If you had a centrifuge in space (like in Rotating wheel space station - Wikipedia) how would you seal a bearing connecting rotating part of the spacecraft from non rotating one so the air wouldn’t come out?
The easiest way is to not have a large surface that needs sealing against vacuum. In the movie 2001: A Space Odyssey the centrifuge onn the Discovery (the long spaceship going to Jupiter) was entirely self-contained. The rotating drum was inside, and the outer skin had no rotating part.
In the same movie, the large Space Station didn’t have any portion of the exterior rotating relative to any other part. If you look closely, the Orion space shuttle that docks with it at the hub has to match both speed and rotation before it can do so, because the docking region is rotating with the rest of the ship.
A better question (as CalM implies) is: why would you want the non-rotating part?
O-rings, or something of the sort. see central tire inflation system for a practical example. Similar example is the self-recoiling air hose reel seen in many auto shops and machine shops: the reel has an o-ring joint at its center so that air can be delivered to the hose without having to twist the bejeezus out of the hose.
OTOH, a slight leak from a tire inflation system or hose reel isn’t a big deal, because the central compressor can pull in as much atmospheric air as is needed to maintain system pressure. If your spacecraft develops a leak halfway to mars, you don’t have an infinite source of air available to keep pumping into your spacecraft the occupants are well and truly screwed.
As CalMeacham says, the safest approach is to design your spacecraft so as to eliminate moving seals like that.
Why seal it at all? The counter-rotating hub can be in vacuum, with a series of air locks in each radial arm.
But it would be much easier to have the docking ship rotate to match the station than have it dampen any relative rotation to mate with the “stationary” hub.
There are dozens of ways to seal large, continuous motion rotary joints. The problem in space is the harsh temperature extremes and radiation. The actual seal only has to contain a few PSI, which is trivial for techniques that routinely handle hundreds to thousands of PSI.
Just offhand, I’d start with a pair of roller bearings handling the actual load, and a Teflon-based flexible seal.
Again why seal it at all, just allow a vacuum in that jointed section with 2 independent air locked systems. Perhaps if needed that chamber could be pressurized but once the need is over for that pressurization the air could be ‘restored’ to the practical limit back into the ship’s storage system with the rest left to bleed through the joint.
Air in that joint would also increase friction and would tend to slow the rotation or require power to maintain rotation.
So you could have both: one environment with gravity and the other with microgravity for experiments etc. (although in the center of rotation force would be small so maybe it would be enough…).
No need for a vacuum seal to do that, though. You could build a rotating section inside a non-rotating exterior (or a counter-rotating section inside a rotating craft) that would accomplish the same thing, without the need for a complex and easy-to-fail rotating vacuum seal.
Although it could be potentially heavier since you need both centrifuge and the “casing”.
I know! You could solve the problem by staying within a breathable atmosphere! 
I think finding work-arounds for a sealed rotary joint misses the OP’s point. No matter how many alternatives you can think of, there are and will be needs to have such rotary couplings in a space environment.
O-ring seals work fine on vacuum chambers, even for ultra-high vacuum (i.e. 10^-8 torr range, where even the tiniest amount of air leakage into the vacuum chamber is unacceptable). Admittedly I’ve only seen tiny ones like these; I’m not sure how well they scale up.
It’s also possible to add redundancy by having multiple seals and differential pumping.
I don’t deny that these are huge complications to add to a space station, but if such a feature were necessary, I think this would be the way to solve it.
But are there solutions perfect? I mean in space you rather can’t afford to loose air, so if they leak even a little it might be not good enough, also how long do they last?
This isn’t clear to me. Where do you absolutely need one, where some solution without such a seal would work?
I’m not being difficult. You could say that a solution using such a seal is “easier”. But in the Real World, it isn’t, precisely because such a seal would be in constant use, and constant assaulted by vacuum such out the plasticizers and space radiation cross-linking and hardening the seal. A big reason the Wankel Rotary engine isn’t more popular is because its seals were constantly in use and in motion, and tended to wear away and leak. So the question of “why do you need such a seal?” is by no means unimportant or trivial. You might call solutions avoiding it “work arounds”, but they may be the only practical ones.
You only need to maintain a pressure differential of one atmosphere; that’s not very difficult. Would you be amazed at something that can be submerged 30 feet underwater without leaking? It’s the same problem.
Actually, all spacecraft leak. They all have airlocks, docking ports, joints, and other seals exposed to vacuum. You can’t expect them all to be “perfect”. And all spacecraft carry extra air to replenish the expected loss from leakage.
In fact, I don’t think any spacecraft or space station today has a closed system. The International Space Station gets its oxygen by electrolysis of water, which is transported from the ground; CO2 and H2 are vented into space(full explanation and diagram here). Compared to that, the amount lost through leakage is small.
No, it’s not a question of being submerged 30 feet under water without leaking. It’s having a seal that can rotate continuously for long periods of time under 30 feet of water without leaking. That’s very different, especially after it’s been in place for several years, rotating the whole time.
I would sleep easily neither under 30 feet of water nor in space with a large rotating seal that was many months old. In either case, I’d build me a nice static seal and put my rotating stuff where it wouldn’t need a constantly rotating seal.
If I HAD to build a rotating seal in space (or underwater), I’d make sure that any places where people lived or worked were protected behind other, static seals, and wouldn’t count on my rotating seal being airtight.
Oh darn. I misread the title; I thought we were going to discuss “STEALING a spacecraft.” As in “Oh boy! Look! The KEYS ARE IN IT!”
Do you have experience with moving seals? I mean maybe they perform well enough. It was only my intuition that something that’s constantly moving so under a lot of friction doesn’t last very long and that’s why I made this thread.
Seems like however they seal submarine propeller shafts would be a great place to start- they go to 400 meters underwater or so; far more pressure to deal with than 1 atmosphere in a spaceship, and I suspect a leak there is a bigger deal than a leak in a spacecraft.