Bridenstine Hits The Ground Running On Gateway
NASA Gateway Program Justification For Other Than Full and Open Competition For The Minimal Habitation Module
“NASA’s decision, based on review of each NextSTEP-2 contractor’s capabilities, to sole source the procurement of the MHM flight unit for the cislunar Gateway to Northrop Grumman Innovation Systems (NGIS) as a follow-on to the originally competitively awarded NextSTEP-2 BAA Appendix A, Habitat Systems studies, Contract NNH15CN76C (See below). … In order to support the mandate to enable a human landing capability in 2024, the MHM must be launched in late 2023 to be delivered to Gateway no later than early 2024. The schedule constraints established by a December 2023 launch dictate that a module be on dock at Kennedy Space Center in mid-2023 for launch processing and integration. Per NASA’s schedule analysis, this typical timeline for module production must already be compressed in order to achieve the 2024 human lunar landing deadline. Given that the NextSTEP-2 contractors advanced designs to a near System Design Review (SDR) fidelity, NASA determined that it must utilize the existing concepts from the NextSTEP-2 Appendix A and use the development done to date to minimize the additional design work necessary to produce a module in time.”
Keith’s note: NASA has been directed by Vice President Pence to truncate NASA’s original plans to land people on the Moon in 2028 to a new date of 2024. That means NASA is going to have to make a number of prompt decisions on some basic aspects of how it accomplishes this 2024 goal. This NASA document makes mention of the fact that NASA is having to compress its procedures in order to meet the deadline set by Vice President Pence. NASA has decided that the only viable solution for a habitation module for the Gateway is to utilize a modified version of Northrop Grumman’s Cygnus cargo spacecraft. This spacecraft (originally developed by Orbital Sciences which was bought by Northrop Grumman) has performed flawlessly each time it has flown, so it is a known, proven design. While it would not be surprising that other companies will protest this sole source decision by NASA, it is hard to argue that other companies could have been able to provide hardware on the dock at KSC when NASA needs it to be there.
The only thing that is missing from this document is the cost of this module which is redacted on page 5 of the original notice posted by NASA. Given the mysterious and ever-changing estimates of how much it will take NASA to meet the 2024 goal it is hard to imagine that this number will remain a secret. Indeed, just last week NASA Administrator Bridenstine openly admitted in congressional testimony that NASA has a chronic problem when it comes to estimating costs and then delivering on them.
Meeting the 2024 lunar landing date is going to be sporty – at a minimum. To his credit Jim Bridenstine has hit the ground running. Gateway has been downsized to a basic initial configuration. Maxar has the propulsion portion of the Gateway and Northrop Grumman now has the initial habitation portion. Orion and its service module exists and SLS is being fabricated albeit behind schedule. Moreover commercial launchers from SpaceX and ULA are ready for procurement to launch components. All that seems to be missing now is a lunar lander. NASA has a long way to go. Many people think that the landing could be done in a simpler fashion. But again, given the lead time Bridenstine has been given he has certainly risen to the challenge. It will be interesting to see who is picked to run HEOMD given that Bridenstine has said that some important decisions are on hold pending those appointments.
Schedule Pressure is now NASA’s primary driver it appears, schedules are being compressed. How has that worked out for us before?
It got us to the moon.
A valid retort, but schedule pressure has also been known to cause catastrophes and lost lives. While we must indeed move our legs quickly in this race we must also be very careful when choosing our steps.
Yes it did, and we haven’t been back in 50 years because the strategy used was a dead end. I watched it transpire first as a young man (12 when Apollo 11 landed) through Skylab, OSTP, and Shuttle.
I think manned exploration of space is a long term proposition (decades not years) and we should plan accordingly. Developing a long term efficient way of getting to and from the moon will surely be useful in going to Mars. Learning how to stay on the moon for long periods of time will also be useful for future missions to Mars. But to keep a station or Hab or whatever you want to call it continuously running will require being able to launch supplies over time. Otherwise we’ll go there half a dozen times, plant the flag, and then forget about it like we did the last time.
We have shown zero capability to develop generations of hardware that build upon the past. You can criticize Soyuz as much as you want, but it’s been upgraded over a fair amount of time and is still out there working properly, reliably and relatively economically.
I’m an engineer, so I enjoy the ‘clean sheet of paper’ approach, but it’s much better in the long run to build on what you’ve already done.
As an aviation example, consider the B-52, designed in the 1950’s, it has been re-purposed for missions that the original designers didn’t consider. It’s been upgraded many times and they’re now looking at upgrading the engines to keep them going another 30 years. That’s the kind of design that’s needed here to allow space exploration to become economically workable in the long term.
I think it’s worth mentioning that those “clean sheet” designs are exactly what you want for experimental vehicles. The evolving design with multiple generations of upgraded designs is what you want for an operational vehicle. NASA, as a research and development organization, is probably biased towards experimental vehicles.
Ok, but at this point in time in space exploration, shouldn’t a lot of the experimental effort be complete? Compare it to the development of aviation. 50 years after the start of the airplane, we were in the mid 1950’s with commercial air travel all over the world. It’s not an exact analogy by any means. Yet I think you could make a case that by now we (USA) should have a reliable proven means of getting people to and from LEO.
When Shuttle was first proposed, all of it’s elements were reusable, with a flyable first stage and a flyable second stage. Than as reality struck the vehicle became less capable.
After Challenger, the Shuttle could no longer carry payloads with a live stage (too risky), so a major portion of it’s reason for existence was shelved.
The Shuttle was also designed to return large payloads from orbit (ESA lab). Except for that mission, the Shuttle only brought back the crew. Had the Shuttle been designed to carry crews only, with a small cargo bay it would have been much smaller, less expensive, safer and would still be flying.
Today, NASA is building and flying experimental aircraft. But that’s a large part of what the first “A” in the name is about. But, as far as I know, they’ve never even considered designing and building an operational aircraft. I think that X vehicle mindset carries over into their spacecraft design and development process. That’s a mindset involving clean sheet designs rather than gradually evolving ones.
If they wanted an operational vehicle, the Shuttle should have been treated as an experimental vehicle (which it was) from the start, and plans for actual cargo hauling or routine access to space for astronauts should have been left to a follow-on vehicle built using experience from the Shuttle. Of course, the budget for that wouldn’t have been approved. They couldn’t even call the first flight article a prototype and do significant redesigns between Columbia and Challenger.
By the way, returning large payloads was used for more than just Spacelab. There were a few other large, two-way science payloads (and Spacelab flew more than once.) Also, STS-51-A retrieved two broken satellites, which were repaired on the ground and then successfully reflown. But that was a very underused capability. Although, even without that, it wouldn’t have been a crew-only vehicle. The size of the payload bay was set by a Department of Defense requirement for launching, not landing, large satellites.
heh. If you want to imagine what it might look like if NASA did try building their own operational aircraft, just look at what happens when they try building their own operational spacecraft… SLS/Orion, cough.
Getting back to Shuttle, the sensible thing to do post-Apollo would have been to switch back to the Saturn 1B for cargo, while developing a small spaceplane to launch on the 1B for crew. The VAB setup would have easily supported back-to-back launches of a cargo then the crew to work on it, and they could have then done most anything Shuttle could at a fraction of the price.
Work toward a fast-turnaround recoverable Saturn 1b in the meantime, and we could have saved 40 years in getting where we are now. (The 1b’s performance was remarkably similar to a Falcon 9’s.)
That’s my point. NASA doesn’t do operational aircraft, and their approach to spacecraft is one suited for experimental vehicles. Even when spacecraft which are claimed to be operational. Even when there are plans for upgraded versions, like the SLS Block 1, 1A and 2, they aren’t evolving designs. Before SLS has ever flown, they know what the changes will be. Not the full design, but Block 1A will be just like Block 1, except for a new, clean-sheet-design upper stage. Block 2 will be just like Block 1A, except for new clean-sheet-design boosters. That’s a contrast to SpaceX, where they didn’t even start deciding on the changes to make for Falcon 9 v 1.1 until they had about a year of experience with v. 1.0.
But I don’t think you could have turned a Saturn 1B into a 1970s (or 80s) equivalent of a Falcon 9. It was designed from the start with performance rather than cost as the first priority. Actually, cost was so far down the list, I’m not sure if it was even on the first page. I don’t think that’s something that’s easy to change by evolving the design. Also, Falcon 9 needs some technology, such as GPS and twenty-first century computers, which just weren’t around in the 1970s.
No fundamental argument with your first paragraph, but I will observe that the NASA approach is suited only to experimental vehicles with extremely generous time and money budgets. Their excessive indulgence in risk-elimination process costs a great deal of both.
RE Saturn 1B, two thoughts: It still would have been far faster and cheaper than what did actually happen. And, I failed to state that I envisioned this done under a continuation of the old fast-development Von Braun Ironworks model, not under the risk averse bureaucracy that followed. That stipulated, I expect the 1B could have been very usefully developed even with mere 70’s technology – there were considerable advances over the early 60’s tech in the 1B available.
But on this sort of counterfactual, best to trade opinions then agree to disagree, as nothing of course is provable.
“Developing a long term efficient way of getting to and from the moon will surely be useful in going to Mars.”
A sensibly organized frequent-flight reusable cislunar transportation/logistics system will be able to throw 10-ton payloads to Mars on third shifts and weekends.
Or, support space assembly of multiple 10-ton payloads over time into a major Mars expedition.
Buy the Moon, get Mars thrown in as a low-cost extra…
While the B-52 is a good example of upgrading an airframe over decades, it was also not our first heavy bomber. Far from it. There were many “clean sheet designs” between the first aerial bomber and the B-52. Launch vehicles and crewed spacecraft have simply not gone through that many iterations yet.
I’m hoping that SpaceX will be successful at creating the first fully reusable TSTO, so that we can build on that. Because I seriously doubt we can build on an existing design that’s expendable and obtain the sort of flight rate that is needed for a truly sustainable program.
It is a difficult problem to solve, since there are so many competing variables. My real concerns are that schedule “pressure” will cause us to go in the direction of a one off that leaves us with no where to go in the future, but to start all over again.
There is also the increased risk. While I’m sure that engineers, managers, etc., won’t ignore risks, when you’re under the pressure of getting something done “on time”, a small problem may not be investigated as it should be.
And people died, remember?
It also got us Apollo 1, the Challenger accident, and the Columbia accident. We should accept schedule pressure when the end is worth the risk but in this case it is not.
The Cygnus is an excellent choice. While NG is the contractor the Cygnus PCM itself is actually produced in Italy by Thales Alenia so you have your international partnership. IIRC, it is the same company that made the Spacehab module for the Space Shuttle and some ISS modules.
The MultiPurpose Logistics Modules, carried by STS and berthed to ISS via the ISSRMS, and Nodes 2 and 3.
I think you mean “Spacelab” not “Spacehab” in the case of the Shuttle article. I don’t think Thales Alenia had anything to do with it, but we’re talking about work started in the 1970s. As far as I can tell, the pressurized module was build by ERNO-VFW Fokker (mostly or entirely German) and the unpressurized part by British Aerospace. Neither of those companies exist anymore. BA’s only merged and reorganized once (now BAE Systems) but ERNO has changed half a dozen times and is now part of ArianeGroup. For all I know, their original divisions of ERNO are scattered all over
EuropaEurope (sorry about the typo…) and Thales may have bought one. That’s as much of a mess as figuring out whether or not Northrup Grumman, as it exists today, had anything to do with building the Apollo lunar lander.Sorry I confused the multipurpose logistics module for Spacelab.
Point being though the Thales Alenia know how to build hab modules.
Is there a multi-port docking hub/node somewhere in the pipeline? How fast might the Italians produce a Tranquility 2?
The sole-source justification says Northrup Grumman will be adding docking ports (among other things) to a Cygnus bus. Two radial and two axial, as opposed to the current one axial port. This is probably going to be as different from a Cygnus as a Dragon 2 is from the original cargo Dragon.
So, there might be an SLS/Orion that can take a crew to and from a transfer point in Near Rectilinear Halo Orbit (NRHO) near Luna in 2024. Maybe even on its second flight – or perhaps third? – and not its first.
There might now also be a minimal hab module with stationkeeping and power at that transfer point in 2024, because, uh, why again? Will the Orion not be able to dock directly with a lander? What’s this station for again?
As for a lander, it looks like delta V from NRHO to Lunar surface and back is roughly 4.7 km/s. Call it 5 km/s with some reserve. It should be easy enough to do that in one single there-and-back stage – LOX/LH2 mass ratio would be just over 3 (IE a bit over 2/3rds initial mass is propellant) while with hypergols it’s just over 5, at Apollo LM Isp.
But it’d be a big single stage lander. The Apollo LM was a bit over 16 tons (just under 5 tons dry) and we’re asking for about 1500 m/s more to get down from NRHO rather than just Low Lunar Orbit. A hypothetical extended-tank Apollo LM would need roughly double the propellant, and would come in at near 28 tons – more than even an SLS could deliver in one go.
And possibly we will want a bit more capacity than an Apollo LM. (And probably we’d have a hard time building anything that ruthlessly lightweight these days anyway. The trend in mass per person per mission-day from the Apollo command module to Orion is upwards.)
So, the lander will have to go up to NRHO in multiple pieces. How those pieces get broken down, that’s the interesting question. Last version of NASA’s plan I saw involved separate descent and ascent vehicles, plus some sort of auxiliary tug, which strikes me as both overly complex and as revealing a somewhat illogical aversion to any tiniest degree of in-space assembly. Lots of rendezvous and docking is OK, but docking then locking modules together into one vehicle is not? Odd.
I suspect an approach that gets one partially-fueled round-trip lander up there then adds separately-shipped propellant – by simple docking with whole “drop tanks” if in-space fluid transfer is regarded as too radical – would lead to lower parts count, thus lower risk of component failure and simpler cheaper development (possibly faster too, with fewer total poles in the tent, each of which might delay things.)
It’d also give the Gateway station something to do (I believe I’ve seen some mention of a small manipulator arm as part of it?) Also the Orion crew could have a much more interesting mission – if NASA’s not confident in automated-docking assembly of the lander at the other end of a couple seconds speed-of-light delay, they could assemble it manually once they arrive.
My axe to grind here of course is that this would be much more useful in the long run than a 3-piece disposa-lander arrangement. And even with a tight deadline, NASA really ought to keep at least one eye on the long view. Do they want to keep going to the Moon long-term, or lock into inherently disposable hardware and endure another couple generations hiatus again?
A few random comments:
The two or three stage design does let you cut down on mass. I agree with you about the mass ratio, but a lot of that final mass can end up being empty propellent tanks and landing gear (and more rockets to push around that dead weight at the acceleration needed for landing and takeoff.) That means less mass for the same amount of usable payload mass.
Unless they can find a way to reuse those stages, I agree with you. This is going nowhere; it’s not a step towards a sustainable presence. The nice thing about some of the concepts is that they could either support expendable landings without in situ refueling or reusable landings with refueling on the surface.
As far as getting it there, with staging, they might be able to manage a single launch. I also remember seeing something about a draft RFI on a commercial LEO-to-Gateway tug.
The trend in mass per person (or person-day) is real. I suspect it’s related to similar trends in cost. Once you insist on very high reliability, people come out of the woodwork with potential failure modes and risks that need to be mitigated. Every new project inherits all the issues from the last one, and invents new ones of it own. That adds both cost and mass.
When it comes to docking versus assembly, I suspect it’s the interface. Assembling modules, as was done with ISS, involves lots of connections, feedthroughs and plumbing. Docking doesn’t. In any case, Gateway won’t have a mechanical arm, at least not for a while. That got deferred juggled the schedule for a minimum capability Gateway in time for a 2024 landing. Which is slightly ironic, since Maxar would have built that arm. Now they get a fast tracked Power and Propulsion module instead.
“The two or three stage design does let you cut down on mass.” And, in an age when an astonishingly cheap all-three-boosters-recovered F9H will put ~40 tons into LEO, allowing ~16 tons delivered to NRHO/Gateway, cutting down on mass should no longer be the be-all and end-all goal overriding longer-term considerations like NOT THROWING THE EXPENSIVE HARDWARE AWAY EVERY FLIGHT. Pound-wise, gigabuck foolish, if I may coin a phrase.
Put another way, propellant delivered to NRHO via F9H is about $6K/lb if you just throw away the LEO-NRHO transfer vehicle. Less on an ongoing basis as you begin to reuse more of the transport elements, get flight rates up into some economies of scale, and eventually get the raw material from closer-by.
Lander hardware delivered to NRHO carries that same ~$6K/lb transport cost (if you don’t use SLS – there, it’s more like $50K/lb) *plus* the production cost of the hardware – these days, tens of thousands per pound and up for NASA space-grade hardware.
IOW, it’s *already* uneconomic to discard a pound of lander hardware to save a few pounds of propellant. And it will only get more so. The world has changed, and the people designing these missions need to wrap their heads around this.
Don’t get me started.
Instead of Gateway how about the Moon Direct with a LEO to lunar surface vessel. Launch it empty, fill it with fuel, board the astronauts and away you go.
Give it a 40 day mission profile and you could even send it to near Earth asteroids once in awhile. Keep the dry mass under 50 tons and you can launch it whole on a FH.
Keep in mind there’s a difference between Gateway the specific SLS/Orion-need-a-destination station-design, and “Gateway” the NRHO orbital location.
The location actually makes considerable sense as a depot/logistics/transfer point to and from Luna. It keeps a fully reusable two-way lander from becoming gargantuan – with LH2, about 5 km/s round trip to the surface with some reserve means about 68% of the total vehicle starting mass is propellant – very doable with a robust reusable vehicle also carrying significant payload.
I’m not familiar with “Moon Direct” details, so if I’m missing something clever here, speak up. But going direct from LEO to Lunar surface and back takes about 6 km/s each way with about the same reserves as my NRHO case. 12 km/s total round trip. Or about 10.5 km/s if you aerobrake on the return? Current SOTA is about half the ~3 km/s excess arrival velocity into LEO being scrubbable via aerobraking… So let’s assume 10.5 km/s total delta V needed for the Moon Direct lander as it leaves LEO.
Ow. That’s a mass ratio of 10.9 – or 91% propellant at mission start. That’s a decent mass ratio for an expendable hydrogen stage with no payload – it’s better than a current Centaur III. Add payload, reusability, landing legs, and aerobrake provisions and you really are talking an, ahem, very advanced piece of lightweight engineering. Ahem.
Put another way, that 50 ton dry mass ship in LEO would require 495 tons of LOX plus LH2 to land on Luna then get back to LEO.
I have to be missing something here. Someone want to tell me what Moon Direct includes that I’m not accounting for?
OK, Bob does assume a functioning propellant source on Luna to support fully reusable Lunar logistics direct to and from LEO.
In my view, he handwaves just a bit too much and stacks up a few too many optimistic assumptions about how quickly and cheaply we get there from here via expendable missions. My view, as in a previous very similar argument we’ve enjoyed (Hi, Bob!) is that the number and cost of expendable missions required to directly reach those sunny uplands isn’t likely sustainable, so we need to get to full reusability much earlier, most likely via an intermediate near-Luna depot/logistics/transfer base.
Bob’s mileage will no doubt vary. Keeps life interesting!
I think (or hope) mechanical structure would be less expensive. But for things like flight instruments, $1-2 million per kilo is typical. That’s much more than your estimate. So I see your point. On the other hand, I’ve had senior engineers tell me flat out that the cost of an entire spacecraft (robotic) scales directly with the size of the fuel tank. And a room full of engineers and scientists didn’t blink at that claim. (I attribute that to bad parametric cost estimates: Past, high-delta v missions have been unusually expensive. But they’ve also been extremely ambitious in several other respects, so you get a false correlation.)
On the other hand, I think we are limited to something like 16 tonnes to lunar orbit (per launch.) I don’t want to use SLS anymore than necessary. Less than 16 tonnes would be better; I think that’s for a fully expended Falcon Heavy, and I’d rather not do that. So I see staging as a way to maximize payload to the surface, not to minimize the amount of fuel use.
But there are plenty of ways to make staging reusable. I could imagine something like an Apollo-style LOR. One vehicle which would be a reusable Gateway to LLO tug, and a separate, reusable LLO to surface (and back) lander. (Actually, I did just imagine that, although it’s almost certainly not original.)
Even civilian airliner airframes average thousands of dollars per pound these days – tens of thousands for engines and avionics. For recent large spacecraft, the overall average is something close to a factor of ten more. Aerospace methods and materials have gotten very sophisticated, even “just a cheap tank” isn’t really, and dollars ain’t what they used to be.
40 tons in LEO gives about 16 tons to NRHO with LOX/LH2, but yes, that includes the transfer vehicle. 4 tons is a reasonable estimate for that, modestly reusable – you might shave it to 3 tons but it’d be fragile. Or you might make it robustly reusable (and with provision to save maybe 1.5 km/s of the 4 km/s return-to-LEO via aerobraking) at nearer 5 tons.
Regardless, yes, 10-12 tons of actual payload delivered from Earth to NRHO via an all-boosters-recovered F9H launch is the practical planning ballpark.
FWIW, the Berlin Airlift supported an entire city for a good part of a year with <10 ton payloads, and significant campaigns in the Southwest Pacific theater in WW II were fought entirely with 5000-lb C-47 cargoes. We should be able to support a very serious Lunar program with one or two ten-ton payloads per month – IF we get our heads out of the past and start planning for what we have now.
But it won’t be NASA that will do it. They have neither the culture, experience or incentive to make it work economically. It is long past time for a public-private corporation to take over the exploration and development of the Moon.
There is some reason to believe, BTW, that commercial outfits can develop useful space vehicles with overall average cost/lb much closer to current commercial airliners ($1K-$2K/lb) than to current government-developed spacecraft (multiple tens of thousands per pound.)
It’s more a matter of far less labor-intensive processes than of any fundamental difference in the technologies incorporated, FWIW. The typical government development process for a major spacecraft ends up burdened with a LOT of people coming out of the woodwork to add their two cents worth, entirely aside from the vast accumulation of mandatory government-procedural bumf.
“Assembling modules, as was done with ISS, involves lots of connections, feedthroughs and plumbing.” The key being, “as was done with ISS”. That was all designed and specced how many decades ago?
Connect controls and sensors via the space equivalent of bluetooth, minimize power and fluid interconnects, and do those you MUST have via some of the modern in-space fluid/power/signal interfaces either already in development or actually entering the market.
And if the result still isn’t completely plug-n-play spaceship, well, you’ll have those two extra astronauts along anyway.
The state of the art in connecting modules _should_ have improved since ISS. But we’re talking about a program which will be putting things in orbit with solid rocket boosters and main engines developed in the 1970s. I have confidence they can make a straight-forward problem very difficult.
Oh, and bluetooth is probably out. I don’t think they even let you use it on airplanes. NASA is even stricter when it comes to electromagnetic interference. To the extent that Reference Publication 1374, a history of examples of serious problems due to EMI, includes a false and completely fabricated story about the cause of the 1967 major fire on the USS Forrestal.
I share your confidence that the established development organization can make any straightforward problem very difficult, if left to follow their ingrained institutional tendencies. The idea of course is to apply a bit of pressure on them to change their ways, even to farm out parts of the problem to others. “To dream, the impossible dream…” 😉
I didn’t mean bluetooth literally, of course. Shudder. But wireless comms can be made reliable and very EMI resistant, especially if you’re limiting the “wireless” section to closely-aligned contact patches at the docking interface. (Even more so if what you’re transferring at those contact patches is a bit higher frequency than RF…)
This sort of thing is already well underway in various places. Not overly burdened with funding, but underway.
The FCC restriction on use of cell phones in flight is roundly ignored. I have been on flights on which passengers could use personal cell phones to call through the aircraft microcell. And the restriction has always been pointless and never supported by actual interference data. During the most critical portion of any flight, take off and landing, the aircraft is flooded by cellular and wifi signals from the ground.
Oh, and I’m still wondering what the two astronauts who stay on Gateway will be doing. This isn’t 1969; we can leave Gateway on autopilot. So why send four people to lunar orbit when only two are going down to the surface? Please don’t say they can teleoperate robotic rovers. Two and a half seconds of two-way light time isn’t very significant.
I’ve actually done some rover teleoperation at a couple seconds simulated speed-of-light delay. It’s doable, but it slows you down to a fraction of the speed you can safely operate at zero lag. Small control input, wait 2.5 seconds for results, correct as needed, rinse-and-repeat endlessly.
But yeah, the real answer here is, the mission is being shaped in a myriad ways by the SLS/Orion they already sank tens of billions into. Not to mention the astronaut cadre they’ve spent many many millions more on. Those “extra” seats will NOT fly empty.
So, might as well plan the mission so the spares have something to do while the lander’s away. And, resting up from being the primaries in bolting the lander together counts.
How autonomous was the rover you teleoperated? My understanding is that JPL’s gotten pretty good with software to autonomously drive short distances (meters), with the operator just identifying a destination with a reasonably clear path from the initial location. And I assume APL will be building Dragonfly for autonomous flight, since I can’t imagine flying a quadcopter with an hour and a half two-way light time (TWLT). I’d think that if a robot can operate autonomously for much longer than the TWLT, the latency wouldn’t be a problem. For lunar rovers operated from Earth, perhaps thirty seconds of autonomy, and that sounds doable.
How autonomous? Not at all. This was a private backyard experiment in the late eighties, using an RC all-terrain truck, an onboard steerable vidcam with transmitter, with delays implemented in the Amiga desktop machine we were using as an operator station. We had a video overlay programmed to give analog visual indication of control-input magnitudes, which let you pretty quickly learn to gauge how much input to apply to achieve a given effect. But zero attempt at autonomy – that was not even a thing at the time <grin> plus we were mainly looking for quick experience with the effect of the delay on manual teleoperations.
Rio Tinto has advanced the technology quite a bit. They are operating a huge iron mine in the Australian outback from Sidney. Blasting, loading, hauling are all done via a satellite uplink. And the operators aren’t Ph.D’s, just regular miners who were retrained for the task. There were a couple of fatal accidents at the start, mining is a dangerous industry, but they worked the issues out and now have a great safety record.
Yup, I recall a really interesting presentation at the last Space Studies Institute conference from a Canadian mining company about remotely operated mining machines that could be just the thing for off-planet extraction projects.
One of the big benefits from remote operation is when you can get it so there are no humans at all near the worksite for routine operations. Can’t accidentally crush or blast or bury someone if they’re not there!
Sounds like Rio Tinto may be doing more direct teleoperation than autonomous machines. Retraining miners as direct teleoperators may well be cheaper for them than developing autonomous equivalent machines – that can be really hard for complex operations with non homogeneous materials. If it remains so longer term, space mining may well require live operators within a small fraction of a second speed-of-light delay.
The upside of autonomous operation in space is that humans are no longer needed. The downside of autonomous operation in space is that … humans are no longer needed.
I just reread a seminal article on AI in a 1975 issue of Analog. The author’s conclusion was that the main obstacle to AI was not the development of a mysterious algorithm for self-awareness but simply getting something approaching the processing capacity of the human brain into a mobile computer. That point is rapidly approaching, and on the software side we have the rapidly advancing field of machine learning. Soon AI will reach something close to human capability. And it will not stop there.
Yes, it’s an evolutionary approach versus trying to make some huge AI leap. That is why lunar mining is far more practical than asteroid mining in the near future.
My understanding is that LOP-G is ISS v2, so wouldn’t there be other research? Besides, the Gateway would also be under construction/commissioning/testing.
My random comments:
1. The 3-piece lander design is no longer a requirement, the latest draft RFP made this part clear, it’s just a baseline/suggestion, companies are free to come up with their own designs. The entire lander is now bid as one thing (integrated lunar lander or something), there is no longer separate bids for descend/transfer stages etc.
2. I believe the 3-piece design comes from the constraint of launch vehicles. Since they’re launching the pieces on commercial LV, the maximum weight of each piece is limited to around 15 metric tons, which is the TLI performance for expendable FH or New Glenn.
3. Manipulator arm on Gateway: Based on current plan this wouldn’t be available until 2027, so can’t rely on it for early landings.
As already noted, taking a 15-ton-per-payload limit as mandating a 15-ton-per-VEHICLE limit is a result of irrational longtime NASA prejudices against any sort of vehicle assembly in space whatsoever, right down to the dirt-simplest dock-modules-together approach.
(Carefully ignoring the LEO restacking/docking of the LM and CSM at the start of every Moon mission NASA ever flew – that’s different, hey look a squirrel!)
Apparently though the resulting Three-Part-Disposable-Lander concept here was so patently absurd that they now have to at least appear to consider other approaches. Good!
Yup, NASA is officially asking for alternate-concept input on landers now. http://www.parabolicarc.com…
Mind, I’ve had some experience with trying to sell NASA on a more effective approach than the “baseline/suggestion” they’ve set their sights on. I will not yet donate my watchful skepticism reserves to Goodwill…
It isn’t even NASA having their sights set on something. They go outside NASA for proposal reviews and evaluations. But these proposals have a page limit, and it’s very important to cover all your bases within that limit. If there is a potential problem, you need to discuss how you’re going to address it or explain why it isn’t really a problem, or something. If the reviewers notice a problem the proposal didn’t even mention, it can really hurt their evaluation.
The proposal also has to show that the whole, underlying concept is viable. If you need to use in situ refueling on the Moon from the first mission, that’s going to be a hard sell. If you think two-stage reusable will work, but it takes in-space, cryogenic propellent transfer, you’ve got to prove that architecture is viable. That takes up the page space you desperately need for addressing the technical details. But if you adopt the possible concept NASA “suggested”, you get a free pass. That would automatically be considered viable, even if all you said was, “This basic approach will work because it’s the one NASA suggested.” In fact, someone could probably sue if they were not selected and the evaluation faulted them for following NASA’s advice.
That said, you can get selected for things that go against the “suggestions” or “exampled” provided. The entire Juno mission was. One of the potential New Frontiers missions people could propose was a “Jupiter Polar Orbiter with Probes”, with the probes specifically being multiple atmospheric probes descending into Jupiter’s atmosphere at different locations. The stated goals involved atmospheric composition deep below the cloud level, and the AO did say it was ok to make those measurements in a different way (_if_ you could convince the reviewers that your different way would actually work.) It turned out probes were much harder than anyone thought, and the Juno team proposed microwave radiometry as an alternative technique. But that was an uphill battle because they had to put a large amount of time and effort into convincing people it would work.
So, why don’t they provide a way to request a limited addition to the proposal page limit in cases where someone is proposing something other than what NASA “suggested”?
My working assumption is they’ve set things up the way they have because they like things the way they are.
Cynically, changing how large procurement processes work is probably harder than putting people on the moon.
Well, yes. Not harder than NASA’s current plan to put people on the Moon, but harder than some of the alternatives which have been suggested. I’ve seen what it takes to make minor changes in the process for contracts and grants worth less than $500,000. Mostly, no one bothers and comes up with cumbersome workarounds.
Well, I don’t think NASA is monolithic enough for “them” to have a single opinion. Some people there probably do like things the way they are. Others might be more than happy to change them.
But in this case, you’re running into Federal Acquisition Regulations. How exactly would you let people apply for extra space to explain why they are doing things in a different way? How do you review that and decide whether or not to do so? I’m certain some people would try to sleaze their way into more space by claiming their proposal was different from the approach suggested in the RFP, even if it was really just a minor variation on that approach.
That might not be impossible, but it would take a whole lot of work to implement it within the FAR requirements. And it would probably add six months (or more) to an already-glacial procurement process that they are trying to speed up.
To be fair, I went back and squinted closely at a very busy NASA chart including the 3-piece lander design, and they did in fact envision docking the three pieces together before using the beast. So, that particular sub-rant, though fun, was wrong.
It still strikes me as an approach that “just sorta growed”, and I think NASA’s time would be far better spend negotiating a bulk-purchase discount for 3-boosters-recovered F9H flights (carrying a transfer stage plus 10-12 tons NRHO payload, or at some point carrying transfer propellant + NRHO payload and meeting in LEO a returning reusable transfer stage) than in trying to squeeze marginally more function per F9H flight out at cost of reusability.
Perhaps there is a concern about descent and ascent loads on a “docked combination” – its one thing to reverse a LM and connect the tunnels in orbit at zero g, with tiny shock/vibe loads. It’s another to take that combination down to the surface and back. I’m not saying the loads are actually a problem or that it can’t be done, but that might be some of the background thinking.
You’d hate to have a hard landing and the stack fall apart.
The docked Apollo LM-CSM stack underwent a fair amount of acceleration during the TLI and LOI burns. Not comparable to LEO launch accelerations, but in the ballpark with Lunar landing.
TLI occurred before the rearrangement, and had a peak acceleration of around 1.4g. LOI was about 2.4 m/s^2 (889 m/s delta V in 360 second burn), or 0.25 g in round numbers.
I found another reference that gives the ascent stage acceleration as 3.5 m/s^2 – about 1/3g.
So those are pretty small accelerations.
What was wrong with 2028?
I think Mr. Bridenstine is correct about political risk. Regardless of what happens in the 2020 election, there is a good chance the Republicans will not be in the White House in 2025. (Why I think that is side issue.) We do tend to see drastic revisions of NASA’s human spaceflight goals when a new President is from a different party than the last one. So 2024 is a good deadline for doing _something_ substantial and which would be a meaningful step forward. I’m not sure landing two people on the Moon is the right something. But the more I think about it, the more that date makes sense.