OK, thanks for the quick replies. (Though I should have been able to
answer the one about solid casings myself, from the positions on the
horizontal vehicle-dry-mass axis. [smacks forehead] )
Some thoughts on your chart, then. (I've attached it again for others'
reference convenience, I trust you won't mind.)
- Background.
I've found it useful over the years to divide aerospace vehicle dry
masses into two broad categories: Structure, and systems. Structure
should be self-explanatory - load-bearing, containment, partitioning,
etc. Systems includes engines, avionics, actuators - any active mechanisms.
Both structures and systems range from high-weight low-cost "dumb" to
light-weight high-cost "smart". "Dumb" often also involves high-margin
and/or low-maintenance. "Smart" often also involves low-margin and/or
high-maintenance. (And both spectrums of course tend to shift to the
right over time.)
And systems at any given point on that spectrum will tend to cost
significantly more and require significantly more maintenance per pound
than structures.
- I propose that the different balance between structure and systems
is actually why all the combat aircraft cluster closely on the vertical
axis somewhat above your "sonic" trendline, while the three transport
aircraft are all somewhat below it - despite one of these three being
the supersonic Concorde.
Consider: Military combat aircraft tend to have to carry both middling
payload (ordnance) plus a lot of high-performance systems - sensors etc.
Transports meanwhile focus primarily on payload, with only the minimum
necessary systems.
So all else being equal transports are going to have a significantly
lower systems-to-structure ratio than combat aircraft, and thus will
also have lower maintenance hours per flight per pound (structures being
lower maintenance per pound than systems, all else equal).
This has some bearing on an SSTO, which like Concorde, will tend to have
very high-performance engines and structure - and the absolute minimum
necessary of other onboard systems.
- Shuttle seems surprisingly low on the vertical Labor Intensity axis
at about the average for space launchers, despite its notoriously high
labor-intensity. But a bit of thought explains that: ~60% of Shuttle's
dry mass is relatively "dumb" steel solid booster casings, which aren't
all that labor-intensive even with NASA in charge. This causes the
labor-hours-per-dry-lb measure to rank it artificially low, relative to
practical labor intensiveness.
- Ditto Ariane 5, to a lesser extent.
- Though that doesn't explain Ariane 4's low labor-intensity. I might
guess two things there: Genuine adherence by its designers to the KISS
principle (this also applies to Soyuz) plus a heartfelt desire to get as
much as possible done in the factory in France rather than on the pad in
French Guiana. (Either or both may apply to 5 to some lesser extent.)
- Most of the mass being steel booster casings didn't help Titan 4
near enough, possibly due to its chief customer's habit of keeping it on
the pad tinkering with things for a year or more before each launch.
- For expendable launchers in general, a great deal of the final
launch-site integration and test labor that's counted here, for reusable
aircraft takes place in the factory and is not counted. This tends to
stretch the vertical difference between space launchers and aircraft
considerably.
In that vein, it would be instructive to see the Falcon 9 plotted on
this graph, first flight separate from reuse flights. Also to see how
the latter changes with the block 5 design and with experience in
general. Alas, yes, SpaceX is not sharing that very useful competitive
data publicly at this point, oh well.
- Atlas Centaur I take it is the old all-balloon-tanks version? Both
ultra-lightweight and high-maintenance, its very high
labor-intensity-per-lb position makes sense.
Though for comparisons between launchers, a chart of labor hours per
payload pound might be instructive. This chart's format implicitly
penalizes ultra-light structures and systems, which may or may not be
requiring more labor hours but certainly have fewer dry pounds. (To
some extent, the same goes for launchers overall versus aircraft.)
- Loose ends, final thoughts: SR-71's higher-than-fighters
labor-intensity-per-lb make sense, as it was pushing considerably harder
to the "smart" end of the structures/systems spectrum than the fighters
listed.
- X-15, even more so - plus the factor that it was also much farther
down the hand-built few-copies experimental vehicle spectrum, where
minimizing routine maintenance hours takes a back seat to just getting
the darn thing to fly.
- DC-X obviously isn't so high on the labor-intensity-per-lb scale
because of its high performance or lightweight structure. There, I
suspect the extremely limited budget exacerbated the just-get-it-flying
factor's complete predominance over providing for easy maintainability.
Anyway, thanks for posting that!
Henry
On 2/17/2018 3:24 PM, William Claybaugh wrote:
Henry:
For launch vehicles, ops begin when it arrives at the launch site (lands, for Shuttle; ferry flights are included in ops).
I’ve bundled the data here, but in some cases I also have a breakout for “moves from processing to launch pad”. More efficient vehicles obviously—in the data—minimize launch pad time.
Words have meaning: dry weight—in this case—means exactly that. No propellant was included in the mass estimate.
I have no ops data on F9; I have counted cars at the launch site which has not proven helpful: there are pretty consistently more cars at that site than at ULA’s Atlas pad, go figure.
Bill
On Sat, Feb 17, 2018 at 3:09 PM Henry Vanderbilt <hvanderbilt@xxxxxxxxxxxxxx <mailto:hvanderbilt@xxxxxxxxxxxxxx>> wrote:
Bill,
"Obviously incorrect" is overdoing the mea culpa. Too cryptic and
context-deficient, perhaps. A hazard of posting late and tired - but
we've all done that at some point.
My apologies in turn for perhaps once or twice visibly enjoying myself
too much during my guessing-what-you-might-mean responses. (Mind, even
guessing wrong I found them useful; I hadn't revisited those numbers in
too long.)
But now that you've posted that chart, there's obviously an interesting
point here: Adjusting labor-hours per flight for overall vehicle dry
mass does lead to some interesting and instructive clusterings, and
potentially clarifies things in the search for factors affecting vehicle
ops costs.
Though I do have two questions about your assumptions there:
- For expendables, where did you draw the line between manufacturing
man-hours and support man-hours? The logical divide would be at the
point where the stages get delivered to the launch site and begin
pre-launch erection and processing. (Mind, on an expendable, I could
see a defensible argument for counting manufacturing hours too.)
- For vehicles with fixed-size large solid boosters associated -
Titan
4, Ariane 5, Shuttle - did you include the dry mass of the solid
casings?
Hmm, make that three questions: What do you have on where Falcon 9 comes
down on this chart? You did mention some such. (And, where might F9
reuse flights fit into that?)
Henry
On 2/17/2018 7:22 AM, William Claybaugh wrote:
> All:
>
> My sincere apologies for this obviously incorrect post.
>
> There is a four order of magnitude difference in *Labor Intensity*
> between subsonic aircraft and space launch systems (see the attached
> chart), not in cost per pound.
>
> By way of explanation (but not excuse), let me admit too putting
in some
> fairly long hours lately on this SSTO study and being in the
middle of
> writing about labor costs when I dashed off this stupid post.
>
> My apologies to all,
>
> Bill
>
> On Fri, Feb 16, 2018 at 2:49 PM, William Claybaugh
> <wclaybaugh2@xxxxxxxxx <mailto:wclaybaugh2@xxxxxxxxx>
<mailto:wclaybaugh2@xxxxxxxxx ;<mailto:wclaybaugh2@xxxxxxxxx>>> wrote:
>
> Henry:
>
> To first order and to date, stuff that goes to orbit costs about
> four orders of magnitude more than subsonic stuff, per pound.
>
> SpaceX has pulled about half an order of magnitude out of that,
> leaving the difference—in their case only—at almost exactly four
> orders of magnitude.
>
> Bill
>
> On Fri, Feb 16, 2018 at 1:28 PM Henry Vanderbilt
> <hvanderbilt@xxxxxxxxxxxxxx
<mailto:hvanderbilt@xxxxxxxxxxxxxx>
<mailto:hvanderbilt@xxxxxxxxxxxxxx
<mailto:hvanderbilt@xxxxxxxxxxxxxx>>> wrote:
>
> Cost-per-airframe/engine pound certainly scales up with
higher
> vehicle
> performance.
>
> Development cost per project has a less linear
relationship with raw
> vehicle performance - other significant variables also apply.
>
> See my previous remarks about the different demands of
achieving a
> profitable performance increment over existing
mature-technology
> ailiner
> competition, versus developing a Good Enough version of a
> radically new
> space transport approach that inherently brings with it a
> significant
> performance edge.
>
> And on the gripping hand, setting up for economic serial
> production of
> hundreds-to-thousands of copies of a big state-of-the-art
> airliner is a
> major expense that developers of advanced rockets
generally avoid.
>
> In fact, SpaceX's investment in reusability can be viewed as
> primarily a
> way to support their high (for the old expendable industry)
> flight rates
> with a much smaller/cheaper booster core production
> establishment than
> they'd otherwise need.
>
> To a first approximation, a successful Mark 1 version
> fast-turnaround
> SSTO space transport will not immediately require mass
> production. More
> like single digit numbers of hand-built copies.
>
> Later marks, as the market radically expands, will be a
> different story.
> But the revenue from the early marks will be there to help
> support
> establishing higher-rate production. Not an issue for
funding the
> initial push to market.
>
> Henry Vanderbilt
>
> On 2/15/2018 3:15 PM, William Claybaugh wrote:
> > Rick:
> >
> > Productivity gains in the aerospace sector have pretty
much
> matched
> > inflation over the period since the 747 was developed;
> accordingly, a
> > large passenger aircraft should cost—in today’s
> dollars—pretty much the
> > same as a 747 cost in then dollars. $1 Billion by your
estimate.
> >
> > The other glaring issue here is that a subsonic
aircraft is not
> > comparable to a Mach 25 spaceship; trying to use the
one to
> estimate the
> > cost of the other guarantees underestimating.
> >
> > Bill
> >
> >
> >
> > On Thu, Feb 15, 2018 at 1:42 PM Rick Wills
<willsrw@xxxxxxxxx <mailto:willsrw@xxxxxxxxx>
> <mailto:willsrw@xxxxxxxxx ;<mailto:willsrw@xxxxxxxxx>>
> > <mailto:willsrw@xxxxxxxxx ;<mailto:willsrw@xxxxxxxxx>
<mailto:willsrw@xxxxxxxxx ;<mailto:willsrw@xxxxxxxxx>>>> wrote:
> >
> > Henry
> >
> > I'll throw my 2 cents in here.
> >
> > $20B should be an upper limit for
spaceplane/launch vehicle
> > development. My estimate is $14B to $17B. A reusable
> orbital
> > launch vehicle may or not be an SSTO but it needs
to be 100%
> > reusable. My rational for the estimate is Boeing
spent
> $1 Billion
> > to develop the 747 with first flight in 1969. Today,
> that's roughly
> > $7B. Rough order of magnitude is double Boeing's
cost;
> than add
> > 20% for cost overruns. I can see why some people
might
> argue $20B
> > to $40B; Boeing Dreamliner is reported to have
cost $30B
> to develop.
> > However, SpaceX could hit 100% reusable with a
reusable
> upper stage.
> >
> > On Monday afternoon, I spoke to freshman
mechanical and
> aerospace
> > engineering students at the University of Dayton
on the
> subject of
> > the Engineering Profession. In my "lessons learned"
> section, I
> > discussed bias. Yep, we all got them. As an
example, I
> discussed
> > my bias about what a reusable orbital launch vehicle
> would like. My
> > long held view was a reusable launch vehicle would be
> "aircraft
> > like": wings, landing gear, etc, and of course a
pilot.
> (full
> > disclosure, I hold a commercial pilot rating and am
> engineer). In
> > preparing for the talk, I realize this bias when
as far
> back as my
> > childhood looking at Pratt & Coggins book "By
Spaceship
> to the
> > Moon". It's 1950 technology but the science is
solid for
> the time.
> > In it, there is a nice drawing of a large winged
> vehicle, they
> > called it a supply ship. The vehicle was taking off
> horizontally
> > with a rocket powered sled. My five year old self
looked
> at that
> > and thought, "that's neat". I now understand the
technical,
> > developmental, political, and financial issues
with these
> sorts of
> > system configurations but the bias was implanted. Now
> Space X
> > comes along and shows how recovering an intact
undamaged
> first stage
> > can return a profit. Biases do die hard, but
it's hard
> to argue
> > with success.
> >
> > Take Care and Be Safe,
> >
> > Rick Wills
> > Still waiting for Buck Rogers
> >
> > -----Original Message-----
> > From: arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>
> <mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>>
> > <mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>
> <mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>>>
> > [mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>
> <mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>>
> > <mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>
> <mailto:arocket-bounce@xxxxxxxxxxxxx
<mailto:arocket-bounce@xxxxxxxxxxxxx>>>] On Behalf Of Henry
> Vanderbilt
> > Sent: Thursday, February 15, 2018 1:54 PM
> > To: arocket@xxxxxxxxxxxxx
<mailto:arocket@xxxxxxxxxxxxx> <mailto:arocket@xxxxxxxxxxxxx
<mailto:arocket@xxxxxxxxxxxxx>>
> <mailto:arocket@xxxxxxxxxxxxx
<mailto:arocket@xxxxxxxxxxxxx> <mailto:arocket@xxxxxxxxxxxxx
<mailto:arocket@xxxxxxxxxxxxx>>>
> > Subject: [AR] Re: SSTO fuels (was Re: SSTO)
> >
> > On 2/13/2018 7:14 PM, William Claybaugh wrote:
> > > I have seen that paper. For something as
technically
> (much less
> > > economically) difficult as SSTO it seems a little
> light: even much
> > > more detailed analysis doesn’t often lead to much
> confidence that I
> > > ought to recommend dropping $20 or $40 billion
on one
> solution
> > over another.
> >
> > My two cents worth: If fielding a useful SSTO space
> transport is
> > costing you $20 to $40 billion, you're doing something
> very wrong.
> >
> > That's the sort of price tag you get by farming it
out to the
> > existing cost-plus government aerospace houses,
> supervised by an
> > existing high-overhead government R&D bureaucracy.
> >
> > At the end of that process you may or may not get a
> useful space
> > transport, but lots of people will have had decades of
> low-stress
> > white-collar job security. Fine if that's your
objective -
> > typically if you're a Congressman and they're your
> constituents - if
> > you actually care about building useful space
> transportation, not so
> > much.
> >
> > Done as previously described, build your own
private team
> up doing
> > methodical risk-reduction then development (as with
> SpaceX and Blue)
> > it should be perhaps a tenth of that.
> >
> > Henry V
> >
> >
> >
>
>
Attachment:
Labor Intensity.pdf
Description: Adobe PDF document