International Journal of Mathematics Trends and Technology (IJMTT) – Volume 51 Number 6 November 2017

Deriving Shape Functions for 9-Noded

Rectangular Element by using Lagrange
Functions in Natural Coordinate System and
Verified
P. Reddaiah#1
#
Professor of Mathematics, Global College of Engineering and Technology, kadapa, Andhra Pradesh, India.

Abstract — In this paper, I derived shape functions II. GEOMETRICAL DESCREPTION

for 9-noded rectangular element by using Lagrange
functions in natural coordinate system and also I Typical nine noded element is shown in Fig1.1
verified two verification conditions for shape
functions. First verification condition is sum of all
the shape functions is equal to one and second
verification condition is each shape function has a
value of one at its own node and zero at the other
nodes. For computational purpose I used
Mathematica 9 Software [2].

Keywords — Rectangualar element, Lagrange

functions, Shape functions.

Fig.1 Typical nine noded rectangular element.

1. INTRODUCTION
III. DERIVING SHAPE FUNCTIONS FOR
In engineering problems there are some basic NINE NODED RECTANGULAR ELEMENT
unknowns. If they are found, the behaviour of the BY USING LAGRANGE FUNCTIONS.
entire structure can be predicted. The basic
unknowns or the field variables which are The natural coordinates of various nodes
encountered in the engineering problems are are as shown in the Fig.1.
displacements in solid mechanics, velocities in fluid
Nodal unknowns
mechanics, electric and magnetic potentials in
Basic unknowns may be displacements
electrical engineering and temperatures in heat flow
for stress analysis, temperatures for heat flow
problems.
problems and the potentials for fluid flow or in the
magnetic field problems. In the problems like truss
In a continuum, these unknowns are infinite. The
analysis, plane stress and plane strain, it is enough if
finite element procedure reduces such unknowns to a
the continuity of only displacements are satisfied,
finite number by dividing the solution region into
since there is no change in the slopes at any nodal
small parts called elements and by expressing the
point. Such problems are classified as ‘Zeroth’
unknown field variables interms of
Continuity problems and are indicated as C0 –
assumed approximating functions (Interpolating
Continutiy problem.
functions/Shape functions) within each element. The
For the C0 Continutiy element in two
approximating functions are defined in terms of field
dimensions
variables of specified points called nodes or nodal
points. Thus in the finite element analysis the Ni  Li ( ) Li ( ) (1)
unknowns are the field variables of the nodal points. Where Li refers to Lagrangian function
Once these are found the field variables at any point
can be found by using interpolation functions/Shape at node i.
functions.

ISSN: 2231-5373 http://www.ijmttjournal.org Page 429

 International Journal of Mathematics Trends and Technology (IJMTT) – Volume 51 Number 6 November 2017

In Fig.1 there are 3 nodes in each direction.

N2 
  1 (  1)  (  1)
Hence n=3 in Lagrange function. 1 2
Lagrange Polynomial in one dimension
N2 
  1 (  1) (  1) (3)
is defined by 2
When i=3
n
x  xm n
y  ym
L k ( x)   , Lk ( y)   n=3, k=3, x= , y  
xk  xm m 1 yk  ym
m 1
mk mk
(1)  N3  L3 ( ) L3 ( )
At node 3, Along  -axis nodes 3,1,2
When i=1  For L3 ( ) we should take nodes 1 & 2 
n=3, k=1, x= , y   At node 3, Along  -axis nodes 3,6,9

(1)  N1  L1 ( ) L1 ( )
 For L3 ( ) we should take nodes 6 & 9 
At node 1, Along  -axis nodes 1,2,3 (  1 )(   2 ) ( 6 )( 9 )
N3 
(3  1 )(3  2 ) (3 6 )(3 9 )
 For L1 ( ) we should take nodes 2 & 3
At node 1, Along  -axis nodes 1,4,7 (   1)(  0) (  0)(  1)
N3 
(1   1)(1  0) (1  0)(1  1)
 For L1 ( ) we should take nodes 4 & 7 
(  2 )(  3 ) ( 4 )( 7 ) N3 
  1   (  1)
N1  1  11  1 (2)
(1  2 )(1  3 ) (1  4 )(1  7 )
(  0)(  1) (  0)(  1) N3 
  1   (  1)
N1 
(1  0)(1  1) (1  0)(1  1)  2 1 2
N1 
 (  1)  (  1)
N 
  1  (  1) (4)
1(2) 1(2) 3
4
 (  1)  (  1) When i=4
N1 
2 2 n=3, k=4, x= , y  
 (  1) (  1) (1)  N4  L4 ( ) L4 ( )
N1  (2)
4 At node 4, Along  -axis nodes 4,5,6
When i=2  For L4 ( ) we should take nodes 5 & 6 
n=3, k=2, x= , y  
At node 4, Along  -axis nodes 4,1,7
(1)  N2  L2 ( ) L2 ( )
At node 2, Along  -axis nodes 2,1,3  For L4 ( ) we should take nodes 1 & 7 
(  5 )(  6 ) ( 1 )( 7 )
 For L2 ( ) we should take nodes 1 & 3 N4 
(4  5 )(4  6 ) (4 1 )(4 7 )
At node 2, Along  -axis nodes 2,5,8
(   0 )(  1) (   1)(  1)
 For L2 ( ) we should take nodes 5 & 8 N4 
(1   0 )(1   1) (0   1)(0  1)
(  1 )(  3 ) ( 5 )( 8 )
N2     1   1 (  1)
(2  1 )( 2  3 ) (2 5 )(2 8 ) N4 
(   1)(  1) (  0)(  1)  1 1  1  0  1 (0  1)
N2     1   1 (  1)
(0   1)(0  1) (1  0)(1  1) N4 
  1 (  1)  (  1)  1 2  1 1
N2     1  1 (  1)
 0  1 1  1 (2) N  (5)
2
3

N2 
  1 (  1)  (  1) When i=5
1 1 2 n=3, k=5, x= , y  

ISSN: 2231-5373 http://www.ijmttjournal.org Page 430

 International Journal of Mathematics Trends and Technology (IJMTT) – Volume 51 Number 6 November 2017

(1)  N5  L5 ( ) L5 ( ) At node 7, Along  -axis nodes 7,4,1

At node 5, Along  -axis nodes 5,4,6  For L7 ( ) we should take nodes 4 & 1
 For L5 ( ) we should take nodes 4 & 6  N7 
(  8 )(  9 ) ( 1 )( 4 )
At node 5, Along  -axis nodes 5,2,8 (7  8 )(7  9 ) (7 1 )(7 4 )
 For L5 ( ) we should take nodes 2 & 8 (  0)(  1) (   1)(  0)
N7 
(  4 )(  6 ) ( 2 )( 8 ) (1  0)(1  1) (1   1)(1  0)
N5 
(5   4 )(5  6 ) (5  2 )(5  8 )    1   1
N7 
(   1)(  1) (   1)(  1)  1 2  1  1 (1)
N5 
(0   1)(0  1) (0   1)(0  1)    1   1
N7 
N5 
  1  1   1 (  1)  1 2  1  11
 0  1 1  0  1 (1)    1    1
N7 
N5 
  1  1   1 (  1)  1 2  2 1
1 1 1 1    1  1
N  (8)
N 
  1  1  1 (  1) (6)
7
4
5
1 When i=8
n=3, k=8, x= , y  
When i=6 (1)  N8  L8 ( ) L8 ( )
n=3, k=6, x= , y   At node 8, Along  -axis nodes 8,7,9
(1)  N6  L6 ( ) L6 ( )  For L8 ( ) we should take nodes 7 & 9 
At node 6, Along  -axis nodes 6,4,5 At node 8, Along  -axis nodes 8,5,2
 For L6 ( ) we should take nodes 4 & 5  For L8 ( ) we should take nodes 5 & 2 
At node 6, Along  -axis nodes 6,3,9 (  7 )(  9 ) ( 2 )( 5 )
N8 
 For L6 ( ) we should take nodes 3 & 9  (8  7 )(8  9 ) (8  2 )(8  5 )
(  4 )(  5 ) ( 3 )( 9 ) (   1)(  1) (   1)(  0)
N6  N8 
(6  4 )(6  5 ) (6  3 )(6  9 ) (0   1)(0  1) (1   1)(1  0)
(   1)(  0) (   1)(  1)
N6  N8 
  1  1   1
(1   1)(1  0) (0   1)(0  1)
 0  1 1 1  1 (1)
N6 
  1    1 (  1)   1  1   1
1  11  0  1 (1) N8 
1 1  2 1
N6 
  1    1 (  1)   1  1  1
 2 1 1 1 N  (9)
2
7

N 
  1    1 (  1) (7)
When i=9
6
2 n=3, k=9, x= , y  
When i=7 (1)  N9  L9 ( ) L9 ( )
n=3, k=7, x= , y   At node 9, Along  -axis nodes 9,7,8
(1)  N7  L7 ( ) L7 ( )  For L9 ( ) we should take nodes 7 & 8
At node 7, Along  -axis nodes 7,8,9 At node 9, Along  -axis nodes 9,6,3
 For L7 ( ) we should take nodes 8 & 9   For L9 ( ) we should take nodes 6 & 3

ISSN: 2231-5373 http://www.ijmttjournal.org Page 431

 International Journal of Mathematics Trends and Technology (IJMTT) – Volume 51 Number 6 November 2017

(  7 )(  8 ) ( 3 )( 6 ) (  1)(  1) (  1)

N9  N2 
(9  7 )(9  8 ) (9 3 )(9 6 ) 2
(   1)(  0) (   1)(  0) (1  1)(1  1)(1)(1  1)
N9  N2 
(1   1)(1  0) (1   1)(1  0) 2
N2  0
N9 
  1    1 (  1)( ) (  1)
1  11 1  1 (1) N3 
4
N9 
  1    1 N3 
(1  1)(1)(1)(1  1)
 2 1  2 1 4
  1    1 N3  0
N 
9 (10)  (  1)(  1)(  1)
4 N4 
IV. VERIFICATION 2
1(1  1)(1  1)(1  1)
(I a) Verification 1st Point N4 
N1  N 2  N3  N 4  N5  N 6  N 7  2
N4  0
 (  1) (  1)
 N8  N 9  (  1)(  1)(  1)(  1)
4 N5 
1
(  1)(  1) (  1) (  1) (  1)
   (1  1)(1  1)(1  1)(1  1)
2 4 N5 
1
 (  1)(  1)(  1) N5  0
 
2
(  1) (  1)(  1)
(  1)(  1)(  1)(  1) N6 
  2
1
(1  1)(1)(1  1)(1  1)
(  1)( )(  1)(  1)  (  1)(  1) N6 
  2
2 4 N6  0
(  1)(  1)(  1) (  1) (  1)
   (  1)(  1)
2 4 N7 
N1  N 2  N3  N 4  N5  N 6  N 7  4
(1)(1  1)(1  1)(1)
 N8  N 9  1 N7 
4
 Sum of all the shape functions is N7  0
equal to one. (  1)(  1)(  1)
N8 
1st point is verified. 2
(1  1)(1  1)(1  1)(1)
(II a) Verification 2nd Point
N8 
2
(i) At Node 1 (-1,-1) N8  0
  1,  =-1
 At Node 1 N1  1, N 2  0, N3  0,
when  =-1,  =-1
 (  1) (  1) N 4  0, N5  0, N 6  0, N 7  0,
N1  N8  0, N9  0
4
1(1  1)(1)(1  1)  At Node 1 2nd condition is verified.
N1  
4 (ii) At Node 2 (0,-1)  =0  =-1
1(2)(1)(2) 4 N1 N2 N3 N4 N5 N6 N7 N8 N9
  1
4 4 Output 0 1 0 0 0 0 0 0 0
N1  1

ISSN: 2231-5373 http://www.ijmttjournal.org Page 432

 International Journal of Mathematics Trends and Technology (IJMTT) – Volume 51 Number 6 November 2017

N1 N2 N3 N4 N5 N6 N7 N8 N9
(iii) At Node 3 (1,-1)  =1, =-1
Output 0 0 0 0 0 0 0 1 0
N1 N2 N3 N4 N5 N6 N7 N8 N9
Output 0 0 1 0 0 0 0 0 0 (ix) At Node 9(1,1)  =1, =1
N1 N2 N3 N4 N5 N6 N7 N8 N9
(iv) At Node 4(-1,0)  =-1, =0
Output 0 0 0 0 0 0 0 0 1
N1 N2 N3 N4 N5 N6 N7 N8 N9
Output 0 0 0 1 0 0 0 0 0 V. AUTHOR’S CONTRIBUTION

1. Deriving Shape functions for 9 noded Lagrange

(v) At Node 5(0,0)  =0, =0 element in natural coordinate system by Lagrange
method.
N1 N2 N3 N4 N5 N6 N7 N8 N9 2. Sum of all the shape functions is equal to one.
3. Each Shape function has a value of one at its own
Output 0 0 0 0 1 0 0 0 0 node and zero at the other nodes.

(vi) At Node 6(1,0)  =1,  0

REFERENCES
[1]. S.S. Bhavikatti, Finite Element Analysis, New
N1 N2 N3 N4 N5 N6 N7 N8 N9 Age International (P) Limited, Publishers,
nd
2 Edition, 2010.
Output 0 0 0 0 0 1 0 0 0 [2]. Mathematica 4, Version number
4.1.0.0 Software, Wolfram Research
(vii) At Node 7(-1,1)  =-1, =1
1988-2000.

N1 N2 N3 N4 N5 N6 N7 N8 N9
Output 0 0 0 0 0 0 1 0 0

(viii) At Node 8(0,1)  =0, =1

ISSN: 2231-5373 http://www.ijmttjournal.org Page 433