I agree, reduced drag is a worthwhile goal. the question is, how much do you actually save? on formula 1 race cars, (v10 engine, 16000+ rpm redline, 900+ hp, 2.5 litres displacement (IIRC), 2 hours between overhauls, budgets for engine development bigger than most state's lotteries, national and/or corporate pride on the line-- this is
not mommies' minivan) some teams pull a vacuum on the crank-case to reduce air drag on all the oddly shaped ironmongery that flails around in there; and they pick up substanial horse-power over what the pump requires.
but they pay a price, too; the vacuum increases oil consumption, as the hot oil is that much closer to boiling, and the vacuum also draws dirt from the outside world into the case seals. So reliablility suffers.
now to the question of pelton wheels: I agree with other posters, that filling the wheelhouse with water would substantially increase drag. But you could provide a calibrated leak, such that the wheelhouse never got to a low enough pressure to support the column of water represented by the tailstock, and avoid that pitfall. And you'd probably gain a bit of efficiency from lower air drag. but you pay for that gain in two ways: 1, you've got to pump out the air in the first place, and 2, you have effectively created a negative head on the tailstock, which makes it harder for the water to leave once it has given up its mechanical energy to the wheel.
If you've already got the wheel at the maximum head you can achieve with your site, then you're probably chasing a wild goose; you'll need some fall to get the tail water to overcome the vacuum. If you haven't optimized your head, then that's the place to start, not more ironmongery.
I think you'd have to have a really large setup to see any meaningful improvement from this. the same amount of effort and money could reduce your electrical losses by the same absolute percentage and get MUCH more real advantage from something the size we here would build.