Hacker News new | comments | show | ask | jobs | submit login

One point that just occurred to me.

If you put this tube inside of another tube, and then pumped most of the air out of the second tube, the vacuum in the inner tube would improve and the whole system would become more efficient. I believe it is linear. For instance if the outer tube reduced atmosphere to 1/3 air pressure, then the energy losses due to pushing air around drop to 1/3 of what they would have been. However local compromise of the outer tube does not significantly compromise the vacuum in the inner tube, which avoids major problems that normally exist in maintaining an evacuated tube.

A lot depends on the flaps that let air out but (mostly) not in. When a plunger goes by, you need it to be open for a very brief period to let air out. I'd be curious whether it is better to have a lot of little flaps that can open/shut fast because they are little, or to connect the flaps to a computer system that can open the flap just in time and close it again promptly.

The smart system in theory can be more efficient, but the straight mechanical system would be massively easier to design and build.

I'm also curious what breakthrough Elon had over the summer that dramatically reduced costs. The things that I'm thinking of involve the design of the plunger. Possibly a spike creating chaotic flow at the right place. Or a plunger design that more efficiently pushes air out. Or some clever design of the behind of the plunger so that the air that is left is not following at high speed. Or even the idea I put above about having 2 shells, with partial evacuation in sections of the outer one.

Hey Ben,

That's pretty clever, so you're using the 'plungers' (carriages?) as the means of partial evacuation and propulsion at the same time...

No true vacuum, just enough to reduce drag to something manageable.

One thing that still bothers me about the whole concept (besides it being a complete guess as to what Musk has in mind of course) is the stopping properties, emergency shutdown.

If you would have to suddenly decelerate in the system I proposed you'd be building up pressure in front of the car which would slow the car down until some maximum pressure i s reached at which point you can breach the walls to relieve it and keep it at that maximum. This provides a nice cushioning effect for emergencies using very little besides the components already in the system.

If you were to even partially evacuate the tube then that would no longer work and you'd need active braking at all times.

Accelerating into the at a speed slightly higher than the loop speed backing into a slug that has just passed might be a way to mitigate this to some extent, then you still have to worry about stopping the train but the braking energy could be dispersed over a much longer distance due to closer packing.

I missed the end of the interview linked by someone else in the other thread where Elon Musk says it has no 'right of way' issues so I think the whole thing is moot but it was a fun exercise anyway.

I've seen his right of way comment. But if you've got something ground based, that is 1200 km long, how can you avoid right of way issues? My guess is that he means to build it in the ocean. There aren't a lot of people living there. If so, it could look like anything you want and there would be no "right of way issues". (Update, it just hit me! If you build this in the ocean, with neutral buoyancy under water, then you don't have to worry about earthquakes or weather! Built on land and able to handle earthquakes is a lot harder.)

But the stopping properties deserve more thought. California has regular earthquakes. What happens if you've got large hunks of metal going 1200 km/h over a section of loop that just was hit by one, what happens?

My suspicion is that you build regular emergency exit tubes. (Many of which might also be good for regular tube service.) The cars are let out at the nearest one. The machinery is at risk, the people are mostly OK.

You then try to stop the plungers with regenerative braking, dumping air in, etc. Or you can build a fast emergency stop, such as having exit holes that you can drop plungers into at high speed, which carries them far enough away from the tube that they won't destroy it when they impact.

This has to be carefully worked through. But the first and most critical piece is the ability to quickly and reliably dump passengers from most of the system. And that shouldn't be that hard to do.

I read somewhere (can't find a quick reference) of a coating that resembles the structure of shark-skin, this reduces drag enormously by creating local turbulence which reduces the shear forces required to separate the layers of fluid adhering to a surface.

A coating like that on the inside of the tube and the outside of the vehicle would be an interesting component but the fluid dynamics of such a set-up are way beyond my abilities to calculate.

This whole thing is all about energy budgets, you'd have to take the amount of fuel it takes a passenger car or rail car to move from one point to another, then take some reasonable savings estimate and then work backwards from there to see how much energy you could expend on friction, drag and so on.

I have a hard time accepting that the friction in a tube while drafting would be that much higher than friction losses to a body of air that is mostly standing still and from playing around with a car and a friend of mine who is a trucker (don't try this at home kids) I've seen how low fuel consumption can go if you are properly in the slip-stream of another vehicle.

There has to be some way to make the numbers work on that. But unless 'no right of way issues' means underground (or ocean bound, which means it would only work between coastal cities and would have a pretty terrible failure mode) the whole thing is off the table anyway.

I can't stop thinking about it though :)

You could brake (at some rate) by keeping the flaps ahead of you shut. The interesting issue is to make damn sure all braking is in sync enough.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact