The basic question is:
"My motor is counting on getting pure liquid propellant. In space, the propellant likes to float around in clinging blobs and bubbles...how do I make sure I'm getting the blob instead of the bubble?"
The expert on this sort of thing is D. E. Jackle Jr., who wrote all of the PMD related Technical Papers at http://www.psi-pci.com
, if you are desperate for answers on in-space propellant tanks, click there now. If you just want a basic overall perspective:
The type of performances for PMDs (Propellant Managment Devices), depend on a number of factors:
- How much propellant do you need in a single off-axis maneuver
- How much accelleration is the off-axis maneuver going to generate when the tank is nearly empty
- From where are the off-axis maneuver accellerations going to be imposed?
- How long between off-axis maneuvers
Why do I keep saying "off-axis"...because an "on-axis" maneuver will pool the propellant right over the outlet, in which case your PMD needs are vastly simplified.
- For on-axis maneuvers, how much propellant will you need before the propellant pool settles? From which come:
- How long will it take the pool to settle?
- How much propellant is your motor gobbling up in the meantime?
So...there are a bunch of demand options:
- Ignition supply (for on-axis maneuvers)
- Specific one-time demand system
- Specific refillable demand system
- Flexible demand system
The specific choices for devices depend primarily on the volume needed to fill the demand of an on-axis ignition or specific off-axis maneuver (most often station keeping(GEO), drag reboost(LEO), or minor course correction(Planetary).) The usual "specific one-time" maneuver is despinning from the top of a PAM type upper stage or after the spin-stabilized firing of a solid-fuelled orbital insertion or apoapsis kick motor. In increasing order of complexity (therefore cost and residuals) and decreasing order of cost and reliability:
- vanes: accellerations less than 7mm/sec2 (hydrazine), often used to refill...
- sponges: accellerations less than 70mm/sec2 (hydrazine), demand volumes less than 4L. Can go with somewhat higher accellerations and lower demand volumes, or vice versa. Vanes and sponges can get residuals down to 0.10-0.25%
- big control sponges: pool falls out under any significant accelleration, but comes back. Used to control the propellant center of gravity and bubble position (Viking Orbiter: pointing accuracy for cameras and ability to vent pressurant if tank pressure is dangerously high) Bigger sponge means more residuals (probably about 1%)
- Trough: any demand volume, accelleration limits depend on ability to refill after maneuver, but can be up to about 1000mm/sec2 (tetroxide)
- Non-refillable trap: one-time demand for off-axis maneuver. Typically up to 500mm/sec2 (hydrazine). At great expense, up to 5000mm/sec2 (hydrazine). If off-axis demand exceeds 10% of tank volume, troughs and galleries become more practical.
- Refillable trap: repeated demand for off-axis maneuver, or ability to hold propellant for on-axis maneuver start while doing off-axis maneuvers from another supply. Typically up to 500mm/sec2 (hydrazine). At great expense, up to 5000mm/sec2 (hydrazine).
- Gallery: up to 100mm/sec2 (tetroxide); 1000mm/sec2 (tetroxide) at great expense. Gallery type devices are often used inside traps and troughs.
- Gallery pickup: a bit like an inflexible version of the hose inside an aerosol can; there can be one or many and don't each have to have their own tube.
- Gallery channel: A vane-like arm with a channel covered by a screen or perforated sheet wrapping around inside the tank walls where propellants need to be picked up
- Gallery liner: a liner around the inside of a tank with perforated sheet or screen covered holes. Performs like poo with lots of residuals (1-2% or more) and complex construction when used in large tanks. Use only for small tanks where building channel and pickup type devices is a pain.
- Diaphragm: All of the liquid propellant is held behind a flexible rubber barrier with little to no gas in it. Flexible demand to very high accellerations (i.e. 2g or more). The downside is that the residuals quantity is second worst behind the gallery liner. Diaphragm rubber is not compatible with nitrogen tetroxide, and so is generally used only in monopropellant systems (because tetroxide has only 27% of the surface tension of hydrazine, bipropellant tanks often have different PMD in the fuel and oxidizer tanks.)
Diaphragms are useless for vapour pressurized systems because the evolved pressurant would wind up inside it. Other vapour pressurized systems must be careful not to evolve (boil) pressurant gas behind the last gas barrier in the outlet. They also need to limit boiled vapour in front of the last gas barrier, in which case it could bubble through if there is too much (i.e. no or too small a liquid path to the outlet.)
So...if you're looking to put a propellant tank on your spacecraft, I hope this gives you a good idea of how to figure one out that can survive all of your spacecraft's needed acrobatics (the highest of which in Orbiter is probably the time-overaccellerated entry which can send you out of the solar system at a high multiple of the speed of lignt.)