Continuum Mechanics PDF

Continuum Mechanics
Introduction
C. Agelet de Saracibar
ETS Ingenieros de Caminos, Canales y Puertos, Universidad Politécnica de Cataluña (UPC), Barcelona, Spain
International Center for Numerical Methods in Engineering (CIMNE), Barcelona, Spain
Introduction > Professors
Professors
Prof. Carlos Agelet de Saracibar
Dr. Ingeniero de Caminos, Canales y Puertos
Professor of Continuum Mechanics and Structural Analysis
Department of Civil and Environmental Engineering
ETS Ingenieros de Caminos, Canales y Puertos
Universidad Politécnica de Cataluña, UPC BarcelonaTech
International Center for Numerical Methods in Engineering (CIMNE)
Office 104 B, Building C1, Campus Norte, UPC BarcelonaTech
Gran Capitán s/n
E-08034 Barcelona, Spain
T +34 93 401 6495
F +34 93 401 1048
E agelet@cimne.upc.edu
W ageletdesaracibar.blogspot.com.es
September 28, 2016 Carlos Agelet de Saracibar 2
Introduction > On-campus Schedule
On-campus Schedule 2016-2017
On-campus Classes Schedule:

Dates: 22 Sep - 7 Oct | M-F | 15:00-18:00
Room: A1-105
Dates: 14, 21 Oct | F | 15:00-18:00

Room: A2-205

Introduction > Methodology
Methodology
Continuum Mechanics: On-campus Course
 Erasmus Mundus Master of Science in Computational
Mechanics
 Master in Numerical Methods in Engineering
 Master in Civil Engineering
 Master in Structural and Construction Engineering
Continuum Mechanics: On-line Course

 Master in Numerical Methods in Engineering

Methodology
Continuum Mechanics: On-campus Course
 Language: english
 Intensive course:
22-09-2016 to 07-10-2016 | M-F | 15:00-18:00 | Room A1 105
14-10-2015 & 21-10-2016 | F | 15:00-18:00 | Room A2 205
 Mandatory class attendance
 Lecture notes in pdf and videos at the Master website
 Interactive lectures in ppt
 Interactive class assignments (CA) and Homeworks (HW), some
of them will be collected and evaluated
 Open book final exam (FE) on 3 Nov 2016 from 11:00-13:30
 Final mark: 30% (CA+HW) + 70% FE
Methodology
Continuum Mechanics: On-line Course
 Video recording of lectures available at the Master website
 Lecture notes in pdf available at the Master website
 Class Assignments (CA) and Homeworks (HW). Solutions
provided for autoevaluation. Not to be collected or evaluated
 Open book final exam (FE) on 3 Nov 2016 from 11:00-13:30
 Final mark: 100% FE (only for students officially registered to
the on-line version of the Master on Numerical Methods)

Methodology
Continuum Mechanics: Final Exam
 Open book final exam
 The exam will take place on 3 Nov 2016 from 11:00-13:30
 Exam will consists of 2 problems + 1 question, similar to the
ones done during the course
 Maximum time allowed: 2.5 hours
 On-campus students: Room A2 102
 On-line students: Exams will be sent by email, being available
at the starting of the evaluation. At the end of the exam,
students must scan and send the exam by email. On-line
students being in Barcelona may do the exam on the campus

Introduction > Contents
Contents
Continuum Mechanics: Library

Contents
Continuum Mechanics: Library

Contents
Continuum Mechanics
 Chapter 1. Introduction to Vectors and Tensors
 Chapter 2. Motions
 Chapter 3. Strains
 Chapter 4. Infinitesimal Strains
 Chapter 5. Stresses
 Chapter 6. Balance Laws
 Chapter 7. Linear Elasticity
 Chapter 8. Plasticity
 Chapter 9. Fluid Mechanics
 Chapter 10. Ideal Fluids
 Chapter 11. Newtonian Fluids
 Chapter 12. Variational Principles
Contents
Continuum Mechanics
 Chapter 1. Introduction to Vectors and Tensors (3h 30m)
 Chapter 2. Motions (4h 30m)
 Chapter 3. Strains (7h)
 Chapter 4. Infinitesimal Strains (3h)
 Chapter 5. Stresses (1h)
 Chapter 6. Balance Laws (9h)
 Chapter 7. Linear Elasticity (6h)
 Chapter 8. Plasticity
 Chapter 9. Fluid Mechanics (30 m)
 Chapter 10. Ideal Fluids
 Chapter 11. Newtonian Fluids (4h)
 Chapter 12. Variational Principles (2h)
Introduction > References
References
Reference Books
 G.A. Holzapfel, Nonlinear Solid Mechanics: A Continuum
Approach for Engineering, Wiley, 2000
 O. Gonzalez, A.M. Stuart, A first course in Continuum
Mechanics, Cambridge Texts in Applied Mechanics, 2008
 J. Bonet, R. Wood, Nonlinear Continuum Mechanics for Finite
Element Analysis, Cambridge University Press, 1997
 J. Marsden, T.J.R. Hughes, Mathematical Foundations of
Elasticity, Courier Dover Publications, 1994
 C. Truesdell, W. Noll, The Nonlinear Field Theories of
Mechanics, Springer-Verlag, 1992
 J. Oliver, C. Agelet de Saracibar, Mecánica de Medios
Continuos para Ingenieros, Ediciones UPC, 2002
Continuum Mechanics
Chapter 1
Introduction to Vectors and Tensors
Introduction to Vectors and Tensors > Contents
Contents
Contents
1. Introduction
2. Algebra of vectors
3. Algebra of tensors
4. Higher order tensors
5. Differential operators
6. Integral theorems

Introduction to Vectors and Tensors > Introduction
Introduction
Tensors
Continuum mechanics deals with physical properties of
materials, either solids or fluids, which are independent of any
particular coordinate system in which they are observed.
Those physical properties are mathematically represented by

tensors.
Tensors are mathematical objects which have the required

property of being independent of any particular coordinate
system.

Introduction
Tensors
According to their tensorial order, tensors may be classified as:
 Zero-order tensors: scalars, i.e. density, temperature, pressure
 First-order tensors: vectors, i.e. velocity, acceleration, force
 Second-order tensors, i.e. stress, strain, strain rate
 Third-order tensors, i.e. piezoelectric tensor
 Fourth-order tensors, i.e. elastic constitutive tensor,
elastoplastic constitutive tensor
First- and higher-order tensors may be expressed in terms of

their components in a given coordinate system.

Introduction
Notation on Printed Documents
 Zero-order tensors: α , a, A
 First-order tensors: α , a, A
 Second-order tensors: α , a, A
 Third-order tensors: 
 Fourth-order tensors: 
Notation on Hand-written Documents

 Zero-order tensors: α , a, A
 First-order tensors: α , a , A
 Second-order tensors: α , a , A

Introduction to Vectors and Tensors > Algebra of Vectors
Algebra of Vectors
Scalar
A physical quantity, completely described by a single real
number, such as the temperature, density or pressure, is called a
scalar and is designated by α , a, A .
A scalar may be viewed as a zero-order tensor.
Vector
A vector is a directed line element in space. A model for physical
quantities having both direction and length, such as velocity,
acceleration or force, is called a vector and is designated by.
α , a, A.
A vector may be viewed as a first-order tensor.

Algebra of Vectors
Sum of Vectors
The sum of vectors yields a new vector, based on the parallelo-
gram law of addition.
The sum of vectors has the following properties,
u+ v = v+u
( u + v ) + w =u + ( v + w )
u+0=u
u + ( −u ) =0
where 0 denotes the unique zero vector with unspecified
direction and zero length.

Algebra of Vectors
Scalar Multiplication
Let u be a vector and α be a scalar. The scalar multiplication α u
gives a new vector with the same direction as u if α > 0 or with
the opposite direction to u if α < 0.
The scalar multiplication has the following properties,
(αβ ) u = α ( β u )
(α + β ) u =α u + β u
α (u + v ) = α u + α v

Algebra of Vectors
Dot Product
The dot (or scalar or inner) product of two vectors u and v,
denoted by u ⋅ v, is a scalar given by,
u⋅v u v cos θ ( u, v ) , 0 ≤ θ ( u, v ) ≤ π
where u is the norm (or length or magnitude) of the vector u,

which is a non-negative real number defined as,
u =(u ⋅ u ) ≥0
12
and θ ( u, v ) is the angle between the non-zero vectors u and v

when their origins coincide.

Algebra of Vectors
Dot Product
The dot (or scalar or inner) product of vectors has the following
properties,
u ⋅ v = v ⋅u
u⋅0 = 0
u ⋅ (α v + β w )= α ( u ⋅ v ) + β ( u ⋅ w )
u ⋅u > 0 ⇔ u ≠ 0
u ⋅ u= 0 ⇔ u= 0
u=⋅ v 0, u ≠ 0, v ≠ 0 ⇔ u ⊥ v

Algebra of Vectors
Unit Vectors
A vector e is called a unit vector if its norm (or lenght or
magnitude) is equal to 1,
e =1

Algebra of Vectors
Orthogonal Vectors
A non-zero vector u is said to be orthogonal (or perpendicular)
to a non-zero vector v if,
⋅ v 0, u ≠ 0, v ≠ 0 ⇔ θ ( u, =
u= v) π 2 ⇔ u ⊥ v
Orthonormal Vectors
A unit vector u is said to be orthonormal (or perpendicular) to a
unit vector v if,
u ⋅ v = 0, u = 1, v = 1 ⇔ θ ( u, v ) = π 2 ⇔ u ⊥ v

Algebra of Vectors
Projection
The projection of a vector u along a direction given by a unit
vector e is a (positive or negative) scalar quantity defined as,
=u ⋅e u cos θ ( u, e ) , 0 ≤ θ ( u, e ) ≤ π
u
θ ( u, e )
e
u cos θ ( u, e )

Algebra of Vectors
Cartesian Basis
Let us consider a three-dimensional Euclidean space with a fixed
set of three right-handed orthonormal basis vectors e1 , e 2 , e3 ,
denoted as Cartesian basis, satisfying the following properties,
x3 ,
e3
e1 ⋅ e 2 = e1 ⋅ e3 = e 2 ⋅ e3 = 0
e1
O e2
x2 , =
e1 e=
2 e=
3 1
x1 ,
A fixed set of three unit vectors which are mutually orthogonal

form a so called orthonormal basis vectors, collectively denoted
as {ei }, and define an orthonormal system.

Algebra of Vectors
Cartesian Components
Any vector u in the three-dimensional Euclidean space is
represented uniquely by a linear combination of the basis
vectors e1 , e 2 , e3 , i.e.
u = u1e1 + u2e 2 + u3e3
where the three real numbers u1 , u2 , u3 are the uniquely
determined Cartesian components of vector u along the given
directions e1 , e 2 , e3, respectively.
Using matrix notation, the Cartesian components of the vector u
can be collected into a vector of components [u ] ∈  3 given by,
[u ] = [u1 u3 ]
T
u2

Algebra of Vectors
Index Notation
Using index notation the vector u can be written as,
u = ∑ i =1 ui ei
3
Leaving out the summation symbol, using the so called Einstein

notation, the vector u can be written in short form as,
u = ui ei
where we have adopted the summation convention introduced

by Einstein.

Algebra of Vectors
Index Notation
 The index that is summed over is said to be a dummy (or
summation) index.
 The same repeated index can appear only twice.
 An index that is not summed over in a given term is called a
free (or live) index.
vi = ai b j c j = ai ( b1c1 + b2 c2 + b3c3 )
v = ai bi c j d j = ( a1b1 + a2b2 + a3b3 )( c1d1 + c2 d 2 + c3d3 )
v=
i ai b j c j ck d k= ai ( b1c1 + b2 c2 + b3c3 )( c1d1 + c2 d 2 + c3d3 )

Algebra of Vectors
Kronecker Delta
The Kronecker delta δ ij is defined as,
1 if i = j
δ ij = 
0 if i ≠ j
The Kronecker delta satisfies the following properties,
=δ ii 3,=
δ ijδ jk δ ik=
, δ ij u j ui
Note that the Kronecker delta plays the role of a replacement

operator.

Algebra of Vectors
Dot Product
The dot (or scalar or inner) product of two orthonormal basis
vectors ei and e j , denoted as ei ⋅ e j , is a scalar quantity (taking
values either 0 or 1) and can be conveniently written in terms of
the Kronecker delta as,
δ ij
ei ⋅ e j =

Algebra of Vectors
Components of a Vector
Taking the basis {ei }, the projection of a vector u onto the basis
vector ei yields the i-th component of u ,
u ⋅ ei= (u e ) ⋅ e =
j j i u j e j ⋅ ei= u jδ ji= ui

Algebra of Vectors
Dot Product
Taking the basis {ei }, the dot (or scalar or inner) product of two
vectors u and v, denoted as u ⋅ v , is a scalar quantity and using
index notation can be written as,
u ⋅ v= ui ei ⋅ v j e j = ui v j ei ⋅ e j = ui v jδ ij = ui vi
= u1v1 + u2 v2 + u3v3

Algebra of Vectors
Euclidean Norm
Taking the basis {ei }, the Euclidean norm (or length or
magnitude) of a vector u, denoted as u , is a non-negative
scalar quantity and, using index notation, can be written as,
(u ⋅ u ) ( uiei ⋅ u je j ) ( uiu jδ ij ) ( uiui )

12 12
u = = = =
12 12
= (u
2
1 +u +u
2
2 3 )
2 12

Assignment 1.1
Assignment 1.1 [Classwork]
Write in index form the following expressions,
(a ⋅ b ) c
v=
v a (b ⋅ c)
=
v= ( a ⋅ b )( c ⋅ d )
v = a ( b ⋅ c )( c ⋅ d )

Algebra of Vectors
Cross Product
The cross (or vector) product of two vectors u and v , denoted
as u × v, is another vector which is perpendicular to the plane
defined by the two vectors and its norm (or length) is given by
the area of the parallelogram spanned by the two vectors,
u× v u v sin θ ( u, v ) e, 0 ≤ θ ( u, v ) ≤ π
where e is a unit vector perpendicular to the plane defined by
the two vectors, u ⋅ e= 0, v ⋅ e= 0, in the direction given by the
right-hand rule, .

Algebra of Vectors
Norm of the Cross Product of two Vectors
The norm (or magnitude or length) of the cross (or vector)
product of two vectors u and v, denoted as u × v , measures
the area spanned by the two vectors and is given by,
=u× v u v sin θ ( u, v ) , 0 ≤ θ ( u, v ) ≤ π
u× v
v
u× v
u
Algebra of Vectors
Permutation Symbol
The permutation (or alternating or Levi-Civita) symbol ε ijk is
defined as,
+1 for even permutations of (i, j , k ), i.e. 123, 231, 312

ε ijk =  −1 for odd permutations of (i, j , k ), i.e. 213, 132, 321
 0 if there is repeated index


Algebra of Vectors
Permutation Symbol
The permutation (or alternating or Levi-Civita) symbol ε ijk may
be written in terms of the Kronecker delta and has the following
properties,
 δ i1 δ i 2 δ i 3   δ ip δ iq δ ir 
δ , ε ε  
ε ijk =
det  j1 δ j2 δ j3  ijk pqr det δ jp δ jq δ jr 
δ k1 δ k 2 δ k 3  δ kp δ kq δ kr 

ε ijpε klp =
δ ik δ jl − δ ilδ jk , ε ikl ε jkl =
2δ ij , ε ijk ε ijk =
2δ ii =
6
ε ijk =
( i − j )( j − k )( k − i )
2
Algebra of Vectors
Cross Product
The cross (or vector) product of two right-handed orthonormal
basis vectors satisfies the following properties,
x3 ,
e1 ×=
e 2 e3 , e 2 ×=
e3 e1 , e3 ×=
e1 e 2 ,
e 2 × e1 =−e3 , e3 × e 2 =−e1 , e1 × e3 =−e 2 ,
e3
e2
O
x2 ,
e1 × e1 = e 2 × e 2 = e3 × e3 = 0.
e1
x1 ,
Taking the right-handed orthonormal basis {ei }, the cross (or

vector) product of two basis vectors ei and e j , denoted as ei × e j ,
can defined as,
ε ijk e k
ei × e j =

Algebra of Vectors
Cross Product
Taking the basis {ei } , the cross (or vector) product of two
vectors u and v, denoted as u × v , is another vector,
perpendicular to the plane defined by the two vectors, and can
be written in index form as,
u × v= ui ei × v j e j = ui v j ei × e j = ui v j ε ijk e k = ε ijk ui v j e k
=( u2v3 − u3v2 ) e1 + ( u3v1 − u1v3 ) e2 + ( u1v2 − u2v1 ) e3
e1 e 2 e3 
= det u1 u2 u3 
 v1 v2 v3 

Algebra of Vectors
Cross Product
The cross (or vector) product of vectors has the following
properties,
u × v =− ( v × u )
u ×=
v 0, u ≠ 0, v ≠ 0 ⇔ u || v
(α u ) × v = u × (α v ) = α ( u × v )
u×(v + w) = u× v + u× w

Assignment 1.2
Assignment 1.2
Using the expression,
u× v v sin 2 θ ( u, v ) , 0 ≤ θ ( u, v ) ≤ π
2 2 2
u
obtain the following relationships between the permutation
symbol and the Kronecker delta,
ε ijpε klp =
δ ik δ jl − δ ilδ jk , ε ipqε jpq =
2δ ij , ε ijk ε ijk =
2δ ii =
6

Algebra of Vectors
Box Product
The box (or triple scalar) product of three right-handed
orthonormal basis vectors ei , e j , e k , denoted as ei ⋅ ( e j × e k ) , is a
scalar quantity (taking values either 1, -1 or 0) and can be
conveniently written in terms of the permutation symbol as,
ε ijk =ei ⋅ ( e j × e k ) =( ei × e j ) ⋅ e k

Algebra of Vectors
Box Product
Taking the orthonormal basis {ei }, the box (or triple scalar)
product of three vectors u, v, w , denoted as u ⋅ ( v × w ), is a
scalar quantity and using index notation can be written as,
u ⋅ ( v × w ) = ui ei ⋅ ( v j e j × wk e k ) = ui v j wk ei ⋅ ( e j × e k )
= ε ijk ui v j wk
 u1 u2 u3 
= det  v1 v2 v3 
 w1 w2 w3 

Algebra of Vectors
Box Product
The box (or triple scalar) product of three vectors u, v, w
represents the volume of the parallelepiped spanned by the
three vectors,
V =u ⋅ ( v × w )
v×w
u
V =u ⋅ ( v × w )
w
v
Algebra of Vectors
Box Product
The box (or triple scalar) product has the following properties,
u ⋅ ( v × w ) =w ⋅ ( u × v ) =v ⋅ ( w × u )
u ⋅ ( v × w ) =−u ⋅ ( w × v )
u ⋅ ( u × v ) =v ⋅ ( u × v ) =0

Algebra of Vectors
Triple Vector Product
Taking the orthonormal basis {ei } , the triple vector product of
three vectors u, v, w , denoted as u × ( v × w ) , is a vector and
using index notation can be written as,
ε ijk ui ( ε pqj v p wq ) e k
u×(v × w) =
= ε kij ε pqj ui v p wq e k
= (δ δ − δ kqδ ip ) ui v p wq e k
kp iq
= ui vk wi e k − ui vi wk e k
= (u ⋅ w ) v − (u ⋅ v ) w

Algebra of Vectors
Taking the orthonormal basis {ei } , the triple vector product of
three vectors u, v, w , denoted as ( u × v ) × w , is a vector and
using index notation can be written as,
ε ijk ( ε pqi u p vq ) w j e k
(u × v ) × w =
= ε jkiε pqi u p vq w j e k
= (δ δ − δ jqδ kp ) u p vq w j e k
jp kq
= u j w j vk e k − v j w j uk e k
= (u ⋅ w ) v − ( v ⋅ w ) u

Algebra of Vectors
The triple vector product has the following properties,
u × ( v × w ) = (u ⋅ w ) v − (u ⋅ v ) w
(u × v ) × w = (u ⋅ w ) v − ( v ⋅ w ) u
u × ( v × w ) ≠ (u × v ) × w

Introduction to Vectors and Tensors > Algebra of Tensors
Algebra of Tensors
Second-order Tensors
A second-order tensor A may be thought of as a linear operator
that acts on a vector u generating a vector v, defining a linear
transformation that assigns a vector v to a vector u,
v = Au
The second-order unit tensor, denoted as 1 , satisfies the

following identity,
u = 1u

Algebra of Tensors
As linear operators defining linear transformations, second-
order tensors have the following properties,
A (α u + v )= α Au + Av
(α A ) u = α ( Au )
( A ± B ) u =Au ± Bu

Algebra of Tensors
Tensor Product
The tensor product (or dyad) of two vectors u and v , denoted
as u ⊗ v , is a second-order tensor which linearly transforms a
vector w into a vector with the direction of u following the rule,
( u ⊗ v ) w =u ( v ⋅ w ) =( v ⋅ w ) u
A dyadic is a linear combination of dyads.

Algebra of Tensors
Tensor Product
The tensor product (or dyad) has the following linear properties,
( u ⊗ v )(α w + x )= α ( u ⊗ v ) w + ( u ⊗ v ) x
(α u + β v ) ⊗ w = α u ⊗ w + β v ⊗ w
(u ⊗ v ) w = (v ⋅ w)u = u(v ⋅ w)
( u ⊗ v )( w ⊗ x ) = ( v ⋅ w ) u ⊗ x = u ⊗ x ( v ⋅ w )
A (u ⊗ v= ) ( Au ) ⊗ v
Generally, the tensor product (or dyad) is not commutative, i.e.
u ⊗ v ≠ v ⊗u
( u ⊗ v )( w ⊗ x ) ≠ ( w ⊗ x )( u ⊗ v )
Algebra of Tensors
Tensor Product
The tensor product (or dyad) of two orthonormal basis vectors ei
and e j , denoted as ei ⊗ e j , is a second-order tensor and can be
thought as a linear operator such that,
( e ⊗ e ) e =( e
i j k j ⋅ e k ) ei =δ jk ei

Algebra of Tensors
Any second-order tensor A may be represented by a linear
combination of dyads formed by the Cartesian basis {ei }
=
A Aij ei ⊗ e j
where the nine real numbers, represented by Aij , are the
uniquely determined Cartesian components of the tensor A
with respect to the dyads formed by the Cartesian basis {ei },
represented by ei ⊗ e j , which constitute a basis second-order
tensor for A .
The second-order unit tensor may be written as,
1 = δ ij ei ⊗ e j = ei ⊗ ei

Algebra of Tensors
Cartesian Components
The ij-th Cartesian component of the second-order tensor A,
denoted as Aij , can be written as,
ei ⋅ Ae j =ei ⋅ ( Akl e k ⊗ el ) e j =ei ⋅ ( Aklδ lj e k )

= ei ⋅ ( Akj e k ) = Akj ei ⋅ e k = Akjδ ik = Aij
The ij-th Cartesian component of the second-order unit tensor 1

is the Kronecker delta,
ei ⋅ 1e j = ei ⋅ e j = δ ij

Algebra of Tensors
Matrix of Components
Using matrix notation, the Cartesian components of any second-
order tensor A may be collected into a 3 x 3 matrix of compo-
nents, denoted by [ A ] , given by,
=
A Aij ei ⊗ e j
 A11 A12 A13 

A A23 
[ ]  21
A = A22
 A31 A32 A33 

Algebra of Tensors
Using matrix notation, the Cartesian components of the second-
order unit tensor 1 may be collected into a 3 x 3 matrix of
components, denoted as [1], given by,
1 0 0 
[1] = 0 1 0
0 0 1 

Algebra of Tensors
The tensor product (or dyad) of the vectors u and v, denoted as
u ⊗ v , may be represented by a linear combination of dyads
formed by the Cartesian basis {ei }
u ⊗=
v ui v j ei ⊗ e j
where the nine real numbers, represented by ui v j , are the

uniquely determined Cartesian components of the tensor u ⊗ v
with respect to the dyads formed by the Cartesian basis {ei },
represented by ei ⊗ e j .

Algebra of Tensors
Using matrix notation, the Cartesian components of any second-
order tensor u ⊗ v may be collected into a 3 x 3 matrix of
components, denoted as [u ⊗ v ], given by,
u ⊗=
v ui v j ei ⊗ e j
 u1v1 u1v2 u1v3 

u v u v u2 v3 
[u ⊗ v ] =
 2 1 2 2
u3v1 u3v2 u3v3 

Algebra of Tensors
Using index notation, the linear transformation v = Au can be
written as,
v = Au
vi ei = ( Aij ei ⊗ e j ) ( uk e k ) = Aij uk ( ei ⊗ e j ) e k = Aij uk δ jk ei
vi ei = Aij u j ei
yielding
= =
v Au , vi Aij u=
j, [ v] [ A ][u ]

Algebra of Tensors
Positive Semi-definite Second-order Tensor
A second-order tensor A is said to be a positive semi-definite
tensor if the following relation holds for any non-zero vector u ≠ 0
u ⋅ Au ≥ 0 ∀u ≠ 0
Positive Definite Second-order Tensor

A second-order tensor A is said to be a positive-definite tensor
if the following relation holds for any non-zero vector u ≠ 0
u ⋅ Au > 0 ∀u ≠ 0

Algebra of Tensors
Transpose of a Second-order Tensor
The unique transpose of a second-order tensor A , denoted as AT ,
is governed by the identity,
v ⋅ AT u =
u ⋅ Av ∀u, v
v= T
i ij u j
A v=
i A ji u j u j Aji vi ∀u j , vi
The Cartesian components of the transpose of a second-order
tensor A satisfy,
AijT :=( ) i e j =e j ⋅ Aei =Aji
A T
=e ⋅ A
ij
T
Using index notation the transpose of A may be written as,

=
AT Aji ei ⊗ e j
Algebra of Tensors
Transpose of a Second-order Tensor
The matrix of components of the transpose of a second-order
tensor A is equal to the transpose of the matrix of components
of A and takes the form,
=
A Aij ei ⊗ e j
=
AT Aji ei ⊗ e j
 A11 A21 A31 

 A
= A ]  A12
 [ = A32 
T T
A22
 A13 A23 A33 

Algebra of Tensors
Symmetric Second-order Tensor
A second-order tensor A is said to be symmetric if the following
relation holds,
A = AT
Using index notation, the Cartesian components of a symmetric
second-order tensor A satisfy,
A= Aij ei ⊗ e j = Aji ei ⊗ e j = AT , Aij = Aji
Using matrix notation, the matrix of components of a symmetric
second-order tensor A satisfies,
[ A ] =
= A  [ A]
T T

Algebra of Tensors
Dot Product
The dot product of two second-order tensors A and B, denoted
as AB , is a second-order tensor such that,
( AB ) u A ( Bu ) ∀u
=

Algebra of Tensors
Dot Product
The components of the dot product AB along an orthonormal
basis {ei } read,
( AB )ij =
Cij = ei ⋅ A ( Be j )
ei ⋅ ( AB ) e j =
ei ⋅ A ( Bkj e k ) =
= ei ⋅ Ae k Bkj
= Aik Bkj
= =
C AB , Cij Aik Bkj

Algebra of Tensors
Dot Product
The dot product of second-order tensors has the following
properties,
(=
AB ) C ( BC ) ABC
A=
A 2 = AA
AB ≠ BA

Algebra of Tensors
Trace
The trace of a dyad u ⊗ v, denoted as tr ( u ⊗ v ) , is a scalar
quantity given by,
tr ( u ⊗ v ) = u ⋅ v = ui vi
The trace of a second-order tensor A , denoted as tr A , is a scalar
quantity given by,
tr = j)
A tr ( Aij ei ⊗ e= Aij tr ( ei ⊗ e j )
= Aij ( ei ⋅ e j )= Aijδ ij
= Aii
= A11 + A22 + A33 = tr [ A ]
Algebra of Tensors
Trace
The trace of a second-order tensor has the following properties,
tr A = tr AT
tr ( AB ) = tr ( BA )
tr ( A + B ) = tr A + tr B
tr (α A ) = α tr ( A )

Algebra of Tensors
Double Dot Product
The double dot product of two second-order tensors A and B ,
denoted as A : B , is a scalar quantity defined as,
= ( ) ( A)
A : B tr=A T
B tr B T
( ) ( )
= tr=AB T
tr BA T
= B:A
or using index notation,
A=
: B A=
ij Bij B=
ij Aij B:A

Algebra of Tensors
Double Dot Product
The double dot product has the following properties,
A=
: 1 1=
: A tr A
A:B = B:A
A : ( BC )
= ( A) : C
=
B T
( ):B
ACT
A : ( u ⊗ v ) =u ⋅ Av =( u ⊗ v ) : A
( u ⊗ v ) : ( w ⊗ x ) = ( u ⋅ w )( v ⋅ x )
( ei ⊗ e j ) : ( ek ⊗ el ) = ( ei ⋅ ek ) ( e j ⋅ el ) = δ ikδ jl

Algebra of Tensors
Euclidean Norm
The euclidean norm of a second-order tensor A is a non-
negative scalar quantity, denoted as A , given by,
(A : A) ( Aij Aij )
12
= = ≥0
12
A

Algebra of Tensors
Determinant
The determinant of a second-order tensor A is a scalar quantity,
denoted as det A , given by,
det A = det [ A ]
 A11 A12 A13 
= det  A21 A22 A23 
 A31 A32 A33 
1
= ε=
ijk A1i A2 j A3 k ε ijk ε pqr Api Aqj Ark
6

Algebra of Tensors
Determinant
The determinant of a second-order tensor has the following
properties,
det AT = det A
det (α A ) = α 3 det A
det ( AB ) = det A det B

Algebra of Tensors
Singular Tensors
A second-order tensor A is said to be singular if and only if its
determinant is equal to zero, i.e.
A is singular ⇔ det A =
0

Algebra of Tensors
Inverse of a Second-order Tensor
For a non-singular second-order tensor A , i.e. det A ≠ 0, there
exist a unique non-singular inverse second-order tensor,
denoted as A −1 , satisfying,
( )=u
AA −1
( A )=u 1u= u ∀u
A −1
yielding,
=
AA −1 −1
A= A 1, Akj−1 Aik=
Aik= −1
Akj δ ij

Algebra of Tensors
Inverse of a Second-order Tensor
The inverse second-order tensor satisfies the following
properties,
(A ) = A
−1 −1
(=
A ) (=
A )
−1 −1 T
T
A −T
(α A ) = α −1A −1
−1
( )
−1 −1 −1
AB = B A
A −2 = A −1A −1
det A −1 = ( det A )
−1

Algebra of Tensors
Orthogonal Second Order Tensor
An orthogonal second-order tensor Q is a linear operator that
preserves the norms and the angles between two vectors,
( u ⋅ Q Qu ) (u ⋅ u )
12
Qu = = =u ∀u
T 12
Qu ⋅ Qv =u ⋅ QT Qv =u ⋅ v 
 ∀u, v
Qv ⋅ Qu =v ⋅ Q Qu =v ⋅ u 
T
and the following relations hold,
QT Q = QQT = 1, Q −1 = QT , det Q = ±1

Algebra of Tensors
Rotation Second Order Tensor
A rotation tensor R is a proper orthogonal second-order tensor,
i.e., is a linear operator that preserves the lengths (norms) and
the angles between two vectors,
( u ⋅ R Ru ) (u ⋅ u )
12
Ru = = =u ∀u
T 12
Ru ⋅ Rv =u ⋅ R T Rv =u ⋅ v 
 ∀u, v
Rv ⋅ Ru =v ⋅ R Ru =v ⋅ u 
T
and satisfies the following expressions,
R=
T
=
R RR T
R −1 R T , det
1, = = R 1

Algebra of Tensors
Symmetric/Skew-symmetric Additive Split
A second-order tensor A can be uniquely additively split into a
symmetric second-order tensor S and a skew-symmetric
second-order tensor W, such that,
A=
S + W, Aij =
Sij + Wij
S=
1
2
( A+A )= T
S ,T
Sij =
1
2
( A
ij + Aji ) = S ji
W=(A − A 1
2
T
)=
−W T
, Wij =( Aij − Aji ) =
1
2
−W ji

Algebra of Tensors
Symmetric/Skew-symmetric Second-order Tensors
The matrices of components of a symmetric second-order
tensor S and a skew-symmetric second-order tensor W, are
given by,
 S11 S12 S13   0 W12 W13 

   
[S ] =  S12 S 22 S 23  , [ W ] =  −W12 0 W23 
 S13 S 23 S33   −W13 −W23 0 

Algebra of Tensors
Skew-symmetric Second-order Tensors
The double dot product of a symmetric second-order tensor S
and a skew-symmetric second-order tensor W gives,
S : W = SijWij
= S12W12 + S13W13 + S 23W23 − S12W12 − S13W13 − S 23W23
=0

Algebra of Tensors
Skew-symmetric Second-order Tensor
A skew-symmetric second-order tensor W can be defined by
means of an axial (or dual) vector ω , such that,
Wu =ω × u ∀u
W = 2ω
and using index notation the following relations hold,
W=
Wij ei ⊗ e j =−ε ijk ωk ei ⊗ e j , Wij =−ε ijk ωk
1 1
ω=
ωk e k =
− ε ijkWij e k , ωk =
− ε ijkWij
2 2

Algebra of Tensors
Skew-symmetric Second-order Tensor
Using matrix notation and index notation the following relations
hold,
 0 W12 W13   0 −ω3 ω2 

 −W =ω 
[ ]  12
W = 0 W23   3 0 −ω1
 −W13 −W23 0   −ω2 ω1 0 

[ω ] =
[ω1 , ω2 , ω3 ] [ −W23 , W13 , −W12 ]
=
T T
=
W 2 (W122 + W132 + W=
23 )
2
2 ω12 + ω32 + ω
=3
2
2ω

Algebra of Tensors
Spherical Second-order Tensor
A spherical second-order tensor A is defined as,
=A α=
1, Aij αδ ij
and has the following properties,
A α=
tr= tr 1 3α , =
Aii αδ
= ii 3α
α 0 0 
0 α 0
[ ] 
A = 
 0 0 α 

Algebra of Tensors
Deviatoric Second-order Tensor
A second-order tensor A is said to be deviatoric if the following
condition holds,
=
tr A 0,=Aii 0

Algebra of Tensors
Spherical/Deviatoric Additive Split
A second-order tensor A can be uniquely additively split into a
spherical second-order tensor A esf and a deviatoric second-
order tensor dev A, such that,
A=
A esf + dev A, Aijesf + ( dev A )ij
Aij =
( tr A ) 1,
1 1
A esf
= A esf
ij Akk δ ij
3 3
A − ( tr A ) 1, ( dev A )ij =
1 1
dev A = Aij − Akk δ ij
3 3

Algebra of Tensors
Eigenvalues/Eigenvectors
Given a symmetric second-order tensor A , its eigenvalues
λi ∈  and associated eigenvectors ni, which define an
orthonormal basis along the principal directions, satisfy the
following equation (Einstein notation does not applies here),
An=
i λi ni , ( A − λi 1) n=i 0
In order to get a non-trivial solution, the tensor A − λi 1 has to
be singular, i.e., its determinant, defining the characteristic
polynomial, has to be equal to zero, yielding,
p ( λi )= det ( A − λi 1)= 0

Algebra of Tensors
Given a symmetric second-order tensor A , its eigenvalues
λi ∈  and associated eigenvectors ni, which define an
orthonormal basis along the principal directions, characterize
the physical nature of the tensor.
The eigenvalues do not depend on the particular orthonormal

basis chosen to define the components of the tensor and,
therefore, the solution of the characteristic polynomial does not
depend on the system of reference in which the components of
the tensor are given.

Algebra of Tensors
Characteristic Polynomial and Principal Invariants
The characteristic polynomial of a second-order tensor A takes
the form,
p (λ ) =det ( A − λ1) =−λ 3 + I1λ 2 − I 2 λ + I 3 =0
where the coefficients of the polynomial are the principal scalar
invariants of the tensor (invariants in front of an arbitrary
rotation of the orthonormal basis) given by,
I1 ( A ) = tr A = λ1 + λ2 + λ3
I2 ( A ) =
1
2
( tr 2
A − tr A 2
) = λ1λ2 + λ2λ3 + λ3λ1
( A ) det
I 3= = A λ1λ2 λ3
Algebra of Tensors
Given a deviatoric part of a symmetric second-order tensor
dev A , its eigenvalues λi′ ∈  and asociated eigenvectors n′i ,
which define an orthonormal basis along the principal
directions, satisfy the following equation (Einstein notation does
not applies here),
An′i λi′n′i ,
dev= ( dev A − λ=
i 1) n i
′ ′ 0
In order to get a non-trivial solution, the tensor dev A − λi′1 has
to be singular, i.e., its determinant, defining the characteristic
polynomial, has to be equal to zero, yielding,
( λi′) det ( dev A − =
p= λi′1) 0

Algebra of Tensors
Characteristic Polynomial and Principal Invariants
The characteristic polynomial of the deviatoric part of a second-
order tensor dev A takes the form,
p (λ′ ) =det ( dev A − λ ′1) =−λ ′3 + I1′λ ′2 − I 2′λ ′ + I 3′ =0
where the coefficients of the polynomial are the pincipal scalar
invariants of the deviatoric part of the tensor (invariants in front
of an arbitrary rotation of the orthonormal basis) given by,
I1′ (=
dev A ) tr=
dev A 0
− tr ( dev A ) =
I 2′ ( dev A ) = − ( λ1′λ1′ + λ2′λ2′ + λ3′λ3′ )
1 2 1
2 2
( dev A ) det
I 3= = dev A λ1′λ2′λ3′
Algebra of Tensors
The eigenvalues/eigenvectors problem for a symmetric second-
order tensor A and its the deviatoric part dev A satisfy the
following relations (Einstein notation does not applies here),
An=
i λi ni , ( A − λi 1) n=i 0
An′i λi′n′i ,
dev= ( dev A − λ= i 1) n i
′ ′ 0
Then,
   
( dev A − λi′1) n′i =
 A −  λi′ + 3 ( tr A )  1  n′i =
( A − λi 1) ni =
1
0
   
yielding,
λi′ + ( tr A ) , ni =
1
λi = n′i
3
Algebra of Tensors
Spectral Decomposition
The spectral decomposition of the secod-order unit tensor takes
the form,
For a second-order unit tensor, all three eigenvalues are equal to
one, λ=1 λ=
2 λ=3 1, and any direction is a principal direction.
Then we may take any orthonormal basis as principal directions
and the spectral decomposition can be written,
1 = ∑ i=1,3 ni ⊗ ni = ni ⊗ ni

Algebra of Tensors
Spectral Decomposition
The spectral decomposition of a symmetric second-order tensor A
takes the form (Einstein notation does not applies here),
=A ∑ i =1,3
λi ni ⊗ ni
If there are two equal eigenvalues λ= λ=
1 λ2 , the spectral
decomposition takes the form,
A = λ (1 − n3 ⊗ n3 ) + λ3n3 ⊗ n3
If the three eigenvalues are equal λ= λ=
1 λ=
2 λ3 , the spectral
decomposition takes the form,
A = λ1 = λ ∑ i =1,3 ni ⊗ ni = λni ⊗ ni
Introduction to Vectors and Tensors > Higher-order Tensors
Third-order Tensors
Third-order Tensors
A third-order tensor, denoted as  , may be written as a linear
combination of tensor products of three orthonormal basis
vectors, denoted as ei ⊗ e j ⊗ e k , such that,
∑ ∑ ∑
3 3 3
= =i 1 =j 1 =k 1
ijk ei ⊗ e j ⊗ e=
k ijk ei ⊗ e j ⊗ e k
A third-order tensor has 33 = 27 components.

Fourth-order Tensors
A fourth-order tensor, denoted as  , may be written as a linear
combination of tensor products of four orthonormal basis
vectors, denoted as ei ⊗ e j ⊗ e k ⊗ el , such that,
 ∑ ∑ ∑ ∑  ijkl ei ⊗ e j ⊗ e k ⊗ e=  ijkl ei ⊗ e j ⊗ e k ⊗ el
3 3 3 3
= =i 1 =j 1 =k 1 =l 1 l
A fourth-order tensor has 34 = 81 components.

Double Dot Product with Second-order Tensors
The double dot product of a fourth-order tensor  with a
second-order tensor A is another second-order tensor B given
by,
B=  : A=  ijkl Akl ei ⊗ e j = Bij ei ⊗ e j , Bij =  ijkl Akl

Transpose of a Fourth-order Tensor
The transpose of a fourth-order tensor  is uniquely defined as
a fourth-order tensor T such that for arbitrary second-order
tensors A and B the following relationship holds,
A, Aij
A :  : B B : T := B
ijkl kl =
B 
T
A
kl klij ij , 
ijkl 
T
klij
The transpose of the transpose of a fourth-order tensor  is

the same fourth-order tensor  ,
( ) T T
=

Tensor Product of two Second-order Tensors
The tensor product of two second-order tensors A and B is a
fourth order tensor  given by,
 = A ⊗ B = Aij Bkl ei ⊗ e j ⊗ e k ⊗ el =  ijkl ei ⊗ e j ⊗ e k ⊗ el ,
 ijkl =( A ⊗ B )ijkl =Aij Bkl
The transpose of the fourth-order tensor = A ⊗ B is given by,
 = ( A ⊗ B ) = B ⊗ A = Bij Akl ei ⊗ e j ⊗ e k ⊗ el
T T
=  ijkl
T
ei ⊗ e j ⊗ e k ⊗ el ,
 =( A ⊗ B )ijkl =( B ⊗ A )ijkl =Bij Akl
T T
ijkl

Fourth-order Identity Tensors
Let us consider the following fourth-order identity tensors,
= δ ik δ jl ei ⊗ e j ⊗ e k ⊗ el
 δ ilδ jk ei ⊗ e j ⊗ e k ⊗ el
=
ˆ= 1 (  +  )= 1 (δ δ + δ δ ) e ⊗ e ⊗ e ⊗ e
ik jl il jk i j k l
2 2

The fourth-order identity tensor  satisfies the following
expressions,
 : A= δ ik δ jl Akl ei ⊗ e j = Aij ei ⊗ e j = A,
 Aijδ ik δ jl e k ⊗ e=
A := l Akl e k ⊗ e=
l A
The fourth-order identity tensor  is symmetric, satisfying,

 : B A=
A:= : B B=: A B :  : A
 ⇒  
= T
A::B = B: :A
T


The fourth-order identity tensor  satisfies the following
expressions,
 : A= δ ilδ jk Akl ei ⊗ e j= Aji ei ⊗ e j= AT

A : = Aijδ ilδ jk Akl e k ⊗ e=
l A e
lk k ⊗ e=l A T
The fourth-order identity tensor  is symmetric, satisfying,

 : B A=
A := : BT B= : AT B :  : A 
 ⇒ =
T
A:  :B = B:  :A
T


The fourth-order identity tensor ̂ satisfies the following
expressions,
ˆ : A =1 (  +  ) : A =1 ( A + AT )
2 2
A : =
ˆ 1
2
A : (  +  )=
1
2
( A + A T
)
The fourth-order identity tensor ̂ is symmetric, satisfying,

B A : (B + B= ) ( )
1 1
A :  :=
ˆ T
B : A + A =
T
B : 
ˆ : A 
2 2  ⇒ 
ˆ 
=
ˆ T
A : ˆ : B = B : ˆ T : A 

Fourth-order Deviatoric Projection Operator Tensor
The fourth-order deviatoric projection operator tensor dev is
defined as,
dev =−
1
 1 ⊗ 1,
3
( ) dev ijkl
1
=δ ik δ jl − δ ijδ kl
3
The deviatoric part of a second-order tensor A can be obtained
as,
 1 
− 1 ⊗ 1  : A = A − ( tr A ) 1,
1
dev A = dev : A = 
 3  3
 
( dev A )ij ( dev )ijkl Akl =
=
1
Aij − ( Akk ) δ ij
 δ ik δ jl − δ ijδ kl  Akl =
 3 
1
3
Algebra of Tensors
Algebra of Tensors
a = u ⋅ v = ui vi , a = ui vi
w = u × v = ε ijk u j vk ei , wi = ε ijk u j vk
a =u ⋅ ( v × w ) =ε ijk ui v j wk , a =ε ijk ui v j wk
A = u ⊗ v = ui v j ei ⊗ e j , Aij = ui v j
= = Aij u j ei , =
v Au vi Aij u j
C=
AB =
Aik Bkj ei ⊗ e j , Cij =
Aik Bkj
=a A=
: B Aij Bij ,=a Aij Bij
 ijkl Akl ei ⊗ e j , Bij =
:A =
B=  ijkl Akl
 = A ⊗ B = Aij Bkl ei ⊗ e j ⊗ e k ⊗ el ,  ijkl = Aij Bkl
Introduction to Vectors and Tensors > Differential Operators
Differential Operators
Nabla
The nabla vector differential operator, denoted as ∇ , is defined
as,
∂
∇ = ei
∂xi
Laplacian
The laplacian scalar differential operator, denoted as ∆ , is
defined as,
∂2
∆= 2
∂xi

Hessian
The hessian symmetric second-order tensor differential
operator, denoted as ∇ ⊗ ∇, is defined as,
∂2
=
∇ ⊗∇ ei ⊗ e j
∂xi ∂x j

Divergence, Curl and Gradient
The divergence differential operator div (·) is defined as,
∂ ()
div (  ) = ∇ ⋅ (  ) = ⋅ ei
∂xi
The curl differential operator curl (·) is defined as,
∂ ()
curl (  ) = ∇ × (  ) = ei ×
∂xi
The gradient differential operator grad (·) is defined as,
∂ () ∂ ()
grad (  ) = ∇ ⊗ (  ) = ⊗ ei = ∇ (  ) = ei
∂xi ∂xi
Laplacian
The laplacian scalar differential operator can be expressed as the
div (grad (·)) operator,
∂2
∆ (  ) = ∇ ⋅∇ (  ) = 2 (  ) = div grad (  )
∂xi
Hessian
The hessian symmetric second order-tensor differential operator
can be expressed as the grad (grad (·)) operator,
∂2 ()
⊗ ∇ ()
∇= ei ⊗ e j grad grad (  )
=
∂xi ∂x j
Gradient of a Scalar Field
∂ () ∂ ()
grad (  ) = ∇ ⊗ (  ) = ⊗ ei = ∇ (  ) = ei
∂xi ∂xi
The gradient of a scalar field is a vector field defined as,
∂φ
grad φ =∇φ = ei =φ,i ei
∂xi

Laplacian of a Scalar Field
The laplacian differential operator ∆ (  ) is defined as,
∂2 ()
∆ (  ) = div grad (  ) = ∇ ⋅∇ (  ) =
∂xi2
The laplacian of a scalar field is a scalar field defined as,
∂ 2φ
∆φ = div grad φ = ∇ ⋅∇φ = 2 = φ,ii
∂xi

Hessian of a Scalar Field
The hessian differential operator ∇ ⊗ ∇ (  ) is defined as,
∂2 ()
⊗ ∇ ()
∇= ei ⊗ e j
∂xi ∂x j
The hessian of a scalar field is a symmetric second-order tensor
field defined as,
∂ 2φ
∇ ⊗∇
= φ e j φ,ij ei ⊗ e j
ei ⊗=
∂xi ∂x j

Gradient of a Vector Field
∂ () ∂ ()
grad (  ) = ∇ ⊗ (  ) = ⊗ ei = ∇ (  ) = ei
∂xi ∂xi
The gradient of a vector field is a second-order tensor field
defined as,
∂u ∂ui
grad u =
∇ ⊗ u = ⊗ e j = ei ⊗ e j =
ui , j ei ⊗ e j
∂x j ∂x j

Curl of a Vector Field
∂ ()
curl (  ) = ∇ × (  ) = ei ×
∂xi
The curl of a vector field is a vector field defined as,
∂u ∂uk ∂uk
curl u= ∇ × u= e j × = e j × e k= ε ijk e= ε ijk uk , j ei
∂x j ∂x j ∂x j
i
If the curl of a vector field is zero, the vector field is said to be

curl-free and there is a scalar field such that,
curl u = 0 ⇒ ∃φ | u = grad φ
Curl of the Gradient of a Scalar Field
The curl differential operator curl (·) and gradient differential
operator grad (·) are defined, respectively, as,
∂ () ∂ ()
curl (  ) = ∇ × (  ) = ei × , grad (  ) = ∇ ⊗ (  ) = ⊗ ei
∂xi ∂xi
The gradient of a scalar field is a vector field defined as,
∂φ
grad φ =∇φ = ei =φ,i ei
∂xi
The curl of the gradient of a vector field is a null vector,
curl grad φ = ∇ × ∇φ = ε ijkφ,kj ei = 0
Divergence of a Vector Field
∂ ()
div (  ) = ∇ ⋅ (  ) = ⋅ ei
∂xi
The divergence of a vector field is a scalar field defined as,
∂u ∂ui ∂ui ∂ui
∇ ⋅ u = ⋅ e j = ei ⋅ e j = δ ij = =
div u = ui ,i
∂x j ∂x j ∂x j ∂xi
If the divergence of a vector field is zero, the vector field is said
to be solenoidal or div-free and there is a vector field such that,
div u = 0 ⇒ ∃ v | u = curl v
Divergence of the Curl of a Vector Field
The divergence differential operator div (·) and curl differential
operator curl (·) are defined, respectively, as,
∂ () ∂ ()
div (  ) = ∇ ⋅ (  ) = ⋅ ei , curl (  ) = ∇ × (  ) = ei ×
∂xi ∂xi
The curl of a vector field is a vector field defined as,
∂u ∂uk ∂uk
curl u= ∇ × u= e j × = e j × e k= ε ijk e= ε ijk uk , j ei
i
The divergence of a curl of a vector field is a null scalar,
div curl u = ∇ ⋅ ( ∇ × u ) = ε ijk uk , ji = 0

Laplacian of a Vector Field
The laplacian differential operator ∆ (  ) is defined as,
∂2 ()
∆ (  ) = ∇ ⋅∇ (  ) =
∂xi2
The laplacian of a vector field is a vector field defined as,
∂ 2ui
∆u = ∇ ⋅∇u = 2 ei = ui , jj ei
∂x j

Divergence of a Second Order Tensor Field
∂ ()
div (  ) = ∇ ⋅ (  ) = ⋅ ei
∂xi
The divergence of a second-order tensor field is a vector field
defined as,
div A = ∇ ⋅ A
∂A ∂Aik ∂Aik
= ⋅e j = ei ⊗ e k ⋅ e j = δ kj ei
∂Aij
= = ei Aij , j ei
∂x j
Curl of a Second-order Tensor Field
∂ ()
curl (  ) = ∇ × (  ) = ei ×
∂xi
The curl of a second-order tensor field is a second-order tensor
field defined as,
curl A = ∇ × A
∂A ∂Akj
=el × =el × ek ⊗ e j
∂xl ∂xl
∂Akj
= ε lki ei ⊗
= e j ε lki Akj ,l ei ⊗ e j
∂xl
Gradient of a Second-order Tensor Field
∂ () ∂ ()
grad (  ) = ∇ ⊗ (  ) = ⊗ ei = ∇ (  ) = ei
∂xi ∂xi
The gradient of a second-order tensor field is a third-order
tensor field defined as,
grad A = ∇ ⊗ A
∂A
= ⊗ ek
∂xk
∂Aij
= ei ⊗ e j ⊗ e= Aij ,k ei ⊗ e j ⊗ e k
∂xk
k

grad φ =∇φ =φ,i ei , ∆φ =∇ ⋅∇φ =φ,ii , ∇ ⊗ ∇φ =φ,ij ei ⊗ e j
div u = ∇ ⋅ u = ui ,i
grad u = ∇ ⊗ u = ui , j ei ⊗ e j ,
curl u = ∇ × u = ε ijk uk , j ei
∆u = ∇ ⋅∇u = ui , jj ei
div A = ∇ ⋅ A = Aij , j ei
grad A = ∇ ⊗ A = Aij ,k ei ⊗ e j ⊗ e k
curl A = ∇ × A = ε lki Akj ,l ei ⊗ e j

Assignments
Assignment 1.3
Establish the following identities involving a smooth scalar field φ
and a smooth vector field v,
(1) div (φ=
v ) φ div v + v ⋅ grad φ
(2) grad (φ v ) =
v ⊗ grad φ + φ grad v

Assignments
Establish the following identities involving the smooth scalar
fields φ and ψ , smooth vector fields u and v , and a smooth
second order tensor field A,
(1) (φ A ) φ div A + A grad φ
div=
(2) div ( A=
T
v) ( div A ) ⋅ v + A : grad v
(3) div ( u × v ) = v ⋅ curl u − u ⋅ curl v
(4) div ( u=
⊗ v) ( grad u ) v + u div v
(5) =
grad (φψ ) ( grad φ )ψ + φ ( gradψ )

Assignments
Assignment 1.5 [Homework]
Establish the following identities involving the smooth scalar
field φ , and smooth vector fields u and v ,
(1) (u ⋅ v )
grad= ( grad T
u ) (
v + grad T
v)u
(2) curl (φ=
v ) grad φ × v + φ curl v
(3) curl ( u × =
v ) u div v − v div u + ( grad u ) v − ( grad v ) u
(4) ∆v grad ( div v ) − curl ( curl v )
=
(5) ∆ ( u ⋅ v ) = ( ∆u ) ⋅ v + 2 grad u : grad v + u ⋅ ∆v

Assignments
Given the vector v =v ( x ) =x1 x2 x3e1 + x1 x2e 2 + x1e3 determine
div v, curl v, grad v, ∆v.

Introduction to Vectors and Tensors > Integral Theorems
Integral Theorems
Divergence or Gauss Theorem
Given a vector field u in a volume V with closed boundary
surface ∂V and outward unit normal to the boundary n, the
divergence (or Gauss) theorem reads,
∫
V
div u dV= ∫ ∇ ⋅ u dV= ∫
V ∂V
u ⋅ n dS
Given a second-order tensor field A in a volume V with closed

boundary surface ∂V and outward unit normal to the boundary n
the divergence (or Gauss) theorem reads,
∫
V
div A dV= ∫ ∇ ⋅ A dV= ∫
V ∂V
An dS

Introduction to Vectors and Tensors > Integral Theorems
Integral Theorems
Curl or Stokes Theorem
Given a vector field u in a surface S with closed boundary ∂S
and outward unit normal to the surface n , the curl (or Stokes)
theorem reads,
∫ ( curl u ) ⋅ n dS =∫ ( ∇ × u ) ⋅ n dS =∫
S S ∂S
u ⋅ dr
where the curve of the line integral must have positive
orientation, such that dr points counter-clockwise when the unit
normal points to the viewer, following the right-hand rule.

Continuum Mechanics
Chapter 2
Kinematics: Motion
Motion > Contents
Contents
Contents
1. Continuum Mechanics
2. Deformation map
3. Material and spatial descriptions
4. Displacement vector field
5. Velocity vector field
6. Acceleration vector field
7. Stationary field
8. Uniform velocity vector field
9. Trajectories
10. Streamlines
11. Material and spatial domains
October 7, 2016 Carlos Agelet de Saracibar 2
Motion > Continuum Mechanics
Continuum Mechanics
Continuum Mechanics
Continuum mechanics is a branch of mechanics that deals with
the analysis of the kinematics and mechanical behavior of
materials, either solids or fluids, modeled as a continuum body,
i.e. at a macroscopic scale.
Continuum mechanics deals with physical properties of

materials, solids or fluids, which are independent of any
particular coordinate system in which they are observed.
These physical properties are the represented by tensors, which

are mathematical objects that have the required property of
being independent of coordinate system.
Continuum Mechanics
Continuum Mechanics
Modeling an object as a continuum body assumes that the
substance of the object completely fills the space it occupies,
ignoring the fact that matter is made of atoms, and so is not
continuous at a microscopic scale.
However at a macroscopic scale, on length scales much greater

than that of inter‐atomic distances, such continuum models are
highly accurate.
Continuum mechanics is a powerful and effective tool to

succesfully describe macroscopic systems using a continuum
approach.
Continuum Mechanics
Continuum Mechanics
A continuum body is viewed as a continuous medium having a
continuous (or at least a piecewise continuous) distribution of
matter in space and time.
A continuum body is determined by macroscopic quantities

which are described by continuous functions with continuous
derivatives.
A continuum body, denoted by  , may be imagined as being a

composition of a (continuous) set of particles (or material
points), represented by P   .
Continuum Mechanics
Continuum Mechanics
Continuum mechanics includes the following key ingredients:
 Kinematics: Motion and deformation of a continuum body
 Stresses: Forces, stresses
 Balance laws: Fundamental laws of physics governing the
motion of a continuum body which are valid for any material
 Constitutive equations: Material characterization of a conti‐
nuum body
Motion > Deformation Map
Configurations
Configurations
Let us consider a continuum body  , which can be viewed as an
infinite set of particles P   , which is embedded in the three‐
dimensional Euclidean space at a given time t.
We introduce a reference frame of rectangular coordinate axes

at a fixed origin O with right‐handed orthonormal basis vectors
ea , a  1, 2,3
x, 3
e3
e2
O
x2 ,
e1
x1,
Configurations
Configurations
As the continuum body  moves in space along the time it
occupies a continuous sequence of geometrical regions denoted
as configurations 0 ,, t , which are uniquely determined at
any time t.
Any particle P   , at any time t, corresponds to a so‐called

geometrical point having a position in the configuration t.
Configurations
Reference or Material Configuration
 Current or Spatial Configuration
time t  0 time t
X x
0 Abstract Configuration
t
P
X  x
X 3 , x3
ê 3
ê 2 X 2 , x2
ê1
X 1 , x1
Configurations
Reference Configuration
The geometrical region 0 with the position of a typical point X
corresponds to a fixed reference time and is denoted as
reference (or material or undeformed) configuration of the
body  .
The point X corresponds to the position occupied by the particle

P   at the reference time. The particle P may be identified
by the position vector (or material or referential position) X of
the point X relative to the fixed origin O.
It is often convenient to call X as the material point X associated

with the particle P   at the fixed reference time.
Configurations
Initial Configuration
A geometrical region at initial time t=0 is referred to as the
initial configuration.
We agree subsequently that the initial configuration coincides

with the reference configuration, hence, we will assume that the
reference time is t=0.
Configurations
Current Configuration
The geometrical region t with the position of a typical point x
corresponds to the current time t>0 and is denoted as current
(or spatial or deformed) configuration of the body  .
The point x corresponds to the position occupied by the particle

P   at the current time t>0. The position vector (or spatial or
current position) relative
x to the fixed origin O can be used as a
label for the associated point x.
It is often convenient to call x as the spatial point x associated

with the particle P   at the current time t>0.
Deformation Map
time t  0 time t
X x
t
0 
P
X  x
Material Map X 3 , x3 Spatial Map
X   0 P  , P    X x   P, t  , P   1  x, t 
1 ê 3
0
ê 2 X 2 , x2
Inverse Deformation Map ê1 Deformation Map
X   0   x, t      x, t  x    01  X  , t     X, t 
X 1 , x1
1 1
Deformation Map
Material Map
A particle P may be identified by the position vector (or material
or referential position) of the point X relative to the fixed origin
O, denoted as , through
X the one‐to‐one material map,
X   0 P  , P   01  X 
Spatial Map
A particle P may be identified by the position vector (or spatial
or current position) of the point x relative to the fixed origin O,
denoted as , through
x the one‐to‐one spatial map,
x   P, t  , P   1  x, t 
Deformation Map
Deformation Map
The composition of the spatial map and the inverse of the
material map, yields the one‐to‐one deformation map defining
the equation of motion given by,
x    01  X  , t     X, t 
Inverse Deformation Map

The composition of the material map and the inverse of the
spatial map, yields the one‐to‐one inverse deformation map
defining the inverse of the equation of motion given by,
X   0  1  x, t     1  x, t 
Deformation Map
Material Coordinates
The vector position of a material point, denoted as , may
X be
written as a linear combination of the orthonormal basis vectors
e1 , e 2 , e3 , i.e., the Cartesian basis, such that,
X  X 1e1  X 2e2  X 3e3  X Ae A

where the components X 1 , X 2 , X 3 are denoted as material
coordinates.
Using matrix notation, the vector of material coordinates,
 X
denoted as , takes the form,
 X   X 1 X3
T
X2
Deformation Map
Spatial Coordinates
The vector position of a spatial point, denoted as , may
x be
written as a linear combination of the orthonormal basis vectors
e1 , e 2 , e3 , i.e., the Cartesian basis, such that,
x  x1e1  x2e 2  x3e3  xa e a

where the components x1 , x2 , x3 are denoted as spatial
coordinates.
Using matrix notation, the vector of spatial coordinates,
 x
denoted as , takes the form,
 x   x1 x3 
T
x2
Material Differential Operators
Material Nabla, Laplacian and Hessian

The material nabla vector differential operator, denoted as , is
defined as,
 
   A1,3 eA  eA
X A X A

The material laplacian scalar differential operator, denoted as ,
is defined as,
2
    
X A2
The material hessian symmetric second‐order tensor differential
operator is defined as,
2
   e A  eB
X AX B
Material Differential Operators
Material Divergence, Curl and Gradient
The material divergence differential operator DIV (∙) is defined
as,
 
DIV            eA
X A
The material curl differential operator CURL (∙) is defined as,
 
CURL           e A 
X A
The material gradient differential operator GRAD (∙) is defined
as,
   
GRAD            eA     eA
X A X A
Spatial Differential Operators

Spatial Nabla, Laplacian and Hessian

The spatial nabla vector differential operator, denoted as , is
defined as,
 
   a 1,3 ea  ea
xa xa
 ,
The spatial laplacian scalar differential operator, denoted as
is defined as,
2
     2
xa
The spatial hessian symmetric second‐order tensor differential
operator is defined as,
2
   e a  eb
xa xb
Spatial Differential Operators

Spatial Divergence, Curl and Gradient
The spatial divergence differential operator div (∙) is defined as,
 
div            ea
xa
The spatial curl differential operator curl (∙) is defined as,
 
curl           ea 
xa
The spatial gradient differential operator grad (∙) is defined as,
   
grad            ea       ea
xa xa
Deformation Map
time t  0 time t
X x
t
0 
P
X  x
Material Map X 3 , x3 Spatial Map
X   0 P  , P   01  X  x   P, t  , P   1  x, t 
ê 3
ê 2 X 2 , x2
Inverse Deformation Map ê1 Deformation Map
X   0   x, t      x, t  x    01  X  , t     X, t 
X 1 , x1
1 1
Deformation Map
Deformation Map
The deformation map x    X,t  has to satisfy the following
conditions,
1. Continuous with continuous derivatives up to the required
continuity degree
2. Consistency condition, i.e. taking t=0 as reference time,
X    X, 0 
3. One‐to‐one map, i.e. there exists the inverse of the deforma‐
tion map,
X   1  x,t 
4. Positive Jacobian, i.e. positive differential of volume,
J : det GRAD   X, t    0
Deformation Map
time t  0 time t
dV dv
X x
0 0 Abstract Configuration  t
0 
P
X  x
X 3 , x3
Inverse Differential Volume Map
ê 3
Differential Volume Map
dV  J 1dv dv  J dV
ê 2 X 2 , x2
Inverse Density Map ê1 Density Map
0  J    J 1 0
X 1 , x1
Deformation Map
Jacobian
The jacobian of the deformation map is a positive real value and
takes the form,
J : det GRAD   X, t    0
The following relation holds,
dv  J dV
Note that at the reference time for t=0,

dv  J dV  dV  J 1
Motion > Material and Spatial Descriptions
Material and Spatial Descriptions
Material Description
Using a material description, any arbitrary property  (of any
tensorial order) involved in the description of a continuum body,
is mathematically described as a function of the material points
(or material vector positions) and time t, i.e.,
X
    X, t 
Spatial Description
Using a spatial description, any arbitrary property  (of any
tensorial order) involved in the description of a continuum body,
is mathematically described as a function of the spatial points
(or spatial vector positions) and time t, i.e.,
x
    x, t 
Material Description
The material description of an arbitrary property  (of any
tensorial order) provides the time‐evolution of the property for a
given particle or material point X and is typically used in solid
mechanics.
    X, t 
Spatial Description
The spatial description of an arbitrary property  (of any
tensorial order) provides the time‐evolution of the property at a
fixed spatial point x and is typically used in fluid mechanics.
    x, t 
Giving the material description of an arbitrary property     X,t 
and the inverse of the motion equation X   1  x,t  , the spatial
description of the property reads,
    X, t     1  x, t  , t     x, t 
Giving the spatial description of an arbitrary property     x,t 
and the motion equation x    X,t  , the material description of
the property reads,
    x, t      X, t  , t     X, t 
Material and Spatial Time Derivatives
Material Time Derivative
    X,t 
Giving the material description of an arbitrary property,
the material time derivative of the property is given by,
d    X, t    X, t 
   
dt t X
t
Spatial Time Derivative
    x,t 
Giving the spatial description of an arbitrary property,
the spatial (or local) time derivative of the property is given by,
   x, t    x, t 
 
t t x t
Motion > Displacement Vector Field
Displacement Vector Field
time t  0 time t
X x
0
u  xX t
X x
X 3 , x3
ê 3
ê 2 X 2 , x2
ê1
X 1 , x1
Motion > Displacement Vector Field
The displacement vector field, denoted as , is
u defined as the
difference between the spatial and material vector positions,
u  xX
The material description of the displacement vector field takes
the form,
u  U  X, t     X, t   X
The spatial description of the displacement vector field takes the
form,
u  u  x, t   x   1  x, t 
Motion > Velocity Vector Field
Velocity Vector Field
The velocity vector field, denoted as , is
v defined as the time
variation of the spatial vector position,
dx
v
dt
The material description of the velocity vector field takes the
form,
  X, t 
v  V  X, t  
t
The spatial description of the velocity vector field takes the
form,
v  V  1  x, t  , t   v  x, t 
Motion > Acceleration Vector Field
Acceleration Vector Field
The acceleration vector field, denoted as , is
a defined as the
second time derivative of the spatial vector position,
d  dx  dv
a  
dt  dt  dt
The material description of the acceleration vector field takes
the form,
 2  X, t  V  X, t 
a  A  X, t   
t 2
t
The spatial description of the acceleration vector field takes the
form,
a  A  1  x, t  , t   a  x, t 
    x, t  ,
Given the spatial description of an arbitrary property,
the material time derivative of the property can be written as,
d  d   x  t  , t    x, t    x, t  dxa  t 
     
dt dt t xa dt
d  d   x  t  , t    x, t 
      grad   x, t    v  x, t 
dt dt t 
Convective time derivative
The material time derivative of an arbitrary property given in

spatial description may be written as the sum of its spatial (or
local) time derivative and its convective time derivative.
Convective Time Derivative
Convective Time Derivative
The convective time derivative of an arbitrary property given in
    x, t  ,
spatial description, may be defined as the difference
between its material time derivative and its spatial (or local)
time derivative, yielding,
  x, t  d    x, t 
 grad   x, t    v  x, t     t  dt  t
The acceleration vector field may be also defined as the material
time derivative of the velocity vector field.
Using a material description for the velocity vector field, the
material description of the acceleration vector field takes the
form,
dv V  X, t 
a   A  X, t 
dt t
dva Va  X, t 
aa    Aa  X, t 
dt t
The acceleration vector field may be also defined as the material
time derivative of the velocity vector field.
Using a spatial description for the velocity vector field, the
spatial description of the acceleration vector field takes the
form,
dv dv  x, t  v  x, t 
a     grad v  x, t   v  x, t   a  x, t 
dt dt t
dva dva  x, t  va  x, t 
aa      grad va  x, t    v  x, t 
dt dt t
va  x, t 
  va ,b  x, t  vb  x, t   aa  x, t 
t
Motion > Kinematics of Deformation
Kinematics of Deformation
Giving the spatial description of an arbitrary property, the material time derivative of the property can be written as,
U  X, t     X, t   X, u  x, t   x   1  x, t 
  X, t 
V  X, t   , v  x, t   V  1  x, t  , t 
t
V  X, t 
A  X, t   , a  x, t   A  1  x, t  , t 
t
v  x, t 
   grad v  x, t   v  x, t 
t
Motion > Assignments
Assignment 2.1
Assignment 2.1
The Cartesian components of the spatial description of the
velocity field is,
vx  x, t   x  z, v y  x, t   z  et  et  , vz  x, t   0
Compute the acceleration at the fixed spatial point with
Cartesian coordinates (1,1,1) at time t=2.
Assignment 2.2
The Cartesian components of the canonical form of a motion
equation, i.e. deformation map, are given by,
x  X, t   Xet , y  X, t   Yet , z  X, t   Z  Xt
1) Compute the acceleration vector field at the fixed spatial

point with Cartesian coordinates (1,1,1).
2) Compute the acceleration vector field at the fixed material
point with Cartesian coordinates (1,1,1).
3) Compute the rate of change of the velocity vector field per
unit of time at the fixed spatial point with Cartesian
coordinates (1,1,1).
Motion > Stationary Field
Stationary Field
Stationary Field
    x,t 
An arbitrary property given in spatial description as
is said to be stationary if and only if the following condition is
satisfied,
   x, t 
  0      x
t t
If an arbitrary property is stationary, its material time derivative
does not needs to be stationary and, in general, will be different
than zero,
d    x 
     grad   x    v  x, t   0
dt t
Motion > Stationary Field
Stationary Velocity Vector Field

Stationary Velocity Vector Field
The velocity vector field is said to be stationary if and only if the
following condition is satisfied,
v v  x, t 
  0  v  v  x
t t
If the velocity vector field is stationary, the acceleration vector
field has to be also stationary, but, in general, different than
zero. Note that the opposite is not true.
v  x 
a  x    grad v  x   v  x    grad v  x   v  x   0
t
Motion > Uniform Velocity Vector Field
Uniform Velocity Vector Field

Uniform Velocity Vector Field
A velocity vector field is said to be uniform if and only if the
following condition is satisfied,
v  v  x, t   v  t  x  t
If the velocity vector field is uniform, the acceleration vector
field has to be also uniform, but, in general, different than zero.
Note that the opposite is not true.
v  t  v  t 
a t     grad v  t   v  t   0
t t
Motion > Trajectories
Trajectories
Trajectories
The motion equation provides the sequence of spatial positions
occupied for any particle at any time, defining a time‐parame‐
trized family of curves denoted as trajectories (or path lines).

time t  0 time t
x
X
0 t
X 3 , x3
X
ê 3 x
x  t     X, t  X
ê 2 X 2 , x2
ê1
X 1 , x1
Differential Equation of the Trajectories

A trajectory can be described in differential form by means of
the spatial velocity vector field, through the following set of
differential equations,
dx  t 
 v  x t  , t 
dt
Integrating the set of differential equations yields,
x    C,t 
where C is a vector of integration constants with Cartesian
components given by,
C  C1 C3 
T
C2

Imposing the consistency condition, taking t=0 as reference
time, yields,
X    C, 0 
Then, the vector of integration constants can be expressed in
terms of the vector of material points, yielding,
C   1  X, 0 
Then, the canonical form of the motion equation reads,
x    C, t     1  X, 0  , t     X, t 
Assignment 2.3
Assignment 2.3
velocity vector field are,
y
vx  x, t   y, v y  x, t   , vz  x, t   z
1 t
Compute the canonical form of the trajectories.
Motion > Streamlines
Streamlines
Streamlines
The streamlines are a time‐dependent family of spatial curves
which at any time t are the envelope of the spatial velocity vector
field, i.e. the velocity vector field is tangent to the streamlines at
any spatial point, at any time t.
Current or Spatial Configuration

time t
x   v  x,t 
Differential Equation of the Streamlines

Differential Equation of the Streamlines
The differential equation of the streamlines may be obtained
imposing the condition that the spatial vector velocity field
v  x    ,t  is tangent to the streamlines x   . The parameter of
the stramlines, denoted as , is  chosen such that the velocity is
equal to the tangent to the streamlines, yielding,
dx   
 v x   ,t 
d
Integrating the differential equations, collecting the integration
constants in vector form, yields the equation of the streamlines,
x    C,  ,t 
Streamlines for a Stationary Motion

Streamlines for a Stationary Motion
If the velocity vector field is stationary the streamlines are
stationary and coincide with the trajectories.
If the velocity vector field is stationary, the trajectories and
streamlines have the same differential equations, yielding,
dx  t  dx   
 v  x t    v  x   
dt d
x    C, t   x    C,  
Assignment 2.4
Assignment 2.4
velocity vector field are,
y
vx  x, t   y, v y  x, t   , vz  x, t   z
1 t
Obtain the equation of the streamlines.
Assignment 2.5
The Cartesian components of the spatial description of a
stationary velocity vector field are,
vx  x, t   y, v y  x, t   y, vz  x, t   z
Obtain the canonical form of the trajectories and the equation of

the streamlines.
Assignment 2.6
The Cartesian components of the streamlines are given by,
x  C1et , y  C2 et , z  C3e2 t

where the parameter  is such that the velocity satisfies the
equation,
dx   
 v x   ,t 
d
Obtain the canonical form of the trajectories, taking t=0 as
reference time.
Motion > Material and Spatial Domains
Material Surface
Material Surface
A material surface is defined by the different positions occupied
in the space by the particles that at the reference time were on a
given surface.
The material description of a material surface may be written
as,
S  X | F  X   0
where the time‐independency of the material function guaran‐
tees that the particles satisfying this equation are always the
same ones, for any time t.
Note that,
F  X, t 
F  X, t   F  X   0
t
Material Surface
Material Surface
The spatial description of the function may be obtained using
the inverse motion equation yielding,
F  X   F  1  x, t    f  x, t 
Additionally,
F  X  df  x, t 
 0
t dt
The spatial description of a material surface may be written as,
 df  x, t  
S  x | f  x, t   0 and  0
 dt 
Spatial Surface
Spatial Surface
A spatial surface is defined by the same fixed spatial points at
any time t. Then, at different times t, different particles will be
on a spatial surface.
The spatial description of a spatial surface may be written as,
S  x | f  x   0
where the time‐independency of the spatial function guaran‐
tees that the spatial points satisfying this equation are always
the same ones, for any time t.
Material Volume
Material Volume
A material volume is a volume defined by a closed material
surface.
A material volume, written in material description, takes the
form,
V  X | F  X   0
and, in spatial description, takes the form,
 df  x, t  
V  x | f  x, t   0 and  0
 dt 
Spatial Volume
Spatial Volume
A spatial volume is a volume defined by a closed spatial surface.
The spatial description of a spatial volume may be written as,
V  x | f  x   0
Assignment 2.7
The Cartesian components of the spatial description of a velocity
vector field are,
z
vx  x, t   2ax, v y  x, t   by, vz  x, t   
t c
where a, b, and c are constants.
(i) Obtain the equation of the trajectories, taking t=0 as
reference time.
(ii) Obtain the equation of the streamlines.
(iii) Determine all the feasible values of the constants a, b, and c,
for which the motion equation has physical sense for any
time t>0.
Assignment 2.7
(iv) Obtain the spatial density at any time t, asuming that

the density at the reference configuration is 0  cte
(v) Obtain at any time t, the spatial equation of the
material surface which at time t=1 is a sphere of radius R
and center in (0,0,0)
Assignment 2.8
The Cartesian components of the spatial description of a
stationary velocity vector field are,
vx  x, t   0, v y  x, t   0, vz  x, t   f  x, y  z
The spatial description of a material surface is given by:

g  x, t   log  xyz    x f  x, y  t  0
(i) Determine the expression of the function f  x, y .

(ii) Obtain the trajectories, taking t=0 as reference time.
(iii) Obtain the streamlines.
Continuum Mechanics
Chapter 3
Kinematics: Strains
Strains > Contents
Contents
Chapter 3 · Strains
1. Tangent deformation map
2. Displacement gradient tensors
3. Strain tensors
4. Volumetric deformation
5. Area deformation
6. Polar decomposition
7. Stretches
8. Variation of angles
9. Assignments
10. Spatial velocity gradient
11. Material time derivatives
Strains > Tangent Deformation Map
Tangent Deformation Map

Reference or Material Configuration ϕϕ Current or Spatial Configuration
Reference Configuration Current Configuration
time t = 0 time t
t=0 t
Q FF
dX P’
dx
P
dS ds Q’
Ω0 Ωt
X x
Inverse Deformation Map X 3 , x3
Deformation Map
X = ϕ −1 ( x,t ) ê 3 x = ϕ ( X,t )
Inverse Tangent Deformation Map ê 2 X 2 , x2 Tangent Deformation Map
ê1
dX = F −1
( x, t ) dx X 1 , x1 dx = F ( X, t ) dX


Deformation Gradient
Let us consider the deformation map given by,
x = ϕ ( X,t )
Differentiating the deformation map, keeping constant the
configuration at time t, using the chain rule, yields,
dx = ( GRAD ϕ ( X, t ) ) dX = ( ∇ ⊗ ϕ ( X, t ) ) dX := F ( X, t ) dX
where the non-symmetric second-order deformation gradient
tensor, denoted as F ( X,t ) , has been introduced as,
F ( X, t ) := ∇ ⊗ ϕ ( X, t ) = GRAD ϕ ( X, t ) , FaA := ϕa , A


Inverse Deformation Gradient
Let us consider the inverse deformation map given by,
X = ϕ −1 ( x,t )
Differentiating the inverse deformation map, keeping constant
the configuration at time t, using the chain rule, yields,
dX = ( grad ϕ −1 ( x, t ) ) dx = ( ∇ ⊗ ϕ −1 ( x, t ) ) dx := F −1 ( x, t ) dx
where the non-symmetric second-order inverse deformation
gradient tensor, denoted as F −1 ( X,t ) , has been introduced as,
F −1 ( x, t ) := ∇ ⊗ ϕ −1 ( x, t ) = grad ϕ −1 ( x, t ) , FAa−1 := ϕ A−1,a


Uniform Deformation Gradient
Let us consider a uniform deformation gradient such that,
dx = F ( t ) dX, dxa = FaA ( t ) dX A

As the deformation gradient is uniform, the deformation map is
linear and it can be easily obtained, integrating, yielding,
x = F ( t ) X + C ( t ) , xa = FaA ( t ) X A + Ca ( t )
where C ( t ) is a vector of integration constants, such that,

assuming the reference time is t=0, satisfies the condition
C ( 0) = 0 .


Giving the spatial description of an Gradient
Deformation arbitrary property, the material time derivative of the property can be written as,
dx = F ( X, t ) dX, dxa = FaA dX A

F ( X, t ) := ∇ ⊗ ϕ ( X, t ) = GRAD ϕ ( X, t ) , FaA = ϕa , A
J := det F ( X, t ) = det ( GRAD ϕ ( X, t ) ) > 0
Inverse Deformation Gradient
dX = F −1 ( x, t ) dx, dX A = FAa−1dxa
F −1 ( x, t ) := ∇ ⊗ ϕ −1 ( x, t ) = grad ϕ −1 ( x, t ) , FAa−1 := ϕ A−1,a
J −1 := det F −1 ( x, t ) = det ( grad ϕ −1 ( x, t ) ) > 0
Example 3.1
Example 3.1
Compute the deformation gradient and inverse deformation
gradient for a motion equation with Cartesian components,
T
ϕ ( X, t )  =  X + Y t Y (1 + t ) Ze 
2 t
The Cartesian components of the deformation gradient are,

1 2Yt 0 
0 1 + t 0 
[ ] 
F = 
0 0 et 
and the jacobian takes the value,
J = det F = (1 + t ) et > 0
Example 3.1
The Cartesian components of the inverse motion equation are,
T
 t y 
ϕ ( x, t )  =  x − y
−1 2
2
−t
ze 
 (1 + t ) 1+ t 
The Cartesian components of the inverse deformation gradient
are,
 t 
1 −2 y 2
0
 (1 + t ) 
 1 
F  = 0
−1
0
 1+ t 
0 0 e−t 
 
 
Strains > Displacement Gradient Tensors
Material Displacement Gradient

Let us consider the material description of the displacement
vector field, given by,
u = U ( X,t )
Differentiating the material description of the displacements,
keeping constant the time t, using the chain rule, yields,
du = ( GRAD U ( X, t ) ) dX = ( ∇ ⊗ U ( X, t ) ) dX := J ( X, t ) dX
where the non-symmetric second-order material displacement
gradient tensor, denoted as J ( X,t ), has been introduced as,
J ( X, t ) := ∇ ⊗ U ( X, t ) = GRAD U ( X, t ) , J aA = U a , A


The material description of the displacement vector field may be
written as
U ( X, t ) = ϕ ( X, t ) − X, U a = ϕa − X a
Taking the material gradient, the material displacement
gradient tensor may be related to the deformation gradient
tensor, yielding,
J ( X, t ) = GRAD U ( X, t ) = GRAD ϕ ( X, t ) − 1 = F ( X, t ) − 1,
J aA = U a , A = ϕa , A − δ aA = FaA − δ aA

Spatial Displacement Gradient

Let us consider the spatial description of the displacement
vector field, given by,
u = u ( x,t )
Differentiating the spatial description of the displacements,
keeping constant the time t, using the chain rule, yields,
du = ( grad u ( x, t ) ) dx = ( ∇ ⊗ u ( x, t ) ) dx := j ( x, t ) dx
where the non-symmetric second-order spatial displacement
gradient tensor, denoted as j ( x,t ) , has been introduced as,
j ( x, t ) := ∇ ⊗ u ( x, t ) = grad u ( x, t ) , j Aa := u A,a


The spatial description of the displacement vector field may be
written as
u ( x, t ) = x − ϕ −1 ( x, t ) , u A = xA − ϕ A−1
Taking the spatial gradient, the spatial displacement gradient
tensor may be related to the inverse deformation gradient
tensor, yielding,
j ( x, t ) = grad u ( x, t ) = 1 − grad ϕ −1 ( x, t ) = 1 − F −1 ( x, t ) ,
j Aa = u A,a = δ Aa − ϕ A−1,a = δ Aa − FAa−1

Displacement Gradient Tensors

Material Displacement
Giving the spatial description of an arbitrary property, Gradient
dU ( X, t ) = J ( X, t ) dX, dU a = J aA dX A
J ( X, t ) = ∇ ⊗ U ( X, t ) = GRAD U ( X, t ) , J aA = U a , A
J ( X, t ) = F ( X, t ) − 1, J aA = FaA − δ aA
du ( x, t ) = j ( x, t ) dx, du A = jAa dxa

j ( x, t ) = ∇ ⊗ u ( x, t ) = grad u ( x, t ) , j Aa = u A,a
j ( x, t ) = 1 − F −1 ( x, t ) , j Aa = δ Aa − FAa−1

Strains > Push-forward / Pull-back
Push-forward / Pull-back
Let us consider an arbitrary scalar field Ɵ such that,
θ = θ ( x, t ) = θ (ϕ ( X, t ) , t ) = Θ ( X, t )
The material gradient of a scalar field can be written as the pull-
back of the spatial gradient of the scalar field given by,
∂Θ ( X, t ) ∂θ ( x, t ) ∂ϕa ( X, t )
( GRAD Θ ( X, t ) ) A
=
∂X A
=
∂xa ∂X A
= ( grad θ ( x, t ) )a FaA ( X, t ) = FAa
T
( X, t ) ( grad θ ( x, t ) )a
GRAD Θ ( X, t ) = FT grad θ ( x, t )
Let us consider an arbitrary scalar field Ɵ such that,
θ = Θ ( X, t ) = Θ (ϕ −1 ( x, t ) , t ) = θ ( x, t )
The spatial gradient of a scalar field can be written as the push-
forward of the material gradient of the scalar field given by,
∂θ ( x, t ) ∂Θ ( X, t ) ∂ϕ A−1 ( x, t )
( grad θ ( x, t ) ) a
=
∂xa
=
∂X A ∂xa
= ( GRAD Θ ( X, t ) ) A FAa−1 ( x, t ) = FaA−T ( x, t ) ( GRAD Θ ( X, t ) ) A
grad θ ( x, t ) = F −T GRAD Θ ( X, t )
Let us consider an arbitrary vector field u such that,
u = u ( x, t ) = u (ϕ ( X, t ) , t ) = U ( X, t )
The material gradient of a vector field can be written as the pull-
back of the spatial gradient of the vector field given by,
∂U a ( X, t ) ∂ua ( x, t ) ∂ϕb ( X, t )
( GRAD U ( X, t ) ) aA
=
∂X A
=
∂xb ∂X A
= ( grad u ( x, t ) )ab FbA ( X, t )
GRAD U ( X, t ) = ( grad u ( x, t ) ) F
u = U ( X, t ) = U (ϕ −1 ( x, t ) , t ) = u ( x, t )
The spatial gradient of a vector field can be written as the push-
forward of the material gradient of the vector field given by,
∂u A ( x, t ) ∂U A ( X, t ) ∂ϕ B−1 ( x, t )
( grad u ( x, t ) ) Aa
=
∂xa
=
∂X B ∂xa
= ( GRAD U ( X, t ) ) AB FBa−1 ( x, t )
grad u ( x, t ) = ( GRAD U ( X, t ) ) F −1
u = u ( x, t ) = u (ϕ ( X, t ) , t ) = U ( X, t )
The material divergence of a vector field can be written in terms
of the material or spatial gradient of the vector field as,
∂U A ( X, t ) ∂u A ( x, t ) ∂ϕa ( X, t )
DIV U ( X, t ) = =
∂X A ∂xa ∂X A
= ( grad u ( x, t ) ) Aa FaA ( X, t ) = ( grad u ( x, t ) ) Aa FAa
T
( X, t )
DIV U ( X, t ) = ( GRAD U ( X, t ) ) : 1 = ( grad u ( x, t ) ) : FT

u = U ( X, t ) = U (ϕ −1 ( x, t ) , t ) = u ( x, t )
The spatial divergence of a vector field can be written in terms of
the material or spatial gradient of the vector field as,
∂ua ( x, t ) ∂U a ( X, t ) ∂ϕ A−1 ( x, t )
div u ( x, t ) = =
∂xa ∂X A ∂xa
= ( GRAD U ( X, t ) )aA FAa−1 ( x, t ) = ( GRAD U ( X, t ) )aA FaA−T ( x, t )
div u ( x, t ) = ( grad u ( x, t ) ) : 1 = ( GRAD U ( X, t ) ) : F −T

Strains > Strain Tensors
Deformation Tensors
timeReference
t = 0 Configuration Current Configuration
time t
t=0 t
Q FF
dX P’
dx
P
dS ds Q’
Ω0 Ωt
X x
Inverse Deformation Map Deformation Map
X 3 , x3
X = ϕ −1 ( x,t ) ê 3 x = ϕ ( X,t )
ê1
dX = F−1 ( x, t ) dx X 1 , x1
dx = F ( X, t ) dX
dS = dX ds = dx
Cauchy-Green Deformation Tensors

Right Cauchy-Green Deformation Tensor
The square of the norm of the differential vector dx may be
written as,
2
ds = dx = dx ⋅ dx = dX ⋅ FT FdX := dX ⋅ CdX,
2
T
ds 2 = dxa dxa = dX A FAa FaB dX B := dX AC AB dX B
where the symmetric positive-definite second-order right
Cauchy-Green deformation tensor, denoted as C , has been
defined as,
C := FT F, C AB := FAa
T
FaB = FaA FaB
with
2
det C := ( det F ) = J 2 > 0
Cauchy-Green Deformation Tensors

Left Cauchy-Green Deformation Tensor
The square of the norm of the differential vector dX may be
written as,
2
dS = dX = dX ⋅ dX = dx ⋅ F −T F −1dx := dx ⋅ b −1dx,
2
dS 2 = dX A dX A = dxa FaA−T FAb−1dxb := dxa bab−1dxb

where the symmetric positive-definite second-order left Cauchy-
Green deformation tensor, denoted as b , has been defined as,
b := FFT , bab := FaA FAb
T
= FaA FbA
with
2
det b := ( det F ) = J 2 > 0

Green-Lagrange Strain Tensor

Green-Lagrange Strain Tensor
Let us consider the following scalar quantity as strain measure,
ds 2 − dS 2 = dx ⋅ dx − dX ⋅ dX
= dX ⋅ ( C − 1) dX := 2dX ⋅ EdX
where the symmetric second-order Green-Lagrange (or
material) strain tensor, denoted as E, has been defined as,
1 1 T 1
E := ( C − 1) = ( F F − 1) = ( J + JT + JT J ) ,
2 2 2
1 1 1
E AB := ( C AB − δ AB ) = ( FaA FaB − δ AB ) = ( J AB + J BA + J CA J CB )
2 2 2
Almansi Strain Tensor

Let us consider the following scalar quantity as strain measure,
ds 2 − dS 2 = dx ⋅ dx − dX ⋅ dX
= dx ⋅ (1 − b −1 ) dx := 2dx ⋅ edx
where the symmetric second-order Almansi (or spatial) strain
tensor, denoted as e , has been defined as,
1 1 1
e := (1 − b ) = (1 − F F ) = ( j + jT − jT j) ,
−1 −T −1
2 2 2
1 1 1
eab := (δ ab − bab ) = (δ ab − FAa FAb ) = ( jab + jba − jca jcb )
−1 −1 −1
2 2 2
Deformation Tensors
Right Cauchy-Green
Giving the spatial description of an arbitrary property, Deformation
the material time derivativeTensor
of the property can be written as,
C := FT F, C AB := FAa
T
FaB = FaA FaB
C := 1 + 2E, C AB := δ AB + 2 E AB
ds 2 := dX ⋅ CdX, ds 2 := dX AC AB dX B
Left Cauchy-Green Deformation Tensor

b := FFT , bab := FaA FAb
T
= FaA FbA
b −1 := F −T F −1 , bab−1 := FaA−T FAb−1 = FAa−1 FAb−1
dS 2 := dx ⋅ b −1dx, dS 2 := dxa bab−1dxb

Strain Tensors
Green-Lagrange Strain the
Giving the spatial description of an arbitrary property, Tensor
material time derivative of the property can be written as,
ds 2 − dS 2 := 2dX ⋅ EdX, ds 2 − dS 2 := 2dX A E AB dX B
ds 2 − dS 2 := 2dx ⋅ edx, ds 2 − dS 2 := 2dxa eab dxb

Strain Tensors
Green-Lagrange Strain the
Giving the spatial description of an arbitrary property, Tensor
material time derivative of the property can be written as,
1 1 T 1
E := ( C − 1) = ( F F − 1) = ( J + J T + JT J ) ,
2 2 2
1 1 1
E AB := ( C AB − δ AB ) = ( FaA FaB − δ AB ) = ( J AB + J BA + J CA J CB )
2 2 2
1 1 1
e := (1 − b ) = (1 − F F ) = ( j + jT − jT j) ,
−1 −T −1
2 2 2
1 1 1
eab := (δ ab − bab ) = (δ ab − FAa FAb ) = ( jab + jba − jca jcb )
−1 −1 −1
2 2 2
Push-forward / Pull-back Maps

Push-forward of a Covariant Second-order Tensor
The push-forward of a covariant second-order tensor is defined
as,
b b
ϕ* ( ) := F −T
()
F −1
The Almansi strain, spatial second-order unit and inverse of the

left Cauchy-Green deformation tensors can be viewed as the
push-forward of the (covariants) Green-Lagrange strain, right
Cauchy-Green deformation and material second-order unit
tensors, respecti-vely, such that,
e = ϕ* ( E ) := F −T EF −1 ,
1 = ϕ* ( C ) := F −T CF −1 ,
b −1 = ϕ* (1) := F −T 1F −1

Pull-back of a Covariant Second-order Tensor
The pull-back of a covariant second-order tensor is defined as,
b b
ϕ ( ) := F ( ) F
* T
The Green-Lagrange strain, right Cauchy-Green deformation

and material second-order unit tensors can be viewed as the
pull-back of the (covariants) Almansi strain, spatial second-order
unit and inverse of the left Cauchy-Green deformation tensors,
respectively, such that,
E = ϕ * ( e ) := FT eF,
C = ϕ * (1) := FT 1F,
1 = ϕ * ( b −1 ) := FT b −1F

time t = 0 time t
t
FF
E, C, 1 e, 1, b −1
Ω0 Ωt
Pull-back Maps X 3 , x3 Push-forward Maps

E = FT eF ê 3
e = F −T EF −1
ê 2 X 2 , x2
C = F 1F T
ê1
1 = F −T CF −1
1 = FT b −1F X 1 , x1
b −1 = F −T 1F −1
Example 3.2
Example 3.2
Compute the Green-Lagrange and Almansi strain tensors for a
motion equation given by,
T
[ x] = ϕ ( X, t )  =  X + Yt , Ye −t
, Ze 
t

Example 3.2
Example 3.2
Compute the Green-Lagrange and Almansi strain tensors for a
motion equation given by,
T
[ x] = ϕ ( X, t )  =  X + Yt , Ye −t
, Ze 
t
The components of the inverse motion, deformation gradient

and inverse deformation gradient are given by,
T
[ X] = ϕ ( x, t )  =  x − yte , ye , ze
−1 t t −t

1 t 0 1 −tet 0
0 e− t   
[ ] 
F = 0  , F  = 0 et
−1
0
0 0 et  0 0 e−t 

Example 3.2
The components of the right Cauchy-Green deformation tensor
and Green-Lagrange strain tensor take the form,
1 0 0  1 t 0  1 t 0
   t e−t 0  0 e−t 0  =  t t 2 + e−2t 0 
[ ]   
C = F T
F =
0 0 et  0 0 et  0 0 e2t 
0 t 0 
1  1 
[ ]  (
E = C − 1 ) 
= t t 2
+ e −2 t
− 1 0 
2  2
0 0 e − 1
2t
Note that at the reference configuration for t=0,

F = 1, C = 1, E = 0
Example 3.2
The components of the left Cauchy-Green deformation tensor
and Almansi strain tensor take the form,
 1 0 0  1 −te
t
0  1 −tet 0 
  0 et   t 
b −1  = F −T F −1  =  −tet et 0 0  =  −te (t 2
+ 1) e2t 0 
 
 0 0 e−t  0 0 e−t   0 0 −2 t 
e 

0 tet 0 
1 −1  1 t 
[e] =  (1 − b ) = te 1 − ( t + 1) e
2 2t
0 
2  2 −2 t 
 0 0 1 − e 
Note that at the reference configuration for t=0,

F = 1, b −1 = 1, e = 0
Strains > Volumetric Deformation
Volumetric Deformation Map

ϕ Current or Spatial Configuration
time t = 0 time t
X dV x
dv
ρ0 ρ
Ω0 Ωt
Inverse Differential Volume Map Differential Volume Map

dV = J −1dv X 3 , x3 dv = J dV
ê 3
Inverse Density Map Density Map
ê 2 X 2 , x2
ρ0 = J ρ ê1 ρ = J −1 ρ0
X 1 , x1

Strains > Volumetric Deformation
Volumetric Deformation
Volumetric Deformation
The volumetric deformation, denoted as e, is a scalar quantity
defined as,
dv − dV dv
e= = −1 = J −1
dV dV
The incompressibility condition, i.e. zero volumetric deforma-
tion, takes the form,
J =1

Strains > Area Deformation
Area Deformation
time t = 0 time t
t=0 t
FF
Ω0 dA Ωt
da
N X
X 3 , x3 nda
dA = dA N ê 3 x
dan = da n
ê 2 X 2 , x2
ê1
X 1 , x1

Area Deformation
Area Deformation
Let us consider a differential of area vector on the reference and
spatial configurations written in terms of the unit outward
normal to the surface on the material and spatial configurations,
respectively, given by,
dA = dA N, da = da n
Taking an arbitrary vector dX , associated differential of volumes
in the material and spatial configurations take the form,
dV = dX ⋅ dA = dX ⋅ dA N, dv = dx ⋅ da = dx ⋅ da n
where
dx = FdX, dv = JdV
Area Deformation
The differential of volumes satisfy the following expression,
dv = da ⋅ dx = da ⋅ FdX = dX ⋅ FT da
= J dV = J dA ⋅ dX = dX ⋅ J dA ∀dX ⇒ FT da = J dA
yielding the relation, known as Nanson’s formula, given by,
da = J F −T dA, da n = dA J F −T N

Strains > Polar Decomposition
Polar Decomposition
Polar Decomposition
For any non-singular second-order tensor, denoted as F , there
exist two unique symmetric positive-definite second-order
tensors, denoted as U and v , and a unique proper orthogonal
second-order tensor, denoted as R , such that,
F = RU = vR, FaA = RaBU BA = vab RbA
where,
12
U = ( F F ) , U = UT , x ⋅ Ux > 0 ∀x ≠ 0
T
T 12
v = ( FF ) , v = vT , x ⋅ vx > 0 ∀x ≠ 0
R = FU −1 = v −1F, R −1 = RT , det R = 1
Polar Decomposition
Polar Decomposition
The polar decomposition of the deformation gradient tensor F ,
reads,
where U is the right (or material) stretch tensor, v is the left (or
spatial) stretch tensor and R is the rotation tensor, such that,
12
U = (F F) T
= C1 2 , U = UT , x ⋅ Ux > 0 ∀x ≠ 0, det U = J
T 12
v = ( FF ) = b1 2 , v = vT , x ⋅ vx > 0 ∀x ≠ 0, det v = J
R = FU −1 = v −1F, R −1 = RT , det R = 1

Polar Decomposition
Polar Decomposition
The rotation tensor R rotates a material line segment dX onto
a unique spatial line segment dx = R dX , such that the norm of
the line segment is preserved.
2 2
dx = dx ⋅ dx = ( R dX ) ⋅ ( R dX ) = dX ⋅ R R dX = dX ⋅ dX = dX
T
The rotation tensor R rotates material line segments dX and dY

onto unique spatial line segments dx = R dX and dy = RdY,
such that the angle between the line segments is preserved.
dx ⋅ dy dX ⋅ RT R dY dX ⋅ dY
cos θ = = = = cos Θ
dx dy dx dy dX dY

Polar Decomposition
Polar Decomposition
The right (or material) stretch tensor U and the left (or spatial)
stretch tensor v satisfy the following pull-back and push-forward
relations with the rotation tensor,
U = R −1F = RT F = RT vR, U AB = RaAvab RbB
v = FR −1 = FRT = RURT , vab = RaAU AB RbB
The right Cauchy-Green tensor C and the left Cauchy-Green
tensor b satisfy the following pull-back and push-forward
relations with the rotation tensor,
C = U 2 = RT vRRT vR = RT v 2 R = RT bR, C AB = RaAbab RbB
b = v 2 = RURT RURT = RU 2 RT = RCRT , bab = RaAC AB RbB
Polar Decomposition
Polar Decomposition
A rigid body motion satisfies the following relations,
F = R ⇔ U = v =1 ⇔ E =e =0
A pure stretch deformation satisfies the following relations,

R =1 ⇔ F = U = v

Polar Decomposition
Polar Decomposition
Any deformation can be seen either as a composition of a right
(or material) stretch, characterized by U, with a rotation,
characterized by R , given by the right polar decomposition,
F = RU, FaA = RaBU BA
dx = FdX = RUdX = R ( UdX ) , dxa = FaA dX A = RaAU AB dX B
or as a composition of a rotation, characterized by R , with a left
(or spatial) stretch, characterized by v , given by the left polar
decomposition,
F = vR, FaA = vab RbA
dx = FdX = vRdX = v ( RdX ) , dxa = FaA dX A = vab RbA dX A
Polar Decomposition

Polar Decomposition

Polar Decomposition
Polar Decomposition
Material Right and Spatial Left Stretch Tensors

12
U = (F F)
12
T
= C , U AB = ( FaA FaB )
12
= C1AB2
T 12
v = ( FF ) 12
= b , vab = ( FaA FbA )
12 12
= bab
Rotation Tensor
R = FU −1 = v −1F, RaA = FaBU BA

−1 −1
= vab FbA
Strains > Stretches
Stretches
time t = 0 time t
t=0 t
Q
T FF
dX P’
dx
P
dS ds Q’
Ω0 Ωt
t
X x
X = ϕ −1 ( x, t ) X 3 , x3
x = ϕ ( X, t )
ê 3
dX = F−1 ( x, t ) dx ê 2 X 2 , x2
dx = F ( X, t ) dX
ê1
dS = dX ds = dx
X 1 , x1
dX = TdS dx = tds
Strains > Stretches
Stretch Vectors
Material Stretch Vector
Let us denote as λT the material stretch vector at a material
point X at time t, along a material direction given by the unit
vector T on the material configuration,
λT ( X, t ) = F ( X, t ) T, λT = FaATA a
Multiplying by dS yields,
dx = λT ( X, t ) dS = F ( X, t ) TdS = F ( X, t ) dX
Taking norms, the stretch, denoted as λ , is defined as,
ds = λT ( X, t ) dS := λ dS ,
12
λ := λT ( X, t ) = ( T ⋅ F FT ) = ( T ⋅ CT ) = (1 + 2T ⋅ ET )
T 12 12

Strains > Stretches
Stretch Vectors
Material Stretch Vector
Let us denote as λT the material stretch vector at a material
point X at time t, along a material direction given by the unit
vector T on the material configuration. The following situations
may arise
λ := λT ( X, t ) > 1 … extension, length increases

λ := λT ( X, t ) = 1 … length does not changes
λ := λT ( X, t ) < 1 … compression, length decreases

Strains > Stretches
Stretch Vectors
Spatial Stretch Vector
Let us denote as λt the spatial stretch vector at a spatial point x
at time t, along a spatial direction given by the unit vector t on
the spatial configuration,
λt ( x, t ) = F −1 ( x, t ) t, λt = FAa−1ta
A
Multiplying by ds yields,
dX = λt ( x, t ) ds = F −1 ( x, t ) tds = F −1 ( x, t ) dx
−1
Taking norms, the inverse stretch, denoted as λ , is defined as,
dS = λt ( x, t ) ds := λ −1ds,
12 12
λ := λt ( x, t ) = ( t ⋅ F F t ) = ( t ⋅ b t ) = (1 − 2t ⋅ et )
−1 −T −1 −1 12

Strains > Stretches
Stretches
Material Stretch
Giving the spatial description Vector the material time derivative of the property can be written as,
of an arbitrary property,
λT ( X, t ) := F ( X, t ) T, λT := FaATA a
ds := λ dS , λ := λT ( X, t )
12
λ := ( T ⋅ F FT ) = ( T ⋅ CT ) = (1 + 2T ⋅ ET )
T 12 12
Spatial Stretch Vector
λt ( x, t ) := F −1 ( x, t ) t, λt := FAa−1ta A
dS := λ −1ds, λ −1 := λt ( x, t )
12 12
λ := ( t ⋅ F F t ) = ( t ⋅ b t ) = (1 − 2t ⋅ et )
−1 −T −1 −1 12

Strains > Stretches
Physical Interpretation of E11 Component

Green-Lagrange Strain Component E11
Let us consider a material segment dX(1) = T(1) dS along the X1-
axis on the material configuration.
The material stretch along the X1 direction will be given by,
(1) 1 2
λ1 := (1 + 2T ⋅ ET ) 12
(1)
= (1 + 2E11 )
The length of the deformed segment will be given by,
12
ds := λ1dS = (1 + 2 E11 ) dS
and the Green-Lagrange component E11 may be interpreted as,
1  ds 2  1  ds 2 − dS 2 
E11 =  2 − 1 =  2 
2  dS  2  dS 
Strains > Stretches
Physical Interpretation of e11 Component

Almansi Strain Component e11
Let us consider a spatial segment dx(1) = t (1) ds along the x1-axis
on the spatial configuration.
The inverse stretch along the x1 direction will be given by,
(1) 1 2
λ := (1 − 2t ⋅ et ) 12
1
−1 (1)
= (1 − 2e11 )
The length of the material segment will be given by,
12
dS := λ ds = (1 − 2e11 ) ds
1
−1
and the Almansi strain component e11 may be interpreted as,

1  dS 2  1  ds 2 − dS 2 
e11 = 1 − 2  =  
2  ds  2  ds 2 
Strains > Stretches
Assignment 3.1
Assignment 3.1
The components of the Almansi strain tensor, with reference
time t=0, are given by,
 0 0 
−tetz
 
[e] =  0 0 0
 tz t 
 −te 0 t ( 2e − e ) 
tz
Compute at the reference time t=0, the length of a material

curve that at time t=2 is a straight line going from point a with
coordinates (0,0,0) to point b with coordinates (1,1,1).

Strains > Stretches
Assignment 3.2
Consider the equations of motion given by,
x = X, y = Y + Z 2t , z = Z + Y 2t
Compute at time t=1 the length of a material curve that at the
reference time t=0 was a straight line going from point A with
coordinates (0,0,0) to point B with coordinates (0,1,1).

Strains > Stretches
Assignment 3.3
Consider the equations of motion given by,
x = X + Yt , y = Y, z = Z
Compute at time t=2 the length of a material curve that at time
t=1 was a curve parametrized as,
x (α ) = 0, y (α ) = α 2 , z ( α ) = α 0 ≤α ≤1

Strains > Stretches
Assignment 3.4
Assignment 3.4
The components of the Green-Lagrange strain tensor, with
reference time t=0, are given by,
 0 tetX 0 
 tX 
[E] = te 0 0 
 0 0 tetY 

Compute at time t=1 the length of a material curve that at the
reference time t=0 was a straight line going from point A (1,1,1)
to point B (2,2,2).

Strains > Variation of Angles
Variation of Angles
time t = 0 time t
t=0 t
dS (2) ds (2) t (2)
T (2) FF dx(2)
R’
dX(2) R P’
P
Θ
θ
Ω0 (1) Q dx(1) Q’
Ωt
dX (1) ds (1)
dS (1) T t (1)
dX( a ) = F−1 ( x, t ) dx( a ) dx( a ) = F ( X, t ) dX( a )

X X 3 , x3 x
ê 3
dX( a ) = T( a ) dS ( a ) ê 2 X 2 , x2 dx( a ) = t ( a ) ds ( a )
ê1
(1) ( 2)
dX ⋅ dX X 1 , x1 dx(1) ⋅ dx( 2)
cos Θ = cos θ =
dS (1) dS ( 2) ds (1) ds ( 2)
Variation of Angles
Spatial Angle
The dot product of the two differential vectors at the spatial
configuration reads,
dx(1) ⋅ dx( 2) = ds(1) ds( 2) cos θ
Alternatively, it may be written in terms of the differential
vectors at the material configuration and using the unit vectors
and the stretches yields,
dx( ) ⋅ dx( ) = dX( ) ⋅ FT FdX( ) = dS ( ) dS ( ) T( ) ⋅ FT FT(
1 2 1 2 1 2 1 2)
(1) ( 2) −1(1) −1( 2 ) (1) T ( 2)

= ds ds λ λ T ⋅ F FT
= ds ( ) ds ( ) λ −1( ) λ −1( ) T( ) ⋅ CT(
1 2 1 2 1 2)
(1) ( 2) −1(1) −1( 2 ) (1) ( 2)

= ds ds λ λ T ⋅ (1 + 2E ) T
Variation of Angles
Spatial Angle
Comparing the two expressions, the angle between the two
segments at the spatial configuration is given by,
cos θ = λ −1(1) λ −1( 2) T(1) ⋅ (1 + 2E ) T( 2)

−1 2 −1 2
(
= 1 + 2T ⋅ ET (1) (1)
) ( 1 + 2T ⋅ ET ( 2) ( 2)
) T(1) ⋅ (1 + 2E ) T( 2)
T(1) ⋅ (1 + 2E ) T(2)
=
1 + 2 T(1) ⋅ E T(1) 1 + 2 T(2) ⋅ ET(2)

Variation of Angles
Material Angle
The dot product of the two differential vectors at the material
configuration reads,
dX( ) ⋅ dX( ) = dS ( ) dS ( ) cos Θ
1 2 1 2
Alternatively, it may be written in terms of the differential

vectors at the spatial configuration and using the unit vectors
and the stretches yields,
(1) ( 2) (1) −T −1 ( 2) (1) ( 2) (1) −T −1 ( 2)
dX ⋅ dX = dx ⋅ F F dx = ds ds t ⋅ F F t
= dS (1) dS ( 2) λ (1) λ ( 2) t (1) ⋅ F −T F −1t ( 2)
(1) ( 2) (1) ( 2 ) (1) −1 ( 2)
= dS dS λ λ t ⋅ b t
= dS ( ) dS ( ) λ ( ) λ ( ) t ( ) ⋅ (1 − 2e ) t (
1 2 1 2 1 2)

Variation of Angles
Material Angle
Comparing the two expressions the angle between the two
segments at the material configuration is given by,
cos Θ = λ (1) λ ( 2) t (1) ⋅ (1 − 2e ) t ( 2)

−1 2 −1 2
(
= 1 − 2t ⋅ et (1) (1)
) ( 1 − 2t ⋅ et ( 2) ( 2)
) t ( ) ⋅ (1 − 2e ) t (
1 2)
t (1) ⋅ (1 − 2e ) t (2)
=
1 − 2 t (1) ⋅ e t (1) 1 − 2 t (2) ⋅ e t (2)

Variation of Angles
Spatial Angle
T(1) ⋅ (1 + 2E ) T(2)
cos Θ = T(1) ⋅ T(2) , cos θ =
1 + 2 T(1) ⋅ E T(1) 1 + 2 T(2) ⋅ E T(2)
Material Angle
t (1) ⋅ (1 − 2e ) t (2)
cos Θ = , cos θ = t (1) ⋅ t (2)
1 − 2 t (1) ⋅ e t (1) 1 − 2 t (2) ⋅ e t (2)


Let us consider material segments dX(1) = T(1) dS (1) and
dX(2) = T(2) dS (2) along the X1- and X2-axis, respectively, on the
material configuration.
The angle between the two segments at the spatial configura-
tion is given by,
T(1) ⋅ (1 + 2E ) T(2) 2 E12
cos θ12 = =
1 + 2T(1) ⋅ ET(1) 1 + 2T(2) ⋅ ET(2) 1 + 2 E11 1 + 2 E22
The Green-Lagrange component E12 may be interpreted as,
1 1 ds (1) ds (2)
E12 = 1 + 2 E11 1 + 2 E22 cos θ12 = (1) (2)
cos θ12
2 2 dS dS

Taking into account that the initial angle between the two
segments at the reference configuration is 90º, the angle
increment may be written as,
π 2 E12
∆θ12 := θ12 − = − arcsin
2 1 + 2 E11 1 + 2 E22
The Green-Lagrange component E12 may be interpreted as,
1 1 ds (1) ds (2)
E12 = − 1 + 2 E11 1 + 2 E22 sin ∆θ12 = − (1) (2)
sin ∆θ12
2 2 dS dS


Let us consider spatial segments dx(1) = t (1) ds and dx(2) = t (2) ds
along the x1- and x2-axis, respectively, on the spatial configu-
ration.
The angle between the two segments at the material configura-
tion is given by,
t (1) ⋅ (1 − 2e ) t (2) −2e12
cos Θ12 = =
1 − 2t (1) ⋅ et (1) 1 − 2t (2) ⋅ et (2) 1 − 2e11 1 − 2e22
The Almansi strain component e12 may be interpreted as,
1 1 dS (1) dS (2)
e12 = − 1 − 2e11 1 − 2e22 cos Θ12 = − (1) (2)
cos Θ12
2 2 ds ds

Taking into account that the deformed angle between the two
segments at the spatial configuration is 90º, the angle increment
may be written as,
π 2e12
∆θ12 := − Θ12 = − arcsin
2 1 − 2e11 1 − 2e22
The Almansi strain component e12 may be interpreted as,
1 1 dS (1) dS (2)
e12 = − 1 − 2e11 1 − 2e22 sin ∆θ12 = − (1) (2)
sin ∆θ12
2 2 ds ds

Assignment 3.5
Assignment 3.5
The equations of motion are given by,
x = X, y = Y , z = Z − Xt
Consider two differential segments which at time t=1 are parallel
to the Cartesian axes x and z. Compute which was the angle
formed by those two segments at the reference time t=0.

Strains > Assignments
Assignment 3.6
Assignment 3.6
The sphere of the figure is subjected to a finite uniform deforma-
tion, with uniform deformation gradient. The motion is such
that,
i. The origin O does not moves
ii. Material points A, B and C move to spatial positions A’, B’ and
C’, where AA’=p>0, BB’=CC’=q>0. z
C
1) Obtain the deformation gradient,
R
Green-Lagrange and Almansi strain C’
tensors and the displacement vector O

B’ B
y
field.
A
2) Obtain the relation between p and q A’

x
if the material is incompressible.
Assignment 3.7
The solid of the figure is subjected to a finite linear displace-
ment field, yielding a uniform deformation gradient, such that,
i. The displacements of the material points A, B and C are zero.
ii. The volume of the solid becomes p√2 Dmes the initial one.
iii. The length of the material segment AE becomes p times the
initial one. z
D F
Obtain the deformation gradient, and
the material and spatial descriptions of a
the displacement vector field. E a y

a A C
B
x
Assignment 3.8
The solid of the figure is subjected to a finite linear displace-
ment field, yielding a uniform deformation gradient, such that,
i. The displacements of the material points O, A and B are zero.
ii. The volume of the solid becomes p times the initial one.
iii. The length of the material segment AC becomes p/√2 times
the initial one. z
iv. The deformed angle formed by OA C
and OC is 45º
a
Obtain the deformation gradient, and
B y
the material and spatial descriptions of a
O a
the displacement vector field. A

Assignment 3.9
Assignment 3.9
The solid of the figure is subjected to a deformation, such that,
i. The displacement is linear on X, Y and Z and skew-symmetric
with respect to the plane Y=0, such that U(X,Y,Z)=-U(X,-Y,Z)
for any X, Y, Z.
ii. The volume of the solid becomes remains constant.
iii. The angle given by OA, OB remains constant, equal to 90º.
iv. The length of the material segment OB Z, w
becomes √2 Dmes the initial one. C
v. The z-displacement of point B is positive. a
Obtain the deformation gradient, the Green- O a B
Lagrange strain tensor and the displacement a
vector field. A
D
X, u
Strains > Spatial Velocity Gradient
Spatial Velocity Gradient

time t = 0 time t
t=0 t
v
Q FF
dX P’
dx
P v + dv
dS ds Q’
Ω0 Ωt
X x
X 3 , x3
ê 3
ê 2 X 2 , x2
ê1
X 1 , x1


Spatial Velocity Gradient Tensor
Let us consider the spacial velocity vector field at a spatial point
and time t, given by,
v = v ( x,t )
Differentiating the spatial velocity field, keeping constant the
configuration at time t, using the chain rule, yields,
dv = ( grad v ( x, t ) ) dx = ( ∇ ⊗ v ( x, t ) ) dx := l ( x, t ) dx
where the non-symmetric second-order spatial velocity gradient
tensor, denoted as l ( x,t ), has been introduced as,
l ( x, t ) := ∇ ⊗ v ( x, t ) = grad v ( x, t ) , lab = va ,b

Deformation and Rotation Rate

Deformation and Rotation Rate Tensors
The spatial velocity gradient tensor can be split into symmetric
and skew-symmetric parts, yielding,
l ( x, t ) = symm l ( x, t )  + skew l ( x, t )  := d ( x, t ) + w ( x, t )
where the symmetric spatial deformation rate tensor, denoted
as d ( x,t ) , has been defined as,
1 1
2
T
2
(
d := symm [ l ] = ( l + l ) = grad v + ( grad v )
T
)
and the skew-symmetric spatial rotation rate tensor, denoted as
w ( x,t ) , has been defined as,
1 1
2
T
2
(
w := skew [ l ] = ( l − l ) = grad v − ( grad v )
T
)
Deformation Rate
Deformation Rate Tensor
The symmetric spatial deformation rate tensor, denoted as d ( x, t ) ,
may be viewed as a measure of the rate of deformation given by,
d d d
dt
( ds − dS ) = ( ds ) = ( dx ⋅ dx ) = dv ⋅ dx + dx ⋅ dv
2 2
dt
2
dt
Using the expression dv = l dx yields,
d d 1
dt
( ds 2
− dS 2
) =
dt
( ds 2
) = 2 dx ⋅
2
( l + l T
) dx = 2 dx ⋅ d dx

Deformation Rate
may be related to the material time derivative of the Green-
Lagrange strain tensor as,
d d 
dt
( ds − dS ) = ( 2dX ⋅ EdX ) = 2dX ⋅ EdX 
2 2
dt



d
( ds 2
− dS 2
) = 2 dx ⋅ d dx = 2 dX ⋅ F T
dFdX 
dt 
= FT dF, d = F −T EF
E −1

Deformation Rate
may be related to the material time derivative of the Almansi
strain tensor as,
d d 
dt
( ds − dS ) = ( 2dx ⋅ edx )
2 2
dt 

= 2 ( dx ⋅ e dx + dv ⋅ edx + dx ⋅ edv ) 

= 2dx ⋅ ( e + l e + el ) dx
T

d 
( ds 2
− dS 2
) = 2 dx ⋅ d dx 
dt 
d = e + lT e + el, e = d − lT e − el
Rotation Rate
Rotation Rate Tensor
The skew-symmetric spatial rotation rate tensor, denoted as w ( x, t ) ,
satisfies the following expressions,
w dx = ω × dx ∀dx, wab dxb = ε abcωb dxc
1 1 1 ∂vc 1
ω = curl v = ∇ × v, ωa = ε abc = ε abc vc ,b
2 2 2 ∂xb 2
where ω is the axial (or dual) rotation rate vector.

Rotation Rate
Rotation Rate Tensor
The components of the skew-symmetric spatial rotation rate
tensor w and the components of the axial (or dual) rotation rate
vector ω , are such that,
 0 w12 w13   0 −ω3 ω2 

−w w23  =  ω3 −ω1 
[ ]  12
w = 0 0
 − w13 − w23 0   −ω2 ω1 0 
ω1   − w23 
ω  =  w 
[ ]  2   13 
ω =
ω3   − w12 

Spatial Velocity
Giving the spatial description of an arbitraryGradient
property, Tensor
dv = l dx, dva = lab dxb

l := grad v = ∇ ⊗ v, lab := va ,b
Deformation and Rotation Rate Tensors
l = sym [ l ] + skew [ l ] := d + w
l + l ) = (∇ ⊗ v + (∇ ⊗ v )
),
1 1 1
d :=
2
( T
2
T
d ab :=
2
( v a ,b + vb ,a )
w := ( l − l ) = ( ∇ ⊗ v − ( ∇ ⊗ v ) ) ,
1 1 1
2 2
T T
wab :=
2
( v a ,b − vb ,a )

Assignment 3.10
Consider two different motions with velocity vector fields given
by,
T
 v  =  V
I I
( X , Y , Z ) = [ Z X Z]
T
 v II  =  v II ( x, y, z )  = [ z x z]
Assuming that the reference time is t=0, obtain for each one of
the motions,
1) The motion equation and the deformation gradient
2) The Green-Lagrange and the Almansi strain tensors
3) The deformation rate tensor

Strains > Material Time Derivatives
Material Time Derivatives

Material Time Derivative of the Deformation Gradient
The material time derivative of the deformation gradient reads,
d  ∂ϕa  ∂  ∂ϕa  ∂va ∂ϕb

FaA =  =   = = lab FbA
dt  ∂X A  ∂X A  ∂t  ∂xb ∂X A
F = l F, FaA = lab FbA

The material time derivative of the inverse deformation gradient
may be computed as follows,
d d −1 d −1
dt
( FF ) = FF + F ( F ) = l + F ( F ) = 0
−1 −1
dt dt
d −1
dt
( F ) = − F −1
l

Material Time Derivative of the Green-Lagrange Strain
The material time derivative of the Green-Lagrange strain tensor
reads,
E = 1 C
= 1 d ( FT F )
2 2 dt
1 T T 1 T
= ( F F + F F ) = F ( l + lT ) F
2 2
= FT dF


Material Time Derivative of the Green-Lagrange Strain
The material time derivative of the Green-Lagrange strain tensor
may be viewed as the pull-back of the (covariant) spatial
deformation rate tensor,
= ϕ * ( d ) = FT dF
E
The spatial deformation rate tensor may be viewed as the push-
forward of the (covariant) material time derivative of the Green-
Lagrange strain tensor
d = ϕ* ( E
) = F −T EF
−1


Material Time Derivative of the Almansi Strain
The material time derivative of the Almansi strain tensor reads,
1 d −1 1 d −T −1
e = −
2 dt
( b )=−
2 dt
( F F )
1  d −T −1 −T d −1 
= −  (F ) F + F ( F )
2  dt dt 
1 T −T −1
= ( l F F + F −T F −1lT )
2
1 T −1 −1
= (l b + b l)
2


Material Time Derivative of the Jacobian
The material time derivative of the Jacobian reads,
d d F dF d F d F

J= F= : = :F = : ( lF )
dt dF dt dF dF
dF −T dF
=JF , = J FaA−T
dF dFaA
d d F dFaA
J= F= = J FaA−T lab FbA = J FbA FAa−1lab = J laa = Jva ,a
dt dFaA dt
J = J div v

Material Time Derivative of the Differential of Volume
The material time derivative of the differential of volume reads,
d d
( dv ) = ( J dV ) = J dV = J div v dV = div v dv
dt dt
Material Time Derivative of the Differential of Area

The material time derivative of the differential of area reads,
d d
( da ) = ( J F −T dA ) = ( ( div v ) 1 − lT ) J F −T dA
dt dt
= ( ( div v ) 1 − lT ) da


Giving the spatial description of an Gradient
Deformation arbitrary property, the material time derivative of the property can be written as,
F = l F
Strain Tensors
= FT dF
E
1 T −1 −1
e = ( l b + b l ) = d − l T e − e l
2
Jacobian
J = J div v

Continuum Mechanics
Chapter 4
Kinematics: Infinitesimal Strains
Infinitesimal Strains > Hypothesis
Infinitesimal Strains
Chapter 4 · Infinitesimal Strains
1. Hypothesis
2. Displacement gradient tensor
3. Infinitesimal strain tensor
4. Variation of volume
5. Polar decomposition
6. Stretches
7. Variation of angles
8. Matrix notation
9. Compatibility equations

Infinitesimal Strains
time t = 0 time t
t=0 t
Q FF
dX P’
dx
P
dS ds Q’
Ω0 Ωt
X x
Inverse Deformation Map X 3 , x3
Deformation Map
X = ϕ −1 ( x,t ) ê 3 x = ϕ ( X,t )
ê1
dX = F −1
( x, t ) dx X 1 , x1 dx = F ( X, t ) dX

Hypothesis
H1. Small Displacements
We assume that the displacements are small, such that,
We do not make any difference between spatial configura-
tion and material configuration,
Ωt Ω 0
We do not make any difference between spatial points and
material points, or spatial coordinates and material
coordinates,
x = ϕ ( X, t ) X, [ x ] [ X]

Hypothesis
We do not make any difference between spatial description
and material description,
γ = γ ( x, t ) = γ (ϕ ( X, t ) , t ) = Γ ( X, t ) γ ( X, t )
γ = Γ ( X, t ) = Γ (ϕ −1 ( x, t ) , t ) = γ ( x, t ) Γ ( x, t )
We do not make any difference between spatial differential
operators and material differential operators,
∇ ∇, ∇ 2 ∇ 2 , ∇ ⊗ ∇ ∇ ⊗ ∇
grad [] GRAD [] , div [] DIV [] , curl [] CURL []
Hypothesis
We do not make any difference between spatial time
derivative and material time derivative,
∂γ ∂γ ( x, t ) ∂γ ( X, t ) d γ
= = γ
∂t ∂t ∂t dt
d γ ∂Γ ( X, t ) ∂Γ ( x, t ) ∂γ
γ = =
dt ∂t ∂t ∂t

Hypothesis
H2. Small Displacements Gradient
We assume that the displacements gradient are small, such
that,
The components of the displacement gradient satisfy,
∂U a ( X, t )
J aA = 1
∂X A
We consider a linear theory, i.e. any non-linear function of the
displacements gradient is linearized,
f ( J ) = f ( J ) J=0 + ∂ J f ( J ) J=0 : J + 0 ( J 2 )
f ( J ) J=0 + ∂ J f ( J ) J=0 : J

Infinitesimal Strains > Displacement Gradient Tensor
Displacement Gradient Tensor

Displacements Gradient Tensor
As we do not make any difference between material and spatial
descriptions and we do not make any difference between
material and spatial gradient, then there is no difference
between the material and spatial displacement gradient,
yielding a single displacement gradient tensor such that,
J := ∇ ⊗ U ( X, t ) = GRAD U ( X, t ) 

j := ∇ ⊗ u ( x, t ) = grad u ( x, t ) 
j J

⇒ ∇ ⊗ u ( x, t ) ∇ ⊗ U ( X, t )

grad u ( x, t ) GRAD U ( X, t )
Infinitesimal Strains > Infinitesimal Strain Tensor
Infinitesimal Strain Tensor

The Green-Lagrange strain tensor is a non-linear function of the
material displacement gradient tensor,
1 T 1 1
2 2
T
(
E := ( F F − 1) = (1 + J ) (1 + J ) − 1 = ( J + JT + JT J )
2
)
The linearization of the Green-Lagrange strain tensor yields,
1 1
E := ( J + J + J J ) ( J + JT ) := ε
T T
2 2
The infinitesimal strain tensor, denoted as ε , may be defined as
the linearized Green-Lagrange strain tensor, yielding,
1 1
ε := ( J + J ) , ε ab := ( ua ,b + ub,a )
T
2 2
Infinitesimal Strains > Infinitesimal Strain Tensor

The Almansi strain tensor is a non-linear function of the spatial
displacement gradient tensor,
1 1 1
2
−T −1
2
(
e := (1 − F F ) = 1 − (1 − j) (1 − j) = ( j + jT − jT j)
T
2
)
The infinitesimal strain tensor, denoted as ε , may be defined as
the linearized Almansi strain tensor, and taking into account that
we do not make any difference between material and spatial
displacement gradient, yields,
1 1 1
ε := ( j + j ) ( J + J ) , ε ab := ( ua ,b + ub,a )
T T
2 2 2

Infinitesimal Strains > Variation of Volume
Variation of Volume
Variation of Volume
The relation between the differential of volume at the spatial
and material configurations is given by,
dv = J dV
where the Jacobian is a non-linear function of the components
of the displacement gradient tensor given by,
1 + u1,1 u1,2 u1,3 
 
J = det F = det (1 + J ) = det  u2,1 1 + u2,2 u2,3 
 u3,1 u3,2 1 + u3,3 

Infinitesimal Strains > Variation of Volume
Variation of Volume
Variation of Volume
The linearization of the Jacobian yields,
J 1 + u1,1 + u2,2 + u3,3 = 1 + div u = 1 + tr ε
Then, within an infinitesimal strains framework, the relation
between the spatial and material differential of volume takes the
form,
dv (1 + div u ) dV = (1 + tr ε ) dV
Note that now, within the infinitesimal strains framework, the
incompressibility condition reads,
J = 1 ⇒ div u = tr ε = 0
Infinitesimal Strains > Polar Decomposition
Polar Decomposition
Polar Decomposition
The polar decomposition of the deformation gradient tensor F ,
reads,
where U is the right (or material) stretch tensor, v is the left (or
spatial) stretch tensor and R is the rotation tensor, such that,
12
U = (F F) T
= C1 2 , U = UT , x ⋅ Ux > 0 ∀x ≠ 0, det U = J
T 12
v = ( FF ) = b1 2 , v = vT , x ⋅ vx > 0 ∀x ≠ 0, det v = J
R = FU −1 = v −1F, R −1 = RT , det R = 1

Polar Decomposition
Linearized Right Stretch Tensor
The right stretch tensor is a non-linear function of the material
displacement gradient tensor given by,
12 12 12
= (F F) = ( 1 + J ) ( 1 + J )  = (1 + J + J + J J )
12 T T T T
U=C
 
The linearization of the right Cauchy-Green tensor and the right
stretch tensor yields,
T
C = F F = (1 + J ) (1 + J ) = 1 + J + JT + JT J
T
1 + J + JT = 1 + 2ε
T 12 1
U=C 12
(1 + J + J ) 1 + ( J + JT ) = 1 + ε
2
Polar Decomposition
Linearized Inverse Right Stretch Tensor
The linearization of the inverse right Cauchy-Green tensor and
the right stretch tensor yields,
T −1
C (1 + J + J
−1
) 1 − ( J + JT ) = 1 − 2ε
−1 2 1
U =C−1 −1 2
(1 + J + J T
) 1 − ( J + JT ) = 1 − ε
2

Polar Decomposition
Linearized Inverse Left Stretch Tensor
The inverse left stretch tensor is a non-linear function of the
spatial displacement gradient tensor given by,
12
−1 1 2 12
= (F F ) = ( 1 − j) ( 1 − j)  = ( 1 − j − j + j j)
−1 −1 2 −T T T T
v =b
 
The linearization of the inverse left Cauchy-Green tensor and
inverse left stretch tensor yields,
T
b = F F = (1 − j) (1 − j) = 1 − j − jT + jT j
−1 −T −1
≅ 1 − j − jT 1 − 2ε
T 12 1
v =b −1 −1 2
≅ (1 − j − j ) ≅ 1 − ( j + jT ) 1 − ε
2
Polar Decomposition
Linearized Left Stretch Tensor
The linearization of the left Cauchy-Green tensor and left stretch
tensor yields,
−1
b (1 − j − j T
) 1 + ( j + jT ) 1 + 2ε
−1 2 1
v=b 12
(1 − j − j T
) 1 + ( j + jT ) 1 + ε
2

Polar Decomposition
Linearized Rotation Tensor
The rotation tensor may be written as a non-linear function of
the material displacement gradient tensor given by,
−1 2
R = FU = FC −1 −1 2
= F (F F) T −1 2
( T
= (1 + J ) (1 + J ) (1 + J ) )
The linearization of the inverse rotation tensor yields,
 1 T  1
R = FU (1 + J )  1 − ( J + J )  1 + ( J − JT ) := 1 + Ω
−1
 2  2
where the skew-symmetric infinitesimal rotation tensor Ω has
been defined as,
1
Ω := ( J − JT )
2
Polar Decomposition
Geometrical interpretation of an infinitesimal rotation
dX Ω dX
dX
R dX (1 + Ω ) dX

Polar Decomposition
The skew-symmetric infinitesimal rotation tensor satisfies the
following expressions,
Ω = −ΩT , Ωab = −ΩTab = Ωba
We may introduce an axial (or dual) infinitesimal rotation vector

such that,
Ω dX = θ × dX ∀dX, Ωab dX b = ε abcθb dX c
1 1 1 ∂uc 1
θ = curl u = ∇ × u, θ a = ε abc = ε abc uc ,b
2 2 2 ∂X b 2

Polar Decomposition
The matrix of components of the skew-symmetric infinitesimal
rotation tensor and the vector of components of the axial (or
dual) infinitesimal rotation vector satisfy,
 0 Ω12 Ω13   0 −θ3 θ2 

 −Ω Ω23  =  θ3 −θ1 
[ ]  12
Ω = 0 0
 −Ω13 −Ω23 0   −θ 2 θ1 0 
θ1   −Ω23 
[θ ] = θ2  =  Ω13 
θ3   −Ω12 
Polar Decomposition
Linearized Polar Decomposition
The linearized, right and left, polar decomposition takes the
form,
F = RU (1 + Ω )(1 + ε ) 1 + ε + Ω
F = vR (1 + ε )(1 + Ω ) 1 + ε + Ω
yielding,
dx = FdX (1 + ε + Ω ) dX
The linearized, either right or left, polar decomposition may be
interpreted as the sum of an infinitesimal deformation (charate-
rized by the infinitesimal strain tensor) and an infinitesimal
rotation (charaterized by the infinitesimal rotation tensor)

Polar Decomposition
Linearized Polar Decomposition
Geometrical interpretation of the linearized polar decomposition
Ω dX ε dX
dX
dX
dx = FdX (1 + ε + Ω ) dX

Strains > Linearized Tensors
Linearized Tensors
Infinitesimal Strain
Giving the spatial description of an arbitrary property,andtheRotation
material time derivativeTensors
of the property can be written as,
1 1 1 1
E e ε := ( J + J ) ( j + j ) , Ω := ( J − J ) ( j − jT )
T T T
2 2 2 2
Linearized Cauchy-Green and Stretch Tensors
C b 1 + 2ε, U v 1 + ε
C−1 b −1 1 − 2ε, U −1 v −1 1 − ε
R 1+ Ω
Infinitesimal Strains > Stretches
Stretches
Linearized Stretch
The stretch at a material point along a direction given by the unit
vector T reads,
ds
λ= = 1+ 2 T ⋅ E T
dS
Using a linear Taylor series expansion, i.e.,
λ ( x ) = 1 + 2x λ ( 0) + λ′ ( 0) x = 1 + x
The linearized stretch takes the form,
λ = 1+ 2 T ⋅ E T 1+ T ⋅ E T 1+ T ⋅ε T

Stretches
Linearized Stretch
The stretch at a spatial point along a direction given by the unit
vector t reads,
ds 1
λ= =
dS 1− 2 t ⋅e t
Using a linear Taylor series expansion, i.e.,
1
λ ( x) = λ ( 0) + λ′ ( 0) x = 1 + x
1− 2x
The linearized stretch takes the form,
1
λ= 1+ t ⋅e t 1+ T ⋅ε T
1− 2 t ⋅e t

Stretches
Stretches along the Cartesian Axes
Taking unit vectors along the Cartesian axes,
T T T
[Tx ] = [1, 0, 0] , Ty  = [0, 1, 0] , [Tz ] = [0, 0, 1]
The stretches along the Cartesian axes take the form,

λx 1 + Tx ⋅ ε Tx = 1 + ε xx , ε xx λx − 1 = ε x
λy 1 + Ty ⋅ ε Ty = 1 + ε yy , ε yy λy − 1 = ε y
λz 1 + Tz ⋅ ε Tz = 1 + ε zz , ε zz λz − 1 = ε z

Infinitesimal Strains > Variation of Angles
Variation of Angles
Linearized Variation of Angles
The angle formed by two material segments along unit vectors
given by T(1) and T(2) in the material configuration, in the
spatial configuration is given by,
T(1) ⋅ (1 + 2E ) T(2)
cos θ =
1 + 2 T(1) ⋅ ET(1) 1 + 2 T(2) ⋅ ET(2)
The linearized expression of the angle takes the form,
cos θ T(1) ⋅ (1 + 2E ) T(2) cos Θ + 2 T(1) ⋅ ε T(2)
cos θ = cos ( Θ + ∆θ ) = cos Θ cos ∆θ − sin Θ sin ∆θ cos Θ − ∆θ sin Θ
2 T(1) ⋅ ε T(2)
∆θ −
sin Θ
Variation of Angles
Linearized Variation of Angles
The angle formed by two spatial segments along unit vectors
(1) (2)
given by t and t in the spatial configuration, in the
material configuration is given by,
t (1) ⋅ (1 − 2e ) t (2)
cos Θ =
1 − 2 t (1) ⋅ e t (1) 1 − 2 t (2) ⋅ e t (2)
The linearized expression for the angle takes the form,
cos Θ t (1) ⋅ (1 − 2e ) t (2) cos θ − 2 t (1) ⋅ ε t (2)
cos θ = cos ( Θ + ∆θ ) = cos Θ cos ∆θ − sin Θ sin ∆θ cos Θ − ∆θ sin Θ
2 t (1) ⋅ ε t (2) 2 T(1) ⋅ ε T(2)
∆θ − −
sin Θ sin Θ
Variation of Angles
Linearized Variation of Cartesian Angles
Taking unit vectors along the Cartesian axes,
T T T
[Tx ] = [1, 0, 0] , Ty  = [ 0, 1, 0] , [Tz ] = [0, 0, 1]
The increment of angles between segments oriented along the
Cartesian axes take the form,
2 Tx ⋅ ε Ty 1
∆θ xy − = −2ε xy , ε xy − ∆θ xy
sin Θ xy 2
2 Tx ⋅ ε Tz 1
∆θ xz − = −2ε xz , ε xz − ∆θ xz
sin Θ xz 2
2 Ty ⋅ ε Tz 1
∆θ yz − = −2ε yz , ε yz − ∆θ yz
sin Θ yz 2
Infinitesimal Strains > Matrix Notation
Matrix Notation
Tensorial and Engineering Matrix Notation
Using tensorial notation, the matrix of components of the sym-
metric infinitesimal strain tensor takes the form,
ε xx ε xy ε xz  ε11 ε12 ε13 
   
[ ]  xy yy yz   12 22 23 
ε = ε ε ε = ε ε ε
ε xz ε yz ε zz  ε13 ε 23 ε 33 
 
Using engineering notation, the matrix of components of the
symmetric infinitesimal strain tensor takes the form,
 εx 1
2
γ xy 1
2
γ xz 
1 
[ε ] =  2 γ xy εy 1
2
γ yz 
 1 γ xz 1
γ ε z 
2 2 yz
Infinitesimal Strains > Matrix Notation
Matrix Notation
Tensorial and Engineering Matrix Notation
Using tensorial notation, the associated vector of components of
the symmetric infinitesimal strain tensor takes the form,
T
[ε ] = ε xx , ε yy , ε zz , 2ε xy , 2ε xz , 2ε yz 
T
= [ε11 , ε 22 , ε 33 , 2ε12 , 2ε13 , 2ε 23 ]
Using engineering notation, the vector of components of the
symmetric infinitesimal strain tensor takes the form,
T
[ε ] = ε x , ε y , ε z , γ xy , γ xz , γ yz 

Infinitesimal Strains > Material Time Derivatives

Material Time Derivative of the Infinitesimal Strain
Within the infinitesimal strains framework, the deformation rate
tensor, the material time derivative of the Green-Lagrange strain
tensor and the material time derivative of the Almansi strain
tensor are approximated by the time derivative of the infinitesi-
mal strain tensor, yielding
e ε
dE

Infinitesimal Strains > Assignments
Assignment 4.1
Assignment 4.1
The solid of the figure is subjected to a uniform strain state
through a uniaxial traction/compression test, such that,
ux (0, y, z ) = 0, ux ( L, y, z ) = δ ∀ y, z
u y ( x, y, z ) = uz ( x, y, z ) = 0 ∀ x, y, z
1) Obtain the motion equations and z
displacement field, indicating the
δ
range of values for δ L.
2) Plot the x-component of the
O
Green-Lagrange, Almansi and x
infinitesimal strain tensors vs the

L
normalized displacement δ L. y

Infinitesimal Strains > Assignments
Assignment 4.2
The solid of the figure is subjected to a uniform strain state
such that: (1) Point A does not moves, point B remains on the x-
axis and point E remains on z-axis; (2) The volume remains
constant; (3) The angle Θxy remains constant; (4) The angle Θ yz
increases in r radians; (5) The length of the segment AF becomes
1+p times the initial one; (6) The length of the segment AC
becomes 1-q times the initial one. z
E F
Obtain the infinitesimal strain tensor, the

infinitesimal rotation tensor and the a
displacements in terms of p, q and r. A

Θyz
D y
Θxy
B C
a
a
October 2, 2013 Carlos Agelet de Saracibar x 45
Infinitesimal Strains > Compatibility Equations
Compatibility Equations
Given an arbitrary smooth displacement vector field u (with the
required continuity degree), it is always possible to get the
associated symmetrical infinitesimal strain tensor defined as,
1 1
2
( T
2
)
ε := ∇ ⊗ u + ( ∇ ⊗ u ) , ε ab = ( ua ,b + ub,a )
The opposite statement is not true.

Given an arbitrary symmetrical infinitesimal strain tensor ε , i.e.
given six arbitrary functions as the matrix components of a
symmetric second-order tensor, it is not always possible to get
the associated displacement vector field such that,
1 1
2
( T
2
)
ε := ∇ ⊗ u + ( ∇ ⊗ u ) , ε ab = ( ua ,b + ub,a )
This gives a system of 6 PDE with 3 unknowns and, hence, the

system may not be integrable and may not have a solution.
In order to get an integrable system of PDE and guarantee the
existence of the displacement vector field, the infinitesimal strain
tensor has to satisfy the compatibility equations.
The compatibility equations for the infinitesimal strain tensor
may be written as,
T T
S := ∇ × ( ∇ × ε ) T
= ∇ × (∇ × ε ) = 0
S := e f × ( ε cdeε bd ,ceb ⊗ ee ), f = ε fbaε cdeε bd ,cf ea ⊗ ee
def ∂ 2ε yy ∂ 2ε zz ∂ 2ε yz ∂ 2ε zz ∂  ∂ε yz ∂ε xz ∂ε xy 
def
S xx = + −2 =0 S xy = − +  + − =0
∂z 2
∂y 2
∂y∂z ∂x∂y ∂z  ∂x ∂y ∂z 
def
∂ 2ε zz ∂ 2ε xx ∂ 2ε xz def ∂ 2ε yy ∂  ∂ε yz ∂ε xz ∂ε xy 
S yy = 2
+ 2
−2 = 0 S xz = − +  − + =0
∂x ∂z ∂x∂z ∂x∂z ∂y  ∂x ∂y ∂z 
2 2
def ∂ 2ε
xx
∂ ε yy ∂ ε xy def ∂ 2
ε xx ∂  ∂ε yz ∂ε xz ∂ε xy 
S zz = + − 2 = 0 S yz = − + − + + =0
∂y 2 ∂x 2 ∂x∂y ∂y∂z ∂x ∂x ∂y ∂z
 
The compatibility tensor S satisfies the following equation,
(
div S = ∇ ⋅ S := ∇ ⋅ ∇ × ( ∇ × ε )
T
)=0
 ∂S xx ∂S xy ∂S xz
 + + =0
 ∂x ∂y ∂z
 ∂S xy ∂S yy ∂S yz
 + + =0
 ∂x ∂y ∂z
 ∂S ∂S yz ∂S zz
 xz
+ + =0
 ∂x ∂y ∂z
Continuum Mechanics
Chapter 5
Stresses
Stresses > Forces
Forces
Chapter 5 · Stresses
1. Forces
2. Cauchy’s stress theorems
3. Stress tensors

Stresses > Forces
Forces
Forces
We consider two types of forces that may act on a continuum
body: body (or mass or volume or internal) forces and surface
forces.
Body (or mass or volume or internal) forces: Forces acting in
the volume of a continuum medium. Typical examples of body
forces are the gravity forces or the electromagnetic forces.
Surface forces: Forces acting on the surface of a continuum
medium due to the interaction with other bodies or the
environtment. Typical examples of surface forces are the
contact forces or applied loads.

Stresses > Forces
Body Forces
Body Forces
Body (or mass or volume or internal) forces may be
characterized by the body forces per unit of mass vector,
denoted as b .
Spatial and material descriptions of the body forces per unit of
mass vector, denoted as b ( x, t ) and B ( X, t ) , respectively, take
the form,
b = b ( x, t ) = b (ϕ ( X, t ) , t ) = B ( X, t )
( )
b = B ( X, t ) = B ϕ −1 ( x, t ) , t = b ( x, t )

Stresses > Forces
Body Forces
Differential Body Forces
The differential body (or mass or volume or internal) force
acting in a differential of volume dv in the spatial configuration
takes the form,
dfv = ρ ( x, t ) b ( x, t ) dv
The differential body (or mass or volume or internal) force

acting in a differential of volume dV in the material configura-
tion takes the form,
dfv = ρ0 ( X ) B ( X, t ) dV

Stresses > Forces
Body Forces
Total Body Forces
The total body (or mass or volume or internal) forces acting in a
spatial volume v of a continuum body, at a time t, may be
written as,
Fv = ∫ ρ ( x, t ) b ( x, t ) dv
v
The total body (or mass or volume or internal) forces acting in a

material volume V of a continuum body, at a time t, may be
written as,
Fv = ∫V
ρ0 ( X ) B ( X, t ) dV

Stresses > Forces
Surface Forces
Differential Surface Forces
The differential surface force acting on a differential of area ds
on the spatial configuration takes the form,
dfs = t ( x, t ) ds
where the Cauchy (or true) traction vector, denoted as t ( x, t ),
represents the spatial description of the surface force per unit of
spatial surface.
The differential surface force acting on a differential of area dS
on the material configuration takes the form,
df s = T ( X, t ) dS
where the first Piola-Kirchhoff (or nominal) traction vector,
denoted as T ( X, t ), represents the material description of the
surface force per unit of material surface.
Stresses > Forces
Surface Forces
Total Surface Forces
The total surface forces acting on a spatial surface ∂v of a
continuum body, at a time t, may be written in terms of the
Cauchy (or true) traction vector as,
F∂v = ∫ t ( x, t ) ds
∂v
The total surface forces acting on a material surface ∂V of a

continuum body, at a time t, may be written in terms of the first
Piola-Kirchhoff (or nominal) traction vector as,
F∂V = ∫ ∂V
T ( X, t ) dS

Stresses > Forces
Forces
Body Forces
Fv = ∫ ρ ( x, t ) b ( x, t ) dv = ∫
v V
ρ 0 ( X ) B ( X, t ) dV
Surface Forces
F∂v = ∫ t ( x, t ) ds = ∫
∂v ∂V
T ( X, t ) dS

Stresses > Cauchy Stress Theorems
The Traction Vector Picture

Cauchy Stress Theorems

First Cauchy Stress Theorem
The Cauchy (or true) traction vector at a spatial point x, at a
given time t, on a spatial surface with unit outward normal n at
the spatial point x, is only a function of the spatial point, the
time t and the unit outward normal at the spatial point x at the
time t,
t = t ( x, t , n ) , ta = ta ( x, t , n )


First Cauchy Stress Theorem
The first Piola-Kirchhoff (or nominal) traction vector at a
material point X , at a given time t, on a material surface with
unit outward normal N at the material point X , is only a
function of the material point, the time t and the unit outward
normal at the material point X at the time t,
T = T ( X, t , N ) , Ta = Ta ( X, t , N )


Second Cauchy Stress Theorem
The Cauchy (or true) traction vector at a spatial point x, at a
given time t, on a spatial surface with unit outward normal n at
the spatial point x, is a linear function of the unit outward
normal at the spatial point x at the time t, satisfying the so
called action-reaction principle,
t = t ( x, t , n ) = −t ( x, t , −n ) , ta = ta ( x, t , n ) = −ta ( x, t , −n )


The first Piola-Kirchhoff (or nominal) traction vector at a
material point X , at a given time t, on a material surface with
unit outward normal N at the material point X , is a linear
function of the unit outward normal at the material point X at
the time t, satisfying the so called action-reaction principle,
T = T ( X, t , N ) = −T ( X, t , −N ) , Ta = Ta ( X, t , N ) = −Ta ( X, t , −N )



Stresses > Stress Tensors
Stress Tensors
Cauchy Stress Tensor
The Cauchy (or true) stress tensor, denoted as σ , is a symmetric
spatial second-order tensor, such that,
t ( x, t , n ) = σ ( x, t ) n, ta ( x, t , n ) = σ ab ( x, t ) nb
First Piola-Kirchhoff Stress Tensor

The first Piola-Kirchhoff (or nominal) stress tensor, denoted as P,
is a non-symmetric two-point second-order tensor, such that,
T ( X, t , N ) = P ( X, t ) N, Ta ( X, t , N ) = PaA ( X, t ) N A

Stress Tensors
time t = 0 time t
n
N
T t
df df
dS ds
df = TdS = t ds
T = df dS t = df ds
T = T ( X, t , N ) = P ( X, t ) N t = t ( x, t , n ) = σ ( x, t ) n
Piola Transformation
The Cauchy (or true) traction vector and the first Piola-Kirchhoff
(or nominal) traction vector are related through the expression,
df = t ds = T dS
Introducing the Cauchy (or true) stress tensor and the first Piola-
Kirchhoff (or nominal) stress tensor, yields,
df = σ n ds = PN dS
and using Nanson’s formula, given by,
nds = J F −T N dS
yields the so called Piola transformation, given by,
P = J σ F −T , PaA = J σ ab FbA−T

Piola Identity
Using the diverge theorem, the following useful identity holds,
∫∂Ω
nds = ∫ 1nds = ∫ div 1dv = 0
∂Ω Ω
Using Nanson’s formula and the divergence theorem yields,
∫ nds = ∫ JF −T NdS = ∫ DIV ( JF −T ) dV = 0

∂Ω ∂Ω0 Ω0
And in local form, we obtain the so called Piola identity given by,
DIV ( JF −T ) = 0, ( aA ) = 0
JF −T
,A

Using the Piola transformation and the Piola identity given by,
P = J σ F −T , DIV ( JF −T ) = 0,
the following expression holds,
DIV P = DIV ( J σ F −T ) = JF −T DIV σ + σ DIV ( JF −T ) = J div σ

yielding the useful expression,
DIV P = J div σ , PaA, A = J σ ab ,b

Symmetry Restriction
The Piola transformation yields to the following relations
between the Cauchy (or true) stress tensor and the first Piola-
Kirchhoff (or nominal) stress tensor,
P = Jσ F −T , σ = J −1PFT
The symmetry of the Cauchy stress tensor, i.e.,
σ = σ T , σ ab = σ ab
T
= σ ba
yields the following symmetry restriction on the first Piola-
Kirchhoff stress tensor,
PFT = FPT , PaA FAb
T T
= PaA FbA = FaA PAb = FaA PbA

Stress Tensors
Cauchy Stress
Giving the spatial description Tensor
of an arbitrary property, the material time derivative of the property can be written as,
σ = σ T , σ ab = σ ab
T
= σ ba
σ = J −1PFT , σ ab = J −1PaA FAbT = J −1PaA FbA
div σ = J −1 DIV P, σ ab,b = J −1 PaA, A
PFT = FPT , PaA FAb

T T
= PaA FbA = FaA PAb = FaA PbA
P = J σ F −T , PaA = J σ ab FbA−T = J σ ab FAb−1
DIV P = J div σ , PaA, A = J σ ab ,b

Stress Tensor Components

Cauchy Stress Tensor Components
The Cauchy (or true) traction vector at a spatial point, along the
Cartesian planes, i.e. on planes with unit vectors along the
Cartesian axes at the spatial configuration, read,
t1 = σ e1 = σ 11e1 + σ 21e 2 + σ 31e3
t 2 = σ e 2 = σ 12e1 + σ 22e 2 + σ 32e3
t 3 = σ e3 = σ 13e1 + σ 23e 2 + σ 33e3
Note that the ab-component of the Cauchy stress tensor may be
computed as,
σ ab = ea ⋅ t b = ea ⋅ σ eb




Using index notation, the matrix of Cartesian components of the
symmetric Cauchy (or true) stress tensor takes the form,
σ 11 σ 12 σ 13  σ 11 σ 12 σ 13 
σ  = σ 
[ ]  21 22 23   12 22 23 
σ = σ σ σ σ
σ 31 σ 32 σ 33  σ 13 σ 23 σ 33 
Being a symmetric second-order tensor, we may collect the six
components into a vector of components, such that,
T
[σ ] = [σ11 σ 22 σ 33 σ 12 σ 13 σ 23 ]


Using engineering notation, the matrix of Cartesian components
of the symmetric Cauchy (or true) stress tensor takes the form,
σ x τ xy τ xz  σ x τ xy τ xz 
   
[σ ] = τ yx σ y τ yz  = τ xy σ y τ yz 
τ zz τ zy σ z  τ xz τ yz σ z 
   
Being a symmetric second-order tensor, we may collect the six
components into a vector of components, such that,
T
[σ ] = σ x σ y σ z τ xy τ xz τ yz 

Other Stress Tensors

Kirchhoff Stress Tensor
The Kirchhoff stress tensor, denoted as τ , is a symmetric spatial
second-order tensor and it may be defined in terms of the
Cauchy stress tensor as,
τ = J σ, τ ab = J σ ab
Second Piola-Kirchhoff Stress Tensor

The second Piola-Kirchhoff stress tensor, denoted as S , is a
symmetric material second-order tensor and it may be defined in
terms of the Kirchhoff stress tensor as,
S = F −1τ F −T , S AB = FAa−1τ ab FbB−T = FAa−1τ ab FBb−1

Push-forward / Pull-back operations

Push-forward / Pull-back Operations
The Kirchhoff stress tensor may be viewed as the push-forward
of the (contravariant) second Piola-Kirchhoff stress tensor,
satisfying,
τ = ϕ* ( S ) = FSFT
The second Piola-Kirchhoff stress tensor may be viewed as the

pull-back of the (contravariant) Kirchhoff stress tensor, satisfying,
S = ϕ * (τ ) = F −1τ F −T


time t = 0 time t
t
FF
S τ
Ω0 Ωt
Pull-back Maps X 3 , x3 Push-forward Maps

ê 3
S = F −1τ F −T τ = FSFT
ê 2 X 2 , x2
ê1
X 1 , x1

Stress Tensors
Cauchy Stress
Giving the spatial description Tensor
σ = J −1τ = J −1PFT = J −1FSFT

Kirchhoff Stress Tensor
τ = Jσ = PFT = FSFT
P = J σ F −T = τ F −T = FS
Second Piola-Kirchhoff Stress Tensor
S = JF −1σ F −T = F −1τ F −T = F −1P

Continuum Mechanics
Chapter 6
Balance Laws
Balance Laws > Contents
Contents
Chapter 6 · Balance Laws
1. Introduction
2. Conservation of mass
3. Reynolds transport theorem
4. Linear momentum balance
5. Angular momentum balance
6. Mechanical energy balance
7. Assignments
8. First law of thermodynamics
9. Second law of thermodynamics
10. Thermodynamic processes
11. Governing equations
Balance Laws > Introduction
Introduction
Introduction
The fundamental laws of continuum mechanics are given by four
conservation/balance laws plus a restriction law.
The four conservation/balance laws are:

Conservation of mass · Mass continuity equation
Linear momentum balance · Cauchy first motion equation
Angular momentum balance · Symmetry of Cauchy stress
First law of thermodynamics · Energy balance equation
The restriction law is given by:

Second law of thermodynamics · Clausius-Planck and heat
conduction inequalities
Balance Laws > Introduction
Introduction
Introduction
The mathematical expressions arising from the fundamental
laws will be given in:
Global (or integral) form
o Global (or integral) spatial form
o Global (or integral) material form
Local (or strong) form
o Local (or strong) spatial form
o Local (or strong) material form

Balance Laws > Conservation of Mass
Conservation of Mass
We assume that during a motion there are neither mass sources
(reservoirs that supply mass), nor mass sinks (reservoirs that
absorb mass), so the mass of a continuum body is a conserved
quantity.
m ( Ω0 ) = m ( Ω ) > 0
Then, the mass is independent of the motion and, hence, the
material time derivative of the mass of a continuum body (or a
material volume) has to be zero,
d
m ( Ω) = 0
dt

Mass Density
The mass at the material (or reference) configuration may be
characterized by a continuous positive scalar field, denoted as
ρ0 = ρ0 ( X ) > 0 , which is a material property called material (or
reference) mass density, such that,
dm ( X ) = ρ0 ( X ) dV > 0
The mass at the spatial (or current) configuration may be
characterized by a continuous positive scalar field, denoted as
ρ = ρ ( x, t ) > 0 , which is called spatial (or current) mass density,
such that,
dm ( x, t ) = ρ ( x, t ) dv > 0
Note that, taking t=0 as reference time,
ρ ( x, 0 ) = ρ ( X, 0 ) = ρ0 ( X ) > 0.
Conservation of Mass: Global Material Form
The mass of a continuum body (or a material volume) is a
conserved quantity,
m ( Ω0 ) = ∫ ρ0 ( X ) dV = ∫ ρ ( x, t ) dv = m ( Ω ) > 0
Ω0 Ω
Using,
dv = J ( X, t ) dV > 0
The global material form of the conservation of mass may be
written as,
∫ Ω0
ρ0 ( X ) dV = ∫ ρ (ϕ ( X, t ) , t ) J ( X, t ) dV > 0
Ω0

Conservation of Mass: Local Material Form
Let us consider the global material form of the conservation of
mass given by,
∫ Ω0
ρ0 ( X ) dV = ∫ ρ (ϕ ( X, t ) , t ) J ( X, t ) dV > 0
Ω0
Localizing the integral expression, the local material form of the

conservation of mass reads,
ρ0 ( X ) = ρ (ϕ ( X, t ) , t ) J ( X, t ) > 0

Conservation of Mass: Global Material Form
The material time derivative of the mass of a continuum body
(or a material volume) has to be zero,
d d
m ( Ω ) = ∫ ρ ( x, t ) dv = 0
dt dt Ω
Using,
dv = J ( X, t ) dV > 0
The global material form of the conservation of mass may be
written as,
d
∫
dt Ω0
ρ (ϕ ( X, t ) , t ) J ( X, t ) dV = 0

Conservation of Mass: Local Material Form
Let us consider the global material form of the conservation of
mass given by,
d
∫
dt Ω0
ρ (ϕ ( X, t ) , t ) J ( X, t ) dV = 0
d
(
∫Ω0 dt ρ (ϕ ( X, t ) , t ) J ( X, t ) dV = 0 )
Localizing the integral expression, the local material form of the
d
dt
(
ρ (ϕ ( X, t ) , t ) J ( X, t ) = 0 )
Conservation of Mass: Global Spatial Form
The mass of a continuum body (or a material volume) is a
conserved quantity,
m ( Ω0 ) = ∫ ρ0 ( X ) dV = ∫ ρ ( x, t ) dv = m ( Ω ) > 0
Ω0 Ω
Using,
dV = J −1 ( x, t ) dv > 0
The global spatial form of the conservation of mass may be
written as,
∫ ρ0 (ϕ −1 ( x, t ) ) J −1 ( x, t ) dv = ∫ ρ ( x, t ) dv > 0
Ω Ω

Conservation of Mass: Local Spatial Form
Let us consider the global spatial form of the conservation of
mass given by,
∫ ρ0 (ϕ −1 ( x, t ) ) J −1 ( x, t ) dv = ∫ ρ ( x, t ) dv > 0
Ω Ω
Localizing the integral expression, the local spatial form of the

ρ0 (ϕ −1 ( x, t ) ) J −1 ( x, t ) = ρ ( x, t ) > 0

Conservation of Mass: Global Spatial Form
The material time derivative of the mass of a continuum body
(or a material volume) has to be zero,
d d
m ( Ω ) = ∫ ρ ( x, t ) dv = 0
dt dt Ω
The global spatial form of the conservation of mass may be
written as,
d
∫
dt Ω
ρ ( x, t ) dv = 0

Let us consider the global spatial form of the conservation of
mass given by,
d
∫
dt Ω
ρ ( x, t ) dv = 0
d d d
dt ∫Ω
ρ dv =
dt ∫Ω0
ρ JdV = ∫Ω0 dt ( ρ J ) dV = ∫Ω0 ( ρ
J + ρ ) dV
J
= ∫ ( ρ + ρ div v ) JdV = ∫ ( ρ + ρ div v ) dv = 0

Ω0 Ω
Localizing the integral expression, the local spatial form of the
conservation of mass, or mass continuity equation, reads,
ρ ( x, t ) + ρ ( x, t ) div v ( x, t ) = 0
The local spatial form of the conservation of mass, or mass
continuity equation, may be written as,
ρ ( x, t ) + ρ ( x, t ) div v ( x, t ) = 0
Using the following expressions,
∂ρ
ρ = + ( grad ρ ) ⋅ v, div ( ρ v ) = ρ div v + ( grad ρ ) ⋅ v
∂t
The local spatial form of the conservation of mass, or mass
continuity equation, may be alternatively written as,
∂ρ ( x, t )
+ div ( ρ ( x, t ) v ( x, t ) ) = 0
∂t
Global and ofLocal
Giving the spatial description Spatialthe Forms
an arbitrary property, material time derivative of the property can be written as,
d
∫
dt Ω
ρ dv = ∫ ( ρ + ρ div v ) dv = 0
Ω
∂ρ
ρ + ρ div v = 0, + div ( ρ v ) = 0
∂t
Global and Local Material Forms
d d
∫
dt Ω0
ρ JdV = ∫
Ω0 dt
( ρ J ) dV = 0
d
( ρ J ) = 0, ρ J = ρ0 > 0
dt
Balance Laws > Reynolds Transport Theorem
Convective Flux
Convective Flux of an Arbitrary Property
Consider an arbitrary property A of a continuum medium and
let us denote as ψ ( x,t ) the spatial description of the amount of
the property per unit of mass, ρ ( x,t ) the spatial density field
and v ( x,t ) the spatial velocity field.
The convective flux of the property A through a fixed spatial
surface with unit normal n ( x,t ) , i.e. the amount of the property
crossing the spatial surface per unit of time due to the
convective flux, is given by,
φA ( t ) = ∫ ρ ( x, t )ψ ( x, t ) v ( x, t ) ⋅ n ( x, t ) ds
s

Convective Flux
Convective Flux of an Arbitrary Property
The net outcoming convective flux of the property A through a
fixed closed spatial surface with unit outward normal n ( x,t ) , i.e.
the net amount of the property A leaving the spatial volume
per unit of time due to the convective flux, i.e. outflow (+) plus
inflow (-), is given by,
φA ( t ) = ∫ ρ ( x, t )ψ ( x, t ) v ( x, t ) ⋅ n ( x, t ) ds
∂v
inflow
v ⋅n ≤ 0
outflow
v ⋅n ≥ 0

Convective Flux
Mass Flux
Given a spatial density field, denoted as ρ ( x,t ) and a spatial
velocity field, denoted as v ( x,t ), the mass flux through a fixed
spatial surface with unit normal n ( x,t ) is given by,
φM ( t ) = ∫ ρ ( x, t ) v ( x, t ) ⋅ n ( x, t ) ds
s
Note that the mass flux may be viewed as a particular case of the
convective flux of an arbitrary property A , setting ψ = 1.
The net outcoming mass flux through a closed surface with unit
outward normal n ( x,t ) is given by,
φM ( t ) = ∫ ρ ( x, t ) v ( x, t ) ⋅ n ( x, t ) ds
∂v

Convective Flux
Volume Flux
Given a spatial velocity field, denoted as v ( x,t ), the volume flux
through a fixed spatial surface with unit normal n ( x,t ) is given
by,
φV ( t ) = ∫ v ( x, t ) ⋅ n ( x, t ) ds
s
Note that the volume flux may be viewed as a particular case of

the convective flux of an arbitrary property A , setting ψ = ρ −1 .
The net outcoming volume flux through a closed surface with
unit outward normal n ( x,t ) is given by,
φV ( t ) = ∫ v ( x, t ) ⋅ n ( x, t ) ds
∂v

Reynolds Transport Theorem

Reynolds Lemma
Consider an arbitrary property A of a continuum medium and
let us denote as ψ ( x,t ) the spatial description of the amount of
the property per unit of mass.
The amount of the property A at the current time t can be
written as,
A ( t ) = ∫ ρψ dv
Ω
The material time derivative of the property A can be written

as,
d
A ( t ) = ∫ ρψ dv

dt Ω

Reynolds Lemma
The material time derivative of the property A may be written
as,
d
A ( t ) = ∫ ρψ dv

dt Ω
d d
= ∫ ρψ JdV = ∫ ( ρψ J ) dV
dt 0Ω Ω 0 dt
d
=∫ ( ρ0ψ ) dV = ∫ ρ0ψ dV
Ω0 dt Ω0
= ∫ ρψ dv
Ω


Reynolds Lemma
The Reynolds Lemma for an arbitrary property A takes the
form,
d
∫
dt Ω
ρψ dv = ∫ ρψ dv
Ω
Note that Reynolds lemma may be directly obtained using mass

conservation, taking into account that,
dm ( x, t ) = ρ ( x, t ) dv > 0


The following key expression holds,
d
ρψ = ( ρψ ) − ρψ

dt
∂
= ( ρψ ) + grad ( ρψ ) ⋅ v − ρψ

∂t
∂
= ( ρψ ) + grad ( ρψ ) ⋅ v + ρψ div v
∂t
∂
= ( ρψ ) + div ( ρψ v )
∂t


The Reynolds transport theorem for an arbitrary property A
may be written as,
d
A ( t ) = ∫ ρψ dv

dt Ω
= ∫ ρψ dv
Ω
∂
=∫ ( ρψ ) dv + ∫Ω div ( ρψ v ) dv
Ω ∂t
∂
=∫ ( ρψ ) dv + ∫∂Ω ρψ v ⋅ nds
Ω ∂t


Reynolds Transport Theorem: Mass Conservation
Taking the mass as a particular case of an arbitrary property, the
Reynolds transport theorem yields a global spatial form of the
conservation of mass and take the form,
d
∫
dt Ω
ρ dv = 0
∂
=∫ ( ρ ) dv + ∫ div ( ρ v ) dv
Ω ∂t Ω
∂
=∫ ( ρ ) dv + ∫ ρ v ⋅ nds
Ω ∂t ∂Ω


Global Spatial
Giving the spatial description Form
d
∫
dt Ω
ρψ dv = ∫ ρψ dv
Ω
∂
=∫ ( ρψ ) dv + ∫Ω div ( ρψ v ) dv
Ω ∂t
∂
=∫ ( ρψ ) dv + ∫ ρψ v ⋅ nds
Ω ∂t ∂Ω
Local Spatial Form

∂
ρψ = ( ρψ ) + div ( ρψ v )
∂t
Assignment 6.1
Consider the spatial description of a velocity field given by,
vx = ye−t , v y = y, vz = 0
The reference time is t=0. The mass density at the reference
configuration is constant. Obtain the spatial mass density and
the mass flux through the open cylindrical surface S of the figure.
A
A

Balance Laws > Linear Momentum Balance
Linear Momentum Balance

Linear Momentum
The linear momentum of a material volume, denoted as M L ( t ),
is defined as a vector-valued function given by,
M L ( t ) = ∫ ρ ( x, t ) v ( x, t ) dv
Ω
= ∫ ρ0 ( X ) V ( X, t ) dV
Ω0


Linear Momentum
Using Reynolds Lemma, i.e. conservation of mass, the material
time derivative of the linear momentum takes the form,
d
M L ( t ) = ∫ ρ ( x, t ) v ( x, t ) dv = ∫ ρ ( x, t ) v ( x, t ) dv

dt Ω Ω
d
= ∫ ρ0 ( X ) V ( X, t ) dV = ∫ ρ0 ( X ) V ( X, t ) dV
dt Ω0 Ω0


Resultant Force
The resultant force acting on a material volume, denoted as F ( t ) ,
is defined as a vector-valued function given by,
F ( t ) = ∫ ρ ( x, t ) b ( x, t ) dv + ∫ t ( x, t ) ds
Ω ∂Ω
= ∫ ρ0 ( X ) B ( X, t ) dV + ∫ T ( X, t ) dS
Ω0 ∂Ω0


Linear Momentum Balance Law
The linear momentum balance law states that the time-
variation of the linear momentum of a material volume is equal
to the resultant force acting on that material volume.
d
ML (t ) = ML (t ) = F (t )

dt
If the continuum body is in equilibrium, the resultant force is

zero and the linear momentum is a conserved quantity,
d
ML (t ) = ML (t ) = 0

dt


Linear Momentum Balance Law: Global Spatial Form
The global spatial form of the linear momentum balance law
can be written as,
d
∫
dt Ω
ρ ( x, t ) v ( x, t ) dv = ∫ ρ ( x, t ) v ( x, t ) dv
Ω
= ∫ ρ ( x, t ) b ( x, t ) dv + ∫ t ( x, t ) ds
Ω ∂Ω


Linear Momentum Balance Law: Global Spatial Form
The surface forces can be written in spatial form as,
∫ ∂Ω
t ds = ∫ σ n ds = ∫ div σ dv
∂Ω Ω
Substituting into the global spatial form yields,

d
∫
dt Ω
ρ ( x, t ) v ( x, t ) dv = ∫ ρ ( x, t ) v ( x, t ) dv
Ω
= ∫ ρ ( x, t ) b ( x, t ) dv + ∫ div σ ( x, t )dv
Ω Ω
= ∫ ( ρ ( x, t ) b ( x, t ) + div σ ( x, t ) )dv
Ω


Linear Momentum Balance Law: Local Spatial Form
Localizing, the local spatial form of the linear momentum
balance law, known as Cauchy’s first equation of motion, can be
written as,
ρ ( x, t ) v ( x, t ) = ρ ( x, t ) b ( x, t ) + div σ ( x, t )
If the resultant force is zero, the local spatial form of the linear
momentum balance law can be written as,
0 = ρ ( x, t ) b ( x, t ) + div σ ( x, t )


Linear Momentum Balance Law: Global Material Form
The global material form of the linear momentum balance law
can be written as,
d
dt ∫Ω0 ∫Ω0 0
ρ 0 ( X ) V ( X, t ) dV = ρ ( X ) ( X, t ) dV
V
= ∫ ρ0 ( X ) B ( X, t ) dV + ∫ T ( X, t ) dS
Ω0 ∂Ω0


Linear Momentum Balance Law: Global Material Form
The surface forces can be written in material form as,
∫
∂Ω0
TdS = ∫
∂Ω0
PN dS = ∫ DIV P dV
Ω0
Substituting into the global material form yields,

d
dt ∫Ω0 ∫Ω0
ρ 0 ( X ) V ( X, t ) dV = ρ 0 ( X ) ( X, t ) dV
V
= ∫ ρ0 ( X ) B ( X, t ) dV + ∫ DIV P ( X, t )dV
Ω0 Ω0
=∫ ( ρ ( X ) B ( X, t ) + DIV P ( X, t ) ) dV
0
Ω0


Linear Momentum Balance Law: Local Material Form
Localizing, the local material form of the linear momentum
balance law can be written as,
ρ0 ( X ) V ( X, t ) = ρ0 ( X ) B ( X, t ) + DIV P ( X, t )
If the resultant force is zero, the local material form of the linear
momentum balance law can be written as,
0 = ρ0 ( X ) B ( X, t ) + DIV P ( X, t )


Global and ofLocal
d
∫
dt Ω
ρ v dv = ∫ ρ v dv = ∫ ρ b dv + ∫ t ds
Ω Ω ∂Ω
ρ v = ρ b + div σ
d
∫
dt 0
Ω
ρ0 v dV = ∫ ρ0 v dV = ∫ ρ0b dV + ∫ T dS
Ω0 Ω0 ∂Ω0
ρ0 v = ρ0b + DIV P

Balance Laws > Angular Momentum Balance
Angular Momentum Balance

Angular Momentum
The angular momentum of a material volume about a fixed
spatial point x0 , denoted as M A ( t ) , is defined as a vector-valued
function given by,
M A ( t ) = ∫ r ( x, t ) × ρ ( x, t ) v ( x, t ) dv
Ω
= ∫ R ( X, t ) × ρ0 ( X ) V ( X, t ) dV
Ω0
where the vector position r = r ( x, t ) = R ( X, t ) is defined as,

r = r ( x, t ) = x − x0 = ϕ ( X, t ) − x0 = R ( X, t )


Angular Momentum
Using Reynolds Lemma, i.e. conservation of mass, the material
time derivative of the angular momentum takes the form,
d
M A ( t ) = ∫ r ( x, t ) × ρ ( x, t ) v ( x, t ) dv

dt Ω
= ∫ r ( x, t ) × ρ ( x, t ) v ( x, t ) dv
Ω
d
= ∫ R ( X, t ) × ρ0 ( X ) V ( X, t ) dV
dt Ω0
∫
= R ( X, t ) × ρ ( X ) V
Ω0
( X, t ) dV
0


Resultant Moment
The resultant moment about a fixed spatial point x0 , denoted as
M ( t ) , is defined as a vector-valued function given by,
M ( t ) = ∫ r ( x, t ) × ρ ( x, t ) b ( x, t ) dv + ∫ r ( x, t ) × t ( x, t ) ds
Ω ∂Ω
= ∫ R ( X, t ) × ρ0 ( X ) B ( X, t ) dV + ∫ R ( X, t ) × T ( X, t ) dS
Ω0 ∂Ω0


ϕ
time t = 0 time t
F dv
Ω0
dV P’
P Ω
dA r = x − x0 ρb
da
r = x − x0
ρ0 B dA X ρ0B
X 3 , x3 x da
N x0 x
ê 3
n
T ê 2 X 2 , x2 t
ê1
T
X 1 , x1

Angular Momentum Balance Law
The angular momentum balance law states that the time-
variation of the angular momentum of a material volume about
a fixed spatial point, is equal to the resultant moment about this
fixed spatial point, acting on that material volume.
d
M A (t ) = M A (t ) = M (t )

dt
If the resultant moment about the fixed spatial point is zero,
then the angular momentum about this spatial point is a
conserved quantity,
d
M A (t ) = M A (t ) = 0

dt

Angular Momentum Balance: Global Spatial Form
The global spatial form of the angular momentum balance law
can be written as,
d
∫
dt Ω
r ( x, t ) × ρ ( x, t ) v ( x, t ) dV = ∫ r ( x, t ) × ρ ( x, t ) v ( x, t ) dv
Ω
= ∫ r ( x, t ) × ρ ( x, t ) b ( x, t ) dv
Ω
+ ∫ r ( x, t ) × t ( x, t ) ds
∂Ω


The moment of the surface forces can be written in spatial form
as,
∫ ∂Ω
r × t ds = ∫ r × σ n ds = ∫ r × div σ dv + ∫ ε abcσ cbea dv
∂Ω Ω Ω
∫∂Ω
ε abc rbtcea ds = ∫ ε abc rbσ cd nd ea ds
∂Ω
= ∫ ε abc ( rbσ cd ),d ea dv

Ω
= ∫ ε abc rbσ cd ,d ea dv + ∫ ε abc rb,dσ cd ea dv

Ω Ω
= ∫ ε abc rb ( ∇ ⋅ σ )c ea dv + ∫ ε abcσ cbea dv

Ω Ω


Substituting into the global spatial form yields,
d
∫
dt Ω
r ( x, t ) × ρ ( x, t ) v ( x, t ) dV = ∫ r ( x, t ) × ρ ( x, t ) v ( x, t ) dv
Ω
= ∫ r ( x, t ) × ρ ( x, t ) b ( x, t ) dv
Ω
+ ∫ r ( x, t ) × div σ ( x, t ) dv
Ω
+ ∫ ε abcσ cb ( x, t ) ea dv
Ω


Angular Momentum Balance: Local Spatial Form
Localizing, the local spatial form of the angular momentum
balance law yields,
r × ρ v = r × ( ρ b + div σ ) + ε abcσ cbea
Using the Cauchy’s first motion equation yields,
ε abcσ cbea = 0
and the local spatial form of the angular momentum balance
law can be written as,
σ ( x, t ) = σ T ( x, t )


Angular Momentum Balance: Global Material Form
The global material form of the angular momentum balance law
can be written as,
d
dt ∫Ω0 ∫Ω0
R ( X, t ) × ρ 0 ( X ) V ( X, t ) dV = R ( X, t ) × ρ 0 ( X ) ( X, t ) dV
V
= ∫ R ( X, t ) × ρ0 ( X ) B ( X, t ) dV
Ω0
+∫ R ( X, t ) × T ( X, t ) dS
∂Ω0


The moment of the surface forces can be written in material
form as,
∫ r × TdS = ∫ r × PNdS = ∫ r × DIV PdV + ∫ ε abc( PFT ) ea dV
∂Ω0 ∂Ω0 Ω0 Ω0 cb
∫
∂Ω0
ε abc rbTcea dS = ∫ ε abc rb PcD N Dea dS
∂Ω0
= ∫ ε abc ( rb PcD ), D ea dV
Ω0
= ∫ ε abc rb PcD , Dea dV + ∫ ε abc rb , D PcDea dV

Ω0 Ω0
= ∫ ε abc rb ( ∇ ⋅ P )c ea dV + ∫ ε abc FbD PcDea dV

Ω0 Ω0
= ∫ r × DIV PdV + ∫ ε abc( PFT ) ea dV

October 8, 2013 Ω0 Carlos Agelet de Saracibar Ω0 cb 57

Substituting into the global material form yields,
d
∫
dt Ω0
r × ρ0 vdV = ∫ r × ρ0 v dV
Ω0
= ∫ r × ρ0 bdV + ∫ r × DIV PdV

Ω0 Ω0
+ ∫ ε abc( PFT ) ea dV
Ω0 cb


Angular Momentum Balance: Local Material Form
Localizing, the local material form of the angular momentum
balance law can be written as,
r × ρ0 v = r × ρ0 b + r × DIV P + ε abc( PFT ) ea
cb
Using the Cauchy’s first motion equation yields,
ε abc( PFT )cbea = 0
and the local material form of the angular momentum balance
law can be written as,
P ( X, t ) FT ( X, t ) = F ( X, t ) PT ( X, t )


Global and ofLocal
d
∫
dt Ω
r × ρ vdV = ∫ r × ρ v dv
Ω
= ∫ r × ρ b dv + ∫ r × t ds
Ω ∂Ω
σ =σT
d
∫
dt Ω0
r × ρ0 vdV = ∫ r × ρ0 v dV
Ω0
= ∫ r × ρ0b dV + ∫ r × T dS
Ω0 ∂Ω0
PFT = FPT
Balance Laws > Mechanical Energy Balance
Kinetic Energy
Kinetic Energy
The global spatial form of the kinetic energy of a continuum
body, denoted as K ( t ) , takes the form,
1 2 1
K ( t ) = ∫ ρ ( x, t ) v ( x, t ) dv = ∫ ρ ( x, t ) v ( x, t ) ⋅ v ( x, t ) dv
2 Ω 2 Ω
The global material form of the kinetic energy of a continuum

body, denoted as K ( t ) , takes the form,
1 2 1
K (t ) =
2
∫ Ω0
ρ0 ( X ) V ( X, t ) dV =
2
∫ Ω0
ρ0 ( X ) V ( X, t ) ⋅ V ( X, t ) dV

Internal Mechanical Power

The internal mechanical power per unit of spatial volume, is the
work done by the stresses per unit of time and unit of spatial
volume, and can be written as,
σ : d = J −1τ : d
= J −1 ( PFT ) : d = J −1PFT : l = J −1P : ( lF ) = J −1P : F
= J −1 ( FSFT ) : d = J −1S : ( FT dF ) = J −1S : E

σ ab d ab = J −1τ ab d ab
= J −1 ( PaA FAb
T
) ab
d = J −1
( aA Ab ) ab
P F T
l = J −1
P ( l F
aA ab bA ) = J −1
PaA FaA
= J −1 ( FaA S AB FBbT ) d ab = J −1S AB ( FAa
T
d ab FbB ) = J −1S AB E AB

The internal mechanical power in the continuum body, denoted
as Pint ( t ) , i.e. work done per unit of time by the stresses, can be
written as,
Pint ( t ) = ∫ σ : d dv
Ω
= ∫ J −1τ : d dv = ∫ τ : d dV
Ω Ω0
= ∫ J −1P : F dv = ∫ P : F dV
Ω Ω0
= ∫ J −1S : E ∫
dv = S : E
dV
Ω Ω0


Global Spatial
Giving the spatial description Form
Pint ( t ) = ∫ σ : d dv
Ω
Global Material Forms
Pint ( t ) = ∫ τ : d dV
Ω0
= ∫ P : F dV
Ω0
= ∫ S:E
dV
Ω0

External Mechanical Power

External Mechanical Power
The external mechanical power, denoted as Pext ( t ) , is the work
done per unit of time by the body forces and surface forces, and
can be written in global spatial and material forms, respectively,
as,
Pext ( t ) = ∫ ρ ( x, t ) b ( x, t ) ⋅ v ( x, t ) dv + ∫ t ( x, t ) ⋅ v ( x, t ) ds
Ω ∂Ω
= ∫ ρ0 ( X ) B ( X, t ) ⋅ V ( X, t ) dV + ∫ T ( X, t ) ⋅ V ( X, t ) dS
Ω0 ∂Ω0

Mechanical Energy Balance

The external mechanical power, denoted as Pext ( t ) , i.e. work
done per unit of time by the body forces and surface forces, can
be written in global spatial form as,
Pext ( t ) = ∫ ρ b ⋅ v dv + ∫ t ⋅ v ds
Ω ∂Ω
The external mechanical power of the surface forces can be

written in global spatial form as,
∫ ∂Ω
t ⋅ v ds = ∫ v ⋅ σ n ds
∂Ω
= ∫ div ( v ⋅ σ ) dv = ∫ v ⋅ div σ dv + ∫ σ : grad v dv

Ω Ω Ω


Then, using the local spatial form of the linear momentum
balance equation, the external mechanical power can be written
in global spatial form as,
Pext ( t ) = ∫ ρ b ⋅ v dv + ∫ t ⋅ v ds
Ω ∂Ω
= ∫ ρ b ⋅ v dv + ∫ v ⋅ div σ dv + ∫ σ : grad v dv
Ω Ω Ω
= ∫ ( ρ b + div σ ) ⋅ v dv + ∫ σ : grad v dv
Ω Ω
dv 1 d
= ∫ ρ ⋅ v dv + ∫ σ : l dv = ∫ ρ v dv + ∫ σ : d dv
2
Ω dt Ω Ω2 dt Ω
d 1 d
= ∫ ρ v dv + ∫ σ :d dv = K ( t ) + Pint ( t )
2
dt Ω 2 Ω dt

The mechanical energy balance states that the external
mechanical power supplied to the continuum body is spent in
changing its kinetic energy and doing an internal mechanical
power.
d
Pext ( t ) = K ( t ) + Pint ( t )
dt


Mechanical Energy Balance: Global Spatial Form
The global spatial form of the mechanical energy balance can be
written as,
Pext ( t ) = ∫ ρ b ⋅ v dv + ∫ t ⋅ v ds
Ω ∂Ω
d 1
= ∫ ρ v dv + ∫ σ :d dv
2
dt Ω 2 Ω


Mechanical Energy Balance: Global Material Form
The global material form of the mechanical energy balance can
be written as,
Pext ( t ) = ∫ ρ0 b ⋅ v dV + ∫ T ⋅ v dS
Ω0 ∂Ω0
d 1
= ∫ ρ0 v dV + ∫ τ :d dV
2
dt Ω0 2 Ω0
d 1
= ∫ ρ0 v dV + ∫ P : F dV
2
dt Ω0 2 Ω0
d 1
= ∫ ρ0 v dV + ∫ S : E dV
2
dt Ω0 2 Ω0


Quasistatic Problem
If the material time derivative of the kinetic energy is zero (or
negligible), the problem is called quasistatic and the mechanical
energy balance reads,
Pext ( t ) = Pint ( t )


Free Vibration Problem
If the external mechanical power is zero (or negligible), the
problem is called free vibration problem and the mechanical
d
0 = K ( t ) + Pint ( t )
dt


Conservative Mechanical System
If both the external and internal mechanical power derive from a
potential, i.e. a potential energy for the external loading and a
strain energy, such that,
d d
Pext ( t ) = − Π ext ( t ) , Pint ( t ) = Π int ( t )
dt dt
the problem is said to be conservative and the mechanical
d d d
− Π ext ( t ) = K ( t ) + Π int ( t )
dt dt dt
d
dt
( K ( t ) + Π int ( t ) + Π ext ( t ) ) = 0

Conservative Mechanical System
The total potential energy, denoted as Π ( t ) , is defined as the
sum of the potential energy of the external loading, denoted as
Π ext ( t ) and the strain energy, denoted as Π int ( t ) , yielding,
d d
dt
( K ( t ) + Π int ( t ) + Π ext ( t ) ) = ( K ( t ) + Π ( t ) ) = 0
dt
In a conservative mechanical system the sum of the kinetic

energy and the total potential energy is a conserved quantity,
K ( t ) + Π ( t ) = constant

Balance Laws > Assignments
Assignment 6.2
Assignment 6.2
A volume flux Q of an incompressible fluid flows in stationary
conditions through the pipeline of the figure. Velocity and
pressure distributions at the sections A and B are uniform. The
pipeline is fixed through a rigid bar OP. The weights of the
pipeline, rigid bar and fluid are neglected.
SB
eB vB
pB
R
pA vA θ
eA
P O
F M
SA

Assignment 6.2
1) Obtain the velocities at the sections A and B in terms of Q.
2) Obtain the reaction force F and moment M at the point O.
3) Obtain the values of the angle θ that maximize and minimize
the reaction at the point O.
4) Obtain the external power needed to keep the volume flux Q
if the fluid is an incompressible ideal fluid with a spherical
stress state given byen by σ = − p1.
SB
eB vB
pB
R
pA vA θ
eA
P O
F M
SA

Assignment 6.2
The local spatial form of the conservation of mass or mass
continuity equation, plus the incompressibility condition yields,
ρ + ρ div v = 0, ρ = 0 ⇒ div v = 0
The global spatial form of the conservation of mass for an
incompressible medium reads,
∫V
div v dV = ∫ v ⋅ n dS
∂V
= ∫ v ⋅ n dS + ∫ v ⋅ n dS
SA SB
= −vA S A + vB S B = 0
Then the velocities at the sections A and B are given by,
vA S A = vB S B = Q ⇒ vA = Q S A , vB = Q S B
Assignment 6.2
The global spatial form of the linear momentum balance for a
stationary motion reads,
d
R/ f = ∫ ρ vdV
dt V
∂
= ∫ ρ vdV + ∫ ρ v ( v ⋅ n ) dS = ∫ ρ v ( v ⋅ n ) dS
∂t V ∂V ∂V
= ∫ ρ v ( v ⋅ n ) dS + ∫ ρ v ( v ⋅ n ) dS
SA SB
2 2
Q Q
= − ρ vA2 S Ae A + ρ vB2 S Be B = − ρ eA + ρ eB
SA SB

Assignment 6.2
The resultant force acting on the volume V of fluid, taking into
account that the weight of the fluid is negligible, reads,
∫
R/ f =
V
ρ bdV + ∫ tdS = ∫ tdS
∂V ∂V
=∫ tdS + ∫ tdS + ∫ tdS

wall SA SB
= R wall / f + R pA + R pB
= R wall / f + pA S Ae A − pB S Be B
Q2 Q2
= −ρ eA + ρ eB
SA SB

Assignment 6.2
The resultant force of the wall of the pipeline acting on the
volume V of fluid, reads,
R wall / f = R / f − pA S Ae A + pB S Be B
 Q2   Q2 
= − ρ + pA S A  e A +  ρ + pB S B  e B
 SA   SB 
Using the action-reaction principle, the resultant force of the
fluid acting on the wall of the pipeline, reads,
R f / wall = −R wall / f
 Q2   Q2 
=ρ + pA S A  e A −  ρ + pB S B  e B
 SA   SB 
Assignment 6.2
The global spatial form of the angular momentum balance about
the point O, for a stationary motion, reads,
d
M O
/f = ∫ r × ρ vdV
dt V
∂
= ∫ r × ρ vdV + ∫ r × ρ v ( v ⋅ n ) dS = ∫ r × ρ v ( v ⋅ n ) dS
∂t V ∂V ∂V
= ∫
SA
r × ρ v ( v ⋅ n ) dS + ∫ r × ρ v ( v ⋅ n ) dS
SB
2
Q
= − ρ vB2 S B Re z = − ρ Re z
SB

Assignment 6.2
The resultant moment about the point O acting on the volume V
of fluid, taking into account that the weight of the fluid is
negligible, reads,
MO/ f =∫ r × ρbdV + ∫ r × tdS = ∫ r × tdS

V ∂V ∂V
= ∫ r × tdS + ∫ r × tdS + ∫ r × tdS

wall SA SB
= MOwall / f + MOpA + MOpB

= MOwall / f + pB S B Re z
Q2
= −ρ Re z
SB
Assignment 6.2
The resultant moment of the wall of the pipeline acting on the
volume V of fluid about the point O, reads,
 Q 2

MOwall / f O
= M / f − pB S B Re z = −  ρ + pB S B  Re z
 SB 
Using the action-reaction principle, the resultant moment of the
fluid acting on the wall of the pipeline about the point O, reads,
MOf / wall = −M Owall / f
 Q2 
=ρ + pB S B  Re z
 SB 

Assignment 6.2
Equilibrium of forces and moments at point O
 Q2 
Rb , f / wall =ρ + pB S B 
 SB  M
 Q2 
O Ra , f / wall =ρ + pA S A 
 SA 
Fa
Fb
O  Q 2

M f / wall =ρ + pB S B  R
 SB 

Assignment 6.2
The equilibrium of forces and moments about the point O on the
pipeline, taking into account that the weight of the pipeline and
the rigid bar are negligible, reads,
R f / wall + W + F = 0
MOf / wall + MWO + M = 0
The reaction force F and moment M, at the point O, read,
F = −R f / wall = − ( ρ Q 2 S A + pA S A ) e A + ( ρQ 2 S B + pB S B ) e B
 Q2 
M = −MOf / wall = − ρ + pB S B  Re z
 SB 

Assignment 6.2
The reactions force F at the point O, reads,
F = − ( ρ Q 2 S A + p A S A ) e A + ( ρ Q 2 S B + pB S B ) e B

>0 >0
The norm of the reaction force F will take a maximum for an

angle such that,
3π
e B = −e A ⇒ θ=
2
The norm of the reaction force F will take a minimum for an
angle such that,
π
eB = e A ⇒ θ =
2
Assignment 6.2
The external mechanical power needed to keep the volume flux
Q, taking into account that the fluid is incompressible, stationary
and the stress state is spherical, is given by,
d d 1
Pext = K + Pint = ∫ ρ v dV + ∫ σ : d dV
2
dt dt V 2 V
∂ 1 1
= ∫ ρ v dV + ∫ ρ v v ⋅ n dS + ∫ − p tr d dV
2 2
∂t V 2 ∂V 2 V
1 1
=∫ ρ v v ⋅ n dS + ∫ ρ v v ⋅ n dS
2 2
SA 2 SB 2
1 3 1 1 
= ρQ  2 − 2 
2  SB S A 
Assignment 6.3
An incompressible fluid flows in stationary conditions through
the pipeline of the figure. Velocities and pressures distributions
are uniform at the sections AE and CD. Pressure on the walls is
assumed to be uniform. There is a basculant barrier BC with a
hinge on B. An horizontal force F, acting on the point C, is
keeping the barrier in vertical position. Body forces in the fluid
are neglected. The weight of the barrier is also neglected.
A B
C F h/2
v1 v2 h/2
y
E D p = patm ≈ 0
x
h
z
h

Assignment 6.3
1) Obtain the velocity v2 at the section CD in terms of the
velocity v1 at the section AE.
2) Obtain the resultant force and moment acting on the fluid at
the point B.
3) Obtain the resultant force and moment of the fluid on the
barrier at the point B.
4) Obtain the force F and the reaction at the point B.
5) Obtain the external mechanical power needed, assuming the
stress tensor in the fluid is spherical,
A
given by σ = − p1 . C
B
F h/2
v1 v2 h/2
y
E D p = patm ≈ 0
x h
z
h

Assignment 6.4
The figure shows the longitudinal section of a pump with a valve
OA of weight W per unit of width (normal to the plane of the
figure). There is a hinge on O. The velocity of the pump is V.
The fluid is incompressible b a
A
and the motion stationary. P αa
O 2
Uniform pressure distribu- B

a/2 a/2
P
tions on the sections OB
2
P W A
1
and AB are p1 and p2=0, O

v
v
2 αa
respectively. Velocity dis- B
tributions at the sections P 1
OB and AB are uniform. Detail
Body forces in the fluid are V

negligible.
Assignment 6.4
1) Obtain the uniform velocities v1 and v2 at the sections OB
and AB, respectively, in terms of the velocity v of the
pumping tool
2) Obtain the resultant of the forces per unit of width given by
the fluid on the valve OA
3) Obtain the moment per unit of width at the point O of the
forces given by the fluid on the valve OA
4) Obtain the weight W of the valve OA per unit of width.
Environmental pressure p2 is neglected.

Balance Laws > First Law of Thermodynamics
External Thermal Power

External Thermal Power: Global Spatial Form
The external thermal power is defined as the amount of heat
that enters into a material volume per unit of time.
We will assume that heat may enter into a material volume due
to:
Internal heat sources characterized by a scalar-valued
function, given in spatial description by r ( x, t ), representing
the heat source per unit of mass.
Non-convective heat flux through the boundary, characterized
by a vector-valued function, given in spatial description by
q ( x,t ) , representing the non-convective outward heat flux
per unit of spatial surface.


External Thermal Power: Global Spatial Form
The global spatial form of the external thermal power is defined
as,
Qext ( t ) = ∫ ρ ( x, t ) r ( x, t ) dv − ∫ q ( x, t ) ⋅ n ds
Ω ∂Ω
= ∫ ρ ( x, t ) r ( x, t ) dv − ∫ div q ( x, t ) dv
Ω Ω
where r ( x, t ) is the heat source per unit of mass and q ( x, t ) is

the non-convective outward heat flux per unit of spatial surface,
both of them given in spatial description.


External Thermal Power: Global Material Form
Using the conservation of mass yields,
∫ ρ ( x, t ) r ( x, t ) dv = ∫
Ω Ω0
ρ0 ( X ) R ( X, t ) dV
Using Nanson’s formula, the heat flux through the spatial

boundary of the continuum body may be written as,
∫
∂Ω
q ⋅ n ds = ∫
∂Ω0
q ⋅ J F −T N dS := ∫
∂Ω0
Q ⋅ N dS
where Q is the nominal heat flux, i.e. heat flux per unit of
material surface, given by,
Q = J F −1q, QA = JFAa−1qa


External Thermal Power: Global Material Form
The global material form of the external thermal power takes
the form,
Qext ( t ) = ∫ ρ0 ( X ) R ( X, t ) dV − ∫ Q ( X, t ) ⋅ N dS
Ω0 ∂Ω0
= ∫ ρ0 ( X ) R ( X, t ) dV − ∫ DIV Q ( X, t ) dV
Ω0 Ω0
where R ( X, t ) is the heat source per unit of mass and Q ( X, t ) is

the nominal heat flux or heat flux per unit of material surface,
both of them given in material description.

Total External Power

Total External Power: Global Spatial Form
The global material form of the total external power, i.e.
mechanical plus thermal external power, can be written as,
Pext ( t ) + Qext ( t ) = ∫ ρ b ⋅ v dv + ∫ t ⋅ v ds
Ω ∂Ω
+ ∫ ρ r dv − ∫ q ⋅ n ds
Ω ∂Ω
d 1
= ∫ ρ v dv + ∫ σ :d dv
2
dt Ω 2 Ω
+ ∫ ρ rdv − ∫ q ⋅ n ds
Ω ∂Ω


Total External Power: Global Material Form (I)
The global material form of the total thermal power, i.e.
Pext ( t ) + Qext ( t ) = ∫ ρ0 b ⋅ v dV + ∫ T ⋅ v dS
Ω0 ∂Ω0
+ ∫ ρ0 r dV − ∫ Q ⋅ N dS
Ω0 ∂Ω0
d 1
= ∫ ρ0 v dV + ∫ τ :d dV
2
dt 0 2
Ω Ω0
+ ∫ ρ0 rdV − ∫ Q ⋅ N dS
Ω0 ∂Ω0


Total External Power: Global Material Form (II)
Ω0 ∂Ω0
+ ∫ ρ0 r dV − ∫ Q ⋅ N dS
Ω0 ∂Ω0
d 1
= ∫ ρ0 v dV + ∫ P : F dV
2
dt Ω0 2 Ω0
+ ∫ ρ0 rdV − ∫ Q ⋅ N dS
Ω0 ∂Ω0


Total External Power: Global Material Form (III)
Ω0 ∂Ω0
+ ∫ ρ0 r dV − ∫ Q ⋅ N dS
Ω0 ∂Ω0
d 1
= ∫ ρ0 v dV + ∫ S : E dV
2
dt Ω0 2 Ω0
+ ∫ ρ0 rdV − ∫ Q ⋅ N dS
Ω0 ∂Ω0

First Law of Thermodynamics

First Postulate. There exist a state function called total energy,
denoted as E ( t ) , such that its material time derivative is equal
to the total external power supplied to the system, i.e. the
external mechanical plus thermal power,
d
E ( t ) := Pext ( t ) + Qext ( t )
dt


Second Postulate. There exist a state function called internal
energy, denoted as U ( t ), which is an extensive property, i.e.
there exist a specific internal energy or internal energy per unit
of mass, denoted as e = e ( x, t ) = E ( X, t ) , such that,
U ( t ) := ∫ ρ ( x, t ) e ( x, t ) dv = ∫ ρ0 ( X ) E ( X, t ) dV
Ω Ω0


Ther first law of thermodynamics states that the material time
derivative of the total energy is equal to sum of the material
time derivative of the kinetic energy and the material time
derivative of the internal energy,
d d d
E (t ) = K (t ) + U (t )
dt dt dt


Using the first postulate and the mechanical energy balance, the
first law of thermodynamics reads,
d d d
E (t ) = K (t ) + U (t )
dt dt dt
= Pext ( t ) + Qext ( t )
d
= K ( t ) + Pint ( t ) + Qext ( t )
dt
yielding the following internal energy balance law,
d
U ( t ) = Pint ( t ) + Qext ( t )
dt

Energy Balance
Energy Balance Law: Global Spatial Form
The internal energy balance law in global spatial form can be
written as,
d
U ( t ) = Pint ( t ) + Qext ( t )
dt
d
∫
dt Ω
ρ e dv = ∫ ρ e dv
Ω
= ∫ σ : d dv + ∫ ρ r dv − ∫ q ⋅ n ds
Ω Ω ∂Ω

Energy Balance
Energy Balance Law: Local Spatial Form
Using the divergence theorem, the internal energy balance law
in global spatial form can be written as,
d
∫
dt Ω
ρ e dv = ∫ ρ e dv
Ω
= ∫ σ : d dv + ∫ ρ r dv − ∫ q ⋅ n ds
Ω Ω ∂Ω
= ∫ σ : d dv + ∫ ρ r dv − ∫ div q dv
Ω Ω Ω
Localizing, the local spatial form of the energy balance law

reads,
ρ e = σ : d + ρ r − div q
Energy Balance
Energy Balance Law: Global Material Form
The internal energy balance law in global material form can be
written as,
d
U ( t ) = Pint ( t ) + Qext ( t )
dt
d
∫
dt Ω0
ρ0 e dV = ∫ ρ0 e dV
Ω0
= ∫ τ : d dV + ∫ ρ0 r dV − ∫ Q ⋅ N dS
Ω0 Ω0 ∂Ω0
= ∫ P : F dV + ∫ ρ0 r dV − ∫ Q ⋅ N dS
Ω0 Ω0 ∂Ω0
= ∫ S:E ∫ 0
dV + ρ r dV −
∫ Q ⋅ N dS
Ω0 Ω0 ∂Ω0
Energy Balance
Energy Balance Law: Global Material Form
Using the divergence theorem, the internal energy balance law
in global material form can be written as,
d
∫
dt Ω0
ρ0 e dV = ∫ ρ0 e dV
Ω0
= ∫ τ : d dV + ∫ ρ0 r dV − ∫ DIV Q dV
Ω0 Ω0 Ω0
= ∫ P : F dV + ∫ ρ0 r dV − ∫ DIV Q dV
Ω0 Ω0 Ω0
= ∫ S:E ∫ 0 ∫
dV + ρ r dV − DIV Q dV
Ω0 Ω0 Ω0

Energy Balance
Energy Balance Law: Local Material Form
Localizing, the internal energy balance law in local material form
can be written as,
ρ0 e = τ : d + ρ0 r − DIV Q
= P : F + ρ0 r − DIV Q
= S:E + ρ r − DIV Q
0
Note that the local material form of the energy balance equation
could have been also obtained from the local spatial form using,
ρ0 = J ρ , DIV Q = J div q, τ : d = P : F = S : E = J σ : d

Energy Balance
Local Spatial
Giving the spatial Form
description of an arbitrary property, the material time derivative of the property can be written as,
ρ e = σ : d + ρ r − div q
Local Material Forms

ρ0 e = τ : d + ρ0 r − DIV Q
= P : F + ρ0 r − DIV Q
= S:E + ρ r − DIV Q
0

Balance Laws > Second Law of Thermodynamics
Second Law of Thermodynamics

First Postulate. There exist a state function called absolute
temperature, denoted as θ = θ ( x, t ) = Θ ( X, t ), which is always a
positive scalar-valued function.
θ = θ ( x, t ) = Θ ( X, t ) > 0
Second Postulate. There exist a state function called entropy,

denoted as H ( t ) , which is an extensive property, i.e. there exist
a specific entropy or entropy per unit of mass, denoted as
η = η ( x, t ) = Ξ ( X, t ) , such that,
H ( t ) = ∫ ρ ( x, t )η ( x, t ) dv = ∫ ρ0 ( X ) Ξ ( X, t ) dV
Ω Ω0


The second law of thermodynamics states that any admissible
thermodynamic process has to satisfy the following inequality,
written in global spatial form as,
d 1 1
H ( t ) ≥ ∫ ρ rdv − ∫ q ⋅ n ds
dt Ωθ ∂Ω θ
or in global material form as,

d 1 1
H (t ) ≥ ∫ ρ0 rdV − ∫ Q ⋅ N dS
dt Ω0 θ ∂Ω0 θ


Admissible thermodynamic processes can be classified as
reversible and irreversible processes.
A thermodynamic process is said to be reversible if the following
condition holds,
d 1 1 1 1
H ( t ) = ∫ ρ rdv − ∫ q ⋅ n ds = ∫ ρ0 rdV − ∫ Q ⋅ N dS
dt Ω θ ∂Ω θ Ω 0 θ ∂Ω 0 θ
A thermodynamic process is said to be irreversible if the

following condition holds,
d 1 1 1 1
H ( t ) > ∫ ρ rdv − ∫ q ⋅ n ds = ∫ ρ0 rdV − ∫ Q ⋅ N dS
dt Ωθ ∂Ω θ Ω0 θ ∂Ω0 θ


Second Law of Thermodynamics: Global Spatial Form
The global spatial form of the second law of thermodynamics
can be written as,
d d
H ( t ) = ∫ ρη dv = ∫ ρηdv
dt dt Ω Ω
1 1
≥ ∫ ρ rdv − ∫ q ⋅ n ds
Ω θ ∂Ω θ


Second Law of Thermodynamics: Global Spatial Form
Using the divergence theorem, the global spatial form of the
second law of thermodynamics can be written as,
d
∫
dt Ω
ρη dv = ∫ ρηdv
Ω
1 1
≥ ∫ ρ rdv − ∫ q ⋅ n ds
Ω θ ∂Ω θ
1 1 
= ∫ ρ rdv − ∫ div  q  dv
Ωθ Ω
θ 
1 1 1
= ∫ ρ rdv − ∫ div q dv + ∫ 2 q ⋅ grad θ dv
Ω θ Ω θ Ω θ

Second Law of Thermodynamics: Local Spatial Form
Localizing, the local spatial form of the second law of
thermodynamics can be written as,
1 1 1
ρη ≥ ρ r − div q + 2 q ⋅ grad θ
θ θ θ
Multiplying by the absolute temperature yields,
1
ρθη ≥ ρ r − div q + q ⋅ grad θ
θ


Clausius-Duhem Inequality: Local Spatial Form
The local spatial form of the Clausius-Duhem inequality states
that the dissipation rate per unit of spatial volume is a non-
negative scalar-valued quantity and it can be written as,
1
D := ρθη − ρ r + div q − q ⋅ grad θ ≥ 0

θ
A stronger assumption is usually introduced, yielding,
1
ρθη − ρ r + div q − q ⋅ grad θ ≥ 0
D := Dint + Dcond =
D ≥0
θ

int
Dcond ≥ 0
1
Dint := ρθη − ρ r + div q ≥ 0, Dcond := − q ⋅ grad θ ≥ 0
θ

Heat Conduction Inequality: Local Spatial Form
The local spatial form of the heat conduction inequality states
that the projection of the heat flux per unit of spatial surface on
the direction of the spatial gradient of the temperature is a non-
positive scalar-valued quantity, i.e. heat flux takes place from the
hot to the cold and not the other way around,
1
Dcond := − q ⋅ grad θ ≥ 0 ⇒ q ⋅ grad θ ≤ 0
θ


Heat Conduction Inequality: Local Spatial Form
Using Fourier law for heat conduction for an isotropic continuum
medium, the second law of thermodynamics yields the following
restriction on the admissible values of the spatial thermal
conductivity parameter,
1
Dcond := − q ⋅ grad θ ≥ 0 ⇒ k grad θ ⋅ grad θ ≥ 0 ⇒ k ≥ 0
θ


Clausius-Planck Inequality: Local Spatial Form (I)
The local spatial form of the Clausius-Planck inequality states
that the internal dissipation rate per unit of spatial volume is a
non-negative scalar-valued quantity and it can be written as,
Dint := ρθη − ρ r + div q ≥ 0
The local spatial form of the Clausius-Planck inequality for a

reversible process takes the form,
Dint := ρθη − ρ r + div q = 0
The local spatial form Clausius-Planck inequality for an
irreversible process takes the form,
Dint := ρθη − ρ r + div q > 0

Clausius-Planck Inequality: Local Spatial Form (II)
Using the local spatial form of the internal energy balance
equation and the Clausius-Planck inequality given by,
ρ e = σ : d + ρ r − div q 

Dint := ρθη − ρ r + div q ≥ 0

ρ r − div q = ρ e − σ : d
the local spatial form of the Clausius-Planck inequality can be
written in terms of the internal energy per unit of mass as,
Dint := σ : d − ρ ( e − θη ) ≥ 0


Clausius-Planck Inequality: Local Spatial Form (III)
Introducing the free energy per unit of mass, denoted as ψ,
defined as,
ψ := e − θη
the local spatial form of the Clausius-Planck inequality can be
written in terms of the free energy per unit of mass as,
Dint := σ : d − ρ (ψ + ηθ ) ≥ 0


Second Law of Thermodynamics: Global Material Form
The global material form of the second law of thermodynamics
states that,
d d
H ( t ) = ∫ ρ0η dV = ∫ ρ0ηdV
dt dt Ω0 Ω
1 1
≥∫ ρ0 r dV − ∫ Q ⋅ N dS
Ω0 θ ∂Ω0 θ


Second Law of Thermodynamics: Global Material Form
Using the divergence theorem, the global material form of the
second law of thermodynamics can be written as,
d
∫
dt Ω0
ρ0η dV = ∫ ρ0ηdV
Ω
1 1
≥∫ ρ0 rdV − ∫ Q ⋅ N dS
Ω0 θ ∂Ω0 θ
1 1 
=∫ ρ0 rdV − ∫ DIV  Q  dV
Ω0 θ Ω0
θ 
1 1 1
=∫ ρ0 rdV − ∫ DIV QdV + ∫ 2
Q ⋅ GRAD θ dV
Ω0 θ Ω0 θ Ω0 θ

Second Law of Thermodynamics: Local Material Form
Localizing, the local material form of the second law of
thermodynamics can be written as,
1 1 1
ρ0η ≥ ρ0 r − DIV Q + 2 Q ⋅ GRAD θ
θ θ θ
Multiplying by the absolute temperature yields,
1
ρ0θη ≥ ρ0 r − DIV Q + Q ⋅ GRAD θ

θ


Clausius-Duhem Inequality: Local Material Form
The local material form of the Clausius-Duhem inequality states
that the dissipation rate per unit of material volume is a non-
negative scalar-valued quantity and it can be written as,
1
D0 := ρ0θη − ρ0 r + DIV Q − Q ⋅ GRAD θ ≥ 0

θ
A stronger assumption is usually considered, yielding,
1
D0 := D0 + D0 = ρ0θη − ρ0 r + DIV Q − Q ⋅ GRAD θ ≥ 0
int cond θ
D0int ≥ 0
D0cond ≥ 0
1
D0 := ρ0θη − ρ0 r + DIV Q ≥ 0, D0 := − Q ⋅ GRAD θ ≥ 0
int cond
θ

Clausius-Planck Inequality: Local Material Form (I)
The local material form of the Clausius-Planck inequality states
that the internal dissipation rate per unit of material volume is a
non-negative scalar-valued quantity and it can be written as,
D0 := ρ0θη − ρ0 r + DIV Q ≥ 0
int
The local material form Clausius-Planck inequality for a

reversible process takes the form,
D0 := ρ0θη − ρ0 r + DIV Q = 0
int
The local material form Clausius-Planck inequality for an

irreversible process takes the form,
D0 := ρ0θη − ρ0 r + DIV Q > 0
int

Clausius-Planck Inequality: Local Material Form (II)
Using the local material form of the internal energy balance
equation and the Clausius-Planck inequality yields,
ρ0 e = τ : d + ρ0 r − DIV Q 

= P : F + ρ0 r − DIV Q 

= S : E + ρ0 r − DIV Q 
D0int
:= ρ0θη − ρ0 r + DIV Q ≥ 0
ρ0 r − DIV Q = ρ0 e − τ : d
= ρ0 e − P : F
= ρ e − S : E
0

Clausius-Planck Inequality: Local Material Form (II)
The local material form of the Clausius-Planck inequality can be
written in terms of the internal energy per unit of mass as,
D0 := τ : d − ρ0 ( e − θη )
int
= P : F − ρ0 ( e − θη )
= S:E − ρ ( e − θη ) ≥ 0
0


Clausius-Planck Inequality: Local Material Form (III)
Introducing the free energy per unit of mass, denoted as ψ,
defined as,
ψ := e − θη
the local material form of the Clausius-Planck inequality can be
written in terms of the free energy per unit of mass as,
D0 := τ : d − ρ0 (ψ + ηθ )
int
(
= P : F − ρ0 ψ + ηθ )
−ρ
= S:E 0 (ψ +ηθ ) ≥ 0
Clausius-Planck Inequality
Local Spatial
Giving the spatial Forms
description of an arbitrary property, the material time derivative of the property can be written as,
Dint := ρθη − ρ r + div q ≥ 0

Dint := σ : d − ρ ( e − θη ) = σ : d − ρ (ψ + ηθ ) ≥ 0
Local Material Forms

D0 := ρ0θη − ρ0 r + DIV Q ≥ 0
int
D0 := τ : d − ρ0 ( e − θη ) = τ : d − ρ0 (ψ + ηθ )

int
= P : F − ρ0 ( e − θη ) = P : F − ρ0 ψ + ηθ ( )
− ρ ( e − θη ) = S : E
= S:E 0
−ρ
0 (ψ +ηθ ) ≥ 0
Balance Laws > Thermodynamic Processes
Adiabatic Process
Adiabatic Process: Local Spatial Form
A thermodynamic process is said to be adiabatic if the net heat
transfer to or from the continuum body is zero.
The internal dissipation rate per unit of spatial volume for an
adiabatic process can be written as,
r = 0, div q = 0 ⇒ Dint := ρθη − ρ r + div q = ρθη ≥ 0
The stress power per unit of spatial volume for an adiabatic

process is equal to the material time derivative of the internal
energy per unit of spatial volume,
Dint := ρ θη = σ : d − ρ ( e − θη ) ≥ 0 ⇒ σ : d = ρ e

Adiabatic Process
Adiabatic Process: Local Material Form (I)
The internal dissipation rate per unit of material volume for an
r = 0, DIV Q = 0 ⇒ D0int := ρ0θη − ρ0 r + DIV Q = ρ0θη ≥ 0
The stress power per unit of material volume for an adiabatic

energy per unit of material volume,
D0 := ρ0 θη = τ : d − ρ0 ( e − θη ) ≥ 0 ⇒ τ : d = ρ0e
int

Adiabatic Process
Adiabatic Process: Local Material Form (II)

D0 := ρ0 θη = P : F − ρ0 ( e − θη ) ≥ 0 ⇒ P : F = ρ0e
int

Adiabatic Process
Adiabatic Process: Local Material Form (III)

D0 := ρ0 θη = S : E − ρ0 ( e − θη ) ≥ 0 ⇒ S : E = ρ0e
int

Isentropic Process
Isentropic Process: Local Spatial Form
A thermodynamic process is said to be isentropic if it takes place
at constant entropy.
isentropic process can be written as,
η = 0 ⇒ Dint := ρθη − ρ r + div q = − ρ r + div q ≥ 0
The stress power per unit of spatial volume for an isentropic
process can be written as,
Dint := − ρ r + div q = σ : d − ρ ( e − θη ) ≥ 0 ⇒
σ : d = ρ e − ρ r + div q
Isentropic Process
Isentropic Process: Local Material Form (I)
η = 0 ⇒ D0 := ρ0θη − ρ0 r + DIV Q = − ρ0 r + DIV Q ≥ 0
int
The stress power per unit of material volume for an isentropic

D0 := − ρ0 r + DIV Q = τ : d − ρ0 ( e − θη ) ≥ 0 ⇒
int
τ : d = ρ0 e − ρ0 r + DIV Q
Isentropic Process
Isentropic Process: Local Material Form (II)
int

D0 := − ρ0 r + DIV Q = P : F − ρ0 ( e − θη ) ≥ 0 ⇒
int
P : F = ρ0e − ρ0 r + DIV Q
Isentropic Process
Isentropic Process: Local Material Form (III)
int

D0 := − ρ0 r + DIV Q = S : E − ρ0 ( e − θη ) ≥ 0 ⇒
int
= ρ e − ρ r + DIV Q
S:E 0 0

Isentropic and Adiabatic Process

Isentropic + Adiabatic Process: Local Spatial Form
A thermodynamic process is said to be isentropic and adiabatic
if it takes place at constant entropy and the net heat flux to or
from the continuum body is zero.
isentropic and adiabatic process is zero and, then, the process is
reversible.
η = 0, r = 0, div q = 0 ⇒ Dint := ρθη − ρ r + div q = 0
The stress power per unit of spatial volume for an isentropic and
adiabatic process is equal to the material time derivative of the
internal energy per unit of spatial volume,
Dint := σ : d − ρ ( e − θη ) = 0 ⇒ σ : d = ρ e
Isentropic/Adiabatic Reversible Process

Isentropic/Adiabatic Reversible Process
If a thermodynamic process is adiabatic and reversible, then the
process is also isentropic,
Dint = 0, r = 0, div q = 0 ⇒ Dint := ρθη − ρ r + div q = 0
If a thermodynamic process is isentropic and reversible, then the
process is also adiabatic,
η = 0, Dint = 0 ⇒ Dint := ρθη − ρ r + div q = 0
For reversible thermodynamic processes, any adiabatic process

is also isentropic, and any isentropic process is also adiabatic .

Isothermal Process
Isothermal Process: Local Spatial Form
A thermodynamic process is said to be isothermal if it takes
place at constant temperature.
isothermal process is given by,
(
θ = 0 ⇒ Dint := σ : d − ρ ψ + ηθ = σ : d − ρψ ≥ 0)

Balance Laws > Governing Equations
Governing Equations: Coupled TM Problem

Governing Equations: Spatial Form
Conservation of mass. Mass continuity
❶ ρ + ρ div v = 0 ❶❸
Balance of linear momentum. Cauchy first motion
❸ div σ + ρ b = ρ v ❾
Balance of angular momentum. Symmetry of Cauchy stress
❸ σ =σT
Balance of energy
❶ ρ e = σ : d + ρ r − div q ❶❸
Clausius-Planck and heat conduction inequalities
Dint := ρ θ η − ρ r + div q ≥ 0, Dcon := −q ⋅ grad θ ≥ 0
❶❶

Constitutive Equations: Spatial Form
Thermo-mechanical constitutive equations
❻ σ = σ ( v, θ , π ) ❶
❶ η = η ( v, θ , π )
Thermal constitutive equation. Fourier law
❸ q = q ( v,θ ) = −k ( v,θ ) grad θ
State equations
❶ e = e ( v, θ )
❶ π = π ( ρ ,θ )


Governing Equations: Material Form (I)
❶ ρ J = ρ0 ❶
❸ DIV P + ρ0 b = ρ0 v ❾❸
Balance of angular momentum. Symmetry restriction 1st P-K
❸ PFT = FPT
Balance of energy
❶ ρ0 e = P : F + ρ0 r − DIV Q ❶❸
D0 := ρ0 θ η − ρ0 r + DIV Q ≥ 0, D0
int cond
:= −Q ⋅ GRAD θ ≥ 0
❶❶

Constitutive Equations: Material Form (I)
❻ PF T
= τ ( v, θ , π ) ❶
❶ η = η ( v, θ , π )
Thermal constitutive equation. Material Fourier law
❸ Q = Q ( v,θ ) = −K ( v,θ ) GRADθ
State equations
❶ e = e ( v, θ )
❶ π = π ( ρ ,θ )


Governing Equations: Material Form (II)
❶ ρ J = ρ0 ❶
❸ ( )
DIV τ F −T + ρ0 b = ρ0 v ❾❸
Balance of angular momentum. Symmetry of Kirchhoff stress
❸ τ =τ T
Balance of energy
❶ ρ0 e = τ : d + ρ0 r − DIV Q ❶❸
D0 := ρ0 θ η − ρ0 r + DIV Q ≥ 0, D0
int cond
:= −Q ⋅ GRAD θ ≥ 0
❶❶

Constitutive Equations: Material Form (II)
❻ τ = τ ( v, θ , π ) ❶
❶ η = η ( v, θ , π )
❸ Q = Q ( v,θ ) = −K ( v,θ ) GRADθ
State equations
❶ e = e ( v, θ )
❶ π = π ( ρ ,θ )


Governing Equations: Material Form (III)
❶ ρ J = ρ0 ❶
❸ DIV ( FS ) + ρ0 b = ρ0 v ❾❸
Balance of angular momentum. Symmetry of 2nd P-K stress
❸ S = ST
Balance of energy
❶ ρ0 e = S : E + ρ0 r − DIV Q ❶❸
D0 := ρ0 θ η − ρ0 r + DIV Q ≥ 0, D0
int cond
:= −Q ⋅ GRAD θ ≥ 0
❶❶

Constitutive Equations: Material Form (III)
❻ τ = τ ( v, θ , π ) ❶
❶ η = η ( v, θ , π )
❸ Q = Q ( v,θ ) = −K ( v,θ ) GRADθ
State equations
❶ e = e ( v, θ )
❶ π = π ( ρ ,θ )

Governing Equations: Mechanical Problem

Mechanical Problem: Spatial Form
❶ ρ + ρ div v = 0 ❶❸
❸ div σ + ρ b = ρ v ❾
❸ σ =σT
Mechanical constitutive equation
❻ σ = σ ( v)


Mechanical Problem: Spatial Form
❶ ρ + ρ div v = 0 ❶❸
❸ div σ + ρ b = ρ v ❻
❻ σ = σ ( v)


Mechanical Problem: Material Form (I)
❶ ρ J = ρ0 ❶
❸ DIV P + ρ0 b = ρ0 v ❾❸
❸ PFT = FPT
❻ PFT = τ ( v )


Mechanical Problem: Material Form (II)
❶ ρ J = ρ0 ❶
❸ ( )
DIV τ F −T + ρ0 b = ρ0 v ❾❸
❸ τ =τ T
❻ τ = τ ( v)


Mechanical Problem: Material Form (II)
❶ ρ J = ρ0 ❶
❸ ( )
DIV τ F −T + ρ0 b = ρ0 v ❻❸
❻ τ = τ ( v)


Mechanical Problem: Material Form (III)
❶ ρ J = ρ0 ❶
❸ DIV ( FS ) + ρ0 b = ρ0 v ❾❸
Balance of angular momentum. Symmetry of 2nd P-K stress
❸ S = ST
❻ S = S ( v)


Mechanical Problem: Material Form (III)
❶ ρ J = ρ0 ❶
❸ DIV ( FS ) + ρ0 b = ρ0 v ❻❸
❻ S = S ( v)

Continuum Mechanics
Chapter 7
Linear Elasticity
Linear Elasticity > Contents
Contents
Chapter 7 · Linear Elasticity
1. Linear elastic model
2. Linear elastic IBVP
3. Solution of the linear elastic IBVP
4. Orthogonal curvilinear coordinates
5. Assignments
6. Linear thermoelastic model
7. Linear thermoelastic IBVP
8. Solution of the linear thermoelastic IBVP
9. Assignments
10. Plane linear elasticity
April 22, 2015 Carlos Agelet de Saracibar 2

Linear Elasticity > Linear Elastic Model
Linear Elastic Model

Infinitesimal Strains Framework
Displacements and gradient of displacements are infinitesimal
No difference between:
o spatial and material configurations
o spatial and material coordinates
o spatial and material descriptions
o spatial and material differential operators
o spatial and material time derivatives
o spatial mass density and material mass density
Linear functions of the gradient of displacements


H1. Adiabatic Processes
We consider that the processes are adiabatic, such that,
r = 0, div q = 0
The internal dissipation rate per unit of volume for an adiabatic
process, given by the Clausius-Planck inequality, can be written
as,
Dint := ρ0θη − ρ0 r + div q = ρ0θη ≥ 0


H2. Isothermal Processes
We consider that the processes are isothermal, such that,
θ = 0
The internal dissipation rate per unit of volume for an isothermal
process, given by the Clausius-Planck inequality, can be written
as,
Dint := ρ0θη − ρ0 r + div q
= σ : ε − ρ0 ( e − θη )
( )
= σ : ε − ρ0 ψ + ηθ = σ : ε − ρ0ψ ≥ 0


H3. Free Energy
We consider that the free energy per unit of volume is a function
of the infinitesimal strain tensor, such that for an isothermal
linear elastic model is a quadratic function of the strain tensor
given by,
1
ρ0ψ = ρ0ψ ( ε ) = ε : C : ε ≥ 0
2
where C is a positive definite symmetric fourth-order tensor,
denoted as isothermal constant elastic constitutive tensor.


As the isothermal elastic constitutive tensor C is a symmetric
fourth-order tensor, the following symmetry condition holds,
ε : C : ε = ε ab Cabcd ε cd = ε ab Ccdabε cd , Cabcd = Ccdab
Furthermore, symmetry of the strain tensor yields the following

symmetry conditions,
ε : C : ε = ε T : C : ε = ε : C : ε T , Cabcd = Cbacd = Cabdc
The symmtery conditions reduce from 81 to 21 the number of

parameters needed to define the elastic constitutive tensor.


Internal Dissipation Inequality
Applying the chain rule, the internal dissipation rate per unit of
volume for an isothermal process, given by the Clausius-Planck
inequality, yields,
Dint := σ : ε − ρ0ψ
∂ψ ( ε )  ∂ψ ( ε ) 
= σ : ε − ρ0 : ε =  σ − ρ0  : ε ≥ 0 ∀ε
∂ε  ∂ε 


Constitutive Equation and Reduced Dissipation
Following Coleman’s method, the internal dissipation rate per
unit of volume for an isothermal process, given by the Clausius-
Planck inequality, must be satisfied for arbitrary thermodynamic
processes, i.e. arbitrary strain rates,
 ∂ψ ( ε ) 
Dint :=  σ − ρ0  : ε ≥ 0 ∀ε
 ∂ε 
yielding the following isothermal linear elastic constitutive
equation and zero internal reduced dissipation rate per unit of
volume, that characterizes a reversible process
∂ψ ( ε )
σ = ρ0 , Dint = 0
∂ε

Constitutive Equation and Reduced Dissipation
The constitutive equation and reduced internal dissipation rate
per unit o volume for an isothermal linear elastic model are given
by,
∂ψ ( ε )
σ = ρ0 = C : ε , σ ab = Cabcd ε cd
∂ε
Dint = 0
Note that, as the internal dissipation rate per unit of volume is

zero for a linear elastic model, any adiabatic process is also
isentropic and any isentropic process is also adiabatic.


Elastic Constitutive Tensor
The isothermal elastic constitutive tensor for a linear elastic
model is given by,
∂ 2ψ ( ε ) ∂ 2ψ ( ε )
C = ρ0 , Cabcd = ρ0
∂ε ⊗ ∂ε ∂ε ab ⊗ ∂ε cd

Isotropic Linear Elastic Model

Isotropic Elastic Constitutive Tensor
A continuum medium is said to be isotropic if at a given point it
has the same material properties in any direction.
For an isotropic linear elastic material model, the physical
property of isotropy is translated into a mathematical property
of isotropy of the isothermal elastic constitutive tensor, yielding,
C = λ 1 ⊗ 1 + 2µ Iˆ , Cabcd = λδ abδ cd + µ (δ acδ bd + δ ad δ bc )
where λ ≥ 0, µ > 0, denoted as isothermal Lamé parameters, are

the two material parameters characterizing an isothermal isotro-
pic linear elastic material model.
Isotropy reduces from 21 to 2 the number of parameters needed
to define the elastic constitutive tensor.

Free Energy for an Isotropic Linear Elastic Model
The free energy per unit of volume for an isothermal isotropic
linear elastic material model can be written as,
1 1 2
ρ0ψ ( ε ) = ε : C : ε = λ ( tr ε ) + µε : ε ≥ 0
2 2


Isotropic Linear Elastic Constitutive Equation
The isothermal constitutive equation for an isotropic linear
elastic material model can be written as,
σ = C:ε
( )
= λ 1 ⊗ 1 + 2µ Iˆ : ε = λ (1 : ε ) 1 + 2µε
= λ ( tr ε ) 1 + 2µε
σ = λ ( tr ε ) 1 + 2µε , σ ab = λε dd δ ab + 2µε ab


The volumetric part of the constitutive equation for an isother-
mal isotropic linear elastic model can be written as,
tr σ = σ : 1 = λ ( tr ε ) 1 : 1 + 2µε : 1 = ( 3λ + 2µ ) tr ε
Introducing the mean stress, volumetric deformation and iso-
thermal bulk modulus, given by,
1 2
σ m := tr σ , e := tr ε , K := λ + µ > 0
3 3
The volumetric part of the constitutive equation for an isother-
mal isotropic linear elastic model can be written as,
σ m = Ke

The deviatoric part of the constitutive equation for an isotropic
linear elastic model can be written as,
dev σ = λ ( tr ε ) dev 1 + 2µ dev ε = 2µ dev ε
Introducing the shear modulus given by,

G := µ > 0
The deviatoric part of the constitutive equation for an isotropic
dev σ = 2G dev ε


Isotropic Linear Elastic Inverse Constitutive Equation
The inverse constitutive equation for an isothermal isotropic
−1
tr ε = ( 3λ + 2µ ) tr σ
λ 1
ε =− ( tr σ ) 1 + σ
( 3λ + 2µ ) 2µ 2µ


Let us introduce the isothermal Young elastic modulus, denoted
as E > 0, isothermal Poisson’s coefficient, denoted as 0 ≤ ν ≤ 1 2,
and isothermal bulk modulus, denoted as K > 0 , such that,
µ ( 3λ + 2µ ) λ 1
E= > 0, 0 ≤ ν = ≤
λ+µ 2 (λ + µ ) 2
νE E
λ= ≥ 0, µ = G = >0
(1 +ν )(1 − 2ν ) 2 (1 +ν )
2 E
K =λ+ µ = >0
3 3 (1 − 2ν )


The inverse constitutive equation for an isothermal isotropic
ν 1 +ν ν 1 +ν
ε = − ( tr σ ) 1 + σ , ε ab = − σ dd δ ab + σ ab
E E E E


Using engineering notation, the Cartesian components of the
inverse constitutive equation for an isothermal isotropic linear
elastic model can be written as,
1 1
(
ε x = σ x −ν (σ y + σ z ) , γ xy = τ xy
E G
)
1 1
ε y = (σ y −ν (σ x + σ z ) ) , γ xz = τ xz
E G
1 1
(
ε z = σ z −ν (σ x + σ y ) , γ yz = τ yz
E G
)


Example 7.1
Let us consider an uniaxial traction test of an isotropic linear
elastic material such that,
σ x > 0, σ y = σ z = τ xy = τ xz = τ yz = 0
y
σx σx
x
1) Obtain the components of the strain tensor

2) Consider as particular cases: (a) ν = 0 ; (b) ν = 1 2


Example 7.1
Given a stress state such that,
σ x > 0, σ y = σ z = τ xy = τ xz = τ yz = 0
the components of the strain tensor take the form,
1
E
( (1
))
1
ε x = σ x −ν σ y + σ z = σ x > 0, γ xy = τ xy = 0
E G
1 ν 1
εy =
E
( ( ))
σ y −ν σ x + σ z = − σ x ≤ 0, γ xz = τ xz = 0
E G
ν
εz =
1
E
( ( ))1
σ z −ν σ x + σ y = − σ x ≤ 0, γ yz = τ yz = 0
E G


(a) Consider the case ν = 0 :
1 ν ν
ε x = σ x > 0, ε y = − σ x = 0, ε z = − σ x = 0
E E E
There is no Poisson effect.
(b) Consider the case ν = 1 2 :

1 ν 1 ν 1
ε x = σ x > 0, ε y = − σ x = − σ x ≤ 0, ε z = − σ x = − σ x ≤ 0
E E 2E E 2E
1 1 1
e = tr ε = ε x + ε y + ε z = σ x − σx − σx = 0
E 2E 2E
The volumetric strain is zero, the volume is preserved, characte-
rizing an incompressible isotropic linear elastic material model.
Linear Elasticity > Linear Elastic IBVP
Isotropic Linear Elastic IBVP

Governing Equations
Let us consider the following governing equations in the space x
time domain Ω× I = Ω× [ 0,T ].
Linear momentum balance · First Cauchy’s motion equation
∂ 2u ∂ 2 ua
div σ + ρ0b = ρ0 2 , σ ab,b + ρ0ba = ρ0 2 in Ω× I
∂t ∂t
Isothermal isotropic linear elastic constitutive equation
σ = λ ( tr ε ) 1 + 2µε , σ ab = λ ε dd δ ab + 2µε ab in Ω× I
Geometrical equations
1 1
s
2
( T
2
)
ε = ∇ u = ∇ ⊗ u + ( ∇ ⊗ u ) , ε ab = ( ua ,b + ub,a ) in Ω× I


Boundary Conditions
Let us consider prescribed displacements and prescribed trac-
tions boundaries, denoted as ∂ u Ω and ∂σ Ω, respectively, such
that,
∂ u Ω ∪ ∂σ Ω = ∂Ω, ∂ u Ω ∩ ∂σ Ω = ∅
with the following bounday conditions for the IBVP:
Prescribed displacements boundary conditions
u = u, u a = u a on ∂ u Ω× I
Prescribed tractions boundary conditions
t = σ n = t , ta = σ ab nb = ta on ∂σ Ω× I


Initial Conditions
Let us consider the following initial conditions for the IBVP,
Initial displacements at time t=0
u t =0 = 0, ua t =0 = 0 in Ω
Initial velocities at time t=0
∂u ∂ua
= v0 , = v0a in Ω
∂t t =0 ∂t t =0


Find the displacements, strains and stresses in Ω× I = Ω× [ 0,T ]
such that the following equations are satisfied:
∂ 2u 
div σ + ρ0b = ρ0 2 
∂t 
σ = λ ( tr ε ) 1 + 2µε  in Ω× I
1 
2
(
ε = ∇ su = ∇ ⊗ u + ( ∇ ⊗ u ) 
T

)
u t =0 = 0 
u = u on ∂ u Ω× I 
∂u  in Ω
t = σ n = t on ∂σ Ω× I = v0 
∂t t =0 

Unicity of the Solution of the Linear Elastic IBVP
The solution of the linear elastic IBVP is unique in strains and
stresses.
The solution of the linear elastic IBVP is unique in displacements
if the boundary conditions are such that arbitrary rigid motions
are not allowed.

Isotropic Linear Elastic BVP

Quasistatic Linear Elastic BVP
A problem is said to be quasistatic if the acceleration term in the
first Cauchy’s motion equation is negligible.
A quasistatic linear elastic BVP does not involves any time
derivative, hence no time integration is involved and initial
conditions are not needed anymore.
Actions on the continuum body (forces, boundary conditions)
may still be a function of time and, then, the response
(displacements, strains, stresses) will be also a function of time.
Here the time plays the role and may be viewed as a loading
parameter.

Isotropic Linear Elastic BVP

Quasistatic Isotropic Linear Elastic BVP
div σ + ρ0b = 0 

σ = λ ( tr ε ) 1 + 2µε  in Ω× I
s 
ε =∇ u 
u = u on ∂ u Ω× I
t =σn = t on ∂σ Ω× I

Linear Elasticity > Solution of the Linear Elastic IBVP
Solution of the Isotropic Linear Elastic IBVP

∂ 2u 
div σ + ρ0b = ρ0 2 
∂t 
σ = λ ( tr ε ) 1 + 2µε  in Ω× I
1 
2
(
ε = ∇ su = ∇ ⊗ u + ( ∇ ⊗ u ) 
T

)
u t =0 = 0 
u = u on ∂ u Ω× I 
∂u  in Ω
t = σ n = t on ∂σ Ω× I = v0 
∂t t =0 

Method of Displacements: Navier Equation
Stresses can be removed from the list of unknowns of the
isothermal isotropic linear elastic IBVP using,
div σ = div ( λ ( tr ε ) 1 + 2µε )

= λ grad ( tr ε ) + 2µ div ε
σ n = ( λ ( tr ε ) 1 + 2µε ) n
= λ ( tr ε ) n + 2µε n


Strains can can be removed from the list of unknowns of the
isothermal isotropic linear elastic IBVP using,
div σ = λ grad ( tr ε ) + 2µ div ε
= ( λ + µ ) grad ( div u ) + µ div ( grad u )
σ n = λ ( tr ε ) n + 2µε n
(
= λ ( div u ) n + µ grad u + ( grad u ) n
T
)


The first Cauchy’s motion equation written in terms of the displa-
cements is denoted as Navier equation and takes the form,
∂ 2u
( λ + µ ) grad ( div u ) + µ div ( grad u ) + ρ0b = ρ0 2 in Ω × I
∂t
∂ 2 ua
( λ + µ ) ub,ba + µua ,bb + ρ0ba = ρ0 2 in Ω× I
∂t


Find the displacements in Ω× I = Ω× [ 0,T ] such that the
following equations are satisfied:
∂ 2u
( λ + µ ) grad ( div u ) + µ div ( grad u ) + ρ0b = ρ0 2 in Ω × I
∂t
( T
λ ( div u ) n + µ grad u + ( grad u ) n = t on ∂σ Ω× I )
∂u
u t =0 = 0, = v0 in Ω
∂t t =0
Solution of the Isotropic Linear Elastic BVP

Quasistatic Isotropic Linear Elastic BVP
Find the displacements in Ω× I = Ω× [ 0,T ] such that the
following equations are satisfied:
( λ + µ ) grad ( div u ) + µ div ( grad u ) + ρ0b = 0 in Ω × I
( T
λ ( div u ) n + µ grad u + ( grad u ) n = t on ∂σ Ω× I )

Solution of the Isotropic Linear Elastic BVP

While the displacements have been obtained, strains and
stresses may be obtained as a post-process of the results,
Geometrical equations
1 1
s
2
( T
2
)
ε = ∇ u = ∇ ⊗ u + ( ∇ ⊗ u ) , ε ab = ( ua ,b + ub,a ) in Ω× I
Isothermal isotropic linear elastic constitutive equations
σ = λ ( tr ε ) 1 + 2µε , σ ab = λ ε dd δ ab + 2µε ab in Ω× I

Linear Elasticity > Orthogonal Curvilinear Coordinates
Orthogonal Curvilinear Coordinates

Cylindrical and Spherical Coordinates

Orthogonal Curvilinear Coordinates

General Remarks
The components of a vector or higher order tensor are equal
to their components in a local cartesian basis linked to the
orthogonal curvilinear coordinates.
The components of the Nabla operator have to be obtained
for each orthogonal curvilinear coordinates system.
Operations involving the Nabla operator, such as grad, div,
curl, in general, cannot be obtained as a matrix operation in
terms of their components, as it happens in cartesian
coordinates, i.e. for the divergence of a vector,
T
∇⋅ u ≠ [∇] [u]

Cylindrical Coordinates

Vector position
x = x ( r ,θ , z ) = r cos θ eˆ x + r sin θ eˆ y + z eˆ z
Physical basis
∂x 
= cos θ eˆ x + sin θ eˆ y 
∂r

∂x 
= −r sin θ eˆ x + r cos θ eˆ y 
∂θ 
∂x 
= ez
ˆ 
∂z 

Euclidean norms of the physical basis vectors
∂x ∂x ∂x
= 1, = r, = 1,
∂r ∂θ ∂z
Local orthonormal basis vectors

eˆ r (θ ) = cos θ eˆ x + sin θ eˆ y 

eˆθ (θ ) = − sin θ eˆ x + cos θ eˆ y 

eˆ z = eˆ z 

Nabla differential operator
∂ 1 ∂ ∂
∇ = er + eθ + e z
∂r r ∂θ ∂z
T
∂ 1 ∂ ∂
[∇ ] = 
 ∂r r ∂θ ∂z 

Navier Equations
∂e 2G ∂ωz ∂ωθ ∂ 2 ur 
( λ + 2G ) − + 2G + ρ0br = ρ0 2 
∂r r ∂θ ∂z ∂t 
1 ∂e ∂ωr ∂ωz ∂ 2uθ 
( λ + 2G ) − 2G + 2G + ρ0bθ = ρ0 2 
r ∂θ ∂z ∂r ∂t 
∂e 2G ∂ (ωθ r ) 2G ∂ωr ∂ 2u z 
( λ + 2G ) − + + ρ0bz = ρ0 2 
∂z r ∂r r ∂θ ∂t 

Navier Equations
1 ∂ ( rur ) 1 ∂uθ ∂u z
e = div u = + +
r ∂r r ∂θ ∂z
1  1 ∂u z ∂uθ  
ωr = −Ωθ z =  −  
2  r ∂θ ∂z  
1  ∂ur ∂uz  
ωθ = −Ω zr =  −  
2  ∂z ∂r  
1  1 ∂ ( ruθ ) 1 ∂ur  
ωz = −Ωrθ =  − 
2  r ∂r r ∂θ  

Components of the Strain Tensor
 ε rr ε rθ ε rz 
[ε ] = ε rθ ε θθ εθ z 

 ε rz εθ z ε zz 
 ∂ur 1  1 ∂ur ∂uθ uθ  1  ∂ur ∂u z  
  + −   + 
∂r 2  r ∂θ ∂r r  2  ∂z ∂r  
 
 1  1 ∂ur ∂uθ uθ  1 ∂uθ ur 1  ∂uθ 1 ∂u z  
=  + −  +  + 
2  r ∂θ ∂r r  r ∂θ r 2  ∂z r ∂θ  
 
 1  ∂ur ∂u z  1  ∂uθ 1 ∂u z  ∂u z 
 2  ∂z ∂r 
+ 

+ 
 
 2 ∂z r ∂θ ∂z 
Spherical Coordinates

Vector position
x = x ( r ,θ , φ ) = r sin θ cos φ eˆ x + r sin θ sin φ eˆ y + r cos θ eˆ z
Physical basis
∂x 
= sin θ cos φ eˆ x + sin θ sin φ eˆ y + cos θ eˆ z 
∂r 
∂x 
= r cos θ cos φ eˆ x + r cos θ sin φ eˆ y − r sin θ eˆ z 
∂θ 
∂x 
= − r sin θ sin φ e x + r sin θ cos φ e y
ˆ ˆ 
∂φ 
Euclidean norms of the physical basis vectors
∂x ∂x ∂x
= 1, = r, = r sin θ ,
∂r ∂θ ∂φ
Local orthonormal basis vectors

eˆ r (θ , φ ) = sin θ cos φ eˆ x + sin θ sin φ eˆ y + cos θ eˆ z 

eˆθ (θ , φ ) = cos θ cos φ eˆ x + cos θ sin φ eˆ y − sin θ eˆ z 

ˆeφ (φ ) = − sin φ eˆ x + cos φ eˆ y 

Nabla differential operator
∂ 1 ∂ 1 ∂
∇ = er + eθ + eφ
∂r r ∂θ r sin θ ∂φ
T
∂ 1 ∂ 1 ∂ 
[∇ ] = 
 ∂r r ∂θ r sin θ ∂φ 

Navier Equations
∂e 2G ∂ ( ωφ sin θ ) 2G ∂ωθ ∂ 2 ur 
( λ + 2G ) − +  + ρ 0br = ρ 0 2
∂r r sin θ ∂θ r sin θ ∂φ ∂t 
1 ∂e 2G ∂ωr 2G ∂ ( rωφ sin θ ) ∂ 2uθ 
( λ + 2G ) − + + ρ b
0 θ
= ρ 0 2 
r ∂θ r sin θ ∂φ r sin θ ∂r ∂t 
1 ∂e 2G ( θ ) 2G ∂ωr
∂ ω r ∂ 2
uφ 
( λ + 2G ) − + + ρ 0bφ = ρ 0 2 
r sin θ ∂φ r ∂r r ∂θ ∂t  

Navier Equations
1  ∂ ( r 2
ur sin θ ) ∂ ( ruθ sin θ ) ∂ ( ruφ ) 
e = div u = 2  + + 
r sin θ  ∂r ∂θ ∂φ 
1 1 ∂ ( uφ sin θ ) 1 ∂uθ  
ωr = −Ωθφ =    −
2  r sin θ ∂θ r sin θ ∂φ  

1  1 ∂ur 1 ∂ ( ruφ )  
ωθ = −Ωφ r =  −  

2  r sin θ ∂φ r ∂r  

1  1 ( θ ) 1 ∂ur 
∂ ru
ωφ = −Ω rθ =  −  
2  r ∂r r ∂θ  

Components of the Strain Tensor
 ε rr ε rθ ε rφ 
 
[ ] ε rθ
ε = εθθ ε θφ 
ε rφ εθφ ε φφ 
∂ur 1  1 ∂ur ∂uθ uθ 

ε rr = ε rθ =  + − 
∂r 2  r ∂θ ∂r r 
1 ∂uθ ur 1  1 ∂ur ∂uφ uφ 
εθθ = + ε rφ =  + − 
r ∂θ r 2  r sin θ ∂φ ∂r r 
1 ∂uφ uθ ur 1  1 ∂uθ 1 ∂uφ uφ 
εφφ = + cot θ + εθφ =  + − cot φ 
r sin θ ∂φ r r 2  r sin θ ∂φ r ∂θ r 
Linear Elasticity > Assignments
Assignment 7.1
Assignment 7.1
A uniform compression pressure, denoted as p, is applied to the
lateral surface of the cylinder of radius R of the figure. The mate-
rial of the cylinder is assumed to be isotropic linear elastic with
Lamé constants λ = µ . There is an horizontal rigid surface on the
top of the cylinder at a distance a<<R.
Body forces and frictional effects are a
neglected. Quasistatic conditions are

assumed.
p p h
(1) Plot p vs δ, where δ is the radial
displacement of the lateral surface
of the cylinder. A
A
(2) Plot p vs σ z at point A.
R R
Assignment 7.1
Assignment 7.1
Boundary conditions will depend on the value of the applied
pressure. Let us consider the following problems,
Problem 1: The cylinder is not in contact with the top horizon-
tal surface. BC on the top surface of the cylinder are zero
tractions. The range of values of the pressure is 0 ≤ p ≤ p *
where p * is the pressure needed for the cylinder just to make
contact (without contact pressure) with the top horizontal
surface.
Problem 2: The cylinder is in contact with the top horizontal
surface. BC on the top surface of the cylinder are zero
incremental vertical displacements (measured from the
displacements at the end of Problem 1) and p ≥ p *.
Assignment 7.1
Problem 1 Problem 2
0 ≤ p ≤ p* p* ≤ p
0 ≤ ∆p = p − p *
a
u ∆u
h h
p ε p
∆p ∆ε ∆p
σ ∆σ
A A
R R R R

Assignment 7.1
Problem 1
We consider the following steps:
Step 1. System of coordinates.
Taking into accoun the geometry of the problem, we select a
cylindrical system of coordinates to solve the problem.

Assignment 7.1
Step 2. Hypotheses on the displacements
Using cylindrical coordinates the displacement field takes the
form,
T
[u] = ur ( r,θ , z ) uθ ( r ,θ , z ) uz ( r ,θ , z ) 
Taking into account the axial symmetry of the geometry, loading
and BC we consider a displacement field of the form,
T
[u] = ur ( r, z ) 0 uz ( r , z ) 
Furthermore, taking into account that the pressure is uniform,
and there is frictionless contact we introduce the following
additional hypotheses, yielding a displacement field of the form,
T
[u] = ur ( r ) 0 u z ( z ) 
Assignment 7.1
Step 3. Navier equations in cylindrical coordinates.
Taking into account quasi-static conditions and body forces
negligible,
∂e 2G ∂ωz ∂ωθ
( λ + 2G ) − + 2G + ρ0br = 0
∂r r ∂θ ∂z
1 ∂e ∂ωr ∂ω z
( λ + 2G ) − 2G + 2G + ρ0bθ = 0
r ∂θ ∂z ∂r
∂e 2G ∂ (ωθ r ) 2G ∂ωr
( λ + 2G ) − + + ρ 0bz = 0
∂z r ∂r r ∂θ

Assignment 7.1
Taking into account the hypotheses introduced on the displace-
ment field,
1 ∂ ( rur ) 1 ∂uθ ∂u z 1 ∂ ( rur ) ∂u z
e = div u = + + = +
r ∂r r ∂θ ∂z r ∂r ∂z
1  1 ∂u z ∂uθ 
ωr = −Ωθ z =  − =0
2  r ∂θ ∂z 
1  ∂ur ∂u z 
ωθ = −Ω zr =  − =0
2  ∂z ∂r 
1  1 ∂ ( ruθ ) 1 ∂ur 
ωz = −Ω rθ =  − =0
2  r ∂r r ∂θ 
Assignment 7.1
Then, the (non-trivial) Navier equations take the form,
∂e ∂  1 ∂ ( rur ( r ) ) ∂u z ( z )  ∂  1 ∂ ( rur ( r ) ) 
=  + =  =0
∂r ∂r  r ∂r ∂z  ∂r  r ∂r 

∂e ∂  1 ∂ ( rur ( r ) ) ∂u z ( z )  ∂  ∂u z ( z ) 
=  + =   =0
∂z ∂z  r ∂r ∂z  ∂z  ∂z 
Integrating the Navier equations yields,
1
ur ( r ) = A1r + A2 , u z ( z ) = A3 z + A4
r

Assignment 7.1
Step 4. Boundary conditions.
The BC on displacements for Problem 1 read,
1
ur ( r ) r = 0 = A1r + A2 =0 ⇒ A2 = 0
r r =0
u z ( z ) z =0 = A3 z + A4 z =0
=0 ⇒ A4 = 0
Substituting into the equations of the components of the
displacement yields,
ur ( r ) = A1r , u z ( z ) = A3 z

Assignment 7.1
The components of the strain tensor take the form,
 A1 0 0
 
[ε ] =  0 A1 0
 0 0 A3 
Taking into account that λ = µ , the components of the stress
tensor take the form,
 4 A1 + A3 0 0 
[σ ] = λ  0 4 A1 + A3 0 

 0 0 2 A1 + 3 A3 
Assignment 7.1
The BC on the traction vector for Problem 1, conveniently
writting the non-trivial component in terms of the components
of the stress tensor yields,
σr r=R
= λ ( 4 A1 + A3 ) = − p
σz z =h
= λ ( 2 A1 + 3 A3 ) = 0
Solving the system of two equations, yields,

3p p
A1 = − , A3 =
10λ 5λ

Assignment 7.1
Step 5. Solution of Problem 1
The solution in displacements, strains and stresses for Problem 1
takes the form,
p T
[u] = [ −3r 0 2 z ]
10λ
 −3 0 0  − p 0 0
p  ,
[ε ] = 0 −3 0 [σ ] =  0 − p 0
10λ  
 0 0 2   0 0 0 

Assignment 7.1
The value of the pressure for which the cylinder comes into
contact with the top surface is given by,
p* 5λ a
uz ( z ) z =h = h=a ⇒ p* =
5λ h
Then, the range of values of the pressure for the solution of the
Problem 1 is given by,
5λ a
0 ≤ p ≤ p* =
h

Assignment 7.1
Step 6. Curve p-δ.
The curve p-δ for Problem 1 is given by,
3p 10λ
δ := ur ( r ) r = R =− R ⇒ p=− δ
10λ 3R
The stress σ zA for Problem 1 is zero for any value of p (within the
range of values of p defining Problem 1).

Assignment 7.1
Problem 2
We have to use an incremental formulation and we may use the
results obtained in Steps 1-3 from Problem 1, yielding an
increment of displacements,
1
∆ur ( r ) = B1r + B2 , ∆u z ( z ) = B3 z + B4
r

Assignment 7.1
The BC on displacements for Problem 2 (imposed on the
reference or undeformed configuration) read,
1
∆ur ( r ) r =0 = B1r + B2 = 0 ⇒ B2 = 0
r r =0
∆u z ( z ) z =0 = B3 z + B4 z =0
= 0 ⇒ B4 = 0
∆u z ( z ) z = h = B3 z z = h = 0 ⇒ B3 = 0
Substituting into the equations of the components of the

displacement yields,
∆ur ( r ) = B1r , ∆u z ( z ) = 0
Assignment 7.1
The components of the incremental strain tensor take the form,
 B1 0 0
0 0 
[ ] 
∆ε = B1
 0 0 0 
Taking into account that λ = µ , the components of the
incremental stress tensor take the form,
 4 B1 0 0 
[ ∆σ ] = λ  0 4 B1 0 
 0 0 2 B1 
Assignment 7.1
The BC on the incremental traction vector for Problem 2,
conveniently writting the non-trivial component in terms of the
component of the incremental stress tensor yields,
∆p
∆σ r r=R
= λ ( 4 B1 ) = −∆p ⇒ B1 = −
4λ

Assignment 7.1
The incremental solution in displacements, strains and stresses
for Problem 2 takes the form,
∆p T
[ ∆u ] = − [ r 0 0]
4λ
1 0 0  2 0 0
∆p  , ∆p  
[ ]
∆ε = − 0 1 0 [ ]
∆σ = − 0 2 0
4λ   2  
0 0 0   0 0 1 

Assignment 7.1
Step 6. Curves Δp-Δδ and ∆σ zA - Δδ
The curve Δp-Δδ for Problem 2 is given by,
∆p 4λ
∆δ := ∆ur ( r ) r = R =− R ⇒ ∆p = − ∆δ
4λ R
The curve ∆σ zA - Δδ for Problem 2 is given by,

∆p
A
∆σ := −
z ⇒ ∆p = −2∆σ zA
2

Assignment 7.1
p ∆p
4λ
∆p = − ∆δ
R
5λ a
p=
h 10λ −∆δ
p=− δ
3R
3Ra −δ
−δ =
2h

Assignment 7.1
p ∆p
∆p = −2∆σ zA
5λ a
p=
h −∆σ zA
−σ zA

Assignment 7.2
Consider a sphere A with radius R1=R and a spherical crown B
with external radius R2=2R. The two spheres are of the same iso-
tropic linear elastic material (Young modulus E>0 and Poisson’s
coefficient ν=0). There is a small uniform gap a<<R between
them. A uniform pressure p is applied on the external surface of
the spherical crown B.
Body forces and frictional effects B p
between the two spheres can be A

neglected. Quasi-static conditions R 2
R 1
can be assumed.

Assignment 7.2
1) Obtain the value of the pressure p for which the two bodies
come into contact.
2) Plot the curve pressure-delta, where delta is defined as
minus the radial displacement of the external surface of the
spherical crown B.

Linear Elasticity > Plane Linear Elasticity
Plane Linear Elasticity

For some problems, i.e. geometry, loading and BCs, one of the
principal stress directions is known “a priori”, it does not play a
role in the formulation, and the problem may be formulated on
the plane defined by the other two principal directions.
Considering a local Cartesian axis, taking the z-axis as the one

associated to the “a-priori” known principal direction, the plane
linear elastic problem may be formulated on the x,y plane.
Variables associated to the z-axis, are either known or they can
be obtained as a post-process of the 2D results.


Two cases of plane linear elastic problems may be considered,
Plane stress problems
Plane strain problems
For each one of them:

(i) We will introduce the hypothesis on the appropriate field,
(ii) We will consider under which conditions those hypothesis
are suitable
(iii) We will obtain the consequences on the remaining fields,
(iv) We will show up some typical examples

Plane Stress Linear Elasticity

Hypothesis on the Stress Field
H1. The components of the stress tensor take the form,
σ x τ xy 0
[σ ] = τ xy σ y 0 
 0 0 0 
H2. The non-zero components of the stress tensor take the form,
σ x = σ x ( x, y ) , σ y = σ y ( x, y ) , τ xy = τ xy ( x, y )


Suitability of the Plane Stress Hypothesis
The thickness is much smaller
than a typical dimension on
the plane of analysis (e << L)
The actions (body forces,
prescribed displacements
and prescribed tractions),
are contained in the plane of
analysis and are independent of the third dimension
The prescribed traction is different than zero, only on the
contour of the thickness of the body


Strain Field
Using the inverse constitutive equations, the strains take the
form,
1 1
ε x = (σ x −νσ y ) , γ xy = τ xy
E G
1
ε y = (σ y −νσ x ) , γ xz = 0
E
1
ε z = − ν (σ x + σ y ) , γ yz = 0
E
ν
εz = −
1 −ν
( ε x +εy )

Strain Field
Then, the strain field takes the form,
 εx 1
γ xy
2
0
1 
[ε ] =  2 γ xy εy 0
 0 0 ε z 

such that,
ε x = ε x ( x, y ) , ε y = ε y ( x, y ) , γ xy = γ xy ( x, y )
and the following relation holds,
ν
ε z ( x, y ) = −
1 −ν
( ε ( x, y ) + ε ( x, y ) )
x y


Plane Stress Constitutive Equation
The constitutive equation for plane stress linear elasticity may be
written in matrix form as,
 
σ x  1 ν 0  ε x 
  E   
σ y  = 2 
ν 1 0  ε y 
τ  1 −ν  1 − ν  γ 
 xy  0 0   xy 
 2 


Displacement Field
The in-plane components of the displacement field are obtained
integrating the in-plane components of the strain field,
∂ux ∂u y ∂ux ∂u y
ε x ( x, y ) = , ε y ( x, y ) = , γ xy ( x, y ) = +
∂x ∂y ∂y ∂x
yielding,
ux = ux ( x, y ) , u y = u y ( x, y )


Displacement Field
The out-of-plane component of the displacements is obtained
integrating the out-of-plane component of the strain field,
ν ∂u z
ε z ( x, y ) = −
1 −ν
( ε x ( x, y ) + ε y ( x , y ) ) =
∂z
yielding,
u z = u z ( x, y , z )


Displacement Field
The plane stress linear elastic problem should be viewed as an
ideal problem, which is not necessarily a particular case of the
3D linear elastic problem.
In particular, the following conditions are not necessarily
satisfied,
∂ux ( x, y ) ∂uz ( x, y, z )
γ xz = + =0
∂z ∂x
∂u y ( x, y ) ∂uz ( x, y, z )
γ yz = + =0
∂z ∂y


Examples
t* Γ σe : t *
Γ σ− : t * = 0
b
Γ σ+ : t * = 0
y superficie media
Plate loaded on its midplane Deep beam

Plane Strain Linear Elasticity

Hypothesis on the Displacement Field
H1. The components of the displacement field take the form,
T
[u] = ux uy 0
H2. The non-zero components of the displacements take the
form,
u x = u x ( x, y ) , u y = u y ( x, y )


Suitability of the Plane Strain Hypothesis
The body to be studied can be generated by
translating the plane of analysis (geometry, forces,
BCs) along the third dimension
The body forces, prescribed displacement
and prescribed tractions are
contained on the plane of analysis
and are independent of the third
dimension
In the central section, the
z-displacement is zero, and the
derivatives of the x-displacement and
y-displacement with respect to z are zero due to symmetry

Strain Field
Using the geometric equations, strains take the form,
∂ux ∂u y ∂ux ∂u y
ε x ( x, y ) = , ε y ( x, y ) = , γ xy ( x, y ) = +
∂x ∂y ∂y ∂x
∂u z
εz = =0
∂z
∂ux ∂uz ∂u y ∂uz
γ xz = + = 0, γ yz = + =0
∂z ∂x ∂z ∂y


Strain Field
Then the strain field takes the form,
 εx 1
2
γ xy 0
1 
[ε ] =  2 γ xy ε y 0
 0 0 0 

such that,
ε x = ε x ( x, y ) , ε y = ε y ( x, y ) , γ xy = γ xy ( x, y )


Stress Field
Using the constitutive equations for linear elasticity, the stresses
take the form,
σ x = λ ( ε x + ε y ) + 2µε x = ( λ + 2G ) ε x + λε y τ xy = G γ xy
σ y = λ ( ε x + ε y ) + 2µε y = ( λ + 2G ) ε y + λε x τ xz = G γ xz = 0
σ z = λ (ε x + ε y ) τ yz = G γ yz = 0

λ
σz =
2(λ + µ )
( σ x + σ y ) = ν (σ x + σ y )


Stress Field
Then, the stress field takes the form,
σ x τ xy 0 
[σ ] = τ xy σ y 0 
 0 0 σ z 
such that,
σ x = σ x ( x, y ) , σ y = σ y ( x, y ) , τ xy = τ xy ( x, y )
λ
σ z = σ z ( x, y ) =
2(λ + µ )
( σ x + σ y ) = ν (σ x + σ y )


Plane Strain Constitutive Equation
The constitutive equation for plane strain linear elasticity may be
written in matrix form as,
 ν 
 1 1 −ν
0 
σ x    ε x 
  E (1 −ν )  ν  
σ y  = 1 − ν 1 0  ε y 
τ  (1 + ν )(1 − 2ν )
 xy    γ xy 
 0 1 − 2ν 
0
 2(1 −ν ) 
λ
σ z = σ z ( x, y ) =
2(λ + µ )
( σ x + σ y ) = ν (σ x + σ y )

Examples
+∞
+∞
Sección transversal
Sección
transversal
Sección transversal

Plane Linear Elastic IVBP

First Cauchy Motion Equation
The first Cauchy motion equation for a plane linear elastic IBVP,
either plane stress or plane strain linear elastic IBVP, may be
written as,
∂σ x ∂τ xy ∂ 2u x 
+ + ρ0bx = ρ0 2 
∂x ∂y ∂t 
2  in Ω × I
∂τ xy ∂σ y ∂ uy
+ + ρ0by = ρ0 2 
∂x ∂y ∂t 


Plane Linear Elastic Constitutive Equation
The constitutive equations for a plane linear elastic IBVP, either
plane stress or plane strain linear elastic IBVP, may be written as,
σ x  1 ν 0  εx 
  E   
σ y  = 2 ν 1 0  ε y 
τ  1 − ν
 xy 
 0 0 (1 −ν ) 2  γ xy 
 E
E = E  E = 1 −ν 2
Plane stress:  Plane strain: 
ν = ν ν =
ν
 (1 −ν )

Kinematic or Geometric Equations
The kinematic or geometric equations for a plane linear elastic
IBVP, either plane stress or plane strain linear elastic IBVP, may
be written as,
∂u x ∂u y ∂u x ∂u y
εx = , εy = , γ xy = +
∂x ∂y ∂y ∂x


Boundary Conditions
Boundary conditions for a plane linear elastic IBVP, either plane
stress or plane strain linear elastic IBVP, may be written as,
1. Prescribed displacements
u x  u *x 
  =  *  on ∂ u Ω × I
u y  u y 
2. Prescribed tractions
σ x τ xy   nx   t x ( x, y , t )
*

τ    = *  on ∂σ Ω × I
 xy σ y   n y   t y ( x, y , t ) 
 

Initial Conditions
Initial conditions for a plane linear elastic IBVP, either plane
stress or plane strain linear elastic IBVP, may be written as,
1. Initial displacements
u = 0 in Ω
2. Initial velocities
∂u
= v 0 in Ω
∂t


Post-process of In-plane Results
While the solution for the plane linear elastic IBVP, either plane
stress or plane strain linear elastic IBVP, has been obtained the
following post-process of results yields,
1. For plane stress linear elastic IBVP (assuming u z z =0 = 0 )
ν ν
εz = −
1 −ν
( ε x + ε y ) , uz = −
1 −ν
( ε x +εy )z
2. For plane strain linear elastic IBVP

λ
σz =
2(λ + µ )
( σ x + σ y ) = ν (σ x + σ y )

Plane Stress States Curves

Traditionally, plane stress states have been graphically represen-
ted using contour lines, such as:
Isostatics (or stress trajectories)
Isoclines
Isobars
Maximum shear lines


Isostatics
The isostatics are defined as the envelops of the principal stress
vector fields.
The isostatics are two orthogonal family of curves which are
tangent to the principal stress axes at each point:
o Isostatics of sigma1: curves
which are tangent to the
largest principal stress vector
field, sigma1
o Isostatics of sigma2: curves
which are tangent to the smallest
principal stress vector field, sigma2


Singular Points and Neutral Points
Singular points: A singular point is characterized by the
following plane stress state:
σ x = σ y , τ xy = 0
Neutral points: A neutral point is a particular case of a
singular point, and is characterized by the following plane
stress state:
σ x = σ y = 0, τ xy = 0
In a singular point, any direction is a principal direction and the
isostatics tend to loose their regularity and can exhibit an abrupt
change of direction


Differential Equation of the Isostatics
The differential equation of the isostatics is obtained as,
2 tan α 2τ xy 2 y′
y′ = tan α , tan ( 2α ) = = =
1 − tan α σ x − σ y 1 − y′2
2
2
σ x −σ y
y′ + y′ − 1 = 0
τ xy

Differential Equation of the Isostatics
The solution of the second-order equation given by,
2
σ x −σ y
y′ + y′ − 1 = 0
τ xy
yields the differential equation of the isostatics, corresponding
to the two orthogonal families of isostatics, given by,
2
σ x −σ y  σ x −σ y 
y′ = − ±   +1
2τ xy  2τ 
 xy 
The integration of the differential equations of the isostatics
yields the algebraic equation of the isostatics.

Isoclines
The isoclines are defined as the locus of the points along which
the principal directions of the stresses have the same directions.
The principal stress vectors in each point of an isocline are paral-

lel to each other, forming an angle θ with the x-axis.
The isoclines can be obtained using photoelastic methods.

Equation of the Isoclines
The equation of the isoclines forming an angle θ can be
obtained setting that the largest principal stress vector must
form an angle α=θ with the x-axis, yielding,
2τ xy
tan ( 2α ) = tan ( 2θ ) =
σ x −σ y
Once the isoclines for different angles θ have been obtained,

the principal stress vector directions are known at a cloud of
points of the body and the isostatics can be obtained as the
envelopes of the principal stress vector directions.


Isobars
The isobars are defined as the locus of points along which the
principal stresses (sigma1 or sigma2) are constant.
The isobars depend on the value, not on the direction, of the

principal stresses.
At each point there are two families of isobars, one for sigma1
and one for sigma2.

Equation of the Isobars
The equations of the isobars for a fixed value of sigma1 = c1 and
sigma2 = c2 are given by,
2
σx +σ y  σ x −σ y  2
σ1 = +   + τ xy = c1
2  2 
2
σx +σ y  σ x −σ y  2
σ2 = −   + τ xy = c2
2  2 


Maximum Shear Lines
The maximum shear lines are defined as the envelops of the
maximum shear stress (in modulus) vector fields.
The maximum shear lines are two orthogonal families of curves
(tau_max and tau_min) which form 45º with the isostatics.


Differential Equation of the Maximum Shear Lines
 π
y′ = tan β = tan  α − 
 4


The differential equation of the maximum shear lines can be
obtained setting,
2 tan β 2 y′
tan ( 2β ) = =
1 − tan β 1 − y′2
2
 π 1 σ x −σ y
tan ( 2β ) = tan  2α −  = − =−
 2 tan ( 2α ) 2τ xy
2
4τ xy
y′ − y′ − 1 = 0
σ x −σ y


The solution of the second order equation given by,
2
4τ xy
y′ − y′ − 1 = 0
σ x −σ y
yields the differential equation of the maximum shear lines,
corresponding to the two orthogonal families of maximum shear
lines, given by,
2
2τ xy  2τ xy 
y′ = ±   + 1
σ x −σ y  σ −σ
 x y 
The integration of the differential equations of maximum shear
lines yields the algebraic equations of maximum shear lines.
Continuum Mechanics
Chapter 8
Plasticity
Plasticity > Contents
Contents
Chapter 8 ∙ Plasticity
1. Introduction
2. 1D Plasticity model
3. 3D Plasticity model
4. Yield surfaces
Plasticity > Introduction
Introduction
Introduction
Plasticity may be characterized by the following main features,
 Nonlinear stress‐strain relationship
 Lack of unicity in the stress‐strain relationship
 Presence of irreversible strains, i.e. plastic strains, in a
loading‐unloading cycle
Here we will consider plasticity within an infinitesimal strains

framework.
Introduction
Space of Principal Stresses
Consider a Cartesian system framework defined by the three
principal stresses, defining the space of principal stresses, such
that the stress state at an arbitrary point of a continuum body,
characterized by the principal stresses  1 ,  2 ,  3 , may be repre‐
sented by a point P  1 ,  2 ,  3  in the space of principal stresses.
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Hydrostatic Stress Axis
The hydrostatic stress axis is defined by the collection of points
belonging to the space of principal sresses that satisfy the
condition  1   2   3   m .
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Octahedral Plane
An octahedral plane is defined as a plane that is orthogonal to
the hydrostatic stress axis. The equation of an octahedral plane
containing the point P is given by,
 1   2   3  constant
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Unit Normal to the Octahedral Plane
n is
The unit normal vector to the octahedral plane, denoted as ,
given by,
1
n  1 1 1
T
3
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Space of Principal Stressess
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Octahedral Normal Stress
 oct is a scalar‐valued
The octahedral normal stress, denoted as ,
quantity defined as the projection of the vector position OP
along the unit normal n,

3 oct : OP  n  OA sgn  m
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Octahedral Normal Stress

3 oct : OP  n  OA sgn  m  3 m
1 1
 oct :  m  I1  tr 
3 3
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Octahedral Normal Stress Surface
The domain in the space of principal stresses satisfying the
condition I1  constant , is the octahedral plane at a distance d of
the origin given by,
3
d  3  oct  3m  I1
3
Introduction
Pure Deviatoric Stress States
Pure deviatoric stress states, satisfying  oct  0 , are characterized
in the space of principal stresses by points located on the
octahedral plane which contains the origin, satisfying the condi‐
tion I1  0 ,
3
 oct  0  d  3  oct  3m  I1  0
3
Introduction
Octahedral Shear Stress
 oct is a positive scalar‐
The octahedral shear stress, denoted as ,
valued quantity defined as,
3 oct : AP
3
1   2   3   m
n
A  m ,  m ,  m  P  1 ,  2 ,  3 
2
O
1
Introduction
Octahedral Shear Stress
 2  2  2
 
2
3 oct  AP  OP  OA
        1   2   3 
2 21 2 2
1 2 3
3
 2 J 2
2
 oct   J 2 
12
Introduction
Octahedral Shear Stress Surface
The domain in the space of principal stresses satisfying the
condition J 2  constant, is the cylinder with axis the hydrostatic
stress axis and radius given by,
R  3  oct  2  J 2 
12
Introduction
Spherical Stress States
Spherical stress states, satisfying  oct  0, are characterized in the
space of principal stresses by points located along the hydros‐
tatic stress axis, satisfying the condition J 2  0 ,
 oct  0  R  3  oct  2  J 2   0
12
Introduction
Principal Stress Space
A stress state at a point of a continuum medium may be
characterized in the principal stress space by a point with
coordinates the principal stresses.
Alternatively, the point in the principal stress space may be
characterized by the stress invariants I1 , J 2 , J 3 :
 The first invariant of the stress tensor fixes
I1 a given
octahedral plane at a distance d  3I1 3 of the origin.
 The second invariant of the deviatoric stress tensor fixes J 2 a
cirumference on the octahedral plane with radius R  2  J 2  
12
and center on the point of the hydrostatic stress axis.

J 3
 The third invariant of the deviatoric stress tensor fixes a
point of the circumference.
Assignment 8.1
Assignment 8.1
Determine the shape of the surface which in the principal stress
space is given by,
aI12  bJ 2  c a, b, c  0
Assignment 8.2
Assignment 8.2
The shape of a surface in the principal stress space is an axisym‐
metric ellipsoide along the hydrostatic stress axis. The center of
the elliposide is at the origin. Semi‐axes along the hydrostatic
stress axis and on the deviatoric plane (octahedral plane passing
through the origin) are a and b, respectively. Obtain the equation
of the surface in terms of the appropriate stress invariants.
Plasticity > 1D Plasticity Model
1D Perfect Plasticity Model

1D Rate‐independent Perfect Plasticity Model
H1. Additive split of the infinitesimal strain
The infinitesimal strain may be additively split into an elastic and
a plastic part,
  e  p
Similarly, the infinitesimal strain rate can be additively split into

an elastic strain rate and a plastic strain rate part,
   e   p

H2. Linear elastic response
The elastic strain satisfies the linear elastic constitutive equation,
   ,   E e  E     p 
e 1
E
where E is the elastic Young’s modulus.
Similarly the elastic strain rate satisfies the linear elastic rate
constitutive equation,
   ,   E e  E     p 
e 1
E

H3. Space of admissible stresses, elastic domain and yield
surface
The space of admissible stresses, elastic domain and yield
surface are defined, respectively, as,
  :    f    0

int     :    f    0 
  :    f    0
f   given by,
where the yield function, denoted as , is
f   :    Y  0
 Y  0 the yield stress of the material.
and is

H4. Plastic flow rule
Plastic flow rule provides the evolution equation for the plastic
strain such that,
if    Y  0   p    0
if    Y  0   p    0
where the plastic multiplier or plastic consistency parameter,
  0 a non‐negative scalar valued quantity.
denoted as , is
Plastic flow rule can be cast into a single expression yielding,
if f    0   p   sgn      f  

H5. Kuhn‐Tucker loading/unloading conditions
Plastic loading and elastic loading/unloading may be characteri‐
zed by the following conditions,
if  0  f    0
if f    0    0
The above conditions may be cast into the so called Kuhn‐Tucker
loading/unloading conditions, yielding,
  0, f    0,  f    0

H6. Plastic consistency condition
Plastic consistency may be characterized by the following condi‐
tions,
if f    0 and   0  f    0
if f    0 and f    0    0
The above conditions may be cast into the so called plastic con‐
sistency conditions, yielding,
if f    0    0, f    0,  f    0

Plastic consistency parameter
The non‐trivial value of the plastic multiplier or plastic consis‐
tency parameter can be obtained imposing the plastic consisten‐
cy condition for the non‐trivial case of plastic loading, yielding,
if f    0 and   0  f    0
Taking the time derivative of the yield function yields,
f     f     sign     sign   E     p 
 sign   E  sign   E sign     sign   E  E
 sign    trial  E  0
  E 1 sign    trial  E 1f    trial ,  trial : E

Plastic loading/elastic unloading from the yield surface can be
determined in terms of the projection of the trial stress rate
along the unit normal to the yield surface,
Elastic unloading:   0

 f     0   f     f    trial  0
trial
      0
f

Plastic loading:    E 1 f    trial  0

 f     0   f     f    trial  E  0
 trial
      0
f


The plastic multiplier or plastic consistency parameter can be
cast into a single expression, covering both the (non‐trivial)
plastic loading and the (trivial) elastic unloading cases, yielding,
  E 1  f    trial   0
1D Plasticity Model
1D Rate‐independent Hardening Plasticity Model
H1. Additive split of the infinitesimal strain
The infinitesimal strain may be additively split into an elastic and
a plastic part,
  e  p
   e   p
H2. Isotropic and kinematic hardening variables
Isotropic and kinematic hardening variables in the strain space
 ,
are denoted, respectively as . Their conjugate variables in
q, q
the stress space are denoted, respectively, as, .
1D Plasticity Model
H3. Linear elastic and linear hardening response
The elastic strain satisfies the linear elastic constitutive equation,
   ,   E e  E     p 
e 1
E
   ,   E e  E     p 
e 1
E
where E is the elastic Young’s modulus.
The isotropic/kinematic hardening variables satisfy the following
linear constitutive hardening equations,
q   H  , q   K
where H and are the
K constant isotropic and kinematic
hardening parameters, respectively.
1D Plasticity Model
H5. Associative plastic flow rule
Associative plastic flow rule equations may be written as,
 p    f  , q, q    sgn   q 
    q f  , q, q   
    q f  , q, q    sgn   q 
Kuhn‐Tucker loading/unloading conditions may be written as,
  0, f  , q, q   0,  f  , q, q   0
1D Plasticity Model
Plastic consistency conditions may be written as,
if f  , q, q   0    0, f  , q, q   0,  f  , q, q   0
1D Plasticity Model
if f  , q, q   0 and   0  f  , q, q   0
f   f    q f q   q f q
  f E     p    q f H    q f K
  f E     f E  f   q f H  q f   q f K  q f   0
1D Plasticity Model
Introducing the trial stress and computing the derivatives of the
yield function with respect to the stress and hardening variables
in the stress space, yields,
 trial : E ,  trial : E
 f  sgn   q  ,  q f  1,  q f   sgn   q 
f   f   trial    E  H  K   0
   E  H  K   f  trial  0
1
1D Plasticity Model
along the unit normal to the yield surface,

 f   trial  0   f   f   trial  0
   0
 f
Plastic loading:    E 1 f   trial  0

 f   trial  0   f   f   trial  E  0
   0
f

1D Plasticity Model
   E  H  K   f  , q, q    trial   0
1
1D Plasticity Model
Continuum elastoplastic tangent modulus
For the (non‐trivial) case of plastic loading, the stress rate and
continuum elastoplastic tangent modulus are given by,
  E e  E  E p  E  E  f
  trial  E  E  H  K   f  trial  f
1
1

  trial  E  E  H  K   trial  1  E  E  H  K 
1
  trial

 E 1 E  E  H  K   1
 : E ep
 E 1  E  E  H  K  
ep 1
E
1D Plasticity Model
Differential constitutive equation
For elastic loading/unloading cases the differential constitutive
equation may be written as,
if f  0 or f  0 and f  0  d  Ed 
For the plastic loading case the differential constitutive equation

may be written as,
if f  0 and f  0  d  E ep d 
Assignment 8.3
Assignment 8.3
The bar‐truss structure of the figure consists of three bars of the
same length L. The three of them are modeled as an elastic‐
perfect plastic material model with elastic Young’s modulus E,
yield stresses  0 in traction and in compression.
10  0

B C
L L
P
e
45º 45º 
O

L 10 e
Assignment 8.3
Assignment 8.3
An increasing vertical load P is applied at the point O, until the
vertical displacement of point O, denoted as δ, takes the value,
0
  20 l
E
Then the struture is fully unloaded, until the load is P=0.
1) Plot the curve P vs δ for the full loading‐unloading cycle.
Indicate the state of the bars (elastic or plastic) during the
full loading‐unloading cycle.
2) Compute the remaining value of the vertical displacement of
point O, denoted as δ, at the end of the loading‐unloading
cycle.

H1. Additive split of the infinitesimal strain tensor
The infinitesimal strain tensor may be additively split into an
elastic strain tensor and a plastic strain tensor,
  e  p
   e   p
H2. Linear elastic response
The linear elastic constitutive equation may be written as,
   :  e   :    p 
   :  e   :     p 
where  is the fourth‐order constant elastic constitutive tensor.

H3. Space of admissible stresses, elastic domain and yield
surface
The space of admissible stresses, elastic domain and yield
surface are defined, respectively, as,
  :    f    0

int     :    f    0 
  :    f    0
f   a scalar‐valued
where the yield function, denoted as , is
function of the stress tensor which may be viewed as a surface in
the space of principal stresses.

H4. Associative plastic flow rule
Associative plastic flow rule, defining the evolution of the plastic
strain tensor, may be written as,
 p    f  
where   0 is the plastic multiplier or plastic consistency
parameter.
Kuhn‐Tucker loading/unloading conditions may be written as,
  0, f    0,  f    0

Plastic consistency condition may be written as,
if f    0    0, f    0,  f    0

if f    0 and   0  f    0
f   f :   f :  :     p    f :  :      f :  : f 
  f :  trial     f :  : f   0
The plastic multiplier or plastic consistency parameter is given
by,
    f :  : f   f : trial  0
1

tensor along the unit normal to the yield surface,

 f :  trial  0   f   f :  trial  0
  f  0

Plastic loading:      f :  : f 1  f : trial  0

 f :  trial  0   f   f :  trial     f :  : f   0
  f  0


    f :  :  f 
1
 f : trial  0

Continuum elastoplastic tangent tensor
For the (non‐trivial) case of plastic loading, the stress rate and
fourth‐order continuum elastoplastic tangent tensor are given
by,
   :  e   :     p    trial    :  f
  trial    f :  : f    :  f    f : trial 
1

     f :  :  f 
1
    
 :  f   f :   : 
 :  ep
: 
 :     f :  : f    :  f     f :  
ep 1

Differential constitutive equation
For elastic loading/unloading cases the differential constitutive
equation may be written as,
if f  0 or f  0 and f  0  d   : d 
For the plastic loading case the differential constitutive equation

may be written as,
if f  0 and f  0  d  ep : d 
Assignment 8.4
Assignment 8.4
Consider the clamped beam of the figure. The material is elastic‐
perfect plastic, with elastic Young’s modulus E, Poisson’s coeffi‐
0
cient   0 and traction/compression yield stress . Body
forces are negligible. The displacements field is assumed to be,
u   z   x  , v  0, w    x 
z, w y, v
M
x, u
h m
Assignment 8.4
Assignment 8.4
1) Show that the displacements field is the solution of the elas‐
tic problem, determine the function   x  and obtain the
strains and stresses.
2) Obtain the maximum elastic moment M e and the correspon‐
ding vertical displacement  e at the end of the beam.
3) Obtain the moment M  M e as a function of the plastified
zone size in an arbitrary transversal section of the beam.
4) Obtain the maximum moment (limit moment) M p when all
the points of the transversal section of the beam have
plastified.
5) Plot the curve M vs δ for 0     .
Plasticity > Yield Surfaces
Von Mises Yield Surface
The Von Mises yield surface may be written as,
3
f    dev    Y  0
2
The Von Mises yield surface, in terms of the stress invariants,
may be written as.
f    F  J 2    3J 2    Y  0
12
In the space of principal stresses the Von Mises yield surface is a
cylinder with axis the hydrostatic stress axis and radius R given
by,
2
R : 3 oct  2  J 2   Y
12
The Von Mises yield surface has the following features:
 Does not depends on the first invariant of the stress tensor
and therefore the intersection of any octahedral plane with
the yield surface gives the same circumference.
 For a positive yield stress, i.e. positive radius, hydrostatic
stress states will never plastify
 Does not depends on the third invariant of the deviatoric
stress tensor and therefore is an axisymmetric surface
 Appropriate for metals, i.e. materials having the same
behaviour in traction and compression
Example 8.1
The stress state at an arbitrary section of a beam is given by,
 x  xy 0
    xy 0 0
 0 0 0
Obtain the expresison of the yield surface using the Von Mises
model.
Example 8.1
 x  xy 0
    xy 0 0
 0 0 0
The mean stress and deviatoric stress tensor are given by,
 23  x  xy 0 
1 1  
 m  tr    x ,  dev      xy  13  x 0 
3 3
 0 0  13  x 

The second invariant of the deviatoric stress tensor is given by,
1
J 2  dev  : dev 
2
14 2 1 2 1 2 2 2  1 2 2
   x   x   x   xy   xy    x   xy
29 9 9  3
The Von Mises yield surface takes the form,
f     3J 2    Y   x2  3 xy2   Y  0
12
Tresca Yield Surface

The Tresca yield surface may be written as,
f    2 max   Y   1   3    Y  0
The Tresca yield surface, in terms of the stress invariants, takes

the form,
f    F  J 2 , J 3   0

In the space of principal stresses, the Tresca yield surface is an
hexagonal prismatic surface with axis the hydrostatic stress axis
and inscribed in the cylinder of Von Mises with radius,
2
R : 3 oct  2  J 2   Y
12

Von Mises and Tresca Yield Surfaces

The Tresca yield surface has the following features:
 Does not depends on the first invariant of the stress tensor
and therefore the intersection of any octahedral plane with
the yield surface gives the same hexagon.
 For a positive yield stress, hydrostatic stress states will never
plastify
 Depends on the third invariant of the deviatoric stress tensor
and therefore it is not axisymmetric surface
 Appropriate for metals, i.e. materials having the same
behaviour in traction and compression

Example 8.2
 x  xy 0
 0
   xy 0
 
 0 0 0
Obtain the expresison of the yield surface using the Tresca
model.

Example 8.2
 x  xy 0
 0
   xy 0
 
 0 0 0
The principal stresses are given by,
1 1 2 2 1 1 2 2
 1   x   x   xy ,  3   x   x   xy
2 4 2 4
The Tresca yield surface takes the form,
f     1   3   x2  4 xy2   Y

The second invariant of the deviatoric stress tensor is given by,
1
J 2  dev  : dev 
2
14 2 1 2 1 2 2 2  1 2 2
   x   x   x   xy   xy    x   xy
29 9 9  3
The Von Mises yield surface takes the form,
f     3J 2    Y   x2  3 xy2   Y  0
12
Mohr‐Coulomb Yield Surface
Moh‐Coulomb Yield Surface
The Mohr‐Coulomb yield surface may be written as,
f     n   n tan   c  0
The normal and shear stresses at the yield surface may be
written in terms of the principal stresses, yielding,
1 1 1
 n   1   3    1   3  sin  ,  n   1   3  cos 
2 2 2
The Mohr‐Coulomb yield surface may be written as,
f     1   3    1   3  sin   2c cos   0
The Mohr‐Coulomb yields surface, in terms of the stress

invariants, takes the form,
F    F  I1 , J 2 , J 3   0
In the space of principal stresses, the Mohr‐Coulomb yield surfa‐
ce is an hexagonal pyramid surface with axis the hydrostatic
stress axis and vertex V on the positive side (traction) of the
hydrostatic axis, at a distance d of the origin given by,
V  c cot  , c cot  , c cot   , d  3c cot 
The Tresca yield surface may be obtained as a particular case of
the Mohr‐Coulomb yield surface setting,
1
  0, c   Y
2
f     1   3    1   3  sin   2c cos 
 2 m sin   2c cos   0
f     1   3   2c   1   3    Y  0
 According to the Mohr‐Coulomb plasticity model, positive
hydrostatic stress states may plastify, but negative hydrostatic
stress states will remain always elastic.
 The Mohr‐Coulomb plasticity model may be suitable for
frictional‐cohesive materials, exhibiting a rather different
behaviour under traction/compression stress states, such as
soils, rocks, concrete.
Drucker‐Prager Yield Surface

The Drucker‐Prager yield surface may be written as,
3
f    3 oct   oct    0
2
where
2sin  6c cos 
 , 
3  3  sin   3  3  sin  
The Drucker‐Prager yield surface, in terms of the stress
invariants, may be written as,
f     I1  J 21 2    0

In the space of principal stresses, the Drucker‐Prager yield surfa‐
ce is an conical surface with axis the hydrostatic stress axis and
vertex V on the positive side (traction) of the hydrostatic axis, at
a distance d of the origin given by,
V  c cot  , c cot  , c cot   , d  3c cot 

The Von Mises yield surface may be obtained as a particular case
of the Drucker‐Prager yield surface setting,
1
  0, c   Y
2
2sin  6c cos  Y
  0   0,   0  
3  3  sin    0 c  Y 2 3  3  sin   3
 0
cY 2
Y
f     I1  J     J  
2
12
2
12
0  3J 21 2   Y  0
3

 According to the Drucker‐Prager plasticity model, positive
hydrostatic stress states may plastify, but negative hydrostatic
stress states will remain always elastic.
 The Drucker‐Prager plasticity model may be suitable for
frictional‐cohesive materials, exhibiting a rather different
behaviour under traction/compression stress states, such as
soils, rocks, concrete.
Yield Surfaces
Von Mises
f    dev    Y   3J 2    Y  0
3 12
2
Tresca
f    2 max   Y   1   3    Y  0
Mohr-Coulomb
f     1   3    1   3  sin   2c cos   0
Drucker-Prager
f    3 oct  32 oct     I1  J 21 2    0

Continuum Mechanics
Chapter 9
Fluid Mechanics
Fluid Mechanics > Contents
Contents
Chapter 9 · Fluid Mechanics
1. Introduction
2. Constitutive equations

Fluid Mechanics > Introduction
Introduction
Hydrostatic Pressure
There exist experimental evidence that the stress state of a fluid
at rest is hydrostatic and it is characterized by a spherical stress
tensor given by,
σ = − p0 1
where p0 > 0 is a positive scalar-valued quantity denoted as
hydrostatic pressure.
The traction vector for a fluid at rest, at a given spatial point, is

the same on any arbitrary plane with unit normal n, and is given
by a compression state along the unit normal,
t = σ n = − p0 1n = − p0n
Introduction
Mean Pressure
The mean pressure, denoted as p, is a scalar-valued quantity
defined as minus the mean stress,
1
p := −σ m = − tr σ
3
For a fluid at rest, the mean pressure is equal to the hydrostatic

pressure,
1 1
p := −σ m = − tr σ = − tr ( − p0 1) = p0
3 3

Introduction
Thermodynamic Pressure
The thermodynamic pressure, denoted as p, is a scalar-valued
quantity, that satisfies the following kinetic state equation,
F ( ρ , p, θ ) = 0
For a fluid at rest, the hydrostatic pressure satisfies the kinetic

state equation and, therefore, it is equal to the thermodynamic
pressure yielding,
p = p = p0
For a fluid in motion the three pressures would be different,
p ≠ p, p ≠ p0 , p ≠ p0
Introduction
Barotropic Fluid
A fluid is said to be barotropic if the kinetic state equation does
not depends on the temperature.
The kinetic state equation for a barotropic fluid may be written
as,
F ( ρ, p) = 0 ⇒ ρ = ρ ( p)
A particular case of barotropic fluid is the incompressible fluid.

The kinetic state equation for an incompressible fluid may be
written as,
F ( ρ ) = 0 ⇒ ρ = ρ0

Fluid Mechanics > Constitutive Equations
Constitutive Equations
Constitutive Equation for Stokes Fluids
The constitutive equation for a Stokes fluid may be written as,
σ = − p1 + f ( d, p,θ ) , σ ab = − pδ ab + f ab ( d, p,θ )
Ideal fluid:
f ( d, p,θ ) = 0
Newtonian fluid:
f ( d, p,θ ) = K 0 ( p, θ ) I1 ( d ) 1 + K1 ( p, θ ) d

Constitutive Equation for Stokes Fluids
Quasi-Newtonian fluid:
f ( d, p,θ ) = K 0 ( I1 ( d ) , I 2 ( d ) , I 3 ( d ) , p, θ ) 1
+ K1 ( I1 ( d ) , I 2 ( d ) , I 3 ( d ) , p, θ ) d
Reiner-Rivlin fluid:
f ( d, p,θ ) = K 0 ( I1 ( d ) , I 2 ( d ) , I 3 ( d ) , p, θ ) 1
+ K1 ( I1 ( d ) , I 2 ( d ) , I 3 ( d ) , p, θ ) d
+ K 2 ( I1 ( d ) , I 2 ( d ) , I 3 ( d ) , p, θ ) dd

Constitutive Equation for Isotropic Newtonian Fluids
The constitutive equation for an isotropic Newtonian fluid may
be written as,
σ = − p1 + λ ( p,θ )( tr d ) 1 + 2µ ( p,θ ) d
where λ ( p,θ ) , µ ( p,θ ) ≥ 0 are two scalar-valued functions
denoted as dynamic viscosities.

Dynamic Viscosities for Isotropic Newtonian Fluids
The dynamic viscosity µ ( p ) ≥ 0 for an isotropic Newtonian fluid
may be written as,
 p
µ ( p ) = µ0 exp   ≥ 0
B
where µ0 ≥ 0 is a thermodynamic pressure-independent viscosity
coefficient and B is a constant, usually bigger enough such that
the dynamic viscosity coefficient µ = µ ( p ) ≥ 0 can be
considered as thermodynamic pressure-independent.

The dynamic viscosity µ (θ ) ≥ 0 for an isotropic Newtonian fluid
may be written as,
 Q 
µ (θ ) = µ0 exp  ≥0
 Rθ 
where µ0 ≥ 0 is a thermodynamic pressure-independent viscosity
coefficient, Q is an activation energy and R is the universal
constant of ideal gases.

The dynamic viscosity µ (θ ) ≥ 0 for an isotropic Newtonian fluid
may be written as,
 Q 
µ (θ ) = µ0 exp   = µ0 exp ( −α (θ − θ0 ) ) ≥ 0
 Rθ 
where the following parameters have been introduced as,
 Q  Q
µ0 = µ0 exp   ≥ 0, α = − 2
 Rθ0  Rθ0

Dynamic Viscosities for Quasi-Newtonian Fluids
Power law model. The dynamic viscosity µ (θ ) ≥ 0 for a Quasi-
Newtonian fluid may be written as,
n −1
µ ( I 2 ( d ) ) = K0 ( 4I 2 ( d ) ) 2
where K0 is the consistency parameter and n is the rate sensiti-

vity coefficient.

Dynamic Viscosities for Quasi-Newtonian Fluids
Carreau model. The dynamic viscosity µ (θ ) ≥ 0 for a Quasi-
n −1
µ ( I 2 ( d ) ) = µ0 (1 + 4λ 2 I 2 ( d ) ) 2 , 0 < n <1
where µ0 is the constant dynamic viscosity parameter, λ is a

model parameter and n is the rate sensitivity coefficient.

Fluid Mechanics > Governing Equations
Governing Equations
Governing Equations
Conservation of mass. Mass continuity equation
❶ ρ + ρ div v = 0 ❶❸
Balance of linear momentum. Cauchy first motion equation
❸ div σ + ρ b = ρ v ❾
❸ σ =σT
Balance of energy
❶ ρ e = σ : d + ρ r − div q ❶❸
❶❶
Governing Equations
Thermo-mechanical constitutive equation for the stresses
❻ σ = − p1 + f ( d, p,θ ) ❶
Thermo-mechanical constitutive equation for the entropy
❶ η = η ( d, p,θ )
❸ q = q ( v,θ ) = −k ( v,θ ) grad θ
Caloric state equation
❶ e = e ( ρ ,θ )
Kinetic state equation
❶ ρ = ρ ( p, θ )
Governing Equations
Mechanical Problem
❶ ρ + ρ div v = 0 ❶❸
❸ div σ + ρ b = ρ v ❾
❸ σ =σT
Mechanical constitutive equation (temperature independent)
❻ σ = − p1 + f ( d, p ) ❶
Kinetic state equation for a barotropic fluid
❶ ρ = ρ ( p)
Governing Equations
Mechanical Problem
❶ ρ + ρ div v = 0 ❶❸
❸ div σ + ρ b = ρ v ❻
❻ σ = − p1 + f ( d, p ) ❶
❶ ρ = ρ ( p)

Governing Equations
Mechanical Problem
❶ ρ + ρ div v = 0 ❶❸
❸ − grad p + div f ( d, p ) + ρ b = ρ v ❶
❶ ρ = ρ ( p)

Governing Equations
Incompressible Mechanical Problem
❶ div v = 0 ❸
❸ div σ + ρ0 b = ρ0 v ❻
❻ σ = − p1 + f ( d, p ) ❶

Governing Equations
❶ div v = 0 ❸
❸ − grad p + div f ( d, p ) + ρ0 b = ρ0 v ❶

Governing Equations
Thermal Problem
Balance of energy
❶ ρ e = σ : d + ρ r − div q ❶❸
❶❶
❶ η = η ( d, p,θ )
❸ q = q ( v,θ ) = −k ( v,θ ) grad θ
❶ e = e ( ρ ,θ )
Governing Equations
Thermal Problem
Balance of energy
❶ ρ e = σ : d + ρ r + div k ( v,θ ) grad θ ( ) ❶❶
Dint := ρ θ η − ρ r − div ( k ( v,θ ) grad θ ) ≥ 0,
❶
Dcon := grad θ ⋅ k ( v,θ ) grad θ ≥ 0
❶ η = η ( d, p,θ )
❶ e = e ( ρ ,θ )

Governing Equations
Incompressible Thermal Problem
Balance of energy
❶ ρ0 e = σ : d + ρ0 r + div k ( v,θ ) grad θ ( ) ❶❶
Dint := ρ0 θ η − ρ0 r − div ( k ( v,θ ) grad θ ) ≥ 0,
❶
❶ η = η ( d, p,θ )
❶ e = e (θ )

Continuum Mechanics
Chapter 10
Ideal Fluids
Ideal Fluids > Contents
Contents
Chapter 10 ∙ Ideal Fluids
1. Constitutive equation
3. Stationary flow
4. Transient curl‐free flow
May 15, 2015 Carlos Agelet de Saracibar 2
Ideal Fluids > Constitutive Equation
Constitutive Equation
Constitutive Equation for Ideal Fluids
The constitutive equation for an ideal fluid may be written as,
   p1,  ab   p ab
Ideal Fluids > Governing Equations
Governing Equations
Governing Equations
 Conservation of mass. Mass continuity
❶    div v  0 ❶❸
 Balance of linear momentum. Cauchy’s first motion
❸ div    b   v ❾
 Balance of angular momentum. Symmetry of Cauchy stress
❸  T
 Balance of energy
❶  e   : d   r  div q ❶❸
 Clausius‐Planck and heat conduction inequalities
int :      r  div q  0, con : q  grad   0
❶❶
Governing Equations
 Thermo‐mechanical constitutive equation for the stresses
❻    p1 ❶
 Thermo‐mechanical constitutive equation for the entropy
❶     d, p, 
 Thermal constitutive equation. Fourier’s law
❸ q  q  v,   k  v,  grad 
 Caloric state equation
❶ e  e   , 
 Kinetic state equation
❶     p,  
Governing Equations
Mechanical Problem
❶    div v  0 ❶❸
❸ div    b   v ❾
 Balance of angular momentum. Symmetry of Cauchy stress
❸  T
 Mechanical constitutive equation (thermal independent)
❻    p1 ❶
 Kinetic state equation for a barotropic fluid
❶     p
Governing Equations
Mechanical Problem
❶    div v  0 ❶❸
❸ div    b   v ❻
 Mechanical constitutive equation
❻    p1 ❶
❶     p
Governing Equations
Mechanical Problem
❶    div v  0 ❶❸
 Balance of linear momentum. Euler equation
❸  grad p   b   v ❶
❶     p
Governing Equations
❶ div v  0 ❸
❸ div   0 b  0 v ❻
 Mechanical constitutive equation
❻    p1 ❶
Governing Equations
❶ div v  0 ❸
❸  grad p  0 b  0 v ❶
Governing Equations
Thermal Problem
❶  e   p div v   r  div q ❶❸
int :      r  div q  0, con : q  grad   0
❶❶
❶     d, p, 
 Thermal constitutive equation. Fourier’s law
❸ q  q  v,   k  v,  grad 
❶ e  e   , 
Governing Equations
Thermal Problem
❶  e   p div v   r  div k  v,  grad    ❶❶
int :      r  div  k  v,  grad    0,
❶
con : grad   k  v,  grad   0
❶     d, p, 
❶ e  e   , 
Governing Equations
❶ 0 e  0 r  div k  v,  grad   ❶❶
int : 0    0 r  div  k  v,  grad    0,
❶
con : grad   k  v,  grad   0
❶     d, p, 
❶ e  e  
Governing Equations
Euler Equation
The first Cauchy’s motion equation for an ideal fluid, arising from
the linear momentum balance law, is denoted as Euler equation,
and takes the form,
 grad p   b   v
Writting the material time derivative of the velocity in terms of
the local time derivative and convective time derivative, yields,
 v 
 grad p   b      grad v  v 
 t 
1 v
 grad p  b    grad v  v
 t
Governing Equations
The convective rate of the velocity may be written as,
 grad v  v   grad v   grad v   v   grad v 

T T
v
 2wv   grad v  v
T
1
 2  v   v
2
Let us assume that body forces derive from a potential such that,

b   grad 
Governing Equations
Let us assume that there exists a function of the thermodynamic
pressure such that,
dp dp 1
P  p   , dP  p   , grad P  p   grad p
  p   p   p
If the fluid is incompressible, then the function of the thermody‐

namic pressure can be explicitly integrated yielding,
dp p dp 1
P  p    , dP  p   , grad P  p   grad p
0 0 0 0
Governing Equations
Euler Equation
The Euler equation (per unit of mass) may be written as,
1 v
 t

v
t
1
 2  v  grad v
2
 
2
 grad P  p   grad  
v
t
1
 2  v  grad v
2
 
2
 1 2  v
 grad  P  p     v    2  v
 2  t
Governing Equations
Euler Equation
The Euler equation (per unit of mass) for the incompressible case
may be written as,
1 v
0 t

v
t
1
 2  v  grad v
2
 
2
1
 grad p  grad  
0
v
t
1
 2  v  grad v
2
 
2
 p 1 2  v
 grad     v    2  v
 0 2  t
Governing Equations
Mechanical Problem
❶    div v  0 ❶❸
❸  1 2  v ❶
 grad  P  p     v    2  v
 2  t
❶     p
Ideal Fluids > Stationary Flow
Governing Equations
Mechanical Problem: Stationary Flow
❶ div   v   0 ❶❸
❸  1 2 ❶
 grad  P  p     v   2  v
 2 
❶     p
Governing Equations
Incompressible Mechanical Problem: Stationary Flow
❶ div v  0 ❸
❸  p 1 2 ❶
 grad     v   2  v
 0 2 
Governing Equations
Bernoulli Equation
The Euler equation (per unit of mass) for the stationary flow may
be written as,
 1 2
 grad  P  p     v   2  v
 2 
Taking the projection of the Euler equation along a streamline,

i.e. taking the dot product with the velocity field, yields,
 1 2 d 1 2
 grad  P  p     v   v    P  p     v   0
 2  dt  2 
Governing Equations
Bernoulli Equation
The Euler equation (per unit of mass) for the stationary flow
remains constant along the points of a streamline, yielding the
Bernoulli equation given by,
1 2
P  p     v  cte along a streamline
2
Furthermore, if the fluid is incompressible, the Bernoulli

equation takes the form,
p 1 2
   v  cte along a streamline
0 2
Governing Equations
Bernoulli Equation
If the fluid is incompressible and the only body forces are those
due to the gravity, the Bernoulli equation given by,
p 1 2
 gz  v  cte along a streamline
0 2
Dividing by the gravity’s acceleration, the Bernoulli equation (in

length’s dimensions) takes the form,
p 1
z v  cte along a streamline
2
0 g 2g
Assignment 10.1
Assignment 10.1
Let us consider a stationary flow of a barotropic fluid through a
pipeline with input section S1  A 2 and output section S2  A.
Velocity and pressure distributions are uniform at both sections.
The kinetic state equation is given by   0 exp   p  . Compute
the velocity v2 at the output section, assuming that the velocity
v1 p1
and pressure at the input section are, respectively, and , and
the pressure at the outward section is p2  0 .
Assignment 10.1
Assignment 10.1
Let us consider a stationary flow of a barotropic fluid through a
pipeline with input section S1  A 2 and output section S2  A.
Velocity and pressure distributions are uniform at both sections.
The kinetic state equation is given by   0 exp   p  . Compute
the velocity v2 at the output section, assuming that the velocity
v1 p1
and pressure at the input section are, respectively, and , and
the pressure at the outward section is p2  0.
The mass continuity equation for a stationary motion may be

written as,
div   v   0
Assignment 10.1
Integrating over the control volume and using the divergence
theorem yields,
 div   v  dV    v  n dS  0
V V
1
   1 v1S1   2 v2 S2    1 v1 A   2 v2 A  0
2
1
 1 v1   2 v2
2
Using the kinetic sate equation yields,
1  0 exp   p1  , 2  0 exp   p2   0
1 1
0 exp   p1  v1  0v2  v2  v1 exp   p1 
2 2
Assignment 10.2
Assignment 10.2
Consider a cylindrical tank of diameter D, open on the top and
with a small lateral circular orifice of diameter d (d<<D) located
at a distance H below the top water level. Determine the
diameter d of the circular orifice needed to obtain a water
volume flux Q.
Assignment 10.2
Assignment 10.2
Consider a cylindrical tank of diameter D, open on the top and
with a small lateral circular orifice of diameter d (d<<D) located
at a distance H below the top water level. Determine the
diameter d of the circular orifice needed to obtain a water
volume flux Q.
Water is modeled as an incompressible ideal fluid. We assume

the only body force is the gravity. As the diameter d of the
lateral orifice is much smaller than the diameter D of the tank, a
stationary flow may be assumed.
Assignment 10.2
Then the Bernoulli equation has to be satisfied along a stream‐
line,
p 1
z v  cte along a streamline
2
0 g 2g
Taking the Bernoulli equation at two different points of the same
streamline, one on the free surface and the other one on the
orifice, and assuming the velocity on the free surface is negligi‐
ble because d<<D, yields,
patm patm 1
H 0  0
2
v
0 g 0 g 2g
v  2 gH
Assignment 10.2
The volume flux is given by,
d2 d2
Q v  2 gH
4 4
To get a volume flux Q, the diameter of the orifice has to be,

4Q
d
 2 gH
Assignment 10.3
Assignment 10.3
Let us consider an incompressible ideal fluid flowing through the
channel of the figure under stationary conditions. The velocities
at the left and right sections are uniform, being 1 m/s and 2 m/s,
respectively. Compute the height H.
v1=1 m/s
h1
H v2=2 m/s
h2
Assignment 10.3
Assignment 10.3
The mass continuity equation for an incompresible medium is
given by,
div v  0
Integrating over the control volume and applying the divergence
theorem yields,
 V
div v dV   v  n dS  0
V
 v1h1  v2 h2  h1  2h2  0  h1  2h2
Assignment 10.3
Applying Bernoulli equation between points A and B belonging to
te streamline on the free surface, being point A the one on the
left section and point B the one on the right section, yields,
patm v12 patm v22
 H  h1    h2 
0 g 2 g 0 g 2g
v22 v12 3 1 3
H  h2  h1    h2    h1 
2g 2g 2g 2 2g
Ideal Fluids > Transient Curl‐free Flow
Governing Equations
Mechanical Problem
❶    div v  0 ❶❸
❸  1 2  v ❶
 grad  P  p     v    2  v
 2  t
❶     p
Governing Equations
Transient Curl‐free Flow
Let us consider a transient curl‐free (or potential) flow such that,
  0  rot v  0     x, t  v  grad 
The transient Euler equation for a curl‐free flow may be written
as,
 1 2    
 grad  P  p     grad     grad    grad  
 2  t  t 
 1  
 grad  P  p     grad   0
2
 2 t 
1 
P  p     grad    f t 
2
2 t
Governing Equations
Transient Curl‐free Flow
Let us introduce a modified potential of velocities such that,
  x, t    x, t 
  f t 
t t
v  x, t   grad   x, t   grad   x, t 
The transient Euler equation for a curl‐free flow may be written

as,
1 
P  p     grad   0
2
2 t
Governing Equations
Mechanical Problem: Curl‐free Flow
❶    div  grad    0 ❶❶
❶ 1  ❶
P  p    grad   0
2
2 t
❶     p
Governing Equations
Incompressible Mechanical Problem: Curl‐free Flow
❶ div  grad    0 ❶
❶ p 1  ❶
  grad   0
2
0 2 t
Continuum Mechanics
Chapter 11
Newtonian Fluids
Newtonian Fluids > Contents
Contents
Chapter 11 · Newtonian Fluids
1. Constitutive equation
3. Boundary conditions
4. Curvilinear coordinates
5. Assignments

Newtonian Fluids > Constitutive Equation
The constitutive equation for an isotropic Newtonian fluid may
be written as,
σ = − p1 + λ ( p,θ )( tr d ) 1 + 2µ ( p,θ ) d
where λ ( p,θ ) , µ ( p,θ ) ≥ 0 are two scalar-valued functions
denoted as dynamic viscosities.

The volumetric part of the constitutive equation for an isotropic
tr σ = − p tr 1 + λ ( p,θ )( tr d ) tr 1 + 2µ ( p,θ ) tr d
= −3 p + ( 3λ ( p,θ ) + 2µ ( p,θ ) ) tr d
= −3 p
 2 
p = p −  λ ( p,θ ) + µ ( p,θ )  tr d
 3 
= p − K ( p,θ ) tr d
where K ( p,θ ) ≥ 0 is a non-negative scalar-valued function
denoted as dynamic bulk viscosity.
Thermodynamic Pressure vs Mean Pressure
The thermodynamic pressure and the mean pressure are related
through the expression,
K ( p, θ ) d ρ
p = p + K ( p,θ ) tr d = p + K ( p,θ ) div v = p −
ρ dt
The thermodynamic pressure will be equal to the mean pressure
under any of the following conditions:
Zero bulk viscosity (Stokes condition): K ( p, θ ) = 0
Incompressible fluid: div v = ρ = 0
Uniform velocity field: v = cte
Fluid at rest: v = 0

Newtonian Fluids > Governing Equations
Governing Equations
Governing Equations
❶ ρ + ρ div v = 0 ❶❸
❸ div σ + ρ b = ρ v ❾
❸ σ =σT
Balance of energy
❶ ρ e = σ : d + ρ r − div q ❶❸
❶❶
Governing Equations
Thermo-mechanical constitutive equation for the stresses
❻ σ = − p1 + λ ( p,θ )( tr d ) 1 + 2µ ( p,θ ) d ❶
❶ η = η ( d, p,θ )
❸ q = q ( v,θ ) = −k ( v,θ ) grad θ
❶ e = e ( ρ ,θ )
Kinetic state equation
❶ ρ = ρ ( p, θ )
Governing Equations
Mechanical Problem
❶ ρ + ρ div v = 0 ❶❸
Balance of linear momentum. Cauchy’s first motion
❸ div σ + ρ b = ρ v ❾
❸ σ =σT
❻ σ = − p1 + λ ( tr d ) 1 + 2µ d ❶
❶ ρ = ρ ( p)
Governing Equations
Mechanical Problem
❶ ρ + ρ div v = 0 ❶❸
❸ div σ + ρ b = ρ v ❻
❻ σ = − p1 + λ ( tr d ) 1 + 2µ d ❶
❶ ρ = ρ ( p)

Governing Equations
❶ div v = 0 ❸
❸ div σ + ρ0 b = ρ0 v ❻
❻ σ = − p1 + 2µ d ❶

Governing Equations
Navier-Stokes Equation
The Navier-Stokes equation is obtained substituting the consti-
tutive and geometric equations into the first Cauchy’s motion
equation, and assuming the viscosities are constants, yields,
div σ + ρ b = div ( − p1 + λ ( tr d ) 1 + 2µ d ) + ρ b
= − grad p + ( λ + µ ) grad ( div v ) + µ div ( grad v ) + ρ b
∂v
= ρ v = ρ + ρ ( grad v ) v
∂t
− grad p + ( λ + µ ) grad ( div v ) + µ div ( grad v ) + ρ b = ρ v

Governing Equations
Mechanical Problem
❶ ρ + ρ div v = 0 ❶❸
Balance of linear momentum. Navier-Stokes equation
− grad p + ( λ + µ ) grad ( div v ) + µ div ( grad v ) + ρ b = ρ v
❸ ❶
❶ ρ = ρ ( p)

Governing Equations
❶ div v = 0 ❸
Balance of linear momentum. Navier-Stokes equation
❸ − grad p + µ div ( grad v ) + ρ0 b = ρ0 v ❶

Governing Equations
Thermal Problem
Balance of energy
❶ ρ e = σ : d + ρ r − div q ❶❸
❶❶
❶ η = η ( d, p,θ )
❸ q = q ( v,θ ) = −k ( v,θ ) grad θ
❶ e = e ( ρ ,θ )
Governing Equations
Stress Power
The stress power per unit of spatial volume for a Newtonian fluid
may be written as,
σ : d = − p1 : d + λ ( tr d ) 1 : d + 2µ d : d
2
= − p tr d + λ ( tr d ) + 2µ d : d
1
2 2 
σ : d = − p tr d + λ ( tr d ) + 2µ  ( tr d ) + dev d : dev d 
3 
 2  2
= − p tr d +  λ + µ  ( tr d ) + 2µ dev d : dev d
 3 
2
= − p tr d + K ( tr d ) + 2µ dev d : dev d
Governing Equations
Thermal Problem
Balance of energy
2
ρ e = − p tr d + K ( tr d ) + 2µ dev d : dev d + ρ r − div q
❶ ❶❸
❶❶
❶ η = η ( d, p,θ )
❸ q = q ( v,θ ) = −k ( v,θ ) grad θ
❶ e = e ( ρ ,θ )
Governing Equations
Thermal Problem
Balance of energy
ρ e = − p tr d + K ( tr d ) + 2µ dev d : dev d + ρ r + div ( k ( v,θ ) grad θ )
2
❶ ❶❶
Dint := ρ θ η − ρ r − div ( k ( v,θ ) grad θ ) ≥ 0,
❶
❶ η = η ( d, p,θ )
❶ e = e ( ρ ,θ )
Governing Equations
Balance of energy
❶ρ0 e = 2µ dev d : dev d + ρ0 r + div k ( v,θ ) grad θ ❶❶ ( )
Dint := ρ0 θ η − ρ0 r − div ( k ( v,θ ) grad θ ) ≥ 0,
❶
❶ η = η ( d, p,θ )
❶ e = e (θ )

Newtonian Fluids > Boundary Conditions
Boundary Conditions
Prescribed Velocity Boundary Conditions
Impenetrability condition
The impenetrability boundary condition may be written as,
( ( ) ( )) ( )
v x, t − v * x, t ⋅ n x, t = 0 ∀x ∈Γ (1)
v , ∀t
where v * ( x,t ) is velocity of the boundary.

If the boundary does not moves, then the impenetrability
condition takes the form,
v ( x, t ) ⋅ n ( x, t ) = 0 ∀x ∈Γ(1)
v , ∀t
The impenetrability condition is used for ideal fluids, while for

viscous fluids a more restrictive boundary condition, denoted as
adherence condition, is used.
Boundary Conditions
Adherence condition
The adherence boundary condition may be written as,
v ( x, t ) − v * ( x, t ) = 0 ∀x ∈Γ(1)
v , ∀t
where v * ( x,t ) is velocity of the boundary.

If the boundary does not moves, then the adherence condition
takes the form,
v ( x, t ) = 0 ∀x ∈Γ(1)
v , ∀t
The adherence condition is used for viscous fluids, such as

Newtonian fluids, while for ideal fluids the less restrictive
impenetrability condition, is used.
Boundary Conditions
Prescribed velocity
The prescribed velocity boundary condition may be written as,
v ( x, t ) = v * ( x, t ) ∀x ∈ Γ(2)
v , ∀t
where v * ( x,t ) is the prescribed velocity.

Boundary Conditions
Prescribed Pressure Boundary Conditions
Prescribed pressure
The prescribed pressure boundary condition may be written as,
p ( x, t ) = p * ( x, t ) ∀x ∈Γ(1)
p , ∀t
where p * ( x, t ) is the prescribed thermodynamic pressure.

Boundary Conditions
Prescribed Pressure Boundary Conditions
Free surface
At the free surface the pressure is prescribed to the environmen-
tal pressure, yielding,
p ( x, t ) = patm ( x, t ) ∀x ∈Γ(2)
p , ∀t
Usually free surfaces are modeled as material surfaces.

Boundary Conditions
Prescribed Traction Boundary Conditions
Prescribed traction
The prescribed traction boundary condition may be written as,
t ( x, t ) = σ ( x, t ) n ( x, t ) = t * ( x, t ) ∀x ∈ Γσ(1) , ∀t

Boundary Conditions
Prescribed Traction Boundary Conditions
Equilibrium at the interface
At the interface between two immiscible fluid the equilibrium
boundary condition may be written as,
t (1) ( x, t ) + t (2) ( x, t ) = 0 ∀x ∈Γσ(1,2) , ∀t
σ (1) ( x, t ) n(1) ( x, t ) + σ (2) ( x, t ) n(2) ( x, t ) = 0 ∀x ∈Γσ(1,2) , ∀t
( σ (1)
( x, t ) − σ (2)
( x, t ) ) n ( x, t ) = 0 ∀x ∈Γ (1,2)
σ , ∀t

Newtonian Fluids > Curvilinear Coordinates


Mass Continuity Equation

Mass Continuity Equation
Mass Continuity Equation in Cartesian Coordinates
∂ρ ∂ ∂ ∂
+ ( ρ vx ) + ( ρ v y ) + ( ρ vz ) = 0
∂t ∂x ∂y ∂z
Mass Continuity Equation in Cylindrical Coordinates

∂ρ 1 ∂ 1 ∂ ∂
+ ( r)
ρ rv + ( θ)
ρ v + ( ρ vz ) = 0
∂t r ∂r r ∂θ ∂z
Mass Continuity Equation in Spherical Coordinates

∂ρ 1 ∂ 1 ∂ 1 ∂
+ 2 ( ρ r vr ) +
2
( ρ vθ sin θ ) + ( ρ vφ ) = 0
∂t r ∂r r sin θ ∂θ r sin θ ∂φ

Incompressible Navier-Stokes Equation

Incompressible Navier-Stokes Equation in
Cartesian Coordinates
∂p  ∂ 2 vx ∂ 2 vx ∂ 2 vx   ∂vx ∂vx ∂vx ∂vx 
− + µ  2 + 2 + 2  + ρ bx = ρ  + vx + vy + vz 
∂x  ∂ x ∂ y ∂ z   ∂t ∂x ∂y ∂z 
∂p  ∂ 2vy ∂ 2vy ∂ 2vy   ∂v y ∂v y ∂v y ∂v y 
− +µ 2 + 2 + 2

+ ρ by = ρ  + vx + vy + vz 
∂y  ∂x ∂y ∂z ∂t ∂ x ∂y ∂ z
   
∂p  ∂ 2 vz ∂ 2 vz ∂ 2 v z   ∂vz ∂vz ∂vz ∂vz 
− + µ  2 + 2 + 2  + ρ bz = ρ  + vx + vy + vz 
∂z  ∂ x ∂ y ∂ z   ∂t ∂ x ∂y ∂z 


∂p  ∂ 1 ∂  1 ∂ vr 2 ∂vθ ∂ vr
2 2

− +µ  ( rvr )  + 2 2 − 2 + 2  + ρ br =
∂r  ∂r  r ∂r  r ∂θ r ∂θ ∂z 
 ∂vr ∂vr vθ ∂vr vθ2 ∂vr 
= ρ + vr + − + vz 
 ∂ t ∂r r ∂θ r ∂z 
1 ∂p  ∂ 1 ∂  1 ∂ vθ 2 ∂vr ∂ vθ 
2 2
− +µ  ( rvθ )  + 2 2 + 2 + 2  + ρ bθ =
r ∂θ  ∂r  r ∂r  r ∂θ r ∂θ ∂z 
 ∂vθ ∂vθ vθ ∂vθ vr vθ ∂vθ 
= ρ + vr + + + vz 
 ∂ t ∂ r r ∂θ r ∂z 
∂p  1 ∂  ∂vz  1 ∂ 2 vz ∂ 2 vz   ∂vz ∂vz vθ ∂vz ∂vz 
− +µ  r +
 2 2
+ 2 
+ ρ bz = ρ  + vr + + v z 
∂z  r ∂r  ∂r  r ∂θ ∂z   ∂ t ∂r r ∂ θ ∂z 

∂p ∂ 1 ∂  1 ∂  ∂vr  1 ∂ 2 vr 2 ∂ 2 ∂vφ 
− + µ   2 ( r 2 vr )  + 2  sin θ +
 2 2 2
− (θ
v sin θ ) 2
− 
∂r  ∂r  r ∂r  r sin θ ∂θ  ∂θ  r sin θ ∂φ r 2 sin θ ∂θ r sin θ ∂φ 
 ∂vr ∂vr vθ ∂vr vφ ∂vr vθ2 + vφ2 
+ ρ br = ρ  + vr + + − 
 ∂t ∂r r ∂θ r sin θ ∂ φ r
 
1 ∂p  1 ∂  2 ∂vθ  1 ∂  1 ∂  1 ∂ 2 vθ 2 ∂vr 2cot θ ∂vφ 
− + µ 2 r + 2  ( vθ sin θ )  + 2 2 2
+ 2 − 2 
r ∂θ  r ∂r  ∂r  r ∂θ  sin θ ∂ θ  r sin θ ∂ φ r ∂θ r sin θ ∂φ 
 ∂vθ ∂vθ vθ ∂vθ vφ ∂vθ vr vθ vφ2 cot θ 
+ ρ bθ = ρ  + vr + + + − 
 ∂t ∂r r ∂θ r sin θ ∂φ r r
 
1 ∂p  1 ∂  2 ∂vφ  1 ∂  1 ∂  1 ∂ 2 vφ 2 ∂vr 2cot θ ∂vθ 
−
r sin θ ∂φ
+ µ  2 r
 ∂ ∂
+
 2
∂θ 
 θ ∂θ
( vφ sin θ ) 
+
 2 2
θ ∂ φ 2
+ 2
θ ∂φ
+ 2
θ ∂φ

 r r  r  r sin r sin r sin r sin 
 ∂vφ ∂v v ∂v v ∂vφ vr vφ vθ vφ cot θ 
+ ρ bφ = ρ  + vr φ + θ φ + φ + + 
 ∂t ∂r r ∂θ r sin θ ∂φ r r
 
Newtonian Fluids > Assignments
Example 11.1
Example 11.1
Let us consider the stationary flow of an incompressible Newto-
nian fluid with dynamic viscosity µ > 0, flowing between two
parallel horizontal plates as it is shown in the figure. Body forces
are considered to be negligible. Obtain the velocity field assu-
ming that the velocity and pressure fields are such that,
vx = v ( y ) , v y = 0, vz = 0
[∇p ] = [ −a 0 0] , a = cte > 0
y
h/2
v(y)
x
h/2

Example 11.1
Example 11.1
The mass continuity equation for an incompressible fluid can be
written as,
ρ + ρ div v = 0, ρ = 0 ⇒ div v = 0
The Navier-Stokes equation for an incompressible Newtonian

fluid under stationary conditions and negligible body forces can
be written as,
 ∂v 
−∇p + ( λ + µ ) grad ( div v ) + µ∇ v + ρ b = ρ 
2
+ ( grad v ) v 
 ∂t 
−∇p + µ∇ 2 v = ρ ( grad v ) v
Example 11.1
The convective derivative of the velocity is zero,
( grad v ) v = 0
yielding,
∂ 2
v( y)
−∇p + µ∇ v = 0 ⇒ a + µ
2
2
=0
∂y
Integrating yields,
a 2
v ( y) = − y + C1 y + C2
2µ
The sticking boundary conditions read,
   
2 2
h ah h h ah h
v  = − + C1 + C2 = 0, v  −  = − − C1 + C2 = 0
2 8µ 2  2 8µ 2
Example 11.1
Solving the system of equations, the integration constants are
given by,
ah 2
C1 = 0, C2 =
8µ
and the velocity field (Poiseuille’s flow) takes the value,
a  h2 2
v( y) =  −y 
2µ  4 
v(y)

Solution Steps
Solution Steps for Incompressible Newtonian Fluids
• Step 1. Select appropriate system of coordinates
• Step 2. Introduce suitable hypothesis on the velocity and
pressure fields
• Step 3. Solve mass continuity equation for an incompressible
fluid
• Step 4. Solve the Navier-Stokes equation for an incompressi-
ble Newtonian fluid to get the velocity and pressure fields.
Then compute deformation rate and stress tensors using
geometric and constitutive equations, respectively.
• Step 5. Apply boundary conditions to determine the integra-
tion constants

Assignment 11.1
An horizontal plate of infinite length and thickness 2a is placed
between two incompressible Newtonian fluids flowing in steady-
state regime between two horizontal plates, as it is shown in the
figure.
F2
B v
h
a x
v/2
a F1
h
A

Assignment 11.1
The distance between the plates is h. The upper plate and the
mid plate have prescribed velocities v* and v*/2, respectively.
The pressure at the points A and B shown on the figure, are pA
and pB, respectively. It is assumed that the flow and its properties
do not depend on x.
1) Compute the velocity, pressure and stress fields in each fluid
2) Compute the forces per unit of surface F1 and F2 acting on
the mid plate and the upper plate, respectively, needed to
keep the described motion

Assignment 11.2
Consider the stationary flow of an incompressible Newtonian
fluid with dynamic viscosity μ, under an horizontal plate of
length L as shown in the figure. Inertial forces, environmental
pressure and body forces in the plate are neglected. The pres-
sure at the point (x=l, y=a) is zero. The volume flux per unit of
width is Q.
y l
d V H
p=0
a Q
a x
Assignment 11.2
The plate remains fixed and horizontal due to the application of
forces per unit of width H and V, at a distance d of the left edge
of the plate. The flow is parallel to the plane x-y.
1) Compute the velocity and pressure fields
2) Compute the horizontal force H and vertical force V (per unit
of width), as well as the distance d, needed to keep the plate
horizontal and fixed
y l
d V H
p=0
a Q
a x
Assignment 11.3
Consider stationary flows of the two immiscid incompressible
Newtonian fluids with dynamic viscosities µ1 , µ2, as shown in the
figure. Body forces in the fluids and environmental pressure are
neglected. The plate has density ρ * and thickness t . On the top
of the plate a tangential force f * per unit of surface is applied.
y f*
t
v* A' B' h
x
µ1
µ2
B
A
α

Assignment 11.3
1) Compute the velocity, pressure and stress fields in each one
of the fluids
2) Compute the volume flux Q accross the circular section BB’
3) Compute the velocity v* of the plate
y f*
t
v* A' B' h
x
µ1
µ2
B
A
α

Continuum Mechanics
Chapter 12
Variational Principles
Variational Principles > Contents
Contents
Chapter 12 · Variational Principles
1. Motivation
2. Functionals
3. Variational Principle
4. Virtual Work Principle
5. Minimum Potential Energy Principle

Variational Principles > Motivation
Motivation
Motivation
The local (or strong) form of the governing equations may not
be appropriate from a computational point of view
The variational principle allows to get the variational (or
weak) form of the local (or strong) form of the governing
equations
The variational principle guarantees that the solution of the
variational (or weak) form is also the solution of the local (or
strong) form of the governing equations
The variational form of the governing equations are more
suitable from a computational point of view

Variational Principles > Functionals
Functionals
Space of Functions and Functionals
Let us consider the space of functions defined as,
U := {u ( x ) : Ω ⊂ »3 → » m }
Elements of the space of functions are functions of arbitrary
tensorial order (scalars, vectors, second-order tensors, etc.)
defined in a subset Ω ⊂ »3 .
A functional F ( u ) is defined as an application of the space of

functions U onto the set of the real numbers » ,
F (u) : U → »

Functionals
Examples
Let us consider a 1D domain Ω := [ a, b] ∈ » and the space of
functions defined as,
U := {u ( x ) : Ω ⊂ » → »}
The following functionals could be defined,

b
F1 ( u ) = ∫ u ( x ) dx
a
b
F2 ( u ) = ∫ ( u ( x ) + u′ ( x ) ) dx
a
b
F3 ( u ) = ∫ f ( x, u ( x ) , u′ ( x ) ) dx
a

Functionals
Gateaux Derivative
Let us consider the space of functions defined as,
U := {u ( x ) : Ω ⊂ »3 → » m }
and the functional given by,
F (u) : U → »
Let us consider a perturbed function, belonging to the same
space of functions, defined as,
u ( x) + ε δ u ( x) ∈ U
where ε is a scalar perturbation parameter and δ u ( x ) ∈ U is a
function representing an arbitrary perturbation direction.

Functionals
Gateaux Derivative
The Gateaux (or directional) derivative of the functional F ( u ) in
the direction δ u , denoted as δ F ( u, δ u ) := DF ( u ) ⋅ δ u , is
defined as,
d
δ F ( u, δ u ) := DF ( u ) ⋅ δ u = F (u + ε δ u)
dε ε =0

Functionals
Gateaux Derivative
Given a family of functionals defined as,
F (u) = ∫ ϕ (u) dΩ + ∫ ψ (u) dΓ
Ω ∂Ω
The Gateaux (or directional) derivative of the functional F ( u ) in
the direction δ u , takes the form,
d
δ F ( u, δ u ) := DF ( u ) ⋅ δ u = F (u + ε δ u)
dε ε =0
d d
=
dε ∫Ω ϕ ( u + ε δ u ) d Ω ε =0 + dε ∫
∂Ω
ψ (u + ε δ u) dΓ
ε =0
∂ϕ ( u ) ∂ψ ( u )
=∫ ⋅δ u dΩ + ∫ ⋅δ u dΓ
Ω ∂u ∂Ω ∂u
Functionals
Gateaux Derivative
The perturbation direction δ u represents an arbitrary variation,
not necessarily infinitesimal, of the function u and it should not
be confused with du which is the differential of the function u
in the sense of the differential calculus.
Nevertheless, the Gateaux (or directional) derivative of a

functional may have the same formalism than the differentiation
of a function.

Functionals
Gateaux Derivative
Let us consider the spaces of functions defined as,
U := {u ( x ) : Ω ⊂ »3 → » m | u ( x ) = u ( x ) on ∂u Ω ⊂ ∂Ω}
U 0 := {δ u ( x ) : Ω ⊂ »3 → » m | δ u ( x ) = 0 on ∂u Ω ⊂ ∂Ω}
and the family of functionals F ( u ) : U → » given by,
F ( u ) = ∫ ϕ ( x, u, ∇u ) d Ω + ∫ ψ ( x, u, ∇u ) d Γ
Ω ∂σ Ω
such that the boundary of the domain satisfies that,

∂ u Ω ∪ ∂σ Ω = ∂Ω, ∂ u Ω ∩ ∂σ Ω = ∅

Functionals
Gateaux Derivative
Let us consider a perturbed function, defined as,
u ( x) + ε δ u ( x) ∈ U
where ε is a scalar perturbation parameter and δ u ( x ) ∈ U 0 is an
arbitrary perturbation direction function.
The Gateaux (or directional) derivative of the functional F ( u )
may be written as,
δ F ( u, δ u ) = DF ( u ) ⋅ δ u
= ∫ E ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Ω + ∫ T ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Γ
Ω ∂σ Ω

Variational Principles > Variational Principle
Variational Principle
Extreme of a Function
Given a real function of a real number, denoted as f ( x ) : » → » ,
the function has a minimum on x0 ∈ » if the following condition
holds,
f ( x0 ) ≤ f ( x ) ∀ x ∈ »
A necessary condition for the function f ( x ) : » → » to have a

minimum on x0 ∈ » is given by,
f ′ ( x0 ) = 0

Extreme of a Functional · Variational Principle
Given a functional, denoted as F ( v ) : U → » , the functional
has a minimum on u ∈ U if the following condition holds,
F (u) ≤ F ( v ) ∀ v ∈ U
A necessary condition for the functional F ( v ) : U → » to have

a minimum on u ∈ U , denoted as the variational principle
associated to this functional, is given by,
δ F ( u, δ u ) := DF ( u ) ⋅ δ u = 0 ∀ δ u ∈ U 0

Extreme of a Functional · Variational Principle
The variational principle associated to the family of functionals
F ( u ) : U → » given by,
F ( u ) = ∫ ϕ ( x, u, ∇u ) d Ω + ∫ ψ ( x, u, ∇u ) d Γ
Ω ∂σ Ω
takes the form,

δ F ( u, δ u ) = DF ( u ) ⋅ δ u
= ∫ E ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Ω
Ω
+∫ T ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Γ = 0 ∀ δ u ∈ U 0
∂σ Ω

Fundamental Theorem of Variational Calculus
The fundamental theorem of variational calculus states that the
variational principle, given by,
δ F ( u, δ u ) = DF ( u ) ⋅ δ u = ∫ E ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Ω
Ω
+∫ T ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Γ = 0 ∀ δ u ∈ U 0
∂σ Ω
is satisfied, if and only if, the following local equations, denoted

as Euler-Lagrange equations and natural boundary conditions
of the variational principle, hold,
E ( x, u ( x ) , ∇u ( x ) ) = 0 ∀x ∈ Ω
T ( x, u ( x ) , ∇u ( x ) ) = 0 ∀x ∈ ∂σ Ω
Example
Given the spaces of functions and the functional defined as,
U := {u ( x ) : [ a, b] ⊂ » → » u ( x ) = u ( x ) on x = a ⊂ ∂Ω}
U 0 := {δ u ( x ) : [ a, b] ⊂ » → » δ u ( x ) = 0 on x = a ⊂ ∂Ω}
b
F ( u ) = ∫ φ ( x, u ( x ) , u′ ( x ) ) dx
a
The variational principle associated to this functional may be

written as
b
δ F ( u, δ u ) = DF ( u ) ⋅ δ u = ∫ E ( x, u ( x ) , u′ ( x ) ) δ u dx
a
+ T ( xb , u ( xb ) , u′ ( xb ) ) δ ub = 0 ∀ δ u ∈ U 0
The perturbed functional takes the form,
b
F ( u + εδ u ) = ∫ φ ( x, u ( x ) + εδ u ( x ) , u′ ( x ) + εδ u′ ( x ) ) dx
a
The Gateaux derivative of the functional takes the form,

d
δ F ( u, δ u ) = F (u + ε δ u )
dε ε =0
b  ∂φ ∂φ 
= ∫  δu + δ u′  dx

a ∂u ∂u′ 
b ∂φ b ∂φ
=∫ δ u dx + ∫ δ u′dx = 0 ∀ δ u ∈ U 0
a ∂u a ∂u ′

The Gateaux derivative of the functional, integrating by parts the
second term and collecting terms, takes the form,
b ∂φ b ∂φ
δ F ( u, δ u ) = ∫ δ u dx + ∫ δ u′dx
a ∂u a ∂u ′
b ∂φ ∂φ ∂φ b d  ∂φ 
=∫ δ u dx + δ ub − δ ua − ∫   δ udx
a ∂u ∂u′ b ∂u′ a 
a dx ∂u ′

 ∂φ d  ∂φ  
b ∂φ
=∫  −    δ u dx + δ ub = 0 ∀ δ u ∈ U 0
a ∂u
 dx  ∂u′   ∂u′ b

The Euler-Lagrange and natural boundary conditions associated
to the variational principle take the form,
 ∂φ d  ∂φ  
b ∂φ
δ F ( u, δ u ) = ∫  −    δ u dx + δ ub
a ∂u
 dx  ∂u′   ∂u′ b
b
= ∫ E ( x, u ( x ) , u′ ( x ) ) δ u dx
a
+ T ( xb , u ( xb ) , u′ ( xb ) ) δ ub = 0 ∀ δ u ∈ U 0
 ∂φ d  ∂φ  
E ( x, u ( x ) , u′ ( x ) ) :=  −    =0
 ∂u dx  ∂u′  
∂φ
T ( xb , u ( xb ) , u′ ( xb ) ) := =0
∂u′ b
Variational Form of Local Governing Equations
Let us consider the local form of the governing equation and
boundary conditions of a continuum mechanics problem, given
by
E ( x, u ( x ) , ∇u ( x ) ) = 0 ∀x ∈ Ω
T ( x, u ( x ) , ∇u ( x ) ) = 0 ∀x ∈ ∂σ Ω
u ( x ) = u ( x ) ∀x ∈ ∂ u Ω

Variational Form of Local Governing Equations
Given the following space of functions,
U := {u ( x ) : Ω ⊂ »3 → » m | u ( x ) = u ( x ) on ∂u Ω ⊂ ∂Ω}
U 0 := {δ u ( x ) : Ω ⊂ »3 → » m | δ u ( x ) = 0 on ∂ u Ω ⊂ ∂Ω}
The variational problem may be written as: Find u ( x ) ∈ U such

that,
δ F ( u, δ u ) = DF ( u ) ⋅ δ u
= ∫ E ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Ω
Ω
+∫ T ( x, u ( x ) , ∇u ( x ) ) ⋅ δ u d Γ = 0 ∀ δ u ∈ U 0
∂σ Ω
Variational Principles > Virtual Work Principle
Virtual Work Principle

Let us consider the local form of the linear momentum balance
equation, i.e. first Cauchy motion equation, and the prescribed
traction and displacement boundary conditions for a continuum
mechanics problem, given by
div σ ( u ) + ρ b = ρ a ( u ) in Ω
σ n = t on ∂σ Ω
u = u on ∂ u Ω


Let us consider the space of admissible displacements, satisfying
the Dirichlet boundary condition, and the space of admissible
virtual displacements, defined as,
U := {u : Ω ⊂ »3 → »3 | u = u on ∂ u Ω ⊂ ∂Ω}
U 0 := {δ u : Ω ⊂ »3 → »3 | δ u = 0 on ∂ u Ω ⊂ ∂Ω}
Setting,
E := div σ ( u ) + ρ b − ρ a ( u ) in Ω
T := t − σ n on ∂σ Ω

The variational problem, known as the virtual work principle,
may be written as: Find an admissible displacement field u ( x ) ∈ U
such that,
δ W ( u, δ u ) = DW ( u ) ⋅ δ u
= ∫ E ⋅δ u dΩ + ∫ T ⋅δ u dΓ
Ω ∂σ Ω
(
= ∫ div σ ( u ) + ρ ( b − a ( u ) ) ⋅ δ u d Ω
Ω
)
+∫ ( t − σ n) ⋅δ u dΓ = 0 ∀δu∈U0
∂ Ω
σ
where δ W ( u, δ u ) is the Gateaux derivative of a functional

and can be viewed as the total virtual work.

Using the divergence theorem yields,
∫Ω
div σ ⋅ δ u d Ω = ∫ div (δ u ⋅ σ ) d Ω − ∫ σ : grad δ u d Ω
Ω Ω
= ∫ δ u ⋅ (σ n ) d Γ − ∫ σ : grad δ u d Ω
∂Ω Ω
=∫ δ u ⋅ (σ n ) d Γ − ∫ σ : grad δ u d Ω
∂σ Ω Ω


such that the total virtual work satisfies,
δ W ( u, δ u ) = DW ( u ) ⋅ δ u
=∫ δ u ⋅ (σ n ) d Γ − ∫ σ : grad δ u d Ω
∂σ Ω Ω
+∫ ρ (b − a (u)) ⋅δ u dΩ + ∫ ( t − σ n) ⋅δ u dΓ
Ω ∂σ Ω
= ∫ ρ (b − a (u)) ⋅δ u dΩ + ∫ t ⋅ δ u d Γ − ∫ σ : grad δ u d Ω
Ω ∂σ Ω Ω
= 0 ∀δu∈U0


such that the total virtual work, defined as the external virtual
work minus the internal virtual work, satisfies,
δ W ( u, δ u ) = δ Wext ( u, δ u ) − δ Wint ( u, δ u ) = 0 ∀ δ u ∈ U 0
where the external and internal virtual works are given,
respectively, by,
δ Wext ( u, δ u ) = ∫ ρ ( b − a ( u ) ) ⋅ δ u d Ω + ∫ t ⋅δ u dΓ
Ω ∂σ Ω
δ Wint ( u, δ u ) = ∫ σ : grad δ u d Ω
Ω

Taking into account the symmetry of the Cauchy stress tensor
and setting,
δε := symm ( grad δ u ) := ∇ sδ u
the virtual work principle may be written as: Find an admissible
displacement field u ( x ) ∈ U , such that,
δ Wint ( u, δ u ) = δ Wext ( u, δ u ) ∀ δ u ∈ U 0
δ Wext ( u, δ u ) = ∫ ρ ( b − a ( u ) ) ⋅ δ u d Ω + ∫ t ⋅δ u dΓ
Ω ∂σ Ω
δ Wint ( u, δ u ) = ∫ σ : δε d Ω
Ω


The first Cauchy motion equation and prescribed traction
boundary conditions, given by,
E := div σ ( u ) + ρ b − ρ a ( u ) in Ω
T := t − σ n on ∂σ Ω
are the Euler-Lagrange equation and natural boundary
conditions, respectively, associated to the virtual work principle,
δ W ( u, δ u ) = 0 ∀ δ u ∈ U 0
The virtual work priciple may be viewed as the variational
principle associated to a functional (not explicitly given), being a
necessary condition to obtain the minimum of this functional.

The virtual work principle is valid for any constitutive material
model (i.e., linear elasticity, nonlinear elasticity, plasticity)
The virtual work principle is not restricted to the infinitesimal
strains theory, being also valid within a finite strain framework

Variational Principles > Minimum Potential Energy Principle
Minimum Potential Energy Principle

Hypothesis
H1. Existence of an elastic potential
Within the framework of the infinitesimal strains, we assume
that there exists an elastic potential per unit of volume such that
the stress tensor may be written as,
σ = ∂ ε ( ρ0 e ( ε ) )
For linear elasticity the elastic potential per unit of volume takes
the form,
1
ρ0e ( ε ) = ε : C : ε ⇒ σ = ∂ ε ( ρ0 e ( ε ) ) = C : ε
2


Hypothesis
H2. Existence of a (modified) body forces potential
We assume that there exists a (modified) body forces potential
per unit of volume such that the modified body forces may be
written as,
ρ0 ( b − a ) = −∂uφ ( u )
For quasistatic conditions, gravity forces and uniform density,

the body forces potential per unit of volume takes the form,
φ ( u ) = − ρ0b ⋅ u ⇒ ρ0b = −∂ uφ ( u )


Hypothesis
H3. Existence of a prescribed traction potential
We assume that there exists a prescribed traction potential per
unit of volume such that the prescribed tractions may be written
as,
t = −∂ uϕ ( u )
For a conservative prescribed traction vector, the prescribed

traction potential per unit of volume takes the form,
ϕ ( u ) = − t ⋅ u ⇒ t = −∂ uϕ ( u )


Total Potential Energy Functional
Let us consider the functional, denoted as total potential
energy, defined as the sum of the elastic energy, body forces
potential energy and prescribed tractions potential energy, given
by,
Π ( u ) = ∫ ρ0e ( ε ) dV + ∫ φ ( u ) dV + ∫ ϕ (u) dΓ
Ω Ω ∂σ Ω
The virtual work principle may be viewed as the variational

principle associated to the functional given by the total potential
energy, being a necessary condition to get a minimum of the
functional, such that,
δΠ ( u, δ u ) = DΠ ( u ) ⋅ δ u = 0 ∀ δ u ∈ U 0

Total Potential Energy Functional
The virtual work principle may be obtained as the variational
principle associated to the functional given by the total potential
energy,
δΠ ( u, δ u ) = DΠ ( u ) ⋅ δ u
∂e ( ε ) ∂φ ( u ) ∂ϕ ( u )
= ∫ ρ0 :δε dV + ∫ ⋅ δ u dV + ∫ ⋅δ u dΓ
Ω ∂ε Ω ∂u ∂σ Ω ∂u
= ∫ σ :δε dV − ∫ ρ0 ( b − a ) ⋅ δ u dV − ∫ t ⋅δ u dΓ = 0 ∀ δ u ∈ U 0
Ω Ω ∂σ Ω
∫ σ :δε dV = ∫
Ω Ω
ρ0 ( b − a ) ⋅ δ u dV + ∫
∂σ Ω
t ⋅δ u dΓ ∀ δ u ∈ U 0

Continuum Mechanics PDF

Uploaded by

Copyright:

Available Formats

Continuum Mechanics PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Continuum Mechanics PDF

Uploaded by

Copyright:

Available Formats

Continuum Mechanics

On-campus Schedule 2016-2017

On-campus Classes Schedule:

Dates: 14, 21 Oct | F | 15:00-18:00

September 28, 2016 Carlos Agelet de Saracibar 3

Continuum Mechanics: On-line Course

September 28, 2016 Carlos Agelet de Saracibar 5

September 28, 2016 Carlos Agelet de Saracibar 7

September 28, 2016 Carlos Agelet de Saracibar 8

September 28, 2016 Carlos Agelet de Saracibar 9

September 28, 2016 Carlos Agelet de Saracibar 10

September 9, 2016 Carlos Agelet de Saracibar 2

Those physical properties are mathematically represented by

Tensors are mathematical objects which have the required

September 9, 2016 Carlos Agelet de Saracibar 3

First- and higher-order tensors may be expressed in terms of

September 9, 2016 Carlos Agelet de Saracibar 4

Notation on Hand-written Documents

September 9, 2016 Carlos Agelet de Saracibar 5

September 9, 2016 Carlos Agelet de Saracibar 6

September 9, 2016 Carlos Agelet de Saracibar 7

September 9, 2016 Carlos Agelet de Saracibar 8

where u is the norm (or length or magnitude) of the vector u,

and θ ( u, v ) is the angle between the non-zero vectors u and v

September 9, 2016 Carlos Agelet de Saracibar 9

September 9, 2016 Carlos Agelet de Saracibar 10

September 9, 2016 Carlos Agelet de Saracibar 11

September 9, 2016 Carlos Agelet de Saracibar 12

September 9, 2016 Carlos Agelet de Saracibar 13

A fixed set of three unit vectors which are mutually orthogonal

September 9, 2016 Carlos Agelet de Saracibar 14

September 9, 2016 Carlos Agelet de Saracibar 15

Leaving out the summation symbol, using the so called Einstein

where we have adopted the summation convention introduced

September 9, 2016 Carlos Agelet de Saracibar 16

September 9, 2016 Carlos Agelet de Saracibar 17

Note that the Kronecker delta plays the role of a replacement

September 9, 2016 Carlos Agelet de Saracibar 18

September 9, 2016 Carlos Agelet de Saracibar 19

September 9, 2016 Carlos Agelet de Saracibar 20

September 9, 2016 Carlos Agelet de Saracibar 21

(u ⋅ u ) ( uiei ⋅ u je j ) ( uiu jδ ij ) ( uiui )

September 9, 2016 Carlos Agelet de Saracibar 22

September 9, 2016 Carlos Agelet de Saracibar 23

September 9, 2016 Carlos Agelet de Saracibar 25

September 9, 2016 Carlos Agelet de Saracibar 27

Taking the right-handed orthonormal basis {ei }, the cross (or

September 9, 2016 Carlos Agelet de Saracibar 29

September 9, 2016 Carlos Agelet de Saracibar 30

September 9, 2016 Carlos Agelet de Saracibar 31

September 9, 2016 Carlos Agelet de Saracibar 32

September 9, 2016 Carlos Agelet de Saracibar 35

September 9, 2016 Carlos Agelet de Saracibar 36

September 9, 2016 Carlos Agelet de Saracibar 38

September 9, 2016 Carlos Agelet de Saracibar 39

September 9, 2016 Carlos Agelet de Saracibar 40

September 9, 2016 Carlos Agelet de Saracibar 41

The second-order unit tensor, denoted as 1 , satisfies the

September 9, 2016 Carlos Agelet de Saracibar 42