Reading assignment:

MANE 4240 & CIVL 4240

Introduction to Finite Elements Lecture notes, Logan 2.2, 3.1

Summary:

Prof. Suvranu De • Linear shape functions in 1D

• Quadratic and higher order shape functions
• Approximation of strains and stresses in an element

Shape functions in 1D

Axially loaded elastic bar

Finite element formulation, takes as its starting point, not the
y A(x) = cross section at x strong formulation, but the Principle of Minimum Potential
b(x) = body force distribution Energy.
F (force per unit length)
x E(x) = Young’s modulus Task is to find the function ‘w’ that minimizes the potential energy
x of the system
x=0 x=L
2
1 L ⎛ dw ⎞ L

2 ∫0
Potential energy of the axially loaded bar corresponding to the Π (w) = EA⎜ ⎟ dx − ∫0 bw dx − Fw(x = L)
⎝ dx ⎠
exact solution u(x)

2
1 L ⎛ du ⎞ L

2 ∫0
Π (u) = EA⎜ ⎟ dx − ∫ bu dx − Fu(x = L) From the Principle of Minimum Potential Energy, that function
⎝ dx ⎠ 0 ‘w’ is the exact solution.

1
 Rayleigh-Ritz Principle
The approximate solution is
Step 1. Assume a solution
u ( x) = a 0ϕ o ( x) + a1ϕ1 ( x) + a 2ϕ 2 ( x) + ...
w( x) = a 0ϕ o ( x) + a1ϕ1 ( x) + a 2ϕ 2 ( x) + ... Where the coefficients have been obtained from step 3

Where ϕo(x), ϕ1(x),… are “admissible” functions and ao, a1,

etc are constants to be determined.
Step 2. Plug the approximate solution into the potential energy
2
1 L ⎛ dw ⎞ L

2 ∫0
Π (w) = EA⎜ ⎟ dx − ∫0 bw dx − Fw(x = L)
⎝ dx ⎠
Step 3. Obtain the coefficients ao, a1, etc by setting
∂Π (w)
= 0, i = 0,1,2,...
∂ai

Finite element idea:

Need to find a systematic way of choosing the approximation
functions. Step 1: Divide the truss into finite elements connected to each
other through special points (“nodes”)
One idea: Choose polynomials!

w( x) = a 0 Is this good? (Is ‘1’ an “admissible” function?) 1 2 3 4

El #1 El #2 El #3
w( x) = a1 x Is this good? (Is ‘x’ an “admissible” function?)

Total potential energy=sum of potential energies of the elements

2
1 L ⎛ dw ⎞ L
Π (w) =
2 ∫0
EA⎜ ⎟ dx − ∫0 bw dx − Fw(x = L)
⎝ dx ⎠

2
 x1=0 x2 x3 x4=L x1=0 x2 x3 x4

El #1 El #2 El #3 El #1 El #2 El #3

Total potential energy Potential energy of element 3:

2
1 L ⎛ dw ⎞ L
Π (w) = ∫ EA⎜ ⎟ dx − ∫0 bw dx − Fw(x = L) 1 x4 ⎛ dw ⎞
2
x4

2 ∫x3
2 0 ⎝ dx ⎠ Π 3 (w) = EA⎜ ⎟ dx − ∫x bw dx − Fw(x = L)
⎝ dx ⎠ 3
Potential energy of element 1:
2
1 x2 ⎛ dw ⎞ x2
Total potential energy=sum of potential energies of the elements
2 ∫x1
Π 1 (w) = EA⎜ ⎟ dx − ∫x bw dx
⎝ dx ⎠ 1

Π (w) = Π 1 (w) + Π 2 (w) + Π 3 (w)

Potential energy of element 2:
2
1 x3 ⎛ dw ⎞ x3

2 ∫x2
Π 2 (w) = EA⎜ ⎟ dx − ∫x bw dx
⎝ dx ⎠ 2

Step 2: Describe the behavior of each element

Recall that in the “direct stiffness” approach for a bar element, we
derived the stiffness matrix of each element directly (See lecture on
Trusses) using the following steps:
TASK 1: Approximate the displacement within each bar as a
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
with k=EA/L
TASK 3: Use the principle of force equilibrium to generate the
stiffness matrix
Now, we will show you a systematic way of deriving the stiffness
matrix (sections 2.2 and 3.1 of Logan).
TASK 1: APPROXIMATE THE DISPLACEMENT WITHIN
EACH ELEMENT
TASK 2: APPROXIMATE THE STRAIN and STRESS WITHIN
EACH ELEMENT
TASK 3: DERIVE THE STIFFNESS MATRIX OF EACH
ELEMENT (next class) USING THE PRINCIPLE OF MIN. POT
ENERGY
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
minimum potential energy, rather than force equilibrium, to derive
the stiffness matrix.

3
TASK 1: APPROXIMATE THE DISPLACEMENT WITHIN
EACH ELEMENT w(x1 ) = a 0 + a1x1 = d1x
w(x 2 ) = a 0 + a 1x 2 = d 2x
Simplest assumption: displacement varying linearly inside each bar

d 2x Solve simultaneously
w(x) = a 0 + a 1x
x2 x1
a0 = d 1x − d 2x
d 1x x El #1 x 2 − x1 x 2 − x1
x1 x2 1 1
a1 = − d 1x + d 2x
x 2 − x1 x 2 − x1
How to obtain a0 and a1?
Hence
w(x1 ) = a 0 + a 1x1 = d1x x2 - x x - x1
w(x) = a 0 + a 1x = d1x + d 2x = N1 (x)d1x + N 2 (x)d 2x
w(x 2 ) = a 0 + a 1x 2 = d 2x x 2 − x1 x 2 − x1



N1 (x) N 2 (x)

“Shape functions” N1(x) and N2(x)

In matrix notation, we write NOTES: PROPERTIES OF THE SHAPE FUNCTIONS

1. Kronecker delta property: The shape function at any node

w(x) = N d (1) has a value of 1 at that node and a value of zero at ALL other
nodes. 1 1
Vector of nodal shape functions x2 - x
N 1 (x) = x - x1
⎡ x - x x - x1 ⎤ N 2 (x) =
x 2 − x1
N = [N 1 (x) N 2 (x)] = ⎢ 2 ⎥
x 2 − x1
⎣ x 2 − x1 x 2 − x1 ⎦ x1 x2 x
El #1
Vector of nodal displacements Check
x2 - x
N1 (x) =
x 2 − x1
⎧d ⎫
d = ⎨ 1x ⎬ ⇒ N1 (x = x1 ) =
x 2 - x1
=1
⎩d 2x ⎭ x 2 − x1
x2 - x2
and N1 (x = x 2 ) = =0
x 2 − x1

4
2. Compatibility: The displacement approximation is continuous 3. Completeness
across element boundaries x -x x-x
w (2) (x) = 3
d 2x + 2
d 3x
w (1) (x) =
x2 - x
d1x +
x - x1
d 2x
x3 − x 2 x3 − x2 N1 (x) + N 2 (x) = 1 for all x
x 2 − x1 x 2 − x1
N1 (x)x1 + N 2 (x)x 2 = x for all x
x2 - x
Use the expressions N1 (x) = ;
x 2 − x1
x - x1
x1 x2 x3 x N 2 (x) =
El #1 El #2 x 2 − x1
At x=x2
x2 - x2 x -x And check
w (1) (x = x 2 ) = d1x + 2 1 d 2x = d 2x x2 - x x - x1
x 2 − x1 x 2 − x1 N1 (x) + N 2 (x) = + =1
x -x x -x x 2 − x 1 x 2 − x1
w (2) (x = x 2 ) = 3 2 d 2x + 2 2 d 3x = d 2x
x3 − x2 x3 − x2 x2 - x x - x1
Hence the displacement approximation is continuous across elements and N1 (x) x1 + N 2 (x) x 2 = x1 + x2 = x
x 2 − x1 x 2 − x1

Rigid body mode Constant strain states

N1 (x) + N 2 (x) = 1 for all x N1 (x)x1 + N 2 (x)x 2 = x at all x

What do we mean by “rigid body modes”? What do we mean by “constant strain states”?

Assume that d1x=d2x=1, this means that the element should Assume that d1x=x1 and d2x=x2. The strain at ANY point (x)
translate in the positive x direction by 1. Hence ANY point within the bar is
(x) on the bar should have unit displacement. Let us see d 2x − d1x x 2 − x1
whether the displacement approximation allows this. ε (x) = = =1
x 2 − x1 x 2 − x1

w(x) = N1 (x)d1x + N 2 (x)d 2x = N1 (x) + N 2 (x) = 1 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
YES! dw(x)
Hence, ε (x) = =1
dx YES!

5
 4. How to write the expressions for the shape functions easily
Completeness = Rigid body modes + Constant Strain states (without having to derive them each time):
Start with the Kronecker delta property (the shape function at
any node has value of 1 at that node and a value of zero at all
other nodes) 1 1
x2 - x x - x1
Compatibility + Completeness ⇒ Convergence N 1 (x) = N 2 (x) =
Ensure that the solution gets better as more elements are introduced x 2 − x1 x 2 − x1
and, in the limit, approaches the exact answer.
x1 x2 x
El #1
Node at which N1 is 0
N1 (x) =
(x 2 - x ) Notice that the length of the element = x -x
(x 2 - x1 ) 2 1

(x1 - x ) (x - x1 ) The denominator is

N 2 (x) = = the numerator evaluated at
(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)
N1 (x) =
(x 2 - x )(x 3 - x ) ⎣ dx ⎦
(x 2 - x1 )(x 3 - x1 ) w(x) = N1 (x)d1x + 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=⎢ ⎥
N 3 (x) =
(x1 - x )(x 2 - x ) 1D and it has three nodes and three ⎣ dx ⎦
(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 + a1x
⎡ 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 − x1 x 2 − x1 ⎦ x 2 − x1
d -d
⎡ -1 1 ⎤ ⎧d 1x ⎫ ε = 2x 1x
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 σ = 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
σ = 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
σ = EB d (3)

7
 Summary

For a linear element

Displacement approximation in terms of shape functions

⎡ x - x x - x 1 ⎤ ⎧d 1x ⎫
u(x) = ⎢ 2 ⎥⎨ ⎬
⎣ x 2 − x 1 x 2 − x 1 ⎦ ⎩d 2x ⎭
Strain approximation
[− 1 1 ]⎧⎨ 1x ⎫⎬
1 d
ε=
x 2 − x1 d
⎩ 2x ⎭
Stress approximation

[− 1 1 ]⎧⎨ 1x ⎫⎬
E d
σ=
x 2 − x1 ⎩d 2x ⎭