4/21/2011

Theory of Elasticity in Geotechnical

Engineering

Lecture

Hadush Seged (Dr.)

April 21, 2011

1

Analysis of Stress in 3D (continued)

Previously we discussed that:
9 The traction (or stress) vector T at a point:
Ti n = Ti (1) n1 + Ti ( 2 ) n2 + Ti ( 3) n3
Ti (1) = [σ 11 σ 12 σ 13 ] ,etc…
⇒ Ti (1) n1 = [σ 11 σ 12 σ 13 ] [n1 ] ,etc…
T

⎡σ 11 σ 12 σ 13 ⎤
T
⎡n1 ⎤
⇒ Ti = T = ⎢⎢σ 21 σ 22 σ 23 ⎥⎥
n ⎢n ⎥ = σ T n
⎢ 2⎥
⎢⎣σ 31 σ 32 σ 33 ⎥⎦ ⎢⎣n3 ⎥⎦
2

1
 4/21/2011

T
⎡σ x τ xy τ xz ⎤ ⎡n1 ⎤
⎢ ⎥ ⎢n ⎥ = σ T n T = σ T n …...(2.1)
⇒ Ti n = T = ⎢τ yx σ y τ yz ⎥ ⎢ 2⎥
⎢τ τ σ ⎥ ⎢⎣n3 ⎥⎦ Cauchy’s formula
⎣ zx zy z ⎦

Stress notations:
σ x σ y σ z τ xy τ xz τ yx τ yz τ zx τ zy
σ xx σ yy σ zz σ xy σ xz σ yx σ yz σ zx σ zy
σ 11 σ 22 σ 33 σ 12 σ 13 σ 21 σ 23 σ 31 σ 32

9 The stress tensor (or matrix) σ is very important and

contains
t i allll the
th information
i f ti needed
d d to
t find
fi d the
th traction
t ti on
every surface that passes through the point in question.

9 By knowing the nine components of σ (where only six are

independent) we can completely determine the infinitely
many tractions that act on the infinitude of surfaces
passing
p g through
g anyyp
point.

9 Cauchy’s discovery created long jam of ideas in mechanics.

Before, researchers had continually struggled with the
description of stress.
4

2
 4/21/2011

Index notation and Summation convention

9 Index notation is a shorthand scheme whereby a whole set of
numbers (elements or components) is represented by a single
symbol with subscripts.

9 For example, the three numbers a1, a2, a3 are denoted by the
symbol ai, where index i will normally have the range 1, 2, 3.

9 In a similar fashion, aij represents the nine numbers a11, a12, a13,
a21, a22, a23, a31, a32, a33.

⎡a1 ⎤ ⎡a11 a12 a13 ⎤

ai = ⎢⎢a2 ⎥⎥ aij = ⎢⎢a21 a22 a23 ⎥⎥
⎢⎣a3 ⎥⎦ ⎢⎣a31 a32 a33 ⎥⎦
5

9 From the consideration of equilibrium of forces acting on

arbitrary
bit small
ll volume
l off material,
t i l the
th componentst off the
th stress
t
vector Ti in the x-coordinate direction can be written as:
3
Ti = ∑ σ ji n j …...(2.2)
j =1

9 Since the subscript i is unspecified, equation (2.2) holds for each

of the three ppossible values corresponding
p g to x,, y, z coordinates.
Index i is therefore called free index.

9 However, index j is called a dummy index because the particular

letter used in this subscript is not important.

3
 4/21/2011

9 Thus: 3 3 3
Ti = ∑ σ ji n j = ∑ σ ki nk = ∑ σ mi nm …...(2.3)
(2 3)
j =1 k =1 m =1
Summation Convention:
9 When a subscript appears twice in the same term then
summation over that subscript from 1 to 3 is implied.
3
9 For
F example:
l aii = ∑ aii = a11 + a22 + a33
i =1
3
aij b j = ∑ aij b j = ai1b1 + ai 2b2 + ai 3b3
j =1
7

free if it appears once.

9 Index =
dummy if it appears twice.

9 An index can only be free or dummy.

4
 4/21/2011

9 Equation (2.3) can be abbreviated as: Ti = σ ji n j …...(2.4)

9 F
From theh moment equilibrium
ilib i
of an element:
τ xy =τ yx τ xz =τ zx τ yz =τ zy
σ xy =σ yx σ xz =σ zx σ yz =σ zy
σ12 =σ 21 σ13 =σ 31 σ 23 =σ 32
σ ji = σ ij …...(2.5)
(2 5)
9 The stress tensor σij is hence symmetric. ⇒ σ = σT
9 Equation (2.4) can be re-written more conveniently and
conventionally as:
Ti = σ ij n j …...(2.6)
9

Normal and Shear Stress Components

x3
9 The stress vector Tin acting on the
surface area element at a point can be
x2
resolved into normal stress and shear
stress components. x1

Normal stress: ⇒ σ n ni
Shear stress: ⇒ S i
9 The magnitude of the normal stress is
clearly Tini. Surface element
Components of traction
⇒ σ n = Ti ni = σ ij n j ni …...(2.7) normal and tangential to
the surface

10

5
 4/21/2011

9 The magnitude of the shear stress is given by:

⇒ Sn = (Si Si ) = (T T − σ ) …...(2.8)
i i
2
n

9 Using equations (2.6) and (2.7), equation (2.8) becomes:

⇒ S n = σ ijσ ik n j nk − (σ ij n j ni ) 2 …...(2.9)

9 Equations (2.7) and (2.9) are used to determine the normal and
shear stress components of a traction vector Ti acting on plane ni.

9 Equations (1.3), (2.7) and (2.9) are all called Cauchy’s equations.

11

9 The vector σnni is in the direction of the normal vector ni, and the
vector Si, lies in the plane formed by the two vectors Ti and ni, or

S i = Ti − σ n ni …...(2.10)
Example 2
The state of stress at a point is represented by the given stress
tensor:
⎡3 4 0⎤
σ ij = ⎢⎢4 9 0⎥⎥
⎢⎣0 0 3 ⎥⎦
For a plane with unit normal ni = (1 2 , 1 2 , 1
2
) , calculate the
normal and shear stress components.
12

6
 4/21/2011

Principal Axes and Stresses

9 For any general state of stress at any point P in a body, there
i t three
exist th mutually
t ll perpendicular
di l planes
l i t P on which
att point hi h the
th
shear stresses vanish. The normal stress components on these
three planes are called principal stresses. The corresponding
three planes are called principal planes.

9 This condition implies that the stress (traction) vector Ti is

perpendicular
p p to the area element and is p parallel to the area
normal ni, that is Ti = λni, where λ is a scalar. Or using equation
(2.6):
σ ij n j = λni …...(2.11)
13

Which can be written as:

σ ij n j − λni = 0 …...(2.12)
Noting that i and j are dummy indexes equation (2.12) can be re-
written as:
σ ij n j − λδ ij n j = (σ ij − λδ ij )n j …...(2.13)
Where δij is called the Kronecker delta.
delta
⎡1 0 0⎤
⎧= 1 if i = j ⇒ δ ij = ⎢⎢0 1 0⎥⎥ …...(2.14)
δ ij = ⎨
⎩= 0 if i ≠ j ⎢⎣0 0 1⎥⎦
14

7
 4/21/2011

9 Equation (2.13) can be written in the expanded matrix form as:

⎡σ 11 − λ σ 12 σ 13 ⎤ ⎡n1 ⎤
⎢σ σ 22 − λ σ 23 ⎥⎥ ⎢n ⎥ = 0
⎢ 21 ⎢ 2⎥ …...(2.15)
⎢⎣σ 31 σ 32 σ 33 − λ ⎥⎦ ⎢⎣n3 ⎥⎦

9 It should be noted that the following relation holds for the unit
normal ni, and its components n1, n2, and n3 which are also called
direction cosines:

ni ni = n1n1 + n2 n2 + n3 n3 = n12 + n22 + n32 =1 …...(2.16)

15

Direction Cosines
y
y
Ay A
β
A 3D γ
2D Ay β α
Az x
α Ax
x
Ax z
Ax
Ax = cos α = l = n1
= cos α = l = n1 A + Ay2 + Az2
2
x
Ax2 + Ay2 Ay
= cos β = m = n2
Ay Ax2 + Ay2 + Az2
= cos β = m = n2
Ax2 + Ay2 Az
= cos γ = n = n3
A + Ay2 + Az2
2
x

16

8
 4/21/2011

9 Since equation (2.16) has to be satisfied, all three components

n1, n2, and n3 can not be zero.
zero

9 Therefore, equation (2.15) possesses solution if and only if the

determinant of the coefficient matrix is zero.

det σ ij − λδ ij = 0 …...(2.17)
9 or:
σ 11 − λ σ 12 σ 13
σ 21 σ 22 − λ σ 23 = 0 …...(2.18)
σ 31 σ 32 σ 33 − λ
17

9 Expanding equation (2.17) leads to the following cubic equation,

which is known as the characteristic equation :

λ3 − I1λ2 + I 2 λ − I 3 = 0 …...(2.18)
9 where,
I1 = σ ii = σ 11 + σ 22 + σ 33 = trσ …...(2.19)

σ 11 σ 12 σ 11 σ 13 σ 22 σ 23
I2 = + +
σ 21 σ 22 σ 31 σ 33 σ 32 σ 33 …...(2.20)
= σ 11σ 22 + σ 11σ 33 + σ 22σ 33 − σ 122 − σ 132 − σ 23
2

18

9
 4/21/2011

σ 11 σ 12 σ 13
I 3 = σ 21 σ 22 σ 23 = det σ …...(2.21)
σ 31 σ 32 σ 33
9 Equation (2.18) has three real roots. These three values are the
three principal stresses σ 1 , σ 2 , σ 3 .

9 If the coordinate system

y aligns
g with the p
principal
p directions, the
stress tensor takes the form:
⎡σ 1 0 0 ⎤
σ = σ ij = ⎢⎢0 σ 2 0 ⎥⎥ …...(2.22)
⎢⎣0 0 σ 3 ⎥⎦
19

9 in which, I1 = σ 1 + σ 2 + σ 3 …...(2.23)
I 2 = σ 1σ 2 + σ 1σ 3 + σ 2σ 3…...(2.24)
I 3 = σ 1σ 2σ 3…...(2.25)
9 the largest principal stress is called major principal stress, and
the smallest minor principal stress. The other value is called
intermediate principal stress. Conventionally, σ 1 ≥ σ 2 ≥ σ 3 .
9 iin special
i l cases, two or all
ll the
h three
h principal
i i l stresses may be
b
numerically equal. The stress condition where all principal
stresses are equal is known as a state of hydrostatic stress,
since it is the condition that exists in a fluid in static
equilibrium.
20

10
 4/21/2011

Stress Invariants

9 The magnitudes and directions of σ 1 , σ 2 , σ 3 depend only on the

l d being
loads b i applied
li d and d cannot be
b influenced
i fl b the
d by h choice
h i off
coordinate axes (x, y, z) used to specify the state of stress at a
point.

9 This means that I1, I2, and I3 given by eqn. (2.18) have the same
magnitudes for all choices of coordinate axes (x, y, z).

9 I1, I2, and I3 are called the first, second and third invariants of
the stress tensor σ ij . (Invariant means something that does not
change even when the coordinate system does.)

21

Another definition of invariants:

9 Equation 2.18 matches the form for the standard eigenvalue
problem in which the principal stresses σ 1 , σ 2 , σ 3 are the
eigenvalues of σij and the direction cosines (n1, n2, n3) define
the associated eigenvector.

9 Since by definition the eigenvalues of σij do not depend on

the choice of the base vector, therefore, the coefficients I1, I2,
and I3 of Eq. (2.18) will not depend on any particular choice of
basis. Hence, I1, I2, and I3 are called the invariants of σij.

22

11
 4/21/2011

Example 3

Find the principal stresses and the principal stress directions

for the given stress tensor:
⎡− 10 15 0 ⎤
σ ij = ⎢⎢ 15 30 0 ⎥⎥
⎢⎣ 0 0 0⎥⎦

23

12