Deck sweeping sails and effective aspect ratio

Please elaborate on the effect of dihedral and how it is defined in the context of heeling.

Is there any degree of "ground effect" taking place due to the closer (and more parallel) proximity to the water?

 

There is something wrong in the table. Heeling moment decreases considerably, but at the same time keel lift increases. Keel lift must be quite equal to heeling moment/heeling arm. If you calculate the heeling arm from that, you get about 10 (feet?) for left column and 6.7 for the right. Can there be such a huge change in center of efforts of keel and sail?

 

Nonsense! Your changes increased the Hydro L/D from the baseline case. You totally violated your claimed desire for apples-apples comparison.

If you want apples-to-apples, start with your baseline and it's calculated pointing ability.

Make the least amount of changes possible, bring the deck up to the sail with no other changes in the boat and then look at the result in pointing and heel.

You could also just bring the whole sail down to the deck.

A third scenario would be to leave the height of the sail peak where it is, and then bring the foot down to the deck. This would be as if the sailor closed the gap with a tarp. (added sail area)

A fourth comparison would be to leave the peak height and bring the sail foot down to the deck while keeping the sail area equal to the original.

There will be theoretical performance improvements to show in some of these scenarios. But anybody can put formulas and numbers in a spreadsheet. The question is whether the results have any resemblance to reality.

The posts in this thread are hinting that for heeled boats, the spreadsheet results inadequately model the flow over the hull and under the sail. However a non heeled boat or land sailer might be more well represented by the spreadsheet.

 

Remember this is a dinghy. The "geometric" center of effort was lowered by 2ft (actually 2.2) as the sail plan was lowered, and due to the new lift distribution, the new center of effort dropped an additional 10". The luff is only 13.5ft long, which is why the drop seems disproportionally big.

The fin keel has a CoE depth of 1.3ft and the shallow keel's is 1.1ft. It corresponds well with what you calculated.

 

And anybody can throw around accusatory comments and pose open ended questions. So where is your contribution to answering this pertinent question, other than one misinterpreted report?

No one is making any claims here, and everybody is clear on the fact that the theory does not seem to match reality.

Please consider your tone before posting again. There is nothing wrong with pointing out an issue or questioning someone's interpretation, but such immature rants are unwelcome on this thread.

 

My contribution is to try to help you understand that the way you are using this spreadsheet renders misleading results.

'To try to help you understand that if you change several things in an an experiment, whether it be by spreadsheet or tow tank, you cannot claim cause and effect in any variable-result unless you have proven no correlation with the other variables. Ideally you change only one thing between trials.

BTW ref post #46 your input for the change in sail lift on the modified boat looks very suspicious. You show a 25% gain in lift force which cannot be accounted for by the change in apparent wind velocity alone.

Intentional or not, what you did was make questionable input to an unproven modification of Speer's spreadsheet.

 

The fact that you do not grasp where the 25% additional lift in the dinghy example comes from, even just in theory, suggests that you have not tried to do any verification but are simply objecting for the sake of doing so. Go and download Vortex 95 - the original - model a sail and see for yourself.
Only once YOU can explain the existence of the additional lift will you be in a position to question other results based on theory, apples vs apples or prickly-pears.

I will nevertheless prepare the scenarios you requested.

 

Actually Tom is up to Vortex2K now. I don't remember if Vortex 95 automatically designs the chord distribution, but if it does there might be a simple explanation for why you are getting such unrealistic big jumps in your lift when closing the gap.

Using his default setting 0 gap the spreadsheet designs a sail with a shape of half an ellipse. If you change the gap to 1, the spreadsheet designs a nearly elliptical sail. Could it be that you didnt realize that? Take a look at the screen grabs. The graphs show the automatically designed sail plan as if the mast were laying on the horizontal axis with the foot of the sail to the left. The line of the graph being the leech.

Do you understand that this spreadsheet automatically changes the shape of the sail?

 

The four scenarios called for above is laid out with the original comparison for reference.

The heeling moment and maximum section lift coefficient of the original dinghy was used as a limiting case.
If heeling moment for a particular scenario exceeded that limit, the assumption is that the sail will be sheeted out to reduce the lift coefficient until the heeling moment has been reduced to within the limit.
Where heeling moment was not limited, the assumption is that the sail will be sheeted in until the limiting lift coefficient is reached.

 

Dihedral = heeling. If you imagine the mirror model under the sea, you get a wing with heaps of dihedral.

What I meant to say, when looking at the photo, was that if I was a bunch of air molecules, travelling with the wind over that sheerline towards the mainsail heeled as is, I would likely flow slightly up and above the clew, out from the leech, rather than forcing myself down towards and under the boom . Some air will always slip under the boom right at the tack already, but the attitude of the boat as shown surely will cut down that tendency. My common sense says that, independently of my CFD sim.

When it comes to the theory/spreadsheet you are discussing, like I already say in my post 67, the lifting line theory is inadequate for modelling sail sealing to the deck. A minimum would be to consider the gap to be the distance between half the height of the freeboard (or the sea surface) and the sail foot, not the distance between the deck and the sail foot. So you can never achieve zero gap (in terms of that theory), unless you bring your sail to the sea surface. The deck is a mere smallish end-plate, not the mirror plane assumed in the lifting line theory.

 

Vortex95 does the same on the Design tab. Real sails do not look like that because of the importance of limiting heel. That is why I am modelling my sails on the Analysis tab where the sail planform is a user-defined shape. I approximate a triangular sail with a bit of roach by using a 1:10 taper ratio and then playing with the sail twist until I get a nice constant Cl distribution. This also then yields a downwash distribution that tapers off linearly - as per Tom Speer's suggestion for least induce drag in a heel-limited case.

Compare the amount of lift generated by the lower third of the sail in the attached screen grabs (only look at the blue lines). They represent the Cl and Lift distributions of the stock Spindrift sail and Scenario 2 in the table (lowered sail as is)

 

Here is the extract from Tom Speer's manual for his spreadsheet. He does make allowance for modelling a free surface, but suggests that the sea be treated as hard ground when modelling a sail. The free surface only comes into play with surface piercing foils travelling above certain speeds.

I don't know enough of the theory of modelling the gap or the effect of a free surface to call anything into question. Aside from the application of theory, how accurate then would a cfd analysis be that treats the sea as a hard surface?

"Surface Effects

When a wing or sail is operating in proximity to a surface, such as the ground or sea, the flowfield is influenced by the surface. For sails, the water is assumed to be a flat, solid surface like the ground.
A wing or sail near such a surface behaves as though it had a virtual twin, mirrored in the surface. This virtual image makes the surface a plane of symmetry, enforcing the condition that there can be no flow through the surface.
For hydrofoils, such as keels, rudders, and hydrofoil boats, the free surface of the water is more complex. At very low speeds and at high speeds, the water surface is approximately flat, and this theory is applicable.
At very low speeds, the gravity forces are large compared to the pressures exerted by the hydrofoil, and the water's surface acts like a solid surface, and the same kind of virtual image is used. This is the "zero Froude number" condition.
The high speed free surface condition acts as though the surface were located half way between the wings of a biplane, with the one wing formed by the hydrofoil, and the other a virtual wing.
The geometry of this virtual wing is the same as the virtual image used for the solid surface, but the lift is directed in the opposite direction. This enforces the condition that the pressure at the surface is constant.
This is the "infinite Froude number condition. For hydrofoils operating at approximately one chord length in depth, this is good approximation for Froude numbers greater than four."

 

OK so there is the reason for the unrealistic jump in lift. With traditional sails we do not have that much control over sail twist from foot to head.

Have you tried to do the same diddling with twist on the original sail?

 

I don't know how well a gap is modelled with that lifting line spreadsheet, but at least it says, that the "The freestream wind is also considered to be uniform in this version." Thus there is no vertical wind gradient. This is very important in your case.

You are bringing the foot of the sail very close to water. The real wind speed will be around 50% of the mast top wind speed there. Thus also the apparent wind AND the angle of attack will be much lower. Due to lower wind speed the force is reduced to less than 50% (depending how much AWS is lower) at the same Cl and the Cl will be lower as well due to lower AoA and the force vector will be more backwards giving more heeling force and less driving force.

If you would make the same comparison with the same model, but including vertical wind gradient, the result would be very different. Now you think you are gaining very much from increased Cl near the foot and loosing nothing at the rest of the sail. In reality, the foot sees much lower apparent wind and has lower angle of attack and thus produces much less driving force.

At the same time the rest of the sail is lower and also sees lower apparent wind and angel of attack.

Thus the total impact might be negative despite the huge benefit in closing the gap to the performance of the planform (according to this model).

Also you are modeling an open dinghy. There is no uniforn deck to seal to. And sealing to deck is not the same as sealing to mirror surface in the model.

 

I used about 12deg washout on the original dinghy to optimise the load. More washout would unload the top part of the sail, while less washout would cause the head to be limited by stall (Cl max) while the rest of the sail is unloaded.

If the gap is sealed without other changes such as aspect ratio (scenarios 1 and 2), the same amount of twist once again yields best results. I have never measured actual twist control range, but based on my own sailing observations seem within reason.

I kept the twist at 12deg for scenarios 3 and 4 even though the head was a little unloaded in each case. Decreasing the twist would load the sail more on top and since heel was already a limiting factor, no additional gain would have been possible.