B5.

9EH2: Mechanical Engineering Science 8

Part 1: Mechanics of Materials
Topic 1: Mechanics of composite materials

Sub-topics:

1.0 Modulus of unidirectional continuous fibre reinforced composites

1.1 Strength of unidirectional continuous fibre reinforced composites
1.2 Discontinuous fibre composites
1.3 Simple laminate theory
1.4 Tutorial solutions

References:
McCrum, Buckley and Bucknall. “Principles of polymer engineering”, Oxford University Press
Voyiadjis and Kattan. “Mechanics of Composite Materials with MATLAB”, Springer Verlag

General definitions:

Matrix: Normally given subscript m, this is the background material into which the reinforcement is
embedded. Matrices can be metal, ceramic or polymer. We will only consider polymers in this module.
Most polymers which are reinforced with fibres are thermosets (as opposed to thermoplastics).

Reinforcement: This can be in the form of particles, flakes or fibres. Here we will consider only fibre
reinforcement, given the subscript f. Fibres can be aligned, random, or arranged in stacks, called
laminae. Fibres can be continuous (passing right through the component) or discontinuous (i.e. short
fibres).

Convention for stress and strain axes relative to fibre direction:

3
2
1

(Aligned) fibre direction

1.0: Modulus of unidirectional, continuous fibre reinforced composites

If a stress acts along direction 1, the fibres and matrix are strained in parallel and therefore will suffer
the same axial strain for a given stress:

1   f   m
1 1

For continuous fibres, the stress in any section is the weighted sum (according to volume fraction Vf) of
the stresses in fibres and matrix:

 1  V f  f  (1  V f ) m
1 1

If both components are linear elastic:

 f  Ef f
1 1
;  m  Em m
1 1

Also, the overall stress:

  1  E1 1

Thus:

E1 1  V f E f  f  1  V f Em m
1
  1

But:
1   f   m 1 1

So:
E1  V f E f  1  V f Em  

If the stress is acting in the 2 direction, the fibres and matrix are in series and can therefore have
different strain, but must carry the same stress:

2  f m 2 2

In this case, the total strain can be calculated as the weighted sum of the strains in the components:

 2  V f  f  1  V f  m
2 2

Again, if the components are linear elastic:

2 f m
E2
Vf
Ef
2
 1Vf  E 2

So:

1 Vf 1Vf  
E mV f  (1  V f ) E f
  
E2 E f Em Em E f

i.e.

Em E f
E2 
E mV f  (1  V f ) E f

1.1: Strength of unidirectional, continuous fibre reinforced composites

The strength of unidirectional composites along the 1-direction can be assessed using the parallel
loading model:
 Fibres

1 Matrix 1

where the stress is shared between the fibres and the matrix in proportion to their volume fractions:

 
1  V f  f  1  V f  m

and the strain in the composite is the same as in the components:

1   f   m

(The subscript “1” is understood for the fibres and matrix.)

Two distinct cases can be identified for composite strengthening; the case of brittle fibres, where the
fibres fail at a lower strain than the matrix, and the case of a brittle matrix, where the matrix fails at a
lower strain than the fibres.

f*
f* Fibre stress-
strain curve
Fibre stress-
strain curve
Stress

Stress

Matrix stress-
strain curve
m* m* Matrix stress-
strain curve

 f* Strain m* m* Strain  f*

1: Brittle fibres 2: Brittle matrix

Brittle fibres:

Here, the fibres fail at a lower strain than the matrix so, the fibres will fail when the composite strain 1
is at the failure strain of the fibres, f*. At low volume fractions of fibres, the matrix has sufficient
load-bearing capacity to continue to carry the entire stress (1, below) and the matrix will continue to
strain until the composite strain reaches the matrix failure strain, m*. Thus, the stress at which the
composite fails is:
 1  (1  V f ) m

(low Vf)

At higher volume fractions of fibres, the matrix is not able to carry the entire load once the fibre failure
strain is reached (2, below). Thus the whole composite fails when the fibres fail and the failure stress is
given by:


1  1  V f  m 
  V f  f (high Vf)

where m is the matrix stress corresponding to a strain of f*.

Matrix fails Fibres fail, and

(1-Vf)*m matrix unable to
1* withstand the stress
(1-Vf)m+Vf)*f
Fibres fail 1* Composite
Composite
stress-strain
stress-strain
curve
curve

Stress
Stress

(1-Vf)m
(matrix component)
Vff (fibre
component)

(1-Vf)m (matrix
Vff (fibre component) component)

 f* Strain m* f*, m Strain m*

1: Low Vf 2: High Vf

It is therefore possible to obtain two curves for the strength of the entire composite series:
Strength, 1*

*f


1  1  V f  m 
  V f  f

*m

1  (1  V f ) m


m

V f V f
Fibre volume fraction, Vf

Brittle matrix:
Here, the matrix fails at a lower strain than the fibres. At low volume fractions of fibres, the whole
composite will fail when the matrix fails (1, below) and thus, the stress at which the composite fails is:
1  1  V f  m


 V f  f (low Vf) 
where f is the fibre stress corresponding to a strain of m*.
 At higher volume fractions of fibres, the fibres are able to carry the entire load after the matrix failure
strain is reached (2, below). Thus the whole composite does not fail until the fibres fail and the failure
stress is given by: 1  V f  f (high Vf)

Fibres fail
Matrix fails, and
fibres unable to
Vf*f
1* withstand the stress 1* Matrix fails
Vff+(1-Vf)*m
Composite
stress-strain

Stress
Stress

Composite
curve
stress-strain
(1-Vf)m curve
(matrix component)
Vff (fibre
component)

Vff (fibre component) (1-Vf)m (matrix

component)
m*, f Strain  f* Strain  f*
m*
1: Low Vf 2: High Vf

Again, we can draw two curves for the strength of the entire composite series:

f*

1  V f  f
Strength, 1*

f’


1  1  V f  m

 V f  f
m*

V f’
Fibre volume fraction, Vf

The changeover from the low to the high volume fraction is given (as before) by the intersection of the
two lines, i.e. when:


m
V f 
 f   m

  f
 1.2 Short Unidirectional Fibre Reinforcement

The most effective strengthening is obtained by using continuous fibres, but this is sometimes not
possible to achieve, for example when using injection moulding. The difference with discontinuous
fibres is that the stress is no longer constant along the fibre, and is transferred to the fibre by shear
across the fibre-matrix interface:

i
 
i
-l/2 0 x l/2

It can be shown (shear-lag theory, Section 6.N.7, McCrum et al) that the stress along the fibre is given
by:



 f  E f 1 1  

 cosh na 2 x 
l 


  cosh na   
  

where

2Gm
n
E f ln 2 R  d

the aspect ratio

a l
d

and 2R is the distance between nearest neighbour fibre centres.

The interface shear stress can also be represented along the fibre:

i 
n
E f 1

sinh na 2 x
l

2 cosh na 
Activity: Using values of na of 1, 3 and 10, try to reproduce the following plots of normalised stress
(f/Ef1 and i/n Ef1) along the fibre length. What is the meaning of a zero normalised interfacial shear
stress, and the meaning of a normalised direct stress of unity?

0.9
na = 1
Normalised stress 0.7
na = 3
0.5
na = 10
0.3
na = 1
0.1
na = 3
-0.5 -0.1 0 0.5
na = 10
-0.3
-0.5
Distance along from centre of fibre (x/l)

Calculating the tensile strength and modulus for discontinuous fibres is slightly different in that a cross-
section does not expose the same stress in each of the fibres:

So that:

1  V f  f  1  V f  m 
and the mean fibre stress which is exposed is given by:

l
1 2
f    f dx
l l
2

Have a go at the integration, and see if you can show that:

 tanh na  
 f  E f 1 1 
 na 

The modulus is obtained by dividing the composite stress, 1, by the composite strain, 1:

 tanh( na) 
E1  1V f E f  (1  V f ) Em , where 1  1   
 na 
As the strain in the fibres is increased, the maximum interfacial shear stress (i/nEf1 = ½) increases
until it reaches the strength of the interface, i*. It can be shown that the interfacial shear stress is
constant for a length given by:

E f  1d

4 i

fmax

i*

 

0 x
 
(debonded
region)

As strain increases, the length of the debonded regions increases. When the strain reaches the fibre
failure strain:
 f
 1   f  Ef

the debonded length will have reached:

 f
 

d
4 i

However, if * is greater than half the length of the fibre, the fibre failure stress cannot be reached
before the fibre becomes debonded from the matrix. Therefore, there is a critical fibre length, below
which strengthening is sub-optimal:

 f
lc  2   d
2 i

The tensile strength of the composite series again depends on volume fraction and whether or not the
fibres are more brittle than the matrix. The following analysis assumes that l  lc

For brittle fibres  m   f (the other case of relatively brittle matrix is very rare):

For low Vf, at a given cross-section, not all the fibres will reach the failure stress, because some will
have failed at the interface:
The proportion of fibres in the pull-out category is lc/l, and the pull-out will be resisted by the friction
stress i*. The fibres in this category will have an average stress of f*/2, and the remaining fibres will
have failed and will be carrying no stress. Thus:

1 
lc
2l

V f  f  1  V f  m


Actually this gives an increased strength over the equivalent continuous fibre composite (by the
amount of frictional resistance offered by the pulled out fibre ends.

At higher volume fractions, composite failure occurs when all the fibres and the matrix are carrying
load:  1  V f  f  1  V f  m

and so the average fibre stress at fibre failure will be:


 lc   lc  f
  f 1  c 
l
 f  1   f 
 l  l 2  2l 

The composite failure stress is then:

 lc 
 1  1  
V f  f  1  V f  m
2l 



Again, the strength of the composite series can be shown in a graph:

f*
Strength, 1*

f*(1-lc/2l)
 l 

  1  c V f  f  1  V f  m

1 
 2l 

f*(lc/2l)

m*
1 
lc

V f  f  1  V f  m


m 2l

Fibre volume fraction, Vf

It can be seen that:

 Discontinuous fibre composites are stronger than the equivalent continuous fibre composites at
low volume fractions of fibres
 The low volume fraction regime is longer for discontinuous fibers
 The strengthening effect of the fibres (in the high volume fraction regime) is dependent upon the
length of the fibres relative to the critical length
 Tutorial 1
Questions on Unidirectional FRPs
(Covering Topics 1.1 and 1.2)

1. The tensile modulus of a thermoset polyester matrix is 3 GPa and that of E-glass reinforcing fibres is
76 GPa. Calculate the tensile modulus for the composite series (Vf = 0 to 1) where the fibres are
continuous and a stress is applied in each of the directions parallel (E1) and transverse (E2) to the fibres.

2. A composite is made from 50% (by volume) of carbon fibres (uniaxial tensile modulus 230 GPa and
uniaxial tensile strength 3200 MPa) continuously uniaxially aligned in an epoxy matrix (uniaxial
tensile modulus 2.4 GPa and uniaxial tensile strength 60 MPa).

(i) Calculate the tensile strength and tensile modulus of the composite.

(ii) Find the minimum volume fraction of fibres required to produce strengthening of the composite and
calculate the tensile modulus at that fibre loading.

3. Calculate the tensile strength for the E-glass / polyester composite series in Q1 where the stress is
applied parallel to the fibre direction. The tensile strengths of the fibres and the matrix are 1800MPa
and 55MPa, respectively.

4. A nylon 6.6 bar (uniaxial tensile modulus 2.7 GPa, shear modulus 1.015 GPa and uniaxial tensile
strength 70 MPa) is reinforced by 20% by volume of carbon fibres (uniaxial tensile modulus 230 GPa
and uniaxial tensile strength 3200 MPa). The bar is produced by injection moulding in such a way that
the fibres (of length 400 m and diameter 6 m) are aligned along the axis of the bar.

(i) Find the mean tensile stress carried by the fibres and the overall stress carried by the bar when a
tensile strain of 10-4 is applied to it.

(ii) Calculate the tensile strength of the bar and the mean fibre stress at failure. Find also the tensile
strengths for composites containing 30% by volume of fibres of the same size and 30% by volume of
fibres of length 600 m. The shear strength of the fibre-matrix interface can be taken to be 32 MPa.
 1.3 Orthotropic laminate analysis for elastic modulus

Laminate analysis is applicable to composites which are made up of a series of plies (laminae) which
are unidirectional reinforced and are oriented in various directions:

z
y
x

As before a set of co-ordinate axes needs to be set up, and these are now the material axes against
which all the orientations are measured. For example, the above laminate has two plies, one of whose
fibres are oriented at 0 to the material axis x, and the other at 90 to the axis, and would be referred to
as a [0/90] laminate. For more complex laminates, a shorthand description has evolved, using a
subscript to designate repeated layers and symmetry. For example:

[0/0/45/-45/90/90/-45/45/0/0] becomes [02/45/90]s

and

[45/-45/0/0/45/-45/0/0/0/0/-45/45/0/0/-45/45] becomes [45/02]2s

The elastic properties of the laminae are normally based on measures of the shear modulus, bulk
modulus and Poisson’s ratios of the fibres and the matrix, respectively; Gm,f, Km,f and m,f, respectively.
The following moduli can be derived:
E1

12
23

k*
E2=E3

G12
E1

G23

k 
 
K m K f  Gm Vm  K f K m  Gm V f
K f  Gm Vm  Km  Gm V f
 E1  EmVm  E f V f

 f   m  1 Km  1 
K f  m f
V V
12   mVm   f V f  
Vm Vf
  1
Kf Km Gm

Vf
G12  Gm 
G f  Gm  
1 Vm
2Gm

 
 
 1   m V f 
G23  Gm 1   for Gf > Gm and Kf > Km
  3 2 V 2  
   1 
 V f

m m

  V f  1 
3


 
 
 1   m V f 
G23  Gm 1   for Gf < Gm and Kf < Km
  3 2 V 2  
   1 
  V f 
m m
  V f3   m  

 m   f   m
where  
1   f  1
1 1 Gf
m  f  and 
3  4 m 3  4 f Gm

4 1 1 4 2 E2,3
Finally,     12 and  23  1
E2,3 G23 k E1 2G23

where Vm and Vf are volume fractions of fibre and matrix, respectively.

The stresses and strains in the laminae are first referred to the laminae axes:

33

11

22

For laminae that are in plane stress, i.e. 13   23   33  0 :

 1  12 
 0 
 11   E1 E1   11 
    12  
  22   
1
0   22 

OR l   S  l 
2   E1 E2  
 12 
 0 1   12 
0
 G12 

where the subscript l refers to lamina co-ordinates and the matrix [S] is the compliance matrix, which
relates stresses and strains in the principal lamina directions, 1 and 2. For plane stress, the remaining
strains can be found to be:
13  23
 23  32  0 , and  33  11   22
E1 E2

The above matrix relationship can be inverted to give:

 l   S 1l   Ql 
where [Q] is known as the reduced lamina stiffness matrix, and is the inverse of the compliance matrix.

Thus, [Q] can be found:

E1
Q11 
Z
E
Q11 Q12 0  Q22  2
Q  Q12 Q22 0 
Z
12 E2
 0 0 Q66  Q12 
Z
E1  12
2
E2
Z
E1
Q66  G12

Most laminates involve laminae with different orientations, so it is necessary to know the lamina
stresses and strains in directions other than the principal ones, 1 and 2.

Defining the laminate co-ordinates x and y:

The stresses can be transformed:

 xx  m 2 n 2  2mn   11 
   2  
 yy    n m 2
2mn   22  OR  x    l 
   mn  mn m 2  n 2   
 xy     12 

where m  cos  and n  sin

Also, the strains can be transformed using a slightly different matrix:

  xx   m 2 n2  mn   11 
    
  yy    n
2
m 2
mn    22  OR  x   l 
2  2mn  2mn m 2  n 2  2 
 xy     12 

Thus, the stress-strain relationship can be written in terms of laminate co-ordinates:

 x   Q  x  
where Q  Q 1

Q11  Q11m 4  Q22 n 4  2m 2 n 2 Q12  2Q66 


Q12  m 2 n 2 Q11  Q22  4Q66   m 4  n 4 Q12 
 
Q11 Q12

Q  Q12 Q22
Q16 
Q26 
  2 2

Q16  Q11m  Q22 n  Q12  2Q66  m  n 2 mn 2

Q16 Q26 Q66  Q22  Q11n 4  Q22 m 4  2m 2 n 2 Q12  2Q66 

 
Q26  Q11n 2  Q22 m 2  Q12  2Q66  m 2  n 2 mn 
Q66  Q11  Q22  2Q12 m 2 n 2  Q66 m 2
n 2 2

m  cos 
n  sin

Equally, the compliance matrix can also be transformed:

 l   S  l   x   S 1 x   S  x 
and so lamina stiffnesses can also be defined:

  xx   S11 S12 S16   xx 

    
  yy   S21 S22 S26   yy 
2   S S66   xy 
 xy   16 S26

Finally, the lamina properties can be defined in terms of either of the reduced stiffness or compliance
matrices as follows:

Q12 2 1 Q122 1
E x  Q11   E y  Q22  
Q22 S11 Q11 S22

1 Q12 Q21 S S
Gxy  Q66   xy     21   12
S66 Q22 Q22 S11 S11

Once the lamina properties are known, it is possible to define the overall properties of the laminate,
subject to the following assumptions:

 the plate thickness is much smaller than the in-plane dimensions

 the strains are much less than unity
 normals to the undeformed surface remain normal to the deformed surface
 vertical deflection does not vary through the section thickness
 stress normal to the plate surface is negligible

A laminate is built up of a series of laminae as follows:

and the theory considers the forces and moments at the edges of an element:
The stress and moment resultants in the element defined in the above diagram can be written in terms
of the mid-surface strain 0 and the curvature  as:

 N xx 
    
N    N yy    Q  0  z  dz  A  0  B 
h

 N  h
 xy 

and

 M xx 
    
M   M yy    Q  0  z  z  dz  B  0  D 
h

M   h
 xy 

which can be summarised as:

 N   A B   0 
   
M   B D   

where:

 A11 A12 A16 B11 B12 B16 

A B26 
 21 A22 A26 B21 B22
 A B   A16 A26 A66 B16 B26 B66 
 B D   B 
   11 B12 B16 D11 D12 D16 
 B21 B22 B26 D21 D22 D26 
 
 B16 B26 B66 D16 D26 D66 

The integrals required to produce the matrices [A], [B] and [D] are not continuous, because the
transformed lamina stiffness matrices do not vary through the lamina thickness and so they can be
written as the following 3  3 matrices:

    
K K
A   Q i zi  zi 1  B  1  Q i zi2  zi21
i 1 2 i 1

D  1  Q i zi3  zi31 

K

3 i 1
Most structural laminates are symmetrical so that cross-coupling between bending and tension is
eliminated (i.e. the [B] matrix is eliminated) and, confining ourselves to extensional behaviour (i.e. no
bending):

 N xx   A11 A12 A16    xx0 

   
 
 N yy    A12 A22 A26    0yy 
N   A A66  2 xy
0 
 xy   16 A26
 

or:

  A
0 1
N  aN
and hence the overall laminate elastic constants can be expressed as:

1 1 1 a12
Ex  Ey  Gxy   xy  
2ha11 2ha22 2ha66 a11

where 2h is the laminate thickness.

 Reminder on matrices, determinants and solution of linear equations

Matrix definition: row  column (mn):

a11 a12 .. a1n

a21 a22 .. a2n
. . .. .
. . .. .
am1 am 2 .. amn

Multiplication:

(mp)  (pn) = (mn), for example:

 a11 a12  a b a b a11b12  a12b22 

   b11 b12   11 11 12 21 
 21 22  
a a   
   21 11
a b  a b
22 21 a21b12  a22b22 
a   b21 b22   a b  a b 
 31 a32   31 11 32 21 a31b12  a32b22 

Inversion:

The reciprocal of a matrix is its adjoint divided by its determinant:

adjA
A1 
A

For a third order matrix:

 a1 b1 c1 
 
A   a2 b2 c2 
a b c3 
 3 3

the determinant:

a1 b1 c1
D3  a2 b2 c2  a1b2c3  a1b3c2  a2b3c1  a2b1c3  a3b1c2  a3b2c1
a3 b3 c3

which can be written:

b2 c2 b1 c1 b1 c1
D3  a1  a2  a3
b3 c3 b3 c3 b2 c2

or:
a2 c2 a1 c1 a1 c1
D3  b1  b2  b3
a3 c3 a3 c3 a2 c2
or:
a2 b2 a1 b1 a1 b1
D3  c1  c2  c3
a3 b3 a3 b3 a2 b2

These secondary determinants (along with their signs) are the co-factors of each element of the matrix,
e.g.:

b2 c2
A1 
b3 c3
a1 c1
B3  
a2 c2
a1 b1
C2  
a3 b3

The adjoint is the transpose of the matrix of the cofactors, i.e.:

 A1 A2 A3 
 
adjA   B1 B2 B3 
C C C 
 1 2 3

Solving linear equations:

If we have n equations with n unknowns, x1, x2, ..., xn:

a11x1  a12 x2  ...  a1n xn  h1

a21x1  a22 x2  ...  a2n xn  h2
............................................
an1x1  an 2 x2  ...  ann xn  hn

we can write these in matrix form:

 a11 .. a1n   x1   h1 
     
 . .. .    .    . 
a     
 n1 .. ann   xn   hn 

or AX = H

which can be solved as: X = A-1H

 Tutorial 2
Exercises in laminate analysis (Topic 1.3)

1. A composite consists of 60% (by volume) continuous, uniaxially aligned glass fibres in a
matrix of epoxy resin. The tensile modulus and Poisson’s ratio for the glass and the epoxy are 76 GPa
and 0.22 and 2.4 GPa and 0.34, respectively. A tensile stress of 150 MPa is applied at 30° to the fibre
direction as shown below:

xx

30°

Calculate the reduced plane-stress stiffness matrix, Q  , for the lamina and hence determine the matrix
 
for the stress and strain components in the x-y co-ordinate system, Q . Invert this matrix to obtain the
compliance and hence determine the strains in the x-x, y-y and x-y directions resulting from the
application of the stress.

Relationship between elastic moduli for isotropic materials (applicable to matrix and fibre properties):

Shear modulus:
E
G
21   

Bulk modulus:
E
K
31  2 

2. A high-modulus graphite / epoxy laminate is made from uniaxial fibre reinforced laminae with
following properties:

E1 = 170 GPa E2 = 12 GPa G12 = 4.5 GPa 12 = 0.30 Vf = 0.6 Thickness = 0.13 mm

using three orientations; 0, 45 and 90.

 
Use the above data to write the Q matrices for each orientation and hence find the [A] matrices for
each of the symmetric laminates: [0/90]s, [45]s and [0/90/+45/-45/90/0]s. Hence determine
the elastic properties Ex, Ey, xy and Gxy for each of the laminates.
 1.4 Tutorial solutions

Tutorial 1
Outline Tutorial Solutions on Unidirectional FRPs
(Covering Topics 1.1 and 1.2)

1. The tensile modulus of a thermoset polyester matrix is 3 GPa and that of E-glass reinforcing fibres is
76 GPa. Calculate the tensile modulus for the composite series (Vf = 0 to 1) where the fibres are
continuous and a stress is applied in each of the directions parallel (E 1) and transverse (E2) to the
fibres.

Using the equations developed for E1 and E2 and substituting values:

 
E1  V f  76  1  V f  3  3  73V f and E2 
 
76  3

228
1  V f  76  V f  3 76  73V f

80
70
60
Modulus

50 E1
40
E2
30
20
10
0
0 0.2 0.4 0.6 0.8 1
Volume fraction fibres

2. A composite is made from 50% (by volume) of carbon fibres (uniaxial tensile modulus 230 GPa and
uniaxial tensile strength 3200 MPa) continuously uniaxially aligned in an epoxy matrix (uniaxial
tensile modulus 2.4 GPa and uniaxial tensile strength 60 MPa).

(i) Calculate the tensile strength and tensile modulus of the composite.

For the tensile modulus

 
E1  V f  E f  1  V f  Em  0.5  230  0.5  2.4  116GPa

For the strength, we need to find out if the composite conforms to the “brittle fibre” or “brittle matrix”
model, so first find the failure strains:
*f   *f E f  1.39  102
and
*m   *m Em  2.5  102

So model is a brittle fibre model. At 50% fibres, this is likely to be a case of high fibre volume fraction,
but can check by determining V'f, the value of Vf at which *1 for the low Vf model is the same as *1
for the high Vf model:
 1  V f  m*  1  V f  m  V f  *f ,
* 
m m
so V f 
 *f   m *
 m

The value of 'm is found from:

  Em*f  2.4  13.9  33.4MPa
m

so V'f = 8.2 × 10-3, and so the composite strength is given by:

 
1*  V f  *f  1  V f  m 
i.e:

1*  0.5  3200  0.5  33.4  1620MPa

(ii) Find the minimum volume fraction of fibres required to produce strengthening of the composite and
calculate the tensile modulus at that fibre loading.

Here we need to find V"f, which is when *1 for the high Vf model equals the strength of the
unreinforced matrix, *m:

*
m  
  V f  *f
 1  V f  m
i.e.

 *  m
 60  33.4
V f  m   8.4 10 3
* 
 f  m 3200  33.4

and the corresponding modulus is:


E1  8.4  103  230  1  8.4  103  2.4  4.31GPa 
which shows that, even for this small fibre addition, the modulus is almost doubled.

3. Calculate the tensile strength for the E-glass / polyester composite series in Q1 where the stress is
applied parallel to the fibre direction. The tensile strengths of fibres and matrix are 1800 MPa and 55
MPa, respectively

As before, first determine the failure strains:

*f  1800 76  103  2.37  102

and so this is a “brittle matrix” composite, where the strength is given by:

At low volume fraction of fibres: 1*  V f  f  1  V f  *m  

and at high volume fraction of fibres: 1*  V f  *f
To find the stress in the fibres when the matrix fails:  f  E f *m  76  1.83  102  1390MPa

so the composite strength is given by the larger of:

 
1*  1390V f  55 1  V f  55  1335V f

and
1*  1800V f

2000
1800
1600
Strength

1400 Low Vf
1200
1000
800 High Vf
600
400
200
0
0 0.5 1

Vf

4. A nylon 6.6 bar (uniaxial tensile modulus 2.7 GPa, shear modulus 1.015 GPa and uniaxial tensile
strength 70 MPa) is reinforced by 20% by volume of carbon fibres (uniaxial tensile modulus 230 GPa
and uniaxial tensile strength 3200 MPa). The bar is produced by injection moulding in such a way that
the fibres (of length 400 m and diameter 6 m) are aligned along the axis of the bar.

(i) Find the mean tensile stress carried by the fibres and the overall stress carried by the bar when a
tensile strain of 10-4 is applied to it.

The mean fibre stress is given by  f  E f 11 , so we need to know the aspect ratio:
a  l d  400 / 6  66.7 and the inter-fibre spacing, R. Assuming fibres to be arranged in a square
array with a fibre at each corner, the relationship between volume (or area) fraction and R is given by:
d 2 
Vf  2R 2  0.2 and so:
2R
  1.982 and so the parameter n can now be found:
4 d 4  0.2

 2Gm   2  1.015 
n       0.1136
 E f  ln2 R d    230  ln1.982 

Now, the fibre length correction factor can be found:

tanhna  tanh0.1136  66.7 

1  1   1  0.868
na 0.1136  66.7

and the mean fibre stress is  f  230  103  104  0.868  20MPa

The total stress in the composite is obtained from:

  
1  V f  f  1  V f  m  0.2  20  0.8  2.7  103  104  4.22MPa

(ii) Calculate the tensile strength of the bar and the mean fibre stress at failure. Find also the tensile
strengths for composites containing 30% by volume of fibres of the same size and 30% by volume of
fibres of length 600 m. The shear strength of the fibre-matrix interface can be taken to be 32 MPa.

 *f 3200
lc  d  6  300m
2 *i 2  32
The critical length for this type of fibre is given by:

so that it will be possible for some fibres to reach their breaking stress at the failure plane. Since the
matrix is a thermoplastic, the model is a brittle fibre one, which is confirmed by calculating the matrix
and fibre failure strains:  *m   *m Em  2.59 102 and  *f  1.39  102

When the fibres fail, the composite strain (1) is equal to the fibre failure strain (above) and the matrix
stress at this point is:
  Em*f  2700  0.0139  37.5MPa
m

whereas the mean fibre stress (allowing for interfacial failures) is:

 lc   300 
 f   *f 1    32001    2000MPa
 2l   800 

and so the stress in the composite when the strain is sufficient to cause fibre breakage is given by:
 
  0.2  2000  0.8  37.5  430MPa .
1  V f  f  1  V f  m

The composite may not fail at this point if the broken fibres slipping within the matrix are able to
withstand a higher stress (low volume fraction case) and the maximum stress carried by the composite
would then be given by:

1 
lc
2l
 
* 
V f  *f  1  V f  m
300
800
0.2  3200  0.8  70  296MPa .

So, the composite strength in this case is 430 MPa.

 300  3200 
More generally, the composite strength is the higher of 1*    70V f  70 (low volume
 2l 
 300  3200 
fraction fibres) and 1*  3162.5  V f  37.5 (high volume fraction fibres):
 2l 
 300
250 l>>lc, low Vf
l>>lc, high Vf
Strength

200
l=400, low Vf
150
l=400, high Vf
100 l=600, low Vf
50 l=600, high Vf
0
0 0.05
Vf

3500
3000 l>>lc, low Vf
2500 l>>lc, high Vf
Strength

2000 l=400, low Vf

1500 l=400, high Vf
1000 l=600, low Vf
500 l=600, high Vf
0
0 0.5 1
Vf
 Tutorial 2
Exercises in laminate analysis Q1 solution

Q1.

xx

30°

From the data: Em = 2.4GPa, Ef = 76GPa m =0.34 and f =0.22

So the remaining matrix and fibre properties are

Shear moduli:

2.4 76
Gm   0.895GPa Gf   31.15GPa
21  0.34 21  0.22

Bulk moduli:

2.4 76
Km   2.5GPa Kf   45.24GPa
31  2  0.34 31  2  0.22

Moduli referred to lamina axes:

E1  EmVm  E f V f  2.4  0.4  76  0.6  46.56

 f   m  1 Km 
K

VmV f
1
 0.12  0.377  0.24
12   mVm   f V f   f 
 0.268   0.260
Vm V
 f  1 8.84  103  0.24  1.117
Kf Km Gm

Vf 0.6
G12  Gm   0.895   3.24GPa
3.31  10 2  0.223
G f  Gm 
Vm
1 
2Gm

K m K f  Gm Vm  K f K m  Gm V f
k 
K f  Gm Vm  K m  Gm V f
2.5  45.24  0.895  0.4  45.24  2.5  0.895  0.6 46.135  27.14  3.395
 
45.24  0.895  0.4  2.5  0.895  0.6 46.135  0.4  3.395  0.6
138.275
  6.75GPa
20.49

1 1 G f
m   0.6097  f   0.4717    34.80
3  4 m 3  4 f Gm
  m   f  m
   0.9076    1.048
1   f  1

and so:

 
 
 1   m V f   1.6097  0.6 
  0.895  1    3.644
G23  Gm 1 
  3 2V 2   

 1.048  1  .1784 
 0.6 

  1  m m
 
V f .8039
 
 V f  1 
3
 

and

4 1 1 4 2 1 1 4  0.262
    12     0.428
E2 G23 k E1 3.644 6.77 46.56

 E2  9.35GPa

Thus, [Q] can be found:

E1
Q11 
Z
E
Q22  2
Q11 Q12 0  Z
Q  Q12 Q22 0  12 E2
Q12 
 0 0 Q66  Z
E1  12
2
E2
Z
E1
Q66  G12

46.56  0.262  9.35

Z  0.986
46.56
46.56  47.2 2.465 0 
Q11   47.2
0.986 Q  2.465 9.48 0 


9.35  0 0 3.24
Q22   9.48
0.986
0.26  9.35
Q12   2.465
0.986
Q66  3.24

For  = -30, m = 0.866 and n = -0.5, so

 0.75 0.25 0.866   0.75 0.25 0.433 

   0.25 0.75  0.866
    0.25 0.75  0.433

 0.433 0.433 0.5   0.866 0.866 0.5 
Using a Matlab routine:

» qbar=th*q*inv(psi)

qbar =

30.4972 9.7378 -12.3659

9.7382 11.6368 -3.9675
-12.3656 -3.9672 10.5130

 
Alternatively, using the formula for the elements of Q :

Q11  47.2  0.562  9.48  6.25  102  2  0.75  0.252.465  2  3.24 

 30.47
Q12  0.187547.2  9.48  4  3.24   0.562  0.0625  2.465
 9.737
Q16  47.2  0.75  9.48  0.25  8.945  0.5   0.433
 12.365
etc .

 30.5 9.73  12.37

Q    9.73 11.63  3.97 
 12.37  3.97 10.5 

» sbar=inv(qbar)

sbar =

0.0746 -0.0372 0.0738

-0.0372 0.1173 0.0005
0.0738 0.0005 0.1824

» sig=[150e-3;0;0]

sig =

0.1500
0
0

» eps=sbar*sig

eps =

0.0112
-0.0056
 0.0111

and so;

 xx  0.0112  yy  0.0056  xy  0.0056

 Tutorial 2
Exercises in laminate analysis Q2 solution

Q2. First we can write [Q] for the laminae (same for all of them):

E1
Q11 
Z
E
Q22  2
Q11 Q12 0  Z
Q  Q12 Q22 0 

12 E2
Q12 
 0 0 Q66  Z
E1  12
2
E2
Z
E1
Q66  G12

170  0.32  12
Z   0.994
170
170 171.1 3.622 0 
Q11   171.1
0.994 Q  3.622 12.07 0 
12  0 0 4.5
Q22   12.07
0.994
0.3  12
Q12   3.622
0.994
Q66  4.5

For  = 0, m = 1 and n = 0, so

1 0 0 1 0 0
  0 1 0    0 1 0
0 0 1 0 0 1

Using a Matlab routine:

» qbar=th*q*inv(psi)

qbar =

171.1000 3.6220 0
3.6220 12.0700 0
0 0 4.5000

as expected.
For  = 90, m = 0 and n = 1, so

0 1 0 
0 1 0 
  1 0 0 
   1 0 0 
0 0  1
0 0  1

Using a Matlab routine:

» qbar=th*q*inv(th)

qbar =

12.0700 3.6220 0
3.6220 171.1000 0
0 0 4.5000

again as expected.

For  = +45, m = 0.7071 and n = 0.7071, so

0.5 0.5  1 0.5 0.5  0.5

  0.5 0.5 1     0.50.5 0.5 
0.5  0.5 0   1  1 0 

So, qbar =

52.1035 43.1035 39.7575

43.1035 52.1035 39.7575
39.7575 39.7575 43.9815

For  = -45, m = 0.7071 and n = -0.7071, so

 0.5 0.5 1 0.5 0.5 0.5 

   0.5 0.5  1    0.5 0.5  0.5
 0.5 0.5 0    1 1 0 

So, qbar =

52.1035 43.1035 -39.7575

43.1035 52.1035 -39.7575
-39.7575 -39.7575 43.9815

Properties of [0/90]s, laminate:

A   Q i zi  zi 1  , so for four laminae, each of thickness 0.13 mm:

K

i 1

A11 = 2  0.13  171.1 + 2  0.13  12.07 = 47.62 (GPa and mm units)

A12 = 2  0.13  3.622 + 2  0.13  3.622 = 1.883 (GPa and mm units)
A16 = 2  0.13  0 + 2  0.13  0 = 0
A21 = 2  0.13  3.622 + 2  0.13  3.622 = 1.883 (GPa and mm units)
A22 = 2  0.13  12.07 + 2  0.13  171.1 = 47.62 (GPa and mm units)
A26 = 2  0.13  0 + 2  0.13  0 = 0
A66 = 2  0.13  4.5 + 2  0.13  4.5 = 2.34 (GPa and mm units)

47.62 1.883 0 

 A  1.883 47.62 0 
 0 0 2.34

Inverting:

» a=inv(A)

a=

0.0210 -0.0008 0
-0.0008 0.0210 0
0 0 0.4274

and so:

1 1 1 1
Ex    91.57GPa Ey    91.57GPa
2ha11 .52  0.021 2ha22 .52  0021
1 1 a12 .0008
Gxy    4.5GPa  xy     0.038
2ha66 0.52  0.4274 a11 .021

Properties of [45]s, laminate:

A   Q i zi  zi 1  , so for four laminae, each of thickness 0.13 mm:

K

i 1

A11 = 4  0.13  52.1035 = 27.094 (GPa and mm units)

A12 = 4  0.13  43.1035 = 22.414 (GPa and mm units)
A16 = 2  0.13  39.7575 + 2  0.13  (-39.7575) = 0
A21 = 4  0.13  43.1035 = 22.414 (GPa and mm units)
A22 = 4  0.13  52.1035 = 27.094 (GPa and mm units)
A26 = 2  0.13  39.7575 + 2  0.13  (-39.7575) = 0
A66 = 4  0.13  43.9815 = 22.870 (GPa and mm units)

27.09 22.41 0 

 A   22.41 27.09 0 
 0 0 22.87

Inverting:

» a=inv(A)

a=

0.1169 -0.0967 0
-0.0967 0.1169 0
0 0 0.0437
and so:

1 1 1
Ex    16.44GPa Ey   16.44GPa
2ha11 .52  0.117 2ha22
1 1 a12 .0967
Gxy    44.0GPa  xy     0.826
2ha66 0.52  0.0437 a11 .117

Properties of [0/90/+45/-45/90/0]s, laminate:

A   Q i zi  zi 1  , so for 12 laminae, each of thickness 0.13 mm, we can make use of the previous
K

i 1
calculations, as we have effectively got two[0/90]s laminates and one [45]s, laminate

A11 = 2  A11[0/90]s + A11[45]s = 2  47.62 + 27.09 = 122.33 (GPa and mm units)

A12 = 2  1.883 + 22.414 = 26.180 (GPa and mm units)
A16 = 2  0 + 0 = 0
A21 = 2  1.883 + 22.414 = 26.180 (GPa and mm units) (GPa and mm units)
A22 = 2  47.62 + 27.09 = 122.33 (GPa and mm units)
A26 = 0
A66 = 2  2.34 + 22.87 = 27.55 (GPa and mm units)

122.33 26.180 0 
 A   26.180 122.33 0 
 0 0 27.55

Inverting:

>> a=inv(A)

a=

0.0086 -0.0018 0
-0.0018 0.0086 0
0 0 0.0363

and so:

1 1 1 1
Ex    74.54GPa E y    74.54GPa
2ha11 1.56  0.0086 2ha22 1.56  .0086
1 1 a12 .0018
Gxy    17.65GPa  xy     0.2093
2ha66 1.56  0.0363 a11 .0086