Performance of the relativity mission superfluid helium flight dewar

Available

online

at www.sciencedirect.com

Pergamon SCIENCE

www.elsevier.com/locate/asr

DIRECT*

doi: lO.l016/SO273-1177(03)00831-7

PERFORMANCE OF THE RELATIVITY HELIUM FLIGHT DEWAR

MISSION SUPERFLUID

R. T. Parmley,G. A. Bell, D. J. Frank, D. 0. Murray, and R. A. Whelan Lockheed Martin Advanced Technology Center, 3251 Hanover Street, Palo Alto, CA 94304, USA

ABSTRACT The world’s largestcapacity helium flight dewarhasbeenassembledfor useon the Relativity Mission, also known as Gravity Probe-B(GP-B). Acceptancetestsperformedinclude (1) weight, (2) proof pressureand leak checks,(3) vacuum bakeout,(4) main tank fill with He I, (5) parasiticheatrate tests,(6) well fill with He I from both the main tank and an externalsupply dewar,(7) well depletion, (8) conditioning the main tank to He II, (9) porousplug tests,(10) heatpulse metertests,(11) transferringHe II from the main tank to the well with a fountain-effectpump, (12) guard tank fill with He I with a nonventedHe II main tank simulating launch pad hold, and (13) guardtank emptying. The measuredperformanceis comparedto the previously launchedIRAS, COBE, and IS0 cryostats.The Relativity Mission spacecraftwill be launchedin the time span1999to 2OOO. 0 2003 Published by Elsevier Ltd on behalf of COSPAR.

INTRODUCTION The Relativity Mission is an orbital test of Einstein’s generaltheory of relativity using gyroscopes.The precessionof the gyroscopeswill measureboth the geodeticeffect (6.6 arcsec/year)through the curved space-time surroundingthe Earth and the motional effect (0.042arcsec/year)due to the rotating Earth draggingspace-timearoundwith it. To achievethe extraordinaryaccuraciesneededto measurethesesmall precessions,it is necessaryto have the gyroscopesoperatingin the following environments:a vacuumof <10-n tot-r,an accelerationlevel of
Q 2003 Published by Elsevier Printed in Great Britain 0273-l 177/$30.00 + 0.00

Ltd on behalf of COSPAR

R. T. Parmley

et al

DEWAR

Fig. 1. The probe houses the science instrument assembly and provides all its service cables and lines.

PAYLOAD ELECTRONIC BOXES NOT SHOWN

DRAG FREE PROOF MASS

DESCRIPTION

The dewar shown in Figure 3 is 3.0 m high, 2.2 m in diameter, and weighs 810 kg dry. The main tank contains 2328 liters of He II at 1.8 K, weighing 338 kg with a vapor ullage of 5%. The guard tank used to simplify ground servicing contains a maximum of 99 liters of He I at 4.2 K, which is reduced to 30 liters prior to launch by boiling off the liquid, using heaters. The guard tank allows the He II main tank to be kept nonvented for 3 months, while the guard tank only has to be filled once a week. Consequently, all of the helium from the main tank is available for use on orbit, as opposed to a system that pumps on the subatmospheric main tank right up until shortly before launch. The 30 liters left in the guard tank allows for two launch aborts plus launch separated by 24-hour intervals with no dewar servicing required.

The main tank is supported by 12 Fig. 2. The probe is inserted into a cold dewar through a helium-purge :d passive orbital disconnect strut air lock. (PODS-V) high-efficiency supports off the vacuum shell, while the guard tank is supported off the composite necktube. The y alumina necktube with a 0.03~mm 15 V-3 Cr-3 Al-3 Sn titanium liner provides a low-heat-leak, vacuumtight cylinder connecting the main tank to the vacuum shell. This cylinder allows the well to be filled with He I during lead bag expansions and probe insertion.

Fig. 3. Cutaway of dewar.

Four perforated 1100 aluminum face sheets with 5052 aluminumhoneycomb-core vapor-cooled shields plus double aluminized Mylar/silk net spacers provide thermal protection inside the evacuated vacuum shell.

Performance

of the GP-B

Helium

Flight

1409

Dewar

A secondsurfacemirror (FOSR) is installed on the vacuum shell to lower its temperaturein orbit. FOSR consistsof a 0.25mm Teflon film silverized on the back side. The silver provideslow solar absorptance while the Teflon providesthe high infraredemittance. The plumbing schematicis shown in Figure 4. The plumbing layout (1) allows the main tank and guard tank to be filled throughthe samebayonetandalsoto bepressuredrainedif required-the tanksarevented throughseparatelines, (2) allows the main tank to be nonventedwhile the ventedguardtank is toppedoff oncea week, (3) hasa bypassline that allows the transferline to be cooleddown prior to main tank servicing, (4) allows the well to be filled with He I from the main tank using heatersor from an externalsupply dewar, (5) allows filling the well with 12-torr gas or He II, if required,using a fountain-effect pump, (6) vents the tanks and lines to space,and (7) providesoverpressureprotection for all trappedvolumes (remotely actuatedcold valve RAV6B is closedonly on orbit, andthe probehasa burstdisc for well relief). A thermal analysisusinga vacuumshell temperatureof 225 K showsthat the major heatloads arefrom the probe (40%), multilayer insulation (220/o),dewar composite necktube(1l%), and dewar wires (ll%), mainly due to the sevenRAV cold valves. ACCEPTANCE TEST RESULTS Dewar Mass The measuredmassof the flight dewaris 810kg. This masscomparesto a calculatedvalueof 807.6kg just prior to the measurement.The massesof all componentswerecloselytrackedthroughoutthe conceptualand detaileddesignprocesses.Severalsignificantprogramdecisionscanbedirectly correlatedto theresultantlow TO THRUSTERS - INTERFACE WITH SPACECRAFT

1

Instellsd

Acontain

only if ths well will

He II during launch

EVACUATION, PURGE PLUMBINWANKS m

l

. rrrm

I

I CAP crep= mr:

VACUUM SHELL

Fig. 4. Plumbing schematic.

1:;

OF

OF MAIN TANK

DRAIN

WITH He I OR

OF MAIN TANK

l

FILL OF GUARD TANK WITH He I ;;A.E3 GUARD TANK WITH

.

;~XilEWN

l

OF BYPASS

m

FILL OF WELL WITH He I AT 4.5 FROM EXTERNAL SOURCE OR MAIN TANK

-

FILL OF WELL WlTH 12 tow GHe OR He II (IF REQUIRED) VENTlNG SPACE

OF TANKS,

OVERPRESSURE FOR AU CLOSED

LINES TO

PROTECTION VOLUMES

K

R. T. Parmley

et al.

mass of the dewar, the principal contributors being lightweighted main tank and vacuum shell components and honeycomb vapor cooled shields. Although selecting waffled domes and ring-stiffened cylindrical sections (Figure 5) and large curved honeycomb panels added 10% to dewar costs, these three areas significantly contributed to the lightweight dewar that has been produced. Table 1 illustrates the savings achieved in key areas.

Fig. 5. Light-weighted

aft vacuum shell.

Also of interest is that, since the weight savings are so significant, a more expensive and riskier option, using aluminum-lithium for the main tank and vacuum shell components, was deemed unnecessary.

Proof Pressure and Leak Checks All tanks, lines, and the vacuum shell were proof pressure tested and leak checked as shown in Table 2. All cold joints (welds, epoxy bonds, Helicoflex seals) were leak tested warm, cycled three times to 77 K and releak tested warm as the dewar was assembled. Following dewar assembly, the tanks and plumbing were leak checked as a system warm and at 4.2 K. No leaks were found. Vacuum Bakeout The dewar was rotated upside down and placed inside a customized, insulated forced air oven (Figure 6). The upside down orientation allowed any minute creep of the aft PODS gaps to be in a favorable direction. The dewar vacuum space was pumped down at a controlled rate to 1 torr in 29 hours while the oven was brought up to 3 17 K and held for 11 days. The temperature was increased to 322 K and held for an additional 6 days, based on the results of creep tests of heated PODS-V supports conducted in parallel. At the end of the bakeout period, the vacuum pressure read 5 x 1O-4 torr at 322 K and 5 x 1O-5 torr when cooled down to 294 K. Water vapor was the predominant gas left. Helium I Tests GSE modules (Figure 7) are connected to the well and vent line with plumbing connections are made to the dewar, to accomplish the following tasks.

lines at Bl, plus electrical

Main Tank Fill with Lia_uid He I. The main tank, guard tank and plumbing lines are evacuated and backfilled three times with GHe to 775 torr. The well, with radiation baffles installed, is evacuated. Supply dewars are connected to the fill-line bayonet B3. First the dewar bypass line is cooled through V13 and RAV2

I Assembly

Table 1. Weight Savings I I Flight Savings Weight (kg) (kg)

Comment

262

80

Waffle domes, ring-stiffened 22 19 aluminum

Guard Tank

15

4

Lightening

Vapor Cooled Shields

72

44

Perforated aluminum honeycomb shields

Vacuum Shell

345

100

Ring stiffened & waffle construction, 22 19 aluminum

Main Tank (including a 5 l-kg proton shield)

cylinder,

holes in support cone

Performance

of the GP-B

Helium

Table 2. Proof Pressure and Leak Checks

Main Tank

0.39

0.25

< 10-g

Well

0.20

0.39

< 10-g

Guard Tank

0.56

0.25

< 10-g

Plumbing

0.56

0.25

< 10-g

Flight

Dewar

1411

with cold gas. RAVl and RAV3 are opened and RAV2 is closed to redirect the cold gas to the main tank; the cooldown rate did not exceed 50 whour. Once the tank temperature is below 5 K, the delta supply pressure is increased to 0.03 MPa to start liquid transfer. After 6 hours, the main tank was 25% full. In later test sequences, the main tank was filled to 97%. The vacuum pressure dropped into the low 10-e torr range with the main tank at 4.2 K.

Well Fill with Liauid He I. Expansion of lead bags in the well to achieve 10-7 gauss * Air equivalent leak rate is 6 x lo-* cc/s. Would deposit 10 A ice magnetic field at the SIA plus insertion or removal of the probe requires the well to be coating on main tank if cold 4 years; no impact on lifetime. filled with 100 liters of He I. Vacuum Shell

0.25

0.10

1 x 10-T”

The simplest method is to transfer helium internally from the main tank to the well. To achieve this transfer, the main tank vent is closed, valve RAV7 is opened (valve RAVBB is always open on the ground) and a 50 watt heater is turned on to pressurize the tank to 785 tom When the well is vented, liquid transfer occurs. Liquid level sensors in the top of the well determine when the well is filled. The transfer rate attained was 0.3 liters/mm. Liquid helium was also transferred to the well from an external supply dewar through V13 and RAVS. Transfer rate was 3 liters/mm. Removal of liquid helium from the well after probe insertion is accomplished by pumping on the well with a Rootes blower/roughing pump system. This operation was demonstrated without the probe installed. Parasitic Heat Rate Test. The parasitic heat rate test is used to establish a baseline performance datum and to validate the thermal mathematical model for the dewar. It is performed with a set of radiation baffles

Fig. 6. Dewar bakeout enclosure with hot-air circulation system attached.

Fig. 7. Dewar with gas module, electrical module, and data acquisition system module.

1412

R. T. Parmley

et al.

installedin the dewarnecktube in placeof the probe,which effectively reducesthe radiationload to on the order of 11mW. This resultant value, when combined with the dewar/probethermal model, gives an updatedlifetime prediction. This test was performedwith the main tank filled to 24% with normal boiling point liquid helium and the well andguardtank evacuated. The parasitic heat rate was measuredafter two weeks with a wet test meter (1% accuracy).The tank temperatureremained stable to 4.280 + 0.0005 K over a 20 hour period with a tank pressureof 776 torr. The remaining temperatureswere 30, 55, 98, 170, and 294 K for the four vapor-cooled shields plus vacuum shell respectively.The vent rate was 0.040 f 0.0016 l/s (6.6 mg/s) which translatesinto a heat rate of 136 mW. This measuredvalue fell within the 30% margin addedon to the calculated heat rate for the dewar test configuration, even if zero radiation was assumedto be coming down the dewar neck. Additional parasitic heat rate tests in 1997 with probe B installed will give a more accurate assessmentof the payload lifetime and hopefully will allow the 30% margin to be reducedin the thermal model. The dewar with the probe meets the orbital lifetime requirementsof >16.5 months. Helium II Tests Conditioninethe Main Tank. Following the He I tests,the main tank was conditionedto He II by pumping on bayonetB 1 with RAV3 open.The tank startedat -85% full at 4.28 K and was pumpeddown to 1.58K and 51% volume in 6 days using a Rootesblower/roughingpump system.The purposeof the test is to demonstratethat the bath temperaturecould be reducedto 1.6 K. The multiple topoffs requiredwith hpoint liquid from supply dewars,to bring the fill level back to 95 to 98%, will be demonstratedlater,prior to the cold vibration testwith probeB installed. PorousPlug Tests.Operationof this porousplug is unique in that it is usedto control the SPHe temperature,operateover a wide rangeof flowrate, andprovide gasfor spacecraftattitudecontrol. Severalsintered stainlessplugs were evaluatedto selectthe oneusedin the dewar(Frank,Yuan 1996).Data weretakenduring normal and chokedflow conditions.The primary requirementsthat affectedthe design of the porous plug wereto operateover a rangeof 4-16 mg/s without choking at 1.8 K, andoperatewith a sufficient back pressureso that the ventedgascan be usedfor the spacecraftattitudecontrol system.The predictednominal flowrate during the sciencephaseof the mission is 6.8 mg/s. The higher flowrates arerequiredduring the early phaseor orbit trim. I Theserequirementsmake the porousplug unique in that the downstreamflow conductanceof the plumbing is not constantas on other spacebornedewars.A heaterin the tank and variable flow conductance valves in the 16 thrustersprovides the temperaturecontrol and flowrates required for attitude control. Increasesin venting areachievedby decreasingthe downstreampressureof the porousplug andmaintaining thebath temperatureby turning on a heater. The plug uses 316L sinteredstainlesssteel porousmedia purchasedfrom Mott Metallurgical Corp. The materialhad a permeability of 3.8 x lo-lo cmz, a bubblepoint of 19.3cm Hg, a thicknessof 0.635cm, and a diameterof 6.9 cm. As the flowrate is increasedby decreasingthe back pressure,the temperaturegradientincreasesuntil the plug achievesa choked condition. The flowrate at this point is referredto as the critical flowrate.As the flowrate is increasedandsubsequentlydecreased,the temperaturegradientgoesthrougha hysteresisloop. Figure 8 showstypical temperaturegradientand back pressureprofiles for variousbath temperatures.The backpressureis measuredin the groundsupportequipmentof the dewar.As expected,the slopeof flowrate versustemperaturedropsin the normal (linear) mode andthe critical flowrate increaseswith bath temperature.Testsat variousinlet hydrostaticheadswereperformedandthe resultswereextrapolatedto zero head

Performance

of the GP-B Helium

Flight

Dewar

14

0

0 0

5

10

15

20

25

30

35

Temp Drop (mK)

40

45

0

5

10 15 20 Flowrate (mgh)

25

Fig. 8. Effect of bathtemperatureon temperaturedrop andbackpressure(0 = 1.7K,O = 1.9K). to determinethe critical flowrateon orbit. Figure 9 showsthat the plug meetsthe requiredflight maximum of 16 mg/s when operatingat 1.8K. Heat PulseMeter Test.The heatpulsemeteris a methodof determiningthe quantity of liquid in a tank by observingthe temperaturerise of the bulk liquid following a shortpulseof heaterpower. For this test, the dewar was vertical and the porousplug was not operating;the main tank pressurewas controlled by pumping via V9 through the vent line. The temperatureof liquid in the main tank was decreasingat 1 mK/hour. The temperatureof the liquid helium is monitoredbefore and after a 12.8watt heaterinput over a 7-minuteon time. The temperatureoffset causedby the energyinput is thus determined and can be used to calculate the amount of liquid helium. With an energyinput of 5800joules and an observedtemperaturerise of 16.1mK, the calculatedmass is 171kg. The referencemass was 167kg, basedon liquid level sensors.The accuracyof the calculatedmassby the heatpulsemethodis affectedby the measurementsof heaterpower,time, and temperaturesplus the stratificationof the vapor in 1 g. The accuracyof the liquid-level sensormeasurementis affectedby the calculatedvolume versusmeasuredliquid level in the main tank, the liquid volume to mass conversion based on temperature/density data, plus the accuracyof theliquid-level sensoritself.

po

For flight operations,33 watts of power will be used.For a full tank, the pulsedurationwill be approximately2.5 minutes, The control system heaterswill be off when pre and post dataaretakento determinethe naturaldrift of the fluid temperature.

(I) 25 H g 20 ii

0

2

4

6

Hydrostatic

8

10

12

Head (cm)

Fig. 9. Effect of hydrostaticheadon critical flowrate(A= 1.7K, 0 = 1.8K, V = 1.9K).

FountainEffect Puma (FEP) Test.The fountain effect pump (FEP) is designedto fill the well with He II from the main tank just prior to launch. This procedureis not a baseline but will be usedonly if the qualification vibration of dewar and probeshow that the lead bagtransition temperaturehas beenexceededwith the well evacuated. The test entailed the operationof the FEP and demonstration of a nominal transferrate of 15 liters in 3 hours. The

1414

R. T. Parmley

et al.

dewar was tilted 60-degoff vertical so that the 50% level of liquid helium coveredthe FEP inlet, The FEP heaterwas then operatedat various levels up to 0.9 watt and the liquid transfer monitored. A total of 24 liters was transferredin close to 6 hours operating time. The maximum transfer rate was 5.9 liter/hour, which meetsthe prelaunchoperatingrequirements.The 50% full main tank temperatureof 1.9K roseat a rate of 0.013 K/hour during the transfer. Guard Tank Fill, Hold, and Deuletion. At VandenbergAir Force Base, Bldg 1610,the main tank will be filled, topped off and conditioned to 95 to 98% with He II at 1.6 K. The main tank is valved off and the guardtank is filled with He I. Once the spacecraftis transferredto the launchpad, the guardtank is topped off oncea week; 24 hours before launch,the guardtank is depletedto 30 liters using heaters.Tests to partially demonstratetheseconditions were performed. The guard tank fill to 100% was accomplishedthrough V13 and RAV2 using a self-pressurizedexternal supply dewar.The entire operation, including line connections,purging, fill, line disconnects,and leak checkstook less than 4 hours.The dewar was in the following condition during guard tank servicing.The well is evacuatedand filled with radiation baffles.The main tank is 50% full of He II at 1.80 K; during the guard tank fill procedure,the bulk He II temperaturein the main tank rose 0.003 K; following guardtank fill, the main tank temperatureroseat a rateof 0.002 K/day. The He I in the guardtank boiled off at the rate of S%/day.The latter rate is probably too optimistic due to nonequilibrium conditions during the 42-hour test period. The coldest vapor-cooledshield temperaturehad stabilized at 4.24 K at the guard tank and 5.04 K at the bottom of the shield. A 4.6 watt heateremptied the guardtank at the rate of 6 liters/hour. PERFORMANCE COMPARISON OF GP-B To PREVIOUS FLIGHT CRYOSTATS Performancecomparisonswere madein threemajor areas. Figure 10 shows the 30% weight savings achievedcomparedto the similarly sized IS0 cryostat. Even larger savingsareachievedif you eliminate the 5 1 kg protonshield deadweight. The thermal performanceimprovement of the GP-B cryostat with the SIA and probeinstalled (Table 3) is due to the use of passiveorbital disconnect struts (PODS-V), the carefully tailored double aluminized Mylar radiation shields (149 each),the silk net low-conductancespacers(3 eachat the tank reducedprogressivelyto 1 eachat the vacuum shell) plus the useof 4 vapor-cooledshields. The remarkablegroundhold performanceof 8 keeping the GP-B main tank nonventedfor 90 days (Table4) is due to the guard tank driving the inner vapor-cooled shield to 4.2 K (top) to 5 K (bottom); the main tank temperaturerises from 1.6 to 1.85 K in this time period. The He I guard tank is filled once a week and no helium is lost from the main tank. Both IRAS and COBE pumped on the main tank up to 26 hours before launch. Consequently,they had less helium for usein orbit. 01

I

I

I

I

I

I

0

50

100

150

200

250

300

6278/Fig 10

He II Weight

CONCLUSIONS 350

(kg)

Figure 10. Dry weight / He II weight ratio comparisonof dewars.

The Relativity Mission dewar acceptance test results show that the dewar meets all GP-B requirements. The use of advanced

Performance

of the GP-B Helium

Flight

Dewar

Table 3. Thermal PerformanceComparisonof He II Flight Cryostats with ScienceInstrumentsTurnedOff (TH = 294 K)

1. 2. 3. 4.

Guardtank not used Dewar boiloff measured:Probe/SIAheatleak calculated.30% margin Aperturecover actively cooledto 5 K with separateHe I flow Aperturecover cooledto 5 K with separatehelium bottle. Refilled every4.5 days

1415

technology in weight reduction techniques,PODS-V supmultilayer insulation, pm vapor-cooled shields, and guard tank design make this dewar the lightest, per unit massof He II stored,the most thermally efficient, and the easiestto service at the launch site (compared to the three previous free flight cryostats that have been launched, IRAS, COBE, and ISO).

Table 4. CryostatsGroundHold Performanceat the LaunchPad Helium Loaded (kg)

Cryostat

GuardTank Percentage Ullage Used? Fill Cycle, Days

Initial Orbital Vent Rate (Comparedto orbit equilibrium value)

Main Tank Nonvented Time (days)

IRAS

72

7

No

-

Above

1.1

COBE

89

7

No

-

Above

1.1

IS0

331

Above

Up to 4.2

Below

up to 90

I GP-B2

I

338

1. Mounted on main tank. Guardtank vapor coolsinner vapor cooledshield to -30 K 2. Basedon dewaracceptancetestdata.Includescalculatedprobeheatleak with 30% margin 3. Guardtank, inner vaporcooled shieldrun at 4.2-5 K

ACKNOWLEDGMENTS The work was performed for StanfordUniversity undersubcontractPR8709 and contract NAS 8-39225 with the NASA Marshall SpaceFlight Center.Bill Fosteris the responsiblemanager,Dick Parmley developedthe dewarconcept,Ed Caveyis the chief designer,DeanReadis the responsibleequipmentengineer, GregBelI is the testconductor,Dave Murray, with JackMix, DanWelsh,Bruce Imai, PaulAyres and Steve VanKeulen,conductedthe testson a 24 hour, 7 day a week basis.Dave Frank performedthe porousplug testsandRich Whelanperformedthe dewarweighing.Dave Doneganis responsiblefor the groundsupport equipment,Art Nakashimais the dewarsystemsengineer,andLarry Reynoldsis the quality engineer.

1416

R. T. Parmley

et al.

REFERENCES Frank, D.J. and S.W.K. Yuan, Experimental Results of Tests Performed on Superfluid Helium Phase Separatorsfor the Relativity Mission, Advancesin CryogenicEngineering,41, 11951201 (1996). Seidel, A. and Th. Passvogel, The IS0 Cryostat; Its Qualification Status, Advances in Cryogenic Engineering,37B, 1327-1340(1991). Seidel, A., Private Communication on IS0 (1995). Shirron, P., Private Communication on COBE (1996). Urbach, A.R., P.V. Mason and W.F. Brooks, ProgressReport on the Infrared Astronomical Satellite Cryogenic System,Advancesin CryogenicEngineering,27,1039-1047 (1982). Urbach, A.R. and P.V. Mason, IRAS Cryogenic System Flight Performance Report, Advances in CryogenicEngineering,29,65 1-659(1983).