Isoparametric Formulation 219

13
Isoparametric Formulation
13.1 INTRODUCTION
The various elements so far we have seen are having straight edges. To take care of curved boundaries
refined meshes are to be used when straight edged elements are employed. Even with refined meshes
analysts were not happy with the results since unnecessary stress concentrations are introduced. Higher
order elements also do not overcome the problem of suitably approximating curved boundaries. The
isoparametric concept brought out by Taig [1] and latter on generalized by B.M. Irons [2] revolutionized
the finite elements analysis and it also helped in properly mapping the curved boundaries. They brought out
the concept of mapping regular triangular and rectangular elements in natural coordinate system, to
arbitrary shapes in global system as shown in Fig. 13.1. In this chapter method of coordinate transformation
of natural coordinates to global coordinate system is presented. The terms isoparametric, super parametric
and subparametrics are defined. The basic theorems on which isoparametric concept is based are explained
and need for satisfying uniqueness theorem of mapping is presented. Assembling of stiffness matrix is
illustrated. For assembling stiffness matrix integration is to be carried out numerically. The Gaussion
integration technique which is commonly employed is explained briefly. To make the procedure clear few
small numerical problems are illustrated and lastly application to structural engineering problems is
presented.
Parent elements in natural Mapped element if global
coordinate system system

x
0

Fig. 13.1 Concept of mapping in isoparametric elements

2 Finite Element
y
4 (–1, 1) 4 (x4, y4) 3 (x3, y3)
3 (1, 1)
P (,x) y
P ()

2 (x2, y2)
1 (– 1, –1) 2 (1, –1) 1 (x1, y1)
x
0

y
7 3

4 7 4 P (,x) y
3 Quadratic
curve
P () 6
8 8
6

5 2
1 5 2 1
x
0

4 10 9 4 10
3 9 3

P ()
11 8 11
P (,x) y 8

12 7 12
7 Cubic curves

1 2 5 6
5 6 1 2
x
0

x
0

Fig. 13.1 (contd)

 Isoparametric 2
y

x
0

Fig. 13.1 (contd)

13.2 COORDINATE TRANSFORMATION

So far we have used the shape functions for defining deflection at any point interms of the nodal
displacement. Taig [1] suggested use of shape function for coordinate transformation form natural local
coordinate system to global Cartesian system and successfully achieved in mapping parent element to
required shape in global system. Thus the Cartesian coordinate of a point in an element may be expressed
as
x = N1 x1 + N2 x2 + ... + Nn xn
y = N1 y1 + N2 y2 + ... + Nn yn
z = N1 z1 + N2 z2 + ... + Nn zn
or in matrix form
(x) N (x) e

where N are shape functions and (x)e are the coordinates of nodal points of the element. The shape functions
are to be expressed in natural coordinate system.
For example consider mapping of a rectangular parent element into a quadrilateral element:

y
4 (–1, 1) 4 (x4, y4) 3 (x3, y3)
3 (1, 1)
P (,x) y
P ()

2 (x2, y2)
1 (– 1, –1) 2 (1, –1) 1 (x1, y1)
x
0
(a) Parent element (b) Mapped element

Fig. 13.2 Wapping of rectangular element in natural local

coordinate system to global cartesian coordinatee system
 2 Finite Element

The parent rectangular element shown in Fig. 13.2 (a) has nodes 1, 2, 3 and 4 and their coordinates are
(–1, –1), (–1, 1), (1, 1) and (1, –1). The shape functions of this element are

N1 
(1   ) (1  , N  (1  ) (1 
2

) )
4 4
(1   ) (1  (1   ) (1   )
)
N3  and N4 
4 4
P is a point with coordinate (, ). In global system the coordinates of the nodal points are
(x1, y1), (x2, y2), (x3, y3) and (x4, y4)
To get this mapping we define the coordinate of point P as
x = N1 x1 + N2 x2 + N3 x3 + N4 x4
and y = N1 y1 + N2 y2 + N3 y3 + N4 y4
Noting that shape functions are such that at node i, Ni = 1 and all others are zero, it satisfy the coordinate
value at all the nodes. Thus any point in the quadrilateral is defined in terms of nodal coordinates.
Similarly other parent elements are mapped suitably when we do coordinate transformation.

13.3 BASIC THEOREMS OF ISOPARAMETRIC CONCEPT

Isoparametric concept is developed based on the following three basic theorems:
Theorem I: If two adjacent elements are generated using shape functions, then there is continuity at the
common edge.

A A

x
0

Fig. 13.3

It may be observed that in the parent element, for any point on edge AB, shape functions Ni = 0 for
nodes not on the edge and Ni exists for nodes on the edge. Hence the final function is the same for the
common edge AB in any two adjacent elements, when we give the same coordinate values for the nodes on
common edge. Hence edge AB is contiguous in the adjacent elements.
 Isoparametric 2

Theorem II: It states, if the shape functions used are such that continuity of displacement is represented in
the parent coordinates, then the continuity requirement, will be satisfied in the isoparametric elements also.
The proof is same as for theorem 1.
Theorem III: The constant derivative conditions and condition for rigid body are satisfied for all isoparametric
elements if,

 N 1 i

Proof: Let the displacement function be

u   1  2x   3y   4
…(13.1)
z
 Nodal displacement at ‘i’th node is given by
ui  1   2 xi   3 yi   4
zi
In finite element analysis we define nodal displacement at any point in the element in terms of nodal
displacement as

u Nu i i

 u   Ni ( 1   2 xi   3 yi   4 zi )

 1  N   N x   N y  
i 2 i i 3 i i 4

N i zi
From the isoparametric concept, we know

Nxx i i

N yy i i

N z z i i …(13.2)

u  i  N 
i 2 x   3y   4 z

Hence if equation 13.2 has to represent equation 13.1 uniquely, then

 N 1 i

The shape functions developed in natural coordinate systems satisfy this requirement. Hence they can
be safely used for isoparametric representation. This theorem is known as convergence criteria for
isoparametric elements.

13.4 UNIQUENESS OF MAPPING

It is absolutely necessary that a point in parent element represents only one point in the isoparametric
element. Some times, due to violent distortion it is possible to obtain undesirable situation of
nonuniqueness. Some of such situations are shown in Fig. 13.4. If this requirement is violated determinant
of Jacobiam matrix (to be explained latter) becomes negative. If this happens coordinate transformation
fails and hence the program is to be terminated and mapping is corrected.
 2 Finite Element

Fig. 13.4 Hon-uniqueness of mapping

13.5 ISOPARAMETRIC, SUPERPARAMETRIC AND

SUBPARAMETRIC ELEMENTS
We have seen that in the finite element analysis with isoparametric elements, shape functions are used for
defining the geometry as well as displacements. If the shape functions defining the boundary and
displacements are the same, the element is called as isoparametric element. For example, in Fig. 13.5 (a)
all the eight nodes are used in defining the geometry and displacement. Thus, in this case

Nodes used for defining geometry Nodes

used for defining displacement
(a) Isoparametric (b) Superparametric (c) Subparametric

Fig. 13.5 Isoparametric, superparametric and subparametric elements

 Isoparametric 2

u  [N] {}e , x = [N] {x}e and y = [N] {y}e where [N] is quadratic shape function of serendipity family.
The element in which more number of nodes are used to define geometry compared to the number of nodes
used to define displacement are known as superparametric element. One such element is shown in Fig.
13.5
(b) in which 8 nodes are used to define the geometry and displacement is defined using only 4 nodes. In the
stress analysis where boundary is highly curved but stress gradient is not high, one can use these elements
advantageously.
Figure 13.5 (c) shows a subparametric element in which less number of nodes are used to define
geometry compared to the number of nodes used for defining the displacements. Such elements can be used
advantageously in case of geometry being simple but stress gradient high.

13.6 ASSEMBLING STIFFNESS MATRIУ

Assembling element stiffness matrix is a major part in finite element analysis. Since it involves coordinate
transformation from natural local coordinate system to Cartesian global system, isoparametric elements
need special treatment. In this article assembling of element stiffness matrix for 4 noded quadrilateral
element is explained in detail. The procedure can be easily extended to higher order elements by using
suitable functions and noting the increased number of nodes.
Figure 13.6 shows the typical parent element and isoparametric quadrilateral element.
y
4 (–1, 1) (x3, y3)
3 (1, 1)

(x4, y4)

(x2, y2)
1 (– 1, –1) 2 (1, –1) (x1, y1)
x
0

Fig. 13.6 Typical isoparametric quadrilateral element

For parent element, the shape functions are

(
1   i )(
1   i )
Ni  …(13.3)
4

i.e., N1 
(1   ) (1  , N  (1   ) (1 
2

) )
4 4
(1  ) (1  (1   ) (1 
) )
N3  and N4 
4 4
We use the above functions for defining the displacement as well as for defining the geometry of any
point within the element in terms of nodal values.
 2 Finite Element

When we use shape functions for the geometry,

ç x1
j
y
jy
ç( x)y jrN j xj
0 y (jy )j
1

j
2
1
0 N2 0 N3 0 N4 2

tyj L 0 N1 0 N2 0 N3 …(13.4)
0 N4
j 三xj j j
4

ty j 4

The above relation helps to determine the (x, y) coordinates of any point in the element when the
corresponding natural coordinates  and  are given.
We are also using the same functions for defining the displacement at any point in the element

ç u1
jv
y j
1

j
uj
ç( u)y jrN 0 y (jv )j
j
2
1
0 N2 0 N3 0 N4 2

tvj L 0 N1 0 N2 0 N3
0 N4
…(13.5)

j j三
u4
tjv j 4

j
In assembling the stiffness matrix we need the derivatives of displacements with respect to global x, y
system. It is easy to find derivatives with respect to local coordinates  and  but it needs suitable assembly
to get the derivatives w.r.t. to global Cartesian system.
The relationship between the coordinates can be computed using chain rule of partial differentiation.
Thus,
 x  y 
 


  x  y


 x  y 
   x   y
 

 y ç x y ç  ç  y
j j y j j j
i.e., j(  j)  jj) x j) j(x …(13.6)
j(   J
j)

j
 x y   Isoparametric 2

jt  j jt y
where J
 jt y j j
…(13.7)
ç y y
x j
(

y
)jj
x

jt 

 2 Finite Element

The matrix [J] shown above is called Jacobian matrix. It relates derivative of the function in local
coordinate system to derivative in global coordinate system. In case of three dimensional problem it is given
by

r x y
j  y

z
 j
J  jj  z j
x 
y  j
j x  z
j
jL  y  j

Now going back to isoparametric quadrilateral element,
Let
J11 J12r y
J  j j
J J
L 21 22

Where J11  x J12  y

 

J21  x J22  y
 


we know, x N i
 N1x1  N2 x2  N3 x3  N4 x4
xi
i1

x N1
J   x N2 N3 N4
 x  x  x
11  1
 2
 3
 4

Similarly J12, J21 and J22 can be assembled.
Then we get
r 4
N 4
N y

j   x   y j
i
i
i
i

J j
jN x  N y jj
i1 i1 …(13.8)
4 i 4 i
i i

L i1
i1

For any specified point the above matrix can be assembled. Now,

j
ç 
(yj )
j

jt
 Isoparametric 2
ç  y
J
j j (x )
jt y j
 2 Finite Element

ç y
çx j j
 (j ) J ( )
1

y j j
 j tj j

ç y
jt y
j r
J* J* y j j
jLjJ* J* jj ( )
11 12 …(13.9)

j j
21 22
t j
where J* , J* , J* and J* are the elements of Jacobian inverse matrix. Since for a given point Jacobian
matrix is11known
12 its
21 inverse2can be calculated and Jacobian inverse matrix is assembled. With this
2
transformation relation known, we can expresses derivatives of the displacements as shown below:

y rJ
u y ç u y
j j *
J12
* 0 0
jjj jjuj jj j
ju j jj
11
x

j j j J
y j
*
J
* 0 0
j( j)
 ( j )j  j 0
21 22

v 0 J *j
j J j v j * …(13.10)

jxv j 11
j jjv j 12
jyjjj j
j 0 0 J J jj j j
* *

L 21 t j 22

The strain displacement relation is given by

çu y
jjx jju
x y r 1 0 0 0 y
()  (  y j)  jj00 10 10 0j1jj (vj y j)
j j j j
t jL xy jx
j v
jt
jj
y

çju y
ju j
rJ *
J* 12 0 0 y j
 jj 0 Jj j
J jj ( )j
j
11
0 *
*  **

LJ
*
J* J …(13.11)
21 22 v
21 22 11 12
 Isoparametric 2
 t j
j j jv
j
 2 Finite Element

4 4

But u   Ni and v  Nv

i 1
i i

ui
i1

çu y rN
y çjvu1 yj
N
N N 0
3
0 4
j u
jj j j jj j
1 2 1
  0  0   2
u N N2 N3 N4 0
j 1
0 0 0

j  j)  j
 (
 
j(v j)

N

N2 N3
j 2

jv j j 1
0

0
 0  j jv j
N u
4 3

N j j j
 N2 0 3
0  0 0 N3 4
4
v u
j0 j j N
1 j
t j L     j j
v
t j 4

Substituting it in equation 13.11 strain displacement matrix [B] is obtained as,

r N1 N2 N3 N4

y
rJ j
0 yj
0  0  0 
0
j
*
J * 0
N j 0
N2
0
N3
0
N4 0
jj
jJ
1

j …(13.12)
(B  jj 0
11 12
J j j 0
0 * 
*22 N1  N2  N3  N
*21 * 0 0
j  j
0 4
J J
 
) LJ
* *
J22 
N j
21 11 12
jj N N2 N3 0
4
1 0  0  
L
0
 j
Then element stiffness matrix is given by

In this case,
k
j B T
D B dV

where t is the thickness.

k t
T
JJ B D B x y …(13.13)

It can be shown that elemental area in Cartesian coordinates (x, y) can be expressed in terms of the area
in local coordinates (, ) as
xy  J   …(13.14)
Where |J| is the determinant of the Jacobian.

 k t
JJ B T D B J   …(13.15)
 Isoparametric 2
Integration is to be performed so as to cover entire area i.e. the limit of integration is from  is form –1
to 1 and  is also from –1 to 1. It is difficult to carryout all the multiplications in equation 13.15 and then
the integration. It is convenient to go for numerical integration.
 2 Finite Element

13.7 NUMERICAL INTEGRATION

1 3
In mathematics numerical integration techniques like trapezoidal rule, Simpson’s rd rule, Simpson’s th
3 8
rule and Gauss quadrate formula are available. In trapezoidal rule, the variation of the function between two
1
sampling point is assumed linear. In Simpson’s rd rule 3 sampling points are used and second degree curve
3
3
is fitted. In Simpson’s th rule 4 sampling points are selected and 3rd degree (cubic) curve is fitted. All
these
8
methods are based on Newton Cotes formula in which values at n equally spaced sampling points are
required to fit in n – 1 degree variation curve.
In finite element analysis Gauss quadrate formula is preferred since in this values at n sampling points
can be used to fit in 2n – 1 degree variation, as the evaluation of functions like BTDBdV is a time
consuming process. In this method, the numerical integration is achieved by the following expression.
1

J()
n
f     wi f …(13.16)
(i )
1 i1

Where wi – weight function and f ( i ) is values of the function at pre determined sampling points.
In Gauss quadrature formula sampling points are cleverly placed. In this, both n sampling points and n
weights are treated as variables to make exact 2n – 1 degree polynomial. This is an open quadrature
formula, the function values need not be known at end points but they must be known at predetermined
sampling points.
The location of sampling points i and weight function wi are determined using Legendre polynomials.
Hence this method is some times called as Gauss Legendre quadrature formula. Table 13.1 shows location
of
Gauss sampling points (i ) and corresponding weight function (wi) for different number (n) of Gauss
integration scheme.

Table 13.1 Location of sampling points and weight functions

in Causs Integration J
1

1
()
f   
i1
n

W i f

n  W1

1 1  0.00000000 W1 = 2.00000000

2 1   2  0.57735027 W1 = W2 = 1.00000000

3 1   3  0.77459667 W1 = W3 = 0.55555556

 2  0.00000000 W2 = 0.88888889

4 1   4  0.86113631 W1 = W4 = 0.34785485

 2   3  0.33998104 W2 = W3 = 0.65214515
 Isoparametric 2

Fig. 13.7 shows the integration scheme for 3 point Gauss integration. It may be noted that the sampling
points are symmetrically placed, all weight are positive and the weights of symmetrically placed points are
same. Depending upon the degree of variation, the number of Gauss points n can be chosen so that
integrations are exact. Since in finite element analysis, the exact degree of variation of the functions like
element stiffness matrix are not known, preliminary investigations may be made by changing n to get stable
result. Many investigators have reported that two point house integration is more than sufficient. The above
scheme may be extended to 2 and 3 dimensional problems also.

f1 ( 1) f2 ( 0) f3 ( 3)

–1 –0.77459667 0 0.77459667 1

Fig. 13.7 Scheme of 3 point Causs integration

For two dimensional problem n = 2 means 2 × 2 = 4 Gaussian points and for three dimensional
problems it works out to be 2 × 2 × 2 = 8. Thus,

(–1, 1) (1, 1) (–1, 1) (1, 1)

2 , 3
1 , 3 3 , 3
( 1, 2 ) ( 2, 2)

2 , 2

1 , 2 3, 2

( 1, 1) ( 2, 1)

1 , 1 2, 1 3 , 1

(–1, –1) (1, –1) (–1, –1) (1, –1)

(a) 2 × 2 points (b) 3 × 3 points

Fig. 13.8 Causs sampling points in two dimensional elements

JJ 1 1

( )
f ,  d d J 1 n
Wi f  , (
)
1 1 1i  1
  d
 2 Finite Element

ç y
 W (j  W f (, )j)
n n


tj
j i

j1 i1 j
n n

   Wj Wi f (, )
j1 i1

For a two dimensional problem, Gauss points for n = 2 and n = 3 are shown in Fig. 13.8.

13.8 NUMERICAL EУAMPLES

Finite element analysis using isoparametric element involves too many calculations and hence not suitable
for hand calculations. For such analysis one has to go for computer analysis by developing programs. With
slight additional statements various elements can be easily incorporated in a single analysis package.
However to make the procedure clear to learners of finite element analysis a small numerical problem is
taken up here.
Example 13.1: Assemble Jacobian matrix and strain displacement matrix corresponding to the Gauss point
(0.57735, 0.57735) for the element shown in Fig. 13.9. Then indicate how do you proceed to assemble
element stiffness matrix.

4 3

10 mm
60°
1 2
60 mm

Fig. 13.9

Solution: The coordinates of node points in Cartesian system are (0, 0), (60, 0), (65.7735, 10) and (5.7735, 10).
The shape functions are
N 
i
1
(1   )(1   )
i i
4
1
N   1  i
 i 4
i ( )
and
Ni

1
 1  ( i
)
 4 i
 Isoparametric 2


N1

1
(1   ) 
1
(1  0.57735)   0.10566
 4 4

N2

1
(1   ) 
1
(1  0.57735)  0.10566
 4 4

N3

1
(1  ) 
1
(1  0.57735)  0.39438
 4 4

N4

1
(1   ) 
1
(1  0.57735)   0.39438
 4 4
Similarly,
N1 1
(1   )
1  0.57735
1
( )   0.10566
   4   4
N 2

1
1 (

1
)
1  0.57735 ( )   0.39438
 4 4
N 3

1
(
1 
1
)
1  0.57735  ( ) 0.39438
 4 4
N 4

1
(
1 
1
)
1  0.57735  ( ) 0.10566
 4 4
The Jacobian Matrix is given by
rj Ni x  Ni y yj
  j i i

J j
j N x N
yji i

jL  
j i

i

 J11 

 N
 x
i
i

= –0.10566
Ni × 0 + 0.10566 × 60 + 0.39438 × 65.7735 – 0.39438 × 5.7735 = 30.0000
J  y
12   i = 0 + 0 + 0.39438 × 10 – 0.39438 × 10 = 0

Ni
J21   xi = 0 – 0.39438 × 60 + 0.39438 × 65.7735 + 0.10566 × 5.7735 = 2.88698

Ni
J22   y
 i = 0 + 0 + 0.39438 × 10 + 0.10566 × 10 = 5.0000
 2 Finite Element

r 30.0000 0 y
 J  j j
L2.88698 5.0000
1 2.88698 y 0.019246 y
r5.0000 r
* 1

0.033333
J
jL
 J j 
0 0.166667
j
30.0000  5.0000
jL
0 30.0000

The strain displacement matrix is given by

rN1 N2 N3 y

jj  0  0
 0
N4
 0
y N4 j
rJ *
J*
0
N
0
j 0
N2
0
N3 0 0
j
j j J j 
1

B j 0
11 12

( j 0 J
jj 0   
j
*
*22 N1 N2 N3 N
*21 * 0 0
 j
J J 0 4
 
) LJ
* *
J22 
N j
21 11 12
jj N N2 N3 0
4
0 0
L
0
1

  
j
Where J*ij are the elements of Jacobian inverse matrix,

r 0.033333 0.019246 0 0 y
 B jj 0 0 0 0.166667  jj
L 0 0.166667 0.033333 0.019246

r0.10566 0 0.10566 0 0.39438 0 0.39438 y 0

jj0.10566 0 0.39438 0 0.39438 0 0.10566
jj 0

j 00 0.10566 0 0.10566 0 0.39438 0 0.39438

j
L 0.10566 0 0.39438 0 0.39438 0 0.10566

r 1.48843 10 y
j j
3
0 0.011112 0 5.55563 103 0 0.022915 0
j 0 0.01761 0 0.06573 0 0.06573 0 0.01761
jj
j 0.01761 0.06573 0.011112 0.06573 5.55563 103 0.01761 0.015179
L 1.48843  103

Answer

Assembling Element Stiffness Matrix

D matrix can be assembled for the plane stress or plane strain as the case may be, by using material
properties
t
2
of the structure. Then the following matrix multiplication may be carried out to assemble B T
D B for

the Gauss point (0.57735, 0.57735). This value when multiplied with weight function (in this case 1) gives
the contribution of the Gauss point (0.57735, 0.57735) to the element stiffness matrix.
 Isoparametric 2

Assembling of the Jacobian matrix and [B] matrix may be carried out for the Gauss point (0.57735,
–0.57735) and then its contribution to stiffness matrix may be found. On the same line assemble the
contribution of remaining two Gauss points. Addition of all the four Gauss points contribution gives the
element stiffness matrix of size 3 × 8.

Example 13.2: Determine the Cartesian coordinate of the point P(  0.5,   0.6) shown in Fig. 13.10.

(7, 7)
(3, 5)
P

(8, 3)
(2, 1)
x
0

Fig. 13.10

Solution: It is given that

  0.5 and   0.6
(1  )(1  ) (1  0.5)(1  0.6)
N1    0.05
4 4
(1  ) (1  ) (1  0.5) (1  0.6)
N2    0.15
4 4
(1  ) (1  ) (1  0.5)(1  0.6)
N3    0.6
4 4
(1   )(1  ) ( )(
1  0.5 1  0.6 )
N4    0.2
4 4

x   Ni xi = 0.05 × 2 + 0.15 × 8 + 0.6 × 7 + 0.2 × 3 = 6.1

y  Ni yi = 0.05 × 1 + 0.15 × 3 + 0.6 × 7 + 0.2 × 5 = 5.7
The Cartesian coordinates of point P are (6.1, 5.7)
Example 13.3: In the element shown in Fig. 13.10, P is the point (6, 5). On this point the load components
in
x and y directions are 8 kN and 12 kN respectively. Determine its nodal equivalent forces.
Solution: We have to first determine the local natural coordinates of point P. We know

x  Ni xi and y   Ni yi
 2 Finite Element

For the quadrilateral element

(
Ni  1   i 1   i )(
) (1  ) (1  )
4
( )(
1 1 )
i.e., N1  N2 
4 4

N3 
(1  ) (1  ) N4 
(1   ) (1   )
4 4

x   Ni xi gives
1
6
4 (1   ) (1  ) 2  (1   ) (1  ) 8  (1  ) (1  ) 7  (1   ) (1  ) 3
 24  2 (1      )  8 (1      )  7 (1      )  3 (1      )
 20  10  0  2
4  10  2 …(1)
or 2  5  

y  Ni yi gives
5
1
4
(
1 1       ) (
 3 1       ) (
 7 1       ) (
 5 1       )
20  16  4  8
or 4  4  8
or 1    2 …(2)
1
From equation (2),  
2

j 1   7j
Substituting it in equation 1, we get
2  5  
J
2 J
or 4  10   1   ( )  9   2

i.e.  2  9  4  0

9 92  4  4
  0.42443
2
1  0.42443
   0.28779
2
 Isoparametric 2

Now, the equivalent load is given by

ç Xy
(F) (tY j
T
N

For Point P.

)
N1 
(1  0.42443) (1   0.10248
0.28779)
4
N2   0.25362
( )(
1  0.42443 1 

N3  0.28779 )  0.45859
4

(
N4  1  0.42443 1  )(  0.18530

0.28779 )
4

(1  0.42443)(1 
0.28779)
4
{F } = [N] {X} T

y r N1 y 0.10248y ç 0.81984y
Fx1
x

F j
j N j j 0.25362j j 2.02896j
)jFF jj jjN (8)  8 jj(0.4585
x2
x3
2 3
j) 3.66872
j
j(
j1.48240
j)
j 9
0 0.1853
t j LNx4 t
4 j t j
F y ç 0.10248 y ç 1.22976 y
Fy2 j
j
y1
j j j j 3.04344 jj
y3
j0.25362 j
)F  ( j0.45859 ) (12) 5.50308
( )
jtF jt0.1853
y4
jt2.22360jj
jj 0 j
Example 13.4: The quadrilateral element shown in Fig. 13.11 is 20 mm thick and is subjected to surface
forces T and T . Determine expressions for its equivalent nodal forces. If T = 10N/mm2 and T = 15
N/mm2, x y x y
determine the numerical values of the nodal forces.

y
Ty
3, (700, 700)
4 (300, 500) Tx
All linear dimensions
In mm
t = 20 mm
2 Finite Element 2 (800, 300)
1 (200, 100)
x
0

Fig. 13.11
 Isoparametric 2

Solution: The element is subjected to load along edge 3–4. We know along edge 3–4,   1

 N1

4
1
(1  )(1  ) 0

N2

1
4
(1   )(1  )  0
N
3

1
(1   )(1  )  1 (1   )
4 2

N 
1
(1  )(1  ) 
1
(1   )
4 4
2
Nodal forces are given by the expressions like

(Fx )  J N (T )ds T
x

J
 t N (T )dl
T
x

We know,

l jj x 7j jj y 7j
(x)  (y)and  
2 2
 j   j
l  2 2

In isoparametric concept, we know

x Nx
Ny
i i
and y  i i

In this case, along line 3–4,

x00
1
2
(1   ) x  (1   ) x
1
2
3 4

 In limiting case,
dx

x

1
(x 
x )
d y 2
3 4

Similarly, y00
1
(1   ) y  1 (1   ) y
3 4
2 2
 In limiting case
dy

1 
y )
(y
4
d 2 3

dl l
   ç( 1 ( )y)ç( 1 ( )y) 1 l
t2 x  x j  t2 y  y j 2
2 2
34
d  3 4 3 4
 2 Finite Element

1
or dl  l d

J
34
2
 Fx ( ) t ()
N T dl
T
x

t J jrj )(
1 00
1 1 )yjj (T x 1
l34 d
t
l
34
1 jr j d
0

j2 j y
2
j J j(11  )TTx j
1 jj1 (1  )jj 4
jL( ) j
L2 1
x

For uniformly distributed load, Tx is constant,

ç Fx1 y ç 0j)y
j j j sincej (1   ) d  (1   )
j( j(0F j)  t l34
x2

4 j2T
J J 1 1
d  2

F x3 x

jt j j2T j
1 1

jF
x4
t j y

ç 0y
t l j j j0 j
Similarly 
34
( )
2 jTx
jtT j
y

j
ç Fy1 y ç 0y
jF j
j( j) N (T )ds  t l34 jj 0 jj
y2
T

( )
jjFJ j

y3
j
2 jTx j
y

tj j
F y4 tjT y

In this problem, j
l34 

(700  300)  (700  500) 447.21 mm

2 2
t = 20 mm
Tx = 10 N/mm2

ç Fx1 y ç 0 y
jj(F j)j  j(j 0 jj)
jF j jj44721.36
x2
 F x3
44721.36
t j t x4 j
 Isoparametric 2

çF y ç 0 y
jj(F jF )j j 0 jj)
y1

y2 j
 (
y3

jtF jj67082.04
y4 t67082.04
jj

QUESTIONS

1. Explain the isoparametric concept in finite element analysis.

2. State and explain the three basic laws on which isoparametric concept is developed.
3. Discuss the convergence criteria for isoparametric elements.
4. Explain the terms isoparametric, subparametric and superparametric elements.
5. Write short notes on
(a) Uniqueness of mapping of isoparametric elements.
(b) Jacobian matrix
(c) Gaussian quadrature integration technique.
6. Explain the isoparametric elements and their advantages.
7. For the isoparametric quadrilateral elements shown in Fig. 13.12, determine
(a) Cartesian coordinates of the point P which has local coordinates   0.57735 and

  0.57735
(Ans. x = 6.36603, y = 4.75088)
(b) Local coordinates of the point Q which has Cartesian coordinates (7, 4)
(Ans.   0.91255 ,   0.21059 )
y
4 (–1, 1) 3 (8, 6)
3 (1, 1)
4 (2, 5)
P. Q.

1 (– 1, –1) 2 (1, –1) 1 (3, 1) 2 (6, 1)

x
0

Fig. 13.12

8. For the element shown in Fig. 13.13, assemble Jacobian matrix and strain displacement matrix
for the Gaussian point (0.57735, 0.57735).