FINITE ELEMENT

ANALYSIS
Dr. Richa Agrawal

MODULE 2 – ELEMENT AND ITS SHAPE FUNCTIONS

Types of Elements

• 1D Element – Rod, Bar, Beam, Pipe, Truss

• 2D Element – Triangular, Quadrilateral
• 3D Element – Tetrahedral, Hexahedral
1D Analysis
• Variation of Primary Variable in one direction.
• Every point is defined by only one coordinate
Types of Nodes
• External Nodes:
• Nodes that occur on the edges / surfaces of the element
• Can be common to one or more elements
• Internal Nodes:
• Nodes that occur inside the element
• Specific to the element
Order and DOF of Element
• Lower and Higher order Element, order of shape functions
• No. of unknown primary Variables at each node is degree
of freedom
Types of Coordinate
• Global Coordinate System
• With reference to the domain
• Applicable to all element
• Local Coordinate System
• With reference to each element
• Independent to each element
• Natural Coordinate System
• Type of local coordinate
• Non-dimensional number
• Used for ease in formulation of stiffness matrix
Shape Functions
𝑢 = 𝑎𝑥ҧ + 𝑏
• 1D Linear Element At node 1, u = 𝑢1 , 𝑥ҧ = 0
∴, 𝑢1 = 𝑏
At node 2, u = 𝑢2 , 𝑥ҧ = ℎ𝑒
∴, 𝑢2 = 𝑎ℎ𝑒 + 𝑏
But b = 𝑢1 ,
∴, 𝑢2 = 𝑎ℎ𝑒 + 𝑢1
𝑢2
∴𝑎=
𝑢2 − 𝑢1
𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔 𝑣𝑎𝑙𝑢𝑒𝑠 𝑜𝑓 𝑎 𝑎𝑛𝑑 𝑏,
𝑢2
𝑢= 𝑥ҧ + 𝑢1
𝑢2 −𝑢1
𝑅𝑒𝑎𝑟𝑟𝑎𝑛𝑔𝑖𝑛𝑔 𝑡ℎ𝑒 𝑡𝑒𝑟𝑚𝑠
𝑥ҧ 𝑥ҧ
𝑢 = 𝑢1 1 − ℎ + 𝑢2 ℎ
𝑒 𝑒
𝑥ҧ 𝑥ҧ
𝜑1 = 1 − ℎ , 𝜑2 = ℎ
𝑒 𝑒
Shape Functions
Characteristics of Shape Functions
• The value of shape function at a specified node will be 1

and zero at other nodes

• Sum of all shape function at a specific point is 1

• No. of nodes is equal to no. of shape functions.

 Alternate method to develop Shape
Functions for an Element of order n
• Element of length he, n no. of nodes, degree of polynomial
function is (n-1)
ℎ𝑒
• Length of each segment within any two nodes will be
(𝑛−1)
• Assuming pilot function:

ℎ𝑒 ℎ𝑒 ℎ𝑒 ℎ𝑒
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − 𝑥ҧ − 2 𝑥ҧ − 3 … . 𝑥ҧ − 𝑛 − 1
(𝑛 − 1) (𝑛 − 1) (𝑛 − 1) (𝑛 − 1)

No. of terms should be equal to no. of nodes

 Linear Element (n = 2)
ℎ𝑒 ℎ𝑒 ℎ𝑒 ℎ𝑒
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − 𝑥ҧ − 2 𝑥ҧ − 3 … . 𝑥ҧ − 𝑛 − 1
(𝑛−1) (𝑛−1) (𝑛−1) (𝑛−1)
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − ℎ𝑒
where A is constant
At node 1, 𝑥ҧ 𝑣𝑎𝑛𝑖𝑠ℎ𝑒𝑠 (Since 𝑥ҧ = 0 𝑎𝑡 𝑛𝑜𝑑𝑒 1)
∴ 𝜑1 = 𝐴 𝑥ҧ − ℎ𝑒 … … … . . (1)
𝑥ҧ = 0, 𝜑1 = 1
∴ 1 = 𝐴(−ℎ𝑒 )
1
∴𝐴=−
ℎ𝑒
1
𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔 𝑖𝑛 1 , 𝜑1 = − 𝑥ҧ − ℎ𝑒
ℎ𝑒
𝑥ҧ
𝜑1 = 1 −
ℎ𝑒
 Linear Element (n = 2)
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − ℎ𝑒 , where A is constant
ഥ − ℎ𝑒 ) 𝑣𝑎𝑛𝑖𝑠ℎ𝑒𝑠 (Since 𝑥ҧ = ℎ𝑒 𝑎𝑡 𝑛𝑜𝑑𝑒 2)
At node 2, (𝑥
∴ 𝜑2 = 𝐴𝑥ҧ … … … . . (2)
𝑥ҧ = ℎ𝑒 , 𝜑2 = 1
∴ 1 = 𝐴(ℎ𝑒 )
1
∴𝐴=
ℎ𝑒
1
𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔 𝑖𝑛 1 , 𝜑2 = 𝑥ҧ
ℎ𝑒
𝑥ҧ
𝜑2 =
ℎ𝑒
 Quadratic Element (n = 3)
ℎ𝑒 ℎ𝑒 ℎ𝑒 ℎ𝑒
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − 𝑥ҧ − 2 𝑥ҧ − 3 … . 𝑥ҧ − 𝑛 − 1
(𝑛−1) (𝑛−1) (𝑛−1) (𝑛−1)

ℎ𝑒
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − 𝑥ҧ − ℎ𝑒
2
At node 1, 𝑥ҧ 𝑣𝑎𝑛𝑖𝑠ℎ𝑒𝑠 (Since 𝑥ҧ = 0 𝑎𝑡 𝑛𝑜𝑑𝑒 1)
ℎ𝑒
∴ 𝜑1 = 𝐴 𝑥ҧ − 𝑥ҧ − ℎ𝑒 … … … . . (1)
2
𝑥ҧ = 0, 𝜑1 = 1
ℎ𝑒
∴1=𝐴 − (−ℎ𝑒 )
2
2
∴𝐴=
ℎ𝑒 2
2 ℎ𝑒
𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔 𝑖𝑛 1 , 𝜑1 = 𝑥ҧ − 𝑥ҧ − ℎ𝑒
ℎ𝑒 2 2
2𝑥ҧ 𝑥ҧ
𝑅𝑒𝑎𝑟𝑟𝑎𝑛𝑔𝑖𝑛𝑔, 𝜑1 = 1 − ℎ 1−ℎ
𝑒 𝑒
 Quadratic Element (n = 3)
ℎ𝑒
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − 𝑥ҧ − ℎ𝑒
2
ℎ𝑒 ℎ𝑒
At node 2, 𝑥ҧ − 𝑣𝑎𝑛𝑖𝑠ℎ𝑒𝑠 (Since 𝑥ҧ = 𝑎𝑡 𝑛𝑜𝑑𝑒 2)
2 2
∴ 𝜑2 = 𝐴𝑥ҧ 𝑥ҧ − ℎ𝑒 … … … . . (2)
ℎ𝑒
𝑥ҧ = , 𝜑2 =1
2
ℎ𝑒 ℎ𝑒
∴1=𝐴 −
2 2
4
∴𝐴=−
ℎ𝑒 2
4
𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔 𝑖𝑛 2 , 𝜑2 = − 2 𝑥ҧ 𝑥ҧ − ℎ𝑒
ℎ𝑒
4𝑥ҧ 𝑥ҧ
𝑅𝑒𝑎𝑟𝑟𝑎𝑛𝑔𝑖𝑛𝑔, 𝜑2 = 1−
ℎ𝑒 ℎ𝑒
 Quadratic Element (n = 3)
ℎ𝑒
𝜑 = 𝐴𝑥ҧ 𝑥ҧ − 𝑥ҧ − ℎ𝑒
2
At node 3, 𝑥ҧ − ℎ𝑒 𝑣𝑎𝑛𝑖𝑠ℎ𝑒𝑠 (Since 𝑥ҧ = ℎ𝑒 𝑎𝑡 𝑛𝑜𝑑𝑒 3)
ℎ𝑒
∴ 𝜑3 = 𝐴𝑥ҧ 𝑥ҧ − … … … . . (3)
2
𝑥ҧ = ℎ𝑒 , 𝜑3 = 1
ℎ𝑒
∴ 1 = 𝐴ℎ𝑒
2
2
∴𝐴=
ℎ𝑒 2
2 ℎ𝑒
𝑆𝑢𝑏𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑛𝑔 𝑖𝑛 3 , 𝜑3 = 𝑥ҧ 𝑥ҧ −
ℎ𝑒 2 2
𝑥ҧ 2𝑥ҧ
𝑅𝑒𝑎𝑟𝑟𝑎𝑛𝑔𝑖𝑛𝑔, 𝜑3 = − 1−
ℎ𝑒 ℎ𝑒
Quadratic Element (n = 3)
2𝑥ҧ 𝑥ҧ
𝜑1 = 1 − 1−
ℎ𝑒 ℎ𝑒
4𝑥ҧ 𝑥ҧ
𝜑2 = 1−
ℎ𝑒 ℎ𝑒
𝑥ҧ 2𝑥ҧ
𝜑3 = − 1−
ℎ𝑒 ℎ𝑒
Cubic Element (n = 4)
3𝑥ҧ 3𝑥ҧ 𝑥ҧ
𝜑1 = 1 − 1− 1−
ℎ𝑒 2ℎ𝑒 ℎ𝑒

𝑥ҧ 3𝑥 𝑥ҧ
𝜑2 = 9 1− 1−
ℎ𝑒 ℎ𝑒 ℎ𝑒

9 𝑥ҧ 3𝑥ҧ 𝑥ҧ
𝜑3 = − 1− 1−
2 ℎ𝑒 ℎ𝑒 ℎ𝑒

𝑥ҧ 3𝑥ҧ 3𝑥ҧ
𝜑4 = 1− 1−
ℎ𝑒 ℎ𝑒 2ℎ𝑒
Properties of Global Stiffness Matrix
• Stiffness matrix is symmetric
• Stiffness matrix is banded matrix i.e. all terms outside of
the band are zero.
• Stiffness matrix has positive diagonal terms.
• Stiffness matrix is singular
• Stiffness matrix has a property where each terms in
column or row sums to zero
• The size of the stiffness matrix is (mxn) where m is
degree of freedom at each node and n is number of
nodes.
Refer class notes for derivation of shape function by
Lagrange Polynomial in Local and Natural Coordinates and
more content