Islamic University of Technology Bangladesh University of

Engineering and Technology

(BUET)

3.00 credits, 3 hrs./week

Prof. Dr. A.F.M. Saiful Amin

Dept. of Civil Engineering
BUET, Dhaka-1000, Bangladesh
 Course

1 Introduction to Finite Element Method 8 Bars and Trusses

Recall 2 Basic Equations in Elasticity Beams and Rigid Frames

from
9
past
3 Matrix Displacement Formulation 10 Discretization of Structures
studies

4 11 Numerical Integration

5 Shape Function 12 Solution Techniques and Convergence

Almost CEE 4711
Stiffness Method
6 Spring Element 13 Basic Finite Element Analysis Techniques
for Dynamic Calculation
7 Iso-parametric Formulation 14 Time Dependent Flow Problem
Module 4
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

In this module, the following topics are discussed:

• Element shapes • Types of nodes
• Order of the element • Types of nodal unknowns
One
• Global and local coordinate • Natural coordinate c
dimensional
systems systems elements

Before taking up mathematical aspect of finite element

analysis, these preliminaries are to be understood.
c
Three c
Two
dimensional dimensional

ELEMENT elements elements

Based on the shapes elements can be classified as

I. One dimensional elements
Shape
II. Two dimensional elements
Functions
III. Three dimensional elements.
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements

1 element These elements are suitable for the analysis of

x one dimensional problem and may be called as
1 2 line elements also.
2 elements
x Figure shows different types of one
1 3 2 dimensional elements.
3 elements
x
1 3 4 2
n elements
x
1 3 4 n −1 n 2
One dimensional elements
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements 2
1
Much less DOF,
5 Less computation time

3
3 4
▪ We need two dimensional elements to solve two dimensional
problems. Common two-dimensional problems in stress analysis are
plane stress, plane strain and plate problems.
▪ Two dimensional elements often used is three noded triangular
element .
▪ It has the distinction of being the first and most used element. These
elements are known as Constant Strain Triangles (CST) or Linear
Displacement Triangles.
1 2

Constant strain triangle

 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
▪ Six noded and ten noded triangular
elements are also used by the analysts.
Mid side or Inside
3 3 What are the
element node exist ▪ Six noded triangular element is known as
to better model the purposes of 5 & 10
curved geometry. no. nodes at Fig. (a) Linear Strain Triangle (LST) or as
8 7 and (b) ? Quadratic Displacement Triangle.
Also can model the
6 5 inertia effect!!
9 10 6 ▪ Ten noded triangular elements are known
as Quadratic Strain Triangles (QST) or
Cubic Displacement Triangles.
1 4 2 1 4 5 2
(a) (b) ▪ One can think of trying the use of still
(a) Linear strain triangle (b) Quadratic strain triangle higher order triangular elements like
Cubic Strain Triangles and Quartic Strain
▪ Stress Triangles.
▪ Strain
Collected from nodes
▪ Deformation
▪ Displacement
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
3 3 y
4 3
10 9 12 11
16
13
11 13 10
8
14 17 19
9
12 7 15
15 18 20 21 8
14
1 2
1 4 5 6 2 1 4 5 6 7 2
o x
(a) (b)

Cubic strain triangle Quadratic strain triangle 4 noded rectangular element

(15 noded) (21 noded)

▪ A simple but less used two dimensional element is the four noded rectangular element whose sides are parallel to the global
coordinate systems.
▪ This systems is easy to construct automatically but it is not well suited to approximate inclined boundaries.
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
Rectangular elements of higher order also can be used. Figure shows a family of Lagrange rectangle in which nodes are in the form
of grid points. Figure shows the family of Serendipity rectangles which are having nodes only along the external boundaries.

Lagrange family rectangular elements Serendipity family rectangular elements

 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
▪ Quadrilateral Elements are also used in finite element analysis.
▪ Initially quadrilateral elements were developed by combining triangular elements. But it has taken backstage after isoparametric
concept was developed.
▪ Iso-parametric concept is based on using same functions for defining geometries and nodal unknowns.
▪ Even higher order triangular elements may be used to generate quadrilateral elements.

y Idealization
3 y y y
3
4 4

2
2
1 1
o x x x x
Quadrilateral element Quadrilateral elements generated using triangular elements

Hyper parametric elements were first used. Simpler methods came later! Mesh
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
y y
Using iso-parametric concept even curved
elements are developed to take care of
Benefits boundaries with curved shapes.

x x
y y We can describe the curved geometry and
deflection using the same elements!!
Curved
element
x x
Curved two dimensional elements
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
These are also known as ring type elements. These elements are useful for the analysis of axi-symmetric problems such as analysis
of cylindrical storage tanks, shafts, rocket nozzles.
Axi-symmetric elements can be constructed from one or two dimensional elements.
• One dimensional axi-symmetric element is a conical frustum
• Two dimensional axi-symmetric element is a ring with a triangular or quadrilateral cross section.

A Rocket Nozzle
Axi-symmetric elements
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
The tetrahedron is the basic element for three-dimensional problems, analogous to the triangle in two-dimensional problems.
• Tetrahedron is having four nodes, one at each corner.
• In three-dimensional analysis, there are two types of elements with eight nodes: the general hexahedron and the rectangular
prism, which is a specific case of a hexahedron often referred to as a brick element.
• Higher order elements can also be used in these cases.
8
4 8
7

4 4
3
3 7

5 3
5
6
6
1 2
1 (d )
(a) 1 2 2
(b) (c )
(a) Tetrahedron element (b) Rectangular prism (brick) element (c) Arbitrary hexahedron element (d) Three dimensional
quadratic element
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
Element Geometry Preferences

2D Element

Other things being equal, prefer

in 2D: Quadrilaterals over Triangles

in 3D: Bricks over Wedges

Quadrilateral Triangular
Wedges over Tetrahedra
9 5
Explain graphically – How?
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements 4 nodes

Dividing a cube into five tetrahedra's

Wedges over Tetrahedra

Middle

5 – tetrahedra ≡ 1 cube
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Elements
Triangles over tetrahedra

Design a gusset plate

 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

• Selected finite points

• To be determined in the finite element analysis
• There are two types of nodes viz.
A) external nodes connected to another element or a boundary 1 3 4 2
B) internal nodes within the element (a)

3
A. External nodes
-Occur on the edges/ surface of the elements 4 7 3
-May be common to two or more elements. 8 7
-Nodes, 1 and 2 in one dimensional element, nodes 1 to 9 in 10
noded triangular element and nodes 1 to 8 in 9 noded lagrangian 8 9 6
9 10 6
element are external nodes.
B. Internal nodes
-Occur inside an element.
1 4 5 2 1 5 2
-There will not be any other element connecting to this node. (b) (c )
-To satisfy the requirement of geometric isotropy while
choosing interpolation functions (a) 4 noded one dimensional element (b) 10 noded triangular element (c) 9
noded Lagrange element

In CE 411, only extreme boundaries are considered. We can

consider other intermediate boundaries by shape function.
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Nodes may also be classified as (i) Primary nodes and (ii) 1,2 - Primary Nodes
3 4 3,4 - Internal Nodes
Secondary nodes. 1 2
(a)
1,2,3,4 - Primary Nodes
I. Primary nodes (external nodes) occur at the ends of 3 1,2,3 - Primary Nodes 5,6,7,8 - Secondary Nodes
4,5,6,7,8,9 - Secondary Nodes 9 - Internal Node
one dimensional elements or at the corners in the 10 - Internal Node 4 7 3
two or three dimensional elements. 8 7

II. Secondary nodes (internal nodes) occur along the 9 10 6 8 9 6

side of an element but not at corners.

1 4 5 2 1 5 2
(b) (c )
In flexibility method / Stiffness method / Moment
Distribution method, only primary or External nodes (a) 4 noded one dimensional element (b) 10 noded triangular element (c) 9 noded
Lagrange element
are considered.
Depends on
provisions
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Unknowns
• Basic unknowns may be displacements (stress analysis), temperatures (heat flow problems) and the potentials (fluid flow or
the magnetic field) .
• In Truss analysis, satisfying the continuity of only displacements are satisfied enough since no change in the slopes at any nodal
point. These are called ‘zeroth’ continuity problems, indicated as C0-continuity problem.
• In case of beams and plates, not only the continuity of displacements, but the slope continuity also should be ensured. Since the
slope is the first derivative of displacement, this type of problems are classified as ‘First order continuity problems and are
denoted as C1 – continuity problems.

 2w 
In exact plate bending analysis even second order
  continuity should be ensured. Hence the actual nodal unknowns in
these problems are,  xy 
w w  2 w
w, , ,
x y xy
where w is displacement. Such problems are classified as C2– continuity problems. In general Cr continuity problems are those in
which nodal unknowns are to be basic unknowns and up to rth derivatives of the basic unknowns.
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Unknowns
Continuity of displacement:

• Most occurred and verified in a stress test

• If bounding condition is not met or encountered, stress
usually flows and there’s no abrupt changes.

Displacement is continuous Discontinuity is not

(Co continuous) expected

Impossible
Co continuity
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Unknowns
Continuity of slope:
Plate

Slope Continuity
(C1 continuity)
Shell

In general:
Displacement
C0

Continuity

Slope C1
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Systems
Introduced by FE
The following terms are commonly referred in FEM: developers

Global Local Natural

coordinates coordinates coordinates

Traditional coordinate system New discovery in FEM

However, there is another term ‘generalized coordinates’ used for defining a polynomial
form of interpolation function. This has nothing to do with the ‘coordinates’ term used
here to define the location of points in the element.

▪ CE 4711
▪ Also the previous modules
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Coordinates

The coordinate system used to define the points in the entire structure is called global coordinate system.
▪ We analyze in global coordinate
▪ We design in local coordinate
y 2 (6) 3 (8) 4

1 (1) 2 (2) 3 (1) (5) (7) ?

(12)
x (3) (9) (11)
x1 x2
(a) 5
(2) (4) (13)
1 x
6 7 8
(b) C T
Cross
Global coordinate system
section

A1 A2
C T
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Coordinates
x1 y

2
y1
For the convenience of deriving element (1)
(2)
properties, in FEM many times for each element x1
a separate coordinate system is used 1 6
x1
1
However the final equations are to be formed in
the common coordinate system i.e. global 2 y1
coordinate system only.
(3)

(4)
x1
y1 6 7
6

Local coordinate system

 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Coordinates
A natural coordinate system is a coordinate system which permits the specification of a point within the
element by a set of dimensionless numbers, whose magnitude never exceeds unity.

Derivation:
It is obtained by assigning weightages to the nodal coordinates in defining the coordinate of any point inside the
element. Hence such system has the property that ith coordinate has unit value at node i of the element and zero
value at all other nodes.

Advantages:
The use of natural coordinate system is advantages in assembling element properties (stiffness matrices), since
closed form integrations formulae are available when the expressions are in natural coordinate systems.

Natural coordinate systems for one dimensional, two dimensional and three dimensional elements are discussed
in the following slides.
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Coordinates
y
Inspired by nature

A (L2, L3)
L3
L2
(x,y) ← Natural
coordinate

Natural coordinates are similar to

Natural coordinates we finding the way on the street!!
use everyday:
How about a google
map?
x
How can birds find ways over long distance? Perhaps they never follow local street signs?
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Coordinates
• Route from BUET to New Market
y

More Practical Solution!

x Cartesian Co-ordinate is of no use !!

 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Natural Global XYZ

Local xyz

In one dimensional problem, the following type of natural coordinate is 1 C P 2

  x
x1 x1 x2 x x2
also used. The natural coordinator ξ for any point in the element is 2
PC 1 0 1
=
defined as  x2 − x1  where P is the point referred and C is the Natural Coordinate ξ
 
 2 
center point of nodes 1 and 2.
1
1
0 2
x≡ 𝜉 −1
y≡η
z≡ζ
Variation of ξ
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Relation Between
η Natural Coordinate

We need to have a conversion system to move between natural coordinate and global coordinate
Y
(1,1)
(0,1) 𝜉
≠ 90o (1,0)
J ≡ Jacobian Matrix
(0,0) x
It is now possible: New quantity that relates
η x≡ 𝜉 between Cartesian and
y≡η natural coordinates
z≡ζ

𝜉
Natural coordinate: does not need to be mutually perpendicular.
≠ 90o

Not necessarily
a right angle
Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Natural

Geocoordinates are not required to be

Mutually perpendicular

Great Circular Arc N

We follow the nature

≠ 90o
LHR London [Heathrow Airport]
NRT Tokyo (Narita)
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Natural
The relationship between natural coordinate ξ and Cartesian coordinate x can be easily expressed as:

PC x − xc
= = , where l is length of the element = x2 − x1
1 1
 
2 2
2 x −x  2 x − x + 2 x1 
= x− 2 1  = x− 2 1 
l 2  l 2 
2 l + 2 x1  2 l
= x−  = x − x −
l 2  l  1
2 
l l
or ,  = x − x1 −
2 2
l
 (1 +  ) = x − x1
2
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Natural
It may be noted that, at node 1 where x = x1,

l
(1 +  ) = x − x1 = 0
2
 = −1

and at node 2, where x = x2, we get

l
(1 +  ) = x2 − x1 = l
2
  =1
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

Natural
The integration formula for integration over entire length is: 1
1
0 2
−1
x2 1
1 2
  pdx =   p d since dx = d (refer equation 4.5)
x1 −1
2 l
Variation of ξ
l 1 1
=  p +1 
2 p +1 −1

= 0, if p is odd
l 1 1
= (2) = , if p is even
2 p +1 p +1
 Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET) Module 4

1. Explanation of the following terms

I. Nodes, primary nodes, secondary nodes and internal nodes
II. Local coordinates, global coordinates, natural coordinates and area coordinates.
III. Higher order elements and lower order elements.

2. Explanation of the terms

I. Constant strain triangle (CST)
II. Linear strain triangle(LST) and
III. Quadratic strain triangles (QST).

3. Derivation of the expressions for natural coordinates for a two noded element
I. In terms of L1 and L2, when range is 0 to 1.
II. In terms of ξ , when range is –1 to 1.

4. Derivation of expressions for natural coordinates in a CST element. Show that they are nothing but area coordinates.
Prof. Saiful Amin, Bangladesh University of Engineering & Technology (BUET)

THANK YOU

35