Ch.

2 Stiffness Method

Chapter 2
Stiffness Method

MS4011 – FEM 1
Ch. 2 Stiffness Method

2.1. Definition of the Stiffness Matrix

Stiffness matrix, kˆ is a matrix such that
fˆ  kˆ dˆ
where kˆ relates local - coordinate ( xˆ, yˆ , zˆ ) nodal displaceme nt dˆ to
local force fˆ of a single element.
The bold notation denotes a matrix or a vector, and the ˆ symbol
denotes quantities referred to a local - coordinate system, set up
to be convenient for the element.
For a continuous medium or structure with n element, a stiffness
matrix K relates global - coordinate ( x, y, z ) nodal displaceme nt
d to global force F of the whole medium or structure.
MS4011 – FEM 2
Ch. 2 Stiffness Method

2.2. Stiffness Matrix for a Spring Element

 Consider a linear spring element shown in Figure 2-1. References point 1 and
2 (nodes) are located at the ends of the element.
 We want to develop a relationship between nodal force and nodal
displacement for a spring element. This relationship will be the stiffness
matrix as follows.
 fˆ1x   k11 k12   dˆ1x 
ˆ   ˆ  (2.2.1)
 f 2 x  k 21 k 22  d 2 x 
where the element kij of the kˆ matrix in Eq. (2.1) are to be determined .
k
1 2 x̂
fˆ1x , dˆ1x fˆ2 x , dˆ2 x
L
Figure 2-1 Linear spring element with positive nodal displacement and force conventions
MS4011 – FEM 3
Ch. 2 Stiffness Method

 Step 1 Select Element Type

Consider a linear spring subjected to resulting nodal forces T directed a long
the spring axial direction x̂ as shown in Figure 2-2, so as to be in
equilibrium.
The spring is represented by labeling nodes at each end. The original distance
between nodes before deformation is denoted by L.

k
T 1 2 T x̂

dˆ1x dˆ2 x
L
k
1 2

Figure 2-2 Linear spring subjected to tensile forces

MS4011 – FEM 4
Ch. 2 Stiffness Method

 Step 2 Select a Displacement Functions

A displacement function û is assumed. Here a linear displacement
varation along the axis of the spring is assumed because a linear
function with specified endpoints has a unique path. Therefore

uˆ  a1  a2 xˆ (2.2.2)

In general, the total number of coefficients a is equal to the total

number of degrees of freedom (dof) associated with the element. Here
the total number of dof is two – an axial displacement at each nodes of
the element. In matrix form Eq. (2.2.2) becomes

 a1 
uˆ  1 xˆ    (2.2.3)
a 2 
MS4011 – FEM 5
Ch. 2 Stiffness Method

We now want to express û as a function of the nodal displacement. We

achieve this by evaluating û at each node and solving for a1 and a2
from Eq. (2.2.2) as follows
uˆ (0)  dˆ  a 1x 1 (2.2.4)
uˆ ( L)  dˆ2 x  a2 L  dˆ1x (2.2.5)
or, solving Eq. (2.5) for a2 ,
dˆ2 x  dˆ1x
a2  (2.2.6)
L
On substituting Eqs. (2.2.4) and (2.2.6) into Eq. (2.2.2), we have

 dˆ2 x  dˆ1x 
uˆ    xˆ  dˆ1x (2.2.7)
L 
 

MS4011 – FEM 6
Ch. 2 Stiffness Method

In matrix form, we express Eq. (2.2.7) as

 xˆ xˆ   dˆ1x 
uˆ  1    (2.2.8)
 L L  dˆ2 x 
 dˆ1x 
or uˆ   N1 N 2    (2.2.9)
dˆ2 x 
xˆ xˆ
Here N1  1  and N 2  (2.2.10)
L L

are called the shape function because the Ni’s express the shape of the
assumed displacement function over the domain of the element when
the ith element degree of freedom has unit value and all other dof are
zero.

MS4011 – FEM 7
Ch. 2 Stiffness Method

In the case, N1 dan N2 are linear functions that have the properties that
N1 = 1 at node 1 and N1 = 0 at node 2, whereas N2 = 1 at node 2 and
N2 = 0 at node 1. Also, N1 + N2 = 1 for any axial coordinate along the
bar.

In addition, the Ni’s are often called interpolation function because

we are interpolating to find the value of a function between given
nodal value. The interpolation function may be different from the
actual function except at the endpoints or nodes where the
interpolation function and actual function must be equal to specified
nodal values.

MS4011 – FEM 8
Ch. 2 Stiffness Method

 Step 3 Define the Strain-displacement and Stress-strain

Relationships

The tensile forces T produce a total elongation (deformation) of the

spring. For the linear spring, T and  are related through Hooke’s law
by.
T  k (2.2.11)
where, because  is the deformatio n of the spring, we have
  uˆ ( L)  uˆ (0) (2.2.12)
Making use of Eq. (2.2.7), Eq. (2.2.11) becomes
  dˆ  dˆ
2x 1x ( 2.2.13)

MS4011 – FEM 9
Ch. 2 Stiffness Method

 Step 4 Derive the Element Matrix and Equations

By the sign convention for nodal forces, we have

fˆ  T and fˆ  T
1x 2x (2.2.14)
Using Eqs. (2.2.11), (2.2.13), and (2.2.14), we have
T   fˆ1x  k (dˆ2 x  dˆ1x )
(2.2.15)
T  fˆ2 x  k (dˆ2 x  dˆ1x )
or rewriting Eq. (2.2.15), we obtain
fˆ1x  k (dˆ1x  dˆ2 x )
(2.2.16)
fˆ  k (dˆ  dˆ )
2x 2x 1x

MS4011 – FEM 10
Ch. 2 Stiffness Method

Now expressing Eq. (2.2.16) in a single matrix equations yields

 fˆ1x   k  k   dˆ1x 
ˆ   ˆ  (2.2.17)
 f 2 x   k k  d 2 x 
This relationship holds for the spring along the xˆ axis. For our basic
definition of a stiffness matrix and use of Eq. (2.2.1) applied to Eq. (2.2.17),
we obtain

ˆk   k  k
 k (2.2.18)
 k 
as the stiffness matrix for a linear spring element. We observe that kˆ is
a symmetric square matrix.

MS4011 – FEM 11
Ch. 2 Stiffness Method

 Step 5 Assemble the Element Equations to Obtain the Total/

Global Equations and Introduce Boundary Equations
The total stiffness matrix and total force vector are assembled using
nodal force equilibrium equations, force/deformation and
compatibility equations from Section 2.2, and the direct stiffness
method described in Section 2.4. This step applies for structures
composed of more than one element such that
N N
K  K  k (e)
and F   F    f (e) (2.2.19)
e 1 e 1
where k and f are now element stiffness and force matrices expressed
in a global reference frame.
The sign  used in this context means that all element matrices must
be assembled properly according to the direct stiffness method
described in Section 2.4.
MS4011 – FEM 12
Ch. 2 Stiffness Method

 Step 6 Solve for the Nodal Displacements

The displacements are then determined by imposing boundary
conditions and solving a system of equations, F = K d, simultaneously.

 Step 7 Solve for the Element Forces

The element forces are determined by back-substitution, applied to
each element, into equation similar to Eq. (2.2.16).

 Step 8 Interpret the results

Finally, the element forces resulted can be analysed to check if all
element forces are below its allowable values.

MS4011 – FEM 13
Ch. 2 Stiffness Method

2.3. Example of a Spring Assemblage

We will consider the specific example of the two-spring assemblage
shown in Figure 2-3. Here we fix node 1, and apply axial forces at node
2 and node 3. The stiffness of spring elements 1 and 2 are k1 and k2
respectively. The x axis is the global axis of the assemblage. In this
case, the local axis of each element coincides with the global axis of the
assemblage.

1 2
1 3 2
x
F3 x F2 x
k1 k2
Figure 2-3 Two-spring assemblage

MS4011 – FEM 14
Ch. 2 Stiffness Method

Using Eq. (2.17) the element stiffness matrix for each element cam be
expressed
for element 1
 f1x   k1  k1   d1x 
     (2.3.1)
 f 3 x   k1 k1  d 3 x 
and for element 2,
 f3 x   k2  k 2  d 3 x 
     (2.3.2)
 f 2 x   k 2 k 2  d 2 x 

Furthermore, element 1 and 2 must remain connected at common node

3 throught the displacement. This is called the continuity or
compatibility requirement. The compatibility requirement yields
(superscript refers to the element number.)
d 3(1x)  d 3( 2x )  d 3 x (2.3.3)
MS4011 – FEM 15
Ch. 2 Stiffness Method

Based on the sign convention for element nodal forces given in Figure 2-1,
we can write nodal equilibrium equations at node 3, 2, and 1 as
F3 x  f 3(x1)  f 3(x2 ) (2.3.4)
F2 x  f 2(x2 ) (2.3.5)
F1x  f1(x1) (2.3.6)
Where F1x results from the reaction at the fixed support. To further
clarify the resulting Eqs. (2.3.4) – (2.3.6), free-body diagrams of each
element and node (using the established sign conventions for element
nodal forces) are shown in Figure 2-4.
f1x(1) f 3x(1) f 3x( 2 ) 2
f 2(x2 )
1 1 3 2

F1x F3 x F2 x
Figure 2-4 Nodal forces consistent with element forces sign convention
MS4011 – FEM 16
Ch. 2 Stiffness Method

Application of Newton’s third law for each node and element gives
F3 x  k1d1x  k1d 3 x  k 2 d 3 x  k 2 d 2 x
F2 x  k 2 d 3 x  k 2 d 2 x (2.3.7)
F1x  k1d1x  k1d 3 x
In matrix form, Eqs. (2.3.7) are expressed by
 F3 x  k1  k 2  k2  k1   d 3 x 
    
 F2 x     k 2 k2 0  d 2 x  (2.3.8)
F    k 0 k1   d1x 
 1x   1

Or rearranging, Eqs. (2.3.8) in numerically increasing order of the nodal

dof, we have
 F1x   k1 0  k1   d1x 
    
 F2 x    0 k2  k 2  d 2 x  (2.3.9)
 F   k  k2 k1  k 2  d 3 x 
 3x   1
MS4011 – FEM 17
Ch. 2 Stiffness Method

Eq. (2.3.9) is now written as the single matrix equation

F  Kd ( 2.3.10)
 F1x   d1x 
   
where  F2 x  is called the global nodal force vector , d 2 x  is called
F  d 
 3x   3x 
the global nodal displaceme nt vector , and
 k1 0  k1 
K   0 k2  k 2  (2.3.11)
 k1  k 2 k1  k 2 
is called the total or global stiffness matrix .

MS4011 – FEM 18
Ch. 2 Stiffness Method

2.4.Assembling the Total Stiffness Matrix by

Superposition (Direct Stiffness Method)
This method is based on “proper superposition” of the individual
element matrices making up a structure.
Referring to the-spring assemblage of Section 2.3, the element stiffness
matrices are given in Eqs. (2.3.1) and (2.3.2) as
d1x d3x d3x d2x
 k1  k1   k2  k2 
k  (1)
 k 
( 2)
 (2.4.1)
 k1 k1   k 2 k 2 
Here the d ix ' s written above the column in the k ' s indicate the degrees
of freedom associated with each element.

MS4011 – FEM 19
Ch. 2 Stiffness Method

To superpose the element matrices, all of them must be expended to the

order of the total structure (spring assemblage) stiffness matrix so that
each element matrix is associated with all the dof of the structure.
To expand each element stiffness matrix to the order of total stiffness
matrix, we simply add rows and columns of zero for the those
displacement not associated with that particular element.
The expanded form of each element equation can be expressed as

for element 1
 1 0  1 d1x   f1x 
(1) (1)

   (1)   (1) 
k1  0 0 0  d 2 x    f 2 x  (2.4.2)
 1 0 1  d 3(1x)   f 3(x1) 

MS4011 – FEM 20
Ch. 2 Stiffness Method
for element 2
0 0 0  d1x   f1x 
( 2) (2)

   
k 2 0 1  1 d 2( 2x)    f 2(x2 )  (2.4.3)
0  1 1  d 3( 2x )   f 3(x2 ) 
Now considering force equilibrium at each node results in
 f1(x1)   0   F1x 
   ( 2)   
 0    f 2 x    F2 x  (2.4.4)
 f (1)   f ( 2)   F 
 3x   3x   3x 
Using Eqs. (2.4.2) and (2.4.3) in Eq. (2.4.4), we obtain
 1 0  1 d1x  0 0 0  d1x   F1x 
(1) ( 2)

     
k1  0 0 0  d 2(1x)   k 2 0 1  1 d 2( 2x)    F2 x  (2.4.5)
 1 0 1  d 3(1x)  0  1 1  d 3( 2x )   F3 x 
MS4011 – FEM 21
Ch. 2 Stiffness Method
Simplifyin g Eq. (2.4.5) results in
 k1 0  k1   d1x   F1x 
 0     
 k2  k 2  d 2 x    F2 x  (2.4.6)
 k1  k 2 k1  k 2  d 3 x   F3 x 
Here the superscrip ts indicating the element numbers associated with
the nodal displaceme nts have been dropped because d1(x1) is really d1x ,
d 2( 2x) is really d 2 x , and, by Eq. (2.3.3) d 3(1x)  d 3( 2x )  d 3 x , the node 3
displaceme nt of the total assemblage . Equation (2.4.6), obtained
through superposit ion, is identical to Eq. (2.3.9).
The method of directly assembling individual element stiffness matrices
to form the total structure stiffness matrix and the the total set of
stiffness equations is called direct stiffness method. It is the most
important step in the FEM.
MS4011 – FEM 22
Ch. 2 Stiffness Method

2.5.Boundary Conditions
We must specify boundary (or support) conditions for structure models
such as the spring assemblage of Figure 2-4, or K will be singular; that
is the determinant of K will be zero and, therefore, its inverse will not
exist. Without specifying adequate kinematic constraints or support
conditions, the structure will be free to move as a rigid body.
Boundary conditions (BC) are of two general type:
1. Homogeneous BC – the most common – occur at locations that are
completely prevented from movement,
2. Non- homogeneous BC occur where finite non-zero values of
displacement are specified, such as the settlement of a support.

MS4011 – FEM 23
Ch. 2 Stiffness Method
In general, specified support conditions are treated mathematically by
partitioning the global equilibrium equations as follows:
Simplifyin g Eq. (2.4.5) results in
 K 11 K 12  d1   F1 
K     (2.5.1)
 21 K 22  d 2  F2 
where we let d1 be the unconstrai ned or free displaceme nt and d 2 be the
the specified displaceme nts nodes. From. (2.5.1), we have
K 11d1  F1  K 12d 2 (2.5.2)
and F2  K 21d1  K 22d 2 (2.5.3)
where F1 are the known nodal forces and F2 are the unknown nodal forces
at the specified displaceme nt nodes. F2 is found from Eq. (2.5.3) after de -
termining d1 from Eq. (2.5.2). In the Eq. (2.5.2), we assume that K 11 is no
longer singular, thus allowing for the determinat ion of d1.
MS4011 – FEM 24
Ch. 2 Stiffness Method

Homogenoue BC
To illustrate the two general types of BC, let us consider Eq. (2.4.6),
derived for the spring assemblage of Figure 2-4. We will first consider
the case of homogenous BC. Hence, all BC are such that the
displacements are zero at certain nodes.Here we have d1x = 0 because
node 1 is fixed. Therefore, Eq. (2.4.6) can be written as.
 k1 0  k1   0   F1x 
 0     
 k2  k 2  d 2 x    F2 x  ( 2 . 5. 4)
 k1  k 2 k1  k 2   d  F 
   3x 
3 x

Equation (2.5.4), written in expanded from becomes

k1 (0)  (0)d 2 x  k1d 3 x  F1x
0(0)  k 2 d 2 x  k 2 d 3 x  F2 x (2.5.5)
 k1 (0)  k 2 d 2 x  (k1  k 2 )d 3 x  F3 x
MS4011 – FEM 25
Ch. 2 Stiffness Method

Consider t he second and third of Eqs. (2.5.5), written in matrix form,

we have
 k2  k 2  d 2 x   F2 x 
 k k  k  d    F  (2.5.6)
 2 1 2  3x   3x 

We have now effectively partitioned off the first column and row of K
and the first row of d and F to arrive at Eq. (2.5.6).
For homogenous BC, Eq. (2.5.6) could have been obtained directly by
deleting the row and column of Eq. (2.5.4) corresponding to the zero-
displacement degree of freedom. Here row 1 and column 1 are deleted
because d1x = 0. However, F1x is not necessary zero and must be
determined as follows.

MS4011 – FEM 26
Ch. 2 Stiffness Method

After solving Eq. (2.5.6)for d 2 x and d 3 x , we have

1 1 1

 k 2   F2 x  k 2 k1 k1   F2 x 

1
d 2 x   k 2
     1 1
  ( 2.5.7 )
 d 3 x   k 2 k1  k 2   F3 x     F3 x 
 k1 k1 
Using Eq. (2.5.7)in the first of Eqs. (2.5.5),we obtain thereaction F1x as
F1x   k1d 3 x ( 2.5.8 )
We can express the unknownnodalforce at node1 (the reaction force)in
terms of applied nodalforces F2 x and F3 x by using Eq. (2.5.7)for d 3 x sub -
stituted into Eq. (2.5.8).The results is
F1x   F2 x  F3 x ( 2.5.9 )

MS4011 – FEM 27
Ch. 2 Stiffness Method

Note for homogenoue BC.

For all homogenous BC, we can delete the rows and culumns
corresponding to the zero-displacement dof from the original set of
equatons and then solve for the unknown displacements.
This procedure is useful for hand calculations. (More practical,
computer-assisted scheme for solving the system of simultaneous
equations will be discussed in another chapter.)
Students are encouraged to review the numerical method to solve the
system of simultaneous equations, i.e. Gauss or Gauss-Seidel method or
other.

MS4011 – FEM 28
Ch. 2 Stiffness Method

Non-homogeneous BC
Next, we consider the case of non-homogenous BC. Hence, some of the
specified displacements are non-zero. For simplicity’s sake, let d1x = ,
where  is a known displacement, in the Eq. (2.4.6). We now have

 k1 0  k1      F1x 
 0     
 k2  k 2  d 2 x    F2 x  ( 2.5.10)
 k1  k 2 k1  k 2  d 3 x   F3 x 
Equation (2.5.10) written in expanded form becomes
k1  0d 2 x  k1d 3 x  F1x
0  k 2 d 2 x  k 2 d 3 x  F2 x (2.5.11)
 k1  k 2 d 2 x  (k1  k 2 )d 3 x  F3 x

MS4011 – FEM 29
Ch. 2 Stiffness Method

Considerin g the second and third of Eqs. (2.5.11) because they have
known right - side nodal forces F2 x and F3 x
0  k 2 d 2 x  k 2 d 3 x  F2 x
(2.5.12)
 k1  k 2 d 2 x  (k1  k 2 )d 3 x  F3 x
Transformi ng the known  to the right side of Eqs. (2.5.12) results in
k 2 d 2 x  k 2 d 3 x  F2 x
(2.5.13)
 k 2 d 2 x  (k1  k 2 )d 3 x  k1  F3 x
Rewritting Eqs. (2.5.13) in matrix form, we have
 k2  k 2  d 2 x   F2 x 
 k     (2.5.14)
 2 k1  k 2  d 3 x  k1  F3 x 

MS4011 – FEM 30
Ch. 2 Stiffness Method

Note for homogenoue BC.

When dealing with non-homogeneous BC, one cannot initially delete
row 1 and column 1 of Eq. (2.5.1)), corresponding to the non-
homogeneous BC, as indicated by the resulting Eq. (2.5.14). Had we
done so, the k1 term in Eq. (2.5.14) would have been neglected,
resulting in an error in the solution for the displacement.
For non-homogeneous BC, we must, in general, transform the terms
associated with the known displacement to the right-side force matrix
before solving for the unknown nodal displacements. This was illustrated
by transforming the k1 term of the second row of Eqs. (2.5.12) to the
right-side of the second row of Eqs. (2.5.13).
Finally, we could solve for the displacement in Eq. (2.5.14) in a manner
similar to that used to solve Eq. (2.5.6).
MS4011 – FEM 31
Ch. 2 Stiffness Method

Some properties of the stiffness matrix (that are also applicable to the
generalization of the finite element method).
1. K is symetric, as is each of the element stiffness matrices.
2. K is singular and thus no inverse exists until sufficient BC are
imposed.
3. The main diagonal terms of K are always positive. Otherwise, a
positive nodal force Fi could produce a negative displacement di – a
behavior contrary to the physical behavior of any actual structure.

Example 2.1

MS4011 – FEM 32
Ch. 2 Stiffness Method

2.6.Potential Energy Approach

One of the alternative methods often used to derive the element
equations and the stiffness matrix for an element is based on the
principle of minimum potential energy (POMPE).
Adantages of the POMPE:
1. More general than previous method (equilibrium equations of nodal and
element).
2. More adaptable for the determination of element equations for
complicated elements (those with large numbers of d.o.f) such as plane
stress/strain element, the axisymetric stress element, the bending plate
element, and the 3D solid stress element).
3. Only applicable for elastic materials (principle of virtual work is
applicable for any material behavior).
4. Has lack of physical insight.

MS4011 – FEM 33
Ch. 2 Stiffness Method

 The total potential energy (TPE), p of a structure is expressed in

term of displacements. In the FE formulation, these will generally be
nodal displacements such that p = p (d1, d2, d3, ..., dn).
 Whenp is minimized with respect to these displacements,
equilibrium equations result. For the spring element, we will show
that the same nodal equilibrium equations result as previously
derived in Section 2.2.
 The POMPE is stated as follows:
Of all the displacements that satisfy the given boundary conditions
of a structure, those that safisfy the equations of equilibrium are
distinguishable by the stationary value of the potential energy. If the
stationary value is a minimum, the equilibrium state is stable.
To explain this principle, we must first explain the concepts of PE
and of a stationary value of a function.
MS4011 – FEM 34
Ch. 2 Stiffness Method

 Total potential energy (PTE) is defined as the sum of the internal

strain energy U and the potential energy of the external force;
thus is

Strain energy PE of external force

 p U  (2.6.1)

Strain energy is the capacity of internal forces (or stresses) to do

work through deformation (strains) in the structure.
Potential energy (PE) of the external force, is the capacity of
external forces such body force, surface traction forces, and applied
nodal forces to do work through deformation of the structure.

MS4011 – FEM 35
Ch. 2 Stiffness Method
 Recall that a linear spring has force related to deformation by F = kx,
where k is the spring constant and x is the deformation of the spring
(Figure 2-5).
F
k
k F
x
x
Figure 2-5 Force-deformation curve for linear spring

The differential internal work (or strain energy) dU in the spring is

the internal force multiplied by the change in displacement through
which the force moves, given by.
dU  F dx (2.6.2)
MS4011 – FEM 36
Ch. 2 Stiffness Method
Now we express F as
F  kx (2.6.3)
Using Eq. (2.6.3) in Eq. (2.6.2), the differenti al strain energy becomes
dU  kx dx (2.6.4)
The total strain energy is then given by
x
U   kx dx (2.6.5)
0
Upon explicit integratio n of Eq. (2.6.5), we obtain
U  12 kx 2 (2.6.6)
Using Eq. (2.6.3) in Eq. (2.6.6), we have
U  12 (kx) x  12 Fx (2.6.7)
Equation (2.6.7) indicates that strain energy is the area under
force - deformatio n curve.
MS4011 – FEM 37
Ch. 2 Stiffness Method
 The PE of external force, being opposite in sign from the external
work expression because the PE of external force is lost when the
work is done by the external force, is given by

   Fx (2.6.8)
Therefore substituti ng Eqs. (2.6.6) and (2.6.8) into (2.6.1),
the TPE becomes
 p  12 kx 2  Fx (2.6.9)

 To apply the POMPE – that is, to minimize ofp – we take the

variation of p defined in general as

 p  p  p
 p  d1  d 2    d n (2.6.10)
d1 d 2 d n

MS4011 – FEM 38
Ch. 2 Stiffness Method

The principles states that equilibrium exists when the di define a

structure state such that p = 0 for arbitrary admissible variations
di from the equilibrium state. An admissible variation is one in
which the displacement field still satisfies the BC and inter-element
continuity.
To satisfy  p  0 (for d i  0), all coefficien ts associated with
the d i must be zero independen tly. Thus
 p  p
 0 (i  1,2,3,  , n) or 0 (2.6.11)
d i  d 
where n equations must be solved for the n values of di that define
static equilibriu m state of the structure.

MS4011 – FEM 39
Ch. 2 Stiffness Method

Equation (2.6.11) shows that for our purposes throughout this text,
we can interpret the variation of p (p) as a compact notation
equivalent to differentiation of p with respect to the unknown nodal
displacements for which p is expressed.
 We now derive the spring element equations and stiffness matrix
using the POMPE by analyzing a linear single-dof spring as shown
in Figure 2-6.
k
1 2
fˆ1x fˆ2 x
L

Figure 2-6 Linear spring subjected to nodal forces

MS4011 – FEM 40
Ch. 2 Stiffness Method
Using Eq. (2.6.9), the TPE becomes
ˆ 1
2
ˆ 
2

 p  k d 2 x  d1x  fˆ1x dˆ1x  fˆ2 x dˆ2 x (2.6.12)
where dˆ  dˆ is the deformatio n of the spring in Eq. (2.6.9).
2x 1x

Simplifyin g Eq. (2.6.12), we obtain

 
 p  12 k dˆ22x  2dˆ2 x dˆ1x  dˆ12x  fˆ1x dˆ1x  fˆ2 x dˆ2 x (2.6.13)
Minimizati on of  p with respect to each nodal displaceme nt
requires that
 p 1
dˆ

 2 k  2dˆ2 x  2dˆ1x  fˆ1x  0
1x
(2.6.14)
 p 1
dˆ
 
 2 k 2dˆ2 x  2dˆ1x  fˆ2 x  0
2x
MS4011 – FEM 41
Ch. 2 Stiffness Method
Simplifyin g Eq. (2.6.14), we have
k (dˆ  dˆ )  fˆ
2x 1x 1x
(2.6.15)
k (dˆ2 x  dˆ1x )  fˆ2 x
In matrix form, we express Eq. (2.6.15) as
 k  k   dˆ1x   fˆ1x 
 k k   ˆ    ˆ  (2.6.16)
  d 2 x   f 2 x 
Since fˆ  kˆ dˆ , we have the stiffness matrix for the spring
element obtained from Eq. (2.6.16) :
ˆk   k  k  (2.6.17)
 k k 
 
As expected, Eq. (2.6.17) is identical to the stiffness matrix
obtained in Section 2.2, Eq. (2.2.18).
MS4011 – FEM 42
Ch. 2 Stiffness Method

 We have developed the FE spring element equations by minimizing

the TPE with respect to the nodal displacements. Now, we show that
the TPE of an entire structure (here an assemblage of spring
elements) can be minimized with respect to each nodal dof and this
minimization results in the same FE equations used for the solution
as those obtained by the direct stiffness method explained in Section
2.4.

Example 2.5

MS4011 – FEM 43
Ch. 2 Stiffness Method

Reference:
1. Logan, D.L., 1992, A First Course in the Finite Element
Method, PWS-KENT Publishing Co., Boston.
2. Imbert, J.F.,1984, Analyse des Structures par
Elements Finis, 2nd Ed., Cepadues.
3. Zienkiewics, O.C., 1977, The Finite Eelement Method,
3rd ed., McGraw-Hill, London.

MS4011 – FEM 44
Ch. 2 Stiffness Method

Example
Consider the spring system shown in the figure. The force at nodes
2 and 3 are known. Node 1 is fixed at the support. Find the nodal
displacements and element internal forces, given that
k1 = 1000 lb/in, k2 = 2000 lb/in
F2x = 1000 lb, F3x = 1000 lb

1 2
1 3 2
x
F3 x F2 x
k1 k2

MS4011 – FEM 45
Ch. 2 Stiffness Method

Exercise
Consider the spring system shown in the figure. The force at nodes
3 is known. Nodes 1 and 4 are fixed at the support. Find the nodal
displacements and element internal forces, given that
k1 = k2 = k3 = 20 N/mm
F3x = 1000 N

1 2 3
1 2 3 4

k1 k 2k F3 x k3
2

MS4011 – FEM 46
Ch. 2 Stiffness Method

Exercise

2
3
1
1 2
k2
P
3
k1 4

k3

MS4011 – FEM 47