A reference publication I often cite. The pdf is available and more complete, including the equation (3) page 29:
Fig. 13 indicates that the power consumption scales with the cube of
the cylinder tangential velocity. Using the analytical formula proposed by (Subramanya, 2005):
Power = Cf . ρ . Utan³ / 2 . Areas (3)
A close agreement with the experimental results is found by setting Cf = 0.007, that can be considered a reasonable value for the friction coefficient. It should be noticed, however, that the actual power consumption of a Flettner rotor is arguably also affected by the functioning of its mechanical systems.
Page 20 (Nomenclature):
Areas Cylinder surface area, π D H
Now if we take a look of the page 4/4 of Norsepower pdf (first link above), we can see (for example for the first rotor) a diameter of 4 m and a height of 18 m, and 225 rpm, leading to a tangential speed (Utan) of about 47 m/s.
With the formula: 0.007 x 1.2 x 47³ / 2 x 72 x π = 98.58 kW, so about 100 kW.
With a wind speed of 10 m/s, 47 m/s tangential speed allow a very good spin (TSR) ratio of 4.7, and a high lift coefficient (Cl) of 8.5 (Fig. 4).
Now, if we take a crosswind Magnus balloon like Wind Fisher , the apparent wind speed becomes the reference wind speed to determine the spin ratio (TSR). If wind speed is 10 m/s, the apparent wind is about 26 m/s (for a reasonable L/D ratio of 2.6). If we want keep the same very good spin ratio (TSR) of 4.7 to keep the high Cl of 8.5, the tangential speed (Utan) becomes about 122 m/s.
And the power consumption increases by the cube of the tangential speed (122³), becoming about 17 times the power consumption with a tangential speed of 47 m/s (47³). In the meantime, the wind speed remained the same, i.e. 10 m/s. The improvement in Cl (the L/D ratio practically does not change) will therefore not be able to compensate for the electricity consumption which will become too higher.
And for an inflatable balloon the power consumption becomes far higher (perhaps 2 or 2.5 times more (by equalizing the parameters), as we can deduce from the curve of power consumption. But with a vertical trajectory with a wind speed of 10 m/s, and with a TSR (spin ratio) of about 1.2, this remains workable.
For AWE use, perhaps systems like turbovoile at a high lift coefficient would be less demanding at high wind speeds, or quite simply crosswind kites with high lift coefficient.