Deck sweeping sails and effective aspect ratio

True, but not quite the case here. The flow separation at the root occurs at a local AoA of barely 5deg. I used the top surface of a NACA4415 airfoil as my leeward surface and just slightly filled in the windward side from the leading edge up to 50% chord. This airfoil only starts stalling at 12deg for the 2D or infinite AR case. Granted, the 12deg is applicable only for a Reynolds number of 3 million
and the model only has a Re of 0.5 - 0.7 million at the foot, depending on AR.

Aside from Reynolds number, the sharp taper and twisted flow would both conspire to cause the head of the sail to stall long before the foot - at least in uniform wind.
Another peculiar effect of the wind gradient is to cause a higher pressure on the upper half of the windward surface that tends to deflect the streamlines on the lower half down towards the deck (not to be confused with "downwash" in the aerodynamic sense). Had this been an aircraft wing, the maximum high pressure would have been right at the foot and all streamlines would have some deflection towards the head, increasing in angle as you get closer to the head.
It is not clear whether this phenomenon has any contribution to the separation on the leeward side. I modelled some "wing fences" which straightened out the flow a bit but had little effect on the separation.

 

Will,
Separation is not always the ugly thing one uses to consider:
Michael Selig in his design philosophy for his hight lift low Reynolds wing section S1223, includes some aft-loading in the 2D section which triggers a little bit of TE separation, but overall the lift is higher & L/D ratio better.

In this case, I'd say candidly that the overlap of 2 orthogonal pressures distributions (1 for the sail +1 for the tramp) is likely to create separation even at the leading edge.

If a Cat-boat without jib, it could be interesting to design the LE of the front cross-beam with this separation issue in mind.

Best regards

EK

 

I recommend that you ignore any AoA values in this study.

If you want to meaningfully compare different sails, you should compare the CD vs CL curves, i.e. the drag polars, since that's what affects boat speed. The AoA values at the design point, or at stall, or whatever, are relatively meaningless in this regard.

If you have a twisted wind profile and there's no unique relative wind direction or magnitude for the CL,CD definition, then you can use the lateral and driving force components instead, since those are unambiguous.

 

Thanks Mark. I only mentioned the AoA (calculated for the apparent wind at the exact height of the foot) because the flow separation seems unusually premature. Your insight on this would be highly valued!

For sail comparison I have been calculating CL and CD based on the windspeed at an arbitrary height of 6ft. I normally look for the sheeting angle that produces maximum attached flow and then also evaluate two more sheeting angles, typically 2 and 4deg less than the critical AoA. These lower angles normally yield better L/D ratios and hence also better overall Drive/Heel ratios which is what I am particularly interested in optimising.

I prefer to use drive/heel rather than drive/lateral force because the CoE is often very different.

 

Will, you posted a pict of a Hornet fighter showing the vortex coming off fuselage strakes. Those strakes (LERX) were designed to manage the flow at the fuselage-wing junction during high AOA flight.
Properly positioned strakes might do some good in your efforts to get rid of the flow detachment at the foot of the sail. Try a mini jib that has roughly the same proportion to the sail as the strake on the Hornet has to it's wing.
Take note though that the strake is there to generate a vortex at high AOA so the benefit might not be there unless the apparent wind is 30 degrees or so.
Eventually you might be able to integrate a vortex generating structure into the deck that ends up having internal volume for storage.

http://www.nasa.gov/centers/armstrong/news/FactSheets/FS-002-DFRC.html#.VVxW5PlViko

http://en.wikipedia.org/wiki/Leading-edge_extension

http://en.wikipedia.org/wiki/Strake_(aeronautics)

http://bagera3005.deviantart.com/art/F-18-High-Angle-of-Attack-90505230

 

Will,

Regarding the minor separation at the leeward bottom TE, I wonder if it would be possible to delete it by cutting the trampoline, just where the separation starts, in order to see if the fresh air from underneath would change something.

May be the point is whether underneath air flow is "fresh air" enought to do the job?
Cheers

EK

 

Will, that huge improvement in "drive" appears to be at odds with the reality that we've seen in eons of experimentation in development classes. While I understand that people in the USA consider that most design is "within the constraints of some class rule or other" there are many classes in other countries where most designs are outside such constraints. As noted before, even in these classes, the experiments with closing the gap do not demonstrate such a difference.

 

I have tried strakes of various sizes without effect, perhaps the local velocity is not enough to set up a strong enough vortex? Vortex generators of the mini-wedge type on the sail's leading edge also had little effect although the mesh might not have been fine enough to capture the mini vortices in detail.
The mini jib increased drive but only by the equivalent amount of its increase in total area - flow separation remained.
I also tried a number of slots in the tramp, some with sizeable scoops underneath - no joy.
One thing that had at least a measurable effect was a 2" gurney flap pointing up from the trailing edge of the tramp.

Now that I have turned the top of the tramp into an ideal smooth surface the flow looks much better (as do the results) at the sheeting angles that used to give separation. I have since discovered another slight change in geometry that further reduced heeling moment without reducing drive.

 

CT, I agree that these results would not be achievable in practice since the model and simulation involve a number of idealised assumptions. It does however serve to highlight the areas that would require particular attention if one is to get the maximum benefit of sealing the gap. A lot of it has nothing to do with the sail or the gap itself.
Just to put this into perspective, my initial comparison with the twist and wind gradient in place only showed 1.3% increase in drive. To even measure that in terms of boat-speed difference in real-world testing conditions would be near impossible, even more so for racing boats that operate at S-L ratios with steeply increasing hydro drag. Most of the subsequent modifications to the model yielded only small improvements, but they start to add up after a while. I have done nearly 130 different simulations and the results quoted are the cumulative effect of many modifications.

It is pretty clear that wind gradient has a huge impact on the results when comparing it to uniform wind-speed results. That means the boat-speed/wind-speed ratio also plays a significant role in how much improvement one can expect. This model is only simulated as doing 5kts in 15kts true wind. The height of the tack compared the the luff length is also quite big, resulting in very little wind twist along the sail.
Other factors that might make the improvement seem unusually large is the fact that I am comparing it to a sail that would hardly be expected to do much upwind sailing at all.

 

CT,
You can have a pretty clear picture of the actual performance pick-up of closing the gap, looking at the relative performance of the C-Cats during last cup in 09/2013 UK.
I was not on the water, but the real time comments and video available at the RSC club-house were the same for all the windward leg.
GroupamaC use to lead by 40s at the windward mark, I don't remenber how long lasted these windward legs, probably between 10 and 20 mn.
Of course, different boats, different foils set-up, have to be considered, but I think this remains a good benchmark.
Regards
EK

 

/\

Erwan, surely the other differences between GroupamaC and the other Cs mean that the benefit of closing the gap cannot be isolated. This is particularly relevant when there are several examples of older UNsuccessful C Class cats that have very closed gaps, and many other classes where craft with closed gaps have not shown any advantage.

 

So I've gone from this

 

To this (a Hoyt jib boom)

 

And I'm toying with the idea of filling in below the jib boom to get some sort (or more at the very least) of seal.

Initially, it seems the quality of disturbed flow (by how it feels to my hand) is very different between the two setups at the foot of the sails- the smaller gap had very lumpy flow that spread out over a large part of both sides of the jib, and the Hoyt jib has a very defined vortex that seems to be almost confined to the lee side in a fairly tight and coherent way. I'll put some telltales on the bottom next sailing season to get a better idea. The main question now seems to be if there is an advantage to a more controlled bottom vortex on the higher foot that outweighs a flap below the boom that either is small so it doesn't get messed up with the lifelines offwind, or big and sacrificial?

FWIW, the Hoyt setup off wind is spectacular. Jibing is easy too...

 

Were you ever able to test your foot-flap?
Your qualitative findings about the steadiness of the foot vortex corresponds well with what I have seen in my simulations. The smaller the gap on the headsail, the more likely it is for the flow around the foot to be disturbed by flow separation from the stem and topsides.