Professional Documents
Culture Documents
By
Textbooks: 1. Advanced Mechanics of Materials; 4th Edition, A.P. Boresi and O.M.
Sidebottom, John Wiley & Sons, 1985
2. Advanced Mechanics of Materials; 2nd Edition, R.D. Cook and W.C. Young,
Prentice Hall, 1999
3. Theory of Elastic Stability; 2nd Edition, S.P. Timoshenko and J.M. Gere,
McGraw-Hill, 1963
4. Theory of Elasticity; 3rd Edition, S.P. Timoshenko and J.N. Goodier, McGraw-
Hill, 1970
5. Theory of Plates and Shells; 2nd Edition, S.P. Timoshenko and S. Woinowsky-
Krieger, McGraw-Hill, 1970
6. Mechanical Behavior of Materials; 2nd Edition, N.E. Dowling, Prentice Hall,
1999
7. Mechanics of Materials; 3th Edition, R.C. Hibbeler, Prentice Hall, 1997
Course Contents:
Chapters Topics
1 Theories of Stress and Strain
2 Stress-Strain Relations
3 Elements of Theory of Elasticity
4 Applications of Energy Methods
5 Static Failure and Failure Criteria
6 Fatigue
7 Introduction to Fracture Mechanics
8 Beams on Elastic Foundation
9 Plate Bending
10 Buckling and Instability
iii
4.5 Second Variation in the Total Potential Energy ................................................... 4-17
4.6 Rayleigh-Ritz Method ........................................................................................... 4-28
4.7 Introduction to Finite Element Method................................................................. 4-35
Chapter 5 Static Failure and Failure Criteria
5.1 Definition of Failure............................................................................................. 5-1
5.2 Modes of Failure ................................................................................................. 5-1
5.3 Failure Criteria .................................................................................................... 5-3
5.4 Comparison of the Failure Criteria ...................................................................... 5-11
Chapter 6 Introduction to Fracture Mechanics
6.1 Introduction ......................................................................................................... 6-1
6.2 Fracture Modes..................................................................................................... 6-2
6.3 Stress and Displacement Field at the Crack Tip .................................................. 6-3
6.4 Stress Intensity Factor (SIF or K ) ...................................................................... 6-7
6.5 Superposition of SIF ............................................................................................ 6-9
6.6 Fracture Toughness (Critical SIF)........................................................................ 6-9
6.7 Strain Energy Release Rate and Its Equivalent to SIF ........................................ 6-13
6.8 Plastic Zone Size ................................................................................................. 6-14
Chapter 7 Fatigue
7.1 Introduction ......................................................................................................... 7-1
7.2 Nomenclature ....................................................................................................... 7-1
7.3 Cyclic Stress-Strain Behavior of Metals ............................................................. 7-3
7.4 Cyclic Stress-Strain Curve .................................................................................. 7-6
7.5 S − N Diagram and Stress Life Relation ............................................................ 7-7
7.6 Fatigue Strength Diagram (Haigh diagram) ........................................................ 7-13
7.7 Endurance Limit Modifying Factor ..................................................................... 7-18
7.8 Fatigue Crack Propagation .................................................................................. 7-19
7.9 Factors Affecting the Fatigue Crack Growth ....................................................... 7-21
Chapter 8 Beams on Elastic Foundation
8.1 Introduction ......................................................................................................... 8-1
8.2 General Theory .................................................................................................... 8-1
8.3 Infinite Beam Subjected to Point Load ............................................................... 8-3
8.4 Beam Supported on Equally Spaced Separated Elastic Supports ....................... 8-8
8.5 Infinite Beam Subjected to a Distributed Load Segment .................................... 8-12
8.6 Semi-infinite beam Subjected to Loads at Its End ............................................... 8-17
iv
Chapter 9 Flat Plates
9.1 Introduction .......................................................................................................... 9-1
9.2 Assumptions and Limitations of Thin Plate with Small Deflection .................... 9-1
9.3 Force-Stress Relations.......................................................................................... 9-2
9.4 Equilibrium Equations ......................................................................................... 9-3
9.5 Kinetics: Strain-Displacement Relations ............................................................ 9-4
9.6 Stress-Strain Relations ......................................................................................... 9-5
9.7 Stress-Deflection Relations .................................................................................. 9-5
9.8 Governing Differential Equations ........................................................................ 9-5
9.9 Boundary Conditions............................................................................................ 9-7
9.10 Double Series Solution of Simply Supported Plates ......................................... 9-14
9.11 Single Series Solution of Simply Supported Plates ........................................... 9-22
Chapter 10 Buckling and Instability
10.1 Introduction ........................................................................................................ 10-1
10.2 Column Buckling ............................................................................................... 10-2
10.3 Plate Buckling ................................................................................................... 10-4
10.4 Plates with Various Boundary Conditions ....................................................... 10-11
10.5 Application of Buckling to Design Problems .................................................. 10-15
References
v
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-1
Chapter 1
Theories of Stress and Strain
1.1 Definition of Stress at a Point
Mechanics of materials is a branch of mechanics that studies
1.) The relationships between the external loads applied to a deformable body and
intensity of internal forces acting within the body.
2.) The deformation and the stability of the body under the action of the external
loads.
In this study, the body is made of the material that is continuous (consist of a
continuum or uniform distribution of matter having no voids) and cohesive (all portions of the
material are connected together without breaks, cracks, and separations).
The external loads that we are interested can be idealized as concentrated force,
surface force, and linear distributed load as shown in Fig. 1.1.
Fig. 1.1
Consider the body subjected to a loading condition as shown in Fig. 1.2a. Under the
action of these loads, the body is deformed and has internal forces to hold the body together.
At a given section, the internal forces are distributed as shown in Fig. 1.2b and we can find
the resultant of the internal forces at a given point O as shown in Fig. 1.2c.
The distribution of the internal forces at a given point on the sectioned area of the
body can be determined by using the concept of stress.
Stress describes the intensity of the internal force on a specific plane or area passing
through a point as shown in Fig. 1.3. It can be classified into two types based on its acting
directions: normal stress and shear stress. Since the stresses generally vary from point to
point, the definitions of stresses must relate to an infinitesimal element.
Normal stress or σ is the intensity of force that acts normal to the area ∆A . If the
normal force or stress pulls on the area element ∆A , it is referred to as tensile stress, whereas
if it pushes on the area ∆A , it is referred to as compressive stress. It can be defined as
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-2
Fig. 1.2
Fig. 1.3
∆Fn
σ = lim
∆A→0 ∆A
Shear stress or τ is the intensity of force that acts tangent to the area ∆A .
∆Ft
τ = lim
∆A→ 0 ∆A
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-3
Fig. 1.4
By passing the imaginary section through the body parallel to the x-y plane as shown
in Fig. 1.4, the stress on the element area ∆A = ∆x∆y can be resolved into stress components
in the rectangular orthogonal Cartesian coordinate axes x, y, and z as σ z , τ zx , and τ zy where
∆Fz
σ z = lim
∆A→ 0 ∆A
∆Fx
τ zx = lim
∆A→ 0 ∆A
∆Fy
τ zy = lim
∆A→0 ∆A
Notation
¾ The first subscript notation refers to the orientation of area ∆A , which is
perpendicular to the subscript notation.
¾ The second subscript notation refers to the direction line of the stress.
By passing the imaginary section through the body parallel to the x-z plane as shown
in Fig. 1.5, we obtain the stress components as σ y , τ yx , τ yz .
By passing the imaginary section through the body parallel to the y-z plane as shown
in Fig. 1.6, we obtain the stress components as σ x , τ xy , τ xz .
If we continue cut the body in this manner by using the corresponding parallel plane,
we will obtain a cubic volume element of material that represents the state of stress acting
around the chosen point on the body as shown in Fig. 1.7.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-4
Fig. 1.5
Fig. 1.6
In array form, we have
Fig. 1.7
1.3 Symmetry of the Stress Array and Stress on an Arbitrarily Oriented Plane
If the stress around the chosen point on the body is constant, some of the stress
components can be reduced by using the force and moment equilibrium for the element.
Normal Stress Components
For a constant state of stress as shown in Fig. 1.8, each of the three normal stress
components must be equal in magnitude, but opposite in direction.
∑F x = 0; σ x (∆y∆z ) − σ ′x (∆y∆z ) = 0
σ x = σ ′x
Similarly, we have
∑F y =0 σ y = σ ′y
∑F z = 0; σ z = σ ′z
Fig. 1.8
Shear Stress Components
For a constant state of stress as shown in Fig. 1.9, pairs of shear stresses on adjacent
faces of the element must have equal magnitude and be directed either toward or away from
the corners of the element.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-6
Fig. 1.9
τ yx = τ ′yx
τ xy = τ yx
Similarly, we have
τ yz = τ ′yz = τ zy = τ ′zy
τ xz = τ ′xz = τ zx = τ ′zx
Therefore, in matrix form,
σ x τ xy τ xz
σ ij = σy τ yz
Sym. σ z
Stress Acting on Arbitrary Plane
The stress vector on the plane that is perpendicular to the x -axis as shown in Fig. 1.10
can be written as
v
σ x = σ x iˆ + τ xy ˆj + τ xz kˆ
In a similar fashion, the stress vector on the planes that are perpendicular to the y -
and z -axes can be written as
v
σ y = τ yx iˆ + σ y ˆj + τ yz kˆ
v
σ z = τ zx iˆ + τ zy ˆj + σ z kˆ
Consider an arbitrary oblique plane P as shown in Fig. 1.11. Let the plane is defined
by a unit normal vector
r
N = liˆ + mˆj + nkˆ
where l = cos α , m = cos β , and n = cos γ . From Fig. 1.12, we have l 2 + n 2 = d 2 and
d 2 + m 2 = 1 . Thus,
l 2 + m2 + n2 = 1
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-7
Fig. 1.10
Fig. 1.11
Fig. 1.12
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-8
dA0 BC = ldAABC
dA0 AC = mdAABC
dA0 AB = ndAABC
Remark:
The area dA0 BC , dA0 AC , and dA0 AB are the projection of area dAABC on the respective
coordinate plane. Let consider the figure of the wedge shown below and compare the volumes
determined from two methods.
1. Let the area associated with the side AC be AAC and corresponding wedge height
AAC
l=
ACB
v
Thus, we can write the stress vector σ P on the oblique plane P by summing the force
vectors acting on the stress element as shown in Fig. 1.11.
v v v v
σ P (dAABC ) = σ x (dA0 BC ) + σ y (dA0 AC ) + σ z (dA0 AB )
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-9
v v v v
σ P = σ xl + σ y m + σ z n
v
σ P = (σ x iˆ + τ xy ˆj + τ xz kˆ)l +
(τ yx iˆ + σ y ˆj + τ yz kˆ)m +
(τ zx iˆ + τ zy ˆj + σ z kˆ)n
v
σ P = (σ x l + τ yx m + τ zx n)iˆ +
(τ xy l + σ y m + τ zy n) ˆj +
(τ xz l + τ yz m + σ z n)kˆ
v
Also, the projections of the stress vector σ P on the x , y , and z -axes may be written
as
v
σ P = σ Px iˆ + σ Py ˆj + σ Pz kˆ
v
Comparing the stress vector σ P , we have
σ Px = σ x l + τ yx m + τ zx n
σ Py = τ xy l + σ y m + τ zy n
σ Pz = τ xz l + τ yz m + σ z n
Normal Stress and Shearing Stress on an Oblique Plane
v
The normal stress on the plane P or σ PN is the projection of the stress vector σ P in
v
the direction of the unit normal vector N . Thus, the magnitude of the normal stress σ PN can
be determined from
v v
σ PN = σ P ⋅ N
= (σ Px iˆ + σ Py ˆj + σ Pz kˆ) ⋅ (liˆ + mˆj + nkˆ)
= σ Px l + σ Py m + σ Pz n
iˆ
v
N = N ˆj
ˆ
k
Then, we have
σ x τ xy τ xz l
σ PN = l m n τ xy σ y τ yz m
τ xz
τ yz σ z n
or,
σ PN = N [σ ]N T
Fig. 1.13
Also, form Fig. 1.13, the magnitude of the shearing stress on plane P or σ PS can be
σ PS = σ P2 − σ PN
2
= σ Px
2
+ σ Py
2
+ σ Pz
2
− σ PN
2
Of all the infinite number of planes passing through point O , there is a set of three
mutually perpendicular planes that the normal stress σ PN has a maximum value called the
principal planes. The maximum normal stress is called the maximum principal stress. On
these planes, the shearing stresses vanish. Also, the three mutually perpendicular axes that are
normal to the three planes are called principal axes.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-11
Fig. 1.14
Let ( x , y , z ) and ( x ′ , y ′ , z ′ ) denote two rectangular coordinate systems with a
common origin as shown in Fig. 1.14. Also, let a general point in space A has coordinate ( x ,
y , z ) and ( x ′ , y ′ , z ′ ) in the respective coordinate system. The direction cosines between
the coordinate axes ( x , y , z ) and ( x ′ , y ′ , z ′ ) can be determined by finding the coordinate
x ′ , y ′ , z ′ of the point A in term of the coordinate x , y , z .
v v
R = OA = xiˆ + yˆj + zkˆ
v
x ′ R ⋅ iˆ′ iˆ ⋅ iˆ′ ˆj ⋅ iˆ′ kˆ ⋅ iˆ′ x
v
y ′ = R ⋅ ˆj ′ = iˆ ⋅ ˆj ′ ˆj ⋅ ˆj ′ kˆ ⋅ ˆj ′ y
z ′ v ˆ iˆ ⋅ kˆ ′ ˆj ⋅ kˆ ′ kˆ ⋅ kˆ ′ z
R ⋅ k ′
x ′ l1 m1 n1 x
y ′ = l 2 m2 n2 y
z ′ l m3 n3 z
3
x′ x
y ′ = [T ] y
z′ z
where [T ] is called the transformation matrix.
As shown in Fig. 1.15, the stress components in the ( x , y , z ) coordinates are
σ x , σ y , σ z , τ xy , τ xz , and τ yz
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-12
Fig. 1.15
From the previous discussions, the normal stress component σ x′ is the components in
the x ′ direction of the stress vector on a plane perpendicular to the x ′ -axis. Thus, from the
σ x τ xy τ xz l
equation σ PN = l m n τ xy σ y τ yz m , we have
τ xz
τ yz σ z n
σ x τ xy τ xz l1
σ x′ = l1 m1 n1 τ xy σ y τ yz m1
τ xz
τ yz σ z n1
σ x τ xy τ xz l3
σ z′ = l3 m3 n3 τ xy σ y τ yz m3
τ xz
τ yz σ z n3
The shear stress component τ x′y′ is the components in the y ′ direction of the stress
σ x τ xy τ xz l 2
τ x′y′ = l1 m1 n1 τ xy σ y τ yz m2
τ xz
τ yz σ z n2
or
σ x τ xy τ xz l1
τ x′y′ = l 2 m2 n2 τ xy σ y τ yz m1
τ xz
τ yz σ z n1
τ x′y′ = σ x l1l 2 + σ y m1 m2 + σ z n1 n2 +
(m1 n 2 + m2 n1 )τ yz + (n1l 2 + n 2 l1 )τ xz + (l1 m2 + l 2 m1 )τ xy
Similarly,
σ x τ xy τ xz l3
τ x′z′ = l1 m1 n1 τ xy σ y τ yz m3
τ xz
τ yz σ z n3
τ x′z′ = σ x l1l3 + σ y m1 m3 + σ z n1 n3 +
(m1 n3 + m3 n1 )τ yz + (n1l3 + n3 l1 )τ xz + (l1 m3 + l3 m1 )τ xy
σ x τ xy τ xz l3
τ y′z′ = l 2 m2 n 2 τ xy σ y τ yz m3
τ xz
τ yz σ z n3
τ y′z′ = σ x l 2 l3 + σ y m2 m3 + σ z n2 n3 +
(m2 n3 + m3 n2 )τ yz + (n 2 l 3 + n3l 2 )τ xz + (l 2 m3 + l3 m2 )τ xy
By analogous to σ PN = N [σ ]N , we can write down the stress vector in the ( x ′ ,
T
σ x′ τ x′y′ τ x′z′ l1 m1 n1 σ x τ xy τ xz l1 l 2 l3
τ x′y′ σ y′ τ y′z′ = l 2 m2 n 2 τ xy σ y τ yz m1 m2 m3
τ x′z′ τ y′z′ σ z′ l 3 m3 n3 τ xz τ yz σ z n1 n2 n3
T
σ x′ τ x′y′ τ x′z′ l1 m1 n1 σ x τ xy τ xz l1 m1 n1
τ x′y′ σ y′ τ y′z′ = l 2 m2 n 2 τ xy σ y τ yz l 2 m2 n2
τ x′z′ τ y′z′ σ z′ l 3 m3 n3 τ xz τ yz σ z l3 m3 n3
or, in short
[σ ′] = [T ][σ ][T ]T
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-14
Fig. 1.16
From section 1-3, since the shearing stresses vanish on the principal plane ABC , the
v
principal stress vector σ N is the only stress vector acting on the principal plane and in the
v
direction of the unit normal vector N to the principal plane.
r
N = liˆ + mˆj + nkˆ
If the infinitesimal area ABC as shown in Fig. 1.16 is defines as dAABC , then,
dA0 BC = ldAABC
dA0 AC = mdAABC
dA0 AB = ndAABC
v
The projection of the principal stress vector σ N along the coordinate ( x , y , z ) are
v
σ N = σ N liˆ + σ N mˆj + σ N nkˆ
From the force equilibrium equations, we have
(σ x − σ N )l + τ xy m + τ xz n = 0
τ xy l + (σ y − σ N )m + τ yz n = 0
τ xz l + τ yz m + (σ z − σ N )n = 0
These 3 equations are linear homogeneous equations. Since all three direction cosines
can not be zero ( l 2 + m 2 + n 2 = 1 ), the system of the linear homogeneous equations has a
nontrivial solution if and only if the determinant of the coefficients of l , m , and n vanish.
Thus, we have
σ x −σ N τ xy τ xz
τ xy σ y −σ N τ yz =0
τ xz τ yz σ z −σ N
σ x τ xy τ xz
I 3 = τ xy σ y τ yz = σ xσ yσ z + 2τ xyτ yzτ xz − (σ xτ yz2 + σ yτ zx2 + σ zτ xy2 )
τ xz τ yz σ z
This cubic polynomial equation has three roots σ 1 , σ 2 , and σ 3 which are the
The magnitudes and directions of the principal stresses σ 1 , σ 2 , and σ 3 for a given
member depend only on the loads being applied to the member. They are independent upon
the choice of initial coordinate system ( x , y , z ) used to specify the state of stress at point 0 .
Thus, the constants I 1 , I 2 , and I 3 must remain the same magnitudes for all the choices of
The direction cosines of the principal normal stresses σ i (i = 1,2,3) can be determined
by:
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-16
1. Substituting any one of the three roots σ i (i = 1,2,3) into any two of the cubic
equation, we obtain
(σ x − σ i )l i + τ xy mi + τ xz ni = 0
τ xy li + (σ y − σ i )mi + τ yz ni = 0
τ xz li + τ yz mi + (σ z − σ i )ni = 0
2. Solving two of the above three equations together with the equation
li2 + mi2 + ni2 = 1 for the direction cosines li , mi , and ni .
For example, if we need to find the direction cosines of the principal stress σ 1 or l1 ,
m1 , and n1 .
- Substituting the principal stress σ 1 into any two of the cubic equation such as
(σ x − σ 1 )l1 + τ xy m1 + τ xz n1 = 0
τ xy l1 + (σ y − σ 1 )m1 + τ yz n1 = 0
l1 m
(σ x − σ 1 ) + τ xy 1 + τ xz = 0
n1 n1
l1 m
τ xy + (σ y − σ 1 ) 1 + τ yz = 0
n1 n1
l1 m
- Solving the equations for and 1
n1 n1
- Substituting the direction cosines l1 and m1 which are the functions of the
direction cosine n1 into the equation l12 + m12 + n12 = 1 , and solving for n1
Two Dimensions
Consider the plate structure subjected only to the external load parallel to the plate as
shown in Fig. 1.17. If the plate is very thin compared to the dimension of the plate, the
stresses σ z , τ xz , and τ yx on an infinitesimal small element far away from the loading points
are approximately equal to zero. In addition, let us assume that the remaining stress
components σ x , σ y , and τ xy are independent of z . This kind of state of stresses on the
infinitesimal small element is called the plane stress as shown in Fig. 1.18.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-17
Fig. 1.17
Fig. 1.18
In the similar fashion as for the coordinate transformation in three dimensions, we can
determine the coordinate transformation matrix of the plane stress as following.
Fig. 1.19
Let ( x , y ) and ( x ′ , y ′ ) denote two rectangular coordinate systems with a common
origin as shown in Fig. 1.19. Also, let a general point in space P has coordinate ( x , y ) and
( x ′ , y ′ ) in the respective coordinate system. The angle between the coordinate axes ( x , y )
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-18
( x ′ , y ′ ) coordinates are
σ x′ m 2 n2 2mn σ x
2
σ y′ = n − 2mn σ y
2
m
− mn mn m 2 − n 2
τ x′y′ τ xy
or
σ x +σ y σ x −σ y
σ x′ = + cos 2θ + τ xy sin 2θ
2 2
σ x +σ y σ x −σ y
σ y′ = − cos 2θ − τ xy sin 2θ
2 2
σ x −σ y
τ x′y′ = − sin 2θ + τ xy cos 2θ
2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-19
Fig. 1.20
Fig. 1.20 shows the variation of the stress σ x′ , σ y′ , and τ x′y′ versus θ for σ y = 0.2σ x
and τ xy = 0.8σ x .
Fig. 1.21
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-21
Fig. 1.22
Consider the stress element subjected to the state of principal stresses σ 1 , σ 2 , and σ 3
as shown in Fig. 1.22a. Also, consider the plane containing any two principal normal stresses,
such as σ 2 and σ 3 . The maximum shear stress in this plane (the 2-3 plane) occurs on the
stress element that is rotated 45 o about the remaining σ 1 principal stress axis as shown in
Fig. 1.22b. Since the principal stress σ 1 is independent from the state of stresses on the 2-3
plane, thus, the shear stress τ max is one of three principal shear stresses on the stress element
2−3
Similarly, the shear stresses τ max and τ max are the principal shear stresses on the 1-3 plane
1−3 1− 2
and the plane 1-2 of the stress element as shown in Fig. 1.22c and 1.22d, respectively. They
have the absolute value of
σ1 −σ 3
τ max =
1−3 2
σ1 −σ 2
τ max =
1− 2 2
One of the τ max , τ max , and τ max is the maximum shear stress that occurs for all
2−3 1−3 1− 2
Each plane of the principal shear stress is also acted upon by a normal stress that is the
same in the two orthogonal directions.
σ2 +σ3
τ max plane; σ τ , 2 −3 =
2−3 2
σ1 + σ 3
τ max plane; σ τ ,1−3 =
1−3 2
σ1 + σ 2
τ max plane; σ τ ,1− 2 =
1− 2 2
Mohr’s Circles for the Principal Planes
Consider a state of stress that has two components of shear stress equal to zero such as
τ xz = τ yz =0 as shown in Fig. 1.23.
Fig. 1.23
Then, the normal stress σ z in the direction normal to the plane of the nonzero
component of shear stress τ xy is one of the principal stresses of the state of stresses shown.
σz =σ3
From the discussions of the plane stress transformation, we can see that the normal
stress σ z does not influence the transformation equations. Thus, in terms of the principal
stresses (no shear stress), the Morh’s circle can be constructed by using any two of the
principal stresses as shown in Fig. 1.24.
Fig. 1.24
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-23
Fig. 1.25
Thus, the octahedral normal stress, σ oct (or the hydrostatic stress) can be determined from the
equation
σ PN = σ x l 2 + σ y m 2 + σ z n 2 + 2mnτ yz + 2nlτ xz + 2lmτ xy
1
σ oct = (σ 1 + σ 2 + σ 3 )
3
In addition, since σ 1 + σ 2 + σ 3 = σ x + σ y + σ z (see Example 1-1), we have
1 I
σ oct = (σ x + σ y + σ z ) = 1
3 3
The the octahedral shearing stress, τ oct can be determined from the equation
σ PS = σ P2 − σ PN
2
= σ Px
2
+ σ Py
2
+ σ Pz
2
− σ PN
2
1 2 1 2 1 2 1
τ oct = σ 1 + σ 2 + σ 3 − (σ 1 + σ 2 + σ 3 ) 2
3 3 3 9
9τ oct
2
= (σ 1 − σ 2 ) 2 + (σ 1 − σ 3 ) 2 + (σ 2 − σ 3 ) 2 = 2 I 12 − 6 I 2
9τ oct
2
= 2 I 12 − 6 I 2
= (σ x − σ y ) 2 + (σ x − σ z ) 2 + (σ y − σ z ) 2 + 6τ xy2 + 6τ xz2 + 6τ yz2
The term σ oct and τ oct are the important quantities as they are used to predict the
1 1 I
σ m = σ oct = (σ x + σ y + σ z ) = (σ 1 + σ 2 + σ 3 ) = 1
3 3 3
Therefore, the plastic behavior of the materials is related mainly to the part of stresses that is
independent of σ m .
where the stress array Tm is called the mean stress tensor or hydrostatic stress tensor, and the
σ x + σ y + σ z
0 0
σ x τ xy τ xz 3
σ x +σ y +σ z
τ yx σ y τ yz = 0 0 +T
3 d
τ zx τ zy σ z σ x +σ y +σ z
0 0
3
σ x +σ y +σ z
σ x − τ xy τ xz
3
σ x +σ y +σ z
Td = τ yx σy − τ yz
3
σ y +σ y +σ z
τ zx τ zy σz −
3
2σ x − σ y − σ z
τ xy τ xz
3
2σ y − σ x − σ z
Td = τ yx τ yz
3
2σ z − σ y − σ x
τ zx τ zy
3
If the stress tensor is the principal stress tensor, we can determine the principal values
of the stress deviator as following
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-25
σ 1 + σ 2 + σ 3
0 0
σ 1 0 0
3
0 σ σ1 + σ 2 + σ 3
2 0 = 0 0 + Td
3
0 0 σ 3 σ1 + σ 2 + σ 3
0 0
3
σ1 + σ 2 + σ 3
σ 1 − 3
0 0
σ1 + σ 2 + σ 3
Td = 0 σ2 − 0
3
σ1 + σ 2 + σ 3
0 0 σ3 −
3
2σ 1 − σ 2 − σ 3
0 0
3
2σ 2 − σ 1 − σ 3
Td = 0 0
3
2σ 3 − σ 1 − σ 2
0 0
3
S1 0 0
Td = 0 S2 0
0 0 S 3
We can see that S1 + S 2 + S 3 = 0 , thus only two of the principal stresses of Td are
independent (we can find the 3rd term from these two).
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-26
Example 1-1
Given a state of stresses at a point with respect to a convenient coordinate system ( x ,
y , z ) be σ x = 100 MPa , σ y = −60 MPa , σ z = 40 MPa , τ xy = 80 MPa , τ yz = τ xz = 0 MPa .
a.) Determine the principal normal stresses and the direction cosine of the principal
normal stresses
b.) Determine the principal shear stresses.
c.) Determine the octahedral normal stresses and octahedral shearing stresses.
d.) Determine the mean and deviator stress.
The principal normal stresses
100 − σ 80 0
80 − 60 − σ 0 =0
0 0 40 − σ
σ 3 − I 1σ 2 + I 2σ − I 3 = 0
where the stress invariants,
I 1 = 100 − 60 + 40 = 80 MPa
(σ x − σ i )l i + τ xy mi + τ xz ni = 0 and τ xz li + τ yz mi + (σ z − σ i )ni = 0
1
m1 = = 0.383 and l1 = 0.924
6.8274
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-27
We can see that the principal axis for σ 1 lie in the x − y plane at the angle of
l 2 = −0.383 m2 = 0.924 n2 = 0
l3 = 0 m3 = 0 n3 = 1
It should be noticed that the principal axis for σ 2 also lie in the x − y plane and is
perpendicular to the principal axis for σ 1 . In addition, the principal axis for σ 3 is coincident
with the original z − axis.
The principal shear stresses
σ 1 − σ 3 133.137 − (−93.137)
τ max = = = 113.14 MPa
1−3 2 2
σ 2 −σ3 40 − (−93.137)
τ max = = = 66.56 MPa
2 −3 2 2
σ 1 − σ 2 133.137 − 40
τ max = = = 46.56 MPa
1− 2 2 2
It should be noted that since there is only one nonzero component of shear stress τ xy ,
the stress normal to the plane of τ xy is one of the principal normal stresses.
σ 2 = σ z = 40 MPa
Then, the Mohr’s circle as shown in Fig. Ex 1-1 below may be used to determine the principal
normal stresses, the direction cosine of the principal normal stresses, and the principal shear
stresses in the x − y plane as for the 2-D problem.
Fig. Ex 1-1
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-28
26.667 0 0
Tm = 0 26.667 0 MPa
0 0 26.667
Example 1-2
Given the state of stresses at the dark spot on the surface of a pressure vessel as shown
in Fig. Ex 1-2a.
a.) Determine the state of principal stresses.
b.) Determine the state of maximum in-plane shear stress.
Fig. Ex 1-2a
From the element right-hand face and the sign convention,
σ x = −20 MPa σ y = +90 MPa τ xy = 60 MPa
σ x + σ y − 20 + 90
σ avg = = = 35 MPa
2 2
2 2
Then, the initial point A (−20,+60) and the center C (35,0) are plotted on the Mohr’s circle as
shown in Fig. Ex 1-2b.
Determine the state of principal stresses
The principal stresses are indicated by the coordinate of points B and D on Mohr’s
circle.
σ 1 = 35 + 81.4 = 116.4 MPa
σ 2 = 35 − 81.4 = −46.4 MPa
The orientation of the element is determined by calculating the counterclockwise
angle 2θ p1 from the radius line AC to BC .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-30
60
2θ p1 = 180 o − φ = 180 o − tan −1 = 132.5 o
55
Fig. Ex 1-2b
Thus, the orientation of the planes that contain the state of the principal stresses is
132.5 o
θ p1 = = 66.3o
2
Fig. Ex 1-2c shows the state of the principal stresses.
Fig. Ex 1-2c
The state of maximum in-plane shear stress
The maximum in-plane shear stress and the average normal stress are identified by the
point E and F on the Mohr’s circle in Fig. Ex 1-2d. Hence, we have
τ max = 81.4 MPa
in − plane
and
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-31
σ avg = 35 MPa
The counterclockwise angle 2θ s11 from the radius line AC to the radius line EC is
2θ s1 = 90 o − φ = 90 o − 47.5 o = 42.5 o
Thus, the orientation of the planes that contain the state of the maximum in-plane shear stress
is
42.5 o
θ s1 = = 21.3o
2
Fig. Ex 1-2d
Fig. Ex 1-2e shows the state of the maximum in-plane shear stress.
Fig. Ex 1-2e
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-32
Fig. 1.26
∂σ x ∂τ xy
∑F x = 0; (σ x +
∂x
dx − σ x ) dydz + ( τ xy +
∂y
dy − τ xy ) dxdz
∂τ xz
+ ( τ xz + dz − τ xz ) dxdy + B x (dxdydz ) = 0
∂z
∂σ x ∂τ xy ∂τ xz
+ + + Bx = 0
∂x ∂y ∂z
where B x is the body force per unit volume in the x direction including the inertia forces.
Similarly,
∑F y = 0;
∂τ xy ∂σ y ∂τ yz
+ + + By = 0
∂x ∂y ∂z
∑F z = 0;
∂τ xz ∂τ yz ∂σ z
+ + + Bz = 0
∂x ∂y ∂z
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-33
Fig. 1.27
If the body is a plane body of uniform thickness as shown in Fig. 1.27, the differential
equations of equilibrium of the three-dimensional body can be reduced to
∂σ x ∂τ xy
+ + Bx = 0
∂x ∂y
∂τ xy ∂σ y
+ + By = 0
∂x ∂y
Equilibrium Equations in Polar Coordinate: Plane Problem
Consider a plane body of uniform thickness in polar coordinate ( r , θ ) as shown in
Fig. 1.28, the differential equations of motion of this plane body can be written as following,
Fig. 1.28
∑F r = 0;
dθ ∂σ r
− σ r (rdθ ) − τ rθ (dr ) − σ θ sin (dr ) + σ r + dr (r + dr )dθ +
2 ∂r
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-34
∂τ rθ ∂σ θ dθ
τ rθ + ∂θ dθ dr − σ θ + ∂θ dθ sin 2 (dr ) + Br rdrdθ = 0
∑ Fθ = 0;
dθ ∂τ
− τ rθ (rdθ ) − σ θ (dr ) − τ rθ sin (dr ) + τ rθ + rθ dr (r + dr )dθ +
2 ∂r
∂σ θ ∂τ rθ dθ
σ θ + ∂θ dθ dr − τ rθ + ∂θ dθ sin 2 (dr ) + Bθ rdrdθ = 0
dθ dθ
Expanding the above two equations, setting sin = , and neglecting the higher-
2 2
order terms, we have the equilibrium equations in the polar coordinate.
∂σ r 1 ∂τ rθ σ r − σ θ
+ + + Br = 0
∂r r ∂θ r
1 ∂σ θ ∂τ rθ τ
+ + 2 rθ + Bθ = 0
r ∂θ ∂r r
1.6 Deformation of a Deformable Body
Fig. 1.29
Consider a deformation body in equilibrium as shown in Fig. 1.29 subjected to
external loads and deformed to a new equilibrium position indicated by the dashed line. The
coordinate of point A on an undeformed is ( x , y , z ) and, after the deformation, the point A
x* = x + u
y* = y + v
z* = z + w
Therefore, in Lagrangian coordinate method, we can write down the relationship between the
coordinate x * , y * , z * and x , y , z in the form of
x * = x * ( x, y , z )
y * = y * ( x, y , z )
z * = z * ( x, y , z )
The functions need to be continuous and differentiable with respect to the independent
variables. Discontinuity implies rupture of the body.
The total differential equation of the coordinate x * , y * , z * can be written as
∂x * ∂x * ∂x *
dx * = dx + dy + dz
∂x ∂y ∂z
∂y * ∂y * ∂y *
dy =*
dx + dy + dz
∂x ∂y ∂z
∂z * ∂z * ∂z *
dz * = dx + dy + dz
∂x ∂y ∂z
1.7 Strain Theory: Principal Strains
Let us define the engineering strain ε E of the line element ds that is transformed to
1 ds * 1
2
M =
2 ds
[ 1
− 1 = (1 + ε E ) 2 − 1 = ε E + ε E2 ]
2 2
Fig. 1.30
v v
By the deformation, the line vector dr1 is moved to be the line vector dr1* as shown in
Fig. 1.30. The vector linked between these two vectors is
v v v
v ∂u ∂u ∂u
u+ dx1 + dy1 + dz1 =
∂x ∂y ∂z
∂u ∂u ∂u ∂v ∂v ∂v
(u + dx1 + dy1 + dz1 )iˆ + (v + dx1 + dy1 + dz1 ) ˆj
∂x ∂y ∂z ∂x ∂y ∂z
∂w ∂w ∂w
+ (w + dx1 + dy1 + dz1 )kˆ
∂x ∂y ∂z
v
Thus, from Fig. 1.30, we can determine the line vector dr1* from the relationship
v v v
v v* v v ∂u ∂u ∂u
u + dr1 = dr + (u + dx1 + dy1 + dz1 )
∂x ∂y ∂z
Then,
v v v
v* v v ∂u ∂u ∂u v
dr1 = dr + (u + dx1 + dy1 + dz1 ) − u
∂x ∂y ∂z
r ∂u ∂u ∂u ∂v ∂v ∂v
dr1* = (dx1 + dx1 + dy1 + dz1 )iˆ + (dy1 + dx1 + dy1 + dz1 ) ˆj
∂x ∂y ∂z ∂x ∂y ∂z
∂w ∂w ∂w
+ (dz1 + dx1 + dy1 + dz1 )kˆ
∂x ∂y ∂z
Similarly, let consider another particle point having an infinitesimal small distance
v
from point A as defined by the line vector dr2 from point A to the particle as shown in Fig.
1.30 and
v
dr2 = dx 2 iˆ + dy 2 ˆj + dz 2 kˆ
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-37
dr2 = dx 22 + dy 22 + dz 22
v v
By the deformation, the line vector dr2 is moved to be the vector dr2* .
r ∂u ∂u ∂u ∂v ∂v ∂v
dr2* = (dx 2 + dx 2 + dy 2 + dz 2 )iˆ + (dy 2 + dx 2 + dy 2 + dz 2 ) ˆj
∂x ∂y ∂z ∂x ∂y ∂z
∂w ∂w ∂w
+ (dz 2 + dx 2 + dy 2 + dz 2 )kˆ
∂x ∂y ∂z
∂u ∂u ∂u ∂v ∂v ∂w
Now, let = u, x , = u, y , = u, z , = v, x , = v, y , K , and = w, z . Then,
∂x ∂y ∂z ∂x ∂y ∂z
in the form of the magnification factor M , we consider the change in the vector
v v v v
multiplication (dot product) of the line vector dr1* and dr2* , and the line vector dr1 and dr2
v v
with respect to the line vector dr1 times dr2 .
v v v v
dr1* .dr2* − dr1 .dr2
v v = [(dx1 + u , x dx1 + u , y dy1 + u , z dz1 )(dx 2 + u , x dx 2 + u , y dy 2 + u , z dz 2 ) +
dr1 .dr2
(dy1 + v, x dx1 + v, y dy1 + v, z dz1 )(dy 2 + v, x dx 2 + v, y dy 2 + v, z dz 2 ) +
(dz1 + w, x dx1 + w, y dy1 + w, z dz1 )(dz 2 + w, x dx 2 + w, y dy 2 + w, z dz 2 ) −
(dx1 dx 2 + dy1 dy 2 + dz1 dz 2 )]/ (dx1 dx 2 + dy1 dy 2 + dz1 dz 2 )
Expanding all terms of the equation, for example,
(dx1 + u , x dx1 + u , y dy1 + u , z dz1 )(dx 2 + u , x dx 2 + u , y dy 2 + u , z dz 2 ) =
………
Then, consider only the numerator of the equation, we have
v* v* v v
dr1 .dr2 − dr1 .dr2 =
(u , x +u , x u , z + v, x v, z + K)dx1 dz 2 +
(u , y u, z + v, z + v, y v, z + K)dy1 dz 2 +
( w, z + w, z + w, 2z + K)dz1 dz 2 ]
Rearranging,
v v v v
dr1* .dr2* − dr1 .dr = 2ε x dx1 dx 2 + γ xy dx1 dy 2 + γ xz dx1dz 2 +
2
[
1 v* v* v v
] 1 1
dr1 .dr2 − dr1 .dr2 = ε x dx1 dx 2 + γ xy dx1 dy 2 + γ xz dx1 dz 2 +
2 2
1 1
γ yx dy1 dx 2 + ε y dy1 dy 2 + γ yz dy1 dz 2 +
2 2
1 1
γ zx dz1 dx 2 + γ zy dz1 dy 2 + ε z dz1 dz 2
2 2
where
∂u 1 ∂u ∂v ∂w
2 2 2
εx = + + +
∂x 2 ∂x ∂x ∂x
∂v 1 ∂u ∂v ∂w
2 2 2
ε y = + + +
∂y 2 ∂y ∂y ∂y
∂w 1 ∂u ∂v ∂w
2 2 2
εz = + + +
∂z 2 ∂z ∂z ∂z
1 1 1 ∂v ∂u ∂u ∂u ∂v ∂v ∂w ∂w
γ xy = γ yx = + + + +
2 2 2 ∂x ∂y ∂x ∂y ∂x ∂y ∂x ∂y
1 1 1 ∂w ∂u ∂u ∂u ∂v ∂v ∂w ∂w
γ xz = γ zx = + + + +
2 2 2 ∂x ∂z ∂x ∂z ∂x ∂z ∂x ∂z
1 1 1 ∂w ∂v ∂u ∂u ∂v ∂v ∂w ∂w
γ yz = γ zy = + + + +
2 2 2 ∂y ∂z ∂y ∂z ∂y ∂z ∂y ∂z
which are called the finite strain-displacement relations.
v v
For a special case when the line vector dr1 and dr2 are identical ( dx1 = dx 2 = dx ,
dy1 = dy 2 = dy , dz1 = dz 2 = dz ), we have the dot product of the vectors
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-39
v v
dr1 .dr2 = dx1 dx 2 + dy1 dy 2 + dz1 dz 2 = ds 2
ds 2 = dx 2 + dy 2 + dz 2
ds = dx 2 + dy 2 + dz 2
v v
By the definition of the direction cosine and let the identical vector dr1 and dr2 have
dx dy dz
the direction cosine of =l, = m , and = n , we have
ds ds ds
v v v v
1 dr1* .dr2* − dr1 .dr2 dx 2 1 dxdy 1 dxdz
v v = ε x 2
+ γ xy 2
+ γ xz +
2 dr1 .dr2 ds 2 ds 2 ds 2
1 dydx dy 2 1 dydz
γ yx 2
+ ε y 2
+ γ yz +
2 ds ds 2 ds 2
1 dzdx 1 dzdy dz 2
γ zx 2 + γ zy +εz 2
2 ds 2 ds 2 ds
Thus,
v v v v
1 (ds * ) 2 − ds 2 1 dr1* .dr2* − dr1 .dr2
M = 2 = v v = ε x l 2 + γ xy lm + γ xz nl + ε y m 2 + γ yz mn + ε z n 2
2 ds 2 d r1 .d r2
or,
1
ε E + ε E2 = ε x l 2 + γ xy lm + γ xz nl + ε y m 2 + γ yz mn + ε z n 2
2
v
Final Direction of Vector r1
v v
As a result of the deformation, vector r1 deforms into the vector r1* . Let the direction
v v
cosines of vector r1 and r1* are
dx dy dz
=l, = m , and =n
dr1 dr1 dr1
dx * dy * dz *
*
= l *
, *
= m *
, and *
= n*
dr1 dr1 dr1
Alternatively, we may write
dx * dr1 * dy
*
dr1 dz * dr1
*
= l , *
= m *
, and *
= n*
dr1 dr1 dr1 dr1 dr1 dr1
∂x * ∂x * ∂x * ∂y * ∂y * ∂y * ∂z * ∂z *
dx * = dx + dy + dz , dy * = dx + dy + dz , and dz * = dx + dy +
∂x ∂y ∂z ∂x ∂y ∂z ∂x ∂y
∂z *
dz , we find
∂z
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-40
dx * ∂u ∂u ∂u
= 1 + l + m+ n
dr1 ∂x ∂y ∂z
dy * ∂v ∂y ∂v
= l + 1 + m + n
dr1 ∂x ∂y ∂z
dz * ∂w ∂w ∂w
= l+ m + 1 + n
dr1 ∂x ∂y ∂z
By using the engineering strain equation dr1 / dr1* = 1 /(1 + ε E ) with the above two
v v
equations, we obtain the final direction cosines of vector r1 when it passes into the vector r1*
under the deformation in the form of
∂u ∂u ∂u
(1 + ε E )l * = 1 + l + m+ n
∂x ∂y ∂z
∂v ∂y ∂v
(1 + ε E )m * = l + 1 + m + n
∂x ∂y ∂z
∂w ∂w ∂w
(1 + ε E )n * = l+ m + 1 + n
∂x ∂y ∂z
Definition of Shear Strain
v v
If originally the vectors dr1 and dr2 having the direction cosine l1 , m1 , n1 and l 2 , m2 ,
n2 , respectively, are normal to each other, by the definition of scalar product of vectors
π
= l1l 2 + m1 m2 + n1 n 2 = 0
cos
2
v v
If the angle between the vectors dr1* and dr2* having the direction cosine l1* , m1* , n1*
and l 2* , m2* , n2* , respectively, is equal to θ * after the deformation, by the definition of scalar
product of vectors
cosθ * = l1*l 2* + m1* m2* + n1* n2*
Then, the angle change can be determined from
v v v v
dr1* .dr2* − dr1 .dr2 = dr1* dr2* cosθ * − 0 = dr1* dr2* cosθ *
By using the definition of the magnification factor where dr1* / dr1 = 1 + ε E1 and
γ 12 =
dr1* dr2*
dr1 dr2
[
cosθ * = (1 + ε E1 )(1 + ε E 2 ) l1*l 2* + m1* m2* + n1* n ]
v v
Using the equation of the final direction of vector r1 when it passes into the vector r1*
and the finite strain-displacement relation, we have the engineering shearing strain between
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-41
v v v v
the vector r1 and r2 as they are deformed into the vector r1* and r2* as shown in Fig. 1.30 in
the form of
γ 12 = 2ε x l1l 2 + γ xy l1 m2 + γ xz n2 l1 + γ yx l 2 m1 + 2ε y m1 m2 + γ yz m1 n2 + γ zx n1l 2 + γ zy m2 n1 + 2ε z n1 n2
If the strain ε E1 and ε E 2 are small and the angle change are small,
π
γ 12 ≈ −θ *
2
and the engineering shearing strain becomes approximately equal to the change in angle
v v
between the vector r1 and r2 .
Strain transformation
The strain tensor,
ε x ε xy ε xz ε x γ xy / 2 γ xz / 2
ε ij = ε yx εy ε yz = γ yx / 2 ε y γ yz / 2
ε zx ε zy ε z γ zx / 2 γ zy / 2 ε z
The strain tensor obeys the tensor law of transformation when the coordinates are
changed as the stress tensor.
The transformation of the strain components from the ( x , y ) coordinates to the ( x ′ ,
y ′ , z ′ ) coordinates,
T
ε x′ ε x′y′ ε x′z′ l1 m1 n1 ε x ε xy ε xz l1 m1 n1
l
ε x′y′ ε y′ ε y′z′ = l 2 m2 n 2 ε xy εy ε yz 2 m2 n2
ε x′z′ ε y′z′ ε z′ l 3 m3 n3 ε xz ε yz ε z l3 m3 n3
Performing the matrix operations, we have
ε x′ = ε x l12 + ε y m12 + ε z n12 + m1 n1γ yz + n1l1γ xz + l1 m1γ xy
γ x′y′ / 2 = ε x l1l 2 + ε y m1 m2 + ε z n1 n2 +
(m1 n 2 + m2 n1 )γ yz / 2 + (n1l 2 + n 2 l1 )γ xz / 2 + (l1 m2 + l 2 m1 )γ xy / 2
γ x′z′ / 2 = ε x l1l3 + ε y m1 m3 + ε z n1 n3 +
(m1 n3 + m3 n1 )γ yz / 2 + (n1l3 + n3 l1 )γ xz / 2 + (l1 m3 + l3 m1 )γ xy / 2
γ y′z′ / 2 = ε x l 2 l3 + ε y m2 m3 + ε z n2 n3 +
(m2 n3 + m3 n2 )γ yz / 2 + (n2 l 3 + n3 l 2 )γ xz / 2 + (l 2 m3 + l 3 m2 )γ xy / 2
Similar to the stress transformation, the transformation of the strain components from
the ( x , y ) coordinates to the ( x ′ , y ′ ) coordinates in two dimension can be performed as
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-42
ε x′ m 2 n2 mn ε x
2
ε y′ = n − mn ε y
2
m
− 2mn 2mn m 2 − n 2
γ x′y′ γ xy
where m = cosθ and n = sin θ .
εx +εy εx −εy γ xy
ε x′ = + cos 2θ + sin 2θ
2 2 2
εx +εy εx −εy γ xy
ε y′ = − cos 2θ − sin 2θ
2 2 2
τ x′y′ εx −εy γ xy
=− sin 2θ + cos 2θ
2 2 2
It should be noted that since the equations of plane strain transformation are
mathematically similar to the equations of plane stress transformation, Mohr’s circle can also
be used to solve problems involving the transformation of strain.
Principal strains
Through any point in an undeformed member, there are three mutually perpendicular
line elements that remain perpendicular under the deformation. The strains of these line
elements are called the principal strains at the point. The maximum and minimum strain can
be determined by using the 2 nd method of calculus of variation.
If we need to determine the maximum and minimum of the function F = F ( x, y, z )
F = ε x l 2 + γ xy lm + γ xz nl + ε y m 2 + γ yz mn + ε z n 2 - λ ( l 2 + m 2 + n 2 -1)
∂F
= 0; 2(ε x − λ )l + γ xy m + γ xz n = 0
∂l
∂F
= 0; γ xy l + 2(ε y − λ )m + γ yz n = 0
∂m
∂F
= 0; γ xz l + γ yz m + 2(ε z − λ )n = 0
∂n
l 2 + m 2 + n 2 =1
For these linear homogenous equations, we may rewrite the first three equations in the
determinant form as
(ε x − λ ) γ xy / 2 γ xz / 2
γ xy / 2 (ε y − λ ) γ yz / 2 = 0
γ xz / 2 γ yz / 2 (ε z − λ )
1
J 2 = ε x ε y + ε y ε z + ε z ε x − (γ xy2 + γ yz2 + γ xz2 )
4
εx γ xy / 2 γ xz / 2
J 3 = γ xy / 2 εy γ yz / 2
γ xz / 2 γ yz / 2 εz
Solving for the values of principal strains λ1 , λ 2 , and λ3 . Then, substituting λi back
into the simultaneous equations, we obtain the direction cosine of the principal plane li , mi ,
and ni , respectively.
J 2 = ε 1ε 2 + ε 2 ε 3 + ε 1ε 3
J 3 = ε 1ε 2 ε 3
In two-dimensions, the principal strains and principal planes can be obtained easily as
2 2
ε +εy εx −εy γ xy
ε1 = x ± +
2 2 2 2
γ xy
tan 2θ p =
εx −εy
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-44
2 2 2
− (ε x − ε y )
tan 2θ s =
γ xy
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-45
Example 1-3
Given a state of strains at a point with respect to a convenient coordinate system ( x ,
y , z ) be ε x = −3000 µε , ε y = 2000 µε , ε z = −2000 µε , γ xy = −5830 µε , γ yz = −670 µε ,
and γ xz = −3000 µε .
a.) Determine the principal normal strains and the direction cosine of the principal
normal strains
b.) Determine the principal shear strains.
The principal normal strains
The strain invariants,
J1 = ε x + ε y + ε z
= − 14.8595(10 −6 )
εx γ xy / 2 γ xz / 2
J 3 = γ xy / 2 εy γ yz / 2
γ xz / 2 γ yz / 2 εz
J 3 = 2.77607(10 −8 )
l1 m
0.00583 + −0.003 1 − 0.00067 = 0
n1 n1
Solving the simultaneous equations, we obtain
l1 m
= −2.576 and 1 = −5.228
n1 n1
γ 1 = 0.00838
γ 2 λ 2 − λ3
=
2 2
γ 2 = 0.00512
γ 3 λ1 − λ 2
=
2 2
γ 3 = 0.00326
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-47
Fig. 1.31
The normal strain at a point of a general testing specimen are usually obtained by
using a cluster of three electrical-resistance strain gauges, arranged in a specified pattern
called strain rosette as shown in Fig. 1.31 and Fig. 1.32.
Fig. 1.32
In general, if we know the angles θ a , θ b , and θ c of the strain rosette with respect to
an axis as shown in Fig. 1.32a and the measured strains ε a , ε b , and ε c , we can determine the
Example 1-4
The state of strains at point A on the bracket as shown in Fig. Ex 1-4 is measured
using the stain rosette as shown. Due to the loadings, the readings from the gauge give
ε a = 60 µε , ε b = 135 µε , and ε c = 264 µε . Determine the in-plane principal strains at the
point and the directions in which they act.
Fig. Ex 1-4
Setting the + x axis as shown and measuring the angles counterclockwise from the
+ x axis to the center-lines of each gauge. We have
θ a = 0 o , θ b = 60 o , and θ c = 120 o
Therefore, we obtain
60(10 −6 ) = ε x cos 2 0 o + ε y sin 2 0 o + γ xy sin 0 o cos 0 o
ε x = 60(10 −6 )
Fig. 1.33
By the definition of normal strain,
L0′2′ − L02 [dx + (u + u, x dx) − u ] − dx ∂u
εx = = = u, x =
L02 dx ∂x
Similarly,
∂v
εy =
∂y
For small displacement analysis, the engineering shear strain
(u + u , y dy ) − u (v + v, x dx) − v ∂u ∂v
γ xy = β 1 + β 2 = + = u , y + v, x = +
dy dx ∂y ∂x
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-52
displacement u with respect to x . The shearing strain γ xy represents the changes in the
original right angles between the line elements 0-1, and 0-2 to the 0′ -1′ and 0′ - 2′ due to the
deformation.
Strain compatibility Relations
Just as stresses must satisfy the equations of equilibrium, the strains must satisfy the
strain compatibility equations in order to describe a physically possible displacement field.
The displacement field must be single-valued, continuous, and has continuous derivatives.
Thus, the material does not overlap itself and no crack appears.
By eliminating the displacement components from the strains equations, we have
∂ 2ε y ∂ 2 ε x ∂ γ xy
2
+ =
∂x 2 ∂y 2 ∂x∂y
∂ 2 ε z ∂ 2 ε x ∂ 2γ xz
+ =
∂x 2 ∂z 2 ∂x∂z
∂ 2ε y ∂ 2 ε z ∂ γ yz
2
+ =
∂z 2 ∂y 2 ∂y∂z
∂ 2 ε z 1 ∂ γ xy 1 ∂ γ yz 1 ∂ 2γ zx
2 2
+ = +
∂x∂y 2 ∂z 2 2 ∂z∂x 2 ∂y∂z
∂ 2ε y1 ∂ 2γ xz 1 ∂ γ xy 1 ∂ γ yz
2 2
+ = +
∂x∂z 2 ∂y 2 2 ∂y∂z 2 ∂x∂y
∂ 2 ε x 1 ∂ γ yz 1 ∂ 2γ xz 1 ∂ γ xy
2 2
+ = +
∂y∂z 2 ∂x 2 2 ∂x∂y 2 ∂x∂z
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-53
Example 1-5
The parallelepiped as shown in Fig. Ex 1-5 is deformed into the shape indicated by the
dashed straight line (small displacements). The displacements are given by the following
relations: u = C1 xyz , v = C 2 xyz , and w = C 3 xyz.
a.) Determine the state of strain at point E when the coordinate of point E * for the
deformed body are (1.504, 1.002, 1.996).
b.) Check if the state of strain as point E is in accordance with the strain
compatibility relations.
Fig. Ex 1-5
The state of strain as point E
The displacements of point E are u = 1.504 − 1.5 = 0.004 m , v = 1.002 − 1 = 0.002 m ,
and w = 1.996 − 2 = −0.004 m . Thus, the displacement relations are in the form of
u 0.004 0.004
C1 = = =
xyz 1.5(1)2 3
0.004
u= xyz
3
In the same manner, we have
0.002
v= xyz
3
0.004
w=− xyz.
3
Thus, the strains at point E are
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 1-54
∂u 0.004 0.004
εx = = yz = (1)2 = 0.00267
∂x 3 3
∂v 0.002 0.002
εy = = xz = (1.5)2 = 0.00200
∂y 3 3
∂w 0.004
εz = =− xy = −0.00200
∂z 3
∂u ∂v 0002 0.004
γ xy = + = yz + xz = 0.00533
∂y ∂x 3 3
∂u ∂w 0.004 0.004
γ xz = + =− yz + xy = −0.00067
∂z ∂x 3 3
∂v ∂w 0.002 0.004
γ yz = + = xy − xz = −0.00300
∂z ∂y 3 3
Substituting the strain equations into the strain compatibility relations, we can see that
the state of strain as point E is in accordance with the strain compatibility relations.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-1
Chapter 2
Stress and Strain Relations
2.1 Concept of Engineering Stress-Strain and True Stress-Strain
Engineering stress-strain behavior is usually determined from monotonic tension test.
Fig. 2.1 shows the free body diagram of the test specimen. Originally, the specimen has the
cross-sectional area of Ao and the length of l o . Under the action of the axial tensile load P ,
the test specimen has the cross-sectional area of A and the length of l .
Fig. 2.1
P
Engineering stress σ=
Ao
l − lo δ
Engineering strain ε= =
lo lo
P
True stress σ~ =
A
l
1 l
True strain ε~ = ∫ dl = ln
l
o
l lo
The use of the true stress and true strain changes the appearance of the monotonic
tension stress-strain curve as shown in Fig. 2.2.
From the engineering stress-strain diagram, the material behavior can be classified
into 4 different ways depending on how the material behave.
Elastic behavior (1st region)
The specimen is called to response elastically if it returns to its original shape or
length after the load acting on it is removed.
In this region, the stress is proportional to the strain from the origin to the proportional
limit (stress has a linear relationship with strain).
If the stress is slightly over to the proportional limit, the material may still respond
elastically. If the stress is increased gradually, the slope of the curve tends to get smaller and
smaller until it reaches the elastic limit.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-2
Fig. 2.2
Strain Hardening (3th region)
When the yielding has ended, a further load increase can be applied to the specimen,
resulting in a curve that rise continuously but becomes flatter until it reaches the ultimate
stress. The rise in the curve is called strain hardening.
Necking (4th region)
At the ultimate stress, the cross-sectional area begins to decrease in a localized region
of the specimen, necking. This phenomenon is caused by slip planes of randomly oriented
crystals formed within the material. As the cross-sectional area is continually decreased, the
load is also gradually decreased, resulting in the stress strain diagram tends to curve
downward until the specimen breaks at the fracture stress.
If a specimen made of ductile material, such as steel and brass, is loaded pass the yield
point A to the plastic region at point A′ and then unloaded, elastic strain is recovered to
point O ′ as the material return to its equilibrium as shown in Fig. 2.3. But, the plastic strain
remains. As a result, the material is subjected to a permanent set.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-3
Fig. 2.3
In general, some heat and energy may be lost as the specimen is unloaded from A′ to
O ′ and loaded again from O ′ to A′ . As a result, the unloaded and loaded curves will have
the shape as shown. The colored area between these curves represents the energy lost and
called hysteresis loop.
True stress-Strain Curve
The true stress is always larger than the corresponding engineering stress, and the
difference may be of a factor of two or more near the end of a tensile test on a ductile
material. True strain based on a length measurement is somewhat smaller than the
corresponding engineering strain. However, once the necking starts, true strain based on an
area measurement are larger.
The total true strain in a tension test can be separated into two components.
1. Linearly elastic strain ε~ that can be recovered upon unloading.
e
2. Nonlinearly plastic strain ε~p that can not be recovered upon unloading.
Pf
σ~ f =
Af
Ao l f
=
A f lo
A
ε~ f = ln o
Af
Ao − A f
Let RA = area reduction of the specimen = . Hence,
Ao
1 A lf
= o =
1 − RA A f lo
1
ε~ f = ln
1 − RA
Since σ~ = K (~
e p ) n , at failure,
σ~ f = K (e~pfailure ) n = K (e~f ) n
σ~ f
K= ~ n
(ε ) f
~
σ ~ n
σ
ε~p = ~ = ~ ε~ f
σf σ f
(ε~ ) n
f
If n , ε~ f , and σ~ f are known, ε~p at a given σ~ can be determined. Also,
1
σ~
σ~ n
ε~total = ε~e + ε~p = + ~ ε~ f
E σ f
~ σ~ f ~ failure
ε f ,total = +εp
E
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-5
Proof
Since the elastic strain is small compared to the plastic strain, we can neglect the
elastic strain.
P P Ao A
σ~ = = =σ o
A Ao A A
l − lo l
ε= = −1
lo lo
l A A A
Since = o , thus, ε = o − 1 and o = 1 + ε . The true stress is
lo A A A
σ~ = σ (1 + ε )
l
Since the nonlinearly elastic strain can be determined from ε~p = ln = ln(1 + ε ) .
lo
K
σ= [ln(1 + ε )]n
1+ ε
At the necking point, the slope of the engineering stress-strain curve is equal to zero.
dσ K n
= [ln(1 + ε )]n − 1 + =0
dε ln(1 + ε )
necking necking
(1 + ε ) −2
Since the 1st and 2 nd term can not be zero, thus, the 3 rd term must be zero.
n = ln(1 + ε ) necking = ε~pnecking
where
B = 0.83 − 0.186 log ε~ (0.15 ≤ ε~ ≤ 3)
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-6
Example 2-1
A tension test was conducted on a specimen of AISI 1020 hot-rolled steel having an
initial diameter of 9.11 mm. The test data is given in Table Ex 2-1 where the length changes
over a 50 mm gage length have been converted to engineering strain ε in the first column.
Loads at corresponding times are given in the second column. Also, diameters for the large
strain portion of the test, measured in the neck when the necking once started, are given in the
third column. After fracture, the gage length had stretched to 68.5 mm.
Table Ex 2-1
Fig Ex 2-1a
The highest load reached was 25.75 kN. Thus, the ultimate tensile strength is
25750
σu = = 395 MPa
π (0.00911) 2 / 4
The percent elongation for 50 mm gage length is
L f − Li 68.5 − 50
100 = 100 = 37%
Li 50
The final diameter was 5.28 mm. Thus, the percent reduction of area is
d i2 − d 2f 9.112 − 5.28 2
100 = 100 = 66.4%
d i2 9.112
The modulus of toughness of AISI 1020 hot-rolled steel can be estimated by the
equation
σ y + σ u 264 + 395 MJ
uf ≈ ε f = 0.366 = 120 3
2 2 m
Since we do not have the data of the length and diameter of the specimen up to the
engineering strain 0.010. Therefore, the true stress-strain diagram in this portion must be
plotted by finding the true stress and true strain from the equation
σ~ = σ (1 + ε )
ε~ = ln(1 + ε )
After that the true stress-strain diagram can be plotted by finding the true stress and true strain
from the equation
P
σ~ =
A
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-8
A
ε~ = ln o
A
From the data of the true stress and true strain, we can plot the true stress-strain curve
as shown in Fig Ex 2-1a.
The true stress-true strain relationship in the plastic region is in the form of
σ~ = K (~
ep )n
or
log σ~ = n log ε~p + log K
which gives a straight line for the log-log true stress and true strain plot as shown in Fig Ex 2-
1b. It should be noted that the true stress used in this plot is the true stress corrected by using
the Bridgeman correction factor B .
Fig Ex 2-1b
The equation is in a form of a straight line on an x − y plot
y = mx + b
From the plot, we can see that the straight line will have a slope of n and intercept at
ε~p = 1 of σ~ = K . By using a graphical or least square method, we obtain
m = n = 0.206
b = 2.7967
Since b = log K , thus, K = 10 b = 626 MPa and the true stress-true strain relationship
in the plastic region is in the form of
σ~ = 626(~
e p ) 0.206
surface forces and during the displacement variations δu , δv , and δw , the body is in static
equilibrium.
For a condition which no net heat flow into the volume V * , the first law of
thermodynamic states that during the displacement variations δu , δv , and δw , the variation
in work of the external forces δWe is equal to the variation of the internal energy δU .
δWe = δU
δW S + δW B = δU
where δWS = the work of the surface forces and δWB = the work of the body forces.
Fig. 2.4
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-10
σ Py = τ xy l + σ y m + τ yz n
σ Pz = τ xz l + τ yz m + σ z n
Fig. 2.5
The force components due the stress vector components acting on the area dS = dAABC
are
dFx = σ Px dS dFy = σ Py dS dFz = σ Pz dS
∫∫∫ ( B δu + B δv + B δw)dV
x y z
V*
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-11
Gauss’s theorem
Let the vector on the surface dS having the total volume dV is S x iˆ + S y ˆj + S z kˆ . The
v
surface dS has the normal vector n = liˆ + mˆj + nkˆ . By using the gauss’s theorem, we have
∂S x ∂S y ∂S z
∫∫∫ ∂x
V
+
∂y
+
∂z
dV = ∫∫ (lS x + mS y + nS z )dS
S
Rewriting the equation of the work δWS into the form of ∫∫ (lS
S
x + mS y + nS z )dS , we
have
δWS = ∫∫ [(σ xδu + τ xy δv + τ xz δw)l + (τ xy δu + σ y δv + τ yz δw)m + (τ xz δu + τ yz δv + σ z δw)n]dS
S*
Let S x = σ x δu + τ xy δv + τ xz δw , S y = τ xy δu + σ y δv + τ yz δw , and S z = τ xz δu + τ yz δv + σ z δw ,
∂S x ∂S y ∂S z ∂ ∂
∫∫∫
V
∂x
+
∂y
+
∂z
dV = ∫∫∫ ∂x (σ x δu + τ xy δv + τ xz δw) +
∂y
(τ xy δu + σ y δv + τ yz δw)
V*
∂
+ (τ xz δu + τ yz δv + σ z δw) dV
∂z
The variation in work of the external forces δWe can be rewritten in the form of
∂ ∂
δ We = ∫∫∫ ∂x (σ x δu + τ xy δv + τ xz δw) +
∂y
(τ xy δu + σ y δv + τ yz δw)
V*
∂
+ (τ xz δu + τ yz δv + σ z δw) + B x δu + B y δv + B z δw dV
∂z
Note that the partial derivatives,
∂ ∂δu ∂σ x ∂σ x
σ x δu = σ x + δu = σ x δε x + δu
∂x ∂x ∂x ∂x
∂ ∂δv ∂τ xy
τ xy δv = τ xy + δv
∂x ∂x ∂x
∂ ∂δw ∂τ
τ xz δw = τ xz + δw xz
∂x ∂x ∂x
∂ ∂δu ∂τ xy
τ xy δu = τ xy + δu
∂y ∂y ∂y
∂ ∂δv ∂σ y ∂σ y
σ y δv = σ y + δv = σ y δε y + δv
∂y ∂y ∂y ∂y
M
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-12
∂ ∂δw ∂σ ∂σ
σ z δw = σ z + δw z = σ z δε z + δw z
∂z ∂z ∂z ∂z
Rearranging the terms, we have
δ We = ∫∫∫ (σ x δε x + σ y δε y + σ z δε z + 2τ xy δε xy + 2τ yz δε yz + 2τ xz δε xz )dV +
V*
∂σ x ∂τ xy ∂τ xz ∂τ xy ∂σ y ∂τ yz
∫∫∫* ∂x ∂y ∂z
+ + + B x
δ u +
∂x
+
∂y
+
∂z
+ B y
δv
V
∂τ ∂τ yz ∂σ z
+ xz + + + B z δw dV
∂x ∂y ∂z
Since the components in the second integral is the differential equilibrium equation of
deformable body which is equal to zero, then,
δ We = ∫∫∫ (σ x δε x + σ y δε y + σ z δε z + 2τ xy δε xy + 2τ yz δε yz + 2τ xz δε xz )dV
V*
or,
δ We = ∫∫∫ (σ *
x δε x + σ y δε y + σ z δε z + τ xy δγ xy + τ yz δγ yz + τ xz δγ xz )dV
V
The internal energy U of the volume V * of the body can be expressed in terms of the
internal-energy density U o as
U = ∫∫∫U o dV
V*
δU o = σ x δε x + σ y δε y + σ z δε z + 2τ xy δε xy + 2τ yz δε yz + 2τ xz δε xz
In the index notation, we have
δU o = σ i δε i
i = 1, 2, 3, 4, 5, 6
Elasticity and internal energy density
For linearly elastic material, the total internal energy in a loaded body is equal to the
total potential energy of the internal forces or elastic strain energy.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-13
Since the strain-energy density U o generally depends on the strain components, the
∂U o ∂U ∂U o ∂U o ∂U ∂U
δU o = δε x + o δε y + δε z + δε xy + o δε yz + o δε xz
∂ε x ∂ε y ∂ε z ∂ε xy ∂ε yz ∂ε xz
σ z δε z + 2τ xy δε xy + 2τ yz δε yz + 2τ xz δε xz , we have
∂U o ∂U o ∂U o
σx = σy = σz =
∂ε x ∂ε y ∂ε z
1 ∂U o ∂U o 1 ∂U o ∂U o 1 ∂U o ∂U o
τ xy = = τ yz = = τ xz = =
2 ∂ε xy ∂γ xy 2 ∂ε yz ∂γ yz 2 ∂ε xz ∂γ xz
∂U o
the form of σ x = . Thus, the strain-energy density in the specimen is
∂ε x
U o = ∫ σ x dε x
Fig. 2.6
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-14
This equation represents the area under the σ x − ε x diagram as shown in Fig. 2.6. In
energy density. The C o is represented by the area above the σ x − ε x curve. Hence,
C o = ∫ ε x dσ x
dC o
εx =
dσ x
In general, the strain components are expressed as functions of the stress components.
ε x = f1 (σ x , σ y , σ z ,τ xy ,τ yz , τ xz )
ε y = f 2 (σ x , σ y , σ z ,τ xy ,τ yz ,τ xz )
ε z = f 3 (σ x , σ y , σ z ,τ xy ,τ yz ,τ xz )
1
ε xy = γ xy = f 4 (σ x , σ y , σ z , τ xy , τ yz ,τ xz )
2
1
ε yz = γ yz = f 5 (σ x , σ y , σ z ,τ xy ,τ yz ,τ xz )
2
1
ε xz = γ xz = f 6 (σ x , σ y , σ z , τ xy , τ yz ,τ xz )
2
Thus, in analogous with the previous discussion, we have
C o = −U o + σ x ε x + σ y ε y + σ z ε z + 2τ xy ε xy + 2τ yz ε yz + 2τ xz ε xz
Differentiating the equation with respect to σ x and using the chain rule, we have
∂U o ∂C ∂σ x ∂ε x ∂σ y ∂ε y ∂σ z ∂ε z
= − o + ε x +σ x + ε y +σ y + ε z +σ z +
∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x
∂τ xy ∂ε xy ∂τ ∂ε ∂τ ∂ε
2 ε xy + τ xy + 2 ε yz yz + τ yz yz + 2 ε xz xz + τ xz xz
∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x
∂U o ∂C ∂ε x ∂ε y ∂ε z ∂ε
= − o + ε x + σ x + σ y + σ z + 2 τ xy xy +
∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x
∂ε yz ∂ε
2 τ yz + 2 τ xz xz
∂σ x ∂σ x
Since U o = U o (ε x , ε y , ε z , ε xy , ε yz , ε xz ) and ε x = f1 (σ x , σ y , σ z ,τ xy ,τ yz , τ xz ) , thus, by
∂U o ∂U o ∂ε x ∂U o ∂ε y ∂U o ∂ε z ∂U o ∂ε xy ∂U o ∂ε yz ∂U o ∂ε xz
= + + + + +
∂σ x ∂ε x ∂σ x ∂ε y ∂σ x ∂ε z ∂σ x ∂ε xy ∂σ x ∂ε yz ∂σ x ∂ε xz ∂σ x
∂U o ∂ε x ∂ε y ∂ε z ∂ε xy ∂ε yz ∂ε
=σx +σy + σz + 2τ xy + 2τ yz + 2τ xz xz
∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x ∂σ x
∂U o
Therefore, the terms in will cancel some terms of the left-handed side of the
∂σ x
previous equation, we have
∂C o
εx =
∂σ x
obtain
∂C o ∂C o 1 ∂C o 1 ∂C o 1 ∂C o
εy = εz = ε xy = ε yz = ε xz =
∂σ y ∂σ z 2 ∂τ xy 2 ∂τ yz 2 ∂τ xz
Fig. 2.7
2.3 Stress-Strain Relations and Strain-Stress Relations
The generalized Hooke’s law relates stresses to strains. Each of the stress components
is a linear function of the strain components.
σ x C11 C12 C13 C14 C15 C16 ε x
σ
y C 21 C 22 C 23 C 24 C 25 C 26 ε y
σ z C C 32 C 33 C 34 C 35 C 36 ε z
31
=
τ yz C 41 C 42 C 43 C 44 C 45 C 46 γ yz
τ C 51 C 52 C 53 C 54 C 55 C 56 γ xz
xz
τ xy C 61 C 62 C 63 C 64 C 65 C 66 γ xy
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-16
coefficients can be shown actually independent for elastic material when the strain energy is
considered. Using the relations between the stress components and the strain energy density,
we have
∂U o
σx = = C11ε x + C12ε y + C13ε z + C14γ yz + C15γ xz + C16γ xy
∂ε x
∂U o
σy = = C 21ε x + C 22 ε y + C 23ε z + C 24γ yz + C 25γ xz + C 26γ xy
∂ε y
∂U o
σz = = C 31ε x + C 32 ε y + C 33ε z + C 34γ yz + C 35γ xz + C 36γ xy
∂ε z
1 ∂U o ∂U o
τ yz = = = C 41ε x + C 42 ε y + C 43ε z + C 44γ yz + C 45γ xz + C 46γ xy
2 ∂ε yz ∂γ yz
1 ∂U o ∂U o
τ xz = = = C 51ε x + C 52 ε y + C 53ε z + C 54γ yz + C 55γ xz + C 56γ xy
2 ∂ε xz ∂γ xz
1 ∂U o ∂U o
τ xy = = = C 61ε x + C 62 ε y + C 63ε z + C 64γ yz + C 65γ xz + C 66γ xy
2 ∂ε xy ∂γ xy
Hence, we can show that the stiffness coefficients are symmetry, C ij = C ji , by using
appropriate differentiation.
∂ 2U o
= C12 = C 21
∂ε x ∂ε y
∂ 2U o
= C13 = C 31
∂ε x ∂ε z
M
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-17
∂ 2U o
= C16 = C 61
∂ε x ∂γ xy
M
∂ 2U o
= C 56 = C 65
∂γ xz ∂γ yz
With the foregoing reduction from 36 to 21 components, the stress-strain relations are
σ x C11 C12 C13 C14 C15 C16 ε x
σ
y C12 C 22 C 23 C 24 C 25 C 26 ε y
σ z C C 23 C 33 C 34 C 35 C 36 ε z
13
=
τ
yz C14 C 24 C 34 C 44 C 45 C 46 γ yz
τ C15 C 25 C 35 C 45 C 55 C 56 γ xz
xz
τ xy C16 C 26 C 36 C 46 C 56 C 66 γ xy
x
1
Fiber
2
y
Matrix
Fig. 2.8
If there is one plane of material property symmetry, the stress-strain relations reduce
to
σ x C11 C12 C13 0 0 C16 ε x
σ
y C12 C 22 C 23 0 0 C 26 ε y
σ z C C 23 C 33 0 0 C 36 ε z
13
=
τ
yz 0 0 0 C 44 C 45 0 γ yz
τ 0 0 0 C 45 C 55 0 γ xz
xz
τ xy C16 C 26 C 36 0 0 C 66 γ xy
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-18
where the plane of symmetry is z = 0 . Such a material is termed monoclinic materials having
13 independent stiffness coefficients.
If there are two orthogonal planes of material property symmetry for a material,
symmetry will exist relative to a third mutually orthogonal plane. Then, the stress-strain
relations reduce to
σ x C11 C12 C13 0 0 0 ε x
σ
y C12 C 22 C 23 0 0 0 ε y
σ z C C 23 C 33 0 0 0 ε z
13
=
τ yz 0 0 0 C 44 0 0 γ yz
τ 0 0 0 0 C 55 0 γ xz
xz
τ xy 0 0 0 0 0 C 66 γ xy
If the material is isotropic in which the mechanical properties of the material are
symmetric on an infinite number of plane, there are only 2 independent stiffness coefficients
and the stress-strain relations are
σ x C11 C12 C12 0 0 0 ε x
σ ε
y C12 C11 C12 0 0 0 y
σ z C C12 C11 0 0 0 ε z
12
=
τ
yz 0 0 0 (C11 − C12 ) / 2 0 0 γ yz
τ 0 0 0 0 (C11 − C12 ) / 2 0 γ
xz xz
τ xy 0 0 0 0 0 (C11 − C12 ) / 2 γ xy
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-19
Deformation characteristic
Let us consider Fig. 2.9. The deformation of the isotropic material is directional
independent. Application of the normal stress causes extension in the direction of the applied
stress and contraction in the perpendicular direction. In addition, shear stress causes only
shearing deformation.
The deformation of the anisotropic is directional dependent. Application of the normal
stress leads not only to extension in the direction of the stress and contraction perpendicular to
it, but to shearing deformation as well. Conversely, shearing stress causes extension and
contraction in addition to the distortion of shearing deformation.
The deformation of the orthotropic material subjected to the normal stress in the
principal material direction is similar to one of the isotropic material. However, due to the
different properties in the two principal directions, the contraction can be either more or less
than the contraction of a similarly loaded isotropic material with the same modulus of
elasticity in the direction of load. In addition, shearing stress causes shearing deformation, but
the magnitude of the deformation is independent of the various modulus of elasticity and
Poisson’s ratios.
Fig. 2.9
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-20
Monoclinic material
The strain-stress relations of the monoclinic material have 13 independent compliance
coefficients. For symmetry about the axis z = 0 ,
ε x S11 S12 S13 0 0 S16 σ x
ε
y S12 S 22 S 23 0 0 S 26 σ y
ε z S S 23 S 33 0 0 S 36 σ z
13
=
γ
yz 0 0 0 S 44 S 45 0 τ yz
γ 0 0 0 S 45 S 55 0 τ xz
xz
γ xy S16 S 26 S 36 0 0 S 66 τ xy
Orthotropic material
The strain-stress relations of the orthotropic material have 9 independent compliance
coefficients.
ε x S11 S12 S13 0 0 0 σ x
ε
y S12 S 22 S 23 0 0 0 σ y
ε z S S 23 S 33 0 0 0 σ z
13
=
γ yz 0 0 0 S 44 0 0 τ yz
γ 0 0 0 0 S 55 0 τ xz
xz
γ xy 0 0 0 0 0 S 66 τ xy
Isotropic material
The strain-stress relations of the isotropic material have 2 independent compliance
coefficients.
ε x S11 S12 S12 0 0 0 σ x
ε σ
y S12 S11 S12 0 0 0 y
ε z S S12 S11 0 0 0 σ z
12
=
γ yz 0 0 0 ( S11 − S12 ) / 2 0 0 τ yz
γ 0 0 0 0 ( S11 − S12 ) / 2 0 τ
xz xz
γ xy 0 0 0 0 0 ( S11 − S12 ) / 2 τ xy
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-22
Fig. 2.10
By integrating the strain energy density equation ( dU o = σ x dε x + σ y dε y + σ z dε z
+ τ xy dγ xy + τ yz dγ yz + τ xz dγ xz ), from the initial to the final state of strain of a body, the strain
1
( C12ε x + C 22 ε y + C 23ε z + C 24γ yz + C 25γ xz + C 26γ xy ) ε y ∫ kdk +
0
1
( C13ε x + C 23ε y + C 33ε z + C 34γ yz + C 35γ xz + C 36γ xy ) ε z ∫ kdk +
0
1
( C14ε x + C 24 ε y + C 34ε z + C 44γ yz + C 45γ xz + C 46γ xy ) γ yz ∫ kdk +
0
1
( C15ε x + C 25ε y + C 35ε z + C 45γ yz + C 55γ xz + C 56γ xy ) γ xz ∫ kdk +
0
1
( C16ε x + C 26 ε y + C 36ε z + C 46γ yz + C 56γ xz + C 66γ xy ) γ xy ∫ kdk
0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-23
1
Uo = ( C11ε x2 + C12ε x ε y + C13ε x ε z + C14 ε x γ yz + C15ε x γ xz + C16ε x γ xy )+
2
1
( C12ε y ε x + C 22ε y2 + C 23ε y ε z + C 24ε y γ yz + C 25ε y γ xz + C 26ε y γ xy )+
2
1
( C13ε z ε x + C 23ε z ε y + C 33ε z2 + C 34 ε z γ yz + C 35ε z γ xz + C 36ε z γ xy )+
2
1
( C14γ yz ε x + C 24γ yz ε y + C 34γ yz ε z + C 44γ yz2 + C 45γ yz γ xz + C 46γ yz γ xy )+
2
1
( C15γ xz ε x + C 25γ xz ε y + C 35γ xz ε z + C 45γ xz γ yz + C 55γ xz2 + C 56γ xz γ xy )+
2
1
( C16γ xy ε x + C 26γ xy ε y + C 36γ xy ε z + C 46γ xy γ yz + C 56γ xy γ xz + C 66γ xy2 )
2
In the index notation, we have
1
Uo = C ij ε i ε j
2
Isotropic and homogenous materials
As previously mentioned, a material is isotropic when the mechanical properties of the
material are symmetric on an infinite number of plane. In other words, its mechanical
properties are invariant under any rotation of coordinates. A material is homogenous when the
mechanical properties of the material are identical for every point in a body. In other words,
its mechanical properties are invariant under any translation of coordinates. Thus, if the
material of an elastic body is isotropic, the strain energy density depends only on the principal
strains ε 1 , ε 2 , and ε 3 , which are invariant under arbitrary rotation.
1 1 1
Uo = C11ε 12 + C12ε 1ε 2 + C13ε 1ε 3 +
2 2 2
1 1 1
C12ε 1ε 2 + C 22ε 22 + C 23ε 2 ε 3 +
2 2 2
1 1 1
C13ε 1ε 3 + C 23ε 2 ε 3 + C 3ε 32
2 2 2
Since the mechanical properties of the isotropic material are symmetric for all planes,
the naming of the principal axes is arbitrary. Thus, the isotropic material has only two distinct
coefficients.
C11 = C 22 = C 33 = C1
C12 = C 23 = C13 = C 2
1 1 1
Uo = C1ε 12 + C1ε 22 + C1ε 32 + C 2 ε 1ε 2 + C 2 ε 1ε 3 + C 2 ε 2 ε 3
2 2 2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-24
C1 2
Uo = (ε 1 + ε 22 + ε 32 ) + C 2 (ε 1ε 2 + ε 1ε 3 + ε 2 ε 3 )
2
Rewriting the term C 2 (ε 1ε 2 + ε 1ε 3 + ε 2 ε 3 ) and noting that
(ab + ac + bc) =
1
2
[
(a + b + c) 2 − (a 2 + b 2 + c 2 ) ]
Hence,
C2 C
C 2 (ε 1ε 2 + ε 1ε 3 + ε 2 ε 3 ) = (ε 1 + ε 2 + ε 3 ) 2 − 2 (ε 12 + ε 22 + ε 32 )
2 2
Then, the strain energy density can be rewritten as
C1 2 C C
Uo = (ε 1 + ε 22 + ε 32 ) + 2 (ε 1 + ε 2 + ε 3 ) 2 − 2 (ε 12 + ε 22 + ε 32 )
2 2 2
C1 − C 2
Let λ = C 2 and G = are elastic constants called Lame’s elastic coefficients.
2
Thus,
1
Uo = λ (ε 1 + ε 2 + ε 3 ) 2 + G (ε 12 + ε 22 + ε 32 )
2
In terms of the strain invariants, J 1 = ε 1 + ε 2 + ε 3 and J 2 = ε 1ε 2 + ε 2 ε 3 + ε 1ε 3 .
Rewriting the second term of the strain energy density equation, we have
G (ε 12 + ε 22 + ε 32 ) = G (ε 1 + ε 2 + ε 3 ) 2 − 2G (ε 1ε 2 + ε 1ε 3 + ε 2 ε 3 )
Then,
1 2
Uo = λJ 1 + GJ 12 − 2GJ 2
2
1
In general, J 1 = ε x + ε y + ε z and J 2 = ε x ε y + ε y ε z + ε z ε x − (γ xy2 + γ yz2 + γ xz2 ) , then,
4
1
Uo = λ (ε x + ε y + ε z ) 2 + G (ε x + ε y + ε z ) 2
2
1
− 2G (ε x ε y + ε y ε z + ε z ε x − (γ xy2 + γ yz2 + γ xz2 ))
4
1
Uo = λ (ε x + ε y + ε z ) 2 + G (ε x2 + ε y2 + ε z2 ) + 2G (ε x ε y + ε y ε z + ε x ε z )
2
1
− 2G (ε x ε y + ε y ε z + ε z ε x − (γ xy2 + γ yz2 + γ xz2 ))
4
1 1
Uo = λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 ))
2 2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-25
∂U o ∂U o ∂U o 1 ∂U o ∂U o 1 ∂U o ∂U o
Since σ x = , σy = , σz = , τ xy = = , τ yz = = ,
∂ε x ∂ε y ∂ε z 2 ∂ε xy ∂γ xy 2 ∂ε yz ∂γ yz
1 ∂U o ∂U o
τ xz = = , consequently,
2 ∂ε xz ∂γ xz
∂ 1 1
σx = [ λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) ]
∂ε x 2 2
σ x = λ (ε x + ε y + ε z ) + 2Gε x = λe + 2Gε x
∂ 1 1
σy = [ λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) ]
∂ε y 2 2
σ y = λ (ε x + ε y + ε z ) + 2Gε y = λe + 2Gε y
∂ 1 1
σz = [ λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) ]
∂ε z 2 2
σ z = λ (ε x + ε y + ε z ) + 2Gε z = λe + 2Gε z
∂ 1 1
τ yz = [ λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) ]
∂γ yz 2 2
τ yz = Gγ yz
∂ 1 1
τ xz = [ λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) ]
∂γ xz 2 2
τ xz = Gγ xz
∂ 1 1
τ xy = [ λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) ]
∂γ xy 2 2
τ xy = Gγ xy
where e ≅ J 1 is the classical small-displacement cubical strain. The stress invariants can be
related with the strain invariants by
I 1 = σ x + σ y + σ z = λe + 2Gε x + λe + 2Gε y + λe + 2Gε z
I 1 = 3λe + 2G (ε x + ε y + ε z )
I 1 = (3λ + 2G ) J 1
I 2 = λ (3λ + 4G ) J 12 + 4G 2 J 2
I 3 = λ2 (λ + 2G ) J 13 + 4λG 2 J 1 J 2 + 8G 3 J 3
Inverting the stress equations, we obtain
τ yz
γ yz = 2ε yz =
G
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-26
τ xz
γ xz = 2ε xz =
G
τ xy
γ xy = 2ε xy =
G
σ x = λ (ε x + ε y + ε z ) + 2Gε x = (λ + 2G )ε x + λε y + λε z
σ y = λ (ε x + ε y + ε z ) + 2Gε y = λε x + (λ + 2G )ε y + λε z
σ z = λ (ε x + ε y + ε z ) + 2Gε z = λε x + λε y + (λ + 2G )ε z
In matrix notation, we have
σ x (λ + 2G ) λ λ ε x
σ y = λ (λ + 2G ) λ ε y
λ λ (λ + 2G ) ε z
σ z
−1
ε x (λ + 2G ) λ λ σ x
ε y = λ (λ + 2G ) λ σ y
λ λ (λ + 2G )
ε z σ z
This matrix inversion can be performed by hand or by computer program such as
Mathematica, and we obtain.
εx =
1
E
[
σ x − ν (σ y + σ z ) ]
εy =
1
E
[
σ y − ν (σ x + σ z ) ]
εz =
1
E
[
σ z − ν (σ x + σ y ) ]
G (3λ + 2G ) λ E
where E = ,ν = , and G = . In matrix notation,
λ +G 2(λ + G ) 2(1 + ν )
1 ν ν
E − − 0 0 0
E E
σ
ε x − ν 1
−
ν
0 0 0 x
ε E E E σ y
y ν ν 1
ε z − − 0 0 0 σ
E E E z
=
γ
yz 0 0 0
1
0 0 τ yz
γ G τ xz
xz 1
γ xy 0 0 0 0 0 τ
G xy
1
0 0 0 0 0
G
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-27
The strain energy density of isotropic material can be found by substituting the strains
1 1
into the equation Uo = λ (ε x + ε y + ε z ) 2 + G ( ε x2 + ε y2 + ε z2 + (γ xy2 + γ yz2 + γ xz2 )) and
2 2
rearranging the terms.
1 2ν 1 2
Uo = (σ x2 + σ y2 + σ z2 ) − (σ xσ y + σ y σ z + σ xσ z ) + (τ xy + τ yz2 + τ xz2 )
2E E 2G
or,
Uo =
1
2
[
σ x ε x + σ y ε y + σ z ε z + τ xy γ xy + τ yz γ yz + τ xz γ xz ]
2.5 Stress-Strain Relations for Isotropic Material: Physical Derivation
Consider a cubic volume element subjected to a state of triaxial normal stress σ x , σ y ,
σ z and associated normal strains ε x , ε y , and ε z are developed in the material. Since the
material is isotropic, the cubic volume element will deform to a rectangular block, no shear
strains is produced in the material. By using the principle of superposition, the deformation of
the cubic volume element subjected to each normal stress can be draw as shown in Fig. 2.11.
Fig. 2.11
First, consider the normal strain of the element in the x direction, caused by separate
application of each normal stress σ x , σ y , and σ z . Under σ x , the cubic volume element
σz
ε ′x′′ = −ν
E
Superimposing these three normal strains, we have the total normal stress σ x of the
εx =
1
E
[
σ x − ν (σ y + σ z ) ]
Likewise, the normal strain in the y and z direction can be determined as
εy =
1
E
[
σ y − ν (σ x + σ z ) ]
εz =
1
E
[
σ z − ν (σ x + σ y ) ]
In the test, application of the shear stress τ xy to the cubic volume element of the
isotropic material only produces the shear strain γ xy in the element as shown in Fig. 2.12a.
Likewise, the shear stresses τ yz and τ zx only produce the shear strain γ yz and γ xz on the
Fig. 2.12
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-29
Example 2-2
The thin-walled cylindrical pressure vessel 10 m long as shown in Fig. Ex2-2 has
closed ends, a wall thickness of 5 mm, and an inner diameter of 3 m. If the vessel is filled
with air to a pressure of 2 MPa , how much do the length, diameter, and wall thickness
change, and in each case is the change an increase or a decrease? The vessel is made of steel
having E = 200 GPa , and ν = 0.30 .
Fig. Ex2-2
Let the x -axis is along the longitudinal axis of the vessel and the z -axis is normal to
the surface. Thus, the y -axis is in the tangential direction.
Since the ratio of the radius to the thickness, r / t , is small, thus,
pr 2(1.5)
σx = = = 300 MPa
2t 2(0.005)
pr 2(1.5)
σy = = = 600 MPa
t (0.005)
The value of σ z varies from − p on the inside wall to zero on the outside wall, thus,
σ z ≈ 0 and we have
1
εx = [300 − 0.3(600 + 0)] = 0.00060
200(10 3 )
1
εy = [600 − 0.3(300 + 0)] = 0.00255
200(10 3 )
1
εz = [0 − 0.3(300 + 600)] = −0.00135
200(10 3 )
∆L ∆(πd ) ∆d ∆t
Since ε x = , εy = = , and ε z = ,
L πd d t
∆L = 0.00060(10)10 3 = +6 mm
∆d = 0.00255(3)10 3 = +7.65 mm
∆t = −0.00135(5) = −6.75(10 −3 ) mm
Thus, there are small increases in length and diameter, and a tiny decrease in the wall
thickness.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-30
Example 2-3
A sample of material subjected to a compressive stress σ z is confined so that it can
not deformed in the y -direction, but deformation is permitted in the x -direction, as shown in
Fig. Ex 2-3. Assuming that the material is isotropic and exhibits linear-elastic behavior.
Determine the following in term of σ z and the elastic constants of the material:
a.) The stress that develops in the y -direction.
b.) The strain in the z -direction.
c.) The strain in the x -direction.
d.) The stiffness E ′ = σ z / ε z in the z -direction. Is this apparent modulus equal to the
elastic modulus E from the uniaxial test on the material? Why or why not?
Fig. Ex 2-3
Since the sample can not deformed in the y -direction, ε y = 0 , and since the
εy =
1
E
[
σ y − ν (σ x + σ z ) ]
0=
1
E
[
σ y − ν (0 + σ z ) ]
σ y = νσ z
The strain in the z -direction is
εz =
1
E
[
σ z − ν (σ x + σ y ) ]
1
εz = [σ z − ν (0 + νσ z )]
E
1 −ν 2
εz = σz
E
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 2-31
εx =
1
E
[
σ x − ν (σ y + σ z ) ]
1
εx = [0 − ν (νσ z + σ z )]
E
ν (1 + ν )
εx = − σz
E
The stiffness E ′ = σ z / ε z is
E
E′ =
1 −ν 2
Thus, the apparent stiffness differs from the elastic modulus E from the uniaxial test. This is
due to the fact that the apparent stiffness is determined by behavior according to the three-
dimensional form of Hooke's law.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-1
Chapter 3
Elements of Theory of Elasticity
3.1 Introduction
In the analysis of a body or structure, the geometry of the structure and the loads are
given. A solution may be obtained by analytical, numerical, and experimental methods. In the
analytical methods, the derivation of the load-stress relations depends on the following
conditions:
1. The equilibrium equations.
2. The compatibility equations.
3. The stress-strain relations.
4. The material responses.
Two different analytical methods used to satisfy the first and second condition are the
method of mechanics of materials and the theory of elasticity method.
The mechanics of materials involve the following steps.
1. The simplified assumptions related to the geometry of the deformation of the
structure are established by using the compatibility equations.
2. Analyze the geometry of the deformation to determine the strain distributions over
a cross section of the structure.
3. Relate the applied loads to the internal stress by using the equilibrium equations.
4. Use the stress-strain relations and the material responses to determine the relations
between the assumed strain distribution and stress distribution over a cross section
of the structure.
5. Relate the applied loads to the displacement of the structure.
The obtained results may be exact, or good approximations, or rough estimate,
depending largely on the accuracy of the assumptions made in the first step.
In the theory of elasticity, the states of stresses and displacements for every point in
the structure are determined by simultaneously satisfy the requirements of equilibrium at
every point, compatibility of all displacements and boundary conditions on stress and
displacement. This method involves no initial assumptions or approximation about the
geometry of the deformation. Thus, the method is more difficult than the mechanics of
materials. However, it is usually used to solve the problems in which the geometry of the
deformation can not be reliably anticipated such as determining the stress concentration
occurred at a hole in a plate.
Often, a practical problem is solves by using both methods simultaneously.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-2
Limitations
In this chapter, we are considering that the material is homogeneous, isotropic, and
linearly elastic, and that the displacements are small.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-3
Fig. Ex 3-1
To determine the states of stresses for every point in the bar shown in Fig. Ex 3-1a, we
are assuming that the bar is in a state of uniaxial stress. This idealization is not exact since
the state of stress at the fixed end of the bar is in three dimensions due to the Poisson’s
effects. However, according to Saint-Venant’s principle, the effects are localized.
1. By using the equilibrium of forces in the longitudinal axis of the differential
element shown in Fig. Ex 3-1c, we have
− Aσ x + A(σ x + dσ x ) + q( x)dx = 0
dσ x q ( x)
+ =0
dx A
This equation is a differential equation of equilibrium, which must be satisfied for
every point in the bar from x = 0 to x = L . Let q = qx and the bar is prismatic, the states of
stresses can be determined by integrating the previously obtained equation.
x
1 1 qx 2
A ∫0
σx = qx dx = + C1
A 2
The integration constant C1 can be found by using the boundary condition of stress: at
qL2
x = 0 , σ x = 0 . Hence, C1 = − and
2A
qx 2 qL2 q 2
σx = − = ( x − L2 )
2A 2A 2A
2. To determine the axial displacement, we need first consider how the differential
element deformed under the load q = qx and, then, obtain the strain-displacement relation.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-4
(u + du ) − u du
εx = =
dx dx
The uniaxial stress-strain relation, σ x = Eε x .
du
σx = E
dx
Substituting σ x into the differential equation of equilibrium, we have
d 2 u qx
E + =0
dx 2 A
Performing the first integration with respect to x and using the displacement
du
boundary condition: at x = L , = 0 , we obtain
dx
x
du 1 qx 2
E = − ∫ qx dx = − + C2
dx A0 2A
qL2
C2 =
2A
du q 2
E == (L − x 2 )
dx 2A
Performing the second integration with respect to x and using the displacement
boundary condition: at x = 0 , u = 0 , we obtain
q 2 x3
x
q
2 A ∫0
Eu = ( L − x )dx =
2 2
L x − + C3
2A 3
C3 = 0
qx 2 x 2
u= L −
2 AE 3
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-5
points are approximately equal to zero. This kind of state of stresses on the infinitesimal small
element is called the plane stress. The strain-stress relations of the plane stress are
Fig. 3.1
1 ν ν
E − − 0 0 0
E E
ε x − ν 1
−
ν
0 0 0 σ x
ε E E E σ y
y ν ν 1
ε z − − 0 0 0 0
E E E
=
γ
yz 0 0 0
1
0 0 0
γ G 0
xz 1
γ xy 0 0 0 0 0 τ xy
G
1
0 0 0 0 0
G
ε x 1 −ν 0 σ x
1
ε y = − ν 1 0 σ y
E 0 2(1 + ν ) τ xy
γ xy 0
ν
εz = − (σ x + σ y )
E
E
G=
2(1 + ν )
and the stress-strain relations of the plane stress are
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-6
σ x 1 ν 0 ε x
E ν 1 ε
σ y = 0 y
1 −ν
2
τ xy 0 0 (1 − ν ) / 2 γ xy
+ =
∂x 2 ∂y 2 ∂x∂y
The differential equations of equilibrium of the plane stress are
∂σ x ∂τ xy
+ + Bx = 0
∂x ∂y
∂τ xy ∂σ y
+ + By = 0
∂x ∂y
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-7
Example 3-2
Given that the body force is negligible. Investigate if the following displacement field
can be a solution of a static plane stress problem
u = a1 ( x 2 − y 2 ) − a 2 y + a3 and v = 2a1 xy + a 4
∂u ∂v
γ xy = + = (−2a1 y − a 2 ) + 2a1 y = −a 2
∂y ∂x
Boundary condition
Boundary condition is the prescribed condition of displacements and forces at the
boundaries of a structure. Consider a body as shown in Fig. 3.2.
Fig. 3.2
1. Geometric or displacement boundary condition pertains to the compatibility
conditions and required that the displacement (including rotation) at the structural
boundary must be satisfied.
Along the boundary AB which is the fixed support, the displacements u and v are
zero.
2. Natural or force boundary condition pertains to the equilibrium conditions and
required that the forces (including moment) at the structural boundary must be
satisfied.
v
From chapter 1, the stress vector σ P on an arbitrary plane in three-dimension is
v
σ P = σ Px iˆ + σ Py ˆj + σ Pz kˆ
where
σ Px = σ x l + τ xy m + τ xz n
σ Py = τ xy l + σ y m + τ yz n
σ Pz = τ xz l + τ yz m + σ z n
From Fig. 3.2, the direction cosine along the boundary BC and AC is l = cos φ ,
m = sin φ , and n = 0 . Thus,
σ Px = σ x cos φ + τ xy sin φ
σ Py = τ xy cos φ + σ y sin φ
( pdA) cos φ
Φx = − = − p cos φ
dA
dFy ( pdA) sin φ
Φy = =− = − p sin φ
dA dA
The force boundary conditions along the boundary BC are
σ Px = Φ x ; − p cos φ = σ x cos φ + τ xy sin φ
At φ = 0 , σx = −p
τ xy = 0
π
At φ = , σ y = −p
2
τ xy = 0
The force boundary conditions along the boundary BC are
At φ = π , σx = 0
τ xy = 0
σy =0
The elasticity solutions of this example must satisfy the above displacement and force
boundary conditions.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-10
Example 3-3
The cantilevered beam as shown in Fig. Ex 3-3 has unit thickness. Neglecting the
body forces, determine
a.) The flexural stress σ x due to the applied load by using the mechanics of material
method of analysis.
b.) The stresses σ y and γ xy by using the obtained the flexural stress σ x .
c.) If the obtained state of stresses is possible for the theory of elasticity method of
analysis?
Fig. Ex 3-3
By the mechanics of material method of analysis, we have
qx 2
y
My 2 3q
σx = = = 3 x2 y
I 2c
3
4c
3
To find τ xy , we use the differential equation of equilibrium of the plane stress where
3q ∂τ xy
xy + =0
2c 3 ∂y
3q
τ xy = − ∫ xydy + C1
2c 3
3q
τ xy = − 3
xy 2 + C1
4c
Stress boundary condition, at y = + c , τ xy = 0 .
3q 2
C1 = xc
4c 3
3q
τ xy = 3
x (c 2 − y 2 )
4c
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-11
To find σ y , we use the differential equation of equilibrium of the plane stress where
3q 2 ∂σ y
(c − y 2 ) + =0
4c 3
∂y
3q 2
σ y = −∫ (c − y 2 )dy + C 2
4c 3
3q 2 y3
σy =− ( c y − ) + C2
4c 3 3
Stress boundary condition, at y = + c , σ y = − q .
3q 3 c 3
−q = − (c − ) + C 2
4c 3 3
q
C2 = −
2
q y3 3y
σy = − − 1
2 2c 3 2c
+ =
∂x 2 ∂y 2 ∂x∂y
But, we need to find the strains from the stresses first. For the plane stress problem,
3q 2
4c 3 x y
ε x 1 −ν 0
1 q y 3 3 y
ε y = − ν 1 0 3 − − 1
E 0 2(1 + ν )
2 2c 2 c
γ xy 0
3q
3 x (c 2 − y 2 )
4c
Then, substituting the obtained strains into the strain compatibility equation, we get
∂ 2ε y ∂ 2 ε x ∂ γ xy
2
+ ≠
∂x 2 ∂y 2 ∂x∂y
Therefore, the state of stresses is impossible for the theory of elasticity method of analysis.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-12
∂3F ∂3F ∂V ∂V
− 2 + 2 + − =0
∂x ∂y ∂x ∂y ∂y ∂y
Therefore, as far as the equations of equilibrium are concerned, the assumed stress functions
constitute a general solution. The problem is reduced to finding a function F = F ( x, y ) that
satisfies the compatibility conditions and the boundary conditions.
The strain-stress relations of the plane stress problems are
∂ 2 F
2 +V
∂y
ε x 1 −ν 0 2
1
∂ F
ε y = − ν 1 0 2 + V
E 0 ∂x
2(1 + ν )
γ xy 0
∂2F
−
∂x∂y
∂ 2ε y ∂ 2 ε x ∂ γ xy
2
∂ 2ε x 1 ∂ 4 F ∂4F ∂ 2V
= 4 − ν 2 2 + (1 − ν ) 2
∂y 2 E ∂y ∂x ∂y ∂y
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-13
∂ 2ε y 1 ∂4F ∂4F ∂ 2V
= − ν + + (1 − ν )
∂x 2 E ∂x 2 ∂y 2 ∂x 4 ∂x 2
∂ 2γ xy 2(1 + ν ) ∂ 4 F
=−
∂x∂y E ∂x 2 ∂y 2
The only compatibility equation not identically satisfied is
∂4F ∂4F ∂4F ∂ 2V ∂ 2V
4 + 2 + + (1 − ν ) 2 + 2=0
∂x ∂x 2 ∂y 2 ∂y 4 ∂x ∂y
∇ 4 F + ∇ 2V = 0
Now, the problem is reduced to the solution of boundary-value problem associated
with biharmonic differential operator ∇ 4 F and harmonic differential operator ∇ 2V which
can be solved by using theory of functions of a complex variable.
If the body forces are absent,
∇4F = 0
The function F = F ( x, y ) is called the Airy stress function which is discovered by Sir
George Biddell Airy in 1863.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-14
Example 3-4
A rectangular block having the mass density ρ stands on a rigid horizontal support
and is loaded by its own weight as shown in Fig. Ex 3-4a. As a proposed solution of this
problem, investigate the equation
y 3 Lx 2
F = − ρg + and V = ρgy
6 2
Determine the displacement u = u ( x, y ) and v = v( x, y ) of the block due to its own weight.
Fig. Ex 3-4a
From the Airy stress function, we obtain
∂2F
σ x = 2 + V = − ρgy + ρgy = 0
∂y
∂2F
σy = + V = − ρgL + ρgy = ρg ( y − L)
∂x 2
∂2F
τ xy = − =0
∂x∂y
The results satisfy the free-surface boundary conditions of the top and the vertical
edges of the block where, for the top edge,
σx = 0
σ y = ρg ( L − L) = 0
τ xy = 0
, for the vertical edge,
σx = 0
σ y = ρg ( y − L)
τ xy = 0
, and for the bottom edge,
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-15
σx = 0
σ y = ρg (0 − L) = − ρgL
τ xy = 0
Determine the displacement u = u ( x, y ) and v = v( x, y ) .
From the stress-strain relations,
ε x 1 −ν 0 0
1
ε y = − ν 1 0 ρg ( y − L )
E 0 2(1 + ν ) 0
γ xy 0
∂u ν
εx = = − ρg ( y − L )
∂x E
∂v ρg
εy = = ( y − L)
∂y E
∂u ∂v
γ xy = + =0
∂y ∂x
Thus, by integration, the displacement functions are in the form of
ν
u = − ρg ( y − L) x + f y
E
ρg y 2
v= ( − Ly ) + f x
E 2
∂u ∂v
Substituting u = u ( x, y ) and v = v( x, y ) into γ xy = + =0
∂y ∂x
ν df y df x
− ρg x + + =0
E dy dx
ν df x df y
− ρg E x + dx + dy = 0
Each expression in the parentheses must be a constant. If not, we could vary x alone
(or y alone) and violate the equality. Thus,
ν df
− ρg x + x = a1
E dx
df y
= − a1
dy
ν ν x2
and f x = ∫ a1 + ρg x dx = a1 x + ρg + a2
E E 2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-16
f y = − ∫ a1 dy = − a1 y + a3
ρg y 2 ν x2 ρg y 2 ν 2
v= ( − Ly ) + a1 x + ρg + a2 = − Ly + x + a1 x + a 2
E 2 E 2 E 2 2
By using the boundary conditions of the block and the symmetry, at x = 0 and y = 0 ,
u = 0 and v = 0 , we obtain
a 2 = 0 and a3 = 0
∂v
At x = 0 and y = 0 , =0,
∂x
∂v ρgν
= x + a1 = 0
∂x E
ρgν
a1 = − x
E
Hence, we get the displacement functions in the form of
ν
u = − ρg ( y − L − 1) x
E
ρg y 2 ν 2
v= − Ly + x − νx
E 2 2
Fig. Ex 3-4b
The deformed shape of the block is shown in Fig. Ex 3-4b. It should be noted that
1.) The block shortens vertically and become wider toward the base due to Poisson's
effect.
2.) All right angles are preserved since γ xy = 0 .
3.) The deflection at the base is incompatible with the rigid horizontal support
4.) According to the Saint-Venant's principle, the solution should be exact for y > 2c .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-17
∂2F
σy = = a2
∂x 2
∂2F
τ xy = − = −b2
∂x∂y
These stress boundary conditions are constant throughout the body. Thus, the stress
a2 2 c
function F = x + b2 xy + 2 y 2 represents a combination of uniform tensions or
2 2
compressions in two perpendicular directions and a uniform shear as shown in Fig. 3.3.
Fig 3.3
Let consider a stress function in the form of a cubic function.
a3 3 b3 2 c d
F= x + x y + 3 xy 2 + 3 y 3
3(2) 2 2 3(2)
where a3 , b3 , c3 , and d 3 are constants. This Airy stress function is satisfied the equation
∂2F
σx = = c3 x + d 3 y
∂y 2
∂2F
σ y = 2 = a3 x + b3 y
∂x
∂2F
τ xy =− = −b3 x − c3 y
∂x∂y
If all coefficients except d 3 are zero and d 3 = 6a1 = a constant,
σ x = 6a1 y σy =0 τ xy = 0
The meaning and the usefulness of the obtained stress boundary condition depend on
the region that we are choosing to consider. Let us consider Fig. 3-4.
Fig. 3.4
1. If we choose the region as shown in Fig. 3.4a, the stress boundary condition
represents a state of the normal stresses due to a pure bending applied at the ends
of the beam.
2. If we choose the region as shown in Fig. 3.4b, the stress boundary condition
represents a state of the normal stresses due to bending plus axial load applied to
the ends of the beam.
3. If we choose the region as shown in Fig. 3.4c, the solution has no practical
interest.
In taking the stress function in the form of quadratic and cubic polynomial equations,
we are completely free in choosing the magnitudes of the coefficients, since the equation
∇ 4 F = 0 is always satisfied whatever values they may have. In the case of polynomials of
higher degrees, the equation ∇ 4 F = 0 is satisfied only if certain relations between the
coefficients are satisfied.
If all coefficients except b3 are zero,
∂2F ∂2F
σx = 0 σy = = b3 y τ xy = − = −b3 x
∂x 2 ∂x∂y
The stress boundary conditions in this case are shown in Fig. 3-5.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-19
Fig 3.5
Let us consider the stress function in the form of a polynomial of the fourth degree,
a4 4 b c d e
F= x + 4 x 3 y + 4 x 2 y 2 + 4 xy 3 + 4 y 4
4(3) 3(2) 2 3(2) 4(3)
∂2F
σy = = a 4 x 2 + b4 xy + c 4 y 2
∂x 2
∂2F b d
τ xy = − = − 4 x 2 − 2c 4 xy − 4 y 2
∂x∂y 2 2
If all coefficients except d 4 are zero and d 4 = 6φ 4 = a constant,
σ x = 6φ 4 xy σy =0 τ xy = −3φ 4 y 2
Fig. 3.6
Fig. 3.6a shows the stress boundary condition associated with the solution on the
chosen region. The solution appears to lack practical interest.
Now, let us remove the shear stress on y = ± c by superimposing the stress function
F = φ 4 ( xy 3 − 3c 2 xy )
This stress function is satisfied the equation ∇ 4 F = 0 and the stress components are
∂2F
σx = = 6φ 4 xy
∂y 2
∂2F
σy = 2 =0
∂x
∂2F
τ xy =− = 3φ 4 (c 2 − y 2 )
∂x∂y
Fig.3.6b shows the stress boundary condition associated with the solution on the
chosen region. They appear to be the stress distribution corresponds to a cantilever beam
having fixed support at x = L and is subjected to both a parabolic distribution of τ xy and pure
bending at x = 0 . Note that, according to the Saint-Venant’s principle, the obtained elasticity
solution must be considered as approximate near the fixed support and loading point
( 2c < x < L − 2c ) due to the Poisson’s effect and stress concentration.
c
2c 3
P = ∫ τ xy dy = 3φ 4 2c 3 − = 4φ 4 c
3
−c 3
My ( Px) y
σx = = = 6φ 4 xy
I 1(2c) 3 / 12
3.5 End Effects
In some problems in which the structure is subjected to loads producing stress
concentrations in the area of loading point, we need to use the Saint-Venant’s principle to
simplify the problems by replacing the loads with a statically equivalent load.
Fig 3.7
Consider a cantilever beam having a unit width as shown in Fig. 3.7a. The
concentrated load P makes the problem difficult to solve since it is difficult to satisfy the
force boundary conditions at the loading point. Hence, the load P is replaced by a quadratic
distribution of surface traction Φ y = −τ xy that is zero at y = ± c and maximum at y = 0 as
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-21
shown in Fig. 3.7b. According to the Saint-Venant’s principle, these two different loading
patterns practically produce no different effect to the beam at a distance of 2c or more.
In addition, the displacement boundary condition at the fixed end of the beam is u = 0
and v = 0 for − c < y < c . This condition makes the elasticity solution difficult to obtain. To
make the problem easier, we assume the initial vertical line at x to warp into the shapes as
shown in Fig. 3.7c. This deformation is due to the shear stress τ xy . Due to the fixed support at
x = L , we may
∂v
1. Set = 0 at y = 0 to make the beam axis horizontal at x = L .
∂x
∂u
2. Set = 0 at y = 0 to make the beam vertical line vertical at x = L .
∂y
3.6 Determination of Displacements from Stresses
If the stress functions are known, we can determine the displacement by integration of
the strain-displacement relations as shown in the following example.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-22
y = 0 . This condition can be satisfied by using the stress components found by using the Airy
stress function in the form of
σ x = 6φ 4 xy σy =0 τ xy = 3φ 4 (c 2 − y 2 )
The constant φ 4 can be found by using the force boundary condition: sum of the
shearing force distribution over the end of the beam must be equal to the load P .
c
2c 3
P = ∫ τ xy dy = 3φ 4 2c 3 − = 4φ 4 c
3
−c 3
P
φ4 =
4c 3
2c 3
Since the moment of inertial of the cross section having a unit thickness is , the
3
stress components equations can be written as
3P Pxy
σx = 3
xy =
2c I
3P y2 P 2
τ xy = (1 − 2 ) = (c − y 2 )
4c c 2I
These stresses can also be obtained directly by using the method of mechanics of
material where
My VQ
σx = and τ xy =
I It
Now, let determine the displacement corresponding to the stresses. Using Hooke’s law
for plane stress, we have
ε x 1 / E − ν / E 0 Pxy / I Pxy /( EI )
ε y = − ν / E 1 / E 0 0 = − νPxy /( EI )
0 1 / G P(c − y ) /(2 I ) P (c − y ) /( 2 IG )
γ xy
0 2 2 2 2
Px 2 y νPxy 2
u= + fy v=− + fx
2 EI 2 EI
where f y and f x are unknown functions of y only and x only, respectively. They may not
be a constant since we are dealing with partial derivative. Substituting the values of u and v
into the shear strain-displacement relation,
Px 2 df y νPy 2 df x P
γ xy = + − + = (c 2 − y 2 )
2 EI dy 2 EI dx 2 IG
Px 2 df x νPy 2 df y Py 2 Pc 2
+ + − + + =
2 EI dx 2 EI dy 2 IG 2 IG
Px 2 df x
In this equation, we can see that F ( x) = + are functions of x only, G ( y ) =
2 EI dx
νPy 2 df y Py 2 Pc 2
− + + are functions of y only, and K = is independent of x and y .
2 EI dy 2 IG 2 IG
F (x) + G ( y ) = K
This equation means that F (x) must be some constant A and G ( y ) must be some
constant B . If not, F (x) and G ( y ) would vary with x and y , respectively. In addition, by
varying x alone or y alone, the equality would be violated. Hence,
Pc 2
A+ B =
2 IG
df x Px 2
= A−
dx 2 EI
df y νPy 2 Py 2
= B+ −
dy 2 EI 2 IG
df x df y
Integrating and , we have
dx dy
Px 3
fx = − + Ax + C
6 EI
νPy 3 Py 3
fy = − + By + D
6 EI 6 IG
Substituting into the expressions of u and v , we have
Px 2 y νPy 3 Py 3
u= + − + By + D
2 EI 6 EI 6 IG
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-24
νPxy 2 Px 3
v=− − + Ax + C
2 EI 6 EI
Assuming that the centroid of the cross section is fixed. Then, the expressions of u
and v are zero at x = L and y = 0 .
D=0
PL3
C= − AL
6 EI
The deflection curve of the beam at y = 0 is
Px 3 PL3
v y =0 = − + − A( L − x)
6 EI 6 EI
From the discussion about the end effects, due to the fixed support at x = L , we may
∂v
1. Set = 0 at y = 0 to make the beam axis horizontal at x = L .
∂x
∂u
2. Set = 0 at y = 0 to make the beam vertical line vertical at x = L .
∂y
∂v
For the first case = 0 at y = 0 and x = L ,
∂x
∂v PL2
=− +A=0
∂x 2 EI
PL2
A=
2 EI
Pc 2
Then, from the equation A + B = ,
2 IG
Pc 2 PL2
B= −
2 IG 2 EI
PL2 Pc 2 PL2 PL3 PL3
Substituting all constants A = , B= − , C= − AL = − , D=0
2 EI 2 IG 2 EI 6 EI 3EI
into the expressions of u and v , we have
Px 2 y νPy 3 Py 3 Pc 2 PL2
u= + − + − y
2 EI 6 EI 6 IG 2 IG 2 EI
Px 3 PL2 x PL3
v y =0 = − + −
6 EI 2 EI 3EI
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-25
PL3
v y =0 =−
x =0 3EI
which is identical to the one obtained by using the mechanics of materials method.
To show the warping of the beam cross section produced by the shearing stress, let us
consider the horizontal displacement u at the support x = L .
νPy 3 Py 3 Pc 2 y
u x=L = − +
6 EI 6 IG 2 IG
∂u νPy 2 Py 2 Pc 2
= − +
∂y x = L 2 EI 2 IG 2 IG
∂u Pc 2 3P
= =
∂y y =0 2 IG 4cG
x = L
3P
This rotation of the cross section is due to the shearing stress τ xy = and having the
4c
clockwise direction.
∂u
Consider the second case where = 0 at y = 0 to make the beam vertical line
∂y
vertical at x = L , we have
∂u ∂ Px 2 y νPy 3 Py 3
= + − + By + D = 0
∂y ∂y 2 EI 6 EI 6 IG
Px 2 νPy 2 Py 2
+ − +B=0
2 EI 2 EI 2 IG
Px 2 νPy 2 Py 2
B=− − +
2 EI 2 EI 2 IG
At y = 0 and x = L ,
PL2
B=−
2 EI
Pc 2
Then, from the equation A + B = ,
2 IG
Pc 2 PL2
A= +
2 IG 2 EI
PL3 Pc 2 L PL3
Substituting into the expressions of v at y = 0 with C = − AL = − − ,
6 EI 2 IG 3EI
we have
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-26
Px 3 Pc 2 x PL2 x Pc 2 L PL3
v y =0 = − + + − −
6 EI 2 IG 2 EI 2 IG 3EI
Px 3 Pc 2 x PL2 x PL3
v y =0 = − − ( L − x) + −
6 EI 2 IG 2 EI 3EI
Px 3 PL2 x PL3
Comparing this equation with the equation v y =0 = − + − previously
6 EI 2 EI 3EI
obtained, we have the deflection of the cantilevered beam increased by
Pc 2 x 3P
( L − x) = ( L − x)
2 IG 4cG
This term is an estimate of the effect of shearing force on the deflection of the beam.
In reality, the cross-section near the fixed support is not free to rotate. Thus, the distribution
of stresses is different from the obtained results. However, if the beam is long compared to the
depth, the results are satisfactory.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-27
Fig. 3.8
Equilibrium Equations in Polar Coordinate
Fig. 3.9
Consider an infinitesimal small element subjected to the state of stresses in the polar
coordinate as shown in Fig. 3.9. If this element is in equilibrium, we have the summations of
the forces in the radial direction and circumferential direction are equal to zero.
∑F r = 0;
dθ ∂σ r
− σ r (rdθ ) − τ rθ (dr ) − σ θ sin (dr ) + σ r + dr (r + dr )dθ +
2 ∂r
∂τ rθ ∂σ θ dθ
τ rθ + ∂θ dθ dr − σ θ + ∂θ dθ sin 2 (dr ) + Br rdrdθ = 0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-28
∑ Fθ = 0;
dθ ∂τ
− τ rθ (rdθ ) − σ θ (dr ) − τ rθ sin (dr ) + τ rθ + rθ dr (r + dr )dθ +
2 ∂r
∂σ θ ∂τ rθ dθ
σ θ + ∂θ dθ dr − τ rθ + ∂θ dθ sin 2 (dr ) + Bθ rdrdθ = 0
dθ dθ
Expanding the above two equations, setting sin = , and neglecting the higher-
2 2
order terms, we have the equilibrium equations in the polar coordinate.
∂σ r 1 ∂τ rθ σ r − σ θ
+ + + Br = 0
∂r r ∂θ r
1 ∂σ θ ∂τ rθ τ
+ + 2 rθ + Bθ = 0
r ∂θ ∂r r
Airy Stress Function
To solve the plane problems in polar coordinate, we begin by finding stresses that
satisfy the equations of equilibrium. However, the two equilibrium equations have three
unknowns stresses. Hence, The number of the possible stress solutions is infinite.
By using coordinate transformation from the Cartesian coordinates ( x , y ) to the polar
coordinates ( r , θ ), we have
σ r m
2
n2 2mn σ x
2
σ θ = n m 2 − 2mn σ y
τ − mn mn m 2 − n 2
rθ τ xy
By taking the inverse of the matrix, we have
σ x m 2 n2 − 2mn σ r
2
σ y = n 2mn σ θ
2
m
mn − mn m 2 − n 2 τ
τ xy rθ
The relations between the Cartesian coordinates ( x , y ) to the polar coordinates ( r ,
θ ) are
r 2 = x2 + y2
y
θ = arctan
x
The derivatives of the relation with respect to x and y are
∂r x ∂r y
= = cosθ = = sin θ
∂x r ∂y r
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-29
∂θ y sin θ ∂θ x cosθ
=− 2 =− = 2 =
∂x r r ∂y r r
Thus, for any function f ( x, y ) that can be written in the polar coordinate as
f (r cosθ , r sin θ ) , we have the first partial derivative of the function f ( x, y ) as
∂f ∂f ∂r ∂f ∂θ ∂f sin θ ∂f
= + = cosθ −
∂x ∂r ∂x ∂θ ∂x ∂r r ∂θ
Also, the second partial derivative of the function f ( x, y ) can be written as
∂2 f ∂ sin θ ∂ ∂f sin θ ∂f
= cosθ − cosθ −
∂x 2
∂r r ∂θ ∂r r ∂θ
∂2 f ∂2 f ∂ 1 ∂f sin θ ∂ ∂f
= cos 2
θ − sin θ cosθ − cosθ
∂x 2
∂r 2
∂r r ∂θ r ∂θ ∂r
sin θ ∂ ∂f
+ sin θ
r ∂θ
2
∂θ
Rearranging the expressions,
∂2 f ∂2 f ∂ 1 ∂f ∂ 1 ∂f 1 ∂f
= cos 2
θ − sin θ cosθ − sin θ cosθ + sin θ cosθ 2 +
∂x 2
∂r 2
∂r r ∂θ ∂θ r ∂r r ∂θ
1 ∂f 1 ∂2 f
sin 2 θ + 2 2
r ∂r r ∂θ
1 ∂f 1 ∂ ∂f ∂ 1 ∂f ∂ 1 ∂f 1 ∂ ∂f
Since the term = − and = , thus,
r ∂θ r ∂r ∂θ
2
∂r r ∂θ ∂θ r ∂r r ∂r ∂θ
we have
∂2 f ∂2 f ∂ 1 ∂f 2 1 ∂f 1 ∂2 f
= cos 2
θ − 2 sin θ cosθ + sin θ +
r ∂r r 2 ∂θ 2
∂x 2
∂r 2
∂r r ∂θ
Similarly, we find
∂2 f ∂2 f ∂ 1 ∂f 2 1 ∂f 1 ∂2 f
= sin 2
θ + 2 sin θ cos θ + cos θ +
r ∂r r 2 ∂θ 2
∂y 2 ∂r 2 ∂r r ∂θ
∂2 f 1 ∂f 1 ∂2 f ∂2 f ∂ 1 ∂f
− = sin θ cosθ + 2 − 2 − (cos 2 θ − sin 2 θ )
∂x∂y r ∂r r ∂θ
2
∂r ∂r r ∂θ
∂2F ∂ 1 ∂F 2 1 ∂F 1 ∂2F
σ x = sin 2 θ + 2 sin θ cos θ + cos θ
r ∂r r 2 ∂θ 2
+
∂r 2 ∂r r ∂θ
∂2F ∂ 1 ∂F 2 1 ∂F 1 ∂2F
σ y = cos 2 θ − 2 sin θ cos θ + sin θ
r ∂r r 2 ∂θ 2
+
∂r 2 ∂r r ∂θ
1 ∂F 1 ∂ 2 F ∂ 2 F ∂ 1 ∂F
τ xy = sin θ cosθ + 2 − 2 − (cos 2 θ − sin 2 θ )
r ∂r r ∂θ
2
∂r ∂r r ∂θ
σ r m
2
n2 2mn σ x
2
σ θ = n m 2 − 2mn σ y
τ − mn mn m 2 − n 2
rθ τ xy
, and rearranging the terms, we have
1 ∂F 1 ∂ 2 F
σr = +
r ∂r r 2 ∂θ 2
∂2F
σθ =
∂r 2
∂ 1 ∂F
τ rθ = −
∂r r ∂θ
Substituting the obtained functions of the stresses into the equilibrium equations,
∂σ r 1 ∂τ rθ σ r − σ θ 1 ∂σ θ ∂τ rθ τ
+ + = 0 and + + 2 rθ = 0 , and neglecting the body forces,
∂r r ∂θ r r ∂θ ∂r r
we can see that the stress functions are satisfied the equations of equilibrium. Thus, the
assumed stress functions constitute a general solution.
Now, the problem is reduced to finding a function F = F (r ,θ ) that satisfies the
compatibility conditions and the boundary conditions.
Next, we will transform the compatibility equation in the Cartesian coordinates,
∂4F ∂4F ∂4F
∇ 4 F = 4 + 2 2 2 + 4 = 0 , to the polar coordinates. Consider the harmonic
∂x ∂x ∂y ∂y
∂2 ∂2 ∂2 f ∂ 1 ∂f 2 1 ∂f 1 ∂2 f
2 + 2 f = cos 2 θ 2 − 2 sin θ cosθ + sin θ + 2
∂x ∂y ∂r ∂r r ∂θ r ∂r r ∂θ
2
∂2 f ∂ 1 ∂f 2 1 ∂f 1 ∂2 f
+ sin 2 θ + 2 sin θ cos θ + cos θ +
r ∂r r 2 ∂θ 2
∂r 2 ∂r r ∂θ
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-31
Thus,
∂2 ∂2 ∂2 1 ∂ 1 ∂2
2 + 2 f = 2 + + 2 f
∂x ∂y ∂r r ∂r r ∂θ 2
This expression tells us that the harmonic operator on the left-handed side, which is in
the Cartesian coordinates, is equivalent to the operator on the right-handed side, which is in
the polar coordinates.
The strain compatibility relation of the plane stress which is in the Cartesian
coordinates is
∂ 2ε y ∂ 2 ε x ∂ γ xy
2
+ =
∂x 2 ∂y 2 ∂x∂y
Substituting the strain-stress relations which are
1 1 2(1 + ν )
εx = (σ x − νσ y ) εy = (σ y − νσ x ) γ xy = τ xy
E E E
into the strain compatibility relation, we have
∂2 ∂2 ∂ τ xy 2
(σ y − νσ x ) + (σ x − νσ y ) = 2 (1 + ν )
∂x 2 ∂y 2 ∂x∂y
∂σ x ∂τ xy
Differentiating the equilibrium equations + = 0 with respect to x and
∂x ∂y
∂τ xy ∂σ y
+ = 0 with respect to y , and adding them, we find
∂x ∂y
∂ 2σ x ∂ σ y ∂ 2τ xy
2
− − =2
∂x 2 ∂y 2 ∂x∂y
Then, substituting it into the strain compatibility relation, we have
∂2 ∂2
2 + 2 (σ x + σ y ) = 0
∂x ∂y
By adding the stress transformation σ x and σ y from the transformation matrix
σ x m 2 n2 − 2mn σ r
2
σ y = n 2mn σ θ , we have
2
m
mn − mn m 2 − n 2 τ
τ xy rθ
σ x +σ y =σ r +σ θ
Consequently, the polar coordinate form of the compatibility equation in term of stress when
the body forces are absent is
∇ 2 (σ r + σ θ ) = 0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-32
∂2 ∂2
2 + 2 (σ r + σ θ ) = 0
∂x ∂y
∂2 ∂2 ∂2 1 ∂ 1 ∂2
Since 2 + 2 f = 2 + + 2 f , we have
∂x ∂y ∂r r ∂r r ∂θ 2
∂2 ∂2 ∂2 1 ∂ 1 ∂2
2 + 2 (σ r + σ θ ) = 2 + + (σ r + σ θ )
∂x ∂y ∂r r ∂r r 2 ∂θ 2
Thus, the transformation of the harmonic differential operator ∇ 2 to the polar
coordinate gives
∂2 1 ∂ 1 ∂2
∇2 = + +
∂r 2 r ∂r r 2 ∂θ 2
or
1 ∂ ∂ 1 ∂2
∇ =2
r +
r ∂r ∂r r 2 ∂θ 2
∂2 1 ∂ 1 ∂2 ∂ 2 F 1 ∂F 1 ∂ 2 F
4
∇ F = 2 + + 2 + + 2 =0
2
∂r r ∂r r 2 ∂θ 2 ∂r r ∂r r ∂θ
Strain Components in Polar Coordinates
Fig. 3.10
Let u = u (r ,θ ) and v = v(r ,θ ) are the radial and tangential (circumferential)
components of the displacement in the polar coordinates as shown in the Fig. 3.10. If the
∂u
radial displacement of the side ad is u , and of the side bc is u + dr ,
∂r
∂u
εr =
∂r
The strain in the tangential direction depends not only on the tangential displacement
v but on the radial direction as well. Assuming that the points a and d have only the
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-33
displacement u , the new length of the arc ad is (r + u )dθ , then, the tangential strain due to
only the displacement u is
(r + u )dθ − rdθ u
=
rdθ r
∂v
The difference in the tangential displacement of the sides ab and cd is dθ . The
∂θ
tangential strain due to only the displacement v is
∂v 1 1 ∂v
dθ =
∂θ rdθ r ∂θ
The total tangential strain is
u 1 ∂v
εθ = +
r r ∂θ
To determine the shearing strain γ rθ , let the element abcd deforms to the position
1 ∂u ∂v v
γ rθ = + −
r ∂θ ∂r r
Stress-strain Relations
Since the stress-strain relations are derived based on the internal energy in the
rectangular coordinate system, they also valid for other orthogonal coordinate systems such
as cylindrical or spherical coordinates. Thus, the stress-strain relations in the polar coordinate
can be written by changing the subscript x to r (radial) and y to θ (circumferential).
ε r 1 / E − ν / E 0 σ r
ε θ = − ν / E 1 / E 0 σ θ
γ 0 0 1 / G τ rθ
rθ
3.8 Stress Distribution Symmetrical about an Axis: Pressurized Cylinder
Consider a circular plate subjected to the internal and external pressure as shown in
Fig. 3.11. The plate may be a slice of a long thick-walled cylinder structure. It has an internal
radius of a and external radius of b . Due to the symmetry of the plate and the loading
condition, the stress components occurred in the small plate element within the radius a and
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-34
b depend on the distance from the center of the plate r only. Let the center of the plate is the
origin of the polar coordinates ( r , θ ). Thus, the compatibility equation becomes
Fig. 3.11
∂ 2 1 ∂ ∂ 2 F 1 ∂F
2 + + =0
∂r r ∂r ∂r 2 r ∂r
If the plate has the hole at the origin, we need to set the constants B to be zero (see
Theory of Elasticity, Timoshenko, p.78) in order to find the condition of stresses in the plate.
A
σr = + 2C
r2
A
σθ = − + 2C
r2
These expressions represent the stress distribution in a hollow cylinder subjected to
uniform pressure on the inner and outer surfaces. From the Fig. 3-11, let the uniform inner
and uniform outer pressures are pi and p o . Then, the boundary conditions are
A
(σ r ) r = a = + 2C = − pi
a2
A
(σ r ) r =b = + 2C = − p o
b2
Solving the simultaneous for A and 2C , we have
a 2 b 2 ( po − pi )
A=
b2 − a2
pi a 2 − p o b 2
2C =
b2 − a2
Therefore, the stress components are
1 a 2 b 2 ( p o − pi ) pi a 2 − p o b 2
σr = +
r2 b2 − a2 b2 − a2
1 a 2 b 2 ( p o − pi ) pi a 2 − p o b 2
σθ = − +
r2 b2 − a2 b2 − a2
Rearranging the equations, we obtain
a2 b2 b2 a2
σ r = pi 2 1 − 2 − p o 2 1 − 2
b − a2 r b − a2 r
a2 b2 b2 a2
σ θ = pi 1 + 2 − p o 2 1 + 2
b2 − a2 r b − a2 r
These stress components are occurred in thick-walled cylinder subjected to the
uniform inner pressure pi and uniform outer pressures p o . The distribution of the stress
components across the thickness of the cylinder is shown in Fig. 3.12a and 3-12b for the case
of internal pressure only and external pressure only, respectively.
Thick-walled Cylinder without External Pressure
If the external pressure p o = 0 , the cylinder is subjected to internal pressure only.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-36
a2 b2
σ r = pi 1 − 2
b2 − a2 r
a2 b2
σ θ = pi 1 + 2
b2 − a2 r
It can be seen that the circumferential stress is always larger than the radial stress at
the same value of r .
Fig. 3.12
If the cylinder as shown in Fig. 3.13 has end cap, the axial stress σ z presents in the
cylinder. The force acting on the end cap due to the internal pressure pi is p iπa 2 . The
a2
σ z = pi
b2 − a2
Fig. 3.13
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-37
The maximum shear stress in this case is occurred at r = a . The radial stress is the
minimum principal stress, σ r = − p i = σ 3 and the circumferential stress is the maximum
a2 + b2
principal stress, σ θ = pi = σ 1 . Therefore, the maximum shear stress is
b2 − a2
σ1 −σ 3 p b2
τ max = = 2i 2
2 b −a
Then, we can determine the radial and circumferential strains by using the strain-stress
relation.
1
εr = [σ r − ν (σ θ + σ z )]
E
1
εθ = [σ θ − ν (σ r + σ z )]
E
1
εz = [σ z − ν (σ r + σ θ )]
E
Finally, we can determine the displacements by using the strain-displacement
relations.
∂u
εr =
∂r
u 1 ∂v
εθ = +
r r ∂θ
1 ∂u ∂v v
γ rθ = + −
r ∂θ ∂r r
Thin-walled Cylinder without External Pressure
For thin-walled cylinder having the ratio of a /(b − a ) = a / t >> 20 and a / b ≈ 1 , we
have
σr = 0
pi a
σθ =
t
pi a
σz =
2t
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-38
Example 3-6
A steel cylinder with end caps is to have the inner radius a = 10 mm and outer radius
of b = 31.3 mm . Under the working pressure of 140 MPa , what is the radius expansion at
r = a and r = b ? Use E = 204 GPa and ν = 0.29 .
At working pressure pi = 140 MPa , we have the state of stresses at r = a = 10 mm as
a2 b2 b2 a2
σ r = pi 1 − 2 − p o 2 1 − 2
b2 − a2 r b − a2 r
σ r = −140 MPa
a2 b2 b2 a2
σ θ = pi 2 1 + 2 − p o 2 1 + 2
b − a2 r b − a2 r
σ θ = 172 MPa
a2
σ z = pi
b2 − a2
σ z = 16 MPa
Circumferential strain at r = a = 10 mm is
1
ε θa = [172 − 0.29(16 − 140)] = 0.00102
204000
Thus,
u 1 ∂v
0.00102 = +
r r ∂θ
Since the rate of change of v with respect to θ is zero due to the symmetry, the radius
expansion at r = a = 10 mm is
u a = 0.00102(10) = 0.0102 mm
In the similar fashion, the radius expansion at r = b = 31.3 mm with σ θb = 31.8 MPa
is
1
u b = bε θb = (31.3) [31.8 − 0.29(16 + 0)] = 0.0042 mm
204000
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-39
Fig. 3.14
Consider the plate having unit thickness and a small circular hole of radius a at the
center of the plate as shown in Fig. 3.14. It is subjected to a uniformly distributed axial tensile
stress σ o in the x − direction.
The stress distribution in the area neighboring the hole will be different from the stress
distribution in the plate without the hole. However, from the Saint-Venant’s principle, the
change is negligible at distances which are large compared with the radius of the hole, a .
Let the radius b is large in comparison to the radius a of the hole so that the stresses
at the b are the same as in the plate without the hole. By using the stress transformation
equation from the Cartesian coordinates ( x , y ) to the polar coordinates ( r , θ ),
σ r m
2
n2 2mn σ x
2
σ θ = n m 2 − 2mn σ y
τ − mn mn m 2 − n 2
rθ τ xy
σo
(σ r ) r =b = σ o cos 2 θ = (1 + cos 2θ )
2
σo
(τ rθ ) r =b = − sin 2θ
2
It can be seen that the stresses can be considered separately into two parts. The first
σo
part is due to the constant components of . The solution of this part is obtained in previous
2
section as
a2 b2 b2 a2
σ r = pi 1 − 2 − p o 2 1 − 2
b2 − a2 r b − a2 r
a2 b2 b2 a2
σ θ = pi 1 + 2 − p o 2 1 + 2
b2 − a2 r b − a2 r
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-40
σo
where pi = 0 and p 0 = − .
2
σo
The remaining part consists of the normal stress cos 2θ and the shearing stress
2
σo
− sin 2θ . The solution of this part can be determined by using the Airy stress function of
2
the form
F = f (r ) cos 2θ
Substituting the function into the compatibility condition,
∂2 1 ∂ 1 ∂2 ∂ 2 F 1 ∂F 1 ∂ 2 F
2 + + 2 2 + + =0
∂r r ∂r r ∂θ 2 ∂r r ∂r r 2 ∂θ 2
d2 1 d 4 d 2 f 1 df 4 f
2 + − 2 2 + − = 0
dr r dr r dr r dr r 2
This is an ordinary differential equation. The general solution of the equation is in the
form of
1
f (r ) = Ar 2 + Br 4 + C +D
r2
Thus, the stress function is
1
F = Ar 2 + Br 4 + C 2 + D cos 2θ
r
The corresponding stress components are
1 ∂F 1 ∂ 2 F 6C 4 D
σr = + 2 = − 2 A + 4 + 2 cos 2θ
r ∂r r ∂θ 2
r r
∂2F 6C
σθ = = 2 A + 12 Br 4 + 4 cos 2θ
∂r 2
r
∂ 1 ∂F 6C 2 D
τ rθ = − = 2 A + 6 Br − 4 − 2 sin 2θ
2
∂r r ∂θ r r
By using the boundary conditions, we can determine the constants of integration. At
the outer boundary where r = b , the remaining stress that produce the stress components
σo σ
(σ r ) r =b and (τ rθ ) r =b are the normal stress cos 2θ and the shearing stress − o sin 2θ .
2 2
σo σ
Thus, the remaining stresses are (σ r ) r =b = cos 2θ and (τ rθ ) r =b = − o sin 2θ .
2 2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-41
σo 6C 4 D
cos 2θ = − 2 A + 4 + 2 cos 2θ
2 b b
6C 4 D σ
2A + 4
+ 2 =− o
b b 2
σo 6C 2 D
− sin 2θ = 2 A + 6 Bb 2 − 4 − 2 sin 2θ
2 b b
6C 2 D σ
2 A + 6 Bb 2 − 4
− 2 =− o
b b 2
At the edge of the hole where r = a , the stress components must be zero since there is
free from external force.
6C 4 D
0 = − 2 A + 4 + 2 cos 2θ
a a
6C 4 D
2A + + =0
a4 a2
6C 2 D
0 = 2 A + 6 Bb 2 − 4 − 2 sin 2θ
b b
6C 2 D
2 A + 6 Ba 2 − − =0
a4 a2
Solving these four simultaneous equations and assuming that the plate is infinitely
large, a / b = 0 , we get
σo
A=−
4
B=0
a4
C=− σo
4
a2
D= σo
2
Substituting the constants of integration into the stress component equation and plus
σo
the stresses component due to the uniform tension stress found in the previous section.
2
σo a2 σo 3a 4 4a 2
σr = 1 − 2 + 1 + 4 − 2 cos 2θ
2 r 2 r r
σo a2 σ o 3a 4
σθ = 1 + 2 − 1 + 4 cos 2θ
2 r 2 r
σ 3a 4 2a 2
τ rθ =− o 1 − 4 + 2 sin 2θ
2 r r
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-42
The plots of the normal stress distribution along the transverse and centerline of the
infinitely large plate is shown in Fig. 3.15.
It can be seen that the normal stress occurred on the edge of the hole where r = a and
π
θ =± has the positive value of
2
(σ θ ) max = 3σ o
For θ = 0 and θ = π , the normal stress on the edge of the hole where r = a has the negative
value of σ o . Thus, σ θ attains a maximum tensile value of three times the uniformly
distributed stress σ o . This value is the largest stress occurs in the plate. Hence the stress
concentration factor at the hole, which is the ratio of the maximum normal stress at the hole
divided by the averaged normal stress at the same point in the absence of the hole, is 3
Fig. 3.15
From Fig. 3.15, it can be seen that the stress σ θ approaches the average values of σ o
at a small distance from the hole. Thus, the high stress gradient or stress concentration is quite
localized in effect according to the Saint-Venant’s principle.
The state of stresses for a circular hole in a plate under other states of plane stress can
be determined by using the principle of superposition. For example, if the previously obtained
the state of stresses is combined with another state of stresses in which everything is rotated
by 90 o , we obtain the state of stresses in equal biaxial tension. Or, if the direction of σ o is
reversed in one of these two solutions, we obtain results for the pure shear.
It should be noted that the obtained stress equations can only be used in the case when
the plate has the diameter that is small compared to the width of the plate. When the diameter
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-43
of the hole is comparable to the width of the plate, researches have shown that the maximum
normal stress should be calculated by using the equation
3κ − 1
σ max = σn
κ + 0.3
where κ is the ratio of the width of the plate to the diameter of the hole, and σ n is the
Fig. 3.16
The resulting stresses due to the concentrated force P can be determined by using the
Airy stress function of the form
F = a1 rθ sin θ
The corresponding stresses can be determined as
1 ∂F 1 ∂ 2 F cosθ
σr = + 2 = 2a1
r ∂r r ∂θ 2
r
∂2F
σθ = 2 = 0
∂r
∂ 1 ∂F
τ rθ = − =0
∂r r ∂θ
The actual stress distribution of σ r is shown in Fig 3.16b. The constant a1 can be
determined by using the summation of the vertical forces to be zero.
π /2
− ∫ (cosθ )σ
−π / 2
r rdθ − P = 0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 3-44
P
a1 = −
π
Thus, the state of stresses is
2 P cosθ
σr = −
πr
σ θ = 0 and τ rθ = 0 .
This distribution of stress is called a simple radial distribution, every polar coordinate
element at a distance r from the point of application of P , being in simple compression in
the radial direction. The stress σ r becomes very large as r becomes small and is undefined
for r = 0 (the stress is said to be singular). For all real materials, yielding will occur in the
neighborhood of the load, resulting in a plastic zone.
If it is assumed that the plastic zone is sufficiently small so that the presence of the
yielding material does not significantly change the elastic solution,
r 2P
= = d y = constant
cosθ πσ y
Chapter 4
Applications of Energy Method
4.1 Degree of Freedom
Degree of freedom is the independent quantities used to define a configuration of a
system that violates neither compatibility conditions nor constraints.
Finite degree of freedom system
Fig. 4.1
Consider the beam having three rigid elements connected with hinges and elastic
springs as shown in Fig. 4.1. In this case, the configuration of the beam can be described by
using two of the independent quantities xi and θ j where i ≠ j such as x1 and x 2 or x1 and
Fig. 4.2
Consider the bar subjected to arbitrary axial load as shown in Fig. 4.2. It requires an
infinite number of degree of freedom to describe its axial displacement since the displacement
varies along the length of the bar.
P ( x) dδ
σ= ε=
A( x) dx
σ = Eε
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-2
P( x ) dδ
=E
A( x ) dx
P ( x ) dx
dδ =
A( x ) E
L P( x )
δ=∫ dx
0 A( x ) E
Fig. 4.3
dW = f cos β (du )
B
W = ∫ f cos βdu
A
π π
If the angle β is − < β < , the force f performs positive work. If the angle β is
2 2
π 3π
<β< , the force f performs negative work. The f sin β does no work. If the system
2 2
is conservative, the work done is independent of the path from point A to point B . Then, if
the integral is independent of the path, the quantity ( f cos β )du is an exact differential of
some function V . Thus, for conservative system,
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-3
B B
W = ∫ f cos βdu = ∫ dV = VB − V A = − ∆V
A A
Fig. 4.4
Consider a three-dimensional deformable body as shown in Fig. 4.4a which is in a
state of undeformed configuration and the external forces have zero potential energy. As the
body slowly deforms, external forces move through true displacements and perform external
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-4
work until the final configuration is reached as shown in Fig. 4.4b. Again, let us assume that
the work done by the external forces in deforming the body is independent of the path from
Fig. 4.4a to Fig. 4.4b. Hence, if We is the total work done by the external forces, dWe is the
exact differential of a potential function Ω which is the potential of the external forces to do
work. The change in Ω from the configuration in Fig. 4.4a to Fig. 4.4b is − We . Therefore,
If the body forces are negligible and the surface forces are represented by a system of
concentrated forces and moments P1 , P2 , K , Pn with the corresponding displacement of ∆ 1 ,
∆2 , K , ∆3 ,
Ω = −( P1 ∆ 1 + P2 ∆ 2 + K + Pn ∆ n )
The internal forces developed in a deformable body also posses a capacity to perform
work. Under the action of external forces, the body is deformed and the stresses are developed
which results in the internal forces. The internal forces perform work while the body are
deformed. If the strained body is allowed to slowly return to its unstrained state, it will return
the work done by the external forces. This capacity of internal forces to perform work as the
body returns to the unstrained state is called strain energy.
For linearly elastic and isotropic material, the strain energy density was derived as
Uo =
1
2
[
σ x ε x + σ y ε y + σ z ε z + τ xy γ xy + τ yz γ yz + τ xz γ xz ]
The total strain energy U of the body is
U = ∫∫∫U o dV
V
δU = σ xδε x
From calculus, the differential change in the strain energy in the body due to an
increment dε x of the strain is
∂U
dU = dε x = σ x dε x
∂ε x
We can see that the symbol δ is not just only the symbol for the virtual quantity. In
fact, it behaves as a variational operator which obeys the rules of operation similar to those
of the first differential operator d . In analogy, if we refer δ as first variation, we can see that
the internal virtual work can be interpreted as the first variation in the strain energy due to
variations in the components of strains.
Similarly, if We is the work done by external forces in a conservative system
Ω = − We
δΩ = −δWe
From the principle of virtual displacements δWe = δU ,
δU = −δΩ
δU + δΩ = 0
δ (U + Ω) = 0
δV = 0
A deformable body is in equilibrium if the first variation in the potential energy of the
system is zero for every virtual displacement consistent with the constraints
Consider a system having two degrees of freedoms x1 and x 2 as shown in Fig. 4.6. In
this case, the total potential energy of the system can be expressed as a function of x1 and x 2 ,
so that are the rate of change of V with respect to x1 and x 2 . Thus, for the structure in
equilibrium, if first x1 is given a variation δx1 and then x 2 a variation δx 2 ,
∂V
δV = δx1 = 0
∂x1
∂V
δV = δx 2 = 0
∂x 2
Since the magnitude of the virtual displacement δx1 and δx 2 are arbitrary,
∂V ∂V
= =0
∂x1 ∂x 2
which mean for equilibrium to exist,
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-6
∂V ∂V
dV = dx1 + dx 2 = 0
∂x1 ∂x 2
Recall from calculus, the total differential of a function vanishes at the critical points
of the function at which the function is a relative maximum or maximum or they may be
saddle points at which the function is minimax. At such point, the function is said to assume a
∂V ∂V
stationary value. Since the equation dV = dx1 + dx 2 = 0 is valid only at the
∂x1 ∂x 2
equilibrium, thus, the principle of stationary potential energy can be stated that
If a structure is in static equilibrium, the total potential energy of the system has a
stationary value.
∂V ∂V
The equation dV = dx1 + dx 2 = 0 can be interpreted another way. Let V is a
∂x1 ∂x 2
continuous function displacement, which is consistent with the boundary conditions. dV is
zero only if the displacements are corresponding to the equilibrium configurations. Thus, the
principle of stationary potential energy can be restated that
Of all the possible displacements which satisfy the boundary conditions of a structural
system, those corresponding to the equilibrium configurations make the total potential energy
assume a stationary value.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-7
Example 4-1
Determine the equilibrium configurations of a system of three bars subjected to the
point load P as shown in Fig. Ex 4-1a. The bars are supported by pins and are joined with
EI
internal hinge and springs having the stiffness of k = .
l
Fig. Ex 4-1a
The system as shown in Fig. Ex 4-1a has two degrees of freedoms x1 and x 2 .
The potential energy of the external force P is equal to the amount of work done by
the force P referred to the reference at support.
2
Ω = − P x1
3
(Negative value of the work done by external force = the external potential energy)
The work done due to the couple moment M o using to move the spring from 0 to φ
as shown in Fig. Ex 4-1b or the strain energy stored in the spring is equal to
φ φ
1 2
∫ M o dθ = ∫ kθ dθ =
0 0
2
kφ .
Fig. Ex 4-1b
For small displacement and rotation, the angle
θ ≈ sin θ ≈ tan θ .
Then, from Fig. Ex 4-1c,
x1 x1 − x 2 2 x1 x 2
θ1 = + = −
l l l l
x 2 x1 − x 2 2 x 2 x1
θ2 = − = −
l l l l
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-8
Fig. Ex 4-1c
Therefore, the strain energy stored in the springs on the beam is
2 2
1 EI 2 x1 x 2 1 EI 2 x 2 x1
U= − + −
2 l l l 2 l l l
The total potential energy is
2 2
2 1 EI 2 x1 x 2 1 EI 2 x 2 x1
V = − P x1 + − + −
3 2 l l l 2 l l l
For the stable equilibrium system, the total potential energy must be minimum.
∂V
=0
∂x1
∂V
=0
∂x 2
5 x2
Solving the simultaneous equation, we have x1 = . Then,
4
10 Pl 3
x1 =
27 EI
8 Pl 3
x2 =
27 EI
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-9
Example 4-2
The three bars plane truss has the configuration as shown in Fig. Ex 4-2. If the bars
have the same EA , determine the stress in each bar.
Fig. Ex 4-2
Let u D and v D are the degree of freedom at joint D . By considering the deformation
of each bar, we can find the relationships of the change in length of each bar δ i and the
δ2 v
ε2 = =− D
L2 L
δ3 1 4 3 1 12 9
ε3 = = − u D − v D = − u D − v D
L3 L3 5 5 L 25 25
L
P2
Strain energy in a uniform bar is U a = ∫ dx . From the Hooke's law, P = AEε .
0
2 AE
Thus,
AEε 2
L
AEL 2
Ua = ∫ dx = ε
0
2 2
Ω = Pu D
Thus, the total potential energy of the system is
V =
AE
2
[ ]
L1ε 12 + L2 ε 22 + L3ε 32 + Pu D
AE 5 12
2 2
16 v D2 5 12 9
V = u D − v D + + − u D − vD + Pu D
2 4 L 25 25 L 3L 25 25
From the principle of stationary potential energy
∂V
=0
∂u D
∂V
=0
∂v D
Fig. 4.5
If small displacement cause no change in potential energy, the system is said to be in
neutral equilibrium, which is also the unstable equilibrium.
Consider the motion of a rigid marble having a weight of W along a smooth contour
in Fig 4.6. The rigid marble has no strain energy.
U =0
Fig 4.6
If x -axis is the reference line, the total potential of a given marble is
V = Ω = Wh(x)
Generalized coordinate is the least number of real independent variables required to
specify the configuration of a system. The rigid bars as shown in Fig. Ex 4-1 require two
generalized coordinates to define their configuration. The virtual displacements in a system,
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-12
which are consistent with the constraints, can be expressed as a function of the variations in
the generalized coordinates
A single marble in Fig. 4.6 can be specified by the generalized coordinate x . Thus, a
virtual displacement following the rigid contour of a marble and having the generalized
coordinate x is a small variation δx . The variation δx produces a variation in V . Since each
marble is in equilibrium,
dh
δV = W δx = 0
dx
dh
which is valid when = 0 at points A , B , C , D , and E . Thus, these points are possible
dx
equilibrium configurations.
From the previous discussion, it may be concluded that
1. The marbles at A and D are in stable equilibrium.
2. The marble at C is in unstable equilibrium.
3. The marble at E is in neutral equilibrium.
The potential energy of the marble at point B is either increased or decreased
depending on the direction of the virtual displacement. Thus, the marble at B is in unstable
equilibrium.
If V (x) corresponds to an equilibrium configuration, V ( x + δx) is the potential energy
of a configuration in the neighborhood of V (x) if δx is sufficiently small. By using Taylor’s
series, we have
dV ( x) 1 d 2V ( x) 1 d 3V ( x)
V ( x + δx ) = V ( x ) + δx + 2
(δ x ) 2
+ 3
(δx) 3 + K
dx 2 dx 3 dx
dV
For equilibrium, = 0 . Thus, the change in V due to δx is V ( x + δx) - V (x) ,
dx
1 d 2V ( x) 1 d 3V ( x)
∆V = 2
(δ x ) 2
+ 3
(δx) 3 + K
2 dx 3 dx
For the system in Fig. 4.6,
1 d 2h 1 d 3h
∆V = W 2
(δ x ) 2
+ 3
(δx) 3 + K
2 dx 3 dx
Thus, the first order variation in the displacements at an equilibrium configuration causes a
change in potential energy of order (δx) 2 .
d 2V
If the term of ∆V is not zero, the sign of ∆V is independent of the sign of the
dx 2
virtual displacement δx since (δx) 2 . Then, V is either a relative maximum or relative
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-13
d 2h
minimum. For the system in Fig. 4.6, the term , representing the curvature of the rigid
dx 2
curve, is positive at A and D and negative at C . Thus, V is a relative minimum at A and
D and a relative maximum at C . Clearly, the marbles at A and D are in stable equilibrium
d 2h
and the marble at C is in unstable equilibrium. At B , = 0 due to flat contour at the
dx 2
d 3h
point. However, if ≠ 0 , the sign of ∆V depends on the sign of the virtual displacement
dx 3
δx due to the term (δx) 3 . At this point, V is a minimax and the marble is in unstable
equilibrium. At E , all derivative of V vanish and ∆V is zero for small virtual displacement.
Thus, The marble at E is in neutral equilibrium.
From these observations, we conclude that
1. If δ 2V > 0 , the system is in stable equilibrium.
2. If δ 2V < 0 , the system is in unstable equilibrium.
3. If δ 2V = 0 , the system is in neutral equilibrium.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-14
Example 4-3
Check the stability of the system of rigid bar subjected to axial load P as shown in
Fig. Ex 4-3.
h
k
Fig. Ex 4-3
Consider a system of rigid bar subjected to axial load P as shown in Fig. Ex 4-3. All
configurations of the system can be described by specifying one independent variable, the
angular coordinate θ . Thus, θ is the generalized coordinate.
The external potential energy Ω of the system is zero for reaction forces plus minus
2 Ph(1 − cosθ ) due to axial force P
Ω = −2 Ph(1 − cosθ )
The internal potential energy or strain energy U of the system is zero for rigid bar
1
k (h sin θ ) in the spring.
2
plus
2
1
k (h sin θ )
2
U=
2
The total potential energy of the system is
1
V (θ ) = k (h sin θ ) 2 − 2 Ph(1 − cosθ )
2
The total potential energy of the system due to variations δθ is
dV (θ ) 1 d 2V (θ ) 1 d 3V (θ )
V (θ + δθ ) = V (θ ) + δθ + (δθ ) +
2
(δθ ) 3 + K
dθ 2 dθ 2
3 dθ 3
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-15
The change in the total potential energy of the system due to a rotation variation δθ is
dV 1 d 2V 1 d 3V
∆V = V (θ + δθ ) − V (θ ) = δθ + (δθ ) 2
+ (δθ ) 3 + K
dθ 2 dθ 2
3 dθ 3
[
∆V = kh 2 sin θ cosθ − 2 Ph sin θ δθ + ] [
1 2
2
]
kh cos 2θ − 2 Ph cosθ (δθ ) 2 + K
kh
Pcr =
2
If θ = 180 o , the term
kh 2 cos 2θ − 2 Ph cosθ = kh 2 + 2 Ph
which is always larger than zero. Thus, the system is always in stable equilibrium when
θ = 180 o .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-16
1
It should be noted that in this case, we can not assume cosθ = 1 − θ 2 in the small
2
displacement analysis.
1 1 1
V = k (hθ ) 2 − 2 Ph 1 − (1 − θ 2 ) = kh 2θ 2 − Phθ 2
2 2 2
[
∆V = kh 2θ − 2 Phθ δθ + ] [
1 2
2
]
kh − 2 Ph (δθ ) 2 + K
which provides only one solution that is θ = 0 o . Thus, for stability analysis, we have to
consider the potential energy of the system to the cubic terms.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-17
If the applied loads are increased, D decreases. When a critical load is reached, δ 2V
is not a positive definite and the system becomes unstable and buckles. Thus, the stability
criterion is
δ 2V = D = 0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-19
Fig. Ex 4-4
The system of rigid bars as shown in Fig. Ex 4-4a has two degree of freedoms. All
configurations of the system can be described by specifying two independent variables, the
angular coordinate θ and ψ . Thus, θ and ψ are the generalized coordinates.
For positive values of θ and ψ , the applied load P is displaced a distance ∆ .
∆ = 2a − a cosθ − a cosψ
Thus, the potential energy of the external force P is
Ω = − P∆
For small values of generalized coordinates θ and ψ and since we are interested to
1
find only the critical load, thus, we let cosθ = 1 − θ 2 . The potential energy of the external
2
force P is
Pa 2
Ω=− (θ + ψ 2 )
2
Since the internal forces and moments are developed in the springs, the system posses
strain energy.
x α
U = 2 ∫ k1 xdx + ∫ kα αdα
0 0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-20
k x2 kα α 2
U = 2 1 +
2 2
where x is the change in length of the linear spring and α is the angle of rotation of the
rotational spring.
From the geometry of Fig. Ex 4-4b, we find
x = a sin θ ≈ aθ
α = ψ −θ
Thus,
1
U = k1 a 2θ 2 + kα (ψ − θ ) 2
2
The total potential energy of the system is
1 1
V (θ ,ψ ) = (kα + 2k1 a 2 − Pa)θ 2 − kα θψ + (kα − Pa)ψ 2
2 2
= c0θ + c1θψ + c 2ψ
2 2
where
1
c0 = (kα + 2k1 a 2 − Pa)
2
c1 = −kα
1
c2 = (kα − Pa)
2
The total potential energy of the system due to variations δθ and δψ is
V (θ + δθ ,ψ + δψ ) = c 0θ 2 + c1θψ + c 2ψ 2
+ [(2c 0θ + c1ψ )δθ + (c1θ + 2c 2ψ )δψ ]
[
+ c0 (δθ ) 2 + c1δθδψ + c 2 (δψ ) 2 ]
The change in total potential energy of the system is
[
∆V = [(2c 0θ + c1ψ )δθ + (c1θ + 2c 2ψ )δψ ] + c 0 (δθ ) 2 + c1δθδψ + c 2 (δψ ) 2 ]
For equilibrium configurations, δV is zero.
∂V
= 0 = 2c 0θ + c1ψ
∂θ
∂V
= 0 = c1θ + 2c 2ψ
∂ψ
Clearly, θ = ψ = 0 is an equilibrium configuration.
∂ 2V ∂ 2V ∂ 2V
δ 2V = (δθ ) 2
+ 2 δθδψ + (δy ) 2
∂θ 2 ∂θ∂ψ ∂ψ 2
[
= 2 c 0 (δθ ) 2 + c1δθδψ + c 2 (δψ ) 2 ]
When the system is stable, the determinant
c0 c1 / 2
D= >0
c1 / 2 c 2
When P reaches the critical load, D = 0 and the system becomes unstable. Thus,
1
c0 c 2 − c12 = 0
4
1 1
Substituting c0 = (kα + 2k1 a 2 − Pa) , c1 = −kα , and c 2 = (kα − Pa) , we have
2 2
(kα − Pa) (kα + 2k1 a 2 − Pa) − kα2 =0
Solving this polynomial equation for P , we have
P1 = [ 1
a
kα + k1 a 2 − kα2 + k12 a 4 ]
P = [k +k a ]
1
2 α + k1 a 2 + kα2 1
2 4
a
Let kα = k1 a 2 = k , then,
k
P1 = 0.586
a
and
k
P2 = 3.414
a
k
Thus, P1 = 0.586 is the critical load for the system. Substituting Pcr = 0.586k / a into the
a
equations 2c0θ + c1ψ = 0 or c1θ + 2c 2ψ = 0 , we find
ψ = 2.414θ
The buckled shape for this case is shown in Fig. Ex 4-4c(a).
Substituting Pcr = 3.414k / a into the equations 2c0θ + c1ψ = 0 or c1θ + 2c 2ψ = 0 , we
find
ψ = −0.414θ
The buckled shape for this case is shown in Fig. Ex 4-4c(b). These two possible buckled
shapes are called the buckling modes of the system.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-22
Fig. Ex 4-4c
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-23
Lagrange multipliers
In many cases when it is not convenient to express V in terms of the least number of
independent variables. Then, the problems become one of minimizing a function whose
variables are constrained by some side relationships. These kinds of problems can be solved
by using the Lagrange multipliers.
In general, when the variables x1 , x 2 , K , x n of a function G ( x1 , x 2 , K, x n ) to be
g 2 ( x1 , x 2 , K , x n ) = 0
M
g m ( x1 , x 2 , K, x n ) = 0
∂G
=0
∂x 2
M
∂G
=0
∂x n
can be determined.
From the previous example, if we do not introduce the equation x = aθ into the
kα α 2
equation U = k1 x 2 + , we have the total potential energy of the system of the form
2
1 1
V (θ ,ψ , x) = (kα − Pa)θ 2 − kα θψ + k1 x 2 + (kα − Pa)ψ 2
2 2
with the additional condition
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-24
x − aθ = 0
We form the new function
V = V (θ ,ψ , x) + λ ( x − aθ )
∂V
= 2 k1 x + λ = 0
∂x
From the last condition, we find that
λ = −2k1 x = −2k1 aθ
Substituting λ into the rest of the equations and rearrange the terms, we have
2c0θ + c1ψ = 0
c1θ + 2c 2ψ = 0
which are obtained before.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-25
Fig. Ex 4-5
For simplicity, neglecting the shear deformation and assuming that the normal stress is
given by
N My
σx = +
A I
and other stresses are zero.
From Hooke’s law, σ x = Eε x . The strain energy in this system is
1
U = ∫∫∫ σ x ε x dV
V
2
1 N My
2
σ x2 L
U = ∫∫∫ dAdx = ∫ ∫ + dAdx
V
2E 2 E 0 A A I
Thus,
1 N2
L
NM M 2 y2
2 E ∫0 ∫A A 2
U= + 2 y + dAdx
AI I2
∫ ydA = 0 and ∫ y dA = I ,
2
Since
A A
N2
L
M2
U = ∫ + dx
0
2 EA 2 EI
If we are interested only the transverse displacement and the stability of the beam-
column, we need to consider only the energy due to changes in curvature. Thus,
2
EI d 2 v
L L
M2
U =∫
2 ∫0 dx 2
dx = dx
0
2 EI
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-26
where v is the transverse displacement. Then, the total potential energy of the external forces
is
L
Ω = − ∫ p( x)vdx − P∆
0
1
dv 2 2 1 dv 2
Since the beam-column’s length ds = 1 + dx ≈ 1 + dx , thus,
dx 2 dx
L
1 dv
2
1 L dv 2
∆ ≅ ∫ dx + dx − dx = ∫ dx
2 dx 2 dx
0 0
Hence,
P dv 2
L
Ω = − ∫ + p ( x)v dx
0 2 dx
The total potential energy of the beam-column is
EI d 2 v 2 P dv 2
L
V = ∫ 2 − − pv dx
0
2 dx 2 dx
Assuming that the elastic curve of the beam-column is in the for of
πx
v = C sin
L
where C is an unknown constant. This equation satisfies all the kinetic boundary conditions
of the beam-column. Since the magnitude of the displacement at any point depends on the
magnitude of C , thus, C is a generalized coordinate. Substituting v into the total potential
energy equation, we have
EI d 2
L 2
πx P d πx
2
πx
V = ∫ 2 C sin − C sin − p C sin dx
0
2 dx L 2 dx L L
L C 2π 2 EIπ 2
V= 2 2 − P − 2 po C
4 L L
If the beam-column is in equilibrium,
dV L 2Cπ 2 EIπ 2
= 2 − P − 2 po = 0
dC 4 L2 L
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-27
po L4
C=
π 2 ( EIπ 2 − PL2 )
po L4 πx
v= sin
π ( EIπ − PL )
2 2 2
L
If the beam-column is in stable equilibrium,
d 2V π 2 EIπ 2
2
= 2 − P > 0
dC 2L L
When P has a value that δ 2V = 0 , the beam-column is no longer in stable
equilibrium. Thus, the critical load of the beam-column-column is
π 2 EI
Pcr =
L2
which is the Euler buckling load of pinned-pinned column.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-28
K , η n are chosen so that the components of displacement satisfy all of the kinematic
conditions. These kinds of function are usually called kinematically admissible functions.
Since the components of displacement u , v , and w are defined in terms of only 3n
independent parameters, these parameters behaves as generalized coordinates, thus, the
system has only 3n degrees of freedom.
From the components of displacement, we can determine the approximate strains and,
then, the total potential energy of the system.
V (u , v, w) = V ( x, y,z,a1 ,K ,an ,b1 ,K ,bn ,c1 ,Kcn )
The variation of the components of displacement can be written in the form of the
variations in the parameters ai , bi , and ci .
n
δu = ∑ φ i δa i
i =1
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-29
n
δv = ∑ψ iδbi
i =1
n
δ w = ∑ η i δ ci
i =1
Hence, the variation of the total potential energy of the system can be written as
n
∂V ∂V ∂V
δV = ∑ δa i + δbi + δci
i =1 ∂ai ∂bi ∂ci
If the system is in equilibrium,
n
∂V ∂V ∂V
∑ ∂a δa i +
∂bi
δbi +
∂ci
δci = 0
i =1 i
for arbitrary values of the variations δai , δbi , and δci . Then, we have
∂V ∂V ∂V
=0 =0 K =0
∂a1 ∂a2 ∂an
∂V ∂V ∂V
=0 =0 K =0
∂b1 ∂b2 ∂bn
∂V ∂V ∂V
=0 =0 K =0
∂c1 ∂c2 ∂cn
These equations are 3n linearly independent simultaneous equations which can be
solved for the unknowns parameters ai , bi , and ci .After solving for the unknowns parameters
In the case of stability analysis, the equations are homogeneous, we then determine the
buckling loads by setting the determinant of the coefficients to be zero.
Some important characteristics of the Rayleigh-Ritz method
1. The accuracy of the assumed displacement is in general increased with an increase
in the number of parameters. However, the exact solutions are rarely obtained.
2. Since the differential equations of equilibrium do not enter the analysis, the
equilibrium is satisfied in an average sense through minimization of the total
potential energy. Thus, in general, the stresses computed do not satisfy the
equations of equilibrium.
3. Although the Rayleigh-Ritz method may provide fairly accurate results for the
displacements, the corresponding stresses may differ significantly from their exact
values since the stresses depend on the derivatives of the displacements.
4. Since the Rayleigh-Ritz method use a finite number of degree of freedom to
approximate the displacements of the system having infinitely many degree of
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-30
freedom, the approximate system is less flexible than the actual system. In the
stability analysis, the Rayleigh-Ritz method always produce the buckling loads
that larger than the exact values.
5. The modes of deformation are often taken as polynomial or trigonometric
functions since they are easy to manipulate.
6. The approximate displacement functions should not omit terms of lower order
since it prevents the convergence of the solutions.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-31
Example 4-6
Find the maximum deflection and moment of the simply supported beam as shown in
Fig. Ex 4-6 by using the Rayleigh-Ritz method and compared with the results from mechanic
of material method of analysis.
Fig. Ex 4-6
Consider the simply supported beam as shown in Fig. Ex 4-6. The elastic curve of the
beam can be assumed as a sine function of
πx
v = a sin
L
πx
where a is an unknown constants. Note that sin satisfy the boundary conditions at x = 0
L
and x = L .
If only the strain energy due to bending is considered, we have the total potential
energy of the beam is of the form
2
EI
L
d 2v L
V=
2 ∫0 dx 2
dx − ∫0 p( x)vdx − Pv L/2
EIπ 4 2 L
V = a − 2 p o a − Pa
4L3
π
Now, we choose the constant a so as to minimize V .
∂V EIπ 4 L
= a − 2 po − P = 0
∂a 2L3
π
4 po L4 2 PL3 2 L3 (2 po + πP)
a= 5 + =
π EI π 4 EI π 5 EI
2 L3 (2 po + πP) πx
v= sin
π 5 EI L
Evaluating the deflection v of the beam at x = L / 2 , we have
4 po L4 2 PL3 po L4 PL3
v L/2 = a = + = +
π 5 EI π 4 EI 76.5EI 48.7 EI
By using the mechanics of materials, we obtain
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-32
po L4 PL3
v L/2 = +
76.8 EI 48 EI
Thus, by using only one parameter, we obtain a maximum deflection, which is only 0.39% in
error in case of uniformly distributed load and 1.46% in error in case of concentrated load.
However, the approximate deflection curve gives a bending moment at x = L / 2 of
d 2 v 4 po L2 2 PL po L2 PL
M L / 2 = − EI = + 2 = +
dx 2 π3 π 7.75 4.93
The first term has 3.15% in error and the second term has 23.37% in error. Note that
the normal stresses are proportion to the bending moment, thus, are in error by the same
percentages.
To obtain more accurate results, let us use two-parameter approximation elastic curve
of the beam as
πx 3πx
v = a sin + b sin
L L
Note that this function also satisfy the kinematic boundary conditions of the beam.
Performing the determination of the total potential energy of the beam as before, and then
∂V ∂V
from the conditions = 0 and = 0 , we find that
∂a ∂b
4 po L4 2 PL3
a= +
π 5 EI π 4 EI
4 po L4 2 PL3
b= −
243π 5 EI 81π 4 EI
In this case, the maximum deflection at x = L / 2 is
po L4 PL3
v L/2 = +
76.8 EI 48.1EI
which coincide with the exact solution. The bending moment at x = L / 2 is
po L2 PL
M L/2 = +
8.05 4.44
The first term has now only 0.63% in error and the second term has 11.00% in error.
We can further reduce the error by increasing the number of parameters in the approximate
displacement function.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-33
Example 4-7
Determine the buckling loads of the fixed end column as shown in Fig. Ex 4-7 by
using the Rayleigh-Ritz method.
Fig. Ex 4-7
Consider the fixed end column as shown in Fig. Ex 4-7. The transverse deflection of
the column can be assumed to be of the form
v = a( x 3 − 3 xL2 + 2 L3 ) + b( x − L) 2
where a and b are the unknown parameters. This function satisfy the boundary condition of
the column at x = L . If we consider only the strain energy of the bending moment, we have
2
EI
L
d 2v
U=
2 ∫0 dx 2 dx
The potential energy of the external force P is
L 2
P dv
2 ∫0 dx
Ω=− dx
2 5 1
V = 6a 2 L3 ( EI − PL2 ) + abL2 (6 EI − PL2 ) + 2b 2 L( EI − PL2 )
5 2 3
If the parameter a and b are to correspond to a stationary value of V ,
∂V 2 5
= 12 L3 ( EI − PL2 )a + L2 (6 EI − PL2 )b = 0
∂a 5 2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-34
∂V 5 1
= L2 (6 EI − PL2 )a + 4 L( EI − PL2 )b = 0
∂b 2 3
A nontrivial solution to this column of homogeneous equations exists only if the
determinant of the coefficients vanishes. Then, we have
EI ( EI ) 2
3P 2 − 104 P + 240 =0
L2 L4
Solving the polynomial equation, we obtain
EI
P1 = 2.486
L2
EI
P2 = 32.181
L2
EI
Thus, the buckling load of the column is Pcr = 2.486 which is only 0.75% larger than the
L2
π 2 EI
exact value of .
4 L2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-35
In the finite element method, there are many approximate functions, each
comparatively simple and each spanning a limited region of the structure. In addition the
degree of freedom are the actual displacements of specific points instead of generalized
coordinates.
Beam Element Formulation
Fig. 4.7
Consider the simply supported prismatic beam having a constant stiffness EI as
shown in Fig. 4.7. First, consider the whole beam as an element in order to see the error that is
occurred due to a coarse element. Then, we will divide the beam into more segments and redo
the analysis.
Selecting an approximate displacement function that closes to the actual displacement
function of the beam in a form of polynomial function
y ( x) = α 1 + α 2 x + α 3 x 2 + α 4 x 3
This function must satisfy the geometrical (displacement) boundary conditions of the
beam.
At x = 0 , y (0) = 0 , α1 = 0
At x = L , y ( L) = α 2 ( L) + α 3 ( L2 ) + α 4 ( L3 ) = 0 ,
α 2 = −α 3 L − α 4 L2
Thus,
y ( x) = α 3 ( x 2 − Lx) + α 4 ( x 3 − L2 x)
dy
= α 3 (2 x − L) + α 4 (3 x 2 − L2 )
dx
d2y
= 2α 3 + 6 xα 4
dx 2
Thus, U = EI (2α 32 L + 6α 3α 4 L2 + 6α 42 L3 )
For the static equilibrium, the total potential energy of the beam must be minimum.
∂V wo L2
=0; 2 EIα 3 + 3EILα 4 = −
∂α 3 12
∂V wo L2
= 0; 2 EIα 3 + 4 EILα 4 = −
∂α 4 12
wo L2
Solving the simultaneous equations for α 3 and α 4 , we have α 3 = − and α 4 = 0
24 EI
Since α 2 = −α 3 L − α 4 L2 , thus
wo L3
α2 =
24 EI
Substituting α 1 to α 4 into the assumed shape function, we have
wo
y ( x) = (− L2 x 2 + L3 x)
24 EI
Comparing the result with the solution by the mechanics of materials,
wo
y ( x) = ( x 4 − 2 L2 x 3 − L3 x)
24 EI
At the mid-span of the beam, we have the deflection by the approximated method and
the deflection by the classical method equal to
L 1 wo L4
y( ) =
2 96 EI
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-37
L 5 wo L4
y( ) =
2 384 EI
The error between these two methods is about 20%. However, we can improve the
solution by using a sine function as previously shown. It should be noted that if we use a
higher degree of the polynomial function, the solution is the same due to the nature of the
beam structure.
Compatibility
Another way to increase the accuracy of the solution is to break the beam down into
more pieces as an example shown in Fig. 4.8.
Fig. 4.8
If we choose to work this way, it is conceivable that each element will have a solution
function that is different from the others. If we are to assemble these elements in a
mathematical sense, there must be some compatibility requirements placed on the function of
the adjacent elements. This function is usually called the compatibility condition of the beam
elements.
The compatibility conditions of a structure require that
1. Within each region, the displacement varies smoothly with no discontinuity.
2. At the boundary between neighboring regions, the displacement matches each
other in a manner consistent with the problem under consideration.
3. At the boundary of the whole structure, the prescribed displacement boundary
conditions such as support conditions are satisfied.
The shape functions of the element i and i + 1 of the beam in Fig. 4.9 are
y i = α 1i + α 2i x + α 3i x 2 + α 4i x 3
y i +1 = α 1i +1 + α 2i +1 x + α 3i +1 x 2 + α 4i +1 x 3
Thus, the compatibility conditions of the junction between the element i and i + 1 of the
beam are
l
y i ( ) = y i +1 (0)
n
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-38
dy i dy i +1
l = 0
dx n
dx
Fig. 4.9
In general case, let consider the element i of a beam having the degree of freedom q1
to q 4 as shown in Fig. 4.10. The shape function of the element i of the beam can be
expressed as
y = α 1 + α 2ξ + α 3ξ 2 + α 4ξ 3
dy
= α 2 + 2α 3ξ + 3α 4ξ 2
dξ
Fig. 4.10
dy
At ξ = 0 , y = q1 and = q 2 . Thus,
dξ
q1 = α 1 + α 2 (0) + α 3 (0) + α 4 (0)
dy
At ξ = L , y = q3 and = q 4 . Thus,
dξ
q3 = α 1 + α 2 L + α 3 L2 + α 4 L3
q 4 = α 1 (0) + α 2 + 2α 3 L + 3α 4 L2
q1 1 0 0 0 α 1
q 0
2 1 0 0 α 2
=
q3 1 L L2 L3 α 3
q 4 0
1 2 L 3L2 α 4
This matrix expressing the relationship between the displacement q of the element i
and the generalized coordinate α of the shape function. It has a physical interpretation as
following:
1. When α 1 = 1 and other α = 0 , q1 = 1 , q 2 = 0 , q3 = 1 , and q 4 = 0
Fig. 4.11
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-40
or symbolically,
{α } = [T ]{q}
The shape function of the beam element can be written in the matrix form as
α 1
α
2
y = 1 x x2 x 3
α 3
α 4
L3 0 0 0 q1
3 1 0 L3 0 0 q 2
y = 1 x x2 x 3
L − 3L − 2 L2 3L − L2 q3
2 L −2 L q 4
q1
q
y = N 1 N2 N3 N 4 2
q 3
q 4
y = N {q} = {q} {N }
T
or
If we plot the shape function N i with respect to the coordinate x , we obtain 4 curves
as shown in Fig. 4.12. Physically, each of the 4 shape functions represents the deflection
curve for the beam element produced by setting the corresponding degree of freedom to be
one and the others to be zero.
Fig. 4.12
The first differentiation (slope) of the shape function of the beam element can be
obtained as
dy dN
= {q}
dx dx
where
dN 6 x 6 x 2 x 3x 2
2
4 x 3x 2 6x 6x 2
dx = − L2 + L3 1−
L
+ 2
L L2
− 3
L
−
L
+ 2
L
The second differentiation (curvature) of the shape function of the beam element can
be obtained as
d 2 y d 2N T d N
2
= {q} = {q} 2
dx 2 dx 2 dx
where
d 2 N 6 12 x 4 6x 6 12 x 2 6x
2 = − 2 + 3 − + − 3 − +
dx L L L L2 L2 L L L2
Let us consider the simply supported beam subjected to a uniformly distributed load as
shown in Fig. 4.13.
By using the finite element method, we divide the beam into four segments. The
positive degrees of freedom xi of each node of the beam are specified as shown. It should be
noted that in this case we use the notation of the degree of freedom xi replacing qi since qi is
the local degree of freedom of the beam element. However, when we consider the whole
beam, we need to use the global degree of freedom xi of the beam.
The total potential energy of a beam element i can be determined from the equation
V (i ) = U (i ) + Ω (i )
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-42
Fig. 4.13
Stiffness Equation of the Beam Element
The strain energy of the beam element can be written in the form of
2
1
L
d2y
U = EI ∫ 2 dx
2 0 dx
L
Ω = − ∫ wo ydy
0
L L
Ω = − ∫ {q}T wo {N }dx = −{q}T ∫ wo {N }dx
0 0
L
Let ∫ wo {N }dx = {Q} , then
0
Ω = −{q}T {Q}
For the beam in this case, we have
L/4
x3 1 x4
x − 2 + 3
L 2L
1 2 2 x3 1 x 4
L/4 2 x − 3 L + 4 L2
{Q} = wo ∫ {N }dx = wo 3 4
0 x − 1 x
L2 2 L3
3 4
− 1 x + 1 x
3 L 4 L2 0
Therefore, the total potential energy of a beam element i can be determined from
1
V (i ) = {q}T [k ]{q} − {q}T {Q}
2
From the principle of stationary (minimum) potential energy, we can determine the value of
the degree of freedom of each node of the beam.
∂V ( i ) ∂
∑ ∂q = ∂q ∑V (i ) = 0
i i
Let us consider the element 2 of the beam, the strain energy stored in this element is
T
x3 k112 k122 k132 k142 x3
2
1 x4 k 21 k 222 k 232 k 242 x 4
U ( 2) =
2 x5 k 312 k 322 k 332 k 342 x5
x 6 2
k 41 k 422 k 432 k 442 x 6
Then, the partial derivative of the strain energy with respect to the degree of freedom x3 to x6
T
1 k112 k122 k132 k142 x3
2
∂U ( 2 ) 1 0 k 21 k 222 k 232 k 242 x 4
= +
∂x3 2 0 k 312 k 322 k 332 k 342 x5
0 2
k 41 k 422 k 432 k 442 x 6
T
x3 0 0 0 0 x3
0
1 x4 0 0 0 x 4
+
2 x5 0 0 0 0 x5
x6
0 0 0 0 x 6
T
x3 k112 k122 k132 k142 1
2 2 2
1 x4 k 21 k 22 k 232 k 24 0
k 312 2
2 x5 k 322 k 332 k 34 0
x6 2 2 2
k 41 k 42 k 432 k 44 0
T
k112 x3
2 x
∂U ( 2)
k12 4
= 2
∂x3 k13 x5
k 2 x6
14
T T
k 212 x3 k12
2
x3
2 x k 2 x
∂U ( 2 ) k 22 4 22 4
= 2 = 2
∂x 4 k 23 x5 k 32 x5
k 2 x 6 k 2 x 6
24 42
T T
k 312 x3 k13
2
x3
2 x k 2 x
∂U ( 2)
k 32 4 23 4
= 2 = 2
∂x5 k 33 x5 k 33 x5
k 2 x6 k 2 x6
34 43
T T
k142 x3 k 41
2
x3
2 x k 2 x
∂U k 24
( 2)
4 42 4
= 2 = 2
∂x6 k 34 x5 k 43 x5
k 2 x 6 k 2 x 6
44 44
The external potential energy of the element 2 is
T
x3 Q3
x Q
4
Ω ( 2) = − 4
x5 Q5
x 6 Q6
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-45
Then, the partial derivative of the external potential energy with respect to the degree of
freedom x3 to x6 can be determined as following
T
1 Q3
0 Q
∂Ω ( 2)
4
= + 0 = −Q3
∂x3 0 Q5
0 Q6
∂Ω ( 2 )
= −Q4
∂x 4
∂Ω ( 2 )
= −Q5
∂x5
∂Ω ( 2 )
= −Q6
∂x6
Thus, the partial derivative of the total potential energy with respect to the degree of
freedom x3 to x6 can be found as
∂V ( 2 )
∂x3
∂V ( 2 ) k112 k122 k132 k142 x3 Q3
2 2
∂x 4 k 21 k 222 k 232 k 24 x 4 − Q4
( 2) = 2
∂V k 31 k 322 k 332 k 342 x5 Q5
∂x5 k 41 2
x6 Q6
2
k 422 3
k 43 k 44
( 2)
∂V
∂x
6
Rewriting the strain energy and the external potential energy of the element 2 in the
global coordinate, we have
U ( 2) , x1
( 2) 0 0 0 0 0 0 0 0 0 0 x1
U , x 2 0
0 0 0 0 0 0 0 0 0 x 2
U ( 2) , x
3 0 0 k112 k122 k132 k142 0 0 0 0 x3
U , x 4 0
( 2 )
0 k 212 2
k 22 k 232 k 242 0 0 0 0 x 4
( 2)
U , x5 0 0 k 312 k 322 k 332 k 342 0 0 0 0 x5
( 2) =
U , x 6 0 0 k 412 2
k 42 k 432 k 442 0 0 0 0 x 6
U ( 2) , x 0 0 0 0 0 0 0 0 0 0 x 7
7
U , x8 0
( 2 ) 0 0 0 0 0 0 0 0 0 x8
( 2) 0 0 0 0 0 0 0 0 0 0 x9
U , x 9
( 2) 0 0 0 0 0 0 0 0 0 0 x10
U , x10
and
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-46
Ω ( 2) , x1
( 2) 0
Ω , x 2 0
Ω ( 2) , x
3 Q3
Ω ( 2) , x 4
( 2) Q4
Ω , x5 Q5
{ }
Ω ( 2 ) , xi = ( 2 ) = −
Ω , x 6 Q6
Ω ( 2) , x 0
7
Ω ( 2) , x8 0
( 2) 0
Ω , x9
( 2) 0
Ω , x10
In general, we have
{U (i )
, xi } 10 x1
= [k i ]10 x10 {x}10 x1
{Ω (i )
, xi } 10 x1
= {Q}10 x1
and the total potential energy of all four elements of the beam is
4
{V } = ∑ [U (i ) + Ω (i ) ]
i =1
For equilibrium, the partial derivative of the total potential energy of the beam with
respect to the degree of freedom xi must be zero.
4
{V , xi } = ∑ [U (i ) , xi + Ω (i ) , xi ] = {0}
i =1
4 4
{V , xi } = ∑ [k i ]{x} + ∑ {Q} = {0}
i =1 i =1
4 4
where ∑ [k i ] is the global stiffness matrix of the beam and
i =1
∑{Q} is the load vector of the
i =1
4
beam. Symbolically, the global stiffness matrix ∑ [k i
] is usually written by using [K ] . In
i =1
details, the partial derivative of the total potential energy of the beam with respect to the
degree of freedom xi can be written as shown in the next page. In this equation, since the
supports are pin and roller, the degrees of freedom x1 and x9 are known to be zero. It should
be noted that the obtained global stiffness matrix [K ] is the same as one that we found by
using the matrix structural analysis.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-47
Q (1)
V , x1 k111 k121 k131 k141 0 0 0 0 0 0 x1 1 0
V , x 1 1 1 1 x Q2(1) 0
2 k 21 k 22 k 23 k 24 0 0 0 0 0 0 2 (1) ( 2)
V , x3 k 31 1 1
k 32 1
k 33 + k112 1
k 34 + k122 k132 k142 0 0 0 0 x3 Q3 + Q1 0
1 (1) ( 2)
V , x 4 k 41 + k 21 + k 222 x 4 Q4 + Q2 0
1 1 2 1
k 42 k 43 k 44 k 232 k 242 0 0 0 0
V , x 0 0 k 312 k 322 k 332 + k113 k 342 + k123 k133 k143 0 0 x5 Q3( 2 ) + Q1( 3) 0
5
= − ( 2) =
V , x 6 0 0 2
k 41 2
k 42 k 432 + k 21
3 2
k 44 + k 22
3 3
k 23 3
k 24 0 0 x6 Q4 + Q2( 3) 0
V , x7 0 0 0 0 k 313 k 323 k 333 + k114 3
k 34 + k124 k134 k144 x7 Q3( 3) + Q1( 4 ) 0
V , x8 0 + k 214 + k 22 x8 Q ( 3) + Q ( 4 ) 0
3 3 3 3 4
0 0 0 k 41 k 42 k 43 k 44 k 234 4
k 24
V , x 0 4 2
0 0 0 0 0 k 314 k 324 k 334 k 344 x9 ( 4 ) 0
9
Q3
V , x10 0 0 0 0 0 0 k 414 k 424 k 434 4
k 44 x10 ( 4 ) 0
Q4
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-48
Let us redefine the degree of freedom of the beam from xi to xi′ as shown in the
figure. The degree of freedom xi′ is arranged so that the known degree of freedom
(displacement at the supports) are numbered first and followed by the unknown degree of
freedom. This kind of set-up will help us to partition the global stiffness matrix, which will
ease the matrix manipulation.
Fig. 4.14
From the Fig. 4.14, we can relate the degree of freedom xi with the degree of freedom
x1 1 0 0 0 0 0 0 0 0 0 x1′
x
2 0 0 1 0 0 0 0 0 0 0 x 2′
x 3 0 0 0 1 0 0 0 0 0 0 x3′
x 4 0 0 0 0 1 0 0 0 0 0 x 4′
x 0 0 0 0 0 1 0 0 0 0 x5′
5
=
x
6 0 0 0 0 0 0 1 0 0 0 x6′
x 7 0 0 0 0 0 0 0 1 0 0 x7′
x8 0 0 0 0 0 0 0 0 1 0 x8′
x 0 1 0 0 0 0 0 0 0 0 x9′
9
x10 0 0 0 0 0 0 0 0 0 1 x10
′
1
V = {x}T [ K ]{x} − {x}T {Q}
2
1
= {x ′} [T ] [K ][T ]{x ′} − {x ′} [T ]{Q}
T T T
2
1
= {x ′}T [ K ′]{x ′} − {x ′}T {Q ′}
2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 4-49
Taking the partial derivative of the total potential energy of the beam with respect to
the degree of freedom xi′ and setting the result to be zero, we have
∂V
= {V , xi′} = [K ′]{x ′} − {P ′} = 0
∂xi′
Partitioning the global stiffness matrix by separating the known displacements at the
supports from the unknown displacements, we have
K R′ ∆ ′ ∆ Ps′ R
K Rx
K ′ − + = {0}
P∆ ′ x ′ P ′ 0
K Px
where ∆ is the known support settlements, x ′ is the unknown displacements, Ps′ is the
known external loads acting on the supports, P ′ is the known external loads, and R is the
unknown support reactions. It should be noted that [K ′] is a symmetric matrix and
[K P′ ∆ ] = [K Rx′ ]T and vice versa. Thus, we can find the unknown displacements x ′ from the
equation
[K P′ ∆ ]{∆} + [K Px′ ]{x ′} − {P ′} = {0}
{x′} = [K Px′ ]−1 {{P ′} − [K P′ ∆ ]{∆}}
Then, we can solve for the unknown support reactions from
[K R′ ∆ ]{∆} + [K Rx′ ]{x ′} − {Ps′} + {R} = {0}
{R} = −[K R′ ∆ ]{∆} − [K Rx′ ]{x ′} + {Ps′}
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-1
Chapter 5
Static Failure and Failure Criteria
5.1 Definition of Failure
Failure can be defined as any changes in the size, shape or material properties of a
structure or mechanical part the render it incapable of satisfactorily performing its intended
functions.
Failure can be caused by the following agents:
Forces: steady, dynamic, transient, cyclic, random
Time: very short, short, long time
Temperature: low, elevated, room, steady, random, cyclic, transient
Environment: Chemical, nuclear, rain, sand
5.2 Modes of Failure
When a structural member is subjected to loads, its response depends not only on the
type of material, but also on the types of loads and environment conditions. Thus, the modes
of failure can be classified as
Yielding failure – The plastic deformation in the structure under operational loads
that is large enough to interfere with the ability of the structure to satisfactorily performing its
intended functions as an example shown in Fig. 5.1.
Fig. 5.1
Force induced elastic deformation – The elastic deformation that is recoverable in a
structure under operational loads become large enough to interfere with the ability of the
structure to satisfactorily performing its intended functions such as stiffness loss as an
example shown in Fig. 5.2.
Fig. 5.2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-2
Ductile failure – The plastic deformation in the structure that exhibits ductile
behavior and is carried to the extreme so that it separates into two or more pieces as an
example shown in Fig. 5.3.
Fig. 5.3
Brittle failure – The elastic deformation in the structure that exhibits ductile behavior
and is carried to the extreme so that it separates into two or more pieces as an example shown
in Fig. 5.4.
Fig. 5.4
Fatigue failure – The separation of a structure into two or more pieces or a certain
size of crack initiation as a result of fatigue load or deformation for a period of time.
Low cycle fatigue: fatigue life < 10 5 cycles
High cycle fatigue: fatigue life ≥ 10 5 cycles
Thermal fatigue
Sonic fatigue
Buckling failure – The deflection of a structure suddenly increased greatly with only
a slight change in load. The buckled part is no longer capable of performing its intended
function as an example shown in Fig. 5.5.
Fig. 5.5
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-3
Creep failure – The plastic deformation in a structure under the influence of stress or
temperature over a period of time and becoming so large enough to interfere with the ability
of the structure to satisfactorily performing its intended function.
5.3 Failure Criteria
The stress analysis itself can not be able to predict the failure of a structure. To know
about how high stress can a structure be sustained or how high the strength of the structure, a
failure criteria is needed.
The criteria discussed in this section will be focused on the failure due to static loads
such as force-induced failure, yielding, ductile, and brittle failure. Once the state of stresses
at a critical point on a structure is determined, the principal stresses can be computed,
and the failure criteria can be used.
5.3.1 Maximum principal normal stress fracture criterion
Experimental observations show that brittle isotropic materials such as cast iron tend
to fail suddenly by fracture without yielding.
Failure will occur when the maximum principal normal stress become equal to or
exceed the maximum normal stress in a simple tension (or compression) test using a specimen
of the same material.
Mathematically, if the material is subjected to plane stress the failure will occur when
σ1
≥ σ ult
σ2
Fig. 5.6
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-4
Any stress falling within the rectangular indicates that the material behave elastically.
Points on the hexagon indicate that the material is failing by separation or fracture.
5.3.2 Maximum shear stress yield criterion
Experimental evidence indicates that, in ductile isotropic material such as mild steel,
slip occurs during yielding along critical oriented planes. This suggests that the maximum
shearing stress play an important role in the failure of the ductile materials.
Failure will occur when the magnitude of the absolute maximum principal shear stress
becomes equal to or exceed the maximum shear stress in a simple tension test using a
specimen of the same material.
Mathematically, if the material is subjected to plane stress the failure will occur when
σ1 −σ 2
τ abs = ≥ τy
max 2
σy −0 σy
τ y = maximum shear strength obtained from the tension test = = .
2 2
Thus, σ1 −σ 2 ≥ σ y
σ σ
± 1 − 2 =1 if σ 1 and σ 2 have the opposite signs
σ y σ y
σ
± 1 = 1
σ y
if σ 1 and σ 2 have the same signs
σ 2
± = 1
σ y
Fig. 5.7
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-5
A plot of these equations is in the hexagonal shape and is shown in Fig. 5.7. Any
stress falling within the hexagon indicates that the material behave elastically. Points on the
hexagon indicate that the material is yielding.
5.3.3 Maximum principal normal strain fracture criterion
Failure will occur when the maximum principal normal strain become equal to or
exceed the ultimate strain in a simple tension (or compression) test using a specimen of the
same material.
This criterion is an improvement over the maximum principal stress criterion, but it
does not reliably predict failure by yielding. In practice, this criterion is rarely used excepting
in the design of thick-walled cylinder. Mathematically, if the material is subjected to plane
stress the failure will occur when
ε 1 ≥ ε ult
or ε 2 ≥ ε ult
σ1 σ σ σ
where ε 1 , ε 2 are the principal strains which are ε 1 = − ν 2 and ε 2 = 2 − ν 1 .
E E E E
σ ult
ε ult = = maximum tensile (or compressive) strain obtained from the tension test
E
Thus, ± (σ 1 − νσ 2 ) = σ ult
± (σ 2 − νσ 1 ) = σ ult
σ2
σ ult
1
σ1
-1 1 σ ult
-1
Fig. 5.8
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-6
1 1 1
U o = σ 1ε 1 + σ 2 ε 2 + σ 3ε 3
2 2 2
Since the strain-stress relations in the form of principal strains and principal stresses
1 1 1
are ε 1 = (σ 1 − νσ 2 − νσ 3 ) , ε 2 = (σ 2 − σ 1 − νσ 3 ) , and ε 3 = (σ 3 − νσ 2 − νσ 1 ) , then, we
E E E
can write the total strain energy density as
Uo =
1
2E
[
σ 12 + σ 22 + σ 32 − 2ν (σ 1σ 2 + σ 2σ 3 + σ 3σ 1 ) ]
Fig. 5.9
From chapter 1, we have the mean stress or hydrostatic stress acting on a stress
σ1 + σ 2 + σ 3
element which is the average of the principal stresses σ avg . = . The
3
corresponding hydrostatic strains can be determined as
1
ε avg . = (1 − 2ν )σ avg .
E
The dilation strain energy density,
1 σ + σ 2 + σ 3 1 − 2ν σ 1 + σ 2 + σ 3
U o ,v = 3 1
2 3 E 3
2
3 1 − 2ν σ 1 + σ 2 + σ 3
U o ,v =
2 E 3
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-7
1 1 +ν
U o,d =
2E 3
(
(σ 1 − σ 2 ) + (σ 2 − σ 3 ) + (σ 3 − σ 1 )
2 2 2
)
The distortion strain energy density in simple tension test at failure is
1 +ν 2
U o ,dy = σ y
3E
Mathematically, if the material is subjected to a general state of stresses the failure
will occur when U o ,d = U o ,dy . Thus,
1 1 +ν 1 +ν
2E 3
(
(σ 1 − σ 2 ) + (σ 2 − σ 3 ) + (σ 3 − σ 1 ) =
2 2 2
) 2
σ y
3E
(σ 1 − σ 2 ) 2 + (σ 2 − σ 3 ) 2 + (σ 3 − σ 1 ) 2 = 2σ y2
If the material is subjected to plane stress, the failure will occur when
(σ 1 − σ 2 ) 2 + (σ 2 ) 2 + (−σ 1 ) 2 = 2σ y2
σ 12 + σ 22 − σ 1σ 2 = σ y2
2 2
σ1 σ1 σ 2 σ 2
− + =1
σ σ σ σ
y y y y
This is an equation of an ellipse. The plot of this equation is shown in Fig. 5.10. Any
stress falling within the ellipse indicates that the material behave elastically. Points on the
ellipse indicate that the material is yielding. Experimental investigations show that this
criterion is best fitted for isotropic materials that fail by yielding or ductile rupture.
Fig. 5.10
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-8
τ oct =
1
3
[
(σ 1 − σ 2 ) 2 + (σ 2 − σ 3 ) 2 + (σ 3 − σ 1 ) 2 ]
1/ 2
or
τ oct =
1
3
[
(σ x − σ y ) 2 + (σ x − σ z ) 2 + (σ y − σ z ) 2 + 6τ xy2 + 6τ xz2 + 6τ yz2 ]
1/ 2
2
τ oct = σy
3
Failure will occur when the maximum octahedral shearing stress become equal to or
exceed the octahedral shearing stress at failure in a simple tension test using a specimen of
the same material.
5.3.6 Coulomb-Mohr fracture criterion
In some brittle materials such as gray cast iron and concrete, the tension and
compression properties are different. The failure of these materials should be predicted by
using the Coulomb-Mohr criterion.
In the Coulomb-Mohr criterion, the fracture is hypothesized to occur on a given plane
in the material when a critical combination of shear and normal stress acts on this plane. The
simplest mathematical relation giving the critical combination of stresses is in the form of
linear relationship. At fracture, we have
τ + µσ = τ i
where τ and σ are the shearing stress and normal stress acting on the fracture plane,
respectively, and µ and τ i are constants for a given material. This equation forms a line on a
plot of σ versus τ and the line has a slope of − µ and intercepts the τ axis at τ i as shown
in Fig. 5.11.
Consider a stress element subjected to the principal stress σ 1 , σ 2 , and σ 3 . The
Mohr’s circle of the state of stresses can be drawn as shown in Fig. The failure is occurred if
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-9
the largest of the three circles touch the line. Thus, the line represents a failure envelope for
the Mohr’s circle.
Fig. 5.11
As shown in the figure, the touching point has the coordinate ( σ ′, τ ′ ) where
σ1 + σ 3 σ1 −σ 3
σ′ = + cos φ
2 2
σ1 −σ 3
τ′ = sin φ
2
Also, the failure planes are occurred where the maximum principal stress acts by a rotation
φ / 2 in either direction and
1
tan φ =
µ
1
Substituting σ ′ , τ ′ , and tan φ = into the relation τ + µσ = τ i , we have
µ
σ 1 − σ 3 + m(σ 1 + σ 3 ) = 2τ u
µ τi
where m = = cos φ and τ u = = τ i sin φ .
1+ µ 2 1+ µ 2
τ + τ + m(τ − τ ) = 2τ u
τ = τu
The τ u is the pure shear necessary to causes fracture. The corresponding largest Mohr’s circle
Fig. 5.12
For uniaxial tensile test, we have σ 1 = σ ut and σ 2 = σ 3 = 0 . Thus,
σ ut − 0 + m(σ ut + 0) = 2τ u
2τ u
σ ut =
1+ m
For uniaxial compression test, we have σ 1 = σ 2 = 0 and σ 3 = −σ uc . Thus,
0 + σ uc + m(0 − σ uc ) = 2τ u
2τ u
σ uc =
1− m
2τ u
It should be noted that σ uc must have a negative value or σ uc = − since it is the
1− m
compressive stress. The fracture planes predicted by the Coulomb-Mohr criterion for uniaxial
tensile test and uniaxial compression test are shown in Fig 5.13.
Fig 5.13
Eliminating τ u , we have
1+ m
σ uc = − σ ut
1− m
Solving for m , we obtain
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-11
σ uc + σ ut
m=
σ uc − σ ut
We can see that for a positive value of m , the strength in tension is predicted to be
less than that in compression.
If the subscript for the principal stresses are assumed to be arbitrary assigned, then,
σ 1 − σ 3 + m(σ 1 + σ 3 ) = 2τ u
σ 2 − σ 3 + m(σ 2 + σ 3 ) = 2τ u
σ 1 − σ 2 + m(σ 1 + σ 2 ) = 2τ u
For plane stress,
σ 1 + mσ 1 = 2τ u
σ 2 + mσ 2 = 2τ u
σ 1 − σ 2 + m(σ 1 + σ 2 ) = 2τ u
The plot of this equation is shown in Fig. 5.14.
Fig. 5.14
It should be noted that the Coulomb-Mohr criterion with the constant m = 0 is
equivalent to the maximum shear stress criterion and Fig. 5.14 will be the same as Fig. 5.7.
5.4 Comparison of the Failure Criteria
Fig. 5.15 shows the experimental results with the failure criteria presented before. It is
concluded that
1. The maximum principal stress criterion is best fitted for isotropic material that is
failed by the brittle fracture.
2. The maximum distortion energy criterion is best fitted for isotropic materials that
fail by yielding or ductile rupture.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-12
Fig. 5.15
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-13
Example 5-1
The stesolid shaft as shown in Fig. Ex 5-1a has a radius of 12.7 mm and is made of
steel having σ y = 250 MPa . Determine if the loadings cause the shaft to fail according to the
maximum shearing stress criterion and the maximum distortion energy criterion.
Fig. Ex5-1
Let the x - axis is in the longitudinal direction of the shaft. The averaged axial stress
due to the axial force is
70
σx = = 138.15 MPa
π (0.0127) 2
The maximum shear stress caused by the torque is
370(0.127)
τ xy = = 115.0 MPa
π (0.0127) 4
2
The stress element at point A is as shown in Fig. Ex 5-1b.
The principal normal stresses due to the state of stresses as shown in Fig. Ex 5-1b are
2
− 138.15 + 0 − 138.15 − 0
σ1 = ± + 115.0
2
2 2 2
σ 1 = 65.07 MPa
σ 2 = −203.23 MPa
Maximum shearing stress criterion
Since the principal normal stresses have the opposite sign,
σ σ ?
± 1 − 2 ≤1
σ y σ y
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-14
65.07 − 203.23
250 − 250 = 1.073 > 1.0
Thus, the loadings cause the shaft to fail according to the maximum shearing stress criterion.
Maximum distortion energy criterion
2 2
σ1 σ1 σ 2 σ 2 ?
− + ≤1
σ σ σ σ
y y y y
2 2
65.07 65.07 − 203.23 − 203.23
− + = 0.940 < 1.0
250 250 250 250
Thus, the loadings do not cause the shaft to fail according to the maximum distortion energy
criterion.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-15
Example 5-2
A circular cylindrical shaft is made of steel with σ Y = 700 MPa , E = 200 GPa , and
ν = 0.29 . The shaft is subjected to a static bending M = 13.0 KN - m and a static torque of
T = 30.0 kN - m as shown in Fig. Ex 5-2. Using the factor of safety of SF = 2.60 , determine
the minimum diameter of the shaft based on the maximum octahedral shearing stress
criterion (or equivalently the maximum distortion energy criterion) and the maximum
shearing stress criterion.
Fig. Ex 5-2
The shaft is subjected to a static bending moment M = 13.0 KN - m and a static torque
of T = 30.0 kN - m . However, due to the factor of safety of SF = 2.60 , the moment and the
torque must be increased by the factor. Thus, if we let the x - axis is in the longitudinal
direction of the shaft, the stresses due to the loadings are
Mc 32( SF ) M
σ x = SF =
I πd 3
Tc 16( SF )T
τ xy = SF =
J πd 3
For the maximum octahedral shearing stress criterion,
2
τ oct (max) = σY
3
1 2
(σ x − σ y ) 2 + (σ x − σ z ) 2 + (σ y − σ z ) 2 + 6τ xy2 + 6τ xz2 + 6τ yz2 = σY
3 3
1 2
2σ x2 + 6τ xy2 = σY
3 3
σ Y = σ x2 + 3τ xy2
or
1/ 3
16( SF )
d min = 4 M 2 + 3T 2
πσ Y
d min = 103 mm
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 5-16
1 σ
σ x 2 + 4τ xy2 = Y
2 2
Substituting the stresses into the obtained equation and rearranging the term, we get
1/ 3
32( SF )
d min = M 2 +T 2
πσ Y
d min = 107 mm
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-1
Chapter 6
Introduction to Fracture Mechanics
6.1 Introduction
Traditionally, the structural design approaches are based on the concept that the
structures must have enough strength, stiffness, and stability to resist the loads.
For the strength criteria, the applied stress must be less than the yielding or ultimate
strength of the material. However, when a crack is occurred in a component of the structure, it
can cause the failure (in the form of fracture) at stresses well below the material's yielding
strength. In this case, a special methodology called fracture mechanics can be used in design
to minimize the possibility of failure.
Fracture mechanics is important in engineering design since cracks and crack-like
flaw occur more frequently than we might expect. For example, the periodic inspections of
large commercial aircraft frequently reveal cracks that must be repaired. Also, they are
commonly occurred in ship structures, in bridge structures and in pressure vessel and piping.
The ability of a given material to resist a crack depends principally on the toughness
of the material. Generally, fracture toughness in some metals such as steel increase with
temperature as shown in Fig. 6.1. Also, there is an especially abrupt change in toughness over
a relative small temperature range such as − 50 o C for A469 steel. The temperature at this
point is called the transition temperature. Thus, the fracture of the steel can be promoted by
the temperature below the transition temperature.
Fig. 6.1
Generally, fracture process can be categorized into three stages.
1. Crack initiation – micromechanics and dislocation theory
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-2
Fig. 6.2
The following questions are important in designing a structure using fracture
mechanics. σ
1. What is the maximum permissible crack size?
2. What is the residual strength as a function of crack size?
3. How long does it take from the maximum detectable crack size to the maximum
permissible crack size?
4. During the period available for crack detection, how often should the structure be
inspected for crack?
6.2 Fracture Modes
Once a crack has been initiated, subsequent crack propagation may occur in several
ways depending on the relative displacement of the particles in the two faces (surfaces) of the
crack. There are three fundamental modes of fracture acting on the crack surface displacement
as shown in Fig. 6.3.
1. Opening mode (Mode I) - The stress acts perpendicular to the crack growth
direction and the crack growth plane. The crack surfaces move directly apart.
2. Shearing mode (Mode II) - The stress acts parallel to the crack growth direction
and the crack growth plane. The crack surfaces move (slide) normal to the crack
edge and remain in the plane of the crack.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-3
3. Tearing mode (Mode III) - The stress acts perpendicular to the crack growth
direction and parallel to the crack growth plane. The crack surfaces move parallel
to the crack edge and remain in the plane of the crack.
The most general cases of crack surface displacements are obtained by superposition
of these basic three modes.
Fig. 6.3
6.3 Stress and Displacement Field at the Crack Tip
In 1950, Irwin showed that the local stresses near the crack tip, as the curvature at the
crack tip goes to zero as shown in Fig. 6.4, are of the form
K
σ ij = f ij (θ )
2πr
and we can see that σ ij → α as r → 0 . Thus, the stress field is a singular stress field with a
singularity of r . The term K is called the stress intensity factor, which defines the
intensities or magnitudes of the singular stress around the crack tip. The expression of K
depends on the fracture modes.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-4
Fig. 6.4
6.3.1 Mode I
Fig. 6.5
The stresses at a point having a distance r and angle θ from the crack tip and for
Mode I as shown in Fig. 6.5 are
KI θ θ 3θ
σx = cos 1 − sin sin
2πr 2 2 2
KI θ θ 3θ
σy = cos 1 + sin sin
2πr 2 2 2
KI θ θ 3θ
τ xy = cos sin cos
2πr 2 2 2
Plane stress σ z = τ yz = τ xz = 0
Plane strain σ z = ν (σ x + σ y )
τ yz = τ xz = 0
The displacements at a point having a distance r from the crack tip and angle θ with
the x axis are
KI r θ 3θ
ux = cos 2 (2 K − 1) − cos 2
8G 2π
u x (θ ) = u x (−θ )
KI r θ 3θ
uy = sin 2 (2 K − 1) − sin 2
8G 2π
u y (θ ) = −u y (−θ )
3 − 4ν plane strain
where K = 3 −ν
1 + ν plane stress
6.3.2 Mode II
Fig. 6.6
The stresses at a point having a distance r from the crack tip and angle θ with the x
axis for Mode II as shown in Fig. 6.6 are
K II θ θ 3θ
σx = − sin 2 + cos cos
2πr 2 2 2
K II θ θ 3θ
σy = cos sin cos
2πr 2 2 2
K II θ θ 3θ
τ xy = cos 1 − sin sin
2πr 2 2 2
Plane stress σ z = τ yz = τ xz = 0
Plane strain σ z = ν (σ x + σ y )
τ yz = τ xz = 0
where K II = τ o πa
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-6
The displacements at a point having a distance r from the crack tip and angle θ with
the x axis for Mode II are
K II 2r θ 3θ
ux = (2 K + 3) sin + sin
8G π 2 2
u x (θ ) = −u x (−θ )
K II 2r θ 3θ
uy = − (2 K − 3) cos − cos
8G π 2 2
u y (θ ) = u y (−θ )
3 − 4ν plane strain
where K = 3 −ν
1 + ν plane stress
Fig. 6.7
The stresses at a point having a distance r from the crack tip and angle θ with the x
axis for Mode III as shown in Fig. 6.7 are
K III θ
τ xz = − sin
2πr 2
K III θ
τ yz = cos
2πr 2
σ z = σ y = σ z = τ xy = 0
where K III = τ o πa
The displacements at a point having a distance r from the crack tip and angle θ with
the x axis for Mode III are
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-7
K III 2r
uz = sin θ and ux = uy = 0
G π
6.4 Stress Intensity Factor (SIF or K )
Stress intensity factors are needed to measure the intensity or magnitude of the
singular stress field in the vicinity of an ideally sharp crack tip in a linear elastic and
isotropic material. This approach is called linear-elastic fracture mechanics (LEFM). The
factors do depend on loading condition, crack size, crack shape, and geometric boundaries.
The general form of the stress intensity factors is given by
K = f .σ πa
where σ = applied stress
a = effective crack length
f = correction factor. For infinity plate, f = 1 .
Thus, stress intensity factor K has a unit in ksi in or MPa m . The solutions of the stress
intensity factors have been obtained for wide variety of problems and published in a
handbook form. The followings are the typical solution for SIF:
6.4.a Center crack
Fig. 6.8
a
K I = σ o πa . f
2b
a πa
f = sec
2b 2b
6.4.b Double edge crack
a
K I = σ o πa . f
b
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-8
2 3
a a a a
f = 1.12 + 0.203 − 1.197 + 1.930
b b b b
Fig. 6.9
6.4.c Single edge crack
Fig. 6.10
a
K I = σ o πa . f
b
2 3 4
a a a a a
f = 1.12 − 0.231 + 10.55 − 21.72 + 30.39
b b b b b
6.4.d Crack under bending
Fig. 6.11
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-9
a
K I = σ o πa . f
b
6M
σ=
b 2t
2 3 4
a a a a a
f = 1.12 − 1.40 + 7.33 − 13.08 + 14
b b b b b
6.5 Superposition of SIF
Stress intensity factor for combined loading can be obtained by the superposition
method, that is, by adding the contribution to K from the individual load components. It is
valid only for combination of the same mode of failure.
Consider an eccentric load applied at a distance e from a centerline of a member with
a single edge crack as shown in Fig. 6.12. This eccentric load is statically equivalent to the
combination of a centrally applied tension load and a bending moment.
The stress intensity factor for the centrally applied tension load is
P a
K Ia = πa . f a
bt b
The stress intensity factor for the bending moment is
6M a
K Ib = 2
πa . f b
b t b
Thus, the total stress intensity factor of this case is
P a a 6e b a
K I = K Ia + K Ib = f + f πa
bt b b b
Fig. 6.12
6.6 Fracture Toughness (Critical SIF)
Fracture toughness K IC is the critical value of the stress intensity factor K . If the
stress intensity factor K occurred in a given material is less than the fracture toughness, the
material will have ability to resist the crack without brittle fracture.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-10
It should be noted that in order to ensure that the state of stress is plane strain for each
of the cases in Section 6-4, the magnitudes of the crack half-length a and the thickness t
should satisfy
2
K
a, t ≥ 2.5 IC
σ
y
Table 6-1 shows K IC at the room temperature for several metals.
Table 6-1
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-11
Example 6-1
Determine the stress intensity factor for the edge-cracked beam having the crack half-
length a of 1.75 in. when subjected to a moment of 100 kips - ft . It the beam was made of an
extremely tough steel that has σ y = 195 ksi and a K IC of 150 ksi. in. . The width of the
beam is 4 in. and the depth of the beam is 12 in. . If the moment applied to the beam was
increased to 400 kips - ft , would this beam fail?
The flexural stress due to the moment 100 kips - ft is
Mc M (b / 2) 6 M 6(100)(12)
σ= = = = = 12.5 ksi
I (tb 3 / 12) tb 2 4(12 2 )
Since the crack half-length a of the beam is 1.75 in. ,
a 1.75
= = 0.14583
b 12
2 3 4
a a a a a
f = 1.12 − 1.40 + 7.33 − 13.08 + 14 = 1.329
b b b b b
The stress intensity factor for the edge-cracked beam is
K I = 12.5 π (1.75) (1.329) = 38.95 ksi. in.
Since the stress intensity factor is larger than K IC of 150 ksi. in. and the flexural stress is
less than the yielding strength σ y = 195 ksi , the beam does fail by fracture.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-12
Example 6-2
A tool as shown in Fig. Ex 6-2 is used to dig up old road beds before replacing them.
Let the tool be made of AISI 4340. The dimensions of the tool are d = 250 mm , b = 60 mm ,
and the width t = 25 mm . Determine the magnitude of the fracture load P for the crack
length of a = 5 mm .
Fig. Ex 6-2
The crack half-length a and the thickness t satisfy the condition
K 2
P 6(280 P)
59 1000 = (1.163) + 2
(1.047) π (5)
25(60) (60 )25
P = 23.17 kN
The total maximum stress is
P 6M 23.17 6(0.280)23.17
σ= + 2 = + = 448 MPa < σ y = 1503 MPa
bt b t 0.025(0.060) 0.060 2 (0.025)
which is in accordance with the assumption of linear elastic behavior.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-13
Fig. 6.13
If the crack moves ahead by a small distance da , while the displacement is held
constant, the stiffness of the member decreases as shown in Fig. 6.13b. This results in the
decreasing in the strain energy by the amount of dU , that is, U decreases due to a release of
this amount of energy.
The strain energy release rate ( G ) is defined as the rate of change of strain energy
with increase in crack area .
∂U 1 ∂U
G=− =−
∂A t ∂a
lb − in lb
where t = thickness of the plate. Since G has a unit in = , G is sometimes
in 2 in
considered as a crack driving force.
For plane stress,
K I2
GI =
E
K II2
G II =
E
1 +ν 2
G III = K III
E
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-14
real materials can not support these theoretical infinite stresses. Thus, upon loading, the crack
tip becomes blunted and a region of yielding or microcracking forms. This region of yielding
is called plastic zone.
Fig. 6.14
For any cases of Mode I loading, the stresses near the crack tip are
KI θ θ 3θ
σx = cos 1 − sin sin
2πr 2 2 2
KI θ θ 3θ
σy = cos 1 + sin sin
2πr 2 2 2
KI θ θ 3θ
τ xy = cos sin cos
2πr 2 2 2
Plane stress σ z = τ yz = τ xz = 0
where K I = σ o πa
For plane stress, the state of stress at the plane of the crack where the angle θ = 0 o is
KI
σx =σy =
2πr
σ z = τ yz = τ xz = τ xy = 0
Since all shear stress along the plane of the crack are zero, σ x , σ y , and σ z are the
principal normal stresses. The maximum shear stress and the maximum octahedral shearing
stress criteria estimate the yielding at
σ x = σ y = σ ys
where σ ys is the yielding strength. Therefore, we obtain the radius of the plastic zone for the
zone size for plane stress condition is larger than that of the plane strain condition.
For cyclic loading, the cyclic plastic zone size can be determined by
For plane stress,
2
1 KI
ry =
8π σ ys
For plane strain,
2
1 KI
ry =
24π σ ys
Hence, we can see that ry cyclic < ry .
static
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 6-16
Example 6-3
A large plate made of 4140 steel ( σ ys = 90 ksi ) containing a 0.2 in. center crack is
a πa
f = sec = 1.0
2b 2b
Chapter 7
Fatigue
7.1 Introduction
Structural members and mechanical parts are often found to have failed under the
action of repeated or fluctuating stresses called fatigue failure. The actual maximum repeated
stresses were well below the ultimate strength of the material and quite frequently evens
below the yield strength. Typical fatigue failures do not involve macroscopic plastic
deformation.
Fatigue failure often begins with a small crack. The initial crack is so minute that it
can not be detected by the naked eyes and even by the X-ray method. The small cracks are
usually developed at high stress gradient area. Once crack is developed, the stress gradient
becomes larger and larger, and the crack progresses more rapidly.
At present, there are three major approaches to analyzing and designing against
fatigue failure. They are stress-based approach, strain-based approach, and fracture mechanic
approach.
Stress-based approach is based on the nominal stresses in the region of the component
being analyzed. The nominal stress that can be resisted under cyclic loading is determined by
considering mean stresses and by making adjustments for the effects of stress risers such as
holes and fillet.
Strain-based approach involves more detailed analysis of the localized yielding that
may occur at stress risers during cyclic loading.
Fracture mechanic approach is used to treat growing crack due to cyclic loading by
using the method of fracture mechanics.
7.2 Nomenclature
Some practical applications involve cycling between maximum and minimum stress
levels that are constant. This is called constant amplitude stressing as shown in Fig. 7.1 and
7.2.
The following nomenclatures for cyclic loading are important.
Mean stress is the average of the maximum stress and the minimum stress.
σ max + σ min
σm =
2
Stress range is the difference between the maximum stress and the minimum stress.
∆σ = σ max − σ min
σ max − σ min
σa =
2
Stress ratio is the ratio of the minimum stress over the maximum stress.
σ min
R=
σ max
Amplitude ratio is the ratio of the stress amplitude over the mean amplitude
σa
a=
σm
Cyclic loading needs two independent variables to specify. Some combinations that
may be used are: σ a and σ m , σ max and R , ∆σ and R , σ max and σ min and σ a and a .
Fig. 7.1
Fig. 7.2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-3
kt S .
also increased from 0 to σ max by following the dashed line of the stress-strain curve. Then,
when we unload the specimen from the strain ε max to ε min , the stress-strain curve follows the
unloaded line, and the stress is decreased from σ max to σ min . Finally, if we reload the
specimen from ε min to ε max , the stress is increased from σ min to σ max by following the reload
curve. It can be seen that there is a loop occurred due to the unloading and reloading the
specimen. This loop is called hystereis loop. It represents measurement of plastic deformation
work done on the material. The area within the loop is the energy per unit volume dissipates
during a cycle.
Fig. 7.3
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-4
Bauschinger Effect
Consider the monotonic tensile stress-strain curve as shown in Fig. 7.4a. If the
specimen is loaded passed the yielding strength σ ot to reach a maximum stress σ max and the
direction of straining is reversed, the stress-strain path that is followed differs from the initial
monotonic one as illustrated in Fig. 7.4b. Yielding on unloading generally occurs prior to the
stress reaching the yield strength σ oc for monotonic compression, as at point A . This early
yielding behavior is called the Bauschinger effect.
Fig. 7.4
The cyclic stress-strain response of metal is dramatically altered due to plastic strain.
It depends on the initial conditions of the specimen such as quenching, tempering, and
annealing, and its testing conditions.
Under the strain-controlled fatigue test, a metal specimen may exhibit the cyclic
stress-strain response as following:
a.) Cyclically hardening
If the stress required to enforce the strain increases on subsequence reversals, the
material undergoes cyclic hardening as shown in Fig. 7.5. In this case, the yield and ultimate
strength of the material are increased. The example of the metal that exhibits this response is
the annealed pure metal.
b.) Cyclically softening
If the stress required to enforce the strain decreases on subsequence reversals, the
material undergoes cyclic softening as shown in Fig. 7.5. In this case, the yield and ultimate
strength of the material are decreased. The example of the metal that exhibits this response is
the cold worked pure metal.
c.) Cyclically stable
Through the cyclic hardening and softening, some intermediate strength levels are
attained which represents a cyclically stable condition. The stable condition is usually reached
in about 20-40% of total fatigue life.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-5
Fig. 7.5
d.) Mixed behavior
A material exhibits cyclic softening at the early stage of fatigue life and then cyclic
hardening at the later stage of fatigue life.
7.3.1 Comparison of Metal Behavior between Monotonic and Cyclic Tests
Cyclic stress-strain curves for several engineering metals are compared with
monotonic tension curves as shown in Fig. 7.6. When the cyclic curve is above the monotonic
one, the material is one that cyclically hardens and when the cyclic curve is below the
monotonic one, the material is one that cyclically softens. A mixed behavior may also occur,
with crossing of the curves indicating softening at some strain levels and hardening at the
others.
The following criteria were proposed by Manson.
S ult
If > 1.4 , cyclically hardening.
S 0.2σ y
S ult
If < 1.2 , cyclically softening.
S 0.2σ y
S ult
If 1.2 < < 1.4 , generally stable, or may hardening or softening.
S 0.2σ y
Fig. 7.6
7.4 Cyclic Stress-Strain Curve
There are several test methods that can be used to develop a cyclic stress-strain curve.
For most metallic material, the controlled strain amplitude fatigue test will generate a
stabilized hysterisis loop. The stress-strain curve is constructed by a sequence of the stabilized
hysterisis loop.
a.) Companion sample method
The cyclic stress-strain curve is constructed by a set of test specimens at various strain
levels as shown in Fig. 7.7. This method is time-consuming and requires a large number of
tests.
Fig. 7.7
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-7
Massing’s hypothesis states that the stabilized hysterisis loop can be obtained by
doubling the cyclic stress-strain curve as shown in Fig. 7.8.
Fig. 7.8
b.) Incremental step method
A cyclic stress-strain curve is constructed by a sequence of gradually increasing or
decreasing strain amplitude in a single test as shown by the strain-time curve in Fig. 7.9.
Fig. 7.9
In general, the material will be stabilized after three to four blocks of loading. After
the incremental step test, the cyclic stress-strain curve will be nearly identical to the one
obtained by connecting the loop tips.
7.5 S − N Diagram and Stress Life Relation
The S − N curve is a plot of alternating stress, S , versus cycles to failure, N f
obtained from the fatigue test. The S − N data are usually presented on a log-log plot with the
actual S − N line representing the mean of the data as shown in Fig. 7.10.
Certain materials exhibit an endurance limit, which is stress level below that the
material has an infinite life. For engineering purpose, the infinite life is usually considered as
10 6 cycles. For most nonferrous metal such as aluminum, there exist no distinct endurance
limit and the S − N curve has a continuous slope. A pseudo-endurance limit for these
materials is taken on the stress value corresponding to 5(10) 8 cycles.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-8
Fig. 7.10
Empirical relations of endurance limit are
S e (ksi) = 0.25( BHN )
Stress-Life relation
Consider a general S − N curve plotted on a log-log coordinate as shown in Fig. 7.11.
Fig. 7.11
For the number of cycle is 10 3 ≤ N f ≤ 10 6 , the S − N curve is a straight line. The
log S = b log N f + C
= b log N f + log 10 C
= log N bf + log 10 C
= log N bf 10 C
Thus, we have
S = 10 C N bf
The exponents b and C are determined by using two end points. In general,
alternating stress level corresponding to a life of 10 3 cycles can be estimated as
S1000 = 0.9 S ult . Also, the endurance limit at 10 6 cycles can be estimated as, S e = 0.5S ult .
Thus,
log S e = b log 10 6 + C = 6b + C
C = log S1000 − 3b
1 S
b = − log 1000 = −0.085
3 Se
2
S1000
C = log = log 1.62S ult
Se
Thus, the mathematical representation of the curve is
S = 1.62 S ult N −f 0.085
Table 7-1
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-11
Example 7-1
From axially loaded fatigue testing under zero mean stress of unnotched AISI 4340
steel specimen, we obtain the stress amplitude and corresponding stress as shown in Table Ex
7-1. Plot the data on the log-log coordinate and determine the constant A and B of the
equation S = A( N f ) B .
Table Ex 7-1
σ a (MPa ) N f (Cycles)
948 222
834 992
703 6004
631 14130
579 43860
524 132150
The plotted data are shown in the Fig Ex 7-1a. They seem to fall along a straight line,
and the first and the last points represent the line well. Using this two point and denoting them
( σ 1 , N f 1 ) and ( σ 2 , N f 2 ), we have
σ 1 = A( N f 1 ) B σ 2 = A( N f 2 ) B
It should be noted that for axially loaded fatigue testing, the nominal stress, S , is
equal to the point stress, σ .
Fig Ex 7-1
Dividing the second equation into the first, and take logarithms of both sides.
B
σ 1 N f 1
=
σ 2 N f 2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-12
σ1 N f1
log = B log
σ2 Nf2
Solving for B ,
σ1
log
σ2 log σ 1 − log σ 2
B= =
N f 1 log N f 1 − log N f 2
log
Nf2
Fig. 7.12
Fig. 7.12 is a fatigue strength diagram for alloy steel, S u = 125 to 180 ksi, axial
loading. Since the tests required to generate a fatigue strength diagram is expensive, several
relationships have been proposed to generate the lines defining the infinite life design region.
The following relationships are commonly used for an infinite life.
a.) Soderberg (USA, 1930)
σa σm
+ =1
Se S y
The curves of each relationship are shown in Fig. 7.13. For finite life, the S e in the
above equations can be replaced with a fully reversed alternating stress level corresponding to
that finite life. It should be noted that
1. The Soderberg model is too conservative and seldom used.
2. The Gerber model is good for ductile material.
3. The Goodman model is good for brittle material.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-14
Fig. 7.13
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-15
Example 7-2
A 30 mm diameter shaft is subjected to the cyclic loading as shown in Fig. Ex 7-2a
where the magnitude of the load varies from Pmin = −0.60 Pmax to Pmax . The shaft is made of
stress-relieved cold-worked SAE 1040 steel having the ultimate stress of 830 MPa , yielding
stress of 660 MPa , the endurance limit of 410 MPa . Determine the magnitude of Pmax based
Fig. Ex 7-2a
Since the steel is a ductile material, we will use the Gerber relation to determine the
magnitude of Pmax .
2
σa σm
+ =1
S e S ult
For the linear elastic behavior of the material under the cyclic loading and we have the
relationship of the minimum load and the maximum load in the form of Pmin = −0.60 Pmax ,
then, we obtain the relationship of the minimum stress and the maximum stress of the form
σ min = −0.60σ max
σ max − σ min
The stress amplitude is σ a = . Then, we have
2
σ max = 1.25σ a
Since σ max = σ m + σ a , the mean stress is
σ m = 0.25σ a
Thus, from the Gerber relation, we have
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-16
2
σ a 0.25σ a
+ =1
410 830
Solving the polynomial equation, we obtain the alternating stress equal to 403.9 MPa
and the maximum stress and the maximum stress equal to
σ max == 1.25σ a = 504.9 MPa
σ min = −0.60σ max = −302.9 MPa
Since σ max < σ y = 660 MPa , the failure would be by fatigue and not by the general
yielding.
The loads need to be multiplied by 1.80 due to the factor of safety. Then, we can draw
the bending moment diagram and torque diagram as shown in Fig. Ex 7-2b.
Fig. Ex 7-2b
The maximum bending moment occurred on the shaft is
The maximum toque occurred on the shaft is 27 P . The moment of inertia and the
polar moment of inertia of the shaft are
πd 4 π (0.030) 4
I= = = 39.76(10 −9 ) m 4
64 64
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-17
πd 4 π (0.030) 4
I= = = 79.52(10 −9 ) m 4
32 32
Let P = Pmax . The flexural stress doe to the bending moment is
For the steel, we use the maximum octahedral shearing stress criteria to predict Pmax .
3τ 2 + σ 2 = σ max
2
2
Pmax = 35.266
Pmax = 5.93 N
Pmin = −3.56 N
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-18
K a = surface factor.
K b = size factor.
K c = load factor.
K d = temperature factor.
K e = other factor.
K a = aS ult
b
Surface finish a b
ksi MPa
Ground 1.34 1.58 -0.085
Machine or cold drawn 2.70 4.51 -0.265
Hot-rolled 14.4 57.7 -0.718
Forged 39.9 272 -0.995
ST
Kd =
S RT
K f = 1 + q ( K t − 1)
q = material parameter
K t = static stress concentration factor
Fig. 7.14
da
Crack propagation rate is defined as crack extension per cycle, . The growth rate
dN
of crack is a function of stress intensity factor.
da
= f (K )
dN
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-20
da
Fig. 7.15 is the typical experimental data plot of the crack propagation rate versus
dN
stress intensity range ∆K .
Fig. 7.15
It can be seen that the plot can be separated into three regions.
Region I: Crack behavior is associated with fatigue crack growth threshold value ∆K th
da
= C (∆K ) m
dN
where C is a constant and m is the slope on the log-log plot, assuming that the decades on
both log scales are the same length. The value of m is important since it indicates the degree
of sensitivity of the growth rate of the stress. For example, if m = 3 , doubling the stress range
∆S doubles the stress intensity range, thus increasing the growth rate by a factor 0f 2 m = 8 .
Region III: The crack growth rate is extremely high and little fatigue life is involved.
7.9 Factors Affecting the Fatigue Crack Growth
Stress ratio effect
σ min
R=
σ max
For a constant K , the more positive R , the higher crack growth rate as shown in Fig.
7.16.
Fig. 7.16
Two crack growth models accounting for the stress ratio effect are
1. Forman’s crack growth model
da ∆K m
=C
dN (1 − R) K c − ∆K
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 7-22
Frequency effect
At normal environmental condition, frequency has little effect on fatigue life for
metallic structure. However, the growth rate will be significantly affected if under an adverse
environment.
Temperature effect
Fatigue life will be reduced if the temperature is increased as shown in Fig. 7.17.
Fig. 7.17
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-1
Chapter 8
Beams on Elastic Foundation
8.1 Introduction
In some applications such as railroad track and the rail of movable crane, the rail does
usually acts as a beam of relatively small flexural stiffness placed on elastic foundation. The
loads are transferred through the beam to the elastic foundation. This rail can be analyzed as a
beam supported by series of discrete elastic springs. However, this analytical method is very
tedious. It is usually more practical to idealize the supports as a continuous elastic foundation.
8.2 General Theory
The response of a beam on elastic foundation can be depicted by a single differential
equation subject to different boundary conditions.
Assumptions
1. The foundation has sufficient strength to prevent failure.
2. The foundation behaves linearly elastic under loads with a small deflection.
3. The beam is fully attached to the foundation.
Consider a beam of infinite length resting on an elastic foundation with infinite length
and subjected to a point load P acting at the origin of the coordinate ( x , y , z ) as shown in
Fig. 8.1a. Under the action of the load P , the beam is deflected as shown in Fig. 8.1c, which
induces a distributed force q between the beam and the foundation.
Consider a free body diagram of an element ∆z as shown in Fig. 8.1b subjected to the
positive shear forces and moments. For small displacement analysis, we have the differential
relation of the displacement and the forces as
dy
=θ
dz
d2y
EI x = −M x 8.1
dz 2
d3y
EI x = −V y
dz 3
d4y
EI x = −q
dz 4
where the distributed reaction force q is positive when acting upward.
For linearly elastic foundation, the distributed force q is linearly proportional to the
deflection y . Thus,
q = ky 8.2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-2
k = bk o 8.3
where k is the elastic coefficient, k o is the elastic foundation modulus, and b is the width of
the foundation.
Fig. 8.1
The k o usually has a value between 20(10 6 ) N/m 2 /m to 200(10 6 ) N/m 2 /m for soil.
d4y
4
+ 4β 4 y = 0 8.5
dz
By using the method of differential equations or by direct substituting into Eq. 8.5, the
general solution of y is
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-3
Due to the symmetry of the beam, we can determine the deflection of the beam for
negative value of z by y (− z ) = y ( z ) .
Fig. 8.2
8.3 Infinite Beam Subjected to Point Load
The constant of integration C 3 and C 4 of the Eq. 8.7 can be determined by using the
α
2 ∫ kCe − βz (sin βz + cos βz )dz = P
0
α α
P
∫ e (sin βz )dz + ∫ e (cos βz )dz =
− βz − βz
0 0
2kC
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-4
1 1 P
+ =
2 β 2 β 2kC
Pβ
C= 8.9
2k
Therefore, the deflection, the slope, the moment, and the shear of the beam can be written as
Pβ − βz
y= e (sin βz + cos βz ) z≥0 8.10
2k
dy Pβ
θ= = [−2 βe − βz (sin βz )]
dz 2k
z≥0 8.11
Pβ 2 − βz
=− [e sin βz ]
k
d2y Pβ 2 d − βz
M x = − EI x = EI x [e (sin βz )]
dz 2 k dz
Pβ 2
= EI x [ βe − βz (cos βz − sin βz )]
k
z≥0
Pβ 3 − βz
= EI x [e (cos βz − sin βz )]
k
P − βz
= [e (cos βz − sin βz )]
4β
8.12
dM d P
Vy = − = − [e − βz (cos βz − sin βz )]
dz dz 4 β
P
=− [ βe − βz (2 cos βz )] z≥0 8.13
4β
P
= − [e − βz cos βz ]
2
Defining
Aβz = e − βz (sin βz + cos βz ) B βz = e − βz sin βz
Then, we have
Pβ
y= Aβz z≥0 8.10
2k
Pβ 2
θ =− Bβz z≥0 8.11
k
P
Mx = C βz z≥0 8.12
4β
P
Vy = − Dβz z≥0 8.13
2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-5
Example 8-1
A rail road uses steel rails ( E = 200 GPa ) with a depth of 184 mm . The distance from
the top of the rail to its centroid is 99.1 mm , and the moment of inertia of the rail is
36.(10 6 ) mm 4 . The rail is supported by ties, ballast, and a road bed that together are assumed
Fig. Ex 8-1a
b.) If a locomotive has 3 wheels per truck equally spaced at 1.70 m , determine the
maximum deflection, maximum bending moment, and maximum flexural stress in
the rail when the load on each wheel is 170 kN .
Since the equations of deflection and bending moment require the value of β ,
k 14
β =4 =4 3 6
= 0.000830 mm -1
4 EI x 4(200)10 (36.9)10
a.) The maximum deflection and the maximum bending moment occur under the load
where
Aβz = C βz = 1.0
Thus,
Pβ 170(10 3 )0.000830
y max = Aβz = (1) = 5.039 mm
2k 2(14)
P 170(10 3 )
M max = C βz = (1) = 51.21 kN - m
4β 4(0.000830)
Fig. Ex 8-1b
2. Under the center wheel as shown in Fig. Ex 8-1c
Fig. Ex 8-1c
For case 1, let the origin of the coordinate be located under one of the end wheel. The
distance to the first wheel z1 = 0 , we have
Aβz1 = C βz1 = 1.0
The distance from the origin to the next wheel is z 2 = 1700 mm , we have
Aβz 2 = 0.2797 C βz 2 = −0.2018
The distance from the origin to the next wheel is z 3 = 3400 mm , we have
Therefore, for this case, we get the maximum deflection and the maximum bending moment
equal to
Pβ
y max = ( Aβz1 + Aβz 2 + Aβz 3 ) = 5.039(1 + 0.2797 − 0.0377) = 6.258 mm
2k
P
M max = (C βz1 + C βz 2 + C βz 3 ) = 51.20(10 6 )(1 − 0.2018 − 0.0752) = 37.02 kN - m
4β
For case 2, let the origin of the coordinate be located under the center wheel. The
distance to the first wheel z1 = 0 , we have
Aβz1 = C βz1 = 1.0
The distance from the origin to either of the end wheel is z 2 = 1700 mm , we have
Aβz 2 = 0.2797 C βz 2 = −0.2018
Therefore, for this case, we get the maximum deflection and the maximum bending moment
equal to
Pβ
y max = ( Aβz1 + 2 Aβz 2 ) = 5.039(1 + 2(0.2797)) = 7.858 mm
2k
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-7
P
M max = (C βz1 + 2C βz 2 ) = 51.20(10 6 )(1 − 2(0.2018)) = 30.54 kN - m
4β
From the calculation, we obtain
y max = 7.858 mm
M max = 37.02 kN - m
Fig. 8.3
Let each spring in Fig. 8.3a has the same spring constant K . The reaction force R
that each spring exert on the beam is directly proportional to the deflection y of the beam at
the section where the spring is attached. Thus,
R = Ky
If l is the spring spacing, the load R can be idealized as uniformly distributed over a
total span l ( l / 2 on either side of the spring) as shown in Fig. 8.3b. If the stepped distributed
loading is approximated by the dashed curve, the approximate distributed load is similar to
the distributed load q of Fig. 8.1a. If the two forces are to be the same, then,
Ky = kyl
K
k=
l
where k is the elastic coefficient for this case. Then, we can use Eq. 8-10 to Eq. 8-13 to find
the deflection, the slope, the moment, and the shear of the beam. However, it has been found
that the solutions are only practically useful when
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-9
π
l≤
4β
The approximate solution for a beam of infinite length, with equally spaced elastic
supports, may be used to obtain a reasonable approximate solution for a sufficiently long
finite length beam as shown in Fig. 8.4a. In general, the end springs do not coincide with the
ends of the beam, but lie at some distance less than l / 2 from the end of the beam as shown in
Fig. 8.4b. Thus, we extend the beam of length L to a beam of length L ′′ , where
L ′′ = ml
and integer m is the number of spring supports. To obtain a reasonable approximate solution,
3π
L ′′ ≥
2β
Fig. 8.4
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-10
Example 8-2
An aluminum alloy I-beam (depth = 100 mm , I x = 2.45(10 6 ) mm 4 , E = 72 GPa ) as
Fig. Ex 8-2
The elastic coefficient,
110
k= = 0.100 N/mm 2
1100
and the value of β ,
0.100
β =4 = 0.000614 mm -1
4(72)10 3 (2.45)10 6
Thus,
Pβ 12(10 3 )0.000614
y max = Aβz = (1) = 36.84 mm
2k 2(0.10)
P 12(10 3 )
M max = C βz = (1) = 4.886(10 6 ) N - m
4β 4(0.000614)
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-11
M max c
σ max = = 99.7 MPa
Ix
Due to the symmetry of the beam, the magnitude of βz , the corresponding Aβz , and
the deflection for the first (point C ), second (point B ), and third (point A ) springs to the
right and left of the load are
Pβ
βl = 0.6754 Aβl = 0.7153 y C = Aβl = 36.84(0.7153) = 26.53 mm
2k
Pβ
2 βl = 1.3508 A2 βl = 0.3094 y B = A2 βl = 36.84(0.3094) = 11.40 mm
2k
Pβ
3βl = 2.0262 A3 βl = 0.0605 y A = A3 βl = 36.84(0.0605) = 2.23 mm
2k
The reaction for each spring can be obtained by using the equation R = Ky and the
results are shown in the Table Ex 8-2.
Table Ex 8-2
Approximate solution Exact solution
Reaction A 245 N − 454 N
Reaction B 1254 N 1216 N
Reaction C 2899 N 3094 N
Reaction D 4052 N 4288 N
y max 36.84 mm 38.98 mm
Comparing the results with the exact results by using the energy method, we can see
that only the reaction at A are considerably in error.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-12
Fig. 8.5
From the displacement solution of the beam subjected to concentrated load, Eq. 8.10,
Pβ − βz
y= e (sin βz + cos βz ) z≥0 8.10
2k
Then,
βw − βz
dy H = e (sin βz + cos βz )dz
2k
By using the principle of superposition, the total deflection due to the distributed load is
a
βw − βz
yH = ∫ e (sin βz + cos βz )dz
0
2k
b
βw − βz
+∫ e (sin βz + cos βz )dz
0
2k
wβ 1
yH =
2k
(
− aβ 1 −bβ
)
β 1 − e cos βa + β 1 − e cos βb ( )
8.11
=
w
2k
[
2 − e − βa cos βa − e − βb cos βb ]
Then, by using the differential relations, the deflection, the slope, the shear force, and
the bending moment of the beam can be determined and simplified as
yH =
w
2k
[2 − Dβa − Dβb ]
wβ
a b
θH = ∫
dy H
dz
dy
dz + ∫ H dz =
dz 2k
Aβa − Aβb [ ] 8.12
0 0
MH =
w
4β 2
[B βa + Bβb ]
VH =
w
4β
[
C βa − C βb ]
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-13
where
Aβa = e − βa (sin βa + cos βa ) B βa = e − βa sin βa
Fig. 8.6
Fig. 8.6 shows the plot of the expressions of the deflection, the slope, the shear force,
and the bending moment of the beam with respect to the z axis. We can see that the
maximum deflection occurs at the center of the segment L ′ . However, the location of the
maximum bending moment may or may not occur at the center of the segment L ′ , depending
on the magnitude of βL ′ .
If βL ′ ≤ π , then, the location of the maximum bending moment is at the center of the
segment L ′ .
w
If βL ′ → α , then, θ → 0 , M x → 0 , V y → 0 , and y → . Therefore, the location of
k
the maximum bending moment of the beam is at either βa = π / 4 or βb = π/ 4 .
If βL ′ > π , then, the location of the maximum bending moment may lie outside the
segment L ′ . However, the maximum value outside the segment L ′ is larger than the
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-14
maximum value within the segment L ′ only 3%. Thus, we may assume that the location of
the maximum bending moment in this case is at π / 4 β from either ends of the uniformly
distributed load within the segment L ′ .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-15
Example 8-3
A long wood beam ( E = 10.0 GPa ) has a rectangular cross section with a depth of
200 mm and a width of 100 mm . It rests on an earth foundation having spring constant of
k o = 0.040 N/mm 3 and is subjected to a uniformly distributed load w = 35.0 N/mm
extending over a length L ′ = 3.61 m . Taking the origin of the coordinate at the center of the
segment L ′ , determine the maximum deflection, the maximum bending stress in the beam,
and the maximum pressure between the beam and the foundation.
The moment of inertia of the beam about x -axis is
I x = 66.67(10 6 ) mm 4
4
β =4 3 6
= 0.001107 mm -1
4(10)10 (66.67)10
From the graph of the deflection of the beam as shown in Fig. 8.6b, the maximum
deflection occurs at the center of segment L ′ . Since a = b = L ′ / 2 ,
L′
βa = βb = β = 2.0
2
Dβa = D βb = −0.0563
y max =
w
2k
[ 35
]
2 − Dβa − Dβb = (2 − 2(−0.0563)) = 9.243 mm
4
The maximum pressure between the beam and the foundation occurs at the point of
the maximum deflection.
q max = k o y max = 0.040(9.243) = 0.370 MPa
They are 4 possible locations at which the maximum bending moment may occur.
However, since the beam is symmetry with respect to the center of the segment L ′ , the
maximum bending moment may be occurred at the center of the segment L ′ or where
VH = 0 .
Since βL ′ = 0.001107(3.61)10 3 = 4.00 > π , the maximum bending moment does not
occur at the center of the segment L ′ . Hence, L ′ , the maximum bending moment will occur
at the location where V H = 0 .
VH =
w
4β
[ ]
C βa − C βb = 0
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-16
C βa = C βb
M max =
−w
4β 2
[
Bβa − Bβb ]
35
= [0.3223 − (−0.0086)]
4(0.001107) 2
= 2.363 kN - m
which is larder than the bending moments occurred inside segment L ′ by about 3%.
The corresponding bending stress is
M max c
σ max = = 3.544 MPa
Ix
It should be noted that if the maximum bending moment is assumed to occur at
π / 4 β , βa = π / 4 and βb = 4 − π / 4 (inside segment L ′ ), we obtain the bending moments
equal to
MH =
w
4β 2
[B βa + Bβb ]
35
= [0.3224 + (−0.0029)]
4(0.001107) 2
= 2.362 kN - m
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-17
In this case the constants of integration C 3 and C 4 can be determined by using the
boundary conditions:
d2y
EI x = −M o
dz 2 z =0
d3y
EI x = −V y = P 8.13
dz 3 z =0
Fig. 8.7
d2y 2β 2
Since = − βz [C 3 cos βz − C 4 sin βz ] ,
dz 2 e
Mo 2β 2 M o
C3 = = 8.14
2 β 2 EI x k
d 3 y 2β 3
Since = βz [C 3 sin βz + C 4 cos βz + C 3 cos βz − C 4 sin βz ] ,
dz 3 e
P 2 βP
C3 + C 4 = =
2 β EI x
3
k
2 βP 2 β 2 M o
C4 = −
k k
Thus, the deflection of the beam in this case is
2 βe − βz
y= [P cos βz − βM o (cos βz − sin βz )] 8.15
k
Rearranging and simplifying the equation, we have
2 Pβ 2β 2 M o
y= Dβz − C βz
k k
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-18
The expressions of the slope, shear force, and bending moment can be found by using
the differential relation.
2 Pβ 2 4β 3 M o
θ =− Aβz + Dβz
k k
P
Mx = − Bβz + M o Aβz
β
V y = − PC βz − 2 M o βBβz
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 8-19
Example 8-4
A steel I-beam ( E = 200 GPa ) has a depth of 102 mm , a width of 68 mm , a moment
of inertia of I x = 2.53(10 6 ) mm 4 , and a length of 4 m . It is attached to a rubber foundation
for which k o = 0.350 N/mm 3 . A concentrated load P = 30.0 kN is applied at one end of the
beam. Determine the maximum deflection, the maximum bending stress in the beam, and
their locations.
The spring coefficient,
k = 68(0.350) = 23.8 N/mm 2
and the value of β ,
23.8
β =4 = 0.001852 mm -1
4(200)10 3 (2.53)10 6
Since
3π
L = 4000 mm > = 2540 mm
2β
the beam can be considered as a long beam.
The maximum deflection occurs at the end where load P is applied ( z = 0 ), since
Dβz is maximum. We have βz = 0 and Dβz = 1.0 .
2 Pβ 2(30)10 3 (0.001852)
y max = Dβz = (1) = 4.67 mm
k 23.8
The maximum bending occurs at z = π / 4 β , where B βz is maximum. This is the same
Chapter 9
Flat Plates
9.1 Introduction
Flat plate is a structural element whose middle surface lies in a flat plane and
subjected to lateral load q . Floor slabs and pavements are the common examples. The plate
can be categorized according to its thickness relative to its other dimensions and according to
its lateral deflection compared to its thickness.
1. Relatively thick plates with small deflections
2. Relatively thin plates with small deflections
3. Very thin plate with large deflection
4. Extremely thin plates or membrane
Fig. 9.1
Under the action of the lateral load q , the midsurface of the flat plate is deflected as
shown in Fig. 9.1a. Fig. 9.1b shows the state of stresses and their distribution in a small
element of the plate. The governing equations of the flat plate can be determined by using the
equilibrium equations, the strain-displacement relations, and the stress-strain relations.
9.2 Assumptions and Limitations of Thin Plate with Small Deflection
In the classical thin-plate theory or Kirchhoff theory, the following assumptions are
applied:
1. The plate is flat and has a constant thickness.
2. The plate has a relatively small thickness compared to the smallest lateral
dimensions.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-2
7. A line normal to the midsurface before loading remains normal to the midsurface
after loading. The transverse shear strain γ yz and γ xz are zero.
Following the assumptions, the lateral deflection of the plate is a function of only
coordinate x and y , and the stresses σ x , σ y , and τ xy are linearly distributed as shown in
Fig. 9.1b.
9.3 Force-Stress Relations
Fig. 9.2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-3
Consider a plate differential element as shown in Fig. 9.1 and Fig. 9.2. The increment
of the bending moments dM x and dM y , the twisting moment dM xy , and the transverse shear
dM x = z (σ x dA)
dM y = z (σ y dA)
dM xy = z (τ xy dA)
dQ x = τ xz dA
dQ y = τ yz dA 9.1
If we define the bending moments M x and M y , the twisting moment M xy , and the
transverse shear forces Q x and Q y per unit length. Then, the differential area dA is
dA = (1)dz , and the moments and the shear forces can be found by integrating the
corresponding moments and shear forces.
t/2
Mx = ∫σ
−t / 2
x zdz
t/2
My = ∫σ
−t / 2
y zdz
t/2
M xy = ∫τ
−t / 2
xy zdz
t/2
Qx = ∫τ
−t / 2
xz dz
t/2
Qy = ∫τ
−t / 2
yz dz 9.2
∂Q x ∂Q y
− Q x dy − Q y dx + qdxdy + Q x + dx dy + Q y + dy dx = 0
∂x ∂y
∂Q x ∂Q y
∂x dx dy + dy dx + qdxdy = 0
∂y
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-4
∂Q x ∂Q y
+ +q = 0 9.3
∂x ∂y
∑M x = 0;
∂M xy ∂M y ∂Q y
− M xy dy − M y dx + M xy + dx dy + M y + dy dx − Q y + dy dxdy
∂x ∂y ∂y
∂Q x dy dy dy
− Q x + dx dy + Q x dy − qdxdy =0
∂x 2 2 2
Neglecting the higher-order term, we have
∂M xy ∂M y
dx dy + dy dx − Q y dxdy = 0
∂x ∂y
∂M xy ∂M y
+ − Qy = 0 9.4
∂x ∂y
∂M x ∂M xy
+ − Qx = 0 9.5
∂x ∂y
Substituting Eq. 9.4 and 9.5 into Eq. 9.3, we have the equation relating the external
lateral load q with the internal resultant bending moments in the form of
∂2M x ∂ 2 M xy ∂ 2 M y
+2 + +q=0 9.6
∂x 2 ∂x∂y ∂y 2
9.5 Kinetics: Strain-Displacement Relations
Fig. 9.3
Consider Fig. 9.3 showing the differential slice of the plate viewed parallel to the y
axis. The displacement in the x axis of the point P is
∂w
u = −z
∂x
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-5
If we view the slice of the plate parallel to the x axis, we have the displacement in the
y axis of the point P is
∂w
v = −z
∂y
Hence, the strains components on the element due to the displacement are
∂u ∂2w
εx = = −z 2
∂x ∂x
∂v ∂2w
εy = = −z 2
∂y ∂y
∂u ∂v ∂2w
γ xy = + = −2 z 9.7
∂y ∂x ∂x∂y
9.6 Stress-Strain Relations
For linear elastic isotropic homogeneous material,
E
σx = (ε x + νε y )
1 −ν 2
E
σy = (ε y + νε x )
1 −ν 2
τ xy = Gγ xy 9.8
E
where G = .
2(1 + ν )
9.7 Stress-Deflection Relations
Substituting Eq. 9.7 into Eq. 9.8, we have
E ∂2w ∂2w
σx = − z 2 − νz 2
1 −ν 2 ∂x ∂y
Ez ∂2w ∂2w
σx = − 2 + ν 9.9a
1 −ν 2 ∂x ∂y 2
Ez ∂2w ∂2w
σy =− 2 + ν 9.9b
1 −ν 2 ∂y ∂x 2
∂2w
τ xy = −2Gz 9.9c
∂x∂y
9.8 Governing Differential Equations
Substituting Eq. 9.9 into the internal force-stress relations, Eq. 9.2, we have the
expressions for the internal forces and the displacement w .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-6
Ez 2 ∂ 2 w ∂2w
t/2 t/2
Mx = ∫ σ x zdz =
−t / 2
∫
−t / 2
−
1 − ν 2 ∂x 2
+ ν dz
∂y 2
E ∂2w ∂2w
t/2
2 ∫
Mx = − 2 + ν z 2 dz
1 −ν 2 ∂x ∂y −t / 2
t/2
t3
Since ∫ z dz = , 2
−t / 2
12
Et 3 ∂2w ∂2w
Mx = − + ν
12(1 − ν 2 ) ∂x 2 ∂y 2
Et 3
Defining D = which is the plate flexural rigidity.
12(1 − ν 2 )
∂2w ∂2w
M x = −D 2 + ν 2 9.10a
∂x ∂y
Similarly,
∂2w ∂2w
M y = −D 2 + ν 2 9.10b
∂y ∂x
t/2
E ∂2w Et 3 ∂ 2 w
2(1 + ν ) ∂x∂y −t∫/ 2
M xy = −2 z 2
dz = −
12(1 + ν ) ∂x∂y
∂2w
M xy = −(1 − ν ) D 9.10c
∂x∂y
Substituting Eq. 9.10 into the equilibrium equation, Eq 9.6, we have the governing
equation for the thin flat plate.
∂2M x ∂ 2 M xy ∂ 2 M y
+2 + +q=0
∂x 2 ∂x∂y ∂y 2
∂4w ∂4w ∂ 4 w q ( x, y )
+2 2 2 + 4 = 9.11
∂x 4 ∂x ∂y ∂y D
The deflection of the midsurface of the flat plate w = w( x, y ) can be determined by
integrating this governing equation. Then, the moment expressions are obtained by
substituting the deflection expressions into the expressions of the moment, Eq. 9.10.
∂2w ∂2w
M x = −D 2 + ν 2
∂x ∂y
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-7
∂2w ∂2w
M y = −D 2 + ν 2
∂y ∂x
∂2w
M xy = −(1 − ν ) D
∂x∂y
The transverse shearing force can be obtained from the Eq. 9.4 and 9.5,
∂M x ∂M xy
Qx = +
∂x ∂y
∂M xy ∂M y
Qy = +
∂x ∂y
Thus,
∂ ∂2w ∂2w
Qx = − D 2 + 2 9.12a
∂x ∂x ∂y
Similarly,
∂ ∂2w ∂2w
Qy = −D + 9.12b
∂y ∂x 2 ∂y 2
a
Simply-supported
x
b Partially
Free
restrained
Fixed
y
z
Fig. 9.4
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-8
Fixed Edges
For the fixed edge, the deflection and slope are zero. Thus,
w y =b = 0
∂w
=0 9.13
∂y y =b
∂2w ∂2w
My = −D 2 + ν 2 = 0
∂y ∂x y =0
y =0
∂2w ∂ ∂w ∂w
The expression ν can be rewritten as ν . The term is the rate of
∂x 2
∂x ∂x ∂x
change of the slope at the boundary. But, the change in slope along the simply support edge
∂2w
y = 0 is always zero. Hence, the quantity ν vanishes and the moment boundary
∂x 2
condition is
∂2w
My = =0 9.14b
y =0 ∂y 2 y =0
Free Edge
At the free edge, the moment and shear are zero. Thus,
Mx x=a
= M xy = Qx x =a
=0
x=a
∂2w ∂2w
2 + ν =0 9.15a
∂x ∂y 2 x =a
The last two boundary conditions can be combined into a single equation. Consider
the Fig. 9.5, Kirchhoff has shown that the moment M xy can be though of as a series of
couples acting on an infinitesimal section. Hence, at any point along the edge
∂M xy
Q′ =
∂y x=a
This equivalent shearing force, Q ′ , must be added to the shearing force Q x acting at
∂M xy
V x = Q x + =0 9.15b
∂y x =a
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-9
obtain
∂3w ∂3w
3 + (2 − ν ) =0 9.15c
∂x ∂x∂y 2 x =a
Fig. 9.5
Partially Restrained Edge
A partially restrained edge occurs in when the plate is connected to the beam as shown
in Fig. 9.6. In this case, the following boundary conditions must be satisfied.
Fig. 9.6
V plate
=V beam
Corner Reactions
It was shown in the derivation of the boundary condition of the shearing force at the
free edge that the torsion moment M xy as shown in Fig. 9.5 can be resolved into a series of
∂2w
R = −2 M xy x =a = 2(1 − ν ) D 9.17
y =b
∂x∂y xy==ab
This equation is usually used to determine the force in corner bolts of rectangular
plates.
Fig. 9.7
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-11
Example 9-1
Determine the moment and reaction for a simply supported rectangular plate of length
a in the x direction and width b in the y direction as shown in Fig. Ex 9-1. The plate is
πx πy
subjected to a sinusoidal lateral load, q = q o sin sin .
a b
Fig. Ex 9-1
The governing differential equation of the plate, Eq. 9.11, is
∂4w ∂4w ∂ 4 w qo πx πy
+ 2 + = sin sin
∂x 4
∂x ∂y
2 2
∂y 4
D a b
For the simply supported plate, the boundary conditions, Eq. 9.14, are
∂2w
w = 0 and = 0 at x = 0 and x = a
∂x 2
∂2w
w = 0 and = 0 at y = 0 and y = b
∂y 2
In order to solve the governing differential equation for the deflection, the assumed
deflection equation must be in the same form as that of the governing equation and must
satisfy the boundary conditions. Thus,
πx πy
w = C sin sin
a b
Substituting the deflection equation into the governing equation, we obtain
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-12
qo
C= 2
1 1
Dπ 4 2 + 2
a b
Thus, the assumed deflection equation is
qo πx πy
w= 2
sin sin
1 1 a b
Dπ 4 2 + 2
a b
Substituting this equation into the force-stress relations, Eq. 9.10, we get
∂2w ∂2w qo 1 ν πx πy
M x = −D 2 + ν 2 = 2 + 2 sin sin
∂x ∂y 2
1 1 a b a b
π 4 2 + 2
a b
∂2w ∂2w qo ν 1 πx πy
M y = −D 2 + ν 2 = 2 + 2 sin sin
∂y ∂x 2
1 1 a b a b
π 4 2 + 2
a b
∂2w q o (1 − ν ) πx πy
M xy = −(1 − ν ) D =− cos cos
∂x∂y 1 1
2
a b
π 4 2 + 2 ab
a b
It should be noted that the maximum M x and M y occur at x = a / 2 and y = b / 2 .
Substituting the deflection equation into the transverse shearing forces, Eq. 9.12, we have
∂ ∂2w ∂2w qo πx πy
Qx = − D + = cos sin
∂x ∂x 2 ∂y 2 1 1
2
a b
πa 2 + 2
a b
∂ ∂2w ∂2w qo πx πy
Qy = −D + = sin cos
∂y ∂x 2 ∂y 2 1 1
2
a b
πb 2 + 2
a b
The reaction on edge x = a can be determined by using the Eq. 9.15b,
∂M xy qo 1 2 − ν πy
V x = Q x + =− 2 + 2 sin
∂y x =a 1 1 a b 2
b
πa 2 + 2
a b
The reaction on edge y = b can be determined by using the equation,
∂M xy qo 1 2 −ν πx
V y = Q y + =− 2 + 2 sin
∂x y =b 1 1 b a 2
a
πb 2 + 2
a b
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-13
The total reaction around the plate can be determined by integrating the reaction
equations from x = 0 to x = a and from y = 0 to y = b and then multiplying by 2 due to
symmetry.
4q o ab 8q o (1 − ν )
total reaction = +
π 2
1 1
2
π 2 + 2 ab
2
a b
The first part of this equation can also be obtained by integrating the applied load over
the total area.
b a
πx πy
∫∫q
0 0
o sin
a
sin dxdy
b
The second part is the summation of the four corner reactions that can be determined
by using the Eq. 9.17. Thus, for example,
∂2w 2q o (1 − ν )
R = −2 M xy x =0 = 2(1 − ν ) D =
y =0
∂x∂y y =0
x = 0 1 1
2
π 2 2 + 2 ab
a b
The positive value of R at the corner x = y = 0 means that the reaction force has the
downward direction. Thus, it indicates that the corners tend to lift up. This action must be
considered when designing the concrete slab. The top corner reinforcements as shown in Fig.
Ex 9-1c are needed to resist these forces.
Fig. Ex 9-1c
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-14
n ′πy
series by sin dy and integrate from y = 0 to y = b . Then we can see that
b
nπy n ′πy
b
∫ sin
0
b
sin
b
dy = 0 when n ≠ n′
nπy n ′πy
b
b
∫ sin
0
b
sin
b
dy =
2
when n = n ′
n ′πy mπx
b
b α
∫ f ( x, y ) sin dy = ∑ q mn′ sin
0
b 2 m =1 a
m ′πx
Multiplying both sides of this equation by sin dx and integrating from x = 0 to
a
x = a , we obtain
m′πx n ′πy
a b
ab
∫ ∫ f ( x, y) sin
0 0
a
sin
b
dxdy =
4
q m′n′
mπx nπy
a b
4
q mn = ∫ ∫
ab 0 0
f ( x, y ) sin
a
sin
b
dxdy 9.19
Example 9-2
Determine the maximum deflection and bending moments a simply supported plate as
shown in Fig. Ex 9-2a due to a uniformly distributed load, q o .
Fig. Ex 9-2a
Since the load is uniformly distributed over the entire plate,
f ( x, y ) = q o
m′πx n ′πy
a b
4q
q mn = o
ab ∫ ∫ sin
0 0
a
sin
b
dxdy
4q o 16q
= (cos mπ − 1)(cos nπ − 1) = 2 o
π mn
2
π mn
where m = 1,3,5, K and n = 1,3,5, K .
Thus, the uniformly distributed load can be represented by the double trigonometric
series as
16q o α α
1 mπx nπy
q ( x, y ) =
π 2 ∑ ∑
m =1, 3, 5,K n =1, 3, 5,K mn
sin
a
sin
b
Substituting this equation and the plate deflection w into the governing differential
∂4w ∂4w ∂4w q
equation of the plate, + 2 2 2 + 4 = , we obtain the coefficient wmn as
∂x 4 ∂x ∂y ∂y D
16qo
wmn = 2
m 2 n 2
π mnD +
6
a b
where m = 1,3,5, K and n = 1,3,5, K . Thus, the deflection of the plate is
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-16
mπx nπy
α α sin sin
16q a b
w= 6 o ∑ ∑
π D m =1,3,5,K n =1,3,5,K m 2 n 2
2
mn +
a b
In this case, the deflection is symmetric with respect to the axes x = a / 2 and
y = b / 2 . The maximum deflection is occurred at the center of the plate in which x = a / 2
and y = b / 2 . Then,
m+n
−1
16q α α
(−1) 2
wmax = 6 o ∑ ∑
π D m =1,3,5,K n =1,3,5,K m 2 n 2
2
mn +
a b
It should be noted that this series converge rapidly to the exact solution. Substituting
the deflection equation into the force-stress relations, Eq. 9.10, we get the bending moments
as
∂2w ∂ 2 w 16q o α α
mπx nπy
M x = −D 2 + ν 2 = 4
∂y π
∑ ∑F mn sin sin
∂x m =1, 3, 5,K n =1, 3, 5,K a b
∂2w ∂ 2 w 16q α α
mπx nπy
M y = −D 2 +ν 2 = 4 o
∂x π
∑ ∑G mn sin sin
∂y m =1, 3, 5,K n =1, 3, 5,K a b
∂2w 16q o (1 − ν ) α α
mπx nπy
M xy = −(1 − ν ) D
∂x∂y
=−
π 4 ∑ ∑ H mn cos
m =1, 3, 5,K n =1, 3, 5,K a
cos
b
2 2
m n
+ν
where Fmn = a b
2
m 2 n 2
mn +
a b
2 2
m n
ν +
Gmn = a b
2
m 2 n 2
mn +
a b
1
H mn = 2
m 2 n 2
ab +
a b
The maximum bending moments are occurred at the center of the plate in which
x = a / 2 and y = b / 2 . Fig. Ex 9-2b shows a plot of the equations of the bending moments by
assuming that ν = 0.3 . The figure also shows a plot of the bending moments M 1 and M 2 that
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-17
are obtained from Mohr’s circle along the diagonal of the plate. It should be noted that M 1
becomes negative near the corner of the plate. This is due to the uplift tendency at the corners.
This uplift is resisted by the reaction R that causes tension at the top portion of the plate near
the corners as mentioned before.
Fig. Ex 9-2b
Finally, substituting x = a / 2 and y = b / 2 into the equation of the bending moment,
we obtain toe maximum bending moment.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-18
Example 9-3
A simply supported square plate of uniform thickness has the width on each side of
500 mm as shown in Fig. Ex 9-3. The plate must carry the uniform lateral pressure
q o = 1.0 MPa without deflecting more than one-fourth its thickness and without exceeding
the allowable normal stress of 250 MPa . Let the material of the plate has E = 200 GPa and
ν = 0.30 and using the maximum octahedral shearing stress criterion, determine the
minimum allowable thickness.
Fig. Ex 9-3
Check the deflection
Due to the symmetry of the plate and loading, the maximum bending moment is
occurred at the center of the plate ( x, y ) = (0.25,0.25) m .
m+n
−1
16q α α
(−1) 2
wmax = 6 o ∑ ∑
π D m =1,3,5,K n =1,3,5,K m 2 n 2
2
mn +
a b
Since this series converge rapidly to the exact solution, we will use the first three
nonzero series terms for which m + n is smallest which are
m = 1 and n = 1
m = 3 and n = 1
m = 1 and n = 3
1+1
−1
3+1
−1
1+ 3
−1
16q (−1) 2
(−1) 2
(−1) 2
wmax = 6 o + + 2
π D 1 2 1 2 2 3 2 1 2
2
1 2 3 2
1(1) + 3(1) + 1(3) +
0 . 5 0 .5
0 . 5 0 . 5
0 . 5 0.5
qo
wmax = 256.573(10 −6 )
D
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-19
From example 9-2, the maximum bending moment at the center of the plate
( x, y ) = (a / 2, b / 2) and the width a = b is
2 2
m n
+ν
16q o α α
a b mπx nπy
Mx =
π 4 ∑ ∑ 2
sin
a
sin
b
m =1, 3, 5,K n =1, 3, 5,K m 2 n 2
mn +
a b
2 2
m n
α α +ν
16qo a a
M max =
π4
∑ ∑ 2
m =1, 3, 5,K n =1, 3, 5,K m 2 n 2
mn +
a a
Since this series converge rapidly to the exact solution, we will use the first three
nonzero series terms for which m + n is smallest which are
m = 1 and n = 1
m = 3 and n = 1
m = 1 and n = 3
Thus, we get
2 2 2 2 2 2
1 + 0.3 1 3 1 1 3
+ 0.3 + 0.3
16q o 0.5 0.5 0.5 0.5 0.5 0.5
M max = 4 + + 2
π 1 2 1 2 2 3 2 1 2
2
1 2 3 2
1(1) + 3(1) + 1(3) +
0.5 0.5 0.5 0.5 0.5 0.5
M max = 11566.3 N - m
The flexural stresses due to the bending moment for unit width of the plate are
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-20
Mc 11566.3(t / 2) 69398
σx =σy =σ = = =
I 1(t 3 ) / 12 t2
Since M xy = 0 , τ xy at the center of the plate is zero and the obtained flexural stresses
are the principal normal stresses. Using the maximum octahedral shearing stress criterion,
σ 12 − σ 1σ 2 + σ 22 = 250 2
2σ 2 − σ 2 = 250 2
69398
2
= 250(10 6 )
t
t = 16.7 mm
Since the thickness based on the maximum octahedral shearing stress criterion is
larger than the maximum moment condition, determine the allowable thickness is at least
t = 16.7 mm
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-21
Example 9-4
Determine the deflection for the simply supported plate subjected to the uniformly
distributed load over the area of rectangular as shown in Fig. Ex 9-4.
By virtual of Eq. 9.18 and Eq. 9.19, the coefficient q mn of the uniformly distributed
Fig. Ex 9-4
Substituting q mn and the plate deflection w into the governing differential equation of
mn +
a b
This equation will be reduced to the same equation in the previous example by setting
c = a , d = b , e = a / 2 , and f = b / 2 . Finally, the deflection of the plate can be determined
from the equation
α α
mπx nπy
w= ∑ ∑
m =1, 3,K n =1, 3,K
wmn sin
a
sin
b
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-22
Each of these parts consists of a single trigonometric series where the unknown function is
determined from the boundary conditions. The homogeneous part is written as
α
mπx
wh = ∑ f m ( y ) sin 9.22
m =1 a
where f m ( y ) is a function of y only. This equation satisfies the simply supported boundary
condition at x = 0 and x = a .
∂ w ∂ w
Substituting wh into the governing differential equation of the plate, +2 2 2
∂x 4
∂x ∂y
∂4w q
+ 4 = , we obtain
∂y D
mπ 4 mπ d f m ( y ) d f m ( y )
2 2 4
mπx
f m ( y ) − 2 2
+ 4 sin =0
a a dy dy a
This equation only satisfies when the bracketed term is equal to zero. Thus,
2 4
d 4 f m ( y) mπ d f m ( y ) mπ
2
− 2 + f m ( y) = 0 9.23
dy 4 a dy 2 a
The solution of this differential equation can be expressed as
f m ( y ) = Fm e Rm y 9.24
where C1m , C 2 m , C 3m , and C 4 m are constants. This equation can also be written as
mπx
a
2
qm ( y) =
a0∫ q( x, y ) sin
a
dx
Substituting the equations of the particular solution w p , Eq. 9.26, and the distributed
Example 9-5
The rectangular plate as shown in Fig. Ex 9-5 is subjected to uniformly distributed
load q o . Determine the deflection of the plate.
Fig. Ex 9-5
Since q( x, y ) = q o which is a constant, the coefficient q m ( y ) of the uniformly
2q o
= (− cos mπ + 1) where m = 1, 3, 5, K .
mπ
4q
= o
mπ
Substituting the distributed load q into Eq. 9.28, we have
2 4
d 4 k m ( y) mπ d k m ( y ) mπ
2
4q o
− 2 + k m ( y) =
dy 4
a dy 2
a mπD
4a 4 q o
km = 5 5 where m = 1, 3, 5,K .
mπ D
The particular solution for the deflection of the plate is
4a 4 q o α
1 mπx
wp =
π 5D
∑
m =1, 3,K m
5
sin
a
The homogeneous solution for the deflection of the plate is obtained from Eq. 9.25.
α
mπy mπy mπy mπy mπx
wh = ∑ Am sinh + Bm cosh + C m y sinh + Dm y cosh sin
m =1 a a a a a
By observing the deflection of the plate due to the uniform load, we can see that the
deflection in the y direction is symmetric about the x axis. Thus, the constants Am and Dm
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 9-25
mπy mπy
must be set to zero since the quantities sinh and y cosh are odd functions as y
a a
varies from positive to negative.
mπx
In addition, m must be set to 1, 3, 5, K in order for the term sin to be symmetric
a
about x = a / 2 . Hence,
α
mπy mπy mπx
wh = ∑
m =1, 3,K
Bm cosh a + C m y sinh a sin a
∂2w
M y = 0 or = 0 at y = ±b / 2
∂y 2
Then, we have
mπb b mπb 4a 4 q o
Bm cosh + C m sinh + 5 5 =0
2a 2 2a mπ D
mπ mπb mπb mπb
Bm a + bC m cosh 2a + C m 2a sinh 2a = 0
Solving these two simultaneous equations, we obtain
2a 3 q o
Cm =
mπb
m 4π 4 D cosh
2a
mπb
4a 4 q o + mπq o a 3b tanh
Bm = 2a
mπb
m 5π 5 D cosh
2a
4 mπb
4a q o + mπq o a b tanh 2a
3
mπy
cosh
mπb a
α m 5π 5 D cosh mπx
w= ∑ 2a sin
m =1, 3,K a
+ 2a 3 q o mπy 4a 4 q o
y sinh + 5 5
mπb a m π D
m π D cosh 2a
4 4
Chapter 10
Buckling and Instability
10.1 Introduction
The selection of structural members is based on three characteristics:
1. strength
2. stiffness
3. stability
Structural instability can occur in numerous situations where the compressive stresses
are present. For example,
Long and slender columns subjected to axial compression can buckle long before the
material reach the ultimate compressive strength.
Thin-walled tubes can wrinkle when subjected to axial compression.
Narrow beams, unbraced laterally, can turn sidewise and collapse under transverse
loads.
Vacuum tank can severely distort under external pressure.
The structural instability and buckling failures are occurred suddenly and dangerous.
For the structural members as shown in Fig. 10.1, we can classify the buckling modes as
followings:
Fig. 10.1
a.) For the column that has limited flexural stiffness but adequate torsional stiffness
subjected to compressive force, the dominant buckling mode is the flexural
buckling.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-2
b.) For the same column in a.) subjected to bending moment, the dominant buckling
mode is the flexural-torsional buckling.
c.) For the same column in a.) subjected to eccentric axial force, the dominant
buckling mode is the flexural-torsional buckling.
d.) For the column that has limited torsional stiffness but adequate flexural stiffness
subjected to compressive force, the dominant buckling mode is the torsional
buckling.
10.2 Column Buckling
Consider an ideal perfectly straight column with pinned supports at both ends as
shown in Fig. 10.2. The column is subjected to axially concentric compressive force P and
deformed as shown. The bending moment due to the axial force P is M = − Pv .
Fig. 10.2
The differential equation for the elastic curve of the column is
d 2v
EI 2 = M
dx
M = − Pv
d 2v
EI 2 = − Pv
dx
d 2v P
+ v=0
dx 2 EI
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-3
This is the homogenous, linear, differential equation of second order with constant
P
coefficients. It can be solved by assuming that k 2 = . Then,
EI
v ′′ + k 2 v = 0
The solution of this equation is in the form of
v = C1 sin kx + C2 cos kx
The constants of integration C1 and C2 can be determined by using the boundary
conditions.
At x = 0 , v = 0 ; C2 = 0
v = C1 sin kx
At x = L , v = 0 ; C1 sin kL = 0
This condition is satisfied either C1 = 0 or sin kL = 0 .
If the constant C1 = 0 , the term sin kL ≠ 0 . Then, the term kL can have any values
and the load P can also be any values since P = k 2 ( EI ) . Thus, C1 = 0 is the trivial solution.
If the term sin kL = 0 and C1 ≠ 0 , then, kL = 0, π, 2 π, 3π,..... . When the term
n 2 π 2 EI
P= n = 1, 2, 3,...
L2
The least force at which a buckled mode is possible is occurred when n = 1 and called
the critical or Euler buckling load.
π 2 EI
Pcr =
L2
where Pcr = critical bucking load. Pcr < Py
E = modulus of elasticity
I = moment of inertia
L = length of the column
Consider a column subjected to axial compressive load. If the column is so slender
that its material is always linear elastic until the critical load is reached at point B as shown
in Fig. 10.3a, the column can behave into two possible ways when subjected to an increasing
axial compressive load. If the column is an ideal column, the column may remain straight
(path BC ). If the column has a slight imperfection, the column may bend (path BD or path
BF ) depending on the analytical approaches. If the column has a larger imperfection, the
response of the column will follow the path OE .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-4
The somewhat slender column behaves similar to the very slender column as shown in
Fig. 10.3b. However, the material of the column will reach the yielding strength, as the
deflection grows larger and larger as shown by the downward curve OBD . If the column has
a larger imperfection, the response of the column will follow the path OE .
For a lesser slender column, the capacity of the column will be influenced by the
yielding strength of the material.
Fig. 10.3
10.3 Plate Buckling
The differential equation of a rectangular plate subjected to lateral load q is obtained
in previous chapter as
∂4w ∂4w ∂4w q
+ 2 + =
∂x 4 ∂x 2 ∂y 2 ∂y 4 D
Plates Subjected to Combined Bending and In-plane Loads
If the plate is additionally loaded in its plane by the compression as shown in the Fig.
10.4, summation of forces in the x direction gives
∂N x ∂N yx
N x + dx dy + N yx + dy dx − N x dy − N y dx = 0
∂x ∂y
∂N x ∂N yx
+ =0
∂x ∂y
Similarly, summation of forces in the y direction provides
∂N y ∂N xy
+ =0
∂y ∂x
In considering the forces in the in the z direction, we must take into account the
deflection of the plate. Due to the curvature of the plate in the xz plane, the projection of the
normal forces N x on the z axis is
∂w ∂N x ∂w ∂ 2 w
− ( N x dy ) + N x + dx + 2 dx dy
∂x ∂x ∂x ∂x
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-5
Fig. 10.4
By neglecting the higher order terms, we have
∂2w ∂N x ∂w
N x 2 dxdy + dxdy
∂x ∂x ∂x
Similarly, the projection of the normal forces N y on the z axis is
∂2w ∂N y ∂w
Ny dxdy + dxdy
∂y 2
∂y ∂y
∂N xy
Due to the shearing forces N xy and N xy + dx , the midsurface of the plate is
∂x
∂w ∂w ∂ 2 w
deformed as shown in Fig. 10.5. Owing to the angle and + dx , the shearing
∂y ∂y ∂x∂y
forces N xy have a projection on the z axis as
∂w ∂N xy ∂w ∂ 2 w
− N xy dy + N xy + dx + dx dy
∂y ∂x ∂y ∂x∂y
∂2w ∂N xy ∂w
N xy dxdy + dxdy
∂x∂y ∂x ∂y
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-6
Fig. 10.5
Similarly for N yx
∂2w ∂N yx ∂w
N yx dxdy + dxdy
∂x∂y ∂y ∂x
Since N yx = N xy , the final expression for the projection of the shearing forces on the
z axis is
∂2w ∂N xy ∂w ∂N yx ∂w
2 N xy dxdy + dxdy + dxdy
∂x∂y ∂x ∂y ∂y ∂x
The total summation of the forces in the z axis is
∂2w ∂N x ∂w ∂2w ∂N y ∂w
Nx dxdy + dxdy + N y dxdy + dxdy
∂x 2 ∂x ∂x ∂y 2 ∂y ∂y
∂2w ∂N xy ∂w ∂N yx ∂w
+ 2 N xy dxdy + dxdy + dxdy + qdxdy = 0
∂x∂y ∂x ∂y ∂y ∂x
∂N x ∂N yx ∂N y ∂N xy
By using the equations + = 0 and + = 0 , we have
∂x ∂y ∂y ∂x
Uo =
1
2
[
σ x ε x + σ y ε y + σ z ε z + τ xy γ xy + τ yz γ yz + τ xz γ xz ]
Thus, the strain energy of a small plate element is
U=
1
[ ]
∫ σ x ε x + σ y ε y + τ xyγ xy dxdydz
2 Vol
Ez ∂2w ∂2w
σx = − 2 + ν
1 −ν 2 ∂x ∂y 2
Ez ∂2w ∂2w
σy =− 2 +ν 2
1 −ν 2 ∂y ∂x
∂2w
τ xy = −2Gz
∂x∂y
and the strain-deflection relations,
∂2w
ε x = −z 2
∂x
∂2w
ε y = −z 2
∂y
∂2w
γ xy = −2 z
∂x∂y
into the equation of the strain energy of a small plate element, we have
D ∂ 2 w ∂ 2 w ∂ 2 w
2 2 2
∂2w ∂2w
U = ∫ 2 + 2 + 2ν 2 + 2(1 − ν ) dxdy
2 ∂x ∂y ∂x ∂y 2 ∂x∂y
D ∂ 2 w ∂ 2 w 2 ∂ 2 w ∂ 2 w ∂ 2 w 2
U= ∫ ∂x 2 + ∂y 2 − 2(1 − ν ) ∂x 2 ∂y 2 − ∂x∂y dxdy
2 Area
Strain Energy Due to In-plane Loads of Plates
The strain energy due to in-plane loads of plates is derived from Fig. 10.6, which
shows the deflection of a unit segment dx . Hence,
2
∂w
dx ′ = dx 2 − dx
∂x
2
1 ∂w
δ x = dx − dx ′ = dx
2 ∂x
or for a unit length
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-8
2
1 ∂w
εx =
2 ∂x
Fig. 10.6
Similarly,
2
1 ∂w
ε y =
2 ∂y
It can also shown that
∂w ∂w
γ xy =
∂x ∂y
Thus, the strain energy due to the in-plane forces is
U= ∫ [N ε
Area
x x ]
+ N y ε y + N xy γ xy dxdy
1 ∂w 2 ∂w
2
∂w ∂w
U = ∫ N x + N y + 2 N xy dxdy
2 Area ∂x ∂y ∂x ∂y
Strain Energy Due to Bending and In-plane Loads of Plates
The total strain energy due to bending and in-plane loads is
D ∂ 2 w ∂ 2 w 2 ∂ 2 w ∂ 2 w ∂ 2 w 2
U= ∫ ∂x 2 + ∂y 2 − 2(1 − ν ) ∂x 2 ∂y 2 − ∂x∂y dxdy
2 Area
1 ∂w 2 ∂w
2
∂w ∂w
+ ∫ N x + N y + 2 N xy dxdy
2 Area ∂x ∂y ∂x ∂y
The total potential energy of the plate is
V = Ω +U
In order for the plate to be in equilibrium, the total potential energy of the plate must
be minimum.
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-9
Example 10-1
Find the buckling stress of a simply supported rectangular plate subjected to the force
N x as shown in Fig. Ex 10-1a.
Fig. Ex 10-1a
Let the deflection of the plate be expressed as
α α
mπx nπy
w( x, y ) = ∑∑ Amn sin sin
m =1 n =1 a b
which satisfies the boundary condition of the plate.
Substituting the deflection equation into the total strain energy and noting that the
∂ 2w ∂ 2w ∂ 2w 2
term 2(1 − ν ) 2 − = 0 , we have
∂x ∂y
2
∂x∂y
b a α α
mπx 2 nπy
2
2 m π n 2π 2
2 2
D
U = ∫ ∫ ∑∑ Amn 2 + 2 sin 2 sin dxdy
2 0 0 m =1 n =1 a b a b
α α
1
b a
2 m π
2 2
2 mπx 2 nπy
+ ∫ ∫ (− N x )∑∑ Amn 2 sin sin dxdy
200 m =1 n =1 a a b
or
π 4 ab α α 2 m 2 n 2 π 2 b
2
N x ∑∑ [m 2 Amn ]
α α
U= D ∑∑ Amn 2 + 2 − 2
8 m =1 n =1 a b 8 a
m =1 n =1
The smallest value of the compressive forces N x ,cr is occurred when n = 1 . The
physical meaning of this is that a plate buckles in such a way that there can be several half-
waves in the direction of compression but only one half-wave in the perpendicular direction.
Thus,
2
π 2D 1 a2
N x ,cr = 2 m +
a m b 2
N cr Et 3
If we substituting σ cr = and D = , we get the critical stress in the form
t 12(1 − ν 2 )
of
π 2E
σ cr = 2
K
b
12(1 − ν )
2
t
2
m a/b
where K = + . The plot of the critical stress is shown in Fig. Ex 10-1b and shows
a/b m
that the minimum value of K is 4.0.
Fig. Ex 10-1b
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-11
Fig. 10.7
Fig. 10.7 shows a rectangular plate simply supported on sides x = 0 and x = a and
subjected to the axial compression N x . The governing differential equation of this plate is
∂2w
This solution satisfies the two boundary conditions w = = 0 at x = 0 and x = a .
∂x 2
Substituting the deflection solution into the governing differential equation, we obtain
d4 f d2 f
4
− A 2 + Bf = 0
dy dy
2 m 2π 2
where A =
a2
m 4π 4 N x m 2π 2
B= −
a4 D a2
The general solution of this fourth order differential equation is
f ( y ) = C1e −αy + C 2 eαy + C 3 cos βy + C 4 sin βy
m 2π 2 N x m 2π 2
where α = +
a2 D a2
m 2π 2 N x m 2π 2
β= − +
a2 D a2
The values of the constants C1 through C 4 are obtained from the boundary conditions
y = 0 and y = b .
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-12
Case 1
Side y = 0 is fixed and side y = b is free.
∂w
At y = 0 , the deflection w = 0 and the rotaion = 0.
∂y
∂2w ∂2w
At y = b , the moment M y = 2 + ν 2 = 0 and the shear
∂y ∂x
∂3w ∂3w
Q = 3 + (2 − ν ) = 0
∂y ∂x∂y 2
From the first boundary condition, we obtain
C1 + C 2 + C 3 = 0
β
f ( y ) = C 3 (cos βy − cosh αy ) + C 4 (sin βy − sinh αy )
α
Substituting f ( y ) into the deflection equation, we have
α
β mπx
w( x, y ) = ∑ C 3 (cos βy − cosh αy ) + C 4 (sin βy − sinh αy ) sin
m =1 α a
Using the last two boundary conditions, we obtain two simultaneous equations. The
critical value of the compressive force, N x , is determined by equating the determinant of
these equations to zero.
1
2 gh( g 2 + h 2 ) cos βb cosh βb = (α 2 h 2 − β 2 g 2 ) sin βb sinh βb
αβ
m 2π 2 m 2π 2
where g = α 2 − ν and h = β 2
+ ν . For m = 1 , the minimum value of the critical
a2 a2
compressive stress is
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-13
π 2E
σ cr = 2
K
b
12(1 − ν )
2
t
For v = 0.25 ,
K min = 1.328
Case 2
Side y = 0 is simply supported and side y = b is free.
Similarly, in this case, the maximum value of K is
b2
K = 0.456 + 2 for ν = 0.25
a
Case 3
Side y = 0 and side y = b are fixed. In this case, the maximum value of K is
K = 7.0 for ν = 0.25
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-14
Example 10-2
Let the plate in Fig. Ex 10-1a be simply supported at x = 0 and x = a , simply
supported at y = 0 , and free at y = b . Determine the required thickness if a = 0.560 m ,
b = 0.430 m , N x = 52500 N/m , ν = 0.25 , σ y = 250 MPa , E = 200 GPa , and factor of safety
= 2.0.
Assuming that t = 6.5 mm . From case 2, the value of K = 1.046 . Then, the critical
compressive stress is
π 2 (200000)
σ cr = 2
(1.046) = 41.93 MPa
0.430
12(1 − 0.25 2 )
0.0065
which is significantly less than the yielding stress σ y = 250 MPa .
b K
= 162
t σy
Fig. 10.8
However, due to the residual stress occurred in the steel member during the
manufacturing, the actual interaction curve is represented by the curve ADC . Therefore,
AISC use a factor of 0.7 to account for this effect and
b K
= 114
t σy
Single Angles
Consider the leg AB of a single angle as shown in Fig. 10.9a as a plate. The plate AB
has a free support at point B and has simply supported support at point A since the point A
b2
can only rotate due to the deflection. Thus, K = 0.456 + and K min = 0.456 . Then,
a2
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-16
b 0.456 76
= 114 =
t σy σy
Double-Angles
Due to the symmetry of the double angles as shown in Fig. 10.9b, the possibility of
rotation of the section under the axial compression load is significantly reduced from the
previous case. Thus, AISC uses the average of the case 2 (simply supported-free) and the
average of the case 1 (fixed-free) and the case 2 (simply supported-free).
0.456 + 1.328
0.456 +
K= 2 = 0.674
2
Therefore, we have
b 0.674 95
= 114 =
t σy σy
Fig. 10.9
Advanced Mechanics of Materials by Dr. Sittichai Seangatith 10-17
Stems of T-section
Consider the stem of the T-section as shown in Fig. 10.9c as a plate. The support at
point A is considered as a fixed support and the support at point B is considered as a free
support. Therefore, the K value is 1.328 for case 1.
b 1.328 132
= 114 =
t σy σy
In practice, the AISC reduces the coefficient of this equation from 132 to 127.
b 4.0 228
= 114 =
t σy σy
The AISC increases the coefficient of this equation from 228 to 238 to match the
experimental results.
Perforated Cover plates
For the perforated plate as shown edge in Fig. 10.9e, the supports of the plate between
the perforation and the edge are assumed to be fixed. This is because the continuous areas
between the perforations add more rigidity to the plate. If the ratio of the dimension a and b
of the perforated plate is equal to one, the value of K is about 7.69. This value is higher than
that obtained in case 3 since it is based on the smallest possible value of K . Thus,
b 7.69 317
= 114 =
t σy σy
b 4.90 253
= 114 =
t σy σy
References
1. Advanced Mechanics of Materials; 4th Edition, A.P. Boresi and O.M. Sidebottom, John
Wiley & Sons, 1985
2. Advanced Mechanics of Materials; 2nd Edition, R.D. Cook and W.C. Young, Prentice
Hall, 1999
3. Theory of Elastic Stability; 2nd Edition, S.P. Timoshenko and J.M. Gere, McGraw-Hill,
1963
4. Theory of Elasticity; 3rd Edition, S.P. Timoshenko and J.N. Goodier, McGraw-Hill,
1970
5. Theory of Plates and Shells; 2nd Edition, S.P. Timoshenko and S. Woinowsky-Krieger,
McGraw-Hill, 1970
6. Mechanical Behavior of Materials; 2nd Edition, N.E. Dowling, Prentice Hall, 1999
7. Mechanics of Materials; 3th Edition, R.C. Hibbeler, Prentice Hall, 1997