01introductiontothefiniteelementmethod Maged 100714091834 Phpapp02

Introduction to the Finite
Element Method
Spring 2010
Introduction to the Finite Element Method

Dr. Mohammad Tawfik
Course Objectives
The student should be capable of writing
simple programs to solve different
problems using finite element method.

Dr. Mohammad Tawfik
Assessment
10% Assignments (1 per week)
20% Quizzes (best 2 out of 3)
Week of 12/11/2006
Week of 20/12/2006
Week of 17/1/2006
20% Course Project
25% Midterm exam (Week of 2/12/2006)
25% Final exam (starting 3/2/2007)

Dr. Mohammad Tawfik
Fundamental Course Agreement
Homework is sent in electronic format (No
hardcopies are accepted)
Computer programs have to written in
MATLAB or Mathematica script
No late homework is accepted
No excuses are accepted for missing a
quiz
Best two out of three quizzes are counted

Dr. Mohammad Tawfik
References
J.N. Reddy, An Introduction to the Finite
Element Method 3rd ed., McGraw Hill, ISBN
007-124473-5
D.V. Hutton, Fundamentals of Finite Element
Analysis 1st ed., McGraw Hill, ISBN
007-121857-2
K. Bathe, Finite Element Procedures, Prentice
Hall, 1996. (in library)
T. Hughes, The finite Element Method: Linear
Static and Dynamic Finite Element analysis,
Dover Publications, 2000. (in library)
Dr. Mohammad Tawfik
Numerical Solution of
Boundary Value Problems
Weighted Residual Methods

Dr. Mohammad Tawfik
Objectives
In this section we will be introduced to the
general classification of approximate
methods
Special attention will be paid for the
weighted residual method
Derivation of a system of linear equations
to approximate the solution of an ODE will
be presented using different techniques

Dr. Mohammad Tawfik
Why Approximate?
Ignorance
Readily Available Packages
Need to Develop New Techniques
Good use of your computer!
In general, the problem does not have an
analytical solution!

Dr. Mohammad Tawfik
Classification of Approximate
Solutions of D.E.s
Discrete Coordinate Method
Finite difference Methods
Stepwise integration methods
Euler method
Runge-Kutta methods
Etc
Distributed Coordinate Method

Dr. Mohammad Tawfik
Distributed Coordinate Methods
Weighted Residual Methods
Interior Residual
Collocation
Galrekin
Finite Element
Boundary Residual
Boundary Element Method
Stationary Functional Methods
Reyligh-Ritz methods
Finite Element method

Dr. Mohammad Tawfik
Basic Concepts
A linear differential equation may be written in the form:
L( f ( x ) ) = g ( x )
Where L(.) is a linear differential operator.
An approximate solution maybe of the form:
n
f ( x ) = ai i ( x )
i =1

Dr. Mohammad Tawfik
Basic Concepts
Applying the differential operator on the approximate
solution, you get:
n

L( f ( x ) ) g ( x ) = L ai i ( x ) g ( x )
i =1
n
= ai L( i ( x ) ) g ( x ) 0
i =1
n
a L( ( x ) ) g ( x ) = R( x )
i =1
i i
Residue

Dr. Mohammad Tawfik
Handling the Residue
The weighted residual methods are all
based on minimizing the value of the
residue.
Since the residue can not be zero over the
whole domain, different techniques were
introduced.

Dr. Mohammad Tawfik
Collocation Method
The idea behind the collocation method is
similar to that behind the buttons of your
shirt!
Assume a solution, then force the residue
to be zero at the collocation points

Dr. Mohammad Tawfik
Collocation Method
R( x j ) = 0
R( x j ) =
a L( ( x ) ) F ( x ) = 0
n
i i j j
i =1

Dr. Mohammad Tawfik
Example Problem

Dr. Mohammad Tawfik
The bar tensile problem
2u
EA 2 + F ( x ) = 0
x

Dr. Mohammad Tawfik
Bar application
2u
EA 2 + F ( x ) = 0
x
n
u ( x ) = ai i ( x )
i =1
n
d 2 i ( x )
EA ai 2
+ F ( x ) = R( x ) Applying the collocation method
i =1 dx
d 2 i ( x j )
+ F(xj ) = 0
n
EA ai 2
i =1 dx

Dr. Mohammad Tawfik
In Matrix Form
k11 k 21 ... k n1 a1 F ( x1 )
k k 22 ... k n 2 a2 F ( x )
12 2 d 2 i ( x )
= kij = EA
dx 2 x= x j
F ( xn )
k1n k 2 n ... k nn an
Solve the above system for the generalized

coordinates ai to get the solution for u(x)

Dr. Mohammad Tawfik
Notes on the trial functions
They should be at least twice
differentiable!
They should satisfy all boundary
conditions!
Those are called the Admissibility
Conditions.

Dr. Mohammad Tawfik
Using Admissible Functions
For a constant forcing function, F(x)=f
The strain at the free end of the bar should
be zero (slope of displacement is zero).
We may use: x
( x ) = Sin
2l

Dr. Mohammad Tawfik
Using the function into the DE:
d ( x)
2
x
2
EA 2
= EA Sin
dx 2l 2l
Since we only have one term in the series,
we will select one collocation point!
The midpoint is a reasonable choice!

2

EA Sin { a1} = { f }
2l 4

Dr. Mohammad Tawfik
Solving:
f 4 2l 2 f l2 f
a1 = = 0.57
EA( 2l ) Sin( 4 )
2 2
EA EA
Then, the approximate l2 f x

u ( x ) 0.57 Sin
solution for this problem is: EA 2l
Which gives the maximum
l2 f
displacement to be: u ( l ) 0.57 ( exact = 0.5)
EA
lf
And maximum strain to be: u x ( 0) 0.9 ( exact = 1.0)
EA
Dr. Mohammad Tawfik
The Subdomain Method (free
reading)
The idea behind the
subdomain method is
to force the integral
of the residue to be
equal to zero on an
subinterval of the
domain

Dr. Mohammad Tawfik
The Subdomain Method
x j +1
R( x ) dx = 0
xj
n x j +1 x j +1
a L( ( x ) ) dx g ( x ) dx = 0
i =1
i i
xj xj

Dr. Mohammad Tawfik
Bar application
2u
EA 2 + F ( x ) = 0
x
n
u ( x ) = ai i ( x )
i =1
n
d 2 i ( x )
EA ai 2
+ F ( x ) = R( x ) Applying the subdomain method
i =1 dx
d i ( x)
n x j +1 2 x j +1
EA ai 2
dx = F ( x ) dx
i =1 xj
dx xj

Dr. Mohammad Tawfik
In Matrix Form
x j+1 d 2 ( x ) x j+1
EA i
2
dx { ai } = F ( x ) dx
x j dx x j


Dr. Mohammad Tawfik
The Galerkin Method
Galerkin suggested that the residue
should be multiplied by a weighting
function that is a part of the suggested
solution then the integration is performed
over the whole domain!!!
Actually, it turned out to be a VERY
GOOD idea

Dr. Mohammad Tawfik
The Galerkin Method
R( x ) ( x ) dx = 0
Domain
j
a ( x ) L( ( x ) ) dx ( x ) g ( x ) dx = 0
i =1
i j i j
Domain Domain

Dr. Mohammad Tawfik
Bar application
2u
EA 2 + F ( x ) = 0
x
n
u ( x ) = ai i ( x )
i =1
n
d 2 i ( x )
EA ai 2
+ F ( x ) = R( x ) Applying Galerkin method
i =1 dx
n
d 2 i ( x )
EA ai j ( x ) 2
dx = j ( x ) F ( x ) dx
i =1 Domain
dx Domain

Dr. Mohammad Tawfik
In Matrix Form
d 2 i ( x )
EA j ( x ) 2
dx { ai } = j ( x ) F ( x ) dx
Domain dx Domain


Dr. Mohammad Tawfik
Same conditions on the functions
are applied
They should be at least twice
differentiable!
They should satisfy all boundary
conditions!
Lets use the same function as in the
collocation method: x
( x ) = Sin
2l

Dr. Mohammad Tawfik
Substituting with the approximate
solution:
n
d 2 i ( x )
EA ai j ( x ) 2
dx = j ( x ) F ( x ) dx
i =1 Domain
dx Domain
2
x x
l
EA a1 Sin Sin dx
2l 0 2l 2l
x
l
= Sin fdx
0 2l
2
l 2l f 16l 2 fl 2
EA a1 = a1 = 0.52
2l 2 EA 3
EA

Dr. Mohammad Tawfik
Substituting with the approximate
solution: (Int. by Parts)
n
d 2 i ( x )
EA ai j ( x ) 2
dx = j ( x ) F ( x ) dx
i =1 Domain
dx Domain
d 2 i ( x )
j ( x) 2
dx
Domain
dx
d i ( x )
l
d j ( x ) d i ( x )
= j ( x) dx
!Zero dx 0 Domain dx dx
2
l 2l f 16l 2 fl 2
EA a1 = a1 = 0.52
2l 2 EA 3
EA

Dr. Mohammad Tawfik
What did we gain?
The functions are required to be less
differentiable
Not all boundary conditions need to be
satisfied
The matrix became symmetric!

Dr. Mohammad Tawfik
Summary
We may solve differential equations using a
series of functions with different weights.
When those functions are used, Residue
appears in the differential equation
The weights of the functions may be determined
to minimize the residue by different techniques
One very important technique is the Galerkin
method.

Dr. Mohammad Tawfik
NOTE
Next Sunday 5/11 (No lecture)
Following week 12/11, Quiz #1 will be held
covering all the material up-to this lecture
Homework #1 is due next week (Electronic
submission of report and code is
mandatory.

Dr. Mohammad Tawfik
Report Should Include
Cover page
Introduction section indicating the
procedure you used with the equations as
implemented in your code
Results section
Observations and Conclusions if any
according to the output of your program.

Dr. Mohammad Tawfik
Homework #1
Solve the beam bending
problem, for beam
displacement, for a simply
supported beam with a load
placed at the center of the
beam using
Collocation Method
Subdomain Method d 4w
Galerkin Method 4
= F ( x)
dx
Use three term Sin series that w(0) = w(l ) = 0
satisfies all BCs
Write a program that produces d 2 w(0) d 2 w(l )
= =0
the results for n-term solution. dx 2
dx 2

Dr. Mohammad Tawfik
Exact Solution
x 3 13 x
w( x) = + 0 < x < 1/ 2
12 60
x3 x 2 7 x 3
= + + 1/ 2 < x < 1
12 4 15 10

Dr. Mohammad Tawfik
The Finite Element Method
2nd order DEs in 1-D

Dr. Mohammad Tawfik
Objectives
Understand the basic steps of the finite
element analysis
Apply the finite element method to second
order differential equations in 1-D

Dr. Mohammad Tawfik
The Mathematical Model
Solve:
d du
a + cu f = 0
dx dx
0< x<L
Subject to:
du
u ( 0 ) = u0 , a = Q0
dx x = L

Dr. Mohammad Tawfik
Step #1: Discretization
At this step, we divide
the domain into
elements.
The elements are
connected at nodes.
All properties of the
domain are defined at
those nodes.

Dr. Mohammad Tawfik
Step #2: Element Equations
Lets concentrate our
attention to a single
element. d du
The same DE applies a + cu f = 0
dx dx
on the element level,
hence, we may follow x1 < x < x2
the procedure for u ( x1 ) = u1 , u ( x2 ) = u 2 ,
weighted residual
methods on the du du
a = Q1 , a = Q2
element level! dx x = x1 dx x = x2

Dr. Mohammad Tawfik
Polynomial Approximation
Now, we may propose an approximate
solution for the primary variable, u(x),
within that element.
The simplest proposition would be a
polynomial!

Dr. Mohammad Tawfik
Interpolating the values
of displacement
knowing the nodal
displacements, we may
write:
u ( x ) = b1 x + b0
u ( x1 ) = u1 = b1 x1 + b0 x2 x x x1
u ( x ) = u1 + u2
u ( x2 ) = u2 = b1 x2 + b0 x2 x1 x2 x1

Dr. Mohammad Tawfik
x2 x x x1
u( x) = u1 + u2
x2 x1 x2 x1
u1
{ }
= 1u1 + 2u2 = 1 2 = ( x ) u e
u2

Dr. Mohammad Tawfik
(contd)
Assuming constant d 2u
a 2 + cu f = 0
dx
domain properties:
x1 < x < x2
Applying the
Galerkin method:
d 2 i ( x )
a j ( x ) 2
ui + c j ( x ) i ( x ) ui j ( x ) f dx = 0
Domain
dx

Dr. Mohammad Tawfik
(contd)
Note that: d 2 i ( x )
a j ( x ) 2
dx
Domain
dx
d i ( x )
x2
d j ( x ) d i ( x )
= a j ( x ) +a dx
dx x1 Domain
dx dx
d 1 ( x ) 1 d 2 ( x ) 1
And: = , =
dx he dx he

Dr. Mohammad Tawfik
(contd)
For i=j=1: (and ignoring boundary terms)
x2 1 x x
2
x2 x
a u
x he 2 he 1
+ c 2
f
he
dx = 0
1
Which gives:
a che fhe
+ u1 =0
he 3 2

Dr. Mohammad Tawfik
(contd)
Repeating for all terms:
a 1 1 che 2 1 u1 fhe 1
1 1 + 6 1 2 u = 2 1
h
e 2
The above equation is called the element

equation.

Dr. Mohammad Tawfik
What happens for adjacent
elements?

Dr. Mohammad Tawfik
Homework #2
Derive the element equation without
ignoring the boundary terms.
What are differences in the element
equation.
The solution should be handed using the
same report format (use equation editor to
write your report).

Dr. Mohammad Tawfik
Finite Element Procedure
1. Connecting Elements
2. Boundary Conditions
3. Solving Equations
Dr. Mohammad Tawfik
Objectives
Learn how the finite element model for the
whole domain is assembled
Learn how to apply boundary conditions
Solving the system of linear equations

Dr. Mohammad Tawfik
Recall
In the previous lecture, we obtained the
element equation that relates the element
degrees of freedom to the externally
applied fields
a 1 1 che 2 1 u1 fhe 1
1 1 + 6 1 2 u = 2 1
h
e 2
k1 k 2 u1 f1
Which maybe written: k =
3 k 4 u2 f 2

Dr. Mohammad Tawfik
TwoElement example
k11 k 21 u11 f11
1 1 1
= 1
k3 k 4 u 2 f 2
k12 k 22 u12 f12

2 2 2
= 2
k3 k 4 u2 f 2
k11 k 21 0 u1 f1 Q1
1 2
k3 k +k
1
4 1
2
k 2 u 2 = f 2 + Q2
0
k32 k 42 u3 f 3 Q
3

Dr. Mohammad Tawfik
Illustration: Bar application
2u
EA 2 + F ( x ) = 0
x
1. Discretization: Divide the bar into N number of

elements. The length of each element will be (L/
N)
2. Derive the element equation from the differential
equation for constant properties an externally
applied force: x
d
EA j d i
2
x he 2 dx dx u j f i dx = 0

1
Dr. Mohammad Tawfik
Performing Integration:
x2
EA d j d i EA 1 1 u1e fhe 1
x he 2 dx dx u j f i dx = 0 e =
he 1 1 u 2 2 1
1
Note that if the integration is evaluated from 0 to he,

where he is the element length, the same results
will be obtained.

Dr. Mohammad Tawfik
TwoElement bar example
EA 1 1
1 u 1
f 1

1 = 1
1

he 1 1 u2 f 2
EA 1 1 u12 f12
2 = 2
he 1 1 u 2 f 2
1 1 0 u1 1 R
EA fhe
1 2 1 u = 2 + 0
he 2
2
0 1 1 u3 1 0


Dr. Mohammad Tawfik
Applying Boundary Conditions

Dr. Mohammad Tawfik
Applying BCs
For the bar with fixed left side and free
right side, we may force the value of the
left-displacement to be equal to zero:
1 1 0 0 1 R
EA fhe
1 2 1 u2 = 2 + 0
he 2
0 1 1 u3 1 0

Dr. Mohammad Tawfik
Solving
Removing the first row and column of the
system of equations:
EA 2 1 u2 fhe 2
=
he 1 1 u3 2 1
Solving:
u2 fhe 2 3
=
u
3 2 EA 4

Dr. Mohammad Tawfik
Secondary Variables
Using the values of the displacements
obtained, we may get the value of the
reaction force:
1 1 0 0 1 R
1 2 1 3 fhe = fhe 2 + 0
2 2
0 1 1 4 fh 1 0

e
2

Dr. Mohammad Tawfik
Secondary Variables
Using the first equation, we get:
3 fhe fhe
= +R
2 2
R = 2 fhe
Which is the exact value of the reaction

force.
Dr. Mohammad Tawfik
Summary
In this lecture, we learned how to
assemble the global matrices of the finite
element model; how to apply the boundary
conditions, and solve the system of
equations obtained.
And finally, how to obtain the secondary
variables.

Dr. Mohammad Tawfik
Homework #3
Problems #3.9 & 3.13 from the text book
Write down a computer code that solves
the problem for N elements.

Dr. Mohammad Tawfik
Bars and Trusses

Dr. Mohammad Tawfik
Objectives

Dr. Mohammad Tawfik
Bar Example (Ex. 4.5.2, p. 187)
Consider the bar shown in the above figure.

It is composed of two different parts. One steel tapered
part, and uniform Aluminum part.
Calculate the displacement field using finite element
method.

Dr. Mohammad Tawfik
Bar Example
The bar may be represented by two
elements.
The stiffness matrices of the two elements
may be obtained using the following
integration:
d 1 1 1
x2
dx d 1 d 2
x2 h2 he2
K e = EA( x ) dx = x EA( x ) e
dx
x1
d
2 dx dx 1
1 1
dx he2 he2
Dr. Mohammad Tawfik
Bar Example
For the Aluminum bar: E=107 psi, and A=1
in2. we get:
107 2 1 1 10 7 1 1
x
2
K Al = dx =
120 x1 1 1 120 1 1
For the Steel bar: E=38107 psi, and

A=(1.5-0.5x/96) in 2
. we get:
0.5 x 1 1 1 1
x
3.107 2 4.75.107
2
K Fe = 1.5 dx = 1 1
96 x1 96 1 1 96
Dr. Mohammad Tawfik
Bar Example
Assembling the Stiffness matrix and
utilizing the external forces, we get:
49.5 49.5 0 u1 0 R

10 4 49.5 57.8 8.33 u2 = 2.105 + 0
0 8.33 8.33 u3 105 0
The boundary conditions may be applied

and the system of equations solved.

Dr. Mohammad Tawfik
Bar Example
Solving, we get: u2 0.061
= in
u3 0.181
For the secondary

variables: R = 30000 lb

Dr. Mohammad Tawfik
Reading Task
Please read and understand examples,
4.5.1 & 4.5.3.

Dr. Mohammad Tawfik
Trusses
A truss is a set of bars that are connected
at frictionless joints.
The Truss bars are generally oriented in
the plain.

Dr. Mohammad Tawfik
Trusses
Now, the problem lies in the
transformation of the local displacements
of the bar, which are always in the
direction of the bar, to the global degrees
of freedom that are generally oriented in
the plain.

Dr. Mohammad Tawfik
Equation of Motion
1 0 1 0 u 1 F 1
0
EA 0 0 0 v1 0
=
h 1 0 1 0 u 2 F 2

0 0 0 0 v 2 0

Dr. Mohammad Tawfik
Transformation Matrix
u1 Cos ( ) Sin( ) 0 0 u1
Sin( ) Cos ( )
v1 0 0 v1
=
u 2 0 0 Cos ( ) Sin( ) u2

v 2
Local DOF 0 0 Sin( ) Cos ( ) v2 Transformed
DOF
{} Local DOF = [T ]{ } Transforme

DOF
d

Dr. Mohammad Tawfik
The Equation of Motion Becomes
Substituting into the
FEM: [ K ][T ]{ } = {F }
Transforming the
forces:
[T ] T
[ K ][T ]{ } = [T ] {F }
T
Finally: [ K ]{ } = { F }
Dr. Mohammad Tawfik
Recall
[ K ] = [T ]T [ K ][T ]
:Where
1 0 1 0
Cos ( ) Sin( ) 0 0
0 0
Sin( )
[T ] =
Cos ( ) 0 0 [ ]
K =
EA
h 1
0 0
0 1 0
0 0 Cos ( ) Sin( )

0 0 Sin( ) Cos ( ) 0 0 0 0

Dr. Mohammad Tawfik
Element Stiffness Matrix in Global
Coordinates
Cos ( ) Sin( ) 0 0 1 0 1 0
Sin( ) Cos ( ) 0 0 0 0 0 0
EA
[K] =
h 0 0 Cos ( ) Sin( ) 1 0 1 0

0 0 Sin ( ) Cos ( ) 0 0 0 0
Cos ( ) Sin( ) 0 0
Sin( ) Cos ( ) 0 0

0 0 Cos ( ) Sin( )

0 0 Sin( ) Cos ( )

Dr. Mohammad Tawfik
Element Stiffness Matrix in Global
Coordinates
1 1
Cos 2
( ) Sin( 2 ) Cos ( )
2
Sin( 2 )
2 2
1 1
EA 2 Sin( 2 ) Sin ( ) Sin( 2 ) Sin ( )
2 2
[K] = 1
2
1

h Cos 2 ( ) Sin( 2 ) Cos 2 ( ) Sin( 2 )
2 2
1 1
Sin( 2 ) Sin ( ) Sin( 2 ) Sin ( )
2 2
2 2

Dr. Mohammad Tawfik
Example: 4.6.1 pp. 196-201
Use the finite element analysis to find the
displacements of node C.

Dr. Mohammad Tawfik
Element Equations
1 0 1 0 0 0 0 0
0 0 1 0 1
0 0 0
[ ]
K1 =
EA
L 1 0 1 0
[ ]
K2 =
EA
L 0 0 0 0

0 0 0 0 0 1 0 1
0.3536 0.3536 0.3536 0.3536

0.3536
[ ]
K3 =
EA 0 .3536 0.3536
L 0.3536 0.3536 0.3536
0. 3536
0.3536

0.3536 0.3536 0.3536 0.3536

Dr. Mohammad Tawfik
Assembly Procedure
1.3536 0.3536 1 0 0.3536 0.3536

0.3536 0 .3536 0 0 0 .3536 0 .3536

EA 1 0 1 0 0 0
[K] =
L 0 0 0 1 0 1
0.3536 0.3536 0 0 0.3536 0.3536

0.3536 0.3536 0 1 0.3536 1.3536

Dr. Mohammad Tawfik
Global Force Vector
F1x F1x
F
1y
1 y F
Remember!
F2 x F2 x
{ F } = = NO distributed load
F F
2 y is applied to a truss
2 y
F3 x P

F3 y 2 P
Dr. Mohammad Tawfik
Boundary Conditions
U1 = V1 = U 2 = V2 = 0
Remove the corresponding rows and columns
EA 0.3536 0.3536 U 3 P
=
L 0.3536 1.3536 V3 2 P
Continue! (as before)
Dr. Mohammad Tawfik
Results
PL 3PL
U 3 = 5.828 , V3 =
EA EA
F1x = P, F1 y = P,
F2 x = 0, F2 y = 3P

Dr. Mohammad Tawfik
Postcomputation
e e P1e Ae Ee 1 1 u1e
P 1P 2
e= e
=
e
=
L 1 1 u 2
Ae Ae P 2
u1 Cos ( ) Sin( ) 0 0 u1

v1 Sin( ) Cos ( ) 0 0 v1
=
u 2 0 0 Cos ( ) Sin( ) u2
v 2 0 0

Sin( ) Cos ( ) v2


Dr. Mohammad Tawfik
Postcomputation
3P P
(1)
= 0, ( 2)
= , ( 3)
= 2
A A

Dr. Mohammad Tawfik
Summary
In this lecture we learned how to apply the
finite element modeling technique to bar
problems with general orientation in a
plain.

Dr. Mohammad Tawfik
Homework #5
Problem 4.27,
Due 13/12/2006 before 9:00am
Problem 4.44,
Due 20/12/2006 before 9:00am

Dr. Mohammad Tawfik
Announcements
Compensation Tutorial for E15:
Next Sunday 17/12/2006 3rd Period in H6
Next Lecture:
Wednesday 20/12/2006 3rd Period in H6
Next Quiz:
Wednesday 20/12/2006 3rd Period in H6
(This Lecture is included)

Dr. Mohammad Tawfik
Term Projects
A problem has got to be solved using the
finite element method
A report is going to be presented by each
group presenting the problem and its
solution

Dr. Mohammad Tawfik
The Report should contain:
Cover page
Project Title
Names of team members
Table of contents
Introduction and literature survey
Introduction to the problem
Historical background and relevance of the problem
Papers and books that presented the problem
Latest achievements in the problem

Dr. Mohammad Tawfik
The finite element derivation
Governing equation
Derivation of the element matrices
Using Glerkin method
Application of Symbolic manipulator to derive the
matrix equations will be appreciated
Solution procedure

Dr. Mohammad Tawfik
The numerical results and verification
Program results
Verification of results compared to published results
Parametric study
Discussion
Observations of the results
Further work that may be performed with the problem
Future developments of the model
References

Dr. Mohammad Tawfik
Evaluation
Report (50%)
Code (30%)
Structured: Functions built, easily modified
Readability: Organization, remarks
Length: The shorter the better
Results (20%)

Dr. Mohammad Tawfik
Projects
Heat transfer in a 2-D heat sink
2-D flow around a blunt body in a wind
tunnel
Vibration characteristics of a pipe with
internal fluid flow
Panel flutter of a beam
Rotating Timoshenko beam/blade

Dr. Mohammad Tawfik
Heat transfer in a 2-D heat sink
The heat sink will have heat flowing from
one side
Convection transfer on the surfaces
Different boundary conditions on the other
three sides
Plot contours of temperature distribution
with different boundary conditions

Dr. Mohammad Tawfik
2-D flow around a blunt body in a
wind tunnel
Potential flow in a duct
Rectangular body with different
Dimensions
Study the effect of the body size on the
flow speed on both sides
Plot contours of potential function,
pressure, and velocity potential

Dr. Mohammad Tawfik
Vibration characteristics of a pipe
with internal fluid flow
Study the change of the natural
frequencies with the flow speed under
different boundary conditions and fluid
density
Indicate the flow speeds at which
instabilities occur

Dr. Mohammad Tawfik
Panel flutter of a beam
A fixed-fixed beam is subjected to flow
over its surface
Plot the effect of the flow speed on the
natural frequencies of the beam
Indicate the speed at which instability
occurs

Dr. Mohammad Tawfik
Rotating Timoshenko beam/blade
Rotating beams undergo centrifugal
tension that results in the change of its
natural frequencies
Study the effect of rotation speed on the
beam natural frequencies and frequency
response to excitations at the root

Dr. Mohammad Tawfik
Teams
2-3 Students teams
Names and selected projects should be
submitted before 4PM on Thursday
21/12/2006

Dr. Mohammad Tawfik
Work Progress
A report should be submitted By 4PM every Wednesday
27/12/2006
The report should contain a preliminary literature survey
Problem statement
Governing equations
10/1/2007
The report should contain a deeper literature survey
The preliminary derivations of the finite element model
17/1/2007
A more mature version of the report should be presented
Preliminary results of the code
List of the program script should be included
24/1/2007
Final version of the report should be presented together with the code

Dr. Mohammad Tawfik
Beams and Frames

Dr. Mohammad Tawfik
Beams and Frames
Beams are the most-used structural
elements.
Many real structures may be
approximated as beam elements
Two main beam theories:
Euler-Bernoulli beam theory
Timoshenko beam theory

Dr. Mohammad Tawfik
Euler-Bernoulli Beam Theory
The main assumption in the Euler-
Bernoulli beam theory is that the beams
thickness is too small compared to the
beam length
That assumption resulted in that the sheer
deformation of the beam may be
neglected without much error in the
analysis
Dr. Mohammad Tawfik
Governing Equation
The equation governing the deformation of
and E-B beam under transverse loading
may be written in the form:
d
2
d w
2
EI ( x ) 2 = F ( x)
2
dx dx

Dr. Mohammad Tawfik
The Thin-Beam Elements
The thin beam element has a special feature,
namely, the two degrees of freedom at each
node are related.

Dr. Mohammad Tawfik
Beam Interpolation Function
w( x) = a1 + a2 x + a3 x + a4 x
2 3
w( x ) = 1 x x 2
x 3
{ a}
w( x ) = H ( x ) { a}
dw( x ) dH ( x )
= { a} = H x ( x ) { a} = 0 1 2 x 3x { a}
2
dx dx

Dr. Mohammad Tawfik
w( 0 ) = w1 = H ( 0 ) { a} w( l ) = w2 = H ( l ) { a}
w' ( 0 ) = w'1 = H x ( 0 ) { a} w' ( l ) = w'2 = H x ( l ) { a}
w1 H ( 0 ) a1
w' H ( 0 ) a
1 x 2
= = [T ]{ a}
w2 H ( l ) a3
w'2 H x ( l ) a4

Dr. Mohammad Tawfik
w1 1 0 0 0 a1
w' 0 1 0 0 a2
1
=
w2 1 l l 2 l 3 a3
w'2 0 2
1 2l 3l a4
1 0 0 0
a1 w1
a 0 1 0 0
2 3 2 3 1 w'1
= 2
a3 l l w2
2
l l
a4 2 1 2 1 w'
l 3 2 2
l2 l3 l
Dr. Mohammad Tawfik
{ a} = [T ] 1 { we }
w( x ) = H ( x ) { a} = H ( x ) [T ]
1
{w } = N ( x ) {w }
e e
4
w( x ) = N i ( x ) wi
i =1

Dr. Mohammad Tawfik
3x 2 2 x 3
1 2 + 3
l l
2 3
x 2 x x
+ 3

N ( x ) = { N ( x )} = 3 x 2 2 x 3
l l
T

l2 l3
x 2 x3
+ 2
l l

Dr. Mohammad Tawfik
Interpolation Functions
1
N1 N3
0.8
0.6
0.4
N(x)
0.2
N2
0
0 0.2 0.4 0.6 0.8 N4 1
-0.2
-0.4
X

Dr. Mohammad Tawfik
Beam Stiffness Matrix
The governing equation is:
d
2
d w2
EI ( x ) 2 = F ( x)
2
dx dx
Using the series solution

4
w( x ) = N i ( x ) wi
i =1

Dr. Mohammad Tawfik
The governing equation becomes
4
d
2
d Ni
2
EI ( x )
2 2
wi F ( x) = R ( x)
i =1 dx dx
Applying Galerkin method:

le le
4 d2 d 2 Ni
0 R ( x ) N j dx =
0 i =1 dx 2
EI ( x ) 2
w
i F ( x )
N j dx
dx
Dr. Mohammad Tawfik
Using integration by parts, twice, and
ignoring the boundary terms, we get:
le
4 d 2
N d 2
N
EI ( x )
0 i =1
j
i
wi F ( x ) N j dx = 0
dx 2
dx 2

In matrix form:
le le
EI ( x ){ N } N dx{ w } = F ( x){ N xx } dx
e
xx xx
0 0

Dr. Mohammad Tawfik
Use of Symbolic Manipulator
Beam Example

Dr. Mohammad Tawfik
Optional Homework #6
Derive the expression for the interpolation
function for a beam in terms of nodal
displacements and slopes.
Try to use a symbolic manipulator to
generate the expressions.
2 4
d w d w
A 2 + EI 4 = F ( x)
dt dx
Dr. Mohammad Tawfik
Two Dimensional Elements

Dr. Mohammad Tawfik
2-D Elements
In this section, we will be introduced to two
dimensional elements with single degree
of freedom per node.
Detailed attention will be paid to
rectangular elements.

Dr. Mohammad Tawfik
For the 2-D BV Problem
Lets consider a problem with a single
dependent variable
We may set one degree of freedom to
each node; say fi.
Further, lets only consider a rectangular
element that is aligned with the physical
coordinates

Dr. Mohammad Tawfik
A Rectangular Element
For the approximation
of a general function
f(x,y) over the element
you need a 2-D
interpolation function
f ( x, y ) = a1 + a2 x + a3 y + a4 xy

Dr. Mohammad Tawfik
Lets follow the same
procedure!

Dr. Mohammad Tawfik
2-D Interpolation Function
f ( x, y ) = a1 + a2 x + a3 y + a4 xy f ( x, y ) = H ( x, y ) { a}
f ( 0,0 ) = f1 = H ( 0,0 ) { a} f ( a,0 ) = f 2 = H ( 0, a ) { a}

f ( a, b ) = f 3 = H ( a, b ) { a} f ( 0, b ) = f 4 = H ( 0, b ) { a}
f1 H ( 0,0 ) a1
f H ( a ,0 ) a
2 2
= = [T ]{ a}
f 3 H ( a, b ) a3
f 4 H ( 0b ) a4
Dr. Mohammad Tawfik
f1 1 0 0 0 a1
f 1 a 0 0 a2
2
=
f 3 1 a b ab a3
f 4 1
0 b 0 a 4 1 0 0 0
a1 1 1
0 0
f1
a a a f
2 1 2
= 0 0
1
f
a3 b b 3
a4 1 1 1 1 f 4
ab ab ab ab
Dr. Mohammad Tawfik
{ }
f ( x, y ) = H ( x, y ) { a} = N ( x, y ) f e
x y xy
1 a b + ab
x xy

N ( x, y ) = { N ( x , y ) } = a ab
T
xy

ab
y xy
b ab

Dr. Mohammad Tawfik
How does this look like?

Dr. Mohammad Tawfik
2-D Interpolation Functions
1 1
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
N1 0.5 N2 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.9 0.9
0.1 0.1
0.6 0.6
0 y 0 y
0.3 0.3
0 0.1 0 0.1
0.2 0.3 0.2 0.3
0.4 0.5 0 0.4 0.5 0
0.6 0.7 0.6 0.7
0.8 0.9 0.8 0.9
x 1 x 1

Dr. Mohammad Tawfik
2-D Interpolation Functions
1 1
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
N3 0.5 N4 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.9 0.9
0.1 0.1
0.6 0.6
0 y 0 y
0.3 0.3
0 0.1 0 0.1
0.2 0.3 0.2 0.3
0.4 0.5 0 0.4 0.5 0
0.6 0.7 0.6 0.7
0.8 0.9 0.8 0.9
x 1 x 1

Dr. Mohammad Tawfik
Example: Laplace Equation
=0
2

2 2
+ 2 =0
x 2
y
= i N i ( x, y ) = N ( x, y ) { }
4
e
i =1

Dr. Mohammad Tawfik
Example: Laplace Equation
= i N i ( x, y ) = N ( x, y ) { e }
4
i =1
Applying the Galerkin method and integrating by parts,
the element equation becomes
[{ N } N x + { }
y ] { }
y =0
e
x N N dA
Area

Dr. Mohammad Tawfik
The Element Equaiton
2( a 2 + b 2 ) a 2 2b 2 a2 b2 2a 2 + b 2
2( a 2 + b 2 )
2 2
1 a 2b 2 2a 2 + b 2 a b e
2
{ } = { 0}
6ab a 2 b 2 2a 2 + b 2 2( a 2 + b 2 ) a 2b
2 2
2
2 a 2
+ b 2
a 2 b2 a 2 2b 2 2( a + b )
2

Dr. Mohammad Tawfik
The Logistic Problem!

Dr. Mohammad Tawfik
The Logistic Problem
In the 2-D problems, the numbering
scheme, usually, is not as straight forward
as the 1-D problem

Dr. Mohammad Tawfik
1-D Example
Element #1 is associated with nodes 1&2
Element #2 is associated with nodes 2&3, etc

Dr. Mohammad Tawfik
2-D Example

Dr. Mohammad Tawfik
2-D Example

Dr. Mohammad Tawfik
For Element #5
Local Node Number Global Node Number
1 5
2 6
3 9
4 8

Dr. Mohammad Tawfik
Contribution of element #5 to global
matrix
1 2 3 4 5 6 7 8 9 10 11 12
5 1,1 1,2 1,4 1,3
6 2,1 2,2 2,4 2,3
8 4,1 4,2 4,4 4,3
9 3,1 3,2 3,4 3,3
10
11
12

Dr. Mohammad Tawfik
A Solution for the Logistics
Problem
One solution of the logistic problem is to
keep a record of elements and the
mapping of the local numbering scheme to
the global numbering scheme in a table!

Dr. Mohammad Tawfik
Elements Register: Global
Numbering
Element Node Number
Number 1 2 3 4
1 1 2 5 4
2 4 5 8 7
3 7 8 11 10
4 2 3 6 5
5 5 6 9 8
6 8 9 12 11
Dr. Mohammad Tawfik
Algorithm for Assembling Global
Matrix
Create a square matrix A;
N*N (N=Number of nodes)
For the ith element 1 2 3 4 5 6 7 8 9 10 11 12
1
Get the element matrix B 2
Element Node Number
For the jth node 3Number
1 2 3 4
4
Get its global number k 5 1 1 2 5 4
For the mth node 6
7
2 4 5 8 7
3 7 8 11 10
Get its global number n 8
4 2 3 6 5
9
Let Akn=Akn+Bjm 10 5 5 6 9 8
Repeat for all m 11
12
6 8 9 12 11
Repeat for all j

Repeat for all i

Dr. Mohammad Tawfik

01introductiontothefiniteelementmethod Maged 100714091834 Phpapp02

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

01introductiontothefiniteelementmethod Maged 100714091834 Phpapp02

Uploaded by

Copyright:

Available Formats

Introduction to the Finite

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Solve the above system for the generalized

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Then, the approximate l2 f x

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Solve the above system for the generalized

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Solve the above system for the generalized

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

2nd order DEs in 1-D

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

The above equation is called the element

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

Introduction to the Finite Element Method

k12 k 22 u12 f12

Introduction to the Finite Element Method

1. Discretization: Divide the bar into N number of