Axially loaded elastic bar

y A(x) = cross section at x

b(x) = body force distribution
F (force per unit length)
x E(x) = Young’s modulus

FEM for 1D Elasticity x=0

x
x=L
Problem Potential energy of the axially loaded bar corresponding to the
exact solution u(x)
2
1 L  du  L
 (u)   EA  dx   bu dx  Fu(x  L)
2 0
 dx  0

Potential energy of the bar corresponding to an admissible

displacement w(x)
2
1 L  dw  L
 (w)  
2 0
EA 
 dx 
dx   bw dx  Fw(x  L)
0

Rayleigh-Ritz Principle
Finite element formulation, takes as its starting point, not the Step 1. Assume a solution
strong formulation, but the Principle of Minimum Potential
Energy. w( x)  a 0 o ( x)  a11 ( x)  a 2 2 ( x)  ...
Task is to find the function ‘w’ that minimizes the potential energy Where o(x), 1(x),… are “admissible” functions and ao, a1,
of the system etc are constants to be determined.
2 Step 2. Plug the approximate solution into the potential energy
1 L  dw  L
 (w)  
2 0
EA 
 dx 
dx   bw dx  Fw(x  L)
0 2
1 L  dw  L
 (w)  
2 0
EA 
 dx 
dx   bw dx  Fw(x  L)
0

From the Principle of Minimum Potential Energy, that function ‘w’ Step 3. Obtain the coefficients ao, a1, etc by setting
is the exact solution.  (w)
 0, i  0,1,2,...
ai

1
 The approximate solution is Need to find a systematic way of choosing the approximation
functions.
u ( x)  a 0 o ( x)  a11 ( x)  a 2 2 ( x)  ...
Where the coefficients have been obtained from step 3 One idea: Choose polynomials!

w( x)  a 0 Is this good? (Is ‘1’ an “admissible” function?)

w( x)  a1 x Is this good? (Is ‘x’ an “admissible” function?)

Finite element idea: x1=0 x2 x3 x4=L

Step 1: Divide the truss into finite elements connected to each El #1 El #2 El #3

other through special points (“nodes”)
Total potential energy
2
1 2 3 4 1 L  dw  L
 (w)  
2 0
EA 
 dx 
dx   bw dx  Fw(x  L)
0
El #1 El #2 El #3
Potential energy of element 1:
2
Total potential energy=sum of potential energies of the elements 1 x2  dw  x2
 1 (w)  
2 x1
EA 
 dx 
dx   bw dx
x1
2
1 L  dw  L
 (w) 
2 0
EA  dx  0 bw dx  Fw(x  L)
 dx 
Potential energy of element 2:
2
1 x3  dw  x3
 2 (w)  
2 2
x
EA  dx  x bw dx
 dx  2

2
 x1=0 x2 x3 x4 Step 2: Describe the behavior of each element

El #1 El #2 El #3 Recall that in the “direct stiffness” approach for a bar element, we

derived the stiffness matrix of each element directly using the
Potential energy of element 3: following steps:

1 x4  dw 
2
x4 TASK 1: Approximate the displacement within each bar as a
 3 (w)  
2 x3
EA 
 dx 
dx   bw dx  Fw(x  L)
x3
straight line
TASK 2: Approximate the strains and stresses and realize that a bar
(with the approximation stated in Task 1) is exactly like a spring
Total potential energy=sum of potential energies of the elements with k=EA/L
TASK 3: Use the principle of force equilibrium to generate the
 (w)   1 (w)   2 (w)   3 (w) stiffness matrix

TASK 1: APPROXIMATE THE DISPLACEMENT WITHIN

Now, we will show you a systematic way of deriving the stiffness EACH ELEMENT
matrix (sections 2.2 and 3.1 of Logan).
Simplest assumption: displacement varying linearly inside each bar
TASK 1: APPROXIMATE THE DISPLACEMENT WITHIN
EACH ELEMENT w(x)  a 0  a 1x d 2x
TASK 2: APPROXIMATE THE STRAIN and STRESS WITHIN
EACH ELEMENT d 1x x El #1
TASK 3: DERIVE THE STIFFNESS MATRIX OF EACH
x1 x2
ELEMENT (next class) USING THE PRINCIPLE OF MIN. POT
ENERGY
How to obtain a0 and a1?
Notice that the first two tasks are similar in the two methods. The
only difference is that now we are going to use the principle of w(x1 )  a 0  a1x1  d1x
minimum potential energy, rather than force equilibrium, to derive w(x 2 )  a 0  a1x 2  d 2x
the stiffness matrix.

3
 w(x1 )  a 0  a 1x1  d1x In matrix notation, we write
w(x 2 )  a 0  a 1x 2  d 2x
w(x)  N d (1)
Solve simultaneously
x2 x1 Vector of nodal shape functions
a0  d 1x  d 2x
x 2  x1 x 2  x1  x - x x - x1 
N  N 1 (x) N 2 (x)   2 
a1  
1
d 1x 
1
d 2x  x 2  x1 x 2  x1 
x 2  x1 x 2  x1
Vector of nodal displacements
Hence
x2 - x x - x1
w(x)  a 0  a 1x  d1x  d 2x  N1 (x)d1x  N 2 (x)d 2x d 
x x x x d   1x 
21 21 d 2x 
N1 (x) N 2 (x)

“Shape functions” N1(x) and N2(x)

NOTES: PROPERTIES OF THE SHAPE FUNCTIONS 2. Compatibility: The displacement approximation is continuous
across element boundaries x -x x-x
1. Kronecker delta property: The shape function at any node w (2) (x)  3
d 2x  2
d 3x
x -x x - x1 x3  x2 x3  x2
has a value of 1 at that node and a value of zero at ALL other w (x)  2
(1)
d1x  d 2x
x 2  x1 x 2  x1
nodes. 1 1
x2 - x x - x1
N1 (x)  N 2 (x) 
x 2  x1 x 2  x1
x1 x2 x
x1 x2 x3 x
El #1 El #1 El #2
Check At x=x2
x -x x2 - x2 x -x
N1 (x)  2 w (1) (x  x 2 )  d1x  2 1 d 2x  d 2x
x 2  x1 x 2  x1 x 2  x1
x 2 - x1 x3 - x2 x -x
 N1 (x  x1 )  1 w (2) (x  x 2 )  d 2x  2 2 d 3x  d 2x
x 2  x1 x3  x 2 x3  x 2
x2 - x2 Hence the displacement approximation is continuous across elements
and N1 (x  x 2 )  0
x 2  x1

4
3. Completeness Rigid body mode

N1 (x)  N 2 (x)  1 for all x N1 (x)  N 2 (x)  1 for all x

N1 (x)x1  N 2 (x)x 2  x for all x
What do we mean by “rigid body modes”?
x2 - x
Use the expressions N1 (x)  ;
x 2  x1 Assume that d1x=d2x=1, this means that the element should
x - x1 translate in the positive x direction by 1. Hence ANY point
N 2 (x)  (x) on the bar should have unit displacement. Let us see
x 2  x1 whether the displacement approximation allows this.
And check
x2 - x x - x1 w(x)  N1 (x)d1x  N 2 (x)d 2x  N1 (x)  N 2 (x)  1
N1 (x)  N 2 (x)   1
x 2  x1 x 2  x 1
x2 - x x - x1 YES!
and N1 (x) x1  N 2 (x) x 2  x1  x2  x
x 2  x1 x 2  x1

Constant strain states Completeness = Rigid body modes + Constant Strain states
N1 (x)x1  N 2 (x)x 2  x at all x

What do we mean by “constant strain states”?

Compatibility + Completeness  Convergence
Assume that d1x=x1 and d2x=x2. The strain at ANY point (x) Ensure that the solution gets better as more elements are introduced
within the bar is and, in the limit, approaches the exact answer.

d 2x  d1x x 2  x1
 (x)   1
x 2  x1 x 2  x1
Let us see whether the displacement approximation allows this.
w(x)  N1 (x)d1x  N 2 (x)d 2x  N1 (x)x1  N 2 (x)x 2  x
dw(x)
Hence,  (x)  1
dx YES!

5
Selecting Approximation Functions for Displacements: 4. How to write the expressions for the shape functions easily
(without having to derive them each time):
Consider the following guidelines, as they relate to the one-dimensional Start with the Kronecker delta property (the shape function at
bar element, when selecting a displacement function. any node has value of 1 at that node and a value of zero at all
1. Common approximation functions are usually Polynomials. other nodes) 1 1
2. The approximation function should be continuous within the bar element.
3. The approximating function should provide inter-element continuity for all x2 - x x - x1
N1 (x)  N 2 (x) 
degrees of freedom at each node for discrete line elements, and along x 2  x1 x 2  x1
common boundary lines and surfaces for two- and three-dimensional
elements. (For the bar element, we must ensure that nodes common to two or more
elements remain common to these elements upon deformation and thus prevent
x1 x2 x
El #1
overlaps or voids between elements.)
Node at which N1 is 0
The linear function is then called a conforming (or compatible) function for the bar
element because it ensures both the satisfaction of continuity between adjacent elements N1 (x)  2
x - x  Notice that the length of the element = x -x
and of continuity within the element. x 2 - x 1  2 1

 x - x   x - x 1  The denominator is
4. The approximation function should allow for rigid-body displacement and N 2 (x)  1 the numerator evaluated at
for a state of constant strain within the element.
x1 - x 2  x 2 - x1 
the node itself

A slightly fancier assumption: TASK 2: APPROXIMATE THE STRAIN and STRESS

displacement varying quadratically inside each bar WITHIN EACH ELEMENT
N1 (x) N 2 (x) N 3 (x)
From equation (1), the displacement within each element

1 w(x)  N d
Recall that the strain in the bar dw
ε
dx

x1 x2 Hence
x
x3
dN 
El #1 ε dBd (2)
x 2 - x x 3 - x   dx 
N1 (x) 
x 2 - x1 x 3 - x1  w(x)  N1 (x)d 1x  N 2 (x)d 2x  N 3 (x)d 3x The matrix B is known as the “strain-displacement matrix”
N 2 (x) 
x1 - x x 3 - x 
x1 - x 2 x 3 - x 2  This is a quadratic finite element in dN 
B
x1 - x x 2 - x  1D and it has three nodes and three  dx 
N 3 (x) 
x1 - x 3 x 2 - x 3  associated shape functions per element.

6
 Displacement is linear
For a linear finite element
w(x)  a 0  a 1x
 x - x x - x1  d 2x
N  N 1 (x) N 2 (x)   2 
 x 2  x1 x 2  x1 
d 1x x El #1
Hence
x1 x2
 -1 1  1
B   1 1 Strain is constant
x
 2  x 1 x 2  x 1  x 2  x1
d 2x - d1x
 -1 1  d1x  ε
εBd   x 2  x1
 x 2  x 1 x 2  x 1  d 2x 
d 2x - d 1x
 x El #1
x 2  x1
x1 x2
Hence, strain is a constant within each element (only for a
linear element)!

Summary
du
Recall that the stress in the bar   Eε  E
dx Inside an element, the three most important approximations in
terms of the nodal displacements (d) are:
Hence, inside the element, the approximate stress is
Displacement approximation in terms of shape functions
  EB d (3)
u(x)  N d (1)

Strain approximation in terms of strain-displacement matrix

For a linear element the stress is also constant inside each element.
This has the implication that the stress (and strain) is discontinuous ε(x)  B d (2)
across element boundaries in general.
Stress approximation in terms of strain-displacement matrix and
Young’s modulus
  EB d (3)

7
 For a linear element Why is the approximation “admissible”?

x1=0 x2 x3 x4=L

Displacement approximation in terms of shape functions El #1 El #2

El #3
 x - x x - x1  d1x  For the entire bar, the displacement approximation is
w(x)   2  
 x 2  x1 x 2  x1  d 2x  w(x)  w (1) (x)  w (2) (x)  w (3) (x)
Strain approximation
Where w(i)(x) is the displacement approximation within element (i).
 1 1  1x 
dw 1 d
ε  Let use set d1x=0. Then, can you seen that the above approximation
dx x 2  x1 d
 2x  does satisfy the two conditions of being an admissible function on
Stress approximation the entire bar, i.e.,
(1) w(x  0)  0
 1 1  1x 
E d
  Eε 
x 2  x1 d 2x  dw
(2) exists
dx

TASK 3: DERIVE THE STIFFNESS MATRIX OF EACH

ELEMENT USING THE RAYLEIGH-RITZ PRINCIPLE Lets see what the matrix

x2

T
Potential energy of element 1: B EB Adx
x1

1 x2 x2 is for a 1D linear element

2 x1
1 (w)    Adx   bw dx
x1

Recall that
Lets plug in the approximation
1
w(x)  N d ε(x)  B d   EB d B  1 1 
x 2  x1

   
1 T Hence 1  1 1
 1 1 
x2 T T x2 T T
1 (d)  d B EB Adx d  d N b dx B EB  E
2 x1 x1
x 2  x1  1  x 2  x1
E  1 E  1  1
   1 1  
x 2  x 1  2
1 x 2  x 1   1 1 
2 

8
 x1
x2 T
B EB Adx 
1
 x2  x1 
2
 1 1 x2
1 1  x1 AEdx 
 
  AEdx  x 1x  11
x2

x1
2 1
2
1
1 
Then why is it necessary to go through this complicated procedure??
1. Easy to handle nonuniform E and A
2. Easy to handle distributed loads
Now, if we assume E and A are constant For nonuniform E and A, i.e. E(x) and A(x), the stiffness matrix of
the linear element will NOT be
x1
x2 T
B EB Adx    AEdx  x 1x  11 11  AE(x
x2

x1
2 1
2
 x )  1 1

2

 x  x  1 1 
2 1

1
2
EA 1 1
AE  1 1  x2  x1  1 1 

 x2  x1  1 1 
Remembering that (x2-x1) is the length of the element, this is the But it will ALWAYS be
stiffness matrix we had derived directly before using the direct x2
k   B EB Adx
T
stiffness approach!! x1

Now lets go back to

Apply Rayleigh-Ritz principle for the 1D linear element
   
1 T  x2 T   x2 T  Π1 (d ) 
1 (d)  d   B EB Adx  d  d
T
 x N b dx   0
2  x1
   1
 d1x  Π 1 (d )
 k   fb
   0
Π1 (d ) d
1 T T  0
 d k d  d fb d 2x 
2
Element stiffness matrix
Recall from linear algebra (Lecture notes on Linear Algebra)
x2
k   B EB Adx
T
1 T T
x1
 1 (d )  d kd  d fb
2
Element nodal load vector due to distributed body force
 1 (d )
x2   kd  fb
f b   N b dx
T
d
x1

9
 Recap of the properties of the element stiffness matrix
Hence x2
k   B EB Adx
T

Π 1 (d )
x1

0
d
1. The stiffness matrix is singular and is therefore non-invertible
2. The stiffness matrix is symmetric
 kd  fb 3. Sum of any row (or column) of k the stiffness matrix is zero!
11

Why?
Exactly the same equation that we had before, except that the
stiffness matrix and nodal force vectors are more general

Sum of any row (or column) of the stiffness matrix is zero

The nodal load vector
Consider a rigid body motion of the element
x2
f b   N b dx
T
1 Element strain ε  0  B d x1
2
d
1 1
b(x)
d1x=1 d2x=1 k d    B EB Adx d
x2

x1
T 1 2 x2
f b   N b dx    1
x1
x2  N ( x ) 

x1
 b dx
 N 2 ( x) 
x2
  B E  Bd Adx
T
d1x d2x
 x2 N ( x) b dx 
 x
x1
 f1x  1 
0
      x12 
0  f 2x  x N 2 ( x) b dx 
 1 
k k  1 0 f1x  
x2
N 1 ( x ) b dx
k d   11 12       x1
k21 k22  1 0 x2 “Consistent” nodal loads
 k11  k12  0 and k21  k22  0 f 2x   N 2 ( x) b dx
x1

10
 Summary: For each element
b(x) /unit length 1 f 2 Displacement approximation in terms of shape functions
1 2 1x
f2x
Replaced by
w(x)  N d
d2x d1x d2x
d1x Strain approximation in terms of strain-displacement matrix
A distributed load is represented by two nodal loads in a ε(x)  B d
consistent manner Stress approximation
e.g., if b=1   EB d
x2 x2 x 2  x1
f1x   N 1 ( x ) b dx   N 1 ( x) dx  Element stiffness matrix
x1 x1 2
x2
x 2  x1 k   B EB Adx
T
x2 x2
f 2x   N 2 ( x) b dx   N 2 ( x) dx  x1
x1 x1 2 Element nodal load vector
Divide the total force into two equal halves and lump them at the
x2
nodes f b   N b dx
T

What happens if b(x)=x? x1

What happens for element #3? What happens for element #3?
2 Now apply Rayleigh-Ritz principle
1 x4  dw  x4
 3 (w)  
2 3
x
EA  dx  x bw dx  Fw(x  L)
 dx  3
Π 3 (d )
0
For element 3 d

 x - x x - x 3  d 3x  0 
w(x)   4    k d  fb   
 x 4  x 3 x 4  x 3  d 4x  F 
 w(x  L)  d 4x
Hence there is an extra load term on the right hand side due to the
The discretized form of the potential energy
concentrated force F applied to the right end of the bar.
 3 (d) 
1 T
2
d  x4

x3
T
B EB Adx d  d  T
 x4

x3
N
T

b dx  Fd 4x
NOTE that whenever you have a concentrated load at ANY
node, that load should be applied as an extra right hand side
term.

11
Step3:Assembly exactly as you had done before, assemble the Problem:
global stiffness matrix and global load vector and solve the 6” E=30x106 psi
resulting set of equations by properly taking into account the r=0.2836 lb/in3
displacement boundary conditions Thickness of plate, t=1”
12”
24”

P=100lb Model the plate as 2 finite elements and

3”
(1)Write the expression for element stiffness
x matrix and body force vectors
(2)Assemble the global stiffness matrix and
load vector
(3)Solve for the unknown displacements
(4)Evaluate the stress in each element
(5)Evaluate the reaction in each support

12