High Altitude Wind Power – A Review

The first international conference on High Altitude Wind Power happened last November in Chico, CA.  That’s a sign this new technology is being taken seriously by academics.  Hence, the rest of us concerned about our energy needs should become acquainted with the topic.  For those unfamiliar with the concept, the idea is simply that winds are stronger and more reliable as you get further from the earth’s surface – peaking near 10,000 meters.  Harnessing high level winds could be an abundant energy source, but for the challenges of the technology.  Various methods of tapping the resource have been suggested.  A few prototypes have been built, and there are a couple of companies on the verge of releasing commercial products.

High Altitude Wind Resources

The winds up at high altitudes are surprisingly strong and steady – much better for power generation than the winds near the surface.  A recent paper by Cristina  Archer and Ken Caldeira looked carefully at National Centers for Environmental Prediction (NCEP) and the Department of Energy (DOE) data and produced high altitude global wind speed maps that factored in the fraction of the time the wind was blowing at speed.  The maps clearly show the jet stream bands in the northern and southern mid latitudes where the strongest and steadiest winds are located.

Wind_power density (kW/m2) that was exceeded 50%, 68%, and 95% of the time during 1979-2006 at 1,000 m (left) and 10,000 m (right) from the NCEP/DOE reanalyses. (Excerpt from [1].)

The energy available from wind scales as the cube of the wind velocity times the density of the air.  Peak energy density in the mid latitudes resides between 5,000 m and 10,000 m elevation.  The best winds are in the jet streams.  There is less advantage to go so high in the equatorial regions.  Surface winds are always attenuated near the ground by friction with the surface.  Just getting above the planetary boundary layer, typically 300 to 500 m, can easily triple the energy density available  compared to winds near the surface.

High Altitude Wind Machines

Various proposed methods for extracting energy from high altitude wind fit into three families of devices.  These are lighter-than-air devices, kite-like systems, and rotating propeller machines.  We will look at the organizations developing these devices and consider the pros and cons of each system.

Lighter than air machines

Magenn Power Inc. is taking orders for their first commercial product, a 100kW device with a price tag of $500,000.  The Magenn machines are large ridged dirigibles that rotate on their tethers.  The machines require lots of helium to stay aloft and they are designed to work between 500 – 1000 ft. elevation.

Magenn is targeting the device as an alternative to diesel generator sets for remote and emergency power applications.  Relatively low-level operation of this systems means that it can operate just about anywhere without regard to airspace restrictions.

Tethered Airfoils  (Kites)

The Europeans seem to like kite power methods.  The most advanced system is being developed by an Italian group.  The researchers at KiteGen have investigated systems that employ steerable kites in several configurations.  In the “yo-yo” configuration, power is generated by the extension of the tether during a fast-flying power stroke.  Then the airfoil is feathered while the tether is reeled back before again executing the power stroke.  A “carousel”  arrangement allows several kites to power a rotating vertical-shaft machine in a continuous manner.

KiteGen Stem 3 MW concept

A key advantage for the kite approach is in the efficient deployment of the active airfoil.  Ground-based windmills require enormous towers and foundations to support the large rotor blades, and only the tips see maximal wind speed and power production.  Kites, on the other hand, can access high winds with a light tether to transfer power from the air foil to the ground.  The entire air foil is subject to maximum wind speed.

A similar design called a Laddermill has been studied by researchers at Delft University in the Netherlands.  Work continues on developing adequate computer models of the kite system.  One of the primary challenges is being able to control the kite in changing wind conditions.  These control problems are formidable, and account for much of the theoretical work being done.

The kite systems are best suited for harnessing the winds below 1000 m elevation.  Machines targeting the upper atmosphere are going after yet another order of magnitude in power density and even steadier winds.  These machines are usually designed with spinning rotors.

Rotating Propeller Machines

Two California companies are investigating “flying electric generators” (FEG) as Sky Windpower calls them.  Bryan Roberts of Sky Windpower is one of the pioneers in the field with papers dating to the late 1970’s, and under his direction a two-rotor prototype was flown in 1980.  Present plans are for a four-rotor demonstration design capable of 240 kW and operation up to 4600 m.

For all power rotor designs, the tether must contain electrical conductors to transfer power from the rotors to the ground.  High voltage transmission is used to minimize the weight of the conductors.  The rotors are powered during launch so the tethered craft can be flown to altitude under power.

The Joby Energy modular turbine concept

The other California start-up is Joby Energy, investigating a modular flying turbine concept.  The company has built several small prototype devices that demonstrate flight control of their platform concept.

Finally, there is Makani Power, Inc., another California start-up that has received $10 million funding from Google.  They have a very capable team, but little information is available about what they plan to do.

Technical Challenges

Most casual observers of high altitude wind power easily point out the obvious problems.  Lightning, air traffic interference, failures causing machines to fall out of the air, and the weight of miles of cable are all frequent criticisms.  Fortunately, these are all engineering challenges that can be tackled with standard engineering practices.  There is no new physics that we need to learn to develop this resource.  Aerodynamics, structural design and materials strength, power electronics, electrical generation, and control theory, are all well-developed disciplines that will be applied to the problem.

The driving motivation is the promise of high returns on investment.  Estimates of the Energy Return of Energy Invested (EROEI) are around 100, better than most fossil fuel resources but without the CO2 problems.  Cost estimates for the electricity produced range from  2 to 5 cents per kW H.  With our insatiable demand for cheap energy, and with a widely dispersed potential solution a few miles overhead, someone will bring it down to earth and make some money along the way – the race is on.


[1] Global Assessment of High-Altitude Wind Power, Cristina L. Archer and Ken Caldeira, Energies 2009, 2(2), 307-319.

[2] Harnessing High Altitude Wind Power, B. W. Roberts, D. H. Shepard, K. Caldeira, M. E. Cannon, D. G. Eccles,  A. J. Grenier, and J. F. Freidin,  IEEE Trans. on Energy Conversion, 2007, 22(1) pp. 136-144.

[3] The Laddermill – Innovative Wind Energy from High Altitudes in Holland and Australia, B. Lansdorp and P. Williams, Windpower, 2006.

[4] Control of tethered airfoils for a new class of wind energy generator, M. Canale, L. Fagiano, M. Ippolito, M. Milanese, 45th. IEEE Conference on Decision and Control, San Diego (CA), USA, 2006.

[5] Design and Construction of a 4 KW Groundstation for the Laddermill, B. Lansdorp and W. J. Ockels, IASTED EuroPES 2007.

[6] KiteGen project: control as key technology for a quantum leap in wind energy generators, M. Canale, L. Fagiano, M. Milanese, M. Ippolito, Proc. of American Control Conference, New York 2007.

[7] Power Kites for Wind Energy Generation – Fast Predictive Control of Tethered Airfoils, M. Canale, L. Fagiano, M. Milanese, IEEE Control Systems Magazine (12) 2007.

[8] Comparison of Two Mathematical Models of the Kite for Laddermill Sail Simulation, A. R. Podgaets and W. J. Ockels, Proc. of the World Cong. on Engineering and Computer Science 2007.

[9] Long-Term Laddermill Modeling for Site Selection, B. Lansdorp, R. Ruiterkamp, P. Williams, and W. Ockels, AIAA Modeling and Simulation Technologies Conference 2008.

[10] High-altitude wind power generation for renewable energy cheaper than oil, L. Fagiano, M. Milanese, and D. Piga, EU Sustainable Development  Conf. 2009.

[11] High altitude wind power: an era of abundance?, Ugo Bardi, The Oil Drum, July 6, 2009.


  1. Good article, Gary. You should offer it for publication in mainstream magazines. Salvational new energy is a hot topic.

    However, I feel that you have cherry-picked objections that can be answered by engineers (someday in the future, using existing science, or something). You have not addressed the vulnerability to deliberate attack, the long deployment time before significant power can be produced nor the reliance on a complexity of orderly, civilized interaction that may increasingly collapse as near-term energy constraints break down the Google-friendly mythos of continual techno-progress.

    The optimism for this technology leaves me skeptical for the same reasons as that surrounding the Bloom Energy fuel cell . Google, which uses a lot of energy, is understandably interested in both technologies, and may actually be able to use them profitably. But as nasty as coal plants or nuke plants or hydro dams are, in various ways, they produce way more power than even the largest kite arrays Google will see in the next decade. While their servers might famously run on “alternative power”, avoiding both soaring mainstream energy costs and possible interruptions in their delivery, the cities of America will never replace liquid fuels, coal and nuclear with kite energy.

    What I see is a niche producer, with individual units having very high cost, which requires high-tech inputs and which is easily broken or perturbed by increasingly uncertain conditions (e.g. weather/ politics / streams of technical innovation and production).

    A further objection, only applicable if the technology were to really become heavily deployed, is that slowing down air currents above the Earth may have unintended consequences. We have already weirded out the natural systems upon which our ancestors have depended throughout the development of civilization, and we are already going to pay. So I instinctively pull back from thoroughly “harnessing” currents that are perhaps analogous to the blood or lymphatic system of a human body.

    All this said, I thank you for introducing me to an interesting concept, and doing it very well. My preference is that we learn to use less energy, rather than spinning ever more rare-earth minerals into complex systems competing with the planet herself. But perhaps there is room to experiment with smaller, less dreadfully efficient models that mountain-dwellers could use to keep a few lights on. Worth a try, anyway.

  2. Chris, Thank you for your thoughtful comment. Your main point that we should figure out how to do with less – I cannot agree more with. Ultimately, the worse possible outcome for High Altitude Wind, is that it actually might work! With more “free” energy we expand our populations further until the limits of the natural system are stretched even further.

    Your other objections I think are less problematic. Decentralized systems are less vulnerable to disruption than large single installations like we have now. If it is easier to get energy from the sky rather than by drilling for oil, a lot of capital can be funneled at HAWP projects. Deployment could be more rapid than you might imagine. This has a lot to do with the fact that there really is nothing new here technically. It is really only an engineering problem. Once GE decides this is where the money is to be made – it can happen overnight.

    The climate change objections were addressed in the Archer and Caldeira paper. The present trajectory for CO2 emissions I find much more worrisome.

  3. I think Chaos Theory will have some application here, forcing engineering to be more robust and difficult.

    Also almost any wind project has to be tethered to the Grid or requires some form of electrical storage. Matching supply to demand will remain difficult, especially if we decentralize.

    I feel solar and fuel cells have greater immediate potential.

    Nice article, very interesting technologies.

    1. Paxton, Indeed, engineering kites to fly in chaotic winds is a challenge. However, if you look at some of the papers from the Italian group, you will discover that they are making real progress. The Archer and Caldeira paper looks at how much storage is required to get various base-load levels for the example of NYC. This is certainly an issue. One potential “distributed storage” solution could be in the form of electric vehicle batteries as EV’s become more important for transportation. As far as solar goes – it is easier so I’m all for it. But average power density for solar is about 0.2kW/m2 in Eugene. Compare to wind at 9km up – blowing 68% of the time with energy density at least 1.5kW/m2, 50%of the time >4kW/m2. Fuel cells are a clever way to burn natural gas – not really an energy/climate solution though certainly and important technology.

  4. What do you guys think of ‘wave energy’ systems, meant to get energy from ocean waves, such s the one planned for deployment off of Reedsport, OR????

    1. Hi Robert,
      I’m sure there is great potential with ocean waves, but I have my reservations about keeping anything functioning that sits in salt water and suffers the relentless beating of Pacific storms. You are limited to regions near ocean coasts, and although the energy density in waves can be large, I think you are going to run out of coast line before you can generate a significant fraction of the nation’s power that way. Typical wave energy density is 20kW/m or which perhaps 25% is harvest-able as electricity – 5kW/m. For the entire Oregon coast thats approx. 500 km x 5 MW/km = 2.5 GW production possible if entire resource used. Present average Oregon power consumption is about 5.5 GW. Waves could be a significant fraction, but can’t scale indefinitely.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s