International Journal of Heat and Mass Transfer 77 (2014) 653661

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

j o u r n a l h o m e p a ge : w w w.e l s e v i e r.c o m / l o c a t e / i j h m t

Effective thermal conductivity of composites

F. Gori , S. Corasaniti
Universit di Roma Tor Vergata, Via del Politecnico 1, 00133 Rome, Italy

a r t i c l e

i n f o

Article history:
Received 31 January 2014
Received in revised form 15 May 2014
Accepted 23 May 2014
Available online 23 June 2014
Keywords: Composite
material Ceramicsilica matrix Fiber
reinforcement
Thermal conductivity
Cubic cell model
Anisotropy degree
Potential anisotropy
Anisotropic efciency

a b s t r a c t
The paper employs a theoretical model to evaluate the thermal conductivity of composites under two
thermal assumptions which allow to solve the heat conduction equation. The composite is made of a
ceramic-silica matrix and a ber reinforcement. Three different materials are investigated as reinforcement, i.e. asbestos, steel and copper. The theoretical effective thermal conductivity is calculated along
the three directions for a non-consumed composite and during its consumption. Numerical solutions
of heat conduction in the composite are carried out. The anisotropy degree of the composite is
investigated for the composite during its consumption. The anisotropic efciency of the composite is
dened as ratio between the anisotropy degree of the composite and the potential anisotropy of the
two materials, i.e. the ratio between the thermal conductivity of the ber and the matrix. The
theoretical model allows to evaluate, under the two thermal assumptions, the anisotropic efciency of
the compos- ites which decreases with the increase of the potential anisotropy reaching a minimum,
which is only dependent on the geometry of the composite.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
The thermal behavior of composites, made of two materials
with different thermal properties is important in several applications as thermal protections, heat shields, heat guides and in
cases where anisotropy is requested. Several researchers have
investi- gated composites with different approaches.
Torquato and Kim [1] computed the effective properties of heterogeneous media with an efcient simulation technique. Havis
et al. [2] used the nite difference method to simulate the thermal
behavior of ber reinforced composites, with the conclusion that
these materials can be considered homogeneous and thermally
anisotropic. Graham and McDowell [3] evaluated the effective
thermal conductivity of a composite made of two different materials by means numerical simulations of the microscopic structure
of the composite. Islam and Pramila [4]
carried out
a
numerical analysis of the transverse thermal conductivity of
composites with imperfect interfaces. Dasgupta and Agarwal [5]
developed a 2-D thermal method based on
series/parallel
arrangement. Gu and Tao [6] studied a periodic composite with
contact resistance with the conclusion that the contact resistance
can change the effective conductivity dramatically. James and
Harrison [7] evaluated the effective thermal conductivity of
composites, made of two

Corresponding author. Tel.: +39 0672597129.

E-mail addresses: fammannati@yahoo.com (F.
uniroma2.it (S. Corasaniti).

Gori),

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.05.047
0017-9310/ 2014 Elsevier Ltd. All rights reserved.

sandra.corasaniti@

materials, simulating numerically the microscopic structure of

the composite for different orientations of the anisotropic materials. Torquato and Rintoul [8] investigated the effect of the interface
on the properties of composite media. Wang [9] determined the
thermal conductivity of a material containing a layer or a staggered
array of thin strips. Shim et al. [10] investigated the thermal
conductivity of carbon ber-reinforced composites with different
ber cross-section types, with emphasis on anisotropy and thermal
diffusivity factors. Lim [11] evaluated the thermal conductivity of
unidirectional continuous ber-composites, particulate composites and laminate sheets. Terron and Sanchez Lavega [12] and
Salazar et al. [13] determined the effective thermal diffusivity of
ber-reinforced composites. Skorokhod [14] studied multi-layer
composites, made of ceramic and metallic sheets, calculating the
effective thermal properties, with the conclusion that the composite material is thermally anisotropic. Felske [15] studied the
thermal conductivity of carbon ber-reinforced composites with
different ber cross-section types. Goyhnche and Cosculluela
[16] presented multi-scale modeling of the effective thermal
conductivity tensor of a stratied composite material made of
carbon bers, phenolic resin, and carbon loads. Al-Nassar [17]
deduced empirical formula to predict the thermal conductivity of
a composite material with estimated air void volume percent and
the relative 3-D problem was modeled using a nite element
analysis. Yuan and Luo [18] studied composites of silicone matrix,
lled with phosphor and used in light emitting diodes packaging,
for predicting the thermal conductivity and experiments were
carried out to validate the model.

142

F. Gori, S. Corasaniti / International Journal of Heat and Mass Transfer 77 (2014) 653661

Nomenclature
Latin
a
b
c
k
L
q
Q_
transfer
s
s1 , s 3
s2
T
Greek
D
n

thickness in x, y directions
thickness in x, y directions
specic heat
thermal conductivity
length
heat ux
heat

composite density
reinforcement volume fraction
anisotropic efciency
potential anisotropy

v
d

Subscripts
f
ber
m
matrix
max
maximum
min
minimum
x, y, z
directions
i, j, k
versors

generic thickness in z direction

matrix thickness
ber thickness
temperature

nite difference
anisotropy degree

A thermal protection composite was studied theoretically and

numerically in [19] during thermal degradation. The thermal
behavior of a composite, made of silica matrix and brous asbestos
reinforcement and used as heat shield, was studied theoretically
and numerically during the thermal ablation of a re-entry mission
at high Mach number [20]. A composite material made of two
plane layers, where one material is silica and the other one is
asbestos, or steel or copper has been investigated theoretically in
[21,22].
This paper employs a theoretical model for the thermal conductivity of composites, subjected to consumption, investigating the
thermal anisotropy behavior. The theoretical approach employs
the cubic cell model with two thermal assumptions, parallel isothermal lines and heat ux lines, originally proposed in [23,24],
later on extended to non-saturated soils [25], compared to experimental data [2629], applied to space problems [3033] and compared to other theoretical models [3440].

s1

s2

a
b

2. Theoretical models

s3

The elementary cubic cell of the material is presented, in threedimensions, in Fig. 1, with the ber reinforcement in the middle

and the ceramicmatrix around. The upper and lower layers of

the matrix have thicknesses s1 and s3, while the intermediate
layer, containing ber and matrix, has thickness s2.
The reinforcement volume fraction is

s2

4a2

/
2
s1 s2 s3 2a b

The heat conduction equation in three dimensions for a non-isotropic material is

1
qx
~q @B qAy C
qz

1
kx kxy kxz
B
C!
r T @ kyx ky kA
rT
yz
kzx kzy kz

The symmetry conditions of the elementary cubic cell of Fig. 1,

kij kji ij, allow to assume kx = ky, kxy = kyx, kzy = kyz and kxz = kzx,
in agreement to [41], and the thermal conductivity tensor can be
reduced to

0
1

kx

B
@0
0

0
ky
0

0
C
0A
kz

a
a

Fig. 1. Elementary cubic cell.

2a b

and parallel heat uxes, or parallel ux lines (PFL), to dene the

thermal conductivity in the generic i direction

ki

Q_ i
Li
DT i Ai

2.1. Parallel isothermal lines (PIL)

The assumption of parallel isothermal lines, PIL, means that the
thermal conductivity in the transversal direction, i.e. on the isothermal plane, is innitely high. The effective thermal conductivities in x and z directions, indicated as kx and kz, are calculated with
the assumption that the temperature at z = 0 or x = 0 is T1, at z = s3
or x = a is T2, at z = s3 + s2 or x = a + b is T3 and at z = s3 + s2 + s1 or
x = 2a + b is T4.
The total heat transfer in the x direction is the same throughout
DT x
each cross section, in steady state, and is

F. Gori, S. Corasaniti / International Journal of Heat and Mass Transfer 77 (2014) 653661

The heat conduction equation can be solved under the two thermal
assumptions of parallel isotherms, or parallel isothermal lines (PIL),

Q_ x kx Ax

Dx

kx

T1
T4
2a b s1 s2 s3 2a b

143

The heat transfer crossing the rst layer of thickness a is

I
Q_
x

km Am

T1
T2
a

T1
T2

kf Af

The heat transfer crossing the rst layer of thickness s1 is

The cross sections, Am and Af are

T1

km 2a b

T2
20

s1

The heat transfer crossing the center layer of thickness s2 is

Am s1 s2 s3 2a b b s2
7

n
2
Q_ II
2
km 4a kf 2a b
z

and

Af b s2

o T
2

4a

T3
21

s2

The heat transfer crossing the last layer of thickness s3 is

Q_z III km 2a
b

giving

Q_

Q_ zI

T3

T4

22

s3
x

km s1 s2 s3 2a b b s2 kf b s2
T1

T2

Eqs. (20)(22) can be expressed with the temperature differences

T1

km s1 s3 2a b kf 2a b s2

T2
T3
b

10

Qx_ III km s1 s2 s3 2a b b s2 kf b s2
T4

T2

T3

T3

_
s1
zQ
T4
2
km 2a b

Eqs. (9)(11) can be expressed with the temperature differences

Q_

a
x

T2

T2

_ II
xQ
T3
kf 2a b s2
km s1 s3 2a b

km s1 s2 s3 2a b b s2 kf b s2

24

4a2

2a b

12

25

The sum of Eqs. (23)(25) gives

T1

km 4a2 kf

III

11

_ II
Q
h z

s2

The heat transfer crossing the last layer of thickness a is

T3

23

km 2a b

The heat transfer crossing the center layer of thickness b is

II_
Q

Q_ z

s3

T2

T1

T4

s1
Q_ z

s3
2

km 2a b

Q_ z

s2

km 4a kf 2a b

26

4a

On the other hand Eq. (19) can be written as

T3

T4

Q_

13

km s1 s2 s3 2a b b s2 kf b s2

14

T1

T4

Q_ x

s2

Q_ x
T4
kx s1 s2 s3

kx s1 s2 s3

s2
km 4a2 kf

s1

2a
Q_ x
km s1 s2 s3 2a b b s2 kf b s2

_
b
Q x

km s1 s3 2a b kf 2a b s2
17

The effective thermal conductivity in the x direction, kx, is nally

Q_ z
2a b

4a2

28

The effective thermal conductivity in the z direction, kz, is nally

kz
16

s3 Q_ z

km 2a b

15

27

s1

and the temperature difference of Eq. (16) can be substituted into

Eq. (15) giving

Q_ x

2a b

s3

kz 2a b

On the other hand Eq. (5) can be written as

T1

km s1 s2 s3 2a b b s2 kf b s2
b
Q_ x

km s1 s3 2a b kf 2a b s2

s3 Q_ z

s1 s2

and the temperature difference of Eq. (27) can be substituted into

Eq. (26) obtaining

s 1
Q_ z

The sum of Eqs. (12)(14) gives

2a

T4

III

T1

s2

s3
29

2a b

s 1 s 3
2
km 2ab

4a2 k

s2
2
4a2
m kf 2ab

2.2. Parallel heat ux lines (PFL)

The assumption of parallel ux lines, (PFL), means that the thermal conductivity in the transversal direction, i.e. orthogonal to the
heat ux, is zero. The effective thermal conductivities in x and z
directions, indicated as kx and kz, are calculated with the assumption that the temperature at z = 0 or x = 0 is T1, at z = s3 or x = a is
T2, at z = s3 + s2 or x = a + b is T3 and at z = s3 + s2 + s1 or x = 2a + b
is T4.
The total heat transfer in the x direction is the sum of the heat

kx

2a s1 s2 s3
km s 1 s 2 s3 2a b b s2 kf b s2 b
b s1 s2 s3
2a b km s1 s3 kf s2

_ the matrix
transfer throughout the matrix, Q m ; the_ ber, Q f , and
_
and ber, Q ,

18

The total heat transfer in the z direction is the same throughout

each cross section, in steady state, and is

Q_z kz Az

DT
kz 2a b 2
Dz

T1

T4

s1 s2 s3

19

_ m _Q f _ Q
Q_ x Q
30
The heat transfer throughout the matrix, Q_ m ; with cross section
(s1 + s3) (2a + b) and thickness (2a + b) is

T 1 T
Q_m km s1 s3 2a b
2a b

31

The heat transfer throughout the layer of ber, Q_ f ; with cross

sec- tion (s2 b) and thickness (2a + b) is

T1
Q_ f kf s2 b

The heat transfer through matrix and ber, Q_ ; can be expressed

in function of the temperature differences from the Eqs. (42)(44)
and
by summation it is obtained

T4
32

2a b

The total heat transfer in the x direction, Q_ , is the same

throughout
each cross section, in steady state. The heat transfer through the
rst layer of matrix, with cross section (2s2 a) and thickness a is

Q_ km

T1
2s2 a

T2

33

h
Q_ 2a
b

T3

34

The heat transfer through the last layer of matrix, with cross section
(2s2 a) and thickness a is

Q_ km 2s2 a

T3

T4

35

The heat transfer, Q_ ; can be expressed with the temperature

differ- ences of Eqs. (33)(35) and by summation it is obtained

Q_ s2 T 1

T4

1
b

km 2a kf

T
T 1
kx 2a b s1 s2 s3 4
2a b
Dx

37

Substitution of Eqs. (33)(35) into Eq. (37) gives

kf
s2 b
s2
2a b
1

1
b

km 2a kf

38

The effective thermal conductivity in the x direction, kx, is nally

kx

b s2
s
2ab 2

km s1 s3 kf
s1 s2 s3

km

b
2a kf

39

The total heat transfer in the z direction is the sum of the heat
trans- fer through the matrix, Q_ m , and the matrix and ber, Q_ ,

Q_ z Q_ m Q_
40

The heat transfer through the matrix, with cross section (4a2) and
thickness (s1 + s2 + s3), is

T1

Q_ m km 4a2

T4

41

s1 s2 s3

s1

sk3

2a

4a

T2

s
1

s2

2a b2

T 1

T 4

45

kf

"

4a

km 2a

s1 s2 s3

s1 s3
ks2
km
f

4a
46

Table 1 reports the thermo-physical properties (assumed constant with the temperature) used in the present theoretical model
and numerical simulations. The matrix is made of ceramic-silica
while three materials are assumed as ber, i.e. asbestos, steel
and copper.
The thermal conductivities of the ber differ by several orders
of magnitude, allowing to study the effect of the ratio between
the thermal conductivities of the ber and the matrix.

Numerical calculations have been performed with a housemade code for the same geometrical conguration. In order to estimate kx and kz heat transfer has been numerically computed by
setting a temperature gradient in the x direction (i.e. adiabatic
walls normal to y and z directions), and z direction.
Energy equation has been solved with a second order nite-differences scheme by using a non-iterative steady-state explicit
method in a grid of hexahedral cells
2

42

@ T @ T @ T
0

@ x2 @ y2 @ z2

47

Once the temperature distribution has been determined, the total

heat ux is calculated by numerical integration on the boundary
where the temperature is imposed. The effective thermal conductivities in the two directions are determined dividing the heat ux
by the correspond temperature gradient

kx

qx
DT x =Lx

48

qz

DT =L
z

T1

3. Thermophysical properties of the materials

The heat transfer through matrix and ber, Q_ is the same throughout
each cross section, in steady state. The heat transfer through the
rst
layer of matrix, with cross section 2a 2 4a2 and thickness s1 is
b
i
h

Q _ km

4. Numerical simulations

DT x

kx s1 s2 s3 km s1 s3

s s

36

The heat transfer in the x direction is then

Q_ x kx Ax

The effective thermal conductivity in the z direction, kz, is nally

T2

km

The heat transfer through the intermediate layer of ber, with cross
section (2s2 a) and thickness b is

Q_ kf 2s2 a

4a2

49

Fig. 2 presents the temperature distribution along the x direction in

case of copper reinforcement without consumption of the composite. The temperature on the hot side is 900 K and on the cold side
500 K. The isotherms are curved around the copper ber, in the center part of the composite, because the thermal conductivities of the

The heat transfer through the intermediate layer of ber, with the
same cross section 2a b2 4a2 but thickness s2 is

Q_ kf 2a
b

4a2

T2
T3
s2

43

The heat transfer through the last layer of matrix, with the same
cross section 2a b2 4a2 but thickness s3 is

Q_ km 2a
b

4a2

T3

T4

s3

44

two materials are different.

Table 1
Thermo-physical properties of silica, asbestos, steel and copper.
Material

Silica
Asbestos
Steel
Copper

2200
2000
7800
8930

[kg/m ]

c [J/(kg K)]

k [W/(m K)]

kf/km

1200
795
500
400

1.7
0.7
33.97
398

0.412
19.98
234.12

10

PIL
PFL
Num.

kx [W/mK]

0
0

Fig. 2. Temperature distribution in a composite with copper reinforcement without

consumption.

3
s [mm]

Fig. 4. Effective thermal conductivity in x direction, kx, for steel ber.

100

5. Theoretical results
5.1. Effective thermal conductivities in the two directions

PIL
PFL
Num.

The theoretical results are presented for a consuming composite

by increasing the thickness s of the consumption from s = 0 mm
to the
complete
consumption,
s = s1 + s2 + s3 = 6 mm,
assuming a = b = 2 mm. Figs. 35 show the theoretical and
numerical values of kx for the two thermal assumptions, PIL and
PFL, vs. the thick- ness s of the consumed composite.
Fig. 3 is relative to a composite made with the ber of asbestos.
The effective thermal conductivity, kx, decreases with the consumption of the rst layer of matrix, because the thermal conductivity of the ber is smaller than the silica one, reaching a
minimum when s = s1 and followed by an increase, up to the maximum value of the silica, when s = s1 + s2. The effective thermal
conductivity in x direction has
higher values for parallel
isothermal
lines, while the numerical simulations are intermediate between
the two thermal assumptions.
Figs. 4 and 5 report the effective thermal conductivity, kx, of a
composite made with the ber of steel or copper, respectively,
which have higher thermal conductivity than the silica matrix.
The effective thermal conductivity, kx, increases with the consumption of the rst layer of silica, because the thermal conductivity of

kx [W/mK]

80

60

40

20

0
0

s [mm]

Fig. 5. Effective thermal conductivity in x direction, kx, for copper ber.

silica is smaller than that of ber, reaching a maximum for s = s1

and followed by a decrease. Increasing the consumption of the central layer, the effective thermal conductivity, kx, decreases reaching
the minimum value of silica when s = s1 + s2. The effective thermal

1.8

1.8
1.7

1.7

k z[W/mK]

kx [W/mK]

1.6
1.6

PIL
PFL
Num.

1.5

PIL
PFL
Num.

1.5
1.4

1.4

1.3

1.3

1.2
0

3
s [mm]

Fig. 3. Effective thermal conductivity in x direction, kx, for asbestos ber.

3
s [mm]

Fig. 6. Effective thermal conductivity in z direction, kz, for asbestos ber.

The thermal conductivities kx and kz are different if the ber is

made of steel or copper because the composite, in the z direction,
is made of layers of silica and ber arranged prevalently in
series,
while the composite, in the x direction, has the layers of silica
and ber arranged prevalently in parallel.

PIL
PFL
Num.

kz [W/mK]

5.2. Anisotropy degree

2

The anisotropy degree xz of the composite is dened as the

ratio
1

0
0

s [mm]
Fig. 7. Effective thermal conductivity in z direction, kz, for steel ber.

PIL
PFL
Num.

kz [ W/mK]

kx
kz

50

It is function of the thickness s of the consumed composite and is

reported in Figs. 911 for the two thermal assumptions, PIL and PFL.
Fig. 9 reports the anisotropy degree of a composite with an
asbestos ber, which is decreasing from the value of the non-consumed composite to a relative minimum when the consumption of
the rst layer is completed, i.e. when s = s1. The consumption of
the central layer allows the anisotropy degree to increase,
reaching a maximum around the middle consumption of the
central layer and followed by a decrease up to minimum
value of 1, when s = s1 + s2.
Figs. 10 and 11 report the variation of the anisotropy degree
with the consumption of the composite when the ber is, respectively, steel or copper. The anisotropy degree increases with the
consumption of the layer of silica because its thermal conductivity
is smaller than the ber one, reaching a maximum at the
complete consumption of the upper layer, i.e. s = s1, and then
decreasing to the value 1 when s = s1 + s2.
Fig. 11 shows that the maximum anisotropy is obtained when
the ber is made of copper, reaching the maximum value of 25
when s = s1, which is an important information, useful in applications where strong anisotropy is required.
5.3. Anisotropic efciency

0
0

s [mm]
Fig. 8. Effective thermal conductivity in z direction, kz, for copper ber.

conductivity in x direction is higher for parallel isothermal lines,

while the numerical simulations are intermediate.
Figs. 68 present the results of the theoretical and numerical
effective thermal conductivities in the z direction kz.
Fig. 6 is relative to a composite with the ber made of asbestos.
The effective thermal conductivity, kz, decreases with the consumption of the rst layer of silica
because its thermal
conductivity is greater than that of asbestos, reaching a
minimum when s = s1 and followed by an increase, up to the
maximum value of the silica, when s = s1 + s2.
Figs. 7 and 8 present the effective thermal conductivity, kz, of a
composite with the ber made of steel or copper, respectively,
which have higher thermal conductivity than the silica matrix.
The theoretical results of Figs. 7 and 8 show an increase of kz with
the consumption of the rst layer of silica, because its thermal
conductivity is smaller than that of the ber, obtaining a maximum
when s = s1, and followed by a decrease. With the increase of the
consumption of the central layer the effective thermal conductivity, kz, decreases reaching the minimum value of the silica when
s = s1 + s2. The effective thermal conductivity in x direction is
higher for parallel isothermal lines, while the numerical simula-

The potential anisotropy of the materials, d, is dened as the

ratio between the ber and the matrix thermal conductivities, i.e.
of the two materials employed in the composite

kf
km

51

The potential anisotropy of the materials, d, allows to evaluate the

ratio between the anisotropy degree of the composite, n, and the
potential anisotropy of the materials, d, and is indicated by v

v d

52

This new parameter, v measures the anisotropic efciency of the

composite, obtained by the employment of the two materials, comparing the anisotropy degree of the composite with the potential
anisotropy of the two materials.
The present theoretical approach allows to evaluate the anisotropic efciency under the two thermal assumptions of PIL and
PFL. The anisotropic efciency under the thermal assumption PIL,
vPIL, can be evaluated by Eqs. (18) and (29) giving

vPIL

nPIL kx =kz PIL

d
kf =km
2

s1 s3 s2 2a b

n
b

2
2

4a d 2a

2a s1 s2 s32
s1 s2 s3 2ab b s2 =ds2 b

4a

b s1 s2 2 s3

2ab s1 s3 =ds2

53
tions are intermediate.

The anisotropic efciency under the thermal assumption PFL, vPFL,

can be evaluated by the Eqs. (39) and (46) giving

1.06

PIL
PFL

1.05
1.04

1.03
1.02
1.01
1.00
0.99
0.98
0

3
s [mm]

Fig. 9. Anisotropy degree of the composite for asbestos ber.

30

PIL
PFL

PIL
PFL

25

20

15
10

5
0

0
0

s1 s3
d

s1 s2 s3
2a b

b s2
s2

2ab
2
2
4a
s1 s2 s3

Fig. 11. Anisotropy degree of the composite for copper ber.

Fig. 10. Anisotropy degree of the composite for steel ber.

nPFL kx =kz PFL

d
kf =km

s [mm]

s [mm]

vPFL

b
d

2a d

54

2.60

4a bbs2
s1 s3

PIL
PFL

2.55

The resulting anisotropic efciencies, vPIL and vPFL, are reported in

Figs. 1214 vs. the consumption of the three composites
investigated.
Fig. 12 is relative to a composite with an asbestos ber, i.e.
with the potential anisotropy of the two materials equal to
d = kf/km = 0.412. The anisotropic efciency v decreases with the
consumption of the rst layer to a minimum when s = s1, then v
increases again up to a maximum, around the middle consumption
of the central layer, and is followed by a decrease to the minimum,
equal to v = 1/d = 1/0.412 = 2.427, when s = s1 + s2.
Fig. 13 is relative to a composite with a steel ber, i.e. with the
potential anisotropy of the two materials equal to d = kf/km = 19.98.
For a potential anisotropy greater than 1, as in Fig. 13, the anisotropic efciency v increases slightly with the consumption of the
rst layer of silica to a maximum, when s = s1, then decreases to

2.50

2.45

2.40
0

s [mm]
Fig. 12. Anisotropic efciency v of the composite with asbestos ber.

0.20

100

PIL
PFL

PIL
PFL
10

0.15

0.10

0.05

0.1

0.00

0.01
0

0.01

10

100

1000

s [mm]

Fig. 15. Anisotropic efciencies, vPIL and vPFL, of the composite vs. the potential
anisotropy, d.

Fig. 13. Anisotropic efciency v of the composite with steel ber.

An important conclusion of Eqs. (55) and (56) is that the minimum

anisotropic efciencies of the composite, vPIL (min) and vPFL (min),
are only function of the geometry of the composite.
Fig. 15 presents the variation of the anisotropic efciencies, vPIL
and vPFL, vs. the potential anisotropy, d, from d = 0.01 until
d = 1000, for the same geometry of the composite of the previous
gures, i.e. s1 = s2 = s2 = s3 = a = b = 2 mm. The decrease of the
anisotropic efciency of the composite v with the increase of the
potential anisotropy d shows that the composite becomes less efcient with the increase of the potential anisotropy of the two
mate- rials. The two anisotropic efciencies are greater than
100 for d = 0.01 and they are practically constant for d > 100. It is
also vis- ible a quasi-linear trend between d = 0.01 and d = 10,
which brings to the conclusion that is not convenient to use
couples of materials with a potential ratio greater than d = 10.

0.20

PIL
PFL
0.15

0.1

0.10

0.05

0.00
0

6. Conclusions

s [mm]
Fig. 14. Anisotropic efciency v of the composite with copper ber.

the minimum value, equal to v = 1/d = 1/19.98 = 0.05, when

s = s 1 + s2 .
Fig. 14 is relative to a composite with a copper ber, i.e. with
the potential anisotropy of the two materials equal to d
= kf/ km = 234.12. For a high potential anisotropy of the materials
d, as in Fig. 14, the anisotropic efciency v increases slightly
with the
consumption of the rst layer of silica to a maximum, when
s = s1, then decreases to the minimum value, equal to v = 1/d =
1/
234.12 = 0.0043 in Fig. 14, when s = s2 + s3.
The general trend of the anisotropic efciencies, vPIL and vPFL, is
to decrease with the increase of the potential anisotropy of the
two materials, d. Eqs. (53) and (54) show that both vPIL and vPFL
tend to
+1 for d tending to 0. If the potential anisotropy d tends to +1, the
anisotropic efciency vPIL, given by Eq. (53), tends to a minimum
value vPIL (min), given by

vPIL min

s2 s1

s3 2a

The anisotropic efciency vPFL, given by Eq. (54), tends to a

minimum value vPFL (min), given by

s2 s1

s3 2a

The two thermal assumptions give the extreme theoretical

values.
The greater values are obtained for PIL thermal
assumption.
The numerical simulations are intermediate between the
two theoretical ones.
The maximum effective thermal conductivity is obtained
for a copper ber and a consumption corresponding to
the rst layer of silica matrix.
The maximum effective thermal conductivity is obtained
for a copper ber in the x direction, kx.

55

s1 s2 s3 2 2a2a b b2

The effective thermal conductivities in x and z directions, kx and

kz, are determined for a ber-reinforced composite. The study is
carried on theoretically under the two assumptions of parallel isothermal lines (PIL) and parallel heat ux lines (PFL), and numerically with a house made code, for three different composites,
made with silica matrix and three different bers.
The composite is investigated when is virgin and during the
increasing consumption. The main conclusions for the effective
thermal conductivity are the followings:

The anisotropy degree is dened as n

sions are the followings:

kx
kz

and the main conclu-

The maximum anisotropy degree is obtained for a copper

vPFL min

s1 s2 s3 s1 s3 2a
b

s2 4ab b

56

ber and a consumption corresponding to the rst layer

of silica matrix.

The maximum anisotropy degree is obtained when the

volume fraction of the ber is maximum.
The anisotropy degree increases with the increase of the
k
potential anisotropy of the materials, d km f .

The anisotropic efciency of the composite is dened as v dn,

can be evaluated theoretically with the two thermal assumptions
and the main conclusions are the followings:

The anisotropic efciency of the composite decreases with

the increase of the potential anisotropy of the two materials, d.
The anisotropic efciency of the composite decreases linearly with the increase of the potential anisotropy of the
two materials, d, from d = 0.01 to d = 10.
The anisotropic efciency reaches a minimum for d > 100,
which is dependent only on the geometrical parameters of
the composite.

Conict of interest
None declared.
References
[1] S. Torquato, I.C. Kim, Efcient simulation technique to compute effective
properties of heterogeneous media, Appl. Phys. Lett. 55 (18) (1989) 1847
1849.
[2] C.R. Havis, G.P. Peterson, L.S. Fletcher, Predicting the thermal conductivity and
temperature distribution in aligned ber composites, J. Thermophys. Heat
Transfer 3 (1989) 416422.
[3] S. Graham, D. McDowell, Numerical analysis of the transverse thermal
conductivity of composites with imperfect interfaces, J. Heat Transfer 125
(2003) 389393.
[4] M.R. Islam, A. Pramila, Thermal conductivity of ber reinforced composite by
FEM, J. Compos. Mater. 33 (18) (1999) 16991715.
[5] A. Dasgupta, R. Agarwal, Orthotropic thermal conductivity of plain-weave
fabric composites, J. Compos. Mater. 26 (18) (1992) 27362750.
[6] G.Q. Gu, R. Tao, Effective thermal conductivity of a periodic composite with
contact resistance, J. Appl. Phys. 64 (6) (1988) 29682972.
[7] B.W. James, P. Harrison, Analysis of the temperature distribution, heat ow and
effective thermal conductivity of homogeneous composite materials with
anisotropic thermal conductivity, J. Phys. D: Appl. Phys. 25 (1992) 12981303.
[8] S. Torquato, M.D. Rintoul, Effect of the interface on the properties of composite
media, Phys. Rev. Lett. 75 (22) (1995) 40674070.
[9] C.Y. Wang, Thermal conductivity of a material containing a layer of thin strips
or a staggered array of thin strips, J. Heat Transfer 121 (1) (1999) 174177.
[10] H. Shim, M. Seo, S. Park, Thermal conductivity and mechanical properties of
various cross-section types carbon ber-reinforced composite, J. Mater. Sci. 35
(9) (2002) 18811885.
[11] T.C. Lim, Unied practical bounds for the thermal conductivity of composite
materials, Mater. Lett. 54 (2002) 152157.
[12] J.M. Terron, A. Sanchez Lavega, On the effective thermal diffusivity of berreinforced composites, Appl. Phys. Lett. 80 (11) (2002) 19031905.
[13] A. Salazar, J.M. Terron, A. Sanchez Lavega, R. Celorrio, On the effective thermal
diffusivity of ber-reinforced composites, Appl. Phys. Lett. 80 (11) (2002)
19031905.
[14] V.V. Skorokhod, Layered composites: structural classication, thermophysical
and mechanical properties, Powder Metall. Met. Ceram. 42 (910) (2003)
437
446.
[15] J.D. Felske, Effective thermal conductivity of composite spheres in a
continuous medium with contact resistance, Int. J. Heat Mass Transfer 47
(2004) 34533461.
[16] J.M. Goyhnche, A. Cosculluela, A multiscale model for the effective thermal
conductivity tensor of a stratied composite material, Int. J. Thermophys. 26
(1) (2005) 191202.
[17] Y.N. Al-Nassar, Prediction of thermal conductivity of air voided-berreinforced composite laminates, Part II: 3D simulation, Heat Mass Transfer
43 (2006) 117122.

[18] C. Yuan, X. Luo, A unit cell approach to compute thermal conductivity of

uncured silicone/phosphor composites, Int. J. Heat Mass Transfer 56 (2013)
206213.
[19] F. Gori, S. Corasaniti, W.M. Worek, W.J. Minkowycz, Theoretical and Numerical
Predictions of the Thermal Conductivity of Thermal Protection Systems, in:
Proceedings of the ASME-ATI-UIT 2010 Conference on Thermal and
Environmental Issues in Energy Systems, Sorrento, Italy, 1619 May, II,
2010, pp. 957962.
[20] F. Gori, J.F. Ciparisse, S. Corasaniti, Thermal anisotropic properties of composite
materials, in: Proceedings of the IEEE Aerospace Conference, Big Sky Montana,
March 310, IEEE Catalog number: CFP12AAC-CDR, 2012.
[21] F. Gori, S. Corasaniti, W.M. Worek, W.J. Minkowycz, Theoretical prediction of
thermal conductivity for thermal protection systems, Appl. Therm. Eng. 49
(2012) 124130.
[22] F. Gori, S. Corasaniti, J.F. Ciparisse, Theoretical prediction of the anisotropic
effective thermal conductivity of composite materials, in: Proceedings of the
ASME International Mechanical Engineering Conference and Exposition,
IMECE, November 915, Houston, Texas, USA, IMECE2012-86114, 2012, pp.
9196.
[23] F. Gori, A theoretical model for predicting the effective thermal conductivity of
unsaturated frozen soils, in: Proceedings of the 4th International Conference
on Permafrost, Fairbanks, USA, 1983, pp. 363368.
[24] F. Gori, On the theoretical prediction of the effective thermal conductivity of
bricks, in: Proceedings of the Eight International Heat Transfer Conference, vol.
II, S. Francisco, USA, 1986, pp. 627632.
[25] F. Gori, S. Corasaniti, Considerations on a theoretical method for the prediction
of the thermal conductivity of porous media, in: 17th UIT National Heat
Transfer Conference, vol. II, 1999, pp. 683694.
[26] F. Gori, C. Marino, M. Pietrafesa, Experimental measurements and theoretical
predictions of the effective thermal conductivity of porous saturated two
phases media, Int. Commun. Heat Mass Transfer 28 (8) (2001) 10911102.
[27] F. Gori, S. Corasaniti, Theoretical Thermal conductivity of soils at high
temperatures, in: Proceedings of the ASME American Society of Mechanical
Engineers, Heat Transfer Division, (Publication) HTD, 369 (2) (2001) 2734,
IMECE, International Mechanical Engineering Congress and Exposition, HTD24152, ASME, 2001, pp. 18.
[28] F. Gori, S. Corasaniti, Theoretical prediction of the thermal conductivity of soils
at moderately high temperatures, J. Heat Transfer ASME 124 (6) (2002) 1001
1008.
[29] F. Gori, S. Corasaniti, Experimental measurements and theoretical prediction of
the thermal conductivity of two-and three-phase water/olivine systems, Int. J.
Thermophys. 24 (5) (2003) 13391353.
[30] F. Gori, S. Corasaniti, Theoretical Prediction of the Effective Thermal
Conductivity of Particulate Materials in Extraterrestrial Conditions and of
Foams at Low Density, Microgravity and Space Station Utilization, 2 (2-3-4)
JuneDecember 2001, pp. 2324.
[31] F. Gori, S. Corasaniti, Theoretical prediction of the thermal conductivity and
temperature variation inside mars soil analogues, Planet. Space Sci. 52 (13)
(2004) 9199.
[32] F. Gori, S. Corasaniti, Thermal properties and temperature variations in
Martian soil analogues, in: N.S. Maravell (Ed.), Space Science, New Research,
Nova Science Publishers, New York, USA, 2006. Chapter 6.
[33] F. Gori, S. Corasaniti, Detection of a dryfrozen boundary inside Martian
regolith, Planet. Space Sci. 56 (2008) 10931102.
[34] V.R. Tarnawski, F. Gori, B. Wagner, G.D. Buchan, Modelling approaches to
predicting thermal conductivity of soils at high temperatures, Int. J. Energy
Res. 24 (5) (2000) 403423.
[35] V.R. Tarnawski, F. Gori, Enhancement of the cubic cell soil thermal
conductivity model, Int. J. Energy Res. 26 (2) (2002) 143157.
[36] V.R. Tarnawski, W.H. Leong, F. Gori, G.D. Buchan, J. Sundberg, Inter-particle
contact heat transfer in soil systems at moderate temperatures, Int. J. Energy
Res. 26 (2002) 13451358.
[37] V.R. Tarnawski, D.J. Cleland, S. Corasaniti, F. Gori, R.H. Mascheroni, Extension of
soil thermal conductivity models to frozen meats with low and high fat
contents, Int. J. Refrig. 28 (6) (2005) 840850.
[38] W.H. Leong, V.R. Tarnawski, F. Gori, G.D. Buchan, J. Sundberg, Inter-particle
contact heat transfer models: an extension to soils at elevated temperatures,
Int. J. Energy Res. 29 (2) (2005) 131144.
[39] F. Gori, S. Corasaniti, New model to evaluate the effective thermal conductivity
of three-phase soils, Int. Commun. Heat Mass Transfer 47 (2013) 16.
[40] F. Gori, S. Corasaniti, Effective thermal conductivity of three-phase soils, in:
Proceedings of ASME 2012 International Mechanical Engineering Conference
and Exposition, IMECE, November 915, Houston, Texas, USA, IMECE201286304, 2012, pp. 23692374.
[41] M. Thomas, N. Boyard, Y. Jarny, D. Delaunay, Estimation of effective thermal
conductivity tensor from composite microstructure images, J. Phys.: Conf. Ser.
135 (2008) 012097.