Show HN: Sail boat simulation – try to sail into the wind
imakesnowflakes.com
imakesnowflakes.com
#1) Which is the front end of the boat? Not that hard to have a pointed end and a not pointy end.
#2) The sail shouldn't go all the way around, nor rotate around its center. If you are approximating something like a Laser it can go through about a 210 degree arc. Other boats, more like 120-160 degrees.
#3) So the wind is coming from the left? Note that almost all things generally show the fluids as streamlines, so at first glance it seemed to me that the wind was coming from the top and the boat was moving to the left.
#4) It's missing the force vector from the boat in the water. That's what really explains what is going on. The force vectors don't add up to the motion shown.
The only way I know how to explain a keel, is like ice skating. you plant one foot, and kick the other foot to the side and back, but the result is forward motion.
Ice skating is impossible with just one leg, and sailing upwind is impossible without both a keel and a sail.
Not true, it's just a little more difficult because it requires better balance. You can fairly easily do a swizzle on one foot shifting from your inside edge to your outside edge, similar to how a boat would tack. Yes, the blade of skate is acting similar to a keel, but no, you don't need two legs to feel the effect.
>It's missing the force vector from the boat in the water. That's what really explains what is going on. The force vectors don't add up to the motion shown.
Yes. You see, I couldn't really isolate that vector. What I am doing is taking the force vector on the sail and taking the projection of it in the direction that the boat is facing. Actually, the forward force comes from the boat trying to move in the direction of the force of the sail. But ends up moving a bit forward because the keel is in a slanted direction to that movement. So even if I show that vector, the viewer probably wont still have any idea regarding where it came from.
AFAIK, sails work like airplane wings (not like socks), creating lift on the convex surface by virtue of faster moving air and lower pressure. Perhaps your model assumes the same, but it's not clear from the way the sail is drawn, and it looks like sailing down wind is just as fast as sailing somewhat across the wind, which is in fact the fastest way to sail.
http://www.thepirateking.com/ships/sail_simulator.htm
@imakesnowflake, great work btw.
Maxxed out 9.25 knots at close haul at 33 knots of wind.
It only lacks the whoosh-rush-boom I recall from turning my Sunfish sharply across the wind.
Didn't capsize.
It works like this. The wind is split by the leading edge of the sail (the mast, headstay, or luff). It is split into two turbulent streams, one on each side of the sail. One stream has slightly less force on the sail so the sail begins to move in that direction. As the flow stabilizes over that side, the wind moves faster on that side than on the turbulent side. This creates lift like an airplane wing. So the wind is actually pulling the sail, not pushing it.
Lift on the sail causes the boat to heel (lean over). Some of the lifting force is counterbalanced by the keel underwater keeping the boat upright. The remainder of the force pushes the boat downwind. However, the flat keel underwater prevents the boat from sliding downwind, with the result that the boat is forced forward.
I recommend you read this book for your next version:
http://www.amazon.com/Sail-Power-Complete-Guide-Handling/dp/...
http://amasci.com/wing/airfoil.html
> As the flow stabilizes over that side, the wind moves faster on that side than on the turbulent side.
The most important rule of sail trim is to avoid turbulence on either side, __especially__ the leeward side. Laminar flow on the leeward side means the sail is deflecting air on that side as well, generating a "pull" in addition to the "push" on the windward side.
The keel generates its righting moment through lift also and not only its wieght btw, since a sailboat moves through the water at a slight angle.
Actually, with the boat heeled over, the lift produced by the keel is both below the hull's center of pressure and producing lift to windward. Its lift would therefore contribute to heeling.
> ...generating a "pull" in addition to the "push" on the windward side.
I tried to overlook this because of the scare quotes around 'pull', but you put them around 'push,' too! :o) Probably the easiest thing to say is that the 'push' is real, and the 'pull' is just 'less push.' That is, air is still pushing on the leeward side of the sail, but with less pressure than on the windward side, and that this reduction in pressure is greater when the flow is attached, as you imply.
Your explanation of the keels lift countering drift, but not heeling, makes perfect sense, and resolves a long-time confusion of mine.
As for "pull", I have a question: Assuming the flow is attached on the leeward side, does that side also turn flow and generate force?
FWIW the words pull and push are used in sailing training to emphasize the importance of keeping the leeward flow attached when going to weather.
Ding ding ding! Sailing is a practical matter, for nearly all sailors. It's nearly a folk activity. The models that are used to transmit and retain that information have more to do with achieving results than with achieving theoretical understanding. The trouble only really comes in when we try to use these pragmatic models as scientific models!
> Assuming the flow is attached on the leeward side, does that side also turn flow and generate force?
Notionally, yes. To be theoretical, attached flow on the leeward side is going to exert less pressure than detached flow would, due to irreversibilities in the detached turbulent flow. To be even stricter, because the equations of subsonic flow are elliptical and not hyperbolic, a mess of detached flow on the leeward side is going to jam up the whole works, even upstream and to windward.
The difficulty comes in when we try to mentally divide the flow between the windward and leeward sides of the sail. (Additional difficulty comes in because the "adverse pressure gradient" is essential to understanding this, but is tricky enough that it would probably do more harm than good to talk about it to sailing students. The first rule of Sailing Club is that you do not talk about the adverse pressure gradient. The second rule of....)
In practice, "[k]eeping the leeward flow attached" really amounts to "not stalling the sail." This is probably more important when sailing to windward because you're running as close to a stall as you can (at as high an angle of attack as you can) in order to point as high as you can. Also, with the sheets reefed in and with the higher apparent wind, the sail is more likely to be distorted. This can result in one part of the sail being stalled. Once one part stalls, the increased pressure in that area is likely to inject turbulent flow into neighboring areas that might otherwise not be stalled.
Another thing to consider is that there's something filling in that area of detached flow. Guess where it comes from? The windward side of the sail via the trailing edge. While the flow has to come off the leech in the direction of the leech, nature is a bit lest strict about what it does right after; It can get turned completely around, in short order, to fill that void behind the sail. The only thing it requires is that its turning radius not be unreasonably small. Congratulations! You've just (1) injected higher-pressure air into the lee, and equivalently (2) stopped turning the flow on the windward side. Because of (2), you can bet that there's an equivalent drop in pressure on the windward side as well.
That was the pragmatic way of understanding things. Now, I'm going to break the first rule of Sailing Club. When a flow is adjacent to a surface, the flow stops at the surface. That is, a 10 knot breeze blowing over your sail comes to a halt at the part of the breeze that contacts the sail. Because the adjacent flow hasn't touched the surface, it tries to maintain its 10 knots. However, nature is sticky. Any time there are adjacent regions of flow traveling at different velocities--any time there is a velocity gradient--viscosity tries to average things out: The flow right next to the sail is nearly stopped; The bit next to that is traveling a bit faster; And so on. (Incidentally, on a larger scale, this is why the wind is stronger aloft than on the deck. You're in the boundary layer, as it's called.) All of this viscous transfer is doing work on the flow. Doing work raises pressure. Because this happens as the flow moves downstream, more work is done on the flow the farther you get from the luff, and the higher the pressure gets. I don't think I need to convince you that fluids like to go away from high pressure to lower pressure. That is, fluids like to follow pressure gradients. And adverse pressure gradient can stop, and even reverse a flow. Note that this is likely to happen at the surface in question. Typically, the flow on the windward side has a stronger favorable pressure gradient to start with, and so can tolerate the adverse contributions of the viscous boundary layer without becoming detached. Because the point of a sail is to have lower pressure on the leeward side, there's less total pressure available to start with to overcome the adverse pressure gradient, and the flow can become detached on the back side of the sail.
This detached condition is called, in aeronautical terms, a "stall".
Just sheet in until the telltales flutter, and then back off until they don't. If that doesn't work, play with the vang until it does. ;)
This is the reason why sails usualy have short strips of canvas on the leeward side - to see if there some turbulence or not. Turbulence depends on sail convexity, it can be changed too, not only sail angle!
> So the wind is actually pulling the sail, not pushing it.
A handy distillation of fluid dynamics is, "Just like you can't push string, you can't suck fluids." That is, fluids don't resist in tension except in very particular very dynamic conditions that are wholly outside of the realms where our intuitions are developed. It looks like you've inadvertently modified the popular "Bernoulli's Principle" theory of lift, which was wrong even before you modified it.
There's no real point in doing anything like a rebuttal, so let me do the best I can, instead, to explain how lift is actually developed and how that applies to the case of a sailing vessel.
The fundamental thing that makes lift possible is the Kutta condition. This states that a real fluid has to leave a cusp in the direction that the cusp faces. There are complicated reasons for this having to do with velocity gradients, viscosity, radius of curvature of the flow, and even relativity, but the layperson can observe, feel, or take on faith that a flow leaves an edge in a direction parallel to the edge. This fact literally turns the flow and results in lift. (Historically and physically, we say that the Kutta condition resolves D'Lambert's Paradox via the Kutta-Jukowski theory of lift, if you want some keywords to look up more information.)
So, the direction of the trailing edge of the flow turns the flow. There are two ways that we can examine the result of this: momentum or pressure. It turns out that they give the exact same result for the lift of an airfoil, but are "intuitive" in different ways. The momentum theory just says, "Lift is just a force that is accounted for by the mass of the air that you're pushing aside. This is what you typically might call 'downwash'," but a bit more mathematically. The pressure theory says, "Look, that's all fine, but that force is experienced as a pressure distribution on a surface, because the only part of the wing that's in the flow is the surface. Let's keep track of (integrate) the pressure over the surface and see what the resulting forces are," but also a bit more mathematically.
So, take your pick of intuitions: Imagine a wing throwing air or fluid aside, or imagine it experiencing a pressure difference from top to bottom. They're both true, and are just different ways of accounting for the same phenomenon, different procedures for accounting for forces and masses as they interact with "control volumes," as engineers and physicists like to say. And remember that both intuitive explanations are driven by the Kutta condition, which is far more interesting, I think, than the erroneous Bernoulli explanation.
So, that's lift. I turn a flow and I get a reaction force. Peachy. How does it apply to sailboats?
Other than the buoyancy force and gravity, every force that we care about on a sailboat is due to the operation of a foil (hydrofoil or airfoil) within a fluid (water or air). The hull of the boat, the rudder, and the keel act like wings of (typically very low) efficiency. Their total lift acts to resist the downwind drift of the boat in the water. The sail acts as a wing. Depending on the point of sail, the lift of the wing acts more or less, literally, in the direction of travel of the boat. (Close to the wind, the majority of the lift is actually wasted in heeling the boat over and trying to slide it downwind.)
I know I said that a rebuttal would be unproductive, but there are a few things I suddenly can't resist.
> However, the flat keel underwater prevents the boat from sliding downwind, with the result that the boat is forced forward.
The "flat keel" is a wing. It prevents this "slide" via lift. It's also not helpful to split the analysis into 'forces' and 'constraints'. The keel produces lift that is a fore that contributes to the balance of forces on the vessel, resulting in motion.
> [The wind] is split into two turbulent streams, one on each side of the sail.
Turbulence is a complete red herring, here. Turbulence has nothing to do with lift (other than causing stalls and complicating the analysis).
> Some of the lifting force is counterbalanced by the keel underwater keeping the boat upright.
Well...the center of mass of the entire boat levered, as it is, upwind of the center of buoyancy, produces a moment that counteracts the moment of the sail's lift.
> For instance, a sailboat is not powered by the wind pushing on the sail.
It kind of really is. It's just more complicated because analyzing fluids requires that we analyze continua, and not just discrete particles. In the end, though, sailboats sail because they exchange momentum with the atmosphere and the water through distributed forces, or pressures.
I've got a small yacht and while I can easily sail into the wind with a full main sail and small jib, there's no way in your simulation to change the angle of the jib (in front of the mast) in relation to the mainsail. The physics are wrong to since the position of the keel is purposely centered behind the sails' COE (otherwise it would be easier to sail backwards than forwards).
I think you need a rudder to counteract the force of the sail swinging the boat around the keel and you need to consider that a full sail (purposely) becomes "wing-shaped" as noted above.
EDITS:
1) I should also note that it's possible in some small boats to sail almost directly into the wind ... I've managed about 15 degrees off the wind in a very lightly loaded Hobie Cat (without a jib).
2) Turbulence is an undesirable effect ... many sails have "tells" sewn into them that show you how the air is moving over the sail. You never want the tell to be wildly gyrating.
So, if I understand you correctly, to put it into one sentence for my friends at the club, air flows along the curved sail which, through pressure differences or momentum, turns the direction of the wind, and the opposing force of the turn causes lift?
My other question is, how does a symmetrical keel then cause lift? Thanks.
You're right about the sail turning the flow. However, the "curve" of the sail is not the only way to turn a flow. In the keel's case, it sees an angle of attack to the flow, and this causes lift. Imagine, for a moment, that the boat isn't heeled over, and we're looking down on the boat as it moves through the water. The bow of the boat, when pointing, points a few degrees upwind of the direction of travel of the boat relative to the water. We go back to the Kutta condition to see that the flow has to turn to point directly off the trailing edge of the keel and see that water is being "thrown" to leeward.
It's a little more complicated than that, even, for keels, especially for full keels. They're still "throwing water," but they're solidly "low aspect ratio" and finite wing effects dominate: Lower lift, higher drag, higher stall angle, softer stalls, attached vortices, and so on. All sorts of fun. There are reasons that race boats tend to have fin keels and daggerboards! :)
There's no reason to complicate it like this. It's simply Newtons laws at work. You accelerate a mass of air in one direction, you get pushed the other.
The over complicated theory you've stated there is actually completely incorrect btw.
http://en.wikipedia.org/wiki/Lift_(force)#False_explanation_...
I'd also love to see centerboard or keel forces since that's what makes the boat go forward.
Thanks for making this. As a teenager I used to sail competitively and this brings good memories. Sailing close hauled under decent wind was incredibly fun.
It reminds me of a car parking simulator [1] I made a while ago.
[1] https://dl.dropboxusercontent.com/u/8554242/dmitri/projects/...
I first thought that they were the sea waves, of a nice visual effect, but they are a part of the simulation.
I learned this from an awesome book, no doubt one that many on HN have heard of, Thinking Physics, by Lewis Carroll Epstein that is worth getting, especially if you have kids and they don't already know basic physics.
An AC-72 might be fastest upwind, but a square-rigged ship definitely isn't.
The more interesting part is that boats can actually sail faster than the wind that powers them! Depending on the surface, from 1.6 times the true wind speed in a catamaran on water, to 5 or more times the wind speed in an iceboat.
You may want to check out the guys at SailX. They are focusing on the simulation of sailboat racing, but they needs some of the same principles that you are trying to simulate. http://www.sailx.com/
For instead of sail-power complete guide, take a look at Aero-Hydrodynamics and the Performance of Sailing Yachts. This is a fun tech-read. http://www.amazon.com/gp/product/0071629106/ref=pd_lpo_sbs_d...
To add to the discussion on sailing faster than the wind, I present Blackbird - which was a land yacht made to sail faster than the wind....DDW! The final design was able to sail 2x the speed of the wind.
This was actually quite a reasonable sailing simulator if I remember correctly. Good enough that when I finally got a chance to sail for real a few years later, I knew what I needed to be doing in terms of positioning sails to obtain various points of sail. Of course, navigating without an overhead map being available, and knowing about ropes were two things that I still had to learn the old-fashioned way :)
A feature idea: would be cool to have a level with "increased difficulty/realism" with just one degree of freedom - the sail. In real world you can not turn the vessel on a whim. And the next level, make the wind gusty, and make it a requirement to keep a certain speed to avoid "drowning", and you have a rather fun windsurfing-simulator game :)
Yes. I originally planned to let the boat turn only using the rudder (the blue line at the end of the boat). But after a while, I felt that it would make very difficult to get it working.
Thanks for sharing, would love to see any improvements toward making it more realistic (as a teaching tool for sailing or something).
However, I did manage to get a small resultant to the left, so woot! :)
Boat: 130
beating to windward while close hauled
wish I could see the keel forces
pretty cool!
Ships can advance into the wind because of:
- drag reduction (hull shape, keep, rudder) : sideways force translates to a diagonal forward movement
- most importantly, lift: more than pushing the sail, the air pulls the sail sliding from the outter side of it (which is why a lot of the cloth doesn't actually suffer much forces in the middle of the sail which is what initial intuition would tell - by far the most tension happens in the corners).
If push is stronger than lift, turbulence on the other side of the sail slows down your boat rather than speeding it up. "Tell tales" stuck in strategic points of the sails allow to quickly tell if the flow is correct, and therefore the trimming is good. Boat speed, flapping and shape also show these forces.
http://i.imgur.com/hZlb34r.gif
http://i.imgur.com/uypQGJJ.jpg
http://i.imgur.com/5tJLUgF.jpg
I don't think there's sail shape or lift in the simulation, but I suspect that the drag is rather low or zero, which allows for some gains into the wind (but real boats do better).
No. It is a very rough simulation (I am not sure if I can even call that). It only takes into account the force of the wind 'hitting' on the sail and the 'drag reduction'.
I am not a physicist or have experience with sailing. I am just a freelance programmer (currently out of work) from India, which means I even haven't seen a real life sail boat. I made this using the stuff I got from Wikipedia and a bit of intuition. So I guess it shows.
But I have a question regarding the boat traveling faster than the wind. Isn't it due to the fact that the lift is a constant force, regardless of the speed at which the boat is traveling? But a boat that is using the force of wind hitting the sails, can only sail as fast as the wind, because at that point, there is no wind hitting the sails.
~~This can't happen if the boat is going straight downhill, because the apparent wind would be zero (as the boat is moving with the wind).~~ [ETA: that's not actually true in the general case. You can't have a sailboat move downwind faster than the wind, but if you have a gear mechanism and a turbine you can do it.]
But when tacking downhill, the movement of air around sails still generates lift.
http://www.croatiacharter.me/images/download/1204484832-boat...
As for the boat traveling faster than the wind, it has more to do with sailing angles than lift proper. The boat can get lift and push from "different air" as it slides perpendicularly with the direction of the wind. It's more of a constant force that produces acceleration until it's cancelled out by forces of resistance.
Typically the fastest points of sail of a boat (varies with hull shape, sails and their configuration) are about 45%-75% from the wind direction.
http://i.imgur.com/61OGxxx.jpg
In polar diagrams 0% is into the wind, so 180% is dead downwind and 120% is about 60% away from dead downwind which is aprox the fastest direction for the boat described by that polar diagram.
This is an even faster boat, build for speed records http://www.youtube.com/watch?v=wnjyusAgk8I
Kids don't question a counter-intuitive: just teach them 'point the boat close to the wind, do this to the sail, and go.'
Isn't the boat a pentagon?
https://news.ycombinator.com/item?id=8427491
I have updated the page to read 'green polygon' now..)