Only between the air and the inside of the tube? Instead of a 100m train with air flowing around it, you get air at length of the track moving along the inside of the tube at the same speed.

I know as good as nothing of aerodynamics, but even if you get the latter flow perfectly laminar, I doubt it is a net win, as the contact area is 1000 times as large or so. But educate me.

Also, when you move cars into and out of the tunnel with the rotating air, you must take care not to 'plug' that air flow. Even if you do that only partly, there will be a pressure front (read: noise, vibrations, energy loss); each car will, for a moment get a sidewind at the speed of the train.

Here is a variant that, to me, makes more sense: it is not a vacuum, but they use the motion of the cars to push air out of the tunnel, so that air pressure drops. I don't have the vaguest idea how that would work, but if you can get it to work, it would be a nice trick.

I know next to nothing as well, but I think that perfectly laminar flow would be worst case. However if the inside of the tube had divots like a golf ball, the slower-moving turbulent vortices would push the faster-moving laminar stream away from the walls, so that the boundary layer doesn't impact the velocity of the central air column nearly as much.

Again: "I think..." I'd be happy if someone who knows more would chime in to tell me why I'm wrong. :-)

Turbulent flow would increase the overall drag in this situation. You're essentially taking energy from the free-flowing middle of the pipe and using it to drag more air along the wall. The reason why turbulence helps a golf ball is because it decreases flow separation http://en.wikipedia.org/wiki/Flow_separation. Separated flow makes a ridiculous amount of drag, and it's worth the extra skin drag to keep the flow attached.

Reply to self: we already have been building tunnels through which we move air at high speed. Looking at https://en.wikipedia.org/wiki/Supersonic_wind_tunnel#Power_r..., I read:

"The power required to run a supersonic wind tunnel is enormous, of the order of 50 MW per square meter of test section cross-sectional area."

A train would be, say, 20 square meters in cross section. That would be a GW of power to operate this. And that likely is a severe underestimate. I am sure one cannot lengthen a wind tunnel to kilometers without lots of power loss, but for now, let's ignore that.

For reference: Looking at https://en.wikipedia.org/wiki/TGV#Rolling_stock, a TGV uses at most 12MW at the relatively low speed of 300 km/h. Also, the three Gorges dam produces 22GW (http://en.wikipedia.org/wiki/List_of_largest_power_stations_...)

I think that's just the power required to accelerate the air. A wind tunnel is a once-through system, not a loop.