hkeithhenson at gmail.com
Mon Oct 11 22:10:12 UTC 2010
This is off NDA so I can go into detail.
For a few years, I was working on a way to reduce the cost of
space-based solar power to the point it could displace fossil fuels.
That's two cents or less per kWh, which is half the price of electric
power from coal, and low enough that (off peak) it can be used to make
synthetic hydrocarbon transport fuels for about a dollar a gallon.
The rough economic analysis is based on a ten-year repayment of
capital cost. Run 80,000 hours in ten years the return is $800 per kW
per penny payment for a kWh. For power satellites, assuming 5kg/kW,
$100 per kg lifted to GEO and about 1/3 of the cost going to
transport, you get the required $1600/kW for 2 cents per kWh.
With the help of Jordin Kare, Howard Davidson, Ron Clark, Spike Jones
and others, by last January I had a proposal that looked like it would
reach $100/kg cost to GEO. The general approach was discussed in an
article in The Oil Drum about a year ago. It proposes huge lasers to
get the average exhaust velocity up to the mission velocity. This
gives a mass ratio to LEO of about 3 and a throughput to GEO upwards
of 100 t per hour.
Late last year Howard became aware of a project an old friend of his,
Ed Kelly, was working on. Ed is best known as a principal with
Transmeta, a company that developed low-power processors some years
ago. Howard introduced me to Ed. I have spent a lot of time going
over Ed's spreadsheets and other details since last January.
In the post-analysis, the reason ordinary ground solar power is so
expensive is the huge amount of materials that are needed because
solar energy is so dilute. (Wind has the same problem.) Ed's
approach, which he named StratoSolar, was to reduce the mass from
hundreds of kg per kW to a few tens of kg by moving the solar
concentrator into the stratosphere as a large, lightweight, buoyant
This has significant advantages over being on the ground.
There are no clouds at 20 km. The winds are light and steady and the
low air density reduces the force on the structure. Because the
primary concentrator can be pointed directly toward the sun, it gives
close to full power whenever the sun is above the horizon. (Rough
pointing--one to two degrees--can be done with combinations of
thrusters, aerodynamic fins and reaction motors, fine pointing by
stepper motors moving the mirror segments.)
They work as far north as Stockholm.
The concentrated sunlight gets to the ground via a hollow light pipe
lined with highly reflective prismatic plastic. Preliminary
optimization for kg/kW leads to a 30-meter diameter light pipe with
less than 10% loss. A larger pipe has lower losses but uses more
total material per kW.
Because the mass is dependent on the pipe diameter and the power
capacity on the area, StratoSolar plants optimize in large sizes,
around 1 GW.
That means the primary collector is a bit over 2 km in diameter and
100-200 meters thick. That gives plenty of room for gasbags to offset
While the concentrator has neutral buoyancy, the light pipe has a lot
of excess buoyancy. If you just think about it as a force diagram,
the buoyancy needs to be 3-4 times the wind force to keep the angle
the light pipe makes with the ground within 15-20 degrees of vertical.
The materials required—aluminum, plastic, steel wire, and hydrogen
(for buoyancy)—are all inexpensive and do not need to be processed to
tighter specifications than the norm for commercial products.
The sunlight is absorbed and converted to heat at the bottom. The
heat is used to run an ordinary, 45%-60%-efficient, one or two stage
power plant. About half the heat during the day is used to heat a
solid heat thermal storage medium. This will provide enough stored
heat to run the plant overnight.
Graphite is a good choice, but any high temperature solid would work.
Cowper blast furnace stoves (regenerators, dating from 1837) produce
air as hot as 1400 deg C, just about the limit for turbine inlet
While stoves for this application are big (typically 70,000 cubic
meters), they are dead simple and should cost well under $100 million
for a GW plant. That cost adds 1/8 of a cent per kWh to the cost of
power. This is less than 1/10th the cost of any other proposed
Our rough estimate for the cost is around $1.2 B per GW, or $1200 per
kW. Using the above ten-year payback, the cost to generate power
should be around 1.5 cents per kWh.
It will take building a few to learn how to manufacture them and get
accurate cost numbers. However, if this is close, it will solve the
long-term energy problems and get the human race off fossil fuels by
simply under pricing them.
Like any other large project, there are a million details. We have
given thought to such topics as ozone, lightning, hydrogen fires,
thunderstorms, icing, interaction with aircraft, high wind loads,
aerodynamic shrouds, UV damage, turbine throttling, maintenance
access, and manufacturing (to name those I can think of at the
Of course, with only a few people working on it, the models are not
very detailed yet.
You can find a PowerPoint presentation if you Google for StratoSolar.
There is a Web site, but the content is not yet up. For those with a
serious interest, there is a 50-page technical tutorial available.
Since the 1970s, US politicians have given lip service to "National
Energy Self-sufficiency." The US has failed to achieve anything,
largely because nobody had a good idea of how to make it work at the
same or lower cost than importing oil. This method might not work
either. However, it passes first-order physics and economics analysis
and seems to deserve serious further study.
More information about the extropy-chat