2003, P.

Joyce

Macromechanical Analysis of a
Lamina

Generalized Hookes Law

ij = Cijkl ij
Cijkl is a 9 x 9 matrix!

2003, P. Joyce

Hookes Law
Assume
linear elastic behavior
small deformations

= E
Uniaxial loading

2003, P. Joyce

Triaxial Stress State

Similarly
the deformations
Uniaxial:
in the y- xzyand
yx = for
dx
x
xz
y = xx + xy + Shears
are
z-directions

yz x

zy

omitted
x

xy yy x
y
=dx= dx
dx

dy
zdy
xx
xz
xy
= x dx = 1dx xy
dx= x
dxzyydx
E
E
z y E= [ y E(Consider
E
shear
x + z )]
yx
E
y zx
separately
y zy
xz yx
x
y
dyxz dy
= = dy
dy
yy =
yz
=

E
E
E
E
E Superimpose
1 x

y
]
)
z = [z ( xthree
+

y uniaxial
z x dz
=

dz
=
zx
=
= = dz
dz
zy
y dz
Ezx
1y
x
zz =
z dz

E
= [ x zz( y + z )] stresses E
E
xx
xzE
2003, P. Joyce

Triaxial Stress State

Strain-Stress Relations

1
x = [ x ( y + z )]
E
1
y = [ y ( x + z )]
E
1
z = [z ( x + y )]
E

2003, P. Joyce

Triaxial Stress State

Stress-Strain Relations

E
[
x =
(1 ) x + ( y + z )]
(1 + )(1 2)
E
[
y =
(1 ) y + ( x + z )]
(1 + )(1 2)
E
[
z =
(1 )z + ( x + y )]
(1 + )(1 2)

2003, P. Joyce

Similarly for shear stresses

2003, P. Joyce

Shear Stress Strain Relationships

Shear stresses are independent of each other
and all axial stresses
Shear strains are independent of each other
and all axial strains
Each obeys a simple linear elastic model

= G
G Shear Modulus

E
G=
2(1 + )

2003, P. Joyce

Shear Stress-Strain Relationships

xy = G xy

E
=
xy
2(1 + )

xz = G xz

E
=
xz
2(1 + )

yz = G yz

E
=
yz
2(1 + )

2003, P. Joyce

Generalized Hookes Law

Compliance matrix
1
E

x
E
y
z
= E
yz 0
zx

xy 0

E
1
E

E
1
E

1
G

1
G

0 x
y

0
z

0 yz
zx

0 xy

1

Strain-stress relations for an isotropic material/matrix form

2003, P. Joyce

Generalized Hookes Law

Stiffness matrix

E (1 )

x (1 2 )(1 + v)

E
y (1 2 )(1 + v)
z
E
=
yz (1 2 )(1 + v)
zx
0

0
xy

E
(1 2 )(1 + v)
E (1 )
(1 2 )(1 + v)
E
(1 2 )(1 + v)

E
(1 2 )(1 + v)
E
(1 2 )(1 + v)
E (1 )
(1 2 )(1 + v)

0
0

0
0

0
x

0 0 0 y

z
0 0 0
yz

G 0 0 zx

0 G 0 xy
0 0 G
0

Stress-strain relations for an isotropic material /matrix form

2003, P. Joyce

Stress-Strain Relations in
Composite Materials
Material

No. of independent
elastic constants

General anisotropic material

81

Anisotropic material (considering

symmetry of stress and strain tensors)

36

Anisotropic material with energy

considerations

21

General orthotropic material

Transversely orthotropic material

Isotropic material

2
2003, P. Joyce

Stress-Strain Relations in
Composite Materials
Orthotropic material ~ has three mutually
perpendicular plans of material symmetry
Specially orthotropic material ~ when the
reference system of coordinates is selected
along principal planes of material symmetry
Transversely isotropic material ~ one of its
principal planes is a plane of isotropy
(properties are the same in all directions.)
2003, P. Joyce

Contracted Notation
Thanks to symmetry of the stress and strain
tensors, the compliance matrix reduces to a
6 x 6 matrix, introduce a contracted notation.
11 = 1, 22 = 2 , 33 = 3 ,
23 = 4 , 31 = 5 ,12 = 6
11 = 1, 22 = 2 , 33 = 1,
2 23 = 4 ,231 = 5 ,212 = 6
C1111 = C11, C1122 = C12, C1133 = C13, C1123 = 2C14, C1131 = 2C15, C1112 = 2C16
C2211 = C21, C2222 = C22, C2233 = C23, C2223 = 2C24, C2231 = 2C25, C2212 = 2C26
2003, P. Joyce

Stress-Strain Relations
for Thin UD Lamina
Assumed to be under a state of plane stress
1 Q11 Q12 0 1
= Q Q

0
22
2 12
2
6 0
0 Q66 6
Fully characterized by 4 independent constants,
Qij ~ reduced stiffnesses

2003, P. Joyce

Reduced Stiffness Matrix?

If the stiffness matrix is the inverse of the
compliance matrix, what is the reduced
stiffness matrix?
Qij = Cij

C i 3C j 3
C33

(i, j = 1, 2, 6)

2003, P. Joyce

Relations Between Mathematical

and Engineering Constants
Q11 =
Q22 =

E1

1 12 21
E2

1 12 21

21 E1
12 E2
Q12 =
=
1 12 21 1 12 21
Q66 = G12

2003, P. Joyce

You Said Four Independent

Constants?
From symmetry of the compliance matrix
ij
Ei

ji
Ej

The above can also be deduced from Bettis

reciprocal law according to which transverse
deformation due to a stress applied in the
longitudinal direction is equal to the longitudinal
deformation due to an equal stress applied in the
transverse direction.
2003, P. Joyce

Stress-Strain Relations
Also expressed in terms of compliances
1 S11
= S
2 12
6 0

S12
S 22
0

0 1
0 2
S 66 6

2003, P. Joyce

Relations between Mathematical

and Engineering Constants
S11 =

1
E1

S 22 =

1
E2

S12 =
S 66 =

12
E1

21
E2

1
G12

Thats Better!
2003, P. Joyce

Sample Calculation
For a graphite/epoxy UD
laminate, find the following:

1.) Compliance matrix

2.) Minor Poissons ratio
3.) Reduced Stiffness Matrix
4.) Strains in the 1-2 coordinate
system if the applied stresses are

12
1

1 = 2 MPa, 2 = -3MPa, 12 = 4 MPa

2003, P. Joyce

Sample Data
Property

Symbol Units

Glass/ Boron Graphite

epoxy /epoxy /epoxy

Fiber
volume
fraction

Vf

0.45

0.50

0.70

Long.
elastic
modulus

E1

GPa

38.6

204

181

Trans.
elastic
modulus

E2

GPa

8.27

18.50

10.30

Major
Poissons
ratio

12

0.26

0.23

0.28

Shear
Modulus

G12

GPa

4.14

5.59

7.17

2003, P. Joyce

Sample Calculation
The compliance matrix elements are
calculated as follows:
S11 =
S12 =

1
1
11
=
=
0
.
5525
(
10
)
9
E1 181(10 )

12
E1

0.28
= 0.1547(10 11 )
9
181(10 )

S 22 =

1
1
10
=
=
0
.
9709
(
10
)
9
E1 10.3(10 )

S 66 =

1
1
9
=
=
0
.
1395
(
10
)
9
G12 7.17(10 )

(all terms have units of Pa-1)

2003, P. Joyce

Sample Calculation
From Bettis reciprocal law:
21
E2

21 =

12
E1
(0.28)
9
(
10
.
3
)(
10
) = 0.01593
181(109 )

2003, P. Joyce

Sample Calculation
The stiffness matrix elements are calculated
as follows:
Q11 =

E1

1 12 21

181(109 )
=
= 181.8(109 )
1 (0.28)(0.01593)

12 E2
21E1
(0.28)(10.3)(109 )
Q12 =
=
=
= 2.897(109 )
1 12 21 1 12 21 1 (0.28)(0.01593)
Q22 =

E2

1 12 21

10.3(109 )
=
= 10.35(109 )
1 (0.28)(0.01593)

Q66 = G12 = 7.17(109 )

(all terms have units of Pa)

2003, P. Joyce

Sample Calculation
The stiffness matrix can also be calculated
by inverting the compliance matrix of Part 1:
0.5525 (10 11 )
[Q ] = 0.1547 (10 11 )

0.1547 (10 11 )
0.9709 (10 10 )
0

9
0.1395 (10 )

181 .8(10 9 ) 2.897 (10 9 )

0

[Q ] = 2.897 (10 9 ) 10 .35(10 9 )

0
Pa
9

0
0
7
.
17
(
10
)

(all terms have units of Pa)

2003, P. Joyce

Sample Calculation
The strains in the 1-2 coordinate system are
calculated as follows:
1 S11
= S
2 12
12 0

S12 )
S 22
0

0 1
0 2
S 66 12

11
1 0.5525 (10 )
= 0.1547 (10 11 )
2
12
0
1 15 .69
= 294 .4 (10 -6 )
2

12 557 .9

0.1547 (10 11 )
0.9709 (10 10 )
0

2(10 6 )

3
(
10
)
0

0.1395 (10 9 ) 4(10 6 )

(10-6 is microstrain)
2003, P. Joyce

Stress-Strain Relations
for Thin Angle Lamina
Generally, a laminate does not consist only of UD laminae because
of their stiffness and strength properties in the transverse direction.
Hence, in most laminates, some laminae are placed at an angle.
2

Global and material axes of an angle lamina.

2003, P. Joyce

Stress-Strain Relations
for Thin Angle Lamina
The axes in the x-y coordinate system are called the global
axes of the off-axes.
The axes in the 1-2 coordinate system are called the
material axes or the local axes, where direction 1 is parallel
to the fibers (also called the longitudinal direction) and
direction 2 is is perpendicular to the fibers (also called the
transverse direction.)
The angle between the two axes is denoted by the angle .

2003, P. Joyce

Stress-Strain Relations
for Thin Angle Lamina
The stress-strain relationship in the 1-2 coordinate system has already
been established.
From Mechanics of Materials, the stresses in the global and material
axes are related to each other through the angle of the lamina, .
x
1

1

[
]
=
T
(
)

y
2
xy
12

Where [T()] is called the transformation matrix and is defined as

c2
[T ( ) ] = s 2
sc

s2
c2
sc

c 2
2 sc

1
2 sc thus [T ( ) ] = s 2
sc
c 2 s 2

2003, P. Joyce

s2
c2
sc

2 sc

2 sc = [T(- ) ]
c 2 s 2

Stress-Strain Relations
for Thin Angle Lamina
Using the stress-strain equation in the
material axes together with the
transformation equation we obtain:
x
1

1

[
]
[
]

T
Q
=
y
2
xy
12

2003, P. Joyce

Stress-Strain Relations
for Thin Angle Lamina
Similarly, the strains in the global and material coordinate
axes are related through the transformation matrix
x
1
= [T ]
y
2
12 xy
12 12

which can be rewritten as

x
1
= [R ][T ][R ]1
y
2
12 xy
12 12

where [R] is the Reuter matrix and is defined as

1 0 0
[R ] = 0 1 0
0 0 2
2003, P. Joyce

Stress-Strain Relations
for Thin Angle Lamina
Multiplying out the first five matrices on the RHS of the previous
equation we obtain the transformed reduced stiffness matrix, [Qxy ]
Thus, [ ]x , y = [Q ]x , y [ ]x , y
Summarizing,

Qxx = c 4Q11 + s 4Q22 + 2c 2 s 2Q12 + 4c 2 s 2Q66

Q yy = s 4Q11 + c 4Q22 + 2c 2 s 2Q12 + 4c 2 s 2Q66
Qxy = c 2 s 2Q11 + c 2 s 2Q22 + (c 4 + s 4 )Q12 4c 2 s 2Q66
Qxs = c 3 sQ11 cs 3Q22 + (cs 3 c 3 s )Q12 + 2(cs 3 c 3 s )Q66
Q ys = cs 3Q11 c 3 sQ22 + (c 3 s cs 3 )Q12 + 2(c 3 s cs 3 )Q66
Qss = c 2 s 2Q11 + c 2 s 2Q22 2c 2 s 2Q12 + (c 2 s 2 ) 2 Q66
The subscript s corresponds to shear stress or strain components in the
x-y system, i.e., s = xy and s = xy
2003, P. Joyce

Strain-Stress Relations for Thin

Angle Lamina
Similarly, the transformed compliance can be obtained:
S xx = c 4 S11 + s 4 S 22 + 2c 2 s 2 S12 + 4c 2 s 2 S 66
S yy = s 4 S11 + c 4 S 22 + 2c 2 s 2 S12 + 4c 2 s 2 S 66
S xy = c 2 s 2 S11 + c 2 s 2 S 22 + (c 4 + s 4 ) S12 4c 2 s 2 S 66
S xs = 2c 3 sS11 2cs 3 S 22 + 2(cs 3 c 3 s ) S12 + (cs 3 c 3 s ) S 66
S ys = 2cs 3 S11 2c 3 sS 22 + 2(c 3 s cs 3 ) S12 + (c 3 s cs 3 ) S 66
S ss = 4c 2 s 2 S11 + 4c 2 s 2 S 22 8c 2 s 2 S12 + (c 2 s 2 ) 2 S 66

Thus, [ ]x , y = [S ]x , y [ ]x , y
2003, P. Joyce

Strain-Stress Relations for Thin

Angle Lamina
How about in terms of Engineering Constants?
If we imagine a series of simple experiments on an element
with sides parallel to the x- and y-axes, we obtain:
1

E
x x
= xy
y E
x
s
xs

Ex

yx

Ey
1
Ey

ys
Ey

2003, P. Joyce

sx

Gxy
x
sy
y
Gxy
s

Gxy

Strain-Stress Relations for Thin

Angle Lamina
What is ? shear coupling coefficient
xs , the first subscript denotes normal
loading in the x-direction; the second
subscript denotes shear strain.
xs = s x
ys = s y
sx = x s
sy = y s
2003, P. Joyce

Strain-Stress Relations for Thin

Angle Lamina
Comparison of equivalent strain-stress relations yields the
following relationships:
S xx =

1
Ex

Ex =

1
S xx

S yy =

1
Ey

Ey =

1
S yy

S ss =

1
G xy

G xy =

1
S ss

S xy = S yx =
S xs = S sx =
S ys = S sy =

xy
Ex

xs
Ex

ys
Ey

=
=

sx
G xy

sy
G xy

yx
Ey

xy =

S yx
S xx

xs =

S sx
;
S xx

ys =

S sy

2003, P. Joyce

S yy

; yx =

S xy
S yy

sx =

S xs
S ss

; yx =

S ys
S ss

Sample Calculation
Find the following for a 60 angle lamina of
graphite/epoxy.
Transformed compliance matrix
Transformed reduced stiffness matrix
Global strains
Local strains
If the applied stresses are x = 2 MPa, y = -3MPa, xy = 4 MPa

2003, P. Joyce

Sample Calculation
From the previous example:

S11 = 0.5525(10 11 )
S12 = 0.1547(10 11 )
S 22 = 0.9709(10 10 )
S 66 = 0.1395(10 9 )

(all terms have units of Pa-1)

2003, P. Joyce

Sample Calculation
The transformed compliance matrix elements are calculated as
follows:
S xx = c 4 S11 + s 4 S 22 + 2c 2 s 2 S12 + 4c 2 s 2 S 66 = 0.8053(10 10 )
S yy = s 4 S11 + c 4 S 22 + 2c 2 s 2 S12 + 4c 2 s 2 S 66 = 0.7878(10 11 )
S xy = c 2 s 2 S11 + c 2 s 2 S 22 + (c 4 + s 4 ) S12 4c 2 s 2 S 66 = 0.3234(10 10 )
S xs = 2c 3 sS11 2cs 3 S 22 + 2(cs 3 c 3 s ) S12 + (cs 3 c 3 s ) S 66 = 0.3475(10 10 )
S ys = 2cs 3 S11 2c 3 sS 22 + 2(c 3 s cs 3 ) S12 + (c 3 s cs 3 ) S 66 = 0.4696(10 10 )
S ss = 4c 2 s 2 S11 + 4c 2 s 2 S 22 8c 2 s 2 S12 + (c 2 s 2 ) 2 S 66 = 0.1141(10 9 )

(All terms are in Pa-1)

2003, P. Joyce

Sample Calculation
Next, invert the transformed compliance matrix [S] to
obtain the transformed reduced stiffness matrix [Q].
0.8053(10 10 ) 0.7878(10 10 ) 0.3234(10 10 )
[Q] = [S ]1 = 0.7878(1010 ) 0.3475(1010 ) 0.4696(1010 )
0.3234(10 10 ) 0.4696(10 10 ) 0.1141(10 9 )

0.2365 0.3246 0.2005

[Q] = 0.3246 1.094 0.5419(1011 )
0.2005 0.5419 0.3674

(all terms in Pa)

2003, P. Joyce

Sample Calculation
The global strains in the x-y plane are given
by [ ]x, y = [S ]x, y [ ]x, y
x 0.8053(10 10 ) 0.7878(10 10 ) 0.3234(10 10 ) 2

(106 )
10
10
10

0
.
7878
(
10
)
0
.
3475
(
10
)
0
.
4696
(
10
)
3

10

10

9
xy 0.3234(10 ) 0.4696(10 ) 0.1141(10 ) 4

x 0.5534(10 4 )

3

0
.
3078
(
10
)
=

y
xy 0.5328(10 3 )

2003, P. Joyce

Sample Calculation
The local strains in the lamina can be calculated using the
Transformation equation.
x
1
= [T ]
y
2
12 xy
12 12

sin 2 60
2 cos 60 sin 60 0.5534(10 4 )
1 cos 2 60

= sin 2 60
2
3

0
.
3078
(
10
)
cos
60
2
cos
60
sin
60

2
12 12 cos 60 sin 60 cos 60 sin 60 cos 2 60 sin 2 60 0.5328(10 3 ) / 2
4
1 0.1367(10 )
= 0.2662(10 3 )

3
12 0.5809(10 )

2003, P. Joyce

Transformation of
Engineering Constants
Flow chart for determination of transformed
elastic constants of UD lamina.
[S]x,y

[S]1,2

[E]1,2

[E]x,y

[Q]x,y

[Q]1,2
2003, P. Joyce

Macromechanical Strength
Parameters
From a macromechanical POV, the strength
of a lamina is an anisotropic property.
It is desirable, for example, to correlate the
strength along an arbitrary direction to some
basic strength parameters (analogous to
micromechanic definitions before.)

2003, P. Joyce

Strength Failure Theories of an

Angle Lamina
Various theories have been developed for studying the failure of an
angle lamina.
Generally based on the normal and shear strengths of a UD lamina.
Need to consider tension and compression
UD lamina has 2 material axes, 1-direction parallel to the fibers and
2-direction which is perpendicular to the fibers.
Hence there are 4 normal strength parameters for UD lamina.

Tensile strength in fiber direction

Transverse tensile strength
Compressive strength in fiber direction
Transverse compressive strength

The fifth strength parameter is the shear strength

2003, P. Joyce

Strength Failure Theories of an

Angle Lamina
Unlike the stiffness parameters, these strength parameters
cannot be transformed directly for an angle lamina.
Hence, the failure theories are based on first finding the
stresses in the material axes and then using these five
strength parameters of a UD lamina to find whether the
lamina has failed.

2003, P. Joyce

Macromechanical Strength
Parameters
Also predict transverse compressive strength and in-plane
shear strength using micromechanics. . .
Failure mechanisms vary greatly with material properties
and type of loading.
Even when predictions are accurate with regard to failure
initiation at critical points, they are only approximate as far
as global failure of the lamina is concerned.
Furthermore, the possible interaction of failure
mechanisms makes it difficult to obtain reliable strength
predictions under a general type of loading.
A macromechanical or phenomological approach to failure
analysis may be preferable.
2003, P. Joyce

Macromechanical Strength
Parameters
This characterization recognizes the fact that most
composite materials have different strengths in tension and
compression.
By convention the sign of the shear stress is immaterial, as
long as the shear strength is referred to the principal
material directions.
Exception, refers to the case when the shear stress is
applied at an angle wrt the principal material directions.
Since most composites have different tensile and
compressive strengths and they are weakest in transverse
tension, it follows that in this case the lamina would be
stronger under positive shear.
2003, P. Joyce

Macromechanical Strength
Parameters

= 6

x = -6
Positive shear stress

Shear stress acting along principal material axes.

2003, P. Joyce

Macromechanical Strength
Parameters

= 6

x = -6
Negative shear stress

Shear stress acting along principal material axes.

2003, P. Joyce

Macromechanical Strength

Parameters

= 6

2 = -s
Positive shear stress

Shear stress acting at 45 wrt principal material axes.

2003, P. Joyce

Macromechanical Strength
Parameters

= -s

2 = s
Negative shear stress

Shear stress acting at 45 wrt principal material axes.

2003, P. Joyce

Macromechanical Failure
Theories
Given a state of stress, the principal stresses and their
directions are obtained by stress transformation
(independent of material properties.)
The principal strains and their directions are obtained by
using the appropriate anisotropic stress-strain relations and
strain transformation.
In general, the principal stress, principal strain, and
material symmetry directions do not coincide.
Since strength varies with orientation, maximum stress
alone is not the critical factor in failure.
2003, P. Joyce

Macromechanical Failure
Theories
An anisotropic failure theory is needed.
Failure criteria for homogeneous isotropic materials, such as

Maximum normal stress (Rankine),

Maximum shear stress (Tresca),
Maximum distortional energy (von Mises),
and so forth are well established.

More than 40 anisotropic theories have been proposed look at the

four most widely used.

2003, P. Joyce

Maximum Stress Failure Theory

Related to the Maximum Normal stress theory by Rankine and the
Maximum Shear stress theory by Tresca.
The stresses acting on a lamina are resolved into the normal and shear
stresses in the material axes.
Failure is predicted in a lamina, if any of the normal or shear stresses
in the material axes are equal to or greater than the corresponding
ultimate strengths of a UD lamina.

( )

1C

ult

( )

< 1 < 1T

ult

( )

, 2C

ult

( )

< 2 < 2T

ult

, (12 )ult < 12 < (12 )ult

Each component of stress is compared with the corresponding strength

and hence does not have an interaction with the others.

2003, P. Joyce

Sample Calculation
Find the off-axis shear strength of a 60 graphite/epoxy
lamina using the Maximum Stress failure criteria.
Assume the following stress state
x = 0, y = 0, xy = ,

Find the stresses along the principal material axes, using

the Transformation Equation.
1 0.2500 0.7500
= 0.7500 0.2500
2
12 0.4330 0.4330
1 0.866
= 0.866
2

12 0.500
2003, P. Joyce

0.8660 0
0.8660 0
0.5000

Sample Calculation
Applying the Maximum Stress Failure Criteria together with strength
data for graphite/epoxy composites from the Data Sheet,we have

1500 < 0.866 < 1500

246 < 0.866 < 40
68 < 0.500 < 68

or
1732 < < 1732
46.19 < < 284.1
136.0 < < 136.0
2003, P. Joyce

Sample Calculation
The off-axis shear strength of a lamina is defined as the minimum of the
positive and negative shear stress which can be applied to an angle lamina
before failure.
Calculations show that xy = 46.19 MPa is the largest magnitude of shear
stress one can apply to the 60 graphite/epoxy composite.
However, the largest positive shear stress one could apply is 136.0 MPa,
and the largest negative shear stress one could apply is 46.19 MPa.
This shows that the maximum magnitude of allowable shear stress in other
than the material axes direction depends on the sign of the shear stress.
This is because the tensile strength perpendicular to the fiber direction is
much lower than the compressive strength perpendicular to the fiber
direction.

2003, P. Joyce

Failure Envelopes
A failure envelope is a 3D plot of the combinations of normal
and shear stresses which can be applied to an angle lamina
before failure.
Drawing 3D graphs is time consuming. . .
One may develop failure envelopes for constant shear stress,
xy, and then use the 2 normal stresses x and y as the 2 axes.
If the applied stress is within the failure envelope, the lamina is
safe; otherwise it has failed.

2003, P. Joyce

Failure Envelopes
For a UD lamina at a given
shear stress loading, the failure
envelope takes the form of a
rectangle as shown.

For a 60 lamina at a given

shear stress loading, the failure
envelope takes the form of a
rectangle as shown.

2
2

( )

( )

2T

2T

ult

( )

1C

( )

1T

ult

ult

( )

1C

( )

1T

ult

ult

( )

2C

xy

( )

ult

2C

2003, P. Joyce

ult

ult

Maximum Strain Failure Theory

Based on the Maximum Normal Strain Theory by St. Venant and the Maximum
Shear Stress Theory by Tresca.
The strains applied to a lamina are resolved into the normal and shear stresses in
the material axes.
Failure is predicted in a lamina, if any of the normal or shear strains in the
material axes are equal to or greater than the corresponding ultimate strains of a
UD lamina.

( )

1C

( )

T
<

<

1
1
ult

( )

C
,

2
ult

( )

T
<

<

2
2
ult

ult

, ( 12 )ult < 12 < ( 12 )ult

The ultimate strains can be found directly from the ultimate strength parameters
and the elastic moduli, assuming the stress-strain response is linear until failure.
Each component of strain is compared with the corresponding ultimate strain
and hence does not have an interaction with the others.
Yields different results from Maximum Stress Failure Theory, because the local
strains in a lamina include the Poissons ratio effect (allows some interaction of
stress components.)
2003, P. Joyce

Maximum Strain Failure Theory

Assume a general biaxial state of stress
on an angle lamina.

2
1

Obtain the stress components along the principal material axes by

stress transformation.
x
1

1

[
]
=
T
(
)

y
2
xy
12

2003, P. Joyce

Maximum Strain Failure Theory

Then the corresponding strain components can be
calculated by means of the lamina stress-strain relations:
1 =
2 =
6 =

21

E1

12

E2

2
E2

1
E1

6
G12

Next calculate the ultimate strains as follows

( ) = (E ) , ( )
T
1 ult

T
1 ult
1

C
1 ult

( ) , ( )
=
E
C
1 ult
1

T
2 ult

( ) , ( )
=
E
T
2 ult
2

2003, P. Joyce

C
2 ult

( ) , ( )
=
E
C
2 ult

12 ult

( 12 )ult
G12

Maximum Strain Failure Theory

Failure subcriteria restated in terms of the
stresses:
( ) when > 0
=
( ) when < 0
( ) when > 0
=
( ) when < 0
1

12

21

T
1 ult
C
1 ult

T
2 ult
C
2 ult

( 12 )ult < 12 < ( 12 )ult

2003, P. Joyce

Maximum Strain Failure Theory

For a 2D state of stress with 6 = 0, the failure envelope takes
the form of a parallelogram with its center off the origin.
y
x

( )

2T

2 21 1 = ( 1T )

ult

( )

1C

( )

1T

ult

( )

1 12 2 = (

C
1

C
2 ult

ult

2 21 1 = ( 1C )

2003, P. Joyce

1 12 2 = ( 1T )

Tsai-Hill Failure Theory

Based on the deviatoric or distortional energy failure
theory of von Mises.
Adapted to anisotropic materials by Hill.
Then adapted to a UD lamina by Tsai.
2
2
2
1 1 2 2 12
<1
T T 2 + T +
(
)

1 ult 1 ult 2 ult 12 ult

( )

( )

( )

Given the global stresses in a lamina, one can find the

local stresses in a lamina and apply the above failure
theory to determine whether or not the lamina has
failed.
2003, P. Joyce

Tsai-Hill Failure Theory

The failure envelope described by the Tsai-Hill criterion is a closed
surface in the 1, 2, 12 space.
Failure envelopes for constant values of
12 22 1 2
have the form
+ 2 2 = 1 k 2
2
F1

Where:

(
(
(
(

F2

)
)
)
)

k = 12 ( 12 )ult

F1

1T ult when 1 > 0

F1 = C
1 ult when 1 < 0
2T ult when 2 > 0
F2 = C
2 ult when 2 < 0

(Modified Tsai-Hill Criterion)

Graphically represents 4 different elliptical arcs joined at the 1, 2 axes.

2003, P. Joyce

Tsai-Hill Failure Theory

Considers the interaction between the 3 UD lamina
strength parameters, unlike the Maximum Stress and
Maximum Strain Theories.
Tsai-Hill Failure Theory is a Unified Theory and hence
does not give the mode of failure like the Maximum Stress
and Maximum Strain Theories.

2003, P. Joyce

Tsai-Wu Failure Theory

Based on a general failure theory for anisotropic materials
first proposed by Goldenblat and Kopnov (1965).
Capable of predicting strength under general states of
stress for which no experimental data are available.
Uses the concept of strength tensors.
Has the form of an invariant formed from stress and strain
tensor components
Has the capability to account for the difference between
tensile and compressive strengths

2003, P. Joyce

Tsai-Wu Failure Theory

Tsai and Wu (1971) proposed a modified tensor polynomial
theory by assuming the existence of a failure surface in the
stress space of the form

H1 1 + H 2 2 + H 6 12 + H11 1 + H 22 2 + H 66 12 + 2 H12 1 2 < 1

2

The coefficients are obtained by applying elementary loading

conditions to the lamina. Thus
H1 =
H11 =

( )

T
1 ult

( )

C
1 ult

( ) ( )
T
1 ult

C
1 ult

H2 =
H 22 =

( )

T
2 ult

( )

C
2 ult

( ) ( )
T
2 ult

C
2 ult

2003, P. Joyce

H6 = 0
H 66 =

( 12 )ult 2

Tsai-Wu Failure Theory

The remaining coefficient H12 must be obtained by
some type of biaxial testing.
Direct biaxial testing is not easy or practical to
perform.
An easier test producing a biaxial state of stress is
the off-axis tensile test.
For = 45 we can measure the off-axis tensile
strength, .
Then, H12 = 2 (H1 + H 2 ) 1 (H11 + H 22 + H 66 )

Again produces an elliptical failure envelope.

2003, P. Joyce

Comparison with Experimental Results

2003, P. Joyce

Comparison with Experimental Results

2003, P. Joyce

Comparison with Experimental Results

2003, P. Joyce

Comparison with Experimental Results

2003, P. Joyce

Comparison with Experimental Results

Observations
The difference between the Maximum Stress and
Maximum Strain Failure Theories and the experimental
results is quite pronounced.
Tsai-Hill and Tsai-Wu Failure Theories are in good
agreement with experimentally obtained results.
The cusps observed in the Maximum Stress and
Maximum Strain Failure Theories correspond to the
change in failure mode.
The variation of the strength as a function of angle is
smooth in the Tsai-Hill and Tsai-Wu Failure Theories.
2003, P. Joyce

Macromechanical Failure Theories

Theory

Physical Basis

Operational
Convenience

Reqd experimental
characterization

Maximum Stress

Tensile behavior of brittle

material

Inconvenient

Few parameters
By simple testing

Inconvenient

Few parameters by
simple testing

Biaxial testing is
needed in addition to
uniaxial testing

No stress interaction
Maximum Strain

Tensile behavior of brittle

material
Some stress interaction

Deviatoric strain
energy

Ductile behavior of
anisotropic materials

Can be programmed

(Tsai-Hill)

Curve fittingfor
heterogeneous brittle
composites

Different functions
required for tensile and
compressive strengths

Interactive tensor
polynomial

Mathematically
consistent

General and
comprehensive;
operationally simple

Tsai-Wu

Reliable curve fitting

2003, P. Joyce

Numerous parameters

Comprehensive
experimental program
needed.

References

Engineering Mechanics of Composite Materials, Daniel, I.M. and Ishai, O., 1994.
Mechanics of Composite Materials, Kaw, A.K., 1997.
Introduction to Composite Materials, Tsai, S. W. and Hahn, H. T., 1980.
Application of Advanced Composites in Mechanical Engineering Designs, Zweben, C.,
Proceedings of the 31st International SAMPE Technical Conference, 1999.

2003, P. Joyce