Now that I’ve wrapped up my Orbital Access Methodologies series, I wanted to share some thoughts about the business and market development side of reusable space transportation. Some of this may be old-hat for many of you, but I figured there are probably some who will find this useful and interesting. I was originally going to write this up in a single post, but I decided it would be best to split this up into a series of articles like I did for Orbital Access Methodologies.
Why Flight Rate Matters: Fixed and Marginal Costs
A common conclusion found in many studies on reusable space transportation is that RLVs need at least 50 flights per year to make economic sense. While there are a lot of assumptions that go into the specific number, the basic idea is that there’s some minimum number of flights you need to make the economics work for RLVs. RLVs typically have higher development and fixed costs, but much lower marginal costs than a similarly sized ELV. The more often you can fly an RLV, the lower your overall flight cost will be because each flight’s percentage of your fixed cost goes down as flight rate goes up.
Fixed costs are those things you have you have to pay for on a monthly or yearly basis regardless of how often you fly. Stuff like facilities, payroll, overhead, capital equipment amortization (including the air frames), etc. Marginal costs on the other hand are what it costs to add one more flight to your manifest. This is stuff like the cost of replacing any expendable components, maintenance and refurbishment costs for the vehicle, launch insurance, mission specific engineering, propellants/consumables, any touch labor not covered under fixed costs, etc.
The fewer flights per year you have, the larger each flight’s share of the fixed costs will be. In fact, at a low enough flight rate, reusable vehicles can sometimes end up costing more per flight than a similarly sized ELV. See the Shuttle as an example of this situation.
Terrestrial airlines are also in a similar situation. They also have high enough fixed costs that they are only able to stay economical is by keeping their vehicles flying as often as possible.
Now, the exact number of flights necessary for a specific RLV to start running in the black will vary a lot depending on the details. The commonly quoted magic number of 50 flights per year mentioned above depends on a lot of assumptions, not all of which may be valid. One typical assumption that may not be valid is that the development cost for an RLV will be much higher than an ELV. This may be true if you develop an RLV using the same processes you would use for an ELV, but there are arguments that there may be ways to use the fact that the vehicle is reusable to actually make development cheaper. Regardless of how the numbers come out though, the fundamental reality is that RLVs need larger flight rates than most existing ELVs see in order to make economic sense.
Achieving Higher Flight Rates: Launch Supply and Demand
In order to achieve higher flight rates, you need both a vehicle capable of high flight rates and enough demand to buy all those flights. You need both parts of the equation in order to make the business case close. The Shuttle is a good example of what happens when you try to do an RLV that doesn’t meet either of those criteria. The Space Shuttle fleet was incapable of coming anywhere near the 50-100 flights a year they needed to get to be economical, and there also weren’t 100 flights per year worth of payloads that the Shuttle could fly. The end result was a very expensive RLV system that flew as infrequently as ELVs and ended up costing several times as much.
Attacking the supply side of the problem mostly involves technology development and maturation. Operability is one key to economical RLVs–If it takes you more than a week to turn around your vehicle, there’s probably something on it that isn’t really robust enough for prime time. If your TPS system for instance takes hundreds of people weeks to inspect, maintain, repair, and qualify for reflight, it’s probably too dangerous and marginal to use on an operational system. It is unclear if the technology we have currently is up to this task, but this is an area where suborbital RLVs are having an important impact. The key will be finding technologies or combinations of technologies that allow you to make engines, TPS, and other systems robust and low-maintenance while still maintaining enough performance to make the rest of the design close.
Now, while us rocket nerds love debating things like the technical aspects of making a low-cost, robust RLV, the demand side is probably even more important. One of the common refrains you hear from industry veterans about RLVs is “where’s the money going to come from to pay for enough payloads?” As Rand Simberg used to say in his usenet tagline back when I was first getting into the whole space thing “Extraordinary launch vehicles require extraordinary markets”. While focusing on the technology side of the problem, and using an “if you build it they will come” approach to handling demand might be a great deal for those who end up buying your company’s bankrupt corpse, it’s probably not the route that ought to be taken if you actually want to make a profit for your original investors.
So, what kind of payload types are best suited for RLVs? Why don’t I start out first by talking about an important payload type that probably isn’t. Satellites–at least as they are done today–are probably not a good fit, for several reasons:
- Not very many satellites are launched per year
- Many of them are going to higher altitude destinations (high LEO, MEO, or even GEO), which most RLVs would have a hard time reaching without an expendable kick stage
- Satellites tend to go into a wide variety of orbits, including a wide range of inclinations, apogees, and perigees. This requires more mission-specific engineering, more time for regulatory compliance such as getting launch licenses (especially if the RLV is using an inland spaceport as may well be the case), and all of this generally means a lot more time between an order and a flight. This may result in a higher margin for satellite flights, but the lead times will be longer
- Satellites tend to require a lot of handholding. Lots of testing and unique integration work that may be not be applicable to any other satellite.
- Most existing satellite developers are very conservative. While price is a factor, perceived risk and insurance costs are also very important.
- Satellites don’t start showing significant demand elasticity with lower prices until the prices have dropped substantially from existing levels.
- Even if the demand does pick up, it will still take several years for that demand to ramp up, since satellite design/build/test programs can often last a long time.
Now, these problems aren’t impossible to solve. Given enough time, the market will adjust to new capabilities, and there may be ways to get higher launch demand out of existing satellite customers, using techniques like the ones Dave Salt has proposed for GEO launches (launch the propellant for the GTO and GEO insertion burns separately from the satellite itself using multiple launches and orbital rendezvous/propellant transfer). But the reality is that for the near future, satellites really aren’t that great of a market for reusable launch vehicles.
When thinking of what the ideal payloads would be for RLVs, I could think of a couple of possible criteria:
- Doesn’t need a lot of handholding, integration work, or mission-specific engineering
- Doesn’t cost tons more than the flight would
- Provides good demand price elasticity
- Is divisible into chunks small enough to be carried by light RLVs (less than 5000lb payload)
- Provides demand for flights on a regular and consistent schedule
- Provides demand for many flights to the exact same destination (for example a station in a resonant orbit).
- Is sufficiently self-similar to allow for many flights reusing the same interfaces, and the same operating procedures
- Is tolerant of risk
- Doesn’t require several years lead-time to develop the payload
As I see it, there are three main types of “RLV Friendly Markets” that I think meet these criteria: people, propellants, and “provisions” (ie light cargo that aren’t self-contained spacecraft or satellties). I’ll give a few thoughts of each of those in the following parts of this series.
Jonathan Goff
Latest posts by Jonathan Goff (see all)
- NASA’s Selection of the Blue Moon Lander for Artemis V - May 25, 2023
- Fill ‘er Up: New AIAA Aerospace America Article on Propellant Depots - September 2, 2022
- Independent Perspectives on Cislunar Depotization - August 26, 2022
–This may be true if you develop an RLV using the same processes you would use for an ELV, but there are arguments that there may be ways to use the fact that the vehicle is reusable to actually make development cheaper–
Comment?
Adam,
Unfortunately, that’s a topic that would take an article or two or three, not just a quick comment. 🙂
~Jon
Adam,
RLV development can use baby steps that don’t require everything to be perfect all the time and on time. An RLV can substitute testing for intensive engineering. It is often quicker and cheaper to whomp up a test article than to spend thousands of engineer hours trying to get that article perfect before it is built. Testing will let the real world teach you things that the simulations never will.
You only have to miss one minor thing on an expendable to lose the vehicle and have to guess the problem before launching the next one. By building the vehicle both reusable and fault tolerant, you get it back in condition to fix the item that caused the problem, and nothing else. While an ELV company is spending months troubleshooting every component that might have caused the problem, they are burning through investment money and interest on that money. Delays in any business are expensive.
The mass ratio considerations are at least as fundamental as those of flight rate.
1) You have to take (RLV+payload) all the way to ~7 kps to leave (payload) in orbit. This is the logical consequence of the “R” in RLV, but it’s still parasitic mass from a “how much do I have to get off the ground?” PoV.
2) it’s hard to imagine any component of a RLV weighing less than its counterparts in a one-use-only ELV — and of course it needs TPS and wings (and/or parachute or lifting-body airframe), and landing gear (or retro rockets or airbags or…), which have no counterpart in ELVs. More non-payload, non-thrust-enhancing mass.
So an RLV has to fly lots not only to amortize higher up-front costs, but because each flight puts less in orbit per kg of GLOW.
argh, close the tag on “less,” pls
Monte,
Nice to hear from you! I’m going to have to figure out how to rope you into being a coblogger here one of these days. 🙂
While I basically agree with your two points, I don’t see how that supports your first sentence. To me, if anything, they reinforce the point I’m trying to make about high flight rate–that high flight rates are absolutely crucial for RLVs. Both being able to achieve them technically, and being able to find enough demand to justify them. But yeah, the technical demands of an RLV also contribute to the need for higher flight rates to make the economics work.
That said, two quibbles.
First regarding your first point, every orbital rocket has to take the rocket + payload into orbit (though for TSTO ELVs or RLVs you only have to haul part of the rocket all the way). The difference with an RLV is that you have more parasitic mass that you have to take up. But even a Centaur or a Falcon still has to drag its upper stage all the way to orbit. They do get to ditch their fairings on the way up, but the fairings are only part of the upper stage mass.
Second, while the RLVs do have subsystems that I agree with you, will probably weigh more than comparable subsystems on an ELV, for the markets I’ll be talking about in the next parts, an ELV payload would need to provide most of that parasitic mass anyway. For instance, if you’re flying people on an ELV, you also need TPS, recovery gear, etc. Which means that for the markets I’ll be talking about, the penalty for RLVs is much lower than for RLVs launching satellites.
~Jon
Right, the first sentence should have been “mass ratio considerations accentuate the need for high flight rates .” And yes, ELV final stages do have to get some non-payload mass to orbit. (That gets into a gray area if it’s a Centaur or Saturn V third stage which still has some function for escape or a higher orbit.) My intention was to clear away two forms of simplistic thinking:
1) the “mighty Saturn nostalgia” that grew up as part of disillusionment with STS: you know, comparing STS’ LEO payload unfavorably to that of an S-V, ignoring the 100,000 kg of the orbiter.
2) the “throw away the jetliner” argument for RLvs: “Just think how expensive air travel would be if you used a 7X7 only once.” The people who advance that as a decisive argument usually omit the footnote: “In this analogy the 7X7 capable of a round trip would be a very significantly different vehicle with greater development cost, smaller payload… and oh yes, you can’t refuel at the destination, so the return trip uses a kinda hairy alternate mode, demanding a whole ‘nother set of features that are dead weight on the first leg…”
I know these are stunningly obvious, but I at least have to keep them near the front of my mind to appreciate just how hard cost-effective RLVs will be. When people sneer at ELVs as “disintegrating totem poles” or at “the artillery model,” there’s an implication that everyone from Peenemunde to KSFC and Baikonur must have been retarded not to simply follow the reusable aviation model from the beginning; just listen to Eugen Sanger, just keep pushing the X-15 higher and faster, yada yada.
NO, I insist; they were fine engineers who hated to throw away precision hardware every bit as much as anyone today — but getting a payload to orbit was just barely possible that way.
Getting it to orbit with something that can return for re-use really is a lot harder. Doing so often enough to make the development worthwhile is harder still. And doing so in a context in which you have to create most of your markets as you go, and carry with you most of the “destination” facilities that are taken for granted with aviation, is even harder than that.
Monte,
Agreed. It is a tough nut to crack, and if I’ve given any impression that I thought it wasn’t, I apologize. My goal in this series was just to point out what sort of markets I think RLVs ought to be focusing on, and what the implications of that are for RLV development.
~Jon
if I’ve given any impression that I thought it wasn’t [tough]…
You haven’t, which is why I like it here. 🙂 I’m insistent about the things that make space hard, not to discourage efforts but because fifty years of underestimating the full scope of the challenge has actually retarded progress. It has done so by inspiring “great leap forward” programs and designs that almost always fail, instead of incremental improvements that accumulate. It has done so by obscuring the great difference between thrilling stunts like Apollo and sustainable capabilities, so ; .
Most insidiously, by feeding great and premature expectations, it has left too much of the public thinking “Space? yeah, we tried that back in the day — didn’t pan out, kinda like jetpacks and flying cars.” And it’s left too many space fans taking it for granted that we should be much farther along — so they look for scapegoats to explain the gap, or the One True Design or One True Program or One True Free Market Bonanza to close it.
The best example of why an RLV might cost less to develop than an ELV is SpaceX’s Falcon 1. Causes of consecutive launch failures (at ~$10M/ per): a corroded B nut, sloshing fuel, and a separation transient. 2 & 3 would have been easily detected in envelope expansion flight testing, and the first may not have been an issue if you could control the first stage return (therefore not requiring a launch in the middle of the ocean).
Jon,
May I make a suggestion for one further entry into your Orbital Access Methodologies posts? Have you considered a SStS/MXER combination? (That’s Single Stage to Sub-oribit / Momentum eXchange – Electrodynamic Reboost). Launch a payload on a suborbital trajectory and where it rendezvous with and is captured by the end of a rotating tether in LEO. The rotation of the tether then picks up the payload and transfers it to a higher orbit. The transfer of momentum from the tether to the payload is then replenished by pushing current through the tether to create a Lorentz force against the Earth’s magnetic field. The same process can be used in reverse to deorbit a payload, thus reducing the requirements for the performance of the TPS.
I don’t think we can really move out into the solar system in large numbers until we learn to master conservation of momentum and energy. Conventional rocketry may be required to get us up out of the gravity well and atmosphere of the Earth, but there’s no reason why we should continue wasting valuable resources to accelerate and decelerate payloads once in space.
If you’d like more info, take a look at the Tethers Unlimited website (http://www.tethers.com/MXTethers.html).
Eric,
I’ve been aware of MXER tethers for many years, but I’m still not sure if the idea makes as much sense once you look at it in detail. I’m pretty confident that it could eventually be made to work. On the one hand, it could potentially allow for much smaller launch vehicles being able to put sizeable payloads into orbit….but it’s the details that worry me. Not just the technical “will it work at all” details, but details like:
-can we get a renedezvous system between the suborbital vehicle and the tether that has a high reliability and safety (ie probably at least 99% reliable)
-can we actually make a MXER tether with a long enough life, and enough robustness to last long enough to make the business close
-how long will it take to “recharge” a MXER tether. Can it be used frequently enough to really make sense? Remember, true RLVs may usher in flight rates that are substantially higher than anything we’ve seen to date in the space launch field.
-can we make the orbital dynamics work so you can use a MXER tether to provide first (or second or third) orbit rendezvous to orbital facilities?
Stuff like that. It’s interesting, but I don’t think I actually have a lot to add to the conversation other than questions.
~Jon
Pingback: Selenian Boondocks » Blog Archive » RLV Markets Part II: The Black Aluminum Analogy
Pingback: Antimatters» Blog Archive » Rearranging a Constellation