Ch05 - Shape Functions
Ch05 - 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
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
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
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
N2
N1
N1
N2
Shape Functions on Natural Coordinates
Hermite Polynomial for Bar Element
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
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
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)
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