Issues with rigid wings and possible solutions (reply to Dave Santos)

Hi Dave, if you’re reading, I won’t be posting in the Yahoo forum, so I’ll reply here. You’re more than welcome to reply here or in the Yahoo forum.

Dave Santos wrote:

If only a rigid wing were unbreakable, weightless, could grow to an infinite size, and be free, I think we would agree that it would be superior to a soft wing because of the higher L/D ratio and other benefits you haven’t mentioned. In reality it doesn’t need to be those things, it only needs to focus on some of them because of the huge advantages they have, like the much greater W/m^2 number, which I understand is important, and the much greater lifetime of the wing, provided it doesn’t crash.

A productive discussion I think would have a narrow focus and think in terms of solutions, not impossibilities, we can’t after all imagine everything, so we can’t say a thing is impossible, unless it breaks the laws of physics.

An issue you point out is, if a rigid wing crashes, it breaks. What might then be ways to make sure it doesn’t crash, or if it does it doesn’t break? What are methods used in aircraft today that lower the chances of crashes and mitigate the effect of crashes?


This is an interesting topic for sure.

  • Soft kite durability is an issue still unsolved. I dont think a power plant where the kites need replacement every few months is likely to be economic viable, nor perhaps environmentally friendly
  • Soft kites will never approach a rigid wing’s fidelity in actuator precision and perhaps also the magnitude of actuated forces. Because a soft kite will change shape. Soft kites need to be passively stable («need») because of this. History has shown that evolving soft kite design is hard and time consuming.

Rigid kites truck vs racecar this is true and a big design limitation making rigid power kite design fundamentally different than aircraft design. Understanding this may probably initiated the recent paper on high Cl biplane wings. Quite a few of the rigid wing actors show elements of high Cl design if you look closely already.

A sailplane can be made to have a L/D ratio of, say 15. By applying advanced aerodynamic analysis, an increase in this number to eg. 20 would directly translate to a faster plane with longer range. Any added weight translates to larger flying speeds. This is manageable.

For kites the story is a bit different. You can have the same L/D 15 wing, then add the tether to get a wing with total L/D 7. Now increase the L/D to 20, and your total L/D is still around 7. Tether drag is dominating kite drag.

You can alleviate this by having a larger wing or higher Cl or both, in which case the wing will easily break the tether unless carefully managed.

The tether drag may be approximted by

F_d = 1/4 (1/2 rho Cd length diameter v^2)

This may be interpreted as tether drag being 1/4 of drag compared to all of the tether moving at the kite’s speed. Though we have the 1/4 factor, we still have F_d proportional to length and v squared.

How high can a rigid wing kite fly without tether drag dominating? You’d have to do the exact numbers for your own rig, but 200 meter is proven (?) fine, and 2 km is too much tether for most I believe.

The tether diameter scales with area not diameter, so for very large kites, you can go higher without as large issues. If your design uses many tethers, the issue is compounded.

Noe back to the truck vs racecar argument. If you instead use a huge soft kite with L/D of only 4, the velocity of the kite is much smaller. In this way you get rid if the bulk if the v^2 part of the tether drag equation. In short you get the same pull wasting less energy on tether drag

Having said this, I am not convinced that this argument would out rigid wing designs in the infeasible bracket. It just pans out as stuff you need to design your rig for, and in the end, limitations in performance.


This. ^

One should not forget kiteswarm concepts like you sketched yourself.

I was trying to describe my view of the racecar vs truck argument. There are many ways around this, requiring AWE to advance a little more than single wing Yoyo. The twin kite yoyo on a Y line is one example of a possible way forward. There are plenty more, I wont go into detail, perhaps they are not all described yet

Soft kite developers propose recycling their polymers by standard processes. Its rigid kiteplanes that have more troublesome recycling issues, especially when they crash and burn.

“Fidelity in actuator precision” is not a requirement of passive-dynamic operation, like kPower’s looping foils. The only required actuation is engaging the load and a kite-killer. AWES that fly actuators are burdened by comparison.

Agreed, “evolving soft kite design is hard and time consuming”, but its the best fun the greatest designers know; what they were born to do. Its fatalistic to think these folks cannot succeed in beating critical rigid wing statistics, at least until graphene airframes and perfected flight automation emerges.

1 Like

Thanks for replying @kitefreak.

Do you have any ideas on this?

As this isn’t my research interest I won’t have many ideas on it, so I asked you and others to comment.

I do have some ideas though. We can for example multiply the probabilities of the failures to get a system that very rarely catastrophically fails. To use made-up numbers, say an automated system fails every 2000 hours, you can add a safety mechanism that works 95 percent of the time, and another that works 95 percent of the time, to get a system that on average can be expected to only fail every 2000 x 20 x 20 = 800,000 hours, or 91 years.

Like with most things in this day and age we don’t have to start from scratch.

Holding up hot fragile kiteplanes under a pilot-lifter prevents crashing, no active flight automation needed. In lulls, the kiteplane calms down and lands soft, while the pilot usually continues to fly, then relaunch the plane when wind recovers.

kPower has done many variations. including flygen. Just let the kiteplane go nuts and harvest the pumping at a groundgen via a low-stretch PTO line. Use an elastic tether to buffer disturbance of the pilot-lifter, it will return excess surge energy to the kiteplane in recovery phases of patterns.


What is PTO?..,.,

Power Take-Off, after the aux power shaft in the rear of a farm tractor

Am I right that most crashes would occur during take off and landing? What are some ways to make that safer?

During flight I assume you could make the requirement to never fly below n meters and have a parachute on board for example so there crashing becomes almost impossible?

Ceterum censeo safety can be an afterthought for any large scale awes. Just secure a perimeter.
The case is different for things like the kitewinder small scale product where that is not an option.
And if a system is very unsafe it’s also unreliable, which makes it unviable.

The Golden Age of Kites established runaway kites as the primary public hazard. Paris’ power was cut in one instance, and a train stopped in another. In our time, airspace safety is a function of aircraft mass and velocity, and a fast massive kiteplane can travel a long way in runaway mode.

Ceterum censeo, it only suffices to secure a perimeter for many-connected (topologically stable) soft-kite networks.

Does anyone have links to AWES architectural scaling barrier predictions better than linked posts from the Old Forum?

Rod and Tallak in a related topic want something better. Sadly, there do not seem to be any in-depth academic studies. There are dozens of rigid-wing AWE start-ups somehow presuming scaling barriers will not prevent their architectures from going big.

Heuristic kPower warnings are not seen compelling nor rigorous enough, even if eventually proven correct.

Another idea I had. I have no opinion on its feasibility:

Don’t make one monolithic wing. Assemble your wing from shorter sections that you attach to each other using connectors that you can release on command.

So after your software failed and your plane goes below its nominal path and can’t recover, and you released the tether from the plane and after your first and back-up parachutes failed to deploy, you can detach the wing sections from each other to on impact with the ground only impart to the wing sections their own kinetic energy. The individual wing sections also then might start acting something like parachutes fluttering to the ground. For kicks you could also add mini-parachutes to the wing sections.

Other benefits this might have: easier and cheaper construction, and less risk in testing. If you also add bridles to each individual wing section you can also reduce strength and thus weight of the wing sections.

1 Like


We call this a “metawing” concept, a “wing of wings”, for several years now. Its a quite common notion, from Mothra to MegaFly. Also, assembly of meta-structure in the sky is long anticipated. The units are called “kixels”.

Moderation here does not allow mention of where these AWE ideas were greatly elaborated, unless a link is provided, page-by-page.

This is a serious scaling method that pilots fully believe in. Automation? Later.

There has been some study on how to fly individual winglets from a rotor after they have broken off…

1 Like

Or another very similar idea is to incorporate failure points in your wing, and forego that release on command part of the idea. So you make your wing (or blade) from multiple section that you then connect to each other (possibly allowing some flexing). The connection points are the failure points.

Random quote:

A bit analogous to rip-stop nylon.

If I were to try to make a rotary system (like @Rodread, @Kitewinder, or @someAWE_cb), this might be how I would construct my first prototypes of a long slender blade.

My blades are held in the hub by rubber bands. The idea is that the rotor throws them off when spinning too fast.

Hasn’t happened yet as RPM control via generator/esc has not failed yet and wind speed has not exceeded 10m/s.

There are plenty of well documented purely mechanical overspeed protection systems for small wind turbines available. I guess my next design would be spring loaded. When the centrifugal force is bigger than spring force the blades move out - while changing their angle of attack.



I learned that many windmills are stall controlled… a simple proven mechanism for non-pitchable blades… (for TRPT rigs)

1 Like