Henry W. Bradhorst Jr.
Photovoltaic Technology
Solar cells have been the workhorse of the space program for nearly all missions lasting longer than a few weeks. Several components are needed for reliable power production from solar cells. Solar cells must be interconnected to provide the requisite voltage and current levels. This matrix must be supported on a substrate such as aluminum honeycomb or a plastic like Kapton. The individual cells also must be covered to provide protection against the electrons and protons found in the Earth's radiation belt and in ejecta from the Sun. Finally, some sort of deployment or erection mechanism must be supplied to extend the solar array from the spacecraft. The mass of the system is made up of these components, along with the power management and distribution system and the storage system needed to provide power during the dark phase.
Currently silicon solar cells are the prime power source for satellite use. Maximum individual efficiency is about 14 percent in volume production of 200-1000 kW. Cell size ranges from 2 by 4 cm to 8 by 8 cm, and the cells cost about $100 per waft. When these cells are mounted in an array, the overall power produced is about 100 W/m2. The largest solar array built to date was that for Skylab and the Apollo Telescope Mount (ATM), with a total power of roughly 20 kW (fig. 8) (Skylab in low Earth). In low Earth orbit, this array should have produced a bus power of 7.5 kW. (Charging efficiency and the cycle of a 60minute day followed by a 40-minute night reduces the average power.) Because one fourth of the array was lost during launch, the total power on orbit was reduced accordingly. The specific power (wafts of electricity produced per kilogram of array mass) of these rigid panels was 10-15 W/kg. When combined with the nickel-cadmium electrochemical energy storage system, the total solar power system had a specific power of approximately 6 W/kg. Silicon arrays also powered the first Apollo lunar surface experiments package (ALSEP) on the Moon.
Present rigid solar arrays, typified by the Tracking and Data Relay Satellite (TDRS) in geosynchronous orbit, have a specific power of 25 W/kg and a cost of about $750/W. Total power is 2.7 kW, which is typical of a communications satellite (see fig. 9) [Tracking and Data Relay Satellite]. A lightweight silicon solar array with a Kapton substrate was tested on the Shuttle in 1984. This array had a specific power of 66 W/kg and was sized to produce 12 kW of power, although only enough cells to produce about 200 W were actually put in place. This array was 102 feet long and 13 feet wide.
Advances expected in the near future include the lightweight, 50-micrometer-thick silicon solar cell blanket. These cells are one-fourth the thickness of conventional cells. The specific power goal for these lightweight arrays is 300 W/kg. These cells and arrays are aimed at applications where mass is critical, such as uses in geosynchronous orbit and exploration of the Moon and the solar system. These cells are also more resistant to the damaging effects of space radiation than thicker silicon solar cells and thus promise longer life in such orbits.
Gallium arsenide (GaAs) solar cells (fig.10) [Structure of solar cell] are being developed as an alternative to silicon cells. These cells have a higher efficiency (17-21%) than silicon cells and are less sensitive to heat. Present production capability is about 10 kW/year. Current costs of GaAs cell arrays are expected to be about $1500/W, with a cost goal of $500/W. Array technology is expected to be similar to silicon cell technology. Gallium arsenide cells were used on the Moon to power the U.S.S.R. lunokhod rover (fig. 11) [lunokhod rover] . Flight of GaAs arrays is expected in the late 1980s.
An emerging technology aimed at achieving lower GaAs array cost is to use sunlight concentration elements 2 inches in diameter and 1/2 inch thick are being developed (fig. 12) [Miniature Cassegrainian Solar Concentrator]. These devices concentrate sunlight about 100 times and illuminate 5-by 5-mm GaAs cells. Because of the small size and novel design, cell operating temperature is about 85 degrees C, not much higher than the 60 degrees C temperature at which a conventional silicon cell array in low Earth orbit operates. The cost of these emerging arrays is expected to be roughly one-third the cost of silicon arrays or about $150-300/W. Alternative optical concepts, such as reflective or retractive Fresnel lenses, are also under study. Gallium arsenide arrays are expected to produce 160-180 W/m2 at a specific power of 25-40 W/kg.
They are also more radiation-resistant than silicon arrays, both inherently and because of the shielding provided by the metallic concentrator element. Furthermore, cover-glass shielding can be provided at little increase in mass. this radiation resistance permits operation in heavy radiation orbits within the Van Allen belt (fig. 13) [Van Allen Radiation Belt] and opens the door to a solar -electric-propelled orbital transfer vehicle (OTV). this technology is being explored for space station applications. It appears feasible to build such arrays in the 500-kW range (up to 1MW with advanced higher efficiency cascade cells). Such power levels enable short trip times from LEO to GEO (several trips per month), and this technology appears suitable for lunar base operation.
Ultralightweight GaAs cell technology has produced a cell only 6 micrometers thick with a 14-percent conversion efficiency and a specific power of 5 kW/kg. When coupled with lightweight array technology, such cells have applicability to GEO and lunar base operations.
An emerging cell technology is the cascade cell, made from combinations of elements from the third and fifth columns of the periodic table. Three junction cells arranged in tandem atop one another may be able to achieve 30-percent conversion efficiency at 100 times solar concentration and at 80oC. If development of these advanced cells is successful, very high power per unit area (approaching 300 W/M2) and a specific power of 75 W/kg appear feasible. These technologies may become available about 1990.
Photovoltaic systems could be used for daytime operation on the lunar surface and for power at stations in GEO or lunar orbit. The specific characteristics required depend on the application. Solar arrays up to 300 kW with silicon planar or GaAs concentrator technology appear reasonable. Ultralightweight arrays based on silicon technology should be available by 1990, with GaAs technology following a few years later.
Operation on the lunar surface adds requirements. First, dust accumulation [Lunar Dust kicked up by the Lunar Rover] on cells or optical surfaces will degrade performance and actual operating temperatures will be greater because of the nearby lunar surface. The dust and lunar environment may also affect the maximum array voltage as a result of arcing phenomena. Finally, arrays must be designed to accommodate the deep temperature cycling of the day night cycle. The most likely use of solar arrays on the lunar surface will be to power daytime-only operations because the mass of known energy storage for the 2-week lunar night is large and makes the total system less attractive than nuclear power systems.
It has been suggested that lunar material could be mined for the production of photovoltaic devices (fig. 14) [Production of Solar Cells from Lunas Material]. The production of high-capacity photovoltaics would be limited by the availability of materials and manufacturing capability in space; thus it is not considered plausible by 2010. However, the use of lunar-derived systems for energy storage should be investigated.
Solar Dynamic Technology (Solar Dynamic Power)
Solar Dynamic systems consist of a mirror that focuses on sunlight on a receiver (which may contain thermal storage) and a Carnot-cycle dynamic conversion system (with heat radiation). (See figure 15 [Sterling Cycle].) The most common conversion cycles studied are the Stirling (fig. 16), Rankine (fig. 17 [Rankine Cycle] ), and Brayton (fig.18 [Brayton Cycle]). All have cycle efficiencies in the 25 - to 35 - percent range. When research on these systems for space use was terminated in the early 1970s, a Brayton system had been tested for a total of 38,000 hours (about 5 years). Commercial low-temperature (750oF) organic Rankine systems have also operated for tens of thousands of hours. Development of Stirling cycles is proceeding under the SP-100 Program, and space station research may support Brayton and Rankine cycle work.
Critical system elements are, first, the heat receiver, especially if it includes thermal storage, and, second, lightweight precision collectors operating at 200- to 1000-times concentration. For lunar, surface operation during the day, no thermal storage is required. As in the electrochemical storage case, extensive amounts of thermal storage would be required to meet the demands of the 2-week nights. If lunar materials having proper thermal characteristics were available for storage (questionable at this time), it is possible that solar dynamic systems could provide complete power night and day. Further study is required to substantiate this possibility.
Studies on solar Brayton cycles for the LEO space station show that a mirror 21 meters in diameter could produce 80 kW, while a mirror 8.2 meters in diameter could produce 10 kW. Were these size systems to be in continuous sunlight, the comparable powers would be roughly 175 and 22 kW, with system specific powers of 13 and 10 W/kg. Because thermal storage is one-half the total system mass, eliminating such storage (for lunar day-only operation) would increase system specific power to 26 and 20 W/kg, respectively. With system improvements (mirrors, receivers, radiators), and including other Carnot-cycle engines, specific powers around 40 W/kg (with no thermal storage) are possible at operating temperatures between 1100 and 1300 K. With space station support and with long-term advanced research support, high-performance solar dynamic systems could be available by the year 2000.
These systems require that the waste heat be rejected. Thermal management (radiators, heat sinks) remains a critical technology for solar thermal dynamic systems, just as it does for nuclear power systems.
Direct Use of Solar Energy
Many industrial processes have substantial need for high quality energy. Such applications as volatilization, evaporation, and melting use thermal energy directly without an electrical intermediary (fig. 19 [Solar Concentrator System on the Lunar Surface]). The basic elements needed are lightweight mirrors and receivers that can collect, distribute, and deliver thermal energy to its point of use. Technology for direct utilization of radiation is being developed for terrestrial applications.
Energy Storage
Energy storage is required to provide power for operations during dark times. The nickel-cadmium battery has been the common energy storage companion for solar cells on satellites. Specific energy densities (energy per unit mass) of 10 Whr/kg are common at the 10- to 20-percent depths of discharge used to provide cycle life. As a rule, the energy storage subsystem is the heaviest and largest part of a solar power system. Furthermore, NiCd batteries are sensitive to overcharge; hence, each cell must be carefully controlled. This need poses additional system constraints as power system voltage increases to the 100-kilowatt level and beyond.
Individual pressure vessel (IPV) nickel-hydrogen battery systems are being developed to provide increased energy densities (fig. 20 [Individual Pressurized Vessel]). These batteries provide about 15-20 Whr/kg for GEO applications. These devices also have applicability to LEO, but they require substantial improvement in cycle life.
There are two high-capacity energy storage systems under consideration for the space station. These are the hydrogen-oxygen regenerative fuel cell (RFC) and the bipolar nickel-hydrogen battery. The former (fig. 21 [Hydrogen-Oxygen Regenerative Fule Cell]) has a specific energy density of about 20 Whr/kg and an expected cycle life of 5-7 years. Operating voltage level appears reasonably unconstrained, allowing 150 to 300 volts. This technology is suitable for lunar surface exploration and use in GEO or lunar orbit.
Technology advances may offer energy densities of 1000 Whr/kg to lunar applications. A fuel cell separates power delivered from energy stored. Power is determined by the area of the plates; energy, by the volume of the reactants. Thus, when energy densities of 1000 Whr/kg are combined with lightweight solar arrays and high voltage power management systems, the overall system promises specific powers near 500 W/kg. It should be noted, however, that the mass of a 1000Whr/kg storage system to provide 100 kW of power during lunar night would be roughly 33 600 kg.
The bipolar NiH2 technology marries battery and fuel cell technologies to the benefit of both. Chief advantages are substantially increased cycle life over IPV NiH2, easy high-voltage battery design by adding more plates, and extremely high discharge capability (20 times charging rate). Bipolar NiH2 systems appear equivalent in mass to state-of-the-art regenerative fuel cells at 100-kW capacities. However, this technology lags that of the hydrogen oxygen RFC by several years. Furthermore, substantial improvement in basic understanding and in plate and separator technology is required before these cells can even begin to approach the 1000-Whr/kg potential of the hydrogen-oxygen regenerative fuel cell.
Two additional systems appear capable of high storage densities. These are the rechargeable lithium battery and the hydrogen-halogen (Br, CI) regenerative fuel cell. Both technologies are in infant stages of development, with issues of materials, cycle life, current densities, separators, and electrolytes. With additional research emphasis, these systems could become available between 1995 and 2000. Because mass is at such a premium on the Moon, and because the energy storage system is the most massive part of a photovoltaic system that supplies continuous power, additional effort should be directed toward innovative energy storage technologies, electrochemical and other.
Flywheels are one example of mechanical energy storage (fig. 22 [Advance flywheel Energy Storage]). Although flywheels probably can store in excess of 100 Whr/kg, the overall systems are still heavy (10.Whr/kg) at present. Although these systems may be capable of long lives, this capability has not yet been demonstrated, nor have all failure modes and safety needs been identified.
Solar dynamic systems also require energy storage for operation during the dark phases of a mission. A number of concepts are being considered. Sensible heat storage (that is, heat stored by the natural heat capacity of the material) in the form of a heat sink mass is one possibliity. Another is the use of a material such as a salt which is melted during the solar phase and allowed to freeze during the dark phase, thereby releasing the heat of fusion. technology development programs are presently under way in the selection of compatible materials and in freeze-thaw phenomena in microgravity.
Whithin the timeframe of this study, it does not appear that the energy storage technology will be affected by nonterrestrial resources. A variety of candidate technologies with high energy densities have been identified (fig. 23 [Energy Storage Opportunities 1997]) and must be considered for future energy storage use in GEO and on the Moon.
Power Management and Distribution
Existing spacecraft power systems are 28 volts dc. This voltage level and type was adequate for the few-kilowatt, dedicated-load missions to date. With the nearly 100-kilowatt electrical power requirements of the space station, however, significantly higher voltage levels and a high-frequency, ac utility-type distribution system are required to deliver this power efficiently to a broad spectrum of national and international users. Compared to existing systems, a 20-kHz ac power management and distribution system provides higher efficiency, lower cost, and improved benefits. The proposed 20-kHz system is based on rapid semiconductor switching, low stored reactive energy, and cycle-by-cycle control of energy flow. This system allows the voltage and wave shapes to be tailored to meet a variety of load. requirements, improves crew safety, and provides compatibility with all types of energy sources photovoltaic, solar dynamic, electrochemical, rotating machines, and nuclear.
Voltage levels on exterior surfaces will likely be set in the 150- to 300-V range by LEO plasma interaction effects. Inside the modules, however, a single-phase, sinusoidal-waveform, 20-kHz distribution system, with a well regulated 220- or 440-V (root mean square) bus, will minimize wiring mass, transformer weight, conversion steps, and parts. Such a distribution system will provide attendant reductions in the sensing and control complexities required by a redundantly distributed power system with multiple energy sources. Component technology and microprocessor-based innovations in system autonomy will be in hand by the early 1990s to enhance the power system. Requirements pertinent to nuclear systems, such as hardening and high temperature operation, are being addressed by the SP-100 Program, under which NASA, the Department of Energy, and the Department of Defense are developing space reactor technology.
As power requirements build to the 1- to 10-megawatt level for future space and lunar base missions, however, it is likely that either the bus voltage must leap to the kilovolt level or current levels must increase with paralleling and phase control. In either case, new semiconductors and other components and more switchgear, cabling, and connectors will be required. Designs for operating in the lunar environment, where dust may provide severe environmental interactions, will be especially critical. Early research into all these types of hardware is warranted. We envision that both ac and dc equipment of various types and voltage levels will be routinely used in orbit and on planetary surfaces.
As in the previous cases, it is unlikely that nonterrestrial resources will affect power management and distribution systems by 2010. Rather, it is the power system that will enable utilization of nonterrestrial resources.
David Buden
Radioisotope Generators
Current status: Radioisotope generators use the spontaneous decay of plutonium-238 as a heat source. The energy has traditionally been converted to electricity by means of thermocouples placed next to the heat source. (See figure. 24 [Radioisotope thermoelectric Generator]). Radioisotope generators have been launched in 21 spacecraft, beginning with the successful flight of a space nuclear auxiliary power (SNAP3A) source in 1961. A summary of launches is shown in table 1 [Summary of Space Nuclear Power Sources Launched by the US 1961-1980].
The technical characteristics of thee radioisotopes generators are listed in table 2 [Radioisotope Generator Characteristics]. Their reliability and long life is demonstrated by the Pioneer satellite, which after 11 years of operation left our solar system still functioning. The recent magnificent pictures of Saturn taken from the Voyager spacecraft powered by radioisotope generators are also testimonials to the longetivity and reliability of this type of power supply. (See figure 25 [Experiments and Spacecraft Powered by the RTG's], figure 25a. [Jupiter and its moons]) and figure 25b [Saturn and Uranus].
Radioisotope thermoelectric generators (RTGs) have been used where long life, high reliability, solar independence, and operation in severe environments are critical. Economic considerations have restrained them from more general use.
Future developments: Improved versions of the RTG will have better performance. However, RTGs will probably be restricted to under 500 W. Higher power levels of maybe 5-10 kWe are possible by using dynamic converters for power conversion. A 1.3-kWe version was tested for several thousand hours before the program was terminated. A revised program to cover the 1-10 kWe range is scheduled to start in 1988. These improved versions using thermocouples and dynamic converters could be used for lunar and Mars rovers and explorations away from lunar camps and bases.
Nuclear Reactor Power Plants
Current status: The current U.S. effort to develop nuclear reactors for space is centered in a program entitled " SP-1 00," which is a joint program of the Department of Defense, the Department of Energy, and NASA. (SP-100 is not an acronym.)
The decision to proceed with the construction of a specific space nuclear power plant was made and a contractor selected in 1986. The program has completed the critical technology development and assessment phase. Activities centered around evaluating promising space reactor concepts and determining which technologies are most likely to achieve the required performance levels. The technology assessment and development phase included defining mission requirements, doing conceptual designs of possible systems, and researching and developing critical technologies.
Following screening by the SP-100 Program of over a hundred potential space nuclear power system concepts, the field was narrowed to three candidate systems which appear to meet the requirements in table 3 [SP-100 goals] without unreasonable technical risks or development time.
One concept uses a fast-spectrum, lithium-cooled, cylindrical, pin-type-fuel-element reactor with thermocouples for power conversion (fig.26 [High-temperature Reactor]) (General Electric Co. 1983). the system is made up of a 12-sided cone structure with a 17-degree cone half angle. the reactor, which is a right-circular cylinder approximately 1 meter in diameter and 1 meter high, is at the apex of the conical structure. It is controlled by 12 rotatable drums, each with a section of absorbing material and a section of reflective material to control the critically level. Control of the reactor is maintained by properly positioning the drums. The reactor outlet temperature is 1350 K.
The shield is mounted directly behind the reactor and consists of both a gamma and a neutron shield. The gamma shield consists of multiple layers of tungsten designed so as to prevent warping. The neutron shield is made up of a series of axial sections with thermal conductors between them. The thermal conductor carries the gamma- and neutron-generated heat to the shield surface, where it is radiated to space. Anticipated temperature levels are 675 K, maximum.
Thermal transport is accomplished by thermoelectrically driven electromagnetic pumps. The thermocouples for the pumps are powered by the temperature drop between the working fluid and the pump radiators. This approach assures pumping of the working fluid as long as the reactor is at temperature, and it facilitates the cool down of the reactor when power is no longer required.
The reactor's thermal interface with the heat distribution system is through a set of heat exchangers. In this way, the reactor system is self-contained, can be fabricated and tested at a remote facility, and can be mated to the power system downstream. Access panels are provided on the main body to facilitate the connection of the heat distribution system to the heat exchanger.
Thermoelectric elements for converting thermal energy to electric power are bonded to the internal surfaces of the heat rejection panels and accept heat from the source heat pipe assembly.
The heat rejection surfaces are beryllium sheets with titanium potassium heat pipes brazed to the surface to distribute and carry the heat to the deployable panels, which are needed for additional heat rejection. The deployable panels are thermally coupled through a heat-pipe-to-heat-pipe thermal joint, which is very similar to the source-heat-pipe-to-heat exchanger joint, made integral by the use of special materials that are self-brazing in orbit. To allow the deployment of the panels, a bellows-like heat pipe section is mounted at the tail end of the heat pipes on the fixed panel. Such a flexible heat pipe has been demonstrated.
The system has a wide range of flexibility. Its output can be expanded either by increasing the thermoelectric efficiency or by increasing the size and weight of the system. The potential for scaling up the system is shown in figure 27 [scaling graph] (Katucki et al. 1984).
A second approach evaluated is an in-core thermionic system with a pumped sodium-potassium eutectic coolant (GA Technologies and Martin Marietta 1983). The general arrangement of this space power system design is shown in figure 28 [In-core Thermionic Power Plant]. The design forms a conical frustum that is 5.8 m long, with major and minor diameters of 3.6 m and 0.7 m. The reactor-converter subsystem includes the reactor, the reflector/control drums, and the neutron shield. The reactor contains the thermionic fuel element (TFE) converters within a cylindrical vessel, which is completely surrounded by control drums.
The hot NaK leaves the reactor at the aft end and the cold NaK is returned to the forward end, thus minimizing differential thermal expansion in the piping. The reactor is also surrounded by an array of long, thin cylindrical reservoirs that collect and retain the fission gases generated in the reactor core during the operating life of the system.
Waste heat is removed from the primary loop through the heat exchanger. The energy is transferred through the heat-sink heat exchanger to heat pipes that form the radiating surfaces for rejection of heat to space.
Within the reactor vessel are 176 TFEs, a grid plate to support the TFEs at one end, a tungsten gamma shield, and the eutectic NaK coolant. Each TFE is welded into the flattop head of the vessel but allowed to move axially in the grid plate. Expansion is expected to be small, since the TFE sheath tubes and reactor vessel are both mada of an alloy of niobium and 1 percent zirconium and their temperatures are nearly the same.
The TFE consists of six cells connected in series with end reflectors of beryllium oxide. Boron carbide neutron absorber is placed at both ends of the fuel element to reduce the thermal neutron flux in the coolant plenums and in the gamma and neutron shields. This reduces activation of the coolant, secondary gamma ray production, and nuclear heating of the lithium hydride shield. The individual cells (see fig. 29 [In-core thermionic Converter]) are connected in series to build up voltage from the 0.4-V cell output. Electrical power is generated in the space between the tungsten emitter and the niobium collector, and the electrical current output is conducted from one cell to the next through the tungsten stem of the emitter and the tantalum transition piece. The UO2 fuel is held in place and supported during launch by a retention device designed to retract when the fuel expands upon heating. The alignment spring at the base of the emitter centers the .emitter in the collector to maintain a uniform interelectrode spacing. It also restrains the emitter against launch vibration to prevent large displacements and limit stresses in the thin stem at the other end of the emitter.
Fission gases are vented from the UO2 fuel to prevent the buildup of pressures that would cause creep deformation of the tungsten emitter and close the interelectrode space. Fission gases are kept separate from the cesium (used to reduce the space charge effect) by the ceramic-to-metal seal and the arrangement of passages through the emitter cap and transition piece.
Reactor control is provided by the rotation of the 20 cylindrical control drums surrounding the reactor. The heat transport subsystem is a single loop that includes all of the NaK plumbing aft of the reactor, the heat-sink heat exchanger, and the radiator. The 100-mm-diameter NaK lines to and from the reactor are routed inside helical grooves in the outer surface of the neutron shield and then pass along the inside surface of the radiator to connect to the heat-sink heat exchanger. The configuration of the NaK lines along the shield is helical, rather than straight, to avoid degradation of the shield performance due to neutron streaming in the pipe channels.
The helical channels in the shield are also occupied by the electrical transmission lines, which are flattened in cross section and are routed over the NaK lines to serve as meteoroid protection. Electromagnetic pumping is used to circulate the NaK during normal operation and during shutdown. Two electromagnetic pumps are provided in the cold leg of the NaK circuit: an annular linear induction pump to serve as the main pump and a parallel thermoelectromagnetic pump (with a check valve) to provide shutdown pumping capability.
The radiator contains two finned heat pipe assemblies, which form a conical frustum when the panels are assembled on the radiator structure. The heat pipes follow the slant height of the core and are deployed fore and aft of the heat-sink heat exchanger, to which they are thermally coupled. The radiator provides environmental protection for the equipment it houses.
Growth is possible by either redesigning the reactor with more TFEs or increasing the emitter temperature (see fig. 30 [Scalability of In-core Thermionic Reactor]) (Katucki et al. 1984). An upper temperature level of about 2000 K is believed to be an operational limit for the tungsten emitter. The third approach uses a Stirling engine to convert to electricity heat from a lower temperature (900 K), fuel-pin-type reactor. This design emphasizes the use of state-of-the-art fuel pins of stainless steel and UO2, with sodium as the working fluid. Such fuel pins have been developed for the breeder reactor program, with 1059 days of operation and 8.5percent burnup demonstrated.
The reactor can be similar in design to the high-temperature reactor, but it utilizes lower temperature materials. In figure 31 [Concept of Stirling Engine conversion] (General Electric Co. 1983), the reactor is constructed as a separate module from the conversion subsystem. Four Stirling engines, each rated to deliver 33 kWe, are included in the design concept to provide redundancy in case of a unit failure. Normally the engines operate at 75 percent of rated power to produce an output of 100 kWe. Each engine contains a pair of opposed-motion pistons, which operate 180 degrees out of phase. This arrangement eliminates unbalanced linear momentum. Each engine receives heat from a pumped loop connected to the reactor vessel.
An alternate arrangement would deliver the heat through an interface heat exchanger with heat pipes between the heat exchanger and the engine. Waste heat is removed from the cooler heads and delivered to a liquid-to-heat-pipe heat exchanger. The heat pipes, in turn, deliver the waste heat to the radiator where it is rejected to space. Figure 32 [Scalability of Stirling Power System Concept] provides performance curves for the Stirling system. A low temperature will meet the goal of 100 kWe. However, growth systems favor combining the Stirling engines with higher temperature reactors both to minimize mass and to reduce heat rejection surface areas.
Figure 33 [Performance Projection graph for Space Nuclear Reactor Power System] summarizes the mass and specific power projected for the 100-kWe class of power plants.
The fast-spectrum, lithium-cooled reactor with thermoelectrics (concept 1) has been selected for the ground demonstration system. Work is continuing on thermionic fuel element development and Stirling engine development for possible use in growth versions of SP-100.
Future developments: Several classes of reactor power plants will be needed in the future to provide adequate energy for lunar camps and base stations, the growth space station and Space Station 2, and electric propulsion. The 50- to 1000-kWe power plant being developed by the SP-100 Program for flight in the early to mid-1990s will meet the power requirements of the growth space station, the lunar surface day/night camp, and nuclear electric propulsion. However, the requirements and designs have been aimed at unmanned systems. these should be reviewed and modified as necessary to meet manned operational requirements. These requirements could include shielding that completely encloses the reactor, additional emphasis on shutdown heat removal and safety systems that are independent and redundant, and considerations of maintainability and disposal.
We anticipate that the early lunar camps and bases will involve the transport of a space station version of the 100-kwe-class power plant with little shielding. The power plant would be arranged to reject heat to space. People would be projected by using lunar materials for the radiation barrier.
Space Station 2, requiring 1 - 10 MWe, would need a new class of reactor plants. Major changes in reactor designs may be called for, such as higher temperatures, refuelability, and maintainability of certain components. Significant improvements in power conversion and heat rejection are also necessary. The power conversion will probably work at a higher temperature; innovative design through in core thermionics is being evaluated as an alternative. Heat rejection will need a deployable system that uses a nonarmored radiator technology. One concept, the liquid droplet radiator, is now being pursued to demonstrate technology feasibility. Other concepts include belts, balloons, and rollup heat pipes. The goal would be to package a 10-MWe power plant in a single Shuttle launch.
The power plant for Space Station can meet the requirements for a manned Mars mission (fig. 34 [Manned Mars Mission configuration]) and for a lunar orbital transfer vehicle using nuclear electric propulsion. For the advanced lunar base, the same power plant could be used. Again, lunar soil could provide shielding. However, if a mining and materials fabrication capability were in place, it could be used to fabricate a specially designed heat rejection subsystem. Doing so could produce a major savings in mass transfer from Earth. Several innovative designs are possible, such as continuous ejection and collection of fluid or solid particles.
Public Safety and the Use of Nuclear Reactors in Space
Policy and goals: The policy of the United States for all U.S. nuclear power sources used in space is to ensure that the probability of release of radioactive materials and the amounts released are such that an undue risk is not presented, considering the benefits of the mission (U.S. Department of Energy 1982). Safety criteria are specified for the design of the SP-100 space nuclear reactor power plant; safety is to be built into the design, not just added on.
The restriction of radiation exposure (DOE 1982) depends on reducing the probability of an accident that might release radioactive materials into the environment and on limiting the magnitude of such a release should one occur.
Space nuclear power applications must keep the radiation exposure of astronauts, occupational workers (e.g., ground support personnel), and members of the general public "as low as reasonably achievable" during all mission phases, normal and abnormal. According to recommended standards (U.N. General Assembly paper 1980), the maximum accumulated doses for closely involved workers and for the general population are those listed in table 4 [Table of Normal Mission Exposure LImits]. Allowable doses for astronauts are generally in the same range as those allowed for radiation workers.
The safety program is designed to protect the public against exposure to radiation levels above established standards. This can be accomplished by preventing accidental reactor criticality and by avoiding releaes of radioactive byproducts into the biosphere in sizes and concentrations that exceed the standards.
Another set of safety goals encompasses the protection of investments in facilities both on the ground and in space. These facilities must be protected both because they are national assets that would be costly to replace and because a failure would produce significant delay us in our national efforts to build the space station. Safety goals and requirements are summarized in table 5 [Saftey and Goals Table].
The safety review process: The United States requires an analysis of each space mission involving nuclear material to assess the potential radiological risk to the biosphere. The process begins when the space mission is defined and the design is conceived. The safety review process continues through launch safety analysis, approval to launch, and proper nuclear power source disposal.
The developer of the nuclear power source is responsible for performing the nuclear safety analyses for the system. Results of these safety analyses are reported at least three times during the development cycle in documents entitled Preliminary Safety Analysis Report (PSAR), Updated Safety Analysis Report (USAR), and Final Safety Analysis Report (FSAR).
The Preliminary Safety Analysis Report is issued 120 days after a design concept is selected. It contains a description of the design, a failure mode analysis, and a nuclear safety analysis. The latter two requirements are based on the safety research data for the development of heat sources, historical heat source design information, and the requirements set forth in the guidelines written by the Department of Energy (DOE). At this stage of system development, the failure mode analysis is based on the response to potential accident environment., and on design limitations established by the guidelines.
The Updated Safety Analysis Report is issued 90 days after the design is set. It is similar in form to the preliminary report. Additional requirements include a description of the mission on which the system is to be used and an update of the failure mode analysis using data from the developmental tests performed to set the design.
The Final Safety Analysis Report issued approximately 1 year before the scheduled launch and is simple in format to the earlier reports. This report provides final system mission, and safety assessment data, factoring in the results of the verification and qualification test programs. Thus, the final assessment is based on the actual mission environments.
The Interagency Nuclear Safety Review Panel (INSRP) is responsible for review of the safety analysis reports at each step of the development process. The end result of the INSRP process is the Safety Evaluation Report (SER). This report evaluates potential human exposures to radiation and the probabilities of exposure during all phases of the mission. The INSRP submits the Safety Evaluation Report to the heads of the Department of Energy, NASA, and the Department of Defense for their review. The head of the agency that wants to fly the nuclear power source must then request launch approval from the President through the Office of Science and Technology Policy. The ultimate authority for launch and use of the nuclear power source lies with the President of the United States.
Figure 35 [Us Safety Review and Launch Approval Process] shows the generalized sequence of events in this flight safety evaluation process. Because safety features are designed into U.S. nuclear power sources from the very beginning, this safety review process is actually an integral part of the overall flight system development.
