Sunteți pe pagina 1din 31

Mechanics of solids & Mechanics of uids.

Part I: Fundamentals
M. Ekh and S. Toll
Division of Material and Computational Mechanics,
Chalmers University of Technology, Goteborg, Sweden.
email: magnus.ekh@chalmers.se
August 29, 2013
Contents
1 Tensors 2
1.1 Index notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 2nd order tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Principal values and principal directions . . . . . . . . . . . . . . 10
1.5 Spatial derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.6 Divergence theorem . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 Stress, motion and deformation 14
2.1 Stress analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Continuum motion . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3 Field equations 24
3.1 Physical conservation principles . . . . . . . . . . . . . . . . . . . 25
3.2 Field variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.3 Continuity equation . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.4 Momentum equation . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.5 Energy equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4 Constitutive equations 29
4.1 Fouriers law of thermal conductivity . . . . . . . . . . . . . . . . 29
4.2 Viscous uids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.3 Elastic solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1
1. Tensors 2
1 Tensors
1.1 Index notation
The index, or indical, notation is a powerful shorthand for linear algebric equa-
tion systems. Consider the equation system
a
1
= T
11
b
1
+T
12
b
2
+T
13
b
3
(1)
a
2
= T
21
b
1
+T
22
b
2
+T
23
b
3
(2)
a
3
= T
31
b
1
+T
32
b
2
+T
33
b
3
(3)
Clearly we may write the exact same thing as a summation,
a
1
=
3

j=1
T
1j
b
j
, (4)
a
2
=
3

j=1
T
2j
b
j
, (5)
a
3
=
3

j=1
T
3j
b
j
. (6)
By introducing a so-called free index i = 1, 2, 3, (free to take values 1, 2 or 3) to
represent the three equations (1), (2) or (3), respectively, we may write
a
i
=
3

j=1
T
ij
b
j
. (7)
To simplify the notation even further, we introduce the convention that sum-
mation is implied over repeated indices (known as Einsteins summation con-
vention), hence
a
i
= T
ij
b
j
. (8)
The second index j is called dummy index, as it is only used to imply summation.
Another way to express the equation system (1)-(3) is to use matrices (in
this example two column matrices 3 1 and a square matrix 3 3):
_
_
a
1
a
2
a
3
_
_
=
_
_
T
11
T
12
T
13
T
21
T
22
T
23
T
31
T
32
T
33
_
_
_
_
b
1
b
2
b
3
_
_
(9)
or also here by using Einsteins summation convention:
[a
i
] = [T
ij
] [b
j
] (10)
Assignment 1 (a) Use index notation to re-write the following expres-
sion: b
11
c
1
d
1
+b
12
c
2
d
1
+b
13
c
3
d
1
+b
21
c
1
d
2
+b
22
c
2
d
2
+b
23
c
3
d
2
+b
31
c
1
d
3
+
b
32
c
2
d
3
+b
33
c
3
d
3
.
(b) Expand a
ijk
b
ik
by giving the terms explicitly.
1.2. Vectors 3
1
2
3
e
1
e
2
e
3
A
B
a
Figure 1: Illustration of vector a.
An example of using Matlab commands for matrix denitions (for T and b)
and multiplication a
i
= T
ij
b
j
is given below:
>> T=[1 2 3; 4 5 6; 7 8 9];
>> b=[1 2 3];
>> a=T*b
a =
14
32
50
1.2 Vectors
For further reading see Reddy 2.2 Vector Algebra.
To describe many physical quantities (such as force, displacement, velocity)
both magnitude and direction must be given. Hence, these quantities can be
described by vectors (1st order tensors) in a 3-dimensional Euclidean space.
By introducing a set of right-handed orthonormal basis vectors { e
1
, e
2
, e
3
} any
vector a =

AB can be expressed as a linear combination these basis vectors, e


i
:
a = a
1
e
1
+a
2
e
2
+a
3
e
3
= a
j
e
j
. (11)
as shown in gure 1. The coecients a
i
or (a
1
, a
2
, a
3
) are the components of
a with respect to the basis e
i
. The length (=Euclidean norm) of a vector
a is denoted a or |a|. For normalized vectors (describing only direction) the
1.2. Vectors 4
following notation is introduced:
e
a
=
a
a
, (12)
whereby a vector a can be written as a = a e
a
. Examples of normalized vectors
are the basis vectors { e
1
, e
2
, e
3
}.
To each pair of vectors a and b there corresponds a positive real number
a b = b a, called the scalar product. The scalar product can be obtained as
(see gure 2):
a b = (a, b) = a b cos , 0 (13)
Another product that is useful is the vector product ab = b a, that also
a b
e
a
b

a
a
b
b
Figure 2: Illustration of scalar product and vector product.
is illustrated in Figure 2. The result is a vector that is orthogonal to the plane
spanned by a and b (with a right-handed system) and has a length
|a b| = a b sin . (14)
By now applying the scalar and vector products to the orthonormal basis
vectors { e
1
, e
2
, e
3
}, the following results are obtained
e
i
e
j
=
ij
(i, j = 1, 2, 3) (15)
and
e
i
e
j
= e
ijk
e
k
, (16)
where

ij
=
_
1 when i = j
0 when i = j
and (17)
e
ijk
=
_
_
_
1 when ijk = 123, 231 or 312
1 when ijk = 321, 213 or 132
0 otherwise
(18)
1.2. Vectors 5
The symbols
ij
and e
ijk
are called the Kronecker delta and the permutation
symbol, respectively. They are linked by the so-called e- identity:
e
ijm
e
klm
=
ik

jl

il

jk
. (19)
It now follows that the scalar product between two vectors a and b may be
written as
a b = a
i
b
j
e
i
e
j
= a
i
b
j

ij
= a
i
b
i
(20)
Similarly, the vector product becomes
a b = a
i
b
j
e
i
e
j
= a
i
b
j
e
ijk
e
k
. (21)
The scalar product has some properties:
1. a b = b a
2. a (b +c) = a b +a c
3. (a) ( b) = a b where and are scalars.
Correspondingly, the vector product has the following properties:
1. a b = b a
2. a (b +c) = a b +a c
3. (a) ( b) = a b where and are scalars.
With the scalar product it is possible to express the length of a vector a = |a|
as follows
a = |a| =

a a =

a
i
a
i
(22)
Scalar multiplication of (11) with e
i
and application of (15) shows that the
components of a are
a
i
= e
i
a (i = 1, 2, 3) . (23)
In a given coordinate system dened by the basis vectors { e
1
, e
2
, e
3
} a vector
a can be represented by a column matrix as follows
[a] = [a
i
] = [ a
1
a
2
a
3
]
T
(24)
An example is the base vector e
1
that can be represented by the following
column matrix
[ e
1
] = [ 1 0 0 ]
T
(25)
Therefore, the scalar multiplication between two vectors can be obtained as
a b = a
i
b
i
= [a]
T
[b] . (26)
A vector must be invariant with respect to coordinate system. Assume two
dierent sets of orthonormal basis vectors { e
1
, e
2
, e
3
} and
_
e

1
, e

2
, e

3
_
. The
vector b can then be written as
b = b
i
e
i
= b

i
e

i
(27)
1.2. Vectors 6
Hence, the components b

i
can be obtained as
b

i
= e

i
e
j
b
j
(28)
In matrix notation this can be written
[b

i
] = [l
ij
] [b
j
] =
_
e

i
e
j

[b
j
] (29)
where the transformation matrix [l
ij
] is orthogonal, i.e. [l
ij
]
T
= [l
ij
]
1
.
Assignment 2 A thin bar OA with masss m and length 5a is attached
without friction in a joint at O. The bar is kept in equilibrium by two
light cables AB and AC acc to the gure. The cable AB is attached to
the bar at B with coordinates (3a; 0; 3a) and the cable AC is attached to
the bar at C with coordinates (a; 2a; 5a).
Determine the forces in
the cables AB and AC
at equilibrium (g is the
acceleration of gravity
in the negative z direc-
tion).
z
x
y
3a
4a
A m
B:(3a; 0; 3a)
C:(a; 2a; 5a)
g
O
Assignment 3 Give the component of the vector a in the rotated co-
ordinate system { e

i
}. This coordinate system is obtained from the co-
ordinate system { e
i
} by rotating around the e
3
axis according to the
gure
/6
e
1
e
2
e

1
e

2
The components of the vector a in the coordinate system { e
i
} is given
as
[1 4 3]
T
.
An example of using Matlab commands for matrix denitions (for a and b)
and scalar multiplication c = a
i
b
i
is given below:
1.3. 2nd order tensors 7
>> a=[1 2 3]; b=[4 5 6];
>> c=a*b
c =
32
Example of Matlab input le to dene e
ijk
-operator and vector product c
k
=
a
i
b
j
e
ijk
:
%definition of epsi
perm=zeros(3,3,3);
for i=1:3
for j=1:3
for k=1:3
%%%
if ( (i==1) & (j==2) & (k==3)) | ((i==2) & (j==3) & (k==1)) | ...
((i==3) & (j==1) & (k==2))
perm(i,j,k)=1;
elseif ( (i==3) & (j==2) & (k==1)) | ((i==2) & (j==1) & (k==3)) | ...
((i==1) & (j==3) & (k==2))
perm(i,j,k)=-1;
end
%%%
end
end
end
%computation of vector product c_k= a_i b_j perm_ijk
a=[1 2 3];
b=[4 5 6];
c=zeros(3,1);
for k=1:3
c(k)=0;
for i=1:3
for j=1:3
c(k)=c(k,1)+a(i)*b(j)*perm(i,j,k);
end
end
end
1.3 2nd order tensors
For further reading see Reddy 2.5 Tensors.
Now consider a linear transformation of a vector b into a
1.3. 2nd order tensors 8
a
b
T
This may be written symbolically as
a = T b. (30)
The linear operator T is called a second-order tensor, or dyad.
An example is the stress tensor from which
the traction vector t( n) can be obtained as
t( n) = n.
t( n)
n
A second-order tensor T is represented in a orthogonal coordinate system
{ e
1
, e
2
, e
3
} as
T = T
ij
e
i
e
j
, (31)
where T
ij
are the nine components of T and e
i
e
j
are the base dyads. The linear
transformation or the scalar multiplication in (30) can now be written as follows
a
i
e
i
= T
ij
e
i
e
j
b
k
e
k
= T
ij
b
j
e
i
or a
i
= T
ij
b
j
. (32)
Equation (32) represents a linear equation system, like that in equation (1)(3),
or, in matrix notation,
[a
1
a
2
a
3
]
T
=
_
_
T
11
T
12
T
13
T
21
T
22
T
23
T
31
T
32
T
33
_
_
[b
1
b
2
b
3
]
T
.
This shows why second-order tensors often are represented by 3 3 matrices.
Note that it is only the components of a, b and T in the coordinate system
{ e
1
, e
2
, e
3
} that are represented by the matrices. If the coordinate system
would be changed then also these components would change.
The components of a second-order tensor T in the coordinate system{ e
1
, e
2
, e
3
}
can be obtained as follows
e
i
T e
j
= e
i
T
kl
e
k
e
l
e
j
= T
ij
. (33)
Transpose T
T
of a 2nd order tensor T is dened as follows:
T
T
= T
ij
e
j
e
i
= T
ji
e
i
e
j
(34)
Many second-order tensors in mechanics are symmetric which means that the
tensor and its transpose are equal e.g. T
T
= T or T
ij
= T
ji
. Another type
is the skew-symmetric second-order tensors. These have the property that the
transpose of the tensor is equal to the tensor with a minus sign, e.g., T
T
= T
or T
ij
= T
ji
. Clearly, for such a tensor the diagonal elements (in a matrix
representation) must be equal to zero.
1.3. 2nd order tensors 9
Open product (also called outer product) between two vectors a and b results
in a 2nd order tensor T (also called dyad) as follows
T = T
ij
e
i
e
j
= ab = a
i
e
i
b
j
e
j
= a
i
b
j
e
i
e
j
(35)
It is noted that the open product is not cummutative i.e. ab = b a in general
but has the following properties:
1. a(b +c) = ab +ab
2. (a) ( b) = ab where and are scalars.
A special 2nd order tensor is the identity tensor :
=
ij
e
i
e
j
= e
1
e
1
+ e
2
e
2
+ e
3
e
3
(36)
with the property a = a.
Multiplication between two 2nd order T and U results in a 2nd order tensor
and is dened as
T U = T
ik
U
kj
e
i
e
j
(37)
and this multiplication has the following properties (note that A B = B A in
general):
1. A (B +C) = A B +A C
2. (A) ( B) = A B where and are scalars.
Further, by applying the transpose operator to such a product we obtain
(A B)
T
= B
T
A
T
. (38)
Double contraction between two 2nd order T and U results in a scalar and
is dened as
T : U = T
ij
U
ij
(39)
which has the following properties:
1. A : B = B : A
2. A : (B +C) = A : B +A : C
3. (A) : ( B) = A : B where and are scalars.
A 2nd order tensor is invariant with respect to coordinate system. Assume
two dierent sets of orthonormal basis vectors { e
1
, e
2
, e
3
} and
_
e

1
, e

2
, e

3
_
. The
2nd order tensor T can then be written as
T = T
ij
e
i
e
j
= T

ij
e

i
e

j
(40)
The components T

ij
can be obtained as
T

ij
= e

i
e
k
T
kl
e
l
e

j
(41)
1.4. Principal values and principal directions 10
In matrix notation this can be written
_
T

ij

= [l
ik
] [T
kl
] [l
jl
]
T
(42)
where the orthogonal transformation matrix [l
ij
] is dened in (29).
The inverse of a second-order tensor T is denoted T
1
. If T is a linear
transformation from b to a then T
1
is a linear transformation in the opposite
direction, i.e.
a = T b b = T
1
a (43)
The components T
1
ij
of the inverse T
1
is obtained by using matrix inversion
of the components T
ij
_
T
1
ij

= [T
ij
]
1
(44)
It is possible to construct tensors of any order (or rank) as follows:
A = A
ijk
e
i
e
j
e
k

In particular, fourth-order tensors are frequently used to, for example, give the
relation (material behavior) between the second-order tensors stress and strain.
Assignment 4 Given the Sherman-Morrison formula:
A
ij
=
ij
+u
i
v
j
then A
1
ij
=
ij


1 +u
k
v
k
u
i
v
j
Show that if A
ij
= B
ij
+u
i
v
j
then
A
1
ij
= B
1
ij


1 +v
k
B
1
kl
u
l
B
1
im
u
m
v
n
B
1
nj
An example of using Matlab commands for matrix denitions (for A and B)
and computing the contraction c = A
ij
B
ij
is given below:
A=[1 2 3; 4 5 6; 7 8 9]; B=[3 2 1; 6 5 4; 9 8 7];
>> c=sum(sum(A.*B))
c =
273
1.4 Principal values and principal directions
For further reading see Reddy 2.5.5 Eigenvalues and Eigenvectors of Tensors.
A second-order tensor can be, as discussed above, thought of a linear trans-
formation between vectors, i.e. a = T b. Important properties of a second-order
1.4. Principal values and principal directions 11
tensor are its eigenvectors (principal directions) end eigenvalues (principal val-
ues). Eigenvectors are dened as vectors that do not rotate upon transformation
with the second-order tensor. If n now is an eigenvector to T this can be illus-
trated as
n
n
T
This can be written as
n = T n or n
i
= T
ij
n
j
. (45)
The eigenvectors n are chosen to be of unit length whereby it is possible to
identify the length of the vector T n as the corresponding eigenvalues .
The way to nd the eigenvalues and eigenvectors is to rewrite (45) as
( T) n = 0 or (
ij
T
ij
) n
j
= 0
i
. (46)
A trivial solution to this equation is that n = 0. However, it is possible to nd
non-trivial solution if ( T) is non-invertible. From linear algebra we know
that then the determinant of the matrix [ T] must be zero, i.e.
det (T ) = 0 (47)
which is called the characteristic equation. An important theorem from linear
algebra is the spectral theorem which states that for symmetric matrices the
eigenvalues are real and the eigenvectors are orthogonal. In the current course
we will only consider eigenvalues and eigenvectors for symmetric second order
tensors (i.e. stress, strain, etc) and for such a tensor the characteristic equation
can be obtained as

T
11
T
12
T
13
T
12
T
22
T
23
T
13
T
23
T
33

= (T
11
) (T
22
) (T
33
) +
T
12
T
23
T
13
+T
13
T
12
T
23
T
2
13
(T
22
) T
2
23
(T
11
) (T
33
) T
2
12
= 0
This third order polynomial equation can be summarized as

3
I
1

2
+I
2
I
3
= 0 (48)
where the invariants of the second-order tensor T were introduced as
I
1
= T
ii
, I
2
= [T
ii
T
jj
T
ij
T
ij
] /2, I
3
= det(T
ij
) (49)
After solving the three eigenvalues
(1)
,
(2)
,
(3)
from (48) we can solve the
corresponding eigenvectors n
(1)
, n
(2)
, n
(3)
from (46).
1.5. Spatial derivatives 12
Assignment 5 a) Use the fact that eigenvectors n
(i)
of a symmetric
2nd order tensor T are orthogonal to show that T can be expressed
as
T =
(1)
n
(1)
n
(1)
+
(2)
n
(2)
n
(2)
+
(3)
n
(3)
n
(3)
b) Given that the exponential function of a scalar and a 2nd order
tensor are dened as:
exp() =

k=0

k
k!
, exp(T) =

k=0
T
k
k!
.
Show that
exp(T) = exp(
(1)
) n
(1)
n
(1)
+exp(
(2)
) n
(2)
n
(2)
+exp(
(3)
) n
(3)
n
(3)
and use this to compute exp(T) where T is represented by the the
following matrix (in a e
i
system)
[T] =
_
_
6 4 0
4 3 0
0 0 2
_
_
.
An example of using Matlab commands for matrix denitions (for A ) and
computing the eigenvalues and eigenvectors given below:
>> A=[1 2 3; 2 4 5; 3 5 6];
>> [n,lambda]=eig(A)
n =
0.7370 0.5910 0.3280
0.3280 -0.7370 0.5910
-0.5910 0.3280 0.7370
lambda =
-0.5157 0 0
0 0.1709 0
0 0 11.3448
1.5 Spatial derivatives
For further reading see Reddy 2.4.2-2.4.3
1.5. Spatial derivatives 13
A tensor eld describes how the tensor depends on the spatial location x
and the time t, e.g.
Scalar eld (such as temperature, pressure) = (x, t) or = (x
i
, t)
Vector eld (such as displacement, velocity, force) u = u(x, t) or u
i
=
u
i
(x
j
, t)
Second-order tensor eld (such as stress, strain) T = T(x, t) or T
ij
=
T
ij
(x
k
, t).
The gradient (dierential vector) operator is dened as
= e
i

x
i
(50)
By applying the gradient operator via an open product (from the left) to a
scalar eld (x, t), a vector eld u(x, t) and a second-order tensor eld T(x, t)
=

x
i
e
i
, u =
u
j
x
i
e
i
e
j
, T =
T
jk
x
i
e
i
e
j
e
k
. (51)
By applying the gradient operator via a scalar product (from the left) to a vector
eld u(x, t) and a second-order tensor eld T(x, t)
u =
u
i
x
i
, T =
T
ij
x
i
e
j
. (52)
This is also called the divergence with the following notation
div(u) = u , div(T) = T. (53)
To further compress the notation we introduce the index form of the gradient
operator,
j
= /x
j
= e
j
, or even more compactly, a subscripted comma
which for example results in:
u =
i
u
j
= u
j,i
, div(T) =
i
T
ij
= T
ij,i
. (54)
Another product with the gradient operator denes the curl of a vector eld
curl(u) = u = e
ijk

i
u
j
e
k
(55)
Later we will also apply from the right to a vector eld u(x, t) which is dened
as
u =
u
i
x
j
e
i
e
j
(56)
Assignment 6 For the scalar eld equation (in 2D) (x
1
, x
2
) = x
2
1
+
(x
2
/)
2
= 0, illustrate = 0 and in the x
1
x
2
plane for = 2
and = 1.
1.6. Divergence theorem 14
1.6 Divergence theorem
For further reading see Reddy 2.4.5
Gauss divergence theorem is an important and useful theorem, which allows
us to convert the volume integral of a divergence into a surface integral, i.e.,
_

udx =
_

n uds, (57)
where S is the closed boundary surface of V , and n is the outward normal unit
vector to S. Equivalently in index notation
_

i
u
i
dx =
_

n
i
u
i
ds. (58)
As a consequence we have the corresponding integral theorems for the divergence
of a tensor eld,
_

j
a
ji
dx =
_

n
j
a
ji
ds, (59)
and for the gradient of a scalar, vector or tensor eld,
_

i
dx =
_

n
i
ds, (60)
_

i
u
j
dx =
_

n
i
u
j
ds, (61)
_

k
a
ij
dx =
_

n
k
a
ij
ds. (62)
In practice, the name divergence theorem refers to all of these results.
Assignment 7 Proove the following formula:
_

ij
dx =
_

n
j

ij

i
ds
_

ij

i
dx.
2 Stress, motion and deformation
2.1 Stress analysis
For further reading see Reddy 4.1-4.3.2
The stress (also called traction) vector t( n) is dened as the force acting on
an area with normal n. In a point of a body the stress vector is dened as
t( n) = lim
S0
F
S
(63)
2.1. Stress analysis 15
d
S
F
n
A property of the stress vector is that it must follow Newtons third law for
action and reaction. Therefore, in the same point of a body the stress vector on
the area with normal n and normal n must be opposite. This means that
t( n) = t( n). (64)
To nd a relation between the normal n and the stress vector t( n) we study
a tetrahedral element:
n
t( n)
x
1
x
2
x
3
The tetrahedron is assumed to have the four surfaces dened as
1. normal n and area A subjected to stress vector t( n),
2. normal e
1
and area A
1
subjected to stress vector t( e
1
),
3. normal e
2
and area A
2
subjected to stress vector t( e
2
),
4. normal e
3
and area A
3
subjected to stress vector t( e
3
).
The relation between the areas A, A
1
, A
2
, A
3
and the components of n can be
conveniently derived by using the divergence theorem (show this at home!):
e
1
A
1
e
2
A
2
e
3
A
3
+ nA = 0. (65)
From these three equations we can identify that A
i
= n
i
A. Next step is now to
study equilibrium of the tetrahedron:
t( n) A+t( e
1
) A
1
+t( e
2
) A
2
+t( e
3
) A
3
= 0.
If we use the relation between the areas and Newtons third law we obtain:
t( n) = t( e
1
) n
1
+t( e
2
) n
2
+t( e
3
) n
3
. (66)
2.1. Stress analysis 16
The second-order stress tensor is dened based on t( e
i
) such that
[
ij
] = [t
j
( e
i
)] (67)
or more explicitly
_
_

11

12

13

21

22

23

31

32

33
_
_
=
_
_
t
1
( e
1
) t
2
( e
1
) t
3
( e
1
)
t
1
( e
2
) t
2
( e
2
) t
3
( e
2
)
t
1
( e
3
) t
2
( e
3
) t
3
( e
3
)
_
_
(68)
This can be graphically shown as (here 2d):
x
1
x
2
t( e
1
)

11

12

22

21
t( e
2
)
To sum up, the relation between the stress vector t and normal vector n is
obtained via the stress tensor as follows:
t = n =
T
n or t
i
= n
j

ji
=
T
ij
n
j
. (69)
This relation is the so-called Cauchys formula. As will be proven later in the
course, the stress tensor is symmetric due to principle of angular momentum i.e.
=
T
and, hence, this relation can be written as t = n. We can conclude
that if the stress tensor is known in a point of the body then it is possible to
compute the stress vector t on any plane through the point. This is called the
Cauchys stress principle.
Often the components of the stress tensor are divided into normal stresses
and shear stresses. The normal stresses are the diagonal components of the
stress tensor i.e.
11
,
22
and
33
whereas the shear stresses are the o-diagonal
components i.e.
12
,
23
and
13
. Note that the terminology normal and shear
components relate to what plane that is chosen. In the gure above the choice
of plane is dened by the normal e
1
or e
2
. In general, the normal component
of the stress on a plane with normal n is obtained from

nn
= n t = n n = : ( n n) =
ij
n
i
n
j
. (70)
Let us now adopt the concept of eigenvalues and eigenvectors for a stress
tensor . The eigenvector is a direction n that is not changed upon a scalar
multiplication with the stress tensor :
2.2. Continuum motion 17
t = n
n

This means that on a plane with the normal being an eigenvector of then the
stress vector t is parallel to the normal i.e. t = n. In other words, on such a
plane only the normal components are non-zero.
Often the stress tensor is additatively decomposed into a deviatoric

and a spherical (hydrostatic) tensor


m
as follows:
=

+
m
or
ij
=

ij
+
m

ij
(71)
with

m
=
kk
/3 and

ij
=
ij

m

ij
. (72)
Assignment 8 Given the stress tensor (here represented in a matrix
format with components in a e
1
, e
2
, e
3
system)
[] =
_
_
30 0 10
0 30 10
10 10 30
_
_
MPa
Compute the corresponding deviatoric stress tensor

. For the devia-


toric stress compute the principal stresses, principal directions, invari-
ants (acc to (49)) and the obtained stress vector on the plane with nor-
mal
n =
1

5
( e
1
+ 2 e
2
) .
2.2 Continuum motion
For further reading see Reddy 3.1-3.3.1, 3.4.2, 5.2.2
The motion of a continuum is shown in gure 3. A material particle P may
be identied by its initial (or reference) position X. The current position x, of
a material particle is then dened by a function
x
i
= x
i
(X, t) (73)
The displacement u of a particle P is dened as
u
i
= x
i
X
i
(74)
A key quantity that describes the deformation of a body is the deformation
2.2. Continuum motion 18
O
P
P
X
x
u
Figure 3: Illustration of motion of a continuum.
gradient F. The deformation gradient describes the relation between a line
element dX in the initial (undeformed) body and the corresponding line element
dx in the current (undeformed) body, i.e.
dx = F dX or F
ij
=
x
i
X
j
=
ij
+
u
i
X
j
or F = x
0
(75)
which is also illustrated in gure 4. Based on the deformation gradient F a
dX
dx
F
Figure 4: Illustration of deformation gradient.
number of strain measures can be dened. An example is the frequently used
Green-Lagrange strain E dened as follows:
(E)
ij
=
1
2
_
F
T
ik
F
kj

ij
_
=
1
2
_
u
i
X
j
+
u
j
X
i
+
u
k
X
i
u
k
X
j
_
(76)
For the special case of small deformations E approaches the usual small strain
2.2. Continuum motion 19
tensor , i.e.
(E)
ij

1
2
_
u
i
X
j
+
u
j
X
i
_

1
2
_
u
i
x
j
+
u
j
x
i
_
=
ij
(77)
Physical quantities can be described in either a Lagrangian (or sometimes
called material) or Eulerian description:
Lagrangian description of scalars, vectors and second-order tensors:
= (X
i
, t) , u
i
= u
i
(X
j
, t) , T
ij
= T
ij
(X
k
, t) (78)
Eulerian description of scalars, vectors and second-order tensors:
= (x
i
, t) , u
i
= u
i
(x
j
, t) , T
ij
= T
ij
(x
k
, t) (79)
The velocity eld has a special status, in that it not only enters in the various
conserved quantities but also denes the motion of the material elements for
which the conservation laws are dened. A material element moves and deforms
according to this velocity eld.
x
i
(X, t)
t
= v
i
(x(X, t) , t) , (80)
noting that v
i
(x, t) is a fundamental continuum eld variable. Next follows
three examples to illustrate the introduced concepts regarding motion.
2.2. Continuum motion 20
Example 1 Simple shear of a quadratic disc (side length h
0
) where the upper
boundary moves horizontally with velocity v
0
:
e
1
e
2
The motion can be expressed as:
_
_
_
x
1
(X
1
, X
2
, X
3
, t) = X
1
+X
2
v
0
t/h
0
x
2
(X
1
, X
2
, X
3
, t) = X
2
x
3
(X
1
, X
2
, X
3
, t) = X
3
whereby the velocity v can be obtained, in Lagrangian description, as:
v
i
=
_
_
X
2
v
0
/h
0
0
0
_
_
By using the expression for the motion the velocity can be written in an Eulerian
description as
v
i
=
_
_
x
2
v
0
/h
0
0
0
_
_
Based on the expression for the motion we can also obtain the deformation
gradient F as
F
ij
=
_
_
1 v
0
t/h
0
0
0 1 0
0 0 1
_
_
and the Green Lagrange strain E
E
ij
=
1
2
_
F
T
ik
F
kj

ij
_
= ... =
1
2
_
_
0 v
0
t/h
0
0
v
0
t/h
0
(v
0
t/h
0
)
2
0
0 0 0
_
_
The displacement vector u = x X is given as:
u
i
=
_
_
X
2
v
0
t/h
0
0
0
_
_
whereby the small strain tensor becomes

ij
=
1
2
_
u
i
X
j
+
u
j
X
i
_
=
_
_
0 (v
0
t/h
0
)/2 0
(v
0
t/h
0
)/2 0 0
0 0 0
_
_
2.2. Continuum motion 21
Example 2 Pure elongation of a quadratic disc (side length h
0
) where the upper
boundary moves vertically with velocity v
0
:
e
1
e
2
The motion can be expressed as:
_
_
_
x
1
(X
1
, X
2
, X
3
, t) = X
1
x
2
(X
1
, X
2
, X
3
, t) = X
2
+X
2
v
0
t/h
0
x
3
(X
1
, X
2
, X
3
, t) = X
3
whereby the velocity v can be obtained, in Lagrangian description, as:
v
i
=
_
_
0
X
2
v
0
/h
0
0
_
_
By using the expression for the motion the velocity can be written in an Eulerian
description as
v
i
=
_
_
0
x
2
v
0
/(h
0
+v
0
t)
0
_
_
Based on the expression for the motion we can also obtain the deformation
gradient F as
F
ij
=
_
_
1 0 0
0 1 +v
0
t/h
0
0 0 1
_
_
and the Green Lagrange strain E
E
ij
=
1
2
_
F
T
ik
F
kj

ij
_
= ... =
_
_
0 0 0
0 (v
0
t/h
0
)
2
/2 +v
0
t/h
0
0
0 0 0
_
_
The displacement vector u = x X is given as:
u
i
=
_
_
0
X
2
v
0
t/h
0
0
_
_
2.2. Continuum motion 22
whereby the small strain tensor becomes

ij
=
1
2
_
u
i
X
j
+
u
j
X
i
_
=
_
_
0 0 0
0 v
0
t/h
0
0
0 0 0
_
_
Example 3 Pure rotation around the left corner of a quadratic disc (side length
h
0
) with rotational velocity :
R
t
e
1
e
2
An arbitrary points initial location in the disc is described by the distance R =
_
X
2
1
+X
2
2
to the left corner and angle
0
= atan(X
2
/X
1
) (from the e
1
axis).
During rotation the angle changes with rotation according to =
0
+t whereby
the motion can be expressed as:
_
_
_
x
1
(X
1
, X
2
, X
3
, t) = R cos()
x
2
(X
1
, X
2
, X
3
, t) = R sin()
x
3
(X
1
, X
2
, X
3
, t) = X
3
which can be (after some manipulations) written as
_
_
_
x
1
(X
1
, X
2
, X
3
, t) = X
1
cos(t) X
2
sin(t)
x
2
(X
1
, X
2
, X
3
, t) = X
1
sin(t) +X
2
cos(t)
x
3
(X
1
, X
2
, X
3
, t) = X
3
The velocity v can be obtained, in Lagrangian and Eulerian description, as :
v
i
=
_
_
(X
1
sin(t) X
2
cos(t))
(X
1
cos(t) X
2
sin(t))
0
_
_
=
_
_
x
2
x
1
0
_
_
Based on the expression for the motion we can also obtain the deformation
gradient F as
F
ij
=
_
_
cos(t) sin(t) 0
sin(t) cos(t) 0
0 0 1
_
_
and the Green Lagrange strain E
E
ij
=
1
2
_
F
T
ik
F
kj

ij
_
= ... =
_
_
0 0 0
0 0 0
0 0 0
_
_
2.2. Continuum motion 23
The displacement vector u = x X is given as:
u
i
=
_

_
X
1
(cos(t) 1) X
2
sin(t)
X
1
sin(t) +X
2
(cos(t) 1)
0
_

_
whereby the small strain tensor becomes

ij
=
1
2
_
u
i
X
j
+
u
j
X
i
_
=
_
_
cos(t) 1 0 0
0 cos(t) 1 0
0 0 0
_
_
3. Field equations 24
Later we will also need time derivatives such as m,

P
i
,

K and

U that enter
in the fundamental conservation equations (85)(88) of a material element that
occupies a spatial domain that changes with time. In order to evaluate such
time derivatives, we need to introduce the concept of material derivative and an
integral theorem known as Reynolds transport theorem. A material derivative
may be thought of as the rate of change of any given eld quantity, as observed
by a particle, i.e., the derivative is taken at a material point, moving with the
velocity v
i
. The material time derivative of a eld quantity, described in an
Lagrangian description, (x, t), denoted as D/Dt,

or d/dt, is
D(X, t)
Dt
=

(X, t) =
(X, t)
t
, (81)
whereas the material time derivative of a eld quantity described in an Eulerian
description is obtained by the chain rule:

(x(X, t), t) =
(x(X, t), t)
t
=
(x, t)
t
+
(x, t)
x
i
x
i
(X, t)
t
. (82)
Thus
d
dt
=

t
+v
i

x
i
. (83)
Reynolds transport theorem relates the material time derivative of a volume
integral, where the integration volume moves with the ow, to the spatial time
derivative, where the volume is xed in space:
d
dt
_

dx =
_

_
d
dt
+
v
i
x
i
_
dx. (84)
Here, is an arbitrary integrand. This theorem, which is given here without
proof, is most useful for deriving the eld equations.
Assignment 9 Use Reynolds transport theorem to show that the rate
of change of volume per unit volume equals the trace of the velocity
gradient (or divergence of velocity), v
i,i
= lim
V 0

V /V.
3 Field equations
For further reading see Reddy 5.1, 5.2.3, 5.3.1, 5.3.3., 5.4.1, 5.4.2.
3.1. Physical conservation principles 25
3.1 Physical conservation principles
We consider a material element (a small part of a solid or uid) occupying a
region at a point of time t. The material element has a mass M, a momentum
P, an angular momentum N, a kinetic energy K and an internal energy U. The
internal energy is simply all energy that is not kinetic, for example thermal or
elastic energy. The material element is governed by the fundamental principles
of conservation of mass, momentum, angular momentum and energy:

M = 0, (85)

P
i
= F
i
, (86)

N
i
= M
i
(87)

K +

U =

W +

H, (88)
where the the quantities on the right are the rates of input of mass momentum
and energy, respectively. Thus F
i
and M
i
are the total force and total moment
(torque), respectively, acting on the region,

W is the work input rate and

H is
the heat input rate. Equation (86) is also known as Newtons second law, and
(88) is known as the rst law of thermodynamics.
3.2 Field variables
The idea of a continuous medium is based on the representation of the above
quantities by piecewise continuous quantities, or elds. One thus introduces a
continuous density eld (x, t), a velocity eld v
i
(x, t) and an internal energy
eld e (x, t), such that
M =
_

dx, (89)
P
i
=
_

v
i
dx, (90)
N
i
=
_

e
ijk
x
j
v
k
dx, (91)
K =
_

1
2
v
i
v
i
dx, (92)
U =
_

e dx. (93)
These expressions are approximations of the real mass, momentum, angular
momentum, kinetic energy and internal energy in , and the approximation is
good as long as is suciently large.
3.2. Field variables 26
The supply terms on the right hand side of (85)(88) may be expressed as
F
i
=
_

f
i
dx +
_

t
i
ds, (94)
M
i
=
_

e
ijk
x
j
f
k
dx +
_

e
ijk
x
j
t
k
ds, (95)

W =
_

f
i
v
i
dx +
_

v
i
t
i
ds, (96)

H =
_

E dx +
_

q
n
ds, (97)
where t is the (vectorial) force per unit surface area acting on the boundary ,
q
n
is the (scalar) outux of heat per unit surface area on , f is the (vectorial)
body force per unit mass (e.g. gravity force), and E is a possible (scalar) internal
heat source per unit mass.
In order that all variables be dened on the same 3-dimensional space, one
must eliminate the surface densities t
i
= t
i
(x, n) and q
n
= q
n
(x, n). Thus the
Cauchy stress theorem states the existence of a tensor
ij
=
ij
(x), called the
Cauchy stress tensor, such that
t
i
(x, n) = n
j

ji
(x) . (98)
Similarly Stokes heat ux theorem states the existence of a heat ux vector
eld q = q (x), such that
q
n
= q
i
(x) n
i
. (99)
Using (98) and (99), and the divergence theorem, equations (94), (96) and (97)
may be expressed as ( below we skip the balance on angular momentum principle
and instead leave the result of it as Problem 10)
F
i
=
_

(f
i
+
ji,j
) dx, (100)

W =
_

_
f
i
v
i
+ (v
i

ji
)
,j
_
dx, (101)

H =
_

(E q
i,i
) dx. (102)
Thus all the quantities in the fundamental laws (85)(88) are now expressed
in terms of 5 eld variables in 3-dimensional space:
, v
i
,
ij
, u, q
i
. (103)
The elds f
i
and E are not considered as variables, but externally prescribed
elds. Since v and q are vectors and is a symmetric tensor (see Problem 10),
the total number of (scalar) variables is 14. We will thus need, in the general
case, 14 (scalar) equations. To solve any given problem, however, one usually
needs only some of those variables and, consequently, only some of the eld
equations.
3.3. Continuity equation 27
Recall the principles of conservation of mass momentum and energy.

M = 0, (104)

P
i
= F
i
, (105)

K +

U =

W +

H. (106)
Using Reynolds transport theorem, the rates of M, P
i
, K and U become

M =
d
dt
_

dx =
_

( +v
i,i
) dx, (107)

P
i
=
d
dt
_

v
i
dx =
_

( v
i
+ v
i
+v
i
v
j,j
) dx, (108)

K =
1
2
d
dt
_

v
i
v
i
dx =
_

_
v
i
v
i
+
1
2
v
i
v
i
( +v
j,j
)
_
dx, (109)

U =
d
dt
_

e dx =
_

( e + e +ev
i,i
) dx. (110)
The supply terms F
i
, dW/dt and dH/dt are given by (101)(102):
F
i
=
_

(f
i
+
ji,j
) dx, (111)

W =
_

_
f
i
v
i
+ (v
i

ji
)
,j
_
dx, (112)

H =
_

(E q
i,i
) dx. (113)
Assignment 10 Use the principle of angular momentum to show that
the stress tensor is symmetric. Hint, see Reddy 5.3.3.
3.3 Continuity equation
Mass conservation (104) together with equation (107) immediately yields
_

( +v
i,i
) dx = 0. (114)
The continuum hypothesis implies that the integration volume V may be taken
as arbitrarily small. This step is called localisation and yields the equation of
mass, or continuity:
+v
i,i
= 0. (115)
Note that this is the continuum version of the mass conservation equation (104).
By combining Reynolds transport theorem and the continuity we obtain the
useful relation
d
dt
_

dx =
_

d
dt
dx. (116)
3.4. Momentum equation 28
for the arbitrary integrand . This can immediately be used for the rates of P
i
,
K and U:

P
i
=
d
dt
_

v
i
dx =
_

v
i
dx, (117)

K =
1
2
d
dt
_

v
i
v
i
dx =
_

v
i
v
i
dx, (118)

U =
d
dt
_

e dx =
_

e dx. (119)
3.4 Momentum equation
Introducing (117) and (111) into Newtons second law (105), and using the
continuity equation, one obtains
_

( v
i

ji,j
f
i
) dx = 0. (120)
Localisation yields the momentum equation:

ji,j
+f
i
= v
i
. (121)
This is the continuum version of Newtons 2nd law, equation (105). The impor-
tant case where v
i
= 0 is called static equilibrium:

ji,j
+f
i
= 0. (122)
3.5 Energy equation
By using (118) and (119) and the momentum equation into (112), the rst law
of thermodynamics (106), becomes
_

[ e v
i,j

ij
+q
i,i
E] dx = 0. (123)
The localised form is known as the energy equation,
e v
i,j

ij
+q
i,i
= E. (124)
This is the continuum version of the 1st law of thermodynamics, equation (106).
Assignment 11 Carry out the steps to obtain (123). Also present the
intermediate results for

K,

U ,

W and

H.
4. Constitutive equations 29
4 Constitutive equations
The eld equations derived above correspond to 5 scalar equations (1 for the
mass equation, 3 for the momentum equation and 1 for the energy equation).
We thus need another 9 equations to solve for all the 14 variables (cf. equa-
tion 103). These additional equations are called constitutive equations, because
they depend on the constitution of the material. The role of the constitutive
equations is to relate the ux quantities and q, to the other variables.
The constitutive equations can be described, in a more general fashion, by
including two more variables: entropy, temperature, and one more equation
(or rather inequality), the second law of thermodynamics. In practice, the role
of the second law of thermodynamics is that it restricts the possible forms of
constitutive equations.
For simplicity, we avoid introducing entropy here. The underlying theory
of constitutive equations is beyond the scope of this course. Here we merely
present the simplest and most common constitutive equations for solids and u-
ids. Notice, however, that there are many other and more complicated types of
material behaviour, and the development of new constitutive equations, material
mechanics, is an important research topic.
We thus only introduce the constitutive equations that we are going to need
later, and the reason for introducing them at this point is to demonstrate that
they indeed close the system of equations of continuum mechanics.
4.1 Fouriers law of thermal conductivity
For further reading see Reddy 6.4.2.
Heat is conducted in the same way in solids and uids. We shall assume
that any heat ux is driven by a temperature gradient so that there exists some
relation,
q
i
= q
i
() with q
i
(0) = 0. (125)
For suciently small temperature gradients, the function q
i
() may be taken
to be linear. If the material is further assumed to be isotropic, then the consti-
tutive equation takes the form
q
i
= k
,i
(126)
where the linear coecient k is the thermal conductivity. Equation (126) is
known as Fouriers law. Since we have here introduced a new dependent eld
variable, , we need an additional relation. In the so-called decoupled theory,
the temperature may be related directly to the internal energy, typically by the
linear relation
e = c
p
(127)
4.2. Viscous uids 30
where the constant c
p
is know as the heat capacity of the material. Then,
usually, the internal energy e is eliminated from the equations, in favour of the
temperature .
4.2 Viscous uids
For further reading see Reddy 6.3.3.
In general for uids it is assumed that the stress is a function of the pressure
and the velocity gradient:
= (p, v) . (128)
It cannot be a function of the velocity itself, since a rigid motion cannot produce
stress. The dependence of the velocity gradient is often introduced via the rate
of strain tensor D
ij
that is dened as follows
D
ij
=
1
2
(v
i,j
+v
j,i
) . (129)
For constitutive modelling of viscous uid it is assumed that the stress
ij
is additatively decomposed according to

ij
= p(, )
ij
+
ij
(130)
where the pressure p follows the ideal gas law p(, ) = R and
ij
is the
viscous stress.
For the case of an isotropic Newtonian viscous uid we assume that the
viscous stress
ij
is linear in terms of the strain rate tensor D
ij
as follows

ij
=

ij
D
kk
+ 2

D
ij
, (131)
Thereby the stress becomes:

ij
= p(, )
ij
+

ij
D
kk
+ 2

D
ij
(132)
The mechanical pressure p
mech
can now be obtained as:
p
mech
=
1
3

ii
= p(, )
_

+
2
3

_
D
kk
(133)
If we introduce the Stokes condition that p
mech
= p(, ) then we can for a
Newtonian viscous uid obtain that

= 2

(134)
where

and D

are the deviatoric stress and deviatoric strain rate tensor,


respectively. Stokes condition means that the pressure in the uid is strain
rate independent.
Assignment 12 Show that Stokes condition leads to equation (134)
and, for a compressible uid D
ii
= 0, give the relation between

and

.
4.3. Elastic solids 31
4.3 Elastic solids
A solid is called elastic if the stress is a unique function of the deformation.
(It cannot be a function of the displacement itself, since a rigid displacement
cannot produce stress.) Thus
= (u) , (135)
where
u(x, t) = x X(x, t) u = v.
Pressure is not needed among the arguments, since the solid is assumed to be
elastically compressible, which implies that the pressure is a function of the
deformation (u). For suciently small displacement gradients, the relation
(135) should be linear. If, in addition, we restrict ourselves to materials whose
response is isotropic, i.e. independent of material orientation, then the consti-
tutive equation must have the form,

ij
=
ij
u
k,k
..

kk
+(u
i,j
+u
j,i
)
. .
2
ij
. (136)
This is Hookes law, in terms of Lames constants, and . The constant is
known as the shear modulus.
The assumption of small displacement gradients also implies, by virtue of
the continuity equation, that
constant. (137)
Assignment 13 The equations
u
i,jj
+ ( +) u
j,ji
+f
i
= u
i
.
are called Naviers equations and may be used to solve elastodynamic
problems with displacement-type boundary conditions. Derive these
equations by combining the momentum equation and Hookes law.

S-ar putea să vă placă și