PMAT 32322 Mathematical Methods [Type here]

Fourier Series
Introduction
In many scientific problems, specially in the study of periodic phenomena in conduction of heat,
optics, acoustics and thermodynamics it is necessary to expend a function 𝑓(𝑥) in a
trigonometric series of sines and cosines in form a series, known as Fourier series.
Some Important Results and Definitions
Definition 01: Periodic Function
A function 𝑓: ℝ → ℝ is said to be periodic if there exists a positive number 𝑇 such that
𝑓(𝑥 + 𝑇) = 𝑓(𝑥) for all real numbers 𝑥 and 𝑇 is called a period of 𝑓(𝑥). If a periodic function
has a smallest positive period 𝑇, then 𝑇 is called the primitive period or simply the period.
sin 𝑥 and cos 𝑥 are periodic functions , each having period 2𝜋 . tan 𝑥 and cot 𝑥 are periodic
functions, each having the period 𝜋.
Definition 02: Even and Odd Functions
A function 𝑓(𝑥) is known as an even function if 𝑓(−𝑥) = 𝑓(𝑥).
For example, 𝑥 4 + 𝑥 2 + 1 , cos 𝑥, |sin 𝑥|, |𝑥| and sin2 𝑥 are all even functions. The graph of
even function is symmetrical about the y-axis.
A function 𝑓(𝑥) is known as an odd function if 𝑓(−𝑥) = −𝑓(𝑥).
For example, 𝑥 3 + 𝑥 , sin 𝑥, and tan 𝑥 are all odd functions. The graph of odd function is
symmetrical about the origin.
We shall use the following results:
𝑙
∫−𝑙 𝑓(𝑥)𝑑𝑥 = 0 , when 𝑓(𝑥) is an odd function.
𝑙 𝑙
∫−𝑙 𝑓(𝑥)𝑑𝑥 = 2 ∫0 𝑓(𝑥)𝑑𝑥 , when 𝑓(𝑥) is an even function.

Some Useful Results:

If 𝑛 is any integer,
𝜋 𝜋
cos 𝑛𝜋 = (−1)𝑛 , sin 𝑛𝜋 = 0 cos(2𝑛 + 1) 2 = 0 sin(2𝑛 + 1) 2 = (−1)𝑛 .

Euler’s Formulae for Fourier Series Expansion of a Function 𝒇(𝒙)

Theorem
Let 𝑓(𝑥) be a periodic function with period 2𝜋. Suppose 𝑓(𝑥) can be represented as a
trigonometric series in the interval 𝑐 < 𝑥 < 𝑐 + 2𝜋, as follows

1
PMAT 32322 Mathematical Methods [Type here]

𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑥 ____(1)
2
1 𝑐+2𝜋 1 𝑐+2𝜋 1 𝑐+2𝜋
Then, 𝑎0 = 𝜋 ∫𝑐 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = 𝜋 ∫𝑐 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 and 𝑏𝑛 = 𝜋 ∫𝑐 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥

where 𝑛 = 1,2,3, ….___(2)

The formulae for 𝑎0 , 𝑎𝑛 and 𝑏𝑛 given by (2) are called Euler’s formulae.
The series (1) in which 𝑎0 , 𝑎𝑛 and 𝑏𝑛 are given by the Euler’s formulae (2) is called the Fourier
series of 𝑓(𝑥). Also the coefficients 𝑎0 , 𝑎𝑛 and 𝑏𝑛 are known as Fourier coefficients.
Note:
In what follows we state the Direchlet’s conditions for the existence of Fourier series expansion
for a periodic function 𝑓(𝑥) in 𝑐 < 𝑥 < 𝑐 + 2𝜋.
Theorem
The function 𝑓(𝑥) of period 2𝜋 defined in (𝑐, 𝑐 + 2𝜋) can be expanded in Fourier series in the
form
𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑥
2

where 𝑎0 , 𝑎𝑛 and 𝑏𝑛 (𝑛 = 1,2,3, … ) are constants given by the Euler’s formulae, provided
(i) 𝑓(𝑥) is periodic with period 2𝜋, single valued and bounded in (𝑐, 𝑐 + 2𝜋).
(ii) 𝑓(𝑥) has at most a finite number if maxima and minima in (𝑐, 𝑐 + 2𝜋).
(iii) 𝑓(𝑥) is piecewise continuous with finite number of discontinuities in (𝑐, 𝑐 + 2𝜋).
Remark:
The condition (i), (ii) and (iii) imposed on 𝑓(𝑥) are sufficient but not necessary, that is, if the
conditions are satisfied the convergence of Fourier series is guaranteed. However, if they are not
satisfied the Fourier series may be or may not converge. The above three conditions are
generally satisfied in problems which are in science and engineering.
Convergence of a Fourier Series
Let 𝑓(𝑥) be a periodic function of period 2𝜋 defined in (𝑐, 𝑐 + 2𝜋) and satisfy Direchlet’s
conditions. Then Fourier series is given by
𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑥 where 𝑎0 , 𝑎𝑛 and 𝑏𝑛 (𝑛 = 1,2,3, … ) are
2
Fourier coefficients converges to
(i) 𝑓(𝑥), if x is a point of continuity.
𝑓(𝑥+0)−𝑓(𝑥−0)
(ii) , if x is a point of discontinuity
2

2
PMAT 32322 Mathematical Methods [Type here]

𝑓(𝑐+0)+𝑓(𝑐+2𝜋−0)
(iii) , if at 𝑥 = 𝑐 or 𝑥 = 𝑐 + 2𝜋 which are end points of interval
2

(𝑐, 𝑐 + 2𝜋).
𝐍𝐨𝐭𝐞:
In the above results (ii) and (iii), we have used the notion of left hand limit and right hand limit
which are denoted and defined as
Left hand limit= 𝑓(𝑎 −) = 𝑓(𝑎 − 0) = lim 𝑓(𝑎 − ℎ).
ℎ→0

Right hand limit= 𝑓(𝑎 +) = 𝑓(𝑎 + 0) = lim 𝑓(𝑎 + ℎ).

ℎ→0

Some Particular Cases of Above Theorem for Direct Applications

(i) If 𝑐 = −𝜋, the interval becomes (−𝜋, 𝜋) and the formulae (1) and (2) reduce to
𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑥 where
2
1 𝜋 1 𝜋 1 𝜋
𝑎0 = 𝜋 ∫−𝜋 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = 𝜋 ∫−𝜋 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 and 𝑏𝑛 = 𝜋 ∫−𝜋 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 .

The above series is further simplified if 𝑓(𝑥) is of following two special forms:
Special Form 01:
When 𝑓(𝑥) is an odd function, then 𝑎0 = 0.
Since cos 𝑛𝑥 is an even function, 𝑓(𝑥) cos 𝑛𝑥 is an odd function and so 𝑎𝑛 = 0. Also , sin 𝑛𝑥 is
an odd function, so 𝑓(𝑥) sin 𝑛𝑥 is an even function. Then, 𝑏𝑛 can be re-written as
1 𝜋 2 𝜋
𝑏𝑛 = 𝜋 ∫−𝜋 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 = 𝜋 ∫0 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥.

Let 𝑓(𝑥) be defined in (−𝜋, 𝜋) be a periodic function with period 2𝜋. Now if 𝑓(𝑥) is an odd
function, then its Fourier series contains only sine terms. Thus, we have
2 𝜋
𝑓(𝑥) = ∑∞
𝑛=1 𝑏𝑛 sin 𝑛𝑥 where 𝑏𝑛 = 𝜋 ∫0 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 where 𝑛 = 1,2,3, ….

Special Form 02:

1 𝜋 2 𝜋
When 𝑓(𝑥) is an even function, then 𝑎0 = 𝜋 ∫−𝜋 𝑓(𝑥)𝑑𝑥 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥.

Since cos 𝑛𝑥 is an even function, 𝑓(𝑥) cos 𝑛𝑥 is an even function and so

1 𝜋 2 𝜋
𝑎𝑛 = 𝜋 ∫−𝜋 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥.

Also , sin 𝑛𝑥 is an odd function, so 𝑓(𝑥) sin 𝑛𝑥 is an odd function. Then, 𝑏𝑛 = 0 .

Let 𝑓(𝑥) be defined in (−𝜋, 𝜋) be a periodic function with period 2𝜋. Now if 𝑓(𝑥) is an even
function, then its Fourier series contains only cosine terms. Thus, we have

3
PMAT 32322 Mathematical Methods [Type here]

𝑎0 2 𝜋 2 𝜋
𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 where 𝑎0 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥 and 𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 where
2
𝑛 = 1,2,3, …..
(ii) If 𝑐 = 0, the interval becomes (0,2𝜋) and the formulae (1) and (2) reduce to
𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑥 where
2
1 2𝜋 1 2𝜋 1 2𝜋
𝑎0 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 and 𝑏𝑛 = 𝜋 ∫0 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥.

The above series is further simplified if 𝑓(𝑥) is of following two special forms:
Special Form 01:
When 𝑓(𝑥) satisfies the condition 𝑓(2𝜋 − 𝑥) = 𝑓(𝑥).Then
1 2𝜋 2 𝜋
𝑎0 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥 .

Now, let 𝐹(𝑥) = 𝑓(𝑥) cos 𝑛𝑥. Then we have

𝐹(2𝜋 − 𝑥) = 𝑓(2𝜋 − 𝑥) cos 𝑛(2𝜋 − 𝑥) = 𝑓(𝑥) cos 𝑛𝑥 = 𝐹(𝑥) and so
1 2𝜋 1 2𝜋 2 𝜋 2 𝜋
𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 = 𝜋 ∫0 𝐹(𝑥)𝑑𝑥 = 𝜋 ∫0 𝐹(𝑥)𝑑𝑥 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 .

Again, let 𝐺(𝑥) = 𝑓(𝑥) sin 𝑛𝑥. Then, we have

𝐺(2𝜋 − 𝑥) = 𝑓(2𝜋 − 𝑥) sin 𝑛(2𝜋 − 𝑥) = −𝑓(𝑥) sin 𝑛𝑥 = −𝐺(𝑥).
1 2𝜋 1 2𝜋
𝑏𝑛 = ∫ 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 = ∫ 𝐺(𝑥)𝑑𝑥
𝜋 0 𝜋 0

Let 𝑓(𝑥) be defined in (0,2𝜋) be a periodic function with period 2𝜋. Now if 𝑓(2𝜋 − 𝑥) = 𝑓(𝑥).,
then its Fourier series contains only cosine terms. Thus, we have
𝑎0 2 𝜋
𝑓(𝑥) = 𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 where 𝑎0 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥 and
2
1 2𝜋 2 𝜋
𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 where 𝑛 = 1,2,3, ……

Special Form 02:

When 𝑓(𝑥) satisfies the condition 𝑓(2𝜋 − 𝑥) = −𝑓(𝑥).Then 𝑎0 = 0 .
Now, let 𝐹(𝑥) = 𝑓(𝑥) cos 𝑛𝑥. Then we have
𝐹(2𝜋 − 𝑥) = 𝑓(2𝜋 − 𝑥) cos 𝑛(2𝜋 − 𝑥) = −𝑓(𝑥) cos 𝑛𝑥 = −𝐹(𝑥) and so
1 2𝜋 1 2𝜋
𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 = 𝜋 ∫0 𝐹(𝑥)𝑑𝑥 = 0 .

Again, let 𝐺(𝑥) = 𝑓(𝑥) sin 𝑛𝑥. Then, we have

𝐺(2𝜋 − 𝑥) = 𝑓(2𝜋 − 𝑥) sin 𝑛(2𝜋 − 𝑥) = 𝑓(𝑥) sin 𝑛𝑥 = 𝐺(𝑥) and so

4
PMAT 32322 Mathematical Methods [Type here]

1 2𝜋 1 2𝜋 2 𝜋 2 𝜋
𝑏𝑛 = ∫ 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 = ∫ 𝐺(𝑥)𝑑𝑥 = ∫ 𝐺(𝑥)𝑑𝑥 = ∫ 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥
𝜋 0 𝜋 0 𝜋 0 𝜋 0

Let 𝑓(𝑥) be defined in (0,2𝜋) be a periodic function with period 2𝜋. Now if 𝑓(2𝜋 − 𝑥) =
−𝑓(𝑥)., then its Fourier series contains only sine terms. Thus, we have
2 𝜋
𝑓(𝑥) = 𝑓(𝑥) = ∑∞
𝑛=1 𝑏𝑛 sin 𝑛𝑥 where 𝑏𝑛 = 𝜋 ∫0 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 .

Example 01:
Find the series of sines and cosines of multiples of x for the function 𝑓(𝑥) = |sin 𝑥| in the
interval [−𝜋, 𝜋].
Solution:
(a) Since 𝑓(𝑥) = |sin 𝑥| is an even function, the required Fourier series is given by
𝑎0
𝑓(𝑥) = |sin 𝑥| = + ∑∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 ___(1)
2
2 𝜋 2 𝜋 2 𝜋 2 2 4
𝑎0 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥 = 𝜋 ∫0 |sin 𝑥|𝑑𝑥 = 𝜋 ∫0 sin 𝑥 𝑑𝑥 = 𝜋 [− cos 𝑥]𝜋0 = 𝜋 (1 + 1) = 𝜋.
2 𝜋 1 𝜋
𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 = 𝜋 ∫0 2 sin 𝑥 cos 𝑛𝑥 𝑑𝑥
1 𝜋
= 𝜋 ∫0 {sin(𝑛 + 1)𝑥 − sin(𝑛 − 1)𝑥}𝑑𝑥

1 cos(𝑛+1)𝑥 cos(𝑛−1)𝑥 𝜋
= 𝜋 [− + ]
𝑛+1 𝑛−1 0

1 cos(𝑛+1)𝜋 cos(𝑛−1)𝜋 1 1
= 𝜋 {− + + 𝑛+1 − 𝑛−1}
𝑛+1 𝑛−1

1 (−1)𝑛+1 (−1)𝑛−1 1 1
= 𝜋 {− + + 𝑛+1 − 𝑛−1}
𝑛+1 𝑛−1

1 (−1)𝑛 (−1)𝑛 1 1
= 𝜋 {− + + 𝑛+1 − 𝑛−1}
𝑛+1 𝑛−1

2{1+(−1)𝑛 }
= .
𝜋(𝑛2 −1)

0 𝑛 𝑖𝑠 𝑜𝑑𝑑
For 𝑛 > 1, 𝑎𝑛 = { −4
𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛 .
𝜋(𝑛2 −1)

Therefore we have
2 4 cos 2𝑛𝑥
𝑓(𝑥) = |sin 𝑥| = 𝜋 − 𝜋 ∑∞
𝑛=1 (2𝑛−1)(2𝑛+1) .

Example 02:
(𝜋−𝑥)
Express 𝑓(𝑥) = in a Fourier series in 0 < 𝑥 < 2𝜋.
2

5
PMAT 32322 Mathematical Methods [Type here]

Solution:
𝜋−(2𝜋−𝑥) −(𝜋−𝑥)
Here 𝑓(2𝜋 − 𝑥) = = = −𝑓(𝑥). Hence the required Fourier series is given by
2 2
𝑥−𝜋 2 𝜋
𝑓(𝑥) = = ∑∞
𝑛=1 𝑏𝑛 sin 𝑛𝑥 , where 𝑏𝑛 = 𝜋 ∫0 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥.
2

2 1 𝜋 1 − cos 𝑛𝑥 − sin 𝑛𝑥 𝜋
𝑏𝑛 = (𝜋) (2) ∫0 (𝜋 − 𝑥) sin 𝑛𝑥 𝑑𝑥 = 𝜋 [(𝜋 − 𝑥) ( ) − (−1) ( )]
𝑛 𝑛2 0

1 𝜋 1
= (𝜋 ) (𝑛 ) = 𝑛.
𝑥−𝜋 sin 𝑛𝑥
Therefore, we have 𝑓(𝑥) = = ∑∞
𝑛=1 .
2 𝑛

Example 03:
−𝜋 −𝜋 < 𝑥 < 0
Express Fourier series expansion of 𝑓(𝑥) defined by 𝑓(𝑥) = { .
𝑥 0<𝑥<𝜋
1 1 1 𝜋2
Deduce that + 22 + 32 + ⋯ = .
12 8

Solution:
The required Fourier series expansion of the given function 𝑓(𝑥) is given by
𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑥 ___(1)
2

1 𝜋 1 0 𝜋
𝑎0 = 𝜋 ∫−𝜋 𝑓(𝑥)𝑑𝑥 = 𝜋 {∫−𝜋(−𝜋)𝑑𝑥 + ∫0 𝑥𝑑𝑥 }
𝜋
1 𝑥2 1 𝜋2 1 −𝜋 2 −𝜋
= 𝜋 {[−𝜋𝑥]0−𝜋 + [ 2 ] } = 𝜋 {−𝜋 2 + }= ( )= .
0 2 𝜋 2 2

1 𝜋 1 0 𝜋
𝑎𝑛 = 𝜋 ∫−𝜋 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 = 𝜋 [∫−𝜋(−𝜋) cos 𝑛𝑥 𝑑𝑥 + ∫0 𝑥 cos 𝑛𝑥 𝑑𝑥 ]

1 cos 𝑛𝑥 0 sin 𝑛𝑥 − cos 𝑛𝑥 𝜋

= 𝜋 {(−𝜋) [ ] + [(𝑥) ( ) − (1) ( )] }(integration by parts in the second
𝑛 −𝜋 𝑛 𝑛2 0
integral)

1 1 1 1 0 𝑛 is even
= 𝜋 {(𝑛2 ) cos 𝑛𝜋 − 𝑛2 } = − (𝑛2 𝜋) (1 − cos 𝑛𝜋) = { −2 𝑛 is odd .
𝑛2 𝜋

1 𝜋 1 0 𝜋
And 𝑏𝑛 = 𝜋 ∫−𝜋 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 = 𝜋 [∫−𝜋(−𝜋) sin 𝑛𝑥 𝑑𝑥 + ∫0 𝑥 sin 𝑛𝑥 𝑑𝑥 ]

1 cos 𝑛𝑥 0 − cos 𝑛𝑥 − sin 𝑛𝑥 𝜋

= 𝜋 {(−𝜋) [ ] + [(𝑥) ( ) − (1) ( )] }(integration by parts in the
𝑛 −𝜋 𝑛 𝑛2 0
second integral)

6
PMAT 32322 Mathematical Methods [Type here]

1 𝜋 𝜋 {1−2(−1)𝑛 }
= 𝜋 {(𝑛) (1 − cos 𝑛𝜋) − 𝑛 cos 𝑛𝜋} = .
𝑛

In [−𝜋, 𝜋], the point 𝑥 = 0 is the only point of discontinuity of the function 𝑓. Therefore, x is
different from 0, (1) reduces to
𝜋 −2 {1−2(−1)𝑛 }
𝑓(𝑥) = − 4 + ∑∞ ∞
𝑛=1 ((2𝑛−1)2 𝜋 ) cos(2𝑛 − 1)𝑥 + ∑𝑛=1 { } sin 𝑛𝑥 or
𝑛

𝜋 2 cos 3𝑥 cos 5𝑥 sin 2𝑥 3 sin 3𝑥 sin 4𝑥

𝑓(𝑥) = − 4 − 𝜋 (cos 𝑥 + + + ⋯ ) + 3 sin 𝑥 − − − + ⋯.
32 52 2 3 4

Functions Having Arbitrary Periods (Change of Interval)

So far we have found Fourier series of functions having period 2𝜋. In many practical problems,
the functions may have arbitrary periods (not necessarily 2𝜋) . With help of a transformation of
the variable , we propose to get Euler’s formulae for Fourier coefficients for functions having
period 2𝑙, where 𝑙 is any positive number.
Let 𝑓(𝑥) be given periodic function in the interval 𝑐 < 𝑥 < 𝑐 + 2𝑙. To change the interval into
one of length 2𝜋. We put
𝜋 𝑧 𝜋𝑥
=𝜋 or 𝑧= ____(1)
𝑙 𝑙
𝜋𝑐
When 𝑥 = 𝑐, 𝑧 = = 𝑑 (say)___(2)
𝑙

𝜋(𝑐+2𝑙) 𝜋𝑐
and when 𝑥 = 𝑐 + 2𝑙, 𝑧 = = + 2𝜋 = 𝑑 + 2𝜋 ___(3)
𝑙 𝑙

𝑙𝑧
Hence the function 𝑓(𝑥) is of period 2𝑙 is transformed to the function 𝑓 ( 𝜋 ) = 𝐹(𝑧), say, of
period 2𝜋 in (𝑑, 𝑑 + 2𝜋). Then 𝐹(𝑧) can be expressed as Fourier series given by
𝑎0
𝐹(𝑧) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝑧 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝑧 where ___(4)
2
1 𝑑+2𝜋 1 𝑑+2𝜋 1 𝑑+2𝜋
𝑎0 = 𝜋 ∫𝑑 𝐹(𝑧)𝑑𝑧 , 𝑎𝑛 = 𝜋 ∫𝑑 𝐹(𝑧) cos 𝑛𝑧 𝑑𝑧 and 𝑏𝑛 = 𝜋 ∫𝑑 𝐹(𝑧) sin 𝑛𝑧 𝑑𝑧___(5)
𝜋𝑥 𝜋𝑑𝑥
Now putting 𝑧 = and 𝑑𝑧 = and noting that when 𝑧 = 𝑑, 𝑥 = 𝑐 and when 𝑧 = 𝑑 +
𝑙 𝑙
2𝜋, 𝑥 = 𝑐 + 2𝑙, (4) and (5) reduce to
𝜋𝑥 𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝐹( 𝑙 )= + ∑∞
𝑛=1 𝑎𝑛 cos ( ) + ∑∞
𝑛=1 𝑏𝑛 sin ( ) ,___(6)
2 𝑙 𝑙
1 𝑐+2𝜋 1 𝑐+2𝜋 𝑛𝜋𝑥 1 𝑐+2𝜋 𝑛𝜋𝑥
𝑎0 = 𝑙 ∫𝑐 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = 𝑙 ∫𝑐 𝑓(𝑥) cos 𝑑𝑥 and 𝑏𝑛 = 𝑙 ∫𝑐 𝑓(𝑥) sin 𝑑𝑥 ___(7)
𝑙 𝑙

Thus the Fourier series for 𝑓(𝑥) in 𝑐 < 𝑥 < 𝑐 + 2𝑙 is given by (6) and the corresponding Fourier
coefficients are given by (7).
Particular Case I :
Fourier series for 𝑓(𝑥) in 0 < 𝑥 < 2𝑙 can be obtained by putting 𝑐 = 0 in (6) and (7). . Thus, we
have
7
PMAT 32322 Mathematical Methods [Type here]

𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos ( ) + ∑∞
𝑛=1 𝑏𝑛 sin ( )__(8)
2 𝑙 𝑙

1 2𝑙 1 2𝑙 𝑛𝜋𝑥 1 2𝑙 𝑛𝜋𝑥
𝑎0 = 𝑙 ∫0 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = 𝑙 ∫0 𝑓(𝑥) cos 𝑑𝑥 and 𝑏𝑛 = 𝑙 ∫0 𝑓(𝑥) sin 𝑑𝑥.__(9)
𝑙 𝑙

Particular Case II :
Fourier series for 𝑓(𝑥) in −𝑙 < 𝑥 < 𝑙 can be obtained by putting 𝑐 = −𝑙 in (6) and (7). . Thus,
we have
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos ( ) + ∑∞
𝑛=1 𝑏𝑛 sin ( )__(10)
2 𝑙 𝑙

1 𝑙 1 𝑙 𝑛𝜋𝑥 1 𝑙 𝑛𝜋𝑥
𝑎0 = 𝑙 ∫−𝑙 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = 𝑙 ∫−𝑙 𝑓(𝑥) cos 𝑑𝑥 and 𝑏𝑛 = 𝑙 ∫−𝑙 𝑓(𝑥) sin 𝑑𝑥.__(11)
𝑙 𝑙

We now discuss two special forms of the above case II.

Special Form I:
Fourier Series of 𝒇(𝒙) in −𝒍 < 𝒙 < 𝒍 when 𝒇(𝒙) is an even function.
𝑛𝜋𝑥 𝑛𝜋𝑥
Note that if 𝑓(𝑥) is even function, then 𝑓(𝑥) cos is an even function whereas 𝑓(𝑥) sin is
𝑙 𝑙
an odd function. Accordingly, (11) may be re-written as
2 𝑙 2 𝑙 𝑛𝜋𝑥
𝑎0 = ∫0 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = ∫0 𝑓(𝑥) cos 𝑑𝑥 and 𝑏𝑛 = 0____(12)
𝑙 𝑙 𝑙

Thus, Fourier series of an even function 𝑓(𝑥) in (−𝑙. 𝑙) is given by

𝑎0 𝑛𝜋𝑥
𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos ( ) __(13)
2 𝑙

Where 𝑎0 and 𝑎𝑛 are given by (12).

Special Form II:
Fourier Series of 𝒇(𝒙) in −𝒍 < 𝒙 < 𝒍 when 𝒇(𝒙) is an odd function.
𝑛𝜋𝑥 𝑛𝜋𝑥
Note that if 𝑓(𝑥) is odd function, then 𝑓(𝑥) cos is an odd function whereas 𝑓(𝑥) sin is an
𝑙 𝑙
even function. Accordingly, (11) may be re-written as
2 𝑙 𝑛𝜋𝑥
𝑎0 = 0, 𝑎𝑛 = 0 and 𝑏𝑛 = 𝑙 ∫0 𝑓(𝑥) sin 𝑑𝑥____(14)
𝑙

Thus, Fourier series of an even function 𝑓(𝑥) in (−𝑙. 𝑙) is given by

𝑛𝜋𝑥
𝑓(𝑥) = ∑∞
𝑛=1 𝑏𝑛 sin ( ) __(15)
𝑙

Where 𝑏𝑛 are given by (14).

Example 04:
Find the Fourier series of 𝑓(𝑥) = 𝑥 2 in (−1,1).

8
PMAT 32322 Mathematical Methods [Type here]

Solution:
Since 𝑓(𝑥) = 𝑥 2 is an even function of x, the required Fourier series is given by
𝑎0
𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos 𝜋𝑥 , where ___(1)
2
1
2 1 𝑥3 2
𝑎0 = 1 ∫0 𝑥 2 𝑑𝑥 = 2 [ 3 ] = 3 .
0

1 sin 𝑛𝜋𝑥 1 − cos 𝑛𝜋𝑥 − sin 𝑛𝜋𝑥

𝑎𝑛 = 2 ∫0 𝑥 2 cos 𝑛𝜋𝑥 𝑑𝑥 = 2 [[(𝑥 2 ) ( )] − (2𝑥) ( ) + 2( )]
𝑛𝜋 0 𝑛2 𝜋 2 𝑛3 3

2 4(−1)𝑛
= 2 (𝑛2𝜋2 ) cos 𝑛𝜋 = .
𝑛2 𝜋 2

1 4(−1)𝑛
From (1), we have 𝑥 2 = 3 + ∑∞
𝑛=1 cos 𝜋𝑥 .
𝑛2 𝜋 2

Example 05:
𝑥 0<𝑥<1 1 1 1 𝜋2
Find Fourier series for 𝑓(𝑥) = { . Deduce + 22 + 32 + ⋯ = .
1−𝑥 1<𝑥<2 12 8

Solution:
𝑎0
𝑓(𝑥) = + ∑∞ ∞
𝑛=1 𝑎𝑛 cos 𝑛𝜋𝑥 + ∑𝑛=1 𝑏𝑛 sin 𝑛𝜋𝑥 ___(1)
2
1 2
2 1 2 𝑥2 𝑥2
𝑎0 = ∫0 𝑓(𝑥)𝑑𝑥 = ∫0 𝑥𝑑𝑥 + ∫1 (1 − 𝑥)𝑑𝑥 = [ 2 ] + {𝑥 − }
0 2 1

1 1
= 2 + (2 − 2 − 1 + 2) = 0.
2 1 2
𝑎𝑛 = ∫0 𝑥 cos 𝑛𝜋𝑥 𝑑𝑥 = ∫0 𝑥 cos 𝑛𝜋𝑥 𝑑𝑥 + ∫1 (1 − 𝑥) cos 𝑛𝜋𝑥 𝑑𝑥

sin 𝑛𝜋𝑥 1 sin 𝑛𝜋𝑥 1 sin 𝑛𝜋𝑥 2 2

= [𝑥 ( ) − ∫0 1 ( ) 𝑑𝑥] + [(1 − 𝑥) ( ) + ∫1 cos 𝑛𝜋𝑥 𝑑𝑥]
𝑛𝜋 𝑛𝜋 0 𝑛𝜋 1
1
sin 𝑛𝜋𝑥 cos 𝑛𝜋𝑥 sin 𝑛𝜋𝑥 cos 𝑛𝜋𝑥 2
= [𝑥 ( )− ((− (𝑛𝜋)2 ))] + [(1 − 𝑥) ( ) + (−1) (− (𝑛𝜋)2
)]
𝑛𝜋 𝑛𝜋 1
0

(−1)𝑛 1 1 (−1)𝑛
= [ 𝑛 2 𝜋 2 − 𝑛 2 𝜋 2 ] + [𝑛 2 𝜋 2 + ]
𝑛2 𝜋 2

0 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛
𝑎𝑛 = { −4
𝑛 𝑖𝑠 𝑜𝑑𝑑 .
𝑛2 𝜋 2
2 1 2
Also similarly 𝑏𝑛 = ∫0 𝑥 sin 𝑛𝜋𝑥 𝑑𝑥 = ∫0 𝑥 sin 𝑛𝜋𝑥 𝑑𝑥 + ∫1 (1 − 𝑥) sin 𝑛𝜋𝑥 𝑑𝑥
1−(−1)𝑛
𝑏𝑛 = .
𝑛𝜋

9
PMAT 32322 Mathematical Methods [Type here]

0 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛
After some simplifications, we have 𝑏𝑛 = { 2 𝑛 𝑖𝑠 𝑜𝑑𝑑 .
𝑛𝜋

4 cos(2𝑛−1)𝑥 2 sin(2𝑛−1)𝑥
Therefore, we have 𝑓(𝑥) = − 𝜋2 ∑∞
𝑛=1 (2𝑛−1)2
+ 𝜋 ∑∞
𝑛=1 (2𝑛−1)
.

Half Range Series

In several engineering and physical applications, sometimes it is required to get the Fourier
series expansion of a function in an interval (0, 𝑙) where 𝑙 is the half the period. Clearly, when
we require to expand a function in the interval (0, 𝑙), then it is immaterial what the function may
be outside the range 0 < 𝑥 < 𝑙. We are free to choose it arbitrarily in the interval (−𝑙, 0).
Half Range Sine Series
Let 𝑓(𝑥) be defined in (0, 𝑙). We now define a new function 𝐹(𝑥) as follows:
𝑓(𝑥) 0<𝑥<𝑙
𝐹(𝑥) = { ___(1)
−𝑓(−𝑥) −𝑙 < 𝑥 < 0
The function 𝐹(𝑥) defined by (1) is an odd function in (−𝑙, 𝑙). Hence the Fourier series of 𝐹(𝑥)
contains only sine terms. Also note that in the interval (0, 𝑙), 𝐹(𝑥) = 𝑓(𝑥) and hence the sine
series of 𝐹(𝑥) gives the sine series of 𝑓(𝑥) in (0, 𝑙). Thus, we have
𝑛𝜋𝑥 2 𝑙 𝑛𝜋𝑥
𝑓(𝑥) = ∑∞
𝑛=1 𝑏𝑛 sin ( ) where 𝑏𝑛 = 𝑙 ∫0 𝑓(𝑥) sin 𝑑𝑥 .
𝑙 𝑙

Particular case:
If 𝑓(𝑥) is defined in (0, 𝜋) so that 𝑙 = 𝜋, then the required half range series is given by
2 𝜋
𝑓(𝑥) = ∑∞
𝑛=1 𝑏𝑛 sin 𝑛𝑥 where 𝑏𝑛 = 𝜋 ∫0 𝑓(𝑥) sin 𝑛𝑥 𝑑𝑥 .

Half Range Cosine Series

Let 𝑓(𝑥) be defined in (0, 𝑙). We now define a new function 𝐺(𝑥) as follows:
𝑓(𝑥) 0<𝑥<𝑙
𝐺(𝑥) = { ___(1)
𝑓(−𝑥) −𝑙 < 𝑥 < 0
The function 𝐺(𝑥) defined by (1) is an even function in (−𝑙, 𝑙). Hence the Fourier series of 𝐺(𝑥)
contains only cosine terms. Also note that in the interval (0, 𝑙), 𝐹(𝑥) = 𝑓(𝑥) and hence the
cosine series of 𝐺(𝑥) gives the cosine series of 𝑓(𝑥) in (0, 𝑙). Thus, we have
𝑎0 𝑛𝜋𝑥
𝑓(𝑥) = + ∑∞
𝑛=1 𝑎𝑛 cos ( )
2 𝑙
2 𝑙 2 𝑙 𝑛𝜋𝑥
𝑎0 = 𝑙 ∫0 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = 𝑙 ∫0 𝑓(𝑥) cos 𝑑𝑥 .
𝑙

10
PMAT 32322 Mathematical Methods [Type here]

Particular case:
If 𝑓(𝑥) is defined in (0, 𝜋) so that 𝑙 = 𝜋, then the required half range series is given by
∞
𝑎0
𝑓(𝑥) = + ∑ 𝑎𝑛 cos 𝑛𝜋𝑥
2
𝑛=1
2 𝜋 2 𝜋
𝑎0 = 𝜋 ∫0 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = 𝜋 ∫0 𝑓(𝑥) cos 𝑛𝑥 𝑑𝑥 .

Example 06:
Find half-range sine series for 𝑒 𝑥 for 0 < 𝑥 < 1.
Solution:
Let the required series be 𝑒 𝑥 = ∑∞
𝑛=1 𝑏𝑛 sin 𝑛𝜋𝑥 . ___(1)

1 𝑒 𝑥 (sin 𝑛𝜋𝑥−𝑛𝜋 cos 𝑛𝜋𝑥) 1

where 𝑏𝑛 = 2 ∫0 𝑒 𝑥 sin 𝑛𝜋𝑥 𝑑𝑥 = 2 [ ]
1+𝑛2 𝜋 2 0

𝑒(−𝑛𝜋 cos 𝑛𝜋) (−𝑛𝜋) 𝑒 𝑎𝑥 (𝑎 sin 𝑏𝑥−𝑏 cos 𝑏𝑥)

= 2[ − 1+𝑛2𝜋2] as ∫ 𝑒 𝑎𝑥 sin 𝑏𝑥 𝑑𝑥 = .
1+𝑛2 𝜋 2 𝑎2 +𝑏 2

2 2𝑛𝜋
= (1+𝑛2𝜋2) (−𝑒𝑛𝜋(−1)𝑛 + 𝑛𝜋) = (1+𝑛2 𝜋2) (1 − 𝑒(−1)𝑛 )..
2
From (1), we have 𝑒 𝑥 = 𝑛𝜋 ∑∞ 𝑛
𝑛=1 (1+𝑛2 𝜋 2 ) (1 − 𝑒(−1) ) sin 𝑛𝜋𝑥 .

Problem:
Find a series of cosines of multiples of x which will represent 𝑥 sin 𝑥 in (0, 𝜋) and show that
2 2 2 𝜋
1 + 1×3 − 3×5 + 5×7 − ⋯ = .
2

**** The End ****

11
PMAT 32322 Mathematical Methods [Type here]

12