Jump to content

Boats and foils comparison


Recommended Posts

16 hours ago, sfigone said:

But if the foil flaps have external pivot points and are just activated by sliding out, then would they not be able to be set for negative lift?

At least, this would respect the aeronautical nomenclature, where flaps only provide positive lift. Otherwise, they should be called ailerons -_-

Link to post
Share on other sites
  • Replies 4.1k
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

Just a few interesting bits of the straight-line performances from today: Upwind /Downwind VMGs - race 1: Upwind /Downwind VMGs - race 2: Same story in both races actually.

Thanks to weta27's pics I have created an approximation of NZ's "BFB v2" foil. Main points: Foil area is almost the same, possibly even a smidge larger. Flaps have increased in area as

OK, it sounds like there's some interest in this topic, so here goes.   Any engineering effort starts by defining the requirements.  From this figure, it looks like the average foil area is 1.64

Posted Images

In aircraft, flaps are deflected symmetrically, while ailerons are deflected differentially.  Sailplanes use negative flap deflection for high speed cruise, so flaps don't only provide positive lift.  The Design Rule requires the two surfaces on a wing to be deflected symmetrically, so they are flaps, not ailerons.

With all the discussion about foil shapes, I think it is useful to start with the attached paper.  However, there are other design philosophies that can be used.  Eppler didn't concern himself with leading edge cavitation at takeoff, and he wasn't trying to maximize the thickness of the foil for structural or ballast considerations, so his approach needs to be modified a bit for AC foils.  

Section_Design_for_Hydrofoil_Wings_with_Flaps.pdf

  • Like 1
Link to post
Share on other sites
17 minutes ago, Basiliscus said:

In aircraft, flaps are deflected symmetrically, while ailerons are deflected differentially.  Sailplanes use negative flap deflection for high speed cruise, so flaps don't only provide positive lift.  The Design Rule requires the two surfaces on a wing to be deflected symmetrically, so they are flaps, not ailerons.

With all the discussion about foil shapes, I think it is useful to start with the attached paper.  However, there are other design philosophies that can be used.  Eppler didn't concern himself with leading edge cavitation at takeoff, and he wasn't trying to maximize the thickness of the foil for structural or ballast considerations, so his approach needs to be modified a bit for AC foils.  

Section_Design_for_Hydrofoil_Wings_with_Flaps.pdf

I'm not seeing that in 15. Foil Flaps

Is it hiding somewhere else?

- Symmetric to each other: yes

- With the same range of motion: yes

- Always the same deflection: ??

Link to post
Share on other sites
9 hours ago, Xlot said:

At least, this would respect the aeronautical nomenclature, where flaps only provide positive lift. Otherwise, they should be called ailerons -_-

Crikey in that case they might be called flaperons. 

  • Like 2
Link to post
Share on other sites
2 hours ago, amc said:

Crikey in that case they might be called flaperons. 

Quite right. And the rudder foil would be termed a stabilator. :)

  • Like 2
Link to post
Share on other sites

Looking at the TNZ latest foils it reminds me of “The area rule “ that all you aero buffs will know relates to cross sectional area consistency to minimise drag is this relevant ( yes I know it has been a bit slow on this topic so come on give it a go!)

Link to post
Share on other sites
2 hours ago, oatsandbeans said:

Looking at the TNZ latest foils it reminds me of “The area rule “ that all you aero buffs will know relates to cross sectional area consistency to minimise drag is this relevant ( yes I know it has been a bit slow on this topic so come on give it a go!)

Would love to but hard when you know fuck all about the technicalities. But I am enjoying the education and it certainly adds to my appreciation of what’s involved. So keep them coming. 
 

 

Link to post
Share on other sites
2 hours ago, oatsandbeans said:

Looking at the TNZ latest foils it reminds me of “The area rule “ that all you aero buffs will know relates to cross sectional area consistency to minimise drag is this relevant ( yes I know it has been a bit slow on this topic so come on give it a go!)

Hmm that is considered for transonic airflow, not sure water is compressible. 

Link to post
Share on other sites

Nice one... I specialty like the confusion with making the loser L hand sign backwards and upside down.... obviously almost as hard as flying an AC75 :)

I still don't really get his ETNZ can balance lift and leeway with a flat foil and a "single" flap.  Surely when going downwind you want to configure foil to get much the same lift but negative leeway. I can see how y foils work for that.

  • Like 2
Link to post
Share on other sites
1 hour ago, sfigone said:

Nice one... I specialty like the confusion with making the loser L hand sign backwards and upside down.... obviously almost as hard as flying an AC75 :)

I thought about cutting that out... but it was too funny, so I had to leave it in!

Link to post
Share on other sites
6 hours ago, sfigone said:

Nice one... I specialty like the confusion with making the loser L hand sign backwards and upside down.... obviously almost as hard as flying an AC75 :)

I still don't really get his ETNZ can balance lift and leeway with a flat foil and a "single" flap.  Surely when going downwind you want to configure foil to get much the same lift but negative leeway. I can see how y foils work for that.

 

NZ doesn't have a "single flap", the rules require one flap per foil wing.

I also don't believe that NZ uses one actuator to control both flaps, as some have suggested. Two actuators can obviously be operated in synch if they wanted to, but I'd bet they do use some differential to adjust leeway etc.

  • Like 1
Link to post
Share on other sites
6 hours ago, amc said:

Hmm that is considered for transonic airflow, not sure water is compressible. 

Water is compressible. It's density can increase with pressure and temperature. But for our sailing-related purposes it can be considered incompressible, as can air.

Link to post
Share on other sites
1 hour ago, MaxHugen said:

NZ doesn't have a "single flap", the rules require one flap per foil wing.

I also don't believe that NZ uses one actuator to control both flaps, as some have suggested. Two actuators can obviously be operated in synch if they wanted to, but I'd bet they do use some differential to adjust leeway etc.

What evidence is there for a split? I cannot see a split on or either side of the symmetry line in any photos or footage. Most significant line I can see in an image is below. 

The rules say two flaps, but the interp says the two flaps may be joined with one foil system linkage. 

However, on W3 and W4 you could see foil system material between the two flaps on the symmetry line. 

Could they have two actuators working in sync? Possibly, but if they dont need the power and the flaps move together through the foil system link, the  why would they bulk out the bulb with more actuators?

Screenshot_20210210-061712_Chrome.jpg

Link to post
Share on other sites

Is there any interest in a series of technical posts that would illustrate the issues in foil wing section design by starting with the E908 section in this paper and modifying it to make it more suitable for the AC75?  It'd be too long for a single post, but I'm thinking that by giving each stage in the process its own post that it might be more digestible.

  • Like 14
Link to post
Share on other sites
8 hours ago, amc said:

Hmm that is considered for transonic airflow, not sure water is compressible. 

Actually something like the area rule is useful for minimizing cavitation and reducing junction drag.  Paul Bieker applied the principle on the AC72 rudders, and Boeing used it on the Jetfoil pods.  But I haven't seen a precise definition of the area rule for the purpose.  The transonic area rule is concerned with the farfield wave drag, so it includes all of the area in the cross section.  For a wing/strut junction, there should probably be some kind of radial weighting of the cross section area because the influence of the junction drops off with distance and the outer parts of the wing don't have a lot of influence on the junction.   

There's probably a PhD dissertation in there for anyone that wants to take it on.

Link to post
Share on other sites
3 hours ago, Mozzy Sails said:

What evidence is there for a split? I cannot see a split on or either side of the symmetry line in any photos or footage. Most significant line I can see in an image is below. 

The rules say two flaps, but the interp says the two flaps may be joined with one foil system linkage. 

However, on W3 and W4 you could see foil system material between the two flaps on the symmetry line. 

Could they have two actuators working in sync? Possibly, but if they dont need the power and the flaps move together through the foil system link, the  why would they bulk out the bulb with more actuators?

I'm not sure what you mean by a "split".

I don't think we have any information to confirm that both flaps always move in unison. If there are advantages in modifying the vertical vs lateral forces or moments via differential flap settings, I think NZ would have added that capability.

Of course it's quite possible that it was decided that any gain was not worthwhile, and used a single actuator.  Can't really say one way or the other.

Plus I have no idea how large a "single body - twin actuator" would actually be?

Link to post
Share on other sites
38 minutes ago, MaxHugen said:

I'm not sure what you mean by a "split".

I don't think we have any information to confirm that both flaps always move in unison. If there are advantages in modifying the vertical vs lateral forces or moments via differential flap settings, I think NZ would have added that capability.

Of course it's quite possible that it was decided that any gain was not worthwhile, and used a single actuator.  Can't really say one way or the other.

Plus I have no idea how large a "single body - twin actuator" would actually be?

If they were split, I would expect to see that split on the trailing edge in some photos. Or light leaking through. 
The ETNZ the only flaps that are not endplate inboard. If they don't move in unison, then I would imagine that would be pretty ugly flow.
Plus, we have the two rule interpretations which are clearly made by or aimed at ETNZ. 

So, a fair amount (although some circumstantial) evidence that they have two flaps joined by 'foil system' to act as one flap. 

However, they could be using a double actuators for extra power... would have to check the rules... perhaps that's what the latest actuator rule interp was all about. https://docs.google.com/a/acofficials.org/viewer?a=v&pid=sites&srcid=YWNvZmZpY2lhbHMub3JnfGFjMzYtb2ZmaWNpYWwtbm90aWNlYm9hcmR8Z3g6ZWQyNmVkMGI1YjRjNDQ4

  • Like 1
Link to post
Share on other sites
21 minutes ago, Mozzy Sails said:

Mozzy - Just want to thank you for all the videos, which are wonderful. Makes me think what a great job they did when they produced the rule. Keep 'em coming.

Link to post
Share on other sites
1 hour ago, Mozzy Sails said:

If they were split, I would expect to see that split on the trailing edge in some photos. Or light leaking through. 
The ETNZ the only flaps that are not endplate inboard. If they don't move in unison, then I would imagine that would be pretty ugly flow.
Plus, we have the two rule interpretations which are clearly made by or aimed at ETNZ. 

So, a fair amount (although some circumstantial) evidence that they have two flaps joined by 'foil system' to act as one flap. 

However, they could be using a double actuators for extra power... would have to check the rules... perhaps that's what the latest actuator rule interp was all about. https://docs.google.com/a/acofficials.org/viewer?a=v&pid=sites&srcid=YWNvZmZpY2lhbHMub3JnfGFjMzYtb2ZmaWNpYWwtbm90aWNlYm9hcmR8Z3g6ZWQyNmVkMGI1YjRjNDQ4

I just interpreted this as a clarification that 2 actuators housed in one body was OK. Nothing there as 'evidence' that 2 flaps are acting as one flap, to me anyway.

Also, the angle of these flaps can be quite minimal to exert an effect, so may not be as ugly as you'd expect.

  • Like 1
Link to post
Share on other sites
On 2/9/2021 at 5:44 AM, Basiliscus said:

In aircraft, flaps are deflected symmetrically, while ailerons are deflected differentially.  Sailplanes use negative flap deflection for high speed cruise, so flaps don't only provide positive lift.  The Design Rule requires the two surfaces on a wing to be deflected symmetrically, so they are flaps, not ailerons.

With all the discussion about foil shapes, I think it is useful to start with the attached paper.  However, there are other design philosophies that can be used.  Eppler didn't concern himself with leading edge cavitation at takeoff, and he wasn't trying to maximize the thickness of the foil for structural or ballast considerations, so his approach needs to be modified a bit for AC foils.  

Section_Design_for_Hydrofoil_Wings_with_Flaps.pdf

My LR-2369 foil profile was estimated months ago by approximating the upper surface of LR's foil from photos, while the lower surface was guesswork. For calculations, the NZ foil was used for Area and the position of the Flap Hinge at 66% from LE. This has been compared with the NACA 64A306 foil studied by Eppler, and has a lot of similarities despite the guesswork involved. The main difference is camber, as the LR-2369 has 2%, compared to the NACA 64A306 with 0.94%.

image.png.7f6b8f9457726919ac15ca6ae98f6906.png

Cavitation. Epplner notes that cavitation is a potential issue at both the LE (leading edge) and Flap Hinge due to low pressure spikes. Both the 64A309 and LR-2369 foils have a relatively small LE radius, which reduces the low pressure spike at the LE, and should therefore reduce the onset of cavitation there.

However, both foils in XFoil display a significant low pressure spike at the Flap Hinge when Flap > 0°, suggesting the importance of maintaining 0° Flap - or close to - at high speeds to avoid cavitation onset. This suction peak also encourages the separation of the boundary area, increasing drag plus hastening cavitation. In Eppler's study, circular arcs of 3% chord width were added to both sides of the foil around the Flap Hinge to mitigate these effects.

Pitch & Angle of Incidence. The AoI is the angle of the foil relative to the MWP. For NZ, the median pitch is around -2.0° from ACWS data, and an assumption could be made that at this pitch, the foil has reached an AoA of 0°, which in turn defines the AoI at +2.0°. By way of comparison, GB has a median pitch of -1.3°.

Leeway. Leeway increases the AoA of the foil. For example, at 22° of foil cant, 2° leeway = 0,7° increase in AoA, 5° leeway = 1.9°.

Data. The following table looks at comparisons of foil vs flap angles required to provide a fixed Lift Force (target) at various speeds, for NZ and GB with their respective foil areas. The columns in brown calculate the CL required for the target Lift, using speed and foil area. The values then derived are limited by the foil/flap angle sensitivity of 0.1° increments.

Set 1 calculates the approximate foil angle required at the speed range, with flaps set to zero.
Set 2 calculates the approximate flap angle required at the speed range, with foil set to zero.
Set 3 attempts to get slightly closer to reality, by setting maximum foil angle to the assumed Angle of Incidence as described above, for lower speeds. At higher speeds, the foil/flap angles were varied to attempt to derive the best L/D ratios.

The data does not consider the leeway induced increases in foil AoA, so these are higher than they would be in reality, especially at low speeds. The target lift does not represent actual conditions.

Comparison of Foil/Flap Angles for a Lift Force (FL) of: 80,000 Newtons
image.png.e31ee9e97746182441c11b9fed628d17.png

Observations/Notes.
Foil Area. Throughout the speed range, the L/D ratio is lower for GB as expected, except at the lowest speed of 20 knots.
Reynolds Number. Although used in XFoil calcs, Re had no quantifiable effect on CL and CD at different boat speeds.
Foil vs Flap. Using flap angle rather than foil angle generally appears to provide a superior L/D ratio, except for GB at higher speeds where they were roughly equivalent.  Using foil+flap angle appears to significantly benefit NZ at the lower speeds.
Foil Thickness. The foil tested was at a thickness of 9% of chord length. An attempt was made to try some profiles at lower thicknesses of 6% and 7%, to see how this might impact GB's foil performance. As surface area would not be reduced by much at all, drag would not be significantly reduced, so no major benefit was actually expected. However, the narrower profiles, including a comparison with the NACA 64-206 (6% thickness) failed to converge using XFoil except in a very limited AoA range, so was not included.

 

[Disclaimer]  This data has a significant degree of assumptions/inaccuracies, may contain errors, and may be missing required factors. I accept NO RESPONSIBILITY should SAAC use any data herein for the AC37 Challenge Yacht which results in a submarine.

  • Like 3
Link to post
Share on other sites
20 minutes ago, MaxHugen said:

 

Cavitation. Epplner notes that cavitation is a potential issue at both the LE (leading edge) and Flap Hinge due to low pressure spikes. Both the 64A309 and LR-2369 foils have a relatively small LE radius, which reduces the low pressure spike at the LE, and should therefore reduce the onset of cavitation there.

However, both foils in XFoil display a significant low pressure spike at the Flap Hinge when Flap > 0°, suggesting the importance of maintaining 0° Flap - or close to - at high speeds to avoid cavitation onset. This suction peak also encourages the separation of the boundary area, increasing drag plus hastening cavitation. In Eppler's study, circular arcs of 3% chord width were added to both sides of the foil around the Flap Hinge to mitigate these effects.

Very interesting, than you for your work. Two questions on the above section. Any views on what speed the cavitation would start to be significant?

When you say flap>0 will cause cavitation, presumably significant negative angles of flap would too?

Link to post
Share on other sites
7 minutes ago, enigmatically2 said:

Very interesting, than you for your work. Two questions on the above section. Any views on what speed the cavitation would start to be significant?

When you say flap>0 will cause cavitation, presumably significant negative angles of flap would too?

No, my 4-month old skills are way below being able to make a cavitation speed prediction!

Only slight negative angles seem to be required at the high speeds, and they may well balance that with foil angle. However XFoil shows a very significant reduction of the "flap hinge spike" when using some negative flap, so reducing the probability of cavitation.

Link to post
Share on other sites

Ah, interesting. So GB gets to zero flap around 10 kts before NZ. Thus if there is cavitation at 45 kts or so (which we know GB has reached) they will suffer less than NZ. 

And because GB can achieve the lift with lower flaps they may lift out at lower speeds, but between 30kts and that cavitation speed NZ has a better l/d

 

 

  • Like 1
Link to post
Share on other sites
8 minutes ago, enigmatically2 said:

Ah, interesting. So GB gets to zero flap around 10 kts before NZ. Thus if there is cavitation at 45 kts or so (which we know GB has reached) they will suffer less than NZ. 

And because GB can achieve the lift with lower flaps they may lift out at lower speeds, but between 30kts and that cavitation speed NZ has a better l/d

Given the very limited accuracy I have, it's difficult to make many conclusions. Note that critical angles at high speeds are actually very low, so the slightest adjustments could make a lot of difference re cavitation.  And of course, the foil profile too is just my guesswork...  :unsure:

Link to post
Share on other sites
6 hours ago, Basiliscus said:

Is there any interest in a series of technical posts that would illustrate the issues in foil wing section design by starting with the E908 section in this paper and modifying it to make it more suitable for the AC75?  It'd be too long for a single post, but I'm thinking that by giving each stage in the process its own post that it might be more digestible.

Great suggestion, I am looking forward to read them, I guess it will about rooftop, recovery and so on.

I remenber "looking" with my fingers, the foil section of EXploder 15 or 22, I noticed 2 concave portions:

The first one was between 60% to 70% of the leading edge on the low pressure side,

and the other one was a bit further aft on the high pressure side,  I candidly interpret that as a kind of "aft loading" as explained by Michael Selig on his workpaper for the high lift wing section S1223.

But I think the actual objective is very different than looking for high lift, instead I have the intuition the aft loading feature aims to create lift at the aft part in order to decrease the lift from the front part of the wing section.

I feel confident we will understand that very soon.

Thanks in advance

EK 

 

1 hour ago, MaxHugen said:

Both the 64A309 and LR-2369 foils have a relatively small LE radius, which reduces the low pressure spike at the LE, and should therefore reduce the onset of cavitation there.

I candidly thought, the smaller the LE radius, the higher the low pressure pike (-Cp)? 

Probably I missed something somewhere ?

Cheers

Link to post
Share on other sites
16 minutes ago, Erwankerauzen said:

I candidly thought, the smaller the LE radius, the higher the low pressure pike (-Cp)? 
 

Perhaps it's just these particular foils... or the angle... Not sure.  I only compared mine (top) to an equivalent NACA generated foil with the same thickness and camber:

image.png.111f539bc25c7445681ab762079499f9.png

Link to post
Share on other sites
9 hours ago, Basiliscus said:

Is there any interest in a series of technical posts that would illustrate the issues in foil wing section design by starting with the E908 section in this paper and modifying it to make it more suitable for the AC75?  It'd be too long for a single post, but I'm thinking that by giving each stage in the process its own post that it might be more digestible.

Of course!

Link to post
Share on other sites
5 hours ago, MaxHugen said:

I'm not sure what you mean by a "split".

I don't think we have any information to confirm that both flaps always move in unison. If there are advantages in modifying the vertical vs lateral forces or moments via differential flap settings, I think NZ would have added that capability.

Of course it's quite possible that it was decided that any gain was not worthwhile, and used a single actuator.  Can't really say one way or the other.

Plus I have no idea how large a "single body - twin actuator" would actually be?

They don't have that ability, because their foil wings are straight. Setting the flaps differently on a straight wing would not change vertical vs lateral forces. The lift vector would still be close to perpendicular to the wing. I don't think that a minimal change in the vector's angle and maybe a minimal shift in the center of effort would pay for the increase in drag. If they wanted to control vertical vs lateral forces with flaps, they'd have anhedral foils like the other teams.

Link to post
Share on other sites
2 hours ago, MaxHugen said:

No, my 4-month old skills are way below being able to make a cavitation speed prediction!

Only slight negative angles seem to be required at the high speeds, and they may well balance that with foil angle. However XFoil shows a very significant reduction of the "flap hinge spike" when using some negative flap, so reducing the probability of cavitation.

I think a lot depends on the hinge mechanism and how smooth you can make the upper surface of the foil/flap interface. Another thing to consider is that changing the pitch and yaw angle of the boat (with the rudder/rudder wing) has similar effects to deflecting the flaps since pitch and yaw determines the AOA of the foil. Of course rotating the boat takes much longer than deflecting the flaps, so they may be deflecting the flaps first to quickly react to changes in conditions, but then the flap goes back to the optimal angle for the highest speed and the boat's pitch and yaw angles are adjusted accordingly.

Link to post
Share on other sites
17 hours ago, Basiliscus said:

Is there any interest in a series of technical posts that would illustrate the issues in foil wing section design by starting with the E908 section in this paper and modifying it to make it more suitable for the AC75?  It'd be too long for a single post, but I'm thinking that by giving each stage in the process its own post that it might be more digestible.

OK, it sounds like there's some interest in this topic, so here goes.  

Any engineering effort starts by defining the requirements.  From this figure, it looks like the average foil area is 1.64 m^2, and Max Hugen has a lift force of 80 kN, so a good design loading would be 49 kN/m^2.  We need the onset of cavitation to be at least 45 kt, and it would be great to be able to get to 50 kt.  The foil should be as thick as possible for the cavitation speed, but I haven't worked out how much volume of steel is needed to meet the ballast requirement.  Let's see what we get for the thickness and that may help drive the decision as to whether to go with a bulb or not.  Let's target 18 kt as the takeoff speed for this first pass around the design spiral.

The plots of pressure along the chord show the upper surface in blue and the lower surface in red.  (The dashed line shows what the pressures would be if the effects of the boundary layer were not included.  The boundary layer displacement thickness is shown added to the shape of the section.)  Cp=0 means the local velocity is equal to the freestream velocity.  Cp=1 means the flow has been brought to a complete stop.  Negative Cp values mean the local velocity is higher than the freestream, and Cp becomes more negative with the square of the local velocity.  So long as the pressure remains greater (Cp more positive) than the vapor pressure of water, cavitation cannot occur.  So there is a threshold Cp value corresponding to the onset of cavitation.  The higher the speed the lower (less negative) this threshold will be.  The speed at which the minimum Cp anywhere on the section equals the threshold for vapor pressure is the incipient cavitation speed.  That's what I will be using for design.  The actual speed at which cavitation bubbles become visible is a few knots higher than that, so this approach is somewhat conservative.

I'm going to use the E908 section as a starting point.  Will it meet the requirements?  The thickness is only 9%.  That's not going to make the structures engineers happy, and it's going to require a big bulb.

The second plot shows the pressure distribution at half a degree angle of attack and zero degrees of flap deflection.  The pressures are fairly flat, and the minimum pressure is at mid chord.  This is where cavitation will begin at high speed.  But, as shown in the third plot, a modest increase in angle of attack forms a pressure peak on the upper side of the leading edge.  This will reach the cavitation threshold at a lower speed.  The fourth plot adds the pressures from 0.5 degree angle of attack to show the changes have mostly been at the leading edge.

The next two plots show what happens at negative angles of attack.  The pressure on the lower surface leading edge drops to being comparable to the upper surface, and then as the angle of attack is lowered further, a pressure peak forms on the underside of the leading edge.  This again lowers the speed for the onset of cavitation.

The E908 section was designed to  use a 20% chord flap, and the next series of plots shows the effect of flap deflection.  A +5.8 degree flap deflection, the design value for the E908, increases the lift on nearly the entire section.  A modest pressure peak has formed at the hinge line.  Deflecting the flap up results in the underside of the hinge line being on the verge of forming a pressure peak.  PLots 11 - 13 show what happens when the flap is deflected 10 degrees.  There are pressure peaks at both the hinge line and leading edge.  Plot 13 shows the pressures at zero flap deflection for comparison.

The next figure shows the drag polar plots for these flap deflections.  The plot on the left is drag vs lift.  In the middle plot, the lines sloping up to the right are lift vs angle of attack.  The nearly horizontal lines are the pitching moment curves.  On the right is the boundary layer transition location.  I've artificially tripped the flow near the leading edge so the boundary layer is fully turbulent.  I'll get into laminar flow later.

Since the E908 is too thin how about making it thicker?  I bumped the thickness up to 12%.  The velocities on both surfaces are increased, especially on the lower surface.  The upper surface would have increased even more, but there is separation at the trailing edge that reduces the lift.  The increased velocity on the upper side means the thicker section will have a lower cavitation onset speed.

The final chart shows the cavitation envelopes for all of these configurations.  If you know the speed and you know the lift coefficient, then you know what the load is per unit area.  The black lines show the loci of constant loading, and are universal.  Since the lift has to support the weight, and the side force is constrained by the righting moment from the foil, as a first approximation the lift can be considered to be constant as the speed changes.  The heavy red line shows the loading corresponding to the design requirement of 49 kN/m^2.  

For each of the section shapes, the speed at which the onset of cavitation corresponds to the minimum Cp as the angle of attack changes is plotted vs the lift at each angle of attack.  This forms the cavitation envelope for that section.  As long as the line corresponding to the loading is inside the envelope, cavitation cannot occur.  The flat right-hand side of the envelope is when the minimum cp is in the middle of the section, and it defines the high-speed limit.  The top side of the envelope shows how the pressure peak on the upper side of the leading edge leads to a rapid reduction in the cavitation speed as the angle of attack is increased.  The bottom side of the envelope is defined by the pressure peak forming on the underside of the leading edge.  

Where the load line intersects the envelope top side sets the minimum takeoff speed.  If the boat tries to take off lower than that speed, cavitation will occur at the leading edge, reducing the lift and increasing the drag.  So cavitation is not just a high-speed phenomenon. 

Even with 10 degrees of flap, the E908 has a takeoff speed that is more than 21 kt.  With zero flap it can get up to almost 42 kt, and deflecting the flap up gets it to 43 kt.  It doesn't meet the requirements for either takeoff or maximum speed.  Increasing the thickness to 12% (peach line) helps a little on takeoff, but makes the high speed limit even worse.  So the E908 is not going to work for the AC75.  

Next, I'll start to modify the E908 to make it more suitable.

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_01.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_02.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_03.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_04.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_05.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_06.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_07.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_08.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_09.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_10.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_11.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_12.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_13.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_14.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_15.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_16.png

plot_e908_t12_p8002rxx_r5e6n1fix25e-3_Page_17.png

E908_Cavitation.png

  • Like 14
Link to post
Share on other sites

^  Awesome, it will take me a while to digest this.  One request... would you mind adding each graphic following each step pls?

FWIW:
I did a rough calc of the area of my 9% thickness profile, at NZ's avg chord width of 0.32m. It worked out to 0.0054082m2.  Using a foil width of 4.0m - 0.4m (for NZ's wide blended bulb) = 3.6m, I get a Volume of approx 0.0195m3.

For stainless steel @ SG= 8 (8000kg/m3), this volume would weigh 156kg in steel, ignoring space required for controls. If the foil weighs ~900kg (excluding fairings), this leaves ~744kg for a bulb. If we estimate a steel "shell" for the bulb at say 44kg, that would require 700kg lead. I think the lower foil arm weight (in steel, as a shell) would also need to be calculated and deducted. I'll take a wild stab and say it weighs 50kg.

That leaves 650kg to be stored inside the bulb. For lead (11,340kg/m3) that would require a volume of about 0.0573m3.  Might try to estimate NZ's bulb volume, but could be too difficult.

Q1.  How do you calculate the pressure per m2?

Q2. What is the vapour pressure of water? I could only find it as 0.0231 atmos.

(I'm trying to understand how your cavitation graph is constructed) :unsure:

[edit] Forgot to mention, I've carefully watched take-of speed for GB and LR during a race, and if boat speed displayed was accurate, their skegs both left the water at ~18.8 knots.

Link to post
Share on other sites
1 hour ago, Tornado-Cat said:

Great to have Basiliscus contributing with others to this thread that is  most amazing for the knowledge shared. I don't pretend to understand everything though !

I'm intently studying this. I now get the first paragraph. :D

  • Like 2
Link to post
Share on other sites
2 hours ago, Kiwing said:

Doubt @Basiliscus would even claim that :D

I wonder if Guillaume Verdier, Nick Holroyd or Martin Fisher take a peek to the thread sometimes and laugh or find it interesting. Best would be to have them exchanging incognito here, even though I very much doubt it.

  • Like 1
Link to post
Share on other sites

Can you imagine the incredible, but credible looking bollocks that their teams disinformation specialists might spread???

  • Like 1
Link to post
Share on other sites
1 hour ago, MaxHugen said:

...Q1.  How do you calculate the pressure per m2?

Q2. What is the vapour pressure of water? I could only find it as 0.0231 atmos.

(I'm trying to understand how your cavitation graph is constructed) :unsure:

[edit] Forgot to mention, I've carefully watched take-of speed for GB and LR during a race, and if boat speed displayed was accurate, their skegs both left the water at ~18.8 knots.

1.  You had the foil lifting 80,000 N.  If the foil planform area is 1.6 m^2, that's 48,780 N/m^2 as the average loading.  The actual pressure at any location is P=Cp*0.5*density*V^2.

2.  Here are some values for vapor pressure: 
10C 1.3KPa  (0.18855psi)
15C 1.67KPa (0.24236psi)
 20C 2.4KPa  (0.34809psi)
25C 3.175KPa (0.461 psi)
ref 1kPa = 0.145038psi

The graph comes from L=CL*0.5*density*V^2*Area.  So  CL = (L/Area)/(0.5*density*V^2).  That defines the loading lines.

The envelopes come from Cp_min as function of CL.  The formula I used for cavitation speed is Vcav = 26.76/SQRT(-Cp_min).  There are better formulas that take into account the increased pressure at depth, but I figure there's always the possibility the foil will be operating near the surface.

Link to post
Share on other sites

So, my take as no sailor, not an engineer. Verdier top priority was to diminish the drag thus,

- T foil to reduce the total length of the foil, vs an anhedral, thus the drag accepting a more difficult foil to control

- Foil positioned as backward as possible for a bigger distance with the neutral point for better control and compensate the inherent difficulty of the T foil

- Very thin end of the foil arm to reduce the drag accepting a lower RM by positioning the weight in the arm fairing

Architects can feel free to contradict me.

 

Link to post
Share on other sites
5 minutes ago, Tornado-Cat said:

{snip}

Architects can feel free to contradict me.

 

I thought you said Anarchists there for a moment... Brace yourself!

Link to post
Share on other sites

Foil Section Design, Part 2

In our previous episode, the Eppler E908 section was found to be lacking with regard to the AC75 requirements.  It was too thin, it didn't go fast enough without cavitating, and it couldn't take off slowly enough without experiencing leading edge cavitation.  I started with the E908 and used Xfoil's inverse design capability to modify the pressure distribution and calculate the shape that would produce those pressures.

The first thing I did was to increase the design flap angle to 10 degrees.  Eppler's approach to handling the pressure peak at the flap hinge was to put the hinge behind the point where the pressure had already started to increase on the undeflected section.  That gave him some breathing room for the pressure peak.  My approach was more direct.  I deflected the flap, causing a big pressure peak.  Then, with the flap deflected, I chopped off the peak to give me a new shape.  When the flap was returned to neutral, this resulted in a depression in the pressure distribution and in the shape.  Then I deflected the flap upward, creating a pressure peak on the bottom at the hinge, and chopped that off, too.  And I did some flattening and smoothing of the rest of the pressure distribution.  Here's what I came up with for the pressures (diamonds indicate the E908 at the same angle of attack):

plot_H140_p8002rxx_r5e6n1fix25e-3_Page_05.thumb.png.c015f20d58f47e3207741c9c7acbebf5.png

The lift is about half that of the E908, but that's OK, because the E908 had way more lift available at high speed than the boat could use, given the design loading.  The minimum pressure is not as great as the E908, which means it has a higher cavitation onset speed.

Here's what the shape looks like compared to the E908:

plot_H140_p8002rxx_r5e6n1fix25e-3_Page_01.thumb.png.6403dec8dc8017d5932bad671183a244.png

The thickness is 12%, which is a big increase over the E908.  Pretty much all that thickness has been added to the bottom surface. The indentations at the flap hinge are evident.  When the flap is deflected, however, the critical surface becomes fair.

Here are what the pressures look like with +5 degree flap deflection, at the same lift coefficient as the basic E908:

plot_H140_p8002rxx_r5e6n1fix25e-3_Page_09.thumb.png.7381cc4a4a78a325beafa9a91fb6ab63.png

It's actually pretty comparable to the E908 with regard to the cavitation speed, despite being so much thicker.

At -5 degrees of flap, the lift is essentially zero:

plot_H140_p8002rxx_r5e6n1fix25e-3_Page_11.thumb.png.882321313fb1cb1ad8f4d591f3ab6285.png

The pressures are rather similar on both surfaces, and the shape is nearly symmetrical.  So the majority of the camber is due to flap deflection or the camber of the flap itself.

With +10 degrees of flap, there's tiny little pressure peak at the hinge line, but the pressure distribution is pretty flat without a significant peak a the leading edge.

plot_H140_p8002rxx_r5e6n1fix25e-3_Page_15.thumb.png.b94c13ec3275b722bef0bc809f121568.png

Here are the cavitation envelopes for the various flap deflections:

H140_Cavitation.thumb.png.3edbe0d7e77484816691bf7f4c1efbfc.png

The cavitation speed at the design loading is up to 43 kt, and it is right in the middle of the cavitation bucket, so the camber is about right.  This is a nice gain over the E908, despite the extra thickness, but it still needs improvement.  No gains on the takeoff speed, though, so that's an area that needs work. 

Here is what the drag polars look like:

plot_H140_p8002rxx_r5e6n1fix25e-3_Page_18.thumb.png.3b9a704279e30d61816c0ad309285c7f.png

The E908 has a little less profile drag, but that's to be expected because it is thinner.  When the boundary layer is fully turbulent, the shape does not have much influence and the drag depends on the thickness more than anything.  The trailing edge separation of the E908 has been eliminated, giving it less drag at high lift and a much higher maximum lift.  However, the maximum lift is not really achievable because the leading edge will cavitate first.

On the whole, though, the H140 is a definite step forward from the E908.  The thickness gain is huge and the high speed cavitation onset is getting there.  Improving the takeoff speed calls for some detailed work on the leading edge.

  • Like 8
Link to post
Share on other sites
6 hours ago, Basiliscus said:

1.  You had the foil lifting 80,000 N.  If the foil planform area is 1.6 m^2, that's 48,780 N/m^2 as the average loading.  The actual pressure at any location is P=Cp*0.5*density*V^2.

2.  Here are some values for vapor pressure: 
10C 1.3KPa  (0.18855psi)
15C 1.67KPa (0.24236psi)
 20C 2.4KPa  (0.34809psi)
25C 3.175KPa (0.461 psi)
ref 1kPa = 0.145038psi

The graph comes from L=CL*0.5*density*V^2*Area.  So  CL = (L/Area)/(0.5*density*V^2).  That defines the loading lines.

The envelopes come from Cp_min as function of CL.  The formula I used for cavitation speed is Vcav = 26.76/SQRT(-Cp_min).  There are better formulas that take into account the increased pressure at depth, but I figure there's always the possibility the foil will be operating near the surface.

1. What is Cp in the equation P=Cp*0.5*density*V^2, sounds like a Coefficient of Pressure, but if so, how is it defined?  And what units is P measured in, is it KPa?

2. What is Cp_min, and -Cp_min, what defines these vals, and what does the value 26.76 represent?

Sorry, 4 months of studying everything to do with the AC75 hasn't got me this far.  But I persevere.

Link to post
Share on other sites
10 hours ago, Basiliscus said:

OK, it sounds like there's some interest in this topic, so here goes.  

Any engineering effort starts by defining the requirements.  From this figure, it looks like the average foil area is 1.64 m^2, and Max Hugen has a lift force of 80 kN, so a good design loading would be 49 kN/m^2.  We need the onset of cavitation to be at least 45 kt, and it would be great to be able to get to 50 kt.  The foil should be as thick as possible for the cavitation speed, but I haven't worked out how much volume of steel is needed to meet the ballast requirement.  Let's see what we get for the thickness and that may help drive the decision as to whether to go with a bulb or not.  Let's target 18 kt as the takeoff speed for this first pass around the design spiral.

 

Thank you so much for this incredible work ! 

Link to post
Share on other sites
8 hours ago, Nutta said:

Can you imagine the incredible, but credible looking bollocks that their teams disinformation specialists might spread???

It would still be less incredible than some of the bollocks that is posted here anyway. Though to be fair much of that doesn't look very credible either

  • Like 1
Link to post
Share on other sites
On 2/10/2021 at 6:25 AM, Basiliscus said:

Actually something like the area rule is useful for minimizing cavitation and reducing junction drag.  Paul Bieker applied the principle on the AC72 rudders, and Boeing used it on the Jetfoil pods.  But I haven't seen a precise definition of the area rule for the purpose.  The transonic area rule is concerned with the farfield wave drag, so it includes all of the area in the cross section.  For a wing/strut junction, there should probably be some kind of radial weighting of the cross section area because the influence of the junction drops off with distance and the outer parts of the wing don't have a lot of influence on the junction.   

There's probably a PhD dissertation in there for anyone that wants to take it on.

Well bu@@er me I thought it was useful !

Link to post
Share on other sites

I think the two Italian enthusiasts are also reading this topic, as they've just issued a video on their views re foils and cavitation!

 

  • Like 1
Link to post
Share on other sites

^ FYI, the airfoil they discuss is the NACA 632-215. I don't know what the subscript "2" refers to, but this foil is described as having 15% thickness, at 34.9% from LE, with a camber of 1.1% at 50% from LE. However, the foil in their pic has been reduced to a thickness of ~11.2%.

For comparison my foil has 9% thickness at 36%, and 2% camber also at 36%. The NACA 64A309 (as adjusted per Eppler) that I also tested has 9% thickness at 32.5%, and 9% camber at 57%.

Interesting that the 2 NACA profiles both have their maximum camber well back, at 50% and 57%.

Foil ref: http://airfoiltools.com/search/index?MAirfoilSearchForm[textSearch]=63215&MAirfoilSearchForm[maxThickness]=&MAirfoilSearchForm[minThickness]=&MAirfoilSearchForm[maxCamber]=&MAirfoilSearchForm[minCamber]=&MAirfoilSearchForm[grp]=&MAirfoilSearchForm[sort]=1&yt0=Search

Link to post
Share on other sites
7 hours ago, MaxHugen said:

1. What is Cp in the equation P=Cp*0.5*density*V^2, sounds like a Coefficient of Pressure, but if so, how is it defined?  And what units is P measured in, is it KPa?

2. What is Cp_min, and -Cp_min, what defines these vals, and what does the value 26.76 represent?

Sorry, 4 months of studying everything to do with the AC75 hasn't got me this far.  But I persevere.

1.  Yes, coefficient of pressure.  You did ask how to calculate the pressure, right?  The local Cp distribution is calculated using Xfoil.  The equation is the definition of Cp.  V is the freestream velocity, not the local velocity.

2.  Cp_min is the minimum Cp anywhere on the section.  -Cp_min has the opposite sign, since the minimum Cp will always be negative.  In the plots of pressure distributions, Cp_min is the upper most point in the plot.  It can be located on either surface, at various points on the chord, depending on the shape and angle of attack.

The value of 26.76 is the result of combining several constants to create an answer with units of knots.

Link to post
Share on other sites
4 hours ago, MaxHugen said:

^ FYI, the airfoil they discuss is the NACA 632-215. I don't know what the subscript "2" refers to, but this foil is described as having 15% thickness, at 34.9% from LE, with a camber of 1.1% at 50% from LE. However, the foil in their pic has been reduced to a thickness of ~11.2%.

For comparison my foil has 9% thickness at 36%, and 2% camber also at 36%. The NACA 64A309 (as adjusted per Eppler) that I also tested has 9% thickness at 32.5%, and 9% camber at 57%.

Interesting that the 2 NACA profiles both have their maximum camber well back, at 50% and 57%.

Foil ref: http://airfoiltools.com/search/index?MAirfoilSearchForm[textSearch]=63215&MAirfoilSearchForm[maxThickness]=&MAirfoilSearchForm[minThickness]=&MAirfoilSearchForm[maxCamber]=&MAirfoilSearchForm[minCamber]=&MAirfoilSearchForm[grp]=&MAirfoilSearchForm[sort]=1&yt0=Search

The NACA 6-series airfoils are good starting points because their design pressure distribution is flat back to the location of the second digit in the designation (30% chord in the case of the NACA 632-215).  But the fact that the flat "rooftop" of the pressure distribution goes right to the leading edge makes them susceptible to forming a leading edge suction peak at high lift, and this leads to leading edge cavitation at low speed. 

A section with the rooftop further back would make an even better starting point, like a NACA 66- or 67- section.  The section would still need to be modified for the flap and at the leading edge.

The camber will be well back, making the section very aft loaded.  Trying to pack in as much thickness as possible with a high cavitation onset speed drives the main part of the section to be almost symmetrical, as you can see already with my example.

  • Like 3
Link to post
Share on other sites
7 hours ago, enigmatically2 said:

It would still be less incredible than some of the bollocks that is posted here anyway. Though to be fair much of that doesn't look very credible either

My cat is offended. He pays attention to stuff.

images?q=tbn:ANd9GcRFaIYlTUyFIS6epr9sUqH

  • Like 1
Link to post
Share on other sites
24 minutes ago, Kiwing said:

@Nutta that boat looks a nice place for a Barby and a beer will watching?      Lucky you?

Nah, I wish...

Link to post
Share on other sites
14 hours ago, Basiliscus said:

Foil Section Design, Part 2

 

Thank you! Very interesting to see how the refinement process would go.  My Xfoil adventures never got me into the inverse design features. In terms of design tools available out there, what do you think the teams are using? Do they develop their own tools or use something off the shelf? It amazes me that despite using the same formulas / laws of physics, the teams ended up with quite different foil sizes and shapes.

  • Like 1
Link to post
Share on other sites
9 hours ago, erdb said:

Thank you! Very interesting to see how the refinement process would go.  My Xfoil adventures never got me into the inverse design features. In terms of design tools available out there, what do you think the teams are using? Do they develop their own tools or use something off the shelf? It amazes me that despite using the same formulas / laws of physics, the teams ended up with quite different foil sizes and shapes.

My guess is Xfoil is used a lot.  These foils are operating a decent Reynolds numbers and are subject to cavitation before they get to high enough angles of attack that they'd stall from flow separation.  So the potential flow theory + integral boundary layer method that Xfoil uses is fairly accurate until cavitation begins.  A RANS code will tell you what a shape will do, but it doesn't tell you what the shape should be.  That's the beauty of Xfoil's inverse design capabilities, and that of the Eppler code that preceded it.  

If the design uses a slotted flap, then MSES is needed to design and analyze the shape.

There are optimizing codes that are wrapped around Xfoil and can iterate a design automatically.  

The 3D effects are important, and Xfoil is strictly 2D.  You can use a panel code to calculate the 3D pressures and see what areas will be cavitation free.  Like Xfoil, a panel code won't tell you what happens after cavitation occurs.  But since the goal is to avoid cavitation, a panel code is still very useful.  You can run Xfoil and a panel code on a laptop and iterate the design very quickly.

Once cavitation occurs, only a RANS code can calculate what the drag rise will be.  

One of Brittain's aviation pioneers, Lord Brabazon, said, "Compared to designing yachts, designing airplanes is child's play."  There's a reason they call it "the art of design."

  • Like 3
Link to post
Share on other sites

Previously, the H140 section was designed to overcome the deficiencies of the Eppler E908 section with regard to the AC75 design requirements.  The H140 made a significant gain in thickness and the high-speed cavitation onset was promising, but it was subject to leading edge cavitation at takeoff speed.  The next iteration, H141, was aimed at addressing takeoff speed.

Here is how negative the minimum pressure coefficient can be to avoid cavitation at each speed.  The minimum pressure can be no more than -2.21 in order to take off at 18 kt.

CavitationThreshold.thumb.png.11e077660f3f41be8908db3d7241c059.png

I first tried to address the takeoff by just modifying the leading edge pressure distribution.  However, this added thickness that ruined the high speed cavitation.  And there was flow separation at the trailing edge of the flap.  The 20% chord flap was just too short.  So I enlarged the flap to 30% chord.  That, plus the leading edge shaping I'd already done, solved the takeoff speed problem.  The high speed cavitation onset occurred at too low a lift coefficient, so I added camber to the flap to get more aft loading.  I couldn't stand to add more lift to the main part of the section without running afoul of high speed cavitation.

Here is what the new shape looks like:

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_1.thumb.png.02275d6728c44cb0507514e6be99e345.png

The nose is a little fuller, which reduces the leading edge suction peak at takeoff.  The depression at the flap hinge is moved forward to 70% chord.  The flap has a bit more hooked shape to give the aft loading.

Here is what the new pressure distribution looks like:

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_2.thumb.png.dc7049dc1f517bf84da972460879ac41.png

And here is the effect of deflecting the flap:

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_4.thumb.png.7bfd75f5a4641742bed3a653ef3bef34.png

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_6.thumb.png.fb1b7073f8e580b1fa9659bcf4db1f80.png

Two degrees of angle of attack gets the 10 degree flap case to takeoff lift:

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_7.thumb.png.08d162e6e7ff38b0b649c8d5ab407d01.png

There's a leading edge suction peak, but it doesn't get to the limit for takeoff.

And, for the sake of completeness, here are the drag polars (assuming fully turbulent boundary layers):

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_8.thumb.png.eb542b3e90244c7dbd586481c5e4a798.png

The cavitation plot shows takeoff is feasible at 18 kt with 10 degrees of flap:

H141_Cavitation.thumb.png.41ede83147b9c9608f1c90468605bc9b.png

The high speed cavitation onset is 40 kt.  That's not going to be competitive.  So the next goal is to extend the high speed cavitation onset without giving up on takeoff and while preserving as much thickness as possible.

 

  • Like 5
Link to post
Share on other sites
2 hours ago, Basiliscus said:

The high speed cavitation onset is 40 kt.  That's not going to be competitive.  So the next goal is to extend the high speed cavitation onset without giving up on takeoff and while preserving as much thickness as possible.

Interesting. Whilst I know that you can change things to improve that, it does suggest that there is a compromise between drag and cavitation and that therefore the boats are likely to be hitting or on the edge of hitting cavitation speeds at times, otherwise they would have given up too much for the compromise

Link to post
Share on other sites
2 hours ago, Basiliscus said:

I first tried to address the takeoff by just modifying the leading edge pressure distribution.  However, this added thickness that ruined the high speed cavitation.  And there was flow separation at the trailing edge of the flap.  The 20% chord flap was just too short.  So I enlarged the flap to 30% chord.  That, plus the leading edge shaping I'd already done, solved the takeoff speed problem.  The high speed cavitation onset occurred at too low a lift coefficient, so I added camber to the flap to get more aft loading.  I couldn't stand to add more lift to the main part of the section without running afoul of high speed cavitation.

The cavitation plot shows takeoff is feasible at 18 kt with 10 degrees of flap:

FWIW, NZ's flap is ~40% of chord, and LR ~36%.

Having watched a couple of good videos of NZ's take-off, they are very level, no doubt using the canoe shapes bustle for minimum drag as they build up speed. After take-off, they maintain level flight for a while as speed rapidly increases, before starting to ease the bow down. Perhaps they are easing the flap angle somewhat during that speed-up stage.

For take-off I think they are using the 2° angle of incidence, plus around 12° of flap.

Link to post
Share on other sites
9 hours ago, Basiliscus said:

The high speed cavitation onset is 40 kt.  That's not going to be competitive.  So the next goal is to extend the high speed cavitation onset without giving up on takeoff and while preserving as much thickness as possible.

Really interesting to see the compromise between take-off speed and top speed. If I understand it correctly, the smaller the foil area is, the harder it gets to find a shape that works well both at take-off and max speed. It also emphasizes the importance of the keel to help lift the boat on foils. This is probably why AM had the most trouble in light winds (once INEOS fixed their problem).

So how bad is it to have cavitation? Let's say the Cp peak at the flap hinge just goes over the limit. Will that immediately screw up the flow completely downstream of that? Is it an on-off kind of thing or there maybe a couple of kts speed range where things get gradually worse.

Is it better to have cavitation at the flap hinge vs at the leading edge? I'd think having cavitation at the leading edge would be more serious affecting the flow over the whole foil.

 

Thanks again for posting all this. It's really eye-opening.

  • Like 1
Link to post
Share on other sites
14 minutes ago, erdb said:

 

So how bad is it to have cavitation? Let's say the Cp peak at the flap hinge just goes over the limit. Will that immediately screw up the flow completely downstream of that? Is it an on-off kind of thing or there maybe a couple of kts speed range where things get gradually worse.

Is it better to have cavitation at the flap hinge vs at the leading edge? I'd think having cavitation at the leading edge would be more serious affecting the flow over the whole foil.

It would be gradual, but you have to include consideration of those peak speeds as they bear away at the top of the mark when they are much faster. If cavitation causes the lift to disappear at those high speeds then it could go wrong fairly quickly. That is one explanation for some of AMs bad bear-aways. At other time it could just be slower.

I think cavitation on the leading edge would slow you more, but if you get cavitation across the flap then control could disappear very quickly

Link to post
Share on other sites
10 minutes ago, enigmatically2 said:

It would be gradual, but you have to include consideration of those peak speeds as they bear away at the top of the mark when they are much faster. If cavitation causes the lift to disappear at those high speeds then it could go wrong fairly quickly. That is one explanation for some of AMs bad bear-aways. At other time it could just be slower.

I think cavitation on the leading edge would slow you more, but if you get cavitation across the flap then control could disappear very quickly

I've watched a couple of videos of cavitation in tank testing, looks like it gets triggered very suddenly at one or more spots and in less than a second spreads over the very low pressure area. Might behave differently depending on the foil profile though.

Link to post
Share on other sites
4 minutes ago, MaxHugen said:

I've watched a couple of videos of cavitation in tank testing, looks like it gets triggered very suddenly at one or more spots and in less than a second spreads over the very low pressure area. Might behave differently depending on the foil profile though.

Sorry, I should have prefixed that first sentence with "I think". Evidently I was wrong

If it starts further back does it spread forward though?

Link to post
Share on other sites
9 minutes ago, enigmatically2 said:

Sorry, I should have prefixed that first sentence with "I think". Evidently I was wrong

If it starts further back does it spread forward though?

I tried unsuccessfully to find these videos again (saw them years ago). One was of a centreboard profile, supported vertically under an endplate representing a yacht hull. Cavitation started at the root of the board, and spread maybe 2/3 of the way down almost instantly.

I don't think a cav zone starting at the flap hinge area would extend forward.

Anyway, what I did find though is a video showing a T foil, and this does exhibit a gradual spread of the cavitation zone from the leading edge. It appears that the foil profile they are using also has a maximum camber a fair way back.

 

Link to post
Share on other sites
3 hours ago, MaxHugen said:

I tried unsuccessfully to find these videos again (saw them years ago). One was of a centreboard profile, supported vertically under an endplate representing a yacht hull. Cavitation started at the root of the board, and spread maybe 2/3 of the way down almost instantly.

I don't think a cav zone starting at the flap hinge area would extend forward.

Anyway, what I did find though is a video showing a T foil, and this does exhibit a gradual spread of the cavitation zone from the leading edge. It appears that the foil profile they are using also has a maximum camber a fair way back.

 

That's a cool video, thanks. The first example with the centerboard may have been ventilation instead of cavitation or maybe a combination of both, no?

Link to post
Share on other sites
11 minutes ago, erdb said:

That's a cool video, thanks. The first example with the centerboard may have been ventilation instead of cavitation or maybe a combination of both, no?

 It was definitely cavitation, I searched following some discussion after AC35, the test tank equipment had no air in it, so that would rule out ventilation. The centreboard had a bit of sweep on it, and was a bit wider at the root, so I don't know if that would make a difference. The vid was already some years old when I saw it, so possibly posted around a decade ago.

 

 

Link to post
Share on other sites

In my last post, I made a cut-and-paste error in the cavitation plot.  Here's the corrected version:

H141_Cavitation.thumb.png.d3a187c4fb94fdeda24605da69417bd5.png

The state of play in that post was the design had the 12% thickness and could take off at 18 kt, but only had a 40 kt top speed.  In order to get to 46 kt, the minimum Cp could be no more than -0.34.

CavitationThreshold.thumb.png.247b8cf156e32389700403c0b6dee41b.png

 

The plan for extending the high-speed cavitation onset was to use aft loading to shift the lift from the main part of the foil to the flap region.  That would bring the upper and lower surface pressure distributions together.

plot_H141_p7001rxx_r5e6n1fix25e-3_Page_2.thumb.png.27e9d9257e75b646ab361c44e11019d4.png

(Sorry for the red annotations.  Yellow just didn't stand out on a white background.)

Here's where that approach led:

plot_H142_p7003rxx_r536n1fix25e-3_Page_1.thumb.png.de5a14530b6a29e56333730dbb74aca5.png

plot_H142_p7003rxx_r536n1fix25e-3_Page_2.thumb.png.03f065b5c5afd0b214e395c74b4892c3.png

There's very little load carried by the main part of the foil.  The pressures are pretty much all due to thickness.  The flap got so thin that I added a blunt trailing edge just to give it some meat.  I expect there would be a lot of anguish on the part of the structures and controls engineers because at high speed, most of the weight of the boat is sitting on the flap.  The hinge moments are going to be horrendous.  But it's still 12% thick.

The high lift case looks OK - the min Cp is greater than -2.0:

plot_H142_p7003rxx_r536n1fix25e-3_Page_3.thumb.png.970ece5fa6102092181b91e3644434c0.png

There's nothing remarkable in the drag polars:

plot_H142_p7003rxx_r536n1fix25e-3_Page_4.thumb.png.9e541f90d4c5e8d0c6a870136e5d7dfd.png

The cavitation envelopes show I've almost met the requirements:

H142_Cavitation.thumb.png.2771498e3f2654e1e76b1a4f628f26ea.png

Takeoff cavitation looks good, but I'm just shy of 45 kt at the high speed end.  

So far, the effort has been fairly straight-forward.  It could even be done using Xfoil by some guy on Sailing Anarchy, ffs.  But now the team needs to make some decisions.

The first thing that can be done is to sharpen the pencil, and do some analysis using RANS to see just what the behavior of the section will be past the incipient cavitation speed.  I'm pretty sure there wouldn't be any visible cavitation at 45 kt, but cavitation will have an effect between there and 50 kt.  Just how bad it would be is going to take some sophisticated CFD.  This section might be the kind of thing a team would use for its first foil to get some experimental data on how bad the cavitation would be and what kind of tradeoffs would be needed for the next set.

The sailors are going to need to weigh in on how important the 45+ speed range is, given the forecast wind conditions for the Match.  If higher speeds will be needed to be competitive, then the thickness ratio needs to come down.

Reducing the thickness ratio can be done in two ways.  One way is to keep the physical thickness and make the foil wider, increasing the area.  This does two things.  It lowers the thickness ratio, which will allow the rooftops in the pressure distributions to come down.  It will also reduce the foil loading.  That will make the takeoff easier and may help the boat stay foilborne in marginal conditions. 

But the added wetted area will add drag.  At the high speed end, this means trading some additional drag at speeds leading up to 45 kt, but delaying the drag rise due to cavitation to higher speeds.  Extra drag can hurt the takeoff performance, too.  Whether or not this is a winning move will have to be examined using the VPP and the race model program.

The other way to reduce the thickness ratio is to make the foil physically thinner.  This means the structural problems become more severe and there's less room for ballast.  It could force the decision as to whether or not to have a bulb.  

And that is one reason why we're seeing different approaches to the foil design among the teams.

  • Like 12
Link to post
Share on other sites
Quote

  I expect there would be a lot of anguish on the part of the structures and controls engineers because at high speed, most of the weight of the boat is sitting on the flap.  The hinge moments are going to be horrendous.  But it's still 12% thick.

Mind boggling

Link to post
Share on other sites
9 hours ago, Basiliscus said:

One way is to keep the physical thickness and make the foil wider, increasing the area. 

In light of your analysis, the ETNZ package seems even more remarkable. Their foils are a lot smaller than those on LR and INEOS, yet they don't seem to have a problem getting up in light air. Their flap to foil area ratio is much higher, so I assume the load on their flaps is through the roof, but apparently they operate both flaps with one actuator, while the others use two.

As for max speed, I wonder how sail design comes into this. The goal is not to break the speed record, but to have higher VMG than the others. So they may be OK with a lower max speed foil as long as they can point closer to the mark while limiting their speed. That depends on L/D ratios of foils and sails. If their sails are more efficient, they can design foils that match that - lower speed, but better angles for higher VMG. Just have to figure out how to survive the mark roundings somehow...

  • Like 3
Link to post
Share on other sites

Data from yesterday's races - no big surprise, but there are some interesting things. I've excluded the first few minutes from Race-1 when INEOS was off the foils.

As was obvious during broadcast, the big difference was upwind. In Race-1, LR was not only faster, but pointed higher for a much better VMG:

152662026_uwbs.png.99ad8972721e2c1cde35e83472561803.png1271877634_uwtwa.png.39a753ecfbae174abc525c2a151150b3.png685854872_uwvmg.png.ec959baf0ff29f8b3367d588e022784a.png

In Race-2, upwind TWAs were similar, but LR was still faster:

628684359_uwbs.png.5b35c6dd28e962bb5f89847d15763d1d.png574075072_uwtwa.png.842ab2a5d656f097ce3883b52f59db6d.png1667023161_uwvmg.png.323b9114010887c9fcc77d87e01e73ac.png

Downwind VMGs were almost identical in both races, but there is a fascinating difference in leeway. LR seems to have two modes both upwind and downwind - a negative and a positive leeway mode - whereas INEOS has only one upwind and one downwind mode with leeway mostly between 0 and +2 deg:

Upwind (Race-1 and Race-2):

1790713258_uwleew.png.75be766b0d3b34ed561cfbff4bcc8447.png624587874_uwleew.png.bbe8fffa6f6592b15022ea00a791c20d.png

Downwind (Race-1 and Race-2):

1871555439_dwleew.png.19f3b0c51548d677941a8536b00dd440.png92698995_dwleew.png.5d633ca4bd75d2478a2d2ea47c200ed8.png

As a reminder leeway is used here as the difference in course sailed vs direction of the boat's centerline. Positive leeway means the bow points upwind from the direction of movement, negative leeway is when the bow points downwind of the direction of movement  - boat seems to be crabbing windward. Current may interfere with this calculation, but I'd expect mostly the same effect on both boats. I don't think LR's double peak in leeway is because of that.

Pitch also changes differently for the two boats. It seems LR pitches forward more as speed increases. upwind pitch in Race-1 was similar, but Race-1 downwind and Race-2 both up/downwind LR pitched forward more.

Race-1:

1984563852_uwpitch.png.5e5f37dea69e62b044610bc015901cff.png701854239_dwpitch.png.90f357d8c9b8830cc527df2aebf631d5.png

Race-2:

84666747_uwpitch.png.45d8a7aa5309027653c5ea675bf4e2f0.png1992046360_dwpitch.png.6978d89f842dd2629b31a7e73ca7f74a.png

Some other differences: INEOS also sails about 0.5m higher than LR. I don't know if this is a calibration issue or it's because of pitch angle (I assume the height sensor is on the bow). INEOS also seems to cant the foils out more both upwind and downwind, and there are some minor differences in how they change the cant when tacking or jibing. If anyone interested, I can post these graphs, too.

Maybe a bit concerning for INEOS, they actually used the shifts a little better than LR in both Race-1 and Race-2 and still lost.

favtack.png.d6df18ba2b72e0b83216790197da59d7.png    favtack.png.de7246b76e9c95c52ac3f44e66b6f2b3.png

  • Like 14
Link to post
Share on other sites