I can’t see any conceivable rationale for the thing.
Well, we already have a missile range, so a spaceport is just another missile range, right? And we sure ain’t getting rich from tourists as it is…
SpaceX apparently used the place for one or two early shots (grasshopper maybe?). They now are crowing about their association with SpaceX. For some reason, SpaceX doesn’t talk about it.
Years ago, I drove from FL to CA on I-10. After West Texas, Las Cruces actually looked good.
Watching their chatter about the thing convinced me that it was getting out of Texas, not getting into Las Cruces, that looked good.
Now, even Virgin is hedging its bet on space tourism: they are now crowing about their evolving launch ability for (micro to very small) satellites sing their 747-400 launch plane.
Right - they are going to fire a two-stage rocket from the soaring height of 35,000’.
From any airport in the world.
Low orbit.
How much does a single cycle of a 747-400 cost? And, by the photo, it looks to have been a passenger plane.
I notice Mr. Rutan is no longer mentioned by Virgin.
The air launched rocket is a proven concept - the Orbital Sciences Pegasus is a 3-stage solid rocket launched from an airplane, and has been very successful. It’s been the default choice for the NASA Small Explorer satellites. Originally it was launched from a B-52, but they switched to a Lockheed L-1011 passenger airliner.
As for Las Cruces, I’ve been on many month-long business trips there (sounding rocket flight from White Sands Missile Range) and I’ve always enjoyed staying there.
Richard Branson was talking up piggy backing small rocket launches from the VMS Eve back in 2009 at the ESA airshow in Oshkosh. I always thought it made sense. Why not try to defray some of the development costs of SpaceShip2 with quick and easy launches?
As for Burt Rutan, the man retired back in 2010 at age 67 and he’s been puttering about doing all the cool things he ever wanted to do I suppose.
Stratolaunch is also developing an air-launched rocket system, using a Scaled Composites carrier aircraft similar to the White Knight Two (but larger).
The Mathematica study was done in 1971, long before the shuttle was even designed. It was not a commitment by NASA. It was a paper study to illustrate per-launch and per-pound payload costs at one theoretical extreme of the usage spectrum.
The shuttle system as actually designed was not built to fly 50 missions per year. This was discussed in detail during the Columbia Accident Investigation Board hearings. Shuttle program manager Robert F. Thompson explained this in his testimony: …Hell, anyone reasonably knew you werenʼt going to fly 50 times a year. The most capability we ever put in the program is when we built the facilities for the Tank at Michoud, we left growth capability to where you could get up to 24 flights a year by producing Tanks, if you really wanted to get that high. We never thought youʼd ever get above 10 or 12 flights a year."
Even in 1972 when the shuttle design envisioned a fully-reusable vehicle and a six-orbiter fleet (vs the four that were built), it was already known it would be hard-pressed to compete economically when expendable launchers from that era. This is very clear in this GAO report to Congress: http://archive.gao.gov/f0302/096542.pdf
The issues affecting operating cost and reusability were not primarily macro-level design elements like wing shape, payload bay size or cross range. Rather they were things like designing for serviceability (access panels, cable routing, etc.) and prioritizing programmatic design for lowest possible fixed costs. IOW to have a chance of succeeding at low cost and rapid reusability, the entire program must be structured around this (low fixed costs, minimum headcount, leanest possible production) as a top priority from inception. In the shuttle’s case this was not done.
This experience led former shuttle program manager John Shannon to call reusability a “myth”. In his presentation to the Augustine Commission He explained when a vehicle is made in very small numbers, you cannot just shut down all manufacturing capacity after it’s built. The vehicle and subsystems require ongoing R&D, fixes, testing, etc. Parts wear out, you have failures, design issues become apparent during use. You essentially have to keep the production line available, even if nothing is being manufactured, along with much of the associated industrial infrastructure. Then you must buy “one of” pieces to fix things which is expensive.
IOW if you only built four 747 airliners and the engines and components had no commonality to other vehicles, they would also be very expensive to operate – despite reuse.
SpaceX plans on a totally different kind of reusability than NASA practiced with the shuttle. Their fixed costs are drastically lower – there are only about 5,000 people in the entire company. They make most items in house.
In the April 18, 2016 news conference for CRS-8, Musk specifically discussed the possible magnitude of savings offered by full and rapid reuse:
“The cost to refuel our rocket…is only about $200,000 to $300,000. But the cost of the rocket itself is $60 million…it’s really quite fundamental…if you’ve got a rocket that can be fully and rapidly reused, it’s somewhere on the order of 100 fold cost reduction – in marginal costs. You still have your fixed costs, but in marginal costs it’s 100 fold reduction”.
So the ratio of marginal (per flight) costs to fixed annual costs significantly affects the amount of benefit from reuse. If fixed costs are low, then reuse is a big win. If fixed costs are high, then reuse doesn’t help much. Since SpaceX has such low fixed costs, they envision reusability as being a major benefit.
Well, the Pegasus as been kind of successful, with 37 successes in 42 attempts, giving a realized success rate of 0.88 and a predicted reliablity of 0.86. That’s pretty good in the commercial world, but about half an order of mangitude less than what would be desireable for mission critical applications. But the real failure of Pegasus is that it hasn’t realized the expected reduction in cost; the CYGNSS mission to be launched this year has a budgetted manifest cost of US$55M, which is in the same range as a Delta II rocket which has 6 to 14 times the payload capacity. Given that the Pegasus can carry up to about 450 kg to LEO, that works out to about US$120k/kg or US$55k/lbm to LEO, which is at the high end of launch costs for any vehicle. It should be noted that this includes a number of special services that Orbital ATK is providing compared to just a bare manifest cost, but the Pegasus has never been as cheap as expected, in part because it has never enjoyed sufficient volume to buy motors in bulk or amoritize the costs of maintaining launch capability and the mission team per annum, even though it requires a relatively small operations crew to launch. It is actually a perfect example of how a novel capability can become technically successful and fiscally non-viable.
I’ll believe Stratolaunch when I see them actually build hardware and fly something. They’ve shifted from concept to concept on an annual basis, realizing each time the challenge of launching a really large (Athena or Delta II) sized rocket horizontally, and the flight rate that would be required to make that viable. It is really just a way to sustain the WhiteKnight2 carrier for Virgin Galactic’s space tourism venture which makes zero fiscal sense.
I can only comment in general terms referencing the AIAA paper by Jonathan Barr. It is most comparable to a scaled up Centaur upper stage, which after a problematic development under the Apollo program has gone on to be a reliable upper stage for a number of different applications. Given the kind of functionality they want to build into ACES, I question that it is goign to be a “Low Cost Stage” that they advertise, and I predict they are going to have development problems with the H2/LOX APU for power generation and I think they are underestimating the challenges of on-orbit propellant transfer, but anything that gets rid of using hydrazine is an improvement in safety and processing, and dealing with propellants and service on-orbit is a technology that needs to be developed and matured. I think the ULA proposal to use it as a lunar lander platform is pretty ridiculous, though; it’s an example of using the single hammer you own to solve every carpentry problem.
I’m well aware of the assumptions that went into the Mathematica, Inc. study and other design architecture studies that went into STS, and the difference in complexity between the STS and the Falcon 9/Falcon Heavy system. The reality of the STS is that every component of the system ended up costing more than expected to maintain, so even if they had been able to achieve the required flight rate (which also would have required a substantially larger support effort and parallel processing facilities) it would not have been breakeven. However, Shuttle also operated at a cost that was an order of magnitude greater than Falcon 9 even in the best estimates, and a general scaling still applies, as shown in the Elias presentation. The fact remains that the majority of the cost in building and operating any conventional launch vehicle isn’t in materials or fabrication, but putting the vehicle together, doing all of the integration testing and repair/swapout of components, and range and launch support operations is the bulk of the cost. Reusing part or all of the vehicle doesn’t substantially reduce these costs unless it can literally be refueled and reflown with minimal testing and no disassembly or replacement save for simple repair of wear items and line replaceable units (LRU). It is noteworthy to look at the cost growth from the early days of Falcon 1.0 to the current Upgrade has having almost doubled per unit mass to LEO, which is almost exactly the trend seen by similar multistage launch vehicles.
The comparison to an airliner misses the cruicial fact that the conditions which engines and propulsion system components like valves see are far beyond any terrestrial application, with achieveable lifetimes measured in the hundreds or few thousands of seconds rather than thousands of hours for aviation components, and the criticality is such that a failure of a single component is often a catastrophic loss of vehicle, an issues SpaceX has addressed with partial redundency in critical systems like avionics, but the propulsion system and tankage are areas where completely indepenent redundancy is impractical if not physically impossible. Achieving requisite lifetime reliability over an operating lifetime of tens or hundreds of thousands of seconds sufficient to get dozens of uses out of a stage requires erring toward robustness and simplicity over highest performance and added complexity that comes with redundancy, and SpaceX has a trend of going for higher performance and more complexity, to wit their use of densified propellants.
I expect that SpaceX will reuse stages, and they may even do so successfully, but I think they’re going to have to make some substantial changes to the fundamental design in order to get dozens of uses out of a stage without essentially rebuilding it, and I do not think they are going to get the kind of costs savings they believe they will achieve based upon what I’ve seen of other attempts and studies of reuse. The optimistic estimates of controlled cost and achieving high reliability are fragile assumptions that are inevitably busted by multiple practical issues, and the more critical performance is to achieving reusability, the harder those assumptions bust. If there is an advantage to their plan to reuse it will be in facilitating a higher flight rate and engendering the nascent commercial space industry into flowering rather than large saving per individual flight.
True, but isn’t the lack of competition a big factor in this? What other options are there for SMEX class satellites? For a science satellite with limited budget, it doesn’t matter that a much larger launcher for slightly more cost; not when there isn’t money to build a larger satellite, and the satellite needs to go into a specific orbit (and cannot be piggybacked with some unrelated satellite.)
If you’re suggesting that Orbital is taking advantage of being the only vehicle in class to artificially raise prices, I can assure you that isn’t the case. They’ve made a profit on Pegasus, but it has never been the sustaining revenue that their targets and drones have been. The bulk of the cost is the Orion motors produced by Thiokol/ATK, which have gotten successively more expensive as ATK has not had enough volume of other business to amortize their capital and operating costs for their solid motor production facilities. Keeping the L-1011 ‘Stargazer’ launch platform operating after Lockheed exited the commercial airliner business and spares became scarce is also a cost that has to be spread across the ever-fewer launches.
Smallsats are more expensive to launch because there roughly the same amount of work to launch a 200 kg payload as a 2000 kg payload, and that’s the major challenge of that industry. There are a number of dedicated smalllsat launch platforms in the works, and hopefully one or two will be successful in getting launch costs down to the mid-seven figures territory. There is no way Pegasus, as currently designed, is going to get there. It’s a good example of a technology that was very innovative but ultimately not fiscally competitive they way it needed to be to transform the market.
I’ve read or skimmed through as many of these studies as I could find, and so far I’m not impressed with the ~60 flight/yr justification.
In most cases, the studies were highly specific to the Shuttle program either as executed, or one of a handful of possible variants. All of them still had the fundamental problem of being centered around the orbiter, which IMHO already invalidates any later analysis. An economically viable reusability program will not have a Shuttle-like orbiter. It is just complete nonsense to put large satellites into a craft qualified for crew and intended to “fly” back to Earth.
Some Orbital studies centered around a more sensible approach, which they dubbed a “Space Taxi” ™. It looks rather like the Sierra Nevada Dream Chaser, which itself was derived from the HL-20 spaceplane concepts. They outline several reusability approaches, but all appear to have a number of drawbacks.
The Orbital slideshow that Stranger linked to was interesting but they were too quick to dismiss a reusable first stage. It correctly notes that the first stage has the lowest payload mass penalty and the greatest cost advantage. But the item “Nearly-insurmountable Recovery Problems Unless Limited in Burnout Velocity” is clearly wrong at this point. The Falcon 9 1.1 has a MECO velocity of ~2 km/s in reusable mode and returns in excellent shape.
In my opinion, what makes SpaceX innovative is that they are pursuing what might be dubbed “opportunistic resuability”. Their system does not require reusability to work from the start, and is economic without it. Their testing program is extremely lean since it uses vehicles already paid for by a customer. Because they (currently) target only first-stage reusability, the conditions are far milder than they would be for an orbital vehicle.
Every other program appears to require billions in upfront costs to even get to the first flight. SpaceX did not (they have spent billions by now, but they only spent hundreds of millions to get to the first Falcon 9 1.0 flight). The reusability costs have been extremely cheap, requiring only a few retrofitted parts and a large boat.
Furthermore, a testing program which can afford to lose many vehicles can move more efficiently than one where vehicles are very precious. You can only expand your flight envelope very gradually when you are worried about losses; if vehicles are free, you can move to full-scale tests as soon as you have even the slightest chance at success. You can run risky experiments that tell you things regardless of the outcome.
I think SpaceX will see a benefit even after a single reuse of their system. It will be small–perhaps less than 5%. But that’s enough to give them an even higher leg up on their competitors than they have already. Lower costs mean more flights; more flights means they can reuse boosters without shrinking their factory (and losing whatever economies of scale they’re achieving).
There are of course limits to this approach. Shotwell has said they expect a ~30% cost reduction. That is obviously not the order-of-magnitude reduction one might hope for, but is significant enough that their competitors have to respond to it.
One cannot run before one walks, and I believe the same is true here. I don’t think it’s possible to have a (successful) full reusability program before one has a partial program. The current approach allows SpaceX to learn the necessary lessons before embarking on a more ambitious program. As Stranger says, rocket reusability is not like airplanes. It is its own thing and requires its own lessons to be learned. We know what a reusability program shouldn’t look like, but no one yet knows what it should look like. Nor does anyone yet know how to run a company that reuses rockets. SpaceX is the closest here, and once they close the loop I think they’ll make rapid progress.
The Mathematic study isn’t presented as the end all, be all of studies on reusability; it is merely the first (that I’m aware of) to evaluate reusability on a cost effectiveness basis, and while it is specifically focused on the concept that became the Space Transportation System, subsequent studies at the same level of fidelity (again, those I’ve seen or worked on) have not contradicted the basic assumptions. Orbital studied and promoted reusability for years, largely championed by Dr. Elias and the other developers of Pegasus, so it isn’t that he has an axe to grind against the concept of reusability.
A 2 km/s velocity at MECO and turnaround is a “limited…burnout velocity” compared to that needed for MEO and GEO missions. We don’t actually know how economic the current Falcon 9v1.1 Upgrade is because SpaceX, as a privately held company, is not obligated to publish per launch costs. We can observe that even before the “Full Thrust” upgrade and the use of densified propellants, the advertised bare manifesting cost for a Falcon 9 flight went from US$35.0M for 10,450 kg to LEO (2011) to Falcon 9 (v1.1 pre-Upgrade) US$61.2M for 13,120 kg to LEO (2015), a growth per unit mass of 40% and does not reflect any “extra” services such as payload contamination protect beyond 10k, extra coupled loads analysis cycles, payload contamination avoidance maneuvers, extended coast, access to manufacturing and integration test data to perform mission assurance analyses, et cetera. I would agree that limiting reusability to just Stage 1 simplifies the problem versus having to protect a stage against orbital-speed reentry conditions, but it still doesn’t follow that this will be reflected in a large savings in launch costs, which again are predominately the labor in handling, testing, and integrating the launch vehicle rather than the invested cost in the hardware.
If there is an advantage to reuse, it is in being able to handle higher flight rates and thus amortizing the fixed costs of maintaining facilities and paying for ground labor unassociated with a particular launch (e.g. all the stand-down time and training, as well as facilities maintenance and repair) rather than a direct savings in not having to build another stage. Unless SpaceX can make a stage that is so robust it can go through many flights with minimal refurbishment–and there is no evidence that the current Falcon 9v1.1 stage 1 can do that other than the examination of the two successfully returned stages after a single flight–the cost savings justification just doesn’t hold water. The ~30% cost reduction is new to me (and while optimistic is within the realm of rationality); previously, I’ve seen both Musk and Shotwell claim that reuse would allow them to relaunch the Falcon 9 for US$8M-10M, which I just don’t believe is possible with the current architecture.
This isn’t to say that either reusability or dramatic cost savings aren’t achievable, even with (mostly) existing technology, but to do so will require the abandonment of some widely held assumptions about space launch vehicle design and operations. One of those assumptions is the long slender tubular shape of a space launch vehicle, which is inherited from silo-launched ICBMs and air-transportable IRBMs. For those applications, fitting inside of an underground silo or a cargo plane fuselage is the most important factor governing the envelope, but it doesn’t apply to launch vehicles in general, and it imposes a form that is extremely mass inefficient, structurally and dynamically weak, limits payload shape and configuration, and provides very limited opportunities for radical improvement. The argument for maintaining this shape in new designs is transportability (so it can be moved by commercial over-the-road transport) and manufacturability (current methods for launch vehicles favor building round tubes) but is by no means the best form for a launch vehicle. A squat conical or pyramidal vehicle provides many opportunities for improvement but has been dismissed because it is too different from current experience or expectation. (Chrysler Aerospace actually proposed a squat conical SSTO with a modular base plug aerospike propulsion system for STS, but was essentially rejected out of hand versus the more ‘conventional’ thrust-augmented winged shuttles that NASA as originally envisioned by Sanger and promoted by Dornberger.)
The trading of highest possible performance for greater reliability and operational simplicity is also something the aerospace community has been reluctant to embrace even though this approach has worked pretty well for the Russians, and taken to the extreme (building pressure-fed rockets to shipbuilding tolerances) offers a genuine reduction in the amount of effort needed to integrate and test the vehicle prior to flight. The counterargument is that these vehicles will be too large to handle on land, which just indicates that the obvious solution is to build them in shipyards on barges and tow them out to sea for launch (or in the extreme case, dispense with the barge and make the vehicle capable of launching directly from the ocean). Although not without issues in dealing with seawater intrusion to the payload, it would allow for a larger volume of flights with almost negligible concerns about interference with air traffic, shipping, or ground hazards even without reusability. And vehicles built to shipbuilding tolerances could also be readily designed to withstand the impact forces of reentry and ocean landing. The cost is a dramatic increase in the amount of propellants required, but since common propellants are a fraction of a percent of current launch costs it is a very favorable trade. I have a family of toy models investigating various methods of cost reduction and design changes to trade performance for simplicity and robustness, and while none of them represents a final launch vehicle design they all indicate that this trade is favorable in terms of cost and launch throughput.
To be fair, SpaceX is making efforts to simplify processing and payload integration. Horizontal integration of payloads and rollout (which they do as the standard operation under their basic manifest cost) is a surprisingly big labor savings that also reduces hazardous lifting ops and can facilitate rapid response in the case of storms or processing problems discovered late in integration flow. So they are taking some good steps to increase efficiency.
Not that much less. For the SES-9 launch, the MECO velocity was ~2.3 km/s. It was their largest payload to date to GEO. They did attempt a barge landing, which was unsuccessful, but the stage did make it to the barge, indicating that reentry was not the issue. And it slowed down enough to not cause dramatic damage.
The 30% is from some comments by Shotwell here. Shotwell said it was too early to set precise prices for a reused Falcon 9, but that if the fuel on the first stage costs $1 million or less, and a reused first stage could be prepared for reflight for $3 million or so, a price reduction of 30 percent – to around $40 million – should be possible.
I agree that they were more optimistic early on, especially when they had that CG video going around with second-stage reuse. They’ve backed away from that and said that the Falcon 9 will likely never have second-stage reuse, and that it requires some rearchitecture.
Yeah, I’m really curious about this too. SpaceX does glean some advantages to flying road-transportable craft. As you say, a different form factor requires sea transport. SpaceX will also have to switch to sea transport for their BFR vehicle regardless of form factor. It seems that once they’re committed to moving pieces around by barge, they may as well go whole hog and choose a squatter form that has better tank efficiency and better ballistic properties for reuse.
Of course, I’d also like to see a Sea Dragon built. The vernier engines were as large as F-1s! And there is something amusing about building your rocket in a shipyard and priced per pound.