Control Engineering

Chapter 2: Modeling in Frequency Domain

Professor Dr. Zobair Ibn Awal

Bangladesh University of Engineering and Technology (BUET)

Dhaka 1000, Bangladesh

May 18, 2025

1 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Chapter 2: Modeling in Frequency Domain
Learning outcomes
1 Find the Laplace transform of time functions and the inverse
Laplace transform (Sections 2.1–2.2)
2 Find the transfer function from a differential equation and
solve the differential equation using the transfer function
(Section 2.3)

2 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Concepts
Fundamental aspects mathematical modeling

3 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Mathematical Modeling

A control system designer uses physical laws, along with

simplifying assumptions, to model the system mathematically.
For example, some of these laws are:
Kirchhoff’s voltage law: The sum of voltages around a closed
path equals zero.
Kirchhoff’s current law: The sum of electric currents flowing
from a node equals zero.
Newton’s laws: The sum of forces on a body equals zero; the
sum of moments on a body equals zero.
Kirchhoff’s and Newton’s laws lead to mathematical models that
describe the relationship between the input and output of dynamic
systems.

4 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Mathematical Modeling
One such model is the linear, time-invariant differential equation:
d m c(t) d m−1 c(t)
am + am−1 + · · · + a0 c(t)
dt m dt m−1
d n r (t) d n−1 r (t)
= bn + bn−1 + · · · + b0 r (t)
dt n dt n−1
Where,
c(t) = Response of the system (function of time)
r (t) = Input to the system (function of time)
m = Order of the response part
n = Odrer of the input part
a and b = Constant coefficients of the differential equation

Many systems can be approximately described by this equation,

which relates the output, c(t), to the input, r (t), by way of the
system parameters, a and b.
5 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
Mathematical Modeling

Example: Maneuvering of a ship in restricted water requires

precise control over the ship. In simplest form it can be considered
that the control input to the ship is rudder angle (δr ) and control
output is the heading (ψ) of the ship. For a given speed the yaw
motion (or heading) of the ship depends on the rudder command
(input), ship’s inertia (I , for yaw direction), angular acceleration
(ψ̈) and angular velocity (ψ̇). The rudder moment generated by a
rudder command (δr ) is proportional to a constant (K ) and equals
summation of the angular moment of inertia and damping moment
of the ship. The angular moment of inertia is proportional to
angular acceleration and damping moment is proportional to
angular velocity of the ship. Construct a mathematical model of
linear differential equation with constant coefficients for heading
response of the ship due to rudder action.

6 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Mathematical Modeling

Solution:
The rudder moment = K × δr
= K δr (t)
The angular moment = I × ψ̈(t)
d2 h i
= I 2 ψ(t)
dt
The damping moment = Damping coefficient (D) × ψ̇(t)
dh i
=D ψ(t)
dt

7 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Mathematical Modeling

Now, balancing the moments of input and response yields the

following linear differential equation with constant coefficients

d2 h i dh i
I ψ(t) + D ψ(t) = K δr (t)
dt 2 dt
The above relationship between input and output can be described
in the following block diagram form as well:

d2 h i dh i
δr (t) −→ I ψ(t) + D ψ(t) = K δr (t) −→ ψ(t)
dt 2 dt

8 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Mathematical Modeling
Hence, we can see a mathematical model of a dynamic system is
defined as a set of equation(s) that represents the dynamics of the
system accurately, or at least fairly well. In studying control
systems one must be able to model dynamic systems in
mathematical terms and analyze their dynamic characteristics.
Note that a mathematical model is not unique to a given system.
A system may be represented in many different ways and,
therefore, may have many mathematical models, depending on
one’s perspective. For example:

d2 h i dh i
δr (t) −→ I ψ(t) + D ψ(t) = K δr (t) −→ ψ(t)
dt 2 dt

dh i h i
δr (t) −→ I r (t) + D r (t) = K δr (t) −→ r (t)
dt

9 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Mathematical Modeling

The dynamics of many systems, whether they are mechanical,

electrical, thermal, economic, biological, and so on, may be
described in terms of differential equations. Such differential
equations may be obtained by using physical laws governing a
particular system — for example, Newton’s laws for mechanical
systems and Kirchhoff’s laws for electrical systems. We must
always keep in mind that deriving reasonable mathematical models
is the most important part of the entire analysis of control systems.

Now, how to solve the problem and get useful results?

10 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Classical method

Example: Solve the differential equation

dx
+ 7x = 5 cos 2t
dt
using classical method. Assume zero initial conditions. Comment
on the limitations of such procedure from control engineering
perspective.

11 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Classical method
Concept
Solution: Let us assume a
particular solution:
Solution =Particular Solution
x = C sin 2t + D cos 2t
+Homogenous Solution

Substitute it in the differential equation yields:

dx
+ 7x = 5 cos 2t
dt
d h i h i
or, C sin 2t + D cos 2t + 7 C sin 2t + D cos 2t = 5 cos 2t
dt
h i h i
or, 2C cos 2t − 2D sin 2t + 7C sin 2t + 7D cos 2t = 5 cos 2t
h i h i
or, 7C − 2D sin 2t + 2C + 7D cos 2t = 5 cos 2t

12 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Classical method
Equating the coefficients gives,
7C − 2D = 0
2C + 7D = 5

From 1st equation, Hence,

2 2
C= D C= D
7 7
2 35
From 2nd equation, or, C = ×
7 53
10
2 ∴C =
2 × D + 7D = 5 53
7
or, 53D = 35 Therefore, particular solution is
35 10 35
∴D= x= sin 2t + cos 2t
53 53 53
13 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
Classical method
Now, for the homogeneous solution, let x = Ae Mt . Putting this in
the equation results in
d h Mt i
e + 7e Mt = 0
dt
⇒Me Mt + 7e Mt = 0

The characteristic polynomial is

M +7=0
∴ M = −7

Hence, the total solution is

!
10 35
x = Ae −7t + sin 2t + cos 2t
53 53

14 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Classical method

Now setting the initial conditions to zero at t = 0,

x =0
dx
=0
dt
35
∴A=−
53
Hence, the solution is
!
35 10 35
x = − e −7t + sin 2t + cos 2t
53 53 53

15 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Classical method

It can be seen from the above procedure that solving a linear

differential equation with constant coefficient is lengthy process.
Also, from control engineering perspective the following issues raise
concerns:
1 When there is a need for addition/removal/modification of
subsystems it becomes mathematical nightmare and obtaining
a solution becomes very difficult.
2 The process does not provide insights into the system
performances and stability.
3 Therefore, such procedure is not suitable for control
engineering design and analysis.

16 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
What is the fundamental problem in control engineering associated
with input-output of a system model?

17 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Differential equation can describe the relationship between the
input and output of a system. The form of the differential
equation and its coefficients are a formulation or description of the
system. Although the differential equation relates the system to its
input and output, it is not a satisfying representation from a
system perspective. Looking at the following equation (Eq. 1.2), a
general, nth-order, linear, time invariant differential equation, we
see that the system parameters, which are the coefficients, as well
as the output, c(t), and the input, r (t), appear throughout the
equation.

d m c(t) d m−1 c(t)

am + a m−1 + · · · + a0 c(t) =
dt m dt m−1
d n r (t) d n−1 r (t)
bn + b n−1 + · · · + b0 r (t)
dt n dt n−1

18 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 We would prefer a mathematical representation such as that shown
in Figure 2.1(a), where the input, output, and system are distinct
and separate parts. Also, we would like to represent conveniently
the interconnection of several subsystems. For example, we would
like to represent cascaded interconnections, as shown in Figure 2.1
(b), where a mathematical function, called a transfer function, is
inside each block, and block functions can easily be combined to
yield Figure 2.1 (a) for ease of analysis and design. This
convenience cannot be obtained with the differential equation.

19 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Therefore, it is essential to have a the system representation
through a viable definition for a function that algebraically relates
a system’s output to its input. This function will allow separation
of the input, system, and output into three separate and distinct
parts, unlike the differential equation. The function will also allow
us to algebraically combine mathematical representations of
subsystems to yield a total system representation.

20 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
Define Laplace Transformation and Inverse Laplace Transformation.

21 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Laplace Transform
Let f (t) be a real function of a The real part σ of a
real variable t defined for t > 0. complex variable s is often
Then Laplace Transform of f (t) written as Re(s) (the real
is as follows: part of s) and the
imaginary part ω as Im(s)
L [f (t)] ≡ F (s) (the imaginary part of s).
Z T The real variable t always
F (s) ≡ lim f (t)e −st dt denotes time.
T →∞,ϵ→0 ϵ
Z ∞ It is noticeable that the
−st lower limit on the integral
= f (t)e dt
0− is t = 0− .
The complex variable s is This definition of the lower
defined by s ≡ σ + jω, where σ limit is sometimes useful in
and √ω are real variable and dealing with functions that
j = −1. are discontinuous at t = 0.

22 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Exercise: Find the Laplace transform of e −t .
Solution: The Laplace transform of e −t is determined as follows:
Z ∞
L e −t = e −t e −st dt
 
−
Z0 ∞
= e −t−st dt
−
Z0 ∞
= e −(1+s)t dt
0−
1 −(1+s)t ∞
 
= − e
1+s 0−
 
1 −∞ 1 0
= − e + e
1+s 1+s
1
= [Re(s) > −1] (Answer)
1+s

23 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Inverse Laplace Transform
The Laplace transform transforms a problem from the real
variable time domain into the complex variable s-domain.
After a solution of the transformed problem has been obtained
in terms of s, it is necessary to invert this transform to obtain
the time domain solution.
The transformation from the s-domain into the t-domain is
called the inverse Laplace transform.

Definition
Let F (s) be the Laplace transform of a function f (t), t > 0. The
following contour integral is called the inverse Laplace transform of
F (s):
Z c+j∞
−1 1
L [F (s)] ≡ f (t) = F (s)e st ds
2πj c−j∞
√
Where, j = −1 and c > σ0 .

24 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Laplace Transformations

Remember the test signals in Chapter 1?

25 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Laplace Transform Theorems

26 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Concepts
Exercise on Laplace transformations.

27 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Impulse Function
Z ∞
L [f (t)] = F (s) = f (t)e −st dt
−
Z0 ∞
L [δt(t)] = ∆(s) = δt(t)e −st dt
0−
Z 0+ Z ∞
−st
= δt(t)e dt + δt(t)e −st dt
0− 0+
Z 0+ Z ∞
= δt(t)e −st dt + 0 × e −st dt
0− 0+
Z 0+
−st
=e |t=0 δt(t)dt = 1 × 1 = 1
0−
Notes: (1) The impulse function is everywhere zero except within
t = 0− and t = 0+ it is infinitely large. (2) The area under the
impulse function is one. (3) e −st is continuous at t=0, that is the
same as saying it is constant from t = 0− to t = 0+ . So we can
replace e −st by its value evaluated at t=0.
28 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
Step Function
Z ∞
L [f (t)] = F (s) = f (t)e −st dt
0−
Z ∞
L [u(t)] = U(s) = u(t)e −st dt
0−
Z ∞
= 1 × e −st dt
0−
1 −st ∞
 
= − e
s 0−
 
1 −∞ 1 0
= − e + e
s s
 
1 1
= − ×0+ ×1
s s
1
= (Answer)
s

29 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Find Laplace transform for each of the following:

1 f (t) = 7 5 f (t) = sin(5t)

2 f (t) = 4t 6 f (t) = cos(−3t)
3 f (t) = t 2 7 f (t) = 4 sin(3t) + 2 cos(9t)
4 f (t) = 3t 2 + 5t + 2 8 f (t) = 3e −4t

30 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Answers:

7 5
1 F (s) = 5 F (s) =
s s2
+ 25
4 s
2 F (s) = 2 6 F (s) = 2
s s +9
2 12 2s
3 F (s) = 3 7 F (s) = 2 + 2
s s + 9 s + 81
6 5 2 1
4 F (s) = 3 + 2 + 8 F (s) = 2
s s s s +4

31 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Concept
Partial fraction review.

32 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Partial Fraction Review: Intro
To find the inverse Laplace transform of a complicated
function, we can convert the function to a sum of simpler
terms for which we know the Laplace transform of each term.
The result is called a partial-fraction expansion.
If F1 (s) = N(s)/D(s), where the order of N(s) is less than
the order of D(s), then a partial-fraction expansion can be
made.
If the order of N(s) is greater than or equal to the order of
D(s), then N(s) must be divided by D(s) successively until
the result has a remainder whose numerator is of order less
than its denominator.
For example, if

s 3 + 2s 2 + 6s + 7
F1 (s) =
s2 + s + 5

33 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Partial Fraction Review: Intro
We must perform the indicated division until we obtain a
remainder whose numerator is of order less than its
denominator. Hence,

s 3 + 2s 2 + 6s + 7 (s 2 + s + 5) + s 3 + s 2 + 5s + 2
F1 (s) = =
s2 + s + 5 s2 + s + 5
s(s 2 + s + 5) + 2 2
=1 + =1+s + 2
s2 + s + 5 s +s +5

Taking the inverse Laplace transform, using Item 1 of Table

2.1, along with the differentiation theorem (Item 7) and the
linearity theorem (Item 3 of Table 2.2),we obtain
 
dδ(t) −1 2
f1 (t) = δ(t) + +L
dt s2 + s + 5

34 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Partial Fraction Review: Intro

Using partial-fraction expansion, we will be able to expand

functions like
2
F (s) = 2
s +s +5
into a sum of terms and then find the inverse Laplace
transform for each term.
We will now consider three cases and show for each case how
an F (s) can be expanded into partial fractions.

35 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Partial Fraction Review: Intro

Study the following cases 1

1 Case 1: Roots of the Denominator of F (s) Are Real and

2
Distinct. Example F (s) = .
(s + 1)(s + 2)
2 Case 2: Roots of the Denominator of F (s) Are Real and
2
Repeated. Example F (s) =
(s + 1)(s + 2)2
3 Case 3: Roots of the Denominator of F (s) Are Complex or
3
Imaginary. Example F (s) = 2
s(s + 2s + 5)

1
Nise (2011) page 37 to 44.
36 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
Case 1: Roots of D(s) are Real and Distinct
Given an F (s) whose denominator has real and distinct roots, a
partial-fraction expansion can be made if the order of N(s) is less
than the order of D(s) as follows:

N(s)
F (s) =
D(s)
N(s)
=
(s + p1 )(s + p2 )...(s + pm )...(s + pn )
K1 K2 Km Kn
= + + ... + + ... +
(s + p1 ) (s + p2 ) (s + pm ) (s + pn )

Example
2 K1 K2
F (s) = = +
(s + 1)(s + 2) (s + 1) (s + 2)

37 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 1: Roots of D(s) are Real and Distinct

To evaluate each residue, Ki , we multiply the expansion by the

denominator of the corresponding partial fraction. Thus, if we
want to find Km , we multiply it by (s + pm ) and get the following:

N(s)
(s + pm )F (s) =(s + pm )
D(s)
N(s)
=(s + pm )
(s + p1 )(s + p2 )...(s + pm )...(s + pn )
K1 K2
=(s + pm ) + (s + pm ) + ...
(s + p1 ) (s + p2 )
Kn
+ Km + ... + (s + pm )
(s + pn )

38 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 1: Roots of D(s) are Real and Distinct

If we let s approach −pm , all terms on the right-hand side of the

previous equation go to zero except the term Km , leaving

N(s)
(s 
+pm ) = Km


(s + p1 )(s + p2 )...
(s 
+pm )...(s + pn )
 

s→−pm

Example

2
(s 
 +1)
 = K1 = 2
(s + 1)(s + 2) s→−1
  

39 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
2
Given the differential equation, ddt y2 + 12 dy
dt + 32y = 32u(t), solve
for y (t) if all initial conditions are zero. Use the Laplace transform.

40 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Taking Laplace transforms on both sides on the following equation
results in:
d 2y dy
2
+ 12 + 32y = 32u(t)
dt  dt
d 2y

dy
⇒L + 12 + 32y = L [32u(t)]
dt 2 dt
⇒[s 2 Y (s) − sy (0) − ẏ (0)] + 12[sY (s) − y (0)] + 32Y (s) = 32U(s)
⇒s 2 Y (s) + 12sY (s) + 32Y (s) = 32U(s)
⇒ s 2 + 12s + 32 Y (s) = 32U(s)
 

32
⇒Y (s) = 2 U(s)
s + 12s + 32
1
For a unit step input U(s) = . Hence,
s
32 32
Y (s) = =
s(s 2 + 12s + 32) s(s + 8)(s + 4)

41 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Applying partial fraction expansion
32 K1 K2 K3
Y (s) = = + +
s(s + 8)(s + 4) s (s + 4) (s + 8)

32
K1 = =1
(s + 4)(s + 8)
s→0

32
K2 = = −2
s(s + 8)
s→−4

32
K3 = =1
s(s + 4)
s→−8
32 1 2 1
Y (s) = = − +
s(s + 8)(s + 4) s (s + 4) (s + 8)
Applying inverse Laplace transform results in:

y (t) = 1 − 2e −4t + e −8t

42 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Exercise 2.4 (Dorf and Bishop):

A laser printer uses a laser beam to print copy rapidly from a

computer. The laser is positioned by a control input r (t), so that
we have
4(s + 50)
Y (s) = 2 R(s)
s + 30s + 200
The input r (t) represents the desired position of the laser beam.
1 If r (t) is a unit step input, find the output y (t).
2 What is the final value of y (t)?

43 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Solution:

44 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 2: Roots of D(s) are Real and Repeated
Given an F (s) whose denominator has real and repeated roots, a
partial-fraction expansion can be made, if the order of N(s) is less
than the order of D(s) and the repeated roots are of multiplicity r
at −p1 , as follows:

N(s)
F (s) =
D(s)
N(s)
=
(s + p1 )r (s
+ p2 )...(s + pn )
K1 K2 Kr Kr +1 Kn
= r
+ r −1
+ ... + + ... +
(s + p1 ) (s + p1 ) (s + p1 ) (s + p2 ) (s + pn )

Example
2 K1 K2 K3
F (s) = 2
= 2
+ +
(s + 1)(s + 2) (s + 2) (s + 2) (s + 1)

45 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 2: Roots of D(s) are Real and Repeated
To find K1 through Kr for the roots of multiplicity greater than
unity, multiply the above expansion by (s + p1 )r :

(s + p1 )r N(s)
(s + p1 )r F (s) =
(s + p1 )r (s + p2 )...(s + pn )
=K1 + (s + p1 )K2 + (s + p1 )2 K3 + ... + (s + p1 )r −1 Kr
Kr +1 (s + p1 )r Kn (s + p1 )r
+ + ... +
(s + p2 ) (s + pn )

Immediately, we can solve for K1 if we let s approach −p1 .

Example
2 K1 K2 K3
= + +
(s + 1)(s + 2)2 (s + 2)2 (s + 2) (s + 1)
2
(s 
+ 2)2 K1(s 
+2)2 K2 (s + 2)2 K3 (s + 2)2
 
⇒ =  + + +
(s + 1)
(s 
+ 2)2 (s 
 + 2)2 (s + 2) (s + 1)

46 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 2: Roots of D(s) are Real and Repeated
We can solve for K2 if we differentiate the expansion with respect
to s and then let s approach −p1 . Subsequent differentiation will
allow us to find K3 through Kr . The general expression for K1
through Kr for the multiple roots is
1 d i−1 F (s)
Ki = i = 1, 2, ..., r ; 0! = 1
(i − 1)! ds i−1 s→−pi

Example
Previously, letting s approach -2 gave K1 = −2. To find K2 we see
that if we differentiate the expansion with respect to s we get

−2 (s + 2)s
= K2 + K3
(s + 1)2 (s + 1)2

K2 is isolated and can be found if we let s approach −2.

47 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 3: Roots of D(s) are Complex or Imaginary

Given an F (s) whose denominator has complex or purely imaginary

roots, a partial-fraction expansion, can be made if the order of
N(s) is less than the order of D(s), here p1 is real and
(s 2 + as + b) has complex or purely imaginary roots, as follows:

N(s) N(s)
F (s) = =
D(s) (s + p1 )(s 2 + as + b)...
K1 (K2 s + K3 )
= + + ...
(s + p1 ) (s 2 + as + b)

The details are given in the text book. However, a simpler method
can be followed similar to Case 1 which is shown in the next slide.

48 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 3: Roots of D(s) are Complex or Imaginary

This technique used for the partial-fraction expansion of F (s) with

real roots in the denominator can be used for complex and
imaginary roots. However, the residues of the complex and
imaginary roots are themselves complex conjugates. Then, after
taking the inverse Laplace transform, the resulting terms can be
identified as
e jθ + e −jθ
= cos θ
2
and
e jθ − e −jθ
= sin θ
2j

49 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Case 3: Roots of D(s) are Complex or Imaginary

50 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Use Laplace Transformations to Solve the Following ODEs:

1 ÿ + 2ẏ + y = 3, Initial conditions ẏ (0) = y (0) = 0.

2 4ÿ − 2y = 7, Initial conditions ẏ (0) = y (0) = 0.
3 3ÿ + 2ẏ + 5y = 3, Initial conditions ẏ (0) = 1, y (0) = 0.
4 ÿ + 3ẏ = −2, Initial conditions ẏ (0) = 1, y (0) = 2.

51 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Answers:

1 y (t) = −3e −t − 3te −t + 3

√ √
7 2 sinh(t 2/2)
2 y (t) =
√ 4 √
3 14e −t/3 sin(t 14/3)
3 y (t) =
7
5 e −3t
4 y (t) = +
6 6

52 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
Explain the fundamentals aspects of Transfer Function.

53 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Transfer Function

The transfer function of a system is defined as the ratio of Laplace

transform of output to the Laplace transform of input where all the
initial conditions are zero.

54 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Some Properties of Transfer Function

1 It is a mathematical model that relates output to its input. If

a transfer function is known then output to a given input can
be predicted.
2 A transfer function is a property of a system not the input.
3 A transfer function includes necessary units but does not
provide insight into the physical/internal structure since it is
only an input/output model.
4 It is possible to determine transfer function experimentally
even if it is unknown analytically.

55 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Some Properties of Transfer Function

Disadvantage
It requires systems to be at zero initial conditions.
Advantages
Indication on system stability
Convenience on system design
Convenience on understanding system dynamics due to
parameter changes
Convenience adding/removing system components

56 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
Find the transfer function represented by the dc(t)
dt + 2c(t) = r (t).
What would be the response of the system for a unit step input?

57 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Solution: Applying Laplace transformation,
dc(t)
+ 2c(t) = r (t)
dt 
dc(t)
⇒L + 2c(t) = L [r (t)]
dt
⇒[sC (s) − c(0)] + 2C (s) = R(s)
⇒sC (s) + 2C (s) = R(s)
⇒(s + 2)C (s) = R(s)
C (s) 1
⇒ = = G (s)
R(s) s +2
For a unit step input the response would be
C (s) = G (s)R(s)
1 1
⇒ C (s) = ×
s +2 s
1 K1 K2
⇒ C (s) = = +
s(s + 2) s s +2
58 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
 Determining the coefficients

1 1 1
K1 = = =
s +2 0+2 2
s→0
1 1 1
K2 = = =−
s −2 2
s→−2

The response is therefore,

1/2 −1/2
+C (s) =
s  s +2 
−1 −1 1/2 −1/2
⇒ L [C (s)] = L +
s s +2
⇒ c(t) = (1/2) × 1 + (−1/2)e −2t
1 1
⇒ c(t) = − e −2t
2 2
1
∴ c(t) = (1 − e −2t )
2
59 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
A Visual Summary

60 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Exercise
Some selected problems.

61 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Skill-Assessment Exercise 2.3: Find the transfer function
C (s)
G (s) = , corresponding to the differential equation
R(s)
d 3c d 2c dc dt 2 dr
3
+ 3 2
+ 7 + 5c = 2
+ 4 + 3r .
dt dt dt dt dt
Skill-Assessment Exercise 2.4: Find the differential equation
2s + 1
corresponding to the transfer function G (s) = 2 .
s + 6s + 2
Skill-Assessment Exercise 2.5: Find the ramp response for
s
a system whose transfer function is G (s) = .
(s + 4)(s + 8)
Example 2.16: Find the transfer function X (s)/F (s) of the
system as shown in the following figure (Figure 2.15).

62 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Skill-Assessment Exercise 2.8: Find the transfer function,
G (s) = X2 (s)/F (s) for the translational mechanical system
shown in the figure below (Figure 2.21).

63 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
Construct a transfer function of a ship maneuvering model.

64 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 The linear ship maneuvering motion equation, alternately second
order Nomoto model, is given as follows (any one of the following
may be considered):

d 2 ψ dψ
T + = K δr
dt 2 dt
dr
T + r = K δr
dt

ψ = Heading
dψ
r = Rate of turn =
dt
K = Rudder constant or Gain
T = Time constant
δr = Rudder command angle

65 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Taking Laplace transform on both sides
h i
L T ψ̈ + ψ̇ = L [K δr ]
⇒T L [ψ̈] + L [ψ̇] = K L [δr ]
⇒T [s 2 Ψ(s) − sψ(0) − ψ̇(0)] + [sΨ(s) − ψ(0)] = K ∆r (s)
⇒Ts 2 Ψ(s) + sΨ(s) = K ∆r (s)
⇒(Ts 2 + s)Ψ(s) = K ∆r (s)
Ψ(s) K
⇒ = 2
∆r (s) (Ts + s)
Ψ(s) K /T
⇒ = 2
∆r (s) (Ts + s)/T

Ψ(s) K /T
∴ G (s) = =
∆r (s) s(s + 1/T )

This is the required transfer function.

66 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Concept
Transfer Function of a Hydraulic System

67 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Hydraulic System
Let us determine the transfer function of a typical hydraulic system.
A hydraulic system works due to liquid pressure difference.

A dashpot (or damper) is shown

in Fig 2.12. Whenever a step
displacement, x is applied to the
piston, a corresponding
displacement y becomes
momentarily equal to x. The
applied force will make oil to flow
through the restriction R and the
cylinder will return to its original
position.

68 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Hydraulic System

Assuming inertia force to be a

negligible, the forces balancing
are represented as

A(P1 − P2 ) = ky ... (1)

Where, A is the piston area, k is

the spring constant and P1 , P2
are the oil pressure existing on
the right and left side of the
piston.

69 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Hydraulic System
The flow rate q through the
restriction with resistance R is
given by
P1 − P2
q= ... (2)
R
The oil is assumed to be
in-compressible (so oil density
ρ = constant). As the mass of oil
flow through the restriction in
time dt must balance with the
change in mass of the left side of
the piston, we have

Aρ(dx − dy ) = qdt

70 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Hydraulic System

Aρ(dx − dy ) = qdt
 
dx dy
⇒Aρ − =q
dt dt
dx dy q
⇒ − =
dt dt Aρ
dx dy P1 − P2
⇒ − = [From (2)]
dt dt RAρ
dx dy ky
⇒ − = [From (1)]
dt dt RA2 ρ
ky dy dx
⇒ 2 + =
RA ρ dt dt
   
ky dy dx
⇒L 2
+ =L
RA ρ dt dt

71 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Hydraulic System
   
ky dy dx
⇒L 2
+ =L
RA ρ dt dt
k
⇒ 2 Y (s) + sY (s) = sX (s)
RA ρ
 
k
⇒ + s Y (s) = sX (s)
RA2 ρ
Y (s) s
⇒ =
X (s) k
+s
RA2 ρ

s
∴ G (s) =
k
s+
RA2 ρ

This is the required transfer function.

72 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain
 Concept
Transfer function of a pneumatic system.

73 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Pneumatic System

A pneumatic system works due to the pressure difference of air or

any other gas. Let us consider a simple pneumatic system with an
actuating valve as shown in the following figure:

Air at a pressure, pi is injected through

the input manifold. The plunger
arrangement has a mass M, B and K
being the coefficients of viscous friction
and spring constant respectively.

74 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Pneumatic System
If A is the area of diaphragm, then the
force exerted on the system will be Api .
Balancing the forces yields the
following:

d 2y dy
M 2
+B + Ky = Api
dt dt
d 2y
 
dy
⇒L M 2 + B + Ky = L [Api ]
dt dt
⇒Ms 2 Y (s) + BsY (s) + KY (s) = API (s)
Y (s) A
⇒ = 2
= G (s)
PI (s) Ms + Bs + K
PI (s) −→ G (s) −→ Y (s)
A
∴ G (s) =
Ms 2 + Bs + K

75 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Concept
Comparison between Hydraulic and Pneumatic systems.

76 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Comparison Between Hydraulic and Pneumatic Systems
No. Hydraulic System Pneumatic System
1. The fluid generally found in The fluid generally found in
hydraulic systems is oil. pneumatic systems is air.
2. Return pipes are always No return pipes are required in
needed in hydraulic systems. pneumatic systems when air is
used.
3. Accuracy of hydraulic actua- Accuracy of pneumatic actua-
tors may be made satisfactory tors is poor at low velocities.
at all velocities.
4. The normal operating pres- The normal operating pres-
sure of hydraulic systems is sure of pneumatic systems is
very much higher than that of very much lower than that of
pneumatic systems. hydraulic systems.
5. Output powers of hydraulic Output powers of pneumatic
systems are considerably high systems are considerably less
than pneumatic systems. than hydraulic systems.

77 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Concept
Transfer function of thermal system.

78 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Thermal System

The following figure shows a water heating system (a thermal

system) where input water temperature is θi and output water
temperature is θo .
The following assumptions are
made to analyze the thermal
system and determine its transfer
function:
The temperature of the
medium is uniform
The tank is insulated from
the surrounding atmosphere.

79 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Thermal System
Furthermore:

Let the steady state temperature

of in flowing water is θi (t) and
that of out flowing water is θo (t).

The steady state heat input rate

from the heater is Hss .

Let the water flow rate be

constant.

Any increase in heat input rate,

∆h(t) will be balanced by
increase in heat outflow rate,
∆h1 (t) and heat storage rate in
the tank, ∆h2 (t).

80 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Thermal System
Using heat balance equation, we
can write
∆h(t) =∆h1 (t) + ∆h2 (t)
∆h1 (t) =QS∆θo (t)

where Q is water flow rate, S is

the specific heat and ∆θo (t) is
the rise of temperature of water
in the tank. Also,
∆θo (t)
∆h1 (t) =
R
where R is the thermal
1
resistance, that is R = QS .

81 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Thermal System

Heat storage in the tank,

d
∆h2 (t) = MS {∆θo (t)}
dt
d
=C {∆θo (t)}
dt
where M is the mass of water in
the tank; C = MS, is the
thermal
 capacitance
 of water and
d
{∆θo (t)} is the rate of rise
dt
of temperature of water in the
tank.

82 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Thermal System
Combining the above equations yield,
∆h(t) = ∆h1 (t) + ∆h2 (t)
∆θo (t) d
⇒∆h(t) = + C {∆θo (t)}
R  dt 
∆θo (t) d
⇒L [∆h(t)] = L + C {∆θo (t)}
R dt
∆Θo (s)
⇒∆H(s) = + CsΘo (s)
R  
1
⇒∆H(s) = ∆Θo (s) + Cs
R
∆Θo (s) 1 R
⇒ = =
∆H(s) 1 RCs + 1
+ Cs
R
Θo (s) R
∴ G (s) = =
H(s) RCs + 1

83 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Exercise
Some additional problems.

84 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Problem:

Krieg, M., & Mohseni, K. (2008, September).

Developing a transient model for squid inspired thrusters,

and incorporation into underwater robot control design.

In 2008 IEEE/RSJ International Conference on Intelligent

Robots and Systems (pp. 3178-3183). IEEE.

Solution:

85 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Question
Explain the Nomoto Model for ship maneuvering motion.
Construct the transfer function of second order Nomoto Model
using conventional symbols. Deduce the expression of time
response of ship heading for a 15 degree rudder command. If the
time constant and rudder constant are taken from Journee and
Pinkster (2002), draw the time response curve to a suitable time
scale.

86 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

Nomoto Model

The linear ship maneuvering motion equation, alternately second

order Nomoto model, is given as follows:

d 2 ψ dψ
T + = Kr δ r
dt 2 dt

ψ = Heading
Kr = Rudder constant or Gain
T = Time constant
δr = Rudder command angle

87 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Taking Laplace transform on both sides
h i
L T ψ̈ + ψ̇ = L [Kr δr ]
⇒T L [ψ̈] + L [ψ̇] = Kr L [δr ]
⇒T [s 2 Ψ(s) − sψ(0) − ψ̇(0)] + [sΨ(s) − ψ(0)] = Kr ∆r (s)
⇒Ts 2 Ψ(s) + sΨ(s) = Kr ∆r (s)
⇒(Ts 2 + s)Ψ(s) = Kr ∆r (s)
Ψ(s) Kr
⇒ =
∆r (s) (Ts 2 + s)
Ψ(s) Kr /T
⇒ =
∆r (s) (Ts 2 + s)/T

Ψ(s) Kr /T
∴ G (s) = =
∆r (s) s(s + 1/T )

This is the required transfer function.

88 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 For a 15 deg rudder command the response is obtained as follows:

Ψ(s) =G (s)∆r (s)

Kr /T 15
= ×
s(s + 1/T ) s
15Kr /T
= 2
s (s + 1/T )

Let us now apply partial fraction expansion

15Kr /T K1 K2 K3
= 2 + +
s 2 (s+ 1/T ) s s (s + 1/T )

89 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Multiplying both sides by s 2 isolates K1 and putting s → 0 gives
the value of K1

15Kr /T K3 s 2
=K1 + K2 s +
(s + 1/T ) (s + 1/T )
∴ K1 =15Kr [s → 0]

To find K2 , let us differentiate the above expression and put s → 0

K3 s 2
   
d 15Kr /T d
= K1 + K2 s +
ds (s + 1/T ) ds (s + 1/T )
−15Kr /T −K3 s 2 2K3 s
⇒ 2
=0 + K2 + +
(s + 1/T ) (s + 1/T )2 (s + 1/T )
∴ K2 = − 15Kr T [s → 0]

90 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 To find K3 , let us isolate it by multiplying both sides of the original
expansion with (s + 1/T ) and putting [s → −1/T ]

15Kr /T K1 K2 K3
= + +
s 2 (s + 1/T ) s2 s (s + 1/T )
15Kr 
(s 
+ 
1/T )/T K1 (s + 1/T ) K2 (s + 1/T ) K3
(s 
+ 
1/T )
⇒ 2  = 2
+ + 
s (s + 1/T )
  s s (s 
 + 1/T )

15Kr /T
⇒ =0 + 0 + K3 [s → −1/T ]
(1/T )2
∴ K3 =15Kr T

Therefore, we obtain the coefficients of partial fraction expansion

as K1 = 15Kr , K2 = −15Kr T and K3 = 15Kr T

15Kr 15Kr T 15Kr T

Ψ(s) = − +
s2 s (s + 1/T )
∴ ψ(t) =15Kr t − 15Kr T + 15Kr Te −t/T

91 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 The rest do it yourself using your own ship.

92 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain

 Thank You!

93 / 93 Professor Dr. Zobair Ibn Awal Chapter 2: Modeling in Frequency Domain