Surface Piercing Foil Damping

[TheUnlogicalOne]

I'm far from a naval architect but I have a couple questions about surface piercing foils and would love some feedback. I considered this briefly when modeled a foiling sailboat a while back. I searched around and couldn't find any information on this specifically.

My understanding is that surface piercing foils tend to give a rough ride over waves. Has anyone ever added small control surfaces to counter act this?

I would envision this being controlled by a fairly simple spring and damper system attached to the control surface. The system could be easily disabled if speed was more important than comfort, removing the extra drag caused by the control surfaces.

The foil is supported by a spring the same way a car's wheel is. The damper is mounted between the control surface and the boat. Any sudden movements of the foil would cause the control surface to defect as the damper takes time to move. Slow changes (such as might be found in a sail boat changing direction) would have almost no affect on the control surface because the damper applies force proportionally to the speed of the change.

Also at high speed the force required to move the control surfaces would be greater so less deflection would occur.

I have a few concerns though:
-Is it possible to design a control surface which, when at speed, will centre to an angle that will produce lift or will it always try to assume the angle which produces the least possible lift?

-I have 3 ideas for mounting the spring/damper assembly but I don't have enough experience (AKA, none) to decide which would be most practical. I recall reading somewhere that a V shaped foil can be unstable as it leaves the water (See images 2 and 3) however it should increase low speed performance (I think).

This is just a very rough idea I had and needs a lot of work. (if it is even feasible)

I have included a few doodles of what I was thinking. Please excuse the poor quality, my tablet is broken and I can't draw with a mouse at all.

The green represents possible locations for the spring/damper and the red is what I envision the control surface would look like.

Thanks in advance,

[tspeer]

Originally Posted by TheUnlogicalOne

...My understanding is that surface piercing foils tend to give a rough ride over waves. Has anyone ever added small control surfaces to counter act this?

The Canadian frigate, the Bras d'Or, used active control surfaces on surface piercing foils.

Active control is essential for fully submerged foils, but surface piercing foils can benefit from active control as well.

I would envision this being controlled by a fairly simple spring and damper system attached to the control surface. The system could be easily disabled if speed was more important than comfort, removing the extra drag caused by the control surfaces.

Hydroptere uses shock absorbers similar to your second arrangement. Although in that case, the whole foil is allowed to move up and down without actuating a flap. When pushed up, the vertical velocity of the foil itself reduces the angle of attack and helps to alleviate the loads.

Why do you think control surfaces would add drag? If you null out the accelerations caused by changing conditions on the foils, the foil is operating at constant lift. That is more likely to reduce drag.

When you design a flapped section for a surface piercing hydrofoil, there are three design conditions to consider. The first is the nominal angle of attack and zero flap deflection. Then there is positive flap deflection and negative angle of attack, and negative flap deflection with positive angle of attack. All three design conditions have the same lift coefficient. The range of angles of attack at which you can maintain a constant lift coefficient by deflecting the flap defines the sea state at which the craft can operate in the platforming mode (flying at constant altitude and letting the waves go by underneath).

...
-Is it possible to design a control surface which, when at speed, will centre to an angle that will produce lift or will it always try to assume the angle which produces the least possible lift?

It is possible, however that would be an over-balanced control surface. It would tend to be unstable, since an increase in angle of attack would tend to increase the flap deflection and increase the lift even more. And vice versa when the angle of attack reversed.

A better solution is to have a stable hinge moment and a spring (or a spring-loaded tab) that pulls down on the surface, opposing the hinge moment. Then, as the speed increases, the surface overcomes the spring and floats up, reducing the lift at the same angle of attack. A sudden increase in angle of attack will also have the surface float up, reducing the impact of the change. This provides a shock-absorber effect.

...I recall reading somewhere that a V shaped foil can be unstable as it leaves the water ...

Static stability - the tendency to return toward an equilibrium state when disturbed - typically requires that the forward foil be more heavily loaded than the aft foil, and the forward foil have a greater heave stiffness than the aft foil. The center of gravity also has to be far enough ahead that a positive pitching moment (bow up) is required to trim.

The higher loading on the front foil ensures that when the angle of attack increases, the aft foil will develop proportionately more lift and generate a negative pitching moment to reduce the angle of attack back toward the trim value.

The heave stiffness must be such that as the craft rises, the foils lose lift and cause the craft to descend back toward the trim height. The higher heave stiffness of the forward foil means that as the craft rises up in a level attitude, the forward foil loses more lift than the aft foil. This causes a bow-down pitching moment that tends to reduce the lift and return toward equilibrium. If the stern foil has a higher heave stiffness, then as the craft rises, the bottom falls out at the stern, pitching the craft up, and this unstable pitch-heave coupling can be greater than the reduction in lift due to the rise itself, causing the craft to rise further. And pitch more, and rise more, until it shoots up out of the water.

The positive moment for trim means that as the craft accelerates, it tends to pitch up. This makes it rise, slowing it because it's now flying uphill.

All these effects deal with the static stability. There's also the issue of dynamic stability. It's not enough to have a tendency to return to the trimmed equilibrium condition. If, when returning to trim, the craft does so too fast, it will overshoot to more than the original disturbance. Which will make it correct back even harder the other way, and gaining energy and getting wilder on each swing.

And finally, there's the issue of trimming out large external loads, such as moments from the sail rig. This is somewhat different from stability itself, but the two often go together. If there's no means of applying trim moments, then what happens is when an external load is applied, the craft will be disturbed until the restoring forces and moments arising from its static stability equal the external load, and a new trim equilibrium is established at a different attitude, height, or speed (or all three). If there's no additional trim control, the stability can soon be overwhelmed by the external loads. On a typical Moth setup, for example, the flap on the forward foil handles stability by giving the main foil the heave stiffness it needs through feedback from the wand. There's another flap on the stern foil that is controlled by the sailor to adjust the trim.

Trim control is especially important for surface piercing foils because trim has a big effect on performance. For example, if a V foil is operated at a constant lift coefficient so the change in lift with speed is entirely compensated for by rising up and reducing area, the reduction in span can result in the induced drag increasing with speed instead of decreasing with speed like it normally does! But if the V foil is trimmed down as speed increases, keeping more of the foil in the water, the reduction in wetted area with speed can balance the (smaller) reduction in span in such a way that the V foil drag does not change with speed. This is something that can't be achieved with a fully submerged foil.

[TheUnlogicalOne]

Thanks for the feed back!

The Bras d'Or is definitely a neat ship. Its too bad its up on blocks but I suppose that's better than being scraped.

Originally Posted by tspeer

When you design a flapped section for a surface piercing hydrofoil, there are three design conditions to consider. The first is the nominal angle of attack and zero flap deflection. Then there is positive flap deflection and negative angle of attack, and negative flap deflection with positive angle of attack. All three design conditions have the same lift coefficient. The range of angles of attack at which you can maintain a constant lift coefficient by deflecting the flap defines the sea state at which the craft can operate in the platforming mode (flying at constant altitude and letting the waves go by underneath).

Maybe it would be possible to have a dial of some sort to control the coefficient of damping to adapt to different conditions. As far as I can see changing the damping coefficient should allow tuning of the response to keep a fairly constant coefficient of lift in a range of conditions.

I was thinking that this setup could be added without significantly changing the static stability of a design. While there is some deflection of the foil over waves it shouldn't be extremely large. Also because just a damper is used the flaps would return to a neutral position no mater what the loading of a given foil was.

I probably was a bit confusing with the V foil question. It relates to the joint between the main foil and the support.

The rational behind option 2 and 3 was to significantly increase the foil area at low speeds but it seems like there may be a problem as the joint leaves the water. There would be a sudden loss of lift as the nacelle around the joint nears the surface and a bouncing effect might occur. Maybe the dampers would deal with it. I'm not sure if there would really be a loss of lift.

Regardless there would be extra drag and if the rotational joint is under water it would likely end up being a bit of a maintenance nightmare.

For example, if a V foil is operated at a constant lift coefficient so the change in lift with speed is entirely compensated for by rising up and reducing area, the reduction in span can result in the induced drag increasing with speed instead of decreasing with speed like it normally does!

Sigh, aerodynamics always gives me a headache. I think I understand. I guess that since the angle of attack increases with speed it wouldn't be possible to just decrease incidence of the tips? They would produce little or negative lift at low speeds. Would a winglet help? If the control flaps are spring loaded then they should help keep the foil trimmed down at high speed.

[tspeer]

Originally Posted by TheUnlogicalOne

...While there is some deflection of the foil over waves it shouldn't be extremely large. Also because just a damper is used the flaps would return to a neutral position no mater what the loading of a given foil was.

The change in loading on a surface piercing foil in a seaway can be tremendous. When the foil enters a crest from a trough, the area can double or more, causing a very sudden increase in lift. Hydroptere's first beams were turned into pretzels in just such an event.

It may be desirable to have considerable deflection of the foil. And it may also be desirable to orient the axis of rotation at an angle to the hull reference line, similar to a "delta3 hinge" on a helicopter rotor. By toeing the hinge line inboard a few degrees, the foil will change its incidence as it rotates up and down. Thus, as the loading increases and drives the foil up, the angle of attack decreases, alleviating some of the load. And as the foil moves down, the angle of attack will increase. This can be used in conjunction with the foil's dihedral, taper and spring stiffness to tune the dynamic response.

It's not possible to use only a damper. There must be some sort of spring support as well. The idea of washing out the flap deflection is a good one, however.

I...Sigh, aerodynamics always gives me a headache. I think I understand. I guess that since the angle of attack increases with speed it wouldn't be possible to just decrease incidence of the tips? They would produce little or negative lift at low speeds. Would a winglet help? If the control flaps are spring loaded then they should help keep the foil trimmed down at high speed.

You would want the angle of attack to decrease with speed, not increase. And at low speeds, you want the entire foil lifting so as to minimize the wetted area of the foil required and to make use of the entire span of the foil to reduce induced drag.

Just as the buoyancy of a hull equals the weight of the boat under steady conditions, regardless of the shape of the hull, the lift from the foils must equal the weight of the boat when flying, regardless of the speed. The lift is proportional to the angle of attack (measured from the zero lift line) times the area times the speed squared. If this product is to remain constant as the speed changes, the angle of attack must decrease or the zero lift line must be changed (this is what a flap does) or the area must decrease (surface piercing foil). Or a combination of these three.

The area has the biggest effect on performance, so there may well be an optimum relationship of steady-state area vs velocity. The flap is the fastest acting of the three, but it has limited authority. Changing the pitch attitude of the boat is slower than moving a flap, but faster than changing altitude. So there are three different time scales that can be exploited in controlling the foiler. The flap can be used for near-instantaneous changes in lift. If the pitch attitude is then used to change the angle of attack, the required flap deflection will naturally wash back out to zero. The ultimate goal of flying the boat in pitch would be to maintain the desired altitude/area for the speed.

This suggests a control law that has three nested loops. The inner loop would be closed on vertical acceleration, and would actuate the flap on the main foil. When commanded with zero vertical acceleration, this would cause the flap to be actuated to smooth out the ride and minimize the vertical disturbance from the waves. Because vertical acceleration is proportional to the square of the encounter frequency, this automatically does a great deal to platform the smaller waves and does not interfere very much with contouring the larger waves.

The middle loop would be closed on pitch rate and actuate primarily the stern foil to change the pitch attitude and augment the pitch damping. The outer loop would be closed on height. The height command would come from a schedule of optimum height versus velocity, possibly with the velocity heavily low-pass filtered so as to ignore the effects of small waves. Since height is measured relative to the water's surface, waves with frequencies lower than the bandwidth of the low-pass filter will be contoured, while waves with higher frequencies will tend to be ignored, resulting in platforming of those wavelengths.

The pitch rate command would be proportional to the error in height. The vertical acceleration command would be a weighted sum of the error in height and the error in pitch rate.

Of course, an alternate way of designing the control laws is just to feedback everything - height, pitch angle, pitch rate, vertical acceleration, and use a linear quadratic regulator design to determine how much of each is fed to the main foil flaps and the stern foil elevator. Then all one has to do is tune the Q and R weighting matrices!

[BMcF]

"This suggests a control law that has three nested loops. The inner loop would be closed on vertical acceleration, and would actuate the flap on the main foil. When commanded with zero vertical acceleration, this would cause the flap to be actuated to smooth out the ride and minimize the vertical disturbance from the waves. Because vertical acceleration is proportional to the square of the encounter frequency, this automatically does a great deal to platform the smaller waves and does not interfere very much with contouring the larger waves.

The middle loop would be closed on pitch rate and actuate primarily the stern foil to change the pitch attitude and augment the pitch damping. The outer loop would be closed on height. The height command would come from a schedule of optimum height versus velocity, possibly with the velocity heavily low-pass filtered so as to ignore the effects of small waves. Since height is measured relative to the water's surface, waves with frequencies lower than the bandwidth of the low-pass filter will be contoured, while waves with higher frequencies will tend to be ignored, resulting in platforming of those wavelengths.

The pitch rate command would be proportional to the error in height. The vertical acceleration command would be a weighted sum of the error in height and the error in pitch rate.

Of course, an alternate way of designing the control laws is just to feedback everything - height, pitch angle, pitch rate, vertical acceleration, and use a linear quadratic regulator design to determine how much of each is fed to the main foil flaps and the stern foil elevator. Then all one has to do is tune the Q and R weighting matrices!
"

wow..just give it all way, Tom!.
A very good synopsis of 'how its actually done" (and that is what our company does do). The most recent hydrofoil control configuration we deployed uses a combination of bow height (sensed by microwave radar), bow acceleration filtered to obtain bow vertical velocity, heave (CG) velocity, roll rate and roll angle, and pitch rate nd pitch angle. Commanded bow height is 'slaved' to real-time vessel speed to aid in the take-off regime and reduce operator load.

All of the active control surfaces have servos that respond to command inputs at least 10 times faster than the motion to be controlled in each degree of freedom.so, if the peak pitch rate of a craft is 3 degrees/sec, then the control surface opposing that motion better be able to respond at 30 deg/sec..and so on. Without getting in to the control theory behind that, it should illustrate why a 'passive' mass-spring-damper approach cannot work. In most cases, it could even be quite unstable and make the motions associated with that degree of freedom quite worse...not better. That said, in theory there is always some (very narrow spectrum and specific) combination of wave encounter frequency and fundamental craft motion for which a spring-damper might appear on paper to work. In practice, however mother nature does not cooperate. Much the same as she seldom provides the conditions where wavepiercing hulls actually pierce waves; moreoften providing conditions where they pitch violently in resonance instead.

[kach22i]

Originally Posted by BMcF

All of the active control surfaces have servos that respond to command inputs at least 10 times faster than the motion to be controlled in each degree of freedom.

Do you use this same 10X rule for the SES (Surface Effect Ship) craft you help design?

[BMcF]

Originally Posted by kach22i

Do you use this same 10X rule for the SES (Surface Effect Ship) craft you help design?

More or less..but the challenge is greater. In that case..for a 35m SES, for example, the vent valve louver control servos must be good to in excess of 20 cycles/second (!) ('good' meaing, in control terminology, the -3db corner frequency of the servo) to effectively deal with the fundamental air cushion resonace and the first longtudinal acoustic mode. In the 35m example..those cushion responses occur at 2 Hz and 5 Hz, respectively. They are, if not actively damped, what cause the 'harsh cobblestone' ride characteristcs so often attributed to SES'.

[kach22i]

Thanks BMcF, I do not mean to hi-jack the thread.

Back to the topic, and just for wacky (but cool) reference to a mechanical suspension system (non-foil).

The spider boat:

SAN FRANCISCO
SPIDER SHIP ON THE BAY
El Cerrito firm unveils the Proteus, 'a new class of vessel'
http://www.sfgate.com/cgi-bin/articl...AGE7NLI001.DTL

Using technology developed by Conti's El Cerrito Marine Advanced Research Inc., the WAM-V is "a new class of watercraft ... that delivers a radically new seagoing experience." It has twin hulls, like a catamaran, connected to each other and a control cabin by four metal legs. The legs ride on titanium springs -- like shock absorbers -- that allow the WAM-V to adjust to the surface of the water -- to flex like knees.

[tspeer]

Originally Posted by kach22i

Do you use this same 10X rule for the SES (Surface Effect Ship) craft you help design?

The factor of ten rule of thumb comes from not wanting to add to the phase lag of the control loop, which would be destabilizing.

Let's say the actuator can be modeled as a third order response to a position command - a pair of high frequency poles determined by the servo loop closed on the control valve, and a lower frequency real pole that is determined by the hydraulic flow to the actuator. The single pole will dominate the response for the frequency range we're talking about.

Here's a table of actuator response vs the ratio of the commanded frequency divided by the corner frequency of the actuator dynamics:
freq magnitude phase, deg
0.01 0.999950 0.5729
0.05 0.998752 2.8624
0.10 0.995037 5.7106
0.20 0.980581 11.3099
0.50 0.894427 26.5651
1.00 0.707107 45.0000

If the actuator is 10 times faster than the bandwidth you're trying to control, it only adds 6 degrees of phase lag. A typical design requirement is to have a minimum of 30 - 45 degrees of phase margin, so this represents as much as 20% of your stability margin. If the actuator is only 5 times faster than what you're trying to control, you've lost a third of your stability margin to the actuator alone. If the actuator is any slower than that, it's going to be way too sluggish to do the job.

The same principle applies to the nested loops in the control law. Each loop needs to be faster than the next outer loop by a factor of around 5 or more. So that outermost height loop is going to have a fairly long period. That's OK, because it is mainly dealing with swells. The faster, inner loops are dealing with ride quality and vehicle stability. But it does drive the actuator to have a high bandwidth when you take on those faster tasks. Something in the range of 6 - 10 hz would not be out of order.

[kach22i]

Thanks to all for the very detailed engineering information. I must confess that some of it has gone over my head and will take several readings to soak in.

When I get an idea, I take it beyond the reasonable just for fun.

From an old post:
High Speed Boat

Originally Posted by kach22i

I'm going to get a little crazy on you for a moment, I trust you don't mind.

1. Even a super thin F-104 style vertical fin connecting the hull to the submerged pod will cause an explosion of water and a heck of a lot of spray.

2. Could the leading edges of the vertical supports also bleed out air?

3. This is the wild idea part; maybe magnalift support between the hull and pods? Some electro-magnets would be lifting, some pulling to maintain a constant distance.

4. Magnetic fields just to far fetched at this point in time, too Sci-Fi?

5. The Shikval is a kind of underwater hovercraft in my eyes, I like the concept. I've read about it for the first time only a few years ago. Perhaps in this forum.

http://s184.photobucket.com/albums/x295/kach22i/

I'm re-posting this because nothing I can think of will act/react faster than the electromagnetic levitation system illustrated above.

[BMcF]

Originally Posted by tspeer

The factor of ten rule of thumb comes from not wanting to add to the phase lag of the control loop, which would be destabilizing.

Let's say the actuator can be modeled as a third order response to a position command - a pair of high frequency poles determined by the servo loop closed on the control valve, and a lower frequency real pole that is determined by the hydraulic flow to the actuator. The single pole will dominate the response for the frequency range we're talking about.

Here's a table of actuator response vs the ratio of the commanded frequency divided by the corner frequency of the actuator dynamics:
freq magnitude phase, deg
0.01 0.999950 0.5729
0.05 0.998752 2.8624
0.10 0.995037 5.7106
0.20 0.980581 11.3099
0.50 0.894427 26.5651
1.00 0.707107 45.0000

If the actuator is 10 times faster than the bandwidth you're trying to control, it only adds 6 degrees of phase lag. A typical design requirement is to have a minimum of 30 - 45 degrees of phase margin, so this represents as much as 20% of your stability margin. If the actuator is only 5 times faster than what you're trying to control, you've lost a third of your stability margin to the actuator alone. If the actuator is any slower than that, it's going to be way too sluggish to do the job.

The same principle applies to the nested loops in the control law. Each loop needs to be faster than the next outer loop by a factor of around 5 or more. So that outermost height loop is going to have a fairly long period. That's OK, because it is mainly dealing with swells. The faster, inner loops are dealing with ride quality and vehicle stability. But it does drive the actuator to have a high bandwidth when you take on those faster tasks. Something in the range of 6 - 10 hz would not be out of order.

Tom has done a good job of explaining the 'why' behind the design and fielding of successful active motion control packages and particularly why they 'seem' to the layman and many vessel owners to have inordinately high power requirements supporting what also seem to be, at first glance, inordinately high actuation rates. Since most vessels have natural motion periods on the order of many seconds, it is often hard to imagine why all the 'stuff' required to control those motions is designed with responses measured in milliseconds. Some range numbers from a 'typical non-SES stabilization system' that might be interesting:

Fin/flap/interceptor/trim tab ('effectors') peak actuation rates: 25-60 deg/sec

Digital servo loop cyle (interrupt) rate: 1-2 KHz

Motion Reference Unit ('gyro') data aquisition and pre-filtering rate: 2 Khz

Control Software 'throughput' rate (end-to-end data rate from processing motion inputs through all control paths to generation of effector command outputs): 200-500 Hz

The ranges above cover a fairly wide range range of vessels types and sizes..a big SWATH requiring slower rates..a fast planing monohull or hydrofoil requiring the fastest..and everything else in between.

[sparky_wap]

For my next electric powered 'rowboat', I was considering using foils to reduce the drag for a higher max speed. After skimming through this thread, I think that ladder foils might be easier than an active control system. Does anyone know how to set up a ladder foil for automatic lift stability? This is a small boat run in 2-3 feet of water weighed down with heavy lead acid batteries. Power level will be between 10-15 HP.

I'm no stranger to control systems but have no budget for home projects. I would prefer a deployable foil that requires no adjustments for speed.

Is there a site to download foil shape layouts?

[tspeer]

Originally Posted by sparky_wap

For my next electric powered 'rowboat', I was considering using foils to reduce the drag for a higher max speed. [This is a small boat run in 2-3 feet of water weighed down with heavy lead acid batteries. Power level will be between 10-15 HP.]

I'm not sure foils will improve your performance at all. I think a long, narrow displacement hull would be best. Foil drag due to lift is very high when trying to fly at low speeds (say, <12 kt).

Besides, ladder foils are going to be really slow when they're stuck in the bottom.

...Is there a site to download foil shape layouts?

Not really. Take a look at the material on foils.org.

[gratko]

I think that the idea has a great potential. You can find something like this at www.americanhydrofoil.com but after many years this site remains unchanged.

[sparky_wap]

Gratko,

Thanks, That's what I was looking for.

Joe