Introduction to the

Stiffness (Displacement) Method

Finite Element Method

By: Dr. Aamir Mubashar
Objectives

 Define the stiffness matrix

 Derive the stiffness matrix for a spring element

 Demonstrate assembly of stiffness matrix into global stiffness

matrix

 Describe and apply different kind of boundary conditions

 Show how penalty method can be used

Stiffness Matrix
 For an element, a stiffness matrix 𝑘 is a matrix such that

𝑓 = 𝑘 𝑑
where 𝑘 relates the nodal displacements 𝑑 to nodal forces 𝑓
of a single element

 For a continuous medium or a structure comprising a series of

elements, stiffness matrix 𝐾 relates nodal displacements 𝑑 in
global coordinates (𝑥, 𝑦, 𝑧) to global forces 𝐹

𝐹 = 𝐾 𝑑
where 𝐾 represents the stiffness matrix of whole assembly
Stiffness Matrix for a Spring Element

 Using direct equilibrium approach, stiffness matrix for a one

dimensional linear spring will be derived
 Spring obeys Hooke’s law
 Resist forces only in the direction of spring

 Nodes 1 and 2 are at the ends of the spring

 Local nodal forces are 𝑓1𝑥 , 𝑓2𝑥 and nodal displacements are 𝑢1 , 𝑢2
 The local axis acts in direction of spring so displacements and
forces can be directly measured
Stiffness Matrix for a Spring Element

 The nodal displacements are called degrees of freedom (dof)

 Positive directions for forces and displacements are taken
along the positive x-axis
 𝑘 is called spring constant or stiffness of the spring

 We need a relationship between nodal forces and nodal

displacements
 Given by stiffness matrix
Stiffness Matrix for a Spring Element

 The stiffness matrix for spring element is of the form

𝑓1𝑥 𝑘 𝑘12 𝑢1
= 11 𝑢2
𝑓2𝑥 𝑘21 𝑘22

 To use this stiffness matrix, element stiffness coefficients 𝑘𝑖𝑗

need to be determined
 𝑘𝑖𝑗 represents force 𝐹𝑖 in the ith degree of freedom due to a
unit displacement 𝑑𝑗 in the jth degree of freedom while all
other displacements are zero
 Now we will use the general steps outlined earlier to derive the
stiffness matrix for the spring element
General Steps of FEM

Step 1 Step 6 Step 7

•Discretise and •Solve for unknown •Solve for element
select the element degree of freedom strains and stresses
types (displacements)

Step 2 Step 5 Step 8

•Select a • Assemble to obtain •Interpret the results
global equations and
displacement apply boundary
function conditions

Step 3 Step 4
•Define the strain / •Derive the element
displacement and stiffness matrix and
stress / strain equations
relationships
Step 1: Element Selection

 We will use a linear spring element which may be part of a

system of springs
 Nodal tensile forces 𝑇 act on spring which may result from
action of adjacent springs
 Original distance between nodes before deformation is 𝐿
Step 2: Select a Displacement Function

 The mathematical function to represent the deformed shape

of the spring element under loading needs to be selected
 The distribution of displacement within the element is assumed
to follow an appropriate mathematical function
 Most common functions used are polynomials
 Spring element resists axial loading only i.e. displacements
degrees of freedom 𝑢1 , 𝑢2 along the 𝑥 direction
 We assume linear displacement variation along the x-axis of
spring, because a linear function with specified end points has
a unique path

𝑢 = 𝑎1 + 𝑎2 𝑥
Step 2: Select a Displacement Function

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

number of degrees of freedom associated with the element
 For spring element, total number of degrees of freedom is two
 In matrix form
𝑎1
𝑢= 1 𝑥 𝑎
2

 We now express 𝑢 in terms of nodal displacements 𝑢1 , 𝑢2 as this

will allow the application of physical boundary conditions
directly on nodal displacements
 Thus evaluating 𝑢 at each node and solving for 𝑎1 , 𝑎2

𝑢 0 = 𝑢1 = 𝑎1
𝑢 𝐿 = 𝑢2 = 𝑎2 𝐿 + 𝑢1
Step 2: Select a Displacement Function

 Writing in the form of 𝑎2

𝑢2 − 𝑢1
𝑎2 =
𝐿

 Thus displacement function 𝑢 can be written as

𝑢2 − 𝑢1
𝑢= 𝑥 + 𝑢1
𝐿

 In matrix form 𝑥 𝑥 𝑢1
𝑢 = 1−
𝐿 𝐿 𝑢2
𝑢1
𝑢 = 𝑁1 𝑁2 𝑢2
𝑥 𝑥
where 𝑁1 = 1 − 𝑁2 =
𝐿 𝐿
Step 2: Select a Displacement Function

 𝑁𝑖 are called shape functions

 𝑁𝑖 represent the shape of the displacement function over the
domain (x-coordinate) when the ith degree of freedom has
unit value and all other degrees of freedom are zero

 𝑁1 and 𝑁2 are linear functions that

have the properties
 𝑁1 = 1 at node 1
 𝑁1 = 0 at node 2
 𝑁2 = 1 at node 2
 𝑁2 = 0 at node 1
 𝑁1 + 𝑁2 = 1 for any axial coordinate
along the bar
Step 2: Select a Displacement Function

 𝑁𝑖 are often called interpolation functions because we are

interpolating to find the value of a function between given
nodal values

 The interpolation function may be different from the actual

function except at the endpoints or nodes

 At endpoints or nodes, the interpolation function and actual

function must be equal to specified nodal values
Step 3: Define the Strain/Displacement
and Stress/Strain Relationships

 The tensile force 𝑇 produce a total elongation (deformation) 𝛿

of the spring

 𝑢1 is a negative value because the direction of displacement is

opposite the positive x-direction
 𝑢2 is a positive value
 The deformation of spring is given by

𝛿 = 𝑢 𝐿 − 𝑢 0 = 𝑢2 − 𝑢1
Step 3: Define the Strain/Displacement
and Stress/Strain Relationships

 Thus total deformation is the difference of nodal displacements

in x-direction

 For a spring, the force can be directly related to deformation,

thus strain/displacement relationship is not necessary

 Stress/Strain relationship can be expressed in terms of force/

displacement relationship

𝑇 = 𝑘𝛿

 or

𝑇 = 𝑘(𝑢2 − 𝑢1 )
Step 4: Derive the Element Stiffness
Matrix and Equations

 For the equilibrium of nodal forces

𝑓1𝑥 = −𝑇
𝑓2𝑥 = 𝑇
 Using the stress/strain relationship

𝑇 = −𝑓1𝑥 = 𝑘(𝑢2 − 𝑢1 ) 𝑇 = 𝑓2𝑥 = 𝑘(𝑢2 − 𝑢1 )

 or
𝑓1𝑥 = 𝑘(𝑢1 − 𝑢2 ) 𝑓2𝑥 = 𝑘(𝑢2 − 𝑢1 )

 In matrix format
𝑓1𝑥 𝑘 −𝑘 𝑢1
=
𝑓2𝑥 −𝑘 𝑘 𝑢2

Stiffness matrix for a

linear spring element
Step 4: Derive the Element Stiffness
Matrix and Equations

𝑘 −𝑘
𝑘 =
−𝑘 𝑘

 This is the local stiffness matrix for the element

 𝑘 is symmetric i.e. 𝑘𝑖𝑗 = 𝑘𝑗𝑖

 𝑘 is square matrix i.e. number of rows equals the number of

columns
Step 5: Assemble to Obtain Global
Equations and Apply Boundary Conditions

 The global stiffness matrix and global force matrix are assembled
using
 Nodal force equilibrium equations
 Force / deformation
 Compatibility equations
 This step applies to structures with more than one element such that

𝑁 𝑁

𝐾 = ෍ 𝑘 (𝑒) 𝐹 = ෍ 𝑓 (𝑒)
𝑒=1 𝑒=1

 Where 𝑘 (𝑒) and 𝑓 (𝑒) are now element stiffness and force
matrices expressed in global reference frame
 σ sign does not imply a simple summation but assembly
according to direct stiffness method
Step 6: Solve for Nodal Displacements

 Displacements are then determined by imposing

boundary conditions such as support conditions and
solving a system of equations simultaneously

𝐹 = 𝐾 𝑢
Step 7: Solve for Element Forces

 Finally the element forces are determined by back

substitution

 This step is applied to each element

Example of a Spring Assembly
Example of a Spring Assembly

 Consider the two spring assembly

 The nodes are numbered as 1, 3 and 2 (kept non sequential for

more generalisation)
 Stiffness of elements are 𝑘1 and 𝑘2
 Node 1 is fixed
 Axial force 𝐹2𝑥 is applied at node 2 and 𝐹3𝑥 is applied at node 3
 The local x-axis of each element coincides with the global x-axis
of the assembly
Example of a Spring Assembly

 For element 1
(1) (1)
𝑓1𝑥 𝑘1 −𝑘1 𝑢1
=
(1)
𝑓3𝑥 −𝑘1 𝑘1 𝑢(1)
3

 For element 2
(2) (2)
𝑓3𝑥 𝑘2 −𝑘2 𝑢3
=
(2)
𝑓2𝑥 −𝑘2 𝑘2 𝑢(2)
2

 Element 1 and 2 must remain connected at common node 3

throughout the displacement. Known as continuity or
compatibility requirement
Example of a Spring Assembly

 The compatibility requirement yields that

(1) (2)
𝑢3 = 𝑢3 = 𝑢3

 Superscript in parenthesis above 𝑢 refers to the element

number
 Subscript identifies the node of displacement and 𝑢3 is the
displacement of node 3 of the total or global spring assembly
 Example of a Spring Assembly

 Free body diagrams for each element and node are shown

 Considering the free body diagrams and the fact that external
forces must equal internal forces at each node
 The force equilibrium equations at nodes 2, 3 and 1 are
(2)
𝐹2𝑥 = 𝑓2𝑥
(1) (2)
𝐹3𝑥 = 𝑓3𝑥 + 𝑓3𝑥
Results from external applied (1)
reaction at the fixed support 𝐹1𝑥 = 𝑓1𝑥
 Example of a Spring Assembly

 We can write force equilibrium equations as

𝐹3𝑥 = −𝑘1 𝑢1 + 𝑘1 𝑢3 + (𝑘2 𝑢3 − 𝑘2 𝑢2 )

𝐹2𝑥 = −𝑘2 𝑢3 + 𝑘2 𝑢2

𝐹1𝑥 = 𝑘1 𝑢1 − 𝑘1 𝑢3

 In matrix form 𝐹3𝑥 𝑘1 + 𝑘2 −𝑘2 −𝑘1 𝑢3

𝐹2𝑥 = −𝑘2 𝑘2 0 𝑢2
𝐹1𝑥 −𝑘1 0 𝑘1 𝑢1

 Rearranging 𝐹1𝑥 𝑘1 0 −𝑘1 𝑢1 Global nodal

𝐹2𝑥 = 0 𝑘2 −𝑘2 𝑢2 displacement
Global 𝐹3𝑥 𝑢3 matrix
−𝑘1 −𝑘2 𝑘1 + 𝑘2
nodal force
matrix Global stiffness
matrix
Assembling the Global Stiffness Matrix by
Superposition (Direct Stiffness Method)

 A more convenient approach for constructing the global

stiffness matrix

 Based on superposition of individual element stiffness matrices

 The element stiffness matrices for the two springs are

𝑢1 𝑢3 𝑢3 𝑢2
𝑘1 −𝑘1 𝑢1 𝑘2 −𝑘2 𝑢3
𝑘 (1) = 𝑘 (2) =
−𝑘1 𝑘1 𝑢3 −𝑘2 𝑘2 𝑢2

 𝑢𝑖 degrees of freedom associated with each row and column

Assembling the Global Stiffness Matrix by
Superposition (Direct Stiffness Method)

 Stiffness matrices are associated with different degrees of

freedom (𝑘 (1) with 𝑢1 and 𝑢3 , 𝑘 (2) with 𝑢2 and 𝑢3 )

 Thus, matrices can not be added in present form

 To superimpose the element matrices, expand them to the

order (size) of the total structure stiffness matrix

 Add rows and columns of zeros for displacements not

associated with that particular element
Assembling the Global Stiffness Matrix by
Superposition (Direct Stiffness Method)

 For element 1, stiffness matrix in expanded form

𝑢1 𝑢2 𝑢3
(1) (1)
1 0 −1 𝑢1 𝑓1𝑥
(1) (1)
𝑘1 0 0 0 𝑢2 = 𝑓2𝑥
−1 0 1 (1) (1)
𝑢3 𝑓3𝑥
(1) (1)
 𝑢2 and 𝑓2𝑥 are not associated with 𝑘 (1)
 For element 2
𝑢1 𝑢2 𝑢3
(2) (2)
0 0 0 𝑢1 𝑓1𝑥
(2) (2)
𝑘2 0 1 −1 𝑢2 = 𝑓2𝑥
0 −1 1 (2) (2)
𝑢3 𝑓3𝑥
Assembling the Global Stiffness Matrix by
Superposition (Direct Stiffness Method)

 Force equilibrium at each node results in

(1)
𝑓1𝑥 0 𝐹1𝑥
(2)
0 + 𝑓2𝑥 = 𝐹2𝑥
(1) (2) 𝐹3𝑥
𝑓3𝑥 𝑓3𝑥
(1) (2)
1 0 −1 𝑢1 0 0 0 𝑢1 𝐹1𝑥
(1) (2)
𝑘1 0 0 0 𝑢2 + 𝑘2 0 1 −1 𝑢2 = 𝐹2𝑥
−1 0 1 (1) 0 −1 1 (2) 𝐹3𝑥
𝑢3 𝑢3
𝑘1 0 −𝑘1 𝑢1 𝐹1𝑥
0 𝑘2 −𝑘2 𝑢2 = 𝐹2𝑥
−𝑘1 −𝑘2 𝑘1 + 𝑘2 𝑢3 𝐹3𝑥

 This is same as was obtained earlier

Assembling the Global Stiffness Matrix by
Superposition (Direct Stiffness Method)

 The two expanded element stiffness matrices can be

directly added to obtain the global stiffness matrix

 The total set of stiffness equations is called direct stiffness

method

 For problems involving a large number of degrees of

freedom, expansion of each stiffness matrix becomes
tedious

 A direct form of direct stiffness matrix is used for this

purpose
Assembling the Global Stiffness Matrix by
Superposition (Direct Stiffness Method)

𝑢1 𝑢3 𝑢3 𝑢2
(1) 𝑘1 −𝑘1 𝑢1 (2) 𝑘2 −𝑘2 𝑢3
𝑘 = 𝑘 =
−𝑘1 𝑘1 𝑢3 −𝑘2 𝑘2 𝑢2

 𝐾 is constructed by directly adding terms associated with

degrees of freedom in 𝑘 (1) and 𝑘 (2) into their corresponding
identical degrees of freedom locations in 𝐾

𝑢1 𝑢2 𝑢3
𝑘1 0 −𝑘1 𝑢1
𝐾 = 0 𝑘2 −𝑘2 𝑢2
−𝑘1 −𝑘2 𝑘1 + 𝑘2 𝑢3
Boundary Conditions
Boundary Conditions

 The stiffness matrix 𝐾 is

 Symmetric
 Singular i.e. determinant of 𝐾 is zero and its inverse does not exist

 Without adequate kinematic constraints or support conditions,

structure will be free to move as a rigid body

 Rigid body motion does not resist any applied load

 In general, number of boundary conditions necessary to make

𝐾 non-singular is equal to possible rigid body mode
Boundary Conditions

 Boundary conditions are associated with nodal

displacements

 Homogeneous: occur at locations that are

completely prevented from movement

 Nonhomogeneous: occur where finite non zero

values of displacements are specified
Homogeneous Boundary Conditions

 Since node 1 is fixed so 𝑢1 = 0. Thus

𝑘1 0 −𝑘1 0 𝐹1𝑥
0 𝑘2 −𝑘2 𝑢2 = 𝐹2𝑥
−𝑘1 −𝑘2 𝑘1 + 𝑘2 𝑢3 𝐹3𝑥

 In expanded form
𝑘1 0 + 0 𝑢2 − 𝑘1 𝑢3 = 𝐹1𝑥
0 0 + 𝑘2 𝑢2 − 𝑘2 𝑢3 = 𝐹2𝑥
−𝑘1 0 − 𝑘2 𝑢2 + 𝑘1 + 𝑘2 𝑢3 = 𝐹3𝑥

 Where 𝐹1𝑥 is the unknown reaction and 𝐹2𝑥 , 𝐹3𝑥 are known
applied loads
 Writing the second and third 𝑘2 −𝑘2 𝑢2 𝐹2𝑥
=
equation in matrix form −𝑘2 𝑘1 + 𝑘2 𝑢3 𝐹3𝑥
Homogeneous Boundary Conditions

 Directly obtain the last matrix equation by deleting the row

and column corresponding to zero displacement dof

 In this example, row 1 and column 1 are deleted

 𝐹1𝑥 is not necessarily zero and can be determined once 𝑢2 and

𝑢3 are solved for. Thus we can write

1 1 1
−1 +
𝑢2 𝑘2 −𝑘2 𝐹2𝑥 𝑘 𝑘1 𝑘1 𝐹2𝑥
𝑢3 = = 2
−𝑘2 𝑘1 + 𝑘2 𝐹3𝑥 1 1 𝐹3𝑥
𝑘1 𝑘1

 Now 𝐹1𝑥 is given by 𝐹1𝑥 = −𝑘1 𝑢3

Homogeneous Boundary Conditions

 The unknown nodal force (reaction) 𝐹1𝑥 can be written in terms

of applied nodal forces

𝐹1𝑥 = −𝐹2𝑥 − 𝐹3𝑥

 For all homogeneous boundary conditions

 Delete the rows and columns corresponding to the zero-displacement

dof from original set of equations

 Solve for the unknown displacements

 Generally useful for hand calculations

Nonhomogeneous Boundary
Conditions

 One or more of the specified displacements are non zero

 Let 𝑢1 = 𝛿 where 𝛿 is a known displacement. Thus

𝑘1 0 −𝑘1 𝛿 𝐹1𝑥
0 𝑘2 −𝑘2 𝑢2 = 𝐹2𝑥
−𝑘1 −𝑘2 𝑘1 + 𝑘2 𝑢3 𝐹3𝑥
Nonhomogeneous Boundary
Conditions

 In expanded form

𝑘1 𝛿 + 0 𝑢2 − 𝑘1 𝑢3 = 𝐹1𝑥
0 𝛿 + 𝑘2 𝑢2 − 𝑘2 𝑢3 = 𝐹2𝑥
−𝑘1 𝛿 − 𝑘2 𝑢2 + 𝑘1 + 𝑘2 𝑢3 = 𝐹3𝑥

 Where 𝐹1𝑥 is now a reaction from support that has moved an

amount 𝛿

 Considering the equations with known right side and

transforming 𝛿 terms to right side

𝑘2 𝑢2 − 𝑘2 𝑢3 = 𝐹2𝑥
−𝑘2 𝑢2 + 𝑘1 + 𝑘2 𝑢3 = 𝐹3𝑥 + 𝑘1 𝛿
Nonhomogeneous Boundary
Conditions

 In matrix form
𝑘2 −𝑘2 𝑢2 𝐹2𝑥
𝑢3 =
−𝑘2 𝑘1 + 𝑘2 𝑘1 𝛿 + 𝐹3𝑥

 In case of nonhomogeneous boundary conditions

 Can not delete row 1 and column 1 corresponding to
nonhomogeneous boundary condition
 Doing so would neglect 𝑘1 𝛿 term and results would be erroneous
 To solve in general
 Transform the terms associated with the known displacements to
the right-side of force matrix before solving for the unknown nodal
displacements

 Now reaction 𝐹1𝑥 becomes

𝐹1𝑥 = 𝑘1 𝛿 −𝑘1 𝑢3
Properties of Stiffness Matrix
 𝐾 is square as it relates the same number or forces and
displacements

 𝐾 is symmetric, as is each of the element stiffness matrices

 𝐾 is singular i.e. its determinant is equal to zero

 No inverse exists till singularity is removed by applying boundary
conditions and prevent rigid body motion

 Main diagonal terms of 𝐾 are always positive

 Otherwise a positive nodal force will create a negative
displacement which is physically not possible

 𝐾 is positive semidefinite i.e. 𝑥 𝑇

𝐾 𝑥 > 0 is for all non zero
vector 𝑥 with real numbers
Examples of Stiffness Method
Example 1
For the spring assembly with arbitrarily numbered nodes shown in
figure, obtain
(a) the global stiffness matrix
(b) the displacements of nodes 3 and 4
(c) the reaction forces at nodes 1 and 2, and
(d) the forces in each spring
A force of 5000 lb is applied at node 4 in the x direction. The spring
constants are given in the figure. Nodes 1 and 2 are fixed
Example 1
 Element stiffness matrices are given by 𝑘 −𝑘
𝑘 =
−𝑘 𝑘
1 3 3 4
𝑘 (1) =
1000 −1000 1 𝑘 (2) =
2000 −2000 3
−1000 1000 3 −2000 2000 4
4 2
Numbers above and right
𝑘 (3) =
3000 −3000 4 are the nodal degrees of
−3000 3000 2 freedom

 Using superposition (stiffness method), global stiffness matrix is

𝑘 = 𝑘 (1) + 𝑘 (2) + 𝑘 (3)

 Example 1

 Global equations are given by

Example 1
 Applying boundary conditions 𝑢1 = 0 and 𝑢2 = 0
 Deleting first two rows and columns of 𝐾 corresponding to
applied displacements
 Also deleting first two rows of 𝐹 and 𝑑

 Global nodal displacements are

Example 1
 To obtain global nodal forces, back substitute the
displacements in global matrix equation

 The sum of the reactions 𝐹1𝑥 and 𝐹2𝑥 is equal in magnitude but
opposite in direction to the applied force 𝐹4𝑥
Example 1
Element forces
 Element 1

Free Body Diagram of Element 1

Example 1
 Element 2

Free Body Diagram of Element 2

Example 1
 Element 3

Free Body Diagram of Element 3

Example 2
For the spring assembly shown in figure, obtain
(a) the global stiffness matrix
(b) the displacements of nodes 2–4
(c) the global nodal forces, and
(d) the local element forces
Node 1 is fixed while node 5 is given a fixed, known displacement
𝛿 = 20.0 𝑚𝑚. The spring constants are all equal to 𝑘 = 200 𝑘𝑁/𝑚
Example 2
 Element stiffness matrices are

 Using superposition (stiffness method), global stiffness matrix is

Example 2
 Thus

 Applying boundary conditions 𝑢1 = 0 and 𝑢5 = 20 𝑚𝑚 = 0.02 𝑚

 Substituting the known global forces 𝐹2𝑥 = 0, 𝐹3𝑥 = 0, and 𝐹4𝑥 = 0
 Partitioning the first and fifth equations
Example 2

 Transposing the product of the appropriate stiffness coefficient

(-200) multiplied by the known displacement (0.02m) to the left
side
Example 2
 Solving we obtain

 To obtain global nodal forces, back substitute the displacements

in global matrix equation
Example 2

Remember

 If the displacement is known at a node in a given direction

(𝛿 = 20.0 𝑚𝑚) then the force (𝐹5𝑥 ) at that same node and in
that same direction is not initially known

 The force is determined after solving for the unknown nodal

displacements

 Next we determine the element forces

Example 2
 Element 1

 Element 2
Example 2
 Element 3

 Element 4
Example 3

This example uses the major concepts discussed

(a) Using the ideas presented earlier for the system of linear elastic
springs shown in figure, express the boundary conditions, the
compatibility or continuity condition and the nodal equilibrium
conditions. Then formulate the global stiffness matrix and
equations for solution of the unknown global displacement
and forces. The spring constants for the elements are 𝑘1 ; 𝑘2 and
𝑘3 ; P is an applied force at node 2

(b) Using the direct stiffness method,

formulate the same global stiffness
matrix and equation as in above
part
Example 3
Part (a)
 The boundary conditions are

 The compatibility condition at node 2 is

 The nodal equilibrium conditions are

Example 3

 The free body diagrams of elements and nodes

 Using the local stiffness matrix equation applied to each

element and compatibility condition equation, obtain the total
or global equilibrium equations as
Example 3

 In matrix form
Example 3

 Making use of boundary conditions, solve for 𝑢2 (deleting rows

and columns 1, 3 and 4)

 Solving for global forces

 These are global reactions

Example 3

Part (b)

 Using stiffness method, the element stiffness matrices are

 Then the global stiffness matrix is

Example 3

 Global stiffness equations are as already obtained

Potential Energy Approach to
Derive Spring Element Equations

Finite Element Method

Potential Energy Approach
 Alternative method to derive element equations and stiffness
matrix based on principle of minimum potential energy
 More generally applicable than method discussed before that
requires nodal and element equilibrium equations along with
stress / strain law
 Better adapted to complex elements with large number of dof
 Important to remember
 Principle of virtual work is applicable for any material behaviour
 Principle of minimum potential energy is applicable only for elastic
materials
 Both principles are part of general category of variational
methods
Potential Energy Approach
 Variational methods allow formulations for field problems,
primarily non structural type such as heat transfer, fluid flow,
electrical potential, etc.
 For other classes for which a variational function is not
definable, weighted residual methods can be used such as
Galerkin’s method

 Total potential energy, 𝜋𝑃 , of a structure is expressed in terms of

displacements
 In FEM, these are nodal displacements

𝜋𝑃 = 𝜋𝑃 (𝑑1 , 𝑑2 , 𝑑3 … 𝑑𝑛 )

 When 𝜋𝑃 is minimised with respect to these displacements,

equilibrium equations result
 Potential Energy Approach

Principle of minimum potential energy

Of all the geometrically possible shapes that a body can

assume, the true one, corresponding to the satisfaction of
stable equilibrium of the body, is identified by a minimum
value of the total potential energy

Total potential energy

The sum of the internal strain energy U and the potential

energy of the external forces, Ω
𝜋𝑃 = 𝑈 + Ω
 Potential Energy Approach

Strain Energy

The capacity of internal forces (or stresses) to do work

through deformations (strains) in the structure

Potential Energy of External Forces

The capacity of forces such as body forces, surface traction

forces, and applied nodal forces to do work through
deformation of the structure
Potential Energy

 Potential energy of an elastic body

𝜋𝑃 = 𝑆𝑡𝑟𝑎𝑖𝑛 𝐸𝑛𝑒𝑟𝑔𝑦 𝑈 + 𝐸𝑥𝑡𝑒𝑟𝑛𝑎𝑙 𝑊𝑜𝑟𝑘 𝑃𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙 (Ω)

 Consider a linear spring

Strain Energy of Linear Spring

 Consider a small change in 𝑑𝑈

displacement 𝑑𝑢
 Differential strain energy of the
𝐹
spring

𝑑𝑈 = 𝐹 𝑑𝑥
𝑥 𝑥 + 𝑑𝑥
 For a linear spring

𝑑𝑈 = 𝑘𝑥 𝑑𝑥

 Total strain energy of the spring

𝑥
1 2
𝑈 = න 𝑘𝑥 𝑑𝑥 = 𝑘𝑥
0 2
Strain Energy of Linear Spring

 Potential energy of a linear spring

𝜋𝑃 = 𝑆𝑡𝑟𝑎𝑖𝑛 𝐸𝑛𝑒𝑟𝑔𝑦 𝑈 + 𝐸𝑥𝑡𝑒𝑟𝑛𝑎𝑙 𝑊𝑜𝑟𝑘 𝑃𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙 (Ω)

 Work done by an external force is simply forces multiplied by

displacement at the point of application of force
 Work potential is negative because potential energy of external
forces is lost when work is done by external forces

1 2
𝜋𝑃 = 𝑘𝑥 − 𝐹𝑥
2
Potential Energy Approach
 A structure can have any deformed shape but for a particular
force it will only attain a unique deformation to attain
equilibrium
 Principle of minimum potential energy implies that the unique
deformation corresponds to the minimum potential energy
 At minimum potential energy, equilibrium state of structure is
stable
 To find minimum potential energy for spring, differentiate w.r.t 𝑥

𝑑𝜋𝑃
=0
𝑑𝑥
𝑘𝑥 − 𝐹 = 0

𝑘𝑥 = 𝐹 Same as Hooke’s law

Potential Energy Approach

 An admissible variation is one in which the displacement field

still satisfies the boundary conditions and inter element
continuity
Potential Energy Approach

Example

For the linear-elastic spring subjected to a force of 1000 lb shown

in figure, evaluate the potential energy for various displacement
values and show that the minimum potential energy also
corresponds to the equilibrium position of the spring
 Example

Potential Energy Approach

 Total potential energy is

 where

 Now minimise 𝜋𝑃 by taking derivative w.r.t 𝑥 as it’s a functions

of only 𝑥 displacement

 Also we have
 Example

Potential Energy Approach

 Back substituting the value of 𝑥

 For given values of F and k we have

 Using the above equation, we search for minimum value of 𝜋𝑃

by plotting against displacement values
 Example

Potential Energy Approach

 𝜋𝑃 has a minimum value at 𝑥 = 2.0 which corresponds to the

equilibrium position
Potential Energy Approach

Spring Element

 Now we derive spring element equations and stiffness matrix

using the principle of minimum potential energy

 Total potential energy

Potential Energy Approach
 Simplifying

 To minimise 𝜋𝑃 , we take partial derivatives w.r.t each nodal

displacement

 Simplifying

 In matrix form 𝑓1𝑥 𝑘 −𝑘 𝑢1

=
𝑓2𝑥 −𝑘 𝑘 𝑢2
Potential Energy Approach

Example

Obtain the total potential energy of the spring assembly and find
its minimum value. The procedure of assembling element
equations can then be seen to be obtained from the minimization
of the total potential energy
 Example

Potential Energy Approach

 The strain energy stored in spring 1

 Where difference in nodal displacements 𝑢3 − 𝑢1 is the

deformation 𝑥 in the spring
 In matrix form

 Above equation shows that strain energy 𝑈 is a quadratic

function of nodal displacements
 Similarly for springs 2 and 3
 Example

Potential Energy Approach

 Total strain energy is a scalar quantity

 The potential energy of the external nodal forces

 In matrix form

 Total potential of the assembly

 Example

Potential Energy Approach

 Minimising 𝜋𝑃 w.r.t each nodal displacement
 Example

Potential Energy Approach

 In matrix form

 Same as was obtained using direct stiffness method

Summary
 Element stiffness matrix 𝑓 = 𝑘 𝑑
 Global or total stiffness matrix 𝐹 = 𝐾 𝑑

 Displacement function assumed for 𝑢 = 𝑎1 + 𝑎2 𝑥

linear spring element
𝑥 𝑥
𝑁
 Shape functions for linear spring element 1 = 1 − 𝑁2 =
𝐿 𝐿

 Basic matrix equation relating nodal forces to nodal displacement

for spring element
𝑓1𝑥 𝑘 −𝑘 𝑢1
=
𝑓2𝑥 −𝑘 𝑘 𝑢2
𝑘 −𝑘
 Stiffness matrix for linear spring element 𝑘 =
−𝑘 𝑘
 Global equations for a spring assembly 𝐹 = 𝐾 𝑑
Summary
 Total potential energy

 For a system of springs

Recommended Reading and
Problem Set
 Chapter 2, A First Course in the finite element method (fifth
edition) by Daryl L. Logan, Cengage Learning, 2012.