All things green are getting a thorough look with oil poised to bust through $100/barrel as Russia, Venezuela, Iraq, Iran, Nigeria and other major exporters view high prices as a time to reap profits and consolidate political control rather than feed the goose that is laying the golden eggs--i.e. invest. One hope for cheap clean energy (energy independence is at our fingertips if we decide to widely deploy coal-to-liquid technology used by SASOL in South Africa and is estimated to cost $35/barrel by Wikipedia) is space based solar power (SBSP). But Taylor Dinerman gets to the key roadblock to SBSP in his Space Review article today, "The chicken and the egg: RLVs and space-based solar power":

The SBSP Study Group universally acknowledged that a necessary pre-requisite for the technical and economic viability of SBSP was inexpensive and reliable access to orbit....

Phase one proposes a strategy that will “Develop new, fully-reusable two-stage, rocket-powered space access systems (aerospaceplanes) for passengers and cargo transport.” The mission is to “Transport passengers and cargo with ‘aircraft-like’ safety and operability.” The report claims that for such systems the TRL is 6–9 [6=space tested systems, 9=debugged commercial vehicle] for a vehicle with a gross weight of 1400 tonnes with the capability of delivering a bit more than 11 tonnes of payload to LEO.

Is this the Falcon IX? At $55 million/launch or $2500/lb, it's still too expensive to be a chicken. Rand thinks that a contract for ten heavy GEO deliveries would do much much better than that using established vehicles.

If it's not the Falcon IX, then what is it? And why? If it costs $15 billion to develop the A380, wouldn't it cost that for an RLV? And what good is it if there aren't eggs aplenty to carry to GEO? Suppose the hypothetical RLV can do 40% cheaper than existing vehicles per pound (like the A380 vs. 737). It would need to carry 10 million pounds to break even. To achieve a 20% cost reduction, it would need to carry 30 million pounds ($75B in lift costs).

Two problems: A) a 20% cost reduction is not enough to set off a demand frenzy--it would just lay an egg; B) $75B in lift costs is 25 years of annual demand.

A credible commitment to want eggs delivered into space whatever the price (e.g., $15 billion/year like everyone's favorite subsidy poster child, ethanol) would spawn several competing providers to get those eggs cheaply into space. $15 billion spent on a technologically successful space plane, on the other hand would a) result in an asset that won't be able to repay its bond debt and b) even totally depreciated would not cross essential price points to make SBSP commercially viable.

And why spend a total of $300 billion to conjure $15 billion/year in demand forever? Vision. $15 billion/year is about 0.1% of GDP. It's affordable if the prize is big enough. In 1976, airmail cost $0.06/pound delivered. In 1926, it was $3.00/pound. If you adjust for inflation, that factor of 50 decrease goes to a factor of 150. That is, a modest demand would likely result in a 10.5% annual takedown rate in the price of spaceflight for the next 50 years. Taking $2500/lb as the starting price, we might end up at $20/lb. or about $3/kwh average price of orbital energy imparted to payload delivered to GEO.

Electric energy prices dropped from $4/kwh in 1892 (in 1992 dollars) to about 9 cents in 1967 so I wouldn't be surprised to see the price to impart energy for space launches drop from the current marginal price of $0.60/kwh or so (fixed costs dominate) for kerosene to less than $0.06 which is around the wholesale price of coal generated electricity today leaving room for another factor of 50 improvement for the subsequent 50 years from 2057-2107. Regenerative braking on a space elevator comes to mind, but that will probably be about as far from what we will have in 2107 as the Wright Flyer is from the A380.

If the price to orbit in 2057 is as cheap as the price to New Zealand today, we might expect to see some feature films shot there, four million people living there and an addition to space-produced GDP of $1 trillion/year (about .4% of solar system GDP in 2057). That produces a good fraction of the tax revenue to justify the investment without any vision or desire for clean energy.

Hi Sam,
I don't want to argue for the economic viability of space solar power on the near term but I can answer this question:

"If it costs $15 billion to develop the A380, wouldn't it cost that for an RLV?"

Only if you are talking about building lots of those RLVs. I've heard this sort of question before, e.g. it takes X billion for a company to bring out a new plane or car so how can you build a new rocket vehicle for so much less? The difference is that those companies are setting up to make many such vehicles. The $15B pays for the creation of the whole infrastructure required to mass produce the A380, i.e. new factory buildings, assembly lines, training of thousands of workers, new tooling, coordination of production of components in many countries, certification, etc., etc.

RLVs are, at least initially, more comparable to the cost to those companies of building prototypes, which they do for a fraction of their mass production setup cost. For example, Eclipse Aviation spent several years and somewhere around a billion dollars to create the assembly lines for the Eclipse 500. Yet, virtually overnight they recently built and flew an Eclipse Concept Jet (www.eclipseconceptjet.com).

Same for cars. Yes, it takes GM several billion dollars to bring out a new model but a few guys working on weekends can build a stock car or dragster that will far out perform that GM sedan.

Maybe it will cost many billions to develop an RLV. (I don't think it will but I have no proof yet to the contrary.) But that will be due to the R in R&D, not for the same reasons that it costs many billions to develop a new mass market airplane or car.

(Rutan has said he hopes to build 50 or so SS2s but he can probably still do that within a small production framework.)

And lets not forget the cost of meeting all those FAA and its Euro counterpart mandated requirements.

I don't see such great promise in near term space solar power.
One could say that if you put the cells in space, they are in total about 4 as efficient as on earth (because of no night and weather constraints).

But you can offset that quite a lot by just adding 3 new solar cells on earth.

Space solar power doesn't get economical until the space launch for one cell costs less than 3 additional cells. And that ain't going to happen anytime soon.

(This all of course had drastic oversimplifications...)

Sam, how do you get to $20/lb to orbit? Using the mature air transportation analogy, which is 3 x fuel cost, I get $45/lb. This assumes rocket fuel cost equals imputed jet fuel cost of $.45/lb and a 3% payload fraction. If a ssto is used - a closer fit to aeroplane operations - then less payload results.

It's hard to see costs to leo dropping below $70/lb.

You seem to assume that lower energy costs results in proportionatley lower total costs, but that doesn't work: Suppose jet fuel costs fell in half today, would the cost of flying drop in half too?
What would happen if the cost of fuel fell to zero?

Seer,
While I think it's a long time before we're going to see prices below $1000/lb, let alone $100/lb, there are some things you aren't considering (regarding the theoretical minimum price):

Most importantly, LOX in bulk is a lot cheaper than aviation fuel. I'm not sure about current prices, but I've heard as cheap as a $0.01-0.05/lb of LOX in large quantities. The mixture ratios of most LOX/Hydrocarbon engines is between 2.5: to as high as 4:1. Which means your effective cost per pound of propellant is going to be lower than the cost of *fuel* for a jet aircraft.

That said...at this point arguments about that are about as relevant as how many angels can dance on the head of a pin....

~Jon

Seer: I get $20/lb to orbit by starting at $2500/lb and applying a 10.5% cost reduction each year for 50 years which is what happend to airmail prices after the airmail acts from 1926-1976.

If we have a space elevator that uses electric energy and regenerative braking on the way down, that might do it. Or perhaps a tethered stratospheric rail gun for cargo launches to again use electricity for launch. Both those methods have a much higher payload fraction. Maybe the second stage will be broken down for scrap, or be edible and become part of payload. Maybe the second stage will be a module that can be added to the space station as structure or living area. Maybe the price of jet fuel will drop to $10/barrel making your math come out to $4.50/lb. I don't know how people achieve 10% cost reductions for airplanes or chips, but if you stoke demand enough, prices will fall steadily after rising initially due to exhausting capacity.

Jon: You saw ethanol plants sprout like corn. What is so fanciful about orbital rockets doing the same thing?

Concorde cost about 1.1 billion UK pounds to develop plus 654 million pounds for 16 production planes in 1977. The exchange rate was all over the place (between $1.60 and $2.60 for the pound around that time). Call it $2/lb. That's about $3.5B or $12B in today's dollars. Would you agree to a probability of success of a US two-stage reusable research development project at 80% or less?

Sam, Jon, my point is that beyond a certain point energy costs don't matter. If it did, then a methane lox vehicle would be substantially lower cost than a kerosene lox vehicle, yet they are very similar in terms of isp, T/W and hence payload.

As I said, what happens if the price of jet fuel falls towards zero, does the cost of flying also tend toward zero?

Let's bring in more engineering and less handwaving.

Kerosene has about 40 MJ/kg. An Atlas V has about 100 t of that, and a payload of about 1.5 t to GSO.
This yields 40 MJ/kg * 100/1.5 = 3 GJ/kg fuel energy spent for placing something in GSO. We disregard the hydrogen in the upper stage, and the lox in both stages, and the rare metals and aluminum used in construction.

Solar power in space is optimistically about 200 W/kg. So it takes 3e9 J / 200 W = 15 million seconds to produce back the fuel energy, or about half a year.
We didn't use any efficiency values, just compared orders of magnitude, whether it might be more efficient to just burn the kerosene instead of using it to launch a power satellite. This shows that at least in theory the powersat might make sense.