FINITE ELEMENT METHOD

ME751
Mechanical IV/II
Chapter 3
Direct Stiffness Method: Discrete
Finite Elements (Spring/Bar)
3.1 INTRODUCTION
The primary characteristics of a finite element are embodied in the
element stiffness matrix.

For a structural finite element, the stiffness matrix contains the

geometric and material behavior information that indicates the
resistance of the element to deformation when subjected to loading.
Such deformation may include axial, bending, shear, and torsional
effects.

For finite elements used in nonstructural analyses, such as fluid

flow and heat transfer, the term stiffness matrix is also used, since
the matrix represents the resistance of the element to change when
subjected to external influences.

2
FINITE ELEMENT METHOD
3.2 SPRING AND BAR ELEMENTS
3.2.1 LINEAR SPRING AS FINITE ELEMENT
A linear elastic spring is a mechanical device capable of supporting
axial loading only and constructed such that, over a reasonable
operating range (meaning extension or compression beyond
undeformed length), the elongation or contraction of the spring is
directly proportional to the applied axial load.
The constant of proportionality between deformation and load is
referred to as the spring constant, spring rate, or spring stiffness,
generally denoted as k, and has units of force per unit length.

3
FINITE ELEMENT METHOD
As an elastic spring supports axial
loading only, we select an element
coordinate system (also known as a
local coordinate system) as an x axis
oriented along the length of the
spring, as shown
The ends of the spring are the nodes and the nodal displacements are
denoted by u1 and u2 and are shown in the positive sense.

If these nodal displacements are known, the total elongation or

contraction of the spring is known as is the net force in the spring.

We require that forces be applied to the element only at the nodes

(distributed forces are accommodated for other element types later),
and these are denoted as f1 and f2 and are also shown in the positive
sense. 4
FINITE ELEMENT METHOD
Assuming that both the nodal displacements are zero when the spring
is undeformed, the net spring deformation is given by

The resultant axial force in the spring is

For equilibrium, f1 + f2 = 0 or f1 = − f2, and we can rewrite Equation

(3.2) in terms of the applied nodal forces as

which can be expressed in matrix form as

5
FINITE ELEMENT METHOD
where

is defined as the element stiffness matrix in the element coordinate

system (or local system), {u} is the column matrix (vector) of nodal
displacements, and {f} is the column matrix (vector) of element
nodal forces
In general, the nodal forces are prescribed and the objective is to
solve for the unknown nodal displacements. Formally, the solution is
represented by

where [ke]−1 is the inverse of the element stiffness matrix.

6
FINITE ELEMENT METHOD
However, this inverse matrix does not exist, since the determinant of
the element stiffness matrix is identically zero. Therefore, the
element stiffness matrix is singular, and this also proves to be a
general result in most cases.

The physical significance of the singular nature of the element

stiffness matrix is found by reexamination of spring, which shows
that no displacement constraint whatever has been imposed on
motion of the spring element; that is, the spring is not connected to
any physical object that would prevent or limit motion of either
node. With no constraint, it is not possible to solve for the nodal
displacements individually. Instead, only the difference in nodal
displacements can be determined, as this difference represents the
elongation or contraction of the spring element owing to elastic
effects. This means rigid body motion. 7
FINITE ELEMENT METHOD
For a single, unconstrained element, if arbitrary forces are applied at
each node, the spring not only deforms axially but also undergoes
acceleration according to Newton’s second law. Hence, there exists
not only deformation but overall motion. If, in a connected system of
spring elements, the overall system response is such that nodes 1 and
2 of a particular element displace the same amount, there is no
elastic deformation of the spring and therefore no elastic force in the
spring.

This physical situation must be included in the element formulation.

The capability is indicated mathematically by singularity of the
element stiffness matrix. As the stiffness matrix is formulated on the
basis of deformation of the element, we cannot expect to compute
nodal displacements if there is no deformation of the element.

8
FINITE ELEMENT METHOD
System Assembly in Global Coordinates
Derivation of the element stiffness matrix for a spring element was
based on equilibrium conditions. The same procedure can be applied
to a connected system of spring elements by writing the equilibrium
equation for each node.

However, rather than drawing free-body diagrams of each node and

formally writing the equilibrium equations, the nodal equilibrium
equations can be obtained more efficiently by considering the effect
of each element separately and adding the element force contribution
to each nodal equation.

The process is described as “assembly” , as we take individual

stiffness components and “put them together” to obtain the system
equations.
9
FINITE ELEMENT METHOD
To illustrate the assembly of element characteristics into global (or
system) equations, consider the system of two linear spring
elements.

10
FINITE ELEMENT METHOD
Assuming the system of two spring elements to be in equilibrium, we
examine free-body diagrams of the springs individually and express
the equilibrium conditions for each spring as

To begin “assembling” the equilibrium equations describing the

behavior of the system of two springs, the displacement compatibility
conditions, which relate element displacements to system
displacements, are written as

The compatibility conditions state the physical fact that the springs
are connected at node 2, remain connected at node 2 after
deformation, and hence, must have the same nodal displacement at
node 2. FINITE ELEMENT METHOD
11
Substituting Equations (3.9) into Equations (3.8), we obtain

These equations are not yet amenable to direct combination, as the

displacement vectors are not the same. We expand both matrix
equations to 3 × 3 as follows (while formally expressing the facts
that element 1 is not connected to node 3 and element 2 is not
connected to node 1):

12
FINITE ELEMENT METHOD
The addition of Equations (3.11) and (3.12) yields

Next, we refer to the free-body diagrams of each of the three nodes.

The equilibrium conditions for nodes 1, 2, and 3 show that

Substituting into Equation (3.13), we obtain the final result:

13
FINITE ELEMENT METHOD
which is of the form [K]{U} = {F}, similar to Equation (3.5).
However, Equation (3.15) represents the equations governing the
system composed of two connected spring elements. By direct
consideration of the equilibrium conditions, we obtain the system
stiffness matrix or global stiffness matrix [K] as

Note that the system stiffness matrix is

• symmetric, as is the case with all linear systems referred to
orthogonal coordinate systems;
• singular, since no constraints are applied to prevent rigid body
motion of the system; and
• the system matrix is simply a superposition of the individual
element stiffness matrices with proper assignment of element
nodal displacements and associated stiffness coefficients to
system nodal displacements. 14
FINITE ELEMENT METHOD
In order for the global system of equations to have a unique solution,
the determinant of the global system matrix must be nonzero.
However, an examination of the global system matrix reveals that
one of its eigenvalues is zero, thus resulting in a zero determinant or
singular matrix. Therefore, the solution is not unique.
The eigenvector corresponding to the zero eigenvalue represents the
translational mode, and the remaining nonzero eigenvalues
represent all of the deformation modes.
Application of Boundary Conditions and Solution
Having formed the assembled equation (3.15) for the system we can
proceed to the solution phase. In most of the cases the solution
“blows up” because the coefficient matrix (the global stiffness
matrix) is singular. The mathematical interpretation of this behavior
is that rows and columns of [K] are linear combinations of each
other [Rank of Stiffness Matrix]. The physical interpretation of
singularity is that there are unsuppressed rigid body motions. 15
FINITE ELEMENT METHOD
To eliminate rigid body motions and render the system nonsingular
we must apply the physical support conditions as displacement
boundary conditions. For the above example if the node 1 is fixed
and P force is applied towards right at node 3, the corresponding
boundary conditions should be

Applying these boundary conditions, Equation (3.15) reduces to

This equation can be solved for unknown displacements U2 and U3.

Substituting U2 and U3 into Equation (3.15), we can determine the
unknown reaction F1.

16
FINITE ELEMENT METHOD
EXAMPLE 3.1
Given: For the spring system shown in Figure E3.1 ,
k1 = 100 N/mm, k2 = 200 N/mm, k3 = 100 N/mm P = 500 N
Find: (a) the global stiffness matrix
(b) displacements of nodes 2 and 3
(c) the reaction forces at nodes 1 and 4.

Figure E3.1

17
FINITE ELEMENT METHOD
EXAMPLE 3.2
Three rigid bodies, 2, 3, and 4, are connected by four springs, as
shown in the Figure E3.2. A horizontal force of 1,000 N is applied
on Body 4. Find the displacements of the three bodies and the
forces (tensile/compressive) in the springs. What is the reaction at
the wall? Assume the bodies can undergo only translation in the
horizontal direction. The spring constants (N/mm) are k1 = 400,
k2 = 500, k3 = 500, k4 = 300.

Figure E3.2

18
FINITE ELEMENT METHOD
3.2 ELASTIC BAR AS FINITE ELEMENT
While the linear elastic spring serves to introduce the concept of the
stiffness matrix, the usefulness of such an element in finite element
analysis is rather limited. Certainly, springs are used in machinery in
many cases and the availability of a finite element representation of
a linear spring is quite useful in such cases. The spring element is
also often used to represent the elastic nature of supports for more
complicated systems.

A more generally applicable, yet similar, element is an elastic bar

subjected to axial forces only. This element, which we simply call a
bar element, is particularly useful in the analysis of both two- and
three dimensional frame or truss structures.

19
FINITE ELEMENT METHOD
Consider an elastic bar of length L to which is affixed a uniaxial
coordinate system x with its origin arbitrarily placed at the left end.
This is the element coordinate system or reference frame.

Denoting axial displacement at any position along the length of the

bar as u(x), we define nodes 1 and 2 at each end as shown and
introduce the nodal displacements u1 = u(x = 0) and u2 = u(x = L).

Recall from elementary strength of materials that the deflection d of

an elastic bar of length L and uniform cross-sectional area A when
subjected to axial load P is given by

20
FINITE ELEMENT METHOD
Using Equation (3.17), we obtain the equivalent spring constant of an
elastic bar as

and could, by analogy with the linear elastic spring, immediately

write the stiffness matrix as Equation (3.6).
To obtain the necessary equilibrium equations relating the
displacements to applied forces, we proceed from displacement to
strain, strain to stress, and stress to loading, as follows. In uniaxial
loading, as in the bar element, we need consider only the normal
strain component, defined as

The axial stress, by Hooke’s law, is then

21
FINITE ELEMENT METHOD
The associated axial force is then given by

Equation (3.21) is now used to relate the applied nodal forces f1 and
f2 to the nodal displacements u1 and u2 as

Equations (3.22) and (3.23) are expressed in matrix form as

Comparison of Equation (3.24) to Equation (3.4) shows that the

stiffness matrix for the bar element is given by

22
FINITE ELEMENT METHOD
EXAMPLE 3.3
For the bar assemblages shown in Figure E3.3, determine the
nodal displacements, the stresses in each element and the
reactions. Use the direct stiffness method for these problems.

Figure E3.3

23
FINITE ELEMENT METHOD