Using upper stages for distributed Artemis - reference figures

[sevenperforce]

The Artemis request for proposals for a Human Landing System (HLS) is conspicuously constrained by the throw weight of any currently-flying commercial HLV. We can't do any HLS component larger than 15 tonnes, NASA says, because 15 tonnes is the max throw to TLI.

In another forum, a few folks have been going over the possibility of launching a lunar sortie vehicle into LEO and then using distributed launch to send it to TLI or beyond. The upper stage of a Falcon Heavy, a Delta IV, an Atlas V, or Vulcan could be launched empty with a docking ring and use its residuals to provide TLI. In some cases, this upper stage could perform the powered lunar flyby, the NRHO insertion, or even the transfer to LLO as imagined by the HLS request.

To save us time in comparing proposals, I went ahead and threw together some reference numbers for what kind of residuals can be expected if someone wanted to use a "naked" HLV upper stage to perform TLI or any number of BLEO ops. To calculate, I've taken published or estimated GTO performance figures for various vehicle configurations and done the math in reverse to calculate staging velocity, then "removed" the payload and calculated forward to get to LEO. As such, these estimates contain some degree of conservatism; the absence of a payload during primary ascent will improve performance marginally. Buoyed by this, I've also ignored the necessary addition of a docking ring: this is likely a safe assumption given that the upper stage will need no payload adapter or decoupling mechanism.

By the time Artemis is realized, Atlas V will likely no longer be flying, but I included it for reference anyway.

Assumptions were 2270 m/s from LEO to GTO, 3200 m/s from LEO to TLI, 430 m/s from TLI to NRHO. I estimated the propellant load of Centaur V by comparing the volume of Centaur V to the volume of the Atlas V Centaur; I estimated the dry mass of Centaur V by subtracting the mass of the RL-10 from the dry mass of the Atlas V Centaur, increasing mass in proportion to fractional surface area, and adding two RL-10s back in.

[TaurusLittrow]

Thanks for the calculations. So the eFH is the only vehicle among those evaluated to place 15 mT in NRHO, which is the wet mass of the Descent Element and Transfer Module as identified in NASA BAAs.

Of course, the Lander Elements could contribute to TLI, which would widen the cLV options.

It would be useful to have similar calculations for New Glenn given its likely "prominent" role in the BO/LM/NG/D lander achitecture.

[spacenut]

Using the existing launchers with docking adapters attached to upper stages could throw more payload to TLI than SLS at a lower price tag.  Especially in expendable mode.  Two to four launches for existing is still way less expensive than SLS. 

Thanks for the chart.  This opens a new can of worms for NASA. 

[sevenperforce]

Thanks for the calculations. So the eFH is the only vehicle among those evaluated to place 15 mT in NRHO, which is the wet mass of the Descent Element and Transfer Module as identified in NASA BAAs.

Of course, the Lander Elements could contribute to TLI, which would widen the cLV options.

They then would need to be refueled at LOP-G, which is yet another bridge. NASA seems to think it is fine to throw a 15-tonne vehicle to TLI, let it brake itself into NHRO, and then top it up. But if that's okay, then why not throw a 50-tonne single-stage lunar lander halfway to GTO, let it use its onboard dV to complete, and then refuel it? It makes no sense.

Using a high-performance upper stage for the lunar flyby and NRHO injection is far more efficient regardless of whether you intend to refuel your vehicle at the gateway or not.

It would be useful to have similar calculations for New Glenn given its likely "prominent" role in the BO/LM/NG/D lander achitecture.

I would have loved to have included New Glenn, but we simply don't know enough about it yet. It claims 13 tonnes to GTO, which puts it slightly under Vulcan 6, DIVH, and FH 2r, but we don't know anything about its dry mass, propellant load, or specific impulse. We know its diameter is 7 meters, and promotional images promulgated by Blue Origin suggest a 19-meter upper stage, which would give it something like eight times the volume of Centaur on Atlas V. But there's simply nothing that gives us any indication of actual dry mass/structural fraction, propellant load, RCS propellant, or specific impulse. Hell, we don't even know propellant mixture ratio. The most I could do is a very gross estimate that it has naked BLEO performance roughly commensurate to Delta IV Heavy. Useful, in a sense, but not really.

Using the existing launchers with docking adapters attached to upper stages could throw more payload to TLI than SLS at a lower price tag.  Especially in expendable mode.  Two to four launches for existing is still way less expensive than SLS. 

Thanks for the chart.  This opens a new can of worms for NASA. 

An expendable Falcon Heavy has naked BLEO throw roughly commensurate to launching ordinary payload on SLS Block 1 Cargo, but falls well short of SLS Block 1B or anything above. That being said, what do we actually need to throw to TLI?

The question, of course, is whether you are going to let the payload brake itself into NRHO, or use the upper stage? Boil-off is not really significant, even for hydrolox, if you are going straight to NRHO. It is way more efficient to use the high-performance upper stage to do NRHO insertion unless, of course, you are an idiot using the EUS.

[w9gb]

I would have loved to have included New Glenn,

I use the old Saturn C-3 data as a “placeholder”,
https://en.wikipedia.org/wiki/Saturn_C-3
since this capability appears to have been Jeff Bezos initial goal for New Glenn.
==
The Vulcan Centaur Heavy (extended length Centaur 5) was not profiled (planned for 2023).
The USAF NSSL requirements for direct GEO was primary driver for design (and replace Delta Heavy).
https://www.spacelaunchreport.com/vulcan.html

April 20, 2018, Tory Bruno revealed that there would be two variations of the new Centaur 5 stage -
an "initial" and a "stretched" version.  These would use RL10-C engines, with an upgrade (5+) to RL10-CX

The Vulcan Centaur 5 version, using the stretched Centaur 5 stage would be denoted "Vulcan Centaur Heavy" in a released graphic.  It would be able to lift 6.8 tonnes to geosynchronous orbit, 15.88 tonnes to geosynchronous transfer orbit, or 35.88 tonnes to low earth orbit.

The Centaur 5 Usable Propellant Mass (tonnes) of 54.43 t,
would be increased to 77.11 t for the “Stretched” Centaur 5 (42% increase).

[TrevorMonty]

I'd launch payload on either Vulcan or Omega as they both have light hydrolox US with around 50mt of fuel capacity. Use NG to refuel them in LEO with 40-45mt of fuel. If just using Vulcan and Omega then use Vulcan for tanker as it has higher LEO performance of +30mt.

Issue with SpaceX is lack of LH support and they don't play well with others.

What ever combination you go for, DL is best done with 2 separate LVs so they can be launched days if not hours apart.

Vulcan + NG should be able to deliver Orion to Gateway, definitely in LM's interest but Boeing wouldn't like it as it bypasses SLS. Not sure Boeing could stop LM from doing this.

[TomH]

I was pleasantly surprised to see so much cooperation in the National Coalition. In that spirit, I will throw out something that I would never before have considered a possibility. It still is something I consider highly improbable, but here it is anyway:

SX Super Heavy with disposable stripped down SS as S2. Payload is ACES which arrives in orbit fully fueled, ready to dock and provide ΔV for TLI and possibly even LOI.

[Steven Pietrobon]

I estimated the dry mass of Centaur V by subtracting the mass of the RL-10 from the dry mass of the Atlas V Centaur, increasing mass in proportion to fractional surface area, and adding two RL-10s back in.

Pressurised tank mass is not proportional to surface area. Its proportional to volume!

https://en.wikipedia.org/wiki/Pressure_vessel

[sevenperforce]

I estimated the dry mass of Centaur V by subtracting the mass of the RL-10 from the dry mass of the Atlas V Centaur, increasing mass in proportion to fractional surface area, and adding two RL-10s back in.

Pressurised tank mass is not proportional to surface area. Its proportional to volume!

https://en.wikipedia.org/wiki/Pressure_vessel

That would certainly be the case for tanks for a pressure-fed engine, but Centaur is not a pressure-fed stage. It is certainly pressed above ambient but not to the degree that it cannot take advantage of the square-cube law.

Even so this is obviously the most uncertain estimate of the table.

[sevenperforce]

I went back over this table because I wanted to re-evaluate Falcon Heavy's performance. I had used GTO figures because that was an apples-to-apples comparison across launch vehicles, but 44.8 tonnes of residuals seemed low given that Falcon Heavy can supposedly deliver 64 tonnes to LEO outright.

Falcon Heavy expendable is quoted at being able to deliver 16.8 tonnes to Mars. Assuming this is the 3600 m/s Mars Transfer Injection, this would require 39.8 tonnes of residuals in LEO, which corresponds to 71.7 tonnes of propellant burned between staging and LEO. If you imagine dropping the payload at staging and proceeding to LEO with a "naked" upper stage, simple application of the rocket equation means you would burn only 62.6 tonnes of propellant to reach LEO and would arrive with 48.9 tonnes of residuals, which is noticeably (though not overwhelmingly) better.

However, SpaceX also advertises delivery direct to Pluto of up to 3.5 tonnes. Pluto requires 8.2 km/s out of LEO, which would take 80.4 tonnes of propellant; this suggests 31.1 tonnes burned and staging just 1,031 m/s short of LEO. If you dropped the payload at staging, you'd only burn 30.2 tonnes to reach LEO and you'd end up with an unbelievable 81.3 tonnes of residuals in LEO for distributed launch.

Is this right? Can this possibly be right?

If that's true, then that's enough to throw up to 47.8 tonnes of distributed-launch payload to TLI. It's enough to deliver 38.3 tonnes of distributed-launch payload direct to LOP-G, if SpaceX made similar "Frankenstage" modifications as with Falcon Heavy's first launch. That's more delivery capability directly to LOP-G than SLS Block 1B can deliver to TLI.

[brickmack]

Do we know if that Pluto figure is for a stock FH? SpaceX has proposed a few unique configurations for high-energy missions before, with either a stretched upper stage or a Star kick stage. The 16.8 tons to Mars figure is definitely for an unmodified expendable FH, per statements from when Red Dragon was still a thing

[sevenperforce]

Do we know if that Pluto figure is for a stock FH? SpaceX has proposed a few unique configurations for high-energy missions before, with either a stretched upper stage or a Star kick stage. The 16.8 tons to Mars figure is definitely for an unmodified expendable FH, per statements from when Red Dragon was still a thing

I have seen us propose a stretched upper stage or a Star kick stage. I have never seen SpaceX suggest it.