10 . SPACECRAFJ T VOL1 . 23NO ,
Droplet Radiator Systems for Spacecraft Thermal Control
Robert T. Taussig* Spectra Technology, Incorporated,^Bellevue, Washington and MattickT. A. J University of Washington, Seattle, Washington
Liquid-droplet radiators (LDRs) can be as much as 5 to 10 times lighter than advanced heat pipe radiators. They are virtually immune to puncture, stow in a small volume, and are easily deployed. Systems studies show that they optimize in the 10- to 100-MW size for liquid lithium, tin, and aluminum and at several hundred kilowatts for NaK and vacuum oils. Larger radiators can be constructed from modules of these sizes. Power systems reoptimize lighmase R th tr sLD lowe d operatfo K 0 r10 rejec t ea t hea5 o t temperature3 e b y ma d san times lighter than those using conventional radiators. Recent collector tests have revealed a splash-free regime of operation promising high collection efficiencies testR spacn si LD . e shoul feasible d b neae th rn ei future .
Nomenclature Introduction A = radiator sheet area (one side) MALL liquid droplets offe vera r y lightweigh rejeco t y ttwa Ar = radiator aspect ratio aste heat from space power systems becaus theif eo r high Bon= d numbe o f rotatinB o r g collector film surface area per unit mass. Lightweight radiators will become c = droplet liquid specific heat a necessity as the need for higher powers in space increases. C = evaporation rate function Missionundew no re consideratiosar n that will require rejec- d = separation distance between adjacent droplet tio f megawattno f wasto s e continuoua hea n o t s basisd an , streams future missions may need to reject hundreds to thousands of E3 = elliptic integral megawatts. The very feasibility of such missions is brought in- Froude= Fnumber f rotatino r g collector film doubo t t because conventional radiatoro heavo to to d e yan sar lengt= f radiatoho Lr r sheet difficult to deploy or assemble in space. drople= tm sheee t evaporation ratunir epe t area Liquid-droplet radiators (LDRs) have been investigatea y db thy = droplet sheet mass flow rate numbe authorsf ro havd 1"an 5 numbee a distinctivf ro e advan- M - total mass of droplet sheet radiator tage thin si s context: mas= f dropletso M ffligh n i s t 1) The radiating surface is invulnerable to puncture by P, Pr = radiated thermal power small particles and by directed energy weapons. r = droplet radius 2) They are lightweight because very little fluid mass is re- Re = Reynolds number of rotating collector film quired for a given radiating area and armoring against punc- drople= t shee s t thickness t requiredturno s ei . T0 = radiator inlet temperature 3) They can be deployed easily by using a mast that expands radiato= rT joutle t temperature to its full length in orbit. Te = effective mean radiator temperature 4) They can be stowed in a small volume, such as a canister ^peak = Peak thermal cycle temperature of liquid and a collapsible structure. drople= t velocitu y 5) The operatn ylona ca r gfo e perio f timdo - e oveap n a r ur = radial flow velocity of rotating collector film propriate temperature interval where evaporative losses are Ue = tip velocity of rotating collector minimal. w = droplet sheet width at droplet generator Droplet radiators are still in the conceptual stage of develop- We = relative droplet Weber number ment. Initial experiments exploring the formation of droplets, evaporate= 18d mass fraction splashing in simulated zero-G conditions, and radiative cool- e0 = intrinsic emissivity of droplet liquid ing from selected droplet materials have been carried outd an , t> €sheet = mean hemispherical emissivity of droplet sheet an investigation of droplet collector technology, described X = separation distance between droplets in a single below, has recently been completed. Preliminary systems stream analysis of the droplet radiator has accompanied these in- drople= t liquip d mass density vestigation guido st experimentae eth l endeavor modeo t d slan 2 pa = droplet sheet area density (kg/m ) the performance (e.g., total system weight, peak power, a = Stefan-Boltzmann constant radiator effectiveness) of various configurations. The systems T> Tmeashee= tn optical dept f dropleho t sheet model include majose eacth f hro radiator components, such optica= l dept70 droplef ho t shee t dropleta t generator as pumps, nozzle arrays, collectors structured an , s thae ar t radiato= rT Llifetim e needed to operate this type of radiator. A description of the angula= r velocitco f rotatinyo g collector systems model and a summary of the major conclusions of that analysis are discussed here. Presented as Paper 84-1797 at the AIAA 19th Thermophysics Con- ference, Snowmass , JunCO ,e 25-28, 1984; submitted Oct , 19841 . ; ConcepR LD e tTh revision received April 1, 1985. Copyright © American Institute of Mattick and Hertzberg have described the basic radiator Aeronautics and Astronautics, Inc., 1985. All rights reserved. 1 2 * Directo f Energo r y Technology. Member AIAA. physics of the liquid-droplet radiator. ' In their model the tFormerly Mathematical Sciences Northwest, Inc. droplets form a thin sheet with variable optical thickness, ^Research Associate Professor, Aeronautics & Astronautics. dependin droplee th n go t number density, droplet sized an , JAN.-FEB. 1986 RADIATOR SYSTEM SPACECRAFR SFO T THERMAL CONTROL 11
sheet thickness. The radiative transfer equation was solved LDR Systems Model assuming isotropic scattering and a constant temperature from preliminarA y systems mode triangulaa f o l r liquid droplet top to bottom of the sheet. The only temperature variation sheet radiator has been developed which incorporates in- considered was in the direction of the droplet velocity due to dividual models of each of the major components and radiative coolin collectore routn gth e o et . Figur showe1 sa calculates self-consistentl liquie yth d flow rate, pressured an , triangular droplet sheet radiator wit rotatinha g collectot a r component mass at each station in the system. As a conse- the focal point. quence e componenth , t sized operatinan s g conditione ar s extreme th t A dropleew limilo f o tt densit vera r yy(o thi n matched throughout the system. sheet), the droplets radiate their heat independently of each initiae Th l radiator configuratio majod an n r components other; this is the lightest-weight version of the droplet radiator showe ar Fign ni heae . 3Th .t exchange t beeno ns incorha r - but als largese oth arean i tr hig Fo .h droplet densities- op e th , porate modee th n di l sinc t dependei specifie th n so c mission tical thickness increases and radiation from the droplets near powed an r system used. the center of the sheet is occluded by neighboring droplets so e tharadiativth t e effectivenes f o individuas l droplets Droplet Sheet Model .decreases. However equivalene th , t emissivit droplee th f yo t The main assumptions underlying the droplet sheet calcula- sheet increases so that the required sheet area decreases. For tion thae sar t 1) materia loss li t primarily through evaporation optical depths much greate poina , r 5 diminishinf tha tr o o n4 g aimind an g inaccuracies radiatoe th ) 2 ; r retains s 100it f %o return s reachei s d wher e sheeeth t becomes heavier without load capacit givea r yfo n lifetime Tliquie L th (years) ) d3 d an ; real gain in emissivity. The optical depth r is defined as droplet temperatur collectoe th t ea vers i r y clos freezingo et . T=irr2S/(d2\). In the case of liquid metal droplets, the sheet The input parameter e totaar s l radiated powe r pea(o r k emissivit largee b n ry ca tha intrinsie nth c material emissivity temperature), droplet material properties (i.e., density, sur- (i.e., infraree 0. 0.1o 1th t near-infrared n 5i an d d spectrum) face tension, viscosity, specific heat, and intrinsic emissivity), because radiation from the interior droplets escapes via multi- optica le sheetth dept f ,o h sheet aspect ratio (i.e.,= ple reflections from neighboring droplets case f th oils eo n I ., width/length), radiator lifetime, aiming accuracy, droplet the droplet emissivity is already high (i.e., 0.5 to 0.85) so that radius, and number of radiators operating in parallel. If the even though radiation fro interioe mth occludeds ri sheee th , t total power radiated and number of parallel radiators are in- higa stils h lha effectiv e intrinsiemissivite th o t e c yemissivdu - put data, a separate calculation must be made of the peak droplee th itf yo t material. (inlet) temperatur e dropletsth f e evaporatioo e Th . n rate basiconcepe R Th cLD tformatio e relieth n so f uniforno m depend radiatoe th n so r temperature opticae th n o l deptd san h droplets using technolog printerst y je relate k in o d.t Uniform- sheete o th f- 2E :1 m3[ (ee= 0f)]C( 7^,7^), wher factoe eth n i r itf dropleyo t sizvelocitied ean s ensures equal cooling during square brackets represent e reabsorptioth s f evaporateo n d droplet travel. Although droplet e outeth rn e o sedgth f o e material by neighboring droplets in the sheet. The factor sheet will cool faster than inside e thosth preliminar r n eo ou , y r( 0C , 7\) increases rapidly with temperatur dependd ean n so calculations indicate that this effec modests i t . Analysie th f so the droplet material chosen.8 Each liquid can be used only electric charge induce droplete space th th n do y e sb plasm a over a specific range of peak temperatures before evaporation and by solar photon bombardment show that deflection of severely limits the lifetime of the radiator. droplet trajectories due to coulombic repulsion will be small under nearly every conceivable circumstance in near-Earth or- bits. On the other hand, atmospheric drag on smaller droplets may prevent such radiators being used for orbital altitudes of less than 250 miles. A possible deployment sequence for LDRs is shown in Fig. 2. In this sequence, a payload plus power system (e.g., a nuclear reactor) and radiator are launched in the shuttle bay. Calculations by Hertzberg and his co-workers have suggested that up to 10 M Watt of power may be included in one shuttle launch by using the SP-100 technology and a droplet radiator.6'7 Once separated fro shuttlee mth chemica,a l rocket raise payloae sth orbito dt , wher radiatoe eth r mast untwists fro s stowemit d forcabled man s snap into plac provido et e structural rigidity. This configuration show o paralletw s l triangular droplet radiators secone ;th d radiator spray- li e sth quid collected at the far end of the first radiator back to the power syste r reheatinmfo avoio gt dheava y return pipe.
Fig. 2 LDR deployment sequence.
COLLECTOR
COLLECTION TROUGH
•DRIVE MOTOR
PRESSURIZING SCOOP DROPLET GENERATOR ROTATING COLLECTOR / COLLECTION PLATE \ STRUCTURE ROTATING SPLASH GUARD SHEET Fig. 1 Triangular LDR with a rotating collector. Fig. 3 Droplet radiator systems model. 12 R.T. TAUSSIG AND A.T. MATTICK J. SPACECRAFT
The following set of equations governs the basic model of PipPumd ean p Model the LDR sheet: The LDR system shown in Fig. 3 relies on a return pipe and a pump. The pump was modeled after liquid metal centrifugal In-flight mass: pump snucleae useth n di r reactor industry. Typically, these cAr large ar e (e.g., 17,000 gal/min wild )an l handle relatively high pressures (e.g., up to 100 or 200 atm). They are driven by elec- where Ar=r0- tric motors, which constitute 80% of the total pump mass. entire Th e uni hermeticalls i t y sealed pume Th . p efficiencies Evaporation loss: 2merLA f$M- f ordee th 85%f on-desig n ro mechanicae o e .Th b n nca l portion e conventionath f o e pum% oth s assumef20 lpwa e b o dt Radiato WL/2= A r area= Q.5A: rL pump weighte conventionath f I . l motor e replacear s y b d 2 samarium-cobalt permanent magnet motors totae th , l pump Radiator length: L = ut = 2iiTcrpAT/3eaT* mass will drop significantly since these device designee b n sca d
for high input powers (tens of megawatts) and will weigh on
2
where T = T[3/(f 3 +/ +/)]* and/= 7*0/7*, the order of 0.1 Ib/kW. 0
e The total mass of the pump, pipe, and coolant filling the Mass flow rate: fnf - (4/3)rupT0A rL pipe is minimized in this model by varying the pipe diameter. For large diameters the pumping power is quite low but the The first equation defines the in-flight mass Mf in terms of fluid inventor d pipan ye mas e substantialar s r smalfo ; l the time it takes a droplet to traverse the length L of the diameter pume sth p powe pumd ran p mass become very large. radiator (droplet velocity w), and the mass flow rate given by the total power radiated in a single radiator Pr divided by the heat losr unispe tdroplete masth f so flightn s i secone Th . d Collector equation relates the evaporation rate me, radiator lifetime TL, Several features have been modele rotatine th r dfo g collec- surfacd an eradiatoe areth fractioe - f th in ao o e t r th n f /o 3 tor. This device uses centrifugal force to transfer fluid col- flight mass which evaporates over the radiator lifetime. The lected from the droplets to the periphery of the collector, factor 2 accounts for losses from both sides of the droplet where the fluid is scooped up by a stationary pipe inserted into sheet thire Th .d equation define radiatoe sth r are termn ai f so a trough formed in the cylindrical splash guard surrounding the radiator length and the aspect ratio Ar. The fourth equa- the collector (see Fig. 1). The radius of the collector depends tion states tha powee tth r radiated whil temperature eth e drops directly on the radiator length and on the aiming accuracy of from T0 to Tl during the droplet's passage from the injector the droplet injector. The radius is required to be large enough collectoe th o t totaequae s ri th lo lt hea t lost fro dropletse mth . reduco t e aperture losse levea t o whicst a l radiatoe hth r will The temperature Te is the average radiation temperature dur- 4 still operat f rt fulLeo a yearld capaciten se (witth t ya h /3My e indroplet'th g s flight (Tj=
Computational Results comparison (assuming 5 kg/m2). Lithium, NaK, and DOW Calculation systemR LD s e sbase th mode n do l have been 70 l system5 oi time 0 1 e — o ar st 5 lighter than heat pipes, carried out to determine overall LDR performance. Figures 4 wherea sattractivs aluminua t LDRn no ti e thin d ei sar m s an an illustratd5 e variation componene th n si totad tan l radiator comparison. However, tin and aluminum LDRs return fluid at masses as functions of the droplet radius and optical depth for temperature 7\ tha cooles i t r tha reture nth n temperature T0 a lithium LDR. In Fig. 4 the initial perception for a fixed constante ofth temperature heat pipe radiator. Thus powea , r radiated power—tha smallee th t droplete th r s arelightee th , r cycle could operate more efficiently with the LDR. the radiator will be—is verified for the complete system, at A significant variation in the LDR—the use of parallel least down to 50-pt size droplets. For example, a threefold radiators—has been incorporate comR LD - e somn di th f eo reductio systen i m mas possibls i usiny eb g 100-/x droplet- sin puter Fig n studie. i Previousl 7) 5 . s 70 s wa (e.g.W t yi DO , stea f 500-do ^ droplets. Figur showe5 effece s th f varyino t g found that with a single sheet, the mass of the return pipe and optica le componen deptth n o h d totaan t l massee th f o s radiator totae Th .l power radiated also variespowee th o rs , per unit mass P/M includes i thin di s figure t smalA . l optical depths (e.g., r
1000
(kg) 500
300
200
100 0 50 0 40 0 30 0 20 0 10 DROPLET RADIUS (microns)
Aiming Accuracy 1 mrad Lifetime 5 years Material Lithium Optical Depth 3.0 Power 6.3 MW th Droplet Reynold. sNo 2000 278°K Aspect Ratio 0.4 Inlet Temperature 510 K componenR LD Fig tota4 .d tan l mas droples sv t radius. power/masR LD Fig 6 . powes sv r radiated. 14 R.T. TAUSSI A.TD G. AN MATTIC K J. SPACECRAFT
the pump to overcome pressure losses dominated the LDR Figur showe8 resulte th s f minimizino s e mase th g th f o s system mass for viscous liquids such as silicone oils. Returning power conversion system plu waste th s e heat radiatore Th . the liquid instea parallea y d b t oppositel bu l y directed sheet radiator mas estimates si aree th ay ddensitb y effectiv e path , e has been foun masR reduco dt sLD significantle eth thesn yi e radiator temperature hemisphericae th d an , l emissivite Th . ye cases. Althoug additionan ha l generato collectod - an r re e ar r smaller area densitdroplee th f yo t radiator allow totae th s l quired, these are downsized since each droplet sheet is smaller system mass minimum to occur at lower cycle temperatures, tha sheee nth t require singla r dfo e traverse. resulting in a higher-efficiency power cycle and a lighter total poinA t desig 10a r 0n fo MW th droplet radiator (actually 18 power system mass. The thermal cycle efficiency is shown at tandem pair f radiatorsso bees )ha n develope givo dt conea - the minimum of each of the cases illustrated. crete picture of the dimensions and operating parameters for curvef o t se sTh p correspondeto advancen a o st d heat pipe such a device. This design is summarized in Table 1. A large radiator havin aren ga a densit emissivitn kg/m5 a f d yo an 2 y numbe f streamo r radiator spe r (~ 10requirede - 6)ar in e Th . of 6 = 0.85 for a blackened metal surface. The bottom curves flight fluid, make-up fluid, and pipe/pump components weigh represent a class of possible LDRs having sheet densities of 0.5 the most. The power per unit mass of 6.8 kW/kg is nearly 7 kg/m2 and emissivities corresponding to an optically thick times higher tha nheaa t pipe radiator operatin same th - t egef a sheet of oil droplets. fective radiator temperatures (
Power Systems Applications It is worthwhile to reoptimize a thermal power cycle when using a droplet radiator because a more efficient cycle with a lower radiating temperature can be obtained with a potential 10 decreas totae th ln e i syste m mass have W .e carrie reopa t d-ou timization of two Brayton cycles and a potassium Rankine cy- e e masseBraytocleth Th . f o s n cycle components (com- pressors, turbines, generators, heat exchangersd an , recuperators) were scaled from a point design study of a closed-loop, space-based GDL laser power system.9 Scaling of the heat sourc head ean t sink heat exchangers (most massive M Vkg/ components dat carrieL s a )wa t froGD accordin e dou m th o gt conventional compact heat exchanger design rules.a r 1Fo 0 fixed peak temperature, Tpea 1644Kk= e remaininth , g cycle parameters were varied to achieve a minimum mass system A\5J* . =0 that would produce 25 MWe. This optimization was carried £ 0.1 -/ out for a range of rejection temperatures of 300 K to 800 K :/ r severafo d lan value f radiatoo s r specific mass, witd an h -/ i without a recuperator in the cycle. t-
0 80 1000 0 60 120 00 40 200
V°K>
Table 1 Baseline 100 MW LDR ih power/masR LD Fig 7 . peas sv k temperature. Specifications Mass radiator Lithium Structure, kg 1,600 5-year lifetime In-flight fluidg k , 2,570 ———— SIMPLE -RECUPERATED BRAYTON Triangular: Base = 33 m Make-up fluid, kg 4,610 Length = 83 m Collectorg k , 870 18 pairs of radiators Generator, kg 1,480 esheet=0-58, rshee3 = t Pipe/pumpg k , 3,500 Aiming accurac mra1 y= d 14,630 Droplet radius = 50 /i Droples t m/ spee 2 2 d= LDR power/mass: 6.8 kW/kg Tin = 528K, rout = 460K (Heat Pipe Power/Mass: Collector radiu 0.3s= 3m 1.08 kW/kg) Pressure drop = 8.5 atm 0.85 kg/m= 5 e , t ,K (a 77 2)=48 Number droplet stream s10 x =3 6 Pumping W powek 4 6 r= 300 400 500 600 700 800 900 Radiated power = 5.5MW MINIMUM CYCLE TEMPERATURE (K) Fig. 8 Reoptimized power cycles. JAN.-FEB. 1986 RADIATOR SYSTEMS FOR SPACECRAFT THERMAL CONTROL 15
based on data taken from the design studies for the Solar vicinite th cu droplea n i rf yo t peak temperatur f 103eo 0K Power Satellite,13 which turnn i , larg,e drath en w dato a base (e.g. Fige , givinse ,6) . gmeaa n effective radiator temperature generated durin e 1960 gth d earl an s y 1970 r potassiufo s m of approximately 729 K. space power systems (for example, see the article by These calculations suggest tha t leasa t threefola t d decrease Zimmerman.14) in power system mass is available for Brayton cycles through A combination of one high-pressure turbine and two low liquid-droplef o - e us e th t radiators. Also mase ,th s minime aar pressure turbine s usei s o modedt e turbinth l e component. shifted to lower radiator temperatures compared to the heat Masse unir spe t electric power generated were taken from Ref. pipe radiator cases eacn i hK . Thi0 ordee 10 sth f shafn ro o s i t 13 studie equated san functionaa o dt turbine l th for r mfo e case. component mass. This function depend turbine th n so e radius Perhap mose sth t important advantag droplef eo t radiators squarelengte th , equivalentlyd hor dan numbee th , stagef ro s is their ability to lower the peak cycle temperature without in- per turbine. The turbine radius can be related directly to the creasin systee gth m mass compare usine on gconvena o dt - square root of the peak cycle temperature divided by the peak tional radiator. That is, lower peak temperatures will increase cycl e woreth pressurky b extracte d an e r pounpe d f o d the droplet radiator mass since the cycle is less efficient, but potassium in the expansion leg of the thermal cycle. The radiatoe th r mas initialls si smalo ys increasln thaca t i t e quite numbe turbinf o r e stage relates si cycle th ef do alsg lo o e toth befort bi a exceedt ei mas e conventionaa s th f so l radiatorr .Fo pressure ratio, assuming approximately equal ratio expanf so - example decreas,a radiaton ei r temperaturK 0 30 eo t fro 0 m50 sior stage nturbine pe Th . e inlet Mach numbe assumes wa r d would permit a drop in peak cycle temperature from 1644 K to constant. 986 K; use of a droplet radiator for the lower temperature cy- The heat source heat exchange complea s ri x assortmenf o t cle would samresule th en i t power system mas heaa s sa t pipe heat absorber tubes, hot manifolds and headers, and the in- radiator for the higher temperature cycle. This example il- ventor liquif yo thesn di e tubes potassiue Th . m boiler compo- lustrate limie ranga sth f to intermediatf eo e power cycles that nent described here alss beeoha n modele e Ref3 th 1 .n di will have both lower peak cycle temperature lowed san r system studie r severafo s l different pressure d temperaturesan s A . masses tha heae nth t pipe case. Fro viewpoine mth thermaf to l chief feature of this component is a strong mass dependence heat source development advantagee th , trule sar y significant on peak pressure temperaturesd an s have W .e modeled this since it is very difficult to gain improvements in the area of componen relatiny b t boilee gth r numbe e masth o st tubef ro s high-temperature material power sfo r systems.15 needed to transport the potassium at any given flow station and to the tube wall thickness and tube length; the fluid inven- Current Statu f Technologso y tor addes yi thio dt s mass numbee Th . tubef ro takes se i b o nt The Air Force and NASA have engaged in a joint program proportiona totae th lo lheat t inpu ttubsystem e e toth th ed ,an investigato t e liquid-droplet radiators. This progra msups i - wall thickness depend peae th kn so cycl e e th pressur n o d ean porting research on droplet generato dropled an r t materials, peak temperature dependence since the creep rupture strength radiative transfer from dense cloud f dropletso s , droplet of the piping has a rather strong temperature dependence. chargeu spaca n pi e environment, droplet radiator systems, This analysis assumes that the actual tube radii will not change and droplet collectors. The droplet generator technology and from the values used in Ref. 13; a more detailed optimization pump component technologies are the most mature of those also should examin impace eth f allowino t tube gth e radio t i needed for droplet radiators. Single streams of uniform change. The trend toward heavier boilers with increasing peak droplete produceb n ca s d ove a widr e rang f orifico e e cycle pressur d temperaturan e s partlyi e , thoug t verno hy diameters and for a variety of liquids. Single droplet streams strongly, offset by the increasing cycle efficiency, so that for a e aimeb n dca with better than milliradian accuracies. given electrical output the amount of heat transferred through However mechanicae th , l control neede accompliso dt h this the boiler decreases compare loweo dt r pressure cycles. same feat with 10o 10t 5 7 droplet streams requirea r fo d Sinc heae eth t sink heat exchange essentialls ri isothern ya - droplet t radiatobeeno ns demonstratedha r o 100t p 0U . mal condense pressurew lo t moderatd a r an s e temperatures, streams have been created with uniformly sized dropleta n si the tub emuce wallb n h sca thinne r tha boilee nth r tubese Th . low-viscosit printingjet y liquiink .d for Curren t laboratoryex- effectiveness of this component is prescribed in terms of the periments using low-vapor-pressure higher-viscosity oils have t inleho t temperature (essentially constant potassium recently created on the order of 100 streams from one orifice temperature as it condenses) and the cold inlet and outlet plate. There is still a considerable challenge to extend this temperature radiatoe th f so r coolant varyiny B . heae - gth ex t technology to a truly large nozzle array. changer effectiveness mase th , s flow throug radiatoe hth n rca A variety of low head pumps have been space-qualified for be varied once the radiator liquid specific heat has been fixed. handling low-temperature fluids in environmental cooling cir- Calculations using a radiator area density of 5 kg/m2 show cuits on the shuttle and for cryogenic liquid rocket fuels. The tharadiatoe th t relativela s i r y small fractio powee th f o rn primary zero-G pum inhibio t e p ar requirement t R LD e th r sfo system mass in this case. Consequently, one would not expect cavitation at the pump inlet and to provide sufficient inlet much benefi thio t t s cycl descreasiny eb radiatoe gth r weight. pressure to overcome the frictional losses at the pump en- The heat source heat exchanger mass is also quite low. The trance. Some of these problems have been accentuated for Ref. 13 studies show that this mass does not become signifi- volatile liquid spropellantse sucth s ha thesn I . e cases, pumps cant until peak pressures exceed 200 psia. From 200 to 500 have employed inducer reduco st e cavitatione case th f th eo n I . psia heae th , t exchanger mass becomes comparabl ture th - o et liquid-droplet radiator, where very low-vapor-pressure liquids bine mass. The turbine mass actually decreases over this same are used, cavitation will not be as serious a problem; rather, range of pressures with the sum of heat exchanger and turbine t positivne e th e suction head (NPSH) r inleo , t pressur- re e masses remaining nearly constant. Because higher pressures quirements probley ke e , mth wil e arealb . Laboratory testf so are associated with higher efficiencies, the radiator mass will pumps drawing liquid from vacuum vessels have shown that decrease. Thus, Fig. 8 illustrates a best case so far as the ordinary gear and diaphragm pumps require very little inlet benefits of lighter radiator masses are concerned; it is even pressurization (e.g., several inches of liquid) for their opera- harder to show benefits from lighter radiators at higher cycle tion with low-vapor-pressure liquids. Similar pumps have been pressures. flight-qualified for low-temperature liquids but not at higher radiatoe Th r temperatures favore thiy db s particular cycle temperatures. are in the range where droplet liquids such as aluminum and significanA systemR t outcomLD e s th analysi f eo thas si t tin might be attractive. A slight minimum in the total system the collector mass can be a large fraction of the system mass mass is shown in Fig. 8 for temperatures near 750 K. The unles aimine sth g accurac droplee th f yo t generato bettes i r r largest powe unir rpe t mass dropleration ti r sfo t radiator- soc (smaller) than 10 mrad. Consequently, there is a dual urgency 16 R.T. TAUSSIG AND A.T. MATTICK J. SPACECRAFT
to demonstrate lightweight collector concepts and to verify droplet velocities were examined. The results have shown that droplet aiming accurac largn yi e arrays. Collector efficiencs yi splashing can be virtually eliminated, within the accuracy of issuey ke splashinf I .a g occur noticeabla t sa ecole leveth - t a l the instrumentation (photographic and visual observation) by lector, the nlose liquib t y fro systede ma m unacceptn th a mt a - operating the droplet stream velocities and plate rotation able rate, limitin lifetime missioge th th f possibld eo nan y con- velocities belo wa critica l relative droplet Weber number. taminating the spacecraft. Collection and splashing ex- These results are shown in Fig. 10. The splash-free regime cor- periments have been completed recently for a small rotating responds to a range of relative Weber numbers depending on collector. Rotation provides an artifical gravity force to fill a the angl f attacdroplee eo th f ko t stream. This range allowsa collectotrouge edge th th f et o h a r wit incomine hth g drople- tli useful operating range for the LDR system, as shown in Fig. quid (se escoo A Fig . 1) p. will remove liquid from that trough, 11, obtained from the LDR systems model for DOW 705. The pressurizing it for the inlet flow to a low NPSH pump to recir- bars at the top of Fig. 11 compare the range of test dimen- culat liquie eth d throug waste hth e heat source- li . e Testth f so sionless parameters to those corresponding to a full-scale quid collectio transpord nan t acros e rotatinth s g collection LDR morA ( . e detailed discussio f thio n s experimens i t plate to the trough of the collector have been quite successful. published elsewhere.5) Additional testing wil requiree b l o dt experimentae Th l setu shows pi Fign i , whic9 . h illustratee sth verif splash-free yth e regim denser efo r droplet spray impinge- vacuum test chambe uppee th closeun a ri r col e d lefth - an tf po ment conditions and to demonstrate repressurization of the lector plate and drive motor assembly. Droplet streams are collected liquid fro collectoe trouge edge mth th th f eo t ha r aimed at the collector plate to determine the conditions under plate. which splashing occurred. Various angle f dropleo s t stream In additio o componennt t technology development, there impact, different film thicknesse rotatine th n o s g plate, dif- severae ar systey lke m issues that concer integratee nth d opera- ferent rotation velocities ranga d f dropleeo an , t sized an s liquid-droplea tio f no t radiator. These include component compatibility wit droplee hth t liqui adjoinind dan g elements of the system; integration of the LDR system with the waste heat source; transien d part-loaan t d respons e heath to t e VACUUM TEST CHAMBER source; orbit adjustment maneuverd an s s (e.g., rapid slewing could result in lost droplets and unbalanced thrusts on the radiator); startup/shutdown conditions; potential spacecraft contaminatio d droplean n t liquid contaminatio y gaseb n s from the spacecraft; stowage; and deployment. Very little wor bees kha n don addreso et s these issue dateo st . DRIVE MOTOR
Triangular Sheet Aspect Ratio =0.4 Droplet Radius = 50 y Droplet Velocity = 11 m/s TACHOMETER Optical Depth =0.5 Aiming Accuracy = 1 mrad Lifetime = 5 years
Re Bo
6 600 T 650 80 T 10i IxlO ROTATING COLLECTOR r h50 -V° 350 Fig 9 Tes. t chambe d closeuan r f o rotatinp g collector test DATA^"* equipment. \ \ 0. • 2.5 x 10~6 1 0 J- 0 DATA FROM THIS FIGURE ROTATING PLATE VELOCITY ____„ 6 /////////////////////////////77T7 SPLASHED ' DROPLETS
Ud, INCIDENT DROPLET VELOCITY
U, RELATIVE DROPLET VELOCITY
0.2 -
100 ym 100 - 0 40 100 0 20 0 0 10 0 6 0 4 0 2 10 POWER RADIATED (kW^) 0 7 0 6 0 5 0 4 0 3 0 2 0 1 0 Fig. 11 Rotating collector parameters vs power radiated (DOW 705);
Becaus^ of the recent success of the collector and droplet References generator tests t seemi , s feasibl demonstrato et escalabla e ver- ^attick A. T. and Hertzberg, A. "Liquid Droplet Radiators for droplea sio f no t radiato r test fo rspacn i s e withi neae nth r Heat Rejectio n Space,i n " Journal f Energy,o Vol , 19815 .. pp , future smalA coulR . e testeLD b dl d withi a nsuitabl e 387-393. enclosure on board the space shuttle. Development of a full- d Hertzbergan 2 Matdc. T . ,A k A., "The Liquid Droplet size radiator could follow within a five-year period after these Radiator—An Ultralight Heat Rejection System for Efficient Energy tests, for application to early low-temperature thermal power Conversion in Space," Preprint IAF-81-185, 32nd Congress of the In- ternational Astronautical Federation, Rome, Italy, Sept. 6-12, 1982. system or, perhaps, for cryogenic refrigeration systems in d Hertzbergan 3Mattic . T . A k, A., "Liquid Droplet Radiator space. Technology Issues," Paper TMP-2, First Symposium on Space Nuclear Power Systems, University of New Mexico, Albuquerque, Conclusions Jan. 11-13, 1984. Systems analysis has shown that a triangular droplet sheet 4Botts, T.E., Powell, J.R. Hornd ,an , F.L., "Magnetic Focusinf go radiator with a rotating collector can be 5 to 10 times lighter Liquid Droplet Sheets for Heat Rejection in Space," Air Force Rocket than an advanced heat pipe radiator operating at the same Propulsion Laboratory Report AFRPLTR-83-072, Dec. 1983. Taussigd 5Mattican . . T.T R . ,, kA "Ne w Thermal Management peak temperature. Radiated power per unit mass maxima oc- d Heaan t Rejection System r Spacfo s e Applications, r ForcAi " e liquir fo dW lithiumM 0 2 o t t a , 5 cualuminum d t ra an n ti d an , Rocket Propulsion Laboratory Report AFRPLTR-84-039, June 1984. NaKd an 5 . Aimin70 W gDO accurac r fo higd W k yan h 0 10 6Hertzberg, A., "Advanced Lightweight Nuclear Power Sources collection efficiency emerg criticas ea l requirementR LD r sfo for Laser Beam Applications, Sym-L " GC Proceedings h 4t e th f o performance. Lower radiator weights can lower the total posium, Sonesa, Italy, Sept. 1982. power system mass by a factor of 3 to 5; these mass minima 7Beals G. et al., "Lightweight Nuclear-Powered OTV Utilizing a occu t approximatela r loweK 0 ry10 rejec t heat temperatures Liquid Droplet Radiator," AIAA Paper 83-1346, AIAA/SAE/ASME 19th Joint Propulsion Conference, Seattle, WA, June 1983. compare heao dt t pipe radiato masR srLD systemsw lo e Th . 8 also allows the peak cycle temperature (e.g., the nuclear reac- Nesmeyanov, A. N., Vapor Pressure of the Elements, Academic Press, New York, 1963. tor temperature) to be reduced by as much as 40% without in- 9Young, W. E. and Kelch, G. W., "Closed-Cycle Gas Dynamic creasing the system mass beyond that required for a more con- Laser Design Investigation," NASA Report No. CR-135130, Jan. ventional heat pipe radiator power system. 1977. Tests of droplet generator and collector concepts have been 10Kays, W. M., and London, A. L., Compact Heat Exchangers, carrie measuro t t dou e aiming accuracy, uniform droplet for- 2nd Ed, McGraw-Hill, New York, 1964. mation, droplet velocity, and collection efficiency. The suc- HPerry, R. H., and Chilton, C. H. Eds., Chemical Engineer's cess of these tests suggests that scalable system demonstrations Handbook, 5th ed. McGraw-Hill, New York, 1973, pp. 3-187. e nea th concep R re carrien b i future LD t n e dou ca to th f, 12Robert D. Gruntz, "A Liquid Metal Rankine Topping Cycle for a followe y testb dn spac i s o verift e y zero-G performance Steam Power Plant," Journal f Spacecraft,o . 859-964Volpp , 4 . . predictions. Rotating collector test data particularn i , , have 13Gregory, D. L., and Woodcock, G. R., Solar Power Satellite, shown recently that a splash-free regime of operation appears System Definition Study, Vol. Ill, "SPS Satellite Systems," NASA- Johnson Space Center, Contrac . NAS9-1596No t , Dec. 1977. exiso t r LDRfo t s using vacuum oils. Further dense droplet 14 stream test needee sar verifo dt y this performance. Zimmerman, W. F., "Alkali Metal Space Power Technology Ap- plicabl Nationao et l Energy Researc Development,d han " Journalf o Energy, Vol. 1, 1977, pp. 159-168. Acknowledgment 15Elliot . G.D , , "Rotary Radiator Reducer sfo d Space Powerplant This work has been supported by the Rocket Propulsion Temperatures," Space Nuclear Power Systems, Vol. 1, edited by M. Laborator Edwardat y ForcAir s e Base, California, under S. El-Genk and M. D. Hoover, Robert E. Krieger Publishing Com- Contrac . F04611-81-K-0040No t . pany, Inc., Melbourne, Australia, 1984.
AIAA Meetings of Interest to Journal Readers"
Meeting Call for Date (Issu f AIAAeo Bulletin whicn i h program will appear) Location Paperst
1986 Feb 11-13 AIAA Aerospace Engineering Los Angeles Airport Hilton Conferenc Shod ean w (AECS) (Dec) Los Angeles, CA April 14-1 6 Ballistic Missile Future Systems Norton AFB, July 85 and Technology Workshop (Feb) Bernardinon Sa A C , May 19-21 AIAA/ASME/ASCE/AHS 27th Structures, Marriott Hotel May 85 Structural Dynamic Materiald san s Conf . (Mar) Antonion Sa X T , Jun4 e2- AIAA/ASME 4th Thermophysics and Sheraton-Boston Hotel Sept 85 Heat Transfer Conference (Apr) Boston, MA June 9-11 AIAA 4th Applied Aerodynamics Inter-Continental Hotel Sept 85 Conference (Apr) San Diego, CA June 16-18 AIAA/SAE/ASME/ASEE 22nd Joint Propulsion Von Braun Civil Center Sep5 8 t Conference (Apr) HuntsvilleL ,A Aug 18-20 AIAA Atmospheric Flight Mechanics Williamsburg Hilton Nov85 Conference (June) Williamsburg, VA
*Fo completa r e listin f AIAgo A meetings currene th e tse , AIAA e issuth f e o Bulletin tissue of AIAA Bulletin in which Call for Papers appeared.