Thursday, October 22, 2009

A 30-year old opinion on Space Power Systems

Just over 30 years ago there was a young science magazine called Omni. It ran from October 1978 through the fall of 1995. In the first April edition (1979) it ran three stories – one, by Mr. Stine, on the Satellite Power System (SPS), one on solar energy in general by Mr. Pohl, and one on an invention by a Mr. Ovshinsky.

I was sufficiently moved by these articles to write a response to the magazine, and I just uncovered that letter. For an amusing perspective on that time I, and since the first two are still around, I thought I would post the thoughts of a rather younger Heading Out (who wasn’t at the time). I will break it into two parts, one of which deals with my thoughts on solar power and the second, which is more of a general comment at the time on energy in general will be posted on Monday.

I am deeply concerned by the tenor of these article since their net promise is that, by implication the same sort of promise is now being made for solar power as was made, less than 20-years ago, by the nuclear industry: “Just give us our heads and you’ll have free energy for life.”

With your indulgence I would like to review the energy scene as I see it, starting with solar energy and working the opposite direction to Mr. Pohl.

Let us begin with the SPS system. If I may quote the article “ . . .the SPS system is generally feasible from the point of view of both technology and economics . . .it is technically feasible to convert that energy to microwaves and to transmit it to the Earth with no negative effects on the environment . . .this would lead to a small (5-Gigawatt or 5-million kW) SPS pilot plant operating in low earth orbit by 1987, giving us the technical and economic answers that we would need to begin construction of a full SPS system with 50 10-GW SPS units in geosynchronous orbit by the year 2000, supplying most of the projected electrical needs of the entire North American continent.”

Let’s do a little arithmetic on this prediction to understand what this means. (My facts, unless stated otherwise, are from the DOE/NASA SPS documentation.)

Firstly a 5 GW satellite would weigh up to 50,000 tons and have an area of 100 km2. Assuming that the shuttle can carry 32 tons means 1,500 plus trips by weight. Assuming an average thickness of 0.1 cm for the structure would give a volume of 100,000 cu. m. The shuttle hold is 91 cu. m. in size, so that, on a volume basis – with no voids –we’d still need about 1,100 trips. Accepting bulking and some larger components, I hope that you will agree that 1,600 trips would not be an out-of-line assumption.

To put the first SPS system in orbit by 1987 would therefore require 200 trips/year starting (in 1979). To then create an additional 50 10-GW stations would require, in the following 13 years, 160,000 trips or 11,500 trips a year.

In regard to the cost for the solar cells Heliotronics have predicted (Electronics, Oct 26, 1978) a cost of $0.25/watt in 8 years, at a 10% level of efficiency – about that considered by the DOE/NASA paper. That level of cost is included in the DOE/NASA reference design which predicts a capital cost of $2,500/kW installed, it also would require an initial R&D phase of $45 billion. (My quibble with these figures are that they are 1977-78 dollars admit no inflation and assume an interest rate of 6%).

The study also considers two other pertinent factors. The first is land for the rectennas. This has been evaluated at an average of 80 sq miles/site; land which could become permanently barred to people and hazardous to the health of all wildlife therein. To quote the report in regard to a full SPS system “Only a small number of sites, relative to population, could be located in either the Northeast or Mid-Atlantic states. Those sites that could be identified were in fairly mountainous areas.” Which means that our scenic wilderness will become hazardous to our health.

The second point which, in its way is more worrisome is the critical materials report. Two systems were evaluated – the silicon system and the gallium arsenide system. The report lists 25 commodities required to provide 2 5-GW satellites a year. It highlights those for which problems might arise. For the silicon option two items – mercury and tungsten are considered critical. The mercury need for 168 tons is a problem since it would require a 10% increase in domestic production. Tungsten at 1,220 tons would require a 25% increase in domestic production. Although gallium requirements are 7 tons, against a current annual production of 8 tons, this is not considered a problem in the silicon option.

The gallium arsenide option, however, has 6 critical items. The most severe of these is the Gallium of which 2,186 tons is required. This is still set against 8 tons of domestic and 7 tons of foreign production per year. It should also consider, as the report does, that the total domestic reserve is only 2,000 tons, with a world reserve of 112,000 tons. A similar requirement for 2,356 tons of arsenic per year meets a similar problem with an annual production of around 23 tons being predicated by the 2000.

The net result of the data to date, I believe, is to show that any realistic use of the SPS system is at least 30-50 years away, as the DOE/NASA study predicts, and will probably only become economic when the material is supplied either from a lunar or asteroidal source? (Incidentally won’t a 100 sq. km. surface act as a solar sail?)

Coming rather rapidly therefore down to Earth where we unfortunately lose a lot of the sun’s power, let us examine the current status of the solar industry. Unfortunately, the arguments about gallium and arsenic still hold, so as foar as the gallium arsenide cell is concerned (with 25% efficiency) the material is unavailable, and so we must, pro forma, accept a 10-15% cell efficiency.

Living in the mid-West I will use Missouri as my initial location for analysis. St. Louis receives about 1,000 Btu/sq ft on a horizontal surface on an average day in March. With a peak demand of about 15 GW this would require, at a 10% efficiency level, an area of 441 sq. miles of collectors or adding 30% for walkways, roads etc about 600 sq. miles.

I could take the analogy further and show that the 76 Quad energy demand of the United States could be satisfied by a collector area of 69,700 sq miles – the area of the state of Missouri – but I’d rather reduce the scale and talk about my back yard.

In January my energy consumption was 5,530 kWh, or an average of 7 kW/hour. Oman and Gelzer, in the Energy Technology Handbook tell me that I need a peak supply of 5 times average demand to get me through the bad spells. My solar cell requirement therefore should be about a 35 kW system. At $0.25/watt (which I would remind you is about 1/30th of current cost) the cells would cost some $8,750. Accepting Oman and Gelzer’s figures for the costs of installation (but multiplying by the necessary factors) I get:

Structure . . . . . . . . . . . . . . . . . . . . .$2,000
Fuel Cell . . . . . . . . . . . . . . . . . . . . . $7,500 ( 35kW peak at $400/kW)
Electrolyzer . . . . . . . . . . . . . . . . . .$2,450 (35 kW peak at $70/kW)
Metal hydride . . . . . . . . . . . . . . . . .$4,000 (4,000 lb FeTi at $0.50 a lb and structure)
Hydrogen and Oxygen storage tanks $2,500 ($300/1,000 scf and controls)
Electric power conditioning/controls$500
Installation . . . . . . . . . . . . . . . . . .$1,000
Total . . . . . . . . . . . . . . . . . . . . .$28,200

I have accepted the arguments that I should use metal hydride rather than batteries on the basis of their projected cost effectiveness. The area of collector that I would require is about 5,000 sq. ft. – which is where I came in because it’s bigger than my house and backyard, and since 8% interest on $28,000 is $2,240 and my total electricity bill last year was $903, I don’t think I, or probably anyone living North of me, can really afford solar electricity just yet, even at $0.25 a watt.

Incidentally in regard to solar heating, Forbes quotes the house built for Mark Hyman in Waltham, Mass where it cost $24,000 for heating only. I don’t think I can afford that option, even with the current tax credits.

If then solar power cannot be used North of me, what are the situations in the South? The general consensus would appear to be that solar systems are best developed in the desert. To which I would like to draw the following points:
Firstly for a 5-MW system we are still talking of somewhere in the region of a 100 sq. mile unit. The equipment for that installation would be installed by heavy construction would be installed by heavy construction equipment that would destroy the surface integrity of the desert lands. We know, from experience, for, example from the Badlands of the Dakotas, that when this occurs the underlying surface can e readily eroded by rain and winds.

The solar collectors will be set up at an angle to the surface and this will act as a wind trap, bringing the wind down across the exposed sand. The result will be to lift the sand into the air, and one can readily anticipate that it will fall across the surface of the collectors.

In a recent paper Hawthorne has shown that such grit, falling from a height of 1 m. will cut the reflectance of a surface by 25% within 10 seconds. One can therefore predict two things; firstly that the collector surfaces will gather considerable amounts of dust which will have to be cleaned off. Secondly that the surfaces will become rapidly scratched loosing their ability to transmit energy because of the impact of the sand particles. Since we are talking about somewhere around 8 million collectors, keeping these clean, and the energy efficient levels up and keeping them protected, will in itself e a major undertaking.

One other consideration is that we have, already, discounted the use of gallium arsenide as a collector. It has the advantage that it can be operated at high temperatures, and since the location of these collectors in the desert, where they get hot, as Mr. Pohl has pointed out this will mean that they need to be cooled, and thus we come back to the ubiquitous mater cooling towers. These symbols of current power station design will thus again be required in the desert.

So that was my opinion, and some data (the real reason for putting up the post) from some 30-years ago – I’ll have a comment on reality and how some of this turned out, after Monday’s post. There is a lot that I got wrong, but I'll come back to that.


  1. Brilliant analysis and critique Heading Out, proving that certain fundamentals are timeless.

  2. Thanks, Morgan:
    Unfortunately some of the arguments about solar powered satellites and power transmission seem to have been lost over the decades, and we have the case of PG & E who are working with Solaren to resurrect the idea.

  3. Interesting. OMNI was an excellent magazine. G . Harry Stine popularized model rocketry, so it's no surprise that he'd be a...booster of SPS as well. Fred Pohl's been publishing SF since the 1930s.

    One of TOD pieces on peak metals had a comment from someone claiming to be an industry insider, saying that gallium production could be jacked up anytime miners got the bug, simply by adding $50 million in kit to existing aluminum works. He sounded like he knew his stuff.

  4. Gallium is a by-product from the production of aluminum. As long as you are producing to the aluminum market, getting the gallium is not that expensive, but if all of a sudden you had to increase overall production just to get the gallium, then it gets really expensive, and you have all that aluminum to get rid of.

  5. I'm not sure where you got the 100 mi² for 5 MW; that looks about 3 orders of magnitude off to me.

    Interesting letter.  I was an old L5 Society member back in the day, and recall that the Glaser group proposed a new heavy-lift vehicle for its SPS concept, and O'Neill suggested getting most of the material from the Moon instead.  Rectenna sizes wouldn't necessarily have been a big deal, as with larger transmitter arrays or shorter wavelengths you could have shrunk them almost arbitrarily.  The big problem would be keeping the peak beam power density low enough to avoid thermal effects in the atmosphere.  Bigger transmitters help there, as the power distribution over the rectenna could be made flatter with a sharper cutoff at the edge.

    The east coast has plenty of ocean, and if nobody would be bothered by a wind farm in it, they wouldn't be bothered by a floating rectenna system either.

    All of this is idle musing, though.  The unity of vision needed to get a project this big done just doesn't exist in the West even if we could carry it off technically.  This may be a good thing, because our ornery nature has kept us from putting too many eggs in one basket more than once (the financial crisis and our reliance on petroleum being two obvious exceptions).

  6. Don't despair too much, one of my colleagues has a funded graduate student who is working on the best ways of excavating the lunar regolith.

    Sometimes, as long as there is still some preparation work going on, a sudden shock can come to the system and projects can get restarted.