ES 240

Solid Mechanics

Homework 6
Due Friday, 2 November
22. Stress-strain relations under the plane strain conditions
Starting from a familiar form of Hookes law in three dimensions, namely, x x y z / E
, xy 21 xy / E , etc., show that under the plane strain conditions the stress-strain relation
takes the following form:
x

0 x
1
E

1
0 y .
y 1 1 2
xy
0
0
0.5 xy

23. Getting weak: derive weak statements from differential equations

If we call the principle virtual work (PVW) a weak statement of equilibrium, we may as well call
the equilibrium equations a strong statement. In class I stated the PVW without telling you how it
is discovered. I dont really know the history, but a more useful thing to tell you is how to
construct a weak statement if you already know the differential equation. This way, you can
extend what you have learned in solid mechanics to other fields of engineering.
Stress field. As an example, start with the equilibrium equations:
ij

bi 0 , in V
x j
ij n j ti , on S

(a1)
(a2)

Instead of requiring that these equations hold for every point in volume and on surface, it is
equivalent that we require that
ij

bi u i dV t i ij n j u i dA 0
(b)

x j

u
for all virtual displacement field
i (i.e., any field what so ever). The first integral extends over

the volume of the body, and the second integral extends over the surface of the body. The left
hand side of (b) is called the weighted residual.
Recall the divergence theorem, and we obtain that
ij u i
ij
u i
ij
dV
ui dV

x j

x j
u i
ij n j u i dA ij
dV
x j

x j

Recall that
1 ui u j
ij

2 x j
xi
and that the stress is a symmetric tensor. We obtain that
u i
ij x j dV ij ij dV .
Combining the above steps, we obtain that

ES 240

Solid Mechanics

ij

ij dV =

b u dV
i

t u dA .
i

(c)

This is PVW. The statements (a), (b) and (c) are all equivalent.
Temperature field. The above method can be applied to other fields. To have some appreciation
of these applications, consider heat conduction in a three dimensional body.
First recall the governing equations. Let T x, y , z , t be the temperature field, and J x, y , z , t
be the vector field of heat flux (i.e., energy across unit area per unit time).
Material law. Fouriers law relates the heat flux to the temperature gradient:
J x k

T
x

J y k

T
y

J z k

T
z

where k is the heat conductivity.

Energy balance. Let be the mass density, and c be the heat capacity (i.e., the energy needed
to increase the temperature per unit mass per unit degree). For a unit volume of the body to
change temperature by dT, the energy needed is cdT . Energy balance requires that
c

T J x J y J z

0
t
x
y
z

in the volume of the body.

The body exchanges heat with the ambient by convection. Let h be the convection coefficient
(energy per area per time per degree). Energy balance requires that
nx J x n y J y nz J z h T Ta

on the surface of the body. Here nx , n y , nz is the unit vector normal to the surface, and Ta is
the ambient temperature.
Action items. (a) Start with a weighted residual and derive the equivalent statement. Let be
an arbitrary scalar field. Show that the energy balance equations are equivalent to requiring that
the equation
T

dV + h T Ta dA =0,
Jx
Jy
Jz
t
x
y
z
hold true for every field . The first integral is over the volume of the body. The second integral

is over the surface of the body. (b) Outline a finite element method for heat conduction.
24. Potential energy and the Rayleigh-Ritz Method
Here is yet another way to do pretty much the same thing. Applying a constant external force to a
body is equivalent to hanging a weight. For example, consider the structure in the figure. Assume
the rope that hangs the weight is inextensible. When the node moves by displacement u, the
weight P drops by the same distance, so that the potential energy of the weight changes by Pu .
Now regard the body and the weights as a single system. The potential energy of this system,
, is the sum of the elastic energy in the body and the potential energy of the weights. We will call
the potential energy due to the fixed load the work potential.
Consider a three-dimension elastic body, subject to the body force b in the volume, and the
traction t on one part of the surface of the body, St . On the other part of the surface, Su , the
displacement is prescribed. The work potential is
WP bi u i dV t i u i dA .
2

ES 240

Solid Mechanics

The first integral extends over the volume of the body.

The second integral extends over the part of the surface
St , where the traction is prescribed. Note that b and t
are known external forces. The displacement field, u, is
unknown, except for its prescribed values on the part of
the surface of the body, Su . For any displacement field
u, one can calculate the work potential. A relation that
maps a field to a number is known as a functional. The
work potential is a functional of the displacement field.
The strain field relates to the displacement field as usual:
ij

1 u i u j

2 x j xi

1
C ijpq ij pq dV .
2

The elastic energy in the body is

The integral extends over the volume of the body. The elastic energy is also a functional of the
displacement field.
Potential energy is a functional of displacement. By definition, the potential energy is the sum
of the elastic energy and the work potential:
u

1
C ijpq ij pq dV bi u i dV t i u i dA .
2

The potential energy is a functional of the displacement field. The body force is prescribed over
the volume of the body, and the traction is prescribed on the surface St . The displacement need
not be the actual displacement field occurring in the body, but can be any field that satisfies the
prescribed displacement on the surface Su . The first two integral extends over the volume of the
body. The third integral extends over the area St .
Principle of minimum potential energy. The principle of minimum potential energy states that,
of all displacement fields that satisfy the prescribed values on the surface of a body, the
displacement field corresponding to equilibrium minimizes the potential energy.
Proof. Let u be the displacement field corresponding to equilibrium, and h be a variation in the
displacement field. The variation is arbitrary, except that h = 0 on Su . For the virtual
displacement u + h, the strain is calculated by inserting u + h in the usual displacement relation.
The linearity requires that u h u h . A direct calculation shows that
u h u

u i hi u p h p
u i u p
1
cijpq
cijpq
dV bi hi dV t i hi dA

2
x j
x q
x j x q

u p hi
hi h p
1
cijpq
dV cijpq
dV bi hi dV t i hi dA

2
x j x q
x q x j

Because u is the equilibrium displacement, ij cijpq u p / x q . According to the PVW, the last
three terms cancel each other. Consequently,
u h u

hi h p
1
cijpq
dV .

2
x j x q

This difference is always positive unless h vanishes everywhere in the body.

ES 240

Solid Mechanics

Rayleigh-Ritz Method. The principle of minimum potential energy calls for a search of the
winner among all displacement fields that satisfy the displacement boundary conditions. Now if
we limit the scope of the search to a subset of the admissible displacement fields, we will not find
the real winner, i.e., the actual displacement field, but an approximate. Of the displacement fields
in the subset, the displacement field that we select minimizes the potential energy.
Here is the Rayleigh-Ritz method that implements the idea. For simplicity, we assume that the
boundary condition that the displacements vanish on Su . Let h1, h2, hn be a set of known
virtual displacement fields. A linear combination is also a virtual displacement field:
u a1h1 a2h 2 ... anh n ,
where a1 , a2 ,..., an are arbitrary numbers. This gives us a family of functions. In the jargon of
linear algebra, we say that h1, h2,, hn is a set of bases, and all their linear combinations form a
space. This is a subspace in that there are other functions that cannot be represented this way.
To select the winner in this subspace, we will vary the coefficients a1 , a2 ,..., an to minimize the
potential energy.
Linearity requires that the strain be given as the sum
u a1 h1 a2 h 2 ... an h n .
A direct calculation gives that
u

where

1
Fi ai ,
Kij ai a j
2 i, j
i

h D h dV
h bdV h tdA .
T

K ij
Fi

T
i

T
i

Minimizing the potential energy, we set / am 0 for m = 1, 2, , n. This gives a set

of algebraic equation:

K
j

ij

a j Fi .

Solve for a1 , a2 ,..., an , and we obtain an approximate equilibrium displacement field:

u a1h1 a2h 2 ... anh n .
Action items. An infinite body has a spherical cavity, radius R and subject to an internal
pressure p.
(a) Express the potential energy of the body as a functional of the radial displacement field u r .
(b) Functions like r 1 , r 2 ,... are all virtual displacement fields. Assume a virtual displacement
field
A
u 2.
r
Use the Rayleigh-Ritz method to determine A.
(c) Calculate the corresponding stress field. Compare with the exact solution to the problem.
Explain your findings.
25. Constant strain triangle node
We want to formulate the finite 1element method to solve a plane-elasticity problem. Divide the

plane into triangular elements. Assign

to each node, and a number
to each element.
x1 , ay1number

node
x2 ,2y2

node
2

4
x

x3 , y3

node

node
3 (0,

(0, 1)

node
1
(1,0)

ES 240

Solid Mechanics

(a) Shape functions. Consider one triangle. Label its three nodes, counterclockwise, as 1, 2, 3.
In the global coordinate system, the three nodes have coordinates x1 , y1 , x2 , y2 , x3 , y3 . Map
a triangle on a different plane, , , to the triangle in the x, y plane. The following functions
map a point in the , plane to a point in the x, y plane:

x N1 x1 N 2 x2 N 3 x3
y N1 y1 N 2 y2 N 3 y3

Show that the shape functions are

N1 ,

N2 ,

N3 1 .

(b) The Jacobian matrix. Calculate the Jacobian matrix of the coordinate transformation,
namely,
x y

J
.
x y

1
Also calculate det J and J .
(c) Displacement field and its gradient. Let the displacements at the three nodes be u1 , v1 ,
u2 , v2 , u3 , v3 . Interpolate the displacement (u,v) of a point inside the triangle as
u N1u1 N 2u 2 N 3u3

v N1v1 N 2 v2 N 3v3

Calculate the displacement gradients in the x, y plane: u / x, u / y , v / x, v / y .

(d) Strain field. The strain column of a point in the element is linear in the nodal displacement
column. We write
Bq .
Work out the entries to the matrix B. You will find the strain field in this element is constant.
This element is known as the constant strain triangle.
(e) Stress field. The stress column is also linear in the displacement column:
DBq .
Give the stiffness matrix.
(f) Stiffness matrix. Calculate the stiffness matrix of the element.
(g) Force column. Calculate the force column of the element.