Scaling Laws in AWES Design

A both relevant and not obvious observation: we would expect in first that the practical scaling limit results from the wind load on the structure. So as a rigid wing increases in scale, as its “vulnerability to even slow impacts” would increase.
Resulting two possibilities:

  • The rigid wing doesn’t scale by 3D law in order to be light. It becomes more fragile and expensive as mentioned.

  • The rigid wing scales by 3D law in order to be robust enough. But in the same time a heavier and stronger wing generates more kinetic energy worsening the effects of even slow impacts.

1 Like

Questions of multi factorial parametric optimisation can now be handled very well by “AI” if the correct initial data and assumptions are given to the problem solver.

In this video you can see how I used grasshopper parametric development language and the galapagos genetic optimisation solver to test various geometry configurations for a torque tube… I set limits on the allowable distance and radius, I asked the solver to find the most likely optimal relationship within those limits to optimise for low line twist angle…
OK a very simple solution and the result was obvious … but given known metrics on wing performance, weight implication relationships, flexion, cost etc… A much better go of this kind of problem could optimise networks of wings to extract energy from an area at lowest cost & danger etc…
@PierreB I believe you were also looking into this recently

@Windy_Skies this is a way to …
Ask a vague question…like …

what are all the ways that you can imagine building a wing, and how large can you reasonably make that? How much would those cost in material? How much power would each of them be able to extract from the wind?

The question attempts to get a quick overview of the thing by asking about everything at once.


The 3 wee rigid wings on the recent Daisy tests were all repeatedly smashed into the ground… All still good enough to be used on the next test model

1 Like

Yes, you would ask those questions separately, as I said right after the part you quoted. An example question would be: what are possible scaling limits of soft single skin kites? A follow up question could be: how exactly would scaling limit x affect soft skin kites? And then, how could you mitigate the effects of x? Slow down to speed up, or break the problem up into its constituent parts to focus your attention and discussion.

I see, to paraphrase… I think you are saying slow down our investigation method breaking it into manageable chunks in order to speed up our solution discovery… Combining our knowledge at a later stage…
Yes it is a fine disciplined approach. Kinda how things have been done…

I was suggesting application of brut force exploration by massive sets of computer analysis attempting to apply known laws to find the performance results of evolved combinations of AWES components… (long sentence)
Kinda like this
Or this

Is it possible to scale by joining available single-skin wings (forming a rotor or a wing)?

1 Like

That’s sort of how the parachutes Storm Dunker makes work

This is the difference between a piece of cloth, a piece of glass, and a toy balloon dropping from a height onto the ground. For the piece of glass the energy of the impact does not travel through the material, for the balloon it does, spreading the force of the impact. Let’s equate the balloon to a tensairity structure.

I said something similar before: I can assume a large piece of cloth will not move much faster than the speed of the wind, I can aim for a tip speed ratio of, let’s say 5, for a rigid wing. To extract the same energy from the cloth, I would then need to make it much larger. The increased area and the higher drag coefficient of the cloth increases the wind load. That’s another possible scaling limit.

The most immediate and enduring scaling limit to me is the price and the environmental impact of the materials I use.

I believe the decision soft vs rigid vs [new tech in between, eg tensairity or similar] is not settled black and white yet. Yes - a mesh containing smaller soft kites can take a lot of punishment on the ground and will not break. Even so, during handling a line may catch something, and giant forces applied to the mesh. Also, using low L/D kites for generation leaves little room for depowering in a storm using L/D reduction (ie. lowering AoA on an airplane style kite)

This is easier avoidable with a more rigid wing. It has higher L/D (otherwise, whats the point) so it has larger depower abilities built in. The rigidness helps handling gusty storm winds without landing.

Now the rigid wing has some issues that have been pointed out; the forklift effect (?) and cubic mass scaling being some, along with higher minimum flying wind speeds (due to more mass).

There are no or few hard objects in the air. The forklift effect only regards handling on the ground. This should be manageable, though still complicating things for rigid wings. Cubic scaling of mass: manageable by optimizing the design, in particular by adding bridles and “antibridles” to support the wing. Optimizing the design might be sufficient to scale to relevant dimensions for utility scale AWE, around 20-30 meter wingspan. The benefits of rigid wings may be more important than this scaling limitation.

Clarifying that a ram-air air-beam such as that used in kites is essentially already a tensairity case, but no air-pump needed, and pressure increases crucially with velocity. Let anyone develop non-ram-air airbeams to test against comparable ram-air versions.

A speed limit to kites is given by tether drag growing exponentially with velocity, so fast wings suffer most, not just from their higher mass-to-power and poorer scaling.

Makani’s rule-of-thumb for equivalent power between soft and rigid wings is 10x area for the soft kite. That’s about right. Keep in mind the swept area of the fast moving wing v. the comparable soft wing is closer to 1 to 1, and the sky is very spacious, so the initial 10x figure is rather misleading by itself. For sure, there is no small solid kite able to function in kite sports like the “10x” wings.

SS kite scaling limit is mostly ground-handling. A grass surface is very desirable and perhaps blowing air under the kite to fluidize its handling could make km2 sizes possible.

Any chance you could clean up what you mean here for ease of readability?
Freel free to delete this message after

The ratio of swept/projected area between a hot fast sweeping kite and larger lumbering kite is closer to 1-to-1 than the 10x projected area comparison of equivalent power. The Sky is larger than land surface and ocean volume, so a 10x larger kite is still a speck in the sky. Its not like a road vehicle, where a 10x vehicle would take up 10 lanes and be unworkable.

Another scaling note from WP Square-Cube article lays out a scaling law for control and lift surface areas-

" Airbus A380: the lift and control surfaces (wings, rudders and elevators) are relatively big compared to the fuselage of the airplane. For example, taking a Boeing 737 and merely magnifying its dimensions to the size of an A380 would result in wings that are too small for the aircraft weight, because of the square–cube rule."

Take “fuselage” to mean “payload” or “power load” for our purposes. An interesting fact is the only reason that jumbo jets grew so large is that the unit-passenger scale remains constant, allowing more scaling than otherwise. This scaling advantage applies to kite aerotecture, but not pure AWE.

1 Like

That claim needs tests results to back it up, and to qualify it. Now it is an unsubstantiated claim stated confidently, which has no place here IMO.

The rest of your post also lacks links to back up the rest of your claims.

The Morse Sled has two ram-air airbeams with embedded whiskers, which is the essential criteria of tensairity as defined in AWES design by Rolf Luchsinger’s circle, and also in Wikipedia: “foundational splinting structure using inflated airbeams”.

Of course no one supplies a proof for every properly confident statement made online. One is best patient in requesting such proof as needed.

That is not an example of a tensairity beam as the air is free to escape and can therefore not resist the load.

This and things like it are tensairity examples and what I am referring to here. LEI kites lack the plank and wires and are therefore completely different:

1 Like

The air in a ram-air air-beam is not fully free to escape. Ram-air pressure prevents escape from the front and newer sled designs are now sealing the rear of the cell. Of course, parafoil cells are usually aft-sealed too, but lack the “splint” [WP].

The analogy is for those who can see the commonality. As for scaling, a valved ram-air airbeam is predicted to both out-scale an airbeam dependent on a flying air-pump system, as well as be simpler and cheaper. The empirical evidence is the preponderance of ram-air kites, but lack of self-pressurizing-aloft airbeam kites.

Old Forum predictions stand on the record, in this case that ram-air tensairity has both analytical and empirical advantages over a blower-system for low mass, low cost, endurance, progressive stiffness with velocity, and so on. Let it be noted that on this New Forum, others seem to predict the blower advantaged.

For clarity of intent in post 32 above Scaling Laws in AWES Design

I think what you have said is there is a rough scale law Makani used … saying a rigid wing flys roughly 10 x as fast, generates roughly 10 x more power per wing area and sweeps roughly 10 x more area
Is that correct?

1 Like


You are confusing Scaling Laws with Netiquette. I am close to quitting this Forum over the Netiquette distractions.

Of course a hot rigid wing sweeps more than a slow wing, and that factor can be weighed against the 10x projected area comparison in judging airspace use.

I hate to come to the defense of Santos, but all he’s saying is a ram-air beam holds its shape by producing internal air pressure that inflates the beam against surface tension of the cloth of the enclosure. It seems like such a simple, non-controversial statement that I don’t see why it would require any further backup or qualification.