Starship On-orbit refueling - Options and Discussion

[Eka]

I'm not saying there must be a hab in the depot.  There could very well be a dedicated repair starship that goes around to the various depots.  That's fine and would accomplish all the testing goals.  But the dedicated repair starship is going look almost identical to the fuel depot and will also be a giant fuel tank to maneuver around where it needs to go.

My expectation is a dedicated repair Starship will have a habitat suitable for a bunch of Teslabots, tools, and a large parts hangar. Yes, I said large. You know, big enough to bring a dead satellite or few back in. Heck, from the outside it'll likely be near identical to a large satellite deployment Starship. Inside it will have places for the Teslabots to "dock". It'll also have tool chests able to be stocked with tools for the repairs expected. The Teslabots will be teleoperated from Earth only when needed. Simple predictable tasks will be automated. Note: Human advisors may tell the Teslabot the tasks to do.

Why put humans in harms way? Spacesuits are also big clumsy things to do work in.

[DanClemmensen]

I'm sure this idea has been mentioned before, but I am not good at searching The NSF forums:

Should Depot be designed for nose-to-tail connection to other Ships?

Depot would have a nose that looks a whole lot like the top of Booster. It would be covered with a disposable fairing for launch. For docking, Ship does the rendezvous and then orients itself pointing away from Depot at a distance of perhaps 50 m. Depot then maneuvers to dock to Ship's tail.

All of Depot's propellant transfer machinery is near the forward end. This includes an SQD that swings out and up to connect to Ship's QD plate. The advantage of this configuration is that all of the Depot-specific pieces are in what would otherwise be the payload area, above the tanks, and there are no docking modifications to Ship. The basic mating interface is "merely" the same surfaces as those between Booster and Ship. Depot may need some active latches, since Ship depends on gravity to stay connected to Booster. Depot may need a 6DOF conformal active latching ring system similar to the one IDSS uses (but 9 m instead of 1 m, what could possibly go wrong?) Depot is the active docking partner and will need fairly powerful 6DOF RCS thrusters.

Since Ship needs no additions for propellant transfer (no funny docking ports) the Ship design is unconstrained by this scheme.

Insane corollaries: Stacked Depots?!  Depot as a third stage (sort of)?!

If Depot nose connects to Ship tail as if it is a Booster, then a Depot can connect to another Depot which is already connected to a Ship. Since the connection uses the same structures on the tail of the ship that are used to connect ship to Booster, it should be reasonably straightforward to design the Depot nose to be able to apply reasonably large continuous thrust to Ship. Thus Depot can boost Depot+Ship until Depot depletes its propellant, Depot+Ship can maneuver as a single unit, just as Booster+Ship do during launch.

This also means that one Depot can refill another Depot without the need for an intermediate Tanker.

All of this is relevant to one of the proposed mission profiles for replacing SLS/Orion. The idea is to send a full Depot and an HLS from LEO to LLO, and then leave Depot  in LLO while HLS lands and then returns to LLO. Depot refills HLS and then Depot and HLS return to LEO. But Depot and HLS can remain attached to each other for the tow transits! And a second Depot can be used to top off the Depot+HLS in a HEEO prior to TLI if needed, so all tanker ops are in VLEO. Crazy.

[TheRadicalModerate]

Insane corollaries: Stacked Depots?!  Depot as a third stage (sort of)?!

If Depot nose connects to Ship tail as if it is a Booster, then a Depot can connect to another Depot which is already connected to a Ship. Since the connection uses the same structures on the tail of the ship that are used to connect ship to Booster, it should be reasonably straightforward to design the Depot nose to be able to apply reasonably large continuous thrust to Ship. Thus Depot can boost Depot+Ship until Depot depletes its propellant, Depot+Ship can maneuver as a single unit, just as Booster+Ship do during launch.

This also means that one Depot can refill another Depot without the need for an intermediate Tanker.

All of this is relevant to one of the proposed mission profiles for replacing SLS/Orion. The idea is to send a full Depot and an HLS from LEO to LLO, and then leave Depot  in LLO while HLS lands and then returns to LLO. Depot refills HLS and then Depot and HLS return to LEO. But Depot and HLS can remain attached to each other for the tow transits! And a second Depot can be used to top off the Depot+HLS in a HEEO prior to TLI if needed, so all tanker ops are in VLEO. Crazy.

Twark and I had a long discussion about something like a depot as a stage a couple of years back.  Search for "pusher/tanker".  The idea was that a full depot would dock nose-to-tail with the target Starship, transfer as much prop as possible into the target, then use the rest of its prop to start the orbital injection.  When it was mostly dry, it would stage off and return to its parking orbit somehow (propulsively or via aerobraking).  Then the target Starship would almost immediately continue the injection burn (no second orbit required to keep the burn relatively impulsive).

TL;DR conclusions:

1) It's a significant risk mitigation for crewed flights, because you never get stuck in an HEEO that could require multiple Van Allen transits, and you can completely forgo a lunar orbit refueling RPOD.

2) If the depot is so big that it can service multiple flights from an HEEO, it's more prop-efficient to let it do so.

3) It's obviously a huge change to the docking configuration and the architecture of the depot.  (BTW I disagree that it doesn't change the plumbing much; the depot will still likely have its plumbing nexus at the QD.  Moving all of that to the nose is a big fat hairy deal.)

[TheRadicalModerate]

Let's confirm whether 1mm/sec is enough to transfer the contents of fuel tanks from one Starship to another.

I think you erring in taking the 1mm/s^2 acceleration proposed for the settling of the propellant as meaning the velocity of the draining propellant out of the tanks occurs at 1mm/s.  Acceleration=/= velocity.

I was relating them by s=1/2at² and v=at, aka the standard kinematic equations.  Did I have a typo where I forgot to put the square sign on the seconds? I don't see it. (edit:  found it thx)

Even with the fix, you seem to be assuming that prop transfer occurs via the hydrostatic pressure generated by settling, which is never going to be true.  Once the prop is settled, it can be either pressure-fed or pumped.  That will happen much, much faster than it would if it were tricking out via the hydrostatics.  So your time horizons will be much shorter.

The trade is always going to be based on whether the prop available in a tanker by launching into a VVLEO is increased enough to be a net improvement over the prop needed to be expended by the depot to maintain the orbit when the tankers and targets aren't there.  That needs to be carefully calculated, probably more carefully than we can calculate from the peanut gallery.  My intuition is that you're better off with a garden-variety VLEO of 300-350km, rather than hanging ten on the edge of an accidental deorbit disaster if your ops go wrong.

[InterestedEngineer]

Let's confirm whether 1mm/sec is enough to transfer the contents of fuel tanks from one Starship to another.

I think you erring in taking the 1mm/s^2 acceleration proposed for the settling of the propellant as meaning the velocity of the draining propellant out of the tanks occurs at 1mm/s.  Acceleration=/= velocity.

I was relating them by s=1/2at² and v=at, aka the standard kinematic equations.  Did I have a typo where I forgot to put the square sign on the seconds? I don't see it. (edit:  found it thx)

Even with the fix, you seem to be assuming that prop transfer occurs via the hydrostatic pressure generated by settling, which is never going to be true.  Once the prop is settled, it can be either pressure-fed or pumped.  That will happen much, much faster than it would if it were tricking out via the hydrostatics.  So your time horizons will be much shorter.

The trade is always going to be based on whether the prop available in a tanker by launching into a VVLEO is increased enough to be a net improvement over the prop needed to be expended by the depot to maintain the orbit when the tankers and targets aren't there.  That needs to be carefully calculated, probably more carefully than we can calculate from the peanut gallery.  My intuition is that you're better off with a garden-variety VLEO of 300-350km, rather than hanging ten on the edge of an accidental deorbit disaster if your ops go wrong.

Oh, I've moved on with the idea of using atmospheric drag to move the prop, esp. after calculating a 100kg/sec flow rate on a 10cm diameter pipe with 1 bar of differential pressure.   Way more efficient.

Here's a snapshot of drag for two orientations, axial and heat shield first.  First, 200km is plenty of time for anything short of a months sitting on station.  Second, we can rotate the ship briefly to get 7 times the drag, settle the prop, rotate to axial, and transfer using pressure differential plus just enough air drag to keep the prop settled.

Then it remains to see whether it's worth it to aim for something lower than 200km.  It takes about 1 ton of extra prop per fuel load to boost the prop + tanker from 150km to 200km.  Worth saving 0.5% of prop or not?  Not sure.

My more general point is the air drag in the 150-220km region is sufficient to keep prop settled, so no need to continuously run thrusters during the prop transfer.

Getting tanker prop from 200km to the proposed of 300km costs 2.5t of prop, or 1.25%..  I don't see the ROI on that, a month in LEO is far more than enough.

https://docs.google.com/spreadsheets/d/1oWjuqb28lLvT9j8n_4NAENN6GJcp-JA8ARuaKPgYzF0/edit?usp=sharing

[TheRadicalModerate]

Here's a snapshot of drag for two orientations, axial and heat shield first.  First, 200km is plenty of time for anything short of a months sitting on station.  Second, we can rotate the ship briefly to get 7 times the drag, settle the prop, rotate to axial, and transfer using pressure differential plus just enough air drag to keep the prop settled.

Some model comments:

1) My atmo densities don't agree very well with yours.  I've been using an atmo model I found here.  Am I missing something?  Do you have a semi-definitive table (at least for average solar cases)? It's not beyond the realm of possibility that I have a bug.

2) For HLS (likely depot's first customer), you should assume a tanker cadence of 1/week.  If you assume about 15 tankers/mission, that's 3.5 months.

3) Until the depot starts getting extremely mission-dense, you also need to plan for it sitting empty for 6-12 months at a time.  You can't let it sit empty at 200km for that long.

4) I'm too stupid to know immediately whether decay time is linear with mass or not.  I kinda think not, which means you might need to model separate fill states of the depot separately, rather than using an average mass.

5) If you're going to use drag to settle, you need to account for RCS thrust to keep the depot orbiting tail-first into the flow.  I wouldn't be surprised if those added up to more than altitude maintenance.

6) If you settle the prop longitudinally, rotating to then settle it axially means it's not settled any more.  It'll slosh all over the place.  Furthermore, settling isn't just a matter of getting it all to one end; it also requires enough acceleration to damp out ripples before they become slosh.¹  I think that's how the 1E-04m/s² number came to be.

_______
¹(Climbs up on soapbox...)  One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

[Stan-1967]

_______
¹(Climbs up on soapbox...)  One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

What?  You haven't done any CFD work examining the surface stability of a slowly boiling cyogenic propellant acting under a variable/non linear hydrostatic pressure gradient, all while counteracted by surface tension effects within a stainless steel container accelerating independently of the coalesced cryogen?   Pffft!! what type of rocket forum is this? 

/s   

Me neither.  I do think the inlet in a pressure fed configuration will need complex baffling to control the liquid column.  Also trying to wrap my head around potential for pump cavitation ( in a pump  configuration).   I think overall a low pressure transfer will be more stable than a high pressure (pump) system at both inlet & outlet.
 
And also, the best part is no part

[InterestedEngineer]

Here's a snapshot of drag for two orientations, axial and heat shield first.  First, 200km is plenty of time for anything short of a months sitting on station.  Second, we can rotate the ship briefly to get 7 times the drag, settle the prop, rotate to axial, and transfer using pressure differential plus just enough air drag to keep the prop settled.

Some model comments:

1) My atmo densities don't agree very well with yours.  I've been using an atmo model I found here.  Am I missing something?  Do you have a semi-definitive table (at least for average solar cases)? It's not beyond the realm of possibility that I have a bug.

Thanks so much for your comments.  I'm using the model recommended above in the thread, which is off of stack exchange and interpolated first by me and verified by ChatGtp, and is for the thermosphere.  (https://space.stackexchange.com/questions/18223/where-can-i-find-data-for-atmospheric-density-vs-altitude).

The link you pointed out doesn't have the thermosphere model, and since its characteristics are quite different than the upper stratosphere, that's probably where the difference lies.

2) For HLS (likely depot's first customer), you should assume a tanker cadence of 1/week.  If you assume about 15 tankers/mission, that's 3.5 months.

That strongly implies they won't be using an aerobrake method right away, which is a problem for the aerobrake methodology.  Any methodology that can't be arrived at iteratively may not be done at all..   I think it's going to be 10 tankers btw, as they'll have Starship V3 running by the time they get around to filling an HLS.

3) Until the depot starts getting extremely mission-dense, you also need to plan for it sitting empty for 6-12 months at a time.  You can't let it sit empty at 200km for that long.

Are we sure they are going to use the depot method yet?

4) I'm too stupid to know immediately whether decay time is linear with mass or not.  I kinda think not, which means you might need to model separate fill states of the depot separately, rather than using an average mass.

You are correct, as air drag applies a force, and a=F/m, so I went for about the middle ground. Alas, three dimensional tables are hard to do in spreadsheets.  The general idea still works as while refueling, the mass increases, but the orbit decreases to denser air, so the acceleration tends to vary less than the change in mass would indicate.

5) If you're going to use drag to settle, you need to account for RCS thrust to keep the depot orbiting tail-first into the flow.  I wouldn't be surprised if those added up to more than altitude maintenance.

Stupid question, why isn't it nose first?  Are you thinking the refueling is nose-to-nose?  I'll have to break out some trig and analyze that, thanks for the feedback.

6) If you settle the prop longitudinally, rotating to then settle it axially means it's not settled any more.  It'll slosh all over the place.  Furthermore, settling isn't just a matter of getting it all to one end; it also requires enough acceleration to damp out ripples before they become slosh.¹  I think that's how the 1E-04m/s² number came to be.

I can play with a glass of water in the kitchen and get it to not slosh by doing very slow rotation.  Why wouldn't that work?

[Narnianknight]

I'm sure this idea has been mentioned before, but I am not good at searching The NSF forums:

Should Depot be designed for nose-to-tail connection to other Ships?

Here's a snapshot of drag for two orientations, axial and heat shield first.  First, 200km is plenty of time for anything short of a months sitting on station.  Second, we can rotate the ship briefly to get 7 times the drag, settle the prop, rotate to axial, and transfer using pressure differential plus just enough air drag to keep the prop settled.

Then it remains to see whether it's worth it to aim for something lower than 200km.  It takes about 1 ton of extra prop per fuel load to boost the prop + tanker from 150km to 200km.  Worth saving 0.5% of prop or not?  Not sure.

My more general point is the air drag in the 150-220km region is sufficient to keep prop settled, so no need to continuously run thrusters during the prop transfer.

¹(Climbs up on soapbox...)  One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

I know it's fun to speculate and be creative, but just a reminder of reality in case we've forgotten: Amit Kshatriya confirmed to Ars Technica that refueling will dock "belly to belly" (dorsal to dorsal), settle propellant with "rocket jets [that] provide 'settling thrust,'" and transfer propellant "using a pressure differential." Until/unless we see updated information, we can assume that this is the way it will be done. Of course, don't let me stop you from pretending to know better than SpaceX if you wish .

https://forum.nasaspaceflight.com/index.php?topic=50157.msg2589661#msg2589661

[Greg Hullender]

I know it's fun to speculate and be creative, but just a reminder of reality in case we've forgotten: Amit Kshatriya confirmed to Ars Technica that refueling will dock "belly to belly" (dorsal to dorsal), settle propellant with "rocket jets [that] provide 'settling thrust,'" and transfer propellant "using a pressure differential." Until/unless we see updated information, we can assume that this is the way it will be done. Of course, don't let me stop you from pretending to know better than SpaceX if you wish .

https://forum.nasaspaceflight.com/index.php?topic=50157.msg2589661#msg2589661

Since I posted that 7 months ago, we haven't heard much of anything from SpaceX or NASA on the topic. It's not a surprise that we fill that vacuum with speculation about other ways to do things.

Looking back at my post, I think the following might merit some discussion:

Once the two Starships come together, they will connect using the same ports SpaceX uses to load propellants on the launch pad. Then, SpaceX will fine-tune tank pressures and fire propellant settling thrusters. "At that point, they'll open up and let the propellant flow."

The propellants will flow from one vehicle to the other using a pressure differential, or "delta," between the donor tank and the recipient tank. This is a simpler solution than relying on pumps.

"The fundamental flow mechanism is because of the pressure delta across the umbilical," Kshatriya said.

If I remember right, the idea here to vent the ullage gas from the target vehicle into space before opening the connection and then counting on the pressure of the gas on the source side being much greater than the pressure from the weight of liquid propellant on the target side so it forces most/all of the propellant from the source to the target rather quickly. I wonder how much propellant gets wasted if you do this though.

One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

I think the plan was to leave the target tanks open to vacuum during fueling, so even if some pressurant goes from the source to the target vehicle, it'll just bubble to the top and vent. It won't increase the pressure on the target side. I'm just wondering how much propellant will be lost this way. (I agree this doesn't jibe with any notion of "fine tuning" the pressures though.)

[InterestedEngineer]

I think the plan was to leave the target tanks open to vacuum during fueling, so even if some pressurant goes from the source to the target vehicle, it'll just bubble to the top and vent. It won't increase the pressure on the target side. I'm just wondering how much propellant will be lost this way. (I agree this doesn't jibe with any notion of "fine tuning" the pressures though.)

That's a bit tricky, because "the top" is actually "the bottom", and bubbling in 0.1mm/s² is going to be veeerrrry slow.

[Greg Hullender]

I think the plan was to leave the target tanks open to vacuum during fueling, so even if some pressurant goes from the source to the target vehicle, it'll just bubble to the top and vent. It won't increase the pressure on the target side. I'm just wondering how much propellant will be lost this way. (I agree this doesn't jibe with any notion of "fine tuning" the pressures though.)

That's a bit tricky, because "the top" is actually "the bottom", and bubbling in 0.1mm/s² is going to be veeerrrry slow.

My impression was that (on the ground) the tanks are filled from the bottom and vent from the top. In orbit, the ullage burns establish the same orientation, just at much lower gravity. Is that not correct?

[Twark_Main]

Here's a snapshot of drag for two orientations, axial and heat shield first.  First, 200km is plenty of time for anything short of a months sitting on station.  Second, we can rotate the ship briefly to get 7 times the drag, settle the prop, rotate to axial, and transfer using pressure differential plus just enough air drag to keep the prop settled.

Some model comments:

1) My atmo densities don't agree very well with yours.  I've been using an atmo model I found here.  Am I missing something?  Do you have a semi-definitive table (at least for average solar cases)? It's not beyond the realm of possibility that I have a bug.

As mentioned, simple exponential models don't extrapolate past the thermosphere.

2) For HLS (likely depot's first customer), you should assume a tanker cadence of 1/week.  If you assume about 15 tankers/mission, that's 3.5 months.

3) Until the depot starts getting extremely mission-dense, you also need to plan for it sitting empty for 6-12 months at a time.  You can't let it sit empty at 200km for that long.

I guess it all depends on when in the "ramp up" you're assuming.

At 200 km you can counteract drag with a couple of Starlink thrusters.

4) I'm too stupid to know immediately whether decay time is linear with mass or not.  I kinda think not, which means you might need to model separate fill states of the depot separately, rather than using an average mass.

It is linear, yes. Everything is self-similar except for a constant multiplier.

5) If you're going to use drag to settle, you need to account for RCS thrust to keep the depot orbiting tail-first into the flow.  I wouldn't be surprised if those added up to more than altitude maintenance.

This is a good point, and it's part of why I advocate for CMGs on certain models of Starship. For ~1 tonne of additional dry mass you gain unlimited free attitude control.

6) If you settle the prop longitudinally, rotating to then settle it axially means it's not settled any more.  It'll slosh all over the place.

Bingo.

Furthermore, settling isn't just a matter of getting it all to one end; it also requires enough acceleration to damp out ripples before they become slosh.¹  I think that's how the 1E-04m/s² number came to be.

Not too worried about ripples, honestly.

_______
¹(Climbs up on soapbox...)  One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

This is a really good point that I haven't seen discussed elsewhere.

[InterestedEngineer]

I think the plan was to leave the target tanks open to vacuum during fueling, so even if some pressurant goes from the source to the target vehicle, it'll just bubble to the top and vent. It won't increase the pressure on the target side. I'm just wondering how much propellant will be lost this way. (I agree this doesn't jibe with any notion of "fine tuning" the pressures though.)

That's a bit tricky, because "the top" is actually "the bottom", and bubbling in 0.1mm/s² is going to be veeerrrry slow.

My impression was that (on the ground) the tanks are filled from the bottom and vent from the top. In orbit, the ullage burns establish the same orientation, just at much lower gravity. Is that not correct?

Attached very crude diagram, you can substitute venting for atmospheric drag, it's the same thing for this discussion

When prop is settled belly-to-belly there's a "gas" gap on the target side.  I show the prop flow going through that gas, but you could (with more parts) route the prop 180 degrees around the tank.

[Greg Hullender]

Attached very crude diagram, you can substitute venting for atmospheric drag, it's the same thing for this discussion

I know you've been talking about using drag, and I kind of like that idea too, but I don't think this is what SpaceX is planning to do. Do you? I think they're planning to accelerate longitudinally, so the propellant will settle at the bottom of the tanks, and I don't think they're planning to depend on drag at all.

Even using drag, I didn't think you were talking about having the vehicles crosswise to the wind during fueling. I thought your intention was to turn them end-on after some period of settling.

[Twark_Main]

off of stack exchange and interpolated first by me and verified by ChatGtp

This utterance makes me terrified for the future of intellectual discourse.   

Actual sources are here:

https://ccmc.gsfc.nasa.gov/models/JB2008~2008/

https://spacewx.com/jb2008/

simplified graph: https://space.stackexchange.com/questions/18223/where-can-i-find-data-for-atmospheric-density-vs-altitude

[Twark_Main]

trying to wrap my head around potential for pump cavitation ( in a pump  configuration).

Ingesting gas isn't "cavitation." Real cavitation is addressed by simply designing the pump correctly.

I think overall a low pressure transfer will be more stable than a high pressure (pump) system at both inlet & outlet.

You're conflating two unrelated variables.

And also, the best part is no part

The best part is no unnecessary part. The trick is all in figuring out which parts are necessary. 

There's another phrase:  "Penny wise, pound foolish."  Yes, very good, like the clown (where do you think it came from?), but after you're done with that, do think about the actual meaning of the phrase and its relevance here. 

[Narnianknight]

Once the two Starships come together, they will connect using the same ports SpaceX uses to load propellants on the launch pad. Then, SpaceX will fine-tune tank pressures and fire propellant settling thrusters. "At that point, they'll open up and let the propellant flow."

The propellants will flow from one vehicle to the other using a pressure differential, or "delta," between the donor tank and the recipient tank. This is a simpler solution than relying on pumps.

"The fundamental flow mechanism is because of the pressure delta across the umbilical," Kshatriya said.

If I remember right, the idea here to vent the ullage gas from the target vehicle into space before opening the connection and then counting on the pressure of the gas on the source side being much greater than the pressure from the weight of liquid propellant on the target side so it forces most/all of the propellant from the source to the target rather quickly. I wonder how much propellant gets wasted if you do this though.

One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

I think the plan was to leave the target tanks open to vacuum during fueling, so even if some pressurant goes from the source to the target vehicle, it'll just bubble to the top and vent. It won't increase the pressure on the target side. I'm just wondering how much propellant will be lost this way. (I agree this doesn't jibe with any notion of "fine tuning" the pressures though.)

Someone had the idea of pumping the destination ullage to the source ullage so as not to waste pressurant. It's not impossible; however, I conjecture that the performance loss of venting the ullage may be more desirable than the complexity of pumping it back through the QD port, assuming active pumping would be necessary for that. As far as propellant losses, some of it has to be vented for prop settling acceleration anyway, but it seems that would be significantly less than what is required for pressure management. Also, I assume both vehicles will have to vent to maintain attitude, and lowering the source ullage pressure isn't desirable. I don't recall if anyone here has forayed into the math of the required pressure differential and associated prop loss vs settling prop acceleration loss. Of course, all that doesn't even take into account prop boil-off between refuelings and the ways it helps or hinders operations.

[Twark_Main]

Once the two Starships come together, they will connect using the same ports SpaceX uses to load propellants on the launch pad. Then, SpaceX will fine-tune tank pressures and fire propellant settling thrusters. "At that point, they'll open up and let the propellant flow."

The propellants will flow from one vehicle to the other using a pressure differential, or "delta," between the donor tank and the recipient tank. This is a simpler solution than relying on pumps.

"The fundamental flow mechanism is because of the pressure delta across the umbilical," Kshatriya said.

If I remember right, the idea here to vent the ullage gas from the target vehicle into space before opening the connection and then counting on the pressure of the gas on the source side being much greater than the pressure from the weight of liquid propellant on the target side so it forces most/all of the propellant from the source to the target rather quickly. I wonder how much propellant gets wasted if you do this though.

One of the reasons I'm on Team Pump instead of Team Pressure Differential is because of the possibility of Slosh Farts.  If you uncover the higher-pressure inlet, ISTM you can drive a lot of pressurant to the other side pretty quickly, which can force you to turn good prop into pressurant, as well as introducing non-trivial operational delays.

On the other hand, with pumps (and low-power ones at that), uncovering the inlet might cause you to have to re-prime the pump (I'm assuming this is a problem that needed to be solved a long time ago), but you can keep the pressures equalized simply by venting the ullage spaces together.

I think the plan was to leave the target tanks open to vacuum during fueling, so even if some pressurant goes from the source to the target vehicle, it'll just bubble to the top and vent. It won't increase the pressure on the target side. I'm just wondering how much propellant will be lost this way. (I agree this doesn't jibe with any notion of "fine tuning" the pressures though.)

Someone had the idea of pumping the destination ullage to the source ullage so as not to waste pressurant. It's not impossible; however, I conjecture that the performance loss of venting the ullage may be more desirable than the complexity of pumping it back through the QD port, assuming active pumping would be necessary for that. As far as propellant losses, some of it has to be vented for prop settling acceleration anyway, but it seems that would be significantly less than what is required for pressure management. Also, I assume both vehicles will have to vent to maintain attitude, and lowering the source ullage pressure isn't desirable. I don't recall if anyone here has forayed into the math of the required pressure differential and associated prop loss vs settling prop acceleration loss. Of course, all that doesn't even take into account prop boil-off between refuelings and the ways it helps or hinders operations.

To square the circle, I imagine you pump the boil-off into CFRP pressure vessels and then use resistojets or methox to improve your I_sp. "The problem is the solution."

[InterestedEngineer]

Attached very crude diagram, you can substitute venting for atmospheric drag, it's the same thing for this discussion

I know you've been talking about using drag, and I kind of like that idea too, but I don't think this is what SpaceX is planning to do. Do you? I think they're planning to accelerate longitudinally, so the propellant will settle at the bottom of the tanks, and I don't think they're planning to depend on drag at all.

Even using drag, I didn't think you were talking about having the vehicles crosswise to the wind during fueling. I thought your intention was to turn them end-on after some period of settling.

Okay I changed the diagram to eliminate all thought or expression of drag.  No prograde direction either.  It's floating in free space as far as anyone can tell, just its own frame of reference.

Did the physics change in regards to propellant flow?  No.  But maybe conversation will be more comfortable for some.

So getting back the the topic at hand, how is belly to belly fuel transferred?  It's hard to describe in words. Make a picture!

If the acceleration is axial, then you've got to run a pipe to the bottom of the ship.  If the acceleration is 90 degrees to axial, then there's little or no pipe outside of what is in the docking connector.

How much pipe until "best part is no part" is violated?

The attached diagram shows axial flow with minimal piping.