While I think that we could do a lunar return for far less than NASA estimates, if allowed to do so without having to use SLS/Orion, or the Gateway, I certainly agree with the second point that Wayne Hale (new head of the NASA Advisory Council) makes, as he takes more than one page from my book.
6 thoughts on “Blood And Money”
Comments are closed.
(copy of a comment submitted to Wayne Hale’s blog)
I am inclined to agree that the intended result of Artemis, incremental development of an eventual reusable and affordably scalable Earth-Moon transport net, is worth it even if costs do come out at the high end of the ranges mentioned so far.
The key issue I see in contention though is, if we take the approach that generates that high end of the estimates, will the job ever be done at all? That approach being to entrust overall project management plus key hardware element developments to the existing NASA human exploration bureaucracy. Which, with all due respect, has gone downhill in the generation since it spent hundreds of billions to eventually fly Station, never mind in the two generations since it developed Shuttle or the three since Apollo.
Seriously. The recent Constellation then SLS/Orion track record is, funding at growing billions a year, in exchange for apparently indefinite year-per-year delays, to develop obsolescent systems we won’t be able to afford to fly more than once a year anyway. Is NASA human exploration even capable of doing better at this point? Bureaucracies do sometimes reach a state of terminal sclerosis as they age. The evidence I see indicates exactly that here (though I’ll certainly listen to reasoned argument otherwise.)
NASA HQ would seem to be aware of this problem, as evidenced by their attempts to manage Artemis via a brand new directorate and to bypass SLS (defeated by Congress), and to bring in commercial developers for landers (jury’s still out.) But at this point, anyone knowledgeable looking at the question of whether NASA should be given $4G-$6G/year in new funding for Artemis has to be aware of the strong possibility that renewed traffic to Luna still won’t result.
We do have a problem here, and a solution doesn’t seem practicable in the current political environment. My best guess is, absent drastic changes, we’ll see Artemis die the traditional death of lip-service – the goals will remain official, but the required new funding (or the permissions to truly radically redirect existing funding) to make it actually happen won’t be there.
Which part? The SLS/Orion/Gateway part or the yearly robotic missions using commercial launchers part? Because a lot of people would be happy if part of it died and it does look like the current administration is setting up the conditions to let that happen.
There are some strange dynamics at play here.
I think you won him over.
Earlier today I was is an Artemis discussion elsewhere (on a non-aerospace blog), which turned into two long comments. They don’t relate to Wayne Hale’s point, but they do address some of the problems with a NASA SLS approach.
****
Artemis itself is probably not going to perform within the schedule and budget because it relies on Orion and SLS. Although Bridenstine was open to Pence’s statement that if SLS can’t get it done by 2024, NASA would use other launchers, there’s just no way for NASA to commit to such a move.
Given the speed at which they need to get the lander contracts rolling, they’ll probably commit to a warmed-over Altair/Apollo style lander. Altair was going to have a crew of four, and is basically an Apollo LM do-over, with one descent engine in a descent stage with splayed legs, and one ascent engine in a stage that sits on top of the descent stage. The descent stage has a top hatch for docking and a side-door and longer ladder for lunar access.
It’s a simple design that uses a vertically stacked pair of stages because that configuration packed easily into the launch vehicle, and “up” in the launch configuration in Florida is the same as “up” for lunar operations because up is up. That puts the docking hatch on the vehicle’s roof, puts the lunar access door on the side, and puts the ascent module on top of the lander so the astronauts have to climb up and down a tall ladder to get to the surface, much like using a third-floor apartment’s fire escape to go back and forth. There are dozens of better layouts that don’t resemble Apollo and which would work much better, but those probably won’t be explored.
And Musk and Bezos may make the whole thing moot. The other day I was crunching numbers on a Starship launched by three Super Heavy boosters into GTO (three boosters gives a tremendous payload gain, without which a Starship can barely even get to GTO) then refueled by two similar Starhip refueling flights to GTO to top it up. Fully fueled in GTO, and reserving enough fuel for the final landing burn on Earth return, it should be able to deliver about 255,000 tonnes of payload to the lunar surface, or about 17 times more payload mass than the uncrewed cargo version of Altair. The Artemis lander won’t even deliver as much payload as Altair because it will be designed for the SLS, which is not as capable as the Constellation program’s Ares V.
Between that and Blue Origin’s program, what NASA does with SLS and Orion may be irrelevant except as a program driver/tech developer/support resource.
***
That got a snarky reply, typical of non-aerospace blogs, so I elaborated on my thinking, which is admittedly speculative and based on the almost random set of unreliable specs that leak onto the Internet.
***
The analysis stemmed from a post about the likely early configuration of Starship/Superheavy, which will likely only have about 150% of the payload capability of Falcon 9 Heavy. Musk could get a bigger increase (163%, by his calculation) by just adding two more side boosters to F9H.
What I find important is that they thought enough about that five-booster configuration to analyze it and come up with performance numbers. It also shows that their basic configuration (no SRB’s) gives their vehicles a growth potential that the SLS must necessarily lack, because an SLS core doesn’t have enough thrust to even lift off the pad. It can’t grow sideways.
So that brought up a discussion that the Superheavy architecture was designed for in-orbit refueling. Well, given the early guesses as to the design, the early architecture is supposed to have a LEO payload capacity of 100,000 kg, and the ship has a dry weight of 87,000 kg. Landing fuel requirements probably set the minimum orbit mass at 103,000 kg.
So the assumption was that a Starship, fully fueled in LEO, could fly a lunar mission. I ran the numbers and found it could deliver -7,400 kg to the lunar surface, which is slightly less than nothing. But someone pointed out that Musk said it might have to be refueled in a high elliptical orbit for such a mission. That made sense, so I ran the numbers for a common elliptical orbit, a GTO (which is the insertion half of putting a geostationary satellite in position, before the satellite does a circularization burn).
The Starship’s Raptor engines have a specific impulse is 380 seconds, and that means the required mass ratio to insert into a geosynchronus transfer orbit is about two, which is nearly the same as the attainable mass ratio of Starship in the simple, straight-from-launch, low-earth orbit configuration. Oddly enough, despite its truly massive size, its high dry mass and return and landing requirements gives the Starship less GTO payload capacity than a Falcon 9. It would be almost useless for launching geostationary satellites.
That also means that Starship/Superheavy can’t refuel a Starship in GTO because they can’t deliver any useful amount of fuel (50 to 100 refueling flights just to top up one ship).
The only way around that fundamental problem (caused by hitting LEO with a high dry weight and low mass ratio), is to use a much bigger booster. The simplest way to do that is to use three instead of one, which is something SpaceX already does with Falcon Heavy.
This leads me to believe that the early configuration of Starship is basically a bigger LEO Space Shuttle (Four times more payload capacity), and the deep space capability will be added to the booster (Super Heavy), not to Starship.
Strangely enough, although landing Starship on the moon on its tail is pretty daft, the surface payload capability is almost exactly the same as if it parked in low lunar orbit and discharged a huge, one-way hypergolic lander with an ISP of 310 seconds and a cargo to landed-weight ratio of 75%.
There are of course some serious stumbling blocks to refueling in GTO. For one, nobody has really ever refueled anything in space yet, much less multiple times. For another, sending the crewed Starship to GTO first would mean the astronauts would be zipping all the way through the Van Allen Belt every five hours while waiting for multiple refueling rendezvous that might be days or weeks apart. Flying through the Van Allen belt twice exposed Apollo crews to a pretty big dose of radiation, so much that they saw spots as their optic nerves fired off. So it’s certainly not a mission plan, just a potential capability.
But what it points out is how obsolete NASA Artemis architecture is. It would take nine to twelve Superheavies (three boosters in each launch) to launch three or four Starships to execute one lunar mission that delivers 255,000 kg of cargo to the lunar surface. That mission would probably deliver more payload than about twenty SLS lunar cargo missions, so each Superheavy booster is accomplishing about twice the useful work (useful pounds to the moon) as the SLS. More importantly, none of those boosters, nor a Starship, nor anything else gets thrown away. They can all be used over and over again, so the only thing being expended is fuel.
In contrast, each SLS mission throws away the incredibly expensive core booster that takes four years to build, two solid rocket boosters, an upper stage that isn’t even flying yet, but which will likely use four expensive RL10 engines, and of course an Orion (which isn’t reusable) and at least a lander descent stage. The cost estimate for launching just a basic SLS Orion mission is somewhere between $1.5 and $2.5 billion dollars. A lunar landing mission might hit $3 to $6 billion, and those are recurring costs for each mission.
The Starship approach might end up with a staggering cost advantage per pound delivered, perhaps as high as 50:1, based on Starhip being more economical than Falcon Heavy, FH vs Falcon 9 in $/LEO-lb, how many boosters have to be launched, etc.
One approach might conceivably lead to commercial lunar development, and even permanent settlement. The other might put twelve more people on the moon before being abandoned due to high cost and low returns.
But Bridenstine is boxed in by all the existing bureaucratic and technical structures of NASA human space flight program, including the Congressional mandates, and there’s really no good way to break out of it. If Starship or Bezo’s New Glenn and New Armstrong start flying, we’d be much better off with NASA attaching themselves to those two programs, providing massive expertise, personnel, and mission support than trying to run a program that by its nature can’t really get past an exploration phase.
****
That analysis was looking at a maximalist architecture of giant re-usable rockets. There are also interesting minimalist architectures where you try to come up with the lightest possible crewed lander to see if you can squeeze useful capabilities out of the cheapest possible launch vehicles.
What all these alternatives require, however, is parallel approaches and out of the box thinking, which is almost the opposite of what large NASA funding levels and a tight schedule will produce. They’ll have to go with a pretty generic architecture that’s in line with existing proposals, based around components they already have (like Orion), with little time or resources available to “explore the space” of possible designs and architectures, since those will mostly document “What we might have done if we weren’t in such a hurry.”
But getting back to Wayne Hales point, the big advantage of taking the commercial route is that any lives at risk would likely be contractors, and as we learned in “Return of the Jedi”, nobody cares how many contractors get blown up in space or we wouldn’t have cheered the destruction of the still unfinished second Death Star.
Whew! I made the all-important topical tie in at the last second.
Interesting.
Yes, not much of a selling point. The example of the Oregon Trail is not a great persuasion tactic. While there will be deaths, the environment is so different that historical comparisons fall apart to some degree. For example, running out of food isn’t something that should happen. We will have new things happen and while some might be catastrophic, I suspect many will be rather mundane and solutions to take a long time to develop and it will all be very very boring.