FairIP index           

Reducing Hard-Automation Risks

AWE flygen kiteplane startups must persevere across the hard-automation "Valley of Death."
In this R & D hell, failure modes emerge in complex dynamical systems as self-organized attractors, often as one-off events never again seen after a simple fix. No company can sustain a relentless loss of custom prototypes, much less serious injuries or fatalities, just to discover critical flukes. Survival depends on resorting as required to elegant low-complexity fail-soft methods.

For this reason experimental flygen kiteplanes are suspended clear of surface impact and allowed to fail repeatedly without serious consequence, just as Hargrave, over a century ago, set up masts to suspend his kite experiments without mishap. So it is that Makani is seen flying tethered foils from a ladder-truck. KiteLab Ilwaco hangs all kinds of tests off a spiderweb of arrow-shot lines festooning old growth forest. But towers are expensive & not a true solution to fly thousands of feet high.

Solutions like lifter kites and tails may not seem sexy, but they do provide high baseline performance to build on. Ballistic chutes are sometimes proposed as a flygen kiteplane fail-safe. These heavy expensive explosive devices even pose their own risks. They require periodic inspection and repacking. Why lug all that packed-away soft wing in the sky when it could be working? A parawing always inflated over the flygen is lighter and cheaper than the ballistic option, and probably safer.

LTA and towers also mitigate control criticality, but with added disadvantages, particularly higher capital-cost. Low-tech kite methods are the superior intermediate AWE solution in many cases. For example, a particular challenge in AWE is scaling up by closely packed element arrays. It may be that kite trains, arches, and meshes critically enable required performance ahead of flocking automation.

AWE testing need not await on a controls team to deliver a functioning autopilot. Testing of subsystem flight dynamics readily proceeds by passive control and human piloting. Passive methods nicely replicate basic active-control flight patterns, like looping and figure-eights, by stable oscillation flight dynamics. For example, tuned Dutch-Roll "Instability" self-flies eights. VTOL operation by passive methods is also elegant; as wind rises, a lifter kite can predictably raise a flygen kiteplane and land it gently on a dying wind. Such a method can eliminate the requirement for variable pitch symmetrical foil turbines.

Should a company's advanced-control R & D track be long delayed, or even fail totally, low-complexity passive-control methods still enable a system to enter market on time and post revenue. Minimal investment in these techniques protects against overoptimistic commitment to fast on-budget automation development. Airworthiness certification and low cost insurability may depend on retaining low-tech kite-lift as a fail-soft foundation. As automation reliability is reached, then the "training wheels" can come off.

KiteLab Group offers hard-automation AWE players a low cost "Plan B," the best of traditional and novel rigging for passive lifting and fail-soft testing of components. Endurance trial flying services of small prototypes, backed by kite-methods, are offered in high availability (~70%) winds on Pacific Coast (KiteLab Ilwaco).

FairIP/CoopIP                                   ~Dave Santos            June 9,2010        M1621

Comment and development of this topic will be occurring here.        All, send notes!

Terms and aspects:   


Related links:

Commentary is welcome:

  • v