I am interested in orbital refueling, not for satellites (though that’s a good idea) but for manned exploration and infrastructure. For instance, to support a lunar base we would want regular flights with reusable craft. The best mission profile would depend on an inter-orbital transfer vehicle that never lands. Instead it shuttles between LEO and lunar rendezvous with landers that stay at the moon and never return to earth. There is a lot of refueling in these schemes. Space-X’s Mars plan for instance, depends heavily on refueling in orbit.

A quick search shows there have been some refueling studies as well as tests and at least some space station refueling. They discuss the problem of boil-off of cryogenic fuels, and list various solutions.

But one thing I have not seen much about is liquid surging and settling. When under thrust it’s not a problem, but in micro gravity, how can we keep the liquid at the drain? Bladders wouldn’t work with most fuels. Centrifugal propellant settling would certainly work, but how? A rotating baffle could spin the liquid in the tank, but it would create turbulence against the walls. Also the way oxygen tanks like to explode, it would be nice to keep machinery out of the tanks. That’s what happened to Apollo 13. One could rotate the whole depot and either have a counter-rotating connection, or spin the ship when it docks, a la 2001. But that’s an awful lot of work just to feed a drain.

Does anyone know how it is envisioned to handle weightless liquids in tanks?

  • $\begingroup$ Posting this as a comment because I am not sure if this could actually work but maybe the fuel flow alone could provide enough "thrust" on the structure to take care of the ullage during transfer. For the initial "hump" before transfer a large enough fuel depot could also be using moving weights instead of auxiliary thrusters to save fuel. Besides, for fuel transfer ullage might not even be as much as an issue because constant fuel flow is not as critical as during thrust. $\endgroup$ Commented Feb 1, 2017 at 7:30
  • $\begingroup$ Thinking about it, maybe there are pumps that can handle gasses in stride. Gulping air would be really bad for a rocket engine but not necessarily so when just moving to another tank. I imagine a pump somehow sorting gas from liquid and returning the gasses to the other end of the tank. Perhaps a centrifugal pump could do this. So maybe just creating a flow is the answer. That would imply a preference for long cylindrical tanks over spheres. Maybe internal parallel baffles would help direct the flow. Just speculating here. $\endgroup$ Commented Feb 1, 2017 at 12:31
  • $\begingroup$ My other thought is maybe moving baffles in cryogenic tanks have become perfected to a point that it's not a problem. Someone who is well aware of the latest in rocket tanks would know about this. $\endgroup$ Commented Feb 1, 2017 at 13:07
  • $\begingroup$ In my opinion, putting the station plus the tank in a mild spin after docking would be the cheapest option. No fuel wasted, no orbit change, whatever saturation is put into the CMGs to start the rotation gets removed once the transfer is complete and the spin arrested, only saturation losses coming from changing moment of inertia as the fuel moves resulting in difference of torque to start and stop the spin. $\endgroup$
    – SF.
    Commented Feb 1, 2017 at 16:45
  • $\begingroup$ Related question: space.stackexchange.com/questions/2446/… $\endgroup$
    – Hobbes
    Commented Feb 26, 2017 at 12:18

3 Answers 3


Shuttle orbital manuevering / attitude system tankage used screens to ensure that fluid always remained at the tank exit. Sufficient fluid would cling to the screens to get the engines started.

The screens are shown in this schematic of an Orbital Maneuvering System tank. In this case the screens are only at the aft because this system always provided thrust in roughly the tank axial direction. (schematic from the Crew Operations Manual, linked below)

enter image description here

The spherical Reaction Control System tanks needed to feed at any (or no) acceleration vector and had screened galleries that girdled the inside of the tanks.

enter image description here


The Shuttle Crew Operations Manual states:

The forward RCS propellant tanks have propellant acquisition devices designed to operate primarily in a low- gravity environment, whereas the aft RCS propellant tanks are designed to operate in both high and low gravity, ensuring adequate propellant flow during all phases of flight. A compartmental tank with individual wire mesh screen devices in both the upper and lower compartments supplies propellant independent of tank load or orientation. A barrier separates the upper and lower compartments in each tank.

It's conceivable that such screen/acquisition gallery technology may be applicable to large-scale free-fall refueling designs. It does add weight to the tankage, but at least has the advantage of being a completely passive system. I don't know if it's been tested in a free-fall, high-flow situation (as I assume the refueling ops would be).

  • 1
    $\begingroup$ Refuelling on Salyut was not high flow. They took over a week to transfer the fuel. This was mostly due to power contraints $\endgroup$
    – Antzi
    Commented Feb 8, 2017 at 4:41

Ullage (i.e. getting the fuel at the end the tank you want it to be at and the pressurant at the other end) for large rocket stages in free-fall is usually handled with small auxiliary thrusters.

For small thrusters, it's practical to use bladder pressurization or some other solution; you use the small thrusters to slightly accelerate the large stage, and the fuel settles at the tail end.

It seems reasonable to use the same strategy for a fuel depot.

  • $\begingroup$ But then the engine ignites and provides the acceleration to keep the fuel in the right place - surely you wouldn't want to be constantly thrusting/accelerating while refuelling? $\endgroup$ Commented Feb 1, 2017 at 13:27
  • 2
    $\begingroup$ We're talking about very small accelerations here, and I believe Roman Reiner's comment above, that the pumping action itself will tend to push things in the right direction, is corrrect, so the ullage thrusters may only be needed at the start. $\endgroup$ Commented Feb 1, 2017 at 16:39
  • $\begingroup$ Right, that makes sense. Interesting physics at play! $\endgroup$ Commented Feb 1, 2017 at 16:43
  • $\begingroup$ @RussellBorogove: I'm afraid the pumping action won't be much of help. The total displacement of craft will be only as much as the center of mass moves with the fuel - over the whole time needed to move the fuel. Maybe a couple meters over good several minutes. Acceleration of order of 0.01g? $\endgroup$
    – SF.
    Commented Feb 2, 2017 at 6:05
  • $\begingroup$ In principle it only needs to be enough to keep the fuel at the right end after the ullage burn puts it there - arbitrarily low acceleration is fine. Dunno how much of a problem slosh and rebound would be in practice. $\endgroup$ Commented Feb 2, 2017 at 7:28

For the Salyut 6 space station, both Progress cargos and the station used bladder tanks.

The tanks were pressurised using nitrogen. The nitrogen could be recovered once the operation was done (at least on the station side).

To vent the pipes, they just let them open, exposed to vacuum for a week then purged with nitrogen.

Source:The Story of Space Station Mir by David M Harland

  • $\begingroup$ I wouldn't have thought that! Really interesting. Apparently they used unsymmetrical dimethylhydrazine and nitrogen tetroxide as fuel for Salyut 6. Obviously there are rubbers or plastics that are compatible. I would imagine that it would be hard to find flexible materials that could work with cryogenic fuels, even ("semi cryogenic" oxygen and methane). My thought is that the cold would make them brittle. I think they use rubber hoses with LOX, but I doubt they flex as much as a bladder would have to. $\endgroup$ Commented Feb 2, 2017 at 18:38
  • $\begingroup$ Not perfect tho. They had leaks (And fixed it!) on Salyut 7 $\endgroup$
    – Antzi
    Commented Feb 3, 2017 at 2:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.