CHAPTER 2

INTRODUCTION TO ELASTICITY

İ. ÖZKOL 1
 2.1 Concept Of Displacement
Consider a material point P at the position
x(x, y, z) before deformation; see Fig. 2.1.
After deformation, P moves to a new
position 𝑃′ (𝑥 ′ , 𝑦 ′ ,𝑧 ′ ). The change of
position during deformation, which is
measured in terms of the displacement
vector u, has three components: u, v, and
w in the x, y, and z directions, respectively.
İ. ÖZKOL 2
İ. ÖZKOL 3
The new location of the point (x,y,z) after the
deformation is given by
𝑋′ = x + u u = 𝑋′ − x
𝑌′ = y + v or v = 𝑌′ − y
𝑍′ = z + w w = 𝑍′ − z (2.1)
Thus, the deformed configuration is uniquely
defined if the displacement components u, v,
and w are given everywhere in the body of
interest.

İ. ÖZKOL 4
Consider an axial member [i.e., a one-
dimensional (1-D) body] of original length
𝐿0 . Assume the axial strain to be uniform
in the member. Then the axial strain
everywhere in the member is calculated
by
Δ𝐿
𝜀= (2.2)
𝐿0

İ. ÖZKOL 5
where ∆𝐿 is the total elongation of the member.
The elongation ∆𝐿 can be regarded as the
difference in displacement 𝑢1 = 𝑢(𝑥1 ) at the right
end and 𝑢0 = 𝑢(𝑥0 ) at the left end (see Fig. 2.2),
i.e.,
∆𝐿 = 𝑢1 − 𝑢0

İ. ÖZKOL 6
The function 𝑢 𝑥 = 𝑢0 + 𝜀0 𝑥 − 𝑥0 gives the axial
displacement at any point ‘x’ in the axial member.
Δ𝐿
If the strain is not uniform, then (2.2) (𝜀 = ) gives
𝐿0
an average strain. To determine the strain at a point,
a small segment 𝐿0 = ∆𝑥 must be considered.
Consider two points 𝑥0 and 𝑥0 + ∆𝑥 that are
separated by a small distance ∆𝑥. Let the
displacements at these two points be
𝑢0 = 𝑢(𝑥0 ) and 𝑢1 = (𝑥0 + ∆𝑥)
respectively.
İ. ÖZKOL 7
The difference in displacement between
these two points is
∆𝑢 = 𝑢1 − 𝑢0 = 𝑢 𝑥0 + ∆𝑥 − 𝑢(𝑥0 ) (2.3)
which can also be regarded as the
elongation of the material between these
two points. The axial strain in this
segment (or at point 𝑥0 ) is defined as
İ. ÖZKOL 8
 Δ𝑢 𝑑𝑢
𝜀= lim = (2.4)
∆𝑥→0 Δ𝑥 𝑑𝑥
Thus, axial strain can be obtained from the
derivative of the displacement function.
If a rod is subjected to a uniform
tension and 𝜀 = 𝜀0 =constant, then
𝑑𝑢
= 𝜀0 𝑥0 ≤ 𝑥 ≤ 𝑥0 + 𝐿0
𝑑𝑥

İ. ÖZKOL 9
Integrate the above equation to obtain
𝑢 = 𝜀0 𝑥 + 𝐶 Let 𝑢 𝑥0 = 𝑢0 ; then, from
the above equation, 𝐶 = 𝑢0 − 𝜀0 𝑥0 and
the displacement function is given by
𝑢 = 𝜀0 𝑥 − 𝑥0 + 𝑢0 (2.5)

İ. ÖZKOL 10
 2.2 STRAIN
Consider two points P and Q in a solid body. The
coordinates of P and Q are (x, y, z) and (x + ∆x, y, z),
respectively. The distance between the two points
before deformation is ∆x; see Fig. 2.3.

İ. ÖZKOL 11
After deformation, let the displacement of P in the
x-direction be u=u(x, y, z) and of Q be 𝑢′ =u(x + ∆x, y, z).
The new distance between these two points
′ ′
(𝑃 and 𝑄 ) in the x-direction after deformation is
(x + ∆x + 𝑢′ ) - (x + u) = ∆x + ∆u (2.6)
where ∆u = 𝑢′ - u is the "change of length" in the x-
direction for material connecting P and Q after
deformation. The strain is defined just as in an axial
member:
Δ𝑢 𝜕𝑢
𝜀𝑥𝑥 = lim = (2.7)
∆𝑥→0 Δ𝑥 𝜕𝑥

İ. ÖZKOL 12
This is the x-component of the normal
strain, which measures the "deformation in
the x-direction" at a point (x,y,z). Similarly,
the y-component and z-component of the
normal strain at the point are given by
𝜕𝑣 𝜕𝑤
𝜀𝑦𝑦 = (2.8) and 𝜀𝑧𝑧 = (2.9)
𝜕𝑦 𝜕𝑧

İ. ÖZKOL 13
Comparing the strain component 𝜀𝑥𝑥 with
the strain in the 1-D case (or in an axial
member), we may interpret 𝜀𝑥𝑥 as the
elongation per unit length of an
"infinitesimal" axial element of the material
at a point (x, y, z) in the x-direction. Similar
interpretations can be given to 𝜀𝑦𝑦 and 𝜀𝑧𝑧

İ. ÖZKOL 14
The three normal strain components are
not sufficient to describe a general state of
deformation in a 3-D body. Additional
shear strain components are needed to
describe the distortional deformation.

İ. ÖZKOL 15
For simplicity, let us consider a 2-D case.
Let P,Q,R be three neighboring points all
lying on the x-y plane as shown in Fig. 2.4.

İ. ÖZKOL 16
 ′ ′ ′
Let 𝑃 ,𝑄 , and 𝑅 be the corresponding
positions after deformation. For no loss of
generality, we assume that P does not
move. In addition, we assume that the
infinitesimal elements 𝑃𝑄 and 𝑃𝑅 do not
experience any elongation. Thus, the
′ ′
positions of 𝑃 and 𝑄 (see Fig. 2.4) are
given by

İ. ÖZKOL 17
 ′
𝑄 : (𝑥 + ∆𝑥, y + ∆𝑣)
′
𝑅 : 𝑥 + ∆𝑢, y + ∆𝑦
′
For 𝑄 , the displacement increment ∆v is
∆𝑣 = 𝑣 𝑥 + ∆𝑥, 𝑦 − 𝑣(𝑥, 𝑦) (2.10)
′
Similarly for 𝑅 , the displacement
increment ∆u can be written as
∆𝑢 = 𝑢 𝑥, 𝑦 + ∆𝑦 − 𝑢(𝑥, 𝑦) (2.11)

İ. ÖZKOL 18
The rotations 𝜃1 and 𝜃2 of elements 𝑃𝑄 and
𝑃𝑅 are assumed to be small and are given by

Δ𝑣 𝜕𝑣 Δ𝑢 𝜕𝑢
𝜃1 = lim = 𝜃2 = lim =
∆𝑥→0 Δ𝑥 𝜕𝑥 ∆𝑦→0 Δ𝑦 𝜕𝑦

The total change of angle between 𝑃𝑄 and

𝑃𝑅 after deformation is defined as the shear
strain component in the x-y plane:

𝜕𝑣 𝜕𝑢
𝛾𝑥𝑦 = 𝛾𝑦𝑥 ≡ 𝜃1 + 𝜃2 =İ. ÖZKOL + (2.12)
𝜕𝑥 𝜕𝑦 19
Similar shear strain components in the y-z
plane and x-z plane are defined as
𝜕𝑤 𝜕𝑣
𝛾𝑧𝑦 = 𝛾𝑦𝑧 ≡ + (2.13)
𝜕𝑦 𝜕𝑧
And
𝜕𝑤 𝜕𝑢
𝛾𝑧𝑥 = 𝛾𝑥𝑧 ≡ + (2.14)
𝜕𝑥 𝜕𝑧

Thus, a general state of deformation at a point

in a solid is described by three normal strain
components 𝜺𝒙𝒙 , 𝜺𝒚𝒚 , 𝜺𝒛𝒛 and three shear
strain components 𝜸𝒙𝒚 , 𝜸𝒚𝒛 , 𝜸𝒙𝒛
İ. ÖZKOL 20
Rigid Body Motion
If a body undergoes a displacement without
inducing strains in the body, then the motion is
a rigid body motion. For instance, the following
displacements represent a rigid body
translational motion and do not yield any
strains
𝑢 = 𝑢0 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑣 = 𝑣0 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝑤 = 𝑤0 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡

İ. ÖZKOL 21
Another rigid body motion is the rigid
body rotation. The following displacements
represent a rigid body rotation in the x-y
plane.
𝑢 = −𝛼𝑦
𝑣 = 𝛼𝑥 (2.15)
𝑤=0
It is easy to verify that no strains are
associated with the above displacement
field.
İ. ÖZKOL 22
Example 2.1
Simple Shear : Consider a 2-D body (a unit square ABCD) in
the x-y plane as shown in Fig. 2.5. After deformation, the
four corner points move to 𝐴′ ,𝐵′ , 𝐶 ′ and 𝐷′ respectively.
Assume that the displacement field is given by
𝑢 = 0,01𝑦 𝑣 = 0,015𝑥 (a)

İ. ÖZKOL 23
Using (2.1) the new position of point A after
deformation is given by
𝑥 ′ = 0 + 𝑢I𝑥=0,𝑦=1 = 0.01
𝑦 ′ = 1 + 𝑣I𝑥=0,𝑦=1 = 1
New coordinates of A ′ : (0.01,1)
Similarly, we obtain the new positions of B, C, and D.
B ′ : (1.01,1.015)
C ′ : (0, 0)
D ′ : (1,0.015)
Since the deformed configuration is linear in x and y, it can be
determined from the new positions A ′, B ′ , C ′, and D ′ as
shown by the dashed lines in Fig. 2.5.
İ. ÖZKOL 24
The strains (normal and shear) corresponding
to the displacements given by (a) are
𝜕𝑢 𝜕𝑣
𝜀𝑥𝑥 = = 0 , 𝜀𝑦𝑦 = =0
𝜕𝑥 𝜕𝑦
𝜕𝑣 𝜕𝑢
𝛾𝑥𝑦 = + = 0,015 + 0,01 = 0,025
𝜕𝑥 𝜕𝑦
This is a simple shear deformation.
The following displacement field which does
not yield any non vanishing strain component
is a rigid body rotation.
U = 0.01y
V = -0,01x
İ. ÖZKOL 25
 2.3 STRESS
For an axial member, the force is always parallel
to the member, and the stress is defined as
𝑃
𝜎= (2.16)
𝐴
where A is the cross-sectional area. If A = 1 unit
area, then 𝜎 = P. The concept of stress can easily
be extended to 3-D bodies subjected to loads
applied in arbitrary directions. Consider an
infinitesimal plane surface of area ∆A with a unit
normal vector n. The total resultant force acting
on this area is F (force is a vector); see Fig. 2.6.
İ. ÖZKOL 26
The stress vector t is defined as
∆𝐹
𝒕= lim (2.17)
∆𝐴→0 ∆𝐴
Thus, t can be considered as the force per unit area
acting on the given plane surface.
İ. ÖZKOL 27
Consider the special plane surface with the
unit normal vector parallel to the x-axis. On
this face, the stress vector t has three
components, which are denoted by 𝝈𝒙𝒙 ,
𝝉𝒙𝒚 , 𝝉𝒙𝒛 as shown in Fig. 2.7. Similarly, on
the y-face and z-face the force intensities
are given by the components of the
respective stress vectors as 𝜎𝑦𝑦 , 𝜏𝑦𝑥 , 𝜏𝑦𝑧
ve 𝜎𝑧𝑧 , 𝜏𝑧𝑥 , 𝜏𝑧𝑦
İ. ÖZKOL 28
İ. ÖZKOL 29
Consider an infinitesimal solid element under
a state of uniform stress. The stress
components on the six faces of this element
are shown in Fig. 2.8. Since the body is in
equilibrium, the six equations of equilibrium
must be satisfied,
σ 𝐹𝑥 = 0, σ 𝐹𝑦 = 0, σ 𝐹𝑧 = 0 (2.18)
σ 𝑀𝑥 = 0, σ 𝑀𝑦 = 0, σ 𝑀𝑧 = 0 (2.19)

İ. ÖZKOL 30
İ. ÖZKOL 31
The force equations (2.18) are obviously
satisfied automatically. To satisfy (2.19), the
following relations among the shear stress
components are necessary.
𝝉𝒙𝒚 = 𝝉𝒚𝒙 , 𝝉𝒚𝒛 = 𝝉𝒛𝒚 , 𝝉𝒙𝒛 = 𝝉𝒛𝒙 (2.20)

Thus, only six stress components are

independent, including three normal stress
components 𝝈𝒙𝒙 , 𝝈𝒚𝒚 , 𝝈𝒛𝒛 and three shear
stress components, say, 𝝉𝒛𝒚 , 𝝉𝒙𝒛 , 𝝉𝒙𝒚
İ. ÖZKOL 32
 2.4 EQUATIONS OF EQUILIBRIUM IN A
NONUNIFORM STRESS FIELD
Consider an infinitesimal element
∆𝑥 × ∆𝑦 × ∆𝑧 in which the stress field is
not uniform. The stress components acting
on the faces of the element are shown in
Fig. 2.9.

İ. ÖZKOL 33
İ. ÖZKOL 34
If the element is in equilibrium, then the six
equations of equilibrium, (2.18) and (2.19), must be
satisfied. Consider one of the equations of
equilibrium, say, σ 𝐹𝑥 = 0. We have

𝜎𝑥𝑥 + ∆𝜎𝑥𝑥 ∆𝑦∆𝑧 − 𝜎𝑥𝑥 ∆𝑦∆𝑧

+ 𝜏𝑦𝑥 + ∆𝜏𝑦𝑥 ∆𝑥∆𝑧 − 𝜏𝑦𝑥 ∆𝑥∆𝑧
+ 𝜏𝑧𝑥 + ∆𝜏𝑧𝑥 ∆𝑥∆𝑦 − 𝜏𝑧𝑥 ∆𝑥∆𝑦 = 0
Dividing the above equation with ∆𝑥∆𝑦∆𝑧, we
obtain
Δ𝜎𝑥𝑥 Δ𝜏𝑦𝑥 Δ𝜏𝑧𝑥
+ + =0 (2.21)
Δ𝑥 Δ𝑦 Δ𝑧
İ. ÖZKOL 35
Taking the limit ∆𝑥, ∆𝑦, ∆𝑧 → 0, the above
equilibrium equation becomes
𝜕𝜎𝑥𝑥 𝜕𝜏𝑦𝑥 𝜕𝜏𝑧𝑥
+ + =0 (2.21)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Similarly, equations σ 𝐹𝑦 = 0 ve σ 𝐹𝑧 = 0 lead
to
𝜕𝜏𝑥𝑦 𝜕𝜎𝑦𝑦 𝜕𝜏𝑧𝑦
+ + =0 (2.22)
𝜕𝑥 𝜕𝑦 𝜕𝑧
and
𝜕𝜏𝑥𝑧 𝜕𝜏𝑦𝑧 𝜕𝜎𝑧𝑧
+ + =0 (2.23)
𝜕𝑥 𝜕𝑦 𝜕𝑧
İ. ÖZKOL 36
It can easily be verified that the moment equations
෍ 𝑀𝑥 = ෍ 𝑀𝑦 = ෍ 𝑀𝑧 = 0

𝝉𝒙𝒚 = 𝝉𝒚𝒙 , 𝝉𝒚𝒛 = 𝝉𝒛𝒚 , 𝝉𝒙𝒛 = 𝝉𝒛𝒙 (2.24)

which are identical to the relation given by
(2.20). Equations (2.21)-(2.24) are the
equilibrium equations of a point in a body.
If a body is in equilibrium, the stress field
must satisfy these equations everywhere
in the body. İ. ÖZKOL 37
 2.5 PRINCIPAL STRESSES
Consider a 2-D state of stress so that
𝜎𝑥𝑥 ≠ 0, 𝜎𝑦𝑦 ≠ 0, 𝜏𝑥𝑦 ≠ 0 and other
stress components vanish. Consider the
wedge shown in Fig. 2.10. The unit normal
vector to the inclined surface is n, and the
stress vector (force per unit area) acting
on this surface is t.

İ. ÖZKOL 38
İ. ÖZKOL 39
From the equilibrium equations
σ 𝐹𝑥 = 0 𝑎𝑛𝑑 σ 𝐹𝑦 = 0 for the wedge,
we obtain
𝑡𝑥 ∆𝑠 = 𝜎𝑥𝑥 ∆𝑦 + 𝜏𝑦𝑥 ∆𝑥
𝑡𝑦 ∆𝑠 = 𝜏𝑥𝑦 ∆𝑦 + 𝜎𝑦𝑦 ∆𝑥 (2.25)
By noting
Δ𝑦
= cos 𝜃 = 𝑛𝑥
Δ𝑠
∆𝑥
= sin 𝜃 = 𝑛𝑦 (2.26)
Δ𝑆 İ. ÖZKOL 40
(2.25) can be expressed in the form
𝑡𝑥 = 𝜎𝑥𝑥 𝑛𝑥 + 𝜏𝑦𝑥 𝑛𝑦
𝑡𝑦 = 𝜏𝑥𝑦 𝑛𝑥 + 𝜎𝑦𝑦 𝑛𝑦 (2.27)
Equations (2.27) can be expressed in
matrix form as
𝜎𝑥𝑥 𝜏𝑦𝑥 𝑛 𝑡𝑥
𝜏𝑥𝑦 𝜎𝑦𝑦 𝑛𝑦 = 𝑡𝑦
𝑥
(2.28)
where the relation 𝜏𝑦𝑥 = 𝜏𝑥𝑦 has been
invoked.
İ. ÖZKOL 41
Using the same method, one can easily derive
the equations for the 3-D case with the result
𝝈𝒙𝒙 𝝉𝒙𝒚 𝝉𝒙𝒛 𝒏𝒙 𝒕𝒙
𝝉𝒙𝒚 𝝈𝒚𝒚 𝝉𝒚𝒛 𝒏𝒚 = 𝒕𝒚 (2.29)
𝝉𝒙𝒛 𝝉𝒚𝒛 𝝈𝒛𝒛 𝒏𝒛 𝒕𝒛
Symbolically, (2.29) can be written as
𝝈 𝒏 = 𝒕 (2.30)

İ. ÖZKOL 42
Equation (2.30) indicates that the stress
matrix 𝜎 can be viewed as a
transformation matrix that transforms the
unit normal vector {n} into the stress
vector {t}, which acts on the surface with
the unit normal {n}.
If we are interested in finding surfaces for
which t is parallel to n, i.e.,
𝒕 =𝝈 𝒏 (2.31)
İ. ÖZKOL 43
where 𝜎 is a scalar, then (2.29) yields a typical
eigenvalue problem:
𝜎 𝑛 = 𝑡 =𝜎 𝑛 (2.32)
( 𝜎 −𝜎 𝐼 ) 𝑛 =0 (2.33)
where 𝐼 is the identity matrix. In order for (2.33) to
have a nontrivial solution for {n}, we require
𝜎 − 𝜎 𝐼 = 0 Or
𝝈𝒙𝒙 − 𝝈 𝝉𝒙𝒚 𝝉𝒙𝒛
𝝉𝒙𝒚 𝝈𝒚𝒚 − 𝝈 𝝉𝒚𝒛 =0 (2.34)
𝝉𝒙𝒛 𝝉𝒚𝒛 𝝈𝒛𝒛 − 𝝈
İ. ÖZKOL 44
Expanding the determinantal equation
(2.34) yields a cubic equation in 𝜎. Since
𝜎 is real and symmetric, there are three
real roots, say 𝜎1 , 𝜎2 , 𝜎3 (see any book on
linear algebra or matrix theory for the
proof). The corresponding
eigenvectors, {n(1)},{n(2)},{n(3)}, can be
shown to be mutually orthogonal.

İ. ÖZKOL 45
These three directions are called principal
directions of stress, and 𝝈𝟏 , 𝝈𝟐 , 𝝈𝟑 are the
corresponding principal stresses.
On the surfaces perpendicular to these
directions, we have

𝒊 𝒊
𝒕 = 𝝈𝒊 𝒏 , 𝑖 = 1,2,3 (2.35)

İ. ÖZKOL 46
This says that on the surface with the unit
normal 𝑛 𝑖 , the stress vector is also normal
to that surface, and its magnitude is 𝜎𝑖 . In other
words, there are no shearing stresses on these
surfaces.
Without loss of generality, we assume
𝜎1 ≥ 𝜎2 ≥ 𝜎3 . Then it can be shown that 𝝈𝟏
and 𝝈𝟑 are the maximum and minimum
normal stresses, respectively, on all surfaces at
a point.
İ. ÖZKOL 47
The proof is given as follows.
On an arbitrary surface with unit normal
vector {n}, let the stress vector be {t}. The
normal component (projection) of {t} on
{n} is given by
𝜎𝑛 = 𝒕. 𝒏 = 𝑡 𝑇 𝑛
𝑇
= 𝜎 𝑛 𝑛
= 𝑛 𝑇 𝜎𝑇 𝑛 = 𝑛 𝑇
𝜎 𝑛 (2.36)

İ. ÖZKOL 48
Choose a coordinate system so that x, y,
and z axes are parallel to the principal
directions of stress, respectively. Then 𝜎
has a simple form as

𝜎1 0 0
𝜎 = 0 𝜎2 0 (2.37)
0 0 𝜎3

İ. ÖZKOL 49
Substituting (2.37) into (2.36), we obtain
𝜎𝑛 = 𝜎1 𝑛𝑥 2 + 𝜎2 𝑛𝑦 2 + 𝜎3 𝑛𝑧 2 (2.38)
If 𝜎1 ≥ 𝜎2 ≥ 𝜎3 , then we have
𝝈𝟏 𝒏𝒙 𝟐 + 𝝈𝟏 𝒏𝒚 𝟐 + 𝝈 𝟏 𝒏𝒛 𝟐 ≥ 𝝈𝒏 ≥ 𝝈𝟑 𝒏𝒙 𝟐 + 𝝈𝟑 𝒏 𝒚 𝟐 + 𝝈𝟑 𝒏𝒛 𝟐
2 2 2
Since 𝑛𝑥 + 𝑛𝑦 + 𝑛𝑧 = 1
(n is a unit vector), it is obvious that
𝜎1 ≥ 𝜎𝑛 ≥ 𝜎3 (2.39)

İ. ÖZKOL 50
Example 2.2 Given the stress matrix
2 2 0
𝜎 = 2 2 0
0 0 1
the eigenvalue problem is
2−𝜎 2 0
2 2−𝜎 0 =0
0 0 1−𝜎
which can be expanded into
𝜎 3 − 5𝜎 2 + 4𝜎 = 0
The three roots for the above equation are obtained as
𝜎1 = 4, 𝜎2 = 1, 𝜎3 = 0
These are the principal stresses
İ. ÖZKOL 51
 1
The unit normal 𝑛 corresponding to
𝜎1 =4 can be obtained by substituting
this value back into the system of
equations (2.33) to obtain

−2𝑛𝑥 1 + 2𝑛𝑦 1 = 0
2𝑛𝑥 1 − 2𝑛𝑦 1 = 0
1
−3𝑛𝑧 =0
İ. ÖZKOL 52
Note that these three equations are not
independent. Thus, only two equations
are available to determine the solution.
Since there are three unknowns, two
equations can determine only up to the
ratios among the three quantities
1 1 1
𝑛𝑥 , 𝑛𝑦 𝑎𝑛𝑑 𝑛𝑧

İ. ÖZKOL 53
 1
However, we note that 𝑛 is a unit vector,
i.e.,
1 2 1 2 1 2
𝑛𝑥 + 𝑛𝑦 + 𝑛𝑧 =1
With this additional equation, the solution
is uniquely obtained as
1 1
𝑛𝑥 1 = , 𝑛𝑦 1 = , 𝑛𝑧 1 = 0
2 2

İ. ÖZKOL 54
Following similar manipulations, the unit
vectors n(2) and n(3) corresponding to 𝜎2
and 𝜎3 , respectively, can be determined.
We have
2 2 2
𝑛𝑥 = 0, 𝑛𝑦 = 0 , 𝑛𝑧 = 1
3 1 3 1 3
𝑛𝑥 = , 𝑛𝑦 =− , 𝑛𝑧 =0
2 2
It is easy to verify that these three
eigenvectors are mutually orthogonal.

İ. ÖZKOL 55
 2.5.1 Shear Stress
The stress vector t can be decomposed into a
normal vector (𝝈𝒏 𝒏) and a tangential vector (𝝉)
that is lying on the surface with the unit normal 𝒏,
i.e.,
𝒕 = 𝝈𝒏 𝒏 + 𝝉 (2.40)
Thus
𝝉 = 𝒕 − 𝝈𝒏 𝒏 (2.41)
Denoting the magnitude of the shear stress vector
by 𝜏, we have
𝜏 2 = 𝑡 2 − 𝜎𝑛 2 = 𝑡𝑥 2 + 𝑡𝑦 2 + 𝑡𝑧 2 − 𝜎𝑛 2 (2.42)
İ. ÖZKOL 56
Let us choose the coordinate system (x,y,z) to be
parallel to the principal directions with
corresponding principal stresses 𝜎1 , 𝜎2 𝑎𝑛𝑑 𝜎3 ,
respectively. With respect to this coordinate system,
the stress matrix [𝜎] assumes the diagonal form as in
(2.37). From (2.29) we obtain
𝑡𝑥 = 𝜎1 𝑛𝑥
𝑡𝑦 = 𝜎2 𝑛𝑦 (2.43)
𝑡𝑧 = 𝜎3 𝑛𝑧
Using (2.43), the magnitude of the stress vector can
be written as
𝑡 2 = 𝑡𝑥 2 + 𝑡𝑦 2 + 𝑡𝑧 2
= (𝜎1 𝑛𝑥 )2 +(𝜎2 𝑛𝑦 )2 +(𝜎3 𝑛𝑧 )2
İ. ÖZKOL
(2.44)
57
Substituting (2.44) and (2.38) into (2.42), we obtain
2 2 2
𝜏2 2 2 2
= 𝜎1 𝑛𝑥 + 𝜎2 𝑛𝑦 + 𝜎3 𝑛𝑧 − 𝜎1 𝑛𝑥 + 𝜎2 𝑛𝑦 + 𝜎3 𝑛𝑧 2

𝜏 2 = 𝑛𝑥 2 1 − 𝑛𝑥 2 𝜎1 2 + 𝑛𝑦 2 1 − 𝑛𝑦 2 𝜎2 2 + 𝑛𝑧 2 1 − 𝑛𝑧 2 𝜎3 2
−2𝜎1 𝜎2 𝑛𝑥 2 𝑛𝑦 2 − 2𝜎2 𝜎3 𝑛𝑦 2 𝑛𝑧 2 − 2𝜎1 𝜎3 𝑛𝑥 2 𝑛𝑧 2 (2.45)
Equation (2.45) can be further simplified by using the
relation
1 − 𝑛𝑥 2 = 𝑛𝑦 2 + 𝑛𝑧 2 .

İ. ÖZKOL 58
We obtain
𝜏 2 = 𝑛𝑥 2 𝑛𝑦 2 (𝜎1 − 𝜎2 )2 +𝑛𝑦 2 𝑛𝑧 2 (𝜎2 − 𝜎3 )2 +𝑛𝑥 2 𝑛𝑧 2 (𝜎3 − 𝜎1 )2 (2.46)
Consider all surfaces that contain the y-axis, namely
surfaces with unit normal vector perpendicular to
the y-axis. For any of these surfaces, we have
𝑛𝑥 ≠ 0, 𝑛𝑦 = 0, 𝑛𝑧 ≠ 0 (2.47)
The magnitude of the shear stress is
𝜏 2 = 𝑛𝑥 2 𝑛𝑧 2 (𝜎3 − 𝜎1 )2 = 1 − 𝑛𝑥 2 𝑛𝑥 2 (𝜎3 − 𝜎1 )2 (2.48)
In deriving (2.48), the equation 𝑛𝑥 2 + 𝑛𝑧 2 = 1 has
been used.
The extremum of 𝜏 occurs at
𝜕 𝜏2
= 0 = 2𝑛𝑥 − 4𝑛𝑥 3 (𝜎3 − 𝜎1 )2 (2.49)
𝜕𝑛𝑥 İ. ÖZKOL 59
This yields the solutions
1
𝑛𝑥 = 0 𝑎𝑛𝑑 𝑛𝑥 = ± (2.50)
2
It can easily be shown that 𝑛𝑥 = 0 leads to the
2 1
minimum value of τ and 𝑛𝑥 = ± yields the
2
maximum shear stress.
2 2
Since 𝑛𝑦 = 0 , and 𝑛𝑥 + 𝑛𝑧 = 1, the solution
1 1
𝑛𝑥 = ± gives 𝑛𝑧 = ± .
2 2
These represent two surfaces making, respectively,
+45 and -45 with respect to the x-axis.
İ. ÖZKOL 60
 1
Substituting 𝑛𝑥 = ± into (2.48), the
2
maximum shear stress is obtained as
1
𝜏𝑚𝑎𝑥 = (𝜎3 − 𝜎1 )2
2
4
1
𝜏𝑚𝑎𝑥 = 𝜎3 − 𝜎1 (2.51)
2
Similar considerations of surfaces containing
the x-axis and the z-axis, respectively, yield
1
𝜏𝑚𝑎𝑥 = 𝜎2 − 𝜎3 (2.52)
2
1
𝜏𝑚𝑎𝑥 = 𝜎1 − 𝜎2 (2.53)
2 İ. ÖZKOL 61
Among these three shear stresses
given by (2.51)-(2.53), the one given
by (2.51) is the true maximum shear
stress if we assume, with no loss of
generality,
𝜎1 ≥ 𝜎2 ≥ 𝜎3

İ. ÖZKOL 62
2.5.2 Revisit of Transformation of Stress
Consider a state of plane stress
𝜎𝑧𝑧 = 𝜏𝑥𝑧 = 𝜏𝑦𝑧 = 0 𝑎𝑛𝑑 𝜎𝑥𝑥 ≠ 0,
′ ′
𝜎𝑦𝑦 ≠ 0 𝑎𝑛𝑑 𝜏𝑥𝑦 ≠ 0 . Let 𝑥 − 𝑦 be
coordinate axes obtained by rotating x-y
axes 𝜃 angle in the counterclockwise
direction; see Fig. 2.11.

İ. ÖZKOL 63
İ. ÖZKOL 64
Consider the surface perpendicular to the 𝑥 ′ -axis. On
this surface, the stress vector t is given by
𝑡𝑥 𝜎𝑥𝑥 𝜏𝑥𝑦 𝑛𝑥
𝑡𝑦 = 𝜏 𝜎 𝑛 (2.54)
𝑥𝑦 𝑦𝑦 𝑦

where 𝒏 = 𝑛𝑥 , 𝑛𝑦 is the unit vector parallel to

the 𝑥 ′ -axis. Let 𝜎𝑥𝑥 ′ , 𝜎𝑦𝑦 ′ 𝑎𝑛𝑑 𝜏𝑥𝑦 ′ be the stress
components in reference
′ ′
to the 𝑥 − 𝑦 coordinates.
′
Noting that 𝜎𝑥𝑥 = 𝜎𝑛 , we have

İ. ÖZKOL 65
 ′
𝜎𝑥𝑥 = 𝜎𝑛 = 𝒕. 𝒏 = 𝑡𝑥 𝑛𝑥 + 𝑡𝑦 𝑛𝑦 (2.55)
Substituting the following relations into (2.55) yields
𝑡𝑥 = 𝜎𝑥𝑥 𝑛𝑥 + 𝜏𝑥𝑦 𝑛𝑦
𝑡𝑦 = 𝜏𝑥𝑦 𝑛𝑥 + 𝜎𝑦𝑦 𝑛𝑦
𝜎𝑥𝑥 ′ = 𝜎𝑥𝑥 𝑛𝑥 2 + 𝜏𝑥𝑦 𝑛𝑥 𝑛𝑦 + 𝜏𝑥𝑦 𝑛𝑥 𝑛𝑦 + 𝜎𝑦𝑦 𝑛𝑦 2
= 𝜎𝑥𝑥 𝑛𝑥 2 + 2𝜏𝑥𝑦 𝑛𝑥 𝑛𝑦 + 𝜎𝑦𝑦 𝑛𝑦 2 (2.56)
Noting 𝑛𝑥 = cos 𝜃 𝑎𝑛𝑑 𝑛𝑦 = sin 𝜃, we rewrite
(2.56) in the form
𝜎𝑥𝑥 ′ = 𝜎𝑥𝑥 𝑐𝑜𝑠 2 𝜃 + 𝜎𝑦𝑦 𝑠𝑖𝑛2 𝜃 + 𝜏𝑥𝑦 sin 2𝜃 (2.57)

İ. ÖZKOL 66
Further use of in (2.57) leads to
2 1 2 1
𝑐𝑜𝑠 𝜃 = 1 + cos 2𝜃 , 𝑠𝑖𝑛 𝜃 = 1 − cos 2𝜃
2 2
′ 1 1
𝜎𝑥𝑥 = 𝜎𝑥𝑥 + 𝜎𝑦𝑦 + 𝜎𝑥𝑥 − 𝜎𝑦𝑦 cos 2𝜃 +
2 2
𝜏𝑥𝑦 sin 2𝜃 (2.58)
The shear stress component 𝜏𝑥𝑦 ′ can be regarded as
the tangential component of the stress vector, i.e.,
𝜏𝑥𝑦 ′ = 𝒕. 𝒔 = 𝑡𝑥 𝑠𝑥 + 𝑡𝑦 𝑠𝑦
where s is the unit vector parallel to the 𝑦 ′ -axis (or
parallel to the surface of interest; see Fig. 2.11), and
its two components are
𝑠𝑥 = − sin 𝜃 𝑎𝑛𝑑 𝑠𝑦 = cos 𝜃
İ. ÖZKOL 67
Thus,
𝜏𝑥𝑦 ′ = 𝜎𝑥𝑥 𝑛𝑥 + 𝜏𝑥𝑦 𝑛𝑦 𝑠𝑥 + 𝜏𝑥𝑦 𝑛𝑥 + 𝜎𝑦𝑦 𝑛𝑦 𝑠𝑦
= −𝜎𝑥𝑥 cos 𝜃 sin 𝜃 + 𝜏𝑥𝑦 𝑐𝑜𝑠 2 𝜃 − 𝑠𝑖𝑛2 𝜃 +
𝜎𝑦𝑦 cos 𝜃 sin 𝜃
′ 1
𝜏𝑥𝑦 = 𝜎𝑦𝑦 − 𝜎𝑥𝑥 sin 2𝜃 + 𝜏𝑥𝑦 cos 2𝜃 (2.59)
2

İ. ÖZKOL 68
The transformation for 𝜎𝑦𝑦 ′ is obtained by
noting that 𝜎𝑥𝑥 ′ is equal to 𝜎𝑦𝑦 ′ if 𝜃 is
replaced by 𝜃 + 𝜋Τ2 . From (2.57) we have
𝝈𝒚𝒚 ′ = 𝝈𝒙𝒙 𝒔𝒊𝒏𝟐 𝜽 + 𝝈𝒚𝒚 𝒄𝒐𝒔𝟐 𝜽 − 𝝉𝒙𝒚 𝐬𝐢𝐧 𝟐𝜽 2.60)
Equations (2.58)-(2.60) are the 2-D stress
coordinate transformation formulas which
can be expressed in matrix form

İ. ÖZKOL 69
as
𝝈𝒙𝒙 ′ 𝒄𝒐𝒔𝟐 𝜽 𝒔𝒊𝒏𝟐 𝜽 𝐬𝐢𝐧 𝟐𝜽 𝝈𝒙𝒙
𝝈𝒚𝒚 ′ = 𝒔𝒊𝒏𝟐 𝜽 𝒄𝒐𝒔𝟐 𝜽 − 𝐬𝐢𝐧 𝟐𝜽 𝝈𝒚𝒚 (2.61)
𝟏 𝟏
𝝉𝒙𝒚 ′ − 𝐬𝐢𝐧 𝟐𝜽 𝐬𝐢𝐧 𝟐𝜽 𝐜𝐨𝐬 𝟐𝜽 𝝉𝒙𝒚
𝟐 𝟐
where 𝜎𝑥𝑥 ′ , 𝜎𝑦𝑦 ′ ve 𝜏𝑥𝑦 ′ are stress components
′ ′
referring to the 𝑥 −𝑦 −coordinate system.

İ. ÖZKOL 70
 2.6 LINEAR STRESS-STRAIN RELATIONS
2.6.1 Strains Induced by Normal Stress
Imagine that a thin cylindrical element (a rod) is
taken out of a solid in the x-direction. Assume
that in this rod, only 𝜎𝑥𝑥 ≠ 0, and all the other
five stress components are vanishing (uniaxial
stress). If stress 𝜎𝑥𝑥 is applied gradually, strains
are produced.
Let us consider for now only the normal strain
component 𝜀𝑥𝑥 . A stress- strain curve as shown
in Fig. 2.12 can be obtained.
İ. ÖZKOL 71
For metals, such a uniaxial stress-strain
curve in general consists of a linear
portion (up to 𝜎𝑥𝑥 = 𝜎𝛾 ) and a nonlinear
portion (𝜎𝑥𝑥 > 𝜎𝛾 ). For the linear portion,
we have
𝜎𝑥𝑥 = 𝐸𝑥 𝜀𝑥𝑥 (2.62)
in which the constant 𝐸𝑥 is called the
Young's modulus in the x-direction of the
solid
İ. ÖZKOL 72
İ. ÖZKOL 73
In the above case of uniaxial stress, lateral
strains 𝜀𝑦𝑦 and 𝜀𝑧𝑧 are also present and are
related to 𝜀𝑥𝑥 as
𝜀𝑦𝑦
𝜀𝑦𝑦 = −𝜈𝑥𝑦 𝜀𝑥𝑥 𝑜𝑟 𝜈𝑥𝑦 = − (2.63)
𝜀𝑥𝑥

𝜀𝑧𝑧
𝜀𝑧𝑧 = −𝜈𝑥𝑧 𝜀𝑥𝑥 𝑜𝑟 𝜈𝑥𝑧 = − (2.64)
𝜀𝑥𝑥
where 𝜈𝑥𝑦 and 𝜈𝑥𝑧 are Poisson's ratios.
The first subscript indicates the loading
direction, and the second subscript indicates
the direction of lateral contraction.
İ. ÖZKOL 74
Using (2.62); (2.63) and (2.64) can be rewritten
as
𝜈𝑥𝑦
𝜀𝑦𝑦 = − 𝜎𝑥𝑥 (2.65)
𝐸𝑥
𝜈𝑥𝑧
𝜀𝑧𝑧 = − 𝜎𝑥𝑥 (2.66)
𝐸𝑥
Similar Young's moduli in the y-direction and z-
direction are introduced in the uniaxial stress-
strain relations as
𝜎𝑦𝑦 = 𝐸𝑦 𝜀𝑦𝑦 (2.67)
𝜎𝑧𝑧 = 𝐸𝑧 𝜀𝑧𝑧 İ. ÖZKOL
(2.68) 75
The corresponding Poisson's ratios are
introduced in the following relations:
𝜈𝑦𝑥
𝜀𝑥𝑥 = −𝜈𝑦𝑥 𝜀𝑦𝑦 = − 𝜎𝑦𝑦
𝐸𝑦
𝜈𝑦𝑧
𝜀𝑧𝑧 = −𝜈𝑦𝑧 𝜀𝑦𝑦 = − 𝜎𝑦𝑦 (2.69)
𝐸𝑦
and
𝜈𝑧𝑥
𝜀𝑥𝑥 = −𝜈𝑧𝑥 𝜀𝑧𝑧 = − 𝜎𝑧𝑧
𝐸𝑧
𝜈𝑧𝑦
𝜀𝑦𝑦 = −𝜈𝑧𝑦 𝜀𝑧𝑧 = − 𝜎𝑧𝑧 (2.70)
𝐸 𝑧
İ. ÖZKOL 76
It should be noted that in a uniaxially stressed
body, say 𝜎𝑥𝑥 ≠ 0 and other 𝜎𝑖𝑗 = 0, shear
strains 𝛾𝑦𝑧 , 𝛾𝑥𝑧 𝑎𝑛𝑑 𝛾𝑥𝑦 may be induced in
addition to the normal strains. In the most
general case, uniaxial stress 𝜎𝑥𝑥 may produce
six strain components,
1 𝜈𝑥𝑦 𝜈𝑥𝑧
𝜀𝑥𝑥 = 𝜎𝑥𝑥 , 𝜀𝑦𝑦 = − 𝜎𝑥𝑥 , 𝜀𝑧𝑧 = − 𝜎𝑥𝑥
𝐸𝑥 𝐸𝑥 𝐸𝑥
𝜂𝑥,𝑦𝑧 𝜂𝑥,𝑥𝑧 𝜂𝑥,𝑥𝑦
𝛾𝑦𝑧 = 𝜎𝑥𝑥 , 𝛾𝑥𝑧 = 𝜎𝑥𝑥 , 𝛾𝑥𝑦 = 𝜎𝑥𝑥
𝐸𝑥 𝐸𝑥 𝐸𝑥
(2.71)
İ. ÖZKOL 77
where the coefficients 𝜂′ s serve a similar
purpose as Poisson's ratios; they are the ratios
of the induced shear strains and the normal
strain 𝜀𝑥𝑥 produced by uniaxial stress 𝜎𝑥𝑥 , i.e.,

𝛾𝑦𝑧 𝛾𝑥𝑧 𝛾𝑥𝑦

𝜂𝑥,𝑦𝑧 = , 𝜂𝑥,𝑥𝑧 = , 𝜂𝑥,𝑥𝑦 =
𝜀𝑥𝑥 𝜀𝑥𝑥 𝜀𝑥𝑥
Again, the first subscript in 𝜂 indicates the
loading direction, and the second set of
subscripts indicates the plane of induced shear
strain. In a similar manner, the strains
produced by uniaxial stresses 𝜎𝑦𝑦 and 𝜎𝑧𝑧 are
İ. ÖZKOL 78
 𝜈𝑦𝑥 1 𝜈𝑦𝑧
𝜀𝑥𝑥 = − 𝜎𝑦𝑦 𝜀𝑦𝑦 = 𝜎𝑦𝑦 𝜀𝑧𝑧 = − 𝜎𝑦𝑦
𝐸𝑦 𝐸𝑦 𝐸𝑦

𝜂𝑦,𝑦𝑧 𝜂𝑦,𝑥𝑧 𝜂𝑦,𝑥𝑦

𝛾𝑦𝑧 = 𝜎𝑦𝑦 𝛾𝑥𝑧 = 𝜎𝑦𝑦 𝛾𝑥𝑦 = 𝜎𝑦𝑦 2.72)
𝐸𝑦 𝐸𝑦 𝐸𝑦
And
𝜈𝑧𝑥 𝜈𝑧𝑦 1
𝜀𝑥𝑥 = − 𝜎𝑧𝑧 𝜀𝑦𝑦 = − 𝜎𝑧𝑧 𝜀𝑧𝑧 = 𝜎𝑧𝑧
𝐸𝑧 𝐸𝑧 𝐸𝑧

𝜂𝑧,𝑦𝑧 𝜂𝑧,𝑥𝑧 𝜂𝑧,𝑥𝑦

𝛾𝑦𝑧 = 𝜎𝑧𝑧 𝛾𝑥𝑧 = 𝜎𝑧𝑧 𝛾𝑥𝑦 = 𝜎𝑧𝑧 2.73)
𝐸𝑧 𝐸𝑧 𝐸𝑧

If all three normal stress components are present,

then the total strains are the sums of the
corresponding strains given by (2.71)-(2.73).
İ. ÖZKOL 79
 2.6.2 Strains Induced by Shear Stress
Consider a state of simple shear with 𝜏𝑥𝑦 ≠ 0
and all other stress components are vanishing.
In the most general solid, all strain components
may be induced by 𝜏𝑥𝑦 . Consider the shear
strain 𝛾𝑥𝑦 induced by 𝜏𝑥𝑦 . In the linear range of
stress-strain relations, we have
1
𝛾𝑥𝑦 = 𝜏𝑥𝑦 (2.74)
𝐺𝑥𝑦
where 𝐺𝑥𝑦 is the shear modulus in the x-y plane.
İ. ÖZKOL 80
Other strains induced by 𝜏𝑥𝑦 can be written as
𝜂𝑥𝑦,𝑥
𝜀𝑥𝑥 = 𝜂𝑥𝑦,𝑥 𝛾𝑥𝑦 = 𝜏𝑥𝑦
𝐺𝑥𝑦
𝜂𝑥𝑦,𝑦
𝜀𝑦𝑦 = 𝜂𝑥𝑦,𝑦 𝛾𝑥𝑦 = 𝜏𝑥𝑦
𝐺𝑥𝑦
𝜂𝑥𝑦,𝑧
𝜀𝑧𝑧 = 𝜂𝑥𝑦,𝑧 𝛾𝑥𝑦 = 𝜏𝑥𝑦 (2.75)
𝐺𝑥𝑦
𝜂𝑥𝑦,𝑦𝑧
𝛾𝑦𝑧 = 𝜂𝑥𝑦,𝑦𝑧 𝛾𝑥𝑦 = 𝜏𝑥𝑦
𝐺𝑥𝑦
𝜂𝑥𝑦,𝑥𝑧
𝛾𝑥𝑧 = 𝜂𝑥𝑦,𝑥𝑧 𝛾𝑥𝑦 = 𝜏𝑥𝑦
𝐺𝑥𝑦
İ. ÖZKOL 81
in which 𝜂𝑥𝑦,𝑦𝑧 and 𝜂𝑥𝑦,𝑥𝑧 are introduced
to represent the interactions among the
shear strains;
and 𝜂𝑥𝑦,𝑥 , 𝜂𝑥𝑦,𝑦 𝑎𝑛𝑑 𝜂𝑥𝑦,𝑧 are the
interactions between the shear strain 𝛾𝑥𝑦
and the normal strains 𝜀𝑥𝑥 , 𝜀𝑦𝑦 𝑎𝑛𝑑 𝜀𝑧𝑧 ,
respectively.

İ. ÖZKOL 82
For simple shear in the y-z and x-z planes, we have
1
𝛾𝑦𝑧 = 𝜏𝑦𝑧 (2.76)
𝐺𝑦𝑧
1
𝛾𝑥𝑧 = 𝜏𝑥𝑧 (2.77)
𝐺𝑥𝑧
Other strains produced by 𝜏𝑦𝑧 and 𝜏𝑥𝑧 , can be
expressed in a form similar to (2.75). If the material is
isotropic, i.e., its mechanical properties are not direction
dependent, all the 𝜼 coefficients vanish and
𝐸𝑥 = 𝐸𝑦 = 𝐸𝑧 = 𝐸
𝜈𝑥𝑦 = 𝜈𝑦𝑥 = 𝜈𝑥𝑧 = 𝜈𝑧𝑥 = 𝜈𝑦𝑥 = 𝜈𝑧𝑦 = 𝜈 (2.78)
𝐺𝑥𝑦 = 𝐺𝑥𝑧 = 𝐺𝑦𝑧 = 𝐺 İ. ÖZKOL 83
 2.6.3 3-D Stress-Strain Relations
The discussion in Sections 2.6.1 and 2.6.2 indicates
that the application of a single stress component can
possibly produce all six strain components. In the
linear range of stress-strain relations, we can write
the strains produced by all six stress components by
using the principle of superposition:

𝜀𝑥𝑥 = 𝑎11 𝜎𝑥𝑥 + 𝑎12 𝜎𝑦𝑦 + 𝑎13 𝜎𝑧𝑧 + 𝑎14 𝜏𝑦𝑧 + 𝑎15 𝜏𝑥𝑧 + 𝑎16 𝜏𝑥𝑦
𝜀𝑦𝑦 = 𝑎21 𝜎𝑥𝑥 + 𝑎22 𝜎𝑦𝑦 + 𝑎23 𝜎𝑧𝑧 + 𝑎24 𝜏𝑦𝑧 + 𝑎25 𝜏𝑥𝑧 + 𝑎26 𝜏𝑥𝑦
𝜀𝑧𝑧 = 𝑎31 𝜎𝑥𝑥 + 𝑎32 𝜎𝑦𝑦 + 𝑎33 𝜎𝑧𝑧 + 𝑎34 𝜏𝑦𝑧 + 𝑎35 𝜏𝑥𝑧 + 𝑎36 𝜏𝑥𝑦
𝛾𝑦𝑧 = 𝑎41 𝜎𝑥𝑥 + 𝑎42 𝜎𝑦𝑦 + 𝑎43 𝜎𝑧𝑧 + 𝑎44 𝜏𝑦𝑧 + 𝑎45 𝜏𝑥𝑧 + 𝑎46 𝜏𝑥𝑦
𝛾𝑥𝑧 = 𝑎51 𝜎𝑥𝑥 + 𝑎52 𝜎𝑦𝑦 + 𝑎53 𝜎𝑧𝑧 + 𝑎54 𝜏𝑦𝑧 + 𝑎55 𝜏𝑥𝑧 + 𝑎56 𝜏𝑥𝑦
𝛾𝑥𝑦 = 𝑎61 𝜎𝑥𝑥 + 𝑎62 𝜎𝑦𝑦 + 𝑎63 𝜎𝑧𝑧 + 𝑎64 𝜏𝑦𝑧 + 𝑎65 𝜏𝑥𝑧 + 𝑎66 𝜏𝑥𝑦
(2.79)
İ. ÖZKOL 84
where 𝒂𝒊𝒋 (i,j=1-6) are elastic compliances(compatibility).
Comparing (2.79) with (2.71)-(2.75) together
with two similar equations for 𝜏𝑦𝑧 and 𝜏𝑥𝑧 , we
can easily relate the elastic compliances to the
engineering moduli 𝑬𝒙 , 𝑬𝒚 , . . . ,𝝂𝒙𝒚 , . . . , 𝑮𝒙𝒚 , .
. . , 𝜼𝒙𝒚,𝒛 , . . .. For example,
1 𝜈𝑦𝑥 𝜈𝑧𝑥
𝑎11 = , 𝑎12 = − , 𝑎13 = −
𝐸𝑥 𝐸𝑦 𝐸𝑧
𝜂𝑦𝑧,𝑥 𝜂𝑥𝑧,𝑥 𝜂𝑥𝑦,𝑥
𝑎14 = , 𝑎15 = , 𝑎16 = (2.80)
𝐺𝑦𝑧 𝐺𝑥𝑧 𝐺𝑥𝑦

İ. ÖZKOL 85
In matrix notation, (2.79) can be expressed as
𝜀𝑥𝑥 𝜎𝑥𝑥
𝜀𝑦𝑦 𝜎𝑦𝑦
𝜀𝑧𝑧 𝜎𝑧𝑧
𝛾𝑦𝑧 = 𝑎𝑖𝑗 𝜏𝑦𝑧 (2.81)
𝛾𝑥𝑧 𝜏𝑥𝑧
𝛾𝑥𝑦 𝜏𝑥𝑦
or symbolically as
𝜀 = 𝑎 𝜎 (2.82)
The inverse relations of (2.82) are given as
𝜎 = 𝑐 𝜀 (2.83)
where
𝒄 = 𝒂 −𝟏 İ. ÖZKOL 86
The elements 𝒄𝒊𝒋 in [c] are called elastic
constants while 𝑬𝒙 , 𝑬𝒚 , . . . ,𝝂𝒙𝒚 . . . , 𝑮𝒙𝒚 , .
. . are called engineering moduli.
From the elastic strain energy consideration,
it can be shown that [a], and thus [c] also,
is a symmetric matrix, i.e.
𝑎𝑖𝑗 = 𝑎𝑗𝑖 𝑖, 𝑗 = 1,2, … . . , 6

İ. ÖZKOL 87
Thus, there are only 21 independent
elastic compliances or, equivalently, 21
elastic constants. This is the maximum
number of independent elastic constants
that linear elastic materials can have.
Most materials possess certain elastic
symmetries that would reduce the
number of independent elastic constants.
The following are two material groups
that are of practical interest.
İ. ÖZKOL 88
Orthotropic Materials
Unidirectional fiber composites can be
regarded as orthotropic materials which
possess three mutually orthogonal planes
of symmetry. The directions perpendicular
to these planes are called the material
principal directions. If the coordinates x, y,
z are set up such that they are parallel to
the material principal directions,
respectively, then the elastic compliance
matrix reduces to İ. ÖZKOL 89
By definition, an orthotropic material has at least 2
orthogonal planes of symmetry, where material properties
are independent of direction within each plane. Such
materials require 9 independent variables (i.e. elastic
constants) in their constitutive matrices. In contrast, a
material without any planes of symmetry is fully anisotropic
and requires 21 elastic constants, whereas a material with an
infinite number of symmetry planes (i.e. every plane is a
plane of symmetry) is isotropic, and requires only 2 elastic
constants.

İ. ÖZKOL 90
A familiar example of an orthotropic material is wood.
In wood, one can define three mutually perpendicular
directions at each point in which the properties are
different. These are the axial direction (along the grain),
the radial direction, and the circumferential direction
İ. ÖZKOL 91
 𝑎11 𝑎12 𝑎13 0 0 0
𝑎21 𝑎22 𝑎23 0 0 0
𝑎31 𝑎32 𝑎33 0 0 0
𝑎𝑖𝑗 = 𝑎44 0 0 (2.84)
0 0 0
0 0 0 0 𝑎55 0
0 0 0 0 0 𝑎66

The elastic compliances 𝑎𝑖𝑗 are related to the engineering moduli as

1 𝜈𝑦𝑥 𝜈𝑧𝑥
𝑎11 = , 𝑎12 = − , 𝑎13 = −
𝐸𝑥 𝐸𝑦 𝐸𝑧
𝜈𝑥𝑦 1 𝜈𝑧𝑦
𝑎21 = − , 𝑎22 = , 𝑎23 = −
𝐸𝑥 𝐸𝑦 𝐸𝑧
𝜈 𝜈𝑦𝑧 1
𝑎31 = − 𝑥𝑧 , 𝑎32 = − , 𝑎33 =
𝐸𝑥 𝐸𝑦 𝐸𝑧
1 1 1
𝑎44 = , 𝑎55 = , 𝑎66 = (2.85)
𝐺𝑦𝑧 𝐺𝑥𝑧 𝐺𝑥𝑦

İ. ÖZKOL 92
Since 𝑎𝑖𝑗 = 𝑎𝑗𝑖 , we have
𝐸𝑦 𝐸𝑧 𝐸𝑧
𝜈𝑦𝑥 = 𝜈𝑥𝑦 , 𝜈𝑧𝑥 = 𝜈𝑥𝑧 , 𝜈𝑧𝑦 = 𝜈𝑦𝑧 (2.86)
𝐸𝑥 𝐸𝑥 𝐸𝑦
Thus, there are nine independent elastic constants
for orthotropic elastic materials.

İ. ÖZKOL 93
Fiber-reinforced composites are regarded
as orthotropic solids. It is customary to
denote the fiber direction as X1 -axis, and
the transverse directions as X2 and X3 . The
elastic moduli are referenced to this
particular coordinate system and denoted
by 𝐸1 , 𝐸2 , 𝐸3 , 𝜈12, 𝜈13 , 𝜈23 , 𝐺23 , 𝐺13 𝑎𝑛𝑑 𝐺12
The following are the in-plane engineering
moduli for some polymeric composites.
İ. ÖZKOL 94
İ. ÖZKOL 95
Note that the Young's modulus in the fiber direction (𝐸1 )
for AS4/3501-6 is 14 times the transverse Young's
modulus (𝐸2 ).
Isotropic Materials The elastic properties of
isotropic materials are invariant with respect to
directions. Thus, isotropy is a special case of orthotropy.
By requiring the conditions given by (2.78), we obtain
1
𝑎11 = 𝑎22 = 𝑎33 =
𝐸
𝜈
𝑎12 = 𝑎13 = 𝑎23 = − (2.87)
𝐸
1
𝑎44 = 𝑎55 = 𝑎66 =
𝐺
İ. ÖZKOL 96
The corresponding elastic constants 𝑐𝑖𝑗 can also be
expressed in terms of engineering moduli as

𝑐11 = 𝑐22 = 𝑐33 = 𝜆 + 2𝐺

𝑐12 = 𝑐13 = 𝑐23 = 𝜆 (2.88)
𝑐44 = 𝑐55 = 𝑎66 = 𝐺

It is evident that the stress-strain relations for

isotropic materials can be expressed in terms of the
Young's modulus E, Poisson's ratio 𝜈 , and shear
modulus 𝐺 . Moreover, it can be shown that these
three quantities are related by
𝐸 =2 1+𝜈 𝐺 İ. ÖZKOL
(2.89) 97
Thus, there are only two independent
elastic constants for isotropic materials.
Aluminum alloys are usually considered
isotropic materials. Typical values of their
elastic moduli are
E=70Gpa, 𝜈 = 0.33

İ. ÖZKOL 98
 2.7 ELASTIC STRAIN ENERGY
An elastic body can store energy in the
form of deformation. This strain energy is
completely released when loads are
removed. Since the strain energy is stored
solely in the form of deformation, it can be
expressed in terms of strain components or
stress components.

İ. ÖZKOL 99
İ. ÖZKOL 100
On the x-face, only the normal stress 𝜎𝑥𝑥 is present. The
corresponding normal strain is 𝜀𝑥𝑥 . The total force acting on
the x-face is ∆𝑦∆𝑧𝜎𝑥𝑥 and the elongation of the element in
the x-direction is ∆𝑥𝜀𝑥𝑥 . The work done by 𝜎𝑥𝑥 is converted
entirely to strain energy. The strain energy stored in the
element is
1 1
∆𝑈 = ∆𝑦∆𝑧𝜎𝑥𝑥 (∆𝑥𝜀𝑥𝑥 ) = (∆𝑉𝜀𝑥𝑥 𝜎𝑥𝑥 ) (2.90)
2 2
where ∆𝑉 is the volume of the element. The 1/2 factor in
(2.90) accounts for the linear stress-strain relations of the
elastic solid. If all three normal stress components are
present, then the strain energy stored in the element is
1
∆𝑈 = ∆𝑉(𝜎𝑥𝑥 𝜀𝑥𝑥 +𝜎𝑦𝑦 𝜀𝑦𝑦 + 𝜎𝑧𝑧 𝜀𝑧𝑧 ) (2.91)
2
İ. ÖZKOL 101
If the infinitesimal element is subjected to the shear
stress 𝜏𝑥𝑧 as shown in Fig. 2.14, the work done (and
thus the strain energy) is

1 1
∆𝑈 = ∆𝑥∆𝑦𝜏𝑥𝑧 𝛾𝑥𝑧 ∆𝑧 = (∆𝑉𝜏𝑥𝑧 𝛾𝑥𝑧 ) (2.92)
2 2

İ. ÖZKOL 102
where the first part on the right-hand side represents the
total force acting on the top face, and the second part
represents the distance the shear force travels during the
deformation. The total strain energy produced by all three
shear stress components is
1
∆𝑈 = ∆𝑉 𝜏𝑥𝑧 𝛾𝑥𝑧 + 𝜏𝑥𝑦 𝛾𝑥𝑦 + 𝜏𝑦𝑧 𝛾𝑦𝑧 (2.93)
2
The total strain energy stored in ∆𝑉 due to normal
and shear stresses is the sum of (2.91) and (2.93).
The strain energy density W is defined as
Δ𝑈
𝑊=
Δ𝑉
1
= (𝜎𝑥𝑥 𝜀𝑥𝑥 +𝜎𝑦𝑦 𝜀𝑦𝑦 + 𝜎𝑧𝑧 𝜀𝑧𝑧 + 𝜏𝑥𝑧 𝛾𝑥𝑧 + 𝜏𝑥𝑦 𝛾𝑥𝑦 +
2
𝜏𝑦𝑧 𝛾𝑦𝑧 ) (2.94)
İ. ÖZKOL 103
Using the matrix representations of the stress-strain
relations (2.82) and (2.83), the strain energy density
can be expressed in various forms as
1 𝑇 1 𝑇 1 𝑇 1 𝑇
𝑊= 𝜎 𝜀 = 𝜀 𝜎 = 𝜀 𝑐 𝜀 = 𝜎 𝑎 𝜎 (2.95)
2 2 2 2

Since W > 0, the above quadratic forms imply(express )

that [c] and [a] must be positive definite. Consequently,
all the diagonal terms 𝑐𝑖𝑖 (𝑖 = 1 − 6) and 𝑎𝑖𝑖 (𝑖 = 1 − 6)
must be positive. For orthotropic materials, this
implies that all the major engineering moduli 𝐸𝑥 , 𝐸𝑦 ,
𝐸𝑧 , 𝐺𝑦𝑧 , 𝐺𝑥𝑧 𝑎𝑛𝑑 𝐺𝑥𝑦 cannot assume negative
values. İ. ÖZKOL 104
 2.8 PLANE ELASTICITY
Many structures under certain types of loading
may yield strain and stress fields that are
independent of, say, the z-direction and have
𝜀𝑧𝑧 = 𝛾𝑦𝑧 = 𝛾𝑥𝑧 = 0 (2.96)
Or
𝜎𝑧𝑧 = 𝜏𝑥𝑧 = 𝜏𝑦𝑧 = 0 (2.97)
over the entire structure. The state of
deformation with (2.96) is called a state of
plane strain parallel to the x-y plane, while
that with (2.97) is called a state of plane stress
parallel to the x-y plane.
İ. ÖZKOL 105
From the strain-displacement relations, the conditions of
(2.96) can be described in terms of the displacement field as
𝑢 = 𝑢(𝑥, 𝑦)
𝑣 = 𝑣(𝑥, 𝑦)
𝑤=0 (2.98)
where u, v, and w are displacement components in
the x, y and z directions, respectively. To produce a
state of plane strain parallel to the x-y plane, the
structure must be uniform in shape in the z-direction,
and loading must be independent of the z-axis. An
example is a hollow cylinder subjected to uniform
internal pressure with both ends constrained to
suppress its movement in the z-direction.
İ. ÖZKOL 106
The plane stress condition given by (2.97), in
general, cannot be exactly realized. It is
often used to approximate the state of stress
in a thin panel subjected to only in-plane
(x-y plane) loading. The corresponding
displacement field is given by
𝑢 = 𝑢(𝑥, 𝑦)
𝑣 = 𝑣(𝑥, 𝑦) (2.99)
𝑤 = 𝜀𝑧𝑧 𝑧
where normal strain 𝜀𝑧𝑧 is independent of the
z-axis.
İ. ÖZKOL 107
 2.8.1 Stress-Strain Relations for Plane
Isotropic Solids
Plane strain and plane stress conditions lead to
simplification in stress-strain relations. Recall that the
3-D stress-strain relations for isotropic solids can be
expressed in the following forms
𝜎 = 𝑐 𝜀 𝒄𝒊𝒋 : elastic constants (2.100)
or
𝜀 = 𝑎 𝜎 𝒂𝒊𝒋 : elastic compliances (2.101)

where
𝐸(1 − 𝑣)
𝑐11 = 𝑐22 = 𝑐33 = 𝜆 + 2𝐺 =
1 + 𝑣 1 − 2𝑣
𝐸𝑣
𝑐12 = 𝑐21 = 𝑐13 = 𝑐31 = 𝑐23 = 𝑐32 =𝜆=
1 + 𝑣 1 − 2𝑣
İ. ÖZKOL 108
𝑐44 = 𝑐55 = 𝑐66 = 𝐺
Other 𝑐𝑖𝑗 = 0
1
𝑎11 = 𝑎22 = 𝑎33 =
𝐸
𝑣
𝑎12 = 𝑎21 = 𝑎13 = 𝑎31 = 𝑎23 = 𝑎32 = −
𝐸
1
𝑎44 = 𝑎55 = 𝑎66 =
𝐺
Other 𝑎𝑖𝑗 = 0

İ. ÖZKOL 109
Plane Strain Using the plane strain conditions,
(2.100) reduces to
𝐸
𝜎𝑥𝑥 = 1 − 𝑣 𝜀𝑥𝑥 + 𝑣𝜀𝑦𝑦
(1+𝑣)(1−2𝑣)
𝐸
𝜎𝑦𝑦 = [𝑣𝜀𝑥𝑥 + (1 − 𝑣)𝜀𝑦𝑦 ] (2.102)
(1+𝑣)(1−2𝑣)
𝑣𝐸
𝜎𝑧𝑧 = [𝜀𝑥𝑥 + 𝜀𝑦𝑦 ]
(1+𝑣)(1−2𝑣)
𝜏𝑦𝑧 = 0
𝜏𝑥𝑧 = 0
𝜏𝑥𝑦 = 𝐺𝛾𝑥𝑦

İ. ÖZKOL 110
It is easy to see that
𝜎𝑧𝑧 = 𝑣(𝜎𝑥𝑥 + 𝜎𝑦𝑦 )
Thus, 𝜎𝑧𝑧 is a dependent quantity. The remaining
nontrivial equations in (2.102) are usually
considered the stress-strain relations for a state of
plane strain. Inverting these equations, we obtain
1+𝑣
𝜀𝑥𝑥 = [ 1 − 𝑣 𝜎𝑥𝑥 − 𝑣𝜎𝑦𝑦 ]
𝐸
1+𝑣
𝜀𝑦𝑦 = [ 1 − 𝑣 𝜎𝑥𝑥 − 𝑣𝜎𝑦𝑦 ] (2.103)
𝐸
1
𝛾𝑥𝑦 = 𝜏𝑥𝑦
𝐺

İ. ÖZKOL 111
Plane Stress The reduced stress-strain relations for a
state of plane stress are readily derived from the 3-D
relations given by (2.101). We have
1
𝜀𝑥𝑥 = [𝜎𝑥𝑥 − 𝑣𝜎𝑦𝑦 ]
𝐸
1
𝜀𝑦𝑦 = [−𝑣𝜎𝑥𝑥 + 𝜎𝑦𝑦 ]
𝐸
𝑣
𝜀𝑧𝑧 = − [𝜎𝑥𝑥 + 𝜎𝑦𝑦 ] (2.104)
𝐸
𝛾𝑦𝑧 = 0
𝛾𝑥𝑧 = 0
1
𝛾𝑥𝑦 = 𝜏𝑥𝑦
𝐺
İ. ÖZKOL 112
Again, from (2.104) we note that
𝑣
𝜀𝑧𝑧 = − [𝜀𝑥𝑥 + 𝜀𝑦𝑦 ] (2.105)
1−𝑣
is a dependent quantity. Inverting the first two
and last equations in (2.104),we obtain
𝐸
𝜎𝑥𝑥 = 2
(𝜀𝑥𝑥 + 𝑣𝜀𝑦𝑦 )
1−𝑣
𝐸
𝜎𝑦𝑦 = 2 (𝑣𝜀𝑥𝑥 + 𝜀𝑦𝑦 ) (2.106)
1−𝑣
𝜏𝑥𝑦 = 𝐺𝛾𝑥𝑦

İ. ÖZKOL 113
These relations are different from the
corresponding plane strain relations given
by (2.102). However, the two sets of
stress-strain relations can be put in a
single expression by introducing the
following parameter defined by
for plane strain: 𝜅 = 3 − 4𝑣
3−𝑣
for plane stress: 𝜅 =
1+𝑣

İ. ÖZKOL 114
In terms of parameter 𝜅 , the stress-strain relations
for both plane strain and plane stress can be
expressed in the form:
3−𝜅
𝜎𝑥𝑥 = 2𝐺 𝜀𝑥𝑥 + (𝜀𝑥𝑥 + 𝜀𝑦𝑦 )
2𝜅−2
3−𝜅
𝜎𝑦𝑦 = 2𝐺 𝜀𝑦𝑦 + (𝜀𝑥𝑥 + 𝜀𝑦𝑦 )
2𝜅−2
𝜏𝑥𝑦 = 𝐺𝛾𝑥𝑦 (2.107)
or
1 3−𝜅
𝜀𝑥𝑥 = 𝜎𝑥𝑥 − (𝜎𝑥𝑥 + 𝜎𝑦𝑦 )
2𝐺 4
1 3−𝜅
𝜀𝑦𝑦 = 𝜎𝑦𝑦 − (𝜎𝑥𝑥 + 𝜎𝑦𝑦 )
2𝐺 4
1
𝛾𝑥𝑦 = 𝜏𝑥𝑦 İ. ÖZKOL (2.108)115
𝐺
 2.8.2 Stress-Strain Relations for
Orthotropic Solids
Fiber-reinforced composite materials are modeled as
orthotropic solids. Furthermore, they are often used
in the form of thin panels for which the plane stress
condition prevails. Let the fiber direction coincide
with the x-axis and the panel be parallel to the x-y
plane. The stress-strain relations for a composite
panel are given by
1 𝑣21
− 0
𝜀𝑥𝑥 𝐸1 𝐸2 𝜎𝑥𝑥
𝑣12 1
𝜀𝑦𝑦 = − 0 𝜎𝑦𝑦 (2.109)
𝐸1 𝐸2
𝛾𝑥𝑦 1
𝜏𝑥𝑦
0 0
𝐺
İ. ÖZKOL 12 116
where E1 is the Young's modulus in the fiber direction
(the longitudinal modulus), E2 is the transverse
Young's modulus, G12 is the shear modulus in the x-y
plane, and 𝑣12 and 𝑣21 are Poisson's ratios. Since the
compliance matrix is symmetric, we have
𝑣21 𝑣12
= (2.110)
𝐸2 𝐸1
Thus, there are only four independent elastic moduli.
Inverting relations (2.109), we have
𝐸1 𝑣12 𝐸1
𝜎𝑥𝑥 0 𝜀𝑥𝑥
∆ ∆
𝜎𝑦𝑦 = 𝑣12 𝐸1 𝐸2 𝜀𝑦𝑦 (2.111)
0
𝜏𝑥𝑦 ∆ ∆ 𝛾𝑥𝑦
0 0 𝐺12
İ. ÖZKOL 117
∆= 1 − 𝑣12 𝑣21
 2.8.3 Governing Equations
To solve plane elasticity problems, three sets of
conditions must be satisfied, i.e., the equilibrium
equations, boundary conditions, and compatibility
equations.
Equilibrium Equations For plane problems, it is easy
to show that the 3-D equilibrium equations (2.21)-
(2.23) reduce to
𝜕𝜎𝑥𝑥 𝜕𝜏𝑥𝑦
+ =0
𝜕𝑥 𝜕𝑦
(2.112)
𝜕𝜏𝑥𝑦 𝜕𝜎𝑦𝑦
+ =0
𝜕𝑥 𝜕𝑦 İ. ÖZKOL 118
Boundary Conditions For plane problems, the
loading stress vector t on the boundary is in
the x-y plane, i.e., tz=0. On the boundary
contour, the applied traction ( stress vector) is
given and the stresses must satisfy the
following boundary conditions:
𝑡𝑥 𝜎𝑥𝑥 𝜏𝑥𝑦 𝑛𝑥
𝑡𝑦 = 𝜏 𝜎 𝑛 (2.113
𝑥𝑦 𝑦𝑦 𝑦
where n=(nx,ny) is the unit normal vector to
the boundary contour of the plane body.
İ. ÖZKOL 119
Compatibility Equation The three strain components
𝜀𝑥𝑥 , 𝜀𝑦𝑦 𝑎𝑛𝑑 𝛾𝑥𝑦 are derived from the two
displacement components u and v as
𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑣
𝜀𝑥𝑥 = , 𝜀𝑦𝑦 = , 𝛾𝑥𝑦 = + (2.114)
𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥
Using the strain-displacement relations in (2.114), we
can derive the following compatibility equation:
𝜕2 𝜀𝑥𝑥 𝜕2 𝜀𝑦𝑦 𝜕2 𝛾𝑥𝑦
+ = (2.115)
𝜕𝑦 2 𝜕𝑥 2 𝜕𝑥𝜕𝑦
Thus, the three strain functions 𝜀𝑥𝑥 , 𝜀𝑦𝑦 𝑎𝑛𝑑 𝛾𝑥𝑦 cannot
be arbitrarily specified; they must satisfy the
compatibility equation (2.115). Otherwise, we may not
be able to find a unique displacement field.
İ. ÖZKOL 120
For isotropic solids, the compatibility equation
(2.115) can be written in terms of stresses by
using the stress-strain relations and equilibrium
equations. We have
2
𝛻 𝜎𝑥𝑥 + 𝜎𝑦𝑦 = 0 (2.116)
where
2 2
𝜕 𝜕
𝛻2 = 2 + 2
𝜕𝑥 𝜕𝑦

İ. ÖZKOL 121
 2.8.4 Solution by Airy Stress Function for
Plane
Isotropic Solids
Consider the possibility of the existence of a
function Ø (x,y) such that
2
𝜕 Φ
𝜎𝑥𝑥 =
𝜕𝑦 2
𝜕2Φ
𝜎𝑦𝑦 =
𝜕𝑥 2
𝜕2 Φ
𝜏𝑥𝑦 = − (2.117)
𝜕𝑥𝜕𝑦 İ. ÖZKOL 122
With the relations given by (2.117), it is easy to verify
that the equilibrium equations (2.112) are satisfied
automatically. Substitution of (2.117) into (2.116)
yields
𝛻2𝛻2Φ = 0
Or, explicitly

𝜕4 Φ 𝜕4 Φ 𝜕4 Φ
+ 2 2 2 + =0 (2.118)
𝜕𝑥 4 𝜕𝑥 𝜕𝑦 𝜕𝑦 4

The above is the compatibility equation in terms of Ø

which is called the Airy stress function. In this form, the
solution to a plane elasticity problem is reduced to
solving (2.118) for Ø from which stresses are derived
from (2.117).
İ. ÖZKOL 123
These stresses are required to satisfy the
boundary conditions. Note that, in using
the Airy stress function, one needs not
worry about equilibrium equations since
they are automatically satisfied. The
solution procedure involves solving the
partial differential equation (2.118) and
satisfying the boundary conditions (2.113).

İ. ÖZKOL 124