The Mac/Windows fanboy battles ain’t got nuthin’ on a bunch of stirred up rocket scientists.
Point out that chemical rockets don’t seem to be able to scale to get us to millions of people out in space, and the wailing and gnashing of teeth commences. Not to mention the ad hominem attacks. I’ve even been told to “Leave the rocketsurgery to those qualified to comment on it”.
No explanations of how my reasoning is wrong, just comments like “that’s stupid” or “
The severe lack of reasoned discourse makes me think I might have struck a nerve.
Dear Mr. Andrews,
I apologize for my colleagues’ rudeness, but they are mostly correct. The problems with chemical rockets have nothing to do with scaling. We don’t know how to make cheap small ones, either. 🙁
The mass of a pressure vessel scales linearly with volume. Aerodynamics favors larger vehicles, and larger launch vehicles tend to cost less per pound than smaller ones. There’s no reason in principle why you couldn’t build a rocket as large as a supertanker and launch it from water as Robert Truax (Seadragon) and Jim Akkerman (Advent) proposed. But you’d never find enough customers at $5000/lb. to fill your manifest. The problem is that all launch vehicles, large and small, are prohibitively expensive for commercial manned spaceflight.
However, you may be correct about having struck a nerve. Here are a couple of dirty little aerospace engineering secrets:
1. Nobody really knows why rockets are so expensive.
2. Nobody will admit that they don’t know.
Rand Simberg claims that it’s because of economies of scale, but I’m not persuaded. I prefer Maxwell Hunter’s “lack of intact abort capability” explanation.
You have an economics background. I would be delighted to hear your reaction to my Launch Cost Rant (a discussion of 19 different explanations I’ve heard):
http://home.earthlink.net/~peter.a.taylor/launch.htm
Once you get into low Earth orbit, I agree that large vehicles need something better than chemical rockets. But for Earth launch, if we don’t know why chemical rockets are so expensive, how do we know that something else is going to be better?
Peter Taylor
I’m a layman on rockets, although I’ve been a “space nut” since I can remember (watched the 1st moon landing live so it’s been a while.) I do know something about aircraft and aircraft maintenance, I was a helicopter mechanic in the Marine Corps.
Economically speaking, I think there is one main reason rockets are so expensive, although everything on your list (except for the first three) contributes to the problem.
Economies of scale in rocket production and operation.
Historically, rockets have been one off builds. One offs can’t use the economies of scale that an industrial production line brings to cost savings. Since World War II even airplanes haven’t been built in true assembly line quantities. A few hundred or even a few thousand built won’t drive the price down much. You have to build tens of thousands or even millions. Then the true economies of scale kick in.
Think of cars. If your car had to be built one at a time, and the factory only made a few hundred a year from all custom parts, it would cost millions of dollars. Sure enough, if you look at Formula 1 cars, those are the prices they deal with.
Anything built as a one off will be orders of magnitude more expensive than the product of an assembly line. Prototypes of everything from cars to electronics are hundreds of times more expensive than the final product. I used to work for a company that designed microchips, and the chips they made to test designs were incredibly expensive because the production runs were so small. When they finalized the design, they shipped it off to one of the big fabs and the chips were made for pennies a piece.
I mentioned airplane production in WWII. The US built over 300,000 airplanes from 1939 to 1945, and on assembly lines like cars. I think that if we made that many rockets in six or seven years, they’d be pretty cheap.
Notice that I’m talking about economies of scale in building the rockets, not in launches. Operational economies of scale have more to do with manpower costs than anything else. Imagine NASA trying to launch a rocket every day of the year. Epic Fail.
One off production methods equal expensive. Sometimes there’s nothing you can do about it. Dams and tall buildings are one offs too, and incredibly expensive, but the expense is amortized through constant use over decades. A new 737 costs around $50 million (Boeing made 5700 as of 2008, an average of 142 a year. Still not what I consider production line quantities.), but an airline will fly it full, at least a couple of times a day, every day for years. A Boeing 777 is estimated at $187 – $250 million yet airlines are buying them because they know they can amortize the costs over thousands of flights. If you could do that with the space shuttle, you’d amortize the $1.7 billion cost of the Endeavor pretty quick, but it would be impossible to fly that pig twice a day, every day for years. Instead we get estimates of $1.3 billion per launch over the life of the program for the space shuttle.
If we get to the point where we can build a rocket that can be launched thousands of times with minimal turn around expense and low operational manpower requirements then they’ll approach airliner economies of scale in operation. On the other hand, if we can build hundreds of thousands of them, the cost per rocket would be lower and we wouldn’t have to worry about amortizing the costs over long term operation.
Frankly I don’t see hundreds of thousands of rockets being built. So the best bet is to come up with a robust design that can hold up with minimal maintenance for thousands of flights.
So I’ll throw it back to you, as an aerospace stress analyst, is it possible to build an airframe and engine system for a rocket that can hold up for the thousands of flights required to amortize it’s costs down to levels comparable with those of airline operations? I think your point about intact abort capability falls in there somewhere.
(My bet is no, but I’d love to be wrong.)
I think the airframe is the easy part. The thermal protection system (TPS) scares me more than the airframe, and the TPS has to not interfere overmuch with being able to inspect the airframe. I wish NASA would spend more money on things like transpiration cooling experiments. Also, I am willing to “cheat” on the airframe design by
using expendable tankage and expendable heat shields.
I don’t know much about engines. The RL-10 (small Pratt & Whitney LOX/LH2 engine used in the Centaur upper stages) is pretty reliable, has decent performance (Isp and T/W), and seems to do quite well in terms of accumulated operating minutes on test stands. But they still have occasional “hard starts” that cause loss of vehicle and payload. XCOR is working on safe engines, at the expense of performance. They won’t say what their Isp or T/W numbers are.
The intact abort issue is another matter. We are talking about several different things here: long service life, low maintenance cost, high reliability, and benign failure modes. The Shuttle’s target service life of 100 flights sounds too short, but amortization over 100 flights is less of a problem than is (1) maintenance cost or (2) the fact that we haven’t been able to get an Orbiter to fly 100 missions without a catastrophic accident. My argument is that (1) the maintenance cost is driven by reliability requirements that are in turn driven by too many catastrophic failure modes, and (2) the expected life of an Orbiter is already limited more by catastrophic failures in other systems than by fatigue or corrosion.
Imagine a trans-oceanic airliner with a novel engine design, no engine-out capability, and no capability to ditch safely. The way the space launch industry has dealt with situations like this historically is by hiring lots of people (Maxwell Hunter’s “marching army”) to try to improve reliability. Hunter wanted to deal with the situation by, in effect, adding an engine-out capability and laying off part of the “army.” We’re not to the point yet of worrying much about making the airframe or engines last longer. We’re looking for a cheaper way to avoid catastrophic failures.
The intact abort problem appears to me to be a subset of the operational amortization issue. As in, if it can’t fail gracefully, it can’t live to continue operating thus cutting short the amortization and ruining the amortization.
I think you’re right about that.