Random Thoughts: Risk, Orbital Rendezvous, and Depots

[Ed: I’m pretty sure I’ve used this argument before, but didn’t see it on the blog, so I figured I’d put it down in writing even if it ends up being repetitive.]

One of the most common criticisms I hear of propellant depots is that we can’t “put unproven technology on the critical path to our return to the Moon.” The idea being that doing something like that would be too risky, and should only be done after we have a basic approach that can do the job. The ironic thing is that had this mentality prevailed during Apollo, we probably wouldn’t be talking about “returning” to the moon, because there’s a good chance that we wouldn’t have made it there in the first place.

In the early days of Apollo, the mission mode had to be selected. The choice was between the massive “direct ascent” method, which while more technically conservative would’ve required a gargantuan (even compared to Saturn V) launcher, and the two approaches that traded riskier new maneuvers (orbital rendezvous) for much smaller launchers.

This debate came to a head in 1962, resulting in picking what was probably the “riskiest” of the three approaches, Lunar Orbit Rendezvous. It’s worth thinking about the context of when this decision was made. This was only five years after the first artificial satellite, Sputnik, had been put in orbit. It was only barely a year after the first US manned spaceflight, Mercury 6. The decision to bet the nation’s moon program on a risky architecture that required orbital rendezvous and docking–in lunar orbit no less–was made 4 years before anyone had actually demonstrated orbital rendezvous and docking. In fact, the spacecraft used for that demo, Gemini, had barely started development the year earlier, and wouldn’t have it’s first manned flight for another two years after the decision was made.

Had the Apollo team decided to take the “low technological risk” approach, and stuck with Direct Ascent and the enormous Nova vehicle, there’s a real chance that the program would’ve been canceled before they had ever even tried to make their first lunar attempt. Given enough time and money, Nova probably could’ve landed people on the Moon–it wasn’t technically impossible. But in the end, there probably wasn’t anywhere near enough time or money to make such a grandiose scheme work. Not only that, but even if it had been forced to work, it would’ve been even less sustainable than the chosen approach.

Now, it isn’t always a good choice to take risky bets by putting unproven technologies on your critical path. Doing so can sometimes be a mistake. But it’s important to realize that in a program like this, risk can’t entirely be avoided. At best risk can be managed, but if you delude yourself into thinking you can avoid technical risk, it often crops up in other forms instead.

I think that in many ways, the decision to make depots a larger part of the NASA lunar architecture going forward is actually less risky than the choice of Lunar Orbit Rendezvous back in 1962. The storage, handling, and transfer of cryogenic propellants on orbit is a lot more mature than the technology for orbital rendezvous and docking was at the point the mission mode decision had to be made in ’62. We have literally decades of launching and flying cryogenic stages in orbit. Hundreds of flights between Centaur, Saturn SIVB, and now the Delta-IV upper stage have demonstrated settled cryogenic handling, pressure control, mass gauging, and even fluid transfer. Projects such as Orbital Express, XSS-11, and others have increased our ability to do tugs, autonomous rendezvous and docking, and have even demonstrated transferring fluids from one spacecraft to another. Most of the pieces needed to make a small, first-generation propellant depot (one big enough to support manned lunar exploration missions) are either proven technologies, or at least well on their way to technical maturation.

I think that when looked at in the context of the decision to go with LOR, the decision whether or not to baseline propellant depots in NASA’s space transportation architecture is even more of a no-brainer. It may not be free of risk, but biting the bullet and aggressively pursuing the development of propellant depots as a vital part of NASA’s continuing human spaceflight activities is just the right thing to do.

The following two tabs change content below.
Jonathan Goff

Jonathan Goff

President/CEO at Altius Space Machines
Jonathan Goff is a space technologist, inventor, and serial space entrepreneur who created the Selenian Boondocks blog. Jon was a co-founder of Masten Space Systems, and is the founder and CEO of Altius Space Machines, a space robotics startup in Broomfield, CO. His family includes his wife, Tiffany, and five boys: Jarom (deceased), Jonathan, James, Peter, and Andrew. Jon has a BS in Manufacturing Engineering (1999) and an MS in Mechanical Engineering (2007) from Brigham Young University, and served an LDS proselytizing mission in Olongapo, Philippines from 2000-2002.
This entry was posted in NASA, Propellant Depots, Space Development, Space Policy, Space Transportation. Bookmark the permalink.

24 Responses to Random Thoughts: Risk, Orbital Rendezvous, and Depots

  1. Martijn Meijering says:

    As usual you have good arguments. A compromise to reduce risk even further would be to use depots only in support of a lunar lander, not for fueling an EDS, and to use hypergolics for the lander. Hypergolics depots have even lower risk than cryogenic ones, and they can coexist with SDLVs. NASA could still use an SDLV to get to L1 by cryogenic propulsion.

    Hypergolic landers are safer than cryogenic ones, both with respect to development risk and operational risk. Hypergolics depots could be operational within years. In LEO they could support refuelable tugs, thus making ISS resupply more efficient. At L1 they would support lunar/near-Earth operations.

    The performance loss due to hypergolics is marginal, both because cryogenics are used to take care of most of the delta-v and because hypergolics allow the propellant to be transported to L1 by Belbruno-like trajectories.

    Later on, once an L1 COTS-D was operational, commercial players could develop cryogenic depots for fueling their own EDSs.

    In this way you could have depots soon (with all the commercial benefits of that) without putting cryogenic depots (and cryogenic landers!) on the critical path to the moon. You would also not have to rely on NASA to develop cryogenic depots.

  2. kert says:

    Not to mention that our flagship up there, the ISS , that was supposed to teach us about future living in space, gets refuelled regularly by the Progress tankers.
    Also, Gemini-Agena demonstrated assembling propulsive stacks on orbit back in 1966.

  3. Adam Greenwood says:

    If there’s no technical risk associated with your project, then there’s no reason at all not to farm it all out to the private sector. This is self-contradictory justification on the part of NASA.

  4. Randy Campbell says:

    Adam; I think you misunderstood the article. NASA has not “justified” propellant depot, they have in fact rejected them as being “too risky” to put into the development path for the Return-to-the-Moon program. At least so far. Jon’s argument is that Propellant Depot(s) are LESS technically risky than risky options that NASA has taken BEFORE to get to the Moon.

    Of course this all falls back on we know this isn’t our parents, or even our grandparent’s “NASA” anymore.

    Of course we ALSO know that the ‘private-sector’ WOULD have this farmed out to them: ULA would get the over-all contract with LM launching a Centaur “fuel-depot” into orbit and Boeing launching re-fueling flights :o)


  5. MG says:

    There are many “flavors” of risk. The propellant depot is a technology risk. I agree with Jon that its technological risk is far less than Apollo’s 1962 LOR selection. I think Nova (Spanish for “(It) Doesn’t Go” was a technological risk all its own, even beyond Saturn. Neither rendezvous ops NOR rockets that large had ever been done.

    In addition to technological risk, there is programmatic risk (hell-o, Ares!). I combine budget, schedule, politics, staffing, etc. in this category. Ares has the image of low technological risk (though I think the reality is otherwise). Its programmatic risk is likely terminal to the project.

    One must also consider the reward potential of the risk. Ares doesn’t offer a reward commensurate with the risk. The propellant depot does — even if it isn’t “ready” for initial lunar missions, it is an enabler for much, much more. Even IF it had a technological risk greater than Ares (and I do not concede that), its reward is greater.

    Congrats on the recent flight success, Jon. ‘Tis a joy to see a bird of your making take wing, I am sure.

  6. MG says:

    D’oh! Now I read Rand Simberg’s related post.

  7. chuck2200 says:

    I agree that NASA’s argument about putting “unproven technology on the critical path” is largely smoke and mirrors, because it just isn’t true. The real reason is they don’t want a depot messing up their plans for the flying Godzilla. And mess up their plans is exactly what would happen because once the depot became operational, there would no longer be any justification whatsoever for building anything so huge.

    In defense of other objections however, I would offer that the moon plans were put on such a tight funding diet schedule that it wouldn’t allow for depot technology to be brought online and matured soon enough to make the schedule. “Pay as You Go” is the worst possible way to mature multiple technologies simultaneously. In that kind of situation, you would want to minimize the number of new (not unproven) technologies that needed to be on the critical path. That is just good Project Management practice. Another way needs to be found to bring the depot to pass. I would suggest creating a launch system big enough to do the mission with two launches, but not so big that the depot becomes its enemy. If done right, the smaller launch system may actually become the catalyst that brings the depot online as part of a follow-on applications program. But that would need to be clearly stated in the plan from the beginning with all parties buying into it.

    President Bush set the stage when he announced the VSE, a wonderful plan, but he hobbled it with unrealistic schedules when he said how it was going to be paid for. Getting to the moon by 2020 and paying for it the way he wanted to just wasn’t going to let a depot come online first. Just something to think about.

  8. gravityloss says:

    The amount of bad logic and unfounded assumptions rules the discussion, if the NSF forum is anything to go by. You Jon have an appreciable amount of patience to explain it why the criticism is so bad.
    I think there is a lot of possible founded criticism for depot architectures. It’s just sad that it hasn’t surfaced. We have to do it ourselves.

  9. Bill White says:

    Jon –

    A question about “dry launched” Earth departure stage fuel tanks.

    Suppose LH2 and LOX have been stockpiled on orbit and a lunar mission launches from Earth. Do you re-fill whatever tanks contained the fuel that propelled the payload to LEO (meaning the 2nd stage Centaur in the case of Atlas V and Delta IV or the Merlin 1C for SpaceX) or do you fill otherwise empty tanks incorporated into the Earth departure stage payload lofted as payload on top of the 2nd stage?

    How much work has been done on this aspect of propellant depot operations? Where would you place the TRL status?

    (Yes, I know Merlins and Kestrels use RP-1)

  10. gravityloss says:

    Something’s broken with comments, I’ve lost posts…

  11. Bill,
    You mean reusing the upper stage as an EDS by tanking it back up once it’s in orbit? If so, that’s actually my preferred approach.

    There are some subtleties depending on how you intend to use the EDS. The thing is most upper stages actually use at least 4 fluids (LOX, LH2 or Kero, Hydrazine for RCS, and Helium for pressurization), and you need to replenish the Hydrazine and Helium as well as the main propellants. If you’re using the stage for a single mission, you can just bring some extra helium and Hydrazine bottles along. That isn’t too complicated, but adds a couple hundred pounds of extra weight. All of the “Lunar Mission Kit” Centaur numbers I’ve used in my various musings have included extra mass for that.

    Alternatively, you could have your depot also hold hydrazine and helium, and transfer those as well. Those have both been demonstrated already in space. Or you could modify your vehicle to get rid of the hydrazine and helium requirements.

    But overall, I think that it would be feasible to retrofit most existing upper stages to allow them to be refueled in LEO and then reused to send stuff beyond LEO. It would take some integration work, but almost all of the actual subsystems/techniques are at a fairly high TRL.


  12. gravityloss says:

    Or maybe I had a dirty word. 🙁
    Anyway. If NSF is anything to go by, the criticism against propellant depots is largely really illogical and founded on baseless assumptions.
    Well, I guess it’s the reaction of the relatively uninformed public whatever new to them you’re talking about.

    We have to really criticize it ourselves to find the weak spots.

    Maybe I’ll start on thursday. I’m pretty busy in between.

  13. Gravityloss,
    A lot of those criticizing propellant depots on NSF are smart individuals, even if they haven’t given much in the way of substantiative arguments so far. But yeah, propellant depots are definitely going to require a lot of work on selling.


  14. Gary C Hudson says:

    If NASA had gone with Nova and Direct Ascent in 1961, my guess is that it would have been dropped or stretched out during the early F1 and CM troubles and then replaced with a two man Gemini lunar landing using EOR and the Saturn IB. Or maybe even a Saturn 1. I expect there would have been a backup plan implemented along with the Nova go-ahead decision that would have maintained the Gemini as an alternate. Perhaps not an overt plan, but a nudge and a wink to McDonnell-Douglas.

    And the program would have had only one lunar landing, most likely after at least one or two failures involving landing. Just my speculation, of course. Perhaps someone should write an alternate history?

  15. Jon, your post made me look for more details. I was unfamiliar with part of the LOR decision recounted here:
    http://oea.larc.nasa.gov/PAIS/Rendezvous.html Perhaps many veteran space buffs already know all the details recounted and more, but the last few paragraphs of the link tell a great story about going around the normal chain-of-command that I hadn’t heard before.

  16. Nels Anderson says:


    You mention that hundreds of flights of the Centaur, S-IVB and Delta IV upper stage have demonstrated “even fluid transfer.” How and when has this been done? I was aware that transfer of non-cryogens was routine but not that cryogens had ever been transferred in space.

  17. Jonathan Goff Jonathan Goff says:

    I was being a little bit coy about the LOX/LH2 transfer. The transfers I was referring to involve transferring liquid through a series of valves and pumps into a “temporary storage container”…..

    ….where they sit for a couple of milliseconds until the flames vaporize them and then they leak out this venturi thingy and vent harmlessly into space….

    Seriously though, engines are far more susceptible to damage from getting gas bubbles and other things, and require far more rapid transfers than you need for propellant tanking. The fact that they can make that work, reliably on demand, is a strong argument that “cryogenic propellant transfer” is a lot more mature than most people realize. The only two things that a relight doesn’t demonstrate are a) the cryogenic fluid coupling (which we have decades of experience doing on the ground where you have to deal with dust, humidity freezing, and liquifying air), and b) holding the propellants in a settled state during the transfer.


  18. MG says:

    “temporary storage container”

    Heh. That Jon is a sly one, isn’t he?

  19. Adam Greenwood says:

    In the aftermath of the Iraq invasion and in the run-up to the stimulus, a commentator, I think it was Megan McArdle, admitted that her support of the Iraq invasion had been flawed because her case for it relied on the Administration doing exactly the right thing. But in reality the Administration will always err, so her case was flawed. Liberal economists, she pointed out, were doing the same thing with the TARP and the stimulus package and the CO2 cap-and-trade bill. They were countering objections by saying, ‘yes, but if we do this and this and this just right, then we can thread the needle and get around that objection.’ But in politics this and this and this never happen. You never thread the needle. A program isn’t worth doing unless it has big margins. Orion/Ares didn’t and doesn’t.

    On the flip side, in business its almost always the secondary revenue streams that make or break you. Just as you don’t want a project where everything has to go right to succeed, you do want a project that can succeed in multiple ways. That’s where you really cash in. As Orion/Ares has developed, its role has become more and more limited. A propellant depot, on the other hand, by nature would enable lots of different kind of capabilities. Its not a luxury buy, its an investment. Its infrastructure.

  20. Adam Greenwood says:

    Adam; I think you misunderstood the article. NASA has not “justified” propellant depot, they have in fact rejected them as being “too risky” to put into the development path for the Return-to-the-Moon program. At least so far. Jon’s argument is that Propellant Depot(s) are LESS technically risky than risky options that NASA has taken BEFORE to get to the Moon.

    No, my point is that if there really was no technical risk associated with the development of Ares/Orion then there was no point in NASA being in charge of developing them.

  21. Andrew Swallow says:

    A propellant depot can help send probes to Mars and other planets plus satellites to GEO. This means that the design of the depot, tug, refuelling tanker and interface can be debugged whilst the Ares V/Jupiter are still being developed. The depot would then be at Technology Readiness Level (TRL) 9 when manned space flight needs them.

    The next major problem is getting private investors interested. A COTS-PD may help.

  22. Martijn Meijering says:

    Another good thing to keep in mind is that keeping something off the critical path doesn’t necessarily mean delaying its development. It can also mean using something else as a backup. Preferably something that already exists so it doesn’t eat up too much development money. Another solution would be to find something else to do until the risky thing is developed. Say keeping the ISS alive indefinitely and migrating to commercial crew taxis. But whatever you do, make sure you develop the new technology and perhaps some competing ideas.

    Constellation is an example of what not to do. To avoid putting cryogenic propellant transfer on the critical path (wise in my opinion) they develop something new which likely takes a whole of lot more money and is much less useful and even downright harmful to commercial development of space. That’s taking one problem and replacing it with a bigger one. But of course the reasons given by Constellation are just rationalisations, pretexts to keep the Shuttle workforce alive and to keep NASA in the launch business.

  23. johnhare john hare says:

    On ARocket the other day one of the participants in the fiasco said roughly, “We have a space station, so what if it is 15 years late and $72B over budget, that doesn’t matter because we have a space station.” As long as anybody involved can make a statement like that, seriously as far as I could tell, you have a situation that will not close.

  24. Martijn Meijering says:

    I wonder if that individual agreed with plans to scuttle said station at the earliest opportunity, which is even more messed up.

Leave a Reply

Your email address will not be published. Required fields are marked *