A few days ago, I mentioned that some of my friends at ULA had passed me copies of some papers they were going to present at the Space 2008 conference a few weeks ago. The paper I want to discuss in this blog article, which can be found in Word Document form here, revolves around using mid-air recovery to reuse the propulsion section from the Atlas V first stage.
I first heard about this concept back around the beginning of the year, but other than a somewhat elliptical reference back in my second Orbital Access Methodologies post, I had held-back on posting until my friends had had a chance to publicly present their concept.
The Rationale
The basic logic behind concept is pretty simple. The RD-180 is derived from the Russian RD-170/171 engines that were designed for use on their Buran and Zenit launch vehicles. They were designed to be reused multiple times, and during the qualifcation test programs were tested to many times the length of a normal stage burn. The RD-180 represents most of the cost of the Atlas-V first stage hardware. It’s also made by a country that’s becoming less than friendly with the US, and it’s one of the key limiting factors for higher flight rates for the Atlas V. The tanks on the other hand are relatively cheap, bulky, and heavy. What this means is that, you can get a lot of cost savings from reusing the RD-180, even if you end up tossing the tankage each time around. The other great thing is that this is a modification that doesn’t cost a lot of capital to develop, and can be profitable at much lower projected flight rates than building a fully reusable vehicle of this size would. This makes it much more realistic in the near-to-medium term.
In order to reap those potential savings, you have to recover the RD-180 in such a way that minimizes the amount of refurbishment work that needs to be done between flights. That’s where mid-air recovery comes in.
The Concept
ULA's Proposed Mid-Air Recovery Scheme for the Atlas-V Propulsion Section
Their proposed scheme would partition the propulsion section of the Atlas-V (which is already largely packaged in this manner anyway) into a recoverable module that could separate from the main tanks using some sort of linear shaped charges. This module would include a hypercone-style inflatable hypersonic decelerator, and some parachutes. At the end of the stage burn, the module would close itself off from the tanks, some purges would be performed to make sure there aren’t any contaminants in the engine, then the shaped charges are blown, severing the tank and separating the module. The hypersonic decelerator slows the vehicle down and keeps heating and reentry forces on the engine module itself down to a minimum, until the vehicle is slow/low enough for the main parachutes to open. As the stage is coming down, a helecopter that is positioned just outside the potential landing zone moves to rendezvous with the slowly falling stage. Using 3GMAR (Third Generation Mid-Air Recovery) techniques developed and refined in work that Vertigo, Inc did with ULA, the helicopter can gently capture the propulsion module by the parachute, and then haul it back to a specially designed recovery cradle located either on land or on a barge out at sea.
By doing this, the recovery loads on the propulsion system can be kept much lower compared with splashdown recovery, and sea water contamination concerns can be greatly reduced or eliminated. This means that most of the stress the engine sees in operation is actually due to the engine firing, not recovery or ground handling. This makes it much easier to reuse the relatively complex staged-combustion engine without the need for extensive refurbishment. In the paper the ULA guys explain a bit about what refurbishment would be needed, and based on that they estimate that you get the most bang-for-the-buck by going for 3 reuses per engines.
There are all sorts of other interesting technical details in the paper, so I’d highly recommend reading it yourself.
Ramifications and Further Thoughts
There are several potential benefits from the proposed scheme:
- Lower Cost: The obvious benefit of this scheme is that it helps reduce one of the key hardware cost items in an Atlas-V launch. While for a single-core Atlas-V launch, the savings would be relatively modest, they would be much larger potentially for Atlas-V Heavy flights or Atlas-V Phase II flights (if either of those ever become a reality).
- Higher Sustainable Flight-Rate: Another benefit is that this allows for higher flight rates of the Atlas-V without requiring drastically higher engine production levels from the Russians. I’ve heard hints that the Russian’s would have a hard time cranking out much more than 10-15 RD-180s per year without spending a decent amount of cash on expanding their production facilities and hiring on more people. If each engine can be used three times, all of the sudden engine manufacturing rate completely supply ceases to be a constraint on launch frequency.
- Less Foreign Dependence: Right now, ULA stocks something like 2-3 years worth of RD-180s in the US at all times. The idea being that if the Russian Government decided to stop allowing ULA to buy RD-180s, that inventory would give the US enough breathing-room for Pratt & Whittney Rocketdyne to bring an Americanized version of the RD-180 into production. Honestly, I’m not sure how seriously Russia takes that threat. There’s enough tricky metallurgy and other issues with making an Amercanized oxygen rich staged-combustion engine like the RD-180, that I’m not sure the Russians really believe we could have an in-sourced version ready to fly before we’d run out of engines. Now, if that 2-3 year stockpile became a 6-9 year stockpile because of the ability to reuse the engines, all of the sudden the threat of in-sourcing becomes more real. The incentives then make it much more likely that Russia would continue to sell us RD-180s at a reasonable price, because the threat of us in-sourcing production is more realistic.
There are also some ways of continuing this line of thought that I think might be worth investigating down the road. A big one is the potential for using mid-air recovery to reuse the Centaur upper stage. The 3GMAR technique is capable of snagging payloads of up to 25klb with existing US helecopters, and the Centaur stage only weighs about 4500lb dry. Now, there are all sorts of caveats on that idea. For one, recovery of a first stage on a suborbital trajectory is much less demanding than recovery of an orbital stage. You need much more serious TPS for the upper stage, landing dispersions will likely be bigger, and the stresses on the stage may be more severe, while additional dry-weight on the upper stage has a much larger impact on payload performance. Not to mention the fact that recovering a modified Centaur stage that goes into GTO (which is quite common) will be even harder, if not impossible.
That said, the idea isn’t entirely without precedent, or even that wacky when you think about it. As it is, ULA was already studying with SpaceHab the ARCTUS concept which would’ve had a Centaur-derived COTS cargo delivery system that could be recovered from space using something similar to the hypercone/MAR concept proposed for the Atlas-V propulsion system. In fact some of the tests they discussed in the paper were tests for that very program. Also, SpaceX is planning on trying to recover the Falcon 9 upper stage. When I spoke with Elon back in January, he hinted that the approach they were looking at at the time was actually pretty straightforward–using ablator panels on the front and some of the sides of the upper stage, followed by parachutes and splashdown recovery. Now, there’s a huge string of if’s between where we stand and SpaceX successfully recovering a Falcon 9 upper stage, but between that and the work done on ARCTUS before the COTS re-bid process was over, it at least looks feasible to scale that up to recovering a Centaur stage.
The Centaur is one of the major cost items on an Atlas V flight. If both it and the first stage propulsion system could be reused a couple of times each, it could make the Atlas V substantially cheaper for moderate flight rates. Also reusing the Centaur looks even better if they ever go to the larger Phase 1 Centaur upper stage concepts. The largest one of those dicussed, a six-RL10 version with ~180klb of propellant has a predicted dry weight of about 15klb. That would leave quite a bit of weight for a recovery system. Of course, that’s assuming that recovering such a big fluffy object wouldn’t pose other problems for the mid-air recovery effort–which probably isn’t a safe assumption.
What does this all mean?
Probably not very much unless either NASA is forced by Congress to ditch Ares-I and focus on making Orion flyable on an EELV, or unless Bigelow manages to close a deal with them for a large number of flights. The current flight rates are so low that the savings provided wouldn’t pay for the added development cost. But what it does mean is that if demand starts increasing, whether due to ISS resupply, Bigelow, propellant depots, or orbital tourism, or several of the above, Atlas-V still has plenty of room for evolution to keep competitive. Would such a recovery/refurbishment scheme get rid of the need for real RLVs? No, but neither would the existence of real RLVs get rid of the need for medium lift launch vehicles.
I’ll get into it more in a later article, but I really think small RLVs and “recoverable” launch vehicles (including both the Atlas concept discussed as well hopefully as Falcon 9) are actually complimentary technologies. But more on that later…
Jonathan Goff
Latest posts by Jonathan Goff (see all)
- Fill ‘er Up: New AIAA Aerospace America Article on Propellant Depots - September 2, 2022
- Independent Perspectives on Cislunar Depotization - August 26, 2022
- Starbright Response to ISAM National Strategy RFC - July 2, 2022
Would recovering an Atlas V heavy mean recovering three separate modules? I’m guessing they’d all be recovered at the same spot at the same time, which sounds dicey.
Tim,
Actually you’d only be recovering two of the modules at the same time. On the EELV heavy designs, the core is throttled so that the strapon boosters burn out much earlier. In fact they tend to burn out much closer to home than a typical 401 stage, which makes recovery easier. Sure you need two helicopters at the same time, which would be trickier, but not entirely unrealistic.
For the core, you might just let it go since it will be splashing down much further downrange (think just off the coast of Africa)–just make sure that it’s always flying an engine on its last (4th?) flight.
~Jon
“all of the sudden engine completely supply ceases to be a constraint on launch frequency.”
?
Benefit #2–the higher flight rate-and benefit #3–less foreign dependence–are in tension.
Anyway, great review. Very helpful.
Adam,
Sorry I clarified that sentence (it wasn’t well worded). Basically, if you can reuse the engines a few times each, the throughput rate of the factory no longer ends up being the limiting factor on flight rate.
And yes, you are right that to a point benefits 2 and 3 are partially an either-or proposition. If the flight rate jumps up a lot, then the inventory won’t last as long (you’ll just get more flights out of the inventory). However, they said that it would be possible to extend the reuse to even more than 3 reuses by doing more testing and refurbishment work. Once you’ve made it possible to reuse the engines, even if the flight rate goes up, you might still be able to get more uses out of them (and hence extend the duration of your inventory), which means you can still get some of benefit 3 even with benefit 2. The only reason they picked 3 is that in a world where you can get RD-180s from the Russian’s that’s where the point of diminishing returns is.
~Jon
It sounds like three reuses per engine is an optimization point based on purchase price of a new engine vs. recovery costs to reuse an engine. This also gives some price leverage in addition to insurance against the Russians cutting off the supply. If the Russians try to raise their prices it becomes economical to reuse the engines more times.
Bob,
Good point.
~Jon
The RD-180 engines mass about 11,000 pounds, put the extra structure, heat shield parachute etc etc you’ve got about 20,000 pounds. What kind of super duper helicopter are you talking about?
Rod
Rod,
Sorry, I didn’t realize they didn’t go into as much detail in this paper on the MAR aspects as they did in their previous paper (you can find it here on the Vertigo website). This is a few years old, but basically they were saying that a Boeing CH-47 Chinook, Sikorsky CH-53E Super Stallion, or Russian-built Mi-26 Halo (or related variants) could all provide at least 22,000lb of MAR capacity, with some of the Mi-26 options providing even more. Read that MAR paper if you have the time–it’s really interesting.
~Jon
About recovering multiple modules even though the staging moment is the same the timing of first the drogue chute and later the parafoil can be staggered (and I would think it should be if nothing else as a safety precaution). also the use of parafoils enables remote or automatic control of the path so as to keep at least a minimum distance.
I read “engine completely supply ceases” as “engine supply completely ceases”.
Maybe SpaceX (and others for all I know, for example AirLaunch and t/Space but some of the stuff could involve all those aiming at suborbital as well) will be interested in a joint venture with the ULA on the concept? Might possibly also cost less to do some of the testing/validation on Falcons first.
Strong bias: DARPA was made for stuff like this, get them on board and put them in charge.
I believe that during EELV competition, it was stated by LM that they would develop a domestic production plant for the RD-180. I take it that this has not happened. Why didn’t they? Is there someone at LM smirking wryly and saying: “Stupid you, you believed us.”
If they were to get this up and running, could the same infrastructure be used for a towback first stage from a different company?
Excellent idea! Modern parafoils can fly instead of just falling, and it’s very easy to catch them with a helicopter. That was demonstrated by the Spacehab Arctus team in miniature hardware.
If a fixed wing craft or an Osprey can be used, it can be based far away from the recovery point as it has lots of speed. I haven’t read your article yet completely to check out how the numbers work.
This all is also very good testing material for the suborbital companies. You’d want lots of flights from high altitude and speed to test the transition between the modes (ballute,plain,drogue,main)
John,
You mean a mid-air recovered tow-back stage? Possibly. The problem is that mid-air recovery itself isn’t cheap, so it’ll set a firm lower limit on the economics of such a system. It makes a lot of sense for bigger launchers, not as much sense for others. But what did you have in mind?
~Jon
GL,
The ARCTUS stuff was done in conjunction with ULA and Vertigo–i.e. it’s the same guys and part of the same overall program. At least that’s my understanding.
~Jon
I’m not as comfortable with air-capture of a 20K lb mass, for either fixed or rotary –even if heavy lifters, the effect of a sudden addition of that much weight on an aircraft’s structure could put a literal crimp in the operation.
BTW, I recall that LM — back when they were MM — once had plans for slapping some chutes and maybe floatation devices on an Atlas II half-stage, to see it they could find, flush & re-fire the engines, despite the salt-water dunking. This was being given some very serious consideration during the olden days of the ALS almost-a-program (engine recovery was popular for alot of companies back then). Anyone know whether they still carried it out in some form?
Brian,
The mid-air recovery stuff isn’t just theoretical–they’ve been working with Vertigo on this for years, doing tests, verifying claims, etc. The entire point of the 3GMAR system (see a comment #9 above for a link to the Vertigo paper) is that it *doesn’t* put a sudden load on the airframe. If they can get this to work, it’ll be far cleaner and easier than ocean recovery, according to the guys doing the research. And their logic is fairly persuasive. Maybe making an Atlas II stage water recoverable wouldn’t have been that hard. But an RD-180? Not likely, IMO.
~Jon
If Atlas is allowed to ignore the requirements for a domestic engine and the Russians bite the Air Force in the … then they get what they deserve.
Jon,
During development and early operations of VTVLs, it might make sense to have a backup chute on board with chopper based in the expected recovery zone. Similar to having paramedics available during certain tests. You hope they are wasting their time and your money. It seems possible that a single save might pay for many deployments early on. But only if someone else has already paid for the system and is recovering some of their costs by providing another service.
John,
You know…that’s not as crazy as it sounds at first. We’re nowhere near the point where something like that would help, but it might just be worth looking into. I hope you don’t start actually charging for your good ideas anytime soon….
~Jon
Dave,
While the original contract terms included Americanized production of the RD-180, the terms AIUI were renegotiated when it became obvious that the flight rate that was used to justify the EELV program wasn’t going to happen. Basically, between the collapse of the commercial comsat market back in 2000 and the whole flap over that former LM person who helped Boeing win most of the government business, the Air Force changed the contract terms. That said, it’s commercially the most beneficial if they can buy the engines from Russia, and having some leverage that makes the Russians less likely to pull the plug is always beneficial.
~Jon
Curious comparison between this LM paper and SpaceX plans. These (professional) people discuss sealing every aperture, including using airbags in the engine bells, even though the plan is to keep the hardware out of the ocean, setting it down gently onto a prepared, dry carrier. SpaceX thinks they can drop stages into the ocean, fish them out of the salt water, rinse it off, and re-certify some of the hardware, perhaps the entire stage, including the heavy but not-so-expensive tanks. Quite a difference.
Comga,
You also have to take into consideration the differences between an RD-180 and a Merlin. The RD-180 is a oxygen-rich, staged-combustion engine run at the highest chamber pressure of any rocket I know of. The Merlin by comparison is a gas-generator, and runs at much lower pressures. It was also designed from the ground up with water recovery in mind, and therefore could choose materials and may very well include features that would ease reuse.
Now, I do agree that the LM approach seems more thorough, and I also think it’s more likely to result in an easier refurbishment job. But we also don’t have all the details of SpaceX’s approach, so reserving at least some judgment seems in order.
~Jon
Undoubtedly, there are differences in the engines, and you have detailed more than I know in one paragraph. There have also never been the release of as many details of the SpaceX recovery method to my knowledge, although they have supposedly had them installed on all three flights. However, the extreme divergence of approaches is remarkable for what is essentially the same problem. Intact recovery vs engine module, sealed vs unsealed, ocean impact vs low impact MAR, salt water dunk vs dry shipdeck. Not to mention a pretty detailed outline of a refurbishment plan for the RD-180 vs “we will see what we recover”. (At least that was the statement in about 2005.) It is just fascinating.
Comga,
I’m a fan of people trying different approaches to see which really works best in practice. I’m not a huge fan of splashdown recovery, but at the time they started, you do have to admit that it was the most mature method for recovering booster stages. Not to mention that Vertigo’s 3GMAR technique came out after SpaceX was already underway. And while I personally agree with you that the MAR approach appears more promising, where doable, it’s also unclear if LM will ever get around to actually implementing this. They’ve got limited funds, and without enough demand to justify the expenditure, it’s unclear that it will ever happen. At least SpaceX’s approach stands a reasonable chance of getting tried in the near future.
That said, the idea of having a rocket land any other way than on plumes of supersonic flame or horizontally using wings just seems somewhat barbaric. 😉
~Jon
I am not really advocating. In fact, it looks like the LM MAR approach is on the high side of sufficiently complex (as SpaceX’s seems on the low side). As such, with having to break the existing booster into two parts, tankage and recoverable propulsion module, the very complex sequentially deployed cone and chutes, the LM proposal has a barrier to entry that they may never be able to surmount.
You may have read my post (#175 in http://forum.nasaspaceflight.com/index.php?topic=13507.165) wondering if perhaps SpaceX wouldn’t have been better served if their primary initial goal was recovery instead of orbit. Perhaps if they had a more robust recovery method, one that somehow attenuated the shock of splashing into the ocean with something to protect the engine, they could have gotten some of the build-a-little, fly-a-little of classical X planes, as you advocate in your discussions of recoverable first stages. Perhaps one or two of the proposed LM techniques could have enhanced SpaceX’s recovery to where it could be relied upon. Imagine a series of flights with under-fueled stages, flying short trajectories and gently spashing down, then being hauled back to the launch pad. (Probably one better equipped than Omelek.) SpaceX certainly wouldn’t have any less results than they have today. Perhaps by next week their actual plan will be proved superior.
PS Do we have any idea of the mass of the F9 first stage? Wonder if it is more or less than 10 mT. The 9 engines must be <4mT. Most of the rest is ~25×3.6m of AlLi isogrid tanks. You can probably estimate that better than I can.
Comga,
Yeah, that’s somewhat similar to how my views on rocket development have evolved. With my original background in Manufacturing Engineering, my initial thoughts on rocket development were to mass produce big dumb boosters. But at some point, I realized that there are not only manufacturing flaws, but design flaws as well, and the only way to make sure you’ve eliminated the design flaws is to fly. It was at that point that I started realizing that having some sort of recoverable first stage would allow you to do that flight test program cheap enough for it to make sense….and thus the slippery slope begins until finally you realize that you really do want to just build an RLV.
Oh, and regarding the F9 FS dry mass, my best guess would be 30klb plus or minus. If you know anyone with an Mi-26 handy, it could work. Falcon 5 would’ve been light enough to mid air recover however….
~Jon
~Jon
Ah, the late, lamented Falcon-(magic number)-5. Do you recall Musk’s original statements on a five engine booster?
Comga,
I was always more a fan of the Falcon V than the Falcon IX. Four-to-six engines really is a sweet-spot for launch vehicle design. But they had a customer come along asking them if they could loft a 20klb payload, and they think they can find more business for an EELV class launcher than a Delta-II class one. Of course, if they do manage to get Falcon IX flying, they always could go back to a five-engine configuration if they upgrade the Merlin with TAN. I guy can dream can’t he?
~Jon
Was that customer other than NASA?
SpaceX now has a much more powerful engine ( 125 vs 71.5 klbf) so that an F5 with the Merlin 1C would have nearly the thrust of the F9 with the Merlin 1A engines, and less mass by 1.6 mT. But…whatever.
While we can dream, the most practical result might be for SpaceX to incorporate some of the LM concepts, shy of MAR, for improving the condition of the recovered F9 first stage. While dunking precision hardware in the ocean may be barbaric, it may yet be possible to make it workable, even if it is not graceful or thrilling.
Comga,
Was that customer other than NASA?
Yeah. From what I’ve heard, I think it was someone in the AFRL that got them to investigate doing an F9 sized vehicle.
SpaceX now has a much more powerful engine ( 125 vs 71.5 klbf) so that an F5 with the Merlin 1C would have nearly the thrust of the F9 with the Merlin 1A engines, and less mass by 1.6 mT. But…whatever.
I know. But part of the upgrade may have been in the works all along for the F9. Not sure though. Pure speculation on my part.
While we can dream, the most practical result might be for SpaceX to incorporate some of the LM concepts, shy of MAR, for improving the condition of the recovered F9 first stage. While dunking precision hardware in the ocean may be barbaric, it may yet be possible to make it workable, even if it is not graceful or thrilling.
Yeah. The first step is finding out if they can get the stage back at all, and what the actual loads are. I figure it’s going to take them several tries on recovery until they have enough data to know what they really need to do to make it work.
~Jon
I guess we have our answer. A successful rocket to orbit for SpaceX. Frankly, it is a relief, as well as a pleasure to watch.
Now we will see if:
The first stage can reenter in a controlled fashion from 200+ km altitude with just parachutes. (No hyper-cone.)
Whether they can find the descending first stage.
How much they can learn from the post-flight analysis of the engines.
Whether any hardware is usable after reentry, impact, and the salt water soak.
I did spend some time (mostly while running) thinking about how SpaceX could deploy something from the front end of the first state that would cover the engines and create something like the hyper-cone. We could discuss this off-line.
There is a statement on
http://forum.nasaspaceflight.com/index.php?topic=13507.570
“Anteres: Another site said that 1st stage recovery was not successful. Needs better TPS.” No basis cited.
Great…. Another wonderful idea stolen from me before I could get it out of my head and onto some sort of media…. I really MUST find a way to sheild myself from these obvious assaults by corperate and other un-licensed telepaths!
Seriously :o) (I can be, I’ve been assured of it :o)
This actually blends well into a ‘concept’ that’s been mulling around in my head. Jon, I’m hopeful you can devote some time to comment on an aspect of ‘design-and-engineering’ in regards to propellant tankage that I’ve been wondering about for years.
There was much discussion over the years on the merits and possibilities of using “expended” propellant tanks as temporary and permanent storage, living, and other facilities on-orbit. From the concept of the Atlas Space Station, through uses of various upper stages to the Shuttle External Tanks, (last most recent being the Space Island Group http://www.spaceislandgroup.com/home.html tank-derived stations) the idea has been to “re-purpose” already existing ‘resources’ on-orbit in the form of expended propellant tanks. Skylab was originally concived as a ‘wet’ station having orbited itself and the crew and materials to turn it into a spartan space habitat, the Space Cruiser http://www.astronautix.com/craft/spauiser.htm proposed by In-Space Operations (ISO) planned to use the 4th stage of a proposed Air-Launched Launch Vehicle as a series of “mini-stations” for the small spaceplanes to visit, similarly Interorbital Systems (IOS) orignally intended to use the expended tankage of their orbital 1.5 stage-to-orbit system for a tourist manned space station.
I’ve seen the idea proposed again and again, but it seems to me the concept is most often met with pretty hostile commentary and a lot of assumptions and rhetoric on why the concept won’t ‘work’ at all. However I’m confused the only “major” drawback I can find is the amount of on-orbit manual labor required to re-fit the tankage as liveable space, and of course the given fact that no such infrastructure is required by any space-faring nation so there is little interest in the idea.
My question I suppose then is this: Is the basic concept flawed or are there engineering, safety, or structural reasons inherent with the idea that make it unworkable?
Thanks in advance for comment and feedback but mostly I’m hoping you take this on in a seperate entry ;o)
Randy Campbell
Inherent
Improved efficiency would be nice to achieve but given the weight of the first stage it would be quite a task for the recovery helicopter.
Pingback: Selenian Boondocks » Blog Archive » Variable Gravity Research Facility (xGRF)