Finite Element Method

CHAPTER‐01 FINITE ELEMENT (FE)

CONCEPTS
Ch‐01 Finite Element Concepts: Chapter Highlights
• What is Finite Elements?
• Simple Examples
• General Solution Steps

• Finite Element Generalization

• General solutions of continuum problems
• Differential and Finite Element Solution Methods
• Solution of simple static problem
• Basic Concepts and Definitions
• Elements, Types and Shapes
• Shape functions & Element Characteristic Equations
• Nodes and Degrees of Freedom (DOF).
• Assembly
• Solution and Boundary Conditions
• Element Types and DOF
• Truss and Beam
• Two Dimensional (2D) elements
• Three Dimensional (3D) and Shell Elements
• General Procedures for Finite Element Analysis
• Brief History of the Development of the FEM
What is Finite Element (FE)?
EXAMPLE 1.1
It is required to calculate the area of the irregular shape given in Figure 1.6. Also, it is
required to calculate the ratio of the bottom leg rectangular area to the total area of
the
structure.
l6

6
l5 5
l4
4

2 3
l3
l7
1

l2 l1
What is Finite Element (FE)? Cont.

APPROACH AND ASSUMPTIONS:

This is a simple example in which we demonstrate the solution steps and use as a
reference for other more realistic examples. Since we don’t know an equation to give
us the area of the shape; divide the shape into smaller regions (later will be called
‘elements’) that we can easily handle (i.e., in this case: find its area). There is no
conceptual constraint to have these elements of the same type and/or shape.

SOLUTION
Step-1: Discretize or Model the Structure
The area is modeled using small but finite number of elements that will be analyzed
separately.

Step-2: Element Characteristic Equations

Consider each element separately and write the characteristic equation for the element. In
this example, the only requirement is to calculate the area and, therefore, we write the area
equation for each element individually.
What is Finite Element (FE)? Cont.

For Element-1:
1 l7
The element characteristic equation is:
𝑎1 = 𝑙2 · 𝑙7
l2
Similarly, for other elements:
𝑎2 = 𝑙2 · (𝑙3 — 𝑙7) ; 𝑎3 = 𝑙1 · (𝑙3 — 𝑙7); 𝑎4 = 𝑙1 · 𝑙4 ; 𝑎5 = 1 𝑙26 · 𝑙5 ; 𝑎6 = 𝑙1 — 𝑙6 · 𝑙5

Step-3: Assembly of Element Equations into Global Equations

In this example, the assembly process is simply the summation of all the areas to
get the total or global area:

i= 1 𝑎 = 𝑙 · 𝑙 + 𝑙 · (𝑙 — 𝑙 )+ 𝑙 · (𝑙 — 𝑙 )+ 𝑙 · 𝑙 + 1 𝑙 · 2𝑙 +
𝐴 = ∑NELM 𝑙1 — 𝑙6 · 𝑙5
i 2 7 2 3 7 1 3 7 1 4 6 5

where N is the total number of elements.

What is Finite Element (FE)? Cont.

Step-4: Solution of Global Equations

In this case, the global gives the total area of the structure.

Step-5: Post Processing

From element equations, we have the area of element-1
𝑎1 = 𝑙2 · 𝑙7,

and the ratio to the total area will be given by 𝑎1⁄𝐴

What is Finite Element (FE)? Cont.

EXAMPLE 1.2
To further clarify the above steps, we consider the example of calculating the area
of a circle and we assume that we do not know the equation for the area, i.e., the
system behavior is unknown. H-version vs p-version

2 3

1 4

8 5
7 6

CONCLUSIONS & DEDUCTIONS FROM

EXAMPLE 1.2
Finite Element Generalization
Continuum Problem

Analytical & Exact

Numerical Methods
Solution Methods

Numerical Solution of Various Finite

Differential and Partial Element Approaches
Differential Equations

dz S c S
V V
a b
Y dx dy Y

X Z X
Z
Infinitesimal size elements leading to Finite size elements leading
differential and partial differential to algebraic equations
equations
Finite Element Generalization (Cont.)
Continuum Problem

Analytical Methods
Exact Solution Methods Numerical Methods

Matrix Techniques
Numerical Solution of
Residual Methods
Differential and Partial
Galerkin and
Differential Equations (DE and
Discrete Element
PDE)
Methods FE FE
FE
Weighted Physical
Variationa
Direct integration Residual Approach
l
Separation of Finite Numerical
variables difference integration Variation of Collocation Force method
Fourier transforms function Least square Displ. Method
Laplace transforms Constrained Galerkin Hybrid
variation Petrov- Ritz-Reissner
Lagrange ;
Galerkin
Penalty

- Have well-established - Suitable for general geometries &

Limited to simple geometries,
simple boundary conditions methods and boundary conditions
- Easy to incorporate nonlinear
and material models. criteria
- Difficult for general effects
- FE equations may be directly obtained
boundary
- Difficult to incorporate from the DE & PDE
- Programming and modeling are time
nonlinear effects
- Difficult to incorporate consuming
- Requires heavy computer resources
different material
models
Finite Element Generalization (Cont.)
SIMPLE ELEMENT Extension of a beam /shaft
EXAMPLE
Modeling/Discretization

P 1 2 2 3 3 4 P
1 Spring
k1 k2 k3 element 2-
nodes/elem
l1, A1 1-DOF/node
l2, A2 l3, A3 3
1 2 2 3 3 4 NNP=4
1 2
k2 k3 NELM=3
k1

𝐸i𝐴i
Element Characteristic Equations 𝑘i =
𝑙i

For element 1: F2(1)  k1 u2  u 1 and F (1)  R  k u  u ,

1 1 1 2 1

For element  2 : F3(2) 

, 2 u3  u 2  (2)
 k 2 u3  u 2 ,
2

For element  3 : k (3)  P , k u  u

F and F (3)  k u  u
4 3 4 3 3 3 4 3
 , and F 
Finite Element Generalization (Cont.)
SIMPLE ELEMENT
EXAMPLE
Assembly, Boundary Conditions & Solution

𝑢4 = 𝑃 1
+ k 2 + k 3 + 𝑢1
1
k1
1
Cannot be solved until we introduce the boundary conditions
𝒖𝟏 = 𝟎

3
1
𝑃= g 4 where 𝐾g = 1 / 1 + 1 + = 1/Σ
𝐾 𝑢 𝑘11 𝑘2 𝑘3
𝑘i
i=1

CONCLUSIONS & DEDUCTIONS FROM

EXAMPLE
Global and element variables
 Boundary conditions and rigid body motion
Finite Element Generalization (Cont.)
EXTENSION TO HEAT TRANSFER
PROBLEMS Tfluid

𝐹 = 𝑘 ∆𝑢; 𝐹i = Tsurface film

coefficient
—𝐹j
dT
Fourier
𝑞 = —𝑘Law: k A ∆𝑇; 𝑞i = —
c𝐴dx = — Lc
thermal
𝑞j k
Conductivities c1
k c2 k c3 k
c4
L Figure 2.2 Heat flow through composite wall
Y
L
q1 q2 x
q1 T1 p
T2 q2
node-1 node-2 X

Thermal conductivity, k t
𝑞1 𝑘t𝐴 𝑇1
Figure 2.3 Simple one-dimensional heat transfer 𝑞2 = 1
conduction element 𝑇2
—1
𝑞1 = k tLA 𝑇1 — 𝑇2 ; 𝑞2 = —𝑞1 = —L 𝑇1 — 𝐿 —1 1
ktA
𝑇2
In-class Self-marking Quiz
It is required to assemble the global equations and find the
displacement at node-2 for the axial members shown in Figure.
The members may be considered as spring elements with one
DOF per node.
Challenge: Try putting the equations in a matrix form
Y
1 P 2 k1 P k2
1 3 R 1
1 2
3 R3
X u1 = 0 u2 ≠ 0 u3 = 0
2
Two axial members with one DOF/node for each

Element-1: 𝐹(1) = 𝑘1 𝑢2 — 𝑢1 𝐹(1) = —𝑘1 𝑢2 — ≡ 𝑅1

𝑢1
2 1
Element−𝟐: 𝐹3(2) = 𝑘2 𝑢3 — 𝑢2 ≡ 𝑅3 𝐹2( 2 ) = —𝑘2 𝑢3 —
Apply BC: 𝑢1 = 𝑢3 = 0 → 𝑘1 + 𝑘𝑢22 𝑢2 = 𝐹(1) + 𝐹(2)
2 2
𝐹 (1) + 𝐹 (2) = 𝑃
From Eqm. At node-2: 2 2
In-class Self-marking Quiz (Cont.)
It is required to assemble the global equations and find the
displacement at node-2 for the axial members shown in Figure.
The members may be considered as spring elements with one
DOF per node.
Challenge: Try putting the equations in a matrix form
Y
1 P 2 k1 P k2
1 3 R 1
1 2
3 R3
X u1 = 0 u2 ≠ 0 u3 = 0
2
Two axial members with one DOF/node for each

1 𝑢1 𝐹1(1) 𝐹2(1) = 𝑘1 𝑢2
𝐸𝑙𝑒𝑚𝑒𝑛𝑡 — 1: 𝑘1 𝑢2 = 1
—1 —
𝐹1 𝑢= —𝑘1 𝑢2 —
(1)
𝐹2(1)
—1 1 𝑢1
(2)
1 𝑢2
𝐸𝑙𝑒𝑚𝑒𝑛𝑡 — 1: 𝐹3( 2 ) = 𝑘2 𝑢3 — 𝑢2
𝑘2
—1 𝐹32 = (2)
𝐹 𝐹2( 2 ) = —𝑘2 𝑢3 —
—1 1 𝑢
3 𝑢2
𝐾11 𝐾12 𝐾13 𝑢1 𝐹1
𝐴𝑠𝑠𝑒𝑚𝑏𝑙𝑦: 𝐾21 𝐾22 𝐾23 𝑢2 = 𝐹2
𝐾31 𝐾32 𝐾33 𝑢3 𝐹3
Basic Concepts and Definitions

Basic Concepts and Definitions

 Elements, Types and Shapes
 Shape functions & Element Characteristic
Equations
 Nodes and Degrees of Freedom (DOF).
 Assembly
 Solution and Boundary Conditions
Basic Concepts and Definitions

Remarks:
There are five basic steps or stages for any type of finite element analysis. These are:
modeling and discretization; formulation and element equations; assembly; boundary
conditions and solution; and finally post processing. In commercial FE programs
these steps are lumped into three stages: modeling; solution and post processing.
The assembly process is quite simple and straight forward. On the other hand, most of
the research done in the finite element area and most of the textbooks written in the
area will involve one or more of the other basic steps/stages.
The global finite element equations for structural analysis may be written as:
𝐾g 𝑢 = 𝑓 where 𝐾g is the global
stiffness matrix of the system, 𝑢 is the vector of displacement response
and 𝑓 is the vector of applied forces.
On the element level, the equation takes the form: 𝐾 ( e ) 𝑢(e) = 𝑓(e) . It should be
noted that at a specific node, the primary element variable 𝑢(e) is the same as the
global variable 𝑢 . However, the force variables 𝑓(e)at a given node have to be
equivalent to the global value '𝑓' at the node.
Element Types and Degrees of Freedom (DOF)

vj
vj
uj uj
Y Y
Y
wj
DOF/Node: u vi
vi DOF/Node: u, v DOF/Node: u, v,
ui uj i w
ui u X
i X wi i
X
One Dimensional Two Dimensional Three Dimensional

FIGURE 1.8 TRUSS

ELEMENTS
P
Can you model this structure with
truss/link elements? Explain your
answer.
Element Types and Degrees of Freedom (DOF): Cont.

vj yj
vj
j uj
Y zj
uj Y
zj xj
vi yi
DOF/Node: u, v, z vi
zi wj
ui
X i ui
zi
wi X
Two-Dimensional xi
Z Three-Dimensional

2-node (linear) 3-node (quadratic) 4-node (cubic)

FIGURE 1.9 BEAM

ELEMENTS
Element Types and Degrees of Freedom (DOF): Cont.
WHY ROTATION
DOF?
vj
Y
uj

vi
DOF/Node: u, v

X
ui
Two dimensional truss

vj
Y
uj
zj
vi
zi DOF/Node: u, v, z
ui
X
Two-Dimensional
Beam
Element for 2D solid structures?

vj
Y
uj

vi
DOF/Node: u, v

X
ui
Two dimensional truss

vj
Y
uj  Not possible to map the geometry?
zj  Must have a solid flat element
vi
zi DOF/Node: u, v, z
ui
X
Two-Dimensional
Beam
Element Types and Degrees of Freedom (DOF): Cont.
Element Types and Degrees of Freedom (DOF): Cont.
Element Types and Degrees of Freedom (DOF): Cont.
General Procedures for Finite Element Analysis

1. Discretization or Modeling and Preprocessing

2. Element Characteristic Equations
3. Assembly of Global Equations
4. Solution of Global Equations
5. Post processing (Other solution quantities)
Phases of Finite Element Analysis in Commercial Programs
Preprocessing Phase
CAD Program
FE pre-processing
Program Geometrical model for
the problem (CAD base)

Geometrical model for Interface with

the problem (FE base) FE program

FE and material model

for the problem

Interface with
FE program
Solution Phase
Calculate element stiffness and
equilibrium equations

Assembly, boundary
conditions and solution phase
Interface with
Post processing program

Post processing
Post processing phase & general display of
results
In-class Self-marking Quiz
For the shown problems, identify:
 Element type to be used
 Sketch the element
 No. of nodes per element
 DOF per node
In-class Self-marking Quiz
For the shown problems, identify:
 Element type to be used
 Sketch the element
 No. of nodes per element
 DOF per node
In-class Self-marking Quiz
For the shown problems, identify:
 Element type to be used
 Sketch the element
 No. of nodes per element
 DOF per node

1.5 in 1.5 in

5 in

6 in

1.5 in
Figure P11
In-class Self-marking Quiz
1.5 in
For the shown problems, identify: 1.5 in
 Element type to be
used
 Sketch the element
 No. of nodes per 5 in
element
 DOF per node
6 in

1.5 in
Figure P11