SPS, 366 GW or 122,000 times the cited production line will be needed. Assuming an 80% learning curve, this will result in a further reduction in cost to Assuming a base pay of $40,000/year (skilled technicians with overhead) and a normal working year of 2000 hrs or $20/hr, this translates to a productivity of .57 kW/hr. At a solar cell weight of 2.5 kg/kW this is equivalent to a productivity of 1.43 kg/hr. These conclusions are based solely on the well established experience and learning trends. However there is every indication that breakthroughs in solar cell manufacturing technology are imminent which show promise of further reduction in cost. Reference 5 presents a summary of past and future potentials in solar cell technology with particular emphasis on extra-terrestrial production. Briefly, ribbon growth machines show considerable promise of achieving low cost solar cell production, but the basic raw material, semiconductor grade silicon (SeG 99.999% pure), currently has costs of the order of $100/kg because the process of converting metallurgical grade silicon to the semiconductor grade is lengthy and costly. Of more promise is the relatively new chemical vapor deposition (CVD) technique which relies on passing a silicon-bearing gas (preferably SiH.,) over a high temperature surface. The gas breaks up at the surface leaving crystalline silicon. This process has the great advantage of using metallurgical grade silicon which is currently available at about $.40/kilogram, material which is abundantly available. However CVD produces a poly-crystalline film and the efficiencies will therefore be lower than in the case of cells produced from single crystals, probably of the order of 8 - 10% depending on thickness and refining techniques used. Reduced efficiency may be a deterrent to the use of CVD cells on earth. Depending on location, approximately ten to twenty times more solar cell area is needed on earth than in space for a given output, and land usage problems alone would therefore dictate the need for high cell efficiencies. In addition, there are almost insurmountable storage problems associated with day/night and weather cycles which mitigate against use of terrestrial solar cells for base load power even at the low potential costs indicated by the recent developments discussed above. However, in space the problem of cell efficiency becomes much less important, particularly if nonterrestrial materials are used. Increasing the size of the solar cell blanket presents no land usage problem and depends only on the tradeoff between increased manpower required to manufacture the larger array vs the reduction in manpower associated with the use of, for example, the CVD process. Reference 5 indicates that, with a proposed highly automated CVD machine, the production of 1 SPS/year would require 250 units operating on an 80% duty cycle. Assuming each unit is attended by one technician per shift, this would require a total of 750 persons for one year to produce the required number of cells or 6.7 kW/hr corresponding to a productivity of 17 kg/hr, as compared to the 1.43 kg/hr arrived at by projecting the experience and learning curves. Evidently the predicted costs in dollar or person hours for the production of solar cells could vary widely depending on the assumptions used. The analyses contained in this paper have attempted to show how these variations will affect the conclusions both as to the ultimate costs of the SPS produced power and the choice of lunar vs terrestrial manufacturing for their production. APPENDIX B. DETERMINATION OF STAGE WEIGHTS TO LEO
RkJQdWJsaXNoZXIy MTU5NjU0Mg==