Sail camber

Doug Halsey · Mar 9, 2016

What's interesting is that although pressure drag contributes to profile drag, pressure drag is more of a symptom rather than the cause. Viscous dissipation is the ultimate cause of profile drag. This video talks about this a bit:

https://www.youtube.com/watch?v=2IbYyvvBQeg
Click to expand...

Not disputing anything in the above quote, but the last few sentences in the video seem misleading :

... you would think that the pressure drag would obviously depend on the pressure distribution on the surface, but in fact all that matters is the viscous shear distributed over the boundary layer integrated in this way.

So we have the paradoxical result that the pressure drag does not in fact depend on the pressure. All that matters is the distribution of the viscous stresses in the boundary layer & wake. So the only way to reduce overall profile drag is to reduce viscous dissipation.

Trying to reduce pressure drag by directly manipulating pressure distributions is really an exercise in futility.
Click to expand...

Since the pressure distribution on the surface plays a significant role in causing the viscous stresses in the boundary layer & wake, wouldn't you still have to say that it's important in determining both the pressure drag & the friction drag, and that manipulating the pressure distribution is far from an exercise in futility?

markdrela · Mar 9, 2016

Doug Halsey said: ↑

Not disputing anything in the above quote, but the last few sentences in the video seem misleading :
Since the pressure distribution on the surface plays a significant role in causing the viscous stresses in the boundary layer & wake, wouldn't you still have to say that it's important in determining both the pressure drag & the friction drag, and that manipulating the pressure distribution is far from an exercise in futility?
Click to expand...

Assuming the airfoil or sail design is reasonable, i.e. without separation, and also the transition location is more or less fixed, then manipulating the pressure distribution via camber, thickness, or other shape changes, will have very little effect on the viscous dissipation and on profile drag. There will be SOME effect because the average surface velocity depends on the lift and the thickness, for example, but these are weak effects. For example, compared to a NACA0001 airfoil at Re=5M, the NACA0006 has 500% more frontal area, but only 13% more drag.

The video comment refers to is the following fallacy I've heard: One can reduce pressure drag by reducing the pressures on on upstream-facing surfaces and increasing the pressure on downstream-facing surfaces. Or correspondingly, by increasing lift on the jib and decreasing lift on the main. This is futile as the video says, because the changes in the sail geometry needed to modify the pressures like this will compensate for the changes in the pressures, for no net effect in the integrated pressure drag.

brian eiland · Mar 9, 2016

markdrela said: ↑

.....
The video comment refers to is the following fallacy I've heard: One can reduce pressure drag by reducing the pressures on on upstream-facing surfaces and increasing the pressure on downstream-facing surfaces. Or correspondingly, by increasing lift on the jib and decreasing lift on the main. This is futile as the video says, because the changes in the sail geometry needed to modify the pressures like this will compensate for the changes in the pressures, for no net effect in the integrated pressure drag.
Click to expand...

So I guess this explanation is totally incorrect??

How do sails work? said:

excerpt...
Sails in Combination
Each sail by itself is much simpler than the combination of a foresail and mainsail as in the sloop rig. The sails are operating so close to each other that they both have significant interaction with the other.
The most interesting feature of this is that the two sails together produce more force to pull the boat than the sum of their forces if they were each alone.

Earlier, upwash was identified as the increase in flow angle immediately upstream of a wing. There is also a corresponding change in angle, called downwash, just behind a wing, where the flow leaving the wing has been turned to an angle lower than the original flow. This is the cause of the well known "badair" that a boat just to windward and behind another boat experiences. approaching the mainsail to be significantly reduced from what it would be otherwise. This decreases the amount of force that the mainsail produces. The observed affect commonly referred to as "backwinding" is partially a result of downwash from the foresail, but is also due to the higher pressure on the windward side of the genoa being very close to the forward, leeward side of the mainsail, causing the flexible material of the mainsail to move away from that higher pressure.

The foresail of a sloop rig operates in the upwash of the mainsail. The wind as far upstream as the luff of a genoa is influenced by the upwash created by the mainsail. Hence, a jib or genoa in front of a mainsail has a higher flow angle than it otherwise would have by itself, causing an increase in the amount of force that the forward sail produces. So, while the mainsail is experiencing detrimental interference from the foresail, the foresail benefits from the interference of the mainsail. Notice that more air is directed around the curved leeward side of the foresail. This causes higher velocity (lower pressure) and more force.
The net result is that the total force of the two-sail system is increased, with the foresail gaining more than the mainsail loses.
Click to expand...

markdrela · Mar 9, 2016

brian eiland said: ↑

So I guess this explanation is totally incorrect??
Click to expand...

I was talking about drag, and specifically pressure drag. Bogataj is talking about lift.

But the lift involves boundary layers nevertheless...

The reason why a head+main combination works better is that it controls the BLs better. I know that Bogataj knows this, although he doesn't say this explicitly in your quoted part. Specifically, the interaction of the two sail elements allows a greater maximum lift to be attained before BL separation sets in. That is its main benefit over a single sail.

Note that if there were no BLs and no flow separation, a single sail would work just as well as a head+main sail, and in fact the sail's camber wouldn't matter either. Therefore, I think it's pointless to argue the aerodynamic pros and cons of different sail configurations without addressing the configuration's effect on BL behavior.

brian eiland · Mar 10, 2016

Speaking of boundary layers, here is an older quote of Tom Speer that I saved:

TomSpeer said:

The finest explanation I've seen of the slot effect was given by A.M.O.
Smith in his 1975 Wright Brothers Lecture ("High Lift Aerodynamics," AIAA
Journal of Aircraft, Vol. 12, No. 6, June, 1975, pp 501-530). This should
be available from any university engineering library.

Smith starts out by considering what limits the amount of lift a section can
provide and shows that there are two basic factors: when the local pressure reaches Mo^2*Cp = -1 [Mo is the freestream Mach number and Cp is the local pressure coefficient], and when the boundary layer separates. The first criterion places an absolute limit on the lift that can be produced regardless of the type of high lift device.

The second, the load-carrying capacity of the boundary layer, is what sets
the limit in nearly all practical cases. For attached flow, the velocity at
the trailing edge is a little slower than, and the pressure a little higher
than for the freestream conditions. Somewhere ahead of the trailing edge
the local velocity is higher and the pressure is lower. So the last thing
the boundary layer has to do as it gets to the trailing edge is to negotiate a region of increasing pressure and decelerating flow. If the increase in pressure is too much the slower flow in the boundary layer literally gets backed up. This is what causes separation - you have flow moving forward meeting flow being pushed backwards, both get brought to a stop at some point, and the only direction they can both go is away from the surface. So there's a limit to how suddenly the flow can be decelerated.

The harder you can decelerate the flow, the longer you can maintain the
higher velocity on the lee side and the more lift you can produce. It's
just like coming up to a stop sign - you can either coast into it from a
long way out, or you can come zooming in and slam on the brakes at the last minute. It turns out that the velocity profile for maximum deceleration of the flow has a concave shape - a steep drop at the beginning when the boundary layer is fresh, and a pronounced flattening out of the profile with distance from the start of the drop (kind of like a car whose brakes fade badly). If the maximum velocity is very high, the deceleration has to start much earlier and more of the airfoil's lee side has to be devoted to it. If the deceleration is started later, then the maximum velocity can't be as high, but the high speed/low pressure can extend over more of the surface and less of the surface has to be spent on the deceleration. Naturally, there's an optimum combination of the two which results in milking as much lift out of the lee side as you can get at a given Reynolds number. This is important for the interplay between main and jib. But put that aside for the moment and consider slot effect.

Here's what Smith says about slot effect:

"A remark about the action of slots comes from a recent NASA report: 'It is well recognized that the usual function of the slot is that of a
boundary-layer control device permitting highly adverse upper surface
pressure gradients to be sustained without incurring severe separation.
This stabilizing influence results from the injection of the high energy
slot flow into the upper surface boundary layer.' A still more recent NASA
report states, 'This leading-edge slat gives the fluid which passes through the gap between the slat and the main airfoil a high velocity.
Consequently, a boundary layer which grows on the upper surface of the main airfoil has more momentum than it would have in the absence of the slat.'

"There are two things wrong with these statements. First of all, the slat
does not give the air in the slot high velocity. If anything, it gives the
air low velocity. Secondly, the air through the slot cannot really be
called high-energy air. All the air outside the actual boundary layers has
the same total head. Properly designed and spaced slats are far enough
apart that each component develops its own boundary layer under the
influence of the main stream, and there is no merging within the slot.
Topologically, the process of boundary-layer development is no different
from that on a biplane. Subject to their particular pressure distributions,
the two boundary layers on a biplane grow, trail off downstream, diffuse,
and finally merge. That is just the process for an airfoil system of two or
more elements so long as merging does not occur within the slot.

"The next paragraphs will elaborate on and confirm what we have just said. There appear to be five primary effects of gaps, and here we speak of properly designed aerodynamic slots.
"1) Slat effect - in the vicinity of the leading edge of a downstream
element, the velocities due to circulation on a forward element, for
example, a slat, run counter to the velocities on the downstream element and so reduce pressure peaks on the downstream element.
"2) Circulation effect - in turn, the downstream element causes the trailing
edge of the adjacent upstream element to be in a region of high velocity
that is inclined to the mean line at the rear of the forward element. Such
flow inclination induces considerably greater circulation on the forward
element.
"3) Dumping effect - because the trailing edge of a forward element 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.
"4) Off-the-surface pressure recovery - the boundary layer from forward
elements 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 wall. Such a
method is more effective than the best possible deceleration in contact with a wall.
"5) Fresh boundary-layer effect - each new element starts out with a fresh boundary layer at its leading edge. Thin boundary layers can withstand stronger adverse gradients than thick ones."

Smith then goes on to discuss each one of the effects in more detail. The
fresh boundary layer effect is typically the only one people mention. The
dumping and circulation effects are responsible for the observation that a
jib is more effective per area than a main. The slat effect is important in
dealing with the flow around the mast.

Brian ask: THEN WHY do we continue to make the mainsail bigger than the jib???

Just because the circulation effect causes the jib to produce more lift on a
per area basis does not necessarily mean that the entire system would be more effective if the jib were made larger than the main. The jib and main can play different roles and the relative importance of each of the five factors above are different for each.

For example, consider a genoa with significant overlap. At the location of
the genoa trailing edge, the flow around the main, in the absence of the
jib, is substantially parallel to the trailing edge of the genoa. This
means the circulation effect is not as important. The circulation effect
can be likened to an increase in the angle of attack, because the presence of the downstream element and the angle of attack both increase the component of the flow normal to the trailing edge. This requires a larger circulation around the jib in order to balance the conditions there, resulting in more lift on the jib. With the large overlap, there's not much lateral component applied to the genoa's leech.

But the dumping velocity effect is very important. The main doesn't produce the sharp leading edge pressure peak of a thin airfoil because the peak is blunted by separation behind the mast. So putting the trailing edge of the genoa back where the highest velocity occurs maximizes the velocity at then trailing edge of the genoa. Since the trailing edge velocity is higher, the jib can carry a higher velocity farther aft on the lee side than it would otherwise. To go back to the braking metaphor, it's like the genoa's boundary layer only has to slow down for a yellow light, but not come to a full stop, so the braking can start later.

For a non-overlapping jib, the situation is different. The mast and main
cause a considerable lateral component at the trailing edge of the jib,
enhancing the circulation effect. The dumping velocity is still important
for the non-overlapping jib, too.

Now sailboats have to operate at a wide range of Reynolds numbers. To get a rough idea of the Reynolds number, take the wind speed in knots, multiply by the chord in feet, and multiply that by 10,000. So a 5 ft chord in a 5 kt wind is operating at a Reynolds number on the order of 250,000. At that Reynolds number, it takes a long ways for the flow to slow down to the sub-freestream velocity at the leech of the main. A laminar boundary layer can't do it at all in the length of the chord - so laminar separation will definitely occur, hopefully with reattachment to form a laminar separation bubble and turbulent boundary layer (most likely just behind the mast).

But the point is, pretty much the entire lee side of the main is fully
engaged in decelerating the flow in light wind/low Reynolds number
conditions. The shorter the main, the lower the maximum velocity on it can be. That means a lower dumping velocity for the jib, too, so more of the pressure recovery has to be done on the jib, which cuts into its lift
production. Is it better to have a short highly-loaded jib with high
dumping velocity and large main with a long pressure recovery and more
off-surface jib pressure recovery? Or is it better to have a large jib,
with less loading and lower dumping velocity, and a shorter main to finish
up the job? It's not a slam-dunk decision either way.

I've not tried to design main/jib combinations, but I have looked at rigid
wing/slotted flap configurations. The data are on my web site. I
considered flap sizes from 20% of the total chord to 50% of the total chord. Especially in the light wind conditions, the bigger flap (hence bigger main) was more effective. The 50/50 combination only starts to get into the genoa/main range, and I didn't look at anything like a slat/wing combo which would be more representative of a jib/main. But the trend did not point to a larger forward element being the better choice.

I would love to be able to look at this issue with the proper tools.
Unfortunately, the only programs I have for multiple elements can't handle separation, so they can't predict maximum lift or the onset of the drag rise due to stall. I had to do some really crude fudging to guess at stall effects. It would take something like MSES to calculate the two-dimensional flow around main and jib (with boundary layer), and that program is out of my price range. I've asked Drela about a multiple-element XFOIL, and he's mthought about it but hasn't pursued it.

The bottom line is that the jib can act as a slat for the main, in which
case a small jib makes sense. The main can act as a slotted flap for the
jib, in which case a large jib makes sense. Deciding between these two
extremes is a quantitative issue that can be answered with experience on the water, wind tunnel test, or computational methods. Each has its strengths and drawbacks.

Cheers,
Tom Speer
Click to expand...

Erwan · Mar 24, 2016

Thank you very much Mr Drela for this private lecture.

Fortunatly I have a very good CFD book, (ISBN-13: 978-0262526449) and your post prompted me to have a thourough reading of chapter 4

But unfortunatly, it still remains a bit hard to chew within a short period of time for a rookie like me. So I have to continue my homework before to be sure to actually understand.

In the exemple provided: Naca 01 vs Naca05 @ 5M reynolds;
Even if the relationship is not exactly linear , the average slope of the curve is really tiny +13% vs +500%

From this example, can we conclude that an appropriate XFOIL simulation @ 0° AoA could be a good proxy to measure the influence of thickness on profil drag ?

Best regards

Erwan

Log in or Sign up

Sail camber

Doug Halsey Senior Member

markdrela Senior Member

brian eiland Senior Member

Attached Files:

How-Sails-Work.pdf

markdrela Senior Member

brian eiland Senior Member

Erwan Senior Member

Overlapping schooner sail plan

Looking for AeroRig sail drawings

Fixed wing sail controls and drawings

Prandtl's lifting-line method for sails

Wing sail questions

rudder design - old sailing vessels

Ideal CE on Different Points of Sail

Bipedal land sailing robot! "Strandbeest"

Log/attempt for open source CFD / FEM framework for mast/sails simulation

Getting Coefficients of lift for single sail boats.