How do sails work?

Back

Separation and Stall
Eventually at some angle of attack, sail sections (and airfoils) experience stall. This occurs when the air flowing around the leeward side of the sail no longer travels on the surface of the sail. The flow separates from the sail resulting in a large loss in lift. Depending on the shape, stall can occur abruptly with a small increase in angle of attack, or more gradually with some indications that the flow is separating from the surface at specific locations first. This is easily seen using telltales (tufts of yarn) that swirl erratically when the flow departs from the desired direction instead of streaming aft when the flow is attached to the sail.

Separation occurs simply because the pressure gradient that the flow is trying to pass through is too extreme. Recall that lift is generated because the flow accelerates around the convexly curved leeward surface of the sail creating low pressure. Eventually, as the flow approaches the back of the sail, the flow must slow down to near its original speed and pressure, since after it leaves the sail it will return to its original state when the sail is no longer there to influence it. This is referred to as pressure recovery.

Another way to think about this is that when air flowing over the leeward side of the sail and air flowing over the windward side of the sail reach the trailing edge, they must have the same pressure, as there will not be anything in between anymore to enable maintaining different pressures. It does not mean that two particles of air that start at the leading edge and travel along different sides of the sail will arrive at the trailing edge at the same time (a common misconception). It simply means that the pressure of air flowing off the top right at the trailing edge of the sail will be equal to the pressure of air flowing off the bottom. This must be true since they are coincident there. That pressure is generally close to the original pressure of the flow prior to being disturbed by the sail.

So, accelerated flow around the leeward side slows down toward the leech in order to provide the necessary matching at the trailing edge as the air is returned toward its original conditions. Overall, air still travels much faster around the leeward side of the sail than the windward side.

As the flow slows down from its accelerated state on the leeward side it yields a pressure gradient along the back of the sail that is increasing from very low pressure (to produce the desired lift) to a much higher pressure toward the leech. The amount of initial acceleration (dependent on angle of attack and shape) and the length of the pressure recovery determine how steep this gradient is. When the increase in pressure that the flow is experiencing becomes too extreme, the flow no longer stays attached to the surface of the sail. It is pushed away by the higher pressure and stall occurs, yielding less lift. This happens when the angle of attack is high and/or the camber is quite deep, causing very high velocity, but necessitating dramatic slowing of the flow, too, with the associated rapid increase in pressure causing separation. It is favorable to slow the flow in a smooth fashion over a longer distance so that there is no steep rise in pressure. This happens most effectively over a long, straighter shape aft, lacking in curvature that would attempt to promote higher velocity.

Pressure Distribution and Curvature
As the air flows around both sides of the sail, its pressure changes with the varying local velocity caused by the curvature of the sail. With the entry angle defined by the oncoming flow direction and the angle of attack governed by avoiding stall, there are still numerous sail shapes that can be established to connect the luff and the leech.

Since the purpose of the sail is to develop force to move the boat in a forward direction, it would be most effective to have as much of the sail as possible operating with the largest possible pressure difference across it. The way to achieve that is to accelerate the air quickly around the curved leading edge of the sail in order to generate low pressure on the leeward side close to the luff and then maintain it back over a significant portion of the sail. This is achieved by imparting high curvature to the front of the sail. Once the flow is accelerated curvature can diminish and the flow will continue quickly around the leeward side of the sail. The back of the sail needs to be flatter in order to allow the flow to gradually decelerate to avoid stall as already described. These details are the basic factors that define the rounded entry with forward draft (position of maximum camber depth) and straight leech profile that has proven fast in typical sailing applications.

It is evident that a sail shape with more curvature aft keeps the air accelerated longer, which may produce a higher amount of total force due to a larger region of negative pressure. The problem is that the negative pressure vectors on the aft portion of the sail are angled more aft than those toward the front of the sail, so the amount of forward force that is produced in the direction that the boat travels is less, while the amount of sideways force that contributes to heeling and leeway is greater. It is also apparent that a shape with its camber aft has a shorter, steeper pressure recovery that will lead to earlier separation and stall.

One more factor to consider is that a rounder leading edge or deeper camber, while producing more force, does so at the expense of having a higher entry angle that requires a higher apparent wind angle in order to fly without luffing. This means that the boat cannot be sailed as close to the wind, which explains why spinnakers can be so full and deep, but becomes a tradeoff when setting an upwind sail. Producing more force versus sailing at a higher angle becomes a subtle optimization. It requires the proper balance between a full, curved leading edge to accelerate the flow, and a flatter, subtly curved leading edge that does not produce as much low pressure to pull the boat but does allow the boat to sail closer to the wind. Boats that sail fast with large amounts of sail area will favor flatter sails than slower boats with less sail area that need to develop more power to move the boat.

Induced Drag
Another factor that must be considered is induced drag. This is the drag that a wing generates when it creates lift. Over most of a wing, the low pressure above the wing is kept isolated from the higher pressure underneath by the physical presence of the wing. At the tip of a wing, where the wing ends, there is nothing preventing air from flowing around the wing tip from the high pressure beneath to the lower pressure above. This results in the standard tip vortex that is often seen spinning off the tips of airplane wings and flaps. When the flow takes this alternate path around the tip instead of over the airfoil surface, energy is expended that does not develop lift, but does cause drag. This is called induced drag and it increases exponentially with lift, so a wing, such as a sail, that is producing substantial lift, experiences much more induced drag than a wing that is producing a lesser amount of lift.

The most effective way to minimize induced drag is to increase span, as induced drag is inversely proportional to the span squared. Highly efficient airplanes like gliders have very high span for the amount of lift they are producing. Winglets are a way to create the effect of higher span without actually increasing the physical span. They are useful when there is an artificial constraint on wingspan (like a draft limitation on a keel).

Spanload
Beyond being heavily dependent on the amount of lift, induced drag is also dependent on how that lift is produced. It has already been explained how taper and sweep affect the upwash along the span of a wing by causing sections along the wing to have different lift levels. Also, varying the amount of camber and the angle of attack of sections along the wing will influence how much lift is generated at various spanwise locations. The distribution of lift along the length of the wing is called the spanload.

It has been found for an isolated wing in untwisted flow that the spanload with the minimum induced drag is an elliptical distribution of lift. This is achieved on an untwisted, unswept wing with an elliptical distribution of area and the same section along its entire span. A spanload can be altered in several ways.

Tapering the wing causes the lift to be reduced outboard because, while more upwash is produced and the outboard sections are loaded more, there is less area to generate lift outboard. Sweeping the wing aft increases the upwash outboard and the lift there because the area is the same and operating at higher angle of attack. Twisting the outboard wing to higher or lower angles will increase or decrease the outboard lift levels, respectively. Finally, adding camber to the outboard wing sections will increase the amount of lift that they produce.

All of these features can be used to modify the spanload with each of the resulting spanloads producing different amounts of induced drag for the same amount of total lift. This is because induced drag is a consequence of how the lift being produced by the wing deteriorates at the wing tip. A wing tip, since it has no more wing outboard of it, cannot sustain lift because it cannot support a pressure difference. Thus, the lift at the very end of a wing must be zero. The inboard portion of the wing produces a significant amount of lift that must diminish toward zero approaching the wing tip. The manner that the lift decreases toward the tip defines the shape of the spanload, and it is the character of that lift distribution that establishes the amount of induced drag.

The exact shape of the optimum spanload for sails in a twisted flowfield varies somewhat from the simple ideal elliptical spanload. Since the flowfield is twisted in a manner that lift toward the top of the sail is oriented more in the direction of the boat than the bottom of the sail (because lift is produced in the direction perpendicular to the local flow direction), it follows that the ideal spanload in twisted flow conditions will be even more highly loaded toward the top than the simple elliptical lift distribution that is optimal for untwisted flow. The extremely tapered planform of typical sails yield spanloads with much less lift toward the top than the elliptical spanload, so are less than optimal. While a genoa has considerable sweep to help reload its top sections, it has very little sail area near its head to develop lift. A mainsail without sweep, particularly on a fractional rig where the top of the mainsail is above the influence of the foresail, does not generate enough lift toward the top to approach the elliptical spanload.

Increasing the chordlength of the top of the sails would be an effective way to create additional lift toward the top and attain a loading closer to optimal. This has been demonstrated to be beneficial through the use of full-length battens, but is not always allowed. Another way to increase the lift levels toward the top of sails is to provide additional camber toward the top to boost the lift being produced up high. Increasing the angle of attack, through decreasing twist, would also increase the loading at the top, but the limitation of stalling the top sections must always be minded, so the sail needs to maintain a certain amount of twist because of the inherently twisted flowfield. It is highly likely that the optimum lift distribution to minimize induced drag is not achieved with typical sails.

Setting Sail
Recognize that the objective of the sails is to create force to pull the boat, but that there can also be a constraint on heel. At some point the stability of the boat or weight of the crew cannot keep the boat sailing at an angle that does not compromise performance, so just using the sails to produce the most force possible is not necessarily the fastest procedure.

In lighter winds, when the sails are struggling to extract enough force from the wind to move the boat fast, the sails should be set such that every section along the height of the sail is working to produce high lift, especially the top sections in order to minimize induced drag. When the wind builds beyond a level that the sails' force causes the boat to heel too much, the sails' characteristics must be modified. There are several options.

Reducing the amount of camber in the entire sail will decrease the amount of force the sail produces, as will decreasing the angle of attack of the entire sail. Implementing these adjustments over the entire sail may or may not be the best alternative for the windier conditions. They reduce the amount of force generated by the sail, but that force is still centered at a similar height. In order to reduce the heeling moment created by the sails to a satisfactory level, the amount of force may decrease to a level that does not pull the boat very fast anymore.

Another approach is to reduce the lift produced by the top of the sail. Reducing the camber of the top of the sail, and/or reducing the angle of attack of the top of the sail through additional twist will affect the sail's force such that the remaining force is centered lower down. A similar reduction in heeling moment as simply reducing the entire sail's force can be achieved through depowering the top of the sail, but while maintaining more total force to pull the boat. The force is centered lower as the bottom of the sail is still trimmed in a fashion that generates substantial lift. This method has the compromise of deviating further from the desired elliptical spanload, as the lift distribution diminishes much more rapidly toward the top of the sail, and causes higher induced drag. The question becomes whether the remaining higher sail force offsets the additional drag component.

A parallel situation occurs with airplanes. Airplanes are not designed to fly with the optimal spanload that yields minimum induced drag because the higher outboard load on the wing would require that the wing be made stronger, hence heavier, to carry that load. It is more efficient to build the airplane lighter and generate more lift on the inboard wing and accept a little more induced drag. This is the same tradeoff that a sailboat experiences in strong wind when heeling becomes a factor and results in a similar, less than optimal spanload in order to maximize performance.

Pointing
With all of these scenarios, the angle that the boat is able to sail to the wind is always a consideration, too. If the angle of attack of the entire sail is decreased, much of it may backwind or luff and the boat may make very little progress due to lack of power. This is a consequence of the original restriction that sails be flexible. The boat can be turned away from the wind in order to fill the sail but that causes the sail to load again and the boat to heel too much making this a less viable option (except to avoid sailing too directly into waves). The component of the boat's speed in the windward direction must be accounted for when considering course variations. Twisting or decambering the top of the sail keeps the bottom of the sail still trimmed at an effective angle of attack, continuing to produce force and allowing the boat to be sailed at a closer angle to the wind.

The correct solution is generally a combination of the various adjustment options and will vary with wind speed and sea conditions. It is also dependent on the characteristics of the boat and rig, and the trim controls available (and probably even with the time interval with which the wind velocity is varying as some adjustments are made more quickly and easily than others). Hopefully, by understanding the lift producing characteristics of the sails and how to manipulate them, a sailor can continually alter the sails in order to produce the most effective force to move the boat in the intended direction.

Summary
It is evident that sails are flexible wings, operating in a twisted flowfield, and in the presence of each other. They produce force by accelerating air over their curved leeward side causing lower pressure on that side of the sail that acts to propel the boat. Camber and angle of attack can be controlled along the span in order to produce lift in many possible ways, some more favorable than others. Stall and luffing provide an upper and lower limit on the useful angles of attack. Taper and sweep affect the upwash of flow approaching the sails beyond the influence of the twisted apparent wind caused by the boat moving through the earth's boundary layer.

Paul Bogataj is an aeronautical engineer, specializing in sailing applications. He was responsible for appendage development for Team Dennis Conner and Young America in the 1995 and 2000 America's Cups, respectively. He approaches sail and keel design from the perspective of using advanced aerodynamic technology and methods that have been developed for the aircraft industry. He previously worked for Boeing and has consulted independently for a variety of sailing design projects. He currently works for North Sails One Design analyzing sails for one-design boats.

He has employed his knowledge of how sailboats function to win the North American Championship of two different classes, and numerous fleet and district championships. Paul combines his practical understanding of sailing from his experience as a successful racing sailor with his awareness of fluid dynamic principles as an engineer to provide explanations of how sailboats work that are understandable to the average sailor.

To contact Paul about this article, click here.

Copyright Paul Bogataj - All rights reserved.