But that's at a pressure of 100Pa (1/1000 atm). How much power does it take to maintain a vacuum of 1/1000atm in a tube hundreds of miles long? It seems this point is skipped over in the whitepaper, so perhaps its negligible.
Edit: yes I was referring to the power to sustain a vacuum in a leaky system, of course no power is required to maintain a vacuum in a perfectly sealed system :)
To me, a 1/1000 atmosphere vacuum system IS a really leaky machine :) Anecdotally, the machine I used to work on would gleefully chew on that for hours idling while we scrambled to find the hiss on the system.
It made an awesome sound when it ground into the atmospheric pressure on initial pumpdown, but at about 1/100 atmo it got bored (mere minutes), and without additional high speed stages the mechanical pumps would backwash at 1/10000 atmo (though it would take a while to get from 1 hPa to 0.1 hPa, many multiples of the time it to to get from 100 hPa to 1 hPa).
Edit: fixed units (0.001hPa is still leaky for base pressures but a completely different thing from 0.001atm).
I used to work with a system of about 1m^3, 2 little mechanical pumps would get you to 10E-6 atm or so. A turbo pump got you down to 10E-9 or there abouts. The door on the thing was about 2.5 feet in diameter, with a big ol' O-ring that would get twisted and tangled all the time (we broke it once, gorilla glue worked just fine to fix it).
I also second the great sound they made. It was like 2 very angry horses going 'Brahbrahbrahbrahbrahrbah...' getting quieter all the time.
If the system will be welded together that'll reduce the leaks quite a bit and the maintenance too, which I think will be the biggest problem. It's not making it work, it's keeping the thing working that will eat up the costs. Eventually, as the thing ages, you'll have bigger and bigger leaks. Figuring out how to deal with that will be a moon-shot on it's own.
To geek out for just a moment on the topic: we had two fix-a-flat substances that required two engineers to agree to use (mostly as a joke interlock - someone had to stop management from slathering it on everything). One was basically putty that you could smoosh over a leaky o-ring (clamp and all); with some work you could seal anything for a few days, but it made an awful mess and it was darn near guaranteed to destroy base pressures later. The other was a vial of clear stuff that was disturbingly like super glue, which would seal any micro leaks like a charm and would hold for a couple pumpdowns. At least, consistently until forgotten.
The machine was so big that it was easy to forget where it was, which o-ring had the clear super glue on it (after a few weeks it was covered in plant filth anyhow, the tags were lost, and too many of the seals had sharpie Xs on them to trust anyhow). It was MASH's meatball surgery, but with low energy plasma physics instead.
I ab-so-freaking-lutely guarantee they will use gorilla glue on the Hyperloop once leak failures become boring and routine. Or super glue. Or giant sheets of Kapton.
Guaranteed. This is going to end up as a union job due to the height requirements and the vacuum. It will be an honestly dangerous job. That means quotas and cutting corners, because there will be humans and it will be 105 out and I just wanna watch the Giants game, man.
I imagine you might be able to locate flaws with a custom capsule along the lines of a pipeline pig[1]. It would be exceptionally hard to actually fix it from the inside whilst travelling, so you either need to do it from the outside (which requires site access and elevation gear), or bypass/shut down the tube and fix internally.
As Balgair notes, you can get away with a lot on the short run, but leaks are notoriously difficult to patch up. Vacuum leaks are super hard to detect at these pressure ranges, where you're lucky to hear or feel them at all. More than likely, your only indication is to leak helium into the system on purpose and use a mass selector on the outside to 'sniff' for ambient helium (which is extremely rare normally). It's terribly slow and time consuming.
You can always take the short cut and just sorta gunk up the leak to mitigate it, but eventually it'll wear out. Vacuum welds are far better, but also take a lot of skill to perform on the spot: any sort of crevasse will act like a leak (albeit a small one - at these pressures it probably won't matter), so the bead needs to be nigh perfect.
Or you just slam another plate on the outside, glue/weld it up and just sorta not worry about it. Heavy pumping covers a multitude of leaks.
So, young man deciding on the Marine Corp or the French Foreign Legion: Just wait 10 years. Those precision welding jobs 50 feet up next to 1 tor vacuum are going to pay out the bank. Union too, and probably never going away if this thing comes to fruition.
On the other side, Mr Musk gains a hell lot of know-how about plugging vacuum leaks quick, reliably and possibly automatically. Where do you think that might come in handy? Hmmm, on trip to Mars maybe?
I would actually expect leaks to be easy to detect. Just run a high-frequency high-voltage plasma discharge detector on a skeleton vehicle. Wherever the discharge quenches you have a leak. Barcodes laser etched during construction could help you localize the leak to within a few centimeters.
Pinholes might be repairable by pressing a sapphire drift against the leak and melting the metal underneath by shining a laser through.
True, but this either has to be done by paid union labor 50 feet off the ground near 1 tor vacuum or by a heck of a robot. The union guys will be certified up the wazoo and paid to match that and persnickety about it. The robot will be and engineering marvel and a hell of a good idea.
I mention in another reply that Mr. Musk also has plans that involve vacuum differences and not a lot of wiggle room. On a trip to Mars a robot hull fixing machine would sure be nice to have.
I'm sitting here in a lab with a vacuum system that is in the process of being repaired for use in testing some new spaceflight components. 1/1000 of the atmosphere, even on this broken, leaky thing, is nothing. Mind you, it's also runnng off of pumps that are a decade or two old and cobbled together from various other labs around here.
How much power does it take to maintain a vacuum of 1/1000atm in a tube hundreds of miles long?
It doesn't require any power, if there are no leaks. The power required will be determined by the amount of leakage that is considered tolerable to the system.
Regular mechanical pumps can haul gas to pressures of a hPa. Any lower, and you're looking at something like turbomolecular pumps. (And efficiency may call for a second stage of specialized mechanical pumps.)
But in my experience, small mechanical leaks are usually of the order of a hPa or so, meaning no high vacuum tech will be needed. Just a fleet of giant blowers and mechanical pumps. (I worked on a wide area sputtering machine for a few years with a few hundred HP of mechanical backing pumps for about a hundred foot long chamber with a 10 sq foot average cross section (order of magnitude error for anonymity ;)) The machine would haul down to 10 or 1 hPa with leaks you could _hear_.
So yeah, it could work practically, especially if it's a welded system and not a giant mess of bolts and o-rings.
Sections 4.2 and 4.5. It estimates the cost of vacuum pumps for the tube and stations at $260 million. I haven't seen if they've broken out costs to maintain the pumps, but it specifically also mentions compensating for minor leaks with higher pump capacity to be able to continue operating and fixing smaller leaks as part of routine maintenance.
Edit: yes I was referring to the power to sustain a vacuum in a leaky system, of course no power is required to maintain a vacuum in a perfectly sealed system :)