artistic merit: 10/10
logic: 2/10.
having the brains to ask some other people "hey, whaddya think of this?" before building something that won't perform as expected: +20 bonus points.
<I'm laughing at myself here. . . I can remember when I'd have gone through 6 hardware iterations of something like this before I realized my theory had a flaw in it. . .>
all other things being equal, a flow that rotates as it travels down a pipe will have a higher velocity relative to the pipe wall than one that does not. this is not an efficient state of affairs.
to see the effect of a water stream rotating as it leaves the nozzle, look no further than the nearest airless squirt bottle. . . the kind used for windowcleaner, etc.
the liquid is caused to rotate so as to improve atomization using centrifugal effects. . . and you can feel the difference in back pressure between 'stream' mode and 'spray' mode when you pull the lever. so, in the case of a nozzle, you would actually dissipate energy with a rotating jet.
in fluid dynamics, you will find that the most efficient way to move a fluid is to do everything you can to reduce turbulence. turbulence, in some ways, resembles a rotating flow; it has the property of transferring energy from the body of fluid out to the 'boundary layer' which is slowed by friction with a surface and viscosity effects.
the usual ways to reduce turbulence losses are to use very smooth surfaces in contact with the fluid, and to minimise the area of that surface. also, reducing the the number of turns and cross-section changes will help. Reducing the velocity gradient is important, too-- if you start by acknowleging that the fluid in contact with a surface (pipe wall, mill blade, wing, duct, stream bed, car body, whatever) is traveling at zero velocity relative to that surface, and think about what viscosity will do to a higher-speed flow further from the surface, the best course of action should be obvious and logical. . . . get as much of the flow as far from any surface as possible.
in our terms, all of ^^^that effectively means using the biggest diameter, smoothest pipe that will flex to follow your routing that you can afford, and to use it in the longest runs between manditory joints that you can manage.
my first choice if I were aiming for maximum energy capture (and hang the cost), would be large-diameter thin-wall copper refrigeration tube, well protected fom dents, and rolled into place from a spool. . . . though sch 40 PVC of the largest practical size would work well, too.