4 ms·

I haven't seen any plausible indication of how they will manage what is called "ullage". Putting propellers on electric motors inside the tanks, to make fluid

I haven't seen any plausible indication of how they will manage what is called "ullage". Putting propellers on electric motors inside the tanks, to make fluid circulate axially around the center line, would propel it to intakes on the circumference, with a growing bubble running down the center. Then they would not need to accelerate the vehicle throughout the process, which alternative seems wholly impractical. It would require clever design of the anti-slosh baffles (never shown in cutaway diagrams) to be compatible with such circulation. It might be tricky to keep the whole tank from switching to tumbling end-over-end.

chris_va 3y ago
They brought me in here to do a job, they asked me to stir the damned tanks, and I stirred the tanks!
- bandyaboot 3y ago
  Oh yeah smart guy? What were the gauges reading before you flipped the switch?
Robotbeat 3y ago
It's as simple as using thrusters to keep the propellant settled. milli-gee thrust (or even lower) has been shown to be sufficient for settling cryo propellant in upper stages during long coasts. SpaceX has been experimenting/analyzing settling of cryo propellant (LOx) during hours-long coasts for years, now. It's a requirement for doing direct-GSO insertion for an upper stage (or even doing any relights or deorbit burns, for that matter), which they've already done. "But wait, it's supposed to be more complicated than that!" No, really, they've been doing this stuff for years, and settling propellant for upper stage coasts really is virtually the same thing as settling for propellant transfer (transferring to the engine(s) for relight vs transferring to another stage... in fact you have to be MORE careful to avoid bubbles when feeding propellant to an engine than you do when transferring propellant to another tank). https://www.youtube.com/watch?v=mVAGoWJuDKk https://www.youtube.com/watch?v=mVAGoWJuDKk They still need to do the whole launch-two-starships-then-dock-them bit, plus quick connect fittings, but they already have years of experience with autonomous rendezvous and docking from operating Dragon at ISS (Dragon 2 has that capability, and it has been used for the uncrewed Dragon 2 demo mission plus the recent cargo Dragon 2 missions, plus I think it's the nominal choice even for crewed missions).
- angiosperm 3y ago
  The difference is that in preparation to fire a rocket, you only need to apply ullage acceleration for a short time because after the rocket lights, that provides the acceleration. When just transferring fluid, you need to accelerate during the entire period of the transfer, because pumping fuel to another tank provides no acceleration of its own. So, you need the acceleration to be small enough not to build up to an inconvenient delta-V over the time it takes to move the fluid. In principle you could apply acceleration for a while in one direction; and then switch to accelerating in the other direction, wait for the fluid to settle at the other end, and pump from that end, alternating back and forth. But it still consumes reaction mass, and introduces delays while switching. Docking vehicles massing hundreds of tons is a much bigger problem than docking Dragon. The Dragon experience might not apply very directly. Presumably the engineers actually involved all understand this (literally rocket science) much better than we do.
  - Robotbeat 3y ago
    You can do the settling thrust near perigee, so the delta-v from settling doesn’t go to waste but increases the apogee of the depot and increases the orbital energy of the propellant. Docking dragon is obviously similar to docking starships (Dragon berthing sensors were tested on Shuttle, broadly similar to an empty Starship). You’re right, tho, that the actual SpaceX engineers (who proposed this and developed Dragon) understand this better than random online folk.

chris_va 3y ago

They brought me in here to do a job, they asked me to stir the damned tanks, and I stirred the tanks!

bandyaboot 3y ago
Oh yeah smart guy? What were the gauges reading before you flipped the switch?

Robotbeat 3y ago

It's as simple as using thrusters to keep the propellant settled. milli-gee thrust (or even lower) has been shown to be sufficient for settling cryo propellant in upper stages during long coasts. SpaceX has been experimenting/analyzing settling of cryo propellant (LOx) during hours-long coasts for years, now. It's a requirement for doing direct-GSO insertion for an upper stage (or even doing any relights or deorbit burns, for that matter), which they've already done. "But wait, it's supposed to be more complicated than that!" No, really, they've been doing this stuff for years, and settling propellant for upper stage coasts really is virtually the same thing as settling for propellant transfer (transferring to the engine(s) for relight vs transferring to another stage... in fact you have to be MORE careful to avoid bubbles when feeding propellant to an engine than you do when transferring propellant to another tank). https://www.youtube.com/watch?v=mVAGoWJuDKk https://www.youtube.com/watch?v=mVAGoWJuDKk They still need to do the whole launch-two-starships-then-dock-them bit, plus quick connect fittings, but they already have years of experience with autonomous rendezvous and docking from operating Dragon at ISS (Dragon 2 has that capability, and it has been used for the uncrewed Dragon 2 demo mission plus the recent cargo Dragon 2 missions, plus I think it's the nominal choice even for crewed missions).

angiosperm 3y ago
The difference is that in preparation to fire a rocket, you only need to apply ullage acceleration for a short time because after the rocket lights, that provides the acceleration. When just transferring fluid, you need to accelerate during the entire period of the transfer, because pumping fuel to another tank provides no acceleration of its own. So, you need the acceleration to be small enough not to build up to an inconvenient delta-V over the time it takes to move the fluid. In principle you could apply acceleration for a while in one direction; and then switch to accelerating in the other direction, wait for the fluid to settle at the other end, and pump from that end, alternating back and forth. But it still consumes reaction mass, and introduces delays while switching. Docking vehicles massing hundreds of tons is a much bigger problem than docking Dragon. The Dragon experience might not apply very directly. Presumably the engineers actually involved all understand this (literally rocket science) much better than we do.
- Robotbeat 3y ago
  You can do the settling thrust near perigee, so the delta-v from settling doesn’t go to waste but increases the apogee of the depot and increases the orbital energy of the propellant. Docking dragon is obviously similar to docking starships (Dragon berthing sensors were tested on Shuttle, broadly similar to an empty Starship). You’re right, tho, that the actual SpaceX engineers (who proposed this and developed Dragon) understand this better than random online folk.