Tuesday, April 24, 2007

Jack and the Beanstalk

I do not much care for beanstalks, for reasons that have nothing to do with the blood of an Englishman.

For those of you who are not space geeks, I should briefly explain what a beanstalk is. A satellite in the right orbit, about 40,000 kilometers up, is "geostationary" - it orbits the Earth in 24 hours, and therefore seems to float at one spot in the sky. If you were aboard a space station in that orbit, and you had a long enough and strong enough wire, you could reel it out till the other end reached the ground. You could then use it as a rope ladder to go up and down with no need of a rocket or anything resembling a rocket. This is a beanstalk.

An alternative term is elevator, for the sensible reason that no one wants to climb the stairs that far. The idea, so far as I know, was first proposed by Arthur C. Clarke - about as good a source as space speculation can have; he was also the discoverer of the geostationary orbit, now used by most communications satellites. (Which is why your satellite TV dish can point in one direction instead of slewing around all the time.)

It's quite a brilliant concept, really, and no great surprise that in the last few years it has become very popular both in science fiction and nonfiction space discussion. So what do I have against beanstalks?

They do have a few technical hitches. No currently known material is strong enough for a wire hanging down from 40,000 km to support its own weight without snapping, let alone carry a load. (However, there are laboratory hints that such materials may be possible.) Reeling the wire down and anchoring it at the ground end could be ... challenging. Devils also tend to lurk in the secondary details, such as operating the elevator cars that would run up and down along the cable or cables.

All of these may well be solvable; a beanstalk would still be a truly stupendous civil engineering project - essentially a railroad suspension bridge, 40,000 kilometers long, turned on-end. Given that an ordinary light-rail line costs at least $10 million/mile, a beanstalk to orbit would be perhaps a trillion-dollar investment.

And that is my beef with beanstalks. At some future date, space traffic might be so heavy that the enormous investment could be justified - the Interstate Highway System probably did not cost all that much less in present-day dollars. The fashionable trend in SF and space speculation, however, is to lo-ball the cost down to as little as $10 billion, and see beanstalks not as an eventual successor to shuttle-like reusable orbiters but as a near-future substitute for them.

On one level this is understandable. The Shuttle has turned out to be a huge disappointment (for reasons I'll discuss in a future post), and the technical challenges in building a truly robust orbital transport are enormous. Yet to pin our hopes for space access instead on a technology never demonstrated on any scale, and perhaps not possible at all, strikes me as a bit like putting our faith in magic beans.


Nyrath the nearly wise said...

Just to be pedantic, the space elevator was first proposed by Konstantin Tsiolkovsky in 1895. Later, Sir Arthur C. Clarke and Charles Sheffield demonstrated telepathy when they independently and simultaneously wrote novels about space elevators. These were "The Fountains of Paradise" and "The Web Between the Worlds" respectively.

The space elevator has a huge capital cost, and is a titanic undertaking, but current boosters have an outrageous cost on the order of thousands of dollars per kilogram of payload delivered to low earth orbit and about $20,000 per kilogram of payload to geosynchronous orbit.

A space elevator could do it for about $100 per kilogram or less.


Rick said...

Nothing pedantic - I had no idea that Tsiolkovsky proposed it (in 1895 no less), though given what a visionary he was it's not really a surprise.

I take figures for cheap space access with a Saturn V payload's worth of salt; space advocates have a long history of lo-balling costs, going back at least to von Braun. But the real challenge is bootstrapping the traffic volume to make any new system pay for its up-front cost.

This is the toughest problem I see even for the pretty conservative TSTO (two stages to orbit) approach that Markus Bauer and I were discussing at sfconsim. Suppose that development cost of a system is $50 billion, but that once built it is entirely free (not even energy cost).

If the required return on development cost is 10 percent (less than in your linked article, though my development cost is higher), the system has to earn $5 billion/year to pay itself off. If traffic is 1000 tons/year - several times current demand - the cost is $5 million/ton, $5000/kg. Lower than current cost even to low orbit, though not drastically so.

Drum up 5000 tons/year of payloads and cost in this idealized case falls to $1 million/ton. Even in messier real-world cases, generally the more you do it the cheaper it gets. But is the potential demand really there?

I'll be talking a lot more about this in future posts!

Anonymous said...

I think you are right that the space elevator enthusiasts are low-balling the cost, but I think it's by a lower factor than you do.

The minimal useful space elevator would be just a rope or ribbon of whatever super-strong material is available. This is considerably less elaborate & so less expensive per km than a light rail line.

Rick said...

You're right - I admit that my mental image is a heavy duty one. But simplifying the cable complicates the tramcars (e.g., if they're not drawing power from the cable, they need some other power feed), and the tramcars in general are non-trivial.

My meta grump is that rocket economics looked pretty good in the Willy Ley days, until we actually started building them. The devils are all in the details, and there are lots of the little buggers.

Z said...

I've always had a pseudo-credulous relationship with space elevators too- and this coming from someone who has worked on climber designs for the X Prize Cup climber challenge. I mean, no one has yet to build a simple suspension bridge out of something as fancy as Kevlar- and the big hope for the future of space travel is a bridge tens of thousands of miles long out of a material that we can't yet make golf clubs or stealth jets out of, and there is no way to make a baby prototype. Some of the more interesting analyses I've seen seem to point to the thing being borderline unmanagable as it coils around the Earth as the payload climbs, and getting around it involves slowing the ascent to be point that you might not ever be able to amortize the cost of the cable.

That being said, I'll keep up with the climber contests, and the steadily creeping outputs of carbon nanotubes- but at the end of the day, you can get into space with jet fuel, jet fuel is cheap, and so are seats on most jet-fuel powered vehicles. :-)

Canageek said...

I've got to imagine that someone has looked into using a railgun as a means of propelling things into space. I wonder what the downsides would be, aside from a massive amount of power needed.

Rick said...

As I understand it, a true railgun (as opposed to a coilgun) is limited to ~5 km/s 'muzzle' velocity, so it would have to launch an entire upper stage. And the acceleration would be extreme.

Coilguns, we don't really know how to build yet. And they would still have the acceleration problem.

Oh yeah, also the vehicle would be going very fast in the lower atmosphere = extreme heating and rapid deceleration.

Kind of a bummer ...

Canageek said...

Too bad. Due to them being involved in my favourite technique (Nuclear magnetic resonance) I've been following work on new superconducting magnets, and there are some cool advances coming in that field.

Rick said...

Nuclear resonance anything is way above my science pay grade, but I imagine it is relevant to coilguns, not just railguns.

Still not very helpful as an Earth surface launch system, but possibly suited to mass drivers for deep space use.

And, yes, to coilgun weapons also.