Chapter 5

Types of Elements & Shape Functions

Hari R. Parajuli
Dr. of Eng., Post Doctorate
Lecturer of Civil Engineering
Types of Finite Elements
 Types of Finite Elements
1D -Spring element
-Bar Element
2D -Truss element
-Beam Element
-Triangular Elements (CST, LST, QST)
-Rectangular/Quadrilateral Element
(Linear, quadratic, cubic etc.)
Lagrangian Element
Serendipidity Element

3D -Truss Element
-Beam element
-Axisymmetric Element
-Tetrahedral Element
-Hexahedral/Brick Element
 Line Elements

Spring element
Bar/Beam Element

Bar in 2D Bar in 3D
Beam Elements

2D

3D
 Plane Elements

Constant strain Triangle Linear strain Triangle

Linear quadrilateral Quadratic quadrilateral

 Higher Order Plane Elements

Lagrangian Element Serendipidity Elements

3D Solid Elements
 Coordinate system

-Global Coordinate system

-Local Coordinate system
-Natural coordinate system
 Natural Coordinate system
A coordinate system that specifies a point in the element by set of dimensionless
parameters. Magnitude of the dimensionless parameters never exceeds unity.
Natural Coordinate system
Natural Coordinate system
Natural Coordinate system
Natural Coordinate system
Isoparametric Elements
Isoparametric Elements
 Shape/Interpolation Function
What is shape/Interpolation Function?
A function which relates the field variable at any point within
the element to the field variables of the nodal points is called
shape function. It is also called interpolation function.
• The values of the field variable computed at the nodes are used to
approximate the values at non-nodal points (that is, in the element interior)
by interpolation of the nodal values.
• For the three-node triangle example, the field variable is described by the
approximate relation φ(x, y) = N1(x, y) φ1 + N2(x, y) φ2 + N3(x, y) φ3 where
φ1, φ2, and φ3 are the values of the field variable at the nodes, and N1, N2,
and N3 are the interpolation functions, also known as shape functions or
blending functions.
• In the finite element approach, the nodal values of the field variable are
treated as unknown constants that are to be determined. The interpolation
functions are most often polynomial forms of the independent variables,
derived to satisfy certain required conditions at the nodes.
• The interpolation functions are predetermined, known functions of the
independent variables; and these functions describe the variation of the
field variable within the finite element.
 Interpolation Functions

Linear Quadratic
 Refinement

h-refinement: In this case, the local density of elements is increased that introduces a
larger number of discrete points in the areas of interest but the order of the interpolation
function remains unchanged. This version is often used due to the simpler approach but
especially ,in three dimensions, the higher number of discrete points is a serious drawback
p-refinement: Here, the order of the interpolation function is increased so that the
characteristics of approximation of the element itself are better. This often reduces the
amount of discrete points that are needed to represent a given function within a tolerable
error level. Nevertheless, there are limits due to oscillating results in case of higher orders.
Therefore shape functions up to a maximum order of three are normally used. For further
increase of accuracy the h-refinement seems to be the better approach.
 Shape/Interpolation Function
Let the displacement field in 2D is represented by u and Ni is shape
function. Here ui, and vi are local nodal displacements. As an example a
displacement field for a triangle can be represented by

u = ∑ Niui v = ∑ Ni vi v3

u = N1u1 + N2u2 + N3u3 3

u3
v1
v = N1v1 + N2v2 + N3v3 u1
1
In matrix form v2
 u1 
v  2 u2
 1
u   N1 0 N2 0 N3 0   u 2 
 = 0  
v   N1 0 N2 0 N 3   v 2 
u3 
 
 v3 
 Polynomial Function

Degree of polynomial Pascal Triangle

Polynomial Function
Derivation of Shape Functions
 Beam Element
Beam element requires satisfying displacement and rotation. This condition is
called C1 continuity. Let v1 and θ1 be the displacement and rotation at node 1
and v2 and θ2 be the displacement and rotation at node 2. Since it has four
nodal displacements, the polynomial should of four coefficients as below

v = α1 +α2 x +α3x2 +α4 x3

∂v
θ = = α2 + 2α3 x + 3α4 x2
∂x
∂v1 ∂v2  v1 
θ1 = θ2 =
∂x ∂x θ 
 
{δ }e =  1
 v2 
θ 2 

Local displacements in
matrix form
 Beam Element
Back substituting  1 0 0 0
 0   v1 
α 1   1 0 0  
α  Or, 3  3  θ1 
{δ } = 1 x x 2  
x 3   2 
{δ } = 1 x x2 
x  − −
2 3 1
−  
 l
2
l l 2
l  v2 
α 3   2 1 2 1  θ2 
α 4   l 3 −
l2 l 3 l 2 
v1 
3  
 3x2 2x3 2x2 x3 2
3x 2x 3
x x θ1 
2
{δ} = 1− 2 + 3 x− + 2 2
− 3 − + 2   {δ } = [ N1 N2 N3 N4 ]{δ }e
 l l l l l l l l v2 
θ2 

Thus,
3x 2 2 x3 2x2 x3 3x 2 2 x3 x 2 x3
N1 = 1 − 2 + 3 N2 = x − + 2 N3 = 2 − 3 N4 = − + 2
l l l l l l l l
 Beam Element

In matrix form −1
α1  1 0 0 0   v1 
 v1  1 0 0  α1 
0 α  0  θ 
Or,  2  1 0 0   1
θ  0 1 0 0  α 2   =  
 1  α3  1 l l2 l3 v2 
 = 3  
v2  1 l l 2
l α3 
 α4  0 
1 2l 3l 2  θ2 
θ2  0 
1 2l 3l 2  α 4 

 1 0 0 0
α1   0 0   v1 
T
α1  l 4 0 −3l 2 2l  v1  1 0

α   2    Or, α  θ 
 2 1 0 l4 −2l2 l  θ1   2  3 2 3 1  1 
 = 4 4      = − 2 − −  
α3  3l − 2l  0 0 3l2 −2l  v2  α 3   l l l2 l   v2 
α4  2    α 4   2 1 2 1  θ 2 
 0 0 −l3 θ
l   2   l 3 − 3
l2 l l 2 
 Beam Element
Plot shapes

N1

N2

N3

N4
 Plane Element-2D CST
v3

u3
Local displacement coordinates is defined by 3
v1
{δ }e = {u1 u2 u3 v1 v2 v3}
T

u1
Linear displacement field is given by 1

u = α1 +α2 x +α3 y v2
CST
u2
u = α4 +α5 x +α6 y 2

Coordinates of vertices of a triangle are (x1, y1), (x2, y2) and (x3, y3)
and nodal displacement

u1 = α1 +α2 x1 +α3 y1
u2 = α1 +α2 x2 +α3 y2
u3 = α1 +α2 x3 +α3 y3
 Plane Element-2D CST

In matrix form Here,

−1
u1  1 x1 y1 α1  α1  1 x1 y1  u1  1 x1 y1 1 1 1
      Or,     u  1 x y = x x x =2A
u2  = 1 x2 y2  α2  α =
 2  2 2  2
1 x y 2 2 1 2 3
u  1 x y  α  α  1 x y  u  1 x3 y3 y1 y2 y3
 3  3 3  3   3  3 3  3
T
α1  x2 y3 − x3 y2 y2 − y3 x3 − x2  u1  Where
  1     a1 = x2 y3 − x3 y2 b1 = y2 − y3
Or, α2  = x3 y1 − x1 y3 y3 − y1 x1 − x3  u2 

α  2A  x y − x y y − y x − x  u  a2 = x3 y1 − x1 y3 b2 = y3 − y1
 3  1 2 2 1 1 2 2 1   3  a = x y − x y b3 = y1 − y2
3 1 2 2 1

T
α1   a1 b1 c1  u1  a1 a2 a3  u1 
Or,
  1     1     c1 = x3 − x2
α2  =  a2 b2 c2  u2  =  b1 b2 b3  u2  c2 = x1 − x3
α  2 A a b c  u  2 A c c c  u  c3 = x2 − x1
 3  3 3 3  3  1 2 3 3
 Plane Element-2D CST

Back substituting

α1  a1 a2 a3 u1 

  1    
{δ} = [1 x y] α2  = [1 x y] b1 b2 b3  u2 
α  2A u 
 3 c c c 
 1 2 3  3 
Or,
u1 
 a1 + b1x + c1 y a2 + b2 x + c2 y a3 + b3 x + c3 y   
{} 
δ =  u2  Where
 2A 2A 2A   a1 + b1 x + c1 y
u3  N1 =
Similarly 2A
a 2 + b2 x + c2 y
Or, u1   v1  N2 =
    2A
{u} = [ N1 N2 N3 ] u2  {v} = [ N1 N2 N 3 ] v2  a + b3 x + c3 y
N3 = 3
u  v  2A
 3  3
 Plot Shapes
CST N1
u1 
u 
 2
u  u ( x, y )   N1 N2 N3 0 0 0  u3 
 = =  
v  v ( x, y )   0 0 0 N1 N2 N3   v1 
v2 
 
v3 

Higher order-6 node triangle

 Shape Functions on Natural Coordinates
1D-Bar Element

N2

N1

N1

N2
 Shape Functions on Natural Coordinates
Hermite Polynomial for Bar Element

Hermite polynomial also satisfies slope thus it can

be used for beam element where displacement as
well as rotation should be satisfied. Hermite
polynomial is given by

Lagrange interpolation

Thus
n n
H (ξ ) = ∑ 1− 2 (ξ − ξi ) L (ξi )  L (ξ ) f (ξi ) + ∑ (ξ − ξi ) L2i (ξ ) f ' (ξi )
'
i
2
i
i =1 i =1
Where L stands for Lagrange interpolation
 Shape Functions on Natural Coordinates
Hermite Polynomial

Condition that should be satisfied is

 Shape Functions on Natural Coordinates
Hermite Polynomial

The coefficients
 Shape Functions on Natural Coordinates

Hermite Polynomial
 Shape Functions on Natural Coordinates
Hermite Polynomial
 Plane Elements-Triangles
CST/LDT
When expressed in Cartesian coordinates, the interpolation
functions for triangular elements are algebraically complex.
Thus, these are often represented in area coordinates. Let
P(x, y) be a point which divides a triangle into three parts
such that coordinates of P is defined by

L1 + L2 + L3 = 1
Since A1 represents L1, it is unity at node 1 and zero at
line 2-3 if the point P moves parallel to the line 2-3. Thus

u1 
u 
 2
u  u ( x, y )   N1 N2 N3 0 0 0  u3 
 = =  
N1 = L1, N2 = L2 and N3 = L3, Thus, v  v ( x, y )   0 0 0 N1 N2 N3   v1 
v2 
 
v3 
 Plane Elements-Triangles
LST/QDT
 Plane Elements-Triangles

QST/CDT- Homework
 Lagrange Interpolation
Lagrange Polynomial 2D- Rectangle

1D-Bar
Lagrange Interpolation-9 node biquadrilateral element
 Serendipidity Elements
It has only edge nodes
Linear Serendipity is similar to Lagrange Element
Quadratic Serendipity Element
Quadratic Serendipity Element
 Quadratic Serendipity Elements

1 1
0 0
-1 -1
1 1 1 1
0 0 0 0
-1 -1 -1 -1

1 1
0 0
-1 -1
1 1 1 1
0 0 0 0
-1 -1 -1 -1

1 1
0.5 0.5
0 0
1 1 1 1
0 0 0 0
-1 -1 -1 -1

1 1
0.5 0
0 -1
1 1 1 1
0 0 0 0
-1 -1 -1 -1
Quadratic Serendipity Element
 Serendipidity Elements

Cubic Serendipity Element- Homework

 Cubic Serendipity Elements
(1−ξ )(1−η) ξ 2 +η2 −  (1+ξ )(1−η) ξ 2 +η2 − 
9 10 9 10
N1 = N2 =
32  9 32  9

(1+ξ )(1+η) ξ 2 +η2 −  (1−ξ )(1+η) ξ 2 +η2 − 

9 10 9 10
N3 = N4 =
32  9 32  9

N5 =
9
32
( )
1−ξ 2 (1−η )(1− 3ξ ) N6 =
9
32
( )
1−ξ 2 (1−η )(1+ 3ξ )

N7 =
9
32
( )
(1−ξ ) 1−η2 (1− 3ξ ) N8 =
9
32
(
(1+ξ ) 1−η2 (1+ 3ξ ) )
(-1, 1) (1, 1)
9
( )
N9 = 1−ξ 2 (1+η)(1+ 3ξ )
4 10 9 3

32
11 8
N10 = (1−ξ 2 ) (1+η )(1− 3ξ )
9
32 O (0, 0)

N11 = (1−ξ ) (1−η2 ) (1+ 3ξ )

9 12 7
32
N12 = (1−ξ ) (1−η2 ) (1− 3ξ )
9 1 5 6 2
(-1, - 1)
32 (1, -1)
 Cubic Serendipity Elements

1 1 1
0 0 0
-1 -1 -1
1 1 1 1 1 1
0 0 0 0 0 0
-1 -1 -1 -1 -1 -1

1 1 1
0 0 0
-1 -1 -1
1 1 1 1 1 1
0 0 0 0 0 0
-1 -1 -1 -1 -1 -1

1 1 1
0 0 0
-1 -1 -1
1 1 1 1 1 1
0 0 0 0 0 0
-1 -1 -1 -1 -1 -1

1 1 1
0 0 0
-1 -1 -1
1 1 1 1 1 1
0 0 0 0 0 0
-1 -1 -1 -1 -1 -1
 Shape Functions for 3D Elements
4 node-Tetrahedral Element
 Shape Functions for 3D Elements
8 node- Brick Element
Shape Functions -Assignment

• Generate member stiffness matrix

and rotation matrix considering
torsion in 3D
• Plot all shape functions
• Use Matlab or GNU-Plot
• Submit in one week