See discussions, stats, and author profiles for this publication at: [Link]

net/publication/389533679

Control Systems

Presentation · March 2025

CITATIONS READS

0 15

1 author:

Daniel Abebe Beyene

Addis Ababa Science and Technology University
8 PUBLICATIONS 3 CITATIONS

SEE PROFILE

All content following this page was uploaded by Daniel Abebe Beyene on 04 March 2025.

The user has requested enhancement of the downloaded file.

 Review on Laplace Transform

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Focus of This Course

• Concentrate on continuous-time systems, primarily modeled through differential equations.

• Mathematical representation may take the form of:
• Transfer function: Relates output and input in the Laplace domain.
• State space representation: Focuses on the internal state using first-order differential
equations (Which will be discussed in Modern Control Systems Course).
• Emphasis on analysis and controller design transfer function domain

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Definition of Laplace Transform

The Laplace transform is an operator that transforms a function of time, f (t), into a new
function of complex variable, F (s), where s = σ + jω.
Z ∞
F (s) = L(f (t)) = f (t)e−st dt
0−

• The lower limit 0− accounts for instantaneous energy transfer.

• The Laplace transform is useful for solving linear differential equations and analyzing
control systems.

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Inverse Laplace Transform

To reverse the operation and take the inverse transform back to the time domain:

L−1 (F (s)) = f (t)

• All functions are defined to be zero for t < 0− .
• The unit step function u(t) is defined as:
(
1 if t ≥ 0
u(t) =
0 if t < 0

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Example 1: Laplace Transform of Exponential Function

Find the Laplace transform of f (t) = e−at u(t), where a is a real constant.
Z ∞ Z ∞
L(e−at u(t)) = e−at e−st u(t)dt = e−(a+s)t dt
0− 0−
The integral exists if ℜ(a + s) > 0. Performing the integration:
1
L(e−at u(t)) =
s+a
The region of convergence (ROC) is σ > −a.

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Linearity Property of Laplace Transform

The Laplace transform is linear:

L(f1 (t) ± f2 (t)) = L(f1 (t)) ± L(f2 (t))

For a constant K:

L(Kf (t)) = KL(f (t))

Combining these properties:

L(Kf1 (t) ± f2 (t)) = KL(f1 (t)) ± L(f2 (t))

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Example 2: Laplace Transform of Unit Step Function

Find the Laplace transform of f (t) = Ku(t), where K is a constant.

Using the linearity property and the Laplace transform of the unit step function:

K
L(Ku(t)) = KL(u(t)) =
s

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Example 3: Laplace Transform of Sine Function

Find the Laplace transform of f (t) = sin(ωt).

Using Euler’s formula:

ejωt − e−jωt
sin(ωt) =
2j
Applying the Laplace transform:
ω
L(sin(ωt)) =
s2 + ω2
Similarly, for the cosine function:
s
L(cos(ωt)) =
s2 + ω 2

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Laplace Transform of Derivatives

The Laplace transform of the first derivative of a function is:

 
df (t)
L = sF (s) − f (0− )
dt
For higher-order derivatives:

dn−1 f (0− )
 n 
d f (t)
L n
= sn F (s) − sn−1 f (0− ) − · · · −
dt dtn−1

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Solving Differential Equations Using Laplace Transform

Solve the differential equation:

d2 x(t) dx(t)
2 2
+7 + 5x(t) = 10u(t)
dt dt
with initial conditions:

dx(0− )
= 1, x(0− ) = −2
dt
Taking the Laplace transform and solving for X(s):

−2s2 − 6s + 5
X(s) =
s(s + 2.5)(s + 1)
Applying the inverse Laplace transform:

x(t) = (2 + 2e−2.5t − 6e−t )u(t)

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Laplace Transform of Integrals

The Laplace transform of an integral:

Z t 
F (s)
L f (x)dx =
0− s
For an integral from −∞ to t:
R 0−
t
−∞ f (x)dx
Z 
F (s)
L f (x)dx = +
−∞ s s

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Time Shifting and Scaling Properties

Time shifting:

L(f (t − a)u(t − a)) = e−sa F (s), a>0

Time scaling:
1 s
L(f (at)) = F , a>0
a a

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Final Value Theorem

The Final Value Theorem allows us to determine the steady-state value of a function f (t) as
t → ∞ using its Laplace transform F (s):

lim f (t) = lim sF (s)

t→∞ s→0

• This theorem is valid only if all poles of sF (s) are in the left half of the complex plane
(i.e., the system is stable).

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Initial Value Theorem

The Initial Value Theorem allows us to determine the initial value of a function f (t) as
t → 0+ using its Laplace transform F (s):

lim f (t) = lim sF (s)

t→0+ s→∞

• This theorem is useful for finding the initial conditions of a system directly from its
Laplace transform.

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Basic Notions
Recovering a Function from Its Transform

Given a Laplace transform Y (s), we want to find the original function y(t).
4 4
• Example: If Y (s) = s−3 , and since L[4e3t ] = s−3 , we conclude:

y(t) = 4e3t

• Uniqueness Theorem: If L[f ] = L[g], then f (t) = g(t) for t ≥ 0.

• The Laplace transform only considers t ≥ 0, so we assume t ≥ 0 in all formulas.

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Inverse Laplace Transform Defined

The inverse Laplace transform of F (s), denoted L−1 [F ], is the function f (t) such that:

L[f (t)] = F (s)

h i
• Example: L−1 4
s−3 = 4e3t .
• Any table of Laplace transforms can also be used for inverse transforms.

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Linearity of the Inverse Transform

The inverse Laplace transform is linear:

L−1 [c1 F1 (s) + c2 F2 (s)] = c1 L−1 [F1 (s)] + c2 L−1 [F2 (s)]
h i
• Example: Find L−1 s21+9 .
h i
• We know L−1 s23+9 = sin(3t), so:
 
−1 1 1
L 2
= sin(3t)
s +9 3

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Partial Fractions

Partial fractions are used to simplify inverse Laplace transforms.

3s2 −28
• Example: Given Y (s) = (s−4)(s 2 +4) , we decompose it as:

A Bs + C
Y (s) = + 2
s−4 s +4
• Solving for A, B, and C, we get:

1 2s + 8
Y (s) = + 2
s−4 s +4
• The inverse transform is:

y(t) = e4t + 2 cos(2t) + 4 sin(2t)

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Inverse Transforms of Shifted Functions

The first shifting identity for Laplace transforms is:

L[eat f (t)] = F (s − a)

For inverse transforms:

L−1 [F (s − a)] = eat f (t)
h i
• Example: Find L−1 1
(s−6)3
.
• Let F (s − 6) = 1 1
(s−6)3
, so F (s) = s3
.
• Then:  
−1 1 1 1
L 3
= e6t · t2 = t2 e6t
(s − 6) 2 2

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

Additional Exercises

• Find the inverse Laplace transform of:

1 1 1 6 5 s
, , , , ,
s−6 s+2 s2 s4 s2 + 25 s2 + 3π 2
• Solve initial-value problems using Laplace transforms:

y ′′ + 9y = 0, y(0) = 4, y ′ (0) = 6

• Use partial fractions to find inverse transforms:

7s + 5 s−1 1
, ,
(s + 2)(s − 1) s2 − 7s + 12 s2 − 4

AASTU [Link]

Review on Laplace Transform Daniel A. (MSc. & Science Ambassador)

 Mathematical Modeling and Control System Problem Formulation

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Transfer Function

• A transfer function provides a mathematical expression that characterizes a system.

• It is defined for linear, time-invariant (LTI) systems.
• The system must be represented by differential equations, and initial conditions must be
zero.

Definition
The transfer function of an LTI system is the ratio of the Laplace transform of the output to
the Laplace transform of the input, assuming zero initial conditions.

Y (s) L{y(t)}
T.F. = G(s) = =
R(s) L{r(t)}

R(s) G(s) Y (s)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Linear Time-Invariant (LTI) Systems

• Linear Systems: Follow the principle of superposition (additivity and homogeneity).

• Time-Invariant Systems: The output does not depend on the time of application of the
input.

Principle of Superposition
• Law of Additivity:
f {x1 (t) + x2 (t)} = y1 (t) + y2 (t)
• Law of Homogeneity:
f {Kx(t)} = Ky(t)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Properties of Transfer Function

• Transfer function is independent of the input magnitude and nature.

• It can be used to determine system stability and output for a given input.
• It is only defined for LTI systems.

Advantages
• Simple algebraic representation.
• Easy determination of system gain, poles, and zeros.
• Simplifies stability analysis.

Disadvantages
• Only valid for LTI systems.
• Does not account for initial conditions or internal system states.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Modeling of Dynamic Systems

Mathematical modeling is a technique for representing dynamic systems through mathematical

expressions, simplifying complex systems to facilitate analysis, prediction, and control of their
behavior. These models describe time-dependent processes, aiding in understanding transient
responses, such as changes from one steady state to another. They provide insights into
system dynamics by characterizing time constants and behaviors.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Why we need Mathematical Modeling?

• Developing training simulators for operator education.

• Supporting ”what-if” analyses, allowing exploration of scenarios (e.g., effects of closing a
valve) using tools like MATLAB or LabVIEW.
• Essential for optimization and control tasks, including:
• Control structure design.
• Controller tuning.
• Model-based control strategies.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Types of Mathematical Models

Mathematical models serve as abstractions that encapsulate the behavior of systems and can
manifest in various forms:
• Empirical Models: Constructed based on observed data, often using statistical methods
to fit data. Effective for predicting system behavior within observed conditions.
• Mathematical Models: Utilize mathematical language to describe relationships among
system variables. Further categorized into:
• Differential Equations: Applicable to continuous systems.
• Difference Equations: Relevant for discrete systems.
• Hybrid Models: Incorporate both continuous and discrete elements.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Modeling of Electrical Systems

Mathematical modeling of electrical systems often relies on fundamental laws such as:
• Ohm’s Law: Relates voltage (vR ), current (iR ), and resistance (R) in a linear circuit:

vR = iR R (1)

• Kirchhoff’s Current Law (KCL): The total current entering a junction equals the total
current leaving the junction.
• Kirchhoff’s Voltage Law (KVL): The sum of the electrical potential differences
(voltage) around any closed network is zero.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Basic Laws Used in Modeling Electrical Systems

• Inductor Voltage-Current Relation: The voltage across an inductor (vL ) is

proportional to the rate of change of current (iL ):

diL
vL = L (2)
dt
where L is the inductance.
• Capacitor Voltage-Current Relation: The current through a capacitor (iC ) is
proportional to the rate of change of voltage (vC ):

dvC
iC = C (3)
dt
where C is the capacitance.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
RLC Circuit

Consider the series RLC circuit below:

R L

vs C vo

Figure: Series RLC circuit

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
RLC Circuit Analysis

From Kirchhoff’s Voltage Law (KVL):

d
vs = iR + L i + vc
dt
d
C vc = i
dt
(4)

Taking the Laplace transform of both equations:

Vs (s) = I(s)R + sLI(s) + Vc (s) (5)

I(s) = sCVc (s) (6)

Substituting I(s) from the second equation into the first:

Vs (s) = sRCVc (s) + s2 LCVc (s) + Vc (s) (7)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Transfer Function

If the goal is the voltage across the capacitor

Vs (s) = LCs2 + RCs + 1 Vc (s)

(8)

Thus, the transfer function G(s) is:

Vc (s) 1
G(s) = = 2
(9)
Vs (s) LCs + RCs + 1

If the goal is the current

I(s) = sCG(s)
Cs
= (10)
LCs2 + RCs + 1

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Magnetic Circuits

Magnetic circuits are analogous to electrical circuits and are used to analyze the behavior of
magnetic fields in various materials. The fundamental concepts of magnetic circuit modeling
include:
• Magnetic flux
• Magnetomotive force (MMF)
• Magnetic reluctance

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Magnetic Flux

Magnetic flux (Φ) is defined as the total magnetic field (B) passing through a given area (A).
It can be expressed mathematically as:

Φ=B·A (11)

where:
• B is measured in teslas (T)
• A is measured in square meters (m2 )
The direction of the magnetic flux is determined by the orientation of the magnetic field lines.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Magnetomotive Force (MMF)

The magnetomotive force (Fmmf ) is the driving force that produces magnetic flux in a
magnetic circuit. It is analogous to voltage in an electrical circuit and is calculated as:

Fmmf = N · i (12)

where:
• N is the number of turns in the coil
• I is the current flowing through the coil, measured in amperes (A)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Magnetic Reluctance

Magnetic reluctance (Rm ) is the opposition to the flow of magnetic flux in a magnetic circuit,
similar to resistance in an electrical circuit. It is defined as:
Fmmf
Rm = (13)
Φ
The unit of magnetic reluctance is ampere-turns per weber (A/Wb). The relationship between
MMF, magnetic flux, and reluctance can be expressed as:

Fmmf = Rm · Φ (14)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Hinged Armature Type Relay

The hinged armature type relay operates as an electromagnetic circuit consisting of:

• A coil (electromagnet)
• An armature (movable part)
• A core

Figure: Hinged armature type relay

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Magnetic Field Strength

The strength of the magnetic field is proportional to the current flowing through the coil (I)
and the number of turns in the coil (N ):

N ·i
B =µ·
l
where:
• B is the magnetic flux density
• µ is the permeability of the core material
• l is the length of the magnetic path

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Electromagnetic Force

The force (Fem ) exerted on the armature can be calculated using the magnetic flux and the
area of the armature:
B2 · A
Fem =
2µ0
where:
• A is the area of the armature
• µ0 is the permeability of free space

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Modeling of Mechanical Systems

Mathematical modeling of mechanical systems involves the use of mathematical expressions to

represent the behavior and dynamics of mechanical components and systems. These models
are crucial for understanding, analyzing, and predicting the performance of mechanical systems
across various applications in engineering.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Basic Laws Used in Modeling Mechanical Systems

The modeling of mechanical systems is grounded in fundamental physical laws, which provide
the framework for understanding how forces and motions interact. The key laws include:
• Newton’s Laws of Motion: These laws describe the relationship between the motion of
an object and the forces acting on it.
• Conservation of Energy: This principle states that energy cannot be created or
destroyed, only transformed from one form to another.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mechanical Transformers

Mechanical transformers, such as levers and gears, are devices that change the direction or
magnitude of forces and motion. These components are fundamental in mechanical systems,
allowing for the efficient transfer and modification of energy.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Levers

A lever amplifies an input force to lift a load. The relationship between the input force (F1 ),
output force (F2 ), input distance (l1 ), and output distance (l2 ) can be described by the law of
the lever:

F1 · l1 = F2 · l2 (15)

l1 F1
F2 l2

Figure: The Lever System

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Gears

Gears transmit torque and change the speed of rotation. The relationship between the input
and output angular velocities (ω1 , ω2 ) and torques (τ1 , τ2 ) can be expressed as:
ω1 r2
= and τ1 · ω1 = τ2 · ω2 (16)
ω2 r1

r1

 2  2
1 1

r2

Figure: A Gear system

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Stiffness Elements (Springs)

Springs are mechanical components that store and release energy. The behavior of springs is
defined by Hooke’s Law:

F = −kx (17)
τ = −kθ (18)

where F and τ are the restoring force and restoring torque, respectively, k is the spring
constant, and x and θ are the linear and angular displacements, respectively.

x 
k
k
F

(a) (b)

Figure: Stiffness Elements: A diagram showing the behavior of springs under load.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Friction Elements (Dampers)

Dampers are devices that dissipate energy, usually in the form of heat, and are used to reduce
oscillations in mechanical systems. The force exerted by a damper is proportional to the
velocity (v) of the moving component:

Fd = −dẋ (19)
τd = −dθ̇ (20)

where Fd and τd are the damping force and damping torque, respectively, b is the damping
coefficient, and ẋ and θ̇ are the linear and angular velocities.

x1 x2 1 2
b b

(a) (b)

Figure: Dampers: A schematic representation of damping mechanisms in mechanical systems.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Solenoid Actuator-Valve System

The 3/2 solenoid actuator valve plays a crucial role in regulating fluid flow through three
ports: one inlet and two outlets.
Spool Valve
Air gap sping
Seated Position Coil Supply Drain

+ Fs
Fem m

Fd
-

Plunger

Drain
Figure: Free body diagram of solenoid
actuator valve
Figure: Solenoid actuator-valve system

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Model of Solenoid Actuator Valve

Applying Newton’s second law of motion, the force acting on the spool valve can be expressed
as:

ΣF = mẍ
Fem − Fs − Fd = mẍ
N i − kx − dẋ = mẍ (21)

where m is mass of the spool, x is displacement of the spool from its equilibrium position, N
is number of turns in the solenoid, i is the current flowing through the solenoid, k is spring
constant and d is damping coefficient.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Transfer Function

Taking the Laplace Transform:

ms2 X(s) + dsX(s) + kX(s) = N I(s) (22)

Simplifying:

(ms2 + ds + k)X(s) = N I(s) (23)

Thus, the transfer function G(s) is:

X(s) N
G(s) = = 2
(24)
I(s) ms + ds + k

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Wind Turbine Generator System
Wind turbine generator systems convert kinetic energy from the wind into electrical energy
through a series of mechanical and electrical processes.

1

wind
 aero r1
b

2

 g
fc

r2

Figure: Wind turbine generator system

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Model of Wind Turbine Generator System

The mathematical model of the system can be divided into two main components: the turbine
system and the generator system.

τaero − bθ̇1 − fc r1 = J1 θ̈1 (25)

fc r2 − b2 θ̇2 − τg = J2 θ̈2 (26)

Solving for fc and rearranging,

   
b2 1 J2
τaero = b1 N + θ̇2 + τg + + J1 N θ̈2
N N N
b2 + b1 N 2
     
N 1
θ̈2 = τaero − τg − θ̇2 (27)
J2 + J1 N 2 J2 + J1 N 2 J2 + J1 N 2

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Transfer Function

Taking the Laplace Transform:

b2 + b1 N 2
     
N 1
s2 θ2 (s) + sθ 2 (s) = τaero (s) − τg (s) (28)
J2 + J1 N 2 J2 + J1 N 2 J2 + J1 N 2

Simplifying:

s(J2 + J1 N 2 ) + (b2 + b1 N 2 )
     
N 1
sθ2 (s) = τaero (s) − τg (s)
J2 + J1 N 2 J2 + J1 N 2 J2 + J1 N 2
(29)

Thus, the transfer function G(s) is:

   
N 1
θ2 (s) = τaero (s) − τg (s)
s2 (J2 + J1 N 2 ) + s(b2 + b1 N 2 ) s2 (J2 + J1 N 2 ) + s(b2 + b1 N 2 )
(30)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Modeling of Electro-mechanical Systems

Electro-mechanical systems can be represented using differential equations that describe the
dynamic behavior of both electrical and mechanical components. A common example of an
electro-mechanical system is a DC motor.
For armature voltage-controlled DC motor with:
• Armature resistance Ra and inductance La .
• Mechanical load with load torque τL , rotational inertia J, and gear ratio rm
rL .
• Mechanical friction coefficient b.
The control input is armature voltage Va , and the output is the load speed ωL .

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
DC Motor Equivalent Circuit

Ra ia La
+ τm , ωm rm
Va M

- τL , ωL
Load

rL
Figure: Armature voltage controlled DC motor with mechanical load

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Model of DC Motor

From electrical characteristics:

τm = kt ia (31)
Eb = kb ωm (32)

Using Kirchhoff’s voltage law:

d
V a = La ia + Ra ia + Eb (33)
dt
From mechanical analysis:
d
τm = J ωL + bωL + τL (34)
dt

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
State Space Representation

Replacing Eb and τm :
d Ra kb 1
ia = − ia − ωm + Va (35)
dt La La La
d kt b 1
ωL = ia − ωL − τL (36)
dt J J J
Gear ratio relation:
rL
ωm = ωL (37)
rm
ω̇m = θm (38)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Transfer Function

Taking the Laplace Transform of the first equation with zero initial conditions:
Ra kb 1
sIa (s) = − Ia (s) − ωm (s) + Va (s)
La La La
1 kb
Ia (s) = Va − ωm (39)
sLa + Ra sLa + Ra

kt b 1
sωL (s) = Ia (s) − ωL (s) − τL
J J J
kt 1
ωL (s) = Ia (s) − τL (40)
Js + b Js + b

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Modeling of Chemical Processes

The dynamic model of chemical systems is derived from the balance equations for total mass,
energy, and component mass (mole). These equations are crucial for understanding and
predicting the behavior of chemical processes. Let’s delve into the concepts and equations
involved in the mathematical modeling of chemical processes.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
General Balance Equation

For a system with a well-defined boundary:

Change in Inventory = (IN − OUT)T ZB + (Generated − Loss)IZB (41)

Note: TZB (through the boundary of the system), IZB (in the boundary of the system).

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Choosing the Right Balance

To formulate the dynamic equation, we must decide the control boundary and which balance
to use:
• Mass, volume, or pressure → mass balance,
• Concentration → component balance,
• Temperature → energy balance,
• Interaction between flow and pressure → mechanical energy balance (momentum balance,
Bernoulli, Newton’s second law).

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mass Balance

The rate of change of mass accumulation in a closed boundary is:

dm
= win − wout (42)
dt
where win and wout are mass flow rates for the entering and exiting streams (bulk flow).
Introducing density:
dV
ρ = ρin qin − ρout qout (43)
dt
where qin and qout are volumetric flow rates.
Note:-Volumetric flow rate q is equal to the product of the flow rate of the fluid (ν) times the
area (a) perpendicular to the flow. (q = ν × a)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mass Balance for Constant Density

For constant density:

dV
= qin − qout (44)
dt
If the interest is the height of the liquid:
dh
A = qin − qout (45)
dt
where A is the cross-sectional area.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Dynamic Component Balance

For an arbitrary component A:

dnA
= FA,in − FA,out + GA (46)
dt
where nA is the inventory of component A, FA,in and FA,out are molar flow rates, and GA is
the net generation in chemical reactions.
With concentrations:
d(c̄A V )
= cA,in qin − cA,out qout + r̄A V [mol A/s] (47)
dt
where c̄A = nA /V , cA,in = FA,in /qin , cA,out = FA,out /qout , and GA = r̄A V .

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Dynamic Energy Balance

The general energy balance over a small duration of time is:

 
dH dV dp
= Hin − Hout + Q + Ws − (pex − p) +V [J/s] (48)
dt dt dt

where H is enthalpy, Q is supplied heat, and Ws is supplied mechanical work. Using specific
enthalpy ℏ = H
m:
 
dℏ dV dp
m = win (ℏin − ℏ) − wout (ℏout − ℏ) + Q + Ws − (pex − p) +V (49)
dt dt dt

Note: Enthalpy ℏ(T, p, f, nj ) is a function of temperature, pressure, phase distribution, and

composition.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Quadruple Tank Process

The quadruple tank process aims to control the liquid

levels in four tanks using two pumps.
• Objective: Control liquid levels in four tanks.
• Pumps:
• Pump 1 and Pump 2 draw liquid from a reservoir.
• Valves:
• Valve 1 (position γ1 ) regulates flow to Tank 1 and
Tank 4.
• Valve 2 (position γ2 ) regulates flow to Tank 2 and Figure: Quadruple tank process
Tank 3.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Flow Rates

The flow rates are given by:

q1in = γ1 k1 V1 q2in = γ2 k2 V2
q3in = (1 − γ2 )k2 V2 q4in = (1 − γ1 )k1 V1

From fluid dynamics, the rate of mass accumulation in each tank is equal to the difference
between mass inflow and mass outflow:

ṁ = ρ(qin − qout ) (50)

where qin is volumetric inflow rate and qout is volumetric outflow rate.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Bernoulli Equation

Using Bernoulli equation:

1 1
P1 + gh1 + v12 = P2 + gh2 + v22 (51)
2 2
If we take the top as point 2 and bottom of the tank as point 1 (h1 = 0), both of them are
exposed to atmospheric pressure so P1 = P2 , similarly since there is no inflow from the top
v2 = 0. The fluid flows out of each tank due to gravity only.
With these assumptions, equation 51 simplifies to:
p
v1 = 2gh2 (52)

Let the outlet have a cross-sectional area a, the outflow rate will be:
p
qout = av = 2gh (53)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Rate of Accumulation of Mass

m m
From basic physics ρ = V = Ah , the rate of accumulation of mass ṁ is:

dh
ṁ = ρA (54)
dt
Combining equation 50 and equation 54, the dynamic equation for each tank will be:

Aḣ = qin − qout (55)

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Dynamic Equations for Quadruple Tank System

The dynamic equations of the quadruple tank system are:

p p
A1 ḣ1 = γ1 k1 V1 + a3 2gh3 − a1 2gh1
p p
A2 ḣ2 = γ2 k2 V2 + a4 2gh4 − a2 2gh2
p
A3 ḣ3 = (1 − γ2 )k2 V2 − a3 2gh3
p
A4 ḣ4 = (1 − γ1 )k1 V1 − a4 2gh4

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Mathematical Modeling of Fluid Systems

Mathematical modeling of fluid systems involves using mathematical expressions to describe

the behavior and dynamics of fluids, whether they are in motion or at rest. These models are
essential for understanding and predicting fluid behavior in various engineering applications,
such as hydraulics, aerodynamics, and process engineering.
Fluid systems typically consist of a flow source or pressure source, energy-storing elements,
and energy-dissipating components, similar to electrical systems.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Basic Laws Used in Fluid Modeling

The modeling of fluid systems is based on fundamental physical laws, including the
conservation of mass, momentum, and energy. Key concepts include:
• Continuity Equation:

∇·v =0 (56)

where v is the fluid velocity vector.

• Bernoulli’s Equation:

1
P + ρv 2 + ρgh = constant (57)
2
where P is the pressure, ρ is the fluid density, v is the flow velocity, g is the acceleration
due to gravity, and h is the height above a reference level.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Basic Laws Used in Fluid Modeling

• Navier-Stokes Equations:
 
∂v
ρ + v · ∇v = −∇P + µ∇2 v + f (58)
∂t

where ρ is the fluid density, p is the pressure, µ is the dynamic viscosity, and f represents
body forces (e.g., gravity).
• Law of Conservation of Mass: This law states that the mass of a closed system must
remain constant over time.

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Key Fluid Properties
Several key properties are used to characterize fluid behavior in mathematical models:
• Fluid Inertia:
∆P
Lf = (59)
∆Q̇
where Lf is the fluid inertia.
• Fluid Capacitance:

∆V
C= (60)
∆P
where C is the fluid capacitance.
• Fluid Bulk Modulus:
∆P
β = ρo (61)
∆ρ
where ρo is a reference fluid density at a nominal pressure and temperature. AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Flow Characteristics

Fluid flow can be categorized into laminar and turbulent flow, each exhibiting different
characteristics and resistances.
• Laminar Flow:
∆P
RL = (62)
Q
• Turbulent Flow:
∆P
RT = (63)
Q2

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Hydraulic Piston and Cylinder System

Hydraulic systems are essential in various industrial applications, facilitating the efficient
transfer of power through the controlled movement of fluids.

Piston Head (A)

Qin CV

Figure: Hydraulic piston and cylinder

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Hydraulic Piston and Cylinder System (Cont.)

Using the law of conservation of mass, we can express the rate of accumulation of mass as the
difference between the mass inflow rate and the mass outflow rate.

ṁ = ρQin (64)

Substituting the mass m with the product of density and volume, m = ρV , we get:

ρV̇ + V ρ̇ = ρQin (65)

Replacing the density ρ using the relationship derived from the bulk modulus of elasticity:

ρo dP
ρ= (66)
β dt

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Hydraulic Piston and Cylinder System (Cont.)

Expressing the rate of change of volume V̇ in terms of the piston head area A and the
displacement of the piston head x:

V̇ = Aẋ (67)

Substituting these expressions into our earlier equation, we arrive at:

βAx
Ṗ = (Qin − Aẋ) (68)
ρ

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Hydraulic Piston and Cylinder System

The simple hydraulic actuator shown in Figure 13 is connected to a load and accompanied by
a restoring spring.

Piston head (A)

k Fatm
Fs
Qin Fh FL
FL Fv

Viscous friction (b)

Figure: Free body diagram of the hydraulic actuator

Figure: Hydraulic actuator

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Hydraulic Actuator System (Cont.)
The mathematical model describing the motion of the actuator is derived using Newton’s first
law of motion:

ΣF = mẍ (69)
Fh − Fatm − Fv − Fs − FL = mẍ (70)

Replacing Fh with P A and substituting Fatm = Patm A:

(P − Patm )A − FL = mẍ + bẋ + kx (71)

To find the pressure required to hold the piston at its nominal position, we can set the velocity
ẋ = 0 and assume x = x0 (the equilibrium position):

FL
P = + Patm (72)
A
This equation indicates that the pressure P must balance the load force per unit area, plus the
atmospheric pressure acting on the piston. AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Weight Measuring System

Simple weight measuring system, with system parameters k = 20N/m, ms = 1kg and
d = 9N s/m.
u = mo g

ms
Scale
x
k d

kx
If the output which is shown on the scale is the mass of the object to be measured (mo = g ).
(use g = 10m/s2 )
AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Shock Absorber

The 3D model and Mechanical System Model of a typical shock absorber is shown in Figures
15a and 15b respectively.
moving piston
fluid

x
u
d
SECTION A-A
spring

k m u
A

(a) 3D Model of Shock Absorber System (b) Mechanical System Model of Shock Absorber

Figure: Shock Absorber System

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Planar Drone
The planar quadrotor possesses two degrees of freedom in position, allowing movement along
the X and Y axes, and one degree of freedom in orientation, enabling rotation about the
vertical axis (yaw).

(a) Planar Drone (b) Planar drone FBD

Figure: Planar Drone

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Projector Screen Positioning System
The DC motor position control system for the remote-controlled projector screen incorporates a
DC motor paired with a non-inverting pre-amplifier to accurately manage the screen’s position.
+ ia
Ra La
+
− +

Rf ωm
Vs Va M Load

Ri
-
-
(a) Armature Voltage Controlled DC Motor (b) Remote Controlled Projector
With Pre-amplifier Screen

AASTU [Link]
Figure: Projector Screen Positioning System
Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Two Degree of Freedom Quarter Car Suspension System

A vehicle suspension system is a complex vibration system having multiple degrees of freedom.
The purpose of the suspension system is to isolate the vehicle body from the road imperfection.

y1
k2 d

y2
k1
z
x
Figure: 2-DOF Quarter Car Suspension Model

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
Ball and Beam System
The objective of ball and beam system is to control the beam’s angle to maintain the ball’s
desired position. The system has two main degrees of freedom: the ball’s position along the
beam and the beam’s angular position.

I,m
x




Figure: Ball and Beam System

AASTU [Link]

Mathematical Modeling and Control System Problem Formulation Daniel A. (MSc. & Science Ambassador)
 Control Systems and Control System Architectures

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
System Poles and Zeros

The transfer function provides a basis for determining important system response
characteristics without solving the complete differential equation. The transfer function is a
rational function in the complex variable s = σ + jω:

bm sm + bm−1 sm−1 + . . . + b1 s + b0
G(s) =
an sn + an−1 sn−1 + . . . + a1 s + a0

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Factored Form of Transfer Function

The transfer function can be factored into poles and zeros:

N (s) (s − z1 )(s − z2 ) . . . (s − zm )
G(s) = =K
D(s) (s − p1 )(s − p2 ) . . . (s − pn )

• zi are the zeros of the system (roots of N (s) = 0).

• pi are the poles of the system (roots of D(s) = 0).

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Behavior at Poles and Zeros

• When s = zi , the numerator N (s) = 0, and the transfer function vanishes:

lim G(s) = 0
s→zi

• When s = pi , the denominator D(s) = 0, and the transfer function becomes unbounded:

lim G(s) = ∞
s→pi

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Real and Complex Poles/Zeros

• All coefficients of N (s) and D(s) are real.

• Therefore, poles and zeros must be either:
• Purely real, or
• Appear in complex conjugate pairs.
• For poles:
pi = σi or pi , pi+1 = σi ± jωi
• Similarly, zeros are either real or appear in complex conjugate pairs.

Pole Zero Plot

The pole and zeros of any system can be plotted on complex plane.

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Stability Analysis
• If the poles of the system has negative real parts, the system is stable.
• If the poles of the system has zero real parts, the system is marginally stable.
• If the poles of the system has positive real parts, the system is unstable.

AASTU [Link]
Figure: Step responses vs. pole locations
Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
System Type and Order
Key Concepts in Control Systems

System Type
The system type refers to the number of integrators (poles at the origin) in the open-loop
transfer function. It is denoted by N :

System Order
The system order is the highest power of s in the denominator of the transfer function.

Relationship
• System type determines steady-state behavior.
• System order affects transient response and stability.

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Control System Overview

• A control system is modeled, and a control algorithm is designed to achieve desired

behavior.
• The controller must account for disturbances, noise, and modeling errors.

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Control Systems Type
• Feedforward Control: Measure the disturbance and compensate for it in the controller.

Feedforward Controller: D(s)

D(s)

D(s)

R(s) + U (s) Y (s)

C(s) G(s)
+
Figure: Feed-forward Control Architecture
AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Control Systems Type(Cont.)

• Open Loop Control: Predefined control actions without feedback.

Open Loop Controller: C(s), G(s)

R(s) U (s) Y (s)

C(s) G(s)

Figure: Open Loop Control Architecture

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Control Systems Type(Cont.)

• Feedback Control: Output is fed back to the input to correct errors.

Feedback Controller: C(s), G(s), H(s)

R(s) E(s) U (s) Y (s)

C(s) G(s)
+
−

Ym (s)
H(s)

Figure: Feed-back Control Architecture

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Control Systems Type(Cont.)

• Feedback system:

Y (s) = C(s)G(s)E(s), where E(s) = R(s) − H(s)Y (s)

Y (s) = C(s)G(s)[R(s) − H(s)Y (s)]

Y (s)(1 + C(s)G(s)H(s)) = C(s)G(s)R(s)
C(s)G(s)
Y (s) = R(s)
1 + C(s)G(s)H(s)
• Closed-loop transfer function:

G(s)C(s)
T (s) =
1 + G(s)C(s)H(s)

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Block Diagram Representation Control System

The general block diagram representation of control system is

Di (s) D0 (s)

R(s) Up (s) Y (s)

C(s) + G(s) +
+
−

Ym (s) N (s)
H(s) +

Figure: General Feed-back Control Architecture

where Di (s) is input disturbance, Do (s) is output disturbance and N (s) is measurement noise.
AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Block Diagram Representation(Cont.)

• System equations:

Y (s) = G(s)Up (s) + Do (s), Up (s) = Di (s) + C(s)E(s)

Y (s) = G(s)(di (s) + C(s)E(s)) + Do (s)

E(s) = R(s) − Ym (s), Ym (s) = H(s)(Y (s) + N (s))
Y (s) = G(s)Di (s) + G(s)C(s)[R(s) − H(s)(Y (s) + N (s))] + Do (s)
Y (s)[1 + G(s)C(s)H(s)] = G(s)Di (s) + G(s)C(s)R(s) − G(s)C(s)H(s)N (s) + Do (s)
G GC −GCH 1
Y = Di + R+ N+ Do
1 + GCH 1 + GCH 1 + GCH 1 + GCH

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Sensitivity Analysis

• Sensitivity definition:
∆T
T ∂T ∂G
S= ∆G
= ·
G
∂G T
where S measures the change in the closed-loop transfer function T due to changes in the
open-loop transfer function.
• For negative feedback the nominal sensitivity function is:

1
S=
1 + GCH
and complementary sensitivity (T = 1 − S):

GCH
T =
1 + GCH

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Sensitivity Analysis for Generalized Control System

• Sensitivities for Negative feedback:

∂Y G ∂Y GC
= , =
∂Di 1 + GCH ∂R 1 + GCH
∂Y −GCH ∂Y 1
= , =
∂N 1 + GCH ∂Do 1 + GCH
• Sensitivities for Positive feedback:
∂Y G ∂Y GC
= , =
∂Di 1 − GCH ∂R 1 − GCH
∂Y −GCH ∂Y 1
= , =
∂N 1 − GCH ∂Do 1 − GCH

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
Key Insights

• Negative Feedback:
• Reduces sensitivity to disturbances.
• Increases robustness to noise.
• Requires |1 + GCH| to be large for disturbance rejection.
• Control Design Trade-offs:
• Reducing disturbance sensitivity increases noise sensitivity.
• Design must balance disturbance rejection and noise immunity.

AASTU [Link]

Control Systems and Control System Architectures Daniel A. (MSc. & Science Ambassador)
 Block Diagram and Signal Flow Graph

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Block Diagrams in Control Systems

• Block diagrams represent the functional relationships between components within a

system
• All system variables are linked to each other through functional blocks
• Blocks represent various elements, and arrows indicate the flow of signals or information
between these elements

R(s) E(s) U (s) Y (s)

C(s) G(s)
+
−

Ym (s)
H(s)

Figure: Feed-back Control Architecture

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Elements of a Control System Block Diagram
• Arrow head pointing towards a block indicates the input
• Arrow head leading away from the block represents the output
• Arrows are referred to as signals
• Plus or minus at each arrow head indicates whether the signal is to be added or subtracted
• Summing points depict the summing or combining of signals
• Take-off points indicate signals extracted for monitoring or feedback

Figure: Basic Elements of Block Diagram

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Purpose of Block Diagrams

• Block diagrams provide a functional block, which is a symbol for the mathematical
operation on the input signal to the block that produces the output
• They offer a visual representation of the system, making it easier to understand the
relationships between different components

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Block Diagram Algebra

Basic operations in Block diagram algebra,

(a) Cascaded Blocks (b) Feedback Blocks

• Cascaded Blocks
• Parallel Blocks
• Feedback Blocks (d) Shift Summing
• Shift Summing Point (c) Parallel Blocks Point

• Shift Take away Point

(e) Shift Take away

Point

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Example

Reduce the following block diagram,

Figure: Example

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Signal Flow Graph

Signal flow graph is a graphical representation of algebraic equations.

• Input Node: It is a node, which has only outgoing branches.
• Output Node: It is a node, which has only incoming branches.
• Mixed Node: It is a node, which has both incoming and outgoing branches.

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Block Diagrams Vs Signal Flow Graphs

Conversion of Block Diagrams into Signal Flow Graphs

• Represent all the signals, variables, summing points and take-off points of block diagram
as nodes in signal flow graph.
• Represent the blocks of block diagram as branches in signal flow graph.
• Represent the transfer functions inside the blocks of block diagram as gains of the
branches in signal flow graph.
• Connect the nodes as per the block diagram. If there is connection between two nodes
(but there is no block in between), then represent the gain of the branch as one.

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
Block Diagrams Vs Signal Flow Graphs(Cont.)

AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
DC Motor in Block Diagram

1 kb
Ia (s) = Va − ωm
La s + Ra La s + Ra
kt 1
ωL (s) = Ia (s) − τL
Js + b Js + b
rL
ωm = ωL
rm

τL (s)
Va (s) 1
Ia (s) τm (s) 1
ωL (s)
+ La s+Ra kt Js+b
−

ωm (s) rL
kb rm

Figure: DC motor block diagram representation AASTU [Link]

Block Diagram and Signal Flow Graph Daniel A. (MSc. & Science Ambassador)
 Time Domain Analysis

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Time Domain Analysis

The analysis of a response of a system for time varying input, is called time domain analysis.
Any system has two types of response for time varying input,
• Transient response: the response of a system at the sudden change in the input. This
response dies out with time
lim ytr (t) = 0
t→∞
• Steady State response: the response of a system that remains after the transient response
dies out.
yss (t)
The total response of a system is the sum of the to responses.

y(t) = ytr (t) + yss (t)

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Standard Inputs used in Time Domain Analysis

Standard inputs are used to test the stability or behavior of a system.

(
1 t=0
• Unit Impulse: δ(t) =
0 otherwise
(
1 t≥0 • Unit Impulse: δ(s) = 1
• Unit Step: u(t) =
0 otherwise • Unit Step: U (s) = 1s
(
t t≥0 • Unit Ramp: R(s) = s12
• Unit Ramp: r(t) = 1
0 otherwise • Unit Parabolic: α(s) = 6s3
(2
t
t≥0
• Unit Parabolic: α(t) = 2
0 otherwise

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Steady-State Error Analysis

The deviation of the system response from the desired output. Using final value theorem,

ess (t) = lim e(t) = lim sE(s)

t→∞ s→0

The desired output of the system is r(t), thus

ess = lim s(R(s) − Y (s))

s→0

1
ess = lim s R(s)
s→0 1 + G(s)

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Steady-State Error Analysis (Cont.)

• Steady-State Error for Unit Step Input:

1
ess = , kp = lim G(s)
1 + kp s→0

• Steady-State Error for Unit Ramp Input:

1
ess = , kv = lim sG(s)
kv s→0

• Steady-State Error for Unit Parabolic Input:

1
ess = , ka = lim s2 G(s)
ka s→0

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

First Order Type 0 System Analysis

• A first order system with unity feedback have a transfer function of a form T (s) = 1
τ s+1 :

R(s) 1
Y (s)
+ τs
−

• The response for impulse input is,

1
Y (s) = T (s)R(s) =
τs + 1
• Inverse Laplace Transform:
1 −1t
y(t) = e τ
τ

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

First Order Type 0 System Analysis(Cont.)

• The steady state error (ess )

R(s)
ess = lim s =0
s→0 1 + G(s)
• The response for step input is,

1 1 1
Y (s) = T (s)R(s) = = −
s(τ s + 1) s τs + 1
• Inverse Laplace Transform:
1 1
y(t) = (1 − e− τ t )u(t)
τ
• The steady state error (ess )

R(s)
ess = lim s = 0.5
s→0 1 + G(s)
AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

First Order Type 0 System Analysis(Cont.)

• The response for ramp input is,

1 1 1 1
Y (s) = = 2− +
s2 (s+ 1) s s τs + 1
• Inverse Laplace Transform:

1 1
y(t) = (t − 1 + e− τ t )u(t)
τ
• The steady state error (ess )

R(s)
ess = lim s =∞
s→0 1 + G(s)

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

First Order Type 0 System Analysis(Cont.)

• Time constant τ is the time it takes for the system to reach 63.2% of its final value.
• For a first-order system:
y(t) = 1 − e−t/τ
• Settling time ts is the time it takes for the system to reach and stay within a certain
percentage of its final value.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

DC Gain

What is DC Gain?
The DC gain of a dynamic system refers to the steady-state output of the system when the
input is a constant (i.e., a DC signal). It is essentially the ratio of the output to the input
when the system has reached equilibrium, and all transient effects have died out. The DC gain
is determined by evaluating the transfer function at s = 0, which corresponds to the
steady-state response.

DC gain for OLTF and CLTF

For first order type 0 system,
• Open-Loop DC Gain:
k
G(0) = =k
0+1
• Closed-Loop DC Gain:
k k
T (0) = =
0+k+1 k+1
AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Second-Order System

• A second-order system is characterized by its natural frequency ωn and damping ratio ζ.

R(s) 2
ωn
Y (s)
+ s(s+2ζωn )
−

• The closed loop transfer function of a second-order system is:

ωn2
T (s) =
s2 + 2ζωn s + ωn2

where ζ is damping ratio/factor and ωn is natural frequency of the system.

• The poles of the system are located at:
p
s = −ζωn ± jωn 1 − ζ2

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Step Response of a Second-Order System

ωn2 1 s + 2ζωn
Y (s) = = − 2
s(s2 + 2ζωn s + ωn2 ) s s + 2ζωn s + ωn2
1 s + ζωn ζωn
= − −
s (s + ζωn ) + ωn (1 − ζ ) (s + ζωn ) + ωn2 (1 − ζ 2 )
2 2 2 2

s a
using this properties cos(at) = s2 +a2
, sin(at) = s2 +a2
and F (s − a) = L{eat f (t)}
!
−ζωn t ζ
y(t) = 1 − e cos(ωd t) + p sin(ωd t)
1 − ζ2

Comparing the output equation to r sin(ωd t + θ) = r sin ωd t cos θ + sin θ cos ωd t

p !
ζ −1 1 − ζ2 1
r sin θ = 1 r cos θ = p → θ = tan r=p
1−ζ 2 ζ 1 − ζ2
AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Step Response of a Second-Order System
Thus the step response of a second-order system is:
e−ζωn t
y(t) = 1 − p sin(ωd t + θ)
1 − ζ2
√ 
p 1−ζ 2
where ωd = ωn 1 − ζ 2 is the damped natural frequency, and θ = tan−1 ζ .

Figure: Second order system step response

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Time Domain Specifications

• Delay Time (td ): Time taken to reach 50% of the final value.

1 + 0.7ζ
td =
ωn
• Rise Time (tr ): Time taken to rise from 10% to 90% of the final value.

π−θ
tr =
ωd
• Peak Time (tp ): Time taken to reach the first peak of the response.
π
tp =
ωd

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Time Domain Specifications (Cont.)

• Maximum Overshoot (Mp ): Is the maximum peak value of the response curve
measured from the final value. ζπ
−√
Mp = e 1−ζ 2

the percentage overshoot is:

− √ ζπ
%Mp = 100 × e 1−ζ 2

• Settling time (ts ): The time required for the response to reach and stay within a certain
percentage (usually 2%) of the final value.
4
ts =
ζωn

for a 5% tolerance:
3
ts =
ζωn
AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Time Domain Specifications for Second Order System

Table: Time domain specification of second order systems

Time domain specification Formula

1 + 0.7ζ
Delay time td =
ωn
π−θ
Rise time tr =
ωd
π
Peak time tp =
 ωd
− √ ζπ
1−ζ 2
Percentage Peak overshoot (%Mp ) Mp = e × 100
4
Settling time with 2% tolerance ts =
ζωn

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Second Order System Analysis
p
The poles of the second order system are s1,2 = −ζωn ± jωn 1 − ζ 2 = −σ ± jωd , the plot of
the poles on complex plane is

jωd

ωn2 = ωd2 + σ 2 s1 • jωd

ωn
ζ = cos(θ) θ
σ
σ = ζωn σ
ωn
s2
• −jωd

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Second Order System Analysis(Cont.)

Analyzing the plot,

Change Effect on Poles Rise Time Settling Time Peak Time Delay Time Max % Overshoot
Increase ωn , Poles move ↓ ↓ ↓ ↑ Constant
keeping ζ radially outward
constant
Increase ωd , Poles move ↓ Constant ↓ ↓ ↓
keeping σ vertically toward
constant infinity
Increase σ, Poles move ↑ ↓ Constant ↓ ↑
keeping ωd horizontally
constant toward infinity
Increase ζ, Poles move along ↑ ↓ ↑ ↓ ↑
keeping ωn the circle and
constant meet on real axis

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Characteristics of Feedback Systems

• Effect on sensitivity to change in feedback gain H(s):

−H(s)G(s) −G(s)
S= = 1
1 + G(s)H(s) + G(s)
H(s)

as the magnitude of H(s) increase, the sensitivity S → −1

• Effect on time constant: For first order system G(s) = k
τ s+1 with feedback gain h,

k
kh+1
T (s) =  
τ
kh+1 s+1

as k increases or h increases time constant decreases.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Characteristics of Feedback Systems(Cont.)

• Effect on Overall gain:

G(s)
T (s) =
1 + G(s)H(s)
1
the magnitude decreases by a factor of 1+G(s)H(s) .
• Effect on stability: for first order k
system G(s) = τ s+1 , with feedback gain h

k
T (s) =
τ s + (1 + kh)

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Characteristic Polynomial and Modes

• The transfer function of a system is given by:

N (s) G(s)
T (s) = =
D(s) 1 + G(s)H(s)

where 1 + G(s)H(s) is the characteristic polynomial.

• The roots of 1 + G(s)H(s) = 0 are called the characteristic modes.
• Example:
s+0
T (s) = 2 =⇒ D(s) = s2 (s + 1)(s + 3)
s (s + 1)(s + 3)
• The characteristic modes are:

u(t), tu(t), e−t u(t), e−3t u(t)

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Approximating Higher Order Systems

• Dominant poles are the poles closest to the imaginary axis.

• They dominate the system’s response.
• The system can be approximated by its dominant poles (slowest Poles).

Dominant Poles Approximation

The dominant pole approximation is a method for approximating a (more complicated) high
order system with a (simpler) system of lower order if the location of the real part of some of
the system poles are sufficiently close to the origin compared to the other poles.

Caution
The DC gain of the exact system and the two approximate systems must be equal. This is
necessary to ensure that the final value of the step response (which is determined by T (0) for
closed loop or G(0) for open loop is unchanged.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Approximating Higher Order Systems(Cont.)

• Reduction of a second order system to first order

βe−αt − αe−βt
 
α·β
G (s) = K y (t) = K 1 −
(s + α) (s + β) β−α

If the magnitude of β is very large compared to α (typically if αβ > 5) we can write write
approximations for the transfer function (assuming s is sufficiently small compared to β),
as well as an approximation for the step response.

βe−αt
 
α·β α
= K 1 − e−αt


G (s) ≈ K =K y (t) ≈ K 1 −
(s + α) (β) (s + α) β

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Approximating Higher Order Systems(Cont.)

• Simplifying Higher Order System

α
(
α · ωn2 s+α , α << ζωn
G (s) = ≈ ω02
(s + α) (s2 + 2ζωn s + ωn2 ) 2 2,
s +2ζωn s+ωn
ζωn << α

In this case the test for the dominant pole compare ”α” against ”ζωn ”. This is because
ζωn is the real part of the complex conjugate root (we only compare the real parts of the
roots when determining dominance because it is the real part that determines how fast
the response decreases).

Caution
The DC gain of the exact system and the two approximate systems must be equal. This is
necessary to ensure that the final value of the step response (which is determined by H(0) is
unchanged.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

First-Order Approximation for Systems with Time Delay

Using Taylor series expansions for small values of s, the transfer function for a time delay can
be expressed as:
e−θ0 s ≈ 1 − θ0 s
or an alternative first-order approximation (Pade approximation) is:
θ0
1− 2s
e−θ0 s ≈ θ0
1+ 2s

or an alternative first-order approximation is:

1 1
e−θ0 s = ≈
eθ0 s 1 + θ0 s
where the time constant has a value of θ0 .

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Skogestad’s “Half Rule”

• Skogestad (2002) proposed a method for approximating higher-order models with multiple
time constants.
• The largest neglected time constant is approximated by adding half of its value to the
existing time delay and the other half to the smallest retained time constant.
• Smaller time constants are approximated as time delays using:

1
e−θ0 s ≈
1 + θ0 s

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Example 1:

Consider the transfer function:

K(−0.1s + 1)e−0.25s
G(s) =
(5s + 1)(3s + 1)(0.5s + 1)

Derive an approximate first-order-plus-time-delay model:

Ke−θs
G̃(s) =
τs + 1
using two methods:
• Taylor series expansions
• Skogestad’s half rule

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Solution: Taylor Series Approximation

Retain the dominant time constant (5). Applying the approximations:

−0.1s + 1 ≈ e−0.1s
1 1
≈ e−3s , ≈ e−0.5s
3s + 1 0.5s + 1
Substitution gives the Taylor series approximation:

Ke−0.1s e−3s e−0.5s e−0.25s Ke−3.85s

G̃(s) = =
5s + 1 5s + 1

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Solution: Skogestad’s Method

• Largest neglected time constant: 3.

• Half of this value is added to the next largest time constant:

τ = 5 + 0.5(3) = 6.5

• The other half provides a new time delay:

θ = 1.5 + 0.1 + 0.5 + 0.25 = 2.35

• The approximate model is:

Ke−2.35s
G̃SK (s) =
6.5s + 1

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Example 2:

Consider the transfer function:

K(−0.1s + 1)e−0.25s
G(s) =
(5s + 1)(3s + 1)(0.5s + 1)

Derive an approximate first-order-plus-time-delay model:

Ke−θs
G̃(s) =
τs + 1
using two methods:
• Taylor series expansions
• Skogestad’s half rule

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Exercise 1:

Consider the transfer function:

20(−0.1s + 1)e−0.25s
G(s) =
(3s + 1)(0.5s + 1)

Derive an approximate first-order-plus-time-delay model:

Ke−θs
G̃(s) =
τs + 1
using two methods:
• Taylor series expansions
• Skogestad’s half rule

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Exercise 2

Consider the transfer function:

20(1 − s)e−s
G(s) =
(12s + 1)(3s + 1)(0.2s + 1)(0.05s + 1)

Use Skogestad’s method to derive:

• A first-order-plus-time-delay model.

Ke−θs
G(s) =
(τ s + 1)
• A second-order-plus-time-delay model.

Ke−θs
G(s) =
(τ1 s + 1)(τ2 s + 1)

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Systems with Zeros

• Zeros in a dynamic system play a crucial role in shaping the system’s response.
• They influence the effect of poles and the overall behavior of the system.
• This presentation discusses two key interpretations of zeros in dynamic systems.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Interpretation 1: Zeros Define the Roles of Poles

• Zeros determine how the poles of the system influence the overall response.
• This can be shown using partial fraction expansion.
• Example: Consider the transfer function:

s+2
G(s) =
(s + 4)(s + 5)
• The zero at s = −2 modifies the contribution of the poles at s = −4 and s = −5.
• Partial fraction expansion reveals how the zero affects the weight of each pole in the
system’s response.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Interpretation 2: Zeros Add Rate of Change

• Zeros can be interpreted as adding the rate of change of the system’s response.
• Without the zero, the system’s response would follow the dynamics defined by the poles.
• The zero introduces an additional derivative-like effect, modifying the response.
• They introduce a ”derivative effect,” making the system respond faster to changes in
input.
• The derivative effect add additional overshoot, thus system with zero have larger
overshoot.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Zeros on the Right-Hand Side of the s-Plane

• Zeros on the right-hand side (RHS) of the s-plane (non-minimum phase zeros) make
control system design more challenging.
• These systems exhibit inverse response:
• The system’s initial response moves in the opposite direction before catching up.
• This behavior complicates control and stabilization efforts.

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Approximation of Higher Order systems with Zero
If the transfer function has zeros in the numerator, you need to be careful before applying the
dominant pole approximation. Consider

(s + 1.1)
G (s) =
(s + 1) (s + 10)

You might think that the pole at s = −1 would be dominant. However, the zero in the
numerator is so close to the pole location, that they almost cancel each other out. In this case
the dominant pole approximation cannot be applied. In fact, because the pole and zero almost
canel out they have little affect on the system, and it can be approximated by
1
G (s) ≈
(s + 10)

Note
It is important to stress that this simplification comes from the fact that the pole and zero
nearly cancel each other out, and not because of the dominant pole approximation.
AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Stability Analysis Using RH Table

For a system with characteristics equation 1 + G(s)H(s) = an sn + an−1 sn−1 + · · · a1 s + a0 ,

the Routh-Hurwitz Table is:
sn an an−2 an−4
s n−1 an−1 an−3 an−5
sn−2 b1 b2 b3
s n−3 c1 c2 c3
.. .. .. ..
. . . .
s0 d1 d2 d3
   
an an−2
1 1 an an−4
where b1 = − an−1 , b2 = − an−1 · · ·, and
an−1 an−3 an−1 an−5
   
a a a a
c1 = − b11 n−1 n−3 , b2 = − b11 n−1 n−5 · · ·
b1 b2 b1 b3

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Stability Analysis Using RH Table(Cont.)

Condition for Stability

• Necessary Condition: All coefficients of the characteristic polynomial must be positive
or negative.
• Sufficient Condition: All elements of the first column of the RH array must be
positive/negative (There should no sigh change).

Special Cases
• One Element is Zero in a row: Temporarily replace the zero with a small positive
number (often denoted as ϵ ) to proceed with the calculations. Evaluate the value as
ϵ → 0, to determine if there is sign change.
• All Elements of a row is Zero: Take the derivative of the polynomial corresponding to
the last non-zero row and use the coefficients from the derivative polynomial to fill in the
new row and proceed with the calculation. This shows that characteristic polynomial
has roots on the imaginary axis.
AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

Exercises

Determine the Stability of the following systems using RH criteria,

• 1 + G(s)D(s) = s4 + 3s3 + 3s2 + 2s + 1 = 0
• 1 + G(s)D(s) = s4 + 2s3 + s2 + 2s + 1 = 0
• 1 + G(s)D(s) = s5 + 3s4 + s3 + 3s2 + s + 3 = 0
• G(s) = s+2
s4 +2s3 +s2 +2s+1

AASTU [Link]

Time Domain Analysis Daniel A. (MSc. & Science Ambassador)

 Root Locus Controller Design

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Root Locus Basics
• The root locus is a graphical method to analyze how the roots of the characteristic
equation change with varying system parameters (e.g., gain k).
• The characteristic equation for a closed-loop system is:

1 + G(s)H(s) = 0

N (s)
1+k = 0 → D(s) + kN (s) = 0
D(s)
• The root locus plots the roots of the characteristic equation as k varies from 0 to ∞.
• if k = 0 the root of the closed loop system is equal to the roots of D(s), meaning the closed
loop poles are equal to open loop system poles.
• if k = ∞ the roots of the closed loop systems are the roots of N (s), meaning the closes loop
poles are equal to open loop system zeros.

Root Locus
Thus as k varies from zero to infinity, the closed loop poles moves from the location of open loop
system poles to the location of open loop system zeros. The path they create it called root locus.
AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Angle & Magnitude Conditions
Angle Condition: For a point s on the root locus, the angle of the open-loop transfer
function G(s)H(s) must satisfy:

∠G(s)H(s) = ±180◦ (2k + 1), k = 0, 1, 2, . . .

Magnitude Condition: The magnitude of the open-loop transfer function G(s)H(s) must
satisfy:
|G(s)H(s)| = 1

Note
If the magnitude of |G(s)H(s)| = 1 and the angle of ∠G(s)H(s) = 180o as a point in a
complex plane the point is in o a root locus.

Questions
• What happens if the magnitude of G(s)H(s) is greater than one (|G(s)H(s)| ≥ 1)?
• What happens if the angle of G(s)H(s) is greater than 180o ( ∠G(s)H(s) ≥ 180o )?
AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Rules for Construction of Root Locus

• Rule 1: Locate the open-loop poles and zeros in the s-plane.

• Rule 2: Determine the number of root locus branches N :

N =P if P ≥ Z

N =Z if P < Z
where P is the number of finite open-loop poles and Z is the number of finite open-loop
zeros.
• Rule 3 Identify real axis root locus branches: A point on the real axis is part of the
root locus if the angle of the open-loop transfer function at that point is an odd multiple
of 180◦ . This occurs if an odd number of open-loop poles and zeros lie to the left of the
point.

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Rules for Construction of Root Locus (Cont.)

• Rule 4 Centroid and Angle of Asymptotes: If P ̸= Z, some root locus branches

approach infinity. The intersection point of asymptotes on the real axis is called the
centroid:
P P
Real part of finite open-loop poles − Real part of finite open-loop zeros
α=
P −Z
The angle of asymptotes is given by:

(2q + 1)180◦
θ= , q = 0, 1, 2, . . . , (P − Z) − 1
P −Z
• Rule 5 Intersection with Imaginary Axis: The intersection points of root locus
branches with the imaginary axis can be found using the Routh array method. If all
elements of any row in the Routh array are zero, the root locus intersects the imaginary
axis.
AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Rules for Construction of Root Locus (Cont.)

• Rule 6 Break-away points and Break-in points: break-away points occur between two
open-loop poles on the real axis, while break-in points occur between two open-loop zeros
on the real axis.
• To find these points:
• Express k in terms of s from the characteristic equation 1 + kG(s)H(s) = 0.
• Differentiate k with respect to s and set it to zero.
• Solve for s to find the break points.

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Rules for Construction of Root Locus (Cont.)

• Rule 7 Angle of Departure and Arrival:

• Angle of Departure (ϕd ) at complex conjugate poles:

ϕd = 180◦ + ϕ

• Angle of Arrival (ϕa ) at complex conjugate zeros:

ϕa = 180◦ − ϕ

where ϕ = ∠G(s)H(s), excluding the pole or zero being considered.

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Summary of steps to construct root locus

• Step 1: Locate the open-loop poles and zeros.

• Step 2: Determine the number of branches of the root locus.
• Step 3: Identify and draw the real axis segments of the root locus.
• Step 4: Find the asymptotes and centroid.
• Step 5: Determine breakaway and break-in points.
• Step 6: Find the intersection of the root locus with the imaginary axis.
• Step 7: Calculate the angle of departure and arrival.

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Example: Root Locus Construction

• Consider a system with:

k(s + 3)
G(s)H(s) =
s2+ 4s + 9
√
• Open-loop poles: s = −2 ± j 5
• Open-loop zero: s = −3
• Number of branches: P = 2, Z = 1
• Real axis segments: Between s = −3 and s = −∞

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Example: Root Locus Construction(Cont.)

• Centroid: P P
Real parts of poles − Real parts of zeros
σ=
P −Z
• For the example:
(−2) + (−2) − (−3)
σ= = −1
2−1
• Asymptotes angles:

(2k + 1)π
θ= , k = 0, 1, . . . , P − Z − 1
P −Z
• For the example:
(2 · 0 + 1)π
θ= =π
1

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Example: Root Locus Construction(Cont.)

• Breakaway and break-in points are found by solving:

dk(s)
=0
ds
• For the example:
s2 + 4s + 9
k(s) = −
s+3
dk(s) (2s + 4)(s + 3) − (s2 + 4s + 9)
=− =0
ds (s + 3)2
• Solving gives:
s = −0.5505 and s = −5.4495
• Evaluating for which value k is positive:
k(s) = 6.899 for s = −5.4495 and k(s) = −2.899 for s = −0.5505, the valid breakaway
and break-in point is s = −5.4495.
AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Example: Root Locus Construction(Cont.)

• The intersection with the imaginary axis is found using the Routh-Hurwitz criterion.

s2 + (4 + k)s + (9 + 3k) = 0

• The auxiliary equation is:

s2 + (9 + 3k) = 0
• Solving gives: √
s = ±j 3
√
thus the intersection with the imaginary axis is is s = ±j 3.

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Example: Root Locus Construction(Cont.)

√
• The angle of departure from a s1 = −2 + j 5 is:
√
θds1 = 180◦ + ∠G(s)H(s)(s + 2 − j 5) = 155.905o
√
• The angle of departure from a s2 = −2 − j 5 is:
√
θds2 = 180◦ + ∠G(s)H(s)(s + 2 + j 5) = 204.095o

• The angle of arrival at s = −3 is

θa@ s=−3 = 180◦ − ∠G(s)H(s)(s + 3) = 180o

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

Example: Root Locus Construction(Cont.)

Example
k(s+3)
• G(s) = s(s+1)(s+2)(s+4)
• G(s) = k
s(s+1)(s+2)(s+4)
• G(s) = k
s(s2 +2s+3)

Effect of adding zero and poles

• Adding poles to a transfer function shift the root locus to the right.
• Adding zeros to a transfer function shift the root locus ti the left.

AASTU [Link]

Root Locus Controller Design Daniel A. (MSc. & Science Ambassador)

 Frequency Domain Analysis

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Introduction to Frequency Response Analysis

• Frequency response analysis focuses on the steady-state response of a system to a

sinusoidal input.
• The output is a sinusoidal signal with the same frequency but different amplitude and
phase.
• The transfer function G(jω) is represented in terms of magnitude and phase.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

What is Frequency Response?

The response of a system can be divided into:

• Transient response: Found using Fourier integrals.
• Steady-state response: The system’s response to a sinusoidal input.
For a Linear Time-Invariant (LTI) system, a sinusoidal input produces a steady-state output
that is also sinusoidal, with the same frequency but different amplitude and phase.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Domain Response

The response of G(s) for cosine input r(t) = ro cos(ωo t) is

s
Y (s) = G(s)R(s) = ro G(s)
s2 + ωo2

Using partial fraction on G(s)

a∗o
   
a1 a1 ao
Y (s) = + + ··· + +
s + p1 s + p2 s − jωo s + jωo
−p1 t −p2 t
+ · · · + ao ejωo t + a∗o e−jωo t



y(t) = a1 e + a2 e
y(s) = ytr (t) + yss (t)

The first term is the transient response, while the second term is steady state response. The
steady state response can be written as

Yss (t) = ao ejωo t + a∗o e−jωo t

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Domain Response(Cont.)

Since ao = Aejϕ , replacing

Yss (t) =Aejϕ ejωo t + Ae−jϕ e−jωo t

!
e(jωo t+ϕ) + e−(jωo t+ϕ)
=A = 2A cos(ωo t + ϕ)
2

Determining the residues

ro G(jωo )
ao = (s − jωo )Y (s)|s=jωo =
2
|ro G(jωo )| |ro G(jωo )|
= ∠G(jωo ) → A = , ϕ = ∠G(jωo )
2 2
thus,
yss (t) = ro |G(jωo )| cos(ωo t + ∠G(jωo ))

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Domain Response(Cont.)

The output signal for a sinusoidal input is:

y(t) = ro |G(jω0 )| sin(ω0 t + ∠G(jω0 ))

• Amplitude: Output amplitude is the product of the input amplitude A and |G(jω0 )|.
• Phase: Output phase is the sum of the input phase and ∠G(jω0 ).
Where:
• ro is the amplitude of the input signal.
• ω0 is the angular frequency of the input signal:

ω0 = 2πf0

where f0 is the frequency of the input signal.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Domain Specifications
The frequency domain specifications include:
• Resonant peak
• Resonant frequency
• Bandwidth
Consider the transfer function of a second-order closed-loop control system:
Y (s) ωn2
T (s) = = 2
R(s) s + 2δωn s + ωn2
Substitute s = jω:
ωn2
T (jω) =
(jω)2 + 2δωn (jω) + ωn2
Simplify:
ωn2 1
T (jω) = =
−ω 2 + 2jδωn + ωn2
  
ω2
1− 2
ωn
+ j 2δω
ωn
ω
Let u = ωn :
1
T (jω) =
(1 − u2 ) + j(2δu) AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Magnitude and Phase of T (jω)

The magnitude of T (jω) is:

1
M = |T (jω)| = p
(1 − u )2 + (2δu)2
2

The phase of T (jω) is:  

−1 2δu
∠T (jω) = − tan
1 − u2

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Resonant Frequency
The resonant frequency ωr is the frequency at which the magnitude of the frequency response
has its peak value for the first time. At ω = ωr , the first derivative of the magnitude of T (jω)
is zero.
Differentiate M with respect to u:
dM 1 − 3 
= − (1 − u2 )2 + (2δu)2 2 2(1 − u2 )(−2u) + 2(2δu)(2δ)

du 2
Simplify:
dM 1 − 3 
= − (1 − u2 )2 + (2δu)2 2 4u(u2 − 1 + 2δ 2 )

du 2
Set dM
du = 0 at u = ur :
1 − 3 
(1 − u2r )2 + (2δur )2 2 4ur (u2r − 1 + 2δ 2 )

0=−
2
This implies:
4ur (u2r − 1 + 2δ 2 ) = 0
Thus:
u2r − 1 + 2δ 2 = 0 ⇒ u2r = 1 − 2δ 2 AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Resonant Frequency

The resonant frequency ωr is given by:

p
ur = 1 − 2δ 2
ωr
Substitute ur = ωn :
ωr p
= 1 − 2δ 2
ωn
Thus: p
ωr = ωn 1 − 2δ 2

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Resonant Peak
The resonant peak Mr is the maximum value of the magnitude of T (jω):
1
Mr = p
(1 − u2r )2 + (2δur )2
√
Substitute ur = 1 − 2δ 2 and 1 − u2r = 2δ 2 :
1
Mr = q √
(2δ 2 )2 + (2δ 1 − 2δ 2 )2

Simplify:
1
Mr = √
2δ 1 − δ 2

• The resonant peak in frequency response corresponds to the peak overshoot in the time
domain transient response for certain values of damping ratio δ.
• Resonant peak and peak overshoot are correlated. AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bandwidth
Bandwidth is the range of frequencies over which the magnitude of T (jω) drops to 70.7% of
its zero-frequency value.
At ω = 0, u = 0:
1
M=p =1
(1 − 0 ) + (2δ(0))2
2 2

• At 3-dB frequency, the magnitude of T (jω) is 70.7% of its value at ω = 0.

1
ω = ωB , M = 0.707(1) = √
2
1 1
⇒M = √ = q
2 (1 − u2b )2 + (2δub )2

Let u2b = x
⇒ 2 = (1 − x)2 + (2δ)2 x
p
−(4δ 2 − 2) ± (4δ 2 − 2)2 + 4
⇒x=
2 AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bandwidth(Cont.)

Consider only the positive value of x.

p
⇒ x = 1 − 2δ 2 + (2 − 4δ 2 + 4δ 4 )
ωb2
Substitute x = u2b = 2
ωn
q p
⇒ ωb = ωn 1 − 2δ 2 + (2 − 4δ 2 + 4δ 4 )

Bandwidth ωb in the frequency response is inversely proportional to the rise time tr in the time
domain transient response.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Domain Analysis
Frequency domain analysis is used to study the steady-state response of a system to a
sinusoidal input. The frequency response of an open loop transfer function G(jω)H(jω), as ω
varies. The behavior is of the response is analyzed by analyzing,
Magnitude of the Transfer Function:

|G(jω)H(jω)|

Phase Angle of the Transfer Function:

∠G(jω)H(jω)

Remark
• In control systems, the analysis of a system’s open-loop transfer function in the frequency
domain can provide insights into the closed-loop time domain behavior of the system.
• Several methods exist to plot the frequency response, each with its own strengths and
weaknesses. AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Response Plotting Methods

• Bode Plot: Two graphs (magnitude in dB and phase in degrees) plotted against log
frequency.
• Nyquist Plot: Polar plot of the real and imaginary parts of the transfer function.
• Nichols Chart: Magnitude (dB) vs. phase (degrees) of the frequency response.
• Polar Plot: Magnitude and phase plotted in polar coordinates.
• Linear Magnitude and Phase Plot: Magnitude and phase plotted on linear scales.
• Spectrogram: Shows how frequency content changes over time.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Frequency Response Plotting Methods(Cont.)
Summary

• Bode Plot: Best for general analysis and control system design.
• Nyquist Plot: Ideal for stability analysis.
• Nichols Chart: Useful for closed-loop system performance.
• Polar Plot: Simple representation for specific frequencies.
• Linear Plot: Easy to interpret for narrow ranges.
• 3D Plot: Comprehensive for multi-dimensional responses.
• Spectrogram: Best for time-varying systems.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot

Note
In the Bode Plot, the magnitude (in decibels) and the phase angle (in degrees) of the
open-loop transfer function G(jω)H(jω) are plotted against frequency using a logarithmic
scale.
An open loop transfer function is composed of proportional gain, simple poles or zeros at
origin, simple poles or zeros on real axis and quadratic poles or zeros,

k(τa s + 1)(τb s + 1) . . . (τm s + 1)

G(s)H(s) =
sn (τ1 s + 1)(τ2 s + 1) . . . (τn s + 1)

The frequency response of the system is:

k(jτa ω + 1)(jτb ω + 1) . . . (jτm ω + 1)

G(jω)H(jω) =
(jω)n (jτ1 ω + 1)(jτ2 ω + 1) . . . (jτn ω + 1)

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot(Cont.)

The Magnitude of the OLTF in Decibel’s (Taking 20 log)

 p q 
2 2 2 2
20 log(|G(jω)H(jω)|) = 20 log k + 20 log( τa ω + 1) + log( τb ω + 1) + . . .
 q q 
2 2 2 2
− 20n log(ω) − 20 log( τ1 ω + 1) + log( τ2 ω + 1) + . . .

p
Approximating log( τa2 ω 2 + 1)
 p


 τ a ω >> 1 log( τa2 ω 2 + 1) ≈ 20 log(τa ω) = 20 log(τa ) + 20 log(ω)

τ >> ω
a 20 log(τa ) + 20 log(ω) ≈ 20 log(τa )


 τa << ω 20 log(τa ) + 20 log(ω) ≈ 20 log(ω)
τ ω << 1 log(pτ 2 ω 2 + 1) ≈ log(1) = 0

a a

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot(Cont.)

The angle ∠G(jω)H(jω) is

∠G(jω)H(jω) = 0 + (tan(τa ω) + tan(τb ω) + · · · + tan(τm ω))

− 90n − (tan(τ1 ω) + tan(τ2 ω) + · · · + tan(τn ω))

τa , τb , . . . τ1 , τ2 are called corner frequencies, the frequencies at which the plot change
direction.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot(Cont.)

• Second Order Quadratic Zero

G(s)H(s) = s2 + 2ξωn s + ωn2 → G(jω)H(jω) = −ω 2 + j2ξωn ω + ωn2
p q
ω 2
Approximating 20 log( (ωn − ω)2 + 4(ξω)2 ) → 20 log(ωn ) + 20 log( (1 − ωn ) + 4(ξ ωωn )2 )
 q
ω ω 2

ωn << 1 20 log(ωn ) + 20 log( (1 − ωn ) + 4(ξ ωωn )2 ) ≈ 20 log(ωn ) + 20 log(1)
q
ω ω 2

ωn >> 1 20 log(ωn ) + 20 log( (1 − ωn ) + 4(ξ ωωn )2 ) ≈ 20 log(ωn ) + 40 log( ωωn )

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot(Cont.)

• Second Order Quadratic Pole

2
ωn ωn2
G(s)H(s) = s2 +2ξω 2 → G(jω)H(jω) = −ω 2 +j2ξω ω+ω 2
n s+ωn n
p q n
ω 2
2 2
Approximating −20 log( (ωn − ω) + 4(ξω) ) → −20 log( (1 − ωn ) + 4(ξ ωωn )2 )
 q
ω ω 2

ωn << 1 −20 log( (1 − ωn ) + 4(ξ ωωn )2 ) ≈ 20 log(1)
q
ω ω 2

ωn >> 1 −20 log( (1 − ωn ) + 4(ξ ωωn )2 ) ≈ −40 log( ωωn )

Corner Frequency
For quadratic poles and zeros, the natural frequency ωn is a corner frequency of the system.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot(Cont.)

• Time Delay: G(s) = e−T s → G(jω) = e−jT ω , the magnitude of |G(jω)| = 1 and the
angle ∠G(jω) − ωT , the phase angle increases as frequency increases.
• Non-minimum phase systems:
τ1 s − 1
G1 (s) =
τ2 s + 1
have larger phase variation than
τ1 s + 1
G2 (s) =
τ2 s + 1

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bode Plot(Cont.)

Bode Plots of Typical Transfer Functions:

• G(s) = ±k → G(jω) = ±k
• G(s) = s → G(jω) = jω
1
• G(s) = 1s → G(jω) = jω
• G(s) = s + a → G(jω) = jω + a
1
• G(s) = s+a 1
→ G(jω) = jω+a
• G(s) = e−T s → G(jω) = e−jT ω
ωn2 2
ωn
• G(s) = s2 +2ξωn s+ωn2 → G(jω) = −ω 2 +j2ξω ω+ω 2
n n
• 2 2 2
G(s) = s + 2ξωn s + ωn → G(jω) = −ω + j2ξωn ω + ωn2
• G1 (s) = ττ12 s−1 τ1 s+1
s+1 Versus G1 (s) = τ2 s+1

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Stability Analysis in Frequency Domain

Gain Margin
• Factor by which the gain of the system is less than the critical value |G(jω)| = 1 or how
much gain can we add to the OLTF gain when ∠G(jω) = 180o and still have stable
system.
• Mathematically: Gain Margin is the ratio of critical gain to the actual gain of the system.

Critical Gain 1
GM = =
Actual Gain |G(jω)|

in DB
GM = −20 log |G(jω)|ωpc
Where ωpc is the gain cross over frequency, the frequency at which ∠G(jω) = 180o

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Stability Analysis in Frequency Domain(Cont.)

Phase Margin
• Factor by which the angle of the system is less than the critical value ∠G(jω) = 180o or
how much phase can we add to the OLTF gain when|G(jω)| = 1 and still have stable
system.
• Mathematically:
P M = 180o + ∠G(jω)|ωgc
Where ωgc is the gain cross over frequency, the frequency at which |G(jω)| = 1.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Stability Analysis in Frequency Domain(Cont.)

Stability
A system is stable if Both phase margin and gain margin are positive.

Note
• If ωgc < ωpc both gain margin and phase margin are positive hence the system is stable.
• If ωgc > ωpc both gain margin and phase margin are negative hence the system is
unstable.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Connecting Time Domain to Bode Plot

• Natural Frequency (ωn ):

• In the Bode plot, ωn is the corner frequency where the magnitude starts to roll off.
• Damping Ratio (ζ):
• Affects the sharpness of the resonance peak in the magnitude plot.
• Higher ζ reduces the peak and increases the phase margin.
• Resonance Peak:
• The peak in the magnitude plot corresponds to the system’s overshoot in the time domain.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Resonance Peak and Overshoot

• The resonance peak in the Bode plot is related to the overshoot in the time domain.
• For a second-order system:

1
Resonance Peak = p
2ζ 1 − ζ 2
• Higher resonance peak ⇒ higher overshoot.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Bandwidth and Settling Time

• Bandwidth (ωBW ):
• The frequency range where the magnitude is above −3 dB.
• For a second-order system:
q p
ωBW ≈ ωn 1 − 2ζ 2 + 4ζ 4 − 4ζ 2 + 2

• Relation to Settling Time:

• Higher bandwidth ⇒ faster settling time.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Phase Margin and Damping Ratio

• Phase Margin (PM):

• The difference between the phase angle and −180◦ at the gain crossover frequency.
• For a second-order system:
P M ≈ 100ζ (for small ζ)
• Relation to Damping Ratio:
• Higher phase margin ⇒ higher damping ratio ⇒ lower overshoot.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Summary

• Time-domain specifications (overshoot, settling time, peak time) are closely related to
Bode plot characteristics.
• Key connections:
• Resonance peak ↔ Overshoot
• Bandwidth ↔ Settling time
• Phase margin ↔ Damping ratio
• Understanding these relationships helps in designing control systems with desired
performance.

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

Example: Bode Plot

Example
• G(s) = k
s(s+1)(1+0.1s)
• G(s) = k
s(0.02s+1)(0.04s+2)

AASTU [Link]

Frequency Domain Analysis Daniel A. (MSc. & Science Ambassador)

 Compensator Design

Daniel A. (MSc. & Science Ambassador)

Department of Electromechanical Engineering

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Compensator Design

• To add a PM (Phase Margin) or Gain Margin (GM) to a system, additional dynamics are
required.

R(s) Y (s)
D(s) G(s)
+
−

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Compensator Transfer Function

The compensator transfer function is given by:

k(s + z)
D(s) = ⇒ modulated form
(s + p)
    
z s/z + 1 τs + 1
D(s) = k = kα
p s/p + 1 ατ s + 1

Condition
 
z
The magnitude determines whether the compensator is a lead or lag compensator
p
 
• Lead Compenstor → zp < 1
 
• Lag Compenstor → zp > 1

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Analysis

The frequency response of the lead compensator is:

  
z jω/z + 1
G(jω) = k
p jω/p + 1

The magnitude in decibels is:

 
z p  p 
20 log (|G(jω)|) = 20 log k + 20 log (jω/z)2 + 1 − 20 log (jω/p)2 + 1
p

Corner Frequencies
ωc1 = z and ωc2 = p since z < p → ωc1 < ωc2

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Analysis(Cont.)

The controller reduce gain for low frequency inputs and add gain to high frequency inputs (i.e
amplification).

Figure: Magnitude Plot

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Analysis (Cont.)

The phase of the compensator is

ω   
ω
∠D(jω) = tan−1 − tan−1
z p

for the first term:

• ω << z the phase of the system is zero
• ω = z the phase of the system is 45o
• ω >> z the phase of the system is 90o
for the second term:
• ω << p the phase of the system is zero
• ω = p the phase of the system is 45o
• ω >> p the phase of the system is 90o

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Analysis (Cont.)

The lead compensator add phase to the system between ωc1 and ωc2 , meaning the phase of
the output leads the input, where the name comes from.

Figure: Phase Plot

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lag Compensator Analysis

The frequency response of the lag compensator is:

  
z jω/z + 1
G(jω) = k
p jω/p + 1

The magnitude in decibels is:

 
z p  p 
20 log (|G(jω)|) = 20 log k + 20 log (jω/z)2 + 1 − 20 log (jω/p)2 + 1
p

Corner Frequencies
ωc1 = z and ωc2 = p since z > p → ωc1 > ωc2

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Analysis(Cont.)

The controller reduce gain for high frequency inputs and add gain to low frequency inputs (i.e
amplification).

Figure: Magnitude Plot

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lag Compensator Analysis (Cont.)

The phase of the compensator is

ω   
ω
∠D(jω) = tan−1 − tan−1
z p

for the first term:

• ω << z the phase of the system is zero
• ω = z the phase of the system is −45o
• ω >> z the phase of the system is −90o
for the second term:
• ω << p the phase of the system is zero
• ω = p the phase of the system is −45o
• ω >> p the phase of the system is −90o

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lag Compensator Analysis (Cont.)

The Lag compensator remove phase from the system between ωc1 and ωc2 , meaning the phase
of the output lags the input, where the name comes from.

Figure: Phase Plot

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead vs. Lag Compensator

Remarks on Lead Compensators

• Improves the phase margin and thus the transient performance (overshoot) of the system.
• The gain crossover frequency is increased. Thus, the speed of the system response is
improved.
• Increases the high frequency gain of the system. This makes the system more susceptible
to noise signals.

Remarks on Lag Compensators

• Allow a high gain at low frequencies and thus improve the steady state error.
• Reduce the gain crossover frequency of the system. Thus, the system response becomes
slower.
• Reduce the high frequency gain of the system. Hence, the high frequency noise
attenuation becomes better.
• The time constant of the lag compensator should not be unnecessarily large. Otherwise,
the system’s response will be too slow. AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead vs. Lag Compensator (Cont.)

Comparison
Parameter Lead Compensator Lag Compensator
Rise Time Decreases (↓) Increases (↑)
Gain Crossover Frequency Increases (↑) Decreases (↓)

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Compensator Design Specifications

Lead Compensator Design

Design Requirements May be:
• Desired steady-state error requirements.
• Desired phase margin requirements ϕd .
or
• Desired closed-loop bandwidth requirements.
• Desired steady-state error requirements.

Note
Sometimes phase margin is may not be directly given, instead maximum percentage overshoot
will be given, then solve for ξ and use the relation ξ = 0.01P M

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Compensator Design Specifications(Cont.)

The phase angle of a compenstor is

ϕ = ∠D(jω) = tan−1 (ωτ ) − tan−1 (αωτ )

the maximum phase we are able to add or remove occurs when

∂ϕ τ ατ
= − =0
∂ω 1 + ω 2 τ 2 1 + α2 ω 2 τ 2
Solving for ω
1
ω= √
τ α
the corresponding maximum angle ϕmax is
 
−1 1−α
ϕmax = sin
1+α

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Compensator Design Specifications(Cont.)

The value of α that will ensure the maximum angle is

1 − sin ϕmax
α=
1 + sin ϕmax
which means the magnitude of the compenstor at ϕmax is,
√
ω2τ 2 + 1 k
|D(jω)| = k √ =√
2 2 2
α ω τ +1 α

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Design Steps
• Step 1: Determine the DC gain kα of D(s)G(s) that meet the design specification.
• Step 2: Determine the gain cross over frequency of the system with the gain kα.
• Step 3: Determine the PM of the system with cross over frequency from step 2.
• Step 4: Determine the maximum phase we need to add,
ϕmax = ϕd − P M + (5o ≈ 10o )
• Step 5: Determine α
1 − sin(ϕmax )
α=
1 + sin(ϕmax )
• Step 6: Determine the gain cross frequency of the compensated system using,
|D(jω)G(jω)| = |D(jω)||G(jω)| = 1
|D(jω)| = √kα
• Step 7: Determine τ using the new cross over frequency from step 6.
1
τ= √
ωgc α AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator Design Steps(Cont.)
If we want to add a phase at some required frequency (We by pass Step 2 ) which is called
bandwidth design method.
• Step 1: Determine the DC gain kα of D(s)G(s) that meet the design specification.
• Step 3: Determine the phase margin of the system at the required frequency.
• Step 4: Determine the maximum phase we need to add,
ϕmax = ϕd − P M + (5o ∼ 10o )
• Step 5: Determine α
1 − sin(ϕmax )
α=
1 + sin(ϕmax )
• Step 6: Determine the gain cross frequency of the compensated system
|D(jω)G(jω)| = 1
|D(jω)| = √k
α
• Step 7: Determine τ using the new cross over frequency from step 6.
1
τ= √
ωgc α AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lag Compensator Design Steps

• Step 1: Determine the DC gain kα of D(s)G(s) that meet the design specification.
• Step 2: Find the phase angle that will make the desired phase margin, with the gain kα.

P M = 180o + ∠(kα)G(jω) − (5o ∼ 10o )

∠(kα)G(jω) = P M − 180o + (5o ∼ 10o )

• Step 3: Find the frequency at which this phase angle occurs and make it the gain cross
over frequency.
• Step 4: Determine α from
|D(jω)G(jω)| = 1
at the gain cross over frequency ωgc and Step 1.
• Step 5: Select τ > 10
ωgc

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Exercise

Given
10
G(s) =
s(s + 1)

1 Design a lead compensator, such that the phase margin P M ≥ 40◦ and the DC gain for a
ramp input is greater than 100.
2 Design a lead compensator that achieve a steady-state error of less than 0.1 for a ramp
input and a maximum overshoot of 25% for a step input.
3 Design a lead compensator, such that the phase margin P M ≥ 40◦ and the DC gain for a
ramp input is greater than 100 at ω = 10 rad/s.
4 Design a lag compensator, such that the phase margin P M ≥ 40◦ and the steady-state
error for a ramp input is less than 0.01.

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

PID Controller Representation

The controller is represented as the ratio of the signal into the plant U (s) and the error signal
E(s):
U (s)
Gc (s) =
E(s)
or
U (s) = Gc (s)E(s)

R(s) E(s) U (s) Y (s)

Gc (s) G(s)
+
−

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

PID Controller Representation (Cont.)

The Parallel realization of PID controller is

U (s) = Gc (s)E(s)

kp
R(s) E(s) ki
U (s) Y (s)
+ G(s)
+ s
−
kd s

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Proportional (P) Control

Output:
u(t) = kp e(t)
Transfer Function:
Gc (s) = kp
Characteristics:
• Output is proportional to the error signal.
• Tends to cause oscillations.

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Integral (I) Control

Output: Z
u(t) = ki e(t) dt

Transfer Function:
ki
Gc (s) =
s
Characteristics:
• Eliminates steady-state error.
• Can lead to slower response times.

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Proportional+Integral (PI) Controllers

Output: Z t
u(t) = kp e(t) + ki e(λ) dλ

Transfer Function:  τi 
Gc (s) = kp 1 +
s
Alternative Form:
k(s + z)
Gc (s) =
s
Characteristics:
• Combines proportional and integral control.
• Pole at the origin, zero at −z.

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Proportional+Derivative (PD) Controllers

Output:
d
u(t) = kp e(t) + kd e(t)
dt
Transfer Function:
Gc (s) = kp (1 + τd s)
Alternative Form:
Gc (s) = k(s + z)
Characteristics:
• Combines proportional and derivative control.
• Zero at −z.

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Proportional+Integral+Derivative (PID) Controllers

Output: Z t
d
u(t) = kp e(t) + ki e(λ) dλ + kd e(t)
dt
Transfer Function:  τi 
Gc (s) = kp 1 + + τd s
s
Alternative Form:
k(s + z1 )(s + z2 )
Gc (s) =
s
Characteristics:
• Combines proportional, integral, and derivative control.
• Pole at the origin, zeros at −z1 and −z2 .

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator vs. PD Controller
 
• If α << 1, |D(jω)| = kα τ s+1
ατ s+1 ≈ kα(τ s + 1) =⇒ PD controller.

Lead Compensator Vs PD Controller

100

80
Magnitude (dB)

60

40
Lead
20 PD

90
Phase (deg)

Lead
45 PD

0
10 -2 10 -1 10 0 10 1 10 2 10 3 10 4 10 5 10 6
Frequency (rad/s)

Figure: Lead Compensator vs PD controller

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead Compensator vs. PD Controller (Cont.)

Lead Compensator: PD Controller:

s+z C(s) = kp (1 + τd s)
D(s) = k , where z < p
s+p
Frequency Domain Characteristics:
Frequency Domain Characteristics:
• Magnitude Plot:
• Magnitude Plot:
  • Low frequency: |D(jω)| ≈ kp
• Low frequency: |D(jω)| ≈ k pz (Proportional term dominates).
• High frequency: |D(jω)| ≈ k • High frequency: |D(jω)| ≈ kd ω
• Phase Plot: (Derivative term dominates).
• Maximum phase lead at ωm = √z · p: • Phase Plot:
• Low frequency: ∠D(jω) ≈ 0◦
 
−1 p−z (Proportional term dominates).
ϕmax = sin • High frequency: ∠D(jω) ≈ 90◦
p+z
(Derivative term dominates).

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Key Differences

Feature Lead Compensator   PD Controller

z
Magnitude Plot Gradual increase from k p to Constant kp at low frequencies,
k linear increase at high frequen-
cies
Phase Plot Positive phase shift with a peak Positive phase shift approaching
at ωm 90◦ at high frequencies
Noise Sensitivity Low (due to pole p) High (due to derivative term)
High-Frequency Gain Limited by pole p Unbounded (increases linearly)

• Lead Compensator:
• Improves phase margin and stability.
• Limits high-frequency gain, reducing noise sensitivity.
• PD Controller:
• Improves transient response and damping.
• Highly sensitive to noise due to unbounded high-frequency gain.
AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lag Compensator vs. PI Controller
 
• If α >> 1, |D(jω)| = kα τ s+1 τ s+1 1


ατ s+1 ≈ kα ατ s = k(1 + τs ) =⇒ PI controller.

Lag Compensator Vs PI Controller

200
Lag
PI
Magnitude (dB)

150

100

50
0
Lag
PI
Phase (deg)

-45

-90
10-5 10-4 10-3 10-2 10-1 100 101 102
Frequency (rad/s)

Figure: Lag Compensator vs PI controller

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lag Compensator vs. PI Controller (Cont.)
Lag Compensator:
s+z PI Controller:
D(s) = k , where z > p
s+p ki
D(s) = kp +
s
Frequency Domain Characteristics:
• Magnitude Plot: Frequency Domain Characteristics:
• Low frequency: |D(jω)| ≈ k • Magnitude Plot:
• Low frequency: |D(jω)| ≈ kωi (integrator
 
• High frequency: |D(jω)| ≈ k pz
dominates).
• Phase Plot: • High frequency: |D(jω)| ≈ kp
• Introduces a negative phase shift (proportional term dominates).
(phase lag). • Phase Plot:
• Maximum phase lag occurs at
√ • Low frequency: ∠D(jω) ≈ −90◦
ωm = z · p:
(integrator dominates).
  • High frequency: ∠D(jω) ≈ 0◦
−1 z−p
ϕmax = sin (proportional term dominates).
z+p
AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Key Differences

Feature Lag Compensator   PI Controller

Magnitude Plot Gradual decrease from k to k z
p High gain at low frequencies ( kωi ),
constant at high frequencies (kp )
Phase Plot Negative phase shift (phase lag) −90◦ at low frequencies, 0◦ at high
frequencies
Purpose Improves steady-state accuracy Improves steady-state accuracy
without affecting stability and eliminates steady-state error
Noise Sensitivity Low (due to pole p) Low (integrator reduces high-
frequency noise)

• Lag Compensator:
• Used to improve steady-state accuracy without significantly affecting transient response.
• Reduces high-frequency gain, improving noise immunity.
• PI Controller:
• Used to eliminate steady-state error (e.g., for step inputs).
• Provides high gain at low frequencies, improving steady-state accuracy.
AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead-Lag Compensator and PID Controller

Consider
τ1 s + 1
D1 (s) = kα
ατ1 s + 1
and
τ2 s + 1
D2 (s) = kβ
βτ2 s + 1
if α << 1 and β >> 1
If the two compensators are cascaded D(s) = D1 (s)D2 (s) the result will be,
 
2 1
D(s) = k α(τ1 s + 1) 1 +
τ2 s
  
2 1 τ1
D(s) = k α τ1 s + + 1+
τ2 s τ2
 
Which is a PID controller with kp = k 2 α 1 + ττ12 , τd = ττ11+τ
τ2
2
1
and τi = τ1 +τ 2
,
AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

 Lead-Lag Compensator Vs PID Controller
200
Magnitude (dB)

150

100 Lead Lag

PID
50
90
Phase (deg)

45
0
-45 Lead Lag
PID
-90
10-4 10-2 100 102 104 106
Frequency (rad/s)

Figure: Lead-Lag Compensator vs PID controller

AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Lead-Lag Compensator vs. PID Controller
Lead-Lag Compensator:

(s + z1 )(s + z2 ) PID Controller:

D(s) = k
(s + p1 )(s + p2 ) ki
D(s) = kp + + kd s
Frequency Domain Characteristics: s
• Magnitude Plot: Frequency Domain Characteristics:
 
• Low frequency: |D(jω)| ≈ k z2
p2 (lag • Magnitude Plot:
effect). • Low frequency: |D(jω)| ≈ kωi (integrator
 
• High frequency: |D(jω)| ≈ k z1
(lead dominates).
p1 • High frequency: |D(jω)| ≈ kd ω
effect). (differentiator dominates).
• Phase Plot: • Phase Plot:
• Introduces a positive phase shift (phase • Low frequency: ∠D(jω) ≈ −90◦
lead) at mid-frequencies. (integrator dominates).
• Introduces a negative phase shift • High frequency: ∠D(jω) ≈ +90◦
(phase lag) at low frequencies. (differentiator dominates).
• Maximum phase lead/lag depends on
pole-zero placement. AASTU [Link]

Compensator Design Daniel A. (MSc. & Science Ambassador)

Key Differences
Feature Lead-Lag Compensator PID Controller
Magnitude Plot Adjusts gain at specific frequencies High gain at low ( kωi ) and high
(lead/lag) (kd ω) frequencies
Phase Plot Positive phase shift (lead) and neg- −90◦ at low frequencies, +90◦ at
ative phase shift (lag) high frequencies
Purpose Improves transient response (lead) Improves transient response,
and steady-state accuracy (lag) steady-state accuracy, and reduces
overshoot
Noise Sensitivity Low (lag reduces high-frequency High (differentiator amplifies high-
noise) frequency noise)
• Lead-Lag Compensator:
• Combines benefits of lead (improves transient response) and lag (improves steady-state
accuracy).
• Reduces high-frequency noise due to lag compensator.
• PID Controller:
• Provides integral action to eliminate steady-state error and derivative action to improve
transient response.
• Sensitive to high-frequency noise due to the derivative term. AASTU [Link]

View publication stats

Compensator Design Daniel A. (MSc. & Science Ambassador)