Starship On-orbit refueling - Options and Discussion

[TheRadicalModerate]

Are there any other options?

The reverse of your #1: the depot launches with a depot-gender QD supplemented with a kit that makes it appear ship-gender to the GSE. The kit is jettisoned after launch.

Wouldn't it make more sense for the tower to have two QD versions instead of the ship?

I guess that should be #5:  GSE has a replaceable QD, which can be either male or female as needed.  As a practical matter, it's male about 95% of the time, when they're not launching depots or, if this turns out to be a thing, inter-depot transit vehicles (which could be either depots or tankers).  (Update:  I had a topological brain fart; transit vehicles don't work.)  I'll add that one.

The downside of this is that GSE fittings are probably pretty finicky, and replacing one for another, even infrequently, could be a decent-sized source of leaks and degraded reliability.

[vanoord]

Twin connectors side-by-side, one 'male', one 'female', the same type fitted to everything.

When two vehicles are brought together, the connectors are the opposite way round, so will go together as intended.

It's basically a modification of a railway Scharfenberg coupler - the image is a Dellner coupler that's a modified version of the Scharfenberg coupler. Obviously it doesn't transfer cryo fluids, but the geometry concept is the same.

[Twark_Main]

Androgynous is potentially the most elegant and maybe it'll eventually show up but cryo coupling is tough. Show me the coupling and it'll push me towards believing.

You don't need to invent an androgynous coupler to make an androgynous interface. You can always line up pairs of male & female connectors across a plane of symmetry.

Edit: sniped by Van Oord

Another advantage is that you get double redundancy for all fluid connections, so if you just can't get one of the two couplers to seal properly you can still transfer propellant (albeit at half the flow rate).

Two half-rated connectors are probably stilil somewhat heavier than having one full-rate connector, but for such a critical failure point it's nice to have a bit of redundancy.

[TheRadicalModerate]

I assume you need to pre-tension the QD to make it pop off cleanly at launch.  Is there something about the male side of the QD that makes that easier than the female side?  Maybe the better question:  is there something about an androgynous QD that would make it difficult to pop off cleanly?

[OTV Booster]

I assume you need to pre-tension the QD to make it pop off cleanly at launch.  Is there something about the male side of the QD that makes that easier than the female side?  Maybe the better question:  is there something about an androgynous QD that would make it difficult to pop off cleanly?

I've been picturing something along the line of dzus fasteners that cam in some force to separate the two sides. Most if not all the fluid connections will be valved immediately inboard of the connectors. At the end of transfer, valve everything off and do not vent the pressure between the mated plates. Mechanical linkages between dzuslike fasteners would both ensure simultaneous action and require only one actuator. The trapped pressurized fluids will enforce QD separation.

[TheRadicalModerate]

I assume you need to pre-tension the QD to make it pop off cleanly at launch.  Is there something about the male side of the QD that makes that easier than the female side?  Maybe the better question:  is there something about an androgynous QD that would make it difficult to pop off cleanly?

I've been picturing something along the line of dzus fasteners that cam in some force to separate the two sides. Most if not all the fluid connections will be valved immediately inboard of the connectors. At the end of transfer, valve everything off and do not vent the pressure between the mated plates. Mechanical linkages between dzuslike fasteners would both ensure simultaneous action and require only one actuator. The trapped pressurized fluids will enforce QD separation.

Sounds great.  So why didn't they build the QDs that way in the first place?  Was the guy who mentioned that androgynous QDs would be really handy when it came to refueling shouted down by "Minimum Viable Product!" or was there something trickier involved?

This is the kind of thing you want to run through about 10,000 cycles, so you work out all the 5σ itty-bitty oscillations that cause something to hang up and pull the rocket over onto your tower.  It could be as simple as the guys in the back room have only gotten to cycle 8000 so far.  Or not.

[Twark_Main]

So why didn't they build the QDs that way in the first place?

Why didn't they build Raptor 3 in the first place? 

We say "iterative design" a lot, but I think the reality is a lot harder for people to wrap their head around.

[TheRadicalModerate]

So why didn't they build the QDs that way in the first place?

Why didn't they build Raptor 3 in the first place? 

We say "iterative design" a lot, but I think the reality is a lot harder for people to wrap their head around.

Iterative design or not, easy things get done right the first time.  So it's not unreasonable to conclude that an androgynous QD is not easy, for some reason.  It's not like nobody thought about refueling when they designed the interface.  Somebody decided that pushing the problem off was the right thing to do.

So, given that it's at least less easy, does this impact how you choose to tackle refueling, which is almost certainly on the critical path between now and Option A?

Seems to me that sdsds's jettisonable female-to-female QD, mounted at launch on the depot's male (gender-reversed) QD, is the easiest thing to do.  But switching over to androgynous is a lot more elegant--and provides a much more flexible set of refueling logistics, since tankers can then ferry prop to lunar orbit, transfer it nearly immediately, and return to EDL.

[wannamoonbase]

So why didn't they build the QDs that way in the first place?

Why didn't they build Raptor 3 in the first place? 

So why didn't they build the QDs that way in the first place?

Why didn't they build Raptor 3 in the first place? 

We say "iterative design" a lot, but I think the reality is a lot harder for people to wrap their head around.

]

It seems many don't appreciate how much longer an iteration is when you're working with things as long as Starship.

If this was some tiny 6 or 7 meter vehicle with only 16 Raptors they could be on version 3 or 4 of the pad hardware by now.

[Greg Hullender]

If you're going to have two depots anyway, you could just make one of them male and one female. And if you have two launch pads, make one of them male and the other female. Then make half of your tankers male and the other half female. Use these tools to fill both depots.

Make the final tanker male (say) and the mission starship female. Fill each from the appropriate depot. Problem solved.

Probably not the idea solution, but it shows that at least there is one.

[TheRadicalModerate]

If you're going to have two depots anyway, you could just make one of them male and one female. And if you have two launch pads, make one of them male and the other female. Then make half of your tankers male and the other half female. Use these tools to fill both depots.

Make the final tanker male (say) and the mission starship female. Fill each from the appropriate depot. Problem solved.

Probably not the idea solution, but it shows that at least there is one.

FWIW, I did a quick model comparing two different conops:

1) Sending a depot to NRHO via BLT (takes a couple of months), using it to refuel an LSS there (SLT reuse of single-LSS missions staged from LEO are both candidates for this), and then return propulsively to LEO.

2) Transferring prop from a full depot to a tanker, going to NRHO via fast transfer (takes 5 days), and transferring the prop to a waiting depot.  Then the tanker returns straight to EDL.¹

The tanker-mediated version is slightly less expensive in terms of prop, but it also has the nice property of being a 10-day mission, instead of a 4-month mission.

All this is a long way of saying that I don't think it's a huge deal to have male depots with everything else--including transit tankers--being female.  It seems a lot easier operationally than segregating the fleet.  Androgynous is probably the optimal solution here, but we don't know why they didn't go that way to begin with.

________
¹I should note that there's a third option, which is to aerobrake the depot back to LEO, if that's possible.  It takes a couple of months if you're killing only 25m/s per pass, which should be fine from a thermal standpoint.  Whether it's fine from a coatings and solar cell standpoint is an open question.

[Greg Hullender]

2) Transferring prop from a full depot to a tanker, going to NRHO via fast transfer (takes 5 days), and transferring the prop to a waiting depot.  Then the tanker returns straight to EDL.¹

If you assume depots (and only depots--other than GSE) are male, then it does simplify some things; a tanker only transfers fuel to/from a depot. A Starship does the same. If you want one tanker to send fuel to another, you need a depot in-between.

That works well for your lunar example, where you just put an extra depot in NRHO, but not so much for other destinations, where you want a "final tanker" and a Starship to fill up from separate depots and then launch into the same HEEO, where the tanker tops off the Starship. This is a really important use case, I think, for destinations beyond the moon, since it not only gets you extra ∆-v, it lets the Starship do a cheap plane change.

I suppose you could try to design a tanker to do a sex-change, but I suspect Elon wouldn't want to fund that. :-)

[TheRadicalModerate]

That works well for your lunar example, where you just put an extra depot in NRHO, but not so much for other destinations, where you want a "final tanker" and a Starship to fill up from separate depots and then launch into the same HEEO, where the tanker tops off the Starship. This is a really important use case, I think, for destinations beyond the moon, since it not only gets you extra ∆-v, it lets the Starship do a cheap plane change.

The interplanetary cases, other than Mars, are pretty rare.  My Mars transfer model is admittedly limited, but everything I've played with shows that it's really hard to get some breakthrough transit capability by throwing lots of departure C3 at the problem.  (It mostly just requires a lot of delta-v to scrub off the arrival C3.)  Given that, VLEO is perfectly fine for filling up martian Starships.

I suspect all the other interplanetary missions are one-offs.  If you need to send a 100t spacecraft to Uranus, you might need a lunar-distance HEEO, but they've got really wide windows, and moving the depot back from the FTO to VLEO is getting close to a rounding error on a mission like that.

The real gnarly HEEO cases are the lunar ones.  I'll be very interested to see what they do for Option A.  v2 LSSes need HEEO to do the whole TLI-BLT-NRHO-loiter-RPOD-LS-NRHO-RPOD-idle set of maneuvers on one tank of prop, but the HEEO can have an apogee that's below the bad parts of Van Allen #1.  (My boiloff guesses are just that, but if the depot can limit boiloff to 200kg/day in LEO, something like 1000km x 200km works fine.)

On the other hand, v3 LSSes can easily fuel up in VLEO and make it just fine.  SpaceX may need to freeze the LSS design early to make NASA happy, in which case there will be a handful of v2 lunar missions.  But if the schedule slips much more, they'll be centering the design around v3 propulsion sections, and HEEO will be unnecessary.

[Greg Hullender]

That would be nice. Have there been any reasonably solid numbers for V3 though?

1) Sending a depot to NRHO via BLT (takes a couple of months), using it to refuel an LSS there (SLT reuse of single-LSS missions staged from LEO are both candidates for this), and then return propulsively to LEO.

So what is "SLT?"

[TheRadicalModerate]

That would be nice. Have there been any reasonably solid numbers for V3 though?

1) Sending a depot to NRHO via BLT (takes a couple of months), using it to refuel an LSS there (SLT reuse of single-LSS missions staged from LEO are both candidates for this), and then return propulsively to LEO.

So what is "SLT?"

Sustainable Lunar Transportation.  It's what happens after Option B for SpaceX, and after Sustainable Lunar Development (SLD) for Blue.  Starts with Artemis VI in the plan of record.

[Nescio Erucis]

Twin connectors side-by-side, one 'male', one 'female', the same type fitted to everything.

When two vehicles are brought together, the connectors are the opposite way round, so will go together as intended.

It's basically a modification of a railway Scharfenberg coupler - the image is a Dellner coupler that's a modified version of the Scharfenberg coupler. Obviously it doesn't transfer cryo fluids, but the geometry concept is the same.

To keep with the sexual metaphor, this type of connection would be "hermaphroditic" rather than "androgynous".
Although I think that rather than the former term, another would be preferable (like "twin connectors" as you mentioned).
Androgynous implies a single, common interface having both male and female properties.

[DanClemmensen]

Twin connectors side-by-side, one 'male', one 'female', the same type fitted to everything.

When two vehicles are brought together, the connectors are the opposite way round, so will go together as intended.

It's basically a modification of a railway Scharfenberg coupler - the image is a Dellner coupler that's a modified version of the Scharfenberg coupler. Obviously it doesn't transfer cryo fluids, but the geometry concept is the same.

To keep with the sexual metaphor, this type of connection would be "hermaphroditic" rather than "androgynous".
Although I think that rather than the former term, another would be preferable (like "twin connectors" as you mentioned).
Androgynous implies a single, common interface having both male and female properties.

Like you , I would prefer to avoid the terms completely. However, If we must use them, two identical connection systems that connect to each other can be called "androgynous" even when the individual subconnectors are not androgynous. This is the terminology used e.g. by the  IDSS standard.
    https://www.internationaldockingstandard.com/download/IDSS_IDD_Revision_F.pdf
I would prefer to use "hermaphroditic" to refer to a Depot with two entirely separate connection systems, one exactly like all other Ships and a second one like the SQD arm.

In particular, If a depot has a 9-meter docking interface as its forward end that looks like the top of a Booster, it can dock to a Ship's engine end. If that Depot has an articulated SQD stored below its docking ring, it can swing it out to connect to the Ship's QD. This Depot gets fueled for launch just like any other Ship, and it can be refilled from another Depot if needed. It refills from tankers using its swing-out SQD, and it refills other ships using its swing-out SQD. It launches using an expendable nosecone.

[Greg Hullender]

In particular, If a depot has a 9-meter docking interface as its forward end that looks like the top of a Booster, it can dock to a Ship's engine end. If that Depot has an articulated SQD stored below its docking ring, it can swing it out to connect to the Ship's QD. This Depot gets fueled for launch just like any other Ship, and it can be refilled from another Depot if needed. It refills from tankers using its swing-out SQD, and it refills other ships using its swing-out SQD. It launches using an expendable nosecone.

But this introduces a lot of new cryogenic plumbing, does it not? Anything involving refueling from the nose instead of the base is going to be a big deal. And cryogenic plumbing with moving parts sounds like something that needs human techs available on site to do repairs.

[DanClemmensen]

In particular, If a depot has a 9-meter docking interface as its forward end that looks like the top of a Booster, it can dock to a Ship's engine end. If that Depot has an articulated SQD stored below its docking ring, it can swing it out to connect to the Ship's QD. This Depot gets fueled for launch just like any other Ship, and it can be refilled from another Depot if needed. It refills from tankers using its swing-out SQD, and it refills other ships using its swing-out SQD. It launches using an expendable nosecone.

But this introduces a lot of new cryogenic plumbing, does it not? Anything involving refueling from the nose instead of the base is going to be a big deal. And cryogenic plumbing with moving parts sounds like something that needs human techs available on site to do repairs.

You may be correct about the complexity. I think the bulk of the repairs we see at Starbase are caused by the launch and not by the moving parts per se. I do not really understand how Ship and Depot can mate without moving plumbing parts. I suspect there are ways to minimize them, but not eliminate them. Note that Booster and Ship already have moving cryogenic plumbing for the TVC engines.

Here is an off-the-wall concept (I cannot use the term "design" here) that minimizes the flex in the plumbing. Put the SQD on a long arm that is mounted on the dorsal side of the Depot. It runs nearly the length of the Depot and is hinged at the bottom. It's range of motion is a few degrees, like the TVCs. The top of the arm is above the  9-meter docking ring and carries the QD plate. The arm is about 45 meters long. The entire arm is launched under a disposable fairing and the top is under the disposable nosecone, which is extended dorsally to cover it.

[Twark_Main]

I do not really understand how Ship and Depot can mate without moving plumbing parts. I suspect there are ways to minimize them, but not eliminate them.

We're in microgravity. The spacecraft is the moving part.

We already have hardware to perform docking and then active 6-axis relative control. It's called NDS. The linear actuators make up a Stewart platform, so you can just "muscle" the two ships into the correct relative orientation and then actively pull them together in a controlled way.

Here is an off-the-tall concept (I cannot use the term "design" here) that minimizes the flex in the plumbing. Put the SQD on a long arm that is mounted on the dorsal side of the Depot. It runs nearly the length of the Depot and is hinged at the bottom. It's range of motion is a few degrees, like the TVCs. The top of the arm is above the  9-meter docking ring and carries the QD plate. The arm is about 45 meters long. The entire arm is launched under a disposable fairing and the top is under the disposable nosecone, which is extended dorsally to cover it.

The usual way to deal with rigid pipe flex is to use a series of expansion loops, as we've seen on Starship before.

https://whatispiping.com/expansion-loop-on-piping-system/