Starship On-orbit refueling - Options and Discussion

[TheRadicalModerate]

Four things you can do with every single Starlink mission:

1) Launch to orbit.
2) Deploy Starlinks.
3) Attempt a refueling RPOD.
4) Attempt an EDL.

My guess is that, once the refueling connectors are finalized, they'll try to do all four, in this order, every single mission.  In many respects, getting refueling to work is more important than getting EDL recovery to work.  You can implement Option A without reusability.  But you can't do it without refueling.

I agree except I'd expect them to start trying things long before anything is finalized.  The equivalent of early F9 parachute recover and ocean landing tests.

For example you can test aspects of bulk handling of liquids by sloshing the ship or dumping fuel to space.  If something turns out to be harder than anticipated there could be a lot of experiments with different ullage thrust or tank baffles or whatever.  These would not need a rendezvous or refueling connectors.

The early connectors might not work or might not work well.  Version 2.0 will be in the pipeline.

They can do whatever they want to a Starship after it has deployed its payload, just as they could do whatever they wanted to F9 cores after they staged.  The scope of the experimentation is limited only by what's ready to be incorporated in a fashion that doesn't risk the payload.

The big decision here will always be how to prioritize refueling experiments vs. EDL experiments.  A refueling experiment that has a high chance of destructive failure precludes an EDL attempt.  But I'd guess that most refueling failures will be RPOD wave-offs, which still allow the EDL.  Still, there's some risk assessment involved.

[sebk]

It's worth remembering that you aren't connecting like-to-like. You are connecting the fuel & oxy discharge lines on the tanker to the fuel/oxy supply lines on the depot; the vent lines on the depot to the pressurisation lines on the tanker.

Why are there separate fill and drain lines?  Seems like you ought to be able to make do with one.  Cryo liquid drains from one tank and fills the other.  Which one is which is a matter of which tank is being pressurized by pumping gas into it.  Presumably, that gas is ullage that you're removing from the tank being filled.

And likewise, since lining up either dorsal-to-dorsal or belly-to-belly means you are connecting left-side to right-side & right-to-left, it should be relatively simple* to allocate male/female connectors in a consistent way that also works for GSE.

*(as these things go.)

If you really do have common fill-drain lines and a ambi-directional pump for each gas line, then the easiest geometry for the connector plate is simply all four lines along the (vertical, x) axis.  No handedness at all.
e top and one near the bottom?

In the same vein as no separate fill & drain lines, why pump gas at all? Pumping gas is less efficient than pumping liquid. But I wouldn't pump either. I'd just vent the ullage on the receiving side. As a side effect by lowering pressure you are lowering the boiling point so you could get superchilled propellant as a side effect (at a cost of boiling off some, of course).

[TheRadicalModerate]

In the same vein as no separate fill & drain lines, why pump gas at all? Pumping gas is less efficient than pumping liquid. But I wouldn't pump either. I'd just vent the ullage on the receiving side. As a side effect by lowering pressure you are lowering the boiling point so you could get superchilled propellant as a side effect (at a cost of boiling off some, of course).

Somebody's going to be pumping something, or you can't fill the target tank to a level higher than that of the source tank.  Pumping liquid is certainly a possibility, in which case you can vent the target tank.  However, if you do that, then you have to create new ullage gas for the source tank.  Note that the mass of the ullage gas is non-trivial; by venting, you're losing multiple tonnes of prop.

One of the things that worries me about pumping liquid is that your inlet pressures are low with only ullage thrust.  Cavitation isn't a big a problem when you're just transferring stuff, but you don't want to get into a position where slosh is uncovering the inlet.  My prejudice for making the system pressure-fed is that the slosh dynamics are a lot simpler.

[eriblo]

There is no "liquid level" that is relevant for propellant transfer, any thrust induced pressure differences are miniscule compared to the likely pressures in the tanks.

Remember that autogenous pressurization means that the tank pressure is self regulating to the vapor pressure except for short time scales:

Increase the temperature in the delivering tank and/or decrease it in the receiving one (through attitude changes end/or electrical heating), settle the propellant and open the valves.

The higher vapor pressure in the warmer tank will push liquid into the colder tank. As the pressure starts to lower in the warmer tank it will be maintained by the liquid evaporating and the pressure increase in the colder tank will be counteracted by gas condensing into the liquid. As long as original temperature difference is large enough to compensate for the lowering temperature in the warm tank (from evaporation and expansion) and increasing temperature in the colder tank (from condensation, compression and the additional warmer propellant) all of the liquid will be transferred. The ullage gas will in effect be transfered in the opposite direction through the liquid.

This situation is the opposite to launch: it is not desirable to develop temperature gradients between the liquid and gas (which can be reduced by spraying, bubbling or other agitation) and the occasional transfer of a gas bubble is unlikely to be an issue.

If a the temperature difference is insufficient by itself it can be supplemented by selective heating/venting of the gas.

[TheRadicalModerate]

There is no "liquid level" that is relevant for propellant transfer, any thrust induced pressure differences are miniscule compared to the likely pressures in the tanks.

I was just arguing that work of some sort had to be done.  Tanks don't spontaneously transfer liquid beyond the point where they're in hydrostatic equilibrium.

Remember that autogenous pressurization means that the tank pressure is self regulating to the vapor pressure except for short time scales:

Autogenous pressurization when the engines aren't running is considerably less efficient.  Could you do the transfer work by heating one side and venting the other?  Sure.  But why would you bother?  An electric gas pump is going to be scads more efficient than dumping a bunch of enthalpy of vaporization into one side and wasting ullage gas on the other.

The higher vapor pressure in the warmer tank will push liquid into the colder tank. As the pressure starts to lower in the warmer tank it will be maintained by the liquid evaporating and the pressure increase in the colder tank will be counteracted by gas condensing into the liquid. As long as original temperature difference is large enough to compensate for the lowering temperature in the warm tank (from evaporation and expansion) and increasing temperature in the colder tank (from condensation, compression and the additional warmer propellant) all of the liquid will be transferred. The ullage gas will in effect be transfered in the opposite direction through the liquid.

This sounds wildly optimistic to me, especially if you're counting on condensation collapse in the target tank to make room.  I don't think that happens without extracting heat from the target tank which is, again, a lot more complicated than a simple gas pump.

The thing that seems to be bothering you is the step of connecting the ullage spaces between the source and target tanks with a pump.  This just isn't rocket science.  It conserves all of the mass, doesn't require heaters or heat pumps, and uses ridiculously high-TRL parts.  Your pump power is dependent solely on the desired transfer rate.  It's easy.

[BT52]

Where energy for pump comes from? I would wildly guess Fuel itself is nice. Then just use raptors turbopumps. Hehe

Yeah anyway does anybody know how much energy would u need for full transfer? Mine hunch is that solar array with battery pack could still take lots of time.

[eriblo]

Remember that autogenous pressurization means that the tank pressure is self regulating to the vapor pressure except for short time scales:

Autogenous pressurization when the engines aren't running is considerably less efficient.  Could you do the transfer work by heating one side and venting the other?  Sure.  But why would you bother?  An electric gas pump is going to be scads more efficient than dumping a bunch of enthalpy of vaporization into one side and wasting ullage gas on the other.

In most cases (delivery tankers) you are going to start out with a tank mostly full of hot high pressure ullage and have on the order of 100 kW of solar heating available during much of the coast. The problem will be to to keep the propellants cold enough...

In other cases you have relatively long periods of time to achieve the desired temperature by adjusting your orientation.

BTW, what is the reason you are suggesting pumping the gas and letting the pressure difference push the liquid instead of vice versa? Cavitation is only an issue if you have low tank pressure to start with and then there is not much gas to bother about anyway...

The higher vapor pressure in the warmer tank will push liquid into the colder tank. As the pressure starts to lower in the warmer tank it will be maintained by the liquid evaporating and the pressure increase in the colder tank will be counteracted by gas condensing into the liquid. As long as original temperature difference is large enough to compensate for the lowering temperature in the warm tank (from evaporation and expansion) and increasing temperature in the colder tank (from condensation, compression and the additional warmer propellant) all of the liquid will be transferred. The ullage gas will in effect be transfered in the opposite direction through the liquid.

This sounds wildly optimistic to me, especially if you're counting on condensation collapse in the target tank to make room.  I don't think that happens without extracting heat from the target tank which is, again, a lot more complicated than a simple gas pump.

The thing that seems to be bothering you is the step of connecting the ullage spaces between the source and target tanks with a pump.  This just isn't rocket science.  It conserves all of the mass, doesn't require heaters or heat pumps, and uses ridiculously high-TRL parts. Your pump power is dependent solely on the desired transfer rate.  It's easy.

Well, I was suggesting that if this approach is workable the best case scenario is no extra hardware and even a small liquid pump, line and sprinkler completely contained in each tank sounds simpler than a blower, separate gas line and  ship to ship coupling. If you want to you could just take the same blower and gas line and direct it into the liquid instead of to the other ship...

There is no need to look at the details of the gas-liquid interactions for most transfers, the temperature management is about controlling the pressure. I.e. if you can maintain the depot at 1 bar vapor pressure and the tankers at 2 bar you can fill it close to 50% purely pneumatically. You then wait for the liquid and gas to equalize (likely much quicker then the time before the next tanker arrives anyway) so you can fill it to 75% and so on...

The exception is transferring a propellant load into a "hot" Starship, i.e. the propellants are at the max pressure boiling point so that the depot/tanker can not increase the pressure further. In this case you might have to vent up to a few tonnes from the target Starship (likely on the order of mass needed for settling).

There might need to be dedicated pumps but I would think that would be due to details like quicker response time or a lack of independent control of the temperatures in the LOX and LCH4 tanks.

[TheRadicalModerate]

Remember that autogenous pressurization means that the tank pressure is self regulating to the vapor pressure except for short time scales:

Autogenous pressurization when the engines aren't running is considerably less efficient.  Could you do the transfer work by heating one side and venting the other?  Sure.  But why would you bother?  An electric gas pump is going to be scads more efficient than dumping a bunch of enthalpy of vaporization into one side and wasting ullage gas on the other.

In most cases (delivery tankers) you are going to start out with a tank mostly full of hot high pressure ullage and have on the order of 100 kW of solar heating available during much of the coast. The problem will be to to keep the propellants cold enough...

In other cases you have relatively long periods of time to achieve the desired temperature by adjusting your orientation.

BTW, what is the reason you are suggesting pumping the gas and letting the pressure difference push the liquid instead of vice versa? Cavitation is only an issue if you have low tank pressure to start with and then there is not much gas to bother about anyway...

The higher vapor pressure in the warmer tank will push liquid into the colder tank. As the pressure starts to lower in the warmer tank it will be maintained by the liquid evaporating and the pressure increase in the colder tank will be counteracted by gas condensing into the liquid. As long as original temperature difference is large enough to compensate for the lowering temperature in the warm tank (from evaporation and expansion) and increasing temperature in the colder tank (from condensation, compression and the additional warmer propellant) all of the liquid will be transferred. The ullage gas will in effect be transfered in the opposite direction through the liquid.

This sounds wildly optimistic to me, especially if you're counting on condensation collapse in the target tank to make room.  I don't think that happens without extracting heat from the target tank which is, again, a lot more complicated than a simple gas pump.

The thing that seems to be bothering you is the step of connecting the ullage spaces between the source and target tanks with a pump.  This just isn't rocket science.  It conserves all of the mass, doesn't require heaters or heat pumps, and uses ridiculously high-TRL parts. Your pump power is dependent solely on the desired transfer rate.  It's easy.

Well, I was suggesting that if this approach is workable the best case scenario is no extra hardware and even a small liquid pump, line and sprinkler completely contained in each tank sounds simpler than a blower, separate gas line and  ship to ship coupling. If you want to you could just take the same blower and gas line and direct it into the liquid instead of to the other ship...

There is no need to look at the details of the gas-liquid interactions for most transfers, the temperature management is about controlling the pressure. I.e. if you can maintain the depot at 1 bar vapor pressure and the tankers at 2 bar you can fill it close to 50% purely pneumatically. You then wait for the liquid and gas to equalize (likely much quicker then the time before the next tanker arrives anyway) so you can fill it to 75% and so on...

The exception is transferring a propellant load into a "hot" Starship, i.e. the propellants are at the max pressure boiling point so that the depot/tanker can not increase the pressure further. In this case you might have to vent up to a few tonnes from the target Starship (likely on the order of mass needed for settling).

There might need to be dedicated pumps but I would think that would be due to details like quicker response time or a lack of independent control of the temperatures in the LOX and LCH4 tanks.

I'm not wedded to pumping gas instead of pumping liquid.  The important part is equalizing the gas pressures.

Let's figure out what we're talking about here.  At O:F=3.5 and a 1200t tanker, we have:

LOX:  933t, density=1142kg/m³, volume=817m³, dynamic viscosity=5.0E-5Pa-s
LCH4: 267t, density=422.6kg/m³, volume=631m³, dynamic viscosity=1.1E-5Pa-s

From the Poiseuille equation for required pressure drop across the two ends of a pipe:

Δp = 8μLQ/(πR⁴)

where:
μ: dynamic viscosity
L: length of the pipe (let's use 50m between the two in/outlets of the tanks)
Q: volumetric flow rate.  For a 2000s fill time, that's 0.41m³/s for LOX, 0.32m³/s for LCH4.
R: radius of the pipe.  10cm?

So:
LOX Δp = 26Pa
LCH4 Δp = 4.5Pa

Multiply both of these by the volumetric rate to get the mechanical power:

Plox = 10.7W
Plch4 = 1.4W

These omit the hydrostatic head, but that's about 150Pa if ullage acceleration is 0.01m/s² for the LOX, and less for the LCH4.  So that's 90W for the LOX pump and less than that for the LCH4 pump.  Maybe 200W total?  You can do the whole job with a 112Wh battery.

This little exercise has convinced me that it is indeed better to pump the liquid.  So all you need to do is equalize gas pressure between the source and target tanks and you're good to go.

BTW:  The hot lift tanker fresh into orbit is not the interesting case, because it only has to pump about 150t of prop.  The interesting case is the depot tanker, which may be pumping up to 1500t of prop into the target LSS or Starship, and it'll be pretty much in thermal equilibrium.

[rakaydos]

While not covered in the recent EDA interview/tour, I would suspect that the following would be near-optimum.

Tanks on starship are already set up for cold gas RCS using ullage propellants. Both tankers and mission craft would have one gender of connector, the tanker would have the other gender, so every transfer outside special mission craft would use a depot. By connecting a tanker to a depot, or a depot to a mission craft, and firing the thrusters on the receiving craft to accelerate toward the filling craft, it creates both a pressure drop in the receiving tank to suck dense liquid propellants in AND microacceleration to settle propellant for draining into feed lines.

All pumping is done using hardware already required for normal operation. "The best part is no part."

[OTV Booster]

Four things you can do with every single Starlink mission:

1) Launch to orbit.
2) Deploy Starlinks.
3) Attempt a refueling RPOD.
4) Attempt an EDL.

My guess is that, once the refueling connectors are finalized, they'll try to do all four, in this order, every single mission.  In many respects, getting refueling to work is more important than getting EDL recovery to work.  You can implement Option A without reusability.  But you can't do it without refueling.

Are you assuming that a tanker can refuel a starlink-dispensing cargo Starship directly? The HLS mission requires a tanker to transfer propellant to a depot and requires the depot to transfer propellant to the HLS. This means that the actual transfer hardware might only be present on the depot and not on the other types. If a tanker can transfer propellant to a Starlink-dispensing Starship, then the transfer hardware must either survive EDL or must remain in space or be expended.

I'm assuming that a depot isn't always necessary.  The transfer hardware needs to be able to survive EDL for the lift tanker if it's to be reused.  And I can't imagine the hardware not being androgynous.

Depots are necessary when lift tanker cadence is low, or if the gap between completion of prop accumulation on-orbit and the launch of the payload mission could be long.  Given that the LSS launch commit process is likely to be fairly involved, that gap could be long, and SpaceX will obviously be fairly low on the learning curve wrt tanker cadence.  But as everybody gets the ops situation wrung out, a naked tanker acting as an accumulator has lots of advantages, especially if it's sent to NRHO, since it can do a direct EDL to return.  A depot can't do that.

By naked tanker do you mean no heat shield? It's been discussed and works if it's worth it. If it's worth it is an open question in my mind. This use moves it a bit in that direction.


I've worked with many types of androgynous electrical and data connections but remember seeing only one androgynous fluid coupling. That's the glad hand connector on Westinghouse air brakes. Do they exist for cryo liquids? It'd be great if they exist but if they don't, it doesn't look like a trivial design problem.

[TheRadicalModerate]

By naked tanker do you mean no heat shield? It's been discussed and works if it's worth it. If it's worth it is an open question in my mind. This use moves it a bit in that direction.

No, that was a bad description.  I was referring to a non-depot tanker, with no insulation or active cryocooling.  Sorry for the confusion.

[OTV Booster]

Remember that autogenous pressurization means that the tank pressure is self regulating to the vapor pressure except for short time scales:

Autogenous pressurization when the engines aren't running is considerably less efficient.  Could you do the transfer work by heating one side and venting the other?  Sure.  But why would you bother?  An electric gas pump is going to be scads more efficient than dumping a bunch of enthalpy of vaporization into one side and wasting ullage gas on the other.

In most cases (delivery tankers) you are going to start out with a tank mostly full of hot high pressure ullage and have on the order of 100 kW of solar heating available during much of the coast. The problem will be to to keep the propellants cold enough...

In other cases you have relatively long periods of time to achieve the desired temperature by adjusting your orientation.

BTW, what is the reason you are suggesting pumping the gas and letting the pressure difference push the liquid instead of vice versa? Cavitation is only an issue if you have low tank pressure to start with and then there is not much gas to bother about anyway...

The higher vapor pressure in the warmer tank will push liquid into the colder tank. As the pressure starts to lower in the warmer tank it will be maintained by the liquid evaporating and the pressure increase in the colder tank will be counteracted by gas condensing into the liquid. As long as original temperature difference is large enough to compensate for the lowering temperature in the warm tank (from evaporation and expansion) and increasing temperature in the colder tank (from condensation, compression and the additional warmer propellant) all of the liquid will be transferred. The ullage gas will in effect be transfered in the opposite direction through the liquid.

This sounds wildly optimistic to me, especially if you're counting on condensation collapse in the target tank to make room.  I don't think that happens without extracting heat from the target tank which is, again, a lot more complicated than a simple gas pump.

The thing that seems to be bothering you is the step of connecting the ullage spaces between the source and target tanks with a pump.  This just isn't rocket science.  It conserves all of the mass, doesn't require heaters or heat pumps, and uses ridiculously high-TRL parts. Your pump power is dependent solely on the desired transfer rate.  It's easy.

Well, I was suggesting that if this approach is workable the best case scenario is no extra hardware and even a small liquid pump, line and sprinkler completely contained in each tank sounds simpler than a blower, separate gas line and  ship to ship coupling. If you want to you could just take the same blower and gas line and direct it into the liquid instead of to the other ship...

There is no need to look at the details of the gas-liquid interactions for most transfers, the temperature management is about controlling the pressure. I.e. if you can maintain the depot at 1 bar vapor pressure and the tankers at 2 bar you can fill it close to 50% purely pneumatically. You then wait for the liquid and gas to equalize (likely much quicker then the time before the next tanker arrives anyway) so you can fill it to 75% and so on...

The exception is transferring a propellant load into a "hot" Starship, i.e. the propellants are at the max pressure boiling point so that the depot/tanker can not increase the pressure further. In this case you might have to vent up to a few tonnes from the target Starship (likely on the order of mass needed for settling).

There might need to be dedicated pumps but I would think that would be due to details like quicker response time or a lack of independent control of the temperatures in the LOX and LCH4 tanks.

There is the seed of a nifty idea here. With the two ships (tanker and depot or depot and SS) along side each other they can orient to allow the supplying ship to shade the receiving ship. The 100kW thermal may not be enough to do the whole job in a reasonable time but it moves things in the right direction.


What we don't have, and I'm incapable of providing, is numbers. One that I can supply is the pressure differential. Six bar max. If one of you kind Sirs who are facile with numbers could noodle a BOE, it would be great. Maybe assume a clear connector aperture equivalent of 250, 500 and 750 sq cm and some reasonable pressure loss across the connector and plumbing. That should be enough to calculate a rough transfer time at six bar. Then we'd have some rough idea of how much assist, of any kind, is needed.


Hmmmm. If the depot needs a sun shield, could it be reoriented and work as a reflector to add heat to the delivering vessel? Probably not a trivial problem. If the transfer can be powered thermally it's more about procedures than gizmos. (except maybe moving the sun shield around)

[OTV Booster]

Remember that autogenous pressurization means that the tank pressure is self regulating to the vapor pressure except for short time scales:

Autogenous pressurization when the engines aren't running is considerably less efficient.  Could you do the transfer work by heating one side and venting the other?  Sure.  But why would you bother?  An electric gas pump is going to be scads more efficient than dumping a bunch of enthalpy of vaporization into one side and wasting ullage gas on the other.

In most cases (delivery tankers) you are going to start out with a tank mostly full of hot high pressure ullage and have on the order of 100 kW of solar heating available during much of the coast. The problem will be to to keep the propellants cold enough...

In other cases you have relatively long periods of time to achieve the desired temperature by adjusting your orientation.

BTW, what is the reason you are suggesting pumping the gas and letting the pressure difference push the liquid instead of vice versa? Cavitation is only an issue if you have low tank pressure to start with and then there is not much gas to bother about anyway...

The higher vapor pressure in the warmer tank will push liquid into the colder tank. As the pressure starts to lower in the warmer tank it will be maintained by the liquid evaporating and the pressure increase in the colder tank will be counteracted by gas condensing into the liquid. As long as original temperature difference is large enough to compensate for the lowering temperature in the warm tank (from evaporation and expansion) and increasing temperature in the colder tank (from condensation, compression and the additional warmer propellant) all of the liquid will be transferred. The ullage gas will in effect be transfered in the opposite direction through the liquid.

This sounds wildly optimistic to me, especially if you're counting on condensation collapse in the target tank to make room.  I don't think that happens without extracting heat from the target tank which is, again, a lot more complicated than a simple gas pump.

The thing that seems to be bothering you is the step of connecting the ullage spaces between the source and target tanks with a pump.  This just isn't rocket science.  It conserves all of the mass, doesn't require heaters or heat pumps, and uses ridiculously high-TRL parts. Your pump power is dependent solely on the desired transfer rate.  It's easy.

Well, I was suggesting that if this approach is workable the best case scenario is no extra hardware and even a small liquid pump, line and sprinkler completely contained in each tank sounds simpler than a blower, separate gas line and  ship to ship coupling. If you want to you could just take the same blower and gas line and direct it into the liquid instead of to the other ship...

There is no need to look at the details of the gas-liquid interactions for most transfers, the temperature management is about controlling the pressure. I.e. if you can maintain the depot at 1 bar vapor pressure and the tankers at 2 bar you can fill it close to 50% purely pneumatically. You then wait for the liquid and gas to equalize (likely much quicker then the time before the next tanker arrives anyway) so you can fill it to 75% and so on...

The exception is transferring a propellant load into a "hot" Starship, i.e. the propellants are at the max pressure boiling point so that the depot/tanker can not increase the pressure further. In this case you might have to vent up to a few tonnes from the target Starship (likely on the order of mass needed for settling).

There might need to be dedicated pumps but I would think that would be due to details like quicker response time or a lack of independent control of the temperatures in the LOX and LCH4 tanks.

I'm not wedded to pumping gas instead of pumping liquid.  The important part is equalizing the gas pressures.

Let's figure out what we're talking about here.  At O:F=3.5 and a 1200t tanker, we have:

LOX:  933t, density=1142kg/m³, volume=817m³, dynamic viscosity=5.0E-5Pa-s
LCH4: 267t, density=422.6kg/m³, volume=631m³, dynamic viscosity=1.1E-5Pa-s

From the Poiseuille equation for required pressure drop across the two ends of a pipe:

Δp = 8μLQ/(πR⁴)

where:
μ: dynamic viscosity
L: length of the pipe (let's use 50m between the two in/outlets of the tanks)
Q: volumetric flow rate.  For a 2000s fill time, that's 0.41m³/s for LOX, 0.32m³/s for LCH4.
R: radius of the pipe.  10cm?

So:
LOX Δp = 26Pa
LCH4 Δp = 4.5Pa

Multiply both of these by the volumetric rate to get the mechanical power:

Plox = 10.7W
Plch4 = 1.4W

These omit the hydrostatic head, but that's about 150Pa if ullage acceleration is 0.01m/s² for the LOX, and less for the LCH4.  So that's 90W for the LOX pump and less than that for the LCH4 pump.  Maybe 200W total?  You can do the whole job with a 112Wh battery.

This little exercise has convinced me that it is indeed better to pump the liquid.  So all you need to do is equalize gas pressure between the source and target tanks and you're good to go.

BTW:  The hot lift tanker fresh into orbit is not the interesting case, because it only has to pump about 150t of prop.  The interesting case is the depot tanker, which may be pumping up to 1500t of prop into the target LSS or Starship, and it'll be pretty much in thermal equilibrium.

Ah kind sir. One step ahead of me. Do your numbers say anything about transfer time?

[TheRadicalModerate]

While not covered in the recent EDA interview/tour, I would suspect that the following would be near-optimum.

Tanks on starship are already set up for cold gas RCS using ullage propellants. Both tankers and mission craft would have one gender of connector, the tanker would have the other gender, so every transfer outside special mission craft would use a depot. By connecting a tanker to a depot, or a depot to a mission craft, and firing the thrusters on the receiving craft to accelerate toward the filling craft, it creates both a pressure drop in the receiving tank to suck dense liquid propellants in AND microacceleration to settle propellant for draining into feed lines.

All pumping is done using hardware already required for normal operation. "The best part is no part."

IMO, it would be a mistake to intermediate everything through a depot.  Depots are either going to be stuck at the highest orbital energy they attain, or they're going to have to spend a huge amount propellant to propulsively return to LEO.  Lift tankers don't have that problem; they can either aerobrake or do direct EDLs for very little return prop.

Note that a lift tanker makes a perfectly good "accumulation tanker", just like a depot does.  The difference is how long it can loiter without a lot of boil-off.  But if you assume that an NRHO refueling mission is a fairly rapid turnaround affair, it probably makes sense for the sustainable LSS missions to use the following conops:

1) Depot is in LEO to start.
2) Launch one fewer lift tankers than are necessary to fill the depot.  (You need a full 1500t of prop, which is likely 10 LTs.)
3) Depot loiters until SLS/Orion launch commit.
4) Launch the last LT.  It receives all prop from the depot, then boosts for NRHO.  (LEO-NRHO for a 1500t LT takes about 1025t of prop.)
5) Upon arrival in NRHO, immediately transfer all available prop (about 450t) to the LSS.
6) After verification of the transfer, launch SLS/Orion or other cislunar transit mission to NRHO ASAP, to minimize boiloff.
7) LT in NRHO returns directly to EDL (takes about 26t of prop).

The trick is to minimize the time between when the LT takes the prop out of the depot and when the Orion transfers the crew to the LSS.

Just FYI, to get a depot back from NRHO to a propulsive LEO insertion (because there's no heat shield or elonerons) takes about 205t of prop.  If this is to be carried by the depot itself, you'll need to be able to carry about 2200t of prop from LEO, which increases your prop to LEO by almost 50%.

Making an androgynous, bi-directional refueling dock/connector doesn't seem like an insurmountable problem to me.  And it provides a lot more mission flexibility.

[TheRadicalModerate]

Ah kind sir. One step ahead of me. Do your numbers say anything about transfer time?

I used 2000 seconds (33 mins 20 secs).  That number was retrieved using the time-honored "rectal extraction" technique.

[TheRadicalModerate]

With the two ships (tanker and depot or depot and SS) along side each other they can orient to allow the supplying ship to shade the receiving ship.

Not in LEO you can't.  I'd have to work out the penumbra from the ship closer to the daylit side of the Earth, but you basically have a 2π steradian plane illuminating the system.  And there's, you know, the Sun on the other side.

[rakaydos]

While not covered in the recent EDA interview/tour, I would suspect that the following would be near-optimum.

Tanks on starship are already set up for cold gas RCS using ullage propellants. Both tankers and mission craft would have one gender of connector, the tanker would have the other gender, so every transfer outside special mission craft would use a depot. By connecting a tanker to a depot, or a depot to a mission craft, and firing the thrusters on the receiving craft to accelerate toward the filling craft, it creates both a pressure drop in the receiving tank to suck dense liquid propellants in AND microacceleration to settle propellant for draining into feed lines.

All pumping is done using hardware already required for normal operation. "The best part is no part."

IMO, it would be a mistake to intermediate everything through a depot.  Depots are either going to be stuck at the highest orbital energy they attain, or they're going to have to spend a huge amount propellant to propulsively return to LEO.  Lift tankers don't have that problem; they can either aerobrake or do direct EDLs for very little return prop.
(trimmed)

Why is this a problem? Are you assuming depots are going to be rare? Put one in every starlink plane to accumulate propellant on routine flights.

My assumtion is that building more depots is cheaper than designing a new connector that doesnt reduce overall vehical performance.

[TheRadicalModerate]

While not covered in the recent EDA interview/tour, I would suspect that the following would be near-optimum.

Tanks on starship are already set up for cold gas RCS using ullage propellants. Both tankers and mission craft would have one gender of connector, the tanker would have the other gender, so every transfer outside special mission craft would use a depot. By connecting a tanker to a depot, or a depot to a mission craft, and firing the thrusters on the receiving craft to accelerate toward the filling craft, it creates both a pressure drop in the receiving tank to suck dense liquid propellants in AND microacceleration to settle propellant for draining into feed lines.

All pumping is done using hardware already required for normal operation. "The best part is no part."

IMO, it would be a mistake to intermediate everything through a depot.  Depots are either going to be stuck at the highest orbital energy they attain, or they're going to have to spend a huge amount propellant to propulsively return to LEO.  Lift tankers don't have that problem; they can either aerobrake or do direct EDLs for very little return prop.
(trimmed)

Why is this a problem? Are you assuming depots are going to be rare? Put one in every starlink plane to accumulate propellant on routine flights.

My assumtion is that building more depots is cheaper than designing a new connector that doesnt reduce overall vehical performance.

I'm assuming that depots can't be moved cheaply, while lift tankers can.

This is not an argument against depots.  Depots are necessary when prop needs to be stored for a long time or when (more likely) it takes a long time to accumulate the amount of prop necessary for a mission.  Both of those things are likely to be true in LEO.  But they're silly if neither of those two missions requirements exists.  As a practical matter, I don't think that either of them will exist in NRHO or any other cislunar orbit, or if they do exist, they won't as operational confidence increases.  And lift tankers will always be the correct way to move prop from LEO to cislunar, if for no other reason than it's a one-way trip for depots, while the lift tankers can return to EDL very cheaply.

The other reason why depots are unlikely to be common is that they can't be sent to arbitrary, idiosyncratic orbits.  Consider, for example a deep space expendable version of Starship, which needs to generate massive C3 or haul an extremely large payload (possibly >500t, assembled in LEO) to a more modest C3.  That will require sending a tanker to a high-energy HEEO with very specific orbital parameters, allowing the deep space Starship to gain a lot of Oberth effect by doing a perigee burn with full tanks.

A depot is effectively expendable for this mission.  Reducing its energy enough for re-use, or moving its inclination and argument of perigee, will cost well more than its value.  On the other hand, a lift tanker can boost to the proper HEEO, fill the target Starship, and do an EDL maneuver for almost nothing.  The depot simply doesn't make sense.  But that requires that a target Starship be able to do a refueling RPOD directly with the lift tanker.

Why would a universal connector reduce vehicle performance?  No matter what, the lift tanker and payload versions of the connectors are going to have to be able to withstand EDL.  What secret sauce is there in a universal version that prevents that from happening?

[OTV Booster]

A glad hand is two parallel faces, each faced with a thin rubber plate. There are matching holes in the middle of each face and rubber to pass compressed air. The latching mechanism is around the outer edge most easily described as something kinda sorta like the lid interface of a pressure cooker.


Face the two connectors together and rotate ~60 deg to lock. The rubber faces scuff against each other while locking. Maybe this can be adapted to cryo, maybe not. I have doubts. Maybe a heater coil under the rubber/Teflon gasket to aid separation while cold. No other style androgynous fluid connector comes to mind.


Image courtesy of [size=78%]https://www.anythingtruck.com/product/810-035042.html[/size]

[OTV Booster]

The core differences between a depot and a tanker, in my mind, is the tanker has all the EDL hardware and the depot has passive and/or active cooling. The depot will need PV and maybe reaction wheels/CMB's to keep tight control on orientation for thermal and, now that I think of it, while moving propellant. Let it be the active member controlling fine attitude while transferring props.


This starts looking like something you think twice about when considering expendability. Cost is only one thing looked at. When production is settled in the depot will be a highly customized variant and will take up a largish hole in the production pipeline. That might or might not be an important issue. Gotta wait and see.