Foiler Design

[nflutter]

just on that point, i remember reading about a project in the UK (geko moth??) where a guy built a set of foils using the H105 section, and reported that the stabilising "squelch" effect as the foil apporached the waters surface was very pronounced. i havent heard of it being as obvious on other confiurations.

 

[Doug Lord]

Foil Drag

Tom, Mark etc-
Since there are such different requirements for light wind takeoff and high speed foiling would either of these ideas have any merit:
1) twist of the foil so that at high speed the outer sections of the foil develop little or no lift or
2) an outer section(each side) of ,say, 8" that was separated from the rest of the foil by a fence(if necessary) and that could be manually rotated slightly at high speed to reduce lift of the rotated section to near zero? The idea being to eliminate the induced drag from a section of foil that is not needed for lift at high speed. The area of the fixed and rotated sections would have to be carefully designed but does the concept appear to have any merit?

 

[tspeer]

Tom, Mark etc-
Since there are such different requirements for light wind takeoff and high speed foiling would either of these ideas have any merit:
1) twist of the foil so that at high speed the outer sections of the foil develop little or no lift or

Twisting the foil at high speed to transfer the lift inboard isn't going to help. The total lift will be unchanged - the lift has to equal the weight. But the effective span will be significantly reduced, increasing the induced drag.

The extra drag at high speed comes from the wetted surface area of the foil. If you make the foil large to take off at a low speed, you are still carrying that area at high speed and incur more profile drag (for a fully submerged foil). This is one of the attractions of a surface piercing foil system. It has lots of area for takeoff, and less area when flying at speed. But class rules have all but precluded the use of surface piercing foils.

2) an outer section(each side) of ,say, 8" that was separated from the rest of the foil by a fence(if necessary) and that could be manually rotated slightly at high speed to reduce lift of the rotated section to near zero? The idea being to eliminate the induced drag from a section of foil that is not needed for lift at high speed. The area of the fixed and rotated sections would have to be carefully designed but does the concept appear to have any merit?

You don't understand the nature of induced drag. Induced drag is inversely proportional to the square of the span of the lifting surface. If you cut the lifting span in half, you incur four times the induced drag. What you want to achieve is for the downwash velocity all along the span to be uniform and as small as possible. Concentrating the lift in the center half of the span means doubling the lift at each section and thus the downwash there (because the total lift still has to support the whole weight of the craft and now it's created by half the span), and you get the double-whammy of stronger downwash acting on stronger lift that produces the span-squared dependence.

What you want to do at high speed is get rid of area. It's like reefing. The total lift on a sail rig cannot exceed the boat's ability to stand up to it, so no matter how hard the wind blows, the lift is essentially determined by the boat's stability - and indeed, most designers size the standing rigging by referring to the boat's stability instead of trying to caclulate the actual air loads. In a high wind, you can flog the head of the sail or you reduce the area by reefing. If the heel angle is the same for both cases, they are both producing the same heeling moment and the induced drag is likely to be comparable. But the reefed sail has less drag than the twisted off sail. I think most sailors have had the experience of being reluctant to reef, and then discovering the boat is actually faster with less sail up. It's the same with hydrofoils with regard to the tradeoff between low takeoff speed and cruising at high speed.

 

[Doug Lord]

Foil Drag

Thanks ,Tom. To help me understand:if the foil tips described earlier were simply discarded (or retracted into the main section of the foil)wouldn't that still concentrate the lift over a shorter span? Could it be possible for the lift coeficient of the shorter foil to still be within the "drag bucket" of the foil section?
Are you saying that to "reef" a foil both the span and chord would have to be shortened to reduce area w/o too much drag?
Thanks for the help...

 

[tspeer]

Thanks ,Tom. To help me understand:if the foil tips described earlier were simply discarded (or retracted into the main section of the foil)wouldn't that still concentrate the lift over a shorter span? Could it be possible for the lift coeficient of the shorter foil to still be within the "drag bucket" of the foil section?
Are you saying that to "reef" a foil both the span and chord would have to be shortened to reduce area w/o too much drag?
Thanks for the help...

Dropping the tips would increase the induced drag, but would also reduce the parasite drag. Since induced drag goes down with speed, you might save more in parasite drag than you incur in induced drag and make the modification worthwhile.

There's no hydrodynamic advantage to reducing the span itself. Span is typically limited by other considerations, like structural strength and stiffness, class rules, minimum chord Reynolds number, etc.

The ideal way to reef would be to reduce the chord and keep the span the same. That is in effect what a ladder foil does, and A.G. Bell even called it "reefing" when rungs of his ladder foils came out of the water.

It's possible to still be in the drag bucket of a section after the tips were discarded, and presumably one would size things so that would be the case. You don't buy anything if you cut the area in half, only to have the profile drag coefficient double.

 

[bennn]

flap??

Hi guys,

Does anybody know what law should be used to calculate the lift and drag of a foil when the flap angle changes.

I mean a formula using maybe the angle of the flap, the general angle of the foil, the percentage of flap on the chord etc...

even if it's very simplified it's fine.

thanks

 

[tspeer]

Hi guys,

Does anybody know what law should be used to calculate the lift and drag of a foil when the flap angle changes...

There are four effects due to flap deflection you need to consider - section lift, profile drag, spanwise lift distribution, and induced drag. The first is the effect of flap deflection on the sectional (2D) lift coefficient. This can be calculated using thin airfoil theory, or (better) a program like XFOIL. It's the only one that has a simple relationship.

The second is the effect of flap deflection on the section drag coefficient. A program like XFOIL is really the only way to do this, as thin airfoil theory doesn't say anything about the section drag - the drag is due to the way the boundary layer development is affected by the change in lift distribution. So download XFOIL and possibly the graphical front end, Profili, which many people find easier to use than XFOIL's command-line interface. Then use XFOIL to get both the section lift and drag due to flap deflection.

Then there's the effect of flap deflection on the spanwise lift distribution due to wake (3D) effects. This will modify the lift at each section along the whole span and induced drag. If your flap is full span and a constant percentage of the chord along the span, then the lift and induced drag will be the same as for the undeflected flap at the same lift coefficient. But for the case where the flap only covers part of the span, or is proportioned differently along the span than the planform shape, the induced drag due to flap deflection will be higher than the corresponding change in angle of attack. The simplest way to estimate the 3D lift and angle of attack is with a lifting line analysis - the effect of flap deflection is to change the zero lift angle of attack of the section, which you can get from XFOIL. The next simplest way to estimate the lift and induced drag is by using a vortex lattice code.

A "2 1/2D" code like AVL will take in the section data, planform shape, and flap deflection, and put it all together for you.

 

[bennn]

okay, but using xfoil, you can only get the Cl and cd with one flap angle, for many general angle of attack... I would like to get the opposite !

Anyway the point of my question is how do you calculate those values during a simulation with matlab or something like that ? what transfer function can be used ?

If anybody did this kind of simul. with a flap...

thanks

 

[Stewpend0us]

Tom, Mark etc-
Since there are such different requirements for light wind takeoff and high speed foiling would either of these ideas have any merit:
1) twist of the foil so that at high speed the outer sections of the foil develop little or no lift or
2) an outer section(each side) of ,say, 8" that was separated from the rest of the foil by a fence(if necessary) and that could be manually rotated slightly at high speed to reduce lift of the rotated section to near zero? The idea being to eliminate the induced drag from a section of foil that is not needed for lift at high speed. The area of the fixed and rotated sections would have to be carefully designed but does the concept appear to have any merit?

When you say "twist of the foil" are you talking about looking at the boat from the side taking the foil and twisting it or looking from the top of the boat twisting it?

Please forgive me for not reading the entire post...yet. If this has been said before I’m very sorry.

Don't Navy fighter jets have this same problem? Need high lift to take off but then want lower lift/drag for higher speeds?

Looking from the top and twisting it would be similar to sweeping the wing back on a fighter jet.

I'm not sure how the difference in the compressibility of air vs. water will affect this comparison but still worth a shot I suppose. The crazy overpoweredness of fighter jets might through it off too..

If my understanding is correct you have a foil that is big enough to produce lift at low speeds but once you’re out of the water and moving it’s creating to much lift/drag to allow you to move faster. I would think that one could design a sweeping foil such that as the speed/lift/drag increased the foil would automatically be pushed back by the increasing drag. In this way the designer could design the foil for a constant drag/lift vs speed instead of an increasing drag/lift vs speed.

Let me know where I messed up!

 

[tspeer]

okay, but using xfoil, you can only get the Cl and cd with one flap angle, for many general angle of attack... I would like to get the opposite !

It's more awkward in XFOIL to slice the problem along lines of constant flap deflection, but you can do it. One way is to store each flap deflection as a separate section shape. Then use a script to flip between them as you calculate one point at a time at a constant angle of attack.

Probably the easiest thing to do is to calculate angle of attack polars, then read them into a spreadsheet and plot them vs flap deflection. XFOIL's polars are ASCII text files, so they read into Excel very simply - just use "space" as the delimiting character. Then create a summary sheet that links to the data in the sheets for each polar file.

Anyway the point of my question is how do you calculate those values during a simulation with matlab or something like that ? what transfer function can be used ?

What's typically done in a simulation is one builds tables of coefficients that are a function of independent variables like angle of attack, sideslip angles, control deflections, etc. Each point in the table is computed one point at a time. For example, in a simulation I'm building at work, we are using five-dimensional tables (5 independent variables, one dependent variable, and a different table for each dependent variable), and I'm planning a wind tunnel test to collect over 23,000 points to populate the tables.

Then you read the tables into the simulation at start-up, or use an autocode generator (which can be as simple as a spreadsheet macro) to embed them directly into the code. As the simulation runs, the dependent variables described by each table are interpolated based on the values of the independent variables at that time in the simulation run. There are efficient interpolation techniques that can handle large tables without getting bogged down, such as starting the search for the appropriate set of knots based on the place in the table from the previous time step, since a continuous system can't make big jumps if you are using an adequate iteration rate.

The number of points you have to compute explodes as you add independent variables if you're doing a full-factorial exploration of the design space. An alternative is to curve fit the data, often called a response surface, and use a design-of-experiments approach to cut down on the number of points you have to calculate to get a decent representation. Then you can evaluate the curve fit instead of having to do table look-ups.

State variable models are used in Matlab to calculate the transfer functions. These are bilinear systems in the form of
xdot = A*x + B*u
y = C*x + D*u

The independent variables in the tables will typically be states or functions of the states. Flap deflections are either treated as input controls ("u" vector), or are the output of subsidiary transfer functions that capture the sensor and actuator dynamics. You will probably have to differentiate the data tables (or curve fits) to get the derivatives of the accelerations (state rates) as a function of the independent variables so you can build the "A" and "B" matrices.

With a feedback control system like a Moth, the flap deflection becomes a function of the states in the control system, often of the form
u = K*(command-x)
So the state model of the closed loop system becomes
xdot = [A - B*K]*x + K*command
y = [C-D*K]x + D*K*command
The closed loop transfer function is calculated from
y/command = [C-D*K]*[s*I-(A - B*K)]^-1*K + D*K

At heart of it all are the tables computed one point at a time.

 

[Doug Lord]

Sweep

I'm really surprised that none of the foilers seem to recognize the need for sweep. It could significantly extend the range of subcavitating operation.

Tom, could you elaborate on this a bit. Say for the example of a Moth foil 120mm(4.72")chord by (914mm)36" . Area=1.18sq.ft. Foil loading @80%=
about 142 lb.per sq.ft. 63412 section. Top speed currently 29.7 knots; minimum foiling speed around 6.5 knots.
1) Do you have an approximate idea what the theoretical top sub-cavitating speed would be for this foil? Approximately how much could sweep increase the range of subcavitating operation? How much sweep?

2) How would sweep impact low speed takeoff?

3) Would the proper sweep for the main and rudder foil allow a decrease in area of the rudder foil due to enhancement of pitch control?

 

[bennn]

A couple of questions...

- What is a classical value of neutral angle of attack for both foils on a moth (I mean at zero deg. boat attitude)

- http://www.fastacraft.com/moulded_foils.html
what the hell is the material they use to build the link between foil and flap on the animation ?

- Tom speer was talking about a bulb that would reduce the junction effect on a T foil. Would it be enough to just to use this kind of smootjing just on the upper part of the foil ? Why have I never seen such a thing on a moth ?

 

[Phil Stevo]

AoA between hull WL and main foil chord line at zero flap deflection is about 1.5 to 2.5 degrees. Rudder foil is similar initially but BR and others have the whole rudder adjustable.

Faster craft used to use kevlar impreganated with rubberised epoxy for the hinge. Others have used rubber, sail cloth, mylar film and metal hinges.

I think the bulb Tom was talking about was copied to the Bladerider. It also increases the foil depth locally and allows the T junction to be built strong enough to be dismantleable.

All this knowledge pre-dates 2007 developments.

 

[bennn]

I'm considering building a twin tip hydroifoil that could be used in kiteboarding. that means it needs to be efficient in both direction (like a snowboard) which is pretty new for a profile....
Any idea on what profile could be appropriate ?
flat bottom ?

 

[MalSmith]

I'm considering building a twin tip hydroifoil that could be used in kiteboarding. that means it needs to be efficient in both direction (like a snowboard) which is pretty new for a profile....
Any idea on what profile could be appropriate ?
flat bottom ?

Tom Speer has designed some sections for proas which might be useful. Go to http://www.basiliscus.com/ProaSections/ProaIndex.html