CHAPTER 1 Introduction

1.1 General Overview

1.2 FEM History
1.3 Computer Programs
1.4 FEM vs FDM
1.5 Review of Matrix
 CHAPTER 1

1.1 General Overview (1)

 Method for obtaining solutions for engineering problems

- Experimental Method
- Analytical Method
- Numerical Method

 Numerical Tool
- FEM: Finite Element Method
- FDM: Finite Difference Method
- BEM: Boundary Element Method
…
 CHAPTER 1

1.1 General Overview (2)

 The finite element method is a powerful numerical method for
solving a variety of continuum mechanics and mathematical
physics.

 Useful for problems with complicated geometries, loadings, and

material properties where analytical solutions cannot be obtained.
 Discretizations
- The finite element method involves subdividing the continuum
body into an equivalent system of smaller bodies or units (finite
elements) interconnected at points common to two or more
elements (nodes or nodal point) and/or boundary lines and/or
surfaces.
 CHAPTER 1

1.1 General Overview (3)

Discretization Example (Bar Structure)

 CHAPTER 1

1.1 General Overview (4)

Solid Element

Discretization Example (2D Solid)

 CHAPTER 1

1.1 General Overview (5)

Discretization Example (3D Solid)

 CHAPTER 1

1.2 FEM History (1)

 Pre-FEM

- Hrenikoff [1941]: Lattice of 1D bars

- McHenry [1943]: Model 3D solids
- Courant [1943]: Variational form
- Argyris and Kelsey [1955]: Energy principle for matrix method
“Energy theorems and structural analysis,” Aircraft engineering, Vol. 26 & 27
 CHAPTER 1

1.2 FEM History (2)

 Beginning of FEM
- Grew out of aerospace industry
: Originated with an engineering group at Boeing Aircraft Corp., Seattle in about 1952.
- Post World War II jets, missiles, space flight
- Need for light weight structures
- Required accurate stress analysis
- Paralleled growth of digital computers
: Cannot be practically used without it…
- Turner, Clough, Martin, Topp [1956]: 2D elements
“Stiffness and deflection analysis of complex structures,” Journal of Aero. Science,
Vol. 23, pp. 805~824
- Clough [1960]: Term “Finite Elements”, Plane problems
 CHAPTER 1

1.2 FEM History (3)

- Zienkiewicz and Chung [1967]: The first book on finite elements
- In late 1960s and early 1970s, finite element analysis was
applied to nonlinear problems and large deformations.
- In late 1980s pre-processors and post-processors were
developed to reduce the time required for input data and
interpretation of analysis results.
- In 1990s softwares were connected to CAD softwares for design
purpose.
- In 2000s FEM programs in parallel processing environment are
used for large scale analysis.
- Today, powerful computers has brought this method within
reach of students and engineers working in small industries.
 CHAPTER 1

1.3 Computer Programs (1)

 Small special purpose program
(truss, frame, plane stress, heat transfer…)
 Large general purpose program (Commercial package)
- NASTRAN (General purpose, PC and workstation)
- ANSYS (General purpose, PC and workstation)
- ABAQUS (General purpose, especially nonlinear problem)
- Dyna-3D (Crash/impact analysis)
- LS-DYNA, CATIA, ADINA….
- HyperMesh (Pre/Post processor)
- PATRAN (Pre/Post processor)
- SPRC/I-DEAS (Complete CAD/CAM/CAE package)
 CHAPTER 1

1.3 Computer Programs (2)

 Three Stages in FEM Programs

- Pre-processing : Create a input file suitable for processing

- Processing : Solve & generate solutions for system equations

- Post-processing : Graphical display of processing results

 CHAPTER 1

1.3 Computer Programs (3) – Block Diagram

PREPROCESSOR

INPUT DATA
Control Data, Materials, Node and Element Element File
Definition, Boundary, Conditions, Loads Load File

FORM ELEMENT [k]

Read Element Data, Calculate Element
Element File
Stiffness Matrix, [k]

FORM SYSTEM [K]

Assemble Element[k]s to Form the System
Stiffness Matrix,[K]

APPLY DISPLACEMENT BOUNDARY CONDITIONS

COMPUTE DISPLACEMENTS
Solve the System Equations
[K]{D}={F} LOAD File
For the Displacements
{D}=[K]-1{F}

COMPUTE STRESSES
Calculate Stresses and Output Displacement,
Files for Postprocessor Plotting Stress Files

POSTPROCESSOR
 CHAPTER 1

1.3 Computer Programs (4) – The Analysis Step-by-Step

Define Analysis Program

Identify Key Variables

Perform Approximate
Engineering Analysis

no Is
finite element analysis
needed?
yes
Develop Conceptual Model

Choose The Computer Program

Prepare First Model

Run The Analysis

Evaluate Results

Prepare Refined Model

Run The Analysis

Evaluate Results

Is
further refinement yes
needed?
no
DEVELOP REPORT AND RECOMMENDATIONS
 CHAPTER 1

1.3 Computer Programs (5)

 Advantages of General Purpose Programs

- Easy input: Preprocessor (MSC/PATRAN, I-DEAS etc.)

- Solves many types of problem (general purpose)
- Modular design: Fluid, Dynamics, Heat etc.
- Can run on PC’s now !!

 Disadvantages of General Purpose Programs

- High development cost

- Less efficient than smaller programs
- Relatively high cost…
 CHAPTER 1

1.4 FEM vs. FDM (1)

 FEM vs. FDM
- While similar in some respects to the finite difference method, the finite
element method is sometimes more versatile and powerful for several
reasons:

a) Whereas the finite difference method can be used to calculate scalar

values, the finite element method can be used to calculate values of
either scalar or vector functions.
For example, the displacements in X, Y and Z directions in a solid body
subjected to external loads, temperature change, etc. can be calculated
simultaneously.

Scalar function T = T ( x, y , z )
Vector function
U ∉ V ( x, y, z )i + V ( x, y, z ) j + W ( x, y, z )k
 CHAPTER 1

1.4 FEM vs. FDM (2)

b) While only rectangular grids can be conveniently used with
the finite difference method, a wide variety of finite element shapes can
be used to subdivide odd shaped areas and volumes.
 CHAPTER 1

Reconsider... How Can Deal with Enormous Information?

Use Matrices in Solving Systems

Definition:
An n by m matrix is an array of objects arranged in n rows and m columns
 CHAPTER 1

1.5 Review of Matrix (1)

1. Determinants and Matrices
(1) Determinant – a SQUARE array of numbers which are
operated on to give ONE number.
A (n×n) determinant may be written as
col .1 col .2 col .n
↑ ↑ ↑
a11 a12  a1n → row 1
D = a 21 a 22  a 2 n → row 2
  
  
a n1 an 2 a nn → row n

= a11 ⋅ A11 − a21 ⋅ A21 + a31 ⋅ A31 −  +  (−1) n ⋅ an1 ⋅ An1

where Aij = the minors

 CHAPTER 1

1.5 Review of Matrix (2)

(2) Matrices – A (n×m) matrix is a RECTANGULAR (not
necessarily square) array of numbers. It cannot be
operated on to give one number.

 a11 a12 a13   a1n 

a
 21 a22 a23   a2 n  i = 1,2 , , m
   
[ ]
aij = 
 

j = 1,2 , , n
m = row number
     
n = column number
     
 
am1 am 2    amn 
 CHAPTER 1

1.5 Review of Matrix (3)

Forms of matrices :
(a) Row Matrix – a matrix consisting of only a single row. (i.e.
m=1)
R = [r1 r12 r13  r1n ] or [r1 r2  rn ]
(b) Column Matrix (Vector) – a matrix consisting of only a single
column. (i.e. n=1)
 C11   C1 
C  C 
C =  21  =  2 
     
   
C m1  C m 
(c) Square matrix – a matrix with the same number of rows as
columns. (m=n).
 K 11 K 12  K 1n 
K K 22  K 2 n 
K =  21
    
 
 K n1 K n2  K nn 
 CHAPTER 1

1.5 Review of Matrix (4)

(d) Diagonal Matrix – a square matrix with zeros everywhere
except on the main diagonal. The main diagonal is the one
running from upper left to lower right. The elements on the
main diagonal are d11 , d 22 , etc.
d11 0 0 
D =  0 d 22 0 
 0 0 d 33 

(e) Unit (Identity) Matrix – (Usually designated I) – a diagonal

matrix with ones on the main diagonal. The unit matrix fulfills
the same function for matrix arithmetic that the number 1 fulfills
for scalar arithmetic.
1 0 0 0
0 1 0 0 
I = 
0 0 1 0 
 
 0 0 0 1 
 CHAPTER 1

1.5 Review of Matrix (5)

Example
Scalar Matrix

a ×1 = 1× a = a AI = I × A = A

(f) Symmetric Matrix – a square matrix that is symmetrical

about the main diagonal for a symmetric matrix A, aij = a ji .

(g) Anti Symmetric (Skew Symmetric) Matrix – a square matrix

for which corresponding elements on opposite sides of the main
diagonal have equal magnitude but opposite signs. For a skew
symmetric matrix A. a ji = − aij .
 CHAPTER 1

1.5 Review of Matrix (6)

Example

 0 − 3 2 − 4 1 3 5
3 
A=
0 1 3  S = 3 2 6
− 2 − 1 0 − 5 5 6 0
 
 4 − 3 5 0 

Anti symmetric Symmetric

Note that since for any element, aij = − a ji , the main diagonal
elements of an anti symmetric matrix are zero.
 CHAPTER 1

1.5 Review of Matrix (7)

2. Matrix algebra
(1) Addition or subtraction of 2 m × n matrices:
 a11 ± b11 a12 ± b12a1n ± b1n 
 
a ± b
 21 21   a 2 n ± b2 n 
a 22 ± b22
(a mn ) ± (bmn ) =     
 
    
a m1 ± bm1   a mn ± bmn 
a m 2 ± bm 2
NOTE: You cannot add or subtract two determinants in the
above manner.
(2) Multiplication by a scalar factor α ( α=real number )
αa11 αa12 αa13  αa1n 
αa αa   αa 
 21 22 2 n 
α ⋅ (a mn ) =      
 
     
αa m1 αa m 2   αa mn 
 CHAPTER 1

1.5 Review of Matrix (8)

(3) Product of two matrices
 a11 a12   a1n   b11 b12   b1m 
a a22   a2 n  b21 b22   b2 m 
 21
(amn )(bnm ) =     ⋅    
   
         
a m1   a mn  bn1 bn 2   bnn 
 am 2

 a11b11 + a12b21 +  + a1nbn1 a11b12 + a12b22 + a13b32 +  + a1nbn1  a11b1m + a12b2 m +  + a1nbnm 
 a b + a b ++ a b a21b12 + a22b22 + a23b32 +  + a2 nbn 2  a21b1m + a22b2 m +  + a2 nbnm 
=  21 11 22 21 2 n n1

    
 
am1b11 + am 2b21 +  + amnbn1 am1b12 + am 2b22 + am 3b32 +  + amnbn 2 am1b1m + am 2b2 m +  + amnbnm 
 CHAPTER 1

1.5 Review of Matrix (9)

The general rule of multiplication of matrices:

(a )(b ) = (c )
ij ji ij
n
where cij = ∑ air brj
r =1

n=total number of columns

NOTE: The multiplication of two matrices is possible only if the

number of columns of the first matrix is equal to the number
of rows of the second. Matrices that satisfy this condition are
called “conformable matrices”. A computer program is provided
below for the multiplication of matrices.
 CHAPTER 1

1.5 Review of Matrix (10)

MATrix LABoratory

MATLAB is an interactive software which has been used recently

in various area of engineering and scientific applications.

% Purpose: Execute a matrix multiplication

% Create the first matrix
>> MatrA = [1 2 3; 4 5 6];
% Create the second matrix
>> MatrB = [1 2 3 4; 5 6 7 8; 9 10 11 12];
% Assign matrix multiplication result to a variable
>> MatrR = MatrA * MatrB;
 CHAPTER 1

1.5 Review of Matrix (11)

Example (Multiplication of matrices):
(a) (square matrix) × (column matrix) = (column matrix):

a1 a2 a3   x1  a1 x1 + a 2 x 2 + a3 x3 
b
 1 b2 b3  ⋅  x 2  =  b1 x1 + b2 x 2 + b3 x3 
 c1 c2 c3   x3   c1 x1 + c 2 x 2 + c3 x3 

(b) (row matrix) × (square matrix) = (row matrix):

a1 a2 a3 
(x1 x2 x3 ) ⋅ b1 b2 b3  =
 c1 c2 c3 
(x1a1 + x2b1 + x3c1 x1a2 + x2b2 + x3c2 x1a3 + x2b3 + x3c3 )
 CHAPTER 1

1.5 Review of Matrix (12)

(c) (row matrix) × (column matrix) = scalar:
 x1 
 
(a1 a2 a3 ) ⋅  x 2  = a1 x1 + a 2 x 2 + a3 x3
x 
 3
3. Matrix inversion (Division of matrices)
Before we can invert a matrix, the following terminologies
will have to be mentioned first:
(1) unit matrix, I:
1 0 0 0
0 1 0 0
I =
0 0 1 0
 
0 0 0 1

I × (aij ) = (aij )
 CHAPTER 1

1.5 Review of Matrix (13)

(2) Inverse matrix:
[ ]
Let A = aij i, j = 1,2,3,  , n
Then A × A−1 = I
Where A −1 = inverse matrix

(3) Minors of a matrix, Aij

Given:

 a11 a12 a13  a1n 

a
 21 a22 a23  a2 n 
[ ]
aij =  a31 a32 a33 a3n 
 
     
am1 am 2 am 3  amn 
 CHAPTER 1

1.5 Review of Matrix (14)

the co-factor, Aij are:
a 22 a 23 a 24  a2n a 21 a 23 a 24  a2n
a32 a33 a34  a3n a31 a33 a34  a3n
A11 = a 42 a 43 a 44 a4n A12 = − a 41 a 43 a 44 a4n
       
am 2 a m3 a m 4  a mn a m1 a m3 a m 4  a mn

Aij = (− 1)
i+ j
M ij
(4) Transposition of Matrix:
By interchanging of rows and columns in a SQUARE matrix, e.g. :
Given:
 a11 a12 a13  a1n 
a a 22 a 23  a 2 n 
 21
[aij ] =  a31 a32 a33 a3 n 
 
     
 a n1 a n3  a nn 
 an 2
 CHAPTER 1

1.5 Review of Matrix (15)

the transposed matrix of aij [ ] is
 a11 a 21 a31  a n1 
a
 12 a 22 a 23  a n 2 
[a ]
ij
T
=  a13 a 23 a33 an3 
 
     
a1n a2n a3n  a nn 

NOTE: the elements on the principal diagonal remain

unchanged.

(5) inversion of a SQUARE matrix:

Step 1: Let A = determinant formed by the element
of a matrix [aij ] :
 CHAPTER 1

1.5 Review of Matrix (16)

 a11 a12 a13  a1n 
a
 21 a22 a23  a2 n 
A = a31 a32 a33 a3n 
 
     
an1 an 2 an 3  ann 

Note: A cannot be zero.

[ ]
Step 2: Let Aij = co-factors of aij and form the following
matrix:
 A11 A12 A13  A1n 
A A22 A23  A2 n 
 21
Aij =  A31 A32 A33 A3n 
 
     
 An1 An3  Ann 
 An 2
 CHAPTER 1

1.5 Review of Matrix (17)

Step 3: Transpose the elements of [A ]
ij :
 A11 A21 A31  An1 
A A22 A32  An 2 
T  12
Aij =  A13 A23 A33 An3 
 
     
 A1n A3n  Ann 
 A2 n
T
Step 4: divide each element of Aij by A :
 A11 A21 A31 An1 

 A A A A 
 
 A12 A22 A32  An 2 
 A A A A 
[ ]
−1 
A −1 = aij =  A13 A23 A33

An 3 
 A A A A 
     
 
 A1n A2 n A3 n Ann 

 A A A A 
 CHAPTER 1

1.5 Review of Matrix (18)

Example
Find the inverse of the following matrix:

1 2 3
[ ]
A = aij = 1 2 4
1 3 5

Step 1:

1 2 3
2 4 2 3 2 3
A = 1 2 4 = 1⋅ − 1⋅ + 1⋅
3 5 3 5 2 4
1 3 5

= (10 − 12 ) − (10 − 9 ) + (8 − 6 ) = −1
 CHAPTER 1

1.5 Review of Matrix (19)

Step 2: the minors of [aij ] are:
2 4 1 4 1 2
A11 = = −2 A12 = − = −1 A13 = =1
3 5 1 5 1 3

2 3 1 3 1 2
A21 = − = −1 A22 = =2 A23 = − = −1
3 5 1 5 1 3

2 3 1 3 1 2
A31 = =2 A32 = − = −1 A33 = =0
2 4 1 4 1 2
− 2 − 1 1 
[ ]
∴ Aij =  − 1 2 − 1
 2 − 1 0 
Step 3:
− 2 − 1 2 
[Aij ]T =  − 1 2 − 1
 1 − 1 0 
 CHAPTER 1

1.5 Review of Matrix (20)

Step 4:
2 1 − 2
[ ]−1 = 1A [Aij ]T
A −1 = aij =  1 − 2 1 
− 1 1 0 

Check if : [aij ]⋅ [aij ]−1 = I

1 2 3  2 1 − 2
[aij ]⋅ [aij ]−1 = 1 2 4 ⋅  1 − 2 1 
1 3 5 − 1 1 0 

1× 2 + 2 × 1 − 3 × 1 1× 1 − 2 × 2 + 3 × 1 − 1× 2 + 2 × 1 + 3 × 0  1 0 0
= 1× 2 + 2 × 1 − 4 × 1 1× 1 − 2 × 2 + 4 × 1 − 1× 2 + 2 × 1 + 4 × 0 = 0 1 0 = I
1× 2 + 3 × 1 − 5 × 1 1× 1 − 3 × 2 + 5 × 1 − 1× 2 + 3 × 1 + 5 × 0  0 0 1

CHECKED!
 CHAPTER 1

1.5 Review of Matrix (21)

MATrix LABoratory

% Purpose: Calculate determinant and inverse matrix

% Create the matrix
>> Mysquare = [1 -3 2; 1 0 1; 0 1 -3];
% Calculate the determinant
>> det(Mysquare);
% Calculate the inverse matrix
>> inv(Mysquare);
 CHAPTER 1

1.6 Equations of Motion (1)

 CHAPTER 1

1.6 Equations of Motion (2)

 CHAPTER 1

1.6 Equations of Motion (3)