[AR] danger (was Re: MSR reactors.)
- From: Henry Spencer <hspencer@xxxxxxxxxxxxx>
- To: Arocket List <arocket@xxxxxxxxxxxxx>
- Date: Sat, 13 Jul 2019 20:27:02 -0400 (EDT)
On Sat, 13 Jul 2019, William Claybaugh wrote:
Apollo was estimated at the time as having a one in five probability of
loss of crew.
The actual flight history argues -- somewhat weakly -- that the risk for a
lunar flight was about one in ten.
There were 9 such flights (Apollos 8, 10, and 11-17), one of which (13)
was a failure. While no crews were lost, the Apollo 13 oxygen-tank burst
came *extremely* close to being fatal; it would have been fatal had it
happened on the trip home, or had any of several other factors been a
little worse. Moreover, the same flight (!) came very close to a
catastrophic launch failure due to the pump-amplified second-stage
center-engine Pogo (which is not well known because the tank burst
overshadowed it and MSFC understandably kept a low profile about it --
13's Flight Evaluation Report has a whole chapter about it, but with such
bland matter-of-fact wording that you have to read and understand the
numbers to realize what a close call it was). There were a few other
lesser incidents on other flights that were a bit worrisome.
That adds up to maybe 1 in 9, which you might as well call "about one in
ten" because there's room to argue about the addition, and the sample size
is too small to be very precise anyway. (If one case turning out
differently would greatly change your estimate, then your estimate can't
be very precise.) It would be a bit surprising if the real odds were as
bad as one in five or as good as one in twenty, but there isn't enough
data to reject such a suggestion very convincingly. There *is* enough to
inspire real doubt that the official crew-survival spec (99%) was met.
A decade ago the then Constellation architecture for earth-Moon human
transport came in at about one in ten.
The initial flight of the Space Shuttle was retrospectively estimated at
one in four; the last flight at one in 63.
Plausible numbers.
However, a caution: all this is not saying that human spaceflight is
*inherently* very dangerous. It's saying that human spaceflight done the
traditional way is inherently very dangerous. Which isn't too surprising,
given that all of those crews were (or would have been) flying on what
should be considered *early test flights* of complex new vehicles, all
also involving major hardware that was flying for the very first time.
We *know* how to do much better. Nobody puts valuable payloads, or crews
other than test pilots, on the first flight of a new aircraft design *or*
the first flight of a new production article of an established design.
Proper flight-test programs for new designs involve hundreds of flights,
not two or three, including systematic envelope expansion and tests of
contingency cases like engine failure, and every single new production
vehicle is test-flown before it's used.
Doing the equivalent for spaceflight should be possible. It would require
vehicles that are designed rather differently, and development programs
that are designed and run very differently. (Notably, the ability to fly
some semblance of a proper flight-test program, at an affordable cost in a
reasonable time, needs to be a non-negotiable design requirement. This
does preclude (a) major expendable components and (b) vehicles that have
to be essentially rebuilt after every flight.)
"Spaceflight is difficult -- today. It is expensive -- today. And the
level of risk remains high -- today. But it need not remain so forever.
We must resist the idea that space is inherently difficult, expensive, and
risky. Aviation once seemed so, too. Today, aviation is efficient and
safe. Space can get there -- if we accept that it can improve, and
realize that it will require hard work, investment, and experimentation."
-- Joseph C. Anselmo (AW&ST Editor-In-Chief), 3/10 Nov. 2014 issue.
Henry
Other related posts:
