SCIENCE SDe~~:,~~:.N1o94~~ AMERICAN ASSOCIATION FOR THE ilDVANCEMENT OF SCIENCE
5 December 1975, Volume 190, Number 4218 SCIENCE Space Colonies and Energy Supply to the Earth Manufacturing facilities in high orbit could be used to build satellite solar power stations from lunar materials. Within this century it may be feasible to establish manufacturing facilities in space, possibly in the vicinity of one of the Lagrange libration points of the earth-moon system (/. 2). Near two of these points, called L4 and LS, there are orbits which are stable under the combined gravitational effects of the earth, the moon, and the sun. A space manufacturing facility (SMF; the terms space community and space colony have also been used to describe such a facility) would be a self-sustaining habitat for a large number of people (of the order of 10' to 105). Its energy needs would be met by solar power, used directly as sunlight for agriculture, as process heat for industry when concentrated by mirrors, or indirectly as electricity. The SM F may be economically more effective than alternative industries on the earth for the construction of products whose end use would be in geosynchronous or higher orbits. Such products, if made on the earth, would have to be lifted by rockets out of the earth's gravitational potential well, which is about 6500 km deep. In contrast, the SM F would obtain the raw materials for its products from the surface of the moon, whose gravitational well is only I/ 20 as deep. As a consequence of the moon's vacuum environment, and of that factor of 20 in energy, a launching device located on the moon could transport material to the SM Fat low cost relative to shipment from the earth. In this article I suggest that solar power stations may be con5 DECEMBER 1975 Gerard K. O'Neill structed at a space colony, and relocated in geosynchronous orbit to supply energy to the earth, at a lower cost than if such stations were to be built on and lifted from the earth. Energy Needs The increasing demand for electricity, the shortage of fuels on the earth, and concern about widespread use of nuclear energy have led to consideration of satellite solar power stations (SSPS's). Glaser (3) has studied the SSPS concept, which is the location in geosynchronous orbit of stations converting solar into electrical energy, to be sent down as microwave power for conversion to direct current or to a power line frequency at the earth's surface. In 1975 the Energy Forecast Working Group of the Institute of Electrical and Electronics Engineers (IEEE) summarized forecasts, by 12 organizations, of the electric generating capacity which will be required by the United States during the years 1975 to 2000 (4). The IEEE summary estimated an increase from about 500 Gw in 1975 (5) to a required capacity of 781 to I070 Gw in 1985, and to a capacity of 1880 to 2250 Gw in 2000. The IEEE estimates therefore correspond to an average construction rate of new generator capacity of about 65,000 Mw/year in 1990 and 115,000 Mw/year in 2000. A study by Associated Universities, Inc. (AUi), predicted a demand for 85,000 Mw/year of new capacity at the turn of the century (6). The discussion that follows is not sensitive to such differences in the estimates. At current prices [typically $450 per kilowatt installed for a coal-fired plant (7)] the forecasts therefore correspond to a market of $30 billion per year in the United States alone in the year 1990, and $40 billion to $50 billion per year a decade later. [The dollar figure may be conservative; the AUi study (6) was based on the assumption that most of the increased capacity during 1985 to 2020 would be powered by nuclear reactors, with higher installed costs (7) of $600 to $1800 per kilowatt in 1972 dollars.] Economic selfinterest would tend to enlarge the market to the wider range beyond national borders; assistance to developing nations in the form of electrical energy would also increase the total production requirement for new power plants. Environmental Effects Each method so far considered for power generation has characteristics which are potentially damaging. Nuclear power produces radioactive wastes and materials convertible for use in nuclear weapons. Coal-fired plants require extensive stripmining to keep them supplied. In the year 2000, electric generation for the United States alone will require the mining of more than 2 x I 09 tons of coal per year, unless alternative sources provide most of the energy needed at that time. Transmission to the earth of the energy generated by an SSPS would require a microwave beam to a central antenna. That may be less desirable environmentally than the high-voltage lines used conventionally at the surface of the earth for the interconnection of large generator plants. Microwave transmission may, though, be more acceptable than the alternatives of nuclear power or strip-mining, and that is an important issue which should be studied carefully. Glaser (3. 8) has stated that the microwave beam intensity outside the antenna site would be low enough to satisfy stringent environmental requirements. Because the conversion of microwave energy The author is professor of physics at Princeton University. Princeton, New Jersey 08540. 943
to direct current could be about 90 percent efficient, an SSPS antenna array would release into the biosphere only one-tenth as much energy as it would deliver for use. In contrast, a fossil-fuel or nuclear plant discards as waste heat about 1.5 times the energy which it delivers to the power lines. The land-use requirements for an SSPS ground antenna array would be only about 1/ 10 to 1/ 100 as great as for direct photovoltaic energy conversion of sunlight, because direct conversion is limited by low efficiency, the day-night cycle, seasonal variation of the day length, atmospheric absorption, and cloud cover. Economic Factors Delay in the realization of the SSPS concept appears to be due mainly to lift costs and power plant mass. The installed cost of an SSPS would depend primarily on four factors: the capital cost per kilowatt of the power plant for converting solar energy to electricity ($/ kw); the specific mass of the power plant in kilograms per kilowatt of output (kg/kw); the cost per kilogram of lifting the power plant from the earth to geosynchronous orbit ($/ kgsy); and the overall efficiency of converting electricity into microwave energy, transmitting it to the earth, and reconverting it into direct current or to a power line frequency (E). The installed capital cost of the installation, per kilowatt of power output from the antenna busbar on the earth, would be approximately I t [$/ kw + (kg/kw) ($/kgsy)] plus interest charges, development costs, and smaller additive terms for the ground antenna and the land it would occupy. Earth-Launched Power Satellites Solution of the microwave transmission problem (9, IO) for the SSPS appears to be progressing well: tests have already demonstrated a transmission efficiency f (direct current to direct current by a microwave link) of 55 percent. The goal is an efficiency of about 63 percent. For an earthlaunched SSPS, the economic problem lies in the remaining factors: capital cost, power plant mass per unit power, and lift cost. Two alternatives for the conversion of solar energy to electric energy in an SSPS ·have been considered: photovoltaic cells (solar panel arrays) and turbogenerators powered by mirror-concentrated sunlight. Glaser (3) has estimated that for an earth-launched SSPS the specific mass for solar panel arrays will have to be reduced to about 0.88 kg/kw. For comparison, the 944 value for photovoltaic solar cell arrays in operational satellites of the last decade has ranged from 78 to 107 kg/kw; one experimental satellite designed as a short-life test vehicle achieved 29 kg/ kw (/ /). For the Solar Electric Propulsion System space probe scheduled to fly in 1984, the specific mass is intended to be 13 kg/kw(/2). Present costs of solar panel arrays for space applications are based on manual assembly techniques and are therefore much too high for application to an SSPS; they are typically $175,000 per kilowatt (3). A more reasonable starting point is the 1971 figure of $5,000 per kilowatt for singlecrystal wafers 5 cm in diameter. The necessary target figure (3) for a competitive SSPS launched from the earth is about $220 per kilowatt, about half the present cost of a large, coal-fired central power station. As an alternative to a photovoltaic array for SSPS power, Woodcock and Gregory (/3) have considered the use of closedcycle helium turbines (/4) driving conventional electric generators. In that alternative the specific mass must be reduced from presently attainable values (IO kg/ kw) to about 5 kg/kw. To achieve that reduction, Woodcock and Gregory have assumed a development program in which the turbine inlet temperature could be increased to a value considerably higher than that used in current practice. For an earth-launched SSPS the cost of lifting components to geosynchronous orbit from the earth would be critically important. For a photovoltaic SSPS of 0.88 kg/kw, the necessary lift cost figure would be $220/kgsy (3). For a turbogenerator SSPS of 5 kg/kw and efficiency f of 70 percent, the performance demands on the lift vehicle would be even more severe: $75/ kgsy (/3). Launch Vehicles An advanced, chemically propelled "space tug" could bring from low earth orbit to geosynchronous orbit, as payload, about one-third of the total payload delivered to low earth orbit from the earth's surface. When the cost of space-tug operations is included, the cost of transport ($/ kgsy) from the earth to geosynchronous orbit can then be taken as roughly four times the cost of transport to low earth orbit. For simplicity, lift cost figures in the following discussion will refer to the overall transport from the surface of the earth to geosynchronous orbit ($/kgsy) and will be taken as four times the cost to low orbit. An additional uncertainty of about ± 30 percent is a consequence of this simplification . The target figure for the space shuttle, planned for operation in the early l980's, is $1400/kgsy, not including development costs of several billion dollars (/ 5). A heavy-lift launch vehicle (70-ton payload) using the same kind of engines that are already being developed for the shuttle, and therefore obtainable without large additional expense for development, is estimated to be capable of achieving $600/ kgsy to $1000/kgsy (/6). To summarize, for an economically viable earth-launched photovoltaic SSPS the specific mass (kg/kw) must be reduced by about a factor of 30 to 60 below the corresponding figure for satellites of the l970's; the lift cost to geosynchronous orbit must be reduced by about a factor of 4 below figures estimated to be attainable in the l980's without large additional development costs; the capital cost ($/kw) must be reduced by a factor of about 30. For an earth-launched turbogenerator SSPS the capital cost must be held equal to that of a present-day coal-fired plant; the specific mass must be reduced to about half the value currently attainable; f must be raised to 0.70; and the lift cost must be reduced by about a factor of I0 below the figure now considered to be attainable in the 1980's without substantial postshuttle development. Table I summarizes the values of these factors which have been assumed in several studies and, where the information is available, the resulting estimate of power cost. Extrapolations to vehicles more advanced than shuttle-derived rockets are necessarily subject to large uncertainties; new developments in engines, heat shields, reusable fuel tanks, and other components would all be needed before their construction. For a very large vehicle, capable of lifting 180 tons to low orbit, estimates of attainable recurring cost range from $80/kgsy to $900/kgsy, and estimates of development cost range from $5 billion to more than $25 billion (/ 7). Power Plant Economics In power generation, the busbar cost is crucial to the achievement of market penetration. Power plants are characterized as base load (operating nearly all the time), intermediate load (operating part of each day), and peak load (operating only during coincidences of maximum industrial and residential demands). Peak-load plants are normally simple and inexpensive to build and, when called into use, generate electricity at a cost up to 60 mills (that is, $0.06 per kilowatt-hour). Intermediateload plants are capitalized more heavily and generate electricity at 20 to 25 mills. SCIENCE, VOL. 190
Eventually photovoltaic solar cells located in the American Southwest may be competitive with one type of intermediate-load service: the supply of energy for air conditioning. Base-load plants (mainly coalfired and nuclear) supply power at IS to 17 mills. Nuclear plants in particular are best suited to base-load service: they must run nearly all the time to amortize the heavy capital investment required for their construction. Once started, a nuclear plant is kept running for another reason also: each time it is turned off there is a risk of component failure due to temperature changes. If electricity could be obtained from an inexhaustible source at 4 to 8 mills, lower even than base-load rates, it could have a profound impact on economic security and independence: residential and industrial heating could then be shifted to electricity, relieving demands on natural gas and oil supplies, and the production of synthetic fuel alternatives to gasoline could become practical. Like a nuclear plant, an SSPS would have to operate nearly all the time to amortize its construction cost. Economic viability of an SSPS would require, therefore, that it operate in base-load service, at rates not over IS to 17 mills. If SSPS power is to have major impact on the problems of energy resources and dependence, a way must be found to build and locate large numbers ofSSPS plants (up to 20 to 40 per year of S-Gw size) and the electricity rates at which they operate must be low enough so that they will achieve market penetration, being chosen for new construction in preference to alternative (coal or nuclear) plants. If those two conditions are not met, SSPS power can be no more than an exotic rarity, classed with hydroelectric and geothermal power among fringe sources (I to S percent) of energy. My purpose in stating these necessary economic conditions is not to discourage the development of a prototype SSPS. Clearly, though, it will be difficult to meet these conditions with SSPS plants built on, and launched from, the earth. In support of the viewpoint that SSPS development is justified nevertheless, I will outline what may be a way to meet the conditions of SSPS mass production and low electrical rates. The Space Manufacturing Alternative The effectiveness of an SMF program for the achievement of economical solar power on the earth would depend on two key elements: the use of lunar materials and the "bootstrap process" - the construction by the first SMF not only of SSPS units but of additional SMF's. The 5 DECEMBER 1975 Table I. Critical factors in satellite power station economics. The numbers assumed in several studies for the factors specific power plant mass. component lift cost from the earth, transmission loss factor (,-1 ) , and interest rate are summarized; in each case a higher number corresponds to a more conservative assumption. Earth-launched SSPS values are from (13) for those with turbogenerators and fro m (3) fo r those with photovoltaic cells. Data in the last column are from th is article. The li ft cost from the earth to geosynchronous orbit is approximately equal to the cost for lift to Lagrange point LS. For base-load service. busbar power costs are now typically 15 to 17 mill/ kwh. Specific Lift SSPS mass cost (kg/ kw) ($/ kgsy) Ea rth -laun ched Turbogenera tor 5 75 Photovoltaic 0.8 220 Built in space from lunar material 10 950 use of lunar materials would circumvent the problem of lift cost ($/ kgsy) and therefore of power plant mass (kg/kw). The bootstrap process would replace linear growth in the number of SSPS units by exponential growth. The establishment of the first SMF would require the transport of 3,000 to I0,000 tons to the lunar surface, and 10,000 to 40,000 tons to LS (2). The structural mass of the SMF has been estimated as IS0,000 tons (18), and the total mass including cosmic-ray shielding could be 2S to 6S times larger. The SM F would be built almost entirely of lunar surface materials. The lunar soil (regolith) as found, unselected, contains 20 to 30 percent metals, 20 percent silicon, and 40 percent oxygen by weight (19). Depending on whether the first SMF were provided at the outset with a massive cosmic-ray shield, or acquired such a shield over a period of years by the accretion of industrial wastes (slag) from the manufacturing operations at LS, the transport machine (mass-driver) for lunar surface materials would be required to lift 80,000 to 700,000 tons per year from the moon to LS. With full-time operation at a cycling rate of 30 kg/sec, the mass-driver previously described (2) would transport 940,000 tons per year. After completion of the SMF, the lunar mass-driver would continue to export raw materials to the SMF site. There, the processing plant already used for SMF construction would continue to produce metals, glass, ceramics, and other materials. In zero or low-gravity construction bays adjacent to the SM F habitat, those materials would be formed into SSPS components. An SSPS built at a space colony would be considerably simpler than one launched from the earth, because the colony-built SSPS could be designed without launch vehicle constraints. Turbogenerators could be fewer and of the most efficient size rather than kept within vehicle limits. Solar reflectors and waste-heat radiators could be built in large sizes and would never have to Interest Initial busbar ,. , rate power cost (%) (mill / kwh) 1.43 8 25 1.54 1.6 10 15 withstand launch accelerations. That is a significant advantage because an SSPS would be mechanically fragile: the specific mass figures of Table I imply an overall average thickness for the SSPS, including solar energy converters, radiators, conductors, mirrors, supports, and transmitting equipment, of only 0.08 to 0.6 mm of aluminum. The linear dimension of the SSPS would be several kilometers, about ten times larger than those of the SM F. On completion, the SSPS would be tested in space close to the construction site. It would then be moved to geosynchronous orbit through the small velocity interval (2.1 km/sec) which separates that orbit from LS. A second mass-driver, similar to the one which by then would have been in operation on the moon for several years, could be used for this task. It would be assembled outside the SM F and attached to the completed power station, to serve as a reaction engine. It would use as reaction mass industrial wastes, possibly liquid oxygen, left over from the processing of materials for the SSPS. As a reaction engine, the massdriver would have an exhaust velocity of 2.4 to 3.7 km/sec and a thrust controllable from zero up to a maximum of several tons. It would be powered by the SSPS during the orbital transfer time of I to 4 months. The economics of SSPS construction at LS requires a fresh viewpoint: in that construction almost no materials or energy from the earth would be required. The colony itself, once established, would be selfsustaining, and its residents would be paid mainly in goods and services produced by the colony. The economic input to the combined colony-SSPS program (Fig. I) is the sum of development and construction costs for the first colony, the cost of lifting the materials needed from the earth for subsequent colonies and for noncolony-built SSPS components, a payment on the earth of $10,000 annually to every colonist, repre945
Table 2. Cost estimates for establishing a first manufacturing facility in space. Cost figures not in parentheses are based on transport rates of $1900/kg from the earth to the lunar surface a.nd $950/kg to Lagrange point LS. Those in parentheses are based on the assumption (/3) of transport rates reduced by a factor of 12, with an additional $10 billion added for vehicle development and no change in administrative costs. In this simplified table, personnel rotation, if required, and material resupply are within the tonnage figures. Ad minisLift to lunar surface Lift to L5 Devel- trative Total Type of estimate Tons $billion Tons Minimal 3,000 5.7 10,000 Intermediate 10,500 20.0 42,000 High 20,000 38.0 80,000 senting that portion of salaries convertible to goods and services on the earth (for subsequent use on visits or; if desired, on retirement), and a carrying charge of 10 percent interest on the outstanding balance in every year of the program (20). That is approximately equivalent to discounted economics with a IO percent discount rate. Cost estimates for the first SMF (Table 2) are based on making an early start, with the shuttle and a shuttle-derived freight vehicle. The lift costs assumed for Table 2 are therefore relatively high, equivalent to $950/kgsy. The time line of Fig. 1 uses the intermediate cost estimate of Table 2. Because of the high interest rate assumed, the cost-benefit analysis is sensitive to the speed of construction of the SMF, 40 20 - 20 Total number of 5000 Mw SSPS stations Years from start of program opment and cost, and con- salaries st ruction (20%) rounded $billion ($billion) ($ bil- ($billion) lion) 9.5 11.0 5.2 31.4 (27.4) 40.0 20.0 16.0 96.0 (50.8) 76.0 40.0 30.8 185 (89.8) and therefore to productivity. The first colony, with a structural mass of about 150,000 tons, is assumed to be constructed in 6 years by a work force of 2,000 people. The corresponding productivity, 13 tons per person per year, is consistent with analyses by D. Morgan of current experience on the earth in materials processing and fabrication (18). I assume that subsequent colonies could be constructed in 2 years. This tripling of production rate would require devoting 4,000 people of a 10,000person colony to new-community construction (compared to 2,000 people available at the construction site during the building of the first SMF) and an increase of efficiency by a modest factor of 1.5. I assume that most of the residents of the 40 Total number of 5000 Mw SSPS stations 10 20 Power cost (mills/kwh) .. .l. '····· 30 Years from start of program Fig. I (left). Estimated costs and benefits of a program for construction of space manufacturing facilities. With assumptions described in the text, the total input capitalization is $178 billion over 14 years, in addition to the costs incurred on the earth for each power satellite produced. Interest of IO percent is paid on the outstanding balance in each year. Power costs, initially 15 mill/kwh, are reduced in steps to achieve market penetration. Fig. 2 (right). Effect of a 7-year delay in the program to wait for advanced lift vehicles~ Although lift costs are assumed to be cut by a factor of 12, the peak funding is reduced by only a factor of 2. Benefits, including energy independence, are delayed by 7 years. 946 early space communities will be employed in production, support services being assisted as far as possible by automation. Later decreases in the employed fraction of the work force are assumed to be compensated by productivity increases. The productivity required at LS for the SSPS would depend on the ratio of kilograms to kilowatts of output at the time of construction. In my assumption, an SSPS supplying 5,000 Mw of electricity at the busbar would have a mass of 80,000 tons; in the study by Woodcock and Gregory (13), 35,000 tons; in the study by Glaser (3), only 11,000 tons. Assuming that the remainder of an SMF work force, 6,000 persons, were committed to SSPS construction, and that two SSPS units were produced per SMF per year, the corresponding requirements for productivity would be 27, 12, and 4 tons per person per year. Most of the production operations of SSPS construction would take place within the weather-free, zero-gravity, enclosed environment of a space community's assembly volume, which should favor high productivity. For 12 to 27 tons per person annually, and a peak production rate of 160 Gw/year of new generator capacity, a total work force of 100,000 to 200,000 people in space would be required. In Fig. I, new colony construction is therefore taken to be halted after the 16th colony, because of market saturation. For the time line of Fig. 1 the benefit/cost ratio would be more than 3. Other time lines, with a variety of input capitalizations, productivities, and interest rates, have also been traced; only extreme cases yield benefit/ cost ratios less than unity. The relative insensitivity of the peak funding requirement to the lift costs assumed can be seen by comparing Figs. I and 2. With the assumptions of Table 2, a reduction of lift costs by a factor of 12 would reduce peak funding by one-half. Because of the exponential growth of the number of SMF's, satellite power could have a strong impact relatively soon. By year 11 from the start of SM F construction, the usable electric energy supplied to the earth by the program could exceed the peak capacity of the Alaska pipeline (2 x 106 barrels a day) (21). Two years later the production rate of SSPS plants could exceed the U.S. annual need for new generating capacity. By year 17, the total energy so far provided from the satellites could exceed the estimated capacity of the Alaskan North Slope (!010 barrels) (21). This discussion has been confined to technical questions. Clearly, though, if an SMF program is initiated, it will have wider impact than the science-oriented space programs that have preceded it. As an enterprise with the potential to return a SCIENCE. VOL. 190
profit and to tap an inexhaustible source of energy, it could be carried out as a joint venture of several or many nations. The worldwide food shortages that have been forecast for the next decades could be alleviated substantially by the provision to developing nations of low-cost energy for the manufacture of agricultural chemicals (22). In the SMF approach, subsidies of that kind to the Third World could be given out of new, nonterrestrial wealth, not requiring sacrifice by donor nations. The data in this article should be considered not as definitive, but as requiring substantiation or correction by additional research. So far, during a year of exposure of the SMF concept to technical review, no major changes in the basic concept have been necessary, but it is almost certain that further work will uncover both unsuspected problems and new technical possibilities. A modest amount of research on the key questions of productivity, life support needs, SMF and SSPS construction methods, and lunar materials transport could substantially improve our knowledge of the cost and time required for the achievement of the first beachhead in space, and of the speed with which the initial investment could be returned. 5 DECEMBER 1975 Summary The feasibility of establishing manufacturing facilities in a high orbit is under discussion. They could be used for the construction of satellite solar power stations from lunar materials. Estimates indicate that this may be considerably more economical than constructing power stations on the earth and lifting them into orbit. References and Notes I. G. K. O'Neill, Nature (London) 250, 636 (1974). 2. ~Phys. Today 27, 32 (September 1974). 3. P. E. Glaser, in Space Shuttle Payloads. Hearing Before the Commillee on Aeronautical and Space Sciences. U.S. Senate. 93rd Congress, Ist Session. on Candidate Missions for the Space Shu11le, 31 October 1973 (Government Printing Office, Washington, D.C., 1973), part 2, pp. 11-62. 4. G. D. Friedlander, IEEE Spectrum 12, 32 (May 1975). 5. Edison Electrical Institute, Statistical Yearbook of the Electricity Utility Industry for 1973 (Edison Electrical Institute, New York, 1973), p. 5, table 1-S. 6. Associated Universities, Inc., Reference Energy Systems and Resource Data for Use in the Assessment of Energy Technologies (AET-8, Associated Universities, Inc., Brookhaven National Laboratory, Upton, N.Y., 1972). 7. W. R. Cherry, Astronaut. Aeronaut. II, 30 (August 1973). 8. P. E. Glaser, NASA Contract. Rep. CR-2357 (1974). 9. W. C. Brown, Proc. IEEE 62, 11 (January 1974). 10. News release, Office of Public Information, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, I May 1975. 11. R. E. Austin and R. Brantley, presentation at NASA Headquarters, Washington, D.C., 17 April 1975 (unpublished). 12. R. E. Austin, personal communication. 13. G. R. Woodcock and D. L. Gregory, AIAA paper 75-640, presented at the AIAA/ AAS (American Institute of Aeronautics and Astronautics/ American Astronomical Society) Conference on Solar Energy for Earth, 24 April 1975. 14. K. Bammert and G. Deuster, paper presented at the ASME (American Society of Mechanical Engineers) Gas Turbine Conference, Zurich, April 1974. 15. "Launch vehicle estimating factors for advance mission planning," NASA Handb. NHB-7100.58 (1973). 16. R. Wilson, in Proceedings of the Conference on Space Colonization, Princeton University, JO May 1974, G. K. O'Neill and R. W. Miles, Eds., m press. 17. H. P. Davis, in Proceedings of the 1975 Princeton University Conference on Space Manufacturing Facilities, G. K. O'Neill, R. W. Miles, R. H. Miles, Eds., in press. Typical values are: nonrecurring, $10 billion; recurring, $220/ kgsy. 18. NASA-Ames/ ASEE (American Society for Engineering Education/Stanford University Summer Study Report, System Design for Space Colonization (NASA-Ames Research Laboratory, Mountain View, Calif., in press). 19. B. Mason and W. G. Melson, The Lunar Rocks (Wiley-lnterscience, New York, 1970), p. 117. 20. G. K. O'Neill, in Future Space Programs 1975, Hearing Before the Subcommillee on Space Science and Applications, Committee on Science and Technology. U.S. House of Representatives. 94th Congress, /st Session. 23 July 1975 (Government Printing Office, Washington, D.C., 1975), p. 111. 21. Exxon Corporation, advertisement in Smithson. Mag. 6, 117 (April 1975). 22. D.R. Safrany, Sci. Am. 231, 64 (October 1974). 23. It is a pleasure to thank P. Glaser, M. Hopkins, D. Morgan, and G. Woodcock for providing useful information not yet published. The latter part of this study was carried out with support from the National Aeronautics and Space Administration. 8 January 1975; revised I October 1975 947
c;. A H I C.~ Suggestions re design sketch: 1) For details on this, the painting which has just been reproduced on the cover of the Dec. 5th issue of Science is quite good. Perhaps the entire object (the zero-gravity assembly) could be made somewhat larger on the painting, for more detail. 2) To emphasize that the ~entral facility (1) is in zero-g, its supports should be shown as very light and spidery, 3) There is an open cylinder going back toward the docking area. It -~is both an air passage and an~v~~ for floating people and objects. ~--,:. There could be detail there if you wish. 4) Windows could show as light hazy color. In actuality would be transparent, and the solar disc would be reflected inside. However, there would be a rib structure too small in scale to be seen in this painting: about 20 inches on a side (rectangular or honeycomb pattern) 5) On the Disc radiator, there could be a structure of radial pipes shown; or this could be more complex, with pipes, fins etc., as on a transistor-radiator. 6) Is this to be the Moon? Could be either earth or moon, but don't show both in the same picture. 7) Here's where both the challenges and the opportunities come. Items 1-6 are by now fairly routine. (7) offers new possibilities. I 'm bother~d that the viewpoint somehow doesn't seem ~o be quite from the right place; its's not clear to me that one will see enough of the equatorial region. The following sketch shows some of what I had in mind: fi,CIN ,. ,.,r,c.c '~."'!:},.•/ SVN A"fG.1.1 IN .,,., NOttTH H•Ml$t'Hlt'f.« (4~). ~CITM HllMICPHIU (~¥ QA.,.I TVJ>6S Foll. ,.-a"-lllf'b TM.&.rn •N••V•.P""'- t1u11.1>1Nf,. A,~4T~•~., SMALL- !»Hof'~ ,_.. tTOf\I&• -J.4', O"- 3."'J "'"' HRA.r\.OW f'AntS1 &.ITT&.e PNZics, '-. ~ ON TMtS 'scALC) R,.TM£tl Lii(. AH IT'AL.IAft' .·~-~ (;; l_ IL&. .. TOW~ t s' .. (J_, w1M:HWI ~IJ.lfT l"l>IVtOfll\I. T"tla6 ",.TO ""' C"°'. 4 .....,,, ,...,. ,,A .. r) ... ,oe~ _, . ., , . , •S• I ,.)C..~ C~'"' ~f6N VIC.&Tlt\'rl.,.., •' ,,,~ ""-." •t. ,. A~vt <00° 47Ho~ ii:r '"' '~'" ~·· ~c.~wl &.IHE!a, ..., . . , ,. . e-.Je> ,,.,, '-•Ua.c.Y "rffJCIC f s1'l-f. ~t I I I I l o .... ao<fOMITM'I scAa.• ~ aooo' ' <-ratt""'P J t'f.J:I. ,,. """''-"~' .. er,ff~ WUIPOW Aft« AS ) , .... \t. I l / ~ - ,.,,:: r\OTATfO,_, Mt$ " •• •': ~s· -.. ,_,rs OF &.Arl'ro06 11 ;.._ &ClCJATO" ., The aim is to make as much use as possible of the fact that this Bernal Sphere geometry gives the freedom of landscaping in a large area, about 3qo meters wide (roughly a quarter-mile) and 1600 meters in circumference. I would emphasize the following: a) Keep everything small-scale. No monumental buildings. As much as possible, diversity and irregularity rather than a repeated pattern. b) Dwellings should be shown mainly as terraced apartments, each with a garden with much vegetation. c) No large roads (because no vehicles except bicycles and perhaps small elec~ric three-wheelers for shopping-parcel carrying). Many small paths and bicycle lanes. d) Much area left as parks, playing fields. e) One stream winding its way along the equator, with beaches on each side. Could show bathers, canoes. f) Possibly small streams cascading down from the higher elevations. There could also be fountains. g) In this geometry one would want to take advantage of the fact that there would be natural "hills". Perhaps a San Francisco landscape, but in smaller scale, with steep paths. -G. K. O'Neill
~ N•..,.. ""'' .. ~ 12/fofr' of " ~L"'' 12./10/''" -Cr
I " (Memo; G. K. O'Neill) . (Supplementary to "Graphics 12/10/75"and accompanying photograph and sketches} , ~ 12/16/75 On graphics for space-colony work: A request has just been received for my testimony before a Senate subcommittee in mid-January. In order to prepare for this, I have had to consider carefully the present state of. knowledge about physiological and mechanical constraints which would control the design of early space habitats. As I have emphasized, the details of habitat design are subsidiary to the key elements of the space-colony concept, but the many requests for information about graphics at this time force some tentative choices ·to be made, subject of course to later modification. The present best guesses obtained from senior ·people in physiology and shield design are that· with selection ( i.e. non-random population) similar to the selection which leads to the workforce on the Alaskan p~peline or _other demanding construction jobs a rotation up to 2 to 3 rpm would be possible. It is guessed ( at this point no solid information exists) that some time daily at a g-level of at least 1/3 earth-normal is necessary, to avoid bone calcium loss. For· economic and logistic reasons, it would be optimum to have an initial, lightweight construction_ station, unshielded against cosmic rays, which could build a habitat (longterm) in such a way that both habitat and construction station would end up shielded. There may be many solutions to this problem; the only one that I have seen so far that satisfies all these conditions, and in addition is economical of mass and provides a final habitat geometry which is spacious, ha~ a clear image of the sun, and would be safe against chain-reaction explosive failures in the event of window breakage is shown in the attached figure. It is only a slight variant of the Bernal-sphere geometry previously discussed. In the variant, the windows occupy the regions from 15 degrees to thirty degrees from the axes. The average sunangle corresponds to about 10:00 am on a June day. All elements of the geometry not shown are identical to the earlier Bernal-sphere studies. A : The initial construction station, a sphere 225 meters in diameter. Gravity at its equator would be 0.48 earth-normal, at a rotation rate of 1.95 rpm. The.volume would be 6 million cubic meters, or enough for up to 2 million square feet of floor space at a ceiling height of 9 feet. This would house the workforce and provide all the moderate~scale working areas for initial construction. In the final state it would be tethered by light cables to the equator of the larger sphere. B : The main sphere, 236 meters in diameter, would provide earth-normal gravity at its equator at a rotation rate of 1.95 rpm. The main difference from the earlier version of the sphere is that the sun-angle would correspond to a time closer to noon, and that the regions between 45 and 60 degrees from the equator, previously occupied by windows, are now usable (high terraced planting, steep paths, etc. Vineyards?) C : These are now planar mirrors, which need not rotate. They are backed by non-rotating shielding. D : These are conical shields, non-rotating, which complete the shield coverage, preventing slant cosmic rays from hitting the land areas. They can be about 12% thinner than the other shields due to slant angle. E : As before, this is a passage of 50 meter diameter, for air circulatio. to the radiator and to agricultural areas (if the latter is desired) and as a zero-gravity corridor for personnel and equipment moving to and from the loading docks. F: This is a zero-gravity region for practical access to (A) and also for recreation (human-powered flight etc.) It could be bounded by light nylon netting cylinders of 50 meter diameter. Low-gravity pools etc. could be either as before, or in (A). is desian is only ~. ~xl06tons.
(J \ \ ' \ \ ' ' \49 \ \ · \ \ \ \ \ \ \ \ \ \ \ I \ \ ,~'' ~ '( I \ I \ I a ' ) 4'\ ' I \ \ ~ ' \ I \ \ I \ \ \I \ I · \ \ •\ I \ \ \ ' 0- ' . I \ I © - . - . . \\ / I I ' ----~-----------+-~-----~~------~r·~-~-'-,~® \:' \© - .__ ~ ··- ~ -· '· I I \ --~,tt,---~~------ . t /· -\ _t_\ . ---- - -~-- ---~__J - - - - ;J / · / / / ./ / # I I I I I " '• I ~-; ' ,' .. ;•/ \ ./ •:"'I \ /'.i / \ I v ,, I I ., , $1 f ,. f , r ' I \ ' \. · \ j \ \ i / \ \ \
RkJQdWJsaXNoZXIy MTU5NjU0Mg==