FOURIER SERIES AND FOURIER INTEGRALS

3.1 Introduction
Periodic functions occur frequently in engineering problems. The representation of these
engineering problems in term of simple Periodic functions, such as sine and cosine is matter of
great practical importance, which leads to Fourier series. These series, named after the physicist
JOSEPH FURIER (1768-1830), are a very powerful tool in connection with various problems
involving ordinary and partial differential equations. Here we shall discuss basic concepts, facts
and techniques in connection with Fourier series. Some illustrative examples and also some
important engineering applications of these series will be included.

3.2 Periodic functions and trigonometric series

Definition: The function 𝑓(𝑥) of a real variable 𝑥 is said to be periodic if there exists a non-zero
number p, independent of 𝑥, such that the equation 𝑓(𝑥 + 𝑝) = 𝑓(𝑥) holds for all values of 𝑥.
The least value of 𝑝 > 0 is called the least period or simply the period of 𝑓(𝑥).

Examples of periodic functions

Example 1. 𝑓(𝑥) = sin 𝑠𝑖𝑛 𝑥 , is a periodic function having period 2π.

Proof: 𝑓(𝑥) = sin 𝑠𝑖𝑛 𝑥 , 𝑓(𝑥 + 2π) = sin 𝑠𝑖𝑛 (𝑥 + 2π) = sin 𝑠𝑖𝑛 𝑥 = 𝑓(𝑥)

𝑓(𝑥 + 4π) = sin 𝑠𝑖𝑛 (𝑥 + 4π) = sin 𝑠𝑖𝑛 𝑥 = 𝑓(𝑥)

………………………………………………

∴𝑓(𝑥) = 𝑓(𝑥 + 2π) = 𝑓(𝑥 + 4π) = ………….. ..... = 𝑓(𝑥 + 2𝑛π)

Thus 𝑓(𝑥) = sin 𝑠𝑖𝑛 𝑥 is a periodic function having period 2π.

1
Similarly, we can show that 𝑓(𝑥) = cos 𝑐𝑜𝑠 𝑥 is also periodic function having period 2π.

Graphs of cosine and sine functions having period 2π.

Example 2. 𝑓(𝑥) = tan 𝑡𝑎𝑛 𝑥 is a periodic function having period π.

Proof: 𝑓(𝑥) = tan 𝑡𝑎𝑛 𝑥 , 𝑓(𝑥 + π) = tan 𝑡𝑎𝑛 (𝑥 + π) = tan 𝑡𝑎𝑛 𝑥 ,

𝑓(𝑥 + 2π) = tan 𝑡𝑎𝑛 (𝑥 + 2π) = tan 𝑡𝑎𝑛 𝑥

………………………………………

𝑓(𝑥 + 𝑛π) = tan 𝑡𝑎𝑛 (𝑥 + 𝑛π) = tan 𝑡𝑎𝑛 𝑥 .

∴𝑓(𝑥) = 𝑓(𝑥 + π) = 𝑓(𝑥 + 2π) = ……… = 𝑓(𝑥 + 𝑛π)

Thus 𝑓(𝑥) = tan 𝑡𝑎𝑛 𝑥 is a periodic function having period 𝜋.

Definition By Trigonometric series we shall mean any series of the form

𝑎0 ∞

2 (
+ ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥
𝑛=1
)
Where the coefficients 𝑎𝑛𝑎𝑛𝑑 𝑏𝑛 are constants.

2
3.3 Fourier series
Definition: The trigonometric series

𝑓(𝑥) = 𝑎0 + 𝑎 cos 𝑐𝑜𝑠 𝑥 + 𝑎2 cos 𝑐𝑜𝑠 2𝑥 + ……….. + 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + …….. + 𝑏1 sin 𝑠𝑖𝑛 𝑥 + 𝑏2 sin 𝑠𝑖𝑛
1

∞
(
= 𝑎0 + ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ……….. (1)
𝑛=1
)
Is a Fourier series if its coefficients 𝑎0, 𝑎𝑛𝑎𝑛𝑑 𝑏𝑛 are given by the following formulas:

π π
1 1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥 = 2π
∫ 𝑓(𝑣)𝑑𝑣
−π −π

π π
1 1
𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 = π
∫ 𝑓(𝑣) cos 𝑐𝑜𝑠 𝑛𝑣 𝑑𝑣
−π −π

π π
1 1
𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 = π
∫ 𝑓(𝑣) sin 𝑠𝑖𝑛 𝑛𝑣 𝑑𝑣
−π −π

∀ 𝑛 = 1, 2, 3, 4, 5, ……………..

Where 𝑓(𝑥) is any singled valued function defined on the interval (− π, π).

The Fourier series can also be written as

𝑎0 ∞
𝑓(𝑥) = 2 (
+ ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥
𝑛=1
)
Where
π
1
𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥
−π

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥, (𝑛 = 0, 1, 2, 3…………..)
−π

π
1
And 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥, (𝑛 = 1, 2, 3, ………….)
−π

3
3.4 Process of determining the coefficient of 𝑎0, 𝑎𝑛, 𝑏𝑛:

(i) Let us determine 𝑎0

Given,
∞
(
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 ∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙∙(1)
𝑛=1
)
Integrating both sides of (1) with respect to 𝑥 from –π to π.

We get,
π π ∞

−π −π 𝑛=1
(
∫ 𝑓(𝑥)𝑑𝑥 = ∫ [𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 ]𝑑𝑥 )
By term by term integration of the series we obtain
π π ∞ π π
⎡ ⎤
∫ 𝑓(𝑥)𝑑𝑥 = ∫ 𝑎0𝑑𝑥 + ∑ ⎢𝑎𝑛 ∫ 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 + 𝑏𝑛 ∫ 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥⎥ ∙∙∙∙∙∙∙(2)
⎢ −π
𝑛=1⎣
⎥
−π −π −π ⎦
Now,
π π π
∫ 𝑎0𝑑𝑥 = 2π𝑎0 𝑎𝑛𝑑 ∫ 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 = ∫ 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥 = 0
−π −π −π

π π
1
∴ ∫ 𝑓(𝑥)𝑑𝑥 = 2π𝑎0, 𝑜𝑟, 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π −π

π
1
= 2π
∫ 𝑓(𝑣)𝑑𝑣
−π

π π
1 1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥 = 2π
∫ 𝑓(𝑣)𝑑𝑣
−π −π

(ii) To determine 𝑎1, 𝑎2, ∙∙∙∙∙∙∙∙∙ in (1), we multiply both sides of (1) by 𝑐𝑜𝑠𝑚𝑥 where 𝑚 is any
fixed positive integer and then integrate both sides of (1) with respect to 𝑥 from –π to π finding.

4
π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑚𝑥 𝑑𝑥
−π

π ∞ π π
⎡ ⎤
= 𝑎0 ∫ 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥 + ∑ ⎢𝑎𝑛 ∫ 𝑐𝑜𝑠𝑛𝑥 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥 + 𝑏𝑛 ∫ 𝑠𝑖𝑛𝑛𝑥 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥⎥
⎢ −π
𝑛=1⎣
⎥
−π −π ⎦
∙∙∙∙∙(3)
π π
𝑁𝑜𝑤, ∫ 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥 = 0, ∫ 𝑠𝑖𝑛𝑛𝑥 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥 = 0
−π −π

π π π
1 1
∫ 𝑐𝑜𝑠𝑛𝑥 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥 = 2
∫ cos 𝑐𝑜𝑠 (𝑛 + 𝑚) 𝑥 𝑑𝑥 + 2
∫ cos 𝑐𝑜𝑠 (𝑛 − 𝑚) 𝑥 𝑑𝑥
−π −π −π

π π
1 1
𝑊ℎ𝑒𝑟𝑒, 2
∫ cos 𝑐𝑜𝑠 (𝑛 + 𝑚) 𝑥 𝑑𝑥 = 0, 2
∫ cos 𝑐𝑜𝑠 (𝑛 − 𝑚) 𝑥 𝑑𝑥 = {0 𝑖𝑓 𝑚≠𝑛 π 𝑖𝑓 𝑚 = 𝑛
−π −π

π π
1
∴ ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑚𝑥 𝑑𝑥 = 𝑎𝑚π 𝑜𝑟, 𝑎𝑚 = π
∫ 𝑐𝑜𝑠𝑚𝑥 𝑑𝑥
−π −π

Now writing 𝑛 for 𝑚, we get

π π
1 1
𝑎𝑛 = π
∫ 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 = π
∫ 𝑐𝑜𝑠𝑛𝑣 𝑑𝑣
−π −π

(iii) Finally, we determine 𝑏1, 𝑏2, ∙∙∙∙∙∙∙∙∙ in (1).

If we multiply (1) by 𝑠𝑖𝑛 𝑚𝑥, where 𝑚 is any fixed positive integer and then integrate with
respect to 𝑥 from –π to π, finding
π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑚𝑥𝑑𝑥
−π

π ∞ π π
= 𝑎0 ∫ 𝑠𝑖𝑛 𝑚𝑥𝑑𝑥 + ∑ [𝑎𝑛 ∫ 𝑐𝑜𝑠𝑛𝑥 𝑠𝑖𝑛 𝑚𝑥 𝑑𝑥 + 𝑏𝑛 ∫ 𝑠𝑖𝑛𝑛𝑥 𝑠𝑖𝑛 𝑚𝑥 𝑑𝑥](4)
−π 𝑛=1 −π −π

π π
𝑁𝑜𝑤, ∫ 𝑠𝑖𝑛 𝑚𝑥𝑑𝑥 = 0, ∫ 𝑐𝑜𝑠𝑛𝑥 𝑠𝑖𝑛 𝑚𝑥 𝑑𝑥 = 0
−π −π

π π π
1 1
∫ 𝑠𝑖𝑛𝑛𝑥 𝑠𝑖𝑛 𝑚𝑥 𝑑𝑥 = 2
∫ cos 𝑐𝑜𝑠 (𝑛 − 𝑚)𝑥 𝑑𝑥 − 2
∫ cos 𝑐𝑜𝑠 (𝑛 + 𝑚)𝑥 𝑑𝑥
−π π π

5
 π
1
𝑊ℎ𝑒𝑟𝑒, 2
∫ cos 𝑐𝑜𝑠 (𝑛 + 𝑚)𝑥 𝑑𝑥 = 0 𝑎𝑛𝑑
π

π
1
2
∫ cos 𝑐𝑜𝑠 (𝑛 − 𝑚) 𝑥 𝑑𝑥 = {0 𝑖𝑓 𝑛≠𝑚 π 𝑖𝑓 𝑛 = 𝑚
−π

Thus from (4), we have

π π
1
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑚𝑥𝑑𝑥 = 𝑏𝑚π, 𝑜𝑟 , 𝑏𝑚 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑚𝑥𝑑𝑥
−π −π

Writing 𝑛 for 𝑚, we have

π π
1 1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑛𝑥𝑑𝑥 = π
∫ 𝑓(𝑣)𝑠𝑖𝑛 𝑛𝑣𝑑𝑣
−π −π

Therefore,
π ∞ π
1 1
𝑓(𝑥) = 2π
∫ 𝑓(𝑣)𝑑𝑣 + π
∑ [{ ∫ 𝑓(𝑣)𝑐𝑜𝑠𝑛𝑣 𝑑𝑣}𝑐𝑜𝑠𝑛𝑥
–π 𝑛=1 −π

π
+ { ∫ 𝑓(𝑣)𝑠𝑖𝑛 𝑛𝑣𝑑𝑣}𝑠𝑖𝑛𝑛𝑥]
−π

π ∞ π
1 1
= 2π
∫ 𝑓(𝑣)𝑑𝑣 + π
∑ [ ∫ 𝑓(𝑣){𝑐𝑜𝑠𝑛𝑣 𝑐𝑜𝑠𝑛𝑥 + sin 𝑠𝑖𝑛 𝑛𝑣 𝑠𝑖𝑛𝑛𝑥}𝑑𝑣]
–π 𝑛=1 −π

π ∞ π
1 1
= 2π
∫ 𝑓(𝑣)𝑑𝑣 + π
∑ ∫ 𝑓(𝑣) cos 𝑐𝑜𝑠 𝑛(𝑥 − 𝑣)𝑑𝑣
–π 𝑛=1 −π

3.5 The Fourier cosine sine series:

Definition Even Function:

A function 𝑓(𝑥) is called even if 𝑓(− 𝑥) = 𝑓(𝑥).

Graphically, an even function is symmetrical about the 𝑦-axis.

If 𝑓(𝑥) is an even function, then

6
π 0 π
∫ 𝑓(𝑥)𝑑𝑥 = ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥
−π −π 0

0 π
= ∫ 𝑓(− 𝑥)𝑑(− 𝑥) + ∫ 𝑓(𝑥)𝑑𝑥
π 0

0 π
=− ∫ 𝑓(𝑥)𝑑(𝑥) + ∫ 𝑓(𝑥)𝑑𝑥
π 0

π π
= ∫ 𝑓(𝑥)𝑑(𝑥) + ∫ 𝑓(𝑥)𝑑𝑥
0 0

π
= 2 ∫ 𝑓(𝑥)𝑑𝑥
0

Thus if 𝑓(𝑥) is even, we have

π
1
𝑎0 = 2π
. 2 ∫ 𝑓(𝑥)𝑑𝑥
0

π
1
= π
∫ 𝑓(𝑥)𝑑𝑥
0

π
1
= π
∫ 𝑓(𝑣)𝑑𝑣
0

Also if 𝑓(𝑥) is an even i.e. 𝑓(− 𝑥) = 𝑓(𝑥), then

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 (𝑛 = 1, 2, 3, ∙∙∙∙∙∙∙∙)
−π

−π
1
= π
∫ 𝑓(− 𝑥)𝑐𝑜𝑠𝑛(− 𝑥)𝑑(− 𝑥)
π

−π
1
=− π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
π

π
1
= π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
−π

7
 π
2
= π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
0

π
2
= π
∫ 𝑓(𝑣)𝑐𝑜𝑠𝑛𝑣 𝑑𝑣
0

But,
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π

−π
1
= π
∫ 𝑓(− 𝑥)𝑠𝑖𝑛𝑛(− 𝑥)𝑑(− 𝑥)
π

−π
1
= π
∫ 𝑓(𝑥). (− 𝑠𝑖𝑛𝑛𝑥)(− 𝑑𝑥)
π

π
1
=− π
∫ 𝑓(𝑥) 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π

=− 𝑏𝑛

∴ 2𝑏𝑛 = 0 or, 𝑏𝑛 = 0

Therefore, if 𝑓(𝑥) is an even, then we have

∞
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛 𝑐𝑜𝑠𝑛𝑥
𝑛=1

π ∞ π
1 2
= π
∫ 𝑓(𝑣)𝑑𝑣 + π
∑ {∫ 𝑓(𝑣)𝑐𝑜𝑠𝑛𝑣 𝑑𝑣} 𝑐𝑜𝑠𝑛𝑥
0 𝑛=1 0

Which represents the function 𝑓(𝑥) in a series of 𝑐𝑜𝑠𝑖𝑛𝑒𝑠 and therefore it is known as 𝑐𝑜𝑠𝑖𝑛𝑒
series in the interval (0, π).

Definition Odd Function:

A function 𝑓(𝑥) is called even if 𝑓(− 𝑥) =− 𝑓(𝑥).

Graphically, an odd function is symmetrical about the origin.

8
When 𝑓(𝑥) is odd, we have
π
1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

−π
1
= 2π
∫ 𝑓(− 𝑥)𝑑(− 𝑥)
π

−π
1
= 2π
∫ 𝑓(𝑥)𝑑𝑥
π

π
1
=− 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

= − 𝑎0

∴ 2𝑎0 = 0 or, 𝑎0 = 0

Also, if 𝑓(𝑥) is odd i.e. 𝑓(− 𝑥) =− 𝑓(𝑥), then

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
−π

−π
1
= π
∫ 𝑓(− 𝑥)𝑐𝑜𝑠𝑛(− 𝑥)𝑑(− 𝑥)
π

−π
1
= π
∫ − 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥(− 𝑑𝑥)
π

−π
1
= π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
π

π
1
=− π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
−π

=− 𝑎𝑛

∴ 2𝑎𝑛 = 0 or, 𝑎𝑛 = 0

π
1
𝐹𝑖𝑛𝑎𝑙𝑙𝑦, 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π

9
 −π
1
= π
∫ 𝑓(− 𝑥)𝑠𝑖𝑛𝑛(− 𝑥)𝑑(− 𝑥)
π

−π
1
= π
∫ − 𝑓(𝑥). − 𝑠𝑖𝑛𝑛𝑥. − 𝑑𝑥
π

−π
1
=− π
∫ 𝑓(𝑥) 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
π

π
1
= π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π

π
2
= π
∫ 𝑓(𝑥) 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
0

π
2
= π
∫ 𝑓(𝑣) 𝑠𝑖𝑛𝑛𝑣 𝑑𝑣
0

Therefore, if 𝑓(𝑥) is odd, then we have

∞
𝑓(𝑥) = ∑ 𝑏𝑛𝑠𝑖𝑛𝑛𝑥
𝑛=1

∞ π
2
= π
∑ {∫ 𝑓(𝑣) 𝑠𝑖𝑛𝑛𝑣 𝑑𝑣}𝑠𝑖𝑛𝑛𝑥
𝑛=1 0

Which represents the function 𝑓(𝑥) in a series of 𝑠𝑖𝑛𝑒𝑠 in the interval (0, π) and therefore it is
known as 𝑠𝑖𝑛𝑒𝑠 series in the interval (0, π).

Half Range Fourier Cosine and Sine Series:

When the Fourier series has only the cosine terms or only the sine terms in the expression we call
the series Half Range Fourier Cosine or Half Range Fourier Sine Series respectively. When
we are interested to find out a half range series corresponding to given a function, the function
must be defined in the interval (0, π), which is the half of the interval (− π, π) and then the
function is specified as odd or even, so that it is clearly defined in the other half of the interval,
namely, (0, π) in such a case we have

10
 π
2
𝑏𝑛 = 0, 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 for half range cosine series and
0

π
2
𝑎𝑛 = 0, 𝑏𝑛 = π
∫ 𝑓(𝑥) 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥 for half range sine series
0

3.7 Dirichlet’s Conditions:

∞
(
Any function 𝑓(𝑥) can be developed as a Fourier series 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 where
𝑛=1
)
𝑎0, 𝑎𝑛 and 𝑏𝑛 are constant, provided.

(i) 𝑓(𝑥) is periodic, single-valued and finite;

(ii) 𝑓(𝑥) has a finite number of discontinuous in any one period;
(iii) 𝑓(𝑥) has at the most finite number of maxima and minima;

In fact the problem of expressing any function 𝑓(𝑥) as a Fourier series depends upon the
1 1
evaluations of the integrals π
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥; π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 . within the limit

(0, 2π), (− π, π) or (∝, ∝ + 2π) according as 𝑓(𝑥) is defined for every value of 𝑥 in
(0, 2π), (− π, π) or (∝, ∝ + 2π).

Change of intervals:
(A) If 𝑓(𝑥) is defined in the interval (− 𝑐, 𝑐) having period 2𝑐, the Fourier series of 𝑓(𝑥) in
the interval (− 𝑐, 𝑐) is given by
∞
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠
𝑛=1
( 𝑛π𝑥
𝑐
+ 𝑏𝑛𝑠𝑖𝑛
𝑛π𝑥
𝑐 )
𝑐
1
𝑊ℎ𝑒𝑟𝑒, 𝑎0 = 2𝑐
∫ 𝑓(𝑥)𝑑𝑥 (𝑛 = 0)
−𝑐

11
 𝑐
1 𝑛π𝑥
𝑎𝑛 = 𝑐
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑐
𝑑𝑥 (𝑛 = 1, 2, 3∙∙∙∙∙)
−𝑐

𝑐
1 𝑛π𝑥
𝑏𝑛 = 𝑐
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑐
𝑑𝑥 (𝑛 = 1, 2, 3∙∙∙∙∙)
−𝑐

(B) If 𝑓(𝑥) is defined in the interval (∝, ∝ + 2𝑐) having period 2𝑐, the Fourier series of
𝑓(𝑥) in the interval (∝, ∝ + 2𝑐) is given by
∞
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠
𝑛=1
( 𝑛π𝑥
𝑐
+ 𝑏𝑛𝑠𝑖𝑛
𝑛π𝑥
𝑐 )
∝+2𝑐
1
𝑊ℎ𝑒𝑟𝑒, 𝑎0 = 𝑐
∫ 𝑓(𝑥)𝑑𝑥 (𝑛 = 0)
∝

∝+2𝑐
1 𝑛π𝑥
𝑎𝑛 = 𝑐
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑐
𝑑𝑥 (𝑛 = 1, 2, 3 ∙∙∙∙∙)
∝

∝+2𝑐
1 𝑛π𝑥
𝑏𝑛 = 𝑐
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑐
𝑑𝑥 (𝑛 = 1, 2, 3 ∙∙∙∙∙)
∝

3.9 Fourier Integral

The Fourier Integral is very useful in the field of electrical communication and forms the basis of
Cauchy’s method for the solution of the partial differential equation.

General Fourier series of a periodic function 𝑓(𝑥) in the interval (− 𝑐, 𝑐) is given by

𝑐 ∞ 𝑐 ∞ 𝑐
1 1 𝑛π𝑡 𝑛π𝑥 1 𝑛π𝑡 𝑛π𝑥
𝑓(𝑥) = 2𝑐
∫ 𝑓(𝑡)𝑑𝑡 + ∑ 𝑐
∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 𝑐
𝑑𝑡. cos 𝑐𝑜𝑠 𝑐
+ ∑ 𝑐
∫ 𝑓(𝑡) sin 𝑠𝑖𝑛 𝑐
𝑑𝑡. sin 𝑠𝑖𝑛 𝑐
−𝑐 𝑛=1 −𝑐 𝑛=1 −𝑐

𝑐 𝑐 ∞
⎡
⇒𝑓(𝑥) =
1
2𝑐
∫ 𝑓(𝑡)𝑑𝑡 +
−𝑐
1
𝑐
∫ 𝑓(𝑡)⎢ ∑ cos 𝑐𝑜𝑠
−𝑐
⎢𝑛=1
⎣
{ 𝑛π𝑥
𝑐
cos 𝑐𝑜𝑠
𝑛π𝑡
𝑐
+ sin 𝑠𝑖𝑛
𝑛π𝑥
𝑐
sin 𝑠𝑖𝑛
𝑛π𝑡
𝑐 }⎤⎥⎥⎦𝑑𝑡
𝑐 𝑐 ∞
⎡
=
1
2𝑐
∫ 𝑓(𝑡)𝑑𝑡 +
−𝑐
1
2𝑐
∫ 𝑓(𝑡)⎢2 ∑ cos 𝑐𝑜𝑠
−𝑐
⎢ 𝑛=1
⎣
{ 𝑛π(𝑥−𝑡)
𝑐 } 𝑑𝑡⎤⎥⎥⎦
𝑐 ∞
⎡
⇒𝑓(𝑥) =
1
2𝑐
∫ 𝑓(𝑡)⎢1 + 2 ∑ cos 𝑐𝑜𝑠
−𝑐
⎢
⎣ 𝑛=1
{ 𝑛π(𝑥−𝑡)
𝑐 } ⎤⎥⎥⎦𝑑𝑡

12
 𝑐 ∞
⎡π
=
1
2π
∫ 𝑓(𝑡)⎢ 𝑐 + ∑ 2.
−𝑐
⎢
⎣ 𝑛=1
π
𝑐
. cos 𝑐𝑜𝑠 { 𝑛π(𝑥−𝑡)
𝑐 } ⎤⎥⎥⎦𝑑𝑡
𝑐 ∞ ∞
⎡π
=
1
2π
∫ 𝑓(𝑡)⎢ 𝑐 cos 𝑐𝑜𝑠 0.
−𝑐
⎢
⎣
{ π
𝑐
(𝑥 − 𝑡) } + ∑
𝑛=1
π
𝑐
cos 𝑐𝑜𝑠 { 𝑛π(𝑥−𝑡)
𝑐 } + ∑
𝑛=1
π
𝑐
cos 𝑐𝑜𝑠 { −𝑛π(𝑥−𝑡)
𝑐 } ⎤⎥⎥⎦𝑑𝑡
𝑐 ∞ ∞
⎡
⇒𝑓(𝑥) =
1
2π
∫ 𝑓(𝑡).
−𝑐
π
𝑐
⎢ ∑ cos 𝑐𝑜𝑠
⎢𝑛=0
⎣
{ 𝑛π(𝑥−𝑡)
𝑐 } + ∑ cos 𝑐𝑜𝑠
𝑛=1
{ −𝑛π(𝑥−𝑡)
𝑐 } ⎤⎥⎥⎦𝑑𝑡.

𝑐 𝑛
⎡ ⎤
=
1
2π
∫ 𝑓(𝑡)⎢ ∑
−𝑐
⎢𝑟=−𝑛
⎣
π
𝑐
cos 𝑐𝑜𝑠 { 𝑟π
𝑐
(𝑥 − 𝑡) ⎥𝑑𝑡
⎥
⎦
}
𝑐 𝑛
1 ⎡ 1 ⎰ 𝑟 ⎱⎤
⇒𝑓(𝑥) = ∫ 𝑓(𝑡)⎢ ∑ cos 𝑐𝑜𝑠 (𝑥 − 𝑡) ⎥𝑑𝑡……………………. (2)
2π ⎢𝑟=−𝑛 𝑐
⎱ 𝑐
⎰⎥
−𝑐 ⎣ π π
⎦
𝑐
Now when, 𝑐→∞, π
→∞ and we have

∞
1 ⎰ 𝑟 ⎱
∑ cos 𝑐𝑜𝑠 (𝑥 − 𝑡 )
𝑟=−∞
𝑐
π
⎱ 𝑐
π
⎰

∞
1
= ∑ cos 𝑐𝑜𝑠 {𝑟∆𝑢(𝑥 − 𝑡)}∆𝑢 𝑤ℎ𝑒𝑟𝑒 ∆𝑢 = 𝑐
𝑟=−∞ π

∞
= ∫ cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)}𝑑𝑢 , {𝑤𝑟𝑖𝑡𝑡𝑖𝑛𝑔 𝑟∆𝑢 = 𝑢 𝑎𝑛𝑑 ∆𝑢 = 𝑑𝑢}
−∞

By the definition of integral as the limit of a sum.

Substituting this value of the sum in the equation (2) we get,

∞ ∞
1
𝑓(𝑥) = 2π
∫ 𝑓(𝑡)𝑑𝑡. ∫ 𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑢……………. (3)
−∞ −∞

This double integral is known as Fourier Integral and holds if x is a point of continuity of (𝑥) .

The second integral in the equation (3) can be written as,

∞ 0 ∞
∫ cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑢 = ∫ cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑢 + ∫ cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑢
−∞ −∞ 0

13
 ∞
= 2 ∫ 𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑢.
0

Thus equation (3) can also be written as,

∞ ∞
1
𝑓(𝑥) = π
∫ 𝑓(𝑡)𝑑𝑡. ∫ 𝑐𝑜𝑠{𝑢(𝑥 − 𝑡)}𝑑𝑢
−∞ 0

∞ ∞
1
= π
∫ 𝑑𝑢 ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡……………………(4)
0 −∞

Which gives another form of the Fourier integral.

When 𝑓(𝑥) is an even function of x that is, 𝑓(− 𝑥) = 𝑓(𝑥), then

∞ 0 ∞
∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑡 = ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡 + ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡
−∞ −∞ 0

[replacing 𝑡 by (− 𝑡) in the first integral]

0 ∞
=− ∫ 𝑓(− 𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 + 𝑡)}𝑑𝑡 + ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡
∞ 0

∞ ∞
= ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 + 𝑡)}𝑑𝑡 + ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡
0 0

∞ ∞
⇒ ∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑡 = 2 ∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 𝑢𝑡 cos 𝑐𝑜𝑠 𝑢𝑥 𝑑𝑡.
−∞ 0

Substituting this result in the equation (4), then

∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑢 ∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 𝑢𝑡 cos 𝑐𝑜𝑠 𝑢𝑥 𝑑𝑡.
0 0

∞ ∞
2
= π
∫ 𝑓(𝑡)𝑑𝑡 ∫ cos 𝑐𝑜𝑠 𝑢𝑡 cos 𝑐𝑜𝑠 𝑢𝑥 𝑑𝑢………………….. (5)
0 0

Which gives Fourier integral, of an even function.

Similarly when 𝑓(𝑥) is an odd function of x, that is, 𝑓(− 𝑥) =− 𝑓(𝑥) we have,

14
∞ 0 ∞
∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑡 = ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡 + ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡
−∞ −∞ 0

On replacing t by (-t) in the first integral on the right hand side we get,
0 0
∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} 𝑑𝑡 = ∫ 𝑓(− 𝑡) cos 𝑐𝑜𝑠 {𝑢(𝑥 + 𝑡)} . (− 𝑑𝑡)
−∞ ∞

∞
=− ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 + 𝑡)}𝑑𝑡
0

∞ ∞
𝑇ℎ𝑢𝑠 ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡 = ∫ 𝑓(𝑡)[cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡)} − cos 𝑐𝑜𝑠 {𝑢(𝑥 + 𝑡)} ]𝑑𝑡
−∞ 0

∞ ∞
⇒ ∫ 𝑓(𝑡)𝑐𝑜𝑠⁡{𝑢(𝑥 − 𝑡)}𝑑𝑡 = 2 ∫ 𝑓(𝑡) sin 𝑠𝑖𝑛 𝑢𝑥 sin 𝑠𝑖𝑛 𝑢𝑡 𝑑𝑡…………………………(6)
−∞ 0

Substituting this relation in the equation (4) we get,

∞ ∞ ∞ ∞
2 2
𝑓(𝑥) = π
∫ 𝑑𝑢 ∫ 𝑓(𝑡) sin 𝑠𝑖𝑛 𝑢𝑡 sin 𝑠𝑖𝑛 𝑢𝑥 𝑑𝑡 = π
∫ 𝑓(𝑡)𝑑𝑡 ∫ sin 𝑠𝑖𝑛 𝑢𝑡 sin 𝑠𝑖𝑛 𝑢𝑥 𝑑𝑢.
0 0 0 0

which is the Fourier integral of an odd function.

15
3.10 Conditions for the convergence of the Fourier series.
We know that a uniformly convergent series of continuous functions is a continuous function.
The Fourier series of a function 𝑓(𝑥) will converge uniformly if the function 𝑓(𝑥) satisfies the
conditions

(𝑖) 𝑓(𝑥) is continuous in the interval − π≤𝑥≤π

(𝑖𝑖) 𝑓(− π) = 𝑓(+ π)

'
(𝑖𝑖𝑖) 𝑓 (𝑥) is sectionally contiouous.

3.11 Differentiation and integration of Fourier series

Continuity of the periodic functions is an important condition for the differentiability of the
Fourier series.

Let 𝑓(𝑥) be a continuous function on the interval − π≤𝑥≤π such that 𝑓(− π) = 𝑓(+ π) and let
𝑓' be piecewie continuous on that interval. Then the Fourier series in the representation

(
𝑓(𝑥) = 𝑎0 + 𝑎𝑛 𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛 𝑠𝑖𝑛𝑛𝑥 ) − π≤𝑥≤π (1)

π π π
1 1 1
Where 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥,𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 and 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π −π −π

''
Is differentiable at each point 𝑥 in the interval − π≤𝑥≤π where 𝑓 (𝑥) exists.
'
Thus 𝑓 (𝑥) is obtained from (1) by term by term differentiation and
∞
'
𝑓 (𝑥) = ∑ 𝑛(− 𝑎𝑛 𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛 𝑠𝑖𝑛𝑛𝑥)
𝑛=1

Again the Fourier series

∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛 𝑠𝑖𝑛𝑛𝑥)
𝑛=1

may be integrated term by term from –π to 𝑥 and the resulting series will converge uniformly to
𝑥
∫ 𝑓(𝑥)𝑑𝑥 provided that 𝑓(𝑥) is sectionally continuous in − π≤𝑥≤π
−π

16
Thus
𝑥 ∞
1
∫ 𝑓(𝑥)𝑑𝑥 = 𝑎0(𝑥 + π) + ∑ 𝑛
[𝑎𝑛( 𝑠𝑖𝑛𝑛𝑥 − 𝑠𝑖𝑛𝑛π) − 𝑏𝑛( 𝑐𝑜𝑠𝑛𝑥 − 𝑐𝑜𝑠𝑛π)]
−π 𝑛=1

∞
= 𝑎0(𝑥 + π) + ∑ [𝑎𝑛𝑠𝑖𝑛𝑛𝑥 − 𝑏𝑛( 𝑐𝑜𝑠𝑛𝑥 − 𝑐𝑜𝑠𝑛π)]
𝑛=1

3.12 Complex from of the Fourier series

From Trigonometry, we have
1 𝑖θ −𝑖θ 1 𝑖θ −𝑖θ
𝑐𝑜𝑠θ = 2
(𝑒 + 𝑒 ) and 𝑠𝑖𝑛θ = 2𝑖
(𝑒 + 𝑒 )

𝑛π𝑥 𝑛π𝑥
Now the complex form of the Fourier series is obtained by expressing 𝑐𝑜𝑠⁡( 𝑐
) and 𝑠𝑖𝑛⁡( 𝑐
)
in exponential form that is the Fourier series.
∞
𝑓(𝑥) = 𝑎0 + ∑ ⎡⎢𝑎𝑛 𝑐𝑜𝑠
𝑛=1⎣
( )+ 𝑏𝑛π𝑥
𝑐 𝑛
𝑠𝑖𝑛 ( )⎤⎥⎦
𝑛π𝑥
𝑐
− 𝑐< 𝑥<𝑐 (1)

can we written in the form

∞

( ) ( )
𝑖𝑛π𝑥 𝑖𝑛π𝑥 𝑖𝑛π𝑥 𝑖𝑛π𝑥
𝑎 − 𝑏𝑛 −
𝑓(𝑥) = 𝑎0 + ∑ ⎡⎢ 2𝑛 𝑒 𝑐
+𝑒 𝑐
+ 2𝑖
𝑒 𝑐
−𝑒 𝑐 ⎤
⎥
𝑛=1⎣ ⎦
∞

( ) ( )
𝑖𝑛π𝑥 𝑖𝑛π𝑥 𝑖𝑛π𝑥 𝑖𝑛π𝑥
𝑎 − 𝑖𝑏𝑛 −
= 𝑎0 + ∑ ⎡⎢ 2𝑛 𝑒 𝑐
+𝑒 𝑐
− 2
𝑒 𝑐
−𝑒 𝑐 ⎤
⎥
𝑛=1⎣ ⎦
∞ 𝑖𝑛π𝑥 𝑖𝑛π𝑥
−
⇒𝑓(𝑥) = 𝑎0 + ∑ ⎡⎢ 2 𝑎𝑛 − 𝑖𝑏𝑛 𝑒 ( ) ( ) ⎤
1 𝑐 1 𝑐
+ 2
𝑎𝑛 + 𝑖𝑏𝑛 𝑒 ⎥
𝑛=1⎣ ⎦
∞ 𝑖𝑛π𝑥 𝑖𝑛π𝑥
𝑐
− 𝑐
= 𝑐0 + ∑ 𝑐𝑛𝑒 + 𝑐−𝑛𝑒
𝑛=1

𝑊ℎ𝑒𝑟𝑒, 𝑐0 = 𝑎0 , 𝑐𝑛 =
1
2 (𝑎𝑛 − 𝑖𝑏𝑛) 𝑎𝑛𝑑 𝑐−𝑛 = 12 (𝑎𝑛 + 𝑖𝑏𝑛)
Equation no (2) is known as the complex form of the Fourier series which can also be written as

17
 ∞ 𝑖𝑛π𝑥
𝑐
𝑓(𝑥) = ∫ 𝑐𝑛𝑒 − 𝑐<𝑥< 𝑐
𝑛=−∞

𝑐 𝑖𝑛π𝑥
1 − 𝑐
𝑊ℎ𝑒𝑟𝑒 𝑐𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑑𝑥, 𝑎𝑛𝑑 𝑛 = 0, ±1, ±2, ………………………..
−𝑐

3.13 Parseval’s Formula

One of the most important properties of Fourier series is Parseval’s Formula or the completeness
relation which gives a relation between the average of the square (or absolute square) of the
function 𝑓(𝑥) and the co-efficients in Fourier series of 𝑓(𝑥).

(A). Particular case

Let 𝑓(𝑥) be a real-valued function of period 2π whose Fourier is

∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛 𝑠𝑖𝑛𝑛𝑥)
𝑛=1

π π π
1 1 1
𝑊ℎ𝑒𝑟𝑒, 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥, 𝑎𝑛𝑑 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π −π −π

Now the average of the square of 𝑓(𝑥) over (− π, π) 𝑖𝑠

π
1 2
2π
∫ [𝑓(𝑥)] 𝑑𝑥
−π

Thus

π π ∞ 2
1 2 1 ⎡ ⎤
∫ [𝑓(𝑥)] 𝑑𝑥 = ∫ ⎢𝑎0 + ∑ (𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥)⎥ 𝑑𝑥
2π 2π ⎢
−π ⎣
⎥
−π 𝑛=1 ⎦

π ∞ π ∞ π
1 2 1 2 2 1 2 2
= 2π
∫ 𝑎0 𝑑𝑥 + 2π
∑ ∫ 𝑎𝑛 𝑐𝑜𝑠 𝑛𝑥𝑑𝑥 + 2π
∑ ∫ 𝑏𝑛 𝑠𝑖𝑛 𝑛𝑥𝑑𝑥
−π 𝑛=1 −π 𝑛=1 −π

+ others terms (which vanish when average is taken)

18
 ∞ π
1 2 π 1 1 2 1
= 2π
. 𝑎0 . [𝑥]−π + 2π
∑ 2
𝑎𝑛 ∫ (1 + 𝑐𝑜𝑠2𝑛𝑥)𝑑𝑥 + 2π
𝑛=1 −π

∞ π
1 1 2
+ 2π
∑ 2
𝑏𝑛 ∫ (1 − 𝑐𝑜𝑠2𝑛𝑥)𝑑𝑥 + 0
𝑛=1 −π

∞ ∞ ∞ ∞
2 1 1 2 π 1 1 2 π 2 1 1 2 1 1 2
= 𝑎0 + 2π
∑ 2
𝑎𝑛 [𝑥]−π + 2π
∑ 2
𝑏𝑛 [𝑥]−π = 𝑎0 + 2π
∑ 2
𝑎𝑛 . 2π + 2π
∑ 2
𝑏𝑛 . 2π
𝑛=1 𝑛=1 𝑛=1 𝑛=1

∞
2 1 2 2
= 𝑎0 + 2
∑ (𝑎𝑛 + 𝑏𝑛 )
𝑛=1

Hence
π ∞
1 2 2 1 2 2
2π
∫ [𝑓(𝑥)] 𝑑𝑥 = 𝑎0 + 2
∑ (𝑎𝑛 + 𝑏𝑛 )
−π 𝑛=1

This is one form of Parseval’s Formula. One can easily verify that this formula is unchanged if
𝑓(𝑥) has period 2𝑐 in place of 2π and its square is averaged over any period of length 2𝑐. Then
we have
𝑐 ∞
1 2 2 1 2 2
2𝑐
∫ [𝑓(𝑥)] 𝑑𝑥 = 𝑎0 + 2
∑ (𝑎𝑛 + 𝑏𝑛 )
−𝑐 𝑛=1

Let ∫(𝑥) be a complex valued function of period 2π,

Fourier series is
∞
𝑖𝑛𝑥
∑ 𝑐𝑛𝑒
𝑛=−∞

π
1 −𝑖𝑛𝑥
𝑊ℎ𝑒𝑟𝑒 𝑐𝑛 = 2π
∫ 𝑓(𝑥)𝑒 (𝑛 = 0, ±1, ±2, ±3………)
−π

Then the average square of 𝑓(𝑥)

π ∞
2 2
1
2𝑐
∫ [𝑓(𝑥)] 𝑑𝑥 = ∑
−π 𝑛=−∞
|𝑐𝑛|

19
(B) General case.
A general form of Parsevals Formula states that if 𝑓(𝑥) and 𝑔(𝑥) are two real-valued functions
of period 2π whose Fourier series are
∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥)
𝑛=1

and
∞
' ' '
𝑔(𝑥) = 𝑎0 + ∑ (𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥)
𝑛=1

respectively.
π
1
𝑊ℎ𝑒𝑟𝑒 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π

π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π

' ' '

Similarly, 𝑎0 , 𝑎𝑛 , 𝑏𝑛 are defined in terms of 𝑔. Then

We have
π ∞
1 ' 1 ' '
2π
∫ 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = 𝑎0𝑎0 + 2
∑ (𝑎𝑛𝑎𝑛 + 𝑏𝑛𝑏𝑛)
−π 𝑛=1

When 𝑓(𝑥) and 𝑔(𝑥) are two complex valued function of period 2π, whose Fourier series are
∞ ∞
𝑖𝑛𝑥 ' 𝑖𝑛𝑥
𝑓(𝑥) = ∑ 𝑐𝑛𝑒 𝑎𝑛𝑑 𝑔(𝑥) = ∑ 𝑐𝑛𝑒
𝑛=−∞ 𝑛=−∞

respectively.

20
 π
1 −𝑖𝑛𝑥
𝑊ℎ𝑒𝑟𝑒, 𝑐𝑛 = 2π
∫ 𝑓(𝑥)𝑒 𝑑𝑥 (𝑛 = 0, ±1, ±2, ±3………)
−π

π
' 1 −𝑖𝑛𝑥
𝑐𝑛 = 2π
∫ 𝑔(𝑥)𝑒 𝑑𝑥 (𝑛 = 0, ±1, ±2, ±3………)
−π

Then we have
π ∞
1 '
2π
∫ 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = ∑ 𝑐𝑛𝑐𝑛.
−π 𝑛=−∞

WORKEDOUT EXAMPLES
−𝑥
Example 1. Obtain the Fourier series for 𝑓(𝑥) = 𝑒 in the interval 0 < 𝑥 < 2π.

Solution: Let
𝑎0 ∞
−𝑥
𝑓(𝑥) = 𝑒 = 2 (
+ ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 (1)
𝑛=1
)
Then
π 2π
1 1 −𝑥
𝑎0= 2π ∫ 𝑓(𝑥)𝑑𝑥 = 2π
∫ 𝑒 𝑑𝑥
−π 0

π 2π
−𝑥 2π= 1 −2π −𝑥
=
1
π [− 𝑒 0 π ] (1 − 𝑒 )𝑎𝑛𝑑, 𝑎𝑛 =
1
π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 =
1
π
∫ 𝑒 𝑐𝑜𝑠𝑛𝑥𝑑𝑥
0 0

2π
1 −𝑥 2π 𝑛 −𝑥
= π
[− 𝑐𝑜𝑠𝑛𝑥. 𝑒 ]0 − π
∫ 𝑒 𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

2π
−𝑥 2π
2
−2π 𝑛 𝑛 −𝑥
= (1 − 𝑒 ) + π
[𝑠𝑖𝑛𝑛𝑥. 𝑒 ]0 − π
∫ 𝑒 𝑐𝑜𝑠𝑛𝑥𝑑𝑥
0

=
1
π (1 − 𝑒−2π) + 0 − 𝑛2𝑎𝑛
2
(
∴ 𝑛 + 1 𝑎𝑛 = ) 1
π (1 − 𝑒−2π)𝑜𝑟, 𝑎𝑛 =
1
π (1 − 𝑒−2π). 𝑛 1+1
2

Finally,
2π 2π
1 1 −𝑥
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 = π
∫ 𝑒 𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0 0

21
 2π 2π
−𝑥 −𝑥
=
1
π [ − 𝑒 𝑠𝑖𝑛𝑛𝑥 ]0 +
𝑛
π
∫ 𝑒 𝑐𝑜𝑠𝑛𝑥𝑑𝑥
0

2π 2π
−𝑥 𝑛2 −𝑥
= 0+
𝑛
π [ − 𝑒 𝑐𝑜𝑠𝑛𝑥𝑑𝑥 ]0 − π
∫ 𝑒 𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

2
(
𝑜𝑟, 𝑛 + 1 𝑏𝑛 = ) 𝑛
π (1 − 𝑒−2π)
∴𝑏𝑛 =
1
π (1 − 𝑒−2π). 𝑛 𝑛+1 2

Now substituting the values of 𝑎0,𝑎𝑛𝑎𝑛𝑑 𝑏𝑛 in (1), we get

∞
−2π −2π
𝑓(𝑥) =
1
2π ( 1 −𝑒 ) + ∑ ⎡⎢ π 1 − 𝑒
1
( ) 𝑛 1+1 𝑐𝑜𝑠𝑛𝑥 + π1 (1 − 𝑒−2π). 𝑛 𝑛+1 . 𝑠𝑖𝑛𝑛𝑥⎤⎥⎦
2 2
𝑛=1⎣

∞ ∞
−2π ⎡1 ⎤
=
1
π ( 1 −𝑒 ) ⎢2 + ∑
⎢
1
2
𝑛 +1
. 𝑐𝑜𝑠𝑛𝑥 + ∑
𝑛 +1
𝑛
2. 𝑠𝑖𝑛𝑛𝑥⎥
⎥
⎣ 𝑛=1 𝑛=1 ⎦

=
1
π (1 − 𝑒−2π)
1
[2 + ( 1
2
𝑐𝑜𝑠𝑥 +
1
5
𝑐𝑜𝑠2𝑥 +
1
10
𝑐𝑜𝑠3𝑥 + … + ) ( 1
2
𝑠𝑖𝑛𝑥 +
2
5
𝑠𝑖𝑛2𝑥 +
3
10 )
𝑠𝑖𝑛3𝑥 + … ]

𝑥
Example 2. Obtain the Fourier series for 𝑓(𝑥) = 𝑒 in the interval − π < 𝑥 < π.

Solution: Let
𝑎0 ∞
𝑥
𝑓(𝑥) = 𝑒 = 2 (
+ ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 (1)
𝑛=1
)
Then
π π
1 1 𝑥
𝑎0= π
∫ 𝑓(𝑥)𝑑𝑥 = π
∫ 𝑒 𝑑𝑥
−π −π

=
1
π (𝑒π − 𝑒π)
π π

( )
π −π
2 𝑒 −𝑒 2𝑆𝑖𝑛ℎπ 1 1 𝑥
𝑎𝑜 = π 2
= π
𝑎𝑛𝑑, 𝑎𝑛 = π
∫ 𝑓(𝑥)𝐶𝑜𝑠𝑛𝑥𝑑𝑥 = π
∫ 𝑒 𝐶𝑜𝑠𝑛𝑥𝑑𝑥
−π −π

2 π
π −π 𝑥 𝑥
=
1
π [ 𝑒 −𝑒 ] 𝐶𝑜𝑠𝑛𝑥 +
𝑛
π
[𝑒 𝑆𝑖𝑛𝑛𝑥]π − π −
𝑛
π
∫ 𝑒 𝐶𝑜𝑠𝑛𝑥𝑑𝑥
−π

22
 ( )(− 1)
π −π
𝑛 2 2 𝑛 2𝑆𝑖𝑛ℎπ
=
2
π
𝑒 −𝑒
2
+ 0 − 𝑛 𝑎𝑛𝑜𝑟, 𝑛 + 1 𝑎𝑛 = (− 1) ( ) π

𝑛
2𝑆𝑖𝑛ℎπ (−1)
𝑎𝑛 = π 2
(𝑛 +1)
Finally,
π π
1 1 𝑥
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑆𝑖𝑛𝑛𝑥𝑑𝑥 = π
∫ 𝑒 𝑆𝑖𝑛𝑛𝑥𝑑𝑥
−π −π

π 2 π
1 𝑥 𝑛 𝑥 𝑛 𝑥 𝑛 𝑥
= π
[𝑒 𝑠𝑖𝑛𝑛𝑥]π − π − π
∫ 𝑒 𝐶𝑜𝑠𝑛𝑥𝑑𝑥 = 0 − π
[𝑒 𝑐𝑜𝑠𝑛𝑥]π − π − π
∫ 𝑒 𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π −π

𝑛 2𝑠𝑖𝑛ℎπ 2 2 𝑛 2𝑆𝑖𝑛ℎπ
=− 𝑛(− 1) π (
− 𝑛 𝑏𝑛𝑜𝑟, 𝑛 + 1 𝑏𝑛 =− 𝑛. (− 1) ) π

𝑛 2𝑆𝑖𝑛ℎπ −𝑛
𝑏𝑛 = (− 1) π (𝑛2+1)
Now substituting the values of 𝑎𝑜, 𝑎𝑛 𝑎𝑛𝑑 𝑏𝑛 in (1) we get

∞ 𝑛
𝑛
. 2𝑠𝑖𝑛ℎπ + ∑ ⎡⎢ π 𝑆𝑖𝑛𝑛𝑥⎤⎥
1 2𝑆𝑖𝑛ℎπ (−1) 2𝑆𝑖𝑛ℎπ 𝑛
𝑓(𝑥) = 2π 2 𝐶𝑜𝑠𝑛𝑥 − (− 1) π 2
𝑛=1⎣ (𝑛 +1) (𝑛 +1) ⎦

=
2𝑠𝑖𝑛ℎπ
π ⎣
1
⎡2 + − ( 1
2
𝑐𝑜𝑠𝑥 +
1
5
𝑐𝑜𝑠2𝑥 −
1
10 ) 1
𝑐𝑜𝑠3𝑥 +... + ( 2 𝑠𝑖𝑛𝑥 −
2
5
𝑠𝑖𝑛2𝑥 +
3
20
𝑠𝑖𝑛3𝑥 − …)⎤
⎦

Example 3: Express 𝑓(𝑥) = 𝑥 as a half range sine series in the interval 0 < 𝑥 < 2.

Solution: The graph of 𝑓(𝑥) = 𝑥 in 0 < 𝑥 < 2 is the line OA. Let us extend the function 𝑓(𝑥)
in the interval − 2 < 𝑥 < 0 . So that the new function is symmetrical about the origin and
therefore represents an odd function in the interval (− 2, 2).

23
Hence the Fourier series for 𝑓(𝑥) will have only the sine terms given by
∞
𝑛π𝑥
𝑓(𝑥) = ∑ 𝑏𝑛𝑠𝑖𝑛 2
(𝑎0 = 0, 𝑎𝑛 = 0)
𝑛=1

Where
𝑐
2 𝑛π𝑥
𝑏𝑛 = 𝑐
∫ 𝑓(𝑥)𝑠𝑖𝑛 2
𝑑𝑥
0
2 2
2 𝑛π𝑥 2 𝑛π𝑥 2 𝑛π𝑥 4 4 𝑛π𝑥
= ∫ 𝑥𝑠𝑖𝑛 𝑑𝑥 = ⎡− 𝑥 𝑐𝑜𝑠 ⎤2 0 + ∫ 𝑐𝑜𝑠 =− 𝑐𝑜𝑠𝑛π + 0 + [𝑠𝑖𝑛 ]2 0
2
0
2 ⎣ 𝑛π 2 ⎦ 𝑛π
0
2 𝑛𝑥 2 2
𝑛π 2

𝑛
4 𝑛 −4(−1)
=− 𝑛π
(− 1) + 0 = 𝑛π

Hence the Fourier sine series for 𝑓(𝑥) = 𝑥 over the half

Range (0, 2) is
∞ 𝑛
−4(−1) 𝑛π𝑥
𝑓(𝑥) = ∑ 𝑛π
𝑠𝑖𝑛 2
𝑛=1

∞ 𝑛+1
4 (−1) 𝑛π𝑥
= π
∑ 𝑛
𝑠𝑖𝑛 2
𝑛=1

4 π𝑥 1 2π𝑥 1 3π𝑥 1 4π𝑥

= π
[𝑠𝑖𝑛 2
− 2
𝑠𝑖𝑛 2
+ 3
𝑠𝑖𝑛 2
− 4
𝑠𝑖𝑛 2
+...]

24
Example 04: Express 𝑓(𝑥) = 𝑥 as a half range cosine series in the interval 0 < 𝑥 < 2.

Solution:
The graph of the function 𝑓(𝑥) = 𝑥 in the interval 0 < 𝑥 < 2 is the line OA. Let us
extend the function f(x) in the interval 0 < 𝑥 < 2.
'
(Shown by the line 𝑂𝐵 ) so that the new function is symmetrical about the y-axis and therefore,
represent an even function in the interval 0 < 𝑥 < 2. Hence the fourier series for f(x) over the
full period (-2,2) will have only cosine terms given by,
𝑎0 ∞
𝑛π𝑥
𝑓(𝑥) = 2
+ ∑ 𝑎𝑛𝑐𝑜𝑠 2
(𝑏𝑛 = 0)
𝑛=1

𝑐 2
2 2
𝑊ℎ𝑒𝑟𝑒 𝑎0 = 𝑐
∫ 𝑓(𝑥)𝑑𝑥 = 2
∫ 𝑥𝑑𝑥 = 2
0 0

𝑐 2 2
2 𝑛π𝑥 2 𝑛π𝑥 2𝑥 𝑛π𝑥 2 2 𝑛π𝑥
𝑎𝑛𝑑, 𝑎𝑛 = 𝑐
∫ 𝑓(𝑥)𝑐𝑜𝑠 2
𝑑𝑥 = 2
∫ 𝑥𝑐𝑜𝑠 2
𝑑𝑥 = [ 𝑛π
𝑠𝑖𝑛 2
] − 𝑛π
∫ 𝑐𝑜𝑠 2
𝑑𝑥
0 0 0 0

4 𝑛π𝑥 2 4 4 𝑛
= 0+ 2 2 [𝑠𝑖𝑛 2
] = 2 2 (cos 𝑐𝑜𝑠 𝑛π − 1 ) = 2 2 {(− 1) − 1}
𝑛π 0 𝑛π 𝑛π

∞
𝑛
∑ ⎡⎢ 2 2 (− 1) − 1 cos 𝑐𝑜𝑠
{ } ⎤
2 4 𝑛π𝑥
𝑇ℎ𝑢𝑠 𝑓(𝑥) = 2 2 ⎥
𝑛=1⎣ 𝑛 π ⎦

= 1 + ⎡⎢− +...⎤⎥
8 π𝑥 8 3π𝑥 8 5π𝑥
2 2 𝑐𝑜𝑠 2
+0− 2 2 𝑐𝑜𝑠 2
+0− 2 2 𝑐𝑜𝑠 2
⎣ 1π 3π 5π ⎦

= 1−
8 ⎡ 1 𝑐𝑜𝑠 π𝑥
+
1
𝑐𝑜𝑠
3π𝑥
+
1
𝑐𝑜𝑠
5π𝑥
+...⎤⎥
2 ⎢ 12 2 2 2 2 2
π ⎣ 3 5 ⎦

2
Example 05: Find the Fourier series expansion of the function 𝑓(𝑥) = 𝑥 in the interval
− π ≤ 𝑥 ≤ π and hence evaluate
∞
1
∑ 2
𝑛=1 𝑛

Solution:
By definition we have
∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 )
𝑛=1

25
 π π π
1 1 1
𝑤ℎ𝑒𝑟𝑒 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 𝑎𝑛𝑑 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π −π −π

π π 2
2 3 3
𝑁𝑜𝑤 𝑎0 =
1
2π
∫ 𝑓(𝑥)𝑑𝑥 =
1
2π
∫ 𝑥 𝑑𝑥 =
1
6π { π − − π ( )} =
π
3
−π −π

2 2
Again since 𝑓(𝑥) = 𝑥 and 𝑓(− 𝑥) = − 𝑥 ( ) = 𝑥2 = 𝑓(𝑥)
2
∴𝑓(− 𝑥) = 𝑓(𝑥). Thus 𝑓(𝑥) = 𝑥 is an even function and so sine terms will vanish i.e. 𝑏𝑛 = 0
π π π
1 1 2 2
𝐹𝑖𝑛𝑎𝑙𝑙𝑦, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 = π
∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 = 0 − 𝑛π
∫ 𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π −π −π

2 2 π 2
= 2 [π cos 𝑐𝑜𝑠 𝑛π − (− π) cos 𝑐𝑜𝑠 𝑛(− π) ] − 3 [sin 𝑠𝑖𝑛 𝑛𝑥] −π = 2 [π cos 𝑐𝑜𝑠 𝑛π − π cos 𝑐𝑜𝑠 𝑛π ]
𝑛π 𝑛π 𝑛π

4 4 𝑛 𝑛
= 2 cos 𝑐𝑜𝑠 𝑛π = 2 (− 1) . Since cos 𝑐𝑜𝑠 𝑛π = (− 1)
𝑛 𝑛
Now putting the value of 𝑎0, 𝑎𝑛 𝑎𝑛𝑑 𝑏𝑛 in (1) we get
∞

{ }
2 𝑛
2 π 4(−1)
𝑓(𝑥) = 𝑥 = 3
+ ∑ 2 cos 𝑐𝑜𝑠 𝑛𝑥 + 0
𝑛=1 𝑛

2
π −cos𝑐𝑜𝑠 𝑥 cos𝑐𝑜𝑠 2𝑥 cos𝑐𝑜𝑠 3𝑥 cos𝑐𝑜𝑠 4𝑥
𝑜𝑟, 𝑓(𝑥) = 3
+ 4{ 2 + 2 − 2 + 2 − …}
1 2 3 4

2
π cos𝑐𝑜𝑠 𝑥 cos𝑐𝑜𝑠 2𝑥 cos𝑐𝑜𝑠 3𝑥 cos𝑐𝑜𝑠 4𝑥
= 3
− 4{ 2 − 2 + 2 − 2 +...}
1 2 3 4

When = 0 , we have
∞

{ }
2 𝑛
π 4(−1)
0= 3
∑ 2
𝑛=1 𝑛

2
π 1 1 1 1
= 3
− 4( 2 − 2 + 2 − 2 +...)
1 2 3 4

{( ) ( )}
2
π 1 1 1 1 1 1 1 1
= 3
− 4 2 + 2 + 2 + 2 +... − 2 2 + 2 + 2 + 2 +...
1 2 3 4 2 4 6 8

π
2
⎰∞ 1 1
∞
1 ⎱
= − 4 ∑ − ∑
3 ⎱𝑛=1 2
𝑛 2
𝑛=1 𝑛
2
⎰

2 ∞ ∞ 2
π 1 1 π
= 3
− 2 ∑ 2 𝑜𝑟, 2 ∑ 2 = 3
𝑛=1 𝑛 𝑛=1 𝑛

26
 ∞ 2
1 π
∴ ∑ 2 = 6
𝑛=1 𝑛

Example 06:
(a) Obtain the Fourier series of the function 𝑓(𝑥){0, − π ≤ 𝑥 ≤ 0 1, 0 ≤ 𝑥 ≤ π
(b) Verify the result found in 6(a) by assuming the complex form of the Fourier series.

Solution of (a):
By definition we have

∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 )
𝑛=1

π π π
1 1 1
𝑤ℎ𝑒𝑟𝑒 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 𝑎𝑛𝑑 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π −π −π

π 0 π
1 1 ⎡ ⎤
𝑁𝑜𝑤 𝑎0 = ∫ 𝑓(𝑥)𝑑𝑥 = ⎢ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥⎥
2π 2π ⎢−π ⎥
−π ⎣ 0 ⎦
0 π
1 ⎡ ⎤ 1 1
= ⎢ ∫ 0𝑑𝑥 + ∫ 1𝑑𝑥⎥ = [0 + π] =
2π ⎢−π ⎥ 2π 2
⎣ 0 ⎦
π
1
Again, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 (𝑛 ≠ 0)
−π
0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + ∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦

0 π
1 ⎡ ⎤
= ⎢ ∫ 0 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + ∫ 1 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
1 π
= 0+ π𝑛
𝑛𝑥] 0 = 0 + 0 = 0

π
1
𝐹𝑖𝑛𝑎𝑙𝑙𝑦, 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π

27
 0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + ∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + ∫ 1 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
1 π
= π
𝑛𝑥] 0

1
=− 𝑛π
[cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0 ]

1 𝑛
=− 𝑛π
[(− 1) − 1]

=
1
𝑛π [1 − (− 1)𝑛] = {𝑜 𝑤ℎ𝑒𝑛 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛 2
𝑛π
𝑤ℎ𝑒𝑛 𝑛 𝑖𝑠 𝑜𝑑𝑑

Now putting the value of 𝑎0, 𝑎𝑛 𝑎𝑛𝑑 𝑏𝑛 in (1) we get

∞
1
𝑓(𝑥) = 2
+ 0 + ∑ 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥
𝑛=1

1
𝑜𝑟, 𝑓(𝑥) = 2
+ 𝑏1 sin 𝑠𝑖𝑛 𝑥 + 𝑏2 sin 𝑠𝑖𝑛 2𝑥 + 𝑏3 sin 𝑠𝑖𝑛 3𝑥 + 𝑏4 sin 𝑠𝑖𝑛 4𝑥 + 𝑏5 sin 𝑠𝑖𝑛 5𝑥 +...

1 2 2 2
= 2
+ π
sin 𝑠𝑖𝑛 𝑥 + 0 + 3π
sin 𝑠𝑖𝑛 3𝑥 + 5π
sin 𝑠𝑖𝑛 5𝑥 +...

1 2 1 1
= 2
+ π
(sin 𝑠𝑖𝑛 𝑥]] + 3
sin 𝑠𝑖𝑛 3𝑥 + 5
sin 𝑠𝑖𝑛 5𝑥 +...)

Solution of (b) : Here we have to expand the function f(x) in the complex Fourier series.
By definition we have,
∞ 𝑖𝑛π𝑥
𝑐
𝑓(𝑥) = ∑ 𝑐𝑛𝑒 , − 𝑐<𝑥<𝑐
𝑛=−∞

𝑐 𝑖𝑛π𝑥
1 𝑐
𝑊ℎ𝑒𝑟𝑒, 𝑐𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑑𝑥 , 𝑛 = 0, ± 1, ± 2, ± 3, ...
−𝑐

Here in our given problem 𝑐 = π

28
 ∞
𝑖𝑛π𝑥
∴𝑓(𝑥) = ∑ 𝑐𝑛𝑒 , −π <𝑥<π
𝑛=−∞

π
1 −𝑖𝑛𝑥
𝑊ℎ𝑒𝑟𝑒, 𝑐𝑛 = 2π
∫ 𝑓(𝑥)𝑒 𝑑𝑥 , 𝑛 = 0, ± 1, ± 2, ± 3, ...
−π

π
1
𝑐0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥 ⎥
2π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤ 1
= ⎢ ∫ 0 𝑑𝑥 + ∫ 1 𝑑𝑥 ⎥ =
2π ⎢−π ⎥ 2
⎣ 0 ⎦
π
1 −𝑖𝑛𝑥
𝐴𝑛𝑑 𝑐𝑛 = 2π
∫ 𝑓(𝑥)𝑒 𝑑𝑥 (𝑛 ≠ 0)
−π

0 π
1 ⎡ −𝑖𝑛𝑥 −𝑖𝑛𝑥 ⎤
= ⎢∫ 𝑒 𝑑𝑥 + ∫ 𝑒 𝑑𝑥 ⎥
2π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ −𝑖𝑛𝑥 −𝑖𝑛𝑥 ⎤
= ⎢∫ 0𝑒 𝑑𝑥 + ∫ 1 𝑒 𝑑𝑥 ⎥
2π ⎢−π ⎥
⎣ 0 ⎦
1 −𝑖𝑛π 0
= −2π𝑖𝑛
[𝑒 −𝑒 ]

1
= −2π𝑖𝑛
[cos 𝑐𝑜𝑠 𝑛π − 𝑖𝑠𝑖𝑛 𝑛π − 1]

1
= { 𝑛π𝑖 𝑤ℎ𝑒𝑛 𝑛 =, ± 1, ± 3, ± 5, ...

Thus

( ) ( )
𝑖𝑥 𝑖3𝑥 𝑖5𝑥 −𝑖𝑥 −𝑖3𝑥 −𝑖5𝑥
1 1 𝑒 𝑒 𝑒 1 𝑒 𝑒 𝑒
𝑓(𝑥) = 2
+ π𝑖 1
+ 0+ 3
+0 + 5
+... + π𝑖 −1
+ 0+ −3
+ 0 + −5
+...
1
=⎡2 +
⎣
1
π𝑖 (𝑒𝑖𝑥 − 𝑒−𝑖𝑥) + 13 (𝑒3𝑖𝑥 − 𝑒−3𝑖𝑥) + 15 (𝑒5𝑖𝑥 − 𝑒−5𝑖𝑥) +...⎤⎦
1 1 1 1
= + ⎡2𝑖 sin 𝑠𝑖𝑛 𝑥 + 2𝑖 sin 𝑠𝑖𝑛 3𝑥 + 2𝑖 sin 𝑠𝑖𝑛 5𝑥 +... ⎤
2 π𝑖 ⎣ 3 5 ⎦
1 2 1 1
= + ⎡sin 𝑠𝑖𝑛 𝑥 + sin 𝑠𝑖𝑛 3𝑥 + sin 𝑠𝑖𝑛 5𝑥 +... ⎤
2 π ⎣ 3 5 ⎦

Which is same as in 6(a).

29
Example: 7.Find the Fourier series expansion of the function,
𝑓(𝑥) = {0, − π < 𝑥 ≤ 0 𝑥, 0 < 𝑥 ≤ π
Hence evaluate
∞
1
∑ 2 .
𝑛=1 (2𝑛−1)

Solution: By the definition of fourier series

∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 )…(1)
𝑛=1

Where,
π π
1 1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
–π –π

And
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥.
–π

Now
π
1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
–π

0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥⎥
2π ⎢–π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0. 𝑑𝑥 + ∫ 𝑥 𝑑𝑥⎥
2π ⎢−π ⎥
⎣ 0 ⎦
2 π
1 ⎡𝑥 ⎤
= 0+ 2π ⎢2⎥
⎣ ⎦0

=
1
4π (π2 − 0) = π
4

Again
π
1
𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥(𝑛 ≠ 0)
−π

30
 0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + ∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 ⎥
π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0. cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + ∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
π
1 0 1
= 0+ π𝑛
[𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]π − π𝑛
∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
0

1 π
= 0+ 2 [cos 𝑐𝑜𝑠 𝑛𝑥 ]0
π𝑛

=
π𝑛
1
2 [cos 𝑐𝑜𝑠 𝑛π − 𝑐𝑜𝑠0] =
1
π𝑛
2 [(− 1)𝑛 − 1]
𝑛
[
Since cos 𝑐𝑜𝑠 𝑛π = (− 1) ]
Finally,
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥.
−π

0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + ∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0. sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + ∫ 𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
π
1 π 1
= 0− π𝑛
[𝑥𝑐𝑜𝑠 𝑛𝑥]0 + π𝑛
∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

1 1 π
=− π𝑛
(π cos 𝑐𝑜𝑠 𝑛π − 0 ) + 2 [sin 𝑠𝑖𝑛 𝑛𝑥 ]0
π𝑛

1
=− 𝑛
cos 𝑐𝑜𝑠 𝑛π + 0 since sin 𝑠𝑖𝑛 𝑛π = 0, sin 𝑠𝑖𝑛 0 = 0
𝑛+1
1 𝑛 (−1)
=− 𝑛
. (− 1) = 𝑛
.

Now putting the values of 𝑎0, 𝑎𝑛, 𝑎𝑛𝑑 𝑏𝑛𝑖𝑛 (1), we get
∞ 𝑛+1
𝑛
𝑓(𝑥) =
π
4
+ ∑[
1
2 { (− 1) − 1 cos 𝑐𝑜𝑠 𝑛𝑥 + } (−1)
𝑛
sin 𝑠𝑖𝑛 𝑛𝑥 ]
𝑛=1 π𝑛

31
=
π
4
+ ⎡⎢ −
⎣ ( π1
2
2 cos 𝑐𝑜𝑠 𝑥 + 0 −
2
π3
2 cos 𝑐𝑜𝑠 3𝑥 + 0 −
π5
2
2 . cos 𝑐𝑜𝑠 5𝑥 + 0 − ……… + ( ) sin𝑠𝑖𝑛 𝑥
1
−
sin𝑠𝑖

=
π
4
−
2
π ( cos𝑐𝑜𝑠 𝑥
1
2 +
cos𝑐𝑜𝑠 3𝑥
3
2 +
cos𝑐𝑜𝑠 5𝑥
5
2 + …… + ) ( sin𝑠𝑖𝑛 𝑥
1
−
sin𝑠𝑖𝑛 2𝑥
2
+
sin𝑠𝑖𝑛 3𝑥
3
−
sin𝑠𝑖𝑛 4𝑥
4
+ …… …… )
Putting, 𝑥 = 0 in the above equation (2), we get
0=
π
4
−
2
π ( cos𝑐𝑜𝑠 0
1
2 +
cos𝑐𝑜𝑠 0
2
3
+
cos𝑐𝑜𝑠 0
5
2 + …… + ) ( sin𝑠𝑖𝑛 0
1
−
sin𝑠𝑖𝑛 0
2
+
sin𝑠𝑖𝑛 0
3
−
sin𝑠𝑖𝑛 0
5
+ …… )
⇒0 =
π
4
−
2
π ( 1
1
2 +
1
3
2 +
1
5
2 + …… + 0 )
⇒
2
π { 1
1
2 +
1
3
2 +
1
5
2 + ...+
1
(2𝑛−1)
2 + ... }= π
4

2
1 1 1 1 π
⇒ 2 + 2 + 2 + ...+ 2 + ...= 8
1 3 5 (2𝑛−1)

∞ 2
1 π
⇒ ∑ 2 = 8
𝑛=1 (2𝑛−1)

Example 8. Find the Fourier series of the function of period –π < 𝑥 < π as follows:
𝑓(𝑥) = {𝑜 𝑤ℎ𝑒𝑛 − π < 𝑥 ≤ 0 𝑠𝑖𝑛𝑥 𝑤ℎ𝑒𝑛 0 < 𝑥 ≤ π

Solution:

By definition of the Fourier series

We have
∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛 𝑛𝑥)………………(1)
𝑛=1

Where
π π
1 1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑛𝑥𝑑𝑥
−π −π

And
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑛𝑥𝑑𝑥
−π

32
Now
π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑑𝑥
−π

0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥⎥
2π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0. 𝑑𝑥 + ∫ 𝑠𝑖𝑛𝑥𝑑𝑥⎥
2π ⎢−π ⎥
⎣ 0 ⎦

1 π
=0+ 2π
[− 𝑐𝑜𝑠𝑥]0

1
=− 2
(𝑐𝑜𝑠π − 1)

1
=− 2π
(− 1 − 1)

1
= π

π
1
𝐴𝑔𝑎𝑖𝑛 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑛𝑥𝑑𝑥
−π

0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑑𝑥 + ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
⎢ ∫ 0. 𝑛𝑑𝑥 + ∫ 𝑠𝑖𝑛𝑥𝑐𝑜𝑠𝑛𝑥𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
π
1
= 0+ 2π
∫ 2𝑐𝑜𝑠𝑛𝑥𝑠𝑖𝑛𝑥𝑑𝑥
0

π
1
= 2π
∫[sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥 − sin 𝑠𝑖𝑛 (𝑛 − 1) 𝑥]𝑑𝑥
0

1 1 π 1 π
= ⎡⎢− [cos 𝑐𝑜𝑠 (𝑛 + 1) 𝑥]0 + [cos 𝑐𝑜𝑠 (𝑛 − 1) 𝑥]0 ⎤⎥
2π ⎣ 𝑛+1 𝑛−1 ⎦
1 1 1
= ⎡− {cos 𝑐𝑜𝑠 (𝑛 + 1) π − 𝑐𝑜𝑠0} + {cos 𝑐𝑜𝑠 (𝑛 − 1) π − 𝑐𝑜𝑠0}⎤
2π ⎣ 𝑛+1 𝑛−1 ⎦
1 1−𝑐𝑜𝑠⁡(𝑛π+π) cos𝑐𝑜𝑠 (𝑛π−π) −1
= ⎡ + ⎤
2π ⎣ 𝑛+1 𝑛−1 ⎦
1 1+𝑐𝑜𝑠𝑛π 1+𝑐𝑜𝑠𝑛π
= ⎡ − ⎤
2π ⎣ 𝑛+1 𝑛−1 ⎦

33
=
1 ⎡ (𝑛−1−𝑛−1)(1+𝑐𝑜𝑠𝑛π) ⎤
2π ⎢ 2 ⎥
⎣ 𝑛 −1 ⎦

=
1 ⎡ −2(1+𝑐𝑜𝑠𝑛π) ⎤
2π ⎢ 2 ⎥
⎣ 𝑛 −1 ⎦
(1+𝑐𝑜𝑠𝑛π)
=− 2
π(𝑛 −1)

𝑛
{1+(−1) }
= 2 (𝑛 ≠ 1)
π(𝑛 −1)

(1+1) 2
When 𝑛 = 2, 𝑎2 =− 2 =− 3π
π(2 −1)

When 𝑛 = 3, 𝑎3 = 0

(1+1) 2
When 𝑛 = 4, 𝑎4 =− 2 =− 15π
π(4 −1)

when 𝑛 = 5, 𝑎5 = 0

(1+1) 2
when 𝑛 = 6, 𝑎4 =− 2 =− 35π
and so on
π(6 −1)

but when
π
1
𝑛 = 1, 𝑎1 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑥𝑑𝑥
−π

π 2 π
1 1 ⎡ (𝑠𝑖𝑛𝑥) ⎤ = 0
= π
∫ 𝑠𝑖𝑛𝑐𝑜𝑠𝑥𝑑𝑥 π ⎢ 2 ⎥
0 ⎣ ⎦0

π
1
𝐹𝑖𝑛𝑎𝑙𝑙𝑦, 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑛𝑥𝑑𝑥.
−π

When
π
1
𝑛 = 1, 𝑏1 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑥𝑑𝑥
−π

0 π
1 ⎡ ⎤
𝑂𝑟, 𝑏1 = ⎢ ∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑥𝑑𝑥 + ∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑥𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0. 𝑠𝑖𝑛𝑥𝑑𝑥 + ∫ 𝑠𝑖𝑛𝑥𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦

34
 π
1 2
= 0+ 2π
∫ 2𝑠𝑖𝑛 𝑥𝑑𝑥.
0

π
1
= 2π
∫(1 − 𝑐𝑜𝑠2𝑥)𝑑𝑥
0

1 1 π
= ⎡𝑥 − 𝑠𝑖𝑛2𝑥⎤
2π ⎣ 2 ⎦0

1 1
= 2π
(π − 0) = 2

π
1
𝐴𝑙𝑠𝑜, 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑛𝑥𝑑𝑥 (𝑛 ≠ 1, 𝑛 = 2, 3, 4…….)
−π

0 π
1 ⎡ ⎤
= ⎢ ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
0 π
1 ⎡ ⎤
= ⎢ ∫ 0. 𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + ∫ 𝑠𝑖𝑛𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥⎥
π ⎢−π ⎥
⎣ 0 ⎦
π
1
= 0+ 2π
∫ 2𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

π
1
= 2π
∫[cos 𝑐𝑜𝑠 (𝑛 − 1) 𝑥 − cos 𝑐𝑜𝑠 (𝑛 + 1) 𝑥]𝑑𝑥
0

1 1 1 π
= ⎡ 𝑛−1 sin 𝑠𝑖𝑛 (𝑛 − 1) 𝑥 − sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥⎤
2π ⎣ 𝑛+1 ⎦0

1
= 2π
. 0 = 0 𝑓𝑜𝑟 (𝑛 ≠ 1, 𝑛 = 2, 3, 4).

Thus 𝑓(𝑥) =
1
π
+ − ( 2
3π
𝑐𝑜𝑠2𝑥 −
2
15π
𝑐𝑜𝑠4𝑥 −
2
35π )
𝑐𝑜𝑠6𝑥 − ………… +
1
2
𝑠𝑖𝑛𝑥 + 0

=
1
π
+
1
2
𝑠𝑖𝑛𝑥 −
2
π ( 𝑐𝑜𝑠2𝑥
3
+
𝑐𝑜𝑠4𝑥
15
+
𝑐𝑜𝑠6𝑥
35
+… )

35
Example 9.Find the Fourier series Expansion of the function f(x)=|𝑥| in the interval[-π,π]
.Hence evaluate the sum
∞
1
∑ 2
𝑛=0 (2𝑛+1)

Solution:

By definition of the Fourier series, we have

∞
(
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 ……………………………. (1)
𝑛=1
)
π π
1 1
𝑊ℎ𝑒𝑟𝑒, 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
−π −π

π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
−π

Now ,by definition |𝑥| = {𝑥, 𝑥 > 0 − 𝑥, 𝑥 < 0 }

Hence the given function f(x)= |𝑥| is an even function and for the even function
𝑏𝑛 = 0, (𝑛 = 1, 2, 3, ……) in the fouerier series expansion (1) of f(x) and

π π π 2 π
1 1 2 2 𝑥
𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥 = π
∫ |𝑥|𝑑𝑥 = π
∫(𝑥)𝑑𝑥 = π
[ 2
] = π
−π −π 0 0

π
1
𝐴𝑔𝑎𝑖𝑛, 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥(𝑛 ≠ 0)
−π

π
1
= π
∫ |𝑥|𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
−π

π
2
= π ∫|𝑥|𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
0

π
2
= π ∫ 𝑥𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
0

36
 π
2 π 2
= π𝑛 [𝑥𝑠𝑖𝑛𝑛𝑥]0 − 𝑛π
∫ 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
0

2 π
=0+ 2 [𝑐𝑜𝑠𝑛𝑥]0
π𝑛

2
= 2 [𝑐𝑜𝑠𝑛π − 𝑐𝑜𝑠0]
π𝑛

2 𝑛
= 2 {(− 1) − 1}
π𝑛

−4
= 2 𝑤ℎ𝑒𝑛 𝑛 = 1, 3, 5 ...
π𝑛

=0,when n=2,4,6….

Now, substituting the values of 𝑎0, 𝑎𝑛,𝑎𝑛𝑑 𝑏𝑛 in(1),we get

f(x)= 2 [
π
( −4
π1
2 + 0 −
4
π3
2 𝑐𝑜𝑠3𝑥 + 0 −
4
π5
2 )
𝑐𝑜𝑠5𝑥 +... + 0]

=2 −
π 4 ⎡ 𝑐𝑜𝑠𝑥 + 𝑐𝑜𝑠3𝑥
+
𝑐𝑜𝑠5𝑥
+....⎤⎥…………………(2)
π ⎢ 12 2 2
⎣ 3 5 ⎦
Putting x=0 in(2),we have

0= 2 −
π 4 ⎡ 𝑐𝑜𝑠0 + 𝑐𝑜𝑠0
+
𝑐𝑜𝑠0
+....⎤⎥
π ⎢ 12 2 2
⎣ 3 5 ⎦

=2 −
π 4 ⎡1 + 1
+
1
+....⎤⎥
π ⎢ 12 2 2
⎣ 3 5 ⎦
2
1 1 1 π π π
Or, 2 + 2 + 2 +.... = 2
* 4
= 8
1 3 5

∞ 2
1 π
Or, ∑ 2 = 8
.
𝑛=0 (1+2𝑛)

Note; the values of 𝑎0, 𝑎𝑛 can also be found in the following way:

π
1
𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥
−π

37
 π
1
= π
∫ |𝑥|𝑑𝑥
−π

0 π
1
= π
[ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥]
−π 0

0 π
1
= π
[ ∫ (− 𝑥)𝑑𝑥 + ∫(𝑥)𝑑𝑥]
−π 0

π π
1
= π
[∫ 𝑥𝑑𝑥 + ∫ 𝑥𝑑𝑥]
0 0

π
2
= π
[∫ 𝑥 𝑑𝑥]
0

2 π
2 𝑥
= π
[ 2
]
0

=π
π
𝑎𝑛 = ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
−π

0 π
1
= π
[ ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 + ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥 𝑑𝑥]
−π 0

0 π π π π
1 1 2 −4
= π
[ ∫ (− 𝑥) 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 + ∫ 𝑥 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥] = π
[∫ 𝑥 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥 + ∫ 𝑥 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥] = π
[∫ 𝑥 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥] 2 𝑤ℎ𝑒
−π 0 0 0 0 π𝑛

=0,when n=2,4,6….

2
Example 10: Find a series of sines and cosines of multiples of 𝑥 which will represent 𝑥 + 𝑥 in
the interval − π < 𝑥 < π .

38
 Hence deduce that
∞ 2
1 π
∑ 2 = 6
𝑛=1 𝑛

Solution: By definition of the Fourier series we have

𝑎𝑜 ∞
𝑓(𝑥) = 2
+ ∑ (𝑎𝑛𝐶𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑆𝑖𝑛𝑛𝑥) −−−−−−−−−−−− (1)
𝑛=1

π π π
1 1 1
𝑊ℎ𝑒𝑟𝑒 𝑎𝑜 = π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥)𝐶𝑜𝑠𝑛𝑥𝑑𝑥 (𝑛 ≠ 0)𝑎𝑛𝑑 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑆𝑖𝑛𝑛𝑥𝑑𝑥
−π −π −π

2
Now given in equation 𝑓(𝑥) = 𝑥 + 𝑥

Where
π π
2
𝑎𝑜 =
1
π
∫ 𝑓(𝑥)𝑑𝑥 =
1
π (
∫ 𝑥 + 𝑥 𝑑𝑥 )
−π −π

2 3 π 2
1 ⎡𝑥 + 𝑥 ⎤ = π
= 2π ⎢2 2 ⎥ 3
⎣ ⎦−π

π π π
2 π π
𝑎𝑛 =
1
π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 =
1
π (
∫ 𝑥 + 𝑥 𝑐𝑜𝑠𝑛𝑥𝑑𝑥 = 0 + ) π𝑛
1
2 [cos 𝑐𝑜𝑠 𝑛𝑥 ]
−π
+
2
π𝑛
2 [𝑛𝑥 ]
−π
−
2
π𝑛 −π
2 ∫ 𝐶𝑜𝑠𝑛𝑥𝑑
−π −π

1 2 2 π 4
= 2 [cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 𝑛π ] + 2 [𝑛π + 𝑛π ] − 3 [sin 𝑠𝑖𝑛 𝑛π ]−π = 2 cos 𝑐𝑜𝑠 𝑛π
π𝑛 π𝑛 π𝑛 𝑛

𝑛 𝑛
=
4
𝑛
2 [
(− 1) ... cos 𝑐𝑜𝑠 𝑛π = (− 1) ]
Finally,
π π π
2 2
𝑏𝑛 =
1
π (
∫ 𝑥 + 𝑥 𝑆𝑖𝑛𝑛𝑥𝑑𝑥 =− ) 1
𝑛π [(𝑥 + 𝑥 ) cos 𝑐𝑜𝑠 𝑛𝑥 ]−π + 1
𝑛π
∫ (1 + 2𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π −π

π
2 2 π π
=−
1
𝑛π [π cos 𝑐𝑜𝑠 𝑛π + π cos 𝑐𝑜𝑠 𝑛π − π cos 𝑐𝑜𝑠 𝑛π + ] 1
𝑛π
[sin 𝑠𝑖𝑛 𝑛𝑥 ]−π +
π𝑛
2
2 [𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]−π − ∫ s
−π

Example: 11. Expand in Fourier series the function 𝑓(𝑥) = 𝑥 sin 𝑠𝑖𝑛 𝑥 in the interval
π 1 1 1 1
–π < 𝑥 < π.Hence deduce that 4 = 2 + 1.3 − 3.5 + 5.7 − ……

39
Solution: By the definition of Fourier series, we have
𝑎 ∞
0
𝑓(𝑥) = 2
+ ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 = 𝑏𝑛𝑆𝑖𝑛 𝑛𝑥)………. (1)
𝑛=1

Where
π π
1 1
𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥(𝑛 ≠ 0)
−π −π

and
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥.
−π

The given function in question is 𝑓(𝑥) = 𝑛𝑥

∴𝑓(− 𝑥) =− 𝑥 sin 𝑠𝑖𝑛 (− 𝑥) =− 𝑥. − sin 𝑠𝑖𝑛 𝑥 = 𝑥 sin 𝑠𝑖𝑛 𝑥 = 𝑓(𝑥).

So 𝑓(𝑥) = 𝑥 sin 𝑠𝑖𝑛 𝑥 is an even function and in this case 𝑏𝑛 = 0.

And
π π π
1 2 2
𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥 = π
∫ 𝑓(𝑥)𝑑𝑥 ⇒ 𝑎0 = π
∫ 𝑥 sin 𝑠𝑖𝑛 𝑥 𝑑𝑥
−π 0 0

π
2 π 2
= π
[− 𝑥 cos 𝑐𝑜𝑠 𝑥 ]0 + π
∫ cos 𝑐𝑜𝑠 𝑥 𝑑𝑥
0

2 2 π
=− π
(π cos 𝑐𝑜𝑠 π − 0) + π
[sin 𝑠𝑖𝑛 𝑥 ]0

=− 2. (− 1). 2

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

When 𝑛 = 1, we have,
π
2
𝑎1 = π
∫ 𝑥 sin 𝑠𝑖𝑛 𝑥 cos 𝑐𝑜𝑠 𝑥 𝑑𝑥
0

π
1
= π
∫ 𝑥 sin 𝑠𝑖𝑛 2𝑥 𝑑𝑥
0

π π
1 𝑥cos𝑐𝑜𝑠 2𝑥 1
= ⎡− ⎤ + ∫ cos 𝑐𝑜𝑠 2𝑥 𝑑𝑥
π ⎣ 2 ⎦0 2π
0

40
 1 1 π
=− 2π
(π cos 𝑐𝑜𝑠 2π − 0) + 4π
[sin 𝑠𝑖𝑛 2𝑥 ]0

1 1
=− 2
+ 0 = 2

Also,
π
2
𝑎𝑛 = π
∫ 𝑥 sin 𝑠𝑖𝑛 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥(𝑛 ≠ 1), 𝑛 = 2, 3, 4, ………
0

π
1
= π
∫ 𝑥(sin 𝑠𝑖𝑛 (𝑛 + 1)𝑥 − sin 𝑠𝑖𝑛 (𝑛 − 1)𝑥 }𝑑𝑥.
0

π π
1 1
= π
∫ 𝑥 sin 𝑠𝑖𝑛 (𝑛 + 1)𝑥 𝑑𝑥 − π
∫ 𝑥 sin 𝑠𝑖𝑛 (𝑛 − 1)𝑥 𝑑𝑥.
0 0

π π
1 𝑥cos𝑐𝑜𝑠 (𝑛+1) 𝑥 π 1 1 π 1
= ⎡− ⎤ + ∫ cos 𝑐𝑜𝑠 (𝑛 + 1)𝑥 𝑑𝑥 + [𝑥 cos 𝑐𝑜𝑠 (𝑛 − 1)𝑥 ]0 − ∫ cos 𝑐
π ⎣ 𝑛+1 ⎦0 (𝑛+1)π
0
(𝑛−1)π (𝑛−1)π
0

1 1 π 1
=− (𝑛+1)π
. π cos 𝑐𝑜𝑠 (𝑛 + 1)π + 0 + 2 [sin 𝑠𝑖𝑛 (𝑛 + 1)𝑥 ]0 + (𝑛−1)π
. π cos 𝑐𝑜𝑠 (𝑛 − 1)π + 0 −
(𝑛+1) π

1 1
=− 𝑛+1
cos 𝑐𝑜𝑠 (𝑛 + 1)π + 𝑛−1
cos 𝑐𝑜𝑠 (𝑛 − 1)π + 0

1 1
= 𝑛+1
cos 𝑐𝑜𝑠 𝑛π − 𝑛−1
cos 𝑐𝑜𝑠 𝑛π

= ( 1
𝑛+1
−
1
𝑛−1 ) cos 𝑐𝑜𝑠 𝑛π
2 2 𝑛 2 𝑛+1
=− 2 cos 𝑐𝑜𝑠 𝑛π =− 2 . (− 1) = 2 (− 1)
𝑛 −1 𝑛 −1 𝑛 −1

Now putting the values of 𝑎0, 𝑎1, 𝑎𝑛 and 𝑏𝑛 in

𝑎0 ∞ ∞
𝑓(𝑥) = 2
+ 𝑎1 cos 𝑐𝑜𝑠 𝑥 + ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + ∑ 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥
𝑛=2 𝑛=1

We get
∞ 𝑛+1
2 1 2(−1)
𝑓(𝑥) = 2
− 2
cos 𝑐𝑜𝑠 𝑥 + ∑ 2 cos 𝑐𝑜𝑠 𝑛𝑥 + 0
𝑛=2 𝑛 −1

1 2 2 2 2 2
= 1− 2
cos 𝑐𝑜𝑠 𝑥 − 2 cos 𝑐𝑜𝑠 2𝑥 + 2 cos 𝑐𝑜𝑠 3𝑥 − 2 cos 𝑐𝑜𝑠 4𝑥 + 2 cos 𝑐𝑜𝑠 5𝑥 − 2 co
2 −1 3 −1 4 −1 5 −1 6 −1

2 2 2 2 2 2
= 1− 4
cos 𝑐𝑜𝑠 𝑥 − 3
cos 𝑐𝑜𝑠 2𝑥 + 8
cos 𝑐𝑜𝑠 3𝑥 − 15
cos 𝑐𝑜𝑠 4𝑥 + 24
cos 𝑐𝑜𝑠 5𝑥 − 35
cos 𝑐𝑜𝑠 6𝑥 +

41
 1 1 1 1 1 1
= 1 − 2⎡ 4 cos 𝑐𝑜𝑠 𝑥 + cos 𝑐𝑜𝑠 2𝑥 − cos 𝑐𝑜𝑠 3𝑥 + cos 𝑐𝑜𝑠 4𝑥 − cos 𝑐𝑜𝑠 5𝑥 + cos 𝑐𝑜𝑠 6𝑥
⎣ 3 8 15 24 35

π
Now Putting 𝑥 = 2
in (2), we get,
𝑓( )=
π
2
π
2 ⎣
𝑠𝑖𝑛
π
= 1 − 2⎡ 4 cos 𝑐𝑜𝑠 2 + 5 cos 𝑐𝑜𝑠 π −
2
1 π 1 1
8
cos 𝑐𝑜𝑠 π −
1
8
cos 𝑐𝑜𝑠
3π
2
+
1
15
cos 𝑐𝑜𝑠 2π
π 1 1 1
Or, 2 = 1 − 2⎡0 − 3 − 0 + 15 − 0 − 35 − 0 + 0 ...⎤
⎣ ⎦
2 3 2 2
= 2 + 3 − 15 + 35 − ...
π 1 1 1 1
Or, 4
= 2
+ 3
− 15 + 35 − ...
π 1 1 1 1
Or, 4
= 2
+ 1.3
+ 3.5 + 5.7 − …

Example 12: Find the fourier series of the function f(x) of period 2π, where 𝑓(𝑥) = 𝑥𝑐𝑜𝑠𝑥 for
–π < 𝑥 ≤ π.

Solution: By definition of the Fourier series, we have

𝑎0 ∝
𝑓(𝑥) = 2 (
+ ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ……………………(1)
𝑛=1
)
π π
1 1
Where, 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥, 𝑛 ≠ 0
−π −π

π
1
and 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π

The given function is 𝑓(𝑥) = 𝑥𝑐𝑜𝑠𝑥

∴𝑓(− 𝑥) =− 𝑥 cos 𝑐𝑜𝑠 (− 𝑥) =− 𝑥𝑐𝑜𝑠𝑥 =− 𝑓(𝑥). And so the function 𝑓(𝑥) = 𝑥𝑐𝑜𝑠𝑥 is an
odd function and in this case 𝑎𝑛 = 0 for 𝑛 = 0, 1, 2, 3, …….

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

thus equation no (1) becomes

∝ ∝
𝑓(𝑥) = ∑ 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 = 𝑏1 sin 𝑠𝑖𝑛 𝑥 + ∑ 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ………(2)
𝑛=1 𝑛=2

π
2
Where 𝑏1 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑥 𝑑𝑥
0

42
 π
2
= π
∫ 𝑥 cos 𝑐𝑜𝑠 𝑥 sin 𝑠𝑖𝑛 𝑥 𝑑𝑥
0

π
1
= π
∫ 𝑥 sin 𝑠𝑖𝑛 2𝑥 𝑑𝑥
0

π
1 π 1
=− 2π
[𝑥 cos 𝑐𝑜𝑠 2𝑥 ]0 + 2π
∫ cos 𝑐𝑜𝑠 2𝑥 𝑑𝑥
0

1 1 π
=− 2π
(π cos 𝑐𝑜𝑠 2π − 0) + 4π
[sin 𝑠𝑖𝑛 2𝑥 ]0

1 1
=− 2
+ 0 = 2

π
2
Also 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥(𝑛 ≠ 1, 𝑛 = 2, 3, 4 ....)
0

π
2
= π
∫ 𝑥 cos 𝑐𝑜𝑠 𝑥 sin 𝑠𝑖𝑛 𝑥 𝑑𝑥
0

π
1
= π
∫ 𝑥[sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥 + sin 𝑠𝑖𝑛 (𝑛 − 1) 𝑥]𝑑𝑥
0

π π
1 1
= π
∫ 𝑥 sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥𝑑𝑥 + π
∫ 𝑥 sin 𝑠𝑖𝑛 (𝑛 − 1) 𝑥𝑑𝑥
0 0

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

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

1 1
=− (𝑛+1)
. π cos 𝑐𝑜𝑠 (𝑛 + 1) π − π(𝑛−1)
. π cos 𝑐𝑜𝑠 (𝑛 − 1) π

1 π 1 π
+ 2 [sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥]0 + 2 [sin 𝑠𝑖𝑛 (𝑛 − 1) 𝑥]0
π(𝑛+1) π(𝑛−1)

1 1
=− (𝑛+1)
. cos 𝑐𝑜𝑠 (𝑛π + π) − (𝑛−1)
cos 𝑐𝑜𝑠 (𝑛π − π) + 0

cos𝑐𝑜𝑠 𝑛π cos𝑐𝑜𝑠 𝑛π
= 𝑛+1
+ 𝑛−1

43
 = ( 1
𝑛+1
+
1
𝑛−1 ) cos 𝑐𝑜𝑠 𝑛π =
2𝑛
2
𝑛 −1
𝑛
. (− 1) . (𝑛 ≠ 1)

𝑛
∴[cos 𝑐𝑜𝑠 𝑛π = (− 1) ]

Now putting the values of 𝑏1 and 𝑏𝑛 (𝑛 ≠ 1) in (2), we get

∝
1 2𝑛 𝑛
𝑓(𝑥) =− 2
sin 𝑠𝑖𝑛 𝑥 + ∑ 2 (− 1) . sin 𝑠𝑖𝑛 𝑛𝑥
𝑛=2 𝑛 −1

1 4 6 8
=− 2
sin 𝑠𝑖𝑛 𝑥 + 3
sin 𝑠𝑖𝑛 2𝑥 − 8
sin 𝑠𝑖𝑛 3𝑥 + 15
sin 𝑠𝑖𝑛 4𝑥 − …

1 2 3 4
=− sin 𝑠𝑖𝑛 𝑥 + 2⎡ 1.3 sin 𝑠𝑖𝑛 2𝑥 − sin 𝑠𝑖𝑛 3𝑥 + sin 𝑠𝑖𝑛 4𝑥 − …⎤
2 ⎣ 2.4 3.5 ⎦

Example 13: Find the series of sines and cosines of multiplies of 𝑥 which represents 𝑓(𝑥) in the
interval –π < 𝑥 < π where
π𝑥
𝑓(𝑥) = {0 𝑤ℎ𝑒𝑟𝑒 − π < 𝑥 < 0 4
𝑤ℎ𝑒𝑟𝑒 0 < 𝑥 < π and deduce that
2
π 1 1
8
= 1+ 2 + 2 +...
3 5

Solution: By definition of the Fourier series, we have

𝑎0 ∞
𝑓(𝑥) = 2
+ ∑ (𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥) .................(1)
𝑛=1

π π π
1 1 1
Where 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥; 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥(𝑛 ≠ 0) and 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 .
−π −π −π

π 0 π
1 1
Now 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥 = π
[ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥]
−π −π 0

π π
1 π𝑥
= π
[ ∫ 0. 𝑑𝑥 + ∫ 4
𝑑𝑥]
−π 0

π 2 2
1 1 𝑥 π
= 0+ π
∫ 𝑥𝑑𝑥 = 4
[ 2
]π 0 = 8
0

44
 π
1
And 𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 + π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π 0

0 π
1 1 π𝑥
= π
∫ 0. 𝑐𝑜𝑠𝑛𝑥𝑑𝑥 + π
∫ 4
𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π 0

π
1
= 0+ 4
∫ 𝑥𝑐𝑜𝑠𝑛𝑥𝑑𝑥
0

π
1 1
= 4𝑛
[𝑥𝑠𝑖𝑛𝑛𝑥]π 0 − 4𝑛
∫ 𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

1
= 0+ 2 [𝑐𝑜𝑠𝑛𝑥]π 0
4𝑛

1 1 𝑛
= 2 [𝑐𝑜𝑠𝑛π − 𝑐𝑜𝑠0] = 2 {(− 1) − 1}
4𝑛 4𝑛

π
1
Finally 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π 0

0 π
1 1 π𝑥
= π
∫ 0. 𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + π
∫ 4
𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π 0

π
1
= 0+ 4
∫ 𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

π
1 1
=− 4𝑛
[𝑥𝑐𝑜𝑠𝑛𝑥]π 0 + 4𝑛
∫ 𝑐𝑜𝑠𝑛𝑥𝑑𝑥
0

1 1
=− 4𝑛
[π𝑐𝑜𝑠𝑛π − 0] + 2 [𝑠𝑖𝑛𝑛𝑥]π 0
4𝑛

π 𝑛 π 𝑛 𝑛
=− 4𝑛
(− 1) + 0 =− 4𝑛
(− 1) ∴[𝑐𝑜𝑠𝑛𝑛π = (− 1) ]

Thus substituting the values of 𝑎0, 𝑎𝑛𝑎𝑛𝑑 𝑏𝑛 in (1) we get,

45
 2 ∞
𝑛 𝑛
𝑓(𝑥) =
π
16
1
+ ∑ [ 4𝑛 (− 1) − 1 𝑐𝑜𝑠𝑛𝑥 + {− { } π
4𝑛
(− 1) }𝑠𝑖𝑛𝑛𝑥]
𝑛=1

2
+ ⎡⎢− 𝑐𝑜𝑠5𝑥 +...⎤⎥ −
π 2 2 2 π 𝑠𝑖𝑛𝑥 1 1 1
= 16 2 𝑐𝑜𝑠𝑥 + 0 − 2 + 0− 2 4
[− 1
+ 2
𝑠𝑖𝑛2𝑥 − 3
𝑠𝑖𝑛3𝑥 + 4
𝑠𝑖𝑛4𝑥 +
⎣ 4.1 4.3 4.5 ⎦

2
=
π
−
1 ⎡ 𝑐𝑜𝑠𝑥 + 𝑐𝑜𝑠3𝑥
+
𝑐𝑜𝑠5𝑥
+...⎤⎥ +
π
⎡
𝑠𝑖𝑛𝑥
−
𝑠𝑖𝑛2𝑥
+
𝑠𝑖𝑛3𝑥
−
𝑠𝑖𝑛4𝑥
+...⎤……(2)
16 2 ⎢ 12 2 2 4 ⎣ 1 2 3 4 ⎦
⎣ 3 5 ⎦

( )
2 2
1 1 π π
Now 𝑓(π) = 2
[𝑓(− π) + 𝑓(π)] = 2
0+ 4
= 8

Therefore, putting 𝑥 = π in (2) we get

2 2
𝑓(π) =
π
=
π
−
1 ⎡ 𝑐𝑜𝑠𝑥 + 𝑐𝑜𝑠3𝑥
+
𝑐𝑜𝑠5𝑥
+...⎤⎥ +
π
⎡
𝑠𝑖𝑛𝑥
−
𝑠𝑖𝑛2𝑥
+
𝑠𝑖𝑛3𝑥
−
𝑠𝑖𝑛4𝑥
+...⎤
8 16 2 ⎢ 12 2 2 4 ⎣ 1 2 3 4 ⎦
⎣ 3 5 ⎦
2 2
Or,
π
=
π
−
1 ⎡− 1
−
1
−
1
− …⎤⎥ + 0
8 16 2 ⎢ 2 2 2
⎣ 1 3 5 ⎦
2
π 1 1 1
Or, 16
= 2
[1 + 2 + 2 +...]
3 5

2
π 1 1
Or, 8
= 1+ 2 + 2 +...
3 5

Example 14: if 𝑓(𝑥) = {1 − 1 𝑤ℎ𝑒𝑛 0 ≤ 𝑥 ≤ π 𝑤ℎ𝑒𝑛 − π ≤ 𝑥 < 0

π 1 1 1 1
expand 𝑓(𝑥) in the Fourier series. Hence deduce 4
= 1− 3
+ 5
− 7
+ 9
− …

Solution: By definition of the Fourier series we have

𝑎𝑜 ∞
𝑓(𝑥) = 2
+ ∑ (𝑎𝑛𝐶𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑆𝑖𝑛𝑛𝑥) --- --- --- --- --- (1)
𝑛=1
π π
1 1
Where 𝑎𝑜 = π
∫ 𝑓(𝑥)𝑑𝑥, 𝑎𝑛 = π
∫ 𝑓(𝑥)𝐶𝑜𝑠𝑛𝑥𝑑𝑥 (𝑛 ≠ 0) and
−π −π
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑆𝑖𝑛𝑛𝑥𝑑𝑥
−π

46
 π 0 π
1 1 1
Now 𝑎𝑜 = π
∫ 𝑓(𝑥)𝑑𝑥 = π
∫ 𝑓(𝑥)𝑑𝑥 + π
∫ 𝑓(𝑥)𝑑𝑥
−π −π 0

0 π
1 1
= π
∫ (− 1)𝑑𝑥 + π
∫(1)𝑑𝑥
−π 0

1 0 1 π
=− π
[𝑥]−π + π
[𝑥]0

1 1
=− π
[0 − π] + π
[π − 0]

= 0

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

0 π
1 1
= π
∫ (− 1) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + π
∫(1) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π 0

1 0 1 π
=− 𝑛π
[sin 𝑠𝑖𝑛 𝑛𝑥 ]−π + 𝑛π
[sin 𝑠𝑖𝑛 𝑛𝑥 ]0

1 1
=− 𝑛π
[sin 𝑠𝑖𝑛 0 − sin 𝑠𝑖𝑛 𝑛π ] + 𝑛π
[sin 𝑠𝑖𝑛 𝑛π − sin 𝑠𝑖𝑛 0 ]

= 0 𝑆𝑖𝑛𝑐𝑒, sin 𝑠𝑖𝑛 0 = 0

Again,
π 0 π
1 1 1
𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π −π 0

0 π
1 1
= π
∫ (− 1) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫(1) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

1 0 1 π
=− 𝑛π
[− cos 𝑐𝑜𝑠 𝑛𝑥 ]−π − 𝑛π
[cos 𝑐𝑜𝑠 𝑛𝑥 ]0

1 1
= 𝑛π
[cos 𝑐𝑜𝑠 0 − cos 𝑐𝑜𝑠 𝑛π ] − 𝑛π
[cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0 ]

2
= 𝑛π
[cos 𝑐𝑜𝑠 0 − cos 𝑐𝑜𝑠 𝑛π ]

=
2
𝑛π [1 − (− 1)𝑛]
47
 Therefore, putting the values of 𝑎𝑜, 𝑎𝑛𝑎𝑛𝑑 𝑏𝑛 in (1), we get

∞
𝑓(𝑥) = 0 + 0 + ∑
𝑛=1
( [1 − (− 1) ]) sin 𝑠𝑖𝑛 𝑛𝑥
2
𝑛π
𝑛

2 2sin𝑠𝑖𝑛 𝑥 2sin𝑠𝑖𝑛 3𝑥 2sin𝑠𝑖𝑛 5𝑥

= ⎡ + 0+ + 0 + +...⎤
π ⎣ 1 3 5 ⎦
2 2sin𝑠𝑖𝑛 𝑥 2sin𝑠𝑖𝑛 3𝑥 2sin𝑠𝑖𝑛 5𝑥
= ⎡ + + +...⎤ --- --- --- --- ---
π ⎣ 1 3 5 ⎦
(2)

Now since 0 < 𝑥 < π

π π
If 𝑥 = 2
, 𝑓(𝑥) = 1, Hence putting 𝑥 = 2
, we get

π 3π 5π

1=
4 ⎡ 2sin𝑠𝑖𝑛 2
+
2sin𝑠𝑖𝑛 2
+
2sin𝑠𝑖𝑛 2
+...⎤⎥
π ⎢ 1 3 5
⎣ ⎦
π 1 1 1
or. = ⎡1 − + − +...⎤
4 ⎣ 3 5 7 ⎦

Example 15: Find the Fourier series expansion of the function below, having period 2π :

𝑓(𝑥) = {− 𝑥 𝑖𝑓 − π < 𝑥 < 0 𝑥 𝑖𝑓 0 < 𝑥 < π

Solution: By the definition of the Fourier series we have

𝑎0 ∞
𝑓(𝑥) = 2
+ ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ) (1)
𝑛=1

π
1
Where 𝑎0 = π
∫ 𝑓(𝑥) 𝑑𝑥,
−π

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π
π
1
and 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π

Now,

48
 π
1
𝑎0 = π
∫ 𝑓(𝑥) 𝑑𝑥
−π
0 π
1 1
= π
∫ 𝑓(𝑥) 𝑑𝑥 + π
∫ 𝑓(𝑥) 𝑑𝑥
−π 0

0 π
1 1
= π
∫ (− 𝑥)𝑑𝑥 + π
∫ 𝑥 𝑑𝑥
−π 0

2 0 2 π
1 ⎡𝑥 ⎤ +1⎡𝑥 ⎤
=− π ⎢2⎥ ⎢ ⎥
⎣ ⎦−π π ⎣ 2 ⎦0
2 2
1 π 1 π
=− π
(0 − 2
)+ π
( 2
− 0)

π π
=2 + 2

=π
π
1
And 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π

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

0 π
1 1
= π
∫ (− 𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + π
∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π 0

0
1 0 1 1 π
=− π𝑛
[𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]−π + π𝑛
∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π𝑛
[𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]0 -
−π
π
1
π𝑛
∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
0

1 0 1 π
=0 - 2 [cos 𝑐𝑜𝑠 𝑛𝑥 ]−π+ 0 + 2 [cos 𝑐𝑜𝑠 𝑛𝑥 ]0
π𝑛 π𝑛

1 1
=- 2 [cos 𝑐𝑜𝑠 0 − cos 𝑐𝑜𝑠 𝑛π] + 2 [cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0]
π𝑛 π𝑛

2
= 2 [cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0]
π𝑛

2 𝑛
= 2 [(− 1) − 1]
π𝑛

49
 π
1
Finally, 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

0 π
1 1
= π
∫ (− 𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫ 𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

0
1 0 1 1 π
= π𝑛 [𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 ]−π − π𝑛
∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 − π𝑛
[𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 ]0 +
−π
π
1
π𝑛
∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

=
1 1 0 1 1 π
π𝑛
[+ π cos 𝑐𝑜𝑠 𝑛π] − 2 [sin 𝑠𝑖𝑛 𝑛𝑥 ]−π − π𝑛
[π cos 𝑐𝑜𝑠 𝑛π] + 2 [sin 𝑠𝑖𝑛 𝑛𝑥 ]0
π𝑛 π𝑛

cos𝑐𝑜𝑠 𝑛π cos𝑐𝑜𝑠 𝑛π
= 𝑛
− 𝑛
+ 0

=0

Now putting the values of 𝑎0 '𝑎𝑛 𝑎𝑛𝑑 𝑏𝑛 in (1) we get

∞
𝑓(𝑥) =
π
2
+ ∑
2
𝑛 {(− 1)𝑛 − 1} cos 𝑐𝑜𝑠 𝑛𝑥 + 0
𝑛=1

+ ⎡⎢− cos 𝑐𝑜𝑠 5𝑥 +...⎤⎥

π 2.2 2.2 2.2
= 2 2 cos 𝑐𝑜𝑠 𝑥 + 0 − 2 cos 𝑐𝑜𝑠 3𝑥 + 0 − 2
⎣ π1 π3 π5 ⎦
π 4 cos𝑐𝑜𝑠 𝑥 cos𝑐𝑜𝑠 3𝑥 cos𝑐𝑜𝑠 5𝑥
= 2
− π
[ 2 + 2 + 2 +...]
1 3 5

Example 16: Find the Fourier series expansion of the function f(x) in the interval (− π, π) where

𝑓(𝑥) = {π + 𝑥 𝑤ℎ𝑒𝑛 − π < 𝑥 < 0 π − 𝑥 𝑤ℎ𝑒𝑛 0 < 𝑥 < π

Solution: By the definition of the Fourier series we have

∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ) (1)
𝑛=1

50
 π
1
Where 𝑎0 = 2π
∫ 𝑓(𝑥) 𝑑𝑥,
−π

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π
π
1
and 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π

Now,
π
1
𝑎0 = 2π
∫ 𝑓(𝑥) 𝑑𝑥
−π
0 π
1 1
= 2π
∫ 𝑓(𝑥) 𝑑𝑥 + 2π
∫ 𝑓(𝑥) 𝑑𝑥
−π 0

0 π
1 1
= 2π
∫ (π + 𝑥)𝑑𝑥 + 2π
∫(π − 𝑥) 𝑑𝑥
−π 0

2 0 2 π
1 ⎡π𝑥 + 𝑥 ⎤ + 1 ⎡π𝑥 − 𝑥 ⎤
= 2π ⎢ 2 ⎥ ⎢ ⎥
⎣ ⎦−π 2π ⎣ 2 ⎦0
2 2
1 2 π 2 π
= 2π
[0 + π − 2
+ 0 +π − 2

1 2
= 2π
*π

π
= 2

π
1
And 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π

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

0 π
1 1
= π
∫ (π + 𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + π
∫(π − 𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
−π 0

=
0 0 π π
1
π
[π ∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + ∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + π ∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 − ∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥]
−π −π 0 0

51
 0
1 π 0 1 0 1
= π
[ 𝑛 [sin 𝑠𝑖𝑛 𝑛𝑥 ]−π + 𝑛
[𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]−π − 𝑛
∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 +
−π
π
π π 1 π 1
𝑛
[sin 𝑠𝑖𝑛 𝑛𝑥 ]0 − 𝑛
[𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]0 + 𝑛
∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥]
0

0 π
= π ⎡⎢0 + [cos 𝑐𝑜𝑠 𝑛𝑥 ]0 ⎤⎥
1 1 1
2 [cos 𝑐𝑜𝑠 𝑛𝑥 ]−π + 0 − 2
⎣ 𝑛 𝑛 ⎦
1
= 2 [cos 𝑐𝑜𝑠 0 − cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 𝑛π + cos 𝑐𝑜𝑠 0 ]
π𝑛

2
= 2 [1 − cos 𝑐𝑜𝑠 𝑛π ]
π𝑛

=
π𝑛
2
2 [1 − (− 1)𝑛]
π
1
Finally, 𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

0 π
1 1
= π
∫ (π + 𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫(π − 𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

0 0 π π
1 ⎡ ⎤
= ⎢π ∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + ∫ 𝑛𝑥 𝑑𝑥 + π ∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 − ∫ 𝑛𝑥 𝑑𝑥⎥
π ⎢ −π ⎥
⎣ −π 0 0 ⎦

0
1 ⎡ π 0 1 0 1 π π 1
= ⎢− [cos 𝑐𝑜𝑠 𝑛𝑥 ]−π + [𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 ]−π + ∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 − [cos 𝑐𝑜𝑠 𝑛𝑥 ]0 + [𝑥 cos 𝑐𝑜𝑠 𝑛
π ⎢ 𝑛 𝑛 𝑛 𝑛 𝑛
⎣ −π

1 ⎡− π π π 1 0 π π π
= π ⎢ 𝑛
+ 𝑛
cos 𝑐𝑜𝑠 𝑛π + 0 − 𝑛
cos 𝑐𝑜𝑠 𝑛π + 2 [sin 𝑠𝑖𝑛 𝑛𝑥 ]−π − 𝑛
cos 𝑐𝑜𝑠 𝑛π + 𝑛
+ 𝑛
cos 𝑐𝑜𝑠 𝑛
⎣ 𝑛

1
= π
x0

= 0

∴ 𝑏𝑛 = 0

52
Hence putting the values of 𝑎0 , 𝑎𝑛 , 𝑏𝑛 in equation (1), we get ,

∞
𝑓(𝑥) =
π
2
+ ∑
2
2 {1 − (− 1)𝑛} cos 𝑐𝑜𝑠 𝑛𝑥 +0
𝑛=1 π𝑛

π 2 2 2
= 2
+ 2 . 2 cos 𝑐𝑜𝑠 𝑥 + 0 + 2 . 2 cos 𝑐𝑜𝑠 3𝑥 + 0 + 2 . 2 cos 𝑐𝑜𝑠 5𝑥 +...
π .1 π .3 π .5

=
π
+
4 ⎡ cos𝑐𝑜𝑠 𝑥 + cos𝑐𝑜𝑠 3𝑥
+
cos𝑐𝑜𝑠 5𝑥
+...⎤⎥
2 π ⎢ 12 2 2
⎣ 3 5 ⎦

Example-17 : If 0≤𝑥≤π . then show that

2 ∞
π cos𝑐𝑜𝑠 2𝑛𝑥
𝑥(π − 𝑥) = 6
− ∑ 2 .
𝑛=1 𝑛

2
Solution : Let, 𝑓(𝑥) = 𝑥(π − 𝑥) = π𝑥 − 𝑥

Since the given interval is 0 ≤ 𝑥 ≤ π . (Half Range), the Fourier series and in the case

𝑎0 ∞
𝑓(𝑥) = 2
+ ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 )
𝑛=1

π
2
Where 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 𝑎𝑛𝑑 𝑏𝑛 = 0
0

π
2
Now 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥
0

π
2 2
= π
∫( π𝑥 − 𝑥 )𝑑𝑥
0

2 3 π
2 𝑥 𝑥
= π
[ π. 2
− 3
]
0

3 3
2 ⎡π − π ⎤
= π ⎢ 2 3 ⎥
⎣ ⎦
3
2 π
= π
. 6

53
 3
π
= 3

π
2
And 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 [𝑛 ≠ 0]
0

π
2 2
= π
∫( π𝑥 − 𝑥 ) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

π π
2 2 2
= π𝑛
[( π𝑥 − 𝑥 ) sin 𝑠𝑖𝑛 𝑛𝑥 ]0 − π𝑛
∫( π − 2𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
0

π π
2 4
= 0− 𝑛
∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π𝑛
∫ 𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
0 0

π
2 π 4 4
= 2 [ 𝑐𝑜𝑥 𝑛𝑥]
0
− 2 [𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 ] + 2 ∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
𝑛 π𝑛 π𝑛 0

2 4 4 π
= 2 [𝑐𝑜𝑥 𝑛π − cos 𝑐𝑜𝑠 0 ] − 2 . π cos 𝑐𝑜𝑠 𝑛π + 3 [ 𝑠𝑖𝑛 𝑛𝑥]0
𝑛 π𝑛 π𝑛

2 2 4
= 2 𝑐𝑜𝑥 𝑛π − 2 − 2 cos 𝑐𝑜𝑠 𝑛π + 0
𝑛 𝑛 𝑛

2 𝑛
= 2 [1 + (− 1) ]
𝑛

2 ∞
𝑛
∑ ⎡⎢− {1 + (− 1) }⎤⎥ cos 𝑐𝑜𝑠 𝑛𝑥
π 2
Thus 𝑓(𝑥) = 6
+ 2
𝑛=1⎣ 𝑛 ⎦

2
π 4 4 4 4
= 6
+ 0 − 2 cos 𝑐𝑜𝑠 2𝑥 + 0 − 2 cos 𝑐𝑜𝑠 4𝑥 + 0 − 2 cos 𝑐𝑜𝑠 6𝑥 + 0 − 2 cos 𝑐𝑜𝑠 8𝑥 +... ....
2 4 6 8

2
=
π
−
4 ⎡cos 𝑐𝑜𝑠 2𝑥 + cos𝑐𝑜𝑠 4𝑥
+
cos𝑐𝑜𝑠 6𝑥
+
cos𝑐𝑜𝑠 8𝑥
+...⎤⎥
6 2 ⎢ 2 2 2
2 ⎣ 2 3 4 ⎦
2
− ⎡⎢ +...⎤⎥
π cos𝑐𝑜𝑠 2𝑥 cos𝑐𝑜𝑠 2.2𝑥 cos𝑐𝑜𝑠 2.3𝑥 cos𝑐𝑜𝑠 2.4𝑥
= 6 2 + 2 + 2 + 2
⎣ 1 2 3 4 ⎦
2 ∞
π cos𝑐𝑜𝑠 2𝑛𝑥
= 6
− ∑ 2
𝑛=1 𝑛

54
 2 ∞
π cos𝑐𝑜𝑠 2𝑛𝑥
Therefore, 𝑥(π − 𝑥) = 6
− ∑ 2
𝑛=1 𝑛

Example-18: Obtain the Fourier series for the function 𝑓(𝑥) given by
2𝑥 2𝑥
𝑓(𝑥) = {1 + π
𝑓𝑜𝑟 − π ≤ 𝑥 ≤ 0 1 − π
𝑓𝑜𝑟 0 ≤ 𝑥 ≤ π

2
1 1 1 π
Hence deduce 2 + 2 + 2 +...= 8
.
1 3 5

Solution: The given function is

2𝑥 2𝑥
𝑓(𝑥) = {1 + π
𝑓𝑜𝑟 − π ≤ 𝑥 ≤ 0 1 − π
𝑓𝑜𝑟 0 ≤ 𝑥 ≤ π

2𝑥
Since 𝑓(− 𝑥) = 1 − π
𝑖𝑛 (− π, 0) = 𝑓(𝑥) 𝑖𝑛 (0, π)

2𝑥
𝑎𝑛𝑑 𝑓(− 𝑥) = 1 + π
𝑖𝑛 (0, π) = 𝑓(𝑥) 𝑖𝑛 (− π, 0)

Therefore, 𝑓(𝑥) is an even function in (− π, π) and every even function is symmetrical about
the 𝑦 − 𝑎𝑥𝑖𝑠. Thus we have,

𝑎0 ∞
𝑓(𝑥) = 2
+ ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 (1) 𝑏𝑛 = 0
𝑛=1
( )
π
2
Where 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥
0

π
=
2
π
∫ 1−
0
( 2𝑥
π ) 𝑑𝑥
2 π
2 2 𝑥
= π
[𝑥 − π
. 2
]
0

2
= π
[π − π]

= 0
π
2
𝑎𝑛𝑑 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

55
 π
=
2
π (
∫ 1 −
0
2𝑥
π ) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
π
2 π 4
= π𝑛
[ 𝑠𝑖𝑛 𝑛𝑥]0 + 2 ∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
π 0

π
4 π 4
= 0 − 2 [ 𝑥 𝑠𝑖𝑛 𝑛𝑥]0 + 2 ∫ sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
π𝑛 π𝑛 0

4 π
= 0 − 2 [ 𝑐𝑜𝑠 𝑛𝑥]0
π𝑛

4
=− 2 2 (𝑐𝑜𝑠𝑛π − 𝑐𝑜𝑠0)
π𝑛

4 𝑛 4 𝑛
=− 2 2 {(− 1) − 1} = 2 2 {1 − (− 1) }
π𝑛 π𝑛

Now putting the values of 𝑎0 and 𝑎𝑛 in equation (1), We get

∞
4 𝑛
𝑓(𝑥) = 0 + ∑ 2 2 {1 − (− 1) }𝑐𝑜𝑠𝑛𝑥
𝑛=1 π 𝑛

4 2𝑐𝑜𝑠𝑥 2𝑐𝑜𝑠3𝑥 2𝑐𝑜𝑠5𝑥

= 2 [ 2 + 0+ 2 + 0+ 2 + ⋯ ⋯⋯ ⋯]
π 1 3 5

8 𝑐𝑜𝑠𝑥 𝑐𝑜𝑠3𝑥 𝑐𝑜𝑠5𝑥

= 2 [ 2 + 2 + 2 + ⋯ ⋯⋯ ⋯]……………..(2)
π 1 3 5

Which is required Fourier series for f(x)

Now putting x=0 in (2),we have

8 𝑐𝑜𝑠0 𝑐𝑜𝑠0 𝑐𝑜𝑠0
𝑓(0) = 1 = 2 [ 2 + 2 + 2 + ⋯ ⋯⋯ ⋯]
π 1 3 5

2
π 1 1 1
𝑜𝑟 8
= 2 + 2 + 2 + ⋯ ⋯⋯ ⋯
1 3 5

Example-19: Show that in Fourier series

56
𝑓(𝑥) = {𝑐𝑜𝑠𝑥 𝑓𝑜𝑟 0≤𝑥≤π − 𝑐𝑜𝑠𝑥 𝑓𝑜𝑟 − π≤𝑥≤0
8 𝑠𝑖𝑛2𝑥 2𝑠𝑖𝑛4𝑥 3𝑠𝑖𝑛6𝑥
Is 𝑓(𝑥) = π
[ 1.3
+ 3.5
+ 5.7
+ ⋯ ⋯⋯ ⋯]

Solution: By definition of the Fourier series we have

π
(
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 ⋯ ⋯⋯ ⋯(1)
𝑛=1
)
Where,
π
1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π

π
1
𝑎𝑛𝑑 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π

Now
π
1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

0 π
1 1
= 2π
∫ 𝑓(𝑥)𝑑𝑥 + 2π
∫ 𝑓(𝑥)𝑑𝑥
−π 0

0 π
1 1
= 2π
∫ (− 𝑐𝑜𝑠𝑥)𝑑𝑥 + 2π
∫ 𝑐𝑜𝑠𝑥𝑑𝑥
−π 0

1 0 1 π 1
=− 2π
[𝑠𝑖𝑛𝑥] −π + 2π
[𝑠𝑖𝑛𝑥] 0 = 2π
×0 = 0

and
π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 + π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π 0

57
 0 π
1 1
= π
∫ (− 𝑐𝑜𝑠𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 + π
∫ 𝑐𝑜𝑠𝑥𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π 0
0 π
1 1
=− 2π
∫ 2𝑐𝑜𝑠𝑛𝑥𝑐𝑜𝑠𝑥𝑑𝑥 + 2π
∫ 2𝑐𝑜𝑠𝑛𝑥𝑐𝑜𝑠𝑥𝑑𝑥
−π 0

0 π
1 1
=− 2π
∫ [cos 𝑐𝑜𝑠 (𝑛 + 1) 𝑥 + cos 𝑐𝑜𝑠 (𝑛 − 1) 𝑥]𝑑𝑥 + 2π
∫[cos 𝑐𝑜𝑠 (𝑛 + 1) 𝑥 + 𝑐𝑜𝑠⁡(𝑛 − 1)𝑥]𝑑𝑥
−π 0

1 sin𝑠𝑖𝑛 (𝑛+1) 𝑥 sin𝑠𝑖𝑛 (𝑛−1) 𝑥 0 1 sin𝑠𝑖𝑛 (𝑛+1) 𝑥 sin𝑠𝑖𝑛 (𝑛−1) 𝑥 π

=− ⎡ + ⎤ −π + ⎡ + ⎤0
2π ⎣ 𝑛+1 𝑛−1 ⎦ 2π ⎣ 𝑛+1 𝑛−1 ⎦
1 1
=− 2π
.0 + 2π
.0

= 0+ 0

= 0

Finally,
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π 0

0 π
1 1
= π
∫ (− 𝑐𝑜𝑠𝑥) 𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + π
∫ 𝑐𝑜𝑠𝑥 𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π 0

0 π
1 1
=− 2π
∫ 2 𝑠𝑖𝑛𝑛𝑥 𝑐𝑜𝑠𝑥 𝑑𝑥 + 2π
∫ 2 𝑠𝑖𝑛𝑛𝑥 𝑐𝑜𝑠𝑥 𝑑𝑥
−π 0

0 π
1 1
=− 2π
∫ [sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥 + sin 𝑠𝑖𝑛 (𝑛 − 1) 𝑥]𝑑𝑥 + 2π
∫[sin 𝑠𝑖𝑛 (𝑛 + 1) 𝑥 + 𝑠𝑖𝑛⁡(𝑛 − 1)𝑥]𝑑𝑥
−π 0

1 cos𝑐𝑜𝑠 (𝑛+1) 𝑥 cos𝑐𝑜𝑠 (𝑛−1) 𝑥 0 1 cos𝑐𝑜𝑠 (𝑛+1) 𝑥 cos𝑐𝑜𝑠 (𝑛−1) 𝑥 π

= ⎡ + ⎤ −π − ⎡ + ⎤0
2π ⎣ 𝑛+1 𝑛−1 ⎦ 2π ⎣ 𝑛+1 𝑛−1 ⎦
1 𝑐𝑜𝑠0 𝑐𝑜𝑠0 cos𝑐𝑜𝑠 (𝑛+1) π cos𝑐𝑜𝑠 (𝑛−1) π 1 𝑐𝑜𝑠⁡(𝑛+1)π 𝑐𝑜𝑠⁡(𝑛−1)π 𝑐𝑜𝑠0 𝑐𝑜𝑠0
= ⎡ 𝑛+1 + − − ⎤− ⎡ + − − ⎤
2π ⎣ 𝑛−1 𝑛+1 𝑛−1 ⎦ 2π ⎣ 𝑛+1 𝑛−1 𝑛+1 𝑛−1 ⎦
1 1 1 cos𝑐𝑜𝑠 (𝑛+1) π cos𝑐𝑜𝑠 (𝑛−1) π
= ⎡ 𝑛+1 + − − ⎤
2π ⎣ 𝑛−1 𝑛+1 𝑛−1 ⎦

58
=
1 ⎡ 2𝑛 + cos𝑐𝑜𝑠 𝑛 π
+
cos𝑐𝑜𝑠 𝑛 π ⎤
π ⎢ 𝑛2−1 𝑛+1 𝑛−1 ⎥
⎣ ⎦

=
1 ⎡ 2𝑛 + 2𝑛
𝑐𝑜𝑠𝑛π⎤⎥
π ⎢ 𝑛2−1 2
⎣ 𝑛 −1 ⎦
1 2𝑛
= π
. 2 [1 + 𝑐𝑜𝑠𝑛π]
𝑛 −1

2𝑛 𝑛
= 2 {1 + (− 1) }
π(𝑛 −1)

Now putting the values of 𝑎0, 𝑎𝑛 and 𝑏𝑛 in equation (1), we get

∞
2𝑛 𝑛
𝑓(𝑥) = 0 + ∑ 2 {1 + (− 1) }𝑠𝑖𝑛𝑛𝑥
𝑛=1 π(𝑛 −1)

2 2 4 6
= π
[0 + 2 . 2𝑠𝑖𝑛2𝑥 + 0 + 2 . 2𝑠𝑖𝑛4𝑥 + 0 + 2 . 2𝑠𝑖𝑛6𝑥 + ⋯ ⋯ ]
2 −1 4 −1 6 −1

8 𝑠𝑖𝑛2𝑥 2𝑠𝑖𝑛4𝑥 3𝑠𝑖𝑛6𝑥

= ⎡ + + + ⋯ ⋯⋯ ⋯⎤
π ⎣ 3 15 35 ⎦
8 𝑠𝑖𝑛2𝑥 2𝑠𝑖𝑛4𝑥 3𝑠𝑖𝑛6𝑥
= ⎡ + + + ⋯ ⋯⋯ ⋯⎤
π ⎣ 1.3 3.5 5.7 ⎦

Example-20: Find the Fourier series for

𝑓(𝑥) = {− π − π < 𝑥 < 0𝑥 0 <𝑥<π

2
π 1 1 1
Hence deduce 8
= 2 + 2 + 2 + ⋯ ⋯⋯ ⋯
1 3 5

Solution: By definition of the Fourier series

we have ,
π
(
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 ⋯ ⋯⋯ ⋯(1)
𝑛=1
)
Where,
π
1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

59
 π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π

π
1
𝑎𝑛𝑑 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π

Now
π
1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

0 π
1 1
= 2π
∫ 𝑓(𝑥)𝑑𝑥 + 2π
∫ 𝑓(𝑥)𝑑𝑥
−π 0

0 π
1 1
= 2π
∫ (− π)𝑑𝑥 + 2π
∫ 𝑥𝑑𝑥
−π 0

2 π
π 0 1 𝑥
=− 2π
[𝑥]−π + 2π
[ 2
]
0

π π
=− 2
+ 4

π
=− 4

π
1
𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑥 + π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑥
−π 0

0 π
1 1
= π
∫ (− π) cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑥 + π
∫ 𝑥 cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑥
−π 0

π
π 0 1 π 1
=− 𝑛
[sin 𝑠𝑖𝑛 𝑛𝑥 ]−π + 𝑛π
[𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 ]0 − 𝑛π
∫ sin 𝑠𝑖𝑛 𝑛𝑥𝑑𝑥
0

1 π
=0+0+ 2 [cos 𝑐𝑜𝑠 𝑛𝑥 ]0
π𝑛

1
= 2 [cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0 ]
π𝑛

60
 1 𝑛
= 2 [(− 1) − 1] .
π𝑛

Finally,
π
1
𝑏𝑛 = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥𝑑𝑥
−π

0 π
1 1
= π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

0 π
1 1
= π
∫ (− π) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 + π
∫ 𝑥 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
−π 0

π
1 0 1 π 1
= 𝑛
[cos 𝑐𝑜𝑠 𝑛𝑥 ]−π − π𝑛
[𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 ]0 + 𝑛π
∫ cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

1 1 1 π
= 𝑛
[cos 𝑐𝑜𝑠 0 − cos 𝑐𝑜𝑠 𝑛π ] − 𝑛π
[π cos 𝑐𝑜𝑠 𝑛𝑥 ] + 2 [sin 𝑠𝑖𝑛 𝑛𝑥 ]0
π𝑛

1 1
= 𝑛
[1 − cos 𝑐𝑜𝑠 𝑛π ] − 𝑛
cos 𝑐𝑜𝑠 𝑛π + 0

1
= 𝑛
[1 − 2 cos 𝑐𝑜𝑠 𝑛π ]

1 𝑛
= 𝑛
[1 − 2(− 1) ].

Now putting the values of 𝑎0, 𝑎𝑛 𝑎𝑛𝑑 𝑏𝑛 in(1),we get

∞ ∞
𝑓(𝑥) =−
π
4
+ ∑
1
2 {(− 1)𝑛 − 1} cos 𝑐𝑜𝑠 𝑛𝑥 + ∑
1
𝑛
𝑛
{1 − 2(− 1) } sin 𝑠𝑖𝑛 𝑛𝑥
𝑛=1 π𝑛 𝑛=1

+ ⎡⎢− cos 𝑐𝑜𝑠 5𝑥 + ⋯⋯⎤⎥ + ⎡ 1 sin 𝑠𝑖𝑛 𝑥 −

π 2 2 2 3 sin𝑠𝑖𝑛 2𝑥 3
=− cos 𝑐𝑜𝑠 𝑥 + 0 − cos 𝑐𝑜𝑠 3𝑥 + 0 − + sin
4
⎣ π.1
2
π.3
2
π.5
2
⎦ ⎣ 2 3

=−
π
−
2 ⎡ cos𝑐𝑜𝑠 𝑥 + cos𝑐𝑜𝑠 3𝑥
+
cos𝑐𝑜𝑠 5𝑥
+ ⋯⋯⎤⎥ + ⎡3 sin 𝑠𝑖𝑛 𝑥 −
sin𝑠𝑖𝑛 2𝑥
+
3sin𝑠𝑖𝑛 3𝑥
−
sin𝑠𝑖𝑛 4𝑥
+ ⋯
⎢ 12
⎦ ⎣
4 π 2 2 2 3 4
⎣ 3 5

One discontinuity occurs at x=0

+ −
𝑓(0) = ( ) 2 ( ) =− ( +) = 0 𝑎𝑛𝑑 𝑓(0−) = π
𝑓 0 +𝑓 0 π
2
{𝑠𝑖𝑛𝑐𝑒 𝑓 0

61
Putting 𝑥 = 0 in the above series
π 2 1 1 1 π
𝑓(0) =− 4
− π
[ 2 + 2 + 2 + ⋯⋯] =− 2
1 3 5

2 1 1 1 π
𝑂𝑟, − π
[ 2 + 2 + 2 + ⋯⋯] =− 4
1 3 5

2
1 1 1 π
𝑂𝑟, 2 + 2 + 2 + ⋯⋯] =− 8
1 3 5

(Ans).

Example-21: Expand 𝑐𝑜𝑠 µ𝑥 in Fourier series in the interval − π < 𝑥 < π, µ≠0, ±1, ±2, …..

Also show that

∞ 4
1 π
∑ 4 = 90
𝑛=1 𝑛

Solution: (First portion)

The given function is 𝑓(𝑥) = cos 𝑐𝑜𝑠 μ𝑥

∴𝑓(− 𝑥) = cos 𝑐𝑜𝑠 µ(− 𝑥) = cos 𝑐𝑜𝑠 μ𝑥 = 𝑓(𝑥), 𝑠𝑖𝑛𝑐𝑒 cos 𝑐𝑜𝑠 (− θ) = cos 𝑐𝑜𝑠 θ

Thus the given function f(x)=cosµx is an even function and in this case in Fourier series 𝑏𝑛 = 0
and
∞
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 …………(1)
𝑛=1

π π
1 2
𝑊ℎ𝑒𝑟𝑒 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥 𝑎𝑛𝑑 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0 0

π
1
𝑁𝑜𝑤 𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥
0

π
1 1 π sin𝑠𝑖𝑛 μπ
= π
∫ cos 𝑐𝑜𝑠 μ𝑥 𝑑𝑥 = μπ
[sin 𝑠𝑖𝑛 μ𝑥 ]0 = πμ
.
0

π
2
𝑎𝑛𝑑 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

62
 π
2
= π
∫ cos 𝑐𝑜𝑠 μ𝑥 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

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

1 sin𝑠𝑖𝑛 (μ+𝑛)𝑥 sin𝑠𝑖𝑛 (μ−𝑛)𝑥 π

= π
[ μ+𝑛
+ μ−𝑛
]
0

1 sin𝑠𝑖𝑛 (μ+𝑛)π sin𝑠𝑖𝑛 (μ−𝑛)π

= ⎡ + ⎤+ 0
π ⎣ μ+𝑛 μ−𝑛 ⎦

=
1 ⎡ (μ−𝑛)sin𝑠𝑖𝑛 (μ+𝑛) π+(μ+𝑛)sin𝑠𝑖𝑛 (μ−𝑛) π ⎤
π ⎢ 2 2 ⎥
⎣ µ −𝑛 ⎦
1
= 2 2 [(μ − 𝑛){sin 𝑠𝑖𝑛 μπ cos 𝑐𝑜𝑠 𝑛π + cos 𝑐𝑜𝑠 μπ sin 𝑠𝑖𝑛 𝑛π } + (μ + 𝑛){sin 𝑠𝑖𝑛 μπ cos 𝑐𝑜𝑠 𝑛π −
π(µ −𝑛 )

1
= 2 2 [(μ − 𝑛 + μ + 𝑛) sin 𝑠𝑖𝑛 μπ cos 𝑐𝑜𝑠 𝑛π + 0 ]
π(µ −𝑛 )

2μ 𝑛
= 2 2 . sin 𝑠𝑖𝑛 μπ cos 𝑐𝑜𝑠 𝑛π {cos 𝑐𝑜𝑠 𝑛π = (− 1) sin 𝑠𝑖𝑛 𝑛π = 0
π(µ −𝑛 )

𝑛 2μsin𝑠𝑖𝑛 μπ
= (− 1) . 2 2
π(µ −𝑛 )

Putting the values of 𝑎0 𝑎𝑛𝑑 𝑎𝑛 in (1),we have

∞
sin𝑠𝑖𝑛 μπ 𝑛 2μsin𝑠𝑖𝑛 μπ
𝑓(𝑥) = μπ
+ ∑ (− 1) . 2 2 . cos 𝑐𝑜𝑠 𝑛𝑥
𝑛=1 π(µ −𝑛 )

=
sin𝑠𝑖𝑛 μπ
+
2μsin𝑠𝑖𝑛 μπ ⎡− cos𝑐𝑜𝑠 𝑥
+
cos𝑐𝑜𝑠 2𝑥
−
cos𝑐𝑜𝑠 3𝑥
+
cos𝑐𝑜𝑠 4𝑥
− ⋯⋯⎤⎥
μπ π ⎢ 2 2 2 2 2 2 2 2
⎣ µ −1 µ −2 µ −3 µ −4 ⎦

=
sin𝑠𝑖𝑛 μπ
−
2μsin𝑠𝑖𝑛 μπ ⎡ cos𝑐𝑜𝑠 𝑥 − cos𝑐𝑜𝑠 2𝑥
+
cos𝑐𝑜𝑠 3𝑥
−
cos𝑐𝑜𝑠 4𝑥
+ ⋯⋯⎤⎥
μπ π ⎢ µ2−12 2 2 2 2 2 2
⎣ µ −2 µ −3 µ −4 ⎦
Proof: (Second portion)
4 2 2
Let us consider a function 𝑓(𝑥) = 𝑥 − 2π 𝑥 in the Fourier series
∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ) ………(1)
𝑛=1

4 2 2
𝑆𝑖𝑛𝑐𝑒 𝑓(− 𝑥) = (− 𝑥) − 2π (− 𝑥)

63
 4 2 2
= (𝑥) − 2π (𝑥) = 𝑓(𝑥).

So f(x) is an even function and in this case in Fourier Series 𝑏𝑛 = 0,

π π
1 2
𝑎0 = π
∫ 𝑓(𝑥)𝑑𝑥 , 𝑎𝑛 = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥.
0 0

π
1 4 2 2
𝑎0 = π
∫(𝑥 − 2π 𝑥 )𝑑𝑥
0

5 3 π
=
1 ⎡ 𝑥 − 2π2 𝑥 ⎤
π ⎢5 3 ⎥
⎣ ⎦0
5 5
1 ⎡π − 2 π ⎤
= π ⎢ 5 3 ⎥
⎣ ⎦
4 4 4
3π −10π −7π
= 15
= 15

And
π
4 2 2
𝑎𝑛 =
2
π (
∫ 𝑥 − 2π 𝑥 𝐶𝑜𝑠𝑛𝑥𝑑𝑥 )
0

π π
4 2 2 3 2
=
2
𝑛π [(𝑥 − 2π 𝑥 𝑠𝑖𝑛𝑛𝑥 ) ]0 − 2
𝑛π (
∫ 4𝑥 − 4π 𝑥 𝑠𝑖𝑛𝑛𝑥𝑑𝑥 )
0

π
3 2
= 0−
8
𝑛π (
∫ 𝑥 − π 𝑥 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥 )
0

π π
3 2 2 2
=
8
π𝑛
2 [(𝑥 − π 𝑥 𝑐𝑜𝑠𝑛𝑥 ) ]0 − 8
π𝑛 0
2 (
∫ 3𝑥 − π 𝑐𝑜𝑠𝑛𝑥𝑑𝑥 )
π π
2 2
= 0−
π𝑛
8
3 [(3𝑥 − π 𝑠𝑖𝑛𝑛𝑥 ) ]0 + π𝑛 0
8
3 ∫ 6𝑥 𝑠𝑖𝑛𝑛𝑥 𝑑𝑥

π
48 π 48
= 0− 4 [𝑥𝑐𝑜𝑠𝑛𝑥]0 + 4 ∫ 𝑐𝑜𝑠𝑛𝑥 𝑑𝑥
π𝑛 π𝑛 0

48 48 π
=− 4 π cos 𝑐𝑜𝑠 𝑛π + 0 + 5 [sin 𝑠𝑖𝑛 𝑛𝑥 ]0
π𝑛 π𝑛

64
 48 48 𝑛 𝑛+1 48
=− 4 cos 𝑐𝑜𝑠 𝑛π + 0 =− 4 (− 1) = (− 1) 4
𝑛 𝑛 𝑛

Putting the values of 𝑎0, 𝑎𝑛 and 𝑏𝑛 in (1) we get

4 ∞
−7π 𝑛+1 48
𝑓(𝑥) = 15
+ ∑ (− 1) 4 cos 𝑐𝑜𝑠 𝑛𝑥
𝑛=1 𝑛

4
+ 48⎡⎢ 4 − + …⎤⎥
−7π 𝑐𝑜𝑠𝑥 𝑐𝑜𝑠2𝑥 𝑐𝑜𝑠3𝑥 𝑐𝑜𝑠4𝑥
𝑜𝑟, 𝑓(𝑥) = 15 4 + 4 − 4
⎣ 1 2 3 4 ⎦

When 𝑥 = 0, we have 𝑓(0) = 0 and

4
+ 48⎡⎢ 4 − + …⎤⎥
7π 𝑐𝑜𝑠0 𝑐𝑜𝑠0 𝑐𝑜𝑠0 𝑐𝑜𝑠0
0 =− 15 4 + 4 − 4
⎣ 1 2 3 4 ⎦
4
= 48⎡⎢ 4 − + …⎤⎥
7π 1 1 1 1
𝑂𝑟, 15 4 + 4 − 4
⎣1 2 3 4 ⎦
4
𝑂𝑟, ⎡⎢ 4 − + …⎤⎥ =
1 1 1 1 7π
4 + 4 − 4 48×15
⎣1 2 3 4 ⎦
4
𝑂𝑟, ⎡⎢ 4 + + …⎤⎥ − 2⎡⎢ 4 + + …⎤⎥ =
1 1 1 1 1 1 1 7π
4 + 4 + 4 4 + 4 48×15
⎣1 2 3 4 ⎦ ⎣2 4 6 ⎦

( )
4
1 1 1 1 2 1 1 1 1 7π
𝑂𝑟, 4 + 4 + 4 + 4 +… − 4 4 + 4 + 4 + 4 + … = 48×15
1 2 3 4 2 1 2 3 4

)( )
4

(
𝑂𝑟, 1 −
1
8
1
4
1
+
1
4
2
+
3
1
4 +
1
4
4 + … =
7π
48×15

∞ 4
7 1 7π
𝑂𝑟, 8
∑ 4 = 48×15
𝑛=1 𝑛

∞ 4
1 π
𝑂𝑟, ∑ 4 = 90
𝑛=1 𝑛

Example 22: Expand

65
 1 1 3 1
𝑓(𝑥) = { 4 − 𝑥 𝑖𝑓 0 < 𝑥 < 2
𝑥 − 4
𝑖𝑓 2
<𝑥<1

in a fourier series of sine terms only.

Solution:

Since we are interested to expand 𝑓(𝑥) in a Fourier series of sine terms only.Let 𝑓(𝑥)

Represent an odd function in (1, − 1) so that

∞
𝑓(𝑥) = ∑ 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛π𝑥
𝑛=1

Where
𝑐
2 𝑛π𝑥
𝑏𝑛 = 𝑐
∫ 𝑓(𝑥)𝑠𝑖𝑛 𝑐
0

1
= 2 ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛π𝑥𝑑𝑥
0

1
2 1
= 2 ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛π𝑥𝑑𝑥 + 2 ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛π𝑥𝑑𝑥
0 1
2

1
1 2 1
= 2⎡
⎣ ( 1
4
− 𝑥 .− ) 1
𝑛π
𝑐𝑜𝑠𝑛π𝑥⎤ −
⎦0
2 2
𝑛π
∫ 𝑐𝑜𝑠𝑛π𝑥 + 2⎡ 𝑥 −
0
⎣ ( 3
4 ) ⎤1 −
⎦
2
1
𝑛π
𝑐𝑜𝑠
𝑛π
2
+
1
4𝑛π

1 𝑛π 1 2 𝑛π 1 1 𝑛π 2 𝑛π
= 2𝑛π
𝑐𝑜𝑠 2
+ 2𝑛π
− 2 2 𝑠𝑖𝑛 2
− 2𝑛π
𝑐𝑜𝑠𝑛π − 2𝑛π
𝑐𝑜𝑠 2
− 2 2 𝑠𝑖𝑛 2
𝑛π 𝑛π

1 4 𝑛π
= 2𝑛π
(1 − 𝑐𝑜𝑠𝑛π) − 2 2 𝑠𝑖𝑛 2
𝑛π

=
1
2𝑛π (1 − (− 1)𝑛) − 𝑛π
4
2 2 𝑠𝑖𝑛
𝑛π
2

2 4 1 4
∴𝑏1 = 2π
− 2 2 = π
− 2 2
1π 1π

2 4 3π
𝑏2 = 0, 𝑏3 = 2.3π
− 2 2 𝑠𝑖𝑛 2
3π

1 4 3π π π
= 3π
+ 2 2 {𝑠𝑖𝑛 2
= 𝑠𝑖𝑛⁡(π + 2
) =− 𝑠𝑖𝑛 2
=− 1
3π

66
 2 4 5π
𝑏4 = 0, 𝑏3 = 2.5π
− 2 2 𝑠𝑖𝑛 2
5π

=
1
5π
−
5π
4
2 2 𝑎𝑛𝑑 𝑠𝑜 𝑜𝑛 {𝑠𝑖𝑛
5π
2 (
= sin 𝑠𝑖𝑛 2π +
π
2 ) = 1

Thus
𝑓(𝑥) = ( 1
π
−
1π
4
2 2 )𝑠𝑖𝑛π𝑥 + ( 1
3π
−
4
2 2
3π ) sin 𝑠𝑖𝑛 3π𝑥 + ( 1
5π
−
4
2 2
5π )+ …
Example 23.
𝑐 𝑐
Find an even function of x which is equal to λ 𝑥 for 0≤𝑥≤ 2
and is λ(𝑐 − 𝑥) for 2
≤𝑥≤𝑐

Solution

Fourier series for the even function 𝑓(𝑥) in the interval (0, 𝑐) is given by the equation
∞
𝑛π𝑥
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠 𝑐
𝑛=1

Where
𝑐
1
𝑎0 = 𝑐
∫ 𝑓(𝑥)𝑑𝑥
0

And
𝑐
2 𝑛π𝑥
𝑎𝑛 = 𝑐
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑐
𝑑𝑥
0

Now
𝑐
1
𝑎0 = 𝑐
∫ 𝑓(𝑥)𝑑𝑥
0

𝑐
2 𝑐
1 1
= 𝑐
∫ 𝑓(𝑥)𝑑𝑥 + 𝑐
∫ 𝑓(𝑥)𝑑𝑥
0 𝑐
2

𝑐
2 𝑐
1 1
= 𝑐
∫ λ𝑥 𝑑𝑥 + 𝑐
∫ λ(𝑐 − 𝑥)𝑑𝑥
0 𝑐
2

67
 𝑐
2 2 2 𝑐
λ 𝑥 λ (𝑐−𝑥)
= 𝑐
[ 2
] − 𝑐
[ 2
]𝑐
0 2

2 2
λ 𝑐 λ 𝑐
= 𝑐
. 8
− 0 − 0 + 𝑐
. 8

λ𝑐
= 4

𝑐
2 𝑛π𝑥
And 𝑎𝑛= 𝑐 ∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑐
𝑑𝑥
0

𝑐
2 𝑐
2 𝑛π𝑥 2 𝑛π𝑥
= 𝑐
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑐
𝑑𝑥 + 𝑐
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑐
𝑑𝑥
0 𝑐
2

𝑐
2 𝑐
2 𝑛π𝑥 2 𝑛π𝑥
= 𝑐
∫ λ𝑥𝑐𝑜𝑠 𝑐
𝑑𝑥 + 𝑐
∫ λ(𝑐 − 𝑥)𝑐𝑜𝑠 𝑐
𝑑𝑥
0 𝑐
2

𝑐
𝑐 2 𝑐
2λ 𝑐 𝑛π𝑥 2 2λ 𝑐 𝑛π𝑥 2λ 𝑐 𝑛π𝑥 𝑐 2λ 𝑐 𝑛π𝑥 𝑐 2λ 𝑐 𝑛π
= . ⎡𝑥𝑠𝑖𝑛 ⎤ − . ∫ 𝑠𝑖𝑛 𝑑𝑥 + . 𝑐. ⎡𝑠𝑖𝑛 ⎤ − . ⎡𝑥𝑠𝑖𝑛 ⎤ + . ∫ 𝑠𝑖𝑛
𝑐 𝑛π ⎣ 𝑐 ⎦0 𝑐 𝑛π
0
𝑐 𝑐 𝑛π ⎣ 𝑐 ⎦𝑐 𝑐 𝑛π ⎣ 𝑐 ⎦𝑐 𝑐 𝑛π
𝑐
2 2
2

𝑐
2λ 𝑐 𝑛π 2λ 𝑐 𝑛π𝑥 2 2λ 𝑛π 2λ𝑐 2λ 𝑐 𝑛π 2
= . 𝑠𝑖𝑛 −0 + ⎡𝑐𝑜𝑠 ⎤ + sin 𝑠𝑖𝑛 𝑛π − 𝑠𝑖𝑛 − sin 𝑠𝑖𝑛 𝑛π + . 𝑠𝑖𝑛 −
𝑛π 2 2 𝑛π 𝑛π ⎣ 𝑐 ⎦0 𝑛π 2 𝑛π 𝑛π 2 2 𝑛

λ𝑐 𝑛π 2λ𝑐 𝑛π 2λ𝑐 2λ𝑐 𝑛π λ𝑐 𝑛π 2λ𝑐 2λ𝑐 𝑛π

= 𝑛π
𝑠𝑖𝑛 2
+ 2 2 𝑐𝑜𝑠 2
− 2 2 + 0− 𝑛π
𝑠𝑖𝑛 2
− 0+ 𝑛π
𝑠𝑖𝑛 2
− 2 2 cos 𝑐𝑜𝑠 𝑛π + 2 2 𝑐𝑜𝑠 2
𝑛π 𝑛π 𝑛π 𝑛π

4λ𝑐 𝑛π 2λ𝑐 2λ𝑐

= 2 2 𝑐𝑜𝑠 2
− 2 2 cos 𝑐𝑜𝑠 𝑛π − 2 2
𝑛π 𝑛π 𝑛π

=
2λ𝑐
𝑛π
2 2 {2𝑐𝑜𝑠 𝑛π
2
− cos 𝑐𝑜𝑠 𝑛π − 1 }
8λ𝑐
2 2 𝑓𝑜𝑟 𝑛≠4𝑚+2, 𝑚=0, 1, 2, 3,…….
𝑛π
= {0 𝑓𝑜𝑟 𝑛≠4𝑚+2(𝑓𝑜𝑟 𝑎𝑙𝑙 𝑜𝑡ℎ𝑒𝑟 𝑣𝑎𝑙𝑢𝑒𝑠 𝑜𝑓 𝑛)

8λ𝑐
2 2 𝑓𝑜𝑟 𝑛=2, 6, 10, 14
𝑛π
= {0 𝑓𝑜𝑟 𝑛≠4𝑚+2(𝑓𝑜𝑟 𝑎𝑙𝑙 𝑜𝑡ℎ𝑒𝑟 𝑣𝑎𝑙𝑢𝑒𝑠 𝑜𝑓 𝑛)

Thus putting the values of 𝑎0 and 𝑎𝑛 in (i), we get

λ𝑐 8λ𝑐 2π𝑥 8λ𝑐 6π𝑥 8λ𝑐 10π𝑥

𝑓(𝑥) = 4
+0 + 2 2 𝑐𝑜𝑠 𝑐
+ 0+ 2 2 𝑐𝑜𝑠 𝑐
+ 0+ 2 2 𝑐𝑜𝑠 𝑐
+ …
2π 6π 10 π

68
 λ𝑐 8λ𝑐 1 2π𝑥 1 6π𝑥 1 10π𝑥
= 4
+ 2 [ 2 𝑐𝑜𝑠 𝑐
+ 2 𝑐𝑜𝑠 𝑐
+ 2 𝑐𝑜𝑠 𝑐
+ …]
π 2 6 10

Which is the required even function of 𝑥.

π π π
Example 24: If 𝑓(𝑥) = { 1 𝑓𝑜𝑟 0≤𝑥≤ 2
0 𝑓𝑜𝑟 𝑥 = 2
− 1 𝑓𝑜𝑟 2
≤𝑥≤π

4 1 1
Show that 𝑓(𝑥) = π
[cos 𝑐𝑜𝑠 𝑥 − 3
cos 𝑐𝑜𝑠 3𝑥 + 5
cos 𝑐𝑜𝑠 5𝑥 − … ]

Proof:

Since the given interval is [0, π]

Which is the half of the interval [− π, π], the fourier series will be a half of Range fourier cosine
series and in this case
∞
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 𝑏𝑛 = 0 …… (1)
𝑛=1
( )
π π
Where 𝑎0 = ∫ 𝑓(𝑥)𝑑𝑥 and 𝑎𝑛 = ∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0 0

π
Now 𝑎0 = ∫ 𝑓(𝑥)𝑑𝑥
0

π
2 π
1 1
= π
∫ 𝑓(𝑥)𝑑𝑥 + π
∫ 𝑓(𝑥)𝑑𝑥
0 π
2

π
2 π
1 1
= π
∫ 1𝑑𝑥 + π
∫ (− 1)𝑑𝑥
0 π
2

π
1 2 1 π
= π
[𝑥]0 − π
[𝑥] π
2

1 π π
= ⎡2 − 0 − π + ⎤
π ⎣ 2 ⎦
1
= π
.0

= 0

69
 π
And 𝑎𝑛 = ∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

π
2 π
2 2
= π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0 π
2

π
2 π
2 2
= π
∫ 1 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 + π
∫ (− 1) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0 π
2

π
2 1 2 1 π
= π
. 𝑛
[sin 𝑠𝑖𝑛 𝑛𝑥 ]02 − π
. 𝑛
[sin 𝑠𝑖𝑛 𝑛𝑥 ] π
2

2 𝑛π 𝑛π
= π𝑛
[sin 𝑠𝑖𝑛 2
− 0 − sin 𝑠𝑖𝑛 𝑛π + sin 𝑠𝑖𝑛 2
]

4 𝑛π
= π𝑛
sin 𝑠𝑖𝑛 2
+ 0

4 𝑛π
= π𝑛
sin 𝑠𝑖𝑛 2

Now putting the values of 𝑎0 and 𝑎𝑛 in (1) we get

∞
4 𝑛π
𝑓(𝑥) = 0 + ∑ π𝑛
sin 𝑠𝑖𝑛 2
cos 𝑐𝑜𝑠 𝑛𝑥
𝑛=1

∞
4 1 𝑛π
= π
∑ 𝑛
𝑠𝑖𝑛 2
cos 𝑐𝑜𝑠 𝑛𝑥
𝑛=1

π 2π 3π 4π 5π
4 sin𝑠𝑖𝑛 2
sin𝑠𝑖𝑛 2
sin𝑠𝑖𝑛 2
sin𝑠𝑖𝑛 2
sin𝑠𝑖𝑛 2
= π
[ 1
cos 𝑐𝑜𝑠 𝑥 + 2
cos 𝑐𝑜𝑠 2𝑥 + 3
cos 𝑐𝑜𝑠 3𝑥 + 4
cos 𝑐𝑜𝑠 4𝑥 + 5
cos 𝑐𝑜𝑠 5𝑥 +

4 1 (−1) 1
= π
[ 1 cos 𝑐𝑜𝑠 𝑥 + 0 + 3
cos 𝑐𝑜𝑠 3𝑥 + 0 + 5
cos 𝑐𝑜𝑠 5𝑥 + 0 + …]

4 1 1
= π
[cos 𝑐𝑜𝑠 𝑥 − 3
cos 𝑐𝑜𝑠 3𝑥 + 5
cos 𝑐𝑜𝑠 5𝑥 − …]

Example 25: Find the complex form of the Fourier series of the periodic function whose
−𝑥
definition in one period is 𝑓(𝑥) = 𝑒 , − 1 < 𝑥 < 1.

Solution:

70
By the definition of Fourier series of the complex form we have
∞ 𝑖𝑛π𝑥 ∞ 𝑖𝑛π𝑥
𝑐
− 𝑐
𝑓(𝑥) = 𝑐0 + ∑ 𝑐𝑛𝑒 + ∑ 𝑐−𝑛𝑒
𝑛=1 𝑛=1

𝑐
1
Where 𝑐0 = 2𝑐
∫ 𝑓(𝑥)𝑑𝑥
−𝑐

𝑐 𝑖𝑛π𝑥
1 −
𝑐𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑐
𝑑𝑥 And
−𝑐

𝑐 𝑖𝑛π𝑥
1 −−
𝑐−𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑐
𝑑𝑥,
−𝑐

Here in our given problem 𝑐 = 1

∞ ∞
𝑖𝑛π𝑥 −𝑖𝑛π𝑥
𝑓(𝑥) = 𝑐0 + ∑ 𝑐𝑛𝑒 + ∑ 𝑐−𝑛𝑒 …(2)
𝑛=1 𝑛=1

1 1
1 1 −𝑖𝑛π𝑥
where 𝑐0 = 2
∫ 𝑓(𝑥)𝑑𝑥, 𝑐𝑛 = 2
∫ 𝑓(𝑥)𝑒 𝑑𝑥
−1 −1

1
1 𝑖𝑛π𝑥
and 𝑐−𝑛 = 2
∫ 𝑓(𝑥)𝑒 𝑑𝑥
−1

−𝑥
Now given 𝑓(𝑥) = 𝑒
1
1 −𝑥 1 −𝑥 1
∴𝑐0 = 2
∫ 𝑒 𝑑𝑥 =− 2
[𝑒 ]−1
−1

or, 𝑐0 =−
1
2 (𝑒1 − 𝑒−1) = 12 (𝑒−1 − 𝑒1) = sin 𝑠𝑖𝑛 ℎ1
1 1
1 −𝑥 −𝑖𝑛π𝑥 1 −(𝑖𝑛π+1)𝑥
𝑐𝑛 = 2
∫𝑒 . 𝑒 𝑑𝑥 = 2
∫ 𝑒 𝑑𝑥
−1 −1

−(𝑖𝑛π+1)𝑥 1
=
1
2
.
1
−(𝑛π𝑖+1) [
𝑒 ]−1
=−
1
2(𝑛π𝑖+1) [𝑒−(𝑖𝑛π+1) − 𝑒(𝑖𝑛π+1)]

71
 =
1
2(𝑛π𝑖+1) [𝑒(𝑖𝑛π+1) − 𝑒−(𝑖𝑛π+1)]
=
1
2(𝑛π𝑖+1) [𝑒𝑛π𝑖. 𝑒1 − 𝑒−𝑛π𝑖. 𝑒−1]
𝑖π
Now 𝑒 = cos 𝑐𝑜𝑠 π + 𝑖 sin 𝑠𝑖𝑛 π =− 1 + 0 =− 1

𝑛π𝑖 𝑛
∴𝑒 = 𝑒 ( π𝑖) = (− 1) .
𝑛

−𝑖π
Similarly, 𝑒 = 𝑐𝑜𝑥 π − 𝑖 sin 𝑠𝑖𝑛 π =− 1 − 0 =− 1

−𝑛π𝑖 −π𝑖 𝑛 𝑛
∴𝑒 = 𝑒 ( ) = (− 1)

1 1 𝑛 −1 𝑛
Therefore, 𝑐𝑛 = 2(𝑛π𝑖+1 )
[𝑒 . (− 1) − 𝑒 . (− 1) ]

𝑛
(−1) 1 1 −1
= 𝑛π𝑖+1
. 2
(𝑒 − 𝑒 )

𝑛
(−1)
= 𝑛π𝑖+1
sin 𝑠𝑖𝑛 ℎ 1

𝑛
(−1) (1−𝑛π𝑖)
= (1+𝑛π𝑖)(1−𝑛π𝑖)
sinh 𝑠𝑖𝑛ℎ 1

𝑛
(−1) (1−𝑛π𝑖)
= 2 2 sinh 𝑠𝑖𝑛ℎ 1
1+𝑛 π

1
1 −𝑥 𝑖𝑛π𝑥
Again, 𝑐−𝑛 = 2
∫ 𝑒 .𝑒 𝑑𝑥
−1

1
1 (𝑛π𝑖−1)𝑥
= 2
∫ 𝑒 𝑑𝑥
−1

1
=
1
2
.
1
(𝑛π𝑖−1) [𝑒(𝑛π𝑖−1)𝑥]−1
=
1
2(𝑛π𝑖−1) [𝑒(𝑖𝑛π−1) − 𝑒−(𝑖𝑛π−1)]
=
1
2(𝑛π𝑖−1) [𝑒𝑛π𝑖. 𝑒−1 − 𝑒−𝑛π𝑖. 𝑒1]
=
1
2(𝑛π𝑖−1) [(− 1)𝑛. 𝑒−1 − (− 1)𝑛. 𝑒1]

72
 𝑛
(−1) −1 1
= 2(𝑛π𝑖−1)
[𝑒 −𝑒 ]

𝑛
(−1)
= (𝑛π𝑖−1)
. sinh 𝑠𝑖𝑛ℎ 1

𝑛
(−1) (1+𝑛π𝑖)
= 2 2 sinh 𝑠𝑖𝑛ℎ 1
1+𝑛 π

∞ ∞
𝑖𝑛π𝑥 −𝑖𝑛π𝑥
Thus 𝑓(𝑥) = 𝑐𝑜 + ∑ 𝑐𝑛𝑒 + ∑ 𝑐−𝑛𝑒
𝑛=1 𝑛=1

∞ 𝑛 ∞ 𝑛
(−1) (1−𝑛π𝑖) 𝑖𝑛π𝑥 (−1) (1+𝑛π𝑖) −𝑖𝑛π𝑥
= 𝑠𝑖𝑛ℎ1 + ∑ 2 2 sinh 𝑠𝑖𝑛ℎ 1. 𝑒 + ∑ 2 2 sinh 𝑠𝑖𝑛ℎ 1 . 𝑒
𝑛=1 1+𝑛 π 𝑛=1 1+𝑛 π

∞ 𝑛
(−1) 𝑖𝑛π𝑥
or, 𝑓(𝑥) = 𝑠𝑖𝑛ℎ1 + ∑ 2 2 sinh 𝑠𝑖𝑛ℎ 1. 𝑒
𝑛=1 1+𝑛 π
∞ 𝑛 ∞ 𝑛 ∞ 𝑛
(−1) 𝑛π𝑖 𝑖𝑛π𝑥 (−1) −𝑖𝑛π𝑥 (−1) 𝑛π𝑖
− ∑ 2 2 sinh 𝑠𝑖𝑛ℎ 1 . 𝑒 + ∑ 2 2 sinh 𝑠𝑖𝑛ℎ 1. 𝑒 + ∑ 2 2 sinh 𝑠𝑖𝑛ℎ 1 . 𝑒
𝑛=1 1+𝑛 π 𝑛=1 1+𝑛 π 𝑛=1 1+𝑛 π

∞ 𝑛
(−1) 𝑖𝑛π𝑥 −𝑖𝑛π𝑥
= 𝑠𝑖𝑛ℎ1 + 𝑠𝑖𝑛ℎ1 ∑ 2 2 [𝑒 −𝑒 ]
𝑛=1 1+𝑛 π
∞ 𝑛
(−1) 𝑛 𝑖𝑛π𝑥 −𝑖𝑛π𝑥
− π𝑖 𝑠𝑖𝑛ℎ1 ∑ 2 2 [𝑒 −𝑒 ]
𝑛=1 1+𝑛 π

∞ 𝑛 ∞ 𝑛
(−1) 2 (−1) 𝑛
= 𝑠𝑖𝑛ℎ1 + 2𝑠𝑖𝑛ℎ1 ∑ 2 2 cos 𝑐𝑜𝑠 𝑛π𝑥 − 2𝑖 π 𝑠𝑖𝑛ℎ1 ∑ 2 2 sin 𝑠𝑖𝑛 𝑛π𝑥
𝑛=1 1+𝑛 π 𝑛=1 1+𝑛 π

∞ 𝑛 ∞ 𝑛
(−1) (−1) 𝑛
= 𝑠𝑖𝑛ℎ1 + 2𝑠𝑖𝑛ℎ1 ∑ 2 2 cos 𝑐𝑜𝑠 𝑛π𝑥 + 2π 𝑠𝑖𝑛ℎ1 ∑ 2 2 sin 𝑠𝑖𝑛 𝑛π𝑥
𝑛=1 1+𝑛 π 𝑛=1 1+𝑛 π

1 1 1
= sinh 𝑠𝑖𝑛ℎ 1 + 2 sinh 𝑠𝑖𝑛ℎ 1 [− 2 2 . cos 𝑐𝑜𝑠 π𝑥 + 2 2 . cos 𝑐𝑜𝑠 2π𝑥 − 2 2 . cos 𝑐𝑜𝑠 3π𝑥 +
1+1 π 1+2 π 1+3 π
1 2 3 4
+ 2π sinh 𝑠𝑖𝑛ℎ 1 [− 2 2 . sin 𝑠𝑖𝑛 π𝑥 + 2 2 . sin 𝑠𝑖𝑛 2π𝑥 − 2 2 . sin 𝑠𝑖𝑛 3π𝑥 + 2 2 . sin
1+1 π 1+2 π 1+3 π 1+4 π

1 − 2 sinh 𝑠𝑖𝑛ℎ 1 ⎡⎢ +...⎤⎥ −

cos𝑐𝑜𝑠 π𝑥 cos𝑐𝑜𝑠 2π𝑥 cos𝑐𝑜𝑠 3π𝑥 cos𝑐𝑜𝑠 4π𝑥 sin𝑠𝑖𝑛 π𝑥
2 2 − 2 2 + 2 2 − 2 2 2π sinh 𝑠𝑖𝑛ℎ 1 [ 2 2
⎣ 1+1 π 1+2 π 1+3 π 1+4 π ⎦ 1+1 π
.

73
 𝑝 𝑝
Example 26. Expand the function 𝑓(𝑥) = 𝑘𝑥 in the interval − 2
<𝑥< 2
in a complex
Fourier series.

Solution : By definition of Fourier series in complex form we have

∞ 𝑖𝑛π𝑥 ∞ 𝑖𝑛π𝑥
𝑐
− 𝑐
𝑓(𝑥) = 𝑐𝑜 + ∑ 𝑐𝑛𝑒 + ∑ 𝑐−𝑛𝑒 (1)
𝑛=1 𝑛=1

𝑐 𝑖𝑛π𝑥
1 −
where 𝑐𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑐
𝑑𝑥
−𝑐

𝑐 𝑖𝑛π𝑥 𝑐
1 𝑐 1
𝑐−𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑑𝑥 𝑎𝑛𝑑 𝑐0 = 2𝑐
∫ 𝑓(𝑥)𝑑𝑥
−𝑐 −𝑐

𝑝
Here in our giver problem 𝑐 = 2

∞ 𝑖2𝑛π𝑥 ∞ 𝑖2𝑛π𝑥
𝑝
− 𝑝
∴𝑓(𝑥) = 𝑐0 + ∑ 𝑐𝑛𝑒 + ∑ 𝑐−𝑛𝑒 (2)
𝑛=1 𝑛=1

𝑝
2 𝑖2𝑛π𝑥
1 −
where 𝑐𝑛 = 𝑝
∫ 𝑓(𝑥)𝑒 𝑝
𝑑𝑥
𝑝
−2

𝑝
2 𝑖2𝑛π𝑥
1 𝑝
𝑐−𝑛 = 𝑝
∫ 𝑓(𝑥)𝑒 𝑑𝑥
𝑝
− 2

𝑝
2
1
𝑎𝑛𝑑 𝑐𝑜 = 𝑝
∫ 𝑓(𝑥)𝑑𝑥
𝑝
−2

Now given 𝑓(𝑥) = 𝑘𝑥

𝑝
2 𝑖2𝑛π𝑥
1 − 𝑝
∴𝑐𝑛 = 𝑝
∫ 𝑘𝑥 𝑒 𝑑𝑥
𝑝
− 2

𝑝 𝑝
𝑖2𝑛π𝑥 2 2 𝑖2𝑛π𝑥
− −
. ⎡⎢𝑥𝑒 ⎤ +
𝑘 𝑝 𝑝 𝑘 𝑝 𝑝
=− 𝑝
. 2𝑖𝑛π ⎥ 𝑝
. 2𝑖𝑛π
∫ 𝑒 𝑑𝑥
⎣ ⎦− 𝑝 −2
𝑝
2

74
 𝑝
𝑖2𝑛π𝑥 2
−
. ⎡⎢𝑒 ⎤
𝑘 𝑝 −𝑖𝑛π 𝑝 𝑖𝑛π 𝑘𝑝
=− ⎡ 𝑒 + 𝑒 ⎤− 𝑝
⎥
2𝑖𝑛π ⎣ 2 2 ⎦ 2 2 2
4𝑖 𝑛 π ⎣ ⎦− 𝑝
2

=−
𝑘𝑝
4𝑖𝑛π [𝑒𝑖𝑛π + 𝑒−𝑖𝑛π] + 𝑘𝑝
4𝑛 π
2 2 [𝑒−𝑖𝑛π − 𝑒𝑖𝑛π]
=−
2𝑘𝑝
4𝑖𝑛π
𝑐𝑜𝑠𝑛π −
𝑘𝑝
4𝑛 π
2 2 [𝑒𝑖𝑛π − 𝑒−𝑖𝑛π]
𝑘𝑝 𝑘𝑝
=− 2𝑖𝑛π
𝑐𝑜𝑠𝑛π − 2 2 2𝑖𝑠𝑖𝑛 𝑛π
4𝑛 π

𝑘𝑝 𝑛 𝑘𝑝 𝑛 𝑛
=− 2𝑖𝑛π
(− 1) − 0 =− 2𝑖𝑛π
(− 1) 𝑠𝑖𝑛𝑐𝑒{𝑠𝑖𝑛 𝑛π = 0 cos 𝑐𝑜𝑠 𝑛π = (− 1)

𝑝
2 𝑖2𝑛π𝑥
1
Similarly, 𝑐−𝑛 = 𝑝
∫ 𝑓(𝑥)𝑒 𝑝
𝑑𝑥
𝑝
− 2

𝑝
2 𝑖2𝑛π𝑥
1 𝑝
= 𝑝
∫ 𝑘𝑥 𝑒 𝑑𝑥
𝑝
−2

𝑝 𝑝
𝑖2𝑛π𝑥 2 2 𝑖2𝑛π𝑥

=
𝑘
.
𝑝 ⎡𝑥𝑒 𝑝 ⎤ − 𝑘
.
𝑝
∫ 𝑒 𝑝
𝑑𝑥
𝑝 2𝑛π𝑖 ⎢ ⎥ 𝑝 2𝑛π𝑖
⎣ ⎦− 𝑝 −
𝑝
2 2
𝑝
𝑖2𝑛π𝑥 2

=
𝑘
⎡𝑝𝑒
𝑛π𝑖
+
𝑝
𝑒
−𝑛π𝑖
⎤− 𝑘𝑝 ⎡𝑒 𝑝 ⎤
2𝑛π𝑖 ⎣2 2 ⎦ 2 2 2
4𝑖 𝑛 π
⎢ ⎥
⎣ ⎦− 𝑝
2

=
𝑘𝑝
4𝑛π𝑖 [𝑒𝑛π𝑖 + 𝑒−𝑛π𝑖] + 4𝑛 π
𝑘𝑝
2 2 [𝑒𝑛π𝑖 − 𝑒−𝑛π𝑖]
2𝑘𝑝 𝑘𝑝
= 4𝑛π𝑖
𝑐𝑜𝑠𝑛π + 2 2 . 2𝑖𝑠𝑖𝑛 𝑛π
4𝑛 π

𝑘𝑝 𝑘𝑝 𝑛
= 2𝑛π𝑖
𝑐𝑜𝑠𝑛π + 0 = 2𝑛π𝑖
(− 1) ,

𝑝 𝑝
2 2
1 1
𝑎𝑛𝑑 𝑐𝑜 = 𝑝
∫ 𝑓(𝑥)𝑑𝑥 = 𝑝
∫ 𝑘𝑥𝑑𝑥
𝑝 𝑝
− 2
−2

( ) = 0.
2 2 2 2
𝑘 ⎡ 𝑥 ⎤ = 𝑘 𝑝 𝑝
= 𝑝 ⎢ 2 ⎥ 𝑝 2𝑝 4
− 4
∴𝑐𝑜 = 0
⎣ ⎦−
2

75
Now putting the values of 𝑐𝑜, 𝑐𝑛 𝑎𝑛𝑑 𝑐−𝑛 in(2), we get

∞ 𝑖2𝑛π𝑥 ∞ −𝑖2𝑛π𝑥
−𝑘𝑝 𝑛 𝑝 𝑘𝑝 𝑛 𝑝
𝑓(𝑥) = 0 + ∑ 2𝑖𝑛π
(− 1) 𝑒 + ∑ 2𝑖𝑛π
(− 1) 𝑒
𝑛=1 𝑛=1

∞ 𝑖2𝑛π𝑥 ∞ −𝑖2𝑛π𝑥
𝑘𝑝 ⎡ −(−1)
𝑛
𝑝 (−1)
𝑛
𝑝 ⎤
= ⎢∑ 𝑒 + ∑ 𝑒 ⎥
2𝑖π ⎢𝑛=1 𝑛 𝑛 ⎥
⎣ 𝑛=1 ⎦
∞ 𝑛 𝑖2𝑛π𝑥 −𝑖2𝑛π𝑥

=−
𝑘𝑝
∑
(−1) ⎡𝑒 𝑝
−𝑒 𝑝 ⎤
2𝑖π 𝑛 ⎢ ⎥
𝑛=1 ⎣ ⎦
∞ 𝑛
𝑘𝑝 (−1) 2𝑛π𝑥
=− 2𝑖π
∑ 𝑛
2𝑖𝑠𝑖𝑛 𝑝
𝑛=1

∞ 𝑛
𝑘𝑝 (−1) 2𝑛π𝑥
=− π
∑ 𝑛
𝑠𝑖𝑛 𝑝
𝑛=1

𝑘𝑝 2π𝑥 1 4π𝑥 1 6π𝑥 1 8π𝑥

=− ⎡− 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − …⎤
π ⎣ 𝑝 2 𝑝 3 𝑝 4 𝑝 ⎦
𝑘𝑝 2π𝑥 1 4π𝑥 1 6π𝑥 1 8π𝑥
= ⎡𝑠𝑖𝑛 − 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − 𝑠𝑖𝑛 + …⎤
π ⎣ 𝑝 2 𝑝 3 𝑝 4 𝑝 ⎦

Example 27. Show that

∞
cos𝑐𝑜𝑠 𝑢𝑥 π −𝑥
∫ 2 𝑑𝑢 = 2
𝑒 ,𝑥 > 0
0 1+𝑢

−𝑥 𝑥
proof: Let 𝑓(𝑥) = {𝑒 , 𝑥 > 0 𝑒 , 𝑥 < 0
𝑥 −𝑥
Then 𝑓(− 𝑥) = {𝑒 , 𝑥 < 0 𝑒 , 𝑥 > 0

Therefore, 𝑓(𝑥) = 𝑓(− 𝑥). Thus 𝑓(𝑥) is an even function. By definition of Fourier integral of
even function for 𝑥 > 0 we have
∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑢 ∫ cos 𝑐𝑜𝑠 𝑢𝑡 cos 𝑐𝑜𝑠 𝑢𝑥𝑑𝑡 (1)
0 0

76
 −𝑥
Putting 𝑓(𝑥) = 𝑒 in (1), we get
∞ ∞
−𝑥 −𝑡
𝑒 = ∫ 𝑑𝑢 ∫ 𝑒 cos 𝑐𝑜𝑠 𝑢𝑡 cos 𝑐𝑜𝑠 𝑢𝑥𝑑𝑡
0 0

∞ ∞
2 −𝑡
= π
∫ cos 𝑐𝑜𝑠 𝑢𝑥𝑑𝑢 ∫ 𝑒 cos 𝑐𝑜𝑠 𝑢𝑡 𝑑𝑡 (2)
0 0

∞ ∞
−𝑡 −𝑡 ∞ −𝑡
Now ∫ 𝑒 cos 𝑐𝑜𝑠 𝑢𝑡 𝑑𝑡 = [− cos 𝑐𝑜𝑠 𝑢𝑡 𝑒 ]0 − ∫ 𝑢𝑠𝑖𝑛 𝑢𝑡 𝑒 𝑑𝑡
0 0

∞
−𝑡 ∞ 2 −𝑡
= 0 + 1 + [𝑢𝑠𝑖𝑛 𝑢𝑡 𝑒 ]0 − 𝑢 ∫ 𝑒 cos 𝑐𝑜𝑠 𝑢𝑡𝑑𝑡
0

∞
2 −𝑡
or, (1 + 𝑢 ) ∫ 𝑒 𝑐𝑜𝑠𝑢𝑡 = 1 + 0 = 1
0

∞
−𝑡 1
or, ∫ 𝑒 cos 𝑐𝑜𝑠 𝑢𝑡 𝑑𝑡 = 2
0 1+𝑢

∞
−𝑡
Substituting the value of ∫ 𝑒 cos 𝑐𝑜𝑠 𝑢𝑡𝑑𝑡 in (2)
0

∞
−𝑥 2 1
we get 𝑒 = π
∫ cos 𝑐𝑜𝑠 𝑢𝑥𝑑𝑢. 2
0 1+𝑢

∞
π −𝑥 cos𝑐𝑜𝑠 𝑢𝑥
or, 2
𝑒 =∫ 2 𝑑𝑢
0 1+𝑢

∞
cos𝑐𝑜𝑠 𝑢𝑥 π −𝑥
or, ∫ 2 𝑑𝑢 = 2
𝑒 (3)
0 1+𝑢

When 𝑥 = 0, the above result in (3) reduces to

∞
𝑑𝑢 −𝑥
∫ 2 =𝑒
0 1+𝑢

3.14 Some physical applications of Fourier series.

77
(A) Square wave (High frequencies)

The analysis of square wave in terms of Fourier components, may occur in electronic
circuits designed to handle sharply rising pulses. Let us define the square wave by the function
𝑓(𝑥) such that

𝑓(𝑥) = {− 0, − π < 𝑥 < 0 ℎ, 0 < 𝑥 < π

Which can be expressed by the Fourier series

∞
𝑓(𝑥) = 𝑎0 + ∑ (𝑎𝑛 cos 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 sin 𝑠𝑖𝑛 𝑛𝑥 ) (1)
𝑛=1

∞
1 ℎ π ℎ
Where 𝑎0 = 2π
∫ ℎ𝑑𝑡 = 2π
[𝑡]0 = 2
,
0

π
1
𝑎𝑛 = π
∫ ℎ cos 𝑐𝑜𝑠 𝑛𝑡𝑑𝑡
0

ℎ π
= π𝑛
[sin 𝑠𝑖𝑛 𝑛𝑡] 0

ℎ
= π𝑛
[sin 𝑠𝑖𝑛 𝑛π − 𝑠𝑖𝑛0] = 0

π
1
and 𝑏𝑛 = π
∫ ℎ sin 𝑠𝑖𝑛 𝑛𝑡𝑑𝑡
0

ℎ π
=− π𝑛
[cos 𝑐𝑜𝑠 𝑛𝑡 ]0

ℎ
=− π𝑛
[cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0]

ℎ
= π𝑛
[1 − cos 𝑐𝑜𝑠 𝑛π]

ℎ 𝑛 𝑛
= π𝑛
[1 − (− 1) [cos 𝑐𝑜𝑠 𝑛π = (− 1) ]

Now putting the values of 𝑎0, 𝑎𝑛 and 𝑏𝑛 in (1), we get

∞
𝑓(𝑥) =
ℎ
2
+ ∑ [0 +
ℎ
π𝑛 {1 − (− 1)𝑛} sin 𝑠𝑖𝑛 𝑛𝑥 ]
𝑛=1

ℎ ℎ 2 2 2
= 2
+ π
[ 1 sin 𝑠𝑖𝑛 𝑛𝑥 + 0 + 3
sin 𝑠𝑖𝑛 3𝑥 + 0 + 5
sin 𝑠𝑖𝑛 5𝑥 +... ]

78
 ℎ 2ℎ sin𝑠𝑖𝑛 𝑥 sin𝑠𝑖𝑛 3𝑥 sin𝑠𝑖𝑛 5𝑥
= 2
+ π
[ 1
+ 3
+ 5
+...]

(B) Full wave Rectifier

We can answer by the Fourier series analysis of the question that how well the out put
of a full wave rectifier approaches pure direct current. The rectifier which may be thought of as
having passed the positive peaks of an incoming sine wave and inverting the negative peaks,
yields the function

𝑓(𝑡) = {sin 𝑠𝑖𝑛 ω𝑡 , 0 < ω𝑡 < π − sin 𝑠𝑖𝑛 ω𝑡 , − π < ω𝑡 < 0

Here 𝑓(𝑡) is an even function and so Fourier series is a Fourier cosine series
∞
𝑓(𝑡) = 𝑎0 + ∑ 𝑎𝑛 cos 𝑐𝑜𝑠 𝑛 ω𝑡 (1)
𝑛=1

π
1
for which 𝑎0 = π
∫ sin 𝑠𝑖𝑛 ω𝑡 𝑑(ω𝑡)
0

1 π
=− π
[cos 𝑐𝑜𝑠 ω𝑡 ]0

1
=− π
[cos 𝑐𝑜𝑠 π − cos 𝑐𝑜𝑠 0 ]

1
=− π
[− 1 − 1]

2
= π

π
2
and 𝑎𝑛 = π
∫ sin 𝑠𝑖𝑛 ω𝑡𝑛 ω𝑡 𝑑(ω𝑡)
0

π
1
= π
∫ 2 cos 𝑐𝑜𝑠 𝑛 ω𝑡 sin 𝑠𝑖𝑛 ω𝑡 cos 𝑐𝑜𝑠 ω𝑡 𝑑(ω𝑡)
0

π
1
= π
∫[sin 𝑠𝑖𝑛 (𝑛 + 1)ω𝑡 − sin 𝑠𝑖𝑛 (𝑛 − 1) ω𝑡𝑑(ω𝑡)
0

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

79
 1 1 1
=− π
[ 𝑛+1 {cos 𝑐𝑜𝑠 (𝑛 + 1)π − 𝑐𝑜𝑠0} − 𝑛−1
{cos 𝑐𝑜𝑠 (𝑛 − 1) π − 𝑐𝑜𝑠0}]

1 1 1
= π
[ 𝑛+1 {− 𝑐𝑜𝑠𝑛π − 𝑐𝑜𝑠0} − 𝑛−1
{− cos 𝑐𝑜𝑠 𝑛π − cos 𝑐𝑜𝑠 0}]

=
1
π { 1
𝑛+1
−
1
𝑛−1 }{cos 𝑐𝑜𝑠 𝑛π + cos 𝑐𝑜𝑠 0}
𝑛 𝑛]
=−
π(𝑛 −1)
{ (− 1) + 1}
2
2 [cos 𝑐𝑜𝑠 𝑛π = (− 1)

Now putting the value of 𝑎0 and 𝑎𝑛 in (1) we get

∞
𝑛
𝑓(𝑡) =
2
π
+ ∑−
2
2
π(𝑛 −1)
{(− 1) + 1} cos 𝑐𝑜𝑠 𝑛 ω𝑡
𝑛=1

2 4 4
= π
+ [0 − 2 cos 𝑐𝑜𝑠 2ω𝑡 + 0 − 2 cos 𝑐𝑜𝑠 4ω𝑡 + 0 − …]
(
π 2 −1 ) (
π 4 −1 )
2 4 cos𝑐𝑜𝑠 2ω𝑡 cos𝑐𝑜𝑠 4ω𝑡
= π
− π
[ 3
+ 15
+...]

Where the original frequency ω has been eliminated and the lowest frequency of
−2
oscillation is 2ω . The high frequency components fall off as 𝑛 , showing that the full wave
rectifier does a fairly good job of approximating the direct current.

(C)Expansion of Riemann zeta function

2
Let 𝑓(𝑥) = 𝑥 ,− π < 𝑥 < π

This is an even function and so we have the Fourier cosine series

∞
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛 𝑐𝑜𝑠𝑛𝑥 ..... (1) (all sine terms vanish)
𝑛=1

π 3 2
1 2 1
where 𝑎0 = ∫ 𝑥 𝑑𝑥 = ⎡𝑥 ⎤π = π
π π ⎢ 3 ⎥0 3
0 ⎣ ⎦

π
2 2
and 𝑎𝑛 = π
∫ 𝑥 𝑐𝑜𝑠𝑛𝑥𝑑𝑥
0

π
2
=
2
𝑛π [𝑥 𝑠𝑖𝑛𝑛𝑥] π
0
−
2
𝑛π
∫ 2𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

80
 π
4 π 4
= 0+ 2 [𝑥𝑐𝑜𝑠𝑛𝑥] 0 − 2 ∫ 𝑐𝑜𝑠𝑛𝑥𝑑𝑥
π𝑛 π𝑛 0

4 4 π
= 2 π𝑐𝑜𝑠𝑛π − 3 [𝑠𝑖𝑛𝑛𝑥] 0
π𝑛 π𝑛

4 𝑛
= 2 (− 1) − 0
𝑛

4 𝑛
= 2 (− 1)
𝑛

Now putting the values of 𝑎0 and 𝑎𝑛 in (1) we get

2 ∞
2 π 4 𝑛
𝑓(𝑥) = 𝑥 = 3
+ ∑ 2 (− 1) 𝑐𝑜𝑠𝑛𝑥
𝑛=1 𝑛

2
2 π 4 4 4
or, 𝑥 = 3
+ [− 2 𝑐𝑜𝑠𝑥 + 2 𝑐𝑜𝑠2𝑥 − 2 𝑐𝑜𝑠3𝑥 +...]
1 2 3

2
− 4⎡⎢ 2 − − …⎤⎥………(2)
π 𝑐𝑜𝑠𝑥 𝑐𝑜𝑠2𝑥 𝑐𝑜𝑠3𝑥
= 3 2 + 2
⎣ 1 2 3 ⎦
Putting 𝑥 = π in (2) we get
2
2
− 4⎡⎢ 2 − − …⎤⎥
π 𝑐𝑜𝑠π 𝑐𝑜𝑠2π 𝑐𝑜𝑠3π
π = 3 2 + 2
⎣ 1 2 3 ⎦
2
− 4⎡⎢ 2 − − …⎤⎥
π −1 1 1
= 3 2 − 2
⎣1 2 3 ⎦
2
2π 1 1 1
or, 3
= 4[ 2 + 2 + 2 +...]
1 2 3

2 ∞
π 1 1 1 1
or, 6
= 2 + 2 + 2 +...= ∑ 2 = ζ(2)
1 2 3 𝑛=1 𝑛

Which ζ(2) is called the Riemann zeta function.

4. 10: Finite Fourier sine and cosine transforms

The finite Fourier sine transforms of 𝐹(𝑥), (0 < 𝑥 < 𝑙) is defined as

81
 𝑙
𝑛π𝑥
𝑓𝑠(𝑛) = ∫ 𝐹(𝑥) 𝑠𝑖𝑛⁡ 𝑙
𝑑𝑥 (𝑖)
0

Where n is an integer.

The function 𝐹(𝑥) is then called the inverse finite Fourier sine transforms of 𝑓𝑠(𝑛) and is given
by
∞
2 𝑛π𝑥
𝐹(𝑥) = 𝑙
∑ 𝑓𝑠(𝑛) sin 𝑠𝑖𝑛 𝑙
(𝑖𝑖)
𝑛=1

The finite Fourier cosine transforms of 𝐹(𝑥), (0 < 𝑥 < 𝑙) is defined as

𝑙
𝑛π𝑥
𝑓𝑐(𝑛) = ∫ 𝐹(𝑥) 𝑐𝑜𝑠⁡ 𝑙
𝑑𝑥 (𝑖𝑖𝑖)
0

Where n is an integer.

The function 𝐹(𝑥) is then called the inverse finite Fourier cosine transforms of 𝑓𝑐(𝑛) and is
given by
∞
1 2 𝑛π𝑥
𝐹(𝑥) = 𝑙
𝑓𝑐(0) + 𝑙
∑ 𝑓𝑐(𝑛) cos 𝑐𝑜𝑠 𝑙
(𝑖𝑣)
𝑛=1

4. 11: Infinite Fourier sine and cosine transforms

The Infinite Fourier sine transform of a function 𝐹(𝑥) of 𝑥 such that 𝑜 < 𝑥 < ∞ is denoted by
𝑓𝑠(𝑛) and is defined as

∞
𝑓𝑠(𝑛) = ∫ 𝐹(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 (𝑖)
0

The function 𝐹(𝑥) is then called the inverse finite Fourier sine transforms of 𝑓𝑠(𝑛) and is given
by
∞
2
𝐹(𝑥) = π
∫ 𝑓𝑠(𝑛) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 (𝑖𝑖)
0

82
The Infinite Fourier cosine transform of a function 𝐹(𝑥) of 𝑥 such that 𝑜 < 𝑥 < ∞ is denoted
by 𝑓𝑐(𝑛) and is defined as

∞
𝑓𝑐(𝑛) = ∫ 𝐹(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 (𝑖𝑖𝑖)
0

The function 𝐹(𝑥) is then called the inverse finite Fourier cosine transforms of 𝑓𝑐(𝑛) and is
given by
∞
2
𝐹(𝑥) = π
∫ 𝑓𝑐(𝑛) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 (𝑖𝑣)
0

Note 1: The Infinite Fourier sine transform and the Infinite Fourier cosine transform are
generally known as Fourier sine transform and Fourier cosine transform respectively.

Note 2: some authors also define Fourier sine transform and Fourier cosine transform in the
following ways respectively:
∞
2
(𝑖)𝐹𝑠{𝑓(𝑥)} = 𝑓𝑠(𝑛) = π
∫ 𝑓(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
0

∞
2
(𝑖𝑖)𝐹𝑐{𝑓(𝑥)} = 𝑓𝑐(𝑛) = π
∫ 𝑓(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥
0

Note 3: some authors also define inverse Fourier sine transform and Fourier cosine transform in
the following ways respectively:
∞
2
(𝑖)𝑓(𝑥) = π
∫ 𝑓𝑠(𝑛) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑛
0

∞
2
(𝑖𝑖)𝑓(𝑥) = π
∫ 𝑓𝑐(𝑛) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑛
0

4. 12: Complex form of the Fourier integral and Fourier transforms

Form the definition of Fourier integral, we have

83
 ∞ ∞
1
𝑓(𝑥) = 2π
∫ 𝑓(𝑡)𝑑𝑡 ∫ cos 𝑐𝑜𝑠 𝑢(𝑥 − 𝑡) 𝑑𝑢 (𝑖)
−∞ −∞

∞ ∞
1 ⎡ ⎤
= ∫ ⎢ ∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 {𝑢(𝑥 − 𝑡) }𝑑𝑡⎥𝑑𝑢 (𝑖𝑖)
2π ⎢
−∞⎣−∞
⎥
⎦

One can easily show that

∞ ∞
1 ⎡ ⎤
∫ ⎢ ∫ 𝑓(𝑡) sin 𝑠𝑖𝑛 {𝑢(𝑥 − 𝑡) }𝑑𝑡⎥𝑑𝑢 = 0 (𝑖𝑖𝑖)
2π ⎢
−∞⎣−∞
⎥
⎦

Adding (𝑖𝑖) and (𝑖𝑖𝑖) we get

∞ ∞
1 ⎡ ⎤
𝑓(𝑥) = ∫ ⎢ ∫ 𝑓(𝑡)𝑢(𝑥 − 𝑡) + 𝑖 sin 𝑠𝑖𝑛 𝑢(𝑥 − 𝑡)} 𝑑𝑡⎥𝑑𝑢
2π ⎢
−∞⎣−∞
⎥
⎦

∞ ∞
1 ⎡ 𝑖 𝑢(𝑥−𝑡) ⎤
= ∫ ⎢ ∫ 𝑓(𝑡)𝑒 𝑑𝑡⎥𝑑𝑢 (𝑖𝑣)
2π ⎢
−∞⎣−∞
⎥
⎦

This is called the complex form of the Fourier integral

Also (𝑖𝑣) can be written as

∞ ∞
1 ⎡ 1 −𝑖 𝑢𝑡 ⎤ 𝑖𝑢𝑥
𝑓(𝑥) = ∫⎢ ∫ 𝑓(𝑡)𝑒 𝑑𝑡⎥𝑒 𝑑𝑢
2π ⎢
−∞⎣
2π ⎥
−∞ ⎦
∞
1 𝑖𝑢𝑥
= ∫ 𝐶(𝑢)𝑒 𝑑𝑢
2π
−∞

∞
1 −𝑖 𝑢𝑡
Where 𝐶(𝑢) = ∫ 𝑓(𝑡)𝑒 𝑑𝑡
2π
−∞

𝑓(𝑥) is called the Fourier transform of 𝐶(𝑢) and 𝐶(𝑢) is called the inverse Fourier transform of
𝑓(𝑥).

Alternative forms of Fourier transform

∞
−𝑖𝑢𝑥
(A) if 𝐹(𝑢) = ∫ 𝑓(𝑥)𝑒 𝑑𝑥 (1)
−∞

84
 ∞
1 𝑖𝑢𝑥
Then 𝑓(𝑥) = 2π
∫ 𝐹(𝑢)𝑒 𝑑𝑢 (2)
−∞

The function 𝐹(𝑢) is called the Fourier transform of 𝑓(𝑥) and is sometimes written as
𝐹(𝑢) = 𝐹{𝑓(𝑥)}

The function 𝑓(𝑥) is called the inverse Fourier transform of 𝐹(𝑢) and is sometimes written as
−1
𝑓(𝑥) = 𝐹 {𝐹(𝑢)}

∞
𝑖𝑢𝑥
(B) if 𝐹(𝑢) = ∫ 𝑓(𝑥)𝑒 𝑑𝑥.....................(1)
−∞

∞
−𝑖𝑢𝑥
Then 𝑓(𝑥) = ∫ 𝐹(𝑢)𝑒 𝑑𝑢…………...(2)
−∞

Then function 𝐹(𝑢) is called the Fourier transform of 𝑓(𝑥) and is sometimes written as

𝐹(𝑢) = 𝐹{(𝑥)}.

Then function 𝑓(𝑥) is called the inverse Fourier transform of 𝑓(𝑢)and is sometimes written as
−1
𝐹(𝑥) = 𝐹 {(𝑢)}

∞
1 −𝑖𝑢𝑥
(C) If 𝑓(𝑥) = ∫ 𝐶(𝑢)𝑒 𝑑𝑡 .....(1)
2π
−∞

∞
1 𝑖𝑢𝑥
Then 𝐶(𝑢) = ∫ 𝑓(𝑡)𝑒 𝑑𝑡…………...(2)
2π
−∞

Then function 𝑓(𝑥)is called the Fourier transform of 𝐶(𝑢)and 𝐶(𝑢) is called the inverse
Fourier transform of 𝑓(𝑥).

Convolution and Convolution theorem :

Definition Convolution:

85
 ∞
1
The function 𝐻(𝑥) = 𝐹 * 𝐺 = ∫ 𝑓(𝑢)𝐺(𝑥 − 𝑢)𝑑𝑢
2π
−∞

is called the Convolution or Faltung of two integrable functions F and G over the interval
(− ∞, ∞).

The Convolution (or Faltung) theorem for Fourier transforms:

Statement: If F{𝑓(𝑥)} and 𝐹{𝑔(𝑥)} are the Fourier transforms of the function 𝑓(𝑥)and 𝑔(𝑥)
respectively then the Fourier transforms of the function convolution of 𝑓(𝑥) and 𝑔(𝑥) is the
product of their Fourier transforms.

i e. 𝐹(𝑓(𝑥)} * 𝑔(𝑥) = 𝐹{𝑓(𝑥). 𝐹{𝑔(𝑥)}

Proof:

By definitionof convolution of two functions 𝑓(𝑥) and 𝑔(𝑥),we have

∞
1
𝑓(𝑥) * {𝑔(𝑥) = ∫ 𝑓(𝑢)𝑔(𝑥 − 𝑢)𝑑𝑢
2π
−∞

∞
1
∴ 𝐹{𝑓(𝑥) * 𝑔(𝑥)} = 𝐹{ ∫ 𝑓(𝑢)𝑔(𝑥 − 𝑢)𝑑𝑢}
2π
−∞

∞ ∞
1 ⎡ 1 ⎤ 𝑖𝑝𝑥
= ∫⎢ ∫ 𝑓(𝑢)𝑔(𝑥 − 𝑢)𝑑𝑢⎥𝑒 𝑑𝑥
2π ⎢
−∞⎣
2π ⎥
−∞ ⎦
∞ ∞
1 ⎡ 𝑖𝑝𝑥 ⎤
= ∫ 𝑓(𝑢)⎢ ∫ 𝑔(𝑥 − 𝑢)𝑒 𝑑𝑥⎥du
2π ⎢−∞ ⎥
−∞ ⎣ ⎦
∞ ∞
1 ⎡ 𝑖𝑝(𝑥−𝑢) 𝑖𝑝𝑢 ⎤
= ∫ 𝑓(𝑢)⎢ ∫ 𝑔(𝑥 − 𝑢)𝑒 . 𝑒 𝑑(𝑥 − 𝑢)⎥du
2π ⎢−∞ ⎥
−∞ ⎣ ⎦
∞ ∞
1 ⎡ 𝑖𝑝𝑢 𝑖𝑝𝑦 ⎤
= ∫ 𝑓(𝑢)⎢𝑒 ∫ 𝑔(𝑦)𝑒 . 𝑑𝑦⎥du where x-u=y
2π ⎢ ⎥
−∞ ⎣ −∞ ⎦
∞ ∞
1 ⎡ 𝑖𝑝𝑢 1 𝑖𝑝𝑥 ⎤
= ∫ 𝑓(𝑢)⎢𝑒 ∫ 𝑔(𝑥)𝑒 𝑑𝑥⎥du
2π ⎢ 2π ⎥
−∞ ⎣ −∞ ⎦

86
 ∞
𝑖𝑝𝑢
=
1
2π
∫ 𝑓(𝑢) 𝑒 [ 𝐹{𝑔(𝑥)} du]
−∞

∞
⎡ 1 𝑖𝑝𝑢 ⎤
=⎢ ∫ 𝑓(𝑢)𝑒 𝑑𝑢⎥ 𝐹{𝑔(𝑥)}
⎢ 2π ⎥
⎣ −∞ ⎦
∞
⎡ 1 𝑖𝑝𝑥 ⎤
=⎢ ∫ 𝑓(𝑥)𝑒 𝑑𝑥⎥ 𝐹{𝑔(𝑥)}
⎢ 2π ⎥
⎣ −∞ ⎦
= 𝐹{𝑓(𝑥). 𝐹{𝑔(𝑥)}

Hence 𝐹(𝑓(𝑥)} * 𝑔(𝑥) = 𝐹{𝑓(𝑥). 𝐹{𝑔(𝑥)}

Alternative form of Definition of convolution:

The convolution of the two functions 𝑓(𝑥) and 𝑔(𝑥) is defined by

∞
𝑓 * 𝑔 = ∫ 𝑓(𝑢)𝑔(𝑥 − 𝑢)𝑑𝑢
−∞

Alternative proof of the convolution theorem:

By definition of Fourier transform,we have

∞ ∞
−𝑖𝑢𝑡 𝑑𝑡 −𝑖𝑢𝑠 𝑑𝑠
𝐹(𝑢) = ∫ 𝑓(𝑡)𝑒 𝑎𝑛𝑑 𝐺(𝑢) = ∫ 𝑔(𝑠)𝑒 }……………….…….……(1)
−∞ −∞

∞ ∞
−𝑖𝑢(𝑡+𝑠)
Then 𝑓(𝑢)𝑔(𝑢) = ∫ ∫ 𝑓(𝑡)𝑔(𝑠)𝑒 𝑑𝑡 𝑑𝑠…......(2)
−∞ −∞

Let 𝑡 + 𝑠 = 𝑥in the double integral (2) which we wish to transform from the variables (𝑡, 𝑠) to
the variables (𝑡, 𝑠)From advanced calcus,we have
∂(𝑡,𝑠)
𝑑𝑡𝑑𝑠 = ∂(𝑡,𝑥)
𝑑𝑡 𝑑𝑠…………………………………..….…..(3)

Where the Jacobian of the transformation is given by

∂𝑡 ∂𝑡
∂(𝑡,𝑠) | | 1 0
∂(𝑡,𝑥)
𝑑𝑡 𝑑𝑠 = | ∂𝑡
∂𝑠
∂𝑥
∂𝑠 | = || 1 1
|= 1
|
| ∂𝑡 ∂𝑥 |

87
 ∞ ∞
−𝑖𝑢𝑥
Then (2) becomes 𝐹(𝑢)𝐺(𝑢) = ∫ ∫ 𝑓(𝑡)𝑔(𝑥 − 𝑡)𝑒 𝑑𝑡 𝑑𝑠
−∞ −∞

∞ ∞
−𝑖𝑢𝑥
= ∫ 𝑒 [ ∫ 𝑓(𝑡)𝑔(𝑥 − 𝑡)𝑑𝑡] 𝑑𝑥
−∞ −∞

⎰∞ ⎱
= ∫ 𝑓(𝑡)𝑔(𝑥 − 𝑡)𝑑𝑡
⎱−∞ ⎰

∞
= 𝐹(𝑓 * 𝑔) 𝑆𝑖𝑛𝑐𝑒 𝑓 * 𝑔 = ∫ 𝑓(𝑡)𝑔(𝑥 − 𝑡)𝑑𝑡 is the convolution of f and g.
−∞

Hence 𝐹(𝑓 * 𝑔) = 𝐹(𝑢)𝐺(𝑢) = 𝐹(𝑓. 𝐺(𝑔)

Parseval’s identity for Fourier transforms (Rayleigh’s theorem or plancherel’s theorem)

Statement: If 𝐹(𝑢)and 𝐺(𝑢) are complex for Fourier transforms of 𝑓(𝑥) and 𝑔(𝑥)
respectively,then
∞ ∞
1
(i) 2π
∫ 𝐹(𝑢)𝐺(𝑢) 𝑑𝑢 = ∫ 𝑓(𝑥)𝑔(𝑥) 𝑑𝑥
−∞ −∞
∞ ∞
1 2 2
(ii) 2π
∫ |𝐹(𝑢)| 𝑑𝑢 = ∫ |𝑓(𝑥)| 𝑑𝑥
−∞ −∞

Where bar denotes the complex conjugate.

Proof: (i) Using the inverse Fourier transformation. We have

∞
1 𝑖𝑢𝑥
𝑔(𝑥) = 2π
∫ 𝐺(𝑢)𝑒 𝑑𝑢……………(1)
−∞

Taking complex conjugates on both sides of (1) , we get

∞
1 −𝑖𝑢𝑥
𝑔(𝑥) = 2π
∫ 𝐺(𝑢)𝑒 𝑑𝑢…….. (2)
−∞

∞ ∞ ∞
1 1 −𝑖𝑢𝑥
2π
∫ 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = ∫ 𝑓(𝑥)[ 2π
∫ 𝐺(𝑢)𝑒 𝑑𝑢]𝑑𝑥
−∞ −∞ −∞

∞ ∞
1 −𝑖𝑢𝑥
= 2π
∫ 𝐺(𝑢)[ ∫ 𝑓(𝑥)𝑒 𝑑𝑥]𝑑𝑢
−∞ −∞

88
 ∞
1
= 2π
∫ 𝐺(𝑢)𝐹(𝑢)𝑑𝑢
−∞

∞ −𝑖𝑢𝑥

Since 𝐹(𝑢) = ∫ 𝑓(𝑥)𝑒 𝑑𝑥

−∞

∞
1
= 2π
∫ 𝐹(𝑢)𝐺(𝑢)𝑑𝑢
−∞

∞ ∞
1
Hence 2π
∫ 𝐺(𝑢)𝐹(𝑢)𝑑𝑢 = ∫ 𝑓(𝑥)𝑔(𝑥)𝑑𝑥
−∞ −∞

∞ ∞
Proof: (ii) From part (i) we have ∫ 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = ∫ 𝑓(𝑥)𝑔(𝑥)𝑑𝑥……………(3)
−∞ −∞

Taking 𝑔(𝑥) = 𝑓(𝑥) in (3) , we get

∞ ∞
1
2π
∫ 𝐹(𝑢 𝐺(𝑢))𝑑𝑢 = ∫ 𝑓(𝑥)𝑓(𝑥)𝑑𝑥
−∞ −∞

Since 𝑔(𝑥) = 𝑓(𝑥) ⇒ 𝐺(𝑢) = 𝐹(𝑢)

∞ ∞
1 2 2
⇒ 2π
∫ |𝐹(𝑢)| 𝑑𝑢 = ∫ |𝑓(𝑥)| 𝑑𝑥
−∞ −∞

Note: one can easily prove the following parseval’s identities for Fourier cosine and sine
transformation :
∞ ∞
2
(iii) π
∫ 𝐹𝑐(𝑢)𝐺𝑐(𝑢)𝑑𝑢 = ∫ 𝑓(𝑢)𝑔(𝑥)𝑑𝑥
−∞ −∞

∞ ∞
2
(iv) π
∫ 𝐹𝑠(𝑢)𝐺𝑠(𝑢)𝑑𝑢 = ∫ 𝑓(𝑢)𝑔(𝑥)𝑑𝑥
0 0

∞ ∞
2 2
(v)
2
π
0
| |
∫ 𝐹𝑐(𝑢) 𝑑𝑢 = ∫ [𝑓(𝑥)] 𝑑𝑥
0

89
 ∞ ∞
2 2
(vi)
2
π
0
| |
∫ 𝐹𝑠(𝑢) 𝑑𝑢 = ∫ [𝑓(𝑥)] 𝑑𝑥
0

WORKED OUT EXAMPLES

2
Example 1: The function 𝑥 is periodic with period 2𝑙 on the interval [− 𝑙, 𝑙] . Find its Fourier
series.
2 2 2
Solution : 𝑓(𝑥) = 𝑥 . 𝑓(− 𝑥) = (− 𝑥) = 𝑥 = 𝑓(𝑥) so 𝑓(𝑥) is an even function and hence
sine terms will vanish. i.e 𝑏𝑛 = 0

∞
𝑛π𝑥
𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛𝑐𝑜𝑠 𝑙
𝑛=1

𝑙
1
Where 𝑎0 = 2𝑙
∫ 𝑓(𝑥)𝑑𝑥 (𝑛 = 0)
−𝑙

𝑙
1 𝑛π𝑥
𝑎𝑛 = 𝑙
∫ 𝑓(𝑥)𝑐𝑜𝑠 𝑙
𝑑𝑥 (𝑛 = 1, 2, 3………..)
−𝑙

𝑙 𝑙
1 2 𝑛π𝑣
Since 𝑓(𝑥) is even, 𝑎0 = 𝑙
∫ 𝑓(𝑣)𝑑𝑣. 𝑎𝑛 = 𝑙
∫ 𝑓(𝑣)𝑐𝑜𝑠 𝑙
𝑑𝑣
0 0

𝑙 3 𝑙
1 2 1 𝑥 1 2
𝑎0 = 𝑙
∫ 𝑥 𝑑𝑥 = 𝑙
[ 3
] = 3
𝑙
0 0

𝑙
2 2 𝑛π𝑥
𝑎𝑛 = 𝑙
∫ 𝑥 𝑐𝑜𝑠 𝑙
𝑑𝑥 (integrating by part )
0

𝑙 𝑙
2 2 𝑛π𝑥 2 𝑙 𝑛π𝑥
= 𝑙
[𝑥 sin 𝑠𝑖𝑛 𝑙
] − 𝑙
. 2 ∫ 𝑥. 𝑛π
𝑠𝑖𝑛 𝑙
𝑑𝑥
0 0

𝑙
4 𝑛π𝑥
= 0− 𝑛π
∫ 𝑥𝑠𝑖𝑛 𝑙
𝑑𝑥
0

𝑙
4 𝑙 𝑛π𝑥 𝑙 4 𝑙 𝑛π𝑥
= 𝑛π
. 𝑛π
[𝑥𝑐𝑜𝑠 𝑙
] − 𝑛π
. 𝑛π
∫ cos 𝑐𝑜𝑠 𝑙
𝑑𝑥
0 0

90
 4𝑙
2
4𝑙 𝑛π𝑥 𝑙
= 2 2 [𝑙𝑐𝑜𝑠𝑛π − 0] − 3 3 [sin 𝑠𝑖𝑛 𝑙
]
𝑛π 𝑛π 0

2
4𝑙 𝑛
= 2 2 (− 1) − 0
𝑛π

2 2
4𝑙 𝑛 4𝑙 𝑛
= 2 2 (− 1) . ∴𝑎𝑛 = 2 2 (− 1)
𝑛π 𝑛π

2 ∞ 2
𝑙 4𝑙 𝑛 𝑛π𝑥
Therefore , 𝑓(𝑥) = 3
+ ∑ 2 2 (− 1) . cos 𝑐𝑜𝑠 𝑙
𝑛=1 𝑛 π

2 2 ∞
𝑙 4𝑙 1 𝑛 𝑛π𝑥
= 3
+ 2 ∑ 2 (− 1) . cos 𝑐𝑜𝑠 𝑙
π 𝑛=1 𝑛

2 2
𝑙 4𝑙 1 𝑛π 1 2𝑛π 1 3𝑛π 1 4𝑛π
= 3
+ 2 [− 2 cos 𝑐𝑜𝑠 𝑙
+ 2 cos 𝑐𝑜𝑠 𝑙
− 2 cos 𝑐𝑜𝑠 𝑙
+ 2 cos 𝑐𝑜𝑠 𝑙
……………]
π 1 2 3 4

2 2
𝑙 4𝑙 1 𝑛π 1 2𝑛π 1 3𝑛π 1 4𝑛π
= 3
− 2 [ 2 cos 𝑐𝑜𝑠 𝑙
− 2 cos 𝑐𝑜𝑠 𝑙
+ 2 cos 𝑐𝑜𝑠 𝑙
− 2 cos 𝑐𝑜𝑠 𝑙
……………]
π 1 2 3 4

Example 2: obtain the Fourier series of the function 𝑓(𝑥) = {1, 0≤𝑥≤π0, − π≤𝑥≤0 and
verify the result by assuming the complex form of Fourier series.

Solution : 𝐷𝑒𝑓𝑛 : The complex form of the Fourier series can be written as ,
∞ 𝑖𝑛π𝑥
𝑐
𝑓(𝑥) = ∑ 𝐶𝑛 𝑒 .− 𝑐 < 𝑥 < 𝑐
𝑛=−∞

𝑐 𝑖𝑛π𝑥
1 𝑐
Where 𝐶𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑑𝑥 𝑎𝑛𝑑 𝑛 = 0, ±1, ±2………..
−𝑐

1st portion
∞
𝑛
By 𝐷𝑒𝑓 , we have 𝑓(𝑥) = 𝑎0 + ∑ 𝑎𝑛 𝑐𝑜𝑠𝑛𝑥 + 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 ……………(1)
𝑛=1
( )
π
1
Where 𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

91
 π
1
𝑎𝑛 = π
∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥
−π

π
1
𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 Now
−π

π 0 π
1 1
𝑎0 = 2π
∫ 𝑓(𝑥)𝑑𝑥 = 2π
[ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥]
−π −π 0

0 π
1
= 2π
[ ∫ 0𝑑𝑥 + ∫ 1𝑑𝑥]
−π 0

1 1
= 2π
[0 + π] = 2

1
∴ 𝑎0 = 2

Again
π
1
𝑎𝑛= π ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 (𝑛≠0)
−π

0 π
1
= π
[ ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥 + ∫ 𝑓(𝑥)𝑐𝑜𝑠𝑛𝑥𝑑𝑥]
−π 0

1 1 π
= π
[0] + 𝑛π
[𝑠𝑖𝑛𝑛𝑥]0 = 0 + 0 = 0

∴𝑎𝑛 = 0

π
1
Finally, 𝑏𝑛 = π
∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
−π

0 0
1
= π
[ ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + ∫ 𝑓(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥]
−π −π

1
= 0− 𝑛π
(𝑐𝑜𝑠𝑛π − 𝑐𝑜𝑠0)

1 𝑛
=− 𝑛π
[(− 1) − 1]

=
1
𝑛π [1 − (− 1)𝑛] = { 𝑛π2 𝑤ℎ𝑒𝑛 𝑛𝑖𝑠 𝑜𝑑𝑑0 𝑤ℎ𝑒𝑛 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛

92
Now putting the values of 𝑎0, 𝑎𝑛, 𝑎𝑛𝑑 𝑏𝑛 in (1) we get

∞
1
𝑓(𝑥) = 2
+ 0 + ∑ 𝑏𝑛𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
𝑛=1

1 2 1 1
= 2
+ π
(𝑠𝑖𝑛𝑥 + 3
𝑠𝑖𝑛3𝑥 + 5
𝑠𝑖𝑛5𝑥………………..)

Second portion:
𝑛
Now we have to expand the function 𝑓(𝑥)𝑖𝑛 the complex Fourier series. By 𝐷𝑒𝑓 we have
∞ 𝑖𝑛π𝑥
𝑐
𝑓(𝑥) = ∑ 𝐶𝑛 𝑒 .− 𝑐 < 𝑥 < 𝑐
𝑛=−∞

𝑐 𝑖𝑛π𝑥
1 𝑐
Where 𝐶𝑛 = 2𝑐
∫ 𝑓(𝑥)𝑒 𝑑𝑥 𝑎𝑛𝑑 𝑛 = 0, ±1, ±2………..
−𝑐

Here in our given problem 𝑐 = π.

∞
𝑖𝑛𝑥
𝑓(𝑥) = ∑ 𝐶𝑛 𝑒 .− π < 𝑥 < π
𝑛=−∞

π
1 −𝑖𝑛𝑥
Where 𝐶𝑛 = 2π
∫ 𝑓(𝑥)𝑒 𝑑𝑥
−π

𝑛 = 0, ±1, ±2………..
π
1
𝐶0 = 2π
∫ 𝑓(𝑥)𝑑𝑥
−π

0 π
1
= 2π
[ ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥]
−π 0

0 π
1 ⎡ ⎤
= ⎢ ∫ 0𝑑𝑥 + ∫ 1𝑑𝑥⎥
2π ⎢−π ⎥
⎣ 0 ⎦
1 1
2π
[0 + π] = 2

1
∴𝐶0 = 2

93
 π
1 −𝑖𝑛𝑥
𝐶𝑛 = 2π
∫ 𝑓(𝑥)𝑒 𝑑𝑥
−π

0 0
1 −𝑖𝑛𝑥 −𝑖𝑛𝑥
= π
[ ∫ 𝑓(𝑥)𝑒 𝑑𝑥 + ∫ 𝑓(𝑥)𝑒 𝑑𝑥]
−π −π

0 π
1 −𝑖𝑛𝑥 −𝑖𝑛𝑥
= 2π
[ ∫ 0𝑒 𝑑𝑥 + ∫ 1𝑒 𝑑𝑥]
−π 0

−𝑖𝑛𝑥
1 𝑒
= 0+ 2π
[ −𝑖𝑛
]

1
=− 2π𝑛𝑖
[𝑐𝑜𝑠𝑛π − 𝑖𝑠𝑖𝑛𝑛π − 1]

1 𝑛
=− 2π𝑛𝑖
[1 − 1]

1
2π𝑛𝑖
𝑤ℎ𝑒𝑛 𝑛 = ±(1, 3, 5……) 𝑎𝑛𝑑 𝑜 𝑤ℎ𝑒𝑛 𝑛 = ±(2, 4, 6, ……)

( )
𝑖𝑥 𝑖3𝑥 𝑖5𝑥 −𝑖𝑥 −𝑖3𝑥 −𝑖5𝑥
1 1 𝑒 𝑒 𝑒 1 𝑒 𝑒 𝑒
𝑓(𝑥) = 2
+ π𝑖 1
+0+ 3
+0 + 5
+ ………. + π𝑖
( −1
+ 0+ −3
+0 + −5
+ 0 + ………)

𝑖𝑥 −𝑖𝑥
=
1
2
+
1
π𝑖 (
[ 𝑒 −𝑒 ) + 13 (𝑒𝑖3𝑥 − 𝑒−𝑖3𝑥) + 15 (𝑒𝑖5𝑥 − 𝑒−𝑖5𝑥) + ………]
1 2 1 1
= 2
+ π
[𝑠𝑖𝑛𝑥 + 3
𝑠𝑖𝑛3𝑥 + 5
𝑠𝑖𝑛5𝑥 + ………]

Hence
1 2 1 1
𝑓(𝑥) = 2
+ π
[𝑠𝑖𝑛𝑥 + 3
𝑠𝑖𝑛3𝑥 + 5
𝑠𝑖𝑛5𝑥 + ………]

Which is same as in the real form.

∞
𝑐𝑜𝑠𝑢𝑥 2 −𝑥
Example:3 Show that ∫ 2 𝑑𝑢 = π
𝑒 ,𝑥 > 0
0 1+𝑢

−𝑥 𝑥
Proof: Let 𝑓(𝑥) = {𝑒 , 𝑥 > 0 𝑒 , 𝑥 < 0
𝑥 −𝑥
Then 𝑓(− 𝑥) = {𝑒 , 𝑥 > 0 𝑒 , 𝑥 < 0

94
Therefore 𝑓(𝑥) = 𝑓(− 𝑥) .Thus 𝑓(𝑥) is an even function. Now by definition of Fourier integral
of an even function for 𝑥 > 0 we have
∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑢 ∫ 𝑓(𝑡) 𝑐𝑜𝑠𝑢𝑥 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡…………(1)
0 0

−𝑥
Putting 𝑓(𝑥) = 𝑒 in (1),we get
∞ ∞
−𝑡 2 −𝑡
𝑒 = π
∫ 𝑑𝑢 ∫ 𝑒 𝑐𝑜𝑠𝑢𝑥 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡
0 0

∞ ∞
2 −𝑡
= π
∫ 𝑐𝑜𝑠𝑢𝑥 𝑑𝑢 ∫ 𝑒 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡
0 0

∞
2 1
= π
∫ 𝑐𝑜𝑠𝑢𝑥 {0 + 2 } 𝑑𝑢
0 1+𝑢

∞
2 1
= π
∫ 𝑐𝑜𝑠𝑢𝑥 2 𝑑𝑢
0 1+𝑢

∞
𝑐𝑜𝑠𝑢𝑥 2 −𝑥
∴∫ 2 𝑑𝑢 = π
𝑒 , 𝑥 > 0.
0 1+𝑢

∞
𝑥𝑆𝑖𝑛 𝑚𝑥 π −𝑚
Example 4. Prove that ∫ 2 𝑑𝑥 = 2
𝑒 , 𝑚>0
0 𝑥 +1

−𝑚 𝑚
Proof: Let {𝑒 , 𝑚 > 0− 𝑒 , 𝑚 < 0
𝑚 −𝑚
Then 𝑓(− 𝑚) = {− 𝑒 , 𝑚 < 0 𝑒 , 𝑚>0
−𝑚 𝑚
Again –𝑓(𝑚) = {𝑒 , 𝑚 > 0− 𝑒 , 𝑚 < 0

∴ 𝑓(− 𝑚) = –𝑓(𝑚)

Thus f(x) is an odd function. Now by definition of Fourier integral of an odd function for 𝑚 > 0
,
∞ ∞
π
We have 𝑓(𝑚) = 2
∫ 𝑑𝑥 ∫ 𝑓(𝑡) 𝑆𝑖𝑛𝑚𝑥 𝑑𝑡……………(2)
0 0

95
 −𝑚
Putting 𝑓(𝑚) = 𝑒 in (2),we get
∞ ∞
−𝑚 π
𝑒 = 2
∫ 𝑑𝑥 ∫ 𝑓(𝑡) 𝑆𝑖𝑛𝑚𝑥 𝑑𝑡
0 0

∞ ∞
π −𝑚
= ∫ 𝑆𝑖𝑛𝑚𝑥 𝑑𝑥 ∫ 𝑒
2
𝑆𝑖𝑛𝑥𝑡 𝑑𝑡
0 0

∞ −𝑡
π 𝑒 (−𝑆𝑖𝑛𝑥𝑡−𝐶𝑜𝑠𝑥𝑡)
= ∫ 𝑆𝑖𝑛𝑚𝑥 𝑑𝑥[
2 2 ∞]0
0 1+𝑥

∞
π 𝑥
= ∫ 𝑆𝑖𝑛𝑚𝑥 𝑑𝑥[ 0 +
2 2 ∞]0
0 1+𝑥

∞
π 𝑆𝑖𝑛𝑚𝑥
= ∫ 2 2 𝑑𝑥
0 1+𝑥

∞
𝑆𝑖𝑛𝑚𝑥 π −𝑚
Or, ∫ 2 𝑑𝑥 = 2
𝑒 , 𝑚 > 0.
0 1+𝑥

−𝑘𝑥
Example 5. Find the Fourier integral of the function 𝑓(𝑥) = 𝑒 when 𝑥 > 0 and
∞
𝑐𝑜𝑠𝑢𝑥 𝑛 π −𝑘𝑥
𝑓(− 𝑥) = 𝑓(𝑥) for 𝑘 > 0 and hence prove that, ∫ 2 2 𝑑𝑢 = (1 + 𝑥) = 2𝑘
𝑒
0 𝑘 +𝑢

Solution: Since 𝑓(− 𝑥) = 𝑓(𝑥), so 𝑓(𝑥) is even and for even function we have the Fourier
integral
∞ ∞
2
𝑓(𝑥) = π
∫ 𝑓(𝑡)𝑑𝑡 ∫ 𝑐𝑜𝑠 𝑢𝑡 𝑐𝑜𝑠 𝑢𝑥𝑑𝑢𝑧
0 0

∞ ∞
2
= π
∫[∫ 𝑓(𝑡)𝑐𝑜𝑠 𝑢𝑡 𝑑𝑡] 𝑐𝑜𝑠 𝑢𝑥𝑑𝑢…………... (1)
0 0

∞
−𝑘𝑡
Now ∫ 𝑓(𝑡)𝑐𝑜𝑠 𝑢𝑡 𝑑𝑡 = 𝑒 𝑐𝑜𝑠 𝑢𝑡 𝑑𝑡
0

−𝑘𝑥 ∞
= ⎡⎢ 2 2 (− 𝑘𝑐𝑜𝑠 𝑢𝑡 + 𝑢𝑠𝑖𝑛 𝑢𝑡)⎤⎥
𝑒
⎣ 𝑘 +𝑢 ⎦0

96
 𝑘
= 0+ 2 2
𝑘 +𝑢

𝑘
= 2 2
𝑘 +𝑢

Thus from (1) ,we get

∞
2 𝑘
𝑓(𝑥) = π
∫ 2 2 𝑐𝑜𝑠 𝑢𝑥𝑑𝑢
0 𝑘 +𝑢

∞
π 𝑘
= 2
∫ 2 2 𝑐𝑜𝑠 𝑢𝑥𝑑𝑢(𝑥 > 0, 𝑘 > 0………………….. (2)
0 𝑘 +𝑢

−𝑘𝑥
Which is the required Fourier integral of the function 𝑓(𝑥) = 𝑒 .
−𝑘𝑥
Again putting 𝑓(𝑥) = 𝑒 in (2)

We get,
∞
−𝑘𝑥 1
𝑒 =∫ 2 2 𝑐𝑜𝑠 𝑢𝑥𝑑𝑥
0 𝑘 +𝑢

∞
𝑐𝑜𝑠𝑢𝑥 π −𝑘𝑥
∴∫ 2 2 𝑑𝑢 = 2𝑘
𝑒
0 𝑘 +𝑢

−𝑘𝑥
Example 6. Find the Fourier integral of the function 𝑓(𝑥)=𝑒 when 𝑥 > 0 and
∞
𝑢 𝑠𝑖𝑛 𝑢𝑥 π −𝑘𝑥
𝑓(− 𝑥) =− 𝑓(𝑥) for 𝑘 > 0 and hence prove that, ∫ 2 2 𝑑𝑢 = 2
𝑒 , 𝑘 > 0.
0 𝑘 +𝑢

Solution: Since 𝑓(− 𝑥) = 𝑓(𝑥).So 𝑓(𝑥)is an odd function for which we have Fourier integral
∞ ∞
𝑓(𝑥) = ∫ 𝑓(𝑡)𝑑𝑡 ∫ 𝑆𝑖𝑛 𝑢𝑡 𝑆𝑖𝑛 𝑢𝑥 𝑑𝑢
0 0

∞ ∞
2 ⎡ ⎤
= ∫⎢∫ 𝑓(𝑡) 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡⎥𝑠𝑖𝑛𝑢𝑥 𝑑𝑢………………(1)
π ⎢
0⎣0
⎥
⎦
∞ ∞
−𝑘𝑥
Now , ∫ 𝑓(𝑡) 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡 = ∫ 𝑒 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡
0 0

−𝑘𝑥 ∞
= ⎡⎢ ⎤
𝑒 (−𝑘 𝑆𝑖𝑛𝑢𝑡−𝑢 𝑐𝑜𝑠 𝑢𝑡)
2 2 ⎥
⎣ 𝑘 +𝑢 ⎦0

97
 𝑢
= 0+ 2 2
𝑘 +𝑢

𝑢
= 2 2
𝑘 +𝑢

Thus from (1) we get,

∞
2 𝑢
𝑓(𝑥) = π
∫ 2 2 𝑠𝑖𝑛 𝑢𝑥 𝑑𝑢
0 𝑘 +𝑢

∞
2 𝑢 𝑠𝑖𝑛 𝑢𝑥
= π
∫ 2 2 𝑑𝑢(𝑥 > 0, 𝑘 > 0)………………. (2)
0 𝑘 +𝑢

−𝑘𝑥
Which is the required Fourier integral of the function 𝑓(𝑥)=𝑒 .
−𝑘𝑥
Again putting 𝑓(𝑥)=𝑒 𝑖𝑛(2). We get,
∞
−𝑘𝑥 2 𝑢 𝑠𝑖𝑛 𝑢𝑥
𝑓(𝑥) = 𝑒 = π
∫ 2 2 𝑑𝑢.
0 𝑘 +𝑢

∞
𝑢 𝑠𝑖𝑛 𝑢𝑥 π −𝑘𝑥
Or, ∫ 2 2 𝑑𝑢 = 2
𝑒 .
0 𝑘 +𝑢

Example7. Find the Fourier integral of the function

1 𝑥
𝑓(𝑥) = {0 𝑤ℎ𝑒𝑛 𝑥 < 0 2
𝑤ℎ𝑒𝑛 𝑥 = 0 𝑒 𝑤ℎ𝑒𝑛 𝑥 > 0

Solution: By the definition of Fourier integral, we have

∞ ∞
1
𝑓(𝑥) = 2π
∫ 𝑓(𝑡)𝑑𝑡 ∫ 𝑐𝑜𝑠 𝑢(𝑥 − 𝑡)𝑑𝑢
−∞ −∞

∞ ∞
1
= π
∫ 𝑓(𝑡)𝑑𝑡 ∫ 𝑐𝑜𝑠 𝑢(𝑥 − 𝑡)𝑑𝑢
−∞ 0

∞ ∞
1
= π
∫ 𝑓(𝑡)𝑑𝑡 ∫(𝑐𝑜𝑠 𝑢𝑥 𝑐𝑜𝑠 𝑢𝑡 + 𝑠𝑖𝑛 𝑢𝑥 𝑆𝑖𝑛 𝑢𝑡) 𝑑𝑢
−∞ 0

98
 ∞ ∞ ∞ ∞
1 ⎡ ⎰ ⎱ ⎰ ⎱ ⎤
= ⎢∫ ∫ 𝑓(𝑡) 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡 𝑐𝑜𝑠 𝑢𝑥 𝑑𝑢 + ∫ ∫ 𝑓(𝑡) 𝑠𝑖𝑛 𝑢𝑡 𝑑𝑡 𝑠𝑖𝑛 𝑢𝑥 𝑑𝑢⎥
π ⎢ 0 ⎱−∞ ⎰ ⎱ ⎰ ⎥
⎣ 0 −∞ ⎦
……………………………………………….. (1)

Now,
∞ 0 ∞
∫ 𝑓(𝑡) 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡 = ∫ 𝑓(𝑡) 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡 + ∫ 𝑓(𝑡) 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡
−∞ −∞ 0

0 ∞
−𝑡
= ∫ 0 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡 + ∫ 𝑒 𝑐𝑜𝑠𝑢𝑡 𝑑𝑡
−∞ 0

−𝑡 ∞
𝑒
= 0 +[ 2 (− 𝑐𝑜𝑠 𝑢𝑡 + 𝑢𝑠𝑖𝑛 𝑢𝑡)]
1+𝑢
0

1
= 2 + 0
1+𝑢

1
= 2
1+𝑢

Similarly,
∞ 0 ∞
∫ 𝑓(𝑡) 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡 = ∫ 𝑓(𝑡) 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡 + ∫ 𝑓(𝑡) 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡
−∞ −∞ 0

0 ∞
−𝑡
= ∫ 0 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡 + ∫ 𝑒 𝑠𝑖𝑛𝑢𝑡 𝑑𝑡
−∞ 0

−𝑡 ∞
𝑒
= 0 +[ 2 (− 𝑠𝑖𝑛 𝑢𝑡 − 𝑢𝑐𝑜𝑠 𝑢𝑡)]
1+𝑢
0

𝑢
= 0+ 2
1+𝑢

𝑢
= 2
1+𝑢

Putting these values in (1), we get

99
 ∞ ∞
1 𝑐𝑜𝑠𝑢𝑥 𝑢 𝑠𝑖𝑛 𝑢𝑥
𝑓(𝑥) = π
[∫ 2 𝑑𝑢 + ∫ 2 𝑑𝑢]
0 1+𝑢 0 1+𝑢

∞
=
1
π
∫
0
( 𝑐𝑜𝑠𝑢𝑥+𝑢 𝑠𝑖𝑛𝑢𝑥
1+𝑢
2 )𝑑𝑢……………………. (2)
Putting x=0 in (2), we get
∞
1 1
𝑓(0) = π
∫ 2 𝑑𝑢
0 1+𝑢

1 ∞
= π
[𝑢 ]0

1
= π
[∞ − 0 ]

1 π
= π
( 2 − 0)

1
= 2

1
So 𝑓(𝑥) = 2
for 𝑥 = 0 is satisfied.

Hence
∞
1 𝑐𝑜𝑠𝑢𝑥+𝑢 𝑠𝑖𝑛𝑢𝑥
𝑓(𝑥) = π
∫( 2 )𝑑𝑢
0 1+𝑢

which is the required Fourier integral of the given function.

−𝑥
Example 8. Find the Fourier sine transform of 𝑒 , 𝑥≥0.

Solution: By the definition of Fourier sine transform of 𝑓(𝑥) for 𝑜 < 𝑥 < ∞, we have
∞
𝑓𝑠(𝑛) = ∫ 𝐹(𝑥) sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
0

∞ −𝑥 ∞ ∞ −𝑥
−𝑥 −𝑥 𝑒 𝑒 cos𝑐𝑜𝑠 𝑛𝑥
𝑓𝑠(𝑛) = ∫ 𝑒 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥 𝑆𝑖𝑛𝑐𝑒 𝐹(𝑥) = 𝑒 = [− 𝑛
cos 𝑐𝑜𝑠 𝑛𝑥 ] − ∫ 𝑛
𝑑𝑥
0 0 0

−𝑥 ∞ ∞ −𝑥
1 𝑒 𝑒 sin𝑠𝑖𝑛 𝑛𝑥
= 0+ 𝑛
−[ 2 sin 𝑠𝑖𝑛 𝑛𝑥 ] − ∫ 2 𝑑𝑥
𝑛 0 𝑛
0

100
 1 1
= 𝑛
− 0− 2 𝑓𝑠(𝑛)
𝑛

1 1
⇒ 𝑓𝑠(𝑛) + 2 𝑓𝑠(𝑛) = 𝑛
𝑛

( )𝑓 (𝑛) =
⇒ 1+
1
𝑛
2
𝑠
1
𝑛

⇒( )𝑓 (𝑛) =
2
𝑛 +1 1
2 𝑛
𝑛 𝑠

2
1 𝑛 𝑛
⇒ 𝑓𝑠(𝑛) = 𝑛
. 2 = 2
𝑛 +1 𝑛 +1

𝑛
∴ 𝑓𝑠(𝑛) = 2
𝑛 +1

−𝑥 𝑛
.Hence the Fourier sin transform of 𝑒 is 2 .
𝑛 +1

𝑛
Example 9. Find the inverse Fourier sine transform of 𝑓𝑠(𝑛) = 2
1+𝑛

Solution: By definition of the inverse Fourier sin transform

we have
∞
2
𝐹(𝑥) = π
∫ 𝑓𝑠(𝑛) sin 𝑠𝑖𝑛 𝑛𝑥𝑑𝑛
0

∞
2 𝑛
= π
∫ 2 sin 𝑠𝑖𝑛 𝑛𝑥𝑑𝑛 ………………………………….. (1)
0 1+𝑛

From the Fourier integral formula of an odd function, we

have
∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑢 ∫ 𝑓(𝑡) sin 𝑠𝑖𝑛 𝑢𝑡 𝑠𝑖𝑛𝑢𝑥 𝑑𝑡
0 0

∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑛 ∫ 𝑓(𝑡) sin 𝑠𝑖𝑛 𝑛𝑡 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑡 ………………………….. (2)
0 0

101
Taking
−𝑥
𝑓(𝑥) = 𝑒

in (2),we have
∞ ∞
−𝑥 2 −𝑡
𝑒 = π
∫ 𝑑𝑛 ∫ 𝑒 sin 𝑠𝑖𝑛 𝑛𝑡 sin 𝑠𝑖𝑛 𝑛𝑥𝑑𝑡
0 0

∞ ∞
2 −𝑡
= π
∫ sin 𝑠𝑖𝑛 𝑛𝑥{∫ 𝑒 sin 𝑠𝑖𝑛 𝑛𝑡 𝑑𝑡 } 𝑑𝑛
0 0

∞
=
2
π
∫ sin 𝑠𝑖𝑛 𝑛𝑥
0
( )𝑑𝑛𝑛
1+𝑛
2

∞
2 𝑛𝑠𝑖𝑛 𝑛𝑥
= π
∫ 2 𝑑𝑛
0 1+𝑛

∞
𝑛 π −𝑥
⇒∫ 2 sin 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑛 = 2
𝑒 ……………………………………. (3)
0 1+𝑛

Combining (1) and (3), we get

2 π −𝑥
𝐹(𝑥) = π
. 2
𝑒

−𝑥
Hence 𝐹(𝑥) = 𝑒 which is the required inverse Fourier sin transform of
𝑛
𝑓𝑠(𝑛) = 2
1+𝑛

.
−𝑥
Example 10. Find the Fourier cosine transform of 𝑒 , 𝑥≥0.

Solution: By definition of Fourier cosine transform of f(x) for 0 < 𝑥 < ∞,

we have
∞
𝑓𝑐(𝑛) = ∫ 𝐹(𝑥) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥……………………………………………(1)
0

∞
−𝑥 −𝑥
∴𝑓𝑐(𝑛) = ∫ 𝑒 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 𝑆𝑖𝑛𝑐𝑒 𝐹(𝑥) = 𝑒
0

102
 −𝑥 ∞ −𝑥
𝑒 sin𝑠𝑖𝑛 𝑛𝑥 𝑒 sin𝑠𝑖𝑛 𝑛𝑥
= ⎡⎢ 𝑛
⎤+ ∫
⎥ 𝑛
𝑑𝑥
⎣ ⎦ 0

∞ ∞
−𝑥 −𝑥
= 0−
𝑛
1
2 [ 𝑒 𝐶𝑜𝑠𝑛𝑥 ]0 −
1
2
𝑛 0
∫𝑒 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥

1 1
=0+ 2 − 2 𝑓𝑐(𝑛)
𝑛 𝑛

1 1
⇒ 𝑓𝑐(𝑛) + 2 𝑓𝑐(𝑛) = 2
𝑛 𝑛

2
𝑛 +1 1
⇒ 2 𝑓𝑐(𝑛) = 2
𝑛 𝑛

1
⇒ 𝑓𝑐(𝑛) = 2
𝑛 +1

−𝑥 1
Hence the Fourier cosine transform of 𝑒 is 2 .
𝑛 +1

1
Example11. Find the inverse Fourier cosine transform of 𝑓𝑐(𝑛) = 2
1+𝑛

Solution: By definition of the inverse Fourier cosine transform

we have,
∞ ∞
2 2 1
𝐹(𝑥) = π
∫ 𝑓𝑐(𝑛) cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑛 = π
∫ 2 cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑛 …………………………. (1)
0 0 1+𝑛

From the Fourier integral formula of an even function, we have

∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑢 ∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 𝑢𝑡 𝑐𝑜𝑠𝑢𝑥 𝑑𝑡
0 0

∞ ∞
2
𝑓(𝑥) = π
∫ 𝑑𝑛 ∫ 𝑓(𝑡) cos 𝑐𝑜𝑠 𝑛𝑡 cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑡 ………………………………(2)
0 0

Taking
−𝑥
𝑓(𝑥) = 𝑒

in (2),we have
∞ ∞ ∞ ∞
−𝑥 2 −𝑡 2 −𝑡
𝑒 = π
∫ 𝑑𝑛 ∫ 𝑒 cos 𝑐𝑜𝑠 𝑛𝑡 cos 𝑐𝑜𝑠 𝑛𝑥𝑑𝑡 = π
∫ cos 𝑐𝑜𝑠 𝑛𝑥{∫ 𝑒 cos 𝑐𝑜𝑠 𝑛𝑡 𝑑𝑡 } 𝑑𝑛
0 0 0 0

103
 ∞
=
2
π
∫ cos 𝑐𝑜𝑠 𝑛𝑥
0
( )𝑑𝑛
1+𝑛
1
2

∞
1 π −𝑥
∴ ∫( 2 ) cos 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑛 = 2
𝑒 ……………………………. (3)
0 1+𝑛

Combining (1) and (3) ,

we get
2 π −𝑥 −𝑥
𝐹[𝑥] = π
. 2
𝑒 =𝑒 .

−𝑥
Hence 𝐹[𝑥] = 𝑒 which is the required inverse Fourier cosine transform of
1
𝐹𝑐(𝑛) = 2
1+𝑛

Example 12. Find the (a) finite Fourier sine transform (b) finite Fourier cosine transform of
the function 𝐹(𝑥) = 2𝑥, 0 < 𝑥 < 4.

Solution: (a) Since 𝑙 = 4,we have

𝑙 4
𝑛π𝑥 𝑛π𝑥
𝑓𝑠(𝑛) = ∫ 𝐹(𝑥)𝑠𝑖𝑛 𝑙
𝑑𝑥 = ∫ 𝐹(𝑥)𝑠𝑖𝑛 4
𝑑𝑥
0 0

4
𝑛π𝑥
= ∫ 2𝑥𝑠𝑖𝑛 4
0

4
4 𝑛π𝑥 8 𝑛π𝑥
= [− 2𝑥. 𝑛π
𝑐𝑜𝑠 4
4] + 0 𝑛π
∫ 𝑐𝑜𝑠 4
𝑑𝑥
0

32 32 𝑛π𝑥
=− 𝑛π
𝑐𝑜𝑠𝑛π + 2 2 [𝑠𝑖𝑛 4
4] 0
𝑛π

32 32
=− 𝑛π
𝑐𝑜𝑠𝑛π + 2 2 (0 − 0)
𝑛π

32
=− 𝑛π
𝑐𝑜𝑠𝑛π

which is the finite Fourier sine transform of 𝐹(𝑥) = 2𝑥

104
(b) If
𝑙
𝑛π𝑥
𝑛 > 0, 𝑓𝑐(𝑛) = ∫ 𝐹(𝑥)𝑐𝑜𝑠 𝑙
𝑑𝑥
0

4
𝑛π𝑥
= ∫ 2𝑥𝑐𝑜𝑠 4
0

4
4 𝑛π𝑥 8 𝑛π𝑥
= [2𝑥. 𝑛π
𝑠𝑖𝑛 4
4] − 0 𝑛π
∫ 𝑠𝑖𝑛 4
𝑑𝑥
0

8 4 𝑛𝑥π 4
= 0− (− )⎡𝐶𝑜𝑠 ⎤
𝑛π 𝑛π ⎣ 4 ⎦0

32
= 2 2 (𝐶𝑜𝑠𝑛𝑛π − 1) which is the finite Fourier Cosine transform of 𝐹(𝑥) = 2𝑥
𝑛π

4
If 𝑛 = 0, 𝑓𝑐(𝑛) = 𝑓𝑐(0) = ∫ 2𝑥𝑑𝑥
0

2 4
2
𝑥
= 2⎡⎢ 2 ⎤⎥ = 4 − 0 = 16
⎣ ⎦0
( )
∴𝑓𝑐(𝑛) = 𝑓𝑐(0) = 16.

Example.13. Find the Fourier transform of 𝑓(𝑥) defined by

𝑓(𝑥) = {1, |𝑥| < 𝑎 0, |𝑥| > 𝑎

And hence evaluate

∞ ∞
𝑆𝑖𝑛 𝑢𝑎 𝐶𝑜𝑠𝑢𝑥 𝑆𝑖𝑛𝑢
(𝑖) ∫ 𝑢
𝑑𝑢 (𝑖𝑖) ∫ 𝑢
𝑑𝑢
−∞ 0

Solution:

By the definition of Fourier transform, we have

∞
−𝑖𝑢𝑡
𝑓(𝑢) = ∫ 𝑓(𝑡)𝑒 𝑑𝑢
−∞

−𝑎 𝑎 ∞
−𝑖𝑢𝑡 −𝑖𝑢𝑡 −𝑖𝑢𝑡
= ∫ 𝑓(𝑡)𝑒 𝑑𝑢 + ∫ 𝑓(𝑡)𝑒 𝑑𝑢 + ∫ 𝑓(𝑡)𝑒 𝑑𝑢
−∞ −𝑎 𝑎

105
 −𝑎 𝑎 ∞
−𝑖𝑢𝑥 −𝑖𝑢𝑥 −𝑖𝑢𝑡
= ∫ 0. 𝑒 𝑑𝑥 + ∫ 1. 𝑒 𝑑𝑥 + ∫ 0. 𝑒 𝑑𝑥
−∞ −𝑎 𝑎

𝑎
−𝑖𝑢𝑥
= 0 + ∫ 1𝑒 𝑑𝑥 + 0
−𝑎

𝑎
=
1
𝑖𝑢 [𝑒−𝑖𝑢𝑥]−𝑎
=−
1
𝑖𝑢 (𝑒−𝑖𝑢𝑎 − 𝑒−𝑖𝑢𝑎)
=
2
𝑢
1
2𝑖 (𝑒−𝑖𝑢𝑎 − 𝑒−𝑖𝑢𝑎)
2
= 𝑢
𝑆𝑖𝑛𝑢𝑎.

Hence
2𝑆𝑖𝑛𝑢𝑎
𝑓(𝑢) = 𝑢
, 𝑢 ≠ 0.

For
𝑎
𝑢 = 0, 𝑤𝑒 𝑜𝑏𝑡𝑎𝑖𝑛 𝑓(𝑢) = ∫ 𝑑𝑥 = 2𝑎
−𝑎

Solution of second portion:

By the definition of the inverse Fourier transform, we have

∞
1 𝑖𝑢𝑥
𝑓(𝑥) = 2π
∫ 𝑓(𝑢)𝑒 𝑑𝑢
−∞

∞
1 𝑖𝑢𝑥
∴ 2π
∫ 𝑓(𝑢)𝑒 𝑑𝑢 = 𝑓(𝑥) = {1, |𝑥| < 𝑎 0, |𝑥| > 𝑎. ………………………(𝑖)
−∞

Now,
∞ ∞
1 𝑖𝑢𝑥 1 2𝑆𝑖𝑛𝑢𝑎 𝑖𝑢𝑥
2π
∫ 𝑓(𝑢)𝑒 𝑑𝑢 = 2π
∫ 𝑢
𝑒 𝑑𝑢
−∞ −∞

∞
1 2𝑆𝑖𝑛𝑢𝑎
= 2π
∫ 𝑢
(𝐶𝑜𝑠𝑢𝑥 + 𝑖𝑆𝑖𝑛𝑢𝑥)𝑑𝑢
−∞

106
 ∞ ∞
1 𝑆𝑖𝑛𝑢𝑎 𝐶𝑜𝑠𝑢𝑥 1 𝑆𝑖𝑛𝑢𝑎 𝑆𝑖𝑛𝑢𝑥
= π
∫ 𝑢
𝑑𝑢 + 2π
∫ 𝑢
𝑑𝑢…………………………………(𝑖𝑖)
−∞ −∞

The integrand in the second integral of (2) on the right side is odd and so this integral is zero.

Hence combining (1) and (2), we get

∞
1 𝑆𝑖𝑛𝑢𝑎 𝐶𝑜𝑠𝑢𝑥
π
∫ 𝑢
𝑑𝑢 = {1, |𝑥| < 𝑎 0, |𝑥| > 𝑎
−∞

∞
𝑆𝑖𝑛𝑢𝑎 𝐶𝑜𝑠𝑢𝑥
∫ 𝑢
𝑑𝑢 = {π, |𝑥| < 𝑎 0, |𝑥| > 𝑎
−∞

If 𝑥 = 0 𝑎𝑛𝑑 𝑎 = 1 then from the above result we have

∞
𝑆𝑖𝑛𝑛𝑢
∫ 𝑢
𝑑𝑢 = π
−∞

∞
𝑆𝑖𝑛𝑢𝑛
𝑂𝑟, 2 ∫ 𝑢
𝑑𝑢 = π
0

, since the integral is even.

∞
𝑆𝑖𝑛𝑛𝑢 π
∫ 𝑢
𝑑𝑢 = 2
.
0

−|𝑥|
Example.14. Find the Fourier transform of 𝑓(𝑥) = 𝑒 where x belongs to (− ∞, ∞)

Solution:

By the definition of Fourier transform, we have

∞
𝑖𝑢𝑡
𝐹(𝑢) = ∫ 𝑓(𝑡)𝑒 𝑑𝑡
−∞

∞
𝑖𝑢𝑥
= ∫ 𝑓(𝑥)𝑒 𝑑𝑥
−∞

0 ∞
𝑖𝑢𝑥 𝑖𝑢𝑥
= ∫ 𝑓(𝑥)𝑒 𝑑𝑥 + ∫ 𝑓(𝑥)𝑒 𝑑𝑥
−∞ 0

107
 0 ∞
𝑥 𝑖𝑢𝑥 −𝑥 𝑖𝑢𝑥
= ∫ 𝑒 𝑒 𝑑𝑥 + ∫ 𝑒 𝑒 𝑑𝑥
−∞ 0

−|𝑥| −𝑥 𝑥
Since 𝑒 = {𝑒 ,𝑥 > 0𝑒 , 𝑥 < 0.
0 ∞
(1+𝑖𝑢) −(1−𝑖𝑢)
= ∫ 𝑒 𝑑𝑥 + ∫ 𝑒 𝑑𝑥
−∞ 0

(1+𝑖𝑢)𝑥 0 −(1−𝑖𝑢)𝑥 ∞
=
1
1+𝑖𝑢 [
𝑒 ]
−∞
+
1
−(1−𝑖𝑢) [ 𝑒 ]0
1 1
= 1+𝑖𝑢
(1 − 0) − 1−𝑖𝑢
(0 − 1)

1−𝑖𝑢+1+𝑖𝑢
= 2 2
1−𝑖 𝑢

2
= 2
1+𝑢

Hence,
2 −|𝑥|
𝐹(𝑢) = 2 which is the required Fourier transform of 𝑓(𝑥) = 𝑒
1+𝑢

Example:15. Find the Fourier transform of

Solution: By definition of the Fourier transform

∞
𝑖𝑢𝑡
𝐹(𝑢) = ∫ 𝑓(𝑡)𝑒 𝑑𝑡
−∞

∞
𝑖𝑢𝑥
= ∫ 𝑓(𝑥)𝑒 𝑑𝑥
−∞

−1 1 ∞
𝑖𝑢𝑥 𝑖𝑢𝑥 𝑖𝑢𝑥
= ∫ 𝑓(𝑥)𝑒 𝑑𝑥 + ∫ 𝑓(𝑥)𝑒 𝑑𝑥 + ∫ 𝑓(𝑥)𝑒 𝑑𝑥
−∞ −1 1

−1 1 ∞
𝑖𝑢𝑥 2 𝑖𝑢𝑥 𝑖𝑢𝑥
= ∫ 0𝑒 𝑑𝑥 ∫ (1 − 𝑥 )𝑒 𝑑𝑥 + ∫ 0𝑒 𝑑𝑥
−∞ −1 1

1
2 𝑖𝑢𝑥
= 0 + ∫ (1 − 𝑥 )𝑒 𝑑𝑥 + 0
−1

108
 1 1 1
2 𝑖𝑢𝑥 1 𝑖𝑢𝑥 𝑖𝑢𝑥 𝑖𝑢𝑥 𝑖𝑢
⎣ (
=⎡ 1 −𝑥 . ) 1
𝑖𝑢
𝑒 ⎤ − ∫ (1 − 2𝑥)
⎦−1 −1
1
𝑖𝑢
𝑒 𝑑𝑥 = 0 +
2
𝑖𝑢
∫ 𝑥𝑒 𝑑𝑥 =
2
𝑖𝑢
𝑥
. [ 𝑖𝑢 𝑒 1] − 1 −
2
2 2
𝑖 𝑢 −1
∫ 𝑒
−1

=
2
2 2
𝑖𝑢
[1𝑒𝑖𝑢 + 1𝑒−𝑖𝑢] + 𝑢
2
2
𝑥
. [ 𝑖𝑢 𝑒
𝑖𝑢𝑥
1] − 1

=−
𝑢
2
2 (𝑒𝑖𝑢 + 𝑒−𝑖𝑢) + 2
𝑖𝑢
3 (𝑒𝑖𝑢 − 𝑒−𝑖𝑢)
=−
4
𝑢
2 .
1
2 (𝑒𝑖𝑢 + 𝑒−𝑖𝑢) + 𝑢
4
3 .
1
2𝑖 (𝑒𝑖𝑢 − 𝑒−𝑖𝑢)
4 4
=− 2 cos 𝑐𝑜𝑠 𝑢 + 3 sin 𝑠𝑖𝑛 𝑢
𝑢 𝑢

4
= 3 (𝑠𝑖𝑛𝑢 − 𝑢𝑐𝑜𝑠 𝑢)
𝑢

Hence
4
𝐹(𝑢) = 3 (𝑠𝑖𝑛𝑢 − 𝑢𝑐𝑜𝑠 𝑢).
𝑢

1 2𝑐𝑜𝑠𝑛π
Or, 𝐴 = 𝑛
(1 − 𝑐𝑜𝑠𝑛π − 1 + 3𝑐𝑜𝑠𝑛π) = 𝑛

Putting the value of A in (4), we get

2
1 2𝑐𝑜𝑠𝑛π −𝑛 𝑡
𝑢 = 𝑢(𝑛, 𝑡) 𝑛
(1 − 3𝑐𝑜𝑠𝑛π) + 𝑛
𝑒

Taking inverse finite Fourier sine transform, we get

∞ ∞ 2 ∞ ∞ 2
2 (1−3𝑐𝑜𝑠𝑛π) 2 2𝑐𝑜𝑠𝑛π −𝑛 𝑡 2 (1−3𝑐𝑜𝑠𝑛π) 4 2𝑐𝑜𝑠𝑛π −𝑛 𝑡
𝑈(𝑥, 𝑡) = π
∑ 𝑛
𝑠𝑖𝑛𝑛π + π
∑ 𝑛
𝑒 𝑠𝑖𝑛𝑛π = π
∑ 𝑛
𝑠𝑖𝑛𝑛π + π
∑ 𝑛
𝑒 𝑠𝑖
𝑛=1 𝑛=1 𝑛=1 𝑛=1

The finite Fourier sine transform of F(x), 0<x<l, is defined as

𝑙
𝑛π𝑥
𝑓𝑠(𝑛) = ∫ 𝐹(𝑥)𝑠𝑖𝑛 𝑙
𝑑𝑥.
0

Here l=π
π
∴ 𝑓𝑠(𝑛) = ∫ 𝐹(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

109
 π
𝑓𝑠{𝐹(𝑥)} = ∫ 𝐹(𝑥)𝑠𝑖𝑛𝑛𝑥𝑑𝑥.
0

π
𝐹(𝑥) = 1 {𝑓𝑠} = ∫ 1𝑠𝑖𝑛𝑛𝑥𝑑𝑥
0

1 π
=− 𝑛
[𝑐𝑜𝑠𝑛π]0

1 1
=− 𝑛
(𝑐𝑜𝑠𝑛π − 1) = 𝑛
(1 − 𝑐𝑜𝑠𝑛π).

∴ Taking inverse finite Fourier sine transform, we get

∞
2 (1−𝑐𝑜𝑠𝑛π)
1= π
∑ 𝑛
𝑠𝑖𝑛𝑛π.
𝑛=1

𝐹(𝑥) = 𝑥
π
𝑓𝑠{𝑥} = ∫ 𝑥𝑠𝑖𝑛𝑛𝑥 𝑑𝑥
0

π
𝑐𝑜𝑠𝑛π π 𝑐𝑜𝑠𝑛π
= [− 𝑥 𝑛
] +∫ 𝑛
𝑑𝑥
0 0

π𝑐𝑜𝑠𝑛π 1 π
= 𝑛
+0 + 2 [𝑠𝑖𝑛𝑛𝑥]0
𝑛

π𝑐𝑜𝑠𝑛π
=− 𝑛

Taking inverse finite Fourier sine transform we get

∞
2 π𝑐𝑜𝑠𝑛π
𝑥= π
∑− 𝑛
𝑠𝑖𝑛𝑛π
𝑛=1

Therefore,
∞
2 (1−3𝑐𝑜𝑠𝑛π)
= π
∑ 𝑛
𝑠𝑖𝑛𝑛π
𝑛=1

∞ ∞
2 (1−𝑐𝑜𝑠𝑛π) 2 2𝑐𝑜𝑠𝑛π
= π
∑ 𝑛
𝑠𝑖𝑛𝑛π + π
∑− 𝑛
𝑠𝑖𝑛𝑛π
𝑛=1 𝑛=1

∞ ∞
2 (1−𝑐𝑜𝑠𝑛π) 2 2 π𝑐𝑜𝑠𝑛π
= π
∑ 𝑛
𝑠𝑖𝑛𝑛π + π
. π
∑− 𝑛
𝑠𝑖𝑛𝑛π
𝑛=1 𝑛=1

110
 2
= 1+ π
.𝑥

Thus
∞
2 (1−3𝑐𝑜𝑠𝑛π) 2
π
∑ 𝑛
𝑠𝑖𝑛𝑛π = 1 + π
.𝑥
𝑛=1

Hence from (5), we have

∞ 2
2𝑥 4 𝑐𝑜𝑠𝑛π −𝑛 𝑡
𝑈(𝑥, 𝑡) = 1 + π
+ π
∑ 𝑛
𝑒 𝑠𝑖𝑛𝑛𝑥
𝑛=1

∞ 𝑛 2
2𝑥 4 (−1) −𝑛 𝑡
= 1+ π
+ π
∑ 𝑛
𝑒 𝑠𝑖𝑛𝑛𝑥
𝑛=1

Example 9: Use the complex form of the Fourier transform to solve the boundary value problem

2
∂𝑈 ∂𝑈
∂𝑡
= 2 , 𝑈(𝑥, 0) = 𝑓(𝑥),
∂𝑥

𝐼𝑈(𝑥, 𝑡 𝐼 < 𝑀, where -∞ < 𝑥 < ∞ , 𝑡 > 0.

Solution: The given partial differential equation is

2
∂𝑈 ∂𝑈
∂𝑡
= 2 (1)
∂𝑥

Taking the complex form of the Fourier transform with respect to x of both sides of (1) we get
∞ ∞ 2
∂𝑈 𝑖α𝑥 ∂𝑈 𝑖α𝑥
∫ ∂𝑡
𝑒 𝑑𝑥 = ∫ 2 𝑒 𝑑𝑥 (2)
−∞ −∞ ∂𝑥

∞
𝑖α𝑥
Let 𝐼(𝑈) = ∫ 𝑈(𝑥, 𝑡)𝑒 𝑑𝑥
−∞

∞
𝑑 ∂𝑈 𝑖α𝑥
∴ 𝑑𝑡
𝐼(𝑈) = ∫ ∂𝑡
𝑒 𝑑𝑥
−∞

∞ 𝑖α𝑥
𝑈
= ∫ 2 𝑒 𝑑𝑥
−∞ ∂𝑥

111
 ∞
𝑖α𝑥 ∂𝑈 ∞ 𝑖α𝑥 ∂𝑈
= [𝑒 ∂𝑡
] − 𝑖α ∫ 𝑒 ∂𝑡
𝑑𝑥
−∞ −∞

∞
𝑖α𝑥 ∂𝑈 ∂𝑈
=0 − 𝑖α ∫ 𝑒 ∂𝑡
𝑑𝑥, 𝑠𝑖𝑛𝑐𝑒 ∂𝑡
→0 𝑎𝑠 𝑥→∞,
−∞

∞ ∞
𝑖α𝑥 2 2 𝑖α𝑥
= − 𝑖α 𝑒 [ 𝑈(𝑥, 𝑡) ] −∞
+𝑖 α ∫ 𝑒 𝑈(𝑥, 𝑡)𝑑𝑥.
−∞

∞
2 𝑖α𝑥
=0-α ∫ 𝑒 𝑈(𝑥, 𝑡)𝑑𝑥 𝑠𝑖𝑛𝑐𝑒 𝑈→0 𝑎𝑠 𝑥→∞
−∞

2
=-α 𝐼(𝑈).

𝑑𝐼(𝑈) 2
∴ 𝑑𝑡
= α 𝐼(𝑈)

𝑑𝐼(𝑈) 2
Or, 𝐼(𝑈)
=− α 𝑑𝑡 (3)

Integrating both sides of (3), we get

2
log 𝑙𝑜𝑔 𝐼(𝑈) =− α 𝑡 + log 𝑙𝑜𝑔 𝑐 , 𝐶 being some constant of integration.
2 2
−α 𝑡 −α 𝑡
Or, log 𝑙𝑜𝑔 𝐼(𝑈) = log 𝑙𝑜𝑔 𝑒 + log 𝑙𝑜𝑔 𝑐 = 𝑒
2
−α 𝑡
∴ 𝐼(𝑈) = 𝐶 𝑒 (4)
2
−α 𝑡
I,e 𝐼{𝑈(𝑥, 𝑡)} = 𝐶 𝑒 (5)

When 𝑡 = 0, 𝑈(𝑥, 0) = 𝑓(𝑥)

0
∴𝐼{𝑈(𝑥, 𝑡)} = 𝐼{𝑓(𝑥)} = 𝐶𝑒 = 𝐶
∞ ∞
𝑖α𝑥 𝑖α𝑢
∴ 𝐶 = 𝐼{𝑓(𝑥)} = ∫ 𝑒 𝑓(𝑥)𝑑𝑥. = ∫ 𝑒 𝑓(𝑢)𝑑𝑢
−∞ −∞

2
−α 𝑡
Thus from (4), we have 𝐼{𝑈(𝑥, 𝑡)} = 𝐼{𝑓(𝑥)}𝑒 ……………………………. (6)

Taking the inverse Fourier transform, we get

112
 ∞ 2 ∞ ∞ 2
1 −α 𝑡 −𝑖α𝑥 1 𝑖α𝑥 −α 𝑡 −𝑖α𝑥
𝑈(𝑥, 𝑡) = 2π
∫ 𝐼{𝑓}𝑒 𝑒 𝑑α. = 2π
∫ [ ∫ 𝑓(𝑥)𝑒 𝑑𝑥]𝑒 𝑒 𝑑α.
−∞ −∞ −∞

∞ ∞ 2
1 𝑖α𝑢 −α 𝑡 −𝑖α𝑥
= 2π
∫ [ ∫ 𝑓(𝑢)𝑒 𝑑𝑢]𝑒 𝑒 𝑑α.
−∞ −∞

∞ ∞ 2
1 −α 𝑡−𝑖α(𝑥−𝑢)
= 2π
∫ 𝑓(𝑢)[ ∫ 𝑒 𝑑α]𝑑𝑢.
−∞ −∞

𝑖(𝑥−𝑢 2 2 2
∞ ∞ ∞ ∞
−𝑡{α+ 2𝑡 }
𝑖(𝑥−𝑢
1 −𝑡{α+ 2𝑡
} 2
(𝑥−𝑢) 1 ⎡ ⎤ − (𝑥−𝑢)
= ∫ 𝑓(𝑢)[ ∫ 𝑒 − 𝑑α]𝑑𝑢. = ∫ 𝑓(𝑢)⎢ ∫ 𝑒 𝑑α⎥𝑒 4𝑡 𝑑𝑢.
2π 4𝑡 2π ⎢−∞ ⎥
−∞ −∞ −∞ ⎣ ⎦
Putting
{
𝑡α+
𝑖(𝑥−𝑢)
2𝑡 }= 𝑦
So that 𝑡𝑑α = 𝑑𝑦
1
∴ 𝑑α = 𝑑𝑦.
𝑡

α=∞
𝐿𝑖𝑚𝑖𝑡𝑠 𝑦=∞} α =− ∞ 𝑦 =− ∞ }

∞ 𝑖(𝑥−𝑢 2
−𝑡{α+ 2𝑡
}
∴ ∫ 𝑒 𝑑α .
−∞

∞
− 𝑖(𝑥−𝑢) 2
=2∫ 𝑒 [√𝑡{α + 2𝑡
] 𝑑α.
0

∞ 2
−𝑦 1
=2∫ 𝑒 √𝑡
𝑑𝑦.
0

∞ 2
2 −𝑦 2 √π π
= √𝑡
∫𝑒 𝑑𝑦 = √𝑡
. 2
= 𝑡
.
0

∞ 2
−𝑦 √π
Since ∫ 𝑒 𝑑𝑦 = 2
.
0

Therefore, from (7), we get

113
 2
+∞ (𝑥−𝑢)
1 √π − 4𝑡
𝑈(𝑥, 𝑡) = 2π
∫ 𝑓(𝑢). √𝑡
.𝑒 𝑑𝑢.
−∞

2
+∞ (𝑥−𝑢)
1 − 4𝑡
∫ 𝑓(𝑢). 𝑒 𝑑𝑢.
2 π𝑡
−∞

Example 10:Use the complex form of the Fourier transform to solve the boundary value problem
2
∂𝑈 ∂𝑈
∂𝑡
= 𝑘 2 , 𝑈(𝑋 , 𝑂) = 𝑓(𝑥), |𝑈( 𝑋 , 𝑡 ) | < 𝑀, Where − ∞ < 𝑥 < ∞, 𝑡 > 0 .
∂𝑋

Solution:

The given partial differential equation is

2
∂𝑈 ∂𝑈
∂𝑡
= 𝑘 2 (1)
∂𝑋

Taking the complex form of the Fourier transformation with respect to x of both sides of (1), we
get
∞ ∞ 2
∂𝑈 𝑖α𝑥 ∂𝑈 𝑖α𝑥
∫ ∂𝑡
𝑒 𝑑𝑥 = ∫ 𝑘 2 𝑒 𝑑𝑥 (2)
−∞ −∞ ∂𝑋

Let,
∞
𝑖α𝑥
𝐼(𝑈) = ∫ 𝑈( 𝑋, 𝑡)𝑒 𝑑𝑥
−∞

𝑑Ƒ(𝑈)
∴ 𝑑𝑡

∞
∂𝑈 𝑖α𝑥
= ∫ ∂𝑡
𝑒 𝑑𝑥
−∞

∞ 2
∂𝑈 𝑖α𝑥
= ∫ 𝑘 2 𝑒 𝑑𝑥 𝑈𝑠𝑖𝑛𝑔 (2)
−∞ ∂𝑥

∞
𝑖α𝑥 ∂𝑈 ∞ 𝑖α𝑥 ∂𝑈
= 𝑘[𝑒 ∂𝑥 ]−∞ − 𝑘 𝑖α ∫ 𝑒 ∂𝑥
𝑑𝑥
−∞

∞
𝑖α𝑥 ∂𝑈
= 0 − 𝑘 𝑖α ∫ 𝑒 ∂𝑥
𝑑𝑥
−∞

114
 ∂𝑈
Since ∂𝑥
→0 𝑎𝑠 𝑥→ ∞

∞
𝑖α𝑥 ∞ 2 2 𝑖α𝑥
= − 𝑘𝑖α[𝑒 𝑈(𝑥 , 𝑡)]−∞ + 𝑘𝑖 α ∫ 𝑈( 𝑋, 𝑡)𝑒 𝑑𝑥
−∞

∞
2 𝑖α𝑥
= 0 − 𝑘α ∫ 𝑈( 𝑋, 𝑡)𝑒 𝑑𝑥
−∞

2
Since 𝑢→0as 𝑥→∞ and 𝑖 =− 1
∞
2 𝑖α𝑥
=− 𝑘α ∫ 𝑈( 𝑋, 𝑡)𝑒 𝑑𝑥
−∞

2
=− 𝐼𝑘α Ƒ(𝑈)

𝑑Ƒ(𝑈) 2
∴ 𝑑𝑡
=− 𝑘α Ƒ(𝑈)

𝑑Ƒ(𝑈) 2
𝑜𝑟 𝑑𝑡
=− 𝑘α 𝑑𝑡 (3)

Integrating both sides of (3), we get

2
log 𝑙𝑜𝑔 Ƒ(𝑈) = − 𝑘α 𝑡 + log 𝑙𝑜𝑔 𝐶, 𝐶 𝑏𝑒𝑖𝑛𝑔 𝑠𝑜𝑚𝑒 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑜𝑓 𝑠𝑜𝑚𝑒 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛.
2
−𝑘α𝑡 −𝑘α 𝑡
𝑜𝑟, log 𝑙𝑜𝑔 Ƒ(𝑈) = 𝑙𝑜𝑔𝑒 + log 𝑙𝑜𝑔 𝐶 = 𝑙𝑜𝑔𝐶𝑒

2
−𝑘α 𝑡
∴Ƒ(𝑈) = 𝐶𝑒 (4)

2
−𝑘α 𝑡
This is, Ƒ(𝑈(𝑥, 𝑡)) = 𝐶𝑒 (5)

𝑊ℎ𝑒𝑛 𝑡 = 0, 𝑈(𝑥, 0) = 𝑓(𝑥)

0
Ƒ𝐼(𝑈(𝑥, 0)) = Ƒ{𝑓(𝑥)} = 𝐶𝑒 = 𝐶

Therefore,
∞
𝑖α𝑥
Ƒ𝐶 = 𝐼{𝑓(𝑥)} = 𝐶 = ∫ 𝑈(𝑥, 0)𝑒 𝑑𝑥
−∞

∞
𝑖α𝑥
= ∫ 𝑓(𝑥)𝑒 𝑑𝑥
−∞

∞
𝑖α𝑥
= ∫ 𝑓(𝑢)𝑒 𝑑𝑢
−∞

115
Thus from (4) we have
2
−𝑘α 𝑡
Ƒ𝐼(𝑈(𝑥, 𝑡)) = Ƒ{𝑓(𝑥)}𝑒 (6)

Taking the inverse Fourier transform, we get

∞ 2 ∞ ∞ 2
1 −𝑘α 𝑡 −𝑖α𝑥 1 𝑖α𝑢 −𝑘α 𝑡 −𝑖α𝑥
𝑈(𝑥, 𝑡) = 2π
∫ Ƒ{𝑓(𝑥)}𝑒 𝑒 𝑑α = 2π
∫ [ ∫ 𝑓(𝑢)𝑒 𝑑𝑢]𝑒 𝑒 𝑑α
−∞ −∞ −∞

∞ ∞ 2
1 −𝑘α 𝑡−𝑖α(𝑥−𝑢)
= 2π
∫ 𝑓(𝑢)[ ∫ 𝑒 𝑑α]𝑑𝑢
−∞ −∞

2
∞ ∞ −𝑡[ 𝑘α+
𝑖(𝑥−𝑢) 2
]−
(𝑥−𝑢)
1 2 𝑘𝑡 4𝑘𝑡
= 2π
∫ 𝑓(𝑢)[ ∫ 𝑒 𝑑α]𝑑𝑢
−∞ −∞

2
(𝑥−𝑢)

=
1
∞
∫ 𝑓(𝑢)[ ∫ 𝑒
∞
⎣ {
−⎡⎢ 𝑡 𝑘α+
𝑖(𝑥−𝑢) 2
2 𝑘𝑡 ⎦}
] 𝑑α⎤⎥𝑒
− 4𝑘𝑡

𝑑𝑢 (7)
2π
−∞ −∞

Putting 𝑡 𝑘α + { 𝑖(𝑥−𝑢)
2 𝑘𝑡 }= 𝑦
1
So that 𝑘𝑡𝑑α = 𝑑𝑦 ∴𝑑α = 𝑑𝑦.
𝑘𝑡

Now,
2
∞
∫ 𝑒
−⎡⎢ 𝑡 𝑘α+
⎣ {
𝑖(𝑥−𝑢) ⎤
2 𝑘𝑡 ⎥⎦ } 𝑑α
−∞

= 2∫ 𝑒
∞ −⎡ 𝑡 𝑘α+ 𝑖(𝑥−𝑢) ⎤
⎢
⎣ 2 𝑘𝑡 ⎥⎦ { } 𝑑α
0

∞ 2
−𝑦 𝑑𝑦
= 2∫ 𝑒
𝑘𝑡
0

∞ 2
2 −𝑦
= ∫𝑒 𝑑𝑦
𝑘𝑡
0

2 π
= . 2
𝑘𝑡

π
= 𝑘𝑡

Therefore from (7) we get

116
 2
∞ (𝑥−𝑢)
1 π − 4𝑘𝑡
𝑈(𝑥, 𝑡) = 2π
∫ 𝑓(𝑢). 𝑘𝑡
.𝑒 𝑑𝑢.
−∞

2
∞ (𝑥−𝑢)
1 π − 4𝑘𝑡
= 2π
. 𝑘𝑡
. ∫ 𝑓(𝑢). 𝑒 𝑑𝑢.
−∞

2
∞ (𝑥−𝑢)
1 − 4𝑘𝑡
= ∫ 𝑓(𝑢). 𝑒 𝑑𝑢. (8)
4π𝑘𝑡
−∞

Now if we change variables from u to z according to the transformation

2
(𝑥−𝑢) 2 𝑥−𝑢
4𝑘𝑡
= 𝑧 𝑜𝑟 , =𝑧
2 𝑘𝑡

𝑢 = (𝑥 − 2 𝑘𝑡𝑧)

𝑑𝑢 =− 2 𝑘𝑡𝑧)

Limits.

u=∞ u=-∞
z=-∞ z=∞

Thus from (8) we have

∞ 2
1 −𝑧
𝑈(𝑥, 𝑡) = ∫ 𝑓(𝑥 − 2𝑧 𝑘𝑡). 𝑒 . − 2 𝑘𝑡𝑑𝑧.
4π𝑘𝑡
−∞

∞ 2
2 𝑘𝑡 −𝑧
=− ∫ 𝑓(𝑥 − 2𝑧 𝑘𝑡). 𝑒 𝑑𝑧
2 π 𝑘𝑡
−∞

∞ 2
1 −𝑧
= ∫ 𝑓(𝑥 − 2𝑧 𝑘𝑡). 𝑒 𝑑𝑧
π
−∞

Example 11: Solve ∂𝑈

∂𝑡
=
2
∂𝑈
2 , 0 < 𝑥 < 6, 𝑡 > 0
∂𝑥

𝑈𝑥(0, 𝑡) = 0, 𝑈𝑥(6, 𝑡) = 0, 𝑈(𝑥, 0) = 2𝑥.

Solution: The given partial differential equation is,

117
 2
∂𝑈 ∂𝑈
∂𝑡
= 2 (1)
∂𝑥

According to the given boundary condition, here the finite Fourier cosine transform is more
useful. Taking the finite Fourier cosine transform (l=6) of both sides of (1) we get,
6 6 2
∂𝑈 𝑛π𝑥 ∂𝑈 𝑛π𝑥
∫ ∂𝑡
𝑐𝑜𝑠 6
𝑑𝑥 = ∫ 2 cos 𝑐𝑜𝑠 6
𝑑𝑥 (2)
0 0 ∂𝑥

6
𝑛π𝑥
Let, 𝑣 = 𝑣(𝑛, 𝑡) = ∫ 𝑈(𝑥, 𝑡)𝑐𝑜𝑠 6
𝑑𝑥 = 𝑓𝑐(𝑈)
0

6
𝑑𝑣 ∂𝑈 𝑛π𝑥
∴ 𝑑𝑡
=∫ ∂𝑡
𝑐𝑜𝑠 6
𝑑𝑥
0

6 2
∂𝑈 𝑛π𝑥
=∫ 2 𝑐𝑜𝑠 6
𝑑𝑥
0 ∂𝑥

Note:
𝑙

{ }
2 2 2 2
∂𝑈 ∂𝑈 𝑛π𝑥 ∂𝑈(𝑙,𝑡) ∂𝑈(0,𝑡) 𝑛π
𝑓𝑐 2 =∫ 2 𝑐𝑜𝑠 𝑙
𝑑𝑥 = ∂𝑥
cos 𝑐𝑜𝑠 𝑛π − ∂𝑥
− 2 𝑓𝑐(𝑈)
∂𝑥 0 ∂𝑥 𝑙

6 2 2 2
∂𝑈 𝑛π𝑥 ∂𝑈(𝑙,𝑡) ∂𝑈(0,𝑡) 𝑛π
∴∫ 2 𝑐𝑜𝑠 6
𝑑𝑥 = ∂𝑥
cos 𝑐𝑜𝑠 𝑛π − ∂𝑥
− 2 𝑓𝑐(𝑈)
0 ∂𝑥 𝑙

2 2
𝑛π
= 𝑈𝑥(6, 𝑡) cos 𝑐𝑜𝑠 𝑛π − 𝑈𝑥(0, 𝑡) − 2 𝑓𝑐(𝑈)
𝑙

2 2
𝑛π
= 𝑈𝑥(6, 𝑡) cos 𝑐𝑜𝑠 𝑛π − 𝑈𝑥(0, 𝑡) − 2 𝑣
𝑙

6
𝑛π𝑥
Since, 𝑓𝑐(𝑈) = ∫ 𝑈(𝑥, 𝑡)𝑐𝑜𝑠 6
𝑑𝑥
0

∴ 𝑓𝑐(𝑈) = 𝑣

2 2
𝑑𝑣 𝑛π
Therefore, 𝑑𝑡
= 𝑈𝑥(6, 𝑡) − 𝑈𝑥(0, 𝑡) − 2 𝑣
6

2 2
𝑑𝑣 𝑛π
Or, 𝑑𝑡
= 0− 0 − 2 𝑣 Since, 𝑈𝑥(0. 𝑡) = 𝑈𝑥(6, 𝑡) = 0
6

118
 2 2
𝑑𝑣 𝑛π
𝑣
=− 2 𝑑𝑡 (3)
6

2 2
𝑛π
Integrating both sides, we get log 𝑙𝑜𝑔 𝑣 = − 2 𝑡 + log 𝑙𝑜𝑔 𝐴 , A being some constant of
6
integration.
2 2 2 2
𝑛π 𝑛π
− 2 𝑡 − 2 𝑡
Or, log 𝑙𝑜𝑔 𝑣 = 𝑙𝑜𝑔 6
+ log 𝑙𝑜𝑔 𝐴 = 𝑙𝑜𝑔𝐴𝑒 6

2 2
𝑛π
− 2 𝑡
𝑣 = 𝐴𝑒 6
(4)
0
When, 𝑡 = 𝑜 R.H.S of (4) 𝐴𝑒 = 𝐴
∞
𝑛π𝑥
When, 𝑡 = 0 R.H.S of (4) = ∫ 𝑈(𝑥, 0) cos 𝑐𝑜𝑠 6
𝑑𝑥
−∞

6
𝑛π𝑥
= ∫ 2𝑥 cos 𝑐𝑜𝑠 6
𝑑𝑥
0

6 6
6 𝑛π𝑥 6 6 𝑛π𝑥 12 𝑛π𝑥
= ⎡2𝑥 𝑠𝑖𝑛 ⎤ −2 ∫ 𝑠𝑖𝑛 𝑑𝑥 = 0 − ∫ 𝑠𝑖𝑛 𝑑𝑥
⎣ 𝑛π 6 ⎦
0
𝑛π
0
6 𝑛π
0
6

12 6 𝑛π𝑥 6
= ⎡cos 𝑐𝑜𝑠 ⎤
𝑛π 𝑛π ⎣ 6 ⎦0

72
= 2 2 (cos 𝑐𝑜𝑠 𝑛π − 1) ⁡
𝑛π

72
Therefore from (4), we get 𝐴 = 2 2 (cos 𝑐𝑜𝑠 𝑛π − 1)
𝑛π

Putting the value of A in (4), we get

2 2
𝑛π
− 𝑡
72 6
2

𝑣 = 𝐴= 2 2 (cos 𝑐𝑜𝑠 𝑛π − 1) 𝑒
𝑛π

Taking the inverse Fourier cosine transform, we get

∞
1 2 𝑛π𝑥
𝐹(𝑥) = 𝑙
𝑓𝑐(0) + 𝑙
∑ 𝑓𝑐(𝑛)𝑐𝑜𝑠 𝑙
𝑛=1

∞
1 2 𝑛π𝑥
That is, 𝑈(𝑥, 𝑡) = 6
𝑓𝑐(0) + 6
∑ 𝑣𝑐𝑜𝑠 𝑙
𝑛=1

119
 2 2
∞ −
𝑛π𝑡
1 1 72 2
𝑛π𝑥
Or, 𝑈(𝑥, 𝑡) = 6
𝑓𝑐(0) + 3
∑ 2 2 (cos 𝑐𝑜𝑠 𝑛π − 1) 𝑒 6
𝑐𝑜𝑠 6
(6)
𝑛=1 𝑛 π

𝑙 6
𝑛π𝑥 𝑛π𝑥
𝑓𝑐(𝑛) = ∫ 𝐹(𝑥)𝑐𝑜𝑠 𝑙
𝑑𝑥 = ∫ 𝐹(𝑥)𝑐𝑜𝑠 6
𝑑𝑥
0 0

𝑙 6
𝑓𝑐(0) = ∫ 𝐹(𝑥)𝑑𝑥 = ∫ 𝑈(𝑥, 0)𝑑𝑥
0 0

6 6
∫ 2𝑥𝑑𝑥 = 𝑥 [ 2]0 = 36
0

Thus from (5) we get,

2 2
∞ −
𝑛π𝑡
36 24 (cos𝑐𝑜𝑠 𝑛π−1) 6
2
𝑛π𝑥
𝑈(𝑥, 𝑡) = 6
+ 2 ∑ 2 𝑒 𝑐𝑜𝑠 6
π 𝑛=1 𝑛

2 2
∞ −
𝑛π𝑡
24 (cos𝑐𝑜𝑠 𝑛π−1) 2
𝑛π𝑥
Or, 𝑈(𝑥, 𝑡) = 6 + 2 ∑ 2 𝑒 6
𝑐𝑜𝑠 6
π 𝑛=1 𝑛

Example 12: Use the method of Fourier transform to determine the displacement 𝑦(𝑥, 𝑡) of an
infinite string, given that the string is initially at rest and that the initial displacement is
𝑓(𝑥), − ∞ < 𝑥 < ∞. Also show that the solution can be put in the form
1
𝑦(𝑥, 𝑡) = 2
[𝑓(𝑥 + 𝑐𝑡) + 𝑓(𝑥 − 𝑐𝑡)]

Solution: Here we have to solve the one-dimensional wave equation

2 2
∂𝑦 2 ∂𝑦
2 =𝑐 2 , − ∞ < 𝑥 < ∞, 𝑡 > 0
∂𝑡 ∂𝑥

Subject to the following initial conditions

𝑦(𝑥, 0) = 𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑑𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 = 𝑓(𝑥)

And 𝑦𝑡(𝑥, 0) = 𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦 = 0

The given partial differential equation is

2 2
∂𝑦 2 ∂𝑦
2 =𝑐 2 …………………………………………………….. (1)
∂𝑡 ∂𝑥

Taking the complex Fourier transform of both sides of (1) , we have

120
∞ 2 2
∂𝑦 −𝑖𝑢𝑥 2 ∂𝑦 −𝑖𝑢𝑥
∫ 2 𝑒 𝑑𝑥 = 𝑐 2 𝑒 𝑑𝑥………………………………(2)
−∞ ∂𝑡 ∂𝑥

𝑛
By the Fourier transform of the derivative of a function we have if 𝐹 (𝑥) is the nth-derivative of
𝐹(𝑥) and the first (𝑛 − 1) derivatives of 𝐹(𝑥) vanish as →±∞ , then
𝑛 𝑛
{ }
𝐹 𝐹 (𝑥) = (− 𝑖𝑢) 𝐹{𝑦(𝑥, 𝑡)} .

Thus from(2), we have

2 ∞
𝑑 −𝑖𝑢𝑥 2 2
2 ∫ 𝑦𝑒 𝑑𝑥 = 𝑐 (− 𝑖𝑢) 𝐹{𝑦(𝑥, 𝑡)}
𝑑𝑡 −∞

2 ∞
𝑑 −𝑖𝑢𝑥 2 2
𝑜𝑟, 2 ∫ 𝑦(𝑥, 𝑡)𝑒 𝑑𝑥 = 𝑐 (− 𝑢) 𝐹{𝑦(𝑥, 𝑡)}
𝑑𝑡 −∞

2
𝑑 2 2
2 𝐹{𝑦(𝑥, 𝑡)} =− 𝑐 𝑢 𝐹{𝑦(𝑥, 𝑡)}
𝑑𝑡

2
𝑑𝑦 2 2
2 =− 𝑐 𝑢 𝑦
𝑑𝑡

Where
∞
−𝑖𝑢𝑥
𝑦 = 𝑦(𝑢, 𝑡) = 𝐹{𝑦(𝑥, 𝑡)} = ∫ 𝑦(𝑥, 𝑡)𝑒 𝑑𝑥
−∞

2
𝑑𝑦 2 2
𝑜𝑟, 2 +𝑐 𝑢 𝑦= 0 (3)
𝑑𝑡

Which is the ordinary second order differential equation whose solution is

𝑦 = 𝑦(𝑢, 𝑡) = 𝐴𝑐𝑜𝑠(𝑐𝑢𝑡) + 𝐵𝑠𝑖𝑛(𝑐𝑢𝑡) (4)

Differentiating both sides with respect to , we get

𝑦𝑡(𝑢, 𝑡) =− 𝐴𝑐𝑢𝑠𝑖𝑛(𝑐𝑢𝑡) + 𝐵𝑐𝑢𝑐𝑜𝑠(𝑐𝑢𝑡) (5)

Also from the initial given conditions, we have

𝑦(𝑥, 0) = 𝑓(𝑥) (6)

𝑦𝑡(𝑥, 0) = 0 (7)

Taking the Fourier transform of (6) and (7) we get

121
 ∞ ∞
−𝑖𝑢𝑥 −𝑖𝑢𝑥
𝑦(𝑢, 0) = ∫ 𝑦(𝑥, 0)𝑒 𝑑𝑥 = ∫ 𝑓(𝑥)𝑒 𝑑𝑥 = 𝑓(𝑢) (𝑠𝑎𝑦)
−∞ −∞

∴𝑦(𝑢, 0) = 𝑓(𝑢)⋯⋯⋯(8)

And
∞ ∞
−𝑖𝑢𝑥 −𝑖𝑢𝑥
𝑦𝑡(𝑢, 0) = ∫ 𝑦𝑡(𝑥, 0)𝑒 𝑑𝑥 = ∫ 0𝑒 𝑑𝑢 = 0
−∞ −∞

∴𝑦𝑡(𝑢, 0) = 0⋯⋯⋯(9)

Putting 𝑡 = 0 in(5) , we have 𝑦𝑡(𝑢, 0) = 𝐵𝑐𝑢

𝑜𝑟, 𝐵𝑐𝑢 = 0 𝑢𝑠𝑖𝑛𝑔 (9)

∴𝐵 = 0 𝑠𝑖𝑛𝑐𝑒 𝑐𝑢≠0

Again putting 𝑡 = 0 in(4), we get

𝑦(𝑢, 0) = 𝐴

∴𝑓(𝑢) = 𝐴 𝑢𝑠𝑖𝑛𝑔 (8)

𝑜𝑟, 𝐴 = 𝑓(𝑢)

Putting the values of 𝐴 and 𝐵 in (4) , we get

𝑦 = 𝑦(𝑢, 𝑡) = 𝑓(𝑢)𝑐𝑜𝑠(𝑐𝑢𝑡)⋯⋯⋯(10)

Taking the inverse Fourier transform of (10) we have

∞
1 𝑖𝑢𝑥
𝑦(𝑥, 𝑡) = 2π
∫ 𝑓(𝑢)𝑐𝑜𝑠(𝑐𝑢𝑡)𝑒 𝑑𝑢
−∞

( )𝑒
𝑖𝑐𝑢𝑡 −𝑖𝑐𝑢𝑡
1 𝑒 +𝑒 𝑖𝑢𝑥
𝑜𝑟, 𝑦(𝑥, 𝑡) = 2π
∫ 𝑓(𝑢) 2
𝑑𝑢
−∞

∞ ∞
⎡ 1 𝑖𝑢(𝑥+𝑐𝑡) 1 𝑖𝑢(𝑥−𝑐𝑡) ⎤
= ⎢ 2π ∫ 𝑓(𝑢)𝑒 𝑑𝑢 + ∫ 𝑓(𝑢)𝑒 𝑑𝑢⎥
⎢ −∞ 2π ⎥
⎣ −∞ ⎦
1
= 2
[𝑓(𝑥 + 𝑐𝑡) + 𝑓(𝑥 − 𝑐𝑡)] (using the definition of inverse Fourier transform)

122
 1
∴𝑦(𝑥, 𝑡) = 2
[𝑓(𝑥 + 𝑐𝑡) + 𝑓(𝑥 − 𝑐𝑡)]

Example 13: A thin membrane of great extend is released from rest in position z=f(x,y).Show
that the displacement at any subsequent tike is given by,
∞ ∞
1 2 2 −𝑖(𝑢𝑥+𝑣𝑦)
𝑧(𝑥, 𝑦, 𝑡) = 2π
∫ ∫ 𝐹(𝑢, 𝑣)𝑐𝑜𝑠 𝑐𝑡 𝑢 + 𝑣 𝑒 𝑑𝑢𝑑𝑣
−∞ −∞

Where 𝐹(𝑢, 𝑣)is the double Fourier transform of 𝑓(𝑥, 𝑦)

Proof: Here the displacement of the membrane is governed by two dimensional wave equations

( )……………………(1)
2 2 2
∂𝑧 2 ∂𝑧 ∂𝑧
2 =𝑐 2 + 2
∂𝑡 ∂𝑥 ∂𝑦

Taking the double Fourier transform of both sides of (1) we get

∞ ∞ ∞ ∞

( )𝑒
2 2 2 2
1 ∂𝑧 𝑖(𝑢𝑥+𝑣𝑦) 𝑐 ∂𝑧 ∂𝑧 𝑖(𝑢𝑥+𝑣𝑦)
2π
∫ ∫ 2 𝑒 𝑑𝑥𝑑𝑦 = 2π
∫ ∫ 2 + 2 𝑑𝑥𝑑𝑦
−∞ −∞ ∂ 𝑡 −∞ −∞ ∂ 𝑥 ∂𝑦

∞ ∞ ∞ ∞

( )
2 2 2 2 2
𝑖(𝑢𝑥+𝑣𝑦) 𝑖(𝑢𝑥+𝑣𝑦) 2 2 2
⇒
𝑑
𝑑𝑡
2
1
2π
∫ ∫ 𝑧𝑒 𝑑𝑥𝑑𝑦 =
𝑐
2π
∫ ∫
∂𝑧
2 +
∂𝑧
2 𝑒 𝑑𝑥𝑑𝑦⟹
𝑑𝑧
2 { }
= 𝑐 (− 𝑖𝑢) + (− 𝑖𝑣) 𝐹{(𝑧(𝑥
−∞ −∞ −∞ −∞ ∂ 𝑥 ∂𝑦 𝑑𝑡

𝑤ℎ𝑒𝑟𝑒 𝑧 = 𝑧(𝑢, 𝑣, 𝑡) = 𝐹{𝑧(𝑥, 𝑦, 𝑡)}

∞ ∞
1 𝑖(𝑢𝑥+𝑣𝑦)
= 2π
∫ ∫ 𝑧𝑒 𝑑𝑥𝑑𝑦
−∞ −∞

2
𝑑𝑧 2 2 2
⟹ 2 =− 𝑐 (𝑢 + 𝑣 ) 𝑧
𝑑𝑡

2
2 2 2
⟹
𝑑𝑧
𝑑𝑡
2 ( )
+ 𝑐 𝑢 + 𝑣 𝑧 = 0…………………………………………………………. (2)

Which is a ordinary differential equation whose solution is

⎰ 2 2 ⎱ ⎰ ⎱
𝑧 = 𝐴 cos 𝑐𝑜𝑠 𝑐 (𝑢 + 𝑣 )𝑡 + 𝐵𝑠𝑖𝑛 𝑐
⎱ ⎰ ⎱ (𝑢2 + 𝑣2)𝑡⎰……………………………….. (3)
The given initial conditions are
∂𝑧
𝑧 = 𝑓(𝑥, 𝑦) 𝑎𝑛𝑑 ∂𝑡
= 0 𝑎𝑡 𝑡 = 0

123
Taking the Fourier transform of these initial conditions, we get
∞ ∞
1 𝑖(𝑢𝑥+𝑣𝑦)
𝑧= 2π
∫ ∫ 𝑓(𝑥, 𝑦)𝑒 𝑑𝑥𝑑𝑦 = 𝐹(𝑢, 𝑣) (4)
−∞ −∞

∞ ∞
1 ∂𝑧 𝑖(𝑢𝑥+𝑣𝑦)
= 2π
∫ ∫ ∂𝑡
𝑒 𝑑𝑥𝑑𝑦
−∞ −∞

∞ ∞
1 ∂𝑧 𝑖(𝑢𝑥+𝑣𝑦)
= 2π
∫ ∫ ∂𝑡
𝑒 𝑑𝑥𝑑𝑦
−∞ −∞

∞ ∞
𝑑 1 𝑖(𝑢𝑥+𝑣𝑦)
= 𝑑𝑡 2π
∫ ∫ 𝑧𝑒 𝑑𝑥𝑑𝑦
−∞ −∞

𝑑𝑧
∴ 𝑑𝑡
= 0 𝑎𝑡 𝑡 = 0

When 𝑡 = 0, 𝑐𝑜𝑚𝑏𝑖𝑛𝑔 (3)𝑎𝑛𝑑 (4), 𝑤𝑒 𝑔𝑒𝑡

𝐴 = 𝐹(𝑢, 𝑣)

2 2
Also,
𝑑𝑧
𝑑𝑡
=− 𝐴𝑐 (𝑢 + 𝑣 ) sin 𝑠𝑖𝑛 𝑐 (𝑢2 + 𝑣2)𝑡 + 𝐵𝑐 2 2 2 2
(𝑢 + 𝑣 ) cos 𝑐𝑜𝑠 𝑐 (𝑢 + 𝑣 )𝑡

𝑑𝑧 2 2
∴0 = ( 𝑑𝑡 ) = 𝐵𝑐 𝑢 + 𝑣
𝑡=0

𝑜𝑟, 𝐵 = 0

Putting the value of A and B in (3), we get

⎰ 2 2 ⎱
𝑧 = 𝐹(𝑢, 𝑣) cos 𝑐𝑜𝑠 𝑐 𝑢 + 𝑣 𝑡 (5)
⎱ ⎰

Now applying the inversion formula for double Fourier transform, we have
∞ ∞
1 2 2 −𝑖(𝑢𝑥+𝑣𝑦)
(𝑥, 𝑦, 𝑡) = 2π
∫ ∫ 𝐹(𝑢, 𝑣) cos 𝑐𝑜𝑠 {𝑐 𝑢 + 𝑣 𝑡} 𝑒 𝑑𝑢𝑑𝑣
−∞ −∞

This is the required displacement at any subsequent time t.

…………………………………..The End…………………………………………

124
125