otherwise wasted

August 20, 2006 2:47 AM

if heat and light can be turned into usable energy, why don't we launch some collectors to orbit the sun and gather truck loads of the stuff that is otherwise dumped into space, then beam it home via microwave?

i hear that our light > usable energy conversion isn't very efficient. What about heat?

How much is lost when you convert to microwave and then back?

generally: why haven't we done this yet?

posted by Tryptophan-5ht to science & nature (25 comments total)

Haven't played SimCity 2000, huh?
posted by Mikey-San at 3:00 AM on August 20, 2006 [4 favorites]

Thermophotovoltaics
Inverse-square law
posted by Pinback at 3:04 AM on August 20, 2006

According to the social energy page on weatherquestions.com (section headed "What About Solar Collectors in Space?"):

[Compared to land-based solar collection, g]reater amounts of sunlight can be collected in outer space, where a satellite can be kept in the sun continuously, and clouds are not a factor. But the roughly factor of three increase in energy intensity is offset by inefficiencies in transmitting that energy to the ground. If it was transmitted as microwave energy, much of the energy would be lost because the large antenna required would not be able to concentrate the energy into a very small area on the ground to be received. Indeed, the antenna required to receive the energy on the ground would be much larger than just putting solar collectors on the ground. There would also be environmental concerns about transmitting huge amounts of energy over such a large area, and the potential dangers of the space transmitter missing the receiving antenna on the ground. Not the least of the problems is the very high cost of launching anything into space.

One other concern in attempting to develop such an energy retrieval system is outlined in "Die Another Day" — such a weapon would invariably attract nutcase megalomaniacs like Gustav Graves, for whom a solar cannon would be so much cosmic catnip.
posted by rob511 at 3:16 AM on August 20, 2006 [1 favorite]

Umm, that should be "According to the solar energy page..."
posted by rob511 at 3:18 AM on August 20, 2006

the economics of solar power generation are why it's not in more widespread use. solar panels cost a lot of money to make. it's not like trying to get them into space is going to make them any cheaper.

also there is plenty of sunlight that makes it to earth's surface, even with nighttime and clouds - if we covered a few counties worth of surface area in utah, we could generate the entire united states' electrical demand. there's no shortage of photons; it's just that non-renewable sources are (at the moment) much much cheaper.
posted by sergeant sandwich at 3:45 AM on August 20, 2006

With the recent talk about space elevators, would it be feasible to put a solar storage collection sat into space, and run a giant "power cord" down to earth rather then mess with microwave transmission?
And if so, if someone tugged on the powercord, would it pull it back to earth?
posted by JonnyRotten at 5:07 AM on August 20, 2006

Yeah, enough sunlight already comes to Earth than to power everything we've got (in fact, there's an awful lot more). The problem with collecting it is cost -- and sending solar panels into space would cost a hell of a lot more than just putting them on everyone's houses or whatever.
posted by reklaw at 5:36 AM on August 20, 2006

People have thought about this, but there are real issues with how to safely get the energy down to Earth. Don't get too close to that microwave beam.
posted by caddis at 6:16 AM on August 20, 2006

The engineering problems are nontrivial. For one thing, getting that much mass into orbit is not within our capabilities right now.

For another, as caddis hints, the microwave beam which brings the power back down to earth can also function as a deathray that can toast cities.

The real answer to your question is that it doesn't make economic sense. If you have a certain amount of money to invest, you can produce a lot more power by using that money to build coal-fired plants than by trying to build solar satellites. And you'll have it sooner, and the regulatory burden will be much lower.
posted by Steven C. Den Beste at 7:14 AM on August 20, 2006

Solar power satellite.
posted by Wet Spot at 7:34 AM on August 20, 2006

DenBeste is right on. "Non trivial" is a generous understatement.

There are at least two major problems with this... engineering and economics.

FWIW, we do this on a supremely small scale. Satellites are routinely powered by solar cells and routinely send miniscule amounts of focused power back to earth in the form of radio signals. While efficient for communications, it's not enough energy to perceptible warm a microbe.

Economically, look at the life cycle cost and contrast it with presently available alternatives. Just a thumbnail sketch of such a scheme would put it out of sane bounds in a microsecond.

There is abundant 'free' energy available here. That's not the problem. Conversion, transmission, storage, polution, infrastructure.... those are the issues. They could use some imagination and speculation.
posted by FauxScot at 7:47 AM on August 20, 2006

Well.. DenBeste is right on about the non-trivial part, but.. Putting aside the fact that we don't have any idea how much mass "that" is, I suspect that we can put enough mass into orbit to collect a significant amount of energy. It isn't collected with mass after all, it is collected with surface area.

I like sergeant sandwich's version - plenty of photons reach the ground. You just have to do the math, cost per sq.ft divided by energy per sq.ft. You probably don't reach 10x more energy collected per square foot in space - after considering transmission efficiency, I'm sure you don't - I'd expect the costs are 100x higher, or more.

It may not always be that way, a space elevator (completely paid for exclusive of and potential energy generation revenue) would radically change the calculation.
posted by Chuckles at 9:39 AM on August 20, 2006

Yeah, let's leave aside that it would cost way more to do what you're suggesting than it would to blanket the world in power plants....

You're making a distinction between heat and light that doesn't make sense in outer space. "Heat" is the average energy of molecules. On earth, heat can be conducted through the air, but in outer space, heat is a result of light energy coming from the sun, which excites molecules in a given object.

The problem with solar power on earth is that there isn't that much energy in the photons that make it here. They'd be a lot more concentrated closer to the sun, but then getting that energy back here...anything that could transmit that much energy (sufficient to make the whole enterprise worthwhile) would be sending back so much power that if something went wrong, as people say, it could cook a city. Not good.

It is also worth noting that any solar array near to the sun would have to contend with temperatures that could cook a lot of electronics.
posted by Dasein at 9:52 AM on August 20, 2006

Furthermore, photovoltaics require maintenance. Which in this case means signing on NASA (or subcontract to cosmonauts). Pretty much nothing but communications and science make economic sense in space (and people will fight pretty hard over whether science makes sense).
posted by nanojath at 2:23 PM on August 20, 2006

I think most of this question is pretty well answered already, but one point I wanted to comment on is the relative efficiency of photovoltaics (light - energy) vs thermoelectric power generation (heat - energy) and also the differences between them. Conveniently, these are the two main thrusts of my research, but I will try my best to refrain from too weighty a tome.

First off: photovoltaics (a/k/a solar cells). The most efficient (and also by far most expensive) solar cells made to date are ~40% efficient. This means that on Earth, they can produce ~400W/m^2. Unfortunately, they are quite cost prohibitive at this time. Typical solar cells are ~10-20% efficient but at a much lower cost. The figure I've usually seen is that we'd need to cover about the area of the state of Nevada to meet all of the US's energy needs.

One seemingly-clever (but fraught with technological difficulties) solution would be to increase the intensity of light hitting the solar cell. You propose sending them into space, but using concentrators to focus more light on the cells is a lot more efficient. Unfortunately, with all this light comes a lot of heat and this can damage the solar cells. Also, increasing the concentration of photons hitting the surface increases the current flowing through the cell rather than the voltage. Some of the vital components of efficient solar cells generally cannot handle very large currents. (If anyone really cares about this in detail, email me.)

As for thermoelectrics, the most efficient currently available are less than 15% efficient (depending on whose numbers you believe, the most efficient might be << 15%). They are generally only used in niche applications where other power sources are impractical. For example, deep space probes cannot use solar because the intensity drops quite quickly with distance from the sun. (See Pinback's link on Inverse Square Law). Instead, a radioisotope thermal generator (RTG) is typically used. In this case, a smallish chunk of plutonium (or something else which gives off heat for many years due to radioactive decay) is put near a thermoelectric generator, which produces electrical energy from the thermal gradient between the Pu and the coldness of space.
posted by JMOZ at 3:55 PM on August 20, 2006

As a practical matter, with current technology extrapolated for the forseeable future, the right way to build a solar satellite is to pass entirely on direct conversion of sunlight to electricity. Instead you build a huge mirror, which can be amazingly flimsy. It's curved, and it concentrates light/heat onto a plain, ordinary boiler. Behind the mirror, in the shadow, is a big radiator. Steam from the boiler is run through a standard steam turbine and runs a normal every day generator.

The expensive part, in terms of mass, is that radiator. The mirror can be flimsy but the radiator cannot be, and it has to be large enough to shed as much heat on a steady-state basis as is being captured by the mirror.

And there's going to be a lot of fluid of some kind in this system, water or something else that can be boiled.

It's going to have to be in geosynchronous orbit, which means the boost cost to get everything up there is a lot greater than for LEO. If you want it to capture enough energy to make any difference at all, it's going to have to be immense. My general thumbnail guideline is that for any new power source to make any difference at all it has to scale up to at least 1% of the power currently used by the US.

Which is to say 1% of 3.5 terawatts, or 35 gigawatts. If you figure the entire system (including the microwave downlink) at 35% efficient (and I doubt it would be that good) then it means your mirror needs to capture 100 gigawatts and your radiator needs to shed 65 gigawatts off into space.

Anyone care to do the "black box radiation" calculation to determine how much radiation surface would be needed running at what temperature in order to radiate that much energy out into space?

And note that the radiator necessarily has to run at high pressure, which means that it is going to have to be made of materials a lot more robust than mylar.

That's why I say that the amount of mass which would need to be boosted into orbit for this would be huge by comparison to anything we've ever launched before.

In fact, the amount of energy that would have to be expended in rocket engines and in equipment manufacture would be huge. How long before the system even paid back its manufacturing cost in energy, let alone paid for itself in bucks?
posted by Steven C. Den Beste at 5:03 PM on August 20, 2006

If, instead, you decide to do it with photovoltaics, then you have a major infrastructure problem. Fact is, there just isn't very much manufacturing capacity out ther producing photovoltaics, and scaling that up isn't necessarily all that easy. Using our 35 gigawatt target again, and borrowing JMOZ's efficiency number of 15%, then again the first thing you have to think about is cooling.

The photovoltaics are going to reflect off a fair amount of the light hitting them, but they're going to absorb a lot, too. You're going to need a cooling system capable of handling somewhere between 50 and 100 gigawatts of waste heat; if you don't have it, the system is going to overheat and the photovoltaics will die.

You have to have a way to move the heat around, and that pretty much means a circulating fluid, which has to operate under pressure. The fluid and the pipes are going to be heavy, very heavy, and so is the radiator.

But getting back to my original point, if you do the math and start with 15% efficiency, and assume that light in space is 3 times light on the surface of the earth (a number I just pulled out of a hat) i.e. 3 kilowatts per square meter and shoot for 35 gigawatts yield then unless I screwed up my math what you get is 450 watts per square meter and thus 78 million square meters of photovoltaics in geosynchronous orbit, plus the cooling system for them.

That gives you your 35 gigawatts before losses in the microwave downlink, which will be at least 50%, so you have to actually double that. Call it 155 million square meters.

If we're producing a thousand square meters of photovoltaics per year now I'd be surprised. I'd bet it's a lot less than that. You're going to have to invest a huge amount of money in very specialized factories to scale up the manufacturing capacity in order to produce that much, at colossal expense.

It just doesn't make economic sense. Coal-fired plants are a lot easier. Nuclear plants should also be, except for the preposterous regulatory burden now which makes them essentially impossible to build in the US.
posted by Steven C. Den Beste at 5:15 PM on August 20, 2006

also there is plenty of sunlight that makes it to earth's surface, even with nighttime and clouds - if we covered a few counties worth of surface area in utah, we could generate the entire united states' electrical demand. there's no shortage of photons; it's just that non-renewable sources are (at the moment) much much cheaper.

No, we couldn't. The problem is that electricity has to be produced at the exact same time as it is consumed. There's no way to store it for later.

Mechanisms exist which can store electrical energy in small quantities, but no reasonable approach exists which could store during the day and then regenerate at night the amount of electrical energy that this country uses during hours of darkness. No approach that makes economic sense comes within six orders of magnitude of making any sense.

California averages about 25 gigawatts of electrical usage during hours of darkness (plus or minus about 20% as a function of seasonal changes)

Figuring average 14 hours of time per day in which the proposed solar cell farm would not be generating power because it would be not be in sufficiently bright direct sunlight then it means California would need to store and regenerate again 14 hours * 3600 s/h * 25 gigajoules/s == 1.26 * 10^15 joules per day.

1260 million million joules. Care to tell me how to do that? And how much would that equipment cost?

Here's what you're not going to use: you're not going to do this with flywheels. You're not going to do it with capacitors. You're not going to do it with storage batteries. You're not going to do it by making hydrogen and then burning it with fuel cells. You're not going to do it by pumping water up hill and then running it through a turbine.

None of those can be scaled up enough just to deal with California's night time electrical usage, let alone what the rest of the country uses.

You'd need 14 hours at something like 200 gigawatts. In other words, about 10^16 joules every night.

Actually, you'd need more than that in winter since the nights are longer and it's colder. No way. Not going to happen.
posted by Steven C. Den Beste at 5:35 PM on August 20, 2006

You wouldn't orbit cheap solar cells, that would be plain silly. So, I think you might go for an efficiency number of 25% or more. Still ridiculous.

I expect getting rid of the waste heat to keep the solar cells running at peak efficiency is very very hard when there are no air molecules to carry it away. Anyone have a useful Theta junction to ambient for a typical space based cell to compare with typical earth based numbers.

Anyway, orbiting a big mylar mirror that's focused on some massive water tank on the ground in Nevada is a pretty interesting idea. Very light. On the other hand, the control systems required would be.. Challenging.

As for electricity storage, peak demand happens to occur while the sun is shining, and the peaks are around 50% higher than the valleys. Since we are only talking about 1% of the market, storage isn't particularly relevant.

Hmm.. Interesting, there is a midday demand dip in winter.
Here is an askme with a bunch of links to demand graphs - What is a kilowatt hour?
posted by Chuckles at 6:37 PM on August 20, 2006

Steven, you actually just might do it with making hydrogen and burning that with fuel cells.

Electricity represents only about 1/3 of present energy use. At least as much again is fuel for transport, and the rest is mostly heat for buildings.

If the transport fleet becomes hydrogen-powered, and if enough generation capacity has grown up to provide enough hydrogen to serve the needs of the transport fleet, then we're already well into the ballpark of the amount of hydrogen required to service off-peak electric generation needs.

For hydrogen production, as long as there's enough average capacity to keep the service stations supplied, instantaneous generation capacity control doesn't matter. It's fine to have generators that only work in the daytime, or when the wind blows, as long as there are enough of them overall.

In fact, as Amory Lovins has been proposing for a long time, off-peak electricity generation could actually be done using the transport fleet itself. Each vehicle would need something like 50kW of fuel cell capacity; just one million vehicles parked overnight with their fuel cells plugged into a reticulated hydrogen supply and a reticulated electricity grid gives you twice your required 25GW(e). And there are many more than a million vehicles in California.

Lovins's other interesting idea is about how to get that hydrogen supply kick-started. He's proposed moving toward heating buildings with the waste heat from natural-gas-to-hydrogen reformers instead of straight natural-gas furnaces. The natural-gas reticulation infrastructure is already in place, and the amount of energy currently used to heat buildings is already in the same ballpark as the amount used to run cars; it's not a huge stretch to see buildings supplying their own associated car fleets with hydrogen for fuel and for grid electricity generation.
posted by flabdablet at 9:01 PM on August 20, 2006

Chuckles, peak electrical usage in New York in winter is at 6PM, when the sun is down.

Flabdablet, you are not going to generate 200 gigawatts with fuel cells.

The problem with all these ideas is scale. It's not that they cannot be made to work small, it's that they cannot plausibly be scaled up enough to really make any substantial difference. That's why I use the 1% criterion as a sanity check. If you can't explain to me how to scale alternative-energy-source-du-jour up to 35 gigawatts, then there's no point in even talking about aesdj because it's nothing more than a technical curiousity.
posted by Steven C. Den Beste at 9:14 PM on August 20, 2006

Flabdablet, you are not going to generate 200 gigawatts with fuel cells.

According to this, one "gallon" of hydrogen gas at one atmosphere pressure contains 40 BTUs. (When, oh when, is America going to switch to MKS like normal people?)

So units, units: a gallon is 3.8 liters. A BTU is 1055 joules. So at one atmosphere, a cubic meter of hydrogen (1000 liters) is 11 million joules.

200 gigawatts for 14 hours is 10^16 joules or 900 million cubic meters of hydrogen.

In other words, you'd have to produce 900 million cubic meters of hydrogen every day while the sun was up, and then burn it all that night.

At 100 atmospheres that's a cube 210 meters on a side. Doesn't sound too bad, does it? Except that we've never built anything gas-tight which is that large that can hold that kind of pressure. If instead you use a series of cubes which are ten meters on a side, you'd need 9,000 of them.

If you store it at 1 atmosphere it's a cube nearly a kilometer on a side. Stored in 10 meter cubes, you'd need 900,000 of them.

This isn't for an energy source; it's for an energy storage system to buffer power generation so that it matches usage. This expense is in addition to the cost of all the photocells or whatever other land-based solar power generation you are trying to use.

Again, the economics don't make any sense. Why bother with something like this when you can build a perfectly decent coal-fired electrical generation plant for a lot less money?
posted by Steven C. Den Beste at 9:32 PM on August 20, 2006

According to this...
posted by Steven C. Den Beste at 9:33 PM on August 20, 2006

Flabdablet, you are not going to generate 200 gigawatts with fuel cells.

Well, if your fuel cells come one per car at 50kW each, that's only four million cars. Given that the number of cars in the USA is of the order of 100 million, it sounds doable to me.

As for storage requirements, here's a snippet from Amory Lovins's "Hypercars®, Hydrogen, and the Automotive Transition":

The supposed difficulty of onboard storage. This has been solved by commercially available filament-wound carbon-fibre tanks lined with an aluminized polyester bladder ... The tanks on the market in 2003 are extremely rugged and safe, have no external high-pressure components, provide normal driving range in efficient platforms, and contain ~8±12 mass percent hydrogen when filled to 345 (US-approved) or 690 (German-approved) bar pressure. Further technical progress will doubtless occur but is not required.

So this is fuel tanks for vehicles, commercially available for the last three years, and they're designed to work at 690 bar (a bar is about an atmosphere) which is over six times the highest pressure you envisioned. Assuming bulk hydrogen storage runs at about the same pressure as car fuel tanks would, that's taken your 210 metre cube down to about 110 metres, or your 9000 ten-metre cubes down to tanker size. And you won't convince me there are less than 9000 fuel tankers operating in the USA right now.

I can't see the scaling required to reach a fully-renewable energy economy as a show-stopper. It's certainly much closer to reality than your rather pessimistic six orders of magnitude.

See, that's the thing about renewable energy stuff: it does scale, because there's way more energy available than we need, and all we have to scale is the infrastructure; and we can use existing mass-production techniques and attitudes to do that, ending up with vast numbers of moderately sized plant instead of small numbers of heroic-sized plant. The big numbers just aren't that scary.

Yes, it's presently much cheaper per kilowatt just to build more coal-fired plants. But I think you'll find that the community is not going to let coal-fired generators continue to socialize their external costs indefnitely.
posted by flabdablet at 10:03 PM on August 20, 2006

peak electrical usage in New York in winter is at 6PM

Ya, but the winter peak isn't that high - I think we are arguing semantics here.

Anyway, the idea of big orbiting focusable mirrors is sounding more and more reasonable to me :P (which means it's time for bed, I think). After all, there isn't any reason to turn the beams off at night. That is, unless you are more concerned with the light pollution than you are with the byproducts of burning coal.
posted by Chuckles at 10:40 PM on August 20, 2006