parametric studies to determine the effect of compliant ... · pdf filenasa technical...
TRANSCRIPT
NASA Technical Memorandum 102465
Parametric Studies to Determine the
Effect of Compliant Layers on MetalMatrix Composite Systems
J.J. Caruso and C.C. Chamis
National Aeronautics and Space Administration _Lewis Research Center
Cleveland, Ohio
........ and
H.C. Brown
Sverdrup Technology, Inc.
NASA Lewis Research Center GroupCleveland, Ohio
Prepared for the35th Symposium and Exhibition
sponsored by the Society for the Advancement of
Material and Process Engineering _Anaheim, California, April 2±5, 1990
(NASA-T_,-IO2G65) PARAMeTrIC eIUOIES TU NgO_I4Zc_DETERMINE TN__I E_F_CT nF COMPLrA_,T LAYEb, S ON
MF_AL ,MATPIX COMPOSIT_ SY_TrMS (NASA) 13 p
CS, CL 11_ Unc1_s
G3/24 0254450
https://ntrs.nasa.gov/search.jsp?R=19900004978 2018-04-23T22:46:52+00:00Z
PARAMETRIC STUDIES TO DETERMINE THE EFFECT OF COMPLIANTLAYERS ON METAL MATRIX COMPOSITE SYSTEMS
J.J. Caruso and C.C. Chamls
National Aeronautics and Space AdministrationLewis Research Center
Cleveland, Ohlo 44135
H.C. Brown
Sverdrup Technology Inc.NASA Lewis Research Center Group
Cleveland, Ohlo 44135
L_
Lr_I
LLJ
ABSTRACT
Computational Slmulatlon studies are conducted to Identify compliant layers
to reduce matrix stresses which result From the coefficient of thermal
expanslon mismatch and the large temperature range over which the current
metal matrix composltes wlll be used. The present study includes varla-
tlons of compllant layers and their properties to determine thelr Influence
on unldlrectlonal composite and constituent response. Two slmulation meth-
ods are used For these studies. The First approach Is based on a three-
dimensional linear Finite element analysls of a nlne-Fiber unldlrect|onal
composlte system. The second approach Is a mlcromechanlcs based nonllnear
computer code developed to determine the behavior of metal matrix composite
system For thermal and mechanlcal loads. The results show that an eFFec-
tlve compliant layer For the SCS 6 (SiC)/TI-24AI-1INb (TI3A1+Nb) and SCS 6
(SIC)/TI-15V-3Cr-3Sn-3AI (TI-15-3) composlte systems should have a modulus
15 percent that of the matrix and a compllant layer coeFFIclent of thermal
expanslon roughly equal to that of the composite system without the CL.
The matrix stress In the 1ongltudinal and the transverse tangent (hoop)
dlrectlon are tensile For the TI3AI+Nb and Ti-15-3 composlte systems upon
cool down From Fabrication. The Fiber longltudinal stress Is compresslve
From Fabrlcatlon cool down. Addition of a recommended compllant layer
will result In a reduction In the composlte modulus.
INTRODUCTION
Metal matrix composite (MMC) are prime candldates For hlgh temperature
applications. They present a unique challenge to today's researchers. The
large operatlng temperature range (greater than 2000°F) and the mismatchbetweenthe thermal expansion coeffIclents (CTE)of the fiber and the
matrix pose great difficulties in fabrication and low thermal cycling resls-tance. Both of these increase the matrix stress to critical proportionsand often lead to falIure. Manysolutions to this problem have been pro-posed (e.g., matching CTEfor fiber and matrix) with little success. One
suggestion is to use a compliant layer (CL) as a buffer between the fiberand the matrix. Thegoal of this CL Is to reduce the matrix stress without
degrading the fiber, the matrix, or the composite. Finding a material todo thls, however, mayprove to be quite a challenge.
In an attempt to flnd such CLs, a parametric study is a natural first
step. It is essential to assemblea working knowledgeof which and/or howcharacterlstlc parameters of the constltuent material effect compositebehavior.
Twocomputatlonal simulation approachesare used in this study to evaluateCLs In MMCsystems for high temperature applications under thermal loads.The first is a three-dimensional linear finite element method. The otheris a nonllnear mIcromechanlcsbasednumerical method.
In attempt to identify characterlstics for suitable compliant layers, com-pTiant layers are evaluated with two commonbut very different metal matri-ces. They are (I) Ti-15V-3Cr-3Sn-3Al (Ti-15-3) and (2) Ti-24A1-11Nb
(Ti3AI+Nb) reinforced with SiC (SCS6) fibers. The SiC/Ti-15-3 compositesystem has beenprocessedwithout matrix cracks (I-3). SiC/TI3AI+Nb, how-ever, does develop matrix cracks durlng processlng (5,6), maklng It a candi-date for a CL. There wlll be somedlscusslon on the Ti-15-3 system;but, the maJorlty of the effort will concentrate on the Ti3Al+Nb system.
The developmentof matrix cracks maybe related to tenslle stresses In spe-clflc dlrectlons. Large tensile hoopstresses in the matrix can cause thetransverse tangent (hoop) cracks observed by Mackay(1) and Ghosn& Lerch
(2). Similarly, large tensile longitudlnal stresses maycause matrix crack-ing perpendicular to the fibers. Thesecracks have not beenobserved. Acompliant layer should reduce the stress in all directions to be an effec-tive methodfor preventing cracks during the fabrication and the thermalcyc11ngof these systems. If the goal Is to ellminate matrix cracklng, acompromlsemayhave to be madebetweenstresses to elimlnate matrix crack-Ing.
METHODS
The FEM calculates the linear composite response at many locations In the
constltuents of the composlte system that cannot be determined through
2
experiments (6-8). The general purpose finite element code MSC/NASTRAN (9)
is used to perform the analysls. The model uses elght-noded hexahedron and
sIx-noded pentahedron elements to form a nlne-cell nlne-flber unidirec-
tlonal composite system (Fig. I). The model provides constituent displace-
ments, stresses, and forces due to thermal loadlng condltlons.
__i"<-.!,! ...-._
Z_X I
C-m
n : ,
y '- ,
ZX_
Z
FIGURE I. - MULTICELL 3-D FINITE ELEMENT MODEL
FOR LOCAL DETAILS.
//
\\
FINITEELEMENT_
COM PONENT
_ _ FINITEELEMENT
GLOBAL STRUCTURAL jANALYSIS
\
LAMINATE /// LAMINATE 1LAMINATE_ MINATE 1
I THEORY / THEORY I I
- METCAN
pl.y___/ _'_ __" PLY I
COMPOSITE / _ ,_ COMPOSITE......... _ _ _ J¢" MICROMECHANICS /
\ MICROMECHANIC5 _THEORY /__ n u '_. THEORY /\ " _ .,-.. PI.._.." /
\ ./UPWARD _ CONSTITUENTS t / TOP-DOWN
INTEGRATED _ _ MATERIAL PROPERTIES / "/ TRACEDP (_,T, t) / OROR _
"SYNTHESIS" --- ....... _ "DECOMPOSITION"
FIGURE 2. - HIGH TEMPERATURECOMPOSITE BEHAVIORCOMPUTATIONALLYSIMULATED.
3
The second approach, the METCAN computer code (Fig. 2), is used to
determine the nonlinear average composite response of unidlrectional and
Iamlnate composite systems (lO). The code's computatlonal simulation
begins wlth the coo] down of the consolidation process and includes any
thermal and/or mechanlcaI processes which occur after consolidation. The
changes In constituent materials resultlng from environmental variations
are taken Into account. In this study a unidlrectiona] MMC system Is evalu-
ated to determine the fiber, matrix, and CL stress. These stresses are
used to determine whether the composite system will survive fabrication
and, if so, the thermal cycles to failure of the system. METCAN also com-
putes composite properties In an attempt to determine the effect of a CL on
them.
EG
C_
Soxxt
TABLE l ROOM TEMPERATURE CONSTITUTIVE
MATERIAL PROPERTIES
pslpsi
In.lln. °Fksl
SIC
Fiber
62.0xi0623.8x106
0.3002.70x10-6
TI3AI+NbMatrix
I].OxlO 6
4.23xl06
0.3006.50xi0-6
65.0
TI-15-3Matrix
14.5x106
5.50x106
0.3204.72x10-6
130.0
RESULTS AND DISCUSSION
Finite Element Method The first phase of this study Is to find, using
the FEM, the longitudinal stresses in the SiC/Ti31+Nb (Table l) for
selected CL's. The first selected CL (2) has a CTE of 4.67 ppm/°F and a
modulus that varies from 1.1 to 22 Msi. Figures 3(a) and (b) shows that
the CL stress Is lower than the matrix stress throughout the range of CL
modulus. In Flg. 3, the stress is normalized per degree Fahrenheit.
(Table 2 summarizes the figures for the readers' benefit). The fiber Is In
compression for the composite system In a fabrlcatlon cooi down through the
range. The longltudlnal stress In the matrix is low on the surface
(Figs. 4(a) and (b)) and Increases as the X/D ratio Increases, where X
is the distance from the free surface of the composlte and D Is the fiber
dlameter. As X/D increases, the matrix stress does not decrease untlI the
compliant layer modulus has decreased to 40 percent of the matrix modulus.
In a11 cases the matrix stress does not significantly decrease until the CL
modulus is 15 percent of the matrix modulus. For the Ti3Al+Nb system the
CL modulus would be 1.7 Msi.
40
---o---- FIBER.__A___ CL (CVR .20)..... O..... MATRIX
20
0(
-20
-40
=_ -60
o"1=. -80
-1000
[,,,lJ
...,-
20/
=: 0
2°l-40
-80
- 100
0 .4 .8 1.2 1.6 2.0
Ec I/Era
FIGURE 3. - ROOM TEMPERATURELONG-
ITUDINALCONSTITUENTMICROSTRESS
(0_iI) FOR A 20 PERCENTFVR SC26/Ti3-AI + Nb AT L/D = 2.21.
35
Ecl/E m
•-..-o--.- 2.0-..in.._ .5..... "_..... .1
---we--- 1.0
30
25,
20{
u. 15
0- 10'
.........i;°7" A_.
, (b)"..
0_- - ;':'_--.-* ....... ,, ....... ,,4.................._ ............. -_...............
_4L 7 ''_ 'T ":--'4 ..................t _0 .5 1.0 1.5 2.0 2.5
x/D
FIGURE 4. - RT LONGITUDINALCON-
STITUENT MICROSTRESS(0/,11)FOR20 PERCENT FVR, 20 PERCENTCVR
SCS-6/Ti3-AI + Nb,
The second specific CL has a higher CTE of 7.5 ppml°F but the same range
for the modulus (l.l to 22 Msl). Figure 5 shows the longltudlnal stress In
the constltuents of the SIC/TI3Al+Nb composite system. At a hlgh CL modu-
lus, the CL stress Is much larger than the matrix stress. As the CL modu-
lus decreases, the compliant layer stress decreases. The matrix stress
Fig- I FVRure
3(a) 0.20
3(b)
4(a)
4(b)
5
6 '.'46
7 .20
8
9
lO
TABLE 2 TEST MATRIX
[Stresses are due to thermal loadln, (excluding Fig. g).]
Composite CVR Compliant Testsystem layer
0.20 TI3AI w/ Longltudinal constituent _icros:r_s_var. Eel CL CTE - 4.57 ppm. X/O . 1.47
SCS61TI3AI
SCS6/TII5
SCS6/Ti3AI
.16
TII5 wl
var. Ecl
Longitudinal _ "",_ns_tuen: ,_Icrostres_
CL CTE = 4.67 ppm, ×ID - 2.21
Longitudinal matrix mlcrostressas X/O increases
Longitudinal CL mlctrostrassas X/O Increases
Longitudinal constltuen: _icros:ressCL CTE - 7.55 ppm, X/D = 2.2l
Longitudinal constltuent mlcrostress,Jar. CL CTE X/D . 2.g4
0.20
Repeat CL
From Flg. 7
GdNbZrTest
0.20 GdNbZr
0
.05,.I,.2
Test
Carbon
Gd
Maximum constituent hccD microstresldue to CL type
Maximum Iongltud!nal constituentmicrostress due to CL type
Change in composite modulus dueto CL type
Thermal cycles to failure asCTE mismatch increases
remalns constant throughout thls range of the CL moduli. The longitudinal
matrix stress does not vary with the two CL CTEs shown In Figs. 3 and 5.
Constituent longltudlna] stresses are predicted from the FEM for a compos-
Ire system that contains several CLs for which the CTE varies inversely
wlth the modulus. For example, when the CL modulus Is halved, the CTE is
doubled. As the CTE decreases and the modulus increases, the CL stress
grows unti] It is slgnlflcantly larger than the matrix stress (Fig. 6).
The CL, even with a high CTE, does not slgnlflcantly reduce the matrix
stress untll the modulus Is 15.percent that of the matrix. Unfortunately,
the CL stress Is very hlgh at th|s polnt. Since the CL stress is large the
CL strength must also be hlgh. A good rule of thumb for the SIC/TI-IS-3
composite system is to have the CL strength twice the matrix strength.
Flndlng a CL that meets thls Is obvlously not very practical since It may
be impossible to Find a material wlth a low modulus and high strength.
Next, the effect of select CL materlals on the SiC/TI3AI+Nb system Is exam-
Ined. Four of the many CL materlals evaluated by Ghosn and Lerch (2) were
6
=.-u_ -40u3
-80
-120E
-160
---o--- FIBER---_-- CL (CVR.20).....[].....MATRIX
80LI0_ _ ....o.O..O°o°..o .°. o.O.O--°'"°""
O_ _ ........
CL CTE = 4.67PPM
•i [ __--=L { {
0 .4 .8 1.2 1.6 2.0
Ecl/Em
FIGURE5. - ROOMTEMPERATURELONGITUDINALCONSTITUENTMIC-
ROSTRESS(0J,ll ) FORA 20 PER-CENTFVRSCS6/Ti3-AI + Nb ATL/D = 2.21.
CVR .01
----0-----FIBER
---c}--- MATRIX
...._ .....CL
,, 992 -_ 80
496 - _. qO6- =vl, e._
_" 0-_ 00 0
-496 -
CVR .16
---4--- FIBER
---m--- MATRIX
-.-A--- CL
_A !
:y ....._..7....._::.._.::._,_.... ----._.
:,,.,=-"'"__-L:._-_......."""'"_:'-I./
-40 " _'_='_'---T--_--,..... i- -'J
0 .2 .4 .6 .8 1.0
Ecl/E m
FIGURE 6, - ROOM_ERATURE CONSTITUENT
M[CROSTRESS0_11 WITH VARYING CL CI'EFOR 46 PERCENTFVR SCS6_i-lS-3 AT
L/D = 2.9q.
chosen for this study (Table 3). Three of the CLs are actual materials, Gd
(gadollnlum), Nb (niobium), and Zr (zlrconlum). One is a fictltlous mate-
rlal deslgned to flnd optlmum CL propertles (called the test case in the
present study). This test case has a CTE of 4.67 ppm/°F and a modulus
varying Form 1.1 to 2 Ms1. The composite systems contalnlng CLs are com-
pared to a reference case of SCS 61TI3AI+Nb having no CL, both at a fiber
volume ratio (FVR) of 20 percent. In this study the FVR Is the Fiber
volume ratlo of the flber without the CL.
The FEN Is used to predlct the hoop microstress SiC/Ti3AI+Nb system. For
most CLs the matrix hoop stress Is less (about 20 percent) than the matrix
stress of the reference case with no CL (Fig. 7). A CL of Zr, however,
increased the stress in the matrix slightly; while the CL stress went into
compression. The CL stresses are slgnlflcant for the fabrication cool down
and, because of thls, the CL strength may need to be close to that of the
matrix or larger.
The longitudinal mlcrostresses in the SIC/Ti3AI+Nb system are determined
from the FEM and plotted per degree Farenhelt. The results show that the
CLs examined Increase (about II percent) the longltudinal matrix stress
From the reference case wlth no CL (Fig. 8). This Increase In 1ongitudlnal
matrix stress may Induce matrix cracklng during fabrication and thermal
cycling. The increase in matrix stress may also reduce Fatigue life, pro-
vided the composite system survives fabrlcatlon.
TABLE3 ROOMTEMPERATURECOMPLIANTLAYERMATERIALPROPERTIES
E psiG pslV
in./in. °F
Gd Zr Nb
8.30xi063.30xi06
0.2605.60xi0-6
13.6xlO65.10xlO 6
0.3322.90xi0 -6
14.9xi065.31x106
0.4024.40xi0-6
496 -
248 -IJ.
O-
-248 -
CL
E_ MATRIX40
IO - --
WICL NO CLCVR = 20
FVR : 20-20
TEST Gd Zf Nb FVR20
COMPLIANT LAYER TYPE
FIGURE 7. - MAXIMUM CONSTITUENT HOOP MICRO-
STRESS ((JO)DUE TO CL TYPE FOR SCS6/
Ti3-AI + Nb.
496 40
248 20
O _ 0
o'-
-248 --20
-q96 -40
CL
MATRIX
....--'-_ i_
_W/CL NO CLCVR = 20
r- ""_FVR= 20
TEST Gd Zr Nb FVR20
COMPLIANT LAYER TYPE
F[GURE 8. - MAXIMUM LONGITUDINAL CONSTITUENT
MICROSTRESS (0_iI) DUE TO CL TYPE FOR SCSG/
Ti3-AI + Nb.
The composite modulus may be affected by the use of CLs. The five systems
from the FEM study have a very small decrease in the longitudinal modulus.
Hence, the composlte modulus variation is small (±4 percent) (Fig. 9). The
two compllant layers that performed the best in this study (Gd, Test) have
a sllghtly lower modulus than the reference. The change in composite modu-
lus could be slgnifIcant when using the recommended CL with a modulus
15 percent that of the matrix.
METCAN
The thermal cycllng resistance Is critical because of the large temperature
envelopes in which MMC systems operate. The multifactor interaction rela-
tionship (MFI) is used to calculate the thermal cycles to fallure. The MFI
is shown for the SIC/TI3AI+Nb composite system. The exponents are deter-
mlned from monolithic matrix experimental data. The matrix material is
fabricated as a composite system Is fabricated.
z ]O-o o.oo
Where S/SF Is the remalnlng ultimate tensile strength of the composite
after thermal cycling over the original ultlmate strength of the composite,
Tm is the melting temperature (2730 °F), Tu is the maximum temperature of
the thermal cycle (1200 °F), TO is the reference temperature (70 °F), SO
Is the room temperature reference ultimate tensile strength of the matrlx
(65 ks|), Op is the longitudlnal residual matrix stress at the end of
cycllng, ac Is the longitudlnal residual matrix stress at the end of con-
solldatlon, N Is the number of cycles at which there Is lO percent remain-
Ing strength, and NF is the number of cycles to fallure. -This equation
is evaluated For the number of cycles to fallure with a lO percent
remaining strength S/SF.
METCAN Is used to investigate the thermal cycles to Failure of FVR 20
SIC/TIAI+Nb and variations of this system. In Four of the composites the
protective carbon coating Is removed From the Fiber, three flbers have the
coating replaced wlth a CL of Gd of different thlcknesses and one without a
coating or CL. The fiber without a coating was assumed to have the same
strength as one with a coatlng (I.e., SCS 6). One addltional composite sys-
tem Is analyzed with the coating. As the CTE mismatch between the Fiber
and the matrix grows, the composlte system became less tolerant of the
cycllng (Fig. lO). The worst case is the composite system that contalns
the flber with the coatlng. The CL sllghtly delayed the Failure as the CTE
mlsmatch increased.
4
I2 2Z
M- -2zw
e,-
_. -{I
MODULUS
TEST Gd Zr Nb FVR20
COMPLIANT LAYER TYPE
FIGURE 9. - CHANGE IN COMPOSITE
MODULUS FOR VARIOUS COMPLIANT
LAYER MATERIALS FOR SCS-6/
Ti3-AI + Nb.
Z
Z
1.0
.8
.6
.4
.2
Gd CVR .20
---o--- Gd CVR .I0
.....A.....Gd CVR .05
- _-- CVR .00----m--- COATING .03
- I!,
r
.21 I . t
.4 .6 .8 1.0
af/am
FIGURE I0. - EFFECT OF CTE MISMATCH
ON THERMAL CYCLES TO FAILUREOF
20 PERCENT _ SICfFI3-AI+ NbWITH DIFFERENTCOATINGS.
9
CONCLUSION
The compliant layers studied appear to be Ineffectlve in reduclng the 1ongl-
tudlnal stress in the Ti3Al+Nb composite system except with very low CL
modull (on the order of 15 percent of the matrix modulus). The reduction
in hoop stress may eliminate the cracking In the hoop direction. It does
not significantly Increase the thermal cycles to failure of the system.
The fiber stress is in compresslon and Is currently not considered the weak
link in thls type of failure.
The best performing compliant layer has a low modulus (around 15 percent of
the matrix) and a CTE lower than the matrix but higher than the fiber.
High CTE CLs cause high stresses in the CL which require high CL strengths.
Low modulus and high strength are not compatible properties for a materlal
to have. The reduction in matrix stresses due to high CTE are not realized
until the CL modulus is 15 percent that of the matrix
modulus.
The compliant layer generally reduces the modulus of the composite but not
by an appreciable amount. If a low modulus compliant layer (15 percent of
the matrix) were used, the reduction could be much greater.
REFERENCES
I. McKay, R.A., "Effect of Fiber Spacing on Interracial Damage in a Metal
Matrix Composlte," submitted to Scripta Metallurgica, 1989.
2. Lerch, B.A., Hull, D.R., and Leonhardt, T.A., "As Received
MIcrostructure of a SIC/TI-15-3 Composite," NASA TM-I00938, 1988.
3. Lerch, B.A., Personal Communication, NASA Lewis Research Center,
Cleveland, OH, 1989.
4. Brlndley, P.K., Bartolotta, P.A., and KIIma, S.J., "Investigation of a
SiC/Ti-24Al-llNb Composite," NASA TM-I00956, 1988.
5. Brlndley, P.K., Bartolotta, P.A., and Mackay, R.A., "Thermal and
Mechanical Fatigue of SIC/Ti3AI+Nb," Hltemp Review 1989, NASA CP-IOO3g,
1989, paper 52.
6. Caruso, J.J., "Application of Finite Element Subststructurlng to
Composite MlcromechanIcs," NASA TM-83729, 1984.
7. Caruso, O.J. and Chamls, C.C., "Superelement Methods Applications to
MlcromechanIcs of High Temperature Metal Matrix Composites," 29th
Structures, Structural Dynamics, and Materials Conference, Part 3,
AIAA, 1988, pp. 138B-1400.
10
8. Caruso, J.J., Trowbrldge, D., and Chamls, C.C., "Flnlte Element
Applications to Explore the EFfects of Partlal Bondlngon Metal MatrixCompositeProperties," 30th Structures, Structural Dynamics, and
Materials ConfereDce, Part l, AIAA, I9B9, pp. 140-154.
9. GockeI, M.A., ed., MSC/NASTRAN: Handbook for Superelement Analysis:
MSC/NASTRAN Verson 61, The MacNeal-Schwendler Corp., Los Angeles, CA,
19B2, p. ].]-l.
IO. Hopkins, D.A. and Chistos, C.C., "A Unique Set of Micromechanics
Equations For High Temperature Metal Matrix Composites," NASA TM-B?I54,
19B5.
It
Report Documentation PageNational Aeronautics andSpace Administration
1. Report No. 2. Government Accession No.
NASA TM- 102465
4. Title and Subtitle
Parametric Studies to Determine the Effect of Compliant
Layers on Metal Matrix Composite Systems
7. Author(s)
J.J. Caruso, C.C. Chamis, and R.C. Brown
9. Performing Organization Name and Address
National Aeronautics and Space AdministrationLewis Research Center
Cleveland, Ohio 44135-3191
12. Sponsoring Agency Name and Address
National Aeronautics and Space Administration
Washington, D.C. 20546-0001
3. Recipient's Catalog No.
5. Report Date
6. Performing Organization Code
8. Performing Organization Report No.
E-5252
10. Work Unit No.
505-63-1B
11. Contract or Grant No.
13. Type of Report and Period Covered
Technical Memorandum
14. Sponsoring Agency Code
15. Supplementary Notes
Prepared for the 35th Symposium and Exhibition sponsored by the Society for the Advancement of Material and
Process Engineering, Anaheim, California, April 2-5, 1990. J.J. Caruso and C.C. Chamis, NASA Lewis
Research Center; H.C. Brown, Sverdrup Technology, Inc., NASA Lewis Research Center Group, Cleveland,Ohio 44135.
16. Abstract
Computational simulation studies are conducted to identify compliant layers to reduce matrix stresses which result
from the coefficient of thermal expansion mismatch and the large temperature range over which the current metal
matrix composites will be used. The present study includes variations of compliant layers and their properties to
determine their influence on unidirectional composite and constituent response. Two simulation methods are used
for these studies. The first approach is based on a three dimensional linear finite element analysis of a 9 fiber
unidirectional composite system. The second approach is a micromechanics based nonlinear computer code devel-
oped to determine the behavior of metal matrix composite system for thermal and mechanical loads. The results
show that an effective compliant layer for the SCS 6 (SiC)/Ti-24AI'-I 1Nb (Ti3AT + Nb) and SCS 6(SiC)/Ti-15V-3Cr-3Sn-3AI (Ti-15-3) composite systems should have modulus 15 percent that of the matrix and a
coefficient of thermal expansion of the compliant layer roughly equal to that of the composite system without the
CL. The matrix stress in the longitudinal and the transverse tangent (hoop) direction are tensile for the Ti3A! + Nband Ti-15-3 composite systems upon cool down from fabrication. The fiber longitudinal stress is compressive
from fabrication cool down. Addition of a recommended compliant layer will result in a reduction in the
composite modulus.
17. Key Words (Suggested by Author(s))
Compliant layer
Metal matrix compositesMicromechanics
18. Distribution Statement
Unclassified - Unlimited
Subject Category 24
19. Security Classif. (of this report) 20, Security Classif. (of this page) 21. No. of pages
Unclassified Unclassified 12
NASAFORM1628OCT86 *For sale by the National Technic_#lInformation Service, Springfield, Virginia 22161
22. Price*
A03