Starship On-orbit refueling - Options and Discussion

[OTV Booster]

The bellows are an complete nonissue: the 'problem' of how to have flexible cryogenic couplings on a rigid but moveable linkage without those bellow being load-bearing is not only a solved problem, but there is a solution currently sat on top of the arm halfway up the orbital launch tower, with said solution having already demonstrated the capability to move around, connect to a ship, and transfer cryogenic propellants.

The arm can handle a small amount of translation, from wind and thermal expansion.  But the amount of torsion on orbit in an emergency is much, much larger.  There's no way that the arm hardware could withstand that.
 

3) Dorsal-to-dorsal nose-to-tail, overlapped docking.  This could be androgynous, but it can also have full plumbing coming out of one payload bay and a stub that just stabilizes things in the other.  (I think I've given up on this one.)

4) Dorsal-to-dorsal, noses in the same direction, but overlapped docking, with the payload bay of one close to the tail QD of the other.  In this case, the single grapple and the plumbing are all that's taking the torsional loads.

OTV Booster seems to be on Team #2.  I think edzieba and I are mostly on Team #4 (although we disagree on whether the torsional loads are a problem).  I used to like #3, but I think #4 does everything it does and is simpler, moments of inertia permitting.

I'm in camp 3, because it means all your drain-propellant-in-orbit-for-transfer plumbing is at the apex of the tanks and out of the way during normal operations, rather than at the nadir where all the drain-propellant-for-propulsion plumbing also needs to be. It also minimises torque about the connection, and minimises the movement of the centre of mass of the system (think of the setup as a heavy glob of propellants that some rigid lightweight shells happen to move around the outside of). It also allows for a 'bumper' to be deployed at the 'unused' (or inactive) nose-to-tail meeting point to aid bracing, if needed.

When I went and actually drew this out, I got horribly confused, but Option #2 starts to look a lot better.  Take a look and see what you think.  Option #3 is really complicated, although it does indeed separate the transfer sources from the engine sumps, and it's the most mechanically stable.  Note that the pros and cons haven't been thought through completely, but are a nice basis for discussion (or the hurling of rotten fruit).

 ¯\_(ツ)_/¯

Nice rendering. I'm curious. Why is commingling the feed and the engine sumps a bad thing? This is how it loads in the ground.

[john smith 19]

When I went and actually drew this out, I got horribly confused, but Option #2 starts to look a lot better.  Take a look and see what you think.  Option #3 is really complicated, although it does indeed separate the transfer sources from the engine sumps, and it's the most mechanically stable.  Note that the pros and cons haven't been thought through completely, but are a nice basis for discussion (or the hurling of rotten fruit).

 ¯\_(ツ)_/¯

The options are complex.  A key question would be how much specific work needs to be done to get from a standard Starship to a tanker (or depot) variant?

SX likes to avoid special versions of stuff as long as possible (F9 and cargo Dragon were flying looong before FH and crew Dragon started flying). Also once SS starts making orbit they will want it to start being productive. That means starlink launches.

This suggests they would prefer options that either a)Can be installed (or removed) as a package or b) Can be permanently installed on all SS because the payload hit is minor or zero.

Both make which specific SS is doing what role an operational decision, rather than a build choice, but b) means that decision can be deferred to (more or less) the day of launch, giving greater flexibility.

How does that change the optics?

[TheRadicalModerate]

When I went and actually drew this out, I got horribly confused, but Option #2 starts to look a lot better.  Take a look and see what you think.  Option #3 is really complicated, although it does indeed separate the transfer sources from the engine sumps, and it's the most mechanically stable.  Note that the pros and cons haven't been thought through completely, but are a nice basis for discussion (or the hurling of rotten fruit).

 ¯\_(ツ)_/¯

The options are complex.  A key question would be how much specific work needs to be done to get from a standard Starship to a tanker (or depot) variant?

SX likes to avoid special versions of stuff as long as possible (F9 and cargo Dragon were flying looong before FH and crew Dragon started flying). Also once SS starts making orbit they will want it to start being productive. That means starlink launches.

This suggests they would prefer options that either a)Can be installed (or removed) as a package or b) Can be permanently installed on all SS because the payload hit is minor or zero.

Sounds right to me.  It also makes me lean more toward Option #2, where the existing plumbing is used, and there's an adapter to allow two female QDs to talk to a male-male adapter.

I think you could make something like the attached deploy out of the payload bay.  At an arm-wave, you have payload bay doors that open and slide out a stowed version of everything you need.  Then the stowed version needs to unfold itself as it extends back along the dorsal surface.  All mechanical engineering that I attempt always turns out to be a kludge, but I can draw a block diagram.

If you do this right, this is a complete "depot kit" that can turn any lift tanker into a depot or, more accurately, a tanker that's enabled to manage the transfer of prop in either direction.  The sole modification you need on vanilla Starships or lunar Starships to use this is a pair of passive posts for the grapples to grab.  (Note that "active" and "passive" refer to Starships with and without depot kits.  They don't denote the direction of fueling.  Depots have to be able to both receive and send prop.)

Ideally, a deployed depot kit could re-stow itself, allowing the tanker to which it was attached to return to EDL.  If that could happen, then there's no need for depots in cislunar orbits, which are likely more varied than the LEO orbit(s) needed for refueling.  A tanker with a stowed depot kit can refuel at the LEO depot, boost to whatever cislunar (or HEEO) orbit is required, deploy its depot kit, refuel whatever LSS or vanilla Starship needs it, stow the kit, and return directly to EDL, where it can be reused as an ordinary tanker.

If the kit can't re-stow, then you have permanent depots in cislunar that can't return.  In this case, you have a passive tanker (no depot kit) that refuels at the LEO depot, boosts to the cislunar orbit (or HEEO) in which a second depot lives, refuels it, and then returns to EDL.  Not quite as good or as flexible, but almost.

[OTV Booster]

The options are complex.  A key question would be how much specific work needs to be done to get from a standard Starship to a tanker (or depot) variant?

SX likes to avoid special versions of stuff as long as possible (F9 and cargo Dragon were flying looong before FH and crew Dragon started flying). Also once SS starts making orbit they will want it to start being productive. That means starlink launches.

This suggests they would prefer options that either a)Can be installed (or removed) as a package or b) Can be permanently installed on all SS because the payload hit is minor or zero.

Sounds right to me.  It also makes me lean more toward Option #2, where the existing plumbing is used, and there's an adapter to allow two female QDs to talk to a male-male adapter.

I think you could make something like the attached deploy out of the payload bay.  At an arm-wave, you have payload bay doors that open and slide out a stowed version of everything you need.  Then the stowed version needs to unfold itself as it extends back along the dorsal surface.  All mechanical engineering that I attempt always turns out to be a kludge, but I can draw a block diagram.

If you do this right, this is a complete "depot kit" that can turn any lift tanker into a depot or, more accurately, a tanker that's enabled to manage the transfer of prop in either direction.  The sole modification you need on vanilla Starships or lunar Starships to use this is a pair of passive posts for the grapples to grab.  (Note that "active" and "passive" refer to Starships with and without depot kits.  They don't denote the direction of fueling.  Depots have to be able to both receive and send prop.)

Ideally, a deployed depot kit could re-stow itself, allowing the tanker to which it was attached to return to EDL.  If that could happen, then there's no need for depots in cislunar orbits, which are likely more varied than the LEO orbit(s) needed for refueling.  A tanker with a stowed depot kit can refuel at the LEO depot, boost to whatever cislunar (or HEEO) orbit is required, deploy its depot kit, refuel whatever LSS or vanilla Starship needs it, stow the kit, and return directly to EDL, where it can be reused as an ordinary tanker.

If the kit can't re-stow, then you have permanent depots in cislunar that can't return.  In this case, you have a passive tanker (no depot kit) that refuels at the LEO depot, boosts to the cislunar orbit (or HEEO) in which a second depot lives, refuels it, and then returns to EDL.  Not quite as good or as flexible, but almost.

Yes! Using the existing plumbing in the way it's designed and oriented to work simplifies a lot.


The adapter you described is essentially the free flyer without free flying. A semi-rhetorical question: is it easier/lighter/cheaper to design/build a mechanism to move the adapter down the hull or to put a small cold gas RCS+translation system on the adapter? It's hard to judge with so many final details TBD, but the question will always be there.


On one hand a manipulating system might have welcome capabilities for other purposes. OTOH, an adapter that can detach and hang around for another campaign, or even move to another orbit, might offer good utility.


Stuffing all the floppy bits back into the cargo bay for return sounds iffy. For that matter, doing stretch tanks and shrinking the cargo bay looks too small for the adapter and everything else, although that really does need tight examination. Your earlier arguments for stretch tanks was on point.


If this rig needs another ride it's a strong argument for a multiple campaign lifetime, which implies free flying capabilities. Besides, free flyer sounds cool. Calling the rig an adapter misses a lot of functionality and is boring. Maybe something like Propellant Transfer and Service Device? Um. PTSD's already taken.

[Twark_Main]

A "gender bender" is an adaptor that's male on one side and female on the other, with plumbing crossed over as needed to make it work. (Otherwise it'll be the mirror image of what you want on one side or the other.) Are we in agreement here?

Nope. It'a trivial to modify the GSE connector to be mirror-symmetrical, with male and female connectors swapped across the plane of symmetry.

This trick works either way you dock (69 or db). It makes no difference: you don't need a "gender bender" component either way.

Obviously we don't see this on Starship today, but equally obvious is that Starship is currently a prototype and not the final version. If we're gonna baseline future changes (eg new parts), let's at least baseline smart changes.

[diagram has a part labeled "gender bender"]

Fortunately there's no need for this.

[TheRadicalModerate]

[diagram has a part labeled "gender bender"]

Fortunately there's no need for this.

But there's probably a need for something, if only to create a ship-to-ship separation that's manageable.  Also, whatever you wind up using still has to function as a genuine QD.  I imagine that an androgynous topology (which I agree is easy) is a little bit dicier to engineer for a clean, reliable separation at liftoff.  Not impossible, but if you need something anyway, why fool with best practice?

[wes_wilson]

[diagram has a part labeled "gender bender"]

Fortunately there's no need for this.

But there's probably a need for something, if only to create a ship-to-ship separation that's manageable.  Also, whatever you wind up using still has to function as a genuine QD.  I imagine that an androgynous topology (which I agree is easy) is a little bit dicier to engineer for a clean, reliable separation at liftoff.  Not impossible, but if you need something anyway, why fool with best practice?

This has probably been shot down somewhere in the prior 28 pages, but I couldn't find it skimming through.

If going the depot approach, would the Super Heavy booster not be a good fit to serve as the depot tanker?  Other threads have indicated it's ssto capable.  It and starship already have ports used for fueling starship while they're both on the OLM.  It and starship already have some clamping mechanisms that hold them together through stage separation. 

Put a sh in orbit, use it as the depot/accumulator.  Filled, it would hold nearly 3 starships worth of fuel/oxidizer. 

Lastly, although it has no heatshield, it has enough dv to reach orbit.  There, it can be refueled by starships and then it would have the same dv as when sitting on the ground.  Which is enough dv to be brought back without a heatshield for occasional and infrequent on-ground servicing.

Has there been an exploration of using SH as the depot?  Seems a closer starting point than using Starship.

[OTV Booster]

[diagram has a part labeled "gender bender"]

Fortunately there's no need for this.

But there's probably a need for something, if only to create a ship-to-ship separation that's manageable.  Also, whatever you wind up using still has to function as a genuine QD.  I imagine that an androgynous topology (which I agree is easy) is a little bit dicier to engineer for a clean, reliable separation at liftoff.  Not impossible, but if you need something anyway, why fool with best practice?

We are back full circle. Androgynous means face to face seals along the lines of a Westinghouse air brake glad hand but cryo rated. Image from Toledospring.com.


Unless there's a cryo variant already available it would be another materials R&D project. Not impossible but not trivial. 

[TheRadicalModerate]

If going the depot approach, would the Super Heavy booster not be a good fit to serve as the tanker?

It's going to be a while before they need to accumulate more prop than will fit in one Starship.  Hence, a Starship will make a perfectly acceptable depot.

When they finally do get to the point where they need more prop than a single Starship will hold, a SuperHeavy would be a pretty terrible waste of perfectly good engines.  Even if they need four or five Starship-sized depots for some massive mission, that's still fewer engines than a single SuperHeavy.

Last but hardly least, turning a SuperHeavy into a depot is a lot of work.  Getting Starship ready to be refueled is probably about 75% of the work you'd need to get a Starship-based depot up and running, but roughly 0% of the work you'd need to get an SH-based depot going.

PS:  Just noticed that you said SH as the tanker, not the depot.  That won't work.  If you get an SH into orbit, it's not coming down.  The reason it can do RTLS is because it's not going that fast at separation.  At reentry from orbital speed, it'll burn up.

[wes_wilson]

If going the depot approach, would the Super Heavy booster not be a good fit to serve as the tanker?

It's going to be a while before they need to accumulate more prop than will fit in one Starship.  Hence, a Starship will make a perfectly acceptable depot.

When they finally do get to the point where they need more prop than a single Starship will hold, a SuperHeavy would be a pretty terrible waste of perfectly good engines.  Even if they need four or five Starship-sized depots for some massive mission, that's still fewer engines than a single SuperHeavy.

Last but hardly least, turning a SuperHeavy into a depot is a lot of work.  Getting Starship ready to be refueled is probably about 75% of the work you'd need to get a Starship-based depot up and running, but roughly 0% of the work you'd need to get an SH-based depot going.

PS:  Just noticed that you said SH as the tanker, not the depot.  That won't work.  If you get an SH into orbit, it's not coming down.  The reason it can do RTLS is because it's not going that fast at separation.  At reentry from orbital speed, it'll burn up.

Thanks, I meant depot, not tanker. 

[Robotbeat]

Im pretty sure they’re using a stretched Starship as a depot. It’s possible that the depot will use the launch tanks, but it’s also possible they’ll use separate tanks which will have better thermal insulation.

[OTV Booster]

[diagram has a part labeled "gender bender"]

Fortunately there's no need for this.

But there's probably a need for something, if only to create a ship-to-ship separation that's manageable.  Also, whatever you wind up using still has to function as a genuine QD.  I imagine that an androgynous topology (which I agree is easy) is a little bit dicier to engineer for a clean, reliable separation at liftoff.  Not impossible, but if you need something anyway, why fool with best practice?

This has probably been shot down somewhere in the prior 28 pages, but I couldn't find it skimming through.

If going the depot approach, would the Super Heavy booster not be a good fit to serve as the depot tanker?  Other threads have indicated it's ssto capable.  It and starship already have ports used for fueling starship while they're both on the OLM.  It and starship already have some clamping mechanisms that hold them together through stage separation. 

Put a sh in orbit, use it as the depot/accumulator.  Filled, it would hold nearly 3 starships worth of fuel/oxidizer. 

Lastly, although it has no heatshield, it has enough dv to reach orbit.  There, it can be refueled by starships and then it would have the same dv as when sitting on the ground.  Which is enough dv to be brought back without a heatshield for occasional and infrequent on-ground servicing.

Has there been an exploration of using SH as the depot?  Seems a closer starting point than using Starship.

The big upside would be the volume. All the difficulties involved with bringing it back and extending the software for a use never intended is the start of the downsides. It's probably technically possible. Maybe worth playing with for mid to long term planning but not for immediate use.


As the design evolves and EDL shows success, non state of the art ships will become available for tanker and depot/accumulator R&D. The first known need for tankers and depot/accumulators is Artemus. They'll only need 'good enough' for that but I think the system will evolve fast from then on. By the time they start mars ops the SH might make sense. Probably use an early mark and let it burn up on reentry.


When they get some history under their belt they'll have a clearer sense of direction.

[Okie_Steve]

At what point does fineness become a problem with all the proposed stretching of 9m booster/ship/depot/flying tanks/etc?

[Twark_Main]

I imagine that an androgynous topology... is a little bit dicier to engineer for a clean, reliable separation at liftoff.

Why?

Androgynous means face to face seals along the lines of a Westinghouse air brake glad hand but cryo rated.

You're confusing androgynous connectors with an androgynous interface plate. You can make an androgynous interface plate (when mirrored across some plane) while still using entirely male-female connectors.

You simply put a female connector in the "mirror image" location of every male connector, and vice-versa.

[LMT]

Eta Space "Cryo-Dock" design, successor to 2024 LOXSAT-1.  (Sound off.)

"Eta Space announces Cryo-Dock™, a full scale commercial propellant depot capable of refueling multiple spacecraft with cryogenic propellants. Cryo-Dock™ design is already underway and is aimed to launch after the LOXSAT-1 mission funded by NASA. Eta Space is focused on cryogenic fluid management (CFM) technologies for the New Space Age and Clean Energy Economy."

[Greg Hullender]

From previous posts, I gather the following are key design constraints:

1) Minimize new plumbing.

2) Don't create anything you don't have to. Modifying existing hardware is better than all-new hardware. In particular, shrinking/stretching tanks is relatively easy.

3) There has to be an ullage burn. Anything else (e.g. rotation) would involve way too much entirely new hardware.

I'll add a couple more that seem obvious to me:

4) There can't be any big protrusions from the sides or front of the vehicle. (It has to be aerodynamic during launch.) Things might get extruded later, but, at launch, the rocket needs to be pretty smooth.

5) It has to be completely automatic; no human beings can be involved in the process except on the ground.

Not being an expert, it's hard to weigh these against each other, but applying them to the excellent drawings The Radical Moderate provided raises at least a few questions. I'd love to hear responses from people who really are experts.

This option, #2, seems far and away the most sensible except for the question of how the gender bender gets put in place. It can't be there at launch time or else it'd violate the no-protrusion rule, and if it's extruded after launch, I think that amounts to making a big change to the vehicle. (But maybe not--like I said; I'm no expert.) If it has to be moved into place, it either needs to fly there (which seems a lot of new hardware!) or else someone needs to put it there (which violates the all-automatic rule).

I don't see how this option (#3) can work at all. Not only does it have lots of new plumbing, I don't see how it can do the ullage burn without spinning like a pinwheel or (maybe) having the acceleration at a crazy angle. (Maybe I just lack imagination.)

This one (option #4) also seems to have way too much new plumbing. This is why I found myself (several posts back) thinking that it might work better to create a new device consisting of just propellant tanks plus a GSE plug and stick that in the cargo bay of a regular cargo starship. (The "source ship" in the picture.") Then you wouldn't have any new plumbing, and you could think of the depot as a really stripped down starship. (One with no engines and much smaller tanks.) So it makes no changes to the base starship at all, the only new construction is a modification of what already exists, and nothing protrudes until the cargo bay doors open. It won't carry as much propellant as a tanker-only starship, but maybe that's not a deal breaker.

You've still got the problem of holding the two vehicles together during fueling. I don't have any ideas for how to make that work, short of hoping that synchronized flying isn't infeasible, given the very low accelerations involved.

[TheRadicalModerate]

We are back full circle. Androgynous means face to face seals along the lines of a Westinghouse air brake glad hand but cryo rated.

I don't think that's a problem, because you're not going face-to-face.  Just use whatever male parts are on the tower-side QD for all the connections on the other side of the center line.  But you'll still need to deal with all the aerodynamics and deployment stuff, to say nothing of the actual disconnect dynamics.

The adapter you described is essentially the free flyer without free flying. A semi-rhetorical question: is it easier/lighter/cheaper to design/build a mechanism to move the adapter down the hull or to put a small cold gas RCS+translation system on the adapter? It's hard to judge with so many final details TBD, but the question will always be there.

If you're using cold gas, then you'll only be flying freely once, to get deployed.  If that's what it takes... meh.  Sure.  But not my first choice.

I think you can do this with extendable masts, if I measured correctly.  See the attached, highly arm-wavy cartoons.

PS:  I realized that I've conflated the terms "boom" and "mast", and also that "mast" implies something that sticks up.  All these booms/masts are are just extendable trusses to push the two hunks of the depot kit to the right spots.

There's also no reason why there have to be two masts.  You could put everything on the back and probably be fine, except it's a long way to extend something without tacking anything down as you go.  Also, putting the solar array where I did has the nice property of not interfering with the elonerons.  You might want to rotate them 90º if you adopt a nose-to-sun orientation to keep the tanks as cool as possible.

[Twark_Main]

This option, #2, seems far and away the most sensible except for the question of how the gender bender gets put in place. It can't be there at launch time or else it'd violate the no-protrusion rule, and if it's extruded after launch, I think that amounts to making a big change to the vehicle. (But maybe not--like I said; I'm no expert.) If it has to be moved into place, it either needs to fly there (which seems a lot of new hardware!) or else someone needs to put it there (which violates the all-automatic rule).

This, but

   • Delete the gender-bender and implement androgynous mirror-symmetrical interface plate

    • Tilt the interface plates at a slight angle, so the Starships have clearance between them

    • For docking the Starships both rotate and translate into position. This is done so the (off-axis) docking kick will exactly null out the rotation+translation. You "just" time-reverse the kinematics to find the exact docking approach maneuver.

[TheRadicalModerate]

This option, #2, seems far and away the most sensible except for the question of how the gender bender gets put in place. It can't be there at launch time or else it'd violate the no-protrusion rule, and if it's extruded after launch, I think that amounts to making a big change to the vehicle.

A couple of passive grappling posts is no big deal.  You can put them in shrouds and have them swing up and lock after you reach orbit.  I suspect that these are table stakes for all of the options.

I don't see how this option (#3) can work at all. Not only does it have lots of new plumbing, I don't see how it can do the ullage burn without spinning like a pinwheel or (maybe) having the acceleration at a crazy angle. (Maybe I just lack imagination.)

If you have both ships burning cooperatively, this isn't a huge deal.  But yeah, it has a lot of new plumbing.

You've still got the problem of holding the two vehicles together during fueling. I don't have any ideas for how to make that work, short of hoping that synchronized flying isn't infeasible, given the very low accelerations involved.

Grapple fixtures.  They're well-understood tech.  See here.  But this does require good enough proximity ops that a simple active arm can reach out and grab them.  But you need two, spaced far enough apart that they can handle worst-case torsion.

[TheRadicalModerate]

This, but

   • Delete the gender-bender and implement androgynous mirror-symmetrical interface plate

    • Tilt the interface plates at a slight angle, so the Starships have clearance between them

    • For docking the Starships both rotate and translate into position. This is done so the (off-axis) docking kick will exactly null out the rotation+translation. You "just" time-reverse the kinematics to find the exact docking approach maneuver.

I assume that when you say "tilt," you're talking about a rotation in the x-y plane, not the x-z plane?  In the x-z plane, the tails will collide.

If you only have one grappling point, or even two spaced too closely together, the torsion is going to be a problem.  Keeping the x-axes aligned lets you grab on in two widely spaced spots.

I don't worry too much about whether you need an adapter or whether your can make direct androgynous contact.  Unless a purely passively cooled depot works, you're going to need some cryocooling.  And unless I'm wrong about the torsion being a problem, you're going to need some active grappling at least a few meters away.  Once that stuff's there, you need to deploy something.  Making "something" include an adapter is no big deal, and provides a nice way for the cryocooling to get ahold of boiloff gases.

I imagine that an androgynous topology... is a little bit dicier to engineer for a clean, reliable separation at liftoff.

Why?

It's certainly doable.  I just don't think you need to sink the engineering time into something with so many moving parts to get it sufficiently reliable--especially given my objections just above.