Paavai Institutions Department of AERO

UNIT-II

METHODS OF ANALYSIS
 Paavai Institutions Department of AERO

S.NO TERM TECHNICAL MEANING LITERAL MEANING

A composite material is a combination of

two materials with different physical and
chemical properties. When they are
consisting of different
1 COMPOSITE combined they create a material which is
parts or materials.
specialised to do a certain job, for instance to
become stronger, lighter or resistant to
electricity.

Tensile strength is a term used to describe

2 TESNILE how much load a material can withstand capable of tension
when loaded in tension
Compression force (or compressive force)
action of applying a force
occurs when a physical force presses inward
3 COMPRESSION to a solid, liquid, or gas so
on an object, causing it to become
that it takes up less space
compacted
a bend or fold in
the material's ability to resist deformation something such as a part of
4 FLEXURAL
under load. the body, or the act of
bending or folding
to measure the ability of an adhesive tape to
resist creep under a constant load, applied
5 SHEAR TEST -
parallel to the surface of the tape and
substrate.
Micro mechanics is the analysis of
composite or heterogeneous materials on the
6 MICROMECHANICS -
level of the individual constituents that
constitute these materials
Macro mechanics is the study of composite
material behavior wherein the material is
7 MACROMECHANICS -
presumed to be homogeneous, and the
effects of the constituent materials are
detected only as averaged apparent
macroscopic a material or a substance
a component properties of thethat
of a composite composite
reinforces
8 FIBER materials that is made from natural
the material
or artificial threads
The matrix binds the fiber reinforcement,
a set of numbers arranged
transfers loads between fibers, gives the
9 MATRIX in rows and columns so as
composite component its net shape and
to form a rectangular array.
determines its surface quality
Paavai Institutions Department of AERO

CONTENT
METHODS OF ANALYSIS

Micro mechanics - Mechanics of materials approach, elasticity approach to determine

material properties; Macro mechanics – Stress-strain relations; Determination of material

properties; Mechanical properties of Composites - Tensile test, compression test, flexural test,

shear test and inter-laminar shear strength.

 UNIT II
Introduction

In this lecture, we are going to develop the 3D constitutive equations. We will start with
the the generalized Hooke’s law for a material, that is, material is generally anisotropic
in nature. Finally, we will derive the constitutive equation for isotropic material, with
which the readers are very familiar. The journey for constitutive equation from
anisotropic to isotropic material is very interesting and will use most of the concepts
that we have learnt in earlier Lecture 9.

The generalized Hooke’s law for a material is given as

(3.1)

where, σij is a second order tensor known as stress tensor and its individual elements are
the stress components. is another second order tensor known as strain tensor and its
individual elements are the strain components. Cijkl is a fourth order tensor known as
stiffness tensor. In the remaining section we will call it as stiffness matrix, as popularly
known. The individual elements of this tensor are the stiffness coefficients for this
linear stress-strain relationship. Thus, stress and strain tensor has ( )9
components each and the stiffness tensor has (=) 81 independent elements. The
individual elements =) 81 are referred by various names as elastic constants,
moduli and stiffness coefficients. The reduction in the number of these elastic constants
can be sought with the following symmetries.

Stress Symmetry:
The stress components are symmetric under this symmetry condition, that is, .
Thus, there are six independent stress components. Hence, from Equation. (3.1) we
write

(3.2)

Subtracting Equation (3.2) from Equation (3.1) leads to the following equation

(3.3)

There are six independent ways to express i and j taken together and still nine
independent ways to express k and l taken together. Thus, with stress symmetry the
number of independent elastic constants reduce to ( ) 54 from 81.

Strain Symmetry:
Composite Materials and Structures-AE 16082
 The strain components are symmetric under this symmetry condition, that is, .
Hence, from Equation (3.1) we write

Subtracting Equation (3.3) from Equation (3.2) we get the following equation

(3.4)

It can be seen from Equation (3.4) that there are six independent ways of expressing i
and j taken together when k and l are fixed. Similarly, there are six independent ways of
expressing k and l taken together when i and j are fixed in Equation (3.4). Thus, there
are independent constants for this linear elastic material with stress and strain
symmetry.

With this reduced stress and strain components and reduced number of stiffness
coefficients, we can write Hooke’s law in a contracted form as

(3.5)

where

(3.6)

Note: The shear strains are the engineering shear strains.

For Equation (3.5) to be solvable for strains in terms of stresses, the determinant of the

stiffness matrix must be nonzero, that is .

The number of independent elastic constants can be reduced further, if there
exists strain energy density function W, given as below.

Strain Energy Density Function (W):

Composite Materials and Structures-AE 16082
The strain energy density function W is given as

(3.7)

with the property that

(3.8)

It is seen that W is a quadratic function of strain. A material with the existence of W with
property in Equation (3.8) is called as Hyperelastic Material.

The W can also be written as

(3.9)

Subtracting Equation (3.9) from Equation (3.7) we get

(3.10)

which leads to the identity . Thus, the stiffness matrix is symmetric. This symmetric
matrix has 21 independent elastic constants. The stiffness matrix is given as follows:

(3.11)

The existence of the function W is based upon the first and second law of thermodynamics.
Further, it should be noted that this function is positive definite. Also, the function W is an
invariant (An invariant is a quantity which is independent of change of reference).

The material with 21 independent elastic constants is called Anisotropic or Aelotropic

Material.
Further reduction in the number of independent elastic constants can be obtained with the use
of planes of material symmetry as follows.

Material Symmetry:
Composite Materials and Structures-AE 16082
It should be recalled that both the stress and strain tensor follow the transformation rule and
so does the stiffness tensor. The transformation rule for these quantities (as given in Equation
(3.1)) is known as follows

(3.12)

where are the direction cosines from i to j coordinate system. The prime indicates the
quantity in new coordinate system.

When the function W given in Equation (3.9) is expanded using the contracted notations for
strains and elastic constants given in Equation (3.11) W has the following form:

(3.13)

Thus, from Equation (3.13) it can be said that the function W has the following form in terms
of strain components:

(3.14)

With these concepts we proceed to consider the planes of material symmetry. The planes of
the material, also called elastic symmetry are due to the symmetry of the structure of
anisotropic body. In the following, we consider some special cases of material symmetry.

Composite Materials and Structures-AE 16082

(A) Symmetry with respect to a Plane:

Let us assume that the anisotropic material has only one plane of material symmetry. A
material with one plane of material symmetry is called Monoclinic Material.
Let us consider the x1- x2 ( x3= 0) plane as the plane of material symmetry. This is shown in
Figure 3.1. This symmetry can be formulated with the change of axes as follows

(3.15)

With this change of axes,

(3.16)

This gives us along with the use of the second of Equation (3.12)

(3.17)

Figure 3.1: Material symmetry about x1-x2 plane

First Approach: Invariance Approach

Now, the function W can be expressed in terms of the strain components . If W is to be

invariant, then it must be of the form

(3.18)

Composite Materials and Structures-AE 16082

 Comparing this with Equation (3.13) it is easy to conclude that

(3.19)

Thus, for the monoclinic materials the number of independent constants are 13. With this
reduction of number of independent elastic constants the stiffness matrix is given as

(3.20)

Second Approach: Stress Strain Equivalence Approach

The same reduction of number of elastic constants can be derived from the stress strain
equivalence approach. From Equation (3.12) and Equation (3.16) we have

(3.21)

The same can be seen from the stresses on a cube inside such a body with the coordinate
systems shown in Figure 3.1. Figure 3.2 (a) shows the stresses on a cube with the coordinate
system x1, x2,x3 and Figure 3.2 (b) shows stresses on the same cube with the coordinate
system . Comparing the stresses we get the relation as in Equation (3.21).

Now using the stiffness matrix as given in Equation (3.11), strain term relations as given in
Equation (3.17) and comparing the stress terms in Equation (3.21) as follows:

Using the relations from Equation (3.17), the above equations reduce to

Composite Materials and Structures-AE 16082

 Noting that , this holds true only when

Similarly,

This gives us the matrix as in Equation (3.20).

Figure 3.2: State of stress (a) in x1, x2, x3 system

(b) with x1-x3 plane of symmetry

Symmetry with respect to two Orthogonal Planes:

Let us assume that the material under consideration has one more plane, say x2-x3 is plane
of material symmetry along with x1-x2 as in (A). These two planes are orthogonal to each
other. This transformation is shown in Figure 3.3.
This can be mathematically formulated by the change of axes as

(3.22)

And

Composite Materials and Structures-AE 16082

 (3.23)

This gives us the required strain relations as (from Equation (3.12)).

or using contracted notations, we can write,

(3.24)

Figure 3.3: Material symmetry about x1-x2 andx2-x3 planes

First Approach: Invariance Approach

We can get the function W simply by substituting in place of and using contracted
notations for the strains in Equation (3.18). Noting that W is invariant, its form in Equation
(3.18) must now be restricted to functional form

(3.25)

From this it is easy to see that

Thus, the number of independent constants reduces to 9. The resulting stiffness matrix is
given as

Composite Materials and Structures-AE 16082

 (3.26)

When a material has (any) two orthogonal planes as planes of material symmetry then that
material is known as Orthotropic Material. It is easy to see that when two orthogonal
planes are planes of material symmetry, the third mutually orthogonal plane is also plane
of material symmetry and Equation (3.26) holds true for this case also.

Note: Unidirectional fibrous composites are an example of orthotropic materials.

Second Approach: Stress Strain Equivalence Approach

The same reduction of number of elastic constants can be derived from the stress strain
equivalence approach. From the first of Equation (3.12) and Equation (3.23) we have

(3.27)

The same can be seen from the stresses on a cube inside such a body with the coordinate
systems shown in Figure 3.3. Figure 3.4 (a) shows the stresses on a cube with the
coordinate system x1, x2, x3 and Figure 3.4 (b) shows stresses on the same cube with the
coordinate system . Comparing the stresses we get the relation as in Equation
(3.27).
Now using the stiffness matrix given in Equation (3.20) and comparing the stress
equivalence of Equation (3.27) we get the following:

This holds true when . Similarly,

This gives us the matrix as in Equation (3.26).

Composite Materials and Structures-AE 16082

 Figure 3.4: State of stress (a) in x1, x2, x3 system
(b) with x1-x2 andx2-x3 planes of symmetry

Alternately, if we consider x1-x3 as the second plane of material symmetry along with x1-x2
as shown in Figure 3.5, then

(3.28)

And

(3.29)

This gives us the required strain relations as (from Equation (3.12))

or in contracted notations, we write

Substituting these in Equation (3.18) the function W reduces again to the form given in
Equation (3.25) for W to be invariant. Finally, we get the reduced stiffness matrix as given
in Equation (3.26).

Composite Materials and Structures-AE 16082

 Figure 3.5: Material symmetry about x1-x2 andx1-x3 planes

The stress transformations for this coordinate transformations are (from the first of Equation
(3.12) and Equation (3.29))

The same can be seen from the stresses shown on the same cube in x1, x2, x3 and
coordinate systems in Figure 3.6 (a) and (b), respectively. The comparison of the
stress terms leads to the stiffness matrix as given in Equation (3.26).

Note: It is clear that if any two orthogonal planes are planes of material symmetry the third
mutually orthogonal plane has to be plane to material symmetry. We have got the same
stiffness matrix when we considered two sets of orthogonal planes. Further, if we proceed in
this way considering three mutually orthogonal planes of symmetry then it is not difficult to
see that the stiffness matrix remains the same as in Equation (3.26).

Transverse Isotropy:
Introduction:

In this lecture, we are going to see some more simplifications of constitutive equation and develop
the relation for isotropic materials.

First we will see the development of transverse isotropy and then we will reduce from it to
isotropy.

First Approach: Invariance Approach

Composite Materials and Structures-AE 16082

This is obtained from an orthotropic material. Here, we develop the constitutive relation for a
material with transverse isotropy in x2-x3 plane (this is used in lamina/laminae/laminate modeling).
This is obtained with the following form of the change of axes.

(3.30)

Now, we have

Figure 3.6: State of stress (a) in x1, x2, x3 system

(b) with x1-x2 and x1-x3 planes of symmetry

From this, the strains in transformed coordinate system are given as:

(3.31)

Composite Materials and Structures-AE 16082

Here, it is to be noted that the shear strains are the tensorial shear strain terms.

For any angle α,

(3.32)

and therefore, W must reduce to the form

(3.33)

Then, for W to be invariant we must have

Now, let us write the left hand side of the above equation using the matrix as given in Equation
(3.26) and engineering shear strains. In the following we do some rearrangement as

Similarly, we can write the right hand side of previous equation using rotated strain components.
Now, for W to be invariant it must be of the form as in Equation (3.33).

1. If we observe the terms containing and in the first bracket, then we

conclude that is unchanged.
2. Now compare the terms in the second bracket. If we have then the first of
Equation (3.32) is satisfied.
3. Now compare the third bracket. If we have , then the third of Equation (3.32) is
satisfied.
4. Now for the fourth bracket we do the following manipulations. Let us assume that
and is unchanged. Then we write the terms in fourth bracket as

Composite Materials and Structures-AE 16082

 To have W to be invariant we need to have so that the third of Equation
(3.32) is satisfied.

Thus, for transversely isotropic material (in plane x2-x3) the stiffness matrix becomes

(3.34)

Thus, there are only 5 independent elastic constants for a transversely isotropic material.

Second Approach: Comparison of Constants

This can also be verified from the elastic constants expressed in terms of engineering
constants like . Recall the constitutive equation for orthotropic material expressed
in terms of engineering constants. For the transversely isotropic materials the following
relations hold.

When these relations are used in the constitutive equation for orthotropic material expressed
in terms of engineering constants, the stiffness matrix relations in Equation (3.34) are
verified.

Isotropic Bodies
If the function W remains unaltered in form under all possible changes to other rectangular
Cartesian systems of axes, the body is said to be Isotropic. In this case, W is a function of the
strain invariants. Alternatively, from the previous section, W must be unaltered in form under
the transformations

Composite Materials and Structures-AE 16082

 (3.35)

and

(3.36)

In other words, W when expressed in terms of must be obtained from Equation (3.33)
simply by replacing by . By analogy with the previous section it is seen that for this to
be true under the transformation Equation (3.35). We can write

And the transformed strains are given as

(3.37)

Thus, for any angle α,

(3.38)

and therefore, W must reduce to the form

Composite Materials and Structures-AE 16082
 (3.39)

Then, for W to be invariant we must have

Now, let us write the left hand side of above equation using the matrix as given in
Equation (3.34) and engineering shear strains. In the following we do some rearrangement as

(3.40)

Similarly, we can write the right hand side of the previous equation using rotated strain
components. Now, for W to be invariant it must be of the form as in Equation (3.39)

1. From the second bracket, if we propose , then we can satisfy the first of
Equation (3.38).
2. From the third bracket, third of Equation (3.38) holds true when

.
3. The fourth bracket is manipulated as follows:

Thus, to satisfy the second of Equation (3.38) we must have . Further, we

should have . From our observation in 2, we can write .

It follows automatically that W is unaltered in form under the transformation in Equation

(3.36).

Thus, the stiffness matrix for isotropic material becomes as

Composite Materials and Structures-AE 16082

 (3.41)

Thus, for an isotropic material there are only two independent elastic constants. It can be
verified that W is unaltered in form under all possible changes to other rectangular coordinate
systems, that is, it is the same function of as it is of when is changed to .

Constitutive Relations for Orthotropic Materials and Stress-Strain Transformations

In the previous lecture we have seen the constitutive equations for various types of (that is,
nature of) materials. There are 81 independent elastic constants for generally anisotropic
material and two for an isotropic material. Let us summarize the reduction of elastic constants
from generally anisotropic to isotropic material.

1. For a generally anisotropic material there are 81 independent elastic constants.

2. With additional stress symmetry the number of independent elastic constants reduces
to 54.

3. Further, with strain symmetry this number reduces to 36.

4. A hyperelastic material with stress and strain symmetry has 21 independent elastic
constants. The material with 21 independent elastic constants is also called as
anisotropic or aelotropic material.

5. Further reduction with one plane of material symmetry gives 13 independent elastic
constants. These materials are known as monoclinic materials.

6. Additional orthogonal plane of symmetry reduces the number of independent elastic

constants to 9. These materials are known as orthotropic materials. Further, if a
material has two orthogonal planes of symmetry then it is also symmetric about third
mutually perpendicular plane. A unidirectional lamina is orthotropic in nature.

7. For a transversely isotropic material there are 5 independent elastic constants. Plane
2-3 is transversely isotropic for the lamina shown in Figure 3.7.

8. For an isotropic material there are only 2 independent elastic constants.

Principal Material Directions:

Composite Materials and Structures-AE 16082

 The interest of this course is unidirectional lamina or laminae and laminate made from
stacking of these unidirectional laminae. Hence, we will introduce the principal material
directions for a unidirectional fibrous lamina. These are denoted by 1-2-3 directions. The
direction 1 is along the fibre. The directions 2 and 3 are perpendicular to the direction 1 and
mutually perpendicular to each other. The direction 3 is along the thickness of lamina. The
principal directions for a unidirectional lamina are shown in Figure 3.7.

Engineering Constants:

The elastic constants which form the stiffness matrix are not directly measured from
laboratory tests on a material. One can measure engineering constants like Young’s modulus,
shear modulus and Poisson’s ratio from laboratory tests. The relationship between
engineering constants and elastic constants of stiffness matrix is also not straight forward.
This relationship can be developed with the help of relationship between engineering
constants and compliance matrix coefficients.
In order to establish the relationship between engineering constants and the compliance
coefficients, we consider an orthotropic material in the principal material directions. If this
orthotropic material is subjected to a 3D state of stress, the resulting strains can be expressed
in terms of these stress components and engineering constants as follows:

Figure 3.7: Unidirectional lamina with principal material directions

(3.42)

whereas the engineering shear strain components are given as

(3.43)

Composite Materials and Structures-AE 16082

Here, are the Young’s moduli in 1, 2 and 3 directions, respectively. Thus,
represents the axial modulus and represent in-plane transverse and out-of-
plane transverse moduli, respectively. Note that axial direction is along the fibre direction.

represents the shear moduli. G12,G13 are the axial shear moduli in two orthogonal planes
that contain the fibers.G23 represents out-of-plane transverse shear modulus. Further, it should
be noted that .

The term represents the Poisson’s ratio. It is defined as follows

(3.44)

where represents the strain in the direction of applied stress and represents the strain
the associated lateral direction. It should be noted that, in general .
We will mimic some (thought) experiments that we actually do in laboratory to extract these
engineering constants. For example, we find engineering constants of a transversely isotropic
lamina.

Experiment 1: The lamina is loaded in traction along the axial direction as shown in Figure
3.8 (a) and the strains in along three principal directions are recorded as the load is varied.
The slope of the axial stress versus axial strain curve yields the axial Young’s modulus .

The ratios give the Poisson’s ratios respectively.

Experiment 2: The lamina is loaded in traction along direction 2. The two views of this
loading case are shown in Figure 3.8 (b). The slope of stress-strain curve in direction 2 gives
the in-plane transverse Young’s modulus . Since, the material is isotropic in 2-3 plane,
is also equal . The strains in all three directions are measured. The ratios

give the Poisson’s ratios , respectively.

Experiment 3: The lamina is loaded in shear in plane 1-2 as shown in Figure 3.8 (c). The
slope of the in-plane shear stress and engineering shear strain curve gives the shear modulus
. Please note that if we load the lamina in 1-3 plane by shear then also we will get this
modulus because the behaviour of material in shear in these two planes is identical. Thus, by
shear loading in plane 1-2 gives .

Experiment 4: The lamina is loaded in shear in 2-3 plane as shown in Figure 3.8(d). The
corresponding shear stress and engineering shear strain curve yields the shear modulus .

Composite Materials and Structures-AE 16082

Note: We will see the experimental details to measure some of these engineering constants in
a chapter on experimental characterization of lamina, laminates, fibres and matrix materials.

Constitutive Equation for an Orthotropic Material:

Now, let us assume that we have measured all the engineering constants of an orthotropic
material along principal directions. With these engineering constants we know the relation
between the strain and stress components as given in Equation (3.42) and Equation (3.43).
Thus, it is easy to see that we can relate the strain components to stress components through
compliance matrix. Let us recall from previous lecture the stiffness matrix for orthotropic
material (Equation (3.26)). The inverse of this matrix (compliance) will have the same form
as the stiffness matrix. Thus, we write the relationship between strain and stress components
using compliance matrix as follows

Figure 3.8: Experiments to extract engineering

constants for a transversely isotropic material

(3.45)

Composite Materials and Structures-AE 16082

Now compare Equation (3.42) and Equation (3.43) with Equation (3.45). This gives us the
compliance coefficients in terms of engineering constants. The coefficients are given in
Equation (3.46).

(3.46)

It should be noted that like stiffness matrix, the compliance matrix is also symmetric. The
compliance matrix given in Equation (3.45) is shown symmetric.

Note: It is known from our elementary knowledge of linear algebra that inverse of a
symmetric matrix is also a symmetric matrix. Since, the stiffness matrix, which is the inverse
of compliance matrix, is symmetric; the compliance matrix has to be symmetric.

Now, let us derive some more useful relations using the symmetry of compliance matrix. If

we compare and we get . Similarly, comparison of and and

comparison of and give two more similar relations. All these relations are given in
Equation (3.47).

(3.47)

or one can write this relation in index form as

(3.48)

The relations in Equation (3.47) or Equation (3.45) are referred to as the reciprocal relations.
These relations are also written as

(3.49)

It should be noted that, in general . From Equation (3.48) we can write

for as

Composite Materials and Structures-AE 16082

 (3.50)

It is known that for transversely isotropic material (in 2-3 plane) is much greater than
and Thus, from the first of Equation (3.47) one can easily see that is much
smaller than . Further, it is clear from the relation that .

Note: Since value of (and may be of other Poisson’s ratios) will be small, the readers are
suggested to use appropriate precision level while calculating (in examinations and writing
computer codes) any data involving these coefficients.

We will get the stiffness matrix by inversion of compliance matrix. Equation (3.46) is
substituted in Equation. (3.45) and the resulting equation is inverted to give the stiffness
matrix of an orthotropic material as

(3.51)

where

(3.52)

is the determinant of stiffness matrix in Equation (3.51). We can write the stiffness matrix for
transversely isotropic material with the following substitutions in the stiffness matrix.

Further, from the resulting, one can reduce the constitutive equation for isotropic material
with following substitutions:

Composite Materials and Structures-AE 16082

The readers should verify these results.

Constraints on Engineering Constants in Orthotropic Materials

For orthotropic materials there are constraints on engineering constants. These constrains
arise due to thermodynamic admissibility. For example, in case of isotropic materials it is
well known that the Young’s modulus and shear modulus are always positive. Further, the
Poisson’s ratio lager than half are not thermodynamically admissible. If these constrains are
violated then it is possible to have a nonpositive strain energy for certain load conditions.
However, for isotropic materials the strain energy must be a positive definite quantity.

In this section, based on the work done by Lempriere we are going to assess the implications
of this thermodynamic requirement (positive definiteness of strain energy) for orthotropic
materials.
The sum of work done by all stress components must be positive, otherwise energy will be
created. This condition imposes a thermodynamic constraint on elastic constants. This
condition requires that both compliance and stiffness matrices must be positive definite. In
other words, the invariants of these matrices should be positive.

Let us look at this condition with physical arguments. For example, consider that only one
normal stress component is applied. Then we can find the corresponding strain component
from the corresponding diagonal entry of the compliance matrix. Thus, we can say that for
the strain energy to be positive definite the diagonal entries of the compliance matrix must be
positive. Thus,

(3.53)

In a similar way, it is possible under certain conditions to have a deformation which will give
rise to only one normal strain component. We can find the corresponding stress using the
corresponding diagonal entry in stiffness matrix. For the strain energy produced by this stress
component to be positive the diagonal entry of the stiffness matrix must be positive. Thus,
this condition reduces to

(3.54)

and the determinant of the compliance matrix must also be positive. That is,

(3.55)
Now, using the reciprocal relations given in Equation (3.49), the condition in Equation (3.54)
can be expressed as

Composite Materials and Structures-AE 16082

 (3.56)

This condition also justifies that the Poisson’s ratio greater than unity is feasible for
orthotropic lamina. Poisson’s ratio greater than unity is sometimes observed in experiments.

The condition in Equation (3.55) can be written as

The terms inside the brackets are positive. Thus, we can write

(3.57)

This condition shows that all three Poisson’s ratios cannot have large positive values and that
their product must be less than half. However, if one of them is negative no restriction is
applied to remaining two ratios.

Let us consider the transverse isotropy as a special case. Let us consider transverse isotropy
in 2-3 plane. Let

(3.58)

Then, the conditions in Equation (3.56) reduce as

(3.59)

and Equation (3.57) (using reciprocal relations in Eq. (3.49)) becomes as

Composite Materials and Structures-AE 16082

 (3.60)

The condition posed by above equation is more stringent than that posed in Equation (3.59).

Note that the quantities and are both positive. Thus, the limits on Poisson’s ratio
in transverse plane are

(3.61)

Further, consider a special case of isotropic material where and . This

simplifies Equation (3.61) to a well known condition

(3.62)

Stress and Strain Transformation about an Axis

Often it is required to transform the stress or strain tensor from one coordinate axes system
to another. For example, if the fibres in a lamina are not oriented along direction x, then we
may need to transform the stress and strain components from principal material directions
1-2-3 to global directions xyz or vice-a-versa. It should be noted that the stress and strain
tensors are second order tensors. Hence, they follow tensor transformation rules.

In this section we are going to introduce two notations. The subscripts 123 will denote a
quantity (like constitutive equation, engineering constants, etc.) in principal material
directions, while subscripts xyz will denote the corresponding quantity in global coordinate
directions.

Figure 3.9: Unidirectional lamina with global xyz

directions and principal material 1-2-3 directions

Composite Materials and Structures-AE 16082

 Let us transform the stress and strain components for the case shown in Figure 3.9. Here,
xy plane is rotated about direction z to 1-2 plane. Here, direction z and direction 3 are in
same directions, that is, along the thickness direction of lamina. The direction cosines for
this axes transformation are as given in Equation (2.2). However, these are again given
below.

Stress Transformation:

Let us do the stress transformation as given in Equation (2.7). In this equation the primed
stress components denote the component in 123 coordinate system. Using the expanded
form of Equation (2.7) and stress symmetry, let us obtain component of stress

Thus, substituting the values of direction cosines from above, we get

The remaining five stress terms (using stress symmetry) on the left hand side are also
obtained in a similar way. Let us write the final form of the relation as

(3.63)

Here, and and is

the stress transformation matrix. Thus, comparing all the terms as in Equation (3.63), we
can write as

Composite Materials and Structures-AE 16082

 (3.64)

where . It should be noted that is not symmetric.

Example 1: Calculate the stiffness and compliance coefficients for transversely isotropic
material AS4/3501 Epoxy. The properties are as given below for a fibre volume fraction of
60%.

Solution:

Unit of all compliance coefficients is 1/GPa.

The corresponding stiffness coefficients are calculated by inversion of the compliance matrix.

Unit of all stiffness coefficients is GPa.

Note: Both stiffness and compliance matrices are symmetric.

Example 2: Transform the stiffness and compliance matrix of Example 1 about axis 3 by an
angle of
Composite Materials and Structures-AE 16082
 = 30°.

Solution:

Approach 1: One can find the transformation matrices and do the matrix
multiplication as given in Equation (3.73) for transformed stiffness matrix and then inverse
this matrix or do the matrix multiplication as given in Equation (3.80) to get the transformed
compliance matrix. The use of Equation (3.73) and Equation (3.80) is suggested because
remembering is not so difficult. Further, their inverse can be easily found with
the help of Equation (3.74).
For

Thus

Composite Materials and Structures-AE 16082

Unit of all transformed stiffness coefficients is GPa.

Unit of all transformed compliance coefficients is 1/GPa.

Approach 2: You can write the expanded form for transformed stiffness and compliance
coefficients in Equation (3.76) and Equation (3.82). However, the readers are suggested to
use this approach only when they are confident of remembering these terms.

Example 3: The coefficients of moisture absorption for T300/5208 composite material are
. Plot the variation these coefficients between
.

Solution:

We have the expression for variation of the coefficients of moisture absorption as

Composite Materials and Structures-AE 16082
where, . We plot the above variation using a computer code. The final
plot is shown in Figure 3.11.

Figure 3.11: Variation of coefficients of moisture

expansion with orientation of fibres

Macroscopic Failure Theories

Introduction

In this lecture we will first present the issues or difficulties in failure theories for composite
as compared to the homogeneous and isotropic materials. Then we will introduce some
failure theories for unidirectional composites. We will conclude this lecture with some
numerical examples.

Many theories are available for predicting failure of composites. These theories predict the
failure of a lamina or laminate. Hence, these theories are called “lamina failure” or “laminate
failure” theories. Further, these theories predict the very first failure in a lamina. Hence, these
theories are popularly known as “first-ply failure” theories.

The macroscopic theories presented here are the early theories. From a designer point of
view, any theory should be applicable at lamina, laminate as well as at component level. The
aims of these theories were to give reasonably accurate prediction of failure as compared to
experimental results and ease of implementation for analysis and design. Hence, some of
theories were based on physical basis. Some of them were just extensions of theories for
homogeneous; isotropic materials to composite materials. While most of the theories
Composite Materials and Structures-AE 16082
 provided mere mathematical expressions such that it gave a best fit of the available
experimental data.

All together, these theories are not good enough to predict the failure at all levels (like
constituent, lamina and laminate). Further, none of them can be used for a general loading
and any composite material as most of them were loadings and materials specific. At present,
a significant progress has been made to address most of these issues.

Issues in Failure Theories for Unidirectional Composites:

The failure theories for unidirectional composites have some difficulties when they are
extended from homogeneous, isotropic materials. In the following, we list some of the issues
related to composite failure theories.

1. The composites are heterogeneous and orthotropic in nature. Hence, the effective
properties in three directions need to be found.
2. The unidirectional laminae are orthotropic in nature. Hence, the strength parameters
(like ultimate stress or strain) in three directions will be different.
3. In a given direction, the strength parameters will be different in tension and
compression.
4. The strengths in normal direction and in shear directions are different.
5. For off axis laminae, the shear strength is different in positive and negative directions.
If one is using the global coordinate system to decide the shear failure, then the
positive and negative shear should be considered carefully. However, in principal
material directions the positive and negative shear has no effect on shear strength.
6. The strength parameters are generally obtained experimentally in principal directions.
Hence, the stresses or strains used in the failure theories should also be in principal
directions. Thus, the transformation of stresses or strains from global coordinate
system to principal material direction in each lamina is imperative.
7. Most of the theories do not give the mode of failure (like fibre breaking, matrix
cracking, etc.). It just mentions that the lamina has failed. Further, they do not give
propagation of damages until final failure.
8. The link between damage and first-ply failure is difficult to establish for failure
theories.

Note: (We refer to point 6 in the above.) One should not transform the strength parameters
from principal coordinate system to global coordinate system to use it in a failure theory.
This transformation from principal to global direction is not known. The strengths should be
obtained by experiments on off-axis laminae.
Macroscopic Failure Theories:

In the following we will present some of the popular macroscopic failure theories used in
design and analysis of composites.

In these theories we will use following quantities and symbols.

1. X denotes the ultimate normal stress magnitude in fibre direction (1-direction).

2. Y denotes the ultimate normal stress magnitude in in-plane transverse direction (2-
direction).

Composite Materials and Structures-AE 16082

 3. Z denotes the ultimate normal stress magnitude in transverse direction (3-direction)
4. Subscript T and C denote tension and compression, respectively.
5. Q, R and S denote the ultimate shear stresses corresponding to 23, 13 and 12 planes.

We will see some definitions related to failure theories.

Strength Ratio (SR):

It is defined as the ratio of maximum load which can be applied such that a lamina does not
fail to the actual load applied. Thus,

(6.1)

This concept can be extended to any failure theory. The strength ratio gives the factor by
which the actual applied load can be increased or decreased upto a lamina failure. For
example, if , it means that the lamina is safe and load applied can be increased by this
factor and if , it means that the lamina is unsafe and the load applied must be
decreased by this factor. It is needless to say that when the condition corresponds to
failure load.

Failure Envelope:

The failure envelope is a surface formed by various combinations of normal and shear
stresses (or strains) that can be applied to a lamina just before it fails. Thus, any state of stress
(or strain) which lies inside the envelope is safe whereas the one which lies on or outside the
envelope is unsafe.

1. Maximum Stress Theory:

This theory is a direct extension of maximum normal stress theory proposed by Rankiene [1]
and maximum stress theory proposed by Tresca [2] for homogeneous, isotropic materials. In
this theory the three normal and three shear stress components are compared with
corresponding ultimate stresses. A given normal stress is compared with corresponding
positive and negative, that is tensile and compressive ultimate stresses. The magnitude of
shear stress is compared with corresponding ultimate shear stress.

Thus, the maximum stress theory results in the following expression for the safe condition.

For normal stresses,

(6.2)

Composite Materials and Structures-AE 16082

For shear stresses,

OR (6.3)

Thus, according to this theory initiation of failure will correspond to one or more inequalities
in Equations (6.2) and (6.3) become an equality. The maximum stress theory can be
represented as intersecting planes in 3D stress space or intersecting lines in 2D stress space.

The stress space is shown as intersecting planes in Figure 6.7. The region inside
this space is regarded as safe, whereas any point on or outside the intersecting planes will be
an unsafe or a failure point. A safe state of stress with normal stresses alone is shown inside
the envelope.
A fully 3D state of stress will represent an envelope or surface in six dimensional stress
space.

The maximum stress theory for planar state of stress is given for normal stresses as

(6.4)

and for shear stress as

(6.5)

Now, consider that an off axis lamina is subjected to an axial stress of . Then, we can
write the maximum stress theory for the planar state of stress for off axis lamina as follows.

Recalling the stress transformation for planar state of stress, we write the stress components
in principal material directions as

(6.6)

Thus, the maximum stress theory for off-axis lamina loaded axially can be written as

Composite Materials and Structures-AE 16082

 Figure 6.7: Failure envelope for normal stress space with an example safe
stress state inside the envelope

(6.7)

2. Maximum Strain Theory:

Maximum strain theory is equivalent to maximum stress theory. This theory is based on
maximum normal strain theory of St. Venant and the strain equivalent of maximum shear
stress theory of Tresca for isotropic materials.

According to this theory, a lamina fails if either of the normal strain exceeds the maximum
allowable strain in tension or compression or any of the shear strain exceeds the maximum
allowable shear strain. The inequalities resulting are:

For normal strains,

(6.8)

For shear stresses,

Composite Materials and Structures-AE 16082

 OR (6.9)

where, and are the ultimate normal strains in tension and compression,
respectively. Further, are ultimate shear strains in 23, 13, 12 planes,
respectively.
Thus, according to this theory initiation of failure will correspond to one or more
inequalities in Equations (6.8) and (6.9) become equality.

The maximum strain theory for planar stress can be expressed as

(6.10)

and for shear strain as

(6.11)

The strains can be obtained from constitutive equation for strains in terms stresses as

(6.12)

These equations can be put in Equation (6.10) and Equation. (6.11). Further, for axial
stress applied we can write the stresses in principal directions as in Equation (6.6).

Note: The maximum stress and maximum strain theories are similar. In both theories there
is no interaction between various components of stress or strain. However, the two theories
yield different results.

3. Tsai-Hill Theory:

This theory is an extension of distortional energy yield criterion of von-Mises [3] for
isotropic materials as applied to anisotropic materials. It is known that total strain energy in a
body is composed of two parts: One is distortion energy which cause change in shape and
second one is dilation energy which causes the change in size or volume. In the von-Mises
criterion it is assumed that the material fails when the maximum distortion energy of the body
exceeds the distortion energy corresponding to yielding of the same material in tension.

Composite Materials and Structures-AE 16082

Hill [4] extended the von-Mises distortion energy criterion of isotropic materials to
anisotropic materials. Later Tsai [5, 6] extended this criterion for anisotropic materials to a
unidirectional lamina. Hence, the theory is called Tsai-Hill theory.

According to this theory the failure takes place when the stress state is such that

(6.13
)

which upon simplifications can be written as

(6.14)

where, F, G, H, L, M and N are the material strength parameters. Thus, any state of stress
which lies inside this envelope is safe and the one which lies on or outside the envelope is
unsafe.

The strength parameters correspond to failure stresses in one dimensional loading. These can
be obtained by a set of thought experiments. For example, consider that for the pure shear
loading in 2-3 plane, that is with , with corresponding shear strength Q and all
other stress components are zero, the Equation (6.14) becomes

(6.15)

Similarly, for the other two shear stress components, we can get

(6.16)

Now the strength parameters F, G and H are obtained by three states of stress. The state of
stress with corresponding strength X and all other stress components being zero, in
Equation (6.14) leads to

(6.17)

Now the conditions and (and other stress components being zero) in Equation
(6.14) result in

(6.18)

Composite Materials and Structures-AE 16082

solving the simultaneous equations in Equation (6.17) and Equation (6.18), we get the
required strength parameters as

(6.19)

Thus, Equation (6.14) becomes

(6.20)

This is Tsai-Hill theory for 3D state of stress. Note that this is quadratic in stress terms with
no linear terms.

Now consider a transversely isotropic material with inplane stresses as the significant
stresses. For this planar state of stress, we have and remaining stress
components are non zero. In this case the failure envelope becomes a three dimensional
space. Thus, the failure condition in Equation (6.14) becomes

(6.21)

Now, using the strength parameters from Equation (6.17), Equation (6.18) and Equation
(6.19), we get

(6.22)

For transverse isotropy, we also have . Thus, the above equation is rearranged as

(6.23)

The above equation gives the Tsai-Hill criterion for failure for planar state of stress. From
Tsai-Hill theory it is clear that it does not differentiate between tension and compression
strengths for normal stresses. Infact, Tsai-Hill theory assumes same strengths in tension
and compression. However, this situation does not occur in case of shear stresses. Thus, for
normal stresses the theory represents a severe limitation that the sign of the normal stresses
should be known a priori and the appropriate strength value should be used for normal
stresses in the failure theory.

Composite Materials and Structures-AE 16082

 It should be noted that unlike maximum stress theory or maximum strain theory Tsai-Hill
theory considers the interaction between three lamina strength parameters or interaction
between stress components.

Further, it should be noted that Tsai-Hill theory is a unified theory and does not give the
mode of failure like the maximum stress and maximum strain theory. However, one can

make a guess of failure mode by calculating the quantities , and . The maximum of
these three values can be said to give the mode of failure.

Note: The right hand side of Equation (6.20) or (6.23) is called as “failure index”.

4. Hoffman Theory

The Hoffman criterion [1] is extension of Tsai-Hill theory. In Tsai-Hill theory the strength
parameters are obtained without considering the difference in values in tension and
compression. However, one should realistically consider the differing tension and
compression strengths that characterize brittle behaviour. This can be done by adding odd
functions of and in the expression of Tsai-Hill criterion. Thus, Equation (6.13)
becomes

(6.24)

where, are the material parameters. These are uniquely determined from
nine basic strength data, namely, three uniaxial tensile strengths, and ; three
uniaxial compressive strengths, and and three shear strength parameters, Q, R
and S. To determine these material strength parameters, we need to do thought experiments
as follows:

First, consider the state of stress such that and all other stress components will be
zero. For shear failure in this mode, we need . Putting this in Equation (6.24), we
get

(6.25)

Similarly, we get the constants

(6.26)

(6.27)

Composite Materials and Structures-AE 16082

 Now, to find remaining constants we apply following state of stress. Let and all
other stress components be zero. For tensile failure, we need for this stress state.
Putting this in Equation (6.24) we get

(6.28)

Similarly for this stress state, the compression failure requires . This results the
Equation (6.24) to give

(6.29)

Now, we have two stress states: First one as and all other stress components are
zero and the second one as and all other stress components are zero. Again, as in
previous case for failure in tension and compression, Equation (6.24) results into following
conditions:

(6.30)

Thus, Equations (6.28), (6.29) and (6.30) give a set of six simultaneous equations in
. Solving these, we get

(6.31)

and

(6.32)

Thus, the Hoffman criterion as given in Equation (6.24) becomes

(6.33)

Composite Materials and Structures-AE 16082

 Now consider transverse isotropy of the material in 2-3 plane. Thus, and .
For shear strength, we have . Then for plane stress condition , the
criterion in Equation. (6.33) becomes

(6.34)

Equation (6.34) represents Hoffman criterion for planar state of stress in transversely
isotropic materials. It should be noted that in this criterion there is no need to check the
sign of the stress components to decide whether a tensile or compressive strength is to be
used.

5. Tensor Polynomial Failure Criterion:

Tensor polynomial criterion attempts to mathematically overcome one of the shortcomings of

the quadratic criteria that they do not account for differences between tensile and
compressive strengths.

In the present lecture we will see the second order tensor polynomial criterion proposed by
Tsai and Wu [2]. This is a complete quadratic tensor polynomial which includes the linear
terms.

The failure surface in the stress space has the following scalar form:

(6.35)

where, and are strength tensors of the second and fourth order, respectively. The
expanded form of the above equation is

(6.36)

It should be noted that the linear term take into account the difference between tensile
stress and compressive stress induced failures. The quadratic terms define an ellipsoid in
the stress space.
The features of this theory are given below:

Composite Materials and Structures-AE 16082

 1. The resulting criterion is a scalar function and thus an invariant. Further, the
interactions between various components are independent unlike in Tsai-Hill theory
where interactions are fixed and in case of maximum stress or maximum strain theory
these interactions are not possible.
2. The strength components are expressed as tensors; one can use the transformation
rules as discussed earlier for their transformations. Further, the invariants of these
strength tenors are also well defined.
3. The property of symmetry of strength tenors and number of independent and non-zero
components can be derived in similar way that we carried out for anisotropic
materials earlier.
4. One can either transform the strength parameters from to and to or
transform to . Most of the existing criteria are limited to specially orthotropic
materials. Such criteria can be applied only by transforming the external stresses to
material axes. However, the transformations of strength criteria cannot be done as this
transformation is not known.
5. Since Equation (6.36) is an invariant, hence, an invariant for any other coordinate
system. This also holds for curvilinear coordinate system as well.
6. The theory has also introduced constraints over the magnitude of the strength
interaction terms in the following manner :

(6.37)

7. The terms represent diagonal terms and are positive terms, whereas the off-
diagonal terms can be positive or negative depending upon their interaction with other
terms. However, their magnitudes are constrained by Equation (6.37). Further, the
inequality in Equation (6.37) is very important as it assures the failure envelope in
Equation (6.35) intersects all stress axes. The surface formed by Equation (6.35) is an
ellipsoid and Equation (6.37) ensures that it is a closed surface unlike a hyperboloid.
It should be noted that the positive definiteness is also imposed for terms.
8. Gol’denbalt and Kopnov [3] proposed a tensorial criterion in a general form of

(6.38)

This form of the equation is more complicated than in Equation (6.35). Further, the size of
the strength terms is enormously very high to handle and the additional terms do not yield
more generality than the linear and quadratic terms.

Note: Statement 4 above equivalently says that the Tsai-Wu strength criterion can be given in
transformed coordinate system. The transformed criterion may be given as

(6.39)

And the strength terms can be transformed using the following relations

Composite Materials and Structures-AE 16082

 (6.40)

However, the transformation of any other strength criteria may not hold true.

Now we will seek simplifications to Equation (6.36) as follows. The strength tensors can be
written in the form

(6.41)

and

(6.42)

where, both strength tensors are assumed to be symmetric with 6 and 21 independent
constants for and , respectively.

The number of independent strength parameters can be further reduced if we have some form
of material symmetry. We consider a special case of specially orthotropic material. Thus, for
specially orthotropic material the terms and will vanish. Further, the off-diagonal
terms which give normal shear coupling like , etc. will also become zero if we
assume that the sign of shear stress does not change the failure stress. Further, with same
reasoning we assume that the shear strengths are also uncoupled leading to
. Thus, with this material symmetry, we have

and

The number of independent strength parameters are now 3 and 9 for and , respectively.
Thus, for orthotropic material the criterion becomes

(6.43)

Composite Materials and Structures-AE 16082

Now we will determine the strength parameters by thought experiments. First, we apply
and other stress components being zero. For this state of stress, the failure in tension
requires . Thus, Equation (6.43) becomes

(6.44)

Similarly, for this state of stress, the failure in compression requires . This results in
Equation (6.43) to become

(6.45)
Equation (6.44) and Equation (6.45) are two simultaneous equations with and as two
unknowns.

Solution of these two equations gives

(6.46)

Likewise, if we apply stress states as and with other stresses being zero, it will
give us the following constants:

(6.47)

(6.48)

Similarly, if we apply , , and with other stress components as zero, as

three separate states of stress, then we get following constants:

(6.49)

So far we have developed expressions for 3 strength terms for and diagonal terms of .
Now the expressions for off-diagonal terms of require combined state of stress to be
applied. The pure axial or shear state of stress will not be sufficient. In other criteria the
interaction terms like are assumed to be dependent or terms like are zero.

There are an infinite combinations of the stresses from which these terms can be obtained.
However, one should choose those combinations which can yield the desired result in a
reliable and easy manner. In the following, we will see typical combinations of stresses to
find .
Composite Materials and Structures-AE 16082
Consider the equivi-biaxial stress state and other stress components are zero.
Putting this in Equation (6.43), we get

(6.50)

Solving for , we get

(6.51)

Similarly, if we apply equivi-biaxial stress states in 1-3 and 2-3 planes, then we get the
following constants:

(6.52)

(6.53)

We can find the constants and by imposing the equivi-biaxial state of stress.
However, practically it is very difficult to impose such a state of stress. Hence, many
researchers have proposed tests on angle specimens to determine these strength
parameters.

Consider the stress state

(6.54)

and other stresses are zero. Here, is the axial tensile strength of specimen. The stress
state in Equation (6.54) is obtained by applying in axial direction for specimen. One
should be able to get the state of stress in Equation (6.54) from our earlier stress
transformation equations. Putting Equation (6.54) in Equation (6.43), we get

(6.55)

which upon solving for gives

(6.56)

Composite Materials and Structures-AE 16082

A similar expression can be derived with compressive strength of specimen. One can
further find this constant using the in-plane shear strength of specimen, which
produces the stress state as

(6.57)

This reduces Equation (6.43) to

(6.58)

The solution from this equation is

(6.59)

Note: In case of anisotropic materials the constant is no longer zero. This can be obtained
by a tension-torque test such that it results in following stress state

(6.60)

Using this in Equation (6.43), we get

(6.61)

which gives

(6.62)

The Tsai-Wu criterion for planar state of stress can be given as

(6.63)

The strength parameters are as given above. If the strength term then, the criterion
becomes

Composite Materials and Structures-AE 16082

 (6.64)

Composite Materials and Structures-AE 16082