... If a high maximum Cp was measured, it always appears that the scale model was rather large with respect to the dimensions of the wind tunnel. This creates tunnel blockage and a Savonius rotor is very sensible to tunnel blockage because it has not only a large thrust but also a large force perpendicular to the wind direction. Tunnel blockage results in a non realistic increase of the Cp value. ...
Thanks. That's good to know - and to check out before committing on what (and how big) to build.
... a Savonius rotor has many other disadvantages like the large amount of material which is needed, the very strong fluctuation of the torque and the thrust for a Savonius rotor with two blades,...
Yes, weight is a big issue. On the other hand, it serves as a flywheel, smoothing the torque variations with position, so it's not all bad.
Stacking three of them, offset by 60 degrees, is reasonably standard practice and should smooth out torque variations. (All but the (3*N)th harmonics cancel.)
... the low rotational speed which requires an accelerating gearing to the generator and the fact that it is almost impossible to limit the rotational speed and thrust at very high wind speeds.
Why bother to limit the speed? Let it spin. Edge (tip) speed is only slightly higher than the wind, not near the speed of sound, so if it's built strongly (or lightly) enough it should be fine.
A well designed and balanced magnet rotor, with good bearings, in a permanent-magnet-alternator motor conversion should be able to handle the additional speed with no problem. Shaft torque limit IS current limit and can be managed electronically. For "gearing" at these low speeds (comparable to a building ventilator fan) I think a cluster of V-belts (with a giant pulley, again ala a building ventilator on the mill side) are entirely adequate. They may be less efficient than a gearbox (90-98%, typically around 95%) but they're entirely adequte - and cheap.
Alternatively, we could go back to the early days of low-speed multi-pole, giant-rotor electrical machines. An axial flux alternator the diameter of the rotor would have a pole-speed the same as, and get as much out of the magnets and copper as, one 1/6th or so the diameter (at the poles) of a horizontal-axis machine running at a TSR of 6. (More, because the coils would have more surface area - crosswise to the wind - and get better cooling, so they could run a somewhat higher current with a given amount of copper.) Fasten the magnets down solidly and you can let it spin to its heart's content.
As for wind loading, it can't be more than that of a wall of the same cross section, right? A structure that would support the wall (or an elevated rotating advertising sign of the same swept area) would support the rotor.
.... A Savonius rotor with a diameter and height of 5 m would have a weight of several tons and the investement in material would be very much higher than for a well designed horizontal axis wind turbine which generates the same power at the same wind speed.
Five meters was a lot bigger than I was planning.
I was thinking of using 3/4" plywood as a starting point. That's 2.13 ppsf. Figure four 5-meter disks for a three-phase (ignoring a possible need for multiple layers for strength in a five meter mill...). Treat the airfoils as two 5x6 meter areas of plywood (assuming PVC pipe for the turns and half-round dowels or airfoil-trailing-edge style wedges for the edges, as appropriate.
Airfoils: 60 square meters = 1376 lb.
Rotor disks: 27 square meters = 615 lb.
Yeah, that's 1991 lb, about a ton of wood, not including any thickening of the disks for such a large mill.
On the other hand, a ten-foot diameter rotor, 12 foot high (i.e. four-by-eight plywood sheets, without cutting, for the bulk of the airfoils / vertical strength members), designed the same way, would only be about 222 lb. It also wouldn't need as much strength in the rotor disks, so a single layer of plywood might be sufficient.