Deck sweeping sails and effective aspect ratio

More washout should also allow increased loading on the sail foot. The Cl peak near 0.75 spanwise distance would reduce and go down the mast.

The result would be higher possible average Cl and because of unloading the top of the sail there would be less heeling.

 

I have tried 17, 22 and 27deg washout. There is indeed a little more total lift for all three cases, but induced drag goes up accordingly and speed actually suffers. Very little though.
Lift coefficient peaks at around 45% of the luff height for 27deg twist. For all three cases, heeling limit kicks in and slightly reduces the maximum section Cl.

 

When talking about doubling the sail aspect ratio through the use of sea surface as a mirror plane, the condition to have the AR doubled (or nearly doubled) is that the resulting lift distribution is a symmetrical and smooth curve across the mirror plane, with no significant kinks in the curve when crossing the hull area.

Now, take a look at the two pictures in the post #60 (http://www.boatdesign.net/forums/hy...ffective-aspect-ratio-52901-4.html#post731469). They are qualitatively correct, the flow will indeed be so turbulent and swirling around the hull and over the deck, and it will produce a gross discontinuity of the lift distribution in the area between the boom and the sea surface. Hence the mirror-image will not yield a double-AR sail, regardless of the absence of the gap. This sketch explains why:



My conclusion is that the flow disturbance and the vorticity created by the hull (and the consequent discontinuity of the lift curve) will by far offset any possible gain obtained by closing the sail gap in the cockpit area.

For the same reason it is not possible to estimate the effect of the presence of the hull through the use of the linear lifting-line theory. Hence, it is not possible to use the Vortex2K spreadsheet for this kind of analysis. You will have to either use a 3-D VLM, or a 3-D CFD. The linear lifting-line theory will just give you a set of wrong numbers in this condition.

Cheers

 

In addition to your twist changes creating an unrealistic jump in sail force there is also the issue of sail stall (detached flow) occurring in the very large turbulence caused by the hull. This detached flow is common even in streamlined aircraft fuselage-wing junctures. By loading the foot of the sail with twist changes you would make this stalled area huge on a real boat. Stall has the nearly the same effect as opening the gap up again. Tom's Vortex spreadsheet does not model this. It is based on the assumption of attached flow everywhere. The following quote is from the first column of the first page of Tom's spreadsheet.

"Lifting line theory is a linear theory. Nonlinear effects, such as stall, are not modeled. This limits its applicablility to low to moderate angles of attack, say 10 degrees or less."

The fact that you are using Cl max of 1.4 indicates angle of attack values far more than 10 degrees from the zero lift line.

Do you now see how you are making unrealistic inputs and drawing bogus conclusions?

 

Even without the disturbance from the hull the water/land/ice can't mirror the planform according to double AR theory, since at the water/land/ice true wind is zero. There is just apparent wind and no driving force can be made due to zero AoA.

 

That is correct, thanks for this note.

 

Actually you have to specify the computational tool even a bit more than that. VLM (vortex lattice method) is also a linear tool that assumes attached flow. CFD (computational fluid dynamics) only indicates that you are doing fluid dynamics calculations with a computer.

The missing specification is "modeling of detached flow" or "flow separation modeling." Examples of this are LES and RANS.

 

The flow over the deck and the deckhouse is not necessarily detached. It presents swirls due to sharp edges (like gunwhale and cabin edges), but that condition is not incompatible with the VLM, especially if the angles of the flow to the deck are not excessively high. Think of the aerodynamics of very low-AR or delta wings, for example, which present similar vortex structures but can be analyzed with the VLM. Of course, if one wants to get into flow details and fine-tuning of the cases, the VLM can also become inadequate when strong 3-D turbulence effects are present.

Ok, but that is bordering the academic pedantry, in this case. The main point of my mentioning the CFD is - the LLT is definitely a wrong tool for this task.

Cheers

 

Once again an unfounded assumption. How many of your claims do you actually check/calculate/simulate before declaring them fact?

The twist affects both the stock model as well as the sealed gap model.

Zero twist - scenario 2 points 9.2deg higher and has 30% more sail lift.

24 deg twist - scenario 2 points 7.6 deg higher and has 16% more sail lift.

While twist does show some effect, both these cases would still constitute a big jump in sail force.

 

And he still continues to ignore the elephant in the living room....

 

I am fully aware of the fact that the theory has limits, and all results are presented with those shortcomings implied.

If we are to quantify the shortcoming of theory then claims such as yours need substantiation. For all I know your claim has merit in real life and I would be the first to acknowledge that, just "show me the numbers".

So what is the elephant? Is it the wind gradient? Is it the free surface that cannot act as a mirror plane? Is it the turbulence of the hull?
It might be any one of them or various combinations depending on application. One thing is certain: you will be no wiser to it by just speculating.

 

Here is a question: is this 10 deg the geometric AoA of the 3D sail to the apparent wind or is it the angle of the local chord at a particular point in the span to the induced flow?

The lower the aspect ratio, the larger the difference between these angles become.

Second question: why only limit AoA and not set a lower limit to AR?

 

Since stall/detached flow is not modeled (that would be the elephant you are ignoring), and subsequently the calculated downwash can be bogus, it is to the apparent wind.

 

Like Joakim very well explained, wind gradient is one big elephant. If you use uniform flow, you exaggerate the loading of the bottom part of of the sail, and that makes closing the gap look better. Most of all, in real the real world, there's hardly any true wind close to the sea, so there's no loading either.

Another elephant is that the sails can only be sealed to the hull, not the sea. Daiquiri's sketch explains well why you cannot model that with Tom's spreadsheet.

Yet another is the sheerline rising when the boat is heeled - you can call that hull turbulence if you like. Here's some turbulence from the hull & the crew for you...

 

Right, now how do we calculate or simulate all these issues such that someone like whitepointer can get a more realistic idea of what results he can expect for his double-ender ketch?