Re: starship-design: Re: One way (again...)

[kuo@bit.csc.lsu.edu (Isaac Kuo)]

L. Parker wrote:
>On Tuesday, December 09, 1997 12:05 PM, Isaac Kuo 
>[SMTP:kuo@bit.csc.lsu.edu] wrote:

>> >It is statistically certain to fail.  I.E. if you are asking a
>> >systems to work longer then the average mean time to failure of
>> >its parts, it will fail without the replacement of those parts.

>Kelly's right it is certain.

Not according to the numbers you supply, and you supply numbers
for _current_ technology.  Not the sort of technology that could
acheive .2c.

>> No it won't.  The average mean time is an average.  The part might
>> fail before, or it might fail later.  It could fail today.  It
>> could fail in a century.

>Which is beside the point. Typical modern MTBFs are pitiful compared to 
>what the mission will require. And you are thinking MTBF of the PART, the 
>system as a whole is LESS than the lifetime of the weakest part.

However, when you have replacement parts this effect is minimized.

>> But there's no inherent reason why we would ask the systems to
>> work longer than their average mean time to failure.  We can
>> bring spares to replace systems before they wear down dangerously.

>No, you can't. You would have to bring the equivalent of four or five whole 
>ships worth of spare parts.

Which isn't an inherent reason, because it's possible to bring four
or five whole ships worth of spare parts.

Increasing the ship mass by 10 or 20 or 200 times is still a bargain
if it saves a mass increase of 10,000 for a return journey.

>> >If the parts are primary structure (remember we'll be shaving
>> >weight margines to get the thing flying) you need major shipyard
>> >facilities.

>> I don't think we'll be shaving weight.  Even at .2c, the
>> thing has _got_ to last at least 20 years or the whole endeavor
>> wasn't worth a damn.

>Which is exactly the point. We have NEVER built any mechanism that was 
>designed to function flawlessly for twenty years without ongoing 
>maintenance. Historical evidence would tend to argue that we can't.

We aren't talking about a _mechanism_ there.  We are talking about
the primary structure.  We've built structures which can last 20 years.

And we are going to be performing ongoing maintenance.

>> >>>Normal systems on that scale usually burn out after 40-50 years.
>> >>>Given the lack of replacement parts (stored parts also don't last
>> >>>forever),

>> >>They don't have to last forever.  They just have to last several
>> >>decades.

>The Air Force places a red tag shelf life on safety harness webbing at five 
>years. This means if it sits on a shelf for five years, the inspector must 
>physically destroy the webbing rather than chance its being used 
>accidentally.

So?  That's a paranoid government regulation.  It doesn't reflect any
MBTF for things hermetically sealed in storage in low pressure inert
gas or vacuum.

>> >Many can't last a few years on the shelf.

>> Like what?  The mission critical systems are:

>> 1. The deceleration rocket systems.  These have to last 2 decades
>>    and there's little margin for spares.  However, after that
>>    they are no longer mission critical.

>Current maximum operational lifetime: 50 hours continuous use, shelf life: 
>none.

Huh?  We don't have _any_ current deceleration rocket system suitable
for a .2c mission.  The rocket systems we do have aren't even remotely
like what would be used for such a mission.

And it's nonsense for shelf life to be lower than the lifetime during
use.

>> 2. Oxygen recycling and CO2 scrubbers.  At least with current
>>    technology, they have a limited expected life span, but
>>    they are relatively lightwieght so many spares can be
>>    carried.  I'm not sure about their shelf life.

>Current maximum operational lifetime: 2 years continuous use, shelf life: 
>none.

2 years continuous use?  Then bring 100 of them.

And what do you mean by no shelf life?  It makes no sense that it
will just crumble to pieces if it's ever turned off.

>> 3. Water recycling.  I'm not sure about this part.
>Current maximum operational lifetime: 2 years continuous use, shelf life: 
>none.

So you bring 100 of them.

I have no idea what this "shelf life: none" means.

>> 4. Food storage.  Irradiated canned food will easily last a couple
>>    hundred years.

>> 5. Spare parts to repair hull problems.  Aluminum nuts, bolts,
>>    welding solder, and wrenches in vacuum storage practically last
>>    forever.  Arc welders also last practically forever since they're
>>    relatively simple devices easy to repair.

>Do YOU know how to repair one? And just in case something happens to you, 
>how about the guy in the next cubicle?

Yes, and if I had to I could probably build one with the spare parts
in storage.  So would everyone on the ship, at least a couple decades
into the mission.

>> 6. Spare solar panels and electrical components.  Last prctically
>>    forever in storage.

>Only, solid state devices have any reasonable shelf life. Unfortunately, 
>their operational life is so short that you will need a LOT of spares.

So you carry lots of spares.  Running a robotics lab has taught me
that practically anything electronic can be revived eventually.

>> >needs repair NOW, you can't just hope it woun't fail for
>> >a decade or two for the last crewman to die.  It almost
>> >certainly will fail in months to years.

>> Why would it almost certainly fail in months or years?  Exactly
>> what mission critical components are certain to fail, even with
>> triple redundancy?  (If there's only one or two crew left,
>> the life support systems will be well below capacity.)

>I already posted a long list of relatively simple, everyday items that must 
>be included on the ship, ALL of which will fail within five years.

None of which seemed to me to be mission critical and all of which
could be built to last.  But we don't for everyday items because we
have no reason to in everyday life.

>Here is a real good question for you: if you fill an air tank with 
>compressed air to 3,000 psi and put it on a shelf then come back five years 
>later, how much air is in the tank? Answer: 14 psi. We can't even build a 
>non-moving air tank that will hold air with no pressure loss for five 
>years. How are you going to keep it in the ship?

By having a really thick hull.  Which you need anyway for radiation
shielding.  Also, the loss of air is proportional to surface area,
while the amount of air is the volume of the crew compartment.  The
square/cube law alone guarantees a large crew compartment will lose
air relatively more slowly than a small air tank.

>> Why would the exploration gear become unservicable so quickly?
>> At the very least, we can expect handheld optical telescopes
>> to last hundreds of years.  Even that alone, at such a close
>> range, is enough to do serious scientific observations impossible
>> from the Solar System.  (Even if we figured out a way to make
>> astronomically huge optical telescopes able to equal their
>> resolution, we could not make fine corona observations since
>> we'd lack the ability to shade out the photosphere.)

>Isaac, would you give up your computer for a pen, paper and slide rule? Do 
>you even know how to use one?

Not unnecessarily but I would do it if I had to.  Yes, I know how
to use pen, paper and slide rule--but I would prefer an abacus because
it's faster and more precise.

Most of my acedemic life I studied mathematics, which mostly involves
lots and lots of proofs by derivations and calculations by hand with
pen and paper (computers just can't handle this sort of thing yet).
It really isn't that bad.

>> >Past
>> >that your need to strip those systems for pars to regulate life support,
>> >medical, etc..

>> Huh?  Keeping the systems alive will be a matter of repairing them
>> with spares.  There's not much commonality between a CO2 scrubber
>> and an IR camera.

>Forget the spares, there aren't going to be any. Do the mass calculations 
>before you bring this one up again. Figure the individual failure rates of 
>each part and the aggregate failure rates of each subsystem and system. 
>Compute average life expectancy for each system, add sufficient spare parts 
>to replace EVERY part in the system as many times as necessary to get there 
>(we won't even bother with getting back for this argument). Then total up 
>the additional mass and recompute fuel requirements.

Before I do such mass calculations, why don't you list for me each and
every part, and the mass ratios of each part relative to each other.

My intuition is that spares will not increase mass by more than 1 or 2
orders of magnitude, since the majority of the payload mass will be in
the highly robust deceleration rocket systems, the thick radiation-proof
hull (which doesn't need much in terms of spares), and the food supplies.

>> Most of the time spent on a manned spaceship, at least currently, is
>> keeping yourself alive.  That's a given.  But really that's not so
>> different from life here on Earth (especially if you're a farmer).

>Until Mir began wearing out this simply wasn't true. Which is a great 
>example and case in point. Even with routine resupply, it is already almost 
>dead. Creeping advanced senility. Which is precisely what Kelly is saying.

The reason Mir started dying is because it was up there for many years
beyond its design lifetime.  It simply wasn't designed to last that
long.
-- 
    _____     Isaac Kuo kuo@bit.csc.lsu.edu http://www.csc.lsu.edu/~kuo
 __|_)o(_|__
/___________\ "Mari-san...  Yokatta...
\=\)-----(/=/  ...Yokatta go-buji de..." - Karigari Hiroshi