot Tom. I'll think I'll consider this my "final" answer for now. No response requested, unless you really want to.


0. Venturi-effect myth: Faster airflow through the slot has been said to increase the rig's power.

For the most part, this does not occur. A venturi is a closed passageway that forces air through a restriction. Air encountering a slot between two sails will also flow around the sails on either side of the slot. There is some variation in flow speed within the slot, but the average speed is slower than freestream.

[Edited per Tom's comment below.]


1. Slat effect: In the vicinity of the mast, the air velocities due to circulation around the jib run counter to the velocities around the main, and so reduce low-pressure peaks behind the mast.

Most aerodynamic elements (sails and wings) are susceptible to separation of the airflow from the leeward side just behind the leading edge. This is caused by strong low-pressure conditions there. With a standard masted mainsail, the flow almost always separates from the leeward side of the mast, drastically reducing power. Reducing the low-pressure condition there makes separation less likely for a wing, and with a masted sail, makes it occur later and helps get the air flowing smoothly again on the leeward side of the main. This allows the main to be sheeted in more and sailed harder.
This is interference between the two sails. It is a lack of circulatory airflow in the slot, rather than a presence of flow.


2. Circulation or "upwash" effect: In turn, the main causes the trailing edge of the jib to be in a region of high velocity that is inclined to the rear of the jib. Such flow inclination induces considerably greater circulation on the jib.

The upwash of the main increases the speed and improves the direction of the apparent wind experienced by the jib, significantly increasing the jib's power.
This requires that the jib be ahead of the main as well as leeward of it.


3. Dumping effect: Because the trailing edge of the jib is in a region of velocity appreciably higher than freestream, the boundary layer "dumps" at a high velocity. The higher discharge velocity relieves the pressure rise impressed on the boundary layer, thus alleviating separation problems or permitting increased lift.

As air passes along the leading portion of the lee side of of the main, it experiences a decrease in pressure (favorable gradient). This reduces the "pile-up" or unsmooth flow of air (boundary layer separation) due to an increase in pressure (adverse gradient) along the weather side of the jib, and therefore keeps the air flowing more smoothly along it in light winds.
The benefit of this effect occurs in the slot, but is caused by an area farther downstream.


4. Off-the-surface pressure recovery: The boundary layer from the jib is dumped at velocities appreciably higher than freestream. The final deceleration to freestream velocity is done in an efficient manner. The deceleration of the wake occurs out of contact with a sail. Such a method is more effective than the best possible deceleration in contact with a sail.

This occurs downstream of the slot, or even downstream of both sails.


5. Fresh-boundary-layer effect: Each sail starts out with a fresh boundary layer at its leading edge. Thin boundary layers can withstand stronger adverse gradients than thick ones.

Multiple sails can be smaller than an equivalent single sail, and therefore have a smaller width (chord). Ideally, this dumps the air before or soon after the boundary layer turns turbulent.
This effect applies to each element individually. It's not a product of their aerodynamic interaction, via the slot or otherwise. It's allowed by the mechanical fact that the two elements are attached to the same vessel, both helping to propel it.

[tspeer]

0. The venturi is not 100% wrong in that the average flow speed does increase from the mouth of the slot to the exit of the slot. Once the massflow through the slot is determined, then it will be slower where there's more cross sectional area than at the narrow slot. You see this in particular near the pressure side of the forward element trailing edge.

But it doesn't mean that if you close down the slot the velocity at the exit of the slot will be faster than if you open the slot. This is because of the change in massflow through the slot. And there can be a significant difference in speed across the slot, too, as you can see in the slotted airfoil pressure distributions posted earlier.

You've pretty much got it for all the rest.

[brian eiland]

The gentleman who most deserves credit for finally getting the explainations of sail aerodynamics corrected, and upon which several excellent books by Tom Whidden and C.A. Marchaj are based has recently updated his website to include many of the technical papers and magazine articles he wrote on the subject originally. I had mentioned his name, Arvel Gentry, in these postings previously, but I could not make a direct reference to his many documents as they were not posted on his site at that time.

From his site, "I got involved in the technical aspect of sailing because I started racing. Reading the sailing books and magazines, I began to realize that most of what was written about the aerodynamics of sails was wrong, or certainly very misleading."

"The explanations for how lift is generated were based on popular myths. The description of the interactions between a jib and mainsail, the "slot effect", did not make much aerodynamic sense."

"I was soon launched on a quest to discover how our sails really worked. Over the years this resulted in a number of technical sailing papers and magazine articles. The technical sailing papers are archived in this section of my web site. My magazine articles can be reached from my Home page."

"If you are interested in sailing aerodynamics and how your sails work, you have come to the right place. All of my sailing technical papers and magazine articles are archived on the Technical Papers and Magazine Articles pages.

http://www.arvelgentry.com/

[brian eiland]

...from a recent issue of Scuttlebutt....
"As regards sails, the canted mast clearly reduces a little of the mainsail's surface area, but the genoa is shifted further back for the same height of rig, optimising aerodynamic efficiency. The difference is made downwind with the larger gennakers. The upwind performance is improved with the forestay shifted back, giving less deflection in the leeward shroud enabling one or two extra degrees to be made up in terms of heading."

Brian wrote:
I noticed this posting, but have not had time to evaluate it. Thought I would post it here until I could get back to it, or if someone else cares to comment.

The entire posting was;
Groupama-2 controls the fleet once again, taking its fourth Grand Prix victory since its launch in June 2004. In Vigo this weekend, Franck Cammas and his ten crew stole the show by winning five of the six legs in rather a light breeze... taking second just the once after a poor start.

There is no debate: the most modern trimaran designed by architects Vincent Lauriot-Prévost and Marc van Peteghem, Martin Fisher, Mike Kermarec and Groupama's shore crew, is unquestionably the fastest multihull. In the 33 legs raced since the Grand Prix of Corsica 2004, Groupama-2's first appearance on the racing circuit, the green trimaran has won 28 legs in the five Grand Prix ! The fine tuning didn't take long: apart from a crack on the hydraulic ram attachment (GP of Corsica 2004) and some delamination on the beam (IB Group Challenge 2005), the trimaran has been perfectly optimised on every level since coming out of the yard in June 2004.

As regards the platform, Groupama-2 has much finer floats, a sharper bow, a much narrower planform x- beam, a mast that cants 60cm more than the other trimarans, smaller profile beams and a weight saving of around 200 kg. As regards the appendages, the green trimaran also has a very thin daggerboard, with a small chord and a large trimtab (rear part of the daggerboard that has its direction altered by up to 8° and represents 28% of the lift surface), some high aspect ratio rudders and very curved foils (a 3 metre radius instead of a 4m one for the other trimarans). As regards sails, the canted mast clearly reduces a little of the mainsail's surface area, but the genoa is shifted further back for the same height of rig, optimising aerodynamic efficiency. The difference is made downwind with the larger gennakers. The upwind performance is improved with the forestay shifted back, giving less deflection in the leeward shroud enabling one or two extra degrees to be made up in terms of heading.

Lighter, Groupama-2 reacts faster to a gust and immediately transforms the acceleration, while the other trimarans take the pressure before transferring it into propulsion. It is also more reactive downwind and pulls away faster out of a manuvre (tack, gybe). With more canvas and lighter construction its power to weight ratio is a little better than the others. With the optimised aero and hydro-dynamics, Groupama-2 is also capable of sailing higher, which is crucial for getting a slight separation during the close contact phases of a start, and a lateral gain when rounding a mark...

All added up, each of these small advantages enable Franck Cammas to remove himself from situations which are beyond the control of any of the other boats and to snatch back a few tenths of a knot when he is sailing
in a stable breeze. He really has to be trapped during a start for Groupama-2 not to win, as has been the case only twice since the beginning of the season : in Marseille and Vigo where he was blocked by his competitors, it took him two laps to get back to the front of the fleet...
-- Laurence Dacoury (Kate Jennings for translation)

[brian eiland]

Originally Posted by brian eiland

Here is another affirmation of the slot effect....look at this new DynaRig proposed for the 285' Maltese Falcon:
<http://www.doylesails.com/newsletter-03-pg8.htm>

Note the sheeting angles for the three sails, and particularly that of the very forward sail. Now why is this?....look at the explantions of the slot effect.

And which sail do you think is providing the greatest drive forward, and which sail is providing the greatest leeway??

Interestingly, they are not trying to account for the wind gradient (twist) with this new design even though it is an extremely tall rig.

I had previously posted this back on thread #47. When I went to look at the referenced website it was no longer there. Meantime I did run across this CFD analsyis of Maltese's rig. http://syr.stanford.edu/HISWA_Tyler_2002.pdf
Have a look at the illustrations toward the end of the paper dealing with optimizing the sheeting angles for the 3 sails

[jam007]

After reading this thread I did a run in JavaFoil. (A 2D foil analyse program similar to XFOIL.)
See picture below.
The angle of attack (for the main) was 14 degrees.
Notice the low pressure near the leading edge of the foresail and the high pressure area just between the leading edges of main and foresail. The relatively high pressure on the lee side of the main and the upwash in font of the foresail and the downwash in front of the main.
Cl for this combination was 2.5. For a single foil at 14 degrees it was 0.9. One foil set at 4 degrees (the attack angle of the foresail) has a Cl of 1.1. The interaction adds an extra 25% of lift.
A single foil with a form approximating the two foils generates a Cl of 1.7 but Cd is much lower 0.03 compared to 0.1 for the two folis. This is in accordance with the supremacy of the sloop or single sail rigg going to windward where Lift/drag ratio is more important than lift alone. Of course real sails are 3D objects but this should enhance the differens rather than diminish it due to induced drag (higher lift is payed for by more induced drag).

[brian eiland]

I’m having a little problem interpreting your interpretation of your JavaFoil analysis run.

Originally Posted by jam007

The angle of attack (for the main) was 14 degrees......
Cl for this combination was 2.5. For a single foil at 14 degrees it was 0.9. One foil set at 4 degrees (the attack angle of the foresail) has a Cl of 1.1. The interaction adds an extra 25% of lift.

You included two sets of ‘single-foil’ conditions in addition to the ‘combination’ condition, and then called for an ‘interaction’ percentage increase. Are you really saying that the ‘interaction’ that would be experienced in the ‘combination’ condition adds an extra 25%? Does the 4 degree statement even belong here?

In other words in the case of 14 degrees angle of attack (for the main), the CL for this combination of two foils was 2.5, while the CL for single foil at this same angle of attack was only 0.9. That appears to be an almost threefold increase, not just 25%?

Originally Posted by jam007

One foil set at 4 degrees (the attack angle of the foresail) has a Cl of 1.1

I don’t understand that this is the angle of attack of the foresail?

Originally Posted by jam007

A single foil with a form approximating the two foils generates a Cl of 1.7 but Cd is much lower 0.03 compared to 0.1 for the two folis. This is in accordance with the supremacy of the sloop or single sail rigg going to windward where Lift/drag ratio is more important than lift alone.

The lift/drag ratios of aero-foils are not the only determining factor in the windward capability of a sailing vessel. One must also consider the vectored direction of the ‘lift forces’ verses the driving forces of the sail-foils. Higher ‘lift’ forces can result in considerable leeway forces and heeling forces. Looking at your attached diagram it appears as though the two foils you show are both practically identical. Yet, it is also quite apparent that the forward foil is significantly more effective at developing lift, and that were these sails on a vessel, this forward sail is much more effective at driving the vessel forward (its lift force is more forwardly directed). AND look at the capability that this forward foil has to heading up higher into the wind if momentarily needed.

Granted you might replace this two-foil combination with a “single foil with the form approximating the two foils” and thus reduce the drag forces per lift, BUT you will lose the higher pointing capability of the combination of two foils, and your lift forces of this single foil will act more to heel you and increase your leeway rather than drive you forward.

[jam007]

Originally Posted by brian eiland

I’m having a little problem interpreting your interpretation of your JavaFoil analysis run.
...

Yet, it is also quite apparent that the forward foil is significantly more effective at developing lift, and that were these sails on a vessel, this forward sail is much more effective at driving the vessel forward (its lift force is more forwardly directed). AND look at the capability that this forward foil has to heading up higher into the wind if momentarily needed.

Granted you might replace this two-foil combination with a “single foil with the form approximating the two foils” and thus reduce the drag forces per lift, BUT you will lose the higher pointing capability of the combination of two foils, and your lift forces of this single foil will act more to heel you and increase your leeway rather than drive you forward.

I will try to clarify my thoghts.
The two foils are identical. The leading foil (foresail) is set at an angle of 10 degrees from the aft (mainsail). I also did runs with one foil. I then got a Cl of 1.1 at 4 degrees angle of attack and Cl 0.9 at 14 degrees. So I then concluded that the interaction of the two foils added an extra 25 %. (Is this a resonable approximation or am I only exposing my ignorance?)

I totaly agree that the fore+main combination not only increases the Cl but in a considerable way the driving force compared to only main. The reason I mentioned sloops and single sail riggs was the earlier diskussion of riggs with many interacting sails. I think the sloop is the optimum for displacement boats and most multis for winward efficency (L/D ratio compared to force generation) due to the reasons you gave.

But I wonder. Is not Cl and Cd measured relative the free flow velocity of the air. This should result in that a large L/D is the same as a more forward directed force. An extreme situation would be with negative Cd if the net force is directed forward of perpendicular to the angel of atack.

What made me post the picture was the graphical illustration of the interaction that You, Tom Speer and other have described and I read about in Marshaj´s books and Gentry´s articles.

Anders M

[tspeer]

Originally Posted by jam007

...
But I wonder. Is not Cl and Cd measured relative the free flow velocity of the air. This should result in that a large L/D is the same as a more forward directed force. An extreme situation would be with negative Cd if the net force is directed forward of perpendicular to the angel of atack....

Yes, lift is defined as the component of the total force that is perpendicular to the relative wind and drag is defined as the component of the total force that is parallel to the wind. Once you know lift and drag, you know everything about the magnitude and direction of the force. There's no such thing as angling the lift vector forward, unless you change the relative wind direction.

But lift and drag aren't the only choices of coordinate system. Normal force and axial force are the components perpendicular and parallel to the chord. Normal force is always larger in magnitude than the lift because drag is always positive and contributes to the normal force. For example, at zero angle of attack, lift and normal force are the same. At 90 degrees angle of attack, drag and normal force are the same. Axial force can be positive (aft), or negative (forward) because the lift vector does point forward as angle of attack increases.

To say, for example, "its lift force is more forwardly directed," is to confuse normal force with lift. Instead, one should say, "its normal force is more forwardly directed." One can't incline the lift forward to increase the drive, one can only reduce the drag

The advantage of using lift and drag is you don't have to know anything about the orientation (angle of attack) of the surface. You can calculate the performance entirely in terms of lift and drag of the topsides and sails, and the lift and drag of the hull and foils. You don't need to know the orientation of the boat (leeway angle) or the orientation of the sails relative to the boat. But often it's useful to know the normal and axial forces, for example when calculating loads or hull trim. Another set of coordinates would be drive (forward) and side force (perpendicular to the boat's plane of symmetry). Given one set of components, you can transform the forces into any other set.

[jam007]

Different coordinate systems can be confusing. Actually often aerofoils have the net force directed forward of the normal of the cord as it is the low pressure area around its forward part that creates the main part of the force.
Example: A NACA 6412 at 10 degrees angle of attack has the net force directed 2 degrees aft of the perpendicular of the flow. That is 8 degrees forward of the normal of the cord.
See picture below.
(Note the lines and vectors are drawn by hand and may not exactly be at the correct angles.)

Anders M

[yokebutt]

Anders,

Is that also true of the same section without camber? Or is that why cambered foils have their low-drag bucket shifted? Please explain!

Jocke.

P.S. For the benefit of the non-scientists and non-native speakers of English here, the word "normal" as used to describe forces in this discussion means perpendicular to, wich in turn means "at right angle to" or "square to".