Sailrocket 2 set to launch

[Doug Lord]

Sail Rocket-amazing stuff:

Just specs? Well no:
1) Moth=66lb + 175 lb crew; SA= 86 sq.ft.. So, Weight(241lb.)/ SA(86sq.ft.)=2.8lb per sq.ft. SA. Sail Loading

2) Hydroptere=15000lb+875crew; SA=6034 sq.ft. So Weight(15875lb)/ 6034sq.ft.=2.63 lb.per sq.ft. Sail Loading

3) TNZ 72- Weight(13940lb.)/ SA(6297 sq.ft)= 2.21lb. per sq.ft. Sail Loading
However, this is downwind SA on the big boats so when I get more accurate info on the jib size for the AC boats I'll take another look.
===============
Now for the really interesting part:
4) Sail Rocket 780lb with one crew/ 236.7 sq.ft.(actual)= 3.3lb per sq.ft Sail Loading.

 

[cavalier mk2]

What are you calling the horizontal wing/strut area?

 

[redreuben]

I've been a way for a few days, I'm just gobsmacked at over 70 knots !
How much more time do they have there ?

 

[Doug Lord]

Sail Rocket-the record run

Here it is-the fastest sailboat on the planet setting that record:
(Note: to see this full screen click on the url above the embedded player, then click on the "full screen" icon on the youtube player. Some of the aerial footage was apparently taken by an rc aircraft with a go-pro camera)
-----
The story that goes with the video: http://sailrocket.com/node/672

http://www.youtube.com/watch?v=pipGWQmerEQ&feature=youtu.be

 

[Doug Lord]

Sail Rocket

The camera plane:

BERNT WITH HIS GOPRO PLANE. THE FOOTAGE IS AMAZING. THIS GUY HAS SKILLS. BLOODY WELL DONE IN 30 KNOTS OF WIND."
http://www.sailrocket.com/node/663

 

[Doug Lord]

Sail Rocket !

Final:

The WSSR Council announces the the establishment of a new World Record.
Record: Outright World and World “B” Division Sailing Speed Record
Venue: Walvis Bay. Namibia.
Name: Paul Larsen. AUS.
Equipment: Vestas SailRocket 2. Inclined rig Hydrofoil Proa.
Date: 24th November 2012. @ 13.32 hrs
Course length: 500 metres
Current: Nil
Elapsed time: 14.85 secs
Speed: 65.45 kts

Comments: Previous Outright and “B” Division Record: 2012. Vestas SailRocket 2. Paul Larsen AUS. Walvis Bay NAM. 59.37 kts

John Reed
Secretary to the WSSR Council

65.45kts (75.27mph)

 

[tspeer]

Sailrocket uses a base ventilated section for its main foil. I don't know what the actual section looks like, but I've made a quick stab a some section designs to see what the issues are.

A base ventilated section is wetted on both sides, but has a thick flat base that is dry. The cavity behind the base of Sailrocket's foil is filled with air, because atmospheric pressure is higher than the vapor pressure of water. So the air doesn't suck back on the base as hard as would be the case if the water were allowed to cavitate and fill the cavity with water vapor. This significantly reduces the drag, because the base drag is a big contributor to the profile drag.

The design philosophy was to have a near constant pressure on both sides, with a gradual acceleration to the trailing edge on the suction side. This minimized the maximum velocity, giving the highest speed without cavitation, and ensured that cavitation would begin at the trailing edge when it did happen. I was aiming at an incipient cavitation speed of 65 kt, since that's what was reported to be the Sailrocket design speed. I did't spend too much time iterating the designs - I'm sure Malcom Barnsley has spent a whole lot more time to get even small improvements!

It quickly became apparent that it would require a very thin foil to avoid cavitation at these speeds, even with a base ventilated foil. A thickness of 7% will almost get you to 60 kt, but you need to go with 5% thickness to get to 65 kt. I've called my two candidate shapes BaseVent03, which has 7.3% thickness, and BaseVent04, which has 5.1% thickness. The two shapes are shown in the first attached figure.

The next two figures show the XFOIL-predicted pressure distribution for BaseVent03 at a lift coefficient of o.25 and BaseVent04 at a lift coefficient of 0.15. These sections are arguably beyond XFOIL's ability to accurately predict the lift and drag, because the large separated base cavity exceeds the assumptions behind XFOIL's potential flow+integral boundary layer approach. However, I think the results are good enough to get a qualitative idea of the engineering issues of this kind of foil.

The first issue is they have a very narrow range of cavitation-resistant operation. The last figure shows the cavitation diagrams for these two sections, compared to the Epper E817 section scaled to 7% and 5% thickness. The Eppler E817 is typical of a high-speed subcavitating hydrofoil section. The base ventilated foils have an incipient cavitation speed that is maybe 5 kt higher than the subcavitating section, but are only better over a narrow range of foil loadings.

The narrow operating range means it's likely that the Sailrocket foil will cavitate at lower speeds if it operates at a higher lift coefficient. This is due to a leading edge suction peak that forms when operated outside the design conditions. And there have been some reports of cavitation making it hard to accelerate to top speeds. Because the loading on the foil is driven by the load on the wing and not the weight of the boat, the foil loading is variable instead of being constant like a vertical-lifting flying foil would be. So it may be possible to approach high speeds at a more or less constant lift coefficient, with both the wing lift and the foil loading going up with the square of the speed as the boat accelerates.

The boundary layer displacement effects tend to increase the velocity ahead of the trailing edge. This would be consistent with the fine pitting Paul reported seeing on the foil after some high speed runs. The bubbles wouldn't have time to grow very much before they were swept off the surface, so it's quite possible that the foil could be operated well beyond the incipient cavitation speeds calculated here.

The key thing would be to keep the lift coefficient near 0.15 so as to stay in the middle of the cavitation bucket and not form a leading edge suction peak. The loading on the foil would be on the order of 1400 lb/sq ft when at speed. The foil needs to be closely matched in size to the wing in order to hit the target loading at the target speeds.

Again, I've no idea if these shapes are what Sailrocket is using. However, I think the design constraints would drive them to something similar.

 

[cavalier mk2]

There are certainly a lot of conflicting possibilities for advancement in the speed stakes. How accurate are these programs in evaluating different prototypes when they are exceeding their parameters?

 

[markdrela]

Sailrocket uses a base ventilated section for its main foil. I don't know what the actual section looks like, but I've made a quick stab a some section designs to see what the issues are.

That's a pretty cool idea. I think you can do better by cambering the aft rear surface, sort of like a smooth Gurney flap, which loads up the bottom surface with no adverse effects on the top. See PDF. This has Cpmin=-0.2 which will go to 60kt before cavitation. But adding 30 deg sweep will increase it to 70 kt. Or it could be thinned and decambered as usual.

Re: Ventilation bubble modeling.
You can reduce the turbulent wake dissipation via the Kdl parameter in the VPAR menu, LAG command. Increasing Kdl from the default 0.9 value will make the wake recirculation bubble mix out more slowly, which I think better mimics a ventilation bubble. Second PDF corresponds to Kdl=4.0 which gives a much longer bubble.

 

[tspeer]

That's a pretty cool idea. I think you can do better by cambering the aft rear surface, sort of like a smooth Gurney flap, which loads up the bottom surface with no adverse effects on the top. See PDF. This has Cpmin=-0.2 which will go to 60kt before cavitation. But adding 30 deg sweep will increase it to 70 kt. Or it could be thinned and decambered as usual.

That's a good idea. I didn't play much with the bottom camber because I was trying to get as much thickness as I could for the speed. There's a big tradeoff between lift and thickness. The top pressures are constrained by cavitation, so the top surface shape is essentially fixed, too. The bottom surface can be carved away to generate lift, but only at the expense of thickness. So the more foil loading you aim for, the less structure there is to support it!

Re: Ventilation bubble modeling.
You can reduce the turbulent wake dissipation via the Kdl parameter in the VPAR menu, LAG command. Increasing Kdl from the default 0.9 value will make the wake recirculation bubble mix out more slowly, which I think better mimics a ventilation bubble. Second PDF corresponds to Kdl=4.0 which gives a much longer bubble.

Ah, the infamous LAG command! Actually, I've not heard of it. Probably because I typically run version 6.94 to avoid the CINC crashes of 6.96 in Windows, and there's no LAG command in 6.94. I see the LAG command is in 6.96 and 6.97. I guess I should start doing more of my XFOIL work in Linux instead of Windows.

Just so I understand the right values, the default value in V6.97 is Klag=5.6, and it produces a short bubble. When I change it to Klag=0.9, it produces a longer bubble. That's the opposite behavior to what you are showing with V6.98. I've attached a pdf of the three cases (Klag=5.6, 4.0, 0.9; transition is fixed at 5% chord).

With the long bubble, the section L/D is almost doubled . Obviously a base ventilated foil is going to be very sensitive to the base pressure. Increasing the base pressure by ventilation is bound to extend the cavity compared to having it filled with water vapor. And that cavity is probably going to be larger than the liquid wake behind the foil at subcavitating speeds. The value of Kdl does change the base pressure, with the shorter cavity having the lower pressure.

Do you think that using the base pressure is a good guide for adjusting Kdl to suit? I would expect that the base pressure is known ahead of time based on atmospheric pressure for the ventilated case and vapor pressure for the cavitating case. So at a given speed, there would be a target Cp for initial wake pressure.

 

[tspeer]

There are certainly a lot of conflicting possibilities for advancement in the speed stakes. How accurate are these programs in evaluating different prototypes when they are exceeding their parameters?

That's a question for Prof. Drela...
(I've wondered myself if there are any validation data for XFOIL with sections that have thick bases.)

 

[cavalier mk2]

The more I learn about these things the more I realize the programs can only consider what they've been directed to. I would think there is a difference in the cavitation between fresh and salt water as one example, or is there an option? Can we get the prof's source code.....

 

[tspeer]

...Can we get the prof's source code.....

XFOIL source code is available from the XFOIL web site.

 

[cavalier mk2]

Thank you!

 

[markdrela]

Just so I understand the right values, the default value in V6.97 is Klag=5.6, and it produces a short bubble. When I change it to Klag=0.9, it produces a longer bubble. That's the opposite behavior to what you are showing with V6.98. I've attached a pdf of the three cases (Klag=5.6, 4.0, 0.9; transition is fixed at 5% chord).

Klag and Kdl are two different parameters.

Klag = 5.6 always.
Kdl = 0.9 for a normal wake. Make a lot bigger for a non-dissipative wake, which is closest to a base ventilation bubble.