totm dec 2023 Artemis Discussion Thread

[sevenperforce]

A further observation...

Given the high dV cost of taking anything from NRO/EM2/DRO down to the surface, there is a strong forcing factor in preference for placing surface assets directly, without a stopover.  

[RCgothic]

28 minutes ago, tater said:

That's the problem with designing a really, really expensive rocket and spacecraft without any particular goal in mind.

This was the original complaint many of us had about SLS, years ago. Not that it was overly expensive. Not that it had a low launch cadence. Not that it was primarily about pork. Not that it was past due and way over budget—though it is all of those things. Our complaint used to be summed up as "rocket to nowhere."

Too big for LEO, to small for anything else. That remains the primary problem. The other issues like cost—those only matter because if it was much cheaper, if it had a much higher cadence on top of that, then we could work around the fact it's simply not the right size via distributed launch.

 

This exactly.

Block 1B can't do Orion and a useful payload, so it's dead. Block 2 could, but it's so far over the horizon as to be effectively never.

So we're stuck with Block1.

Block 1 requires commercial support. Many launches of commercial support, with rendezvous somewhere around the moon. And if you're requiring commercial support then commercial rockets are more than capable of both putting the CSM in orbit *and* delivering separately a naked second stage with residuals to dock to and serve as an earth departure stage, and of doing it again for a lander. It's more moving parts, sure. More moving parts than rendezvousing at a gateway in LLO? Nope.

So what exactly is the point of Block1?

How many flights does it take to man-rate Falcon Heavy? 7? Guaranteed to cost less than EM1.

[sevenperforce]

Let's drill down on this optimization problem a little farther.

Suppose we take the following assumptions as givens:

• SLS will never launch anything but Orion and a uselessly small payload;
• Orion will never reach anything but NRO/EM-2/DRO;
• Orion will always need an abort-to-Earth option; and
• Our goal is lengthy, manned polar surface missions.

With these assumptions, we articulate our forcing function: we want our missions to achieve maximum cadence with the lowest cost. The following factors will impact this:

• Cost amortization won't help us much because NASA will be depending on commercial partners, who will absorb a good deal of the development cost.
• Component reuse (whether via commonality of components or actual hardware reuse) can help to improve mission cadence even if NASA does not benefit in absolute cost. 
• Mission risk is folded into mission cadence; if you have an architecture that cannot accommodate schedule slip, your overall mission cadence drops.

Are we agreed so far?

I think, given the high dV cost of going to the Orion rendezvous stopover, all surface components MUST be emplaced in advance using direct Hohmann transfer rather than stopping off midway. 

Given these variables, component commonality seems strongly preferred. I think that one of the major elements which can be broadly useful in mission design is a cryogenic, high-performance, expendable tug with no more than a week of persistence. Such a tug could be used in two different ways: both to deliver human-rated mission components from LEO to the Orion rendezvous point and to deliver surface components from LEO to a suborbital lunar staging point. 

Any components used to deliver crew from the Orion rendezvous point (with or without involving LOP-G) to the lunar surface need to be as small as possible, due to the significant delta-v costs of braking into the rendezvous point and then proceeding to the lunar surface. Going from TLI directly to low lunar orbit costs 900 m/s, while going from TLI to LOP-G (or equivalent) and then to low lunar orbit costs 1,160 m/s, a 30% dV penalty. If the cryogenic tug is very large, it could be used to deliver very large surface assets as well as delivering an entire descent/ascent mission stack to the Orion rendezvous point. If it is small, it could be used to deliver smaller surface assets and individual components of the descent/ascent mission stack. This approach offers a nice constraint on mission building; whatever is delivered to the lunar surface will have a specified mass ratio with respect to whatever is delivered to the Orion staging point.

[sevenperforce]

For reference: if we assume that a surface component delivered to a suborbital lunar staging point will need to provide the last 300 m/s of its own braking (not including margins for hover and touchdown), then a cryogenic tug will need to provide a total of 5.67 km/s to take it from LEO to lunar suborbital staging. Crew modules (whether delivered in piecemeal or all at once) will need to get a total of 3.63 km/s from LEO to the Orion rendezvous point.

• If the cryogenic tug is kerolox, then surface assets can be up to 55% the mass of whatever is delivered to the Orion rendezvous point.
• If the cryogenic tug is methalox, then surface assets can be up to 58% the mass of whatever is delivered to the Orion rendezvous point.
• If the cryogenic tug is hydrolox, then surface assets can be up to 63% the mass of whatever is delivered to the Orion rendezvous point.

That should give us a good starting point.

[jinnantonix]

On 2/13/2020 at 2:12 AM, sevenperforce said:

Let's drill down on this optimization problem a little farther.

Suppose we take the following assumptions as givens:

• SLS will never launch anything but Orion and a uselessly small payload;
• Orion will never reach anything but NRO/EM-2/DRO;
• Orion will always need an abort-to-Earth option; and
• Our goal is lengthy, manned polar surface missions.

With these assumptions, we articulate our forcing function: we want our missions to achieve maximum cadence with the lowest cost. The following factors will impact this:

• Cost amortization won't help us much because NASA will be depending on commercial partners, who will absorb a good deal of the development cost.
• Component reuse (whether via commonality of components or actual hardware reuse) can help to improve mission cadence even if NASA does not benefit in absolute cost. 
• Mission risk is folded into mission cadence; if you have an architecture that cannot accommodate schedule slip, your overall mission cadence drops.

Are we agreed so far?

I think you have nailed the constraints for Artemis perfectly.
 

Quote

Given these variables, component commonality seems strongly preferred. I think that one of the major elements which can be broadly useful in mission design is a cryogenic, high-performance, expendable tug with no more than a week of persistence. Such a tug could be used in two different ways: both to deliver human-rated mission components from LEO to the Orion rendezvous point and to deliver surface components from LEO to a suborbital lunar staging point. 

In many of the speculations around best HLS design, it has been pointed out that the most efficient design is a 3-stage lander using a transit vehicle for the first stage to get from NRHO to LLO.  Much talk of re-usability, nuclear engines, etc for this transit stage.  However beefing this up so it can act as a general purpose tug for various LEO to cislunar operations, and going cryogenic and expendable, makes sense.  Being able to use this to craft to deliver payloads directly to the lunar surface would be an alternative to the lander carrying payloads, and allows the SLS/Orion/HLS design to focus on just humans (and all it's safety overheads).

So a potential high level mission design for Artemis 3 could be:

• (1) Commercial carrier launches tug and static payload direct to lunar surface.  Includes rover, science instruments, emergency habitation etc.
• (2) Commercial carrier launches tug + HLS (lander ascent and descent vehicle) to NHRO or lowly elliptical staging point
• (3) SLS/Orion delivers humans for LOR at staging point, shortly after HLS system arrival.  
• Humans transfer to lander, and tug delivers lander to LLO or suborbital path to lunar surface landing site, tug expended
• Expendable HLS descent vehicle carries the humans to the lunar surface
• After surface operation HLS ascent vehicle has sufficient dV to return directly to the stage point.  Re-usable AV?
• Transfer to Orion and return to Earth.

So, one SLS launch and two large commercial launches, with one LOR and no fuel transfer required to complete the mission.  If the AV is reusable, launch 2 would carry fuel + pumping facilities and provisions instead of the AV.

Edited February 13, 2020 by jinnantonix

[tater]

ACES would be a transformational system if they just built the bloody thing already.

ACES can deliver ~35t to LLO expendable. More to Gateway.

It can likely bring something like 20t to Gateway—and still have the dv to return to LEO propulsively. More if it was rigged for aerobraking. This is only really useful for return mass (whatever that would do to mass to cislunar to leave residual props) as retanking requires another ACES, so might as well expend it at that point.

Some other upper stage with this concept in mind (IVF) should really be a focus. Dumping EUS in favor of an agnostic upper stage tech built along these lines would be a far better way to spend money.

 

That's all for the nominal 5.4m ACES, BTW. Bigger is possible.

[jinnantonix]

ACES has the rights specs.  But ULA is busy getting the Vulcan ready, and doesn't appear to be interested in bidding for any of the contracts.

[tater]

1 hour ago, jinnantonix said:

ACES has the rights specs.  But ULA is busy getting the Vulcan ready, and doesn't appear to be interested in bidding for any of the contracts.

The problem is that ULA won't compete with one of their parent companies.

 

[sevenperforce]

On 2/12/2020 at 6:33 PM, jinnantonix said:

I think you have nailed the constraints for Artemis perfectly.

Thank you!

On 2/12/2020 at 6:33 PM, jinnantonix said:

In many of the speculations around best HLS design, it has been pointed out that the most efficient design is a 3-stage lander using a transit vehicle for the first stage to get from NRHO to LLO.  Much talk of re-usability, nuclear engines, etc for this transit stage.  However beefing this up so it can act as a general purpose tug for various LEO to cislunar operations, and going cryogenic and expendable, makes sense.  Being able to use this to craft to deliver payloads directly to the lunar surface would be an alternative to the lander carrying payloads, and allows the SLS/Orion/HLS design to focus on just humans (and all it's safety overheads).

The reason for the three-stage approach was that NASA seemed to be dead set on assembly at LOP-G, where each component would be boosted to TLI by its cryogenic launch vehicle and then brake itself into NRO for rendezvous with LOP-G. The three-stage design was forced solely because no commercial LV has a large enough lunar throw to get a component large enough to brake itself into NRO and then perform LLO transfer and lunar descent with the ascent vehicle. If we had larger commercial launch vehicles, a two-component, separate-launch architecture or even a monolithic two-stage architecture would be vastly preferred.

Introducing an expendable, cryogenic tug on the front end rather than a hypergolic tug on the back end changes a lot of the math. Depending on the size of the tug and the capability of your launch vehicles, there are numerous ways to solve the problem:

• Constellation Redux
  • Launch descent and ascent modules, already mated, into LEO
  • Launch cryogenic tug to LEO and rendezvous
  • Cryogenic tug performs TLI
  • Either cryogenic tug or descent module performs insertion at NRO
• Apollo-C
  • Launch an ascent module and a descent module to LEO in two separate launches
  • Mate the two modules together in LEO
  • Launch the cryogenic tug to LEO and rendezvous
  • Cryogenic tug throws stack to TLI
  • Either cryogenic tug or descent module performs insertion at NRO
• Artemis Lite
  • Use a commercial vehicle to send lightweight ascent vehicle to TLI
  • Ascent vehicle brakes for its own insertion at NRO
  • Use a commercial vehicle to send lightweight descent stage to LEO
  • Launch cryogenic tug to LEO and rendezvous
  • Cryogenic tug throws stack to TLI
  • Cryogenic tug performs insertion at NRO, rendezvous with ascent stage
  • Orion reaches NRO on SLS
  • Cryogenic tug uses residuals for transfer to LLO
• Crasher Chaos
  • Use a commercial vehicle to send a reusable crew capsule to TLI where it brakes itself into NRO
  • Use a commercial vehicle to place a landing/ascent stage in LEO
  • Launch cryogenic tug to LEO and rendezvous
  • Tug performs TLI and injection; loiters until Orion arrives
  • Tug performs transfer and most of descent

In the last version the tug needs better loiter or ZBO.

On 2/12/2020 at 6:33 PM, jinnantonix said:

So a potential high level mission design for Artemis 3 could be:

• (1) Commercial carrier launches tug and static payload direct to lunar surface.  Includes rover, science instruments, emergency habitation etc.
• (2) Commercial carrier launches tug + HLS (lander ascent and descent vehicle) to NHRO or lowly elliptical staging point
• (3) SLS/Orion delivers humans for LOR at staging point, shortly after HLS system arrival.  
• Humans transfer to lander, and tug delivers lander to LLO or suborbital path to lunar surface landing site, tug expended
• Expendable HLS descent vehicle carries the humans to the lunar surface
• After surface operation HLS ascent vehicle has sufficient dV to return directly to the stage point.  Re-usable AV?
• Transfer to Orion and return to Earth.

So, one SLS launch and two large commercial launches, with one LOR and no fuel transfer required to complete the mission.  If the AV is reusable, launch 2 would carry fuel + pumping facilities and provisions instead of the AV.

One thing to note is that no commercial launch provider is going to be able to send both the HLS and the tug to LEO. The cryogenic tug will need to be launched separately.

The cool thing is that if the tug is delivering surface assets, you don't necessarily have to have a 1-to-1 correspondence between the surface asset placements and the missions. You can put stuff down way in advance, and you might have missions without any surface-related launches at all.

[tater]

That NASA slide I posted that shows the different lander types (1 stage, 2 stage, 3 stage) is predicated on a large lander (Ascent stage min 9 tons). This is presumably related to:

1. Abort to NRO requirements.

2. Crew capacity.

3. Surface stay requirements (including abort timing, 1 per 6.5 days).

I think much of this is related to crew size, and life support for extended stays (and having to double this for even 1 launch window abort—if there was an issue at 6.5 days, they'd need another 6.5 days to try again). My solution up the thread could be added to the excellent list just posted by @sevenperforce. A much smaller lander system designed for one of the profiles (1, 2, or 3 stage) that delivers the crew to a hab that was already placed, and doesn't have to leave the surface, ever. About 54% of the mass of the Apollo LM was propellant/consumables. For that 9 tonne Ascent stage, it needs to have a dry mass of no more than ~3.6t to be able to make it back to Gateway (~1.7X the LM dry mass). A smaller Ascent stage could deliver crew to a pre-deployed habitat. Secure lander. Suit up, egress. Do initial EVA. Return to hab module, live there, EVA over days. Return to lander to go to Gateway. If the lander was slightly larger than the Apollo LM (crew of 3-4), say 2.4t, it could have a wet mass of only 6t. That drops the Descent stage substantially, to the point a 2-stage can be delivered by commercial LVs.

FH can throw what, 21-22t to TLI?

This would allow an 8t (dry mass) habitat to be landed on the lunar surface assuming hypergolic props. 9.5t dry if CH₄ is the fuel. more like 11t dry if it was hydrolox. This is a decent sized hab. On the small end of those numbers, we have a hab that is 7X the volume of the LM. I used the volume/mass numbers from the ISS airlock module for that calc, since a decent chunk of that vol needs to be the airlock. If the airlock portion took the same 34m2, that would leave 2X the Apollo LM as crew hab volume (that could likely be optimized, didn't really need to be for ISS, the module was well below what Shuttle could carry).

Edited February 15, 2020 by tater

[tater]

Useful chart:

[sevenperforce]

4 hours ago, tater said:

This would allow an 8t (dry mass) habitat to be landed on the lunar surface assuming hypergolic props. 9.5t dry if CH₄ is the fuel. more like 11t dry if it was hydrolox. This is a decent sized hab.

If you have a high-performance cryogenic tug performing both TLI and the initial braking burn, then you can launch a very respectably-sized hab (maybe even a hab and pressurized rover combo) into LEO, loiter until the tug shows up to ferry it to suborbital lunar descent, then use its own hypergolic propulsion to perform the final braking and landing.

[tater]

Just now, sevenperforce said:

If you have a high-performance cryogenic tug performing both TLI and the initial braking burn, then you can launch a very respectably-sized hab (maybe even a hab and pressurized rover combo) into LEO, loiter until the tug shows up to ferry it to suborbital lunar descent, then use its own hypergolic propulsion to perform the final braking and landing.

Yeah, a tug is transformative.

[sevenperforce]

1 minute ago, tater said:

Yeah, a tug is transformative.

Doesn't even have to be ACES. Can be any cryo stage with multiple restarts, a docking ring, pointing support, and at least 3-4 days of loiter.

Edited February 15, 2020 by sevenperforce

[tater]

Just now, sevenperforce said:

Doesn't even have to be ACES. Can be any cryo stage with multiple restarts, a docking ring, pointing support, and at least 3-4 days of loiter.

Yeah, Could even be a stage 2 mod for extant second stages.

[sevenperforce]

2 minutes ago, tater said:

Yeah, Could even be a stage 2 mod for extant second stages.

That was my thought. Grab an existing second stage, mod it slightly, and slap it on top of your preferred launch vehicle. The upcoming M10-based cryogenic stage for Vega-E would be a methalox option. You could also use a mildly modified DCSS or Centaur. Heck, you could even put a cluster of Electron upper stages into play, if desired.

[tater]

"Naked FH" stage 2, NG, yeah, many possibilities. Bottom line is send hab direct, or via EOR with tug. Other CLPS flights land rovers, etc. Crew lander becomes minimal vehicle to safely transport crew to the surface and return. The monkey wrench in the system is likely abort modes. Given the (awful) placement of the orbital facility in a highly elliptical orbit relative to the polar base, I wonder about what the abort constraints are. Ie: if they have to abort during their descent to the surface, can they in fact rendezvous immediately with Gateway, or are they stuck in a position where they have to achieve some holding orbit, then wait for the next Gateway pass in 6+ days?

Edited February 15, 2020 by tater

[sevenperforce]

21 hours ago, tater said:

Bottom line is send hab direct, or via EOR with tug. Other CLPS flights land rovers, etc. Crew lander becomes minimal vehicle to safely transport crew to the surface and return.

I think that's broadly the best plan.

21 hours ago, tater said:

The monkey wrench in the system is likely abort modes. Given the (awful) placement of the orbital facility in a highly elliptical orbit relative to the polar base, I wonder about what the abort constraints are. Ie: if they have to abort during their descent to the surface, can they in fact rendezvous immediately with Gateway, or are they stuck in a position where they have to achieve some holding orbit, then wait for the next Gateway pass in 6+ days?

One of the reasons for placing LOP-G in NRO was this exactly. Assuming a polar surface mission, NRO is the best abort destination at almost any point. The only better destination for abort during descent is a frozen low polar orbit, but since Orion has no chance of hitting LLO and getting back to Earth entry interface, it was out of the question.

If you're on the surface at either pole, you can abort to NRO at literally any moment. Some phasing is better than others in terms of dV margins, but it's all pretty close. I am fairly sure that you can abort during descent at any time too without phasing issues.

[tater]

13 minutes ago, sevenperforce said:

If you're on the surface at either pole, you can abort to NRO at literally any moment. Some phasing is better than others in terms of dV margins, but it's all pretty close. I am fairly sure that you can abort during descent at any time too without phasing issues.

Gotcha, so it's just a dv issue. The constraint is that the ascent stage needs enough dv to deal with all possible abort modes (which including crew requirements is within NASA's minimum 9t ascent stage assumption).

That's assuming it's actually a base, and not flags and footprints.

[jadebenn]

On 2/11/2020 at 3:29 PM, RCgothic said:

Block 1B can't do Orion and a useful payload, so it's dead.

10 tons is significant. It's toughly equivalent to the capacity of an expendable FH launch when you factor in that Orion means you don't need to budget mass for a tug in the CPL.

On 2/12/2020 at 9:12 AM, sevenperforce said:

• SLS will never launch anything but Orion and a uselessly small payload;

I wouldn't bet on this. Remember, Doug Loverro is reviewing the Artemis architecture, and is projected to release the results of his deep dive in mid-March. We might be seeing some changes...

Edited February 16, 2020 by jadebenn

[tater]

1 hour ago, jadebenn said:

10 tons is significant. It's toughly equivalent to the capacity of an expendable FH launch when you factor in that Orion means you don't need to budget mass for a tug in the CPL.

It's significant, but not terribly useful. Certainly not given the dev (cost and time) required to get EUS functional. Dev alone will rack up billions—more than enough for many commercial LV launches (dozens?). By useful it's simple math: Can SLS B1b fly a complete lunar surface mission in one flight, yes or no? If the answer is no, then the mission still requires multiple launches—in for a penny, in for a pound.

 

1 hour ago, jadebenn said:

I wouldn't bet on this. Remember, Doug Loverro is reviewing the Artemis architecture, and is projected to release the results of his deep dive in mid-March. We might be seeing some changes...

Well, maybe they can make sense of it, but the fundamental limit is the rocket equation. The Orion capsule is too heavy. If Orion is on top, SLS is never capable of doing a useful crew mission by itself. CST-100 has a total mass of 13t (including the little SM). The capsule (had to find mass for just that) is certainly lighter than Orion, and the useful interior volume is similar.  Dump Orion, SLS is more useful, else keep Orion, and regardless of the mission, distributed launch is required.

Look, if SLS had an even halfway reasonable cost, it wouldn't matter. The issue is that a single SLS launch puts up 2X or more the payload, but costs 20X as much to do so. $/kg.

For the same payload mass, 300M vs 10X that (3 billion a flight is being kind to SLS, BTW, the 300M is retail, you have to include all program costs, even ignoring dev). Assembly at ISS, for example, is 100% doable for a lunar mission. NASA could buy commercial launches in bulk and get a deal (spread around to all providers), then it's a matter of how many hundred tons they could fly to LEO for the price of a single SLS flight.

[tater]

 

[tater]

The TLDR version is that the plan Berger links to is basically giving the store to Boeing, but happily JB says that is not the plan.

[RCgothic]

Wow, that's a major public rebuke.

[tater]

So this is F9/FH with an extended fairing. This is of course combined with the ability to do vertical integration at 39A.

FH loft to LEO seems to be:

Fully expended: 63.8 t

Center core expended: 57.4 t (per Elon tweet)

Full reuse: 18 -28.4 t (nothing explicit said about it, so an estimate range is what I see).

At the upper end of full reuse, Orion could be sent to LEO on FH. Without question it could with the center core expended. Regardless of Orion the larger fairing permits a broader range of possible payloads. Looks to be about like the ~12m long Atlas V fairing.