Académique Documents
Professionnel Documents
Culture Documents
CONTINUUM MECHANICS
FOR ENGINEERS
141 movies
How to interact with the media ...
1 Choose the slide with the subject you are interested in
How to interact with the media ...
2 To watch the movie explaining the slide, click on
the link-to-YouTube icon
How to interact with the media ...
3 To reach the book-section dealing with the slide,
click on the link-to-book icon
How to interact with the media ...
4 To return to the slide, click on the link-to-slide icon
Acknowledgements
Prof. Carlos Agelet (CIMNE/UPC)
Dr. Manuel Caicedo (CIMNE/UPC)
Dr. Eduardo Car (CIMNE)
Prof. Eduardo Chaves (UCLM)
Dr. Ester Comellas (CIMNE)
Dr. Alex Ferrer (CIMNE/UPC)
Prof. Alfredo Huespe (CIMNE/UNL/UPC)
Dr. Oriol Lloberas-Valls (CIMNE/UPC)
Dr. Julio Marti (CIMNE)
2
Overview (cont’d)
1.4. Time Derivatives Lecture 6
1.4.1. Material and Local Derivatives
1.4.2. Convective Rate of Change Lecture 7
1.4.3. Example Lecture 8
3
Overview (cont’d)
1.7. Trajectory or Pathline
1.7.1. Equation of the Trajectories
1.7.2. Example
1.8. Streamlines Lecture 11
1.8.1. Equation of the Streamlines
1.8.2. Trajectories and Streamlines
1.8.3. Example
1.8.4. Streamtubes
4
1.1 Definition of the Continuous
Medium
Ch.1. Description of Motion
5
The Concept of Continuum
Microscopic scale:
Matter is made of atoms which may be grouped in
molecules.
Matter has gaps and spaces.
Macroscopic scale:
Atomic and molecular discontinuities are disregarded.
Matter is assumed to be continuous.
6
Continuous Medium or Continuum
Matter is studied at a macroscopic scale: it completely
fills the space, there exist no gaps or empty spaces.
7
Exceptions to the Continuous Medium
8
Continuum Mechanics
Study of the mechanical behavior of a continuous
medium when subjected to forces or displacements, and
the subsequent effects of this medium on its
environment.
It divides into:
General Principles: assumptions and consequences
applicable to all continuous media.
Constitutive Equations: define the mechanical behavior of
a particular idealized material.
9
1.2 Equations of Motion
Ch.1. Description of Motion
10
Material and Spatial points,
Configuration
A continuous medium is formed by an infinite number of
particles which occupy different positions in space during
their movement over time.
MATERIAL POINTS: particles
SPATIAL POINTS: fixed spots in space
11
Configurations of the Continuous
Medium
Ω0: non-deformed (or reference) Ω or Ωt: deformed (or present)
configuration, at reference time t0. configuration, at present time t.
Γ0 : non-deformed boundary. Γ or Γt : deformed boundary.
X : Position vector of a particle at
reference time. x : Position vector of the same
particle at present time.
ϕ ( X,t ) Γ
t0= 0 → reference time t ∈ [ 0, T ] → current time
Γ0
Initial, reference Ω
or undeformed
configuration
Ω0 X x
Present or deformed
configuration
12
Material and Spatial Coordinates
The
position vector of a given particle can be
expressed in:
Non-deformed or Reference Configuration
X1 X
[ X] X=
= 2
Y ≡ material coordinates (capital letter)
X Z
3
13
Equations of Motion
The motion of a given particle is described by the evolution
of its spatial coordinates (or its position vector) over time.
= x ϕ= (particle label, t ) x ( particle label, t )
= xi ϕi ( particle label, t ) i ∈ {1, 2,3}
14
Inverse Equations of Motion
The inverse equations of motion give the material
coordinates as a function of the spatial ones.
ϕ ( X,t ) Γ
Γ0
Ω0 X ϕ −1 ( x, t ) x
not
= X ϕ= −1
( x, t ) X ( x, t )
= X i ϕi −1 ( x1 , x2 , x3 , t ) i ∈ {1, 2,3}
15
Mathematical restrictions for φ and φ-1
defining a “physical” motion
Consistency condition
ϕ ( X, 0 ) = X , as X is the position vector for t=0
Continuity condition
ϕ ∈ C , φ is continuous with continuous derivatives
1
Biunivocity condition
φ is biunivocal to guarantee that two particles do not occupy
simultaneously the same spot in space and that a particle does not
occupy simultaneously more than one spot in space.
Mathematically: the “Jacobian” of the motion’s equations should
be different from zero: ∂ϕ ( X, t ) ∂ϕi
= J = det ≠0
∂X ∂X j
∂X ∂X j (to be proven)
16
Example
The spatial description of the motion of a continuous medium is given by:
= x1 X= e 2t
x Xe 2t
1
x ( X, t=
) ≡ x2 X 2e−2t = ≡ y Ye −2t
x = + 2t z = + 2t
3 5 X 1t X 3 e 5 Xt Ze
17
x1 = X 1e 2t
Example - Solution
x ( X, t ) ≡ x2 = X 2 e −2t
x 5 X t + X e 2t
=
3 1 3
18
x1 = X 1e 2t
Example - Solution
x ( X, t ) ≡ x2 = X 2 e −2t
x 5 X t + X e 2t
=
3 1 3
X 1 = x1e −2t
X ≡ ϕ −1 ( x, t ) =
X2 = x2 e 2t
X
= −2 t −4 t
3 x3e − 5tx1e
19
1.3 Descriptions of Motion
Ch.1. Description of Motion
20
Descriptions of Motion
Themathematical description of the particle properties
can be done in two ways:
21
Material or Lagrangian Description
The physical properties are described in terms of the
material coordinates and time.
22
Spatial or Eulerian Description
The physical properties are described in terms of the spatial
coordinates and time.
23
Example
The equation of motion of a continuous medium is:
=x X − Yt
x =x ( X, t ) ≡ y =Xt + Y
z = − Xt + Z
Find the spatial description of the property whose material description is:
X +Y + Z
ρ ( X,Y,Z,t ) =
1+ t2
24
=x X − Yt
Example - Solution x =x ( X, t )
≡ y =Xt + Y
z =
− Xt + Z
J =1 + t 2 > 0
25
=x X − Yt
Example - Solution x =x ( X, t )
≡ y =Xt + Y
z =
− Xt + Z
y − xt x + xt 2 + yt − xt 2 x + yt
X =x + Yt =x + 2
t= =
1 + t 1 + t 2
1+ t2
x + yt z + zt 2 + xt + yt 2
z =− Xt + Z ⇒ Z =z + Xt =z + 2
t=
1+ t 1+ t2
x + yt
X =
1+ t2
y − xt
X ≡ ϕ ( x, t ) =
−1
Y =2
1+ t
z + zt 2 + xt + yt 2
Z =
1+ t2
26
Example - Solution
Calculate the property in its spatial description:
x + yt
X =
1+ t2
y − xt
X ≡ ϕ ( x, t ) =
−1
Y =2
1+ t
z + zt 2 + xt + yt 2
Z =
1+ t2
x + yt y − xt z + zt + xt + yt
2 2
+
2
+
2
X + Y + Z 1+ t 1+ t 1+ t2 x + y + yt + yt 2 + z + zt 2
( X,Y,Z,t ) =
ρ= =
1+ t2 1+ t2 (1 + t 2 )
2
X +Y + Z x + y (1 + t + t 2 ) + z (1 + t 2 )
ρ ( X,Y,Z,t ) = ⇒ ρ ( x, y,z,t )
=
1+ t2 (1 + t )2 2
27
1.4 Time Derivatives
Ch.1. Description of Motion
28
Material and Local Derivatives
The
time derivative of a given property can be defined
based on the:
Material Description Γ(X,t) TOTAL or MATERIAL DERIVATIVE
Variation of the property w.r.t. time following a specific particle
in the continuous medium.
∂Γ ( X,t ) partial time derivative of the
material derivative ≡ →
∂t material description of the propery
29
Convective Derivative
Remember: x=x(X,t), therefore, γ(x,t)=γ(x(X,t),t)=Γ(X,t)
The material derivative can be computed in terms of
spatial descriptions:
d
not not D ∂Γ( X, t )
material derivative = → γ ( x, t )
= =γ ( x, t ) =
dt Dt ∂t
∂γ ( x, t ) ∂γ ∂xi ∂γ ( x, t ) ∂γ ∂x
γ ( x ( X, t ) ,=
t)
d
= + ⋅ = + ⋅ =
∂t ∂x ∂t ∂t ∂x ∂t
i
dt
∇ γ ∂x ∇γ ( x,t ) v ( x,t )
∂t
v (x,t )⋅∇γ (x,t )
i i
31
Example
Given the following equation of motion:
x = X + Yt + Zt
x(X, t ) ≡ y = Y + 2 Zt
z = Z + 3 Xt
32
x = X + Yt + Zt
Example - Solution
x(X, t ) ≡ y = Y + 2 Zt
z = Z + 3 Xt
ρ(x, t ) = 3x + 2 y + 3t
d ρ ( x, t ) ∂ρ ( X, t )
= =3 + 3Y + 7 Z
dt ∂t
( )
x = x X,t
33
x = X + Yt + Zt
Example - Solution
x(X, t ) ≡ y = Y + 2 Zt
z = Z + 3 Xt
ρ(x, t ) = 3x + 2 y + 3t
3 X
( ) ( ) ( )
T T
∇ρ ( x= ∂ρ x , t ∂ρ x , t ∂ρ x , t ∂ ∂ ∂
,t) , , = ∂x ( 3 x + 2 y + 3t ) , ∂y ( 3 x + 2 y + 3t ) , ∂z ( 3 x + 2 y + 3t )=
∂ x ∂ y ∂ z
3
= [3,= 2, 0] 2
T
0
34
x = X + Yt + Zt
Example - Solution
x(X, t ) ≡ y = Y + 2 Zt
z = Z + 3 Xt
ρ(x, t ) = 3x + 2 y + 3t
Option #2
The material derivative is obtained:
3
d ρ ( x, t )
=3 + [Y + Z , 2 Z , 3 X ] 2 =3 + 3Y + 3Z + 4 Z
dt
x = x( X,t ) v
T
0
ρ
v⋅(ρ )
d ρ ( x, t )
=3 + 3Y + 7 Z
dt x=x X,t
( )
35
1.5 Velocity and Acceleration
Ch.1. Description of Motion
36
Velocity
Time derivative of the equations of motion.
Material description of the velocity: REMARK
Time derivative of the equations of motion Remember the
∂x ( X, t ) equations of motion
(
V X, t ) = are of the form:
∂t
not
V X, t ∂xi ( X, t ) = ( X, t ) x ( X, t )
x ϕ=
i ( )
= i ∈ 1, 2,3
∂t
37
Acceleration
Material time derivative of the velocity field.
Material description of acceleration:
Derivative of the material description of velocity:
∂V ( X, t )
(
A X, t ) =
∂t
A X, t ∂Vi ( X, t )
i ( )
= i ∈ 1, 2,3
∂t
Spatial description of acceleration:
A(X,t) is expressed in terms of x using the inverse equations of motion:
A ( X ( x, t ) , t ) a ( x, t )
Or a(x,t) is obtained directly through the material derivative of v(x,t):
dv ( x, t ) ∂v ( x, t )
=
a ( x, t ) = + v ( x, t ) ⋅∇v ( x, t )
dt ∂t
a ( x, t ) = dvi ( x, t ) = ∂vi ( x, t ) + v ( x, t ) ⋅ ∂vi ( x, t ) i ∈ 1, 2,3
i dt ∂t
k
∂xk
38
Example
Consider a solid that rotates at a constant angular velocity ω and has the
following equation of motion:
= x R sin (ωt + φ)
x ( R, φ , t ) →
label of = y R cos (ωt + φ)
particle
→ (non - canonical equations of motion)
Find the velocity and acceleration of the movement described in both, material
and spatial forms.
39
x = R sin(ωt + φ)
Example - Solution
y = R cos(ωt + φ )
40
=x X cos ( ωt ) + Y sin ( ωt )
Example - Solution y =− X sin ( ωt ) + Y cos ( ωt )
( ωt ) Y ( cos 2 ( ωt ) + sin 2 ( ωt ) )
x sin ( ωt ) + y cos=
=1
42
= x X cos ( ωt ) + Y sin ( ωt )
Example - Solution x ( X, t ) →
y =− X sin ( ωt ) + Y cos ( ωt )
∂x ( X, t )
Velocity in material description is obtained from V ( X, t ) =
∂t
∂x ∂
=
∂x ( X, t ) ∂t ∂t
( X cos ( ωt ) + Y sin ( ωt ) )
V ( X, t ) =
=
∂t ∂y = ∂
( − X sin ( ωt ) + Y cos ( ωt ) )
∂t ∂t
43
=X x cos ( ωt ) − y sin ( ωt )
Y x sin ( ωt ) + y cos ( ωt )
=
Example - Solution x X cos ( ωt ) + Y sin ( ωt )
=
y =− X sin ( ωt ) + Y cos ( ωt )
v x ω y
v (=
x, t ) =
v y −ω x
44
− X ω sin ( ωt ) + Y ω cos ( ωt )
Example - Solution V ( X, t ) =
− X ω cos ( ω t ) − Y ω sin ( ω t )
∂V ( X, t )
Acceleration in material description is obtained applying: A ( X, t ) =
∂t
∂Vx
∂V ( X, t ) ∂t
=− X ω2
cos ( ω t ) − Y ω2
sin ( ωt )
A ( X, t ) =
=
∂t ∂Vy = X ω2 sin ( ωt ) − Y ω2 cos ( ωt )
∂t
Ax X cos ( ωt ) + Y sin ( ωt )
2
A ( X, t ) = = −ω
Ay − X sin ( ω t ) + Y cos ( ω t )
45
X x cos ( ωt ) − y sin ( ωt )
Example - Solution =
Y x sin ( ωt ) + y cos ( ωt )
=
( X ( x, t ) , t )
a ( x, t ) A=
=
( x cos ( ωt ) − y sin ( ωt ) ) cos ( ωt ) + ( x sin ( ωt ) + y cos ( ωt ) ) sin ( ωt )
= −ω 2
− ( x cos ( ωt ) − y sin ( ωt ) ) sin ( ωt ) + ( x sin ( ωt ) + y cos ( ωt ) ) cos ( ωt )
=0 =1
ax −ω x
2
a (=
x, t ) = 2
a y −ω y
46
v x ω y
Example - Solution v (=
x, t ) =
v y −ω x
−ω x
2
a ( x, t ) = 2
−ω y
47
1.6 Stationarity and Uniformity
Ch.1. Description of Motion
48
Stationary properties
A property is stationary when its spatial description is not
dependent on time. REMARK
χ ( x,t ) = χ ( x ) In certain fields, the
term steady-state is
The local derivative of a stationary property is zero. more commonly used.
∂χ ( x, t )
=χ ( x, t ) χ=
(x) 0
∂t
49
Example
Consider a solid that rotates at a constant angular velocity ω and has the
following equation of motion:
= x R sin (ωt + ϕ )
= y R cos (ωt + ϕ )
We have obtained:
Velocity in spatial description
v x ω y stationary
v (=
x, t ) =
v y −ω x
Velocity in material description
Vx − X ω sin ( ωt ) + Y ω cos ( ωt )
V (X
=, t ) =
Vy − X ω cos ( ω t ) − Y ω sin ( ω t )
50
Uniform properties
A property is uniform when its spatial description is not
dependent on the spatial coordinates.
χ ( x, t ) = χ ( t )
51
1.7 Trajectory (path-line)
Ch.1. Description of Motion
52
Trajectory or pathline
A trajectory or pathline is the locus of the positions
occupied by a given particle in space throughout time.
REMARK
A trajectory can also be defined as
the path that a particle follows
through space as a function of time.
53
Equation of the trajectories
The equation of a given particle’s trajectory is obtained
particularizing the equation of motion for that particle, which is
identified by it material coordinates X*.
x ( t ) ϕ=( X , t ) X = X * φ (t )
=
i (t )
x= ϕ i ( X, t )= φi (t ) i ∈
X = X*
55
Example - Solution
We integrate the velocity field:
dx ( t )
= v x ( x, t )= ωy
dx ( t )
= v ( x, t )
dt
dt dy ( t ) = v ( x, t ) = −ωx
dt y
d 2 y (t ) dx ( t )
2
= −ω = −ω2
y (t ) y′′ + ω2 y = 0
dt dt
56
Example - Solution
The characteristic equation: r 2 + ω2 = 0
Has the characteristic solutions: rj =± i ω j ∈ {1, 2}
And the solution of the problem is:
(t ) Real Part {Z1eiwt + Z 2 e −=
y= iwt
} C1 cos ( ωt ) + C2 sin ( ωt )
And, using dy = −ωx , we obtain
dt
=− ( −C1ω sin ( ωt ) + C2 ω cos ( ωt ) )
1 dy 1
x =−
ω dt ω
57
x ( C1 , C=2,t) C1 sin ( ωt ) − C2 cos ( ωt )
Example - Solution
y ( C1 , C=
2,t) C1 cos ( ωt ) + C2 sin ( ωt )
58
1.8 Streamline
Ch.1. Description of Motion
59
Streamline
The streamlines are a family of curves which, for every
instant in time, are the velocity field envelopes.
time – t0 time – t1
Y Y
REMARK
Two streamlines can
never cut each other.
Is it true?
X X
Streamlines are defined for any given time instant and change with
the velocity field.
REMARK
The envelopes of vector field are the curves whose tangent vector
at each point coincides (in direction and sense but not necessarily
in magnitude) with the corresponding vector of the vector field.
60
Equation of the Streamlines
The equation of the streamlines is of
the type:
vz
d x d y dz dx
= = = d= λ ( ds ) = v vy
vx v y vz dλ
vx
Where each group ( C1′, C2′ , C3′ ) identifies a streamline x(λ) whose
points are obtained assigning values to the parameter λ.
For each time instant t* a new family of curves is obtained.
61
Trajectories and Streamlines
For a stationary velocity field, the trajectories and the
streamlines coincide – PROOF:
1. If v(x,t)=v(x):
Eq. trajectories:
dx ( t )
= v=
dt
( x (t ) , t ) x φ ( C1 , C2 , C3 , t )
Eq. streamlines:
dx ( λ )
= v=
dλ
( x ( λ ) , t *) x φ ( C1 , C2 , C3 , λ , t *)
62
Trajectories and Streamlines
For a stationary velocity field, the trajectories and the
streamlines coincide – PROOF:
2. If v(x,t)=v(x) the envelopes (i.e., the streamline) of the field do not
vary throughout time.
63
Trajectories and Streamlines
The inverse is not necessarily true: if the trajectories and the
streamlines coincide, the velocity field is not necessarily
stationary – COUNTER-EXAMPLE: at
Given the (non-stationary) velocity field: v ( t ) = 0
0
64
Example
Consider the following velocity field:
xi
=vi i ∈ {1, 2,3}
1+ t
Obtain the equation of the trajectories and the streamlines associated to this
vector field.
Do they coincide? Why?
65
Example - Solution
dx ( t )
= v (x (t ) , t )
dt
Eq. trajectories:
dxi ( t ) v x ( t ) , t i ∈
= i ( )
dt
Introducing the velocity field and rearranging:
dxi xi dxi dt
= i ∈ = i ∈
dt 1 + t xi 1 + t
The solution:
xi = Ci (1 + t ) i ∈
66
Example - Solution
dx ( λ )
= v ( x ( λ ) , t *)
dλ
Eq. streamlines: dx t
i ( ) v x ( λ ) , t * i ∈
= i ( )
d λ
Introducing the velocity field and rearranging:
dxi xi dxi d λ
= i ∈ = i ∈
dλ 1+ t xi 1 + t
Integrating both sides of the expression:
λ
+ Ki
λ
= 1+ t Ki 1+ t
xi e= e e
1 1 λ
∫ xi i ∫ 1 + t d λ
dx = ln=
xi
1+ t
+ Ki
i ∈ = Ci
The solution: λ
1+ t
=xi Ci e
i ∈
67
Streamtube
A streamtube is a surface composed of streamlines
which pass through the points of a closed contour fixed
in space.
69
1.9 Control and Material Surfaces
Ch.1. Description of Motion
81
Control Surface
A control surface is a fixed surface in space which does not
vary in time.
=Σ: { x=
f ( x, y, z ) 0}
82
Material Surface
A material surface is a mobile surface in the
space constituted always by the same particles.
In the reference configuration, the surface Σ0 will be defined in terms
of the material coordinates:
=Σ0 : { X=
F ( X , Y , Z ) 0}
The set of particles (material points)
belonging the surface are the same at all times
In spatial description F ( X ,=
Y , Z ) F ( X (x, t ), Y (x, t ), Z (=
x, t ) ) f=
(x, t ) f ( x, y, z , t )
=Σt : { x=
f ( x, y , z , t ) 0 }
The set of spatial points belonging to the the surface depends on time
The material surface moves in space
84
Material Surface
Necessary and sufficient condition for a mobile surface in space,
implicitly defined by the function f ( x, y, z, t ) , to be a material
surface is that the material derivative of the function is zero:
Necessary: if it is a material surface, its material description does not depend
on time:
d ∂F ( X)
f ( x, t ) → f ( x( X, t ), t ) =
F ( X, t ) 0 =f ( x, t ) = = 0
dt ∂t
Sufficient: if the material derivative of f(x,t) is null:
d ∂F ( X, t )
f ( x, t ) → f ( x( X, t ), t ) =F ( X, t ) 0 = f ( x, t ) = F ( X, t ) ≡ F ( X)
dt ∂t
The surface Σt := { x f ( x, t ) = 0 } = { X F ( X ) = 0 }
contains always the same set the of particles (it is a material surface)
85
Control Volume
A control volume is a group of fixed points in space
situated in the interior of a closed control surface, which
does not vary in time.
V := { x | f (x ) ≤ 0}
REMARK
The function f(x) is defined
so that f(x)<0 corresponds
to the points inside V.
87
Material Volume
A material volume is a (mobile) volume enclosed inside a
material boundary or surface.
In the reference configuration, the volume V0 will be defined in terms
of the material coordinates:
=V0 : { X | F ( X ) ≤ 0 }
The set of spatial points belonging to the the volume depends on time
The material volume moves in space along time
89
Material Volume
A material volume is always constituted by the same
particles. This is proved by reductio ad absurdum:
90
Chapter 1
Description of Motion
rs
ee
s gin
1.1 Definition of the Continuous Medium
t d le En
r
A continuous medium is understood as an infinite set of particles (which form
ba
ge ro or
eS m
part of, for example, solids or fluids) that will be studied macroscopically, that
ci
f
is, without considering the possible discontinuities existing at microscopic level
ra
C d P cs
(atomic or molecular level). Accordingly, one admits that there are no discon-
b
a
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.
s gin
uous medium at time t, denoted by Ωt , is defined as the locus of the positions
t d le En
occupied in space by the material points (particles) of the continuous medium at
r
the given time.
ba
ge ro or
eS m
A certain time t = t0 of the time interval of interest is referred to as the ref-
ci
f
erence time and the configuration at this time, denoted by Ω0 , is referred to as
ra
C d P cs
initial, material or reference configuration1 .
b
a
i
Consider now the Cartesian coordinate system (X,Y, Z) in Figure 1.1 and the
an an n
le
the position vector X of a particle occupying a point P in space (at the reference
liv or ec
.A
where the components (X1 , X2 , X3 ) are referred to as material coordinates (of the
d
uu
er
⎡ ⎤
tin
X1
⎢ ⎥ de f
on
.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.
s gin
Mathematically, this requires the definition of a function that provides for each
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-
r
ticle can be chosen as the label that univocally characterizes it and, thus, the
ba
ge ro or
eS m
equation of motion
ci
f
ra
not
x = ϕ (particle,t) = ϕ (X,t) = x (X,t)
C d P cs
b
a
(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
Remark 1.1. There are different alternatives when choosing the la-
C
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
• The Jacobian of the transformation J = det = > 0.
∂X ∂X
ee
s gin
The physical interpretation of this condition (which will be studied later) is
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.
r
ba
ge ro or
eS m
ci
f
ra
Remark 1.2. The equation of motion at the reference time t = 0 re-
C d P cs
b
a
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
∂ (xyz)
le
∂ xi
liv or ec
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
not
7The two-index operator Delta Kronecker = δi j is defined as δi j = 0 when i = 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
not ⎢ ⎥ ⎢ ⎥
s gin
x (X,t) ≡ ⎣ x2 = X2 e−2t ⎦ = ⎣ y = Y e−2t ⎦
x3 = 5X1t + X3 e2t z = 5Xt + Ze2t
t d le En
r
ba
Obtain the inverse equation of motion.
ge ro or
eS m
ci
f
ra
Solution
C d P cs
b
a
i
The determinant of the Jacobian is computed as
an an n
y ha
∂ x1 ∂ x1 ∂ x1
le
liv or ec
∂ X1 ∂ X2 ∂ X3 e2t 0 0
M
.A
∂ xi ∂ x
∂ x2 ∂ x2 = 0 e−2t 0 = e2t = 0.
J = =
2
m
∂ Xj ∂ X1 ∂ X2 ∂ X3
d
5t 0 e 2t
uu
∂ x3 ∂ x3 ∂ x3
e
X Th
∂X ∂ X2 ∂ X3
er
tin
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
not ⎢ ⎥ ⎢ ⎥
X = ϕ −1 (x,t) ≡ ⎢ X
⎣ ⎦ ⎣
2
⎥=⎢ x 2 e 2t ⎥.
⎦
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
described by a certain function ρ (•,t) : R3 × R+ → R+ , where the argument (•)
ee
in ρ (•,t) represents the material coordinates,
s gin
ρ = ρ (X,t) = ρ (X1 , X2 , X3 ,t) . (1.7)
t d le En
Here, if the three arguments X ≡ (X1 , X2 , X3 ) are fixed, a specific particle is being
r
ba
ge ro or
eS m
followed (see Figure 1.3) and, hence, the name of material description.
ci
f
ra
C d P cs
1.3.2 Spatial Description
b
a
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
s gin
X +Y + Z
t d le En
ρ (X,Y, Z,t) = .
1 + t2
r
ba
ge ro or
Solution
eS m
ci
f
ra
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
a
i
an an n
x=X
y ha
not
x = X (X, 0) ≡ y = Y .
le
liv or ec
z=Z
M
.A
∂x ∂x ∂x
uu
1 −t 0
X Th
∂ X ∂Y ∂Z
er
tin
∂ xi ∂ y ∂ y
∂y
J = = = t 1 0 = 1 + t 2 = 0
∂ X j ∂ X ∂Y
on
.O
∂Z −t 0 1
∂z
C
∂z ∂z
©
∂ X ∂Y ∂Z
and the inverse equation of motion is given by
⎡ ⎤
x + yt
⎢X = ⎥
⎢ 1 + t2 ⎥
⎢ ⎥
not ⎢ y − xt ⎥
X (x,t) ≡ ⎢ Y = ⎥.
⎢ 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
ee
s gin
1.4 Time Derivatives: Local, Material and Convective
t d le En
The consideration of different descriptions (material and spatial) of the proper-
r
ties of the continuous medium leads to diverse definitions of the time derivatives
ba
ge ro or
eS m
of these properties. Consider a certain property and its material and spatial de-
ci
scriptions,
f
ra
Γ (X,t) = γ (x,t) ,
C d P cs
(1.10)
b
a
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
∂t
.O
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
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
s gin
= + = + · =
dt ∂t ∂ xi ∂t ∂t ∂ x
∂t
(1.13)
t d le En
∂ γ (x,t) ∂ γ v (x,t)
= + · v (x,t) ,
r
∂t ∂x
ba
ge ro or
eS m
ci
f
where the definition of velocity as the derivative of the equation of motion (1.5)
ra
C d P cs
with respect to time has been taken into account,
b
a
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)
d
uu
dt ∂t
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
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,
s gin
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
r
description with respect to time,
ba
ge ro or
eS m
ci
∂ρ
f
ra
= 3Y + 7Z + 3 .
C d P cs
∂t
b
a
i
an an n
dρ ∂ρ
M
= + v · ∇ρ
.A
with
dt ∂t
m
∂ρ ∂x
d
∂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)
⎪
s gin
⎪ ∂
⎩ Vi (X,t) = i (X,t) i ∈ {1, 2, 3}
x
∂t
t d le En
and, if the inverse equation of motion X = ϕ −1 (x,t) is known, the spatial de-
r
ba
ge ro or
scription of the velocity can be obtained as
eS m
ci
f
ra
v (x,t) = V (X (x,t) ,t) . (1.17)
C d P cs
b
a
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)
⎪ ∂V
⎩ Ai (X,t) = i (X,t) i ∈ {1, 2, 3}
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
Solution
f
ra
C d P cs
b
a
The equation of motion can be rewritten as
i
an an n
y ha
.A
and, since for t = 0, X = R sin φ and Y = R cos φ , the canonical form of the
m
uu
e
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
not ωy
v (x,t) = V (X (x,t) ,t) ≡ .
−ωx
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)
s gin
Acceleration in spatial description:
t d le En
Replacing the canonical form of the equation of motion into the material
r
description of the acceleration results in
ba
ge ro or
eS m
ci
f −ω 2 x
ra
not
a (x,t) = A (X (x,t) ,t) ≡ .
C d P cs
b
a
−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) =
∂t
m
dt ⎡ ⎤
d
uu
∂
e
∂ ⎢ ⎥
⎢ ∂x ⎥
X Th
ωy
er
not
tin
≡ + ωy , −ωx ⎢ ⎥ ωy , −ωx ,
∂t −ωx ⎣ ∂ ⎦
on
.O
∂y
⎡ ⎤
C
∂ ∂
⎢ (ωy) (−ωx) ⎥ −ω 2x
0 ⎢ ⎥
+ 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)
s gin
v (x,t) = v (x) ⇐⇒ =0. (1.20)
∂t
t d le En
r
ba
Remark 1.5. The non-dependence on time of the spatial description
ge ro or
eS m
ci
(stationarity) assumes that, for a same point in space, the property
f
ra
being considered does not vary along time. This does not imply that,
C d P cs
b
a
for a same particle, such property does not vary along time (the ma-
i
an an n
v (x,t) is stationary,
le
v (x,t) ≡ v (x) = v (x (X,t)) = V (X,t) ,
liv or ec
M
.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
The velocity field in Example 1.4 is v (x) ≡ [ωy , −ωx]T . Therefore, it is a
not
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.
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
dv (x) ∂ v (x)
a (x) = = + v (x) · ∇v (x) = v (x) · ∇v (x) ,
M
.A
dt ∂t
m
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.
s gin
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
r
ba
ge ro or
can be obtained by formulating the system of differential equations that imposes
eS m
ci
that, for each point in space x, the velocity vector is the time derivative of the
f
ra
parametric equation of the trajectory defined in (1.21), i.e.,
C d P cs
b
a
⎧
i
⎪ dx (t)
an an n
⎪
⎨ = v (x (t) , t) ,
y ha
dt
M
.A
e
er
tin
(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
Example 1.6 – Given the velocity field in Example 1.5, v (x) ≡ [ωy , −ωx]T ,
not
Solution
Using expression (1.22), one can write
⎧
⎪
⎪ dx (t)
dx (t) ⎨ = vx (x,t) = ωy ,
= v (x,t) =⇒ dt
⎪
⎩ dy (t) = vy (x,t) = −ωx .
dt ⎪
rs
dt
ee
s gin
This system of equations is a system with crossed variables. Differentiating
the second equation and replacing the result obtained into the first equation
t d le En
yields
d 2 y (t) dx (t)
= −ω = −ω 2 y (t) =⇒ y + ω 2 y = 0 .
r
ba
ge ro or
dt 2 dt
eS m
ci
f
The characteristic equation of this second-order differential equation is
ra
C d P cs
r2 + ω 2 = 0 and its characteristic solutions are r j = ±iω j ∈ {1, 2}.
b
a
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
The solution for x (t) is obtained from dy/dt = −ωx , which results in
M
.A
er
tin
.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
s gin
that same point in space.
t d le En
r
Remark 1.6. In general, the velocity field (in spatial description) will
ba
ge ro or
eS m
be different for each instant of time (v ≡ v (x,t)). Therefore, one
ci
f
must speak of a different family of streamlines for each instant of
ra
C d P cs
time (see Figure 1.6).
b
a
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 (C1 ,C2 ,C3 ), provides the parametric expression of the streamlines,
ee
x = φ (C1 ,C2 ,C3 , λ , t ∗ ) ,
s gin
(1.27)
xi = φi (C1 ,C2 ,C3 , λ , t ∗ ) i ∈ {1, 2, 3} .
t d le En
Each triplet of integration constants (C1 ,C2 ,C3 ) identifies a streamline whose
r
ba
ge ro or
points, in turn, are obtained by assigning values to the parameter λ . For each
eS m
ci
time t ∗ a new family of streamlines is obtained.
f
ra
C d P cs
b
a
i
an an n
tories and streamlines coincide. This can be proven from two differ-
le
liv or ec
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.
s gin
1.9.1 Equation of the Streamtube
t d le En
Streamlines constitute a family of curves of the type
r
ba
x = f (C1 ,C2 ,C3 , λ , t) .
ge ro or
(1.28)
eS m
ci
f
ra
The problem consists in determining, for each instant of time, which curves
C d P cs
b
a
of the family of curves of the streamlines cross a closed line, which is fixed in the
i
space Γ , whose mathematical expression parametrized in terms of a parameter s
an an n
y ha
is
Γ := x = g (s) .
le
(1.29)
liv or ec
To this aim, one imposes, in terms of the parameters λ ∗ and s∗ , that a same point
M
.A
d
(1.30)
X Th
can be isolated,
C
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
ee
s gin
1.10 Streaklines
t d le En
r
∗
ba
ge ro or
Definition 1.9. A streakline, relative to a fixed point in space x
eS m
named spill point and at a time interval ti ,t f named spill period,
ci
f
ra
is the locus of the positions occupied at time t by all the particles
C d P cs
b
a
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)
s gin
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
r
time.
ba
ge ro or
eS m
ci
f
ra
C d P cs
Example 1.7 – Given the equation of motion
b
a
⎧
i
an an n
⎨ x = (X +Y )t 2 + X cost ,
y ha
⎩ y = (X +Y ) cost − X ,
le
liv or ec
M
.A
obtain the equation of the streakline associated with the spill point x∗ = (0, 1)
for the spill period [t0 , +∞).
m
d
uu
e
Solution
X Th
er
tin
The material coordinates of a particle that has occupied the spill point at time
on
τ are given by
.O
⎧
⎪ −τ 2
C
⎪
⎨ X = ,
0 = (X +Y ) τ + X cos τ
2
©
=⇒ τ 2 + cos2 τ
1 = (X +Y ) cos τ − X ⎪
⎩ Y = τ + cos τ .
⎪ 2
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 ⎬
τ + cos τ 2
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 −τ 2
not ⎢
x= 2 t + 2
2 cost ⎥
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
s gin
fact that, in the stationary case, the trajectory follows the envelope of
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.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
1.11 Material Surface
i
an an n
y ha
le
liv or ec
.A
d
uu
e
X Th
.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
surface in space, defined implicitly by a function f (x, y, z,t) = 0, to
s gin
be material (to be always constituted by the same particles) is that
the material derivative of f (x, y, z,t) is null,
t d le En
r
d f (x,t) ∂ f
ba
ge ro or
= +v·∇f = 0 ∀x ∈ Σt ∀t .
eS m
∂t
ci
dt
f
ra
C d P cs
The condition is necessary because, if the surface is a material sur-
b
a
face, its material description will not depend on time (F ≡ F (X))
i
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
.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 .
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
M
.A
∂f ∂f ∂η
uu
df
e
= + v · ∇ f where =− and
X Th
∂t ∂t ∂t
er
dt
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
rs
Obviously, a control surface is occupied by the different particles of the contin-
ee
uous medium along time (see Figure 1.10).
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
.A
m
er
tin
.O
material surface.
C
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.
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
s gin
a) A stationary velocity field implies that the spatial description of velocity does
t d le En
not depend on time,
∂ v (x,t)
= 0 =⇒ v (x) .
r
ba
∂t
ge ro or
eS m
ci
The acceleration is the material derivative of the velocity, therefore
f
ra
C d P cs
b
a
∂ v (x,t)
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)
s gin
a (x,t) = = + v (x,t) · ∇v (x,t) ,
dt ∂t
t d le En
where
∂ v not ! T
r
≡ 0, z e t − e−t , 0 and
ba
ge ro or
∂t
eS m
⎡ ⎤
ci
f
ra
∂
C d P cs
⎢ ⎥ ⎡ ⎤
b
a
⎢ ∂x ⎥
i
⎢ ⎥ 1 0 0
an an n
not ⎢ ∂ ⎥ ! ⎢ ⎥
y ha
∇v ≡ ⎢ ⎢ ⎥ x−z , z e t + e−t , 0 =⎢ 0⎥
⎥ ⎣ 0 0 ⎦ , such that
le
⎢ ∂y ⎥
liv or ec
⎢ ⎥ −1 (e t + e−t ) 0
⎣ ∂ ⎦
M
.A
∂z
m
0]T .
not
v · ∇v ≡ [x − z , 0,
uu
e
X Th
er
tin
.O
not ! T
a ≡ x − z , z e t − e−t , 0
C
and, for the given point at the given instant of time, the acceleration is
not T
a (x = x∗ , t = 2) ≡ 0 , e2 − e−2 , 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
c) An observer located in point (1, 1, 1) that measures the accelerations as the
ee
difference between velocities at this point per unit of time.
s gin
t d le En
Solution
r
ba
a) The spatial description of the acceleration in point x∗ = (1, 1, 1) must be
ge ro or
eS m
ci
obtained,
f
ra
∂ V (X (x∗ ,t) ,t)
C d P cs
a (x = x∗ , t) = A (X (x∗ ,t) , t) = .
b
a
∂t
i
an an n
⎡ ⎤
le
liv or ec
⎢ 0 ⎥
M
.A
⎢ ⎥
∂ x (X,t) ⎢ ⎥
not ⎢ ⎥
V (X,t) ≡ ⎢ 1 ((Y + Z) e t − (Y − Z) e−t ) ⎥ .
m
V (X,t) = =⇒
d
∂t ⎢2 ⎥
uu
⎢ ⎥
e
⎣1 ⎦
X Th
er
−t
((Y + Z) e + (Y − Z) e )
t
tin
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 !
Ay = (Y + Z) e t + (Y − Z) e−t = y 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
z]T
not
a (x,t) ≡ [0 , y,
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.
s gin
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
r
ba
ge ro or
eS m
⎧
ci
⎪
⎪ f
ra
⎪ X =x
C d P cs
⎪
⎪
b
a
⎪
⎨
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
not T
A (X∗ ,t) ≡ 0 , e t , e t .
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
s gin
" ! #
Σt := x | F (x, y, z,t) = e−2t x2 + y2 + z2 e−t −C = 0 ,
2
t d le En
r
ba
where C = 0 is a constant, determine ϕ (t) considering that the particles on this
ge ro or
eS m
ci
surface are always the same.
f
ra
C d P cs
b
a
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 ,
∂t
m
dt
d
uu
e
where
!
X Th
∂F
er
tin
∂t $ %
on
2 T
.O
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
!!
Since C − e−2t x2 + y2 = 0 cannot be satisfied for ∀x and ∀t because C is a
constant, the only possibility left is
ϕ (t) = t .
rs
not
v (x,t) ≡ ze ,t
, vz
1+t
ee
s gin
and the surface ϕ (x,t) = x − z (1 + t) e t + k = 0 (where k is a constant), which
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
r
streamlines.
ba
ge ro or
eS m
ci
f
b) The equation of the streakline and the position of its initial and final points
ra
if the spill point is x∗ and the spill period is t ∈ [t1 ,t2 ].
C d P cs
b
a
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
∂ϕ
X Th
dϕ
er
tin
= + v · ∇ϕ = 0 ∀x ∈ Σt ∀t ,
dt ∂t
on
.O
∂ϕ
∇ϕ ≡ [1 , 0 , −e t (1 + t)]T
C
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
s gin
expression results in
t d le En
dx
= C1 e−t e t = C1 =⇒ x = C1t +C2 .
r
dt
ba
ge ro or
eS m
ci
Finally, the y-component is
f
ra
C d P cs
b
a
dy y
= =⇒ y = C3 (1 + t) .
i
an an n
dt 1+t
y ha
⎧
M
.A
⎨ x (0) = C2 = X
⎪
y (0) = C3 = Y
m
⎪
⎩
d
z (0) = C1 = Z
uu
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
s gin
λ
y = C3 e 1+t .
z = C1 e−λ
t d le En
r
ba
ge ro or
eS m
ci
b) To obtain the equation of the streakline it is enough to take the equation of
f
ra
motion and impose x∗ = x (X, τ), where τ is a time belonging to the spill period.
C d P cs
b
a
⎧
i
an an n
∗
⎨ x = X + Zτ
⎪
y ha
y∗ = Y (1 + τ)
le
⎪
liv or ec
⎩ ∗
z = Ze−τ
M
.A
⎧
uu
X = x∗ − Zτ = x∗ − z∗ τeτ
e
⎪
⎪
⎨
X Th
y∗
er
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 ⎧
⎪ ∗
⎪ ⎨x = x
⎪
s gin
⎨ 1 + t
Initial point: y = y∗ Final point: y = y∗
⎪ + ⎪
t d le En
⎪ 1 t ⎩
⎪
⎪
1
z = z∗
⎩ z = z∗ e (t1 −t)
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
er
tin
on
ii) t ≥ t2
.O
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 (t1 −t) z = z∗ e (t2 −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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
ee
s gin
1.2 – Given the following velocity field in material description
not T
t d le En
v ≡ Ae At X1 , BtX1 , CX3 ,
r
ba
ge ro or
with A, B and C constants, obtain its spatial description and the conditions A, B
eS m
and C must fulfill for the motion to be feasible for 0 < t < ∞.
ci
f
ra
C d P cs
b
a
1.3 – Tracer fluid is injected at point (1, 1, 1) of the interior of a fluid from time
i
an an n
le
x = C1 eλt , y = C2 eλt , z = C3 e2λt
liv or ec
M
.A
determine the equation of the streakline, indicating its initial and final points for
t = 5.
m
d
uu
e
er
tin
not T
v ≡ ye−t , ze t , 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
40 C HAPTER 1. D ESCRIPTION OF M OTION
rs
ee
s gin
t d le En
r
ba
1.6 – A certain motion is defined by the velocity field
ge ro or
eS m
ci
vx = ye−t ; vy = y ;f vz = 0 .
ra
C d P cs
b
a
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
CH.2. DEFORMATION AND
STRAIN
Multimedia Course on Continuum Mechanics
Overview
Introduction Lecture 1
Deformation Gradient Tensor
Material Deformation Gradient Tensor Lecture 2
Inverse (Spatial) Deformation Gradient Tensor
Lecture 3
Displacements Lecture 4
Displacement Gradient Tensors
Strain Tensors
Green-Lagrange or Material Strain Tensor Lecture 5
Euler-Almansi or Spatial Strain Tensor
Variation of Distances
Stretch Lecture 6
Unit elongation
Variation of Angles Lecture 7
2
Overview (cont’d)
Physical interpretation of the Strain Tensors Lecture 8
Material Strain Tensor, E
Spatial Strain Tensor, e Lecture 9
Polar Decomposition Lecture 10
Volume Variation Lecture 11
Area Variation Lecture 12
Volumetric Strain Lecture 13
Infinitesimal Strain
Infinitesimal Strain Theory
Strain Tensors
Stretch and Unit Elongation Lecture 14
Physical Interpretation of Infinitesimal Strains
Engineering Strains
Variation of Angles
3
Overview (cont’d)
Infinitesimal Strain (cont’d)
Polar Decomposition Lecture 15
Volumetric Strain
Strain Rate
Spatial Velocity Gradient Tensor Lecture 16
Strain Rate Tensor and Rotation Rate Tensor or Spin Tensor
Physical Interpretation of the Tensors
Material Derivatives Lecture 17
Other Coordinate Systems
Cylindrical Coordinates Lecture 18
Spherical Coordinates
4
2.1 Introduction
Ch.2. Deformation and Strain
5
Deformation
Deformation: transformation of a body from
a reference configuration to a current configuration.
6
2.2 Deformation Gradient Tensors
Ch.2. Deformation and Strain
7
Continuous Medium in Movement
Ω0: non-deformed (or reference) Ω or Ωt: deformed (or present)
configuration, at reference time t0. configuration, at present time t.
X : Position vector of a particle at x : Position vector of the same particle
reference time. at present time.
ϕ ( X,t )
t0 Q t
Reference or Ωt
dX Q’
non-deformed dx
Ω0 P P’
X x Present or
deformed
8
Fundamental Equation of Deformation
The Equations of Motion:
not
xi i ( X1, X 2 , X 3 , t )
ϕ= xi ( X 1 , X 2 , X 3 , t ) i ∈ {1, 2,3}
not
= ( X, t ) x ( X, t )
x ϕ=
Differentiating w.r.t. X :
∂xi ( X, t ) not
dxi = dX j Fij ( X, t ) dX j i, j ∈ {1, 2,3}
∂X j
∂x ( X, t ) not Fundamental equation
d=
x X F ( X, t ) ⋅ dX
⋅ d= of deformation
∂X
(material) deformation
gradient tensor
9
Material Deformation Gradient Tensor
F(X,t):
is a primary measure of deformation
characterizes the variation of relative placements in the neighbourhood of a
material point (particle).
=dx F ( X, t ) ⋅ dX
10
Inverse (spatial) Deformation Gradient
Tensor
The inverse Equations of Motion:
not
( x1 , x2 , x3 , t ) X i ( x1 , x2 , x3 , t )
X i ϕ= i
−1
i ∈ {1, 2,3}
not
= ( x, t ) X ( x, t )
X ϕ= −1
Differentiating w.r.t. x :
∂X i ( X, t ) not
=dX i = dx j Fij−1 ( x, t ) dx j i, j ∈ {1, 2,3}
∂x j
∂X ( x, t ) not
=
dX =⋅ dx F −1 ( x, t ) ⋅ dx
∂x
11
Inverse (spatial) Deformation Gradient
Tensor
The spatial (or inverse) deformation gradient tensor:
= dX F −1 ( x, t ) ⋅ dx
∂
F ( x, t ) ≡ X(x, t ) ⊗ ∇
−1 REMARK x
The spatial Nabla 1
−1 ∂X i ∂
=
Fij i, j ∈ {1, 2,3} operator is defined as: [∇] =
∂x j ∂ x2
∇≡ eˆ i ∂
∂X 1 ∂X 1 ∂X 1 ∂xi
= ∇T ∂x1 ∂x2 ∂x3 x3
X1
∂ ∂ ∂ ∂X 2 ∂X 2 ∂X 2
F = [ X ⊗ ∇ ] = X 2 ∂x
−1
=
∂x2 ∂x3 ∂x1 ∂x2 ∂x3
X 3 1
∂X 3 ∂X 3 ∂X 3
F-1(x,t): = X
∂x1 ∂x2 ∂x3
is a primary measure of deformation
characterizes the variation of relative placements in the neighbourhood of a
spatial point.
It is not the spatial description of the material deformation gradient tensor
12
Properties of the Deformation
Gradients
The spatial deformation gradient tensor is the inverse of the material
deformation gradient tensor:
∂xi ∂X k ∂xi
= = δ ij F ⋅ F −1 = F −1 ⋅ F = 1
∂X k ∂x j ∂x j
∂x
If there is no motion, =
x X and =
F −1
= F= 1 .
∂X
13
Example
Compute the deformation gradient and inverse deformation gradient tensors
for a motion equation with Cartesian components given by,
X + Y 2t
[ x] Y (1 + t )
=
Zet
1.
Using the results obtained, check that F ⋅ F −1 =
14
X + Y 2t
[ ] ( )
Example - Solution =x Y 1 +
Zet
t
The Cartesian components of the inverse motion equation will be given by,
y 2t
X= x − 2 yt
( + )
2
1 t 1 0
(1 + t )
y
[ X=] ϕ −1 ( x, t =
) =
f (=
( ) x, t ) 0 1 + t
Y F X ( x , =
t ), t 0
1+ t
0 et
Z = ze − t f ( x ,t )
0
15
Example - Solution
The Cartesian components of the inverse deformation gradient tensor are,
2 yt
1 − 0
( + )
2
1 t
1
F −1 ( x, t ) = 0 0
1+ t
0 0 e − t
1:
And it is verified that F ⋅ F −1 =
2 yt 2 yt 2 yt
1 − 0 1 − + 0
0
2 yt
(1+ t) (1 + t ) (1 + t )
2 2 2
1
(1 + t ) 1 0 0
1 1+ t 0 1 0=
F ⋅ F −=
1
0 1 + t 0 ⋅ 0 0= 0 0 = [1]
0 1+ t 1+ t
0 et 0 0 1
0 0
0 e−t 0 et e − t
16
2.3 Displacements
Ch.2. Deformation and Strain
17
Displacements
Displacement: relative position of a particle, in its current (deformed)
configuration at time t, with respect to its position in the initial
(undeformed) configuration.
Displacement field: displacement of all the particles in the continuous
medium.
Material description (Lagrangian form):
( X, t ) x ( X, t ) − X
U= t0 t
U i ( X, t ) = xi ( X, t ) − X i i ∈ {1, 2,3} Ω
P U=u P’
Spatial description (Eulerian form): Ω0 X
x
u ( x, t )= x − X ( x, t )
ui ( x, t ) =
xi − X i ( x, t ) i ∈ {1, 2,3}
18
Displacement Gradient Tensor
U= ( X, t ) x ( X, t ) − X Taking partial derivatives of U w.r.t. X :
∂U i ( X, t ) ∂xi ( X, t ) ∂X i
U i ( X , t ) = xi ( X , t ) − X i i ∈ {1, 2,3}
def
= − = Fij − δ ij = J ij
∂X j ∂X j ∂X
j
∂U i δij
J = = Fij − δ ij i, j ∈ {1, 2,3} Fij
∂X j
ij
Material Displacement
def
Gradient Tensor
(
J X , t ) = U ( X, t ) ⊗ ∇ = F − 1
Taking partial derivatives of u w.r.t. x :
u ( x, t )= x − X ( x, t ) ∂ui (x, t ) ∂xi ∂X i (x, t ) def
= − = δ ij − Fij − δ ij = jij
−1
ui ( x, t ) =
xi − X i ( x, t ) i ∈ {1, 2,3} ∂x j ∂x j
∂x j
δij Fij−1
∂ui
ij ∂x =
j = δ ij − Fij−1 i, j ∈ {1, 2,3} Spatial Displacement Gradient
j
Tensor
def
j ( x, t ) = u ( x, t ) ⊗ ∇ =1 − F
−1
20
Strain Tensors
F characterizes changes of relative placements during motion but is not
a suitable measure of deformation for engineering purposes:
It is not null when no changes of distances and angles take place, e.g.,
in rigid-body motions.
21
Strain Tensors
Consider
F ( X,t )
t0 Q t d=
x F ⋅ dX
Ω dxi = Fij dX j
dS dX dx Q’
ds
d=
X F -1 ⋅ dx
Ω0 P P’
X x dX i = Fij−1dx j
22
Strain Tensors
d=
X F -1 ⋅ dx d=
x F ⋅ dX
dX i = Fij−1dx j dxi = Fij dX j
=
dS dX ⋅ dX =
ds dx ⋅ dx
( ) = ⋅ = [ ] ⋅ [ ] = [ ⋅ ] ⋅ [ F ⋅ dX ] = dX ⋅ FT ⋅ F ⋅ dX
2 T T
ds dx dx dx dx F dX
( ds ) dx
= = = =
2 T
k dx k F ki dX F
i kj dX j dX F F
i ki kj dX j dX i ik Fkj dX j
F
REMARK
not
( dS ) = dX ⋅ dX =[ dX ] ⋅ [ dX ] = F ⋅ dx ⋅ F ⋅ dx = dx ⋅ F ⋅ F ⋅ dx
2 T −1 T −1 −T −1 The convention
not
T
(•) −1 =
(•) −T
( dS )2 dX
= = k dX k Fki−1 dxi=
Fkj−1 dx j dxi Fki−=
1 −1
Fkj dx j dxi Fik−T Fkj−1dx j is used.
23
Green-Lagrange Strain Tensor
( ds ) = dX ⋅ FT ⋅ F ⋅ dX ( dS =
) dX ⋅ dX
2 2
Subtracting:
( ds ) − ( dS ) = dX ⋅ FT ⋅ F ⋅ dX − dX ⋅ dX = dX ⋅ FT ⋅ F ⋅ dX − dX ⋅ 1 ⋅ dX = dX ⋅ ( FT ⋅ F − 1 ) ⋅ dX = 2 dX ⋅ E ⋅ dX
2 2
def
= 2E
( )
E is symmetrical:
(2
1 T
) 2 F ⋅ ( F ) − 1= 2 ( F ⋅ F − 1=) E
1 T T T T 1 T
T
E=
T
F ⋅ F − 1=
=Eij E ji i, j ∈ {1,2,3}
24
Euler-Almansi Strain Tensor
( ds )= dx ⋅ dx ( dS ) =dx ⋅ F −T ⋅ F −1 ⋅ dx
2 2
Subtracting:
( ds ) − ( dS ) = dx ⋅ dx − dx ⋅ F −T ⋅ F −1 ⋅ dx = dx ⋅ 1 ⋅ dx − dx ⋅ F −T ⋅ F −1 ⋅ dx =
2 2
def
= 2e
= dx ⋅ (1 − F ⋅ F ) ⋅ dx = 2 dx ⋅ e ⋅ dx
−T −1
def
= 2e
The Euler-Almansi or Spatial Strain Tensor is defined:
e ( x, t ) =
1
2
( 1 − F −T
⋅ F −1
)
e ( x, t ) =
ij
1
2
( δ ij − Fki−1 Fkj −1 ) i, j ∈ {1, 2,3}
e is symmetrical: e =
T 1
2
( −T −1 T 1 T
2
−1 T
( −T T 1
2
)
1 − F ⋅ F ) = 1 − ( F ) ⋅ ( F ) = (1 − F −T ⋅ F −1 ) = e
( ds ) − ( dS )= 2 dX ⋅ E ⋅ dX
= 2 dx ⋅ e ⋅ d x
2 2
26
Strain Tensors in terms of Displacements
Substituting F −1= 1 - j and F= J + 1 into
E=
2
( F ⋅ F − 1) and e= 1 (1 − F −T ⋅ F −1 ) :
1 T
2
1 1
E= (1 + J T ) ⋅ (1 + J ) − 1= J + J T + J T ⋅ J
2 2
Eij= 1 ∂U i ∂U j ∂U k ∂U k
+ + i, j ∈ {1, 2,3}
2 ∂X j ∂X i ∂X i ∂X j
1
1 − (1 − jT ) ⋅ (1 − j) = j + jT − jT ⋅ j
1
e= 2
2
∂u
eij= 1 ∂ui + j − ∂uk ∂uk i, j ∈ {1, 2,3}
2 ∂x j ∂xi ∂xi ∂x j
27
Example
For the movement in the previous example, obtain the strain tensors in the
material and spatial description.
X + Y 2t
[ x] Y (1 + t )
=
Zet
29
Example - Solution
The deformation gradient tensor and its inverse tensor have already been
obtained: 2 yt
1 − 0
( + )
2
1 t
1 2Yt 0
0 1 + t 0 1
F=
F −1 =
0 0
1+ t
0 0 e t
0 0 e − t
30
Example - Solution
The spatial strain tensor :
2 yt
2 yt 1−1 − 0
1 − (1 + t )
2
0
0 (1 + t )
2
1 0
1 2
2
1 1 2 yt =
e=
1
(1 − F−T ⋅ F −1 )= 2 yt
1 − −
1
0 ⋅ 0
1
0 = −
(1 + t ) 2
1− −
2 yt
+
(1 + t ) 2 1 + t
0
(1+ t) 1+ t 1+ t
2
2 2 2
e − t 0 0 e−t −t −t
0 0
0 0 1− e e
2 yt
0 − 0
(1 + t )
2
2
2 yt 1 2
1 2 yt
= − 1− − 0
2 (1 + t ) 2 (1 + t ) 2 1 + t
0 0 1 − e −2 t
31
X + Y 2t
[ x] Y (1 + t )
Example - Solution =
Zet
2 yt
0 − 0 −
2Yt
(1 + t )
2
0 (1 + t )
0
2
1 2 yt 2 yt 1 2 2Yt 1
2 2
e ( x, t ) =
1 2Yt
− 1− − 0 e ( X, t ) = − 1 − − 0
2 (1 + t )
(1 + t ) 1 + t 2 (1 + t ) (1 + t ) 1 + t
2 2
y Y (1 + t )
=
0 0 1− e −2 t
0 0 1 − e −2t
spatial description material description
Observe that E ( x, t ) ≠ e ( x, t ) and E ( X, t ) ≠ e ( X, t ).
32
2.5 Variations of Distances
Ch.2. Deformation and Strain
33
Stretch
The stretch ratio or stretch is defined as:
def P´Q´ ds
stretch = λ T= λ=
t = (0 < λ < ∞)
PQ dS
t0 Q t t
dS dX Ω Q’ REMARK
dx
ds The sub-indexes (●)T and
(●)t are often dropped. But
Ω0 P P’
X x one must bear in mind that
stretch and unit elongation
always have a particular
direction associated to them.
34
Unit Elongation
The extension or unit elongation is defined as:
def ∆ PQ ds − dS
unit elongation = ε=
T ε=
t =
PQ dS
t0 Q t t
dS dX Ω Q’ REMARK
dx
ds The sub-indexes (●)T and
(●)t are often dropped. But
Ω0 P P’
X x one must bear in mind that
stretch and unit elongation
always have a particular
direction associated to them.
35
Relation between Stretch and Unit
Elongation
The stretch and unit elongation for a same point and direction are
related through:
ds − dS ds
ε= = − 1= λ − 1 ( −1 < ε < ∞ )
dS dS
=λ
= (ε 0 ) =
If λ 1= ds dS : P and Q may have moved in time but have kept
the distance between them constant.
36
Stretch and Unit Elongation in terms of
the Strain Tensors
Considering:
( ds ) − ( dS )= 2 dX ⋅ E ⋅ dX
2 2
dX = T dS
( ds ) − ( dS )= 2 dx ⋅ e ⋅ dx dx = t ds
2 2
Then:
× 1
( dS ) 2 = λ2
2
ds
( ds ) − ( dS ) = 2 ( dS ) T ⋅ E ⋅ T − 1= 2 T ⋅ E ⋅ T
2 2 2
dS REMARK
2
dS E(X,t) and e(x,t)
( ds ) − ( dS ) = 2 ( ds ) t ⋅ e ⋅ t 1− = 2 t ⋅e ⋅ t
2 2 2
ds contain information
× 1
( λ)
2
( ds )2 = 1 regarding the stretch
and unit elongation for
1
λ= any direction in the
λ = 1+ 2 T ⋅ E ⋅ T 1− 2 t ⋅e ⋅ t differential neighbour-
ε = λ −1 = 1 + 2 T ⋅ E ⋅ T −1 1 hood of a point.
ε = λ −1 = −1
1− 2 t ⋅e ⋅ t
37
2.6 Variation of Angles
Ch.2. Deformation and Strain
38
Variation of Angles
dx( ) = t ( ) ds ( )
1 1 1
(1) (1) (1)
dX = T dS
dx( ) = t ( ) ds (
2 2 2)
( 2) ( 2) ( 2) T(1)
dX =T dS t
t0
Q t(1)
t(2)
T(2)
Θ Ω R’ θ
dS(1) Q’
R ds(2) ds(1)
dS(2)
Ω0 P P’
X x
39
Variation of Angles
d x( 2) =
x(1) ⋅ d ds (1) ds ( 2) cos θ
dx(1)
T
dx ( 2 )
(1)
dx = F ⋅ dX (1) T(1) ⋅ (1 + 2E ) ⋅ T( 2)
cos θ =
( 2) ( 2) 1 + 2 T(1) ⋅ E ⋅ T(1) 1 + 2 T( 2 ) ⋅ E ⋅ T( 2 )
dx = F ⋅ dX
= F ⋅ dX (1) ⋅ F ⋅ dX (2) = dX (1) ⋅ ( FT ⋅ F ) ⋅ dX (2)
T
(1) (2)
dx ⋅ dx
2E+1
dX(1) = T(1) dS (1)
dX( ) = T( ) dS ( )
2 2 2
1 1
dx (1)=
⋅ dx (2) T ⋅ ( 2E + 1 ) ⋅ T = ds (1)ds (2) (1) (2) T(1) ⋅ ( 2E + 1) ⋅=
(1) (1) (2) ( 2)
dS
dS T(2) ds (1)ds (2) cos θ
ds(1) ds(2) λ λ
(1) (2)
λ λ
λ = 1+ 2 T ⋅ E ⋅ T
T(1) ⋅ (1 + 2E ) ⋅ T( 2)
(1) (1)
1+ 2 T ⋅ E ⋅ T 1 + 2 T( 2 ) ⋅ E ⋅ T( 2 )
40
Variation of Angles
41
Variation of Angles
X(1) =
d X( 2) dS (1) dS ( 2) cos Θ
⋅ d
dX(1)
T
dX( 2)
t (1) ⋅ (1 − 2e ) ⋅ t ( 2)
dX=(1)
F −1 ⋅ dx(1) cos Θ =
( 2) −1
dX= F ⋅ dx
( 2) 1 − 2 t (1) ⋅ e ⋅ t (1) 1 − 2 t ( 2) ⋅ e ⋅ t ( 2)
dX (1) ⋅ dX (2) = F −1 ⋅ dx (1) ⋅ F −1 ⋅ dx (2) = dx (1) ⋅ ( F −T ⋅ F −1 ) ⋅ dx (2)
T
1−2e
dx(1) = t (1) ds (1)
( 2) ( 2) ( 2)
dx = t ds
REMARK λ=
1
E(X,t) and e(x,t) contain information 1− 2 t ⋅e ⋅ t
regarding the variation in angles t (1) ⋅ (1 − 2e ) ⋅ t ( 2)
between segments in the differential
(1) (1)
neighbourhood of a point. 1− 2 t ⋅e ⋅ t 1 − 2 t ( 2) ⋅ e ⋅ t ( 2)
42
Example
Let us consider the motion of a continuum body such that the spatial description
of the Cartesian components of the spatial Almansi strain tensor is given by,
0 0 −tetz
[ e ( x, t ) ] = 0 0 0
tz
−te 0 t ( 2etz − et )
Compute at time t=0 (the reference time), the length of the curve that at time
t=2 is a straight line going from point a (0,0,0) to point b (1,1,1).
The length of the curve at time t=0 can be expressed as,
1
∫ λ ( x, t ) ds
B b
=L ∫=
A
dS
a
= ds
λ
43
Example - Solution
The inverse of the stretch, at the points belonging to the straight line going
from a(0,0,0) to b(1,1,1) along the unit vector in the direction of the straight
line, is given by,
1
λ ( x, t ) = → λ −1 ( x, t ) = 1 − 2t ⋅ e ( x, t ) ⋅ t
1 − 2t ⋅ e ( x, t ) ⋅ t
Where the unit vector is given by,
1
[t ] = [1 1 1]
T
3
Substituting the unit vector and spatial Almansi strain tensor into the expression
of the inverse of the stretching yields,
2 t
λ −1
( x, t=) 1 + te
3
44
Example - Solution
The inverse of the stretch, which is uniform and therefore does not depends on
the spatial coordinates, at time t=2 reads,
4
λ −1 ( x, 2=
) 1 + e2
3
Substituting the inverse of the stretch into the integral expression provides the
length at time t=0,
4 2 4 2 (1,1,1)
( )
b b
∫Γ ∫a λ ∫a 3 ∫
−1
L= dS = x , 2 ds = 1 + e ds =1 + e ds =+
3 4e 2
3
(0,0,0)
= 3
45
2.7 Physical Interpretation of E and e
Ch.2. Deformation and Strain
46
Physical Interpretation of E
Consider the components of the material strain tensor, E:
E XX E XY E XZ E11 E12 E13
=
E = E XY EYY EYZ E E22 E23
12
E XZ EYZ EZZ E13 E23 E33
Stretching of
1 dS λ= 1 + 2E11 the material in
1
T (1)
≡ 0 dX ≡ dS T (1)
=
0
0 0
the X-direction
47
Physical Interpretation of E
Similarly, the stretching of the material in the Y-direction and the Z-
direction:
λ1 = 1 + 2 E11 ε X = λ X − 1= 1 + 2 E XX − 1
λ2 = 1 + 2 E22 ε Y = λY − 1= 1 + 2 EYY − 1
λ3 = 1 + 2 E33 ε Z = λZ − 1= 1 + 2 EZZ − 1
48
Physical Interpretation of E
Consider the angle between a segment parallel to the X-axis and a
segment parallel to the Y-axis, the angle is:
T ( ) ⋅ ( 1 + 2E ) ⋅ T ( )
1 2
cos θ =
1 + 2 T(1) ⋅ E ⋅ T(1) 1 + 2 T( 2 ) ⋅ E ⋅ T( 2 )
reference 1 0 T(1) ⋅ T( 2) =
0
T ⋅ E ⋅ T(1) =
(1)
configuration T(1) = 0 T( 2) = 1 E11
0
0
T(1) ⋅ E ⋅ T( 2) =E12
( 2) ( 2)
T ⋅E ⋅T = E22
2 E12
cos θ =
1 + 2 E11 1 + 2 E 22
deformed
configuration
2 E XY π 2 E XY
θ ≡ θ xy =arccos = − arcsin
1 + 2 EXX 1 + 2 EYY 2 1 + 2 EXX 1 + 2 EYY
49
Physical Interpretation of E
π 2 E XY
θ ≡ θ xy = − arcsin
2 1 + 2 EXX 1 + 2 EYY
2 E XY
θ xy − Θ
∆Θ XY =
=− arcsin
1 + 2 EXX 1 + 2 EYY
XY
π
2
reference
configuration
deformed
configuration
50
Physical Interpretation of E
Similarly, the increment of the final angle w.r.t. its initial value for
couples of segments oriented in the direction of the coordinate axes:
2 E XY
∆Θ XY = − arcsin
1 + 2 EXX 1 + 2 EYY
2 E XZ
∆Θ XZ = − arcsin
1 + 2 EXX 1 + 2 EZZ
2 EYZ
∆ΘYZ = − arcsin
1 + 2 EYY 1 + 2 EZZ
reference λ1 dX= 1 + 2 E XX dX
configuration
λ2 dY= 1 + 2 EYY dY
λ3 dZ= 1 + 2 EZZ dZ
2 E XY
∆Θ XY = − arcsin
1 + 2 EXX 1 + 2 EYY
2 E XZ
∆Θ XZ = − arcsin
1 + 2 EXX 1 + 2 EZZ
2 EYZ
∆ΘYZ = − arcsin
1 + 2 EYY 1 + 2 EZZ
52
Physical Interpretation of e
Consider the components of the spatial strain tensor, e:
exx exy exz e11 e12 e13
e ≡ exy eyy eyz = e12 e22 e23
exz eyz ezz e13 e23 e33
1 ds 1 Stretching of
λ1 = the material in
t (1) ≡ 0 dx ≡ 0 1 − 2e11
0 0 the x-direction
53
Physical Interpretation of e
Similarly, the stretching of the material in the y-direction and the z-
direction: 1 1
λ1 = ⇒ ε x = λx − 1 = −1
1 − 2e11 1 − 2exx
1 1
λ2 = ⇒ ε y = λy − 1= −1
1 − 2e22 1 − 2eyy
1 1
λ3 = ⇒ ε z = λz − 1 = −1
1 − 2e33 1 − 2ezz
54
Physical Interpretation of e
Consider the angle between a segment parallel to the x-axis and a
segment parallel to the y-axis, the angle is:
t (1) ⋅ (1 − 2e ) ⋅ t ( 2)
cos Θ =
reference deformed 1 − 2 t (1) ⋅ e ⋅ t (1) 1 − 2 t ( 2) ⋅ e ⋅ t ( 2)
configuration configuration
1 0 t (1) ⋅ t ( 2) = 0
(1) (1)
t ⋅ e ⋅t = e11
t (1) = 0 t (2) = 1
0 0 t (1) ⋅ e ⋅ t ( 2) =e12
( 2)
t ⋅e ⋅t = ( 2)
e22
−2 e12
cos Θ =
1 − 2 e11 1 − 2 e 22
π 2exy
Θ ≡ Θ XY= + arcsin
2 1 − 2 exx 1 − 2 eyy
55
Physical Interpretation of e
π 2exy
Θ ≡ Θ XY= + arcsin
2 1 − 2 exx 1 − 2 eyy
The increment of the angle in the reference configuration w.r.t. its value
in the deformed one:
2exy
∆θ xy = θ xy − Θ XY = − arcsin
1 − 2 exx 1 − 2 eyy
π
2
reference deformed
configuration configuration
56
Physical Interpretation of e
Similarly, the increment of the angle in the reference configuration w.r.t.
its value in the deformed one for couples of segments oriented in the
direction of the coordinate axes:
π 2exy
∆θ xy = − Θ XY = − arcsin
2 1 − 2 exx 1 − 2 eyy
π 2exz
∆θ xz = − Θ XZ = − arcsin
2 1 − 2 exx 1 − 2 ezz
π 2eyz
∆θ yz = − ΘYZ = − arcsin
2 1 − 2 eyy 1 − 2 ezz
57
Physical Interpretation of e
In short, reference
configuration
deformed
configuration
dx
= 1 − 2exx dx
λ1
dy
= 1 − 2eyy dy
λ2
dz
= 1 − 2ezz dz
λ3
π 2exy
∆θ xy = − Θ XY = − arcsin
2 1 − 2 exx 1 − 2 eyy
π 2exz
∆θ xz = − Θ XZ = − arcsin
2 1 − 2 exx 1 − 2 ezz
π 2eyz
∆θ yz = − ΘYZ = − arcsin
2 1 − 2 eyy 1 − 2 ezz
58
2.8 Polar Decomposition
Ch.2. Deformation and Strain
59
Polar Decomposition
Polar Decomposition Theorem:
“For any non-singular 2nd order tensor F there exist two unique
positive-definite symmetrical 2nd order tensors U and V, and a unique
orthogonal 2nd order tensor Q such that: ”
not
left polar
=
U F ⋅F
T
decomposition
not
V = F ⋅ FT F = Q ⋅ U = V ⋅Q
Q= F ⋅ U −1 =V −1 ⋅ F right polar
decomposition
The decomposition is unique. REMARK
Q: Rotation tensor An orthogonal 2nd
order tensor verifies:
U: Right or material stretch tensor
QT ⋅ Q = Q ⋅ QT = 1
V: Left or spatial stretch tensor
60
Properties of an orthogonal tensor
An orthogonal tensor Q when multiplied (dot product) times
a vector rotates it (without changing its length): y= Q ⋅ x
y has the same norm as x:
y = y ⋅ y = [ y ] [ y ] = [Q ⋅ x ] ⋅ [Q ⋅ x ] = x ⋅ Q
⋅Q⋅x = x
2 T T T 2
when Q is applied on two vectors x(1) and x(2), with the same origin,
the original angle they form is maintained:
T T
(1) (2)
y y
y = Q⋅x
(1) (1)
y ⋅y
(1)
x ⋅ QT ⋅ Q ⋅ x (2)
(2)
(1)
x (1) ⋅ x (2)
= = = cos α
y (2)= Q ⋅ x (2) y (1) y (2) y (1) y (2) x (1) x (2)
61
Polar Decomposition of F
Consider the deformation gradient tensor, F:
F = Q ⋅ U = V ⋅Q
stretching
rotation
dx =F ⋅ dX =( V ⋅ Q ) ⋅ dX =V ⋅ ( Q ⋅ dX )
(not)
F (•) ≡ stretching rotation (•)
rotation
stretching
dx =F ⋅ dX =( Q ⋅ U ) ⋅ dX =Q ⋅ ( U ⋅ dX )
REMARK
For a rigid body motion:
not
F(•) ≡ rotation stretching (•) U= V= 1 and Q = F
62
2.9 Volume Variation
Ch.2. Deformation and Strain
63
Differential Volume Ratio
Consider the variation of a differential volume associated to a particle
P:
=
dV0 ( dX
(1) ( 2)
× dX (3)
)=
⋅ dX
dX 1(1) dX 2(1) dX 3(1)
deformed ( 2) ( 2) ( 2)
configuration = det
= dX 1 dX 2 dX 3 M
( 3) ( 3) ( 3)
dX 1 dX 2
dX 3
M
reference
configuration
=
dVt ( dx( ) × dx( ) )=
1 2
⋅ dx( )
3
= =
M ij dX (i )
j mij dx (ji )
64
Differential Volume Ratio
Consider now: dx (i ) =
F ⋅ dX ( i ) i ∈ {1, 2,3} → Fundamental eq. of deformation
(i )
dx j = Fjk ⋅ dX k(i ) i, j ∈ {1, 2,3}
= =
M ij dX (i )
j and mij dx (ji )
m=
ij dx (j=
i)
Fjk dX k(=
i)
Fjk M ik= M ik FkjT = M ⋅ FT
m
Then:
dVt = m = M ⋅ FT = M FT = F M = F dV0 dVt = F dV0
dV
0
65
2.10 Area Variation
Ch.2. Deformation and Strain
66
Surface Area Ratio
Consider the variation of a differential area associated to a particle P:
dA := dAN → material vector "differential of area" → dA = dA
da := dan → spatial vector "differential of area" → da = dA
Deformed (current)
configuration ( 3)
dV0 = dH dA = d
X
⋅
N dA =
Reference (initial) dH
configuration =dX(3) ⋅ N
dA =d A ⋅ d X ( 3)
dA
dVt = dh da = x( ) ⋅ n da =
3
d
dh
=dx ( 3 ) ⋅ n
da =d a ⋅ d x ( 3)
da
67
Surface Area Ratio
Consider now:
da ⋅ dx (
3)
dV=
t
dx (3)= F ⋅ dX( )
3
dVt = F dV0
dV= dA ⋅ dX( )
3
0
( 3) ( 3) ( 3)
dVt = F d
A ⋅
dX =
d a ⋅ F ⋅ dX ∀dX ⇒ F dA = da ⋅ F
F dV dV0 dVt
0
da = F ⋅ dA ⋅ F −1 =
da n F N ⋅ F −1dA da F N ⋅ F −1 dA
=
dA = N dA
da = n da
68
2.11 Volumetric Strain
Ch.2. Deformation and Strain
69
Volumetric Strain
Volumetric Strain:
def dV ( X, t ) − dV ( X, t0 ) not dVt − dV0
e ( X, t ) =
dV ( X, t ) dV0
dVt = F dV0
F dV0 − dV0
e=
dV0
=
e F −1 REMARK
The incompressibility condition (null
volumetric strain) takes the form
e = J −1 = 0 ⇒ J = F = 1
70
2.12 Infinitesimal Strain
Ch.2. Deformation and Strain
71
Infinitesimal Strain Theory
The infinitesimal strain theory (also called t0 t
small strain theory) is based on the Ω
simplifying hypotheses: P u P’
Ω0 X
x
Displacements are very small w.r.t. the
typical dimensions in the continuum
medium,
As a consequence, u << ( size of Ω0 ) and the reference and deformed configurations are
considered to be practically the same, as are the material
not
and spatial coordinates:
Ω ≅ Ω0 and x = X+u ≅ X ( X, t ) u ( X, t ) ≡ u ( x, t )
U=
xi = X i + ui ≅ X i not
U i ( X, t ) =ui ( X, t ) ≡ ui ( x, t ) i ∈ {1, 2,3}
72
Infinitesimal Strain Theory
The material and spatial coordinates coincide, x = X +
u ≅X
≈0
Even though it is considered that u cannot be neglected when calculating
other properties such as the infinitesimal strain tensor ε.
There is no difference between the material and spatial
differential operators:
symb
∂ ∂
∇ = ˆ
e = eˆ i =
∇
∂ ∂
i
X i x i
J (=
X, t ) U( X, t )= ⊗ ∇ u(x, t )=
⊗ ∇ j(x, t )
=ε J + J=T
j + j=
T
(u ⊗ ∇ + ∇ ⊗ =
u ) ∇ su
2 2 2 ε is a symmetrical tensor and its components
ε ij =1 ∂ui ∂u j are infinitesimal: | ε ij |<< 1, ∀i, j ∈ {1, 2,3}
+ i, j ∈ {1, 2,3}
2 ∂x j ∂xi
74
Stretch and Unit Elongation
Stretch in terms of the strain tensors:
1
λT = 1 + 2
T⋅⋅T
E λt =
x 1− 2 t ⋅e ⋅ t
For a segment parallel to the x-axis, the stretch and unit elongation are:
λ ≅ 1+ t ⋅ε ⋅t
ε11 ε12 ε13 1
reference t ⋅ ε ⋅ t [1 0 0] ⋅ ε12 ε 22 ε 23 =
= ⋅ 0 ε11
configuration ε13 ε 23 ε 33 0
The diagonal components of the infinitesimal strain tensor are the unit
elongations of the material when in the x, y and z-directions.
ε xx ε xy ε xz
ε = ε xy ε yy ε yz
ε xz ε yz ε zz
77
Physical Interpretation of
Infinitesimal Strains
Consider the angle between a segment parallel to the X-axis and a
segment parallel to the Y-axis, the angle is Θ = π2 . XY
Applying: π 2E
θ ≡ θ xy = − arcsin XY
2 1 + 2 EXX 1 + 2 EYY
E XX = ε xx
E XY = ε xy
EYY = ε yy
π 2ε xy π π
reference
θ xy = − arcsin ≅ − arcsin 2ε xy = − 2ε xy
2 1 + 2ε xx 1 + 2ε yy 2 2
configuration ≈2ε xy
≈1 ≈1
REMARK
The Taylor linear series expansion of arcsin x yields
( x ) + ... = x + O ( x 2 )
d arcsin
arcsin ( x ) ≅ arcsin ( 0 ) +
dx x =0
78
Physical Interpretation of
Infinitesimal Strains
π
θ xy ≅ − 2ε xy
2
79
Physical Interpretation of
Infinitesimal Strains
In short,
reference deformed
configuration configuration
ε xx = ε x 1
ε xy =− ∆θ xy
2
ε yy = ε y 1
ε xz =− ∆θ xz
2
ε zz = ε z 1
ε yz =− ∆θ yz
2
80
Engineering Strains
Using an engineering notation, instead of the scientific notation, the
components of the infinitesimal strain tensor are
REMARK
Positive longitudinal strains indicate
increase in segment length.
cos(Θ + ∆θ ) =
T(1) ⋅ T(2) + 2T(1) ⋅ ε ⋅ T(2)
82
Variation of Angles
cos(Θ + ∆θ=) T(1) ⋅ T(2) + 2T(1) ⋅ ε ⋅ T(2)
T(1) and T(2) are unit vectors in the directions of the original segments,
therefore, T(1) ⋅ T(2)= T(1) T(2) cos Θ
= cos Θ
T(1) ≈ t (1)
Also, cos(Θ + ∆θ=) cos Θ ⋅ cos ∆θ − sinΘ ⋅ sin∆=
θ cos Θ − sinΘ ⋅ ∆θ T (2)
≈ t (2)
Θ ≈θ
≈1 ≈ ∆θ
2T(1) ⋅ ε ⋅ T(2) 2t (1) ⋅ ε ⋅ t (2)
θ cos Θ + 2T ⋅ ε ⋅ T
cos Θ − sinΘ ⋅ ∆= (1) (2)
∆θ =− =−
sinΘ sinθ
REMARK
The Taylor linear series expansion of sin x
and cos x yield
( x ) + ... = x + O ( x 2 )
d sin
sin ( x ) ≅ sin ( 0 ) +
dx x =0
( x ) + ... =1 + O ( x 2 )
d cos
cos ( x ) ≅ cos ( 0 ) +
dx x =0
83
Polar Decomposition
Polar decomposition in finite-strain problems:
left polar
not
decomposition
=
U FT ⋅ F
not
V = F ⋅ FT ⇒ F = Q ⋅ U = V ⋅Q
Q= F ⋅ U −1 =V −1 ⋅ F right polar
decomposition
REMARK
In Infinitesimal Strain Theory
x ≈ X , therefore,= ∂x
F ≈1
∂X
84
Polar Decomposition
In Infinitesimal Strain Theory:
U = FT F = (1 + J ) ⋅ ( 1 + J ) =
T
1 + J + JT + JT ⋅ J ≈ 1 + J + JT = 1 +
1
2
( J + JT )
<< J =x
=ε
U= 1 + ε
infinitesimal strain tensor
Similarly, REMARK
The Taylor linear series expansion of 1 + x
U = (1 + ε ) = 1 − ε = 1 − ( J + J T )
−1 1
and (1 + x ) yield
−1 −1
=x 2
dλ
x =1 + x + O ( x 2 )
1
λ ( x ) = 1 + x ≅ λ ( 0) +
=ε dx x =0 2
U −1 = 1 − ε λ ( x ) = (1 + x ) ≅ λ ( 0 ) +
−1 dλ
x =1 − x + O ( x 2 )
dx
infinitesimal x =0
strain tensor
85
Polar Decomposition
(1 + J ) ⋅ 1 − ( J + JT ) =+
1 J − ( J + J T ) − J ⋅ ( J + J T ) =+ ( J − JT )
1 1 1 1
F U −1 =
Q =⋅ 1
2 2 2 2
<< J = Ω
Q= 1 + Ω
The infinitesimal rotation tensor Ω is defined: REMARK
The antisymmetric or
def 1 1 def
Ω = (J − J T )= (u ⊗ ∇ − ∇ ⊗ u)= ∇ a u skew-symmetrical
2 2 gradient operator is
Ω 1 ∂ui ∂u j defined as:
= − << 1 i, j ∈ {1, 2,3}
ij 2 ∂x j ∂xi 1
0 Ω12 −Ω31
∇ a (•=
) [(•) ⊗ ∇ − ∇ ⊗ (•)]
2
The diagonal terms of Ω are zero: [ Ω ] = −Ω12 0 Ω 23
Ω31 −Ω 23 0
It can be expressed as
an infinitesimal rotation vector θ, ∂u3 ∂u2
∂x − ∂x
θ1 −Ω 23 2 3
REMARK
1 ∂u1 ∂u3 def 1
θ ≡ θ 2 = −Ω31 = − = ∇×u Ω is a skew-symmetric
θ −Ω 2 ∂x3 ∂x1 2 tensor and its components
3 12 ∂u2 ∂u1
− are infinitesimal.
∂x1 ∂x2
86
Polar Decomposition
From any skew-symmetric tensor Ω, it can be extracted a vector θ (axial
vector of Ω) exhibiting the following property:
Ω ⋅r =θ×r ∀r
As a consequence:
The resulting vector is orthogonal to r.
87
Proof of θ×r =
Ω⋅r ∀r
The result of the dot product of the infinitesimal rotation tensor, Ω, and a
generic vector, r, is exactly the same as the result of the cross product of
the infinitesimal rotation vector, θ, and this same vector.
0 Ω12 −Ω31 θ1 −Ω 23 r1
[Ω ] = −Ω12 0 Ω23 → θ ≡ θ 2 = −Ω31 ⇒ Ω ⋅ r =θ × r ∀r = r2
θ −Ω r
Ω31 −Ω 23 0 3 12 3
Proof:
eˆ1 eˆ 2 eˆ 3 eˆ1 eˆ 2 eˆ 3 Ω12 r2 − Ω31r3
θ × r = det θ1 θ 2 θ3 = det −Ω 23 −Ω12 =
not
−Ω31 −Ω12 r1 + Ω 23 r3
r1 r2 r3 Ω r −Ω r
r1 r2 r3 31 1 23 2
88
Polar Decomposition
Using:
J= F − 1
=ε
1
2
( J + JT ) 1 J =+
F =+ 1
1
2
( J + JT ) + ( J − JT )
1
2 F = 1+ε+Ω
Q= 1 + Ω =ε = Ω
REMARK
The infinitesimal rotation tensor
characterizes the rotation and, in the
small-strain context, maintains angles
89 and distances.
Volumetric Deformation
The volumetric strain:
=
e F −1
Considering: F= Q ⋅ U and U= 1 + ε
1 + ε xx ε xy ε xz
F = Q ⋅ U = Q U = U = 1 + ε = det ε xy 1 + ε yy ε yz =
ε xz ε yz 1 + ε zz
= 1 + ε xx + ε yy + ε zz + O ( ε 2 ) ≈ 1 + Tr ( ε )
= Tr ( ε )
e = Tr ( ε )
90
2.13 Strain Rate
Ch.2. Deformation and Strain
REMARK
We are no longer assuming an
infinitesimal strain framework
91
Spatial Velocity Gradient Tensor
Consider the relative velocity between two points in space at a given
(current) instant:
∂v
=
v P′ v= (x, t ) v ( x1 , x2 , x3 , t ) dv = ⋅ dx =⋅
l dx
∂
x
dv (x, t ) = v Q′ − v P′ = v ( x + dx, t ) − v ( x, t ) l
∂vi
=
dvi = dx j lij dx j
∂x j
lij i, j ∈ {1, 2,3}
( )= v ⊗ ∇
def ∂v x, t
l ( x , t )=
Spatial velocity ∂x
gradient tensor = lij ∂vi i, j ∈ {1, 2,3}
∂x j
92
Strain Rate and Rotation Rate (or Spin)
Tensors
The spatial velocity gradient tensor can be split into a symmetrical and
a skew-symmetrical tensor:
l= v ⊗ ∇
sym [ l ] + skew [ l ] =
l= :d + w
l ∂vi i, j ∈ 1, 2,3
=
ij ∂x { }
j
93
Physical Interpretation of d
The strain rate measures the rate of deformation of the square of the
differential length ds in the spatial configuration,
d d d d dx dx
( ds(t =
)) ( dx ⋅ d=x) ( dx ) ⋅ dx + dx ⋅ ( d=
x ) d ⋅ dx + dx ⋅ d = dv ⋅ dx + dx ⋅ dv
2
dt dt dt dt
dt
dt
dv = l ⋅ dx =v =v
1
= d (l + lT )
2
d
dt
( ds(t ) )
2
= (
d x ⋅ l ) ⋅ d x + d x ⋅ ( l ⋅ dx ) =
T
dx ⋅ lT + l ⋅ dx = 2dx ⋅ d ⋅ dx
dv
T dv = 2d
d
dt
( (ds (t )) 2 − (
dS ) 2 ) =
d
dt
( 2dX ⋅ E ( X, t ) ⋅ dX ) = 2dX ⋅
dE
dt
⋅ dX =
d
dt
( (ds (t )) 2 )
constant
2dX⋅E X,t ⋅dX notation
= E
94
Physical Interpretation of d
⋅ dX = dx ⋅ d ⋅ dx
dX ⋅ E
dx= F ⋅ dX
⋅ dX = dx ⋅ d ⋅ dx =
dX ⋅ E [
dx ] [ d ][ dx ] = [ F ⋅ dX ] [d ][ F ⋅ dX ] =
T
T
dX ⋅ ( FT ⋅ d ⋅ F) ⋅ dX
F⋅dX F⋅dX T T
dX F F dX
There is a direct relation between the material derivative of the material strain
tensor and the strain rate tensor but they are not the same.
and d will coincide when in the
E
reference configuration F |t =t0 = 1 .
REMARK
Given a 2 order tensor A,
nd
95
Physical Interpretation of w
To determine the (skew-symmetric) rotation rate (spin) tensor only three
different components are needed:
0 w12 w13
1 ∂vi ∂v j
w ij = − i, j ∈ {1, 2,3} [ w ] = − w12 0 w23
2 ∂x j ∂xi − w13 − w23 0
96
Physical Interpretation of w
It can be proven that the equality ω× r =w ⋅ r ∀r holds true.
Therefore:
ω is the angular velocity of a rotation movement.
ω x r = w · r is the rotation velocity of the point that
has r as its position vector w.r.t. the rotation centre.
dv =d ⋅ dx + w ⋅ dx
97
2.14 Material time Derivatives
Ch.2. Deformation and Strain
99
Deformation Gradient Tensor F
The material time derivative of the deformation gradient tensor,
∂xi ( X, t ) REMARK
=
Fij ( X, t ) i, j ∈ {1, 2,3} The equality of cross derivatives
∂X j
applies here: ∂ 2 (•) = ∂ 2 (•)
d ∂µi µ j ∂µ j µi
dt
dFij ∂ ∂xi ( X, t ) ∂ ∂xi ( X, t ) ∂Vi ( X, t ) ∂vi ( x ( X, t ) ) ∂x k
= = = = = lik Fkj
dt ∂t ∂X j ∂X j ∂t ∂X j ∂xk ∂X j
=Vi ( X, t ) =lik =Fkj
dF notation
dt = F= l ⋅ F
dF=ij
F= lik Fkj i, j ∈ {1, 2,3}
dt ij
100
Inverse Deformation Gradient Tensor F-1
The material time derivative of the inverse deformation gradient
tensor,
1
F ⋅ F −1 = REMARK
Do not mistake the material derivative
d of the inverse tensor for the inverse of
dt the material derivative of the tensor:
d ( F −1 )
( F ( X,t ) ) ≠ ( F
( X,t ) )
d −1 dF −1 d −1
(F ⋅ F ) = ⋅F + F⋅ = 0 −1
dt dt dt dt
d ( F −1 ) dF −1
⇒ F⋅ =− ⋅ F =−F ⋅ F −1
dt dt
Rearranging terms,
d ( F −1 )
−F −1 ⋅ F ⋅ F −1 =
= −F −1 ⋅ l ⋅ F ⋅ F −1 =
−F −1 ⋅ l
dt d ( F −1 )
= l⋅F =1 = −F −1 ⋅ l
dt
−1
dFij
dt = − F −1
ik lkj i, j ∈ {1, 2,3}
101
Strain Tensor E
The material time derivative of the material strain tensor has already
been derived for the physical interpretation of the deformation rate
tensor:
E= F ⋅d ⋅F
T
=
E
2
( F ⋅ F − 1)
1 T
F = l ⋅ F
d F=T
F T ⋅ lT
dt
dE 1 T
dt
= E=
2
( F ⋅ F + FT ⋅ F ) =
2
( F ⋅ l ⋅ F + FT ⋅ l ⋅ F ) = FT ⋅ ( l + lT ) ⋅ F = FT ⋅ d ⋅ F
1 T T 1
2
d
= FT ⋅ d ⋅ F
E
102
Strain Tensor e
The material time derivative of the spatial strain tensor,
e=
1
2
( 1 − F −T ⋅ F −1 )
F=−1
F −1 ⋅ l
d F −T= lT ⋅ F −T
dt
− ( F −T ⋅ F −1 + F −T ⋅ F −1 ) =( lT ⋅ F −T ⋅ F −1 + F −T ⋅ F −1 ⋅ l )
de 1 1
e =
=
dt 2 2
e = ( l ⋅ F ⋅ F + F −T ⋅ F −1 ⋅ l )
1 T −T −1
2
103
Volume differential dV
The material time derivative of the volume differential associated to a
given particle,
d F d F dFij −1 dFij −1
= = F F= F F F lik
d dt dFij dt
ji
dt
kj ji
( dV=) ( ∇ ⋅ v ) F dV0 lik Fkj ( F⋅F−1 ) δ
= ki
dt
= dV ki
∂v i dF
= F=
lii F = F ∇⋅v = F ∇ ⋅ v = (∇ ⋅ v ) F
d ∂xi
dt
( dV (x, t ) )= ∇ ⋅ v(x, t ) dV (x, t ) ∇⋅v
dt
104
Area differential vector da
The material time derivative of the area differential associated to a
given particle,
da ( x( X, t ), t ) = F ( X, t ) ⋅ dA( X) ⋅ F −1 ( X, t ) = F ⋅ dA ⋅ F −1
d
dt
dA ⋅ F −1 + F ⋅ dA ( F −1 )
d dF d
(t )
da=
dt dt dt
= F ∇⋅v = −F −1 ⋅ l
d
( da ) = ( ∇ ⋅ v ) F dA ⋅ F −1 − F dA ⋅ F −1 ⋅ l
dt
= da = da
d
( da ) = da ( ∇ ⋅ v ) − da ⋅ l = da ⋅ 1(∇ ⋅ v) − da ⋅ l = da ⋅ ( (∇ ⋅ v)1 − l )
dt da⋅1
105
2.15 Other Coordinate Systems
Ch.2. Deformation and Strain
106
Curvilinear Orthogonal Coord. System
A curvilinear coordinate system is defined by:
The coordinates, generically named {a, b, c}
Its vector basis, {eˆ a , eˆ b , eˆ c } , formed by unit vectors eˆ=
a eˆ=
b e=
ˆ c 1.
REMARK
A curvilinear orthogonal coordinate system can be seen as a mobile Cartesian
coordinate system { x′, y′, z ′} , associated to a curvilinear basis {eˆ a , eˆ b , eˆ c } .
108
Curvilinear Orthogonal Coord. System
A curvilinear orthogonal coordinate system can be seen as a mobile
Cartesian coordinate system {eˆ a , eˆ b , eˆ c } , associated to a curvilinear basis
{ x′, y′, z′} .
The components of a vector and a tensor magnitude in the curvilinear
orthogonal basis will correspond to those in the given Cartesian local
system:
va v x′ Taa Tab Tac Tx′x′ Tx′y′ Tx′z′
v ≡ v b ≡ v y′ T ≡ Tba Tbb Tbc ≡ Ty′x′ Ty′y′ Ty′z′
Tca Tcc Tz′x′ Tz′z′
v c v z′ Tcb Tz′y′
109
Cylindrical Coordinate System
← z− coordinate line
← θ − coordinate line x = r cos θ
x( r , θ , z ) ≡ y = r sin θ
← r− coordinate line z = z
∂eˆ r ∂eˆθ
= eˆθ = −eˆ r
∂θ ∂θ
dV = r dθ dr dz
110
Cylindrical Coordinate System
Nabla operator ∂
∂r
∂ 1 ∂ ∂ 1 ∂
∇= eˆ r + eˆθ + eˆ z ⇒ ∇≡
∂r r ∂θ ∂z r ∂θ x = r cos θ
∂ x( r , θ , z ) ≡ y = r sin θ
∂z z = z
Displacement vector
ur
u= u r eˆ r + uθ eˆθ + u z eˆ z ⇒ u= uθ
u z
Velocity vector vr
v= v r eˆ r + vθ eˆθ + v z eˆ z ⇒ u= vθ
v z
111
Cylindrical Coordinate System
Infinitesimal strain tensor
ε x′x′ ε x′y′ ε x′z′ ε rr ε rθ ε rz
ε=
1
2
{
[u ⊗ ∇ ] + [u ⊗ ∇ ]
T
}
≡ ε x′y′ ε y′y′ ε y′z′ =
ε
rθ εθθ εθ z
ε x′z′ ε y′z′ ε z′z′ ε rz εθ z ε zz
∂u 1 1 ∂u r ∂uθ uθ
ε rr = r ε= + −
2 r ∂θ
rθ
∂r ∂r r
1 ∂uθ u r 1 ∂u r ∂u z
εθθ
= + =ε rz +
r ∂θ r 2 ∂ z ∂r
∂u 1 ∂uθ 1 ∂u z
ε zz = z εθ z
= +
∂z
2 ∂z r ∂θ
x = r cos θ
x( r , θ , z ) ≡ y = r sin θ
z = z
112
Cylindrical Coordinate System
Strain rate tensor
d x′x′ d x′y′ d x′z′ d rr d rθ d rz
d=
1
2
{
[v ⊗ ∇] + [v ⊗ ∇]
T
}
≡ d x′y′ d y′y′
d y′z′ =
d
rθ dθθ dθ z
d x′z′ d y′z′ d z′z′ d rz dθ z d zz
∂v 1 1 ∂v r ∂vθ vθ
d rr = r d= + −
2 r ∂θ
rθ
∂r ∂r r
1 ∂vθ v r 1 ∂v r ∂v z
=
dθθ + =
d rz +
r ∂θ r 2 ∂ z ∂r
∂v 1 ∂vθ 1 ∂v z
d zz = z = +
∂z dθ z
2 ∂z r ∂θ
x = r cos θ
x( r , θ , z ) ≡ y = r sin θ
z = z
113
Spherical Coordinate System
x = r sin θ cos φ
=x x ( r ,θ , ϕ ) =
≡ y r sin θ sin φ r− coordinate line →
z = r cos θ
dV = r 2 sin θ dr dθ dφ
114
Spherical Coordinate System
Nabla operator ∂
∂ r
∂ 1 ∂ ∂ 1 ∂
eˆ φ ⇒ ∇ ≡ x = r sin θ cos φ
1
∇= eˆ r + eˆθ +
∂r r ∂θ r sin θ ∂φ r ∂θ = x x ( r , θ , φ ) ≡=
y r sin θ sin φ
z = r cos θ
1 ∂
r sin θ ∂φ
Velocity vector vr
v= v r eˆ r + vθ eˆθ + vφ eˆ φ ⇒ u= vθ
vφ
115
Spherical Coordinate System
Infinitesimal strain tensor
ε x′x′ ε x′y′ ε x′z′ ε rr ε rθ ε rφ
ε=
1
2
{[u ⊗ ∇ ] + [u ⊗ ∇ ]
T
}
≡ ε x′y′ ε y′y′ ε y′z′ =
εθ r εθθ εθφ
ε x′z′ ε y′z′ ε z′z′ ε rφ εθφ ε φφ
∂u r
ε rr =
∂r
1 ∂uθ u r
εθθ
= + x = r sin θ cos φ
r ∂θ r
1 ∂uφ uθ u =x x ( r , θ , φ ) ≡=
y r sin θ sin φ
ε ϕϕ= + cotgφ + r z = r cos θ
r sin θ ∂φ r r
1 1 ∂u r ∂uθ uθ
ε= + −
2 r ∂θ
rθ
∂r r
1 1 ∂u r ∂uφ uφ
=ε rφ + −
2 r sin θ ∂φ ∂r r
1 1 ∂uθ 1 ∂uφ uφ
=εθφ + − φ
2 r sin θ ∂φ r ∂θ
cotg
r
116
Spherical Coordinate System
Deformation rate tensor
d x′x′ d x′y′ d x′z′ d rr d rθ d rφ
d=
1
2
{[v ⊗ ∇] + [v ⊗ ∇]
T
}
≡ d x′y′ d y′y′
d y′z′ =
d rθ dθθ dθφ
d x′z′ d y′z′ d z′z′ d rφ dθφ dφφ
∂v r
d rr =
∂r x = r sin θ cos φ
1 ∂vθ v r =x x ( r , θ , φ ) ≡=
y r sin θ sin φ
=
dθθ + z = r cos θ
r ∂θ r
1 ∂vφ vθ v
dφφ = + cotgϕ + r
r sin θ ∂φ r r
1 1 ∂v r ∂vθ vθ
d= + −
2 r ∂θ
rθ
∂r r
1 1 ∂v r ∂vφ vφ
= + −
2 r sin θ ∂φ
d rφ
∂r r
1 1 ∂vθ 1 ∂vφ vφ
= + − φ
2 r sin θ ∂φ r ∂θ
dθφ cotg
r
117
Chapter 2
Strain
rs
ee
s gin
2.1 Introduction
t d le En
r
ba
ge ro or
eS m
ci
f
Definition 2.1. In the broader context, the concept of deformation no
ra
C d P cs
longer refers to the study of the absolute motion of the particles as
b
a
i
seen in Chapter 1, but to the study of the relative motion, with respect
an an n
le
liv or ec
M
.A
d
uu
reference configuration Ω0 occupies the point in space P in the present config-
er
tin
.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.
ee
s gin
Equation (2.2) defines the material deformation gradient tensor F (X,t) 1 .
t d le En
⎧
⎪ not
⎨F = x⊗∇
r
ba
Material deformation
ge ro or
∂ xi
eS m
(2.3)
gradient tensor ⎪
⎩ Fi j = i, j ∈ {1, 2, 3}
ci
∂ Xj f
ra
C d P cs
b
a
i
an an n
⎡ ⎤
∂ x1 ∂ x1 ∂ x1
le
liv or ec
⎡ ⎤ ⎢ ⎥
x ⎢ ∂X ∂ X2 ∂ X3 ⎥
⎢ 1⎥ ∂
M
⎢ 1 ⎥
.A
∂ ∂ ⎢ ∂ x2 ∂ x2 ∂ x2 ⎥
[F] = x ⊗ ∇ = ⎢ ⎥
⎣ x2 ⎦ ∂ X1 , ∂ X2 , ∂ X3 = ⎢ ⎥ . (2.4)
⎢ ∂ X1 ∂ X3 ⎥
m
∂ X2
⎢ ⎥
d
x3 ⎣ ∂x ∂ x3 ⎦
uu
T ∂ x3
e
3
X Th
∇
er
∂ ∂ X2 ∂ X3
tin
[x] X1
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
s gin
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
⎧
r
ba
⎪
ge ro or
not
⎨ F−1 = X ⊗ ∇
eS m
ci
Spatial deformation
∂ Xi f (2.7)
ra
gradient tensor ⎪ −1
⎩ Fi j = i, j ∈ {1, 2, 3}
C d P cs
∂xj
b
a
i
an an n
y ha
le
liv or ec
(2.6) and (2.7) as F−1 , is in effect the inverse of the (material) defor-
.A
∂ xi ∂ Xk ∂ xi not
uu
e
= = δi j =⇒ F · F−1 = 1 ,
X Th
∂ Xk ∂ x j ∂xj
er
tin
Fik F −1
on
.O
kj
C
∂ Xi ∂ xk ∂ Xi not
F−1 · F = 1 .
©
= = δi j =⇒
∂ xk ∂ X j ∂ Xj
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 = 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
[X] ∂ x1 ∂ x2 ∂ x3
rs
ee
Example 2.1 – At a given time, the motion of a continuous medium is defined
s gin
by ⎧
⎨ x1 = X1 − AX3
⎪
t d le En
x2 = X2 − AX3 .
⎪
r
⎩
ba
x3 = −AX1 + AX2 + X3
ge ro or
eS m
ci
f
Obtain the material deformation gradient tensor F (X,t) at this time. By
ra
C d P cs
b
a
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
T ⎢
⎥ ∂
uu
∂ ∂
e
not
F = x ⊗ ∇ ≡ [x] ∇ = ⎣ ⎢ ⎥
X2 − AX3 ⎦ ∂ X1 , ∂ X2 , ∂ X3
X Th
er
tin
−AX1 + AX2 + X3
on
.O
⎡ ⎤
−A
C
1 0
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 ⎢ ⎥
X (x,t) ≡ ⎢ ⎥
⎣ X2 = A x1 + 1 − A x2 + Ax3 ⎦ .
2 2
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
s gin
⎡ ⎤⎡ ⎤ ⎡ ⎤
0 −A 1 + A2 −A2
t d le En
1 A 1 0 0
not ⎢ ⎥⎢ ⎥ ⎢ ⎥ not
F · F−1 ≡ ⎣ 0 1 −A ⎦ ⎣ A2 1 − A2 A⎦ = ⎣0 1 0⎦ ≡ 1 .
r
ba
ge ro or
eS m
−A A 1 A −A 1 0 0 1
ci
f
ra
C d P cs
b
a
i
an an n
y ha
2.3 Displacements
le
liv or ec
M
.A
particle.
e
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
ee
s gin
Figure 2.2: Displacement of a particle.
t d le En
r
2.3.1 Material and Spatial Displacement Gradient Tensors
ba
ge ro or
eS m
ci
Differentiation with respect to the material coordinates of the displacement vec-
f
ra
C d P cs
tor Ui defined in (2.9) results in
b
a
i
∂Ui ∂ xi ∂ Xi
an an n
de f
= − = Fi j − δi j = Ji j , (2.11)
y ha
∂ Xj ∂ Xj ∂ Xj
le
liv or ec
Fi j δi j
M
.A
⎧
uu
⎪
e
de f
⎨ J (X,t) = U (X,t) ⊗ ∇ = F − 1
X Th
er
Material displacement
tin
.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
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
⎧
⎪
⎨ u = j · dx
⎪ ∂ ui (2.16)
⎩ dui = dx j = ji j dx j i, j ∈ {1, 2, 3}
rs
∂xj
ee
s gin
2.4 Strain Tensors
t d le En
Consider now a particle of the continuous medium that occupies the point in
r
space P in the material configuration, and another particle Q√in its differen-
ba
ge ro or
eS m
dS = dX · dX) from
ci
tial neighborhood separated a segment dX (with length
√
f
ra
the previous paticle, being dx (with length ds = dx · dx) its counterpart in
C d P cs
b
a
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
≡ [dx]T [dx] = [F · dX]T [F · dX] ≡ dX · FT · F · dX
not
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 ,
⎧ −1 T −1
⎪
⎪ (dS) 2
= dX · dX
not
≡ [dX] T
[dX] = F · dx F · dx =
⎪
⎪
⎪
⎨ ≡ dx · F−T · F−1 · dx ,
not
(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
ee
s gin
Figure 2.3: Differential segments in a continuous medium.
t d le En
r
2.4.1 Material Strain Tensor (Green-Lagrange Strain Tensor)
ba
ge ro or
eS m
ci
Subtracting expressions (2.18) and (2.19) results in
f
ra
C d P cs
b
a
(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
M
.A
de f
= 2E
m
tensor as follows.
e
X Th
er
⎧
tin
⎪
Material ⎨ E (X,t) = 1 FT · F − 1
on
.O
2 (2.21)
(Green-Lagrange)
⎪ 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.
⎧
s gin
⎪
Spatial ⎨ e (x,t) = 1 1 − F−T · F−1
t d le En
(Almansi) 2 (2.23)
⎪
strain tensor ⎩ ei j (x,t) =
1
δi j − Fki−1 Fk−1 i, j ∈ {1, 2, 3}
r
2 j
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
Remark 2.4. The spatial strain tensor e is symmetric. Proof is ob-
an an n
⎧
⎪
T = 1 1 − F−T · F−1 T = 1 1T − F−1 T · F−T T =
le
⎪
liv or ec
⎪
⎪ e
⎨ 2 2
M
.A
1 −T −1
⎪
⎪ = 1−F ·F =e,
⎪ 2
m
⎪
⎩e = e
d
ij ji i, j ∈ {1, 2, 3} .
uu
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
2A3
2
−2A − 2A −2A
ee
3 2A3 2
s gin
Observe that E = e.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Remark 2.5. The material strain tensor E and the spatial strain ten-
b
a
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
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
Variation of Distances: Stretch and Unit Elongation 51
1 1
E= 1 + JT · (1 + J) − 1 = J + JT + JT · J
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
ee
1 1
s gin
e= 1 − 1 − jT · (1 − j) = j + jT − jT · j
2 2
t d le En
(2.25)
1 ∂ ui ∂ u j ∂ uk ∂ uk
ei j = + − i, j ∈ {1, 2, 3}
r
2 ∂ x j ∂ xi ∂ xi ∂ x j
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
.A
sponding positions in the present configuration are given by the points in space
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
er
tin
on
.O
is the length of the deformed differential segment P Q per unit of
length of the original differential segment PQ.
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
ee
s gin
Figure 2.4: Differential segments and unit vectors in a continuous medium.
t d le En
r
ba
ge ro or
eS m
ci
f
Definition 2.4. The unit elongation, elongation ratio or extension of
ra
C d P cs
a material point P (or a spatial point P ) in the material direction T
b
a
i
5
(or spatial direction t ) is the increment of length of the deformed
an an n
y ha
.A
m
Δ 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
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.
ee
s gin
t d le En
2.5.1 Stretches, Unit Elongations and Strain Tensors
r
ba
ge ro or
Consider equations (2.21) and (2.22) as well as the geometric expressions
eS m
ci
dX = T dS and dx = t ds (see Figure 2.4). Then,
f
ra
C d P cs
⎧
b
a
⎪
⎪ (ds)2 − (dS)2 = 2 dX · E · dX = 2 (dS)2 T · E · T
i
⎪
an an n
⎪
⎨
y ha
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
ds t ds t
m
√
X Th
er
tin
ds 2 λ = 1 + 2T · E · T
−1 = λ −1 = 2 T·E·T ⇒
2
√ (2.30)
dS ε = λ − 1 = 1 + 2T · E · T − 1
on
.O
C
λ
©
1
λ=√
dS 2 1 2
1 − 2t · e · t
1− = 1− = 2 t·e·t ⇒ 1 (2.31)
ds λ ε = λ −1 = √ −1
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
not ⎢ ⎥
s gin
e (x,t) ≡ ⎣ 0 0 0 ⎦.
−tetz t (2etz − et )
t d le En
0
r
Calculate the length, at time t = 0, of the segment that at time t = 2 is recti-
ba
ge ro or
eS m
linear and joins points a ≡ (0, 0, 0) and b ≡ (1, 1, 1).
ci
f
ra
C d P cs
b
a
Solution
i
an an n
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
1 ds 1
uu
λ=√ = =⇒ dS = ds .
e
λ
X Th
1 − 2 t · e · t dS
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
Variation of Angles 55
√
To obtain the stretch in the direction t ≡ [1, 1, 1]T / 3, the expression t · e · t
not
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=! =⇒ λ" =! =√ .
ee
1 + 23 tet t=2
1+ 3e
4 2 3 + 4e2
s gin
The length at time t = 0 of the segment AB is
t d le En
# B # b # b
1 1 1 1√
r
lAB = dS = ds = ds = lab =
ba
3
ge ro or
λ λ λ λ
eS m
A a a
ci
f
ra
lab
C d P cs
b
a
i
and replacing the expression obtained above for the stretch at time t = 2
an an n
y ha
finally results in $
le
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
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
Figure 2.5: Angles between particles in a continuous medium.
ci
f
ra
C d P cs
⎧ ⎧
b
a
⎨ dX(1) = dS(1) T(1) ⎨ dx(1) = ds(1) t(1) ,
i
an an n
=⇒ (2.33)
y ha
⎧
⎧
m
⎪
⎪ 1
d
λ (1)
e
=⇒ (2.34)
X Th
⎪
tin
⎩ dS(2) = 1
ds(2) .
λ (2)
on
.O
Expanding now the scalar product6 of the vectors dx(1) and dx(2) ,
C
T
= F · dX(1) F · dX(2) ≡ dX(1) · FT · F · dX(2) = dX(1) · (2E + 1) · dX(2)
not
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
In an analogous way, operating on the reference configuration, the angle Θ
ee
between the differential segments dX(1) and dX(2) (in terms of t(1) , t(2) and e )
s gin
is obtained,
t d le En
t(1) · (1 − 2e) · t(2)
cosΘ = $ $ .
r
(2.38)
ba
ge ro or
1 − 2t(1) · e · t(1) 1 − 2t(2) · e · t(2)
eS m
ci
f
ra
C d P cs
b
a
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.
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)
EXZ EY Z EZZ E13 E23 E33
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.
s gin
t d le En
Consequently,
r
ba
⎡ ⎤⎡ ⎤
ge ro or
eS m
E11 E12 E13 1
ci
⎣ f⎦ ⎣ 0 ⎦ = E11 .
ra
not T
T · E · T ≡ [T] [E] [T] = [1, 0, 0] E12 E22 E23 (2.40)
C d P cs
b
a
E13 E23 E33 0
i
an an n
y ha
.A
√ √ √
d
√ √ √
X Th
er
(2.41)
√ √ √
λ3 = 1 + 2E33 = 1 + 2EZZ ⇒ εZ = λZ − 1 = 1 + 2EZZ − 1
on
.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-
s gin
rations.
t d le En
r
Consider now the angle between segments PQ (parallel to the X1 -axis) and PR
ba
ge ro or
eS m
(parallel to the X2 -axis), where Q and R are two particles in the differential neigh-
ci
f
ra
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)
a
i
between the segments in the reference configuration is known, the angle θ in
an an n
y ha
the present configuration can be determined using (2.37) and taking into ac-
count their orthogonality ( T(1) · T(2) = 0 ) and the equalities T(1) · E · T(1) = E11 ,
le
liv or ec
.A
cos θ = $ $
d
uu
(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
ΔΘXY = θxy − ΘXY = − arcsin √ √ . (2.44)
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
ΔΘXY = − arcsin √ √
1 + 2EXX 1 + 2EYY
2EXZ
ΔΘXZ = − arcsin √ √ . (2.45)
1 + 2EXX 1 + 2EZZ
2EY Z
ΔΘY Z = − arcsin √ √
1 + 2EYY 1 + 2EZZ
rs
ee
Remark 2.10. The components EXY , EXZ and EY Z (or E12 , E13 and
s gin
E23 ) of the tensor E (denoted angular strains) contain the informa-
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.
r
ba
ge ro or
eS m
• If EXY = 0 : The deformation does not produce a variation in the
ci
f
angle between the two segments initially oriented in the direc-
ra
C d P cs
tions X and Y .
b
a
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
ee
s gin
Figure 2.8: Physical interpretation of the material strain tensor.
t d le En
r
ented in the direction of the coordinate axes in the present configuration,
ba
ge ro or
eS m
ci
f
ra
C d P cs
1 1 1
b
a
λ1 = √ =√ ⇒ εx = √ −1
i
1 − 2e11 1 − 2exx 1 − 2exx
an an n
y ha
1 1 1
λ2 = √ =$ ⇒ εy = $ −1 ,
le
(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
Δ θxy = −ΘXY = − arcsin √ $
2 1 − 2exx 1 − 2eyy
π 2exz .
Δ θxz = −ΘXZ = − arcsin √ √ (2.48)
2 1 − 2exx 1 − 2ezz
π 2eyz
Δ θyz = −ΘY Z = − arcsin $ √
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.
ee
s gin
2.8 Polar Decomposition
t d le En
The polar decomposition theorem of tensor analysis establishes that, given a
r
ba
second-order tensor F such that |F| > 0, there exist an orthogonal tensor Q 7 and
ge ro or
eS m
ci
two symmetric tensors U and V such that8
f
ra
C d P cs
not √
⎫
b
a
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
rotation
©
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
stretching
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:
s gin
• When applied on any vector x, the result is another vector
y = Q · x with the same modulus,
t d le En
y
2 = y·y ≡ [y]T ·[y] = [Q · x]T ·[Q · x] ≡ x·QT · Q ·x = x·x =
x
2 .
not not
r
ba
ge ro or
eS m
1
ci
f
ra
• The result of multiplying (mapping) the orthogonal tensor Q to
C d P cs
b
a
two vectors x(1) and x(2) with the same origin and that form an
i
an an n
.A
d
uu
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
ee
s gin
t d le En
Figure 2.10: Polar decomposition.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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) ⎥
dV0 = dX(1) × dX(2) · dX(3) = det ⎢ (2) (2) ⎥
⎣ dX1 dX2 dX3 ⎦ = |M| ,
rs
ee
(3) (3) (3)
dX1 dX2 dX3
s gin
[M]
t d le En
⎡ (1) (1) (1)
⎤
r
dx1 dx2 dx3
ba
ge ro or
⎢ (2) ⎥
eS m
dVt = dx(1) × dx(2) · dx(3) = det ⎢ (2) (2) ⎥
⎣ dx1 dx2 dx3 ⎦ = |m| ,
ci
(2.53)
f
ra
C d P cs
(3) (3) (3)
b
a
dx1 dx2 dx3
i
an an n
[m]
y ha
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 (2.54)
m
d
uu
e
er
tin
.O
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
ee
s gin
Figure 2.11: Variation of a volume differential element.
t d le En
r
2.10 Area Variation
ba
ge ro or
eS m
ci
f
Consider an area differential dA associated with a particle P in the reference
ra
C d P cs
configuration and its variation along time. To define this area differential, con-
b
a
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
e
space P and will have an area differential da associated with it which, in turn,
X Th
normal vector in the spatial configuration. Consider also the positions of the
other particles Q , R and S and their relative position vectors dx(1) , dx(2) and
on
.O
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)
dH dA (2.56)
(3)
dVt = dh da = dx · n da = dx · n da = da · dx(3)
(3)
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.
s gin
t d le En
Comparing the first and last terms12 in (2.57) and considering that the relative
r
position of particle S can take any value (as can, therefore, vector dX(3) ), finally
ba
ge ro or
eS m
ci
yields
f
ra
da · F = |F| dA =⇒ da = |F| dA · F−1 .
C d P cs
(2.58)
b
a
i
an an n
To obtain the relation between the two area differential scalars, dA and da,
y ha
( (
M
.A
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
ee
s gin
Figure 2.13: Infinitesimal strain in the continuous medium.
t d le En
r
ba
ge ro or
eS m
In accordance with the first hypothesis, the reference configuration Ω0 and
ci
f
the present configuration Ωt are very close together and are considered to be
ra
C d P cs
b
indistinguishable from one another. Consequently, the material and spatial co-
a
i
ordinates coincide and discriminating between material and spatial descriptions
an an n
y ha
.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 "
" " ∀ i, j ∈ {1, 2, 3} .
"∂xj "
1 (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
⎪
⎩
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
s gin
1 ∂ ui ∂ u j ∂ uk ∂ uk ∼ 1 ∂ ui ∂ u j (2.64)
⎪
⎪ e ij = + − = + .
⎪
⎪ ∂ ∂ ∂ ∂ ∂ ∂
t d le En
⎪ 2 x j x i x i x j 2 x j xi
⎩
1
r
ba
ge ro or
eS m
ci
Equations (2.63) and (2.64) allow defining the infinitesimal strain tensor (or
f
ra
small strain tensor) ε as13
C d P cs
b
a
⎧
i
an an n
⎪
⎪ 1 not
y ha
⎪
⎨ε = J + JT = ∇s u
le
Infinitesimal 2
liv or ec
(2.65)
strain tensor ⎪
⎪ 1 ∂ ui ∂ u j
⎪ ε
⎩ ij = + i, j ∈ {1, 2, 3}
M
.A
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.
s gin
t d le En
Solution
r
The equation of motion is given by
ba
ge ro or
eS m
⎧
ci
⎨ x1 = X1 − AX3
⎪ f
ra
C d P cs
b
a
x2 = X2 − AX3 ,
⎪
i
⎩
an an n
x3 = −AX1 + AX2 + X3
y ha
le
liv or ec
⎡ ⎤
.A
U1 = −AX3
not ⎢ ⎥
m
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
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
−2A − 2A 3 3 −2A 2
rs
2A
ee
Neglecting
4 the second-order
and higher-order infinitesimal terms
A
A3
A2
A results in
s gin
⎡ ⎤ ⎡ ⎤
t d le En
0 0 −A 0 0 −A
not ⎢ ⎥ not ⎢ ⎥
0 −A ⎦ =⇒ E = e = ε ,
r
E≡⎣ 0 0 −A ⎦ and e≡⎣ 0
ba
ge ro or
eS m
−A A −A A
ci
0 0
f
ra
C d P cs
b
a
which is in accordance with Remark 2.14.
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
1 + 2t · ε · t ∼
tin
λt = = 1+t·ε ·t
(2.66)
on
x
.O
εt = λt − 1 = t · ε · t
C
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
s gin
parallel to the coordinate axis x1 ≡ x. The stretch λx and the unit elongation εx
in this direction are, according to (2.66) with t = [1, 0, 0]T ,
t d le En
λx = 1 + t · ε · t = 1 + εxx =⇒ εx = λx − 1 = εxx .
r
(2.68)
ba
ge ro or
eS m
ci
This allows assigning to the component εxx ≡ ε11 the physical meaning of unit
f
ra
elongation εx in the direction of the coordinate axis x1 ≡ x. A similar interpre-
C d P cs
b
a
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-
M
.A
the coordinate directions x and y, respectively. Then, these two segments form
d
an angle Θxy = π/2 in this configuration. Applying (2.43), the increment in the
uu
e
er
tin
π εxy ∼
Δ θxy = θxy − = −2 arcsin $ $ = −2 arcsin εxy = −2εxy ,
on
.O
2 1 + 2εxx 1 + 2εyy
εxy
C
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 = − Δ θxy ; εxz = − Δ θxz ; εyz = − Δ 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
engineering notation
⎡
scientific notation
⎤
⎡ ⎤ ⎡ ⎤
ε11 ε12 ε13 εxx εxy εxz εx 2 γxy 2 γxz
1 1
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
ε22 ε23 ⎦ ≡ ⎣ εxy εyy εyz ⎦ ≡ ⎢ ⎥
not
ε ≡ ⎣ ε12
rs
⎣2
1
γxy ε y
1
2 γyz ⎦ (2.72)
ee
ε13 ε23 ε33 εxz εyz εzz
2 γxz 2 γyz εz
1 1
s gin
t d le En
r
Remark 2.17. The components in the main diagonal of the strain ten-
ba
ge ro or
eS m
sor (named longitudinal strains) are denoted by ε(•) and coincide
ci
f
ra
with the unit elongations in the directions
of the coordinate axes.
C d P cs
Positive values of longitudinal strains ε(•) > 0 correspond to an
b
a
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
T
ε ∈ R6 ε = εx , εy , εx , γxy , γxz , γyz
(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
1
1
ee
where T(1) and T(2) are the unit vectors
s gin
( ( (in the( directions of PQ and PR and,
(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 Δ θ , the follow-
ing holds true16 .
r
ba
ge ro or
eS m
ci
f
cos θ = cos (Θ + Δ θ ) = cosΘ · cos Δ θ − sinΘ · sin Δ θ =
ra
C d P cs
b
a
≈1 ≈ Δθ
i
an an n
y ha
= cosΘ
le
liv or ec
.A
≈1 ≈1
d
uu
er
tin
.O
Δθ = − =− , (2.76)
©
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 .
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.
s gin
t d le En
ing (2.12) and the infinitesimal character of the components of the tensor J
r
(see (2.61)), tensor U in (2.49) can be written as17
ba
ge ro or
eS m
$ $ ⎫
ci
U = FT · F = (1 + JT ) · (1 + J) = f ⎪
ra
⎪
C d P cs
⎪
⎬
b
a
$ $
i
1
= 1 + J + J + J · J ≈ 1 + J + J = 1 + J + J ⎪ =⇒ U = 1 + ε .
an an n
T T T T
⎪
y ha
⎪
2
⎭
J ε
le
liv or ec
(2.77)
M
.A
1
uu
U−1 = (1 + ε )−1 = 1 − ε = 1 −
e
J + JT . (2.78)
X Th
2
er
tin
.O
⎫
1 ⎪
C
−1
Q = F · U = (1 + J) · 1 − J + J T = ⎪
⎪
⎪
⎪
©
2 ⎬
1 1 1 =⇒ Q = 1 + Ω .
= 1 + J − J + JT − J · J + JT = 1 + J − JT ⎪ ⎪
⎪
2 2 2 ⎪ ⎪
⎭
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
⎧
ee
⎨ Ω T = 1 J − JT T = 1 JT − J = −Ω Ω
s gin
2 2 .
⎩ Ω = −Ω i, j ∈ {1, 2, 3}
ji ij
t d le En
Consequently, the terms in the main diagonal of Ω are zero, and its
r
ba
ge ro or
eS m
matrix of components has the structure
ci
⎡ f ⎤
ra
C d P cs
0 Ω12 −Ω31
b
a
Ω] = ⎣ −Ω12 Ω23 ⎦ .
i
[Ω 0
an an n
y ha
Ω31 −Ω23 0
le
liv or ec
M
.A
⎡ ⎤
on
.O
Infinitesimal ∂ u 3 ∂ u2
rotation vector: ⎡ ⎤ ⎡ ⎤ ⎢ − ⎥
C
θ1 −Ω23 ⎢ ∂ x2 ∂ x3 ⎥
⎢ ⎥ 1
©
not ⎢ ⎥ ⎢ ⎥ 1⎢ ⎥ de f
θ ≡ ⎣ θ2 ⎦ = ⎣ −Ω31 ⎦ = ⎢ ∂ u1 − ∂ u3 ⎥ = ∇ × u . (2.81)
2 ⎢ ∂ x3 ∂ x1 ⎥ 2
θ3 −Ω12 ⎢ ⎥
⎣ ∂u ∂ u1 ⎦
2
−
∂ 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,
ee
⎡ ⎤⎡ ⎤ ⎡ ⎤
s gin
0 Ω12 −Ω31 r1 Ω12 r2 − Ω31 r3
not ⎢ ⎥⎢ ⎥ ⎢ ⎥
Ω · r ≡ ⎣ −Ω12 0 Ω23 ⎦ ⎣ r2 ⎦ = ⎣ −Ω12 r1 + Ω23 r3 ⎦ ,
t d le En
Ω31 −Ω23 0 r3 Ω31 r1 − Ω23 r2
r
ba
ge ro or
⎡ ⎤ ⎡ ⎤
eS m
ci
ê1 ê2 ê3 ê1 ê2 ê3
f
ra
not ⎢ ⎥ ⎢ ⎥
C d P cs
θ × r ≡ ⎣ θ1 θ2 θ3 ⎦ = ⎣ −Ω23 −Ω31 −Ω12 ⎦ =
b
a
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.
s gin
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
r
ba
ge ro or
stretching transforms this vector into vector dx as follows.
eS m
ci
f
ra
stretching rotation
C d P cs
b
a
dx = F · dX = (1 + ε + Ω) · dX = ε · dX + (1 + Ω) · dX
i
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
ee
s gin
t d le En
2.12 Volumetric Strain
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Definition 2.6. The volumetric strain is the increment produced by
b
a
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
⎢ ⎥
s gin
|F| = |Q · U| = |Q| |U| = |U| = |1 + ε | = det ⎣ εxy 1 + εyy εyz ⎦ ,
εxz εyz 1 + εzz
t d le En
(2.86)
r
ba
where (2.77) has been considered. Taking into account that the components of ε
ge ro or
eS m
ci
are infinitesimal, and neglecting in the expression of its determinant the second-
f
ra
order and higher-order infinitesimal terms, results in
C d P cs
b
a
⎡ ⎤
i
an an n
⎢ ⎥
|F| = det ⎣ εxy 1 + εyy εyz ⎦ = 1 + εxx + εyy + εzz +O ε 2 ≈ 1 + Tr (εε ) .
le
liv or ec
.A
(2.87)
m
Then, introducing (2.87) into (2.85) yields, for the infinitesimal strain case
d
⎫
uu
e
er
tin
.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)
⎪
⎪
⎪ ∂ vi
s gin
⎪
⎩ dvi = dx j = li j dx j i, j ∈ {1, 2, 3}
∂xj
t d le En
where the spatial velocity gradient tensor l (x,t) has been introduced.
r
ba
⎧
ge ro or
eS m
⎪ de f ∂ v (x,t)
ci
⎪ l (x,t) =
⎪
⎪ f
ra
⎪
⎨ ∂x
C d P cs
b
a
Spatial velocity
l = v⊗∇
i
(2.91)
gradient tensor ⎪
an an n
⎪
⎪ ∂ vi
y ha
⎪
⎪ i, j ∈ {1, 2, 3}
⎩ li j =
∂xj
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
82 C HAPTER 2. S TRAIN
rs
ij i, j ∈ {1, 2, 3}
2 ∂ x j ∂ xi
ee
rate (2.93)
⎪
tensor ⎪
⎪
⎪
⎡ ⎤
s gin
⎪
⎪ d 11 d 12 d 13
⎪
⎪
⎪
⎪ [d] = ⎣ d12 d22 d23 ⎦
⎩
t d le En
d13 d23 d33
r
ba
ge ro or
eS m
ci
and w is an antisymmetric tensor denominated rotation rate tensor or spin ten-
f
ra
sor, whose expression is
C d P cs
b
a
i
⎧
an an n
⎪
⎪ 1 1
y ha
de f
⎪
⎪ w = skew (l
l ) = l − l T not
= (v ⊗ ∇ − ∇ ⊗ v) = ∇a v
⎪
⎪
le
⎪ 2 2
liv or ec
Rotation⎪
⎪
⎨ w = 1 ∂ vi − ∂ v j
⎪
i, j ∈ {1, 2, 3}
M
.A
rate ij
2 ∂ x j ∂ xi (2.94)
(spin) ⎪
⎪ ⎡ ⎤
tensor ⎪
m
⎪
⎪ 0 w12 −w31
d
⎪
⎪
uu
⎪
⎪ [w] = ⎣ −w12 0 w23 ⎦
e
⎪
⎩
X Th
er
w31 −w23 0
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
Strain Rate 83
rs
dt dt
ee
Replacing (2.2) into (2.97) results in22
s gin
.
dX · E · dX = dx · d · dx ≡ [dx]T [d] [dx] = [dX]T FT · d · F [dX]
not
t d le En
. .
=⇒ dX · FT · d · F − E · dX = 0 ∀ dX =⇒ FT · d · F − E = 0
r
ba
ge ro or
eS m
.
ci
E = FT · d · F .
f (2.98)
ra
C d P cs
b
a
i
an an n
y ha
.
the material strain tensor E (X,t), providing a physical interpreta-
M
.A
exactly the same. Both tensors will coincide in the following cases:
uu
"
e
= 1.
er
tin
t=t0
∂x
on
∂X
C
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 = 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.
ee
s gin
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
r
defined using only three different components), the vector
ba
ge ro or
eS m
⎡ ⎤
ci
∂ v 2 ∂ v3 f
ra
⎢−
C d P cs
− ⎥ ⎡ ⎤
b
a
⎢ ⎥
⎢ ∂ x3 ∂ x2 ⎥ −w23
i
an an n
not 1 ⎢ ∂ v 3 ∂ v1 ⎥
1 1
ω = rot (v) = ∇ × v ≡ ⎢ ⎥=⎢ ⎥
y ha
⎢ − − ⎥ ⎣ −w31 ⎦ (2.99)
2 2 2⎢ ∂ x1 ∂ x3 ⎥
le
⎢ ⎥ −w12
liv or ec
⎣ ∂ v 1 ∂ 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)
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
23 Observe 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.
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ω ).
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
ee
s gin
where (2.91) has been taken into account for the velocity gradient tensor l . To
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
r
d dF −1 d F−1
ba
ge ro or
−1 −1
F · F = 1 =⇒ F·F = ·F +F· =0
eS m
ci
−1 dt dt dt
f
ra
d F .
C d P cs
=⇒ = −F−1 · F · F−1 = −F−1 · l · F · F−1 = −F−1 · l =⇒
b
a
dt
i
an an n
l ·F 1
y ha
−1
le
liv or ec
d F
= −F−1 · l
M
dt
.A
(2.103)
dFi−1
= Fik−1 lk j
j
m
i, j ∈ {1, 2, 3}
d
dt
uu
e
X Th
er
tin
.O
1 T . 1 .T .
©
dE
E= F · F − 1 =⇒ =E= F · F + FT · F =
2 dt 2
1 T T 1
= F · l · F + FT · l · F = FT · l + l T · F = FT · d · F
2 2
.
=⇒ E = F · d · F .
T 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 d −1
e= 1−F ·F ⇒ =e=− F · F + F−T · 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
ee
The volume differential dV (X,t) associated with a certain particle P varies
s gin
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
r
d d |F|
ba
ge ro or
dV (X,t) = |F (X,t)| dV0 (X) =⇒ dV (t) = dV0 .
eS m
(2.106)
ci
dt dt
f
ra
C d P cs
Therefore, the material derivative of the determinant of the deformation gradient
b
a
tensor |F| is27
i
an an n
y ha
dt dFi j dt dt
M
.A
lik Fk j [ F·F−1 ] =δ
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
©
27 The 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.
ee
s gin
pression, results in
t d le En
d d d |F| d −1
(da) = |F| · dA · F−1 = dA · F−1 + |F| · dA =
r
F
ba
ge ro or
dt dt dt
dt
eS m
ci
|F| ∇ · v −F−1 · l
f
ra
C d P cs
b
a
= (∇ · v) |F| dA · F−1 − |F| dA · F−1 · l =⇒
i
an an n
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
s gin
basis {êa , êb , êc } (
êa
=
êb
=
êc
= 1), whose components are
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
r
orientation of the curvilinear basis changes at each point in space
ba
ge ro or
(êm ≡ êm (x) m ∈ {a, b, c}). Therefore, for the purposes here, an
eS m
ci
f
orthogonal curvilinear coordinate system can be considered as a mo-
ra
bile Cartesian coordinate system {x , y , z } associated with a curvi-
C d P cs
b
a
linear basis {êa , êb , êc } (see Figure 2.25).
i
an an n
y ha
le
liv or ec
M
.A
er
tin
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
va vx Taa Tab Tac Tx x Tx y Tx z
on
.O
not ⎢ ⎥ ⎢ ⎥ not ⎢ ⎥ ⎢ ⎥
v ≡ ⎣ vb ⎦ ≡ ⎣ vy ⎦ T ≡ ⎣ Tba Tbb Tbc ⎦ ≡ ⎣ Ty x Ty y Ty z ⎦
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, ∇
ee
T
s gin
∂ 1 ∂ ∂ ∂
not 1 ∂ ∂
∇ = êr + êθ + êz =⇒ ∇≡ , , (2.111)
∂r r ∂θ ∂z ∂r r ∂θ ∂z
t d le En
r
ba
ge ro or
⎡ ⎤
eS m
x = r cos θ
ci
f x (r, θ , z) ≡ ⎣ y = r sin θ ⎦
not
ra
C d P cs
b z=z
a
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
u ≡ [ur , uθ , uz ]T
not
u = ur êr + uθ êθ + uz êz =⇒ (2.112)
v ≡ [vr , vθ , vz ]T
not
v = vr êr + vθ êθ + vz êz =⇒ (2.113)
rs
not ⎢ ⎥ ⎢ ⎥
ε= (u ⊗ ∇) + (u ⊗ ∇)T ≡ ⎣ εx y εy y εy z ⎦ = ⎣ εrθ εθ θ εθ z ⎦
ee
2
εx z εy z εz z εrz εθ z εzz
s gin
∂ ur 1 ∂ uθ ur ∂ uz
t d le En
εrr = εθ θ = + εzz =
∂r r ∂θ r ∂z
r
ba
ge ro or
1 1 ∂ ur ∂ uθ uθ ∂ ur ∂ uz
eS m
1
εrθ = + − εrz = +
ci
2 r ∂θ ∂r f ∂z ∂r
ra
r 2
C d P cs
b
a
1 ∂ u θ 1 ∂ uz
i
an an n
εθ z = + (2.114)
2 ∂z r ∂θ
y ha
le
liv or ec
.A
Figure (2.26).
m
⎡ ⎤ ⎡ ⎤
er
tin
.O
2
C
∂ vr 1 ∂ vθ vr ∂ vz
drr = dθ θ = + dzz =
∂r r ∂θ r ∂z
1 1 ∂ vr ∂ vθ vθ 1 ∂ v r ∂ 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
• Nabla operator, ∇
ee
s gin
∂ 1 ∂ 1 ∂ ∂not 1 ∂ 1 ∂ T
∇ = êr + êθ + êφ =⇒ ∇≡ , ,
∂r r ∂θ r sin θ ∂ φ ∂ r r ∂ θ r sin θ ∂ φ
t d le En
(2.117)
r
ba
ge ro or
eS m
ci
• Displacement vector, u, and velocity vector, v f
ra
C d P cs
b
a
not T
i
u = ur êr + uθ êθ + uφ êφ =⇒ u ≡ ur , uθ , uφ
an an n
(2.118)
y ha
not T
le
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
⎡ ⎤
x = r sin θ cos φ
on
.O
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
ee
1 1 ∂ ur ∂ uθ uθ 1 ∂ ur ∂ uφ uφ
s gin
1
εrθ = + − εrφ = + −
2 r ∂θ ∂r r 2 r sin θ ∂ φ ∂r r
t d le En
1 1 ∂ uθ 1 ∂ uφ uφ
r
εθ φ = + − cot φ (2.120)
ba
ge ro or
2 r sin θ ∂ φ r ∂θ r
eS m
ci
f
ra
The components of ε are presented on the corresponding differential element in
C d P cs
b
a
Figure 2.28.
i
an an n
y ha
le
• Strain rate tensor, d
liv or ec
⎡ ⎤ ⎡ ⎤
M
.A
2
uu
er
tin
∂ vr 1 ∂ vθ vr
drr = dθ θ = +
on
∂r r ∂θ
.O
r
1 ∂ vφ vθ
C
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
ee
s gin
t d le En
r
ba
ge ro or
Figure 2.28: Differential element in spherical coordinates.
eS m
ci
f
ra
C d P cs
b
a
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.
s gin
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.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
# A # 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 =⇒ Δ Φ xy = −γxy = −2ε xy = −q =⇒ εxy = .
π ⎪
⎪ 2
final angle = + Δ Φ xy ⎭
2
In addition, F is symmetric and the z-axis is a principal direction, therefore
⎡ ⎤
∂ ux ∂ ux ∂ ux
rs
⎡ ⎤ ⎢1+ ∂x ∂y ∂z ⎥
⎢ ⎥
ee
F11 F12 0
not ⎢ ⎥ not not ⎢ ∂ uy ∂ uy ∂ uy ⎥
F ≡ ⎣ F12 F22 0 ⎦ ≡ 1 + J ≡ ⎢ ⎢ ⎥,
s gin
1+ ⎥
⎢ ∂ x ∂ y ∂ z ⎥
0 0 F33 ⎣ ∂ uz ∂ uz ∂ uz ⎦
t d le En
1+
∂x ∂y ∂z
r
ba
ge ro or
eS m
which reveals the nature of the components of the displacement vector,
ci
⎧ f
ra
C d P cs
⎪ ∂ ∂
bux (x, y) ,
a
⎪
⎪ u x u y
⎨ = = 0 =⇒
i
an an n
∂z ∂z uy (x, y) ,
y ha
⎪
⎪ ∂ u z ∂ uz
⎪
⎩ = = 0 =⇒ uz (z) .
le
liv or ec
∂x ∂y
M
.A
d
1 ∂ ux ∂ u z
uu
εxz = + = 0 =⇒ εxz = 0
e
2 ∂z ∂x
X Th
er
tin
1 ∂ ux ∂ u z
εxz = + = 0 =⇒ εxz = 0
on
.O
2 ∂z ∂x ⎫
∂ uz ⎬
C
εzz = = λz − 1
∂z =⇒ εzz = s
©
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
not ⎢ ⎥ not ⎢ ⎥
dA0 ≡ ⎣ 0 ⎦ =⇒ dA0 · F−1 ≡ ⎢ ⎣
0 ⎥.
⎦
ee
1
dA0 dA0
s gin
1+s
t d le En
Then, taking into account that |F| = Tr (εε ) + 1, and neglecting second-order
terms results in
r
ba
⎫
ge ro or
eS m
dA = (1 + r) dA0 ⎬
ci
f
=⇒ εyy = r − p .
ra
1
C d P cs
dA = (1 + p + s + εyy ) dA0 ⎭
b
a
1+s
i
an an n
y ha
q
M
.A
⎢p 2
0⎥
not ⎢ q ⎥
ε ≡⎢ 0⎥⎥ .
m
⎢2 r− p
d
⎣ ⎦
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 Δ Φ
1 is satisfied and, in this problem, Δ Φ = π/4 ≈ 0.7854.
s gin
Observation: strains are dimensionless; in engineering, small strains are usually
t d le En
considered when these are of order 10−3 − 10−4 .
r
ba
ge ro or
eS m
b) The conditions in the statement of the problem must be imposed one by one:
ci
f
1. Considering that F (X,t) = F (t) and knowing that dx = F · dX, the latter
ra
C d P cs
b
can be integrated as
a
i
# # #
an an n
y ha
⎡ ⎤ ⎡ ⎤
F11 F12 F13 C1
M
.A
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
ee
0 0 F33
s gin
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
r
ba
ge ro or
expression can be integrated and the determinant of F can be moved outside the
eS m
ci
integral, # # #
f
ra
C d P cs
Vf = dV f = |F| dV0 = |F| dV0 = |F| V0 .
b
a
i
V V0 V0
an an n
le
liv or ec
3. The condition in the statement imposes that lAC, f inal = √p lAC, initial .
2
M
.A
⎡ ⎤⎡ ⎤ ⎡ ⎤
X Th
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
p p √
= √ lAC = √ 2 a = p a .
2 2
Therefore,
!
(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
4. The condition in the statement imposes that AOC f inal = 45◦ = π/4.
rs
Considering dX(1) ≡ [1, 0, 0] and dX(2) ≡ [0, 0, 1], the corresponding vectors
not not
ee
in the spatial configuration are computed as
s gin
⎡ ⎤⎡ ⎤ ⎡ ⎤
1 0 1 1 1
(1) not ⎢ ⎥⎢ ⎥ ⎢ ⎥
t d le En
(1)
dx = F · dX ≡ ⎣ 0 1 0 ⎦ ⎣ 0 ⎦ = ⎣ 0 ⎦ ,
r
0 0 p 0 0
ba
ge ro or
eS m
ci
⎡
⎤⎡ ⎤ ⎡ ⎤
f
ra
0 1 0 1
1
C d P cs
(2) not ⎢ ⎥⎢ ⎥ ⎢ ⎥
b
a
(2)
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) 2
M
.A
◦
cos AOC f inal = cos 45 = "" (1) "" "" (2) "" =
dx dx 2
m
is imposed, with
uu
e
.O
such that √
C
1 2 1
$ = =√ =⇒ p = ±1 .
©
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
which allows determining the displacement field in material and spatial descrip-
tions as
⎡ ⎤ ⎡ ⎤
rs
Z z
not ⎢ ⎥ not ⎢ ⎥
ee
U (X,t) = x − X ≡ ⎣ 0 ⎦ and u (x,t) ≡ ⎣ 0 ⎦ .
s gin
0 0
t d le En
c) The graphical representation of the deformed tetrahedron is:
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Solution
C d P cs
b
a
a) A uniform deformation implies that F (X,t) = F (t) , ∀t, X. The deformation
i
an an n
# #
M
.A
∂ U (X,t)
J= =⇒ dU = J dX =⇒ dU = J dX
∂X
m
# #
uu
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
UZ = 0 , ∀ X, Y, Z =⇒ J31 = J32 = J33 = C3 = 0
ee
s gin
From the second boundary condition:
"
t d le En
UX "X=0 = 0 , ∀ Y, Z =⇒ J12 = J13 = C1 = 0
r
ba
ge ro or
From the third boundary condition:
eS m
ci
" f
ra
δ
C d P cs
UX "X=L = δ , ∀ Y, Z =⇒ J11 L = δ ⇒ J11 =
b
a
L
i
an an n
y ha
Finally,
⎡ ⎤ ⎡ ⎤
⎡ ⎤ δ
le
δ
liv or ec
0 0 0 ⎢L ⎥X
⎢ L ⎥ not ⎢ ⎥ not ⎢ ⎥
M
.A
J≡⎢ ⎥
not
⎣0 0 0⎦; C ≡ ⎣0⎦ =⇒ U (X) = J · X + C ≡ ⎢ 0 ⎥ .
⎣ ⎦
m
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 ⎢ ⎥
F = 1+J ≡ ⎢
L ⎥ =⇒ |F| = 1 + δ > 0 =⇒ δ > −L .
⎣ 0 1 0 ⎦ 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
Spatial strain tensor: e= 1 − F−T · F−1
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: E=
F ·F−1
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
⎡ ⎤ ⎡ ⎤
s gin
δ
EXX 0 0 ⎢L 0 0 ⎥
not ⎢ ⎥
⎥ with EXX = δ + 1 δ
2
not ⎢ ⎥
t d le En
E≡⎢ ⎣ 0 0 0 ⎦ ; ε ≡ ⎢ 0 0 0 ⎥
L 2L 2 ⎣ ⎦
r
ba
ge ro or
0 0 0 0 0 0
eS m
ci
f
ra
C d P cs
b
a
i
an an n
le
liv or ec
Here,
M
.A
• EXX is a second-order
parabola that contains the
m
er
EXX = −1/2.
tin
.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
rs
a) The material description of v1 and the spatial description of v2 (consider
ee
t = 0 is the reference configuration).
s gin
b) The density distribution in both cases (consider ρ0 is the initial density).
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
r
velocity field v1 .
ba
ge ro or
eS m
d) Repeat c) for configurations close to the reference configuration (t → 0).
ci
f
ra
C d P cs
e) Prove that the two strain tensors coincide for the conditions stated in d).
b
a
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
er
tin
on
.O
⎡ ⎤
©
0 tetX 0
⎢ ⎥
E≡⎢ 0 ⎥
not
⎦.
tX
⎣ te 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:
s gin
U (X,Y, Z) = −U (X, −Y, Z)
t d le En
∀ X,Y, Z
r
ba
ge ro or
eS m
ci
3) Under said displacement field,
f
ra
C d P cs
the volume of the element in the
b
a
figure does not change, its an-
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.
s gin
3) The angle θ increases its value
in r (given in radians).
t d le En
Under these conditions, deter-
r
ba
mine:
ge ro or
eS m
ci
c) The strain tensor, the rotation
f
ra
C d P cs
vector and the displacement
b
a
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.)
s gin
such that points A, B and C move to positions A , B and C , respectively. Point
O does not move. Determine:
t d le En
a) The deformation gradient tensor in terms of p and q.
r
ba
ge ro or
b) The equation of the deformed external surface of the sphere. Indicate which
eS m
ci
type of surface it is and draw it.
f
ra
C d P cs
c) The material and spatial strain tensors. Obtain the value of p in terms of q
b
a
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,
the results of c) and d) coincide.
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
CH.3. COMPATIBILITY
EQUATIONS
Multimedia Course on Continuum Mechanics
Overview
Introduction Lecture 1
2
3.1 Compatibility Conditions
Ch.3. Compatibility Equations
3
Introduction
Given a displacement field, the corresponding strain field is
found:
1 ∂U i ∂U j ∂U k ∂U k
U ( X,t ) E= + + i, j ∈ {1, 2,3}
2 ∂X j ∂X i ∂X i ∂X j
ij
1 ∂ui ∂u j
u ( x,t ) ε ij = + i, j ∈ {1, 2,3}
2 ∂x j ∂xi
ε ( x,t ) u ( x, t )
4
Compatibility Conditions
Given an (arbitrary) symmetric second order tensor field, ε ( x,t ),
a displacement field, u ( x, t ), fulfilling ∇ s u(x, t ) = ε ( x, t ) cannot
always be obtained:
1 ∂ui ∂u j 6 PDEs OVERDETERMINED
ε ij = + i, j ∈ {1, 2,3} 3 unknowns SYSTEM
2 ∂x j ∂xi
Incompatible
strain field
6
3.2 Compatibility Equations of a
Potential Vector Field
Ch.3. Compatibility Equations
7
Preliminary example: Potential Vector Field
8
Potential Field
v ( x,t ) φ ( x,t ) such that ∇φ ( x, t ) =
v ( x, t )
In component form,
∂φ ( x, t ) ∂φ ( x, t )
vi ( x, t ) = vi ( x, t ) − =
0 i ∈ {1, 2,3} 1 3unknown
eqns.
∂xi ∂xi
OVERDETERMINED
Differentiating once these expressions with respect to x j : SYSTEM
∂vi ∂ φ ( x, t )
2
9
Schwartz Theorem
The Schwartz Theorem about symmetry of second partial
derivatives guarantees that, given a continuous function
Φ ( x1 , x2 ,..., xn ) with continuous derivatives, the following holds
true:
∂Φ ∂Φ 2 2
= ∀i, j
∂xi ∂x j ∂x j ∂xi
10
Compatibility Equations
Considering the Schwartz Theorem,
∂v x ∂ 2φ ∂v x ∂ 2φ ∂v x ∂ 2φ
= = =
∂x ∂x 2
∂y ∂x∂y ∂z ∂x∂z
∂v y ∂ 2φ ∂v y ∂ 2φ ∂v y ∂ 2φ
= = =
∂x ∂y∂x ∂y ∂y 2 ∂z ∂y∂z
∂v z ∂ 2φ ∂v z ∂ 2φ ∂v z ∂ 2φ
= = =
∂x ∂z∂x ∂y ∂z∂y ∂z ∂z 2
In this system of 9 equations, only 6 different 2nd derivatives of the
unknown φ ( x,t ) appear: ∂ 2 φ ∂ 2 φ ∂ 2 φ ∂ 2 φ ∂ 2 φ ∂ 2φ
, , , , and
∂x 2
∂y 2
∂z 2
∂x∂y ∂x∂z ∂y∂z
They can be eliminated and the following identities are obtained:
∂v x ∂v y ∂v x ∂v z ∂v y ∂v z
= = =
∂y ∂x ∂z ∂x ∂z ∂y
11
Compatibility Equations
A scalar function φ ( x,t ) which satisfies ∇φ ( x, t ) =
v ( x, t ) will exist if
the vector field v ( x,t ) verifies:
∂v y ∂v x def
− ==
0 Sz
∂x ∂y eˆ1 eˆ 2 eˆ 3
Sx
∂v x ∂v z def
∂ ∂ ∂
− ==
0 Sy where S ≡ S y ≡ ≡ ∇×v
∂z ∂x S ∂x ∂y ∂z
∂v z ∂v y def z v vy vz
− ==
0 Sx x
∂y ∂z
REMARK
A functional relation can be
INTEGRABILITY ∇ × v =0
(COMPATIBILITY) established between these
EQUATIONS ∂vi ∂v j three equations.
∂x − ∂x = 0 i, j ∈ {1, 2,3}
of a potential
vector field j i ∇ ⋅ (∇ × v ) =
0
12
3.3 Compatibility Conditions for
Infinitesimal Strains
Ch.3. Compatibility Equations
13
Infinitesimal strains case
The infinitesimal strain field can be written as:
∂u x 1 ∂u x ∂u y 1 ∂u x ∂u z
+ +
∂x 2 ∂y ∂x 2 ∂z ∂x
ε xx ε xy ε xz
∂u y
1 ∂u y ∂u z
ε xy ε yy ε yz =
ε = × +
ε ∂y 2 ∂z ∂y
ε ε
xz yz zz
∂u z
symmetrical ×
∂z
6 PDEs
3 unknowns
14
Infinitesimal strains case
The infinitesimal strain field can be written as:
∂u 1 ∂u x ∂u y
ε xx − x 0
= ε xy − = + 0
∂x 2 ∂y ∂x
∂u 1 ∂u ∂u
ε yy − y 0
= ε xz − =
x
+ z 0
∂y 2 ∂z ∂x
∂u 1 ∂u y ∂u z
ε zz − z 0
= ε yz − = + 0
∂z 2 ∂z ∂y
15
Compatibility Conditions
The compatibility conditions for the infinitesimal strain field are
obtained through double differentiation (single differentiation is
not enough).
∂u
∂ 2 ε xx − x
∂x
= 6
∂ x , ∂y , ∂z , ∂xy, ∂xz , ∂yz
2 2 2
equations
6x6=36
equations
1 ∂u y ∂u z
∂ ε yz −
2
+
2 ∂z ∂y
=
6
∂ x 2 , ∂y 2 , ∂z 2 , ∂xy, ∂xz , ∂yz equations
16
Compatibility Conditions
The compatibility conditions for the infinitesimal strain field are
obtained through: ∂ 2ε xx ∂ 3ux ∂ 2ε yz 1 ∂ 3u y ∂ 3u z
= = +
∂x 2 ∂x 3 ∂x 2 2 ∂z∂x 2 ∂y∂x 2 18 equations for
ε xy ,ε xz ,ε yz
∂ ε xx
2
∂ ux3
∂ ε yz 1 ∂ u y ∂ u z
2 3 3
= = +
∂y 2 ∂x∂y 2 ∂y 2 2 ∂z∂y 2 ∂y 3
∂ 2ε xx ∂ 3u x ∂ 2ε yz 1 ∂ 3u y ∂ 3u z
= ... = +
∂z 2 ∂x∂z 2 ∂z 2 2 ∂z 3 ∂y∂z 2
18 equations for
ε xx ,ε yy ,ε zz ∂ 2ε xx ∂ 3u x ∂ 2ε yz 1 ∂ 3u y ∂ 3u z
= = +
∂x∂y ∂x 2 ∂y ∂x∂y 2 ∂z∂x∂y ∂y 2 ∂x
∂ 2ε xx ∂ 3u x ∂ 2ε yz 1 ∂ 3u y ∂ 3u z
= = +
∂x∂z ∂x 2 ∂z ∂x∂z 2 ∂z 2 ∂x ∂y∂x∂z
∂ 2ε xx ∂ 3u x ∂ 2ε yz 1 ∂ 3u y ∂ 3u z
= = +
∂y∂z ∂x∂y∂z ∂y∂z 2 ∂z 2∂y ∂y 2∂z
17
Compatibility Conditions
All the third derivatives of u x , u y and u z appear in the
equations: 3
∂ ux
= 10 derivatives
∂x , ∂x y, ∂x z , ∂y , ∂y x, ∂y z , ∂z , ∂z x, ∂z y, ∂xyz
3 2 2 3 2 2 3 2 2
∂ 3u y
= 10 derivatives
∂x , ∂x y, ∂x z , ∂y , ∂y x, ∂y z , ∂z , ∂z x, ∂z y, ∂xyz
3 2 2 3 2 2 3 2 2
∂ 3u z
= 10 derivatives
∂x , ∂x y, ∂x z , ∂y , ∂y x, ∂y z , ∂z , ∂z x, ∂z y, ∂xyz
3 2 2 3 2 2 3 2 2
18
30
Compatibility Equations
∂ 3ui
Eliminating the 30 unknowns , , 6 equations (involving
∂x j ∂xk ∂xl
only strain derivatives) are obtained:
def ∂ 2ε yy ∂ 2ε ∂ 2
ε yz COMPATIBILITY
S xx = + zz
−2 = 0 EQUATIONS
∂z 2
∂y 2
∂y∂z for the infinitesimal
def
∂ ε zz ∂ ε xx
2 2
∂ 2ε xz strain tensor
S yy = ∂x 2 + ∂z 2 − 2 ∂x∂z = 0
def ∂ 2ε ∂ ε ∂ ε xy S = ∇ × (ε × ∇ ) = 0
2 2
S zz = xx
+
yy
−2 =0
∂ y 2
∂ x 2
∂ x ∂ y
def ∂ 2ε zz ∂ ∂ε yz ∂ε xz ∂ε xy
S xy = − +
∂x
+ − = 0
∂x ∂y ∂ z ∂ y ∂z
def ∂ ε yy ∂ ∂ε yz ∂ε xz ∂ε xy
2
S xz = − +
− + = 0
∂ ∂
x z ∂ ∂
y x ∂ y ∂ z
def ∂ 2ε xx ∂ ∂ε yz ∂ε xz ∂ε xy
S yz = − + −
+ + = 0
∂y∂z ∂x ∂x ∂y ∂z
19
Compatibility Equations
The six equations are not functionally independent. They satisfy
the equation,
∇ ⋅ S = ∇ ⋅ (∇ × (ε × ∇ )) = 0
In indicial notation:
∂S xx ∂S xy ∂S xz
+ + =
0
∂x ∂y ∂z
∂S xy ∂S yy ∂S yz
+ + =
0
∂x ∂y ∂z
∂S xz ∂S yz ∂S zz
+ + =
0
∂x ∂y ∂z
20
Compatibility Equations
The compatibility equations can be expressed in terms of the
permutation operator, eijk .
Or, alternatively:
ε ij ,kl + ε kl ,ij − ε ik , jl −=
ε jl ,ik 0 i, j , k , l ∈ {1, 2,3}
REMARK
Any linear strain tensor (1st order polynomial) with respect to
the spatial variables will be compatible and, thus, integrable.
21
3.4 Integration of the Infinitesimal
Strain Tensor
Ch.3. Compatibility Equations
22
Preliminary Equations
Rotation tensor Ω ( x,t ) :
1
=Ω skew(u ⊗ = ∇) (u ⊗ ∇ − ∇ ⊗ u)
2
= 1 ∂ui ∂u j
Ω − i, j ∈ {1, 2,3}
ij 2 ∂x j ∂xi
θ1 −Ω 23 −Ω yz 0 −θ3 θ2
θ = ∇ × u = θ 2 = −Ω =
[Ω(θ )] = θ3 −θ1
1
31 −Ω
zx 0
2
θ3 −Ω12 −θ 2 θ1 0
−Ω xy
23
Preliminary Equations
Differentiating Ω ( x,t ) with respect to xk :
= Ωij
Introducing the definition of θ ( x,t ) , the components of J ( x,t )
are rewritten:
= j 1= j 2= j 3
∂u x ∂u x ∂u x
i = 1: =ε xx =ε xy − θ3 =ε xz + θ 2
∂x ∂y ∂z
∂u y ∂u y ∂u y
i = 2: =ε xy + θ3 =ε yy =ε yz − θ1
∂x ∂y ∂z
∂u z ∂u z ∂u z
i = 3: =ε xz − θ 2 =ε yz + θ1 =ε zz
∂x ∂y ∂z
26
Integration of the Strain Field
The integration of the strain field ε ( x,t ) is performed in two
steps:
1. Integration of derivative of θ ( x,t ) using the1st order PDE system
derived for ∇θ1 , ∇θ 2 and ∇θ3 . The solution will be of the type:
θ=i θi ( x, y, z, t ) + ci ( t ) i ∈ {1, 2,3}
The integration constants ci ( t ) can be obtained knowing the value of
the rotation vector in some points of the medium (boundary conditions).
2. Known ε ( x,t ) and θ ( x,t ) , u is integrated using the 1st order PDE
REMARK system derived for u ⊗ ∇ . The solution will be:
If the compati- ui = ui ( x, y, z , t ) + ci′ ( t ) i ∈ {1, 2,3}
bility equations
The integration constants ci′ ( t ) can be obtained knowing the value of
are satisfied, the displacements in some point of space (boundary conditions)
these equations
will be integra-
ble. 27
Integration of the Strain Field
The integration constants that appear imply that an integrable
strain tensor ε ( x,t ) will determine the movement in any instant not
of
not
time except for a rotation c(t ) = θˆ (t ) and a translation c′(t ) = uˆ (t ) :
( x, t ) θ ( x, t ) + θ ( t )
θ= ˆ
ε ( x, t )
( x, t ) u ( x, t ) + uˆ ( t )
u=
A displacement field can be constructed from this uniform rotation
and translation: u∗ (= ˆ (θˆ(t )) ⋅ x + uˆ (t )
x, t ) Ω
u∗ ⊗ ∇ = Ω
ˆ
S * ∗ ∗
∇ ( u =
) 1
2 (u ⊗ ∇ + (u ⊗ ∇ )T
)= 1
2 (Ω
ˆ +Ω
ˆ T )= 0
28
3.5 Integration of the Deformation
Rate Tensor
Ch.3. Compatibility Equations
29
Compatibility Equations in a
Deformation Rate Field
There is a correspondence between
u v
ε (u) d( v )
1 ∂ui ∂u j 1 ∂vi ∂v j
=ε ij + =dij +
2 ∂x j ∂xi 2 ∂x j ∂xi
1 ∂ui ∂u j 1 ∂vi ∂v j
=
Ωij − =
w ij −
2 ∂x j ∂xi 2 ∂x j ∂xi
1 1
=
θ ∇ ×u =
ω ∇×v
2 2
30
Example
Obtain the velocity field corresponding to the deformation rate tensor:
0 tety 0
d ( x, t ) = tety 0 0
0 0 tetz
31
0 tety 0
Example - Solution d ( x, t ) = tety
0
0
0
0
tetz
u
Consider the correspondence: v
ε(u) d( v )
1 1
=
θ ∇ ×u =
ω ∇×v
2 2
Take the expressions derived for ∇θ1 , ∇θ 2 and ∇θ3 substitute θ ( x,t ) with ω ( x,t )
and ε ( x,t ) with d ( x,t ) :
∂ω1 ∂d xz ∂d xy
= − =0−0
∂x ∂y ∂z
∂ω1 ∂d yz ∂d yy
∇ω1 = − = 0−0 ω1 ( t ) =
C1 ( t )
∂y ∂y ∂z
∂ω ∂d ∂d zy
1 =zz − =0−0
∂z
∂y ∂z
32
0 tety 0
Example - Solution d ( x, t ) = tety
0
0
0
0
tetz
∂ω2 ∂d xx ∂d xz
∂x = ∂z
−
∂x
=0−0
∂ω2 ∂d xy ∂d yz
∇ω2 = − = 0−0 ω2 ( t ) =
C2 ( t )
∂y ∂z ∂x
∂ω2 ∂d xz ∂d zz
= − =0−0
∂z ∂z ∂x
∂ω3 ∂d xy ∂d xx
= − =0−0
∂x ∂x ∂y
∂ω3 ∂d yy ∂d xy
∇ω3 = − = 0 − t 2 ety ω3 ( y, t ) =
∫ −t 2 ty
e dy =
−te ty
+ C3 ( t )
∂y ∂x ∂y
∂ω ∂d yz ∂d
3 = − xz = 0−0
∂z ∂x ∂y
33
ω1 =C1 ( t )
∂v x
= d xy − ω= tety − ( −tety=
) 2tety v x ( y=
,t) ∫ 2te =
ty
dy 2ety + C1′ ( t )
∂y
3
∂v x
= d xz + ω2 = 0 + 0
∂z
35
0 0 tety 0
{ω ( x, t )} = 0 ; d ( x, t ) = tety
Example - Solution −tety
0
0
0
0
tetz
∂v y
= d= 0 v y ( t ) = C2′ ( t )
∂y
yy
∂v y
= d yz − ω1 = 0 − 0
∂z
∂v z
= d xz − ω2 = 0 − 0
∂x
∂v z
= d yz + ω1 = 0 + 0 v z ( z , t=
) ∫ te tz
=
dz e tz
+ C3′ ( t )
∂y
∂v z
= d= tetz
∂z
zz
36
v x 2ety + C1′ ( t )
=
So,
= et 2ety + C1′ ( t )
v x 2=
x=(1,1,1) C1′ ( t ) = 0
v=y e=
t
C2′ ( t ) C2′ ( t ) = et
tz C3′ ( t ) = 0
v=
z e=t
e + C3
′ ( t )
x=(1,1,1)
Therefore, for any point, 2ety
t
{ ( )} e
v x , t =
etz
37
Chapter 3
Compatibility Equations
rs
ee
s gin
3.1 Introduction
t d le En
Given a sufficiently regular displacement field U (X,t), it is always possible to
r
ba
ge ro or
eS m
find the corresponding strain field (for example, the Green-Lagrange strain field)
ci
f
by differentiating this strain field with respect to its coordinates (in this case, the
ra
C d P cs
material ones)1 ,
b
a
i
an an n
2 ∂ X j ∂ Xi ∂ Xi ∂ X j 2 (3.1)
le
i, j ∈ {1, 2, 3}.
liv or ec
M
.A
In the infinitesimal strain case, given a displacement field u (x,t), the strain
field
m
d
1 ∂ ui ∂ u j not 1
uu
(3.2)
2 ∂ x j ∂ xi
X Th
2
er
tin
is obtained.
on
.O
The question can be formulated in reverse, that is, given a strain field ε (x,t),
is it possible to find a displacement field u (x,t) such that ε (x,t) is its infinites-
C
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
not
1 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).
s gin
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,
r
ba
ge ro or
therefore, for there to exist a displacement field from which it comes.
eS m
ci
f
ra
C d P cs
b
a
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
s gin
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
r
ba
∂ vx ∂ 2 φ ∂ vx ∂ 2φ ∂ vx ∂ 2φ
ge ro or
eS m
= 2 , = , = ,
ci
∂x ∂x ∂y ∂ x∂ y ∂z ∂ x∂ z
f
ra
C d P cs
∂ vy ∂ 2φ ∂ vy ∂ 2 φ ∂ vy ∂ 2φ
b
a
= , = 2 , = , (3.5)
i
∂x ∂ y∂ x ∂y ∂y ∂z ∂ y∂ z
an an n
y ha
∂ vz ∂ 2φ ∂ vz ∂ 2φ ∂ vz ∂ 2 φ
= , = , = 2 ,
le
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 .
s gin
t d le En
r
3.3 Compatibility Conditions for Infinitesimal Strains
ba
ge ro or
eS m
ci
f
Consider the infinitesimal strain field ε (x,t) with components
ra
C d P cs
b
a
1 ∂ ui ∂ u j not 1
i
an an n
2
le
liv or ec
.A
∂ u x 1 ∂ u x ∂ uy 1 ∂ ux ∂ uz
⎡ ⎤ ⎢ ∂x 2 ∂y + ∂x + ⎥
m
⎢ ⎥
⎢ ⎥ ⎢ 1 ∂ uy ∂ u z ⎥
uu
∂ uy
e
∂y 2 ∂z
er
tin
.O
∂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-
nents of the displacement vector u (x,t) ≡ [ux , uy , uz ]T . In general, this problem
not
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,
∂ ux
∂ εxx −
2
∂x
= 6 equations
∂ x , ∂ y , ∂ z , ∂ xy, ∂ xz, ∂ yz
2 2 2
.. .. (3.12)
. .
rs
1 ∂ uy ∂ u z
∂ 2 εyz − +
ee
2 ∂z ∂y
= 6 equations ,
s gin
∂ x2 , ∂ y2 , ∂ z2 , ∂ xy, ∂ xz, ∂ yz
t d le En
providing a total of 36 equations,
r
ba
∂ 2 εxx ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
ge ro or
eS m
= = +
ci
∂ x2 ∂ x3 ∂ x2 2 ∂ z∂ x2 ∂ y∂ x2
f
ra
C d P cs
∂ εxx
2 ∂ 3 ux ∂ 2 εyz 1 ∂ 3 uy ∂ 3 uz
b
a
= = +
i
∂ y2 ∂ x∂ y2 ∂ y2 2 ∂ z∂ y2 ∂ y3
an an n
y ha
∂ εxx
2 ∂ 3 ux ∂ εyz 1 ∂ uy
2 3
∂ 3 uz
= ··· = +
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
∂ x∂ z ∂ x ∂ z ∂ x∂ z 2 ∂ z2 ∂ x ∂ y∂ x∂ z
e
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
(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
∂ 2 εi j /(∂ xk ∂ xl ). After the corresponding algebraic operations, the resulting equa-
ee
tions are
s gin
⎧
⎪
⎪ de f ∂ εyy
2
∂ 2 εzz ∂ 2 εyz
t d le En
⎪
⎪ = + − =0
⎪
⎪ S xx 2
⎪
⎪ ∂ z2 ∂ y2 ∂ y∂ z
r
⎪
⎪
ba
ge ro or
⎪
eS m
⎪
⎪ de f ∂ εzz
2 ∂ 2 εxx ∂ 2 εxz
⎪
ci
⎪ S = + − 2 =0
⎪ yy
f
ra
⎪
⎪ ∂ x2 ∂ z2 ∂ x∂ z
C d P cs
⎪
⎪
b
a
⎪
⎪ ∂ 2 εyy ∂ 2 εxy
de f ∂ εxx
2
⎪
i
⎪
an an n
⎨ Szz = + − 2 =0
∂ y2 ∂ x2 ∂ x∂ y
y ha
Compatibility
equations ⎪ (3.16)
le
⎪
⎪ ∂ 2 εzz ∂ ∂ εyz ∂ εxz ∂ εxy
liv or ec
⎪
⎪ S
de f
= − + + − =0
⎪
⎪ xy
∂ x∂ y ∂ z ∂ x ∂y ∂z
M
⎪
.A
⎪
⎪
⎪
⎪
⎪
⎪ de f ∂ εyy
2
∂ ∂ εyz ∂ εxz ∂ εxy
m
⎪
⎪ Sxz = − + − + =0
d
⎪
⎪ ∂ x∂ z ∂ y ∂ x ∂y ∂z
uu
⎪
⎪
e
⎪
⎪
X Th
⎪ de f ∂ εxx
2 ∂ ∂ εyz ∂ εxz ∂ εxy
er
⎪
tin
⎩ Syz = − + − + + =0
∂ y∂ z ∂ x ∂x ∂y ∂z
on
.O
which constitute the compatibility equations for the infinitesimal strain tensor ε .
C
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
⎪
⎪ ∂x ∂y ∂z
⎪
⎪
ee
⎪
⎪
⎪
⎪ ∂S ∂S ∂S
⎩ xz + yz + zz = 0
s gin
∂x ∂y ∂z
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Remark 3.5. The three-index operator denominated permutation op-
C d P cs
b
a
erator is given by
i
an an n
⎧
y ha
⎪
⎪ 0 → if an index is repeated,
⎪
⎪ i = j or i = k or j = k
le
⎪
⎪
liv or ec
⎨
1 → positive (clockwise) direction of the indexes,
ei jk =
M
.A
⎪
⎪ i, j, k ∈ {123, 231, 312}
⎪
⎪
⎪ −1 → negative (counterclockwise) direction of the indexes,
m
⎪
⎩
d
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.
ee
s gin
t d le En
r
3.4 Integration of the Infinitesimal Strain Field
ba
ge ro or
eS m
ci
3.4.1 Preliminary Equations
f
ra
C d P cs
b
Consider the rotation tensor Ω (x,t) for the infinitesimal strain case (see Chap-
a
i
an an n
⎧
⎪ 1
le
⎨ Ω = (u ⊗ ∇ − ∇ ⊗ u) ,
liv or ec
2 (3.20)
1 ∂ ui ∂ u j
M
.A
⎪
⎩ 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
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
θ1 −Ω23 −Ωyz
on
.O
1 1 not ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
θ = rot u = ∇ × u ≡ ⎣ θ2 ⎦ = ⎣ −Ω31 ⎦ = ⎣ −Ωzx ⎦ . (3.21)
C
2 2
θ3 −Ω12 −Ωxy
©
⎡ ⎤
0 Ω12 −Ω31
not ⎢ ⎥
3 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
∂ Ωi j 1 ∂ ∂ ui ∂ u j 1 ∂ 2 uk 1 ∂ 2 uk
= − + − =
∂ 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
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-
s gin
lows. ⎧
⎪
⎪ ∂ θ1 ∂ Ωyz ∂ εxz ∂ εxy
t d le En
⎪
⎪
⎪
⎪ =− = −
⎪
⎪ ∂x ∂x ∂y ∂z
r
⎪
⎨
ba
ge ro or
eS m
∇θ1 ∂ θ1 ∂ Ωyz ∂ εyz ∂ εyy
ci
=− = − (3.24)
⎪
⎪ f
ra
⎪ ∂y ∂y ∂y ∂z
C d P cs
⎪
⎪
b
a
⎪
⎪ ∂ θ1 ∂ Ωyz ∂ εzz ∂ εzy
⎪
i
⎪
an an n
⎩ =− = −
∂z ∂z ∂y ∂z
y ha
le
⎧
liv or ec
⎪
⎪ ∂ θ2 ∂ Ωzx ∂ εxx ∂ εxz
M
⎪
.A
⎪
⎪ =− = −
⎪
⎪ ∂x ∂x ∂z ∂x
⎪
⎨
m
=− = − (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
⎪
⎪
⎪
⎪ ∂ Ωxy ∂ εyz ∂ εxz
⎪ ∂ θ3
⎪
⎩ =− = −
∂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
⎩ 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 .
ee
s gin
j=1 j=2 j=3
∂ ux ∂ ux ∂ ux
i=1: = εxx = εxy − θ3 = εxz + θ2
t d le En
∂x ∂y ∂z
∂ uy ∂ uy ∂ uy (3.28)
r
ba
i=2: = εxy + θ3 = εyy = εyz − θ1
ge ro or
eS m
∂x ∂y ∂z
ci
∂ uz ∂ uz f ∂ uz
ra
C d P cs
i=3: = εxz − θ2 = εyz + θ1 = εzz
b
a
∂x ∂y ∂z
i
an an n
y ha
le
3.4.2 Integration of the Strain Field
liv or ec
M
.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 ci (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
ee
s gin
Remark 3.8. The presence of the integration constants in (3.29) and
t d le En
(3.30) shows that an integrable strain tensor, ε (x,t), determines the
r
not
motion of each instant of time except for a rotation c (t) = θ̂θ (t) and
ba
ge ro or
eS m
not
a translation c (t) = û (t).
ci
f
ra
θ (x,t) = θ̃θ (x,t) + θ̂θ (t)
C d P cs
ε (x,t)
b
a
u (x,t) = ũ (x,t) + û (t)
i
an an n
y ha
From these uniform rotation θ̂θ (t) and translation û (t) the displace-
le
liv or ec
ment field
M
u∗ (x,t) = Ω̂ u∗ ⊗ ∇ = Ω̂
.A
Ω (t) x + û (t) =⇒ Ω
m
er
tin
1 1
ε ∗ (x,t) = ∇s u∗ = (u∗ ⊗ ∇ + ∇ ⊗ u∗ ) = Ω̂ Ω +
ΩT = 0 ,
Ω̂
2 2
on
.O
Ω
−Ω̂
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.
ee
s gin
Solution
t d le En
Infinitesimal rotation vector
r
Posing the systems of equations defined in (3.24) through (3.26) results in
ba
ge ro or
eS m
ci
∂ θ1 ∂ θ1 ∂ θ1
f
ra
=0 ; =0 ; =0 ⇒ θ1 = C1 (t) ,
C d P cs
∂x ∂y ∂z
b
a
i
∂ θ2 ∂ θ2 ∂ θ2
an an n
3 3
= −3xz ; =0 ; = − x2 ⇒ θ2 = − x2 z +C2 (t) ,
y ha
∂x ∂y ∂z 2 2
le
∂ θ3 ∂ θ3 3 ∂ θ3
liv or ec
3
=0 ; = ; =0 ⇒ θ3 = y +C3 (t) .
∂x ∂y ∂z
M
2 2
.A
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 +C1 (t) ,
∂x ∂y ∂z
∂ u2 ∂ u2 ∂ u2
=y ; =x ; =0 ⇒ u2 = xy +C2 (t) ,
∂x ∂y ∂z
∂ u3 ∂ u3 ∂ u3
= 3x2 z ; =0 ; = x3 ⇒ u3 = x3 z +C3 (t) .
rs
∂x ∂y ∂z
ee
and imposing that u (x,t)|x=(0,0,0)T ≡ [3t, 0, 0]T yields
not
s gin
not T
t d le En
C1 (t) = 3t ; C2 (t) = C3 (t) = 0 =⇒ u (x) ≡ 4x2 − y2 + 3t , xy , x3 z .
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
3.5 Compatibility Equations and Integration of the Strain
i
an an n
Rate Field
y ha
le
Given the definitions of the infinitesimal strain tensor ε , the infinitesimal rotation
liv or ec
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
not ⎢ ⎥
rs
d (x,t) ≡ ⎢ ⎣ 0 0 te y +1⎥ .
⎦
ee
0 te + 1
y 0
s gin
For x = 0, ω 0 ≡ [t − 1, 0, 0]T and v0 ≡ [t, 0, t]T f or ∀t is satisfied.
not not
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Solution
C d P cs
b
a
The problem is solved by integrating the corresponding differential equations,
i
an an n
⎫ ⎧
le
liv or ec
u⎪⎬ ⎪
⎨v
M
.A
ε ⇐⇒ d
⎪
⎭ ⎪
⎩
θ ω
m
d
uu
e
er
tin
∂ ω1 ∂ ω1 ∂ ω1
on
∂x ∂y ∂z
C
∂ 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 C3 = 0
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
∂ v1 ∂ v1 ∂ v1
= tetx ; =0 ; =0 ⇒ v1 = C1 (t) + etx ,
∂x ∂y ∂z
rs
∂ v2 ∂ v2 ∂ v2
=0 ; =0 ; =2 ⇒ v2 = C2 (t) + 2z ,
ee
∂x ∂y ∂z
s gin
∂ v3 ∂ v3 ∂ v3
=0 ; = 2tey ; =0 ⇒ v3 = C3 (t) + 2tey .
∂x ∂y ∂z
t d le En
The boundary conditions are imposed for x = 0,
r
ba
ge ro or
⎡ ⎤ ⎡ ⎤ ⎧
eS m
ci
t 1 +C1 ⎨ C1 = t − 1
⎪
f
ra
not ⎢ ⎥ ⎢ ⎥
C d P cs
v0 ≡ ⎣ 0 ⎦ = ⎣ C2 ⎦ =⇒ C2 = 0 ,
b
a
⎪
⎩
i
2t +C3 C3 = −t
an an n
t
y ha
le
and the spatial description of the velocity field is
liv or ec
⎡ ⎤
M
.A
etx + t − 1
not ⎢ ⎥
m
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
s gin
3.2 – Determine the spatial description of the velocity field that corresponds to
t d le En
the strain rate tensor
⎡ ⎤
r
ba
tetz
ge ro or
0 0
eS m
not ⎢ ⎥
ci
d (x,t) ≡ ⎢ ty
f ⎥.
ra
⎣ 0 te 0 ⎦
C d P cs
b
a
tetz 0 0
i
an an n
y ha
for z = 0 : vx = vz = 0 , ∀t , x, y
M
.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
CH.4. STRESS
Multimedia Course on Continuum Mechanics
Overview
Forces Acting on a Continuum Body Lecture 1
Cauchy’s Postulates Lecture 2
Stress Tensor Lecture 3 Lecture 4
Stress Tensor Components
Scientific Notation
Lecture 5
Engineering Notation
Sign Criterion
Properties of the Cauchy Stress Tensor
Cauchy’s Equation of Motion Lecture 6
Principal Stresses and Principal Stress Directions
Mean Stress and Mean Pressure
Lecture 7
Spherical and Deviatoric Parts of a Stress Tensor
Stress Invariants
2
Overview (cont’d)
Stress Tensor in Different Coordinate Systems
Cylindrical Coordinate System Lecture 8
Spherical Coordinate System
Mohr’s Circle Lecture 9
Mohr’s Circle for a 3D State of Stress
Lecture 10
Determination of the Mohr’s Circle
Mohr’s Circle for a 2D State of Stress
2D State of Stress Lecture 11
Stresses in Oblique Plane
Direct Problem
Lecture 12
Inverse Problem
Mohr´s Circle for a 2D State of Stress Lecture 13
3
Overview (cont’d)
Mohr’s Circle a 2D State of Stress (cont’d)
Construction of Mohr’s Circle
Mohr´s Circle Properties Lecture 14
The Pole or the Origin of Planes Lecture 15
Sign Convention in Soil Mechanics Lecture 16
Particular Cases of Mohr’s Circle Lecture 17
4
4.1. Forces on a Continuum Body
Ch.4. Stress
5
Forces Acting on a Continuum Body
Forces acting on a continuum body:
Body forces.
Act on the elements of volume or mass inside the body.
“Action-at-a-distance” force.
E.g.: gravity, electrostatic forces, magnetic forces
fV = ∫ ρ b ( x, t ) dV body force per unit
V
mass
Surface forces. (specific body forces)
6
4.2. Cauchy’s Postulates
Ch.4. Stress
7
Cauchy’s Postulates
1. Cauchy’s 1st postulate. REMARK
The traction vector t remains unchanged The traction vector (generalized to
for all surfaces passing through the point P internal points) is not influenced by
and having the same normal vector n at P . the curvature of the internal surfaces.
t = t ( P, n )
8
4.3. Stress Tensor
Ch.4. Stress
9
Stress Tensor
The areas of the faces of the tetrahedron
are:
S1 = n1S
n ≡ {n 1 , n 2 , n 3 }
T
S 2 = n2 S with
S3 = n3 S
*
1 2
The surface normal vectors of the planes perpendicular to the axes are
n1 = −eˆ1 ; n2 = −eˆ 2 ; n3 = −eˆ 3 REMARK
Following Cauchy’s fundamental lemma: The asterisk indicates an
t ( x, −eˆ i ) =
not
−t ( x, eˆ i ) =
− t (i ) ( x ) i ∈ {1, 2,3} mean value over the area.
10
Mean Value Theorem
Let f : [ a, b ] → R be a continuous function on the closed interval
[a, b] , and differentiable on the open interval ( a, b ) , where a < b .
Then, there exists some x* in ( a, b ) such that:
f ( x* )
1
f ( x ) dΩ
Ω Ω∫
=
11
Stress Tensor
From equilibrium of forces, i.e. Newton’s 2nd law of motion:
R=∑ fi =∑ mi ai ∫ ρ b dV + ∫ t dS =∫ a ρ ∫ ρ a dV
dV =
i i V ∂V V dm V
resultant
body forces
(1) ( 2) ( 3)
∫
V
ρ b dV + ∫ t
S
dS + ∫ − t
S1
dS + ∫ − t
S2
dS + ∫ dS =∫ ρ a dV
− t
S3 V
resultant
Considering the mean value theorem, surface forces
1
Si ni S i ∈ {1, 2,3} and V = Sh ,
Introducing=
3
1 1
( ρ b)* h S + t*S − t (1)*n1S − t ( 2)*n2S − t (3)*n3S = ( ρ a)* hS
3 3
12
Stress Tensor
If the tetrahedron shrinks to point O,
x*Si → xO (
lim t (i )* x*Si , eˆ i =
h →0 )
t (i ) ( O,eˆ i ) i ∈ {1, 2,3}
1 1
lim= ( ρ b )*
h =
h →0
lim ( ρ a )*
h 0
h →0 3
3
= t (1) = t ( 2) = t ( 3)
1 (1)* ( 2 )* ( 3)* 1
( ρ b) h + t − t n1 − t n2 − t n3 =( ρ a)* h
* *
t ( O, n ) − t (i ) ni =
0
3 3
= t ( O, n )
13
Stress Tensor
Considering the traction vector’s Cartesian components :
=t (i ) ( P ) t =
(i )
ˆ σ ij eˆ j
j ( P) e j t ( P=
, n ) t (i ) ni ⇒
i, j ∈ {1, 2,3}
σ ij ( P ) = t j ( P )
(i )
t j ( P=
,n) t=
(i )
j ni niσ ij
σ ij
Cauchy’s Stress Tensor t ( P, n )= n ⋅ σ ( P)
σ σ ij eˆ i ⊗ eˆ j
=
P
t (1) t ( 2) t ( 3)
14
Stress Tensor
REMARK 1
The expression t ( P, n )= n ⋅ σ ( P ) is consistent with Cauchy’s postulates:
t ( P, n )= n ⋅ σ
t ( P, n ) =
−t ( P, −n )
t ( P, −n ) =−n ⋅ σ
REMARK 2
The Cauchy stress tensor is constructed from the traction vectors on three
coordinate planes passing through point P.
σ 11 σ 12 σ 13
σ ≡ σ 21 σ 22 σ 23
σ 31 σ 32 σ 33
Yet, this tensor contains information on the traction vectors acting on any plane
(identified by its normal n) which passes through point P.
15
4.4.Stress Tensor Components
Ch.4. Stress
16
Scientific Notation
Cauchy’s stress tensor in scientific notation
σ 11 σ 12 σ 13
σ ≡ σ 21 σ 22 σ 23
σ 31 σ 32 σ 33
17
Engineering Notation
Cauchy’s stress tensor in engineering notation
σ x τ xy τ xz
σ ≡ τ yx σ y τ yz
τ zx τ zy σ z
Where:
σ a is the normal stress acting on plane a.
τ ab is the tangential (shear) stress acting on the plane perpendicular to
the a-axis in the direction of the b-axis.
18
Tension and compression
The stress vector acting on point P of an
arbitrary plane may be resolved into:
a vector normal to the plane (σ n = σ n)
an in-plane (shear) component which acts on the plane.
(τ n ; τ n = τ )
The sense of σ n with respect to n defines the normal stress character:
σ= σ n ⋅n >0 tensile stress (tension)
<0 compressive stress (compression)
The sign criterion for the stress components is:
positive (+) tensile stress
σ ij or σ a negative (−) compressive stress
positive (+) positive direction of the b-axis
τ ab negative (−) negative direction of the b-axis
19
4.5.Properties of the Cauchy Stress
Tensor
Ch.4. Stress
20
Cauchy’s Equation of Motion
Consider an arbitrary material volume,
Cauchy’s equation of motion is:
∇ ⋅ σ + ρ= b ρ a ∀x ∈ V
∂σ ij
∂x + ρ b=
j ρ a j j ∈ {1, 2,3}
i
In engineering notation: b ( x, t ) x ∈ V
∂σ x ∂τ yx ∂τ zx t* ( x, t ) x ∈ ∂V
+ + + ρ bx =ρ ax
∂x ∂y ∂z
∂τ xy ∂σ y ∂τ zy REMARK
+ + + ρ by =ρ ay
∂x ∂y ∂z Cauchy’s equation of motion is derived
∂τ xz ∂τ yz ∂σ z from the principle of balance of linear
+ + + ρ bz =ρ az momentum.
∂x ∂y ∂z
21
Equilibrium Equations
For a body in equilibrium a = 0 ,
Cauchy’s equation of motion becomes
∇ ⋅ σ + ρ b = 0 ∀x ∈ V
∂σ internal equilibrium
ij + ρ b = 0 j ∈ {1, 2,3}
∂xi j equation
22
Cauchy’s Equation of Motion
Taking into account the symmetry of the
Cauchy Stress Tensor,
Cauchy’s equation of motion
∇ ⋅ σ + ρ b = σ ⋅ ∇ + ρ b = ρ a ∀x ∈ V
∂σ ij ∂σ ji
∂x + ρ b = + ρbj = ρ a j j ∈ {1, 2,3}
∂
j
i xi
b ( x, t ) x ∈ V
Boundary conditions
t* ( x, t ) x ∈ ∂V
n ⋅ σ= σ ⋅ n= t (x, t ) ∀x ∈ ∂V
*
ni σij = σ ji ni = t j ( x, t ) ∀x ∈ ∂V i, j ∈ {1, 2,3}
*
23
Principal Stresses and Principal
Stress Directions
Regardless of the state of stress, it is always possible to choose a
special set of axes (principal axes of stress or principal stress
directions) so that the shear stress components vanish when the
stress components are referred to this system.
The three planes perpendicular to the principal axes are the
principal planes.
The normal stress components in the principal planes are the
principal stresses. σ 33
x3 σ 31 σ 32 x3
σ 13 x3′
σ 1 0 0
σ 23
σ 11 σ 12 σ 21 σ x1′ σ3
[σ ] = 0 σ 2 0 22
σ1
0 0 σ 3
σ2
x1 x1
x2 x2
x2′
24
Principal Stresses and Principal
Stress Directions
The Cauchy stress tensor is a symmetric 2nd order tensor so it will
diagonalize in an orthonormal basis and its eigenvalues are
real numbers.
For the eigenvalue λ and its corresponding eigenvector v :
σ⋅v =λv [ σ − λ 1] ⋅ v =0 INVARIANTS
not characteristic
det [ σ − λ 1] = σ − λ 1 =0 λ 3 − I 1 ( σ )λ 2 − I 2 ( σ )λ − I 3 ( σ ) =
0 equation
σ 33
λ1 ≡ σ 1 x3 σ 31 σ 32 x3
x3′
λ2 ≡ σ 2 σ 13 σ 23
λ3 ≡ σ 3 σ 11 σ 12 σ 21 σ x1′ σ3
σ1
22
REMARK σ2
x1 x1
The invariants associated with a x2 x2
tensor are values which do not change x2′
25
with the coordinate system being used.
Mean Stress and Mean Pressure
Given the Cauchy stress tensor σ and its principal stresses, the
following is defined:
σ 11 σ 12 σ 13
Mean stress
σ ≡ σ 21 σ 22 σ 23
1 1 1
σ m= Tr (σ )= σ ii = (σ 1 + σ 2 + σ 3 ) σ 31 σ 32 σ 33
3 3 3
Mean pressure
REMARK
1 In a hydrostatic state of stress, the
− (σ 1 + σ 2 + σ 3 )
−σ m =
p=
3 stress tensor is isotropic and, thus,
its components are the same in
A spherical or hydrostatic any Cartesian coordinate system.
state of stress: σ 0 0 As a consequence, any direction
is a principal direction and the
σ= σ= σ3 σ ≡ 0 σ 0 =
σ1
1 2
stress state (traction vector) is the
0 0 σ same in any plane.
26
Spherical and Deviatoric Parts of a
Stress Tensor
The Cauchy stress tensor σ can be split into:
σ = σ sph + σ ′
The spherical stress tensor:
Also named mean hydrostatic stress tensor or volumetric stress tensor or
mean normal stress tensor.
Is an isotropic tensor and defines a hydrostatic state of stress.
Tends to change the volume of the stressed body
1 1
σ=
sph : σ =
m 1 Tr ( σ=) 1 σ ii 1 REMARK
3 3 The principal directions of a stress tensor
The stress deviator tensor: and its deviator stress component coincide.
Is an indicator of how far from a hydrostatic state of stress the state is.
Tends to distort the volume of the stressed body
σ ′ dev
= = σ σ − σ m1
27
Stress Invariants
Principal stresses are invariants of the stress state:
invariant w.r.t. rotation of the coordinate axes to which the stresses are
referred.
The principal stresses are combined to form the stress invariants I :
I1 = Tr ( σ ) = σ ii = σ 1 + σ 2 + σ 3 REMARK
I 2 =( σ : σ − I12 ) =
1 The I invariants are obtained
− (σ 1σ 2 + σ 1σ 3 + σ 2σ 3 )
2 from the characteristic equation
I 3 = det ( σ ) of the eigenvalue problem.
These invariants are combined, in turn, to obtain the invariants J :
J=
1 I=
1 σ ii REMARK
The J invariants can be
J2 =
2
(
1 2
I1 + 2 I 2 ) = σ ijσ ji = ( σ : σ )
1
2
1
2 expressed in the unified form:
J=
3
3
(
1 3
I1 + 3 I I
1 2 + 3 I =
3 ) 1
3
Tr ( σ ⋅ σ ⋅ σ
= )
1=
3
σ σ σ
ij jk ki
J i
1
i
Tr( )
σ i
i ∈ {1, 2,3}
28
Stress Invariants of the Stress
Deviator Tensor
The stress invariants of the stress deviator tensor:
= I1′ Tr
= ( σ′ ) 0
=I 2′
1
2
( )
σ′ : σ′ − I12 = σ 12′ σ 12′ + σ 13′ σ 13′ + σ 23
′ σ 23′
J 2′ =
1 2
2
( )
I1′ + 2 I 2′ =
1
I 2′ = ( σ′ : σ′ )
2
J 3′ =
3
( )
I1′ + 3I1′I 2′ + 3I 3′ = I 3′ = Tr ( σ′ ⋅ σ′ ⋅ σ′ ) = (σ ij′σ ′jkσ ki′ )
1 3 1
3
1
3
29
4.6. Stress Tensor in Different
Coordinate Systems
Ch.4. Stress
30
Stress Tensor in a Cylindrical
Coordinate System
The cylindrical coordinate system is defined by:
dV = r dθ dr dz
x = r cos θ
x( r , θ , z ) ≡ y = r sin θ
z = z
The components of the stress tensor are then:
σ x´ τ x´ y ´ τ x´ z ´ σ r τ rθ τ rz
τ
=σ =
τ x´ y ′ σ y ´ τ y ´ z ´ rθ σ θ τ θ z
τ x´ z´ τ y´ z´ σ z´ τ rz τ θ z σ z
31
Stress Tensor in a Spherical
Coordinate System
dV = r 2 sen θ dr dθ dϕ
The cylindrical coordinate system is defined by:
x = r sen θ cos φ
x ( r ,θ , ϕ ) ≡ y = r sen θ sen φ
The components of the stress tensor are then: z = r cos θ
σ x´ τ x´ y´ τ x´ z´ σ r τ rθ τ rφ
σ ≡ τ x´ y′ σ y´ τ y´ z´ = τ
rθ σ θ τ θφ
τ x´ z´ τ y´ z´ σ z´ τ rφ τ φθ σ φ
32
4.7. Mohr´s Circle
Ch.4. Stress
33
Mohr’s Circle
Introduced by Otto Mohr in 1882.
Mohr´s Circle is a two-dimensional graphical representation of
the state of stress at a point that:
will differ in form for a state of stress in 2D or 3D.
illustrates principal stresses and maximum shear stresses as well as stress
transformations.
is a useful tool to rapidly grasp
the relation between stresses for a
given state of stress.
34
4.8. Mohr´s Circle for a 3D State of
Stress
Ch.4. Stress
35
Determination of Mohr’s Circle
Consider the system of Cartesian axes linked to the principal
directions of the stress tensor at an arbitrary point P of a
continuous medium: x3
ê3
The components of the stress tensor are ê1 σ3
σ 1 0 0 x1
σ ≡ 0 σ 2 0 with σ2
0 0 σ 3 ê2
x2
The components of the traction vector are
σ 1 0 0 n1 σ 1 n1
t = σ ⋅ n = 0 σ 2 0 n2 = σ n
2 2
0 0 σ 3 n3 σ 3 n3
where n is the unit normal to the base associated to the principal directions
36
Determination of Mohr’s Circle
The normal component of stress σ is
n1
σ = t ⋅ n = [σ 1n1 , σ 2 n2 , σ 3n3 ] n2 = σ 1n12 + σ 2 n22 + σ 3n32
T n3
t
n σ n= σ ⋅n
The squared modulus of the traction vector is
37
Determination of Mohr’s Circle
The previous system of equations can be written as a matrix
equation which can be solved for any couple
σ 12 σ 22 σ 32 n12 σ 2 + τ 2
2
σ
1 σ 2 σ 3 n
2 = σ
1 1 1 n3 1
2
A x b
0 ≤ n12 ≤ 1
A feasible solution for x ≡ n12 , n22 , n32 requires that for the
T
0 ≤ n2 ≤ 1
2
38
Determination of Mohr’s Circle
The system σ 12 σ 22 σ 32 n12 σ 2 + τ 2
2
σ
1 σ 2 σ 3 n
2 = σ
1 1 1 n3 1
2
A x b
can be re-written as
( I ) → σ 2 + τ 2 − (σ 1 + σ 3 ) σ + σ 1σ 3 −
A
n12 =
0
(σ 1 − σ 3 )
( II ) → σ 2 + τ 2 − (σ 2 + σ 3 ) σ + σ 2σ 3 −
A
n22 =
0
(σ 2 − σ 3 )
( III ) → σ 2 + τ 2 − (σ 1 + σ 2 ) σ + σ 1σ 2 −
A
n32 =
0
(σ 1 − σ 2 )
with (σ 1 − σ 2 ) (σ 2 − σ 3 ) (σ 1 − σ 3 )
A=
39
Determination of Mohr’s Circle
Consider now equation ( III ) :
σ 2 + τ 2 − (σ 1 + σ 2 ) σ + σ 1σ 2 −
A
0 with
n32 = (σ 1 − σ 2 ) (σ 2 − σ 3 ) (σ 1 − σ 3 )
A=
(σ 1 − σ 2 )
It can be written as: = a
1
(σ 1 + σ 2 )
2
(σ − a ) + τ 2 =
2
R 2 with 1
(σ 1 − σ 2 ) + (σ 2 − σ 3 ) (σ 1 − σ 3 ) n32
2
=
R
4
which is the equation of a semicircle of center C3 and radius R3 :
1
=C3 (σ 1 + σ 2 ) , 0
2 REMARK
A set of concentric semi-circles is
1
( σ 1 − σ 2 ) + (σ 2 − σ 3 ) (σ 1 − σ 3 ) n32
2
R=
3
4 obtained with the different values of
n3 with center C3 and radius R3 ( n3 ) :
n32 = 0 =
R3min
1
2
(σ 1 − σ 2 )
n32 = 1 R3max =
1
2
(σ 1 + σ 2 ) − σ 3
40
Determination of Mohr’s Circle
Following a similar procedure with ( I ) and ( II ) , a total of three
semi-annuli with the following centers and radii are obtained:
1
1
C1 [ (σ 2 + σ 3 ) , 0]
=
=
R1min (σ 2 − σ 3 )
2
2
a1 = σ 1 − a1
R1max
1
1
C2 [ (σ 1 + σ 3 ) , 0]
= =
R2
max
(σ 1 − σ 3 )
2 2
a2 = σ 2 − a2
R2min
1 1
=
C3 [ (σ 1 + σ 2 ) , 0] =
R3min (σ 1 − σ 2 )
2
2
a3 = σ 3 − a3
R3max
41
Determination of Mohr’s Circle
Superposing the three annuli,
45
2D State of Stress
3D general state of stress 2D state of stress
σ x τ xy τ xz σ x τ xy 0
σ ≡ τ yx σ y 0
σ ≡ τ yx σ y τ yz
τ zx τ zy σ z 0 0 σ z
3D problem
REMARK σ x τ xy
σ≡
In 2D state of stress problems, the τ yx σ y
principal stress in the disregarded
direction is known (or assumed) a priori. 2D (plane)
problem
46
Stresses in a oblique plane
Given a plane whose unit normal n forms an angle θ with the
x axis,
Traction vector
σ x τ xy cos θ σ x cos θ + τ xy sin θ
t = σ ⋅n = = τ cos θ + σ sin θ
τ
σ θ
xy
sin
xy y
y
σ n
Normal stress
σ x +σ y σ x −σ y
σ θ =t ⋅ n = + cos ( 2θ ) + τ xy sin ( 2θ ) cos θ
=
sin θ
2 2 n= m − cos θ
sin θ
Shear stress
σ x −σ y
τ θ =t ⋅ m = sin ( 2θ ) − τ xy cos ( 2θ )
2
48
Direct Problem
In the x´and y´axes, τ α = 0 then,
σ x −σ y
τα
= sin ( 2α ) − τ xy cos ( 2=
α) 0
2
τ xy
tan ( 2α ) =
σ x −σ y
2
Using known trigonometric relations,
1 τ xy
sin ( 2α ) =
± =
±
This equation has two solutions:
1 σ x −σ y
2
1+ 2
tg ( 2α ) + τ xy 1. α1 ( sign "+ ")
2
2
π
σ x −σ y 2. α2 =
α1 + ( sign " − ")
2
1 These define the principal stress
cos ( 2α ) =
± =
± 2
1 + tg ( 2α )
2
σ x −σ y
2
directions.
+ τ xy
2
2 (The third direction is perpendicular to
the plane of analysis.)
49
Direct Problem
The angles θ = α1 and θ = α 2 are then introduced into the equation
σ x +σ y σ x −σ y
σθ = + cos ( 2θ ) + τ xy sin ( 2θ )
2 2
σx +σ y σ x −σ y
2
σ 1 = + + τ xy
2
2 2
σα →
σx +σ y σ x −σ y
2
σ 2 = − + τ xy
2
2 2 θ ≡α
50
Inverse Problem
Given the directions and principal stresses σ 1 and σ 2, to find the
stresses in a plane characterized by the angle β :
Take the equations
Replace , , and θ ≡ β
to obtain:
σ1 + σ 2 σ1 − σ 2
σβ
= + cos ( 2 β )
2 2
σ −σ
τ β = 1 2 sin ( 2β )
2
51
Mohr’s Circle for a 2D State of Stress
52
Mohr’s Circle for a 2D State of Stress
53
Construction of Mohr’s Circle
Interactive applets and animations:
by M. Bergdorf:
http://www.zfm.ethz.ch/meca/applets/mohr/Mohrcircle.htm
54
Mohr’s Circle’s Properties
A. To obtain the point in Mohr’s Circle representative of the state of
stress on a plane which forms an angle β with the principal stress
direction σ 1:
1. Begin at the point on the circle (representative of the plane where σ 1
acts).
2. Rotate twice the angle in the sense σ 1 → σ β .
3. This point represents the shear and normal stresses at the desired plane
(representative of the stress state at the plane where acts).
3.
2. 1.
55
Mohr’s Circle’s Properties
B. The representative points of the state of stress on two
orthogonal planes are aligned with the centre of Mohr’s Circle:
π
This is a consequence of property A as β=
2 β1 + .
2
56
Mohr’s Circle’s Properties
C. If the state of stress on two orthogonal planes is known, Mohr’s
Circle can be easily drawn:
1. Following property B, the two points representative of these planes will
be aligned with the centre of Mohr’s Circle.
2. Joining the points, the intersection with the σ axis will give the centre of
Mohr’s Circle.
3. Mohr’s Circle can be drawn.
3.
1.
2.
57
Mohr’s Circle’s Properties
D. Given the components of the stress tensor in a particular
orthonormal base, Mohr’s Circle can be easily drawn:
This is a particular case of property C in which the points
representative of the state of stress on the Cartesian planes is known.
1. Following property B, the two points representative of these planes will
be aligned with the centre of Mohr’s Circle.
2. Joining the points, the intersection with the σ axis will give the centre of
Mohr’s Circle.
3. Mohr’s Circle can be drawn. 3.
1.
2.
σ x τ xy
σ=
τ xy σ y
58
Mohr’s Circle’s Properties
The radius and the diametric points of the circle can be obtained:
σ x τ xy
σ=
τ
xy σ y
σx −σ y
2
= + τ xy
2
R
2
59
Mohr’s Circle’s Properties
Note that the application of property A for the point
representative of the vertical plane implies rotating in
the sense contrary to angle.
σ x τ xy
σ=
τ xy σ y
60
The Pole or the Origin of Planes
The point called pole or origin of planes in Mohr’s circle has the
following characteristics:
Any straight line drawn from the pole will intersect the Mohr circle at a
point that represents the state of stress on a plane parallel in space to
that line.
2.
1.
61
The Pole or the Origin of Planes
The point called pole or origin of planes in Mohr’s circle has the
following characteristics:
If a straight line, parallel to a given plane, is drawn from the pole, the
intersection point represents the state of stress on this particular plane.
1.
2.
62
Sign Convention in Soil Mechanics
The sign criterion used in soil mechanics, is the inverse of the one
used in continuum mechanics:
In soil mechanics,
negative (−) tensile stress
σβ
positive (+) compressive stress
63
Sign Convention in Soil Mechanics
For the same stress state, the principal stresses will be inverted.
σ 1* = −σ 2
τ β* = −τ β σ 2* = −σ 1
σ β* = −σ β β *= β +
π
continuum mechanics 2
soil mechanics
65
Particular Cases of Mohr’s Circles
Hydrostatic state of stress
66
Chapter 4
Stress
rs
ee
s gin
4.1 Forces Acting on a Continuum Body
t d le En
r
Two types of forces that can act on a continuous medium will be considered:
ba
ge ro or
eS m
body forces and surface forces.
ci
f
ra
C d P cs
4.1.1 Body Forces
b
a
i
an an n
y ha
Definition 4.1. The body forces are the forces that act at a distance
le
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
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 ΔVi
s gin
that contain the particle P, and the corresponding sequence of body
forces fΔVi , there exists the limit
t d le En
fΔVi
ρb (x,t) = lim
ΔVi →0 ΔVi
r
ba
ge ro or
and, in addition, it is independent of the sequence of volumes con-
eS m
ci
sidered.
f
ra
C d P cs
b
a
i
an an n
y ha
Earth’s surface, obtain the value of the total resultant of the body forces in
terms of the gravitational constant g.
M
.A
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
b (x,t) ≡ [ 0 , 0 , −g ]T
not
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
s gin
surface forces acting on the boundary ∂V of volume V can be written as
fS = t (x,t) dS .
t d le En
(4.2)
∂V
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Remark 4.2. In the definition of surface forces given in (4.2), the ex-
b
a
istence of the vector of surface forces per unit of surface t (x,t) (trac-
i
an an n
forces fΔ Si are considered (see Figure 4.3), there exists the limit
M
.A
fΔ S i
t (x,t) = lim
Δ Si →0 Δ 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
s gin
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
r
and an arbitrary surface containing point P and with a unit normal vector n at
ba
ge ro or
eS m
this point, which divides the continuous medium into two parts (material vol-
ci
f
umes). The surface forces due to the contact between volumes will act on the
ra
C d P cs
imaginary separating surface, considered now a part of the boundary of each of
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
Remark 4.3. Consider a particle P of a continuous medium and dif-
eS m
ci
ferent surfaces that contain this point P such that they all have the
f
ra
same unit normal vector n at said point. In accordance with Cauchy’s
C d P cs
b
a
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
ee
s gin
4.3 Stress Tensor
t d le En
4.3.1 Application of Newton’s 2nd Law to a Continuous Medium
r
ba
ge ro or
Consider a discrete system of particles in motion such that a generic particle i
eS m
of this system has mass mi , velocity vi and acceleration ai = dvi /dt. In addition,
ci
f
ra
a force fi acts on each particle i, which is related to the particle’s acceleration
C d P cs
b
a
through Newton’s second law3 ,
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
R= ρb dV + t dS = a dm = ρa dV . (4.5)
V
Vt ∂Vt M
ρdV t
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
s gin
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
r
the tetrahedron is introduced in Figure 4.7 as well as the corresponding normals
ba
ge ro or
eS m
with which they are associated.
ci
f
ra
According to Cauchy’s second postulate (see Definition 4.4), the traction vec-
C d P cs
b
a
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
.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
ee
s gin
Figure 4.7: Traction vectors on an elemental tetrahedron.
t d le En
r
ba
ge ro or
eS m
ci
f
Remark 4.4. The mean value theorem establishes that, given a
ra
C d P cs
(scalar, vectorial o tensorial) function that is continuous in the in-
b
a
i
terior of a (compact) domain, the function reaches its mean value
an an n
continuous in Ω ,
le
liv or ec
∃ 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
∗
t(i) (x) dS = t(i) Si i ∈ {1, 2, 3} , t (x) dS = t∗ S ,
ee
s gin
Si S
(4.8)
ρ (x) b (x) dV = ρ ∗ b∗V ρ (x) a (x) dV = ρ ∗ a∗V .
t d le En
and
r
V V
ba
ge ro or
eS m
ci
Applying now (4.5) on the tetrahedron considered, results in
f
ra
C d P cs
b
a
ρb dV + t dS + t dS + t dS + t dS =
i
an an n
y ha
V S S1 S2 S3
(4.9)
le
= ρb dV + t dS + −t(1) dS + −t(2) dS + −t(3) dS = ρa dV,
liv or ec
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
∗ ∗ ∗
ρ ∗ b∗ V + t∗ S − t(1) S1 − t(2) S2 − t(3) S3 = ρ ∗ a∗ V .
er
tin
(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 PP = 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
s gin
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
The traction vector t(1) can be written in terms of its corresponding Cartesian
r
ba
ge ro or
eS m
components (see Figure 4.9) as
ci
f
ra
C d P cs
t(1) = σ11 ê1 + σ12 ê2 + σ13 ê3 = σ1i êi . (4.15)
b
a
i
an an n
Operating in an analogous manner on traction vectors t(2) and t(3) (see Fig-
y ha
t(2) = σ21 ê1 + σ22 ê2 + σ23 ê3 = σ2i êi (4.16)
M
.A
(4.17)
d
uu
e
er
tin
.O
(i)
σi j (P) = t j (P) i, j ∈ {1, 2, 3}
C
(4.19)
©
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
le
liv or ec
M
.A
t (P, n) = n · σ (P)
er
(4.20)
tin
on
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
ee
s gin
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
r
ba
means of (4.20), the stress tensor σ (P) is seen to contain informa-
ge ro or
eS m
ci
tion on the traction vectors corresponding to any plane (identified by
f
ra
its normal n) that contains this point.
C d P cs
b
a
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)
σ31 σ32 σ33
and each component σi j can be characterized in terms of its indices:
− Index i indicates the plane on which the stress acts (plane perpendicular to
the xi -axis).
− Index j indicates the direction of the stress (direction of the x j -axis).
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).
s gin
t d le En
4.3.3.2 Engineering Notation
In engineering notation, the components of the Cauchy stress tensor (see Fig-
r
ba
ge ro or
ure 4.13) are written as
eS m
⎡ ⎤
ci
σx τxy τxz
f
ra
not ⎢ ⎥
C d P cs
σ ≡ ⎣ τyx σy τyz ⎦
b
a
(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.
− The component τab is the tangential (shear) stress acting on the plane per-
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
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.
The sign criterion for the visible faces is
s gin
positive (+) ⇒ tension
t d le En
Normal stresses σi j or σa and
negative (−) ⇒ compression
r
ba
ge ro or
eS m
ci
positive (+) ⇒ direction of b-axis
Tangential stresses τab f
ra
C d P cs
negative (−) ⇒ opposite direction to b-axis
b
a
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
s gin
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
t∗ (x,t) acts on ∂V . The acceleration vector field of the particles is a (x,t) and
r
ba
ge ro or
eS m
the Cauchy stress tensor field is σ (x,t) (see Figure 4.16).
ci
f
ra
C d P cs
b
a
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
⎧
s gin
Internal ⎨ ∇ · σ + ρb = 0 ∀x ∈ V
∂ σi j
t d le En
equilibrium (4.26)
equation ⎩ ∂ x + ρb j = 0 j ∈ {1, 2, 3}
r
i
ba
ge ro or
eS m
ci
f
which is known as the internal equilibrium equation of the continuous medium.
ra
C d P cs
Cauchy’s equation of motion is derived from the principle of balance of linear
b
a
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
e
∀x ∈ ∂V
er
tin
equation at (4.27)
the boundary ni σi j = t ∗j j ∈ {1, 2, 3}
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
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) ∀x ∈ ∂V
rs
ni σi j = σ ji ni = t j∗ j ∈ {1, 2, 3}
ee
s gin
t d le En
r
Example 4.2 – A continuous medium moves with a velocity field whose spa-
ba
ge ro or
eS m
tial description is v (x,t) ≡ [z, x, y]T . The Cauchy stress tensor is
not
ci
f
ra
⎡ ⎤
C d P cs
y g (x, z,t) 0
b
a
not ⎢ ⎥
i
an an n
σ ≡ ⎣ h (y) z (1 + t) 0 ⎦ .
y ha
0 0 0
le
liv or ec
Determine the functions g, h and the spatial form of the body forces b (x,t)
M
.A
d
uu
Solution
e
X Th
er
tin
.O
h (y) = C ,
σ =σ T
=⇒ h (y) = g (x, z,t) =⇒
g (x, z,t) = C ,
C
∇ · σ + ρb = ρa
=⇒ b=a.
∇·σ = 0
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
∂z
ee
s gin
the acceleration
⎡ ⎤
0 1 0
t d le En
a = v · ∇v ≡ [z, x, y] ⎣ 0 0 1 ⎦ = [y, z, x]
not
r
ba
ge ro or
1 0 0
eS m
ci
f
ra
is obtained. Finally, the body forces are
C d P cs
b
a
i
an an n
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
not ⎢ ⎥
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 145
rs
ee
Figure 4.17: Diagonalization of the stress tensor.
s gin
t d le En
r
Definition 4.5. The principal stress directions are the directions, as-
ba
ge ro or
sociated with the axes {x , y , z }, in which the stress tensor diago-
eS m
ci
nalizes.
f
ra
C d P cs
The principal stresses are the eigenvalues of the stress tensor
b
a
(σ1 , σ2 , σ3 ). In general, they will be assumed to be arranged in the
i
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 .
(σ
er
(4.31)
tin
The solution to this system will not be trivial (will be different to v = 0) when
on
.O
not
det (σ σ − λ 1| = 0 .
σ − λ 1) = |σ (4.32)
Equation (4.32) is a third-grade polynomial equation in λ . Since tensor σ
is symmetric, its three solutions (λ1 ≡ σ1 , λ2 ≡ σ2 , λ3 ≡ σ3 ) are real. Once the
eigenvalues have been found and ordered according to the criterion σ1 ≥ σ2 ≥ σ3 ,
the eigenvector v(i) can be obtained for each stress σi by resolving the system in
(4.31),
σ − σi 1) · v(i) = 0
(σ i ∈ {1, 2, 3} . (4.33)
This equation provides a non-trivial solution of the eigenvectors v(i) , orthogo-
nal between themselves, which, once it has been normalized, defines the three
elements of the base corresponding to the three principal directions.
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
ee
s gin
Definition 4.6. The mean stress is the mean value of the principal
t d le En
stresses.
r
1
ba
σm = (σ1 + σ2 + σ3 )
ge ro or
eS m
3
ci
f
ra
C d P cs
b
a
i
an an n
Considering the matrix of components of the stress tensor in the principal stress
y ha
1 1
σm = σ) .
(σ1 + σ2 + σ3 ) = Tr (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
3
.O
C
0 0 σ
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
s gin
The stress tensor σ can be split7 into a spherical part (or component) σ sph and
a deviatoric part σ ,
t d le En
σ = σ sph + σ . (4.35)
r
ba
ge ro or
eS m
spherical deviatoric
ci
part part
f
ra
C d P cs
b
a
The spherical part is defined as
i
an an n
⎡ ⎤
σm 0
y ha
0
de f 1 not ⎢ ⎥
le
σ sph : = σ ) 1 = σ m 1 ≡ ⎣ 0 σm 0 ⎦ ,
Tr (σ (4.36)
liv or ec
3
0 0 σm
M
.A
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
not ⎢ ⎥ ⎢ ⎥
σ = σ − σ sph ≡ ⎣ τxy σy τyz ⎦ − ⎣ 0 σm 0 ⎦
on
.O
(4.37)
τxz τyz σz 0 σm
C
0
©
resulting in
⎡ ⎤ ⎡ ⎤
σx − σm τxy τxz σx τxy τxz
not ⎢ ⎥ ⎢ ⎥
σ ≡ ⎣ τxy σy − σm τyz ⎦ = ⎣ τxy σy τyz ⎦ . (4.38)
τxz τyz σz − σm τxz τyz σz
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
(see (4.37) and Remark 4.11).
ee
s gin
t d le En
r
Remark 4.13. The principal directions of the stress tensor and of its
ba
ge ro or
eS m
deviatoric tensor coincide. Proof is trivial considering that, from Re-
ci
mark 4.11, the spherical part σ sph is diagonal in any coordinate sys-
f
ra
C d P cs
tem. Consequently, if σ diagonalizes for a certain basis in (4.37), σ
b
a
i
will also diagonalize for that basis.
an an n
y ha
le
liv or ec
M
.A
er
tin
on
.O
C
8 The 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 , the corresponding J
s gin
invariants are (see Remark 4.14 and equations (4.39) to (4.44))
⎧
⎪
t d le En
⎫ ⎪
⎪ J1 = I1 = 0
⎪
⎪
J1 = I1 = 0 ⎪
⎬ ⎪
⎨
r
ba
ge ro or
=⇒ σ =⇒ J2 = I2 = 1 (σ 1
eS m
J2 = I2 σ : σ ) = σ i j σ ji
ci
⎪
⎭ ⎪
⎪
⎪ f2 2
ra
J3 = I3 ⎪
⎪
C d P cs
⎪ 1
b
⎩ J3 = I3 = σ i j σ jk σ ki
a
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 {x , y , z } defined as dextrorotatory are considered at this point.
on
.O
⎡ ⎤ ⎡ ⎤
©
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
x (r, θ , z) ≡ ⎣ y = r sin θ ⎦
not
z=z
rs
ee
Figure 4.18: Cylindrical coordinates.
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
x (r, θ , φ ) ≡ ⎣ y = r sin θ sin φ ⎦
not
z = z cos θ
rs
ee
s gin
Figure 4.20: Spherical coordinates.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
Consider now the normal component σ n = σ n, where σ is the normal com-
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,
σ > 0 tension ,
r
ba
σn = σ ·n
ge ro or
(4.47)
eS m
σ < 0 compression .
ci
f
ra
C d P cs
Consider now the tangential component τ n , of which only its module is of inter-
b
a
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
er
Figure 4.23. If the infinite number of planes that contain point P are now con-
tin
sidered (characterized by all the possible unit normal vectors n(i) ) and the corre-
on
sponding values of the normal stress σi and tangential stress τi are obtained and,
.O
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
rs
σ1 0 0 n1 σ1 n1
not ⎢ ⎥⎢ ⎥ ⎢
ee
⎥
t = σ · n ≡ ⎣ 0 σ2 0 ⎦ ⎣ n2 ⎦ = ⎣ σ2 n2 ⎦ , (4.51)
s gin
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-
r
sociated with the principal stress directions. In view of (4.51), the normal com-
ba
ge ro or
eS m
ponent of the stress (σ ), defined in (4.47), is
ci
⎡ ⎤ f
ra
C d P cs
n1
b
a
⎢ ⎥
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
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-
s gin
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.,
0 ≤ n21 ≤ 1 , 0 ≤ n22 ≤ 1 0 ≤ n23 ≤ 1 .
r
and
ba
ge ro or
eS m
Every pair (σ , τ) that leads to a solution x that satisfies this require-
ci
f
ra
ment will be considered a feasible point of the half-space σ − τ,
C d P cs
b
a
which is representative of the stress state on a plane that contains P.
i
The locus of feasible points (σ , τ) is named feasible zone of the half-
an an n
y ha
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
⎪
⎪ (σ − σ )
⎪
⎪
2 3
(4.57)
⎪ A
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)
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
s gin
expression of R3 in (4.59) is positive, these values are obtained for n23 = 0 (the
minimum radius) and n23 = 1 (the maximum radius).
t d le En
1
r
n23 = 0 =⇒ R3min = (σ1 − σ2 )
ba
ge ro or
2
eS m
(4.60)
ci
1f
ra
n23 =1 =⇒ R3max = (σ1 + σ2 ) − σ3
C d P cs
b
a
2
i
an an n
This process is repeated for the other two equations, (I) and (II), in (4.57),
M
.A
resulting in: ⎧
&1 ' ⎨ R min = 1 (σ − σ )
m
1 2 3
− Equation (I) : C1 = (σ2 + σ3 ), 0 =⇒ 2
d
2 ⎩ R max = |σ − a |
uu
e
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
ee
s gin
Figure 4.25: Feasible region.
t d le En
r
⎧
ba
ge ro or
&1 ' ⎨ R min = 1 (σ − σ )
eS m
ci
2 1 3
− Equation (II) : C2 = (σ1 + σ3 ), 0 =⇒
f 2
ra
2 ⎩ R max = |σ − a |
C d P cs
b
a
2 2 2
a2
i
an an n
⎧
y ha
2 ⎩ R max = |σ − a |
M
.A
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
ee
s gin
The Mohr’s circle for the defined stress state is drawn and the given points
are marked in the half-space σ − τ.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
s gin
the two dimensions associated with the x1 - and x2 -axes (or x- and y-axes), as
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
( ) ( )
r
σ11 σ12 σx τxy
ba
ge ro or
eS m
not
σ≡ = (4.63)
ci
σ12 σ22 τxy σy
f
ra
C d P cs
b
a
and the components of the traction vector
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).
s gin
• Unit tangent vector m: generates a clockwise rotation with re-
spect to point P.
t d le En
• Angle θ : defined as counterclockwise.
r
ba
ge ro or
eS m
ci
f
ra
Consider σ , the stress tensor at a given point, whose components are defined
C d P cs
b
a
in a Cartesian base, ( )
i
an an n
σx τxy
y ha
not
σ≡ . (4.65)
τxy σy
le
liv or ec
Using (4.64), the traction vector on the given point, which belongs to the plane
M
.A
considered, is
m
( )( ) ( )
d
not
t = σ ·n ≡ = .
e
(4.66)
X Th
.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)
2
s gin
= σx sin θ cos θ − σy sin θ cos θ + τxy sin θ − cos θ ,2
t d le En
which can be rewritten as10
r
ba
ge ro or
eS m
σx + σy σx − σy
ci
σθ = + cos (2θ ) + τxy sin (2θ )
f
ra
2 2
C d P cs
(4.69)
b
a
σx − σy
i
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 τxy
rs
τα = sin (2α) − τxy cos (2α) = 0 =⇒ tan (2α) = σ − σ ,
ee
2 x y
2
s gin
1 τxy
sin (2α) = ± % = ± * ,
t d le En
1
2
1+ 2 σx − σy
tan (2α) + τxy
2
ba
(4.70)
ge ro or
2
eS m
ci
f
ra
σx − σy
C d P cs
b
a
1
i
cos (2α) = ± + = ± * 2 .
an an n
1 + tan2 (2α)
2
y ha
σx − σy
+ τxy
2
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
⎧
©
*
⎪
⎪
⎪
⎪ σ x + σy σx − σy 2
⎪
⎨ σ1 = + + τxy
2
2 2
σα → *
(4.72)
⎪
⎪ 2
⎪
⎪ σ + σ σ − σ
⎪
⎩ σ2 =
x y
−
x y
+ τxy
2
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.
ee
s gin
4.7.3 Inverse Problem
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
r
ba
analyzed. The stress state on any plane is characterized by the angle β that
ge ro or
eS m
ci
forms the unit normal vector of the plane with the principal stress direction cor-
f
ra
responding to σ1 . As a particular case, the components of the stress tensor on
C d P cs
b
a
an elemental rectangle associated with the system of axes x − y can be obtained
i
an an n
Consider now the Cartesian system x − y , associated with the principal stress
le
directions (see Figure 4.30). Applying (4.69) with σx = σ1 , σy = σ2 , τx y = 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
s gin
σ1 + σ2 2 σ1 − σ 2 2
σ− +τ =2
. (4.76)
t d le En
2 2
r
Note that this equation, which will be valid for any value of the angle β , or,
ba
ge ro or
eS m
in other words, for any arbitrarily oriented plane that contains the point, corre-
ci
f
sponds to a circle with center C and radius R in the plane σ − τ given by (see
ra
C d P cs
b
a
Figure 4.31)
i
σ1 + σ2 σ1 − σ2
an an n
C= ,0 and R = . (4.77)
y ha
2 2
le
liv or ec
.A
in Figure 4.31.
m
The inverse proposition is also true: given a point of Mohr’s circle with co-
d
ordinates (σ , τ), there exists a plane that contains P whose normal and tangent
uu
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 .
s gin
t d le En
σ1 + σ2
σ−
2 σ −a
r
cos (2β ) =
=
ba
ge ro or
eS m
σ1 − σ2 R
ci
f
ra
2 (4.78)
C d P cs
b
a
τ τ
i
sin (2β ) =
=
an an n
σ1 − σ R
y ha
2
le
2
liv or ec
M
These expressions uniquely define the angle β between the normal direction to
.A
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
ee
s gin
t d le En
Figure 4.33: Representative point associated with angle β in Mohr’s circle.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
Figure 4.36: First property of the pole of Mohr’s circle.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
12Note 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).
s gin
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.
r
ba
ge ro or
eS m
2) Consider now an arbitrary plane whose normal direction forms an angle θ
ci
f
ra
with the horizontal direction (see Figure 4.39, left) and consider also the
C d P cs
normal and tangent stresses, σθ and τθ , respectively, according to this plane.
b
a
i
Assuming that the major principal stress direction σ1 forms an angle α with
an an n
.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:
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
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
r
ba
and determine, thus, the diameter of the circle.
ge ro or
eS m
The pole is obtained as the intersection of the lines that are parallel to the two
ci
f
planes inclined at 45◦ and that contain the points that they represent. Once
ra
C d P cs
b
a
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-
tal plane (2, 1). The same procedure is repeated for a vertical plane to obtain
le
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
ee
s gin
4.7.7 Mohr’s Circle with the Soil Mechanics Sign Criterion
The sign criterion, with respect to the normal and tangent stresses, used in soil
t d le En
mechanics is the inverse of the one used in continuum mechanics (see Fig-
r
ure 4.40). The differences are:
ba
ge ro or
eS m
ci
• The positive stresses in soil mechanics are in the opposite direction (normal
f
ra
C d P cs
stresses are positive when they are compressive, and the direction of the pos-
b
a
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
− cos (2β ∗ )
s gin
−σ2∗ + σ1∗ (4.80)
−τβ∗ = sin (2β ∗ − π) ,
t d le En
2
− sin (2β ∗ )
r
ba
ge ro or
eS m
and, operating on these expressions finally results in
ci
f
ra
σ1∗ + σ2∗ σ1∗ − σ2∗
C d P cs
σβ∗ =
b cos (2β ∗ ) ,
a
+
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
⎡ ⎤ ⎧
σm 0 0 ⎪
⎨ σ1 = σm + σ1
not ⎢ ⎥
σ= σ sph + σ ; σ sph ≡ ⎣ 0 σm 0 ⎦ =⇒ σ2 = σm + σ2
⎪
⎩
spherical deviator 0 0 σm σ3 = σm + σ3
part part
rs
ee
s gin
t d le En
r
ba
ge ro or
eS m
Figure 4.43: Mohr’s circle for a stress state and its deviator.
ci
f
ra
C d P cs
b
a
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.
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
s gin
t d le En
r
ba
ge ro or
eS m
ci
Determine:
f
ra
C d P cs
1) The expression of the forces per unit of mass acting on the solid.
b
a
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
( ) ( )
on
.O
1 not 1 ∂ ∂ xy 5y 1 y+5
b = − ∇ · σ =⇒ b ≡ − , =− .
ρ ρ ∂x ∂y ρ
C
5y 4x 0
©
2) The normal (σ ) and tangent (τ) components of the body forces acting on the
boundary are given by
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
s gin
( )( )
1 xy 5y 1 xy + 5y
t d le En
not
t1 = n1 · σ ≡ √ [1, 1] =√ .
2 5y 4x 2 5y + 4x
r
ba
ge ro or
eS m
ci
Then, the corresponding normal and tangent components of the body forces are
f
ra
( )
C d P cs
b
a
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] √ = (−4x + xy) .
M
.A
2 2 −1 2
m
This is now particularized for the x and y values corresponding to the boundary,
d
uu
er
⎧
tin
⎪
⎪ 1
⎨ σ1 = 10 − 5x − x2 with x ∈ [0, 1] ,
on
.O
2
⎪
⎩ τ1 = 1
−3x − x2 with x ∈ [0, 1] .
⎪
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],
⎧
⎨ σ2 = 4x with x ∈ [0, 1] ,
rs
ee
⎩τ = 0.
2
s gin
t d le En
Boundary 3
r
The traction vector for this surface is
ba
ge ro or
( ) ( )
eS m
ci
−xy
f xy 5y
ra
not
t3 = n3 · σ ≡ [−1, 0] = .
C d P cs
b −5y
a
5y 4x
i
an an n
y ha
Then, the corresponding normal and tangent components of the body forces are
( )
le
liv or ec
not −1
σ3 = t3 · n3 ≡ [−xy, −5y] = xy ,
M
.A
0
m
( )
d
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
ee
s gin
⎧ ⎧
⎨ σ2 = σy with x ∈ [0, 1] ⎨ σ3 = σx with x = 0
t d le En
⎩τ = τ with y = 0 ⎩ τ = −τ = −5y with y ∈ [0, 1]
r
2 xy 3 xy
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
Solution
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.
r
ba
By means of the definition of the deviatoric stress tensor,
ge ro or
eS m
ci
σ = σ − σm 1 =⇒ f
σzz = σzz − σm = a − a = 0 =⇒ σzz = 0
ra
C d P cs
b
a
i
is deduced. The fact that the trace is an invariant and that the trace of the devia-
an an n
σ ) = 0, results in
toric stress tensor is zero, Tr (σ
y ha
le
liv or ec
σzz = σ2 = 0 ,
σxx + σyy = 0 =⇒
M
.A
σ1 + σ3 = 0 .
m
Finally, the radius of the major circumference (between σ1 and σ3 ) is determined
d
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.
ee
s gin
t d le En
Solution
r
ba
The following property of the Mohr’s circle in 3D must be taken into account to
ge ro or
eS m
solve this problem.
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
2. σx is the intermediate principal stress, which results in the following
r
ba
ge ro or
Mohr’s circle.
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
X Th
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
ee
s gin
t d le En
r
Solution
ba
ge ro or
eS m
ci
f
The problem is solved following the same steps in all three cases, which are:
ra
C d P cs
b
a
Step 1: Draw the Mohr’s circle corresponding to the stress state. Even for the
i
stress states in which only two different pairs of points (σ , τ) belonging to the
an an n
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
these planes and, thus, the angles α and β are given directly by the orientation
C
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
(c)
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
Solution
ba
ge ro or
eS m
ci
Stress state II on the vertical plane must be found to be able to add states I and
f
ra
II together.
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
σ = 3
σ = 7
The values of τ 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 and
ee
α.
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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 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 must be
©
√
τ = −3 2 .
Since there are two possible values of τ , two values of α will exist, each cor-
responding to one of the τ 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
ee
s gin
t d le En
r
Determining the inclination of these two lines will result directly in the possible
ba
ge ro or
eS m
values of α.
ci
f
ra
C d P cs
√ & √ '
b
a
τ = −3 2 ⇒ α + = 180◦ − arctan 1+3 2 141◦
i
an an n
&√ '
y ha
√
−τ = 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
ee
s gin
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
r
direction of σ1 is τmax = 2. Obtain all the values of σ1 , σ2 and σ3 that make
ba
ge ro or
eS m
possible the stress state σ = 2 and τ = 2 on a certain plane for the following
ci
cases (separately). f
ra
C d P cs
b
a
a) The maximum shear stress in planes parallel to the principal stress di-
i
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
4.5 – Given states I and II, determine the possible values of σ and τ for which
ci
f
state III = I + II verifies that the principal stress σ2 is positive and its direction
ra
C d P cs
forms a 30◦ angle with the y-axis.
b
a
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
ee
s gin
a) There do not exist tensile stresses on any plane.
b) There do not exist compressive stresses on any plane.
t d le En
c) The maximum shear stress (τmax ) is less than 2.
r
d) It is a pure shear stress state.
ba
ge ro or
eS m
e) It is a hydrostatic stress state.
ci
f
ra
C d P cs
b
a
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
CH.5. BALANCE PRINCIPLES
Multimedia Course on Continuum Mechanics
Overview
Balance Principles Lecture 1
Convective Flux or Flux by Mass Transport Lecture 2 Lecture 3
Local and Material Derivative of a Volume Integral Lecture 4
Conservation of Mass
Spatial Form Lecture 5
Material Form
Reynolds Transport Theorem
Lecture 6
Reynolds Lemma
General Balance Equation Lecture 7
Linear Momentum Balance
Global Form Lecture 8
Local Form
2
Overview (cont’d)
Angular Momentum Balance
Global Spatial Lecture 9
Local Form
Mechanical Energy Balance
External Mechanical Power
Lecture 10
Mechanical Energy Balance
External Thermal Power
Energy Balance
Thermodynamic Concepts Lecture 11
First Law of Thermodynamics
Internal Energy Balance in Local and Global Forms Lecture 12
Second Law of Thermodynamics Lecture 13
Reversible and Irreversible Processes
Lecture 14
Clausius-Planck Inequality
Lecture 15
3
Overview (cont’d)
Governing Equations
Governing Equations
Lecture 16
Constitutive Equations
The Uncoupled Thermo-mechanical Problem
4
5.1. Balance Principles
Ch.5. Balance Principles
5
Balance Principles
The following principles govern the way stress and deformation vary in
the neighborhood of a point with time.
REMARK
The conservation/balance principles: These principles are always
Conservation of mass valid, regardless of the type of
material and the range of
Linear momentum balance principle
displacements or deformations.
Angular momentum balance principle
Energy balance principle or first thermodynamic balance principle
6
5.2. Convective Flux
Ch.5. Balance Principles
7
Convection
The term convection is associated to mass transport, i.e., particle
movement.
Properties associated to mass will be transported with the mass when
there is mass transport (particles motion) convective transport
amount of A crossing S
ΦS =
unit of time
8
Convective Flux or
Flux by Mass Transport
Consider:
An arbitrary property A of a continuum medium (of any tensor order)
The description of the amount of the property per unit of mass, Ψ ( x,t )
(specific content of the property A ) .
The volume of particles dV crossing a
differential surface dS during the
interval [t , t + dt ] is
dV =dS ⋅ dh =v ⋅ n dt dS
= ρ dV
dm = ρ v ⋅ n dSdt
Then,
The amount of the property crossing the differential surface per unit of
time is: Ψ dm
d ΦS = = ρ Ψ v ⋅ n dS
dt
9
Convective Flux or
Flux by Mass Transport
Consider:
An arbitrary property A of a
continuum medium (of any tensor order) inflow
v ⋅n ≤ 0
outflow
The specific content of A (the amount v ⋅n ≥ 0
of A per unit of mass) Ψ ( x,t ) .
Then,
The convective flux of A through a spatial surface, S , with unit
normal n is:
Φ S ( t )= v is velocity
∫
s
ρ Ψ v ⋅ n dS Where:
ρ is density
10
Convective Flux
REMARK 1
The convective flux through a material surface is always null.
REMARK 2
Non-convective flux (conduction, radiation). Some properties can be
transported without being associated to a certain mass of particles. Examples of
non-convective transport are: heat transfer by conduction, electric current flow,
etc.
Non-convective transport of a certain property is characterized by the non-
convective flux vector (or tensor) q ( x,t ) :
∫ q ⋅ n dS ; convective flux =
non - convective flux =
s ∫ ρψ v ⋅ n dS
s
11
Example
Compute the magnitude and the convective flux Φ S which correspond to the
following properties:
a) volume
b) mass
c) linear momentum
d) kinetic energy
12
Φ S ( t )=
Example - Solution ∫ ρ Ψ v ⋅ n dS
s
The magnitude “property content per unit of mass” is volume per unit of
mass, i.e., the inverse of density:
V 1
Ψ
= =
M ρ
The convective flux of the volume of the particles V through the surface S is:
1
Φ S= ∫ρ
s ρ
v ⋅ n dS= ∫
s
v ⋅ n dS VOLUME FLUX
13
Φ S ( t )=
Example - Solution ∫ ρ Ψ v ⋅ n dS
s
The magnitude “property per unit of mass” is mass per unit of mass, i.e., the
unit value:
M
Ψ
= = 1
M
The convective flux of the mass of the particles M through the surface S is:
Φ
= S ∫ρ
s
1 v ⋅n =
dS ∫s
ρ v ⋅ n dS MASS FLUX
14
Φ S ( t )=
Example - Solution ∫ ρ Ψ v ⋅ n dS
s
The magnitude “property per unit of mass” is mass times velocity per unit of
mass, i.e., velocity:
Mv
=
Ψ = v
M
The convective flux of the linear momentum of the particles M v through the
surface S is:
=ΦS ∫ ρ v ( v ⋅ n ) dS
s
MOMENTUM FLUX
15
Φ S ( t )=
Example - Solution ∫ ρ Ψ v ⋅ n dS
s
16
5.3. Local and Material Derivative
of a Volume Integral
Ch.5. Balance Principles
17
Derivative of a Volume Integral
Consider:
An arbitrary property A of a continuum medium (of any tensor order)
The description of the amount of the property per unit of volume
(density of the property A ), µ ( x,t )
REMARK
and Ψ are related
The total amount of the property through .
in an arbitrary volume, V, is:
Q (t )
Q ( t ) = ∫ µ ( x, t ) dV
V
Q ( t + ∆t )
The time derivative of this volume integral is:
Q ( t + ∆t ) − Q ( t )
Q′ ( t ) = lim
∆t → 0 ∆t
18
Local Derivative of a Volume Integral
Consider: Q (t )
The volume integral Q ( t ) = ∫ µ ( x, t ) dV
V Q ( t + ∆t ) Control
Volume, V
The local derivative of Q ( t ) is:
local not ∂ ∫V µ ( x, t + ∆t ) dV − V∫ µ ( x, t ) dV REMARK
derivative
= ∫ µ ( x, t ) dV ∆lim
∂t V t →0 ∆t The volume is fixed in
space (control volume).
It can be computed as:
∂ Q ( t + ∆t ) − Q ( t ) ∫ µ ( x, t + ∆t ) dV − ∫ µ ( x, t ) dV
= ∫ µ ( x , t ) dV =
lim lim V = V
∂t V ∆t → 0 ∆t ∆t → 0 ∆t
∫ [µ ( x, t + ∆t ) − µ ( x, t )]dV µ ( x, t + ∆t ) − µ ( x, t ) ∂µ ( x, t )
lim=
∆t → 0
V
∆t ∫V ∆
= lim
t →0
∆ t
dV ∫V ∂t dV
∂µ x,t
∂t
19
Material Derivative of a Volume
Integral
Consider:
The volume integral Q ( t ) = ∫ µ ( x, t ) dV
V
material
not d
µ ( x, t ) dV
dt Vt ∫≡V
derivative =
µ ( x, t + ∆t ) dV − ∫ µ ( x, t ) dV REMARK
= lim
∫ V ( t +∆t ) V (t )
The volume is mobile in space
∆t → 0 ∆t
and can move, rotate and
It can be proven that: deform (material volume).
d ∂ ∂µ dµ
( ) µ dV + ∫ ∇ ⋅ ( µ v )= ∫V ∂t + ∇ ⋅ ( µ v ) =
dt Vt∫≡V ∂t V∫ ∫V dt + µ∇ ⋅ v dV
µ x , t = dV dV dV
V
material local convective
derivative of derivative of derivative of
the integral the integral the integral
20
5.4. Conservation of Mass
Ch.5. Balance Principles
21
Principle of Mass Conservation
It is postulated that during a motion there are neither mass
sources nor mass sinks, so the mass of a continuum body is a
conserved quantity (for any part of the body).
( t ) M ( t + ∆t ) > 0
M=
Where:
=M (t ) ∫ ρ ( x, t ) dV ∀∆Vt ⊂ Vt
∆Vt
=
M ( t + ∆t ) ∫∆Vt +∆t
ρ ( x, t + ∆t ) dV ∀∆Vt +∆t ⊂ Vt +∆t
22
Conservation of Mass in Spatial Form
Conservation of mass requires that the material time derivative of
the mass M ( t ) be zero for any region of a material volume,
M ( t + ∆t ) − M ( t ) d
M ′ (=
t ) lim = ∫ ρ dV
= 0 ∀∆V ⊂ V , ∀t
∆t → 0 ∆t dt ∆V ⊂V ≡V t t
∂
→ ∫ ρ F ( X , t ) dV
= 0 ∀∆V0 ⊂ V0 , ∀t
∂t
0
∆V0 ⊂V0
25
dρ
Reynolds Lemma dt
+ ρ∇ ⋅ v =0
Consider:
An arbitrary property A of a continuum medium (of any tensor order)
The spatial description of the amount of the property per unit of
mass, ψ ( x,t ) (specific contents ofA )
The amount of the property A in the continuum body at time t
for an arbitrary material volume is: Q ( t ) = ∫ ρψ dV
Vt =V
∫ ∇ ⋅ v dV= ∫ n ⋅ vdS
V ∂V
= ∫ v ⋅ n dS
∂V
27
dψ ∂ ( ρψ )
∫ρ dt
=
dV ∫V ∂t dV + ∂∫V ρ ψ v ⋅ n dS
Reynolds Transport Theorem
V
ê3
ê 2 ∫ ρψ
V
dV
ê1
28
Reynolds Transport Theorem
∂ dψ REYNOLDS TRANSPORT
∫
∂t V
ρ ψ dV= ∫ρ
V
dt
dV − ∫ ρ ψ v ⋅ n dS
∂V
THEOREM (integral form)
∂ dψ dψ
∫
∂t V
ρ ψ dV= ∫ρ dt
dV − ∫ ρ ψ v ⋅ n dS ρ
dt
V ∂V ∂V
∂
= ∫ ( ρ ψ ) dV = ∫ ∇ ⋅ (ρ ψ v ) dV
V
∂t V
dV
∂ dψ
∫ ∂t
ρ ψ ) dV
(= ∫ [ρ
dt
− ∇ ⋅ ( ρ ψ v )] dV ∀∆V ⊂ V ∀t
∆V ⊂V ∆V ⊂V ê 3
ê 2 ∫ ρψ
V
dV
∂ dψ ê1
(ρ =
ψ) ρ − ∇ ⋅ ( ρ ψ v ) ∀x ∈ V ∀t
∂t dt
REYNOLDS TRANSPORT
THEOREM (local form)
29
5.6. General Balance Equation
Ch.5. Balance Principles
30
General Balance Equation
Consider:
An arbitrary property A of a
continuum medium (of any tensor order)
The amount of the property per
unit of mass, ψ ( x,t )
The rate of change per unit of time
of the amount of A in the control volume V is due to:
a) Generation of the property per unit mas and time time due to a source: k A ( x, t )
b) The convective (net incoming) flux across the surface of the volume. source term
c) The non-convective (net incoming) flux across the surface of the volume: jA (x, t )
So, the global form of the general balance equation is: non-convective
flux vector
∂
∂t V∫ ∫ ρ kA
ρ ψ=
dV dV − ∫ ρ ψ v ⋅ n dS − ∫ jA ⋅ n dS
V ∂V
∂V
a b c
31
∂
∫
∂t V
ρ ψ=
dV ∫ ρ kA dV − ∫ρψ v ⋅ n dS − ∫ jA ⋅ n dS
General Balance Equation
V ∂V ∂V
∂
= ∫V ∂t ( ρ ψ ) + ∇ ⋅ ( ρ ψ v ) = ∫ ( ρ kA − ∇ ⋅ jA ) dV
dV
V
dψ
=ρ (Reynolds Theorem)
dt
dψ
∫ ρ= dV ∫ ( ρ kA − ∇ ⋅ jA ) dV ∀∆V ⊂ V ∀t
∆V ⊂V
dt ∆V ⊂V
32
dψ
ρ = ρ kA − ∇ ⋅ jA
Example
dt
∂ dψ
( ρψ
= ) ρ − ∇ ⋅ ( ρ ψ v ) ∀x ∈V
∂t dt
33
5.7. Linear Momentum Balance
Ch.5. Balance Principles
34
Linear Momentum
in Classical Mechanics
Applying Newton’s 2nd Law to the discrete system formed by n
particles, the resulting force acting on the system is:
n n n
dv i
R ( t )= ∑
= fi ∑ m=
a ∑ m =
=i 1 =i 1 i = i
i 1
i
dt
Resulting force mass conservation
on the system
principle: dmi = 0
d n n
dmi dP ( t ) dt
=
=
∑ mi v i − ∑
dt i 1 =i 1 dt
vi =
dt
= P ( t ) linear momentum
For a system in equilibrium, = R 0, ∀t :
dP ( t )
=0 P ( t ) = cnt CONSERVATION OF THE
dt LINEAR MOMENTUM
35
Linear Momentum n
P ( t ) = ∑ mi v i
in Continuum Mechanics
i =1
=P (t ) v ( x, t ) d M ∫ ρ ( x, t ) v ( x, t ) dV
∫=
M V
d M = ρ dV
36
Linear Momentum Balance Principle
The time-variation of the linear momentum of a material volume is
equal to the resultant force acting on the material volume.
dP ( t ) d
= = ∫ ρ v dV R ( t )
dt dt Vt
surface forces
∫ t dS =∫ n ⋅ σ dS =∫ ∇ ⋅ σ dV
∂V ∂V V
So, the global form is rewritten:
∫ ρ b dV + ∫ t dS =
∆V ⊂V ∂∆V ⊂V
d
= ∫ ( ρ b=
+∇ ⋅ σ ) dV ∫ ρ v dV ∀∆V ⊂ V , ∀t
38 ∆V ⊂V
dt ∆Vt ⊂Vt ≡V
Local Form of the
Linear Momentum Balance Principle
Applying Reynolds Lemma to the global form of the principle:
d dv
∫ ( ∇=
⋅ σ + ρ b ) dV = ∫ ρ v dV ∫ ρ dV ∀∆V ⊂ V , ∀t
∆V ⊂V
dt ∆Vt ⊂Vt ≡V ∆V ⊂V
dt
∆V → dV (x, t )
dv(x, t )
∇ ⋅ σ (x, t ) + ρ=
b(x, t ) ρ = ρ a(x, t ) ∀x ∈ V , ∀t
dt
LOCAL FORM OF THE LINEAR
MOMENTUM BALANCE
(CAUCHY’S EQUATION OF MOTION)
39
5.8. Angular Momentum Balance
Ch.5. Balance Principles
40
Angular Momentum
in Classical Mechanics
Applying Newton’s 2nd Law to the discrete system formed by n
particles, the resulting torque acting on the system is:
n n
dv i
) ∑ ri × f=i ∑ ri × mi =
MO ( t=
=i 1 =i 1 dt
=0 n
d n n
dri d dL
=
dt
∑
=i 1 =i 1
ri × m v
i i − ∑ dt
× m v=
i i
=i 1 dt
∑ ri × m v=
i i
dt
= L (t )
= vi
angular momentum
d L (t )
MO (t ) =
dt
For a system in equilibrium, M = O 0, ∀t :
d L (t )
= 0 ∀t L ( t ) = cnt CONSERVATION OF THE
dt ANGULAR MOMENTUM
41
Angular Momentum
in Continuum Mechanics
The angular momentum of a material volume Vt of a continuum
medium with mass M is:
L (t ) =
∫ r × v ( x, t ) d M =
∫ x × ρ ( x, t ) v ( x, t ) dV
M ≡x V
d M = ρ dV
42
Angular Momentum Balance Principle
The time-variation of the angular momentum of a material volume
with respect to a fixed point is equal to the resultant moment with
respect to this fixed point.
d L (t ) d
= ∫ r × ρ v dV = M O ( t )
dt dt Vt ≡V ≡x
M O ( t ) =∫ r × ρ b dV + ∫ r × t dS
V ∂V
torque due to
surface forces
43
Global Form of the
Angular Momentum Balance Principle
The global form of the angular momentum balance principle:
d
∫V (r × ρ b) dV + ∂∫V (r × t) dS
= ∫
dt Vt ≡V
(r × ρ v ) dV
= ∫ ( r × σT ) ⋅ ∇ dV
V
44
Global Form of the
Angular Momentum Balance Principle
Applying Reynolds Lemma to the right-hand term of the global
form equation:
Reynold's
Lemma
d d ↓ d
∫ r × ρ v dV
= ∫ ρ ( r × v ) dV = ∫ρ ( r × v ) dV
=
dt Vt ≡V dt Vt ≡V V
dt
dr =0 dv dv
= ∫ ρ × v + r × dV = ∫V r × ρ dt dV
dt dt
V
=v
Then, the global form of the balance principle is rewritten:
dv
∫ r × ( ρ b + ∇ ⋅σ ) + ijkσ jk eˆ i dV=
V
∫ r×ρ
V
dt
dV
45
Local Form of the
Angular Momentum Balance Principle
Rearranging the equation:
=0 (Cauchy’s Eq.)
dv
∫
∆V ⊂V
r ×
∇ ⋅ σ + ρ b − ρ
dt
+ m = 0
dV
∫
∆V ⊂V
= 0 ∀∆V ⊂ V , ∀t
m( x, t )dV
Localizing
m(x=
, t) 0 mi ijkσ=
= jk 0 ; i, j , k ∈ {1, 2,3} ; ∀x ∈ Vt , ∀t
= i 1 ⇒ 123 σ 23 + 132 σ 32= 0 ⇒ σ 23= σ 32
σ 11 σ 12 σ 13
=1 =−1
i 2 ⇒ 231 σ 31 + 213 σ 13= 0 ⇒ σ 31= σ 13
= σ ≡ σ 12 σ 22 σ 23
=1 =−1 σ 13 σ 23 σ 33
=i 3 ⇒ 312 σ 12 + 321 σ 21= 0 ⇒ σ 12= σ 21
=1 =−1
= σ ( x, t ) σ T ( x, t ) ∀x ∈ Vt , ∀t
SYMMETRY OF THE CAUCHY’S STRESS TENSOR
46
5.9. Mechanical Energy Balance
Ch.5. Balance Principles
47
Power
Power, W (t ) , is the work performed in the system per unit of
time.
48
External Mechanical Power
The external mechanical power is the work done by the body
forces and surface forces per unit of time.
In spatial form it is defined as:
Pe ( t ) = ∫ ρ b ⋅ v dV + ∫ t ⋅ v dS
V ∂V
dr
ρ b ⋅ dV =⋅ ρ b v dV
dt
=v
dr
t ⋅ dS =t ⋅ v dS
dt
=v
Mechanical Energy Balance
Using t= n ⋅ σ and the Divergence Theorem, the traction
contribution reads,
Divergence
Theorem
↓
∫∂V
=
t ⋅ v=
dS ∫∂V n ⋅ (σ ⋅ v ) dS V
= ∫ ( ∇ ⋅ σ ) ⋅ v + σ : ∇v dV
∫ ∇ ⋅ (σ ⋅ v ) dV V
n⋅σ =l
Taking into account the identity: l= d +
w spatial velocity
=0 skew gradient tensor
σ=
:l σ :d +σ :w symmetric
So, ∫
∂V
t ⋅ v dS= ∫ ( ∇ ⋅ σ ) ⋅ v dV + ∫ σ : d dV
V V
50
dv
∇ ⋅σ + ρ b =
ρ
Mechanical Energy Balance dt
dv
=∫ ( ∇ ⋅ σ + ρ b ) ⋅ v dV + ∫ σ : d dV =∫ ρ ⋅ v dV + ∫ σ : d dV
V dt
V V V
= ρ dv d 1 d 1
dt = ρ ( v= ⋅ v ) ρ ( v )
2
dt 2 dt 2 v = v
Reynold's
Lemma
d 1 2 ↓ d 1 2
Pe ( t=) ∫ ρ ( v )dV + ∫ σ : d dV = ∫ ρ ( v )dV + ∫ σ : d dV
V
dt 2 V
dt V 2 V
51
Mechanical Energy Balance. Theorem of
the expended power. Stress power
d 1 2
Pe ( t ) = ∫ ρ b ⋅ v dV + ∫ t ⋅ v dS = ∫ ρ v dV + ∫ σ : d dV
V ∂V dt Vt ≡V 2 V
d
Pe ( t )
= K ( t ) + Pσ Theorem of the expended
dt mechanical power
REMARK
The stress power is the mechanical power entering the system which is not spent
in changing the kinetic energy. It can be interpreted as the work by unit of time
done by the stress in the deformation process of the medium.
A rigid solid will produce zero stress power ( d = 0 ) .
52
External Thermal Power
The external thermal power is incoming heat in the continuum
medium per unit of time.
The incoming heat can be due to:
Non-convective heat transfer across the
volume’s surface.
incoming heat
− ∫ q(x, t ) ⋅ n dS =
unit of time
∂Vheat conduction
flux vector
53
External Thermal Power
The external thermal power is incoming heat in the continuum medium
per unit of time.
In spatial form it is defined as:
Qe (=
t) ∫ ρ r dV − ∫ q ⋅ n dS
= ∫ ( ρ r − ∇ ⋅ q) dV
V
∂V
V
= ∫∂V n⋅q dS
= ∫V (∇⋅q) dV
where:
q ( x, t ) is the non-convective heat flux vector per unit of spatial surface
r ( x, t ) is the internal heat source rate per unit of mass.
54
Total Power
The total power entering the continuous medium is:
d 1 2
=
Pe + Qe ∫
dt Vt ≡V 2
ρ v dV + ∫ σ : d dV + ∫ ρ r dV − ∫ q ⋅ n dS
V V ∂V
55
5.10. Energy Balance
Ch.5. Balance Principles
56
Thermodynamic Concepts
A thermodynamic system is a macroscopic region of the continuous
medium, always formed by the same collection of continuous matter
(material volume). It can be:
ISOLATED SYSTEM OPEN SYSTEM Thermodynamic space
MATTER
HEAT
57
Thermodynamic Concepts
A thermodynamic process is the energetic development of a
thermodynamic system which undergoes successive thermodynamic states,
changing from an initial state to a final state
→ Trajectory in the thermodynamic space.
If the final state coincides with the initial state, it is a closed cycle process.
2. There exists a function U ( t ) named the internal energy of the system, such that:
It is an extensive property, so it can be defined in terms of a specific internal energy (or
internal energy per unit of mass) u ( x, t ):
U ( t ) := ∫ ρ u dV REMARK
d E and d Kare exact differentials,
V
The variation of the total energy of the system is:
U d E − d K.
therefore, so is d=
d d d Then, the internal energy is a
= E (t ) K (t ) + U (t )
dt dt dt state function.
60
Global Form of the
Internal Energy Balance
Introducing the expression for the total power into the first
postulate: =K
d d 1 2
= E (t ) ∫ ρ v dV + ∫ σ : d dV + ∫ ρ r dV − ∫ q ⋅ n dS
dt dt Vt ≡V 2 V V ∂V
du
⇒ =∫
∆V ⊂V
ρ
dt
dV ∫
∆V ⊂V
σ : d dV + ∫
∆V ⊂V
ρ r dV − ∫
∆V ⊂V
∇ ⋅ q dV ∀∆V ⊂ V ∀t
62
Second Law of Thermodynamics
The total energy is balanced in all thermodynamics processes
following:
dE dK dU
Pe ( t ) + Qe ( t ) = = +
dt dt dt
In an isolated system (no work can enter or exit the system)
dE dU dK
Pe ( t ) + Qe ( t ) = =0 + =
0
dt dt dt
63
Second Law of Thermodynamics
The concept of energy in the first law does not account for
the observation that natural processes have a preferred
direction of progress. For example:
S (t ) S ( ) (t ) + S ( (t )
e)
=
i
dS dS ( ) dS ( )
i e
= +
dt dt dt
Entropy generated by interaction with the outside medium:
S ( ) = ∫ ρ s( ) ( x, t ) dV
i i
S ( e ) = ∫ ρ s( e ) ( x, t ) dV
V
68
Second Law of Thermodynamics
dS ( )
e
If one establishes, r q
=Γ e =∫ ρ dV − ∫ ⋅ n dS
dt V
θ ∂V
θ
dS (i ) dS dS ( e ) dS r q
= − = − ∫ ρ dV − ∫ ⋅ n dS ≥ 0 ∀∆V ⊂ V ∀t
dt dt dt dt ∆V ⊂V θ ∂∆V ⊂V
θ
69
Local Spatial Form of the
Second Law of Thermodynamics
The previous eq. can be rewritten as:
d (i ) d r q
= ∫
dt ∆Vt ⊂Vt ≡V
ρ s dV ∫
dt ∆Vt ⊂Vt ≡V
ρ s dV − ∫ ρ
∆V ⊂V θ
dV − ∫
∂∆V ⊂V
θ
⋅ n dS ≥0
∀∆V ⊂ V ∀t
ds (i ) ds r q
∫ =
∆V ⊂V
ρ
dt
dV ∫
∆V ⊂V
ρ
dt
dV − ∫ ρ dV − ∫ ∇ ⋅ dV ≥ 0 ∀∆V ⊂ V ∀t
∆V ⊂V θ ∆V ⊂V θ
Then, the local spatial form of the second law of thermodynamics is:
ds (i ) ds r q Local (spatial) form of the 2nd
ρ = ρ − ρ − ∇ ⋅ ≥ 0 ∀x ∈ V , ∀t Law of Thermodynamics
dt dt θ θ (Clausius-Duhem inequality)
= reversible process
> irreversible process
70
Local Spatial Form of the
Second Law of Thermodynamics
Considering that, q 1 1 REMARK
∇ ⋅ ( )= ∇ ⋅ q − 2 q ⋅ ∇θ
θ θ θ (Stronger postulate)
Internally generated entropy can
The Clausius-Duhem inequality can be written as (i )
be generated locally, slocal , or by
= s( ) = s
i
(i )
thermal conduction, scond , and
ds (i ) ds r 1 1 both must be non-negative.
= − + ∇ ⋅q − q ⋅ ∇θ ≥ 0
dt dt θ ρθ ρθ 2
(i ) (i )
= slocal = scond
CLAUSIUS-PLANCK HEAT FLOW
r 1
s − + ∇ ⋅ q ≥0 −
1
q ⋅ ∇θ ≥ 0 INEQUALITY
θ ρθ INEQUALITY ρθ
2
ρθ silocal
=: ρθ s − ρ r + ∇ ⋅ q ≥ 0
yields,
ρθ s + (σ : d − ρ u ) ≥ 0 − ρ ( u − θ s ) + σ : d ≥ 0
Clausius-Planck Inequality
in terms of the
specific internal energy
72
5.11. Governing Equations
Ch.5. Balance Principles
74
Governing Equations in Spatial Form
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 Continuity Equation.
1 eqn.
ρ u σ : d + ρ r − ∇ ⋅ q
= Energy Balance.
1 eqn.
First Law of Thermodynamics.
75
Governing Equations in Spatial Form
The fundamental governing equations involve the following variables:
ρ density 1 variable
v velocity vector field 3 variables
σ Cauchy’s stress tensor field 9 variables
u specific internal energy 1 variable
q heat flux per unit of surface vector field 3 variables
σ = σ ( v, θ , ζ ) Thermo-Mechanical
6 eqns.
Constitutive Equations.
s = s ( v, θ , ζ ) Entropy
Constitutive Equation. 1 eqn.
u = f ( ρ , v, θ , ζ ) Heat
State Equations. (1+p) eqns.
Fi ( ρ ,θ=
, ζ ) 0 i ∈ {1, 2,..., p} Kinetic
(19+p) PDE +
set of new thermodynamic
{ }
variables:ζ = ζ 1 , ζ 2 ,..., ζ p .
(19+p) unknowns
REMARK 1 REMARK 2
The strain tensor is not considered an unknown as they These equations are
can be obtained through the motion equations, i.e., ε = ε ( v ). specific to each material.
77
The Coupled
Thermo-Mechanical Problem
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 Continuity Mass Equation.
1 eqn.
Energy Balance.
1 eqn.
First Law of Thermodynamics.
79
The Uncoupled
Thermo-Mechanical Problem
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 Continuity Mass Equation.
1 eqn.
Energy Balance.
1 eqn.
First Law of Thermodynamics.
Thermal
problem
Second Law of Thermodynamics.
2 restrictions.
Clausius-Planck Inequality.
80
The Uncoupled
Thermo-Mechanical Problem
Then, the variables involved in the mechanical problem are:
ρ density 1 variable
Mechanical v velocity vector field 3 variables
variables
σ Cauchy’s stress tensor field 6 variables
u specific internal energy 1 variable
q heat flux per unit of surface vector field 3 variables
Thermal
variables
θ absolute temperature 1 variable
s specific entropy 1 variable
81
Chapter 5
Balance Principles
rs
ee
s gin
5.1 Introduction
t d le En
r
Continuum Mechanics is based on a series of general postulates or principles
ba
ge ro or
eS m
that are assumed to always be valid, regardless of the type of material and the
ci
f
range of displacements or deformations. Among these are the so-called balance
ra
C d P cs
principles:
b
a
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
ee
s gin
Figure 5.1: Convective transport in the continuous medium.
t d le En
r
ba
ge ro or
eS m
ci
f
Definition 5.1. The convective flux (or mass transport flux) of a
ra
C d P cs
generic property A through a control surface S is the amount of A
b
a
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].
ee
s gin
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,
r
these particles will have followed a pathline dx = v dt, such that at the instant of
ba
ge ro or
eS m
time t + dt they will occupy a new position in space. Taking now into account all
ci
f
ra
the particles that have crossed dS in the time interval [t, t + dt], these will occupy
C d P cs
b
a
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
Finally, the amount of A crossing dS in the time interval [t, t + dt] is calculated
tin
.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.
convective flux
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
A ≡V and Ψ = lead to ΦS = v · n dS = volume flow rate .
ρ
rs
S
ee
b) If the property A is the mass, then Ψ is the mass per unit of mass, that
s gin
is, the unit. Therefore,
t d le En
A ≡ M and Ψ = 1 lead to ΦS = ρ v · n dS .
r
ba
ge ro or
eS m
S
ci
f
ra
c) If the property A is the linear momentum (= mass × velocity), then Ψ
C d P cs
b
a
is the linear momentum per unit of mass, that is, the velocity. Therefore,
i
an an n
y ha
A ≡ m v and Ψ = v lead to ΦS = ρ v (v · n) dS .
le
liv or ec
S
(Note that in this case Ψ and the convective flux ΦS are vectors).
M
.A
d) If the property A is the kinetic energy then Ψ is the kinetic energy per
m
X Th
er
1 1 1
tin
.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
ee
s gin
Figure 5.4: Net outflow through a closed control surface.
t d le En
r
ba
ge ro or
eS m
ci
f
Remark 5.2. The convective flux of any property through a material
ra
C d P cs
b
surface is always null. Indeed, the convective flux of any property is
a
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 + Δt, the following
s gin
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.
r
ba
ge ro or
eS m
2) A material volume that at time t occupies the spatial volume Vt ≡ V is
ci
f
considered and, thus, the volume occupies different positions in space
ra
C d P cs
b
along time.
a
i
an an n
Different values of the amount Q (t + Δt) are obtained for each case, and com-
y ha
puting the difference between the amounts Q (t + Δt) and Q (t) when Δt → 0
le
yields
liv or ec
Q (t + Δt) − Q (t)
Q (t) = lim , (5.8)
M
.A
Δt→0 Δt
resulting in two different definitions of the time derivative, which lead to the
m
er
.O
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 ∂
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.
s gin
The amount Q of the generic property A in the control volume V at times t
t d le En
and t + Δt is, respectively,
r
ba
ge ro or
eS m
Q (t) = μ (x,t) dV and Q (t + Δt) = μ (x,t + Δt) dV . (5.9)
ci
f
ra
C d P cs
V V
b
a
i
Using (5.9) in addition to the concept of time derivative of Q (t) results in3
an an n
y ha
∂ 1
le
Q (t) = μ (x,t) dV = lim Q (t + Δt) − Q (t) =
liv or ec
∂t Δt→0 Δt
⎛ ⎞
M
V
.A
1 ⎝
= lim μ (x,t + Δt) dV − μ (x,t) dV ⎠ =
m
Δt→0 Δt
d
uu
V V
e
X Th
er
= lim dV = dV ,
Δt→0 Δt ∂t
on
.O
V V
∂ 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.
s gin
t d le En
The content Q of the generic property A in the material volume Vt at times t and
r
t+Δt is, respectively,
ba
ge ro or
eS m
ci
Q (t) = μ (x,t) dV f
Q (t + Δt) = μ (x,t + Δt) dV .
ra
and (5.12)
C d P cs
b
a
Vt Vt+Δt
i
an an n
y ha
d Q (t + Δt) − Q (t)
M
.A
V ≡V
Vt ⎛ t
⎞ (5.13)
d
uu
e
1 ⎝
= lim μ (x,t + Δt) dV − μ (x,t) dV ⎠ .
X Th
er
Δt→0 Δt
tin
Vt+Δt 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 + Δt,
⎧
⎪
⎪ xt = ϕ (X,t) → (dx1 dx2 dx3 )t = |F (X,t)| (dX1 dX2 dX3 ) ,
⎪
⎪
⎪
⎨
dVt dV0
⎪
⎪ xt+Δt = ϕ (X,t + Δt) → (dx1 dx2 dx3 )t+Δt = |F (X,t + Δt)| (dX1 dX2 dX3 ) ,
⎪
⎪
⎪
⎩
dVt+Δt dV0
(5.14)
4 Note that the integration domains are now different at times t and t + Δt.
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.
ee
s gin
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
r
μ̄ (X,t + Δt)
ba
ge ro or
eS m
d 1
ci
μ (x,t) dV = lim μ (x (X,t + Δt) ,t + Δt) |F (X,t + Δt)| dV0
f
ra
dt Δt→0 Δt
C d P cs
b
a
Vt V0
− μ (x (X,t) ,t) |F (X,t)| dV0 =
i
an an n
y ha
V0
μ̄ (X,t)
le
liv or ec
.A
Δt→0 Δt
V0
∂ d
m
∂t
e
dt
d
X Th
er
tin
= 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|
μ (x,t) dV = μ |F| dV0 = |F| + μ dV0 =
dt dt dt dt
Vt V0 V0
|F| ∇ · v
dμ dμ
= + μ∇ · v |F| dV0 = + μ∇ · v dV ,
dt
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 ,
d d dμ
μ (x,t) dV
not
= μ (x,t) dV = + μ∇ · v dV . (5.17)
dt dt dt
Vt Vt ≡V Vt ≡V V
rs
dt
Vt ≡V V
ee
∇ · (μv) (5.18)
s gin
∂μ ∂
= dV + ∇ · (μv) dV = μ dV + ∇ · (μv) dV ,
∂t ∂t
t d le En
V V V V
r
where the expression of the local derivative (5.11) has been taken into account.
ba
ge ro or
eS m
Then, (5.18) produces the expression of the material derivative of a volume in-
ci
f
ra
tegral.
C d P cs
b
a
i
an an n
d ∂
μ (x,t) dV = μ dV + ∇ · (μv) dV
le
liv or ec
dt ∂t (5.19)
Vt ≡V V V
M
.A
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
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.
s gin
5.4 Conservation of Mass. Mass continuity Equation
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Definition 5.4. Principle of conservation of mass. The mass of a
C d P cs
b
a
continuous medium (and, therefore, the mass of any material vol-
i
an an n
le
liv or ec
.A
in space Vt and Vt+Δt , respectively (see Figure 5.7). Consider also the spatial
m
description of the density, ρ (x,t). The mass enclosed by the material volume V
d
er
tin
.O
Vt Vt+Δt
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
which must be satisfied for Vt and, also, for any partial material volume ΔVt ⊂ Vt
ee
that could be considered. In particular, it must be satisfied for each of the ele-
s gin
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
r
ba
ge ro or
eS m
dρ dρ (x,t)
ci
+ ρ∇ · v dV = + ρ (x,t) ∇ · v (x,t) dV (x,t) = 0
f
ra
dt dt
C d P cs
dV (x,t)
b
a
∀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
(5.24)
d
uu
dρ
e
+ ρ∇ · v = 0 dV ∀x ∈ Vt , ∀t
X Th
dt
er
tin
which constitutes the so-called mass continuity equation. Replacing the expres-
on
.O
(5.24) results in
©
∂ρ ∂ρ
+ v · ∇ρ + ρ∇ · v = 0 =⇒ + ∇ · (ρv) = 0 , (5.25)
∂t
∂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 ,
s gin
dρ dρ 1 d |F|
t d le En
+ ρ∇ · v dV = +ρ dV =
dt dt |F| dt
r
Vt
Vt
1 d
ba
ge ro or
d |F|
eS m
1 dρ
= |F| +ρ dV = ρ |F| dV =
ci
|F| dt dt f |F| dt
ra
C d P cs
Vt
Vt
|F| dV0
b
a
d
i
an an n
ρ |F|
y ha
dt
∂
le
= ρ (X,t) |F (X,t)| dV0 ∀ΔV0 ⊂ V0 , ∀t ,
liv or ec
∂t
M
.A
V0
(5.27)
m
where the integration domain is now the volume in the reference configura-
d
tion, V0 . Given that (5.27) must be satisfied for each and every part ΔV0 of V0 , a
uu
e
er
tin
∂
ρ (X,t) |F (X,t)| = 0 ∀X ∈ V0 , ∀t
on
.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
Q (t) = ρΨ dV . (5.30)
ee
Vt ≡V
s gin
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
material derivative of a volume integral (with μ = ρΨ ) results in
r
ba
ge ro or
eS m
ci
d (ρΨ )
d
f
ra
Q (t) = ρΨ dV = + ρΨ ∇ · v dV . (5.31)
C d P cs
b
dt dt
a
Vt ≡V μ
i
V
an an n
y ha
grouping terms and introducing the mass continuity equation (5.24) yields
M
.A
d dΨ dρ
ρΨ dV = ρ +Ψ + ρΨ ∇ · v dV =
dt dt dt
m
Vt ≡V V
d
uu
dΨ dρ (5.32)
e
= ρ +Ψ + ρ∇ · v dV =⇒
X Th
er
dt dt
tin
V
=0 (mass continuity eqn.)
on
.O
C
Reynolds’ Lemma
©
d
ρΨ dV = ρ
dΨ
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.
s gin
t d le En
The variation of the amount of property A in the material volume Vt , which in-
r
stantaneously coincides at time t with the control volume V (Vt ≡ V ), is given by
ba
ge ro or
eS m
expression (5.19) of the material derivative of a volume integral (with μ = ρΨ )
ci
f
ra
and by (5.11),
C d P cs
b
a
i
d ∂ (ρΨ )
an an n
ρΨ dV = dV + ∇ · (ρΨ v) dV . (5.35)
y ha
dt ∂t
Vt ≡V
le
V V
liv or ec
.A
(5.35) results in
m
Reynolds’
uu
e
d Lemma dΨ ∂ (ρΨ )
X Th
ρΨ dV = ρ dV = dV + ∇ · (ρΨ v) dV =
er
tin
dt dt ∂t
Vt ≡V V V V
on
.O
Divergence
Theorem ∂ (ρΨ )
= dV + ρΨ v · n dS ,
C
∂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.
∇ · 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),
s gin
dΨ ∂ (ρΨ )
t d le En
ρ dV = dV + ∇ · (ρΨ v) dV ∀ΔV ⊂ V =⇒
dt ∂t
r
V V V (5.38)
ba
ge ro or
eS m
dΨ ∂ (ρΨ )
ci
ρ = + ∇ · (ρΨ v) ∀x ∈ V =⇒
∂t f
ra
dt
C d P cs
b
a
i
an an n
(5.39)
∂ (ρΨ )
le
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
ee
s gin
t d le En
Figure 5.9: An arbitrary control volume used in the definition of the global form of the
r
general balance equation.
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
Consider an arbitrary control volume V (see Figure 5.9). Then, the variation
i
per unit of time of property A in volume V will be due to
an an n
y ha
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
unit of time
.O
V
C
amount of A exiting through ∂V per convective flux
©
ρΨ v · n dS = ,
unit of time
∂V
amount of A exiting through ∂V per non-convective flux
jA · n dS = ,
unit of time
∂V
(5.41)
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
s gin
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
r
∂
ba
ge ro or
ρΨ dV = ρkA dV − ∇ · (ρΨ v) dV − ∇ · jA dV =⇒
eS m
ci
∂t
f
ra
V V V V
(5.43)
C d P cs
b
a
∂
(ρΨ ) + ∇ · (ρΨ v) dV = (ρkA − ∇ · jA ) dV ∀ΔV ⊂ V
i
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
∂
e
dΨ
(ρΨ ) + ∇ · (ρΨ v) = ρ = ρkA − ∇ · jA
X Th
er
∂t
tin
dt
(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 .
dt
s gin
t d le En
r
ba
ge ro or
eS m
Example 5.2 – Particularize the local spatial form of the general balance
ci
f
equation for the case in which property A is associated with the mass.
ra
C d P cs
b
a
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 = m i ai = m i = (mi vi ) (5.45)
s gin
dt dt
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
r
is equal to the variation of the linear momentum of the particle.
ba
ge ro or
eS m
Applying now Newton’s second law to the discrete system formed by n par-
ci
ticles results in
f
ra
C d P cs
b
a
n n n
dvi d n
d P (t)
R (t) = ∑ fi = ∑ mi ai = ∑ mi ∑ mi vi
i
= = . (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-
vation of mass (dmi /dt = 0) has been used. Equation (5.46) expresses that the
m
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
d P (t) n
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-
s gin
uous medium is equal to the variation per unit of time of its linear
momentum,
d P (t)
t d le En
d
R (t) = = ρ v dV .
dt dt
r
ba
Vt
ge ro or
eS m
ci
f
ra
C d P cs
b
a
The resultant of all the forces acting on the continuous medium defined above
i
an an n
le
liv or ec
R (t) = ρb dV + t dS . (5.48)
M
.A
V ∂V
m
body surface
d
forces forces
uu
e
X Th
.O
C
d (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
ee
s gin
t d le En
Figure 5.11: Forces acting on a material volume of the continuous medium.
r
ba
ge ro or
eS m
ci
5.7.2 Local Form of the Balance of Linear Momentum f
ra
C d P cs
b
a
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
ρv dV = ρb dV + n · σ dS = ρ dV ⎪
⎪
⎪
dt
dt ⎪
⎪
M
.A
Vt ≡V V ∂V V ≡V ⎬
t t
Divergence =⇒ (5.50)
m
Theorem ⎪
⎪
⎪
d
n · σ dS = ∇ · σ dV ⎪
⎪
uu
⎭
e
X Th
∂V V
er
tin
dv
=⇒ (∇ · σ + ρb) dV + ρ ∀ΔV ⊂ V
on
.O
dV (5.51)
dt
C
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)
s gin
dt
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
r
ba
ge ro or
eS m
⎫
ci
n n n
dvi ⎪
MO (t) = ∑ ri × fi = ∑ ri × mi ai = ∑ ri × mif ⎪
ra
⎪
dt ⎪
C d P cs
⎪
⎪
b
a
i=1 i=1 i=1 ⎪
⎪
⎬
i
an an n
n n n
d dri dvi
∑ × = ∑ dt ×m + ∑ × =⇒
y ha
r m v v r m
dt ⎪
i i i i i i i
dt i=1 i=1
⎪
⎪
le
i=1
⎪
⎪
liv or ec
vi ⎪
⎪
⎪
⎭
M
.A
(5.54)
=0
m
d n
dL (t)
uu
=⇒ MO (t) = ∑ ri × mi vi =
e
X Th
dt dt
er
tin
i=1
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
s gin
is defined as
L = r × v d M = r × ρ v dV (5.55)
t d le En
M ρ dV V
r
ba
ge ro or
eS m
and the continuous version of the postulate of balance of angular momentum is
ci
f
ra
obtained as follows.
C d P cs
b
a
i
an an n
y ha
.A
dL (t) d
d
MO (t) = = r × ρ v dV
uu
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)
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)
r × ρ v dV = r × ρ b dV + r × t dS
dt
s gin
Vt ≡V V ∂V
t d le En
r
ba
ge ro or
5.8.2 Local Spatial Form of the Balance of Angular Momentum
eS m
ci
f
ra
The procedure followed to obtain the local spatial form of the balance equation
C d P cs
b
a
is detailed below.
i
an an n
le
liv or ec
d d d
r × ρv dV = ρ (r × v) dV = ρ (r × v) dV =
M
.A
dt dt dt
Vt ≡V Vt ≡V V
dr
m
dv dv (5.58)
= ρ ×v dV + ρ r × dV = r × ρ dV ,
d
uu
dt dt dt
e
V
V V
X Th
er
v
tin
=0
on
.O
r × t dS = r × n · σ dS = [r] × [n · σ ]T dS =
∂V n·σ ∂V ∂V
(5.59)
Divergence
Theorem
= (r × σ ) · n dS
T
= r × σ T · ∇ 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
218 C HAPTER 5. BALANCE P RINCIPLES
where the component r × σ T · ∇ i is computed as
symb
∂ ∂
r×σT ·∇ = ei jk x j σrk = ei jk x j σrk =
i
∂ xr ∂ xr
σkr
T
(5.60)
∂xj ∂ σrk
= ei jk σrk + ei jk x j = ei jk σ jk + [r × ∇ · σ ]i i ∈ {1, 2, 3} .
∂ xr ∂ xr
[r × ∇ · σ ]i mi
δ jr
rs
ee
Introducing now (5.60) in (5.59) produces
s gin
r × t dS = m dV + (r × ∇ · σ ) dV
t d le En
∂V V V (5.61)
r
ba
ge ro or
eS m
mi = ei jk σ jk i, j, k ∈ {1, 2, 3}
ci
f
ra
C d P cs
b
a
i
and, finally, replacing (5.58) and (5.61) in (5.57) yields
an an n
y ha
dv
r×ρ dV = r × ρb dV + m dV + (r × ∇ · σ ) dV .
le
(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
e
X Th
dv
er
r × ∇ · σ + ρb − ρ dV + m dV = 0
tin
dt
V
V
on
.O
=0 (5.63)
C
=⇒ m dV = 0 ∀ΔV ⊂ 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
s gin
translating into the symmetry of the Cauchy stress tensor17 .
t d le En
Local spatial form of the
r
principle of balance of angular momentum
ba
ge ro or
(5.66)
eS m
ci
σ = σT
f
ra
C d P cs
b
a
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
ee
s gin
Definition 5.8. The mechanical power entering the continuous
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.
r
ba
ge ro or
eS m
ci
f
Consider the continuous medium shown in Figure 5.14 is subjected to the ac-
ra
C d P cs
tion of body forces, characterized by the vector of body forces b (x,t), and of
b
a
surface forces, characterized by the traction vector t (x,t). The expression of the
i
an an n
le
liv or ec
Pe = ρ b · v dV + σ · v) dS .
t · v dS = ρ b · v dV + n · (σ (5.68)
M
.A
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
d+w =0
s gin
Replacing (5.70) in (5.69) yields
t d le En
σ · v) dS =
n · (σ (∇ · σ ) · v dV + σ : d dV . (5.71)
r
ba
ge ro or
eS m
∂V V V
ci
f
ra
C d P cs
Introducing (5.71) in (5.68), the mechanical power entering the continuous
b
a
results in19
i
medium
an an n
y ha
Pe = ρ b · v dV + t · v dS = ρ b · v dV + (∇ · σ ) · v dV + σ : d dV =
le
liv or ec
V ∂V V V V
dv
M
.A
= (∇ · σ + ρ b) · v dV + σ : d dV = ρ · v dV + σ : d dV =
dt
V
m
V V V
d
d 1 d 1 2
ρ v · v dV + σ : d dV = ρ v dV + σ : d dV .
uu
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
Pe = ρ b · v dV + t · v dS = ρv dV + σ : d dV
dt 2 (5.73)
V ∂V Vt ≡V V
mechanical
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
v·v =
1 dv 1 dv dv
·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,
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 d
ee
not 1 2 1 2
kinetic energy = K = ρ v dV =⇒ = ρ v dV .
2 dt dt 2
s gin
V V
b) creating stress power,
t d le En
de f
stress power = σ : d dV .
r
ba
ge ro or
V
eS m
ci
f
ra
C d P cs
b
a
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
rs
q · n dS =
unit of time
ee
∂V (5.74)
s gin
amount of heat entering the medium
− q · n dS =
unit of time
t d le En
∂V
r
ba
ge ro or
eS m
ci
f
Remark 5.9. A typical example of non-convective flux is heat trans-
ra
C d P cs
fer by conduction phenomena. Heat conduction is governed by
b
a
i
Fourier’s Law, which provides the vector of heat flux by (non-
an an n
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
224 C HAPTER 5. BALANCE P RINCIPLES
rs
ee
Figure 5.16: Internal heat sources.
s gin
t d le En
• Internal heat sources
r
ba
ge ro or
eS m
Heat can be generated (or absorbed) in the interior of the continuous
ci
f
medium due to certain phenomena (chemical reactions, etc.). Consider a
ra
C d P cs
scalar function r (x,t) that describes in spatial form the heat generated by
b
a
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
heat generated by the internal sources
M
.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
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-
s gin
dynamic variables. In a hyperspace (thermodynamic space) defined by the
thermodynamic variables μi i ∈ {1, 2, ... , n} (see Figure 5.17), a thermo-
t d le En
dynamic state is represented by a point.
• Thermodynamic process: the energetic development of a thermodynamic
r
ba
ge ro or
eS m
system that undergoes successive thermodynamic states, changing from an
ci
f
initial state at time tA to a final state at time tB (it is a path or continuous
ra
C d P cs
segment in the thermodynamic space), see Figure 5.18.
b
a
• Closed cycle: A thermodynamic process in which the final thermodynamic
i
an an n
state coincides with the initial thermodynamic state (all the thermodynamic
y ha
• 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
e
X Th
er
.O
(5.78)
C
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 5.19: Closed cycle.
ee
s gin
Consider also a given thermodynamic process A → B in the space of the
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
r
path (thermodynamic process) A → B are known by means of
ba
ge ro or
eS m
ci
f
B
ra
C d P cs
φB = φA + δφ .
b
a
(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
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
⎪
δ φ = f1 (μ1 , ... , μn ) dμ1 + . . . + fn (μ1 , ... , μn ) dμn ⎪ ⎬
ee
∂ fi (μ1 , ... , μn ) ∂ f j (μ1 , ... , μn ) ⇔ δ φ = dφ .
s gin
= ∀i, j ∈ {1, ... , n} ⎪
⎪
⎭
∂ μj ∂ μi
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
If the differential form (5.78) is an exact differential, (5.79) results in
b
a
i
an an n
B
y ha
B
φ B = φA + dφ = φA + Δ 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,
©
A " A
δ φ = dφ = Δ φ = 0 .
A
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 =⇒ =
f 2 ≡ μ1 ∂ f2 ∂ μ 2 ∂ μ1
=1
∂ μ1
rs
Then, δ φ is not an exact differential (see Remark 5.10) and φ is not a state
ee
function.
s gin
t d le En
5.10.2 First Law of Thermodynamics
r
ba
ge ro or
Experience shows that the mechanical power (5.73) is not an exact differential
eS m
ci
and, therefore, the mechanical work performed by the system in a closed cycle
f
ra
C d P cs
is not null. The same happens with the thermal power (5.76).
b
a
"
i
an an n
δ φ1 = Pe dt =⇒ Pe dt = 0
y ha
" (5.81)
le
δ φ2 = Qe dt =⇒ Qe dt = 0
liv or ec
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
t
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)
rs
Variation of Mechanical Thermal
total energy work work
ee
s gin
2) There exists another state function U , named internal energy of the system,
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
r
ba
ge ro or
eS m
ci
U=f ρu dV . (5.84)
ra
C d P cs
b
a
V
i
an an n
b) The variation of the total energy of the system E is equal to the sum of
y ha
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
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
= + = ρv dV + σ : d dV + ρr dV − q · n dS
dt dt dt dt 2
rs
V V V ∂V
ee
dK dU
s gin
dt dt
(5.86)
t d le En
r
ba
ge ro or
eS m
ci
Global form of the internal energy balance f
ra
C d P cs
b
a
dU d (5.87)
i
= ρu dV = σ : d dV + ρr dV − q · n dS
an an n
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
X Th
er
V
on
.O
− a variation per
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
ee
s gin
The first law of thermodynamics leads to a balance equation that must be ful-
filled for all the physical processes that take place in reality,
t d le En
dE dU dK
Pe + Qe = = + .
r
(5.90)
ba
ge ro or
dt dt dt
eS m
ci
f
ra
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).
a
i
an an n
Therefore, the energy balance equation (5.90), established by the first law of
thermodynamics, imposes that any variation of internal energy d U /dt must be
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
is, it does not establish any restriction that indicates if an imaginary and arbitrary
e
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.
s gin
Figure 5.23
t d le En
Consider now the following two processes:
r
1) At a certain instant of time the brake acts, the rotation speed of the wheel ω
ba
ge ro or
eS m
decreases and, thus, so does its kinetic energy (d K < 0). On the other hand,
ci
f
due to the friction between the brake and the wheel, heat is generated and
ra
C d P cs
there is an increase of the internal energy (d U > 0). Experience shows
b
a
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
d
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
s gin
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.
r
ba
ge ro or
eS m
ci
f
Definition 5.11. A thermodynamic process A → B is a reversible
ra
C d P cs
b
a
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).
A thermodynamic process A → B is an irreversible process when
le
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
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 (5.92)
= ρs dV ≥ ρ dV − · n dS
s gin
dt dt θ θ
Vt ≡V V ∂V
t d le En
r
is satisfied, where:
ba
ge ro or
eS m
− The sign = corresponds to reversible processes.
ci
f
− The sign > corresponds to irreversible processes.
ra
C d P cs
b
a
− 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
ee
s gin
Similarly to (5.93), the new source term r/θ and non-convective flux vec-
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
r
ba
ge ro or
(heat/unit of temperature) entering V r q
eS m
= ρ dV − · n dS .
ci
(5.94)
unit of time f θ θ
ra
C d P cs
V ∂V
b
a
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
≥ ρ dV − · n dS
on
.O
dt θ θ (5.95)
∂V
C
V
©
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
rs
= ρ dV − · n dS (5.97)
dt θ θ
ee
V ∂V
s gin
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
⎛ ⎞
r
ba
ge ro or
dS(i) dS dS(e) dS ⎝
eS m
r q
= − = − ρ dV − · n dS⎠ ≥ 0 .
ci
(5.98)
θ θ f
ra
dt dt dt dt
C d P cs
V ∂V
b
a
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
ee
S(i) = ρs(i) dV and S(e) = ρs(e) dV (5.101)
s gin
V V
t d le En
dS dS(i) dS(e)
S = S(i) + S(e) = =⇒
+ (5.102)
r
dt dt dt
ba
ge ro or
eS m
and introducing Reynolds’ Lemma (5.33) in (5.102) yields
ci
f
ra
C d P cs
dS(i) ds(i)
b
a
d
= ρs(i) dV = ρ dV ,
i
an an n
dt dt dt
y ha
Vt ≡V V
(5.103)
dS(e) ds(e)
le
d
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
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
ρs(i) dV = ρs dV − ρ dV − · n dS ≥ 0
dt dt θ θ
Vt ≡V Vt ≡V ∂V
rs
V
ee
Applying Reynolds’ Lemma (5.33) (on the first and second integral of the left-
s gin
hand term in (5.106)) and the Divergence Theorem (on the last term) yields
ds(i) q
t d le En
ds r
ρ dV = ρ dV − ρ dV − ∇ · dV ≥ 0 ∀ΔV ⊂ V
dt dt θ θ
r
ba
ge ro or
V V V V
eS m
(5.107)
ci
f
ra
and localizing in (5.107), the local form of the second law of thermodynamics
C d P cs
b
a
or Clausius-Duhem equation is obtained.
i
an an n
y ha
(Clausius-Duhem inequality)
(5.108)
ds r q
M
.A
ds(i)
ρ = ρ − ρ −∇· ≥0 ∀x ∈ V, ∀t
dt dt θ θ
m
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
.(i) .(i)
slocal scond
rs
ee
s gin
Local internal generation of entropy
(Clausius-Planck inequality)
(5.111)
t d le En
.(i) . r 1
slocal = s − + ∇·q ≥ 0
θ ρθ
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
Internal generation of entropy by heat conduction
i
an an n
.(i) 1
y ha
(5.112)
scond = − q · ∇θ ≥ 0
ρθ 2
le
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.
s gin
t d le En
r
ba
ge ro or
Remark 5.16. In the context of Fourier’s Law of heat conduction,
eS m
q = −K ∇θ (see Remark 5.9), expression (5.112) can be written as
ci
f
ra
!
C d P cs
b
a
q · ∇θ ≤ 0
i
=⇒ −K |∇θ |2 ≤ 0 =⇒ K ≥ 0
an an n
q = −K∇θ
y ha
le
liv or ec
.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)
s gin
Differentiating (5.116) with respect to time results in
t d le En
. . . . . . . .
r
ψ = u − sθ − sθ =⇒ u − θ s = ψ + sθ (5.117)
ba
ge ro or
eS m
ci
f
and, replacing (5.117) in (5.115), yields the Clausius-Planck equation in terms
ra
C d P cs
b
of the Helmholtz free energy,
a
i
. .
an an n
.(i) . .
ρθ slocal = −ρ (u − θ s) + σ : d = −ρ ψ + sθ + σ : d ≥ 0 .
y ha
(5.118)
le
liv or ec
M
.A
d
uu
e
X Th
.
er
For the infinitesimal strain case, d = ε (see Chapter 2, Remark 2.22), and re-
tin
.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 ⎪
⎪
+ρ =0 ⎭
ee
dt ∂ xi
s gin
t d le En
2) Balance of linear momentum. Cauchy’s equation.
⎫
r
⎪
ba
ge ro or
dv ⎪
eS m
∇ · σ + ρb = ρ ⎪
⎬
ci
dt
f
→ 3 equations
ra
(5.122)
C d P cs
∂ σ ji ⎪
⎪
b
dvi
a
+ ρbi = ρ i ∈ {1, 2, 3} ⎪
⎭
i
∂xj
an an n
dt
y ha
le
liv or ec
.A
!
σ = σT
m
→ 3 equations
d
(5.123)
uu
er
tin
on
.O
⎪
⎪
©
du
ρ = σ : d + (ρr − ∇ · q) ⎪ ⎬
dt
→ 1 equation (5.124)
du ∂ qi ⎪⎪
⎪
ρ = σi j di j + ρr − ⎭
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
rs
ρθ ∂ xi
ee
These add up to a total of 8 partial differential equations (PDEs) and two re-
s gin
strictions. Counting the number of unknowns that intervene in these equations
results in25 ⎫
t d le En
ρ → 1 unknown ⎪ ⎪
⎪
⎪
⎪
r
v → 3 unknowns ⎪ ⎪
ba
ge ro or
⎪
⎪
eS m
⎪
σ → 9 unknowns ⎪
ci
⎪
⎬ f
ra
C d P cs
u → 1 unknown
b
19 unknowns
a
⎪
⎪
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)
i ∈ {1, 2, 3} ⎪
⎭
©
qi = −K
∂ 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.
s gin
8) Thermodynamic equations of state.
t d le En
⎫
⎪
Caloric ⎪
r
u = g (ρ, ε (v) , θ , μ ) ⎬
ba
ge ro or
eqn. of state
eS m
→ (1 + p) eqns.
ci
⎪
Kinetic f ⎪
i ∈ {1, 2, ... , p} ⎭
ra
Fi (ρ, θ , μ ) = 0
C d P cs
b
eqns. of state
a
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
s gin
set of equations are considered 27
t d le En
⎪
⎪
dρ ⎪
(1 eqn) ⎪
Mass continuity ⎪
+ ρ∇ · v = 0 ⎪
r
equation: ⎪
⎪
ba
dt
ge ro or
⎪
eS m
⎪
⎬
ci
dv f (3 eqn) ⎪ → 10 equations ,
ra
∇ · σ + ρb = ρ
C d P cs
Cauchy’s equation:
⎪
b
a
dt ⎪
⎪
⎪
i
⎪
an an n
⎪
⎪
σ , ε (v)) = 0
fi (σ ⎪
y ha
(6 eqn) ⎪
Mechanical
⎭
constitutive equations: i ∈ {1, ... , 6}
le
liv or ec
(5.129)
M
.A
ρ (x,t) → 1 unknown ⎪
d
⎬
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.
27 For 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
ee
s gin
t d le En
Solution
r
a) The statement is false because a material surface is always constituted by
ba
ge ro or
eS m
the same particles and, therefore, cannot be crossed by any particle throughout
ci
f
its motion. For this reason, the mass flux across a material surface is always null,
ra
C d P cs
independently of the motion being stationary or not.
b
a
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 ⎪ ⎬
∂t
m
=⇒ ∇ · (ρv) = 0 .
d
⎪
⎪
uu
∂ρ ⎪
e
Stationary flux =⇒ =0 ⎪
⎭
X Th
er
∂t
tin
on
C
∇ · (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
ee
s gin
t d le En
r
Solution
ba
ge ro or
eS m
ci
Taking into account the Reynolds Transport Theorem (5.39) and that the prob-
f
ra
lem is in steady-state regime, the forces acting on the fluid are
C d P cs
b
a
d ∂
i
∑ Fext/ f = ρv dV = (ρv) dV + ρv (n · v) dS = ρv (n · v) dS .
an an n
dt V V ∂t
y ha
∂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
∑ ext/ f
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
s gin
a pipe from end A (with cross-section SA ) to end B (with cross-section SB < SA ).
The pipe is secured at point O by a rigid element P − O.
t d le En
Determine:
r
ba
ge ro or
eS m
a) The entry and exit velocities vA and vB in terms of the flow rate.
ci
f
b) The values of the angle θ that maximize and minimize the reaction force
ra
C d P cs
b
F at O, and the corresponding values of said reaction force.
a
i
c) The values of the angle θ that maximize and minimize the reaction mo-
an an n
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
∇·v = 0 ⇐⇒ ∇ · v dV = 0 ∀V . [1]
V
The adequate integration volume must now be defined. To this aim, a control
volume such that its boundary is a closed surface must be found (S = ∂V ) to be
rs
able to apply the Divergence Theorem,
ee
s gin
∇ · v dV = n · v dS ∀V [2]
t d le En
V ∂V
r
where n is the outward unit normal vector in the boundary of the volume V .
ba
ge ro or
eS m
Then, by means of [1] and [2], the conclusion is reached that the net outflow
ci
across the contour of the control volume is null, f
ra
C d P cs
b
a
i
n · v dS = 0 ∀V .
an an n
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
SA and SB since n · v = 0 on the walls of the pipe, that is, n and v are perpendic-
er
tin
.O
n · v dS = n · v dS + n · v dS = (−eA ) · vA eA dS + eB · vB eB dS = 0
C
∂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
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 ∂
ρv dV = ρv dV + ρv (n · v) dS . [5]
ee
dt ∂t
V V ∂V
s gin
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
r
ba
ge ro or
of the pipe. Therefore, according to [4] and [5],
eS m
ci
f
ra
C d P cs
b
a
R= ρv (n · v) dS + ρv (n · v) dS =
i
an an n
SA SB
y ha
le
= ρvA eA (−eA · vA eA ) dS + ρvB eB (eB · vB eB ) dS
liv or ec
M
.A
SA SB
m
er
tin
.O
1 1
R = −ρQ − eA + eB .
2
C
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
R= ρb dV + t dS = t dS = t dS + t dS + t dS =
V ∂V ∂V SA SB Slat
= 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.
rs
R p/ f = R − pA eA dS − pB (−eB ) dS
ee
SA SB
s gin
R p/ f = −ρv2A SA eA + ρv2B SB eB − pA SA eA + pB SB eB
t d le En
r
R p/ f = − ρv2A + pA SA eA − ρv2B + pB SB eB [7]
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
Introducing [3], R p/ f can be expressed in terms of Q,
a
i
an an n
2
Q2
y ha
Q
R p/ f =− ρ + pA SA 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).
s gin
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
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
The vector (−eA ) is fixed and does not depend on θ but eB does vary with θ . The
e
X Th
er
scalars FA and FB are constant values. Therefore, the maximum and minimum
tin
one another, respectively. That is, when the vectors (−eA ) and eB are parallel to
.O
each other. Taking into account [3] and [8], the maximum and minimum values
C
− Minimum value of F
π
θ=
2
1 1
|F|min = ρQ 2
− + 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
Mliq = r × ρb dV + r × t dS = r × ρv dV , [9]
dt
s gin
V ∂V 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
r
ba
ge ro or
Transport Theorem (5.39), yields
eS m
ci
f
ra
d ∂
C d P cs
r × ρv dV = r × ρv dV + (r × ρv) (n · v) dS .
b
[10]
a
dt ∂t
i
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
m
[11]
uu
e
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
Mliq = r × ρb dV + r × t dS = r × t dS =
s gin
V ∂V ∂V
t d le En
= r × t dS + r × t dS − r × t dS =
r
SA SB Slat
ba
ge ro or
eS m
= 0+ R pB ez dS + M p/ f = R pB SB ez + M p/ f ,
ci
f
ra
C d P cs
SB
b
a
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
©
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
W= ρv dV + σ : d dV [14]
dt 2
V V
rs
This is proven as follows.
ee
s gin
1 T
σ : d = −p 1 : d = −p Tr (d) = −p Tr l +l =
t d le En
2
r
⎡ ⎤
ba
ge ro or
∂ v x ∂ vx ∂ v x
eS m
ci
⎢ ⎥
f
⎢ ∂x ∂y ∂z ⎥
ra
⎢ ⎥
C d P cs
⎢
b ⎥
a
= −p Tr (ll ) = −p Tr ⎢ ∂ vy ∂ vy ∂ vy ⎥ =
i
⎢ ∂x ∂y ∂z ⎥
an an n
⎢ ⎥
y ha
⎣ ∂v ∂v ∂v ⎦
z z z
le
liv or ec
∂x ∂y ∂z
M
.A
∂ vx ∂ v y ∂ v z
= −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
©
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
W= ρv (n · v) dS + ρv (n · v) dS =
2 2
SA SB
1 2 1 2 1 1
= 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
ee
s gin
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
r
ba
ge ro or
hinge B and is maintained in vertical position by the action of force F.
eS m
ci
f
ra
C d P cs
b
a
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.
s gin
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,
r
ba
ge ro or
per unit of width (normal to the plane of the figure), is W . Consider a steady-
eS m
ci
state motion, driven by the velocity of the piston V and the internal uniform
f
ra
pressure P1 . The external uniform pressure is P2 .
C d P cs
b
a
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.
ee
s gin
5.4 – A perfect and incompressible fluid flows through the pipe junction shown
in the figure below. The junction is held in place by a rigid element O − D.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
CH.6. LINEAR ELASTICITY
Multimedia Course on Continuum Mechanics
Overview
Hypothesis of the Linear Elasticity Theory Lecture 1
Linear Elastic Constitutive Equation
Lecture 2
Generalized Hooke’s Law
Elastic Potential Lecture 3
Isotropic Linear Elasticity
Isotropic Constitutive Elastic Constants Tensor
Lecture 4
Lamé Parameters
Isotropic Linear Elastic Constitutive Equation
Young’s Modulus and Poisson’s Ratio
Lecture 5
Inverse Isotropic Linear Elastic Constitutive Equation
Spherical and Deviator Parts of Hooke’s Law Lecture 6
Limits in Elastic Properties Lecture 7
2
Overview (cont’d)
The Linear Elastic Problem
Governing Equations Lecture 8
Boundary Conditions
The Quasi-Static Problem Lecture 9
Solution
Displacement Formulation
Lecture 10
Stress Formulation Lecture 11
Saint-Venant’s Principle
Uniqueness of the solution
Lecture 12
Linear Thermoelasticity
Hypothesis of the Linear Elasticity Theory Lecture 13
Linear Thermoelastic Constitutive Equation
Inverse Constitutive Equation
Lecture 14
Thermal Stress and Strain
3
Overview (cont’d)
Thermal Analogies
Solution to the linear thermoelastic problem Lecture 15
1st Thermal Analogy
2nd Thermal Analogy Lecture 16 Lecture 17
4
6.1 Hypothesis of the Linear
Elasticity Theory
Ch.6. Linear Elasticity
5
Hypothesis of the Linear Elastic Model
The simplifying hypothesis of the Theory of Linear Elasticity are:
8
Hypothesis of the Linear Elastic Model
1. ‘Infinitesimal strains and deformation’ framework
the displacements are infinitesimal:
material and spatial configurations or coordinates are the same
≈0
x= X + u x≈X
material and spatial descriptions of a property & material and spatial
differential operators are the same:
x=X γ ( x, t ) =
γ ( X, t ) =
Γ ( X, t ) =
Γ ( x, t )
∂ (•) ∂ (•)
= ∇ ( • )= ∇ ( • )
∂X ∂x
∂x
the deformation gradient =
F ≈1 F ≈ 1, so the current spatial
∂X
density is approximated by the density at the reference configuration.
ρ 0 ρt F ≈ ρt
=
Thus, density is not an unknown variable in linear elastic problems.
9
Hypothesis of the Linear Elastic Model
1. ‘Infinitesimal strains and deformation’ framework
the displacement gradients are infinitesimal:
The strain tensors in material and spatial configurations collapse
into the infinitesimal strain tensor.
E ( X, t ) ≈ e ( x, t ) =
ε ( x, t )
10
Hypothesis of the Linear Elastic Model
2. Existence of an unstrained and unstressed reference state
It is assumed that there exists a reference unstrained and unstressed
neutral state, such that,
ε 0 ( x ) ε=
= ( x, t 0 ) 0
σ 0 ( x ) σ=
= ( x, t0 ) 0
The reference state is usually assumed to correspond to the reference
configuration.
11
Hypothesis of the Linear Elastic Model
3. Isothermal and adiabatic (=isentropic) processes
In an isothermal process the temperature remains constant.
θ ( x, t ) ≡ θ ( x, t0 ) ≡ θ 0 ( x ) ∀x
Q=
e ∫ρ
V
0 r dV − ∫ q ⋅ n dS=
∂V
0 ∀∆V ⊂ V
internal
heat conduction REMARK
from the exterior An isentropic process is an
sources
ρ0 r − ∇ ⋅ q= 0 ∀x ∀t idealized thermodynamic
process that is adiabatic,
isothermal and reversible.
12
6.2 Linear Elastic Constitutive
Equation
Ch.6. Linear Elasticity
13
Hooke’s Law
R. Hooke observed in 1660 that, for relatively small deformations
of an object, the displacement or size of the deformation is
directly proportional to the deforming force or load.
σ
F= k ∆l F ∆l
=E E
A l
F ε
σ= E ε
14
Generalized Hooke’s Law
This proportionality is generalized for the multi-dimensional case
in the Theory of Linear Elasticity.
σ ( x, t ) =
C(x) : ε ( x, t )
Generalized
= σ Cijkl ε kl i, j ∈ {1, 2,3} Hooke’s Law
ij
It constitutes the constitutive equation of a linear elastic material.
The 4th order tensor is the constitutive elastic constants tensor:
Has 34=81 components.
Has the following symmetries, reducing the tensor to 21 independent
components:
minor Cijkl = C jikl REMARK
symmetries The current stress at a point
Cijkl = Cijlk depends only on the current strain
at the point, and not on the past
major Cijkl = C klij
symmetries history of strain states at the point.
15
Elastic Potential
The internal energy balance equation for the (adiabatic) linear
elastic model is
global form stress power heat transfer rate
d d ρ0 u
dt V∫
ρ 0=
u dV ∫V = dt
dV ∫ σ : d dV + ∫ ( ρ 0 r − ∇ ⋅ q ) dV
V
= ε V
internal energy infinitesimal
strains REMARK
(Vt ≡ V ∀t )
local form The rate of strain tensor is related to the
d material derivative of the material strain
( ρ 0=
u ) σ : ε + ρ r − ∇ ⋅ q tensor through: E = FT ⋅ d ⋅ F
dt
= ε and F = 1 .
In this case, E
Where:
u is the specific internal energy (energy per unit mass).
r is the specific heat generated by the internal sources.
q is the heat conduction flux vector per unit surface.
16
Elastic Potential
The stress power per unit of volume is an exact differential of the
internal energy density, û , or internal energy per unit of volume:
d duˆ ( x, t ) REMARK
( ρ 0 u )= = uˆ= σ : ε
dt dt The symmetries of the
uˆ constitutive elastic
Operating in indicial notation: constants tensor are used:
duˆ 1 Cijkl = C jikl
σ
=
:
ε
=
ε σ
=
ε C ε = ( ε ij Cijkl εkl + ε ij Cijkl εkl ) = minor
dt C:ε
ij
ij ij ijkl kl
2 symmetries Cijkl = Cijlk
Cijkl ε kl i ↔k
j ↔l
major
= ( ε ij Cijkl εkl + εkl C klij εij ) = ( ε ij Cijkl εkl + εij Cijkl ε kl ) =
1 1 Cijkl = C klij
symmetries
2 2
Cijkl d (ε C ε )
dt ij ijkl kl
1d
= εij Cijkl εkl duˆ d 1
= (ε : C : ε)
2 dt dt dt 2
ε: C:ε
17
Elastic Potential
duˆ 1 d
= σ=: ε (ε : C : ε)
dt 2 dt
Consequences:
1. Consider the time derivative of the internal energy in the whole
volume:
d d d ˆ
∫V dt ( ) ( ) U (t )
dt V∫ ∫ σ : ε dV
ˆ
u =
x , t dV ˆ
u =
x , t dV = stress
dt V power
REMARK
The stress power, in elastic materials is an
exact differential of the internal energy Uˆ .
Then, in elastic processes, we can talk of the
elastic energy Uˆ(t ) .
18
Elastic Potential
duˆ 1 d
= σ=: ε (ε : C : ε)
dt 2 dt
Consequences:
2. Integrating the time derivative of the internal energy density,
1
=uˆ ( x, t ) ε ( x, t ) : C : ε ( x, t ) + a ( x )
2
and assuming that the density of the internal energy vanishes at the
neutral reference state, uˆ ( =
x, t0 ) 0 ∀x:
=0 1 1
uˆ ( ε )
1
ε ( x, t0 ) : C : ε ( x, t0 ) + a ( x ) =a (x) =0 ∀x = =ε
:C:ε σ (ε ) : ε
2 2 σ 2
Due to thermodynamic reasons the internal energy is assumed
always positive 1
=uˆ ( ε ) ε : C : ε > 0 ∀ε ≠ 0
2
19
σ ( x, t ) = C( x ) : ε ( x, t )
Elastic Potential
The internal energy density defines a potential for the stress
tensor, and is thus, named elastic potential. The stress tensor can
be computed as
= σ=
T
σ
∂uˆ ( ε( x, t )) ∂ 1 1 1 1 ∂û ( ε )
= ( ε : C : ε) = C : ε + ε : C= ( σ + σ)= σ σ=
∂ε ∂ε 2 2 2 2 ∂ε
=σ
∂σ (ε) ∂ uˆ ( ε ) ∂ ( C : ε )
2
∂ 2uˆ ( ε )
= = = C Cijkl =
∂ε ∂ε ⊗ ∂ε ∂ε ∂εij ∂ε kl
20
6.3 Isotropic Linear Elasticity
Ch.6. Linear Elasticity
21
Isotropic Constitutive Elastic Constants
Tensor
An isotropic elastic material must have the same elastic properties
(contained in C ) in all directions.
All the components of C must be independent of the orientation of the
chosen (Cartesian) system C must be a (mathematically) isotropic
tensor.
C= λ 1 ⊗ 1 + 2 µ I
Cijkl =λδ ijδ kl + µ (δ ik δ jl + δ ilδ jk ) i, j , k , l ∈ {1, 2,3}
Where:
1
[I ]ijkl
I is the 4th order unit tensor defined as= δ ik δ jl + δ ilδ jk
2
λ and µ are scalar constants known as Lamé‘s parameters or coefficients.
REMARK
The isotropy condition reduces the number of independent elastic constants from 21 to 2.
22
Isotropic Linear Elastic Constitutive
Equation
Introducing the isotropic constitutive elastic constants tensor
C= λ 1 ⊗ 1 + 2 µ I into the generalized Hooke’s Law σ = C : ε ,
in index notation:
σ ij Cijkl=
= ε kl ( λδ δ
ij+ µ (δ ik δ jl + δ ilδ jk ) =
kl )
ε kl
= εij
1 1
= λδ ij δ kl ε kl + 2 µ δ ik δ jl ε kl + δ ilδ jk = ε kl λTr (ε)δ ij + 2 µεij
2 2
= εll = Tr (ε ) = ε=ji ε ij
1 1
= = εij + εij εij
2 2
23
Elastic Potential
If the constitutive equation is,
=σ λ Tr ( ε ) 1 + 2 µ ε Isotropic linear elastic
constitutive equation.
λδ ij ε ll + 2µ ε ij i, j ∈ {1, 2,3}
σ ij = Hooke’s Law
24
Inversion of the Constitutive Equation
1. ε is isolated from the expression derived for Hooke’s Law
=σ λ Tr ( ε ) 1 + 2 µ ε =
ε
1
2µ
( σ − λ Tr ( ε ) 1 )
25
Inverse Isotropic Linear Elastic
λ 1
Tr ( σ ) 1 +
Constitutive Equation ε=−
2 µ ( 3λ + 2 µ ) 2µ
σ
In engineering notation: εy =
1
E
(σ y − ν ( σ x + σ z ) ) 1
γ xz = τ xz
G
εz =
1
E
(
σ z − ν (σ x + σ y ) ) 1
γ yz = τ yz
G
26
Young’s Modulus and Poisson’s Ratio
Young's modulus E is a measure of the stiffness of an elastic
material. It is given by the ratio of the uniaxial stress over the
uniaxial strain.
µ ( 3λ + 2µ )
E=
λ+µ
λ
ν=
2 (λ + µ )
27
Example
Consider an uniaxial traction test of an isotropic linear elastic material such
that:
σx > 0 E, ν
σ= σ= τ= τ= τ= 0 y
σx σx
y z xy xz yz
σx σx
x
Obtain the strains (in engineering notation) and comment on the results
obtained for a Poisson’s ratio of ν = 0 and ν = 0.5 .
28
εx =
1
(
σ x − ν (σ y + σ z ) ) γ xy =
1
τ xy
σx > 0 E G
Solution (σ y − ν ( σ x + σ z ) )
1 1
εy = γ xz = τ xz
σ=
y σ=z τ=
xy τ=
xz τ=
yz 0 E G
εz =
1
E
(
σ z − ν (σ x + σ y ) ) 1
γ yz = τ yz
G
For ν = 0 :
1
=εx = σx γ xy 0 1 There is no Poisson’s effect
E = σx
=εx
γ xy 0
ν
E and the transversal normal
εy =
− σx γ xz =
0 =ε y 0= γ xz 0 strains are zero.
E
=ε z 0= γ yz 0
ν
εz =
− σx γ yz =
0
E
For ν = 0.5 :
1 The volumetric deformation is
εx =
1
σx γ xy 0 =ε = σ γ 0
E
x
E
x xy
zero, tr ε = ε x + ε y + ε z = 0, the
0 material is incompressible
1
εy =
−
0.5
σx γ xz =
0 ε y =
− σx γ xz =
E 2E and the volume is preserved.
1
εz =
−
0.5
σx γ yz =
0 ε z =
− σx γ yz =
0
E 2 E
29
Spherical and deviatoric parts of
Hooke’s Law
The stress tensor can be split into a spherical, or volumetric, part
and a deviatoric part:
1
σ= m1
sph : σ= Tr (σ ) 1
3 σ σ m1 + σ ′
=
σ ′ dev
= = σ σ − σ m1
30
Spherical and deviatoric parts of
Hooke’s Law
Operating on the volumetric strain:
e = Tr ( ε )
ν 1 +ν
− Tr ( σ ) 1 +
ε= σ
E E
ν 1 +ν
− Tr ( σ ) Tr (1 ) +
e= Tr ( σ )
E E = 3σ m
=3
K : bulk modulus
3 (1 − 2ν ) E (volumetric strain modulus)
e= σm σm = e
E 3 (1 − 2ν ) def
2 E
K= λ+ µ=
3 3(1 − 2ν )
The spherical parts of the stress and strain tensor are directly
related: σ = K e
m
31
Spherical and Deviator Parts of
Hooke’s Law
ν 1 +ν
Introducing=
σ σ m 1 + σ ′ into ε= − Tr ( σ ) 1 + σ :
E E
ν 1 +ν
− Tr (σ m 1 + σ ′ ) 1 +
ε= (σ m 1 + σ ′ )
E E
ν ν =0 1 +ν 1 +ν 1 +ν 3ν 1 +ν
− σ mTr (1 ) 1 − Tr (σ ′ ) 1 + σ m
= 1+ σ′ = − σ
m 1 + σ′
E =3 E E E E E E
E
Taking into account that σm = e :
3 (1 − 2ν ) Comparing this
1 − 2ν 1 E 1 +ν 1 1 +ν with the expression
ε=
e 1 + σ ′ = e 1 + σ′
E 3 (1 − 2ν ) E 3 E =ε
1
e 1 + ε´
3
1 +ν 1 1 1 +ν
= = ε´= σ′
E 2µ 2G E
The deviatoric parts of the stress and strain tensor are related
component by component:
= σ ′ 2G ε′= σ ij′ 2Gε ij′ i, j ∈ {1, 2,3}
32
Spherical and deviatoric parts of
Hooke’s Law
The spherical and deviatoric parts of the strain tensor are
directly proportional to the spherical and deviatoric parts
(component by component) respectively, of the stress tensor:
σm = K e σ ij′ = 2G ε ij′
33
Elastic Potential
The internal energy density û ( ε ) defines a potential for the stress
tensor and is, thus, an elastic potential:
REMARK
1
uˆ ( ε ) = ε : C : ε ∂uˆ ( ε ) The constitutive elastic constants
=σ = :ε
2 ∂ε tensor C is positive definite due
to thermodynamic considerations.
Plotting û ( ε ) vs. ε :
There is a minimum for ε = 0 :
∂uˆ ( ε )
= (C : ε) =0
∂ε ε =0 ε =0
∂ 2uˆ ( ε ) ∂ 2uˆ ( ε )
= = C= C
∂
=
ε ⊗ ∂ ε ∂
ε 0=
ε ⊗ ∂ ε ε 0
ε =0
34
Elastic Potential
The elastic potential can be written as a function of the spherical
and deviatoric parts of the strain tensor:
ε : C : ε = (C : ε) : ε = σ : ε
1 1 1
uˆ ( ε )
= =ε : C :=
ε σ:ε λTr ( ε ) 1 + 2 µε : ε =
2 2 2
=σ
1 1
= e 1 + ε´
: e 1 + ε ´ =
1 = Tr ( ε ) = e 1 = e 2
3 3
= λTr ( ε ) 1 : ε + µε= :ε λTr ( ε ) + µ ε : ε
2
1 2 2
2 2 = e 1 : 1 + e1 : ε ´ + ε´= : ε´
9 3
3 Tr ( ε′ ) =0
K
1 1 1 2 1
uˆ ( ε ) = λ e 2 + µ e 2 + μ ε´: ε´= λ + µ e 2 + μ ε´: ε´ = e 2 + ε´: ε´
2 3 2 3 3
Elastic potential in terms of the
1
uˆ ( ε ) = K e + µ ε´: ε´≥ 0 spherical and deviatoric parts
2
2 of the strains.
35
Limits in the Elastic Properties
The derived expression must hold true for any deformation
process: 1
uˆ ( ε ) = : K e 2 + µ ε´ : ε´≥ 0
2
Consider now the following particular cases of isotropic linear
elastic material:
Pure spherical deformation process
1
ε (1) = e 1 1
3 uˆ (1)
= K e2 ≥ 0 K >0 bulk modulus
2
ε′(1) = 0
Pure deviatoric deformation process
ε ( 2 ) = ε′ Lamé’s second
=
uˆ( 2)
µ ε´: ε´≥ 0 µ >0
e( 2 ) = 0 parameter
REMARK
ε′=
: ε′ ε ij ε ij ≥ 0
36
Limits in the Elastic Properties
K and µ are related to E and ν through:
E E
=K >0 µ= G= >0
3 (1 − 2ν ) 2 (1 +ν )
REMARK
In rare cases, a material can
Poisson’s ratio has a non-negative value, have a negative Poisson’s ratio.
E Such materials are named
>0
2 (1 +ν ) E≥0 Young’s auxetic materials.
modulus
ν ≥0
Therefore,
E
>0
3 (1 − 2ν ) 0 ≤ν ≤
1
Poisson’s ratio
E≥0 2
37
6.4 The Linear Elastic Problem
Ch.6. Linear Elasticity
38
Introduction
The linear elastic solid is subjected to body forces and prescribed
tractions:
Initial actions:
b ( x, 0 )
t =0
t ( x, 0 )
b ( x, t )
Actions
through time: t ( x, t )
39
Governing Equations
The Linear Elastic Problem is governed by the equations:
1. Cauchy’s Equation of Motion.
Linear Momentum Balance Equation.
∂ 2 u ( x, t )
∇ ⋅ σ ( x, t ) + ρ 0 b ( x, t ) =
ρ0
∂ t2
2. Constitutive Equation.
Isotropic Linear Elastic Constitutive Equation. This is a PDE system of
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
= 15 eqns -15 unknowns:
u ( x,t ) 3 unknowns
3. Geometrical Equation. ε ( x,t ) 6 unknowns
Kinematic Compatibility. σ ( x,t ) 6 unknowns
ε ( x, t )= ∇ S u ( x, t )=
1 Which must be solved in
(u ⊗ ∇ + ∇ ⊗ u )
2 the R 3 × R +space.
40
Boundary Conditions
Boundary conditions in space
Affect the spatial arguments of the unknowns
Are applied on the boundary Γ of the solid,
which is divided into three parts:
Prescribed displacements on Γ u :
u(x, t ) = u* (x, t )
∀x ∈ Γu ∀t
=ui ( x, t ) ui ( x, t ) i ∈ {1, 2,3}
*
41
Boundary Conditions
42
Boundary Conditions
Boundary conditions in time. INTIAL CONDITIONS.
Affect the time argument of the unknowns.
Generally, they are the known values at t = 0 :
Initial displacements:
u ( x, 0=
) 0 ∀x ∈ V
Initial velocity:
∂u ( x, t ) not
= u (= x, 0 ) v 0 ( x ) ∀x ∈ V
∂t t =0
43
The Linear Elastic Problem
Find the displacements u ( x,t ) , strains ε ( x,t ) and stresses σ ( x,t )
such that
∂ 2 u ( x, t )
∇ ⋅ σ ( x, t ) + ρ 0 b ( x, t ) =
ρ0 Cauchy’s Equation of Motion
∂ t2
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
= Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= (u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary conditions in space
σ ⋅n
Γσ : t * =
u ( x, 0 ) = 0
Initial conditions (Boundary conditions in time)
u ( x, 0 ) = v 0
44
Actions and Responses
The linear elastic problem can be viewed as a system of actions or data
inserted into a mathematical model made up of the EDP’s and
boundary conditions , which gives a response (or solution) in
displacements, strains and stresses.
b ( x, t )
u ( x, t )
t* ( x, t ) Mathematical
model ε ( x, t )
u* ( x, t )
EDPs+BCs σ ( x, t )
v0 ( x )
not
not RESPONSES = R ( x,t )
ACTIONS = A ( x,t )
Generally, actions and responses depend on time. In these cases, the
problem is a dynamic problem, integrated in R3 × R + .
In certain cases, the integration space is reduced to R3 . The problem is
termed quasi-static.
45
The Quasi-Static Problem
A problem is said to be quasi-static if the acceleration term can
be considered to be negligible.
∂ 2 u ( x, t )
=a ≈0
∂t 2
46
The Quasi-Static Problem
Find the displacements u ( x,t ) , strains ε ( x,t ) and stresses σ ( x,t )
such that ∂ u ( x,t ) 2
ρ0 ≈0
∂ t2
∇ ⋅ σ ( x, t ) + ρ0b ( x, t ) =
0 Equilibrium Equation
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
= Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= (u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary Conditions in Space
σ ⋅n
Γσ : t * =
u ( x, 0 ) = 0
Initial Conditions
u ( x, 0 ) = v 0
47
The Quasi-Static Problem
The quasi-static linear elastic problem does not involve time
derivatives.
Now the time variable plays the role of a loading descriptor: it describes the
evolution of the actions.
b ( x, λ ) u ( x, λ )
Mathematical
t* ( x, λ ) model ε ( x, λ )
u* ( x, λ ) EDPs+BCs σ ( x, λ )
not not
ACTIONS = A ( x, λ ) RESPONSES = R ( x, λ )
48
Example
Consider the typical material strength problem where a cantilever beam is
subjected to a force F ( t ) at it’s tip.
For a quasi-static problem,
49
Solution of the Linear Elastic Problem
To solve the isotropic linear elastic problem posed, two approaches can
be used:
Displacement formulation - Navier Equations
Eliminate σ ( x,t ) and ε ( x,t ) from the general system of equations. This
generates a system of 3 eqns. for the 3 unknown components of u ( x,t ) .
Useful with displacement BCs.
Avoids compatibility equations.
Mostly used in 3D problems.
Basis of most of the numerical methods.
50
Displacement formulation
∂ 2 u ( x, t )
∇ ⋅ σ ( x, t ) + ρ 0 b ( x, t ) =
ρ0 Cauchy’s Equation of Motion
∂ t2
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
= Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= (u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary Conditions in Space
σ ⋅n
Γσ : t * =
u ( x, 0 ) = 0
Initial Conditions
u ( x, 0 ) = v 0
The aim is to reduce this system to a system with u ( x,t ) as the only unknowns.
Once these are obtained, ε ( x,t ) and σ ( x,t ) will be found through substitution.
51
Displacement formulation
Introduce the Constitutive Equation into Cauchy’s Equation of
motion:
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
=
∂ 2u
λ ∇ ⋅ [Tr (ε) 1] + 2µ∇ ⋅ ε + ρ0b =ρ0 2
∂ 2 u ( x, t ) ∂t
∇ ⋅ σ ( x, t ) + ρ 0 b ( x, t ) =
ρ0
∂ t2
Consider the following identities:
= ( ∇ ( ∇ ⋅u ) )i
∂ ∂ ∂uk ∂ ∂uk ∂
∇ ⋅ ( Tr ( ε) 1)=
i
∂x j
( )
ε11 δ ij=
∂x j ∂x δ ij=
∂x ∂x =
∂xi
( ∇ ⋅ u=)
k i k
= ∇⋅u
=∇ ( ∇ ⋅ u ) i i ∈ {1, 2,3}
∇ ⋅ (Tr ( ε ) 1 ) =
∇ ( ∇ ⋅u )
52
Displacement formulation
Introduce the Constitutive Equation into Cauchy’s Equation of
motion:
=σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
∂ 2u
λ ∇ ⋅ [Tr (ε) 1] + 2µ∇ ⋅ ε + ρ0b =ρ0 2
∂ 2 u ( x, t ) ∂t
∇ ⋅ σ ( x, t ) + ρ 0 b ( x, t ) =
ρ0
∂ t2
Consider the following identities:
( )
= ∇2 u = ( ∇ ( ∇ ⋅ u ) )i
i
∂ε ij
( ∇ ⋅ ε )i = =
∂ 1 ∂ui ∂u j 1 ∂ 2ui
+ = +
1 ∂
∂x j ∂x j 2 ∂x j ∂xi 2 ∂x j ∂x j 2 ∂xi
∂u j
1 2
=
∂x j 2 ( )
∇ u +
i
1 ∂
2 ∂xi
( ∇ ⋅ u )=
1 1 = ∇⋅u
= ∇ 2u + ∇ ( ∇ ⋅ u ) i ∈ {1, 2,3}
2 2 i
1 1
∇=
⋅ε ∇ (∇ ⋅ u) + ∇ 2u
2 2
53
Displacement formulation
Introduce the Constitutive Equation into Cauchy’s Equation of
Movement:
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
=
∂ 2u
λ ∇ ⋅ [Tr (ε) 1] + 2µ∇ ⋅ ε + ρ0b =ρ0 2
∂ 2 u ( x, t ) ∂t
∇ ⋅ σ ( x, t ) + ρ 0 b ( x, t ) =
ρ0
∂ t2
Replacing the identities:
∇ ⋅ (Tr ( ε ) 1 ) =
∇ ( ∇ ⋅u ) ∇=
⋅ε
1 1
∇ (∇ ⋅ u) + ∇ 2u
2 2
1 1 2 ∂ 2u 2nd order
Then, λ ∇ ( ∇ ⋅ u ) + 2 µ ∇ (∇ ⋅ u) + ∇ u + ρ0b =
ρ0 2 PDE system
2 2 ∂t
∂ 2u
( λ + µ ) ∇ ( ∇ ⋅ u ) + µ∇ u + ρ0b = ρ0 2
2
The Navier Equations ∂t
are obtained: ( λ + µ ) u j , ji + µ ui , jj + ρ
= 0 bi ρ0ui i ∈ {1, 2,3}
54
Displacement formulation
The boundary conditions are also rewritten in terms of u ( x,t ) :
σ ( x, t ) λTr ( ε ) 1 + 2 µ ε
= = ∇⋅u
=t* λ (Tr ( ε ) ) n + 2 µ ε ⋅ n
1
t = σ ⋅n
* = ∇ S u= (u ⊗ ∇ + ∇ ⊗ u )
2
t=
*
λ (∇ ⋅ u ) n + µ (u ⊗ ∇ + ∇ ⊗ u ) ⋅ n
The BCs are now:
u = u*
on Γu
=ui ui i ∈ {1, 2,3}
*
λ ( ∇ ⋅ u ) n + µ ( u ⊗ ∇ + ∇ ⊗ u ) ⋅ n = t* REMARK
on Γσ The initial conditions
λ uk .k ni + µ ( ui , j n j + u j ,i n j ) =
ti* i ∈ {1, 2,3}
remain the same.
55
Displacement formulation
Navier equations in a cylindrical coordinate system:
∂e 2G ∂ω z ∂ωθ ∂ 2 ur x = r cos θ
( λ + 2G ) − + 2G + ρ br = ρ 2
∂r r ∂θ ∂z ∂t x( r , θ , z ) ≡ y = r sin θ
1 ∂e ∂ωr ∂ω z ∂ 2uθ z = z
( λ + 2G ) − 2G + 2G + ρ bθ =ρ 2
r ∂θ ∂z ∂r ∂t
∂e 2G ∂ 2G ∂ωr ∂ 2u z
( λ + 2G ) − ( rωθ ) + + ρ bz =
ρ 2
∂z r ∂r r ∂θ ∂t
1 1 ∂u z ∂uθ
ωr = −Ωθ z = −
2 r ∂θ ∂z
1 ∂ ur ∂ u z dV = r dθ dr dz
Where: ωθ = −Ω zr = −
2 ∂z ∂r
1 1 ∂ ( ruθ ) 1 ∂ur 1 ∂ 1 ∂u ∂u
ω z = −Ω rθ =
2 r ∂r
−
r ∂θ
=
e ( rur ) + θ + z
r ∂r r ∂θ ∂z
56
Displacement formulation
Navier equations in a spherical coordinate system:
∂e 2G ∂ 2G ∂ωθ ∂ 2 ur
( λ + 2G ) −
∂r r sin θ ∂θ
(ωϕ sin θ ) + r sin θ ∂ϕ + ρ br =
ρ 2
∂t x = r sin θ cos ϕ
=x x ( r ,θ , ϕ ) =
≡ y r sin θ sin ϕ
( λ + 2G ) ∂e − 2G ∂ωr ∂ ∂u
2
r ∂θ
+
2G
r sin θ ∂ϕ r sin θ ∂r
( )
rωϕ sin θ + ρ bθ =
ρ 2θ
∂t
z = r cos θ
( λ + 2G ) ∂e − 2G ∂ rω + 2G ∂ωr + ρ b = ∂ 2uϕ
( θ) ϕ ρ 2 dV = r 2 sin θ dr dθ dϕ
r sin θ ∂ϕ r ∂r r ∂θ ∂t
1 1 ∂ 1 ∂uθ
ωr = −Ωθϕ =
2 r sin θ ∂θ
( uϕ sin θ ) −
r sin θ ∂φ
Where: 1 1 ∂ur 1 ∂ ( ruϕ )
ωθ = −Ωϕ r = −
2 r sin θ ∂φ r ∂r
11 ∂ 1 ∂u
ωϕ = −Ω rθ = ( ruθ ) − r
2 r ∂r r ∂θ
∂ 2 ∂ ∂
∂r ( r ur sin θ ) + ∂θ ( ruθ sin θ ) + ∂ϕ ( ruϕ )
1
=e
r sin θ
2
57
Stress formulation
Equilibrium Equation
∇ ⋅ σ ( x, t ) + ρ0b ( x, t ) =
0
(Quasi-static problem)
ν 1 +ν
ε ( x,t ) =
− Tr ( σ ) 1 + σ Inverse Constitutive Equation
E E
1
ε ( x, t )= ∇ S u ( x, t )= (u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary Conditions in Space
σ ⋅n
Γσ : t * =
The aim is to reduce this system to a system with σ ( x,t ) as the only unknowns.
Once these are obtained, ε ( x,t ) will be found through substitution and u ( x,t ) by
integrating the geometric equations.
REMARK
For the quasi-static problem, the time variable plays the role of a loading factor.
58
Stress formulation
Taking the geometric equation and, through successive derivations,
the displacements are eliminated:
∂ 2ε ij ∂ 2ε kl ∂ 2ε ik ∂ 2ε jl Compatibility Equations
+ − − = 0 i, j , k , l ∈ {1, 2,3} (seen in Ch.3.)
∂xk ∂xl ∂xi ∂x j ∂x j ∂xl ∂xi ∂xk
59
Stress formulation
The boundary conditions are:
Equilibrium Equations: ∇ ⋅ σ + ρ0b =0
This is a 1st order PDE system, so they can act as boundary conditions of
the (2nd order PDE system of the) Beltrami-Michell Equations
Prescribed stresses on : Γσ =
σ ⋅ n t* on Γσ
60
Stress formulation
Once the stress field is known, the strain field is found by substitution.
ν 1 +ν
− Tr ( σ ) 1 +
ε ( x,t ) = σ
E E
The calculation, after, of the displacement field requires that the
geometric equations be integrated with the prescribed displacements
on Γu: 1
ε=
( x) ( u ( x) ⊗ ∇ + ∇ ⊗ u ( x) ) x ∈ V
2
=
u ( x ) u* ( x ) ∀x ∈ Γu
REMARK
This need to integrate the second system is a considerable disadvantage with
respect to the displacement formulation when using numerical methods to solve
the lineal elastic problem.
61
Saint-Venant’s Principle
From A. E. H. Love's Treatise on the mathematical theory of elasticity:
“According to the principle, the strains that are produced in a body by the application,
to a small part of its surface, of a system of forces statically equivalent to zero force
and zero couple, are of negligible magnitude at distances which are large compared
with the linear dimensions of the part.” REMARK
This principle does not have a
Expressed in another way: rigorous mathematical proof.
“The difference between the stresses caused by statically equivalent load systems is
insignificant at distances greater than the largest dimension of the area over which
the loads are acting.”
u (I) ( x P , t ) ≈ u (II) ( x P , t )
ε (I) ( x P , t ) ≈ ε (II) ( x P , t ) ∀P | δ >>
σ (I) ( x P , t ) ≈ σ (II) ( x P , t )
62
Saint-Venant’s Principle
Saint Venant’s Principle is often used in strength of materials.
It is useful to introduce the concept of stress:
The exact solution of this problem is very complicated.
64
6.5 Linear Thermoelasticity
Ch.6. Linear Elasticity
65
Hypothesis of the Linear Thermo-elastic
Model
The simplifying hypothesis of the Theory of Linear Thermo-
elasticity are:
1. Infinitesimal strains and deformation framework
Both the displacements and their gradients are infinitesimal.
σ 0 ( x ) σ=
= ( x, t0 ) 0
3. Isentropic and adiabatic processes – no longer isothermal !!!
Isentropic: entropy of the system remains constant
Adiabatic: deformation occurs without heat transfer
66
Hypothesis of the Linear Thermo-Elastic
Model
3. (Hypothesis of isothermal process is removed)
The process is no longer isothermal so the temperature changes
throughout time: not
θ ( x, t ) ≠ θ ( x, 0 ) = θ0
∂θ ( x, t )
θ ( x, t )
= ≠0
∂t
We will assume the temperature field is known.
Q=e ∫ ρ r dV − ∫ q ⋅ n dS=
∂V
0 ∀∆V ⊂ V
internal V
heat conduction
sources from the exterior
ρ r − ∇ ⋅ q= 0 ∀x ∀t
67
Generalized Hooke’s Law
The Generalized Hooke’s Law becomes:
σ ( x, t ) C : ε ( x, t ) − β (θ=
= − θ 0 ) C : ε ( x, t ) − β ∆θ Generalized Hooke’s Law for
σ ij= Cijkl ε kl − βij (θ − θ 0 ) i, j ∈ {1, 2,3} linear thermoelastic problems
Where
is the elastic constitutive tensor.
θ ( x,t ) is the absolute temperature field.
θ 0 = θ ( x,t0 ) is the temperature at the reference state.
β is the tensor of thermal properties or constitutive thermal
constants tensor.
It is a positive semi-definite symmetric second-order tensor.
REMARK
A symmetric second-order tensor A is positive semi-definite
when zT·A·z > 0 for every non-zero column vector z.
68
Isotropic Constitutive Constants Tensors
An isotropic thermoelastic material must have the same elastic
and thermal properties in all directions:
must be a (mathematically) isotropic 4th order tensor:
C= λ 1 ⊗ 1 + 2 µ I
Cijkl =λδ ijδ kl + µ (δ ik δ jl + δ ilδ jk ) i, j , k .l ∈ {1, 2,3}
Where:
1
th =
I is the 4 order symmetric unit tensor defined as [ I ] δ ik δ jl + δ ilδ jk
ijkl
2
λ and µ are the Lamé parameters or coefficients.
69
Isotropic Linear Thermoelastic
Constitutive Equation
Introducing the isotropic constitutive constants tensors β = β 1 and
C= λ 1 ⊗ 1 + 2 µ I into the generalized Hooke’s Law, σ = C : ε − β (θ − θ 0 )
(in indicial notation)
σ=
ij Cijkl ε kl −βij (θ − θ 0=) ( λδ δ
ij kl )
+ µ (δ ik δ jl + δ ilδ jk ) ε kl − β (θ − θ 0 ) δ ij=
= ε ij
1 1
= λδ ij δ kl ε kl + 2 µ δ ik δ jl ε kl + δ ilδ jk ε kl − β (θ − θ 0 ) δ ij
=ε 2 = ε= ε
2 = ∆θ
ll ji ij
= ε ij
The resulting constitutive equation is,
σ λ Tr ( ε ) 1 + 2 µ ε − β ∆θ 1
= Isotropic linear thermoelastic
σ
=
ij λδ ε
ij ll + 2 µ ε ij − β ∆θ δ ij i , j ∈ {1, 2,3}
constitutive equation.
70
Inversion of the Constitutive Equation
1. ε is isolated from the Generalized Hooke’s Law for linear
thermoelastic problems:
=ε C −1 : σ + ∆θ C −1
:β
σ= C : ε − β ∆θ
α
2. The thermal expansion coefficients tensor α is defined as:
def
α = C −1 : β
71
Inverse Isotropic Linear Thermoelastic
Constitutive Equation
For the isotropic case:
−1 ν 1 +ν
C =− 1 ⊗ 1 + I
E E 1 − 2ν
→ α =
C −1
: ( β 1) = β1
ν 1 +ν
C −1 =
ijkl
−
E
δ δ
ij kl +
E
( δ δ
ik jl + δ δ
il jk ) i , j , k .l ∈ {1, 2,3}
E
ν 1 +ν
ε = − Tr ( σ ) 1 + σ + α ∆θ 1
Inverse isotropic linear thermo
E E
elastic constitutive equation.
ε = ν 1 +ν
− σ δ + σ ij + α ∆θ δ ij i, j ∈ {1, 2,3}
ij E
ll ij
E
73
Thermal Strain
Similarly, by comparing the inverse constitutive equations,
ν 1 +ν Inverse isotropic linear elastic
ε=− Tr ( σ ) 1 + σ
E E constitutive eq.
ν 1 +ν
ε=− Tr ( σ ) 1 + σ + α ∆θ 1 Inverse isotropic linear thermoelastic
E E nt
=ε = ε t
constitutive eq.
the decomposition is made:
=
ε ε nt + ε t
Where:
ε nt is the non-thermal strain: the strain produced if there is no
temperature increment.
ε t is the thermal strain: the “corrector” strain due to the
temperature increment.
74
Thermal Stress and Strain
The thermal components appear when thermal processes are considered.
NON-THERMAL THERMAL
TOTAL
COMPONENT COMPONENT
σ nt = C : ε σ t = ∆θ β
=
σ σ nt − σ t Isotropic material: Isotropic material:
= σ nt λTr (ε) 1 + 2 µ ε σ t= β ∆θ 1
ε nt = C −1 : σ ε t = ∆θ α
=
ε ε nt + ε t Isotropic material:
ν 1 +ν Isotropic material:
− Tr (σ ) 1 +
ε =
nt
σ ε t= α ∆θ 1
E E
These are the equations used
in FEM codes. =σ C=: ε nt C : ε − ε t
= −1
ε C= : σ nt C −1 : σ + σ t
75
Thermal Stress and Strain
REMARK 1
In thermoelastic problems, a state of zero strain in a body does not necessarily
imply zero stress.
ε=0 → σ nt =0
σ =−σ t =− β∆θ 1 ≠ 0
REMARK 2
In thermoelastic problems, a state of zero stress in a body does not necessarily
imply zero strain.
σ =0 → ε nt =0
ε= α θ1≠0
ε t =∆
76
6.6 Thermal Analogies
Ch.6. Linear Elasticity
77
Solution to the Linear Thermoelastic
Problem
To solve the isotropic linear thermoelastic problem posed thermal
analogies are used.
The thermoelastic problem is solved like an elastic problem and then,
the results are “corrected” to account for the temperature effects.
They use the same strategies and methodologies seen in solving
isotropic linear elastic problems:
Displacement Formulation - Navier Equations.
Stress Formulation - Beltrami-Michell Equations.
Two basic analogies for solving quasi-static isotropic linear
thermoelastic problems are presented:
1st thermal analogy – Duhamel-Neumann analogy.
2nd thermal analogy
78
1st Thermal Analogy
The governing eqns. of the quasi-static isotropic linear thermoelastic
problem are:
∇ ⋅ σ ( x, t ) + ρ0b ( x, t ) =
0 Equilibrium Equation
( x, t ) C : ε ( x, t ) − β ∆θ 1
σ= Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= ( u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary Conditions in Space
Γσ : t * =
σ ⋅n
79
1st Thermal Analogy
The actions and responses of the problem are:
not
(I )
( x, t )
not
ACTIONS = A RESPONSES = R ( ) ( x, t )
I
b ( x, t )
u ( x, t )
t * ( x, t ) Elastic model ε ( x, t )
u * ( x, t ) EDPs+BCs σ ( x, t )
∆θ ( x, t )
REMARK
∆θ ( x,t ) is known a priori, i.e., it is
independent of the mechanical response.
This is an uncoupled thermoelastic problem.
80
1st Thermal Analogy
To solve the problem following the methods used in linear elastic
problems, the thermal term must be removed.
The stress tensor is split into =
σ σ nt − σ t and replaced into the governing
equations:
Momentum equations
=σ σ nt − σ t ∇= ⋅ σ ∇ ⋅ σ nt − ∇ ⋅ σ=t
∇ ⋅ σ nt − ∇ ( β ∆θ )
β ∆θ 1
∇ ⋅ σ + ρ0 b =0
∇ ⋅ σ nt + ρ0 bˆ =
0
∇ ⋅ σ nt + ρ0 b − ∇ ( β ∆θ ) =
1
0
ρ 0 1
b − ∇ ( β ∆θ )
bˆ =
not ρ0
= bˆ
81
1st Thermal Analogy
Boundary equations:
σ σ nt − σ t
=
σ nt ⋅ n − σ t ⋅ n =t* σ nt ⋅ n = tˆ*
σ ⋅n =t *
Γσ :
σ nt ⋅ n = t* + σ t
⋅ n = t* + ( β∆θ ) n tˆ* = t* + ( β∆θ ) n
β∆θ 1⋅n tˆ*
ANALOGOUS PROBLEM – A linear elastic problem can be solved as:
1
∇ ⋅ σ nt + ρ0bˆ = b − ∇ ( β∆θ )
0 with bˆ = Equilibrium Equation
ρ0
=
σ nt
C= : ε λTr (ε) 1 + 2 µ ε Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= (u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u = u*
Boundary Conditions in Space
Γσ : σ nt ⋅ n =tˆ* with tˆ* = t* + β∆θ n
82
1st Thermal Analogy
The actions and responses of the ANALOGOUS NON-THERMAL PROBLEM are:
not
( II )
( x, t )
not
ACTIONS = A RESPONSES = R ( ) ( x, t )
II
bˆ ( x, t ) u ( x, t )
Elastic model ε ( x, t )
tˆ* ( x, t )
EDPs+BCs σ nt ( x, t )
u* ( x, t )
RESPONSES
u u 0 0 def
R ( ) (x, t ) − R ( ) (x, t ) = ε − ε = 0 = 0 =R ( )
I II III
σ σ nt σ − σ nt − β∆θ 1
t
= −σ
84
1st Thermal Analogy
u ( x, t ) u ( x, t ) u=0
R ( I ) ( x, t ) ε ( x, t ) R(
II )
( x, t ) ε ( x, t )
R(
III )
( x, t ) ε=0
σ ( x, t ) σ nt ( x, t ) σ=−σ t =− ( β∆θ ) 1
86
2nd Thermal Analogy
The governing equations of the quasi-static isotropic linear thermoelastic
problem are:
∇ ⋅ σ ( x, t ) + ρ0b ( x, t ) =
0 Equilibrium Equation
ε ( x, t ) C -1 : σ ( x, t ) + α ∆θ 1
= Inverse Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= ( u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary Conditions in Space
Γσ : t * =
σ ⋅n
87
2nd Thermal Analogy
The actions and responses of the problem are:
not
(I )
( x, t )
not
ACTIONS = A RESPONSES = R ( ) ( x, t )
I
b ( x, t )
u ( x, t )
t * ( x, t ) Elastic model ε ( x, t )
u * ( x, t ) EDPs+BCs σ ( x, t )
∆θ ( x, t )
REMARK
∆θ ( x,t ) is known a priori, i.e., it is
independent of the mechanical response.
This is an uncoupled thermoelastic problem.
88
2nd Thermal Analogy
The assumption is made that ∆θ (x, t ) and α (x) are such that the thermal
strain field ε t = α∆θ 1 is integrable (satisfies the compatibility equations).
If the thermal strain field is integrable, there exists a field of thermal
displacements, ut ( x, t ), which satisfies:
1 t
ε t ( x, t ) = (α ∆θ ) 1 = ∇ S ut =( u ⊗ ∇ + ∇ ⊗ ut )
2
1 ∂ uit ∂ u tj
(α ∆θ ) δ ij =
ε ijt = + i, j ∈ {1, 2,3}
2 ∂ x j ∂ xi
REMARK
The solution u t ( x, t ) is determined except for a rigid body motion
characterized by a rotation tensor Ω ∗ and a displacement vector c* .
The family of admissible solutions is u t (=x, t ) u ( x, t ) + Ω ∗ ⋅ x + c* .
This movement can be arbitrarily chosen (at convenience).
Then, the total displacement field is decomposed by defining:
def
u =
( x , t ) u( x , t ) − u t ( x , t )
nt
=
u u nt + ut
89
2nd Thermal Analogy
To solve the problem following the methods used in linear elastic
problems, the thermal terms must be removed.
The strain tensor and the displacement vector splits,=
ε ε nt + ε t and =
u u nt + ut
are replaced into the governing equations:
Geometric equations:
ε= ∇ S u= ∇ S ( u nt + ut =
) ∇ S u nt + ∇
u= ∇
S t
u + εt
S nt
εt εnt ε nt = ∇ S u nt
=
ε ε nt + ε t
Boundary equations:
u = u*
u nt + ut =
u* Γu : u =u* u nt =
u* − u t
=
u u nt + ut
90
2nd Thermal Analogy
∇ ⋅ σ + ρ 0b =0 Equilibrium Equation
ε nt = C -1 : σ Inverse constitutive Equation
ε nt = ∇ S u nt Geometric Equation
Γu : u nt =
u* − u t
Boundary Conditions in space
Γσ : σ ⋅ n =t*
91
2nd Thermal Analogy
The actions and responses of the ANALOGOUS PROBLEM are:
not
( II )
( x, t )
not
ACTIONS = A RESPONSES = R ( ) ( x, t )
II
b ( x, t ) u nt ( x, t )
Elastic model ε nt ( x, t )
t* ( x, t ) EDPs+BCs σ ( x, t )
u nt u* ( x, t ) − ut ( x, t )
=
92
2nd Thermal Analogy
If the actions and responses of the original and analogous problems are
compared:
b b 0
ACTIONS u∗ u∗ − ut ut def
(I ) ( II )
A (x, t ) − A (x, t ) =
* * =
− = A ( III )
t t 0
∆ θ 0 ∆ θ
RESPONSES
u u u ut
nt t
(I ) ( II ) nt t def ( III )
R (x, t ) − R (x, t ) = ε − ε = ε = α ∆θ 1 = R
σ σ 0 0
93
2nd Thermal Analogy
b ( x, t ) b b = 0
t* ( x, t ) = u∗ − u t
u u ∗ = ut ( x, t )
A ( I ) ( x, t ) u ( x, t )
* A(
II )
( x, t ) t*
A ( III ) ( x, t ) t* = 0
∆θ ( x, t ) ∆θ =0 ∆ θ ( x, t )
u ( x, t ) u nt ( x, t ) u = u t ( x, t )
R ( I ) ( x, t ) ε ( x, t ) R ( II ) ( x, t ) ε nt ( x, t )
R(
III )
( x, t ) ε= ε=t
(α ∆θ ) 1
σ ( x, t ) σ ( x, t ) σ=0
95
2nd Analogy in structural analysis
u x = u xt = α∆θ x u x = u xnt
ε x = ε xt = α∆θ ε x = ε xnt
Γu : ux =
u xt α∆θ
= Γu : u x =u*x − u xt =−α∆θ
x =
x =
0
96
Thermal Analogies
Although the 2nd analogy is more commonly used , the 1st analogy
requires less corrections.
The 2nd analogy can only be applied if the thermal strain field
is integrable.
It is also recommended that the integration be simple.
The particular case
Homogeneous material:= α (x) const
=. α
Lineal thermal increment: ∆θ = ax + by + cz + d
is of special interest because the thermal strains are:
ε t = α ∆θ 1 = linear polinomial
and trivially satisfy the compatibility conditions (involving second
order derivatives).
97
Thermal Analogies
In the particular case
Homogeneous material: =α (x) const
=. α
Constant thermal increment: ∆θ (x) = ∆θ
const. =
the integration of the strain field has a trivial solution because
the thermal strains are constant ε t = ∆θ α 1 = const. , therefore:
rigid body motion
(can be chosen arbitrarily:
u ( x, t ) = α ∆θ x + Ω ⋅ x + c
t ∗ ∗
at convenience)
The thermal displacement is:
ut ( x, t )= α ∆θ x x + ut = x + α ∆θ x = (1 + α ∆θ ) x
HOMOTHECY
(free thermal expansion)
98
6.7 Superposition Principle
Ch.6. Linear Elasticity
99
Linear Thermoelastic Problem
The governing eqns. of the isotropic linear thermoelastic problem are:
∇ ⋅ σ ( x, t ) + ρ0b ( x, t ) =
0 Equilibrium Equation
( x, t ) C : ε ( x, t ) − β ∆θ 1
σ= Constitutive Equation
1
ε ( x, t )= ∇ S u ( x, t )= ( u ⊗ ∇ + ∇ ⊗ u ) Geometric Equation
2
Γu : u =
u*
Boundary Conditions in space
Γσ : t * =
σ ⋅n
u ( x, 0 ) = 0
Initial Conditions
u ( x, 0 ) = v 0
100
Linear Thermoelastic Problem
Consider two possible systems of actions:
b ( ) ( x, t ) b( ( x, t )
2)
1
t *( ) ( x, t ) t*( ) ( x, t )
1 2
∆θ (1) ( x, t ) ∆θ ( 2) ( x, t )
v 0( ) ( x ) v 0( ) ( x )
1 2
R ( ) ( x,t ) ≡
1
ε ( ) ( x, t )
1
R(
2)
( x,t ) ≡ ε ( ) ( x, t )
2
σ ( ) ( x, t ) σ ( ) ( x, t )
1 2
101
Superposition Principle
The solution to the system of actions= A (3) λ (1) A (1) + λ ( 2) A ( 2) where
λ (1)and λ ( 2) are two given scalar values, is
= R (3) λ (1) R (1) + λ ( 2) R ( 2) .
102
6.8 Hooke’s Law in Voigt Notation
Ch.6. Linear Elasticity
103
Stress and Strain Vectors
Taking into account the symmetry of the stress and strain tensors,
these can be written in vector form: εx
1 1 ε
εx 2
γ xy γ xz
2
y
ε x ε xy ε xz def ε
=
ε xy ε y ε yz
ε =
not . 1
γ εy
1
γ yz {ε} γ z ∈ R 6
=
2 xy 2 xy
ε xz ε yz ε z
γ xz
1 γ xz 1
γ yz εz
2 2 γ yz
REMARK VOIGT
The double contraction ( σ : ε ) is NOTATION σ x
σ
transformed into the scalar (dot) y
σ x τ xy τ xz
product ({σ} ⋅ {ε} ) : def σ
vectors
σ ≡ τ xy σ y τ yz {σ} τ z ∈ R 6
=
τ xz τ yz σ z xy
σ=
:ε {σ} ⋅ {ε} σ ij ε ij = σ iε i τ xz
2nd order
tensors τ yz
104
Inverse Constitutive Equation
The inverse constitutive equation is rewritten:
ν 1 +ν
ε=− Tr ( σ ) 1 + σ + α ∆θ 1 {ε} = Cˆ −1 ⋅ {σ} + {ε}
t
E E
105
Hooke’s Law
By inverting the inverse constitutive equation, Hooke’s Law in terms of the
stress and strain vectors is obtained:
ν ν
1 1 −ν 1 −ν
0 0 0
Where Ĉ is an elastic
ν 1
ν
0 0 0
constants matrix : 1 −ν 1 −ν
ν ν
1 0 0 0
E (1 −ν ) 1 −ν 1 −ν
C=
ˆ 1 − 2ν
(1 +ν )(1 − 2ν ) 0 0 0 0 0
2 (1 −ν )
1 − 2ν
0 0 0 0 0
2 (1 −ν )
0 1 − 2ν
0 0 0 0
2 (1 −ν )
106
Chapter 6
Linear Elasticity
rs
ee
s gin
6.1 Hypothesis of the Linear Theory of Elasticity
t d le En
r
The linear theory of elasticity can be considered a simplification of the general
ba
ge ro or
eS m
theory of elasticity, but a close enough approximation for most engineering ap-
ci
f
plications. The simplifying hypotheses of the linear theory of elasticity are
ra
C d P cs
b
a
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
x = X+ u =⇒ x≈X
e
(6.1)
X Th
er
tin
≈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-
s gin
tive, etc.) no longer make sense in infinitesimal elasticity.
Likewise, the spatial Nabla differential
operator ( ∇ ) is indistin-
t d le En
guishable from the material one ∇ ¯ ,
r
ba
ge ro or
∂ (•) ∂ (•)
eS m
= =⇒ ¯ (•) .
∇ (•) = ∇
ci
∂X ∂x f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
.A
ρ0 = ρt |F| ≈ ρt ,
X Th
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
temperature θ (x,t) that is constant along time,
s gin
θ (x,t) ≡ θ (x) .
t d le En
r
Adiabatic processes are those that take place without heat generation
ba
ge ro or
eS m
at any point and instant of time.
ci
f
ra
C d P cs
b
a
Heat generated inside a domain V per unit of time:
i
an an n
Qe = ρr dV − q · n dS = 0 ∀ΔV ⊂ V
y ha
le
V ∂V
liv or ec
=⇒ 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
→ minor symmetries
Ci jkl = Ci jlk
ee
(6.7)
s gin
Ci jkl = Ckli j → major symmetries
t d le En
Consequently, the number of different constants in the tensor of elastic constants
C is reduced to 21.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Remark 6.3. An essential characteristic of the elastic behavior
b
a
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
©
(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
s gin
Here, U (t) is the internal energy of the material volume considered.
t d le En
r
ba
ge ro or
Remark 6.4. The stress power (in the case of linear elasticity) is an
eS m
ci
exact differential,
f
ra
C d P cs
. dU
b
a
stress power = σ : ε dV = .
i
dt
an an n
V
y ha
le
liv or ec
.A
i↔k
j↔l
m
1. .
d
d û not . . . .
uu
dt 2
X Th
er
1. 1 .
tin
. .
= 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
(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
=0
1 1
Density of
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
s gin
symmetries in (6.7).
⎧
t d le En
⎪
⎪ ∂ û (εε ) 1
⎪
⎪ 1 1 1
⎨ = C :ε + ε :C = C :ε + C :ε =C :ε =σ
r
∂ε
ba
2 2 2 2
ge ro or
eS m
⎪
ci
⎪
⎪ ∂ û (εε ) 1
⎪ 1 1
f 1
ra
⎩ = Ci jkl εkl + εkl Ckli j = Ci jkl εkl + Ci jkl εkl = Ci jkl εkl = σi j
C d P cs
∂ εi j
b
2 2 2 2
a
i
⎧ (6.14)
an an n
⎪
y ha
⎪
⎪ ε
∂ û (ε )
⎨ =σ
le
∂ε
liv or ec
=⇒ (6.15)
⎪
⎪ ∂ û ε
(ε )
M
⎪
.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
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 .
s gin
Consider, for example, the systems {x1 , x2 , x3 } and {x1 , x2 , x3 } in Figure 6.2,
the constitutive equation for these two systems is written as
t d le En
r
ba
σ ] = [C
{x1 , x2 , x3 } =⇒ [σ C] : [εε ]
ge ro or
eS m
ci
(6.17)
f
ra
C d P cs
{x1 , x2 , x3 } =⇒ σ]
[σ C]
= [C : [εε ]
b
a
i
an an n
y ha
and, for the case of an isotropic material, the components of C in both sys-
tems must be the same ( [C C] = [C C] ). Therefore, the aforementioned definition
le
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
constants ⎪
⎪
⎪
⎩ i, j, k, l ∈ {1, 2, 3}
on
.O
C
Here, λ and μ are known as Lamé’s constants, which characterize the elastic
©
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
s gin
equation,
1
t d le En
1
σi j = Ci jkl εkl = λ δi j δkl εkl +2μ δik δ jl εkl + δil δ jk εkl . (6.19)
2 2
r
ba
ge ro or
εll εi j ε ji = εi j
eS m
ci
f εi j
ra
C d P cs
b
a
⎧
i
an an n
Hooke’s law.
M
.A
m
The constitutive equation (6.20) provides the stresses in terms of the strains. To
X Th
er
tin
.O
⎪
⎬
3 =⇒
(i = j) =⇒ σii = λ εll δii +2μεii = (3λ + 2μ) εll ⎪
⎪
⎪
⎪ (6.21)
⎭
3
1
=⇒ Tr (εε ) = σ) .
Tr (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 =− σ)1+
Tr (σ 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 ⎪
⎪
Poisson’s ratio ν= ⎭
2 (λ + μ)
⎧ (6.23)
⎪
⎪
rs
⎪
⎪ νE
⎨λ =
ee
(1 + ν) (1 − 2ν)
=⇒
⎪
s gin
⎪
⎪ E
⎪
⎩μ = =G shear (elastic) modulus
2 (1 + ν)
t d le En
r
Equation (6.22) can be expressed in terms of E and ν, resulting in the inverse
ba
ge ro or
eS m
Hooke’s law.
ci
f
ra
⎧
C d P cs
b
a
⎪
⎪ ν 1+ν
⎪
i
⎪ ε = − Tr (σ σ)1+ σ
an an n
Inverse constitutive ⎨⎪ E E
y ha
⎪
⎪ E E
⎪
⎩ i, j ∈ {1, 2, 3}
M
.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
⎪ ⎪ γxy = =0
ee
⎪ εx = ⎪
⎪
⎨ E ⎪
⎨ G
σ τ
s gin
σy = σz = 0 =⇒ εy = −ν x τxy = τxz = τyz = 0 =⇒ γxz = xz = 0
⎪
⎪ E ⎪
⎪ G
⎪ ⎪
⎩ εz = −ν σx
⎪ ⎩ γyz = τyz = 0
⎪
t d le En
E G
r
ba
ge ro or
eS m
Therefore, due to these strains, the rectangular cuboid subjected to an uni-
ci
f
ra
axial tensile test, shown in the figure below, stretches in the x-direction and
C d P cs
b
a
contracts in the y- and z-directions.
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 ) = − σ ) Tr (1) +
Tr (σ σ) =
Tr (σ σ) =
Tr (σ
E E E
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
s gin
3 3 (1 − 2ν)
Introducing (6.26), (6.27) and (6.29) in (6.24), results in
t d le En
r
ν 1+ν 1 − 2ν 1+ν
ba
ge ro or
ε = − 3σm 1 + (σm 1 + σ ) = σm 1 + σ =
eS m
E
ci
E E E
f
ra
E
C d P cs
b
e
a
3 (1 − 2ν)
i
an an n
1 1 1
= e1+ σ =⇒ ε = e 1 + ε = e 1 + σ
le
3 E 3 3 E
liv or ec
M
.A
1+ν 1 1
=⇒ ε = 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 and ε ) 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
©
(6.31)
σ = 2Gεε
Deviatoric part
σ i j = 2Gε 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
ee
s gin
Thermodynamic considerations allow proving that the tensor of elastic constants
C is positive-definite5 , and, thus,
t d le En
ε :C :ε >0 ∀εε = 0 . (6.32)
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Remark 6.7. As a consequence of (6.32), the elastic potential is al-
b
a
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
er
tin
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 )
©
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
ee
s gin
Consider the expression of the elastic potential (6.16) and the constitutive
equation (6.20), then,
t d le En
1
r
1 1
û (εε ) = ε : C : ε = σ : ε = λ Tr (εε ) 1 + 2μεε : ε =
ba
ge ro or
eS m
2 2 2
ci
f (6.33)
ra
1 1
= λ Tr (εε ) 1 : ε +μεε : ε = λ Tr2 (εε ) + μεε : ε .
C d P cs
b
a
2 2
i
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
1 2
m
1
û (εε ) = λ Tr (εε ) + μεε : ε = λ e2 + μεε : ε .
d
(6.34)
2
uu
2
e
e
X Th
er
tin
.O
1 1 1 2
e 1 + ε = e2 1 : 1 + e 1 : ε + ε : ε =
C
ε :ε = e1+ε :
3
©
3 3 9
3 Tr (εε ) = 0 (6.35)
1
= e2 + ε : ε .
3
Replacing (6.35) in (6.34),
1 2 1 2 1 2
û (εε ) = λ e + μe + μεε : ε = λ + μ e2 + μεε : ε . (6.36)
2 3 2 3
K
6 The trace of a deviatoric tensor is always null, Tr (εε ) = 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 : ε ≥ 0 (6.37)
2
rs
3 =⇒ û(1) = Ke2 ≥ 0 =⇒ K > 0 (6.38)
(1) ⎪ 2
ee
ε = 0 ⎭
s gin
b) A pure deviatoric deformation process7
t d le En
ε (2) = ε
r
=⇒ û(2) = μεε : ε ≥ 0 =⇒ μ > 0
ba
(6.39)
ge ro or
(2)
eS m
e =0
ci
f
ra
C d P cs
Equations (6.38) and (6.39) lead to
b
a
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
> 0⎪
⎬
e
2 (1 + ν)
X Th
=⇒ E >0,
er
tin
⎪
⎭
ν ≥0
on
.O
⎫ (6.41)
E
> 0⎪
⎬
C
3 (1 − 2ν) 1
=⇒ 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 : ε = ε i j ε 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.
ee
s gin
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-
r
ba
ge ro or
tions characterized by the vector of body forces b (x,t) in the interior of the vol-
eS m
ci
ume V and the traction vector t (x,t) on the boundary ∂V . The set of equations
f
ra
that allow determining the evolution along time of the displacements u (x,t),
C d P cs
b
a
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)
s gin
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).
These are of the type (•) (x, y, z,t), and, thus, must be solved in the R3 × R+
r
ba
ge ro or
eS m
space. The problem is well defined when the adequate boundary conditions are
ci
provided. f
ra
C d P cs
b
a
i
6.6.2 Boundary Conditions
an an n
y ha
.A
(6.46)
uu
e
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.
s gin
t d le En
• Boundary Γuσ : prescribed displacements and tractions11
r
ui (x,t) = u∗i (x,t)
ba
ge ro or
eS m
i, j, k ∈ {1, 2, 3}, i =
j ∀x ∈ Γuσ ∀t (6.49)
σ jk (x,t) · nk = t ∗j (x,t)
ci
f
ra
C d P cs
b
a
i
an an n
le
liv or ec
Solution
M
.A
The different types of boundary conditions in space are illustrated in the fol-
lowing figure of a beam.
m
d
uu
e
X Th
er
tin
on
.O
C
11In Γuσ certain components (components i ) have prescribed displacements while the others
(components j ) have the traction vector prescribed.
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
s gin
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-
r
tial equations given in Section 6.6.1 and the boundary conditions described in
ba
ge ro or
eS m
Section 6.6.2, provides the response or solution in the form of the displacement
ci
f
field u (x,t), the deformation field ε (x,t) and the stress field σ (x,t).
ra
C d P cs
b
a
i
⎫
an an n
b (x,t) ⎪ ⎧
y ha
⎪
⎬ MAT HEMAT ICAL ⎨ u (x,t)
u∗ (x,t)
le
⇒ ⇒ ε (x,t)
liv or ec
∗ MODEL :
t (x,t) ⎪⎪ PDEs + BCs ⎩ (6.51)
⎭ σ (x,t)
M
.A
v0 (x)
m
not
Responses = R (x,t)
d
not
Actions = A (x,t)
uu
e
X Th
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
and the time variables R3 × R+ . However, in certain cases, the integration
space can be reduced in one dimension, the one corresponding to time. This is
the case for the so-called quasi-static problems.
∂ 2 u (x,t)
a= ≈0.
∂t 2
rs
This hypothesis is acceptable when the actions are applied slowly.
ee
In
2such case, the variation of the actions A along time is slow
s gin
∂ A /∂t 2 ≈ 0 and, due to the continuous dependency of the results
on
2the data, the variation of the response R along time is also small
t d le En
∂ R /∂t 2 ≈ 0 . Consequently, the second derivative of the response
r
is considered negligible and, in particular,
ba
ge ro or
eS m
ci
∂ 2 u (x,t)
f
ra
≈0.
C d P cs
∂t 2
b
a
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 )
⎭ PDEs + BCs ⎩ (6.55)
t∗ (x, λ ) σ (x, λ )
rs
not not
Actions = A (x, λ ) Responses = R (x, λ )
ee
s gin
In other words, for each value of the actions (characterized by a fixed value of
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.
r
ba
ge ro or
eS m
Example 6.3 – Application to a typical problem of strength of materials.
ci
f
ra
C d P cs
b
a
Solution
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
Now, if the evolution along time of λ (t) can take any form, the value of
tin
.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
ee
s gin
6.7 Solution to the Linear Elastic Problem
t d le En
r
The linear elastic problem can be typically solved following two different ap-
ba
ge ro or
eS m
proaches:
ci
f
ra
a) Displacement formulation
C d P cs
b
a
b) Stress formulation
i
an an n
Their names are directly related to which is the main unknown being considered
y ha
.A
er
tin
on
.O
∂ 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
u (x,t) intervenes as an unknown. The first step consists in replacing the consti-
rs
tutive equation in the Cauchy’s equation, both given in (6.56).
ee
s gin
∂ 2u
∇ · σ + ρ0 b = ∇ · λ Tr (εε ) 1 + 2μεε + ρ0 b = ρ0 2
∂t
t d le En
(6.59)
∂ 2u
r
=⇒ λ ∇ · Tr (εε ) 1 + 2μ∇ · ε + ρ0 b = ρ0 2
ba
ge ro or
∂t
eS m
ci
f
ra
Consider the following identities13 .
C d P cs
b
a
i
an an n
∂ εi j ∂1 ∂ ui ∂ u j 1 ∂ 2 ui 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
∂ ∂ ∂ ul ∂ ∂ ul
C
∇ · Tr (εε ) 1 = (εll δi j ) = δi j = =
©
i ∂xj ∂ x j ∂ xl ∂ xi ∂ xl
∂
= (∇ · u) = ∇ (∇ · u) i i ∈ {1, 2, 3}
∂ xi
∇ · Tr (εε ) 1 = ∇ (∇ · u)
(6.61)
de f
13 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
in Γσ of (6.57) results in
s gin
t∗ = σ · n = λ Tr (εε ) 1 + 2μεε · n = λ (Tr (εε )) n + 2μεε · n =
t d le En
(6.63)
r
= λ (∇ · u) n + 2μ ∇S · u · n = λ (∇ · u) n + μ (u ⊗ ∇ + ∇ ⊗ u) · n
ba
ge ro or
eS m
ci
f
ra
and the boundary conditions in space (6.57) expressed in terms of the displace-
C d P cs
b
a
ments are obtained.
i
an an n
u = u∗
y ha
in Γu
ui = u∗i
le
i ∈ {1, 2, 3}
liv or ec
(6.64)
λ (∇ · u) n + μ (u ⊗ ∇ + ∇ ⊗ u) · n = t∗
M
.A
in Γ σ
λ ul,l ni + μ (ui, j n j + u j,i n j ) = ti∗ i, j ∈ {1, 2, 3}
m
d
uu
The initial conditions (6.58) remain unchanged. Integrating the system (6.62)
e
X Th
yields the displacement field u (x,t). Differentiation of this field and substitution
er
tin
in the geometric equation of (6.56) produces the strain field ε (x,t), and, finally,
replacing the strain in the constitutive equation results in the stress field σ (x,t).
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
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
s gin
ωr = −Ωθ z = −
2 r ∂θ ∂z
t d le En
1 ∂ ur ∂ uz
ωθ = −Ωzr = −
r
2 ∂z ∂r
ba
ge ro or
eS m
ci
1 1 ∂ (ruθ ) 1 ∂ ur
ωz = −Ωrθ = − f
ra
C d P cs
2 r ∂r r ∂θ
b
a
i
an an n
1 ∂ (rur ) 1 ∂ uθ ∂ uz
e= + +
y ha
r ∂r r ∂θ ∂z
le
liv or ec
M
.A
⎡ ⎤
x = r cos θ
x (r, θ , z) ≡ ⎣ y = r sin θ ⎦
m
not
d
z=z
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
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
r sin θ ∂ φ ∂r r ∂θ ∂t
rs
r
ee
where
' (
s gin
1 1 ∂ uφ sin θ 1 ∂ uθ
ωr = −Ωθ φ = −
r sin θ ∂θ r sin θ ∂ φ
t d le En
2
' (
r
1 ∂ ur 1 ∂ ruφ
ba
1
ge ro or
ωθ = −Ωφ r = −
eS m
2 r sin θ ∂ φ r ∂r
ci
f
ra
C d P cs
b
a
1 1 ∂ (ruθ ) 1 ∂ ur
ωz = −Ωrθ = −
i
an an n
2 r ∂r r ∂θ
' (
y ha
1 ∂ r2 ur sin θ ∂ (ruθ sin θ ) ∂ ruφ
le
e= 2 + +
liv or ec
r sin θ ∂r ∂θ ∂φ
M
.A
m
⎡ ⎤
d
x = r sin θ cos φ
uu
e
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
ee
s gin
Γu : u = u∗
Boundary conditions in space (6.68)
Γσ : t∗ = σ · n
t d le En
r
ba
ge ro or
where the inverse constitutive (6.24) (strains in terms of stresses) has been con-
eS m
ci
sidered in (6.67).
f
ra
The starting point of the stress formulation is the geometric equation of (6.67)
C d P cs
b
a
from which, by means of successive differentiation, the displacements are elim-
i
an an n
.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
ν
©
1
∇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
s gin
2 (6.73)
⎪
⎩ u (x) = u∗ (x) ∀x ∈ Γu
t d le En
r
Thus, the system of second-order PDEs must be integrated in R3 .
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
R (x,t) ≡ ⎣ ε (x,t) ⎦
not
©
σ (x,t)
to the linear elastic problem posed in (6.42) to (6.44) is unique.
Proof
Consider the actions defined by A (x,t) ≡ [b (x,t) , u∗ (x,t) , t∗(x,t) , v0 (x)]T ,
not
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
Γσ Γu = 0) / act on the linear elastic problem schematically represented in Fig-
s gin
ure 6.10.
The possible solutions R (x,t) ≡ [u (x,t) , ε (x,t) , σ (x,t)]T to the linear elas-
not
t d le En
tic problem must satisfy the equations:
r
ba
ge ro or
∂ 2u
eS m
∇ · σ + ρ0 b = ρ0 Cauchy’s equation
ci
∂t 2
f
ra
C d P cs
σ = λ Tr (εε ) 1 + 2μεε (6.74)
b
a
Constitutive equation
i
an an n
1
ε = ∇S u = (u ⊗ ∇ + ∇ ⊗ u) Geometric equation
y ha
2
le
liv or ec
M
Γu : u = u∗
.A
d
uu
e
X Th
er
u (x, 0) = 0
tin
.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) ⎦
σ (1) (x,t) σ (2) (x,t) (6.77)
which, therefore, must satisfy equations (6.74) to (6.76) and are the elastic re-
sponses to the same action A (x,t) ≡ [b (x,t) , u∗ (x,t) , t∗ (x,t) , v0 (x)]T . Con-
not
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
+ (x,t) = ∇ · σ (2) (x,t) − σ (1) (x,t) = ∇ · σ (2) − ∇ · σ (1) =
∇·σ
ee
s gin
' ( ' (
∂ 2 u(2) ∂ 2 u(1) (6.79)
= −ρ0 b + ρ0 − −ρ0 b + ρ0 =
t d le En
∂t 2 ∂t 2
r
ba
ge ro or
∂ 2 u(2) ∂ 2 u(1) +
∂ 2u
eS m
= ρ0 − ρ = ρ
ci
0 0
∂t 2 ∂t 2
f ∂t 2
ra
C d P cs
b
a
• Constitutive equation17
i
an an n
y ha
(6.80)
= C : ε (2) − ε (1) = C : ε+
M
.A
m
• Geometric equation
d
uu
e
X Th
(6.81)
= ∇S u(2) − u(1) = ∇S u +
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
292 C HAPTER 6. L INEAR E LASTICITY
rs
⎪
⎨u + (x, 0) = u(2) (x, 0) − u(1) (x, 0) = 0 − 0 = 0
ee
(6.84)
⎪ + (x, 0) .
⎩ ∂u
⎪
s gin
. .
+ (x, 0) = u(2) (x, 0) − u(1) (x, 0) = v0 − v0 = 0
=u
∂t
t d le En
Consider now the calculation of the integral
r
ba
ge ro or
eS m
= 0 in Γσ Divergence
ci
. Theorem .
. f
ra
+ ·u + · u + ·u
C d P cs
n· σ + dS = n·σ + dS = ∇· σ + dV = 0 , (6.85)
b
a
i
∂V Γu Γσ
an an n
= 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
⎪
⎪ . . . + . . T
⎪
⎪ ∂ 2u
⎨∇· σ + ·u + = ∇·σ + ·u ++σ + : ∇u + = ρ0 2 · u ++σ + : ∇u +
m
∂t
d
uu
⎪ . ∂ σ+ . . .
e
⎪
⎪ ∂ ∂ +
u ∂ +
2u . ∂ +
u
⎪
X Th
i j j j j
⎩ σ+i j u+j = u+j + σ+i j = ρ0 2 u+j + σ+ ji i, j ∈ {1, 2, 3}
er
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 . . . .
©
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
. .
.
∂2u + . ∂2u+ . 1 1 ∂ + +·u
∂ u
v·++
v
ρ0 2 · u+ = ρ0 2 · u+ = ρ0 = ρ0 =
∂t ∂t 2 ∂t 2 ∂t (6.88)
d 1 2 ∂2u+ . d 1 2
= ρ0 +
v =⇒ + = ρ0
ρ0 2 · u +
v .
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
.
.
d 1 2
ee
+
∇· σ ·u + dV = ρ0 +
v dV + σ + : ε+ dV = 0 =⇒
dt 2
s gin
V V
. V
d 1
ρ0 +
v2 dV + σ+ : ε+ dV = 0 =⇒ (6.89)
t d le En
dt 2
V V
r
ba
ge ro or
+ /dt d U+/dt
eS m
dK
ci
f
ra
C d P cs
d + +
+ d U+
b
a
dK
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 ⎪
+ = v0 dV = 0 ⎪
d
K ρ0+ v dV =
2
ρ0 + v0 · + ⎪
⎪
uu
2 2 ⎪
⎪
e
t=0 t=0
. ⎪
⎬
X Th
V V
er
+0 = 0
tin
u
⇒ +
K + +
U =0
+ 1 ⎪
⎪ t=0
U = û (x,t) dV = ε+ : C : ε+ dV = 0 ⎪
on
⎪
.O
t=0 t=0 2
t=0 t=0 ⎪
⎪
V V ⎪
⎭
C
=0
©
(6.91)
and the integration of (6.90) with the initial condition (6.91) leads to
+ + U+ = 0
K ∀t ≥ 0 , (6.92)
where
+= 1
K ρ0 +
v2 dV ≥ 0 ∀t ≥ 0 . (6.93)
2
V
≥0
de f
19 v| = +
Here, the definition |+ 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
ε+ (x,t) : C : ε+ (x,t) ≥ 0 ∀x ∈ V ∀t ≥ 0 =⇒
rs
1 (6.95)
U+ = ε+ : C : ε+ dV ≥ 0 ∀t ≥ 0 .
ee
2
V
s gin
Then, comparing (6.94) and (6.95) necessarily leads to
t d le En
U+ (t) ≤ 0 1
r
+
∀t ≥ 0 =⇒ U (t) = +
ε :C:+ ε dV = 0 ∀t ≥ 0 . (6.96)
ba
ge ro or
+
eS m
U (t) ≥ 0 2
ci
fV
ra
C d P cs
b
a
Considering once more the positive-definite condition of tensor C 20 ,
i
an an n
y ha
1
U+ = ε+ : C : ε+ dV = 0 ∀t ≥ 0 =⇒ ε+ : C : ε+ = 0 ∀x , ∀t ≥ 0 (6.97)
le
2
liv or ec
V ≥0
M
.A
ε+ : C : ε+ = 0 ⇐⇒ ε+ (x,t) = 0 ∀x , ∀t ≥ 0
uu
(6.98)
e
X Th
er
tin
.O
1 ∂ u+i ∂ u+j
©
ε+ (x,t) = ∇ · u
S
+ = 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
+ (x,t) = Ω
u + · x + +c with
rotation translation
⎡ ⎤ ⎡ ⎤ (6.101)
0 −θ+3 θ+2 c+1
+ not ⎢ ⎥ not ⎢ ⎥
Ω ≡ ⎣ θ+3 0 −θ+1 ⎦ and +c ≡ ⎣ c+2 ⎦ ,
−θ+2 θ+1 c+3
rs
0
ee
where Ω + is an antisymmetric tensor (rotation tensor dependent on three con-
s gin
stants θ+1 , θ+2 and θ+3 ) and +c is a constant vector equivalent to a translation. Ulti-
mately, the solution (6.100) to the system (6.101) are the displacements u + (x,t)
t d le En
compatible with a null strain ε+ (x,t) = 0, which correspond to a rigid body
r
+ and +c are determined by imposing the
ba
motion. The integration constants in Ω
ge ro or
eS m
boundary conditions (6.82) (+ u (x,t) = 0 ∀x ∈ Γu ), therefore, if the rigid body
ci
f
ra
motion is impeded through the restrictions in Γu , one obtains Ω + = 0 and +c = 0.
C d P cs
b
a
In conclusion,
i
an an n
y ha
+ (x,t) = Ω
u + · x ++c
=⇒ u + (x,t) = u(2) − u(1) = 0 =⇒ u(2) = u(1) .
le
liv or ec
+
Ω ≡ 0 ; +c ≡ 0
M
.A
(6.102)
Finally, replacing (6.99) in (6.80) yields
m
d
uu
σ =⇒ (6.103)
X Th
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
ee
s gin
Figure 6.11: Saint-Venant’s principle.
t d le En
r
ba
ge ro or
eS m
boundary characterized by the traction vector t∗ (see Figure 6.11). These ac-
ci
tions will lead f
to a solution or response in displacements, strains and stresses,
ra
not T
C d P cs
R (I) (x,t) ≡ u(I) (x,t), ε (I) (x,t) , σ (I) (x,t) . Consider now a part Γ̂ of the
b
a
i
boundary Γσ Γ̂ ⊂ Γσ of said medium, whose typical dimension is , and re-
an an n
y ha
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
not T
.A
Saint-Venant’s principle states that, for the points belonging to the domain Ω
d
that are sufficiently far from the boundary Γ̂ , the solution in both cases is prac-
uu
e
er
tin
⎫
u(I) (x p ,t) ≈ u(II) (x p ,t) ⎪
on
.O
⎬
ε (I) (x p ,t) ≈ ε (II) (x p ,t) ∀P δ . (6.105)
C
⎪
⎭
©
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
distributed tensile loads in the end sections σ = F/A (system (II)), the elastic
rs
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
s gin
of materials (uniformly distributed stresses in all the piece, σx = F/A). At
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
also allows dimensioning the strength characteristics of the piece for practical
r
ba
ge ro or
eS m
purposes.
ci
f
ra
C d P cs
b
a
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) = θ (x, 0) = θ0 ,
. ∂ θ (x,t) (6.106)
θ (x,t) = = 0 .
∂t
Nevertheless, the hypothesis that the processes are adiabatic (slow) is maintained
and, thus,
ρ0 r − ∇ · q ≈ 0 . (6.107)
rs
6.10.1 Linear Thermoelastic Constitutive Equation
ee
s gin
Hooke’s law (6.6) in this case is generalized to
t d le En
σ = C : ε − β (θ − θ0 )
, (6.108)
r
ba
σi j = Ci jkl εkl − βi j (θ − θ0 ) i, j ∈ {1, 2, 3}
ge ro or
eS m
ci
f
ra
Here, C is the tensor of elastic constants defined in (6.7), θ (x,t) is the tempera-
C d P cs
b
a
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
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
β i 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 ) = Δ 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 Δ θ 1
σi j = λ εll δi j + 2μεi j − β Δ θ δi j i, j ∈ {1, 2, 3}
rs
Equation (6.111) can be inverted as follows.
ee
⎧
s gin
−1 −1
⎪
⎨σ = C : ε −Δθ β ⇒ ε = C : σ +Δθ C : β = C −1 : σ + Δ θ α
α
t d le En
⎪
⎩ de f −1
α = C : β → Tensor of thermal expansion coefficients
r
ba
ge ro or
(6.112)
eS m
ci
where α is a second-order (symmetric) tensor involving six thermal properties
f
ra
C d P cs
named coefficients of thermal expansion. For an isotropic case, in agreement
b
a
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
ε =− σ)1+
Tr (σ σ + αΔ θ 1 (6.113)
m
E E
d
uu
ν 1+ν
e
εi j = − σll δi j + σi j + αΔ θ δi j i, j ∈ {1, 2, 3}
X Th
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 Δ θ 1 = σ nt − σ t
σ nt σt
⎧ (6.115)
⎪
⎨ Non-thermal stress → σ nt de f
= λ Tr (εε ) 1 + 2μεε
rs
⎪
⎩ Thermal stress de f
→ σt = β Δθ 1
ee
s gin
Here, σ nt represents the stress produced if there do not exist any thermal phe-
nomena and σ t is named thermal stress and acts as the “correcting” stress due
t d le En
to the thermal increment.
r
A similar operation can be performed on the inverse constitutive equations
ba
ge ro or
eS m
for the linear elastic and linear thermoelastic cases of (6.24) and (6.113), re-
ci
spectively, resulting in
f
ra
C d P cs
b
a
ν 1+ν
i
σ)1+
ε = − Tr (σ σ + α Δ θ 1 = ε nt + ε t
an an n
E E t
y ha
ε nt ε
le
liv or ec
⎧ (6.116)
⎪ ν 1+ν
M
.A
⎨ Non-thermal strain → ε nt de f
σ)1+
= − Tr (σ σ
E E
⎪
m
⎩ Thermal strain de f
→ εt = α Δθ 1
d
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 = Δθ β
σ = σ nt − σ t Isotropic material: Isotropic material: (6.117)
σ nt = λ Tr (εε ) 1 + 2μεε σt = β Δθ 1
ε nt = C −1 : σ εt = Δθ α
ε = ε nt + ε t Isotropic material: Isotropic material: (6.118)
ν 1+ν εt = α Δθ 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
state of null stress in said point. In effect, for ε = 0 in (6.117),
s gin
σ t = −β Δ θ 1 = 0 .
ε = 0 =⇒ σ nt = 0 =⇒ σ = −σ
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Δ θ = 0
C d P cs
σ t = −β Δ θ 1
σ = −σ
b
a
ε =0
i
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 Δ θ 1 = 0 .
C
Δ θ = 0
ε = εt = α Δθ 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
ee
s gin
Figure 6.12: Actions on a continuous medium.
t d le En
r
ba
6.11 Thermal Analogies
ge ro or
eS m
ci
f
ra
The thermal analogies arise from the search of procedures to solve the linear
C d P cs
b
a
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
Γu and Γσ act the prescribed displacements u∗ (x,t) and a traction vector t∗ (x,t),
tin
respectively.
on
.O
⎧
©
⎪
⎪
⎪
⎪ ∇ · σ + ρ0 b = 0 Equilibrium equation
⎨
Governing
σ = C : ε −β Δθ 1 Constitutive equation
equations ⎪
⎪
⎪
⎪
⎩ ε = ∇S u Geometric equation (6.121)
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) ⎪
⎪ + ⎩ (6.122)
⎭ PDEs BCs σ (x,t)
Δ θ (x,t)
Responses = R (I) (x,t)
not
Actions = A (I) (x,t)
not
rs
ee
To be able to apply the resolution methods typical of the liner elastic problem
s gin
developed in Section 6.7, the thermal term in the equations of the thermoelastic
problem (6.121) must be eliminated (at least, in appearance). To this aim, the
t d le En
σ t is replaced in (6.121) as follows.
decomposition of the stress tensor σ = σ nt −σ
r
ba
a) Equilibrium equation
ge ro or
eS m
ci
f
ra
σ = σ nt − σ t =⇒
C d P cs
b
a
∇·σ = ∇ · σ nt − ∇ · σ = ∇ · σ nt − ∇ (β Δ θ )
t
i
(6.123)
an an n
y ha
β Δθ1
le
liv or ec
1
∇ · σ + ρ0 b = 0 =⇒ ∇ · σ nt + ρ0 b − ∇ (β Δ θ ) = 0
M
.A
ρ0
(6.124)
m
not
= b̂
d
uu
e
=⇒ ∇ · σ + b̂ = 0
nt
X Th
er
tin
.O
⎪
⎪
©
⎪ 1
⎨ b̂ (x,t) = b (x,t) − ∇ (β Δ θ )
ρ0 (6.125)
⎪
⎪ 1 ∂ (β Δ θ )
⎪
⎩ b̂i (x,t) = bi (x,t) − i ∈ {1, 2, 3}
ρ0 ∂ 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
s gin
=⇒ σ nt · n − σ t · n = t∗ =⇒
σ · n = t∗
t d le En
(6.129)
∗ ∗
σ nt · n = t∗ + σ · n = t + β Δ θ n =⇒
t
Γσ : σ · n = t̂
nt
r
ba
ge ro or
eS m
β Δθ 1·n t̂∗
ci
f
ra
where t̂∗ (x,t) is a pseudo-traction vector defined by
C d P cs
b
a
i
an an n
t̂∗ = t∗ + β Δ θ n .
y ha
(6.130)
le
liv or ec
.A
⎧
⎪
⎪ ∇ · σ nt + ρ0 b̂ = 0
m
⎪
⎪
Equilibrium
d
⎪ 1
⎪ with b̂ = b − ∇ (β Δ θ )
uu
⎪ 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 Δ 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 ⎩ nt (6.132)
t̂∗ (x,t) σ (x,t)
not (II)
Responses = R(II) (x,t)
not
Actions = A (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 b − b̂ ∇ (β Δ θ )
⎢ ⎥
⎢ u∗ ⎥ ⎢ 0 ⎥ ⎢ ρ0
s gin
not ⎢ u ⎥ ⎥ de f (III)
∗
(I)
A −A (II) ⎢ ⎥ ⎢ ⎥ ⎢
A ≡ ⎣ ∗ ⎦−⎣ ∗ ⎦ = ⎣ ∗ ∗ ⎦ = ⎢ ⎥ ⎢ 0 ⎥=A (x,t)
t t̂ t − t̂ ⎥
⎣ −β Δ θ n ⎦
t d le En
Δθ 0 Δθ Δθ
r
⎡ ⎤ ⎡
ba
⎤
ge ro or
⎡ ⎤ ⎡ ⎤
eS m
ci
u u 0 0
⎢ ⎥ ⎢
f ⎥ de f (III)
ra
R(II) ≡ ⎣ ε ⎦ − ⎣ ε ⎦ = ⎢
R (I) −R ⎥=⎣
not
⎦=R (x,t)
C d P cs
⎣ 0 ⎦ 0
b
a
σ σ nt
σ −σ −β Δ θ 1
i
nt
an an n
y ha
−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 Δθ1
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
s gin
t d le En
⎡ ⎤
b (x,t) ⎡ ⎤
r
⎢ u∗ (x,t) ⎥ u (x,t)
ba
ge ro or
⎢ ⎥ ⎢ ⎥
eS m
⎢ ∗ ⎥ ⎣ ε (x,t) ⎦
ci
⎣ t (x,t) ⎦
f
ra
σ (x,t)
C d P cs
Δ θ (x,t)
b
a
i
an an n
le
liv or ec
⎡ ⎤
M
.A
1
b̂ = b − ∇ (β Δ θ )
⎢ ρ0 ⎥ ⎡ ⎤
⎢ ⎥ u (x,t)
m
⎢ ⎥
u∗ (x,t)
d
⎢ ⎥ ⎢ ⎥
⎢ ⎥ ⎣ ε (x,t) ⎦
uu
⎢ ⎥
e
⎢ ∗ ∗
t̂ = t + β Δ θ n ⎥ σ nt (x,t)
X Th
⎣ ⎦
er
tin
Δθ = 0
on
.O
⎡ ⎤
+ = 1 ∇ (β Δ θ )
b ⎡ ⎤
⎢ ρ0 ⎥ u=0
⎢ ⎥
⎢ +
u ∗=0 ⎥ ⎢ ⎥
⎢ ⎥ ⎣ ε =0 ⎦
⎢ +∗ ⎥
⎣ t = −β Δ θ n ⎦ σ = −β Δ θ 1
Δ 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
⎩ ε = ∇S u Geometric equation (6.135)
ee
s gin
Boundary Γu : u = u∗
conditions Γσ : σ · n = t∗
t d le En
r
which constitute the actions (data) A (x,t) and responses (unknowns) R (x,t) of
ba
ge ro or
eS m
the problem.
ci
f
ra
C d P cs
b
a
⎫
⎧
i
b (x,t) ⎪
an an n
∗
⎪
⎬ MAT HEMAT ICAL ⎨ u (x,t)
y ha
u (x,t)
⇒ MODEL : ⇒ ε (x,t)
t∗ (x,t) ⎪
le
⎪ ⎩
liv or ec
.A
Responses = R (I) (x,t)
not
Actions = A (I) (x,t)
not
m
d
uu
e
X Th
er
tin
.O
α (x) and the thermal increment Δ 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
rigid body motion
s gin
The rigid body motion may be chosen arbitrarily (in the form which
t d le En
is most convenient for the resolution process).
r
ba
ge ro or
eS m
ci
f
Once the thermal displacements have been defined, a decomposition of the
ra
C d P cs
total displacements into their thermal and non-thermal parts can be performed
b
a
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
uu
e
X Th
er
∇ · σ + ρ0 b = 0 (6.139)
on
.O
C
c) Geometric equation
⎫
ε = ∇S u = ∇S (unt + ut ) = ∇S unt + ∇S ut = ∇S unt + ε t ⎬
=⇒ ε nt = ∇S unt
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
ee
⎧
s gin
⎪
⎪ Equilibrium
⎪
⎪ ∇ · σ + ρ0 b = 0
⎪
⎪ equation
t d le En
⎪
⎪
⎪
⎨ nt
ε = C −1 : σ =
r
Governing
ba
Constitutive
ge ro or
ν 1+ν
eS m
equations ⎪
⎪ = − Tr (σ σ)1+ σ equation
ci
⎪
⎪
⎪ E E f
ra
⎪ (6.144)
⎪
C d P cs
⎪
⎪
b
a
⎩ ε nt = ∇S unt Geometric
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
Actions = A (II) (x,t) Responses = R (II) (x,t)
not not
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
not ⎢ u ⎥ ⎢ ∗
∗ t⎥ ⎢ t ⎥
A(II) ≡ ⎢
A (I) −A ⎥ ⎢ u − u ⎥ ⎢ u ⎥ de f (III) (x,t)
⎣ t∗ ⎦ − ⎣ t∗ ⎦ = ⎣ 0 ⎦ = A
Δθ 0 Δθ
(6.146)
⎡ ⎤ ⎡ nt ⎤ ⎡ t ⎤ ⎡ t
⎤
u u u u
⎢ ⎥ de f
(I) R ≡ ⎣ ε ⎦ − ⎣ ε ⎦ = ⎣ ε ⎦ = ⎣ α Δ θ 1 ⎦ = R (III) (x,t)
(II) not
R −R nt t
rs
σ σ
ee
0 0
s gin
where equations (6.138) and (6.118) have been taken into account.
t d le En
r
ba
Remark 6.15. It can be directly verified that, in (6.146), R(III) is the
ge ro or
eS m
ci
response corresponding to the system of actions A (III) in the ther-
f
ra
C d P cs
moelastic problem (6.135).
b
a
i
an an n
y ha
Therefore, the original problem (I) can be interpreted as the sum (superposi-
le
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
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) ⎡ ⎤
u (x,t)
ee
⎢ u∗ (x,t) ⎥
⎢ ⎥ ⎢ ⎥
⎢ ∗ ⎥ ⎣ ε (x,t) ⎦
s gin
⎣ t (x,t) ⎦
σ (x,t)
Δ θ (x,t)
t d le En
r
(I) Thermoelastic (original)
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
⎡ ⎤
i
b (x,t) ⎡ ⎤
an an n
⎢ u∗ − ut ⎥ unt (x,t)
y ha
⎢ ⎥ ⎢ nt ⎥
⎢ ∗ ⎥ ⎣ ε (x,t) ⎦
le
⎣ t (x,t) ⎦
liv or ec
σ (x,t)
Δθ = 0
M
.A
m
er
tin
⎡ ⎤
b=0 ⎡ ⎤
on
.O
⎢ ⎥ u = ut (x,t)
⎢ +∗ = ut
u ⎥ ⎢ ⎥
⎣ ε = α Δθ1 ⎦
C
⎢ ⎥
⎣ +t∗ = 0 ⎦
©
σ =0
Δ 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 Δ 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-
static structure); and 3) the displacement of the beam’s right end is recovered
rs
until it is brought again to zero.
ee
This procedure coincides exactly with the application of the second thermal
s gin
analogy in which the thermal displacement field ut is defined by the thermal
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 Δ θ and, when
recovering the displacement at this end, the boundary condition
r
ba
ge ro or
eS m
Γu : u = u∗ − ut = −ut ,
ci
f
ra
C d P cs
b
a
0
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
ee
s gin
t d le En
Remark 6.18. In the case involving
r
• a homogeneous material (αα (x) = const. = α)
ba
ge ro or
eS m
• a constant thermal increment (Δ θ = const.)
ci
f
ra
C d P cs
the integration of the thermal strain field ε t = Δ θ α1 = const. is triv-
b
a
i
ial, resulting in
an an n
y ha
Ω ∗ · x + c∗
ut (x,t) = α Δ θ x +Ω ,
le
liv or ec
.A
where the rigid body motion can be chosen arbitrarily (see Re-
m
er
tin
ut (x,t) = α Δ θ x =⇒ x + ut = x + α Δ θ x = (1 + α Δ 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.
ee
s gin
6.12 Superposition Principle in Linear Thermoelasticity
t d le En
Consider the linear thermoelastic problem in Figure 6.16 and its corresponding
r
governing equations
ba
ge ro or
eS m
ci
∂ 2u f
ra
C d P cs
∇ · σ + ρ0 b = ρ0 Cauchy’s equation
b
a
∂t 2
i
an an n
y ha
C:ε
M
.A
1
ε = ∇S u = (u ⊗ ∇ + ∇ ⊗ u) Geometric equation
2
m
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
⎪
⎪ + ⎩
Δ θ (x,t) ⎪
⎪ PDEs BCs σ (x,t) (6.151)
ee
⎪
v0 (x) ⎭
s gin
not
Responses = R (x,t)
not
Actions = A (x,t)
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Remark 6.19. The different (scalar, vector, tensor and differential)
C d P cs
b
a
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
ee
A (3) = λ (1) A (1) + λ (2) A (2)
s gin
(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)
r
ba
ge ro or
eS m
ci
f
In other words, the solution to the linear thermoelastic problem
ra
C d P cs
when considering a linear combination of different systems of ac-
b
a
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
account the linearity of the different operators (see Remark 6.19) yields
er
tin
a) Cauchy’s equation
on
.O
C
∂ 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 Δ θ (3) 1 = λ (1) σ (1) − C : ε (1) − β Δ θ (1) 1 +
= 0
λ (2) σ (2) − C : ε (2) − β Δ θ (2) 1 = 0
=0
rs
(6.155)
ee
s gin
c) Geometric equation
t d le En
ε (3) − ∇S u(3) = λ (1) ε (1) − ∇S u(1) +λ (2) ε (2) − ∇S u(2) = 0
r
ba
(6.156)
ge ro or
=0 =0
eS m
ci
f
ε (3) = ∇S u(3)
ra
C d P cs
b
a
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
=0 =0 (6.157)
m
Γu : u(3) = u∗(3)
d
uu
e
X Th
er
tin
e) Boundary condition in Γσ
on
.O
=0 =0 (6.158)
©
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
=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-
s gin
tors”. These are constructed in R6 , systematically arranging the elements of the
upper triangle of the matrix of components of the corresponding tensor in the
t d le En
following manner25 .
⎡ ⎤
r
ba
ge ro or
σx
eS m
⎢ ⎥
ci
⎡ ⎤ f ⎢ σy ⎥
ra
σx τxy τxz ⎢ ⎥
C d P cs
⎢ ⎥
b
a
not ⎢ ⎥ de f ⎢ σz ⎥
i
σ ≡ ⎣ τxy σy τyz ⎦ → {σ σ} = ⎢ ⎥ (6.160)
an an n
⎢ τxy ⎥
⎢ ⎥
y ha
τxz τyz σz ⎢τ ⎥
⎣ xz ⎦
le
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 ⎥
εx εxy εxz ⎢ 2 2 ⎥ ⎢ ⎥
C
⎢ ⎥ ⎢ε ⎥
not ⎢ ⎥ not ⎢ 1 ⎥ ⎢ z⎥
©
de f
ε ≡ ⎣ εxy εy εyz ⎦ = ⎢ γxy εy 1 γyz ⎥ → {εε } = ⎢ ⎥ (6.161)
⎢2 2 ⎥ ⎢ γxy ⎥
εxz εyz εz ⎢ ⎥ ⎢ ⎥
⎣1 1 ⎦ ⎢γ ⎥
γxz γyz εz ⎣ xz ⎦
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
rs
nitions in (6.160) and (6.161).
ee
s gin
The inverse constitutive equation (6.113),
t d le En
ν 1+ν
r
ε =− σ)1+
Tr (σ σ + αΔ θ 1 , (6.162)
ba
ge ro or
eS m
E E
ci
f
ra
can now be rewritten in terms of the stress and strain vectors as
C d P cs
b
a
i
C−1 · {σ
{εε } = Ĉ σ } + {εε }t , (6.163)
an an n
y ha
⎡ ⎤
M
.A
1 −ν −ν
⎢ 0 0 0 ⎥
⎢ E E E ⎥
m
⎢ ⎥
d
⎢ −ν 1 −ν ⎥
uu
⎢ ⎥
e
⎢ 0 0 0 ⎥
E
X Th
⎢ E E ⎥
er
tin
⎢ −ν −ν 1 ⎥
⎢ ⎥
⎢ 0 0 0 ⎥
on
not ⎢ ⎥
.O
E E E
C−1 ≡ ⎢
Ĉ ⎥ (6.164)
⎢ ⎥
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 Δ 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 Δθ
⎡ ⎤ ⎢ ⎥
⎢α Δθ ⎥
⎢α Δθ 0 0 ⎥ ⎢
⎢
⎥
⎥
⎢ ⎥ t de f ⎢ α Δ θ ⎥
εt ≡ ⎢ ⎥
not
⎢ 0 α Δ θ 0 ⎥ → {εε } = ⎢ ⎥. (6.165)
⎣ ⎦ ⎢ 0 ⎥
⎢ ⎥
0 0 α Δθ ⎢ 0 ⎥
⎣ ⎦
0
rs
the stress and strain vectors,
ee
Hooke’s law
s gin
in terms of the {σ C · {εε } − {εε }t
σ } = Ĉ (6.166)
t d le En
stress and strain vectors
r
ba
ge ro or
eS m
C is the matrix of elastic constants.
where Ĉ
ci
⎡ f ⎤
ra
C d P cs
ν ν
b
a
⎢ 1 0 0 0 ⎥
⎢ ⎥
i
1−ν 1−ν
an an n
⎢ ⎥
⎢ ν ν ⎥
y ha
⎢ 1 0 0 0 ⎥
⎢ 1−ν 1−ν ⎥
⎢ ⎥
le
⎢ ν ⎥
liv or ec
⎢ ν ⎥
⎢ 1 0 0 0 ⎥
M
⎢ 1−ν 1−ν ⎥
.A
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
2 (1 − ν)
tin
⎢ ⎥
⎢ ⎥
⎣ 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.
ee
s gin
Solution
t d le En
r
a) According to the second law of thermodynamics (5.114),
ba
ge ro or
eS m
. .
ci
ρ0 θ sloc
i
= ρ0 θ s − (ρ0 r − ∇ · q) ≥ 0 .
f
ra
C d P cs
b
a
All processes are reversible in the case of a thermoelastic material and, thus, the
i
an an n
. .
le
ρ0 θ sloc
i = ρ θ s − (ρ r − ∇ · q) = 0 .
liv or ec
0 0 [1]
M
.A
.
An isentropic process (entropy remains constant) is characterized by s = 0. On
m
ρ0 r − ∇ · q = 0 .
er
tin
on
.
©
ρ0 θ s − (ρ0 r − ∇ · q) = 0 =⇒ ρ0 r − ∇ · q = 0 .
=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.
=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),
εi j, kl + εkl, i j − εik, jl − ε jl, ik = 0 i, j, k, l ∈ {1, 2, 3} .
Given that these involve second-order derivatives of the components of the strain
tensor with respect to x, y and z, they will be automatically satisfied if α = const.
and Δ θ = const., or if α Δ θ is linear in x, y and z (which is the definition of a
linear thermoelastic material). Therefore, the statement is true.
rs
ee
s gin
Problem 6.2 – An isotropic linear elastic solid is subjected to a constant pres-
sure of value p on all of its external boundary, in addition to a thermal incre-
ment of Δ θ = θ (x, y, z) in its interior. Both actions cancel each other out such
t d le En
that no displacements are observed in the solid. Obtain the value of Δ θ in each
r
ba
point of the solid.
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
⎨ t∗ = −p n ⎨u
tin
in Γσ
Actions: Responses: ε
⎪ u∗ = 0 ⎪
on
in Γu
.O
⎪
⎪ ⎩
⎪
⎩Δθ = Δθ σ
C
P ROBLEM III
This problem is solved first since its solution is trivial.
⎧
⎪ 1
⎪
⎪ bIII = β ∇ (Δ θ ) ⎧
⎪ ρ
⎪
⎨ ∗ ⎨ uIII = 0
⎪
t
Actions: III = −β Δ θ n in Γ σ Responses: ε III = 0
⎪
⎪ ∗ =0 ⎪
⎩
⎪
⎪ u in Γ u σ III = −β Δ θ 1
⎪
⎩
III
Δ θIII = Δ 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 ∇ (Δ θ )
⎪
⎪ ρ
⎨ ∗
Actions: tII = (−p + β Δ θ ) n in Γσ
⎪
⎪
⎪
⎪ u∗II = u∗ = 0 in Γu
⎪
⎩
Δ 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 =⇒
s gin
bII = 0 =⇒ β ∇ (Δ θ ) = 0 =⇒ Δ θ is uniform
t d le En
In addition, uII = 0 also results in
r
ba
ge ro or
eS m
ci
1 f
ra
ε II = (uII ⊗ ∇ + ∇ ⊗ uII ) = 0 ,
C d P cs
b
a
2
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 Δ θ ) n = 0 ∀n =⇒ −p + β Δ θ = 0 ,
m
d
uu
er
tin
p
Δθ = .
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
Hypotheses:
le
1) Young’s modulus: E
liv or ec
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 Δ 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.
ee
The first thermal analogy (see Section 6.11.1) and the superposition principle
s gin
(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
r
ba
ge ro or
The actions in problem III, the trivial problem, are
eS m
ci
f
ra
1
C d P cs
bIII = β Δθ) .
∇ · (β
b
a
ρ
i
an an n
y ha
bIII = 0 .
M
.A
β Δ θ n = −β Δ θ n.
2) Prescribed stresses in Γσ : t = −β
X Th
er
tin
.O
uIII = 0
C
ε III = 0 [1]
©
σ III = −β Δ θ 1
P ROBLEM II
The actions in problem II, the analogous problem, are
1
β Δθ) .
bII = b − ∇ · (β
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
ee
In addition, uz (z) = 0 will be imposed in all points since no information on the
s gin
top and bottom surfaces of the cylindrical shell is given. Boundary conditions
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 = 0 were
considered. Therefore, the displacement vector
r
ba
ge ro or
eS m
uII (r, z) ≡ [ur (r) , 0 , 0 ]T
not
ci
f
ra
C d P cs
b
a
is adopted. Navier’s equation (6.62) will be used to solve this problem,
i
an an n
∂ 2 uII
y ha
.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]
∂r r ∂θ ∂z ∂t
on
.O
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-
s gin
duced to
∂e ∂ 1 ∂
t d le En
(λ + 2G) = 0 =⇒ E (r ur ) = 0 .
∂r ∂r r ∂r
r
ba
ge ro or
Integrating this last expression leads to
eS m
ci
f
ra
1 ∂ ∂
C d P cs
(r ur ) = 2A =⇒ (r ur ) = 2Ar =⇒ r ur = Ar2 + B
b
a
r ∂r ∂r
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
to this displacement vector is easily obtained by means of the geometric equa-
m
tion (6.3),
uu
⎡ ⎤
e
X Th
B
er
A− 2
tin
0 0
⎢ r ⎥
not ⎢ ⎥
ε II (r) ≡ ⎢ 0 A + B 0 ⎥ . [5]
on
.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 Δ θ n = t∗ + β Δ 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 Δ θ ,
s gin
which, replacing the value of the radial stress from [6] and, consider-
t d le En
ing [3], results in
r
ba
ge ro or
eS m
B
A− = α Δθ .
ci
[7]
4R2 f
ra
C d P cs
b
a
i
an an n
B OUNDARY CONDITION AT r = R
y ha
problem,
M
.A
tII = σ II · n = t∗ + β Δ θ n = t∗ + β Δ θ n .
m
er
t∗ = 0, since, for this phase, problem I has no loading on the lateral walls.
tin
.O
σrr (r = R) = β Δ θ ,
which, replacing the value of the radial stress from [6] and, consider-
ing [3], results in
B
A− = α Δθ . [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 Δ θ r , 0 , 0
⎡ ⎤
α Δθ 0 0
not ⎢ ⎥
ε II ≡ ⎣ 0 α Δ θ 0 ⎦
0 0 0 [10]
⎡ ⎤
Eα Δ θ
rs
0 0
not ⎢ ⎥
ee
σ II ≡ ⎣ 0 Eα Δ θ 0 ⎦
s gin
0 0 0
t d le En
Taking into account the superposition principle (see Section 6.12), and
r
expressions [1], [3] and [10], the original problem is solved for the first
ba
ge ro or
eS m
phase.
not T
ci
u ≡ α Δθr, 0, 0 f
ra
C d P cs
b
a
⎡ ⎤
i
α Δθ 0 0
an an n
not ⎢ ⎥
y ha
ε ≡ ⎣ 0 α Δθ 0 ⎦
le
liv or ec
0 0 0 [11]
⎡ ⎤
M
.A
0 0 0
not ⎢ ⎥
m
σ ≡⎣ 0 0 0 ⎦
d
uu
e
0 0 −Eα Δ θ
X Th
er
tin
To obtain the value of Δ 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 Δ 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
Δθ∗ = . [12]
2αR
b) First, the value Δ 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 Δ 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-
s gin
stants A and B must be determined by means of the boundary conditions.
Stresses can be imposed on the internal lateral wall of the cylindrical shell
t d le En
and displacements, on its external lateral wall.
r
ba
ge ro or
B OUNDARY CONDITION AT r = 2R
eS m
ci
f
If r = 2R and according to the boundary conditions in Γu of the analogous
ra
C d P cs
b
a
problem,
i
ur (r = 2R) = 0 .
an an n
y ha
B
A 2R + =0 . [13]
M
.A
2R
m
d
uu
B OUNDARY CONDITION AT r = R
e
X Th
er
tin
.O
tII = σ II · n = t∗ + β Δ θ n = t∗ + β Δ θ n .
C
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 Δ θ ,
which, replacing the value of the radial stress from [6], and consider-
ing [3], results in
B
A− 2 = α Δθ . [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 Δθ and B = − α Δ 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
1
not 4R2
uII ≡ α Δθ r − , 0, 0
5 r
rs
⎡
⎤
ee
1 4R2
⎢ 5 α Δ θ 1 + r2 0 0 ⎥
s gin
⎢
⎥
not ⎢ 1 4R2 ⎥
ε II ≡ ⎢ 0 α Δθ 1− 2 0 ⎥
⎢ ⎥
t d le En
⎣ 5 r ⎦
r
[16]
ba
0 0 0
ge ro or
eS m
ci
⎡
f ⎤
ra
C d P cs
1 4R2
b
a
⎢ 5 Eα Δ θ 1 + r2 0 0 ⎥
i
⎢ ⎥
an an n
not ⎢ 4R2 ⎥
y ha
σ II ≡ ⎢ 1 ⎥
⎢ 0 Eα Δ θ 1 − 0 ⎥
le
⎣ 5 r2 ⎦
liv or ec
0 0 0
M
.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. !
"T
X Th
er
4R2
tin
not 1
u ≡ α Δθ r − , 0, 0
5 r
on
.O
⎡
⎤
4R2
C
1
⎢5 α Δ θ 1 + 0 0 ⎥ [17a]
©
⎢ r2
⎥
not ⎢ 1 4R 2 ⎥
ε ≡⎢ 0 α Δθ 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
⎡
⎤
4 R2
⎢ 5 Eα Δ θ −1 + r2 0 0 ⎥
⎢
⎥
not ⎢ 4 R 2 ⎥
σ ≡⎢ 0 Eα Δ θ −1 − 2 0 ⎥ [17b]
⎢ 5 r ⎥
⎣ ⎦
0 0 −Eα Δ 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 Δ θ = Δ θ ∗ ,
ee
uinitial = u f irst phase (Δ θ = Δ θ ∗ ) ,
s gin
ε initial = ε f irst phase (Δ θ = Δ θ ∗ ) , [18]
t d le En
σ initial = σ f irst phase (Δ θ = Δ θ ∗ ) .
r
ba
ge ro or
In fact, the variable Δ θ in [17] is not a total thermal increment but the
eS m
ci
difference in temperature at the moment corresponding to Δ θ ∗ , that is,
f
ra
C d P cs
b
a
Δθ = Δθ −Δθ∗ .
i
[19]
an an n
y ha
Then, considering [17], [18] and [19], the actual displacements, strains
le
liv or ec
.A
[20]
d
uu
er
tin
.O
to that of the second phase. In this way, the total displacement will be null.
C
a
δ1 = ur (r = R, Δ θ = Δ θ ∗ ) = α Δ 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
Δ θ ∗∗ = Δ θ ∗∗ + Δ θ ∗ = + =⇒ Δ θ ∗∗ =
6αR 2αR 3αR
Now, the curve δ −Δ θ can be plotted, where δ is the displacement of the internal
radius of the cylindrical shell.
rs
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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 − Δ 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
s gin
t d le En
r
ba
ge ro or
eS m
Additional hypotheses:
ci
f
ra
C d P cs
1) Weights can be neglected.
b
a
2) Lamé’s constants: λ = μ
i
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
ee
s gin
t d le En
r
6.3 – Two solid cylinders composed of different elastic materials are vertically
ba
ge ro or
eS m
superimposed and confined between two infinitely rigid walls. The cylinders are
ci
f
subjected to the external pressures p and α p (p > 0, α > 0) as shown in the
ra
C d P cs
figure.
b
a
i
an an n
b) Indicate the boundary conditions that need to be applied for the different
M
c) Assuming a constant value α such that the contact surface between the
m
two cylinders does not have a vertical displacement, calculate the inte-
d
uu
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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 Δ 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
Additional hypotheses:
f
ra
C d P cs
Assume the displacements are of the form u = [u (x) , v (y) , w (z) ]T .
b
1)
a
i
2) Linear elastic material
an an n
y ha
3) λ =μ
le
4) The weight of the rail can be neglected.
liv or ec
M
.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
CH.7. PLANE LINEAR
ELASTICITY
Multimedia Course on Continuum Mechanics
Overview
Plane Linear Elasticity Theory Lecture 1
Plane Stress
Simplifying Hypothesis
Strain Field Lecture 2
Constitutive Equation
Displacement Field
The Linear Elastic Problem in Plane Stress Lecture 3
Examples
Plane Strain
Simplifying Hypothesis
Strain Field
Constitutive Equation Lecture 4
Stress Field
The Linear Elastic Problem in Plane Stress
Examples
2
Overview (cont’d)
The Plane Linear Elastic Problem
Lecture 5
Governing Equations
Representative Curves
Isostatics or stress trajectories
Isoclines
Isobars Lecture 6
Maximum shear lines
3
7.1 Plane Linear Elasticity Theory
Ch.7. Plane Linear Elasticity
4
Plane Linear Elasticity
For some problems, one of the principal directions is known
a priori:
Due to particular geometries, loading and boundary conditions
involved.
The elastic problem can be solved independently for this direction.
Setting the known direction as z, the elastic problem analysis is
reduced to the x-y plane PLANE ELASTICITY
7
7.2 Plane Stress
Ch.7. Plane Linear Elasticity
8
Hypothesis on the Stress Tensor
Simplifying hypothesis of a plane stress linear elastic problem:
1. Only stresses “contained in the x-y plane” are not null
σ x τ xy 0
[σ ]xyz ≡ τ xy σ y 0
0 0 0
σ x = σ x ( x, y , t )
σ y = σ y ( x, y , t ) REMARK
τ xy = τ xy ( x, y, t ) The name “plane stress” arises
from the fact that all (not null)
stress are contained in the x-y
plane.
9
Geometry and Actions in Plane Stress
These hypothesis are valid when:
The thickness is much smaller than the typical dimension associated to the
plane of analysis: e << L
The actions b ( x,t ) , u* ( x,t ) and t* ( x,t ) are contained in the plane of
analysis (in-plane actions) and independent of the third dimension, z.
t* ( x,t ) is only non-zero on the
contour of the body’s thickness:
10
Strain Field in Plane Stress
The strain field is obtained from the inverse Hooke’s Law:
2(1 + ν )
(σ x − νσ y )
1
εx = γ xy =
2ε xy = τ xy
ν 1 +ν E E
− Tr ( σ ) 1 +
ε= σ
E E
σ z =0
1
εy =
E
(σ y − νσ x ) γ xz =
2ε xz =
0
τ xz =0
ν
τ yz =0 − (σ x + σ y )
εz = γ yz =
2ε yz =
0
E
As σ x = σ x ( x, y , t )
ε = ε ( x, y , t )
σ y = σ y ( x, y , t )
(σ y − νσ x )
1 E
εy = γ xz =
2ε xz = =σy ε y + νε x
(1 − ν )
0 2
E
ν
− (σ x + σ y )
εz = γ yz =
2ε yz =
0 τ xy =
E
γ xz
E plane
2 (1 + ν )
=C stress
σ x 1 ν 0 ε x
E ν 1
σ y = 0 ε y
τ 1 −ν
2
1 −ν γ
xy xy
0 0
2
= {σ} = {ε}
Constitutive equation
{σ} C stress ⋅ {ε}
plane
in plane stress =
(Voigt’s notation)
12
Displacement Field in Plane Stress
The displacement field is obtained from the geometric equations,
ε ( x, t ) = ∇ S u ( x, t ) . These are split into:
Those which do not affect the displacement u z :
∂u x
ε x ( x, y , t ) =
∂x
∂u y Integration u x = u x ( x, y , t )
ε y ( x, y , t ) =
∂y in . u y = u y ( x, y , t )
∂u x ∂u y
γ xy ( x, y=
, t ) 2ε= +
∂y ∂x
xy
∂u x ( x, y ) ∂uz ∂uz
γ xz ( x, y , t ) = 2ε xz =
∂z ∂x
+ =
∂x
= 0 Contradiction !!!
=0
⇒ uz ( z, t )
∂u y ( x, y ) ∂uz ∂uz
γ yz ( x, y , t ) = 2ε yz = + = = 0
∂ z ∂y ∂y
=0
13
The Linear Elastic Problem in
Plane Stress
The problem can be reduced to the two dimensions of the plane of
analysis.
The unknowns are:
ε x σ x
u x
u ( x, y , t ) ≡ {ε}( x, y, t ) ≡ ε y {σ}( x, y, t ) ≡ σ y
u y γ τ
xy xy
The additional unknowns (with respect to the general problem) are either null,
or independently obtained, or irrelevant:
σ=
z τ=
xz τ=
yz γ=
xz γ=
yz 0 REMARK
This is an ideal elastic problem because it
ν
εz = −
1 −ν
(ε x + ε y ) cannot be exactly reproduced as a particular
case of the 3D elastic problem. There is no
u z ( x, y , z , t ) does not appear guarantee that the solution to u x ( x, y, t ) and
in the problem
u y ( x, y, t ) will allow obtaining the solution to
u z ( x, y, z , t ) for the additional geometric eqns.
14
Examples of Plane Stress Analysis
3D problems which are typically assimilated to a plane stress state are
characterized by:
One of the body’s dimensions is significantly smaller than the other two.
The actions are contained in the plane formed by the two “large” dimensions.
Slab loaded on
Deep beam
the mean plane
15
7.3 Plane Strain
Ch.7. Plane Linear Elasticity
16
Hypothesis on the Displacement Field
Simplifying hypothesis of a plane strain linear elastic problem:
1. The displacement field is
u x
u = u y
0
u x = u x ( x, y , t )
u y = u y ( x, y , t )
17
Geometry and Actions in Plane Strain
These hypothesis are valid when:
The body being studied is generated by moving the plane of analysis
along a generational line.
The actions b ( x,t ) , u* ( x,t ) and t* ( x,t ) are contained in the plane
of analysis and independent of the third dimension, z.
In the central section, considered as the “analysis section” the
following holds (approximately) true:
uz = 0
∂u x
=0
∂z
∂u y
=0
∂z
18
uz = 0
∂u x
19
Stress Field in Plane Strain
Introducing the strain tensor into Hooke’s Law
= (σ λTr ( ε ) 1 + 2Gε ) and
operating on the result yields:
σ x = λ (ε x + ε y ) + 2Gε x τ xy = G γ xy
σ y = λ (ε x + ε y ) + 2Gε y τ xz = G γ xz = 0
= ( λ + 2G ) ε y + λε x
σ z = λ (ε x + ε y ) = v (σ x + σ y ) τ yz = G γ yz = 0
ε x = ε x ( x, y , t )
As ε y = ε y ( x, y , t )
σ = σ ( x, y , t )
ε z = ε z ( x, y , t )
γ xy = γ xy ( x, y , t )
σ x τ xy 0
And the stress tensor
σ z ν (σ x + σ y )
σ ( x, y, t ) ≡ τ xy σ y 0 with=
for plane strain is: 0 0 σ z
20
Constitutive equation in Plane Strain
Introducing the values of the strain tensor into the constitutive equation
and operating on the result yields:
E (1 −ν ) ν
σ x =( λ + 2G ) ε x + λε y = ε + εy
(1 +ν )(1 − 2ν ) 1 −ν
x
σ λ Tr ( ε ) 1 + 2 µ ε σ y =( λ + 2G ) ε y + λε x =
E (1 −ν )
ε +
ν
ε
(1 +ν )(1 − 2ν ) 1 −ν
y x
E
τ xy G=
= γ xy γ xy
2 (1 +ν )
plane
=C strain
ν
1 0
1 −ν
{σ} C strain ⋅ {ε}
plane
σ x ε x =
E (1 − ν ) ν
σ y = 1 0 ε y
τ (1 +ν )(1 − 2ν ) 1 −ν γ Constitutive equation
xy 0 1 − 2ν xy in plane strain
= {σ}
0
2 (1 −ν ) = {ε} (Voigt’s notation)
21
The Lineal Elastic Problem in
Plane Strain (summary)
The problem can be reduced to the two dimensions of the plane of
analysis.
The unknowns are:
ε x σ x
u x
u ( x, y , t ) ≡ {ε}( x, y, t ) ≡ ε y {σ}( x, y, t ) ≡ σ y
u y γ τ
xy xy
The additional unknowns (with respect to the general problem) are either null
or obtained from the unknowns of the problem:
uz = 0
ε=
z γ=
xz γ=
yz τ=
xz τ=
yz 0
σ z ν (σ x + σ y )
=
22
Examples of Plane Strain Analysis
3D problems which are typically assimilated to a plane strain state are
characterized by:
The body is generated by translating a generational section with actions
contained in its plane along a line perpendicular to this plane.
The plane strain hypothesis ( ε=
z γ=
xz γ=
yz 0 ) must be justifiable. This typically
occurs when:
1. One of the body’s dimensions is significantly larger than the other two.
Any section not close to the extremes can be considered a symmetry
plane and satisfies:
uz = 0
∂u x u x
=0
∂z u = u y
∂u y 0
=0
∂z
2. The displacement in z is blocked at the extreme sections.
23
Examples of Plane Strain Analysis
3D problems which are typically assimilated to a plane strain state are:
24
7.4 The Plane Linear Elastic Problem
Ch.7. Plane Linear Elasticity
25
Plane problem
A lineal elastic solid is subjected to body forces and prescribed
traction and displacement
Actions:
t *
( x, y, t )
On Γσ : t = *
* x
y
t ( x , y , t )
x ( x, y , t )
u *
On Γu : u* = *
y (
u x , y , t )
bx ( x, y, t )
On Ω: b=
( )
y
b x , y , t
∂σ x ∂τ xy ∂τ xz ∂ 2u x
+ + + ρ bx =
ρ 2
∂x ∂y ∂z ∂t
∂τ xy ∂σ y ∂τ yz ∂ 2u y
+ + + ρ by =
ρ 2
∂x ∂y ∂z ∂t
∂τ xz ∂τ yz ∂σ z ∂ 2u z
+ + + ρ bz =
ρ 2
∂x ∂y ∂z ∂t
27
Governing Equations
The Plane Linear Elastic Problem is governed by the equations:
2. Constitutive Equation (Voigt’s notation).
Isotropic Linear Elastic Constitutive Equation.
σ ( x, t ) = C : ε
2D
{σ}= C ⋅ {ε}
σ x εx 1 ν 0
With {σ} ≡ σ y , {ε} = ε y and C=
E
2
ν 1 0
τ γ 1 −ν
xy xy 0 0 (1 −ν ) 2
E
E=
PLANE E=E PLANE 1 −ν 2
STRESS ν =ν STRAIN ν
ν =
(1 −ν )
28
Governing Equations
The Plane Linear Elastic Problem is governed by the equations:
3. Geometrical Equation.
Kinematic Compatibility.
1
ε ( x, t )= ∇ S u ( x, t )= (u ⊗ ∇ + ∇ ⊗ u )
2
2D
This is a PDE system of
∂u
εx = x 8 eqns -8 unknowns:
∂x
∂u u ( x,t ) 2 unknowns
εy = y
∂y ε ( x,t ) 3 unknowns
γ=
∂u x ∂u y
+
σ ( x,t ) 3 unknowns
∂y ∂x
xy
Which must be solved in
the 2 × + space.
29
Boundary Conditions
Boundary conditions in space
Affect the spatial arguments of the unknowns
Are applied on the contour Γ of the solid,
which is divided into:
Prescribed displacements on Γ u :
x ( x, y , t )
u *
= u *
u = *
* x
y
u = u *
y ( x , y , t )
30
Boundary Conditions
INTIAL CONDITIONS (boundary conditions in time)
Affect the time argument of the unknowns.
Generally, they are the known values at t = 0 :
Initial displacements:
ux
u ( x, =
y ,0 ) = 0
u
y
Initial velocity:
∂u ( x, y , t ) u x v x
≡ u ( x, y ,0 ) =
= = v 0 ( x, y )
∂t t =0
u
y y v
31
Unknowns
The 8 unknowns to be solved in the problem are:
1
u x εx γ xy
2 σ x τ xy
u ( x, y , t ) = ε ( x, y , t ) ≡ σ ( x, y , t ) ≡
u y 1 γ τ σ
εy xy y
2 xy
σz ν σx +σ y
PLANE STRAIN = ( )
32
7.5 Representative Curves
Ch.7. Plane Linear Elasticity
33
Introduction
Traditionally, plane stress states where graphically represented
with the aid of the following contour lines:
Isostatics or stress trajectories
Isoclines
Isobars
Maximum shear lines
Others: isochromatics, isopatchs, etc.
34
Isostatics or Stress Trajectories
System of curves which are tangent to the principal axes of stress
at each material point .
They are the envelopes of the principal stress vector fields.
There will exist two (orthogonal) families of curves at each point:
Isostatics σ 1 , tangents to the largest principal stress.
Isostatics σ 2 , tangents to the smallest principal stress.
REMARK
The principal stresses are orthogonal
to each other, therefore, so will the two
families of isostatics orthogonal to
each other.
35
Singular and Neutral Points
Singular point: characterized by the stress state
σx =σy
τ xy = 0
Neutral point: characterized by the stress state
σ=
x σ=
y τ=
xy 0
Mohr’s Circle of
REMARK
a singular point
In a singular point, all directions
are principal directions. Thus, in
singular points isostatics tend to
Mohr’s Circle of loose their regularity and can
a neutral point abruptly change direction.
36
Differential Equation of the Isostatics
Consider the general equation of an isostatic curve: y = f ( x )
2τ xy 2 tg α
=
tg ( 2α ) =
σ x − σ y 1 − tg 2 α
dy
α
tg= = y′
dx
2τ xy 2 y′ σ x −σ y
= ( y′ ) + y′ − 1 =0
2
σ x −σ y 1 − ( y′ ) τ xy
2
37
Isoclines
Locus of the points along which the principal stresses are in the
same direction.
The principal stress vectors in all points of an isocline are parallel to
each other, forming a constant angle θ with the x-axis.
38
Equation of the Isoclines
To obtain the general equation of an isocline with angle θ , the
principal stress σ 1 must form an angle α = θ with the x-axis:
ϕ ( x, y )
2τ xy For each value of θ , the equation of
Algebraic equation
tg ( 2θ ) =
of the isoclines σ x −σ y the family of isoclines parameterized
in function of θ is obtained:
y = f ( x, θ )
REMARK
Once the family of isoclines is
known, the principal stress
directions in any point of the
medium can be obtained and,
thus, the isostatics calculated.
39
Maximum shear lines
Envelopes of the maximum shear stress (in modulus) vector fields.
They are the curves on which the shear stress modulus is a maximum.
Two planes of maximum shear stress correspond to each material
point, τ max and τ min .
These planes are easily determined using Mohr’s Circle.
REMARK
The two planes form a 45º
angle with the principal
stress directions and, thus,
are orthogonal to each
other. They form an angle
of 45º with the isostatics.
42
Equation of the maximum shear lines
Consider the general equation of a slip line y = f ( x ), the relation
2τ xy π π 1
tan 2α = and β= α + tan ( 2 β ) =
tan 2α − = −
σ x −σ y 4 2 tan 2α
Then,
1 σ x −σ y 2 tan β
tan ( 2 β ) =
− =
− =2
tan ( 2α ) 2τ xy 1 − tan β σ x −σ y 2 y′
− =2
tan (=
dy not
β ) = y′
2τ xy 1 − ( y′ )
dx
4τ xy
( y′ ) − y′ − 1 =0
2
σ x −σ y
43
Equation of the maximum shear lines
Solving the 2nd order eq.: φ ( x, y ) Known this function,
the eq. can be integrated
Differential 2 to obtain a family of
2τ xy 2τ xy
equation of the y'= ±
σ −σ + 1 curves of the type:
slip lines σ x −σ y x y
=y f ( x) + C
44
Chapter 7
Plane Linear Elasticity
rs
ee
s gin
7.1 Introduction
t d le En
r
As seen in Chapter 6, from a mathematical point of view, the elastic problem
ba
ge ro or
eS m
consists in a system of PDEs that must be solved in the three dimensions of
ci
space and in the dimension associated with time R3 × R+ . However, in certain
f
ra
C d P cs
situations, the problem can be simplified so that it is reduced
to two dimensions
b
a
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
analysis is reduced to the x-y plane and, hence, the name plane elasticity used to
uu
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
s gin
the plane of analysis x-y,
eL. (7.3)
t d le En
b) The actions (body forces b (x,t), prescribed displacements u∗ (x,t) and trac-
r
ba
ge ro or
tion vector t∗ (x,t) ) are contained within the plane of analysis x-y (its z-
eS m
ci
f
component is null) and, in addition, do not depend on the third dimension,
ra
C d P cs
⎡ ⎤ ⎡ ⎤
b
a
bx (x, y,t) u∗x (x, y,t)
i
an an n
not ⎢ ⎥ ⎢ ⎥
b≡⎢ ⎥, ⎢ u∗ (x, y,t) ⎥ ,
y ha
Γ ∗ not
≡
⎣ by (x, y,t) ⎦ u : u ⎣ y ⎦
le
liv or ec
0 −
M
.A
⎡ ⎤ (7.4)
tx∗ (x, y,t)
m
not ⎢ ∗ ⎥
d
⎣y ⎦
e
X Th
−
er
tin
c) The traction vector t∗ (x,t) is only non-zero on the boundary of the piece’s
on
.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
ee
s gin
Figure 7.1: Example of a plane stress state.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Remark 7.1. The piece with the actions defined by (7.4) and (7.5) is
C d P cs
b
a
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 ⎢ ⎥⎢ ⎥ ⎢ ⎥
d
n≡⎣ 0 ⎦ , σ · n ≡ ⎣ τxy σy 0 ⎦ ⎣ 0 ⎦ = ⎣ 0 ⎦ ,
uu
e
±1 ±1
X Th
0 0 0 0
er
tin
• Edge Γσe
on
.O
⎡ ⎤ ⎡ ⎤⎡ ⎤ ⎡ ⎤
C
not ⎢ ⎥ not ⎢ ⎥⎢ ⎥ ⎢
n ≡ ⎣ ny ⎦ , σ (x, y,t)·n ≡ ⎢ τ σ 0 ⎥ ⎣ ny ⎦ = ⎣ ty (x, y,t) ⎥
⎦,
⎣ xy y ⎦
0 0 0 0 0 0
1 The 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.
s gin
t d le En
r
ba
ge ro or
7.2.1 Strain Field. Constitutive Equation
eS m
ci
f
ra
Consider now the linear elastic constitutive equation (6.24),
C d P cs
b
a
i
ν 1+ν ν 1
an an n
ε =− σ)1+
Tr (σ σ)1+
σ = − Tr (σ 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
not ⎢ ⎥
ε (x, y,t) ≡ ⎢ 1 γxy εy 0 ⎥ with εz = − (εx + εy ) (7.9)
⎣2 ⎦ 1−ν
0 0 εz
rs
ee
E E
σx = (εx + νεy ) , σy = (εy + νεx ) ,
s gin
1 − ν2 1 − ν2 (7.10)
E
t d le En
and τxy = γxy ,
2 (1 + ν)
r
ba
ge ro or
eS m
which can be rewritten as
ci
⎡ ⎤⎡ ⎤
⎡ ⎤ f
ra
C d P cs
σx 1 ν
⎥ ⎢ εx ⎥
0
b
a
⎢ ⎥ ⎢
i
⎢σ ⎥= E ⎢ ⎥⎢ ⎥
an an n
plane
⎣ y ⎦ 1 − ν2 ⎢ ν 1 0 ⎥ ⎣ εy ⎦ =⇒ σ } = C stress · {εε } .
{σ
⎣ ⎦
y ha
1−ν
τxy γxy
le
0 0
liv or ec
2
M
σ}
{σ {εε }
.A
plane
C stress
(7.11)
m
d
uu
e
X Th
.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 ux = ux (x, y,t)
∂ uy =⇒ . (7.13)
εy (x, y,t) = ⎪ uy = uy (x, y,t)
∂y ⎪
⎪
∂ 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
σz = 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
ee
s gin
In the central section (which is a plane of symmetry with respect to the z-axis)
the conditions
t d le En
∂ ux ∂ uy
uz = 0 , =0 =0
r
and (7.19)
ba
∂z ∂z
ge ro or
eS m
ci
f
ra
are satisfied and, thus, the displacement field in this central section is of the form
C d P cs
b
a
⎡ ⎤
i
ux (x, y,t)
an an n
not ⎢ ⎥
y ha
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 ∂ u y ∂ uz
rs
γxy (x, y,t) = + , γyz (x, y,t) = + =0.
∂y ∂x ∂z ∂y
ee
s gin
Therefore, the structure of the strain tensor is3
t d le En
⎡ 1 ⎤
εx γxy 0
r
⎢ 2 ⎥
ba
ge ro or
not ⎢ ⎥
eS m
ε (x, y,t) ≡ ⎢ 1 γxy εy 0 ⎥ . (7.22)
ci
⎣2 ⎦
f
ra
C d P cs
b
a
0 0 0
i
an an n
y ha
le
Consider now the lineal elastic constitutive equation (6.20)
liv or ec
M
.A
er
(7.24)
.O
C
Considering (7.21) and (7.24), one concludes that stresses do not depend on the
σ = σ (x, y,t)). On the other hand, the stress σz in (7.24)
z-coordinate either (σ
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
σx = (λ + 2G) εx + λ εy = εx + εy ,
s gin
(1 + ν) (1 − 2ν) 1−ν
t d le En
E (1 − ν) ν (7.27)
σy = (λ + 2G) εy + λ εx = εy + εx ,
(1 + ν) (1 − 2ν) 1−ν
r
ba
ge ro or
eS m
ci
E
and τxy = G γxy = f
γxy .
ra
2 (1 + ν)
C d P cs
b
a
i
an an n
⎡ ν ⎤
⎡ ⎤ ⎡ ⎤
le
liv or ec
1 0
σx ⎢ 1−ν ⎥ εx
⎢ ⎥ (1 − ν) ⎢ ν ⎥ ⎢ ⎥
M
.A
⎢ σy ⎥ = E ⎢ ⎥ ⎢ εy ⎥ ⇒
1 0
⎣ ⎦ (1 + ν) (1 − 2ν) ⎢ 1 − ν ⎥ ⎣ ⎦
⎣ 1 − 2ν ⎦
m
τxy γxy
d
0 0
2 (1 − ν)
uu
e
(7.28)
X Th
σ}
{σ {εε }
er
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) ≡ σ (x, y,t)} ≡ ⎣ σy ⎦ ,
, {εε (x, y,t)} ≡ ⎣ εy ⎦ and {σ (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
ee
s gin
Figure 7.4: The plane linear elastic problem.
t d le En
r
ba
ge ro or
eS m
in which the additional unknowns with respect to the general problem are either
ci
f
null or can be calculated in terms of those in (7.29),
ra
C d P cs
b
a
uz = 0 , εz = γxz = γyz = τxz = τyz = 0 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 } ≡ ⎣ σy ⎦ ,
{σ {εε } ≡ ⎣ εy ⎦ ; σ } = C · {εε } ,
{σ (7.32)
τxy γxy
where the constitutive matrix C can be written in a general form, from (7.11)
and (7.28), as
Ē = E
rs
⎡ ⎤ Plane stress
ee
1 ν̄ 0 ν̄ = ν
⎢ ⎥ ⎧
s gin
not Ē ⎢ ⎥
C≡ ⎢ ν̄ 1 0 ⎥ ⎪
⎪ E (7.33)
1 − ν̄ 2 ⎣ ⎨ Ē =
1 − ν̄ ⎦
t d le En
0 0 Plane strain 1−ν 2
⎪
⎩ ν̄ = ν
⎪
2
r
ba
ge ro or
1−ν
eS m
ci
f
ra
c) Geometric equation
C d P cs
b
a
i
an an n
∂ ux ∂ uy ∂ ux ∂ uy
y ha
εx = , εy = , γxy = + (7.34)
∂x ∂y ∂y ∂x
le
liv or ec
M
.A
d
uu
e
er
Γu : u∗ ≡ Γσ : t∗ ≡
tin
not not
,
u∗y (x, y,t) ty∗ (x, y,t)
on
.O
(7.35)
C
σx τxy nx
©
t∗ = σ · n ,
not not
σ≡ , n≡
τxy σy ny
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 )
1−ν
s gin
(7.38)
Plane strain → σz = ν (σx + σy )
t d le En
r
ba
ge ro or
eS m
7.5 Problems Typically Assimilated to Plane Elasticity
ci
f
ra
C d P cs
7.5.1 Plane Stress
b
a
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
stress state. The slab loaded on its mean plane and the deep beam of Figure 7.5
M
.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)
s gin
∂z ∂z
which result in the initial condition of the plane strain state (7.17),
t d le En
⎡ ⎤ ⎡ ⎤
ux ux (x, y,t)
r
ba
not ⎢ ⎥ ⎢ ⎥
ge ro or
eS m
u ≡ ⎣ uy ⎦ = ⎣ uy (x, y,t) ⎦ . (7.40)
ci
f
ra
uz 0
C d P cs
b
a
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.
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
s gin
t d le En
Considering the definition of the envelope of a vector field, isostatics are, at each
r
point, tangent to the two principal directions and, thus, there exist two families
ba
ge ro or
eS m
of isostatics:
ci
f
− Isostatics σ1 , tangent to the direction of the largest principal stress.
ra
C d P cs
− Isostatics σ2 , tangent to the direction of the smallest principal stress.
b
a
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
σx = σy and τxy = 0
on
.O
and its Mohr’s circle is a point on the axis σ (see Figure 7.11).
C
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
Figure 7.11: Singular and neutral points.
ci
f
ra
C d P cs
b
a
i
an an n
y ha
Remark 7.3. All directions in a singular point are principal stress di-
le
rections (the pole is the Mohr’s circle itself, see Figure 7.11). Conse-
liv or ec
.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) = = ⎬
σx − σy 1 − tan2 α ⇒ 2τxy = 2y ⇒
dy not ⎪
⎪ σx − σy 1 − (y )2
tan α = =y ⎪
⎭ (7.41)
dx
σx − σy
(y )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 y , the differential equation of
s gin
the isostatics is obtained.
t d le En
Differential σx − σy 2
r
σ − σ
ba
x y
ge ro or
→ y =− ± +1
eS m
equation (7.42)
2τxy 2τxy
ci
of the isostatics
f
ra
C d P cs
ϕ (x, y)
b
a
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
τxy = 3x y = 0 =⇒
2
⎪
⎪
⎪
⎪ 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),
2
dy σx − σy σx − σy
y = =− ± +1 ,
dx 2τxy 2τxy
rs
⎪
⎪ dy x
⎨ = x2 − y2 = C1
ee
dx y =⇒ integrating =⇒ .
⎪ xy = C2
s gin
⎪
⎪ dy −y
⎩ =
dx x
t d le En
Therefore, the isostatics are two families of equilateral hyperboles orthogo-
r
ba
ge ro or
nal to each other.
eS m
ci
f
On the line of singular points x = 0 (which divides the plate in two regions)
ra
C d P cs
the isostatics will brusquely change their slope. To identify the family of
b
a
isostatics σ1 , consider a point in each region:
i
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).
s gin
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
r
ba
ge ro or
direction, that is,
eS m
ci
f
ra
C d P cs
b
a
Algebraic equation 2τxy
i
tan (2θ ) =
an an n
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
s gin
t d le En
Definition 7.4. Isobars are the locus of points in the plane of analy-
r
ba
sis with the same value of principal stress σ1 (or σ2 ).
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
Two families of isobars will cross at each point of the plane of analysis: one
an an n
corresponding to σ1 and another to σ2 . Note that the isobars depend on the value
y ha
.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
⎧
⎪
⎪ 2
on
.O
⎪
⎪ σ + σ σ − σ
⎪
⎪ σ1 =
x y
+
x y
+ τxy
2 = const. = c
⎪
⎪ 1
C
⎪ 2 2
⎪
Algebraic ⎪
©
⎪
⎨
equation ϕ1 (x, y)
(7.46)
of the ⎪⎪
isobars ⎪⎪
⎪ σx + σ y σx − σy 2
⎪
⎪ σ2 = − + τxy
2 = const. = c
⎪
⎪ 2 2
2
⎪
⎪
⎪
⎩
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
ee
s gin
7.6.4 Maximum Shear Stress or Slip Lines
t d le En
r
ba
ge ro or
eS m
Definition 7.5. Maximum shear stress lines or slip lines are the en-
ci
f
velopes of the directions that, at each point, correspond with the
ra
C d P cs
b
maximum value (in modulus) of the shear (or tangent) stress.
a
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
d
uu
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.
s gin
t d le En
where α is the angle formed by the principal stress direction σ1 with the
r
ba
horizontal direction. Consequently, considering the general equation of a slip
ge ro or
eS m
line, y = f (x), the expression (7.48) and the relation tan (2α) = 2τxy / (σx − σy )
ci
f
ra
yields
C d P cs
⎫
b
a
1 σx − σy 2 tan β ⎪
i
tan (2β ) = − = = ⎬
an an n
dy not ⎪
⎭
tan β = =y
le
liv or ec
dx (7.49)
M
.A
σx − σy 2y 4τxy
=⇒ (y )2 − y − 1 = 0.
m
− =
d
2τxy 1 − (y )2 σx − σy
uu
e
X Th
Solving the second-order equation in (7.49) for y provides the differential equa-
er
tin
.O
C
Differential 2
2τ 2τxy
©
equation of the xy
y =− ± +1 (7.50)
max. shear stress σx − σy σx − σy
or slip lines
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
ee
s gin
Solution
t d le En
a) A singular point is defined as:
r
ba
ge ro or
eS m
ci
f
ra
"
C d P cs
The stress state is
b
a
σ1 = σ2 represented by a point.
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
which can take any value, the principal stress direction will form an angle θ
m
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
ee
s gin
Solution
t d le En
1) The Mohr’s circle for the stress state σx = 0 ; σy = b > 0 ; τxy = 0 is:
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
ee
s gin
t d le En
2) The Mohr’s circle for the stress state σx = 0 ; σy = 0 ; τxy = m y , m > 0 is:
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.
s gin
t d le En
7.2 – Plot the isostatics in the transversal section of the cylindrical shell shown
r
below. Assume a field of the form:
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
CH.8. PLASTICITY
Multimedia Course on Continuum Mechanics
Overview
Introduction Lecture 1
Previous Notions Lecture 2 Lecture 3
Principal Stress Space
Lecture 4
Normal and Shear Octahedral Stresses
Stress Invariants
Lecture 5
Effective Stress
Rheological Friction Models
Elastic Element
Frictional Element
Elastic-Frictional Model
2
Overview (cont’d)
Rheological Friction Models (cont’d)
Frictional Model with Hardening
Elastic-Frictional Model with Hardening
Phenomenological Behaviour
Notion of Plastic Strain
Notion of Hardening Lecture 6
Bauschinger Effect
Elastoplastic Behaviour
1D Incremental Theory of Plasticity
Additive Decomposition of Strain Lecture 7
Hardening Variable
Yield Stress, Yield Function and Space of Admissible Stresses Lecture 8 Lecture 9
Constitutive Equation Lecture 9
Elastoplastic Tangent Modulus
Lecture 10
Uniaxial Stress-Strain Curve
3
Overview (cont’d)
3D Incremental Theory of Plasticity
Additive Decomposition of Strain
Hardening Variable Lecture 11
Yield Function
Loading - Unloading Conditions and Consistency Conditions
Constitutive Equation
Lecture 12
Elastoplastic Constitutive Tensor
Yield Surfaces
Lecture 13
Von Mises Criterion
Tresca Criterion
Mohr-Coulomb Criterion Lecture 14
Drucker-Prager Criterion
4
8.1 Introduction
Ch.8. Plasticity
5
Introduction
A material with plastic behavior is characterized by:
A nonlinear stress-strain relationship.
The existence of permanent (or plastic) strain during a
loading/unloading cycle.
Lack of unicity in the stress-strain relationship.
6
Previous Notions
PRINCIPAL STRESSES
Regardless of the state of stress, it is always possible to choose a special
set of axes (principal axes of stress or principal stress directions) so
that the shear stress components vanish when the stress components are
referred to this system.
The three planes perpendicular to the principle axes are the principal
planes.
The normal stress components in the principal planes are the principal
stresses. σ 33
x3 σ 31 σ 32 x3
σ 13 x3′
σ 1 0 0 σ 23
σ 11 σ 12 σ 21 σ x1′ σ3
[σ ] = 0 σ 2 0 22
σ1
0 0 σ 3 σ2
x1 x1
x2 x2
x2′
7
Previous Notions
PRINCIPAL STRESSES
The Cauchy stress tensor is a symmetric 2nd order tensor so it will diagonalize
in an orthonormal basis and its eigenvalues are real numbers.
Computing the eigenvalues λ and the corresponding eigenvectors v:
σ ⋅ v λv
= [σ −=
λ 1] ⋅ v 0 σ 11 − λ σ 12 σ 13
not
det [ σ − λ 1] = =
σ − λ 1 σ 12 σ 22 − λ =
σ 23 0
σ 13 σ 23 σ 33 − λ
INVARIANTS
σ 33
λ 3 − I1λ 2 − I 2 λ − I 3 =
characteristic x3 σ 31 σ 32 x3
0 equation σ 13 x3′
σ 23
λ1 ≡ σ 1 σ 11 σ 12 σ 21 σ x1′ σ3
σ1
22
λ2 ≡ σ 2
λ3 ≡ σ 3 σ2
x1 x1
x2 x2
x2′
8
Previous Notions
STRESS INVARIANTS
Principal stresses are invariants of the stress state.
They are invariant w.r.t. rotation of the coordinate axes to which the
stresses are referred.
The principal stresses are combined to form the stress invariants I :
I1 = Tr ( σ ) = σ ii = σ 1 + σ 2 + σ 3 REMARK
I 2 =( σ : σ − I12 ) =
1
− (σ 1σ 2 + σ 1σ 3 + σ 2σ 3 ) The I invariants are obtained
2 from the characteristic equation
I 3 = det ( σ ) of the eigenvalue problem.
These invariants are combined, in turn, to obtain the invariants J:
J=
1 I=
1 σ ii REMARK
The J invariants can be
J2 =
2
(
1 2
I1 + 2 I 2 ) = σ ijσ ji = ( σ : σ )
1
2
1
2 expressed the unified form:
1 = Ji
1
( )
Tr σ i i ∈ {1, 2,3}
J=
3
3
( I1 + 3I1 I 2 + 3I3=) 3 Tr ( σ ⋅ σ ⋅ σ=) 3 σ ijσ jkσ ki
1 3 1
i
9
Previous Notions
SPHERICAL AND DEVIATORIC PARTS OF THE STRESS TENSOR
Given the Cauchy stress tensor σ and its principal stresses, the following is
defined:
Mean stress
1 1 1
σ m= Tr (σ )= σ ii = (σ 1 + σ 2 + σ 3 )
3 3 3
REMARK
Mean pressure
In a hydrostatic state of stress, the
1
p= −σ m =− (σ 1 + σ 2 + σ 3 ) stress tensor is isotropic and, thus,
3 its components are the same in
any Cartesian coordinate system.
A spherical or hydrostatic
As a consequence, any direction
state of stress: σ 0 0 is a principal direction and the
σ=
1 σ=
2 σ3 σ ≡ 0
σ 0 = σ 1 stress state (traction vector) is the
0 0 σ same in any plane.
10
Previous Notions
SPHERICAL AND DEVIATORIC PARTS OF THE STRESS TENSOR
The Cauchy stress tensor σ can be split into: σ = σ sph + σ ′
The spherical stress tensor:
Also named mean hydrostatic stress tensor or volumetric stress tensor or
mean normal stress tensor.
Is an isotropic tensor and defines a hydrostatic state of stress.
Tends to change the volume of the stressed body
1 1
σ=
sph : σ=m 1 Tr ( σ=) 1 σ ii 1
3 3
The stress deviatoric tensor:
Is an indicator of how far from a hydrostatic state of stress the state is.
Tends to distort the volume of the stressed body
σ ′ dev
= = σ σ − σm 1
11
Previous Notions
STRESS INVARIANTS OF THE STRESS DEVIATORIC TENSOR
The stress invariants of the stress deviatoric tensor:
=I1′ Tr =( σ′ ) 0
=I 2′
1
2
(
σ′ : σ′ − I12 )
′ = (σ ij′σ ′jkσ ki′ )
1
I 3′ = det ( σ′ ) = σ 11′ σ 22
′ σ 33
′ + 2σ 12′ σ 23
′ σ 13′ − σ 12′2σ 33
′ − σ 23
′2σ 11′ − σ 13′2σ 22
3
J 2′ =
1 2
2
( )
I1′ + 2 I 2′ =
1
I 2′ = ( σ′ : σ′ )
2
J 3′ =
3
( )
I1′ + 3I1′I 2′ + 3I 3′ = I 3′ = Tr ( σ′ ⋅ σ′ ⋅ σ′ ) = (σ ij′σ ′jkσ ki′ )
1 3 1
3
1
3
12
Previous Notions
EFFECTIVE STRESS
The effective stress or equivalent uniaxial stress σ is the scalar:
3 3
σ= 3J=
'
2 σ′ij σ′ij= σ´: σ´
2 2
13
Example
Calculate the value of the equivalent uniaxial stress for an uniaxial state of
stress defined by:
E, ν
y
σ u σ 0x0
σx
σ ≡ 0 0 0
σu σu
0 0 0
x
14
σ u 0 0
σ ≡ 0 0 0
Example - Solution 0 0 0
1 σu σ u
Mean stress: =
σm =
Tr (σ ) 0 0
3 3
σ m 0 0 3
σ
Spherical and deviatoric parts σ sph ≡ 0 σ m 0 =0 u
0
3
0 0 σ m
of the stress tensor: 0 σ u
0
3
2
3 σu 0 0
σ u − σ m 0 0
≡ 0 0 = 0 0
1
σ′ =−
σ σ sph −σ m − σu
3
0 0 −σ m
0 1
0 − σu
3
3 3 2 4 1 1 32
σ= σ′ij σ′ij= σu ( + + =) σu σ = σu
2 2 9 9 9 23
15
8.2 Principal Stress Space
Ch.8. Plasticity
16
Principal Stress Space
The principal stress space or Haigh–Westergaard stress space is
the space defined by a system of Cartesian axes where the three
spatial axes represent the three principal stresses for a body
subject to stress:
σ1 ≥ σ 2 ≥ σ 3
17
Octahedral plane
Any of the planes perpendicular to the hydrostatic stress axis is a
octahedral plane. 1
1
Its unit normal is n = 1 .
3
1 σ1 ≥ σ 2 ≥ σ 3
18
Normal and Shear Octahedral Stresses
Consider the principal stress space:
The normal octahedral stress is defined as:
1/ 3
3σ oct = OA = OP ⋅ n= [σ 1 , σ 2 , σ 3 ] 1/ 3 =
1/ 3
3
= (σ 1 + σ 2 + σ=
3) 3 σm
3
I1
σ=
oct σ=
m
3
19
Normal and Shear Octahedral Stresses
Consider the principal stress space:
The shear or tangential octahedral
stress is defined as:
3τ oct = AP
= AP = OP − OA = (σ 12 + σ 22 + σ 32 ) −
2 2 2
3τ 2
oct
1
− (σ 1 + σ 2 + σ 3 ) =
2
2 J 2'
3 Alternative forms of τ oct :
1/2
1 2 1 2
τ= σ + σ 2
+ σ 2
− ( σ + σ + σ )
3
oct 1 2 3 1 2 3
3
2
[ J2 ] ( ) ( ) ( )
1 2 1/2
′
τ oct
2 2
=
12
τ oct
= σ − σ + σ − σ + σ − σ
3 3
1 2 2 3 1 3
3
20
Normal and Shear Octahedral Stresses
In a pure spherical stress state:
1
σ= σ 1 → σm= 3σ → σ esf= σ=
1 σ
3
σ′ =σ − σ esf = 0 J 2′ = 0
A pure spherical stress state is
τ oct = 0 located on the hydrostatic stress axis.
σ = σ′ σ m Tr
= = (σ′) 0
(σ ) Tr= σ oct = 0
21
Stress Invariants
Any point in space is unambiguously defined by the three
invariants:
The first stress invariant I1 characterizes the distance from the origin to
the octahedral plane containing the point.
22
Projection on the Octahedral Plane
The projection of the principal stress space on the octahedral
plane results in the division of the plane into six “sectors”:
These are characterized by the different principal stress orders.
23
8.4 Phenomenological Behaviour
Ch.8. Plasticity
38
Notion of Plastic Strain
PLASTIC STRAIN
ε= ε e + ε p
elastic limit: σ e
LINEAR ELASTIC
BEHAVIOUR
σ = Eεe
39
Bauschinger Effect
Also known as kinematic hardening.
σf
σ e + Kε
K
σe
−σ e + K ε
K
−σ e
41
Elastoplastic Behaviour
Considering the phenomenological behaviour observed,
elastoplastic materials are characterized by:
42
8.5 1D Incremental Plasticity Theory
Ch.8. Plasticity
43
Introduction
The incremental plasticity theory is a mathematical model used
to represent the evolution of the stress-strain curve in an
elastoplastic material.
Developed for 1D but it can be generalized for 3D problems.
REMARK
This theory is
developed under
the hypothesis of
infinitesimal strains.
44
Additive Decomposition
of Strain
Total strain can be split into an elastic (recoverable) part, ε ,
e
and an inelastic (unrecoverable) one, ε :
p
σ
ε= ε + ε
e p
where ε = e
E elastic modulus or
Young modulus
Also,
dσ
dε dε + dε
= e p
where dε =e
45
Hardening Variable
The hardening variable, α , is defined as: REMARK
The sign ( • ) function is:
dα = sign (σ ) d ε p
46
Yield Stress and Hardening Law
Stress value, σ f , threshold for the material exhibiting plastic
behaviour after elastic unloading + elastic loading
It is considered a material property.
For ε p= α= 0 σ f= σ e HARDENING LAW
σ=
f σ e + H ′α
σ = σ f (α )
dσ f = H ′dα
σ ≤ σ f (α )
47
Yield Function
The yield function, F (σ , α ) , characterizes the state of the material:
F (σ , α ) ≡ σ − σ f (α )
Space of
F (σ , α ) < 0 F (σ , α ) = 0 admissible
stresses
ELASTIC STATE ELASTO-PLASTIC STATE
{
σ ∈ R F (σ , α ) < 0
Eσ := } {
∂Eσ := σ ∈ R F (σ , α ) = 0 }
ELASTIC DOMAIN YIELD SURFACE
INITIAL ELASTIC
DOMAIN: {
Eσ0 := σ ∈ R F (σ , 0 ) ≡ σ − σ e < 0 }
48
Space of Admissible Stresses
Any admissible stress state must belong to the space of
admissible stresses, Eσ (postulate):
Eσ = Eσ ∂Eσ =
{σ ∈ R F (σ , α ) ≤ 0}
=
F (σ , α ) ≡ σ − σ f (α )
Space of
admissible
stresses
REMARK
Eσ ≡ −σ f (α ), σ f (α )
49
Constitutive Equation
The following situations are defined:
ELASTIC REGIME
σ ∈ Eσ dσ = E d ε
E H′
1 EH ′
=dσ = dε dε
1 1
+ E + H′
E H′ E ep
ELASTOPLASTIC EH ′
dσ = E d ε
ep E =
ep
TANGENT MODULUS E + H′
52
Uniaxial Stress-Strain Curve
Following the constitutive equation defined
ELASTOPLASTIC
REGIME
dσ = E ep d ε
ELASTIC
REGIME
dσ = E d ε
REMARK
Plastic strain is generated only
during the plastic loading process.
53
Role of the Hardening Modulus
The value of the hardening modulus, H ′, determines the following
situations:
EH ′
E =
ep
E + H′
H′ > 0
Linear elasticity Plasticity with
strain hardening H′ < 0
Plasticity with
strain softening
Perfect plasticity
55
Plasticity in Real Materials
In real materials, the stress-strain curve shows a combination of
the three types of hardening modulus.
H′ < 0
H′ = 0 H′ > 0
56
8.6 3D Incremental Theory
Ch.8. Plasticity
57
Introduction
The 1D incremental plasticity theory can be generalized to a
multiaxial stress state in 3D.
58
ε= ε e + ε p
Additive Decomposition
1D → e σ
ε = E
of Strain
Total strain can be split into an elastic (recoverable) part, ε ,
e
and an inelastic (unrecoverable) one, ε :
p
ε= ε e + ε p where ε e = C −1 : σ
constitutive elastic
(constant) tensor
Also,
dε dε + dε
= e p
where d ε e = C −1 : dσ
59
1D → d ε p =λ sign (σ )
Hardening Variable =d α
dα = λ with α ∈ [ 0, ∞ )
∂σ
60
Yield Function 1D → F (σ , α ) ≡ σ − σ f (α )
Yield stress
F (σ , α ) < 0 F (σ , α ) = 0
ELASTIC STATE ELASTOPLASTIC STATE
Eσ : {σ }
F (σ , α ) < 0 =
∂Eσ : {σ (σ , α ) 0
F= }
ELASTIC DOMAIN YIELD SURFACE
Space of
INITIAL ELASTIC
=
DOMAIN: E 0
σ : {σ F (σ , 0 ) < 0 } admissible =
stresses
Eσ Eσ ∂Eσ
61
Loading-Unloading Conditions and
Consistency Condition
Loading/unloading conditions (also known as Karush-Kuhn-
Tucker conditions):
λ≥0 ; F ( σ, α ) ≤ 0 ; λ F ( σ, α ) =
0
Consistency conditions:
For F ( σ, α ) =
0 → λ dF ( σ, α ) =
0
∂G (σ , α ) ELASTOPLASTIC
F=0; dF < 0 λ= 0; d ε = λ p
=0 ELASTIC UNLOADING
∂σ
∂G (σ , α ) ELASTOPLASTIC
= λ =
0; d ε p
λ = 0 NEUTRAL LOADING
∂σ
=F 0;=dF 0
λ > 0; d ε= ∂G (σ , α ) ELASTOPLASTIC
p
λ ≠0
∂σ LOADING
=F 0; dF > 0 IMPOSSIBLE
62
Constitutive Equation
The following situations are defined:
ELASTIC REGIME ( F < 0)
σ ∈ Eσ dσ = C : d ε
=
ELASTOPLASTIC REGIME – PLASTIC and dF (σ , α ) 0)
LOADING ( F 0=
σ ∈ ∂Eσ
dσ = C ep : d ε
dF (σ , α ) = 0
ELASTOPLASTIC
CONSTITUTIVE TENSOR
∂G ∂F
Cijpq C rskl
∂σ pq ∂σ rs
Cijkl
ep
=
Cijkl − i, j , k , l , p, q, r , s ∈ {1, 2,3}
∂F ∂G
H′+ C pqrs
∂σ pq ∂σ rs
64
8.7 Failure Criteria: Yield Surfaces
Ch.8. Plasticity
65
Introduction
The initial yield surface, ∂Eσ0 , is the external boundary of the initial
elastic domain Eσ for the virgin material
0
The state of stress inside the yield surface is elastic for the virgin material.
When in a deformation process, the stress state reaches the yield surface, the
virgin material looses elasticity for the first time: this is considered as a failure
criterion for design. Subsequent stages in the deformation process are not
considered.
J 3′ =
3
(
1 3
) 1
3
1
(
I1′ + 3I1′I 2′ + 3I 3′ = I 3′ = Tr ( σ′ ⋅ σ′ ⋅ σ′ ) = σ ij′σ ′jkσ ki′
3
)
The elastoplastic behavior will be isotropic.
67
F ( σ ) ≡ φ( σ ) − σ e = 0
Von Mises Criterion
The yield surface is defined as: REMARK
The Von Mises criterion
F (σ ) ≡ σ ( ) − σ e =
0 depends solely on the
second deviator stress
Where σ (σ ) = 3J 2′ is the effective stress. invariant.
(often termed the Von-Mises stress)
2
The shear octahedral stress is, by definition, τ oct = [ J2 ] .
′
12
3
Thus, the effective stress is rewritten:
3 3
σ (σ )
3
[ J 2′ ] = τ oct = 3 τ oct
= τ oct
12
2 2 2
3
And the yield surface is given by: F (σ ) ≡ τ oct − σ e =
0
2
68
3
Von Mises Criterion F (σ ) ≡
2
τ oct − σ e =
0
F (σ) ≡ F ( J 2′ ) − σ e
REMARK
The Von Mises Criterion is adequate for metals, where
hydrostatic stress states have an elastic behavior and
failure is typically due to deviatoric stress components.
69
Example
Consider a beam under a composed flexure state such that for a beam section
the stress state takes the form,
σ x τ xy 0
σ
[σ ] = τ xy 0 x 0
σx
0 0 0
70
Example - Solution
The mean stress is: 1 σx
=σm =Tr ( σ )
3 3
The deviator part of the stress tensor is:
σ x − σ m τ xy 0 23 σ x τ xy 0
σ′ =σ − σ esf ≡ τ xy −σ m 0 = τ xy − 13 σ x 0
0 0 −σ m 0 0 − 13 σ x
71
Example - Solution
The uniaxial effective stress is:
σ (=
σ) J 2′
3= σ x2 + 3τ xy2
F (σ ) ≡ 3 J 2′ − σ e =
0 σ x2 + 3τ xy2 σe
=
co
(comparison stress)
72
F ( σ ) ≡ φ( σ ) − σ e = 0
Tresca Criterion
Also known as the maximum shear stress criterion, it establishes
that the elastic domain ends when:
σ1 − σ 3 σ e
= τ max = F (σ ) ≡ (σ 1 − σ 3 ) − σ e =
0
2 2
Plane parallel to axis σ 2
73
Tresca Criterion F (σ ) ≡ (σ 1 − σ 3 ) − σ e =
0
F (σ) ≡ F ( J 2′ , J 3′ ) − σ e =
0
REMARK
The Tresca yield surface is appropriate for metals, which have an elastic behavior under
hydrostatic stress states and basically have the same traction/compression behavior.
74
Example
Obtain the expression of the Tresca criterion for an uniaxial state of stress
defined by:
E, G
y
σ u 0 0
σ ≡ 0
σ0x 0 σx
σu σu
0 0 0
x
76
σ u 0 0
σ ≡ 0 0 0
Example - Solution 0 0 0
Consider:
σ1 = σ u
σu ≥ 0 F (σ ) = (σ 1 − σ 3 ) − σ e = σ u − σ e = σ u − σ e
σ3 = 0
σu
σ1 = 0
σu < 0 F (σ ) =
(σ 1 − σ 3 ) − σ e =
−σ u − σ e =σu −σe
σ3 = σu
σu
77
Example
Consider a beam under a composed flexure state such that for a beam section
the stress state takes the form,
σ x τ xy 0
σ
[σ ] = τ xy 0 x 0
σx
0 0 0
78
Example - Solution
The principal stresses are:
1 1 1 1
σ 1 = σ x + σ x2 + τ xy2 , σ 3 = σ x − σ x2 + τ xy2
2 4 2 4
F (σ ) ≡ (σ 1 − σ 3 ) − σ e =
0
1 1 2 2 1 1 2 2
σ e = σ 1 − σ 3 = σ x + σ x + τ xy − σ x − σ x + τ xy
2 4 2 4
σ x2 + 4τ xy2 σe
=
co
(comparison stress)
79
Mohr-Coulomb Criterion
It is a generalization of the Tresca criterion, by including the
influence of the first stress invariant.
In the Mohr circle’s plane, the Mohr-Coulomb yield function takes
the form, cohesion internal friction angle
τ = c − σ tan φ
REMARK
The yield line cuts the
normal stress axis at a
positive value, limiting the
materials tensile strength.
80
Mohr-Coulomb Criterion
Consider the stress state for which the yield point is reached:
τ A = R cos φ
σ1 + σ 3
σA
= + R sin φ
2
σ1 − σ 3 σ + σ 3 σ1 − σ 3
τ A + σ A tg φ − c 0
= cos φ + 1 + = sin φ tg φ − c 0
2 2 2
(σ 1 − σ 3 ) + (σ 1 + σ 3 ) sin φ − 2c cos φ = 0 REMARK
For φ 0=
= and c σ e / 2 ,
F (σ ) ≡ (σ 1 − σ 3 ) + (σ 1 + σ 3 ) sin φ − 2c cos φ =
0 the Tresca criterion is
recovered.
81
Mohr-Coulomb Criterion
F (σ) ≡ F ( I1 , J 2′ , J 3′ ) =
0
REMARK
The Mohr-Coulomb yield surface is appropriate for frictional cohesive
materials, such as concrete, soils or rocks which have considerably
different tensile and compressive values for the uniaxial elastic limit.
82
Drucker-Prager Criterion
It is a generalization of the Von Mises criterion, by including the
influence of the first stress invariant.
The yield surface is given by the expression: REMARK
For φ 0=
= and c σ e / 2 ,
F (σ ) ≡ 3ασ m + [ J 2′ ] − β =
1/2
0
the Von Mises criterion
is recovered.
Where:
2sin φ 6c cos φ σ 1 + σ 2 + σ 3 I1
=α = ; β =; σm =
3 ( 3 − sin φ ) 3 ( 3 − sin φ ) 3 3
83
Drucker-Prager Criterion
F (σ) ≡ F ( I1 , J 2′ )
REMARK
The Drucker-Prager yield surface, like the Mohr-Coulomb one, is appropriate for
frictional cohesive materials, such as concrete, soils or rocks which have
considerably different tensile and compressive values for the uniaxial elastic limit.
84
Chapter 8
Plasticity
rs
ee
s gin
8.1 Introduction
t d le En
r
The elastoplastic models (constitutive equations) are used in continuum mechan-
ba
ge ro or
eS m
ics to represent the mechanical behavior of materials whose behavior, once cer-
ci
f
tain limits in the values of the stresses (or strains) are exceeded, is no longer rep-
ra
C d P cs
resentable by means of simpler models such as the elastic ones. In this chapter,
b
a
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
369
370 C HAPTER 8. P LASTICITY
diagonalization
rs
ee
s gin
Figure 8.1: Diagonalization of the stress tensor.
t d le En
Since σ is a symmetrical second-order tensor, it will diagonalize in an orthonor-
r
ba
ge ro or
eS m
mal base and all its eigenvalues will be real numbers. Then, consider a system
ci
of Cartesian axes {x , y , z } associated with a base in which σ diagonalizes. Its
f
ra
C d P cs
matrix of components in this base is
b
a
i
⎡ ⎤
an an n
σ1 0 0
y ha
⎢ ⎥
σ ]x y z = ⎣ 0 σ 2 0 ⎦ ,
[σ
le
(8.2)
liv or ec
0 0 σ3
M
.A
the directions associated with the axes {x , y , z } are named principal directions
uu
e
er
.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-
s gin
sulting from the computation of (8.5)) is
⎧
t d le En
⎪
⎪ I = Tr (σσ ) = σii = σ1 + σ2 + σ3
⎨ 1
1
r
I stress
ba
I2 = σ : σ − I12 = − (σ1 σ2 + σ1 σ3 + σ2 σ3 )
ge ro or
(8.7)
eS m
invariants ⎪
⎪ 2
ci
⎩
σ ) = σ1 σ2 σ3
I3 = det (σ f
ra
C d P cs
b
a
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
⎪
⎪ σ)
uu
⎪
⎪ J1 = I1 = σii = Tr (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)
= (σ σ : σ ) = Tr (σ σ ·σ)
C
invariants ⎪
⎪ 2 2
⎪
©
⎪
⎪
⎪
⎪ 1
3 1
⎪
⎪ = + + = σi j σ jk σki =
⎪
⎪ J3 I1 3I1 I2 3I3
⎪
⎪ 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
s gin
Given the stress tensor σ , the mean stress σm is defined as
t d le En
I1 1 1 1
σm = σ ) = σii = (σ1 + σ2 + σ3 )
= Tr (σ (8.9)
r
ba
3 3 3 3
ge ro or
eS m
ci
f
ra
and the mean pressure p̄ as
C d P cs
b
a
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 ,
liv or ec
M
.A
σ = σ sph + σ , (8.11)
m
er
tin
de f 1
σ sph : = Tr (σ σ ) 1 = σm 1
on
.O
⎡3 ⎤
C
σm 0 0 (8.12)
not ⎢ ⎥
©
σ sph ≡ ⎣ 0 σm 0 ⎦
0 0 σm
⎡ ⎤
σx − σm τxy τxz
not ⎢ ⎥
σ = σ − σ sph ≡ ⎣ τxy σy − σm τyz ⎦ . (8.13)
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 , named I and J invari-
ants, respectively, are derived from (8.7), (8.8), (8.9) and (8.13).
⎧
⎪
⎪
⎪
⎪ J1 = I1 = 0
⎪
⎪
⎪
⎨
J stress 1 1
invariants ⎪ J2 = I2 = (σ σ : σ ) = σij σ ji (8.14)
⎪
⎪ 2 2
⎪
⎪
⎪
⎪ 1
⎩ J3 = I3 = σ σ σ
3 i j jk ki
rs
ee
s gin
Remark 8.2. It is easily proven that the principal directions of σ co-
t d le En
incide with those of σ , that is, that both tensors diagonalize in the
r
same base. In effect, working in the base associated with the princi-
ba
ge ro or
pal directions of σ , i.e., the base in which σ diagonalizes, and, given
eS m
ci
that σ sph is a hydrostatic tensor and, thus, is diagonal in any base,
f
ra
then σ also diagonalizes in the same base (see Figure 8.2).
C d P cs
b
a
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
ee
s gin
Example 8.1 – Compute the value of the equivalent uniaxial stress (or effec-
tive stress) σ̄ for an uniaxial stress state defined by
t d le En
⎡ ⎤
r
σu 0 0
ba
ge ro or
not ⎢ ⎥
eS m
σ ≡⎣ 0 0 0 ⎦ .
ci
f
ra
C d P cs
0 0 0
b
a
i
an an n
y ha
Solution
le
liv or ec
.A
1
σ) = u .
Tr (σ σm =
3 3
m
⎡ ⎤
X Th
σu
er
⎡ ⎤
tin
0 0
σm 0 0 ⎢ 3 ⎥
not ⎢ ⎥ ⎢ σu ⎥
on
σ sph ≡ ⎣ 0 σm 0 ⎦ = ⎢ ⎥
.O
⎢ 0 3 0 ⎥
⎣ σu ⎦
C
0 0 σm
©
0 0
3
and the deviatoric component results in
⎡ ⎤
⎡ ⎤ 2
σu 0 0
σu − σm 0 0 ⎢3 ⎥
⎢ ⎥ ⎢⎢ 1 ⎥
0 ⎥
not
σ = σ − σ sph ≡ ⎣ 0 −σm 0 ⎦ = ⎢ 0 − σu ⎥.
⎣ 3
1 ⎦
0 0 −σm
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
ee
s gin
Consider a system of Cartesian axes in R3 {x ≡ σ1 , y ≡ σ2 , z ≡ σ3 } such that
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).
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Definition 8.1. The hydrostatic stress axis is the locus of points in
b
a
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.
s gin
t d le En
Definition 8.2. The octahedral plane Π is any of the planes that
r
ba
ge ro or
eS m
are perpendicular to the hydrostatic stress axis (see Figure 8.4). The
ci
equation of an octahedral plane is f
ra
C d P cs
σ1 + σ2 + σ3 = const.
b
a
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
ee
s gin
t d le En
Figure 8.5: Definitions of the normal and shear octahedral stresses.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
Remark 8.4. The normal octahedral stress σoct informs of the dis-
a
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-
√ the same value of σoct is the octahedral plane
le
cipal stress space with
liv or ec
.A
√
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 ⎢ √ ⎥ ⎪
(σ1 + σ2 + σ3 ) = 3σm ⎪
OA = OP · n ≡ σ1 , σ2 , σ3 ⎣ 1/ 3 ⎦ = 3
⎬
√ 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), J2 , as
ee
2 √ ⎫
AP = 2 J =⇒ AP = 2 (J )1/2 ⎬
s gin
2 2
√ =⇒ (8.18)
AP = 3 τ ⎭
t d le En
oct
r
ba
ge ro or
eS m
2 1/2
τoct = (J )
ci
(8.19)
f 3 2
ra
C d P cs
b
a
i
Alternative expressions of τoct in terms of the value of J2 in (8.14) are
an an n
y ha
1/2
1 1
le
2
τoct = √ σ1 + σ2 + σ3 − (σ1 + σ2 + σ3 )
2 2 2
liv or ec
and
3 3
(8.20)
M
.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 = σ −σ
σ sph = 0 ⇔ J2 = 0 ⇔ τoct = 0 .
C
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
tance (= 3 τoct ) from the hydrostatic stress axis to the point.
ee
Thus, it places point P on a certain circle in the octahedral
s gin
plane with center in the hydrostatic stress axis and radius
√ √
3 τoct = 2 (J2 )1/2 .
t d le En
− The third deviatoric stress invariant J3 characterizes the position
r
ba
of the point on this circle by means of an angle θ (J3 ).
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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 , J2 and J3 .
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
s gin
is, the curve obtained with the same function f (σ1 , σ2 , σ3 ) = 0 but
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
r
ba
ge ro or
of symmetry with respect to each of the axis in Figure 8.7.
eS m
ci
f
ra
C d P cs
b
a
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).
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
s gin
in the Coulomb model) whose value cannot be exceeded.
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.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
M
.A
m
d
uu
e
er
tin
on
|σ | < σe → Δ ε = 0
.O
|σ | = σe → Δ ε
= 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-
s gin
creasing (Δ σ < 0) and unloading be- Figure 8.11: Stress-strain curve
gins (σ < σe ). Further strain increase is for a loading-unloading-reloading
t d le En
automatically halted (Δ ε = 0). This sit- cycle in a frictional rheological
uation is maintained until stress is can- model.
r
ba
celed (σ = 0) at point 3. Note that, if
ge ro or
eS m
ci
the process was to be halted at this point, the initial state of null stress would be
f
ra
recovered but not the initial state of null strain. Instead, a permanent or residual
C d P cs
b
a
strain would be observed (ε
= 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
(Δ 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 (Δ ε < 0), which progressively reduces the accumulated
on
.O
strain. Finally, at point 5, the initial strain state is recovered, but not the original
stress state. Beyond this point, if unloading was imposed, there would be a cor-
C
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
Δ ε = Δ εe + Δ ε f ⎭ of strain
rs
ee
s gin
Frictional element
Elastic element
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
Δ ε = Δ εe
• |σ | < σe =⇒ Δ ε = 0 =⇒ Δ ε = Δ ε =⇒
f e
on
.O
Δ σ = EΔ ε
C
The frictional element does not deform for stresses |σ | < σe , therefore all
©
σ
• |σ | = σe =⇒ Δ ε f
= 0 =⇒ ε = + ε f =⇒
E
|σ | = σe
Δ ε = Δ ε f =⇒ Δ ε e = 0 =⇒ Δ 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,
at point 1, the strain is ε = ε e = σe /E. The value of σe at the end of this elastic
ee
section justifies its denomination as elastic limit.
s gin
Section 1 − 2:
t d le En
σe
ε= +ε f
r
|σ | = σe =⇒ Δ ε
= 0 =⇒
f
ba
E
ge ro or
eS m
Δε = Δε f > 0
ci
f
ra
C d P cs
b
This section corresponds to the frictional loading during which no deformation
a
i
is generated in the elastic element (no elastic strain is generated) and all incre-
an an n
y ha
|σ | < σe =⇒ Δ ε f = 0 =⇒ Δ ε = Δ 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
strain is ε = ε f
= 0. That is, the strain generated by the frictional element during
X Th
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 =⇒ Δ ε f = 0 =⇒ Δ ε = Δ 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
|σ | = σe =⇒ Δ ε =
f
0 =⇒ E
⎩Δε = Δε 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
a loading-unloading-reloading cycle in
s gin
an elastic-frictional rheological model.
t d le En
r
8.4.4 Frictional Model with Hardening
ba
ge ro or
eS m
ci
Consider the rheological model in Figure 8.14 composed of an elastic element
f
ra
(characterized by the parameter H , which will be denoted as hardening mod-
C d P cs
b
a
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
frictional element (σ (1) ) plus the stress in the elastic element (σ (2) ).
le
liv or ec
M
.A
σ = σ (1) + σ (2)
m
ε= εe = εf
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 387
rs
ee
b) Elastic element
s gin
σ (2) = H ε e = H ε
t d le En
(8.25)
Δ σ (2) = H Δ ε e = H Δ ε
r
ba
ge ro or
eS m
c) Combining (8.24) and (8.25) leads to
ci
f
ra
C d P cs
(1) (2)
= σ − H ε
b
a
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) Δε f = Δε = 0
• σ < σe ⇐⇒ σ − H ε < σe =⇒
Δ σ (2) = H Δ ε e = H Δ ε = 0
m
uu
Δ σ = Δ σ (1)
e
X Th
=⇒
er
tin
Δε = 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 ε = σe =⇒
⎩ σ (2) = σ − σ (1)
⎪
=⇒ Δ σ (2) = Δ σ = H Δ 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:
Δ σ (2) = EΔ ε = 0
(1)
σ < σe =⇒ Δ ε = 0 =⇒
Δ σ (1) = Δ 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 ε < σe .
rs
ee
Section 1 − 2:
s gin
σ = σe + σ (2)
(1)
σ = σe =⇒
t d le En
Δ σ = Δ σ (2) = H Δ ε
r
ba
ge ro or
eS m
This is a loading section in which all stress is absorbed by the elastic element.
ci
f
In global terms, the model increases its capacity to resist stress (the model is
ra
C d P cs
said to suffer hardening) proportionally to the increment of strain, being the
b
a
proportionality factor the hardening modulus H . This section is characterized
i
an an n
by the condition
y ha
σ − H ε = σe .
le
liv or ec
Section 2 − 3:
M
.A
Δ σ (1) = Δ σ
m
(1)
d
σ < σe =⇒ Δ ε = 0 =⇒
uu
Δ 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 ε < σe .
Section 3 − 4:
σ = −σe + σ (2)
(1)
σ = σe =⇒
Δ σ = Δ σ (2) = H Δ 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 ε = σe .
Beyond this point, relaxation of the stress in the frictional element leads to the
original state at point 0.
rs
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
Δε = Δ εe + Δ ε f of strain
(8.27)
σ = σe = σ f
Δσ = Δσe = Δσ 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.
s gin
t d le En
where σ e and σ f represent, respectively, the stresses sustained by the elastic
r
ba
element and the frictional model with hardening. Combining now the behav-
ge ro or
eS m
ci
ior of an elastic element (see Figure 8.8) with that of the frictional model with
f
ra
hardening in Figure 8.14, yields the following situations:
C d P cs
b
a
i
Δε f = 0
an an n
• σ − H ε f < σe =⇒
=⇒ Δ σ = EΔ ε
y ha
Δ σ = Δ εe
le
liv or ec
The frictional element with hardening does not deform and the increment of
M
.A
as elastic process.
d
uu
• σ − H ε f = σe
e
X Th
er
tin
⎧
⎨ σ > 0 and Δ σ > 0
on
Δ σ = Δ σ f = H Δ ε f
.O
a) σ Δ σ > 0 ⇐⇒ or =⇒
⎩ Δ σ = Δ σ e = EΔ ε e
C
1 1 E + H
=⇒ Δ ε = Δ ε e + Δ ε f = Δσ + Δσ = Δσ
E H EH
⎧
⎨ Δ σ = Ee f Δ ε
=⇒
⎩ Ee f = E H
E + H
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 Δ σ and the strain increment Δ 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-
s gin
ferentiated.
t d le En
Section 0 − 1 and section 2 − 3:
r
σ − H ε f < σe =⇒ Δ σ = EΔ ε
ba
ge ro or
eS m
ci
f
ra
Correspond to elastic processes.
C d P cs
b
a
Section 1 − 2 and section 3 − 4:
i
an an n
y ha
σ − H ε f = σe
le
=⇒ Δ σ = E e f Δ ε
liv or ec
σ Δσ > 0
M
.A
uu
Point 2:
e
σ − H ε f = σe
X Th
er
=⇒ Δ σ = EΔ ε
tin
σ Δσ < 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
ee
s gin
Figure 8.17: Stress-strain hardening curve for a loading-unloading-reloading cycle in an
elastic-frictional model with hardening.
t d le En
r
ba
8.5 Elastoplastic Phenomenological Behavior
ge ro or
eS m
ci
f
Consider a steel bar of length and cross-section A subjected to a tensile force F
ra
C d P cs
at its extremes. The stress in the bar will be σ = F/A (see Figure 8.18) and the
b
a
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
strain is no longer recovered during an ensuing unloading to null stress (point 3).
on
.O
There appears, thus, a remaining strain named plastic strain, ε p . However, dur-
C
incrementally elastic (Δ σ = E Δ 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
ee
s gin
t d le En
Figure 8.19: Response to loading-unloading-reloading cycles in an uniaxial tensile load-
ing test.
r
ba
ge ro or
eS m
ci
f
ra
process. From this point on, the behavior is no longer incrementally elastic (as
C d P cs
b
a
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
pears in the form of permanent strain in point 5, and, also, additional elastic
liv or ec
strain ε e is produced, understood as the part of the strain that can be recovered
M
.A
Consider a sample of virgin material (a material that has not suffered previous
er
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
ee
s gin
curve of the stretched material
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
curve without the Bauschinger effect
b
a
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-
s gin
tion to several dimensions requires the in- Figure 8.21: Uniaxial stress-strain
troduction of more abstract concepts.
t d le En
curve for an elastoplastic model.
r
ba
ge ro or
eS m
8.6.1 Additive Decomposition of Strain. Hardening Variable
ci
f
ra
The total strain ε is decomposed into the sum of an elastic (or recoverable)
C d P cs
b
a
strain ε e , governed by Hooke’s law, and a plastic (or irrecoverable) strain ε p ,
i
an an n
y ha
⎧ ⎧
⎪
⎨ ⎪
⎨
le
Additive decomposition ε = ε + ε =⇒ dε = dε + dε
e p e p
liv or ec
σ dσ (8.28)
⎪ ⎪
M
.A
of strain ⎩ εe = ⎩ dε e =
E E
m
d
uu
e
er
tin
.O
⎧
⎨ dα = sign (σ ) dε
⎪ p
C
Hardening dα ≥ 0 (8.29)
©
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
εp εp
s gin
dε p ≥ 0 =⇒ α = |dε p | = dε p = ε p .
t d le En
0 0
r
ba
ge ro or
plastic strain may decrease and its value no longer coincides with
eS m
ci
that of the hardening variable α.
f
ra
C d P cs
b
a
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
er
tin
on
.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
s gin
parameter
t d le En
r
ba
ge ro or
eS m
admissible
ci
f
ra
stress space
C d P cs
b
a
i
an an n
y ha
le
liv or ec
.A
m
! "
X Th
er
tin
.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
with the yield function σy (α) such that (see Figure 8.22)
s gin
Ēσ ≡ [−σy (α) , σy (α)] .
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
8.6.3 Constitutive Equation
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 + dσ , α + dα) = F (σ , α) + dF (σ , α) > 0
=0 >0
rs
ee
and, in agreement with (8.36), the stress state σ + dσ is not admis-
s gin
sible.
t d le En
r
ba
8.6.4 Hardening Law. Hardening Parameter
ge ro or
eS m
ci
The hardening law provides the evolution of the yield stress σy (α) in terms of
f
ra
C d P cs
the hardening variable α (see Figure 8.22). Even though the aforementioned
b
a
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
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 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 sign (σ ) dε p = 0 =⇒ dε p = dσ . (8.42)
H
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 =⇒ .
1 1 ⎩ E ep = E H
ee
+ E + H
E H
s gin
t d le En
8.6.6 Uniaxial Stress-Strain Curve
r
ba
The constitutive equation defined by expressions (8.37) to (8.39) allows obtain-
ge ro or
eS m
ci
ing the corresponding stress-strain curve for an uniaxial process of loading-
f
ra
unloading-reloading (see Figure 8.23) in which the following sections are ob-
C d P cs
b
a
served.
i
an an n
Section 0 − 1:
y ha
le
|σ | < σe =⇒ σ ∈ Eσ =⇒ Elastic regime
liv or ec
M
.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
.O
dF (σ , α) = 0 in plastic loading
C
Section 2 − 3 − 2:
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
Figure 8.23: Uniaxial stress-strain curve for a loading-unloading-reloading cycle con-
b
a
sidering the incremental theory of plasticity.
i
an an n
y ha
le
liv or ec
M
.A
%
d
F (σ , α) ≡ |σ | − σy (α) = 0 =⇒ σ ∈ ∂ Eσ
uu
Elastic unloading
e
in section 2 − 3
X Th
dF (σ , α) < 0
er
tin
%
F (σ , α) ≡ |σ | − σy (α) = 0 =⇒ σ ∈ ∂ Eσ
on
.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.
s gin
t d le En
r
Remark 8.14. Note the similarity between the stress-strain curve in
ba
ge ro or
eS m
Figure 8.23 and the one obtained with the elastic-frictional rheologi-
ci
f
cal model with hardening in section 8.4.5 (Figure 8.17). The friction
ra
C d P cs
strain in said model is equivalent to the plastic strain in the incre-
b
a
i
mental theory of plasticity.
an an n
y ha
le
liv or ec
M
.A
ing (8.43),
X Th
H
er
tin
E ep = E
E + H
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 in the definition of the slope E ep .
rs
ee
8.7 Plasticity in 3 Dimensions
s gin
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
r
ingredients, that is:
ba
ge ro or
eS m
ci
1) Additive decomposition of strain
f
ra
C d P cs
b
a
Additive ε = εe +ε p dεε = dεε e + dεε p
i
an an n
decomposition =⇒ (8.44)
y ha
.A
⎧
X Th
er
σ , α)
⎨ dε p = λ ∂ G (σ
tin
⎪
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
σ , α) < 0}
Elastic domain Eσ := { σ | F (σ
Initial (8.46)
σ , 0) < 0}
E0σ := { σ | F (σ
elastic domain
rs
Yield surface σ , α) = 0}
∂ Eσ := { σ | F (σ
ee
#
s gin
Admissible Ēσ = Eσ σ , α) ≤ 0}
∂ Eσ = { σ | F (σ
stress state
t d le En
where φ (σσ ) ≥ 0 is denoted as the equivalent uniaxial stress, σe is the elastic
r
ba
ge ro or
eS m
limit obtained in an uniaxial test of the material (it is a material property) and
ci
σy (α) is the yield stress. The hardening parameter H plays the same role as
f
ra
C d P cs
in the uniaxial case and determines the expansion or contraction of the elastic
b
a
domain Eσ , in the stress space, as α grows. Consequently,
i
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 =⇒ elastic
→ Perfect plasticity
domain
Eσ = E0σ
on
.O
C
Loading-unloading → λ ≥ 0 ; F (σ
σ , α) ≤ 0 ; λ F (σ
σ , α) = 0
conditions
(8.48)
Consistency σ , α) = 0 ⇒ λ F (σ
→ If 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
σ = C : dεε
=⇒ dσ (8.50)
ee
σ , α) < 0
dF (σ
s gin
t d le En
• Elastoplastic regime in plastic loading
%
r
ba
σ ∈ ∂ Eσ
ge ro or
eS m
=⇒ dσσ = C ep : dεε (8.51)
ci
σ , α) = 0
dF (σ f
ra
C d P cs
b
a
where C ep is known as the elastoplastic constitutive tensor which, after certain
i
an an n
le
liv or ec
∂G ∂F
M
.A
⊗ C: :C
σ , α) = C −
C ep (σ ∂σ ∂σ
m
∂F ∂G
d
H + :C :
uu
∂σ ∂σ
e
X Th
(8.52)
er
tin
∂G ∂F
Ci jpq Crskl
∂ σ pq ∂ σrs
on
.O
H + C pqrs
∂ σ pq ∂ σrs
©
σ ) ≡ φ (σ
E0σ := { σ | F (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 ,
σ ) ≡ φ (σ
∂ E0σ := { σ | F (σ σ ) − σe = 0} . (8.54)
Given that the initial elastic domain contains the origin of the stress space
σ = 0), every loading process in any point of the medium will include an elas-
(σ
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
s gin
different materials of interest in engineering.
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,
r
ba
ge ro or
eS m
F (σσ ) ≡ F I1 , J2 , J3 ,
ci
(8.55)
f
ra
C d P cs
b
and, since the criterion σ1 ≥ σ2 ≥ σ3 is adopted, its definition only affects the
a
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
σ ) − σe = 0}
∂ E0σ := { σ | φ (σ
d
uu
e
X Th
σ ) − σe < 0}
E0σ := { σ | φ (σ
er
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
Von Mises criterion: σ ) ≡ σ̄ (σ
F (σ σ ) − σe = 3 J2 − σe = 0 (8.56)
σ ) = 3 J2 is the effective stress (see Remark 8.3). An alternative
where σ̄ (σ
expression is obtained taking (8.19) and (8.20) and replacing them in (8.56),
rs
which produces
ee
1 1/2
s gin
σ ) ≡ √ (σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ1 − σ3 )2
F (σ − σe = 0 . (8.57)
2
t d le En
The graphical representation of the von Mises yield surface is shown in Fig-
r
ure 8.27.
ba
ge ro or
eS m
ci
f
ra
C d P cs
2
b
a
R= σe 2
i
3 R= σe
an an n
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.
ee
s gin
Solution
An uniaxial tensile loading case is characterized by the stress state
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
⎡ ⎤
b
a
σu 0 0
i
an an n
not ⎢ ⎥
σ ≡⎣ 0 0 0 ⎦ .
y ha
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
σ ) ≡ σ̄ (σ
F (σ σ ) − σe = |σu | − σe .
uu
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
s gin
not ⎢ ⎥ 1
σ ≡ ⎣ τxy 0 0⎦ =⇒ σm = σx =⇒
3
t d le En
0 0 0
r
ba
ge ro or
eS m
⎡ ⎤
ci
2
σx τxy f
ra
0
⎢ 3 ⎥
C d P cs
not ⎢
b ⎥
a
1 ⎢ 1
0 ⎥
i
σ = σ − σx 1 ≡ ⎢ τxy − σx ⎥.
an an n
3 ⎣ 1 ⎦
3
y ha
0 0 − σx
le
liv or ec
3
M
.A
1 1 4 2 1 2 1 2 1
J2 = σ : σ = σx + σx + σx + τxy + τxy = σx2 + τxy .
d
2 2 2
uu
2 2 9 9 9 3
e
X Th
er
tin
And the effective stress is obtained for the von Mises criterion,
on
.O
σ̄ = 3 J2 = σx2 + 3τxy 2 =⇒ F (σ σ ) < 0 =⇒ σ̄ < σe =⇒
C
σco = σx2 + 3τxy e .
2 <σ
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
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.
s gin
It follows from (8.58) that the Tresca criterion can be written as
t d le En
σ ) ≡ (σ1 − σ3 ) − σe = 0
F (σ (8.59)
r
Tresca criterion:
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
M
.A
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
ee
s gin
Figure 8.29: Tresca criterion in the principal stress space.
t d le En
r
ba
ge ro or
Figure 8.29 shows the yield surface corresponding to the Tresca criterion in
eS m
ci
the principal stress space, which results in an hexahedral prism whose axis is the
f
ra
C d P cs
hydrostatic stress axis.
b
a
i
an an n
y ha
Remark 8.19. Since the Tresca criterion does not depend on the first
le
stress invariant (and, therefore, on the stress σoct , see (8.16)), the cor-
liv or ec
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 , J2 , J3 ), is on said yield surface, all the points in the stress
uu
e
space with the same values of J2 and J3 will also be on this surface.
X Th
er
tin
.O
On the other hand, the dependency on the two invariants J2 and J3 ,
prevents (unlike in the case of the von Mises criterion) the surface
C
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
not ⎢ ⎥
ee
σ ≡⎣ 0 0 0 ⎦ .
s gin
0 0 0
t d le En
For the case σu ≥ 0 ,
%
r
ba
σ1 = σ u
ge ro or
eS m
=⇒ F (σ1 , σ2 , σ3 ) = (σ1 − σ3 ) − σe = σu − σe = |σu | − σe .
ci
σ3 = 0 f
ra
C d P cs
b
a
For the case σu < 0,
i
an an n
%
y ha
σ1 = 0
le
=⇒ F (σ1 , σ2 , σ3 ) = (σ1 − σ3 ) − σe = −σu − σe = |σu | − σe .
liv or ec
σ 3 = σu
M
.A
And the initial elastic domain is then characterized in the same way as in the
m
er
F (σ
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
Yield Surfaces. Failure Criteria 413
rs
ee
s gin
Figure 8.30: Representation of the Mohr-Coulomb criterion using Mohr’s circle in three
t d le En
dimensions.
r
ba
ge ro or
eS m
ci
The end of the elastic behavior (failure) in an increasing load process takes
f
ra
C d P cs
place when the first point in the Mohr’s circle (corresponding to a certain plane)
b
a
reaches the aforementioned yield line.
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-
pacity to withstand tensile loads.
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
⎧
⎪ σ − σ3
C
⎪
⎨ τA = R cos φ = 1 cos φ
σ1 − σ3
©
R= =⇒ 2
⎪
⎩ σA = σ1 + σ3 + R sin φ = σ1 + σ3 + σ1 − σ3 sin φ
2 ⎪
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
ee
s gin
t d le En
Mohr-Coulomb criterion:
(8.63)
σ ) ≡ (σ1 − σ3 ) + (σ1 + σ3 ) sin φ − 2 c cos φ = 0
F (σ
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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-
d
uu
er
length whose axis is the hydrostatic stress axis (see Figure 8.32). The
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
Yield Surfaces. Failure Criteria 415
√
3 c cot φ
rs
ee
s gin
Figure 8.32: Mohr-Coulomb criterion in the principal stress space.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Remark 8.23. In soil mechanics, the sign criterion of the normal
C d P cs
b
a
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.
s gin
t d le En
r
ba
ge ro or
eS m
Remark 8.24. Following certain algebraic operations, the Mohr-
ci
f
Coulomb criterion can be written in terms of the three stress invari-
ra
C d P cs
ants.
b
a
i
an an n
F σ
(σ ) ≡ F I , J , J
y ha
Mohr-Coulomb criterion: 1 2 3
le
liv or ec
M
.A
m
er
tin
.O
C
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
1/2
3
σ ) ≡ αI1 + J2
F (σ − β = 3 ασoct + τoct − β = F I1 , J2 = 0 . (8.66)
2
rs
ee
Remark 8.26. The independence on the third stress invariant J3 es-
s gin
tablishes that, if a certain point in the stress space belongs to the
yield surface, all the other points with the same value of the stress
t d le En
invariants I1 and J2 also belong to this surface, independently of the
r
value of the third stress invariant J3 . Given that the constant values of
ba
ge ro or
eS m
these invariants correspond to points of the octahedral plane placed
ci
f
at a same distance from the hydrostatic stress axis (see Figure 8.6),
ra
C d P cs
b
it can be concluded that the yield surface is a surface of revolution
a
i
around this axis.
an an n
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
ee
s gin
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.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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 2 ) = 0
rs
c) a I12 + b τoct
2 = c with a, b and c strictly positive
ee
s gin
t d le En
Solution
r
ba
ge ro or
a) In this case, there is a condition on the mean stress since
eS m
ci
f
I1 = σ1 + σ2 + σ3 = 3σm .
ra
C d P cs
b
a
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
3 σm
X Th
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 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
ee
s gin
c) The representation of a plane defined by a given point of the yield surface
and the hydrostatic stress axis is:
t d le En
r
ba
ge ro or
eS m
Point
ci
f
ra
C d P cs
b
a
i
an an n
y ha
.A
√ ⎫ ⎧ √
√ 3 ⎬ ⎨ I1 = 3 x
m
d=x= 3 σoct = I1
d
=⇒ R
uu
√ 3
⎭ ⎩ τoct = √
e
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
©
by2 ⎜ x ⎟ ⎜ y ⎟
a I1 + b τoct = c =⇒ 3ax +
2 2 2
= 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
⎜ x ⎟ ⎜ y ⎟ ⎜ z ⎟
⎜ c ⎟ +⎜ ⎟ ⎜ ⎟
⎝ ⎠ ⎝ 3c ⎠ + ⎝ 3c ⎠ = 1 .
3a b b
rs
ee
hydrostatic stress axis
s gin
t d le En
r
ba
ge ro or
eS m
Problem 8.2 – Graphically determine, indicating the most significant values,
ci
f
ra
the cohesion and internal friction angle of an elastoplastic material that follows
C d P cs
b
a
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
ee
s gin
Problem 8.3 – The following properties of a certain material have been exper-
imentally determined:
t d le En
1) In a hydrostatic compressive regime, the material never plasticizes.
r
ba
ge ro or
2) In a hydrostatic tensile regime, the virgin material plasticizes for a value of
eS m
ci
the mean stress σm = σ ∗ .
f
ra
C d P cs
b
a
3) In an uniaxial tensile regime, the virgin material plasticizes for a tensile
i
stress value σu .
an an n
y ha
σ = σ : σ = a σm + b .
.A
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 = 3σ∗
hydrostatic
stress axis
rs
ee
s gin
√
σ | = σ : σ = 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
r
tensile case and, thus, has no deviatoric stresses,
ba
ge ro or
eS m
ci
+ b = 0 =⇒ aσ ∗ + b = 0 .
σ | = 0 =⇒ aσm
|σ f [1]
ra
C d P cs
∗ σm =σ
b
a
i
an an n
The procedure is repeated for the uniaxial tensile case, whose deviatoric stresses
y ha
are now
le
liv or ec
⎡ ⎤ ⎡ ⎤
σu 0 0 2 0 0
M
.A
not ⎢ ⎥ 1 not σu ⎢ ⎥
σ ≡ ⎣ 0 0 0 ⎦ and σ sph = σu 1 =⇒ σ ≡ ⎣ 0 −1 0 ⎦ .
m
3 3
d
0 0 0 0 0 −1
uu
e
X Th
er
σ | = a σm + b produces
Then, applying the yield criterion |σ
tin
on
.O
2 2 1
σ |
|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
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
= 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:
s gin
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
r
ba
ge ro or
shown in Figure C is obtained. Determine:
eS m
ci
f
b) The values of the elastic load and displacement, Pe and δe .
ra
C d P cs
b
c) The values of the plastic load and displacement, Pp and δp .
a
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
2
σ̄ = (3 J2 ) 2 J2 = Tr σ
rs
2
1
ee
σ = σ − σ sph σ sph = σm 1 σm = σ)
Tr (σ
3
s gin
For this problem in particular,
t d le En
⎡ ⎤
0 τ 0
r
ba
not ⎢ ⎥
ge ro or
eS m
σ ≡⎣ τ 0 0 ⎦ ,
ci
f
ra
0 0 0
C d P cs
b
a
i
so σm = 0 and, therefore, σ sph = 0, leading to σ = σ . Then,
an an n
y ha
⎡ ⎤
le
τ2 0 0
liv or ec
2 not ⎢ ⎥ √
σ ≡ ⎣ 0 τ2 0 ⎦ =⇒ J2 = τ 2 =⇒ σ=
σ̄ 3τ .
M
.A
0 0 0
m
d
uu
Considering that material (1) plasticizes when τe = τ ∗ and material (2), when
e
X Th
τe = 2τ ∗ , then
er
tin
√
Material 1 =⇒ σe = 3 τ ∗ ,
on
.O
√
Material 2 =⇒ σe = 2 3 τ ∗ .
C
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
ee
s gin
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
r
ba
ge ro or
they act, considering the magnitude perpendicular to the plane of the paper as
eS m
ci
the unit value.
f
ra
C d P cs
b
h A h A h ∗ h ∗
a
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 = τ1 + τ2 = τ + 2τ ⎬ Pp = τ ∗ h and
2 2 2 2 =⇒ 2
2τ ∗ ⎪
⎪ τ ∗h
δ p = γ1B h = γ2B h = h ⎭ δp = 2 = 2δe .
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 B , which corresponds to a deformation
ee
of 3δe when the plastic load Pp is maintained constant (perfectly plastic regime).
s gin
Consider first the material (1). Unloading takes place starting at B and, accord-
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
r
since this is independent of the material being under loading or unloading con-
ba
ge ro or
eS m
ditions. Thus, to determine point C it is enough to draw a straight line that crosses
ci
point B and is parallel to OA, until the value −τ ∗ is reached.
f
ra
C d P cs
The same occurs in the case of material (2), with the difference that when the
b
a
line parallel to OA is drawn to cross point B , 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 ⎪
h D h D h h 3τ ⎪ 3τ ∗ h
s gin
PD = τ1 + τ2 = (−τ ∗ ) + (−2τ ∗ ) = − ⎪
⎬ PD = − and
2 2 2 2 2 ⇒ 2
∗
t d le En
τ ⎪
⎪ τ ∗h
δD = γ2D h = − h ⎪
⎭ δD = − = −δe .
r
G G
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
θ
d
not T
u (r, θ , z) ≡ [ur , uθ , uz ] = 0 , rz , 0 ,
uu
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 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
ee
s gin
t d le En
Figure B
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
⎢ φ r ⎥
not ⎢ ⎥
ε ≡⎢ 0 0 ⎥ .
⎣ φr
2h ⎦
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
⎢ ⎥
s gin
not ⎢ Gφ r ⎥
σ ≡⎢ 0 0 ⎥ .
⎣ h ⎦
t d le En
Gφ r
0 0
h
r
ba
ge ro or
eS m
ci
Plotting the σrr and τθ z components of the stress tensor in terms of the radius r
f
ra
C d P cs
yields:
b
a
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
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
higher stress, τe = 2τ ∗ . In addition, the external surface of the cylinder (r = R)
rs
suffers larger stresses, and this surface is composed of material (1). Therefore,
ee
plasticization will initiate when
s gin
τ ∗h
Gφe R
t d le En
τ = τ∗ =⇒ = τ∗ =⇒ φe =
r=R; φ =φe h GR
r
ba
ge ro or
eS m
is satisfied. This is the value of the rotation angle at the free end of the cylinder
ci
f
required for plasticization to initiate in the exterior material points of the cylinder
ra
C d P cs
(material (1)). The corresponding moment is obtained by replacing φe in [1],
b
a
i
an an n
πτ ∗ R3
y ha
πGR4
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
stresses cannot exceed the value τe , which corresponds to the onset of plasticity.
e
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
2π R 2π R/2
s gin
∗ Gφ1 r
M1 = rτ r dr dθ + r r dr dθ =
h
t d le En
0 R/2 0 0
R R/2
φ1
r
∗ 31 ∗ 3
ba
= 2πτ r dr + 2πG
2
r3 dr =⇒ M1 = πτ R .
ge ro or
eS m
h 48
ci
R/2 0
f
ra
C d P cs
b
a
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.
4τ ∗ h
Gφ2 R
τ = 2τ ∗ =⇒ = 2τ ∗ =⇒ φ2 = .
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
.A
2π R 2π R/2
3
m
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 , J2 and J3 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
Intersection with octahedral
ci
plane at (0, 0, 0)
f
ra
C d P cs
b
a
i
an an n
y ha
8.2 – An elastoplastic material is subjected to a pure shear test (I) and an uni-
le
axial tensile test (II). Plasticization occurs, respectively, at τ = a and σ = b.
liv or ec
Determine the values of the cohesion and internal friction angle assuming a
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
Problems and Exercises 435
rs
ee
s gin
t d le En
Infinitely rigid
material
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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)
Pi R3 − Pe R3e Pi − Pe R3i R3e
s gin
⎢ ⎥ C1
u=⎣ 0 ⎦ ur = Cr + 2 ; C =
i 3 ; C1 =
r E Re − R3i 2E R3e − R3i
t d le En
0
⎡ ⎤
r
ba
σrr 0
ge ro or
0
eS m
⎢ ⎥ 2C1 C1
ci
σ = ⎣ 0 σθ θ 0 ⎦ σrr = E C − 3 ; σθ θ = σφ φ = E C + 3
f
ra
r r
C d P cs
b
a
0 0 σφ φ
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
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 Δ 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 − Δ 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
s gin
⎢ ⎥ 2C1 C1
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
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
8.7 – A cylinder of radius R and height h is subjected to an exterior load P and
b
a
a uniform temperature increment Δ θ .
i
an an n
y ha
constants.
M
.A
er
.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
CH.9. CONSTITUTIVE
EQUATIONS IN FLUIDS
Multimedia Course on Continuum Mechanics
Overview
Introduction
Fluid Mechanics Lecture 1
What is a Fluid?
Pressure and Pascal´s Law Lecture 3
Constitutive Equations in Fluids
Lecture 2
Fluid Models
Newtonian Fluids
Constitutive Equations of Newtonian Fluids Lecture 4
Relationship between Thermodynamic and Mean Pressures
Components of the Constitutive Equation Lecture 5
Stress, Dissipative and Recoverable Power
Dissipative and Recoverable Powers Lecture 6
Thermodynamic Considerations
Limitations in the Viscosity Values
2
9.1 Introduction
Ch.9. Constitutive Equations in Fluids
3
What is a fluid?
Fluids can be classified into:
Ideal (inviscid) fluids:
Also named perfect fluid.
Only resists normal, compressive stresses (pressure).
No resistance is encountered as the fluid moves.
5
9.2 Pressure and Pascal’s Law
Ch.9. Constitutive Equations in Fluids
6
Pascal´s Law
Pascal’s Law:
In a confined fluid at rest, pressure acts equally in all
directions at a given point.
7
Consequences of Pascal´s Law
In fluid at rest:
there are no shear stresses
only normal forces due to pressure are present.
σ = − p0 1
− p0δ ij i, j ∈ {1, 2,3}
σ ij =
8
Pressure Concepts
Hydrostatic pressure, p0 : normal compressive stress exerted on a
fluid in equilibrium.
9
Pressure Concepts
Barotropic fluid: pressure depends only on density.
F ( ρ , p ) 0=
= p f (ρ)
F ( ρ , p,θ ) ≡ F ( ρ ) = ρ − k = 0 ρ = k = const.
10
9.3 Constitutive Equations
Ch.9. Constitutive Equations in Fluids
11
Reminder – Governing Eqns.
Governing equations of the thermo-mechanical problem:
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 Continuity Equation.
1 eqn.
− ρ ( u − θ s ) + σ : d ≥ 0 Clausius-Planck
Inequality. Second Law of
1 2 restrictions
− 2 q ⋅ ∇θ ≥ 0 Heat flux Thermodynamics.
ρθ Inequality.
19 scalar unknowns: ρ , v , σ , u , q , θ , s .
12
Reminder – Constitutive Eqns.
Constitutive equations of the thermo-mechanical problem:
Thermo-Mechanical
σ = σ ( v, θ , ζ ) Constitutive Equations.
6 eqns.
Entropy
s = s ( v, θ , ζ ) Constitutive Equation. 1 eqn.
(19+p) PDE +
(19+p) unknowns Thermal Constitutive Equation. Fourier’s
q=q (θ ) =− K ∇θ Law of Conduction. 3 eqns.
u = f ( ρ , v, θ , ζ ) Caloric
State Equations. (1+p) eqns.
Fi ( ρ ,θ=
,ζ ) 0 i ∈ {1, 2,..., p} Kinetic
13
Constitutive Equations
Constitutive equations
Together with the remaining governing equations, they are used to
solve the thermo/mechanical problem.
− p 1 + f ( d, ρ , θ )
σ=
− pδ ij + f ij ( d, ρ , θ ) i, j ∈ {1, 2,3}
σ ij =
15
9.4. Newtonian Fluids
Ch.9. Constitutive Equations in Fluids
16
Constitutive Equations of Newtonian
Fluids
Mechanic constitutive equations:
σ= − p1 + C : d
σ ij =− pδ ij + Cijkl d kl i, j ∈ {1, 2,3}
where C is the 4th-order constant (viscous) constitutive tensor.
C = λ1 ⊗ 1 + 2 µ I
Assuming:
Cijkl =λδ ijδ kl + µ (δ ik δ jl + δ ilδ jk )
an isotropic medium
the stress tensor is symmetrical i, j , k , l ∈ {1, 2,3}
Substitution of C into the constitutive equation gives:
σ= − p1 + λTr ( d ) 1 + 2 µ d REMARK
σ ij =− pδ ij + λ dllδ ij + 2 µ dij i, j ∈ {1, 2,3} λ and µ are not necessarily constant.
Both are a function of ρ and θ .
17
Relationship between Thermodynamic
and Mean Pressures
Taking the mechanic constitutive equation,
σ ij =
− pδ ij + λ dllδ ij + 2 µ dij i, j ∈ {1, 2,3}
Setting i=j, summing over the repeated index, and noting that
δ ii = 3 , we obtain
−3 p + ( 3λ + 2µ ) dll =
σii = 1
−3 p ( p = − σ ii )
3
−3 p Tr (d )
bulk viscosity κ
2
p = p + (λ + µ )Tr ( d ) = p + κ Tr ( d ) κ= λ + µ
2
3 3
18
Relationship between Thermodynamic
and Mean Pressures
Considering the continuity equation,
dρ 1 dρ
+ ρ∇ ⋅ v = 0 ∇⋅v = −
dt ρ dt
And the relationship
p= p + κ Tr ( d )
∂v i
Tr (d ) = d ii = =∇⋅v
∂xi
REMARK κ dρ
For a fluid at rest, v= 0 p= p= p0 p= p + κ ∇ ⋅ v= p −
dρ
ρ dt
For an incompressible fluid,
= 0= p p
dt
2
For a fluid with , κ
=0 λ= − µ p=p
Stokes' 3
condition
19
9.5 Components of the
Constitutive Equations
Ch.9. Constitutive Equations in Fluids
20
Components of the Constitutive
Equation
Given the Cauchy stress tensor, the following may be defined:
σ= − p1 + λTr ( d ) 1 + 2 µ d = σ σ sph + σ′
= − p1
SPHERICAL PART – mean pressure p= p − κ ∇ ⋅ v= p − κ Tr ( d )
DEVIATORIC PART
− p1 + λTr ( d ) 1 + 2 µ d =
− p1 + σ′ σ′ =( p − p ) 1 + λTr ( d ) 1 + 2 µ d
p= p − κ Tr ( d ) 2
2 κ= λ + µ
σ =−(λ + µ ) Tr ( d ) 1 + λTr ( d ) 1 + 2 µ d
′ 3
3
1 deviatoric part of the rate of
σ′ =2 µ (d − Tr ( d ) 1) = 2 µ d′ strain tensor
3
=d′
21
Components of the Constitutive
Equation
Given the Cauchy stress tensor, the following may be defined:
SPHERICAL PART – mean pressure p
p= p − κ ∇ ⋅ v= p − κ Tr ( d ) p
−κ
σ ′ = 2 µ d′
2µ
The stress tensor is then dij′
1
σ =Tr ( σ ) 1 + σ′ = − p 1 + σ′ REMARK
3
= −3 p from the definition Note that κ is not a
of mean pressure function of d, while
µ = µ (d ).
22
9.6 Stress, Dissipative and
Recoverable Powers
Ch.9. Constitutive Equations in Fluids
23
Reminder – Stress Power
Mechanical Energy Balance:
d 1 2
Pe ( t ) = ∫ ρ b ⋅ v dV + ∫ t ⋅ v dS = ∫ ρ v dV + ∫ σ : d dV
V ∂V dt Vt ≡V 2 V
24
Dissipative and Recoverable Powers
Stress Power = ∫ σ : d dV
1
V =d Tr (d)1 + d′
3
σ= − p1 + σ′
1
(
σ : d =− p1 + σ ′ ) ( )
: Tr d 1 + d ′ =
3
=3 = Tr =( d′ ) 0
1 1
=− p Tr ( d ) 1 : 1 + σ′ : d′ − p1 : d′ + Tr ( d ) σ′ : 1 =
3 3 = Tr =( σ′ ) 0
− p Tr ( d ) + σ′ : d′
=
σ ′ = 2 µ d′ − pTr ( d ) + κ Tr 2 ( d ) + 2 µ d′ : d′
σ :d =
p= p − κ Tr ( d )
RECOVERABLE STRESS DISSIPATIVE STRESS
POWER, WR. POWER, 2WD.
25
Dissipative and Recoverable Parts of
the Cauchy Stress Tensor
Associated to the concepts of recoverable and dissipative
powers, the Cauchy stress tensor is split into:
− p 1 + λTr ( d ) 1 + 2 µ d
σ=
RECOVERABLE DISSIPATIVE
PART, σ R . PART, σ D .
− pTr ( d ) =
WR = − p1 : d =
σR : d
2WD = κ Tr 2 ( d ) + 2 µ d′ : d′ = σ D : d REMARK
For an incompressible fluid,
WR = − p Tr ( d ) =
0
26
Thermodynamic considerations
Specific recoverable stress power is an exact differential,
1 1 dG
=
W σR =
:d → (exact differential)
ρ R
ρ dt
Then, the recoverable stress work per unit mass in a closed cycle is
zero: B ≡ A 1 B≡ A
1
B≡ A
∫
A
ρ
WR dt = ∫
A
ρ
σ R : d dt = ∫
A
dG = GB ≡ A − GA = 0
28
Thermodynamic Considerations
According to the 2nd Law of Thermodynamics, the dissipative
power is necessarily non-negative,
2WD ≥ 0 2WD = κ Tr 2 ( d ) + 2 µ d′ : d′ = 0 d= 0
In a closed cycle, the work done by the dissipative stress per unit
mass will, in general, be different to zero:
B≡B
1
∫
A
ρ
σ D : d dt > 0
2WD > 0
29
Limitations in the Viscosity Values
The thermodynamic restriction,
2WD = κ Tr 2 ( d ) + 2 µ d′ : d′ ≥ 0
introduces limitations in the values of the viscosity parameters
κ , λ and µ :
30
Chapter 9
Constitutive Equations in Fluids
rs
ee
s gin
9.1 Concept of Pressure
t d le En
r
Several concepts of pressure are used in continuum mechanics (hydrostatic pres-
ba
ge ro or
eS m
sure, mean pressure and thermodynamic pressure) which, in general, do not co-
ci
incide.
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
Figure 9.2: Mohr’s circle of the stress tensor of a fluid at rest.
ci
f
ra
C d P cs
b
a
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
ee
s gin
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
r
ba
Cartesian coordinate system used.
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
le
A new thermodynamic pressure variable, named thermodynamic pressure and
liv or ec
.A
m
.O
C
Example 9.1
The ideal gas law is a typical example of kinetic equation of state:
F (p, ρ, θ ) ≡ p − ρRθ = 0 =⇒ p = ρRθ ,
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
Fluid at rest : p0 = p̄ = p
Generally, in a fluid in motion the hydrostatic pressure, the mean
pressure and the thermodynamic pressure do not coincide.
rs
ee
s gin
Remark 9.5. A barotropic fluid is defined by a kinetic equation of
state in which the temperature does not intervene.
t d le En
Barotropic fluid : F (p, ρ) = 0 =⇒ p = f (ρ) =⇒ ρ = g (p)
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
Thermo-mechanical
σ = −p 1 + f (d, ρ, θ ) 6 equations (9.2)
constitutive equation:
Entropy
s = s (d, ρ, θ ) 1 equation (9.3)
rs
constitutive equation:
ee
s gin
c) Thermodynamic constitutive equations or equations of state
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-
r
ba
ge ro or
tion for the thermodynamic pressure.
eS m
ci
Caloric equation of f
ra
u = g (ρ, θ )
C d P cs
state:
b
a
Kinetic equation of 2 equations (9.4)
F (ρ, p, θ ) = 0
i
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
⎨ q = −k · ∇θ
e
X Th
Fourier’s
er
∂θ
tin
3 equations (9.5)
law: ⎩ qi = 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.
ee
s gin
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
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
9.4 Constitutive Equation in Newtonian Fluids
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
pendency of the stress tensor σ on the strain rate tensor d is obtained as a result
er
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 +CC:d σ =C:ε
s gin
σi j = −p δi j + Ci jkl dkl σi j = Ci jkl εkl
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Remark 9.9. The parameters λ and μ physically correspond to the
C d P cs
b
a
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
−3 p̄ 2
p = p̄ + λ + μ Tr (d) = p̄ + K Tr (d)
ee
3
s gin
K
(9.11)
t d le En
where K is denoted as bulk viscosity.
r
ba
ge ro or
eS m
ci
2
f
Bulk viscosity : K = λ + μ
ra
(9.12)
C d P cs
3
b
a
i
an an n
y ha
dρ 1 dρ
+ ρ∇ · v = 0 =⇒ ∇·v = − (9.13)
ρ dt
M
.A
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 =⇒ λ =− μ =⇒ p = p̄
ee
3
s gin
t d le En
r
ba
ge ro or
9.4.2 Constitutive Equation in Spherical and Deviatoric Components
eS m
ci
Spherical part f
ra
C d P cs
b
a
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
d
uu
in its spherical and deviator components, and replacing in the constitutive equa-
e
X Th
er
1
on
.O
σ= σ ) 1 + σ = − p̄ 1 + σ = −p 1 + λ Tr (d) 1 + 2μ d =⇒
Tr (σ
3
C
−3 p̄
©
σ = ( p̄ − p) 1 + λ Tr (d) 1 + 2μd = λ − K Tr (d) 1 + 2μd =⇒
−K Tr (d) 2
λ+ μ
3
2 1
σ = − μ Tr (d) 1 + 2μd = 2μ d − Tr (d) 1 =⇒
3 3
d
(9.17)
2 Stokes’ condition is assumed in certain cases because the results it provides match the
experimental observations.
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 = 2μd (9.18)
rs
1
σ = − p̄ 1 + σ d= Tr (d) 1 + d ,
ee
and (9.19)
3
s gin
and replacing in the expression of the stress power density (stress power per unit
of volume) σ : d, results in3
t d le En
r
1
ba
σ : d = (− p̄ 1 + σ ) :
ge ro or
Tr (d) 1 + d =
eS m
3
ci
f
ra
1 1
= − p̄ Tr (d) 1 : 1 +σ σ : d − p̄ 1 : d + Tr (d) σ : 1 =
C d P cs
b
a
3
3
(9.20)
i
an an n
3 Tr d = 0 Tr σ = 0
y ha
= − p̄ Tr (d) + σ : d .
le
liv or ec
M
.A
σ : d = − p − K Tr (d) Tr (d) + 2μ d : d .
d
(9.21)
uu
e
X Th
er
tin
on
.O
(9.22)
©
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
ee
⎩ 2W = K Tr2 (d) + 2μ d : d = σ : d ,
D D (9.25)
s gin
σ R + σ D ) : d = σ R : d + σ D : d = WR + 2WD .
σ : d = (σ
t d le En
r
ba
ge ro or
eS m
ci
f
Remark 9.11. In an incompressible fluid, the recoverable power is
ra
C d P cs
null. In effect, since the fluid is incompressible, dρ/dt = 0 , and
b
a
i
considering the mass continuity equation (5.24),
an an n
y ha
1 dρ
∇·v = − = 0 = Tr (d) =⇒ WR = −p Tr (d) = 0 .
le
liv or ec
ρ dt
M
.A
m
d
uu
e
X Th
er
tin
.O
dK dK
C
Pe = + σ : d dV = + σ R : d dV + σ D : d dV
©
dt dt
V V V
dK
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
WR dt = σ R : d dt = dG = GB≡A − GA = 0 ,
ee
(9.27)
ρ ρ
s gin
A A A
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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,
B
1
σ D : d dt > 0 . (9.29)
ρ
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
0 = Pe = + WR dV + 2WD dV =⇒
ee
dt
s gin
V =0 V
dK
t d le En
d 1 2
= 2 WD dV < 0
ρv dV = −
dt dt 2
r
ba
>0
ge ro or
V V
eS m
ci
f
and, therefore, the fluid looses (dissipates) kinetic energy and the velocity of
ra
C d P cs
its particles decreases.
b
a
i
an an n
y ha
le
9.4.5 Limitations in the Viscosity Values
liv or ec
M
.A
d
uu
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) = 0 ; d = 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
⎡ ⎤
i
⎡ ⎤ 1 ∂ vx
an an n
0 0
y ha
vx (y) ⎢ 2 ∂y ⎥
⎢ ⎥ ⎢ ⎥
v (x, y) = ⎢ ⎥ ; ⎢ ∂ ⎥
le
d=⎢ 1 v
0 ⎥=d
x
liv or ec
⎣ 0 ⎦ 0
⎣ 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
CH.10. FLUID MECHANICS
Multimedia Course on Continuum Mechanics
Overview
Governing Equations
Newtonian Fluids Lecture 1
Barotropic fluids
Boundary Conditions
BC in velocities
BC in pressures Lecture 12
Mixed BC
BC on free surfaces
3
10.1. Governing Equations
Ch.10. Fluid Mechanics
4
Reminder – Governing Eqns.
Balance equations of the thermo-mechanical problem:
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 Continuity Equation.
1 eqn.
− ρ ( u − θ s ) + σ : d ≥ 0 Clausius-Planck
Inequality. Second Law of
1 2 restrictions
− 2 q ⋅ ∇θ ≥ 0 Heat flux Thermodynamics.
ρθ Inequality.
5
Newtonian Fluids
Constitutive equations of the thermo-mechanical problem in a
Newtonian fluid:
Thermo-Mechanical
− p1 + λTr ( d ) 1 + 2µ d
σ= Constitutive Equations.
6 eqns.
Entropy
s = s ( d, ρ , θ ) Constitutive Equation. 1 eqn.
12 PDE
Thermal Constitutive Equation. Fourier’s
q = − K θ Law of Conduction. 3 eqns.
u = f ( ρ ,θ ) Caloric
State Equations. 2 eqns.
F ( ρ , p, θ ) = 0 Kinetic
6
Barotropic Fluids
A barotropic fluid is characterized by the
kinetic state equation:
F ( ρ , p, θ ) = 0 ρ = ρ ( p)
The uncoupled mechanical problem becomes:
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 Continuity Equation.
1 eqn.
Thermo-Mechanical
− p1 + λTr ( d ) 1 + 2µ d
σ= Constitutive Equations.
6 eqns.
8
Hydrostatic stress state vs. Hydrostatic
problem
Uniform velocity, v ( x ,t ) ≡ v ( t ) :⇒ v = ⊗ v = v ⊗ = 0
1
d = ∇S v = [v ⊗ ∇ + ∇ ⊗ v ] = 0
2
HYDROSTATIC
Thus, − p1 + λTr ( d ) 1 + 2 µ d
σ= σ=− p 1 ⇒ p =p STRESS SATE
=0 =0
Uniform and stationary velocity, v ( x ,t ) ≡ cnt :
dv ∂v
a= = + v ⋅∇ v= 0
dt ∂t
HYDROSTATIC
Thus, σ = − p0 1 Tr ( σ ) = −3 p0 p= p= p0 CASE
9
Hydrostatic Problem
A hydrostatic problem ( v ( x ,t ) ≡ cnt ) is characterized by:
Conservation of Mass.
ρ + ρ ∇ ⋅ v = 0 ρ ( X, t ) = ρ0 ( X ) 1 eqn.
Continuity Equation.
Thermo-Mechanical
− p1 + λTr ( d ) 1 + 2µ d
σ= σ = − p0 1 Constitutive Equations.
6 eqns.
Substituting the constitutive and the continuity eqn. into the Cauchy
eqn.:
−∇p0 + ρ0b =0 FUNDAMENTAL
σ =− p0 1 ⇒ ∇ ⋅ σ =∇ ⋅ (− p0 1) =−∇p0 → ∂p0 EQUATION OF
− ∂x + ρ0=
bi 0 i ∈ {1, 2,3} HYDROSTATICS
i
10
Gravity forces. Triangular pressure
distribution
For a fluid subjected to gravity forces,
−∇p0 + ρ0b =0 0
∂p0 b ( x, t ) = 0
− ∂x + ρ0=
bi 0 i ∈ {1, 2,3} − g
i
the momentum eq. can be written as:
∂p0 ( x, y , z ) ∫ dx
− =
0 p0 ( x, y, z ) ≡ p0 ( y, z )
∂x
∂p0 ( y , z ) ∫ dy p0 ( y, z ) ≡ p0 ( z )
− =
0
∂y
dp0 ( z ) ∫ dz − ρ0 gz + C
p0 =
− − ρ0 g =
0
dz
If the surface pressure is considered zero, then: =p0 ρ0 g ( h − z )
=
p0 z = h 0 − ρ0 g=
h+C 0 =C ρ0 g h Triangular
pressure
distribution
11
Archimedes´ Principle
Any fluid applies a buoyant force (up-thrust) to an object that is
partially or completely immersed in it.
The magnitude of the buoyant force is equal to
the weight of the fluid displaced by the object.
The resultant of the buoyant force on a floating
object acts at the center of mass of the displaced
fluid (center of buoyancy).
12
Archimedes´ Principle - Proof
Consider a solid with a volume V and density ρ within a fluid in a
hydrostatic case. Then,
The traction vector on the solid boundary :
t =σ ⋅ n =− p0 1 ⋅ n =− p0 n
p0 ( z ) ρ0 g ( h − z )
=
R= ∫ t dS= ∫ − p ( z )n dS
∂V ∂V
0
13
Archimedes´ Principle R= ∫ t dS= ∫ − p ( z )n dS
0
−∇p0 + ρ 0b =
0
∂V
Consider the same fluid without the solid in it, and replaced by fluid .
Then,
Pressures on the boundary of the “replacing” fluid are the same than in the
immersed solid case (and, therefore, the resulting force, R)
Volume of the displaced
The divergence theorem can be applied: fluid=V
(The pressure distribution is now continuous in space)
0
0 0
R = ∫ − p0 ndS = ∫ −∇p0 dV ≡ − ρ 0 0 ∫ dV =
ρ
∂V V −ρ b − g
V gV
0
0
V W
b
Finally, R= E eˆ z= W eˆ z ⇒ E= W
Up-thrust on the body ( E ) = weight of the
fluid displaced by the body (W )
14
Archimedes´ Principle −∇p0 + ρ 0b =
0
(second part proof)
Consider the moment of the up-thrust forces at the center of mass
(center of gravity, CG) of the volume of displaced fluid:
Divergence
Theorem
M GE = ∫ x × ( − p n ) dS= ∫ x × (−
0 p ) dV =
0
∂V V
=∇p0
− ∫ ( x × p0 ) dV
=
V
− ∫ ( x × ρ0b ) dV =
M GE = −MWG =
0
V
= MWG Moment of the weight of the displaced fluid with
respect to its center of gravity (by definition it must
be zero)
The up-thrust force, E, passes through the CG
of the volume of the displaced fluid (center of buoyancy).
15
Equilibrium of Floating Solids
The equilibrium can be:
16
10.3. Fluid Dynamics.
Barotropic Perfect Fluids
Ch.10. Fluid Mechanics
17
Barotropic Perfect Fluids
A perfect fluid is a Newtonian fluid with null viscosity, µ= λ= 0 :
− p1 + λTr ( d ) 1 + 2µ d
σ= σ = − p1 hydrostatic
stress state
Therefore, ⋅ σ =−p
σ:d = − pTr ( d )
− p1 : d =
18
Barotropic Perfect Fluids: Field Equations
The mechanical problem for a barotropic perfect fluid:
Conservation of Mass.
ρ + ρ ⋅ v =0
Continuity Equation.
1 eqn.
5 scalar unknowns: ρ , v , p .
19
Barotropic Perfect Fluids: Field
Equations
The thermal problem for a barotropic perfect fluid:
20
Bernoulli’s Trinomial
Consider a barotropic fluid with potential body forces:
0
T
∂φ ∂φ ∂φ 0
φ(
x, t ) =
gz → b ( x, t ) =
−φ ( x, t ) =
− =
∂x ∂y ∂z
Body forces − g
potential
1 dv
−p + ρ =
b ρ v → − p + =
b
ρ
dt
∂v + v⋅∇v
∂v
∂t
−P − φ= + 2ω × v + ( 12 v 2 )
1 ∂t
p = P ; b = − ∇φ
ρ
v ⋅ v = 2ω × v + 12 v 2 )
Bernoulli’s Trinomial
Rearranging, ∂v EQ. OF MOTION for a
− P + φ + 12 v 2 = + 2ω × v barotropic perfect fluid
∂t under potential body forces
22
Barotropic perfect fluid with potential forces:
Steady state solution
The equation of motion for a steady flow becomes:
∂v 1
− P + φ + 12 v 2 = + 2ω × v − ∇ P + φ + v 2=
2ω × v
∂t 2
23
Barotropic perfect fluid with potential forces:
Steady state solution
Then, the equation of motion along the considered streamline , Γ , reads:
dx ds
=
= M ( x( s ) ) ds dt
− P + φ + 12 v 2 = 2ω × v
− ( M ) ⋅ v= ( 2ω × v ) ⋅ v
=0
∇M ( x ( s ) ) ⋅ = = 0
dx dM
∀x ∈ Γ
ds ds M=
(x) cnt ∀x ∈ Γ
∂M
∂x
Γ: x =
x( s )
( )
Beurnoulli’s trinomial P + φ + 12 v 2 =x cnt ∀x ∈ Γ
remains constant along
the same streamline.
24
Barotropic perfect fluid with potential
(gravitational)forces: steady state solution
Incompressible fluid:
ρ ( p) =
p 1 1 p p
ρ= ρ0 =
cnt P( p) ≡ ∫ dp = ∫0
dp =
0 ρ
( )
p ρ 0 ρ0
ρ0
Potential gravitational forces:
0
b (x) = 0
−φ ( x ) = φ (x) =
gz
− g
p
P + φ + 12 v 2 (=
) ρ
x + gz + 2 v=
1 2
cnt ∀x ∈ Γ
0
p 1 v 2 def BERNOULLI’S
z+ + = H= cnt ∀x ∈ Γ THEOREM
ρ0 g 2 g
25
Barotropic perfect fluid with potential
(gravitational)forces: steady state solution
Bernoulli’s Theorem can be interpreted as:
pressure velocity
head head
elevation p 1 v 2 def
z+ + = H= cnt ∀x ∈ Γ
ρ0 g 2 g
velocity total or
head energetic
Piezometric or head
pressure hydraulic head, h
head
elevation
27
Example - Solution
p1 v12 p2 v22
z1 + + =z2 + +
S1 = cross section area at 1 ρ0 g 2 g ρ0 g 2 g
p=
1 p=
2 patm ≈ 0
S2
h v1 ≈ 0 ( S1 ≥ S2 → v=
1 v 2 ≈ 0)
S
1
≈0
h
v22
S 2 = cross section area at 2 z=
1 z2 + =v2 2g ( z1 − z2 )
2g
Velocity at the bottom hole of the tank: v 2 = 2gh
28
Barotropic perfect fluid with potential forces:
Transient solution
The equation of motion for an unsteady flow is:
∂v
− P + φ + 12 v 2 = + 2ω × v
∂t
∇ × v( x,=
t) 0 ∀x , ∀t
29
Barotropic perfect fluid with potential forces
and irrotational flow: Transient solution
In an irrotational flow:
1
× v( x, t )= 0 ∀x , ∀t ω( x, t )=
× v( x, t )= 0 ∀x , ∀t
2
There will exist a scalar function (named velocity potential χ ( x, t ) ) which
satisfies: v ( x,t ) = χ ( x, t )
t ∇= χ v ( x,t )
χ ∇=
( x, t ) χ ( x, t ) − ∫ ϕ (τ ) dτ
def
χ= ∂χ ∂χ
0 = − ϕ (t )
∂t ∂t
Finally,
∂χ Differential equation of
P + φ + ( ∇χ ) + = 0
1 2
∀x , ∀t
2 ∂t hydraulic transients
31
Barotropic perfect fluid with potential forces:
Transient solution in irrotational flows
The mechanical problem for a potential (irrotational) flow:
v ( x,t ) = ∇χ ( x, t )
Conservation of Mass.
ρ + ρ ⋅ (χ ( x, t ) ) =
0 ρ + ρ 2 χ =
0 1 eqn.
Continuity Equation.
∂χ
P (ρ, p) + φ + ( ) ∂t = 0
1 2 Linear Momentum Balance.
χ + ∀x , ∀ t 1 eqn.
2 Hydraulic Transients Equation.
32
Incompressible perfect fluid with potential
forces: transient solution in irrotational flows
In an incompressible flow:
dρ
=0 ρ = ρ0
dt
( x, t ) ∫ ρ =
P=
1 1
dp = ∫ dp
p
0 ( p) ρ 0 0 ρ0
33
Fluid Mechanics Equations
The mechanical problem for a potential (irrotational) and
incompressible flow:
Conservation of Mass.
ρ + ρ ⋅ (χ ( x, t ) ) =
not
0 ∇ 2 χ =∆χ =0 Continuity Equation.
1 eqn.
∂χ
p 1
( ) Linear Momentum Balance.
2
+φ + χ + = 0 ∀x , ∀ t 1 eqn.
ρ0 2 ∂t Hydraulic Transients Equation.
34
10.4. Newtonian Viscous Fluids
Ch.10. Fluid Mechanics
35
Governing Equations
Governing equations of the general fluid mechanics problem:
Conservation of Mass.
ρ + ρ ⋅ v =0 Continuity Equation. 1 eqn.
Energy Balance.
ρ u= σ : d + ρ r − ⋅ q First Law of Thermodynamics. 1 eqn.
Mechanical Constitutive
− p1 + λTr ( d ) 1 + 2 µ d
σ= Equations.
6 eqns.
Entropy
s = s ( d, θ , ρ ) Constitutive Equation.
1 eqn.
Lemma 2.
⋅ (α 1 ) =
α Where α ( x, t ) is a scalar function.
− p1 + λ Tr ( d ) 1 + 2 µ d
σ= ⋅ σ= ⋅ ( − p1 + λTr (d )1 + 2 µ d )=
= ⋅v =−p + λ (Tr (d )) + µ∆v + µ( ⋅ v )
(⋅v )
37
Navier-Stokes Equations
Then, the linear momentum balance equation is rearranged:
−p + ( λ + µ ) ( ⋅ v ) + µ∆v + ρ b =
dv dv
⋅σ + ρb =ρ ρ
dt dt
− p + ( λ + µ ) ( ⋅ v ) + µ∆ v + ρ b = ρ
dv
dt
NAVIER-STOKES
∂ ∂2v j ∂ 2
+ (λ + µ )
− p vi dv
+µ = + ρ bi ρ i i, j ∈ {1, 2,3} EQUATIONS
∂xi ∂xi ∂x j ∂x j ∂x j dt
The Navier-Stokes equations are essentially the equation of motion
(Cauchy’s equation) expressed solely in terms of velocity and pressure.
38
Energy Equation du
ρ = σ : d + ρ r −⋅q
dt
− pTr ( d ) + κ Tr 2 ( d ) + 2 µ d′ : d′
σ :d =
RECOVERABLE DISSIPATIVE
POWER, WR . POWER, 2WD .
du ∂v ∂ ∂θ ∂v
ρ =− p i + ρr + (K ) + κ ( i ) 2 + 2 µ dij′ dij′ i, j ∈ {1, 2,3}
dt ∂xi ∂xi ∂xi ∂xi
ENERGY EQUATION
The energy equation is essentially the energy balance equation expressed
solely in terms of velocity and pressure.
40
Reduced System of Equations
Governing equations of the general fluid mechanics problem are
reduced to:
Conservation of Mass.
ρ + ρ ⋅ v =0 Continuity Equation. 1 eqn.
Momentum Balance.
−p + ( λ + µ ) ( ⋅ v ) + µ∆v + ρ b =
ρ v Navier-Stokes Equations. 3 eqns.
Entropy
s = s ( v, θ , ρ ) Constitutive Equation.
1 eqn.
{
}
−p − − ( λ + µ ) ( ⋅ v ) + µ∆v + ρ b − ρ
dv
dt
=0
>0
Forces due to Viscous forces due
the pressure Body Inertial
to the contact with forces forces
gradient neighbour particles
42
Energy Equation: Physical
Interpretation
The energy equation can be physically interpreted as:
du
ρ =− p ⋅ v + ρ r + ⋅ ( K θ ) + κ Tr 2 (d ) + 2 µ d′ : d′
dt
43
Fluid Mechanics Equations ∂ρ
+ ⋅ ( ρ v) =
0
in Curvilinear Coordinates ∂t
CONTINUITY EQUATION
Cartesian Coordinates
∂ρ ∂ ∂ ∂
+ ( ρ vx ) + ( ρ v y ) + ( ρ vz ) =
0
∂t ∂x ∂y ∂z
Cylindrical Coordinates
∂ρ 1 ∂ 1 ∂ ∂
+ ( r)
ρ r v + ( θ)
ρ v + ( ρ vz ) =
0
∂t r ∂r r ∂θ ∂z
Spherical Coordinates
∂ρ 1 ∂ ∂ ∂
+ 2 ( ρ r 2 vr ) + ( ρ vϕ ) =
1 1
( ρ vθ sin θ ) + 0
∂t r ∂r r sin θ ∂θ r sin θ ∂ϕ
44
Fluid Mechanics Equations =aρa = − ∇p + ( λ + µ ) ( )
∇ ∇ ⋅ v + µ∆v + ρ b
∂v
+ v ⋅ ∇v
in Curvilinear Coordinates
; ∇⋅v =0
∂t
θ component ∂vθ
ρ + vr
∂vθ vθ ∂vθ
+ +
vϕ ∂vθ v r vθ vϕ2 cotg θ
+ −
1 ∂p
= − +
∂t ∂r r ∂θ r sin θ ∂ϕ r r r ∂θ
1 ∂ 2 ∂vθ 1 ∂ 1 ∂ 1 ∂ 2 vθ 2 ∂v r 2 cotg θ ∂vϕ
+µ 2 r + 2 ( vθ sin θ ) + 2 2 + 2 − 2 + ρ bθ
r ∂r ∂r r ∂θ sin θ ∂θ r sin θ ∂ ϕ 2
r ∂θ r sin θ ∂ϕ
φ component ρ ∂vϕ + v ∂vϕ + vθ ∂vϕ + vϕ ∂vϕ + vϕ v r + vθ vϕ cotg θ =
−
1 ∂p
+
∂t ∂r r ∂θ r sin θ ∂ϕ r sin θ ∂ϕ
r
r r
1 ∂ 2 ∂vϕ 1 ∂ 1 ∂ ∂ 2 vϕ ∂v r 2 cotg θ ∂vθ
+µ 2 r
r ∂r ∂
+
2
∂θ
θ ∂θ
( vϕ sin θ )
+
2 2
1
θ ∂ ϕ 2
+ 2
2
θ ∂ϕ
+ 2
r sin θ ∂ϕ
+ ρ bϕ
r r sin r sin r sin
47
Fluid Mechanics Equations σ=− p1 + λTr ( d ) 1 + 2 µ d
in Curvilinear Coordinates
2
λ= K - µ ; Tr (d= ) ⋅ v
3
∂v 2 ∂v x ∂v y
σ x µ 2 x − (·v ) − p + K ·v
= τ xy = τ yx = µ +
∂x 3 ∂y ∂x
∂v y 2 ∂v y ∂v z
σ y = µ 2 − (∇·v ) − p + K ∇·v τ yz = τ zy = µ +
∂y 3 ∂z ∂y
∂v 2 ∂v ∂v
σ z = µ 2 z − (∇·v ) − p + K ∇·v τ zx = τ xz = µ z + x
∂z 3 ∂x ∂z
48
Fluid Mechanics Equations σ=− p1 + λTr ( d ) 1 + 2 µ d
in Curvilinear Coordinates
2
λ= K - µ ; Tr (d= ) ⋅ v
3
∂v 2
σ r µ 2 r − (·v ) − p + K ·v
= ∂ v 1 ∂v r
τ rθ = τ θr = µ r θ +
∂r 3
∂r r r ∂θ
1 ∂vθ v r 2 ∂v 1 ∂v z
=σ θ µ 2 + − ( ·v ) − p + K ·v τ θz = τ zθ = µ θ +
r ∂θ r 3 ∂z r ∂θ
∂v 2 ∂v ∂v
σ z µ 2 z − (·v ) − p + K ·v
= τ zr = τ rz = µ z + r
∂z 3 ∂r ∂z
1 ∂ 1 ∂vθ ∂v z
=
·v ( rv r ) + +
r ∂r r ∂θ ∂z
49
NOTE: for incompressible fluids ( ⋅ v =0)
Fluid Mechanics Equations σ=− p1 + λTr ( d ) 1 + 2 µ d
in Curvilinear Coordinates
2
λ= K - µ ; Tr (d= ) ⋅ v
3
1 ∂vθ v r 2 sinθ ∂ v φ 1 ∂v θ
=σ θ µ 2 + − ( ·v ) − p + K ·v τ θφ = τ φθ = µ
sinθ + r sinθ ∂φ
r ∂θ r 3 r ∂θ
1 ∂vϕ v r vθ cotg θ 2
σϕ µ 2 + + −
3 ( ·v ) − p + K ·v
r sin θ ∂ϕ r r 1 ∂v r ∂ vφ
τ φr = τ rφ = µ + r
r sinθ ∂φ ∂r r
1 ∂ 2 ∂ 1 ∂ vϕ
·v =
r 2 ∂r
( r v r ) + r sin θ ∂θ ( vθ sin θ ) + r sin θ ∂ϕ
1
50
NOTE: for incompressible fluids ( ⋅ v =0)
10.5. Boundary Conditions
Ch.10. Fluid Mechanics
61
Boundary Conditions in Velocities
Prescribed velocities
Velocities are known in a certain part of the control volume
boundary, Γ v :=v ( x, t ) v ( x, t ) ∀x ∈ Γ
v
Impervious walls
Part of the boundary of control volume, Γ vn , which can be mobile, is
impervious (it cannot be penetrated by the fluid).
The normal component of the relative fluid/wall velocity, v r ≡ v − v* , is
considered null.
( v − v )=
*
⋅n 0 ∀x ∈ Γ v n
v n ( x,t ) = v ⋅ n = v* ⋅ n ∀x ∈ Γ v n
63
Boundary Conditions in Velocities
Adherent walls
In a viscous fluid in contact with a wall the fluid is considered to adhere to
the wall.
The relative fluid/wall velocity, v r , is considered null.
v r ( x, t ) = v − v* = 0 ∀x ∈ Γ v
=
v v* ∀x ∈ Γ v
64
Boundary Conditions in Pressures
Prescribed tractions
The traction vector’s value is prescribed in certain parts of the control
volume contour Γσ :
t ( x, t )= σ ⋅ n= t* ( x, t ) ∀x ∈ Γσ
=
p ( x , t ) p * ( x , t ) ∀x ∈ Γ p
65
Mixed Boundary Conditions
Prescribed traction vector and velocities
Pressure and the tangential component of the velocity, v t , are prescribed :
66
Boundary Conditions on Free Surfaces
67
Boundary Conditions on Free Surfaces
∂η ∂η ∂η
v z ( x,=
t) + vx + vy ∀x ∈ Γ fs
∂t ∂x ∂y
68
Boundary Conditions on Free Surfaces
p (=
x, t ) Patm ∀x ∈ Γ fs
This allows identifying the position of the free surface once the
pressure field is known:
=Γ fs : {x | p ( x,=
t ) − Patm 0 }
69
10.6. Laminar and Turbulent Flows
Ch.10. Fluid Mechanics
70
Laminar Flow
Flow persists as unidirectional movement.
Particles flow in parallel layers which do not mix.
A flow’s laminar character is identified by the Reynolds number: Re < 1000
The governing equations of the fluid mechanics problem are valid for this
type of flow.
def
V ×L
Re =
ν
71
Turbulent Flow
High values of the Reynolds number.
Highly distorted and unstable flow:
Stress and velocity at a given spatial point fluctuate randomly and very
fast, along time, about a mean value.
Specific models (turbulence models) are used to characterize this regime.
72
Chapter 10
Fluid Mechanics
rs
ee
s gin
10.1 Governing Equations
t d le En
r
A fluid is a particular case of continuous medium that is characterized by its
ba
ge ro or
eS m
specific set of constitutive equations. Consequently, the fluid mechanics problem
ci
is defined by the following equations:
f
ra
C d P cs
b
a
a) Balance Equations
i
an an n
le
liv or ec
dρ
+ ρ∇ · v = 0 (1 equation) (10.1)
M
dt
.A
m
dv
X Th
er
∇·σ +ρ b = ρ
tin
(3 equations) (10.2)
dt
on
.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
s gin
q = −K ∇θ (1 equation) (10.7)
t d le En
r
c) Thermodynamic equations of state
ba
ge ro or
eS m
ci
8) Caloric equation of state f
ra
C d P cs
b
a
u = u (ρ, θ )
i
(1 equation) (10.8)
an an n
y ha
le
liv or ec
.A
⎫
X Th
er
ρ → 1 unknown ⎪
tin
⎪
⎪
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
1 Note 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-
s gin
vene in the thermo-mechanical consti-
tutive equation (10.5), the governing
t d le En
equations of the (uncoupled) mechani-
Figure 10.1: Density depends on
r
cal problem in a Newtonian fluid are de-
ba
ge ro or
the thermodynamic pressure in a
eS m
fined as
ci
barotropic fluid.
f
ra
C d P cs
b
a
1) Mass continuity equation
i
an an n
dρ
y ha
+ ρ∇ · v = 0 (1 equation) (10.12)
le
dt
liv or ec
M
.A
2) Cauchy’s equation
m
dv
∇·σ +ρ b = ρ
uu
(3 equations) (10.13)
e
dt
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
456 C HAPTER 10. F LUID M ECHANICS
rs
spatial point being considered and is only a function of time. Therefore,
ee
1
s gin
d = ∇S v =(v ⊗ ∇ + ∇ ⊗ v) = 0 . (10.17)
2
t d le En
Then, the constitutive equation (10.14) is reduced to
r
σ = −p 1 + λ Tr (d) 1 + 2 μ d =⇒ σ = −p1 ,
ba
ge ro or
(10.18)
eS m
ci
=0 =0 f
ra
C d P cs
b
a
which indicates that the stress state is hydrostatic (see Figure 10.2). In addi-
i
an an n
σ ) = −3 p̄ = −3p
Tr (σ =⇒ p̄ = p .
le
(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 hydrostatic
dt ∂t (10.20)
⎪ case
σ = −p0 1 =⇒ p̄ = p = p0 ⎭
rs
ee
c) Fluid at rest: v (x,t) ≡ const. = 0
s gin
A particular case of hydrostatics is that of a fluid at rest with null velocity.
t d le En
r
10.2.1 Hydrostatic Equations
ba
ge ro or
eS m
ci
f
The hydrostatic problem is governed by the following equations:
ra
C d P cs
b
a
1) Constitutive equation
i
an an n
y ha
σ = −p0 1
(10.21)
le
σi j = −p0 δi j i, j ∈ {1, 2, 3}
liv or ec
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)
s gin
dt
Introducing (10.21) and (10.22) in (10.23),
t d le En
⎧
⎪
⎪
⎪
r
⎨ ∇ · σ = ∇ · (−p0 1) = −∇p0
ba
ge ro or
eS m
ci
⎪
⎪ ∂ σi j ∂ ∂ p0 f
ra
⎪
⎩ ∇·σ j = = (−p0 δi j ) = − = − ∇p0 j j ∈ {1, 2, 3}
C d P cs
∂ xi ∂ xi ∂xj
b
a
i
an an n
(10.24)
y ha
le
liv or ec
⎧
⎪
⎨ −∇p0 + ρ0 b = 0
M
.A
i ∈ {1, 2, 3}
d
∂ xi
uu
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.
ee
s gin
⎧
⎪
⎪ ∂ p0 (x, y, z)
⎪
⎪ − =0 =⇒ p0 (x, y, z) ≡ p0 (y, z)
t d le En
⎪
⎪ ∂x
⎪
⎨
r
∂ p0 (y, z)
ba
ge ro or
− =0 =⇒ p0 (y, z) ≡ p0 (z) (10.27)
eS m
⎪
⎪ ∂y
ci
⎪
⎪
⎪ f
ra
⎪
⎩ − ∂ p0 (z) − ρ0 g = 0
⎪
C d P cs
=⇒ p0 = −ρ0 g z +C
b
a
∂z
i
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.
ee
s gin
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,
r
ba
ge ro or
even though its velocity and acceleration are assumed to be small enough to
eS m
ci
ensure a hydrostatic state in the fluid. On the other hand, Figure 10.6 b) shows
f
ra
the same fluid without the solid, such that the volume occupied by said solid in
C d P cs
b
a
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)
ee
s gin
and the resultant R of the forces the fluid exerts on the solid is
t d le En
R= t dS = −p0 n dS . (10.31)
r
∂V ∂V
ba
ge ro or
eS m
ci
Note now that, since the hydrostatic pressure distribution is the same in both
f
ra
cases of Figure 10.6, this resultant is the same as the one obtained in case b)
C d P cs
b
a
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
M
.A
∂V V
d
uu
e
er
tin
R= −∇p0 dV = −p0 b dV = − ρ0 b dV = W êz = E êz , (10.33)
on
.O
V V V
C
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,
upward buoyant force = weight of the displaced fluid , (10.34)
E W
whereby the first part of Archimedes’ Principle is proven.
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
ee
s gin
Figure 10.7: Forces acting on the volume of displaced fluid.
t d le En
r
ba
ge ro or
eS m
c) Vertical line of application of the upward buoyant force
ci
f
ra
Consider now the moment MG
C d P cs
E of the upward buoyant force E with respect to
b
a
the center of gravity, G, of the volume of displaced fluid (see Figure 10.72 ),
i
an an n
⎧
y ha
⎪
⎪
Divergence
⎪ Theorem
le
⎪
liv or ec
⎪
⎪ G = × (−p = × (−p ∇) = − x × ∇p0 dV
⎪ M
⎪ E x 0 n) dS x 0 dV
⎪
M
.A
⎪
⎪
⎪
⎪ ∂V V V
⎪
⎪
⎨ G ∂
m
ME i = − ei jk x j p0 nk dS = − ei jk x j p0 dV =
d
uu
∂ xk
e
⎪
⎪ ∂V
⎪ V
X Th
⎪
er
⎪
tin
⎪
⎪ ∂xj ∂ p0 ∂ p0
⎪
⎪ = − e p dV − e x dV = − ei jk x j dV
⎪
⎪
i jk
∂ xk
0 i jk j
∂ xk ∂ xk
on
.O
⎪
⎪ V
⎪
⎪
V V
⎩ ei jk δ jk = i ∈ {1, 2, 3}
C
ei j j = 0
©
(10.35)
and replacing the fundamental equation of hydrostatics (10.25) in (10.35) finally
yields
MG
E =− (x × ∇p0 ) dV = − (x × ρ0 b) dV = −MW
G
=0, (10.36)
V V
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.
s gin
Solution
t d le En
Consider a floating medium, in equilibrium, and the following two situations:
r
ba
a) The center of gravity of the solid (center of thrust) is below the center of
ge ro or
eS m
gravity of the displaced fluid (center of buoyancy).
ci
f
ra
C d P cs
In this case, any perturbation (inclination) tends to create a moment M = W d
b
a
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
s gin
improved flotation stability of this boat.
t d le En
r
ba
ge ro or
eS m
10.3 Fluid Dynamics: Barotropic Perfect Fluids
ci
f
ra
In the most common case, the velocity is not uniform nor stationary (v ≡ v (x,t)),
C d P cs
b
a
and, therefore, in general, the acceleration will not be null (a (x,t) = 0). In con-
i
an an n
sequence,
the divergence of the velocity (∇ · v = 0) and the gradient of the ve-
y ha
not
locity ∇ ⊗ v = ∇v = 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.
d
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:
s gin
a) Mechanical problem
t d le En
1) Mass continuity equation
r
ba
ge ro or
eS m
dρ
ci
+ ρ∇ · v = 0 (1 equation) (10.38)
f
ra
dt
C d P cs
b
a
i
an an n
le
dv
liv or ec
.A
m
The mechanical
problem is composed of 5 equations and 5 unknowns ρ (x,t) , v (x,t),
on
.O
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
(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
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-
s gin
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).
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
10.3.2 Resolution of the Mechanical Problem under Potential Body
an an n
le
Consider now the mechanical problem for the particular case of potential body
liv or ec
.A
m
In the particular case of a gravitational potential with the line of action along the
er
tin
.O
⎡ ⎤
C
0
not ⎢ ⎥
©
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
p
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)
⎪
⎪ ∂ P̂ 1
rs
⎩ ∇P i = ∇p i = ∇p i i ∈ {1, 2, 3}
∂p ρ (p)
ee
s gin
leading to
1
∇P = ∇p .
t d le En
(10.48)
ρ (p)
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
Lemma 10.2. The convective term of the acceleration can be written
i
an an n
as
y ha
1 2
ω ×v+∇
v · ∇v = 2ω v ,
le
2
liv or ec
.A
where 2ω
m
d
uu
e
Proof
X Th
er
∂vj ∂ v j ∂ vi ∂ vi ∂ vi
on
.O
v · ∇v j = vi = vi − +vi = 2 vi w ji +vi =
∂ xi ∂ xi ∂ x j ∂xj
∂xj
C
−wi j
©
2 w ji
∂ vi ∂ vi
= −2 vi wi j + vi = 2 ei jk vi ωk +vi =
∂xj
∂xj (10.49)
e jkivi ω k
∂ 1 1 2
= 2 e jki vi ωk + vi vi = 2ω ω ×v j + ∇ v ,
∂ x j 2
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
ω ×v+∇
v · ∇v = 2ω v . (10.50)
2
rs
ee
dv ∂ v ∂v 1 2
− ∇P − ∇φ = = + v · ∇v = ω ×v+∇
+ 2ω v , (10.52)
s gin
dt ∂t ∂t 2
t d le En
where the result (10.50) has been taken into account. Finally, (10.52) is rewritten
as
r
∂v
ba
1 2
ge ro or
− ∇P + ∇φ + ∇ = + 2ωω ×v .
eS m
v (10.53)
∂t
ci
2
f
ra
C d P cs
b
a
i
an an n
1 2 ∂v (10.54)
−∇ P + φ + v = + 2ωω ×v
M
.A
2 ∂t
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
s gin
Figure 10.9). Projecting (multiplying) equation (10.53) in the direction tangent
to the streamline, t, results in
t d le En
1 2
r
− ∇ P + φ + v = 2ω ω × v =⇒ − (∇M) · v = (2ω ω × v) · v = 0 (10.56)
ba
ge ro or
eS m
2
ci
=0
dx ds f
ra
M (x)
C d P cs
b
a
ds dt
i
⎫
an an n
⎪
y ha
dx dM ⎪
⎪
∇M (x (s)) · = = 0⎪
⎬
le
liv or ec
ds ds ∀x ∈ Γ =⇒ M (x) = const. ∀x ∈ Γ
⎪
M
⎪
.A
∂ M (x (s)) dxi dM ⎪
= =0⎪ ⎭
∂ xi ds ds
m
(10.57)
uu
e
er
tin
1
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.
ee
s gin
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-
r
ba
ge ro or
acteristics:
eS m
ci
f
ra
a) The fluid is incompressible
C d P cs
b
a
i
ρ = ρ (p) = ρ0 = const. (10.59)
an an n
y ha
p p
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
elevation pressure velocity total
head head height
rs
Remark 10.4. The expression in (10.63) constitutes the so-called
ee
Bernoulli’s theorem (for an incompressible perfect fluid under
s gin
gravitational forces and in steady-state regime), which establishes
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-
r
ure 10.10).
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
Solution
C d P cs
b
a
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 + +
ρ0 g 2 g
ρ0 g 2 g
on
.O
= h
= 0
=0 =0 =0
C
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 ∂v
−∇ P + φ + v2 = ω ×v .
+ 2ω (10.64)
2 ∂t
rs
• with potential body forces
ee
• and irrotational flow.
s gin
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
r
ba
ge ro or
point of this fluid.
eS m
ci
f
ra
C d P cs
b
a
i
In other words, an irrotational flow has a null vorticity vector.
an an n
y ha
⎧
le
⎨ ∇ × v (x,t) = 0
liv or ec
⎩ 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
Note that, in this case, the vector field v (x,t) is determined in terms of the scalar
C
velocity potential χ (x,t) (which becomes the main unknown of the problem).
©
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
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 + = ϕ (t) .
rs
(10.69)
2 ∂t
ee
Defining a modified velocity potential χ̄ (x,t) of the form
s gin
⎧
t ⎪
⎨ ∇χ̄ = ∇χ = v (x,t)
t d le En
de f
χ̄ (x,t) = χ (x,t) − ϕ (τ) dτ ⇒ ∂ χ̄ ∂χ (10.70)
r
⎪
⎩ = − ϕ (t)
ba
ge ro or
eS m
0 ∂t ∂t
ci
f
ra
C d P cs
and replacing (10.70) in (10.69) produces
b
a
i
an an n
1 ∂χ 1 ∂ χ̄
y ha
P + φ + v2 + − ϕ (t) = 0 =⇒ P + φ + (∇χ̄)2 + =0 ∀x ∀t ,
2 ∂t
2 ∂t
le
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
©
∇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,
• irrotational flow
ee
• and incompressible flow.
s gin
Since the flow is incompressible, (10.46) and (10.72) allow determining7
t d le En
⎧
⎪ p
⎪
r
⎪ 1 p
⎨ P (p) =
ba
d p̄ =
ge ro or
eS m
dρ ρ ( p̄) ρ0
= 0 =⇒ ρ = ρ0 =⇒
ci
(10.76)
dt ⎪
⎪
0
f
ra
⎪
C d P cs
⎩ ∇2 χ̄ = Δ χ̄ = 0
not
b
a
i
an an n
y ha
.A
∂ 2 χ̄
Δ χ̄ = =0 (10.77)
m
∂ xi ∂ xi
d
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
Δ (•) = ∇ · ∇ (•) = ∇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
ρ = σ : d+ρ r−∇·q Energy balance (1 eqn.)
rs
dt
ee
σ = −p1 + λ Tr (d) 1 + 2μ d Mechanical constitutive equation (6 eqns.)
s gin
t d le En
s = s (d, θ , ρ) Entropy constitutive equation (1 eqn.)
r
ba
ge ro or
eS m
q = −K∇θ
ci
Heat conduction equation (3 eqns.)
f
ra
C d P cs
b
a
i
u = u (ρ, θ ) Caloric equation of state (1 eqn.)
an an n
y ha
le
F (ρ, p, θ ) = 0
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 = Δ 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 ∂ vi ∂ v j 1 ∂ 2 vi 1 ∂ 2v j
∇·d = di j = + = + =
j ∂ xi ∂ xi 2 ∂ x j ∂ xi 2 ∂ xi ∂ x j 2 ∂ xi ∂ xi
1 ∂ ∂ vi 1 ∂ 2 v j 1 ∂ 1
= + = (∇ · v) + Δ v j =
2 ∂ x j ∂ xi 2 ∂ xi ∂ xi 2 ∂ x j 2
Δvj [Δ v] j
∇·v [∇ (∇ · v)] j
1 1
= Δ v + ∇ (∇ · v) j ∈ {1, 2, 3}
rs
2 2 j
ee
(10.80)
s gin
1 1
∇ · d = Δ v + ∇ (∇ · v) (10.81)
2 2
t d le En
r
ba
ge ro or
eS m
ci
f
Lemma 10.4. Given a scalar function α (x,t), the following is sat-
ra
C d P cs
b
isfied.
a
i
∇ · (α 1) = ∇α
an an n
y ha
le
liv or ec
Proof
M
.A
∂ (αδi j ) ∂α ∂α
m
∇ · (α 1) = = δi j = = ∇α i i ∈ {1, 2, 3} (10.82)
d
i ∂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)) +μΔ v + μ∇ (∇ · v) (10.84)
⎪
⎨
∇ (∇ · v)
⎪
⎪ dv
⎪
⎩ ∇ · σ + ρb = −∇p + (λ + μ) ∇ (∇ · v) + μΔ 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) + μΔ 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
ee
s gin
10.4.2 Energy Equation
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
r
of (10.79). To this aim, the definition of stress power in a Newtonian fluid (see
ba
ge ro or
eS m
Chapter 9) is recovered,
ci
f
ra
C d P cs
σ : d = WR + 2WD = −p ∇ · v + K Tr2 (d) + 2μ d : d ,
b
(10.86)
a
i
an an n
where d 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
dt
X Th
er
tin
Energy equation
on
.O
du
= −p∇ · v + ρ r + ∇ · (K ∇θ ) + K Tr2 (d) + 2μ d : d
C
ρ
dt
©
(10.89)
2WD
du ∂ vi ∂ ∂θ ∂ vi 2
ρ = −p +ρ r+ K +K + 2μ di j di 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)
+μΔ v + ρb = ρ
s gin
dt
du
ρ = −p∇ · v + ρ r + ∇ · (K ∇θ ) +
t d le En
dt Energy balance (1 eqn.)
+K Tr2 (d) + 2μ d : d
r
ba
ge ro or
eS m
ci
u = u (ρ, θ ) f
ra
Caloric equation of state (1 eqn.)
C d P cs
b
a
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
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)
s gin
that acts on a volume differential of the fluid in motion as follows.
t d le En
−∇p − − (λ + μ) ∇ (∇ · v) + μΔ v + ρb + ρa = 0
r
ba
ge ro or
(10.93)
eS m
forces due to viscous forces exerted by body inertial
ci
the pressure forces forces
f
the contact between particles
ra
gradient
C d P cs
(= 0 when λ = μ = 0)
b
a
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 d : d
dt
2WD
rs
dt unit of volume and unit of time
ee
d (dV )
variation of volume
s gin
∇ · v = dt =
dV unit of volume and unit of time
t d le En
d (dV ) mechanical work of the thermodynamic pressure
p =
r
−p∇ · v = − dt unit of volume and unit of time
ba
ge ro or
dV
eS m
(see Figure 10.12 and footnote8 )
ci
f
ra
C d P cs
heat generated by the internal sources and conduction
b
a
ρ r + ∇ · (k ∇θ ) =
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
s gin
and introducing (10.94) in (10.90) results in the governing equations detailed in
Table 10.2.
t d le En
r
ba
ge ro or
eS m
∇·v = 0
ci
Mass continuity equation
Mechanical
f
ra
C d P cs
Problem
b
a
dv
−∇p + μ Δ v + ρ0 b = ρ0
i
Navier-Stokes equation
an an n
dt
y ha
du
le
Energy balance ρ0 = ρ0 r + ∇ · (K ∇θ ) + 2μ d : d
liv or ec
Thermal dt
Problem
M
.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) + μΔ v + ρb = ρ
3 dt
du
Energy balance ρ = −p∇ · v + ρ r + ∇ · (K ∇θ ) + 2μ d : d
dt
rs
ee
Kinetic equation of state F (ρ, p, θ ) = 0
s gin
2
Constitutive equation σ = −p 1 − μ Tr (d) 1 + 2μ d
t d le En
3
r
ba
ge ro or
eS m
ci
Table 10.3: Governing equations in Stokes fluids.
f
ra
C d P cs
b
a
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
ρ0 = ρ0 r + ∇ · (K ∇θ )
s gin
Energy balance
Thermal dt
Problem
t d le En
Caloric equation of state u = u (ρ0 , θ )
r
ba
ge ro or
σ = −p0 1
eS m
Constitutive equation
ci
f
ra
C d P cs
b
a
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,
d
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.
ee
s gin
cannot be penetrated by said fluid. The mathematical expression of this con-
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
r
ba
ge ro or
eS m
be null (see Figure10.14),
ci
f
ra
vn (x,t) = v · n = v∗ · n ∀x ∈ Γvn =⇒
C d P cs
b
a
i
an an n
fluid wall
(10.99)
y ha
vr · n = (v − v∗ ) · n = 0 .
le
∀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,
s gin
thus, the relative velocity between the fluid and the wall vr is null.
t d le En
vr (x,t) = v − v∗ = 0 ∀x ∈ Γv =⇒ v = v∗ ∀x ∈ Γv (10.100)
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
on a part of the boundary Γp is written as
ee
s gin
p (x,t) = p∗ (x,t) ∀x ∈ Γp . (10.103)
t d le En
r
ba
ge ro or
eS m
10.5.3 Mixed Boundary Conditions
ci
f
ra
C d P cs
In certain cases (such as the entrance and exit sections of pipes) the pressure (a
b
a
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
s gin
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
r
to a free surface is that such a surface is a material surface (constituted always
ba
ge ro or
eS m
by the same particles). This hypothesis implicitly establishes certain boundary
ci
f
conditions on the velocity field in the material surface Γf s . In effect, considering
ra
C d P cs
b
the free surface in Figure 10.18,
a
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 ∂φ ∂η ∂η ∂η ∂φ
m
= + v · ∇φ = − − vx − vy + vz =0, (10.105)
d
dt ∂t ∂t ∂x ∂y ∂z
uu
e
X Th
=1
er
tin
∂η ∂η ∂η
on
.O
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
ee
s gin
Figure 10.19: Free surface of an embankment dam.
t d le En
Equation (10.107) allows identifying, in certain cases, the position of the free
r
ba
ge ro or
surface (once the pressure field is known) as the locus of points in the fluid in
eS m
ci
which the pressure is equal to the atmospheric pressure.
f
ra
C d P cs
b
a
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
s gin
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
r
ba
ge ro or
understood as being in a situation in which the velocity v (x,t), at a given point
eS m
ci
in space, randomly and rapidly fluctuates along time about a mean value v̄ (x,t)
f
ra
C d P cs
(see Figure 10.21). This situation is defined as turbulent flow (or regime).
b
a
Even though the equations of the fluid mechanics problem in general, and the
i
an an n
.A
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 ∂ v z
rs
σy = 2μ −p τyz = τzy = μ + (10.110)
∂y ∂z ∂y
ee
∂ vz ∂ v z ∂ vx
s gin
σz = 2μ −p τzx = τxz = μ +
∂z ∂x ∂z
t d le En
r
Cylindrical coordinates
ba
ge ro or
eS m
ci
∂ vr f ∂ vθ 1 ∂ vr
ra
σr = 2μ −p τrθ = τθ r = μ r +
C d P cs
∂r ∂r r r ∂θ
b
a
i
an an n
1 ∂ v θ vr ∂ v θ 1 ∂ vz
σθ = 2μ + −p τθ z = τzθ = μ +
y ha
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
e
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
∂ v θ 1 ∂ vr
τrθ = τθ r = μ r +
∂r r r ∂θ
sin θ ∂ vφ 1 ∂ vθ
τ θ 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 = r vr + (vθ sin θ ) + (10.114)
r ∂r
2 r sin θ ∂ θ r sin θ ∂ φ
rs
ee
s gin
10.7.2 Continuity Equation
Cartesian coordinates
t d le En
∂ρ ∂ ∂ ∂
r
+ (ρvx ) + (ρvy ) + (ρvz ) = 0
ba
(10.115)
ge ro or
eS m
∂t ∂ x ∂y ∂z
ci
f
ra
C d P cs
b
a
Cylindrical coordinates
i
an an n
∂ρ 1 ∂ 1 ∂ ∂
y ha
.A
Spherical coordinates
m
∂ρ 1 ∂ 2 1 ∂ 1 ∂
d
∂t r ∂ r r sin θ ∂ θ r sin θ ∂ φ
e
X Th
er
tin
on
.O
Cartesian coordinates
2
∂p ∂ vx ∂ 2 vx ∂ 2 vx ∂ vx ∂ vx ∂ vx ∂ vx
− +μ + + + ρb x = ρ + v x + vy + vz
∂x ∂ x2 ∂ y2 ∂ z2 ∂t ∂x ∂y ∂z
2
∂p ∂ vy ∂ vy ∂ vy
2 2 ∂ vy ∂ vy ∂ vy ∂ vy
− +μ + + 2 + ρby = ρ + vx + vy + vz
∂y ∂ x2 ∂ y2 ∂z ∂t ∂x ∂y ∂z
2
∂p ∂ vz ∂ vz ∂ vz
2 2 ∂ 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
∂p ∂ 1 ∂ 1 ∂ 2 vr 2 ∂ vθ ∂ 2 vr
− +μ (rvr ) + 2 − + + ρbr =
∂r ∂r r ∂r r ∂ θ 2 r2 ∂ θ ∂ z2
∂ vr ∂ vr vθ ∂ vr v2θ ∂ vr
=ρ + vr + − + vz
∂t ∂r r ∂θ r ∂z
1∂p ∂ 1 ∂ 1 ∂ 2 vθ 2 ∂ vr ∂ 2 vθ
rs
− +μ (rvθ ) + 2 + 2 + + ρbθ =
r ∂θ ∂r r ∂r r ∂θ2 r ∂θ ∂ z2
ee
∂ vθ ∂ vθ vθ ∂ vθ vr vθ ∂ vθ
s gin
=ρ + vr + + + vz
∂t ∂r r ∂θ r ∂z
t d le En
∂p 1 ∂ ∂ vz 1 ∂ 2 vz ∂ 2 vz
r
− +μ + 2 + 2 + ρbz =
ba
ge ro or
r
eS m
∂z r ∂r ∂r r ∂θ2 ∂z
ci
f
ra
∂ vz ∂ vz vθ ∂ vz ∂ vz
C d P cs
=ρ + vr + + vz
b
a
∂t ∂r r ∂θ ∂z
i
an an n
(10.119)
y ha
le
Spherical coordinates
liv or ec
M
.A
#
∂p ∂ 1 ∂ 2 1 ∂ ∂ vr 1 ∂ 2 vr
m
− +μ r v + sin θ + +
d
r
∂r ∂ r r2 ∂ r r2 sin θ ∂ θ ∂$θ r2 sin2 θ ∂ φ 2
uu
e
∂ 2 ∂ vφ
X Th
2
er
tin
.O
vφ ∂ vr vθ + vφ
2 2
∂ vr ∂ vr vθ ∂ vr
=ρ + vr + + −
C
∂t ∂r r ∂ θ r sin θ ∂ φ r
©
#
1∂p 1 ∂ ∂ vθ 1 ∂ 1 ∂
− +μ 2 r 2 + 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 r ∂ θ r2 sin θ ∂ φ
2
$
vφ ∂ vθ vr vθ vφ cot θ
2
∂ vθ ∂ vθ vθ ∂ vθ
=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 ∂
− +μ 2 r2 + 2 vφ sin θ +
r sin θ ∂ φ r ∂r ∂r r ∂ θ sin θ ∂ θ $
1 ∂ vφ 2
2 ∂ vr 2 cot θ ∂ vθ
+ 2 2 + 2 + + ρbϕ =
r sin θ ∂ φ 2 r sin θ ∂ φ r2 sin θ ∂ φ
∂ vφ ∂ v φ v θ ∂ 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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.)
ρ0
rs
ee
Determine, for a steady-state regime, the exit pressure P2 in terms of the other
s gin
variables shown in the figure.
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
d
uu
ρ dV = 0 .
e
dt
X Th
er
tin
.O
∂ ∂
C
d
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
v 1 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
s gin
t d le En
r
Problem 10.2 – Determine the value per unit of thickness of the horizontal
ba
ge ro or
eS m
force F that must be applied on point B of the semicircular floodgate shown
ci
f
in the figure such that the straight line AB remains vertical. The floodgate can
ra
C d P cs
rotate around the hinge A and separates two different height levels, h and αh,
b
a
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
s gin
by the floodgate. Thus, posing the equilibrium of momentum with respect to the
center of this circumference (see Figure A) results in
t d le En
FR = HR =⇒ H=F.
r
ba
ge ro or
eS m
Imposing now the equilibrium of horizontal forces, knowing that the fluids exert
ci
f
a horizontal pressure with a triangular distribution (see Figure B), yields
ra
C d P cs
b
a
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
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
.
p δ2 S1 . 2 1 ρ S1 2 .
s gin
H+ + = 0+0+ δ =⇒ p = − 1 δ 2 − ρg H
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).
r
ba
ge ro or
Then,
eS m
p = const. ∀x = H =⇒ F = p S1 .
ci
f
ra
.
C d P cs
Finally, the force F is related to δ in the following manner.
b
a
i
an an n
# 2 $
y ha
ρ S1 .
F = S1 − 1 δ 2 − ρg H S1
le
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
H YPOTHESES :
b
a
i
an an n
.A
m
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
not T
C
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 − A . 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
∂p
ee
y − component =⇒ 0=− ⇒ p = p (x) .
∂y
s gin
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
r
ba
ge ro or
which the pressure term depends solely on x whilst the velocity term depends
eS m
ci
exclusively on y. Consequently, these terms must be constant.
f
ra
⎫
C d P cs
b
a
dp d 2 vx ⎪
=μ 2 ⎬
i
d 2 vx
an an n
dp
dx dy =⇒ =μ = k = const.
y ha
⎪
⎭ dx dy2
f (x) = f (y)
le
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
dy 2 μ 2μ
er
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 ⎥
⎢ ⎥
σ (x, y) ≡ ⎢ ⎥
not
⎢μ k ⎥
⎣ 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
rs
P RESSURE BOUNDARY CONDITIONS
ee
In the fluid with density μ1 , the pressure is prescribed for y = h or, directly,
s gin
since p does not depend on y (because the weight of the fluid is neglected), the
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.
r
ba
ge ro or
eS m
4. W = ρ ∗ gt is the weight of a section of the plate with unit length, according
ci
f
ra
to the x- and z-axis. Here, patm = 0 has been considered.
C d P cs
b
a
5. p1 = ρ ∗ gt cos α , ∀x is the projection on the y-axis. Since a unit length has
i
an an n
6. p1 = p2 ∀x is the continuity condition for the pressure in the
le
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
tween the two fluids does not affect the complete tensor σ . Instead, only the
er
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
n1 ≡ [0 , −1 , 0]T n2 ≡ [0 , 1 , 0]T .
not not
and
Hence, the shear stresses must satisfy:
1
2
7. τxy = τxy
y=0 y=0
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
s gin
%
y=0
=⇒ μ1 B1 = μ2 B2
t d le En
k1 = k2 = 0
r
ba
ge ro or
Solving and replacing these values in the expressions for the pressure, velocity
eS m
ci
and stress obtained in b) results in
f
ra
C d P cs
b
a
i
v∗ y μ1
an an n
1+ h μ2
μ2
le
liv or ec
v∗
μ1 v∗ y τxy
1 = τ2 = μ = const.
M
1
.A
xy
v2x (y) = +1 μ1
μ2 1 + μ1 h h 1+
μ2
m
μ2
d
uu
e
X Th
er
tin
unit element of the plate is posed. Three forces act on this element:
.O
1) The force f ∗ that pushes the plate in the positive direction of the x-axis.
C
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∗
s gin
f ∗ = ρ ∗ gt sin α + μ1
μ1
1+ h
t d le En
μ2
r
and, isolating v∗ , produces the velocity in terms of the shear force,
ba
ge ro or
eS m
ci
f
ra
μ1
C d P cs
h
v∗
b
(f ∗ − ρ ∗ gt sin α)
a
= 1+ .
μ1 μ2
i
an an n
y ha
To compute the volume flow rate that flows across the surface B−B , one must
le
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 B was considered, that is,
m
d
uu
h
e
X Th
q= v · n dS = v · n dS = vx (y) dy .
er
tin
BB curved BB straight −h
on
.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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
i
an an n
Figure A Figure B
y ha
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−A 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
will be assumed that vx = 0 in all the zone ABCD of
s gin
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
r
ba
region if there is a high velocity.
ge ro or
eS m
In short, the velocity and pressure fields are assumed to be of the form
ci
f
ra
not T
C d P cs
v ≡ 0 , vy (x, y) , 0
b
and p = p (x, y) .
a
i
an an n
∂ vy
M
.A
=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
∂p ∂ vy
2
⎪
⎭ ∂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
s gin
P RESSURE BOUNDARY CONDITIONS
t d le En
3. p (y) = patm = 0
r
ba
ge ro or
y=m+h
eS m
ci
f
ra
C d P cs
VOLUME FLOW RATE BOUNDARY CONDITIONS
b
a
i
an an n
In an incompressible fluid the entrance and exit volume flow rates are the same,
y ha
.A
S
.
The piston descends at a velocity δ and, thus, its cross-section is introduced into
m
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
a
Qout = 2 v · n dS = 2 vy (x) dx .
Sa 0
Finally, equating the entrance and exit volume flow rates results in:
a .
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
a a
rs
k 2 k a3 a2 .
2 vy (x) dx = 2 x +C2 x dx = 2 +C2 = δL =⇒
ee
2μ 2μ 3 2
s gin
0 0
.
6μ . δ L
t d le En
k = − 3 δ (a + L) and C2 = 2+3
a a a
r
ba
ge ro or
eS m
ci
f
ra
6μ .
C d P cs
p (y) =0 =⇒ k (m + h)+C1 = 0 =⇒
b C1 = δ (a + L) (m + h)
a
a3
i
y=m+h
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
. 2 δ
.
X Th
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 ⎥
σ ≡ ⎢ μ y −p 0 ⎥ where
⎢ ∂x
⎣
⎥
⎦ ∂ vy . x L 2
μ = μ δ −6 3 (a + L) + 3 2 + .
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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
postive sign
f
ra
criterion
C d P cs
b
a
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
. a
e
L 2
X Th
∗ ∗
er
τ1 = τ2 = −τxy = −μ δ −6 3 (a + L) + 3 2 + =⇒
tin
x=a a a a
on
.O
.
μδ L
τ1∗ = τ2∗ = 3 +4 .
C
a a
©
.
Imposing the equilibrium of forces (since δ is a constant velocity),
F = Lσ ∗ + h τ1∗ + h τ2∗ =⇒
2
. 2μh 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
ee
s gin
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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 W 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
s gin
placed at a distance h from the plate, as shown in the figure. The plate and the
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.
r
b) The integration constants, by applying the adequate boundary conditions.
ba
ge ro or
eS m
c) The forces per unit of surface F1 and F2 exerted on the plate and the top
ci
f
ra
boundary needed to produce the described motion.
C d P cs
b
a
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
ee
s gin
t d le En
r
ba
ge ro or
Additional hypotheses:
eS m
ci
f
1) The flow is assumed to be parallel to the x-y plane.
ra
C d P cs
2) Inertial forces can be neglected.
b
a
i
3) The volume flow rate, Q, and the components of the force, V and H, are
an an n
4) The weight of the plate and the atmospheric pressure are negligible.
le
liv or ec
M
.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
s gin
slow to neglect the inertial forces.
2) The effect of the lateral walls (fluid-
t d le En
wall friction effects) can be neglected.
r
3) The velocity field varies linearly with
ba
ge ro or
eS m
the distance to the inferior surface.
ci
f
4) Assume a steady-state regime.
ra
C d P cs
b
a
i
an an n
10.9 – The cross-section of a cylindrical piston ABA B that slides inside a con-
y ha
in the figure. The motion of the piston, at a velocity δ , causes the fluid to flow
through the pipe DED E .
M
.A
a) Determine the pressure and velocity fields of the fluid in zone DED E in
m
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 BB 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
CH.11. VARIATIONAL
PRINCIPLES
Multimedia Course on Continuum Mechanics
Overview
Introduction Lecture 1
Functionals
Lecture 2
Gâteaux Derivative
Extreme of a Functional Lecture 3
Variational Principle
Lecture 4
Variational Form of a Continuum Mechanics Problem
Virtual Work Principle Lecture 5
Virtual Work Principle
Interpretation of the VWP Lecture 6
VWP in Engineering Notation
Minimum Potential Energy Principle
Hypothesis Lecture 7
Potential Energy Variational Principle
2
11.1. Introduction
Ch.11. Variational Principles
3
Computational Mechanics
In computational mechanics problems are solved by cooperation
of mechanics, computers and numerical methods.
This provides an additional approach to problem-solving, besides the
theoretical and experimental sciences.
Includes disciplines such as solid mechanics, fluid dynamics,
thermodynamics, electromagnetics, and solid mechanics.
6
11.2. Functionals
Ch.11. Variational Principles
8
Definition of Functional
Consider a function space X : R
b
∫ u ( x)dx
=X: {u ( x ) : Ω ⊂ R 3
→R m
} F (u )
b
a
u ( x ) : [ a, b ] → R
9
Definition of Gâteaux Derivative
Consider :
=
a function space X : {u( x ) : Ω ⊂ R 3 → R m }
the functional F ( u ) : X → R
a perturbation parameter ε ∈ R
a perturbation direction η( x ) ∈ X
u (x) Ω
P P’
Ω0 ε η(x)
u (x) + ε η(x)
10
Definition of Gâteaux Derivative
The Gâteaux derivative of the functional F ( u ) in the η direction is:
( F ( u + εη ) )
d
( u; η) :
δF= P’
dε ε=0 F (u )
t=0
t
u (x) Ω
P P’
Ω0 ε η(x)
u (x) + ε η(x)
REMARK not
The perturbation direction is often denoted as η = δu .
Do not confuse δu(x) with the differential du(x) .
δu(x) is not necessarily small !!!
11
Example
Find the Gâteaux derivative of the functional t=0 t
F ( u )=: ∫ ϕ (u ) d Ω + ∫ φ (u ) d Γ u (x) Ω
Ω ∂Ω
P P’
Ω0 ε δu ( x )
u ( x ) + ε δu ( x )
Solution :
d d d
δF ( u; δ=
u) F ( u + εδu )= ∫ ϕ ( u + εδu ) d Ω + ∫ φ ( u + εδu ) d Γ=
dε d
ε=0 = ε Ω ε 0 =
d ε ∂Ω ε 0
∂ϕ( u) ∂φ( u)
δF (=
u; δu ) ∫Ω ∂u ⋅ δu d Ω + ∫∂Ω ∂u ⋅ δu d Γ
12
Gâteaux Derivative with boundary
conditions
Consider the function space V :
Γu Γ σ =∅
=V: {u ( x ) u ( x ) : Ω → R m=
; u ( x ) x∈Γ u* ( x )
u
}
=u *
13
Gâteaux Derivative in terms of
Functionals
Consider the family of functionals
Γu Γ σ =∅
=F (u ) ∫ φ (x, u(x), ∇u(x))d Ω
Ω
REMARK
The example showed that for F ( u )=: ∫ φ ( u ) d Ω + ∫ ϕ ( u ) d Γ , the
Ω ∂Ω
∂φ( u) ∂ϕ( u)
F (u)
Gâteaux derivative is δ= ∫Ω ∂u ⋅ δu d Ω + ∫∂Ω ∂u ⋅ δu d Γ .
14
Extrema of a Function
A function f ( x ) : R → R has a local minimum (maximum) at x0
Necessary condition:
df ( x) not
= f= ′ ( x0 ) 0
dx x = x0
Local minimum
The same condition is necessary for the function to have extrema
(maximum, minimum or saddle point) at x0 .
15
Extreme of a Functional. Variational
principle
A functional F ( u ) : V → R has a minimum at u ( x ) ∈ V
δF ( u; δu ) = 0 ∀δ u | δ u x∈Γ = 0
u
δ F ( u; δ u=
) ∫ E(u) ⋅ δ u d Ω + ∫ T(u) ⋅ δ u d Γ= 0 ∀δ u
Ω Γσ
δ u x∈Γ = 0
Variational Principle u
16
11.3.Variational Principle
Ch.11. Variational Principles
17
Variational Principle
Variational Principle:
δ F ( u; δ u ) = ∫ E ⋅ δ u d Ω + ∫ T ⋅ δ u d Γ = 0 ∀δ u REMARK
Ω Γσ
δ u x∈Γ = 0 Note that δ u
is arbitrary.
u
18
Example
Find the Euler-Lagrange equations and the natural and forced boundary
conditions of the functional
b
19
Example - Solution
Find the Euler-Lagrange equations and the natural and forced boundary
conditions of the functional
b
(u )
F= ∫ φ x, u ( x ) , u′ ( x ) dx with u ( x=
) x =a u=
(a) p
a
Solution :
First, the Gâteaux derivative must be obtained.
The function u ( x ) is perturbed:
u ( x ) → u ( x ) + ε η ( x ) not
∀η ( x ) ≡ δ u ( x ) | η ( a ) = ηa =0
u ( x ) → u ( x ) + ε η ( x )
′ ′ ′
This is replaced in the functional:
b
F (u=
+εη) ∫ φ x, u ( x ) + ε η , u ′ ( x ) + ε η ′ dx
a
20
b
F ( u ) = ∫ φ x, u ( x ) , u ′ ( x ) dx
Example - Solution u ( x=
a
(a) p
) x =a u=
ηa = 0
The Gâteaux derivative is re-written as:
b
(u) φ x, u ( x ) , u′ ( x ) dx ; u ( a )
∫= p
a
∂φ d ∂φ
b ∂φ
δ (u; η= ) δ (u; δ u ) = ∫ [ − ( )]δ udx + δ ub
a ∂u dx ∂u′ ∂u′ b
≡δ u
21
Example - Solution
Therefore, the Variational Principle takes the form
b ∂φ d ∂φ ∂φ ∀δ u
δ (u; δ u ) = ∫ − δ u dx + δ u =
0
a ∂u dx ∂u′ ∂u ′ b δ ua = 0
b
∂φ d ∂φ
E ( x, u , u ′ ) ≡ − = 0 ∀x ∈ ( a, b ) Euler-Lagrange Equations
′
∂u dx ∂u
∂φ Natural (Newmann)
T ( x, u , u ′ ) ≡ =
0 boundary conditions
∂u ′ x =b
Essential (Dirichlet)
u ( x) x = a ≡ u (a) =
p
boundary conditions
22
Variational Form of a Continuum
Mechanics Problem
Consider a continuum mechanics problem with local or strong
governing equations given by,
Euler-Lagrange equations
23
Variational Form of a Continuum
Mechanics Problem
The variational form of the continuum mechanics problem consists
in finding a field u ( x ) ∈ X where
=
V: {u ( x ) : V ⊂ R → R u=
3
( x ) u ( x ) on Γ }
m ∗
u
V= {δ u(x) : V ⊂ R → R δ u(x=
0
3
) 0 on Γ }
m
u
fulfilling:
24
Variational Form of a Continuum
Mechanics Problem
REMARK 1
The local or strong governing equations of the continuum mechanics
are the Euler-Lagrange equation and natural boundary conditions.
REMARK 2
The fundamental theorem of variational calculus guarantees that the
solution given by the variational principle and the one given by the local
governing equations is the same solution.
25
11.4. Virtual Work Principle
Ch.11. Variational Principles
26
Governing Equations
Continuum mechanics problem for a body:
Cauchy equation
∂ 2u ( x, t )
∇ ⋅ σ ( x, t ) + ρ0b ( x, t ) =
ρ0 in V
∂t 2
Boundary conditions
u ( x, t ) u∗ ( x, t ) on Γu
=
σ ( x, t ) ⋅ n (=
x, t ) t ∗ ( x, t ) on Γσ
(ε (∇u),t)
27
Variational Principle
The variational principle consists in finding a displacement field
, where= V : {u ( x, t ) : V ⊂ R 3 → R m u=
( x, t ) u∗ ( x, t ) on Γu }
such that the variational principle holds,
∂ 2u
δ W ( u; δ u=
) ∫ [∇ ⋅ σ + ρ (b − 2 )] ⋅ δ u dV + ∫ ( t∗ − σ ⋅ n) ⋅ δ u d Γ= 0 ∀δ u ∈ V0
∂t
V Γσ
=T
=E
where V0=: {δ u ( x ) : V ⊂ R3 → R m δ u ( x=) 0 on Γu }
Note:
is the space of admissible displacements.
is the space of admissible virtual displacements (test functions).
The (perturbations of the displacements ) δ u are termed virtual
displacements.
28
Virtual Work Principle (VWP)
The first term in the variational principle
=a
∂ 2u
) ∫ [∇ ⋅ σ + ρ (b − 2 )] ⋅ δ u dV + ∫ ( t∗ − σ ⋅ n ) ⋅ δ u d Γ= 0 ∀δ u ∈ V0
δ W ( u; δ u=
∂t
V Γσ
=T
=E
Considering that ( ∇ ⋅ σ ) ⋅ δ u= ∇ ⋅ ( σ ⋅ δ u ) − σ : ∇ sδ u
and (applying the divergence theorem):
∫ ( ∇ ⋅ σ ) ⋅ δ=u ⋅ dV ∫ ( n ⋅ σ) ⋅ δ u dΓ − ∫ δ u dV
σ : ∇ s
V Γσ V
29
Virtual Work Principle (VWP)
REMARK 1
The Cauchy equation and the equilibrium of tractions at the boundary
are, respectively, the Euler-Lagrange equations and natural boundary
conditions associated to the Virtual Work Principle.
REMARK 2
The Virtual Work Principle can be viewed as the variational principle
associated to a functional W ( u ) , being the necessary condition to find
a minimum of this functional.
30
Interpretation of the VWP
The VWP can be interpreted as:
δ W ( u=
;δ u ) ∫ ρ ( b
− a ) ⋅ δ u dV + ∫ t *
⋅ δ u d Γ − ∫ σ : ( ∇ s
) dV 0
δ u=
V Γσ V
b* δε
pseudo - virtual
body forces strains
Work by the
Work by the pseudo-body virtual strain.
forces and the contact forces. Internal virtual
External virtual work work
δW ext δ W int
δ W ( u; δ u ) = δ W ext − δ W int = 0 ∀δ u ∈ V0
31
VWP in Voigt’s Notation
Engineering notation uses vectors instead of tensors:
σ x δε x δε x
σ δε δε
y y y
σ z not δε
z δε z
{σ} ∈ R ; {σ} =τ
6
{δε } ∈ R ; {δε} =δγ =2δε
6
σ : δε ={σ} ⋅ {δε} ={δε} ⋅ {σ}
xy xy xy
τ δγ 2δε
xz xz xz
τ yz δγ yz 2δε yz
32
11.5. Minimum Potential Energy Principle
Ch.11. Variational Principles
33
Hypothesis
An explicit expression of the functional W in the VWP can only be
obtained under the following hypothesis:
1. Linear elastic material. The elastic potential is:
1 ∂uˆ (ε )
=
uˆ (ε ) ε:C:ε =σ = C:ε
2 ∂ε
2. Conservative volume forces. The potential for the quasi-static
case (a = 0) under gravitational ∂φ(u)
forces and constant density is: φ ( u ) =
− ρ b ⋅ u =−ρ b
∂u
3. Conservative surface forces. The potential is: ∂G(u)
G ( u ) =−t ∗ ⋅ u =−t ∗
∂u
Then a functional, total potential energy, can be defined as
U ( u=
) ∫ uˆ (ε )dV + ∫ φ ( u ) dV + ∫ G ( u ) dS
V V Γσ
U ( u; δ u )
δ= ∫ σ : δε dV − ∫ ρ ( b − a ) ⋅ δ u dV − ∫ ⋅ δ u dΓ ∀δ u ∈ V0
*
t
V V Γσ
35
Minimization of the Potential Energy
The VWP is obtained as the variational principle associated with
this functional U , the potential energy.
deriving from a
The potential energy is potential
1
=U(u) ∫V 2 ε (u ) : C : (u ) dV − ∫V ρ ( b − a (u ) ) ⋅ u dV − ∫Γ ⋅ u d Γ
t *
δ U(u; δ u) = 0 ∀δ u ∈ V0
36
Chapter 11
Variational Principles
rs
ee
s gin
11.1 Governing Equations
t d le En
r
Variational calculus is a mathematical tool that allows working with the so-
ba
ge ro or
eS m
called integral or weak form of the governing differential equations of a problem.
ci
f
Given a system of differential equations, which must be verified in local form
ra
C d P cs
(point by point) for a certain domain, the variational principles allow obtain-
b
a
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
d
uu
e
X Th
er
tin
C
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
b b
F (u) = u (x) dx , G (u) = u (x) dx
a a
b
rs
and H (u) = F x, u (x) , u (x) dx .
ee
a
s gin
t d le En
r
ba
ge ro or
Definition 11.2.
Consider the (scalar, vector
or tensor) function
eS m
space X := u (x) u (x) : R3 ⊃ Ω → Rn on a domain Ω and a
ci
f
ra
functional F (•) : X → R.
C d P cs
b
a
Consider the two functions u, η ∈ X and the (perturbation) param-
i
an an n
.A
de f d
d
δ F (u; η ) = F (u + ε η ) .
uu
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
δ F (u; δ u) =
F (u + εδ u) = φ (u + εδ u) dΩ +
dε ε=0 dε ε=0
Ω
rs
d
+ ϕ (u + εδ u) dΓ =
ee
dε ε=0
∂Ω
s gin
⎡ ⎤
t d le En
⎢ ∂ φ (u + εδ u) d (u + εδ u) ⎥
⎢ ⎥
r
=⎢ · dΩ ⎥ +
ba
ge ro or
⎣ ∂u dε ⎦
eS m
Ω
ci
⎡ δu f ε =0 ⎤
ra
C d P cs
b
a
⎢ ∂ ϕ (u + εδ u) d (u + εδ u) ⎥
i
an an n
⎢ ⎥
+⎢ · dΓ ⎥ =⇒
y ha
⎣ ∂u dε ⎦
le
∂Ω
liv or ec
δu ε =0
M
.A
⎡ ⎤
m
∂ φ (u) ∂ ϕ (u)
δ⎣ ϕ (u) dΓ ⎦ =
d
φ (u) dΩ + · δ u dΩ + · δ u dΓ
uu
∂u ∂u
e
X Th
Ω ∂Ω Ω ∂Ω
er
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
ee
s gin
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
r
space V (u + ε δ u ∈ V). In this case, if u + ε δ u ∈ V,
ba
ge ro or
eS m
ci
f
ra
(u + ε δ u) = u∗ =⇒ u + ε δ u = u∗ =⇒ ε δ u = 0
C d P cs
b
a
x∈Γu x∈Γ x∈Γu x∈Γu
i
u
an an n
u∗
y ha
and the perturbation δ u must satisfy δ u =0.
le
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
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
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
b
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
s gin
u(x) → u(x) + ε η(x)
not
∀η (x) ≡ δ u (x) η (a) = ηa = 0 =⇒
t d le En
u (x) → u (x) + ε η (x)
r
ba
ge ro or
eS m
b
ci
F (u + ε η) = f
φ x, u (x) + ε η, u (x) + ε η dx
ra
C d P cs
b
a
a
i
an an n
b
le
liv or ec
d ∂φ ∂φ
δ F (u; η) = F (u + ε η) = η + η dx .
M
.A
dε ε=0 ∂u ∂u
a
m
b x=b b
er
tin
∂φ ∂φ d ∂φ
η dx = η − η dx =
∂ u ∂ u x=a dx ∂ u
on
.O
a a
b
∂ φ
C
∂φ d ∂φ
= ηb − ηa − η dx =
©
∂u ∂ u
dx ∂ u
x=b x=a
=0 a
b
∂φ d ∂φ
= ηb − η dx ,
∂u dx ∂ u
x=b a
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, u ) ≡ − ∀x ∈ (a, b)
∂ u dx ∂ u
.
∂ φ
T (x, u, u ) ≡
∂u
x=b
rs
11.1.2 Extrema of the Functionals. Variational Principles.
ee
Euler-Lagrange Equations
s gin
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
r
ba
ge ro or
f (x0 ) ≤ f (x) ∀x ∈ R . (11.4)
eS m
ci
f
ra
The necessary condition for f to have an extrema (maximum, minimum or sad-
C d P cs
b
a
dle point) at x = x0 is known to be
i
an an n
d f (x)
y ha
not
= f (x0 ) = 0 .
le
(11.5)
dx
liv or ec
x=x0
M
.A
when
e
F (u) ≤ F (v) ∀v ∈ V ,
X Th
(11.6)
er
tin
.O
δ F (u; δ u) = 0 ∀δ u δ u =0 . (11.7)
x∈Γu
Variational principle
δ F (u; δ u) = E · δ u dΩ + T · δ u dΓ = 0 (11.8)
Ω Γσ ∀δ u 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
Governing Equations 519
E (x) = 0 ∀x ∈ Ω
⇐⇒
T (x) = 0 ∀x ∈ Γσ
rs
ee
s gin
Proof (indicative1 )
t d le En
Consider the following choice for δ u (x).
⎧
r
⎨ E (x) ∀x ∈ Ω
ba
ge ro or
eS m
ci
δ u (x) = 0 ∀x ∈ Γu
⎩ f
ra
C d P cs
T (x) ∀x ∈ Γσ
b
a
i
an an n
le
liv or ec
.A
Ω
≥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-
s gin
tions denoted as natural or Neumann boundary conditions. Together
with the conditions (11.1),
t d le En
u (x) = u∗ (x) ∀x ∈ Γu ,
r
ba
ge ro or
eS m
named essential or Dirichlet boundary conditions, they define a sys-
ci
f
tem whose solution u (x) is an extreme of the functional F.
ra
C d P cs
b
a
i
an an n
y ha
le
liv or ec
b
d
with
e
x=a
X Th
er
a
tin
on
.O
Solution
C
b
∂φ d ∂φ ∂ φ
δ F u; δ u = − δ u dx + δ ub ,
∂ u dx ∂ u ∂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 :
∂φ d ∂φ
E (x, u, u ) ≡ − =0 ∀x ∈ (a, b)
∂ u dx ∂ u
rs
x=b
ee
Essential boundary conditions
:
s gin
u (x) = u (a) = p
x=a
t d le En
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
11.2 Virtual Work Principle (Theorem)
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)
V := ut (x) : V → R3 ut (x) = 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
ee
s gin
Figure 11.2: Definition of the material volume Vt .
t d le En
Consider now the variational principle
r
ba
ge ro or
eS m
ci
f
∇ · σ (u) + ρ (b − a (u)) · δ u dV +
ra
δ W (u; δ u) =
C d P cs
b
a
V
i
E
an an n
(11.14)
y ha
+ t∗ − σ (u) · n · δ u dΓ = 0 ; ∀δ u (x) δ u =0,
le
liv or ec
x∈Γu
Γσ
M
.A
T
m
X Th
er
Virtual displacements: δ u : V → R3 δ u
tin
=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,
s gin
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) .
r
2 2
ba
ge ro or
eS m
(11.18)
ci
Introducing (11.18) in (11.17)3 produces
f
ra
C d P cs
b
a
i
σ · δ u) − σ : (δ u ⊗ ∇) =
(∇ · σ ) · δ u = ∇ · (σ
an an n
y ha
σ · δ u) − σ : ∇s δ u − σ : ∇a δ u
= ∇ · (σ =⇒ (11.19)
le
liv or ec
=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
(∇ · σ ) · δ u dV = σ · δ u) dV −
∇ · (σ σ : ∇s δ u dV =
C
V V V
= σ · δ u) dΓ −
n · (σ σ : ∇s δ u dV =
∂V =Γu Γσ
V
= (n · σ ) · δ u dΓ + (n · σ ) · δ u dΓ − σ : ∇s δ u dV =⇒
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
(∇ · 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
δ W (u; δ u) = ∇ · σ + ρ (b − a) · δ u dV + t∗ − (σ
σ · n) · δ u dΓ =
rs
V Γσ
ee
∗
= (∇ · σ ) · δ u dV + ρ (b − a) · δ u dV + t · δ u dΓ − σ · n) · δ u dΓ =
(σ
s gin
V V Γσ Γσ
∗
t d le En
=− σ : ∇ δ u dV +
s
ρ (b − a) · δ u dV + t · δ u dΓ = 0 =⇒
r
V V Γσ
ba
ge ro or
eS m
(11.23)
ci
f
ra
Virtual Work Principle
C d P cs
b
a
δ W (u; δ u) = ρ (b − a) · δ u dV + t∗ · δ u dΓ
i
an an n
(11.24)
y ha
V
Γσ
le
∀δ u (x) δ u = 0
liv or ec
− σ : ∇s δ u dV = 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
s gin
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).
r
ba
ge ro or
Consider, in addition, the virtual (fictitious) configuration Vt+δt corresponding
eS m
to time t + δt, separated from the real configuration by a virtual displacement
ci
f
ra
field (11.15)
C d P cs
b
a
i
an an n
Virtual displacements: δ u δ u =0 . (11.25)
y ha
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
Internal δ Wint = σ : δ ε dV = σ : ∇s δ u dV .
ee
(11.27)
virtual work:
s gin
V V
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-
r
val [t, t + δt], the work performed by these forces (external virtual work) results
ba
ge ro or
eS m
in
ci
f
ra
C d P cs
b
a
External δ Wext = ρ (b − a) · δ u dV = t∗ · δ u dV
i
an an n
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
δW = σ : δ ε dV − ρ (b − a) · δ u dV + t∗ · δ u dΓ = 0
on
.O
V V Γσ
Total
virtual
C
work Internal
virtual External virtual
work δ Wint work (δ Wext )
©
(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 z ⎥
σ} ≡ ⎢ z ⎥ {δ ε } ∈ R6 ; {δ ε } ≡ ⎢ z ⎥ = ⎢
not not
σ } ∈ R6 ; {σ
{σ ⎥
⎢ τxy ⎥ ⎢ δ γxy ⎥ ⎢ 2δ εxy ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣ τxz ⎦ ⎣ δ γxz ⎦ ⎣ 2δ εxz ⎦
rs
τyz δ γyz 2δ εyz
ee
(11.30)
s gin
They satisfy the equality
"
t d le En
σ : δ ε = {σ σ } · {δ ε } = {δ ε } · {σ
σ} ,
r
(11.31)
ba
ge ro or
σi j δ εi j = σm δ εm = δ εm σm i, j ∈ {1, 2, 3} , m ∈ {1, ..6} .
eS m
ci
f
ra
C d P cs
Replacing (11.31) in the VWP (11.29) results in
b
a
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
Total V V Γσ
virtual
m
work Internal
virtual External virtual
d
(11.32)
e
X Th
=⇒ δ W = δ Wint − δ Wext
er
tin
.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
2
s gin
potential: 2
σ (11.33)
∂ û (εε )
t d le En
=C :ε =σ
∂ε
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
2) The body forces ρb∗ (x,t) are conservative
b
a
i
That is, these body forces derive from a potential φ (u) and, thus,
an an n
y ha
le
∂ φ (u)
liv or ec
= −ρb∗ = −ρ (b − a) . (11.34)
∂u
M
.A
m
d
uu
e
er
stant density,
on
.O
b (x,t) ≡ [ 0, 0, −g ]T = const.
not
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.
s gin
∂u
In this case, the potential of the surface forces is
t d le En
∂ G (u)
r
G (u) = −t∗ · u = −t∗ .
ba
=⇒
ge ro or
eS m
∂u
ci
f
ra
C d P cs
b
a
i
an an n
y ha
Under the above circumstances, the following functional, named total poten-
le
liv or ec
.A
Potential energy
m
d
uu
(11.36)
X Th
er
Γσ
tin
V V
Total
potential
energy Elastic Potential energy Potential energy 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
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
boundary (11.13), is now the Gateaux variation of the potential
s gin
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
r
1
ba
ε : C : ε dV − ρ (b − a) · u dV − t∗ · u dΓ ,
ge ro or
U (u) =
eS m
2
ci
V V
f Γσ
ra
û (εε )
C d P cs
b
a
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.
s gin
t d le En
r
Solution
ba
ge ro or
eS m
ci
A linear elastic material is a particular type of hyperelastic material and, thus,
f
ra
there must exist an elastic potential of the type
C d P cs
b
a
∂W
i
an an n
∃ 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
d
uu
e
∂ Φ (u)
X Th
er
∃ Φ (u) ρb = − =⇒ δ Φ = −ρb · δ u
tin
∂u
on
.O
Now, the given expression of the Virtual Work Principle can be rewritten as
C
δ W dV0 + δ Φ dV0 + δ G dΓ = 0 ∀ δu δu = 0 in Γu ,
©
V0 V0 Γσ
⎛ ⎞
δ⎝ W dV0 + Φ dV0 + G dΓ ⎠ = 0 ∀ δu δu = 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 ∀ δu δu = 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
U (u) = ε (u) : C : ε (u) dV0 − ρb · (u) dV0 − t∗ · u dΓ and
ee
2
V0 V0 Γσ
s gin
1
U (u + δ u) = ε (u + δ u) : C : ε (u + δ u) dV0 − ρb · (u + δ u) dV0
t d le En
2
r
V0 V0
ba
ge ro or
t∗ · (u + δ u) dΓ .
eS m
−
ci
f
ra
Γσ
C d P cs
b
a
Taking into account
i
an an n
ε (u + δ u) = ε (u) + ε (δ u)
y ha
M
.A
1 1
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
t∗ · δ u dΓ
X Th
+ ε (δ u) : C : ε (δ u) dV0 − ρb · δ u dV0 −
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
U (u + δ u) − U (u) = σ : δ ε dV0 + ε (δ u) : C : ε (δ u) dV0
2
V0 V0
− 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
U (u + δ u) − U (u) = ε (δ u) : C : ε (δ u) dV0 .
2
V0
rs
ee
U (u + δ u) − U (u) ≥ 0
s gin
t d le En
and, thus, the potential energy is seen to have a minimum in the equilibrium
state.
r
ba
ge ro or
eS m
ci
f
ra
C d P cs
b
a
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
TENSOR ALGEBRA
Multimedia Course on Continuum Mechanics
Overview
Introduction to tensors Lecture 1
Integral Theorems
Lecture 7
References
2
Introduction
SCALAR ρ , θ , ...
v
VECTOR v, f , ...
MATRIX σ , ε, ...
? C, ...
4
Concept of Tensor
A TENSOR is an algebraic entity with various components
which generalizes the concepts of scalar, vector and matrix.
5
Order of a Tensor
The order of a tensor is given by the number of indexes
needed to specify without ambiguity a component of a tensor.
6
Cartesian Coordinate System
Given an orthonormal basis formed by three mutually
perpendicular unit vectors:
eˆ1 ⊥ eˆ 2 , eˆ 2 ⊥ eˆ 3 , eˆ 3 ⊥ eˆ1
Where:
=eˆ1 1=
, eˆ 2 1=
, eˆ 3 1
Note that
1 if i = j
=
eˆ i ⋅ eˆ j = δ ij
0 if i ≠ j
7
Indicial or (Index) Notation
Tensor Algebra
10
Tensor Bases – VECTOR
A vector v can be written as a unique linear combination of the
three vector basis eˆ i for i ∈ {1, 2,3} .
v
v = v1eˆ1 + v 2eˆ 2 + v3eˆ 3
v3
In matrix notation:
v1
[ v ] = v2 v1
v3 v2
In index notation:
11
Tensor Bases – 2nd ORDER TENSOR
A 2nd order tensor A can be written as a unique linear combination
of the nine dyads eˆ i ⊗ eˆ j ≡ eˆ i eˆ j for i, j ∈ {1, 2,3} .
12
Tensor Bases – 2nd ORDER TENSOR
= A11 ( eˆ1 ⊗ eˆ1 ) + A12 ( eˆ1 ⊗ eˆ 2 ) + A13 ( eˆ1 ⊗ eˆ 3 ) +
A
+ A21 ( eˆ 2 ⊗ eˆ1 ) + A22 ( eˆ 2 ⊗ eˆ 2 ) + A23 ( eˆ 2 ⊗ eˆ 3 ) +
+ A31 ( eˆ 3 ⊗ eˆ1 ) + A32 ( eˆ 3 ⊗ eˆ 2 ) + A33 ( eˆ 3 ⊗ eˆ 3 )
In matrix notation:
A11 A12 A13
[ A ] = A21 A22 A23
A31 A32 A33
In index notation:
=A ∑ Aij ( eˆ i ⊗ eˆ j )
ij
tensor as a
physical entity
13
Tensor Bases – 3rd ORDER TENSOR
A 3rd order tensor A can be written as a unique linear combination
of the 27 tryads eˆ i ⊗ eˆ j ⊗ eˆ k ≡ eˆ i eˆ j eˆ k for i, j , k ∈ {1, 2,3}.
A A 111 ( eˆ1 ⊗ eˆ1 ⊗ eˆ1 ) + A 121 ( eˆ1 ⊗ eˆ 2 ⊗ eˆ1 ) + A 131 ( eˆ1 ⊗ eˆ 3 ⊗ eˆ1 ) +
=
+ A 211 ( eˆ 2 ⊗ eˆ1 ⊗ eˆ1 ) + A 221 ( eˆ 2 ⊗ eˆ 2 ⊗ eˆ1 ) + A 231 ( eˆ 2 ⊗ eˆ 3 ⊗ eˆ1 ) +
+ A 311 ( eˆ 3 ⊗ eˆ1 ⊗ eˆ1 ) + A 321 ( eˆ 3 ⊗ eˆ 2 ⊗ eˆ1 ) + A 331 ( eˆ 3 ⊗ eˆ 3 ⊗ eˆ1 ) +
+ A 112 ( eˆ1 ⊗ eˆ1 ⊗ eˆ 2 ) + A 122 ( eˆ1 ⊗ eˆ 2 ⊗ eˆ 2 ) + ...
Alternatively, this could have been written as:
14
Tensor Bases – 3rd ORDER TENSOR
A A 111 ( eˆ1 ⊗ eˆ1 ⊗ eˆ1 ) + A 121 ( eˆ1 ⊗ eˆ 2 ⊗ eˆ1 ) + A 131 ( eˆ1 ⊗ eˆ 3 ⊗ eˆ1 ) +
=
+ A 211 ( eˆ 2 ⊗ eˆ1 ⊗ eˆ1 ) + A 221 ( eˆ 2 ⊗ eˆ 2 ⊗ eˆ1 ) + A 231 ( eˆ 2 ⊗ eˆ 3 ⊗ eˆ1 ) +
A 311 ( eˆ 3 ⊗ eˆ1 ⊗ eˆ1 ) + A 321 ( eˆ 3 ⊗ eˆ 2 ⊗ eˆ1 ) + A 331 ( eˆ 3 ⊗ eˆ 3 ⊗ eˆ1 ) +
+
+ A 112 ( eˆ1 ⊗ eˆ1 ⊗ eˆ 2 ) + A 122 ( eˆ1 ⊗ eˆ 2 ⊗ eˆ 2 ) + ...
In matrix notation:
A
A 112113 AA 123 A 133
A 132
A
111 A 122 A
A A 213 AA 223 AA 233
121 131
[A ] = A211212 A 221222 A 231232
A A 313 AA 323 AA 333
A311312 A 321322 A 331332
15
Tensor Bases – 3rd ORDER TENSOR
A A 111 ( eˆ1 ⊗ eˆ1 ⊗ eˆ1 ) + A 121 ( eˆ1 ⊗ eˆ 2 ⊗ eˆ1 ) + A 131 ( eˆ1 ⊗ eˆ 3 ⊗ eˆ1 )
=
+ A 211 ( eˆ 2 ⊗ eˆ1 ⊗ eˆ1 ) + A 221 ( eˆ 2 ⊗ eˆ 2 ⊗ eˆ1 ) + A 231 ( eˆ 2 ⊗ eˆ 3 ⊗ eˆ1 ) +
+ A 311 ( eˆ 3 ⊗ eˆ1 ⊗ eˆ1 ) + A 321 ( eˆ 3 ⊗ eˆ 2 ⊗ eˆ1 ) + A 331 ( eˆ 3 ⊗ eˆ 3 ⊗ eˆ1 ) +
+ A 112 ( eˆ1 ⊗ eˆ1 ⊗ eˆ 2 ) + A 122 ( eˆ1 ⊗ eˆ 2 ⊗ eˆ 2 ) + ...
In index notation:
=A ∑ A ( eˆ ⊗ eˆ =
ijk
ijk ⊗ eˆ )
i j k
= A ijk ( eˆ i ⊗ eˆ j ⊗ eˆ k ) ≡ A ijk eˆ i eˆ j eˆ k
tensor as a
physical entity
16
Repeated-index (or Einstein’s) Notation
The Einstein Summation Convention: repeated Roman indices are
summed over. 3
i is a mute
index
ai bi = ∑ ai bi = a1b1 + a2b2 + a3b3
i =1
i is a talking 3
index and j is a Aij b j = ∑ Aij b j = Ai1b1 + Ai 2b2 + Ai 3b3
mute index j =1
A “MUTE” (or DUMMY) INDEX is an index that does not appear in a
monomial after the summation is carried out (it can be arbitrarily changed
of “name”).
A “TALKING” INDEX is an index that is not repeated in the same
monomial and is transmitted outside of it (it cannot be arbitrarily changed
of “name”). REMARK
An index can only appear up
to two times in a monomial.
18
Repeated-index (or Einstein’s) Notation
∂xi i =1 ∂xi ∂x j ∂x j j =1 ∂x j ∂x j j =1 ∂x j
i ,i ij , j
5. The number of talking indices indicates the order of the tensor result
19
Kronecker Delta δ
The Kronecker delta δij is defined as:
1 if i = j
δ ij =
0 if i ≠ j
Both i and j may take on any value in
Only for the three possible cases where i = j is δij non-zero.
0 if i ≠ j (δ12 =
δ13 =
δ 21... = 0)
REMARK
δ ij = δ ji Following Einsten’s notation: δ ii = δ11 + δ 22 + δ 33 = 3
Kronecker delta serves as a replacement operator:
=δ ij u j u=
i , δ ij A jk Aik
20
Levi-Civita Epsilon (permutation) e
The Levi-Civita epsilon eijk is defined as:
21
Example
Prove the following expression is true:
eijk eijk = 6
23
Example - Solution
k =1 k =2 k =3
eijk eijk = e111e111 + e112e112 + e113e113 + j =1
=1
+ e121e121 + e122e122 + e123e123 + j = 2
i =1 = −1
+ e131e131 + e132e132 + e133e133 + j = 3
= −1
+ e211e211 + e212e212 + e213e213 +
i=2 + e221e221 + e222e222 + e223e223 +
=1
+ e231e231 + e232e232 + e233e233 +
=1
+ e311e311 + e312e312 + e313e313 +
= −1
i=3 + e321e321 + e322e322 + e323e323 +
24
Vector Operations
Tensor Algebra
25
Vector Operations
Sum and Subtraction. Parallelogram law.
a+b = b+a = c c=
i ai + bi
a−b =d d=
i ai − bi
Scalar multiplication
αa =
b=α a1eˆ1 + α a2eˆ 2 + α a3eˆ 3 bi = α ai
26
Vector Operations
Scalar or dot product yields a scalar
where θ is the angle
u⋅v =u v cos θ between the vectors u and v
In index notation:
i =3
u ⋅ v= ui eˆ i ⋅ v j eˆ j= ui v j eˆ i ⋅ eˆ j= ui v jδ ij= ui v i = ∑ i i = [ u] [ v ]
T
u v
i =1
u v δ ij = 0 (i ≠ j )
= 1=
Norm of a vector ( j i)
u =(u ⋅ u ) =( uiui )
12 12
u = u ⋅ u = ui eˆ i ⋅ u j eˆ j = ui u jδ ij = ui ui
2
27
Vector Operations
Some properties of the scalar or dot product
u ⋅ v = v ⋅u
u⋅0 = 0
u ⋅ (α v + β w )= α ( u ⋅ v ) + β ( u ⋅ w ) Linear operator
u ⋅u > 0 u≠0
=
u ⋅u 0 = u 0
u=⋅ v 0, u ≠ 0, v ≠ 0 u⊥v
28
Vector Operations
Vector product (or cross product ) yields another vector
c =a × b =−b × a
c = a b sin θ where θ is the angle
between the vectors a and b
0 ≤θ ≤π
In index notation:
=
c c= ˆ eijk a j bk eˆ i
i ei ⇒ =
ci eijk a j bk i ∈
i =1 i=2 i=3
eˆ1 eˆ 2 eˆ 3
= det a1 a2 a3
symb
c=( a2b3 − a3b2 ) eˆ1 + ( a3b1 − a1b3 ) eˆ 2 + ( a1b2 − a2b1 ) eˆ 3
a2b3 + e132 a3b2 e231 a3b1 + e213 a1b3 e312 a1b2 + e321 a2b1 b1 b2 b3
e
123
=1 =−1 =1 = −1 =1 = −1
29
Vector Operations
Some properties of the vector or cross product
u × v =− ( v × u )
u ×=
v 0, u ≠ 0, v ≠ 0 u || v
u × ( a v + b w ) = au × v + b u × w Linear operator
30
Vector Operations
Tensor product (or open or dyadic product) of two vectors:
A = u ⊗ v ≡ uv
Also known as the dyad of the vectors u and v, which results in a 2nd
order tensor A.
Deriving the tensor product along an orthonormal basis {êi}:
A = ( u ⊗ v ) = ( ui eˆ i ) ⊗ ( v j eˆ j ) = ui v j ( eˆ i ⊗ eˆ j ) = Aij ( eˆ i ⊗ eˆ j )
31
Vector Operations
Some properties of the open product:
(u ⊗ v ) ≠ ( v ⊗ u )
( u ⊗ v ) ⋅ w =u ⊗ ( v ⋅ w ) =u ( v ⋅ w ) =( v ⋅ w ) u
u ⊗ (α v + β w ) = α u ⊗ v + β u ⊗ w Linear operator
( u ⊗ v )( w ⊗ x ) = ( u ⊗ x )( v ⋅ w )
u ⋅ ( v ⊗ w ) = (u ⋅ v ) ⊗ w = (u ⋅ v ) w = w (u ⋅ v )
32
Example
Prove the following property of the tensor product is true:
u ⋅ ( v ⊗ w ) = (u ⋅ v ) ⊗ w
33
Tensor Operations
Tensor Algebra
39
Tensor Operations
Summation (only for equal order tensors)
A+B =B+A =C C=
ij Aij + Bij
40
Tensor Operations
Dot product (.) or single index contraction product
A ⋅ b =
c ci = Aij b j Index “j” disappears (index
2nd 1st 1st contraction)
order order order
A ⋅ B =
C Cik = Aij B jk Index “j” disappears (index
2nd 2nd 2nd contraction)
order order order
REMARK
A⋅B ≠ B⋅A A⋅A =
A2
41
Tensor Operations
Some properties:
A ⋅ (α b + β c )= α A ⋅ b + β A ⋅ c Linear operator
1⋅ u = u ⋅1 = u
1 = δ ij e j ⊗ ei = ei ⊗ ei 1 0 0
[1] = 0 1 0
[1]ij = δ ij
0 0 1
42
2nd Order Tensor Operations
Some properties:
1⋅ A = A = A ⋅1
A ⋅ ( B + C) = A ⋅ B + A ⋅ C
A ⋅ ( B ⋅ C ) =( A ⋅ B ) ⋅ C = A ⋅ B ⋅ C
A⋅B ≠ B⋅A
43
Example
When does the relation n ⋅ T = T ⋅ n hold true ?
44
2nd Order Tensor Operations
Transpose
A11 A12 A13
[ A ] = A21 A22 A23 [ AT ]
A11
A
A21
A22
A31
A32
(A )
T T
=A
12 ( A ⋅ B ) =BT ⋅ AT
T
49
2nd Order Tensor Operations
Some properties
A : B= Tr ( AT ⋅ B )= Tr ( BT ⋅ A )= Tr ( A ⋅ BT )= Tr ( B ⋅ AT )= B : A
1= =
: A Tr A A :1
A : ( B ⋅ C ) = ( BT ⋅ A ) : C = ( A ⋅ CT ) : B
A : ( u ⊗ v ) =u ⋅ ( A ⋅ v )
(u ⊗ v ) : ( w ⊗ x) = (u ⋅ w ) ⋅ ( v ⋅ x)
REMARK
A : B= C : B ⇒ A= C
50
2nd Order Tensor Operations
51
Tensor Operations
Norm of a tensor is a non-negative real number defined by
(A : A) ( Aij Aij )
12
= = ≥0
12
A
A ⋅⋅ B= Tr ( A ⋅ B )= Tr ( B ⋅ A )= B ⋅⋅A
REMARK
1 ⋅⋅A= TrA= A ⋅⋅1
A : B ≠ A ⋅⋅B
Unless one of the two
tensors is symmetric.
52
Example
Prove that:
A : B Tr ( AT ⋅ B )
=
⋅⋅ B Tr ( A ⋅ B )
A=
53
2nd Order Tensor Operations
Determinant yields a scalar
A11 A12 A13
[ A ] det A21 A23 eijk A1i A
1
= =
det A det A22 = = A
2 j 3k eijk epqr Api Aqj Ark
6
A A32 A33
Some properties: 31
REMARK
det ( A ⋅ B=
) det A ⋅ det B The tensor A is SINGULAR if and
det AT = det A only if det A = 0.
det (α A ) = α 3 det A A is NONSINGULAR if det A ≠ 0.
Inverse
There exists a unique inverse A-1 of A when A is nonsingular, which
satisfies the reciprocal relation:
A ⋅ A −1 ==
1 A −1 ⋅ A
A=
−1
ik Akj A= −1
ik Akj δ ij i, j , k ∈ {1, 2,3}
55
Example
Prove that det A = eijk A1i A2 j A3.k
57
Differential Operators
Tensor Algebra
64
Differential Operators
A differential operator is a mapping that transforms a field
v ( x ) , A ( x ) ... into another field by means of partial derivatives.
The mapping is typically understood to be linear.
Examples:
Nabla operator
Gradient
Divergence
Rotation
…
65
Nabla Operator
The Nabla operator is a differential operator
“symbolically” defined as:
symbolic
∂ symb. ∂
∇= = eˆ i
∂x ∂xi
In Cartesian coordinates, it can be used as a (symbolic) vector
on its own:
∂
∂
1x
symb.
∂
[∇ ] =
∂x2
∂
∂
3
x
66
Gradient
The gradient (or open product of Nabla) is a differential
operator defined as:
Gradient of a scalar field Φ(x):
Yields a vector
symb.
∂ ∂Φ
[ ∇Φ ]i [
= ∇ ⊗ Φ ]i [ ]i
= ∇ Φ =
∂xi
Φ =
∂xi
i ∈ {1, 2,3}
∂Φ
∇Φ = [∇Φ ] eˆ = ∂Φ eˆ ∇Φ = eˆ i
i i
∂xi
i
∂xi
Gradient of a vector field v(x):
Yields a 2nd order tensor
symb.
∂ ∂v j
[ ∇ ⊗ v ] = [ ∇ ] [ v ] = vj = i, j ∈ {1, 2,3}
ij i j
∂xi ∂xi
∂v j
∇v = ∇ ⊗ v = [∇ ⊗ v ] eˆ ⊗ eˆ = ∂v j =
∇v eˆ i ⊗ eˆ j
eˆ i ⊗ eˆ j ∂xi
ij i j
∂xi
67
Gradient
Gradient of a 2nd order tensor field A(x):
Yields a 3rd order tensor
symb.
∂ ∂A jk
[∇A ]ijk = [∇ ⊗ A ]ijk = [∇ ]i [ A ] jk = A jk = i, j , k ∈ {1, 2,3}
∂xi ∂xi
∇A = ∇ ⊗ A = [∇ ⊗ A ] eˆ ⊗ eˆ ⊗ eˆ = ∂A jk eˆ ⊗ eˆ ⊗ eˆ
ijk i j k
∂xi
i j k
∂A jk
=
∇A eˆ i ⊗ eˆ j ⊗ eˆ k
∂xi
68
Divergence
The divergence (or dot product of Nabla) is a differential
operator defined as :
Divergence of a vector field v(x):
Yields a scalar
symb.
∂ ∂vi ∂vi
∇ ⋅ v = [∇ ]i [ v ]i = vi = ∇⋅v =
∂xi ∂xi ∂xi
Divergence of a 2nd order tensor A(x):
Yields a vector
symb.
∂ ∂A ij
∇ ⋅ A j = [∇ ]i [ A ]ij =
[ ] A ij = j ∈ {1, 2,3}
∂x i ∂xi
∇ ⋅ A = [∇ ⋅ A ] eˆ = ∂A ij eˆ ∂A ij
j j
∂xi
j
∇⋅A = eˆ j
∂xi
69
Divergence
70
Rotation
The rotation or curl (or vector product of Nabla) is a differential
operator defined as:
Rotation of a vector field v(x):
Yields a vector
[ symb. symb.
∂ ∂ vk
∇ × v=]i e [ ∇ ] [ v ]
= e =
v e i ∈ {1, 2,3}
∂ ∂x j
ijk j k ijk k ijk
x j
∂ vk
∇ × v = [∇ × v ]i eˆ = e ∂ v k eˆ ∇ × v = eijk eˆ i
i ijk
∂x j
i
∂x j
Rotation of a 2nd order tensor A(x):
Yields a 2nd order tensor
[ symb.
∂ ∂ A kl
=
]
∇ × A il e= A kl eijk i, j , k ∈ {1, 2,3}
∂ ∂
ijk
x j x j
∂ A kl
∇ × A = [∇ × A ]il eˆ ⊗ eˆ = e ∂ A kl eˆ ⊗ eˆ = ∇ × A eijk
∂x j
eˆ i ⊗ eˆ l
i l ijk
∂x j
i l
71
Rotation
72
Differential Operators - Summary
[∇ ⊗ Φ ]i = [∇ ⊗ v ]ij = [∇ ⊗ A ]ijk =
GRADIENT ∂Φ ∂v ∂A jk
= [∇Φ ]i = [∇v ]ij = j
= [ ]ijk
=∇ A =
∂xi ∂xi ∂xi
∂vi ∂A ij
DIVERGENCE ∇⋅v = [∇ ⋅ A ] j =
∂xi ∂xi
73
Example
Given the vector v =v ( x ) =x1 x2 x3eˆ1 + x1 x2eˆ 2 + x1eˆ 3
determine ∇ ⋅ v, ∇ × v, ∇v.
74
Example - Solution
x1 x2 x3
v =v ( x ) =x1 x2 x3eˆ1 + x1 x2eˆ 2 + x1eˆ 3
[ v ] = x1 x2
x1
Divergence:
∂vi
∇⋅v =
∂xi
75
Example - Solution
Divergence: x1 x2 x3
∇⋅v =
∂vi [ v ] = x1 x2
∂xi
x1
In matrix notation:
x1 x2 x3 symb
T
∂ ∂ ∂
, x1 x2 =
symb symb
∇ ⋅ v = [∇ ] [ v ] = ,
T
76
Integral Theorems
Tensor Algebra
81
Divergence or Gauss Theorem
Given a field A in a volume V with closed boundary surface
∂V and unit outward normal to the boundary n , the
Divergence (or Gauss) Theorem states:
∫ ∇ ⋅A
V
=
dV ∫ ∂V
n ⋅A dS
∫ A ⋅∇=
V
dV ∫
∂V
A ⋅ n dS
Where:
A represents either a vector field ( v(x) ) or a tensor field ( A(x) ).
82
Generalized Divergence Theorem
Given a field A in a volume V with closed boundary surface
∂V and unit outward normal to the boundary n , the
Generalized Divergence Theorem states:
∫ ∇ ∗A
V
=
dV ∫∂V
n ∗A dS
∫ A ∗∇=
V
dV ∫
∂V
A ∗ n dS
Where:
∗ represents either the dot product ( · ), the cross product ( ) or the
tensor product ( ).
A represents either a scalar field ( ϕ(x) ), a vector field ( v(x) ) or a
tensor field ( A(x) ).
83
Example
Use the Generalized Divergence Theorem to show that
∫S
xi n j dS = V δ ij
nj
where xi is the position vector of nj .
∫∂V
A ∗ n=
dS ∫ A ∗∇ dV
V
85
References
Tensor Algebra
88
References
89