Trevor Brown proposes spherical solar power satellites.
This isn’t a new idea. I wrote a paper on it back in the early nineties for an SPS conference, and I think that Geoff Landis has done some work on it as well (for instance, here’s a report of a talk that he gave on it at the 1996 ISDC, which was the last one that I attended prior to by going to Dallas two years ago — ctrl-F for “spherical”). It does vastly simplify the design issues, because it is no longer necessary to point the panels at the sun. One of the comments there needs some elaboration:
While the surface area of the sphere facing the Sun matches your calculations, the whole side would not be available for power generation. The so-called Beta angle, or the Sun angle, affects the total amount of power converted. Also, while a sphere would not need rigid station-keeping and attitude control to collect solar energy, the transmitter back to Earth certainly will. Also, a large spherical structure would be more taxing on a station-keeping/attitude control system than a more planar design. These caveats in mind, this is a creative alternative.
With regard to the needed area, the beta-angle effect means that at any orientation, you’re only getting the effective solar panel area of the cross-section of the sphere. That is, while the hemisphere has twice the area of the circular cross section, the non-zero beta angle of all points except that at the center of the illuminated area means that you need twice the solar panel that would be necessary if it were a flat circle. Add to this the fact that you have just as much area on the side in shadow, and it means that you need four times the total solar panel area to get the equivalent collection capacity of a pointed flat plate. So you have to postulate very cheap panels for this to make economic sense. But if you can get them, the simplification of the design is worth a lot.
As for pointing the transmitter, that’s actually not so tough a job. You hang it down below the sphere, and it will remain vertical, due to gravity gradient restoring torques. You could point it with control cables all around its circumference, attached to the sphere. In addition to inflating it, I also considered putting a charge on its surface to keep it spherical, but it would take a lot of ions, particularly for a big one, and inflating is probably a better solution, though subject to leaks, and the need for gas resupply.
One other point. I actually considered a fleet of them in MEO, continuously switching from one rectenna to the next as they orbit, to reduce the size of the transmitter antenna, which gets kind of humungous out at GEO.