Given computational capability, each oscillator can predict its drift and use this information to calculate and correct the expected drift. The control system will only be needed to calculate the difference between the calculated oscillator drift and the actual drift. Using this technique, it may be possible to keep the oscillator stability to Af/f<1012. Given the capabilities of a super-computer, it would be possible to keep an overall dynamic model of the system, with information on the characteristics of each oscillator continuously updated from the information available. The master computer could then be used to update the correction factor for each oscillator. More extensive measurement and calculation on the receiving antenna could simplify this procedure considerably. In principle the receiving antenna can generate information on the phase and intensity of the received beam at every location on the ground antenna. This results in a two-dimensional map of two independent variables, which can be used to readjust the transmitter phases to minimize the spot size. The transmitting antenna has 5 local variables: the phase and intensity of each local oscillator, and the x, y, and z location of each oscillator on the (flexible) structure. These are not independent degrees of freedom. The oscillator phase delay can be exactly translated into an equivalent distance. The oscillator location is further subject to two constraints, since the transmitter elements are arranged in a two-dimensional array. This leads to at most two independent degrees of freedom for each oscillator. The two-dimensional information from the receiver are thus, in principle, sufficient to calculate the transmitter phase corrections needed. Once the nearly-correct phase is established, with perhaps a dynamic model of the transmitter, it should be possible for the fine phase adjustment to be generated relatively rapidly. It should be noted that the phasing difficulty is not significantly increased if the transmitting array is to beam to more than one receiving site. In this case the transmitter output is the linear superposition of that which would be required for each individual output beam. This could be accomplished by having each receiving station send a pilot beam, or alternatively, each ground receiver could send phase and intensity information to the master controller, which then optimizes the output phase to put the required amount of power on each receiver. If the exact location of each receiver is known and the dynamic model of the system is good enough, the wavefront measurement and computation could all be done at one receiver. Discussion Table 3 shows a comparison of the mass of an integrated thin-film solar power satellite compared to the SPS system baselined in the 1980’s. Using the “conservative technology extrapolation the reduction in weight is more than a factor of ten; assuming a more advanced technology, a reduction in weight by over a factor of a hundred is possible. Since the integrated design provides a transmitting aperture the size of the photovoltaic array, not only is there potential for reduced mass, but there is also the possibility to achieve “economic breakeven” at a considerably reduced power level. The baseline SPS concept used a 1 square kilometer transmitting antenna and a 102
RkJQdWJsaXNoZXIy MTU5NjU0Mg==