What were the other tanks for?
The ESA page doesn't say, does anyone know and have a link?
The European Space Agency is planning what it thinks is a world-first transfer of fuel between tanks on an orbiting satellite that wasn't designed to do the job. The agency's venerable XMM-Newton X-ray 'scope has been orbiting Earth since 1999, thanks in part to daily engine burns that keep its orbit stable. While the 'scope …
At a guess I would say for attitude control thrusters that unload momentum from the reaction wheels to keep their speed within sensible bounds.
Some satellites use torquing coils to do that instead of chemical engines (basically pull against the Earth's magnetic field as needed) - not sure if there are any special reasons for XMM not using that, or maybe it has both and they managed to hardly ever use thruster fuel?
>The main difference being that while NASA would make a film sending Bruce Willis up to it, ESA would make a film with Benedict Cumberbatch thinking how to solve it from the ground<
Actually they made that film (Space Cowboys) which starred Clint Eastwood & several other well-known pensionable actors. Utter rubbish if you look at it with an even slightly critical eye, but quite enjoyable otherwise
XMM is an x-ray telescope with a small optical camera (which I briefly worked on nearly 20years ago) so no liquid coolant so generally a very long life.
With x-ray telescopes the life is generally limited by the operating system you wrote the data analysis software packages on. It's cheaper to launch a new spacecraft than to port all the analysis stuff from VMS to Ultrix, to Solaris, to Linux to Windows10
If tank-to-tank transfer wasn't designed in then I'd love to know how it is even remotely possible to retro-fit it - why would there be a path from one tank to another for any other purpose?
Is there an engine that shares both the non-empty tanks and the main tank?
Will shifting the fuel around affect the orbit?
God*, I love this stuff!
*Just a figure of speech.
Probably it's using a fuel selector and cutoff valves in a "non approved" manner. (Fuel flows out of full tank through normal "output" line, gets to tank selector valves, gets blocked from flowing to normal place, flows into empty tank through "output" line of empty valve.) There's a lot of valving and piping for redundancy in case a valve gets stuck. I can imagine that some imaginative use of valves can create a route from a full to an empty tank. All you then need is a pressure differential to get things flowing.
It won't be a retrofit (a bit tough to do for a satellite in orbit, at least now the shuttle's are museum displays), but I'd imagine it'll be lowering the pressure in the main tank as much as they can, overriding some interlocks and forcing a valve or three open to create a path for the flow between the tanks, presumably via pipework that isn't normally supposed to be open via that route/condition.
But it's also wonderful to see yet another piece of space hardware far outliving its designed lifetime, although the cynic in me does have to wonder if they are applying Mr Scott's principle on their lifetime estimates...
If you can get an engine with fuel up there alongside the satellite (which is probably spinning to maintain direction, just to complicate matters) you don't need to transfer fuel. Just use the engine you brought along to keep the satellite in orbit. All you need then is a way of attaching the engine+fuel to the satellite.
After you've gone to all the trouble of doing the rendezvous, anything else is basically a rounding error. You need an engine with fuel to do the rendezvous. All I'm saying is that you might as well use the hardware you brought along to do the job rather than inventing some new, wonderful and potentially disastrously misconceived scheme.
Note that Landsat 7 wasn't designed to be refueled, so they're going to have to robotically deal with the "permanently sealed" tanks.
"NASA has a builder to construct a five-ton spacecraft to catch up with the aging Landsat 7 Earth observation satellite and refuel it in 2020, employing robotic tools mastered in years of rehearsals on the International Space Station."
http://spaceflightnow.com/2016/12/09/nasa-selects-builder-for-robotic-satellite-servicing-mission/
Short of some kind of servicing mission (no way that's going to happen less than 5 years) or they come up with some clever operations hack.
Which looks like what they are planning to do.
Over the years ESA and NASA have come up with some amazing (and amazingly complex) procedures to make probes and satellites either do things they were not designed to do or compensate for parts that have failed, through a mix of very clever simulations followed by software uploads to implement the plan.
I suspect that this will involve some clever valve sequencing coupled with some patches to override certain safety features to do with tank pressurization.
I'll with them good luck