2.

Fourier Series

Many phenomena that are studied in engineering are periodic in nature. For example the
current and voltage in an alternating current circuit, the displacement, velocity and acceleration
of the piston and many parameters in a vibrating system are all periodic. In order to solve such
problems it is very often desirable that these functions are represented in infinite trigonometric
series.
We would hope that such a modeling would give adequate representation over the whole cycle
of periodicity rather than the local nature of the Taylor series representation.
Many wave phenomena are also periodic and it is well known, especially through acoustics that a
wave can in general be decomposed or analyzed into several distinct waves of different
frequencies.

Definition

a0
A series of Sines and Cosines of the form  a1Cosx  a2Cos 2 x...  b1Sinx  b2 Sin 2 x  .......
2
a0  
=   anCosnx   bn Sinnx ,where a0 , a1 ,..., b1, b2 ,... are constants is called a Fourier
2 n 1 n 1
series.

We can show that a periodic function can be represented as a fourier series.

A periodic function f(x) whose period is 2  can be represented in the form

f(x) = c0  c1 Sin( x   )  c2 Sin(2 x   2 )  ..........................(1)

approximately provided that

(i) f(x) is single- valued
(ii) f(x) is not infinite (i.e. infinite series is convergent).

Note:
A function f(x) is said to be periodic with period p if for all x,
f(x+p) = f(x), p is a positive constant. The least value of p satisfying the above relationship
is called the period of f(x).

Second term of (1) can be written as

c1 Sin( x  1 )  c1CosxSin1  c1 SinxCos 1
= a1Cosx  b1 Sinx
( where a1  c1 Sin1 , b1  c1Cos1 )
Converting the remaining terms in a similar way f(x) becomes

12
 a0
where c 0 has been written as .(which would be shown later).
2

1
i.e. f(x) = a0   (a k Coskx  bk Sinkx)................................(2)
2 k 1
i.e. A periodic function can be represented as a fourier series.

Consider the function Asin(nx), where A and n are constants. This represents a wave where the
amplitude of the wave is A, and the period will be 2/n, where n is the wavenumber. If we
increase the wavenumber, then the function oscillated more rapidly, so sin(10x) has ten peaks
and troughs between x=0 and x=2 whereas sin(x) only has one peak and one trough.

So the terms of a Fourier series represent waves of different periods (or wavelengths), with each
successive term giving a more rapidly oscillating contribution than the previous.

The fundamental idea is that the large-scale features of any given periodic function can be
roughly approximated by a sine or cosine wave, with suitably chosen amplitude. The shorter-
scale features are more accurately captured by the subsequent terms in the Fourier series. In this
way we can recreate the function as accurately as we choose, by taking more and more terms,
thus including shorter and shorter scales.

Fourier Series is used in the analysis of signals in electronics.

We will see functions like the following, which approximates a saw-tooth signal:

Taking one extra term in the series each time and drawing separate graphs, we have:

f(t) = 1 (first term of the series):

13
f(t) = 1 + 2 sin t (first 2 terms of the series):

f(t) = 1 + 2 sin t - sin 2t (first 3 terms of the series):

14
We say that the infinite Fourier series converges to the saw tooth curve.

That is, if we take more and more terms, the graph will look more and more like a saw tooth. If
we could take an infinite number of terms, the graph would look like a set of saw teeth...

Note:
(i) Whether the series converges or not will depend on the value of x chosen and
the coefficients a k and b k .

(ii) If the series converges in any closed interval [c,c+2  ], the periodic nature of
the Cosine and Sine functions guarantees convergence for all values of x.
(iii) We try to approximate a given function in the fourier series. For that it is
necessary to know the values a 0 , a1 ,........, an and b1 , b2 ,......... which are
called the fourier coefficients.

15
Definition:
If f(x) is defined on an interval    x   with f(x)=f(x+2  ). We
define the fourier series of f(x) on [-  ,  ] as the series

a0 
  (a k Coskx  bk Sinkx) ------------ (3)
2 k 1

1
where a0 
  f ( x)dx


1
ak 
 
 f ( x)Coskxdx, k  1,2......

1
bk 
  f ( x)Sinkxdx, k  1,2....... and

fourier series of f (x) is denoted

by f s (x).

In otherwords f (x) = f s (x). approximately.

Proof is shown later.

That is, if the coefficients of the series (1) are determined in the case where the representation
of a given function in the form (1) is valid then the series with these coefficients is called
Fourier series of the function.

i.e. we are approximating a periodic function by an infinite trignometric series.

Integrals useful to calculate fourier coefficients:

 
(i)  Cosnxdx  0

and  Sinnxdx  0

 
1
(ii)  CosmxCosnxdx 

2 
{Cos(m  n) x  Cos(m  n) x}dx

1 
= 2 
 (Cos 2nx  1)dx   , m  n

 0, m  n


16
  
1
(iii)  SinmxSinnxdx 

2 
{Cos (m  n) x  Cos (m  n)}dx

 0, m  n
=
 , m  n
 
1
(iv)  SinmxCosnx 

2 
{Sin (m  n) x  Sin (m  n) x}dx  0

in all cases.
Same result holds for other limits provided they differ by 2  .

Proof of equation (3)

Consider,

1
f(x) = a0   (a k Coskx  bk Sinkx) …………………(1)
2 k 1
Integrating both sides from   to  ,
 
1 1
 f ( x)dx   2a0 dx  2 a0 .2

1
a0 
  f ( x)dx


To find a1 , multiply by Cos x and integrate,

 



f ( x)Cosxdx   a1Cos 2 xdx  a1


a1   f ( x)Cosxdx


To find a2 , a3 .........., an multiply (1) by Cos 2x, Cos 3x,…… and integrate between   to
 and the general term is,
 



f ( x)Cosnxdx   anCos 2 nxdx  an


1
an 
  f ( x)Cosnxdx


To find bn multiply by Sin nx and integrate,

17
 



f ( x) Sinnxdx   bn Sin 2 nxdx  bn


1
bn 
  f ( x)Sinnxdx


Note: a n and bn may be written

2
1
an 
  f ( x)Cosnxdx
0
2
1
bn 
  f ( x)Sinnxdx
0

Example:
Develop f(x) in a fourier series in the interval  , if

0 ,  x  0
f(x) = 
1 ,0  x  

a0 
Let f s ( x)    (a n Cosnx  bn Sinnx )
2 n1
 10 
Then, a 0 
1 
 f ( x)dx    0 dx   1dx  1
    0 

1
an 
  f ( x)Cosnxdx


1 0 1
=  0Cosnxdx   Cosnxdx
  0

1  Sinnx 
0
  n  0
=

 
1  1   Cosnx 1
bn   f ( x) Sinnxdx =  Sinnxdx  
  0   n  0


1
1  Cosn , n  1,2,....
n


1
n

1  (1) n 

18
 2 2
b1  , b2  0, b3 
, b4  0
 3
1 2 1 1 
 f s ( x)    Sinx  Sin3x  Sin5 x  .......
2  3 5 

1 1 2
Note: The partial sums of the series are s1  , s2   Sinx ,
2 2 

1 2 1
s3   ( Sinx  sin 3x)
2  3

The graphs of f(x) and partial sums are shown below.

It is evident from the graphs that the accuracy with which S n represents f (x) increases with n.

Exercise 16
Find the fourier series for

(1) f(x) = x+2,   x  . .

x
(2) f(x) = Cos ,  x  
2

19
Odd and even functions

Definition: Function f (x) is odd iff f ( x)   f ( x) and

Function f (x) is even iff f ( x)  f ( x) .

Note: (i) Graph of an odd function is symmetrical about the origin whereas the
graph of an even function is symmetrical about the y-axis.

(ii) (a) A product of two even or odd functions is even.

(b)A product of an even and odd functions is odd.
( c) For an even function f (x)
 



f ( x)dx  2 f ( x)dx
0

(d) For an odd function f (x)



 f ( x)dx  0


a0 
(e) Fourier series of an even function is,   an cos nx .
2 n 1


(f)Fourier series of an odd function is  b sin nx
n 1
n

Examples: Find fourier series of the following functions.

 
 1 ,    t   2
 1 ,  t  0   
(1) f(t) =  (2) f(t) =  1 ,   t 
 1 ,0  t    2 2
 1 , 
t 
 2

20
Dirichlet’s theorem :
If f(x) is single valued and bounded periodic function which in any one period has a finite
number of discontinuities and a finite number of maxima and minima , then

p
a0
S p ( x)    (an cos nx  bn sin nx) converges to f(x) as p   at values of x for
2 n 1
which f(x) is continuous and to
1
 f ( x  0)  f ( x  0) (i.e.
2
the average of the right – hand and left- hand limits of f(x) ) at points of discontinuity.

Exercise 17
(i) Given that f ( x)  x  x 2 for    x   with f ( x  2 )  f ( x). Show that
 2
1 1
1   ........
6 22 32

  ,  x  0
(ii) Find fourier series for f (x) , if f ( x)   .
 x ,0  x  
1 1 1 2
Deduce that 2  2  2  ........  . Graph the fourier series.
1 3 5 8

Half range series

We can often make use of the oddness or evenness of a function giving rise to Sine series or a
Cosine series respectively. In the solution of some partial differential equations, the boundary
conditions may restrict us to a series which contains only Sine terms. We shall therefore need to
investigate how to manufacture an odd function or an even function, given a function which
may be neither.

Suppose f(x) is defined on [0,  ] , and we want a Cosine series. We extend f(x) to a new
function g(x) defined on [-  ,  ] as follows

 f ( x) ,0  x  
g(x) = 
 f ( x) ,   x  0

a0 
Then g(x) is even in [-  ,  ] . Hence g s (x) =   a n Cosnx
2 n 0

21
 
2
Where a0 
  g ( x)dx
0

2
an 
  g ( x)Cosnxdx
0

As g(x) = f(x) in (0,  ), therefore in this interval

a0 
f s (x) =   a n Cosnx …………..(1)
2 n 1


2
a0 
  f ( x)dx
0

2
an 
  f ( x)Cosnxdx
0

This is called fourier half-range cosine series for x  (0,  ) . It is denoted by

f cs (x).

2
Therefore, f s ( x)  f cs ( x) for x  (0,  ). Where, a0 
  f ( x)dx ,
0

2
an 
  f ( x) cos nxdx .
0

g (x) is called even periodic extension of f (x). Thus the series converges to f (x) at all points
on (0,  ) where f (x) is continuous and at all points of discontinuity the series converges to
the average of its left and right hand limits.

Suppose we want a Sine series. We can extend f(x) to an odd function h(x) on [-  ,  ] .

 f ( x) 0 x 
h(x) = 
 f ( x) ,  x  0

hs ( x)   bn Sinnx
n 1


2
Here b n =
  g ( x)Sinnxdx
0

Since h(x) = f(x) in (0,  ), therefore in this interval


f s (x) =  b Sinnx
n 1
n ……………(2)

22
 
2
where b n =
  f ( x)Sinnxdx
0

This is called fourier half-range sine series and it is denoted by f ss (x).

i.e. f s ( x)  f ss ( x). h(x) is called odd periodic extension of f (x).
Thus the series converges to f (x) at all points on (0,  ) where f (x) is continuous and at all
points of discontinuity the series converges to the average of its left and right hand limits.

Thus a function f(x) , whether even or odd or neither, defined over the interval (0,  ) is
capable of these two distinct expansions.
Series (1) and (2) with coefficients defined above are called Fourier half range cosine series
and sine series of f(x) respectively.

Exercise 18
1. Expand f ( x)  sin x in a fourier cosine series in 0  x   . Graph the fourier series of
the even periodic extension of f (x).

2. Represent the following by a fourier sine series in the given region and obtain the
graph of the fourier series of the odd periodic extension of the function.

 
t ,0  t 
f ( x)   2
 
 , t 
2 2

2.5 Fourier series of general period

So far only functions with period 2  have been considered. In practice it is often necessary to
find a fourier series of f (x) defined over the interval -L to L or 0 to 2L.

The fourier series of f (x) defined on (-L,L) with period 2L is,

a0  nx nx
f s ( x)    (a n Cos  bn Sin )
2 n1 L L

nx nx
L L L
1 1 1
Where a0 
L L
f ( x)dx , an  
L L
f ( x) Cos
L
dx , bn   f ( x) Cos
L L L
dx

23
 The fourier series of f (x) defined on (0,2L) with period 2L is,

a0  nx nx
f s ( x)    (a n Cos  bn Sin )
2 n1 L L

nx nx
2L 2L 2L
1 1 1
Where a0 
L 
0
f ( x)dx , an 
L 
0
f ( x) Cos
L
dx , bn 
L  f ( x) Cos
0
L
dx

Exercise 19

1. Find the fourier series expansion of the following periodic function of period 1

1 1
 2  x , 2  x  0
f ( x)   .
1 1
  x ,0  x 
2 2


2. Expand f ( x)  Cosx, 0  x  in a fourier Sine series. Graph the fourier expansion
2
of odd periodic extension of f (x).

x , x  1
3. Find a fourier Cosine series for f ( x)   on (o, 2).Graph the fourier series of the
2 , x  1
even extension of the function function.

24
25