F. T.

problems:

1. Find the inverse Fourier transform of 𝑓(̅ 𝑝) = 𝑒 −|𝑝|𝑦 where 𝑦 ∈ (−∞, ∞)

𝑒 𝑝𝑦 , 𝑝 ≤ 0
Solution: since 𝑒 −|𝑝|𝑦 = {
𝑒 −𝑝𝑦 , 𝑝 > 0
1 ∞
So 𝐹 −1 {𝑓 (̅ 𝑝)} = ∫ 𝑓 (̅ 𝑝)𝑒 𝑖𝑝𝑥 𝑑𝑝
√2𝜋 −∞

0 ∞
1
= [∫ 𝑓(̅ 𝑝)𝑒 𝑖𝑝𝑥 𝑑𝑝 + ∫ 𝑓(̅ 𝑝)𝑒 𝑖𝑝𝑥 𝑑𝑝]
√2𝜋 −∞ 0

0 ∞
1
= [∫ 𝑒 𝑝𝑦 𝑒 𝑖𝑝𝑥 𝑑𝑝 + ∫ 𝑒 −𝑝𝑦 𝑒 𝑖𝑝𝑥 𝑑𝑝]
√2𝜋 −∞ 0

0 ∞
1
= [∫ 𝑒 (𝑖𝑥+𝑦)𝑝 𝑑𝑝 + ∫ 𝑒 −(−𝑖𝑥+𝑦)𝑝 𝑑𝑝]
√2𝜋 −∞ 0

0 ∞
1 1 (𝑖𝑥+𝑦)𝑝 1 −(−𝑖𝑥+𝑦)𝑝
= {[ 𝑒 ] +[ 𝑒 ] }
√2𝜋 𝑖𝑥 + 𝑦 −∞
−𝑦 + 𝑖𝑥 0

1 1 1 1 1 1 √2𝑦
= [ − ]= [ + ]=
√2𝜋 𝑦 + 𝑖𝑥 −𝑦 + 𝑖𝑥 √2𝜋 𝑦 + 𝑖𝑥 𝑦 − 𝑖𝑥 √𝜋(𝑦 2 + 𝑥 2 )
1
2. Use convolution theorem to find 𝐹 −1 {6+5𝑖𝑠−𝑠2}

Solution: 6 + 5𝑖𝑠 − 𝑠 2 = (3 + 𝑖𝑠)(2 + 𝑖𝑠)

1 1 1
Hence 𝐹 −1 {6+5𝑖𝑠−𝑠2 } = 𝐹 −1 {2+𝑖𝑠 . 3+𝑖𝑠} = 𝐹 −1 {𝑓(̅ 𝑠). 𝑔̅ (𝑠)}

1 1
where 𝑓(̅ 𝑠) = 2+𝑖𝑠 , 𝑔̅ (𝑠) = 3+𝑖𝑠

1 1, 𝑡>0
Since 𝐹{𝑒 −𝑎𝑡 𝐻(𝑡)} = 𝑎+𝑖𝑠, where H(t) ={
0, 𝑡<0

1 1
so 2+𝑖𝑠 = 𝐹{𝑒 −2𝑡 𝐻(𝑡)}& 3+𝑖𝑠 = 𝐹{𝑒 −3𝑡 𝐻(𝑡)}

Therefore f(t) = 𝑒 −2𝑡 𝐻(𝑡) 𝑎𝑛𝑑 𝑔(𝑡) = 𝑒 −3𝑡 𝐻(𝑡)

Using convolution theorem we get

 ∞
1 1 1
𝐹 −1 { . }= ∫ 𝑓(𝑢)𝑔(𝑡 − 𝑢)𝑑𝑢
2 + 𝑖𝑠 3 + 𝑖𝑠 √2𝜋 −∞
∞
1
= ∫ 𝑒 −2𝑢 𝐻(𝑢)𝑒 −3(𝑡−𝑢) 𝐻(𝑡 − 𝑢)𝑑𝑢
√2𝜋 −∞

∞
1
= 𝑒 −3𝑡 ∫ 𝑒 𝑢 𝐻(𝑢)𝐻(𝑡 − 𝑢)𝑑𝑢
√2𝜋 −∞

1, 𝑢 > 0, 𝑡 − 𝑢 > 0
Now𝐻(𝑢)𝐻(𝑡 − 𝑢) = {
0, 𝑢 ≤ 0 𝑜𝑟 𝑡 − 𝑢 ≤ 0 𝑜𝑟 𝑏𝑜𝑡ℎ

Hence

1 1
𝐹 −1 { . }
2 + 𝑖𝑠 3 + 𝑖𝑠
0 𝑡 ∞
1 −3𝑡
= 𝑒 ∫ 𝑒 𝐻(𝑢)𝐻(𝑡 − 𝑢)𝑑𝑢 + ∫ 𝑒 𝐻(𝑢)𝐻(𝑡 − 𝑢)𝑑𝑢 + ∫ 𝑒 𝑢 𝐻(𝑢)𝐻(𝑡 − 𝑢)𝑑𝑢
𝑢 𝑢
√2𝜋 −∞ 0 𝑡

𝑡
1 −3𝑡
= 𝑒 ∫ 𝑒 𝑢 𝑑𝑢
√2𝜋 0

1 1 1
= 𝑒 −3𝑡 [𝑒 𝑢 ]𝑡0 = 𝑒 −3𝑡 (𝑒 𝑡 − 1) = (𝑒 −2𝑡 − 𝑒 −3𝑡 )
√2𝜋 √2𝜋 √2𝜋

LAPLACE TRANSFORMATION

Introduction: One of the difficulties with the Fourier transform is that the integral
defining the transform will not converge for some useful engineering signals. An
alternative approach is to redefine the transform. This leads to the concept of
Laplace transform which exists for larger range of signals than the Fourier
transform. Laplace transform has also other advantages for system analysis where
initial conditions can be easily included to give system response. Some other
typical applications are as follows: Transient and steady state analysis of electrical
circuits, analysis of impact and mechanical vibration, analysis of structural
engineering problems such as deflection of beams, columns etc.

Definition: There are two types of Laplace transform. The transform defined by
∞
𝐹(𝑠) = ∫ 𝑓(𝑡)𝑒 −𝑠𝑡 𝑑𝑡
−∞

where s is the complex variable, is known as two sided or bilateral Laplace

transform of the function f(t) provided the integral exist.

For some functions the above transform causes problems of converges. This can be
almost avoided by restricting the range of integration to between 0 to ∞ and
considering f(t)=0 to t<0. Thus the transformation defined by
∞
𝒇̅(𝒔) = ∫𝟎 𝒆−𝒔𝒕 𝒇(𝒕)𝒅𝒕

where t>0 and s is complex variable, is known as one sided or unilateral or simply
Laplace transform of the function f(t) provided the integral exist.

L.T. OF ELEMENTARY FUNCTIONS:

I. 𝑓(𝑡) = 1
−𝑠𝑡 ∞
∞ 𝑒 1
𝑓 (̅ 𝑠) = ∫0 𝑒 −𝑠𝑡 . 1𝑑𝑡 = [ ] = , 𝑠>0
−𝑠 0 𝑠

II. 𝑓(𝑡) = 𝑒 𝑎𝑡
∞ ∞ ∞
𝑒 −(𝑠−𝑎)𝑡 1
̅ −𝑠𝑡 𝑎𝑡
𝑓 (𝑠) = ∫ 𝑒 . 𝑒 𝑑𝑡 = ∫ 𝑒 −(𝑠−𝑎)𝑡
𝑑𝑡 = [ ] = ,𝑠 > 𝑎
0 0 −(𝑠 − 𝑎) 0 𝑠−𝑎

III. 𝑓(𝑡) = 𝑡 𝑛 , 𝑛 > −1

∞
𝑓(̅ 𝑠) = ∫ 𝑒 −𝑠𝑡 . 𝑡 𝑛 𝑑𝑡
0

∞ 𝑘 𝑛 𝑑𝑘 𝑑𝑘
= ∫0 𝑒 −𝑘 ( ) , (put 𝑠𝑡 = 𝑘 𝑠𝑜 𝑑𝑡 = )
𝑠 𝑠 𝑠
 ∞
1
= ∫ 𝑒 −𝑘 𝑘 𝑛 𝑑𝑘
𝑠 𝑛+1 0

Γ(𝑛 + 1)
= , 𝑖𝑓 𝑛 > −1 𝑎𝑛𝑑 𝑠 > 0
𝑠 𝑛+1
IV. 𝑓(𝑡) = cos 𝑎𝑡
∞ 𝑠
𝑓 (̅ 𝑠) = ∫0 𝑒 −𝑠𝑡 . cos 𝑎𝑡 𝑑𝑡 =
𝑠 2 +𝑎2

V. 𝑓(𝑡) = sin 𝑎𝑡
∞
𝑎
𝑓 (̅ 𝑠) = ∫ 𝑒 −𝑠𝑡 . sin 𝑎𝑡 𝑑𝑡 =
0 𝑠 2 + 𝑎2

VI. 𝑓(𝑡) = cosh 𝑎𝑡

∞ ∞
𝑒 𝑎𝑡 + 𝑒 −𝑎𝑡
𝑓 (̅ 𝑠) = ∫ 𝑒 −𝑠𝑡 . cosh 𝑎𝑡 𝑑𝑡 = ∫ 𝑒 −𝑠𝑡 . 𝑑𝑡
0 0 2

1 ∞ −(𝑠−𝑎)𝑡 ∞
= [∫ 𝑒 𝑑𝑡 + ∫ 𝑒 −(𝑠+𝑎)𝑡 𝑑𝑡]
2 0 0
∞
1 𝑒 −(𝑠−𝑎)𝑡 𝑒 −(𝑠+𝑎)𝑡
= [ + ]
2 −(𝑠 − 𝑎) −(𝑠 + 𝑎) 0

1 1 1 𝑠
= [ + ]= 2
2 𝑠−𝑎 𝑠+𝑎 𝑠 − 𝑎2
VII. 𝑓(𝑡) = sinh 𝑎𝑡
∞ ∞
𝑒 𝑎𝑡 − 𝑒 −𝑎𝑡
𝑓 (̅ 𝑠) = ∫ 𝑒 −𝑠𝑡 . sinh 𝑎𝑡 𝑑𝑡 = ∫ 𝑒 −𝑠𝑡 . 𝑑𝑡
0 0 2

1 ∞ −(𝑠−𝑎)𝑡 ∞
= [∫ 𝑒 𝑑𝑡 − ∫ 𝑒 −(𝑠+𝑎)𝑡 𝑑𝑡]
2 0 0
∞
1 𝑒 −(𝑠−𝑎)𝑡 𝑒 −(𝑠+𝑎)𝑡
= [ − ]
2 −(𝑠 − 𝑎) −(𝑠 + 𝑎) 0
 1 1 1 𝑎
= [ − ]= 2
2 𝑠−𝑎 𝑠+𝑎 𝑠 − 𝑎2
PROPERTIES OF L.T.

1. Linearity property:
̅
If 𝐿[𝑓(𝑡)] = 𝑓 (𝑠) and 𝐿[𝑔(𝑡)] = 𝑔̅ (𝑠) then

𝐿{𝑎𝑓(𝑡) ± 𝑏𝑔(𝑡)} = 𝑎𝐿[𝑓(𝑡)] ± 𝑏𝐿[𝑔(𝑡)] = 𝑎𝑓(̅ 𝑠) ± 𝑏𝑔̅ (𝑠)

2. First shifting property:

̅
𝐿[𝑓(𝑡)] = 𝑓 (𝑠) ̅ − 𝑎)
then 𝐿[𝑒 𝑎𝑡 𝑓(𝑡)] = 𝑓(𝑠

Proof:
∞
𝐿[𝑒 𝑓(𝑡)] = ∫ 𝑒 −𝑠𝑡 𝑒 𝑎𝑡 𝑓(𝑡)𝑑𝑡
𝑎𝑡
0
∞
= ∫ 𝑒 −(𝑠−𝑎)𝑡 𝑓(𝑡)𝑑𝑡
0
∞
= ∫0 𝑒 −𝑝𝑡 𝑓(𝑡)𝑑𝑡, (put 𝑠 − 𝑎 = 𝑝)

= 𝑓(̅ 𝑝) = 𝑓 (𝑠
̅ − 𝑎)

𝑓(𝑡) 𝐿[𝑒 𝑎𝑡 𝑓(𝑡)]

𝑡𝑛 𝑛!
(𝑠 − 𝑎)𝑛+1
cos 𝑏𝑡 𝑠−𝑎
(𝑠 − 𝑎)2 + 𝑏 2
sin 𝑏𝑡 𝑏
(𝑠 − 𝑎)2 + 𝑏 2
cosh 𝑏𝑡 𝑠−𝑎
(𝑠 − 𝑎)2 − 𝑏 2
sinh 𝑏𝑡 𝑏
(𝑠 − 𝑎)2 − 𝑏 2

3. Change of scale:
 1 𝑠
̅
If 𝐿[𝑓(𝑡)] = 𝑓 (𝑠) then 𝐿[𝑓(𝑎𝑡)] = 𝑓 ̅ ( )
𝑎 𝑎

Proof:
∞
We know that 𝐿[𝑓(𝑡)] = 𝑓 (̅ 𝑠) = ∫0 𝑒 −𝑠𝑡 𝑓(𝑡)𝑑𝑡
∞
𝐿[𝑓(𝑎𝑡)] = ∫ 𝑒 −𝑠𝑡 𝑓(𝑎𝑡)𝑑𝑡
0
𝑠𝑝
1 ∞ 𝑑𝑝
= ∫0 𝑒 − 𝑎 𝑓(𝑝)𝑑𝑝 put 𝑎𝑡 = 𝑝, 𝑑𝑡 =
𝑎 𝑎

1 𝑠
= 𝑓̅ ( )
𝑎 𝑎
1 1 1 2
Ex: Let 𝐿[sinh 𝑡] = then𝐿[sinh 2𝑡] = =
𝑠 2 −1 2 ( 𝑠 )2 −1 𝑠 2 −4
2

𝑠
𝑠+1 1 3+1 𝑠+3
Ex: Let 𝐿[𝑓(𝑡)] = , then 𝐿[𝑓(3𝑡)] = =
𝑠 2 −2 3 ( 𝑠 )2 −2 𝑠 2 −18
3

4. Change of scale shifting:

1 𝑠−𝑏
If 𝐿[𝑓(𝑡)] = 𝑓 (̅ 𝑠) then 𝐿[𝑒 𝑏𝑡 𝑓(𝑎𝑡)] = 𝑓 ̅ ( )
𝑎 𝑎

For example,
𝑠−2
𝑠 2𝑡 1 3 𝑠−2
𝐿[cos 𝑡] = then 𝐿[𝑒 cos 3𝑡] = 𝑠−2 = (𝑠−2)2
𝑠 2 +1 3 ( )^2+1 +32
3

EXISTENCE CONDITIONS

L.T. does not exist for all functions. If it exists it is uniquely determined. The
following are the conditions to be satisfied:

Let f(t) be the given function. If

1. f(t) is piecewise continuous on every finite interval

2. f(t) satisfy the following inequality:
 |𝑓(𝑡)| ≤ 𝑏. 𝑒 𝑎𝑡 for all 𝑡 ≥ 0 for some constants a and b,
then 𝐿[𝑓(𝑡)] exists.

The function which satisfies the condition (2) is called of exponential order.
For examplecosh 𝑡 < 𝑒 𝑡 ∀ 𝑡 > 0, 𝑡 𝑛 < 𝑛! 𝑒 𝑡 , (𝑛 = 0,1,2, … ) ∀ 𝑡 > 0
2 2
But 𝑒 𝑡 > 𝑏𝑒 𝑎𝑡 whatever may be a and b. so 𝐿{𝑒 𝑡 } does not exist.
1
Similarly does not have L.T.
𝑡

Examples:

1. Find the L.T of

(a) cos 3𝑡 cos 4𝑡
(b) sin √𝑡

Solution:
1
(a) Here cos 3𝑡 cos 4𝑡 = (cos 7𝑡 + cos 𝑡)
2

1
∴ 𝐿[cos 3𝑡 cos 4𝑡] = (𝐿[cos 7𝑡] + 𝐿[cos 𝑡])
2
1 𝑠 𝑠 𝑠(𝑠 2 + 25)
= ( 2 + )= 2
2 𝑠 + 49 𝑠 2 + 1 (𝑠 + 49)(𝑠 2 + 1)
3 5
(√𝑡) (√𝑡)
(b) 𝐿[sin √𝑡] = 𝐿 [√𝑡 − + − ⋯]
3! 5!
1 1 3 1 5
= 𝐿 [𝑡 2 ] − 𝐿 [𝑡 2 ] + 𝐿 [𝑡 2 ] − ⋯
3! 5!
3 5 7
Γ( ) 1 Γ(2) 1 Γ(2) Γ(𝑛+1)
2
= − + − ⋯,[ L(tn) = , 𝑖𝑓 𝑛 > −1 𝑎𝑛𝑑 𝑠 > 0 ]
3
3! 52 5! 72 𝑠 𝑛+1
𝑠2 𝑠 𝑠

1 3 1 5 3 1
2
√𝜋 1 2 . 2 . √𝜋 1 (2 . 2 . 2 . √𝜋)
= 3 − + 7 −⋯
𝑠 2
6 𝑠 52 120 𝑠 2

√𝜋 1 1 1 2 1 1 3
= 3 [1 − + ( ) − ( ) + ⋯ ]
2𝑠 2 4𝑠 2! 4𝑠 3! 4𝑠
 √𝜋 −
1
= 3𝑒
4𝑠

2𝑠 2
Example 2.Find the L.T of

a. cosh 𝑎𝑡 sin 𝑎𝑡
b. 𝑡 sin 𝑎𝑡
c. cosh(5𝑡 + 2)

Solution:
𝑒 𝑎𝑡 +𝑒 −𝑎𝑡
a. cosh 𝑎𝑡 =
2
1
𝐿[cosh 𝑎𝑡 sin 𝑎𝑡] = (𝐿(𝑒 𝑎𝑡 sin 𝑎𝑡) + 𝐿[𝑒 −𝑎𝑡 sin 𝑎𝑡])
2
1 𝑎 𝑎 𝑎(𝑠 2 + 2𝑎2 )
= ( + )= 4
2 (𝑠 − 𝑎)2 + 𝑎2 (𝑠 + 𝑎)2 + 𝑎2 𝑠 + 4𝑎4

1
b. 𝐿{𝑡𝑒 𝑖𝑎𝑡 } = (𝑠−𝑖𝑎)2

(𝑠+𝑖𝑎)2
So 𝐿[𝑡(cos 𝑎𝑡 + 𝑖 sin 𝑎𝑡)] = (𝑠2
+𝑎2 )2

(𝑠 2 − 𝑎2 ) + 2𝑖𝑎𝑠
𝐿[𝑡 cos 𝑎𝑡] + 𝑖 𝐿[𝑡 sin 𝑎𝑡] =
(𝑠 2 + 𝑎2 )2

Equating real and imaginary part

2𝑎𝑠
𝐿[𝑡 sin 𝑎𝑡] =
(𝑠 2 + 𝑎2 )2
𝑒 5𝑡+2 +𝑒 −(5𝑡+2)
c. cosh(5𝑡 + 2) =
2

𝑒2 𝑒 −2
So 𝐿{cosh(5𝑡 + 2)} = 𝐿[𝑒 5𝑡 ] + 𝐿[𝑒 −5𝑡 ]
2 2

𝑒2 1 𝑒 −2 1
= . + .
2 𝑠−5 2 𝑠+5
INVERSE L.T USING METHOD OF PARTIAL FRACTIONS
 ̅
If 𝐿[𝑓(𝑡)] = 𝑓 (𝑠) then 𝐿−1 [𝑓(̅ 𝑠)] = 𝑓(𝑡) is said to be inverse Laplace
transform.

Some inverse L.T

̅
𝑓(𝑠) 𝑓(𝑡) = 𝐿−1 [𝑓(̅ 𝑠)]
1 1
𝑠
1 𝑡 𝑛−1
, 𝑛 = 1,2,3, ..
𝑠𝑛 (𝑛 − 1)!
1 𝑒 𝑎𝑡
𝑠−𝑎
1 1
sin 𝑎𝑡
𝑠 2 + 𝑎2 𝑎
𝑠 cos 𝑎𝑡
𝑠 2 + 𝑎2
1 1
sinh 𝑎𝑡
𝑠 2 − 𝑎2 𝑎
𝑠 cosh 𝑎𝑡
𝑠 2 − 𝑎2
1 1 𝑎𝑡
𝑒 sin 𝑏𝑡
(𝑠 − 𝑎)2 + 𝑏 2 𝑏
𝑠−𝑎 𝑒 𝑎𝑡 cos 𝑏𝑡
(𝑠 − 𝑎)2 + 𝑏 2
1
𝑎𝑡
𝑡 𝑛−1
(𝑠 − 𝑎)𝑛 𝑒 .
(𝑛 − 1)!