Sunday, June 28, 2009

Power from Space

Just recently the California utility company PG&E announced a program to beam energy from space, where it would be collected and beamed back to Earth. The program has a target of 200 MW of power by 2016.
The project is expected to cost around $2 billion, which will mainly go towards the R&D of the base station and launching the satellites. SolarEn CEO Gary Spirnak has complete confidence in the concept and the company’s ability to develop this system. In fact, he projects that they will be able to generate 1.2 to 4.8 gigawatts of power at a price that is comparable to other forms of renewable energy. PG&E is also committed to the idea and has entered into a 15 year contract with SolarEn to produce enough power for 250,000 homes.

This isn’t the only SSP under development now - Japanese Aerospace Exploration Agency (JAXA) is also working on a similar system, but instead of radio waves, they will transmit power via laser beam. Both companies ideas seem a little far-fetched, but if either of them succeed, it could mean huge things for renewable energy generation.
Now this whole idea is not new, back in my archives I have more than a foot of shelf space devoted to reports on the Satellite Power System (SPS) that was developed by NASA and DOE back in 1978. And, for your interest, I thought I would pull out the Concept Development volume and review the system that was planned back in those pioneering days. The initial work required that the concept be fleshed out, and to do this Boeing Aerospace and Rockwell International were under contract to the Marshall Space Flight Center. Out of that work came a whole series of reports that makes up this part of my bookshelf. For simplicity I will just quote from the Concept summary today.

Back then the goal was to generate a system that could be fielded by 2000, then more than 20 years in the future. The ideas built on an original concept that Peter Glaser had suggested in an article in Science, back in 1968. At the time they concluded that microwaving the energy back to Earth would be more efficient than using lasers to transmit the power.
The following target guidelines and assumptions were built into those plans:

The system would be operational in 2000.
At that time the system would add two 5 GW satellite systems a year to ultimately reach a total of 300 GW..
The ground receiving antennas (rectennas) are sized to receive a 5 GW feed.
The satellites would be placed in geosynchronous orbit.
The systems would operate at a frequency of 2.45 GHz.
The intensity of the microwave signal is not to exceed 23 mW/sq. cm at the center and 1 mW/sq. cm at the edge of the rectenna.
System life is 30 years.
Needed (but unavailable technology) would be needed by 1990 to get the system up in time.
All materials will come from the Earth, and there will be no launch failures.
Costs will be derived in 1997 dollars.
The arrays on the satellite will be built on a graphite composite, and two options were considered, a straight single crystal silicon PV array nd a gallium-aluminum- arsenide solar cell with a concentration ratio of 2. An efficiency of 7% for the conversion efficiency of the array was assumed. Thus to get a 5 GW output the solar panels must be large enough to capture some 70 GW of solar power.
The gallium arsenide option would result in a solar blanket area of 26.52 sq km a reflector area of 53 sq km and a total platform area of 55 sq. km. With a single silicon crystal concept gives a platform size of 54 sq km, roughly equivalent.
A microwave antenna some 1 km in diameter will be used to transmit the power back to Earth.
It is assumed that a satellite could be constructed every 6-months, and depending on type this would require a space-borne construction crew that would be in the range of 555 to 715 individuals.
The cells are assumed to heat in the sun, and this improves efficiency so that at 125 deg C the cells are assumed to reach 18.2% efficiency. The gallium cell design was anticipated to cost some $71 a square meter.
The single crystal silicon would use a 50 micron sheet of solar cells, with a borosilicate glass cover. The efficiency is assumed to be 17% though it will require some monitoring and laser annealing. Costs for this alternative were calculated at $35 a sq. m. Small thrusters are to be mounted on the array to allow it to move away from approaching space debris.
The power collected from the array will be transmitted to the microwave transmission station with the transmitting system broken down into 7220 sub-arrays, each 10 m on a side. The sub-arrays are phased to give a coherent beam focused on the center of the rectenna.
In sizing the system, it was recognized that it would need to be a modified design to accommodate a maximum heat build up in the array, and a maximum rate (23 mW/sq cm) at which power could be concentrated through the ionosphere. (Heating the ionosphere above this level could cause signal interference and efficiency loss).
To convert the DC-RF power some 101,552 tubes would be needed. The power beam would operate in the 2400 to 2500 MHz frequency band
It was recognized in the report that there would be some RFI effects due to re-scattering of the radiation. The rectenna would have an area of 78.5 sq km. be open faced, to allow air passage, and will need a set of rectifying diodes to convert the energy back into phased electrical power. Some 7% of the energy is assumed to be lost due to heat. The microwave system efficiency is assumed to be 63%.

The reduced cost of the single crystal silicon array is made up in the greater mass that would be required for that option (51 million kg, as opposed to 34 million kg for the gallium arsenide option.

Because of the location of the satellites in Geosynchronous orbit (GEO) the material for the construction would be lifted first to Low Earth Orbit (LEO) and then moved to GEO. The lift out of the gravity well would require a vehicle, anticipated to weigh 11,000 tons with a payload to LEO of 424 tons. Given the number of workers required to build a station, special transportation vehicles would need to be built, and each would have a carrying capacity of 75 people to get them into LEO. To move out to GEO a special transit vehicle would be needed, and the one designed would carry up to 160 passengers at a time.
The cargo would be carried out to GEO using special carriers powered electrically using ioon bombardment thrusters, with cryogenic argon as the propellant. (So that the round trip to GEO would take around 180 days. For personnel a faster vehicle would shorten the trip to a few hours )

The logistics of the support operation are also described. Bear in mind that each array will need 10^11 solar cells, 10^5 klystrons, and 10^10 dipoles. The system would need about a million tons of hydrocarbon propellants to move between locations. For 2 5 GW stations the vehicles would need 375 surface to LEO launches (225 with the gallium arsenide)

Just for set-up alone it is anticipated that it would take 6 months to build the LEO base, after the vehicles become available. 3 months would then be needed to built the first transfer vehicle, and another 3 months for the next two. It would take a year to build and configure the entire fleet needed. It is expected to take 9 months to build the GEO base., thus it would only be 2 years after the start that work could begin on the first SPS station. Once built it would require some 5 to 20 people would be required per SPS for maintenance.

There are many more details in the reports, but this will give you some sense of scale of the original project. It will be interesting to see how the current plan compares with this older one.