Starship On-orbit refueling - Options and Discussion

[Paul451]

I hadn't seen this before. Is it in one of the other threads?
 The depot looks big enough for two starship loads.

Not quite. But (as discussed throughout this thread) it is enough for the depot to refuel in circular-LEO, then push itself into the highest (safe) eccentric orbit which LSS can reach from launch. (Even a LEO/GEO eccentric orbit should be possible. If we ignore the VA belts.)

The benefit is that LSS takes on a full load of propellant that "contains" the higher energy of that eccentric orbit. (Or rather, when the LSS does a perigee burn to TLI, it gets an Oberth boost, reducing the propellant cost to TLI.) This solves a lot of the concerns about mass budgets for the lunar landing architecture; especially the "sustainable" model.

From a mission architecture POV, it doesn't greatly increase risks (again ignoring the VA belts), because LSS still only does one RPOD for refuelling. Plus all depot-filling risks are completed and stable before LSS launches, which is before SLS/Orion launches, so the risk of mission delay isn't drastically increased.

[edit: words are hard]

[Bryan Hayward]

From a mission architecture POV, it doesn't greatly increase risks (again ignoring the VA belts), because LSS still only does one RPOD for refuelling. Plus all depot-filling risks are completed and stable before LSS launches, which is before SLS/Orion launches, so the risk of mission delay isn't drastically increased.

Re: VA belts - there is a spot between the belts that has a much lower radiation load. I suspect that is where they will do this "high" orbit refuel. It doesn't make sense to do it inside an Allen Belt if you don't have to.

[InterestedEngineer]

From a mission architecture POV, it doesn't greatly increase risks (again ignoring the VA belts), because LSS still only does one RPOD for refuelling. Plus all depot-filling risks are completed and stable before LSS launches, which is before SLS/Orion launches, so the risk of mission delay isn't drastically increased.

Re: VA belts - there is a spot between the belts that has a much lower radiation load. I suspect that is where they will do this "high" orbit refuel. It doesn't make sense to do it inside an Allen Belt if you don't have to.

The sweet spot is about 4 Earth Radii (4 ER), which is about 20,000km - 25,000km.   That gets you 2km/sec of deltaV boost at periapsis (added to LEO), which is about what is needed.

One still traverses the lower belts an additional time.

OTOH, there's plenty of mass for Starship to add radiation protection far in excess of what Orion's mass budget can do.

https://docs.google.com/spreadsheets/d/18EEOrUn7T_JUXHeNxYsa4M5896ogaUgBfOaNANfdrck/

[TheRadicalModerate]

I hadn't seen this before. Is it in one of the other threads?
 The depot looks big enough for two starship loads.
 *Teslarati*

That comes from an IAC talk that some of the NASA people gave, and it is referenced up-thread here somewhere.

I'm still skeptical that the depot is actually stretched, because it doesn't need to be stretched.¹  Mod whatever amount of boiloff is to be expected, it'll never need to be able to transfer more prop than however big the target's main tanks are.

Something that stretched has implications for the GSE.  Unless you can move the hard points down into the center of the tanks somewhere, I don't think the chopsticks can raise it up high enough to stack it on the booster.  Seems like a lot of work to make mid-tank hard points structurally sound enough to carry the (somewhat larger) dry mass.

______________
¹A few weasel words required here:
a) If you need to do two quick refills in a row, a big depot is handy.
b) If you have so much traffic that you have continuous transfers in and out of the depot, bigger is probably better.

However, I don't see either of those scenarios occurring until well after Option A.  There is a case in Option B refueling where, if your lift tankers only have 1200t tanks, you need two of them to haul prop out to NRHO for reuse, and doing so in rapid succession may have some operational simplicity. But 1200t tankers are kinda silly, because rearranging ring segments, domes, and inter-tank bulkheads is pretty easy, which yields a 1650t tanker with the same dry mass.  Then, even if you can't fill it completely full at launch, you can in LEO, and you can do the entire service mission to NRHO to get the LSS topped up for the next Artemis mission in one LEO-NRHO-transferProp-EDL pass.

[TheRadicalModerate]

From a mission architecture POV, it doesn't greatly increase risks (again ignoring the VA belts), because LSS still only does one RPOD for refuelling. Plus all depot-filling risks are completed and stable before LSS launches, which is before SLS/Orion launches, so the risk of mission delay isn't drastically increased.

From my soi disant "HEEO hater" perch, the VA belts¹ aren't the only issue.  If you're not circular, then your argument of perigee is actually important.  If you stick a depot up into that eccentric orbit, you dramatically narrow the set of departure windows it can service.  So, yeah, you've kept your RPOD risks low, but you've introduced window risks, which in turn increase your boiloff risks. 

On top of that, any RAAN misalignments become more painful, because you'll have to do plane changes for TLI at higher speeds, which increases their delta-v consumption.

There is a way to minimize some of this:  Put the depot in LEO, load it up, then have it boost itself into the target HEEO just before the LSS is due to launch--or even after it's launched and sitting in a VLEO parking orbit.  The the LSS can top itself off, and the depot can propulsively return to LEO for the next mission.

That takes some of the risk out, but if there's a failed RPOD, you consume a lot more prop re-kludging the ω and RAAN than if you just dealt with giving the LSS big enough tanks to do LEO-NRHO-LS-NRHO in the first place.  And you can rearrange segments to get a 1500t tankage on the LSS and still have a 2m-high cylindrical payload bay for airlocks and hatches, plus whatever ogive space the crew module doesn't take up.

However, Twark has mentioned that up-thread that re-jiggering the tanks on the LSS this way does indeed raise the landing center of mass.  I think I got about 2m higher.  That increases lunar landing risk.

As usual, it's a very rich trade space.

__________
¹I got a lot less certain about VA belt risks when I noticed that Arty 2 plans to put the crew into a 300x60,000km HEEO for checkout, before continuing on to the free-return using the Orion OMS pod.  If NASA's OK with dipping their astronauts four times through the VA (once for HEEO insertion, once pre-TLI, once post-TLI, and once for TEI-to-EDL), then maybe this isn't such a big deal.  Of course, exposing one crew to these risks, which are balanced against the risks of an ICPS underperformance leaving them with bad abort options if TLI goes wrong, is different from exposing every crew to them.

[TheRadicalModerate]

I hadn't seen this before. Is it in one of the other threads?
 The depot looks big enough for two starship loads.
 *Teslarati*

https://forum.nasaspaceflight.com/index.php?topic=50157.msg2406262#msg2406262

Depot seems to be roughly 61.5 meters tall.

So, while you've got your pixel-counting hat on, how much clearance is there before the chopsticks hit the top of their travel?  And if it's not enough, can you move the depot's chopstick hard points so that they're attached to a spot where there's nothing but sheet metal and stringers?

[TheRadicalModerate]

From a mission architecture POV, it doesn't greatly increase risks (again ignoring the VA belts), because LSS still only does one RPOD for refuelling. Plus all depot-filling risks are completed and stable before LSS launches, which is before SLS/Orion launches, so the risk of mission delay isn't drastically increased.

Re: VA belts - there is a spot between the belts that has a much lower radiation load. I suspect that is where they will do this "high" orbit refuel. It doesn't make sense to do it inside an Allen Belt if you don't have to.

If all you're doing is refueling an uncrewed LSS, then it matter--much--if your apogee is inside VA #1.

I got pelted with various brickbats when I suggested that having a single refueling RPOD might be a constraint for an Option A mission.  Most assume that you'd refuel once in LEO, then top off in HEEO, before continuing on to do HEEO-NRHO-LS-NRHO.  If that's the case, you might as well refuel in something GTO-like an have done with it.  However, RPODs have risks attached to them, and if you do want to limit things to a single refueling for Option A, then you're left with an apogee that's constrained by how high the LSS can get on the prop that it launched with.  That may easily be inside VA #1.

As for crewed refueling, you may be able to refuel in the gap, as long as you don't have RPOD problems.  If you do, then you dip your crew through the VA #1 twice for every additional orbit, until they decide to abort.  And even if everything goes peachy, you're still increasing crew exposure from a pair of VA #1 transits to four.

[Paul451]

If NASA's OK with dipping their astronauts four times through the VA (once for HEEO insertion, once pre-TLI, once post-TLI, and once for TEI-to-EDL), then maybe this isn't such a big deal.  Of course, exposing one crew to these risks, which are balanced against the risks of an ICPS underperformance leaving them with bad abort options if TLI goes wrong, is different from exposing every crew to them.

As for crewed refueling,

Not sure why you keep bringing this up. Crew xfer is in NRHO. LSS refuelling is uncrewed.

[TheRadicalModerate]

If NASA's OK with dipping their astronauts four times through the VA (once for HEEO insertion, once pre-TLI, once post-TLI, and once for TEI-to-EDL), then maybe this isn't such a big deal.  Of course, exposing one crew to these risks, which are balanced against the risks of an ICPS underperformance leaving them with bad abort options if TLI goes wrong, is different from exposing every crew to them.

As for crewed refueling,

Not sure why you keep bringing this up. Crew xfer is in NRHO. LSS refuelling is uncrewed.

Option A/B LSS is almost certainly the first application, but this is a general Starship refueling thread.  And beyond Option A and B, crewed LSS missions staged out of LEO, and eventually EDL-enabled missions launched direct from the ground, are all going to need a refueling architecture.

[TomH]

So, while you've got your pixel-counting hat on, how much clearance is there before the chopsticks hit the top of their travel?  And if it's not enough, can you move the depot's chopstick hard points so that they're attached to a spot where there's nothing but sheet metal and stringers?

In that the tower is modular, how difficult would it be to remove the top, add another section, and put the top back on? If there are multiple pads in the future, how difficult would it be to simply make the newer towers taller and launch variants of differing height from those taller towers?

[sdsds]

Some reflection and reconsideration of Starship refilling now that Artemis 1 is through TLI seems timely.

The success of Artemis 1 will likely mean NASA will continue to drink the Kool-Aid awhile longer regarding the Artemis 3 schedule, so they will now want SpaceX to demonstrate Starship refilling ASAP, since in NASA's mind refilling might be the long pole for an Artemis 3 landing. That pressure could increase the priority at SpaceX of refilling as compared to the priority of Starlink launches. (Personally I would like that outcome, even without fully believing the reasoning is correct.)

HLS Starship is going to require some special modifications that require work, even for an uncrewed landing demonstration mission. Thus I prognosticate a mission requiring refilling before LEO departure will appear on the Starship manifest prior to the launch of the first HLS Starship. (Again personally, I'm hoping that demonstration mission is Mars-bound in 2024.)

[DanClemmensen]

Some reflection and reconsideration of Starship refilling now that Artemis 1 is through TLI seems timely.

The success of Artemis 1 will likely mean NASA will continue to drink the Kool-Aid awhile longer regarding the Artemis 3 schedule, so they will now want SpaceX to demonstrate Starship refilling ASAP, since in NASA's mind refilling might be the long pole for an Artemis 3 landing. That pressure could increase the priority at SpaceX of refilling as compared to the priority of Starlink launches. (Personally I would like that outcome, even without fully believing the reasoning is correct.)

HLS Starship is going to require some special modifications that require work, even for an uncrewed landing demonstration mission. Thus I prognosticate a mission requiring refilling before LEO departure will appear on the Starship manifest prior to the launch of the first HLS Starship. (Again personally, I'm hoping that demonstration mission is Mars-bound in 2024.)

Recall that we have seen several allusions to a demonstration of on-orbit fuel transfer to be done between two tanks on the same ship, so we might see that before we see actual ship-to-ship transfer.

[Twark_Main]

Surely if you were going to lower the lifting hardpoints, the "natural" location would be the common dome seam. This area is already stronger, so it would require the least reinforcing mass.


SpaceX has shown that they're willing to bring in a portable crane for one-time (or otherwise "rare") operations, so I wouldn't lose any sleep over it. If altering the Starship lift points is somehow insurmountable (unlikely IMO), they already have a ready-made backup plan.

[DanClemmensen]

Surely if you were going to lower the lifting hardpoints, the "natural" location would be the common dome seam. This area is already stronger, so it would require the least reinforcing mass.


SpaceX has shown that they're willing to bring in a portable crane for one-time (or otherwise "rare") operations, so I wouldn't lose any sleep over it. If altering the Starship lift points is somehow insurmountable (unlikely IMO), they already have a ready-made backup plan.

Also recall that Depot won't land, so these lift points are only used to lift the empty Depot gently and under controlled conditions. They are not used to catch a descending Depot. Since Depot launches are rare, they can also rig an external harness to allow the chopsticks to do the lift.

[Barley]

If NASA's OK with dipping their astronauts four times through the VA (once for HEEO insertion, once pre-TLI, once post-TLI, and once for TEI-to-EDL), then maybe this isn't such a big deal.  Of course, exposing one crew to these risks, which are balanced against the risks of an ICPS underperformance leaving them with bad abort options if TLI goes wrong, is different from exposing every crew to them.

As for crewed refueling,

Not sure why you keep bringing this up. Crew xfer is in NRHO. LSS refuelling is uncrewed.

Option A/B LSS is almost certainly the first application, but this is a general Starship refueling thread.  And beyond Option A and B, crewed LSS missions staged out of LEO, and eventually EDL-enabled missions launched direct from the ground, are all going to need a refueling architecture.

By then there will have been dozens, perhaps hundreds, of refuelings.  Maybe the handwringing over RPODs will go away.  Just like the handwringing over load-and-go.

[OTV Booster]

Some reflection and reconsideration of Starship refilling now that Artemis 1 is through TLI seems timely.

The success of Artemis 1 will likely mean NASA will continue to drink the Kool-Aid awhile longer regarding the Artemis 3 schedule, so they will now want SpaceX to demonstrate Starship refilling ASAP, since in NASA's mind refilling might be the long pole for an Artemis 3 landing. That pressure could increase the priority at SpaceX of refilling as compared to the priority of Starlink launches. (Personally I would like that outcome, even without fully believing the reasoning is correct.)

HLS Starship is going to require some special modifications that require work, even for an uncrewed landing demonstration mission. Thus I prognosticate a mission requiring refilling before LEO departure will appear on the Starship manifest prior to the launch of the first HLS Starship. (Again personally, I'm hoping that demonstration mission is Mars-bound in 2024.)

Recall that we have seen several allusions to a demonstration of on-orbit fuel transfer to be done between two tanks on the same ship, so we might see that before we see actual ship-to-ship transfer.

How would we know if this happens if it isn't announced? Speculation has it that the first test would be between the headers and main tanks. There'd be no special hardware and the only external ops difference would be ullage settling. I doubt this would be apparent to a casual observer.

[DanClemmensen]

Some reflection and reconsideration of Starship refilling now that Artemis 1 is through TLI seems timely.

The success of Artemis 1 will likely mean NASA will continue to drink the Kool-Aid awhile longer regarding the Artemis 3 schedule, so they will now want SpaceX to demonstrate Starship refilling ASAP, since in NASA's mind refilling might be the long pole for an Artemis 3 landing. That pressure could increase the priority at SpaceX of refilling as compared to the priority of Starlink launches. (Personally I would like that outcome, even without fully believing the reasoning is correct.)

HLS Starship is going to require some special modifications that require work, even for an uncrewed landing demonstration mission. Thus I prognosticate a mission requiring refilling before LEO departure will appear on the Starship manifest prior to the launch of the first HLS Starship. (Again personally, I'm hoping that demonstration mission is Mars-bound in 2024.)

Recall that we have seen several allusions to a demonstration of on-orbit fuel transfer to be done between two tanks on the same ship, so we might see that before we see actual ship-to-ship transfer.

How would we know if this happens if it isn't announced? Speculation has it that the first test would be between the headers and main tanks. There'd be no special hardware and the only external ops difference would be ullage settling. I doubt this would be apparent to a casual observer.

If it's a formal milestone of the HLS contract, then we might not "see" it, but we will eventually see evidence of the progress payment. SpaceX or NASA might even say something about it.

[TheRadicalModerate]

By then there will have been dozens, perhaps hundreds, of refuelings.  Maybe the handwringing over RPODs will go away.  Just like the handwringing over load-and-go.

But that's why you want to keep crewed flights in mind when you're figuring this out.  It's not just RPOD hand-wringing.  There's radiation hand-wringing involved, too.

And even if the prox ops and docking become completely bulletproof, I'm not sure what happens with rendezvous.  AFAIK, eccentric orbit rendezvous, especially single-orbit eccentric rendezvous (again, radiation and crews--bad!) isn't exactly a thing right now.  I suspect that phasing is relatively straightforward, but you'll also have ω and RAAN adjustments that will be more interesting.  (Note:  ω--argument of perigee--isn't a thing for a circular orbit, but it is for an eccentric orbit.)

The big trade, which has to be made pretty early in the program, is whether sticking with 1200t mains for LSS and HEEO refueling is acceptable, or whether it's better to go to 1500t+ mains and circular VLEO refueling.  If they decide to go down the latter path, they don't need to worry about HEEO refueling for a long time, if ever.

[OTV Booster]

By then there will have been dozens, perhaps hundreds, of refuelings.  Maybe the handwringing over RPODs will go away.  Just like the handwringing over load-and-go.

But that's why you want to keep crewed flights in mind when you're figuring this out.  It's not just RPOD hand-wringing.  There's radiation hand-wringing involved, too.

And even if the prox ops and docking become completely bulletproof, I'm not sure what happens with rendezvous.  AFAIK, eccentric orbit rendezvous, especially single-orbit eccentric rendezvous (again, radiation and crews--bad!) isn't exactly a thing right now.  I suspect that phasing is relatively straightforward, but you'll also have ω and RAAN adjustments that will be more interesting.  (Note:  ω--argument of perigee--isn't a thing for a circular orbit, but it is for an eccentric orbit.)

The big trade, which has to be made pretty early in the program, is whether sticking with 1200t mains for LSS and HEEO refueling is acceptable, or whether it's better to go to 1500t+ mains and circular VLEO refueling.  If they decide to go down the latter path, they don't need to worry about HEEO refueling for a long time, if ever.

In one of Eric's articles on the Artemus launch he mentioned that angle changed throughout the launch window. I think they have the mechanics down pat but might need tight restrictions on the exact launch time for any particular HEEO. That said, big tanks and VLEO sounds like a better idea.


Remind me. With all the variations discussed, I've lost track. If they do the big tanks and VLEO, how far can they get? Refuel on lunar orbit before or after landing?  Which lunar orbit?

[TheRadicalModerate]

Remind me. With all the variations discussed, I've lost track. If they do the big tanks and VLEO, how far can they get? Refuel on lunar orbit before or after landing?  Which lunar orbit?

If I assume a 95t dry LSS, 15t crew module, 7t deployable payload, and Isp=368s (with at least two R2SL's running), a 1500t LSS can just do VLEO-NRHOviaBLT-100dayLoiter-crewXfer-LS-10dayMission-NRHO-crewXfer.

That also assumes 150kg/day boiloff in NRHO, and 500kg/day boiloff on the lunar surface near the poles, and 2% delta-v FPR for all crewed segments.

The mission actually doesn't close if you do fast transit to NRHO.

You can easily do a mission with 1200t tanks if you're willing to refuel both in LEO and something like an LEO+2000m/s HEEO.  But AFAICT, that's not the plan.