America in Space the First Decade - Spacecraft Power

8/8/2019 America in Space the First Decade - Spacecraft Power

1/26


2/26


3/26

by Will iam R. Corliss

National Aeronautics and Space Administration, Washington, D.C. 20546


4/26

With the exception of a few special-purpose Earthsatellites, the lifeblood of every manned and unmannedspacecraft is electrical power.It is the reliable flow and availability of electricalpower that allows man to extend his personal venturessafely beyond the atmosphere and keeps unmannedscientific payloads serving as useful tools for spaceexploration and applications.Electric power is essential to space communications,guidance, control, tracking, telemetry, life-supportsystems, sensors, data handling and storage, and toassure the proper functioning of countless experimentaland housekeeping systems and subsystems aboardoperating spacecraft.

sources of energy and to see how best they can beconverted to reliable spacecraft power.The research and technology of power-generatingsystems illustrates a seldom recognized goal of NASA-to assure this Nation a freedom of choice; thechoice, in this case, being that of going where we wishto go in the atmosphere or in space. Technicalcapability is the key to such freedom.

Oran W. Nicks (acting)Associate Administrator for Advanced Researchand Technology

It remains the task of the National Aeronautics andSpace Administration, since NASA's founding in 1958,to fully investigate the chemical, nuclear and solar


5/26

Table0%Contents

Who Needs Power? 1Requirements and Profiles BHow to Choose the Best Power Plant 2

Anatomy of a Space Power Plant 3The Vital Parts 3Interfaces 4

Batteries 4Fuel Cells 6Solar Cells 7RTGs 11Nuclear Fission Power Plants in Space 14


6/26


7/26

SpacecraPowerWhoNeedsPower?Requirements and ProMesModern civilization exists because abundant powerno.urishes it-especially electrical power. Take awayelectrical power and the wheels stop, as they didduring the Great Northeast Blackout of 1965.Power is even more critical on a spacecraft. Withoutelectricity, a spacecraft becomes "dark," which in thelanguage of tracking engineers means that thespacecraft has lost its radio beacon and telemetrytransmitter. Without electricity, spacecraft tell usnothing about outer space, or themselves. Spacecraftare electrical machines just as surely as a TV set.A few watts of electrical power will bring picturesof the Martian surface and measurements from theVan Allen belts; a kilowatt will sustain an astronaut.The quantity of power needed by a spacecraft dependsupon the job to be done. The more power required,the bigger and heavier the space power plant. Becauseweight and space are always coveted by spacecraftengineers for other equipment, the actual power levelon any specific spacecraft is a matter of negotiationand compromise. In the spacecraft design process,everyone has to agree upon a weight and spacebudget as well as a power budget: so much of eachcommodity for experiments, so much for communica-tions, so much for attitude control, and so on.

modern satellites and space probes usually consumeless than 100 electrical watts of power. In fact, over90% of the 700-plus spacecraft launched since 1957have operated successfully on less power than thatconsumed by an ordinary household light bulb.The less than 10% of the spacecraft requiring morethan 100 watts fall into two classes:1. The observatory-class satellites, such as the OrbitingGeophysical Observatory and the Nimbus weathersatellite, which are large and relatively sophisticatedspacecraft. These spacecraft generally draw between100 and 500 watts-a still modest requirement interrestrial terms.2. The manned spacecraft, such as those in theGemini and Apollo programs, which must supportmen. From one to four kilowatts suffice for thesespacecraft.NASA has studied more ambitious missions to theouter planets that would consume tens and evenhundreds of kilowatts-mostly for electrical propulsion.In general, however, spacecraft are misers withelectrical power when compared with the averageAmerican household.

The first satellites orbited in late 1957 and 1958 Spacecraft do not draw power at a steady rate likeneeded no more than a watt or two to radio their a light bulb or radio. Refrigerators and washingsensor readings back to Earth. Today's satellites and machines are better examples because they operateprobes need more power because they make many in cycles or sequences of events, each event demand-more measurements, because they are often much ing a different quantity of power. The same is truefarther away from Earth, because they are bigger and on spacecraft. A typical cycle for a scientific satellitemore complex, and because some must sustain is illustrated by its record-and-readout operation.astronauts. Despite their higher metabolic rates,

1 Mariner VI, an instrumented spacecraft designed topass within 2000 miles of the surface of Mars to nvestigateth e planet's physical chara cteristics and the possibility ofextraterrestrial life.


8/26

Most of the time a scientific satellite's telemetrytransmitter is silent while its instruments fill up thespacecraft's tape recorder with data. As the satellitepasses over a ground station, it receives a commandto readout its tape recorder. The telemetry trans-mitter switches on and delivers a burst of tape recorderdata lasting several minutes. More electrical energyis drawn during readout than during the recordingmode. The scientific satellite power prode (powerplotted versus time) thus shows peaks and valleys.Other spacecraft functions add to the complexity ofthe power profde. The power plant designer has tofind extra power somewhere for these spurts ofactivity.How to Choose the Best Power PlantOnce a spacecraft designer knows approximately howmuch power he wants and how much weight andspace he will be allowed for the power plant, he canbegin to look at other factors that idu en ce powerplant performance. One of the most influential ispower plant lifetime. Obviously, he wants the powerplant to provide power reliably during the plannedlife of the spacecraft-usually six months to a yearfor unmanned craft. However, the best engineer inthe world cannot guarantee that a specific power plantwill survive for, say, a year in space-or even fora single day, for that matter. What he can state withsome assurance is that a given power plant will havea probability of, say, 95% of lasting a year in space.The 95% figure is called the reliability of thepowerplant.

An engineer could build a power plant that wouldlast ten years with a 99% probability; but it wouldbe a monster and would cost a great deal. Thethree factors of cost, weight, and reliability (lifetime)are mutually dependent; changing one changes theothers.NASA has sponsored considerable research aimed atreducing power plant weight and increasing reliability.Better power plant equipment is always reaching themarket. For example, one could hand pick the bestsolar cells coming off a production line and fabricatea solar cell power supply that would be significantlylighter than one made from run-of-the-mill cells.The best products cost more money, though, justas they do at the grocery store. The constructionof a light weight, low cost, reliable space power plantis a continual challenge to the engineer.

2 Typical power profile for a scientific satellite, showing theperiodic requirements for high bursts of power.

Data Mode= 9.22Min.Data Mode-Record Calibration Mode= 0.64Min.Low Power Transmitter OnI I Calibration Mode-Playback

ata Mode-Playback Calibration Mode40AU)5m5 30

(U

a 20

10

20 40 60 80 100 -120 140 160 180 200 220 240Time (Orbit-Minutes)


9/26

3 Block diagram of a space power plant, showing typicalcomponents. Energy source generatesP watts; energyconverter has an efficiency of e.

The Vital PartsWhy not put a few dry cells on a satellite and send itoff? Unfortunately, satellites are not as simple asChristmas toys. Although spacecraft usually consumeonly a few watts, they must operate for many months-far longer than dry cells could sustain them.Batteries sufIice for a few short-lived satellites andsounding rockets; but better sources of energy areneeded for most space missions. This observationintroduces the heart of the space power plant: theenergy source, which is the first of four importantsections of the typical space power plant.In the early days of space technology, engineers lookedat all energy sources from clock springs to nuclearreactors. In the decade that has passed, only threeenergy Eources have proven good enough for actualuse on spacecraft:1. Chemical energy sources: both batteries and fuelcells are important in spacecraft power generation.In these energy sources, part of the chemical energyin fuels and oxidizers is converted into electricitythrough chemical reactions, such as combustion.2. Solar energy sources: solar cells are the basis formost space power plants. Here, the energy in theSun's photons is converted directly into electricity.In principle, sunlight could be converted into heatfirst and then into electricity, but high performancespace power plants using solar heat have not beenbuilt yet for operational use.3. Nuclear energy sources: the energy created by thedisintegration (decay) of radioisotopes or the fissionof uranium can be converted into heat and then toelectricity. Nuclear space power plants have seen onlylimited use in space to date.Electricity is generated directly in batteries, fuel cells,and solar cells without the intervention of theturbines and generators familiar in terrestrial electricalpower plants. However, if heat is the basic productof the energy source, as in radioisotope energysources, the heat must be converted into electricity viathermoelectric elements, a turbogenerator, or someother thermal energy conversion section. This secondpart of the space power plant is called the powerconversion section. It is present primarily in thenuclear space power plants.

1 (Watts)

11-e) P (Watts)

Waste Heat Radiated ToSpace Environment

4Load

Whenever one form of energy is converted intoanother, some energy is inevitably wasted. Thiswaste energy usually appears in the form of heat. Ifthe waste heat is not rejected from the power plant,temperatures will quickly rise to catastrophic levels,just as they do in an automobile engine that has lostits water coolant. The third critical part of anyspace power plant is thus a heat rejection section..The fourth and final cornerstone of the power plantfoundation gets little publicity because it is takenfor granted. This is the power conditioning section.Power conditioning means tailoring the power tomeet the needs of the consumer. We require steady,reliable power in our homes; the spacecraft is nodifferent. The power conditioning section must createand maintain a steady voltage on the spacecraft busbars if all parts of the spacecraft are to work well.Frequently, power conditioning involves changing


10/26

low voltage direct current (DC) to high voltage DC,say, 28 volts, a rather common bus bar voltage onNASA spacecraft. The same sort of problem isfaced in terrestrial electrical engineering as millionsof transformers on utility poles bear witness.InterfacesA space power plant has interfaces with all the otherspacecraft subsystems; that is, it can affect them andvice versa. A sort of Golden Rule on spacecraftstates that one component should not infringe uponthe rights of another. Radiation from a nuclearpower plant should not, by way of illustration, causeGeiger counters to start discharging in an experimentdesigned to measure Van Allen belt radiation. Norshould waste heat from a power plant be allowed todamage adjacent temperature-sensitive equipment.On the positive side, a power plant can sometimeslend a helping hand across spacecraft interfaces. If thepower plant produces considerable waste heat, it canbe used to warm cold-sensitive equipment elsewhereon the spacecraft. Spacecraft heating is particularlydesirable on missions striking far out into the solarsystem away from the Sun's warmth or on instrumentpackages left by astronauts on the cold, dark sideof the Moon.Some other interfaces are important in power plantdesign. The large wing-like solar-cell panels sotypical of modern spacecraft must not interfere withthe fields of view of scientific instruments, such asspectroscopes analyzing the Sun's radiation. Moreobscure is the magnetic interface. Coils of wire inthe power plant generate magnetic fields whenelectricity flows through them. Many scientific space-craft carry extremely sensitive magnetometers intheir payloads to measure the very weak magneticfields in outer space. The interference is so severe thatthe magnetometer is often located on the end of aspecial boom far away from the power plant.

BatteriesBatteries depend upon chemical fuels for energy.There is no need to unfold solar-cell paddles or pointthe spacecraft toward the Sun. Chemicals keep onreleasing their energy whether the spacecraft is insunlight or in the Moon's shadow. They keep onreleasing energy as long as they last; and herein lies therub. Chemicals for all their good points are notconcentrated sources of energy. You have to keep

filling up the gas tank, and there are no gasoline stationsin outer space; all fuel has to be onboard at the launchpad. Nevertheless, the chemical battery has foundnumerous applications in space power plants.One hundred fifty years ago, the battery was the onlycontinuous source of electricity. A battery is also anenergy reservoir; that is, some batteries can storeenergy, release it on demand, store energy again, andrelease it again. When a battery is utilized as acontinuous source of energy and is not rechargeable, itis called a primary battery. When it is used as anenergy reservoir and is rechargeable, it is termed asecondary battery. A few spacecraft and many sound-ing rockets have employed primary batteries. Mostsatellites, however, and all space probes require anenergy reservoir to help meet their uneven powerprofile demands and to store energy for use when thespacecraft cannot extract power from sunlight.Secondary batteries are integral parts of most solar-cell power plants because they can store up the extrapower generated by the solar cells while they are insunlight and then release it when the spacecraftswings into the Earth's shadow. Together the solarcell and battery can generate whatever electricalpower is required by the power profile for a longperiod of time-something neither could do separately.A battery consists of two electrodes of dissimilarmaterials separated by an electrolyte. When anelectrical load is connected across the terminals, ionsmigrate through the electrolyte, carrying negativecharges in one direction and positive charges in theother. Some typical battery chemical reactions arepresented in the table below.When NASA first began using batteries in its space-craft, it found that the usual commercial products wereinadequate. They gave off too much gas duringoperation; they leaked excessively under zero-gconditions; they could not survive very many charge-discharge cycles. NASA had to invest a great dealof research and development effort before it hadlightweight batteries that functioned well in space.


11/26

Characteristics of Spacecraft Batteries

Chemical Reaction *Nickel-cadmium 2NiOOH+Cd+2H2(h 10-25 High cycling life,2Ni(OH),+Cd(OH), neglig ible gassing,can be overcharged

Good for high tem-perature operation.Always used as a

*These reactions are rather idealized; cell chemistry is not nearly so clear cut.t When used as primary batteries.

Explorer I, the first U.S. satellite, carried mercurybatteries for its power source. Because of ExplorerI's very modest power requirements, the batterieswere able to keep the small transmitter functioningfor almost four months before they went dead. IfExplorer I were designed today, a solar cell-batterycombination would probably be used; in 1958 thetried-and-true battery was considered a better riskthan the newly developed solar cells. (Historicallyspeaking, sounding rockets had been using batteriesfor years before Explorer I.) As long as a space

4 A typical silver-zinc battery employed in space powerplants. Space batteries must be carefully sealed for operationin a vacuum and under zero-g conditions.


12/26

mission lasts less than a month or so and po.werrequirements are modest (under 100 watts), batteriesare the best choice from a weight and cost standpoint.Batteries alone have powered several of NASA'searly scientific satellites. For those with short missions,such as Explorer X and Vanguards I1 and 111, batterieswere the obvious choice. There are, moreover, afew missions where solar cells interfere with thescientific experiments. In making direct measurementsof the upper atmosphere, for example, solar-cellpaddles would distort the flow of the thin air aroundthe high-speed satellite. Some satellites in NASA'sAtmosphere Explorer series have been built inspherical shape to simplify the calculation of the dragforces due to the tiny bit of air remaining at satellitealtitudes. Solar-cell paddles would have complicatedthe computations immensely. Explorers XIII, XVI,and XXIII, in the NASA Micrometeoroid Explorerseries, also employed batteries rather than solar cells.Here, scientists feared that solar cells would havegot in the way of the micrometeoroid detectors thatcovered the external skins of the satellites.

FuelCellsThe fuel cell is a close relative of the battery. Themajor difference between the two is that the fuel cellis supplied with fuel and oxidizer from externaltanks and rejects the reaction products, while thebattery must make do with the chemicals sealed intoit during manufacture.The hydrogen-oxygen cell is typical of most fuelcells. In essence, it combines hydrogen and-oxygen toform water, heat, and electricity. The fuel andoxidizer are supplied continuously and the excess water,the combustion product, is drawn off as it is formed,creating a long-lived battery.In the hydrogen-oxygen fuel cell, the reacting gasesare pumped into porous metal electrodes underpressure. At the electrode surfaces they combine tocreate water. At the anode, the hydrogen electrode,electrons are freed and travel through the externalelectrical load. At the cathode, these electrons arereturned to the fuel cell. Power will be generatedfor the load as long as the external fuel/oxidizersupply lasts.

5 Schematic showing the operation of a hydrogen-oxygenfuel cel l. The hydrogen and oxygen gases enter throughporous electrodes.The manned orbital missions in NASA's Mercury andGemini programs lasted between an hour and a halfand two weeks. Batteries would seem ideally suitedto these short applications; but, in actuality, fuelcells were selected for some of the longer Geminiflights. Why? The main difference between a longGemini mission and a short scientific satellitemission is not time but power level, or, equivalently,total energy. Ten watts may be sufficient for asmall scientific satellite, but the Mercury and Geminispacecraft needed from a few hundred to over athousand watts. In terms of total energy required(watt-hours) rather than power (watts), a mannedorbital mission is perhaps more than 100times more demanding than a short-lived unmannedsatellite.Batteries are really energy packages. In a one-poundmercury battery, we have, say, 30 watt-hours ofenergy; in a two-pound battery, 60 watt-hours, andso on. Much of the weight is actually in the packageitself, the heavy electrodes, and the electrolyte.Weight would be saved if dozens of energy packages(batteries) could be replaced by a continuouslyfueled battery, with only one set of electrodes, onecase, etc.; in short, a fuel cell. Whenever the totalenergy requirements exceed approximately 10,000watt-hours, fuel cells are usually lighter than batteries.For this reason, NASA has developed fuel cells foruse in manned spacecraft (including the Apollocraft) and the 30-day Biosatellite flight.


13/26

6 The Gemini fuel cell. Hydrogen and oxygen gases were"burned" in this fuel cel l to produce electr ici ty. Thecombustion product was water.

When one looks at practical fuel cells, one finds alot of valves and plumbing draped around the powerplant. In this sense, fuel cells are more complexthan the simple battery with its two-terminal case.In the Gemini hydrogen-oxygen fuel cell, for example,the pressures of the two feed gases must be carefullyregulated before they enter the fuel cell. Between thegas tanks and the fuel cell are valves, gauges, meters.and, of course, pipes. Fuel cells also generateoxidation combustion products, which must be drawnoff and stored. However, the water from hydrogen-oxygen cells is a valuable commodity on mannedmissions; on Apollo missions it is used for drinkingand food preparation. Fuel cells are not as simpleas batteries, but the weight they save overcomesthis objection.

SolarCells

The Sun's power flux at the Earth's orbit is a steady130 watts per square foot. With this power density,why do spacecraft have to spread large, solar-cell-covered wings to capture enough sunlight to generatea few tens of watts? Part of the answer lies in thelow power conversion efficiency of solar cells, whichis only about 10% in average cells. As a consequence,90% of the power in sunlight is not used; andsolar-cell areas must be at least ten times thoseexpected on the basis of power flux alone.Another problem with solar energy is its directionality;in other words, objects cast shadows. One shadowcaster is the Earth itself. Satellites may spend up tohalf their time in the Earth's shadow cone. Therefore,they must carry extra solar cells which generatesurplus power and store it in batteries for use duringthe long shadow periods. Shadows cause problems onthe Moon, too. Even when spacecraft are in fullsunlight, some solar cells may not be pointing directly


14/26

at the Sun-a situation that again lowers efficiency.Further, if the spacecraft is spinning, as many do forpurpose of stabilization, the solar cells swing in andout of the sunlight and will not be effective all thetime. Planetary probes have to cope with the attenua-tion of sunlight with distance. For example, asolar-cell power plant generating 100 watts as it leavesthe vicinity of the Earth will produce only 45 wattswhen it arrives at Mars.Despite these d~ cu l t i e s , olar cells are chosen foralmost every satellite and space probe. It is oftensaid that solar power plants are dominant on space-craft because sunlight is free-no need to carry energysources along. Solar power isn't really free becauseengineers have to work hard to convert sunlight intoelectricity efficiently. Because of this developmentwork-much of it NASA-sponsored-solar-cell powerplants are cheaper, more reliable, and lighter inweight than chemical and nuclear power plants inmost space applications.Let us look at solar cells in more detail. They arelittle thin wafers of silicon, black and shiny inappearance, and 1 x 2 or 2 x 2 cm in size (the twostandard sizes). Solar cells are sometimes calledsolar batteries because when sunlight (or almost anykind of light) shines on the silicon surfaces, voltagesare developed between the tops and bottoms of thewafers. When a load is connected, electrical powerwill flow into it as long as the sunlight persists.Most solar cells are made from silicon, an elementmuch like carbon in its chemical affinities (there arefour electrons in the outer shell of each atom). Butpure silicon is not used in making solar cells. A fewatoms of boron are distributed throughout thecrystal structure. The boron atoms act as electronacceptors; they have only three electrons in theirouter shells rather than four like the silicon atoms.The boron atoms try to take on extra electrons sothat they will have four in their outer shells like thesurrounding silicon atoms. The spots where freeelectrons will be accepted are called holes or vacancies.To complete the construction of a solar cell, one sideof the silicon wafer is treated with phosphorus to adepth of a micron or two." Phosphorus in contrast

* A micron is a millionth of a meter. Other semiconductor materials,such as germanium, are also used to make solar cells.

to boron is an electron donor, having five rather thanfour outer electrons. Boron and phosphorus arecalled p-type and n-type materials (for positive andnegative), respectively. Together, the two typesof silicon make a sandwich. The thin region dividingthe two silicon layers containing the different kindsof impurities is called a p-n junction.The p-n junction is where the action is. The excesselectrons on the phosphorus side and the holes onthe boron side set up an electrostatic field betweenthem. In fact, the holes act just like positive electriccharges. If carriers of electric current-free electronsand mobile holes-are created in the vicinity of thejunction, they will flow under the influence of theelectrostatic field that has been created by the boron andphosphorus impurity atoms. The absorption of asolar photon in the vicinity of the junction usuallycreates an electron-hole pair when the photon knocksone of the outer electrons off an atom in the crystallattice. These pairs attract each other and wouldquickly recombine if the established electrostatic fielddid not whisk the electrons away to an externalload. Eventually, electrons are returned from theload where they have done useful work as part of thesolar-cell electric current; then they recombine withthe holes.The typical solar cell has an area of about a halfsquare inch and converts roughly 10% of the Sun'senergy into electricity. Each cell, therefore, generatesonly a few hundredths of a watt. A space powerplant of any consequence employs thousands of thelittle rectangles. By connecting a string of severaldozen together in series, a bus bar voltage of 28volts can be attained. More strings are connected inparallel to provide the desired current.Solar cells present a challenge to both the spacecraftdesigner and the power plant engineer, who mustjointly find many square feet of area for the cells.Two methods of mounting solar cells are popular:shingle-mounting and flat-mounting. Shingling is amost accurate description; the cell edges are lappedand the shingled strips attached to a substrate orfoundation with an adhesive. Because the substratedoes not provide enough mechanical support, thestrips of cells are next attached to a lightweight supportof plastic or metal. Flat-mounted cells do not overlapand, if a defective cell is discovered, it can be replacedeasily. Adjacent flat-mounted cells are usuallyconnected in parallel by metal bus bars running along


15/26

Solar Cell Solder Adhesive Optical Coatina

Adhesive

oneycomb Core

7 Shingled solar cells bonded to an aluminu m honeycombsupport.

two opposite edges. Dozens, sometimes hundreds,of cells are shingle-mounted or flat-mounted intosubassemblies called modules. 8 Explorer XXX I I , an Atmosphere Explorer, showing body-mounted solar cells.For attaching solar-cell modules to the spacecraftproper, three approaches are in vogue today: body-mounting, paddle-mounting, and panel-mounting.Body-mounted cells are cemented either directly to thespacecraft skin or to light metallic sheets that arethen attached to the spacecraft. Body-mounting isfavored for small, spin-stabilized satellites. Patches ofsolar cells are body-mounted symmetrically on space-craft facets around the spin axis so that the overallpower level will remain relatively constant as thesatellite spins. Because some body-mounted cells willalways be in the shadow of the spacecraft at anygiven instant, the overall efficiency of a body-mountedsolar-cell power plant may be only 1-2%, eventhough individual cells can attain 10%.Medium-sized spacecraft, such as NASA's Inter-planetary Monitoring Platform (IMP) series, cannotoffer enough body area to meet the spacecraft's powerdemands. Consequently, NASA engineers haveresorted to the familiar wing-like solar-cell paddles.Solar-cell paddles are lightweight structures with cellsbonded to both faces. During launch, the paddlesare folded up within the launch vehicle shroud. Oncein space, the shroud is blown off by a small charge ofexplosive and the paddles are deployed-often bythe centrifugal force created by the spacecraft's spin-and locked into position. The sole function of the


16/26

0.005 - m Expanded SilverMesh Conductor0.0038 - m Epox y ImpFiber Glass Insulator

Aluminum Honevcomb

Solar Cells & Covers

at ed

k// 0 .64- cm ~e l l s ,~0 .0018- CMWall

9 Details of construction for the IMP solar paddles.paddles is to add to the surface area of the space-craft. with paddles, too, a large fraction of the solar 10 The scientific satellite Explorer XVIII, showing typicalsolar cell paddles in deployed position. Body of satellite is 28cells will still be in the shade, and the overall inches in diamete r. The ball on the end of the boom containsefficiency of the power plant will be only 1-2%. a magnetometer.


17/26

11 The OGO solar panels shown in their folded position.Solar-cell power plants are tailor-made for eachspacecraft; there is no typical model. The OGO-oriented solar panels will be described in some detailto illustrate the engineering behind one particularspace power plant. The OGO power requirement wassome 500 watts, considerably more than that of mostspacecraft. OGO solar cells were the standardl x 2cm variety, with 0.15-cm glass covers. Thecells were assembled in modules of 112 ( 7 x 16 ) cellsand bonded to beryllium plates. One hundred fortymodules were attached to each of two panels. Thetotal number of cells employed was over 31,000.Modules were interconnected to provide a bus barvoltage of 28 volts.Batteries have almost inevitably accompanied solarcells into space. Secondary batteries act as energyaccumulators that help the power plant meet thepeaks and valleys in the desired power profile. Whenthe spacecraft moves into the shadow of Earth, Moon,or planet, the batteries become the sole source ofelectrical power. On OGO, two silver-cadmium batterieswere installed. Each contained 24 cells and weighedabout 24 pounds. When the solar cells (but not thepower-conditioning equipment) were included, theOGO power plant weighed approximately 174 pounds.This figure corresponds to about 3 watts per pound

at the beginning of satellite life. Near the end ofOGO's planned life, the solar cells had been degradedby space radiation and the abrasion of micrometeoroidsto the point where the power plant produced onlyabout 2 watts per pound.

An RTG is a radioisotope thermoelectric generator;the need for the shorter name is obvious. Radio-isotopes find application in outer space because, on aweight basis, radioisotope fuel releases roughly 1000times more energy than the best chemical fuels, andthis energy production does not depend upon the Sun.Radioisotopes generate heat when unstable nucleidisintegrate or decay into more stable nuclei.Plutonium-238-the most often-used nuclear fuelin space-decays into uranium-234 by emitting analpha particle (a doubly ionized helium atom).The kinetic energies of the alpha particle and recoilingnucleus are quickly converted into thermal energy.Radioisotopes, therefore, are basically heat sources.Although radioisotopes are very concentrated sourcesof thermal energy, they have the unsettling anduncontrollable property of disappearing exponentiallywith time. Each species of radioisotope has a halflife,which is that length of time in which half thenuclei in any given sample will decay and yield uptheir potential energies. For example, a gram ofplutonium-238 will be half gone in 89 years (thehalf life of plutonium-238) and three-quarters gone in178 years. With an 89-year half life, this decreasein thermal power level is a minor problem on spacemissions lasting less than 10 years, which meansalmost all space missions. But with a radioisotopefuel such as cerium-144, which has a half life ofonly 290 days, the loss of thermal power during aspace mission can be serious.The half-life problem, however, is not as limiting asthe high cost and scarcity of good radioisotope fuel.These factors have restricted the use of radioisotopesin space power plants to a handful of missions.The U.S. Navy has powered several of its Transitnavigation satellites with RTGs varying in power levelfrom 2 to 25 watts. NASA has also installed anRTG called SNAP-19 * on one of its Nimbus

* SNAP = Systems for Nuclear Auxiliary Power, a joint NASA-AECProgram. -


18/26

weather satellites. Speaking generally, RTGs arecompetitive with solar-cell power plants on a weightbasis, but they are more costly due to their expensivefuel. For this reason, RTGs find applicationwherever, for some reason, solar cells cannot do thejob well. Potential RTG applications include satellitesthat must operate in the Van Allen belts, lunarprobes that must generate power during the longlunar night, probes to Jupiter and other planets farfrom the Sun, and probes that go so close to theSun that solar cells would become overheated andcease to operate efficiently.In the center of the typical RTG, there is a thickcylindrical fuel capsule, which serves as the heatsource. Surrounding the fuel capsule are thermoelectricenergy converters that change some of the heat flowingoutward from the fuel capsule into electricity. Theheat that escapes conversion into electricity (usuallysome 95% of the radioisotope-generated heat) nextflows out to the shell of the RTG, where it is radiatedaway. To understand the RTG better, we needexplanations of the thermoelectric effect and theprocess of thermal radiation.The thermoelectric element is a close relative of thethermocouple, the first device used by man to convertheat into electricity directly without moving parts.12. Idealized RTG, showing heat flowing from fuel capsulethrough thermoelectric converters to the outer shell, wherewaste heat is radiated away.

at Flow

ThermalInsulation

If one takes two dissimilar metal wires, joins them,and heats the junction, one can measure a voltageacross the cool ends of the wires. This is the thermo-electric efJect. In the thermocouple, the magnitude ofthe voltage measured is a measure of the temperaturedifferential . The key to the practical produc-tion of electrical power from the thermoelectriceffect is the substitution of semiconductor materialsfor the metal wires of the thermocouple. For onewire, a p-type semiconductor is used; for the other,an n-type semiconductor, just as in the solar cell,which converts solar radiation instead of heat directlyinto electricity.If the semiconductor materials, say, lead telluride,are fashioned into squat cylinders and joined thermallyand electrically to a metal plate, a miniature thermo-electric generator is created. If the plate-called thehot junction-is heated, the electrons and holes inthe semiconductors will both be driven toward thecold ends of the cylinders. An electrical load can beconnected acrcss the cold ends and electrical powerwill be delivered to it as electrons flow through itto combine with the holes in the p-type "leg" of thethermoelectric couple.

13 When the hot junction of a fhermoelectric couple isheated, electrons and holes are driven tow ard the cold ends ofthe semiconduc tor cylinders. Electrons flow through the loadand neutralize the positive holes.In the RTG, the unconverted heat or waste heat mustbe disposed of or the power plant will soon melt.The only practical way to get rid of waste heat inouter space is to radiate it away to cold space. The


19/26

Stefan-Boltzmann law, which describes thermalradiation-whether it be from a hot flatiron or anRTG-states that the rate at which heat will beradiated is proportional to the difference between thefourth powers of the absolute temperatures of theradiating surface and the object absorbing the radiation.In the RTG case, the hot temperature is that of theRTG casing or radiator fins, while the low temperatureis that of space itself. The rate at which heat isradiated from the RTG is also proportional to thearea of the radiating surface and its emissivity; that is,the surface's ability to emit thermal radiation,What is the "temperature of space" referred to in theStefan-Boltzmann law? If the RTG is located farout in interstellar space, the temperature of spacewould be only a few degrees above absolute zerobecause only the weak heat fluxes from the distantstars would impinge upon it. But if the RTG isexposed to the hot Sun and warm Earth, as it is insatellite orbit, the temperature of space is close toroom temperature. Fortunately, the temperature ofthe RTG surface needs to be only a few tens ofdegrees above the temperature of space to radiatewaste heat readily, because the difference between thefourth powers of two numbers increases much morerapidly than the difference of first powers.

14 Closeup of two SNAP-19's mounted on the Nimbusweather satellite. Note the large fins needed to radiate awaywaste heat. SNAP-19 generated about 25 wa tts of electricpower and weighed about 30 pounds.15 Artist's concep t of the SNAP-27 radioisotope power plantin use w ith an Apollo experime nt on the Moon's surface.An RT G could operate successfully during the lunar night,powering the helical telemetery antenna just in back of it.

structure. Because RTGs are generally low-powereddevices, the rejection of waste heat is not nearly theproblem that it is with the large nuclear space powerplants employing nuclear reactors as heat sources.

RTGs are simple, with no moving parts; they generatetheir electrical power in any environment, since therate of radioactive decay is not affected by heat,sunlight, Van Allen belt radiation, or any otherenvironmental factor. These advantages have ledNASA and the AEC to develop two RTGs, SNAP-19and SNAP-27, for space use. Two SNAP-19 unitswere launched on Nimbus B in May 1968, but thesatellite failed to attain orbit. Each of the SNAP-19units was fueled with plutonium-238, weighed 30pounds, and would have provided 25 watts in spaceto augment the solar-cell power supply of NimbusB. Each RTG was 11 inches long and 22 inches indiameter (including fins) ; hey were mountedaxially on top of one another. The lifetime expectedfor SNAP-19 was about 5 years. SNAP-27 is

If an RTG generates more than about 5 electricalwatts, there will be enough waste heat to require theaddition of radiator fins to the outer surface of the


20/26

intended for use during and after manned landingson the Moon. It will power instruments and telemetrytransmitters during the cold lunar night. SNAP-27also employs plutonium-238 fuel, but it generates 65watts, while weighing only 65 pounds, a significantimprovement over SNAP-19. SNAP-27 has acylindrical envelope 18 inches long and 18 inches indiameter. Its expected lifetime is 5 yearsAs NASA undertakes the more difficult missions to theouter planets and under the clouds of Venus, RTGswill find more and more missions where sunlight isweak or nonexistent and where they are clearly superiorto solar cells.

Nuclear FissionPower Plantsin SpaceNuclear fission reactors, like radioisotope fuelcapsules, are extremely compact heat sources. Whenthe nucleus of uranium-235 is split by a neutron, thekinetic energies of the fragments quickly appear asheat energy. The amount of energy tied up in onepound of uranium-235 is equivalent to that in 1400tons of coal. The key to applying the nuclear fissionreactor to space power is the controlled release ofthis immense energy and its partial conversion intoelectricity.An important characteristic of the nuclear reactor is itsrequirement for a certain minimum mass of fissionablefuel before a heat-releasing chain reaction cancommence. A nuclear reactor power plant designedfor space use may weigh 900 pounds and generate500 watts, but it would not weigh a great deal less ifit produced only 1 watt. Because of this irreducibleminimum size, reactor space power plants will beconsigned to future missions requiring high powerlevels-usually over 10 kilowatts.Compared with the large extant supply of uranium-235,suitable radioisotope fuels for space power plantsare rare. Little wonder that space power engineersstudied nuclear fission plants early. The NASA-AECSNAP program has investigated a number of reactorspace power plants ranging from 500 watts to 1000kilowatts. One of these power plants, SNAP-lOA,has already flown in an orbital test; another, SNAP-8,is still under development.

16 SNAP-1OA. The reactor is perched on the top of the conethat forms the radiator surface. During launch, the wholepower plant was protected by a shroud.

SNAP-lOA, like the RTGs, employed thermoelectricelements to convert nuclear heat into electricity. Theheart of SNAP-1OA was its nuclear reactor, consistingof uranium-zirconium fuel elements surrounded bya neutron reflector of beryllium. When the movablereflector pieces closed in around the reactor fuelcore, a heat-producing chain reaction began. Theheat was removed by a circulating liquid sodium-potassium alloy called NaK. The NaK conveyed theheat to a conical assembly of tubes that were bondedto the hot junctions of hundreds of thermoelectricelements. The outer surface covering the coldjunctions of the thermoelectric elements formed thepower plant radiator. The continually flowing streamof NaK was propelled by an electromagnetic pump,which used magnetic force to accelerate the metallicfluid.

Na K


21/26

POWER-CONVERSION UNITElectrical Power, Minimum 550 WattsAv. Hot Junction Tem p. 51 3C (955 F)Av . Radiato r Temp. 288 C (550 F)Efficiency , Conversion 1.83%Voltage 29.8 Volts

-Sche matic of the SNAP-1OA reactor-thermoelectric power plant.

SNAP-1OA was launched on an Atlas-Agena rocketfrom Vandenberg Air Force Base, California, onApril 3, 1965, into an 800-mile polar orbit. Thepower plant generated some 500 watts during the test,which lasted 43 days until telemetry contact waslost with the satellite. SNAP-1OA weighed about960 pounds, yielding about half a watt for each pound.The overall conversion efficiency was less than 2 %.It should be emphasized, though, that SNAP- 1OAwas a test power plant, not one intended for operationaluse. It was in fact the first nuclear reactor everlaunched into space.

SNAP-8 will not employ thermoelectric elements;rather the decision was made to try dynamic conversionof heat into electricity. In dynamic conversion, heatis first converted into the motion of a working fluid,say, steam; the high velocity steam then hits aturbine, which is connected to an electrical generatorby a shaft. As the turbine turns, electricity isgenerated. This is the same basic process employedin most commercial coal- and oil-fired electricalpower plants.

Mercury will replace water in SNAP-8; mercuryvapor will drive the SNAP-8 turbine rather thanSNAP-8 is a 35-kilowatt nuclear fission power plantand is in a power range where nuclear fission poweris clearly superior to radioisotopes, chemicals, andsolar cells on a weight basis. SNAP-8 has not beenflown; it is still in the development stage.


22/26

(2) EXPANSION OF HOT WORKING FLUID

HEATADDITION

Load

Waste HeatRadiator(3) WASTE HEATREMOVAL

w(4 ) LIQUID PUMP ORGAS COMPRESSOR

18 A heat engine cycle using a working fluid. In theRankine cycle, the fluid boils in the heat source andcondenses in the radiatorsteam. Because the mercury fluid and vapor travel ina continuous circuit from the heat source, to theturbine, to the radiator, the word cycle is applied tothese types of dynamic conversion systems. Inparticular, mercury and water cycles are termedRankine cycles, after William J. M. Rankine, aScottish engineer. The Rankine cycle can be brokendown into four parts: (1 Heat addition at thesource (the reactor in SNAP-8); (2) the expansionstep, where the vapor created by boiling during heataddition expands through the turbine and turns it;(3) removal of the waste heat at the radiator, whichcondenses the vapor to a liquid; and (4 ) repressur-ization of the fluid by a pump. The liquid thenreturns to the heat source for reheating. Of course,all four steps are going on all the time but in differentportions of the fluidivapor circuit.

With SNAP-8, the problem of waste heat disposalbecomes more critical. The amount of waste heat willbe five to ten times the 35 kilowatts of electricitygenerated. Simple RTG-type finned radiators aregrossly inadequate. With an operational SNAP-8power plant, the spacecraft will require wing-likeradiators that make the craft look like some strangeairplane. Large solar-cell power plants present thesame general appearance, but the objective there is tocollect solar radiation rather than to dispose ofwaste heat derived from an internal fuel supply.Some other interesting features of SNAP-8 are itsthree liquid metal circuits and single circuit oforganic fluid. The first liquid metal circuit, called theprimary loop, uses liquid NaK to carry heat from thereactor to a heat exchanger where the heat istransferred to liquid mercury, occupying the secondaryloop. The liquid mercury boils in the heat exchanger,creating the hot vapor that drives the turbine.Instead of going directly to the radiator for condensa-tion, the mercury vapor is cooled and condensed in


23/26

19 Artist's concep t of the SNAP-8 fligh t model. It is plannedfor developmen t of 30 to 60 kilowatts of electrical power.

another heat exchanger. The tertiary fluid used to When completely developed, SNAP-8 will probablycool the mercury vapor is NaK again. The organic weigh about 10,000 pounds in flight co~ltiguration.coolant loop cools various portions of the power plant Because it will generate 35 kilowatts, it will obtainwhere liquid metals would be undesirable. There are 3 watts for each pound-a much higher figure thanseveral engineering reasons for this complexity; one those mentioned for other space power plants. Overallis that mercury cannot be sent through the reactor thermal efficiency will be about 8%.directly because its affinity for neutrons would stop thefission chain reaction, NaK is used instead; anotherreason is that mercury vapor condensation in a hugeradiator under zero-g conditions is difficult, insteadfluid NaK conveys the heat of condensation to theradiator.

What missions would need SNAP-8's 35 kilowatts? The history of spacecraft design has been one of


24/26

NASA's mission planners foresee SNAP-8s being power-pinching because there has never been enoughused on large orbital space stations and on lunar bases. power for all desired spacecraft components andNASA has even begun exploratory development instruments. If spacecraft engineers had kilowattswork on components for space power plants that rather than watts at their disposal, they wouldwill generate hundreds of kilowatts; here, the applica- undoubtedly find many things to do with the extration would be on advanced missions employing power. They would welcome this k i d of affluence.electrical engines, such as ion rockets.20 Artist's concept of a Moon city, conforming to actualscientif ic projections. In upper left is a nuclear power station.

fr U. S. GOVERNMENT PRINTING OFFICE : 970 0 - 402-126


25/26

I DATEDUEI

AdditionalReadingFor titles of books and teaching aids related to thesubjects discussed in this booklet, see NASA'seducational publication E P-48, AerospaceBibliography.Information concerning other educationalpublications of the N ational Aeronautics andSpace Administration may be obtained from theEducational Programs Division, Code FEYOfficeof Public Affairs, NASA, Washington, D.C. 20546

For sale by the Superintendent of DocumentsU.S. Government P rinting Office, Washington, D.C. 20402Prtce 36 cents


26/26