Engineering Math 2: Laplace Transform - Session 3 - Tutorial Solutions

Convolution Integrals

1. Solve the following equations using the convolution theorem:

𝑎. ) 𝑦(𝑡) + 2𝑒 ∫ 𝑦(𝑢) 𝑒 −𝑢 𝑑𝑢 = 𝑒 𝑡
𝑡

𝑏. ) 𝑦(𝑡) + 2 ∫ 𝑦(𝑢) 𝑒 𝑡−𝑢 𝑑𝑢 = 𝑡𝑒 𝑡

0

Solution:
a.)
𝑡

𝑦(𝑡) + 2 ∫ 𝑦(𝑢) 𝑒 𝑡−𝑢 𝑑𝑢 = 𝑒 𝑡

0

Using convolution theorem:

ℒ {∫ 𝑓(𝑢)𝑔(𝑡 − 𝑢)𝑑𝑢} ⇒ ℒ{𝑓(𝑡) ∗ 𝑔(𝑡)} ⇒ 𝐹(𝑠)𝐺(𝑠)

0

Hence:
𝑡

ℒ {∫ 𝑦(𝑢) 𝑒 𝑡−𝑢 𝑑𝑢} = ℒ{𝑦(𝑡) ∗ 𝑔(𝑡)}

0

𝑦(𝑡) + 2 𝑦(𝑡) ∗ 𝑒 𝑡 = 𝑒 𝑡
1 1
𝑦̅(𝑠) + 2𝑦̅(𝑠) =
𝑠−1 𝑠−1
2 1
𝑦̅(𝑠) (1 + )=
𝑠−1 𝑠−1
𝑠+1 1
𝑦̅(𝑠) ( )=
𝑠−1 𝑠−1
1
𝑦̅(𝑠) =
(𝑠 + 1)
𝑦(𝑡) = 𝑒 −𝑡
b.)
 𝑡

𝑦(𝑡) + 2 ∫ 𝑦(𝑢) 𝑒 𝑡−𝑢 𝑑𝑢 = 𝑡𝑒 𝑡

0

Solution:
𝑡

𝑦(𝑡) + 2 ∫ 𝑦(𝑢) 𝑒 𝑡−𝑢 𝑑𝑢 = 𝑡𝑒 𝑡

0

Using convolution theorem:

𝑦(𝑡) + 2 𝑦(𝑡) ∗ 𝑒 𝑡 = 𝑡𝑒 𝑡
1 1
𝑦̅(𝑠) + 2𝑦̅(𝑠) =
𝑠−1 (𝑠 − 1)2
2 1
𝑦̅(𝑠) (1 + )=
𝑠−1 𝑠−1
𝑠+1 1
𝑦̅(𝑠) ( )=
𝑠−1 (𝑠 − 1)2
1
𝑦̅(𝑠) =
(𝑠 + 1)(𝑠 − 1)
1 𝐴 𝐵
𝑦̅(𝑠) = = +
(𝑠 + 1)(𝑠 − 1) (𝑠 + 1) (𝑠 − 1)
Using Heaviside cover up technique:
1 1
𝐴=− and 𝐵 =
2 2
Hence:
1 1 1
𝑦̅(𝑠) = ( − )
2 (𝑠 − 1) (𝑠 + 1)
The inverse transform becomes and hence a final solution is:
1
𝑦(𝑡) = (𝑒 𝑡 − 𝑒 −𝑡 ) = sinh(𝑡)
2

You may also decide to expand the denominator of the right hand side:
1 1
𝑦̅(𝑠) = = 2
(𝑠 + 1)(𝑠 − 1) 𝑠 − 1
1
𝑦(𝑡) = sinh(𝑡) = (𝑒 𝑡 − 𝑒 −𝑡 )
2
2. For the integral equation
𝑡

𝑓(𝑡) = 1 − 4 ∫ 𝑓 (𝑡 − 𝑢) 𝑒 2𝑢 𝑑𝑢
0
 Using the convolution theorem, show that:
(𝑠 − 2)
𝐹(𝑠) = and 𝑓(𝑡) = 2𝑒 −2𝑡 − 1
𝑠(𝑠 + 2)

Solution:

𝑓(𝑡) = 1 − 4 ∫ 𝑓 (𝑡 − 𝑢) 𝑒 2𝑢 𝑑𝑢
0

𝑡
It is obvious that ∫0 𝑓 (𝑡 − 𝑢) 𝑒 2𝑢 𝑑𝑢 is a convolution integral. Hence, we rewrite the

equation thus:

𝑓(𝑡) = 1 − 4𝑓(𝑡) ∗ 𝑒 2𝑡

Note that both are functions of 𝑡.

Take Laplace transform of the rewritten function noting that:

ℒ{𝑓(𝑡) ∗ 𝑔(𝑡)} ⇒ 𝐹(𝑠)𝐺(𝑠)

1
ℒ{𝑓(𝑡)} = 𝐹(𝑠); ℒ{𝑒 2𝑡 } =
𝑠−2

1 1
𝐹(𝑠) = − 4𝐹(𝑠)
𝑠 𝑠−2

Group similar terms together by moving the second term on the right hand side to the
left:

1 1
𝐹(𝑠) + 4𝐹(𝑠) ( )=
𝑠−2 𝑠

4 1
𝐹(𝑠) (1 + )=
𝑠−2 𝑠

(𝑠 + 2) 1
𝐹(𝑠) =
(𝑠 − 2) 𝑠

(𝑠 − 2)
𝐹(𝑠) = 𝑄𝐸𝐷!
𝑠(𝑠 + 2)
 (𝑠 − 2) 𝐴 𝐵 𝐴(𝑠 + 2) + 𝐵𝑠
𝐹(𝑠) = = + =
𝑠(𝑠 + 2) 𝑠 𝑠 + 2 𝑠(𝑠 + 2)

𝑠: 𝐴 + 𝐵 = 1 ⟹ 𝐴 = 1 − 𝐵;

𝑠 0 : 2𝐴 = −2 ⟹ 𝐴 = −1

𝐵 =1−𝐴=2

Hence:

2 1
𝐹(𝑠) = −
𝑠+2 𝑠

Using the Laplace table and carrying out an inverse transform, we have:

𝑓(𝑡) = 2𝑒 −2𝑡 − 1 𝑄𝐸𝐷!

3. Using the convolution theorem, show that for:

1
𝑃(𝑠) =
𝑠(𝑠 − 4)2

𝑓 ∗ 𝑔 = ∫ 𝜏𝑒 4𝜏 𝑑𝜏
0

1 1
Hint let 𝐹(𝑠) = and 𝐺(𝑠) =
𝑠 (𝑠 − 4)2

Solution:

1
𝑃(𝑠) =
𝑠(𝑠 − 4)2

ℒ −1 {𝐹(𝑠)𝐺(𝑠)} = ℒ −1 {𝐹(𝑠)} ∙ ℒ −1 {𝐺(𝑠)} ⟹ 𝑓 ∗ 𝑔

1 1
ℒ{𝑓(𝑡)} = 𝑎𝑛𝑑 ℒ{𝑔(𝑡)} =
𝑠 (𝑠 − 4)2

1 1
𝑓(𝑡) = ℒ −1 { } = 1; 𝑔(𝑡) = ℒ −1 { } = 𝑡𝑒 4𝑡
𝑠 (𝑠 − 4)2
 𝑡

𝑓 ∗ 𝑔 = ∫ 𝑓(𝑡 − 𝜏)𝑔(𝜏)𝑑𝜏
0

𝑓(𝑡 − 𝜏) = 𝑓(𝑡) = 1

𝑓 ∗ 𝑔 = ∫ 𝜏𝑒 4𝜏 𝑑𝜏
0

𝑡
−1
1
𝑝(𝑡) = ℒ { } = 𝑓 ∗ 𝑔 = ∫ 𝜏𝑒 4𝜏 𝑑𝜏
𝑠(𝑠 − 4)2
0

1 1 1
𝑝(𝑡) = 𝑡𝑒 4𝑡 − 𝑒 4𝑡 +
4 16 16

𝑒 4𝑡 −4𝑡
𝑝(𝑡) = (𝑒 + 4𝑡 − 1)
16

Integro-Differential Equations

4. Solve the integro differential equation:

𝑡
𝑑𝑥
− 4𝑥 + 5 ∫ 𝑥𝑑𝑡 = 1; 𝑥(0) = 0
𝑑𝑡
0

Solution:

Recall that:

𝑡
𝑥̅
ℒ {∫ 𝑥𝑑𝑡} =
𝑠
0

Introducing the subsidiary equation and using the above rule, we find the Laplace
transform of each term in the integro-differential equation:

5𝑥̅ 1
𝑠𝑥̅ − 𝑥(0) − 4𝑥̅ + =
𝑠 𝑠

𝑠 2 𝑥̅ − 4𝑠𝑥̅ + 5𝑥̅ = 1

1
𝑥̅ =
𝑠2 − 4𝑠 + 5
Using completing the square method:

1
𝑥̅ =
(𝑠 − 2)2 + 1

We can now carry out the inverse transform:

𝑥(𝑡) = 𝑒 2𝑡 sin 𝑡

5. Solve the integro differential equation:

𝑡
𝑑𝑥
+ 6𝑥 + 9 ∫ 𝑥𝑑𝑡 = 𝑢(𝑡); 𝑥(0) = 0
𝑑𝑡
0

Note- use the first integration theorem to evaluate the Laplace transform of the integral:

9𝑥̅ 1
𝑠𝑥̅ − 𝑥(0) + 6𝑥̅ + =
𝑠 𝑠

𝑠 2 𝑥̅ + 6𝑠𝑥̅ + 9𝑥̅ = 1

𝑥̅ (𝑠 2 + 6𝑠 + 9) = 1

1 1
𝑥̅ = =
(𝑠 2 + 6𝑠 + 9) (𝑠 + 3)2

𝑥(𝑡) = 𝑡𝑒 −3𝑡

Transfer Functions

6. What is the transfer function of 𝐺(𝑠) the block diagram shown below:

𝑠2 − 1
(𝑠 − 1)2

R C
𝑠2 + 1
(𝑠 − 1)2
Transfer function is given by:

𝐶(𝑠)
𝐺(𝑠) = = 𝐺1 + 𝐺2
𝑅(𝑠)

𝑠2 − 1 (𝑠 2 + 1) 𝑠 2 − 1 + 𝑠 2 + 1
𝐺(𝑠) = + =
(𝑠 − 1)2 (𝑠 − 1)2 (𝑠 − 1)2

2𝑠 2
=
(𝑠 − 1)2

7. A second order differential equation that defines an engineering system is given by:

𝑑2 𝑦 𝑑𝑦
2
−4 + 4𝑦 = 𝑖(𝑡); 𝑌 ′ (0) = 𝑌(0) = 0
𝑑𝑡 𝑑𝑡

If this system has a discontinuous input defined by the Dirac delta function 𝑖(𝑡) =
4𝛿(𝑡 − 2), find the system’s output 𝑦(𝑡)?

𝑑2 𝑦 𝑑𝑦
2
−4 + 4𝑦 = 𝑖(𝑡); 𝑌 ′ (0) = 𝑌(0) = 0
𝑑𝑡 𝑑𝑡
𝑠 2 𝑌(𝑠) − 𝑠𝑌(0) − 𝑌 ′ (0) − 4(𝑠𝑌(𝑠) − 𝑌(0)) + 4𝑌(𝑠) = 𝐼(𝑠)
𝑌 ′ (0) = 𝑌(0) = 0
𝑠 2 𝑌(𝑠) − 4𝑠𝑌(𝑠) + 4𝑌(𝑠) = 𝐼(𝑠)
𝑌(𝑠)(𝑠 2 − 4𝑠 + 4) = 𝐼(𝑠)
𝑌(𝑠) 1
𝐺(𝑠) = = 2
𝐼(𝑠) (𝑠 − 4𝑠 + 4)
𝑖(𝑡) = 4𝛿(𝑡 − 2)
𝐼(𝑠) = 4𝑒 −2𝑠
4𝑒 −2𝑠
𝑌(𝑠) = 𝐺(𝑠)𝐼(𝑠) =
(𝑠 2 − 4𝑠 + 4)
1
𝑌(𝑠) = 4𝑒 −2𝑠 ( )
(𝑠 − 2)2
Using the second shifting theorem:
ℒ −1 {𝑒 −𝑎𝑠 𝑦̅(𝑠)} = 𝐻(𝑡 − 𝑎)𝑦(𝑡 − 𝑎)
1
𝑦(𝑡) = 4ℒ −1 {𝑒 −2𝑠 ( )}
(𝑠 − 2)2
𝑦(𝑡) = 𝐻(𝑡 − 2) ∙ 4(𝑡 − 2)𝑒 2(𝑡−2)
8. The transfer function of a microelectromechanical system is given as:

1
𝐺(𝑠) =
𝑠2 − 3𝑠 − 1

What is the equation of motion describing the behaviour of the system in the time domain
for a sinusoidal forcing function of amplitude 3 and period = 𝜋 ?

Hint: 𝑌(𝑠) = 𝐼(𝑠)𝐺(𝑠) and we have been given the system’s input in the question.

2𝜋
Also, 𝜔 = 2𝜋𝑓 = 𝑇
= 2, so we have all the information for ℒ −1 {𝐼(𝑠)}.

Solution:

1
𝐺(𝑠) =
𝑠2 − 3𝑠 − 1

𝐼(𝑠)
𝑌(𝑠) = 𝐼(𝑠)𝐺(𝑠) =
𝑠2 − 3𝑠 − 1

𝑌(𝑠)(𝑠 2 − 3𝑠 − 1) = 𝐼(𝑠)

𝑠 2 𝑌(𝑠) − 3𝑠𝑌(𝑠) − 𝑌(𝑠) = 𝐼(𝑠)

Using Laplace transform of derivatives with initial values 𝑌(0) = 𝑌 ′ (0) =0, then we can
evaluate the inverse Laplace transform of the system.

𝑑2 𝑦 𝑑𝑦
𝑠 2 𝑌(𝑠) = , 3𝑠𝑌(𝑠) = 3 , 𝑌(𝑠) = 𝑦(𝑡)
𝑑𝑡 2 𝑑𝑡

From the question, we know that

𝑖(𝑡) = 𝐴 sin 𝜔𝑡 = 3 sin(2𝑡)

2𝜋
𝐴 = Amplitude, 𝑇 = 𝜋, 𝜔 = 2𝜋𝑓 = =2
𝑇

Therefore, the original equation of motion is given by:

𝑑2 𝑦 𝑑𝑦
2
−3 − 𝑦(𝑡) = 3 sin(2𝑡)
𝑑𝑡 𝑑𝑡

9. An RC network is modelled by the equation:

 𝑑𝑉
𝑅𝐶 + 𝑉 = 𝑒(𝑡)
𝑑𝑡

where 𝑣(𝑡) is the system response and the time varying input is 𝑒(𝑡). Assume the initial
condition 𝑉(0) = 0.

a.) Determine the transfer function 𝐺(𝑠) for the system.

b.) What is the time response 𝑉(𝑡) of the system for an impulse input 𝑒 = 𝛿(𝑡)?

Solution:

a.) The transfer function 𝐺(𝑠)

𝑑𝑉
𝑅𝐶 + 𝑉 = 𝑒(𝑡)
𝑑𝑡

𝑑𝑉 𝑉 1
+ = 𝑒(𝑡)
𝑑𝑡 𝑅𝐶 𝑅𝐶

1 1
𝑠𝑉̅ − 𝑉(0) + 𝑉̅ = 𝐸̅
𝑅𝐶 𝑅𝐶

1 1
𝑉̅ (𝑠 + )= 𝐸̅
𝑅𝐶 𝑅𝐶

Output 𝑉̅
𝐺(𝑠) = =
Input 𝐸̅
1 1
𝐺(𝑠) = ( )
𝑅𝐶 𝑠 + 1⁄𝑅𝐶

b.) Time response 𝑉(𝑡) of the system for an impulse input 𝑒 = 𝛿(𝑡)

𝑉̅ = 𝑉(𝑠) = 𝐺(𝑠)𝐸(𝑠)

𝐸(𝑠) = ℒ{𝛿(𝑡)} = 1

1 1
𝑉(𝑠) = ( )
𝑅𝐶 𝑠 + 1⁄𝑅𝐶

1 −𝑡/𝑅𝐶
𝑣(𝑡) = 𝑒
𝑅𝐶