Dynamic soaring
en.wikipedia.org
en.wikipedia.org
P.S. bring ski goggles—you’ll need them if the conditions are good, you won’t see without them if it’s a 400mph day—which is more and more common these days!
Some other interesting HN comments:
https://news.ycombinator.com/item?id=32739679 https://news.ycombinator.com/item?id=38653688
https://newatlas.com/aircraft/dynamic-soaring-speed-record-s...
Edit: very cool. I watched the video referenced in the HN comments (BTD10: The 835kph Sailplane and Dynamic Soaring) and I get it.
Say you have a glider with a 2 m/s headwind. If you cross into a wind mass that's a 10m/s headwind, you've just gained 8 m/s of airspeed for free. Rinse and repeat.
> Dynamic Soaring as a Means to Exceed the Solar Wind Speed
https://arxiv.org/abs/2211.14643
Related video:
On the other hand Swifts apparently can stay up for months. They eat flying insects so can stay in the air. https://www.latimes.com/science/sciencenow/la-sci-sn-common-...
Trying to do so in the boundary layer at ground level would be extremely dangerous, if at all possible, but Renner's insight was that there is sometimes an exploitable windshear at a safe altitude, on the boundary of an inversion.
From the article: "Renner has been attempting dynamic soaring for about four years, and the difficulties can be gauged from the fact that he has achieved only four really successful flights in that time."
[1] Flight International, 22 October 1977 reprinted here (PDF): https://www.sac.ca/index.php/en/free-flight-magazine-2/1970s...
It's sort of like how if you push someone on a swing a little bit each oscillation you can keep adding energy even though your pushes aren't particularly fast or energetic.
What actually happens in any of the turns is that the glider undergoes a change of velocity equal to twice its original airspeed, in the direction opposite to which it was originally flying. This is simply a consequence of it flying an energy-conserving path (we are ignoring drag and also assuming the turn is made in level flight, so there's no change of potential energy.)
When measured in a frame of reference moving with the wind, with W as the wind speed and A as the air speed, a glider flying upwind has a velocity of -A, and after turning 180 degrees has a velocity of +A. The magnitude of both velocities is the same.
From the ground frame of reference, it has gone from W-A to W+A, so the magnitude has increased, from the difference to the sum. In both cases, the change in velocity is +2A, regardless of how strong the wind speed is or whether there is any wind. The appearance of there being an acceleration when the wind is blowing but not when it is still results from considering the change of ground speeds before and after, instead of change in velocities (measured in any inertial frame of reference, as long as it is the same one before and after.)
It is not easy to see intuitively how this allows the kinetic energy of the glider to keep on increasing, but perhaps this will help: the 180 degree turns are like elastic collisions (i.e. without energy loss) between the glider and the air. From this perspective, we can make an analogy with a bat and an elastic ball: if you bounce the ball between the bat and the floor, striking it downwards each time it bounces back, it will keep on picking up speed: it rebounds from each blow you give it faster than it was coming up, and it keeps that newly-acquired speed when it bounces back off the floor. In this analogy, the bat stands for the wind at altitude, and the floor stands for the still air below.
This seems to be begging the question. Why does it get a velocity of +A? It seems obvious if you assume the inertial frame of reference of the wind, but from the ground it's very surprising. Why is the frame of reference of the wind the "right one"?
Perhaps a better way to ask is: when the second turn happens, why don't we go back to -A? It seems the "correct frame of reference" is now the still air? But why?
It's easy for me to intuitively understand that there's a source of energy at this windspeed gradient but it's hard for me to understand that this technique can extract it.
Why is the frame of reference of the wind the "right one"?
There isn't a right one in the sense of being the only one giving the right answer, it's just by far the easiest one to work with, and then you can convert the result to any other inertial frame via the Galilean transformations [1]. In the frame in which the wind speed is zero (and only in that frame) you can calculate the path of the glider while turning by using the well-known solution for uniform circular motion [2], in which the centripetal force is always perpendicular to the direction of motion, and the speed of the moving object is constant.
The difference between the solution in that frame and the ground frame is nothing more than the difference between drawing a semicircle on a piece of paper, and going through the same motions on a piece of paper that is moving.
It's easy for me to intuitively understand that there's a source of energy at this windspeed gradient but it's hard for me to understand that this technique can extract it.
The turning glider exerts a force on the airmass it is turning in (Newton's third law), which is itself accelerated, in a direction that lessens the difference in speed between the two airmasses.
Your middle question will take more time to address, and I plan to come back to it tomorrow.
[1] https://en.wikipedia.org/wiki/Galilean_transformation
[2] https://phys.libretexts.org/Bookshelves/University_Physics/B...
We can go through the sequence of events as shown in the article's animation, with a bit more explanation at each step. Like the article, I will ignore both drag and the fact that the glider is alternately trading kinetic energy (speed) for potential energy (height) and vice-versa, as it rises and falls. In practice, both of these will have a effect, but if the wind gradient is sufficiently strong, they will be small enough to ignore.
For vectors (velocity and acceleration, but not speed), I will use the convention that rightwards is positive. There are two frames of reference that are relevant: the one moving with the upper airmass, and the one pinned to the lower one (which, as there is no wind in the lower region, is also the ground frame of reference.) At each step, I will calculate the glider's velocity in each frame of reference (we can always measure the glider's velocity in either frame, regardless of whether it is in the corresponding airmass.) A key point to bear in mind is that the glider's airspeed is, by definition, the magnitude of its velocity in the frame of reference in which the air surrounding it is not moving.
1) The animation starts with the glider ascending in the lower layer at some airspeed A₀. Here, its velocity in the lower, ground frame is -A₀, and in the upper frame, which moves with the wind and has a velocity W relative to the ground frame, the glider's velocity is -(A₀ + W).
2) It crosses the boundary between the two airmasses, encountering a headwind. We are ignoring drag, so it experiences no deceleration and still has the same velocities as before in either frame, but now the frame for calculating the airspeeed is the upper one, so its airspeed has increased to A₀ + W. There is nothing strange about this: it is just what we feel when we walk out of a sheltered doorway into a windy street.
3) The glider reverses direction. On the basis of what was said in the previous reply, the airspeed remains the same, and the velocity in the local (upper) frame is now A₀ + W. Applying a Galilean transformation, the velocity in the ground frame (which, from the glider's perspective, is pinned to a headwind) is now A₀ + 2W. Again, there is nothing strange here: if you watch birds circling in a strong wind, they appear to slow down as they turn into the wind, and pick up speed as they turn downwind. Note that, in either frame of reverence, the change in velocity in the turn is +2(A₀ + W).
4) The glider crosses the airmass boundary in the opposite direction. Because of the reversal of direction, this also feels like encountering a headwind with speed W, and its airspeed increases by that amount, becoming the magnitude of its velocity in the ground frame, A₀ + 2W.
This is the step that seems most like cheating, I think: why doesn't crossing back through the boundary reverse the effect of crossing it in the first place? It seems to violate some basic rule of symmetry, tit-for-tat or no-free-lunch. I don't think there is any everyday experience like it, but from working through the example in detail, we can see that this must be the case, and that our intuitions have been nullified by the glider's reversal of direction between the crossings.
5) The glider reverses direction again. As before, the airspeed remains the same: A₀ + 2W. The glider is back where it started, but going faster.
-------------------------- velocity --------------------------
Step ground frame upper airmass frame change from previous step airspeed
1 -A₀ -(A₀ + W) N/A A₀
2 cross windshear -A₀ -(A₀ + W) 0 A₀ + W
3 turn A₀ + 2W A₀ + W 2(A₀ + W) A₀ + W
4 cross windshear A₀ + 2W A₀ + W 0 A₀ + 2W
5 turn -(A₀ + 2W) -(A₀ + 3W) -2(A₀ + 2W) A₀ + 2W
The way one turns a glider is to bank it in the direction one wishes to turn, which causes the lift to have a horizontal component perpendicular to the direction of motion of the glider. [1]
One problem with trying to calculate the path in the ground frame directly is that the glider is only pointing in the direction it is moving when going directly upwind or downwind. For example, suppose the glider is going westwards across the ground, in a wind from the north with a windspeed equal to the glider's ground speed. The glider will be pointing north-west (and encountering an airflow - the relative wind - from the north-west) and the horizontal component of the lift will be pointing south-west. This is inclined forwards with respect to the glider's direction of motion in the ground frame, and the acceleration it causes, while being the same in any frame of reference, will, in this frame, result in the glider speeding up - and this is what an observer on the ground will see: as the glider turns downwind, it picks up speed relative to the observer.
[1] https://www.boldmethod.com/learn-to-fly/aerodynamics/the-aer...
> if we adopt the premise that the centripetal force is always perpendicular to the tangential speed
This also seems to be an extremely important element. I think the underlying thing that I wasn't getting was why the "pivot point" of the rotation moves the way it does.
If someone makes software that can seek out and make use of lift (using various information sources), it could pave the way for very low-energy drones.
EDIT: Interesting.... seems there is something already that handles the "circling and staying in thermals" part.
https://ardupilot.org/plane/docs/soaring.html
https://arxiv.org/abs/1802.08215 "ArduSoar: an Open-Source Thermalling Controller for Resource-Constrained Autopilots"
It would be a pretty good mode of transport if the general population could be trusted to responsibly and competently fly aircraft. Very low carbon.
It's considered 3-4x more dangerous than driving based on the annual fatality rate.
Not that it isn't cool, it just won't become mainstream due to safety alone.
As a sport it's absolutely breathtaking, it might be the best single thing that I've ever done. Please do mind the safety, though, it is not easy and it is not safe even though it may seem so after you've done your first ~hundred hours. It is a lot safer than paragliding because you'll crack the landing gear or the tail and not your spine on a bad landing.
Many years ago I did both skydiving and paragliding. For skydiving, in approx 70 jumps I had one slightly twisted ankle (which was actually an earlier injury that had weakened the ankle) and saw one situation where a jumper was stretchered away (they landed on their backside because they were trying to hit a target with outstretched legs. I also had a parachute malfunction and safely deployed my reserve. In paraglidimg, over a shorter timeframe, two people in the club received bad injuries (broken leg and broken hip) and another smashed their helmet in a collision with a dry-stone wall.
The problem with paragliding seemed to be that canopies could collapse at very low altitude due to air turbulence (rotors) around the tops of hills where flights started and usually ended. In contrast, in parachuting from planes, there was 'plenty' of time to deal with parachute opening malfunctions (and to open reserves), and it was rare for canopies to then fail due to wind or terrain.
I have been watching with interest the advance of traction kites in commercial shipping. The description of Airseas's implementation[0] resembles dynamic soaring but details are wanting:
> “What differentiates it from other wind solutions,” says Bernatets, “is that the wing is not just pulled by the wind and countered by the ship.” Instead, it flies in figure-of-eight loops, which multiply the pulling effect of the airflow to give what he calls “crazy power.”
Whether or not it is caused by dynamic soaring, it is true that such kites pull much harder in the same wind while turning and a figure 8 pattern would maximize that without twisting up the lines.
My understanding is that the incentives don't encourage the traction kites' installation because the owners of the boats don't pay for fuel. Perhaps that friction can be overcome as the fuel costs continue to rise.
[0] https://www.cnn.com/2023/06/30/travel/airseas-giant-kites-sh...
NZ is great for this.