3 1176 00166 6792 NASA-TM-8153919800018861 NASA TechnicalMemorandum81539 COMPOSITE WALL CONCEPT FOR HIGH TEMPERATURE TURBINE SHROUDS--HEAT TRANSFER ANALYSIS Francis S. Stepka and Lawrence P. Ludwig Lew& Research Center Cleveland, Ohio t df th '_-- Prepare or e ..,,:_J, .,.,, _, Aerospace Congress ,....:.:. sponsored by the Society of Automotive Engineers ': ._ Los Angeles, California, October 13-16, 1980 _ https://ntrs.nasa.gov/search.jsp?R=19800018861 2018-04-24T06:56:58+00:00Z
Transcript
3 1176 00166 6792
NASA-TM-8153919800018861
NASA TechnicalMemorandum81539
COMPOSITE WALL CONCEPTFOR HIGH TEMPERATURETURBINE SHROUDS--HEATTRANSFER ANALYSIS
Francis S. Stepka and Lawrence P. LudwigLew& Research CenterCleveland, Ohio
t
d f th '_--Prepare or e ..,,:_J, .,.,,
_, Aerospace Congress ,....:.:.sponsored by the Society of Automotive Engineers ':
A heat transfer analysis was made of acomposite wall shroud consisting of a ceram-ic thermal barrier layer bonded to a porousmetal layer which, in turn, is bondeO to ametal base. The porous metal layer servesto mitigate the strain differences betweenthe ceramic and the metal base. Variouscombinations of ceramic and porous metallayer thicknesses and of porous metal densi-ties and thermal conductivities were inves-tigated to determine the layer thicknessesrequired to maintain a limiting temperaturein the porous metal layer. Analysis showeothat the composite wall offered significantair cooling flow reductions compared to anall-impingement air-cooled all-metal shroud.
Stepka and Ludwig
THE WALLSHROUDSabove high pressure turbineblades of current aircraft gas turbine en-gines are metallic structures and need to becooled to reduce thermal distortions, crack-ing, and oxidation. This cooling is ob-tained by air bled from the engine compres-sor. Since the bleeding of the air from thecompressor reduces engine cycle efficiency,reduction of the cooling air requirement isdesired.
The use of a layer of ceramic on thehot gas-side of the shroud structure, asstudied in references (i to 3)* can signifi-cantly reduce the coolant flow requirementand/or reduce the shroud metal tempera-tures. Wear measurement of conventionalshrouds of large, high bypass turbofan en-gines indicate that local metal removalcaused by turbine blade rub is generally inthe range of 0.76 mm(0.030 in.) deep;therefore, in a composite shroud, the ceram-ic layer thickness must be at least 0.76 mm(0.030 in.) to preclude exposure of themetal support structure by blade rubs.
Several approaches have been investiga-ted in regard to adherence of thermallysprayed ceramics to the metal support struc-ture. The simplest approach is to thermallyspray ceramics directly on the solid metalsupport structure. However, data in refer-ence (3) indicate that the adherence of ce-ramics is poor when thick ceramic layers inthe range of 22 mm(0.08 in.) are sprayeddirectly on the metal substrate. Two possi-ble reasons for poor adherence are (1) largethermal stresses through the thick ceramiclayers under transient operation and (2)high stress due to abrupt change in materialthermal expansion at the ceramic/metal in-terface. These effects are mitigated andthe adherence of a thick ceramic layer isenhanced by employing graded ceramic/metallayers (4) in which the thermal expansioncoefficient is tailored by changing the per-centage of metal in the intermediate lay-ers. However, reference (4) reports thatexcessive stresses exist in the ceramic toplayer and a means to build in beneficialresidual stresses is needed. Another ap-
proach, the one of concern in this paper, is Stepka and Ludwigto use a compliant, generally low densityand low modulus, interlayer between the ce-ramic and metal base. In this composite 2wall concept, the compliant layer acts tomitigate the strain difference between the
*Numbers in parentheses designateReferences at end of paper.
ceramic layer and the metal base. Possibleinterlayer materials are various types ofporous metals such as felt, woven and foammetals. Thermal shock studies (5) revealedthat this compliant concept is more effective in reducing thermal stresses than thegraded layer concept.
The objective of the study reportedherein was to analytically examine the variables that affect the design of a compositewall shroud consisting of a metal base, aninterlayer of porous metal and an outerlayer of yttria stabilized zirconia.
Based on considerations of low oxidation and long life, the maximum allowabletemperatures of the porous metal of the composite shroud, and the all-metal shroud wereset at 1144 K (16000 F) for current materi~ls and 1200 K (17000 F) for advancedmaterials. Both the composite and all~metal
shrouds were assumed to only be impingementair cooled. The gas and coolant conditionsassumed were those of an advanced gas turbine.
The overall thickness of the compositewall shroud was kept the same as on typicalall-metal shroud (5.59 mm (0.220 in.)). Thesolid metal wall thickness of the compositeshroud was held constant at 2.03 mm (0.080in.) to maintain structural integrity. Thevariables investigated for the compositeshroud were (1) ceramic thicknesses from 0.5to 3.06 mm (0.02 to 0.12 in.), (2) corresponding porous-metal thicknesses from 3.06to 0.5 mm (0.12 to 0.02 in.), (3) porousmetal density from 10 to 50 percent of afully dense material, and (4) two porousmetals with thermal conductivities that differed by a factor as much as 6, and (5) ratios of cooling airflow to turbine gas flowfrom near zero to 0.03.
The data are presented as curves oftemperatures through the composite shroudfor various coolant- to gas-flow ratios forselected thicknesses of layers and for various porous-metal densities. Comparisons aremade between the various composite shroudcombinations and the all- metal shroud.
SYMBOLS
AAc
parameter, see eq. (4)impingement cooling airflow
areablade chordconstantsdiameter of impingement
cooling holes
Stepka and Ludwig
3
dhKmNuPrRRcReTVWXc
z
Subscripts:
bcgi
oPplp2wx
porous metal relative densityheat transfer coefficientthermal conductivityexponent, see eqs. (3) and (6)Nusse It numberPrandtl numberthermal conductance, KITcoolant- to gas-flow ratioReynolds numbertemperatureve loc ityfl ow ratedistance between impingement
see eq. (1)outside (toward gas side)porous metalporous-metal type 1porous-metal type 2metal support structureflow rate, cross flow air
ANALYSIS AND CONDITIONS
HEAT BALANCE - An element of the composite turbine shroud, shown in Fig. 1, wasanalyzed for the assumed engine conditionsand geometry shown in Table 1. Radiationwas neglected and heat flow was assumed onedimensional through the ceramic thermalbarrier, the porous-metal interlayer, andthe metal wall. The effective gas and cooling air temperature Tg and Tc wereassumed equal to their respective total temperatures. For these assumptions the heatflow equations are
hg(Tg - Tbo) = Rb(Tbo - Tbi) = Rp(Tpo - Tpi)
= Rw(Two - Twi) = hc(Twi - Tc) (1)
This set of equations was used in a computerprogram to calculate the desired surface andinterface temperatures. Equations (1)required gas- to-surface and surface-to-
Stepka and Ludwig
4
(7)
coolant heat transfer coefficients. Thegas-to-surface heat transfer coefficient wasobtained from the equation of heat transferreported in reference (6). This equation,developed from experimental heat transferstudies on a turbine shroud, is
NUg = hgC/Kg = 0.052 RegO.8(l - 20°. 8) (2)
where 6 is the tip-clearance to blade-spanratio. The gas-side Reynolds number Rewas evaluated at an assumed average gas ~achnumber of 0.8, the characteristic dimensionof the blade chord, and the gas propertiesnear the shroud surface. The transport gasproperties were obtained from the data ofreference (7).
Impingement cooling of the metal wallwas assumed. For this cooling method, thecoolant-side heat transfer coefficient wasobtained from the correlation from reference(8), which in the notation of this report is
Nu = h D /K = AS Rem PrO•33(Z/D )0.091 (3)c . c c c c c cwhere, for an assumed 3000 < Re c < 30 000,
A = eXP[0.026(Xc/Dc)2 - 0.8259(Xc/Dc) - 0.398~
(4 )
B = 1{1 + 0.4696(WxI/WcDdo.965] (5)
m= -0.00252(Xc/Dc)2 + 0.06849(Xc/Dc ) + 0.50699
(6 )
The following fixed geometry values wereassumed for the previous equations: ratioof jet-to-wall distance to hole diameterIIDc = 15, a ratio of hole spacing-todiameter (Xc/Dc) = 10, a ratio of crossflowto jet-flow (Wx/Wc) = 1.0, and a hole diameter Dc = 0.51 mm (0.02 in.). Substituting these values into equations (4) to (6)and then into equation (3) along with appropriate coolant properties values from reference (7) results in
hc
= c ReO. 941 c
wherec1 = 3.73x10-2 W/m2/K
The coolant Reynolds number was determinedfrom the following equation:
Stepka and Ludwig
5
Re c = WcDc/Ac~c
where for the given values of the coolanttemperature, assumed geometry of theimpingement holes, diameter of the turbineshroud, and shroud width gives
wherec2 = 2.0x104 sec/kg
(8)
(9)
This equation is then substituted into equation (7), which for assumed turbine gas flowrate and values of coolant- to gas-flowratio Rc, and provides the needed valuesof coolant-side heat transfer coefficients.
COMPOSITE SHROUD CONFIGURATIONS ANDMATERIALS - Fig. 1 depicts the compositeshroud configuration consisting of a metalsupport, an interlayer of porous metal, anda sprayed ceramic layer which is exposed tothe turbine gas flow. The overall radialthickness of the shroud was 5.59 mm (0.220in.) and was based on the consideration ofreplacing a specified all-metal shroud withceramic composite shrouds. Analysis of therelative structural strengths of the shroudswas not made since it was considered outsidethe scope of the present paper. The raaialthickness of the metal support base was selected to be 2.03 mm (0.080 in.) for allconfigurations. Therefore the combined radial thickness of the ceramic and porousmetal layers was 3.56 mm (0.140 in.) for allconfigurations. The thicknesses of the ceramic and porous-metal layers were variedand the temperatures and cooling flow requirements were determined. The porousmetal thickness was varied from 0.5 to 3.06mm (0.02 to 0.12 in.) which corresponds to aceramic thickness variation of 3.06 to 0.5mm (0.12 to 0.02 in.) (see Table 2). Thedensity of the porous-metal interlayer wasalso an independent variable in the studyand was varied from 0.1 to 0.5 of solidmetal.
The values of thermal conductivityneeded for the conductance terms in equation(1) for the metal wall (MAR-M-509) and theceramics (yttria stabilized zirconia) wereobtained from references (9) and (10), respectively. Two different porous metalscomposed of the same materials (FeNiCrA1Y)were considered; they differed in structureand thermal conductivity. Porous-metal 1was a felt-type material and porous-metal 2was an open-cell foam-type material. The
Stepka and Ludwig
6
thermal conductivity for porous-metal type 1was obtained from the following equationwhich was fitted to the data of reference(11) for various relative densities dpand average temperatures Tp:
-41 (5.4x10 T)
Kpl = 4.5 dp•38 e P W/m/K (10)
since there was no empirical data for thethermal conductivity for porous metal 2, itwas assumed that the thermal conductivitywas a linear function of material density.The published thermal conductivity data for100 percent dense FeNiCrA1Y alloy (12} wasadjusted for density by the following equation:
Kp2 = dp(6.38 + 0.018 Tp) W/m/K (11)
RESULTS AND DISCUSSION
The results of the heat transfer analysis of composite shroud designs and comparisons with an all-metal shroud (where bothshrouds were impingement air-cooled) arepresented in Figs. 2 to 4. The compositeshroud is illustrated in Fig. 1, and theassumed engine conditions and geometry aregiven in Table 1.
COMPARISON OF COMPOSITE AND ALL-METALSHROUDS - The calculated results for theall-metal shroud for coolant- to gas flowratios as high as 0.06 are shown in Fig.2(a). The figure shows little reduction inmetal temperature with increasing coolantflow ratio. The calculations showed thatthe specified maximum metal temperature of1144 K (16000 F) could not be obtainedeven with a coolant- to gas-flow ratio ashigh as 0.10. Therefore, for the conditionsof the analysis, impingement cooling aloneis not suffic~ent or practical for the allmetal shroud. The impingement cooling wouldneed to be supplemented by film-cooling, orthe shroud would require low thermal conductivity rub material on the gas side to reduce the heat flux, metal temperatures, andcooling flow ratio.
By way of comparison, Fig. 2(b) showsthe temperatures in the composite shroud.The composite shroud in Fig. 2(b) consistedof 1.78 mm thickness of ceramic and a 1.78mm thickness of porous-metal 2 with a 0.2density. The metal temperatures for thecomposite shroud are significantly lowerthan for the all-metal shroud at the same or
Stepka and Ludwig
7
lower coolantflow ratios. Data in Fig.2(b) for the compositeshroud indicatethatthe maximum allowablemetal temperatureisreachedat a coolant- to gas-flowratio of0.02. The data in Fig. 2(b) show the verylarge insulativeeffect of the ceramic. Forexample, at a coolant-to gas-flowratio of0.02 the temperaturedrop throughthe ceram-ic layer is 426 K (7660 F). Another ob-servationis that the combined insulativeeffect of the ceramic layer and porous-metallayer causes the gas-sideceramicsurfacetemperatureto be very near the gas tempera-ture.
EFFECTSOF POROUS-METALDENSITYANDTHERMALCONDUCTIVITY- Fig. 3 shows the tem-peraturesof compositeshrouds (ceramicandporous-metallayerseach 1.78 mm thick) as afunctionof coolingflow ratio for porous-metal types I and 2 (with differentdensitiesand with inherentdifferencesin
thermalconductivities). For the two densi-ties and for the range of porous-metaltem-peraturesin Fig. 3, it can be determinedfrom equations(10) and (11) that porous-metal 2 has a thermalconductivityfour tosix times higher than porous-metal1. For adensityof 0.2 and a temperatureof 1144 K(16000F), porous-metal2 has a thermalconductivity5.9 higher than porous-metal1.
Fig. 3 shows the effect of increasingthe densitiesof the porousmaterialsfrom0.2 to 0.5. As density is increased,theporous-metalgas-sidetemperatureis re-duced; this occurs becausethermalconduc-tivity increaseswith density,thus result-ing in increasedheat transferfrom the por-ous metal. The beneficialeffect of thegreaterthermalconductivityof porous-metal2 compared to porous-metal1 is apparentwhen comparingFig. 3. Fig. 3 also showsthat the coolant- to gas-flowratio requiredto obtain the 1144 K (16000F) maximumporous-metaltemperatureis 0.02 for porous-metal 2 with a 0.2 density. However,porous-metalI with a 0.2 densitycould notbe cooled to this temperatureeven with veryhigh coolantflows, if the densityofporous-metal1 was increasedto 0.5, the
maximum allowableporous-metaltemperature Stepka and Ludwigcould be obtainedat a coolantflow ratio of0.024.
In the design of a compositeshroud, 8the configurationparametersincludethelayer thicknessesas well as the porousmetal densityand thermalconductivity. Theporous-metaldensity,layer thickness,andstructure(felt,foam, or woven), in addi-
tion to their effect on heat flow and shroudtemperatures,are factorswhich will affectthe stress,adherence,and durabilityof thecompositewall. Although stress analysisisnot addressedin this paper, a genera]ob-servationfrom thermalstress and cycliclife considerationsis that lower porous-metal densitieswith their associatedlowermodulus of elasticityare desired. Fromaconsiderationof handlingand structuralintegrity,a lower limit on the densityofthe porouswall is assumedto be about 0.2.
EFFECTSOF LAYER THICKNESS- Fig. 4presentsthe resultsof the analysisin aform which shows how porous-metaltempera-ture varieswith changes in thicknessesofthe ceramicand porousmetal layersfor agiven coolantflow ratio.
The effect of the differentthermalconductivitiesof the two differentporousmaterialsis apparentfrom Fig. 4. A gener-al observationis that, as the ceramic layerthicknessdecreases,the porous-metalden-sity must increaseto maintain a given tem-perature. This is due to the fact that ahigher thermalconductivityin the porous-metal,is requiredto accommodatethe in-creasedheat flux and to maintain the allow-able temperaturelimitson the porousmetalat 1144 K (1600° F) for currentmaterialsor 1200 K (17000F) for advancedmaterials.
Fig. 4 shows that the higher theallowableporous-metaltemperature(1200 K(17000F)) is, the lower the porous-metaldensitiesat a given ceramiclporous-metalthicknessratio can be. InspectionofFig. 4 also revealsthat the porous layertemperaturedecreaseswith decreasingporouslayer thicknessand increasingceramiclayerthicknessfor a constantporous layer den-sity. As the porous-metaldensity is in-creased,the porous-metaltemperatureisdecreasedfor given ceramicand porous-metalthicknesses. As an example,for a 0.2 den-sity and a 1144 K (1600° F) temperaturelimit,porous-metalI would need to be about0.91 mm (0.036 in.) thick with a layer ofceramic2.64 mm (0.104 in.) thick to satisfythe heat load and coolingconditionsselec-
ted. On the other hand, porous-metal2 with Stepka and Ludwigthe same densitywould be 1.78 mm (0.07 in.)thick with a 1.78 mm (0.07 in.) thick layerof ceramic. The choice,as stated in theprevioussection,would be influencedby 9thermalstress considerationswhich dictatea selectionof the lowermodulusporouslayer,that is, porous-metal 2 with
the thickerporousmetal layersand thecorrespondinglythinnerceramic layer.
SUMMARYOF RESULTS
The analysisprovideda basis for eval-uating the effects of variableson the de-sign of compositeturbineshroudswhich con-sisted of the metal case wall, an interlayerof a porousmetal, and an outer layer ofceramic (yttriastabilizedzirconia). Theresultswere as follows:
1. Significantreductionsin the cool-ing-airto gas-flowratio are indicatedforthe compositeshroudscomparedto an all-metal shroud that was only impingementaircooled. This is based on the same maximumallowabletemperaturefor the all-metalshroud and for the porousmetal interlayerof the compositeshroud.
2. The good insulatingpropertiesofthe ceramicsignificantlyreducedthe tem-peraturesof the porousmetal and supportwall, but also caused the gas-sidesurfacetemperatureof the ceramicto be essentiallyat the gas temperature.
3. For a given porous metal densityandcoolant-to gas-flowratio, decreasingthethicknessof the porous-metaland in-creasingceramicthicknessresulted in lowersupportwall temperatures.
4. To maintain given allowableinter-layer temperaturesand coolant- to gas-flowratios,porous-metaldensityor thermalconductivitymust increaseas the ratio ofthe thicknessof the ceramic-to-porousmetaldecreases.
CONCLUDINGREMARKS
In general,thermalcycle lifeconsid-erations (refs.5 and 13) would indicatemore compliant,lower densitiesof porousmetal for the compositewall designs. Alsothe porous layer thicknessmust be largeenough to providethe needed strain isola-tion betweenthe ceramic layer and the metalbase. Thereforea high thermalconductivityat a low densityin the porous layer isneeded. Based on the foregoingand an as-sumed lower limiton porous-metaldensity Stepkaand Ludwig(consideringthe compositewall handlingandstructureintegrity)a 1.78 mm (0.07 in.)thicknessof porousmaterial2 with a den- 10sity of 0.2 and a 1.78 mm (0.07 in.) thick-ness of ceramicappearsto be a good compos- \
ite wall configurationsfor the assumedcon-ditions.
The insulatingpropertyof the ceramiclayer causes the ceramicsurfacetemperatureto be almostequal to the engine gas temper-ature. A detrimentalresult of the highsurfacetemperaturewould be an increaseinradiationto turbineparts. This may re-quire supplementalcoolingof these parts.
. A beneficialeffectof the high ceramictem-peraturemay be improvedabradabilityof theceramicduring a blade rub.
REFERENCES
i. L. T. Shiembob,"Developmentof aPlasmaSprayedCeramicGas Path Seal forHigh PressureTurbineApplication," Prattand WhitneyAircraftGroup, East Hartford,CT, PWA-5569-12, May 1978. (NASA CR-135387.)
2. R. C. Bill, L. T. Shiembob, andO. L. Stewart, "Development of Sprayed Ce-ramic Seal System for Turbine Gas Path Seal-ing," NASATM-79022, 1978.
3. R. C. Bill and D. W. Wisander,"Preliminary Study of Cyclic Thermal ShockResistance of Plasma-Sprayed Zirconium OxideTurbine Obter Air Seal Shrouds," NASATM-73852, 1977.
4. C. M. Taylor and R. C. Bill,Thermal Stresses in a Plasma-Sprayed CeramicGas Path Seal," AIAA Paper 78-93, Jan.1978. Journal of Aircraft, Vol. 16, April1979, pp. 239-246.
5. R. C. Bill, D. W. Wisander, andD. E. Brewe, "Preliminary Study of MethodsProviding Thermal Shock Resistance toPlasma-Sprayed Ceramics Gas Path Seals."NASA TP-1561, 1980.
6. A. G. Karimova,V. I. Lokai, andN. S. Tkachenko,"Investigationof Heat Re-leaseFrom a Gas to the Elementsof a Tur-bine Body." IzvestiyaVUZ AviatsionnayaTekhnika,Vol. 16, 1973, pp. 114-119.
7. Hippensteele,Steven A.; andColliday,RaymondS.: ComputerProgramforObtainingThermodynamicand TransportProp-erties of Air and Productsof CombustionofASTM-A-1Fuel and Air. NASA TP-1160,1978.
8. R. E. Gaugler,TACT 1: A Computer• Programfor the TransientThermalAnalysis
of a Cooled TurbineBlade or Vane Equipped Stepka and Ludwigwith a CoolantInsert. I - User's Manual.NASA TP-1271,1978.
9. C. H. Liebertand F. S. Stepka, 11"PotentialUse of CeramicCoatingas a Ther-mal Insulationon Cooled TurbineHardware,"NASA TM X-3352,1976.
10.C. H. Liebert and F. S. Stepka,IIIndustry Tests of NASA Ceramic Thermal Barrier Coating,1I NASA TP-1425, 1979.
11. W. P. Jarvi and A. R. Erickson,IIDevelopment of Improved Abradable Compressor Gas Path Seal. 1I Brunswick Corp.,Deland, FL, ER-382, July 1978. (AFML-TR-78101, AD-A072171.).
12. Y. S. Touloukian, ed., IIThermophysical Properties of Matter, Vol. 1: ThermalConductivity - Metallic Elements and Al10ys.1I New York, NY: IFI/Plenum, 1970.
13. F. E. Kennedy and R. C. Bill, IIThermal Stress Analysis of Ceramic Gas-Path SealComponents for Aircraft Turbines. 1I NASATP-1437, 1979.
Table 1. - Assumed Engine Conditions andGeometry
0.8
• 0.02
756 (900)
113.6 (250)
1144 (1600)1200 (1700)
Gas total temperatureat shroud, K (o F). • 1589 (2400)
Gas total pressure, atm ••.•••••• 25Gas flow rate, kg/sec
(lb/sec) •••• '••Turbine tip diameter,
cm (in.). • • • • • • 96.5 (38)Blade chord, cm (in.) •••••• 3.05 (1.2)Blade tip clearance
to-span ratio •••••Gas average absolute
Mach number • • • • •Cooling air temperature
to shroud, K (O F) ••••Allowable porous-metal
temperatures, K (O F):Current material ••••Advanced material •••
Francis S. Stepka and Lawrence P. Ludwig E-40210. Work Unit No.
9. Performing Organization Name and Address
National Aeronautics and Space Administration 11. Contract or Grant No.Lewis Research Center
Cleveland, Ohio 44135 13. Type of Report and Period Covered
12. Sponsoring Agency Name and Address TechnicalMemorandumNational Aeronautics and Space Administration 14. Sponsoring Agency Code
Washington, D.C. 20546
15. Supplementary Notes
Prepared for the Aerospace Congress sponsored by the Society of Automotive Engineers, Los
Angeles, California, October 13-16, 1980.
16. Abstract
A heat transfer analysis was made of a composite wall shroud consisting of a ceramic thermal
barrier layer bonded to a porous metal layer which, in turn, is bonded to a metal base. The
porous metal layer serves to mitigate the strain differences between the ceramic and the metalbase. Various combinations of ceramic and porous metal layer thicknesses and of porous metal
densities and thermal conductivities were investigated to determine the layer thicknesses re-
quired to maintain a limiting temperature in the porous metal layer. Analysis showed that the
composite wall offered significant air cooling flow reductions compared to an all-impingement
air-cooled all-metal shroud.
17. Key Words (Suggested by Author(s)) 18. Distribution Statement
Gas path sealing Ceramic shroud Unclassified - unlimitedTurbine shroud Ceramic seal STAR Category 07
Turbine tip sealing
19. Security Classif. (of this report) I 20. Security Classif. (of this page) I 21. No. of Pages 22. Price*
Unclassified I Unclassified I* Forsalebythe NationalTechnicalInformationService,Springfield,Virginia22161
National _,eronautics and SPECIAL FOURTH CLASS MAIL Postageand Fees Paid
__._nace Administration BOOK National Aeronautics andSpace Administration
Washington, D.C. NASA45120546Official Business
Penalty for Private Use, $300
__A POSTMASTER: If Undeliverable (Section 158Postal Manual) Do Not Return
