FE FORMULATION FOR

STRUCTURAL 1D BAR
[Subjected to axial loading]
Concept of Interpolation Function:
Trial solution represented in terms of functional values at nodal end points with the help of
interpolation functions: [1-x/l] and [x/l]
% age contribution of the function value uk to the PCT
values within the sub-domain: Interpolation function-1

uˆk ( x)  [1  x / l ]  uk  [ x / l ]  uk 1

% age contribution of the function value uk 1 to

the PCT values within the sub-domain: Interpolation
function-2 kth
Local coordinate system (x) uˆk ( x) x l / 2  [1  l / 2l ]  uk  [l / 2l ]  0.5  uk  0.5  uk 1
uˆk ( x) x 0  [1  0 / l ]  uk  [0 / l ]  uk 1  1 uk  0  uk 1

uˆk ( x) x l  [1  l / l ]  uk  [l / l ]  0  uk  1 uk 1 uˆk ( x) x l / 3  [1  l / 3l ]  uk  [l / 3l ]  0.7  uk  0.3  uk 1

A generic representation of interpolation of the field variable within an element valid for any
 x x
number of elements may be developed as: uk ( x)  [1  ]uk  [ ]uk 1  [ N ]{u}k
Where, [N]: is the Shape function array and {u(cap)}k: is l l
the DOF array k u
 x x  x x
uk ( x)  [1  ]uk  [ ]uk 1  [ N ]{u}  uk (x) 
k
[(1  ) ( )]
l l l l uk 1
As the Trial solution contains two unknown coefficients; we need two numbers of weighing
functions for formation of weak form based WRS. These two numbers of weighing functions are
to be selected from the expression of the trial solution.
 The trial solution can be represented in terms of generalized form as below:
x x
ˆuk ( x)  uk [1  ]  uk 1[ ]    C1 N1  C2 N 2
l l
 x x
uk ( x)  uk [1  ]  uk 1[ ]  C1 N1  C2 N 2
x l l
W1 ( x)  N1  [1  ] 0 < x < l
l
x
W2 ( x)  N 2  [ ] 0 < x < l
l
Step 5: Solving for unknown coefficients by putting the trial solution in the weak form
Accordingly Weak form of the Weighted Residual Statement for the kth finite element can be
converted from GCS to LCS as below:
X k 1 X
duˆ ( X ) X k 1 duˆk dWk k 1

 [Wk ( X ) AE k
dX
]X k  
Xk
AE 
dX dX
 dX   Wk ( X )  q( X )  dX  0
Xk
Substituting: X  x  Xk X  X k : x  0; X  X k 1 : x  X k 1  X k  l
duˆk ( x) l l duˆk dWk l

 [Wk ( x) AE ]0   AE   dx   Wk ( x)  q( x)  dx  0
dx 0 dx dx 0
Now lets us try to calculate the unknown coefficients by solving
the weak formulation based WR statements.

kth
 duˆk ( x) l l duˆk dWk l

 [Wk ( x) AE ]0   AE   dx  Wk ( x)  q( x)  dx  0
dx 0 dx dx 0
 x x
As the Trial solution contains two unknown coefficients: uk ( x)  [1  ]uk  [ ]uk 1
l l
Weak form of the weighted residual statement has to be made into two different equations:
l  l 
duk dW1 duk l
0 AE dx  dx  dx  0 W1 ( x)  q( x)  dx  [W1 ( x) AE dx ]0 .........(1)
l  l 
duk dW2 duk l
0 dx dx
AE   dx  0 2
W ( x )  q ( x )  dx  [W 2 ( x ) AE
dx
]0 .........(2)
x x  x x
Substituting: W1 ( x)  N1  [1  ] W2 ( x)  N 2  [ ] uk ( x)  [1  ]uk  [ ]uk 1
l l  l l
u  u 1
l l
du dW
LHS Equation1   ( AE k )( 1 )dx   AE k 1 k ( )dx
0
dx dx 0
l l
  
uk 1  uk  1 u
AE   AE k
 AE ( )l  (uk  uk 1 )  [1  1] 
l l l l uk 1
  
uk 1  uk 1
l l
duk dW2
LHS Equation2   ( AE )( )dx   AE ( )dx
0
dx dx 0
l l
   
uk
u u 1 AE   AE
 AE k 1 k ( )  l  (uk 1  uk )  [1 1] 
l l l l uk 1
l l
x q0l
RHS Equation1 firstterm
 0 1
W ( x )  q0  dx  0 l 0
[1  ]  q  dx 
2
l l
x q0l
RHS Equation 2 firstterm
 0 2
W ( x )  q0  dx  0 l 0
[ ]  q  dx 
2

duk l x
RHS Equation1 sec ond term  [W 1( x) AE ]0  [(1  ) P]l0   P0
dx l

duk l x
RHS Equation2 sec ond term  [W 2( x) AE ]0  [( ) P]l0  Pl
dx l
We have used the fact that the term: AE(du dx) stands for the axial force “P” in the bar at a
particular cross section. Thus P0 and Pl stands for the axial forces at either end of finite element.
The above weak form equations have been brought to the form:

uk q0l
AE   q0l AE   P0
(uk  uk 1 )   P0 [1  1] 
l 2 l uk 1 2
AE   q0l 
(uk 1  uk )   Pl uk q0l
l 2 AE   Pl
[1 1] 
l uk 1 2