Theory of Vibration

Single Degree of Freedom (SDOF)

System (7)
Periodic Loading
Lecture 8
M.Sc. in Earthquake Engineering
Thapathali Campus, Institute of Engineering
2020
 Fourier Analysis
Discrete Continuous

Fourier Fourier
Series Integral

Fast Discrete
Fourier
Fourier Fourier
Transform
Transform Transform

Page 2
 Outline
• Periodic Function
• Fourier Cosine and Sine Series
• Periodic Function with Period 2L
• Odd and Even Functions
• Half Range Fourier Cosine and Sine Series
• Complex Notation for Fourier Series

Page 3
 Periodic Function
• Definition: Periodic Function
A function f(x) is said to be periodic with
period T if for all x

f ( x  T )  f ( x)
f(x)

T
Page 4
 Periodic Function

• f(x+p)=f(x), f(x+np)=f(x)
• If f(x) and g(x) have period p, then the
function H(x)=af(x)+bg(x) , also has
the period p
• If a period function of f(x) has a
smallest period p (p >0), this is often
called the fundamental period of f(x)

Page 5
 Periodic Function
• Example
– Cosine Functions: cosx, cos2x, cos3x, …
– Sine Functions: sinx, sin2x, sin3x, …
– eix, ei2x, ei3x, …
– e-ix, e-i2x, e-i3x, …

Page 6
 Fourier Cosine and Sine Series
• A function f(x) is periodic with period 2
and

f ( x )  a0  a1 cos x  a2 cos 2 x  a3 cos 3x  

 b1 sin x  b2 sin 2 x  b3 sin 3x  

 a0   an cos nx  bn sin nx 
n 1

Page 7
Fourier Cosine and Sine Series
(Euler formulas)
• Then
1 
a0 
2 

f ( x )dx

1 
an 
 

f ( x ) cos nxdx

1 
bn 
 

f ( x ) sin nxdx

Page 8
 Periodic Function with Period 2L
• A periodic function f(x) with period 2L


 nx nx 
f ( x )  a0    an cos  bn sin 
n 1  L L 

f(x)

2L

Page 9
 Periodic Function with Period 2L
• Then
1 L
a0 
2L L
f ( x ) dx

1 L nx
an 
L 
L
f ( x ) cos
L
dx

1 L nx
bn 
L 
L
f ( x ) sin
L
dx

Page 10
 Odd and Even Functions
• A function f(x) is said to be even if

f ( x)  f ( x)

• A function f(x) is said to be odd if

f ( x)   f ( x)

Page 11
 Odd and Even Functions
Even Function Odd Function
f ( x)  f ( x) f ( x)   f ( x)

f(x) f(x)

x x

Page 12
 Odd and Even Functions
• Property
L L
L
f ( x )dx 2  f ( x )dx, if f ( x ) is even
0

L
 L
f ( x )dx 0, if f ( x ) is odd

• The product of an even and an odd

function is odd.
Page 13
 Odd and Even Functions
• Fourier Cosine Series

nx
f ( x )  a0   an cos , if f ( x ) is even
n 1 L

• Fourier Sine Series


nx
f ( x )   bn sin , if f ( x ) is odd
n 1 L

Page 14
 Sum of Functions
• The Fourier coefficients of a sum f1+f2 are
the sum of the corresponding Fourier
coefficients of f1 and f2.

• The Fourier coefficients of cf are c times

the corresponding Fourier coefficients of f.

Page 15
Complex Notation for Fourier Series

f ( x )  a0   a
n 1
n cos nx  bn sin nx 


f ( x)   n
c e
n  
inx

1 

 inx
cn  f ( x )e dx
2 

Page 16
Complex Notation for Fourier Series
(Loading)
  2n 2n 
p(t )  a0    an cos t  bn sin t

n 1  Tp Tp 


p(t )  c n exp in 1t  (a)
n  

p(t ) exp in 1t dt

1 Tp
cn 
Tp 0
(b)

Page 17
 Arbitrary Non-periodic Loading
• Coefficients Cn obtained from the
equation for the interval 0 < t < Tp
would actually define the periodic
function shown in the figure by P(t)

the solid lines as well as the

dashed lines.
• The spurious loadings could be
eliminated by extending the
loading period to infinity.
0
• Rewrite the earlier two equations -Tp -Tp/2 Tp/2 Tp Tp/2
t

in slightly modified form using the

following notation:
1  1  Arbitrary loading represented by Fourier series
 
Tp 2 2

cn  c n   c n 
1
n 1  n   n Tp 2
Page 18
 Fourier Integral
With the new symbols, the Fourier series equations
become

 
pt    c expi t  (c)
2
n n
n  

t T p / 2

c n   Tp cn   pt exp i t dt n (d)

t  T p / 2

Where advantage has been taken of the fact that the limits of the
integral are arbitrary so long that they span one complete loading
period Page 19
 Fourier Integral and Fourier
Transform
Now if the loading period is extended to infinity (Tp →∞), the frequency
increment becomes an infinitesimal (Δϖ → dϖ) and the discrete
frequencies ϖn become a continuous function ϖ. Thus in the limit,
the Fourier series expression of Equation (c) becomes Fourier integral:

1 
pt    c n  expit d (e)
2  
In which the harmonic-amplitude function is given by


c    pt exp it dt (f)

t  
Page 20
 Fourier Integral and Fourier
Transform
• The two integrals of Equations (e) and (d)
are known as Fourier transform pair
• The time function can be derived from the
frequency function or vice versa by
equivalent process.
• A necessary condition for the existence of
the Fourier transform is that the following
integral must be finite


 pt dt
 Page 21
 Numerical Analysis
• To make practical use of the Fourier analysis
(Frequency domain approach), it is necessary to
formulate it in terms of numerical-analysis
procedure.
• Two phases of numerical formulation:
1. Derivation of Discrete Fourier Transform
(DFT) expressions which correspond to the
Fourier integrals or Fourier transfer pair, i.e.,
Equations (e) and (d)
2. Development of an efficient numerical
technique for evaluating the DFTs.
Page 22
Discrete Fourier Transform (DFT)
• Assume that the loading is periodic of period Tp. This constitutes an
approximation in the treatment of an arbitrary general loading but is
necessary in order to replace the infinite time integral of Equation (d)
by a finite sum. It also serves to define the lowest frequency that is
considered in the analysis.
ϖ1 = Δ ϖ = 2π/Tp
• The load period is divided into N equal time increments Δt, and the
load is defined for the discrete times tm = m Δt. Then, the exponential
term in Equation (c) becomes

 nm 
exp i ntm   exp inmt   exp 2i 
 N 
Accordingly Equation (c) takes the discrete form

 N 1
 nm 
pt m    c   exp  2i 
2
n
n 0  N 
Page 23
Discrete Fourier Transform (DFT)
• The corresponding discrete
expression for the amplitude function
c(ϖn) can be obtained by merely
substituting the sum of a finite series
of discrete terms for the integral of
Equation (d) to obtain the result:
N 1
 nm 
c n   t  pt m  exp  2i 
m 0  N 

Page 24
Complex Notation for Fourier Series
• A periodic function f(x) with period 2L

 inx
f ( x)  c e
n  
n
L

inx
1 L 
cn  
2L L
f ( x )e L
dx

Page 25
Fourier Cosine and Sine Integrals

• Example 1  0,  L  x  1

f ( x )  1, 1  x  1
0, 1 x  L

f ( x  2 L)  f ( x )

Page 26
 Fourier Cosine and Sine Integrals

1 1 1
a0  
2 L 1
f ( x )dx 
L
1 1 nx 2 sin( n / L) 2 sin  n
an   f ( x ) cos dx  
L 1 L L n / L L n

1
x
sin(x)

x Page 27
 Fourier Cosine and Sine Integrals

sin x
x

x
Page 28
Fourier Cosine and Sine Integrals

Page 29
Fourier Cosine and Sine Integrals

Page 30
Fourier Cosine and Sine Integrals

L

2 sin 
A( ) 
 

Page 31
 Gibb’s Phenomenon
sin 

u
• Sine Integral Sin( u)  d
0 

Page 32
 Gibb’s Phenomenon
• Gibb’s Phenomenon

Page 33
 Gibb’s Phenomenon
2 a cos x sin 
f ( x) 
 0 
d

1 a sin(  x )  sin(  x ) 

 0 
d

1 a sin(  x ) 1 a sin(  x )

 0 
d  
 0 
d

1 ( x 1) a sin t 1 ( x 1) a sin t


 0 t
dt 
 
0 t
dt


1
Si(a ( x  1))  Si (a ( x  1))

Page 34
Fourier Cosine and Sine Integrals

• Fourier Cosine Integral of f(t)

2 
A( ) 
 
0
f (t ) cos(t )dt


f (t )   A( ) cos(t ) d
0

Page 35
 Fourier Cosine and Sine Integrals

• Fourier Sine Integral of f(t)

2 
B( ) 
 
0
f (t ) sin(t )dt


f (t )   B( ) sin(t ) d
0

Page 36
 Fourier Integrals

• Fourier Integral of f(t)

f (t )   A( ) cos(t )  B( ) sin(t ) d



1 
A( ) 
 

f (t ) cos(t )dt

1 
B( ) 
 

f (t ) sin(t )dt

Page 37