[AR] Re: SSTO fuels (was Re: SSTO)

  • From: Henry Vanderbilt <hvanderbilt@xxxxxxxxxxxxxx>
  • To: arocket@xxxxxxxxxxxxx
  • Date: Fri, 16 Feb 2018 21:13:46 -0700

On 2/16/2018 2:49 PM, William Claybaugh wrote:

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.

"About four orders of magnitude more" is at least relatively precise - that's an average multiplier of 10^4, 10,000X, give or take.

But "stuff that goes to orbit" and "subsonic stuff" both could mean a lot of different things here. Could you pin "stuff" down a bit?

In the meantime, for my best guess from context at what "stuff" you have in mind - we've been talking reusable orbital launchers and comparing them to subsonic airliners - 10000X more expensive for the orbital iron seems seriously high.

Mind, you still could be talking about either development costs, or about cost of the actual "stuff" - vehicle structure and systems - that leaves the ground. But in both cases, 10,000X more for the space stuff seems implausible.

Let's start with a look at some recent examples of actual vehicle weights and costs, for a current airliner and a current launch vehicle. I'll look at both overall vehicle numbers, and at engines broken out separately. (Rough numbers - no point in worrying very far past the first decimal when we're talking four orders of magnitude.)

The 787 airliner weighs ~250,000 lbs dry, and costs ~$250m each. Easy math, that's ~$1000/lb overall.

A GE GEnx 787 engine weigh in at ~13,000 lb dry, and lists for $22m. That's ~$1700/lb.

Meanwhile, an Atlas 5 with Centaur upper stage weighs about 51,000 lb dry, and lists for $108m lately. That's ~$2100/lb.

Hmm. 2.1 times the cost per pound of a 787. We seem to be missing a few orders of magnitude... Let's see if they're hiding in the engines.

The Atlas 5 booster engine is a Russian RD-180, ~12,000lb dry, about $12m each. ~$1000/lb. Hmmmmm...

But wait, let's be fair. ULA got a smoking long-term deal from the Russians on RD-180 a long while back. Let's bump that cost to what I've seen estimated for a US-manufactured RD-180 clone, ~$50m each, and add 20% more for inflation, ~$60m each. Bringing us to $5000/lb for RD-180, or about 3 times GE GEnx's $1700/lb.

OK, we're only missing three and a half orders of magnitude there.

Let's look at the Atlas 5 upper stage engine, an Aerojet RL-10A5. No published price, but Aerojet has reportedly been severely jacking up the tariff on RL-10 in recent years - I've heard to even more than the RD-180 costs.

Let's guess at $15m for an RL-10A5, which has dry mass of ~370lb, for a cost of, wow, ~$40,000/lb. So we're up to ~24 times 787 engine $/lb.

We've gotten the discrepancy under three orders of magnitude! Mind, it took a prime example of dinospace-monopoly all-the-traffic-will-bear pricing to do it, but still. Only a bit over a factor of 400 off.

And I've gone on too long already, so I'll just blast through development costs.

787 development cost exclusive of manufacturing is ~$16 billion, or ~$64,000 per dry airframe pound of one vehicle.

Atlas 5 development cost between L-M and USAF was in the neighborhood of $2 billion, IIRC. To be fair, let's bump that to $3 billion, to account for a share of Centaur upper stage development costs plus miscellaneous other no-doubt hidden system costs.

That's ~$59,000 per Atlas 5 dry airframe pound.  Versus ~$64,000/lb for 787.

Four of our orders of magnitude are missing.


All snark aside, this bears out what I've been driving at elsewhere. Development costs for advanced rockets are not automatically some large multiplier of development costs for advanced airliners.

The actual development cost for advanced aerospace vehicles in general tends to scale strongly with dry weight of the vehicle, times a per-pound multiplier related to just how hard you're pushing the state-of-the-art to maximize performance and minimize weight.

And in order to make money on new airliners, they're pushing both of those factors pretty darn hard.

There are also other factors that actually favor rockets, like not having to optimize everything for high-volume production, and that would favor a first-gen reusable SSTO, like not having to press nearly as far up all available diminishing-returns curves to be competitive.

I reiterate: Developing a useful first-gen SSTO space transport as a private venture? Some billions, yes. Tens of billions, no.

IF you avoid heading down too many of the classic dinospace blind alleys, of course.

Henry Vanderbilt






On Fri, Feb 16, 2018 at 1:28 PM Henry Vanderbilt <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>>> 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>>] 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>>
     >     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
     >
     >
     >


Other related posts:

  1. Home
  2. arocket
  3. Archive
  4. 02-2018