FINITE ELEMENT METHOD

ME751
Mechanical IV/II
Chapter 4

Continuum Problems
(Variational Principle & Need for Weighted Integral
Statement)
4.1 INTRODUCTION
Problems in engineering and science fall into two fundamentally
different categories, depending on which point of view we adopt.

One point of view is that all matter consists of discrete particles that
retain their identity and nature as they move through space. Their
position at any instant is given by coordinates in some reference
system, and these coordinates are functions of time ─ the only
independent variable for any process. This viewpoint, known as the
rigid body viewpoint, is the basis for Newtonian particle mechanics.

2
FINITE ELEMENT METHOD
The other viewpoint, the one that we shall use, stems from the
continuum rather than the discrete molecular or particle approach to
nature. In the continuum viewpoint we say that all bodies of interest
are continuous at all points in space to the extent that all field
quantities describing the state of the body are sufficiently
differentiable in the independent variables of the continuum, space,
and time. This allows us to focus our attention on one point in space
and time and observe the phenomena occurring there.

Continuum problems are concerned with the fields of temperature,

stress, mass concentration, displacement and potentials, to name just
a few examples. These problems arise from phenomena in nature that
are approximately characterized by partial differential equation and
their boundary conditions.

3
FINITE ELEMENT METHOD
4.2 SOME METHODS FOR SOLVING CONTINUUM
PROBLEMS
There are a number of approaches to the solution of linear and
nonlinear boundary value problems, and they range from completely
analytical to completely numerical.

Some of these methods are:

1. Direct Integration (Exact Solutions)
• Separation of variables
• Similarity solutions
• Fourier and Laplace transforms

4
FINITE ELEMENT METHOD
2. Approximate Solutions
• Perturbation
• Power series
• Probability schemes
• Method of weighted residuals
• Ritz method
• Finite difference techniques
• Finite element method

5
FINITE ELEMENT METHOD
4.3 VARIATIONAL FORMULATIONS
Often continuum problems have a different, but equivalent
formulation: a differential formulation and a variational formulation.

In the differential formulation, the problem is to integrate a

differential equation or a system of differential equations subjected to
specified boundary conditions.

In the variational formulation the problem is to find the unknown

function or functions that extremize (minimize or maximize) or make
stationary a functional or system of functionals subjected to the same
specified boundary conditions.

The two formulations are equivalent because the functions that

satisfy the differential equation system and its boundary conditions
also extremize or make stationary functionals.
6
FINITE ELEMENT METHOD
4.4 NEED FOR WEIGHTED-INTEGRAL STATEMENTS
In almost all approximate methods used to determine the solution of
differential and/or integral equations, we seek a solution in the form

………. (1)

where u represents the solution of a particular differential equation

and associated boundary conditions, UN is its approximation that is
represented as a linear combination of unknown parameters cj and
known functions fj of position x in the domain W on which the
problem is posed.

The approximate solution UN is completely known only when cj are

known. Thus, we must find a means to determine cj such that UN
satisfies the equations governing f.
7
FINITE ELEMENT METHOD
If somehow we can find fN that satisfies the differential equation at
every point x in the domain W and conditions on the boundary G of
W, then UN(x) = u(x), which is the exact solution of the problem.
Of course, approximate methods are not about problems for which
exact solutions can be determined by some methods of mathematical
analysis; the role of approximate methods is to find an approximate
solution of the problems that do not admit analytical solutions.

When the exact solution cannot be determined, the alternative is to

find a solution UN that satisfies the governing equations in an
approximate way.

In the process of satisfying the governing equations approximately,

we obtain N algebraic relations among the N parameters c1, c2, c3,
…, cN.
8
FINITE ELEMENT METHOD
 Consider the problem of solving the differential equation
………. (2a)

subjected to the boundary conditions

………. (2b)

where a(x), c(x) and f(x) are known functions, u0 and Q0 are known
parameters, and u(x) is the function to be determined. The set of a(x),
c(x), f(x), u0 and Q0 is called problem data.

We seek an approximate solution over the entire domain in

the form

………. (3)
9
FINITE ELEMENT METHOD
where the cj are coefficients to be determine and fj(x) and f0(x) are
functions chosen such that the specified boundary conditions of the
problem are satisfied by the N-parameter approximate solution UN.

Note that the particular form of in equation (3) has two parts: one
containing the unknowns that is termed the homogeneous
part and the other is the nonhomogeneous part (f0) that has the sole
purpose of satisfying the specified boundary conditions of the
problem.

Since (f0) satisfies the boundary conditions, the sum must

satisfy, for arbitrary, the homogeneous form of the boundary
conditions

10
FINITE ELEMENT METHOD
To be more specific, let L = 1, u0 = 1, Q0 = 0, a(x) = x, c(x) = 1, f(x)
= 0 and N =2. Then we choose the approximate solution in the form

that satisfies the boundary conditions, equation (2b), of the problem

for any values of c1 and c2 because

………. (4)
To make U2 satisfy the differential equation (2a), we must have

………. (5) 11
FINITE ELEMENT METHOD
Since this expression must be zero for any value of x, the coefficients
of the various powers of x must be zero

These relations are inconsistent; hence, there

is no solution to the equations.

On the other hand, we can require the approximate solution UN to

satisfy the differential equation (2a) in the weighted-integral sense,

………. (6)

where  denotes the left side of the equality in equation (5) and is
called the residual,
………. (7)
12
FINITE ELEMENT METHOD
wi(x) is called a weight function. From equation (6), we obtain as
many linearly independent equations as there are linearly
independent functions wi(x).

For example, in the present example, if we take w1 = 1 and w2 = x,

we obtain

On simplifying,
222 100
c1 = c2= -
23 23

which provides two linearly independent equations for c1 and c2.

13
FINITE ELEMENT METHOD
4.5 ELEMNETS OF CALCULUS OF VARIATIONS
Calculus of variation is a branch of mathematics that deals with
extrema and stationary behavior of functionals. There are many
problems that can be posed only as finding the extremum of
functionals, such as the Brachistochrone problem, geodesic problem,
and isoperimetric problem.

The field of solid and structural mechanics heavily depends on the

use of energy principles that are stated in terms of finding the
extrema or stationary values of functionals to construct finite
element models (e.g., the principle of minimum total potential
energy, the principle of total complementary energy, and the Hu-
Washizu and Hellinger-Reissner mixed variational principles).

The direct variational methods like the Ritz method use variational
principles to obtain approximate solutions directly, bypassing the
derivation of the governing equations. 14
FINITE ELEMENT METHOD
4.5.1 Variational Operator and First Variation
Consider the functional F(x, u, u’). For an arbitrarily fixed value of
independent variable x, F depends on u and u’.

The change ev in u, where e is a

constant and v is a function, is
called the variation of u is denoted
by du, as shown in Figure.
………. (8)

where the operator d is called the

variational operator.

15
FINITE ELEMENT METHOD
The variation du of a function u represents an admissible change in
the function u(x) at a fixed value of the independent variable x. If u is
specified at a point (usually on the boundary), the variation of u
should satisfy the homogeneous form of the boundary conditions of
u. The variation du represents a virtual but admissible change in u.
Associated with this change in u (i.e., when u is changed to u + ev),
there is change in F,
………. (9)

Expanding in powers of e gives (treating u + ev and u’ + ev’ as

dependent functions)

………. (10)
16
FINITE ELEMENT METHOD
The first variation of F is defined by

………. (11)

Thus, the variation of F can be written in terms of the variations of

the dependent variable u and its derivatives u’.

17
FINITE ELEMENT METHOD
Note the analogy between the first variation (11), and the total
differential of F,
………. (12)

Since x is not varied during the variation of u to u + du, dx = 0 and

the analogy between dF and dF becomes apparent, i.e., d acts as a
differential operator with respect to dependent variables.

The above discussion can be extended to two dimensions and to

functions F that depend on more than one dependent variable in two
or more dimensions. Let F = F(x, y, u, v, ux,vx, uy, vy), where, u =
u(x,y) and v = v(x,y) are dependent variables, and ux = ¶u/ ¶x, uy =
¶u/ ¶y and so on. The first variation of F is given by

18
FINITE ELEMENT METHOD
It can easily be verified that the laws of variation of sums, products,
ratios, powers, and so forth are completely analogous to the
corresponding laws of differentiation. For example, F1 = F1(u) and F2
= F2(u), then

Furthermore, the variational operator can commute with differential

and integral operators (as long as the coordinates x and y are the
fixed coordinates):

19
FINITE ELEMENT METHOD
4.5.2 Fundamental Lemma of Variational Calculus
The fundamental lemma of calculus of variations can be stated as
follows: for any integrable function G(x), if the statement

holds for any arbitrary function h(x) for all x in (a, b), then it follows
that G(x) = 0 in (a, b).

4.5.3 The Euler Equations

As stated earlier, certain problems are formulated as one of seeking
the extremum of functionals (i.e., functions of dependent unknowns
of the problem). For example, problems of solid mechanics can be
formulated as one of minimizing the total potential energy of the
system.
20
FINITE ELEMENT METHOD
Typically, the total potential energy (functional) is written in terms
of the displacement field and applied loads. Then it is useful to
derive the differential equations that govern the displacement field
from this minimum principle.

Consider, for example, the problem of finding a function u = u(x)

such that
………. (20)

and

………. (21)

is a minimum.

21
FINITE ELEMENT METHOD
The necessary condition for I to attain a minimum gives

22
FINITE ELEMENT METHOD
The boundary term vanishes because du is zero at x = a and x = b.
The fact that du is arbitrary inside the interval (a, b) and yet the
equation should hold implies, by the fundamental lemma of the
calculus of variations, that the expression in the square brackets is
zero identically:
………. (22)

Equation (22) is called the Euler equation of the functional in

equation (21). Of all the admissible functions, the one that satisfies
(i.e., the solution of) equation (22) is the true minimizer of the
functional

23
FINITE ELEMENT METHOD
4.5.3 Natural and Essential Boundary Conditions
Consider the problem of minimizing the functional in equation
(7.21) subjected to no end conditions [hence, an element v of the set
of admissible variation is arbitrary even at the end points, i.e., v(a) ¹
0 and v(b) ¹ 0]. Then the functional I(u) has the form

………. (23)

where Qa and Qb are the known values. The necessary conditions for
I to attain a minimum yields

……. (24)

24
FINITE ELEMENT METHOD
Now suppose that ¶F/ ¶u’ and v are selected such that
……. (25)

Then using the fundamental lemma of the calculus of variations, we

obtain the same Euler equation (22).

Equations in (25) are satisfied identically for any of the following

combinations:

………. (26)

25
FINITE ELEMENT METHOD
 The requirement that v = 0 an end point is equivalent to the
requirement that u is specified (to be some value) at that point. The
end conditions in (26) are classified into two types: essential
boundary conditions, which require v (and possibly its derivatives) to
vanish at the boundary, and natural boundary conditions, which
require the specification of the coefficient of v (and possibly its
derivatives).
Problems in which all of the
boundary conditions are of
the essential type are called
Dirichlet boundary value
problems, and those in which
all of the boundary conditions
are of the natural type are
Mixed boundary value problems are called Neumann boundary
those in which both essential and value problems.
natural boundary conditions are
26
specified. FINITE ELEMENT METHOD
EXAMPLE 4.1
Derive Euler’s Equation for a functional I defined by

EXAMPLE 4.2
Find the governing differential equation of a string under
tension T subjected to lateral load q for which potential energy is
given by

27
FINITE ELEMENT METHOD
EXAMPLE 4.3
A 1-Dimensional beam of length L, is subjected to a lateral
distributed load of intensity q per unit length. The governing
differential equation is given by

Derive the variational form of the problem.

28
FINITE ELEMENT METHOD