Shape Functions

Gokarna Bahadur Motra

FEM: Displacement Method
Choice of Displacement Functions:
• Displacement function is the beginning point for the structural analysis by finite element
method. This function represents the variation of the displacement within the element.
• Based on the problem to be solved, the displacement function needs to be approximated in the
form of either linear or higher-order function. A convenient way to express it is by the use of
polynomial expressions.
1. Convergence Criteria of Shape (Displacement) Functions:
• For convergence of the FEM solution, following three conditions should be fulfilled by the
assumed displacement function.
a. The displacement function must be continuous within the elements. Continuous function can
be obtained by selecting an appropriate polynomial. 1-D polynomial function of n-degree can
be selected as:
u ( x)  a0  a1 x  a2 x  a3 x  ...  an x
2 3 n

b. The displacement function must ensure rigid body displacements of the element. The constant
terms used in the polynomial (a0 to an) ensure this condition.
c. The displacement function must ensure the constant strain within the element. This can be
achieved by selecting infinitely small element. For this, the displacement function should
include appropriate terms.
FEM: Displacement Method
2. Compatibility Conditions:
• Shape/Displacement function should be compatible between neighboring
elements. Elements should be continuous and elements should not overlap, u ( x, y )  a0  a1 x  a2 y
when it deforms. u ( x, y )  a0  a1 x  a2 y  a3 x 2  a4 xy  a5 y 2
• Elements confirming convergence requirements and compatibility
conditions are known as conforming or compatible elements.
3. Geometric invariance:
• Deformation shape should be same when local coordinate system changes.
For this, polynomial should be complete. If it is not possible, incomplete
polynomial can be formed, but it must be balanced. For 2-D problems,
balanced polynomial can be formed with the help of “Pascal Triangle”. For
a four term polynomial, the invariance is ensured as follows:
u ( x, y )  a0  a1 x  a2 y  a3 xy
• The geometric invariance is obtained by maintaining symmetry,
while selecting the terms in Pascal’s Triangle.

Pascal’s Triangle
Shape Functions
• In FEA, the true deformation inside an element in not known, so the variations of
deformations inside an element are expressed using nodal deformation u ( x) 
using interpolation function.
N i ( x)ui

• u(x) is the deformation at any point inside the element and u ( x)   Ni ( x)ui are the
nodal deformations. Interpolating function is a single valued continuous polynomial of
n-degree.
• Interpolation function (Ni) is also known as “Shape function”. For linear interpolation,
n will be 1 and for quadratic interpolation n will become 2 and so on.
Types of Interpolation Function
(i) Lagrange interpolation and (ii) Hermitian interpolation
• Lagrange interpolation assumes that the function takes on the same values as the given
function at specified points. In Hermitian interpolation function, the slopes of the
function also take the same values as the given function at specified points.
Shape Functions
Degree of Continuity
• Let u(x) is an piecewise interpolation function in a finite element mesh. The
interpolation function u(x) should vary smoothly within the element. The transition
between adjacent elements may not be smooth.
• The term Cm is used to define the continuity of a piecewise displacement. A function
Cm is continuous if its derivative up to and including degree m are inter-element
continuous. For example, for one dimensional problem, u= u(x) is C0 continuous if u
is continuous, but its derivative is (u,x) is not.
• u= u(x) is C1 continuous if u and u’(x) are continuous, but u,xx is not.
• C0 element is used to model plane and solid body and C1 element is used to model
beam, plate and shell like structure, where inter-element continuity of slope is
necessary to ensure.
• Shape function for bar element, u1 ( x)  a0  a1 x, is C0 continuous, since u1,x is
discontinuous. Function u2 ( x)  a0  a1 x  a2 x 2, first derivative, u2 , x  a1  2a2 x and
second derivative, u2 , xx  2a2 , is discontinuous, so the function u2 is C1 continuous.
Shape Functions
Isoparametric elements
• If the shape functions (Ni) used to represent the variation of geometry of the element are the same as
the shape functions (N´i) used to represent the variation of the displacement, then the elements are
called isoparametric elements. For example, the coordinates (x,y) inside the element are defined by
the shape functions (Ni) and displacement (u,v) inside the element are defined by the shape functions
(N´i) as:
x  N i xi u  N i ui
'

y  N i yi v  N i'vi
• If Ni = N´i, then the element is called isroparametric. Fig. a) shows the two dimensional 8 node
isoparametric element.
• If the geometry of element is defined by shape
functions of order higher than that for the
variation of displacements, then the elements
are called superparametric (Fig. b).
• If the geometry of element is defined by shape
functions of order lower than that for representing
the variation of displacements then the elements
are called subparametric (Fig. c). a) Isoparametric b) Superparametric c) Subparametric
0 – Geometry of the element and Displacement
Element Types
One dimensional elements
(a) Two node element u ( x)  a0  a1 x u ( x)  a0  a1 x  a2 x 2
(b) Three node element

Two dimensional elements

(a) Triangular element
(b) Rectangular element
(c) Quadrilateral element
(d) Quadrilateral formed by two triangles
(e) Quadrilateral formed by four triangles
Element Types

• Simple element shapes

3. Three dimensional elements.
(a) Tetrahedron
(b) Rectangular brick element
Pascal Pyramid
• Generation of geometric
isotropic, symmetric 3-D
elements.
Polynomial Forms
• 1-D Elements P  x x
0
u ( x)  a0  a1 x x
 x  x
u ( x)  1   u1    u2  N1  x  u1  N 2  x  u2 Cartesian Coordinates
   
u ( )  a0  a1 P   
0 x
u ( )  1    u1   u2  N1   u1  N 2   u2 
• Higher Order Polynomials
1
Natural Coordinates
u ( x)  a0  a1 x  a2 x 2 1 2 3
2 2
 v1 2 v2
Polynomial Forms 1
P  x x
• Beam Elements 0
x
Polynomial Forms
• Beam Elements
2-D: Triangular Elements
• Used to model 2-D plane structures.
• The triangular elements can be used in axisymmetric 3-
D cases as well as in structural analyses involving out-
of-plane bending, as in plate and shell structures. Here, a) 3-node linear b) 6-node quadratic c)
the nodal degrees of freedom include first derivatives of 10-node cubic 2-D triangular elements
the field variable and the field variable itself.
• Assuming that only 1 degree of freedom for each node,
field variable can be expressed as:
  x, y   a0  a 1 x  a2 y

  x1 , y1   a0  a1 x1  a2 y1  1 1 x1 y1   a0  1 
    
  x2 , y2   a0  a1 x2  a2 y2  2  1 x2 y2   a1   2 
  x3 , y3   a0  a1 x3  a2 y3  3
1 x3 y3  a2  3 
2-D: Triangular Elements 1 x1 y1   a0  1 
    
• Inverting above system, coefficients can be 1 x2 y2   a1   2 
evaluated. 1 x3 y3  a2  3 
1
a0  1  x2 y3  x3 y2   2  x3 y1  x1 y3   3  x1 y2  x2 y1  
2A y 1 x1 y1
1
a1  1  y2  y3   2  y3  y1   3  y1  y2   1
x2 1
2A y2 x1 A  1 x2 y2
y3 2
1 1 x3 y3
a2  1  x3  x2   2  x1  x3   3  x2  x1   x3
2A
  x2 y3  x3 y2    y2  y3  x   x3  x2  y  1 
  
1  
  x, y     3 1 1 3  3
  x y  x y  y  y1 x   1 3  2 
x  x y  
2A   
  x1 y2  x2 y1    y1  y2  x   x2  x1  y  3 

N1  x, y    x2 y3  x3 y2    y2  y3  x   x3  x2  y 
N 2  x, y    x3 y1  x1 y3    y3  y1  x   x1  x3  y 
N 3  x, y    x1 y2  x2 y1    y1  y2  x   x2  x1  y 
2-D: Triangular Elements
• The field variable to take on a constant
value, as per the completeness requirement,
and that the first partial derivatives with
respect to the independent variables x and y
are constants. Local Coordinates
• The gradients of the field variable are
constant in both coordinate directions. For a
planar structural element, this results in
constant strain components.
• In structural applications, the three-node
triangular element is commonly known as a
constant strain triangle (CST).
2-D: Triangular Elements

• Difficult to obtain stiffness matrix in Cartesian

coordinates.
• Area Coordinates
A1 A2 A3
L1  ; L2  ; L3 
A A A
• The area coordinates are not independent,
since
L1  L2  L3  1
2-D: Triangular Elements
• Relation between two coordinates system

Displacement field
• Derivatives wrt Cartesian coordinates
Six Node Triangular Elements
• Complete polynomial

• Shape function
must have zero values at 2, 3, 4, 5 and 6
nodes.
 1  1 1 1
N1  c1L1  L1    1  c1  11    c1  2 N 4  c4 L1L2  1  c4    c4  4
 2  2 2 2
N 4  4 L1L2
 1
N1  2 L1  L1    N1  L1  2 L1  1
 2
• Integration in area coordinates
 L1=0 L2=0 L3=1
10-noded Triangular Element
L1=1/3 L2=1/3
L2=2/3

 1  2  1  2 9 L1=2/3 L3=1/3
N1  c1L1  L1   L1    1  c1  11  1    c1  L1=1 L2=1
 3  3  3  3 2
9  1  2
N1  L1  L1   L1   L3=0
2  3  3
Six Node Triangular Elements
• Integration
Rectangular Elements
• Polynomial for 4-node element as
incomplete, but symmetric to ensure
geometric isotropy
  x, y   a0  a1 x  a2 y  a3 xy
Rectangular Elements
• The interpolation functions in terms of the nodal η η
coordinates is algebraically complex. 2b
ξ ξ
• For the rectangular element, natural coordinates
2a
(normalized coordinates or serendipity
coordinates) in ξ and η will simplify the derivation
of shape functions. 1
xx yy x1  x2 y1  y2 N1  ,   1   1   
 ;  x y 4
a b 2 2
1
1 N 2  ,   1   1   
N1  ,   c1 1   1     1  c1 1  11  1  c1  4
4
1
  x, y     ,   N1  , 1  N 2  , 2  N 3  , 3  N 4  , 4 N 3  ,   1   1   
4
1
N 4  ,   1   1   
4
Higher Order Rectangular Elements
η
• For eight node element, a complete polynomial (-1,0)
cannot be constructed, but two type of ξ
incomplete, symmetric cubic polynomials can be (0,-1)
constructed:
N 5  ,   c5 1   1   1   
1
 1  c5 1  0 1  0 1  1  c5 
N1  ,   c1 1   1       1 2
1
1 N 5  ,   1   1   1   
 1  c1 1  11  1 1  1  1  c1  2
4
N1  ,  
1
1   1       1 N 5  ,  
1
2
1 2  1   
4
1
N 2  ,   1   1        1 N 6  ,  
1
2
1    1   
2

4
1
N 3  ,   1   1       1
4
N 7  ,  
1
2
1 2  1   
1
N 4  ,  
4
1   1       1 N8  ,  
1
2
1    1   
2
Three Dimensional Elements
• Four-node Tetrahedral Element
Volume Coordinates:
V1 V V V
L1  ; L2  2 ; L3  3 ; L4  4
V V V V
• Volume coordinates are not independent, since
• V1 + V2 + V3 + V4 = V
• If point P corresponds to node 1, V1 = V, V2 = V3 = V4 = 0
• L1 = 1, L2 = L3 = L4 = 0 at node 1.
• P moves away from node 1, V1 decreases linearly, since the volume of
a tetrahedron is directly proportional to its height (the perpendicular
distance from P to the plane defined by nodes 2, 3, and 4) and the area
of its base (the triangle formed by nodes 2, 3, and 4). On any plane
parallel to the base triangle of nodes 2, 3, 4, the value of L1 is
constant.
• If P lies in the plane of nodes 2, 3, 4, the value of L1 is zero. Identical
observations apply to volume coordinates L2, L3, and L4. So the
volume coordinates satisfy all required nodal conditions for
interpolation functions.
Three Dimensional Elements
Three Dimensional Elements
• Eight-node brick elements

N1  c 1 1   1   1   
1
 1  c1 1  11  11  1  c1 
8
1
N1  1   1   1   
8
Lagrange and Serendipity Elements
• Lagrange interpolation polynomials can be used to derive shape functions in a very simple way. This
method is suitable to derive shape function for elements having higher order of nodes.
• The Lagrange interpolation function at node i is defined by

f    
n
   

j       ...       ...   
1 2 i 1 i 1 n
i
          ...      ...   
j 1
j i
i j i 1 i 2 i i 1 i i 1 i n

• The function fi(ξ) gives the Lagrange interpolation function for ith node, and ξj denotes ξ coordinate of
jth node in the element. Substituting ξ = ξj, and j ≠ i, the value of the function fi(ξ) will be equal to zero.
• Substituting ξ = ξi, the numerator will be equal to denominator and hence fi(ξ) produces a value of
unity.
• Since, Lagrange interpolation function for ith node includes product of all terms except jth term; for an
element with n nodes, fi(ξ) will have n-1 degrees of freedom.
• For one-dimensional elements with n-nodes, shape function can be defined as . Ni(ξ)=fi(ξ).
Shape function for two node bar element
• Consider the natural coordinate of the center of 0 P  x x
the element as 0, and the natural coordinate of xi xj
x
the nodes 1 and 2 are -1 and +1, respectively.
The relation between natural and Cartesian
coordinates can be expressed as: 0 P   
2  x  x1    1 
x   1
 1
1 1
• Using Lagrange Polynomial, shape function for
2-node bar element can be written as:
N1  f1   
    2    1 1
  1   
1   2   1  1 2
N 2  f 2   
   1 

   1 1
 1   
 2  1  1  (1)  2
Shape function for three node bar element
• Quadratic shape functions

N1  f1   
    2   3 

   1 1
  1   
1   2 1  3   1 1  1 2
N 2  f 2   
   1   3    1  1
  1   2 
 2  1  2  3  1 1
N 3  f 3   
   1    2    1  1
     1
 3 1  3 2     2
      2 1
Shape functions for 2-D elements using Lagrange Interpolation Function
The Lagrange interpolation function for a rectangular element can be obtained from
the product of appropriate interpolation functions in the ξ direction fi(ξ) and η
direction fi(η).
Thus

Generation of N1
Generation of N2

Generation of N3
Generation of N4
 Serendipity Elements

• Higher order Lagrange elements contain internal nodes, which do not

contribute to the inter-element connectivity. However, these can be
eliminated by condensation procedure which needs extra computation.
The elimination of these internal nodes results in reduction in size of the
element matrices.
• Shape functions of two dimensional elements which contain nodes only
on the boundaries can be developed easily by inspection.
• The elements having nodes only at boundaries are called serendipity
elements (see Figs.).
Shape Function for 8 Node Rectangular Element (Serendipity Element)
- (Q8) – Quadratic in natural coordinates

Serendip, name for the island of Sri Lanka (Ceylon). It is best known to speakers of
English through the word “serendipity,” which was invented in the 18th century by
Horace Walpole. Walpole was inspired by a Persian fairy tale, “The Three Princes of
Serendip,” whose heroes often made discoveries by chance.

Corner nodes
At node 3, (,) = (1, 1) and N3 = 1, while at all other
nodes: N3 =0.

Following trial function satisfies the Kronecker delta

property and can be considered as possible shape
function.
Shape functions can be expressed in general form as
Mid side nodes 5 and 7
At node 5, (,) = (0, -1) and N5 = 1, while at all other nodes: N5 =0.
At node 7, (,) = (0, 1) and N7 = 1, while at all other nodes: N7 =0.
Trial functions for

Mid side nodes 6 and 8

At node 6, (,) = (1,0) and N6 = 1, while at all other nodes: N6 =0.
At node 8, (,) = (-1, 0) and N8 = 1, while at all other nodes: N8 =0.
Trial functions for
Shape function for Nine Node Rectangular Element (Lagrange Element) -
Quadratic –Two Terms
Note on Lagrange and Serendipity elements
Higher order Lagrange elements contains internal nodes, which do not
contribute to the inter-element connectivity as N9 = (1-2) (1-η2) = 0 on
all element boundaries.
It may be observed that the Lagrange elements have a better degree of
completeness in polynomial function compared to serendipity
elements. Therefore, Lagrange elements produce comparatively faster
and better accuracy .
 Seven node rectangular element (degrading Technique)

Displacement along the edge 1–4 is linear

 Shape Function Four-Node Tetrahedral Element (Linear) – Volume
Coordinate

The nodes are numbered 1–4 as

per the convention that node 1
can be selected arbitrarily and
nodes 2–4 are then specified in a
counter clockwise direction from
node 1.

This convention is the same as used by most commercial finite

element analysis software and is very important in assuring
geometrically correct tetrahedrons.

Point P(x , y, z) is an arbitrary point in the tetrahedron

 V1 = vol(P234),V2 = vol(P134),V3 = vol(P124),V4 = vol(P123)

The volume coordinates are defined as

If point P corresponds to node 1,

V1 = V, V2 = V3 = V4 = 0. L1 = 1, L2 = L3 = L4 = 0 at node 1

Volume coordinates satisfy all required nodal conditions for

interpolation functions
Note:
•Tetrahedral element is most useful in modelling irregular
geometries.
•However, the tetrahedral element is not particularly amenable to
use in conjunction with other element types, strictly as a result of
the nodal configurations.
•The field variable representation, as given is a linear function of
the Cartesian coordinates. Therefore, all the first partial derivatives
of the field variable are constant.

•In structural applications, the tetrahedral element is a constant

strain element. They are much too stiff.
Shape Function 10-Node Tetrahedral Element (Quadratic)
- Volume Coordinate
Trial Function for N1 (Corner Node)

i) Since L1= 0 at 2,3,4,8,9,10 and L1 = 1/2 at 5,6,7

N1 = C L1 (L1-1/2)

ii) Since N1 should be 1 at node 1 and L1=1; C = 2

N1 = 2 L1 (L1-1/2) = L1 (2L1-1)

In general Ni = Li (2 Li - 1)
Trial Function for N5 (Mid Node)

i. Since L2= 0 at 1,3,4,6,7,9 and L1 = 0 at 2 and 8

N5 = C L1L2

ii) Since N5 should be 1 at node 5, L1=1/2,L2= 1/2; C = 4

N5 =4 L1L2

Hence shape functions can be written as:

Pascal Pyramid to choose terms for the polynomial
 Brick Elements (Hexahedron) - Linear – 8 Nodes

The displacement component u can be expressed as:

u = a0 + a1 + a2  + a3+ a4   + a5  + a6   + a7  

 ,  , varies from -1 to +1

The above equation is incomplete but symmetric (from Pascal pyramid)

At Node 7: (, , = 1,1,1andN71

N7 1 1) 1) /8 = 1 at Node 7

= 0 at other nodes
as either or  or  will be = -1

The above expression for shape function satisfies Kroneckar delta property
and continuity exists due to linear variation of the shape function.

Let 0 i , 0  i and 0  i

 Lagrange Interpolation

Shape Function for Node 1

Substituting

1 1 1  1and 2  2 2 1

Likewise other shape functions can be derived.

• It may be observed that the Lagrange
elements have a better degree of
completeness in polynomial function
compare to serendipity elements. Therefore,
Lagrange elements produce comparatively
faster and better accuracy.

Equations of lines for 2-D 8-

node elements
Solid (3-D) Elements
• Two basic families of three-dimensional
elements as in Two-dimensional case.
• Extension of triangular elements will produce
tetrahedrons in three dimensions.
• Rectangular parallelepipeds are generated on
the extension of rectangular elements.
• Derivation of shape functions for such three
dimensional elements in Cartesian coordinates
are algebraically quite cumbersome. Commonly used 3-D solid elements for
• The shape functions for the two basic finite element analysis.
elements of the tetrahedral and parallelepipeds
families will be derived using natural
coordinates.
Solid (3-D) Elements

• The polynomial expression of the field variable

in three dimensions must be complete or
incomplete but symmetric to satisfy the
geometric isotropy requirements.
• Completeness and symmetry can be ensured
using the Pascal pyramid which is shown in Fig.
• Each independent variable must be of equal
strength in the polynomial.
• The geometric isotropy is not an absolute
requirement for field variable representation to
derive the shape functions. Pascal Pyramid
Tetrahedral Elements
• The coordinates of the nodes are defined both in Cartesian
and volume coordinates. Point P(x, y, and z) as shown in
Fig. is an arbitrary point in the tetrahedron.

• Here, L1, L2, L3, L4 are the set of natural coordinates inside
the tetrahedron and are defined as follows

• where Vi is the volume of the sub element which is bound by

point and face and is the total volume of the element. For Relationship between Cartesian
example L1 may be interpreted as the ratio of the volume of and natural coordinates
the sub element P234 to the total volume of the element
1234. The volume of the element is given by the
determinant of the nodal coordinates as follows:
Tetrahedral Elements
• The inverse relation is given by
Isoparametric formulation
Mapping techniques using interpolation functions

Analysis of irregular geometry is not possible by using

straight line elements. For the continuum having curve
boundary as in the Fig., use of straight line elements at the
boundaries is not possible.
To model complex boundaries, different curved elements
are required.

• Mapping a straight line, curve, and quadrilateral areas

with straight and curved edges.
• Requirement for valid mapping guidelines for mapped
element shapes.
• Numerical examples
Elements Mapping
Coordinate Transformation
The geometry of an element may be expressed
in terms of the interpolation functions as
where, n = nos of nodes, N=Interpolation Functions
x, y, z = Coordinates of nodal points of the element

Field variable variation in the element as

• Same shape functions are used for both the field variable and description of
element geometry, the method is known as isoparametric mapping.
• Elements defined by such a method is known as an isoparametric element.
• This method can be used to transform the natural coordinates of a point to the
Cartesian coordinate system and vice versa.
The main objective is to define suitable
elements, called parent or master elements,
and use the nodal coordinates of the actual
element as data for interpolation over the
master element.

The idea is first explained for mapping lines and

then is extended to the quadrilaterals.

The same concept works for three-dimensional

solid elements as well.
 Mapping a Straight Line

Consider two straight lines as shown in the figure above.

The master line is two units long and is lying along the  coordinate.

The given line is of an arbitrary length and orientated an x, y coordinate system.

The x, y coordinates of the given straight line can be considered as two sets of data
points for the master line and following linear interpolation can be established.
Hence x() and y() represent the mapping of the straight line to
the master line.
The coefficients used to map the lines are nothing but the shape
functions of linear bar element used to interpolate displacement.

Shape functions can be used to describe the displacement as well as

geometry also. This forms the basis of isoparametric elements used
in finite element analysis.
 Mapping a Curve

As an example, consider a quadratic curve in the xy plane

described in terms of coordinates at three points as shown in
the figure above.

The master element is defined: with nodes at = -1, 0, 1.

For interpolating between the three data points we use the

quadratic Lagrange interpolation functions as follows:
Example 1: The data points on a quadratic curve are (1, 1), (2,3.5), (4, 5).
Map this curve to a straight line two units long.

Using the above equations, for various

values of  between -1 to 1, x,y values
are generated .
It can be observed that the given points
are mapped appropriately into the
master element and one to one
correspondence exists.

Each point on the master element maps

into a unique point on the given curve
Example 2 : The data points on a quadratic curve are (1, 1), (1.25,3.5), (4, 5).
Map this curve to a straight line two units long.

The following points are noted.

The y values are acceptable as they are

continuously increasing from -1 to +1.

However, x values between  = -1 to 0 are not

acceptable as they are not in the range 1 to 1.25
and for different values of  same values of x are
obtained.

The mapping is not correct as many to one

correspondence exist.

Each point on the master element does not map

into a unique point on the given curve.
 Condition for good mapping of curves
In the previous examples the coordinates are given in increasing
order.

Hence the mapping coordinates should always be increasing i.e.,

derivative of the mapping functions must be positive (dx/d and
dy/d > 0).

Case 1: Example 1
The mapping functions are obtained as
Case 2: Example 2
The mapping functions are obtained as
 General requirement for location of the middle point for a
quadratic curve to get good mapping

The general formula for mapping of a quadratic curve is given by:

is the length of horizontal projection of curve.

Considering the positive requirement of the derivative for good mapping, we

have
Hence, it can be stated that the derivative will be positive over the entire range
of  (-1 to 1), if,

Case 1: Example 1
The data points on a quadratic curve are (1, 1), (2,3.5), (4, 5)

The condition is satisfied. Hence, results in good mapping.

Case 2: Example 2
The data points on a quadratic curve are (1, 1), (1.25,3.5), (4, 5)

The condition is not satisfied. Hence, results in bad mapping.

Mapping Quadrilateral Areas
The concept used to map lines and curves can be extended easily to
quadrilateral areas.

In this case the master area is 2x2 square in (,) coordinate system.
Quadrilaterals with Straight Sides

The shape functions can be written as:

 The mapping functions can be written as:

In compact form the above equations can be written as:

Provides the required mapping from the (  , ) to (x,y) and is known as

isoparametric mapping.
Let’s check for appropriate mapping as follows

Hence, line corresponding to (x1,y1) and (x2,y2) in general

quadrilateral area maps to line corresponding to (-1,-1) and (1,-1)
of master area.

Similarly it is possible to verify that each point on the master area

maps to a unique point on the quadrilateral area.
Example 3: Determine the Cartesian coordinate of the
point P (ξ= 0.8, η= 0.9) as shown in Fig

Shape functions at point P:

(x,y) coordinates of point P:

 Concept of Jacobian Matrix
• Derivatives of the interpolation functions with respect to the global coordinates are
necessary to formulate the element stiffness matrices.
• Both element geometry and variation of the shape functions are represented in terms of
the natural coordinates of the parent element, some additional mathematical obstacle
arises.
• For example, to evaluate strain vector, the operator matrix is a function of x and y, but
the interpolation function is with ξ and η. So, the operator matrix need to be transformed
for taking derivative with ξ and η.
• The relationship between two coordinate systems can be computed by using the chain
rule of partial differentiation as In matrix form

Elements of Jacobian
 Restrictions on mapping of areas

The derivative of the mapping function should be positive over the range of the master line.
Likewise, a similar criterion can be established for quadrilateral areas.
Since, the mapping function consists of two variables we use the chain rule for partial
differentiation as

In matrix form:

[J] s called the Jacobian matrix which establishes the relation between
(x,y) and (,)
Likewise, expressions can be developed for other terms in [J]
and can be written as:

Thus, for the mapping to be valid,

|J| > 0 for all (,) in the range(-1,1 to -1,1)
Example 4: Calculate the Jacobian matrix for four node two
dimensional quadrilateral element corresponding to the
point  = = 0.57735 as shown in Fig.
Example 5: Evaluate the Jacobian matrix [J] for the four-node element (square
with sides equal to 1 rotated by 45 degrees)
Example 6: Check the validity of isoparametric mapping
for the element shown in Figure
|J| > 0, for all values of (,) in the range. Hence, the mapping is valid.
|J|=0 for  = --4 which is clearly outside the area.
Example 7: Check the validity of isoparametric mapping for
the element shown in Figure
|J| < 0, for all (,) =(1-1). Hence, the mapping is invalid

|J |= 0 for  =  + 0.4 which is clearly inside the area

Example 8: Check the validity of isoparametric mapping
for the element shown in Figure
|J| > 0, for all (,) . Hence, the mapping is valid.

|J|= 0 for  = 5.818013(0.1075 + 0.615625) which is clearly outside

the area
Example 9: Check the validity of isoparametric mapping for
the element shown in Figure
|J| < 0, for (,) =(1-1). Hence, the mapping is invalid.

|J |= 0 for  = 1.402844 + 0.990521 which is clearly inside the area.

 Triangular elements by collapsing quadrilateral Elements

Assign same global nodes to two corner nodes of the element.

Substituting nodal coordinates
 Quadrilaterals with Curved Sides

Using appropriate interpolation functions, it is possible to map quadrilaterals

with one or more sides curved.

The master element side that corresponds to a curve will have as many
equally spaced nodes as those needed to define the curve on the actual area.

For example, for a side represented by a quadratic curve, there will be three
nodes on the master element placed at - 1, 0, and 1.

Thus, if all sides of an element are quadratic curves, then there will be a
total of eight nodes.
All four sides are quadratic curves
The master area is an eight noded square.

The shape functions are that of eight node rectangular element

First three sides are quadratic curves
The master area is a seven noded square.

The shape functions are that of seven node rectangular element

Opposite sides are quadratic curves
The master area is a six noded square.
The shape functions are that of six node
rectangular element derived using degrading
technique.
Two adjacent sides are quadratic curves
The master area is a six noded square. The shape functions are
that of six node rectangular element derived using degrading
technique
First side quadratic curve
The master area is a five noded square. The shape functions are
that of five node rectangular element derived using degrading
technique
Guidelines for Mapped Element Shapes

In general, the element geometry is invalid if the Jacobian is

either zero or negative anywhere in the element.

Problem also arises when the Jacobian matrix is nearly singular

either due to round off errors or due to badly shaped elements.

To avoid problems due to badly shaped elements, it is

recommended that the inside angles in quadrilateral elements
be > 15˚ and < 165˚ as illustrated in Figure.
Mapping Solid elements with straight and curved sides
With Straight sides
The shape functions can be written as:

The geometry of an element may be expressed in terms of the

interpolation functions as follows.
Following relation can be established between (x,y,z) and (,,)

[J] is called the Jacobian matrix

 Need of Numerical Integration
• The integrations in FEM are quite complicated and it is not possible to find a
closed form solutions to those problems.
• Exact and explicit evaluation of the integral associated to the element matrices and
the loading vector is not always possible because of the algebraic complexity of
the coefficient of the different equation (i.e., the stiffness influence coefficients,
elasticity matrix, loading functions etc.).
• In FEA evaluation of the integrals in one, two and three dimensional are necessary
to compute element stiffness and element load vector.

• Approximate solutions using numerical techniques are available in mathematics

for solving definite integration problems, including, mid-point rule, trapezoidal-
rule, Simpson’s 1/3rd rule, Simpson’s 3/8th rule and Gauss Quadrature formula.
• Among these, Gauss Quadrature technique is most useful one for solving problems
in FEM and discussed here in detail.
 Numerical Integration- Gauss Quadrature
• Line or one-dimensional integrals: One point, two
point and three point formula.
• Procedure and numerical examples.
• Area or two-dimensional Integrals: procedure and
numerical examples.
• Volume or three-dimensional Integrals: procedure and
numerical examples
 Numerical Integration- Gauss Quadrature

When mapping is used to develop finite elements, especially higher order

elements, the integrands become quite complicated and it is not possible to
evaluate the required integrals in a closed form.

We must use numerical integration in these situations.

Simple numerical integration techniques such as the trapezoidal rule and

Simpson's formula do not work that well in finite element computations.

The Gauss quadrature, also called the Gauss-Legendre quadrature, is more

suitable for finite element applications and has become the standard tool for
developing mapped elements.
Line or one-Dimensional Integrals
In the Gauss quadrature, the basic derivation assumes that the integral of
a function f() is to be evaluated over an interval -1    1.

The main idea is to represent the integral in the following form:

The locations of Gauss points and the appropriate weights are

determined by requiring that the above equation gives exact integrals
for constant, linear, quadratic, etc., terms in a polynomial.
One point formula

w1 and s1 are determined by making this formula to give exact

integrals for the first two terms in the linear polynomial
f(s) = 1+ s

Hence, one point formula gives

exact integrals for linear
polynomials beyond which the
results will be approximate.
Two point formula

w1, s1 and w2, s2 are determined by making this formula to give

exact integrals for the first four terms in the cubic polynomial :
f(s) = 1+ s +s2+s3
 The four conditions can be satisfied with

Hence, the two-point formula will give exact integrals for polynomials
up to the third order, beyond which the results will be approximate.
Three point formula

w1 , s1 ; w2 , s2 and w3 , s3 are determined by making this formula to

give exact integrals for the first six terms in the fifth degree
polynomial
f(s) = 1+ s +s2+s3 +s4+s5

Hence, the three-point formula will give exact integrals for

polynomials up to the fifth order, beyond which the results will be
approximate.
Note:
Gauss integration procedure yield exact values for polynomials of
order  (2n-1) where n = number of Gauss points

Example 1: Evaluate the integral using one, two and three point formula
and compare with exact solution.

One point formula:

 Two point formula:

Three point formula:

Exact:

Note: Two point formula itself gives exact values since the order of polynomial
 2x2-1=3
Example 2: Evaluate the integral using one and two point formula and
compare with exact solution.

One point formula:

Two point formula:

Three point formula:

Transformation to evaluate integrals with different limits

To evaluate this as per Gauss integration procedure, following transformation

is necessary

Let z be the change in variable then

Example 3: Evaluate the integral using one and two point formula and
compare with exact solution.

One point formula:

Two point formula:

Exact:

Example 4: Evaluate the integral using one, two point formula and
compare with exact solution

One point formula:

Two point formula:

Exact:
 Area or two-dimensional Integrals
The one-dimensional Gauss quadrature formule extend easily to two-
dimensional integrals, if the integration region is a 2 x 2 square with the
origin at the center.

Gauss Product rules are obtained by successive application of one

dimensional rule.

The total number of points is (m x n).

Usually the same number of points are used in both directions,
1 x 1 = 1, 2 x 2= 4, 3 x 3 = 9
One point rule:

is exact for a product of two linear polynomials.

Two point rule:

is exact for a product of two cubic polynomials.

Example 5: Evaluate the integral using one and two point
formula and compare with exact solution.

One point rule:

Two point rule:

Exact
Example 6: Evaluate the integral using one and two
point formula and compare with exact solution.

Let p and r be the change variables corresponding to x and y

Similarly
One point rule:

Two point rule

Exact: 224/9 = 24.888

Volume or three-dimensional Integrals
The integration region now must be a 2 x 2 x 2 cube with
the origin at the center.

Gauss Product rules are obtained by successive application

of one dimensional rule.

The total number of points is (m x n x p).

Usually the same number of points is used in both
directions,
1 x 1 x 1 = 1, 2 x 2 x 2 = 8, 3 x 3 x 3 = 27.