EXPLORING FOURIER TRANSFORM

AND ITS APPLICATIONS

Mathematics Project (MTH600)

Submitted By(Group 4) Supervised By

Group ID F2307DAF47 Kiran Naseem Aslam

Fatima Asif Bc190409175 Supervisor
Husna Saleem Bc190404934 M Phil
Mathematics
Maria Rafiq Bc190404932 Virtual University of Pakistan
Rimsha Bc190409351
Tamseela Rahim Bc200412966
Saeed Abdullah Bc190206881

1
 Fourier Transform and its Applications
In this chapter, we will discuss Fourier transform, its relation with the Fourier series, its
properties, Fourier transform of derivative, convolution and its theorem, Fourier sine and
cosine transform, and Fourier transform application.

2.1 Fourier Transform

Given an integrable function f (x) for −∞ < x <∞ . We can associate with it another
function F (k ) of variable k, ¿ by the relation

+∞
1
F (k )= ∫
√ 2 π −∞
ikx
e f (x)dx (2.1.1)

The function F (k ) is called the Fourier transform of f (x), and f (x)is called the inverse
Fourier transform of F (k ). It can be shown that

+∞
1
f (x)=
√ −∞
2 π
∫ −ikx
e F (k )dk (2.1.2)

2.1.1 Fourier series and Fourier transform

The Fourier series representation of a periodic place wise smooth function over the
interval (−l , l) leads to the integral representation of the same function as l →∞ and the
index n in the Fourier series → ∞ . The condition of periodicity is replaced by the
condition of absolute integrability for the function f (x) over (−∞ , ∞ ). This can be seen
as follows.

We start with the complex from of the Fourier series representation for the function f (x),

+∞
f (x)= ∑ cn e inπx/ l, −l< x <l (2.1.3)
n=−∞

Where the complex Fourier coefficients c n are given by

+l
1
c n= ∫ f ( x)e−inπx/ l dx
2 l −l
(2.1.4)

nπ lk
Now we consider the situation in which l →∞ . Let =k then n= and the increment
l π
l∆k
∆ n in n will be given by i.e,
π

l∆k π
∆ n= or ∆ k=
π l

Where ∆ n=1. In the limit l →∞ , ∆ k → 0. In view of this we can rewrite (2.1.3) as

2
 +∞
l
f ( x )= ∑ c n ∆ ne inπx / l =∑ cl ( k )
π
∆ k e ikx (2.1.5)
n=−∞ k

Where we have put c n=c(lk/ π )=cl(k ) to show the dependence of the coefficients c n on l
and k.

Similarly (2.1.4) can be written as

+l
1
c (lk /π )=cl(k )= ∫ f (x)e dx
ikx
2 l −l

+l
l 1
cl(k)= ∫ f (x)e dx
−ikx
(2.1.6)
π 2 π −l

Equations (2.1.5) and (2.1.6) correspond to each other in the same way as equations
(2.1.3) and (2.1.4) do. Now, we let l →∞ so that k now in comes a continuous variable
and assuming that sum goes over into the Riemann integral, we have from (2.1.5)

+∞
f (x)=∫ c(k )e dk
ikx
(2.1.7)
−∞

Where c (k)=lim
l→∞
(l/π )c (lk /π ). Also from (2.1.6)

+l
1
c (k)= ∫ e dx
−ikx
(2.1.8)
2 π −l

To conform the notation followed in this research, we further set

1
c ( k)= F(−k )
√2 π

then (2.1.7) and (2.1.8) become

+∞
1
F (k )= ∫
√ 2 π −∞
ikx
f (x)e dx

and

+∞
1
f (x)=
√ −∞
2 π
∫ −ikx
F (k )e dk

2.1.3 Example

Find the Fourier transform of the box function

f (x)=
{
1 ,|x| ≤ a , a>0
0 ,|x|> a

Solution

3
 +∞
1
F {f (x )}= ∫
√ 2 π −∞
ikx
e f (x )dx

a +a +∞
1
¿ ∫
√2 π −∞
0 ×e dx + ∫ 1× e dx + ∫ 0 × e dx
ikx ikx ikx

−a −a

)|
+a

(
ikx iak −iak
1 e 1 e −e
¿ 0+ =
√2 π ik −a √2 π ik

¿
2 e iak−e−iak
√2 π 2 ik
=
2 sinka
π k √
2.1.4 Example

a
Find the Fourier transform of g(x )= 2 , a> 0.
2
(x +a )

Figure 2.1

Solution

The function g(x ) as well as its derivative are continuous over interval (−∞ ,+∞ ), and
+∞

the integral ∫ g(x )dx is absolutely integrable. Therefore, the Fourier transform of the
−∞

given function must exist.

+∞
1 a
F {g (x)}=G(k )= ∫ ikx
e 2 2 dx
√ 2 π −∞ x + a
+∞ ikx
a e
¿ ∫
√2 π −∞ x +a 2
2
dx

ikx
a e
¿ ∮
√ C x +a 2
2 π 2
dx

Where C is a closed contour consisting of the X-axis and a semicircle (of infinite radius)
in the upper of lower half-plane. Now

ikz=ik(ℜ z+i ℑ z )=ik (x+iy)=ikx−ky

Therefore e ikz=eikx−ky → 0 when y →0 for k > 0.

4
The same will also → 0 for k < 0 if y →−∞ . We want to choose the contour C in such a

way that the integral ∮ g(x )dx is zero. In one case the contour will lie in the upper-half
C

plane whereas in the other case it will lie in the lower-half plane, see figure (2.1).
Therefore

ikx
a e a
G(k)= ∮
√ 2 π C x +a
2 2
dx=
√2 π
×2 π i × S

Where S denotes the sum of residues of the poles in the contour.

(i). where k > 0, we take the contour as a semicircle in the upper-half plane. Therefore,
the residue of the function at z=ia is given by

−kz
ikz 1 e
residue= lim e = , (k > 0)
x →ia z +ia 2ia

Hence

a e−kz √ π −ka
G(k)= ×(2 πi)× = e (k > 0)
√2 π 2ia 2

(ii). When k < 0, we take the semicircle in the lower-half plane.

ka
1 e
Residue of the function at −ia=lim eikz =
x→ ia z−ia −2 ia

Therefore

a e ka −√ π ka
G(k)= ×(2 πi)× = e
√2 π −2 ia 2

Combining the two results, we have

a −|k|a
G(k)= e , for all k
√2 π

2.2 Properties of Fourier Transforms

a). Linearity property

It is a linear transformation; both F and F−1 are linear.

b). Conjugation property

If f (x) is real, then F (−k )=F (k ) , (where the bar symbol denotes the complex
conjugate).

Proof

5
 +∞
1
F (k )= ∫
√ 2 π −∞
ikx
f (x)e dx

and therefore

+∞
1
F (k )=
√ −∞
2 π
∫ −ikx
f (x)e dx

Also

+∞
1
F (−k )= ∫
√ 2 π −∞
−ikx
f (x) e dx

which proves that F (−k )=F (k ) .

c). Real and complex values of the Fourier Transform

a) If f (x) is real and even, F (k ) is real.

b) If f (x) is real and odd, F (k ) is pure imaginary.
c) If f (x) is complex, then F {f (− x)}=F(k ).

d). Attenuation property

ax
F {e f ( x) }=F ( k−ai)

Proof
+∞
1
ax
F e f ( x )= ∫
√2 π −∞
ikx ax
e e f ( x ) dx

+∞
1
¿ ∫
√2 π −∞
e
(ik +a) x
f ( x ) dx

+∞
1
¿ ∫
√2 π −∞
e
i(k−ia)x
f ( x ) dx

¿ F (k −ai)

d). Shifting properties

ika
(i) F {f (x −a)=e F (k )
ikx
(ii) F {e f ( x )}=F ( k +a)

Proof of (i)
+∞
1
F {f (x −a)= ∫
√2 π −∞
ikx
e f ( x−a ) dx

Put x−a=x ' , which implies dx=dx '

6
 +∞
1
∫ f (x )d x
'
ik (x +a) ' '
F {f (x −a)= e
√2 π −∞
+∞
1
∫ e e f (x )d x
'
ika ik x ' '
¿
√2 π −∞
+∞
1
∫ e f (x )d x
'
ika ik x ' '
¿e
√ −∞
2 π

ika
¿ e F (k )

Proof of (ii)
+∞
1
ikx
F {e f ( x )}= ∫
√ 2 π −∞
iax ikx
e e f ( x ) dx

+∞
1
¿
√ −∞
2 π
∫ e
i(k+a )x
f ( x ) dx

¿ F (k + a)

2.3 Fourier transforms of Derivatives

The Fourier transforms of derivative of a function f ( x ) whose Fourier transforms exists

are given by

'
F {f (x )}=(−ik)F ( k ) (2.3.1)

where f ( x ) is supposed to tend to zero as x → ± ∞ .

'' 2
F {f (x)}=(−ik ) F (k ) (2.3.2)

Where f ( x ) , f ' ( x) are supposed to tend to zero as x → ± ∞ . And

n n
F {f (x )}=(−ik) F (k ) (2.3.3)

Where f ( x ) , f ' ( x) ,…. f (n−1) (x )→ 0 as x → ± ∞ .

Proof

For (2.3.1) we have

+∞
1
'
F {f (x )}= ∫
√ 2 π −∞
ikx '
e f ( x) dx

[ ]
+∞
1 +∞
e f (x)|−∞− ∫ f (x)(ik )e dx
ikx ikx
¿
√2 π −∞

[ ]
+∞
1
0+(−ik ) ∫ f (x) e dx
ikx
¿
√2 π −∞

7
 +∞
1
¿(−ik) ∫
√ 2 π −∞
ikx
f (x)e dx=(−ik) F (k )

For (2.3.2)

[∫ ]
+∞
'' 1 ikx ' '
F {f ( x ) }= e f ( x ) dx
√2 π −∞

[ ]
+∞
1 +∞
e f ( x )|−∞−ik ∫ e f ( x ) dx
ikx ' ikx '
¿
√2 π −∞

[ ]
+∞
1
0+(−ik ) ∫ e f ( x ) dx
ikx '
¿
√2 π −∞

+∞
1
¿(−ik)
√ −∞
2 π
∫ ikx '
e f ( x ) dx

'
¿(−ik) F {f ( x ) }=(−ik)(−ik)F ( k )

Or finally

'' 2 2
F {f ( x ) }=(−ik ) F (k )=−k F (k )

For (2.3.3) we have

+∞
1
n
F {f ( x ) }=
√ −∞
2 π
∫ ikx n
e f ( x ) dx

[ ]
+∞
1 +∞
e f ( x )|−∞−(ik) ∫ e f ( x ) dx
ikx n−1 ikx n−1
¿
√2 π −∞

[∫ ]
+∞
(−ik) ikx n−1
¿ e f ( x ) dx
√2 π −∞

[ ]
+∞
(−ik) ikx n−2 +∞
e f ( x )|−∞ −(ik) ∫ e f ( x ) dx
ikx n−2
¿
√2 π −∞

[ ]
+∞
(−ik)2
¿
√2 π
∫ eikx f n−2 ( x ) dx
−∞

Continuing this process, we get

+∞
n
F {f ( x ) }=(−ik)
n
∫ f ' ( x ) e ikx dx=(−ik)n F(k )
−∞

2.3.1 Fourier transform of functions of the form x n f ( x )

Let n be a positive integer and f ( x ) a piecewise continuous function on the integral

+∞
[−l,+l] for every positive l. suppose that ∫ |x n f (x)|dx coverges. Then
−∞

8
 ( )
n
n n d n n
F {f ( x ) }=(−i) F (k)=(i) F (k )
dk

Solution
+∞
1
F ( k )= ∫
√2 π −∞
ikx
e f ( x ) dx

And differentiation successively w.r.t. k we obtain

' '' 2 2
F ( k )=i F { xf ( x)}, F ( k )=i F {x f (x )}

n n n
F ( k )=i F {x f ( x )}

From the last relation we obtain

n 1 n −n n
F { x f (x)}= n F ( k )=i F ( k )
i

2.3.2 Fourier transform of an integral

+∞

We suppose that f (x) is a piecewise continuous on (−∞ ,+∞ ) and that ∫ |f ( x )|dx< ∞.
−∞

Also let F ( k )=F {f (x )}. Then

{∫ }
x
−1
f (x )d x =
' '
F F (k )
−∞ ik

Proof

Let

x
y (x )=∫ f ( x ) d x
' '

−∞

Then by the Leibniz rule, g' ( x)=f ( x), whenever f (x) is continuous. From the defining
relation for g(x ), g(−∞)=0.

Now using the definitions

+∞ +∞
1 1
F ( k )= ∫
√2 π −∞
ikx
e f ( x ) dx and F ( 0 )= ∫ f ( x ) dx
√2 π − ∞
We have

+∞
lim g(x )= ∫ f ( x ) dx=√ 2 π F (0)=0
x→ ∞ −∞

Now using the result

F {g ' (x ) }=−ik F { g(x ) }

Now, we have

9
 F ( k )=F { f (x ) }=F {g ' (x ) }=−ik F { g(x ) }

{∫ }
x
f (x )d x
' '
¿−ik F
−∞

Which gives

{∫ }
x
1
f (x )d x =
' '
F F (k )
−∞ −ik

2.4 Convolution

If f (x) and g(x ) are functions of x defined over the interval (−∞ ,+∞ ) then their
convolution, denoted by f∗g is defined as follows

+∞
1
f∗g= ∫ f ( η) g(x−η)d η
√2 π −∞
We can show that g∗f =f ∗g.

Now consider

+∞
1
f∗g= ∫ f ( η) g(x−η)d η
√2 π −∞
+∞
1
¿− ∫
√2 π −∞
' ' '
f (x−x )g(x )d x , where x−η=x '

+∞
1
¿ ∫
√2 π −∞
' ' '
g(x )f (x −x ) d x =g∗f

In a similar manner we can prove the following properties of the convolution

f∗(g∗h)=(f∗g)∗h

And

f∗(g+h)=f∗g+ f ∗h

2.4.1 Convolution Theorem

If F (k ) and G(k) are Fourier transforms of f (x) and g(x ), then

F {f ∗g }=F(k )G(k)

or

−1
F { F( k )G( k) }=f ( x)∗g ( x)

Proof
+∞
1
−1
F { F(k )G(k)}= ∫
√2 π −∞
−ikx
e F (k )G (k )dk

10
 { }
+∞ +∞
1 1
∫ ∫
'
−ikx ik x ' '
e F (k ) e g (x )d x dk
√2 π −∞ √ 2 π −∞

Where we have used the definition of the Fourier transform of g(x ).

Changing the order of integration, we have

+∞
1 1
−1
F { F(k )G(k)}=
√2 π
∫ ¿¿
√ 2 π −∞
Now by definition

+∞
1
∫
√2 π −∞
e
−ik (x−x ¿¿')F (k)dk
=f (x−x ¿¿ ')¿ ¿

We find that

+∞
1
−1
F { F(k )G(k)}=
√ −∞
2 π
∫ ' ' '
f (x−x )g (x )d x =f∗g

Hence

−1
F { F( k )G( k) }=f ( x)∗g ( x)

Which is equivalent to

F (k )G (k )=F {f ∗g }

2.5 Fourier Integral Theorem

+∞

If f (x) is a real function defined over (−∞ ,+∞ ), and the integral ∫ f ( x )dx is absolutely
−∞

convergent, then

∞ +∞
1
f (x)= ∫ dk ∫ cosk (ξ−x )f (ξ )dξ
π 0 −∞

Proof
+∞

Since the integral ∫ f ( x )dx is absolutely convergent, its Fourier transform

−∞

and the inverse both exist. Therefore

+∞
1
f ( x )=
√ −∞
2 π
∫ −ikx
e F ( k ) dk

Splitting the infinite integral in parts and using the fact that f (x) is real, we
have

[ ]
∞ 0
1
f ( x )=
√2 π
∫e −ikx
F ( k ) dk + ∫ e
−ikx
F ( k ) dk
0 −∞

11
 [ ]
∞ 0
1
∫e F ( k ) dk + ∫ e
'

F (−k ) (−d k ) , k '=−k

−ikx −i k x ' '
¿
√2 π 0 −∞

[∫ ]
∞ 0
1
F ( k ) dk + ∫ e
'

F (−k ) d k
−ikx ik x ' '
¿ e
√2 π 0 −∞

Since f (x) is real, by conjugate property, we must have F (−k )=F (k ). Therefore, we can
write

∞
1
f (x)= ∫ [ e−ikx F ( k )+ eikx F ( k ) ] dk
√2 π 0
Now

∞
1
∫ f (x )d x
'
−ik x ' '
F (k )= e
√ 2 π −∞
Therefore

∞
1
∫ f (x )d x
'
−ikx −ik ( x−x ) ' '
e F (k )= e
√ 2 π −∞
and taking complex conjugate of both sides

∞
1
∫ f (x )d x
'
ikx ik (x−x ) ' '
e F (k )= e
√ −∞
2 π

Adding the last two equations, we have

∞
1
∫ f ( x ) [e ] d x'
' '
−ikx ikx ' ik (x−x ) −ik (x−x )
e F (k )+e F (k )= +e
√ 2 π −∞
∞
1
¿ ∫
√2 π −∞
f ( x ) 2cosk (x−x )d x
' ' '

Substituting in the equation for f (x) , we obtain

∞ ∞
1
f (x)= ∫ ∫ f ( x ' ) cosk (x −x' )d x ' dk
π 0 −∞

2.6 Fourier Sine and Cosine Transforms

If a function f (x) is defined over the interval ¿, then we can define its Fourier sine
transform F s (k ) or F s {f (x)} as

√
∞
2
F s (k )= ∫ f (x)sinkx dx
π 0
(2.5.1)

Similarly, the Fourier cosine transform is defined as

√
∞
2
F c (k )≡ F c {f (x)}= ∫ f (x )coskx dx (2.5.2)
π 0

12
The inverse relations to (2.5.1) and (2.5.2) are given by

√
∞
2
f (x)≡ F { F s (k )}= ∫ F s (k )sinkx dk
−1
s (2.5.3)
π 0

and

√
∞
2
−1
f (x)≡ F c { F c (k)}= ∫ F (k )coskx dk
π 0 c

2.6.1 Justification For the Definition

The above definitions follow directly from the definitions of complex or exponential
Fourier transform. Let a real function f (x) be defined over ¿. its even extension f e ( x)
over the whole real line can be defined as

f e ( x )=
{ f (x), for 0 ≤ x <∞
f (−x) , for−∞< x <0

Now, the Fourier transform of f c (x) is given by

+∞
1
F {f c (x )}≡ F c (k )= ∫
√2 π −∞
ikx
e f c ( x ) dx

[∫ ]
+∞ 0
1
e f ( x ) + ∫ e f (−x ) dx
ikx ikx
¿
√2 π 0 −∞

+∞
1
=
√ 0
2 π
∫ [ eikx +e−ikx ] f (x)dx

Or

√
∞
2
F c (k )= ∫ f (x )coskx dx
π 0

Similarly, if we define he odd extension of f (x) over the whole real line as

f o (x )=
{ f ( x), for 0 ≤ x< ∞
−f (−x ) , for−∞ < x< 0

and perform similar calculations, we at the definition of the Fourier sine transform given
above.

2.6.2 Fourier Sine and Cosine Transforms of Derivatives

To calculate Fourier sine and cosine transforms of first order derivatives, we assume that
(i) f (x) is real and (ii) |f ( x )|→ 0 as x → ∞ . Then

√
∞
2
'
F c {f (x) }= ∫ f ' (x)coskx dx
π 0

13
 √[ ]
∞
2
coskxf ( x )|0 + k ∫ f ( x ) sinkx dx
∞
¿
π 0

√ √ √
∞
2 2 2
¿
π
× 0−
π
f (0)+k ∫ f ( x ) sinkx dx
π 0

Therefore

F c {f ' (x) }=−

√ 2
π
f ( 0)+ k F s (k )

Similarly

√
∞
2
'
F s {f (x )}= ∫
π 0
'
f ( x )sinkx dx

√
∞
2
f ( x ) sinkx|0 −k ∫ f ( x ) coskx dx
∞
¿
π 0

√ √ √
∞
2 2 2
¿−
π
× 0−
π
×0−k ∫ f ( x ) coskx dx
π 0

Therefore

'
F s {f (x )}=−k Fc (k)

To calculate Fourier Sine and Cosine transforms of second order derivatives, we assume
further that (iii) |f ' (x)|→ 0 as x → ∞ . Then

√
∞
2
F s {f (x)}= ∫ f (x)sinkx dx
'' ''
π 0

√[ ]
∞
2 ∞
sinkx f (x)|0 −k ∫ f (x ) coskx dx
' '
¿
π 0

√[ ]
∞
2
0−k ∫ f (x)coskx dx
'
¿
π 0

√
∞
2
¿−k ∫ f ' (x)coskx dx
π 0

'
¿−k F c {f (x)}

¿−k − [√ 2
π
f (0)+ k F s (k ) ]
Therefore

F s {f '' (x)}=
√ 2
π
kf (0)−k 2 F s ( k )

Similarly

14
 √
∞
2
F c {f ( x)}= ∫ f (x )coskx dx
'' ''
π 0

√[ ]
∞
2 ∞
coskx f (x )|0 +k ∫ f (x ) sinkx dx
' '
¿
π 0

√ √
∞
2 ' 2
¿ 0−
π
f (0)+ k ∫
π 0
'
f ( x) sinkx dx

¿−
√ 2 '
π
f (0)+ k F s {f ' (x )}

¿−
√ 2 '
π
f (0)+ k [−k F c (k ) ]

F c {f ' ' ( x)}=−

√ 2 '
π
f (0)−k 2 F c ( k )

2.6.3 Example

Calculate Fourier sine transform of the function f (x)= {sinx0,,0x≤>xπ≤ π ,

Solution

√
∞
2
F s {f (x)}= ∫ f (x) sinkx dx
π 0

√ √
π ∞
2 2
¿ ∫
π 0
f (x )sinkx dx + ∫ f (x )sinkx dx
π π

√ √∫
π ∞
2
¿ ∫ sinx sinkx dx + 2π
π 0
0 sinkx dx
π

√
π
2
¿ ∫ sinx sinkx dx
π 0

√
π
1 2
¿−
2
∫ [ cos (k +1)x −cos( k−1) x ] dx
π 0

π
1
¿ ∫ [−cos(k +1)x +cos (k −1) x ] dx
√2 0

{ }|
π
1 −sin(k +1) x sin(k−1) x
¿ +
√2 k +1 k −1 0

¿
√2 [
1 −sin(k +1)π sin(k−1) π
k +1
+
k−1 ]
1 1
¿ [−(k−1) sin(k + 1) π +(k +1) sin(k−1)π ]
√2 k 2−1

15
 1 1
¿ [(k−1)sinkπ−(k +1)sinkπ ]
√2 k 2−1
1 1
¿ (−2 sinkπ )
√2 k 2−1

¿
√ 2 2 sinkπ
π k 2−1

2.6.4 Example

Calculate Fourier sin transform of the function f (x)=e−x cosx

Solution

By definition

√
∞
2
F s (e cosx)= ∫ e cosx sinkx dx
−x −x
π 0

1
Using the result sinAcosB= [ sin( A+ B)+sin( A−B) ] , we have
2

1
sinkxcosx= [ sin(k +1) x +sin(k−1)x ]
2

Therefore

√
∞
1 2
−x
F s (e cosx )= ∫ e−x { sin(k+ 1) x+ sin(k −1)x } dx
2 π 0

Now, we use the formula

ax
∫ eax sinbxdx= a2e+b2 {asinbx−bcosbx }

and obtain

∞ −x
e
∫ e sin (k +1)xdx =
−x
1+¿ ¿
0

1
¿ 0− 2
(0−(k +1))
k + 2 k +2

k +1
¿ 2
k +2 k +2

Similarly,

∞ −x
e
∫ e sin (k −1) xdx=
−x
1+ ¿ ¿
0

1
¿ 0− 2
(0−(k−1))
k −2 k +2

16
 k−1
¿ 2
k −2 k +2

Therefore

F s (e− x cosx)=
1
[ k +1
+ 2
k −1
√ 2 π k + 2 k +2 k −2 k +2
2 ]
3 3
1 2k 2 k
¿ =
√2 π k + 4 √ π k 4 + 4
4

2.7 Applications of Fourier Transform

2.7.1 Use of Fourier Sine and Cosine Transforms in B.V / I.V Problems

The B.V / I.V problems usually consist of second order PDEs along with I.Cs and B.Cs.
If the x coordinate extends from 0 to ∞ , as happens in heat conduction problems
involving conducting rod of infinite length, or in vibration problems involving strings of
infinite length, then we may use sine or cosine transforms in the solution of such
problems. The choice of sine or cosine transform is dictated by the type of boundary
condition. From the formulas

F s {u' ' ( x)}=

√ 2
π
ku (0)−k 2 U s (k )

and

F c {u ' ' ( x )}=−

√ 2 '
π
k u (0)−k 2 U c ( k)

Where F s and F c denote Fourier Sine and Cosine transform operators.

We can see that if the problems contain B.C of the form u(0 ,t )=c where c is a
constant, then we use sin transform. On the other hand, if the problem contain a B.C of
the form u x (0 , t)=constant , then we use cosine transform.

2.7.2 Example

Solve the potential equation for the potential u(x , y) in the semi-infinite strip

0< x < c , y >0 that satisfies the following conditions

u(0 , y )=0 , u y (x , 0)=0, u x (c , y )=f ( y ) (1)

Solution

The potential equation is given by

2 2
∂ u ∂ u
2
+ 2 =0 , 0< x < c , y >0 (2)
∂x ∂y

The second B.C in (1) suggests that we take Fourier cosine transform w.r.t. y. Therefore

17
 F c {u xx }+ F c {u yy }=0

Or on using formulas for Fourier transforms of derivatives, we have

d2
dx 2
U c (x , k )−
2
π √
u y (x ,0)−k 2 U c ( x , k )=0

Or using the second B.C. in (1) we have

( )
2
d 2
2
−k U c (x ,k )=0
dx

Whose general solution can be written as

kx −kx
U c (x , k )=c 1 ( y )e +c 2 ( y )e (3)

Now, we transform the first and third B.Cs in (1) and obtain

d
U c (0 , k )=0, U (c , k )=F c (k ) (4)
dx

Using the B.Cs in (4), we obtain from (3)

c 1 +c 2=0 , c 1 k e kc −c 2 k e−kc =F c ( k ) (5)

On solving equations in (5), we have

F c (k )
c 1=−c2= =c
k (e kc + e−kc )

Therefore, on substitution in (3)

kx −kx
F c (k )e F c ( k )e
U c (x , k )= −
2k cosh ck 2 k cosh ck

F c (k ) kx −kx
¿ (e −e )
2 k cosh ck

sinh kx
¿ F c (k )
k cosh ck

On taking inverse cosine Fourier transform, we obtain

u(x , k )=F c { U c (x , k ) }
−1

√
∞
2 F c (k )sinh kx cos ky
¿ ∫
π 0 k cosh ck

√ √
∞ ∞
2 sinh kx cos ky 2
¿ ∫
π 0 k cosh ck
. ∫
π 0
' ' '
f ( y )cosk y d y dk

∞ ∞ '
2 sinh kx cos kycosk y
¿ ∫∫
' '
f ( y )d y dk
π 0 0 k cosh ck

18
2.7.3 Use of complex Fourier transform in solving B.V / I.V problems

When we have an I.V / B.V problem in which the space coordinate x extends
over the whole real line, we may use complex Fourier transform. For the
validity of the formula for the derivatives it is required that both the
unknown function u and its partial derivatives approach zero as x goes to ± ∞
.

Example

Solve the problem by means of the Fourier transform method

2
ut =α u xx , −∞ < x <+∞ t >0 (1)

u x {x , 0 }=f ( x),|u(x , 0)|< ∞ ,−∞ < x <+∞ (2)

Solution

We assume that both u and u x → 0 as x → ± ∞ , and further that both u and f (x), are
absolutely integrable over (−∞,+∞). Taking the Fourier transform of both sides of (1),
and denoting the Fourier transform of u(x , t) w.r.t. x by U ( k ,t ), we obtain the first order
ODE

dU (k ,t )
=−α 2 k 2 U (k , t) (3)
dt

General solution of (3) is given by

2 2

U ( k ,t )=c e−α kt
(4)

The I.C in (2) can be transformed as U ( k , 0)=F(k ). From this condition and (4), we
obtain c=F (k ). Hence
2 2

U ( k ,t )=F (k )e−α k t

Taking the inverse Fourier transform, we have

+∞
1
∫
2 2
−ikx −α kt
u(x , t)= e F (k ) e dk
√ 2 π −∞
or

[ ]
+∞ +∞
1 1
∫ ∫
2 2 '
−ikx −α kt ik x ' '
u(x , t)= e e . e f (x )d x dk
√ 2 π −∞ √2 π −∞

or on changing order of integration

[∫ ]
+∞ +∞
1
u(x , t)= ∫
√ 2 π −∞
'
f (x )
'
exp[−ik(x −x )−−α k t ]dk d x
2 2 '

−∞

19
 +∞
1
u(x , t)= ∫
√ 2 π −∞
' '
f (x )I (x , x )d x
'
(5)

Where

+∞
I (x , x )= ∫ exp [−ik (x−x )−α k t ]dk
' ' 2 2

−∞

+∞
¿ ∫ exp[−ik α−β k ]dk
2

−∞

Where β=α 2 t and u=x−x '. By completing the squares, we can express the gaussian
integral as

2 +∞
−u
) ∫ exp ¿ ¿
'
I (x , x )=exp(
4 β −∞

Hence on substituting in (5), we have

( )
+∞ ' 2
1 1 −( x , x )
u(x , t)= ∫ f (x )exp 4 α 4 t2 d x '
'
2π √ πt α −∞
2

Add more examples

2.7.4 Real life Applications

The Fourier transform is a mathematical tool that decomposes a complex signal into a
sum of simple sine and cosine waves. This transformation is widely used in various fields
and has numerous real-life applications. Here are some examples:

1. Signal Processing

Audio Processing

Fourier transform is extensively used in audio signal processing for tasks like
compression (e.g., MP3), equalization, and filtering.

Image Processing

In image analysis, Fourier transforms are employed for tasks such as image
compression, pattern recognition, and filtering.

Communication Systems

Fourier analysis is crucial for signal modulation, demodulation, and encoding in

communication systems.

2. Electrical Engineering

Power Systems Analysis

Fourier analysis helps in understanding and analyzing electrical waveforms in

power systems, aiding in the detection and correction of power quality issues.

20
 Control Systems

Fourier transform is used in the analysis of control systems to study system

behavior in the frequency domain.

3. Physics

Quantum Mechanics

In quantum mechanics, the Fourier transform is utilized to transform wave

functions between position and momentum representations.

Crystallography

Fourier analysis is used to determine the structure of crystals by analyzing X-ray

diffraction patterns.

4. Medical Imaging

MRI (Magnetic Resonance Imaging)

Fourier transform is a fundamental technique in MRI, used to convert raw data

into images by analyzing the frequency components of signals.

CT (Computed Tomography)

Fourier transform is applied in the reconstruction of images from projections in

CT scans.

5. Mechanical Engineering

Vibration Analysis

Fourier analysis is employed to analyze vibrations in mechanical systems, helping

identify and mitigate potential issues.

Structural Analysis

In structural engineering, Fourier transforms are used to analyze the response of

structures to various forces.

6. Economics and Finance

Time Series Analysis

Fourier analysis is applied in the analysis of economic and financial time series
data to identify underlying trends and patterns.

7. Geophysics

Seismology

Fourier transform is used to analyze seismic signals, helping in the study of the
Earth's subsurface structure.

21
8. Chemistry

Spectroscopy

Fourier transform is widely used in various spectroscopic techniques, such as

infrared (FTIR) and nuclear magnetic resonance (FT-NMR) spectroscopy.

9. Computer Science

Image and Audio Compression

Fourier transform is employed in various compression algorithms, such as JPEG

for images and MP3 for audio.

Fourier transform's applications are broad and diverse, spanning across multiple
disciplines. It provides a powerful tool for analyzing and understanding complex signals
and systems in different domains.

2.8 Integral Equation

An equation which involves the unknow variable under integral sign is call integral
equation.

2.8.1 Example
b

∫ x ' u ( x ' )d x ' =1 (2.7.1)

a

∫ K (x , x ¿¿ ' )u( x' )d x ' =f ( x) ¿ (2.7.2)

a

2.8.2 Use of Fourier Transforms in Integral Equations

Example

Solve the integral equation

+∞
y (x )=f (x)+ λ ∫ e
−ixz
y (z)dz (2.7.3)
−∞

Solution

We will use the definition of Fourier transform and its inverse as given above. Denoting
the Fourier transform of f (x) by f (k ), (2.7.3) can be written as

y (x )=f (x)+ λ √2 π y (−x) (2.7.4)

Where y (−x ) is the Fourier transform of y (z ).

Taking the Fourier transfrom of both sides of (2.7.4), we have

+∞
1
y (k)=f ( k )+ λ √ 2 π ∫
√2 π −∞
ikx
e y (−x)dx

22
 +∞
1
¿ f (k )+ λ √ 2 π ∫
√2 π −∞
−ikx
e y (x ) dx (2.7.5)

Using the definition of inverse Fourier transform, (2.7.5) becomes

y (k)=f ( k )+ √ 2 π λy (k ) (2.7.6)

From (2.7.6)

y (−x )=f (−x)+ √ 2 π λy(−x) (2.7.7)

Substitutaing for y (−x ) from (2.7.7) into (2.7.4), we have

y (x )=f (x)+ λ √ 2 π [ f (−x )+ √ 2 π λy (−x) ]

¿ f (x)+ λ √2 π f (−x )+ 2 π λ y (−x )

2
(2.7.8)

From (2.7.8)

y (−x )=f (−x)+ λ √ 2 π f ( x )+ 2 π λ y ( x )

2

Substituating for y (−x ) into (2.7.8), we finally obtain

y (x )=f (x)+ λ √ 2 π f (−x )+ 2 π λ2 × [ f (−x )+ λ √ 2 π f (x )+2 π λ 2 y (x ) ]

From the last equation

3
y ( x ) ( 1−4 π λ )=f ( x ) + λ √ 2 π f (−x )+ 2 π λ f (−x ) + ( 2 π ) λ f ( x )
2 4 2 23

or

[ ]
3
1
y ( x )= 2 4
f ( x ) + λ √ 2 π f (−x ) +2 π λ2 f (−x )+ ( 2 π ) 2 λ3 f ( x )
1−4 π λ

The solution given above is unique as long as 1−4 π 2 λ4 ≠ 0 i.e λ is not equal to one of the
eigenvalues of the associated homogenoue IE.

23
References
[1]. Kammler, D. W. (2007). A first course in Fourier Analysis. Cambridge
University Press.

[2]. Joseph Fourier, J.-B. (1878). The Analytical Theory of Heat. Cambridge
University Press.

[3]. Rayleigh, J. W. (1894). Theory of sound. London, -96

[4]. Wiener, N. (1930). Generalized harmonic analysis. Acta Math. 117–258

[5]. E. Oran Brigham. (1988). The Fast Fourier Transform and Its Applications.
Pearson.

[6]. Daubechies, I. (1992). Ten Lectures on Wavelets. SIAM.

[7]. Proakis, J. G. (1996). Digital Signal Processing Principles Algorithms and

Applications.

[8]. Edward Charles Titchmarsh. (1988). Introduction to the Theory of Fourier

Integrals.

[9]. Lathi, B. P. (2010). Linear systems and signals. Oxford University Press.

[10]. Priestley, M. B. (1982). Spectral Analysis and Time Series, Two-Volume Set.
Academic Press.

[11]. Haykin, S. S. (1983). Communication Systems. John Wiley & Sons.

[12]. Cormen, T. H., Leiserson, C. E., & Rivest, R. L. (1990). Introduction to

Algorithms. Mit Press.

[13]. Press, W. H. (1989). Numerical recipes in Pascal : the art of scientific computing.
Cambridge University Press.

24