Professional Documents
Culture Documents
CM PDF
CM PDF
FOR ENGINEERS
THEORY AND PROBLEMS
Second Edition
X. Oliver
C. Agelet de Saracibar
Xavier Oliver
Carlos Agelet de Saracibar
Continuum Mechanics
for Engineers
Theory and Problems
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Continuum Mechanics for Engineers. Theory and Problems
First edition: September 2016
Second edition: March 2017
Cite as:
X. Oliver and C. Agelet de Saracibar, Continuum Mechanics
for Engineers. Theory and Problems, 2nd edition, March 2017,
doi:10.13140/RG.2.2.25821.20961.
Open Access This book is distributed under the terms of the Creative Commons
Attribution Non-Commercial No-Derivatives (CC-BY-NC-ND) License, which
permits any noncommercial use, distribution, and reproduction in any medium
of the unmodified original material, provided the original author(s) and source
are credited.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Foreword
This book was born with the vocation of being a tool for the training of engi-
neers in continuum mechanics. In fact, it is the fruit of the experience in teaching
this discipline during many years at the Civil Engineering School of the Tech-
nical University of Catalonia (UPC/BarcelonaTech), both in undergraduate de-
grees (Civil Engineering and Geological Engineering) and postgraduate degrees
(Master and PhD courses). Unlike other introductory texts to the mechanics of
continuous media, the work presented here is specifically aimed at engineering
students. We try to maintain a proper balance between the rigor of the math-
ematical formulation used and the clarity of the physical principles addressed,
although always putting the former at the service of the latter. In this sense, the
essential vector and tensor operations use simultaneously the indicial notation
(more useful for rigorous mathematical proof) and the compact notation (which
allows for a better understanding of the physics of the problem). However, as the
text progresses, there is a clear trend towards compact notation in an attempt to
focus the reader’s attention on the physical component of continuum mechanics.
The text content is intentionally divided into two specific parts, which are pre-
sented sequentially. The first part (Chapters 1-5) introduces fundamental and de-
scriptive aspects common to all continuous media (motion, deformation, stress
and conservation-balance equations). In the second (Chapters 6 to 11), specific
families of the continuous medium are studied, such as solids and fluids, in an
approach that starts with the corresponding constitutive equation and ends with
the classical formulations of solid mechanics (elastic-linear and elasto-plastic)
and fluid mechanics (laminar regime). Finally, a brief incursion into the varia-
tional principles (principle of virtual work and minimization of potential energy)
is attempted, to provide the initial ingredients needed to solve continuum me-
chanics problems using numerical methods. This structure allows the use of this
text for teaching purposes both in a single course of about 100 teaching hours
or as two different courses: the first based on the first five chapters dedicated to
the introduction of the fundamentals of continuum mechanics and, the second,
specifically dedicated to solid and fluid mechanics. The theoretical part in every
chapter is followed by a number of solved problems and proposed exercises so
v
vi
Xavier Oliver
and
Carlos Agelet de Saracibar
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
1 Description of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Definition of the Continuous Medium . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Equations of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Descriptions of Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1 Material Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.2 Spatial Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Time Derivatives: Local, Material and Convective . . . . . . . . . . . . . 8
1.5 Velocity and Acceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.6 Stationarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.7 Trajectory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7.1 Differential Equation of the Trajectories . . . . . . . . . . . . . . . 16
1.8 Streamline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.8.1 Differential Equation of the Streamlines . . . . . . . . . . . . . . . . 18
1.9 Streamtubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.9.1 Equation of the Streamtube . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.10 Streaklines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.10.1 Equation of the Streakline . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.11 Material Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.12 Control Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.13 Material Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.14 Control Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Problems and Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2 Strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.2 Deformation Gradient Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.2.1 Inverse Deformation Gradient Tensor . . . . . . . . . . . . . . . . . . 43
2.3 Displacements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.3.1 Material and Spatial Displacement Gradient Tensors . . . . . 46
vii
viii Contents
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Contents ix
4 Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.1 Forces Acting on a Continuum Body . . . . . . . . . . . . . . . . . . . . . . . . 127
4.1.1 Body Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.1.2 Surface Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.2 Cauchy’s Postulates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.3 Stress Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.3.1 Application of Newton’s 2nd Law to a Continuous Medium132
4.3.2 Stress Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
4.3.3 Graphical Representation of the Stress State in a Point . . . 138
4.4 Properties of the Stress Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.4.1 Cauchy Equation. Internal Equilibrium Equation . . . . . . . . 141
4.4.2 Equilibrium Equation at the Boundary . . . . . . . . . . . . . . . . . 142
4.4.3 Symmetry of the Cauchy Stress Tensor . . . . . . . . . . . . . . . . 142
4.4.4 Diagonalization. Principal Stresses and Directions . . . . . . . 144
4.4.5 Mean Stress and Mean Pressure . . . . . . . . . . . . . . . . . . . . . . 146
4.4.6 Decomposition of the Stress Tensor into its Spherical
and Deviatoric Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.4.7 Tensor Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.5 Stress Tensor in Curvilinear Orthogonal Coordinates . . . . . . . . . . . 149
4.5.1 Cylindrical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.5.2 Spherical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
4.6 Mohr’s Circle in 3 Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
4.6.1 Graphical Interpretation of the Stress States . . . . . . . . . . . . 151
4.6.2 Determination of the Mohr’s Circles . . . . . . . . . . . . . . . . . . . 153
4.7 Mohr’s Circle in 2 Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.7.1 Stress State on a Given Plane . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.7.2 Direct Problem: Diagonalization of the Stress Tensor . . . . . 161
4.7.3 Inverse Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
4.7.4 Mohr’s Circle for Plane States (in 2 Dimensions) . . . . . . . . 162
4.7.5 Properties of the Mohr’s Circle . . . . . . . . . . . . . . . . . . . . . . . 164
4.7.6 The Pole of Mohr’s Circle . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.7.7 Mohr’s Circle with the Soil Mechanics Sign Criterion . . . . 171
4.8 Mohr’s Circle for Particular Cases . . . . . . . . . . . . . . . . . . . . . . . . . . 172
4.8.1 Hydrostatic Stress State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
4.8.2 Mohr’s Circles for a Tensor and its Deviator . . . . . . . . . . . . 173
4.8.3 Mohr’s Circles for a Plane Pure Shear Stress State . . . . . . . 173
Problems and Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
x Contents
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Contents xi
8 Plasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
8.2 Previous Notions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
8.2.1 Stress Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
8.2.2 Spherical and Deviatoric Components of the Stress Tensor 372
8.3 Principal Stress Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
8.3.1 Normal and Shear Octahedral Stresses . . . . . . . . . . . . . . . . . 376
8.4 Rheological Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
xii Contents
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Contents xiii
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 1
Description of Motion
rs
n ee
gi
1.1 Definition of the Continuous Medium
t d le En
ar
A continuous medium is understood as an infinite set of particles (which form
ar s
ge ro or
eS m
ib
part of, for example, solids or fluids) that will be studied macroscopically, that
ac
f
is, without considering the possible discontinuities existing at microscopic level
C d P cs
(atomic or molecular level). Accordingly, one admits that there are no discon-
b
i
tinuities between the particles and that the mathematical description of this
an an n
le
liv or ec
.A
particle along time. In general, these functions and their derivatives are required
e
X Th
to be continuous.
er
tin
on
.O
1
2 C HAPTER 1. D ESCRIPTION OF M OTION
Ω0 – reference configuration
t0 – reference time
Ωt – present configuration
t – present time
rs
ee
Figure 1.1: Configurations of the continuous medium.
n
gi
t d le En
uous medium at time t, denoted by Ωt , is defined as the locus of the positions
occupied in space by the material points (particles) of the continuous medium at
ar
ar s
the given time.
ge ro or
eS m
ib
A certain time t = t0 of the time interval of interest is referred to as the ref-
ac
f
erence time and the configuration at this time, denoted by Ω0 , is referred to as
C d P cs
initial, material or reference configuration1 .
b
i
an an n
Consider now the Cartesian coordinate system (X,Y, Z) in Figure 1.1 and the
y ha
.A
where the components (X1 , X2 , X3 ) are referred to as material coordinates (of the
d
uu
er
tin
X1
on
de f
.O
not
X ≡ [X] = X2 = material coordinates. (1.2)
C
X3
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Equations of Motion 3
rs
The motion of the particles of the continuous medium can now be described
ee
by the evolution of their spatial coordinates (or their position vector) along time.
n
Mathematically, this requires the definition of a function that provides for each
gi
particle (identified by its label) its spatial coordinates xi (or its spatial position
t d le En
vector x) at successive instants of time. The material coordinates Xi of the par-
ar
ticle can be chosen as the label that univocally characterizes it and, thus, the
ar s
ge ro or
eS m
ib
equation of motion
ac
(
f not
C d P cs
x = ϕ (particle,t) = ϕ (X,t) = x (X,t)
b (1.5)
i
xi = ϕi (X1 , X2 , X3 ,t) i ∈ {1, 2, 3}
an an n
y ha
is obtained, which provides the spatial coordinates in terms of the material ones.
le
liv or ec
The spatial coordinates xi of the particle can also be chosen as label, defining
the inverse equation of motion6 as
M
.A
not
(
X = ϕ −1 (x,t) = X (x,t) ,
m
(1.6)
d
er
tin
.O
C
Remark 1.1. There are different alternatives when choosing the la-
©
bel that characterizes a particle, even though the option of using its
material coordinates is the most common one. When the equation of
motion is written in terms of the material coordinates as label (as in
(1.5)), one refers to it as the equation of motion in canonical form.
5 Whenever possible, uppercase letters will be used to denote variables relating to the refer-
ence configuration Ω0 and lowercase letters to denote the variables referring to the current
configuration Ωt .
6 With certain abuse of notation, the function will be frequently confused with its image.
Hence, the equation of motion will be often written as x = x (X,t) and its inverse equation as
X = X (x,t).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
4 C HAPTER 1. D ESCRIPTION OF M OTION
rs
∂ ϕ (X,t) not ∂ ϕ (X,t)
• The Jacobian of the transformation J = det = > 0.
ee
∂X ∂X
n
The physical interpretation of this condition (which will be studied later) is
gi
that every differential volume must always be positive or, using the principle of
t d le En
mass conservation (which will be seen later), the density of the particles must
always be positive.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Remark 1.2. The equation of motion at the reference time t = 0 re-
b
sults in x (X,t)|t=0 = X. Accordingly, x = X, y = Y , z = Z is the
i
an an n
equation of motion at the reference time and the Jacobian at this in-
y ha
J (X, 0) =
= det = det [δi j ] = det 1 = 1.
∂ (XY Z) ∂ Xj
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
7 not
The two-index operator Delta Kronecker = δi j is defined as δi j = 0 when i 6= j and δi j = 1
when i = j. Then, the unit tensor 1 is defined as [1]i j = δi j .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Equations of Motion 5
rs
ee
x1 = X1 e2t x = Xe2t
n
not
x (X,t) ≡ x2 = X2 e−2t = y = Y e−2t
gi
x3 = 5X1t + X3 e2t z = 5Xt + Ze2t
t d le En
ar
ar s
Obtain the inverse equation of motion.
ge ro or
eS m
ib
ac
Solution f
C d P cs
b
i
The determinant of the Jacobian is computed as
an an n
y ha
le
∂ x1 ∂ x1 ∂ x1
liv or ec
∂ X1 ∂ X2 ∂ X3 e2t 0 0
M
.A
∂ xi ∂ x ∂ x2 ∂ x2 = 0 e−2t 0 = e2t 6= 0.
J = = 2
m
∂ Xj
∂ X1 ∂ X2 ∂ X3
d
2t
uu
5t 0 e
e
∂ x3 ∂ x3 ∂ x3
X Th
er
tin
∂X ∂ X2 ∂ X3
1
on
.O
The sufficient (but not necessary) condition for the function x = ϕ (X,t) to
be biunivocal (that is, for its inverse to exist) is that the determinant of the
C
Jacobian of the function is not null. In addition, since the Jacobian is positive,
©
the motion has physical sense. Therefore, the inverse of the given spatial
description exists and is determined by
X1 x1 e−2t
X = ϕ −1 (x,t) ≡
not
2t
X2
= x 2 e .
X3 x3 e−2t − 5tx1 e−4t
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
6 C HAPTER 1. D ESCRIPTION OF M OTION
rs
In the material description8 , a given property (for example, the density ρ) is
ee
described by a certain function ρ (•,t) : R3 × R+ → R+ , where the argument (•)
in ρ (•,t) represents the material coordinates,
n
gi
ρ = ρ (X,t) = ρ (X1 , X2 , X3 ,t) . (1.7)
t d le En
ar
Here, if the three arguments X ≡ (X1 , X2 , X3 ) are fixed, a specific particle is being
ar s
ge ro or
eS m
ib
followed (see Figure 1.3) and, hence, the name of material description.
ac
f
C d P cs
1.3.2 Spatial Description
b
i
an an n
In the spatial description9 , the focus is on a point in space. The property is de-
y ha
.A
Then, when the argument x in ρ = ρ (x,t) is assigned a certain value, the evolu-
m
tion of the density for the different particles that occupy the point in space along
d
uu
time is obtained (see Figure 1.3). Conversely, fixing the time argument in (1.8)
e
X Th
Obviously, the direct and inverse equations of motion allow shifting from one
on
.O
C
Figure 1.3: Material description (left) and spatial description (right) of a property.
8 Literature on this topic also refers to the material description as Lagrangian description.
9 The spatial description is also referred to as Eulerian description.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Descriptions of Motion 7
rs
z = −Xt + Z
ee
Obtain the spatial description of the property whose material description is
n
gi
X +Y + Z
t d le En
ρ (X,Y, Z,t) = .
1 + t2
ar
ar s
ge ro or
eS m
Solution
ib
ac
f
The equation of motion is given in the canonical form since in the reference
C d P cs
configuration Ω0 its expression results in
b
i
an an n
y ha
x=X
" #
not
x = X (X, 0) ≡ y = Y .
le
liv or ec
z=Z
M
.A
uu
∂x ∂x ∂x
e
X Th
er
∂ X ∂Y ∂Z
1 −t 0
tin
∂ xi ∂ y ∂ y ∂y
= t 1 0 = 1 + t 2 6= 0
J = =
on
.O
∂ X j ∂ X ∂Y ∂Z
−t 0 1
C
∂z ∂z ∂z
©
∂ X ∂Y ∂Z
and the inverse equation of motion is given by
x + yt
X =
1 + t2
not
X (x,t) ≡ Y = y − xt .
1 + t 2
z + zt 2 + xt + yt 2
Z=
1 + t2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
8 C HAPTER 1. D ESCRIPTION OF M OTION
x + yt + y + z + zt 2 + yt 2
ρ (X,Y, Z,t) ≡ = ρ (x, y, z,t) .
(1 + t 2 )2
rs
n ee
1.4 Time Derivatives: Local, Material and Convective
gi
t d le En
The consideration of different descriptions (material and spatial) of the proper-
ar
ties of the continuous medium leads to diverse definitions of the time derivatives
ar s
ge ro or
eS m
ib
of these properties. Consider a certain property and its material and spatial de-
ac
scriptions,
f
C d P cs
Γ (X,t) = γ (x,t) , (1.10)
b
i
an an n
in which the change from the spatial to the material description and vice versa
y ha
is performed by means of the equation of motion (1.5) and its inverse equa-
le
tion (1.6).
liv or ec
M
.A
er
written as10
tin
not ∂ γ (x,t)
local derivative = .
on
.O
∂t
The material derivative of a property is its variation along time fol-
C
not d ∂Γ (X,t)
material derivative = Γ= .
dt ∂t
10 The expression ∂ (•,t)/∂t is understood in the classical sense of partial derivative with
respect to the variable t.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Time Derivatives: Local, Material and Convective 9
However, taking the spatial description of the property γ (x,t) and considering
the equation of motion is implicit in this expression yields
γ (x,t) = γ (x (X,t) ,t) = Γ (X,t) . (1.11)
Then, the material derivative (following a particle) is obtained from the spatial
description of the property as
not d ∂Γ (X,t)
material derivative = γ (x (X,t) ,t) = . (1.12)
dt ∂t
rs
Expanding (1.12) results in11
ee
dγ (x (X,t) ,t) ∂ γ (x,t) ∂ γ ∂ xi ∂ γ (x,t) ∂ γ ∂ x
n
= + = + · =
gi
dt ∂t ∂ xi ∂t ∂t ∂ x |{z}
∂t
(1.13)
t d le En
∂ γ (x,t) ∂ γ v (x,t)
ar
= + · v (x,t) ,
ar s
ge ro or
∂t ∂x
eS m
ib
ac
f
where the definition of velocity as the derivative of the equation of motion (1.5)
C d P cs
b
with respect to time has been taken into account,
i
an an n
∂ x (X,t)
y ha
The deduction of the material derivative from the spatial description can be
M
.A
generalized for any property χ (x,t) (of scalar, vectorial or tensorial character)
as12
m
dχ (x,t) ∂ χ (x,t)
uu
dt ∂t
X Th
er
tin
| {z } | {z } | {z }
material local convective
derivative derivative derivative
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
10 C HAPTER 1. D ESCRIPTION OF M OTION
Solution
rs
The material description of the property is obtained introducing the equation
ee
of motion into its spatial description,
n
gi
ρ (X,Y, Z,t) = 3 (X +Y t + Zt)+2 (Y + 2Zt)+3t = 3X +3Y t +7Zt +2Y +3t .
t d le En
The material derivative is then calculated as the derivative of the material
ar
description with respect to time,
ar s
ge ro or
eS m
ib
ac
∂ρ
f
= 3Y + 7Z + 3 .
C d P cs
∂t
b
i
an an n
dρ ∂ρ
M
.A
= + v · ∇ρ with
dt ∂t
m
∂ρ ∂x
= [Y + Z, 2Z, 3X]T and ∇ρ = [3, 2, 0]T .
uu
=3, v=
e
∂t ∂t
X Th
er
tin
.O
dρ
= 3 + 3Y + 7Z
C
dt
©
is obtained. Note that the expressions for the material derivative obtained
from the material description, ∂ ρ/∂t, and the spatial description, dρ/dt, co-
incide.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Velocity and Acceleration 11
rs
∂ x (X,t)
V (X,t) =
ee
∂t (1.16)
n
Vi (X,t) = i (X,t) i ∈ {1, 2, 3}
∂ x
gi
∂t
t d le En
and, if the inverse equation of motion X = ϕ −1 (x,t) is known, the spatial de-
ar
ar s
ge ro or
scription of the velocity can be obtained as
eS m
ib
ac
f
v (x,t) = V (X (x,t) ,t) . (1.17)
C d P cs
b
i
an an n
y ha
field.
M
.A
m
acceleration is given by
X Th
er
tin
∂ V (X,t)
A (X,t) =
on
.O
∂t (1.18)
C
∂t
and, through the inverse equation of motion X = ϕ −1 (x,t), the spatial descrip-
tion is obtained, a (x,t) = A (X (x,t) ,t). Alternatively, if the spatial description
of the velocity is available, applying (1.15) to obtain the material derivative of
v (x,t),
dv (x,t) ∂ v (x,t)
a (x,t) = = + v (x,t) · ∇v (x,t) , (1.19)
dt ∂t
directly yields the spatial description of the acceleration.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
12 C HAPTER 1. D ESCRIPTION OF M OTION
Example 1.4 – Consider the solid in the figure below, which rotates at a
constant angular velocity ω and has the expression
(
x = R sin (ωt + φ )
y = R cos (ωt + φ )
as its equation of motion. Find the velocity and acceleration of the motion
described both in material and spatial forms.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
Solution
f
C d P cs
The equation of motion can be rewritten as
b
i
an an n
y ha
(
x = R sin (ωt + φ ) = R sin (ωt) cos φ + R cos (ωt) sin φ
le
liv or ec
.A
and, since for t = 0, X = R sin φ and Y = R cos φ , the canonical form of the
m
( (
x = X cos (ωt) +Y sin (ωt) X = x cos (ωt) − y sin (ωt)
X Th
er
tin
and .
y = −X sin (ωt) +Y cos (ωt) Y = x sin (ωt) + y cos (ωt)
on
.O
∂x
∂ x (X,t) not = −Xω sin (ωt) +Y ω cos (ωt)
V (X,t) = )≡ ∂t
∂t
∂y
= −Xω cos (ωt) −Y ω sin (ωt)
∂t
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Velocity and Acceleration 13
rs
" #
X cos (ωt) +Y sin (ωt)
ee
= −ω 2
−X sin (ωt) +Y cos (ωt)
n
gi
Acceleration in spatial description:
t d le En
Replacing the canonical form of the equation of motion into the material
ar
description of the acceleration results in
ar s
ge ro or
eS m
ib
ac
f
" #
not −ω 2 x
C d P cs
a (x,t) = A (X (x,t) ,t) ≡ .
b −ω 2 y
i
an an n
y ha
This same expression can be obtained if the expression for the velocity v (x,t)
le
liv or ec
and the definition of material derivative in (1.15) are taken into account,
M
.A
dv (x,t) ∂ v (x,t)
a (x,t) = = + v (x,t) · ∇v (x,t) =
m
dt ∂t
d
uu
e
∂
X Th
er
∂ ωy
∂x
tin
not
≡ + ωy , −ωx ωy , −ωx ,
∂t −ωx
∂
on
.O
∂y
C
©
∂ ∂
(ωy) (−ωx)
" # " #
0 −ω 2x
+ ωy , −ωx ∂ x ∂x
not
a (x,t) ≡ = .
0 ∂ ∂ −ω 2 y
(ωy) (−ωx)
∂y ∂y
Note that the result obtained using both procedures is identical.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
14 C HAPTER 1. D ESCRIPTION OF M OTION
1.6 Stationarity
According to the above definition, and considering the concept of local deriva-
tive, any stationary property has a null local derivative. For example, if the ve-
rs
locity for a certain motion is stationary, it can be described in spatial form as
ee
∂ v (x,t)
n
v (x,t) = v (x) ⇐⇒ =0. (1.20)
gi
∂t
t d le En
ar
ar s
ge ro or
Remark 1.5. The non-dependence on time of the spatial description
eS m
ib
(stationarity) assumes that, for a same point in space, the property
ac
f
being considered does not vary along time. This does not imply that,
C d P cs
b
for a same particle, such property does not vary along time (the ma-
i
an an n
v (x,t) is stationary,
le
liv or ec
.A
labeled X1 and X2 that have varying densities along time, when oc-
uu
e
cupying a same spatial point x (at two different times t1 and t2 ) their
X Th
er
tin
.O
C
That is, for an observer placed outside the medium, the density of
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Trajectory 15
Solution
not
The velocity field in Example 1.4 is v (x) ≡ [ωy , −ωx]T . Therefore, it is a
case in which the spatial description of the velocity is not dependent on time
and, thus, the velocity is stationary. Obviously, this implies that the velocity
of the particles (whose motion is a uniform rotation with respect to the origin,
with angular velocity ω) does not depend on time (see figure below). The
rs
direction of the velocity vector for a same particle is tangent to its circular
trajectory and changes along time.
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
dv (x) ∂ v (x)
M
.A
appears due to the change in direction of the velocity vector of the particles
d
uu
er
tin
1.7 Trajectory
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
16 C HAPTER 1. D ESCRIPTION OF M OTION
rs
by a trajectory characterized by the value of the label (material coordinates) X.
Then, the equation of motion defines a family of curves whose elements are the
ee
trajectories of the various particles.
n
gi
1.7.1 Differential Equation of the Trajectories
t d le En
Given the velocity field in spatial description v (x,t), the family of trajectories
ar
ar s
ge ro or
can be obtained by formulating the system of differential equations that imposes
eS m
ib
that, for each point in space x, the velocity vector is the time derivative of the
ac
f
parametric equation of the trajectory defined in (1.21), i.e.,
C d P cs
b
i
an an n
dx (t)
= v (x (t) , t) ,
y ha
Find x (t) := dt (1.22)
le
liv or ec
dt
.A
(
X Th
(1.23)
x = φi (C1 ,C2 ,C3 , t) i ∈ {1, 2, 3} .
on
.O
constants (C1 ,C2 ,C3 ). Assigning a particular value to these constants yields a
©
and, replacing into (1.23), the canonical form of the equation of the trajectory,
X = φ (C1 (X) ,C2 (X) ,C3 (X) , t) = ϕ (X,t) , (1.25)
is obtained.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Trajectory 17
not
Example 1.6 – Given the velocity field in Example 1.5, v (x) ≡ [ωy , −ωx]T ,
obtain the equation of the trajectory.
Solution
Using expression (1.22), one can write
dx (t)
dx (t)
= vx (x,t) = ωy ,
= v (x,t) =⇒ dt
dt
dy (t) = vy (x,t) = −ωx .
rs
dt
n ee
This system of equations is a system with crossed variables. Differentiating
gi
the second equation and replacing the result obtained into the first equation
t d le En
yields
d 2 y (t) dx (t)
= −ω 2 y (t) =⇒ y00 + ω 2 y = 0 .
ar
= −ω
ar s
ge ro or
dt 2 dt
eS m
ib
ac
f
The characteristic equation of this second-order differential equation is
C d P cs
r2 + ω 2 = 0 and its characteristic solutions are r j = ±iω j ∈ {1, 2}.
b
i
Therefore, the y component of the equation of the trajectory is
an an n
y ha
y (t) = Real Part C1 eiwt +C2 e−iwt = C1 cos (ωt) +C2 sin (ωt) .
le
liv or ec
M
The solution for x (t) is obtained from dy/dt = −ωx , which results in
.A
(
x (C1 ,C2 ,t) = C1 sin (ωt) −C2 cos (ωt) ,
uu
e
X Th
.O
x (C1 ,C2 , 0) = X ,
that is, (
x (C1 ,C2 , 0) = −C2 = X ,
y (C1 ,C2 , 0) = C1 = Y .
Finally, the equation of motion, or the equation of the trajectory, in canonical
form (
x = Y sin (ωt) + X cos (ωt)
y = Y cos (ωt) − X sin (ωt)
is obtained.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
18 C HAPTER 1. D ESCRIPTION OF M OTION
1.8 Streamline
Definition 1.7. The streamlines are a family of curves that, for every
instant of time, are the velocity field envelopes13 .
rs
According to its definition, the tangent at each point of a streamline has the same
ee
direction (though not necessarily the same magnitude) as the velocity vector at
n
that same point in space.
gi
t d le En
ar
Remark 1.6. In general, the velocity field (in spatial description) will
ar s
ge ro or
eS m
ib
be different for each instant of time (v ≡ v (x,t)). Therefore, one
ac
f
must speak of a different family of streamlines for each instant of
C d P cs
time (see Figure 1.6).
b
i
an an n
y ha
le
liv or ec
.A
Consider a given time t ∗ and the spatial description of the velocity field at this
m
a certain parameter λ . Then, the vector tangent to the streamline is defined, for
e
X Th
er
tin
on
.O
C
13The envelopes of a vector field are the family of curves whose tangent vector has, at each
point, the same direction as the corresponding vector of the vector field.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Streamline 19
each value of λ 14 , by dx (λ )/dλ and the vector field tangency condition can be
written as follows.
dx (λ )
= v (x (λ ) , t ∗ ) ,
Find x (λ ) := dλ (1.26)
dxi (λ ) = vi (x (λ ) , t ∗ ) i ∈ {1, 2, 3} .
dλ
The expressions in (1.26) constitute a system of first-order differential equa-
tions whose solution for each time t ∗ , which will depend on three integration
rs
constants (C10 ,C20 ,C30 ), provides the parametric expression of the streamlines,
ee
(
x = φ (C10 ,C20 ,C30 , λ , t ∗ ) ,
n
(1.27)
gi
xi = φi (C10 ,C20 ,C30 , λ , t ∗ ) i ∈ {1, 2, 3} .
t d le En
Each triplet of integration constants (C10 ,C20 ,C30 ) identifies a streamline whose
ar
ar s
ge ro or
eS m
ib
points, in turn, are obtained by assigning values to the parameter λ . For each
ac
time t ∗ a new family of streamlines is obtained.
f
C d P cs
b
i
an an n
y ha
ent viewpoints:
M
.A
• The fact that the time variable does not appear in (1.22) or (1.26)
m
er
tin
both systems must be, therefore, the same, except for the name
of the parameter used in each type of curves.
on
.O
14 It is assumed that the value of the parameter λ is chosen such that, at each point in space
x, not only does dx (λ )/dλ have the same direction as the vector v (x,t), but it coincides
therewith.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
20 C HAPTER 1. D ESCRIPTION OF M OTION
1.9 Streamtubes
In non-stationary cases, even though the closed line does not vary in space, the
rs
streamtube and streamlines do change. On the contrary, in a stationary case, the
ee
streamtube remains fixed in space along time.
n
gi
1.9.1 Equation of the Streamtube
t d le En
Streamlines constitute a family of curves of the type
ar
ar s
ge ro or
x = f (C1 ,C2 ,C3 , λ , t) . (1.28)
eS m
ib
ac
f
The problem consists in determining, for each instant of time, which curves
C d P cs
b
of the family of curves of the streamlines cross a closed line, which is fixed in the
i
an an n
is
le
Γ := x = g (s) . (1.29)
liv or ec
To this aim, one imposes, in terms of the parameters λ ∗ and s∗ , that a same point
M
.A
(
uu
(1.30)
X Th
er
.O
s∗ = s∗ (C1 ,C2 , t) ,
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Streaklines 21
rs
n ee
1.10 Streaklines
gi
t d le En
ar
ar s
∗
ge ro or
Definition 1.9. A streakline, relative to a fixed
point in space x
eS m
ib
named spill point and at a time interval ti ,t f named spill period,
ac
is the locus of the positions occupied at timefT t by all the particles
C d P cs
b
that have occupied x∗ over the time τ ∈ [ti ,t] ti ,t f .
i
an an n
y ha
le
liv or ec
The above definition corresponds to the physical concept of the color line
M
.A
(streak) that would be observed in the medium at time t if a tracer fluid were
injected at spill point x∗ throughout the time interval ti ,t f (see Figure 1.8).
m
d
uu
e
X Th
er
tin
on
.O
C
Figure 1.8: Streakline corresponding to the spill period τ ∈ ti ,t f .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
22 C HAPTER 1. D ESCRIPTION OF M OTION
rs
and replacing (1.33) into the equation of motion (1.5) results in
ee
\
x = ϕ (f (τ) , t) = g (τ, t) τ ∈ [ti ,t] ti ,t f . (1.34)
n
gi
Expression (1.34) is, for each time t, the parametric expression (in terms of
t d le En
parameter τ) of a curvilinear segment in space which is the streakline at that
ar
time.
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Example 1.7 – Given the equation of motion
b
i
an an n
x = (X +Y )t 2 + X cost ,
y ha
le
y = (X +Y ) cost − X ,
liv or ec
M
.A
obtain the equation of the streakline associated with the spill point x∗ = (0, 1)
m
Solution
X Th
er
tin
The material coordinates of a particle that has occupied the spill point at time
on
.O
τ are given by
C
) −τ 2
2 X = ,
©
0 = (X +Y ) τ + X cos τ
τ 2 + cos2 τ
=⇒
1 = (X +Y ) cos τ − X 2
Y = τ + cos τ .
τ 2 + cos2 τ
Therefore, the label of the particles that have occupied the spill point from
the initial spill time t0 until the present time t is defined by
−τ 2
X= 2 2
τ + cos τ
\
τ ∈ [t0 ,t] [t0 , ∞) = [t0 ,t] .
τ 2 + cos τ
Y= 2
τ + cos2 τ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Material Surface 23
Then, replacing these into the equation of motion, the equation of the streak-
line is obtained,
cos τ 2 −τ 2
x= 2 t + 2 cost
not
x = g (τ,t) ≡ τ + cos2 τ τ + cos2 τ τ ∈ [t0 ,t] .
cos τ −τ 2
y= 2 cost − 2
τ + cos2 τ τ + cos2 τ
rs
Remark 1.8. In a stationary problem, the streaklines are segments of
ee
the trajectories (or of the streamlines). The rationale is based on the
n
fact that, in the stationary case, the trajectory follows the envelope of
gi
the velocity field, which remains constant along time. If one consid-
ers a spill point x∗ , all the particles that occupy this point will follow
t d le En
portions (segments) of the same trajectory.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
1.11 Material Surface
an an n
y ha
le
liv or ec
.A
d
uu
e
X Th
er
.O
The spatial description of the surface is obtained from the spatial description
of F (X (x,t)) = f (x, y, z,t) as
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
24 C HAPTER 1. D ESCRIPTION OF M OTION
rs
ee
Remark 1.11. The necessary and sufficient condition for a mobile
n
surface in space, defined implicitly by a function f (x, y, z,t) = 0, to
gi
be material (to be always constituted by the same particles) is that
t d le En
the material derivative of f (x, y, z,t) is null,
ar
d f (x,t) ∂ f
ar s
ge ro or
+v·∇f = 0 ∀x ∈ Σt ∀t .
eS m
=
ib
dt ∂t
ac
f
C d P cs
The condition is necessary because, if the surface is a material sur-
b
i
face, its material description will not depend on time (F ≡ F (X))
an an n
and, therefore, its spatial description will have a null material deriva-
y ha
tive. The condition of sufficiency is based on the fact that, if the ma-
le
liv or ec
scription will not depend on time (F ≡ F (X)) and, therefore, the set
.A
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Material Surface 25
Example 1.8 – In ocean waves theory, the condition that the free surface
of the fluid in contact with the atmosphere is a material surface is imposed.
This restriction implies that the free surface is always composed of the same
particles, which is a reasonable hypothesis (especially in deep waters). De-
termine how this condition is stated in terms of the velocity field of the fluid.
Solution
Assuming that z = η (x, y,t) defines the elevation of the sea surface with re-
rs
spect to a reference level, the free surface of the water will be given by
ee
f (x, y, z,t) ≡ z − η (x, y,t) = 0 .
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
d
uu
df ∂f ∂f ∂η
e
= + v · ∇ f where =− and
X Th
er
dt ∂t ∂t ∂t
tin
∂f ∂f ∂f T
∂f ∂f ∂f
on
.O
not
v · ∇ f ≡ [vx , vy , vz ] , , = vx + vy + vz .
∂x ∂y ∂z ∂x ∂y ∂z
C
Then,
df ∂f ∂η ∂η ∂η
= +v·∇f = − − vx − vy + vz = 0
dt ∂t ∂t ∂x ∂y
and, isolating vz leads to
∂η ∂η ∂η
vz = + vx + vy .
∂t ∂x ∂y
Therefore, the material surface condition results in a condition on the vertical
component of the velocity field.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
26 C HAPTER 1. D ESCRIPTION OF M OTION
Σ := {x | f (x, y, z) = 0} . (1.37)
rs
Obviously, a control surface is occupied by the different particles of the contin-
ee
uous medium along time (see Figure 1.10).
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
m
er
tin
.O
material surface.
C
Vt := {x | f (x,t) ≤ 0} , (1.39)
15 It is assumed that function F (X) is defined such that F (X) < 0 corresponds to points in
the interior of V0 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Control Volume 27
where F (X) = f (x (X,t) ,t) is the function that describes the material surface
that encloses the volume.
rs
ticles.
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
d
uu
e
er
tin
on
.O
16 It is assumed that function f (x) is defined such that f (x) < 0 corresponds to points in the
interior of V .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
28 C HAPTER 1. D ESCRIPTION OF M OTION
rs
Figure 1.12: A control volume is occupied by different particles along time.
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 29
P ROBLEMS
Problem 1.1 – Justify whether the following statements are true or false.
a) If the velocity field is stationary, the acceleration field is also stationary.
b) If the velocity field is uniform, the acceleration field is always null.
rs
ee
Solution
n
gi
a) A stationary velocity field implies that the spatial description of velocity does
t d le En
not depend on time,
∂ v (x,t)
ar
= 0 =⇒ v (x) .
ar s
ge ro or
∂t
eS m
ib
ac
The acceleration is the material derivative of the velocity, therefore
f
C d P cs
∂ v (x,t)
b
i
a (x,t) = + v (x,t) · ∇v (x,t) = v (x) · ∇v (x) .
an an n
∂t
y ha
le
The resulting expression does not depend on time. Thus, the statement is true.
liv or ec
M
.A
b) A uniform velocity field implies that the spatial description of velocity does
not depend on the spatial coordinates,
m
d
uu
v (x,t) =⇒ v (t) .
e
X Th
er
tin
.O
∂ v (x,t) ∂ v (t)
a (x,t) = + v (x,t) · ∇v (x,t) = ,
C
∂t ∂t
©
where the expression used for the gradient of the velocity field is
∂ vi (t)
[∇v (t)]i j = =0.
∂xj
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
30 C HAPTER 1. D ESCRIPTION OF M OTION
Solution
Since the velocity field is given in its spatial expression and the acceleration is
requested for a point x∗ = (1, 1, 1)T , the equation of motion is not needed. One
rs
can simply apply
ee
dv (x,t) ∂ v (x,t)
n
a (x,t) = = + v (x,t) · ∇v (x,t) ,
gi
dt ∂t
t d le En
where
∂ v not
ar
T
z e t − e−t ,
≡ 0,
ar s
0 and
ge ro or
eS m
ib
∂t
ac
f
C d P cs
∂
b
i
∂x 1 0 0
an an n
y ha
z e t + e−t ,
not ∂
∇v ≡ x−z , 0 =
0 0 , such that
0
le
∂y
liv or ec
−1 (e t + e−t ) 0
M
.A
∂
∂z
m
not
v · ∇v ≡ [x − z , 0, 0]T .
uu
e
X Th
er
tin
.O
T
a ≡ x − z , z e t − e−t , 0
not
C
and, for the given point at the given instant of time, the acceleration is
T
a (x = x∗ , t = 2) ≡ 0 , e2 − e−2 ,
not
0 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 31
rs
ee
c) An observer located in point (1, 1, 1) that measures the accelerations as the
difference between velocities at this point per unit of time.
n
gi
t d le En
Solution
ar
ar s
ge ro or
a) The spatial description of the acceleration in point x∗ = (1, 1, 1) must be
eS m
ib
ac
obtained,
f
∂ V (X (x∗ ,t) ,t)
C d P cs
a (x = x∗ , t) = A (X (x∗ ,t) , t) =
b .
i
∂t
an an n
y ha
0
M
.A
∂ x (X,t)
V (X,t) ≡ 1 ((Y + Z) e t − (Y − Z) e−t ) .
m
not
V (X,t) = =⇒
d
∂t
uu
2
e
X Th
1
er
t −t
tin
((Y + Z) e + (Y − Z) e )
2
on
.O
©
0
∂ V (X,t) not
1
A (X,t) = ≡ ((Y + Z) e t + (Y − Z) e−t ) .
∂t 2
1
((Y + Z) e t − (Y − Z) e−t )
2
Careful observation of the expression obtained reveals that
1
(Y + Z) e t + (Y − Z) e−t = y
Ay = and
2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
32 C HAPTER 1. D ESCRIPTION OF M OTION
1
(Y + Z) e t − (Y − Z) e−t = z .
Az =
2
Therefore, the spatial description of the acceleration field is
not
a (x,t) ≡ [0 , y, z]T
and, for x = x∗ ,
a (x∗ ,t) ≡ [0 , 1 , 1]T .
not
rs
NOTE: In case one does not realize that Ay = y and Az = z, this same result can
ee
be obtained by replacing into the material expression of the acceleration field
the inverse equation of motion as follows.
n
gi
) (
y + z = (Y + Z) e t Y + Z = (y + z) e−t
t d le En
=⇒
y − z = (Y − Z) e−t Y − Z = (y − z) e t
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
X =x
b
i
an an n
1
Y = ((y + z) e−t + (y − z) e t )
y ha
2
le
1
liv or ec
Z = ((y + z) e−t − (y − z) e t )
2
M
.A
m
er
tin
T
A (X∗ ,t) ≡ 0 , e t , e t
not
on
.O
.
C
c) The difference between the spatial velocities per unit of time must be ob-
tained, for point x∗ = (1, 1, 1),
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 33
Problem 1.4 – Given the spatial description of the velocity field in Cartesian
coordinates,
rs
v ≡ [x , y , zϕ (t)]T
not
ee
and the surface
n
gi
n 2
o
Σt := x | F (x, y, z,t) = e−2t x2 + y2 + z2 e−t −C = 0 ,
t d le En
ar
ar s
ge ro or
where C 6= 0 is a constant, determine ϕ (t) considering that the particles on this
eS m
ib
surface are always the same.
ac
f
C d P cs
b
i
an an n
Solution
y ha
The function F defines the material surface Σt := {x | F (x, y, z,t) = 0}. The nec-
le
liv or ec
.A
dF ∂F
= + v · ∇F = 0 ∀x ∈ Σt ∀t ,
m
dt ∂t
d
uu
e
where
X Th
∂F
er
tin
2
= −2e−2t x2 + y2 − 2tz2 e−t ,
∂t h
on
2 T
.O
i
not
∇F ≡ 2xe−2t , 2ye−2t , 2ze−t , and
C
2
v · ∇F = 2x2 e−2t + 2y2 e−2t + 2z2 e−t ϕ (t) .
©
2 C − e−2t x2 + y2 (ϕ (t) − t) = 0
∀x ∀t .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
34 C HAPTER 1. D ESCRIPTION OF M OTION
ϕ (t) = t .
rs
not t
v (x,t) ≡ ze , , vz
1+t
n ee
and the surface ϕ (x,t) = x − z (1 + t) e t + k = 0 (where k is a constant), which
gi
is known to be a material surface, determine:
t d le En
a) The equation of the trajectory in canonical form and the equation of the
ar
ar s
streamlines.
ge ro or
eS m
ib
ac
f
b) The equation of the streakline and the position of its initial and final points
C d P cs
if the spill point is x∗ and the spill period is t ∈ [t1 ,t2 ].
b
i
an an n
y ha
le
Solution
liv or ec
.A
the velocity field must be completed. To find vz , the information given about
surface ϕ is used. The necessary and sufficient condition for this surface to be a
m
material surface is
uu
e
X Th
dϕ
er
∂ϕ
tin
= + v · ∇ϕ = 0 ∀x ∈ Σt ∀t ,
dt ∂t
on
.O
∂ϕ
C
∇ϕ ≡ [1 , 0 , −e t (1 + t)]T
not
where = −z (e t + e t (1 + t)) ,
©
∂t
and v · ∇ϕ = ze t − vz e t (1 + t) .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 35
Now, this field must be integrated to obtain the equation of the trajectory since
dx/dt = v (x,t). Applying the equality for each component and particularizing
for the velocity field determined yields
dx dy y dz
= ze t , = and = −z .
dt dt 1+t dt
Note that the x-component depends on the z-coordinate. Then, the z-coordinate
must be determined first,
dz
= −z z = C1 e−t .
rs
=⇒
dt
ee
Replacing the expression found for z into the x-component and integrating the
n
expression results in
gi
t d le En
dx
= C1 e−t e t = C1 =⇒ x = C1t +C2 .
ar
dt
ar s
ge ro or
eS m
ib
Finally, the y-component is
ac
f
C d P cs
dy y
b
i
= =⇒ y = C3 (1 + t) .
an an n
dt 1+t
y ha
.A
x (0) = C2 = X
y (0) = C3 = Y
m
uu
z (0) = C1 = Z
e
X Th
er
tin
.O
x = X + Zt
C
y = Y (1 + t) .
©
z = Ze−t
The equation of the streamlines is found by integrating the velocity field with
respect to λ , that is, dx (λ )/dλ = v (x (λ ) ,t). As in the case of the equation of
the trajectory, the z-component must be determined before the x-component,
dz
= −z =⇒ z = C1 e−λ .
dλ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
36 C HAPTER 1. D ESCRIPTION OF M OTION
rs
ee
x = −C1 e (t−λ ) +C2
n
λ
y = C3 e 1+t .
gi
e−λ
t d le En
z = C1
ar
ar s
ge ro or
eS m
ib
b) To obtain the equation of the streakline it is enough to take the equation of
ac
f
motion and impose x∗ = x (X, τ), where τ is a time belonging to the spill period.
C d P cs
b
i
an an n
∗
x = X + Zτ
y ha
y∗ = Y (1 + τ)
le
liv or ec
∗
z = Ze−τ
M
.A
d
uu
e
X = x∗ − Zτ = x∗ − z∗ τeτ
X Th
er
y∗
tin
Y=
1+τ
on
.O
Z = z∗ eτ
C
Replacing these into the equation of motion results in the equation of the streak-
line,
x = x∗ − z∗ (τ − t) eτ
1+t
y = y∗ .
1+τ
z = z∗ e(τ−t)
Consider the physical concept of the streakline as the color line that would be
observed in the medium if a tracer fluid were injected at the spill point through-
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 37
out the spill period. Then, for each time t, the streakline can be visualized in
terms of the parameter τ, which gives the position in space of the colored parti-
cles. It is verified that, as expected, x = x∗ for t = τ, since it corresponds to the
time in which the streakline is crossing the spill point. Now, the streakline must
be delimited for each time t.
rs
τ = t1 while the last one is the one crossing the spill point at τ = t.
ee
x = x∗ − z∗ (t1 − t) e t1
∗
n
x = x
gi
1 + t
Initial point: y = y∗ Final point: y = y∗
t d le En
1 + t1
z = z∗
z = z∗ e (t1 −t)
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
ii) t ≥ t2
C
The first colored point in the streakline is the same as in the previous case,
©
τ = t1 , but the last point is now τ = t2 . The streakline has now “moved away”
from the spill point.
x = x∗ − z∗ (t1 − t) e t1 x = x∗ − z∗ (t2 − t) e t2
1+t 1+t
Initial point: y = y∗ Final point: y = y∗
1 + t1
1 + t2
z = z∗ e (t2 −t)
z = z∗ e (t1 −t)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
38 C HAPTER 1. D ESCRIPTION OF M OTION
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 39
E XERCISES
rs
can cross each other at one or more points.
n ee
1.2 – Given the following velocity field in material description
gi
t d le En
not T
v ≡ Ae At X1 , BtX1 , CX3 ,
ar
ar s
ge ro or
with A, B and C constants, obtain its spatial description and the conditions A, B
eS m
ib
and C must fulfill for the motion to be feasible for 0 < t < ∞.
ac
f
C d P cs
b
1.3 – Tracer fluid is injected at point (1, 1, 1) of the interior of a fluid from time
i
an an n
le
z = C3 e2λt
liv or ec
x = C1 eλt , y = C2 eλt ,
M
.A
determine the equation of the streakline, indicating its initial and final points for
t = 5.
m
d
uu
e
er
tin
T
v ≡ ye−t , ze t , 0 .
not
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
40 C HAPTER 1. D ESCRIPTION OF M OTION
rs
n ee
gi
t d le En
ar
ar s
1.6 – A certain motion is defined by the velocity field
ge ro or
eS m
ib
ac
vx = ye−t ; vy = y f ; vz = 0 .
C d P cs
b
i
Determine:
an an n
y ha
a) The equation of the trajectory in canonical form and the equation of the
le
streamlines.
liv or ec
.A
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 2
Strain
rs
n ee
gi
2.1 Introduction
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Definition 2.1. In the broader context, the concept of deformation no
C d P cs
b
longer refers to the study of the absolute motion of the particles as
i
seen in Chapter 1, but to the study of the relative motion, with respect
an an n
y ha
.A
d
uu
e
0
er
.O
relative positions with respect to this particle in the reference and present times
given by dX and dx, respectively. The equation of motion is given by
C
(
not
x = ϕ (X,t) = x (X,t)
not . (2.1)
xi = ϕi (X1 , X2 , X3 ,t) = xi (X1 , X2 , X3 ,t) i ∈ {1, 2, 3}
41
42 C HAPTER 2. S TRAIN
rs
Figure 2.1: Continuous medium in motion.
n ee
Equation (2.2) defines the material deformation gradient tensor F (X,t) 1 .
gi
t d le En
not
F = x⊗∇
ar
ar s
Material deformation
ge ro or
eS m
ib
∂ xi (2.3)
gradient tensor Fi j = i, j ∈ {1, 2, 3}
ac
∂ Xj f
C d P cs
b
i
an an n
le
∂ x1 ∂ x1 ∂ x1
liv or ec
∂X ∂ X2 ∂ X3
x
M
1
.A
h i 1 ∂ ∂ ∂
∂ x2 ∂ x2 ∂ x2
[F] = x ⊗ ∇ = x2 ∂ X1 , ∂ X2 , ∂ X3 =
. (2.4)
m
∂ X1 ∂ X2 ∂ X3
d
uu
x3 | }
e
{z ∂x ∂ x3 ∂ x3
h iT 3
X Th
er
| {z }
tin
[x] ∇ ∂ X1 ∂ X2 ∂ X3
on
.O
C
1 Here, the symbolic form of the material Nabla operator, ∇ ≡ ∂ êi /∂ Xi , applied to the
not
expression of the open or tensor product, [a ⊗ b]i j = [a b]i j = ai b j , is considered.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Deformation Gradient Tensor 43
rs
∂ Xi (2.6)
dXi = dx j = Fi−1
j dx j i, j ∈ {1, 2, 3} .
ee
∂xj
n
gi
The tensor defined in (2.6) is named spatial deformation gradient tensor or in-
verse (material) deformation gradient tensor and is characterized by2
t d le En
ar
ar s
ge ro or
not
F−1 = X ⊗ ∇
eS m
ib
Spatial deformation
ac
gradient tensor −1 ∂ Xi f i, j ∈ {1, 2, 3}
(2.7)
C d P cs
Fi j =
b
∂xj
i
an an n
y ha
le
liv or ec
.A
(2.6) and (2.7) as F−1 , is in effect the inverse of the (material) defor-
mation gradient tensor F. The verification is immediate since3
m
d
uu
∂ xi ∂ Xk ∂ xi not
e
= = δi j =⇒ F · F−1 = 1 ,
X Th
er
∂ Xk ∂ x j ∂xj
tin
|{z} |{z}
Fik F −1
on
.O
kj
C
∂ Xi ∂ xk ∂ Xi not
F−1 · F = 1 .
©
= = δi j =⇒
∂ xk ∂ X j ∂ Xj
|{z} |{z}
Fik−1 Fk j
2 Here, the symbolic form of the spatial Nabla operator, ∇ ≡ ∂ êi /∂ xi , is considered. Note
the difference in notation between this spatial operator ∇ and the material Nabla ∇.
3 The two-index operator Delta Kronecker δ is defined as δ = 1 if i = j and δ = 0 if
ij ij ij
i 6= j. The second-order unit tensor 1 is given by [1]i j = δi j .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
44 C HAPTER 2. S TRAIN
rs
ee
Example 2.1 – At a given time, the motion of a continuous medium is defined
n
by
gi
x1 = X1 − AX3
t d le En
x2 = X2 − AX3 .
ar
ar s
ge ro or
x3 = −AX1 + AX2 + X3
eS m
ib
ac
f
Obtain the material deformation gradient tensor F (X,t) at this time. By
C d P cs
b
means of the inverse equation of motion, obtain the spatial deformation gra-
i
dient tensor F−1 (x). Using the results obtained, verify that F · F−1 = 1.
an an n
y ha
le
Solution
liv or ec
M
.A
X1 − AX3
d
uu
h iT
e
not ∂ ∂ ∂
F = x ⊗ ∇ ≡ [x] ∇ = X2 − AX3 ∂ X1 , ∂ X2 , ∂ X3
X Th
er
tin
−AX1 + AX2 + X3
on
.O
C
1 0 −A
©
not
F≡ 0 1 −A .
−A A 1
The inverse equation of motion is obtained directly from the algebraic inver-
sion of the equation of motion,
X1 = 1 + A2 x1 − A2 x2 + Ax3
not 2 2
X (x,t) ≡ X2 = A x1 + 1 − A x2 + Ax3 .
X3 = Ax1 − Ax2 + x3
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Displacements 45
rs
A −A 1
ee
Finally, it is verified that
n
gi
1 + A2 −A2
t d le En
1 0 −A A 1 0 0
F · F−1 ≡ 0
not not
1 −A A2 1 − A2 A = 0 1 0 ≡ 1 .
ar
ar s
ge ro or
eS m
ib
−A A 1 A −A 1 0 0 1
ac
f
C d P cs
b
i
an an n
y ha
2.3 Displacements
le
liv or ec
M
.A
particle.
X Th
er
tin
on
.O
joins the points in space P (initial position) and P (position at the present time t)
©
of the particle (see Figure 2.2). The displacement of all the particles in the con-
tinuous medium defines a displacement vector field which, as all properties of
the continuous medium, can be described in material form U (X,t) or in spatial
form u (x,t) as follows.
(
U (X,t) = x (X,t) − X
(2.9)
Ui (X,t) = xi (X,t) − Xi i ∈ {1, 2, 3}
(
u (x,t) = x − X (x,t)
(2.10)
ui (x,t) = xi − Xi (x,t) i ∈ {1, 2, 3}
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
46 C HAPTER 2. S TRAIN
rs
n ee
Figure 2.2: Displacement of a particle.
gi
t d le En
ar
2.3.1 Material and Spatial Displacement Gradient Tensors
ar s
ge ro or
eS m
ib
ac
Differentiation with respect to the material coordinates of the displacement vec-
f
C d P cs
tor Ui defined in (2.9) results in
b
i
an an n
∂Ui ∂ xi ∂ Xi de f
= − = Fi j − δi j = Ji j ,
y ha
(2.11)
∂ Xj ∂ Xj ∂ Xj
le
liv or ec
|{z} |{z}
Fi j δi j
M
.A
d
uu
e
de f
J (X,t) = U (X,t) ⊗ ∇ = F − 1
X Th
er
Material displacement
tin
∂ Xj
.O
C
©
U = J · dX
∂Ui (2.13)
dUi =
dX j = Ji j dX j i, j ∈ {1, 2, 3}
∂ Xj
Similarly, differentiation with respect to the spatial coordinates of the expres-
sion of ui given in (2.10) yields
∂ ui ∂ xi ∂ Xi de f
= − = δi j − Fi−1
j = ji j , (2.14)
∂xj ∂xj ∂xj
|{z} |{z}
δi j Fi−1
j
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Strain Tensors 47
rs
∂xj
n ee
gi
2.4 Strain Tensors
t d le En
Consider now a particle of the continuous medium that occupies the point in
ar
ar s
space P in the material configuration, and another particle Q√in its differen-
ge ro or
eS m
ib
tial neighborhood separated a segment dX (with length dS = dX · dX) from
ac
√
f
the previous paticle, being dx (with length ds = dx · dx) its counterpart in
C d P cs
b
the present configuration (see Figure 2.3). Both differential vectors are related
i
an an n
through the deformation gradient tensor F (X,t) by means of equations (2.2) and
y ha
(2.6),
le
liv or ec
dx = F · dX and dX = F−1 · dx ,
M
.A
(2.17)
dxi = Fi j dX j and dXi = F −1 dx j i, j ∈ {1, 2, 3} .
m
ij
d
uu
e
Then,
X Th
er
tin
(ds)2 = dx · dx not not
≡ [dx]T [dx] = [F · dX]T [F · dX] ≡ dX · FT · F · dX
on
.O
(2.18)
(ds)2 = dxk dxk = Fki dXi Fk j dX j = dXi Fki Fk j dX j = dXi F T Fk j dX j
ik
C
or, alternatively4 ,
T −1
2 not T −1
(dS) = dX · dX ≡ [dX] [dX] = F · dx F · dx =
not
≡ dx · F−T · F−1 · dx ,
(2.19)
(dS)2 = dXk dXk = Fki−1 dxi Fk−1 −1 −1
j dx j = dxi Fki Fk j dx j =
= dxi Fik−T Fk−1
j dx j .
T not
4 The convention (•)−1 = (•)−T is used.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
48 C HAPTER 2. S TRAIN
rs
n ee
Figure 2.3: Differential segments in a continuous medium.
gi
t d le En
ar
2.4.1 Material Strain Tensor (Green-Lagrange Strain Tensor)
ar s
ge ro or
eS m
ib
Subtracting expressions (2.18) and (2.19) results in
ac
f
C d P cs
b
(ds)2 − (dS)2 = dX · FT · F · dX − dX · dX =
i
an an n
= dX · FT · F · dX − dX · 1 · dX =
y ha
le
= dX · FT · F − 1 · dX = 2 dX · E · dX ,
(2.20)
liv or ec
| {z }
M
.A
de f
= 2E
m
tensor as follows.
X Th
er
tin
E (X,t) = 1 FT · F − 1
Material
on
.O
(Green-Lagrange) 2 (2.21)
1
strain tensor Ei j (X,t) = Fki Fk j − δi j i, j ∈ {1, 2, 3}
C
2
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Strain Tensors 49
rs
which implicitly defines the spatial strain tensor or Almansi strain tensor as
ee
follows.
n
e (x,t) = 1 1 − F−T · F−1
gi
Spatial
t d le En
(Almansi) 2 (2.23)
1
strain tensor ei j (x,t) =
δi j − Fki−1 Fk−1 i, j ∈ {1, 2, 3}
ar
2 j
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
Remark 2.4. The spatial strain tensor e is symmetric. Proof is ob-
an an n
le
T = 1 1 − F−T · F−1 T = 1 1T − F−1 T · F−T T =
liv or ec
e
2 2
M
.A
1 −T −1
= 1−F ·F =e,
2
m
d
i, j ∈ {1, 2, 3} .
e = e
uu
ij ji
e
X Th
er
tin
on
.O
Example 2.2 – Obtain the material and spatial strain tensors for the motion
C
in Example 2.1.
©
Solution
The material strain tensor is
1 0 −A 1 0 −A 1 0 0
1 T not 1
E (X,t) = F · F − 1 ≡ 0 1 A 0 1 −A − 0 1 0 =
2 2
−A −A 1 −A A 1 001
2
A −A2 −2A
1
= −A2 A2 0
2
−2A 0 2A2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
50 C HAPTER 2. S TRAIN
rs
= A2 + 2A4 A2 − 2A4 2A3 .
2
ee
−2A − 2A 3 2A 3 −2A 2
n
gi
Observe that E 6= e.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Remark 2.5. The material strain tensor E and the spatial strain ten-
b
sor e are different tensors. They are not the material and spatial de-
i
an an n
le
(ds)2 − (dS)2 = 2dX · E · dX = 2dx · e · dx ,
liv or ec
M
.A
d
uu
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Variation of Distances: Stretch and Unit Elongation 51
1 1
1 + JT · (1 + J) − 1 = J + JT + JT · J
E=
2 2 (2.24)
1 ∂Ui ∂U j ∂Uk ∂Uk
Ei j = + + i, j ∈ {1, 2, 3}
2 ∂ X j ∂ Xi ∂ Xi ∂ X j
rs
n ee
1 1
gi
1 − 1 − jT · (1 − j) = j + jT − jT · j
e=
2 2
t d le En
(2.25)
1 ∂ ui ∂ u j ∂ uk ∂ uk
ei j = + − i, j ∈ {1, 2, 3}
ar
ar s
2 ∂ x j ∂ xi ∂ xi ∂ x j
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
le
Consider now a particle P in the reference configuration and another particle
liv or ec
.A
sponding positions in the present configuration are given by the points in space
0 0
P and Q such that the distance between the two particles in the reference con-
m
figuration, dS, is transformed into ds at the present time. The vectors T and t are
uu
e
0 0
the unit vectors in the directions PQ and P Q , respectively.
X Th
er
tin
on
.O
0 0
is the length of the deformed differential segment P Q per unit of
length of the original differential segment PQ.
0 0
de f PQ ds
Stretch = λT = λt = = (0 < λ < ∞) . (2.26)
PQ dS
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
52 C HAPTER 2. S TRAIN
rs
n ee
Figure 2.4: Differential segments and unit vectors in a continuous medium.
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Definition 2.4. The unit elongation, elongation ratio or extension of
C d P cs
0
a material point P (or a spatial point P ) in the material direction T
b
i
5
(or spatial direction t ) is the increment of length of the deformed
an an n
0 0
y ha
.A
m
∆ PQ ds − dS
er
de f
tin
.O
C
Equations (2.26) and (2.27) allow immediately relating the values of the unit
©
elongation and the stretch for a same point and direction as follows.
ds − dS ds
ε= = −1 = λ − 1 (⇒ −1 < ε < ∞) (2.28)
dS dS
|{z}
λ
5 Often, the subindices (•)T and (•)t will be dropped when referring to stretches or unit
elongations. However, one must bear in mind that both stretches and unit elongations are
always associated with a particular direction.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Variation of Distances: Stretch and Unit Elongation 53
rs
P and Q has shortened with the deformation of the medium.
n ee
gi
t d le En
2.5.1 Stretches, Unit Elongations and Strain Tensors
ar
ar s
ge ro or
Consider equations (2.21) and (2.22) as well as the geometric expressions
eS m
ib
ac
dX = T dS and dx = t ds (see Figure 2.4). Then,
f
C d P cs
b
(ds)2 − (dS)2 = 2 dX · E · dX = 2 (dS)2 T · E · T
i
an an n
y ha
|{z} |{z}
dS T dS T (2.29)
le
(ds)2 − (dS)2 = 2 dx · e · dx = 2 (ds)2 t · e · t
liv or ec
M
.A
|{z} |{z}
ds t ds t
m
√
er
tin
ds 2 λ = 1 + 2T · E · T
2
−1 = λ −1 = 2 T·E·T ⇒ √ (2.30)
dS
on
.O
ε = λ − 1 = 1 + 2T · E · T − 1
| {z }
C
λ
©
1
λ=√
dS 2
2
1 1 − 2t · e · t
1− = 1− = 2 t·e·t ⇒ 1 (2.31)
ds λ ε = λ −1 = √ −1
| {z } 1 − 2t · e · t
1/λ
These equations allow calculating the unit elongation and stretch for a given
direction (in material description, T, or in spatial description, t ).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
54 C HAPTER 2. S TRAIN
Remark 2.7. The material and spatial strain tensors, E (X,t) and
e (x,t), contain information on the stretches (and unit elongations)
for any direction in a differential neighborhood of a given particle,
as evidenced by (2.30) and (2.31).
rs
ee
0 0 −tetz
n
not
e (x,t) ≡ 0 0 0 .
gi
−tetz t (2etz − et )
t d le En
0
ar
ar s
Calculate the length, at time t = 0, of the segment that at time t = 2 is recti-
ge ro or
eS m
ib
linear and joins points a ≡ (0, 0, 0) and b ≡ (1, 1, 1).
ac
f
C d P cs
Solution
b
i
an an n
y ha
linear and the positions of its extremes A and B (see figure below) are not
M
.A
known. To determine its length, (2.31) is applied for a unit vector in the di-
rection of the spatial configuration t,
m
d
uu
1 ds 1
e
λ=√ = =⇒ dS = ds .
X Th
1 − 2 t · e · t dS
er
λ
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Variation of Angles 55
not √
To obtain the stretch in the direction t ≡ [1, 1, 1]T / 3, the expression t · e · t
is computed first as
0 0 −tetz 1
not 1 1 1 t
t·e·t ≡ √ [1, 1, 1] 0 0 0 1 √ = − te .
3 3 3
−tetz 0 t (2etz − et ) 1
rs
1 1 3
λ=q =⇒ =q =√ .
λ
ee
1 + 32 tet t=2
1+ 4 2 3 + 4e2
3e
n
gi
The length at time t = 0 of the segment AB is
t d le En
1√
Z B Z b Z b
1 1 1
ar
ar s
lAB = dS = ds = ds = lab = 3
ge ro or
eS m
ib
A a λ λ a λ λ
ac
f | {z }
lab
C d P cs
b
i
an an n
and replacing the expression obtained above for the stretch at time t = 2
y ha
finally results in
le
p
lAB = 3 + 4e2 .
liv or ec
M
.A
m
d
uu
er
tin
Consider a particle P and two additional particles Q and R, belonging to the dif-
ferential neighborhood of P in the material configuration (see Figure 2.5), and
on
.O
0 0 0
the same particles occupying the spatial positions P , Q and R . The relationship
C
between the angles that form the corresponding differential segments in the ref-
©
and using the definitions of the unit vectors T(1) , T(2) , t(1) and t(2) that establish
the corresponding directions in Figure 2.5,
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
56 C HAPTER 2. S TRAIN
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Figure 2.5: Angles between particles in a continuous medium.
ac
f
C d P cs
b
i
dX(1) = dS(1) T(1) dx(1) = ds(1) t(1) ,
an an n
=⇒ (2.33)
y ha
.A
1
d
dS(1) = ds(1) ,
ds(1) = λ (1) dS(1)
uu
e
=⇒ λ (1) (2.34)
X Th
er
dS(2) = 1
ds(2) .
λ (2)
on
.O
Expanding now the scalar product6 of the vectors dx(1) and dx(2) ,
C
h iT h i
not
dS(1) dS(2) cos θ = dx(1) dx(2) cos θ = dx(1) · dx(2) ≡ dx(1) dx(2) =
h iT h i
not
= F · dX(1) F · dX(2) ≡ dX(1) · FT · F · dX(2) = dX(1) · (2E + 1) · dX(2)
1 1
= dS(1) T(1) · (2E + 1) · T(2) dS(2) = (1) ds(1) T(1) · (2E + 1) · T(2) (2) ds(2) =
λ λ
(1) (2) 1 1 (1) (2)
= ds ds T · (2E + 1) · T ,
λ (1) λ (2)
(2.35)
6 The scalar product of two vectors a and b is defined in terms of the angle between them, θ ,
as a · b = |a| · |b| cos θ .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Physical Interpretation of the Strain Tensors 57
rs
ee
In an analogous way, operating on the reference configuration, the angle Θ
between the differential segments dX(1) and dX(2) (in terms of t(1) , t(2) and e )
n
gi
is obtained,
t d le En
t(1) · (1 − 2e) · t(2)
ar
cosΘ = p . (2.38)
ar s
p
ge ro or
1 − 2t(1) · e · t(1) 1 − 2t(2) · e · t(2)
eS m
ib
ac
f
C d P cs
b
i
Remark 2.8. Similarly to the discussion in Remark 2.7, the material
an an n
and spatial strain tensors, E (X,t) and e (x,t), also contain informa-
y ha
process. These facts will be the basis for providing a physical inter-
.A
d
uu
e
X Th
er
tin
.O
C
Consider a segment PQ, oriented parallel to the X1 -axis in the reference config-
uration (see Figure 2.6). Before the deformation takes place, PQ has a known
length dS = dX.
0 0
The length of P Q is sought. To this aim, consider the material strain tensor
E given by its components,
EXX EXY EXZ E11 E12 E13
not
E ≡ EXY EYY EY Z = E12 E22 E23 . (2.39)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
58 C HAPTER 2. S TRAIN
rs
ee
Figure 2.6: Differential segment in the reference configuration.
n
gi
t d le En
Consequently,
ar
ar s
ge ro or
eS m
ib
E11 E12 E13 1
ac
not T
T · E · T ≡ [T] [E] [T] = [1, 0, 0] E12 E22 E23 f 0 = E11 . (2.40)
C d P cs
b
E13 E23 E33 0
i
an an n
y ha
.A
√ √ √
d
uu
√ √ √
X Th
er
tin
.O
C
Remark 2.9. The components EXX , EYY and EZZ (or E11 , E22 and
E33 ) of the main diagonal of tensor E (denoted longitudinal strains)
contain the information on stretch and unit elongations of the dif-
ferential segments that were initially (in the reference configuration)
oriented in the directions X, Y and Z, respectively.
• If EXX = 0 ⇒ εX = 0 : No unit elongation in direction X.
• If EYY = 0 ⇒ εY = 0 : No unit elongation in direction Y .
• If EZZ = 0 ⇒ εZ = 0 : No unit elongation in direction Z.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Physical Interpretation of the Strain Tensors 59
rs
ee
Figure 2.7: Angles between differential segments in the reference and present configu-
n
rations.
gi
t d le En
ar
ar s
Consider now the angle between segments PQ (parallel to the X1 -axis) and PR
ge ro or
eS m
ib
(parallel to the X2 -axis), where Q and R are two particles in the differential neigh-
ac
f
0 0 0
borhood of P in the material configuration and P , Q and R are the respective
C d P cs
b
positions in the spatial configuration (see Figure 2.7). If the angle (Θ = π/2)
i
an an n
the present configuration can be determined using (2.37) and taking into ac-
le
count their orthogonality ( T(1) · T(2) = 0 ) and the equalities T(1) · E · T(1) = E11 ,
liv or ec
.A
m
cos θ = p
uu
p
1 + 2T(1) · E · T(1) 1 + 2T(2) · E · T(2)
e
(2.42)
X Th
er
tin
2E12
=√ √ ,
1 + 2E11 1 + 2E22
on
.O
π 2EXY
θ ≡ θxy = − arcsin √ √ . (2.43)
2 1 + 2EXX 1 + 2EYY
The increment of the final angle with respect to its initial value results in
2EXY
∆ΘXY = θxy − ΘXY = − arcsin √ √ . (2.44)
|{z} 1 + 2EXX 1 + 2EYY
π/2
Analogous results are obtained starting from pairs of segments that are ori-
ented in different combinations of the coordinate axes, resulting in
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
60 C HAPTER 2. S TRAIN
2EXY
∆ΘXY = − arcsin √ √
1 + 2EXX 1 + 2EYY
2EXZ
∆ΘXZ = − arcsin √ √ . (2.45)
1 + 2EXX 1 + 2EZZ
2EY Z
∆ΘY Z = − arcsin √ √
1 + 2EYY 1 + 2EZZ
rs
ee
Remark 2.10. The components EXY , EXZ and EY Z (or E12 , E13 and
n
E23 ) of the tensor E (denoted angular strains) contain the informa-
gi
tion on variation of the angles between the differential segments that
t d le En
were initially (in the reference configuration) oriented in the direc-
tions X, Y and Z, respectively.
ar
ar s
ge ro or
eS m
ib
• If EXY = 0 : The deformation does not produce a variation in the
ac
f
angle between the two segments initially oriented in the direc-
C d P cs
tions X and Y .
b
i
• If EXZ = 0 : The deformation does not produce a variation in the
an an n
y ha
.A
d
uu
e
X Th
er
tin
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Physical Interpretation of the Strain Tensors 61
rs
n ee
Figure 2.8: Physical interpretation of the material strain tensor.
gi
t d le En
ar
ented in the direction of the coordinate axes in the present configuration,
ar s
ge ro or
eS m
ib
ac
f
C d P cs
1 1 1
λ1 = √ =√ ⇒
b εx = √ −1
i
1 − 2e11 1 − 2exx 1 − 2exx
an an n
y ha
1 1 1
λ2 = √
le
=p ⇒ εy = p −1 , (2.47)
liv or ec
.A
1 1 1
λ3 = √ =√ ⇒ εz = √ −1
m
er
while the components outside the main diagonal (angular strains) contain infor-
tin
mation on the variation of the angles between the differential segments oriented
on
.O
π 2exy
∆ θxy = −ΘXY = − arcsin √ p
2 1 − 2exx 1 − 2eyy
π 2exz . (2.48)
∆ θxz = −ΘXZ = − arcsin √ √
2 1 − 2exx 1 − 2ezz
π 2eyz
∆ θyz = −ΘY Z = − arcsin p √
2 1 − 2eyy 1 − 2ezz
Figure 2.9 summarizes the physical interpretation of the components of the spa-
tial strain tensor.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
62 C HAPTER 2. S TRAIN
rs
Figure 2.9: Physical interpretation of the spatial strain tensor.
n ee
gi
2.8 Polar Decomposition
t d le En
The polar decomposition theorem of tensor analysis establishes that, given a
ar
ar s
second-order tensor F such that |F| > 0, there exist an orthogonal tensor Q 7 and
ge ro or
eS m
ib
two symmetric tensors U and V such that8
ac
f
C d P cs
not √
b
U = FT · F
i
an an n
not √
y ha
V = F·F T =⇒ F = Q · U = V · Q . (2.49)
le
liv or ec
Q = F · U−1 = V−1 · F
M
.A
This decomposition is unique for each tensor F and is denominated left polar
m
Considering now the deformation gradient tensor and the fundamental re-
er
tin
.O
stretching
C
z }| {
rotation
©
z }| {
dx = F · dX = (V · Q) · dX = V · (Q · dX) (2.50)
not
F (•) ≡ stretching ◦ rotation (•)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Polar Decomposition 63
rotation
z }| {
stretching
z }| {
dx = F · dX = (Q · U) · dX = Q · (U · dX) (2.51)
not
F (•) ≡ rotation ◦ stretching (•)
rs
rotation tensor and the mapping y = Q · x is denominated rotation.
ee
A rotation has the following properties:
n
• When applied on any vector x, the result is another vector
gi
y = Q · x with the same modulus,
t d le En
kyk2 = y·y ≡ [y]T ·[y] = [Q · x]T ·[Q · x] ≡ x·QT · Q ·x = x·x = kxk2 .
not not
ar
ar s
ge ro or
| {z }
eS m
ib
1
ac
f
• The result of multiplying (mapping) the orthogonal tensor Q to
C d P cs
b
two vectors x(1) and x(2) with the same origin and that form an
i
an an n
.A
=
=
= cos α .
y(1)
y(2)
y(1)
y(2)
x(1)
x(2)
m
d
uu
e
er
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
64 C HAPTER 2. S TRAIN
rs
n ee
gi
t d le En
Figure 2.10: Polar decomposition.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
.A
by U = V = 1 and Q = F.
uu
e
X Th
er
tin
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Volume Variation 65
tion are dx(1) = F · dX(1) , dx(2) = F · dX(2) and dx(3) = F · dX(3) . Obviously, the
relations (
dx(i) = F · dX(i)
(i) (i) (2.52)
dx j = Fjk dXk i, j, k ∈ {1, 2, 3}
are satisfied. Then, the volumes10 associated with a particle in both configura-
tions can be written as
(1) (1) (1)
dX1 dX2 dX3
(2) (2) (2)
dV0 = dX(1) × dX(2) · dX(3) = det
rs
dX1 dX2 dX3 = |M| ,
ee
(3) (3) (3)
dX1 dX2 dX3
n
| {z }
gi
[M]
t d le En
(1) (1) (1)
ar
dx1 dx2 dx3
ar s
ge ro or
eS m
ib
(2) (2) (2)
dVt = dx(1) × dx(2) · dx(3) = det
dx1 dx2 dx3 = |m| , (2.53)
ac
f
C d P cs
(3) (3) (3)
b
dx1 dx2 dx3
i
an an n
| {z }
y ha
[m]
le
liv or ec
(i) (i)
where Mi j = dX j and mi j = dx j . Considering these expressions,
M
.A
(i) (i)
mi j = dx j = Fjk dXk = Fjk dMik = dMik FkTj =⇒ m = M · FT
m
(2.54)
d
uu
e
er
tin
dVt = |m| = M · FT = |M| FT = |F| |M| = |F| dV0
on
.O
|{z}
dV0 =⇒ dVt = |F|t dV0
C
©
dVt = dV (x (X,t) ,t) = |F (X,t)| dV0 (X, 0) = |F|t dV0
(2.55)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
66 C HAPTER 2. S TRAIN
rs
n ee
gi
Figure 2.11: Variation of a volume differential element.
t d le En
ar
2.10 Area Variation
ar s
ge ro or
eS m
ib
ac
f
Consider an area differential dA associated with a particle P in the reference
C d P cs
configuration and its variation along time. To define this area differential, con-
b
i
sider two particles Q and R in the differential neighborhood of P, whose relative
an an n
positions with respect to this particle are dX(1) and dX(2) , respectively (see Fig-
y ha
ure 2.12). Consider also an arbitrary auxiliary particle S whose relative position
le
liv or ec
.A
differential area, dA, is defined. The module of vector dA is dA and its direction
is the same as that of the unit normal vector in the material configuration N.
m
0
e
space P and will have an area differential da associated with it which, in turn,
X Th
er
normal vector in the spatial configuration. Consider also the positions of the
0 0 0
on
.O
other particles Q , R and S and their relative position vectors dx(1) , dx(2) and
dx(3) .
C
The volumes dV0 and dVt of the corresponding parallelepipeds can be calcu-
©
lated as
dV0 = dH dA = dX(3) · N dA = dX(3) · N dA = dA · dX(3)
| {z } |{z}
dH dA (2.56)
(3)
dVt = dh da = dx · n da = dx · n da = da · dx(3)
(3)
| {z } |{z}
dh da
and, taking into account that dx(3) = F · dX(3) , as well as the expression for
change in volume (2.55), results in
da · F · dX(3) = da · dx(3) = dVt = |F| dV0 = |F| dA · dX(3) ∀dX(3) . (2.57)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Infinitesimal Strain 67
rs
ee
Figure 2.12: Variation of an area differential.
n
gi
t d le En
Comparing the first and last terms12 in (2.57) and considering that the relative
ar
ar s
position of particle S can take any value (as can, therefore, vector dX(3) ), finally
ge ro or
eS m
ib
yields
ac
f
da = |F| dA · F−1 .
C d P cs
da · F = |F| dA =⇒ (2.58)
b
i
an an n
To obtain the relation between the two area differential scalars, dA and da,
y ha
.A
d
uu
er
tin
.O
12 Here, the following tensor algebra theorem is taken into account: given two vectors a and
b, if the relation a · x = b · x is satisfied for all values of x, then a = b.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
68 C HAPTER 2. S TRAIN
rs
n ee
gi
Figure 2.13: Infinitesimal strain in the continuous medium.
t d le En
ar
ar s
ge ro or
eS m
ib
In accordance with the first hypothesis, the reference configuration Ω0 and
ac
f
the present configuration Ωt are very close together and are considered to be
C d P cs
b
indistinguishable from one another. Consequently, the material and spatial co-
i
ordinates coincide and discriminating between material and spatial descriptions
an an n
y ha
x = X+u ∼ = X U (X,t) not
= u (X,t) ≡ u (x,t)
M
.A
=⇒
xi = Xi + ui ∼
= Xi Ui (X,t) not
= ui (X,t) ≡ ui (x,t) i ∈ {1, 2, 3}
m
(2.60)
d
uu
er
tin
∂ ui
∂xj 1
∀ i, j ∈ {1, 2, 3} . (2.61)
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Infinitesimal Strain 69
1 T + JT · J ∼ 1 J + JT ,
E = J + J =
2 2
1 ∂ ui ∂ u j ∂ uk ∂ uk ∼ 1 ∂ ui ∂ u j (2.63)
Ei j = + + = + ,
2 ∂ x j ∂ xi ∂ xi ∂ x j 2 ∂ x j ∂ xi
| {z }
1
where the infinitesimal character of the second-order term (∂ uk ∂ uk /∂ x j ∂ xi ) has
been taken into account. Operating in a similar manner with the spatial strain
tensor,
rs
1 T − jT · j ∼ 1 j + jT = 1 J + JT ,
ee
e = j + j =
2 2 2
n
1 ∂ ui ∂ u j ∂ uk ∂ uk ∼ 1 ∂ ui ∂ u j (2.64)
gi
e ij = + − = + .
t d le En
2 ∂ x j ∂ x i ∂ x i ∂ x j 2 ∂ x j ∂ x i
| {z }
ar
1
ar s
ge ro or
eS m
ib
Equations (2.63) and (2.64) allow defining the infinitesimal strain tensor (or
ac
small strain tensor) ε as13 f
C d P cs
b
i
an an n
1
y ha
not
ε = J + J T = ∇s u
le
Infinitesimal 2
liv or ec
(2.65)
strain tensor
1 ∂ ui ∂ u j
M
.A
= + i, j ∈ {1, 2, 3}
ε
ij
2 ∂ x j ∂ xi
m
d
uu
e
X Th
er
tin
.O
strain tensor.
E (x,t) = e (x,t) = ε (x,t)
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
70 C HAPTER 2. S TRAIN
Example 2.4 – Determine under which conditions the motion in Example 2.1
constitutes an infinitesimal strain case and obtain the infinitesimal strain ten-
rs
sor for this case. Compare it with the result obtained from the spatial and
ee
material strain tensors in Example 2.2 taking into account the infinitesimal
strain hypotheses.
n
gi
t d le En
Solution
ar
The equation of motion is given by
ar s
ge ro or
eS m
ib
ac
f
x1 = X1 − AX3
C d P cs
b
x2 = X2 − AX3 ,
i
an an n
x3 = −AX1 + AX2 + X3
y ha
le
liv or ec
.A
U1 = −AX3
m
not
U (X,t) = x − X ≡ U2 = −AX3 .
d
uu
e
U3 = −AX1 + AX2
X Th
er
tin
.O
finitesimal (A 1). Now, to obtain the infinitesimal strain tensor, first the
displacement gradient tensor J (X,t) = j (x,t) must be computed,
C
−AX3
0 0 −A
not ∂ ∂ ∂
J = U⊗∇ ≡ −AX3 ∂ X1 , ∂ X2 , ∂ X3 = 0 0 −A .
−AX1 + AX2 −A A 0
−A 0 0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Infinitesimal Strain 71
The material and spatial strain tensors obtained in Example 2.2 are, respec-
tively,
2
A −A2 −2A
not 1
E (X,t) ≡ −A2 A2 0 and
2
−2A 0 2A2
−3A2 − 2A4 A2 + 2A4 −2A − 2A3
not 1
e (X,t) ≡ A2 + 2A4 A2 − 2A4 2A3 .
2
3 3 2
rs
−2A − 2A 2A −2A
ee
Neglecting the second-order and higher-order infinitesimal terms
n
A4 A3 A2 A results in
gi
t d le En
0 0 −A 0 0 −A
ar
not not
E≡ 0 0 −A and e≡ 0 0 −A =⇒ E = e = ε ,
ar s
ge ro or
eS m
ib
−A A 0 −A A 0
ac
f
C d P cs
which is in accordance with Remark 2.14.
b
i
an an n
y ha
Considering √ the general expression (2.30) of the unit elongation in the direction
M
.A
yields
uu
e
X Th
√
er
tin
λt = 1 + 2t · ε · t ∼
= 1+t·ε ·t
| {z } (2.66)
on
.O
x
εt = λt − 1 = t · ε · t
C
14
√ √
1 + x around x = 0 is 1 + x = 1 + x/2 + O x2 .
The Taylor series expansion of
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
72 C HAPTER 2. S TRAIN
rs
ee
Consider a differential segment PQ oriented in the reference configuration
n
parallel to the coordinate axis x1 ≡ x. The stretch λx and the unit elongation εx
gi
in this direction are, according to (2.66) with t = [1, 0, 0]T ,
t d le En
ar
λx = 1 + t · ε · t = 1 + εxx =⇒ εx = λx − 1 = εxx . (2.68)
ar s
ge ro or
eS m
ib
ac
This allows assigning to the component εxx ≡ ε11 the physical meaning of unit
f
C d P cs
elongation εx in the direction of the coordinate axis x1 ≡ x. A similar interpre-
b
tation is deduced for the other components in the main diagonal of the tensor
i
an an n
Given now the components outside the main diagonal of ε , consider the dif-
.A
the coordinate directions x and y, respectively. Then, these two segments form
d
uu
an angle Θxy = π/2 in this configuration. Applying (2.43), the increment in the
e
er
tin
π εxy ∼
on
p
2 1 + 2εxx 1 + 2εyy | {z }
' εxy
C
| {z } | {z }
'1 '1
©
(2.70)
where the infinitesimal character of εxx , εyy and εxy has been taken into account.
Consequently, εxy can be interpreted from (2.70) as minus the semi-increment,
produced by the strain, of the angle between the two differential segments ini-
tially oriented parallel to the coordinate directions x and y. A similar interpre-
tation is deduced for the other components εxz and εyz ,
1 1 1
εxy = − ∆ θxy ; εxz = − ∆ θxz ; εyz = − ∆ θyz . (2.71)
2 2 2
15 The Taylor series expansion of arcsin x around x = 0 is arcsin x = x + O x2 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Infinitesimal Strain 73
rs
1 1
ε ≡ ε12 ε22 ε23 ≡ εxy εyy εyz ≡ (2.72)
2 γxy ε y 2 γyz
ee
ε13 ε23 ε33 εxz εyz εzz 1 1
2 γxz 2 γyz εz
n
gi
t d le En
ar
ar s
Remark 2.17. The components in the main diagonal of the strain ten-
ge ro or
eS m
ib
sor (named longitudinal strains) are denoted by ε(•) and coincide
ac
f
with the unit elongations in the directions of the
coordinate axes.
C d P cs
b
Positive values of longitudinal strains ε(•) > 0 correspond to an
i
increase in length of the corresponding differential segments in the
an an n
y ha
reference configuration.
le
liv or ec
M
.A
m
er
.O
de f
h iT
ε ∈ R6 ε = εx , εy , εx , γxy , γxz , γyz
| {z } | {z } (2.73)
longitudinal angular
strains strains
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
74 C HAPTER 2. S TRAIN
rs
| {z } | {z }
1 1
ee
where T(1) and T(2) are the unit vectors
n
in the
directions of PQ and PR and,
gi
(1) (2)
(1)
(2)
therefore, the relation T · T =
T
T
cosΘ = cosΘ is fulfilled. Con-
t d le En
sidering the infinitesimal character of the components of ε and ∆ θ , the follow-
ar
ing holds true16 .
ar s
ge ro or
eS m
ib
ac
f
cos θ = cos (Θ + ∆ θ ) = cosΘ · cos ∆ θ − sinΘ · sin ∆ θ =
C d P cs
b
| {z } | {z }
≈1 ≈ ∆θ
i
an an n
y ha
= cosΘ
le
liv or ec
z }| {
T(1) · T(2) +2T(1) · ε · T(2) (2.75)
= cosΘ − sinΘ · ∆ θ = p =
M
.A
p
1 + T(1) · ε · T(1) 1 + T(2) · ε · T(2)
m
| {z } | {z }
≈1 ≈1
d
uu
e
er
tin
.O
C
sinΘ sin θ
where the infinitesimal character of the strain has been taken into account and,
thus, it follows that T(1) ≈ t(1) , T(2) ≈ t(2) and Θ ≈ θ .
and cos x = 1 + O x2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Infinitesimal Strain 75
rs
ee
Figure 2.15: Variation of the angle between two differential segments in infinitesimal
strain.
n
gi
t d le En
ing (2.12) and the infinitesimal character of the components of the tensor J
ar
(see (2.61)), tensor U in (2.49) can be written as17
ar s
ge ro or
eS m
ib
ac
p p
U = FT · F = (1 + JT ) · (1 + J) = f
C d P cs
b
i
p 1
= 1 + J + J + J · J ≈ 1 + J + J = 1 + J + J =⇒ U = 1 + ε .
p
an an n
T T T
T
y ha
| {z } 2
| {z }
J
le
liv or ec
ε
(2.77)
M
.A
d
uu
1
U−1 = (1 + ε )−1 = 1 − ε = 1 −
e
J + JT .
(2.78)
X Th
2
er
tin
.O
1
C
−1 T
Q = F · U = (1 + J) · 1 − J + J =
©
2
1 1 1 =⇒ Q = 1 + Ω .
= 1 + J − J + JT − J · J + JT = 1 + J − JT
2 |2 |2 {z }
{z }
J Ω
(2.79)
17
√ √
The Taylor series expansions of tensor 1 + x around x = 0 is 1 + x = 1 + x/2 + O x2 .
18 The Taylor series expansions of tensor (1 + x)−1 around x = 0 is (1 + x)−1 = 1 − x +
O x2 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
76 C HAPTER 2. S TRAIN
rs
Remark 2.19. The tensor Ω is antisymmetric. Indeed,
ee
Ω T = 1 J − JT T = 1 JT − J = −Ω
n
Ω
2 2 .
gi
Ω = −Ω i, j ∈ {1, 2, 3}
t d le En
ji ij
ar
Consequently, the terms in the main diagonal of Ω are zero, and its
ar s
ge ro or
eS m
ib
matrix of components has the structure
ac
f
C d P cs
0 Ω12 −Ω31
b
i
Ω] = −Ω12
[Ω 0 Ω23 .
an an n
y ha
Ω31 −Ω23 0
le
liv or ec
M
.A
.O
Infinitesimal ∂ u3 ∂ u2
rotation vector: −
C
∂ x2 ∂ x3
θ1 −Ω23
©
1
1
de f
θ ≡ θ2 = −Ω31 = ∂ u1 − ∂ u3 = ∇ × u . (2.81)
not
2 ∂ x3 ∂ x1 2
θ3 −Ω12
∂u
2 ∂ u1
−
∂ x1 ∂ x2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Infinitesimal Strain 77
rs
Figure 2.16) coincide. Indeed,
n ee
0 Ω12 −Ω31 r1 Ω12 r2 − Ω31 r3
gi
not
Ω · r ≡ −Ω12 0 Ω23 r2 = −Ω12 r1 + Ω23 r3 ,
t d le En
Ω31 −Ω23 0 r3 Ω31 r1 − Ω23 r2
ar
ar s
ge ro or
eS m
ib
ac
ê1 ê2 ê3 ê1 ê2 ê3
f
C d P cs
not
θ × r ≡ θ1 θ2 θ3 = −Ω23 −Ω31 −Ω12 =
b
i
an an n
r1 r2 r3 r1 r2 r3
y ha
Ω12 r2 − Ω31 r3
le
liv or ec
= −Ω12 r1 + Ω23 r3 .
M
.A
Ω31 r1 − Ω23 r2
m
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
78 C HAPTER 2. S TRAIN
rs
ee
Figure 2.16: Product of the infinitesimal rotation vector and tensor on a vector r.
n
gi
t d le En
Consider now a differential segment dX in the neighborhood of a particle P
in the reference configuration (see Figure 2.17). In accordance with (2.82), the
ar
ar s
ge ro or
stretching transforms this vector into vector dx as follows.
eS m
ib
ac
stretching f
C d P cs
rotation
b
z }| { z }| {
i
dx = F · dX = (1 + ε + Ω) · dX = ε · dX + (1 + Ω) · dX
an an n
(2.83)
y ha
le
F (•) ≡ stretching (•) + rotation (•)
liv or ec
M
.A
m
d
uu
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Volumetric Strain 79
rs
n ee
gi
t d le En
2.12 Volumetric Strain
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Definition 2.6. The volumetric strain is the increment produced by
b
i
the deformation of the volume associated with a particle, per unit of
an an n
le
liv or ec
M
.A
dV (X, 0) dV0
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
80 C HAPTER 2. S TRAIN
• Infinitesimal strain
Considering (2.49) and recalling that Q is an orthogonal tensor (|Q| = 1), yields
rs
ee
1 + εxx εxy εxz
n
|F| = |Q · U| = |Q| |U| = |U| = |1 + ε | = det εxy 1 + εyy εyz ,
gi
t d le En
εxz εyz 1 + εzz
(2.86)
ar
ar s
ge ro or
where (2.77) has been considered. Taking into account that the components of ε
eS m
ib
are infinitesimal, and neglecting in the expression of its determinant the second-
ac
f
order and higher-order infinitesimal terms, results in
C d P cs
b
i
an an n
1 + εxx εxy εxz
y ha
| {z }
εxz εyz 1 + εzz Tr (εε )
M
.A
(2.87)
m
Then, introducing (2.87) into (2.85) yields, for the infinitesimal strain case
d
uu
e
dVt = (1 + Tr (εε )) dV0
X Th
er
tin
dVt − dV0 =⇒ e = Tr (εε ) . (2.88)
e= = |F| − 1
on
.O
dV0
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Strain Rate 81
Then,
∂v
rs
dv = · dx = l · dx
∂x
ee
, (2.90)
n
∂ vi
dvi = dx j = li j dx j i, j ∈ {1, 2, 3}
gi
∂xj
t d le En
where the spatial velocity gradient tensor l (x,t) has been introduced.
ar
ar s
ge ro or
eS m
ib
de f ∂ v (x,t)
ac
l (x,t) = f ∂x
C d P cs
b
Spatial velocity
l = v⊗∇
i
(2.91)
an an n
gradient tensor
y ha
∂ vi
li j =
i, j ∈ {1, 2, 3}
le
∂xj
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
82 C HAPTER 2. S TRAIN
rs
Strain ij i, j ∈ {1, 2, 3}
ee
rate 2 ∂ x j ∂ xi (2.93)
tensor
n
d11 d 12 d 13
gi
[d] = d12 d22 d23
t d le En
d13 d23 d33
ar
ar s
ge ro or
eS m
ib
and w is an antisymmetric tensor denominated rotation rate tensor or spin ten-
ac
sor, whose expression is f
C d P cs
b
i
an an n
1 1
y ha
de f T not
= (v ⊗ ∇ − ∇ ⊗ v) = ∇a v
w = skew l
(l ) = l − l
le
2 2
liv or ec
Rotation
w = 1 ∂ vi − ∂ v j
M
i, j ∈ {1, 2, 3}
.A
rate ij
2 ∂ x j ∂ xi (2.94)
(spin)
m
w12 −w31
d
tensor
0
uu
e
er
tin
w31 −w23 0
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Strain Rate 83
rs
dt dt
ee
Replacing (2.2) into (2.97) results in22
n
gi
. not
dX · E · dX = dx · d · dx ≡ [dx]T [d] [dx] = [dX]T FT · d · F [dX]
t d le En
. .
=⇒ dX · FT · d · F − E · dX = 0 ∀ dX =⇒ FT · d · F − E = 0
ar
ar s
ge ro or
eS m
ib
.
E = FT · d · F .
ac
f (2.98)
C d P cs
b
i
an an n
y ha
.
the material strain tensor E (X,t), providing a physical interpreta-
M
.A
the same equation reveals that tensors d (x,t) and E (X,t) are not
d
exactly the same. Both tensors will coincide in the following cases:
uu
e
X Th
• In the reference configuration: t = t0 ⇒ F = 1.
er
tin
t=t0
∂x
on
.O
22 Here, the following tensor algebra theorem is used: given a second-order tensor A, if
x · A · x = 0 is verified for all vectors x 6= 0, then A ≡ 0.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
84 C HAPTER 2. S TRAIN
Figure 2.20: Differential segment between particles of the continuous medium along
rs
time.
n ee
gi
2.13.4 Physical Interpretation of the Rotation Rate Tensor
t d le En
Taking into account the antisymmetric character of w (which implies it can be
ar
defined using only three different components), the vector
ar s
ge ro or
eS m
ib
ac
f
∂ v2 ∂ v3
C d P cs
− −
b
∂ x3 ∂ x2
i
−w23
an an n
1 1 not 1 ∂ v3 ∂ v1
y ha
−w12
∂ v1 ∂ v2
M
− −
.A
∂ x2 ∂ x1
m
er
tin
ω ×r = w·r ∀r (2.100)
on
.O
rotation motion, and ω × r = w · r as the rotation velocity of the point that has r
©
as the position vector with respect to the rotation center (see Figure 2.21). Then,
considering (2.90) and (2.92),
dv = l · dx = (d + w) · dx = d · dx + w · dx , (2.101)
| {z } | {z }
stretch rotation
velocity velocity
which allows describing the relative velocity dv of the particles in the neigh-
borhood of a given particle P (see Figure 2.22) as the sum of a relative stretch
23Observe the similarity in the structure of tensors Ω and θ in Section 2.11.6 and of tensors
w and ω seen here.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Material Time Derivatives of Strain and Other Magnitude Tensors 85
rs
ee
Figure 2.21: Vorticity vector.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
d
uu
e
velocity (characterized by the strain rate tensor d) and a relative rotation velocity
X Th
er
.O
Magnitude Tensors
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
86 C HAPTER 2. S TRAIN
dF not .
= F = l ·F
dt (2.102 (cont.))
dFi j .
= Fi j = lik Fk j i, j ∈ {1, 2, 3}
rs
dt
n ee
where (2.91) has been taken into account for the velocity gradient tensor l . To
gi
obtain the material time derivative of tensor F−1 , the time derivative of the iden-
tity F · F−1 = 1 is performed25 .
t d le En
ar
d F−1
ar s
d dF −1
ge ro or
−1 −1
eS m
ib
F · F = 1 =⇒ F·F = ·F +F· =0
dt dt dt
ac
d F−1
. f
C d P cs
= −F−1 · F · F−1 = −F−1 · l · F · F−1 = −F−1 · l =⇒
=⇒
b
i
dt
an an n
|{z} | {z }
l ·F 1
y ha
le
d F−1
liv or ec
= −F−1 · l
M
.A
dt (2.103)
dFi−1
j
m
= Fik−1 lk j i, j ∈ {1, 2, 3}
d
dt
uu
e
X Th
er
tin
.O
1 T dE . 1 .T .
F · F + FT · F =
E= F · F − 1 =⇒ =E=
2 dt 2
1 T T 1
= F · l · F + FT · l · F = FT · l + l T · F = FT · d · F
2 2
. T
| {z }
=⇒ E = F · d · F . 2d (2.104)
25 The material time derivative of the inverse tensor d F−1 /dt must not be confused with
. −1
the inverse of the material derivative of the tensor: F . These two tensors are completely
different tensors.
26 Observe that the result is the same as the one obtained in (2.98) using an alternative pro-
cedure.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Material Time Derivatives of Strain and Other Magnitude Tensors 87
Using (2.23) and (2.103) for the spatial strain tensor e yields
1 −T −1
de . 1 d −T
−1 −T d −1
e= 1−F ·F ⇒ =e=− F ·F +F · F =
2 dt 2 dt dt
1 T −T −1
= l · F · F + F−T · F−1 · l
2
. 1 T −T −1
=⇒ e = l · F · F + F−T · F−1 · l . (2.105)
2
rs
2.14.3 Volume and Area Differentials
n ee
The volume differential dV (X,t) associated with a certain particle P varies
gi
along time (see Figure 2.23) and, in consequence, it makes sense to calculate
t d le En
its material derivative. Differentiating (2.55) for a volume differential results in
ar
d |F|
ar s
d
ge ro or
eS m
dV (X,t) = |F (X,t)| dV0 (X)
ib
=⇒ dV (t) = dV0 . (2.106)
dt dt
ac
f
C d P cs
Therefore, the material derivative of the determinant of the deformation gradient
b
i
tensor |F| is27
an an n
y ha
dt dFi j dt dt | {z }
M
.A
|{z}
[F·F −1 ] =δ
lik Fk j ki ki
m
∂ vi d |F|
uu
∂ xi dt
X Th
er
tin
where (2.102) and (2.91) have been considered. Introducing (2.107) into (2.106)
on
.O
d
©
27The derivative of the determinant of a tensor A with respect to the same tensor can be
written in compact notation as d |A|/dA = |A| · A−T or, in index notation, as d |A|/dAi j =
|A| · A−1
ji .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
88 C HAPTER 2. S TRAIN
rs
Figure 2.23: Variation of the volume differential.
n ee
gi
pression, results in
t d le En
d d d |F| d
|F| · dA · F−1 = dA · F−1 + |F| · dA F−1 =
ar
(da) =
ar s
ge ro or
dt dt dt } |dt {z }
eS m
ib
| {z
ac
|F| ∇ · v
f −F−1 · l
C d P cs
= (∇ · v) |F| dA · F−1 − |F| dA · F−1 · l b
i
=⇒
an an n
| {z } | {z }
y ha
da da
le
liv or ec
d
(da) = (∇ · v) da − da · l = da · ((∇ · v) 1 − l ) , (2.109)
M
.A
dt
m
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Motion and Strains in Cylindrical and Spherical Coordinates 89
rs
Remark 2.23. An orthogonal curvilinear coordinate system (gener-
ee
ically referred to as {a, b, c}), is characterized by its physical unit
n
basis {êa , êb , êc } (kêa k = kêb k = kêc k = 1), whose components are
gi
orthogonal to each other (êa · êb = êa · êc = êb · êc = 0), as is also the
t d le En
case in a Cartesian system. The fundamental difference is that the
ar
orientation of the curvilinear basis changes at each point in space
ar s
ge ro or
eS m
ib
(êm ≡ êm (x) m ∈ {a, b, c}). Therefore, for the purposes here, an
ac
f
orthogonal curvilinear coordinate system can be considered as a mo-
C d P cs
bile Cartesian coordinate system {x0 , y0 , z0 } associated with a curvi-
b
i
linear basis {êa , êb , êc } (see Figure 2.25).
an an n
y ha
le
liv or ec
M
.A
er
tin
va vx 0 Taa Tab Tac Tx0 x0 Tx0 y0 Tx0 z0
on
.O
not not
v ≡ vb ≡ vy0 T ≡ Tba Tbb Tbc ≡ Ty0 x0 Ty0 y0 Ty0 z0
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
90 C HAPTER 2. S TRAIN
rs
• Nabla operator, ∇
n ee
T
1 ∂ 1 ∂
gi
∂ ∂ ∂
not ∂
∇ = êr + êθ + êz =⇒ ∇≡ , , (2.111)
t d le En
∂r r ∂θ ∂z ∂r r ∂θ ∂z
ar
ar s
ge ro or
eS m
ib
x = r cos θ
ac
f not
x (r, θ , z) ≡ y = r sin θ
C d P cs
b z=z
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Motion and Strains in Cylindrical and Spherical Coordinates 91
rs
1
not
ε= (u ⊗ ∇) + (u ⊗ ∇)T ≡ εx0 y0 εy0 y0 εy0 z0 = εrθ εθ θ εθ z
ee
2
n
εx0 z0 εy0 z0 εz0 z0 εrz εθ z εzz
gi
∂ ur 1 ∂ uθ ur ∂ uz
t d le En
εrr = εθ θ = + εzz =
∂r r ∂θ r ∂z
ar
ar s
ge ro or
eS m
ib
1 1 ∂ ur ∂ uθ uθ 1 ∂ ur ∂ uz
εrθ = + − εrz = +
ac
2 r ∂θ ∂r r f 2 ∂z ∂r
C d P cs
b
i
1 ∂ uθ 1 ∂ uz
an an n
εθ z = + (2.114)
y ha
2 ∂z r ∂θ
le
liv or ec
.A
Figure (2.26).
m
d
uu
er
tin
dx0 x0 dx0 y0 dx0 z0 drr drθ drz
1
not
(v ⊗ ∇) + (v ⊗ ∇)T ≡ dx0 y0 dy0 y0 dy0 z0 = drθ dθ θ dθ z
on
.O
d=
2
C
∂ vr 1 ∂ vθ vr ∂ vz
drr = dθ θ = + dzz =
∂r r ∂θ r ∂z
1 1 ∂ vr ∂ vθ vθ 1 ∂ vr ∂ vz
drθ = + − drz = +
2 r ∂θ ∂r r 2 ∂z ∂r
1 ∂ vθ 1 ∂ vz
dθ z = + (2.115)
2 ∂z r ∂θ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
92 C HAPTER 2. S TRAIN
rs
ee
• Nabla operator, ∇
n
∂ T
1 ∂ 1 1 ∂ 1
gi
∂ ∂ ∂not
∇ = êr + êθ + êφ =⇒ ∇≡ , ,
t d le En
∂r r ∂θ r sin θ ∂ φ ∂ r r ∂ θ r sin θ ∂ φ
(2.117)
ar
ar s
ge ro or
eS m
ib
ac
• Displacement vector, u, and velocity vector, v f
C d P cs
b
i
not T
an an n
le
not T
v = vr êr + vθ êθ + vφ êφ =⇒ v ≡ vr , vθ , vφ (2.119)
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
x = r sin θ cos φ
not
x (r, θ , φ ) ≡ y = r sin θ sin φ
C
z = z cos θ
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Motion and Strains in Cylindrical and Spherical Coordinates 93
∂ ur 1 ∂ uθ ur
εrr = εθ θ = +
∂r r ∂θ r
1 ∂ uφ uθ ur
rs
εφ φ = + cot φ +
r sin θ ∂ φ r r
n ee
1 1 ∂ ur ∂ uθ uθ 1 1 ∂ ur ∂ uφ uφ
εrθ = + − εrφ = + −
gi
2 r ∂θ ∂r r 2 r sin θ ∂ φ ∂r r
t d le En
1 1 ∂ uθ 1 ∂ uφ uφ
ar
εθ φ = + − cot φ (2.120)
ar s
ge ro or
2 r sin θ ∂ φ r ∂θ r
eS m
ib
ac
f
The components of ε are presented on the corresponding differential element in
C d P cs
Figure 2.28.
b
i
an an n
y ha
le
• Strain rate tensor, d
liv or ec
M
.A
dx0 x0 dx0 y0 dx0 z0 drr drθ drφ
1
not
(v ⊗ ∇) + (v ⊗ ∇)T ≡ dx0 y0 dy0 y0 dy0 z0 = drθ dθ θ dθ φ
m
d=
d
2
uu
e
er
tin
∂ vr 1 ∂ vθ vr
drr = dθ θ = +
on
.O
∂r r ∂θ r
C
1 ∂ vφ vθ vr
dφ φ = + cot φ +
©
r sin θ ∂ φ r r
1 1 ∂ vr ∂ vθ vθ 1 1 ∂ vr ∂ vφ vφ
drθ = + − drφ = + −
2 r ∂θ ∂r r 2 r sin θ ∂ φ ∂r r
1 1 ∂ vθ 1 ∂ vφ vφ
dθ φ = + − cot φ (2.121)
2 r sin θ ∂ φ r ∂θ r
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
94 C HAPTER 2. S TRAIN
rs
n ee
gi
t d le En
ar
ar s
ge ro or
Figure 2.28: Differential element in spherical coordinates.
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 95
P ROBLEMS
Problem 2.1 – A deformation that takes place in a continuous medium has the
following consequences on the triangle shown in the figure below:
1. The segment OA increases its initial length in (1 + p).
2. The angle AOB decreases in q radians its initial value.
rs
3. The area increases its initial value in (1 + r).
ee
4. p, q, r, s 1.
n
gi
The deformation is uniform and the z-axis is one of the principal directions of
t d le En
the deformation gradient tensor, which is symmetric. In addition, the stretch in
this direction is known to be λz = 1 + s. Obtain the infinitesimal strain tensor.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
Solution
C
A uniform deformation implies that the deformation gradient tensor (F) does
©
not depend on the spatial variables. Consequently, the strain tensor (E) and the
stretches (λ ) do not depend on them either. Also, note that the problem is to be
solved under infinitesimal strain theory.
The initial and final lengths of a segment parallel to the x-axis are related as
follows.
Z A Z A
OA f inal = λx dX = λx dX = λx OAinitial
O O =⇒ λx = 1 + p
OA f inal = (1 + p) OAinitial
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
96 C HAPTER 2. S TRAIN
Also, an initial right angle (the angle between the x- and y-axes) is related to its
corresponding final angle after the deformation through
π
initial angle =
q
2 =⇒ ∆ Φ xy = −γxy = −2ε xy = −q =⇒ εxy = .
π 2
final angle = + ∆ Φ xy
2
In addition, F is symmetric and the z-axis is a principal direction, therefore
rs
∂ ux ∂ ux ∂ ux
1+ ∂x
ee
F11 F12 0 ∂y ∂z
not not not ∂ uy ∂ uy ∂ uy
n
F ≡ F12 F22 0 ≡ 1 + J ≡ 1+ ,
gi
∂ x ∂ y ∂ z
0 0 F33
∂ uz ∂ uz ∂ uz
t d le En
1+
∂x ∂y ∂z
ar
ar s
ge ro or
eS m
ib
which reveals the nature of the components of the displacement vector,
ac
f
C d P cs
(
∂ u x ∂ uy
b
ux (x, y) ,
i
= = 0 =⇒
an an n
∂z ∂z uy (x, y) ,
y ha
∂ uz ∂ uz
le
= = 0 =⇒ uz (z) .
liv or ec
∂x ∂y
M
.A
1 ∂ ux ∂ uz
uu
εxz = + = 0 =⇒ εxz = 0
e
2 ∂z ∂x
X Th
er
tin
1 ∂ ux ∂ uz
εxz = + = 0 =⇒ εxz = 0
on
.O
2 ∂z ∂x
∂ uz
C
εzz = = λz − 1
=⇒ εzz = s
©
∂z
λ = 1+s z
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 97
rs
dA0 · F−1 ≡
not not
0
dA0 ≡ 0 =⇒ .
ee
1
dA0 dA0
n
1+s
gi
t d le En
Then, taking into account that |F| = Tr (εε ) + 1, and neglecting second-order
terms results in
ar
ar s
ge ro or
eS m
ib
dA = (1 + r) dA0
ac
f
1 =⇒ εyy = r − p .
C d P cs
b
dA = (1 + p + s + εyy ) dA0
1+s
i
an an n
y ha
q
M
.A
p 0
2
not q
m
ε ≡ r− p 0 .
d
2
uu
e
0 0 s
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
98 C HAPTER 2. S TRAIN
Then,
a) Justify why the infinitesimal strain theory cannot be used here.
b) Determine the deformation gradient tensor, the possible values of p and the
displacement field in its material and spatial forms.
c) Draw the deformed solid.
Solution
rs
a) The angle AOC changes from 90◦ to 45◦ therefore, it is obvious that the
ee
deformation involved is not infinitesimal. In addition, under infinitesimal strain
theory ∆ Φ 1 is satisfied and, in this problem, ∆ Φ = π/4 ≈ 0.7854.
n
gi
Observation: strains are dimensionless; in engineering, small strains are usually
t d le En
considered when these are of order 10−3 − 10−4 .
ar
ar s
ge ro or
eS m
ib
b) The conditions in the statement of the problem must be imposed one by one:
ac
f
1. Considering that F (X,t) = F (t) and knowing that dx = F · dX, the latter
C d P cs
b
can be integrated as
i
an an n
Z Z Z
y ha
M
not not
with F ≡ F21 F22 F23 and C ≡ C2 ,
m
er
.O
0 0 0
C
not
©
0 = [F] 0 + C =⇒ C ≡ 0
0 0 0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 99
rs
F ≡ 0 1 F23 .
ee
0 0 F33
n
gi
2. The condition in the statement imposes that V f inal = pVinitial .
t d le En
Expression dV f = |F| dV0 allows to locally relate the differential volumes at
different instants of time. In this case, F is constant for each fixed t, thus, the
ar
ar s
ge ro or
expression can be integrated and the determinant of F can be moved outside the
eS m
ib
integral,
ac
Z Z Z
f
C d P cs
|F| dV0 = |F| dV0 = |F| V0 .
b
Vf = dV f =
i
V V0 V0
an an n
y ha
3. The condition in the statement imposes that lAC, f inal = √p lAC, initial .
2
M
.A
er
tin
1 0 F13 0 a F13
not
xC = F · XC ≡ 0 1 F23 0 = a F23 and
on
.O
0 0 F33 a ap
C
lAC, f inal = lA0C0 = [a F13 , a F23 , ap] − [a, 0, 0] = [a (F13 − 1) , a F23 , ap] =
q q
= (a (F13 − 1))2 + (a F23 )2 + (ap)2 = a (F13 − 1)2 + F23 2 + p2 =
p p √
= √ lAC = √ 2 a = p a .
2 2
Therefore,
q
(F13 − 1)2 + F23
2 + p2 = p ⇒ (F − 1)2 + F 2 = 0 ⇒ F = 1; F = 0
13 23 13 23
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
100 C HAPTER 2. S TRAIN
0 0 p
4. The condition in the statement imposes that AOC f inal = 45◦ = π/4.
rs
not not
Considering dX(1) ≡ [1, 0, 0] and dX(2) ≡ [0, 0, 1], the corresponding vectors
ee
in the spatial configuration are computed as
n
gi
1 0 1 1 1
t d le En
(1) (1) not
dx = F · dX ≡ 0 1 0 0 = 0 ,
ar
0 0 p 0 0
ar s
ge ro or
eS m
ib
ac
0 f
1 0 1
1
C d P cs
(2) (2) not
b
dx = F · dX ≡ 0 1 0 0 = 0 .
i
an an n
y ha
0 0 p 1 p
le
liv or ec
Then, √
dx(1) · dx(2)
M
2
.A
◦
cos AOC f inal = cos 45 = (1) (2) =
dx dx 2
m
is imposed, with
uu
e
X Th
er
tin
(1) (2) p
dx = 1 , dx = 1 + p2 and dx(1) · dx(2) = 1
on
.O
such that √
C
1 2 1
= =√ =⇒ p = ±1 .
©
p
1 + p2 2 2
But |F| = p > 0, and, consequently, p = 1. Then, the deformation gradient tensor
is
1 0 1
not
F ≡ 0 1 0 .
0 0 1
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 101
z 0 0 1 Z Z
which allows determining the displacement field in material and spatial descrip-
tions as
rs
Z z
ee
not not
U (X,t) = x − X ≡ 0 and u (x,t) ≡ 0 .
n
0 0
gi
t d le En
c) The graphical representation of the deformed tetrahedron is:
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
UY = UZ = 0 , ∀ X, Y, Z
UX X=0 = 0 , ∀ X, Y
UX X=L = δ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
102 C HAPTER 2. S TRAIN
c) The possible values (positive and negative) that δ may take. Justify the an-
swer obtained.
d) The material and spatial strain tensors and the infinitesimal strain tensor.
e) Plot the curves EXX − δ /L, exx − δ /L and εx − δ /L for all possible values
of δ , indicating every significant value.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
Solution f
C d P cs
b
a) A uniform deformation implies that F (X,t) = F (t) , ∀t, X. The deformation
i
an an n
.A
∂ U (X,t)
Z Z
J= =⇒ dU = J dX =⇒ dU = J dX
∂X
m
d
uu
Z Z
e
=⇒ dU = J dX =⇒ U = J · X + C (t) .
X Th
er
tin
where C (t) is an integration constant. Then, the general expression of the mate-
on
.O
b) Using the previous result and applying the boundary conditions given in the
statement of the problem will yield the values of J and C.
Boundary conditions:
UY = UZ = 0 , ∀ X, Y, Z ⇒ Points only move in the X-direction.
UX X=0 = 0 , ∀ Y, Z ⇒ The YZ plane at the origin is fixed.
UX X=L = δ , ∀ Y, Z ⇒ This plane moves in a uniform manner
in the X-direction.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 103
If the result obtained in a) is written in component form, the equations and con-
clusions that can be reached will be understood better.
rs
∀ X, Y, Z
ee
UZ = 0 , =⇒ J31 = J32 = J33 = C3 = 0
n
From the second boundary condition:
gi
t d le En
UX X=0 = 0 , ∀ Y, Z =⇒ J12 = J13 = C1 = 0
ar
ar s
ge ro or
From the third boundary condition:
eS m
ib
ac
f
C d P cs
δ
b
UX X=L = δ , ∀ Y, Z =⇒ J11 L = δ ⇒ J11 =
L
i
an an n
y ha
Finally,
le
δ
liv or ec
δ
X
0 0 0 L
L
M
.A
0 0 0 0
d
0
uu
e
X Th
er
tin
c) In order to justify all the possible positive and negative values that δ may
on
.O
take, the condition |F| > 0 must be imposed. Therefore, the determinant of F
must be computed,
C
δ
1+ 0 0
not L =⇒ |F| = 1 + δ > 0 =⇒
F = 1+J ≡ 0 1 0 δ > −L .
L
0 0 1
d) To obtain the spatial and material strain tensors as well as the infinitesimal
strain tensor, their respective definitions must be taken into account.
1
1 − F−T · F−1
Spatial strain tensor: e=
2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
104 C HAPTER 2. S TRAIN
1 T
Material strain tensor: F ·F−1
E=
2
1 T
Infinitesimal strain tensor: ε = J ·J
2
Applying these definitions using the values of F and J calculated in b) and c),
the corresponding expressions are obtained.
exx 0 0 ,
δ 1 δ2 δ 2
rs
not
e ≡ 0 0 0 with exx = + 1+
L 2 L2 L
ee
0 0 0
n
δ
gi
EXX 0 0 L 0 0
2
with EXX = δ + 1 δ
t d le En
not not
E≡ ; ≡
0 0 0 2
ε 0 0 0
L 2L
ar
ar s
ge ro or
0 0 0 0 0 0
eS m
ib
ac
f
C d P cs
b
i
an an n
le
liv or ec
Here,
M
.A
• EXX is a second-order
parabola that contains the
m
er
tin
EXX = −1/2.
• εx is the identity straight line
on
.O
origin).
©
It can be concluded, then, that for small δ /L strains the three functions have a
very similar behavior and the same slope at the origin. That is, the same result
will be obtained with any of the definitions of strain tensor. However, outside
this domain (large or finite strains) the three curves are clearly different.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 105
E XERCISES
3z T
T
not x 2y X
not 2Y 3Z
v1 ≡ , , and v2 ≡ , , .
1+t 1+t 1+t 1+t 1+t 1+t
Determine:
rs
a) The material description of v1 and the spatial description of v2 (consider
ee
t = 0 is the reference configuration).
n
b) The density distribution in both cases (consider ρ0 is the initial density).
gi
c) The material and spatial descriptions of the displacement field as well as
t d le En
the material (Green-Lagrange) and spatial (Almansi) strain tensors for the
ar
velocity field v1 .
ar s
ge ro or
eS m
ib
d) Repeat c) for configurations close to the reference configuration (t → 0).
ac
f
C d P cs
e) Prove that the two strain tensors coincide for the conditions stated in d).
b
i
an an n
y ha
x = X +Y t , y=Y , z=Z.
M
.A
Obtain the length at time t = 2 of the segment of material line that at time t = 1
m
y (α) = α 2 ,
X Th
.O
0 tetX 0
not
tX
E≡ te 0 0 .
0 0 tetY
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
106 C HAPTER 2. S TRAIN
x=X , y=Y , z = Z − Xt .
Calculate the angle formed at time t = 0 by the differential segments that at time
t = t are parallel to the x- and z-axes.
rs
2) It is antisymmetric with re-
ee
spect to plane Y = 0, that is, the
following is satisfied:
n
gi
U (X,Y, Z) = −U (X, −Y, Z)
t d le En
ar
∀ X,Y, Z
ar s
ge ro or
eS m
ib
ac
3) Under said displacement field,
f
C d P cs
the volume of the element in the
figure does not change, its an-
b
i
an an n
.A
Determine:
d
uu
e
a) The most general expression of the given displacement field, such that condi-
X Th
er
tin
.O
deformation gradient tensor and the material strain tensor. Draw the de-
C
formed shape of the element in the figure, indicating the most significant
©
values.
c) The directions (defined by their unit vectors T) for which the deformation is
reduced to a stretch (there is no rotation).
NOTE: Finite strains must be considered (not infinitesimal ones).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 107
2.6 – The solid in the figure undergoes a uniform deformation such that points
A, B and C do not move. Assuming an infinitesimal strain framework,
a) Express the displacement field in terms of “generic” values of the stretches
and rotations.
b) Identify the null components of the strain tensor and express the rotation
vector in terms of the stretches.
rs
2) The volume becomes (1 + q)
ee
times its initial value.
n
3) The angle θ increases its value
gi
in r (given in radians).
t d le En
Under these conditions, deter-
ar
ar s
ge ro or
mine:
eS m
ib
ac
c) The strain tensor, the rotation
f
C d P cs
vector and the displacement
b
i
field in terms of p, q and r.
an an n
y ha
.A
glected.
m
d
uu
2.7 – The solid in the figure undergoes a uniform deformation with the following
e
consequences:
X Th
er
tin
not move.
C
mains constant.
3) The angle θxy remains con-
stant.
4) The angle θyz increases in r
radians.
5) The segment AF becomes
(1 + p) times its initial length.
6) The area of the triangle
ABE becomes (1 + q) its ini-
tial value.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
108 C HAPTER 2. S TRAIN
Then,
a) Express the displacement field in terms of “generic” values of the stretches
and rotations.
b) Identify the null components of the strain tensor and express the rotation
vector in terms of the stretches.
c) Determine the strain tensor, the rotation vector and the displacement field in
terms of p, q and r.
NOTE: The values of p, q and r are small and its second-order infinitesimal
terms can be neglected.
rs
ee
2.8 – The sphere in the figure undergoes a uniform deformation (F = const.)
n
such that points A, B and C move to positions A0 , B0 and C0 , respectively. Point
gi
O does not move. Determine:
t d le En
a) The deformation gradient tensor in terms of p and q.
ar
ar s
ge ro or
b) The equation of the deformed external surface of the sphere. Indicate which
eS m
ib
type of surface it is and draw it.
ac
f
C d P cs
c) The material and spatial strain tensors. Obtain the value of p in terms of q
b
i
when the material is assumed to be incompressible.
an an n
y ha
d) Repeat c) using infinitesimal strain theory. Prove that when p and q are small,
le
the results of c) and d) coincide.
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 3
Compatibility Equations
rs
n ee
gi
3.1 Introduction
t d le En
ar
Given a sufficiently regular displacement field U (X,t), it is always possible to
ar s
ge ro or
eS m
ib
find the corresponding strain field (for example, the Green-Lagrange strain field)
ac
f
by differentiating this strain field with respect to its coordinates (in this case, the
C d P cs
material ones)1 ,
b
i
an an n
1 ∂Ui ∂U j ∂Uk ∂Uk not 1
y ha
Ei j = + + = Ui, j +U j,i +Uk,i Uk, j (3.1)
2 ∂ X j ∂ Xi ∂ Xi ∂ X j 2
le
liv or ec
i, j ∈ {1, 2, 3}.
M
.A
In the infinitesimal strain case, given a displacement field u (x,t), the strain
m
field
d
1 ∂ ui ∂ u j not 1
uu
e
2 ∂ x j ∂ xi 2
er
tin
is obtained.
on
.O
The question can be formulated in reverse, that is, given a strain field ε (x,t),
C
is it possible to find a displacement field u (x,t) such that ε (x,t) is its infinites-
©
imal strain tensor? This is not always possible and the answer provides the so-
called compatibility equations.
Expression (3.2) constitutes a system of 6 (due to symmetry) partial differen-
tial equations (PDEs) with 3 unknowns: u1 (x,t), u2 (x,t), u3 (x,t). This system
is overdetermined because there exist more conditions than unknowns, and it
may not have a solution.
Therefore, for a second-order symmetric tensor ε (x,t) to correspond to a
strain tensor (and, thus, be integrable and there exist a displacement field from
which it comes) it is necessary that this tensor verifies certain conditions. These
conditions are denominated compatibility conditions or equations and guarantee
1 not
Here, the simplified notation ∂Ui /∂ X j = Ui, j is used.
109
110 C HAPTER 3. C OMPATIBILITY E QUATIONS
rs
the continuity of the continuous medium during the deformation process (see
ee
Figure 3.1).
n
gi
t d le En
Definition 3.1. The compatibility conditions are conditions that a
second-order tensor must satisfy in order to be a strain tensor and,
ar
ar s
ge ro or
therefore, for there to exist a displacement field from which it comes.
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
.A
m
d
uu
e
X Th
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Preliminary Example: Compatibility Equations of a Potential Vector Field 111
rs
∂φ ∂φ
vz = =⇒ vz − =0,
ee
∂z ∂z
n
gi
which corresponds to a system of PDEs with 3 equations and 1 unknown
(φ (x,t)), thus, the system is overdetermined and may not have a solution.
t d le En
Differentiating once (3.4) with respect to (x, y, z) yields
ar
ar s
ge ro or
∂ vx ∂ 2 φ ∂ vx ∂ 2φ ∂ vx ∂ 2φ
eS m
ib
= 2 , = , = ,
ac
∂x ∂x ∂y
f
∂ x∂ y ∂z ∂ x∂ z
C d P cs
∂ 2φ ∂ vy ∂ 2 φ ∂ 2φ
b
∂ vy ∂ vy (3.5)
= , = 2 , = ,
i
an an n
∂x ∂ y∂ x ∂y ∂y ∂z ∂ y∂ z
y ha
∂ vz ∂ 2φ ∂ vz ∂ 2φ ∂ vz ∂ 2 φ
le
= , = , = 2 ,
liv or ec
∂x ∂ z∂ x ∂y ∂ z∂ y ∂z ∂z
M
.A
∂ 2φ ∂ 2φ ∂ 2φ ∂ 2φ ∂ 2ϕ ∂ 2φ
er
tin
, , , , and . (3.6)
∂ x2 ∂ y2 ∂ z2 ∂ x∂ y ∂ x∂ z ∂ y∂ z
on
.O
So, they can be removed from the original system (3.5) and 3 relations, named
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
112 C HAPTER 3. C OMPATIBILITY E QUATIONS
Remark 3.3. The 3 compatibility equations (3.7) or (3.8) are not in-
dependent of one another and a functional relation can be established
rs
between them. Indeed, applying the condition that the divergence of
ee
the rotational of a vector field is null2 , ∇ · (∇ × v) = 0 .
n
gi
t d le En
ar
ar s
3.3 Compatibility Conditions for Infinitesimal Strains
ge ro or
eS m
ib
ac
f
Consider the infinitesimal strain field ε (x,t) with components
C d P cs
b
i
1 ∂ ui ∂ u j not 1
an an n
2 ∂ x j ∂ xi 2
le
liv or ec
.A
∂ ux 1 ∂ ux ∂ uy 1 ∂ ux ∂ uz
∂x 2 ∂y + ∂x +
m
2 ∂z ∂x
d
∂ uy 1 ∂ uy ∂ uz
e
er
∂y 2 ∂z ∂y
tin
εxz εyz εzz
∂ uz
on
.O
(symm) ×
∂z
C
(3.10)
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Compatibility Conditions for Infinitesimal Strains 113
Equation (3.11) is a system of 6 PDEs with 3 unknowns, which are the compo-
not
nents of the displacement vector u (x,t) ≡ [ux , uy , uz ]T . In general, this problem
will not have a solution unless certain compatibility conditions are satisfied. To
obtain these conditions, the equations in (3.11) are differentiated twice with re-
spect to their spatial coordinates,
2 ∂ ux
∂ εxx −
∂x
2 2 2
= 6 equations
∂ x , ∂ y , ∂ z , ∂ xy, ∂ xz, ∂ yz
.. .. (3.12)
. .
rs
1 ∂ uy ∂ uz
ee
∂ 2 εyz − +
2 ∂z ∂y
n
= 6 equations ,
∂ x2 , ∂ y2 , ∂ z2 , ∂ xy, ∂ xz, ∂ yz
gi
t d le En
providing a total of 36 equations,
ar
ar s
∂ 2 εxx ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
ge ro or
eS m
ib
= = +
ac
∂ x2 ∂ x3 ∂ x2 2 ∂ z∂ x2 ∂ y∂ x2
f
C d P cs
2 ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
b
∂ εxx
= = +
i
∂ y2 ∂ x∂ y2 ∂ y2 2 ∂ z∂ y2 ∂ y3
an an n
y ha
2 2 3
∂ 3 ux ∂ 3 uz
∂ εxx ∂ εyz 1 ∂ uy
= ··· = +
le
∂ z2 ∂ x∂ z2 ∂ z2 2 ∂ z3 ∂ y∂ z2
liv or ec
∂ 2 εxx ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
M
.A
= 2 = + 2 (3.13)
∂ x∂ y ∂ x ∂ y ∂ x∂ y 2 ∂ z∂ x∂ y ∂ y ∂ x
m
∂ 2 εxx ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
d
= 2 = +
uu
e
∂ x∂ z ∂ x ∂ z ∂ x∂ z 2 ∂ z2 ∂ x ∂ y∂ x∂ z
X Th
er
∂ 2 εxx ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
tin
= = +
∂ y∂ z ∂ x∂ y∂ z ∂ y∂ z 2 ∂ z2 ∂ y ∂ y2 ∂ z
on
.O
| {z } | {z }
(18 eqns for εxx , εyy , εzz ) (18 eqns for εxy , εxz , εyz )
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
114 C HAPTER 3. C OMPATIBILITY E QUATIONS
rs
do not appear, but there will be 21 second derivatives of the strain tensor
∂ 2 εi j /(∂ xk ∂ xl ). After the corresponding algebraic operations, the resulting equa-
ee
tions are
n
gi
2 ∂ 2 εyz
∂ 2 εzz
t d le En
de f ∂ εyy
Sxx = + − 2 =0
∂ z2 ∂ y2 ∂ y∂ z
ar
ar s
ge ro or
eS m
ib
2 ∂ 2 εxx ∂ 2 εxz
de f ∂ εzz
ac
= + − =0
S 2
f
yy
∂ x2 ∂ z2 ∂ x∂ z
C d P cs
b
2 ∂ 2 εyy ∂ 2 εxy
i
de f ∂ εxx
an an n
Szz = + − 2 =0
y ha
Compatibility ∂ y2 ∂ x2 ∂ x∂ y
equations (3.16)
le
2ε
liv or ec
de f ∂ zz ∂ ∂ εyz ∂ ε xz ∂ ε xy
Sxy = − + + − =0
M
.A
∂ x∂ y ∂ z ∂ x ∂y ∂z
2
m
∂ x∂ z ∂ y ∂ x ∂y ∂z
uu
e
X Th
2
er
tin
de f ∂ εxx ∂ ∂ εyz ∂ εxz ∂ εxy
Syz = −
+ − + + =0
∂ y∂ z ∂ x ∂x ∂y ∂z
on
.O
C
which constitute the compatibility equations for the infinitesimal strain tensor ε .
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Compatibility Conditions for Infinitesimal Strains 115
rs
+ + =0
∂x ∂y ∂z
ee
∂S ∂S ∂S
xz + yz + zz = 0
n
gi
∂x ∂y ∂z
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 3.5. The three-index operator denominated permutation op-
C d P cs
erator is given by
b
i
an an n
y ha
0 → if an index is repeated,
le
i = j or i = k or j = k
liv or ec
1 → positive (clockwise) direction of the indexes,
M
.A
ei jk =
i, j, k ∈ {123, 231, 312}
m
i, j, k ∈ {132, 321, 213}
uu
e
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
116 C HAPTER 3. C OMPATIBILITY E QUATIONS
rs
ε (t) is integrable.
n ee
gi
t d le En
ar
3.4 Integration of the Infinitesimal Strain Field
ar s
ge ro or
eS m
ib
ac
3.4.1 Preliminary Equations
f
C d P cs
b
Consider the rotation tensor Ω (x,t) for the infinitesimal strain case (see Chap-
i
an an n
1
le
liv or ec
Ω = (u ⊗ ∇ − ∇ ⊗ u) ,
2 (3.20)
M
.A
1 ∂ ui ∂ u j
Ωi j =
− i, j ∈ {1, 2, 3} .
2 ∂ x j ∂ xi
m
d
uu
and the infinitesimal rotation vector θ (x,t), associated with said rotation tensor,
e
X Th
defined as3
er
tin
−Ω23 −Ωyz
on
.O
θ1
1 1 not
θ = rot u = ∇ × u ≡ θ2 = −Ω31 = −Ωzx . (3.21)
C
2 2
©
θ3 −Ω12 −Ωxy
0 Ω12 −Ω31
3 not
The tensor Ω is skew-symmetric, i.e., Ω ≡ −Ω12 0 Ω23 .
Ω31 −Ω23 0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Integration of the Infinitesimal Strain Field 117
Adding and subtracting in (3.22) the term ∂ 2 uk /(2 ∂ xi ∂ x j ) and rearranging the
expression obtained results in
1 ∂ 2 uk 1 ∂ 2 uk
∂ Ωi j 1 ∂ ∂ ui ∂ u j
= − + − =
∂ xk 2 ∂ xk ∂ x j ∂ xi 2 ∂ xi ∂ x j 2 ∂ xi ∂ x j
∂ 1 ∂ ui ∂ uk ∂ 1 ∂ u j ∂ uk ∂ εik ∂ ε jk (3.23)
= + − + = − .
∂ x j 2 ∂ xk ∂ xi ∂ xi 2 ∂ xk ∂ x j ∂xj ∂ xi
| {z } | {z }
εik ε jk
rs
This expression can now be used to calculate the Cartesian derivatives of the
ee
components of the infinitesimal rotation vector, θ (x,t), given in (3.21), as fol-
n
lows.
gi
t d le En
∂ θ1 ∂ Ωyz ∂ εxz ∂ εxy
=− = −
∂x ∂x ∂y ∂z
ar
ar s
ge ro or
eS m
ib
∂ θ1 ∂ Ωyz ∂ εyz ∂ εyy (3.24)
ac
∇θ1 =− −
∂y ∂y
=
∂y f ∂z
C d P cs
b
i
∂ θ1 ∂ Ωyz ∂ εzz ∂ εzy
an an n
=− = −
y ha
∂z ∂z ∂y ∂z
le
liv or ec
M
.A
d
∂ θ2 ∂ Ωzx ∂ εxy ∂ εyz
uu
∇θ2 =− = − (3.25)
e
∂y ∂y ∂z ∂x
X Th
er
tin
∂ θ2 ∂ Ωzx ∂ εxz ∂ εzz
=− = −
on
.O
∂z ∂z ∂z ∂x
C
©
∂ θ3 ∂ Ωxy ∂ εxy ∂ εxx
=− = −
∂x ∂x ∂x ∂y
∇θ3 ∂ θ3 ∂ Ωxy ∂ εyy ∂ εxy (3.26)
=− = −
∂y ∂y ∂x ∂y
∂ θ3
∂ Ωxy ∂ εyz ∂ εxz
=− = −
∂z ∂z ∂x ∂y
Assume the value of the infinitesimal rotation vector θ (x,t) is known and,
through it by means of (3.21), the value of the infinitesimal rotation tensor
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
118 C HAPTER 3. C OMPATIBILITY E QUATIONS
Ω (x,t) is also known. Then, the displacement gradient tensor J (x,t) (see Chap-
ter 2, Section 2.11.6) becomes
∂ u (x,t)
J= = ε +Ω
∂x
∂ ui 1 ∂ ui ∂ u j 1 ∂ ui ∂ u j
Ji j = = + + − = εi j + Ω i j
∂xj 2 ∂ x j ∂ xi 2 ∂ x j ∂ xi
| {z } | {z } i, j ∈ {1, 2, 3} .
εi j Ωi j
(3.27)
Finally, writing in explicit form the different components in (3.27) and taking
rs
into account (3.21), the following is obtained4 .
n ee
j=1 j=2 j=3
gi
∂ ux ∂ ux ∂ ux
t d le En
i=1: = εxx = εxy − θ3 = εxz + θ2
∂x ∂y ∂z
ar
∂ uy ∂ uy ∂ uy (3.28)
ar s
ge ro or
i=2: = εxy + θ3 = εyy = εyz − θ1
eS m
ib
∂x ∂y ∂z
ac
∂ uz ∂ uz f ∂ uz
C d P cs
i=3: = εxz − θ2 = εyz + θ1 = εzz
∂x ∂y
b ∂z
i
an an n
y ha
le
liv or ec
.A
Consider ε (x,t) is the infinitesimal strain field one wants to integrate. This op-
eration is performed in two steps:
m
1) Using (3.24) through (3.26), the infinitesimal rotation vector θ (x,t) is in-
uu
e
er
tin
.O
of time, can be determined if the value (or the evolution along time) of the
infinitesimal rotation vector at some point of the medium is known.
2) Once the infinitesimal strain tensor ε (x,t) and the infinitesimal rotation
vector θ (x,t) are known, the displacement field u (x,t) is integrated. The
system of first-order PDEs defined in (3.28) is used, resulting in
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Integration of the Infinitesimal Strain Field 119
Again, the integration constants c0i (t) that appear, which, in general, will be
a function of time, are determined when the value (or the evolution along
time) of the displacements at some point of space is known.
rs
n ee
gi
t d le En
Remark 3.8. The presence of the integration constants in (3.29) and
(3.30) shows that an integrable strain tensor, ε (x,t), determines the
ar
not
ar s
motion of each instant of time except for a rotation c (t) = θ̂θ (t) and
ge ro or
eS m
ib
0 not
a translation c (t) = û (t).
ac
f
C d P cs
θ (x,t) = θ̃θ (x,t) + θ̂θ (t)
ε (x,t)
b
i
u (x,t) = ũ (x,t) + û (t)
an an n
y ha
From these uniform rotation θ̂θ (t) and translation û (t) the displace-
le
liv or ec
ment field
M
.A
u∗ (x,t) = Ω̂
Ω (t) x + û (t) =⇒ u∗ ⊗ ∇ = Ω̂
Ω
m
er
tin
1 1
ε ∗ (x,t) = ∇s u∗ = (u∗ ⊗ ∇ + ∇ ⊗ u∗ ) = Ω̂ ΩT = 0 ,
Ω + |{z}
Ω̂
2 2
on
.O
−Ω̂
Ω
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
120 C HAPTER 3. C OMPATIBILITY E QUATIONS
rs
and Ω (x,t)|x=[0,0,0]T = 0.
n ee
Solution
gi
t d le En
Infinitesimal rotation vector
ar
Posing the systems of equations defined in (3.24) through (3.26) results in
ar s
ge ro or
eS m
ib
ac
∂ θ1
=0 ;
∂ θ1
=0 ;
∂ θ1
=0 f ⇒ θ1 = C1 (t) ,
C d P cs
b
∂x ∂y ∂z
i
an an n
∂ θ2 ∂ θ2 ∂ θ2 3 3
= −3xz ; =0 ; = − x2 ⇒ θ2 = − x2 z +C2 (t) ,
y ha
∂x ∂y ∂z 2 2
le
liv or ec
∂ θ3 ∂ θ3 3 ∂ θ3 3
=0 ; = ; =0 ⇒ θ3 = y +C3 (t) .
M
∂x ∂y 2 ∂z 2
.A
m
er
tin
T
3 3
on
.O
not
C1 (t) = C2 (t) = C3 (t) = 0 =⇒ θ (x) ≡ 0 , − x2 z , y
2 2
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Compatibility Equations and Integration of the Strain Rate Field 121
Displacement vector
Posing, and integrating, the systems of equations in (3.28) produces
∂ u1 ∂ u1 ∂ u1
= 8x ; = −2y ; =0 ⇒ u1 = 4x2 − y2 +C10 (t) ,
∂x ∂y ∂z
∂ u2 ∂ u2 ∂ u2
=y ; =x ; =0 ⇒ u2 = xy +C20 (t) ,
∂x ∂y ∂z
∂ u3 ∂ u3 ∂ u3
= 3x2 z ; =0 ; = x3 ⇒ u3 = x3 z +C30 (t) .
rs
∂x ∂y ∂z
ee
not
and imposing that u (x,t)|x=(0,0,0)T ≡ [3t, 0, 0]T yields
n
gi
T
t d le En
not
C1 (t) = 3t ; C2 (t) = C3 (t) = 0 =⇒ u (x) ≡ 4x2 − y2 + 3t , xy , x3 z .
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
3.5 Compatibility Equations and Integration of the Strain
i
an an n
Rate Field
y ha
le
liv or ec
Given the definitions of the infinitesimal strain tensor ε , the infinitesimal rotation
tensor Ω and the infinitesimal rotation vector θ , there exists a clear correspon-
M
.A
dence between these magnitudes and a) the strain rate tensor d, b) the rotation
rate (or spin) tensor w and c) the spin vector ω given in Chapter 2. These corre-
m
er
tin
u v
on
.O
ε (u) d (v)
C
1 ∂ ui ∂ u j 1 ∂ vi ∂ v j
©
εi j = + di j = +
2 ∂ x j ∂ xi 2 ∂ x j ∂ xi
⇐⇒ (3.31)
1 ∂ ui ∂ u j 1 ∂ vi ∂ v j
Ωi j = − wi j = −
2 ∂ x j ∂ xi 2 ∂ x j ∂ xi
1 1
θ = ∇×u ω = ∇×v
2 2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
122 C HAPTER 3. C OMPATIBILITY E QUATIONS
To integrate this field, the same procedure as that seen in Section 3.4.2 can be
used, replacing ε by d, u by v, Ω by w and θ by ω . Certainly, this integration
can only be performed if the compatibility equations in (3.16) are satisfied for
the components of d (x,t).
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 123
P ROBLEMS
Problem 3.1 – Determine the spatial description of the velocity field that cor-
responds to the strain rate tensor
tetx 0 0
rs
not
y +1 .
d (x,t) ≡ 0 0 te
ee
0 y
te + 1 0
n
gi
not not
For x = 0, ω 0 ≡ [t − 1, 0, 0]T and v0 ≡ [t, 0, t]T f or ∀t is satisfied.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
Solution f
C d P cs
b
i
The problem is solved by integrating the corresponding differential equations,
an an n
le
liv or ec
u v
M
.A
ε ⇐⇒ d
m
θ ω
d
uu
e
er
tin
∂ ω1 ∂ ω1 ∂ ω1
on
.O
∂ ω2 ∂ ω2 ∂ ω2
©
=0 ; =0 ; =0 ⇒ ω2 = C2 (t) ,
∂x ∂y ∂z
∂ ω3 ∂ ω3 ∂ ω3
=0 ; =0 ; =0 ⇒ ω3 = C3 (t) .
∂x ∂y ∂z
The boundary conditions are imposed for x = 0,
t −1 t +C1 C1 = −1
not
ω 0 ≡ 0 = C2 =⇒ C2 = 0 ,
0 C3 = 0
C3
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
124 C HAPTER 3. C OMPATIBILITY E QUATIONS
0
Velocity vector
∂ v1 ∂ v1 ∂ v1 0
= tetx ; =0 ; =0 ⇒ v1 = C1 (t) + etx ,
∂x ∂y ∂z
rs
∂ v2 ∂ v2 ∂ v2 0
=0 ; =0 ; =2 ⇒ v2 = C2 (t) + 2z ,
ee
∂x ∂y ∂z
n
∂ v3 ∂ v3 ∂ v3 0
= 2tey ; ⇒ v3 = C3 (t) + 2tey .
gi
=0 ; =0
∂x ∂y ∂z
t d le En
The boundary conditions are imposed for x = 0,
ar
ar s
ge ro or
eS m
ib
0 0
t 1 +C1 C1 = t − 1
ac
f
C d P cs
not 0 0
v0 ≡ 0 = C2 =⇒ C2 = 0 ,
b
i
0 0
an an n
t 2t +C3 C3 = −t
y ha
le
and the spatial description of the velocity field is
liv or ec
M
.A
etx + t − 1
m
not
v (x) ≡ 2z .
d
uu
e
2tey − t
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 125
E XERCISES
3.1 – Deduce the displacement field that corresponds to the infinitesimal strain
tensor
0 tety 0
not ty
ε (x,t) ≡ te 0 0 .
0 0 tetz
rs
ee
At point (1, 1, 1) , u ≡ [2et , et , et ]T and θ ≡ [0, 0, −tet ]T is verified.
not not
n
gi
3.2 – Determine the spatial description of the velocity field that corresponds to
t d le En
the strain rate tensor
ar
ar s
ge ro or
0 0 tetz
eS m
ib
ac
not
f
d (x,t) ≡ 0 te ty 0 .
C d P cs
b
tetz 0 0
i
an an n
y ha
(
for z = 0 : vx = vz = 0 , ∀t ,
M
x, y
.A
for y = 1 : vy = 0 , ∀t , x, z
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 4
Stress
rs
n ee
gi
4.1 Forces Acting on a Continuum Body
t d le En
ar
Two types of forces that can act on a continuous medium will be considered:
ar s
ge ro or
eS m
ib
body forces and surface forces.
ac
f
C d P cs
4.1.1 Body Forces
b
i
an an n
y ha
le
Definition 4.1. The body forces are the forces that act at a distance
liv or ec
.A
d
uu
e
X Th
er
tin
on
.O
C
127
128 C HAPTER 4. S TRESS
Consider b (x,t) is the spatial description of the vector field of body forces
per unit of mass. Multiplying the vector of body forces b (x,t) by the density ρ,
the vector of body forces per unit of volume ρb (x,t) (density of body forces) is
obtained. The total resultant, fV , of the body forces on the material volume V in
Figure 4.1 is Z
fV = ρb (x,t) dV . (4.1)
V
Remark 4.1. In the definition of body forces given in (4.1), the exis-
rs
tence of the vector density of body forces ρb (x,t) is implicitly ac-
ee
cepted. This means that, given an arbitrary sequence of volumes ∆Vi
n
that contain the particle P, and the corresponding sequence of body
gi
forces f∆Vi , there exists the limit
t d le En
f∆Vi
ρb (x,t) = lim
∆Vi →0 ∆Vi
ar
ar s
ge ro or
and, in addition, it is independent of the sequence of volumes con-
eS m
ib
ac
sidered.
f
C d P cs
b
i
an an n
y ha
Earth’s surface, obtain the value of the total resultant of the body forces in
M
d
uu
e
X Th
er
tin
on
.O
C
Solution
Assuming a system of Cartesian axes (see figure above) such that the x3 -
axis is in the direction of the vertical from the center of the Earth, the vector
field b (x,t) of gravitational force per unit of mass is
not
b (x,t) ≡ [ 0 , 0 , −g ]T
and, finally, the vector of body forces is
h Z iT
R not
fV = ρb (x,t) dV ≡ 0 , 0 , − ρg dV .
V
V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Forces Acting on a Continuum Body 129
Definition 4.2. The surface forces are the forces that act on the
boundary of the material volume considered. They can be regarded
as produced by the contact actions of the particles located in the
boundary of the medium with the exterior of this medium.
Consider the spatial description of the vector field of surface forces per unit of
rs
surface t (x,t) on the continuous medium shown in Figure 4.2. The resultant
ee
force on a differential surface element dS is t dS and the total resultant of the
n
surface forces acting on the boundary ∂V of volume V can be written as
gi
Z
t d le En
fS = t (x,t) dS . (4.2)
ar
∂V
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Remark 4.2. In the definition of surface forces given in (4.2), the ex-
b
i
istence of the vector of surface forces per unit of surface t (x,t) (trac-
an an n
forces f∆ Si are considered (see Figure 4.3), there exists the limit
M
.A
f∆ Si
t (x,t) = lim
∆ Si →0 ∆ Si
m
er
tin
on
.O
C
1 In literature, the vector of surface forces per unit of surface t (x,t) is often termed traction
vector, although this concept can be extended to points in the interior of the continuous
medium.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
130 C HAPTER 4. S TRESS
rs
ee
4.2 Cauchy’s Postulates
n
gi
Consider a continuous medium on which body and surface forces are acting (see
t d le En
Figure 4.4). Consider also a particle P in the interior of the continuous medium
ar
and an arbitrary surface containing point P and with a unit normal vector n at
ar s
ge ro or
eS m
ib
this point, which divides the continuous medium into two parts (material vol-
ac
f
umes). The surface forces due to the contact between volumes will act on the
C d P cs
imaginary separating surface, considered now a part of the boundary of each of
b
i
these material volumes.
an an n
Consider the traction vector t that acts at the chosen point P as part of the
y ha
boundary of the first material volume. In principle, this traction vector (de-
le
liv or ec
fined now at a material point belonging to the interior of the original continuous
M
.A
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Cauchy’s Postulates 131
t = t (P, n)
rs
n ee
gi
t d le En
ar
ar s
ge ro or
Remark 4.3. Consider a particle P of a continuous medium and dif-
eS m
ib
ferent surfaces that contain this point P such that they all have the
ac
f
same unit normal vector n at said point. In accordance with Cauchy’s
C d P cs
b
postulate, the traction vectors at point P, according to each of these
i
an an n
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
132 C HAPTER 4. S TRESS
Definition 4.4. Cauchy’s 2nd postulate - action and reaction law es-
tablishes the traction vector at point P of a continuous medium, ac-
cording to a plane with unit normal vector n, has the same magnitude
and opposite direction to the traction vector at the same point P ac-
cording to a plane with unit normal vector −n at the same point (see
Figure 4.4).
t (P, n) = −t (P, −n)
rs
n ee
4.3 Stress Tensor
gi
t d le En
4.3.1 Application of Newton’s 2nd Law to a Continuous Medium
ar
ar s
ge ro or
Consider a discrete system of particles in motion such that a generic particle i
eS m
ib
of this system has mass mi , velocity vi and acceleration ai = dvi /dt. In addition,
ac
f
a force fi acts on each particle i, which is related to the particle’s acceleration
C d P cs
through Newton’s second law3 ,
b
i
an an n
y ha
fi = mi ai . (4.3)
le
liv or ec
Then, the resultant R of the forces that act on all the particles of the system is
M
.A
R = ∑ fi = ∑ mi ai . (4.4)
m
i i
d
uu
e
The previous concepts can be generalized for the case of continuous mediums
X Th
er
tin
.O
ous medium with total mass M, on which external forces characterized by the
vector density of body forces ρb (x,t) and the traction vector t (x,t) are acting,
C
whose particles have an acceleration a (x,t), and that occupies at time t the space
©
volume Vt results in
Z Z Z Z
R= ρb dV + t dS = a dm = ρa dV . (4.5)
Vt M
|{z} V
∂Vt t
| {z } | {z } ρdV
Resultant of Resultant of
the body the surface
forces forces
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Stress Tensor 133
rs
(outward) normals −ê1 , −ê2 and −ê3 , respectively. Through geometric consid-
ee
erations, the relations
n
gi
S1 = n1 S S2 = n2 S S3 = n3 S (4.6)
t d le En
can be established. The notation for the traction vectors on each of the faces of
ar
ar s
the tetrahedron is introduced in Figure 4.7 as well as the corresponding normals
ge ro or
eS m
ib
with which they are associated.
ac
f
According to Cauchy’s second postulate (see Definition 4.4), the traction vec-
C d P cs
b
tor on a generic point x belonging to one of the surfaces Si (with outward nor-
i
mal −êi ) can be written as
an an n
y ha
not
t (x, −êi ) = −t (x, êi ) = −t(i) (x) i ∈ {1, 2, 3} .
le
(4.7)
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
134 C HAPTER 4. S TRESS
rs
n ee
Figure 4.7: Traction vectors on an elemental tetrahedron.
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 4.4. The mean value theorem establishes that, given a
C d P cs
(scalar, vectorial o tensorial) function that is continuous in the in-
b
i
terior of a (compact) domain, the function reaches its mean value
an an n
continuous in Ω ,
le
liv or ec
Z
∃ x∗ ∈ Ω | f (x) dΩ = Ω · f (x∗ )
M
.A
Ω
m
sion.
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Stress Tensor 135
In virtue of the mean value theorem, the vector field t(i) (x), assumed to be
continuous in the domain Si , attains its mean value in the interior of this domain.
∗
Let x∗sI ∈ Si be the point where the mean value is reached and t(i) = t(i) x∗sI
this mean value. Analogously, the vectors t∗ = t x∗S , ρ ∗ b∗ = ρ (xV∗ ) b (xV∗ )
and ρ ∗ a∗ = ρ (xV∗ ) a (xV∗ ) are the mean values corresponding to the vector fields:
traction vector t (x) in S, density of body forces ρb (x) and inertial forces ρa (x),
respectively. These mean values are attained, again according to the mean value
theorem, at points x∗s ∈ S and xV∗ ∈ V of the interior of the corresponding do-
mains. Therefore, one can write
rs
Z Z
∗
t(i) (x) dS = t(i) Si t (x) dS = t∗ S ,
ee
i ∈ {1, 2, 3} ,
n
Si S
(4.8)
gi
Z Z
ρ (x) b (x) dV = ρ ∗ b∗V ρ (x) a (x) dV = ρ ∗ a∗V .
t d le En
and
ar
V V
ar s
ge ro or
eS m
ib
Applying now (4.5) on the tetrahedron considered, results in
ac
f
C d P cs
b
Z Z Z Z Z
ρb dV + t dS + t dS + t dS + t dS =
i
an an n
y ha
V S S1 S2 S3
Z Z Z Z Z Z (4.9)
le
−t(1) dS + −t(2) dS + −t(3) dS = ρa dV,
liv or ec
= ρb dV + t dS +
M
.A
V S S1 S2 S3 V
m
where (4.7) has been taken into account. Replacing (4.8) in (4.9), the latter can
d
∗ ∗ ∗
er
(4.10)
on
.O
Introducing now (4.6) and expressing the total volume of the tetrahedron as
V = Sh/3, the equation above becomes
C
1 ∗ ∗ ∗ ∗ ∗ 1
ρ b h S + t∗ S − t(1) n1 S − t(2) n2 S − t(3) n3 S = ρ ∗ a∗ h S =⇒
3 3 (4.11)
1 ∗ ∗ ∗ (1)∗
(2)∗
(3)∗ 1 ∗ ∗
ρ b h + t − t n1 − t n2 − t n3 = ρ a h .
3 3
Expression (4.11) is valid for any tetrahedron defined by a plane with unit
normal vector n placed at a distance h of point P. Consider now an infinites-
imal
tetrahedron,
also in the neighborhood of point P, by making the value
of PP0 = h tend to zero but maintaining the orientation of the plane constant
(n=constant). Then, the domains Si , S and V in (4.11) collapse into point P (see
Figure 4.7). Therefore, the points of the corresponding domains in which the
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
136 C HAPTER 4. S TRESS
and, in addition,
1 ∗ ∗ 1 ∗ ∗
lim ρ b h = lim ρ a h =0. (4.13)
h→0 3 h→0 3
rs
ee
Taking the limit of (4.11) and replacing expressions (4.12) and (4.13) in it
leads to
n
gi
t (P, n) − t(1) n1 − t(2) n2 − t(3) n3 = 0 =⇒ t (P, n) − t(i) ni = 0 . (4.14)
t d le En
ar
The traction vector t(1) can be written in terms of its corresponding Cartesian
ar s
ge ro or
eS m
ib
components (see Figure 4.9) as
ac
f
C d P cs
t(1) = σ11 ê1 + σ12 ê2 + σ13 ê3 = σ1i êi .
b
(4.15)
i
an an n
Operating in an analogous manner on traction vectors t(2) and t(3) (see Fig-
y ha
(4.16)
.A
m
t(3) = σ31 ê1 + σ32 ê2 + σ33 ê3 = σ3i êi (4.17)
d
uu
e
er
tin
.O
(i)
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Stress Tensor 137
Figure 4.9: Decomposition of the traction vector t(1) into its components.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
le
liv or ec
M
.A
(i)
t (P, n) = ni t(i) =⇒ t j (P, n) = ni t j (P) = ni σi j (P) i, j ∈ {1, 2, 3} =⇒
d
uu
e
X Th
er
.O
σ = σi j êi ⊗ ê j . (4.21)
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
138 C HAPTER 4. S TRESS
Figure 4.11: Traction vectors for the construction of the Cauchy stress tensor.
rs
n ee
gi
Remark 4.7. In accordance with (4.18) and (4.21), the Cauchy stress
t d le En
tensor is constructed from the traction vectors according to three co-
ordinate planes that include point P (see Figure 4.11). However, by
ar
ar s
ge ro or
means of (4.20), the stress tensor σ (P) is seen to contain informa-
eS m
ib
tion on the traction vectors corresponding to any plane (identified by
ac
its normal n) that contains this point. f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
faces oriented in accordance to the Cartesian planes and in which the corre-
uu
e
sponding traction vectors are decomposed into their normal and tangent compo-
X Th
er
tin
.O
notation. In this notation, the matrix of components of the stress tensor is written
as
σ11 σ12 σ13
not
σ ≡ σ21 σ22 σ23 (4.22)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Stress Tensor 139
rs
ee
Figure 4.12: Graphical representation of the stress tensor (scientific notation).
n
gi
t d le En
4.3.3.2 Engineering Notation
In engineering notation, the components of the Cauchy stress tensor (see Fig-
ar
ar s
ge ro or
ure 4.13) are written as
eS m
ib
ac
f
σx τxy τxz
C d P cs
b
not
σ ≡ τyx σy τyz (4.23)
i
an an n
τzx τyz σz
y ha
le
and each component can be characterized as follows:
liv or ec
.A
the a-axis.
m
− The component τab is the tangential (shear) stress acting on the plane per-
d
uu
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
140 C HAPTER 4. S TRESS
rs
the same direction as the positive base vector and the faces can be seen in the
ee
figure) and the negative or hidden faces.
n
The sign criterion for the visible faces is
gi
positive (+) ⇒ tension
t d le En
Normal stresses σi j or σa and
negative (−) ⇒ compression
ar
ar s
ge ro or
eS m
ib
ac
positive (+) ⇒ direction of b-axis
Tangential stresses τab f
C d P cs
negative (−) ⇒ opposite direction to b-axis
b
i
an an n
Figure 4.13 (on the visible faces of the parallelepiped) correspond to positive
le
liv or ec
.A
er
tin
on
.O
C
σ n = σn
> 0 tension
σ = t·n
< 0 compression
4 It is obvious that the negative values of the components of the stress tensor will result in
graphical representations of opposite direction to the positive values indicated in the figures.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Properties of the Stress Tensor 141
rs
ee
4.4 Properties of the Stress Tensor
n
gi
Consider an arbitrary material volume V in a continuous medium and its bound-
t d le En
ary ∂V . The body forces b (x,t) act on V and the prescribed traction vector
ar
t∗ (x,t) acts on ∂V . The acceleration vector field of the particles is a (x,t) and
ar s
ge ro or
eS m
ib
the Cauchy stress tensor field is σ (x,t) (see Figure 4.16).
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
142 C HAPTER 4. S TRESS
rs
If the system is in equilibrium, the acceleration is null (a = 0), and (4.24) is
ee
reduced to
n
gi
Internal ∇ · σ + ρb = 0 ∀x ∈ V
t d le En
equilibrium ∂ σi j (4.26)
equation ∂ x + ρb j = 0 j ∈ {1, 2, 3}
ar
i
ar s
ge ro or
eS m
ib
ac
f
which is known as the internal equilibrium equation of the continuous medium.
C d P cs
Cauchy’s equation of motion is derived from the principle of balance of linear
b
i
momentum, which will be studied in Chapter 5.
an an n
y ha
le
4.4.2 Equilibrium Equation at the Boundary
liv or ec
Equation (4.20) is applied on the boundary points taking into account that the
M
.A
(
Equilibrium
X Th
equation at (4.27)
the boundary ni σi j = t ∗j j ∈ {1, 2, 3}
on
.O
C
σ = σT
(4.28)
σi j = σ ji i, j ∈ {1, 2, 3}
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Properties of the Stress Tensor 143
Remark 4.8. The symmetry of the stress tensor allows the Cauchy’s
equation (4.24) and the equilibrium equation at the boundary (4.27)
to be written, respectively, as
∇ · σ + ρb = σ · ∇ + ρb = ρa ∀x ∈ V
∂σ ∂ σ ji
i j + ρb j = + ρb j = ρa j j ∈ {1, 2, 3}
∂ xi ∂ xi
n · σ = σ · n = t∗ (x,t)
rs
∀x ∈ ∂V
∗
ee
ni σi j = σ ji ni = t j j ∈ {1, 2, 3}
n
gi
t d le En
ar
Example 4.2 – A continuous medium moves with a velocity field whose spa-
ar s
ge ro or
eS m
ib
not
tial description is v (x,t) ≡ [z, x, y]T . The Cauchy stress tensor is
ac
f
C d P cs
b
y g (x, z,t) 0
i
an an n
not
σ ≡ h (y) z (1 + t) 0 .
y ha
0 0 0
le
liv or ec
M
.A
Determine the functions g, h and the spatial form of the body forces b (x,t)
that generate the motion.
m
d
uu
Solution
e
X Th
er
tin
.O
T h (y) = C ,
σ =σ =⇒ h (y) = g (x, z,t) =⇒
C
g (x, z,t) = C ,
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
144 C HAPTER 4. S TRESS
rs
∂
ee
∂z
n
the acceleration
gi
t d le En
0 1 0
not
a = v · ∇v ≡ [z, x, y] 0 0 1 = [y, z, x]
ar
ar s
ge ro or
1 0 0
eS m
ib
ac
is obtained. Finally, the body forces are f
C d P cs
b
i
an an n
not
b (x,t) = a (x,t) ≡ [y, z, x]T .
y ha
le
liv or ec
M
.A
er
tin
then, its matrix of components in the Cartesian basis {x, y, z} (see Figure 4.17),
on
.O
σx τxy τxz
C
not
σ ≡ τyx σy τyz . (4.29)
©
0 0 σ3 {x0 , y0 , z0 }
5 A theorem of tensor algebra guarantees that all symmetric second-order tensor diagonalizes
in an orthonormal basis and its eigenvalues are real.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Properties of the Stress Tensor 145
rs
ee
Figure 4.17: Diagonalization of the stress tensor.
n
gi
t d le En
ar
Definition 4.5. The principal stress directions are the directions, as-
ar s
ge ro or
sociated with the axes {x0 , y0 , z0 }, in which the stress tensor diago-
eS m
ib
ac
nalizes.
f
C d P cs
The principal stresses are the eigenvalues of the stress tensor
b
i
(σ1 , σ2 , σ3 ). In general, they will be assumed to be arranged in the
an an n
form σ1 ≥ σ2 ≥ σ3 .
y ha
le
liv or ec
M
.A
To obtain the principal stress directions and the principal stresses, the eigen-
value problem associated with tensor σ must be posed. That is, if λ and v are an
m
σ ·v = λv σ − λ 1) · v = 0 .
er
=⇒ (σ (4.31)
tin
The solution to this system will not be trivial (will be different to v = 0) when
on
.O
not
σ − λ 1) = |σ
det (σ σ − λ 1| = 0 . (4.32)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
146 C HAPTER 4. S TRESS
rs
4.4.5 Mean Stress and Mean Pressure
n ee
gi
Definition 4.6. The mean stress is the mean value of the principal
t d le En
stresses.
ar
1
ar s
ge ro or
σm = (σ1 + σ2 + σ3 )
eS m
ib
3
ac
f
C d P cs
b
i
an an n
Considering the matrix of components of the stress tensor in the principal stress
y ha
1 1
σm = (σ1 + σ2 + σ3 ) = Tr (σ
σ) . (4.34)
M
.A
3 3
m
d
uu
e
Definition 4.7. The mean pressure is the mean stress with its sign
X Th
er
tin
changed.
not 1
mean pressure = p̄ = −σm = − (σ1 + σ2 + σ3 )
on
.O
3
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Properties of the Stress Tensor 147
rs
4.4.6 Decomposition of the Stress Tensor into its Spherical and
ee
Deviatoric Parts
n
The stress tensor σ can be split7 into a spherical part (or component) σ sph and
gi
a deviatoric part σ 0 ,
t d le En
σ = σ sph + σ 0 . (4.35)
ar
ar s
ge ro or
|{z} |{z}
eS m
ib
spherical deviatoric
part part
ac
f
C d P cs
The spherical part is defined as
b
i
an an n
y ha
σm 0 0
1
le
de f not
σ ) 1 = σm 1 ≡ 0 σm 0 ,
liv or ec
σ sph : = Tr (σ (4.36)
3
M
.A
0 0 σm
m
where σm is the mean stress defined in (4.34). According to definition (4.35), the
d
uu
er
tin
σx τxy τxz σm 0 0
on
(4.37)
C
τxz τyz σz 0 0 σm
©
resulting in
σx 0 τxy 0 τxz 0
σx − σm τxy τxz
σ0 ≡
not 0
σy − σm = τxy σy 0 τyz 0 . (4.38)
τxy τyz
τxz τyz σz − σm τxz 0 τyz 0 σz 0
6 A tensor is defined as isotropic when it remains invariant under any change of orthogonal
basis. The general expression of an isotropic second-order tensor is T = α1 where α can be
any scalar.
7 This type of decomposition can be applied to any second-order tensor.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
148 C HAPTER 4. S TRESS
rs
tor of how far from a hydrostatic stress state the present state is
(see (4.37) and Remark 4.11).
n ee
gi
t d le En
ar
Remark 4.13. The principal directions of the stress tensor and of its
ar s
ge ro or
eS m
ib
deviatoric tensor coincide. Proof is trivial considering that, from Re-
ac
f
mark 4.11, the spherical part σ sph is diagonal in any coordinate sys-
C d P cs
tem. Consequently, if σ diagonalizes for a certain basis in (4.37), σ 0
b
i
will also diagonalize for that basis.
an an n
y ha
le
liv or ec
M
.A
Tr σ 0 = Tr σ − σ sph = Tr (σ
σ ) − Tr σ sph = 3σm − 3σm = 0 .
X Th
er
tin
on
.O
C
I1 = Tr (σ
σ ) = σii = σ1 + σ2 + σ3 , (4.39)
1
σ : σ − I12 = − (σ1 σ2 + σ1 σ3 + σ2 σ3 ) ,
I2 = (4.40)
2
I3 = det (σ
σ) . (4.41)
8The tensor invariants are scalar algebraic combinations of the components of a tensor that
do not vary when the basis changes.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Stress Tensor in Curvilinear Orthogonal Coordinates 149
Any combination of the I invariants is, in turn, another invariant. In this manner,
the J invariants
J1 = I1 = σii , (4.42)
1 2 1 1
J2 = I1 + 2I2 = σi j σ ji = (σ σ : σ) , (4.43)
2 2 2
1 3 1 1
J3 = I1 + 3I1 I2 + 3I3 = Tr (σ σ · σ · σ ) = σi j σ jk σki , (4.44)
3 3 3
are defined.
rs
ee
Remark 4.15. For a purely deviatoric tensor σ 0 , the corresponding J
n
invariants are (see Remark 4.14 and equations (4.39) to (4.44))
gi
t d le En
J1 0 = I1 0 = 0
J1 = I1 = 0
ar
ar s
ge ro or
=⇒ σ 0 =⇒ J2 0 = I2 0 = 1 (σ 1
eS m
ib
J2 = I2 σ 0 : σ 0 ) = σ 0 i j σ 0 ji
ac
J3 = I3
f
2 2
C d P cs
1
b
0
J3 = I3 0 = σ 0 i j σ 0 jk σ 0 ki
i
3
an an n
y ha
le
liv or ec
M
.A
Figure 4.18). A physical (orthonormal) basis {êr , êθ , êz } and a Cartesian system
of local axes {x0 , y0 , z0 } defined as dextrorotatory are considered at this point.
on
.O
σx0 τx0 y0 τx0 z0 σr τrθ τrz
not
σ ≡ τx0 y0 σy0 τy0 z0 = τrθ σθ τθ z . (4.45)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
150 C HAPTER 4. S TRESS
x = r cos θ
not
x (r, θ , z) ≡ y = r sin θ
z=z
rs
ee
Figure 4.18: Cylindrical coordinates.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
er
tin
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 3 Dimensions 151
x = r sin θ cos φ
not
x (r, θ , φ ) ≡ y = r sin θ sin φ
z = z cos θ
rs
n ee
Figure 4.20: Spherical coordinates.
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
152 C HAPTER 4. S TRESS
rs
n ee
Consider now the normal component σ n = σ n, where σ is the normal com-
gi
ponent of the stress on the plane, defined in accordance with the sign criterion
t d le En
detailed in Section 4.3.3.3,
ar
σ > 0 tension ,
ar s
ge ro or
σn = σ ·n (4.47)
eS m
ib
σ < 0 compression .
ac
f
C d P cs
b
Consider now the tangential component τ n , of which only its module is of inter-
i
est,
an an n
The stress state on the plane with unit normal vector n at the point considered
le
liv or ec
.A
σ ∈R
(σ , τ) → (4.49)
τ ∈ R+
m
d
uu
e
er
tin
Figure 4.23. If the infinite number of planes that contain point P are now con-
sidered (characterized by all the possible unit normal vectors n(i) ) and the corre-
on
.O
sponding values of the normal stress σi and tangential stress τi are obtained and,
finally, are represented in the half-space mentioned above, a point cloud is ob-
C
tained. One can then wonder whether the point cloud occupies all the half-space
©
n1 → (σ 1 , τ 1 )
n2 → (σ 2 , τ 2 )
· · ·
ni → (σ i , τ i )
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 3 Dimensions 153
0 0 σ3
rs
σ1 0 0 n1 σ1 n1
ee
not
t = σ · n ≡ 0 σ2 0 n2 = σ2 n2 , (4.51)
n
gi
0 0 σ3 n3 σ3 n3
t d le En
where n1 , n2 , n3 are the components of the unit normal vector n in the basis as-
ar
sociated with the principal stress directions. In view of (4.51), the normal com-
ar s
ge ro or
eS m
ib
ponent of the stress (σ ), defined in (4.47), is
ac
f
C d P cs
n1
b
i
not
t · n ≡ [σ1 n1 , σ2 n2 , σ3 n3 ] n2 = σ1 n21 + σ2 n22 + σ3 n23 = σ (4.52)
an an n
y ha
n3
le
liv or ec
.A
The modules of the traction vector and of its normal and tangential components
uu
e
er
tin
.O
where (4.53) has been taken into account. Finally, the condition that n is a unit
C
Equations (4.54), (4.52) and (4.55) can be summarized in the following ma-
trix equation.
2 2 2 2 2
σ1 σ2 σ3 n1 σ + τ2
σ1 σ2 σ3 n22 = σ =⇒ A · x = b (4.56)
1 1 1 n23 1
| {z } | {z } | {z }
A x b
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
154 C HAPTER 4. S TRESS
rs
ee
Remark 4.16. Only the solutions of system (4.56) whose compo-
n
gi
not T
nents x ≡ n21 , n22 , n23 are positive and smaller than 1 will be fea-
t d le En
sible (see (4.55)), i.e.,
ar
0 ≤ n21 ≤ 1 , 0 ≤ n22 ≤ 1 and 0 ≤ n23 ≤ 1 .
ar s
ge ro or
eS m
ib
Every pair (σ , τ) that leads to a solution x that satisfies this require-
ac
f
ment will be considered a feasible point of the half-space σ − τ,
C d P cs
b
which is representative of the stress state on a plane that contains P.
i
an an n
space σ − τ.
le
liv or ec
M
.A
Consider now the goal of finding the feasible region. Through some algebraic
m
A
e
σ 2 + τ 2 − (σ1 + σ3 ) σ + σ1 σ3 − n2 = 0 (I)
X Th
(σ1 − σ3 ) 1
er
tin
A
σ 2 + τ 2 − (σ2 + σ3 ) σ + σ2 σ3 − n22 = 0 (II)
on
.O
(σ2 − σ 3 )
A (4.57)
C
σ 2 + τ 2 − (σ1 + σ2 ) σ + σ1 σ2 − n23 = 0
(III)
©
(σ1 − σ2 )
Given, for example, equation (III) of the system in (4.57), it is easily verifiable
that it can be written as
1
(σ − a)2 + τ 2 = R2 with a = (σ1 + σ2 )
2 (4.58)
r
1 2
and R = (σ1 − σ2 ) + (σ2 − σ3 ) (σ1 − σ3 ) n23 ,
4
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 3 Dimensions 155
rs
points occupy a certain region of this half-space. This region is delimited by
ee
the maximum and minimum values of R3 (n3 ). Observing that the radical in the
n
expression of R3 in (4.59) is positive, these values are obtained for n23 = 0 (the
gi
minimum radius) and n23 = 1 (the maximum radius).
t d le En
1
ar
n23 = 0 =⇒ R3min = (σ1 − σ2 )
ar s
ge ro or
2
eS m
ib
(4.60)
ac
n23
1 f
R3max = (σ1 + σ2 ) − σ3
C d P cs
=1 =⇒
b2
i
an an n
This process is repeated for the other two equations, (I) and (II), in (4.57),
M
.A
resulting in:
R min = 1 (σ − σ )
1
m
2 3
1
− Equation (I) : C1 = (σ2 + σ3 ), 0 =⇒ 2
d
2
uu
R max = |σ − a |
e
| {z } 1 1 1
X Th
a1
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
156 C HAPTER 4. S TRESS
rs
n ee
gi
Figure 4.25: Feasible region.
t d le En
ar
ar s
ge ro or
R min = 1 (σ − σ )
eS m
ib
1 1 3
2
ac
− Equation (II) : C2 =
2 f
(σ1 + σ3 ), 0 =⇒ 2
C d P cs
R max = |σ − a |
b
| {z } 2 2 2
a2
i
an an n
y ha
R min = 1 (σ − σ )
1
le
1 2
3
liv or ec
.A
|
{z } 3 3 3
a3
For each case, a feasible region that consists in a semi-annulus defined by the
m
minimum and maximum radii is obtained. Obviously, the final feasible region
uu
e
er
Figure 4.26 shows the final construction that results of the three Mohr’s semi-
tin
circles that contain points σ1 , σ2 and σ3 . It can also be shown that every point
on
.O
within the domain enclosed by the Mohr’s circles is feasible (in the sense that
the corresponding values of σ and τ correspond to stress states on a certain plane
C
The construction of Mohr’s circle is trivial (once the three principal stresses
are known) and is useful for discriminating possible stress states on planes, de-
termining maximum values of shear stresses, etc.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 157
rs
Solution
n ee
The Mohr’s circle for the defined stress state is drawn and the given points
gi
are marked in the half-space σ − τ.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
Only the points belonging to the gray zone represent stress states (feasible
uu
e
er
tin
on
.O
stress state9 in which one of the principal stress directions is known (or assumed)
a priori. In these cases, the Cartesian axis x3 (or z-axis) is made to coincide with
said principal direction (see Figure 4.25) and, thus, the components of the stress
tensor can be written as
σ11 σ12 0 σx τxy 0
not
σ ≡ σ12 σ22 0 = τxy σy 0 . (4.61)
0 0 σ33 0 0 σz
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
158 C HAPTER 4. S TRESS
Consider now only the family of planes parallel to the x3 -axis (therefore, the
component n3 of its unit normal vector is null). The corresponding traction vec-
tor is
t1 σ11 σ12 0 n1
t (P, n) = σ · n =⇒ t2 = σ12 σ22 0 n2 (4.62)
0 0 0 σ33 0
and its component t3 vanishes. In (4.61) and (4.62) the components of the stress
tensor, σ , of the unit normal vector defining the plane, n, and of the traction
rs
vector, t, associated with direction x3 are either well known (this is the case for
σ13 , σ23 , n3 or t3 ), or do not intervene in the problem (as is the case for σ33 ). This
ee
circumstance suggests ignoring the third dimension and reducing the analysis to
n
the two dimensions associated with the x1 - and x2 -axes (or x- and y-axes), as
gi
indicated in Figure 4.27. Then, the problem can be defined in the plane through
t d le En
the components of the stress tensor
ar
" # " #
ar s
ge ro or
σ11 σ12 σx τxy
eS m
ib
not
σ≡ = (4.63)
ac
σ12 σ22
fτxy σy
C d P cs
and the components of the traction vector b
i
an an n
y ha
t2 σ12 σ22 n2
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 159
Remark 4.17. The unit normal vector n, the unit tangent vector m,
and the angle θ in Figure 4.28 have the following positive directions
associated with them.
rs
• Unit normal vector n: towards the exterior of the plane (with re-
ee
spect to the position of point P).
n
• Unit tangent vector m: generates a clockwise rotation with re-
gi
spect to point P.
t d le En
• Angle θ : defined as counterclockwise.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Consider σ , the stress tensor at a given point, whose components are defined
in a Cartesian base,
b
i
an an n
" #
σx τxy
y ha
not
σ≡ . (4.65)
le
τxy σy
liv or ec
Using (4.64), the traction vector on the given point, which belongs to the plane
M
.A
considered, is
m
not σx τxy
e
t = σ ·n ≡ = . (4.66)
X Th
τxy σy
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
160 C HAPTER 4. S TRESS
and
rs
not sin θ
τθ = t · m ≡ [σx cos θ + τxy sin θ , τxy cos θ + σy sin θ ] =
ee
− cos θ (4.68)
n
2 2
= σx sin θ cos θ − σy sin θ cos θ + τxy sin θ − cos θ ,
gi
t d le En
which can be rewritten as10
ar
ar s
ge ro or
eS m
ib
σx + σy σx − σy
ac
σθ =
2
+
2 f
cos (2θ ) + τxy sin (2θ )
C d P cs
(4.69)
b
i
σx − σy
an an n
2
le
liv or ec
M
.A
Direct problem
m
d
uu
e
X Th
er
tin
on
.O
C
Inverse problem
10 The following trigonometric relations are used here: sin (2θ ) = 2 sin θ cos θ ,
cos2 θ = (1 + cos (2θ ) ) / 2 and sin2 θ = (1 − cos (2θ ) ) / 2.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 161
σx − σy
rs
τxy
τα = sin (2α) − τxy cos (2α) = 0 =⇒ tan (2α) = σ − σ ,
ee
2 x y
2
n
gi
1 τxy
sin (2α) = ± r = ± s ,
t d le En
1 2
1+ 2 σx − σy
ar
2
+ τxy
tan (2α)
ar s
(4.70)
ge ro or
2
eS m
ib
ac
f
C d P cs
σx − σy
1 b
i
2
an an n
cos (2α) = ± p = ± s .
y ha
1 + tan2 (2α)
σx − σy
2
2
+ τxy
le
liv or ec
2
M
.A
Equation (4.70) provides two solutions (associated with the + and − signs) α1
m
and α2 = α1 + π/2, which define the two principal stress directions (orthogonal)
d
uu
to the plane being analyzed11 . The corresponding principal stress directions are
e
X Th
σx + σy σx − σy
on
.O
s
σx − σy 2
+
σ x σy 2
σ1 = + + τxy
2 2
σα → s (4.72)
2
−
σ x + σy σx σy 2
σ2 =
− + τxy
2 2
11 The third principal stress direction is the direction perpendicular to the plane being ana-
lyzed (z- or x3 -axis), see (4.61) and Figure 4.27.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
162 C HAPTER 4. S TRESS
rs
Figure 4.30: Inverse problem.
n ee
4.7.3 Inverse Problem
gi
t d le En
The problem consists in obtaining the stress state on any plane given the prin-
cipal stresses and the principal stress directions σ1 and σ2 in the plane being
ar
ar s
ge ro or
analyzed. The stress state on any plane is characterized by the angle β that
eS m
ib
forms the unit normal vector of the plane with the principal stress direction cor-
ac
f
responding to σ1 . As a particular case, the components of the stress tensor on
C d P cs
b
an elemental rectangle associated with the system of axes x − y can be obtained
i
an an n
Consider now the Cartesian system x0 − y0 , associated with the principal stress
le
directions (see Figure 4.30). Applying (4.69) with σx0 = σ1 , σy00 = σ2 , τx0 y0 = 0
liv or ec
and θ ≡ β results in
M
.A
m
σ1 + σ2 σ1 − σ2
d
σβ = + cos (2β )
uu
e
2 2 (4.73)
X Th
er
tin
σ1 − σ2
τβ = sin (2β )
2
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 163
all the possible stress states for planes that contain the point being analyzed, the
ensuing procedure is followed.
Considering a reference system that coincides with the principal stress di-
rections (as in Figure 4.30) and characterizing the inclination of the planes
by means of the angle β with the principal stress direction σ1 , one obtains
from (4.73)
σ + σ2 σ1 − σ2
σ − 1
= cos (2β )
2 2 (4.75)
τ = σ1 − σ2 sin (2β )
rs
2
ee
and, squaring both equations and adding them up results in
n
σ1 + σ2 2 σ1 − σ2 2
gi
2
σ− +τ = . (4.76)
t d le En
2 2
ar
ar s
Note that this equation, which will be valid for any value of the angle β , or,
ge ro or
eS m
ib
in other words, for any arbitrarily oriented plane that contains the point, corre-
ac
f
sponds to a circle with center C and radius R in the plane σ − τ given by (see
C d P cs
Figure 4.31)
b
i
σ1 + σ2 σ1 − σ2
an an n
C= ,0 and R = . (4.77)
y ha
2 2
le
liv or ec
in Figure 4.31.
m
The inverse proposition is also true: given a point of Mohr’s circle with co-
d
uu
ordinates (σ , τ), there exists a plane that contains P whose normal and tangent
e
stresses are σ and τ, respectively. In effect, using (4.75) the following trigono-
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
164 C HAPTER 4. S TRESS
rs
ee
Figure 4.32: Interpretation of the angle β .
n
gi
t d le En
σ1 + σ2
σ−
σ −a
ar
2
ar s
cos (2β ) = =
ge ro or
eS m
ib
σ1 − σ2 R
ac
2 f (4.78)
C d P cs
b
i
τ τ
an an n
sin (2β ) = =
σ1 − σ2
y ha
R
le
2
liv or ec
M
.A
These expressions uniquely define the angle β between the normal direction to
the plane and the principal stress direction σ1 . The plane obtained corresponds
m
er
tin
.O
a) Obtaining the point in Mohr’s circle that is representative of the stress state
C
on a plane whose normal direction forms an angle β with the principal stress
©
direction σ1 .
Take a representative point of the plane on which the principal stress direc-
tion σ1 acts (point (σ1 , 0)) and rotate an angle 2β in the direction going
from σ1 to σβ (see Figure 4.32 and Figure 4.33).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 165
rs
n ee
gi
t d le En
Figure 4.33: Representative point associated with angle β in Mohr’s circle.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
Figure 4.34: Representative points for two orthogonal planes in Mohr’s circle.
d
uu
e
X Th
er
tin
c) Mohr’s circle can be drawn if the stress state on two orthogonal planes is
known.
on
.O
thogonal planes in plane σ − τ are aligned with the center of Mohr’s circle.
©
Therefore, joining both points provides the intersection with the σ -axis that
corresponds to the center of the circle. Since two additional points of the
circle are known, the circle can be drawn.
d) Mohr’s circle can be drawn if the components of the stress tensor in a certain
orthonormal base are known.
This is a particular case of property c) in which the points representative
of a stress state on Cartesian planes are known (see Figure 4.35). Note, in
this figure, how the radius and the diametrical points of the circle can be
obtained. In addition, note that the application of property a) on the point
representative of the plane perpendicular to the x-axis implies moving in the
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
166 C HAPTER 4. S TRESS
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
Figure 4.35: Calculation of the radius and diametrical points of Mohr’s circle for a stress
an an n
y ha
.A
d
uu
e
er
tin
on
.O
• Any straight line drawn from the pole P will intersect Mohr’s cir-
cle at a point A that represents the stress state on a plane parallel
in space to that line (see Figure 4.36).
• The inverse is also verified, that is, if a straight line, paral-
lel to a given plane, is drawn from the pole P, the intersection
point B represents the stress state on this particular plane (see
Figure 4.37).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 167
rs
n ee
Figure 4.36: First property of the pole of Mohr’s circle.
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
.O
Proof
C
Consider the stress tensor at the point being analyzed and its graphical rep-
resentation on the Cartesian planes of Figure 4.38 (left)12 denoted as plane A
(vertical plane) and plane B (horizontal plane). A and B are the corresponding
points in the Mohr’s circle drawn in Figure 4.38 (right).
1) Assuming property a) is verified, the pole of Mohr’s circle can be obtained
by drawing a vertical line from point A (parallel to plane A). Then, the pole P
is located at the intersection of this line with the Mohr’s circle. Also, drawing
a horizontal line from point B (parallel to plane B) determines the location of
12 Note that, following the sign criterion of Mohr’s circle, the tangent stress on plane A is
τ = −τxy .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
168 C HAPTER 4. S TRESS
rs
ee
Figure 4.38: Proof of the properties of the pole of Mohr’s circle (1).
n
gi
t d le En
the pole at the intersection of this line with the Mohr’s circle. The same point
P is obtained in both cases, as is verified in the Figure 4.38.
ar
ar s
ge ro or
eS m
ib
2) Consider now an arbitrary plane whose normal direction forms an angle θ
ac
f
with the horizontal direction (see Figure 4.39, left) and consider also the
C d P cs
b
normal and tangent stresses, σθ and τθ , respectively, according to this plane.
i
Assuming that the major principal stress direction σ1 forms an angle α with
an an n
y ha
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Figure 4.39: Proof of the properties of the pole of Mohr’s circle (2).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 169
3) Consider the Mohr’s circle and the pole P obtained in step 1) (see Figure 4.39,
right)13 . Using property a) of Section 4.7.5, point C can be obtained. This
point is representative of the Mohr’s circle that corresponds to the plane con-
sidered, obtained by rotating from point M a double angle equal to 2 (θ − α)
such that the angle MOC is 2 (θ − α). By construction, angle AOM is 2α
and angle AOC, the sum of both, is 2 (θ − α) + 2α = 2θ . The arc included
by this angle is AMC = 2θ . Then, the angle semi-inscribed in APC, which
includes arc AMC, will be θ , which proves that the straight line PC is paral-
lel to the trace of the plane considered. Since this plane could be any plane,
the validity of the property is proven.
rs
ee
Example 4.4 – Calculate the stresses acting on state III = I + II:
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
Solution
d
uu
e
To be able to add states I and II, the stresses must act on the same planes.
X Th
er
tin
Since the two states present planes with different orientations, the stresses
acting in state II must be found for the planes given in state I. To this aim, the
on
.O
13 The following geometric properties are used here: a) the value of a central angle of a circle
is the same as the arc it includes; and b) the value of an angle semi-inscribed in a circle is
equal to half the arc it includes.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
170 C HAPTER 4. S TRESS
rs
n ee
gi
t d le En
To draw the circle, planes a and b are represented since their stress states are
known. The corresponding points in the Mohr’s circle belong to the abscissa
ar
ar s
and determine, thus, the diameter of the circle.
ge ro or
eS m
ib
The pole is obtained as the intersection of the lines that are parallel to the two
ac
f
planes inclined at 45◦ and that contain the points that they represent. Once
C d P cs
b
the pole is determined, a horizontal line is drawn from it, whose intersection
i
an an n
with the Mohr’s circle (because it is tangent to the point, the intersection in
y ha
this case is the same pole) determines the point representative of the horizon-
le
tal plane (2, 1). The same procedure is repeated for a vertical plane to obtain
liv or ec
point (2, −1). With this information, state II can be reconstructed on the hor-
M
.A
izontal and vertical planes. Then, the stresses obtained are added to those of
state I to finally obtain state III.
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle in 2 Dimensions 171
rs
n ee
4.7.7 Mohr’s Circle with the Soil Mechanics Sign Criterion
gi
t d le En
The sign criterion, with respect to the normal and tangent stresses, used in soil
mechanics is the inverse of the one used in continuum mechanics (see Fig-
ar
ar s
ure 4.40). The differences are:
ge ro or
eS m
ib
ac
• The positive stresses in soil mechanics are in the opposite direction (normal
f
C d P cs
stresses are positive when they are compressive, and the direction of the pos-
b
itive tangent stresses is defined by a counterclockwise rotation with respect
i
an an n
to the plane).
y ha
• The sign criterion for angles is the same (counterclockwise angles are posi-
le
liv or ec
tive).
M
.A
Consider the fundamental expressions in (4.73), which are the starting point
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
172 C HAPTER 4. S TRESS
Continuum mechanics: σβ , τβ , σ1 , σ2 , β
∗
σβ = −σβ
∗
τβ = −τβ
(4.79)
Soil mechanics: σ1∗ = −σ2
σ ∗ = −σ1
2∗
β = β + π/2
rs
−σβ∗ = + cos (2β ∗ − π) ,
ee
2 2 | {z }
− cos (2β ∗ )
n
−σ2∗ + σ1∗ (4.80)
gi
−τβ∗ = sin (2β ∗ − π) ,
t d le En
2 | {z }
− sin (2β ∗ )
ar
ar s
ge ro or
eS m
ib
and, operating on these expressions finally results in
ac
σ1∗ + σ2∗ σ1∗ − σ2∗f
C d P cs
σβ∗ = +
b cos (2β ∗ ) ,
i
2 2
an an n
(4.81)
σ ∗ − σ2∗
y ha
Note that the fundamental expressions in (4.81), obtained on the basis of the sign
M
.A
criterion in soil mechanics, are the same as those in (4.73), obtained on the basis
m
Mohr’s circle and the determination of its properties is the same in both cases.
uu
e
X Th
er
tin
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mohr’s Circle for Particular Cases 173
0
σm 0 0 σ1 = σm + σ1
σ sph + σ 0
not
σ= ; σ sph ≡ 0 σm 0 =⇒ σ2 = σm + σ2 0
σ3 = σm + σ3 0
|{z} |{z}
0 0 σm
spherical deviator
part part
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Figure 4.43: Mohr’s circle for a stress state and its deviator.
ac
f
C d P cs
b
i
an an n
le
liv or ec
M
.A
er
tin
.O
The proof is immediate from the construction criteria of the Mohr’s circle (see
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
174 C HAPTER 4. S TRESS
rs
Figure 4.44: Mohr’s circle for a plane pure shear stress state.
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 175
P ROBLEMS
Problem 4.1 – The solid below is subjected to the following stress state in
equilibrium.
" #
xy 5y
rs
not
σ≡ (in MPa)
ee
5y 4x
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Determine:
ac
f
C d P cs
1) The expression of the forces per unit of mass acting on the solid.
b
i
2) The expression of the normal and tangent components of the forces act-
an an n
ing on the boundary, indicating their sign according to the Mohr’s cir-
y ha
cle criterion.
le
liv or ec
M
.A
m
Solution
d
uu
e
1) The expression of the body forces is obtained directly from the internal equi-
X Th
er
" # " #
.O
1 not 1 ∂ ∂ xy 5y 1 y+5
b = − ∇ · σ =⇒ b ≡ − , =− .
C
ρ ρ ∂x ∂y 5y 4x ρ 0
©
2) The normal (σ ) and tangent (τ) components of the body forces acting on the
boundary are given by
where n and m are the unit normal vector and the unit tangent vector of the
boundary, respectively. The boundary of the solid can be divided into three parts,
according to their n and m vectors:
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
176 C HAPTER 4. S TRESS
rs
ee
Boundary 1
The traction vector for this surface is
n
gi
" #" #
1 xy 5y 1 xy + 5y
t d le En
not
t1 = n1 · σ ≡ √ [1, 1] =√ .
2 5y 4x 2 5y + 4x
ar
ar s
ge ro or
eS m
ib
Then, the corresponding normal and tangent components of the body forces are
ac
f
C d P cs
b
" #
not 1 1 1 1
i
σ1 = t1 · n1 ≡ √ [xy + 5y, 5y + 4x] √ = (4x + 10y + xy) ,
an an n
2 2 1 2
y ha
le
liv or ec
" #
not 1 1 1 1
τ1 = t1 · m1 ≡ √ [xy + 5y, 5y + 4x] √
M
= (−4x + xy) .
.A
2 2 −1 2
m
This is now particularized for the x and y values corresponding to the boundary,
uu
er
tin
1
10 − 5x − x2
σ1 = with x ∈ [0, 1] ,
on
.O
2
C
Boundary 2
The traction vector for this surface is
" # " #
not xy 5y −5y
t2 = n2 · σ ≡ [0, −1] = .
5y 4x −4x
Then, the corresponding normal and tangent components of the body forces are
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 177
" #
not 0
σ2 = t2 · n2 ≡ [−5y, −4x] = 4x ,
−1
" #
not −1
τ2 = t2 · m2 ≡ [−5y, −4x] = 5y .
0
This is now particularized for the x and y values corresponding to the boundary,
that is, for y = 0 and x ∈ [0, 1],
rs
σ2 = 4x with x ∈ [0, 1] ,
ee
τ = 0.
2
n
gi
t d le En
Boundary 3
ar
The traction vector for this surface is
ar s
ge ro or
eS m
ib
" # " #
ac
not
t3 = n3 · σ ≡ [−1, 0] f xy 5y
=
−xy
.
C d P cs
b 5y 4x −5y
i
an an n
y ha
Then, the corresponding normal and tangent components of the body forces are
le
liv or ec
" #
not −1
M
σ3 = t3 · n3 ≡ [−xy, −5y]
.A
= xy ,
0
m
" #
uu
0
e
not
τ3 = t3 · m3 ≡ [−xy, −5y] = −5y .
X Th
er
1
tin
This is now particularized for the x and y values corresponding to the boundary,
on
.O
σ3 = 0 ,
τ = −5y with y ∈ [0, 1] .
3
Note that the results for boundaries 2 and 3 could have been obtained by direct
comparison since they are a horizontal and a vertical surface, respectively:
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
178 C HAPTER 4. S TRESS
rs
n ee
gi
σ2 = σy with x ∈ [0, 1] σ3 = σx with x = 0
t d le En
ar
τ = τ with y = 0 τ = −τ = −5y with y ∈ [0, 1]
ar s
2 xy 3 xy
ge ro or
eS m
ib
ac
f
C d P cs
b
Finally, the expression of the normal and tangent components of the forces act-
i
an an n
ing on the boundary of the solid are drawn, indicating the most significant val-
y ha
ues.
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 179
rs
n ee
Solution
gi
t d le En
Note that the only difference there will be between the two circles is that one
will be translated a distance σm with respect to the other.
ar
ar s
By means of the definition of the deviatoric stress tensor,
ge ro or
eS m
ib
ac
σ 0 = σ − σm 1 =⇒ f
σzz0 = σzz − σm = a − a = 0 =⇒ σzz0 = 0
C d P cs
b
i
is deduced. The fact that the trace is an invariant and that the trace of the devia-
an an n
σ 0 ) = 0, results in
toric stress tensor is zero, Tr (σ
y ha
le
liv or ec
(
0 0 σzz0 = σ20 = 0 ,
σxx + σyy = 0 =⇒
M
.A
σ10 + σ30 = 0 .
m
Finally, the radius of the major circumference (between σ10 and σ30 ) is determined
uu
e
through the application of condition 3). The two Mohr’s circles are shown below.
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
180 C HAPTER 4. S TRESS
rs
obtain all the possible Mohr’s circles corresponding to this state, indicating the
values of the principal stresses.
n ee
gi
t d le En
Solution
ar
ar s
The following property of the Mohr’s circle in 3D must be taken into account to
ge ro or
eS m
ib
solve this problem.
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Circle number:
1 − corresponds to planes parallel to the principal stress direction of σ3 .
2 − corresponds to planes parallel to the principal stress direction of σ1 .
3 − corresponds to planes parallel to the principal stress direction of σ2 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 181
rs
n ee
gi
t d le En
2. σx is the intermediate principal stress, which results in the following
ar
ar s
ge ro or
Mohr’s circle.
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
conditions 2) and 3) cannot be satisfied at the same time since they refer
on
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
182 C HAPTER 4. S TRESS
Problem 4.4 – Determine the values of α and β for which the following stress
states are possible, considering that σ > 0 and τ = 0.5σ .
rs
n ee
gi
t d le En
ar
ar s
Solution
ge ro or
eS m
ib
ac
f
The problem is solved following the same steps in all three cases, which are:
C d P cs
b
Step 1: Draw the Mohr’s circle corresponding to the stress state. Even for the
i
an an n
stress states in which only two different pairs of points (σ , τ) belonging to the
y ha
Mohr’s circle are given, the circle can be drawn taking into account that it must
le
be symmetric with respect to the longitudinal axis.
liv or ec
Step 2: Identify the pole. In all cases, a straight horizontal line is drawn, which
M
.A
must contain the point of the Mohr’s circle corresponding to the horizontal plane.
m
Then, the pole is identified as the point where the line crosses the circle again.
d
The horizontal plane is used to identify the pole because, of the three planes
uu
e
shown for each stress state, it is the only one with a known orientation.
X Th
er
tin
Step 3: Draw a straight line joining the pole and the two (σ , τ) points corre-
on
.O
Step 4: The schematic description of the stress states on the three planes can be
redrawn with the appropriate inclination.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 183
(a)
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
(b)
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
184 C HAPTER 4. S TRESS
rs
n ee
(c)
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 185
rs
n ee
gi
t d le En
ar
Solution
ar s
ge ro or
eS m
ib
Stress state II on the vertical plane must be found to be able to add states I and
ac
II together. f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
The Mohr’s circle of state II will allow determining the normal and shear stress
C
on the vertical plane. The known stress state on the horizontal plane (4, −3)
©
τ =3 .
Now, a third point belonging to the Mohr’s circle must be obtained in order to be
able to draw the complete circle. Because there exists only one pole and it must
belong to the Mohr’s circle, finding this point will allow completing the circle. A
straight horizontal line (parallel to the horizontal plane) is draw at point (4, −3),
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
186 C HAPTER 4. S TRESS
which corresponds to the stress state on a horizontal plane. Another straight line,
parallel to the other plane with a known stress state, the plane inclined at 45◦
in the counterclockwise direction, is drawn passing through the corresponding
stress state, (4, 3). The point where these to lines meet provide the pole of the
Mohr’s circle, which is found to be at (−2, −3):
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
Once these three points are known, the Mohr’s circle can be drawn. Before cal-
le
liv or ec
culating the stress state on the vertical plane, the value of σ is sought. To obtain
the stress state on the plane inclined at 45◦ in the clockwise direction, a straight
M
.A
line must be drawn, parallel to this plane, that crosses the pole.
m
d
uu
e
X Th
er
tin
on
.O
C
This results in a line tangent to the pole, therefore, the stress state corresponding
to the pole is also the stress state on this plane and
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 187
σ = −2 .
Finally, a vertical line is drawn from the pole and the intersection of this line with
the Mohr’s circle provides the stress state on the vertical plane, which results
in (−2, 3).
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
188 C HAPTER 4. S TRESS
σ0 = 3
σ 00 = 7
The values of τ 0 and α remain to be found. To this aim, the Mohr’s circle of
stress state III must be drawn. The points corresponding to the known stress
states on the vertical and horizontal planes are marked on the σ − τ space and,
in a procedure analogous to the one used for the Mohr’s circle of state II, the
pole is obtained. The circle can now be drawn and simple trigonometry allows
rs
calculating its center at (2, 0), which will be useful in the calculation of τ 0 and
ee
α.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
.O
tive value of τ 0 and another corresponding to the same value but with a negative
C
sign. Following the sign criterion for the Mohr’s circle, and to be consistent with
©
the directions drawn in the figure representing state III, the value of τ 0 must be
√
τ 0 = −3 2 .
Since there are two possible values of τ 0 , two values of α will exist, each cor-
responding to one of the τ 0 values. To obtain the values of α, a straight line is
drawn from the pole to each of the points representing the possible stress states
of the plane inclined at α in a clockwise direction.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 189
rs
n ee
gi
t d le En
ar
ar s
Determining the inclination of these two lines will result directly in the possible
ge ro or
eS m
ib
values of α.
ac
f
C d P cs
√
b
√
τ 0 = −3 2 ⇒ α + = 180◦ − arctan 1+3 2 ' 141◦
i
an an n
y ha
√ √
−τ 0 = 3 2 ⇒ α − = arctan 2−1 ' 8◦
le
liv or ec
3
M
.A
The two possible configurations of stress state III are pictured below.
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
190 C HAPTER 4. S TRESS
E XERCISES
rs
n ee
gi
4.2 – The following is known of the stress state in a point of a continuous
t d le En
medium. The maximum shear stress in planes parallel to the principal stress
ar
direction of σ1 is τmax = 2. Obtain all the values of σ1 , σ2 and σ3 that make
ar s
ge ro or
eS m
ib
possible the stress state σ = 2 and τ = 2 on a certain plane for the following
ac
cases (separately). f
C d P cs
b
i
a) The maximum shear stress in planes parallel to the principal stress di-
an an n
rection of σ2 is τ2max = 2.
y ha
b) The maximum shear stress in planes parallel to the principal stress di-
le
liv or ec
rection of σ3 is τ3max = 0.
M
.A
c) The maximum shear stress in planes parallel to the principal stress di-
rection of σ2 is τ2max = 4.
m
d
uu
e
X Th
er
tin
.O
belonging to P.
C
a) σ = 4 and τ = 2.
©
b) σ = 4 and τ = 1.
c) σ = 7 and τ = 0.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 191
4.4 – Obtain, in terms of τ, the principal stresses and the value of the maximum
shear stress of the state that results from the sum of states I and II.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
4.5 – Given states I and II, determine the possible values of σ and τ for which
ac
f
state III = I + II verifies that the principal stress σ2 is positive and its direction
C d P cs
forms a 30◦ angle with the y-axis.
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
192 C HAPTER 4. S TRESS
4.6 – Determine all the possible values of τ ∗ for which the stress state that is the
sum of states I and II verifies the following conditions (separately).
rs
n ee
a) There do not exist tensile stresses on any plane.
gi
b) There do not exist compressive stresses on any plane.
t d le En
c) The maximum shear stress (τmax ) is less than 2.
ar
d) It is a pure shear stress state.
ar s
ge ro or
eS m
ib
e) It is a hydrostatic stress state.
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 5
Balance Principles
rs
n ee
gi
5.1 Introduction
t d le En
ar
Continuum Mechanics is based on a series of general postulates or principles
ar s
ge ro or
eS m
ib
that are assumed to always be valid, regardless of the type of material and the
ac
f
range of displacements or deformations. Among these are the so-called balance
C d P cs
principles:
b
i
an an n
• Conservation of mass
y ha
.A
er
tin
.O
C
in the medium, which derives from the motion of its particles. The continuous
medium is composed of particles, some of whose properties are associated with
the amount of mass: specific weight, angular momentum, kinetic energy, etc.
Then, when particles move and transport their mass, a transport of the these
properties occurs, named convective transport (see Figure 5.1).
Consider A, an arbitrary (scalar, vector or tensor) property of the continuous
medium, and Ψ (x,t), the description of the amount of said property per unit of
mass of the continuous medium. Consider also S, a control surface, i.e., a surface
fixed in space (see Figure 5.2). Due to the motion of the particles in the medium,
these cross the surface along time and, in consequence, there exists a certain
amount of the property A that, associated with the mass transport, crosses the
control surface S per unit of time.
193
194 C HAPTER 5. BALANCE P RINCIPLES
rs
n ee
gi
Figure 5.1: Convective transport in the continuous medium.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Definition 5.1. The convective flux (or mass transport flux) of a
C d P cs
generic property A through a control surface S is the amount of A
b
i
that, due to mass transport, crosses the surface S per unit of time.
an an n
y ha
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mass Transport or Convective Flux 195
Figure 5.3: Cylinder occupied by the particles that have crossed dS in the time inter-
rs
val [t, t + dt].
n ee
gi
To obtain the mathematical expression of the convective flux of A through the
t d le En
surface S, consider a differential surface element dS and the velocity vector v of
the particles that at time t are on dS (see Figure 5.3). In a time differential dt,
ar
ar s
ge ro or
these particles will have followed a pathline dx = v dt, such that at the instant of
eS m
ib
time t + dt they will occupy a new position in space. Taking now into account all
ac
f
the particles that have crossed dS in the time interval [t, t + dt], these will occupy
C d P cs
b
a cylinder generated by translating the base dS along the directrix dx = v dt, and
i
an an n
dV = dS dh = v · n dt dS . (5.1)
le
liv or ec
Since the volume (dV ) of the particles crossing dS in the time interval
M
.A
[t, t + dt] is known, the mass crossing dS in this same time interval can be ob-
tained by multiplying (5.1) by the density,
m
dm = ρ dV = ρv · n dt dS . (5.2)
uu
e
X Th
er
tin
Finally, the amount of A crossing dS in the time interval [t, t + dt] is calculated
by multiplying (5.2) by the function Ψ (amount of A per unit of mass),
on
.O
Ψ dm = ρ Ψ v · n dt dS . (5.3)
C
Dividing (5.3) by dt yields the amount of the property that crosses the differ-
ential control surface dS per unit of time,
Ψ dm
d ΦS = = ρ Ψ v · n dS . (5.4)
dt
Integrating (5.4) over the control surface S results in the amount of the property
A crossing the whole surface S per unit of time, that is, the convective flux of the
property A through S.
Z
convective flux
o
ΦS = ρ Ψ v · n dS (5.5)
of A through S
S
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
196 C HAPTER 5. BALANCE P RINCIPLES
Example 5.1 – Compute the magnitude Ψ and the convective flux ΦS corre-
sponding to the following properties: a) volume, b) mass, c) linear momen-
tum, d) kinetic energy.
Solution
a) If the property A is the volume occupied by the particles, then Ψ is the
volume per unit of mass, that is, the inverse of the density. Therefore,
1
Z
A ≡V and Ψ = lead to ΦS = v · n dS = volume flow rate .
rs
ρ
S
ee
b) If the property A is the mass, then Ψ is the mass per unit of mass, that
n
gi
is, the unit. Therefore,
t d le En
Z
A ≡ M and Ψ = 1 lead to ΦS = ρ v · n dS .
ar
ar s
ge ro or
eS m
ib
S
ac
f
c) If the property A is the linear momentum (= mass × velocity), then Ψ
C d P cs
b
is the linear momentum per unit of mass, that is, the velocity. Therefore,
i
an an n
y ha
Z
A ≡ m v and Ψ = v lead to ΦS = ρ v (v · n) dS .
le
liv or ec
S
M
(Note that in this case Ψ and the convective flux ΦS are vectors).
.A
m
d) If the property A is the kinetic energy then Ψ is the kinetic energy per
d
er
tin
1 1 1
Z
A ≡ m |v|2 and Ψ = |v|2 lead to ΦS = ρ |v|2 (v · n) dS .
2 2 2
on
.O
S
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Mass Transport or Convective Flux 197
rs
n ee
gi
Figure 5.4: Net outflow through a closed control surface.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 5.2. The convective flux of any property through a material
C d P cs
b
surface is always null. Indeed, the convective flux of any property is
i
associated, by definition, with the mass transport (of particles) and,
an an n
y ha
.A
d
uu
e
X Th
er
tin
.O
1 Unless stated otherwise, when dealing with closed surfaces, the positive direction of the
unit normal vector n is taken in the outward direction of the surface.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
198 C HAPTER 5. BALANCE P RINCIPLES
rs
V
ee
To compute the content of property A at a different time t + ∆t, the following
n
gi
two situations arise:
t d le En
1) A control volume V is considered and, therefore, it is fixed in space and
crossed by the particles along time.
ar
ar s
ge ro or
eS m
ib
2) A material volume that at time t occupies the spatial volume Vt ≡ V is
ac
f
considered and, thus, the volume occupies different positions in space
C d P cs
b
along time.
i
an an n
Different values of the amount Q (t + ∆t) are obtained for each case, and com-
y ha
puting the difference between the amounts Q (t + ∆t) and Q (t) when ∆t → 0
le
yields
liv or ec
Q (t + ∆t) − Q (t)
Q0 (t) = lim , (5.8)
M
.A
∆t→0 ∆t
resulting in two different definitions of the time derivative, which lead to the
m
er
tin
.O
Z
©
Q (t) = µ (x,t) dV ,
V
is the time derivative of Q (t) when the volume V is a volume fixed
in space (control volume), see Figure 5.5. The notation
not ∂ R
local derivative = µ (x,t) dV
∂t V
will be used.
2µ is related to Ψ = (amount of A)/(unit of mass) through µ = ρ Ψ and has the same tensor
order as the property A .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Local and Material Derivatives of a Volume Integral 199
rs
ee
Figure 5.5: Local derivative of a volume integral.
n
gi
The amount Q of the generic property A in the control volume V at times t
t d le En
and t + ∆t is, respectively,
ar
ar s
ge ro or
Z Z
eS m
ib
Q (t) = µ (x,t) dV and Q (t + ∆t) = µ (x,t + ∆t) dV . (5.9)
ac
f
C d P cs
V V
b
i
Using (5.9) in addition to the concept of time derivative of Q (t) results in3
an an n
y ha
1
Z
∂
le
0
Q (t) = µ (x,t) dV = lim Q (t + ∆t) − Q (t) =
liv or ec
∂t ∆t→0 ∆t
M
.A
V
1
Z Z
m
∆t→0 ∆t
uu
e
V V
X Th
er
tin
.O
V | {z } V | {z }
∂ µ (x,t) local
C
derivative
©
∂t of µ
(5.10)
which yields the mathematical expression of the local derivative of a volume
integral.
3 Note that the integration domain does not vary when the volume V is considered as a control
volume and, therefore, is fixed in space.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
200 C HAPTER 5. BALANCE P RINCIPLES
rs
dt Vt
ee
will be used.
n
gi
t d le En
The content Q of the generic property A in the material volume Vt at times t and
ar
t+∆t is, respectively,
ar s
ge ro or
eS m
ib
Z Z
ac
Q (t) = µ (x,t) dV and f
Q (t + ∆t) = µ (x,t + ∆t) dV . (5.12)
C d P cs
Vt
b Vt+∆t
i
an an n
y ha
Q (t + ∆t) − Q (t)
M
d
Z
.A
Q0 (t) = µ (x,t) dV = lim =
dt ∆t→0 ∆t
m
Vt Vt ≡V
(5.13)
d
uu
e
1
Z Z
= lim µ (x,t + ∆t) dV − µ (x,t) dV .
X Th
er
tin
∆t→0 ∆t
Vt+∆t Vt
on
.O
suitable for one of the two integrals in (5.13), which lead to the same integra-
©
tion domain in both expressions. These variable substitutions are given by the
equation of motion x = ϕ (X,t), particularized for times t and t + ∆t,
xt = ϕ (X,t) → (dx1 dx2 dx3 )t = |F (X,t)| (dX1 dX2 dX3 ) ,
| {z } | {z }
dVt dV0
xt+∆t = ϕ (X,t + ∆t) → (dx1 dx2 dx3 )t+∆t = |F (X,t + ∆t)| (dX1 dX2 dX3 ) ,
| {z } | {z }
dVt+∆t dV0
(5.14)
4 Note that the integration domains are now different at times t and t + ∆t.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Local and Material Derivatives of a Volume Integral 201
rs
Figure 5.6: Material derivative of a volume integral.
n ee
gi
where the identity dVt = |F (X,t)| dV0 has been taken into account. The variable
t d le En
substitutions in (5.14) are introduced in (5.13), resulting in
ar
ar s
µ̄ (X,t + ∆t)
ge ro or
eS m
ib
d 1
Z Z z }| {
ac
dt
µ (x,t) dV = lim
f
µ (x (X,t + ∆t) ,t + ∆t) |F (X,t + ∆t)| dV!0
C d P cs
∆t→0 ∆t
b
Z
Vt V0
− µ (x (X,t) ,t) |F (X,t)| dV0
i
=
an an n
| {z }
y ha
V0
µ̄ (X,t)
le
liv or ec
.A
∆t→0 ∆t
V0 |
m
{z }
d
d
∂
µ̄ (X,t) |F (X,t)| = µ (x,t) |F (x,t)|
uu
e
∂t dt
X Th
er
tin
d
Z
= µ |F| dV0 .
dt
on
.O
V0 (5.15)
Finally, expanding the last integral in (5.15) 5 and considering the equality
C
!
d d dµ d |F|
Z Z Z
µ (x,t) dV = µ |F| dV0 = |F| + µ dV0 =
dt dt dt dt
Vt V0 V0 | {z }
|F| ∇ · v
Z Z
dµ dµ
= + µ∇ · v |F| dV0 = + µ∇ · v dV ,
dt | {z } dt
V0 dVt Vt
(5.16)
5 The change of variable xt = ϕ (X,t) is undone here.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
202 C HAPTER 5. BALANCE P RINCIPLES
that is6 ,
Z
d d dµ
Z Z
not
µ (x,t) dV = µ (x,t) dV = + µ∇ · v dV . (5.17)
dt dt dt
Vt Vt ≡V V
Vt ≡V
rs
dt ∂t
Vt ≡V V | {z }
ee
∇ · (µv) (5.18)
n
Z Z Z Z
∂µ ∂
gi
= dV + ∇ · (µv) dV = µ dV + ∇ · (µv) dV ,
∂t ∂t
t d le En
V V V V
ar
where the expression of the local derivative (5.11) has been taken into account.
ar s
ge ro or
eS m
ib
Then, (5.18) produces the expression of the material derivative of a volume in-
ac
tegral. f
C d P cs
b
i
an an n
d
Z Z Z
∂
le
µ (x,t) dV = µ dV + ∇ · (µv) dV
liv or ec
dt ∂t (5.19)
Vt ≡V V V
M
.A
| {z } | {z } | {z }
material local convective
derivative derivative
m
derivative
d
uu
e
X Th
er
tin
.O
tion 1.4) also appears here when differentiating integrals in the con-
©
6 The expression
d
Z
µ (x,t) dV
dt
Vt ≡V
denotes the time derivative of the integral over the material volume Vt (material derivative of
the volume integral) particularized at time t, when the material volume occupies the spatial
volume V .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Conservation of Mass. Mass continuity Equation 203
rs
ee
Figure 5.7: Principle of conservation of mass in a continuous medium.
n
gi
t d le En
5.4 Conservation of Mass. Mass continuity Equation
ar
ar s
ge ro or
eS m
ib
ac
f
Definition 5.4. Principle of conservation of mass. The mass of a
C d P cs
b
continuous medium (and, therefore, the mass of any material vol-
i
an an n
le
liv or ec
M
.A
description of the density, ρ (x,t). The mass enclosed by the material volume V
d
uu
er
tin
Z Z
M (t) = ρ (x,t) dV and M (t + ∆t) = ρ (x,t + ∆t) dV . (5.20)
on
.O
Vt Vt+∆t
C
satisfied.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
204 C HAPTER 5. BALANCE P RINCIPLES
rs
ee
which must be satisfied for Vt and, also, for any partial material volume ∆Vt ⊂ Vt
that could be considered. In particular, it must be satisfied for each of the ele-
n
gi
mental material volumes associated with the different particles in the continuous
medium that occupy the differential volumes dVt . Applying (5.22) on each dif-
t d le En
ferential volume dVt ≡ dV (x,t) yields7
ar
ar s
ge ro or
eS m
ib
Z
dρ dρ (x,t)
+ ρ∇ · v dV = + ρ (x,t) ∇ · v (x,t) dV (x,t) = 0
ac
dt dt f
C d P cs
b
dV (x,t)
∀x ∈ Vt , ∀t
i
an an n
dρ
y ha
=⇒ + ρ∇ · v = 0 dV ∀x ∈ Vt , ∀t
dt
le
liv or ec
(5.23)
M
.A
dρ
e
+ ρ∇ · v = 0 dV ∀x ∈ Vt , ∀t
X Th
dt
er
tin
on
which constitutes the so-called mass continuity equation. Replacing the expres-
.O
(5.24) results in
©
∂ρ ∂ρ
+ v · ∇ρ + ρ∇ · v = 0 =⇒ + ∇ · (ρv) = 0 , (5.25)
∂t | {z } ∂t
∇ · (ρv)
7 This procedure, which allows reducing a global (or integral) expression such as (5.22) to
a local (or differential) one such as (5.24), is named in continuum mechanics localization
process.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Conservation of Mass. Mass continuity Equation 205
∂ρ
+ ∇ · (ρv) = 0
∂t
∂ ρ ∂ (ρvi ) ∀x ∈ Vt , ∀t (5.26)
+ =0 i ∈ {1, 2, 3}
∂t ∂ xi
∂ ρ ∂ (ρvx ) ∂ (ρvy ) ∂ (ρvz )
+ + + = 0
∂t ∂x ∂y ∂z
rs
5.4.2 Material Form of the Principle of Conservation of Mass
ee
From (5.22)8 ,
n
gi
Z Z
dρ dρ 1 d |F|
t d le En
+ ρ∇ · v dV = +ρ dV =
dt dt |F| dt
ar
Vt Vt
ar s
ge ro or
eS m
ib
1 dρ d |F| 1 d
Z Z
= |F| +ρ dV = ρ |F| dV =
ac
|F| dt dt f |F| dt
C d P cs
|{z}
Vt Vt
b |F| dV0
| {z }
i
d
an an n
ρ |F|
y ha
Z dt
le
∂
liv or ec
.A
V0
(5.27)
m
where the integration domain is now the volume in the reference configura-
d
uu
tion, V0 . Given that (5.27) must be satisfied for each and every part ∆V0 of V0 , a
e
er
tin
∂
on
ρ (X,t) |F (X,t)| = 0 ∀X ∈ V0 , ∀t
.O
∂t
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
206 C HAPTER 5. BALANCE P RINCIPLES
rs
Z
Q (t) = ρΨ dV . (5.30)
n ee
Vt ≡V
gi
The variation along time of the content of property A in the material volume
t d le En
Vt is given by the time derivative of Q (t), which using expression (5.17) of the
ar
material derivative of a volume integral (with µ = ρΨ ) results in
ar s
ge ro or
eS m
ib
ac
Z
f
d d (ρΨ )
Z
0
Q (t) = ρΨ dV = + ρΨ ∇ · v dV . (5.31)
C d P cs
b
dt |{z} dt
i
Vt ≡V µ V
an an n
y ha
grouping terms and introducing the mass continuity equation (5.24) yields
M
.A
Z
d dΨ dρ
Z
ρΨ dV = ρ +Ψ + ρΨ ∇ · v dV =
dt dt dt
m
Vt ≡V V
d
!
uu
(5.32)
dΨ dρ
Z
e
= ρ +Ψ + ρ∇ · v dV =⇒
X Th
er
dt dt
tin
V | {z }
=0 (mass continuity eqn.)
on
.O
C
Reynolds’ Lemma
©
d dΨ
Z Z
ρΨ dV = ρ dV . (5.33)
dt dt
Vt ≡V V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Balance Equation. Reynolds Transport Theorem 207
rs
ee
Figure 5.8: Reynolds Transport Theorem.
n
gi
t d le En
The variation of the amount of property A in the material volume Vt , which in-
ar
ar s
stantaneously coincides at time t with the control volume V (Vt ≡ V ), is given by
ge ro or
eS m
ib
expression (5.19) of the material derivative of a volume integral (with µ = ρΨ )
ac
and by (5.11), f
C d P cs
b
i
d ∂ (ρΨ )
an an n
Z Z Z
ρΨ dV = dV + ∇ · (ρΨ v) dV . (5.35)
y ha
dt ∂t
le
Vt ≡V V V
liv or ec
.A
(5.35) results in
m
Reynolds’Z
uu
e
d Lemma dΨ ∂ (ρΨ )
Z Z Z
X Th
ρΨ dV = dV = dV + ∇ · (ρΨ v) dV =
er
ρ
tin
dt dt ∂t
Vt ≡V V V V
on
.O
Divergence Z
Theorem ∂ (ρΨ )
Z
C
= dV + ρΨ v · n dS ,
∂t
©
V ∂V
(5.36)
which can be rewritten as follows.
10The Divergence Theorem provides the following relation between a volume integral and a
surface integral of a tensor A.
Z Z
∇ · A dV = n · A dS ∀V ,
V ∂V
where n is the outward unit normal vector in the boundary of the volume V .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
208 C HAPTER 5. BALANCE P RINCIPLES
rs
The local form of the Reynolds Transport Theorem can be obtained by local-
ee
izing in (5.36),
n
gi
dΨ ∂ (ρΨ )
Z Z Z
t d le En
ρ dV = dV + ∇ · (ρΨ v) dV ∀∆V ⊂ V =⇒
dt ∂t
ar
V V V (5.38)
ar s
ge ro or
eS m
ib
dΨ ∂ (ρΨ )
= + ∇ · (ρΨ v) ∀x ∈ V =⇒
ac
ρ
dt ∂t f
C d P cs
b
i
an an n
(5.39)
le
∂ (ρΨ ) dΨ
liv or ec
=ρ − ∇ · (ρΨ v) ∀x ∈ V
∂t dt
M
.A
m
d
uu
er
tin
.O
that there exists an internal source that generates property A and that this prop-
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
General Expression of the Balance Equations 209
rs
n ee
gi
t d le En
Figure 5.9: An arbitrary control volume used in the definition of the global form of the
ar
ar s
general balance equation.
ge ro or
eS m
ib
ac
f
C d P cs
b
Consider an arbitrary control volume V (see Figure 5.9). Then, the variation
i
an an n
1) the generation of property A per unit of time due to the source term,
le
liv or ec
.A
That is,
uu
e
X Th
er
tin
.O
unit of time
V
C
ρΨ v · n dS = ,
unit of time
∂V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
210 C HAPTER 5. BALANCE P RINCIPLES
and the expression of the balance of the amount of property A in the control
volume V results in
rs
per unit of time generation convective flux non-
ee
convective flux
n
gi
Using the Divergence Theorem and (5.11), the global form of the general
balance equation (5.42) can be written as
t d le En
ar
ar s
ge ro or
Z Z Z Z
∂
eS m
ib
ρΨ dV = ρkA dV − ∇ · (ρΨ v) dV − ∇ · jA dV =⇒
ac
∂t
V V V f V (5.43)
C d P cs
b
Z Z
∂
i
(ρΨ ) + ∇ · (ρΨ v) dV = (ρkA − ∇ · jA ) dV ∀∆V ⊂ V
an an n
∂t
y ha
V V
le
liv or ec
and localizing in (5.43), the local spatial form of the general balance equation
M
.A
m
∂ dΨ
(ρΨ ) + ∇ · (ρΨ v) = = ρkA − ∇ · jA
X Th
ρ
er
tin
∂t dt
| {z } | {z } |{z} | {z } (5.44)
dΨ variation of the variation variation
on
.O
is obtained, where the local form of the Reynolds Transport Theorem (5.39) has
been taken into account.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
General Expression of the Balance Equations 211
rs
ation of this property,
ee
dΨ
ρ = ρkA .
n
dt
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Example 5.2 – Particularize the local spatial form of the general balance
ac
f
equation for the case in which property A is associated with the mass.
C d P cs
b
i
Solution
an an n
y ha
.A
er
tin
.O
dΨ ∂ρ
©
ρ = + ∇ · (ρv) = 0 ,
dt ∂t
which is one of the forms of the mass continuity equation (5.26).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
212 C HAPTER 5. BALANCE P RINCIPLES
rs
ee
dvi d
fi = mi ai = mi = (mi vi ) (5.45)
n
dt dt
gi
The linear momentum of the particle12 is defined as the product of its mass
t d le En
by its velocity (mi vi ). Then, (5.45) expresses that the force acting on the particle
ar
is equal to the variation of the linear momentum of the particle.
ar s
ge ro or
eS m
ib
Applying now Newton’s second law to the discrete system formed by n par-
ac
ticles results in
f
C d P cs
n n n
dvi
bd n
d P (t)
i
R (t) = ∑ fi = ∑ mi ai = ∑ mi = ∑ mi vi = . (5.46)
an an n
dt dt dt
y ha
momentum
M
.A
Note that, again, to obtain the last expression in (5.46), the principle of conser-
m
vation of mass (dmi /dt = 0) has been used. Equation (5.46) expresses that the
d
resultant R of all the forces acting on the discrete system of particles is equal
uu
e
to the variation per unit of time of the linear momentum P of the system. This
X Th
er
tin
.O
C
n
d P (t)
R (t) = 0 ∀t =⇒
dt
=0 =⇒ ∑ mi vi = P = const. ,
i=1
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Balance of Linear Momentum 213
rs
Definition 5.5. Principle of balance of linear momentum. The resul-
ee
tant R (t) of all the forces acting on a material volume of the contin-
n
uous medium is equal to the variation per unit of time of its linear
gi
momentum,
t d le En
d P (t) d
Z
R (t) = = ρ v dV .
dt dt
ar
ar s
ge ro or
Vt
eS m
ib
ac
f
C d P cs
b
The resultant of all the forces acting on the continuous medium defined above
i
an an n
le
Z Z
liv or ec
R (t) = ρb dV + t dS . (5.48)
M
.A
V ∂V
| {z } | {z }
m
body surface
d
forces forces
uu
e
X Th
er
.O
C
d
Z Z Z
(5.49)
ρb dV + t dS = ρv dV
dt
V ∂V Vt ≡V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
214 C HAPTER 5. BALANCE P RINCIPLES
rs
n ee
gi
t d le En
Figure 5.11: Forces acting on a material volume of the continuous medium.
ar
ar s
ge ro or
eS m
ib
ac
5.7.2 Local Form of the Balance of Linear Momentum f
C d P cs
b
i
Using Reynolds’ Lemma (5.33) on (5.49) and introducing the Divergence The-
an an n
orem, results in
y ha
le
d dv
liv or ec
Z Z Z Z
ρv dV = ρb dV + n · σ dS = ρ dV
dt dt
M
.A
|{z}
Vt ≡V V ∂V t Vt ≡V
Divergence Z =⇒ (5.50)
m
Theorem
Z
d
n · σ dS = ∇ · σ dV
uu
e
X Th
V
er
∂V
tin
dv
Z Z
on
.O
V V
©
and, localizing in (5.51), yields the local spatial form of the balance of linear
momentum, also known as Cauchy’s equation13 .
13 The Cauchy equation (already stated, but not deduced, in Chapter 4 ) is, thus, identified as
the local spatial form of the balance of linear momentum.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Balance of Angular Momentum 215
rs
Figure 5.12
Mi is15
ee
dvi
Mi = ri × fi = ri × mi ai = ri × mi (5.53)
n
dt
gi
Extending the previous result to the discrete system formed by n particles, the
t d le En
resultant moment about the origin MO of the forces acting on the system of
particles is obtained as16
ar
ar s
ge ro or
eS m
ib
n n n
ac
MO (t) = ∑ ri × fi = ∑ ri × mi ai = ∑ ri × mi f dvi
C d P cs
dt
i=1 i=1
b i=1
i
an an n
n n n
d dri dvi
y ha
∑ r i × m i v i = ∑ dt ×m i v i + ∑ r i × mi =⇒
dt i=1 dt
le
i=1 |{z} i=1
liv or ec
vi
M
.A
| {z }
(5.54)
=0
m
n
uu
d dL (t)
e
=⇒ MO (t) = ∑ ri × mi vi =
X Th
dt dt
er
tin
i=1
| {z }
Angular
on
.O
momentum L
C
Equation (5.54) expresses that the resultant moment MO of all the forces act-
©
ing on the discrete system of particles is equal to the variation per unit of time
of the moment of linear momentum (or angular momentum), L , of the system.
This postulate is named principle of balance of angular momentum.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
216 C HAPTER 5. BALANCE P RINCIPLES
rs
Result (5.54) can be extended to a continuous and infinite system of particles
ee
(the continuous medium, see Figure 5.13). In such case, the angular momentum
n
is defined as
gi
Z Z
L = r × v d M = r × ρ v dV (5.55)
t d le En
|{z}
M ρ dV V
ar
ar s
ge ro or
eS m
ib
and the continuous version of the postulate of balance of angular momentum is
ac
obtained as follows. f
C d P cs
b
i
an an n
y ha
.A
dL (t)
d
d R
r × ρ v dV
uu
MO (t) = =
e
dt dt Vt ≡V
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Balance of Angular Momentum 217
The resultant moment of the forces acting on the continuous medium (mo-
ment of the body forces and moment of the surface forces) is (see Figure 5.13)
Z Z
MO (t) = r × ρ b dV + r × t dS , (5.56)
V ∂V
then, the global form of the principle of balance of the angular momentum re-
sults in:
rs
ee
d (5.57)
Z Z Z
r × ρ v dV = r × ρ b dV + r × t dS
n
dt
Vt ≡V
gi
V ∂V
t d le En
ar
ar s
ge ro or
eS m
ib
5.8.2 Local Spatial Form of the Balance of Angular Momentum
ac
f
The procedure followed to obtain the local spatial form of the balance equation
C d P cs
is detailed below.
b
i
an an n
le
liv or ec
d d d
Z Z Z
r × ρv dV = ρ (r × v) dV = ρ (r × v) dV =
M
.A
dt dt dt
Vt ≡V Vt ≡V V
m
dr
dv dv
Z Z Z
(5.58)
d
= ρ ×v dV + ρ r × dV = r × ρ dV ,
uu
dt dt dt
e
V V V
X Th
|{z}
er
v
tin
| {z }
=0
on
.O
C
Z Z Z
r × t dS = r × n · σ dS = [r] × [n · σ ]T dS =
|{z}
∂V n·σ ∂V ∂V
(5.59)
DivergenceZ
Theorem
Z
T
r × σ T · ∇ dV ,
= (r × σ ) · n dS =
∂V V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
218 C HAPTER 5. BALANCE P RINCIPLES
r × σ T · ∇ i is computed as
where the component
h i symb ∂ ∂
r × σ T · ∇ = ei jk x j σrk
= ei jk x j σrk =
i |{z} ∂ xr ∂ xr
T
σkr
(5.60)
∂xj ∂ σrk
= ei jk σrk + ei jk x j = ei jk σ jk + [r × ∇ · σ ]i i ∈ {1, 2, 3} .
∂ xr ∂ xr
| {z } | {z }
mi
|{z}
δ jr [r × ∇ · σ ]i
rs
ee
Introducing now (5.60) in (5.59) produces
n
gi
Z Z Z
r × t dS = m dV + (r × ∇ · σ ) dV
t d le En
V V (5.61)
ar
∂V
ar s
ge ro or
eS m
ib
mi = ei jk σ jk i, j, k ∈ {1, 2, 3}
ac
f
C d P cs
b
i
and, finally, replacing (5.58) and (5.61) in (5.57) yields
an an n
y ha
dv
Z Z Z Z
le
r×ρ dV = r × ρb dV + m dV + (r × ∇ · σ ) dV . (5.62)
liv or ec
dt
V V V V
M
.A
Reorganizing the terms in (5.62) and taking into account Cauchy’s equation (5.52)
m
X Th
dv
Z Z
er
tin
r × ∇ · σ + ρb − ρ dV + m dV = 0
dt
V V
on
.O
| {z }
=0 (5.63)
C
Z
=⇒ m dV = 0 ∀∆V ⊂ V .
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Power 219
rs
(5.65)
ee
which results in the local spatial form of the balance of angular momentum
n
translating into the symmetry of the Cauchy stress tensor17 .
gi
t d le En
Local spatial form of the
ar
ar s
principle of balance of angular momentum
ge ro or
(5.66)
eS m
ib
σ = σT
ac
f
C d P cs
b
i
an an n
y ha
le
5.9 Power
liv or ec
M
.A
m
er
that can be quantified as the ability to perform work per unit of time.
tin
Then, for a system (or continuous medium) the power W (t) entering
on
.O
unit of time
In some cases, but not in all, the power W (t) is an exact differential of a function
E (t) that, in said cases, receives the name of energy,
d E (t)
W (t) = . (5.67)
dt
17 The symmetry of the Cauchy stress tensor (already stated, but not deduced, in Chapter 4 )
is, thus, identified as the local spatial form of the balance of angular momentum.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
220 C HAPTER 5. BALANCE P RINCIPLES
Here, it is assumed that there exist two procedures by which the continuous
medium absorbs power from the exterior and performs work per unit of time
with this power
− Mechanical power, by means of the work performed by the mechanical
actions (body and surface forces) acting on the medium.
− Thermal power, by means of the heat entering the medium.
rs
n ee
Definition 5.8. The mechanical power entering the continuous
gi
medium, Pe , is the work per unit of time performed by all the (body
t d le En
and surface) forces acting on the medium.
ar
ar s
ge ro or
eS m
ib
ac
f
Consider the continuous medium shown in Figure 5.14 is subjected to the ac-
C d P cs
tion of body forces, characterized by the vector of body forces b (x,t), and of
b
i
surface forces, characterized by the traction vector t (x,t). The expression of the
an an n
le
liv or ec
Z Z Z Z
Pe = ρ b · v dV + t · v dS = ρ b · v dV + n · (σ
σ · v) dS . (5.68)
M
.A
|{z}
V ∂V n·σ V ∂V
m
d
uu
e
X Th
er
tin
on
.O
C
dr
©
ρb · dV = ρb · vdV
dt
dr
t· dS = t · vdS
dt
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Power 221
rs
σ : l = σ : d+σ : w = σ : d . (5.70)
ee
|{z} | {z }
d+w =0
n
gi
Replacing (5.70) in (5.69) yields
t d le En
Z Z Z
n · (σ
σ · v) dS = (∇ · σ ) · v dV + σ : d dV . (5.71)
ar
ar s
ge ro or
eS m
ib
∂V V V
ac
f
C d P cs
Introducing (5.71) in (5.68), the mechanical power entering the continuous
results in19 Z b
i
medium
an an n
Z Z Z Z
y ha
Pe = ρ b · v dV + t · v dS = ρ b · v dV + (∇ · σ ) · v dV + σ : d dV =
le
liv or ec
Z V ∂V V V V
dv
Z Z Z
M
.A
= (∇ · σ + ρ b) · v dV + σ : d dV = ρ · v dV + σ : d dV =
dt
m
Z V Z V Z V VZ
d
d 1 d 1 2
uu
ρ v · v dV + σ : d dV = ρ v dV + σ : d dV .
e
dt 2 dt 2
X Th
er
V V V V
tin
(5.72)
And applying Reynolds’ Lemma (5.33) in (5.72), the mechanical power entering
on
.O
d 1 2
Z Z Z Z
Pe = ρ b · v dV + t · v dS = ρv dV + σ : d dV
dt 2 (5.73)
|{z} V ∂V Vt ≡V V
mechanical | {z } | {z }
power entering K = kinetic stress
the medium energy power
18 The tensor σ is symmetric and the tensor w is antisymmetric. Consequently, their product
σ : w = 0.
is null,σ
19 The expression d 1 1 dv 1 dv dv
v·v = ·v+ v· = · v is used here, in addition to the
dt 2 2 dt 2 dt dt
notation v · v = |v|2 = v2 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
222 C HAPTER 5. BALANCE P RINCIPLES
Definition 5.9. The balance of mechanical energy states that the me-
chanical energy entering the continuous medium,
Z Z
Pe = ρ b · v dV + t · v dS
V ∂V
is invested in:
a) modifying the kinetic energy of the particles in the continuous
rs
medium,
dK
ee
1 2 d 1 2
Z Z
not
kinetic energy = K = ρ v dV =⇒ = ρ v dV .
2 dt dt 2
n
V V
gi
b) creating stress power,
t d le En
Z
de f
stress power = σ : d dV .
ar
ar s
ge ro or
V
eS m
ib
ac
f
C d P cs
b
i
Remark 5.8. Considering (5.73), the stress power can be defined as
an an n
y ha
the part of the mechanical power entering the system that is not used
in modifying the kinetic energy. It can be interpreted as the work per
le
liv or ec
.A
fore, the stresses do not perform mechanical work and the stress
d
uu
power is null. In this case, all the mechanical power entering the
e
X Th
.O
C
The heat entering the medium can be produced by two main causes:
a) Heat entering the medium due to the (non-convective) heat flux across the
boundary corresponding to the material volume. Note that, since the vol-
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Power 223
ume is a material volume, the heat flux due to mass transport (convective) is
null and, thus, all the heat flux entering the medium will be non-convective.
b) The existence of heat sources inside the continuous medium.
• Non-convective heat flux
Consider the spatial description of the vector of non-convective heat flux
per unit of surface, q (x,t). Then, the net non-convective heat flux across
the boundary of the material volume is (see Figure 5.15)
amount of heat exiting the medium
Z
rs
q · n dS =
unit of time
ee
∂V (5.74)
n
amount of heat entering the medium
Z
− q · n dS =
gi
unit of time
t d le En
∂V
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 5.9. A typical example of non-convective flux is heat trans-
C d P cs
fer by conduction phenomena. Heat conduction is governed by
b
i
Fourier’s Law, which provides the vector of heat flux by (non-
an an n
le
liv or ec
o
Fourier’s Law of q (x,t) = −K∇θ (x,t) ,
heat conduction
M
.A
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
224 C HAPTER 5. BALANCE P RINCIPLES
rs
ee
Figure 5.16: Internal heat sources.
n
gi
t d le En
• Internal heat sources
ar
ar s
ge ro or
eS m
ib
Heat can be generated (or absorbed) in the interior of the continuous
ac
f
medium due to certain phenomena (chemical reactions, etc.). Consider a
C d P cs
b
scalar function r (x,t) that describes in spatial form the heat generated by
i
the internal sources per unit of mass and unit of time (see Figure 5.16).
an an n
y ha
Then, the heat entering the system, per unit of time, due to the existence of
internal heat sources is
le
liv or ec
.A
ρr dV = . (5.75)
unit of time
m
V
d
uu
Consequently, the total heat entering the continuous medium per unit of
e
X Th
time (or thermal power Qe ) can be expressed as the sum of the contributions
er
tin
.O
R R
Qe = ρr dV − q · n dS . (5.76)
the medium
©
V ∂V
Then, considering (5.73) and (5.76), the total power entering the continu-
ous medium can be written as follows.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Energy Balance 225
rs
• Thermodynamic state: a thermodynamic state is defined when a certain
ee
value is assigned to the state variables and, therefore, to all the thermo-
n
dynamic variables. In a hyperspace (thermodynamic space) defined by the
gi
thermodynamic variables µi i ∈ {1, 2, ... , n} (see Figure 5.17), a thermo-
t d le En
dynamic state is represented by a point.
ar
• Thermodynamic process: the energetic development of a thermodynamic
ar s
ge ro or
eS m
ib
system that undergoes successive thermodynamic states, changing from an
ac
f
initial state at time tA to a final state at time tB (it is a path or continuous
C d P cs
segment in the thermodynamic space), see Figure 5.18.
b
i
• Closed cycle: A thermodynamic process in which the final thermodynamic
an an n
y ha
state coincides with the initial thermodynamic state (all the thermodynamic
variables recover their initial value), see Figure 5.19.
le
liv or ec
• State function: any scalar, vector or tensor function φ (µ1 , ... , µn ) of the
M
.A
i ∈ {1, 2, ... , n} and a function φ (µ1 , ... , µn ) of said variables implicitly defined
X Th
er
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
226 C HAPTER 5. BALANCE P RINCIPLES
rs
Figure 5.18: Thermodynamic space. Figure 5.19: Closed cycle.
n ee
Consider also a given thermodynamic process A → B in the space of the
gi
thermodynamic variables. Equation (5.78) provides the value of the function
t d le En
not not
φ (µ1B , ... , µnB ) = φB when its value φ (µ1A , ... , µnA ) = φA and the corresponding
ar
path (thermodynamic process) A → B are known by means of
ar s
ge ro or
eS m
ib
ac
ZBf
C d P cs
φB = φA +
b δφ . (5.79)
i
an an n
A
y ha
However, (5.79) does not guarantee that the result φB is independent of the path
le
liv or ec
.A
that the function φ : Rn → R defined by (5.79) is univocal (see Figure 5.20) and,
thus, that there exists a single image φ (µ1 , ... , µn ) corresponding to each point
m
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Energy Balance 227
rs
ee
δ φ = f1 (µ1 , ... , µn ) dµ1 + . . . + fn (µ1 , ... , µn ) dµn
⇔ δ φ = dφ .
n
∂ fi (µ1 , ... , µn ) ∂ f j (µ1 , ... , µn )
= ∀i, j ∈ {1, ... , n}
gi
∂ µj ∂ µi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
If the differential form (5.78) is an exact differential, (5.79) results in
i
an an n
ZB
y ha
h iB
φB = φA + dφ = φA + ∆ φ (5.80)
le
liv or ec
A
A
M
.A
and the value φB is independent of the integration path. Then, function φ is said
m
to be a state function that depends only on the values of the state variables and
d
uu
er
tin
on
.O
and the integral along the complete closed cycle of the differential
©
δ φ is null,
ZA I h iA
δ φ = dφ = ∆ φ = 0 .
A | {z A}
=0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
228 C HAPTER 5. BALANCE P RINCIPLES
Solution
Following (5.78),
∂ f1
f1 ≡ 4µ2 =4 ∂ f1 ∂ f2
=⇒ ∂ µ2 =⇒ 6=
f2 ≡ µ1 ∂ f2 ∂ µ2 ∂ µ1
=1
∂ µ1
rs
Then, δ φ is not an exact differential (see Remark 5.10) and φ is not a state
ee
function.
n
gi
t d le En
5.10.2 First Law of Thermodynamics
ar
ar s
ge ro or
Experience shows that the mechanical power (5.73) is not an exact differential
eS m
ib
and, therefore, the mechanical work performed by the system in a closed cycle
ac
f
C d P cs
is not null. The same happens with the thermal power (5.76).
b
i
an an n
I
δ φ1 = Pe dt =⇒ Pe dt 6= 0
y ha
I (5.81)
le
liv or ec
δ φ2 = Qe dt =⇒ Qe dt 6= 0
M
.A
However, there exists experimental evidence that proves that the sum of the me-
m
chanical and thermal powers, that is, the total power entering the system (5.77)
d
(see Figure 5.21), is, in effect, an exact differential and, thus, a state function E
uu
e
er
tin
Zt
on
.O
t0
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Energy Balance 229
dE
= Pe + Qe
dt
dE = Pe dt + Qe dt (5.83)
|{z} |{z} | {z }
rs
Variation of Mechanical Thermal
total energy work work
n ee
2) There exists another state function U , named internal energy of the system,
gi
such that
t d le En
a) It is an extensive property21 . Then, a specific internal energy u (x,t)
(or internal energy per unit of mass) can be defined as
ar
ar s
ge ro or
eS m
ib
Z
ac
U= f ρu dV . (5.84)
C d P cs
b V
i
an an n
y ha
b) The variation of the total energy of the system E is equal to the sum of
the variation of the internal energy U and the variation of the kinetic
le
liv or ec
energy K.
M
.A
dE = dK + dU
(5.85)
m
|{z} |{z}
d
Exact Exact
uu
differential differential
e
X Th
er
tin
Remark 5.12. Note that, since the total energy E and the internal en-
on
.O
21 A certain property is extensive when the complete content of the property is the sum of the
content of the property in each of its parts. An extensive property allows defining the content
of this property per unit of mass (specific value of the property) or per unit of volume (density
of the property).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
230 C HAPTER 5. BALANCE P RINCIPLES
dE dK dU d 1 2
Z Z Z Z
= + = ρv dV + σ : d dV + ρr dV − q · n dS
rs
dt dt dt dt 2
V V V ∂V
ee
| {z } | {z }
dK dU
n
gi
dt dt
(5.86)
t d le En
ar
ar s
ge ro or
eS m
ib
ac
Global form of the internal energy balance f
C d P cs
dU d
b (5.87)
Z Z Z Z
i
ρr dV − q · n dS
an an n
= ρu dV = σ : d dV +
dt dt
y ha
Vt ≡V V V ∂V
le
liv or ec
M
.A
Remark 5.13. From (5.87) it follows that any variation per unit of
m
Z
er
tin
.O
− aZ variation per
Z unit of time of the content of heat in the medium,
C
ρr dV − q · n dS.
©
V ∂V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Reversible and Irreversible Processes 231
Finally, localizing in (5.88) results in the local spatial form of the internal energy
balance.
rs
5.11 Reversible and Irreversible Processes
n ee
The first law of thermodynamics leads to a balance equation that must be ful-
gi
filled for all the physical processes that take place in reality,
t d le En
dE dU dK
ar
Pe + Qe = = + . (5.90)
ar s
ge ro or
dt dt dt
eS m
ib
ac
f
In particular, if an isolated system22 is considered, the time variation of the total
C d P cs
b
energy of the system will be null (d E /dt = 0 ⇒ the total energy is conserved).
i
an an n
Therefore, the energy balance equation (5.90), established by the first law of
y ha
.A
What the first law of thermodynamics does not establish is whether this (ki-
netic and internal) energy exchange in an isolated system can take place equally
m
in both directions or not (d U /dt = −d K/dt > 0 or d U /dt = −d K/dt < 0). That
d
uu
e
is, it does not establish any restriction that indicates if an imaginary and arbitrary
X Th
er
tin
on
.O
C
22 An isolated thermodynamic system is a system that cannot exchange energy with its
exterior. In a strict sense, the only perfectly isolated system is the universe, although one can
think of quasi-isolated or imperfectly isolated smaller systems.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
232 C HAPTER 5. BALANCE P RINCIPLES
rs
− a brake that can be applied on the wheel at
ee
a certain instant of time.
n
Figure 5.23
gi
t d le En
Consider now the following two processes:
ar
1) At a certain instant of time the brake acts, the rotation speed of the wheel ω
ar s
ge ro or
eS m
ib
decreases and, thus, so does its kinetic energy (d K < 0). On the other hand,
ac
f
due to the friction between the brake and the wheel, heat is generated and
C d P cs
there is an increase of the internal energy (d U > 0). Experience shows
b
i
that this process, in which the internal energy increases at the expense of
an an n
decreasing the kinetic energy23 , can take place in reality and, therefore, is
y ha
2) Maintaining the brake disabled, at a certain instant of time the wheel spon-
M
.A
taneously increases its rotation speed ω and, thus, its kinetic energy in-
creases (d K > 0). According to the first law of thermodynamics, the in-
m
ternal energy of the system will decrease (d U < 0). However, experience
uu
e
shows that this (spontaneous) increase of speed never takes place, and nei-
X Th
er
tin
ther does the decrease in the amount of heat of the system (which would be
reflected in a decrease in temperature).
on
.O
the example is not a feasible physical process. More generally, only thermo-
©
dynamic processes that tend to increase the internal energy and decrease the
kinetic energy, and not the other way round, are feasible for the system under
consideration.
It is concluded, then, that the first law of thermodynamics is only applicable
when a particular physical process is feasible, and the need to determine when a
particular physical process is feasible, or if a physical process is feasible in one
direction, in both or in none, is noted. The answer to this problem is provided
by the second law of thermodynamics.
23 The wheel, being a non-deformable medium, has null stress power (see Remark 5.8) and
all the variation of internal energy of the system derives from a variation of its heat content
(see Remark 5.13).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Second Law of Thermodynamics. Entropy 233
rs
ee
The previous considerations lead to the classification, from a thermodynamic
n
gi
point of view, of the possible physical processes in feasible and non-feasible pro-
cesses and, in addition, suggest classifying the feasible processes into reversible
t d le En
and irreversible processes.
ar
ar s
ge ro or
eS m
ib
ac
f
Definition 5.11. A thermodynamic process A → B is a reversible
C d P cs
b
process when it is possible to return from the final thermodynamic
i
an an n
state B to the initial thermodynamic state A along the same path (see
y ha
Figure 5.24).
le
A thermodynamic process A → B is an irreversible process when
liv or ec
.A
d
uu
e
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
234 C HAPTER 5. BALANCE P RINCIPLES
2) There exists a state function named entropy S with the following character-
istics:
a) It is an extensive variable. This implies that there exists a specific
entropy (entropy per unit of mass) s such that
entropy
Z
s= =⇒ S= ρs dV . (5.91)
unit of mass
V
b) The inequality
rs
Integral form of the second law of thermodynamics
ee
dS d r q
Z Z Z
(5.92)
n
= ρs dV ≥ ρ dV − · n dS
dt dt
gi
θ θ
Vt ≡V V ∂V
t d le En
ar
is satisfied, where:
ar s
ge ro or
eS m
ib
− The sign = corresponds to reversible processes.
ac
f
− The sign > corresponds to irreversible processes.
C d P cs
b
− The sign < cannot occur and indicates that the corresponding pro-
i
an an n
le
5.12.2 Physical Interpretation of the Second Law of Thermodynamics
liv or ec
M
.A
.O
These terms allow computing the amount of heat per unit of time entering a
C
material volume Vt , which at a certain instant of time occupies the spatial volume
©
Consider now a new magnitude defined as heat per unit of absolute temper-
ature in the system. If θ (x,t) is the absolute temperature, the amount of said
magnitude will be characterized by
a) A source term r/θ corresponding to the generation of heat per unit of ab-
solute temperature, per unit of mass and unit of time.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Second Law of Thermodynamics. Entropy 235
b) A non-convective flux vector q/θ of the heat per unit of absolute temper-
ature.
rs
unit of time θ θ
n ee
Similarly to (5.93), the new source term r/θ and non-convective flux vec-
gi
tor q/θ allow computing the amount of heat per unit of absolute temperature
t d le En
entering the material volume per unit of time as
ar
ar s
ge ro or
eS m
(heat/unit of temperature) entering V q
ib
r
Z Z
= ρ dV − · n dS . (5.94)
ac
unit of time f θ θ
C d P cs
V ∂V
b
i
an an n
Observing now (5.94), the second term in this expression is identified as the
y ha
.A
heat per unit of temperature entering the system per unit of time.
m
d
uu
e
X Th
dS r q
Z Z
on
≥ dV − · n dS
.O
ρ
dt θ θ (5.95)
C
V ∂V
©
| {z }
amount of the property
“heat / unit of absolute temperature”
entering the domain V per unit of time
Consider now the decomposition of the total entropy of the system S into two
distinct components:
• S(i) : entropy generated (produced) internally by the continuous medium. Its
generation rate is dS(i) /dt.
• S(e) : entropy generated by the interaction of the continuous medium with
its exterior. Its variation rate is dS(e) /dt.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
236 C HAPTER 5. BALANCE P RINCIPLES
dS dS(e) dS(i)
= + (5.96)
dt dt dt
Now, if one establishes that the variation rate of the entropy generated by
the interaction with the exterior coincides with the magnitude heat per unit of
absolute temperature in (5.93),
dS(e) r q
Z Z
rs
= ρ dV − · n dS (5.97)
dt θ θ
ee
V ∂V
n
gi
and, taking into account (5.95) to (5.97), the variation per unit of time of the
t d le En
internally generated entropy results in
ar
ar s
ge ro or
eS m
ib
dS(i) dS dS(e) dS r q
Z Z
= − = − ρ dV − · n dS ≥ 0 . (5.98)
ac
dt dt dt dt θ θ f
C d P cs
b
V ∂V
i
an an n
y ha
le
Remark 5.14. According to (5.98), the internally generated en-
liv or ec
.A
with the exterior and the variation of entropy due to interaction with
uu
e
the exterior is null, (dS(e) /dt = 0). In this case, the second law of
X Th
er
tin
.O
= ≥0
dt dt
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Second Law of Thermodynamics. Entropy 237
rs
Z Z
ee
S(i) = ρs(i) dV and S(e) = ρs(e) dV (5.101)
n
V V
gi
t d le En
dS dS(i) dS(e)
S = S(i) + S(e) = =⇒
+ (5.102)
ar
dt dt dt
ar s
ge ro or
eS m
ib
and introducing Reynolds’ Lemma (5.33) in (5.102) yields
ac
f
C d P cs
dS(i) d
b ds(i)
Z Z
ρs(i) dV =
i
= ρ dV ,
an an n
dt dt dt
y ha
Vt ≡V V
(5.103)
le
dS(e) d ds(e)
Z Z
liv or ec
(e)
= ρs dV = ρ dV .
dt dt dt
M
.A
Vt ≡V V
m
exterior) is associated with the variation of the magnitude heat per unit of
e
X Th
dS(e) r q
Z Z
on
.O
= ρ dV − · n dS . (5.104)
dt θ θ
C
V ∂V
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
238 C HAPTER 5. BALANCE P RINCIPLES
dS(i) dS dS(e)
= − ≥0
dt dt dt
! (5.106)
d d r q
Z Z Z Z
ρs(i) dV = ρs dV − ρ dV − · n dS ≥0
dt dt θ θ
rs
Vt ≡V Vt ≡V V ∂V
ee
Applying Reynolds’ Lemma (5.33) (on the first and second integral of the left-
n
hand term in (5.106)) and the Divergence Theorem (on the last term) yields
gi
!
t d le En
ds(i) ds r q
Z Z Z Z
ρ dV = ρ dV − ρ dV − ∇ · dV ≥ 0 ∀∆V ⊂ V
ar
dt dt θ θ
ar s
ge ro or
V V V V
eS m
ib
(5.107)
ac
f
and localizing in (5.107), the local form of the second law of thermodynamics
C d P cs
or Clausius-Duhem equation is obtained.
b
i
an an n
y ha
(Clausius-Duhem inequality)
(5.108)
M
.A
ds(i) ds r q
ρ = ρ − ρ −∇· ≥0 ∀x ∈ V, ∀t
dt dt
m
θ θ
d
uu
e
X Th
er
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Second Law of Thermodynamics. Entropy 239
. . r 1 1
s(i) = s − + ∇·q − 2 q · ∇θ ≥ 0 (5.110)
θ ρθ ρθ
| {z } | {z }
.(i) .(i)
slocal scond
rs
n ee
Local internal generation of entropy
gi
(Clausius-Planck inequality)
(5.111)
t d le En
.(i) . r 1
slocal = s − + ∇·q ≥ 0
ar
θ ρθ
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
Internal generation of entropy by heat conduction
i
an an n
y ha
.(i) 1 (5.112)
scond = − q · ∇θ ≥ 0
le
ρθ 2
liv or ec
M
.A
m
d
uu
e
er
tin
.O
q · ∇θ ≤ 0 ,
C
which establishes that the non-convective heat flux, q, and the tem-
perature gradient, ∇θ , are vectors that have opposite directions (their
dot product is negative). In other words, (5.112) is the mathemati-
cal expression of the experimentally verified fact that heat flows by
conduction from the hottest to the coldest parts in the medium (see
Figure 5.24), characterizing as non-feasible those processes in which
the contrary occurs.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
240 C HAPTER 5. BALANCE P RINCIPLES
rs
ee
Figure 5.25: Heat flux is opposed to the thermal gradient.
n
gi
t d le En
ar
ar s
ge ro or
Remark 5.16. In the context of Fourier’s Law of heat conduction,
eS m
ib
q = −K ∇θ (see Remark 5.9), expression (5.112) can be written as
ac
f
C d P cs
b
)
q · ∇θ ≤ 0
i
=⇒ −K |∇θ |2 ≤ 0 =⇒ K ≥ 0
an an n
q = −K∇θ
y ha
le
liv or ec
physical meaning.
.A
m
d
uu
e
X Th
er
tin
.O
tinuum mechanics.
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Second Law of Thermodynamics. Entropy 241
rs
ee
de f
ψ = u − sθ . (5.116)
n
gi
t d le En
Differentiating (5.116) with respect to time results in
. . . . . . . .
ar
ar s
ψ = u − sθ − sθ =⇒ u − θ s = ψ + sθ (5.117)
ge ro or
eS m
ib
ac
f
and, replacing (5.117) in (5.115), yields the Clausius-Planck equation in terms
C d P cs
b
of the Helmholtz free energy,
i
an an n
.(i) . . . .
y ha
ρθ slocal = −ρ (u − θ s) + σ : d = −ρ ψ + sθ + σ : d ≥ 0 . (5.118)
le
liv or ec
M
.A
−ρ ψ + sθ + σ : d ≥ 0
d
uu
e
X Th
.
er
tin
For the infinitesimal strain case, d = ε (see Chapter 2, Remark 2.22), and re-
placing in (5.119) results in
on
.O
C
. . . . (5.120)
−ρ ψ + sθ + σ : ε ≥ 0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
242 C HAPTER 5. BALANCE P RINCIPLES
rs
dρ ∂ vi
+ρ =0
ee
dt ∂ xi
n
gi
t d le En
2) Balance of linear momentum. Cauchy’s equation.
ar
ar s
ge ro or
dv
eS m
ib
∇ · σ + ρb = ρ
ac
f
dt → 3 equations (5.122)
C d P cs
b
∂ σ ji dvi
+ ρbi = ρ i ∈ {1, 2, 3}
i
an an n
∂xj dt
y ha
le
liv or ec
.A
)
m
σ = σT
d
→ 3 equations (5.123)
uu
er
tin
on
.O
©
du
= σ : d + (ρr − ∇ · q)
ρ
dt
→ 1 equation (5.124)
du ∂ qi
= σi j di j + ρr −
ρ
dt ∂ xi
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Continuum Mechanics Equations. Constitutive Equations 243
. .
−ρ (u − θ s) + σ : d ≥ 0
→ 1 restriction
. .
−ρ (u − θ s) + σ d ≥ 0
ij ij
1 (5.125)
− 2 q · ∇θ ≥ 0
ρθ → 1 restriction
1 ∂θ
− 2 qi ≥0
rs
ρθ ∂ xi
ee
These add up to a total of 8 partial differential equations (PDEs) and two re-
n
strictions. Counting the number of unknowns that intervene in these equations
gi
results in25
t d le En
ρ → 1 unknown
ar
ar s
v → 3 unknowns
ge ro or
eS m
ib
ac
σ → 9 unknowns
f
C d P cs
b
u → 1 unknown 19 unknowns
i
an an n
q → 3 unknowns
y ha
θ → 1 unknown
le
liv or ec
s → 1 unknown
M
.A
Therefore, it is obvious that additional equations are needed to solve the prob-
m
lem. These equations, which receive the generic name of constitutive equations
d
uu
and are specific to the material that constitutes the continuous medium, are
e
X Th
er
tin
.O
q = −K ∇θ
C
∂θ → 3 equations (5.126)
©
qi = −K i ∈ {1, 2, 3}
∂ xi
25The six components of the strain rate tensor d in (5.124) and (5.125) are not considered
unknowns because they are assumed to be implicitly calculable in terms of the velocity v by
means of the relation d (v) = ∇s v (see Chapter 2, Section 2.13.2).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
244 C HAPTER 5. BALANCE P RINCIPLES
rs
where µ = µ1 , ... , µ p are a set of new thermodynamic variables (p new
ee
unknowns) introduced by the thermo-mechanical constitutive equations.
n
gi
8) Thermodynamic equations of state.
t d le En
Caloric
o
ar
ar s
u = g (ρ, ε (v) , θ , µ )
ge ro or
eqn. of state
eS m
ib
→ (1 + p) eqns.
ac
Kinetic
o
Fi (ρ, θ , µ ) = 0 f
i ∈ {1, 2, ... , p}
C d P cs
b
eqns. of state
i
(5.128)
an an n
lem.
M
.A
m
symmetry of the stress tensor, the energy balance and the inequalities
e
X Th
are valid and general for all the continuous medium, regardless of the
material that constitutes the medium, and for any range of displace-
on
.O
being studied (solid, fluid, gas) and differentiate them from one an-
other.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Continuum Mechanics Equations. Constitutive Equations 245
rs
stitutive equations (5.127), and that, in addition, said constitutive equations do
ee
not involve new thermodynamic variables (µ µ = {0})./ In this case, the following
n
set of equations are considered 27
gi
t d le En
dρ
Mass continuity
+ ρ∇ · v = 0
(1 eqn)
ar
ar s
equation:
dt
ge ro or
eS m
ib
ac
f
dv
(3 eqn) → 10 equations ,
C d P cs
Cauchy’s equation: ∇ · σ + ρb = ρ
dt
b
i
an an n
y ha
fi (σ
σ , ε (v)) = 0
Mechanical
(6 eqn)
constitutive equations: i ∈ {1, ... , 6}
le
liv or ec
(5.129)
M
.A
d
ρ (x,t) → 1 unknown
uu
e
er
tin
σ (x,t) → 6 unknowns
on
.O
The problem defined by equations (5.129) and (5.130) constitutes the so-
C
called mechanical problem, which involves the variables (5.130) (named me-
©
chanical variables) that, moreover, are the real interest in many engineering
problems.
The mechanical problem constitutes, in this case, a system of reduced differ-
ential equations, with respect to the general problem, and can be solved inde-
pendently of the rest of equations of said problem.
27For simplicity, it is assumed that the symmetry of the stress tensor (5.123) is already
imposed. Then this equation is eliminated from the set of equations and the number of un-
knowns of σ is reduced from 9 to 6 components.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
246 C HAPTER 5. BALANCE P RINCIPLES
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 247
P ROBLEMS
Problem 5.1 – Justify whether the following statements are true or false.
a) The mass flux across a closed material surface is null only when the
motion is stationary.
b) The mass flux across a closed control surface is null when this flux is
rs
stationary.
n ee
gi
t d le En
Solution
ar
a) The statement is false because a material surface is always constituted by
ar s
ge ro or
eS m
ib
the same particles and, therefore, cannot be crossed by any particle throughout
ac
f
its motion. For this reason, the mass flux across a material surface is always null,
C d P cs
independently of the motion being stationary or not.
b
i
an an n
b) The statement is true because the application of the mass continuity equation
y ha
M
.A
∂ρ
Mass continuity equation =⇒ + ∇ · (ρv) = 0
m
∂t =⇒ ∇ · (ρv) = 0 .
d
uu
e
∂ρ
Stationary flux =⇒ =0
X Th
er
tin
∂t
on
Z Z
∇ · (ρv) = 0 =⇒ ∇ · (ρv) dV = ρv · n dS = 0 .
©
V ∂V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
248 C HAPTER 5. BALANCE P RINCIPLES
rs
n ee
gi
t d le En
ar
Solution
ar s
ge ro or
eS m
ib
Taking into account the Reynolds Transport Theorem (5.39) and that the prob-
ac
f
C d P cs
lem is in steady-state regime, the forces acting on the fluid are
d R b
i
R ∂
an an n
R R
∑ Fext/ f = ρv dV = (ρv) dV + ρv (n · v) dS = ρv (n · v) dS .
dt V
y ha
V ∂t ∂V S
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Note that the velocity vector of the fluid along the surfaces Slat−1 and Slat−3 is
perpendicular to the outward unit normal vector of the volume that encloses the
fluid, therefore, v · n = 0. The same happens in the walls of the disc.
The vectors v and n in sections S2 and S4 are not perpendicular but, because
there exists symmetry and v is perpendicular to F, they do not contribute com-
ponents to the horizontal forces. Therefore, the only forces acting on the fluid
are Z Z
F
∑ ext/ f = ρv (n · v) dS = ρve (−e · ve) dS = −ρv2 Se .
∂V S
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 249
On the other hand, the external force, the pressure of the water jet and the atmo-
spheric pressure (which is negligible) also act on the fluid,
rs
ee
Problem 5.3 – A volume flow rate Q circulates, in steady-state regime, through
n
a pipe from end A (with cross-section SA ) to end B (with cross-section SB < SA ).
gi
The pipe is secured at point O by a rigid element P − O.
t d le En
Determine:
ar
ar s
ge ro or
eS m
ib
a) The entry and exit velocities vA and vB in terms of the flow rate.
ac
f
b) The values of the angle θ that maximize and minimize the reaction force
C d P cs
b
F at O, and the corresponding values of said reaction force.
i
an an n
c) The values of the angle θ that maximize and minimize the reaction mo-
y ha
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Hypotheses:
1) The water is a perfect fluid (σi j = −p δi j ) and incompressible.
2) The weight of the pipe and the water are negligible.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
250 C HAPTER 5. BALANCE P RINCIPLES
Solution
a) The incompressible character of water implies that the density is constant
for a same particle and, therefore, dρ/dt = 0. Introducing this into the mass
continuity equation (5.24), results in
Z
∇·v = 0 ⇐⇒ ∇ · v dV = 0 ∀V . [1]
V
The adequate integration volume must now be defined. To this aim, a control
rs
volume such that its boundary is a closed surface must be found (S = ∂V ) to be
able to apply the Divergence Theorem,
n ee
Z Z
gi
∇ · v dV = n · v dS ∀V [2]
t d le En
V ∂V
ar
where n is the outward unit normal vector in the boundary of the volume V .
ar s
ge ro or
eS m
ib
Then, by means of [1] and [2], the conclusion is reached that the net outflow
ac
across the contour of the control volume is null, f
C d P cs
Z
b
i
an an n
n · v dS = 0 ∀V .
y ha
∂V
le
liv or ec
The volume the defined by the water contained inside the pipe between the cross-
M
.A
direction of the flow of water. Then, the following expression is deduced. Note
uu
e
SA and SB since n · v = 0 on the walls of the pipe, that is, n and v are perpendic-
er
tin
.O
Z Z Z Z Z
C
n · v dS = n · v dS + n · v dS = (−eA ) · vA eA dS + eB · vB eB dS = 0
©
∂V SA SB SA SB
=⇒ −vA SA + vB SB = 0 =⇒ vA SA = vB SB = Q
It is verified, thus, that the flow rate at the entrance and exit of the pipe are the
same,
Q Q
vA = ; vB = . [3]
SA SB
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 251
b) The balance of linear momentum equation (5.49) must be applied to find the
value of the force F,
d
Z Z Z
R= ρb dV + t dS = ρv dV , [4]
dt
V ∂V V
where R is the total resultant of the forces acting on the fluid. On the other
hand, expanding the right-hand term in [4] by means of the Reynolds Transport
Theorem (5.39), yields
rs
d
Z Z Z
∂
ρv dV = ρv dV + ρv (n · v) dS . [5]
ee
dt ∂t
n
V V ∂V
gi
The problem is being solved for a steady-state regime, i.e., the local derivative
t d le En
of any property is null. In addition, the flow is known to exist solely through
sections SA and SB since n and v are perpendicular to one another on the walls
ar
ar s
ge ro or
of the pipe. Therefore, according to [4] and [5],
eS m
ib
ac
f
C d P cs
Z Z
R= ρv (n · v) dS +
b ρv (n · v) dS =
i
an an n
SA SB
y ha
Z Z
le
liv or ec
.A
SA SB
m
er
tin
.O
2 1 1
C
R = −ρQ − eA + eB .
SA SB
©
Now the different forces that compose R must be analyzed. According to the
statement of the problem, body forces can be neglected (b = 0). Therefore, only
surface forces must be taken into account, that is, the forces applied on the
boundary of the control volume (SA , SB and Slat , where this last one corresponds
to the lateral surface of the walls),
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
252 C HAPTER 5. BALANCE P RINCIPLES
Z Z Z Z Z Z
R= ρb dV + t dS = t dS = t dS + t dS + t dS =
V ∂V ∂V SA SB Slat
Z Z
= pA eA dS + pB (−eB ) dS + R p/ f .
SA SB
Here, R p/ f represents the forces exerted on the fluid by the walls of the pipe,
which initially are unknown but can be obtained using [6] as follows.
Z Z
rs
R p/ f = R − pA eA dS − pB (−eB ) dS
ee
SA SB
n
gi
R p/ f = −ρv2A SA eA + ρv2B SB eB − pA SA eA + pB SB eB
t d le En
ar
R p/ f = − ρv2A + pA SA eA − ρv2B + pB SB eB
ar s
[7]
ge ro or
eS m
ib
ac
f
C d P cs
b
Introducing [3], R p/ f can be expressed in terms of Q,
i
an an n
2 2
y ha
Q Q
R p/ f =− ρ + p A S A eA − ρ + pB SB eB .
le
SA SB
liv or ec
M
.A
Now the relation between R p/ f and the unknown being sought, F, must be
found. To this aim, the action and reaction law is considered, and the pipe and
m
the rigid element P − O are regarded as a single body. Under these conditions,
d
uu
er
tin
R f /p = −R p/ f .
on
.O
Since it is the only action on the body, and taking into account that the weight
C
R f /p + F = 0 =⇒ F = −R f /p = R p/ f
Introducing [7], the value of F is finally obtained as
F = − ρv2A + pA SA eA + ρv2B + pB SB eB .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 253
There are two possible ways of obtaining the maximum and minimum of |F| in
terms of θ :
1) Determine the expression of |F| and search for its extremes by imposing
that its derivative is zero (this option not recommended).
rs
2) Direct method, in which the two vectors acting in the value of F are
ee
analyzed (this option developed below).
n
gi
According to [7], the value of F depends on the positive scalar values FA and FB ,
which multiply the vectors (−eA ) and eB , respectively.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
The vector (−eA ) is fixed and does not depend on θ but eB does vary with θ . The
X Th
er
tin
scalars FA and FB are constant values. Therefore, the maximum and minimum
values of F will be obtained when FA and FB either completely add or subtract
on
.O
one another, respectively. That is, when the vectors (−eA ) and eB are parallel to
each other. Taking into account [3] and [8], the maximum and minimum values
C
− Minimum value of F
π
θ=
2
2 1 1
|F|min = ρQ − + pB SB − pA SA
SB SA
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
254 C HAPTER 5. BALANCE P RINCIPLES
− Maximum value of F
3π
θ=
2
1 1
|F|min = ρQ2 + + pB SB + pA SA
SB SA
rs
d
ee
Z Z Z
Mliq = r × ρb dV + r × t dS = r × ρv dV , [9]
dt
n
V V
gi
∂V
t d le En
where Mliq is the resultant moment of the moments acting on the fluid. On
the other hand, expanding the right-hand term in [9] by means of the Reynolds
ar
ar s
ge ro or
Transport Theorem (5.39), yields
eS m
ib
ac
d
Z
∂
Z
f Z
C d P cs
r × ρv dV = r × ρv dV + (r × ρv) (n · v) dS .
b
[10]
dt
i
∂t
an an n
V V ∂V
y ha
null. Again, n and v are perpendicular to one another on the walls of the pipe
M
Z Z
d
SA SB
X Th
er
tin
.O
of the velocities in the middle point of each cross-section since the velocity
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 255
Q2
Mliq = −ρ v2B R SB ez = −ρ R ez [12]
SB
The following step consists in studying the contributions of the body forces,
which in this case are null (b = 0), and of the surface forces.
rs
ee
Z Z Z
Mliq = r × ρb dV + r × t dS = r × t dS =
n
gi
V
Z Z ∂V Z ∂V
t d le En
= r × t dS + r × t dS − r × t dS =
ar
SA SB Slat
ar s
ge ro or
eS m
ib
Z
= 0+ R pB ez dS + M p/ f = R pB SB ez + M p/ f ,
ac
f
C d P cs
SB
b
i
an an n
where M p/ f is the moment exerted by the pipe on the fluid. To determine its
y ha
M p/ f = Mliq − R pB SB ez = −ρ v2B R SB ez − R pB SB ez ,
M
.A
Q2
m
M p/ f = −R SB ρ v2B + pB
ez = −R ρ + pB SB ez . [13]
d
SB
uu
e
X Th
er
tin
Introducing the action and reaction law will allow obtaining the moment exerted
by the fluid on the pipe,
on
.O
M p/ f = −M f /p .
C
Considering the pipe and the rigid element P − O as a single body in equilibrium
©
M f /p + M = 0 =⇒ M = −M f /p = M p/ f .
Note that this result does not depend on the angle θ and, therefore, its module
will have a constant value.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
256 C HAPTER 5. BALANCE P RINCIPLES
d) To determine the value of the power W needed to provide a volume flow rate
Q the balance of mechanical energy equation (5.73) is used.
d 1 2
Z Z
W= ρv dV + σ : d dV [14]
dt 2
V V
rs
ee
This is proven as follows.
n
gi
1 T
σ : d = −p 1 : d = −p Tr (d) = −p Tr l +l =
t d le En
2
ar
ar s
ge ro or
eS m
ib
∂ vx ∂ vx ∂ vx
ac
f
∂x ∂y ∂z
C d P cs
b
= −p Tr (ll ) = −p Tr ∂ vy ∂ vy ∂ vy =
i
an an n
∂x ∂y ∂z
y ha
∂v ∂v ∂v
le
z z z
liv or ec
∂x ∂y ∂z
M
.A
∂ vx ∂ vy ∂ vz
= −p + + = −p ∇ · v = 0 ,
m
∂x ∂y ∂z
d
uu
e
where [1] has been applied in relation to the incompressibility condition, to con-
X Th
er
tin
.O
d 1 2 1 2 1 2
Z Z Z
∂
©
W= ρv dV = ρv dV + ρv (n · v) dS .
dt 2 ∂t 2 2
V V ∂V
And, again, considering the problem is in steady-state regime and that n and v
are perpendicular to one another on the walls of the pipe, the expression of the
incoming power W is determined.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 257
1 2 1 2
Z Z
W= ρv (n · v) dS + ρv (n · v) dS =
2 2
SA SB
1 2 1 2 1 1
Z Z
= ρv (−vA ) dS + ρvB (vB ) dS = ρv3A SA + ρv3B SB
2 A 2 2 2
SA SB
rs
W = ρQ3 2
− 2
2 SB SA
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
258 C HAPTER 5. BALANCE P RINCIPLES
E XERCISES
rs
only when the density at the initial time is uniform.
n ee
gi
5.2 – The figure below shows the longitudinal cross-section of a square pipe.
t d le En
Water flows through this pipe, entering through section AE and exiting through
section CD. The exit section includes a floodgate BC that can rotate around
ar
ar s
ge ro or
hinge B and is maintained in vertical position by the action of force F.
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Determine:
a) The exit velocity v2 in terms of the entrance velocity v1 (justify the ex-
pression used).
b) The resultant force and moment at point B of the actions exerted on the
fluid by the interior of the pipe.
c) The resultant force and moment at point B of the actions exerted by the
fluid on floodgate BC.
d) The value of the force F and the reactions the pipe exerts on flood-
gate BC.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 259
rs
5) Perfect fluid: σi j = −pδi j
ee
6) The weights of the fluid and the floodgate are negligible.
n
gi
t d le En
5.3 – The figure below shows the longitudinal cross-section of a pump used to
inject an incompressible fluid, fitted with a retention valve OA whose weight,
ar
ar s
ge ro or
per unit of width (normal to the plane of the figure), is W . Consider a steady-
eS m
ib
state motion, driven by the velocity of the piston V and the internal uniform
ac
pressure P1 . The external uniform pressure is P2 . f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Determine:
a) The uniform velocities v1 and v2 in terms of V (justify the expression
used).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
260 C HAPTER 5. BALANCE P RINCIPLES
b) The resultant force, per unit of width, exerted by the fluid on the valve OA.
c) The resultant moment about O, per unit of width, exerted by the fluid on
the valve OA.
d) The value of W needed for the valve OA to maintain its position (as
shown in the figure) during the injection process.
Additional hypotheses:
1) The body forces of the fluid are negligible.
2) Perfect fluid: σi j = −pδi j
rs
Perform the analysis by linear meter.
n ee
5.4 – A perfect and incompressible fluid flows through the pipe junction shown
gi
in the figure below. The junction is held in place by a rigid element O − D.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Determine:
a) The entrance velocities (vA and vB ) and the exit velocity (vC ) in terms
of the volume flow rate Q (justify the expression used).
b) The resultant force and moment at O of the actions exerted on the fluid
by the interior of the pipes in the junction.
c) The reaction force and moment at D of the rigid element.
d) The power W of the pump needed to provide the volume flow rates indi-
cated in the figure.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 261
Additional hypotheses:
1) The weights of the fluid and the pipes are negligible.
5.5 – The front and top cross-sections of an irrigation sprinkler are shown in
the figure below. A volume flow rate Q of water enters through section C at a
pressure P and exits through sections A and B at an atmospheric pressure Patm .
The flow is assumed to be in steady-state regime.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
Determine:
M
.A
b) The resultant force and moment at point O of the actions exerted on the
d
uu
er
c) The reaction that must be exerted on point O to avoid the sprinkler from
tin
.O
e) The power needed to provide a volume flow rate 2Q, considering that
W ∗ is the power of the pump needed to provide a volume flow rate Q.
Additional hypotheses:
1) Incompressible fluid
2) Perfect fluid: σi j = −pδi j
3) The weights of the sprinkler and the water inside it are negligible.
4) SA = SB = S and SC = S∗
5) m = Iα
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 6
Linear Elasticity
rs
n ee
gi
6.1 Hypothesis of the Linear Theory of Elasticity
t d le En
ar
The linear theory of elasticity can be considered a simplification of the general
ar s
ge ro or
eS m
ib
theory of elasticity, but a close enough approximation for most engineering ap-
ac
f
plications. The simplifying hypotheses of the linear theory of elasticity are
C d P cs
b
a) Infinitesimal strains. The displacements and its gradients are small, see
i
an an n
Chapter 2.
y ha
.A
ordinates cannot be distinguished from each other either, see Figure 6.1.
m
d
uu
e
x = X+ u =⇒ x≈X (6.1)
X Th
er
tin
|{z}
≈0
on
.O
C
Figure 6.1: Small displacements are considered in the linear theory of elasticity.
263
264 C HAPTER 6. L INEAR E LASTICITY
rs
and all references to the spatial and material descriptions (in addition
ee
to any associated concepts such as local derivative, material deriva-
n
tive, etc.) no longer make sense in infinitesimal elasticity.
gi
Likewise, the spatial Nabla differential operator ( ∇ ) is indistin-
t d le En
¯ ,
guishable from the material one ∇
ar
ar s
ge ro or
eS m
ib
∂ (•) ∂ (•) ¯ (•) .
= =⇒ ∇ (•) = ∇
ac
∂X ∂x f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
ρ0 = ρt |F| ≈ ρt ,
er
tin
.O
problems.
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Linear Elastic Constitutive Equation. Generalized Hooke’s Law 265
rs
ee
Definition 6.1. Isothermal processes are those that take place at a
n
temperature θ (x,t) that is constant along time,
gi
t d le En
θ (x,t) ≡ θ (x) .
ar
Adiabatic processes are those that take place without heat generation
ar s
ge ro or
eS m
ib
at any point and instant of time.
ac
f
C d P cs
b
Heat generated inside a domain V per unit of time:
i
an an n
Z Z
Qe = ρr dV − q · n dS = 0 ∀∆V ⊂ V
y ha
le
V
liv or ec
∂V
=⇒ ρr − ∇ · q = 0 ∀x ∀t
M
.A
batic.
d
uu
e
X Th
er
tin
on
.O
Hooke’s Law
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
266 C HAPTER 6. L INEAR E LASTICITY
σ (x,t) = C : ε (x,t)
Generalized (6.6)
Hooke’s law σ = C ε i, j ∈ {1, 2, 3}
ij i jkl kl
rs
Ci jkl = C jikl
→ minor symmetries
ee
Ci jkl = Ci jlk (6.7)
n
Ci jkl = Ckli j → major symmetries
gi
t d le En
Consequently, the number of different constants in the tensor of elastic constants
ar
C is reduced to 21.
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Remark 6.3. An essential characteristic of the elastic behavior
b
i
(which is verified in (6.5)) is that the stresses at a certain point and
an an n
time, σ (x,t), depend (only) on the strains at said point and time,
y ha
.A
m
Consider the specific internal energy u (x,t) (internal energy per unit of mass)
X Th
er
tin
and the density of internal energy û (x,t) (internal energy per unit of volume),
which related through
on
.O
û (x,t) = ρ0 u (x,t) ,
C
û
©
z }| { (6.8)
du du d (ρ0 u) d û
ρ ≈ ρ0 = = ,
dt dt dt dt
where ρ0 ≈ ρ (see Remark 6.2) has been taken into account. Consider now the
energy equation in its local form2 ,
du d û . d û .
ρ0 = = σ : d + ρ0 r − ∇ · q = σ : ε =⇒ =σ :ε , (6.9)
dt dt dt
2 .
The identity d = ε , characteristic of the infinitesimal strain case, is considered here.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Linear Elastic Constitutive Equation. Generalized Hooke’s Law 267
where the adiabatic nature of the deformation process (ρ0 r − ∇ · q = 0) has been
considered. Then, the global (integral) form of the energy equation in (6.9) is
obtained by integrating over the material volume V .
rs
ee
V
n
gi
Here, U (t) is the internal energy of the material volume considered.
t d le En
ar
ar s
ge ro or
eS m
ib
Remark 6.4. The stress power (in the case of linear elasticity) is an
ac
exact differential,
f
C d P cs
. dU
b
R
stress power = σ : ε dV = .
i
dt
an an n
V
y ha
le
liv or ec
.A
i↔k
m
j↔l
d
d û not . . . . 1. z. }| {
uu
dt 2
X Th
er
tin
1. . 1 . .
= εi j Ci jkl εkl + εkl Ckli j εi j = εi j Ci jkl εkl + εi j Ci jkl εkl =
on
2 2
.O
1d 1d
C
= εi j Ci jkl εkl = (εε : C : ε ) ,
2 dt 2 dt
©
(6.11)
where the symmetries in (6.7) have been taken into account. Integrating the ex-
pression obtained and imposing the condition that the density of internal energy
û (x,t0 ) in the neutral state be null3 (for t = t0 ⇒ ε (x,t0 ) = 0) produces the
density of internal energy.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
268 C HAPTER 6. L INEAR E LASTICITY
1
û (x,t) = (εε (x,t) : C : ε (x,t)) + a (x)
2 =⇒
û (x,t0 ) = 0 ∀x
(6.12)
1
=⇒ ε (x,t0 ) : C : ε (x,t0 ) + a (x) = a (x) = 0 ∀x
2 | {z }
=0
Density of
o 1 1
internal energy û (εε ) = (εε : C : ε ) = εi j Ci jkl εkl (6.13)
2 2
rs
ee
Now, (6.13) is differentiated with respect to ε , considering once more the
n
symmetries in (6.7).
gi
t d le En
∂ û (εε ) 1 1 1 1
= C :ε + ε :C = C :ε + C :ε =C :ε =σ
ar
ar s
2 2 2 2
ge ro or
∂ε
eS m
ib
ac
∂ û (εε ) 1 1 1
f 1
= Ci jkl εkl + εkl Ckli j = Ci jkl εkl + Ci jkl εkl = Ci jkl εkl = σi j
C d P cs
b
∂ εi j 2 2 2 2
i
(6.14)
an an n
y ha
∂ û (ε )
ε
=σ
le
liv or ec
=⇒ ∂ε (6.15)
û (ε )
M
∂ ε
.A
= σi j i, j ∈ {1, 2, 3}
∂ εi j
m
for the stresses (which are obtained by differentiation of this potential), named
X Th
er
tin
elastic potential.
on
.O
1 1
û (εε ) = 2 ε : C :ε = σ :ε
C
| {z } 2
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Isotropy. Lamé’s Constants. Hooke’s Law for Isotropic Linear Elasticity 269
rs
The elastic properties of a linear elastic material are contained in the tensor of
elastic constants C defined in (6.6) and (6.7). Consequently, the components of
ee
this tensor must be independent of the orientation of the Cartesian system used4 .
n
Consider, for example, the systems {x1 , x2 , x3 } and {x1 0 , x2 0 , x3 0 } in Figure 6.2,
gi
the constitutive equation for these two systems is written as
t d le En
ar
ar s
ge ro or
{x1 , x2 , x3 } =⇒ [σ
σ ] = [C
C] : [εε ]
eS m
ib
(6.17)
ac
f
C d P cs
{x1 0 , x2 0 , x3 0 } =⇒ σ]0
[σ C] 0
= [C : [εε ] 0
b
i
an an n
y ha
and, for the case of an isotropic material, the components of C in both sys-
C] 0 ). Therefore, the aforementioned definition
le
tems must be the same ( [C C] = [C
liv or ec
of isotropy, which has a physical character, translates into the isotropic charac-
M
.A
d
C = λ 1 ⊗ 1 + 2µI
uu
e
Tensor of elastic
X Th
er
(6.18)
tin
constants Ci jkl = λ δi j δkl + µ δik δ jl + δil δ jk
i, j, k, l ∈ {1, 2, 3}
on
.O
C
Here, λ and µ are known as Lamé’s constants, which characterize the elastic
behavior of the material and must be obtained experimentally.
4 A tensor is isotropic if it maintains its components in any Cartesian coordinate system. The
most general expression of a fourth-order isotropic tensor is C = λ 1⊗1+2µI , ∀λ , µ. Here,
the fourth-order
symmetric (isotropic) unit tensor I is defined by means of its components as
[I]i jkl = δik δ jl + δil δ jk /2.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
270 C HAPTER 6. L INEAR E LASTICITY
rs
ee
Introducing (6.18) in (6.6) results in the isotropic linear elastic constitutive
n
equation,
gi
t d le En
1 1
σi j = Ci jkl εkl = λ δi j δkl εkl +2µ δik δ jl εkl + δil δ jk εkl . (6.19)
ar
2 | {z } 2 | {z }
ar s
ge ro or
| {z }
ε ji = εi j
eS m
ib
εll εi j
ac
|
f {z
εi j
}
C d P cs
b
i
an an n
Constitutive eqn. for a σ = λ Tr (εε ) 1 + 2µεε
y ha
.A
m
The constitutive equation (6.20) provides the stresses in terms of the strains. To
X Th
er
tin
.O
Tr (σ
σ ) = λ Tr (εε ) Tr (1) +2µ Tr (εε ) = (3λ + 2µ) Tr (εε )
©
| {z }
3 =⇒
(i = j) =⇒ σii = λ εll δii +2µεii = (3λ + 2µ) εll
(6.21)
|{z}
3
1
=⇒ Tr (εε ) = Tr (σ
σ) .
(3λ + 2µ)
b) ε is isolated from (6.20) and introduced in (6.21),
1 1 λ 1
ε =− λ Tr (εε ) 1 + σ =− Tr (σ
σ)1+ σ. (6.22)
2µ 2µ 2µ (3λ + 2µ) 2µ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Isotropy. Lamé’s Constants. Hooke’s Law for Isotropic Linear Elasticity 271
The new elastic properties E (Young’s modulus) and ν (Poisson’s ratio) are de-
fined as follows.
Young’s modulus or µ (3λ + 2µ)
E=
tensile (elastic) modulus
λ +µ =⇒
λ
ν=
Poisson’s ratio
2 (λ + µ)
(6.23)
rs
νE
λ =
ee
(1 + ν) (1 − 2ν)
=⇒
n
E
gi
µ = =G
shear (elastic) modulus
2 (1 + ν)
t d le En
ar
ar s
Equation (6.22) can be expressed in terms of E and ν, resulting in the inverse
ge ro or
eS m
ib
Hooke’s law.
ac
f
C d P cs
b
ν 1+ν
i
ε = − Tr (σ σ)1+
an an n
σ
E E
Inverse constitutive
y ha
equation for an isotropic ν 1+ν (6.24)
le
εi j = − σll δi j +
liv or ec
i, j ∈ {1, 2, 3}
.A
m
d
uu
Finally, (6.24) is rewritten, using engineering notation for the components of the
e
er
tin
1 1
on
.O
1 1
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
272 C HAPTER 6. L INEAR E LASTICITY
Solution
From (6.25) one obtains
rs
σx τxy
ee
εx =
γxy = =0
E G
n
σ τ
σy = σz = 0 =⇒ εy = −ν x τxy = τxz = τyz = 0 =⇒ γxz = xz = 0
gi
E G
t d le En
εz = −ν σx γyz = τyz = 0
E G
ar
ar s
ge ro or
eS m
ib
Therefore, due to these strains, the rectangular cuboid subjected to an uni-
ac
f
axial tensile test, shown in the figure below, stretches in the x-direction and
C d P cs
contracts in the y- and z-directions.
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
Consider the decomposition of the stress tensor σ and the deformation tensor ε
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Hooke’s Law in Spherical and Deviatoric Components 273
ν 1+ν 1 − 2ν
e = Tr (εε ) = − Tr (σ
σ ) Tr (1) + Tr (σ
σ) = Tr (σ
σ) =
E | {z } E E | {z }
3 3σm (6.28)
3 (1 − 2ν)
= σm
E
E
σm = e = Ke
rs
=⇒ 3 (1 − 2ν) (6.29)
ee
de f 2 E
K = λ + µ = = bulk modulus
n
3 3 (1 − 2ν)
gi
Introducing (6.26), (6.27) and (6.29) in (6.24), results in
t d le En
ar
ar s
1+ν 1 − 2ν 1+ν 0
ge ro or
ν
(σm 1 + σ 0 ) =
eS m
ib
ε = − 3σm 1 + σm 1 + σ =
E E E |{z} E
ac
E f
C d P cs
b
e
3 (1 − 2ν)
i
an an n
y ha
.A
1+ν 0 1 0 1 0
=⇒ ε 0 = σ = σ = σ .
m
E 2µ 2G
d
uu
e
Equations (6.29) and (6.30) relate the spherical part (characterized by the mean
X Th
er
tin
σ 0 and ε 0 ) of the
stress σm and the volumetric strain e) and the deviatoric part (σ
stress and strain tensors as follows.
on
.O
C
σm = Ke Spherical part
©
σ 0 = 2Gεε 0
(6.31)
Deviatoric part
σ 0 i j = 2Gε 0 i j i, j ∈ {1, 2, 3}
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
274 C HAPTER 6. L INEAR E LASTICITY
rs
6.5 Limits in the Values of the Elastic Properties
n ee
Thermodynamic considerations allow proving that the tensor of elastic constants
gi
C is positive-definite5 , and, thus,
t d le En
ε :C :ε >0 ∀εε 6= 0 . (6.32)
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Remark 6.7. As a consequence of (6.32), the elastic potential is al-
b
i
ways null or positive,
an an n
1
y ha
û (εε ) = ε : C : ε ≥ 0 .
2
le
liv or ec
M
.A
m
d
uu
Remark 6.8. The elastic potential has a minimum at the neutral state,
e
X Th
1 ∂ û (εε ) ∂ 2 û (εε )
û (εε ) = ε : C : ε , σ = = C : ε and =C .
on
.O
2 ∂ε ∂ε ⊗∂ε
Then, for ε = 0,
C
∂ û (εε )
=0 =⇒ û (εε ) has an extreme
∂ ε ε =0 (maximum-minimum) at ε = 0.
∂ 2 û (εε )
= C =⇒ The extreme is a minimum.
∂ε ⊗∂ε ε =0 |{z}
positive-
definite
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Limits in the Values of the Elastic Properties 275
rs
Figure 6.4: Elastic potential.
n ee
Consider the expression of the elastic potential (6.16) and the constitutive
gi
equation (6.20), then,
t d le En
ar
1 1 1
ar s
ge ro or
û (εε ) = ε : C : ε = σ : ε = λ Tr (εε ) 1 + 2µεε : ε =
eS m
ib
2 2 2
ac
1 1 f (6.33)
C d P cs
= λ Tr (εε ) 1 : ε +µεε : ε = λ Tr2 (εε ) + µεε : ε .
2 2
b
i
|{z}
an an n
Tr (εε )
y ha
le
Expression (6.33) can also be written in terms of the spherical and deviatoric
liv or ec
components of strain6 ,
M
.A
m
1 2 1
d
2 | {z } 2
e
e
X Th
er
tin
.O
1 1 1 2
C
ε :ε = e1+ε :0 e 1 + ε = e2 1 : 1 + e 1 : ε 0 + ε 0 : ε 0 =
0
©
3 3 9 |{z} 3 | {z }
3 Tr (εε 0 ) = 0 (6.35)
1
= e2 + ε 0 : ε 0 .
3
Replacing (6.35) in (6.34),
1 2 1 2 0 0 1 2
û (εε ) = λ e + µe + µεε : ε = λ + µ e2 + µεε 0 : ε 0 . (6.36)
2 3 2 3
| {z }
K
6 The trace of a deviatoric tensor is always null, Tr (εε 0 ) = 0.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
276 C HAPTER 6. L INEAR E LASTICITY
1
û (εε ) = Ke2 + µεε 0 : ε 0 ≥ 0 (6.37)
2
rs
3 =⇒ û(1) = Ke2 ≥ 0 =⇒ K > 0 (6.38)
ee
(1)
ε0 = 0
2
n
gi
b) A pure deviatoric deformation process7
t d le En
)
ε (2) = ε 0
ar
=⇒ û(2) = µεε 0 : ε 0 ≥ 0 =⇒ µ > 0
ar s
(6.39)
ge ro or
eS m
ib
(2)
e =0
ac
f
C d P cs
Equations (6.38) and (6.39) lead to
b
i
an an n
E E
K= >0 and µ =G= >0 (6.40)
y ha
3 (1 − 2ν) 2 (1 + ν)
le
liv or ec
which are the limits in the values of the elastic constants K and G. Experience
M
.A
d
uu
E
e
> 0
X Th
2 (1 + ν)
er
=⇒ E >0,
tin
ν ≥0
on
.O
(6.41)
E
C
> 0
1
3 (1 − 2ν)
©
=⇒ 0≤ν ≤ .
2
E ≥0
7 The double contraction or double dot product of a tensor by itself is always equal or greater
than zero: ε 0 : ε 0 = ε 0 i j ε 0 i j ≥ 0.
8 In rare cases, a material can have a negative Poisson’s ratio. Such materials are named
auxetic materials.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
The Linear Elastic Problem 277
Initial actions:
(
b (x, 0)
t =0 →
t (x, 0)
rs
Figure 6.5: Linear elastic problem.
n ee
gi
6.6 The Linear Elastic Problem
t d le En
Consider the linear elastic solid9 in Figure 6.5, which is subjected to certain ac-
ar
ar s
ge ro or
tions characterized by the vector of body forces b (x,t) in the interior of the vol-
eS m
ib
ac
ume V and the traction vector t (x,t) on the boundary ∂V . The set of equations
f
C d P cs
that allow determining the evolution along time of the displacements u (x,t),
b
strains ε (x,t) and stresses σ (x,t) is named linear elastic problem.
i
an an n
y ha
.A
∂ 2 u (x,t)
uu
e
∇ · σ (x,t) + ρ0 b (x,t) = ρ0
X Th
∂t 2
er
tin
(3 equations) (6.42)
∂ σi j ∂ 2u j
+ ρ0 b j = ρ0 2 j ∈ {1, 2, 3}
on
.O
∂ xi ∂t
C
9 Here, linear elastic solid refers to a continuous medium constituted by a material that obeys
the linear elastic constitutive equation.
10 The symmetry of the stress and strain tensors entails that only six of the nine equations
are different from one another. In addition, when listing the unknowns, only the different
components of these tensors will be considered.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
278 C HAPTER 6. L INEAR E LASTICITY
1
ε (x,t) = ∇S u (x,t) = (u ⊗ ∇ + ∇ ⊗ u)
2 (6 equations) (6.44)
1 ∂ ui ∂ u j
εi j = + i, j ∈ {1, 2, 3}
2 ∂ x j ∂ xi
rs
• u (x,t) (3 unknowns)
ee
• ε (x,t) (6 unknowns) (6.45)
• σ (x,t) (6 unknowns)
n
gi
and constitute a system of partial differential equations (PDEs). The system is
t d le En
composed of 15 differential equations with the 15 unknowns listed in (6.45).
ar
These are of the type (•) (x, y, z,t), and, thus, must be solved in the R3 × R+
ar s
ge ro or
eS m
ib
space. The problem is well defined when the adequate boundary conditions are
ac
provided. f
C d P cs
b
i
an an n
.A
Γuσ = Γ ≡ ∂V ,
S S
Γu Γσ
d
(6.46)
uu
e
Γσ = {0}
T T T
Γu Γσ = Γu Γuσ = Γuσ / .
X Th
er
tin
These allow defining the boundary conditions in space, that is, those conditions
on
.O
that affect the spatial arguments (x, y, z) of the unknowns (6.45) of the problem.
C
)
u (x,t) = u∗ (x,t)
∀x ∈ Γu ∀t (6.47)
ui (x,t) = u∗i (x,t) i ∈ {1, 2, 3}
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
The Linear Elastic Problem 279
rs
ee
Figure 6.6: Boundary conditions in space.
n
gi
t d le En
• Boundary Γuσ : prescribed displacements and tractions11
ar
ui (x,t) = u∗i (x,t)
ar s
ge ro or
eS m
ib
i, j, k ∈ {1, 2, 3}, i 6
= j ∀x ∈ Γuσ ∀t (6.49)
σ jk (x,t) · nk = t ∗j (x,t)
ac
f
C d P cs
b
i
an an n
le
liv or ec
Solution
M
.A
The different types of boundary conditions in space are illustrated in the fol-
m
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
280 C HAPTER 6. L INEAR E LASTICITY
rs
6.6.3 Quasi-Static Problem
ee
The system of equations (6.42) to (6.50) can be visualized, from a mechanical
n
gi
point of view, as a system of actions or data (the body forces b (x,t), the traction
vector t∗ (x,t), the prescribed displacements u∗ (x,t) and the initial velocities
t d le En
v0 (x)) that, introduced into a mathematical model composed of the differen-
ar
ar s
tial equations given in Section 6.6.1 and the boundary conditions described in
ge ro or
eS m
ib
Section 6.6.2, provides the response or solution in the form of the displacement
ac
f
field u (x,t), the deformation field ε (x,t) and the stress field σ (x,t).
C d P cs
b
i
an an n
b (x,t)
y ha
MAT HEMAT ICAL u (x,t)
* +
∗
u (x,t)
le
liv or ec
⇒ MODEL : ⇒ ε (x,t)
t∗ (x,t) PDEs + BCs (6.51)
M
.A
σ (x,t)
v0 (x)
| {z }
m
| {z } not
d
er
tin
In the most general case12 , both the actions and the responses depend on time
(see Figure 6.7) and the system of PDEs must be integrated over both the space
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
The Linear Elastic Problem 281
∂ 2 u (x,t)
a= ≈0.
∂t 2
rs
ee
This hypothesis is acceptable when the actions are applied slowly.
In such case, the variation of the actions A along time is slow
n
∂ 2A /∂t 2 ≈ 0 and, due to the continuous dependency of the results
gi
on the data, the variation of the response R along time is also small
t d le En
∂ 2R /∂t 2 ≈ 0 . Consequently, the second derivative of the response
ar
is considered negligible and, in particular,
ar s
ge ro or
eS m
ib
ac
∂ 2 u (x,t)
f
≈0.
C d P cs
∂t 2
b
i
an an n
y ha
le
liv or ec
The governing differential equations are reduced to the following in the case
of a quasi-static problem:
M
.A
d
uu
e
∂ 2 u (x,t)
X Th
∂t 2
on
.O
C
b) Constitutive equation
©
1
ε (x,t) = ∇S u (x,t) = (u ⊗ ∇ + ∇ ⊗ u) (6.54)
2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
282 C HAPTER 6. L INEAR E LASTICITY
b (x, λ ) *
MAT HEMAT ICAL
+
u (x, λ )
∗
u (x, λ ) ⇒ MODEL : ⇒ ε (x, λ )
t∗ (x, λ )
PDEs + BCs
σ (x, λ ) (6.55)
rs
| {z } | {z }
not not
Actions = A (x, λ ) Responses = R (x, λ )
n ee
In other words, for each value of the actions (characterized by a fixed value of
gi
λ ∗ ), A (xλ ∗ ), a response R (x, λ ∗ ) is obtained. Varying the value of λ ∗ produces
t d le En
a family of actions and its corresponding family of responses.
ar
ar s
ge ro or
eS m
ib
Example 6.3 – Application to a typical problem of strength of materials.
ac
f
C d P cs
Solution
b
i
an an n
Consider a cantilever beam subjected to a force F (t) at its free end. Under
y ha
of the type λ F ∗ , the response (deflection at its free end) can be computed as
M
.A
F ∗l3
δ (λ ) = λ .
m
3EI
d
uu
e
er
tin
Now, if the evolution along time of λ (t) can take any form, the value of
δ (t) = δ (λ (t)) corresponding to each instant of time only depends on the
on
.O
corresponding value of λ .
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Solution to the Linear Elastic Problem 283
rs
n ee
gi
6.7 Solution to the Linear Elastic Problem
t d le En
ar
The linear elastic problem can be typically solved following two different ap-
ar s
ge ro or
eS m
ib
proaches:
ac
a) Displacement formulation f
C d P cs
b) Stress formulation
b
i
an an n
Their names are directly related to which is the main unknown being considered
y ha
.A
er
tin
on
.O
C
∂ 2u
∇ · σ + ρ0 b = ρ0 Cauchy’s equation
∂t 2
σ = λ Tr (εε ) 1 + 2µεε Constitutive equation (6.56)
1
ε = ∇S u = (u ⊗ ∇ + ∇ ⊗ u) Geometric equation
2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
284 C HAPTER 6. L INEAR E LASTICITY
)
Γu : u = u∗
Boundary conditions in space (6.57)
Γσ : t∗ = σ · n
)
u (x, 0) = 0
. Initial conditions (6.58)
u (x, 0) = v0
The aim is to pose a reduced system in which only the displacement field
rs
u (x,t) intervenes as an unknown. The first step consists in replacing the consti-
tutive equation in the Cauchy’s equation, both given in (6.56).
n ee
∂ 2u
gi
∇ · σ + ρ0 b = ∇ · λ Tr (εε ) 1 + 2µεε + ρ0 b = ρ0 2
∂t
t d le En
(6.59)
∂ 2u
ar
=⇒ λ ∇ · Tr (εε ) 1 + 2µ∇ · ε + ρ0 b = ρ0 2
ar s
ge ro or
eS m
ib
∂t
ac
Consider the following identities13 . f
C d P cs
b
i
an an n
1 ∂ 2 ui
∂ εi j 1 ∂ ui ∂ u j 1 ∂ ∂uj
y ha
∂
∇·ε i = = + = + =
∂xj ∂xj 2 ∂ x j ∂ xi 2 ∂ x j ∂ x j 2 ∂ xi ∂ x j
le
liv or ec
1 1 ∂ 1 1
M
.A
= ∇2 u i + (∇ · u) = ∇2 u + ∇ (∇ · u) i ∈ {1, 2, 3}
2 2 ∂ xi 2 2 i
m
1 1
uu
∇ · ε = ∇ (∇ · u) + ∇2 u
e
X Th
2 2
er
tin
(6.60)
on
.O
C
∂ ∂ ∂ ul ∂ ∂ ul
∇ · Tr (εε ) 1 i = (εll δi j ) = δi j = =
©
∂xj ∂ x j ∂ xl ∂ xi ∂ xl
∂
= (∇ · u) = ∇ (∇ · u) i i ∈ {1, 2, 3}
∂ xi
∇ · Tr (εε ) 1 = ∇ (∇ · u)
(6.61)
13
de f
The Laplace operator of a vector v is defined as ∇2 v i = ∂ 2 vi /(∂ x j ∂ x j ).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Solution to the Linear Elastic Problem 285
rs
The boundary conditions can also be written in terms of the displacements as
ee
follows. Replacing the constitutive equation of (6.56) in the boundary conditions
n
in Γσ of (6.57) results in
gi
t∗ = σ · n = λ Tr (εε ) 1 + 2µεε · n = λ (Tr (εε )) n + 2µεε · n =
t d le En
(6.63)
ar
= λ (∇ · u) n + 2µ ∇S · u · n = λ (∇ · u) n + µ (u ⊗ ∇ + ∇ ⊗ u) · n
ar s
ge ro or
eS m
ib
ac
f
and the boundary conditions in space (6.57) expressed in terms of the displace-
C d P cs
ments are obtained.
b
i
an an n
y ha
u = u∗
in Γu
le
ui = u∗i i ∈ {1, 2, 3}
liv or ec
(6.64)
λ (∇ · u) n + µ (u ⊗ ∇ + ∇ ⊗ u) · n = t∗
M
.A
in Γ
λ ul,l ni + µ (ui, j n j + u j,i n j ) = ti∗
σ
i, j ∈ {1, 2, 3}
m
d
uu
The initial conditions (6.58) remain unchanged. Integrating the system (6.62)
e
X Th
er
yields the displacement field u (x,t). Differentiation of this field and substitution
tin
in the geometric equation of (6.56) produces the strain field ε (x,t), and, finally,
on
replacing the strain in the constitutive equation results in the stress field σ (x,t).
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
286 C HAPTER 6. L INEAR E LASTICITY
Cylindrical coordinates
∂ e 2µ ∂ ωz ∂ ωθ ∂ 2 ur
(λ + 2µ) − + 2µ + ρbr = ρ 2
∂r r ∂θ ∂z ∂t
1 ∂e ∂ ωr ∂ ωz ∂ 2 uθ (6.65)
(λ + 2µ) − 2µ + 2µ + ρbθ = ρ 2
r ∂θ ∂z ∂r ∂t
∂ e 2µ ∂ (rωθ ) 2µ ∂ ωr ∂ 2 uz
(λ + 2µ) − + + ρbz = ρ 2
∂z r ∂r r ∂θ ∂t
rs
where
ee
1 1 ∂ uz ∂ uθ
n
ωr = −Ωθ z = −
gi
2 r ∂θ ∂z
t d le En
1 ∂ ur ∂ uz
ωθ = −Ωzr = −
ar
2 ∂z ∂r
ar s
ge ro or
eS m
ib
1 1 ∂ (ruθ ) 1 ∂ ur
ac
ωz = −Ωrθ = − f
C d P cs
2 r ∂r r ∂θ
b
i
an an n
1 ∂ (rur ) 1 ∂ uθ ∂ uz
y ha
e= + +
r ∂r r ∂θ ∂z
le
liv or ec
M
.A
x = r cos θ
m
not
x (r, θ , z) ≡ y = r sin θ
d
uu
z=z
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Solution to the Linear Elastic Problem 287
Spherical coordinates
∂e 2µ ∂ ωφ sin θ 2µ ∂ ωθ ∂ 2 ur
(λ + 2µ) − + + ρbr = ρ 2
∂ r r sin θ ∂θ r sin θ ∂ φ ∂t
1 ∂e 2µ ∂ ωr 2µ ∂ rωφ sin θ ∂ 2 uθ (6.66)
(λ + 2µ) − + + ρbθ = ρ 2
r ∂ θ r sin θ ∂ φ r sin θ ∂r ∂t
1 ∂ e 2µ ∂ (rωθ ) 2µ ∂ ωr ∂ 2 uφ
(λ + 2µ) − + + ρbφ = ρ 2
rs
r sin θ ∂ φ r ∂r r ∂θ ∂t
ee
where
n
!
gi
11 ∂ uφ sin θ 1 ∂ uθ
ωr = −Ωθ φ = −
t d le En
2
r sin θ ∂θ r sin θ ∂ φ
ar
!
ar s
1 1 ∂ ur 1 ∂ ruφ
ge ro or
eS m
ib
ωθ = −Ωφ r = −
2 r sin θ ∂ φ r ∂r
ac
f
C d P cs
b
1 1 ∂ (ruθ ) 1 ∂ ur
i
ωz = −Ωrθ = −
an an n
2 r ∂r r ∂θ
y ha
!
∂ r2 ur sin θ
∂ (ruθ sin θ ) ∂ ruφ
le
1
liv or ec
e= 2 + +
r sin θ ∂r ∂θ ∂φ
M
.A
m
uu
x = r sin θ cos φ
e
not
x (r, θ , φ ) ≡ y = r sin θ sin φ
X Th
er
tin
z = z cos θ
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
288 C HAPTER 6. L INEAR E LASTICITY
∇ · σ + ρ0 b = 0 Equilibrium equation
ν 1+ν
ε = − Tr (σ σ)1+ σ Inverse constitutive equation (6.67)
E E
1
ε = ∇S u = (u ⊗ ∇ + ∇ ⊗ u) Geometric equation
rs
2
n ee
)
Γu : u = u∗
gi
Boundary conditions in space (6.68)
Γσ : t∗ = σ · n
t d le En
ar
ar s
ge ro or
where the inverse constitutive (6.24) (strains in terms of stresses) has been con-
eS m
ib
sidered in (6.67).
ac
f
The starting point of the stress formulation is the geometric equation of (6.67)
C d P cs
b
from which, by means of successive differentiation, the displacements are elim-
i
an an n
le
εi j, kl + εkl, i j − εik, jl − ε jl, ik = 0 i, j, k, l ∈ {1, 2, 3} . (6.69)
liv or ec
M
.A
Then, the equations of the problem are deduced in the following manner:
a) The constitutive equation of (6.67) is replaced in the compatibility equa-
m
tions (6.69).
uu
e
er
tin
.O
Beltrami-Michell equation
C
1 ν
∇2 σi j + σll,i j = − δi j (ρ0 bl ), l − (ρ0 bi ), j − (ρ0 b j ), i (6.70)
1+ν 1−ν
i, j ∈ {1, 2, 3}
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Unicity of the Solution to the Linear Elastic Problem 289
The integration of the system in (6.70) yields the stress field σ (x). Substi-
tution of the stresses in the inverse constitutive equation of (6.67) results in
the strains ε (x). However, to obtain the displacement field u (x), the geomet-
ric equations must be integrated, taking into account the boundary conditions
rs
in Γu 15 .
ee
ε (x) = 1 u (x) ⊗ ∇ + ∇ ⊗ u (x)
x∈V
n
2 (6.73)
gi
u (x) = u∗ (x) ∀x ∈ Γu
t d le En
ar
Thus, the system of second-order PDEs must be integrated in R3 .
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
Remark 6.10. The need to integrate the second system (6.73) (when
i
an an n
.A
m
er
tin
on
u (x,t)
C
not
R (x,t) ≡ ε (x,t)
©
σ (x,t)
to the linear elastic problem posed in (6.42) to (6.44) is unique.
Proof
not
Consider the actions defined by A (x,t) ≡ [b (x,t) , u∗ (x,t) , t∗S(x,t) , v0 (x)]T ,
in the domains V , Γu , Γσ and V , respectively, (satisfying Γσ Γu = ∂V and
15An analytical procedure to integrate these geometric equations was provided in Chapter 3,
Section 3.4.2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
290 C HAPTER 6. L INEAR E LASTICITY
rs
ee
T
Γσ Γu = 0) / act on the linear elastic problem schematically represented in Fig-
n
ure 6.10.
gi
not
The possible solutions R (x,t) ≡ [u (x,t) , ε (x,t) , σ (x,t)]T to the linear elas-
t d le En
tic problem must satisfy the equations:
ar
ar s
ge ro or
∂ 2u
eS m
ib
∇ · σ + ρ0 b = ρ0 Cauchy’s equation
ac
∂t 2
f
C d P cs
(6.74)
σ = λ Tr (εε ) 1 + 2µεε
b Constitutive equation
i
an an n
1
ε = ∇S u = (u ⊗ ∇ + ∇ ⊗ u) Geometric equation
y ha
2
le
liv or ec
M
)
.A
Γu : u = u∗
Boundary conditions in space (6.75)
Γσ : t∗ = σ · n
m
d
uu
e
X Th
er
)
tin
u (x, 0) = 0
. Initial conditions (6.76)
u (x, 0) = v0
on
.O
C
The unicity of the solution is proven as follows. Suppose the solution is not
©
unique, that is, there exist two different solutions to the problem,
(1) (2)
u (x,t) u (x,t)
not not
R (1) (x,t) ≡ ε (1) (x,t) and R (2) (x,t) ≡ ε (2) (x,t)
which, therefore, must satisfy equations (6.74) to (6.76) and are the elastic re-
not
sponses to the same action A (x,t) ≡ [b (x,t) , u∗ (x,t) , t∗ (x,t) , v0 (x)]T . Con-
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Unicity of the Solution to the Linear Elastic Problem 291
rs
ee
e (x,t) = ∇ · σ (2) (x,t) − σ (1) (x,t) = ∇ · σ (2) − ∇ · σ (1) =
∇·σ
n
gi
! !
∂ 2 u(2) ∂ 2 u(1) (6.79)
= −ρ0 b + ρ0 − −ρ0 b + ρ0 =
t d le En
∂t 2 ∂t 2
ar
ar s
ge ro or
∂ 2 u(2) ∂ 2 u(1) ∂ 2u
eS m
ib
e
= ρ0 − ρ = ρ
ac
0 0
∂t 2 ∂t 2
f ∂t 2
C d P cs
• Constitutive equation17 b
i
an an n
y ha
(6.80)
= C : ε (2) − ε (1) = C : εe
M
.A
m
• Geometric equation
uu
e
X Th
er
(6.81)
= ∇S u(2) − u(1) = ∇S u
on
.O
e
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
292 C HAPTER 6. L INEAR E LASTICITY
rs
e (x, 0) = u(2) (x, 0) − u(1) (x, 0) = 0 − 0 = 0
u
ee
(6.84)
e (x, 0) .
n
∂u . .
e (x, 0) = u(2) (x, 0) − u(1) (x, 0) = v0 − v0 = 0
=u
gi
∂t
t d le En
Consider now the calculation of the integral
ar
ar s
ge ro or
eS m
ib
= 0 in Γσ Divergence
ac
Z . Z z }| { . f
TheoremZ .
C d P cs
n· σ
e ·u
e dS = n·σe · u e dS = ∇· σ
e ·u
e dV = 0 , (6.85)
b
i
|{z}
an an n
S
∂V Γu Γσ = 0 in Γu V
y ha
le
where the conditions (6.82) and (6.83) have been applied. Operating on (6.85)
liv or ec
results in
M
.A
. . . ∂ 2u e . . T
∇· σ e ·u e = ∇·σ e ·u e = ρ0 2 · u
m
e+σ e : ∇u e+σ e : ∇u
e
d
∂t
uu
. .
e
. ∂ σe . u 2u . u
X Th
∂ i j ∂ j ∂ j ∂
e ji j
er
i, j ∈ {1, 2, 3}
e e e
σei j uej = uej + σei j = ρ0 2 uej + σ
tin
∂ xi ∂ xi ∂ xi ∂t ∂ xi
(6.86)
on
.O
where the condition (6.79) has been considered. On the other hand18 ,
C
. . .
©
. T . 1 . . 1 .
∇ue =u e⊗∇ = e⊗∇+∇⊗u
u e + u e⊗∇−∇⊗u e = εe + Ω
e =⇒
2| 2
. . .
{z } | {z }
S a
.
εe = ∇ u e e =∇ u
Ω e
. T . . . T .
e : ∇u
σ e =σ e : εe + σ
e :Ω e =⇒ e : ∇u
σ e =σ e : εe .
| {z }
=0
(6.87)
.
18The fact that σ
e is a symmetric tensor and Ω
e is an antisymmetric one is considered here,
. .
which leads to σ e = σei j Ω
e :Ω e i j = 0.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Unicity of the Solution to the Linear Elastic Problem 293
In addition19 ,
v2
. .
e
. z}|{
e·u
∂ u
e . e . 1
∂2u ∂2u 1 ∂ e v·eve
ρ0 2 · ue = ρ0 2 · ue = ρ0 = ρ0 =
∂t ∂t 2 ∂t 2 ∂t (6.88)
∂2ue .
d 1 2 d 1 2
= ρ0 v
e =⇒ ρ0 2 · u
e = ρ0 v
e .
dt 2 ∂t dt 2
Replacing (6.88) and (6.87) in (6.86), and the resulting expression in (6.85), and
taking into account the definition of internal energy U given in (6.10) produces
rs
Z . Z
d 1 2
Z .
ee
∇· σ ·u
e e dV = ρ0 v dV + σ
e e : εe dV = 0 =⇒
dt 2
n
V V V
gi
d
Z
1 . Z
v2 dV + σ
ρ0 e e : εe dV = 0 =⇒ (6.89)
t d le En
dt 2
V V
ar
ar s
ge ro or
| {z } | {z }
eS m
ib
dK
e /dt d Ue/dt
ac
f
C d P cs
dK
e d Ue d e e
b
i
+ = K + U = 0 ∀t ≥ 0 . (6.90)
an an n
dt dt dt
y ha
le
liv or ec
Note, though, that at the initial time t = 0 the following is satisfied (see (6.10),
M
.A
1 1
Z Z
d
2
K v0 · e
e =
v dV = v0 dV = 0
ρ0e ρ0 e
uu
e
V V
er
tin
e0 = 0
u
⇒ K + U =0
e e
1
Z Z
t=0
U = û (x,t) dV =
on
εe : C : εe dV = 0
.O
e
t=0 t=0 2 | {z
t=0 t=0
V V
C
}
=0
©
(6.91)
and the integration of (6.90) with the initial condition (6.91) leads to
K
e + Ue = 0 ∀t ≥ 0 , (6.92)
where
1
Z
K
e= v2 dV ≥ 0
ρ0 e ∀t ≥ 0 . (6.93)
2 |{z}
V
≥0
19 de f
Here, the definition |e
v| = e
v is used.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
294 C HAPTER 6. L INEAR E LASTICITY
εe (x,t) : C : εe (x,t) ≥ 0 ∀x ∈ V ∀t ≥ 0 =⇒
rs
1 (6.95)
Z
Ue = εe : C : εe dV ≥ 0 ∀t ≥ 0 .
ee
2
V
n
gi
Then, comparing (6.94) and (6.95) necessarily leads to
t d le En
)
Ue (t) ≤ 0
ar
1
Z
ar s
∀t ≥ 0 =⇒ U (t) = ε :C:e ε dV = 0 ∀t ≥ 0 . (6.96)
ge ro or
e
eS m
ib
e
U (t) ≥ 0
e 2
ac
fV
C d P cs
b
Considering once more the positive-definite condition of tensor C 20 ,
i
an an n
y ha
1
Z
Ue = εe : C : εe dV = 0 ∀t ≥ 0 =⇒ εe : C : εe = 0 ∀x , ∀t ≥ 0 (6.97)
le
2 | {z }
liv or ec
V ≥0
M
.A
d
uu
εe : C : εe = 0 ⇐⇒ εe (x,t) = 0 ∀x , ∀t ≥ 0 (6.98)
e
X Th
er
tin
.O
1 ∂ uei ∂ uej
©
S
εe (x,t) = ∇ · u
e = 0 =⇒ + =0 i, j ∈ {1, 2, 3} , (6.100)
2 ∂ x j ∂ xi
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Saint-Venant’s Principle 295
which is a system of six homogeneous and first-order PDEs. Its integration leads
to the solution21
e (x,t) = Ω
u e ·x + e c with
|{z} |{z}
rotation translation
(6.101)
0 −θe3 θe2 ce1
e not
Ω ≡ θe3
0 −θe1
and e
not
c ≡ ce2 ,
rs
−θe2 θe1 0 ce3
ee
where Ω e is an antisymmetric tensor (rotation tensor dependent on three con-
n
gi
stants θe1 , θe2 and θe3 ) and e
c is a constant vector equivalent to a translation. Ulti-
t d le En
mately, the solution (6.100) to the system (6.101) are the displacements u e (x,t)
compatible with a null strain εe (x,t) = 0, which correspond to a rigid body
ar
ar s
ge ro or
motion. The integration constants in Ω e and ec are determined by imposing the
eS m
ib
boundary conditions (6.82) (e u (x,t) = 0 ∀x ∈ Γu ), therefore, if the rigid body
ac
f
C d P cs
motion is impeded through the restrictions in Γu , one obtains Ω e = 0 and e c = 0.
In conclusion, b
i
an an n
y ha
)
e (x,t) = Ω
u e · x +e c
le
=⇒ u e (x,t) = u(2) − u(1) = 0 =⇒ u(2) = u(1) .
liv or ec
Ω≡0 ; e
e c≡0
M
.A
(6.102)
Finally, replacing (6.99) in (6.80) yields
m
d
uu
e
er
tin
.O
u(2) = u(1)
C
σ (2) = σ (1)
21 This solution can be obtained applying the methodology used in Chapter 3, Section 3.4.2
to integrate the strain field.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
296 C HAPTER 6. L INEAR E LASTICITY
rs
n ee
gi
Figure 6.11: Saint-Venant’s principle.
t d le En
ar
ar s
ge ro or
eS m
ib
boundary characterized by the traction vector t∗ (see Figure 6.11). These ac-
ac
tions will lead f
to a solution or response in displacements, strains and stresses,
C d P cs
not T
b
R (I) (x,t) ≡ u(I) (x,t), ε (I) (x,t) , σ (I) (x,t) . Consider now a part Γ̂ of the
i
an an n
place the system of actions applied on the boundary, t(I) , by another system,
le
t(II) , that is statically equivalent to t(I) 22 , without modifying the actions on the
liv or ec
rest of Γσ . Modifying the actions in this way will presumably result in the new
M
.A
not T
responses R (II) (x,t) ≡ u(II) (x,t) , ε (II) (x,t) , σ (II) (x,t) .
m
Saint-Venant’s principle states that, for the points belonging to the domain Ω
d
uu
that are sufficiently far from the boundary Γ̂ , the solution in both cases is prac-
e
X Th
on
ε (I) (x p ,t) ≈ ε (II) (x p ,t) ∀P δ ` . (6.105)
C
σ (I) (x p ,t) ≈ σ (II) (x p ,t)
In other words, if the distance δ between the point being considered and the
part of the boundary in which the actions have been modified is large in com-
parison with the dimension ` of the modified zone, the response in said point is
equivalent in both cases.
22 Two systems of forces t(I) and t(II) are said to be statically equivalent if the resultant
(forces and moments) of both systems is the same.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Linear Thermoelasticity. Thermal Stresses and Strains 297
Solution
Consider a beam (or prismatic piece) with a cross-section A subjected to
a tensile point force F in its ends, as shown in the figure below. The exact
solution to the original elastic problem (system (I)) is extremely complicated,
especially in the vicinity of the points of application of the point forces. If
the forces F are now replaced by a statically equivalent system of uniformly
rs
distributed tensile loads in the end sections σ = F/A (system (II)), the elastic
solution to the corresponding problem is extremely simple and coincides (for
ee
a Poisson’s ratio of ν = 0) with the axial stress solution provided by strength
n
of materials (uniformly distributed stresses in all the piece, σx = F/A). At
gi
a far enough distance from the beam’s ends (once or twice the edge), Saint-
t d le En
Venant’s principle allows approximating solution (I) with solution (II), and
ar
also allows dimensioning the strength characteristics of the piece for practical
ar s
ge ro or
eS m
ib
purposes.
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
298 C HAPTER 6. L INEAR E LASTICITY
not
θ (x,t) 6= θ (x, 0) = θ0 ,
. ∂ θ (x,t) (6.106)
θ (x,t) = 6= 0 .
∂t
Nevertheless, the hypothesis that the processes are adiabatic (slow) is maintained
and, thus,
ρ0 r − ∇ · q ≈ 0 . (6.107)
rs
ee
6.10.1 Linear Thermoelastic Constitutive Equation
n
Hooke’s law (6.6) in this case is generalized to
gi
t d le En
σ = C : ε − β (θ − θ0 )
, (6.108)
ar
ar s
ge ro or
σi j = Ci jkl εkl − βi j (θ − θ0 ) i, j ∈ {1, 2, 3}
eS m
ib
ac
f
C d P cs
Here, C is the tensor of elastic constants defined in (6.7), θ (x,t) is the tempera-
b
ture field, θ0 (x) = θ (x, 0) is the distribution of temperatures in the neutral state
i
an an n
le
liv or ec
(
Tensor of thermal β = βT
(6.109)
M
.A
properties βi j = β ji i, j ∈ {1, 2, 3}
m
d
uu
er
tin
(
C = λ 1 ⊗ 1 + 2µI
on
.O
Ci jkl = λ δi j δkl + µ δik δ jl + δil δ jk i, j, k, l ∈ {1, 2, 3}
C
(6.110)
©
(
β = β1
βi j = β δi j i, j ∈ {1, 2, 3}
where now a single thermal property β appears in addition to the elastic con-
stants λ and µ. Replacing (6.110) in the constitutive equation (6.108) and defin-
not
ing (θ − θ0 ) = ∆ θ , yields
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Linear Thermoelasticity. Thermal Stresses and Strains 299
Constitutive equation of an
isotropic linear thermoelastic material
(6.111)
σ = λ Tr (εε ) 1 + 2µεε − β ∆ θ 1
σi j = λ εll δi j + 2µεi j − β ∆ θ δi j i, j ∈ {1, 2, 3}
rs
ee
Equation (6.111) can be inverted as follows.
n
−1
−1 : β = C −1 : σ + ∆ θ α
σ = C : ε −∆θ β ⇒ ε = C : σ +∆θ C
gi
| {z }
t d le En
α
de f −1
α = C : β → Tensor of thermal expansion coefficients
ar
ar s
ge ro or
(6.112)
eS m
ib
ac
where α is a second-order (symmetric) tensor involving six thermal properties
f
C d P cs
named coefficients of thermal expansion. For an isotropic case, in agreement
b
with (6.111) and (6.24), and after certain algebraic manipulation, one obtains
i
an an n
y ha
le
Inverse constitutive equation of an
liv or ec
.A
ν 1+ν (6.113)
ε =− Tr (σ
σ)1+ σ + α∆ θ 1
m
E E
d
uu
e
ν 1+ν
X Th
εi j = − σll δi j + σi j + α∆ θ δi j i, j ∈ {1, 2, 3}
er
tin
E E
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
300 C HAPTER 6. L INEAR E LASTICITY
σ = λ Tr (εε ) 1 + 2µεε − β ∆ θ 1 = σ nt − σ t
| {z } | {z }
σ nt σt
(6.115)
Non-thermal stress → σ nt de
f
= λ Tr (εε ) 1 + 2µεε
rs
de f
Thermal stress → σt = β ∆θ 1
n ee
Here, σ nt represents the stress produced if there do not exist any thermal phe-
gi
nomena and σ t is named thermal stress and acts as the “correcting” stress due
t d le En
to the thermal increment.
ar
A similar operation can be performed on the inverse constitutive equations
ar s
ge ro or
eS m
ib
for the linear elastic and linear thermoelastic cases of (6.24) and (6.113), re-
ac
spectively, resulting in
f
C d P cs
ν 1+ν
b
i
ε = − Tr (σ
σ)1+ σ + α ∆ θ 1 = ε nt + ε t
an an n
E E
y ha
| {z } | {zt }
nt ε
le
ε
liv or ec
(6.116)
1+ν
M
.A
Non-thermal strain → ε nt de f ν
= − Tr (σ
σ)1+
σ
E E
m
de f
d
→ εt = α ∆θ 1
Thermal strain
uu
e
X Th
er
tin
In conclusion, the stress and strain tensors in linear thermoelasticity can be de-
composed into
on
.O
Non-thermal Thermal
C
σ nt = C : ε σt = ∆θ β
σ = σ nt − σ t Isotropic material: Isotropic material: (6.117)
σ nt = λ Tr (εε ) 1 + 2µεε σt = β ∆θ 1
ε nt = C −1 : σ εt = ∆θ α
ε = ε nt + ε t Isotropic material: Isotropic material: (6.118)
ν 1+ν εt = α ∆θ 1
ε nt = − Tr (σ σ)1+ σ
E E
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Linear Thermoelasticity. Thermal Stresses and Strains 301
where the thermal components appear due to the thermal processes being taken
into account. The following expressions result from (6.117) and (6.118).
ε nt = C −1 : σ σ = C : ε nt = C : ε − ε t
=⇒ (6.119)
σ nt = C : ε ε = C −1 : σ nt = C −1 : σ + σ t
=⇒ (6.120)
rs
Remark 6.11. Unlike what occurs in elasticity, in the thermoelastic
ee
case a state of null strain in a point of a medium does not imply a
n
state of null stress in said point. In effect, for ε = 0 in (6.117),
gi
ε = 0 =⇒ σ nt = 0 =⇒ σ = −σ
σ t = −β ∆ θ 1 6= 0 .
t d le En
ar
ar s
ge ro or
eS m
ib
ac
∆ θ 6= 0 f
C d P cs
σ t = −β ∆ θ 1
σ = −σ
b
i
ε =0
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
er
tin
.O
σ = 0 =⇒ ε nt = 0 =⇒ ε = ε t = α ∆ θ 1 6= 0 .
C
∆ θ 6= 0
ε = εt = α ∆θ 1
σ =0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
302 C HAPTER 6. L INEAR E LASTICITY
rs
n ee
gi
Figure 6.12: Actions on a continuous medium.
t d le En
ar
ar s
6.11 Thermal Analogies
ge ro or
eS m
ib
ac
f
The thermal analogies arise from the search of procedures to solve the linear
C d P cs
b
thermoelastic problem using the strategies and methodologies developed in Sec-
i
tion 6.7 for the linear elastic problem (without considering thermal effects).
an an n
y ha
Two analogies are presented in this section which, for the sake of simplicity,
are restricted to the isotropic quasi-static problem, although they can be directly
le
liv or ec
.A
Consider the continuous medium in Figure 6.12 on which the body forces b (x,t)
uu
e
er
tin
Γu and Γσ act the prescribed displacements u∗ (x,t) and a traction vector t∗ (x,t),
respectively.
on
.O
∇ · σ + ρ0 b = 0 Equilibrium equation
Governing
equations σ = C : ε −β ∆θ 1 Constitutive equation
(6.121)
ε = ∇S u
Geometric equation
(
Boundary Γu : u = u∗
conditions Γσ : σ · n = t∗
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Thermal Analogies 303
which compose the actions (data) A (x,t) and responses (unknowns) R (x,t) of
the problem24 .
b (x,t) *
MAT HEMAT ICAL u (x,t)
+
∗
u (x,t)
⇒ MODEL : ⇒ ε (x,t)
t∗ (x,t) PDEs + BCs (6.122)
σ (x,t)
∆ θ (x,t)
| {z }
| {z } not
not Responses = R (I) (x,t)
Actions = A (I) (x,t)
rs
ee
To be able to apply the resolution methods typical of the liner elastic problem
n
developed in Section 6.7, the thermal term in the equations of the thermoelastic
gi
problem (6.121) must be eliminated (at least, in appearance). To this aim, the
t d le En
decomposition of the stress tensor σ = σ nt −σ
σ t is replaced in (6.121) as follows.
ar
ar s
ge ro or
a) Equilibrium equation
eS m
ib
ac
σ = σ nt − σ tf =⇒
C d P cs
b
i
t
∇·σ = ∇ · σ nt − ∇ · σ = ∇ · σ nt − ∇ (β ∆ θ ) (6.123)
an an n
y ha
|{z}
β ∆θ1
le
liv or ec
1
∇ · σ + ρ0 b = 0 =⇒ ∇ · σ nt + ρ0 b − ∇ (β ∆ θ ) = 0
M
.A
ρ0
m
| {z }
not (6.124)
d
= b̂
uu
e
nt
=⇒ ∇ · σ + b̂ = 0
X Th
er
tin
.O
©
1
b̂ (x,t) = b (x,t) − ∇ (β ∆ θ )
ρ0 (6.125)
1 ∂ (β ∆ θ )
b̂i (x,t) = bi (x,t) − i ∈ {1, 2, 3}
ρ 0 ∂ xi
b) Constitutive equation
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
304 C HAPTER 6. L INEAR E LASTICITY
ε = ∇2 u (6.127)
d) Boundary condition in Γu
Γu : u = u∗ (6.128)
e) Boundary condition in Γσ
rs
ee
σ = σ nt − σ t
n
=⇒ σ nt · n − σ t · n = t∗ =⇒
σ · n = t∗
gi
t d le En
(6.129)
∗ ∗
σ nt ·n = t∗ + t
σ · n = t + β ∆ θ n =⇒ nt
Γσ : σ · n = t̂
ar
ar s
ge ro or
eS m
ib
| {z } | {z }
β ∆θ 1·n t̂∗
ac
f
C d P cs
where t̂∗ (x,t) is a pseudo-traction vector defined by
b
i
an an n
t̂∗ = t∗ + β ∆ θ n .
y ha
(6.130)
le
liv or ec
.A
∇ · σ nt + ρ0 b̂ = 0
m
Equilibrium
d
1
uu
with b̂ = b − ∇ (β ∆ θ ) equation
e
ρ0
X Th
er
Governing
tin
equations Constitutive
σ nt = C : ε = λ Tr (εε ) 1 + 2µεε
equation
on
.O
(6.131)
ε = ∇S u Geometric
C
equation
©
(
Boundary Γu : u = u∗
conditions Γσ : σ nt · n = t̂∗ with t̂ = t + β ∆ θ n
which constitutes the so-called analogous problem, a linear elastic problem that
can be solved with the methodology indicated for this type of problems in Sec-
tion 6.7 and characterized by the following actions and responses.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Thermal Analogies 305
b̂ (x,t) * MAT HEMAT ICAL + u (x,t)
∗
u (x,t) ⇒ MODEL : ⇒ ε (x,t)
PDEs + BCs (6.132)
t̂∗ (x,t)
nt
σ (x,t)
| {z } | {z }
not (II) not
Actions = A (x,t) Responses = R(II) (x,t)
Comparing the actions and responses of the original problem (6.122) with
those of the analogous problem (6.132), reveals the difference between them to
rs
be
ee
1
b − b̂ ∇ (β ∆ θ )
b b̂
n
∗ u∗ 0 ρ0
not u de f (III)
gi
(I)
A −A (II)
A ≡ ∗ − ∗ = ∗ ∗ =
0 =A (x,t)
t t̂ t − t̂
t d le En
−β ∆ θ n
∆θ 0 ∆θ
ar
∆θ
ar s
ge ro or
eS m
ib
0
u u 0
ac
R (I) −R
not
R(II) ≡ ε − ε =
f de f (III)
C d P cs
0 = 0 =R (x,t)
b
nt
i
σ σ σ −σ nt −β ∆ θ 1
an an n
y ha
| {z }
−σσt
le
(6.133)
liv or ec
.A
m
er
tin
.O
C
Equation (6.133) suggests that the original problem (I) may be interpreted as
©
STATE (II) (to be solved): analogous elastic state in which the temperature
does not intervene and that can be solved by means of elastic procedures.
+
STATE (III) (trivial): trivial thermoelastic state in which the responses
R(III) (x) given in (6.133) are known without the need of any calculations.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
306 C HAPTER 6. L INEAR E LASTICITY
Once STATE (II) is computed, the solution to the original thermoelastic prob-
lem of STATE (I) is obtained as
(I) (II)
Solution to the u = u
original thermoelastic ε = ε (II)
(I)
(6.134)
problem (I) (II)
σ = σ −β ∆θ1
The procedure to solve the thermoelastic problem based on the first thermal
analogy is summarized as a superposition of states in Figure 6.13.
rs
ee
STATE ACTION RESPONSE
n
gi
t d le En
b (x,t)
ar
u (x,t)
ar s
ge ro or
u∗ (x,t)
eS m
ib
ε (x,t)
∗
ac
f t (x,t)
σ (x,t)
C d P cs
∆ θ (x,t)
b
i
an an n
y ha
.A
1
b̂ = b − ∇ (β ∆ θ )
ρ0
m
u (x,t)
u∗ (x,t)
d
uu
ε (x,t)
e
∗ ∗
t̂ = t + β ∆ θ n
X Th
σ nt (x,t)
er
tin
∆θ = 0
on
.O
e = 1 ∇ (β ∆ θ )
b
ρ0 u=0
∗=0
u
e
ε =0
e∗
t = −β ∆ θ n
σ = −β ∆ θ 1
∆ θ (x,t)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Thermal Analogies 307
rs
(6.135)
ε = ∇S u
Geometric equation
n ee
(
Boundary Γu : u = u∗
gi
conditions Γσ : σ · n = t∗
t d le En
ar
ar s
which constitute the actions (data) A (x,t) and responses (unknowns) R (x,t) of
ge ro or
eS m
ib
the problem.
ac
f
C d P cs
b
i
b (x,t) *
an an n
MAT HEMAT ICAL u (x,t)
+
y ha
∗
u (x,t)
⇒ MODEL : ⇒ ε (x,t)
le
t∗ (x,t)
liv or ec
PDEs + BCs
σ (x,t) (6.136)
∆ θ (x,t)
M
.A
| {z }
| {z } not
not Responses = R (I) (x,t)
m
er
tin
.O
α (x) and the thermal increment ∆ θ (x,t) are such that the thermal
strain field
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
308 C HAPTER 6. L INEAR E LASTICITY
rs
rotation translation
ee
| {z }
rigid body motion
n
gi
The rigid body motion may be chosen arbitrarily (in the form which
t d le En
is most convenient for the resolution process).
ar
ar s
ge ro or
eS m
ib
ac
f
Once the thermal displacements have been defined, a decomposition of the
C d P cs
total displacements into their thermal and non-thermal parts can be performed
b
i
as follows.
an an n
y ha
de f
unt (x,t) = u (x,t) − ut (x,t) =⇒ u = unt + ut (6.138)
le
liv or ec
To eliminate the thermal term in the equations that constitute the thermoe-
M
.A
result in
e
X Th
er
∇ · σ + ρ0 b = 0
on
(6.139)
.O
C
c) Geometric equation
ε = ∇S u = ∇S (unt + ut ) = ∇S unt + ∇S ut = ∇S unt + ε t
=⇒ ε nt = ∇S unt
|{z}
εt
ε = ε nt + ε t
(6.141)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Thermal Analogies 309
d) Boundary condition in Γu
u = u∗
=⇒ Γu : unt = u∗ − ut (6.142)
u = unt + ut
Γσ : σ · n = t∗ (6.143)
rs
Equations (6.139) to (6.143) allow rewriting the original problem (6.135) as
n ee
gi
Equilibrium
∇ · σ + ρ0 b = 0
equation
t d le En
ε = C −1 : σ =
nt
ar
Governing
ar s
Constitutive
ge ro or
eS m
ib
equations ν 1+ν equation
= − Tr (σ σ)1+
ac
σ
f
E E (6.144)
C d P cs
b
Geometric
ε nt = ∇S unt
i
an an n
equation
y ha
(
Γu : u = u∗ − ut
le
Boundary
liv or ec
conditions Γσ : σ · n = t∗
M
.A
m
which constitutes the so-called analogous problem, a linear elastic problem char-
d
er
tin
nt
b̂ (x,t) u (x,t)
on
MAT HEM.
.O
* +
u∗ (x,t) − ut (x,t) ⇒ MODEL : ⇒ ε nt (x,t)
C
| {z } | {z }
not not
Actions = A (II) (x,t) Responses = R (II) (x,t)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
310 C HAPTER 6. L INEAR E LASTICITY
Comparing the actions and responses of the original problem (6.136) and the
analogous problem (6.145), reveals the difference between them to be
b b 0
∗ ∗ t t
A (I) −A
not u
A(II) ≡ u − u u de f (III) (x,t)
t∗ − t∗ = 0 = A
∆θ 0 ∆θ
(6.146)
nt t t
u u u u
(I) not de f
(II)
R ≡ ε − ε = ε = α ∆ θ 1 = R (III) (x,t)
rs
R −R nt t
ee
σ σ 0 0
n
gi
where equations (6.138) and (6.118) have been taken into account.
t d le En
ar
ar s
ge ro or
Remark 6.15. It can be directly verified that, in (6.146), R(III) is the
eS m
ib
response corresponding to the system of actions A (III) in the ther-
ac
f
C d P cs
moelastic problem (6.135).
b
i
an an n
y ha
le
Therefore, the original problem (I) can be interpreted as the sum (superposi-
liv or ec
.A
STATE (II) (to be solved): analogous elastic state in which the temperature
m
does not intervene and that can be solved by means of elastic procedures.
uu
e
+
X Th
er
R (III) (x) given in (6.146) are known without the need of any calculations.
on
.O
C
Once STATE (II) is computed, the solution to the original thermoelastic prob-
©
where ut is known from the integration process of the thermal strain field
in (6.137). The procedure to solve the thermoelastic problem based on the sec-
ond thermal analogy is summarized as a superposition of states in Figure 6.14.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Thermal Analogies 311
rs
b (x,t)
ee
u∗ (x,t) u (x,t)
ε (x,t)
n
∗
t (x,t)
gi
σ (x,t)
∆ θ (x,t)
t d le En
ar
(I) Thermoelastic (original)
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
b (x,t)
unt (x,t)
y ha
u∗ − ut
nt
ε (x,t)
le
∗
liv or ec
t (x,t)
σ (x,t)
∆θ = 0
M
.A
m
er
tin
b=0
on
.O
u = ut (x,t)
e∗ = ut
u
C
ε = α ∆θ1
et∗ = 0
©
σ =0
∆ θ (x,t)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
312 C HAPTER 6. L INEAR E LASTICITY
Example 6.5 – Solve the problem of a beam fully-fixed at its ends and sub-
jected to a constant thermal increment ∆ θ using the second thermal analogy.
Solution
The classic procedure followed in strength of materials to solve this problem
consists in the superposition (sum) of the following situations: 1) The struc-
ture is initially considered to be hyperstatic; 2) the right end is freed to allow
for thermal expansion, which takes place with null stresses (since it is an iso-
rs
static structure); and 3) the displacement of the beam’s right end is recovered
until it is brought again to zero.
ee
This procedure coincides exactly with the application of the second thermal
n
analogy in which the thermal displacement field ut is defined by the thermal
gi
expansion of the piece with its right end freed (state III). Said expansion
t d le En
produces a displacement in the right end of value u|x=` = α ∆ θ ` and, when
ar
recovering the displacement at this end, the boundary condition
ar s
ge ro or
eS m
ib
Γu : u = u∗ − ut = −ut ,
ac
f
C d P cs
|{z}
0
b
i
an an n
plied.
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Thermal Analogies 313
rs
n ee
gi
t d le En
Remark 6.18. In the case involving
ar
ar s
• a homogeneous material (α
ge ro or
α (x) = const. = α)
eS m
ib
• a constant thermal increment (∆ θ = const.)
ac
f
C d P cs
the integration of the thermal strain field ε t = ∆ θ α1 = const. is triv-
b
i
ial, resulting in
an an n
y ha
Ω ∗ · x + c∗
ut (x,t) = α ∆ θ x +Ω ,
le
liv or ec
| {z }
rigid body motion
M
.A
where the rigid body motion can be chosen arbitrarily (see Re-
m
er
tin
ut (x,t) = α ∆ θ x =⇒ x + ut = x + α ∆ θ x = (1 + α ∆ θ ) x ,
on
.O
which means that STATE (III) in the second thermal analogy (see
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
314 C HAPTER 6. L INEAR E LASTICITY
rs
thermal increment.
n ee
6.12 Superposition Principle in Linear Thermoelasticity
gi
t d le En
Consider the linear thermoelastic problem in Figure 6.16 and its corresponding
ar
governing equations
ar s
ge ro or
eS m
ib
ac
∂ 2u f
C d P cs
∇ · σ + ρ0 b = ρ0 Cauchy’s equation
∂t 2
b
i
an an n
y ha
| {z }
C:ε
M
.A
1
ε = ∇S u = (u ⊗ ∇ + ∇ ⊗ u) Geometric equation
m
2
d
uu
e
X Th
er
)
Γu : u = u∗
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Superposition Principle in Linear Thermoelasticity 315
)
u (x, 0) = 0
. Initial conditions (6.150)
u (x, 0) = v0
rs
∆ θ (x,t) PDEs + BCs
σ (x,t) (6.151)
ee
v0 (x)
n
| {z }
not
gi
| {z }
not
Responses = R (x,t)
Actions = A (x,t)
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 6.19. The different (scalar, vector, tensor and differential)
C d P cs
b
operators that intervene in the governing equations of the problem
i
an an n
(6.148) to (6.150) are linear, that is, given any two scalars a and b,
y ha
le
∇ · (•) → linear =⇒ ∇ · (a x + b y) = a ∇ · x + b ∇ · y ,
liv or ec
M
.A
C : (•) → linear =⇒ C : (a x + b y) = a C : x + b C : y ,
m
∇S (•) → linear =⇒ ∇S (a x + b y) = a ∇S x + b ∇S y ,
uu
e
X Th
er
∂2 ∂ 2 (a x + b y) ∂ 2x ∂ 2y
tin
(•) → linear =⇒ = a + b .
∂t 2 ∂t 2 ∂t 2 ∂t 2
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
316 C HAPTER 6. L INEAR E LASTICITY
rs
The solution (response) to the system of actions
ee
A (3) = λ (1) A (1) + λ (2) A (2)
n
gi
(where λ (1) and λ (2) are any two scalars) is
t d le En
R (3) = λ (1) R (1) + λ (2) R (2)
ar
ar s
ge ro or
eS m
ib
ac
f
In other words, the solution to the linear thermoelastic problem
C d P cs
when considering a linear combination of different systems of ac-
b
i
tions is the same linear combination of the individual solutions to
an an n
le
liv or ec
M
.A
Proof
m
Replacing the actions A (3) = λ (1) A (1) + λ (2) A (2) and the responses
d
uu
R = λ (1) R (1) + λ (2) R (2) in the equations of the problem, and taking into
(3)
e
X Th
er
account the linearity of the different operators (see Remark 6.19) yields
tin
a) Cauchy’s equation
on
.O
C
| {z } | {z }
∂ 2 u(1) ∂ 2 u(2)
ρ0 ρ0
∂t 2 ∂t 2
∂ 2 λ (1) u(1) + λ (2) u(2) ∂ 2 u(3)
= ρ0 = ρ 0
∂t 2 ∂t 2
∂ 2 u(3)
∇ · σ (3) + ρ0 b(3) = ρ0
∂t 2
(6.154)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Superposition Principle in Linear Thermoelasticity 317
b) Constitutive equation
σ (3) − C : ε (3) − β ∆ θ (3) 1 = λ (1) σ (1) − C : ε (1) − β ∆ θ (1) 1 +
| {z }
= 0
λ (2) σ (2) − C : ε (2) − β ∆ θ (2) 1 = 0
| {z }
=0
rs
(6.155)
n ee
c) Geometric equation
gi
t d le En
ε (3) − ∇S u(3) = λ (1) ε (1) − ∇S u(1) +λ (2) ε (2) − ∇S u(2) = 0
ar
ar s
| {z } | {z }
ge ro or
=0 =0 (6.156)
eS m
ib
ac
f
ε (3) = ∇S u(3)
C d P cs
b
i
an an n
d) Boundary condition in Γu
y ha
le
u(3) − u∗(3) = λ (1) u(1) − u∗(1) +λ (2) u(2) − u∗(2) = 0
liv or ec
M
.A
| {z } | {z }
=0 =0 (6.157)
m
Γu : u(3) = u∗(3)
d
uu
e
X Th
er
tin
e) Boundary condition in Γσ
on
.O
σ (3) · n − t∗(3) = λ (1) σ (1) · n − t∗(1) +λ (2) σ (2) · n − t∗(2) = 0
C
| {z } | {z }
(6.158)
©
=0 =0
Γσ : σ (3) · n = t∗(3)
f) Initial conditions
. (3)
. (1)
. (2)
u(3) (x, 0) − v0 = λ (1) u(1) (x, 0) − v0 +λ (2) u(2) (x, 0) − v0 = 0
| {z } | {z }
=0 =0
. (3)
u(3) (x, 0) = v0
(6.159)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
318 C HAPTER 6. L INEAR E LASTICITY
not T
Consequently, R (3) = λ (1) R (1) + λ (2) R (2) ≡ u(3) , ε (3) , σ (3) is the solu-
tion to the thermoelastic problem subjected to the actions: A (3) = λ (1) A (1) +
λ (2) A (2) (QED).
rs
and to “economize” in writing, only these six different components are used in
ee
engineering, and they are expressed in the form of the stress and strain “vec-
n
tors”. These are constructed in R6 , systematically arranging the elements of the
gi
upper triangle of the matrix of components of the corresponding tensor in the
t d le En
following manner25 .
ar
ar s
ge ro or
eS m
ib
σx
ac
f
σy
C d P cs
σx τxy τxz
b
de f σz
i
not
σ ≡ τxy σy τyz → {σ σ} = (6.160)
an an n
τxy
y ha
τxz τyz σz
le
τ
xz
liv or ec
τyz
M
.A
The same arrangement is followed in the case of the strains, with the particularity
m
that the strain vector {εε } is constructed using the angular strains γxy = 2 εxy ,
uu
e
er
tin
εx
on
.O
1 1
εx γxy γxz εy
2 2
C
εx εxy εxz
©
ε
de f
ε ≡ εxy εy εyz = γxy εy 1 γyz → {εε } = (6.161)
not not 1 z
2 2 γxy
εxz εyz εz
1
1 γ
xz
γxz γyz εz
2 2 γyz
25 The notation {x} is used to denote the vector in R6 constructed from the symmetric ten-
sor x.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Hooke’s Law in terms of the Stress and Strain “Vectors” 319
σ :ε = {σ
σ } · {εε } ⇐⇒ σi j εi j = σi εi
| {z } | {z }
second-order vectors
tensors
rs
nitions in (6.160) and (6.161).
n ee
gi
The inverse constitutive equation (6.113),
t d le En
1+ν
ar
ν
ε =−
ar s
Tr (σ
σ)1+ σ + α∆ θ 1 , (6.162)
ge ro or
eS m
ib
E E
ac
f
can now be rewritten in terms of the stress and strain vectors as
C d P cs
b
i
C−1 · {σ
{εε } = Ĉ σ } + {εε }t ,
an an n
(6.163)
y ha
.A
1 −ν −ν
0 0 0
E E E
m
d
−ν 1 −ν
uu
e
0 0 0
X Th
E E E
er
tin
−ν −ν 1
0 0 0
on
.O
E E E
C−1 ≡
not
(6.164)
Ĉ
C
1
0
©
0 0 0 0
G
1
0 0 0 0 0
G
1
0 0 0 0 0
G
and {εε }t is a thermal strain vector defined by means of an adequate translation
of the thermal strain tensor ε t = α ∆ θ 1,
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
320 C HAPTER 6. L INEAR E LASTICITY
α ∆θ
α ∆θ
α ∆θ 0 0
t de f α ∆ θ
not
εt ≡
0 α ∆ θ 0 → {εε } = . (6.165)
0
0 0 α ∆θ 0
0
rs
the stress and strain vectors,
ee
Hooke’s law
n
C · {εε } − {εε }t
gi
in terms of the {σ
σ } = Ĉ (6.166)
t d le En
stress and strain vectors
ar
ar s
ge ro or
eS m
ib
where Ĉ
C is the matrix of elastic constants.
ac
f
C d P cs
b
ν ν
1 0 0 0
i
1−ν 1−ν
an an n
y ha
ν ν
1−ν 1 0 0 0
1−ν
le
liv or ec
ν ν
1 0 0 0
M
.A
1−ν 1−ν
not E (1 − ν)
(6.167)
C≡
Ĉ
(1 + ν) (1 − 2ν) 1 − 2ν
m
0 0 0 0 0
d
2 (1 − ν)
uu
e
1 − 2ν
X Th
0 0 0 0 0
er
tin
2 (1 − ν)
1 − 2ν
on
0 0 0 0 0
.O
2 (1 − ν)
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 321
P ROBLEMS
Problem 6.1 – Justify whether the following statements are true or false.
a) The terms isentropic and adiabatic are equivalent when dealing with a
thermoelastic material.
b) The second thermal analogy is always applicable to linear thermoelas-
rs
tic materials.
n ee
gi
Solution
t d le En
ar
a) According to the second law of thermodynamics (5.114),
ar s
ge ro or
eS m
ib
i . .
ac
= ρ0 θ s − (ρ0 r − ∇ · q) ≥ 0 .
ρ0 θ sloc
f
C d P cs
b
All processes are reversible in the case of a thermoelastic material and, thus, the
i
an an n
. .
le
i = ρ θ s − (ρ r − ∇ · q) = 0 .
liv or ec
ρ0 θ sloc 0 0 [1]
M
.A
.
An isentropic process (entropy remains constant) is characterized by s = 0. On
m
er
ρ0 r − ∇ · q = 0 .
tin
on
.
©
ρ0 θ s − (ρ0 r − ∇ · q) = 0 =⇒ ρ0 r − ∇ · q = 0 .
|{z}
=0
Conversely, if an adiabatic process is assumed, and its mathematical expression
is introduced in [1], the definition of an isentropic process is obtained,
. .
ρ0 θ s − (ρ0 r − ∇ · q) = 0 =⇒ s=0.
| {z }
=0
In conclusion, the statement is true.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
322 C HAPTER 6. L INEAR E LASTICITY
b) The second thermal analogy is not always applicable. The condition that the
thermal strain field be integrable must be verified, that is, the thermal strain field
ε t (x,t) must satisfy the compatibility conditions (3.19),
rs
n ee
Problem 6.2 – An isotropic linear elastic solid is subjected to a constant pres-
gi
sure of value p on all of its external boundary, in addition to a thermal incre-
t d le En
ment of ∆ θ = θ (x, y, z) in its interior. Both actions cancel each other out such
that no displacements are observed in the solid. Obtain the value of ∆ θ in each
ar
ar s
point of the solid.
ge ro or
eS m
ib
ac
f
C d P cs
b
i
Solution
an an n
y ha
The first thermal analogy described in Section 6.11.1 will be applied. To this
le
liv or ec
aim, the original problem I is decomposed into the sum of problems II and III as
described in Figure 6.13.
M
.A
m
P ROBLEM I
d
uu
e
b=0
X Th
er
u
tin
t∗ = −p n
in Γσ
Actions: Responses: ε
on
.O
u∗ = 0 in Γu
σ
C
∆θ = ∆θ
©
P ROBLEM III
This problem is solved first since its solution is trivial.
1
bIII = β ∇ (∆ θ )
ρ uIII = 0
∗
t
Actions: III = −β ∆ θ n in Γ σ Responses: ε III = 0
u ∗ =0 in Γ
u σ III = −β ∆ θ 1
III
∆ θIII = ∆ θ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 323
P ROBLEM II
1
bII = β ∇ (∆ θ )
∗ ρ
Actions: tII = (−p + β ∆ θ ) n in Γσ
u∗II = u∗ = 0 in Γu
∆ θII = 0
To solve problem II, Navier’s equation (6.62) is taken into account, together with
the fact that uII = 0.
rs
ee
(λ + µ) ∇ (∇ · uII ) + µ∇2 uII + ρbII = 0 =⇒
n
gi
bII = 0 =⇒ β ∇ (∆ θ ) = 0 =⇒ ∆ θ is uniform
t d le En
ar
In addition, uII = 0 also results in
ar s
ge ro or
eS m
ib
ac
1 f
(uII ⊗ ∇ + ∇ ⊗ uII ) = 0 ,
C d P cs
ε II =
2
b
i
an an n
le
liv or ec
Since the traction vector t∗II is defined in terms of the stress tensor σ II ,
M
.A
σ II · n = t∗II = (−p + β ∆ θ ) n = 0 ∀n =⇒ −p + β ∆ θ = 0 ,
m
d
uu
er
tin
p
∆θ = .
β
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
324 C HAPTER 6. L INEAR E LASTICITY
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
Hypotheses:
le
liv or ec
1) Young’s modulus: E
2) Poisson’s coefficient: ν = 0
M
.A
er
.O
C
Solution
©
First phase
The cylindrical shell has not come into contact with the rigid walls of the
cavity. The boundary condition on the lateral walls, both internal and ex-
ternal, will be null radial stress. The two cylinders will come into contact
when
ur (r = 2R) = a .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 325
Second phase
The cylindrical shell and the rigid walls of the cavity are in contact and,
therefore, the boundary condition on the external lateral wall is different
than that of the first phase. In this case, a null radial displacement will
be imposed. Nonetheless, the internal wall will retain the same boundary
condition as in the previous phase.
A positive ∆ θ will reduce the internal radius since the external radius
cannot increase because it is limited by the infinitely rigid walls of the
cavity. Then, the only possibility is that the cylindrical shell continues
expanding inwards. There will be a point in which the internal radius,
rs
which had increased in the first phase, will recover its initial value.
n ee
The first thermal analogy (see Section 6.11.1) and the superposition principle
gi
(see Section 6.12) will be applied. To this aim, the original problem (problem I)
t d le En
is decomposed into the sum of problems II and II as described in Figure 6.13.
P ROBLEM III
ar
ar s
ge ro or
eS m
The actions in problem III, the trivial problem, are
ib
ac
1 f
C d P cs
b
bIII = ∇ · (β
β ∆θ) .
i
ρ
an an n
y ha
bIII = 0 .
M
.A
2) Prescribed stresses in Γσ : t = −β
β ∆ θ n = −β ∆ θ n.
X Th
er
tin
.O
uIII = 0
C
ε III = 0 [1]
©
σ III = −β ∆ θ 1
P ROBLEM II
The actions in problem II, the analogous problem, are
1
bII = b − ∇ · (β
β ∆θ) .
ρ
Here, b = 0 because the weight of the cylinder is assumed to be negligible and
the second term is zero, as seen in problem III. Therefore,
bII = 0 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
326 C HAPTER 6. L INEAR E LASTICITY
rs
not
uII (r, z) ≡ [ur (r) , 0 , uz (z)]T .
ee
In addition, uz (z) = 0 will be imposed in all points since no information on the
n
top and bottom surfaces of the cylindrical shell is given. Boundary conditions
gi
in displacements cannot be imposed for these surfaces because there is no way
t d le En
to determine the integration constants of uz that would appear if uz 6= 0 were
considered. Therefore, the displacement vector
ar
ar s
ge ro or
eS m
ib
not
uII (r, z) ≡ [ur (r) , 0 , 0 ]T
ac
f
C d P cs
b
is adopted. Navier’s equation (6.62) will be used to solve this problem,
i
an an n
y ha
∂ 2 uII
(λ + µ) ∇ (∇ · uII ) + µ∇2 uII + ρ0 bII = ρ0 =0.
le
∂t 2
liv or ec
M
.A
Note that the problem requires working in cylindrical coordinates and, thus, the
equation must be adapted to this system of coordinates. Given the simplifications
m
introduced into the problem, only the radial component of the equation will
d
uu
er
tin
∂ e 2G ∂ ωz ∂ ωθ ∂ 2 ur
(λ + 2G) − + 2G + ρbr = ρ 2 , [2]
on
.O
∂r r ∂θ ∂z ∂t
C
1 ∂ ur ∂ uz
ωθ = + =0,
2 ∂z ∂r
1 1 ∂ (r uθ ) 1 ∂ ur
ωz = − =0,
2 r ∂r r ∂θ
1 ∂ (r ur ) 1 ∂ uθ ∂ uz 1 ∂ (r ur )
e= + + = .
r ∂r r ∂θ ∂z r ∂r
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 327
rs
ee
The problem can be considered to be a quasi-static and, taking into account
bII = 0 and the relations derived in [3], the Navier’s stokes equation [2] is re-
n
gi
duced to
∂e ∂ 1 ∂
t d le En
(λ + 2G) = 0 =⇒ E (r ur ) = 0 .
∂r ∂r r ∂r
ar
ar s
ge ro or
Integrating this last expression leads to
eS m
ib
ac
1 ∂ ∂ f
C d P cs
(r ur ) = 2A =⇒ (r ur ) = 2Ar =⇒ r ur = Ar2 + B
r ∂r ∂r
b
i
[4]
an an n
T
y ha
B not B
=⇒ ur = Ar + =⇒ uII (r) ≡ Ar + , 0 , 0 ,
le
r r
liv or ec
M
.A
where A and B are the integration constants. The strain tensor corresponding
m
tion (6.3),
uu
e
X Th
B
er
tin
A− 2 0 0
r
ε II (r) ≡ 0 A + B 0 .
not
on
[5]
.O
r2
C
0 0 0
©
First phase
The integration constants A and B must be determined by means of the
boundary conditions. Stresses can be imposed in both lateral walls of the
cylindrical shell as follows.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
328 C HAPTER 6. L INEAR E LASTICITY
B OUNDARY CONDITION AT r = 2R
If r = 2R and according to the boundary conditions in Γσ of the analogous
problem,
tII = σ II · n = t∗ + β ∆ θ n = t∗ + β ∆ θ n .
Here, the following is known:
n = [1 , 0 , 0 ]T : outward unit normal vector.
t∗ = 0, since, for this phase, problem I has no loading on the lateral walls.
σ II is given by [5] and [6].
rs
Therefore, the boundary condition is reduced to
ee
σrr (r = 2R) = β ∆ θ ,
n
gi
which, replacing the value of the radial stress from [6] and, consider-
t d le En
ing [3], results in
ar
ar s
ge ro or
eS m
ib
B
A− = α ∆θ . [7]
ac
4R2 f
C d P cs
b
i
an an n
B OUNDARY CONDITION AT r = R
y ha
le
If r = R and according to the boundary conditions in Γσ of the analogous
liv or ec
problem,
M
.A
tII = σ II · n = t∗ + β ∆ θ n = t∗ + β ∆ θ n .
m
er
tin
t∗ = 0, since, for this phase, problem I has no loading on the lateral walls.
σ II is given by [5] and [6].
on
.O
σrr (r = R) = β ∆ θ ,
which, replacing the value of the radial stress from [6] and, consider-
ing [3], results in
B
A− = α ∆θ . [8]
R2
From [7] and [8], the values
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 329
are obtained. Now, replacing [9] in [4], [5] and [6] results in the displace-
ments, strains and stresses of the analogous problem.
not T
uII ≡ α ∆ θ r , 0 , 0
α ∆θ 0 0
not
ε II ≡ 0 α ∆ θ 0
0 0 0 [10]
rs
Eα ∆ θ 0 0
ee
not
σ II ≡ 0 Eα ∆ θ 0
n
0 0 0
gi
t d le En
Taking into account the superposition principle (see Section 6.12), and
ar
expressions [1], [3] and [10], the original problem is solved for the first
ar s
ge ro or
eS m
ib
phase.
ac
f
not T
u ≡ α ∆θr, 0, 0
C d P cs
b
i
an an n
α ∆θ 0 0
y ha
not
ε ≡ 0 α ∆θ 0
le
liv or ec
0 0 0 [11]
M
.A
0 0 0
m
not
σ ≡ 0 0 0
d
uu
e
0 0 −Eα ∆ θ
X Th
er
tin
To obtain the value of ∆ θ ∗ for which the external lateral walls of the cylindrical
on
.O
shell and the rigid walls of the cavity come into contact, it is enough to impose
C
that
ur (r = 2R) = a =⇒ α ∆ θ ∗ 2R = a .
©
Then, the temperature field required for the external lateral walls of the cylindri-
cal shell and the rigid walls of the cavity to come into contact is
a
∆θ∗ = . [12]
2αR
b) First, the value ∆ θ ∗∗ for which the internal radius recovers its initial position
will be determined. To this aim, the same geometry as in the initial problem will
be used, but now there will exist contact between the cylindrical shell and the
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
330 C HAPTER 6. L INEAR E LASTICITY
rigid walls of the cavity, which corresponds to the second phase defined in the
previous section. So, a new problem must be solved, with the same geometry as
before but considering different boundary conditions.
Second phase
The first phase will be obviated in this section, but one must bear in mind
that the solid now starts from a state that results from the previous phase,
that is, it has already suffered certain displacements, strains, stresses and
thermal increments. The variable ∆ θ will be used.
rs
As before, the first thermal analogy will be applied. Problem III remains
unchanged and, thus, so does its result [1]. Therefore, problem II must be
ee
solved with the same expressions [4], [5] and [6]. The integration con-
n
stants A and B must be determined by means of the boundary conditions.
gi
Stresses can be imposed on the internal lateral wall of the cylindrical shell
t d le En
and displacements, on its external lateral wall.
ar
ar s
ge ro or
eS m
ib
B OUNDARY CONDITION AT r = 2R
ac
f
If r = 2R and according to the boundary conditions in Γu of the analogous
C d P cs
problem,
b
i
an an n
ur (r = 2R) = 0 .
y ha
B
A 2R + =0 .
M
[13]
.A
2R
m
d
uu
B OUNDARY CONDITION AT r = R
e
X Th
er
tin
.O
tII = σ II · n = t∗ + β ∆ θ n = t∗ + β ∆ θ n .
C
not
n ≡ [−1 , 0 , 0 ]T : outward unit normal vector.
t∗ = 0, since, for this phase, problem I has no loading on the lateral walls.
σ II is given by [5] and [6].
Therefore, the boundary condition is reduced to
σrr (r = R) = β ∆ θ ,
which, replacing the value of the radial stress from [6], and consider-
ing [3], results in
B
A− 2 = α ∆θ . [14]
R
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 331
1 4
A = α ∆θ and B = − α ∆ θ R2 [15]
5 5
are obtained. Introducing now [15] in [4], [5] and [6] results in the dis-
placements, strains and stresses of the analogous problem.
T
4R2
1
not
uII ≡ α ∆θ r − , 0, 0
5 r
rs
ee
4R2
1
5 α ∆ θ 1 + r2 0 0
n
gi
4R2
not
ε II ≡ 1
0 α ∆θ 1− 2 0
t d le En
5 r
ar
[16]
ar s
0 0 0
ge ro or
eS m
ib
ac
f
C d P cs
4R2
1
5 Eα ∆ θ 1 + r2 b0 0
i
an an n
4R2
y ha
not
σ II ≡ 1
0 Eα ∆ θ 1 − 0
le
5 r2
liv or ec
M
0 0 0
.A
Taking into account the superposition principle (see Section 6.12), and
m
expressions [1], [3] and [16], the original problem is solved for the second
uu
e
phase.
X Th
T
er
4R2
tin
not 1
u ≡ α ∆θ r − , 0, 0
5 r
on
.O
4R2
C
1
α ∆ θ 1 + 0 0 [17a]
©
5 r2
2
not
ε ≡ 1 4R
0 α ∆θ 1− 2 0
5 r
0 0 0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
332 C HAPTER 6. L INEAR E LASTICITY
R2
4
5 Eα ∆ θ −1 + r2 0 0
2
not
σ ≡ 4 R
0 Eα ∆ θ −1 − 2 0 [17b]
5 r
0 0 −Eα ∆ θ
Note that, up to this point, the second phase has been solved assuming
an initial neutral state. In reality, this phase starts from the final state of
the first phase, which has the displacements, strains, stresses and thermal
rs
increments corresponding to ∆ θ = ∆ θ ∗ ,
ee
uinitial = u f irst phase (∆ θ = ∆ θ ∗ ) ,
n
ε initial = ε f irst phase (∆ θ = ∆ θ ∗ ) ,
gi
[18]
t d le En
σ initial = σ f irst phase (∆ θ = ∆ θ ∗ ) .
ar
ar s
ge ro or
eS m
ib
In fact, the variable ∆ θ in [17] is not a total thermal increment but the
difference in temperature at the moment corresponding to ∆ θ ∗ , that is,
ac
f
C d P cs
b
∆θ = ∆θ −∆θ∗ .
i
[19]
an an n
y ha
Then, considering [17], [18] and [19], the actual displacements, strains
le
liv or ec
.A
usecond phase = uinitial + u ∆ θ ,
m
ε second phase = ε initial + ε ∆ θ , [20]
d
uu
σ second phase = σ initial + σ ∆ θ .
e
X Th
er
tin
.O
to that of the second phase. In this way, the total displacement will be null.
C
a
δ1 = ur (r = R, ∆ θ = ∆ θ ∗ ) = α ∆ θ ∗ R = . [21]
2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 333
5a a 4a
∆ θ ∗∗ = ∆ θ ∗∗ + ∆ θ ∗ = + =⇒ ∆ θ ∗∗ =
6αR 2αR 3αR
Now, the curve δ −∆ θ can be plotted, where δ is the displacement of the internal
radius of the cylindrical shell.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
c) Expressions [11] and [17] must be used to plot the curves σrr − ∆ θ ,
m
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
334 C HAPTER 6. L INEAR E LASTICITY
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 335
E XERCISES
rs
b) Curve p − σA , where σA is the stress normal to the bottom contact sur-
ee
face at point A.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Additional hypotheses:
ac
f
C d P cs
1) Weights can be neglected.
b
i
2) Lamé’s constants: λ = µ
an an n
y ha
.A
m
d
uu
e
X Th
er
tin
on
.O
C
6.2 – The solid sphere A with external radius R1 and the solid spherical B, with
external radius R2 are composed of the same material. The external surface of
A and the internal surface of B are separated by a very small distance “a”
(a R1 and a R2 ).
a) Determine what value of the uniform normal pressure p shown in the
figure is required for the two surfaces to be in contact.
b) Plot, indicating the most significant values, the curve p − δ , where δ is
the shortening of R2 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
336 C HAPTER 6. L INEAR E LASTICITY
Additional hypotheses:
1) Young’s modulus: E
2) Lamé’s constants: λ = µ
3) R1 = R
4) R2 = 2 R
rs
n ee
gi
t d le En
ar
6.3 – Two solid cylinders composed of different elastic materials are vertically
ar s
ge ro or
eS m
ib
superimposed and confined between two infinitely rigid walls. The cylinders are
ac
f
subjected to the external pressures p and α p (p > 0, α > 0) as shown in the
C d P cs
b
figure.
i
an an n
b) Indicate the boundary conditions that need to be applied for the different
M
.A
two cylinders does not have a vertical displacement, calculate the inte-
uu
e
er
tin
on
.O
C
Additional hypotheses:
©
1) Top cylinder: λ1 = µ1
2) Bottom cylinder: λ2 = µ2
3) The friction between the cylin-
ders and between the cylinders
and the walls is assumed to be
null.
4) Weights can be neglected.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 337
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
Data:
y ha
Ri = 1
le
liv or ec
Re = 2
M
.A
ν =0
m
E (Young’s modulus)
d
uu
e
X Th
er
tin
Additional hypotheses:
1) The constitutive law of the elastic gasket is p∗ = K δ ∗ , where p∗ is the pres-
on
.O
sure acting on the gasket, δ ∗ is the shortening of its thickness and K is its
C
elastic modulus.
©
2) e Ri
3) A plane strain behavior in an infinitesimal strain framework may be as-
sumed.
6.5 – The figure below schematizes the layout of a railway rail composed of
straight rails of length “L”, separated by an elastic gasket with elastic modu-
lus K. Due to symmetry and construction considerations, it can be assumed that
the section x = 0 suffers no longitudinal displacements and the inferior part of
the rail suffers no vertical displacements. A constant thermal increment ∆ θ is
imposed in all points of the rail.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
338 C HAPTER 6. L INEAR E LASTICITY
a) Obtain the displacement, strain and stress fields in terms of the corre-
sponding integration constants.
b) Indicate the boundary conditions that must be applied to determine the
integration constants.
c) Determine the integration constants and obtain the corresponding dis-
placement, strain and stress fields.
d) Particularize these results for the cases K = 0 (open junction) and
K → ∞ (continuous rail).
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Additional hypotheses:
ac
f
C d P cs
Assume the displacements are of the form u = [u (x) , v (y) , w (z) ]T .
b
1)
i
an an n
3) λ =µ
le
liv or ec
.A
6.6 – A solid cylinder with radius R and height h is placed between two in-
m
finitely rigid walls, fitting perfectly between them without producing any stress.
d
uu
er
tin
.O
Additional hypotheses:
1) Material properties: λ = µ and
α = α (r) = α0 + α1 r
2) The friction between the cylinder
and the walls is negligible.
3) Weights can be neglected.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 7
Plane Linear Elasticity
rs
n ee
gi
7.1 Introduction
t d le En
ar
As seen in Chapter 6, from a mathematical point of view, the elastic problem
ar s
ge ro or
eS m
ib
consists in a system of PDEs that must be solved in the three dimensions of
ac
space and in the dimension associated with time R3 × R+ . However, in certain
f
C d P cs
situations, the problem can be simplified so that it is reduced to two dimensions
b
in space in addition to, obviously, the temporal dimension R2 × R+ . This sim-
i
an an n
plification is possible because, in certain cases, the geometry and boundary con-
y ha
a direction of the problem) such that solutions independent of this dimension can
be posed a priori for this elastic problem.
M
.A
relevant direction (assumed constant) coincides with the z-direction. Then, the
d
uu
analysis is reduced to the x-y plane and, hence, the name plane elasticity used to
e
denote such problems. In turn, these are typically divided into two large groups
X Th
er
tin
.O
For the sake of simplicity, the isothermal case will be considered here, even
C
339
340 C HAPTER 7. P LANE L INEAR E LASTICITY
2) The non-zero stresses (that is, those associated with the x-y plane) do not
depend on the z-variable,
σx = σx (x, y,t) , σy = σy (x, y,t) and τxy = τxy (x, y,t) . (7.2)
rs
ee
a) The thickness e is much smaller than the typical dimension associated with
n
the plane of analysis x-y,
gi
eL. (7.3)
t d le En
b) The actions (body forces b (x,t), prescribed displacements u∗ (x,t) and trac-
ar
ar s
ge ro or
tion vector t∗ (x,t) ) are contained within the plane of analysis x-y (its z-
eS m
ib
ac
f
component is null) and, in addition, do not depend on the third dimension,
C d P cs
b
u∗x (x, y,t)
i
bx (x, y,t)
an an n
y ha
not ∗ not
b≡ by (x, y,t) , Γ u : u ≡ u∗ (x, y,t) ,
y
le
liv or ec
0 −
M
.A
(7.4)
tx∗ (x, y,t)
m
not ∗
Γσ = Γσ+ Γσ− Γσe : t∗ ≡
S S
.
uu
X Th
er
−
tin
on
c) The traction vector t∗ (x,t) is only non-zero on the boundary of the piece’s
.O
thickness (boundary Γσe ), whilst on the lateral surfaces Γσ+ and Γσ− it is null
C
0
+ − ∗ not
[
Γσ Γσ : t ≡ 0 . (7.5)
0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Plane Stress State 341
rs
n ee
gi
Figure 7.1: Example of a plane stress state.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 7.1. The piece with the actions defined by (7.4) and (7.5) is
C d P cs
b
compatible with the plane stress state given by (7.1) and (7.2), and
i
an an n
.A
0 σx τxy 0 0 0
m
not not
n≡ 0 , σ · n ≡ τxy σy 0 0 = 0 ,
uu
e
X Th
±1 0 0 0 ±1 0
er
tin
• Edge Γσe
on
.O
C
nx σx τxy 0 nx tx (x, y,t)
©
not not
n ≡ ny , σ (x, y,t)·n ≡ ny = ty (x, y,t)
,
τ xy σy 0
0 0 0 0 0 0
1The fact that all the non-null stresses are contained in the x-y plane is what gives rise to the
name plane stress.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
342 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
ee
Figure 7.2: Plane stress state.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
7.2.1 Strain Field. Constitutive Equation
ac
f
Consider now the linear elastic constitutive equation (6.24),
C d P cs
b
i
1+ν 1
an an n
ν ν
ε =− Tr (σ
σ)1+ σ = − Tr (σ
σ)1+ σ, (7.6)
y ha
E E E 2G
le
liv or ec
which, applied on the stress state in (7.1) and in engineering notation, provides
the strains (6.25)2
M
.A
m
1 1 1
d
E E G
X Th
er
tin
1 1 1 (7.7)
εy = (σy − ν (σx + σz )) = (σy − νσx ) γxz = τxz = 0 ,
E E G
on
.O
1 ν 1
C
where the conditions σz = τxz = τyz = 0 have been taken into account. From (7.2)
and (7.7) it is concluded that the strains do not depend on the z-coordinate either
(εε = ε (x, y,t)). In addition, the strain εz in (7.7) can be solved as
ν
εz = − (εx + εy ) . (7.8)
1−ν
2 The engineering angular strains are defined as γxy = 2 εxy , γxz = 2 εxz and γyz = 2 εyz .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Plane Stress State 343
In short, the strain tensor for the plane stress case results in
1
εx γxy 0
2 ν
ε (x, y,t) ≡ 1 γxy εy 0
not
with εz = − (εx + εy ) (7.9)
2 1−ν
0 0 εz
rs
ee
E E
σx = (εx + νεy ) , σy = (εy + νεx ) ,
n
1 − ν2 1 − ν2
gi
(7.10)
E
t d le En
and τxy = γxy ,
2 (1 + ν)
ar
ar s
ge ro or
eS m
ib
which can be rewritten as
ac
f
C d P cs
1 ν 0
b
σx εx
i
E
an an n
plane
σy = ν 1 0 εy =⇒ σ } = C stress · {εε } .
{σ
y ha
1 − ν2 1−ν
le
τxy 0 0 γxy
liv or ec
| {z } | {z 2 } | {z }
M
.A
{σ
σ} plane {εε }
C stress
(7.11)
m
d
uu
e
X Th
er
.O
1
C
2
can be decomposed into two groups:
1) Those that do not affect the displacement uz (and are hypothetically inte-
grable in R2 for the x-y domain),
∂ ux
εx (x, y,t) =
integration
∂x
in R2
(
∂ uy ux = ux (x, y,t)
εy (x, y,t) = =⇒ . (7.13)
∂y
uy = uy (x, y,t)
∂ ux ∂ uy
γxy (x, y,t) = 2εxy = +
∂y ∂x
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
344 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
Observation of (7.1) to (7.14) suggests considering an ideal elastic plane
ee
stress problem reduced to the two dimensions of the plane of analysis and char-
n
acterized by the unknowns
gi
t d le En
εx σx
u
u (x, y,t) ≡ x , {εε (x, y,t)} ≡ εy and {σ
not not not
σ (x, y,t)} ≡ σy , (7.15)
ar
ar s
uy
ge ro or
eS m
ib
γxy τxy
ac
f
C d P cs
b
in which the additional unknowns with respect to the general problem are either
i
null, or can be calculated in terms of those in (7.15), or do not intervene in the
an an n
y ha
reduced problem,
ν
le
σx = τxz = τyz = γxz = γyz = 0 , εz = − (εx + εy ) ,
liv or ec
1−ν (7.16)
M
.A
d
uu
e
er
tin
.O
solution of the reduced plane stress ux (x, y,t) and uy (x, y,t) will al-
low obtaining a solution uz (x, y, z,t) for the rest of components of
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Plane Strain 345
rs
0 0 0
n ee
In the central section (which is a plane of symmetry with respect to the z-axis)
gi
the conditions
t d le En
∂ ux ∂ uy
ar
uz = 0 , =0 and =0 (7.19)
ar s
ge ro or
∂z ∂z
eS m
ib
ac
f
are satisfied and, thus, the displacement field in this central section is of the form
C d P cs
b
i
an an n
ux (x, y,t)
y ha
not
u (x, y,t) ≡ uy (x, y,t) . (7.20)
le
liv or ec
0
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
346 C HAPTER 7. P LANE L INEAR E LASTICITY
∂ ux ∂ uz
εx (x, y,t) = , εz (x, y,t) = =0,
∂x ∂z
∂ uy ∂ ux ∂ uz
εy (x, y,t) = , γxz (x, y,t) = + =0, (7.21)
∂y ∂z ∂x
∂ ux ∂ uy ∂ uy ∂ uz
rs
γxy (x, y,t) = + , γyz (x, y,t) = + =0.
∂y ∂x ∂z ∂y
n ee
Therefore, the structure of the strain tensor is3
gi
t d le En
1
γxy 0
ar
εx
ar s
2
ge ro or
eS m
ib
ε (x, y,t) ≡ 1 γxy εy 0 .
not
(7.22)
ac
2
f
C d P cs
b
0 0 0
i
an an n
y ha
le
liv or ec
.A
er
tin
.O
Considering (7.21) and (7.24), one concludes that stresses do not depend on the
z-coordinate either (σ
σ = σ (x, y,t)). On the other hand, the stress σz in (7.24)
can be solved as
λ
σz = (σx + σy ) = ν (σx + σy ) (7.25)
2 (λ + µ)
3 By analogy with the plane stress case, the fact that all non-null strains are contained in the
x-y plane gives rise to the name plane strain.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Plane Strain 347
and the stress tensor for the plane strain case results in
σx τxy 0
not
σ (x, y,t) ≡ τxy σy 0 with σz = −ν (σx + σy ) , (7.26)
0 0 σz
rs
ee
E (1 − ν) ν
n
σx = (λ + 2G) εx + λ εy = εx + εy ,
(1 + ν) (1 − 2ν) 1−ν
gi
t d le En
E (1 − ν) ν (7.27)
σy = (λ + 2G) εy + λ εx = εy + εx ,
ar
(1 + ν) (1 − 2ν) 1−ν
ar s
ge ro or
eS m
ib
E
ac
and τxy = G γxy = f
γxy .
C d P cs
2 (1 + ν)
b
i
an an n
le
ν
liv or ec
1 0
σx 1−ν εx
M
.A
E (1 − ν) ν
σy = 1 0 εy ⇒
(1 + ν) (1 − 2ν) 1 − ν
m
1 − 2ν
d
τxy 0 0 γxy
uu
2 (1 − ν)
e
| {z } | {z }
(7.28)
X Th
{σ
σ} {εε }
er
| {z }
tin
plane
C strain
on
.O
C
plane
σ } = C strain · {εε } .
{σ
©
Similarly to the plane stress problem, (7.20), (7.21) and (7.26) suggest con-
sidering an elastic plane strain problem reduced to the two dimensions of the
plane of analysis x-y and characterized by the unknowns
" # εx σx
not ux not not
u (x, y,t) ≡ , {εε (x, y,t)} ≡ εy and {σ
σ (x, y,t)} ≡ σy , (7.29)
uy
γxy τxy
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
348 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
n ee
gi
Figure 7.4: The plane linear elastic problem.
t d le En
ar
ar s
ge ro or
eS m
ib
in which the additional unknowns with respect to the general problem are either
ac
f
null or can be calculated in terms of those in (7.29),
C d P cs
uz = 0 , εz = γxz = γyz = τxz = τyz = 0
b and σz = ν (σx + σy ) . (7.30)
i
an an n
y ha
le
liv or ec
.A
In view of the equations in Sections 7.2 and 7.3, the linear elastic problem for
m
the plane stress and plane strain problems is characterized as follows (see Fig-
uu
e
ure 7.4).
X Th
er
tin
Equations 4
on
.O
a) Cauchy’s equation
C
∂ σx ∂ τxy ∂ 2 ux
+ + ρbx = ρ 2
∂x ∂y ∂t (7.31)
∂ τxy ∂ σy ∂ 2 uy
+ + ρby = ρ 2
∂x ∂y ∂t
4 The equation corresponding to the z-component either does not intervene (plane stress), or
is identically null (plane strain).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
The Plane Linear Elastic Problem 349
b) Constitutive equation
σx εx
not not
{σ
σ } ≡ σy , {εε } ≡ εy ; {σ
σ } = C · {εε } , (7.32)
τxy γxy
where the constitutive matrix C can be written in a general form, from (7.11)
and (7.28), as
(
rs
Ē = E
Plane stress
ee
1 ν̄ 0 ν̄ = ν
n
not Ē
gi
C≡ ν̄ 1 0 (7.33)
E
1 − ν̄ 2
Ē =
t d le En
1 − ν̄ 1−ν 2
0 0 Plane strain
2
ar
ν̄ = ν
ar s
ge ro or
eS m
ib
1−ν
ac
f
C d P cs
c) Geometric equation
b
i
an an n
∂ ux ∂ uy ∂ ux ∂ uy
y ha
εx = , εy = , γxy = + (7.34)
le
∂x ∂y ∂y ∂x
liv or ec
M
.A
d
uu
e
" # " #
u∗x (x, y,t) tx∗ (x, y,t)
X Th
er
tin
not not
Γu : u∗ ≡ , Γσ : t∗ ≡
u∗y (x, y,t) ty∗ (x, y,t)
on
.O
(7.35)
C
" # " #
nx
©
e) Initial conditions
.
u (x, y,t) =0, u (x, y,t) = v0 (x, y) (7.36)
t=0 t=0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
350 C HAPTER 7. P LANE L INEAR E LASTICITY
Unknowns
" # 1 " #
not ux not
εx γxy σ τ
x xy
u (x, y,t) ≡ , ε (x, y,t) ≡ 1 2 , σ (x, y,t) not
≡ (7.37)
uy γxy εy τ xy σy
2
rs
ee
ν
Plane stress → εz = (εx + εy )
n
1−ν (7.38)
gi
Plane strain → σz = ν (σx + σy )
t d le En
ar
ar s
ge ro or
eS m
ib
7.5 Problems Typically Assimilated to Plane Elasticity
ac
f
C d P cs
7.5.1 Plane Stress
b
i
an an n
The stress and strain states produced in solids that have a dimension consider-
y ha
ably inferior to the other two (which constitute the plane of analysis x-y) and
le
liv or ec
whose actions are contained in said plane are typically assimilated to a plane
M
stress state. The slab loaded on its mean plane and the deep beam of Figure 7.5
.A
are classic examples of structures that can be analyzed as being in a plane stress
m
state. As a particular case, the problems of simple and complex bending in beams
d
uu
lems.
X Th
er
tin
on
.O
C
Figure 7.5: Slab loaded on its mean plane (left) and deep beam (right).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems Typically Assimilated to Plane Elasticity 351
rs
thus, satisfies the conditions
ee
∂ ux ∂ uy
uz = 0 , = 0 and =0, (7.39)
n
∂z ∂z
gi
which result in the initial condition of the plane strain state (7.17),
t d le En
ux ux (x, y,t)
ar
ar s
ge ro or
eS m
ib
not
u ≡ uy = uy (x, y,t) . (7.40)
ac
uz f 0
C d P cs
b
i
Examples of this case are a pipe under internal (and/or external) pressure
an an n
y ha
(see Figure 7.6), a tunnel (see Figure 7.7) and a strip foundation (see Fig-
ure 7.8).
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
2) The length of the piece in the longitudinal direction is reduced, but the dis-
placements in the z-direction are impeded by the boundary conditions at the
end sections (see Figure 7.9).
In this case, the plane strain hypothesis (7.17) can be assumed for all the
transversal sections of the piece.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
352 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
ee
Figure 7.7: Tunnel.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Representative Curves of Plane Elasticity 353
rs
Definition 7.1. The isostatics or stress trajectories are the envelopes
ee
of the vector field determined by the principal stresses.
n
gi
t d le En
Considering the definition of the envelope of a vector field, isostatics are, at each
ar
point, tangent to the two principal directions and, thus, there exist two families
ar s
ge ro or
eS m
ib
of isostatics:
ac
f
− Isostatics σ1 , tangent to the direction of the largest principal stress.
C d P cs
b
− Isostatics σ2 , tangent to the direction of the smallest principal stress.
i
an an n
In addition, since the principal stress directions are orthogonal to each other,
y ha
both families of curves are also be orthogonal. The isostatic lines provide infor-
le
mation on the mode in which the flux of principal stresses occurs on the plane
liv or ec
of analysis.
M
.A
d
uu
e
X Th
er
state
on
.O
σx = σy and τxy = 0
C
and its Mohr’s circle is a point on the axis σ (see Figure 7.11).
©
σx = σy = τxy = 0
and its Mohr’s circle is the origin of the σ − τ space (see Fig-
ure 7.11).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
354 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Figure 7.11: Singular and neutral points.
ac
f
C d P cs
b
i
an an n
y ha
Remark 7.3. All directions in a singular point are principal stress di-
le
liv or ec
rections (the pole is the Mohr’s circle itself, see Figure 7.11). Conse-
quently, the isostatics tend to loose their regularity in singular points
M
.A
d
uu
e
X Th
er
tin
.O
Consider the general equation of an isostatic line y = f (x) and the value of the
angle formed by the principal stress direction σ1 with respect to the horizontal
C
2τxy 2 tan α
tan (2α) = =
0
σx − σy 1 − tan2 α ⇒ 2τxy = 2y
⇒
dy not 0
σx − σy 1 − (y0 )2
tan α = =y
(7.41)
dx
σx − σy 0
(y0 )2 + y −1 = 0
τxy
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Representative Curves of Plane Elasticity 355
rs
ee
and solving the second-order equation (7.41) for y0 , the differential equation of
n
the isostatics is obtained.
gi
t d le En
s
σx − σy 2
ar
Differential −
ar s
σx σ y
ge ro or
0
→ y =− ± +1
eS m
ib
equation (7.42)
2τxy 2τxy
ac
of the isostatics
f
|
C d P cs {z }
b
ϕ (x, y)
i
an an n
y ha
.A
y = f (x) +C . (7.43)
m
er
tin
.O
Solution
The singular points are defined by σx = σy and τxy = 0 . Then,
(
σx = −x3 = 0
x = 0 =⇒ ∀y
σy = 2x3 − 3xy2 = 0
2
τxy = 3x y = 0 =⇒ (
σx = −x3
y = 0 =⇒ =⇒ x = 0
σy = 2x3 − 3xy2 = 2x3
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
356 C HAPTER 7. P LANE L INEAR E LASTICITY
Therefore, the locus of singular points is the straight line x = 0. These singu-
lar points are, in addition, neutral points (σx = σy = 0).
The isostatics are obtained from (7.42),
s 2
0 dy σx − σy σx − σy
y = =− ± +1 ,
dx 2τxy 2τxy
rs
dy x
= x2 − y2 = C1
ee
dx y =⇒ integrating =⇒ .
n
dy −y xy = C2
gi
=
dx x
t d le En
Therefore, the isostatics are two families of equilateral hyperboles orthogo-
ar
ar s
ge ro or
nal to each other.
eS m
ib
ac
f
On the line of singular points x = 0 (which divides the plate in two regions)
C d P cs
the isostatics will brusquely change their slope. To identify the family of
b
i
isostatics σ1 , consider a point in each region:
an an n
y ha
.A
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Representative Curves of Plane Elasticity 357
7.6.2 Isoclines
Definition 7.3. Isoclines are the locus of the points in the plane of
analysis along which the principal stress directions form a certain
angle with the x-axis.
It follows from its definition that in all the points of a same isocline the principal
rs
stress directions are parallel to each other, forming a constant angle θ (which
ee
characterizes the isocline) with the x-axis (see Figure 7.13).
n
gi
7.6.2.1 Equation of the Isoclines
t d le En
The equation y = f (x) of the isocline with an angle θ is obtained by establishing
that the principal stress direction σ1 forms an angle α = θ with the horizontal
ar
ar s
ge ro or
direction, that is,
eS m
ib
ac
f
C d P cs
Algebraic equation b 2τxy
i
tan (2θ ) =
an an n
| {z }
le
ϕ (x, y)
liv or ec
M
.A
y = f (x, θ ) , (7.45)
uu
e
X Th
er
tin
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
358 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
ee
7.6.3 Isobars
n
gi
t d le En
Definition 7.4. Isobars are the locus of points in the plane of analy-
ar
ar s
ge ro or
sis with the same value of principal stress σ1 (or σ2 ).
eS m
ib
ac
f
C d P cs
b
i
Two families of isobars will cross at each point of the plane of analysis: one
an an n
y ha
corresponding to σ1 and another to σ2 . Note that the isobars depend on the value
le
of σ1 , but not on its direction (see Figure 7.14).
liv or ec
M
.A
The equation that provides the value of the principal stresses (see Chapter 4) im-
d
plicitly defines the algebraic equation of the two families of isobars y = f1 (x, c1 )
uu
e
and y = f2 (x, c2 ),
X Th
er
tin
s
on
.O
2
σ x + σ y σx − σy 2 = const. = c
σ1 = + + τxy
1
C
2 2
©
Algebraic
| {z }
equation
ϕ1 (x, y)
(7.46)
of the s
σx − σy 2
σx + σy
isobars 2 = const. = c
σ2 = − + τxy 2
2 2
| {z }
ϕ2 (x, y)
which leads to (
y = f1 (x, c1 )
(7.47)
y = f2 (x, c2 )
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Representative Curves of Plane Elasticity 359
rs
n ee
gi
7.6.4 Maximum Shear Stress or Slip Lines
t d le En
ar
ar s
ge ro or
eS m
ib
Definition 7.5. Maximum shear stress lines or slip lines are the en-
ac
f
velopes of the directions that, at each point, correspond with the
C d P cs
b
maximum value (in modulus) of the shear (or tangent) stress.
i
an an n
y ha
le
liv or ec
M
.A
Remark 7.5. At each point of the plane of analysis there are two
planes on which the shear stresses reach the same maximum value
m
(in module) but that have opposite directions, τmax and τmin . These
uu
e
a 45◦ angle with the principal stress directions (see Figure 7.15).
er
tin
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
360 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
ee
Figure 7.15: Maximum shear stress planes.
n
gi
t d le En
where α is the angle formed by the principal stress direction σ1 with the
ar
ar s
horizontal direction. Consequently, considering the general equation of a slip
ge ro or
eS m
ib
line, y = f (x), the expression (7.48) and the relation tan (2α) = 2τxy / (σx − σy )
ac
yields f
C d P cs
1 σx − σy
b2 tan β
i
tan (2β ) = − = =
an an n
tan (2α) 2τxy 1 − tan2 β =⇒
y ha
dy not 0
le
tan β = =y
liv or ec
dx (7.49)
M
.A
σx − σy 2y0 4τxy
m
− = =⇒ (y0 )2 − y0 − 1 = 0.
d
2τxy 1 − (y0 )2 σx − σy
uu
e
X Th
er
Solving the second-order equation in (7.49) for y0 provides the differential equa-
tin
.O
C
Differential s 2
2τ 2τxy
©
equation of the 0 xy
max. shear stress y =− ± +1 (7.50)
σx − σy σx − σy
or slip lines | {z }
ϕ (x, y)
If the function ϕ (x, y) in (7.50) is known, this differential equation can be in-
tegrated and the algebraic equation of the two families of orthogonal curves
(corresponding to the double sign in (7.50)) is obtained.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Representative Curves of Plane Elasticity 361
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
362 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 363
P ROBLEMS
Problem 7.1 – Justify whether the following statements are true or false.
a) If a plane stress state has a singular point, all the isoclines cross this
point.
b) If a plane stress state is uniform, all the slip lines are parallel to each
rs
other.
n ee
gi
Solution
t d le En
a) A singular point is defined as:
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
The stress state is
b represented by a point.
σ1 = σ2
i
an an n
τ =0
y ha
le
liv or ec
M
.A
Therefore, all directions are principal stress directions and, given an angle θ
m
which can take any value, the principal stress direction will form an angle θ
d
with the x-axis. Then, an isocline of angle θ will cross said point and, since this
uu
e
holds true for any value of θ , all the isoclines will cross this point. Therefore,
X Th
er
tin
.O
b) A uniform stress state implies that the Mohr’s circle is equal in all points of
the medium, therefore, the planes of maximum shear stress will be the same in
C
all points. Then, the maximum shear stress lines (or slip lines) will be parallel to
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
364 C HAPTER 7. P LANE L INEAR E LASTICITY
1) σx = 0 ; σy = b > 0 ; τxy = 0
2) σx = 0 ; σy = 0 ; τxy = m y , m > 0
Plot for each state the isostatics and the slip lines, and indicate the singular
points.
rs
n ee
gi
Solution
t d le En
1) The Mohr’s circle for the stress state σx = 0 ; σy = b > 0 ; τxy = 0 is:
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 365
rs
There do not exist singular points for this stress state.
n ee
gi
t d le En
2) The Mohr’s circle for the stress state σx = 0 ; σy = 0 ; τxy = m y , m > 0 is:
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
366 C HAPTER 7. P LANE L INEAR E LASTICITY
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 367
E XERCISES
σx = σy
τxy = ax
σy = b
rs
(a > 0 , b > 0)
ee
Plot the isostatics and the slip lines, and indicate the singular points.
n
gi
t d le En
7.2 – Plot the isostatics in the transversal section of the cylindrical shell shown
ar
ar s
below. Assume a field of the form:
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
B
y ha
ur = Ar + r ; A > 0, B > 0
le
liv or ec
uθ = 0
M
.A
uz = 0
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 8
Plasticity
rs
n ee
gi
8.1 Introduction
t d le En
ar
The elastoplastic models (constitutive equations) are used in continuum mechan-
ar s
ge ro or
eS m
ib
ics to represent the mechanical behavior of materials whose behavior, once cer-
ac
f
tain limits in the values of the stresses (or strains) are exceeded, is no longer rep-
C d P cs
resentable by means of simpler models such as the elastic ones. In this chapter,
b
i
these models will be studied considering, in all cases, that strains are infinitesi-
an an n
mal.
y ha
.A
during the loading process is not recovered during the unloading process.
uu
e
X Th
er
tin
.O
The concepts in this section are a review of those already studied in Sec-
C
τxz τyz σz
369
370 C HAPTER 8. P LASTICITY
diagonalization
rs
n ee
Figure 8.1: Diagonalization of the stress tensor.
gi
t d le En
ar
Since σ is a symmetrical second-order tensor, it will diagonalize in an orthonor-
ar s
ge ro or
eS m
ib
mal base and all its eigenvalues will be real numbers. Then, consider a system
of Cartesian axes {x0 , y0 , z0 } associated with a base in which σ diagonalizes. Its
ac
f
C d P cs
matrix of components in this base is
b
i
an an n
σ1 0 0
y ha
le
σ ]x0 y0 z0 = 0 σ2 0 ,
[σ (8.2)
liv or ec
0 0 σ3
M
.A
the directions associated with the axes {x0 , y0 , z0 } are named principal directions
uu
e
er
tin
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Previous Notions 371
λ 3 − I1 λ 2 − I2 λ − I3 = 0 , (8.6)
where the coefficients I1 (σi j ), I2 (σi j ) and I3 (σi j ) are certain functions of the
components σi j of the tensor σ expressed in the coordinate system {x, y, z}. Yet,
the solutions to (8.6), which will be a function of its coefficients (I1 , I2 , I3 ), are
the principal stresses that, on the other hand, are independent of the system of
axes chosen to express σ . Consequently, said coefficients must be invariant with
rs
respect to any change of base. Therefore, the coefficients I1 , I2 and I3 are denoted
ee
as I stress invariants or fundamental stress invariants and their expression (re-
n
sulting from the computation of (8.5)) is
gi
t d le En
I = Tr (σσ ) = σii = σ1 + σ2 + σ3
1
ar
I stress 1
ar s
ge ro or
σ : σ − I12 = − (σ1 σ2 + σ1 σ3 + σ2 σ3 )
I2 = (8.7)
eS m
ib
invariants 2
ac
I3 = det (σ
σ ) = σ1 σ2 σ3 f
C d P cs
b
i
an an n
y ha
Obviously, any scalar function of the stress invariants will also be an invariant
le
liv or ec
and, thus, new invariants can be defined based on the I stress invariants given
in (8.7). In particular, the so-called J stress invariants are defined as
M
.A
m
d
uu
J1 = I1 = σii = Tr (σ σ)
e
X Th
er
tin
1 2 1
J2 = I1 + 2I2 = σi j σ ji =
on
.O
2 2
J stress 1 1 (8.8)
C
invariants = (σ σ : σ ) = Tr (σ σ ·σ)
2 2
©
1 3 1
J3 = I1 + 3I1 I2 + 3I3 = σi j σ jk σki =
3 3
1
= Tr (σ σ ·σ ·σ)
3
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
372 C HAPTER 8. P LASTICITY
rs
ee
8.2.2 Spherical and Deviatoric Components of the Stress Tensor
n
Given the stress tensor σ , the mean stress σm is defined as
gi
t d le En
I1 1 1 1
σm = = Tr (σ
σ ) = σii = (σ1 + σ2 + σ3 ) (8.9)
ar
ar s
3 3 3 3
ge ro or
eS m
ib
ac
and the mean pressure p̄ as f
C d P cs
b
p̄ = −σm . (8.10)
i
an an n
y ha
The Cauchy stress tensor can be decomposed into a spherical part (or com-
le
ponent), σ sph , and a deviatoric one, σ 0 ,
liv or ec
M
.A
σ = σ sph + σ 0 , (8.11)
m
d
uu
er
tin
de f 1
on
σ sph : = Tr (σ σ ) 1 = σm 1
.O
3
C
σm 0 0 (8.12)
©
not
σ sph ≡ 0 σm 0
0 0 σm
σx − σm τxy τxz
not
σ 0 = σ − σ sph ≡ σy − σm . (8.13)
τxy τyz
τxz τyz σz − σm
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Previous Notions 373
Finally, the I and J invariants of the deviatoric tensor σ 0 , named I 0 and J 0 invari-
ants, respectively, are derived from (8.7), (8.8), (8.9) and (8.13).
J10 = I10 = 0
J 0 stress
1 0 1 (8.14)
invariants J20 = I20 = (σ σ : σ 0 ) = σi0j σ 0ji
2 2
J0 = I0 = 1 σ 0 σ 0 σ 0
3 3
3 i j jk ki
rs
n ee
gi
Remark 8.2. It is easily proven that the principal directions of σ co-
t d le En
incide with those of σ 0 , that is, that both tensors diagonalize in the
ar
same base. In effect, working in the base associated with the princi-
ar s
ge ro or
eS m
ib
pal directions of σ , i.e., the base in which σ diagonalizes, and, given
ac
f
that σ sph is a hydrostatic tensor and, thus, is diagonal in any base,
C d P cs
then σ 0 also diagonalizes in the same base (see Figure 8.2).
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Figure 8.2: Diagonalization of the spherical and deviatoric parts of the stress tensor.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
374 C HAPTER 8. P LASTICITY
rs
n ee
Example 8.1 – Compute the value of the equivalent uniaxial stress (or effec-
gi
tive stress) σ̄ for an uniaxial stress state defined by
t d le En
ar
σu 0 0
ar s
ge ro or
eS m
ib
not
σ ≡ 0 0 0 .
ac
f
C d P cs
0 0 0
b
i
an an n
y ha
Solution
le
liv or ec
.A
1 σ
σ) = u .
Tr (σ σm =
3 3
m
er
σu
tin
0 0
σm 0 0 3
on
.O
not σu
σ sph ≡ 0 σm 0 = 0 3 0
C
0 0 σm
σu
©
0 0
3
and the deviatoric component results in
2
σu 0 0
σu − σm 0 0 3
0 not 1
σ = σ − σ sph ≡ 0 −σm 0 = 0 − σu 0 .
3
0 0 −σm
1
0 0 − σu
3
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Principal Stress Space 375
σ̄ = |σu | .
rs
8.3 Principal Stress Space
n ee
Consider a system of Cartesian axes in R3 {x ≡ σ1 , y ≡ σ2 , z ≡ σ3 } such that
gi
each stress state, characterized by the values of the three principal stresses
t d le En
σ1 ≥ σ2 ≥ σ3 , corresponds to a point in this space, which is known as the prin-
cipal stress space1 (see Figure 8.3).
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Definition 8.1. The hydrostatic stress axis is the locus of points in
b
the principal stress space that verify the condition σ1 = σ2 = σ3 (see
i
an an n
Figure 8.3). The points located on the hydrostatic stress axis repre-
y ha
.A
m
d
uu
e
X Th
er
tin
.O
(σ1 = σ2 = σ3 )
= bisector of the 1st octant
C
1 The principal stress space is also known as the Haigh-Westergaard stress space.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
376 C HAPTER 8. P LASTICITY
rs
ee
Figure 8.4: The hydrostatic stress axis and the octahedral plane.
n
gi
t d le En
ar
Definition 8.2. The octahedral plane Π is any of the planes that
ar s
ge ro or
eS m
ib
are perpendicular to the hydrostatic stress axis (see Figure 8.4). The
ac
equation of an octahedral plane is f
C d P cs
b
σ1 + σ2 + σ3 = const.
i
an an n
not 1
n ≡ √ [1, 1, 1]T .
le
liv or ec
3
M
.A
m
d
uu
e
er
tin
.O
ure 8.5). Now, the octahedral plane Π containing point P is considered. The
C
intersection of the hydrostatic stress axis with said plane defines point A.
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Principal Stress Space 377
rs
n ee
gi
t d le En
Figure 8.5: Definitions of the normal and shear octahedral stresses.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
Remark 8.4. The normal octahedral stress σoct informs of the dis-
i
an an n
tance between the origin O of the principal stress space and the oc-
y ha
tahedral plane that contains point P. The locus of points in the prin-
le
cipal stress space with
√ the same value of σoct is the octahedral plane
liv or ec
.A
er
space with the same value of τoct is a cylinder whose axis is the hy-
tin
√
drostatic stress axis and whose radius is 3 τoct .
on
.O
C
The distance OA can be computed as the projection of the vector OP on the
unit normal vector of the octahedral plane, n,
√
1/ 3 √
√ √
not
OA = OP · n ≡ σ1 , σ2 , σ3 1/ 3 = 3
(σ1 + σ2 + σ3 ) = 3σm
√ 3 ⇒
1/ 3
√
OA = 3 σoct
(8.15)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
378 C HAPTER 8. P LASTICITY
I1
σoct = σm = (8.16)
3
rs
of the second invariant of the deviatoric stress tensor in (8.14), J20 , as
ee
2 √
AP = 2 J 0 =⇒ AP = 2 (J 0 )1/2
n
2 2
gi
√ =⇒ (8.18)
t d le En
AP = 3 τ oct
ar
ar s
ge ro or
r
eS m
ib
2 0 1/2
τoct = (J ) (8.19)
ac
f3 2
C d P cs
b
i
Alternative expressions of τoct in terms of the value of J20 in (8.14) are
an an n
y ha
1/2
1 1
le
2 2 2 2
liv or ec
.A
1 1/2
τoct = √ (σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ1 − σ3 )2 .
m
3 3
d
uu
e
X Th
er
tin
.O
σ = σ sph = σm 1 ⇔ σ 0 = σ −σ
σ sph = 0 ⇔ J20 = 0 ⇔ τoct = 0 .
C
σ = σ 0 ⇔ σm = Tr (σ
σ ) = Tr σ 0 = 0 ⇔
σoct = 0 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Principal Stress Space 379
rs
√ deviatoric stress invariant J2 characterizes the dis-
tance (= 3 τoct ) from the hydrostatic stress axis to the point.
ee
Thus, it places point P on a certain circle in the octahedral
n
plane with center in the hydrostatic stress axis and radius
gi
√ √
3 τoct = 2 (J20 )1/2 .
t d le En
− The third deviatoric stress invariant J30 characterizes the position
ar
ar s
ge ro or
of the point on this circle by means of an angle θ (J30 ).
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
er
tin
on
.O
C
Figure 8.6: Univocal definition of a point by means of the invariants I1 , J20 and J30 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
380 C HAPTER 8. P LASTICITY
Remark 8.7. Figure 8.7 shows the projection of the principal stress
space on an octahedral plane Π . The division of the stress space
into six sectors can be observed in this projection. Each sector is
characterized by a different ordering of the principal stresses and the
sectors are separated by the projections on the plane of the bisectors
σ2 = σ3 , σ1 = σ3 and σ1 = σ2 .
Selecting the criterion σ1 ≥ σ2 ≥ σ3 automatically reduces the fea-
sible work domain to the sector marked in gray in the figure. The
intersection of any surface of the type f (σ1 , σ2 , σ3 ) = 0 with the
rs
plane Π is reduced to a curve in said sector.
ee
This curve can be automatically extended to the rest of sectors, that
n
is, the curve obtained with the same function f (σ1 , σ2 , σ3 ) = 0 but
gi
considering the different orderings of the principal stresses can be
t d le En
easily plotted, by considering the symmetry conditions with respect
to the bisector planes. The resulting curve presnts, thus, three axes
ar
ar s
ge ro or
of symmetry with respect to each of the axis in Figure 8.7.
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Rheological Models 381
rs
ure 8.8). The model establishes a proportionality between stress and strain, both
ee
in loading and unloading, being the constant E the proportionality factor (see
Figure 8.8).
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
er
tin
.O
Consider a solid block placed on a rough surface (see Figure 8.9) and subjected
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
382 C HAPTER 8. P LASTICITY
rs
this device (analogous to the displacement δ ). This rheological model includes
ee
a frictional device characterized by a limit value σe (analogous to the role of µ N
n
in the Coulomb model) whose value cannot be exceeded.
gi
Figure 8.11 shows the stress-strain curve corresponding to the frictional rhe-
t d le En
ological model for a loading-unloading-reloading cycle, which can be split into
the following sections.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
er
tin
on
.O
|σ | < σe → ∆ ε = 0
|σ | = σe → ∆ ε 6= 0
C
|σ | > σe → impossible
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Rheological Models 383
rs
Section 2 − 3: At point 2, the tendency
ee
of the stress is inverted, stress starts de-
n
creasing (∆ σ < 0) and unloading be- Figure 8.11: Stress-strain curve
gi
gins (σ < σe ). Further strain increase is for a loading-unloading-reloading
t d le En
automatically halted (∆ ε = 0). This sit- cycle in a frictional rheological
uation is maintained until stress is can- model.
ar
ar s
ge ro or
celed (σ = 0) at point 3. Note that, if
eS m
ib
the process was to be halted at this point, the initial state of null stress would be
ac
f
recovered but not the initial state of null strain. Instead, a permanent or residual
C d P cs
b
strain would be observed (ε 6= 0). This reveals that, in this model, the trajec-
i
an an n
tory of the stress-strain curve is different in the loading and unloading regimes
y ha
and that the deformation process is (from a thermodynamic point of view) irre-
le
versible in character.
liv or ec
M
.A
Section 3 − 4: Beyond point 3, the sign of the stress is inverted and stress be-
comes compressive. However, since |σ | < σe , no changes in strain are observed
m
(∆ ε = 0).
d
uu
e
Section 4−5: At point 4, the criterion |σ | = σe is satisfied and the model enters a
X Th
er
tin
loading regime again. The element flows at a constant stress value σ = −σe , gen-
erating negative strain (∆ ε < 0), which progressively reduces the accumulated
on
.O
strain. Finally, at point 5, the initial strain state is recovered, but not the original
C
stress state. Beyond this point, if unloading was imposed, there would be a cor-
©
responding decrease in stress until the cycle was closed at point 0. Conversely,
the loading regime could continue, generating a permanent negative strain.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
384 C HAPTER 8. P LASTICITY
stress will act on both of them. On the other hand, the total strain can be decom-
posed into the sum of the strain experienced by the elastic element (ε e ) plus the
strain experienced by the frictional device ε f . The same logic can be applied
at incremental level.
σ = σe = σ f
σ
ε = εe + ε f = + ε f Additive decomposition (8.22)
E
∆ ε = ∆ εe + ∆ ε f of strain
rs
n ee
gi
Frictional element
Elastic element
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
m
Taking into account the stress-strain behavior of each basic element that com-
d
uu
er
tin
(
f e ∆ ε = ∆ εe
• |σ | < σe =⇒ ∆ ε = 0 =⇒ ∆ ε = ∆ ε =⇒
on
.O
∆ σ = E∆ ε
C
The frictional element does not deform for stresses |σ | < σe , therefore all
strains are absorbed by the elastic element.
σ
• |σ | = σe =⇒ ∆ ε f 6= 0 =⇒ ε = + ε f =⇒
E
(
|σ | = σe
∆ ε = ∆ ε f =⇒ ∆ ε e = 0 =⇒ ∆ σ = 0
All strain increments are absorbed by the frictional element with a null in-
crement of stress.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Rheological Models 385
• |σ | > σe
rs
This section corresponds to the elastic loading phase. At the end of the loading,
ee
at point 1, the strain is ε = ε e = σe /E. The value of σe at the end of this elastic
section justifies its denomination as elastic limit.
n
gi
Section 1 − 2:
t d le En
( σe
ε= +ε f
ar
f
ar s
|σ | = σe =⇒ ∆ ε 6= 0 =⇒ E
ge ro or
eS m
ib
∆ε = ∆ε f > 0
ac
f
C d P cs
b
This section corresponds to the frictional loading during which no deformation
i
is generated in the elastic element (no elastic strain is generated) and all incre-
an an n
y ha
|σ | < σe =⇒ ∆ ε f = 0 =⇒ ∆ ε = ∆ ε e
M
.A
This section corresponds to the elastic unloading. At the end of the unloading,
m
at point 3, the initial state of null stress is recovered (σ = 0). Consequently, the
d
uu
strain is ε = ε f 6= 0. That is, the strain generated by the frictional element during
er
tin
the frictional loading section 1 − 2 is not recovered during this phase of stress
relaxation to zero. This allows qualifying the frictional component of strain ε f
on
.O
Section 3 − 4:
©
|σ | < σe =⇒ ∆ ε f = 0 =⇒ ∆ ε = ∆ ε e
This section corresponds to the elastic reloading phase, similar to section 0 − 1
but with a compressive stress (σ < 0). The frictional component of strain is not
modified during the reloading and the final value, at point 4, of the elastic strain
is ε e = −σe /E.
Section 4 − 5:
ε = − σe + ε f
f
|σ | = σe =⇒ ∆ ε =6 0 =⇒ E
∆ε = ∆ε f < 0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
386 C HAPTER 8. P LASTICITY
rs
initial state 0.
ee
Figure 8.13: Stress-strain curve for
n
a loading-unloading-reloading cycle in
gi
an elastic-frictional rheological model.
t d le En
ar
8.4.4 Frictional Model with Hardening
ar s
ge ro or
eS m
ib
Consider the rheological model in Figure 8.14 composed of an elastic element
ac
f
(characterized by the parameter H 0 , which will be denoted as hardening mod-
C d P cs
b
ulus) and a frictional element (characterized by the elastic limit σe ) placed in
i
an an n
the same strain, while the total stress in the model is the sum of the stress in the
le
frictional element (σ (1) ) plus the stress in the elastic element (σ (2) ).
liv or ec
M
.A
(
σ = σ (1) + σ (2)
m
= εe =εf
X Th
ε
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Rheological Models 387
rs
ee
b) Elastic element
n
gi
(
σ (2) = H 0 ε e = H 0 ε
t d le En
(8.25)
∆ σ (2) = H 0 ∆ ε e = H 0 ∆ ε
ar
ar s
ge ro or
eS m
ib
c) Combining (8.24) and (8.25) leads to
ac
f
C d P cs
b
= σ − H 0 ε
(1) (2)
σ = σ − σ (8.26)
i
an an n
y ha
In agreement with (8.24) and (8.25), the following situations can be estab-
le
lished regarding the rheological model:
liv or ec
M
.A
(
(1) 0
∆ε f = ∆ε = 0
• σ < σe ⇐⇒ σ − H ε < σe =⇒
∆ σ (2) = H 0 ∆ ε e = H 0 ∆ ε = 0
m
d
uu
(
e
∆ σ = ∆ σ (1)
X Th
=⇒
er
tin
∆ε = 0
on
.O
All the stress is absorbed by the frictional device and strain is null.
C
©
σ (1) = σe
• σ (1) = σe ⇐⇒ σ − H 0 ε = σe =⇒
σ (2) = σ − σ (1)
=⇒ ∆ σ (2) = ∆ σ = H 0 ∆ ε
All stress increments are totally absorbed by the elastic element.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
388 C HAPTER 8. P LASTICITY
Section 0 − 1:
(
(1) ∆ σ (2) = E∆ ε = 0
σ < σe =⇒ ∆ ε = 0 =⇒
∆ σ (1) = ∆ σ
In this section all the stress is absorbed by the frictional element. At the end of
the section, at point 1, the strain is ε = 0 and the stress is σ = σe . This section
is characterized by the condition
σ − H 0 ε < σe .
rs
ee
Section 1 − 2:
n
gi
(
(1) σ = σe + σ (2)
σ = σe =⇒
t d le En
∆ σ = ∆ σ (2) = H 0 ∆ ε
ar
ar s
ge ro or
eS m
ib
This is a loading section in which all stress is absorbed by the elastic element.
ac
f
In global terms, the model increases its capacity to resist stress (the model is
C d P cs
said to suffer hardening) proportionally to the increment of strain, being the
b
proportionality factor the hardening modulus H 0 . This section is characterized
i
an an n
by the condition
y ha
σ − H 0 ε = σe .
le
liv or ec
M
Section 2 − 3:
.A
m
(
∆ σ (1) = ∆ σ
d
(1)
σ < σe =⇒ ∆ ε = 0 =⇒
uu
∆ σ (2) = 0
e
X Th
er
tin
In this section the stress in the frictional element decreases with a null incre-
on
.O
ment of strain and keeping the stress constant in the elastic element. This state
is maintained until stress is totally inverted in the frictional element. Thus, at
C
σ − H 0 ε < σe .
Section 3 − 4:
(
(1)
σ = −σe + σ (2)
σ = σe =⇒
|{z} ∆ σ = ∆ σ (2) = H 0 ∆ ε
−σ e
The situation is symmetrical with respect to section 1 − 2, with the elastic ele-
ment decreasing the stress it can bear, until the stress becomes null at point 3,
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Rheological Models 389
where σ (1) = −σe and σ (2) = 0. This section is characterized by the condition
σ − H 0 ε = σe .
Beyond this point, relaxation of the stress in the frictional element leads to the
original state at point 0.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
d
uu
e
X Th
Combining now an elastic element, with elastic modulus E, in series with the
on
.O
(
ε = εe + ε f
→ Additive decomposition
∆ε = ∆ εe + ∆ ε f of strain
( (8.27)
σ = σe = σ f
∆σ = ∆σe = ∆σ f
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
390 C HAPTER 8. P LASTICITY
rs
ee
Figure 8.16: Elastic-frictional model with hardening.
n
gi
t d le En
where σ e and σ f represent, respectively, the stresses sustained by the elastic
ar
ar s
ge ro or
element and the frictional model with hardening. Combining now the behav-
eS m
ib
ior of an elastic element (see Figure 8.8) with that of the frictional model with
ac
f
hardening in Figure 8.14, yields the following situations:
C d P cs
b
i
(
an an n
0
∆ε f = 0
• σ − H ε f < σe =⇒
y ha
=⇒ ∆ σ = E∆ ε
∆ σ = ∆ εe
le
liv or ec
The frictional element with hardening does not deform and the increment of
M
.A
as elastic process.
d
uu
e
• σ − H 0 ε f = σe
X Th
er
tin
σ > 0 and ∆ σ > 0
on
(
.O
∆ σ = ∆ σ f = H 0∆ ε f
a) σ ∆ σ > 0 ⇐⇒ or =⇒
C
∆ σ = ∆ σ e = E∆ ε e
σ < 0 and ∆ σ < 0
©
1 1 E + H0
=⇒ ∆ ε = ∆ ε e + ∆ ε f = ∆σ + 0 ∆σ = ∆σ
E H EH 0
∆ σ = Ee f ∆ ε
=⇒ 0
Ee f = E H
E + H0
The strain increment is absorbed by the two elements of the model (the
frictional one with hardening and the elastic one). The relation between
the stress increment ∆ σ and the strain increment ∆ ε is given by the
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Rheological Models 391
rs
ee
Figure 8.17 shows the stress-strain curve corresponding to the model for a
loading-unloading-reloading cycle, in which the following sections can be dif-
n
ferentiated.
gi
t d le En
Section 0 − 1 and section 2 − 3:
ar
σ − H 0 ε f < σe =⇒ ∆ σ = E∆ ε
ar s
ge ro or
eS m
ib
ac
Correspond to elastic processes. f
C d P cs
b
i
Section 1 − 2 and section 3 − 4:
an an n
y ha
(
σ − H 0 ε f = σe
le
=⇒ ∆ σ = E e f ∆ ε
liv or ec
σ ∆σ > 0
M
.A
d
uu
Point 2:
e
(
σ − H 0 ε f = σe
X Th
er
tin
=⇒ ∆ σ = E∆ ε
σ ∆σ < 0
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
392 C HAPTER 8. P LASTICITY
Elastic loading
Elastic
rs
n ee
Figure 8.17: Stress-strain hardening curve for a loading-unloading-reloading cycle in an
gi
elastic-frictional model with hardening.
t d le En
ar
ar s
8.5 Elastoplastic Phenomenological Behavior
ge ro or
eS m
ib
ac
f
Consider a steel bar of length ` and cross-section A subjected to a tensile force F
C d P cs
b
at its extremes. The stress in the bar will be σ = F/A (see Figure 8.18) and the
i
corresponding strain can be estimated as ε = δ /`, where δ is the lengthening
an an n
y ha
of the bar. If the bar is subjected to several loading and unloading cycles, the
response typically obtained, in terms of stress-strain curve σ − ε, is as indicated
le
liv or ec
in Figure 8.19.
M
.A
Observation of the first cycle reveals that, as long as the stress does not ex-
ceed the value σe (denoted as elastic limit) in point 1, the behavior is linear
m
elastic, characterized by the elastic modulus E (σ = Eε), and there do not ex-
d
uu
ist irrecoverable strains (in a possible posterior unloading, the strain produced
e
X Th
For stress values above σe , the behavior ceases to be elastic and part of the
on
strain is no longer recovered during an ensuing unloading to null stress (point 3).
.O
There appears, thus, a remaining strain named plastic strain, ε p . However, dur-
C
incrementally elastic (∆ σ = E ∆ ε). The same occurs with the posterior reload-
ing 3 − 2, which produces an incrementally elastic behavior, until the stress
reaches, in point 2, the maximum value it will have achieved during the loading
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Elastoplastic Phenomenological Behavior 393
second unloading
first unloading
rs
n ee
gi
t d le En
Figure 8.19: Response to loading-unloading-reloading cycles in an uniaxial tensile load-
ing test.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
process. From this point on, the behavior is no longer incrementally elastic (as
b
if the material remembered the maximum stress to which it has been previously
i
an an n
again that, during section 2 − 4, additional plastic strain is generated, which ap-
le
liv or ec
pears in the form of permanent strain in point 5, and, also, additional elastic
strain ε e is produced, understood as the part of the strain that can be recovered
M
.A
er
Consider a sample of virgin material (a material that has not suffered previous
tin
states of inelastic strain) subjected to an uniaxial tensile test and another sample
on
Figure 8.20, for both tests are symmetrical with respect to the origin. That is, in
©
the tensile test the response is elastic up to a value of σ = σe (tensile elastic limit)
and in the compressive response the answer is also elastic up to a value of σ =
−σe (compressive elastic limit), being the rest of both curves (for an assumed
regime of monotonous loading) also symmetrical. In this case, the stress-strain
curve of the virgin material is said to be symmetrical in tension and compression.
Suppose now that a specimen that has been previously subjected to a his-
tory of plastic strains3 , for example a tensile loading-unloading cycle such as
the 0 − 1 − 2 − 3 cycle shown in Figure 8.19, undergoes now a compressive
test. Consider also that σy > σe is the maximum stress the material has been
3 This procedure is known as cold hardening and its purpose is to obtain an apparent elastic
limit that is superior to that of the virgin material σy > σe .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
394 C HAPTER 8. P LASTICITY
rs
n ee
gi
curve of the stretched material
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
curve without the Bauschinger effect
b
i
an an n
y ha
le
liv or ec
M
.A
d
uu
e
X Th
er
tin
.O
1) Unlike in the elastic case, there does not exist unicity in the
stress-strain relation. A same value of strain can correspond to
infinite values of stress and vice-versa. The stress value depends
not only on the strain, but also on the loading history.
2) There does not exist a linear relation between stress and strain.
At most, this linearity may be incremental in certain sections of
the deformation process.
3) Irrecoverable or irreversible strains are produced in a loading-
unloading cycle.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Incremental Theory of Plasticity in 1 Dimension 395
rs
means of piece-wise approximations using
elastic region
elastic and inelastic regions such as the
ee
ones shown in Figure 8.21. The generaliza-
n
tion to several dimensions requires the in-
gi
Figure 8.21: Uniaxial stress-strain
troduction of more abstract concepts.
t d le En
curve for an elastoplastic model.
ar
ar s
ge ro or
eS m
ib
8.6.1 Additive Decomposition of Strain. Hardening Variable
ac
f
The total strain ε is decomposed into the sum of an elastic (or recoverable)
C d P cs
b
strain ε e , governed by Hooke’s law, and a plastic (or irrecoverable) strain ε p ,
i
an an n
y ha
le
e p e p
liv or ec
Additive decomposition ε = ε + ε =⇒ dε = dε + dε
dσ (8.28)
M
.A
of strain σ
εe =
dε e =
E E
m
d
uu
e
er
tin
.O
p
dα = sign (σ ) dε
C
Hardening (8.29)
©
dα ≥ 0
variable α
α p =0
ε =0
4 Up to a certain point, these models may be inspired, albeit with certain limitations, in the
elastic-frictional rheological models described in section 8.4.
5 Here, the sign operator is used, which is defined as x ≥ 0 ⇐⇒ sign (x) = +1 and
x < 0 ⇐⇒ sign (x) = −1.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
396 C HAPTER 8. P LASTICITY
rs
both variables coincide,
ee
p p
Zε Zε
n
dε p ≥ 0 =⇒ α = |dε p | = dε p = ε p .
gi
t d le En
0 0
ar
ar s
ge ro or
eS m
plastic strain may decrease and its value no longer coincides with
ib
ac
that of the hardening variable α.
f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
The elastic domain in the stress space is defined as the interior of the domain
enclosed by the surface F (σ , α) = 0,
m
d
uu
e
(8.30)
X Th
.O
The initial elastic domain E0σ is defined as the elastic domain corresponding
©
An additional requirement of the initial elastic domain is that it must contain the
null stress state,
0 ∈ E0σ =⇒ F (0, 0) < 0 , (8.32)
and this is achieved by defining a yield function of the type
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Incremental Theory of Plasticity in 1 Dimension 397
where σy (α) > 0 is known as the yield stress. The initial value (for α = 0)
of the yield stress is the elastic limit σe (see Figure 8.22) and the function
σy (α) : R+ → R+ is named hardening law.
rs
ee
hardening
n
parameter
gi
t d le En
ar
ar s
ge ro or
eS m
ib
admissible
ac
stress space f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
m
er
tin
Yield surface: ∂ Eσ := σ ∈ R | F (σ , α) ≡ |σ | − σy (α) = 0 (8.34)
on
.O
The elastic domain Eσ together with its boundary ∂ Eσ determine the admissible
C
and it is postulated that any feasible (admissible) stress state must belong to
the admissible stress space Ēσ . Considering the definitions of elastic domain
in (8.30), yield surface in (8.34) and admissible stress space in (8.35), the fol-
lowing is established.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
398 C HAPTER 8. P LASTICITY
σ in the elastic domain
F (σ , α) < 0 ⇐⇒ |σ | < σy (α) ⇐⇒
(σ ∈ Eσ )
σ on the yield surface (8.36)
F (σ , α) = 0 ⇐⇒ |σ | = σy (α) ⇐⇒
(σ ∈ ∂ Eσ )
rs
Remark 8.10. Note how, in (8.35), the admissible stress space de-
ee
pends on the hardening variable α. The admissible domain evolves
n
with the yield function σy (α) such that (see Figure 8.22)
gi
Ēσ ≡ [−σy (α) , σy (α)] .
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
8.6.3 Constitutive Equation
b
i
an an n
To characterize the response of the material, the following situations are defined:
y ha
le
• Elastic regime
liv or ec
M
.A
σ ∈ Eσ =⇒ dσ = Edε (8.37)
m
d
uu
er
tin
)
σ ∈ ∂ Eσ
=⇒ dσ = Edε (8.38)
on
.O
dF (σ , α) < 0
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Incremental Theory of Plasticity in 1 Dimension 399
F (σ , α) ≡ |σ | − σy (α) = 0 .
F (σ + dσ , α + dα) = F (σ , α) + dF (σ , α) > 0
| {z } | {z }
=0 >0
rs
ee
and, in agreement with (8.36), the stress state σ + dσ is not admis-
n
sible.
gi
t d le En
ar
ar s
8.6.4 Hardening Law. Hardening Parameter
ge ro or
eS m
ib
ac
f
The hardening law provides the evolution of the yield stress σy (α) in terms of
C d P cs
the hardening variable α (see Figure 8.22). Even though the aforementioned
b
i
hardening law may be of a more general nature, it is common (and often suffi-
an an n
le
liv or ec
.A
d
uu
e
er
tin
.O
)
©
σ ∈ ∂ Eσ =⇒ F (σ , α) ≡ |σ | − σy (α) = 0
=⇒
dF (σ , α) = 0 (8.41)
d |σ | − dσy (α) = 0 =⇒ sign (σ ) dσ − H 0 dα = 0 ,
where (8.40) has been taken into account. Introducing the first expression of
(8.29) in (8.41) yields
1
sign (σ ) dσ − H 0 sign (σ ) dε p = 0 =⇒ dε p = dσ . (8.42)
H0
6 The property d |x|/dx = sign (x) is used here.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
400 C HAPTER 8. P LASTICITY
rs
dσ = dε =⇒ 0 .
1 1 E ep = E H
ee
+ E + H0
E H0
n
gi
t d le En
8.6.6 Uniaxial Stress-Strain Curve
ar
ar s
The constitutive equation defined by expressions (8.37) to (8.39) allows obtain-
ge ro or
eS m
ib
ing the corresponding stress-strain curve for an uniaxial process of loading-
ac
f
unloading-reloading (see Figure 8.23) in which the following sections are ob-
C d P cs
served.
b
i
an an n
Section 0 − 1:
y ha
le
liv or ec
.A
From (8.37), dσ = Edε and the behavior is linear elastic, defining an elastic
region in the stress-strain curve.
m
d
uu
e
Section 1 − 2 − 4:
X Th
er
tin
)
F (σ , α) ≡ |σ | − σy (α) = 0 =⇒ σ ∈ ∂ Eσ Elastoplastic regime
=⇒
on
.O
dF (σ , α) = 0 in plastic loading
C
Section 2 − 3 − 2:
From (8.37), dσ = Edε and the behavior is linear elastic, defining an elastic
region in the stress-strain curve.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Incremental Theory of Plasticity in 1 Dimension 401
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Figure 8.23: Uniaxial stress-strain curve for a loading-unloading-reloading cycle con-
b
i
sidering the incremental theory of plasticity.
an an n
y ha
le
liv or ec
M
.A
are distinguished:
d
)
F (σ , α) ≡ |σ | − σy (α) = 0 =⇒ σ ∈ ∂ Eσ
uu
Elastic unloading
e
X Th
in section 2 − 3
er
dF (σ , α) < 0
tin
)
on
F (σ , α) ≡ |σ | − σy (α) = 0 =⇒ σ ∈ ∂ Eσ
.O
Plastic loading in
dF (σ , α) = 0 section 2 − 4
C
Remark 8.13. Note that plastic strain is only generated during the
plastic loading process in the elastoplastic region (see Figure 8.24).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
402 C HAPTER 8. P LASTICITY
Elastoplastic
region
Elastic region
rs
ee
Figure 8.24: Generation of plastic strain in the elastoplastic region.
n
gi
t d le En
ar
Remark 8.14. Note the similarity between the stress-strain curve in
ar s
ge ro or
eS m
ib
Figure 8.23 and the one obtained with the elastic-frictional rheologi-
ac
f
cal model with hardening in section 8.4.5 (Figure 8.17). The friction
C d P cs
b
strain in said model is equivalent to the plastic strain in the incre-
i
mental theory of plasticity.
an an n
y ha
le
liv or ec
M
.A
ing (8.43),
X Th
er
H0
tin
E ep = E
E + H0
on
.O
ure 8.25):
©
7 Plasticity with strain softening presents a specific problematic regarding the uniqueness of
the solution to the elastoplastic problem, which is beyond the scope of this text.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Plasticity in 3 Dimensions 403
Figure 8.25: Role of the hardening parameter H 0 in the definition of the slope E ep .
rs
ee
8.7 Plasticity in 3 Dimensions
n
gi
The incremental theory of plasticity developed in one dimension in section 8.6
t d le En
can be generalized to a multiaxial stress state (three dimensions) using the same
ar
ingredients, that is:
ar s
ge ro or
eS m
ib
1) Additive decomposition of strain
ac
f
C d P cs
b
( (
Additive
i
ε = εe +ε p dεε = dεε e + dεε p
an an n
decomposition =⇒ (8.44)
y ha
.A
er
tin
dε p = λ ∂ G (σ
σ , α)
Flow rule ∂σ (8.45)
on
.O
dα = λ α ∈ [0, ∞)
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
404 C HAPTER 8. P LASTICITY
Elastic domain Eσ := { σ | F (σ
σ , α) < 0}
Initial (8.46)
E0σ := { σ | F (σ
σ , 0) < 0}
elastic domain
rs
Yield surface ∂ Eσ := { σ | F (σ
σ , α) = 0}
n ee
Admissible Ēσ = Eσ
S
∂ Eσ = { σ | F (σ
σ , α) ≤ 0}
gi
stress state
t d le En
ar
where φ (σσ ) ≥ 0 is denoted as the equivalent uniaxial stress, σe is the elastic
ar s
ge ro or
eS m
ib
limit obtained in an uniaxial test of the material (it is a material property) and
ac
σy (α) is the yield stress. The hardening parameter H 0 plays the same role as
f
C d P cs
in the uniaxial case and determines the expansion or contraction of the elastic
b
i
domain Eσ , in the stress space, as α grows. Consequently,
an an n
y ha
le
liv or ec
.A
Contraction
m
of Eσ with α
uu
e
X Th
Constant
er
tin
H 0 = 0 =⇒ elastic domain
→ Perfect plasticity
Eσ = E0σ
on
.O
C
Loading-unloading → λ ≥ 0 ; F (σ
σ , α) ≤ 0 ; λ F (σ
σ , α) = 0
conditions
(8.48)
Consistency → If F (σ
σ , α) = 0 ⇒ λ F (σ
σ , α) = 0
condition
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 405
rs
)
σ ∈ ∂ Eσ
=⇒ dσ
σ = C : dεε
ee
(8.50)
dF (σ
σ , α) < 0
n
gi
t d le En
• Elastoplastic regime in plastic loading
ar
ar s
)
ge ro or
σ ∈ ∂ Eσ
eS m
ib
=⇒ dσσ = C ep : dεε (8.51)
ac
dF (σ
σ , α) = 0 f
C d P cs
b
where C ep is known as the elastoplastic constitutive tensor which, after certain
i
an an n
le
liv or ec
M
.A
∂G ∂F
⊗ C: :C
C ep (σ
σ , α) = C − ∂σ ∂σ
m
∂F ∂G
d
H0 + :C :
uu
e
∂σ ∂σ
X Th
(8.52)
er
tin
∂G ∂F
Ci jpq Crskl
on
.O
∂ σ pq ∂ σrs
Cep
i jkl = Ci jkl − i, j, k, l ∈ {1, 2, 3}
∂F ∂G
C
H0 + C pqrs
©
∂ σ pq ∂ σrs
E0σ := { σ | F (σ
σ ) ≡ φ (σ
σ ) − σe < 0} (8.53)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
406 C HAPTER 8. P LASTICITY
and determines a domain in the stress space delimited by the initial yield sur-
face ∂ E0σ ,
∂ E0σ := { σ | F (σ
σ ) ≡ φ (σ
σ ) − σe = 0} . (8.54)
Given that the initial elastic domain contains the origin of the stress space
(σ
σ = 0), every loading process in any point of the medium will include an elas-
tic regime (as long as the trajectory of the stresses remains inside E0σ , see Fig-
ure 8.26) that will end at the instant in which said trajectory reaches the yield
surface ∂ E0σ . The initial yield surface plays then the role of indicating the in-
stant of failure (understood as the end of the elastic behavior) independently
of the possible post-failure (plastic) behavior that initiates beyond this instant.
rs
Thus, the importance of the initial yield surface and the interest in formulat-
ee
ing the mathematical equations that adequately determine this surface for the
n
different materials of interest in engineering.
gi
With the aim of defining the yield surface independently of the reference
t d le En
system (isotropic material)8 , even if formulated in the principal stress space, its
mathematical equation is typically defined in terms of the stress invariants,
ar
ar s
ge ro or
eS m
ib
σ ) ≡ F I1 , J20 , J30 ,
F (σ (8.55)
ac
f
C d P cs
b
and, since the criterion σ1 ≥ σ2 ≥ σ3 is adopted, its definition only affects the
i
an an n
first sector of the principal stress space and can be automatically extended, due
y ha
to symmetry conditions (see Remark 8.7), to the rest of sectors in Figure 8.7.
le
liv or ec
M
.A
m
∂ E0σ := { σ | φ (σ
σ ) − σe = 0}
d
uu
e
X Th
er
E0σ := { σ | φ (σ
σ ) − σe < 0}
tin
on
.O
C
8 An isotropic elastoplastic behavior is characterized by the fact that the yield surface, under-
stood as an additional ingredient of the constitutive equation, is independent of the reference
system.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 407
q
0
Von Mises criterion: σ ) ≡ σ̄ (σ
F (σ σ ) − σe = 3 J2 − σe = 0 (8.56)
q
0
where σ̄ (σ
σ ) = 3 J2 is the effective stress (see Remark 8.3). An alternative
expression is obtained taking (8.19) and (8.20) and replacing them in (8.56),
rs
which produces
ee
1 1/2
n
σ ) ≡ √ (σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ1 − σ3 )2
F (σ − σe = 0 . (8.57)
gi
2
t d le En
The graphical representation of the von Mises yield surface is shown in Fig-
ar
ure 8.27.
ar s
ge ro or
eS m
ib
ac
f
r
C d P cs
2
R=
b
σe r
2
i
3 R=
an an n
σe
3
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
408 C HAPTER 8. P LASTICITY
Example 8.2 – Compute the expression of the von Mises criterion for an
rs
uniaxial tensile loading case.
n ee
Solution
gi
An uniaxial tensile loading case is characterized by the stress state
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
σu 0 0
i
an an n
not
y ha
σ ≡ 0 0 0 .
le
0 0 0
liv or ec
M
.A
The effective stress is known to be σ̄ = |σu | (see Example 8.1) and, replacing
in the expression of the von Mises criterion (8.56), yields
m
d
uu
σ ) − σe = |σu | − σe .
σ ) ≡ σ̄ (σ
F (σ
e
X Th
er
tin
Thus, the initial elastic domain is characterized in the same way as in unidi-
mensional plasticity seen in Section 8.6.2, by the condition
on
.O
C
Example 8.3 – Compute the expression of the von Mises criterion for a stress
state representative of a beam under composed flexure.
Solution
The stress state for a beam under composed flexure is
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 409
rs
ee
σx τxy 0
n
1
gi
not
σ ≡ τxy 0 0 =⇒ σm = σx =⇒
3
t d le En
0 0 0
ar
ar s
ge ro or
eS m
ib
ac
2
f
σx τxy 0
C d P cs
3
b
1 1
i
0 not
σ = σ − σx 1 ≡ τxy − σx 0 .
an an n
3 3
y ha
1
0 0 − σx
le
liv or ec
3
0
M
.A
0 1 0 0 1 4 2 1 2 1 2 1
d
2 2 9 9 9 3
e
X Th
er
tin
And the effective stress is obtained for the von Mises criterion,
on
.O
q q
0
σ̄ = 3 J2 = σx2 + 3τxy 2 =⇒ F (σ σ ) < 0 =⇒ σ̄ < σe =⇒
C
q
σco = σx2 + 3τxy
2 <σ
e .
q
The comparison stress, σco = σxx 2 + 3τ 2 , which can be regarded as a scalar
xy
for comparison with the uniaxial elastic limit σe , is commonly used in the
design standards of metallic structures.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
410 C HAPTER 8. P LASTICITY
rs
dimensions. In a loading process in which this circle increases starting from
the origin, the elastic behavior ends when the circle with radius τ max becomes
ee
tangent to the straight line τ = τ max = σe /2.
n
It follows from (8.58) that the Tresca criterion can be written as
gi
t d le En
σ ) ≡ (σ1 − σ3 ) − σe = 0 (8.59)
ar
Tresca criterion: F (σ
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
.A
0 0
Tresca criterion: F (σ σ ) ≡ (σ1 − σ3 ) − σe ≡ F J2 , J3
m
d
uu
e
X Th
er
tin
on
.O
C
Figure 8.28: Representation of the Tresca criterion using Mohr’s circle in three dimen-
sions.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 411
von Mises
Tresca
rs
n ee
gi
Figure 8.29: Tresca criterion in the principal stress space.
t d le En
ar
ar s
ge ro or
Figure 8.29 shows the yield surface corresponding to the Tresca criterion in
eS m
ib
ac
the principal stress space, which results in an hexahedral prism whose axis is the
f
C d P cs
hydrostatic stress axis.
b
i
an an n
y ha
Remark 8.19. Since the Tresca criterion does not depend on the first
le
liv or ec
stress invariant (and, therefore, on the stress σoct , see (8.16)), the cor-
responding yield surface does not depend on the distance from the
M
.A
origin to the octahedral plane containing the point (see Remark 8.4).
m
ants (I1 , J20 , J30 ), is on said yield surface, all the points in the stress
uu
e
space with the same values of J20 and J30 will also be on this surface.
X Th
er
tin
.O
On the other hand, the dependency on the two invariants J20 and J30 ,
C
prevents (unlike in the case of the von Mises criterion) the surface
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
412 C HAPTER 8. P LASTICITY
Example 8.4 – Compute the expression of the Tresca criterion for an uniaxial
tensile loading case.
Solution
An uniaxial tensile load case is characterized by the stress state
rs
σu 0 0
ee
not
σ ≡ 0 0 0 .
n
0 0 0
gi
t d le En
For the case σu ≥ 0 ,
ar
ar s
)
ge ro or
σ1 = σu
eS m
ib
=⇒ F (σ1 , σ2 , σ3 ) = (σ1 − σ3 ) − σe = σu − σe = |σu | − σe .
ac
σ3 = 0 f
C d P cs
b
i
For the case σu < 0,
an an n
y ha
)
σ1 = 0
le
liv or ec
.A
And the initial elastic domain is then characterized in the same way as in the
m
er
tin
.O
C
τ = c − σ tan φ . (8.60)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 413
rs
n ee
gi
Figure 8.30: Representation of the Mohr-Coulomb criterion using Mohr’s circle in three
t d le En
dimensions.
ar
ar s
ge ro or
eS m
ib
ac
The end of the elastic behavior (failure) in an increasing load process takes
f
C d P cs
place when the first point in the Mohr’s circle (corresponding to a certain plane)
reaches the aforementioned yield line.
b
i
an an n
The shear stress in this plane, τ, becomes smaller as the normal stress σ in the
y ha
plane increases. It therefore becomes obvious that the behavior of the model un-
le
liv or ec
.A
stress axis in the positive side of these stresses, limiting thus the material’s ca-
m
state for which plasticization initiates. In such case, the corresponding Mohr’s
X Th
er
tin
circle is defined by the major and minor principal stresses and is tangent to the
yield line at point A (see Figure 8.31), verifying
on
.O
C
σ − σ3
τA = R cos φ = 1
cos φ
©
σ1 − σ3 2
R= =⇒
2
σA = σ1 + σ3 + R sin φ = σ1 + σ3 + σ1 − σ3 sin φ
2 2 2
(8.61)
and, replacing (8.61) in (8.60), results in
τA = c − σA tan φ =⇒ τA + σA tan φ − c = 0 =⇒
σ1 − σ3 σ1 + σ3 σ1 − σ3
cos φ + + sin φ tan φ − c = 0 =⇒ (8.62)
2 2 2
(σ1 − σ3 ) + (σ1 + σ3 ) sin φ − 2 c cos φ = 0 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
414 C HAPTER 8. P LASTICITY
Figure 8.31: Deduction of the expression for the Mohr-Coulomb criterion using Mohr’s
circle.
rs
n ee
gi
t d le En
Mohr-Coulomb criterion:
(8.63)
ar
σ ) ≡ (σ1 − σ3 ) + (σ1 + σ3 ) sin φ − 2 c cos φ = 0
F (σ
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
le
σ ) ≡ (σ1 − σ3 ) + (σ1 + σ3 ) sin φ − 2 c cos φ = 0 ,
F (σ
liv or ec
M
.A
into account symmetry conditions, to the other five sectors (see Re-
uu
e
er
length whose axis is the hydrostatic stress axis (see Figure 8.32). The
tin
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 415
√
3 c cot φ
rs
n ee
gi
Figure 8.32: Mohr-Coulomb criterion in the principal stress space.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 8.23. In soil mechanics, the sign criterion of the normal
C d P cs
b
stresses is the opposite to the one used in continuum mechanics
i
(σ ≡ −σ , see Chapter 4) and, thus, σ1 ≡ −σ3 and σ3 ≡ −σ1 . Then,
an an n
y ha
.A
er
tin
on
.O
C
Figure 8.33: Representation of the Mohr-Coulomb criterion using Mohr’s circle in three
dimensions and soil mechanics sign criterion.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
416 C HAPTER 8. P LASTICITY
rs
Figure 8.34: Mohr-Coulomb criterion in the principal stress space, using soil mechanics
ee
sign criterion.
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Remark 8.24. Following certain algebraic operations, the Mohr-
ac
f
Coulomb criterion can be written in terms of the three stress invari-
C d P cs
b
ants.
i
an an n
0 0
F (σ ) ≡ F I , J , J
y ha
Mohr-Coulomb criterion: σ 1 2 3
le
liv or ec
M
.A
m
er
tin
.O
C
0 1/2
Drucker-Prager criterion: σ ) ≡ 3 ασm + J2
F (σ −β = 0 (8.64)
where
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Yield Surfaces. Failure Criteria 417
2 sin φ 6 c cos φ σ1 + σ2 + σ3 I1
α=√ ,β=√ and σm = = , (8.65)
3 (3 − sin φ ) 3 (3 − sin φ ) 3 3
being c and φ the cohesion and the internal friction angle, respectively, which are
considered to be material properties. Considering (8.16) and (8.18), the criterion
can be rewritten as
0 1/2 r
3 0
σ ) ≡ αI1 + J2
F (σ − β = 3 ασoct + τoct − β = F I1 , J2 = 0 . (8.66)
2
rs
ee
0
Remark 8.26. The independence on the third stress invariant J3 es-
n
tablishes that, if a certain point in the stress space belongs to the
gi
yield surface, all the other points with the same value of the stress
t d le En
0
invariants I1 and J2 also belong to this surface, independently of the
0
ar
value of the third stress invariant J3 . Given that the constant values of
ar s
ge ro or
eS m
ib
these invariants correspond to points of the octahedral plane placed
ac
f
at a same distance from the hydrostatic stress axis (see Figure 8.6),
C d P cs
b
it can be concluded that the yield surface is a surface of revolution
i
around this axis.
an an n
y ha
static stress axis (see Figure 8.5 and Figure 8.35). The distance from
M
.A
the origin
√ of the principal stress space to the vertex of the cone is
d = 3 c cot φ . It can be verified that the Drucker-Prager surface has
m
er
tin
on
.O
C
Drucker-
Prager
Mohr-
Coulomb
√
3 c cot φ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
418 C HAPTER 8. P LASTICITY
rs
n ee
gi
Remark 8.28. In soil mechanics, where the sign criterion for the nor-
t d le En
mal stresses is inverted, the yield surface for the Drucker-Prager cri-
terion is as indicated in Figure 8.36.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
.A
m
d
uu
e
X Th
er
tin
on
.O
Drucker-
Prager
C
Mohr-
©
Coulomb
Figure 8.36: Drucker-Prager criterion in the principal stress space, using soil mechanics
sign criterion.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 419
P ROBLEMS
Problem 8.1 – Justify the shape the yield surface will have in the principal
stress space for each of the following cases:
a) f I12 = 0
b) f (J 0 2 ) = 0
rs
c) a I12 + b τoct
2 = c with a, b and c strictly positive
n ee
gi
t d le En
Solution
ar
ar s
ge ro or
a) In this case, there is a condition on the mean stress since
eS m
ib
ac
f
I1 = σ1 + σ2 + σ3 = 3σm .
C d P cs
b
i
Then, the yield surface is an octahedral plane whose distance to the origin is
an an n
y ha
imposed by the first stress invariant. However, because this invariant is squared,
there are two octahedral planes, one in each direction of the hydrostatic stress
le
liv or ec
axis.
M
.A
m
d
uu
e
√
X Th
3 σm
er
tin
√
3 σm
on
.O
C
b) Here, the distance between a given stress state and an hydrostatic stress
state is imposed. So, the yield surface is a cylinder with circular section in the
octahedral planes,
3 2 √
J 0 2 = τoct =⇒ distance = 3 τoct .
2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
420 C HAPTER 8. P LASTICITY
√
3 τoct
rs
n ee
c) The representation of a plane defined by a given point of the yield surface
gi
and the hydrostatic stress axis is:
t d le En
ar
ar s
ge ro or
eS m
ib
Point
ac
f
C d P cs
b
i
an an n
y ha
.A
√ √
√
m
3 I1 = 3 x
d
d = x = 3 σoct = I1
=⇒
uu
√ 3 R
e
τoct = √
R = y = 3 τoct
X Th
3
er
tin
are deduced and replacing these values in the given expression of the yield sur-
on
.O
face results in
C
2 2
©
2 2 2 by2 rx ry
a I1 + b τoct = c =⇒ 3ax + = c =⇒ c + 3c = 1 .
3
3a b
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 421
2 2 2
rx ry rz
c +
3c + 3c = 1 .
3a b b
rs
ee
hydrostatic stress axis
n
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Problem 8.2 – Graphically determine, indicating the most significant values,
ac
f
the cohesion and internal friction angle of an elastoplastic material that follows
C d P cs
b
the Mohr-Coulomb yield criterion using the following information:
i
an an n
plasticizes at σ = σA .
le
liv or ec
.A
plasticizes at σ = σB .
m
d
uu
e
X Th
er
Solution
tin
In the uniaxial tensile stress state, the Mohr’s circle will cross the origin and the
on
.O
value σ = σA in the horizontal axis. However, for the triaxial isotensile stress
C
state, the Mohr’s circle will degenerate to a point in this axis, σ = σB . Thus, the
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
422 C HAPTER 8. P LASTICITY
σA /2
φ = arcsin and c = σB tan φ .
σB − σA /2
rs
n ee
Problem 8.3 – The following properties of a certain material have been exper-
gi
imentally determined:
t d le En
1) In a hydrostatic compressive regime, the material never plasticizes.
ar
ar s
ge ro or
eS m
ib
2) In a hydrostatic tensile regime, the virgin material plasticizes for a value of
the mean stress σm = σ ∗ .
ac
f
C d P cs
b
3) In an uniaxial tensile regime, the virgin material plasticizes for a tensile
i
an an n
stress value σu .
y ha
.A
σ = σ 0 : σ 0 = a σm + b .
m
Plot the yield surface, indicating the most significant values, and calculate the
d
er
tin
on
.O
Solution
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 423
stress axis and whose vertex is in the tensile part of this axis:
√ √
3 σm = 3 σ ∗
hydrostatic
stress axis
rs
n ee
√
gi
σ 0 | = σ 0 : σ 0 = a σm + b
To calculate the values of a and b, the yield criterion |σ
t d le En
is applied on the vertex of the cone, which corresponds with the hydrostatic
ar
tensile case and, thus, has no deviatoric stresses,
ar s
ge ro or
eS m
ib
ac
|σ 0
σ | = 0 =⇒ aσm
f
+ b = 0 =⇒ aσ ∗ + b = 0 . [1]
C d P cs
∗ σm =σ
b
i
an an n
The procedure is repeated for the uniaxial tensile case, whose deviatoric stresses
y ha
are now
le
liv or ec
M
σu 0 0 2 0 0
.A
1 not σu
σ ≡ 0 0 0 and σ sph = σu 1 =⇒ σ 0 ≡
not
0 −1 0 .
m
3 3
d
0 0 0 0 0 −1
uu
e
X Th
er
tin
σ 0 | = a σm + b produces
Then, applying the yield criterion |σ
on
.O
r r
2 2 1
σ 0| =
|σ σu =⇒ σu = a σu + b . [2]
C
3 3 3
©
Equations [1] and [2] allow determining the desired values of a and b as
r r
2 2
σu σu σ ∗
3 3
a= σ and b=− σ .
u u
−σ∗ −σ∗
3 3
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
424 C HAPTER 8. P LASTICITY
Problem 8.4 – The metallic component PQRS has a thickness “e” and is com-
posed of two different materials, (1) and (2), considered to be perfect elasto-
plastic materials. The component is subjected to a pure shear test by means of
the machine shown in Figure A, such that the uniform stress and strain states
produced are
δ
εx = εy = εz = 0 , γxz = γyz = 0 , γxy = γ = ,
h
σx = σy = σz = 0 , τxz = τyz = 0 and τxy = τ 6= 0 .
When a component exclusively composed of one of the materials is tested sep-
rs
arately, a τ − γ curve of the type shown in Figure B is obtained for both mate-
ee
rials. Determine:
n
gi
a) The elastic limit that will be obtained in separate uniaxial tensile tests of
each material, assuming they follow the von Mises criterion.
t d le En
When the component composed of the two materials is tested, the P − δ curve
ar
ar s
ge ro or
shown in Figure C is obtained. Determine:
eS m
ib
ac
f
b) The values of the elastic load and displacement, Pe and δe .
C d P cs
b
c) The values of the plastic load and displacement, Pp and δp .
i
an an n
le
liv or ec
rigid frame
M
.A
m
d
uu
e
X Th
er
tin
on
.O
Figure A
C
H YPOTHESES :
Material (1)
G = G and τe = τ ∗
Material (2)
G = G and τe = 2 τ ∗
Figure B Figure C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 425
Solution
a) In an uniaxial state of stress, plasticization according to the von Mises crite-
rion is known to begin when (see Example 8.2)
σ̄ = σe ,
where σ̄ is the effective stress and σe is the elastic limit. In addition, the follow-
ing relations seen in this chapter, are known to hold.
1 1
σ̄ = (3 J20 ) 2 J20 = Tr σ 0 2
rs
2
1
ee
σ 0 = σ − σ sph σ sph = σm 1 σm = Tr (σ
σ)
3
n
gi
For this problem in particular,
t d le En
0 τ 0
ar
ar s
ge ro or
eS m
ib
not
σ ≡ τ 0 0 ,
ac
0 0 0 f
C d P cs
b
i
an an n
le
τ2 0 0
liv or ec
2 √
σ0 J02 = τ 2
not
M
≡ 0 τ2 0
.A
=⇒ =⇒ σ= 3τ .
σ̄
0 0 0
m
d
uu
Considering that material (1) plasticizes when τe = τ ∗ and material (2), when
e
X Th
er
τe = 2τ ∗ , then
tin
√
Material 1 =⇒ σe = 3 τ ∗ ,
on
.O
√
C
Material 2 =⇒ σe = 2 3 τ ∗ .
©
b) The elastic load Pe and the elastic displacement δe determine the end of the
elastic regime in the component. The statement of the problem indicates that
when the materials are tested separately, the τ − γ curve in Figure B is obtained,
where τe = τ ∗ in material (1) and τe = 2τ ∗ in material (2). It is also known that G
is the same in both materials, that is, they have the same slope in their respective
τ − γ curves.
Now, to determine the combined behavior of these materials in the metallic
component, one can assume that the behavior will be elastic in this component
as long as both materials are in their corresponding elastic domain. Therefore,
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
426 C HAPTER 8. P LASTICITY
since the elastic interval of material (1) is smaller, then this material will define
the elastic domain of the whole component (up to point A in Figure C).
rs
n ee
gi
To obtain the value of the elastic force, equilibrium of forces is imposed for the
t d le En
force Pe and the stresses each material has at point A. Note that equilibrium is im-
posed on forces, therefore, stresses must be multiplied by the surface on which
ar
ar s
ge ro or
they act, considering the magnitude perpendicular to the plane of the paper as
eS m
ib
the unit value.
ac
f
C d P cs
b
h A h A h ∗ h ∗
Pe = τ + τ = τ + τ =⇒ Pe = h τ ∗
i
2 1 2 2
an an n
2 2
y ha
two materials,
τ∗
M
.A
δe = γ1A h = γ2A h =⇒ δe = h .
G
m
d
uu
e
c) To obtain the plastic values Pp and δ p one must take into account that, at
X Th
er
tin
point A, material (1) begins to plasticize, while material (2) initiates plasticiza-
tion at point B. Therefore, the behavior of the complete component will be per-
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 427
and, imposing the equilibrium and compatibility equations, yields the values of
Pp and δ p .
h B h B h ∗ h ∗ 3
Pp = τ ∗ h and
Pp = τ1 + τ2 = τ + 2τ
2 2 2 2 =⇒ 2
2τ ∗ τ ∗h
δ p = γ1B h = γ2B h = δp = 2 = 2δe .
h
G G
rs
d) The coordinates of points A and B have already been obtained. The statement
of the problem gives the value of point B0 , which corresponds to a deformation
ee
of 3δe when the plastic load Pp is maintained constant (perfectly plastic regime).
n
Consider first the material (1). Unloading takes place starting at B0 and, accord-
gi
ing to the information given, this material plasticizes when it reaches a value
t d le En
of −τ ∗ . The slope of the curve is still the value of the material parameter G
ar
since this is independent of the material being under loading or unloading con-
ar s
ge ro or
eS m
ib
ditions. Thus, to determine point C it is enough to draw a straight line that crosses
ac
f
point B0 and is parallel to OA, until the value −τ ∗ is reached.
C d P cs
The same occurs in the case of material (2), with the difference that when the
b
line parallel to OA is drawn to cross point B0 , this line must be extended to the
i
an an n
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
428 C HAPTER 8. P LASTICITY
To obtain the load and displacement values at point C, the equilibrium and com-
patibility equations are imposed. Taking into account the τ and γ values obtained
at point C in the curves above yields
∗h
h C h C h h τ τ ∗h
PC = τ1 + τ2 = (−τ ∗ ) + (0) = −
PC = − and
2 2 2 2 2 =⇒ 2
∗
τ τ ∗h
δC = γ1C h = h δC = = δe .
G G
rs
Repeating the procedure for point D results in
ee
∗h
3τ ∗ h
n
h D h D h h 3τ
PD = τ1 + τ2 = (−τ ∗ ) + (−2τ ∗ ) = −
PD = − and
gi
2 2 2 2 2 ⇒ 2
t d le En
∗
τ τ ∗h
δD = γ2D h = − h δD = − = −δe .
ar
G G
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
Problem 8.5 – Consider the solid cylinder shown in Figure A, which is fully
y ha
fixed at its base and has a torsional moment M applied on its top end. The
le
cylinder is composed of two materials, (1) and (2), which have an elastoplas-
liv or ec
.A
T
d
not T θ
uu
u (r, θ , z) ≡ [ur , uθ , uz ] = 0 , rz , 0 ,
e
h
X Th
er
tin
where φ is the rotation of the section at the free end of the cylinder. Assuming
on
.O
a) The strain and stress tensors, ε and σ , in cylindrical coordinates and elas-
©
tic regime. Plot, indicating the most significant values, the σrr − r and
τθ z − r curves for a cross-section of the cylinder at height z. Schematically
represent the stress distribution of τθ z in this cross-section.
b) The value of φ = φe (see Figure C) for which plasticization begins in at least
one point of the cylinder, indicating where it begins and the corresponding
value of the moment
Z M = Me .
NOTE: M = r τθ z dS
S
c) The minimum value of φ = φ1 for which material (1) has totally plasticized
and the corresponding value of M = M1 (see Figure C). Schematically rep-
resent the stress distribution in a cross-section at this instant.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 429
d) The minimum value of φ = φ2 for which material (2) has totally plasticized
and the corresponding value of M = M2 (see Figure C). Schematically rep-
resent the stress distribution in a cross-section at this instant.
e) The asymptotic value of M = M p (= plastic moment) corresponding to the
plasticization of the complete cross-section. Schematically represent the
stress distribution in a cross-section at this instant.
rs
n ee
gi
t d le En
Figure B
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
Figure A
.A
Figure C
m
H YPOTHESES :
d
er
tin
on
.O
C
Solution
©
a) The infinitesimal strain tensor is calculated directly from the given displace-
ment field, both in cylindrical coordinates,
0 0 0
not φ r
ε ≡ 0 0 .
2h
φr
0 0
2h
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
430 C HAPTER 8. P LASTICITY
rs
ee
0 00
n
not Gφ r
gi
σ ≡ 0 0 .
h
t d le En
Gφ r
0 0
h
ar
ar s
ge ro or
eS m
ib
ac
f
Plotting the σrr and τθ z components of the stress tensor in terms of the radius r
C d P cs
yields:
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
The stresses are linear and do not depend on the z-coordinate of the cross-section
considered. Thus, the distribution of stresses in any cross-section (z = const.) of
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 431
Z2πZR ZR
Gφ r3 πGR4
Gφ r
Z
M= r τ (r) dS = r r dr dθ = 2π dr = φ. [1]
h h 2h
S 0 0 0
This is the relation between the moment and the rotation angle (M − φ ) at the
free end of the cylinder when the two materials behave elastically.
Material (1) starts to plasticize first at τe = τ ∗ , since material (2) plasticizes at a
rs
higher stress, τe = 2τ ∗ . In addition, the external surface of the cylinder (r = R)
ee
suffers larger stresses, and this surface is composed of material (1). Therefore,
plasticization will initiate when
n
gi
Gφe R τ ∗h
t d le En
= τ∗ =⇒ = τ∗ =⇒ φe =
τ
r=R; φ =φe h GR
ar
ar s
ge ro or
eS m
ib
is satisfied. This is the value of the rotation angle at the free end of the cylinder
ac
f
required for plasticization to initiate in the exterior material points of the cylinder
C d P cs
b
(material (1)). The corresponding moment is obtained by replacing φe in [1],
i
an an n
y ha
πGR4 πτ ∗ R3
Me = M (φe ) = φe =⇒ Me = .
le
2h 2
liv or ec
M
.A
c) If the material were elastic, the slope of the stresses τ would increase with φ
m
(remaining, though, linear with r), but since the material is now elastoplastic,
uu
e
stresses cannot exceed the value τe , which corresponds to the onset of plasticity.
X Th
That is, material (1) has a perfectly plastic distribution of stresses while material
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
432 C HAPTER 8. P LASTICITY
This is the minimum value of the rotation angle at the free end of the cylinder
required for material (1) to be completely plasticized.
In order to compute the corresponding moment, relation [1] between M and
φ is no longer valid here because material (1) behaves elastoplastically while
material (2) behaves completely elastically. The moment acting on the cylinder
rs
is now
ee
Z2π ZR Z2π ZR/2
n
∗ Gφ1 r
M1 = rτ r dr dθ + r r dr dθ =
gi
h
t d le En
0 R/2 0 0
ZR ZR/2
ar
φ1 31 ∗ 3
ar s
∗ 2
r3 dr
ge ro or
= 2πτ r dr + 2πG =⇒ M1 = πτ R .
eS m
ib
h 48
ac
R/2 0
f
C d P cs
b
i
an an n
d) Material (2) starts plasticizing for τe = 2τ ∗ , which does not correspond with
y ha
.A
m
d
uu
e
X Th
er
tin
on
.O
C
The following condition is imposed to obtain the value of the rotation angle.
Gφ2 R 4τ ∗ h
= 2τ ∗ =⇒ = 2τ ∗ =⇒ φ2 = .
τ
r=R/2; φ =φ2 2h GR
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 433
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
Mp = r τ ∗ r dr dθ + r 2τ ∗ r dr dθ =⇒ M p = πτ ∗ R3 .
d
4
uu
e
0 R/2 0 0
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
434 C HAPTER 8. P LASTICITY
E XERCISES
8.1 – Formulate in terms of the stress invariants I1 , J20 and J30 the equation of
the yield surface that, in the principal stress space, is a spheroid (ellipsoid of
revolution) with semi-axes a and b.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Intersection with octahedral
ac
f plane at (0, 0, 0)
C d P cs
b
i
an an n
y ha
8.2 – An elastoplastic material is subjected to a pure shear test (I) and an uni-
le
liv or ec
.A
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 435
rs
n ee
gi
t d le En
Infinitely rigid
material
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
Figure A
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
Figure B
Additional hypotheses:
1) Young’s modulus, E, and Poisson’s coefficient, ν.
2) Elastic limit, σe .
3) Thickness of the component, b.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
436 C HAPTER 8. P LASTICITY
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
8.5 – Consider a solid sphere with radius R1 encased inside a spherical shell
d
uu
with interior radius R1 and exterior radius R2 . The sphere and the shell are
e
X Th
composed of the same material and are initially in contact without exerting any
er
tin
.O
Determine:
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 437
Additional hypotheses:
1) Material properties:
− Young’s modulus, E, and Poisson’s coefficient, ν = 0.
− Thermal constant, α.
− Yield stress, σy .
− Radii, R1 = 1 and R2 = 3.
2) The body forces are negligible.
3) The displacement and stress fields of a spherical shell with interior radius Ri
and exterior radius Re subjected to an interior pressure Pi and an exterior
rs
pressure Pe are, for ν = 0:
ee
ur (r)
n
C1 Pi R3i − Pe R3e Pi − Pe R3i R3e
gi
u= 0 ur = Cr + 2 ; C = ; C =
1
E R3e − R3i 2E R3e − R3i
r
t d le En
0
ar
ar s
ge ro or
σrr 0 0
eS m
ib
2C1 C1
ac
σ = 0 σθ θ 0
r f
σrr = E C − 3 ; σθ θ = σφ φ = E C + 3
r
C d P cs
0 0 σφ φ
b
i
an an n
y ha
le
liv or ec
8.6 – Consider a solid sphere with radius R1 and composed of material (1), en-
M
.A
cased inside a spherical shell with interior radius R1 , exterior radius R2 and
composed of material (2). The sphere and the shell are initially in contact with-
m
out exerting any pressure on each other. An exterior pressure P is applied simul-
uu
e
er
tin
.O
b) Obtain the stress state of the shell and the sphere for these values.
©
c) Under these conditions, compute, for each value of the pressure P, the value
of ∆ θ ∗ for which plasticization initiates at some point of the sphere or the
shell, according to the von Mises and Mohr-Coulomb criteria. Plot the cor-
responding P − ∆ θ ∗ curves (interaction graphs).
Additional hypotheses:
1) Material properties:
− Young’s moduli, E (1) = E (2) = E, and Poisson’s coefficients, ν (1) = ν (2) = 0.
− Thermal constants, α (1) = 2α and α (2) = α.
(1) (2)
− Yield stresses, σy = σy = σy .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
438 C HAPTER 8. P LASTICITY
rs
0
ee
σrr 0 0
n
2C1 C1
gi
σ = 0 σθ θ 0 σrr = E C − 3 ; σθ θ = σφ φ = E C + 3
r r
t d le En
0 0 σφ φ
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
8.7 – A cylinder of radius R and height h is subjected to an exterior load P and
a uniform temperature increment ∆ θ .
b
i
an an n
y ha
constants.
M
.A
er
tin
.O
Additional hypotheses:
1) Material properties:
− Cohesion value, C, and internal friction angle, φ = 30◦ .
− Thermal constant, β .
− Lamé parameters, λ = µ.
2) The body forces are negligible.
3) The friction between the cylinder and the ground can be neglected.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 9
Constitutive Equations in Fluids
rs
n ee
gi
9.1 Concept of Pressure
t d le En
ar
Several concepts of pressure are used in continuum mechanics (hydrostatic pres-
ar s
ge ro or
eS m
ib
sure, mean pressure and thermodynamic pressure) which, in general, do not co-
ac
incide.
f
C d P cs
b
i
9.1.1 Hydrostatic Pressure
an an n
y ha
le
liv or ec
.A
stress.
uu
e
X Th
er
tin
In accordance with Pascal’s law, the stress state of a fluid at rest is characterized
on
.O
σ = −p0 1
, (9.1)
σi j = −p0 δi j i, j ∈ {1, 2, 3}
439
440 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
Figure 9.2: Mohr’s circle of the stress tensor of a fluid at rest.
ac
f
C d P cs
b
i
an an n
y ha
and its representation in the Mohr’s plane is a point (see Figure 9.2).
Consequently, any direction is a principal stress direction and the
M
.A
er
tin
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Concept of Pressure 441
Remark 9.2. In a fluid at rest, the mean pressure p̄ coincides with the
hydrostatic pressure p0 ,
1
σ = −p0 1 =⇒ σm = (−3p0 ) = −p0 =⇒ p̄ = p0 .
3
Generally, in a fluid in motion the mean pressure and the hydrostatic
pressure do not coincide.
rs
n ee
gi
Remark 9.3. The trace of the Cauchy stress tensor is a stress invari-
t d le En
ant. Consequently, the mean stress and the mean pressure are also
stress invariants and, therefore, their values do not depend on the
ar
ar s
Cartesian coordinate system used.
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
le
A new thermodynamic pressure variable, named thermodynamic pressure and
liv or ec
.A
m
er
.O
C
Example 9.1
The ideal gas law is a typical example of kinetic equation of state:
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
442 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
rs
n ee
Remark 9.5. A barotropic fluid is defined by a kinetic equation of
gi
state in which the temperature does not intervene.
t d le En
Barotropic fluid : F (p, ρ) = 0 =⇒ p = f (ρ) =⇒ ρ = g (p)
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
.A
F (p, ρ, θ ) ≡ ρ − k = 0
m
er
tin
on
.O
Here, the set of equations, generically named constitutive equations, that must
be added to the balance equations to formulate a problem in fluid mechanics
(see Section 5.13 in Chapter 5) is considered. These equations can be grouped
as follows:
a) Thermo-mechanical constitutive equation
This equation expresses the Cauchy stress tensor in terms of the other ther-
modynamic variables, typically the thermodynamic pressure p, the strain
rate tensor d (which can be considered an implicit function of the velocity,
d (v) = ∇S v), the density ρ and the absolute temperature θ .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Constitutive Equation in Viscous Fluids 443
rs
constitutive equation:
n ee
c) Thermodynamic constitutive equations or equations of state
gi
t d le En
These are typically the caloric equation of state, which defines the specific
internal energy u, and the kinetic equation of state, which provides an equa-
ar
ar s
ge ro or
tion for the thermodynamic pressure.
eS m
ib
ac
Caloric equation of f
u = g (ρ, θ )
C d P cs
state:
Kinetic equation of
b 2 equations (9.4)
i
F (ρ, p, θ ) = 0
an an n
state:
y ha
le
liv or ec
.A
The most common one is Fourier’s law, which defines the heat flux by con-
m
duction q as
d
uu
e
q = −k · ∇θ
X Th
Fourier’s
er
tin
∂θ 3 equations (9.5)
law: q i = ki j i ∈ {1, 2, 3}
∂xj
on
.O
which is a property of the fluid. For the isotropic case, the thermal conductiv-
©
σ = −p 1 + f (d, ρ, θ )
, (9.6)
σi j = −p δi j + fi j (d, ρ, θ ) i, j ∈ {1, 2, 3}
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
444 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
rs
be considered.
n ee
gi
Remark 9.7. The perfect fluid hypothesis is frequently used in hy-
draulic engineering, where the fluid under consideration is water.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
9.4 Constitutive Equation in Newtonian Fluids
b
i
an an n
y ha
σ = −p 1 +C
C:d
M
.A
, (9.7)
σi j = −p δi j + Ci jkl dkl i, j ∈ {1, 2, 3}
m
d
uu
e
er
pendency of the stress tensor σ on the strain rate tensor d is obtained as a result
tin
.O
fourth-order tensor.
C
(
C = λ 1 ⊗ 1 + 2µI
©
(9.8)
Ci jkl = λ δi j δkl + µ δik δ jl + δil δ jk i, j, k, l ∈ {1, 2, 3}
1 Note that the thermal dependencies of the constitutive equation are not considered here and,
thus, the name mechanical constitutive equations.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Constitutive Equation in Newtonian Fluids 445
(
Constit. eqn. of σ = −p 1 + λ Tr (d) 1 + 2µ d
an isotropic (9.10)
Newtonian fluid σi j = −p δi j + λ dll δi j + 2µ di j i, j ∈ {1, 2, 3}
Remark 9.8. Note the parallelism that can be established between the
constitutive equation of a Newtonian fluid and that of a linear elastic
solid (see Chapter 6):
rs
Newtonian fluid Linear elastic solid
ee
( (
σ = −p 1 +C C:d σ =C:ε
n
σi j = −p δi j + Ci jkl dkl σi j = Ci jkl εkl
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
Remark 9.9. The parameters λ and µ physically correspond to the
b
viscosities, which are understood as material properties. In the most
i
an an n
general case, they may not be constant and can depend on other ther-
y ha
modynamic variables,
le
liv or ec
.A
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
446 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
σ = −p 1 + λ Tr (d) 1 + 2µ d =⇒
rs
| {z }
−3 p̄
2
p = p̄ + λ + µ Tr (d) = p̄ + K Tr (d)
ee
3
n
| {z }
gi
K
(9.11)
t d le En
where K is denoted as bulk viscosity.
ar
ar s
ge ro or
eS m
ib
ac
2
f
Bulk viscosity : K = λ + µ (9.12)
C d P cs
3
b
i
an an n
y ha
dρ 1 dρ
+ ρ∇ · v = 0 =⇒ ∇·v = − (9.13)
M
.A
dt ρ dt
m
∂ vi
X Th
∂ xi
on
.O
K dρ
p = p̄ + K∇ · v = p̄ − (9.15)
ρ dt
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Constitutive Equation in Newtonian Fluids 447
rs
2
K=0
ee
=⇒ λ =− µ =⇒ p = p̄
3
n
gi
t d le En
ar
ar s
ge ro or
9.4.2 Constitutive Equation in Spherical and Deviatoric Components
eS m
ib
ac
Spherical part f
C d P cs
b
From (9.15), the following relation is deduced.
i
an an n
y ha
p̄ = p − K ∇ · v = p − K Tr (d) (9.16)
le
liv or ec
M
.A
Deviatoric part
m
Using the decomposition of the stress tensor σ and the strain rate tensor d
uu
e
in its spherical and deviator components, and replacing in the constitutive equa-
X Th
er
1
on
.O
σ= σ ) 1 + σ 0 = − p̄ 1 + σ 0 = −p 1 + λ Tr (d) 1 + 2µ d =⇒
Tr (σ
3 | {z }
C
−3 p̄
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
448 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
σ 0 = 2µd0 (9.18)
rs
1
σ = − p̄ 1 + σ 0 Tr (d) 1 + d0 ,
ee
and d= (9.19)
3
n
gi
and replacing in the expression of the stress power density (stress power per unit
of volume) σ : d, results in3
t d le En
ar
1
ar s
ge ro or
σ : d = (− p̄ 1 + σ 0 ) : Tr (d) 1 + d =0
eS m
ib
3
ac
1 f 1
σ 0 : d0 − p̄ 1 : d0 + Tr (d) σ 0 : 1 =
C d P cs
= − p̄ Tr (d) 1 : 1 +σ
3
b 3 (9.20)
i
|{z} | {z } | {z }
an an n
3 0 0
Tr d = 0 Tr σ = 0
y ha
= − p̄ Tr (d) + σ 0 : d0 .
le
liv or ec
M
.A
σ : d = − p − K Tr (d) Tr (d) + 2µ d0 : d0 .
d
(9.21)
uu
e
X Th
er
tin
on
.O
(9.22)
©
| {z } | {z }
recoverable power dissipative power
WR 2WD
3 The property that the trace of a deviator tensor is null is used here.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Constitutive Equation in Newtonian Fluids 449
Associated with the concepts of recoverable and dissipative powers, the re-
coverable and dissipative parts of the stress tensor, σ R and σ D , respectively, are
defined as
rs
WR = −p Tr (d) = −p 1 : d = σ R : d ,
ee
2W = K Tr2 (d) + 2µ d0 : d0 = σ : d ,
D D (9.25)
n
gi
σ : d = (σ
σ R + σ D ) : d = σ R : d + σ D : d = WR + 2WD .
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Remark 9.11. In an incompressible fluid, the recoverable power is
C d P cs
null. In effect, since the fluid is incompressible, dρ/dt = 0 , and
b
i
considering the mass continuity equation (5.24),
an an n
y ha
1 dρ
le
∇·v = − = 0 = Tr (d) =⇒ WR = −p Tr (d) = 0 .
liv or ec
ρ dt
M
.A
m
d
uu
e
X Th
er
tin
.O
C
dK dK
Z Z Z
Pe = + σ : d dV = + σ R : d dV + σ D : d dV
©
dt dt
V V V
dK
Z Z
Pe = + WR dV + 2WD dV ,
dt
V V
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
450 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
rs
B≡A B≡A B≡A
1 1
Z Z Z
ee
WR dt = σ R : d dt = dG = GB≡A − GA = 0 , (9.27)
ρ ρ
n
A A A
gi
which justifies the denomination of WR as recoverable power.
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
er
tin
2) The second law of thermodynamics allows proving that the dissipative power
on
.O
and, therefore, in a closed cycle the work performed per unit of mass by the
dissipative stresses will, in general, not be null,
ZB
1
σ D : d dt > 0 . (9.29)
ρ | {z }
A
2WD > 0
This justifies the denomination of 2WD as (non-recoverable) dissipative power.
The dissipative power is responsible for the dissipation (or loss of energy) phe-
nomenon in fluids.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Constitutive Equation in Newtonian Fluids 451
Solution
The recoverable power in an incompressible fluid is null (see Remark 9.11).
In addition, the dissipative power 2WD is known to be always non-negative
(see (9.28)). Finally, applying the balance of mechanical energy (see Re-
mark 9.12) results in
rs
dK
Z Z
ee
0 = Pe = + WR dV + 2WD dV =⇒
dt
n
|{z}
V =0 V
gi
t d le En
dK d 1 2
Z Z
= ρv dV = −
2 WD dV < 0
dt dt 2
ar
ar s
|{z}
ge ro or
V V >0
eS m
ib
ac
f
and, therefore, the fluid looses (dissipates) kinetic energy and the velocity of
C d P cs
its particles decreases.
b
i
an an n
y ha
le
9.4.5 Limitations in the Viscosity Values
liv or ec
M
.A
d
uu
. (9.30)
X Th
er
tin
.O
for all velocity fields v) that the fluid may possibly have. Therefore, it must be
©
verified for any arbitrary value of the strain rate tensor d = ∇S (v). Consider, in
particular, the following cases:
a) The strain rate tensor d is a spherical tensor.
In this case, from (9.30) results
Tr (d) 6= 0 ; d0 = 0 =⇒ 2WD = K Tr2 (d) ≥ 0 =⇒
2 (9.31)
K=λ+ µ ≥0
3
such that only the non-negative values of the bulk viscosity K are feasible.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
452 C HAPTER 9. C ONSTITUTIVE E QUATIONS IN F LUIDS
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
1 ∂ vx
0 0
y ha
vx (y) 2 ∂y
le
1 v 0
∂
liv or ec
v (x, y) = 0 ; d=
x
0 0 =d
2 ∂y
M
.A
0
0 0 0
m
d
uu
e
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 10
Fluid Mechanics
rs
n ee
gi
10.1 Governing Equations
t d le En
ar
A fluid is a particular case of continuous medium that is characterized by its
ar s
ge ro or
eS m
ib
specific set of constitutive equations. Consequently, the fluid mechanics problem
ac
is defined by the following equations:
f
C d P cs
a) Balance Equations
b
i
an an n
y ha
dρ
+ ρ∇ · v = 0 (1 equation) (10.1)
M
dt
.A
m
dv
er
tin
.O
3) Energy balance
C
du
ρ = σ : d+ρ r−∇·q (1 equation) (10.3)
dt
453
454 C HAPTER 10. F LUID M ECHANICS
b) Constitutive Equations
5) Thermo-mechanical constitutive equation
rs
ee
7) Law of heat conduction
n
gi
q = −K ∇θ (1 equation) (10.7)
t d le En
ar
ar s
c) Thermodynamic equations of state
ge ro or
eS m
ib
ac
8) Caloric equation of state f
C d P cs
b
i
u = u (ρ, θ ) (1 equation) (10.8)
an an n
y ha
le
liv or ec
.A
er
tin
ρ → 1 unknown
v → 3 unknowns
on
.O
σ → 6 unknowns
C
u → 1 unknown
©
→ 17 unknowns . (10.10)
q → 3 unknowns
θ → 1 unknown
s → 1 unknown
p → 1 unknown
1Note that the strain rate tensor d is not considered an unknown since it is an implicit function
of the velocity field v.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Governing Equations 455
of equations, denoted as the mechanical problem, may be posed and solved sep-
arately for a reduced number of unknowns (mechanical variables).
Consider the case of a barotropic fluid, which is characterized by the fact that
the temperature does not intervene in the kinetic equation of state. Then,
Kinetic equation F (ρ, p) = 0 =⇒ ρ = ρ (p) , (10.11)
of state
rs
of the thermodynamic pressure (see
ee
Figure 10.1). Assuming, in addition,
that the temperature does not inter-
n
vene in the thermo-mechanical consti-
gi
tutive equation (10.5), the governing
t d le En
equations of the (uncoupled) mechani-
Figure 10.1: Density depends on
ar
cal problem in a Newtonian fluid are de-
ar s
ge ro or
the thermodynamic pressure in a
eS m
ib
fined as
ac
barotropic fluid.
f
C d P cs
1) Mass continuity equation
b
i
an an n
y ha
dρ
+ ρ∇ · v = 0 (1 equation) (10.12)
le
dt
liv or ec
M
.A
2) Cauchy’s equation
m
dv
uu
dt
X Th
er
tin
.O
C
ρ → 1 unknown
v → 3 unknowns
→ 11 unknowns . (10.16)
σ → 6 unknowns
p → 1 unknown
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
456 C HAPTER 10. F LUID M ECHANICS
rs
spatial point being considered and is only a function of time. Therefore,
ee
1
n
d = ∇S v =(v ⊗ ∇ + ∇ ⊗ v) = 0 . (10.17)
gi
2
t d le En
Then, the constitutive equation (10.14) is reduced to
ar
ar s
σ = −p 1 + λ Tr (d) 1 + 2 µ d
ge ro or
=⇒ σ = −p1 , (10.18)
eS m
ib
|{z} |{z}
ac
=0 =0 f
C d P cs
b
which indicates that the stress state is hydrostatic (see Figure 10.2). In addi-
i
an an n
le
σ ) = −3 p̄ = −3p
Tr (σ =⇒ p̄ = p . (10.19)
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Hydrostatics. Fluids at Rest 457
dv ∂ v
a= = + v · ∇v = 0
dt ∂t hydrostatic (10.20)
case
σ = −p0 1 =⇒ p̄ = p = p0
rs
ee
c) Fluid at rest: v (x,t) ≡ const. = 0
n
gi
A particular case of hydrostatics is that of a fluid at rest with null velocity.
t d le En
ar
ar s
10.2.1 Hydrostatic Equations
ge ro or
eS m
ib
ac
f
The hydrostatic problem is governed by the following equations:
C d P cs
1) Constitutive equation
b
i
an an n
y ha
σ = −p0 1
(10.21)
le
liv or ec
σi j = −p0 δi j i, j ∈ {1, 2, 3}
M
.A
d
uu
e
er
tin
.O
cal structure of the stress tensor in (10.21), which ensures that all
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
458 C HAPTER 10. F LUID M ECHANICS
3) Cauchy’s equation
rs
ee
dv
∇·σ +ρ b = ρ (10.23)
n
dt
gi
Introducing (10.21) and (10.22) in (10.23),
t d le En
ar
ar s
∇ · σ = ∇ · (−p0 1) = −∇p0
ge ro or
eS m
ib
ac
∂ σi j ∂ ∂ p0
f
C d P cs
∇·σ j = = (−p0 δi j ) = − = − ∇p0 j j ∈ {1, 2, 3}
b
∂ xi ∂ xi ∂xj
i
an an n
(10.24)
y ha
le
liv or ec
M
.A
−∇p0 + ρ0 b = 0
Fundamental equation (10.25)
of hydrostatics
m
− ∂ p0 + ρ0 bi = 0 i ∈ {1, 2, 3}
d
uu
∂ xi
e
X Th
er
tin
on
.O
Consider the particular case, which is in fact very common, of the body forces
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Hydrostatics. Fluids at Rest 459
rs
Figure 10.4: Gravitational Force.
n ee
gi
∂ p0 (x, y, z)
t d le En
− =0 =⇒ p0 (x, y, z) ≡ p0 (y, z)
∂x
ar
∂ p0 (y, z)
ar s
ge ro or
− =0 =⇒ p0 (y, z) ≡ p0 (z) (10.27)
eS m
ib
∂y
ac
f
C d P cs
− ∂ p0 (z) − ρ0 g = 0
b
=⇒ p0 = −ρ0 g z +C
i
∂z
an an n
y ha
For a case such as the one shown in Figure 10.5, in which the surface pressure
le
(height z = h) is considered null, the solution (10.26) results in
liv or ec
M
.A
p0 = 0 ⇒ −ρ0 g h +C = 0 ⇒ C = ρ0 g h ⇒ p0 = ρ0 g (h − z) ,
z=h
m
(10.28)
uu
e
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
460 C HAPTER 10. F LUID M ECHANICS
rs
center of gravity of the volume of the displaced fluid.
n ee
gi
To prove Archimedes’ principle, consider the situations in Figure 10.6. On the
t d le En
one hand, Figure 10.6 a) illustrates a solid with volume V and density ρ in
the interior of a fluid of density ρ0 . The solid is not necessarily in equilibrium,
ar
ar s
ge ro or
even though its velocity and acceleration are assumed to be small enough to
eS m
ib
ensure a hydrostatic state in the fluid. On the other hand, Figure 10.6 b) shows
ac
f
C d P cs
the same fluid without the solid, such that the volume occupied by said solid in
b
Figure 10.6 a) is occupied here by an identical volume of fluid.
i
an an n
y ha
le
liv or ec
Volume of
M
.A
displaced fluid
m
d
uu
e
X Th
er
tin
on
.O
C
Figure 10.6: a) Solid submerged in a fluid and b) volume of the displaced fluid.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Hydrostatics. Fluids at Rest 461
p0 (z) = ρ0 g (h − z)
(10.29)
σ = −p0 1
Note that the hydrostatic pressure and the stress state in the fluid are the same
for equivalent points of the fluid in the cases a) and b) of Figure 10.6.
rs
t = σ · n = −p0 1 · n = −p0 n (10.30)
n ee
and the resultant R of the forces the fluid exerts on the solid is
gi
Z Z
t d le En
R= t dS = −p0 n dS . (10.31)
ar
ar s
∂V ∂V
ge ro or
eS m
ib
Note now that, since the hydrostatic pressure distribution is the same in both
ac
f
C d P cs
cases of Figure 10.6, this resultant is the same as the one obtained in case b)
b
for the forces that the rest of the fluid exerts on the volume of displaced fluid,
i
an an n
with the particularity that, because the pressure distribution is constant in space
y ha
on (10.30), resulting in
M
.A
Z Z
R= −p0 n dS = −∇p0 dV . (10.32)
m
∂V V
uu
e
er
tin
Z Z Z
R= −∇p0 dV = −p0 b dV = − ρ0 b dV = W êz = E êz , (10.33)
on
.O
V V V
C
| {z }
©
W êz
where E is the upward buoyant force acting on the submerged solid and W is the
weight of the displaced fluid (see Figure 10.6 b) ). That is,
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
462 C HAPTER 10. F LUID M ECHANICS
Volume of
displaced fluid
rs
n ee
gi
Figure 10.7: Forces acting on the volume of displaced fluid.
t d le En
ar
ar s
ge ro or
eS m
ib
c) Vertical line of application of the upward buoyant force
ac
Consider now the moment MG f
C d P cs
E of the upward buoyant force E with respect to
b
the center of gravity, G, of the volume of displaced fluid (see Figure 10.72 ),
i
an an n
y ha
Divergence
le
TheoremZ
liv or ec
Z Z
G
M E = x × (−p 0 n) dS = x × (−p 0 ∇) dV = − x × ∇p0 dV
M
.A
V V
∂V
m
Z Z
∂
d
G
ME i = − ei jk x j p0 nk dS = − ei jk x j p0 dV =
uu
∂ xk
e
X Th
∂V V
er
tin
∂xj ∂ p0 ∂ p0
Z Z Z
= − e i jk p0 dV − e x
i jk j dV = − ei jk x j dV
on
∂ xk ∂ xk ∂ xk
.O
V | {z } V V
i ∈ {1, 2, 3}
C
ei jk δ jk =
©
ei j j = 0 (10.35)
and replacing the fundamental equation of hydrostatics (10.25) in (10.35) finally
yields
Z Z
MG
E =− (x × ∇p0 ) dV = − G
(x × ρ0 b) dV = −MW =0, (10.36)
V V
| {z }
G
MW
2 Without loss of generality, the origin of the system of Cartesian axes can be placed at G.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Hydrostatics. Fluids at Rest 463
where MW G is the moment of the weight of the displaced fluid with respect to its
center of gravity G, which, considering the definition of the center of gravity,
is null. Consequently, the moment of the upward buoyant force E with respect
to the center of gravity of the volume of displaced fluid is also null. Then, it
is concluded that the vertical line of application of the upward buoyant force
crosses said center of gravity, as established by the second part of Archimedes’
principle.
rs
centers of gravity of the solid and the corresponding volume of displaced
ee
fluid affect the nature of this equilibrium.
n
gi
Solution
t d le En
Consider a floating medium, in equilibrium, and the following two situations:
ar
ar s
a) The center of gravity of the solid (center of thrust) is below the center of
ge ro or
eS m
ib
gravity of the displaced fluid (center of buoyancy).
ac
f
C d P cs
In this case, any perturbation (inclination) tends to create a moment M = W d
b
i
in the sense that tends to recover the initial state of equilibrium. It is, thus, a
an an n
center of gravity
le
recovering
liv or ec
of the displaced
fluid moment
M
.A
perturbation
m
d
uu
e
X Th
er
tin
center of gravity
on
.O
of the solid
C
b) The center of gravity of the solid (center of thrust) is above the center of
gravity of the displaced fluid (center of buoyancy).
In this case, any perturbation (inclination) tends to create a moment M = W d
in the sense that tends to capsize the floating solid, that is, it tends to move
the solid further away from the initial state of equilibrium. It is, thus, a case
of unstable flotation equilibrium.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
464 C HAPTER 10. F LUID M ECHANICS
center of gravity
of the displaced fluid
rs
ee
Placing weights (ballasts) on the keel of a boat responds to the search of
n
improved flotation stability of this boat.
gi
t d le En
ar
ar s
ge ro or
eS m
ib
10.3 Fluid Dynamics: Barotropic Perfect Fluids
ac
f
C d P cs
In the most common case, the velocity is not uniform nor stationary (v ≡ v (x,t)),
b
and, therefore, in general, the acceleration will not be null (a (x,t) 6= 0). In con-
i
an an n
sequence,
the divergence of the velocity (∇ · v 6= 0) and the gradient of the ve-
y ha
not
locity ∇ ⊗ v = ∇v 6= 0 will not be null either.
le
liv or ec
M
.A
by the fact that the viscosities λ and µ (see (10.14)) are null.
uu
e
X Th
er
tin
.O
C
)
σ = −p 1 + λ Tr (d) + 2µ d
©
=⇒ σ = −p 1
λ =µ =0
(10.37)
(
∇ · σ = −∇ p
=⇒
σ : d = −p 1 : d = −p Tr (d)
3 A hydrostatic stress state (the stress tensor is spherical) should not be confused with a
hydrostatic motion regime (the velocity is uniform or null).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: Barotropic Perfect Fluids 465
rs
Taking into account the hypotheses of a perfect and a barotropic fluid, the equa-
ee
tions governing a fluid dynamics problem are reduced to:
n
gi
a) Mechanical problem
t d le En
1) Mass continuity equation
ar
ar s
ge ro or
eS m
ib
dρ
ac
+ ρ∇ · v = 0 (1 equation) (10.38)
dt f
C d P cs
b
i
an an n
le
dv
liv or ec
.A
m
.O
p (x,t) that can be solved uncoupled from the thermal problem.
C
b) Thermal problem
1) Fourier’s law
q = −K ∇θ ⇒ ∇ · q = −K ∇ · (∇θ ) = −K ∇2 θ (3 equations)
(10.41)
2) Energy balance
du
ρ = −p∇ · v +ρ r + K∇2 θ (1 equation)
dt | {z } | {z } (10.42)
σ :d −∇ · q
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
466 C HAPTER 10. F LUID M ECHANICS
The thermal
problem is defined by 5 equations and 5 unknowns q (x,t), θ (x,t),
u (x,t) and can be solved once the mechanical problem has been solved and the
velocity field v (x,t), the density ρ (x,t) and the pressure p (x,t) are known.
rs
cludes the thermal conductivity K between the viscosities (in a gen-
ee
eralized sense) of the problem. The definition of a perfect fluid as
a fluid without viscosity results, in this context, in the cancella-
n
gi
tion of the thermal conductivity (K = 0), therefore (10.41) leads to
q = −K∇θ = 0 and the thermal problem is reduced to the equa-
t d le En
tions (10.42) and (10.43).
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
10.3.2 Resolution of the Mechanical Problem under Potential Body
an an n
y ha
Consider now the mechanical problem for the particular case of potential body
forces (the body forces derive from a potential φ ),
M
.A
m
In the particular case of a gravitational potential with the line of action along the
er
tin
.O
C
0
©
not
φ (x, y, z,t) = g z =⇒ b = −∇φ ≡ 0
. (10.45)
−g
∇p = ρ ∇P .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: Barotropic Perfect Fluids 467
Proof
Defining the function P (x,t) as
Zp
1
P (x,t) = P̂ (p (x,t)) = d p̄ , (10.46)
ρ ( p̄)
0
then, it will satisfy
∂ P (x,t) ∂ P̂ ∂ p
=
∂ xi ∂ p ∂ xi (10.47)
1
rs
∂ P̂
i ∈ {1, 2, 3}
∇P i =
∇p i = ∇p i
∂p ρ (p)
n ee
leading to
gi
1
t d le En
∇P = ∇p . (10.48)
ρ (p)
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
Lemma 10.2. The convective term of the acceleration can be written
i
an an n
as
y ha
1 2
v · ∇v = 2ω
ω ×v+∇ v ,
le
2
liv or ec
M
.A
d
uu
e
Proof
X Th
er
.O
∂vj ∂ v j ∂ vi ∂ vi ∂ vi
v · ∇v j = vi = vi − +vi = 2 vi w ji +vi =
∂ xi ∂ xi ∂ x j ∂xj ∂xj
C
|{z}
−wi j
©
| {z }
2 w ji
∂ vi ∂ vi
= −2 vi wi j + vi = 2 ei jk vi ωk +vi =
∂xj | {z } ∂xj (10.49)
e jkivi ω k
∂ 1 1 2
= 2 e jki vi ωk + vi vi = 2ω ω ×v j + ∇ v ,
| {z } ∂ x j 2 |{z} 2 j
ω × v] j
[2ω v · v = v2
j ∈ {1, 2, 3}
4The following results,
previously obtained in Chapter 2, are used here:
w ji = −wi j = ∇a v ji = (∂ v j /∂ xi − ∂ vi /∂ x j ) /2 , wi j = −ei jk ωk and v2 = |v|2 = v · v.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
468 C HAPTER 10. F LUID M ECHANICS
which leads to
1 2
v · ∇v = 2ω
ω ×v+∇ v . (10.50)
2
rs
ee
dv ∂ v ∂v 1 2
− ∇P − ∇φ = = + v · ∇v = ω ×v+∇
+ 2ω v , (10.52)
n
dt ∂t ∂t 2
gi
t d le En
where the result (10.50) has been taken into account. Finally, (10.52) is rewritten
as
ar
ar s
1 2 ∂v
ge ro or
eS m
ib
− ∇P + ∇φ + ∇ v = + 2ωω ×v . (10.53)
2
ac
∂t
f
C d P cs
b
i
an an n
1 2 ∂v (10.54)
−∇ P + φ + v = + 2ωω ×v
M
.A
2 ∂t
| {z }
m
Bernoulli’s trinomial
d
uu
e
X Th
er
tin
Equation (10.54) is the particular form adopted by the balance of linear momen-
tum (Euler’s equation (10.39)) in barotropic perfect fluids subjected to potential
on
.O
body forces.
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: Barotropic Perfect Fluids 469
rs
ee
and a streamline5 Γ : x = x (s) parametrized in terms of its arc-length s (see
n
Figure 10.9). Projecting (multiplying) equation (10.53) in the direction tangent
gi
to the streamline, t, results in
t d le En
1 2
ar
ar s
− ∇ P + φ + v = 2ω ω × v =⇒ − (∇M) · v = (2ω ω × v) · v = 0 (10.56)
ge ro or
eS m
ib
2 |{z} | {z }
ac
| {z } dx ds f =0
C d P cs
M (x)
b
ds dt
i
an an n
y ha
dx dM
∇M (x (s)) ·
le
= = 0
liv or ec
ds ds ∀x ∈ Γ =⇒ M (x) = const. ∀x ∈ Γ
M
.A
∂ M (x (s)) dxi dM
= =0
m
∂ xi ds ds
d
(10.57)
uu
e
er
tin
1 2
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
470 C HAPTER 10. F LUID M ECHANICS
rs
Figure 10.9: Parametrized streamline.
n ee
gi
10.3.3.1 Solution in Steady-State Regime for an Incompressible Fluid under
t d le En
Gravitational Forces
Consider now the particular case of a barotropic fluid with the following char-
ar
ar s
ge ro or
acteristics:
eS m
ib
ac
a) The fluid is incompressible f
C d P cs
b
i
ρ = ρ (p) = ρ0 = const. (10.59)
an an n
y ha
Zp Zp
1 1 p
M
.A
P (x,t) = d p̄ = d p̄ = (10.60)
ρ ( p̄) ρ0 ρ0
m
0 0
d
uu
e
X Th
.O
C
0
©
not
φ = gz b = −∇φ ≡ 0 (10.61)
−g
Introducing (10.60) and (10.61) in Bernouilli’s trinomial (10.58) yields
p 1 p 1 v2 de f
+ g z + v2 = const. =⇒ z + + = H = const. ∀x ∈ Γ
ρ0 2 ρ0 g 2 g
(10.62)
The terms in (10.61) have dimensions of length (height) and may be inter-
preted in the following manner.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: Barotropic Perfect Fluids 471
Bernoulli’s theorem
p 1 v2 de f (10.63)
z + + = H = const. ∀x ∈ Γ
ρ0 g 2 g
|{z} |{z} |{z} |{z}
elevation pressure velocity total
head head height
rs
ee
Remark 10.4. The expression in (10.63) constitutes the so-called
Bernoulli’s theorem (for an incompressible perfect fluid under
n
gravitational forces and in steady-state regime), which establishes
gi
that the sum of the elevation, the pressure head and the velocity head
t d le En
is constant in every point belonging to a same streamline (see Fig-
ar
ure 10.10).
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
named hydraulics. Since, in general, the body forces are of the gravi-
M
.A
d
uu
e
X Th
er
tin
on
.O
C
velocity head
pressure head
elevation
streamline
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
472 C HAPTER 10. F LUID M ECHANICS
Example 10.2 – Determine the velocity of the water exiting the tank through
a small lateral hole placed at a distance h below the top surface of the water.
Consider the top of the tank is open and neglect the atmospheric pressure.
Assume a steady-state regime.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
Solution f
C d P cs
b
i
The fluid in this problem (water) is an incompressible perfect fluid in steady-
an an n
state regime under gravitational forces and, thus, Bernoulli’s theorem can be
y ha
applied.
le
liv or ec
.A
at point B of the exit hole (shown in the figure above). Applying Bernoulli’s
theorem between points A and B, and taking into account that the velocity
m
of the free surface in the tank is practically null and that its cross-section is
d
uu
er
tin
pA 1 v2 pB 1 v2B
zA + + A = zB + +
|{z} ρ0 g 2 g |{z} ρ0 g 2 g
on
.O
1 v2 √
h+0+0 = 0+0+ =⇒ v= 2gh .
2 g
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: Barotropic Perfect Fluids 473
1 2 ∂v
−∇ P + φ + v = ω ×v .
+ 2ω (10.64)
2 ∂t
rs
• with potential body forces
ee
• and irrotational flow.
n
gi
t d le En
Definition 10.4. The motion (or flow) of a fluid is said to be irrota-
tional (or potential) if the rotational of the velocity field is null at any
ar
ar s
ge ro or
point of this fluid.
eS m
ib
ac
f
C d P cs
b
i
In other words, an irrotational flow has a null vorticity vector.
an an n
y ha
le
liv or ec
∇ × v (x,t) = 0
Irrotational flow ∀x ∀t (10.65)
M
ω (x,t) = 1 ∇ × v (x,t) = 0
.A
2
m
d
uu
If the flow is irrotational, it is inferred from (10.65) that there exists a scalar
e
X Th
.O
C
Note that, in this case, the vector field v (x,t) is determined in terms of the scalar
©
velocity potential χ (x,t) (which becomes the main unknown of the problem).
Replacing the conditions (10.65) and (10.66) in (10.64) yields
1 ∂v ∂v ∂ ∂χ
− ∇ P + φ + v2 = + 2 ω ×v = = ∇χ (x,t) = ∇ =⇒
2 ∂t |{z} ∂t ∂t ∂t
=0
(10.67)
6 It can be proven that, given an irrotational vector field v (x,t), that is, a vector field
that satisfies ∇ × v = 0, there exists a scalar function χ (x,t) (potential function) such that
v = ∇χ (x,t). Obviously, since ∇ × ∇ (•) ≡ 0, then ∇ × v = ∇ × ∇χ (x,t) = 0 is satisfied.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
474 C HAPTER 10. F LUID M ECHANICS
1 2 ∂χ
∇ P+φ + v + = ∇M (x,t) = 0
2 ∂t
| {z }
M (x,t) ∀x ∀t (10.68)
∂ M (x,t)
=0 i ∈ {1, 2, 3}
∂ xi
This equation can be trivially integrated, resulting in
1 ∂χ
M (x,t) = P + φ + v2 +
rs
= ϕ (t) . (10.69)
2 ∂t
ee
Defining a modified velocity potential χ̄ (x,t) of the form
n
gi
t d le En
Zt ∇χ̄ = ∇χ = v (x,t)
de f
χ̄ (x,t) = χ (x,t) − ϕ (τ) dτ ⇒ (10.70)
ar
∂ χ̄ ∂χ
ar s
− ϕ (t)
ge ro or
=
eS m
ib
0
∂t ∂t
ac
f
C d P cs
and replacing (10.70) in (10.69) produces
b
i
an an n
1 ∂χ 1 ∂ χ̄
y ha
P + φ + v2 + − ϕ (t) = 0 =⇒ P + φ + (∇χ̄)2 + = 0 ∀x ∀t ,
2 2
le
|∂t {z } ∂t
liv or ec
∂ χ̄
M
.A
∂t
(10.71)
m
er
tin
.O
dρ dρ dρ
+ ρ∇ · v = + ρ ∇ · (∇χ̄) = 0 =⇒ + ρ∇2 χ̄ = 0 (10.72)
C
dt dt dt
©
| {z }
∇2 χ̄
1 ∂ χ̄
P (ρ, p) + φ + (∇χ̄)2 + =0 ∀x ∀t (10.73)
2 ∂t
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: Barotropic Perfect Fluids 475
rs
• with potential body forces,
ee
• irrotational flow
• and incompressible flow.
n
gi
Since the flow is incompressible, (10.46) and (10.72) allow determining7
t d le En
Zp
ar
1 p
ar s
ge ro or
P (p) =
d p̄ =
eS m
ib
dρ ρ ( p̄) ρ0
= 0 =⇒ ρ = ρ0 =⇒ (10.76)
ac
dt 0
f
C d P cs
not
∇2 χ̄ = ∆ χ̄ = 0
b
i
an an n
y ha
.A
∂ 2 χ̄
(10.77)
m
∆ χ̄ = =0
d
∂ xi ∂ xi
uu
e
X Th
er
tin
.O
p 1 ∂ χ̄
+ φ + (∇χ̄)2 + =0 ∀x ∀t (10.78)
C
ρ0 2 ∂t
©
7 Here, the differential operator named Laplace operator or Laplacian of (•) is defined as
not
∆ (•) = ∇ · ∇ (•) = ∇2 (•) = ∂ 2 (•)/∂ xi ∂ xi .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
476 C HAPTER 10. F LUID M ECHANICS
dρ
+ ρ∇ · v = 0 Mass continuity equation (1 eqn.)
dt
dv
∇ · σ + ρb = ρ Balance of linear momentum (3 eqns.)
dt
du
= σ : d+ρ r−∇·q Energy balance (1 eqn.)
rs
ρ
dt
ee
σ = −p1 + λ Tr (d) 1 + 2µ d Mechanical constitutive equation
n
(6 eqns.)
gi
t d le En
s = s (d, θ , ρ) Entropy constitutive equation (1 eqn.)
ar
ar s
ge ro or
eS m
ib
q = −K∇θ Heat conduction equation (3 eqns.)
ac
f
C d P cs
b
i
u = u (ρ, θ ) Caloric equation of state (1 eqn.)
an an n
y ha
le
liv or ec
.A
(10.79)
which constitute a system of 17 equations and 17 unknowns. This system is too
m
er
tin
.O
of (10.79) expressed solely in terms of the velocity field v (x,t) and the pres-
©
sure p (x,t).
Lemma 10.3. The divergence of the strain rate tensor d (x,t) is re-
lated to the velocity field v (x,t) by
1 1
∇ · d = ∆ v + ∇ (∇ · v) .
2 2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: (Newtonian) Viscous Fluids 477
Proof
1 ∂ 2 vi 1 ∂ 2v j
∂ ∂ 1 ∂ vi ∂ v j
∇·d j = di j = + = + =
∂ xi ∂ xi 2 ∂ x j ∂ xi 2 ∂ xi ∂ x j 2 ∂ xi ∂ xi
1 ∂ ∂ vi 1 ∂ 2 v j 1 ∂ 1
= + = (∇ · v) + ∆ v j =
2 ∂ x j ∂ xi 2 ∂ xi ∂ xi 2 ∂ x j 2 |{z}
[∆ v] j
|{z} | {z } | {z }
∇·v ∆vj [∇ (∇ · v)] j
1 1
= ∆ v + ∇ (∇ · v) j ∈ {1, 2, 3}
rs
2 2 j
ee
(10.80)
n
1 1
gi
∇ · d = ∆ v + ∇ (∇ · v) (10.81)
2 2
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
Lemma 10.4. Given a scalar function α (x,t), the following is sat-
C d P cs
b
isfied.
i
∇ · (α 1) = ∇α
an an n
y ha
le
liv or ec
Proof
M
.A
∂ (αδi j ) ∂α ∂α
m
∇ · (α 1) i = = δi j = = ∇α i i ∈ {1, 2, 3} (10.82)
d
∂xj ∂xj ∂ xi
uu
e
X Th
er
tin
∇ · (α 1) = ∇α (10.83)
on
.O
C
linear momentum of (10.79), and taking into account (10.81) and (10.83) leads
to
σ = −p1 + λ Tr (d) 1 + 2µ d
dv =⇒
∇ · σ + ρb = ρ
dt
∇ · σ = −∇p + λ ∇ (Tr (d)) +µ∆ v + µ∇ (∇ · v) (10.84)
| {z }
∇ (∇ · v)
dv
∇ · σ + ρb = −∇p + (λ + µ) ∇ (∇ · v) + µ∆ v + ρb = ρ
dt
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
478 C HAPTER 10. F LUID M ECHANICS
Navier-Stokes equation
dv
−∇p + (λ + µ) ∇ (∇ · v) + µ∆ v + ρb = ρ
dt (10.85)
∂p ∂ 2v
j ∂ 2v i dvi
− + (λ + µ) +µ + ρbi = ρ ; i ∈ {1, 2, 3}
∂ xi ∂ xi ∂ x j ∂ x j∂ x j dt
rs
n ee
10.4.2 Energy Equation
gi
The aim is to eliminate σ and q from the energy balance of (10.79) by replacing
t d le En
in this equation the mechanical constitutive equation and the entropy equation
ar
of (10.79). To this aim, the definition of stress power in a Newtonian fluid (see
ar s
ge ro or
eS m
ib
Chapter 9) is recovered,
ac
f
C d P cs
σ : d = WR + 2WD = −p ∇ · v + K Tr2 (d) + 2µ d0 : d0 ,
b
(10.86)
i
an an n
where d0 is the deviatoric part of the strain rate tensor. Fourier’s law is also
y ha
recovered,
le
liv or ec
q = −K ∇θ =⇒ ∇ · q = −∇ · (K ∇θ ) . (10.87)
M
.A
du
d
ρ
e
dt
X Th
er
tin
Energy equation
on
.O
du
C
dt | {z } (10.89)
2WD
∂ vi 2
du ∂ vi ∂ ∂θ
ρ = −p +ρ r+ K +K + 2µ di0 j di0 j
dt ∂ xi ∂ xi ∂ xi ∂ xi
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: (Newtonian) Viscous Fluids 479
rs
−∇p + (λ + µ) ∇ (∇ · v) + Balance of linear momentum
(3 eqns.)
ee
dv (Navier-Stokes)
+µ∆ v + ρb = ρ
n
dt
gi
du
ρ = −p∇ · v + ρ r + ∇ · (K ∇θ ) +
t d le En
dt Energy balance (1 eqn.)
+K Tr2 (d) + 2µ d0 : d0
ar
ar s
ge ro or
eS m
ib
ac
u = u (ρ, θ ) f
Caloric equation of state (1 eqn.)
C d P cs
b
i
an an n
y ha
(10.90)
M
.A
can be uncoupled from the thermal part in the set of equations of (10.79), result-
uu
e
er
tin
dρ
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
480 C HAPTER 10. F LUID M ECHANICS
rs
ee
can be interpreted as a component of the system of forces (per unit of volume)
n
that acts on a volume differential of the fluid in motion as follows.
gi
t d le En
−∇p − − (λ + µ) ∇ (∇ · v) + µ∆ v + ρb + ρa = 0
ar
ar s
ge ro or
| {z } } |{z} |{z} (10.93)
eS m
ib
| {z
forces due to viscous forces exerted by body inertial
ac
the pressure forces forces
gradient f
the contact between particles
C d P cs
(= 0 when λ = µ = 0)
b
i
an an n
direction.
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: (Newtonian) Viscous Fluids 481
Each of the terms in the energy equation (10.89) can also be given a physical
interpretation, as indicated in Table 10.1.
du
ρ = −p∇ · v + ρ r + ∇ · (k ∇θ ) + K Tr2 (d) + 2µ d0 : d0
dt | {z }
2WD
rs
dt unit of volume and unit of time
ee
d (dV )
variation of volume
n
∇ · v = dt =
gi
dV unit of volume and unit of time
t d le En
d (dV ) mechanical work of the thermodynamic pressure
p =
ar
−p∇ · v = − dt
ar s
unit of volume and unit of time
ge ro or
dV
eS m
ib
(see Figure 10.12 and footnote8 )
ac
f
C d P cs
heat generated by the internal sources and conduction
ρ r + ∇ · (k ∇θ ) =
b
i
unit of volume and unit of time
an an n
y ha
.A
d
uu
e
X Th
er
tin
on
.O
C
8 Here, the relation d (dV )/dt = (∇ · v) dV is used (see Section 2.14.3 in Chapter 2).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
482 C HAPTER 10. F LUID M ECHANICS
rs
dρ ∇ · v = Tr (d) = 0
+ ρ∇ · v = 0
ee
dt
n
gi
and introducing (10.94) in (10.90) results in the governing equations detailed in
t d le En
Table 10.2.
ar
ar s
ge ro or
eS m
ib
Mass continuity equation ∇·v = 0
ac
Mechanical
f
C d P cs
Problem
b dv
i
Navier-Stokes equation −∇p + µ ∆ v + ρ0 b = ρ0
an an n
dt
y ha
du
le
Energy balance = ρ0 r + ∇ · (K ∇θ ) + 2µ d0 : d0
liv or ec
ρ0
Thermal dt
M
Problem
.A
d
uu
Constitutive equation σ = −p 1 + 2µ d
e
X Th
er
tin
on
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Dynamics: (Newtonian) Viscous Fluids 483
dρ
Mass continuity equation + ρ∇ · v = 0
dt
1 dv
Navier-Stokes equation −∇p + µ∇ (∇ · v) + µ∆ v + ρb = ρ
3 dt
du
Energy balance ρ = −p∇ · v + ρ r + ∇ · (K ∇θ ) + 2µ d0 : d0
dt
rs
ee
Kinetic equation of state F (ρ, p, θ ) = 0
n
gi
2
Constitutive equation σ = −p 1 − µ Tr (d) 1 + 2µ d
t d le En
3
ar
ar s
ge ro or
eS m
ib
Table 10.3: Governing equations in Stokes fluids.
ac
f
C d P cs
b
i
an an n
.A
Table 10.4.
m
d
uu
dρ
e
dt
er
tin
dv
on
.O
du
©
Constitutive equation σ = −p 1
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
484 C HAPTER 10. F LUID M ECHANICS
10.4.5.4 Hydrostatics
In this case, the following conditions apply (see (10.20)):
dv
a= = 0 , ∇ · v = 0 , ρ = ρ0 , p = p0 and σ = −p0 1 , (10.97)
dt
and, thus, (10.90) is reduced to the equations described in Table 10.5.
rs
Problem
ee
du
n
Energy balance ρ0 = ρ0 r + ∇ · (K ∇θ )
Thermal dt
gi
Problem
t d le En
Caloric equation of state u = u (ρ0 , θ )
ar
ar s
ge ro or
eS m
ib
Constitutive equation σ = −p0 1
ac
f
C d P cs
b
i
Table 10.5: Governing equations in hydrostatics.
an an n
y ha
le
10.5 Boundary Conditions in Fluid Mechanics
liv or ec
M
.A
The governing equations of the fluid mechanics problem presented in the previ-
ous sections require adequate boundary conditions to be solved correctly. In gen-
m
eral, the spatial (or Eulerian) description is used in fluid mechanics problems,
uu
er
the aforementioned spatial boundary conditions are applied. Even though there
tin
are different boundary conditions, and these often depend on the type of problem
on
being studied, the most common types of boundary conditions are summarized
.O
below.
C
b) Impermeability condition
Usually, part of the boundary of the control volume V is composed of imper-
meable walls, Γvn , which are assumed to be impervious to fluid, that is, they
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Boundary Conditions in Fluid Mechanics 485
rs
Figure 10.13: Velocity boundary conditions: prescribed velocity.
n ee
cannot be penetrated by said fluid. The mathematical expression of this con-
gi
dition is denoted as impermeability condition and it establishes that the rela-
t d le En
tive velocity of the fluid, vr , with respect to the impermeable wall (assumed
mobile and with a velocity v∗ ) in the direction normal to the boundary must
ar
ar s
ge ro or
eS m
ib
be null (see Figure10.14),
ac
vn (x,t) = v · n = v∗ · n f
C d P cs
∀x ∈ Γvn =⇒
b
i
|{z} | {z }
an an n
fluid wall
(10.99)
y ha
le
vr · n = (v − v∗ ) · n = 0 ∀x ∈ Γvn .
liv or ec
M
.A
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
486 C HAPTER 10. F LUID M ECHANICS
c) Adherence condition
rs
In viscous fluids in contact with an impermeable wall, due to the effect of
ee
viscosity, the fluid is assumed to adhere to the wall (see Figure 10.15) and,
n
thus, the relative velocity between the fluid and the wall vr is null.
gi
t d le En
vr (x,t) = v − v∗ = 0 ∀x ∈ Γv =⇒ v = v∗ ∀x ∈ Γv (10.100)
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
.O
Under certain circumstances, only a part of the traction vector such as the ther-
modynamic pressure is prescribed. In effect, for a Newtonian fluid,
σ = −p1 + λ Tr (d) 1 + 2µ d =⇒
(10.102)
t = σ · n = −p n + λ Tr (d) n + 2µ d · n ,
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Boundary Conditions in Fluid Mechanics 487
which exposes how the thermodynamic pressure p is a part of the normal com-
rs
ponent of the traction vector t. The prescription of the thermodynamic pressure
ee
on a part of the boundary Γp is written as
n
gi
p (x,t) = p∗ (x,t) ∀x ∈ Γp . (10.103)
t d le En
ar
ar s
ge ro or
eS m
ib
10.5.3 Mixed Boundary Conditions
ac
f
C d P cs
In certain cases (such as the entrance and exit sections of pipes) the pressure (a
b
part of the normal component of the traction vector) and the tangential com-
i
an an n
ponents of the velocity (which are assumed to be null, see Figure 10.17) are
y ha
prescribed.
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
488 C HAPTER 10. F LUID M ECHANICS
free surface
rs
ee
Examples of free surface9 are the surface of the sea (see Figure 10.18) or the
n
gi
surface that separates the saturated and unsaturated parts of an embankment dam
(see Figure 10.19).
t d le En
A hypothesis with a clear physical sense that is frequently used in relation
ar
to a free surface is that such a surface is a material surface (constituted always
ar s
ge ro or
eS m
ib
by the same particles). This hypothesis implicitly establishes certain boundary
ac
f
conditions on the velocity field in the material surface Γf s . In effect, considering
C d P cs
b
the free surface in Figure 10.18,
i
an an n
le
liv or ec
and imposing the material character of the free surface (null material derivative,
see Section 1.11 in Chapter 1),
M
.A
dφ
m
∂φ ∂η ∂η ∂η ∂φ
= + v · ∇φ = − − vx − vy + vz =0, (10.105)
d
dt ∂x ∂y ∂z
uu
∂t ∂t
e
|{z}
X Th
=1
er
tin
on
.O
∂η ∂η ∂η
vz (x,t) = + vx + vy ∀x ∈ Γf s . (10.106)
∂t ∂x ∂y
C
This condition establishes the dependency of the vertical component of the ve-
locity vz on the other components vx and vy .
Another boundary condition frequently imposed on free surfaces is that, in
these surfaces, the thermodynamic pressure is known and equal to the atmo-
spheric pressure10 ,
p (x,t) = Patm ∀x ∈ Γf s (10.107)
9 In general, in fluid mechanics problems in which free surfaces appear, the position of these
surfaces is not known and their geometrical characteristics become an unknown of the prob-
lem.
10 The value of the atmospheric pressure is generally neglected (P
atm ≈ 0).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Laminar and Turbulent Flows 489
saturated
rs
n ee
Figure 10.19: Free surface of an embankment dam.
gi
t d le En
Equation (10.107) allows identifying, in certain cases, the position of the free
ar
ar s
ge ro or
surface (once the pressure field is known) as the locus of points in the fluid in
eS m
ib
which the pressure is equal to the atmospheric pressure.
ac
f
C d P cs
b
i
an an n
free surface
le
liv or ec
M
.A
d
uu
e
er
tin
.O
flow (or regime). Basically, laminar flow is physically characterized by the fact
C
that the fluid moves in parallel layers that do not mix (see Figure 10.20).
©
Vortex
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
490 C HAPTER 10. F LUID M ECHANICS
de f V ×L
Reynolds number: Re =
ν
V = characteristic velocity of the fluid
(10.109)
L = characteristic length of the domain
ν = kinematic viscosity (ν = µ/ρ)
rs
such that small values of the Reynolds number characterize laminar flows.
ee
10.6.2 Turbulent Flow
n
gi
When the velocity increases and the viscosity decreases, the Reynolds num-
t d le En
ber (10.109) increases. For increasing values of this number, the initially lam-
inar flow is seen to distort and become highly unstable. The flow can then be
ar
ar s
ge ro or
understood as being in a situation in which the velocity v (x,t), at a given point
eS m
ib
in space, randomly and rapidly fluctuates along time about a mean value v̄ (x,t)
ac
f
C d P cs
(see Figure 10.21). This situation is defined as turbulent flow (or regime).
b
Even though the equations of the fluid mechanics problem in general, and the
i
an an n
.A
the so-called turbulence models. These models are based on isolating the mean
d
values of the velocity and pressure fields from their fluctuations and, then, the
uu
e
governing equations of the problem are obtained in terms of these mean values.
X Th
er
tin
on
.O
C
Figure 10.21: Variation of the velocity along time in laminar and turbulent flows.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Mechanics Formulas 491
Cartesian coordinates
∂ vx ∂ vx ∂ vy
σx = 2µ −p τxy = τyx = µ +
∂x ∂y ∂x
∂ vy ∂ vy ∂ vz
rs
σy = 2µ −p τyz = τzy = µ + (10.110)
ee
∂y ∂z ∂y
n
∂ vz ∂ vz ∂ vx
σz = 2µ −p τzx = τxz = µ +
gi
∂z ∂x ∂z
t d le En
ar
Cylindrical coordinates
ar s
ge ro or
eS m
ib
ac
f
∂ vr ∂ vθ 1 ∂ vr
−p
C d P cs
σr = 2µ τrθ = τθ r = µ r +
∂r
b ∂r r r ∂θ
i
an an n
1 ∂ vθ vr ∂ vθ 1 ∂ vz
y ha
σθ = 2µ + −p τθ z = τzθ = µ +
r ∂θ r ∂z r ∂θ
le
liv or ec
∂ vz ∂ vz ∂ vr
σz = 2µ −p τzr = τrz = µ +
M
.A
∂z ∂r ∂z
(10.111)
m
1 ∂ 1 ∂ vθ ∂ vz
d
r ∂r r ∂θ ∂z
X Th
er
tin
Spherical coordinates
on
.O
∂ vr
C
σr = 2µ −p
∂r
©
1 ∂ vθ vr
σθ = 2µ + −p (10.113)
r ∂θ r
1 ∂ vφ vr vθ cot θ
σφ = 2µ + + −p
r sin θ ∂ φ r r
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
492 C HAPTER 10. F LUID M ECHANICS
∂ vθ 1 ∂ vr
τrθ = τθ r = µ r +
∂r r r ∂θ
sin θ ∂ vφ 1 ∂ vθ
τθ φ = τφ θ = µ + (10.113 (cont.))
r ∂ θ sin θ r sin θ ∂ φ
1 ∂ vr ∂ vφ
τφ r = τrφ = µ +r
r sin θ ∂ φ ∂r r
1 ∂ 2 1 ∂ 1 ∂ vφ
∇·v = 2
r vr + (vθ sin θ ) + (10.114)
r ∂r r sin θ ∂ θ r sin θ ∂ φ
rs
n ee
10.7.2 Continuity Equation
gi
Cartesian coordinates
t d le En
ar
∂ρ ∂ ∂ ∂
ar s
+ (ρvx ) + (ρvy ) + (ρvz ) = 0
ge ro or
(10.115)
eS m
ib
∂t ∂ x ∂y ∂z
ac
f
C d P cs
Cylindrical coordinates
b
i
an an n
y ha
∂ρ 1 ∂ 1 ∂ ∂
+ (ρrvr ) + (ρvθ ) + (ρvz ) = 0 (10.116)
le
∂t r ∂r r ∂θ ∂z
liv or ec
M
.A
Spherical coordinates
m
1 ∂ 1 1
d
∂ρ ∂ ∂
ρr2 vr +
+ 2 (ρvθ sin θ ) + ρvφ = 0 (10.117)
uu
e
∂t r ∂ r r sin θ ∂ θ r sin θ ∂ φ
X Th
er
tin
on
.O
Cartesian coordinates
2
∂ vx ∂ 2 vx ∂ 2 vx
∂p ∂ vx ∂ vx ∂ vx ∂ vx
− +µ + + + ρb x = ρ + v x + v y + vz
∂x ∂ x2 ∂ y2 ∂ z2 ∂t ∂x ∂y ∂z
2 2 2
∂p ∂ vy ∂ vy ∂ vy ∂ vy ∂ vy ∂ vy ∂ vy
− +µ + + 2 + ρby = ρ + vx + vy + vz
∂y ∂ x2 ∂ y2 ∂z ∂t ∂x ∂y ∂z
2 2 2
∂p ∂ vz ∂ vz ∂ vz ∂ vz ∂ vz ∂ vz ∂ vz
− +µ + + 2 + ρbz = ρ + vx + vy + vz
∂z ∂ x2 ∂ y2 ∂z ∂t ∂x ∂y ∂z
(10.118)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Fluid Mechanics Formulas 493
Cylindrical coordinates
1 ∂ 2 vr 2 ∂ vθ ∂ 2 vr
∂p ∂ 1 ∂
− +µ (rvr ) + 2 − + + ρbr =
∂r ∂r r ∂r r ∂ θ 2 r2 ∂ θ ∂ z2
∂ vr vθ ∂ vr v2θ
∂ vr ∂ vr
=ρ + vr + − + vz
∂t ∂r r ∂θ r ∂z
1 ∂ 2 vθ 2 ∂ vr ∂ 2 vθ
1∂p ∂ 1 ∂
rs
− +µ (rvθ ) + 2 + 2 + + ρbθ =
r ∂θ ∂r r ∂r r ∂θ2 r ∂θ ∂ z2
ee
∂ vθ ∂ vθ vθ ∂ vθ vr vθ ∂ vθ
n
=ρ + vr + + + vz
gi
∂t ∂r r ∂θ r ∂z
t d le En
1 ∂ 2 vz ∂ 2 vz
ar
∂p 1 ∂ ∂ vz
ar s
ge ro or
− +µ r + 2 + 2 + ρbz =
eS m
ib
∂z r ∂r ∂r r ∂θ2 ∂z
ac
∂ vzf ∂ vz vθ ∂ vz ∂ vz
C d P cs
b
=ρ + vr + + vz
∂r r ∂θ ∂z
i
∂t
an an n
(10.119)
y ha
le
Spherical coordinates
liv or ec
M
.A
1 ∂ 2 vr
∂p 1 ∂ 2 1 ∂ vr
m
∂ ∂
− +µ r v + sin + +
d
r θ
∂r ∂r r2 ∂ r r2 sin θ ∂ θ r2 sin2 θ ∂ φ 2
uu
∂!θ
e
X Th
2 2 ∂ vφ
er
∂
tin
.O
2 2
!
∂ vr ∂ vr vθ ∂ vr vφ ∂ vr vθ + vφ
=ρ + vr + + −
C
∂t ∂r r ∂ θ r sin θ ∂ φ r
©
1∂p 1 ∂ 2 ∂ vθ 1 ∂ 1 ∂
− +µ 2 r + 2 (vθ sin θ ) +
r ∂θ r ∂r ∂r r ∂ θ sin θ ∂ θ !
1 ∂ 2 vθ 2 ∂ vr 2 cot θ ∂ vφ
+ 2 2 + 2 − + ρbθ =
r sin θ ∂ φ 2 r ∂ θ r2 sin θ ∂ φ !
2
∂ vθ ∂ vθ vθ ∂ vθ vφ ∂ vθ vr vθ vφ cot θ
=ρ + vr + + + −
∂t ∂r r ∂θ r sin θ ∂ φ r r
(10.120)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
494 C HAPTER 10. F LUID M ECHANICS
1 ∂p 1 ∂ ∂ vφ 1 ∂ 1 ∂
r2
− +µ 2 + 2 vφ sin θ +
r sin θ ∂ φ r ∂r ∂r r ∂ θ sin θ ∂ θ !
2
1 ∂ vφ 2 ∂ vr 2 cot θ ∂ vθ
+ 2 2 + 2 + + ρbϕ =
r sin θ ∂ φ 2 r sin θ ∂ φ r2 sin θ ∂ φ
∂ vφ ∂ vφ vθ ∂ vφ vφ ∂ vφ vφ vr vθ vφ
=ρ + vr + + + + cot θ
∂t ∂r r ∂θ r sin θ ∂ φ r r
(10.120 (cont.))
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 495
P ROBLEMS
Problem 10.1 – The barotropic fluid flowing inside the pipe shown in the figure
below has the following kinetic equation of state.
ρ
p = β ln (β and ρ0 const.)
rs
ρ0
ee
Determine, for a steady-state regime, the exit pressure P2 in terms of the other
n
variables shown in the figure.
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
Solution
M
.A
The global spatial form of the mass continuity equation (5.22) states
m
d
Z
uu
ρ dV = 0 .
e
dt
X Th
er
tin
.O
C
d
Z Z Z Z Z
∂ ∂
ρ dV = ρ dV + ρv · n dS =⇒ ρ dV + ρv · n dS = 0 ,
©
dt ∂t ∂t
V V ∂V V ∂V
Applying this last expression to the problem described in the statement produces
−ρ1 v1 S1 + ρ2 v2 S2 = 0 =⇒ ρ1 v1 S1 = ρ2 v2 S2 .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
496 C HAPTER 10. F LUID M ECHANICS
Finally, isolating the density from the given kinetic equation of state,
ρ
p = β ln =⇒ ρ = ρ0 e p/β ,
ρ0
and introducing it into the previous one produces
v1 S1
ρ0 e P1 /β v1 S1 = ρ0 e P2 /β v2 S2 =⇒ e (P2 −P1 )/β = =⇒
v2 S2
rs
v1 S1
P2 = P1 + β ln .
ee
v2 S2
n
gi
t d le En
ar
Problem 10.2 – Determine the value per unit of thickness of the horizontal
ar s
ge ro or
eS m
ib
force F that must be applied on point B of the semicircular floodgate shown
ac
f
in the figure such that the straight line AB remains vertical. The floodgate can
C d P cs
b
rotate around the hinge A and separates two different height levels, h and αh,
i
of a same fluid.
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
H YPOTHESES :
C
Solution
The only forces acting on the floodgate are the pressure forces of the fluids,
the force F and the reaction in A (horizontal component H and vertical compo-
nent V).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 497
Figure A Figure B
Since the pressure exerted by the fluids is perpendicular to the surface of the
rs
floodgate and the floodgate is circular, the resultant force obtained by integrating
ee
the pressures on the surface are applied at the center of the circumference defined
n
by the floodgate. Thus, posing the equilibrium of momentum with respect to the
gi
center of this circumference (see Figure A) results in
t d le En
FR = HR =⇒ H=F.
ar
ar s
ge ro or
eS m
ib
Imposing now the equilibrium of horizontal forces, knowing that the fluids exert
ac
f
a horizontal pressure with a triangular distribution (see Figure B), yields
C d P cs
b
i
1 1 1
an an n
2 2 4
le
liv or ec
M
.A
m
Problem 10.3 – Determine the relation between the. force F applied on the
d
uu
er
tin
on
.O
C
H YPOTHESES :
1) Assume the fluid is an incompressible perfect fluid in steady-state regime.
2) The atmospheric pressure is negligible.
3) S1 and S2 are the cross-sections.
4) The density of the fluid is ρ.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
498 C HAPTER 10. F LUID M ECHANICS
Solution
The stress state of a perfect fluid is known to be of the form σ = −p 1 (see
Section 9.3 in Chapter 9). The mass continuity equation (5.22) is applied to
obtain the relation between the velocities of the fluid,
S1 S1 .
v1 S1 = v2 S2 =⇒ v2 = v1 = δ .
S2 S2
Taking into account Bernoulli’s theorem (10.63) between an arbitrary point in
contact with the piston and another at the exit cross-section, both belonging to a
rs
same streamline, results in
ee
. !
S1 . 2 1
2
p δ2
ρ S1 .
n
H+ + = 0+0+ δ =⇒ p = − 1 δ 2 − ρg H
gi
ρg 2g S2 2g 2 S2
t d le En
Therefore, p must be constant for any point in contact with the piston (x = H).
ar
ar s
ge ro or
Then,
eS m
ib
p = const. ∀x = H =⇒ F = p S1 .
ac
. f
C d P cs
b
Finally, the force F is related to δ in the following manner.
i
an an n
y ha
2 !
ρ
S1 .
le
F = S1 − 1 δ 2 − ρg H S1
liv or ec
2 S2
M
.A
m
d
uu
e
Problem 10.4 – A shear force f ∗ per unit of surface acts on an rigid plate of
X Th
er
tin
indefinite size with density ρ ∗ and thickness t. The plate slides at a velocity v∗
in the longitudinal direction on a plane inclined at an angle α with respect to
on
.O
the horizontal longitudinal direction. Between the plate and the inclined plane
C
are two distinct and immiscible Newtonian fluids with viscosities µ1 and µ2 ,
©
which are distributed into two layers with the same thickness h.
a) Establish the generic form of the pressure and velocity fields and argue the
hypotheses used to determine them.
b) Integrate the corresponding differential equations and obtain, except for the
integration constants, the distribution of pressures and velocities in each
fluid.
c) Indicate and justify the boundary conditions that must be applied to deter-
mine the above integration constants.
d) Completely determine the pressure and velocity fields as well as the stresses
in each fluid. Plot the distribution of each variable (velocities, pressure and
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 499
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
H YPOTHESES :
b
i
an an n
d
uu
e
X Th
Solution
er
tin
a) Note that the z-dimension, perpendicular to the plane of the paper, does not
on
.O
intervene in the problem. Thus, the pressure and velocity fields are reduced to
C
not T
p = p (x, y) and v = v (x, y) ≡ vx (x, y) , 0 , 0 .
©
In fact, vx does not depend on x either since the velocity should be the same in
all the cross-sections of the type A − A0 . If this is not acknowledged a priori, the
mass continuity equation (5.22) may be imposed, considering the incompress-
ible nature of the fluids, as follows.
dρ ∂ vx ∂ vy
+ ρ∇ · v = 0 =⇒ ∇·v = 0 =⇒ + = 0 , but vy = 0
dt ∂x ∂y
∂ vx not T
=⇒ =0 =⇒ vx = vx (y) =⇒ v ≡ vx (y) , 0 , 0
∂x
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
500 C HAPTER 10. F LUID M ECHANICS
rs
∂x ∂y
ee
∂p
y − component =⇒ 0=− ⇒ p = p (x) .
∂y
n
gi
The pressure p only depends on x and the component vx of v only depends on y.
t d le En
Therefore, the partial derivatives in the equation for the x-component can be
replaced by total derivatives. In this way, an equality of functions is obtained in
ar
ar s
ge ro or
which the pressure term depends solely on x whilst the velocity term depends
eS m
ib
ac
exclusively on y. Consequently, these terms must be constant.
f
C d P cs
d 2 vx
b
dp
i
=µ 2 d 2 vx
an an n
dp
dx dy =⇒ =µ = k = const.
y ha
dx dy2
le
f (x) = f (y)
liv or ec
M
.A
dp
=k =⇒ p (x) = k x + A
dx
m
d 2 vx
uu
k k 2
e
= =⇒ vx (y) = y + By + C
X Th
2 2µ
er
dy µ
tin
.O
Table10.2 is used,
C
σx = σy = σz = −p (x)
©
∂ vx (y) =⇒
τxy = τyx = µ
∂y
k
−k x − A µ µ
y+B 0
not
σ (x, y) ≡ k
µ y+B −k x − A 0
µ
0 0 −k x − A
where the constants in these expressions (k, A, B, C) are different for each fluid.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 501
2. v2x (y)y=−h = 0 , since the inclined plane does not move and µ > 0.
3. v1x (y)y=0 = v2x (y)y=0 , which is the continuity condition for v at the inter-
face between the two fluids.
rs
P RESSURE BOUNDARY CONDITIONS
ee
In the fluid with density µ1 , the pressure is prescribed for y = h or, directly,
n
since p does not depend on y (because the weight of the fluid is neglected), the
gi
pressure p1 is prescribed in the whole domain of this fluid. The value of p1
t d le En
corresponds to the pressure that the plate exerts on the fluid with density µ1 ,
which is the projection of the plate’s weight in the direction of the y-axis.
ar
ar s
ge ro or
eS m
ib
4. W = ρ ∗ gt is the weight of a section of the plate with unit length, according
ac
f
to the x- and z-axis. Here, patm = 0 has been considered.
C d P cs
b
5. p1 = ρ ∗ gt cos α , ∀x is the projection on the y-axis. Since a unit length has
i
an an n
le
6. p1 = p2 ∀x is the continuity condition for the pressure in the
liv or ec
y=0 y=0
interface between the two fluids.
M
.A
m
The continuity condition for stresses that must be imposed in the interface be-
X Th
er
tween the two fluids does not affect the complete tensor σ . Instead, only the
tin
.O
t1 = − t2
C
y=0 y=0
©
must be satisfied. Considering that the unit normal vector n is the exterior normal
vector, then
not not
n1 ≡ [0 , −1 , 0]T and n2 ≡ [0 , 1 , 0]T .
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
502 C HAPTER 10. F LUID M ECHANICS
k1 2
h + B1 h + C1 = v∗
2µ1
k2 2
h − B2 h + C2 = 0
2µ2
C1 = C2
(
k1 = 0
k1 x + A1 = ρ ∗ gt cos α , ∀x =⇒
A1 = ρ ∗ gt cos α
(
rs
k1 = k 2 = 0
k1 x + A1 = k2 x + A2 , ∀x =⇒
ee
A1 = A2 = ρ ∗ gt cos α
n
gi
y=0
=⇒ µ1 B1 = µ2 B2
t d le En
k1 = k2 = 0
ar
ar s
ge ro or
Solving and replacing these values in the expressions for the pressure, velocity
eS m
ib
and stress obtained in b) results in
ac
f
C d P cs
b
i
v∗
y µ1
an an n
µ1 h µ2
1+
le
µ2
liv or ec
1 = τ2 = µ v∗
v∗ τxy = const.
M
1
.A
µ1 y xy
v2x (y) = +1 µ1
µ2 1 + µ1 h h 1+
µ2
m
µ2
uu
e
X Th
er
tin
unit element of the plate is posed. Three forces act on this element:
.O
C
1) The force f ∗ that pushes the plate in the positive direction of the x-axis.
©
2) The projection of the plate’s own weight in the direction of the x-axis. This
force pulls the plate in the negative direction of the x-axis.
3) The shear force of the fluid on the plate, which resists the motion of the plate
and, thus, acts in the negative direction of the x-axis.
To determine the sign criterion of this last force, the stresses acting on an element
of the fluid domain are drawn:
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 503
rs
Posing the equilibrium of forces yields
ee
v∗
n
f ∗ = ρ ∗ gt sin α + µ1
gi
µ1
1+ h
t d le En
µ2
ar
and, isolating v∗ , produces the velocity in terms of the shear force,
ar s
ge ro or
eS m
ib
ac
f
C d P cs
h µ1
v∗ = 1+
b (f ∗ − ρ ∗ gt sin α) .
i
µ1 µ2
an an n
y ha
le
To compute the volume flow rate that flows across the surface B−B0 , one must
liv or ec
take into account that the fluids are incompressible and, thus, the volume flow
M
.A
rate crossing the curved surface is the same as if a straight segment joining B
and B0 was considered, that is,
m
d
uu
Zh
e
Z Z
X Th
er
q= v · n dS = v · n dS = vx (y) dy .
tin
.O
Then, replacing the expressions found in d) for the velocity vx and integrating
C
1 µ1
q = v∗ h + .
2 µ1 + µ2
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
504 C HAPTER 10. F LUID M ECHANICS
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
Figure A Figure B
le
liv or ec
a) Determine the pressure and velocity fields in the zone of the fluid shown in
Figure B (zone ABCD), except for the integration constants.
M
.A
b) Indicate and justify the boundary conditions that must be applied to deter-
m
c) Completely determine the pressure and velocity fields in zone ABCD of the
e
X Th
fluid.
er
tin
d) Determine the expression of the stress tensor in zone ABCD of the fluid.
e) Assuming that the stress σy in the surface A−A0 is uniform and equal
on
.O
. to
the stress in point A, prove there exists a relation of the form F.= η δ (t),
C
where F is the force per unit of length applied on the piston and δ (t) is the
©
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 505
Solution
a) The problem is not defined in the z-direction, the direction perpendicular to
the plane of the paper, and, thus, is independent of the z variable. Then, consider
the bidimensional situation
not T
v ≡ vx (x, y) , vy (x, y) , 0 .
rs
For convenience, an additional approximate hypoth-
ee
esis is introduced to further simplify the problem: it
n
will be assumed that vx = 0 in all the zone ABCD of
gi
the fluid. However, the streamlines have, in fact, the
t d le En
approximate form shown in the figure to the right.
It is even possible that vortexes are formed in this
ar
ar s
region if there is a high velocity.
ge ro or
eS m
ib
In short, the velocity and pressure fields are assumed to be of the form
ac
f
C d P cs
T
b
not
v ≡ 0 , vy (x, y) , 0 and p = p (x, y) .
i
an an n
y ha
.A
∂ vy
=0 =⇒ vy = vy (x) .
∂y
m
d
uu
Then, the velocity remains constant for a same vertical line since the spatial
e
X Th
.O
∂p
0=− =⇒ p = p (y) ∂p ∂ 2 vy
∂x =⇒ = µ
2
∂ vy
∂p ∂y ∂ x2
0=− +µ 2
∂y ∂x
The term in the right-hand side of the equation depends solely on x and the one
in the left-hand side depends only on y, therefore both terms must be constant.
∂p
=k =⇒ p = ky +C1
∂y
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
506 C HAPTER 10. F LUID M ECHANICS
∂ 2 vy ∂ vy k 1k 2
µ =k =⇒ = x +C2 =⇒ vy (x) = x +C2 x +C3
∂ x2 ∂x µ 2µ
rs
.
2. vy (x) = −δ , ∀y, again, since there is no relative displacement.
ee
x=a
n
gi
P RESSURE BOUNDARY CONDITIONS
t d le En
3. p (y) = patm = 0
ar
ar s
ge ro or
y=m+h
eS m
ib
ac
f
C d P cs
VOLUME FLOW RATE BOUNDARY CONDITIONS
b
i
an an n
In an incompressible fluid the entrance and exit volume flow rates are the same,
y ha
Q= v · n dS .
M
.A
S
.
m
The piston descends at a velocity δ and, thus, its cross-section is introduced into
d
the fluid, pushing it upwards. Then, the entrance volume flow rate can be defined
uu
e
as (velocity · surface),
X Th
.
er
tin
Qin = δ · L .
on
.O
On the other hand, the exit volume flow rate, flowing in the space left between
the piston and the lateral walls, is determined by means of the general expression
C
Z Za
Qout = 2 v · n dS = 2 vy (x) dx .
Sa 0
Finally, equating the entrance and exit volume flow rates results in:
Za .
4. 2 vy (x) dx = δ L
0
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 507
.
. 1k 2 δ k
vy (x) = −δ =⇒ vy (a) = a +C2 a =⇒ C2 = − − a
x=a 2µ a 2µ
Za Za
rs
k 2
k a3 a2
.
2 vy (x) dx = 2 x +C2 x dx = 2 +C2 = δL =⇒
ee
2µ 2µ 3 2
n
0 0
gi
.
6µ .
L
t d le En
and δ
k = − 3 δ (a + L) C2 = 2+3
a a a
ar
ar s
ge ro or
eS m
ib
ac
f 6µ .
C d P cs
b
p (y) =0 =⇒ k (m + h)+C1 = 0 =⇒
C1 = δ (a + L) (m + h)
i
y=m+h a3
an an n
y ha
le
Introducing these values in the expressions for the pressure and velocity obtained
liv or ec
in a) results in:
M
.A
m
6µ .
d
p = p (y) = δ (a + L) (m + h − y)
uu
a3
e
.
X Th
. 2 δ
er
3 L
tin
vy (x) = − 3 (a + L) δ x + 2+3 x
a a a
on
.O
C
d) The stresses in zone ABCD of the fluid are computed by means of the con-
stitutive equation in Cartesian coordinates of Table 10.2. Using the expressions
for the pressure and velocity fields obtained in c) yields
∂ vy
−p µ ∂ x 0
not ∂v where
σ ≡ µ y −p
∂x 0
∂ vy . x L 2
= µ δ −6 3 (a + L) + 3 2 + .
µ
0 0 −p ∂x a a a
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
508 C HAPTER 10. F LUID M ECHANICS
C OMMENT
When the piston descends, the steady-state regime hypothesis is, in fact, not
completely rigorous since, at some point, the piston will reach the lowest point
of its trajectory and the flow will vary. To
. be able to apply this hypothesis, either
(m + h) must be a very large length or δ must be a very low velocity.
e) The stresses acting on the piston must be computed to obtain the resultant
forces and, then, the equilibrium of forces is applied to determine the expression
for F. These stresses are:
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
postive sign
ac
criterion f
C d P cs
b
i
an an n
6µ .
le
σ ∗ = −σy = p (y) = k m +C1 =⇒ σ ∗ = 3 (a + L) hδ .
liv or ec
A y=m a
M
.A
In the lateral surfaces, due to symmetry, τ1∗ = τ2∗ and, therefore, only τ1∗ needs
m
to be computed,
d
uu
e
. a L 2
X Th
er
∗ ∗
tin
τ1 = τ2 = −τxy = −µ δ −6 3 (a + L) + 3 2 + =⇒
x=a a a a
on
.O
.
µδ L
τ1∗ = τ2∗
C
= 3 +4 .
a a
©
.
Imposing the equilibrium of forces (since δ is a constant velocity),
F = Lσ ∗ + h τ1∗ + h τ2∗ =⇒
. 2µh
2
L L
F = ηδ with η = 3 2 +6 +4
a a a
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 509
E XERCISES
10.1 – Compute the horizontal and vertical components of the resultant of the
actions, per unit of length, exerted by the water on the gravity dam shown in the
figure.
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
10.2 – The wall of a tank has a valve that rotates about point O as shown in the
i
an an n
figure. Compute the resultant force and moment, per unit of thickness, that the
y ha
fluid exerts on the valve. The weight of the valve can be neglected.
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
10.3 – Determine the weight of the ballast W0 required at the bottom of the crate
shown in the figure, whose weight is W, such that it is maintained afloat in stable
equilibrium.
N OTE: The water has a density ρ and the weights are per unit of thickness.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
510 C HAPTER 10. F LUID M ECHANICS
rs
10.5 – A plate of indefinite size and thickness 2a separates two incompressible
ee
Newtonian fluids that move between two rigid boundaries of indefinite length
n
placed at a distance h from the plate, as shown in the figure. The plate and the
gi
top boundary move at velocities v/2 and v, respectively. Determine:
t d le En
a) The pressure, velocity and stress fields in terms of the integration constants.
ar
b) The integration constants, by applying the adequate boundary conditions.
ar s
ge ro or
eS m
ib
c) The forces per unit of surface F1 and F2 exerted on the plate and the top
ac
boundary needed to produce the described motion. f
C d P cs
b
d) The dissipated energy, per unit of time and of surface perpendicular to the
i
an an n
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
Additional hypotheses:
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 511
rs
n ee
gi
t d le En
ar
ar s
ge ro or
Additional hypotheses:
eS m
ib
ac
f
1) The flow is assumed to be parallel to the x-y plane.
C d P cs
b
2) Inertial forces can be neglected.
i
3) The volume flow rate, Q, and the components of the force, V and H, are
an an n
y ha
.A
m
er
tin
ible Newtonian fluid, with viscosity µ1 , is contained inside the cylindrical shell.
Determine:
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
512 C HAPTER 10. F LUID M ECHANICS
rs
Additional hypotheses:
ee
1) The rotation of the disc is sufficiently
n
slow to neglect the inertial forces.
gi
2) The effect of the lateral walls (fluid-
t d le En
wall friction effects) can be neglected.
ar
3) The velocity field varies linearly with
ar s
ge ro or
eS m
ib
the distance to the inferior surface.
ac
f
4) Assume a steady-state regime.
C d P cs
b
i
an an n
10.9 – The cross-section of a cylindrical piston ABA0 B0 that slides inside a con-
y ha
in the figure. The motion of the piston, at a velocity δ , causes the fluid to flow
M
.A
b) Indicate and justify the boundary conditions that must be applied to deter-
X Th
er
tin
mine the value of the integration constants. Determine these constants and
the complete expressions of the pressure and velocity fields.
on
.O
d) Assuming that the stress normal to the surface BB0 in the fluid is constant
©
Additional hypotheses:
1) The body forces of the fluid and weight
of the piston can be neglected.
2) Assume a steady-state regime.
3) The atmospheric pressure is negligible.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Chapter 11
Variational Principles
rs
n ee
gi
11.1 Governing Equations
t d le En
ar
Variational calculus is a mathematical tool that allows working with the so-
ar s
ge ro or
eS m
ib
called integral or weak form of the governing differential equations of a problem.
ac
f
Given a system of differential equations, which must be verified in local form
C d P cs
(point by point) for a certain domain, the variational principles allow obtain-
b
i
ing an integral or weak formulation (global, in the domain), whose imposition,
an an n
isfied. Integral formulations are of particular interest when treating and solving
le
liv or ec
.A
m
er
tin
.O
With certain language misuse, one could say that the functional F (u) is a scalar
function whose arguments are functions u (x).
513
514 C HAPTER 11. VARIATIONAL P RINCIPLES
Zb Zb
F (u) = u (x) dx , G (u) = u0 (x) dx
a a
Zb
rs
F x, u (x) , u0 (x) dx .
and H (u) =
ee
a
n
gi
t d le En
ar
ar s
ge ro or
Definition 11.2.
Consider the (scalar, vector
or tensor) function
eS m
ib
space X := u (x) u (x) : R3 ⊃ Ω → Rn on a domain Ω and a
ac
functional F (•) : X → R. f
C d P cs
b
i
Consider the two functions u, η ∈ X and the (perturbation) param-
an an n
.A
de f d
uu
δ F (u; η ) = F (u + ε η ) .
e
dε ε=0
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Governing Equations 515
Solution
d d
Z
δ F (u; δ u) = F (u + εδ u)
= φ (u + εδ u) dΩ +
dε ε=0 dε ε=0
Ω
rs
d
Z
+ ϕ (u + εδ u) dΓ =
ee
dε ε=0
n
∂Ω
gi
t d le En
Z ∂ φ (u + εδ u) d (u + εδ u)
ar
= · dΩ +
ar s
ge ro or
∂u dε
eS m
ib
Ω
ac
| {z }
δu f ε =0
C d P cs
b
i
an an n
Z ∂ ϕ (u + εδ u) d (u + εδ u)
+ · dΓ =⇒
y ha
∂u dε
le
liv or ec
∂Ω | {z }
δu ε =0
M
.A
m
∂ φ (u) ∂ ϕ (u)
Z Z Z Z
d
δ φ (u) dΩ + ϕ (u) dΓ = · δ u dΩ + · δ u dΓ
uu
e
∂u ∂u
X Th
er
Ω ∂Ω Ω ∂Ω
tin
Note, in this case, the formal similarity of obtaining the Gateaux derivative
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
516 C HAPTER 11. VARIATIONAL P RINCIPLES
\
Γu Γσ = 0/
rs
n ee
gi
Remark 11.2. When computing the Gateaux derivative, a condition,
established in the definition itself, on the perturbation η ≡ δ u is that
t d le En
the perturbed function u + ε δ u must belong to the same function
ar
space V (u + ε δ u ∈ V). In this case, if u + ε δ u ∈ V,
ar s
ge ro or
eS m
ib
ac
f
(u + ε δ u) = u∗ =⇒ u + ε δ u = u∗ =⇒ ε δ u = 0
C d P cs
x∈Γu x∈Γ
b x∈Γu x∈Γu
i
| {z }u
an an n
u∗
y ha
le
and the perturbation δ u must satisfy δ u =0.
liv or ec
x∈Γu
M
.A
m
Based on the family of functions (11.1), consider now the following family
uu
e
of functionals
X Th
er
tin
Z Z
on
.O
Ω Γσ
©
where the functions φ and ϕ are regular enough to be integrable in the domains
Ω and Γσ , respectively. Assume, in addition, that, through adequate algebraic
operations, the Gateaux derivative of F (u) can be written as
Z Z
δ F (u; δ u) = E (x, u (x) , ∇u) · δ u dΩ + T (x, u (x) , ∇u) · δ u dΓ
. (11.3)
Ω Γσ
∀δ u; δ u = 0
x∈Γσ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Governing Equations 517
Example 11.3 – Obtain the Gateaux derivative, in the format given in (11.3),
of the functional
Zb
0
F (u) = φ x, u (x) , u (x) dx with u (x) = u (a) = p .
x=a
a
Solution
The given functional is a particular case of the functional in (11.2), reduced
to a single dimension with ϕ ≡ 0, Ω ≡ (a, b), Γu ≡ a and Γσ ≡ b.
rs
ee
Perturbing the function u (x) and replacing in the functional yields
n
gi
)
u(x) → u(x) + ε η(x)
not
∀η (x) ≡ u (x) η (a) = ηa = 0 =⇒
t d le En
0 0 0
δ
u (x) → u (x) + ε η (x)
ar
ar s
ge ro or
eS m
ib
Zb
ac
F (u + ε η) = f
φ x, u (x) + ε η, u0 (x) + ε η 0 dx
C d P cs
a
b
i
an an n
le
Zb
liv or ec
d ∂φ ∂φ 0
δ F (u; η) = F (u + ε η) = η + 0 η dx .
M
.A
dε ε=0 ∂u ∂u
a
m
er
Zb x=b Zb
tin
∂φ 0 ∂φ d ∂φ
η dx = η − η dx =
∂ u0 ∂ u0 x=a dx ∂ u0
on
.O
a a
Zb
C
∂φ ∂ φ d ∂φ
©
= 0 ηb − ηa − η dx =
∂ u0 |{z} dx ∂ u0
∂u
x=b x=a a
=0
Zb
∂φ d ∂φ
= 0 ηb − η dx ,
∂u dx ∂ u0
x=b a
Zb ∂ φ
d ∂φ ∂ φ
δ F u; δ u = − δ u dx + 0 δ ub ,
|{z} ∂ u dx ∂ u0 ∂u
η a x=b
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
518 C HAPTER 11. VARIATIONAL P RINCIPLES
∂φ d ∂φ
E (x, u, u0 ) ≡ − ∀x ∈ (a, b)
∂ u dx ∂ u0
.
∂ φ
T (x, u, u0 ) ≡ 0
∂u
x=b
rs
ee
11.1.2 Extrema of the Functionals. Variational Principles.
Euler-Lagrange Equations
n
gi
Consider a real function of a real variable f (x) : R → R. This function is said to
t d le En
have a minimum at x = x0 when
ar
ar s
ge ro or
f (x0 ) ≤ f (x) ∀x ∈ R . (11.4)
eS m
ib
ac
f
The necessary condition for f to have an extrema (maximum, minimum or sad-
C d P cs
dle point) at x = x0 is known to be
b
i
an an n
y ha
d f (x) not 0
le
= f (x0 ) = 0 . (11.5)
liv or ec
dx
x=x0
M
.A
when
e
X Th
.O
δ F (u; δ u) = 0 ∀δ u δ u
=0 . (11.7)
x∈Γu
Variational principle
Z Z
δ F (u; δ u) = E · δ u dΩ + T · δ u dΓ = 0 (11.8)
∀δ u δ u =0
Ω Γσ
x∈Γu
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Governing Equations 519
E (x) = 0 ∀x ∈ Ω
⇐⇒
T (x) = 0 ∀x ∈ Γσ
rs
n ee
gi
Proof (indicative1 )
t d le En
Consider the following choice for δ u (x).
ar
ar s
ge ro or
E (x) ∀x ∈ Ω
eS m
ib
δ u (x) = 0 ∀x ∈ Γu
ac
f
C d P cs
∀x ∈ Γσ
T (x)
b
i
an an n
le
Z Z
liv or ec
.A
Ω | {z } Γσ | {z }
≥0 ≥0
m
Q.E.D.
uu
e
X Th
.O
C
Euler-Lagrange equations
(11.9)
©
1 This proof is not rigorous and is provided solely as an intuitive indication of the line of
reasoning followed by the theorem’s proof.
2 Strictly speaking, (11.8) is a variational equation or the weak form of a differential prob-
lem.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
520 C HAPTER 11. VARIATIONAL P RINCIPLES
rs
ee
constitute a set of boundary conditions on these differential equa-
n
tions denoted as natural or Neumann boundary conditions. Together
gi
with the conditions (11.1),
t d le En
u (x) = u∗ (x) ∀x ∈ Γu ,
ar
ar s
ge ro or
eS m
ib
named essential or Dirichlet boundary conditions, they define a sys-
ac
f
tem whose solution u (x) is an extreme of the functional F.
C d P cs
b
i
an an n
y ha
le
liv or ec
.A
Zb
d
φ x, u (x) , u0 (x) dx
uu
x=a
X Th
er
a
tin
on
.O
Solution
C
Zb ∂ φ
d ∂φ ∂ φ
δ F u; δ u = − δ u dx + 0 δ ub ,
|{z} ∂ u dx ∂ u0 ∂u
η a x=b
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Virtual Work Principle (Theorem) 521
Euler-Lagrange equations :
0 ∂φ d ∂φ
E (x, u, u ) ≡ − =0 ∀x ∈ (a, b)
∂ u dx ∂ u0
rs
∂u
x=b
ee
Essential boundary conditions
:
n
gi
u (x) = u (a) = p
t d le En
x=a
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
11.2 Virtual Work Principle (Theorem) b
i
an an n
y ha
forces t∗ (x,t) on the boundary Γσ (see Figure 11.2). Consider also the func-
M
.A
tional spaceV of all the admissible displacements, which satisfy the boundary
condition ux∈Γu = u∗ .
m
d
uu
e
X Th
(11.11)
= ut∗ (x)
V := ut (x) : V → R3 ut (x)
on
.O
x∈Γu
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
522 C HAPTER 11. VARIATIONAL P RINCIPLES
rs
n ee
Figure 11.2: Definition of the material volume Vt .
gi
t d le En
Consider now the variational principle
ar
ar s
ge ro or
eS m
ib
ac
f
Z
δ W (u; δ u) = ∇ · σ (u) + ρ (b − a (u)) · δ u dV +
C d P cs
b
i
V | {z }
an an n
E (11.14)
y ha
Z
t∗ − σ (u) · n · δ u dΓ = 0 ; ∀δ u (x) δ u
le
+ =0,
liv or ec
x∈Γu
Γσ
M
| {z }
.A
T
m
er
tin
Virtual displacements: δ u : V → R3 δ u =0 (11.15)
x∈Γu
on
.O
Euler-Lagrange
E ≡ ∇ · σ + ρ (b − a) = 0 ∀x ∈ Ω ,
equations:
(11.16)
Natural boundary T ≡ t∗ − σ · n = 0 ∀x ∈ Γσ ,
conditions:
that is, Cauchy’s equation (11.12) and the equilibrium condition at the bound-
ary (11.13).
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Virtual Work Principle (Theorem) 523
rs
and the splitting of δ u ⊗ ∇ into its symmetrical part, ∇s δ u, and its skew-
ee
symmetric part ∇a δ u,
n
gi
δ u ⊗ ∇ = ∇s δ u + ∇a δ u
with
t d le En
de f 1 de f 1
∇s δ u = (δ u ⊗ ∇ + ∇ ⊗ δ u) and ∇a δ u = (δ u ⊗ ∇ − ∇ ⊗ δ u) .
ar
2 2
ar s
ge ro or
eS m
ib
(11.18)
Introducing (11.18) in (11.17)3 produces
ac
f
C d P cs
b
i
(∇ · σ ) · δ u = ∇ · (σ
σ · δ u) − σ : (δ u ⊗ ∇) =
an an n
y ha
σ · δ u) − σ : ∇s δ u − σ : ∇a δ u
= ∇ · (σ =⇒ (11.19)
le
liv or ec
| {z }
=0
M
.A
m
σ · δ u) − σ : ∇s δ u
(∇ · σ ) · δ u = ∇ · (σ . (11.20)
d
uu
e
X Th
Integrating now (11.20) over the domain V and applying the Divergence Theo-
er
tin
rem yields
on
.O
Z Z Z
(∇ · σ ) · δ u dV = ∇ · (σ
σ · δ u) dV − σ : ∇s δ u dV =
C
V Z V Z V
= n · (σ
σ · δ u) dΓ − σ : ∇s δ u dV =
ZV
S
∂V =Γu Γσ
Z Z
= (n · σ ) · δ u dΓ + (n · σ ) · δ u dΓ − σ : ∇s δ u dV =⇒
|{z}
Γu =0 Γσ V
(11.21)
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
524 C HAPTER 11. VARIATIONAL P RINCIPLES
Z Z Z
(∇ · σ ) · δ u dV = (n · σ ) · δ u dΓ − σ : ∇s δ u dV , (11.22)
V Γσ V
where the condition δ ux∈Γu = 0 (see (11.15)) has been taken into account. Fi-
nally, introducing (11.20) in the original form of the variational principle (11.14)
results in
Z Z
t∗ − (σ
δ W (u; δ u) = ∇ · σ + ρ (b − a) · δ u dV + σ · n) · δ u dΓ =
rs
V Z Γσ
ee
Z Z Z
∗
= (∇ · σ ) · δ u dV + ρ (b − a) · δ u dV + t · δ u dΓ − σ · n) · δ u dΓ =
(σ
n
gi
V Z VZ ΓσZ Γσ
t d le En
s ∗
=− σ : ∇ δ u dV + ρ (b − a) · δ u dV + t · δ u dΓ = 0 =⇒
ar
V V Γσ
ar s
ge ro or
eS m
ib
(11.23)
ac
Virtual Work Principle f
C d P cs
b
Z Z
t∗ · δ u dΓ
i
δ W (u; δ u) = ρ (b − a) · δ u dV +
an an n
(11.24)
y ha
V Γσ
le
Z
liv or ec
σ : ∇s δ u dV = 0
− ∀δ u (x) δ u = 0
M
x∈Γu
.A
V
m
principle and maintains the same Euler-Lagrange equations and boundary con-
e
X Th
ditions (11.16), is known as the Virtual Work Principle (or Theorem) (VWP).
er
tin
on
.O
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Virtual Work Principle (Theorem) 525
rs
ee
11.2.1 Interpretation of the Virtual Work Principle
n
gi
Consider the continuous medium in the present configuration Vt at time t sub-
jected to the fictitious body forces b∗ (x,t) = b (x,t) − a (x,t) and the real sur-
t d le En
face forces t∗ (x,t) (see Figure 11.3), and suffering the real stresses σ (x,t).
ar
ar s
ge ro or
Consider, in addition, the virtual (fictitious) configuration Vt+δt corresponding
eS m
ib
to time t + δt, separated from the real configuration by a virtual displacement
ac
field (11.15) f
C d P cs
b
i
an an n
. (11.25)
y ha
Virtual displacements: δ u δ u =0
x∈Γu
le
liv or ec
M
.A
m
t − present configuration
uu
e
t + δt − virtual configuration
X Th
er
tin
on
.O
C
Figure 11.3: Continuous medium subjected to fictitious body forces and real surface
forces.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
526 C HAPTER 11. VARIATIONAL P RINCIPLES
Under infinitesimal strain kinematics, the virtual strains associated with the
virtual displacements (11.25) are
and, assuming that the stresses σ (x,t) remain constant along the time inter-
val [t, t + δt], the virtual strain work (internal virtual work) performed by the
medium during this interval is
rs
Z Z
Internal δ Wint = σ : ∇s δ u dV
ee
σ : δ ε dV = . (11.27)
virtual work:
n
V V
gi
Likewise, assuming that both the pseudo-body forces b∗ (x,t) and the surface
t d le En
forces t∗ (x,t) remain constant during the virtual strain process in the inter-
ar
ar s
val [t, t + δt], the work performed by these forces (external virtual work) results
ge ro or
eS m
ib
in
ac
f
C d P cs
b
Z Z
External δ Wext =
i
ρ (b − a) · δ u dV = t∗ · δ u dV
an an n
| {z }
V
b∗ Γσ
le
liv or ec
and, comparing the VWP (11.24) with expressions (11.27) and (11.28), the
M
.A
d
uu
e
er
tin
Z Z Z
δW = σ : δ ε dV − ρ (b − a) · δ u dV + t∗ · δ u dΓ = 0
on
.O
|{z}
Total V V Γσ
virtual | {z } | {z }
C
(11.29)
=⇒ δ W = δ Wint − δ Wext = 0
for any kinematically admissible
change inthe virtual configuration
δ u =0
x=Γu
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Virtual Work Principle (Theorem) 527
11.2.2 Virtual Work Principle in terms of the Stress and Strain Vectors
The vectors of stress {σ σ } and virtual strain {δ ε } can be extracted from the
symmetrical tensors of stress, σ , and virtual strain, δ ε = ∇s δ u, in (11.29) as
follows.
σx δ εx δ εx
σ δε δε
y y y
σ δ ε δ ε
σ} ≡ z {δ ε } ∈ R6 ; {δ ε } ≡ z = z
not not
σ } ∈ R6 ; {σ
{σ
δ γxy 2δ εxy
τxy
rs
τxz δ γxz 2δ εxz
ee
τyz δ γyz 2δ εyz
(11.30)
n
gi
They satisfy the equality
t d le En
(
σ : δ ε = {σ σ } · {δ ε } = {δ ε } · {σ
σ} ,
ar
(11.31)
ar s
ge ro or
σi j δ εi j = σm δ εm = δ εm σm i, j ∈ {1, 2, 3} , m ∈ {1, ..6} .
eS m
ib
ac
f
C d P cs
Replacing (11.31) in the VWP (11.29) results in
b
i
an an n
y ha
σ } dV − ρ (b − a) · δ u dV + t∗ · δ u dΓ = 0
δ W = {δ ε } · {σ
M
.A
|{z}
Total V V Γσ
virtual | {z } | {z }
m
work δ Wint
work (δ Wext )
uu
(11.32)
e
X Th
er
=⇒ = δ Wint − δ Wext
tin
δW
for any kinematically admissible
on
.O
δ u =0
©
x=Γu
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
528 C HAPTER 11. VARIATIONAL P RINCIPLES
rs
ee
Elastic 1 1
û (εε ) = ε :C :ε = σ :ε
n
potential: 2 | {z } 2
gi
σ (11.33)
t d le En
∂ û (εε )
=C :ε =σ
ar
∂ε
ar s
ge ro or
eS m
ib
ac
f
C d P cs
2) The body forces ρb∗ (x,t) are conservative
b
i
That is, these body forces derive from a potential φ (u) and, thus,
an an n
y ha
le
liv or ec
∂ φ (u)
= −ρb∗ = −ρ (b − a) . (11.34)
M
∂u
.A
m
d
uu
e
er
tin
.O
not
b (x,t) ≡ [ 0, 0, −g ]T = const. and ρ (x,t) = const.
C
4 The restriction to the linear elastic problem can be made less strict and be extended to the
case of hyperelastic materials in a finite strain regime.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Potential Energy. Minimum Potential Energy Principle 529
∂ G (u)
t∗ = − . (11.35)
∂u
rs
ee
∂ t∗
=0.
n
∂u
gi
In this case, the potential of the surface forces is
t d le En
ar
∂ G (u)
ar s
G (u) = −t∗ · u = −t∗ .
ge ro or
=⇒
eS m
ib
∂u
ac
f
C d P cs
b
i
an an n
y ha
Under the above circumstances, the following functional, named total poten-
le
liv or ec
.A
Potential energy
m
Z Z Z
uu
(11.36)
X Th
er
| {z }
tin
Total V V Γσ
potential | {z } | {z } | {z }
energy Elastic Potential energy Potential energy of
on
.O
∂ û ∂ φ (u) ∂ G (u)
Z Z Z
δ U (u; δ u) = : ∇S (δ u) dV + · δ u dV + · δ u dΓ =
∂ ε | {z } ∂{zu } ∂{zu }
V |{z} V | Γσ |
σ δε −ρ (b − a) −t∗
Z Z Z
= σ : δ ε dV − ρ (b − a) · δ u dV − t∗ · δ u dΓ ; ∀δ u δ u =0,
x∈Γu
V V Γσ
(11.37)
where (11.33) to (11.35) have been taken into account.
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
530 C HAPTER 11. VARIATIONAL P RINCIPLES
rs
The variational principle (11.38), which is still the weak form of
ee
Cauchy’s equation (11.12) and the equilibrium condition at the
n
boundary (11.13), is now the Gateaux variation of the potential
gi
energy functional U (u) in (11.36). Consequently, this functional,
t d le En
which for the case of constant body and surface forces takes the form
ar
1
Z Z Z
ar s
ε : C : ε dV − ρ (b − a) · u dV − t∗ · u dΓ ,
ge ro or
U (u) =
eS m
ib
2
ac
V | {z } V
f Γσ
C d P cs
û (εε )
b
i
an an n
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 531
P ROBLEMS
rs
prove the Minimum Potential Energy Principle for a linear elastic material
ee
under infinitesimal strain regime.
n
gi
t d le En
ar
Solution
ar s
ge ro or
eS m
ib
A linear elastic material is a particular type of hyperelastic material and, thus,
ac
there must exist an elastic potential of the type f
C d P cs
b
i
an an n
∂W
∃ W (εε ) = σi j =⇒ δ W = σi j δ εi j = σ : δ ε .
y ha
∂ εi j
le
liv or ec
.A
∂ G (u)
∃ G (u) t∗ = − =⇒ δ G = −t∗ · δ u
m
∂u
uu
e
∂ Φ (u)
X Th
er
tin
∃ Φ (u) ρb = − =⇒ δ Φ = −ρb · δ u
∂u
on
.O
Now, the given expression of the Virtual Work Principle can be rewritten as
C
Z Z Z
δ W dV0 + δ Φ dV0 + δ G dΓ = 0 ∀ δu δu = 0 in Γu ,
©
V0 V0 Γσ
Z Z Z
δ W dV0 + Φ dV0 + G dΓ = 0 ∀ δ u δ u = 0 in Γu .
V0 V0 Γσ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
532 C HAPTER 11. VARIATIONAL P RINCIPLES
leads to
δU = 0 ∀ δ u δ u = 0 in Γu ,
which is the same as stating that U has an extreme at u. To prove that this extreme
is a minimum, consider
1 ∂ 2W
W (εε ) = ε : C : ε where Ci jkl = .
2 ∂ εi j ∂ εkl
rs
1
Z Z Z
U (u) = ε (u) : C : ε (u) dV0 − ρb · (u) dV0 − t∗ · u dΓ and
ee
2
V0 V0
n
Γσ
gi
1
Z Z
t d le En
U (u + δ u) = ε (u + δ u) : C : ε (u + δ u) dV0 − ρb · (u + δ u) dV0
2
ar
V0 V0
ar s
ge ro or
Z
eS m
ib
− t∗ · (u + δ u) dΓ .
ac
f
C d P cs
Γσ
ε (u + δ u) = ε (u) + ε (δ u)
y ha
le
results in the following expression for the subtraction U (u + δ u) − U (u):
liv or ec
M
.A
1 1
Z Z
U (u + δ u) − U (u) = ε (u) : C : ε (δ u) dV0 + ε (δ u) : C : ε (u) dV0
m
2 2
d
V0 V0
uu
e
1
Z Z Z
X Th
ε (δ u) : C : ε (δ u) dV0 − ρb · δ u dV0 − t∗ · δ u dΓ
er
+
tin
2
V0 V0 Γσ
on
.O
Introducing
C
ε (u) : C : ε (δ u) = ε (δ u) : C : ε (u) = σ : δ ε
reduces the subtraction to
1
Z Z
U (u + δ u) − U (u) = σ : δ ε dV0 + ε (δ u) : C : ε (δ u) dV0
2
V0 Z V0 Z
− ρb · δ u dV0 − t∗ · δ u dΓ .
V0 Γσ
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Problems and Exercises 533
yields
1
Z
U (u + δ u) − U (u) = ε (δ u) : C : ε (δ u) dV0 .
2
V0
rs
n ee
U (u + δ u) − U (u) ≥ 0
gi
t d le En
and, thus, the potential energy is seen to have a minimum in the equilibrium
state.
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
rs
n ee
gi
t d le En
ar
ar s
ge ro or
eS m
ib
ac
f
C d P cs
b
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961
Bibliography
rs
n ee
gi
S. S. Antman. Nonlinear Problems of Elasticity. Springer-Verlag, New York, 1995.
t d le En
R. J. Asaro and V. A. Lubarda. Mechanics of solids and materials. Cambridge University
Press, New York, 2006.
ar
ar s
ge ro or
Y. Başar and Weichert. Nonlinear continuum mechanics of solids: fundamental concepts and
eS m
ib
perspectives. Springer Verlag, 2000.
ac
f
R. C. Batra. Elements of continuum mechanics. AIAA Education Series, 2006.
C d P cs
T. Belytschko, W. K. Liu, and B. Moran. Non-linear finite elements for continua and struc-
b
i
tures. Wiley, 2002.
an an n
253, 1956.
le
J. Bonet and R. D. Wood. Nonlinear continuum mechanics for finite element analysis. Cam-
liv or ec
.A
P. Chadwick. Continuum mechanics concise theory and problems. George Allen & Unwin
Ltd, 1976.
m
Diego, 1994.
uu
e
er
W. F. Chen and D. J. Han. Plasticity for structural engineers. Springer-Verlag, New York,
1988.
on
.O
M. A. Crisfield. Non-linear finite element analysis of solids and structures. Volumes 1 and 2.
John Wiley & Sons, New York, 1997.
C. S. Desai and H. J. Siriwardane. Constitutive laws for engineering materials with emphasis
on geological materials. Prentice-Hall, Inc., New Jersey, 1984.
Y. C. Fung. Foundations of solids mechanics. Prentice Hall Inc., New Jersey, 1965.
Y. C. Fung. A first course in continuum mechanics. Prentice Hall Inc., New Jersey, 1977.
A. E. Green and P. M. Naghdi. A general theory of an elasto-plastic continuum. Arch. Rat.
Mech. Anal., 18, 1965.
M. E. Gurtin. An introduction to continuum mechanics. Academic Press, Inc., New York,
1996.
P. Haupt. Continuum mechanics and theory of materials. Springer-Verlag, Gemany, 2002.
R. Hill. The mathematical theory of plasticity. Oxford University Press, Oxford, 1950.
G. A. Holzapfel. Nonlinear solid mechanics. John Wiley & Sons Ltd., England, 2000.
L. M. Kachanov. Introduction to continuum damage mechanics. Springer, Netherlands, 1986.
535
536 B IBLIOGRAPHY
R. Lai, W. M., D., and E. Krempl. Introduction to continuum mechanics. Pergamon Press,
1978.
C. Lanczos. The variational principles of mechanics. Dover Publications, Inc., New York,
1970.
L. D. Landau and E. M. Lifshits. Fluid mechanics. Pergamon Press, 1959.
J. Lemaitre and J. L. Chaboche. Mechanics of solids materials. Cambridge University Press,
Cambridge, 1990.
A. E. H. Love. A treatise on the mathematical theory of elasticity. Cambridge University
Press, London, 1944.
J. Lubliner. Plasticity theory. Macmillan Publishing Company, New York, 1990.
L. E. Malvern. Introduction to the mechanics of a continuous medium. Prentice-Hall, New
Jersey, 1969.
rs
J. E. Marsden and T. J. R. Hughes. Mathematical foundations of elasticity. Dover Publica-
tions, Inc., New York, 1983.
ee
G. E. Mase. Mecánica del medio continuo. McGraw-Hill, USA, 1977.
n
A. Mendelson. Plasticity: Theory and application. Robert E. Krieger Publishing Company,
gi
New York, 1968.
R. W. Ogden. Non-linear elastic deformations. Dover Publications, Inc., New York, 1984.
t d le En
D. F. Parker. Fields, flows and waves: An introduction to continuum models. Springer-Verlag,
ar
London, 2003.
ar s
ge ro or
W. Pilkey and W. Wunderlich. Mechanics of structures. Variational and computational
eS m
ib
methods, 1992.
ac
S. B. Pope. Turbulent flows. IOP Publishing, 2001. f
C d P cs
b
A. Romano, R. Lancellota, and A. Marasco. Continuum mechanics using Mathematica:
i
fundamentals, applications, and scientific computing. Birkhauser, Boston, 2006.
an an n
.A
Germany, 1997.
F. M. White. Fluid Mechanics. McGraw-Hill, New York, 2010.
m
O. C. Zienkiewicz and R. Taylor. The finite element method. McGraw-hill London, 1977.
d
uu
e
X Th
er
tin
on
.O
C
X. Oliver and C. Agelet de Saracibar Continuum Mechanics for Engineers.Theory and Problems
doi:10.13140/RG.2.2.25821.20961