MANE 4240 & CIVL 4240

Introduction to Finite Elements

Finite element formulation for
1D elasticity using the
Rayleigh-Ritz Principle
Prof. Suvranu De
Reading assignment:

Lecture notes, Logan 3.10
Summary:

Stiffness matrix and nodal load vectors for 1D elasticity
problem
Axially loaded elastic bar
A(x) = cross section at x
b(x) = body force distribution
(force per unit length)
E(x) = Youngs modulus
x
y
x=0
x=L
x
F
Potential energy of the axially loaded bar corresponding to the
exact solution u(x)
L) Fu(x dx bu dx
dx
du
EA
2
1
(u)
0 0
2
=
|
.
|

\
|
= H
} }
L L
Potential energy of the bar corresponding to an admissible
displacement w(x)
L) Fw(x dx bw dx
dx
dw
EA
2
1
(w)
0 0
2
=
|
.
|

\
|
= H
} }
L L
Finite element idea:

Step 1: Divide the truss into finite elements connected to each
other through special points (nodes)

El #1
El #2
El #3
1 2 3
4
L) Fw(x dx bw dx
dx
dw
EA
2
1
(w)
0 0
2
=
|
.
|

\
|
= H
} }
L L
Total potential energy=sum of potential energies of the elements
El #1
El #2
El #3
x
1
=0 x
2

x
3

x
4
=L

dx bw dx
dx
dw
EA
2
1
(w)
2
1
2
1
2
1
} }

|
.
|

\
|
= H
x
x
x
x
L) Fw(x dx bw dx
dx
dw
EA
2
1
(w)
0 0
2
=
|
.
|

\
|
= H
} }
L L
Total potential energy
Potential energy of element 1:
dx bw dx
dx
dw
EA
2
1
(w)
3
2
3
2
2
2
} }

|
.
|

\
|
= H
x
x
x
x
Potential energy of element 2:
El #1
El #2
El #3
x
1
=0 x
2

x
3

x
4

Total potential energy=sum of potential energies of the elements
Potential energy of element 3:
(w) (w) (w) (w)
3 2 1
H + H + H = H
L) Fw(x dx bw dx
dx
dw
EA
2
1
(w)
4
3
4
3
2
3
=
|
.
|

\
|
= H
} }
x
x
x
x
Step 2: Describe the behavior of each element

In the direct stiffness approach, we derived the stiffness matrix
of each element directly (See lecture on Springs/Trusses).

Now, we will first approximate the displacement inside each
element and then 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 (this class) USING THE RAYLEIGH-RITZ
PRINCIPLE

Inside an element, the three most important approximations in
terms of the nodal displacements (d) are:
d B E = o
(1)
Displacement approximation in terms of shape functions
d N w(x) =
d B (x) =
Strain approximation in terms of strain-displacement matrix
(2)
Stress approximation in terms of strain-displacement matrix and
Youngs modulus
(3)
Summary
The shape functions for a 1D linear element
2x
1 2
1
1x
1 2
2
d
x x
x - x
d
x x
x - x
w(x)

=
Within the element, the displacement approximation is
x
x
1
x
2
El #1
1 2
2
1
x x
x - x
(x) N

=
1 2
1
2
x x
x - x
(x) N

=
1
1
Displacement approximation in terms of shape functions
)
`

=
2x
1x
1 2
1
1 2
2
d
d
x x
x - x
x x
x - x
w(x)
Strain approximation
Stress approximation
For a linear element
| |
)
`

= =
2x
1x
1 2
d
d
1 1
x x
1
dx
dw

| |
)
`

= =
2x
1x
1 2
d
d
1 1
x x
E
E o
For the entire bar, the displacement approximation is
Why is the approximation admissible?
El #1
El #2
El #3
x
1
=0 x
2

x
3

x
4
=L

(x) w (x) w (x) w w(x)
(3) (2) (1)
+ + =
exists
dx
dw
(2)
0 0) w(x (1) = =
Where w
(i)
(x) is the displacement approximation within element (i).
Let use set d
1x
=0. Then, can you seen that the above approximation
does satisfy the two conditions of being an admissible function on
the entire bar, i.e.,

2 2
1 1
1
1
(w) dx bw dx
2
x x
x x
A o c H =
} }
Potential energy of element 1:
TASK 3: DERIVE THE STIFFNESS MATRIX OF EACH
ELEMENT USING THE RAYLEIGH-RITZ PRINCIPLE
Lets plug in the approximation
d N w(x) =
( ) ( )
2 2
1 1
T
1
1
(d) d B EB dx d d N b dx
2
x x
T T T
x x
A H =
} }
d B (x) =
d B E = o
2
1
T
B EB dx
x
x
A
}
Lets see what the matrix
is for a 1D linear element
Recall that
| | 1 1
x x
1
B
1 2

=
Hence
| |
( )
| |
( )
(

=
)
`

)
`

=
1 1
1 1
x x
E
1 1
1
1
x x
E
1 1
x x
1
E
1
1
x x
1
B E B
2
1 2
2
1 2
1 2 1 2
T
( )
( )
( )
2 2 2
1 1 1
T
2 2
2 1 2 1
1 1 1 1
1 1
B EB dx AEdx AEdx
1 1 1 1
x x x x
x x x
x x x
A

( (
= =
( (



} } }
Remembering that (x
2
-x
1
) is the length of the element, this is the
stiffness matrix we had derived directly before using the direct
stiffness approach!!
( )
( ) ( )
( )
2 2
1 1
T
2 1
2 2
2 1 2 1
2 1
1 1 1 1
AE(x x ) 1
B EB dx AEdx
1 1 1 1
x x x x
1 1
AE
1 1 x x
x x
x x
A

( (
= =
( (

(
=
(


} }
Now, if we assume E and A are constant
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
For nonuniform E and A, i.e. E(x) and A(x), the stiffness matrix of
the linear element will NOT be
( )
2 1
1 1
EA
1 1 x x

(
(


But it will ALWAYS be
2
1
T
B EB dx
x
x
k A =
}
Now lets go back to
2 2
1 1
T
1
1
(d) d B EB dx d d N b dx
2
1
d d d
2
b
x x
T T T
x x
k f
T T
b
A
k f
| |
| |
|
|
H =
|
|
|
|
\ .
\ .
=
} }
2
1
T
B EB dx
x
x
k A =
}
Element stiffness matrix
Element nodal load vector due to distributed body force
dx b N
2
1
}
=
x
x
T
b
f
Apply Rayleigh-Ritz principle for the 1D linear element

=
c
c
=
c
c
0
d
) d (
0
d
) d (
2x
1
1x
1
b
b
T T
f k
f k
=
c
H c

= H
d
d
) d (
d d d
2
1
) d (
1
1
0
d
) d (
1
=
c
c

Recall from linear algebra (Lecture notes on Linear Algebra)

Hence
0
d
) d (
1
=
c
c
b
f k = d
Exactly the same equation that we had before, except that the
stiffness matrix and nodal force vectors are more general
Recap of the properties of the element stiffness matrix
1. The stiffness matrix is singular and is therefore non-invertible
2. The stiffness matrix is symmetric
3. Sum of any row (or column) of the stiffness matrix is zero!
2
1
T
B EB dx
x
x
k A =
}
Why?
11
k
Consider a rigid body motion of the element
1
2
d
1x
=1
d
2x
=1
)
`

=
1
1
d
d B 0 = =
Element strain
( )
( )
2
1
2
1
T
T
d B EB dx d
B E Bd dx
0
0
x
x
x
x
k A
A
=
=

=
`
)
}
}
0 and 0
0
0
1
1
d
22 21 12 11
22 21
12 11
= + = +
)
`

=
)
`

=
k k k k
k k
k k
k
Sum of any row (or column) of the stiffness matrix is zero
The nodal load vector
1
2
d
1x

d
2x

b(x)

=
)
`

)
`

= =
}
}
} }
dx b ) (
dx b ) (
dx b
) (
) (
dx b N
2
1
2
1
2
1
2
1
2
1
2
1
2
1
x
x
x
x
x
x
x
x
x
x
b
x N
x N
f
f
x N
x N
f
dx b N
2
1
}
=
x
x
T
b
f
dx b ) (
dx b ) (
2
1
2
1
2 2
1 1
}
}
=
=
x
x
x
x
x
x
x N f
x N f
Consistent nodal loads
1
2
d
1x

d
2x

b(x) /unit length
1
2
d
1x

d
2x

Replaced by
f
1x

f
2x

A distributed load is represented by two nodal loads in a
consistent manner
e.g., if b=1
2
x
dx ) ( dx b ) (
2
x
dx ) ( dx b ) (
1 2
2 2 2
1 2
1 1 1
2
1
2
1
2
1
2
1
x
x N x N f
x
x N x N f
x
x
x
x
x
x
x
x
x
x

= = =

= = =
} }
} }
Divide the total force into two equal halves and lump them at the
nodes
What happens if b(x)=x?
d B E = o
Displacement approximation in terms of shape functions
d N w(x) =
d B (x) =
Strain approximation in terms of strain-displacement matrix
Stress approximation
Summary: For each element
2
1
T
B EB dx
x
x
k A =
}
Element stiffness matrix
dx b N
2
1
}
=
x
x
T
b
f
Element nodal load vector
What happens for element #3?
L) Fw(x dx bw dx
dx
dw
EA
2
1
(w)
4
3
4
3
2
3
=
|
.
|

\
|
= H
} }
x
x
x
x
For element 3
)
`

=
4x
3x
3 4
3
3 4
4
d
d
x x
x - x
x x
x - x
w(x)
4x
d L) w(x = =
The discretized form of the potential energy
( ) ( )
4 4
3 3
T
3 4x
1
(d) d B EB dx d d N b dx Fd
2
x x
T T T
x x
A H =
} }
What happens for element #3?
0
d
) d (
3
=
c
c
)
`

+ =
F
0
d
b
f k
Now apply Rayleigh-Ritz principle
Hence there is an extra load term on the right hand side due to the
concentrated force F applied to the right end of the bar.

NOTE that whenever you have a concentrated load at ANY
node, that load should be applied as an extra right hand side
term.
Step3:Assembly exactly as you had done before, assemble the
global stiffness matrix and global load vector and solve the
resulting set of equations by properly taking into account the
displacement boundary conditions
Problem:
x
24
3
6
P=100lb
12
E=30x10
6
psi
=0.2836 lb/in
3
Thickness of plate, t=1
Model the plate as 2 finite elements and

(1)Write the expression for element stiffness
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
Finite element model
x
12
12
1
2
3
El #1
El #2
P=100lb
Element # Node 1 Node 2
1 1 2
2 2 3
Node-element connectivity chart
Stiffness matrix of El #1
( )
12 12
(1) T
2
0 0
1 1
B EB dx ( )
1 1 (12)
E
k A A x dx
(
= =
(

} }
12 12 12
3
0 0 0
( ) (6 0.125 ) (6 0.125 ) 63 A x dx t x dx t x dx in = = =
} } }
( )
(1)
6
2
1 1 1 1
63 13.125 10
1 1 1 1 (12)
E
k
( (
= =
( (


Solution (1)
Stiffness matrix of El #2
( )
24 24
(2) T
2
12 12
1 1
B EB dx ( )
1 1 (12)
E
k A A x dx
(
= =
(

} }
24 24 24
3
12 12 12
( ) (6 0.125 ) (6 0.125 ) 45 A x dx t x dx t x dx in = = =
} } }
( )
(2)
6
2
1 1 1 1
45 9.375 10
1 1 1 1 (12)
E
k
( (
= =
( (


Now compute the element load vector due to distributed body
force (weight)
dx b N
2
1
}
=
x
x
T
b
f
x
12
6
4.5
x
6 - 0.125x
( )
12 12
(1)
0 0
12
0
(1)
12
1
(1)
0
2
( )
N b dx N A dx
N dx
( )
(6 0.125 ) dx
( )
33
0.2836
30
9.3588
8.508
T T
b
T
A x
f
A
N x
t x
N x
lb
lb

= =
=

=
`
)

=
`
)

=
`
)
} }
}
}
For element #1
x
12
1
2
El #1
) (
) 1 (
2
x N
) (
) 1 (
1
x N
12
) (
12
12
) (
) 1 (
2
) 1 (
1
x
x N
x
x N
=

=
Superscript in parenthesis indicates
element number
( )
24 24
(2)
12 12
24
12
(2)
24
2
(2)
12
3
( )
N b dx N A dx
N dx
( )
(6 0.125 ) dx
( )
24
0.2836
21
6.8064
5.9556
T T
b
T
A x
f
A
N x
t x
N x
lb
lb

= =
=

=
`
)

=
`
)

=
`
)
} }
}
}
For element #2
x
12
12
1
2
3
El #1
El #2
) (
) 2 (
2
x N
) (
) 2 (
3
x N
12
12
) (
12
24
) (
) 2 (
3
) 2 (
2

=
x
x N
x
x N
Solution (2) Assemble the system equations
6
13.125 13.125 0
10 13.125 22.5 9.375
0 9.375 9.375
K

(
(
=
(
(

9.3588
8.508 6.8064
5.9556
0
100
0
9.3588
115.3144
5.9556
b
concentrated load
b
concentrated load
f f f
f lb
f lb
f lb
= +


= +
`

)


=
`

)


=
`

)
Hence we need to solve
1x 1
6
2x
3x
13.125 13.125 0 d 9.3588
10 13.125 22.5 9.375 d 115.3144
0 9.375 9.375 d 5.9556
R +
(

(
=
` `
(

(

) )
R
1
is the reaction at node 1.
Notice that since the boundary condition at x=0 (d
1x
=0) has not been
taken into account, the system matrix is not invertible.
Incorporating the boundary condition d
1x
=0 we need to solve the
following set of equations
2x 6
3x
d 22.5 9.375 115.3144
10
d 9.375 9.375 5.9556
(
=
` `
(

) )
Solution (3)
5
2
5
3
0.92396 10
0.98749 10
x
x
d
in
d

=
` `

) )
Solution (4) Stress in elements
| |
(1) (1)
(1)
1x
2x 2 1
6
2x 1x
d
1 1
d
30 10
d d 0
12
23.099
EB d
E
x x
psi
o =

=
`

= =
=
Solve to obtain
Notice that since we are using linear elements, the stress within
each element is constant.
In element #1
| |
( )
(2) (2)
(2)
2x
3x 3 2
6
3x 2x
d
1 1
d
30 10
d -d
12
1.5882
EB d
E
x x
psi
o =

=
`

=
=
In element #2
Solution (5) Reaction at support
Go back to the first line of the global equilibrium equations
1
130.6288 R lb =
1x 1
6
2x
3x
13.125 13.125 0 d 9.3588
10 13.125 22.5 9.375 d 115.3144
0 9.375 9.375 d 5.9556
R +
(

(
=
` `
(

(

) )
Check
x
24
3
6
12
P=100lb
R
1
24
1
0
24
0
( )
100 (6 0.125 )
130.6288
x
x
R P A x dx
t x dx
lb

=
=
= +
= +
=
}
}
The reaction at the wall from force
equilibrium in the x-direction
(The ve sign indicates that the force is in the ve x-direction)
Problem: Can you solve for the displacement and stresses
analytically?

Check out
9 2 7
9 2 7 6
4.727 10 9.487 10 0 12
4.727 10 2.0797 10 8.89 10 12 24
anal
x x for x
u
x x for x

+ s <

=

+ + s s

Stress
6
( ) 30 10
anal anal
anal
du du
x E
dx dx
o = =
0 5 10 15 20
0
0.2
0.4
0.6
0.8
1
1.2
x 10
-4
x (in)
D
i
s
p
l
a
c
e
m
e
n
t

(
i
n
)
Analytical solution
Finite element solution
Comparison of displacement solutions
Notice:
1. Slope discontinuity at x=12 (why?)
2. The finite element solution does not produce the exact
solution even at the nodes
3. We may improve the solution by
(1) Increasing the number of elements
(2) Using higher order elements (e.g., quadratic instead of
linear)
0 5 10 15 20
-5
0
5
10
15
20
25
30
x (in)
S
t
r
e
s
s

(
p
s
i
)
Finite element solution
Analytical solutions
Comparison of stress solutions
The analytical as well as the finite element stresses are
discontinuous across the elements