A Morphing Radiator for High-Turndown Thermal Control of Crewed Space Exploration Vehicles

Thomas J. Cognata1 Business or Academic Affiliation 1, City, State, Zip Code

Darren Hardtl2 Business or Academic Affiliation 2, City, Province, Zip Code, Country

and

Rubik Sheth3, Craig Dinsmore4 Business or Academic Affiliation 3, City, State, Zip Code

Spacecraft designed for missions beyond low earth orbit (LEO) face a difficult thermal control challenge, particularly in the case of crewed vehicles where the thermal control system (TCS) must maintain a relatively constant internal environment temperature despite a vastly varying external thermal environment and despite heat rejection needs that are contrary to the potential of the environment. A thermal control system is in other words required to reject a higher heat load to warm environments and a lower heat load to cold environments, necessitating a quite high turndown ratio. A modern thermal control system is capable of a turndown ratio of on the order of 12:1, but for crew safety and environment compatibility these are massive multi-loop fluid systems. This paper discusses the analysis of a unique radiator design which employs the behavior of shape memory alloys (SMA) to vary the turndown of, and thus enable, a single-loop vehicle thermal control system for space exploration vehicles. This design, a morphing radiator, varies its shape in response to facesheet temperature to control view of space and primary surface emissivity. Because temperature dependence is inherent to SMA behavior, the design requires no accommodation for control, instrumentation, nor power supply in order to operate. Thermal and radiation modeling of the morphing radiator predict a turndown ranging from 11.9:1 to 35:1 independent of TCS configuration. Stress and deformation analyses predict the desired morphing behavior of the concept. A system level mass analysis shows that by enabling a single loop architecture this design could reduce the TCS mass by between 139 kg and 225 kg. The concept is demonstrated in proof-of-concept benchtop tests.

1 Insert Job Title, Department Name, Address/Mail Stop, and AIAA Member Grade for first author. 2 Insert Job Title, Department Name, Address/Mail Stop, and AIAA Member Grade for third author. 3 Insert Job Title, Department Name, Address/Mail Stop, and AIAA Member Grade for fourth author (etc). 4 Insert Job Title, Department Name, Address/Mail Stop, and AIAA Member Grade for fourth author (etc).

https://ntrs.nasa.gov/search.jsp?R=20140017124 2020-06-28T18:01:56+00:00Z

Page2of15

I. Introduction Spacecraftdesignedformissionsbeyondlowearthorbit(LEO)faceadifficultthermalcontrolchallenge,particularlyinthecaseofcrewedvehicles.Thethermalcontrolsystemmustmaintainarelativelyconstantinternalenvironmenttemperaturedespiteavastlyvaryingexternalthermalenvironment.Theexternalenvironmentmaypresentasinktemperaturethatrangesfrom70Kintrans‐planetarycoast(TPC)toapproximately228Kinplanetarysurfaceoperations(PSO),andheatrejectionneedsthatarecontrarytothechangeincapacityoftheenvironment.Inotherwords,thethermalcontrolsystem(TCS)isrequiredtorejectahigherheatloadtowarmenvironmentsandalowerheatloadtocoldenvironments,necessitatingaquitehighturndownratio.

ThisthermaldesignneediscomplicatedbythechallengeoftransportingheatfromcrewedportionsofthevehicletotheheatrejectionportionofaTCSwithouttheendangermentofcrewmembersbyworkingfluidtoxicityorbypotentialreactionoftheworkingfluidwithothervehiclesystemsduetoaleak.ThisdrivesthecurrentstateoftheartTCSdesign–atwo‐loopsystemwithapropylene‐glycol/waterworkingfluidcontainedinaninnerTCSloopandwithalowfreezingpointrefrigerantinaloopwhollyexternaltothecabinwhichisabletooperateovertheverywiderangeoftemperaturesthattheradiatorissubjectto.ThisdesignhasthedrawbackofaddingsignificantmassandcomplexitytotheTCS,interfacialinefficienciestoheattransferbetweentheloops,andtheduplicationoffluidhandlingequipmentsuchaspumpsandexpansiontanks.Previoustrades[19,20]haveshownthatatwo‐loopTCSmaybeapproximately25%heavierthanasimilarlyperformingsingleloopsystem.

Thisworkdescribesauniquelouvre‐derivativeradiatorwhichemploystheshapememorybehaviorofNitinoltoproduceveryhighturn‐downsratioscapableofenablingsingle‐loopthermalcontrolofavehicleusingpropylene‐glycolorsimilarcommon,non‐toxic,highfreezing‐pointworkingfluids.

Figure 1 A representative mission profile illustratingvarying thermal environment and heat rejection.

Page3of15

II. The Morphing Radiator Thistechnologyemploystheshapememoryeffectinconjunctionwithanintegralbiasloadtoactuatethesurfaceofaradiator.ThebasicfunctionofthisconceptisillustratedinFigure7.Theradiatortakesoneoftwobasicshapesdependingonthetemperatureofthefacesheet.Whenhottheradiatorwilltakeitsfullyextendedausteniticshape,shownatthetopofthefigure.Whencoldtheradiatortakesitsdeformedmartensiticshape,thesemi‐circleshowninthebottomofthefigure.Byvaryingshapethisconceptalterstheviewfactoroftheradiatortospace,maximizingviewinthehotcaseandminimizingitinthecoldcase.Thethermaleffectofthisgeometricactuationismultipliedbyselectivesurfaceemissivity,wherethetopsurfacehasahighemissivityandthebottomsurfacealowemissivity.

AnillustrationofanarrayofsuchpanelsisshowninFigure6.Thisdemonstratestheproportionalturndownresponseofthisconceptinaparallelflowradiatorconfiguration.Hotfluidenterstheinletheadertothisarrayatthetopleftofthefigurefromwhichitenterseachflowtube.Workingfluidtemperature,andthusradiatorpaneltemperature,decreasesalongthelengthofeachflowtubeasheatisrejectedthroughradiation.

Inahotenvironment,asillustratedbytheleftmostflowtubeandpanelsinFigure6,theworkingfluidandpaneltemperatureremainsabovetheshapememorytransitiontemperaturesoallpanelsareinthehotstate.Thus,heatisrejectedthroughthegreatestpossibleradiatorarea.Inacoolenvironment,however,representedbytherightmostflowtubeandpanelsinFigure6,panelsbegintakingthecoldshapeandlimitingheatrejectiontowardtheendoftheflowtubeastheworkingfluidtemperaturefallsbelowthetransitiontemperatureoftheshapememoryalloy.AsaresultofsuchbehaviorthisradiatordesignwilltendtomaintainaminimumsystemfluidtemperatureinthevicinityoftheMftemperatureoftheshapememoryalloy,abehaviorthatcanvastlysimplifycontroldesignofthethermalcontrolsystemforavehicle.Thetargetcontroltemperatureforasingle

Figure 3 A conceptual illustration of the technology. Top shows the austenitc shape at high temperature and below is the deformed martensitic shape at low temperature. A central tube carries a thermal working fluid.

Figure 2 An illustration of an array of panels in a parallel flow radiator configuration. The array illustrates the passive proportional turn-down of such a system design; as heat is rejected and the working fluid cools, downstream panels take the cold shape and limit the heat rate. The result is a thermal control system which is responsive to the heat rejection needs of the vehicle.

Page4of15

loopthermalcontrolsystemusingaPGWworkingfluidisinthevicinityof‐10C,anidealfitformedicalgradenitinol.

ThistechnologymaybeintegratedintoasingleloopthermalcontrolsysteminplaceofthetraditionalradiatorshowninFigure9.Noaccommodationisneededforcontrol,powersource,orinstrumentationastheactuationofthistechnologyispassiveandafunctionoftheinherenttemperaturedependenceoftheshapememorymaterial.

III. Turn‐down and Thermal Analysis  Afirstorderevaluationofperformanceofthisconceptisfoundthroughthedefinitionofradiationheatrejection,Qi Fi i A Tsurface

4 Tspace4 ,

whereQiistheheatrate,Fiistheviewfactoroftheradiator,itheemissivityoftheactivearea,theStephan‐Boltzmannconstant,Athetotalradiatorsurfacearea,Tsurfacethetemperatureoftheradiatorsurface,andTspacethesinktemperature.Theconceptturndowninthisidealcaseisaratioofthehighandlowstates,or

R Fi,space i Tsurface

4 Tsin k4

Hot

Fi,space i Tsurface4 Tsin k

4 Cold

wherethesubscriptireferstoasurface,e.g.topandbottom,thesubscriptHotreflectsvaluesofthehighheatrejectionshapeandhotenvironmentandthesubscriptColdreflectsvaluesofthelowheatrejectionshapeandcoldenvironment.

Thisconceptiscapableofa15:1turndownbasedontheparametersandtheenvironmentsdescribedinFigure1andTable1.Table1showstheviewfactorandemissivityforeachsurfaceinbothhotandcoldcases.EmissivityofthetopsurfaceistypicalofasilverTefloncoating.[23]EmissivityofthebottomsurfaceistypicalofhighlypolishedmetalorofMLIblanketsformediumareaapplications.[24]Thetopsurfaceviewfactorinthecoldcaseiscalculatedasthatoftheinnersurfaceofacylinderthroughitsopenends.[25]Bottomsurfaceviewfactorsassumetheradiatortobebodymountedtoavehicle.[26,27].

Table 1 Defining characteristics.

Hot ColdFtop,space 1 0.031εtop 0.81Fbottom,space <0.001 0.560εbottom 0.015Tsink 230K 70KTsurface 293KRadius rLength 32rWidth 2πr

Figure 4 Concept A Thermal control system design. A single loop thermal control system integrating Concept A as the Radiator.

Page5of15

Aturndownof60:1becomespossiblewithshieldslocatedateachendofthecylinderformedbythecoldshapeasthesedrivethecoldvalueofFtop,spacetowardzero.Thisrangeofturndowngreatlyexceedsthecurrentstateoftheartandindicatesthedesignroomavailabletomeetmissionbasedturn‐downneedswhereconditionsdonotmatchanalyticalassumptions.Infact,itisusefultonotethattheturndownratioisaproductoftwoturn‐downterms,thephysicalturn‐downwhichiscomposedoftheviewfactorandemissivityandaconsequenceofdesign,andtheenvironmentalturn‐downwhichisdescribedbytemperatures.Thephysicalturn‐downis96:1withendshields,25:1without.

AfinitedifferencerepresentationwascreatedusingThermalDesktopwithFLUINTinordertopredicttheconceptturndownwithhigherfidelityandwithasimulatedfluidloop.Therepresentationincludesaparameterizedradiatorgeometry,asimplifiedthermalcontrolfluidloop,andsimplifiedgeometryrepresentingavehicle.

Figure10andFigure11showthecoldandhotcaseanalysisgeometries,respectively.Thesefiguresshowtheradiatorandfluidpath;thevehiclegeometryandthebalanceofthethermalcontrolloopishiddenforclarity.ThemodelisbuiltusingThermalDesktopnativegeometry,asopposedtofiniteelementgeometricapproximations,sothatcalculationsare

Figure 6 Thermal Desktop finite element representation of the hot shape. The model is parametric: this shows geometry given the maximum radius.

Figure 5 Thermal Desktop finite element representation of the cold shape. The model is parametric: this shows geometry given the minimum radius.

Page6of15

mathematicallyaccurate.[28]ThermalDesktopemploysaMonteCarloraytracingmethodtocalculatetheviewfactorofeachelementofthegeometryforhighconfidenceradiationheattransferprediction.

Theworkingfluidinthefluidloopisa50/50ratioofpropyleneglycoltowaterandisbasedonapropertydeckwhichincludesexperimentallymeasureddataatlowtemperaturesapproachingthefluid’sglassystate.Asetofdesignoptimizationstudiesareperformedtodeterminetotalareaextentoftheradiator,panelwidth(ortubespacing),andpanellengththatmeettherepresentativemissionneedsdescribedbyFigure1.Table3outlinesthekeyinputsandselectedperformanceparametersthatdefinethemodel.Thesystemflowrateissettoprovidearadiatorinlettemperatureof40Cgivenasystemsetpointtemperatureof1.7C.ThesesystemtemperaturesareselectedbasedontherangesemployedinpreviousvehiclestudiesincludingseveralofOrion[1],Altair[2],andMMSEV[30,31].Thefluidmodelusespathduplicationtoapportiontheappliedvehicleheatloadandflowratetothesinglepanelgeometryasifitwereoneinanarrayofparallelpanels.

IV. Stress and Deformation Analysis Withregardtotheshapememoryalloyconstitutivemodel,werequireone

thatcapturesallthepertinentthermalandmechanicaleffectsofinterest.Thisincludesinparticularthemartensitictransformationinitiatedandmaintainedbysomecombinationofstressandtemperatureandtheeffectsofheattransfer.Tothisend,wehavechosenthemodelfromthelegacyworkofLagoudas,Hartl,andcoworkers[34].Itisphenomenologicalinitsformulationandcalibrationandconsiderstheaveragethermo‐mechanicalresponseofagivenrepresentativevolumeelement(RVE)byfirstpostulatingRVE‐averagedenergyanddissipationpotentials.Thisthermodynamicallyorientedapproachallowsforstraightforwardformulationofcouplingsbetweentheenergeticsofheattransferandthoseofdeformation,whicharesocriticalintheanalysisofSMAs.Further,wehavehereintakentheapproachofHartlandLagoudas[35],whichconsiderstheeffectsoflargestructuraldeformations(rotations)aswillbecommoninthecurrentwork.

Table 3 Thermal modeling parameters.

Parameter Hot ColdFluidTin 40C 8.1CQload 5.8kW ≤1kWTotalArea 34.8m2 Panelwidth 24cmPanelLength 1.2mSegmentlength 10cmÝm 175kg/hcp(@10C) 3354J/kg‐Knpanels 121Tsetpoint 1.7°CkApanel,x(unitlength) ≥0.152W‐m/K

Table 2 Composite facesheet properties

Thickness 0.64mmkxpanel 360W/m‐Kky,zpanel 0.21W/m‐KkApanel,x 0.230W‐m/Kfacesheet 1875kg/m3”panel 2.49kg/m2

Page7of15

Theoverallhierarchicalanalysistoolisflexibleandmodular;differentsimulationprocessmanagers,FEAtools,orconstitutivemodelscanbesubstitutedatanytime.Inthecurrentwork,weutilizeaprovencombinationoftheAbaqusUnifiedFEASuiteandusermaterialsubroutines[33].Abaqusisusedtoperformpreprocessing,processing,andpost‐processingoperations,whiletheusermaterialsubroutine(UMAT)associatedwiththeconstitutivemodelhasbeencodedinFortran.

Thestressanddeformationanalysisfocusedonheattransferfromacentralrodthatsimulatestheflowpipeandintotheattached“winglets”oftheconcept,whichincludeSMAmaterialregions.Thewingletsareformedinaninitiallystress‐freecylindricalshape.SufficientheatingofSMAmaterialplacedaroundtheouterdiameterofthecylinderleadstostrainrecoveryandthusflatteningofthewinglet,increasingtheheatrejectioneffectivenessofthestructureasaradiator.ThecapabilitiesoftheSMAtoaffectthisbehaviorandthecapabilityoftheelasticcompositelaminatetowithstanditwithoutfailurewillbeassessedinthisanalysis.

Thefiniteelementmodelforthisconceptconsidersthecoupledthermalandmechanicalresponsesofallbodiesinvolved.Insuchanalysis,localizedandevolvingthermalconditionsresultingfromthephysicsofheattransferdrivethelocaltransformation‐induceddeformationresponseintheSMA,whichthenleadstodeformationinothermaterialregionsdependingonimposedmechanicalcouplingconstraints(i.e.,contactormeshties).Themodelconsideredtwokeystructuralcomponents:theflowtubeandtheflexiblewinglet.Theflowtubewasmodeledasasimplealuminumtubeforreasonstobedescribed,whilethewingletitselfwasmodeledasalaminateshell.TwospecificdesignoptionswereconsideredwithregardtotheconfigurationoftheactiveSMAcomponent(s).

OneoptionconsideredtwobundlesofSMAwirestiedtothepetaledgesandincontactwiththeouterlayerofthewingletelsewhere(SMAWireoption).Themodelassociatedwiththisdesignutilized540linear3‐Delements(C3D8T)fortheflowtube,825quadraticshellelements(S8RT)forthecompositewinglet,and33quadraticshellelements(S8RT)fortheSMAwirebundles.

ThesecondconsideredtheuseofanSMAthinsheetorfoil,whichenterstheanalysisasanadditional(thermallyactive)laminainthewingletcompositelayup(SMAFilmoption).AnalysisofthisdesignismuchmorecomputationallyefficientandalsohighlightstheflexibilityoftheimplementedSMAmodelindirectlycalculatedintra‐laminarSMAresponse,whichassumesplanestressconditions.Themodelassociatedwiththisdesignutilized540linear3‐Delements(C3D8T)fortheflowtubeand825quadraticshellelements(S8RT)forthecompositewingletthatincludesanSMAfilmlayer.

Page8of15

Modelsforbothdesignoptionsassumedlengthwisesymmetrywithrespecttotheflowtubeforthepurposesofcomputationalefficiency.TheanalysismodelfortheSMAwireoptionincludingadetailofthecompositelayupisshowninFigure10.Thebottomlaminainthelayupasshowncorrespondstotheinnerdiameterofthecurledwinglet.ThemodelfortheSMAfilmdesignoptionisverysimilar.TheSMAwireregionsareremovedandanadditionallaminaisaddedtotheshellcompositelayup.ThisaugmentedandannotatedlayupisdetailedinFigure11.

Thetimehistoryofthermalconditionsusedinthisanalysisreflectsthepreliminaryassumptionthatthewingletresponseisdrivencompletelybythetemperatureoftheflowtube(i.e.,atemperatureboundarycondition),andthatthistemperaturevariesnegligiblyoverthelengthofasinglewinglet.Basedonthisassumption,thecurrentthermallycoupledanalysisconsideredacaseinwhichtheflowtubetemperaturewasincreasedinaspatiallyhomogeneousmannerfromT=263KtoT=313Kover60sfollowedbya10sdwell.Duetothisevenheating,thetubewasmodeledasasimplesolidaluminumbody.ThetransientpropagationofheatoutwardfromtheflowtubeandintotheSMAfilmlayerinthewingletsinducesthephasetransformation,whichleadstostrainrecoveryanddeformationoftheinitiallycylindricalwinglettowardaflatconfiguration.

ThecriticalmaterialbehaviorsofinterestinthisstudywereassociatedwithboththeflexiblecompositeradiatorwingletsandtheSMAwire/filmthatdrivesthebendingdeformationofthosewinglets.Thecompositewasassumedtobeconstructedfromcarbonfiber‐basedandglassfiber‐basedunidirectionallaminae,andtheselaminapropertieswerecarefullyobtainedfrommicromechanicalcalculationsconsideringfiberandepoxymatrixproperties,themselvesobtainedfromanumberofsourcesincludingcommercialcompositessoftwaredatabases(theepoxy)andpublishedmanufacturerdata(thefibers).K1100fiberswerechosenespeciallyfortheirhighthermalconductivityinthefiberdirection.Glassfiberlaminaewereusedtoprovidestructuralintegrityinthetransversedirection.ThepropertiesofthecarbonfiberandglassfiberlaminaearegiveninTable 7.

SMA Wire Flat Bundles (6 wires ea.)

SMA/CompositeTie Constraint

SMA/CompositeSliding Contact

Figure 7 - Coupled thermo-structural finite element model of Concept A (SMA wire option) with composite layup detail.

Ply 1(Inner

winglet diameter)

Ply 5

Ply 6(SMA; Outer winglet diameter)

Ply 4

Ply 3

Ply 2

Figure 8 – Winglet composite layup for the coupled thermo-structural finite element model of Concept A (SMA film option)

Page9of15

Table 4 - Material properties for K1100 carbon fiber-based lamina

Material Property

K-1100 carbon fiber

S2 glass fiber

ρ 1812 kg/m3 1969 kg/m3 E1 557 GPa 54.4 GPa E2 6.23 GPa 15.9 GPa ν12 0.318 0.252 G12 0.451 GPa 5.81 GPa G13 0.451 GPa 5.81 GPa G23 0.305 GPa 5.69 GPa

k11=k33 594 W/m/K 0.861W/m/K k22 0.0 W/m/K 0.0 W/m/K c 1000 J/kg/K 1000 J/kg/K

Table 5 - Composite layup of flexible radiator winglet as used in the analysis of Concept A (SMA foil option adds a 6th ply).

Ply Material Depth (mm)

Angle

1 60% S2 Glass in 5250 Epoxy 0.127 90° 2 60% K1100 in 5250 Epoxy 0.127 45° 3 60% K1100 in 5250 Epoxy 0.127 0° 4 60% K1100 in 5250 Epoxy 0.127 45° 5 60% S2 Glass in 5250 Epoxy 0.127 90°

Table 6 - Material properties of the SMA wire as used in the analysis of Concept A.

Material Property

Value

ρ 6450 kg/m3 GA 70 GPa GM 30 GPa

Ms, Mf -14°C, -40°C As, Af 5°C, 32°C

CA, CM 6 MPa/K, 7MPa/K(@ 300 MPa)

Hmax 3.9% k 0.013/MPa

n1, n2, n3, n4 0.5, 0.5, 0.5, 0.5 k 22 W/m/K c 329 J/kg/K

Determininganappropriatecompositelayupwasamatterofiterativeanalysis.ThethermalteamsetakAminimumof0.152W‐m/K,whichismetusing0.64mmthicknessandatleasttwocarbonfiberlaminaealignedperpendiculartotheflowtube.Itwasdeterminedthatamoreeffectivethreecarbonfiberlaminaedesigncouldbeemployedwithoutsubstantiallaminafiberfailureiftwosuchlaminaewerealignedat45°anglesrelativetotheflowtube.ThefinallayupfortheConceptAanalysisconsideredhereinisgivenasdescribed

inTable8.Notethatfiber“Angle”isheregivenrelativetothecircumferentialdirectionofthecylindricalwinglet(i.e.,a90°isalignedwiththeflowtube).

ThepropertiesoftheSMAwirewereestimatedfromtheresponseofcommerciallyavailablepre‐trainedmaterial.Thetransformationtemperatures,however,wereartificiallyadjustedtomatchtherangerequiredforoperationof

theradiatoraccordingtothethermalanalysis.ThisreflectstheneedinlaterdevelopmentstudiestoproperlychooseandcharacterizetheappropriateSMAalloysystem.Notethatmaterialwiththerequiredtemperaturesisexpectedtobewidelyavailableduetotheprevalenceofmedical‐gradeNiTi,whichtransformsintherequiredtemperaturerange.ThepropertiesusedinthecurrentanalysisaregiveninTable10

Page10of15

V. Analysis Results Thethermalmodelindicatesthattheconceptturndownapproaches11.9and35.2foropenendandclosedendgeometry,whereclosedendreferstoashieldplacedoverbothendsofthecylinderformedbythecoldshape.Themodelfurthershowsthatanygapbetweentheedgesofeachsideofthefacesheetwheretheymeetovertheflowtubeinthecoldshapehasaninfluenceonthemaximumturndown.Figure13showstheturndownfromaperfectcoldshape,or0mmgap,to15mm(0.59in)gap.Theseturndownvaluesarecalculatedbyvaryingthecoldshaperelativetoahotshape

havingaradiusofcurvatureof1m.Ontheotherhand,asseeninFigure14,turndownisnotstronglysensitivetoaradiusofcurvatureinthehotshapegreaterthan~0.1m.Thevaluesforthisfigurearecalculatedbyvaryingthehotshapewithrespecttoacoldshapehavingnogap.Bothrelationshipsindicatethatemphasisneedstobeplacedprimarilyuponcoldshapedeformationandbiasspringstiffnessinthenextphaseofdesignwhilethecurvatureofthehightemperatureshapeisafairlyflexibledesignparameter.

Stressanddeformationanalysisshowsthatevenafteronlyashort(60s)periodofheating,thereconfigurableradiatorpanelcanbeseentoopensubstantially,withthemajorityoftheflatteningdeformationinducedbytheSMAconcentratedneartheflowtubeheatsource.ThiscanbeseeninFigure11.Thecontoursinthisfigureshowtheaxialstressintheglassfiberlaminaetransversetothefiberdirection,whichcorrespondstothelargeststresscomponentinducedbythebending.Thelimitsofthecontourlegendareequivalenttothecalculatedvaluesforaxialstressesatfailureinthisdirection.

Figure 9 Sensitivity to gap in the cold case. Shows sensitivity of turndown to the gap between the end of each side of the facesheet for this concept, with and without end shields.

Figure 10 Sensitivity to curvature for the hot case. Shows sensitivity in turndown to the hot case curvature of the facesheet. A curvature of 0.1 is approximately a half circle.

Page11of15

Ply 1(Tension critical)

Ply 5(Comp. critical)

-170 MPa ≤ σ22 ≤ 85 MPa -170 MPa ≤ σ22 ≤ 85 MPa

Figure 11 - Global heated deformation and local stress results (S22) in the glass fiber-based laminae (contour plot bounds correspond to stress limits).

SimilarresultscanalsobeseeninFigure12.Thecontoursinthisfigureshowtheaxialstressinthecarbonfiberlaminaealignedwiththefiberdirection,whichcorrespondstothelargeststresscomponentinducedbythebending.Asintheabove,thelimitsofthecontourlegendareequivalenttothecalculatedvaluesforaxialstressesatfailureinthisfiber‐aligneddirection.

-690 MPa ≤ σ11 ≤ 1800 MPa -690 MPa ≤ σ11 ≤ 1800 MPa -690 MPa ≤ σ11 ≤ 1800 MPa

Ply 2(Tension critical)

Ply 3

Ply 4(Comp. critical)

Figure 12 - Global heated deformation and local stress results (S11) in the carbon fiber-based laminae (contour plot bounds correspond to stress limits).

Benchtop Demonstration 

Twobenchtopdemonstrationswereperformedtoshowproofofconcept.ThefirstwasperformedatTexasA&MusinghightemperatureSMAmaterialandheatersonametalwingletstructureinordertodemonstratethedesiredactuation.ThesecondwasperformedattheJohnsonSpaceCenterusingSMAwithatransitiontemperatureinthetargetrangeandonascalecompositebasedwingletstructure.

Thedemonstrationsareintendedto:

Reflectasinglewingletsectionofthefullscalethermalmodelato Approximatelyhalf‐scaleatTexasA&Mlabo ApproximatelyfullscaleatJSClab

Demonstrateconceptactuationundero Labambientconditionsusingheaterinducedtemperaturechange

Page12of15

o Targetapplicationtransitiontemperatureusinglowtemperatureenvironmentchamber.

Besimpleinconfiguration, Basedonthermallyinducedtransformationstrainsgeneratedby

commerciallyavailable(i.e.,fullystabilized)SMAwire.TheTexasA&MdemowasfabricatedasschematicallydescribedinFigure14.Thebasicstructuralmaterialconsistedofstainlesssteelfoil(0.002inthick)ontowhichalayerofthermalgraphitesheet(0.004inthick)wasadheredforgreatlyincreasedthermalconductivity.Thetotalplanardimensionsofthetwo‐plylaminatewere5inlongby2inwide.Itwasdeterminedusingpreliminaryfiniteelementanalysisthattwo0.3mmdiameterwiresrecovering4%(contractile)strainwouldhavesufficientstrengthanddeformationtomorphthewingletasneeded.Thewireswereinstalledinan“x”configurationasshowninFigure14,theendofeachwireterminatingbypassingthroughseveralsmallholesattheendsofthesheet,whichprovidedsufficientfrictiontopreventwireslipping.Asmalldropletofsuperglueappliedtoeachholefurthersecuredthewire.Afterassembly,eachsheetwasrolledinthelongdirectiontoformanopencirclesectionwithadiameterof~1.6in.

Thefullthermaltestassemblyconsistedoftwomorphingwingletsattachedtoanaluminumrodwithasquarecross‐section(0.25inx0.25in),whichitselfwasheldbyabenchviceatoneendandwaswrappedwithresistiveheatingtapeontheother.Athermocouplewasinsertedatthecenterofeachofthetwowinglet/rodinterfaces(seeFigure15),thefirst(left)onebeingusedtocontroltheheateroutput,andthesecondprovidinginformationonanydisparitiesintemperaturesbetweenwinglets1and2.AnIRcamera(Testo885‐1ThermalImager)wasusedtoqualitativelyrecordbothlocalizedheatingandglobaldeformations.

2.0 in

5.0 in

SMA Wires (x2, D= 0.3mm)

Layup: -SS Foil (0.002 in) -Therm. Graphite (0.004 in) -SMA wires

Roll assemblyinto tube of

diameter = 1.6 in,SMA wire to outside

Figure 14 - A schematic showing the proof-of-concept morphing radiator winglet design

9 in 2.5 in

2 in (x2) 1 in

Winglet

1 Winglet 2

TC1* TC2

SMA Wires (D= 0.3mm)

Heater(4 in long)

Figure 13 - Test setup for the morphing radiator winglet demonstration

Page13of15

0

20

40

60

80

100

120

140

0 20 40 60 80 100

Temperature

(°C)

Time (min)

Set Point (Winglet 1) Winglet 1 Temp. Winglet 2 Temp.

Figure 15 - Setpoint and measured temperatures on the morphing radiator demonstration

Thedemonstrationteststeppedheatertemperaturefrom50°Cthrough140°Cin10°Cincrementsallowingwinglettemperaturestostabilizeateachincrement.After30minatthepeaktemperature,theheaterisremovedandthesystemallowedtocool.TheresultingwinglettemperaturescanbeseeninFigure22.Notethesubstantialdisparitybetweenthetemperaturesofthetwowingletsduetotheeffectivenessofthefirstintransferringheatenergyoutofthealuminumrodandintotheambientenvironment.

Thoughsimilarinmanyrespects,theJSCdemodiffersfromthatperformedatTexasA&Minthatitisintendedtoshowboththeplausibilityofthecompositematerialforuseinaflexiblefacesheetandactuationinthetargetapplicationtemperaturerange.Thewingletisfabricatedfroma5‐plycompositelayupsimilartotheconceptdesignbutwithasomewhatlargercoldshapediameterduetomoldmaterialsthatwereonhandtofabricatetheunsprungshape.Theshapememorywireusedhasalowtransitiontemperaturenearfreezing.

Inthistestthewingletandwirearesoakedatbelow‐20Cduringandfollowingassembly.Theassembledwingletisthenremovedfromthelowtemperatureenvironmentchamberandplacedonatableatroomtemperature.ThesequenceshowninFigure24depictstheactuationofthefacesheetfrombelowfreezingtoanearlyflathotshapethatspans12inches.Notethewhiteareasonthefacesheetindicatingfrostinthetopframe.Bythesecondframemostfrosthasmeltedbutsomeremainsuntilthefinalframe,whereuponmostoftheactuationhasoccurred.

Theactuationobservedinbothtestsisshowninduringthecontrolledheatingproceduresuccessfullydemonstratedthattheself‐morphingradiatorofconceptAwasinfactfeasibleatthescaleconsidered.ThermalandIRimagesattheinitialandfinalstateaswellasatthe100°Cand140°CsetpointsareshowninFigure17,andimagesatallsetpointsareavailable.Itisclearthatasthecenterofthefirstwingletreaches100°C,noticeablethermallyinducedmorphinghasoccurred,thoughthetemperatureofthesecondwingletisinsufficienttoinducedactuation.However,asthefirstwingletreaches~140°Candexhibitssubstantialdeformationtowardaflat(open)configuration,thesecondwinglethasbeguntoopenaswell.Uponremovaloftheheaterpower,thesystemcoolstowardroomtemperatureonceagainandthewingletsbegintorecovertheirinitialshape.

Page14of15

a) Initial Condition

b) Set Point: 100°C

c) Set Point: 140°C

d) Heater Removed

Figure 17 - A series of images showing the actuation of a proof-of-concept for the morphing radiator concept.

VI. Conclusions Auniqueconceptforhighturndownthermalcontrolofmannedspacevehiclesthatutilizesshapememoryalloyshasbeenexploredthroughanalysisandbenchtopprototypetesting.Resultsindicatethat:

Highthermalcontrolsystemturndownratiosof35.2:1arepredictedinalikelymissionscenariowithasimple,singlefluid‐loopthermalcontrolsystem.

Physicalturndownprovidedbytheradiatorcanapproach96:1. Turndowncapabilityismostsensitivetocold‐shapegeometrywheregapsin

thegeometryraisetheviewfactoroftheradiatingsurfacetospace. Limitedturndownsensitivitytohotshapeaffordsflexibilityincontinued

developmentofthetechnology Thedesiredactuationbehaviorhasbeenmodeledandtheconceptshown

capableofreproducingtherangeofactuationneededinresponsetofluidlooptemperaturevariationalone.

ThedesiredactuationbehaviorhasbeenreproducedinbenchtopprototypesoftheconcepttestedatTexasA&MUniversityandattheNASAJohnsonSpaceCenter.

Thestrengthofthistechnologyisitsentirelypassivebehaviorinresponsetochangingenvironmentandvehicleheatloads.Thisbehaviormeansthatthis

Figure 16 - low temperature with composite facesheet proof-of-concept demonstration

Page15of15

technologycanproduceveryhighturn‐downswhichenablesingle‐loopthermalcontrolwithoutactivecontrol,powerdraw,orinstrumentation.Byenablingsingleloopthermalcontrol,italsopromisestoreducethermalcontrolsystemlaunchmassbyapproximately25%.FortheOrionCEVthisamountstoaprojectedmassreductionofbetween139kgand225kg.

VII. References [1]Ochoa,D.,Vonau,W.,andEwert,M.,"AComparisonbetweenOne‐andTwo‐LoopATCSArchitecturesProposedforCEV,"SAEInt.J.Aerosp.4(1):344‐350,2011,doi:10.4271/2009‐01‐2458[2]Navarro,Moses,“LunarLander1‐Loopversus2‐LoopActiveThermalControlTradeStudy,”ESCGroup,ESCG‐4470‐07‐TEAN‐DOC‐0099,DeliveredtoNASAAug24,2007[23]JSC/ESCGguidance[24]Donabedian,M.,Gilmore,D.,et.al,“Insulation,”SpacecraftThermalControlHandbook,VolumeI:FundamentalTechnologies,2ndedition,editedbyD.Gilmore,2002,P165[25]RohsenowW.M.,etal.,HandbookofHeatTransfer,3rded,McGraw‐Hill,p.7.84[26]Incropera,F.&Dewitt,D.Fundamentalsofheatandmasstransfer,5thedition.JohnWileyandSons,2002.p.794.[27]RohsenowW.M.,etal.,HandbookofHeatTransfer,3rded,McGraw‐Hill,p.7.80.[28]ThermalDesktoppaper…[33]D.Hartl,D.Lagoudas,F.Calkins,“AdvancedMethodsfortheAnalysis,Design,andOptimizationofSMA‐BasedAerostructures,”SmartMaterialsandStructures,Vol.20,094006,2011.[34]D.Lagoudas,D.Hartl,Y.Chemisky,L.Machado,P.Popov,“ConstitutiveModelforPolycrystallineShapeMemoryAlloyswithSmoothTransformationSurfaces,”InternationalJournalofPlasticity,Vol.32–33,pp.155–183,2012.[35]D.Hartl,D.Lagoudas,“ConstitutiveModelingandStructuralAnalysisConsideringSimultaneousPhaseTransformationandPlasticYieldinShapeMemoryAlloys,”SmartMaterialsandStructures,Vol.18,No.10,2009.[36]SiegelandHowell,Thermalradiationheattransfer,4thedition,p.843