Fourier Tranform

Dirichlet conditions
Fourier Transform
The Dirichlet condition for periodic function are;
Syllabus:- i) The function must be periodic.
Definition, Fourier integral, Fourier transform, inverse ii) The function has finite number of discontinuities in
transform, Fourier transform of derivatives, convolution each period
(mathematical statement only), Parseval’s theorem iii) The function has finite number of maxima and
(statement only), Applications minima in each period.
iv) The function must converge over any period. That is
Fourier series 𝑇

∫|𝑓(𝑡)| 𝑑𝑡 is finit𝑒
Any periodic function 𝑓(𝑡) having period T satisfying
0
Dirichlet condition can be expressed by the following series
∞
For periodic function 𝑓(𝑥) having period L satisfying
𝑎0 Dirichlet condition can be expressed by the following series
𝑓(𝑡) = + ∑(𝑎𝑛 cos 𝑛𝜔𝑡 + 𝑏𝑛 sin 𝑛𝜔𝑡) ∞
2 𝑎0 2𝜋𝑛𝑥 2𝜋𝑛𝑥
𝑛=1
𝑓(𝑥) = + ∑ (𝑎𝑛 cos ( ) + 𝑏𝑛 sin ( ))
2 𝐿 𝐿
Where 𝑛=1

𝑇 Where
1
𝑎0 = ∫ 𝑓(𝑡)𝑑𝑡 , 𝑎𝑛 𝑥0 +𝐿
𝑇 1
0
𝑇 𝑎0 = ∫ 𝑓(𝑥)𝑑𝑡 ,
2 𝐿
𝑥0
= ∫ 𝑓(𝑡) cos 𝑛𝜔𝑡 𝑑𝑡
𝑇 𝑥0 +𝐿
0
2 2𝜋𝑛𝑥
𝑇 𝑎𝑛 = ∫ 𝑓(𝑥) cos ( ) 𝑑𝑥
2 𝐿 𝐿
& 𝑏𝑛 = ∫ 𝑓(𝑡) sin 𝑛𝜔𝑡 𝑑𝑡 𝑥0
𝑇
0 𝑥0 +𝐿
2 2𝜋𝑛𝑥
& 𝑏𝑛 = ∫ 𝑓(𝑥) sin ( ) 𝑑𝑥
𝐿 𝐿
𝑥0

pg. 1 Prof. Avadhut Surekha Prakash Manage,

Bhaurao Kakatkar College, Belgaum
Fourier Tranform

Fourier integral Fourier Sine and Cosine transform

For periodic function 𝑓(𝑥) defined in interval[−𝑙, 𝑙] is 1) Fourier Cosine transform
represented in integral form as
1 ∞ ∞
𝑓(𝑥) = 𝜋 ∫0 𝑑𝜔 ∫0 𝑓(𝑡) cos 𝜔(𝑡 − 𝑥) 𝑑𝑡 If 𝑓(𝑡) is even function ie. 𝑓(−𝑡) = 𝑓(𝑡)
Then Fourier cosine transform is given by
Is called as Fourier integral.
If cos 𝜔(𝑡 − 𝑥) is replaced in complex form, 2 ∞
𝐹𝐶 (𝜔) = √ ∫ 𝑓(𝑡) cos 𝜔𝑡 𝑑𝑡
1 ∞ ∞ 𝜋 0
𝑓(𝑥) = 2𝜋 ∫−∞ 𝑒 𝑖𝜔𝑥 𝑑𝜔 ∫−∞ 𝑓(𝑡) 𝑒 −𝑖𝜔𝑡 𝑑𝑡
And inverse Fourier cosine transform is given by
is called Fourier complex integral.
2 ∞
𝑓(𝑡) = √ ∫ 𝐹𝐶 (𝜔) cos 𝜔𝑡 𝑑𝜔
Fourier transform 𝜋 0
The complex form of Fourier integral
1 ∞ ∞ 2) Fourier Sine transform
𝑓(𝑥) = 2𝜋 ∫−∞ 𝑒 𝑖𝜔𝑥 𝑑𝜔 ∫−∞ 𝑓(𝑡) 𝑒 −𝑖𝜔𝑡 𝑑𝑡
If 𝑓(𝑡) is odd function ie. 𝑓(−𝑡) = −𝑓(𝑡)
This may be expressed as ,
Then Fourier sine transform is given by
∞
1
𝑓(𝑥) = ∫ 𝑒 𝑖𝜔𝑥 𝐹(𝜔)𝑑𝜔 … … . . (1) 2 ∞
√2𝜋 −∞ 𝐹𝑆 (𝜔) = √𝜋 ∫0 𝑓(𝑡) sin 𝜔𝑡 𝑑𝑡
1 ∞
Where 𝐹(𝜔) = ∫ 𝑒 −𝑖𝜔𝑡
√2𝜋 −∞
𝑓(𝑡)𝑑𝑡 … … … (2) And inverse Fourier sine transform is given by
The function 𝐹(𝜔) is called Fourier transform of 𝑓(𝑡) and 𝑓(𝑡)
2 ∞
is called as inverse transform of 𝐹(𝜔). 𝑓(𝑡) = √ ∫ 𝐹𝑆 (𝜔) sin 𝜔𝑡 𝑑𝜔
𝜋 0
Eqn (1) is transform of frequency function 𝐹(𝜔) into position
function 𝑓(𝑥).
Eqn (2) is transform of time function 𝑓(𝑡) into frequency
function 𝐹(𝜔).
pg. 2 Prof. Avadhut Surekha Prakash Manage,
Bhaurao Kakatkar College, Belgaum
Fourier Tranform

Properties of Fourier transform vi) Convolution property ( Convolution Theorem) :-

i) Linear property :- If function 𝐹(𝜔) is called Fourier The convolution of two function 𝑓(𝑡) and 𝑔(𝑡) over the
transform of 𝑓(𝑡) & interval (−∞ , ∞) is defined as
If 𝑓(𝑡) = 𝑎1 𝑓1 (𝑡) + 𝑎2 𝑓2 (𝑡) + ………….+𝑎𝑛 𝑓𝑛 (𝑡) = 1 ∞
𝑓(𝑡) *𝑔(𝑡) = ∫ 𝑓(𝑢) 𝑔(𝑡 − 𝑢) 𝑑𝑡
∑ 𝑎𝑛 𝑓𝑛 (𝑡), √2𝜋 −∞

then the Fourier transform of 𝑓(𝑡) is given by Convolution theorem states that “If 𝐹(𝜔) and 𝐺(𝜔) be the
Fourier transforms of 𝑓(𝑡) & 𝑔(𝑡) respectively, then the Fourier
𝐹(𝜔) = 𝑎1 𝐹1 (𝜔) + 𝑎2 𝐹2 (𝜔) + ….+𝑎𝑛 𝐹𝑛 (𝜔) = ∑ 𝑎𝑛 𝐹𝑛 (𝜔) transform of the convolution of 𝑓(𝑡) & 𝑔(𝑡) is the product of
their Fourier transform”. That is

ii) Change of scale property:- If function 𝐹(𝜔) is called Fourier 𝐹. 𝑇 [𝑓(𝑡) ∗ 𝑔(𝑡)] = 𝐹. 𝑇 [𝑓(𝑡)] ∗ 𝐹. 𝑇 [𝑔(𝑡)] = 𝐹(𝜔)*𝐺(𝜔)
transform of 𝑓(𝑡), then Fourier transform of 𝑓(𝑎𝑡) is
1 𝜔
𝐹 (𝑎 )
𝑎 Parsvel’s theorem
iii) Shifting property:- If function 𝐹(𝜔) is called Fourier If 𝐹(𝜔) and 𝐺(𝜔) be the Fourier transforms of 𝑓(𝑡) & 𝑔(𝑡)
transform of 𝑓(𝑡), then Fourier transform of 𝑓(𝑡 − 𝑎) is respectively then,
𝑒 𝑖𝜔𝑎 𝐹(𝜔) ∞ ∞
i) ∫−∞ 𝑓(𝑡) 𝑔∗ (𝑡) 𝑑𝑡 = ∫−∞ 𝐹(𝜔) 𝐺 ∗ (𝜔) 𝑑𝑡
iv) Conjugate property :- If function 𝐹(𝜔) is called Fourier Where 𝐺 ∗ (𝜔) is Fourier transform of 𝑔∗ (𝑡) and 𝑔∗ (𝑡)is
transform of 𝑓(𝑡), then Fourier transform of complex complex conjugate of 𝑔(𝑡).
conjugate of 𝑓(𝑡) (ie. 𝑓 ∗ (𝑡)) is . 𝐹 ∗ (−𝜔) ( complex ∞ ∞
conjugate of 𝐹(−𝜔)) ii) ∫−∞ [𝑓(𝑡)2 ]𝑑𝑡 = ∫−∞ [𝐹(𝜔)2 ]𝑑𝜔
ie. 𝐹. 𝑇 [𝑓 ∗ (𝑡)] =. 𝐹 ∗ (−𝜔)

v) Modulation property :- If function 𝐹(𝜔) is called Fourier

transform of 𝑓(𝑡), then Fourier transform of 𝑓(𝑡) cos 𝑎𝑡 is
1
[𝐹(𝜔 + 𝑎) + 𝐹(𝜔 − 𝑎)]
2

pg. 3 Prof. Avadhut Surekha Prakash Manage,

Bhaurao Kakatkar College, Belgaum
Fourier Tranform

Derivative of Fourier transform 1 ∞

𝐹1 (𝜔) = [𝑓(𝑡)𝑒 −𝑖𝜔𝑡 ]−∞ + 𝑖𝜔𝐹(𝜔)
1 ∞ √2𝜋
𝐹(𝜔) = ∫ 𝑒 −𝑖𝜔𝑡 𝑓(𝑡)𝑑𝑡
√2𝜋 −∞ As 𝑡 → ∞, 𝑒 −𝑖𝜔𝑡 → 0 & 𝑡 → −∞, 𝑒 −𝑖𝜔𝑡 → ∞ . Therefore for
Differentiating both side with respect to 𝜔 existence of limit , the function 𝑓(𝑡) should be much faster
decrease to zero as 𝑡 → ∞ than 𝑡 → −∞ . Then in such case
𝑑𝐹(𝜔) 1 𝑑 ∞ −𝑖𝜔𝑡
= ∫ 𝑒 𝑓(𝑡)𝑑𝑡
𝑑𝜔 √2𝜋 𝑑𝜔 −∞ 1 ∞
[𝑓(𝑡)𝑒 −𝑖𝜔𝑡 ]−∞ = 0
1 ∞ 𝛿𝑦 √2𝜋
= ∫ [𝑒 −𝑖𝜔𝑡 𝑓(𝑡)]𝑑𝑡
√2𝜋 −∞ 𝛿𝜔 ∴ 𝐹1 (𝜔) = 𝑖𝜔𝐹(𝜔)
∞1
= −𝑖 ∫ 𝑡 𝑓(𝑡) 𝑒 −𝑖𝜔𝑡 𝑑𝑡 Similarly 𝐹𝑛 (𝜔) be the Fourier transform of nth derivative of
√2𝜋 −∞
function 𝑓(𝑡) then
= −𝑖 𝐹. 𝑇 [𝑡 𝑓(𝑡)]
𝐹𝑛 (𝜔) = (𝑖𝜔)𝑛 𝐹(𝜔)
If 𝐹(𝜔) differentiated n times then
𝑑𝑛 (𝜔)
== (−𝑖 )𝑛 𝐹. 𝑇 [𝑡 𝑛 𝑓(𝑡)]
𝑑𝜔 𝑛
Application of Fourier transform
1) Evaluation of integrals
Fourier transform of derivative
Using Fourier transform certain integrals may be obtained.
Let 𝐹1 (𝜔) be the Fourier transform of first derivative of
Ex. Find the Fourier sine and cosine transform of
function 𝑓(𝑡) then ∞ cos 𝜔𝑡
𝑓(𝑡) = 𝑒 −𝑝𝑡 , 𝑝 > 0 . Hence evaluate ∫0 𝑑𝜔 and
∞ 𝑝2 +𝜔2
1
𝑑𝑓(𝑡) −𝑖𝜔𝑡 ∞ 𝜔 sin 𝜔𝑡
𝐹1 (𝜔) = ∫ 𝑒 𝑑𝑡 ∫0 𝑑𝜔
√2𝜋 −∞ 𝑑𝑡 𝑝2 +𝜔2

Integrating by parts, we get Solution:- Given 𝑓(𝑡) = 𝑒 −𝑝𝑡 , 𝑝 > 0

∞ By taking cosine transform of 𝑓(𝑡)
1 ∞ 𝑖𝜔
𝐹1 (𝜔) = [𝑓(𝑡)𝑒 −𝑖𝜔𝑡 ]−∞ + ∫ 𝑓(𝑡)𝑒 −𝑖𝜔𝑡 𝑑𝑡
√2𝜋 √2𝜋 −∞

pg. 4 Prof. Avadhut Surekha Prakash Manage,

Bhaurao Kakatkar College, Belgaum
Fourier Tranform

2 ∞ 2 ∞
𝐹𝐶 (𝜔) = √ ∫ 𝑓(𝑡) cos 𝜔𝑡 𝑑𝑡 𝐹𝑆 (𝜔) = √ ∫ 𝑓(𝑡) sin 𝜔𝑡 𝑑𝑡
𝜋 0 𝜋 0

2 ∞
= √𝜋 ∫0 𝑒 −𝑝𝑡 cos 𝜔𝑡 𝑑𝑡
2 ∞
2 𝑃 = √ ∫ 𝑒 −𝑝𝑡 sin 𝜔𝑡 𝑑𝑡
= √𝜋 … . (1) 𝜋 0
𝑝2 +𝜔2

By taking inverse cosine transform 𝐹𝐶 (𝜔) 2 𝜔

= √𝜋 𝑝2 +𝜔2 … . (3)

2 ∞
𝑓(𝑡) = √ ∫ 𝐹 (𝜔) cos 𝜔𝑡 𝑑𝜔 By taking inverse sine transform 𝐹𝐶 (𝜔)
𝜋 0 𝐶
2 ∞
𝑓(𝑡) = √ ∫ 𝐹 (𝜔) sin 𝜔𝑡 𝑑𝜔
𝜋 0 𝑆
2 ∞ 2 𝑃
𝑓(𝑡) = √ ∫ √ cos 𝜔𝑡 𝑑𝜔
𝜋 0 𝜋 𝑝 + 𝜔2
2

2 ∞ 2 𝜔
𝑓(𝑡) = √ ∫ √ sin 𝜔𝑡 𝑑𝜔
𝜋 0 𝜋 𝑝2 + 𝜔 2
2𝑃 ∞ cos 𝜔𝑡
𝑓(𝑡) = ∫ 𝑑𝜔
𝜋 0 𝑝2 + 𝜔 2
2 ∞ 𝜔 sin 𝜔𝑡
𝑓(𝑡) = ∫ 𝑑𝜔
∞
𝜋 0 𝑝2 + 𝜔 2
cos 𝜔𝑡 𝜋 𝜋 −𝑝𝑡
∴∫ 𝑑𝜔 = 𝑓(𝑡) = 𝑒 … … (2)
0 𝑝2 + 𝜔 2 2𝑃 2𝑃
∞
𝜔 sin 𝜔𝑡 𝜋 𝜋
∴∫ 𝑑𝜔 = 𝑓(𝑡) = 𝑒 −𝑝𝑡 … … (4)
0 𝑝2 + 𝜔 2 2 2
By taking sine transform

pg. 5 Prof. Avadhut Surekha Prakash Manage,

Bhaurao Kakatkar College, Belgaum
Fourier Tranform

2) Fourier transform to solve differential equation 𝐹(𝜔) = (𝜆2 −𝜔2 + 2𝑖𝑘𝜔)𝑋(𝜔)

Ex. To illustrate the application of Fourier transform in solving 𝐹(𝜔)
Or 𝑋(𝜔) = (𝜆2 −𝜔2+2𝑖𝑘𝜔) … … … (1)
differential equations, we consider the motion of a damped
1 ∞
harmonic oscillator. Where 𝑋(𝜔) = ∫ 𝑒 −𝑖𝜔𝑡 𝑥(𝑡)𝑑𝑡
√2𝜋 −∞
The differential equation of motion of damped harmonic
Taking Fourier inverse transform, The solution for eqn (1) is
oscillator is
given by
𝑑2 𝑥 𝑑𝑥
+ 2𝑘 𝑑𝑡 + 𝜆2 𝑥 = 𝑓(𝑡) … … (1) 1 ∞
𝑑𝑡 2 𝑥(𝑡) = ∫ 𝑒 𝑖𝜔𝑡 𝑋(𝜔)𝑑𝜔
1 ∞ √2𝜋 −∞
𝑖𝜔𝑡
𝑓(𝑡) = ∫ 𝑒 𝐹(𝜔)𝑑𝜔 ∞
√2𝜋 −∞ 1 𝐹(𝜔)𝑒 𝑖𝜔𝑡
𝑥(𝑡) = ∫ 𝑑𝜔
& √2𝜋 −∞ (𝜆2 −𝜔 2 + 2𝑖𝑘𝜔)
∞
1
𝐹(𝜔) = ∫ 𝑒 −𝑖𝜔𝑡 𝑓(𝑡)𝑑𝑡
√2𝜋 −∞ 3) Solution for boundary value problems
∞
1 𝑑2𝑥 𝑑𝑥 The Fourier transform may be applied to solve certain
𝐹(𝜔) = ∫ 𝑒 −𝑖𝜔𝑡 ( + 2𝑘 + 𝜆2 𝑥(𝑡)) 𝑑𝑡
√2𝜋 −∞ 𝑑𝑡 2 𝑑𝑡 boundary problems like one dimensional heat flow, one
∞
dimensional heat equation, etc.
𝑑 2 𝑥 −𝑖𝜔𝑡
1 2𝑘 ∞ 𝑑𝑥 −𝑖𝜔𝑡
𝐹(𝜔) = ∫ 𝑒 𝑑𝑡 + ∫ 𝑒 𝑑𝑡 𝛿𝑢(𝑥,𝑡) 𝛿 2 (𝑥,𝑡)
√2𝜋 −∞ 𝑑𝑡 2 √2𝜋 −∞ 𝑑𝑡 Ex. Solve 𝛿𝑡
= 𝛿𝑥 2
,> 0 , 𝑡 > 0 ; subject to
𝜆2 ∞ 1; 0 < 𝑥 < 1
+ ∫ 𝑥(𝑡)𝑒 −𝑖𝜔𝑡 𝑑𝑡 conditions 𝑖) 𝑢(0, 𝑡) 𝑖𝑖) 𝑢(𝑥, 0) = {
√2𝜋 −∞ 0; 𝑥 ≥ 1

𝐹(𝜔) = 𝑋2 (𝜔) + 2𝑘𝑋1(𝜔) + 𝜆2 𝑋(𝜔) 𝑖𝑖𝑖) 𝑢(𝑥, 𝑡) 𝑖𝑠 𝑏𝑜𝑢𝑛𝑑𝑒𝑑

We Know, 𝑋𝑛 (𝜔) = (𝑖𝜔)𝑛 𝑋(𝜔)

∴ 𝐹(𝜔) = (𝑖𝜔)2 𝑋(𝜔) + 2𝑘(𝑖𝜔)1 𝑋(𝜔) + 𝜆2 𝑋(𝜔) Solution : - The given differential equation is

𝐹(𝜔) = −𝜔2 𝑋(𝜔) + 2𝑖𝑘𝜔𝑋(𝜔) + 𝜆2 𝑋(𝜔) 𝛿𝑢 𝛿 2𝑥

=
𝛿𝑡 𝛿𝑥 2
pg. 6 Prof. Avadhut Surekha Prakash Manage,
Bhaurao Kakatkar College, Belgaum
Fourier Tranform

∞
𝛿𝑢𝑠 2
= √ 𝜔 [{𝑢(0, 𝑡)}∞
0 + 𝜔 ∫ 𝑢 sin 𝜔𝑥 𝑑𝑥 ]
Taking Fourier sine transform on both side 𝛿𝑡 𝜋 0

2 ∞ 𝛿𝑢 2 ∞ 𝛿 2𝑥 Using condition (i)

√ ∫ sin 𝜔𝑥 𝑑𝑥 = √ ∫ sin 𝜔𝑥 𝑑𝑥
𝜋 0 𝛿𝑡 𝜋 0 𝛿𝑥 2 𝛿𝑢𝑠
= −𝜔2 𝑢𝑠 … … … . . (2)
𝛿𝑡
𝛿 2 ∞ After rearranging
√ ∫ 𝑢 sin 𝜔𝑥 𝑑𝑥
𝛿𝑡 𝜋 0
𝛿𝑢𝑠
∞ ∞ = −𝜔2 𝛿𝑡
2 𝛿𝑢 𝛿𝑢 𝑢𝑠
= √ [{ sin 𝜔𝑥} − 𝜔 ∫ cos 𝜔𝑥 𝑑𝑥]
𝜋 𝛿𝑥 0 0 𝛿𝑥 Integrating, we get
log 𝑢𝑠 = −𝜔2 𝑡 + log 𝐴 ( A is constant)
𝛿𝑢
As 𝑥 → ∞; 𝛿𝑥 → 0 and 𝑥 → 0 ; sin 𝜔𝑥 → 0 Taking antilog on both side
∞ 2𝑡
𝛿𝑢
Hence{𝛿𝑥 sin 𝜔𝑥} = 0 & 𝑢𝑠 = 𝐴𝑒 −𝜔
0
At t = 0,
2 ∞
√ ∫ 𝑢 sin 𝜔𝑥 𝑑𝑥 = 𝑢𝑠 … … (1) 𝑢𝑠 (𝜔, 0) = 𝐴 … … . . (3)
𝜋 0
Using eqn (1), we have
∞
𝛿𝑢𝑠 2 𝛿𝑢
∴ = −√ 𝜔 ∫ cos 𝜔𝑥 𝑑𝑥
𝛿𝑡 𝜋 0 𝛿𝑥
2 ∞
𝑢𝑠 (𝜔, 0) = 𝐴 = √ ∫ 𝑢(𝑥, 0) sin 𝜔𝑥 𝑑𝑥
𝜋 0
∞
𝛿𝑢𝑠 2 ∞
= √ 𝜔 [{𝑢 cos 𝜔𝑥}0 + 𝜔 ∫ 𝑢 sin 𝜔𝑥 𝑑𝑥 ] 1 ∞
𝛿𝑡 𝜋 0
2
=√ [∫ 1 sin 𝜔𝑥 𝑑𝑥 + ∫ 0 sin 𝜔𝑥 𝑑𝑥 ]
𝜋 0 1
For 𝑢 → 0 𝑎𝑠 𝑥 → ∞

pg. 7 Prof. Avadhut Surekha Prakash Manage,

Bhaurao Kakatkar College, Belgaum
Fourier Tranform

2 cos 𝜔𝑥 1
𝐴=√ [ ] +0
𝜋 𝜔 0

2 1 − cos 𝜔
𝐴 = −√
𝜋 𝜔

Applying Inverse Fourier sine transform, we get

2 ∞
𝑢(𝑥, 𝑡) = √ ∫ 𝑢𝑠 sin 𝜔𝑥 𝑑𝜔
𝜋 0

2 ∞ 1−cos 𝜔
= √𝜋 ∫0 sin 𝜔𝑥 𝑑𝜔
𝜔

This is the required solution.

pg. 8 Prof. Avadhut Surekha Prakash Manage,

Bhaurao Kakatkar College, Belgaum