Finite Element Analysis

Somenath Mukherjee

Chief Scientist
CSIR-Central Mechanical Engineering Research Institute
Durgapur – 713209, West Bengal.

2023
1
 Acknowledgements
My sincere thanks go to my revered mentors
of Finite Element Method

• Prof. Dipak Sengupta,

(Deptt. of Aerospace Engg and Applied Mechanics, IIEST, Shibpur
Formerly known as Bengal Engineering College)

• Dr. Gangan Prathap,

(Outstanding Scientist, CSIR National Institute of Interdisciplinary Science
and Technology, Trivandrum.
Formerly Senior Principal Scientist, NAL, Bangalore,
Scientist-in-charge, CMMACS, Bangalore,
Vice-Chancellor, Cochin University,
Director, NISCAIR, New Delhi.)
- Somenath Mukherjee
2
Introduction to the Finite Element Method

3
 What is FEM ?
• The Finite Element Method is a Numerical tool to solve
differential equations (approximately) through
discretization.
• Steps of FEM:
(1) Discretisation of the continuum into elements.
(2) Element formulation using approximate solution for the field variable of
the differential equation.
(3) Element assembly through connecting nodal degrees of freedom.
(4) Solution of the system of equations of the global domain, incorporating
boundary conditions.
(5) Post processing of solutions of variables (nodal displacements) of the global
FE equations to compute desired fields (strain/stress) of the domain.
Domain boundary

element
• Requirements of acceptable FE solutions: Convergence of
solutions to more accurate results by using more elements (i.e. through finer
4
discretization).
5
 Element formulation
• Using approximate solutions for field variables
• Variational principles for equilibrium
Element Equilibriu m : F = K  
e e e

ND
or Fi _ DOF =  kij . j ND = Total element DOF
j =1

Element Load Vector : F = F

e e
applied + R e
Connectivity 
kij=Force required at d.o.f. i for unit displacement for d.o.f. j,
with all other d.o.f. locked.

F =  F 
NE
Assembly of elements: Global Force : G e

e =1

K  =  K 
NE
G e
Global Stiffness :
e =1

    
Solve for displacements: F G = K G  G

Solve with Boundary conditions satisfied

6
Estimate element Strains/Stresses from nodal displacements
 How do finite elements capture the solutions ?
(FEA Paradigm: the “Analysis of FEM”)
Traditional view of FEM:
The Finite Element Method is a displacement based (equilibrium searching )
method. For simple determinate structures, FEM employs approximate
displacement functions and yet captures the exact displacements at the nodes.

Though the above assumption is true for many simple cases, it is NOT true for elements
with varying sectional properties used for statically indeterminate structural systems
(case of pollution error in fixed-fixed bars of varying sections).

Alternative interpretation of FEM results (the BEST-FIT Paradigm)

Although FEM is viewed as a displacement based (stiffness) method of analysis,
it effectively works by “sensing” the exact/converged solution of the differential
equations and providing an approximate solution with the least error possible.
7
FE solutions are the best-fits to the exact strains !
Finite element discretization and stress analysis of a segment of a bicycle frame

8
Stress Concentration Factor in a loaded plate with a hole

Von Mises Stress Distributions around the hole

3 oct 3 1
 VM =  =  ( p1 − p2 )2 + ( p2 − p3 )2 + ( p3 − p1 )2
2 2 3

Von Mises Criterion for stress failure beyond yield point:

 VM   Y
 Octahedral plane and octahedral stresses
At a stressed point, the octahedral plane is one that is
equally inclined to the three principal planes.
 1
z  1oct =  2 oct =  3 oct = 
p3 T p1 3
C
Traction on octahedral plane
ν  p1 0 0   1 
oct
 p1 
p2 p2
B y
T =   
oct p oct
=  0 p2
 oct  1  
0   2 . =  p2 
o  0 0 p3   3 
oct 3 
A
 p3 
x p1

p3
(T )oct 2
=
3
(
1 2 2
p1 + p2 + p3
2
) …(16a)

T
 oct   p1 0 0   1 
oct

 1 oct   oct  1
Normal stress on  n
oct
  T 
=  oct
T oct
=  2  0
 p2 0   2  = ( p1 + p 2 + p3 )
3
octahedral plane  oct   0 0 p3   3 
oct
 3 
…(16b)
( )
n
oct 2 1
= ( p1 + p 2 + p3 )
9
2

Shear stress on  oct = (T ) − ( )

oct 2
n
oct 2
=
1
3
( p1 − p 2 )2 + ( p 2 − p3 )2 + ( p3 − p1 )2 10
…(16c)
octahedral
plane:
 Free vibration analysis of an aircraft
wing using Euler beam elements.

A typical subsonic aircraft wing

7.105 m Leading edge

2.1469m
Trailing edge
Root 0.849 m Tip

(a) (b)

Length of the wing divided

in 22 elements for FE
formulation

(c) (d)

Normal mode shapes of the wing of the aircraft

(a) First bending mode in y direction (natural frequency 7.2165 Hz)
(b) Second bending mode in y direction (natural frequency 21.138 Hz)
(c) Third bending mode in y direction (natural frequency 50.405 Hz) 11
(d) First torsional mode (natural frequency 56.8296 Hz)
Brief History of Finite Elements

12
The Greeks had conceived the idea of realization through
discretization(!) . How  was discovered.

b D

Perimeter nb
Limn→ = Limn→ = 3.14159.... = 
Diameter D
How ?
nb
Limn→ = Limn→ [n. sin(  / n)] = 
D 13
Concepts of Variational Mathematical foundations of FEM in the
Principles in the 20th century
19th century

G. Strang
Lord Rayleigh W. Ritz
O C Zienkiewicz,
John Argiris
&
Ray W Clough
Richard Courant 14
B Galerkin
 Lecture 1

Analysis of Skeletal Systems

(Direct Approach)

Somenath Mukherjee

Chief Scientist,
CSIR-CMERI,
Durgapur, India

15
 Analysis of Skeletal Systems
(Direct Approach)

Chapters
1. Analysis of a system of assembly of springs
2. Simple bar element
3. Simple Classical Euler beam element
4. Transformation of co-ordinates;
5. Analysis of Plane Trusses and Plane Frames

16
 Chapter 1
Analysis of a system of assembly of
springs.

17
 1.1 A single isolated spring element
k
1 2
u1 ,F1 /\/\/\/\/\/\ u2 ,F2
Convention: Right hand direction is positive for forces/displacements.
What is the relationship between the nodal displacements (u1,u2) and the
nodal forces (F1,F2) for a linear spring at the equilibrium condition ?

We express the linear force-displacement relationship of the spring as

 F1   k11 k12  u1  (1.1)

 =    or {F e } = [ K e ]{d e }
 F2  k 21 k 22  u2 
where
 k11 k12 
[K ] = 
e
 = spring element stiffness matrix
k 21 k 22 
kij is the force necessary at degree of freedom i for a unit
displacement at degree of freedom j, with all degrees of 18
freedom except j locked.
 k 2
u1=1, 1
/\/\/\/\/\/\ u2=0,
F1=k11=k F2=k21= -k
1 k 2
u1=0,
/\/\/\/\/\/\ u2=1,
F1=k12=-k F2=k22= k
Hence the stiffness matrix for the spring element is

 k11 k12   k − k
[K ] = 
e
 = − k (1.2)
k k
 21 22   k 

The force-displacement relationship for the spring element is

 F1   k − k  u1 
 =    or {F e } = [ K e ]{d e } (1.3)
F2  − k k  u2 

F 1= ku1 − ku2
Expanding (1.3)
F2 = −ku1 + ku2 19
 1.2 A connected spring element
k
Adjacent element 1 2 Adjacent element
or support /\/\/\/\/\/\ or support
F1( applied ) F2 ( applied )

F1( applied ) k F2 ( applied )

/\/\/\/\/\/\
R1 R2
From free body diagram
of the element − R1 − R2
F1 = F1( applied) + R1
F2 = F2 ( applied) + R2 R1 & R2 are end reactions acting on the element

i.e.
 F1( applied) + R1   k − k  u1  (1.4)
F + =    or {F e } + {R e } = [ K e ]{d e }
 2 ( applied) R2 − k k  u2  20
 1.3 An assembly of spring elements
Fixed k1 k2 k3 4 R4
end; R 1 2 3
1 /\/\/\/\/\/\ /\/\/\/\/\/\ /\/\/\/\/\/\
u1=0 u4=0
F1( applied ) F2 ( applied ) F3( applied ) F4 ( applied )
Assembly by adding corresponding parameters of same degree of freedom
Assembly :
 F1( applied) + R1 = ?   k1 − k1 0 0  u1 = 0 
    u = ? 
 F 2 ( applied)  − k1 k1 + k 2 − k2 0   2 
 = 
 F 3 ( applied)   0 − k2 k 2 + k3 − k 3   u3 = ? 
 F4 ( applied) + R4 = ?  0 0 − k3

k3  u4 = 0
 
{F G } + {RG } = [ K G ]{uG } (1.5)
Note: For every degree of freedom, you know either the displacement, or the
force. For every kinematic (displacement) boundary condition, there is an
associated unknown force (reaction). 21
 k1 k2 k3
R1 1 2 3 4 R4
/\/\/\/\/\/\ /\/\/\/\/\/\ /\/\/\/\/\/\ u4=a4
u1=a1
F1( applied ) F2 ( applied ) F3( applied ) F4 ( applied )

How do we incorporate arbitrary boundary conditions ?

u1=a1, u4=a4
 F1( applied) + R1 = ?   k1 − k1 0 0   u1 = a1 
  
 F 2 ( applied)  − k1 k1 + k 2 − k2 0   u2 = ? 
 =  
 F 3 ( applied)   0 − k2 k 2 + k3 − k 3   u3 = ? 
 F4 ( applied) + R4 = ?  0 0 − k3

k3  u4 = a4 
 
Solve
 F2 ( applied ) − (−k1 )u1  k1 + k 2 − k 2  u2 
F  =  u 
 3( applied ) − ( − k u
3) 4  − k k + k
  2 2 3  3 
or
{Fnew } = [ K new ]{uunknown} u1 = a1 , u4 = a4 (1.6)
22
 k1 k2 k3
R1 1 2 3 4 R4
/\/\/\/\/\/\ /\/\/\/\/\/\ /\/\/\/\/\/\ u4=a4
u1=a1
F1( applied ) F2 ( applied ) F3( applied ) F4 ( applied )

Solving for displacements Boundary Reactions

u1 = a1, u 2 = a4
{RG } = [ K G ]{uG } − {F G }
Method 1 :
 F2 ( applied ) − (−k1 )u1  k1 + k 2 − k 2  u2   R1   u1   F1( applied ) 
F =   0 u   F 
 3( applied ) − ( − k )u
3 4   − k2 k 2 + k3  u3    G  2   2 ( applied ) 
  = [ K ]  −  F 
 0 u
 3   ( applied ) 
3
 R4  u 4   F4 ( applied ) 
 
Method 2:
 a1   1 0 0 0   u1  (1.7)
F   0  u2 
 2 ( applied )  − k1 k1 + k 2 − k2
 =  
 3( applied )   0
F − k2 k 2 + k3 − k3  u3  R1 = k1u1 − k1u2 − F1( applied )
 a4   0 0 0

1  u4 
R4 = −k3u3 + k3u4 − F4( applied )
23
 k1 k2 k3
R1 1 2 3 4 R4
/\/\/\/\/\/\ /\/\/\/\/\/\ /\/\/\/\/\/\ u4=0
u1=0
F2( applied ) = 10kN F3( applied ) = 20kN
Example 1.1:
Solve for joint displacements and fixed support reactions for the spring
assembly under the given loading. Given data:
k1=1200 kN/m, k2=1800 kN/m, k3=1500 kN/m
F2=10 kN, F3= 20 kN
10   3000 − 1800 u2 
Use equation (1.4) and solve:  =  
20 − 1800 3300  u3 
Displacements: u2=0.0104 m, u3= 0.0117 m
R1 = 0 − 1200u2 − 0
Use equation (1.5) : R4 = −1500u3 + 0 − 0

Support Reactions: R1= -12.43 kN ( ) , R4= -17.57 kN ( )

Note: Applied forces and displacements are rightwards (positive). Support
reactions are leftwards (negative); opposite to the direction shown in the figure.
Equilibrium: F2+F3+R1+R4=0 24
 k1 k2 k3
R1 1 2 3 4 R4
/\/\/\/\/\/\ /\/\/\/\/\/\ /\/\/\/\/\/\
u1=0
u4=0.05
Example 1.2:
Solve for joint displacements and support reactions for the spring assembly if
the left end is fixed and the right end is pulled out by u4=0.05 m. No other
forces are applied.
k1=1200 kN/m, k2=1800 kN/m, k3=1500 kN/m

 0   3000 − 1800 u2 

Use equation (1.4) and solve:  =  u 
1500.u 4  − 1800 3300  3 
Displacements: u2=0.0203 m, u3= 0.0338 m
R1 = −1200u2
Use equation (1.5) :
R4 = −1500u3 + 1500u4
Support Reactions: R1= -24.32 kN ( ) , R4=24.32 kN ( )
Note: One should apply a force of 24.32 kN rightward at the right end to
achieve an end displacement of u4=0.05 m rightward. 25
 Chapter 2
Simple bar element

26
 2.1 Classical formulation of a bar under axial distributed
loading intensity q(x)
x,u(x)
L
Fixed end q
d  du 
Governing differential equation: −  EA  = q(x)
dx  dx  (2.1)
du
Strain: =
dx
du
Stress:  = E = E
dx
du (2.2)
Stress resultant : P = A = EA
dx
L 2
1  du  (2.3)
Strain Energy: U =  EA  dx
2 0  dx 
27
2.2 A simple bar element as an equivalent spring element
Le
1 2
u1,F1 u2,F2
EAe=Elastic rigidity
1 /\/\/\/\/\/\ 2
u1,F1 u2,F2
k=EAe/Le
E=Elastic Modulus Ae=Constant section area
Equilibriu m :
 F1   EAe / Le − EAe / Le   u1  e  1
u
 = e   = [ K ]  (2.4)
 F2  − EA / L
e e e
EA / L  u 2  u2 
[ K e ] = bar element stiffness matrix

For connected element: {F e } + {R e } = [ K e ]{d e } (2.5)

28
 2.3 An assembly of bar elements as spring elements
1 2 3 4 R4
R1
L1 L3
F1( applied ) F2 ( applied ) L2 F3( applied ) F4 ( applied )
Fixed
k k2 k3 R4
end; R 1 1 2 3 4
1 /\/\/\/\/\/\ /\/\/\/\/\/\ /\/\/\/\/\/\
u1=0 u4=0
F1( applied ) F2 ( applied ) F3( applied ) F4 ( applied )
Equivalent spring stiffnesses k1=EA1/L1 k2=EA2/L2 k3=EA3/L3
Assembly :
 F1( applied) + R1 = ?   k1 − k1 0 0  u1 = 0 
    u = ? 
 F 2 ( applied)  − k1 k1 + k 2 − k2 0   2 
 = 
 F 3 ( applied)   0 − k2 k 2 + k3 − k 3   u3 = ? 
 F4 ( applied) + R4 = ?  0 0 − k3

k3  u4 = 0
 

{F G } + {RG } = [ K G ]{d G } (2.6)

29
Global stiffness matrix
 R1 1 L1 2 L2 3 L3 4
R4

Fixed F2( applied ) = 1(kN ) F3( applied ) = 2(kN )

Fixed
end; end;
k1 k2 k3
u1=0 R 1 2 3 4 R u4=0
1 /\/\/\/\/\/\ /\/\/\/\/\/\ 4
/\/\/\/\/\/\

Example 2.1: 1(kN ) 2(kN )

Solve for joint displacements and fixed support reactions for the bar
assembly under the given loading. F2=1 kN, F3= 2 kN.
Young’s Modulus E=200 GPa,
A1=6(10-6)m2, A2=9(10-6)m2, A3=7.5(10-6)m2, L1=L2=L3= 0.1m.

Equivalent spring stiffnesses

k1=EA1/L1= 12000 kN/m , k2=EA2/L2=18000 kN/m k3=EA3/L3 =15000 kN/m
Solution as in Example 1 (for springs)

u2=0.000104 m, u3= 0.000117 m R1= -1.243 kN ( ) , R4= -1.757 kN ( )

30
 2.4 Strain and stress resultant in bar element

Element Strain (approximated as constant in the element)

u2 − u1  − 1 1  u1 
 =
h
= e  
L e
L Le  u2 
 h = [ B]d e 
(2.7)
[ B] = Strain − displaceme nt matrix

Element Stress Resultant Le

(approximated as constant in the element)
u1 1 2 u2

P = EA .[ B]d 
e e e
EAe=Elastic rigidity
In general
 = [ D][ B]d 
h e
(2.8)
[ D] = Element elastic rigidity matrix

31
 2.5 Finite Element Formulation of the bar element
Governing differential equation for the bar
Le element with only nodal (point) loads :
1 2
u1,F1 u2,F2
d 2u h
EA 2 = 0 (2.9)
x = 0,  = −1 x = L , = 1
e
dx
Approximation of Displacement (assumed linear polynomial)
2x
u h ( ) = N1 ( )u1 + N 2 ( )u2  = e −1
L
 
u h ( ) = N  d e (2.10)

Shape functions (interpolations between nodes):

N1 ( ) = (1 −  ) / 2 N2 ( ) = (1 +  ) / 2 (2.11)

Axial Strain u2 − u1  − 1 1  u1 

h = = e  
L e
L Le  u2  (2.12)
 h = [ B]d e 
32
[ B] = Strain − displaceme nt matrix
Example 2.2: Analysis of a bar under uniformly distributed load q
using 3 equal elements
Fixed end; L
u1=0
q
e Le=L/3 Le Le Discretization
R1 = −q(3L ) 1 2 3 4
= −qL
e
F1( applied ) = qL / 2 F2( applied ) = qLe F3( applied ) = qLe F = qLe
/2
4 ( applied )

FE Equation after assembly of the three elements

 F1( applied) + R1 = ?  k1 − k1 0 0  u1 = 0 
   
 F 2 ( applied)  − k1 k1 + k 2 − k2 0  u2 = ?
 =  
 F3 ( applied)   0 − k2 k 2 + k3 
− k3 u3 = ?
   
 F 4 ( applied)   0 0 − k3 k3  u4 = ?

F + R = K  
G
app
G G G
 Example 2.2 (continued):
Fixed end; L
u1=0
q
e Le=L/3 Le Le Discretization
R1 = −q(3L ) 1 2 3 4
= −qL
F1( applied ) = qLe / 2 F2( applied ) = qLe F3( applied ) = qLe F e
4 ( applied ) = qL / 2

Hence the FE Equation for the entire bar:

( )
 qLe =1 / 2 + R1 = ? 
(
 e =1 e=2
 qL / 2 + qL / 2  ) ( 
)  L
with  Le = 
 e=2 
( qL / 2 + qL ) (
e =3
/ 2  )  3
 qLe =3 / 2(  )
(
 EA / Le =1 ) (
− EA / Le =1 ) 0 0  u1 = 0 

=
− EA / L (
e =1
) ( ) (
EA / Le =1 + EA / Le = 2 ) (
− EA / Le = 2 ) 0

u
  2 = ? 

 0 (
− EA / Le = 2 ) ( ) (
EA / Le = 2 + EA / Le =3 ) (− EA / Le =3  u3 = ?
)
 
 0 0 (
− EA / Le =3 ) ( EA / Le =3 )  u4 = ?
(qL / 6) + R1 = ?  (3EA / L ) − (3EA / L ) 0 0  u1 = 0
 ( qL / 3)  − (3EA / L ) (6 EA / L ) − (3EA / L ) 0  u = ?
     2 
  =  
 ( qL / 3)   0 − (3 EA / L ) (6 EA / L ) − (3 EA / L ) u
 3 = ? 
 

 (qL / 6)   0 0 − (3EA / L ) (3EA / L )   u4 = ?
FE Solution for nodal displacements
2  2 −1  u2  u2  2
5 2
5 / 18
qL   3EA   u    qL   qL  
2
  =  − 1 2 − 1  3  u3  = 8
  = 8 / 18 
6   L u  u  18EA 9 EA 9 / 18
1
  
 − 1 1 
 4   4    
FE Solution for element strains: Analytical solution for the strain (linear in x):
u 2 − u1 u 2 − 0 5qL L qL  x 
 e =1 = = = Le = = 1 − 
L e
(L / 3) 6 EA 3 EA  L 
u −u qL
 e=2 = 3 e 2 = Solution for Support Reaction:
L 2 EA − qL 
 0 
 =
e =3 u 4 − u3
=
qL
=
qLe R = K  − F 
G G G G 
=
0


Le
6 EA 2 EA   35
 0 
 h
p

h2=b2+p2

Pythagoras
37
 2.6 FEA as best fit ?
The best fit strain vector is the orthogonal projection of
the analytical strain vector, onto a given subspace.
A best fit  to  satisfies the
following Theorem (Pythagoras)
2 2 2
 − =  −
 = Best − fit
B FEA strain is the best-fit strain at the element level.
h 2 2 h 2
 =h
 − =  − (2.13)
Norm : a =  a, a  [D]=elastic rigidity of element
 a Db
T
Inner − Product :  a, b = (EA for the axially loaded bar)
element


2
=    [ D] dx = 2U
T
 h 2
=    [ D] dx = 2U
h T h h
 − h 2
 
T
 
=   −  h [ D]  −  h dx
x x x

The Energy of the Error= Error of the Energies

We need a paradigm to show - 38
How FEA turns out to be the best fit to the analytical solution.
 Chapter 3
Simple Classical Euler beam
element

39
Galileo was the first person to Euler derived the equation for
perform experiments on a the thin beam
cantilever beam 40
3.1 Assumptions of Simple (Euler-Bernoulli) Thin Beam Theory
•The deflection of the neutral axis of the beam is small.

•Beam material is within the linear elastic limit (Hooke’s law applies).

• Plane sections normal to the beam axis before bending remain so even after bending.

This automatically implies that the transverse shear deformation in the beam is zero.
(Approximately true for thin beams)

(This is perfectly true for pure bending only, where the bending moment remains constant
along the beam length)
C
x y y
y Positive Directions
z, w(x) z
w1,F1 w2,F2
(a) (b)
dθ
Le
My My 1 2
θ1,M1
z θ2,M2
dx
(c)
41
 3.2 Classical formulation of a the Euler beam under
transverse distributed loading intensity q(x)
R q(x)
y x
M
z,w V
M dθ z z M V+dV
M+dM
dx
NA Equilibrium:
dM dV d 2M
V= , q=−  − 2
=q
dx dx dx
du ( R + z )d − Rd  z
= = =
Strain: dx R.d R
z
Stress:  = E = E
R
Governing differential equation:
d 2  d 2 w 
NA passes through section centroid:
2
EI 2 = q
P =   .dA =
E dx  dx 
R A
z.dA = 0  z=0
A
(3.1)
Bending moment and M =   .z.dA =
E 2 EI z M
R A
z .dA =  =E = z I =  z 2 .dA
bending stress: A
R R I A
42
 3.3 Basic formulation, properties and use of
the Euler beam element (using direct method)
w1,F1 w2,F2
Le EIe=Elastic
bending rigidity
θ1,M1 1 2
EIe θ2,M2

{F e app} + {R e } = [ K e ]{d e } (3.2)

 F1   k11 k12 k13 k14   w1 

 M  k k 24   1 
 1   21 k 22 k 23
 =  
 F2  k31 k32 k33 k34  w2 
M 2  k 41 k 42 k 43

k 44   2 

kij is the force necessary at degree of freedom i for a unit

displacement at degree of freedom j, with all degrees of
freedom except j locked. 43
3.3.1 Derivation of the Element Stiffness Matrix

44
 (
 k11   12 EI / Le 
3
) k11 = −k 31
k   2 
(
 21   − 6 EI / Le 
K1  =   =  )  k 21 + k 41
k
  
31 − 12 EI / (
L e 3
)
 k11 = −
(
k 41   − 6 EI / L 
e
)
2 Le
 

( )
 k12  − 6 EI / Le 
2
k12 = −k 32
k   
K 2  =  22  =  4 EI / Le 2 
e
k 22 + k 42
k 32   6 EI / L( )  k12 = −
Le
k 42   2 EI / L e
 

( )
 k13  − 12 EI / Le 
3
k13 = −k 33
k   2 
 23   6 EI / Le
K 3  =   =  ( ) 
 k 23 + k 43
k
 33   12 EI / L( )
e 3
 k13 = −
Le
 ( )
k 43   6 EI / Le 2



( )
 k14  − 6 EI / Le 
2
k14 = −k 34
k   
k 24 + k 44
 24   2 EI / Le 
K 4  =   =   k14 = −
k
 34   6 EI / ( )
Le 2
 Le
k 44   4 EI / Le  45
 
Element Stiffness Matrix for both conventions A and Anti-A


(
 12 EI / Le 3 ) ( )
− 6 EI / Le
2
( )
− 12 EI / Le
3
− 6 EI / Le  ( )
2


 
Ke = 
(
− 6 EI / L )
e 2
4 EI / L e
6 EI / L( )
e 2 e
2 EI / L  …(3.3A)

(
− 12 EI / Le 3 ) ( )
6 EI / Le
2
12 EI / Le( )3
6 EI / Le 
2
( )
(
 − 6 EI / L )
e 2
2 EI / L e
6 EI / L( )
e 2
4 EI / L 
e

Observations
• The stiffness matrix is symmetric. kij=kji [ K e ] = [ K e ]T …(3.4)
• C1= - C3 and R1= - R3
• C1=(C2+C4)/Le
• R1=(R2+R4)/Le (but Why?) det[ K e ] = 0 Rank [ K e ] = 2
…(3.5) 46
Element Stiffness Matrix for convention B


( )
 12 EI / Le 3 ( )
6 EI / Le
2
(
− 12 EI / Le )
3
( )
6 EI / Le
2


 
Ke = 
6 EI /( )
Le 2
4 EI / Le (
− 6 EI / Le
2
) 2 EI / Le 

( )
− 12 EI / Le 3 ( )
− 6 EI / Le
2
(
12 EI / Le
3
) ( )
− 6 EI / Le
2

 …(3.3B)
( )
 6 EI / Le
2
2 EI / Le (
− 6 EI / Le
2
) 4 EI / Le 

Observations
• The stiffness matrix is symmetric. kij=kji [ K e ] = [ K e ]T …(3.4)
• C1= - C3 and R1= - R3
• C1=(C2+C4)/Le
• R1=(R2+R4)/Le (but Why?) det[ K e ] = 0 Rank [ K e ] = 2
…(3.5) 47
3.3.2 Properties of the Element Stiffness Matrix


(
 12 EI / Le 3 ) ( )
− 6 EI / Le
2
( )
− 12 EI / Le
3
( ) 
− 6 EI / Le
2

 
K =
e  (
− 6 EI / Le
2
) 4 EI / Le ( )
6 EI / Le
2
2 EI / Le 

(
− 12 EI / Le 3 ) ( )
6 EI / Le
2
( )
12 EI / Le
3
( )
6 EI / Le
2 

(
 − 6 EI / Le
2
) 2 EI / Le ( )
6 EI / Le
2
4 EI / Le 
48
49
50
3.3.3 Element Assembly: Summation of Element Stiffness Matrices by common DOF
Element assembly is performed by adding stiffness and force terms corresponding to the
Common DOF

Example: A tip loaded cantilever with 2 elements

P
d2 d4 d6=0
Fixed end
Element 1 Element 2

d1 d3 d5=0

k11e =1 k12e=1 k13e =1 k14e =1 0 0   d1   P   0 

 e =1      e =1 
k 21
e =1
k 22 e =1
k 23 e =1
k 24 0 0   d 2   0   R2 = 0 
k31
e =1 e =1
k32 e =1
k33 + k11e = 2 e =1
k34 + k12e =2 k13e = 2 k14e= 2   d 3   0   R3 + R1 = 0
e =1 e=2

 e =1 e=2    =   +  e=1 e=2 

  4   0   R4 + R2 = 0
e =1 e =1 e=2 e =1 e=2 e=2
k 41 k 42 k 43 + k 21 k 44 + k 22 k 23 k 24 d
 0 0 e=2
k31 e=2
k32 e=2
k33 e=2  
k34 d 5 = 0  0   R3e = 2 = R5G 
      
k 44  d 6 = 0  0   R4 = R6 
e=2 e=2 e=2 e=2 e=2 G
 0 0 k 41 k 42 k 43

[ K G ]{d G } = {Fapp
G
} + {R G } …(3.6)
or
[ K G ]{d G } = {F G } 51
A segment of a FORTRAN code for assembly of the element stiffness matrices and load vectors.

52
 3.3.4 Incorporation of Boundary Conditions on the assembled structure

The global stiffness matrix [KG] is singular => RIGID BODY displacement permitted.
For Flexible Modes, we need to give a minimum of boundary constraints

det[ K G ] = 0 …(3.7)

For every DOF, we either know (a priori) the force (externally applied)

or

the displacement (specified constraint value).

- This is a fundamental requirement for all boundary value problems.

53
Example: Kinematic and Kinetic Boundary Constraints for a tip loaded cantilever

P
d2 d4 d6=0
Fixed end
Element 1 Element 2

d1 d3 d5=0 Kinetic constraint

k11e =1 k12e=1 k13e =1 k14e =1 0 0   d1 = ?   P   0 

 e =1      e =1 
k 21
e =1
k 22 e =1
k 23 e =1
k 24 0 0   d 2 = ?   0   R2 = 0 
k31
e =1 e =1
k32 e =1
k33 + k11e = 2 e =1
k34 + k12e =2 k13e = 2 k14e= 2   d 3 = ?   0   R3 + R1 = 0
e =1 e=2

 e =1 e=2    =   +  e=1 e=2 

k 41
e =1
k 42 e =1
k 43 + k 21e=2 e =1
k 44 + k 22e=2 e=2
k 23 k 24   d 4 = ?   0   R4 + R2 = 0
 0 0 e=2
k31 e=2
k32 e=2
k33 e=2  
k34 d 5 = 0  0   R3e = 2 = R5G = ? 
      e=2 
 0 0 e=2
k 41 e=2
k 42 e=2
k 43 e=2
k 44   6 d = 0   0 
    4  R = R6
G
= ? 
Kinematic constraint

[ K G ]{d G } = {Fapp
G
} + {RG }
54
 Two Methods to incorporate boundary conditions

A. Elimination Method B. Penalty Method

The global force and the A large number, (Penalty

stiffness matrices are modified Stiffness) is added to the
to incorporate the boundary corresponding diagonal
conditions. element of the stiffness
matrix.
Solutions of Nodal Displacement
Vector satisfy geometric boundary Solutions for the nodal
conditions exactly. DOF displacements satisfy
geometric conditions
approximately.

Larger penalty stiffness

gives better results.

55
 Computer Implementation of the Elimination Method

Let Nb = Total number of boundary constraints

k(j)= k-th DOF associated with j-th boundary constraint (j=1,2…Nb)
d(k)=ak(j) = specified boundary value for the k-th DOF

Fi * = ak ( j ) if i = k ( j)
Nb
Modify the Global Force Vector:
Fi * = F G
i , app −  K iG,k ( j ) ak ( j ) if i  k ( j)
j =1

Modify the Stiffness Matrix

Initialize: [ K *] = [ K G ]

Modify: K kk* = 1.0; K ki* = K ik* = 0 for i  k

56
 Specified kinematic constraints
Modified global force vector

 F1* = F1Gapp − K15G a5 − K16G a6  d5 = a5 d 6 = a6

 G 
 F2 * = F2 app − K 25a5 − K 26a6 
G G

 F * = F G − K G a − K G a  Geometric & Force boundary

F * =  3 3 app
G
35 5 36 6
 conditions for the given cantilever
 F4 * = F4 app − K 45a5 − K 46a6 
G G

 F5 * = a5  a5 = 0 a6 = 0
 
 F1,app = P F2,app = F3,app = F4,app = 0
G G G G
 F6 * = a6

Modified global stiffness matrix

 K11G K12G K13G K14G 0 0 Solution of the following
 G G G G  equation gives the global
 K 21 K 22 K 23 K 24 0 0 displacement vector
 K 31
G G
K 32 G
K 33 G
K 34 0 0
[ K *] =  G G G G  [ K *]{d G } = {F *}
 K 41 K 42 K 43 K 44 0 0
 0 0 0 0 1 0
  ...(3.8)
 0 0 0 0 0 1 57
How the Elimination Algorithm works

[ K *]{d G } = {F *}

0  d1   F1* = F1app − K15G a5 − K16G a6 

G
 K11G K12G K13G K14G 0
 G    G 
 K 21
G
K 22 G
K 23 G
K 24 0 0  2   F2 * = F2 app − K 25a5 − K 26a6 
d G G

 K 31
G G
K 32 G
K 33 G
K 34 0 0 d 3   F3 * = F3Gapp − K 35
G
a5 − K 36
G
a6 
 G G G G   =  
 K 41 K 42 K 43 K 44 0 0 d 4   F4 * = F4 Gapp − K 45 G
a5 − K 46a6 
G

 0 0 0 0 1 0 d 5   F5 * = a5 
    
 0 0 0 0 0  d 
1  6   F6 * = a6 

From the last two rows, it is evident that this automatically gives the required
values (specified a priori) for the constrained DOFs

d5 = a5 d 6 = a6
The first four rows then are employed for the solutions of the other unknown
displacements.

58
A segment of a FORTRAN code for incorporation of the kinematic boundary conditions.

59
 Computer Implementation of the Penalty Method

Let Nb = Total number of boundary constraints

k(j)= k-th DOF associated with j-th boundary constraint (j=1,2…Nb)
d(k)=ak(j) = specified boundary value for the k-th DOF

Modify the Global Force Vector:

Fi * = Cak ( j ) if i = k ( j)
C is a very large number
Fi * = Fi ,Gapp if i  k ( j) (Artificial Stiffness)

Modify the Stiffness Matrix

Initialize: [ K *] = [ K G ]

Modify: K kk* = K kkG + C for i = k

60
 Specified kinematic constraints
Modified global force vector

 F1* = F1Gapp  d5 = a5 d 6 = a6
 G 
 F2 * = F2 app 
 F * = F G 
F * =  3 3 app
G 
Geometric & Force boundary
conditions for the given cantilever
 F4 * = F4 app 
 F * = Ca  a5 = 0 a6 = 0
 5 5

 F6 * = Ca6  F1G,app = P F2G,app = F3G,app = F4G,app = 0

Modified global stiffness matrix

Solution of the following
K G
11 K G
12 K G
13 K G
14 K G
15 K G
16
 equation gives the global
 G G G G G G  displacement vector
K 21 K 22 K 23 K 24 K 25 K 26 
K 
[ K *]{d G } = {F *}
G G G G G G
K K K K K
[ K *] =  31
G
32
G
33
G
34
G
35
G
36
G 
K 41 K 42 K 43 K 44 K 45 K 46 
K G
K G
K G
K G G
K 55 +C K G 
 51 52 53 54 56

G
 K
61 K G
62 K G
63 K G
64
G
K 65 K 66 + C 
G
...(3.9) 61
How the Penalty Algorithm works

[ K *]{d G } = {F *}
N

K
l =1
G
kl d l + Cd k = Cak k-th row operation where k is a constrained DOF.

For very large values of the Penalty parameter C

Cd k  Cak k = 5,6 the required specified constraints are approached
d 5  a5 d 6  a6 closely for the constrained DOFs

K16G   d1   F1app 
G
 K11G K12G K13G K14G K15G
 G G G G G G    G 
 K 21 K 22 K 23 K 24 K 25 K 26  d 2   F2 app 
 K 31
G G
K 32 G
K 33 G
K 34 G
K 35 G
K 36  d 3   F3G 
 G G G G G G    =  app 
 K 41 K 42 K 43 K 44 K 45 K 46  d 4   F4 Gapp 
K G G
K 52 G
K 53 G
K 54 G
K 55 +C G
K 56  d 5   Ca 
 51    5 
G
 K 61
G
K 62 G
K 63 G
K 64 G
K 65 K 66 + C  d 6   Ca6 
G

The first four rows then are employed for the solutions of the other unknown
displacements , wherein d5 and d6 are made to approach the required specified
values forcefully !
62
3.3.5 Method of extracting support reactions at the geometrically constrained DOFs

For the global structure:

{R } = [ K ]{d } − {F }
G G G G
app …(3.10a)

At an element level:

{R } = [ K ]{d } − {F }
e e e e
app …(3.10b)

63
Example 3.1

64
65
 On assembly of the two elements, the global equations are obtained
0  0 
M   0 
   
 P   R3 + R3 = 0
e =1 e=2

  +  e =1 e=2 
0
   4 R + R 4 = 0 
 0   R5 = ?  e = 2

   
 0   R6e = 2 = ? 

 12 − 6 Le − 12 − 6 Le 0 0   d1 = ?

− 6 L
e
( )
4 Le
2
6 Le 2 Le ( )
2
0

0  d 2 = ? 
EI  − 12 6 Le 12 + 12 6 Le − 6 Le − 12 − 6 Le   d 3 = ? 
= 3  2  
( )
Le − 6 L
e
( )
2 Le
2
6 Le − 6 Le ( )
2
4 Le + 4 Le ( ) 2
6 Le ( )
2 Le   d 4 = ?
 0 0 − 12 6 Le 12 6 Le  d 5 = 0
 2  
 0 0 − 6 Le 2 Le( )
2
6 Le ( )
4 Le  d 6 = 0

or  0   0   12 − 6 Le − 12 − 6 Le 0 0   d1 = ?
M   0 
   

 − 6 Le
( )
4 Le
2
6 Le ( )
2 Le
2
0

0  d 2 = ? 
 P   0  EI  − 12 6 Le 24 0 − 12 − 6 Le   d 3 = ? 
 + = e 3  2  
0
   0 ( )
 L − 6 L
e
( )
2 Le
2
0 ( )
8 Le
2
6 Le ( )
2 Le   d 4 = ?
 0   R5e = 2 = ?  0 0 − 12 6 Le 12 6 Le  d 5 = 0
   e=2    
 0   R6 = ?  0 0 − 6 Le ( )
2 Le
2
6 Le ( )e 2
4 L  d 6 = 0

F + R = K d 
G
app
hG G G 66
 Solutions for nodal displacements: d1 =
( )
5P Le
3

+
( )
2M Le
2

=
5PL3 ML2
+
 e L
 L = 
6 EI EI 48EI 2 EI  2

d2 =
( )
P Le
+
2
2MLe PL2 ML
= +
2 EI EI 8EI EI

d3 =
( )
P Le
3

+
M Le ( )
2

=
PL3 ML2
+
3EI 2 EI 24 EI 8EI

d4 =
( )
P Le
+
2
MLe PL2 ML
= +
2 EI EI 8EI 2 EI

Observation: FE Solutions for nodal displacements agree exactly with analytical values

The reaction vector can be obtained from the relation

R = K d − F 
hG G G G
app


0
0


 
 0 
R  hG 
= 0


 
 −P 
−  PL + M 
  2  67
68
Example 3.2 Finite Element Analysis of a cantilever with tip load with multiple elements

P = 1000 N E = 200 GPa

L=6 m A = 0.01 m 2 I = 33.33 10 −6 m4

Observations:
• Exact displacements are obtained at the
at the nodes even with a single element.

• Finer discretization has no effect on the

accuracy of the solutions.

Why ?
• FEM displacement is cubic, just as the
exact (analytical) solution for the
concentrated, nodal loads.

For nodal loads alone, a single element can

therefore capture exact displacement, stress
and strain fields.

Since the FEM solution coincides with the

exact one, no further improvement in the
results can be expected with finer meshing. 69
 3.4 Equivalent nodal loads for the Euler beam element
under distributed loading
How do we take care of distributed loading over the element?
We need to find equivalent nodal loads that replace the distributed loading.

Adage: Finite Elements can sense only nodal loads

Governing differential equation for nodal (point) loads (FEA) d 4 wh
EI =0
Displacement Approximation (cubic polynomial in x dx 4 ..(3.11)
within element) wh = ax 3 + bx 2 + cx + d
BM (linear in x within element) & SF (constant within element)
w2,F2
w1,F1
Equilibrium equation for a beam element with distributed load
Le
F + R  = K d 
e he e e
θ2,M2

When all the element DOF are locked, θ1,M1

F + R = 0
e he
 F = −R ; d = 0
e he e
…(3.12)

Equivalent nodal loads for distributed loading is negative of the 70

fixed end reactions
F = −R 
e he
when d = 0
e
q

Le=L/2 Le=L/2
Fixed end reactions on element ends
-q(L/2)/2 -q(L/2)/2 -q(L/2)/2 -q(L/2)/2
q q

q(L/2)2 /12 -q(L/2)2 /12 q(L/2)2 /12 -q(L/2)2 /12

Equivalent nodal loads

q(L/2)/2 q(L/2)/2 q(L/2)/2 q(L/2)/2
q(L/2)2 /12
-q(L/2)2 /12 -q(L/2)2 /12 q(L/2)2 /12

q(L/2)/2 q(L/2) q(L/2)/2

q(L/2)2 /12
-q(L/2)2 /12 71
72
 3.5 Displacement, strain and stress resultant interpolation
in the Euler beam element
Governing differential equation d 4 wh
EI =0
for nodal (point) loads in FEA : dx 4

Solution (even for distributed loading) ! w1,F1 w2,F2

bending displacement is assumed to be Le
cubic in x,
θ1,M1
w = ax + bx + cx + d
h 3 2
…(3.13) θ2,M2

2 h
1 d w
bending curvature of neutral − axis,  h = = (linear )
R dx 2
EI d 2 wh
bending moment of sec tion, M =h
= EI (linear )
R dx 2

…(3.14)
73
 Shape Functions for displacement in the Euler Beam element
(Convention A)

dw h
At node 1, x = 0, w (0) = w1 ,
h
= −1
dx x =0

dw h
At node 2, x=L, e
w ( L ) = w2 ,
h e
= − 2
dx x = Le

1 0 0 0  a0   w1 
0 − 1 0 0 a   
   1  =  1  w1,F1 w2,F2
1 Le Le
2
( ) ( )
Le
3
 a 2  w2 

0 − 1 − 2 L
e
( )
− 3 Le
2
 a3   2 
Le

Aa = d e  θ1,M1
θ2,M2

wh ( x) = 1 x x 2 x a = 1 x x x A d 
3 2 3 −1 e

= N1 N2 N3 N4  d 
e

= N  d 
e …(3.15)
74
Shape Functions for displacement in the Euler Beam element (Convention A)
w h ( ) = N1 w1 + N 21 + N 3 w2 + N 4 2
…(3.16)
N1 ( ) = 1 − 3 2 + 2 3 N 2 ( ) = − Le (1 −  ) 2
…(3.17A)
N 3 ( ) = 3 − 2
2 3
N 4 ( ) = L  (1 −  )
e 2

These are the Hermite Cubics (for C1 continuity elements)

 = x / Le is the non-dimensional coordinate in the element

w1,F1 w2,F2
Le

θ1,M1 θ2,M2

75
Bending Curvature (Generic Strain) :

d 2 wh ( ) 1 d 2 wh ( )
 =
h
=
dx 2 (Le )2 d 2
 w1 
 1   
=  
 (−6 + 12 ) − Le (6 − 4) (6 − 12 ) − Le (6 − 2)  1  = [ B]{d e }
 (Le )2 
  w2 
 2 

Bending Moment (Generic Stress Resultant) : …(3.18)

M h = EI yy  h = EI yy [ B]{d e } …(3.19)

76
Shape Functions for displacement in the Euler Beam element (Convention B)

w ( ) = N1 w1 + N 21 + N 3 w2 + N 4 2
h
…(3.16)

N1 ( ) = 1 − 3 2 + 2 3 N 2 ( ) = + Le (1 −  ) 2
N 3 ( ) = 3 2 − 2 3 N 4 ( ) = − Le 2 (1 −  ) …(3.17B)

These are the Hermite Cubics (for C1 continuity elements)

 = x / Le is the non-dimensional coordinate in the element

Example 3.3

(One element)

(Two elements)

Analytical BM distribution
78
One element solution:
 q0 Le / 2   0   12 − 6 Le − 12 − 6 Le   d1 = ?

( )e 2  
− q0 L / 12  0 
 
EI − 6 L
e
( )
4 Le
2
6 Le ( ) 2 
2 Le   d 2 = ? 
  +  e=1 = e 3   
 q0 L / 2   R3 = ? L  − 12
e
( ) 6 Le 12 6 Le  d 3
2
= 0
( )
 q0 Le 2 / 12  R4 e=1 = ? − 6 L
e
( )
2 Le
2
6 Le ( )
4 Le  d 4 = 0

 d1  q0 (Le ) / 8EI 

4

   = 
d
 2  0 q (L )
e 3
/ 6 EI 

Displacement Interpolation in the element
( )
w h ( ) = 1 − 3 2 + 2 3 d1 − Le (1 −  ) 2 d 2
Bending Moment Interpolation in the element

e 2 
M = h − EI
  d1 

(−6 + 12 ) − L (6 − 4)   = −q0 L  − 
e 1
( )
(L )
e 2
d 2   2 12 

x   x  x  
2 2
q0 L4 
w( ) = 1 −   3 + 2 +   
24 EI  L   L L 
Analytical Solutions:
2
d 2w q0 L2 x
M ( ) = − EI 2 = −   79
dx 2 L
CANTILEVER BEAM ANALYSIS USING A SINGLE EULER BEAM
ELEMENT OF LENGTH L
Uniformly distributed loading is q per unit length.

80
Two elements solution:

81
Two elements solution (continued):
 q0 Le / 2   0 

 0− q( )
Le 2
/ 12
 
  0 

 q0 L e   0 
 + 
 0   0 
 q0 L / 2   R5 = ?
e

   
( ) 2
 q0 Le / 12   R6 = ?

 12 − 6 Le − 12 − 6 Le 0 0   d1 = ? 

 − 6 Le
( )
4 Le
2
6 Le ( )
2 Le
2
0

0  d 2 = ?


EI  − 12 6 Le 12 + 12 6 Le − 6 Le − 12 − 6 Le   d 3 = ? 
=  2  
( )
e 3
L − 6 L
e
( )
2 Le
2
6 Le − 6 Le ( ) 2
4 Le + 4 Le ( )2
6 Le ( )
2 Le   d 4 = ? 
 0 0 − 12 6 Le  d 5 = 0 
  
 0 0 − 6 Le ( )
2 Le
2
 d
  6 = 0 

d1 =
( )
q0 2 Le
4

=
q0 L4
, d2 =
( )
q0 2 Le
3

=
q0 L3
8EI 8EI 6 EI 6 EI
Solutions Nodal DOF

d3 =
( )
17q0 2 Le
4

=
17q0 L4
, d4 =
( )
7q0 2 Le
3

=
7q0 L3
384 EI 384 EI 48EI 48EI 82
CANTILEVER BEAM ANALYSIS USING TWO EULER BEAM ELEMENTS
Uniformly distributed loading is q per unit length.

Le=L/2 Le=L/2
x

q ( Le ) 2 Bending strain
12 EI Analytical (Curvature)
(quadratic)
(M / EI ) = qx
2
/ 2 EI qL2
2 EI
FEM
(linear) q ( Le ) 2
12 EI
Shear Force
qL
83
Example 3.4 Finite Element Analysis of a cantilever loaded uniformly with distributed
loading, using multiple elements

q = 100 N / m E = 200 GPa

L=6 m A = 0.01 m 2
I = 33.33 10 −6 m4
Example 3.4 (continued)
Finite Element Analysis of a cantilever loaded uniformly with distributed
loading, using multiple elements

Lateral Deflection w (mm) at the nodal points from FEM analysis

with various number of elements and comparison with exact values.

x (m) w w( N = 1) w( N = 2) w( N = 4) w( N = 6)
(Exact)
0 0 0 0 0 0
1 0.1206 0.1207
1.5 0.2563 0.2565
2 0.4300 0.4304
3 0.8606 0.8615 0.8615 0.8614
4 1.3600 1.3610
4.5 1.623 1.625
5 1.8906 1.8920
6 2.432 2.432 2.432 2.432 2.432
Example 3.4 (continued)
Finite Element Analysis of a cantilever loaded uniformly with distributed
loading, using multiple elements

Slopes (in 10 -3 radians) at the nodal points from FEM analysis with
various number of elements and comparison with exact values.

x (m)   ( N = 1)  ( N = 2)  ( N = 4)  ( N = 6)
(Exact)
0 0 0 0 0 0
1 -0.227 -0.227
1.5 -0.312 -0.312
2 -0.380 -0.380
3 -0.473 -0.473 -0.473 -0.473
4 -0.520 -0.520
4.5 -0.532 -0.532
5 -0.537 -0.537
6 -0.540 -0.540 -0.540 -0.540 -0.540
Example 3.4 (continued)

This picture shows how

the FEA strains (linear)
converge to the analytical
solution (quadratic) by
best-fitting the exact
element strain (curvature)
with increasing number
of elements.

FEA strains in the

elements are projections
(shadows) of the exact
strain field.

Does FEA somehow

‘sense’ the exact strain
‘a priori’ and tries
to yield results
in a best-fit manner ?
 3.6 FEA as best fit Projection of Strain

h =

B h 2 2 h 2
 − =  −

…(3.20)
Norm : a =  a, a 

 Db
 T
Inner − Product :  a, b = a
element

[D] is the elastic rigidity of the element

(EI for the Euler beam)

The Energy of the Error= Error of the Energies

88
 Chapter 4
Transformation of co-ordinates
for the Plane Truss and Plane Frame
Elements

89
 4.1 What is meant by transformation of co-ordinates ?

Consider a vector A in the plane, being

observed from two reference frames y* a2 y A
S and S*
A = a1i + a 2 j = a1*i * + a 2* j* a2 * a1* x*

A = i
a 
j 1  = i * *

 a1* 
j  *
a 2  a1 x
a 2  
A = i 
j{ A} = i * 
j* { A* }
Invariance of magnitude of a
Using trigonometry:
vector upon transformation
 a1*   cos  sin    a1 
 * =    { A* }T { A* } = { A}T { A}
a2  − sin  cos   a2 
{ A}T [T ]T [T ]{ A} = { A}T { A}
{ A*} = T {A} (4.1)  [T ]T [T ] = [ I ]

T −1
The transformation matrix is orthogonal [T ] = [T ] (4.2)
90
-for isometric transformation
 4.2 Transformation of element matrices
It is often required to transform the equilibrium equations of
an element from its own local co-ordinates to a global co-ordinate
system, common to all elements of the domain.

 
{F e } = [ K e ]  e in local co-ordinates
[Te] is the element transformation matrix
{F e } = [T e ]{F eG } {d e } = [T e ]{d eG }
[T e ]{F eG } = [ K e ][T e ]{d eG }

Pre-multiply both sides by [Te]T

[T e ]T [T e ]{F eG } = [T e ]T [ K e ][T e ]{d eG }
Transformation rule
{F } = [T ] [ K ][T ]{d }
eG e T e e eG

[ K eG ] = [T e ]T [ K e ][T e ]
because [T e ]T = [T e ]−1 {F eG } = [T e ]−1{F e } = [T e ]T {F e }
{F eG } = [ K eG ]]{d eG } (4.3)
in global co-ordinates 91
 4.3 Transformation of the 4 DOF bar (plane truss) element

 
{F e } = [ K e ] d e
in local co-ordinates
 F e1, L  1 0 −1 0 u e1, L 
    
 0  EAe  0 0 0 0  v e1, L 
 e = e  
 F 2, L  L − 1 0 1 0 u e 2, L 
(4.4)
 0   
0 0 0 0  v e1, L 
Transformation of co-ordinates [Te] : element transformation matrix
 u e1, L   u e1,G   cos  sin  0 0 
 e   e  − sin  cos  0 0 
 v 1, L  e  v 1,G  [T ] = 
e
(4.5b)
 e  = [T ] e   0 0 cos  sin  
u 2, L  u 2,G   
 0 0 − sin  cos  
 v e 2, L  v e 2,G  (4.5a)
    [T e ]T = [T e ]−1
Transformation rule
{F eG } = [ K eG ]{d eG }
[ K eG ] = [T e ]T [ K e ][T e ]
e −1
in global co-ordinates
{F } = [T ] {F } = [T ] {F }
eG e e T e 92
(4.6)
 4.4 Transformation of the 6 DOF frame element (in the plane)
Convention B is adopted here

A typical plane frame structure

 P1   cos  sin  0 0 0 0  FX 1   u1, L   u1,G 

F  v  v 
 1
− sin 
 cos  0 0 0 0  FY 1   1, L   1,G 
 M 1   0 0 1 0 0 0  M 1  1, L   
e  1,G 
  =     = [T ] 
 P2   0 0 0 cos  sin  0  FX 2  u u
 F2   0  2, L   2 ,G 
0 0 − sin  cos  0  FY 2 
      v2 , L   v2 ,G 
M 2  L  0 0 0 0 0 1  M 2 G    
 2, L  L  2,G G

{F e } = [T e ]{F eG } & {d e } = [T e ]{d eG } …(4.7)

93
  
{F e } = [ K e ] d e in local co-ordinates …(4.8)
 EA / Le 0 0 − EA / Le 0 0 

 0 ( )
12 EI / Le
3
( )
6 EI / Le
2
0 ( )
− 12 EI / Le
3
( )
6 EI / Le 
2 

 0
[K ] = 
e ( )
6 EI / Le
2
4 EI / Le 0 ( )
− 6 EI / Le
2
2 EI / Le 

− EA / L
e
0 0 EA / Le 0 0 
 0

( )
− 12 EI / Le
3
( )
− 6 EI / Le
2
0 ( )
12 EI / Le
3
( )
− 6 EI / Le 
2


 0 ( )
6 EI / Le
2
2 EI / Le 0 ( )
− 6 EI / Le
2 e
4 EI / L 

- element’s local stiffness matrix (Convention B) …(4.9)

Transformation rule
[ K eG ] = [T e ]T [ K e ][T e ]
[T e ]T = [T e ]−1 …(4.10)
e −1
{F } = [T ] {F } = [T ] {F }
eG e e T e

in global co-ordinates {F eG } = [ K eG ]]{d eG } …(4.11) 94

 y
2 3 4.5 Inclined supports
Global displacement vector in xy co-ordinate system

u = u1x , u1 y ,u2 x , u2 y ,u3 x , u3 y ,u4 x , u4 y , T


1 Inclined support at node 4
x 4 New Global displacement vector
(with components at node 4 along x’y’ co-ordinate system)

u ' = u1x , u1 y ,u2 x , u2 y ,u3 x , u3 y,U 4 x' ,U 4 y ', T

Step 1. Determine the required transformation matrix [T] Step 2. Transform Global
 u1x  1 0 0 0 0 0 0 0   u1x  Stiffness Matrix
 u    
 1 y  0 1 0 0 0 0 0 0   u1 y 
 u 2 x  0 0 1 0 0 0 0 0  u2 x  [ K 'G ] = [T ][ K G ][T ]T
    
 u 2 y  0 0 0 1 0 0 0 0  u2 y 
 =   or {u '} = [T ]{u}
u
 3 x  0 0 0 0 1 0 0 0   u3 x 
 Step 3. Apply the boundary
 u3 y   0 0 0 0 0 1 0 0  u3 y  condition
    
 U 4 x '  0 0 0 0 0 0 cos  sin   u4 x 
 U 4 y' = 0
U 4 y ' = 0 0 0 0 0 0 0 − sin  cos   u4 y 
  

…(4.12) 95
 4.6 The 12 DOF Three-dimensional Space Frame Element
Three-dimensional Transformation Matrix
t 0 0 0
0 cos  x , X cos  x ,Y cos  x , Z 
t 0 0  
[T e ] =  ; t = cos  y , X cos  y ,Y cos  y , Z 
0 0 t 0
   cos  z , X cos  z ,Y cos  z , Z 
0 0 0 t
x X 
   
y
  = [t ]Y 
z Z 
  LOCAL  GLOBAL
Transformation of element Stiffness and Force from local to global co-ordinates:

[ K eG ] = [T e ]T [ K e ][T e ]
{F eG } = [T e ]T {F e }

[ disp. w (along z axis) & rotation ϴy positive directions with Convention Anti-A
96
& disp. v (along y axis) & rotation ϴz positive directions with Convention B ]
Element Stiffness Matrix of the 3D frame element in local-coordinates:
[ disp. w (along z axis) & rotation ϴy positive directions with Convention Anti-A
& disp. v (along y axis) & rotation ϴz positive directions with Convention B

[K e ] =

97
Typical Space Frame Structures

98
 Chapter 5

Analysis of Plane Truss and Plane Frame

(For MATLAB commands, programs from the following reference are

used,
Peter Kattan. “MATLAB Guide to Finite Elements”, Springer 2008)

99
 Example 1: Analysis of a Plane Truss (pinned joints)
3
Analyse the given truss as shown.

3 Section rigidity EA is constant

2 10 kN
L for all members.
1 10 kN
1 L 2 Element Stiffness
Element 1 , α=00 [ K e G ] = [T e ]T [ K e L ][T e ] By equations
(4.5b) & (4.6)
for α=00
2 4 1 0 −1 0
0 0 0 0
1 3 EA 
[ K e =1G ] =
1 L 2 L − 1 0 1 0
 
0 0 0 0 100
3 Element 2 , for α=1350

 (1 / 2) (−1 / 2) (−1 / 2) (1 / 2) 
3 (−1 / 2) (1 / 2) 
2 10 kN [ K e=2G ] =
EA  (1 / 2 ) ( −1 / 2) 
L 2 L (−1 / 2) (1 / 2) (1 / 2) (−1 / 2)
 
 (1 / 2) ( −1 / 2) ( −1 / 2) (1 / 2 ) 
1 L 2 10 kN

6 6 Element 3 , for α=-900

5 3 5
3 0 0 0 0
0 1 0 − 1
2 3 e =3 EA 
[ K G] =
L L 0 0 0 0
4 2  
3 1 1 0 − 1 0 1
2
101
3
Element Assembly
3
2 10 kN [ K G ] = [ K e=1G ] + [ K e=2G ] + [ K e=3G ]
L
10 kN
1
1 L 2 Global stiffness matrix
1 0 −1 0 0 0 
0 1 0 0 0 − 1 

− 1 0 1 + 1  − 1 −
1 1 
  2 2 2 2 2 2 2 2 
EA  1 1 1 1 
[K ] =
G
0 0 − − 
L  2 2 2 2 2 2 2 2 
0 0 1 1 1 1 
 − −
2 2 2 2 2 2 2 2 
 
 0 −1 1 1 1  1 
− − 1 + 
 2 2 2 2 2 2  2 2 
102
 3
Incorporating geometric boundary
3 conditions and applied forces
2 10 kN
L
10 kN
1
L
[ K ]{u } = {F } + {R }
G G G G
1 2
Applied forces Reactions
1 0 −1 0 0 0 
0 1 0 0 0 − 1 
  1 = 0   0   R1 
− 1 0 1 + 1  − 1 −
1 1     R 
  2 2 2 2 2 2 2 2   2 = 0   0   2
EA  1   3 = ?  10 ( N )  0 
4
1 1 1
0 0 − −  = 4 + 
L  2 2 2 2 2 2 2 2   4 = ?  10 ( N )  0 
0 0 1 1 1 1 
− −  5 = 0  0   R5 
 2 2 2 2 2 2 2 2      
   6 = 0  0   R6 

 0 −1 1 1 1  1 
− − 1 + 
 2 2 2 2 2 2  2 2 
103
 Solution of equations
 1  1 
 1 +  −   3 = ? 10 4 ( N )
EA  2 2  2 2
  = 4 
3 L  1 1   4 = ? 10 ( N )
−
 2 2 2 2 
3
2 10 kN
L
10 kN  3 = ?  2.0  10 4 L 
1  =  
1 L 2  4 = ? 4.8281 EA 

 2  10 4 L   4.8281  10 4 L 
 3 =    4 =  
 EA   EA 

104
Evaluation of support reactions {RG } = [ K G ]{u G } − {F G }
1 0 −1 0 0 0 
0 1 0 0 0 −1  1 = 0 
 R1      0 
− 1 0 1 + 1  − 1 1 1   2 = 0  
R  − 
 2     2  10 4
L   0
 2 2 2 2 2 2 2 2
  3 =
    4 
 0  EA  1 1 1 1   EA   − 10 ( N )
  =  0 0 − − 
0 L  2 2 2 2 2 2 2 2   4.8281  10 4 L  10 4 ( N )
1   4 = 
 
 R5   1 1 1
 0 0 − −   EA   0 
  2 2 2 2 2 2 2 2    
 R6    5 = 0  0 
 0 −1 1 1 1  1    

− − 1 +   6 =0
 2 2 2 2 2 2  2 2  

-10 kN  R1  − 2  10 4 ( N )
10 kN R   
 2  0 
 0   0 
10 kN  = 
0
   0 
-20 kN  R5   10 4 ( N ) 
10 kN    
 R6   − 10 4 ( N ) 
105
 Element Stress Resultants

Step 1: Get local displacement vector for the element

{ e L } = [T e ]{ e G }

Step 2: Get element strain

u 2 − u1
 =
h

Le

Step 3: Evaluate element stresses and stress resultants

 u2 − u1   u2 − u1 
 = E = E  e ,
h h
P =  A = EA
h
e 
 L   L 

106
Stress Resultant for element 1
Local displacement vector for the element
{ e L } = [T e ]{ e G }
for α=00 e =1
 0   0 
0    
e =1
u1  1 0 0 0 0
v  0     
 1 1 0 0   2  10 4
L     2  10 4
L 
  =   3 =
 EA    =    
0 0     EA 
u2  0 1   
v2     4 . 8281  10 4
L   4.828110 L 
4
0 0 0 1  4 =  
  EA  EA

    
Stress resultant in the element
 u 2 − u1 
P = EA  = +2  10( N )
 L 
Positive, therefore tensile in nature 107
 Stress Resultant for element 2
Local displacement vector for the element
0
{ e
L } = [T ]{ G }
e e
for α=135
e=2
  2 10 4 L     2 10 4 L  
   3 =  EA  
e=2      
u1  (−1 / 2 ) (1 / 2 ) 0 0
v         EA  
 1 (−1 / 2 ) (−1 / 2 ) 0 0   =  4.828110 L   6.828110 4 L 
4
  =  4   =  − 
 (−1 / 2 ) (1 / 2 )  
u2  
0 0
  EA   2 .EA 
v2  (−1 / 2 ) (−1 / 2 )  0   0 
 0 0
   
 0   0 

Stress resultant in the element

 u2 − u1 
P = EA  = − 2 10 4 ( N )
 2L 
Negative, therefore compressive in nature

Stress Resultant in element 3 in zero

108
 P1 20
 e=1 = = kN / m 2
= 2  10 5
kN / m 2

Calculation of A 110 − 4
member stresses P − 2 10
 e=2 = 2 = kN / m 2
= − 1. 414  10 5
kN / m 2

Tensile: Positive
A 110 −4
Compressive: Negative P 0
 e =3 = 3 = kN / m 2
= 0 kN / m 2

A 110 − 4 109
MATLAB commands for Example 1 Analysis of Truss (Units in m and kN)
E = 210 Gpa = 210 109 N / m 2 = 210 106 kN / m 2
3 A = 1 cm 2 = 110− 4 m2
L = 2 m  L1 = L3 = L = 2m L2 = L 2 = 2.828 m
3 2 10 kN
L >> E=210e6
E = 2.1000e+08
1 10 kN
1 L 2 >> A=1e-4
A = 1.0000e-04

>> L1=2
L1 = 2

>> L2=2.828
L2 = 2.8280

>> L3=2
L3 = 2
>> k1=PlaneTrussElementStiffness(E,A,L1,0)
k1 =

10500 0 -10500 0
0 0 0 0
-10500 0 10500 0
0 0 0 0

>> k2=PlaneTrussElementStiffness(E,A,L2,135)
k2 =

3712.9 -3712.9 -3712.9 3712.9

-3712.9 3712.9 3712.9 -3712.9
-3712.9 3712.9 3712.9 -3712.9
3712.9 -3712.9 -3712.9 3712.9

>> k3=PlaneTrussElementStiffness(E,A,L3,-90)
k3 =

3.9366e-29 -6.4292e-13 -3.9366e-29 6.4292e-13

-6.4292e-13 1.0500e+04 6.4292e-13 -1.0500e+04
-3.9366e-29 6.4292e-13 3.9366e-29 -6.4292e-13
6.4292e-13 -1.0500e+04 -6.4292e-13 1.0500e+04

111
>> K=zeros(6,6)
K=

0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0

112
>> K=PlaneTrussAssemble(K,k1,1,2)
K=

10500 0 -10500 0 0 0
0 0 0 0 0 0
-10500 0 10500 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0

>> K=PlaneTrussAssemble(K,k2,2,3)
K=

1.0500e+04 0 -1.0500e+04 0 0 0
0 0 0 0 0 0
-1.0500e+04 0 1.4213e+04 -3.7129e+03 -3.7129e+03 3.7129e+03
0 0 -3.7129e+03 3.7129e+03 3.7129e+03 -3.7129e+03
0 0 -3.7129e+03 3.7129e+03 3.7129e+03 -3.7129e+03
0 0 3.7129e+03 -3.7129e+03 -3.7129e+03 3.7129e+03

>> K=PlaneTrussAssemble(K,k3,3,1)
K=

1.0500e+04 -6.4292e-13 -1.0500e+04 0 -3.9366e-29 6.4292e-13

-6.4292e-13 1.0500e+04 0 0 6.4292e-13 -1.0500e+04
-1.0500e+04 0 1.4213e+04 -3.7129e+03 -3.7129e+03 3.7129e+03
0 0 -3.7129e+03 3.7129e+03 3.7129e+03 -3.7129e+03
-3.9366e-29 6.4292e-13 -3.7129e+03 3.7129e+03 3.7129e+03 -3.7129e+03
6.4292e-13 -1.0500e+04 3.7129e+03 -3.7129e+03 -3.7129e+03 1.4213e+04
113
>> k=K(3:4,3:4)
k=

1.4213e+04 -3.7129e+03
-3.7129e+03 3.7129e+03

>> f=[10;10]
f=

10
10

>> u=k\f
u=

1.9048e-03
4.5981e-03

114
>> U=[0;0;u;0;0]
U=
0
0
1.9048e-03
4.5981e-03
0
0

>> F=[0;0;f;0;0]
F=
0
0
10
10
0
0

>> R=K*U-F
R=
-20.0000
0
0.0000
0
10.0000
-10.0000 115
 >> u1=[U(1);U(2);U(3);U(4)]
u1 =
0
0
1.9048e-03
4.5981e-03

>> u2=[U(3);U(4);U(5);U(6)]
u2 =
1.9048e-03
4.5981e-03
0
0
Calculation of member stresses
>> u3=[U(5);U(6);U(1);U(2)]
Tensile stresses are positive
u3 =
Compressive stresses are negative
0
0
P1 20 0
 e=1 = = −4
kN / m 2 = 2 105 kN / m 2
A 110 0
P − 2 10
 e=2 = 2 = −4
kN / m 2 = −1.414 105 kN / m 2
>> sigma1=PlaneTrussElementStress(E,L1,0,u1)
A 110
P 0 sigma1 = 2.0000e+05
 e =3 = 3 = kN / m 2
= 0 kN / m 2
>>sigma2=PlaneTrussElementStress(E,L2,135,u2)
A 110 − 4
sigma2 = -1.4142e+05
>> sigma3=PlaneTrussElementStress(E,L3,-90,u3)
116
sigma3 = 0
Example 2: Analysis of a Plane Frame (rigid joints)

Analyse the given frame as shown.

Section rigidities EA and EI are constant for all members.

1.5 kN/m
2 kN

3kN/m 2m

4m

117
The sectional properties of the members are identical and their values are calculated here.
Area of the section A = bt = 10  2 cm 2 = 20 cm 2 = 20  10 −4 m 2
−8
Area moment of inertia I = bt / 12 = 10  2 / 12 cm = 20 / 3 cm = 6.67  10
3 3 4 4
m4
Young’s Modulus of Steel E = 200 GPa = 200  10 9 Pa

118
Equivalent nodal loads from applied distributed loading

1 kNm 3 kN 3 kN
3 kN
2
2 3
1
3 kN 2 kNm 2 kNm
1 kNm

3 kN 3 kN
2-1=1 kNm 2 kNm

3 kN
2 3
2m
1 4
3 kN 4m 119
 Equivalent nodal loads for the full frame
3 kN 3 kN
2-1=1 kNm 2 kNm

3 kN 2 kN
2 3
2m
1 4
3 kN 4m
1 kNm
5
8 9
6 Degrees of freedom
4 7 (of positive sense)
Free DOF
2 11 Fixed DOF
1 10
120
3 12
Nodal Force-Displacement relationship in an Euler Beam
Element
(Convention B, 6 DOF in the plane, in local co-ordinates)

 
{F eL } = [ K eL ] d eL

 EA / Le 0 0 − EA / Le 0 0 

 0 ( )
12 EI / Le
3
( )
6 EI / Le
2
0 ( )
− 12 EI / Le
3
( )
6 EI / Le 
2 

 0
[K ] = 
eL ( )
6 EI / Le
2
4 EI / Le 0 ( )
− 6 EI / Le
2
2 EI / Le 

 − EA / Le
0 0 EA / Le 0 0 
 0

( )
− 12 EI / Le
3
( )
− 6 EI / Le
2
0 ( )
12 EI / Le
3
( )
− 6 EI / Le 2

 0 ( )
6 EI / L e 2
2 EI / Le 0 ( )
− 6 EI / L e 2 e
4 EI / L 

{d eL } = [T e ]{d eG }
{F eL } = [T e ]{F eG }

121
 Transformation of element stiffness matrices
from local to global co-ordinates
Element Stiffness in
Global Co-ordinates:
[ K ] = [T ] [ K ][T ]
eG e T eL e

Transformation Matrix  cos  sin  0 0 0 0

of an element: − sin  cos  0 0 0 0

 0 0 1 0 0 0
[T ] = 
e

 0 0 0 cos  sin  0
 0 0 0 − sin  cos  0
 
 0 0 0 0 0 1

α=-900
for element 3
Element Assembly [ K G ] = [ K e=1,G ] + [ K e=2,G ] + [ K e=3,G ]
Solution of the system of [ K ]{u } = {F } + {R }
G G G G

equations
 1 = 0   3 103 ( N )   R1 = ? 
 = 0    
R = ? 
 2   0   2 
  3 = 0  − 110 ( Nm)  R3 = ? 
3

     

 4 = ?   3  10 3
( N )   0 
  5 = ?   − 3 10 ( N )   0 
3
     
 = ?  − 110 ( Nm)  0 
G  6
3
[ K ] = + 

 7 = ?   2 10 ( N )  
3
0 
  8 = ?   − 3 103 ( N )   0 
     

 9 = ?   2 10 ( Nm)  
3
0 
 = 0  0   R = ?
 10
    10 
11 = 0   0   R11 = ?
 = 0  0   R = ?
 12     12
123 
 Results : for Axial Stiffness of the members very large compared to their
Flexural Stiffness
EAe EI e E is in N/m2,

Le
Le 3
( ) Ie is in m4, Ae is in m2,
Le is in m.
Nodal Displacement Reactions:
 1 = 0 
  = 0   R1 = −4.6 103 (N ) 
 2  
R = 2.25  10 3
( N )

 3 = 0   2 
   R3 = 3.4 10 3
( Nm) 
  4 = 0 . 2 ( m )   
 0 
 5 = −0.1125  10 − 4
( m) 
   0 
  6 = −0.12 (rad )   

  = 0 . 2 ( m )


 
{R G } = [ K G ] u G

− {F G } = 
0 

7
 0 
 8 = −0.1876 10 − 4
( m)   0 
   
  9 = −0 . 03 ( rad )   0 
 10 = 0   R = −3.4 103 ( N )
   10

 11 = 0   R11 = 3.75 10
3
(N ) 
   R = 3.6 103 ( Nm)
 12 = 0   12
124
Equilibrium under External Forces

4m

1.5 kN/m
2 kN

3kN/m 2m
3.4 kNm 3.6 kNm
3.4 kN
4.6 kN

2.25 kN 3.75 kN
125
 Bending Moment distribution
Step 1: Get local displacement vector for the element
{d eL } = [T e ]{d eG }

Step 2: Evaluate Bending Moment distribution

Finite Elements sense only nodal loads.
Approximation of displacement field by cubic polynomial of C1
continuity in non-dimensional co-ordinate ξ=x/Le.
Bending Curvature /Bending Moment distribution by FEA will be linear
in ξ.

 w1e 
 e
   1 
 
2 h 2 h
d w ( ) 1 d w ( ) 1
h = = = ( −6 + 12 ) − Le
( 6 − 4) ( 6 − 12 ) − Le
( 6 − 2)  e
dx 2 Le
2
d 2 Le
2
w2 
 e 
 2 
EI d 2 wh
bending moment distributi on M =
h
= EI
R dx 2
126
FEA captures the linear best-fit 2 3
solutions to the exact bending
moments in the members. M

BM in kNm
Member 1, (1-2)
3
Exact : M ( x) =
2
(2 x − x 2 ) + (1.6 x − 3.4) Mh
FEA : M ( x ) = 1 . 6 x − 2 .4
h

2 2

 M ( x)dx =  M ( x)dx = − 1.6

h

0 0
BM in kNm 2

Member 2, (2-3)
1.5 -3.2 kNm
Exact : M ( x) = (4 x − x 2 ) − (0.75 x + 0.2)
2 3
FEA : M ( x) = −0.75 x + 1.8
h
-
4 4

 M ( x)dx =  M ( x)dx =1.2

h

0 0 Member 3, (3-4) Mh=M

Mh
M The Energy of the Error +
= Error of the Energies 4
1 M −Mh
2
= M
2
− Mh
2
3.6 kNm
127
M
2
=  M 2 dx = (2 EI )U Mh
2
=  M h dx = (2 EI )U h
2
M −M h
2
( ) 2
=  M − M h dx = (2 EI )U error
e e e
 Corrected Bending Moment distribution in the members
1.8 kNm

+
2 kNm -
-0.2 kNm -1.2 kNm
+0.8 kNm -0.2 kNm 1 kNm
-3.2 kNm
2 kNm
2 2 kNm
+ 2 3

1 kNm 2 kNm shift of FE moment

towards hogging to counteract -3.2 kNm
-3.4 kNm sagging curvature from
- 1
1 kNm equivalent end moments
-
3
applied for distributed loading
-2.4 kNm
1 kNm shift of FE moment Bending Moment Mh
towards hogging to counteract by FEA: +
sagging curvature from 4
equivalent end moments Corrected BM
applied for distributed loading at nodal points 3.6 kNm
are encircled; 128
MATLAB Commands for example 2: Analysis of Frame (Units in m & KN)

>> E=200e6
E = 2.0000e+08

>> A=20e-04
A = 2.0000e-03

>> I=6.67e-08
I = 6.6700e-08

>> L1=2
L1 = 2

>> L2=4
L2 = 4

>> L3=2
L3 = 2

129
>> k1=PlaneFrameElementStiffness (E,A,I,L1,90)
k1 =

2.0010e+01 1.2245e-11 -2.0010e+01 -2.0010e+01 -1.2245e-11 -2.0010e+01

1.2245e-11 2.0000e+05 1.2252e-15 -1.2245e-11 -2.0000e+05 1.2252e-15
-2.0010e+01 1.2252e-15 2.6680e+01 2.0010e+01 -1.2252e-15 1.3340e+01
-2.0010e+01 -1.2245e-11 2.0010e+01 2.0010e+01 1.2245e-11 2.0010e+01
-1.2245e-11 -2.0000e+05 -1.2252e-15 1.2245e-11 2.0000e+05 -1.2252e-15
-2.0010e+01 1.2252e-15 1.3340e+01 2.0010e+01 -1.2252e-15 2.6680e+01

>> k2=PlaneFrameElementStiffness(E,A,I,L2,0)
k2 =

1.0000e+05 0 0 -1.0000e+05 0 0
0 2.5013e+00 5.0025e+00 0 -2.5013e+00 5.0025e+00
0 5.0025e+00 1.3340e+01 0 -5.0025e+00 6.6700e+00
-1.0000e+05 0 0 1.0000e+05 0 0
0 -2.5013e+00 -5.0025e+00 0 2.5013e+00 -5.0025e+00
0 5.0025e+00 6.6700e+00 0 -5.0025e+00 1.3340e+01

>> k3=PlaneFrameElementStiffness(E,A,I,L3,-90)
k3 =

2.0010e+01 -1.2245e-11 2.0010e+01 -2.0010e+01 1.2245e-11 2.0010e+01

-1.2245e-11 2.0000e+05 1.2252e-15 1.2245e-11 -2.0000e+05 1.2252e-15
2.0010e+01 1.2252e-15 2.6680e+01 -2.0010e+01 -1.2252e-15 1.3340e+01
-2.0010e+01 1.2245e-11 -2.0010e+01 2.0010e+01 -1.2245e-11 -2.0010e+01
1.2245e-11 -2.0000e+05 -1.2252e-15 -1.2245e-11 2.0000e+05 -1.2252e-15
2.0010e+01 1.2252e-15 1.3340e+01 -2.0010e+01 -1.2252e-15 2.6680e+01 130
>> K=zeros(12,12)
K=

0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0

131
>> K=PlaneFrameAssemble(K,k1,1,2)
K=

Columns 1 through 10:

2.0010e+01 1.2245e-11 -2.0010e+01 -2.0010e+01 -1.2245e-11 -2.0010e+01 0 0 0 0

1.2245e-11 2.0000e+05 1.2252e-15 -1.2245e-11 -2.0000e+05 1.2252e-15 0 0 0 0
-2.0010e+01 1.2252e-15 2.6680e+01 2.0010e+01 -1.2252e-15 1.3340e+01 0 0 0 0
-2.0010e+01 -1.2245e-11 2.0010e+01 2.0010e+01 1.2245e-11 2.0010e+01 0 0 0 0
-1.2245e-11 -2.0000e+05 -1.2252e-15 1.2245e-11 2.0000e+05 -1.2252e-15 0 0 0 0
-2.0010e+01 1.2252e-15 1.3340e+01 2.0010e+01 -1.2252e-15 2.6680e+01 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0

Columns 11 and 12:

0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
132
>> K=PlaneFrameAssemble(K,k2,2,3)
K=

Columns 1 through 10:

2.0010e+01 1.2245e-11 -2.0010e+01 -2.0010e+01 -1.2245e-11 -2.0010e+01 0 0 0 0

1.2245e-11 2.0000e+05 1.2252e-15 -1.2245e-11 -2.0000e+05 1.2252e-15 0 0 0 0
-2.0010e+01 1.2252e-15 2.6680e+01 2.0010e+01 -1.2252e-15 1.3340e+01 0 0 0 0
-2.0010e+01 -1.2245e-11 2.0010e+01 1.0002e+05 1.2245e-11 2.0010e+01 -1.0000e+05 0 0
0
-1.2245e-11 -2.0000e+05 -1.2252e-15 1.2245e-11 2.0000e+05 5.0025e+00 0 -2.5013e+00 5.0025e+00
0
-2.0010e+01 1.2252e-15 1.3340e+01 2.0010e+01 5.0025e+00 4.0020e+01 0 -5.0025e+00 6.6700e+00
0
0 0 0 -1.0000e+05 0 0 1.0000e+05 0 0 0
0 0 0 0 -2.5013e+00 -5.0025e+00 0 2.5013e+00 -5.0025e+00 0
0 0 0 0 5.0025e+00 6.6700e+00 0 -5.0025e+00 1.3340e+01 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0

Columns 11 and 12:

0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0 133
0 0
>> K=PlaneFrameAssemble(K,k3,3,4)
K=

Columns 1 through 10:

2.0010e+01 1.2245e-11 -2.0010e+01 -2.0010e+01 -1.2245e-11 -2.0010e+01 0 0 0 0

1.2245e-11 2.0000e+05 1.2252e-15 -1.2245e-11 -2.0000e+05 1.2252e-15 0 0 0 0
-2.0010e+01 1.2252e-15 2.6680e+01 2.0010e+01 -1.2252e-15 1.3340e+01 0 0 0 0
-2.0010e+01 -1.2245e-11 2.0010e+01 1.0002e+05 1.2245e-11 2.0010e+01 -1.0000e+05 0 0 0
-1.2245e-11 -2.0000e+05 -1.2252e-15 1.2245e-11 2.0000e+05 5.0025e+00 0 -2.5013e+00 5.0025e+00 0
-2.0010e+01 1.2252e-15 1.3340e+01 2.0010e+01 5.0025e+00 4.0020e+01 0 -5.0025e+00 6.6700e+00 0
0 0 0 -1.0000e+05 0 0 1.0002e+05 -1.2245e-11 2.0010e+01 -2.0010e+01
0 0 0 0 -2.5013e+00 -5.0025e+00 -1.2245e-11 2.0000e+05 -5.0025e+00 1.2245e-11
0 0 0 0 5.0025e+00 6.6700e+00 2.0010e+01 -5.0025e+00 4.0020e+01 -2.0010e+01
0 0 0 0 0 0 -2.0010e+01 1.2245e-11 -2.0010e+01 2.0010e+01
0 0 0 0 0 0 1.2245e-11 -2.0000e+05 -1.2252e-15 -1.2245e-11
0 0 0 0 0 0 2.0010e+01 1.2252e-15 1.3340e+01 -2.0010e+01

Columns 11 and 12:

0 0
0 0
0 0
0 0
0 0
0 0
1.2245e-11 2.0010e+01
-2.0000e+05 1.2252e-15
-1.2252e-15 1.3340e+01
-1.2245e-11 -2.0010e+01
2.0000e+05 -1.2252e-15
-1.2252e-15 2.6680e+01
134
>> k=K(4:9,4:9)
k=

1.0002e+05 1.2245e-11 2.0010e+01 -1.0000e+05 0 0

1.2245e-11 2.0000e+05 5.0025e+00 0 -2.5013e+00 5.0025e+00
2.0010e+01 5.0025e+00 4.0020e+01 0 -5.0025e+00 6.6700e+00
-1.0000e+05 0 0 1.0002e+05 -1.2245e-11 2.0010e+01
0 -2.5013e+00 -5.0025e+00 -1.2245e-11 2.0000e+05 -5.0025e+00
0 5.0025e+00 6.6700e+00 2.0010e+01 -5.0025e+00 4.0020e+01

>> f=[3;-3;-1;2;-3;2]
f=

3
-3
-1
2
-3
2

>> u=k\f
u=

1.9991e-01
-1.1250e-05
-1.1995e-01
1.9989e-01
-1.8750e-05
-2.9982e-02 135
>> U=[0;0;0;u;0;0;0] >> R=K*U-F
U= R=
0
0
-4.6001
0 2.2500
0.1999 3.4001
-0.0000 0
-0.1199 0.0000
0.1999 0.0000
-0.0000 0.0000
-0.0300
0
0
0 -0.0000
0 -3.3999
3.7500
>> F=[3;0;-1;f;0;0;0] 3.5999
F=

3
0
-1
3
-3
-1
2
-3
2
0
0 136
0
>> u1=[U(1);U(2);U(3);U(4);U(5);U(6)] >> f1=PlaneFrameElementForces(E,A,I,L1,90,u1)
u1 = f1 =

0 2.2500
0 1.6001
0 2.4001
0.1999 -2.2500
-0.0000 -1.6001
-0.1199 0.8000

>> u2=[U(4);U(5);U(6);U(7);U(8);U(9)] >> f2=PlaneFrameElementForces(E,A,I,L2,0,u2)

u2 = f2 =

1.9991e-01 1.3999
-1.1250e-05 -0.7500
-1.1995e-01 -1.8000
1.9989e-01 -1.3999
-1.8750e-05 0.7500
-2.9982e-02 -1.2000

>> u3=[U(7);U(8);U(9);U(10);U(11);U(12)] >> f3=PlaneFrameElementForces(E,A,I,L1,-90,u3)

u3 = f3 =

0.1999 3.7500
-0.0000 3.3999
-0.0300 3.2000
0 -3.7500
0 -3.3999
0 3.5999
137
>> PlaneFrameElementAxialDiagram(f1,L1)
>> PlaneFrameElementAxialDiagram(f2,L2)
>> PlaneFrameElementAxialDiagram(f3,L3)
>> PlaneFrameElementShearDiagram(f1,L1)
>> PlaneFrameElementShearDiagram(f2,L2)
>> PlaneFrameElementShearDiagram(f3,L3)
>> PlaneFrameElementMomentDiagram(f1,L1)
>> PlaneFrameElementMomentDiagram(f2,L2)
>> PlaneFrameElementMomentDiagram(f3,L3)

138
Stress Resultants in Element 1

139
Stress Resultants in Element 2

140
Stress Resultants in Element 3

141
 Summary of Lecture 1
• A brief history of the genesis of Finite Element Method is presented.
• The elementary spring element has been employed to illustrate the basic matrix methods of finite
element analysis. Furthermore, it has been shown how the spring element formulation can be
extended to the development of the axially loaded bar element.

• Simple and direct equilibrium methods for establishing stiffness matrices of the spring element,
axially load bar element and the Euler beam element have been shown.

•Basic theories of the axially loaded elastic bar and the transversely loaded beam have been briefly
discussed.

•Mathematical properties of the stiffness matrices and transformation matrices have been discussed
with proper explanations.

•Various methods of incorporation of the geometric boundary conditions have been discussed.

• Duality in each DOF of finite elements has been highlighted.

• Detailed computation in finite element analyses of axially loaded bars, transversely loaded beams,
loaded plane trusses and frames have been shown.

• Convergence of FE results towards analytical solutions with finer meshing has been demonstrated
effectively with special reference to the best-fit paradigm of FEA.

•Special MATLAB programs have been employed to demonstrate the FEA procedures for analysis of
the plane truss and frame elements
142
 Bibliography
Variational Principles of Classical & Computational Solid Mechanics
1. C L Dym, I H Shames. Solid Mechanics- A Variational Approach. McGraw Hill.
2. T R Tauchert. Energy Principles in Structural Mechanics . McGraw Hill.
3. M F Rubinstein. Structural Systems- Statics, Dynamics & Stability, Prentice Hall.
4. C Lanczos. The Variational Principles of Mechanics. Dover.
Linear Algebra
1. H Edwards, D E Penny. Elementary Linear Algebra. Prentice Hall.
2. G Strang. Linear Algebra. Thomson Brooks/Cole.
Finite Element Method
1. R D Cook, D S Malkus, M E Plesha. Concepts and Applications of Finite Element
Analysis. John Wiley & Sons.
2. T R Chandrapatla, A D Belegundu. Finite Elements in Engineering. Eastern Economy
Edition, Prentice Hall.
3. K H Huebner, DL Dewhirst, D E Smith, T G Byron. The Finite Element Method for
Engineers. John Wiley & Sons, Inc.
4. Peter Kattan. MATLAB Guide to Finite Elements. Springer.
5. George R Buchanan. Finite Element Analysis. Schaum Outlines, McGraw Hill.
6. C T F Ross. Finite Element Methods in Structural Mechanics. Ellis Horwood Limited.
7. E Hinton, D R J Owen. Finite Element Programming. Academic Press
8. G Strang, G J Fix. An analysis of the Finite element Method. Prentice Hall.
9. G Prathap. The Finite Element Method in Structural Mechanics. Kluwer Academic Press.
10. J N Reddy. An Introduction to the Finite Element Method. McGraw Hill.
11. T J R Hughes. The Finite Element Method. Dover.
12. K J Bathe. Finite Element Procedures. Prentice Hall.
13. O C Zienkiewicz, K Morgan. Finite Elements & Approximations. Dover
143
 REFERENCES: Journal papers/ reports published by our FEA team
1. S. Mukherjee and G. Prathap 2001 17 (6), pp 385-393. Communications in Numerical Methods in
Engineering. Analysis of Shear Locking in Timoshenko beam elements using a function space approach.
2. S. Mukherjee and G. Prathap 2002 Sadhana.27(5) 507-526. Analysis of delayed convergence in the three
noded isoparametric Timoshenko beam element using the function space approach.
3. G. Prathap and S. Mukherjee 2003 Current Science, 85(17), pp 989-994. The engineer grapples with
Theorem 1.1 and Lemma 6.3 of Strang and Fix.
4. H. Mishra and S. Mukherjee 2004 Sadhana 29(6), pp 573-588. Examining the best-fit paradigm in FEM at
element level.
5. S. Mukherjee, P. Jafarali and G. Prathap 2005 Journal of Sound and Vibration. Vol 285(3), pp 615-635. A
variational basis for error analysis in finite element elastodynamics.
6. K. Sangeeta , Somenath Mukherjee, and Gangan Prathap 2005 Structural Engineering and Mechanics
Vol 21(5), pp 539-551. A function space approach to study rank deficiency and spurious modes in finite
elements.
7. K. Sangeeta, Somenath Mukherjee and Gangan Prathap 2006 International Journal for Computational
Methods in Engineering Science & Mechanics, Vol 7, pages 1-12. Conservation of the Best-Fit Paradigm at
Element Level.
8. P. Jafarali, M Ammen, S. Mukherjee, G. Prathap, 2007 Journal of Sound and Vibration, Vol 299(2), pp 196-
211. Variational Correctness in Timoshenko beam finite element elastodynamics.
9. S. Mukherjee, P. Jafarali 2010. Communication in Numerical Methods in Engineering (also known as
International Journal of Numerical Methods in Biomedical Engineering). Vol 26; pp 1246-1262. Prathap’s best-fit
paradigm and optimal strain recovery points in indeterminate tapered bar analysis using linear element.
10. S. Mukherjee and S. Manju 2011 Structural Engineering and Mechanics, Vol 38 (3). An improved
parametric formulation for the variationally correct distortion immune three-noded bar element.
11. S Manju and S. Mukherjee 2019 Structural Engineering and Mechanics, Vol.69(6), pages 615- 626.
Function space formulation of the 3-noded distorted Timoshenko metric beam element.
144
12. S. Mukherjee and G. Prathap 1999 NAL Project Document ST 9931. Nonlinear Analysis of beams.
Thank you

145