INTRODUCTION TO AUTOMATIC CONTROL

1. AUTOMATIC CONTROL

1.1 INTRODUCTION

Control systems are an integral part of modern society. Numerous applications are
around us: the rocket fire and the space shuttle lifts off to earth orbit; a self-guided vehicle
delivering material to work stations in aerospace assembly plant glides along the floor seeking
destination.

These are just a few examples of automatically control system that we can create.

Within our own bodies numerically control systems such as the pancreas which regulates our
blood sugar. In time of “fight or flight” our adrenaline increases along with heart rate causing
more oxygen to be delivered to our cells.

Our eyes follow a moving object to keep it in view; our hands grasp the object and place it
precisely at a predetermined location.

1.2 DEFINTION

A control system consists of subsystem and processes (or plant) assembled for the purpose of
controlling the outputs of the processes. For example, a furnace produces heat output as a
result of the floor of fuel. In this process, subsystems called fuel valves and fuel valves
actuators are used to regulate the temperature of a room by controlling the heat output from
the furnace.

Others subsystems such as thermostats, which act as sensor, measure the room temperature.

In its sum from, in a control system provides an output or response for a given input or
stimulus.

Input, stimulus Output, response

Control system
Desired response Actualresponse

AVANTAGES OF CONTROL SYSTEMS

With control systems we can move large equipment with precision. We can point huge
antennas to wand the farthest reaches of the universe to pick up faint radius signals,
controlling these antennas by hand would be impossible.

Because of control system, elevators carry up quickly to our destination, automatically

stopping at the right floor.
We build control systems for four primary reasons:

1 Power amplification
2 Remote control
3 Convenience of input form
4 Compensation for disturbance

1) A BRIEF HISTORY OF CONTROL SYSTEM

Feedback control systems are older than humanity. Numerous biological control systems
built into the earless inhabitant of our planet. Let us now look at a brief history of human-
designed control systems.

- Liquid-level control

The Greeks began engineering feedback system around 300 BC. A water clock
invented by ktesibios operated by having water trickle into a measuring container at a
constant rate could be used to tell time.

- Steam Pressure and Temperature Controls

Regulation of steam pressure began around 1681 with Denis Papin's invention of the
safety valve. The concept was further elaborated on by weighting the valve top. If the
upward Pressure from the boiler exceeded the weight, steam was released, and the
pressure decreased. If it did not exceed the weight, the valve did not open, and the
pressure inside the boiler increased.

- Speed Control

In the eighteenth century, James Watt invented the flyball speed governor to control
the speed of steam engines.
 - Stability

Alexandr Michailovich Lyapunov also contributed to the development and formulation of

today's theories and practice of control system stability. During the second half of the 1800s,
the development of control systems, based on the steering and stabilizing of ships. In1874,
Henry Bessemer, using a gyroscope to sense a ship's motion and applying power generated by
the ship's hydraulic system, moved the ship to keep it stable.

Twentieth-Century Developments

It was not until the early 1900s that automatic steering of ships was achieved. In1922,
the Sperry Gyroscope Company installed an automatic steering system that used the elements
of compensation and adaptive control to improve performance. However, much of the general
theory used today to improve the performance of automatic control systems is attributed to
Nicholas Minorsky, a Russian born in 1885. It was his theoretical development applied to the
automatic steering of ships that led to what we call to day proportional-plus-integral-plus-
derivative (PID).

In the late 1920 s and early 1930s, H.W.Bode and H.Nyquist at Bell Telephone
Laboratories developed the analysis of feedback amplifiers. These contributions evolved into
sinusoidal frequency analysis and design techniques currently used for feedback control
system.

In1948, Walter R. Evans, working in the aircraft industry, developed a graphical

technique to plot the roots of a characteristic equation of a feedback system whose parameters
changed over a particular range of values.

Contemporary Applications

Today, control systems find widespread application in the guidance, navigation, and control
of missiles and spacecraft, as well as planes and ships at sea. For example, modern ships use a
combination of electrical, mechanical, and hydraulic components to develop rudder
commands in response to desired heading commands. The rudder commands, in turn, result in
a rudder angle that steers the ship.

2) RESPONSE CHARACTERISTIC, AND SYSTEM CONFIGURATION

As noted earlier, a control system provides an input or response for a give input. The input
represents a desired response; the output is the actual response. For example, when the 4th
floor button of an elevator is pushed on the ground floor, the elevator rises to the 4th floor with
speed and floor-leveling accuracy designed for passengers’ comfort.
 3) OPEN-LOOP AND CLOSED-LOOP SYSTEMS

4.1) Open-loop system

An open loop system is a system with an input, output and a process, where the output
doesn’t influence the input.

The distinguishing characteristic of an open-loop system is that it cannot compensate for any
disturbance.

4.2) Closed-loop system

A closed-loop system is a system where the output is feedback to the system.

Closed-loop systems have the advantages of greater accuracy than the open-loop systems,
they are less sensitive to noise disturbance. Transient response and steady-state error can be
controlled more conveniently.

4) ANALYSIS AND DESIGN OBJECTIVES

The three major objectives of feedback analysis, and design and desired transient response
steady-state error and achieving stability.

- Desired transient response:

Not too fast, not too slow
- Steady-state response:
Find a compensator to reduce the steady-state error.
- Stability:
Transient (natural) response should decay to zero as time t →∞
- Total response
 CHAPTER 2 : MODELING IN
THE FREQUENCY DOMAIN

2.1. Introduction
In this chapter, we will develop mathematical model from schematic of physical systems. The
method of transfer function in the frequency domain will be discussed.

As we proceed, we will notice that in every case the step in developing a mathematical model
is to apply the fundamental physical laws of science.

For example, when we model electrical network, ohm’s law and Kirchhoff’s laws will be
applied.

When we study mechanical systems, we use network’s laws as the fundamental guiding
principles (we will sum the forces or torques).

Les Equations Différentielles Ordinaires (EDO) Linéaire à coefficient constant

Une fois le schéma dessiné, le concepteur utilise des lois physiques, telles que les lois de Kirchhoff
pour les réseaux électriques et de Newton pour les systèmes mécaniques, ainsi que des hypothèses
simplificatrices, pour modéliser mathématiquement le système. Ces lois sont :

La loi du courant de Kirchhoff : la somme des courants électriques circulant à partir d'un nœud est
égale à zéro.

Les lois de Newton : La somme des forces sur un corps est égale à zéro ; et la somme des moments
sur un corps est égale à zéro.

Les lois de Kirchhoff et de Newton conduisent à des modèles mathématiques qui décrivent la relation
entre l'entrée et la sortie des systèmes dynamiques. L'un de ces modèles est l'équation différentielle
linéaire, invariante dans le temps, Eq. (1.2) :
From these equations, we will obtain the relationship between the system’s input and output.

The differential equations representation are transformed to transfer function representation

using Laplace transform.

Exemple : système linéaire à coefficient constants

Soit le système d’entrée u et de sortie y décrit par l’équation différentielle suivante :

𝑦̈ + 𝑎1 𝑦̇ + 𝑎0 𝑦 = 𝑏𝑢

2.2. Laplace transforms (Review)

2.2.1. Definition
+∞
𝐿{𝑓(𝑡)} = 𝐹(𝑠) = ∫ 𝑓(𝑡)𝑒 −𝑠𝑡 𝑑𝑡
0

Where 𝑠 = 𝜎 + 𝑗𝜔, a complex variable.

F(s) is called Laplace transform.

2.2.2. Laplace transform table

2.2.3. Laplace transforms theorems
❖ Theorem
2.2.4. Inverse Laplace transform

𝑓(𝑡) = 𝐿−1 {𝐹(𝑠)}

By table :

1
𝐿−1 { } = 𝑢(𝑡)
𝑠
𝜔
𝐿−1 { 2 } = sin(𝜔𝑡) 𝑢(𝑡)
𝑠 + 𝜔2
𝑛!
𝐿−1 { } = 𝑡 𝑛 𝑢(𝑡)
𝑠 𝑛+1
By partial-fraction expansion
Example:

Find 𝑳−𝟏 {𝒀(𝒔)}

(𝑠+2)(𝑠+4) 𝐴 𝐵 𝐶
• 𝑌(𝑠) = 𝑠(𝑠+1)(𝑠+3) = 𝑠 + 𝑠+1 + 𝑠+3

(𝑠 + 2)(𝑠 + 4) (0 + 2)(0 + 4) 8
𝐴 = 𝑠𝑌(𝑠)|𝑠=0 = | = =
(𝑠 + 1)(𝑠 + 3) 𝑠=0 (0 + 1)(0 + 3) 3

(𝑠 + 2)(𝑠 + 4) (−1 + 2)(−1 + 4) 3

𝐵 = (𝑠 + 1)𝑌(𝑠)|𝑠=−1 = | = =−
𝑠(𝑠 + 3) 𝑠=−1
−1(−1 + 3) 2

(𝑠 + 2)(𝑠 + 4) (−3 + 2)(−3 + 4) 1

𝐶 = (𝑠+3)𝑌(𝑠)|𝑠=−3 = | = =−
𝑠(𝑠 + 1) 𝑠=−3
−3(−3 + 1) 6

8 1 3 1 1 1
𝑌(𝑠) = × − × − ×
3 𝑠 2 𝑠+1 6 𝑠+3

8 −1 1 3 1 1 1
𝐿−1 {𝑌(𝑠)} = 𝐿 { } − 𝐿−1 { } − 𝐿−1 { }
3 𝑠 2 𝑠+1 6 𝑠+3

𝟖 𝟑 𝟏
𝑳−𝟏 {𝒀(𝒔)} = 𝒖(𝒕) − 𝒆−𝒕 𝒖(𝒕) − 𝒆−𝟑𝒕 𝒖(𝒕)
𝟑 𝟐 𝟔

(𝑠+2)(𝑠+4) 𝐴 𝐵 𝐶 𝐷
• 𝑌(𝑠) = = + + +
𝑠(𝑠+1)2 (𝑠+3) 𝑠 𝑠+1 (𝑠+1)2 𝑠+3

(𝑠 + 2)(𝑠 + 4) (0 + 2)(0 + 4) 8
𝐴 = 𝑠𝑌(𝑠)|𝑠=0 = 2
| = 2
=
(𝑠 + 1) (𝑠 + 3) 𝑠=0 (0 + 1) (0 + 3) 3

(𝑠 + 2)(𝑠 + 4) (−1 + 2)(−1 + 4) 3

𝐶 = (𝑠 + 1)2 𝑌(𝑠)|𝑠=−1 = | = =−
𝑠(𝑠 + 3) 𝑠=−1
−1(−1 + 3) 2

(𝑠 + 2)(𝑠 + 4) (−3 + 2)(−3 + 4) 1

𝐷 = (𝑠+3)𝑌(𝑠)|𝑠=−3 = 2
| = 2
=
𝑠(𝑠 + 1) 𝑠=−3
−3(−3 + 1) 12
𝑑 (𝑠 + 2)(𝑠 + 4) 𝑑 𝑠 2 + 6𝑠 + 8
𝐵= ( ) = ( 2 )
𝑑𝑠 𝑠(𝑠 + 3) 𝑠=−1
𝑑𝑡 𝑠 + 6𝑠 𝑠=−1
(2𝑠 + 6)(𝑠 2 + 6𝑠) − (𝑠 2 + 6𝑠 + 8)(2𝑠 + 6) 32
𝐵= 2 2
| =−
(𝑠 + 6𝑠) (25)2
𝑠=−1

8 1 32 1 3 1 1 1
𝑌(𝑠) = × − × − × 2
+ ×
3 𝑠 625 𝑠 + 1 2 (𝑠 + 1) 12 𝑠 + 3
 8 −1 1 32 −1 1 3 1 1 1
𝐿−1 {𝑌(𝑠)} = 𝐿 { }− 𝐿 { } − 𝐿−1 { 2
} + 𝐿−1 { }
3 𝑠 625 𝑠+1 2 (𝑠 + 1) 12 𝑠+3

𝟖 32 −𝒕 𝟑 𝟏 −𝟑𝒕
𝑳−𝟏 {𝒀(𝒔)} = 𝒖(𝒕) − 𝒆 𝒖(𝒕) − 𝒕𝒆−𝒕 𝒖(𝒕) + 𝒆 𝒖(𝒕)
𝟑 625 𝟐 𝟏𝟐

2.2.5. The transfer function

We are now ready to formulate the system representation by establishing a via definition for
the function that algebraically relates a system’s output c(t) to its input r(t).

Let us begin by writing a general nth order linear time –invariant differential equation:

𝑑𝑛 𝑐(𝑡) 𝑑 𝑛−1 𝑐(𝑡) 𝑑 𝑚 𝑟(𝑡) 𝑑 𝑚−1 𝑟(𝑡)

𝑎𝑛 + 𝑎𝑛−1 + ⋯ + 𝑎0 𝑐(𝑡) = 𝑏𝑚 + 𝑏𝑚−1 + ⋯ + 𝑏0 𝑟(𝑡)
𝑑𝑡 𝑛 𝑑𝑡 𝑛−1 𝑑𝑡 𝑚 𝑑𝑡 𝑚−1
𝑎𝑖 ’s and 𝑏𝑖 ’s are constant coefficients.

Taking the Laplace transform of both sides assuming zero initial conditions.

We have:

(𝑎𝑛 𝑆 𝑛 + 𝑎𝑛−1 𝑆 −1𝑛 + ⋯ + 𝑎0 )𝐶(𝑠) = (𝑏𝑚 𝑆 𝑚 + 𝑏𝑚−1 𝑆 𝑚−1 + ⋯ + 𝑏0 )𝑅(𝑠)

𝐶(𝑠) 𝑏𝑚 𝑆 𝑚 + 𝑏𝑚−1 𝑆 𝑚−1 + ⋯ + 𝑏0

𝐺(𝑠) = = (∗∗)
𝑅(𝑠) 𝑎𝑛 𝑆 𝑛 + 𝑎𝑛−1 𝑆 −1𝑛 + ⋯ + 𝑎0

Equation (**) separates the output C(s), the input R(s) and the system.

The ratio of polynomial in s on the right is called the transfer function G(s).

Block diagram representation

R(s) G(s) C(s)

Notice that the denominator of G(s) is the characteristic polynomial of the differential
equation and 𝑎𝑛 𝑆 𝑛 + 𝑎𝑛−1 𝑆 −1𝑛 + ⋯ + 𝑎0 = 0 is called the characteristic equation differential
equation.

The output is obtained by: 𝐶(𝑠) = 𝑅(𝑠)𝐺(𝑠)

Example 1:
 𝑑𝑐(𝑡)
+ 2𝑐(𝑡) = 𝑟(𝑡)
𝑑𝑡

Taking Laplace transform 𝑠𝐶(𝑠) + 2𝐶(𝑠) = 𝑅(𝑠) ⇒ (𝑠 + 2)𝐶(𝑠) = 𝑅(𝑠)

𝐶(𝑠) 1
𝐺(𝑠) = =
𝑅(𝑠) 𝑆 + 2

Example 2:

Taking the input like the step function:

1
𝑟(𝑡) = 𝑢(𝑡) ⇒ 𝑅(𝑠) =
𝑠
Find c(t):

1 1 1⁄2 1⁄2
𝐶(𝑠) = 𝑅(𝑠)𝐺(𝑠) = = −
𝑠𝑠+2 𝑠 𝑠+2
So
1 1
𝐶(𝑠) = 𝑢(𝑡) − 𝑒 −2𝑡 𝑢(𝑡)
2 2
Example 3:

2𝑠 + 1
𝐺(𝑠) = ⇒ (𝑠 2 + 6𝑠 + 2)𝐶(𝑠) = (2𝑠 + 1)𝑅(𝑠)
𝑠 2 + 6𝑠 + 2
𝑠 2 𝐶(𝑠) + 6𝑠𝐶(𝑠) + 2𝐶(𝑠) = 2𝑠𝑅(𝑠) + 𝑅(𝑠)

𝑑2 𝑐(𝑡) 𝑑𝑐(𝑡) 𝑑𝑟(𝑡)

2
+6 + 2𝑐(𝑡) = 2 + 𝑟(𝑡)
𝑑𝑡 𝑑𝑡 𝑑𝑡
Exemple :

Soit à calculer le niveau de l’eau ℎ(𝑡) dans un récipient de base A avec un flux entrant et un flux
sortant. Les flux entrant et sortant sont donnés par 𝑞(𝑡) = 𝑤𝑒 −𝑡 et 𝑘ℎ(𝑡) respectivement, où

𝑤 et 𝑘 sont des constantes de proportionnalité. La conservation de masse conduit à une équation

différentielle linéaire de premier ordre pour le niveau de l’eau ℎ(𝑡).

𝑑ℎ(𝑡)
𝐴 𝑑𝑡
+ 𝑘ℎ(𝑡) = 𝑞(𝑡) = 𝑤𝑒 −𝑡

La condition initiale est ℎ(𝑡 = 0) = ℎ𝑜 = 4

Les valeurs numériques sont :

1 1
𝑘 = 2 𝑚2 /𝑠 , 𝑤 = 4 𝑚2 /𝑠 et 𝐴 = 4 𝑚2 , .

La solution s’effectue en trois étapes, en utilisant la transformation de Laplace.

𝑑ℎ(𝑡)
𝑇𝐿 [𝐴 + 𝑘ℎ(𝑡) = 𝑞(𝑡) = 𝑤𝑒 −𝑡 ]
𝑑𝑡

𝑑ℎ(𝑡)
𝐴. 𝑇𝐿 { } + 𝑘𝑇𝐿{ℎ(𝑡)} = 𝑇𝐿{𝑞(𝑡)} = 𝑤𝑇𝐿{𝑒 −𝑡 }
𝑑𝑡
𝑑ℎ(𝑡)
𝐴. 𝑇𝐿 { } + 𝑘𝑇𝐿{ℎ(𝑡)} = 𝑇𝐿{𝑞(𝑡)} = 𝑤𝑇𝐿{𝑒 −𝑡 }
𝑑𝑡
𝑤
𝐴(𝑠𝐻(𝑠) − ℎ𝑜 ) + 𝑘𝐻(𝑠) = 𝑄(𝑠) =
𝑠+1

Soit, avec les valeurs numériques :

1 1 4
(𝑠𝐻(𝑠) − 4) + 𝐻(𝑠) = 𝑄(𝑠) =
4 2 𝑠+1

Solution algébrique dans le domaine image :

1 4 𝑠+5
(𝑠 + 2)𝐻(𝑠) = +1=
4 𝑠+1 𝑠+1

D’où l’on tire

4(𝑠 + 5)
𝐻(𝑠) =
(𝑠 + 1)(𝑠 + 2)

On doit effectuer la décomposition en éléments simples pour 𝐻(𝑠) :

4(𝑠 + 5) 𝐴 𝐵
𝐻(𝑠) = = +
(𝑠 + 1)(𝑠 + 2) (𝑠 + 1) (𝑠 + 2)
Pour calculer A, on multiplie cette équation par (𝑠 + 1) et on pose 𝑠 = −1 :

4(𝑠 + 5) 4(−1 + 5)
𝐴= | = = 16
(𝑠 + 2) 𝑠=−1 (−1 + 2)

De même, on obtient pour B :

4(𝑠 + 5) 4(−2 + 5)
𝐴= | = = −12
(𝑠 + 1) 𝑠=−2 (−2 + 1)

Ainsi, 𝐻(𝑠) s’écrie sous forme :

16 12
𝐻(𝑠) = −
(𝑠 + 1) (𝑠 + 2)

Transformation inverse vers le domaine original à l’aide des tables des transformées.

16 12
ℎ(𝑡) = 𝑇𝐿 { } − 𝑇𝐿 { }
(𝑠 + 1) (𝑠 + 2)

ℎ(𝑡) = 16𝑒 −𝑡 − 12𝑒 −2𝑡

2.3. Electric Network Transfer Function

1. Modélisations des Systèmes Electriques

1.1.Circuit RL forcé

R : La résistance du circuit, en ohm (𝛀)

L : L’inductance de la bobine, en henrys (H)

𝒗(𝒕): La tension appliquée au circuit, en volts (V)

𝒊(𝒕) : L’intensité du courant électrique, en ampère (A)

La loi de maille :

𝑣𝐿 (𝑡) + 𝑣𝑅 (𝑡) = 𝑣(𝑡)

𝑑𝑖(𝑡)
𝐿 𝑑𝑡
+ 𝑅𝑖(𝑡) = 𝑣(𝑡)
 1.2.Circuit RLC en série avec boucle Simple

R : La résistance du circuit, en ohm (𝛀)

L : L’inductance de la bobine, en henrys (H)

C : La capacité électrique du condensateur en farads(F)

𝒗(𝒕): La tension appliquée au circuit, en volts (V)

𝒊(𝒕) : L’intensité du courant électrique, en ampère (A)

La loi de maille :

𝑣𝐿 (𝑡) + 𝑣𝑅 (𝑡) + 𝑣𝐶 (𝑡) = 𝑣(𝑡)

𝑡
𝑑𝑖(𝑡) 1
𝐿 + 𝑅𝑖(𝑡) + ∫ 𝑖(𝜏)𝑑(𝜏) = 𝑣(𝑡)
𝑑𝑡 𝐶
0

Soit 𝒒(𝒕) la charge électrique du condensateur, en coulombs, (C)

𝑑𝑞(𝑡)
On a : 𝑖(𝑡) =
𝑑𝑡

𝑑 2 𝑞(𝑡) 𝑑𝑞(𝑡) 1
𝐿 𝑑𝑡 2
+𝑅 𝑑𝑡
+ 𝐶 𝑞(𝑡) = 𝑣(𝑡)

𝑞(𝑡) = 𝐶𝑣𝑐 (𝑡)

𝑑 2 𝑣𝑐 (𝑡) 𝑑𝑣𝑐 (𝑡)

𝐿𝐶 𝑑𝑡 2
+ 𝑅𝐶 𝑑𝑡
+ 𝑣𝑐 (𝑡) = 𝑣(𝑡)
 𝑑 2 𝑣𝑐 (𝑡) 𝑑𝑣𝑐 (𝑡)
𝐿𝐶 𝑑𝑡 2
+ 𝑅𝐶 𝑑𝑡
+ 𝑣𝑐 (𝑡) = 𝑣(𝑡)

𝑑 2 𝑣𝑐 (𝑡) 𝑅 𝑑𝑣𝑐 (𝑡) 1 1

𝑑𝑡 2
+𝐿 𝑑𝑡
+ 𝐿𝐶 𝑣𝑐 (𝑡) = 𝐿𝐶 𝑣(𝑡)

𝑅 1 1
𝑣̈𝑐 (𝑡) + 𝐿 𝑣̇𝑐 (𝑡) + 𝐿𝐶 𝑣𝑐 (𝑡) = 𝐿𝐶 𝑣(𝑡)

Fonction de Transfert
Méthode de La transformée de Laplace
En prenant la transformée de Laplace de l’équation différentielle, en ignorant les conditions
initiales on obtient :
𝑅 1 1
𝑣̈𝑐 (𝑡) + 𝐿 𝑣̇ (𝑡) + 𝐿𝐶 𝑣𝑐 (𝑡) = 𝐿𝐶 𝑣(𝑡)

𝑅 1 1
𝑻𝑳 [𝑣̈𝑐 (𝑡) + 𝐿 𝑣̇ (𝑡) + 𝐿𝐶 𝑣𝑐 (𝑡)] = 𝑻𝑳 [𝐿𝐶 𝑣(𝑡)]
𝑅 1 1
𝑻𝑳[𝑣̈𝑐 (𝑡)] + 𝑇𝐿 [ 𝐿 𝑣̇ (𝑡)] + 𝑻𝑳 [𝐿𝐶 𝑣𝑐 (𝑡)] = 𝑻𝑳 [𝐿𝐶 𝑣(𝑡)]
𝑅 1 1
𝑠 2 𝑉𝑐 (𝑠) + 𝑠𝑉𝑐 (𝑠) + 𝑉 (𝑠) = 𝑉(𝑠)
𝐿 𝐿𝐶 𝑐 𝐿𝐶

𝑅 1
(𝑠 2 + 𝐿 𝑠 + 𝐿𝐶) 𝑉𝑐 (𝑠) = 𝑉(𝑠)

Obtention de la fonction de transfert :

𝟏
𝑽𝒄 (𝑠) 𝑳𝑪 𝟏
𝐺(𝑠) = = 𝑹 𝟏 = (𝑳𝑪𝑠2 +𝑹𝑪𝑠+𝟏)
𝑉(𝑠) (𝑠2 + 𝑠+ )
𝑳 𝑳𝑪

Modèle d’états :

𝑹 𝟏 𝟏
𝒗̈ 𝒄 (𝒕) + 𝑳 𝒗̇ (𝒕) + 𝑳𝑪 𝒗𝒄 (𝒕) = 𝑳𝑪 𝒗(𝒕)

𝑹 𝟏 𝟏
𝒗̈ 𝒄 (𝒕) = − 𝑳 𝒗̇ (𝒕) − 𝑳𝑪 𝒗𝒄 (𝒕) + 𝑳𝑪 𝒗(𝒕)
𝑥(𝑡) = 𝒗𝒄 (𝑡): Tension aux bornes du condensateur
𝑥̇ (𝑡) = 𝒗̇ 𝒄 (𝑡):
𝑥̈ (𝑡) = 𝒗̈ 𝒄 (𝑡):

Le modèle masse-ressort amortisseur est un système de 2ième ordre avec un dégrée de liberté,
donc nous devons choisir deux variables d’états : 𝑥1 , 𝑥2

On pose :

𝑥1 = 𝑥
𝑥2 = 𝑥̇

Les dérivées des variables 𝑥1 , 𝑥2 donnent :

𝑥̇ 1 = 𝑥̇
𝑥̇ 2 = 𝑥̈

Ce qui donne en utilisant

𝑹 𝟏 𝟏
𝒗̈ 𝒄 (𝒕) = − 𝒗̇ (𝒕) − 𝒗𝒄 (𝒕) + 𝒗(𝒕)
𝑳 𝑳𝑪 𝑳𝑪

𝑹 1 𝟏
𝑥̈ = − 𝑳 𝑥̇ − 𝑳𝑪 𝑥 + 𝑳𝑪 𝑓

𝑥̇ 1 = 𝑥2
𝑹 1 𝟏
𝑥̇ 2 = − 𝑳 𝑥̇ − 𝑳𝑪 𝑥 + 𝑳𝑪 𝑓 (*)

Ce système de deux équations de 1ier ordre peut être écrit sous forme matricielle comme :
𝑥̇ = 𝐴𝑥 + 𝐵𝑢
𝑦 = 𝐶𝑥 + 𝐷𝑢 (**)
𝑥1
Ou 𝑥 = [𝑥 ]
2
A : matrice carrée de dimension 2
y : le vecteur de sortie du système

Ecrivons (*) sous forme de (**)

𝑥̇ 𝑥1
[ 1] = [ ] [𝑥 ] + [ ] 𝑓
𝑥̇ 2 2

La matrice vide correspond à la matrice A qui est en remplir. En utilisant la 1 ier équation
différentielle on a puisque 𝑥̇ 1 = 𝑥2 :
 𝑥̇ 0 1 𝑥1 0
[ 1] = [ ] [𝑥 ] + [ ] 𝑓
𝑥̇ 2 2

En utilisant la 2ième équation différentielle :

𝑹 1 𝟏
𝑥̇ 2 = − 𝑳 𝑥̇ − 𝑳𝑪 𝑥 + 𝑳𝑪 𝑓

𝑥̇ 0 1 𝑥1 0
[ 1] = [ ][ ]+ [ ]𝑓
𝑥̇ 2 −1/𝐿𝐶 −𝑅/𝐿 𝑥2 1/𝐿𝐶

En supposant que y=𝑥1 , on peut l’écrire sous forme matricielle comme :

𝑥
𝑦=[ ] [𝑥1 ]
2

𝑥1
𝑦 = [1 0] [𝑥 ] + 0𝑢
2

Le modèle d’états devient :

𝑥̇ 0 1 𝑥1 0
[ 1] = [ ] [𝑥 ] + [ ]𝑓
𝑥̇ 2 −1/𝐿𝐶 −𝑅/𝐿 2 1/𝐿𝐶
𝑥1
𝑦 = [1 0] [𝑥 ] + [0]𝑢
2

Où
0 1 0
𝐴=[ ] ; 𝐵=[ ]
−1/𝐿𝐶 −𝑅/𝐿 1/𝐿𝐶

𝐶 = [1 0] ; 𝐷 = 0

1. Modélisation des Systèmes Electromécanique

2. Modélisation des Systèmes Electromécanique

Maintenant, nous passons à des systèmes hybrides de variables électriques et mécaniques, les
systèmes électromécaniques. Les applications des systèmes à composants électromécaniques sont
les commandes de robots, les suiveurs solaires et satellitaires, et les commandes de position de
bandes et de disques d'ordinateur.
Un moteur est un composant électromécanique qui produit une sortie de déplacement pour une
entrée de tension, c'est-à-dire une sortie mécanique générée par une entrée électrique. Nous allons
dériver la fonction de transfert pour un type particulier de système électromécanique, le
servomoteur à courant continu commandé par l'induit (Mablekos, 1980).

2.1.Moteur à Courant Continu avec charge:

Dans la figure 2.35(a), un champ magnétique est développé par des aimants permanents fixes ou un
électroaimant fixe appelé champ fixe. Un circuit rotatif appelé l'induit, dans lequel le courant 𝑖𝑎 (𝑡),,
traverse ce champ magnétique à angle droit et ressent une force, 𝐹 = 𝐵𝑙𝑖𝑎 (𝑡), où B est l'intensité du
champ magnétique et 𝑙 la longueur du conducteur. Le couple qui en résulte fait tourner le rotor,
l'élément rotatif du moteur.

Un autre phénomène se produit au niveau du moteur : Un conducteur se déplaçant à angle droit par
rapport à un champ magnétique génère une tension aux bornes du conducteur égale à 𝑒 = 𝐵𝑙𝑣, où e
est la tension et v est la vitesse du conducteur normale au champ magnétique. Comme l'induit qui
transporte le courant tourne dans un champ magnétique, sa tension est proportionnelle à la vitesse.
Ainsi, la tension de l'induit est proportionnelle à la vitesse,

𝑑𝜃𝑚 (𝑡)
𝑣𝑏 (𝑡) = 𝐾𝑏
𝑑𝑡

Nous appelons 𝑣𝑏 (𝑡) la force électromotrice (f.é.m.) ; 𝑘𝑏 est une constante de proportionnalité
𝑑𝜃𝑚 (𝑡)
appelée constante de f.é.m. et 𝑑𝑡
= 𝜔(𝑡) est la vitesse angulaire du moteur. En prenant la
transformée de Laplace, on obtient :

𝑉𝑏 (𝑠) = 𝐾𝑏 𝑠𝜃𝑚 (𝑠)

La relation entre le courant d'induit, 𝑖𝑎 (𝑡),, la tension d'induit appliquée, 𝑒𝑎 (𝑡), et la f.é.m. a, 𝑣𝑏 (𝑡),
est trouvée en écrivant une équation en boucle autour du circuit d'induit transformé de Laplace (voir
figure 3.5(a)) :

𝑅𝑎 𝐼𝑎 (𝑠) + 𝐿𝑎 𝑠𝐼𝑎 (𝑠) + 𝑉𝑏 (𝑠) = 𝐸𝑎 (𝑠)

Le couple développé par le moteur est proportionnel au courant d'induit ; donc,

𝑇𝑚 (𝑠) = 𝐾𝑡 𝐼𝑎 (𝑠)

where Tm is the torque developed by the motor, and 𝐾𝑡 is a constant of proportionality, called the
motor torque constant, which depends on the motor and magnetic field characteristics. In a
consistent set of units, the value of 𝐾𝑡 is equal to the value of 𝐾𝑏 . Rearranging Eq. (2.147) yields

où 𝑇𝑚 est le couple développé par le moteur, et 𝐾𝑡 est une constante de proportionnalité, appelée
constante de couple du moteur, qui dépend des caractéristiques du moteur et du champ
magnétique. Dans un ensemble cohérent d'unités, la valeur de 𝐾𝑡 est égale à la valeur de 𝐾𝑏 . Le
réarrangement de l'équation (2.147) donne

𝐼𝑎 (𝑠) = 𝑇𝑚 (𝑠)/𝐾𝑡

Pour trouver la fonction de transfert du moteur, nous substituons d'abord les équations (2.145) et
(2.148) à (2.146), ce qui donne

(𝑅𝑎 + 𝐿𝑎 𝑠)𝑇𝑚 (𝑠)/𝐾𝑡 + 𝐾𝑏 𝑠𝜃𝑚 (𝑠) = 𝐸𝑎 (𝑠)

Nous devons maintenant trouver les termes 𝜃𝑚 (𝑠) de 𝑇𝑚 (𝑠) ; et nous devons séparer les variables
d'entrée et de sortie et obtenir la fonction de transfert, 𝜃𝑚 (𝑠)/𝐸𝑎 (𝑠).

La figure 2.36 montre une charge mécanique équivalente typique sur un moteur. 𝐽𝑚 est l'inertie
équivalente à l'induit et comprend à la fois l'inertie de l'induit et, comme nous le verrons plus loin,
l'inertie de la charge réfléchie sur l'induit.
𝐷𝑚 is the equivalent viscous damping at the armature and includes both the armature viscous
damping and, as we will see later, the load viscous damping reflected to the armature. From Figure
2.36,

En remplaçant l'équation (2.150) par l'équation (2.149)

(𝑅𝑎 +𝐿𝑎 𝑠)(𝐽𝑚 𝑠2 +𝐷𝑚 𝑠)𝜃𝑚 (𝑠)

+ 𝐾𝑏 𝑠𝜃𝑚 (𝑠) = 𝐸𝑎 (𝑠)
𝐾𝑡

Si nous supposons que l'inductance de l'induit, 𝐿𝑎 , est faible par rapport à la résistance de l'induit,
𝑅𝑎 , qui est habituelle pour un moteur à courant continu, Eq. (2,151) devient

𝑅
[𝐾𝑎 (𝐽𝑚 𝑠 + 𝐷𝑚 ) + 𝐾𝑏 ] 𝑠𝜃𝑚 (𝑠) = 𝐸𝑎 (𝑠)
𝑡

Après simplification, la fonction de transfert souhaitée, 𝜃𝑚 (𝑠)/𝐸𝑎 (𝑠), s'avère être

𝜃𝑚 (𝑠) 𝐾𝑡 /(𝑅𝑎 𝐽𝑚 )
=
𝐸𝑎 (𝑠) 𝑠 [𝑠 + 1 (𝐷 + 𝐾𝑡 𝐾𝑏 ) ]
𝐽𝑚 𝑚 𝑅𝑎

Même si la forme de l'équation (2.153) est relativement simple, à savoir

𝜃𝑚 (𝑠) 𝐾
=
𝐸𝑎 (𝑠) 𝑠(𝑠 + 𝛼 )

Le lecteur peut être préoccupé par la manière d'évaluer les constantes. Examinons d'abord les
constantes mécaniques, 𝐽𝑚 et 𝐷𝑚 .
Considérons la figure 2.37, qui montre un moteur avec une inertie 𝐽𝑎 et un amortissement 𝐷𝑎 au
niveau de l'induit entraînant une charge composée d'une inertie 𝐽𝐿 et d'un amortissement 𝐷𝐿 . En
supposant que toutes les valeurs d'inertie et d'amortissement indiquées sont connues, 𝐽𝐿 et
𝐷𝐿 peuvent être renvoyées à l'induit comme une inertie et un amortissement équivalents à ajouter à
𝐽𝑎 et 𝐷𝑎 , respectivement. Ainsi, l'inertie équivalente, 𝐽𝑚 , et l'amortissement équivalent, 𝐷𝑚 , à l'induit
sont

𝑁1 2
𝐽𝑚 = 𝐽𝑎 + 𝐽𝐿 ( )
𝑁2
𝑁1 2
𝐷𝑚 = 𝐷𝑎 + 𝐷𝐿 ( )
𝑁2

Maintenant que nous avons évalué les constantes mécaniques, 𝐽𝑚 et 𝐷𝑚 , qu'en est-il des constantes
électriques dans la fonction de transfert de Eq. (2.153) ? Nous montrerons que ces constantes
peuvent être obtenues par un test du moteur sur dynamomètre, où un dynamomètre mesure le
couple et la vitesse d'un moteur sous la condition d'une tension appliquée constante. Développons
d'abord les relations qui dictent l'utilisation d'un dynamomètre. En substituant les équations (2.145)
et (2.148) à l'équation (2.146), avec 𝐿𝑎 = 0, on obtient

𝑅𝑎
𝑇 (𝑠) + 𝐾𝑏 𝑠𝜃𝑚 (𝑠) = 𝐸𝑎 (𝑠)
𝐾𝑡 𝑚

En prenant la transformée de Laplace inverse, on obtient

𝑅𝑎
𝑇 (𝑡) + 𝐾𝑏 𝜔𝑚 (𝑡) = 𝑒𝑎 (𝑡)
𝐾𝑡 𝑚

If a dc voltage, 𝑒𝑎 , is applied, the motor will turn at a constant angular velocity, 𝑣𝑚 , with a constant
torque, 𝑇𝑚 . Hence, dropping the functional relationship based on time from Eq. (2.157), the
following relationship exists when the motor is operating at steady state with a dc voltage input:

Si une tension continue, 𝑒𝑎 , est appliquée, le moteur tournera à une vitesse angulaire constante, 𝑣𝑚 ,
avec un couple constant, 𝑇𝑚 . Par conséquent, si l'on laisse tomber la relation fonctionnelle basée sur
le temps à partir de l'équation (2.157), la relation suivante existe lorsque le moteur fonctionne en
régime permanent avec une entrée de tension continue :

𝑅𝑎
𝑇 + 𝐾𝑏 𝜔𝑚 = 𝑒𝑎
𝐾𝑡 𝑚

En résolvant pour 𝑇𝑚 , on obtient :

 𝐾𝑏 𝐾𝑡 𝜔𝑚 𝐾𝑡
𝑇𝑚 = − + 𝑒
𝑅𝑎 𝑅𝑎 𝑎

L'équation (2.159) est une ligne droite, 𝑇𝑚 par rapport au 𝑣𝑚 , et est illustrée à la figure 2.38.

Ce tracé est appelé courbe couple-vitesse. L'interception de l'axe du couple se produit lorsque la
vitesse angulaire atteint zéro. Cette valeur de couple est appelée le couple de décrochage, 𝑇𝑠𝑡𝑎𝑙𝑙 .
Ainsi,

𝐾𝑡
𝑇𝑠𝑡𝑎𝑙𝑙 = 𝑒
𝑅𝑎 𝑎

La vitesse angulaire qui se produit lorsque le couple est nul est appelée vitesse à vide, ω_(no-load).
Ainsi,

𝑒𝑎
𝜔no−load =
𝐾𝑏

Les constantes électriques de la fonction de transfert du moteur peuvent maintenant être trouvées
dans les équations (2.160) et (2.161) comme

𝐾𝑡 𝑇𝑠𝑡𝑎𝑙𝑙
=
𝑅𝑎 𝑒𝑎
Et
𝑒𝑎
𝐾𝑏 =
𝜔no−load

Les constantes électriques, 𝐾𝑡 /𝑅𝑎 et 𝐾𝑏 , peuvent être trouvées à partir d'un essai au dynamomètre
du moteur, qui donnerait 𝑇𝑠𝑡𝑎𝑙𝑙 et 𝜔no−load pour un 𝑒𝑎 donné.

2.2. Exemple:
Compte tenu du système et de la courbe couple-vitesse de la figure 2.39(a) et (b), trouvez la fonction
de transfert, 𝜃𝐿 (𝑠)/𝐸𝑎 (𝑠).

3. TIME RESPONSE
In this chapter, you will learn the following:

- How to find the time response from the transfer function;

- How to use poles and zeros to determine the response of control system;

- How to describe quantitatively the transient response of first and second order system.

3.1. Introduction

After the engineer abstains a mathematical representation of a system, the system is

analyzed for it transient and study state response to see if their characteristic.
3.2. Poles, zeros and system response

The output response of a system is the sum of two responses: the forced response and
the natural response. Although many techniques, such as solving a differential equation or
taking the inverse Laplace transform, enable us to evaluate this output response, these
techniques are laborious and time-consuming. Productivity is aided by analysis and design
techniques that yield result s in a minimum of time.

The concept of poles and zeros, fundamental to the analysis and design of control
systems, simplifies the evaluation o f a system's response.

3.2.1. Poles of a Transfer Function

The poles of a transfer function are:

1) The values of the Laplace transform variable , s, that cause the transfer function to become
infinite ;
2) Any roots of the denominator of the transfer function that is common to roots of the
numerator (transmission, poles and zeros) cancellation. In this cause, we will be interested
only with definition 1.
3) The roots of a characteristic polynomial of the denominator

3.2.2. Zeros of transfer functions

The zeros of a transfer function are:

1) The value s of the Laplace transform variables, that cause the transfer function to become
zero, or
2) Any root s of the numerator of the transfer function that is common to roots of the
denominator.

Strictly speaking, the zeros of a transfer function satisfy part (1) of this definition.

3.2.3. Poles and zeros of a first-order system

𝑠+2
Given the transfer function G(s)= 𝑠+5

𝑃𝑜𝑙𝑒: 𝑠 = −5, 𝑍𝑒𝑟𝑜: 𝑠 = −2

3.3. Unit step time response to a first order system

To show the properties of the pole and zeros, let us find the unit step response of the system.
 𝐶(𝑠)
𝐺(𝑠) = ⇒ 𝐶(𝑠) = 𝐺(𝑠) ⋅ 𝑅(𝑠)
𝑅(𝑠)
𝑠+2
⇒ 𝐶(𝑠) =
𝑠(𝑠 + 5)
2 3
𝐴 𝐵
⇒ 𝐶(𝑠) = + = 5+ 5
𝑠 𝑠+5 𝑠 𝑠+5
2 3
⇒ 𝐶(𝑡) = 𝑢(𝑡) + ⋅ 𝑒 −5𝑡 ⋅ 𝑢(𝑡)
5 5
2 3 −5𝑡
⇒ 𝐶(𝑡) = + ⋅ 𝑒
5 5
1) A pole of the input function generates the form of the forced response.
2) A pole of the transfer function generates the form of the natural response.
− t
3) A pole on the real axis generates an exponential response of the form e , where − is
the pole location on the real axis. Thus, the farther to the left a pole is on the negative
real axis, the faster the exponential transient response will decay to zero.
4) The zeros and poles generate the amplitudes for both the forced and natural response.

Résumé:
Time response 1st Order 2nd Order
Transfer Function 𝑎 𝜔𝑛 2
𝑠+𝑎 𝑠 2 + 2ζ𝜔𝑛 s + 𝜔𝑛 2
c(t) = 1 − e−at 𝑒 −𝜁𝜔𝑛 𝑡
Time response to a step 𝑐(𝑡) = 1 − cos(𝜔𝑛 √1 − 𝜁 2 𝑡 − 𝜑)
function: c(t) √1 − 𝜁 2
2.2 𝑇𝑟 =𝑇𝑟0.9 − 𝑇𝑟0.1
Rise time: Tr 𝑇𝑟 =
𝑎

--- −𝑙𝑛(%OS/100)
ζ=
Damping ratio: 𝜍 √𝜋2 +𝑙𝑛2 (%OS/100)

Peak time: Tp ----

𝜋
𝑇𝑝 =
𝜔𝑛 √1 − ζ2
Time constant: 𝜏 𝜏 = 1/𝑎 ---
𝜁𝜋
--- −
Percent overshoot : %OS √1−𝜁2
%OS= 𝑒 × 100

Settling time: Ts 4 𝑇𝑠 =
𝟒
𝑇𝑠 = 𝜻𝜔𝑛
𝑎

A first-order system without zeros can be described by the transfer function.

1
If the input is a unit step, where 𝑅(𝑠) = 𝑠 , the Laplace transform of the step response
is C(s), where
To sketch the response 𝑐(𝑡), we need three parameters:

▪ Time constant: 𝜏
1
We call 𝜏 = 𝑎 the time constant of the response. The time constant can be described as
the time for 𝑒 −𝛼𝑡 to decay to 37% of its initial value. Alternately, it is the time for 1- e−at to
rise to 63 % of its final value.
1
When 𝑡 = 𝜏 = 𝑎 we have:

The Time constant can be considered as a transient response specification for a first- order
system.

▪ Rise time: 𝑇𝑟

Rise time is defined as the time for the waveform to go from 0.1 to 0.9 of its final value.
Rise time is found by solving equation c(t)= 1- e−at for the difference in time at c(t) = 0. 9 and
c(t)= 0.1. Hence,

▪ Settling Time :𝑇𝑠

The Settling time is defined as the time for the response to reach, and stay within
tolerance =  n% of its final value.

At 𝑛% of the final value: 100% − 𝑛% = 1 − 𝑒 −𝑎𝑇𝑠

Example:
3
At 𝑇𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒 = 5% ⇒ 100% − 5% = 95% = 1 − 𝑒 −𝑎𝑇𝑠 ⇒ 𝑇𝑠 = 𝑎 = 3𝜏 (𝜏 = 1/𝑎)

4
At 𝑇𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒 = 2% ⇒ 100% − 2% = 98% = 1 − 𝑒 −𝑎𝑇𝑠 ⇒ 𝑇𝑠 = 𝑎 = 4𝜏 (𝜏 = 1/𝑎)
 Note: If the tolerance is not specified we will suppose T𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒 = 2%

3.4. Unit step time response to a first order system

Example

Write the output, 𝑐(𝑡), in general terms. Specify the forced and natural parts of the solution.

Solution:

By inverse Laplace transform we get:

 3.1. Introduction

Let us now extend the concepts of poles and zeros and transient response to second order
systems.

A second-order system can display characteristics much like a first-order system, or,
depending on component values, display damped or pure oscillations for its transient
response.

Let us consider this transfer function:

By assigning appropriate values to parameters a and b, we can show all possible second-order
transient responses.

3.2. General Second-Order System

Time response 1st Order 2nd Order
Transfer Function 𝑎 1 𝐾𝜔𝑛 2
=
𝑠 + 𝑎 𝑠/𝑎 + 1 𝑠 2 + 2ζ𝜔𝑛 s + 𝜔𝑛 2
1 𝐾
= =
1 + 𝜏𝑠 2ζ 1
1 + 𝜔 s + 𝜔 2 𝑠2
𝑛 𝑛
Time constant: 𝜏 𝜏 = 1/𝑎 ---
Time response to a step function: c(t) = 1 − e−at 𝑐(𝑡)
c(t) 𝑒 −𝜁𝜔𝑛 𝑡
=1− cos(𝜔𝑛 √1 − 𝜁 2 𝑡
√1 − 𝜁 2
− 𝜑)
with
ζ
𝜑 = tan−1 ( ),
√1 − ζ 2

4 𝑇𝑠 =𝜻𝜔
𝟒
Settling time: Ts 𝑇𝑠 = 𝑛
𝑎
2.2 𝑇𝑟 =𝑇𝑟0.9 − 𝑇𝑟0.1
Rise time: Tr 𝑇𝑟 =
𝑎
--- −𝑙𝑛(%OS/100)
ζ=
Damping ratio: ζ √𝜋 2 +𝑙𝑛2 (%OS/100)

Peak time: Tp ----

𝜋
𝑇𝑝 =
𝜔𝑛 √1 − ζ2
 𝜁𝜋
--- −
Percent overshoot : %OS √1−𝜁2
%OS= 𝑒 × 100

There are two quantitative specifications for characterizing second order system:

the natural frequency:𝜔𝑛 and the damping ratio:𝜁

▪ Natural Frequency : 𝜔𝑛

The natural frequency of a second-order system is the frequency of oscillation of the system
without damping.

▪ Damping ratio:
𝑏
Consider the general system 𝐺(𝑠) = 𝑠2+𝑎𝑠+𝑏

The poles of the system 𝑎𝑟𝑒 𝑠1,2 = −𝜎 ± 𝑗𝜔𝑑

𝑠1 = −𝜎 + 𝑗𝜔𝑑 , 𝑠2 = −𝜎 − 𝑗𝜔𝑑

(𝑠 − 𝑠2 )(𝑠 − 𝑠1 ) = [𝑠 − (−𝜎 − 𝑗ωd )][𝑠 − (−𝜎 + 𝑗ωd )]

(𝑠 − 𝑠2 )(𝑠 − 𝑠1 ) = 𝑠 2 + 2𝜎𝑠 + 𝜎 2 + 𝜔𝑑 2

𝜔𝑛2 = (𝜎 2 + 𝜔𝑑 2 ) ,𝜔𝑛 = √(𝜎 2 + 𝜔𝑑 2 )

(𝑠 − 𝑠2 )(𝑠 − 𝑠1 ) = 𝑠 2 + 2𝜎𝑠 + 𝜔𝑛2

𝐸𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙 𝑑𝑒𝑐𝑎𝑦 |𝜎|

𝜍= = ⇒ 𝜎 2 = 𝜍𝜔𝑛
𝑁𝑎𝑡𝑢𝑟𝑎𝑙 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝜔𝑛

𝜔𝑛2 = (𝜎 2 + 𝜔𝑑 2 ) ⇒ 𝜔𝑑2 = 𝜔𝑛 2 − 𝜎 2 ⇒ 𝜔𝑑 2 = 𝜔𝑛 2 (1 − 𝜍 2 )

𝑠 2 + 2𝜎𝑠 + 𝜎 2 + 𝜔𝑑 2 = 𝑠 2 + 2𝜍𝜔𝑛 𝑠 + 𝜔𝑛 2 𝜍 2 + 𝜔𝑛 2 (1 − 𝜍 2 )
 = 𝑠 2 + 2𝜍𝜔𝑛 𝑠 + 𝜔𝑛 2
𝑏 𝜔 2
G(s)=𝑠2 +𝑎𝑠+𝑏 = 𝑠2 +2𝜍𝜔𝑛 s+𝜔 2
𝑛 𝑛

𝑎
𝑎 = 2𝜍𝜔𝑛 𝜍 = 2√𝑏
By identification, we get{ 2 ⇒{
𝑏 = 𝜔𝑛 𝜔𝑛 = √𝑏

3.3. Classification of the poles with the damping 

s2+2𝜍𝜔𝑛 s+𝜔𝑛 2 =(s+𝜍𝜔𝑛 )2-𝜔𝑛 2 𝜍 2 + 𝜔𝑛 2 = 0

=(s+𝜍𝜔𝑛 )2+𝜔𝑛 2 (1 − 𝜍 2 )=0

let (𝑠 + 𝜍𝜔𝑛 )2 = 𝜔𝑛 2 (𝜍 2 − 1)

Poles location term of 𝝇

3.4. Step time response of the 2nd order systems

3.4.1. Derivation of the step time response : c(t)

𝐶(𝑆)
𝐺(𝑠) = 𝑅(𝑠) → 𝐶(𝑠) = 𝑅(𝑠)𝐺(𝑠)

1 𝜔 2
𝐶(𝑠)= 𝑠 𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔 2
𝑛 𝑛

𝜔 2 𝐾1 𝐾 𝑠+𝐾3
C(s) =𝑠(𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔 2) = 2
+ 𝑠2 +2𝜁𝜔 2
𝑛 𝑛 𝑠 𝑛 𝑠+𝜔𝑛

𝜔 2 𝜔 2
𝐾1 = 𝑠𝐶(𝑠) = 𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔 2
=𝜔𝑛2 = 1
𝑛 𝑛 𝑛

s=0 s=0

1 2 𝐾 𝑠+𝐾3 𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔𝑛 2 +𝐾2 𝑠+𝐾3 𝑆 (1+𝐾2 )𝑠2 +(2𝜁𝜔𝑛 +𝐾3 )𝑠+𝜔𝑛 2
C(s) = 𝑠 + 𝑠2 +2𝜁𝜔 2
= =
𝑛 𝑠+𝜔𝑛 𝑠(𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔𝑛 2 ) 𝑠(𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔𝑛 2 )

By identification

1+𝐾2 = 0 ⇒ 𝐾2 = −1

2𝜁𝜔𝑛 + 𝐾3 = 0 ⟹ 𝐾3 = −2𝜁𝜔𝑛
1 −𝑠−2𝜁𝑊𝑛 1 𝑠+2𝜁𝜔𝑛 1 𝑆+ζ𝜔𝑛 𝜁𝜔𝑛
Therefore C(s) =𝑠 + 𝑠2 +2𝜁𝜔 2 = 𝑆 − 𝑠2 +2𝜁𝜔 2 = 𝑠 − 𝑠2 +2𝜁𝜔 2 − 𝑠2 +2𝜁𝜔 2
𝑛 𝑠+𝜔𝑛 𝑛 𝑠+𝜔𝑛 𝑛 𝑠+𝜔𝑛 𝑛 𝑠+𝜔𝑛

1 𝑠+𝜁𝜔𝑛 𝜁 𝜔𝑛 √1−𝜁 2
C(s) = 𝑠 − (𝑠+𝜁𝜔 2 2 2
− (𝜁+𝜁𝜔𝑛 )2 +𝜔𝑛 2 (1−𝜁 2 )
𝑛 ) +𝜔𝑛 (1−𝜁 ) √1−𝜁 2

𝜁
c (t) =𝐿−1 (𝐶(𝑠)) = 1 − 𝑒 −𝜁𝜔𝑛𝑡 𝑐𝑜𝑠(𝜔𝑛 √1 − 𝜁 2 𝑡) − 𝑒 −𝜁𝜔𝑛𝑡 𝑠𝑖𝑛(𝜔𝑛 √1 − 𝜁 2 𝑡)
√1−𝜁 2

𝜁
Let = sin 𝜙; √1 − 𝜁 2 = 𝑐𝑜𝑠𝜙; 𝜙 = tan−1 ( )
√1−𝜁 2

since 𝑐𝑜𝑠 (𝑎 − 𝑏) = 𝑐𝑜𝑠𝑎𝑐𝑜𝑠𝑏 + 𝑠𝑖𝑛𝑎𝑠𝑖𝑛𝑏

𝑒 −𝜁𝜔𝑛 𝑡
c(t)=1 − cos(𝜔𝑛 √1 − 𝜁 2 𝑡 − 𝜑)
√1−𝜁 2
3.4.2. Definition of other parameters specification for an underdamped system (0< 𝜻 < 𝟏)

We have defined two parameters associated with 2nd order system: ζ et 𝜔𝑛 other
parameters associated with the underdamped response are rise time, peak time , percent
overshoot and setting time.

Rise time Tr

The time required for waveform to go from 0.1 to the value 0.9 of the final value.

Peak time Tp

The time required to reach the first or maximum peak

Percent overshoot %OS

The amount that the waveform overshoots the steady-state, or final , value at the peak time ,
expressed as a percentage of the steady-state value.

Settling timeTs

The time required for the transient's damped oscillation to reach and stay within ±n % of the
steady-state value.

These information can help a designer determine if the speed and the nature of the
response do or do not degrade the performance of the system. For example, the speed of an
entire computer system depends on the time it takes for a hard drive head to reach steady state
and read data; passenger comfort depends in part on the suspension system of a car an d the
number of oscillations it goes through after hitting a bump.

3.4.3. Evaluation of the parameters

▪ Evaluation of Peak time Tp
 𝜔 2
𝑛
C(s) =𝑠(𝑠2+2𝜁𝑊𝑛𝑠+𝜔 2)
assuming zero initial condition
𝑛

𝜔 2
L (𝑐̇ (t)) =sC(s) =𝑠2 +2𝜁𝜔𝑛 𝑠+𝜔 2
𝑛 𝑛

𝜔𝑛
𝜔𝑛 √1−ζ2
√1−ζ2 𝜔𝑛 𝜔𝑛 √1−ζ2
L (ċ (t)) =sC(s) =(s+ζ𝜔 2 2 (1−ζ2 )
=
𝑛 ) +𝜔𝑛 √1−ζ2 (s+ζ𝜔𝑛 )2 +𝜔𝑛 2 (1−ζ2 )

𝜔2
But L{𝑠𝑖𝑛𝜔𝑡} = 𝑠2 +𝜔2

𝜔𝑛 𝜔𝑛 √1−ζ2 𝜔𝑛 𝜔𝑛 √1−ζ2
ċ (t) = L−1 { } = L−1 {(s+ζ𝜔 }
√1−ζ2 (s+ζ𝜔𝑛 )2 +𝜔𝑛 2 (1−ζ2 ) √1−ζ2 2 2 2
𝑛 ) +𝜔𝑛 (1−ζ )

𝜔2 𝜔2
But 𝑠𝑖𝑛𝜔𝑡 = 𝐿−1 {𝑠2 +𝜔2} and 𝑒 −𝑎𝑡 𝑠𝑖𝑛𝜔𝑡 = 𝐿−1 {(𝑠+𝑎)+𝜔2}

ζ𝜔𝑛
Therefore: ċ (t) = e−ζ𝜔𝑛t sin(𝜔𝑛 √1 − ζ2 t)
√1−𝜁 2

For max 𝑐̇ (t) =0, sin(𝜔𝑛 √1 − ζ2 t) = 0, 𝜔𝑛 √1 − 𝜁 2 𝑡 = 𝑛𝜋; 𝑛 = 0; 1; 2;……….

nπ
𝑡=
𝜔𝑛 √1−ζ2

Letting :

1) 𝑛 = 0 ⇒ 𝑡 = 0 𝑡ℎ𝑒 𝑓𝑖𝑟𝑠𝑡 𝑝𝑜𝑖𝑛𝑡 𝑜𝑛 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒. ( the origin)

𝜋
2) n=1⇒ 𝑡 = 𝑇𝑝 The first peak which occurs at the peak time, 𝑇𝑝 , 𝑇𝑝 =
𝜔𝑛 √1−𝜁 2
2π
3) n=2⇒ 𝑡 = 𝑇2𝑝 The 2nd peak which occurs at the peak time, 𝑇2𝑝 , 𝑇2𝑝 =
𝜔𝑛 √1−ζ2

▪ Evaluation of %OS

𝐶𝑚𝑎𝑥 −𝐶𝑓𝑖𝑛𝑎𝑙
The percent overshoot %OS is given by %𝑂𝑆 = × 100
𝐶𝑓𝑖𝑛𝑎𝑙

𝜁
c(t)=1-𝑒 −𝜁𝜔𝑛𝑡 (cos 𝜔𝑛 √1 − 𝜁 2 𝑡 + sin 𝜔𝑛 √1 − 𝜁 2 𝑡
√1−𝜁 2

𝜁𝜋
−
√1−𝜁2 𝜁
𝑐𝑚𝑎𝑥 =c (𝑇𝑝 )=1-𝑒 (cos 𝜋 + sin 𝜋)
√1−𝜁 2

𝜁𝜋
−
√1−𝜉2
𝑐𝑚𝑎𝑥 = 1 + 𝑒
 𝜁𝜋
−
√1−𝜁2
%OS= 𝑒

▪ Evaluation of 𝑻𝒔

In order to find the settling time, we must find the time for which c(t) reaches and stays within
±n% of the steady-state value, that amplitude of c(t) at the 𝑇𝑠 .

1
c(t)=1 − e−ζ𝜔𝑛t cos( 𝜔𝑛 √1 − ζ2 t − φ) at 𝑇𝑠
√1−ζ2

e−ζ𝜔𝑛𝑇𝑠 e−ζ𝜔𝑛 𝑇𝑠
100% − n% = 1 − ⟹ n% =
√1−ζ2 √1−ζ2

Solving that we have

𝑒 −𝜁𝜔𝑛𝑇𝑠 = 𝑛%√1 − 𝜁 2 − 𝜁𝜔𝑛 𝑇𝑠 = ln (𝑛%√1 − 𝜁 2 )

−𝐥𝐧 (𝒏%√𝟏−𝜻𝟐 ) 𝟒
𝑇𝑠 = , 𝑇𝑠 =𝜻𝜔
𝜻𝜔𝑛 𝑛

▪ Evaluation of 𝑻𝒓

A precise analytical relationship between rise time and damping ratio, 𝜁, cannot be found .
However, using a computer the rise time can be found
1
c (t)=1 − e−ζ𝜔𝑛t cos( 𝜔𝑛 √1 − ζ2 t − ϕ)
√1−ζ2

ζ
ϕ = tan−1 ( ),
√1 − ζ2

𝑇𝑟 =𝑇𝑟0.9 − 𝑇𝑟0.1

Where:

1
0.1 − e−ζ𝜔𝑛t cos( 𝜔𝑛 √1 − ζ2 𝑇𝑟0.1 − ϕ)
√1 − ζ2

1
0.9 − e−ζ𝜔𝑛t cos( 𝜔𝑛 √1 − ζ2 𝑇𝑟0.9 − ϕ)
√1 − ζ2

Example 1
100
Let a 2nd order transfer function: G(s) =𝑠2 +15𝑠+100

Find a) 𝑇𝑝 ; 𝑏) %OS; c) 𝑇𝑠 and d) sketch the time response

Solution:
 a)
𝜔𝑛 2 = 100 ⟹ 𝜔𝑛 = 10
15 15
2𝜁𝜔𝑛 = 15 ⟹ 15 ⟹ 𝜁 = 2𝜔 = 20 = 0.75.
𝑛

𝜋
𝑇𝑝 = = 0.475
10√1 − 0.752

b)
2
%𝑂𝑆 = 𝑒 −𝜋𝜁√1−𝜁 = 0.02837 = 2.837%

c)

0.05√1 − 0.752
𝑇𝑠 5% = − ln ( ) = 0.45𝑠
0.75 × 10

0.02√1 − 0.752
𝑇𝑠2% = − ln ( ) = 0.57𝑠
0.75 × 10

d)

Example 2

Given the system shown

Given the system shown in Figure, find J and D to yield 20% overshoot and a settling
time of 2 seconds for a step input of torque T(t).

Solution:

First, the transfer function for the system is

,

3.4.4. Relating peak time ,%𝑂𝑆, 𝑇𝑠 𝑡𝑜 𝜔𝑛 𝑎𝑛𝑑 𝜁

𝑠1,2 = −𝜔𝑛 𝜁 ± 𝑗𝜔𝑛 √1 − 𝜁 2 = −𝜎𝑑 ± 𝑗𝜔𝑑

Pythagore:

√(−𝜁𝜔𝑛 )2 + 𝜔𝑛 2 (1 − 𝜁 2 ) = √𝜁 2 𝜔𝑛 2 + 𝜔𝑛 2 − 𝜁 2 𝜔𝑛 2 = √𝜔𝑛 2 = 𝜔𝑛

cos 𝜃 = 𝜁
𝜋 𝜋
𝑇𝑝 = =𝜔 ,
√1−𝜁 2 𝜔𝑛 𝑑

−ln (%𝑛√1 − 𝜁 2 ) −ln (%𝑛√1 − 𝜁 2 )

𝑇𝑠 = =
𝜁𝜔𝑛 𝜎𝑑

• 𝜔𝑑 and 𝜎𝑑 𝑎𝑟𝑒 𝑡ℎ𝑒 𝑚𝑎𝑔𝑛𝑖𝑡𝑢𝑑𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑖𝑚𝑎𝑔𝑖𝑛𝑎𝑟𝑦 𝑎𝑛𝑑 𝑡ℎ𝑒 𝑟𝑒𝑎𝑙 𝑝𝑎𝑟𝑡 𝑎𝑓 𝑡ℎ𝑒 𝑝𝑜𝑙𝑒𝑠.

𝜔𝑑 : damped frequency of oscillation

𝜎𝑑 :Exponential frequency of oscillation

𝑇𝑝 is inversely proportional to 𝜔𝑑 .

Horizontal line

Horizontal line is line of constant imaginary value there are also of constant peak.

j𝜔𝑛 √1 − 𝜁 2 =j𝜔𝑑
 -j𝜔𝑛 √1 − 𝜁 2 = −𝑗𝜔𝑑

Vertical line

−l n(𝑛%√1−𝜁 2 )
𝑇𝑠 = Inversely proportional to 𝜎𝑑
𝜎𝑑

𝑠1 j𝜔𝑑

𝑠2 -j𝜔𝑑

Radial line

The radial line cos θ = ζ is the time of constant since %OS is only function of ζ

j𝜔𝑑

−𝑗𝜔𝑑

4.7- System response with additional poles

If a system has more than two poles or has zeros. We cannot use the formula to calculate
the performance specifications that we derived for less than or equal to two poles and without
zeros. However, under certain conditions a system with more than two poles or with zeros can
be approximated as a second order system. Let
us consider a 3rd order system with poles −σd + j𝜔𝑑 = −ζ𝜔𝑛 ±
j𝜔𝑛 √1 − ζ2 and the real pole at − αr .

1
𝐺(𝑠) =
(𝑠 − (−𝜁𝜔𝑛 − 𝑗𝜔𝑑 ))(𝑠 − (−𝜁𝜔𝑛 + 𝑗𝜔𝑑 )(𝑠 + 𝛼𝑟 )
1
=
(𝑠 + 𝜁𝜔𝑛 + 𝑗𝜔𝑑 )(𝑠 + 𝜁𝜔𝑛 − 𝑗𝜔𝑑 )(𝑠 + 𝛼𝑟 )
1
Step response by partial fraction expansion of 𝑆 𝐺(𝑠)
Three cases :

𝛼𝑟 ≈ 𝜁 𝜔𝑛 𝛼𝑟 ≫ 𝜁 𝜔𝑛 𝛼𝑟 ⟶ +∞

Conditions for approximation 𝐷𝑒 −𝛼𝑟 𝑡

1- 𝛼𝑟 ≫ 𝜁 𝜔𝑛 or amplitude (residue) of 𝛼𝑟 becomes zero as 𝛼𝑟 ⟶ +∞

Example:
Determine the validity of the 2nd order system approximation for each of the two TF
700 360
a) 𝐺(𝑠) = b) 𝐺(𝑠) =
(𝑆+15)(𝑆 2 +4𝑠+100) (𝑆+4)(𝑆 2 +2𝑆+90)

700
a) G(s)= (𝑆+15)(𝑆 2 +4𝑆+100)

700 700 700

C(s)= 𝑆(𝑆+15)(𝑆 2 +4𝑆+100) = 𝑆(𝑆+15)(𝑆+2)2 −4+100 = 𝑠(𝑠+15)[(𝑠+2)2 +96]

𝐴 𝐵(𝑆+2)+96 𝐷
C(s)= 𝑆 + (𝑠+2)2 +96
+ 𝑆+15 ; 𝜁𝑤𝑛 = 1; 15 ≫ 2 the 2nd order approximation is valid

360 360
b) G(s)= (𝑆+4)(𝑆 2 +2𝑆+90 = (𝑆+4)[(𝑆+1)2 −1+100] ;

𝜁𝑤𝑛 = 1; 4 ≅ 1 ; 2nd order approximation is not valid

360 360
𝐷 = lim =
𝑠⟶4 𝑠(𝑠 2 + 2𝑠 + 90) −312
Chapter 4: Reduction of multiple subsystems

Objective and introduction

In this chapter we will learn the following:

- How to reduce bloc diagram of multiple subsystem to a single block representing

the transfer function from input to output.

We have been working with the individual subsystem represented by a block diagram with
its inputs and outputs.

More complicated systems however are represented by the interconnection of many

subsystems.

Since the response of a single transfer function can be calculated we want to represent
multiple subsystems as a single transfer function we can then apply the analytical of the
previous chapter and obtain the transient response information about the entire system block
diagram algebra will be used to reduce block diagram.

4.1. Block diagram

System is represented as a block diagram with an input, and output and a transfer function.
 To interconnect multiple subsystems we need new elements, summing junctions and pick
off point.

Figure 1: Summing junction

Figure 2: Pick off point

4.2. Reduction to a single block diagram

▪ Cascade form or series forms

Figure 3: Cascadedsubsystems

Figure 4: Equivalent transfer function

▪ Parallelform
 Figure 5: Parrallel system

Figure 6: Equivalent transferfunction

▪ Feedback

The feedback form the basis for study of control system engineering

Figure 7: Feedback control system

Figure 8: Equivalent transfer function

Proof

𝐸(𝑠) = 𝑅(𝑠) − 𝐶(𝑠) ⋅ 𝐻(𝑠) (1)

Since 𝐶(𝑠) = 𝐸(𝑠) ⋅ 𝐺(𝑠) (2)

 Substitute (2) in (1)

𝐸(𝑠) = 𝑅(𝑠) − 𝐸(𝑠) ⋅ 𝐺(𝑠) ⋅ 𝐻(𝑠)

𝐸(𝑠) ⋅ (1 + 𝐺(𝑠) ⋅ 𝐻(𝑠)) = 𝑅(𝑠)
𝐸(𝑠) 1
=
𝑅(𝑠) 1 + 𝐺(𝑠) ⋅ 𝑅(𝑠)

𝐶(𝑠) 1
From (2) =
𝐺(𝑠)⋅𝑅(𝑠) 1+𝐺(𝑠)⋅𝐻(𝑠)

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

The product 𝐺(𝑠) ⋅ 𝐻(𝑠) is called the open loop transfer function

𝐺(𝑠)
1+𝐺(𝑠)⋅𝐻(𝑠)
is called the close loop transfer function

1
1+𝐺(𝑠)⋅𝐻(𝑠)
is called error transfer function

4.3. Moving block to create familiar form

In this subsection we will discuss block move that can be made in order to establish familiar
form (series, parallel and feedback) when they almost exist.

In particular we will explain how to move block left and right past summing junctions and
pick off points.
Example: Reduce the block diagram to a single transfer function

Figure 9: Block diagram for example

Resolution

We solve the problem following the steps in the figures below

Figure 9. 1: Collapse summingjunction

4.4. Moving block to create familiar form

4.4.1. Block diagram algebra for summing junction

▪ Equivalent forms for moving a block to the left past a summing junction

▪ Equivalent forms for moving a block to the right past a summing junction
 4.4.2. Block diagram algebra for pickoff point
▪ Equivalent forms for moving a block to the left past a pickoff point

▪ Equivalent forms for moving a block to the right past a pickoff point

Chapter 5: Stability
5.1. INTRODUCTION
For on linear and time varying systems, the study of stability is a complex and
often difficulty subject.

In this section we will consider only linear-time invariant, LTI (constant

parameters) systems for which we have the following condition of stability.
a) Condition of stability
A linear-time invariant is said to be stable if all the roots of the
denominator polynomial of the transfer function have negative real parts (i.e.
they are all in the left half s-plane) other wise the system is unstable response.
b) Generality
Consider the LTI system where denominator’s polynomial of the transfer
function leads the characteristic equation.

𝑆𝑛 + 𝑎𝑛−1 𝑆𝑛−1 + 𝑎𝑛−2 𝑆𝑛−2 + ⋯ … + ⋯ . . +𝑎0 (∗)

Assume that the roots{Pi} of the characteristic equation are real or

complex, but are distinct.

Equation (*) shows up as the denominator in the transfer function for the
system as follows before any cancellation of poles by zeros is made.
C(S) 𝑆𝑛 +𝑏𝑛−1 𝑆𝑛−1 + 𝑏𝑛−2 𝑆𝑛−2 +⋯…+⋯..+𝑏0
G(s) = =
R(S) 𝑆𝑛 + 𝑎𝑛−1 𝑆𝑛−1 + 𝑎𝑛−2 𝑆𝑛−2 +⋯…+⋯..+𝑎0

C(s) K ∏m
i=1(s−zi )
G(s) = = ∏m
R(s) i=1(s−pi )

The solution of the differential equation may be written rising partial

fraction expansion.

c(t) = ∑ K i × e𝑝𝑖 t (**)

Where {𝑝𝑖 } are the roots of the characteristic equation and {K i } depend on the
initial condition and zero location.

The system is stable iff (necessary and sufficient every term in the
equation (**) goes to zero as t→ ∞. This happen if all the poles of the system
are strictly in the left-hand plan when Re {𝑝𝑖 } < 0.
 If any poles are repeated, the response must be changed from that of
equation (**) by including a polynomial in t in place of K i , but the conclusion is
the same. This is called internal stability.

If the system has any poles in the right plane, it is unstable. Hence jw-axis
is the stability boundary between asymptotically stable and unstable.

LHP RHP

Example:
7 7
+ R(S) C(S)
𝑆(𝑆 + 1)(𝑆 + 2) 𝑆 3 + 3𝑆 2 + 2𝑆 + 7

s1 = -9.16 – 1.04i

s2 = -0.16 +1.04i The system is stable

s3 = -2.071

3 3
+ R(S) C(S)
𝑆(𝑆 + 1)(𝑆 + 2) 𝑆3 + 3𝑆 2+ 2𝑆 + 3

-
s1 = 0.04+ 1.5i

s1 =0.04 +1.5i The system is unstable

s1 = -3.08

I. ROUTH-HURWITZ CRITERION

A necessary condition (but not sufficient) for stability is that all the coefficients
are missing (are zero) or negative, the system will have poles located outside the
LHP.

Once the elementary conditions have been satisfied, we need more powerful
test or criterion called Routh-Hurwitz criterion.

In this section, we will learn a method that yields stability information without
the need to solve for closed-loop poles.

Using this method, we can tell how many closed-loop poles are in the LHP, in
the RHP and on jw-axis.

The method required two steps:

1) Generate a table called Routh table

2) Interpret the table

Example: Let consider the system below assuming ai> 0.

𝑁(𝑆)
𝑎4 𝑆4 + 𝑎3 𝑆3 + 𝑎2 𝑆 2 + 𝑎1 𝑆1 + 𝑎0
 𝑠4 𝑎4 𝑎2 𝑎0 𝑎4 𝑎2
|𝑎 𝑎1 |
3
𝑠3 𝑎3 𝑎1 0 𝑏1 =
𝑎3
𝑠2 𝑏1 𝑏2 0 𝑎4 𝑎0
|𝑎 0|
3
𝑠1 𝑐1 𝑐2 0 𝑏2 = −
𝑎3
𝑠0 𝑏2 0 0 𝑎3 𝑎1 𝑎3 0
|𝑏 | |
1 𝑏2 | 𝑏1 0
𝑐1 = − 𝑐2 = −
𝑎3 𝑏1

Interpretation

1. The number of roots or the polynomial that are in the RHP is equal to the
number of sign change in the 1st column
2. A system is stable iff all the elements in the 1st column of the table are
positive (No sign change).

Note:

➢ A pattern of +,-,+ is counted as two sign changes

➢ One sign change is + to – or – to +.

Example 1: Find the stability

 3 + 7
+
𝑠(𝑠 + 1)(𝑠 + 2)
𝑠(𝑠 + 1)(𝑠 + 2)
- -

3 3
≡ 𝑠3 + 3𝑠 2+ 2𝑠 + 3 ≡ 𝑠3 + 3𝑠 2 + 2𝑠 + 3

P(s) = 𝑠 3 + 3𝑠 2 + 2𝑠 + 3 P(S) = 𝑠 3 + 3𝑠 2 + 2𝑠 + 7

𝑠3 1 2 0 𝑠3 1 2 0

𝑠2 3 3 0 𝑠2 3 7 0

𝑠1 1 0 0 𝑠1 −1 0 0
3
𝑠0 3 0 0 𝑠0 7 0 0

No sign change in the 1st column, so the Two sign change so the system is no stable.
system is stable.

Example 2: Make a Routh table and tell how many root of the following
polynomial are in the RHP and in the LHP.

P(s) = 3𝑠 7 + 9𝑠 6 + 6𝑠 5 + 4𝑠 4 + 7𝑠 3 + 8𝑠 2 + 2𝑠 + 6
 𝑠7 3 6 7 2 0

𝑠6 9 4 8 6 0

𝑠5 14 13 0 0 0
3 3
𝑠4 61 8 6 0 0
−
14
𝑠3 787 392 0 0 0
61 61
𝑠2 8004 6 0 0 0
787
𝑠1 1581 0 0 0 0
−
1334
𝑠0 6 0 0 0 0

II. ROUTH-HURWITZ CRITERION: SPECIAL CASES

Two special cases can occur:

1) The Routh-Hurwitz table sometimes will have a zero only in the

first column of a row
2) The Routh table sometimes will have entire row that consist of zero

ZERO ONLY ON THE FIRST COLUMN OF A ROW

If the 1st element of a row is zero, division by zero would be required to

form the next row. To avoid this phenomenon an epsilon 𝜀 is assigned to
replace the zero in the first column.

The value of 𝜀 is then allowed to approach zero from either positive or

negative side after which the signs of the entries in the first column be
determinate.

Example: Determinate the stability of the closed-loop transfer bellow

10
G(S) =
𝑆5 +2𝑆 4 +3𝑆 3 + 6𝑆 2 +5𝑆+3
𝑠5 1 3 5 0

𝑠4 2 6 3 0

𝑠3 0 7/2 0 0

𝑠2

𝑠1

𝑠0

𝑠5 1 3 5 0

𝑠4 2 6 3 0

𝑠3 𝜀 7/2 0 0

𝑠2 6𝜀 − 7 3 0 0
𝜀
𝑠1 𝑑1 0 0 0

𝑠0 3 0 0 0

2 6
−| | (7 − 6𝜀) 6𝜀 − 7
𝜀 7/2
𝑐1 = =− =
𝜀 𝜀 𝜀

𝑐2 = 3

𝑠5 1 3 5 0

𝑠4 2 6 3 0

𝑠3 𝜀 7/2 0 0

𝑠2 6𝜀 − 7 3 0 0
𝜀
𝑠1 6𝜀 2 − 42𝜀 + 49 0 0 0
14 − 12𝜀
 𝑠0 3 0 0 0 +

𝜀 7/2
− |6𝜀−7 | 42𝜀−49 6𝜀2 −42𝜀+49
3 (3𝜀 − ) 6𝜀2 − 42𝜀 + 49
𝜀 2𝜀 2
𝑑1 = 6𝜀−7 =− 6𝜀−7 =− 6𝜀−7 =−
12𝜀 − 14
𝜀 𝜀 1
6𝜀2 − 42𝜀 + 49
=
14 − 12𝜀

𝑠5 1 3 5 0 +

𝑠4 2 6 3 0 +

𝑠3 0+ 7/2 0 0 +

𝑠2 −∞ 3 0 0 -

𝑠1 7 0 0 0 +
2
𝑠0 3 0 0 0 +

6𝜀 − 7 7
lim+ = − =−∞
𝑥→0 𝜀 0

6𝜀2 − 42𝜀 + 49 49 7
lim+𝑑1 = lim+ = =
𝑥→0 𝑥→0 14 − 12𝜀 14 2

ENTIRE ROW OF ZERO

Example:
10
T(S) =
𝑆 5 +7𝑆 4 +6𝑆 3 + 42𝑆 2 +8𝑆+56
 𝑆5 1 6 8

𝑆4 7 42 56

𝑆3 0 0 0

𝑆2

𝑆1

𝑆0

PROCEDURE

1. Verify that there is entire row of zero at row I

2. Write the purely even or purely odd polynomial P(s)
𝑑
3. Take the derivative of row I by the coefficient of P(s): 𝑃(𝑠)
𝑑(𝑠)
𝑑
4. Replace the zero coefficient of row i by the coefficient of 𝑃(𝑠)
𝑑(𝑠)
5. Continue filling the table and deduce the stability

P(s) = 7𝑠 4 + 42𝑠 2
𝑑
𝑃(𝑠) = 28𝑠 3 + 84𝑠
𝑑𝑥

𝑆5 1 6 8

𝑆4 7 42 56

𝑆3
28 84 0
𝑆 2 21 56 0

𝑆 1 28
3
𝑆 0 56

7 42
| | (7 × 84 − 42 × 28) (588 − 1176)
𝑐1 = − 28 84 =− =− = 21
28 28 28
7 56
| |
𝑐2 = − 28 0 = 56
28

28 84
| | (28 × 56 − 84 × 21) (1568 − 1764) 196
𝑑1 = − 21 56 = − =− =
21 21 21 21
28 × 7 28
= =
3×7 3

21 56
| 28 0 |
3
𝑒1 = − 28 = 56
3
 𝑆5 1 6 8

𝑆4 7 42 58
1 6 8
𝑆3 0 0
4 12
𝑆2 3 8

𝑆1 4/3

𝑆0 8

P(s) = 𝑠 4 + 6𝑠 2
𝑑
𝑃(𝑠) = 4𝑠 3 + 12𝑠
𝑑𝑥

1 6
| |
𝑐1 = − 4 12 = − (12 − 24) = 12 = 3
4 4 4
1 8
| |
𝑐2 = − 4 0 = 32 = 8
4 4

4 12
| |
𝑑1 = − 3 8 = − (32 − 36) = 4
3 3 3
3 8
| 4 0|
3
𝑒1 = − 4 =8
3

The system is marginally stable’ cause there is no sign change

 An entire row of zero will appear in a routh table when purely even or
purely odd polynomial as a factor the original polynomial.

(𝑠 4 + 6𝑠 2 + 8)(𝑠 + 𝑎) = 𝑠 5 + 𝑎𝑠 4 + 6𝑠 3 + 6𝑎𝑠 2 + 8𝑠 + 8𝑎

By identification

𝑠 5 + 7𝑠 4 + 6𝑠 3 + 42𝑠 2 + 8𝑠 + 56

a=7

STABILITY VERSUS PARAMETER RANGER

Routh table is also useful in determinate the range of parameters for which a
feedback system remains stable.

Example1: Find range for which this system is stable.

𝐾 𝐾
+ ≡
𝑠(𝑠 + 1)(𝑠 + 2) 𝑠 3 + 3𝑠 2 + 2𝑠 + 𝐾

𝑠3 1 2

𝑠2 3 K

𝑠1 6−𝐾 0
3
𝑠0 K 0

Since each element of the first column needs to be positive for the system to be stable, then:
 6−𝐾
>0 ⇒𝐾<6
{ 3
𝐾>0
⇒0<𝐾<6

If K= 6 ⇒ the system is marginally stable

Example:
𝐾(𝑆+20)
𝐾(𝑠 + 20) 𝑆(𝑆+2)(𝑆+3)
+
𝑠(𝑠 + 2)(𝑠 + 3) ≡ 𝐾(𝑆+20)
1 + 𝑆(𝑆+2)(𝑆+3)
-

P(s) = 𝑠 3 + 5𝑠 2 + (6 + 𝐾)𝑠 + 20𝐾

𝑠3 1 6+K

𝑠2 5 20K

𝑠1 6-3K

𝑠0 20K

The system is stable iff 6-3K >0 and 20K>0

⇒0<K<2

Example 2: For the closed loop system

20
𝑃(𝑠) =
𝑠 8 + 𝑠 7 + 12𝑠 6 + 22𝑠 5 + 39𝑠 4 + 59𝑠 3 + 48𝑠 2 + 38𝑠 + 20
Tell how many poles are in the LHP and on the jw-axis
CHAPTER 6 :STEADY-STATE ERRORS
Objectives

In this chapter we will learn the following:

• How to find the steady-state error.

• How to specify a system’s steady-state error.
• How to find steady-state for disturbance inputs.
• How to design system parameters to meet steady-state error
performance specification.

6.1. Introduction

Definition and test inputs

Steady-state error is the difference between the input and the output for a
prescribed test input as t ∞.
Let us assume a position control system (a satellite position or an antenna
position). Where the output position follows the input commanded position.
Step inputs represent constant position and thus are useful in determining
the ability of a control system to position itself with respect to a stationary
target such as a satellite in geostationary orbit.

Ramp inputs represent constant velocity inputs to a position control system by

their linearly increasing amplitudes; therefore can be used to text a system’s
ability to follow a linearly increasing input or equivalently to track a constant
velocity target. For example a position control system that tracks a satellite
that moves across the sky at a constant angular velocity, would be tested with
a ramp input.
Finally paraboles, whose second derivative are constant, represent constant
acceleration input to position control system and can be used to represent
acceleration targets such as to determine the steady-state error of an
accelerating.

Application to stable systems

Our discussion of steady-state is limited to state where the natural response

approaches zero as t ∞
Unstable systems represent loss of control in the steady-state and are
acceptable for use at all.

Evaluating steady-state errors for an input

Evaluating steady-state error with a ramp input

7.2- Steady-state error for (unity) feed bock in terms of the open loop
transfer function

The error E(s) can be written E(s) =R(s)-C(s) (1)

But C(s) =E(s).G(s) (2)

(2) in (1) E(s) =R(s)-E(s) G(s) E(s) + E(s) G(s) =R(s)

(1+G(s)).E(s) =R(s)
E(s) 1
=
𝑅(𝑠) 1+𝐺(𝑠)

𝑅(𝑠)
E(s) =
1+𝐺(𝑠)

Applying the final value theorem

e (∞) = lim 𝑒(𝑡) = lim 𝑠𝐸(𝑠)
𝑡→∞ 𝑠→0

𝑠.𝑅(𝑆)
e (∞) =lim
𝑠→0 1+𝐺(𝑆)

1
Steady-state error for a step input R(S)=
𝑆
1
𝑆(𝑆) 1
e (∞) =estep(∞) =lim =lim 1+𝐺(𝑆)
𝑆→0 1+𝐺(𝑆) 𝑆→0

1
e (∞) =
1+ lim 𝐺(𝑠)
𝑆→0

lim G(s) =G (0) is called the dcgain: the value of the forward transfer function
𝑠→0
at zero frequency (s=0).

In order to have zero steady- state error the dcgain ought to infinite (G(0)=
∞ ) with n≥1

(𝑠+𝑧1 )(𝑠 +𝑧2 )…

( G(s) = ) that is at least one pole must be at the origin
𝑠𝑛( 𝑠+𝑝1 )(𝑠+𝑝2 )…
1
(one integrator must be present in the forward loop).
𝑆

Steady-state error of a ramp input

1
𝑠( 2 ) 1
e (∞) =eramp (∞) =lim
1+𝐺(𝑠)
𝑆
= lim 𝑠+𝑠𝐺(𝑠)
𝑠→0 𝑠→0

1
e (∞) =
lim s𝐺(𝑠)
𝑆→0

In order to have zero steady-state error for a ramp input, we must have

lim 𝑠𝐺(𝑠) = ∞
𝑠→0

Since G(s) is of the form:

(𝑠+𝑧1 )(𝑠 +𝑧2 )…

G(s) = n≥2
𝑆 𝑛 (𝑠+𝑝1 )(𝑠+𝑝2 )…

1
Steady-state error for a parabolic input: 𝑅 =
𝑆3

1
𝑠( 3 )
e (∞) =eparabola(∞) =lim
𝑠→0 1+𝐺(𝑠) 𝑠→0
1
= lim 𝑠2 +𝑠2𝐺(𝑠)
𝑠
 1
e (∞) =
lim 𝑠 2 𝐺(𝑠)
𝑠→0

In order to have zero steady- state error for a parabolic input we must have:
lim 𝑠2 𝐺(𝑠) = ∞
𝑠→0

Which means that at least 3 integrators (n≥2) then?

𝑧 𝑧 𝑧 𝑧
lim 𝑠2 𝐺(𝑆) = 1 2…
𝑠→0 𝑝1 𝑝2 …
e(∞)=𝑝 1𝑝2…… = constant.
1 2

If there is only one integrator or less

(𝑠+𝑧1 )(𝑠+𝑧2 )…
G(s) =
𝑆(𝑠+𝑝1 )(𝑠+𝑝2 )…

𝑆(𝑠+𝑧1 )(𝑠 +𝑧2 )…

𝑠 2 G(s) = lim 𝑠2 𝐺(𝑠) = 0 e (∞) =∞
(𝑠 +𝑝1 )(𝑠+𝑝2 )… 𝑠→0

Example 1: Find the steady-state error for inputs of 5u(t), 5tu(t), and 5t2u(t)

For a step input

For a ramp input

For a parabolic input

Example 2:

5
For the step input R(s) =
𝑆

5
For a ramp input R(s) =
𝑆2

5
For a parabolic input R(s) =
𝑆3

7-3 Static error constants and system type

Just as we defined damping ratio, natural frequency setting time, percent

overshoot and so as a performance specifications for transient response,
we define here parameters that can be used as steady-state error
performance specifications. They are called static error constants.

Definitions

Step input

Kp=lim𝐺(𝑠): position constant (dcgain)

𝑠→0

1
e (∞) = estep(∞) =
1+Kp
Ramp input

Kv=lim 𝑠𝐺(𝑠): velocity constant

𝑆→0

1
e (∞) = eramp (∞)=
Kv

Parabolic input

Ka=lim 𝑠2 𝐺(𝑠):acceleration constant

𝑆→0

1
e (∞) = epara(∞)=
Ka

System type

We define the system’s type to be number n of pure integrator in the

forward loop. Therefore a system with n=0 is a type 0. If n=1 or n=2 the
corresponding type is type 1 and type 2.

Type 0 Type 1 Type 2

Input Steady- Static Static Static

state error Error error Error error Error
error constant constant constant
Step u(t) 1 Kp= cst 1 Kp = ∞ 0 𝐾𝑝 = ∞ 0
1 + Kp 1 + Kp
Ramp 1 K v =0 ∞ K v =cst 1 K v =∞ 0
tu(t) Kv Kv
Parabola 1 K a =0 ∞ K a =0 ∞ K a =cst 1
t² u(t) Ka Ka

Steady-state error of a ramp input

Type 0 Type 1 Type 2

Step 1 0 0
1 + Kp
Ramp ∞ 1 0
Kv
Parabola ∞ ∞ 1
Ka

Step Ramp Parabola

Type 0 1 ∞ ∞
1 + Kp
Type 1 0 1 ∞
Kv
Type 2 0 0 1
Ka

Example
1000 (s+8)
G(s) =
(s+7)(s+9)

a) Evaluate Kp, Kv, Ka.

b) Use your answers to find the steady-state errors for the standard step,
ramp and parabolic input.

Resolution
8000
a) Kp= lim𝐺(𝑠)=
𝑆→0 63

Kv= lim 𝑠𝐺(𝑠) = 0

𝑆→0

Ka= lim 𝑠2 𝐺(𝑠)= 0

𝑠→0

b)
 Static error Error
constant

Step Kp=
8000 63
= 0,0078
63 8063
Ramp 0 ∞
Parabola 0 ∞

7-4-Steady-state error specifications

Static error constant Kp, Kv orKacan be used to specify the steady state error
characteristic of control system.

There is a wealth of information contained within a specification of a static

error constant.

For example of a control system has the specifications Kv = 1000 we can draw
several conclusions.

1. The system is stable (Kv = constant)

2. The system is of type 1
3. A ramp input is the test input (Kv = constant)
1
4. The steady state error between the ramp and the output is
Kv

Example 1

What information is contained in the specification Kp = 1000

• Stable
• Type 0
• Step input

Example 2

Gain design to meet a steady-state error specification given the control system,
find the value of K so that there is 10 % error in the steady-state.

The system is of type 1 and the error is consant so we have Kv.

5K
Kv= lim 𝑠𝐺(𝑠) =
𝑠→0 336
 1 336 10 3360
e(∞) = = = ═> K= = 672
Kv 5K 100 5

7-5 Steady-state error for disturbance

Feedback is also used to compensate for disturbances or unwanted input that

enters a system.

C(s) = E(s).G1(s).G2(s) + D(s).G2(s) (1)

but C(s) = R(s) –E(s) (2)

(2) in (1) R(s)-E(s) = E(s).G1(s).G2(s) + D(s).G2(s)

E(s).(1+G1(s).G2(s)) = R(s)- D(s).G2(s)

1 G2 (𝑠)
E(s) = R(s) - D(s)
1+G1 (𝑠)G2 (𝑠) 1+G1 (𝑠)G2 (𝑠)

1
: transfer function relating E(s) to R(s)
1+G1 (𝑠)G2 (𝑠)

G2 (𝑠)
− : transfer function relating E(s) to D(s)
1+G1 (𝑠)G2 (𝑠)

Example

1) Find the steady-error component due to a step disturbance

 2) Evaluate the steady-state error component due to a step disturbance.

Solutions

1)
1
1 𝑆(𝑆+25)
𝐸 (𝑠 ) = 100 𝑅 (𝑠 ) − 100 𝐷(𝑆)
1+ 1+
𝑠(s+25) 𝑆(𝑆 +25)

𝑠(s + 25) 1
𝐸 (𝑠 ) = 𝑅 (𝑠 ) − 𝐷(𝑠)
𝑠(𝑠+25) + 100 𝑠(𝑠+25) + 100

2)

𝑠+4 𝑠+2
𝐸 (𝑠 ) = 𝑅 (𝑠 ) − 𝐷(𝑠)
1001𝑠 + 2004 1001𝑠 + 2004

7-6Compensator (controller)

The goal of a compensator is to guarantee the dynamic and static behavior of a

process according to the design specification.

E(S) U(S)
Controller Plant

The controller which determine the control command U(s), depends on the
internal structure of the control.
PID controller

The most popular controller used in industries is the

proportional+integral+derivative controller (PID). The input-output
τ de(t)
relationship of a PID controller is u(t) = Kp e(t) + KI ∫0 e(t)dt+ KD
dt

In the Laplace transform terminology (Initial conditions =0)

U(s) KI
= Kp+ + KD s = D(s)
E(s) s

Where:

Kp is the proportional constant

KI is the integral constant

KD is the derivative constant

Proportional control

When a feedback control signal is a linearly proportional to error, we call the

result proportional feedback control.
U(s)
= D(s) = Kp
E(s)

D(S)=KP G(S)

Integral control

when a feedback control signal is proportional to the integral of the error

we call the result integral feedback control.
U(s) KI
= = D(s)
E(s) s
 KI
= D(s) G(S)
s

Derivative control

When a feedback control signal is proportional to the derivative of the

error we call the result derivative feedback control.
U(s)
= KD s = D(s)
E(s)

KD s = D(s) G(S)

Proportional integral and derivative control

When a feedback control signal is linearly, derivatively and integrally

proportional to the error we call the result proportional to the error we
call the result proportional, integral and derivative control.
U(s) KI
= D(s) = Kp+ + KD s
E(s) s

KI
D(s) = Kp+ + KD s G(S)
s
Proportional integral control

When a feedback control signal is linearly and integrally proportional to

the error we call the result proportional to the error we call the result
proportional and integral control.
U(s) KI
= D(s) = Kp+
E(s) s

KI
D(s) = Kp+ G(S)
s

Proportional and derivative

When a feedback control signal is linearly and derivatively proportional

to the error we call the result proportional to the error we call the result
proportional and derivativefeedback control.
U(s)
= D(s) = Kp+ KD s
E(s)

D(s) = Kp+ KD s G(S)

K
Kp s+KI Kp (s+ I )
KI Kp KI
D(s) = Kp+ = = = Kp(1+ )
s s s Kp s
1
= K (1+ )
Ti s

K=Kp
 KP
Ti = : integration time constant (min)
KI

(KP, KI) → (K, TI)

The PI controller can be approximated by phase lag controller.

KI s+z
K
⏟p + = D(s) ≈ K where z > p
s ⏟
s+p
PI phase lag

KD
D(s) = Kp+ KD s =Kp(1+ s) = (1+TD s)K
Kp

K=KP
KP
TD= : derivative time constant
KD

(KP, KD) → (K, TD)

The PD controller can be approximated by phase lead controller.

s+z
K
⏟p + K D = D(s) ≈ K where z < p.
⏟
s+p
PI phase lead

1
D(s) = K(1+ + TD s) where K=KP
TI s

K
TD= I = integral time constant
KP

K
TI = D = derivative time constant
KP

(KP, KD, KI) → (K,TI,TD)

The PID controller can be approximated by lead + lag controller (notch

controller)

KI (s+z )(s+z ) zlag < Plag

Kp+ + KD s ≈ K (s+Plag)(s+Plead ) where {
s lag lead zlead < Plead
 Name Effects on steady Effects on stability
state relative
K P
K
(s+α) PI Improves Degrades
s
K
(s2 +α1 s+α0 ) PID Improves Improves
s (somewhat)
K
(s+α)
α>β Lag Improves Degrades
s+β
(somewhat) (somewhat)
K
(s+α)
α<β Lead Degrades Improves
s+β
(somewhat) (somewhat)
K (s+α) PD Degrades Improves
s2 +2εw1 s+w1 ² Lead + lag Improves Improves
K
s2 +2εw2 s+w2 ²
(notch) (somewhat) (somewhat)

STRUCTURE OF PID CONTROLLER

Type Block diagram Equation

Serie 1
D(s) = K(1+ )(1+TDs)
Ti
P I D

Parallel P D(s) = K+
1
+TDs
Ti 𝑠

1
Mixte D D(s) = K(1+ +TDs)
Ti 𝑠

I
Chapter10: FREQUENCY RESPONSE TECHNIQUE
In this chapter you will learn the following

• How to plot frequency response.

• How to use frequency response to analyze stability.
• How to use frequency response to analyze a system’s transient and
steady-state error performance.
• How to use frequency response to design.
• The gain to meet stability specification.

I. INTRODUCTION

Frequency response methods were developed Nyquist and Bode in the

1930s.We will first discuss the concept of frequency response, define frequency
response derive analytical expressions for the frequency response plot frequency
response, develop ways to sketching the frequency response and then apply the
concept to control system analysis and design.

II. The concept of frequency response

Ms. are the amplitude of the sinusoids.

𝛷𝑠 are the phases angles of the sinusoids.

The steady-state output is:

𝑀𝑂 (𝑤)∠𝛷𝑂 (𝑤) = 𝑀𝑖 (𝑤)𝑀(𝑤)∠𝛷𝑖 (𝑚) + Φ(w)

The system function is given by:

𝑀𝑂 (𝑤)
𝑀(𝑤) = : The magnitude frequency response
𝑀𝑖 (𝑤)

Φ(w) = 𝛷𝑂 (𝑤) − 𝛷𝑖 (𝑤) : The phase frequency response

𝑀(𝑤)∠Φ(w) : Frequency response

𝑀(𝑤)∠Φ(w) = 𝑀(𝑤)𝑒 𝑖Φ(w) = 𝑀(𝑤) cos Φ(w) + 𝑖𝑀(𝑤) sin Φ(w)

Example:
𝑠+1
𝐺(𝑠) =
s
𝑗𝜔+1 1 𝑗
𝐺(𝑗𝜔) = G(s) = =1+ =1−
𝑆 = 𝑗𝑤 𝑗𝜔 𝑗𝜔 𝜔

2
√ 12
1 𝜔 2 + 1 √𝜔 2 + 1
|𝐺(𝑗𝜔)| = +( ) =√ =
𝜔 𝜔2 𝜔
−1
1
𝐴𝑟𝑔(𝐺(𝑗𝜔)) = 𝐴𝑟𝑐𝑡𝑔 ( 𝜔 ) = −𝐴𝑟𝑐𝑡𝑔( )
1 𝜔
 III. ANALYTICAL EXPRESSION FOR FREQUENCY RESPONSE

Consider the following transfer function:

(𝑠 + 𝑧1) (𝑠 + 𝑧2 ) … (𝑠 + 𝑧𝑘 )
𝐺 (𝑠) = K
𝑠 𝑛 (𝑠 + 𝑝1 )(𝑠 + 𝑝2 ) … (𝑠 + 𝑝𝑖 )
The magnitude frequency response is the product of the magnitude
frequency of each term.
𝐺 (𝑠)|𝑠=𝑗𝜔 = 𝐺(𝑗𝜔)

|𝑠+𝑧1 ||𝑠+𝑧2 |…|𝑠+𝑧𝑘 |

|𝐺(𝑗𝜔)| = 𝐾
𝑠𝑛 |𝑠+𝑝1 ||𝑠+𝑝2 |…|𝑠+𝑝𝑘 | 𝑠 → 𝑗𝜔

The phase frequency response is the argument of 𝐺 (𝑠) at s=j𝜔

|𝑠 + 𝑧1 ||𝑠 + 𝑧2 | … |𝑠 + 𝑧𝑘 |
𝐴𝑟𝑔[𝐺(𝑗𝜔)]|s=j𝜔 = 𝐴𝑟𝑔[𝐾 ]|
|𝑠𝑛 ||𝑠 + 𝑝1 ||𝑠 + 𝑝2 | … |𝑠 + 𝑝𝑘 |
s=j𝜔

𝐴𝑟𝑔[𝐺(𝑗𝜔)]|s=j𝜔
= 𝐴𝑟𝑔(𝐾) + 𝐴𝑟𝑔(𝑠 + 𝑧1 ) + 𝐴𝑟𝑔(𝑠 + 𝑧2 ) + ⋯ + −𝑛𝐴𝑟𝑔(𝑠)
− 𝐴𝑟𝑔(𝑠 + 𝑝1 ) − 𝐴𝑟𝑔(𝑠 + 𝑝2 ) − ⋯ 𝐴𝑟𝑔(𝑠 + 𝑝𝑘 )

1. First order system

The log-magnitude and phase frequency response curves as function of log w

are called Bode plots or Bode Diagram.
𝑘
Let G(s) =
1+𝜏𝑠
1
Define 𝑤𝑐 = the cutting frequency
𝜏

𝐾
G(s) = 𝐺(𝑠) =
𝑠 = 𝑗𝜔 1+𝜏𝑠 𝑠 = 𝑗𝜔
Magnitude:
 𝐾
|𝐺(𝑗𝜔)| =
2
√1 + ( 𝜔 )
𝜔𝑐

The logarithmic gain: (in decibel)

𝐺𝑑𝑏 = 20 log|𝐺(𝑗𝜔)|

𝐾
𝐺𝑑𝑏 = 20 log
2
√1 + ( 𝜔 )
[ 𝜔𝑐 ]
1
𝜔 2 2
𝐺𝑑𝑏 = 20 [log 𝐾 − log (1 + ( ) )]
𝜔𝑐

1 𝜔 2
𝐺𝑑𝑏 = 20 [log 𝐾 − log (1 + ( ) )]
2 𝜔 𝑐

𝜔 2
𝐺𝑑𝑏 = 20 log 𝐾 − 10 log (1 + ( ) )
𝜔𝑐

Phase:

𝐾 𝜔
𝐴𝑟𝑔[𝐺(𝑗𝜔)] = 𝐴𝑟𝑐𝑡𝑔 ( 𝜔) = 𝐴𝑟𝑔(𝐾) − 𝐴𝑟𝑔(1 + 𝑗 )
1+𝑗 𝜔𝑐
𝜔𝑐
𝜔
𝐴𝑟𝑔[𝐺(𝑗𝜔)] = 0 − 𝐴𝑟𝑐𝑡𝑔( )
𝜔𝑐

𝜔
Φ(𝜔) = 𝐴𝑟𝑔[𝐺(𝑗𝜔)] = −𝐴𝑟𝑐𝑡𝑔( )
𝜔𝑐
The logarithmic gain and phase are calculated for values of frequency between 0
and ∞ ,expressed in rad/s
 𝜔𝑐
𝜔𝑐 (rad/s) 0 2 𝜔𝑐 2𝜔𝑐 10𝜔𝑐 ∞
2𝐾 𝐾 𝐾
𝐾
|𝐺(𝑗𝑤)| K √3 √2 √5 √101 0

|𝐺(𝑗𝑤)| 20 log 𝐾 -1+20 log 𝐾 -3+20 log 𝐾 -7 -20+20 log 𝐾 −∞

𝐺𝑑𝑏 for K=2 0 5 3 -1 -14 −∞

𝐴𝑟𝑔[𝐺(𝑗𝑤)] 0 -26.5° -45° -63 ,43° -84,3° -90°

Bode stability criterion

1. The gain margin is found by using the phase to get the frequency 𝜔𝜋
when the phase angle is −180°.
1
2. 𝐺𝑚 = 20 log ⌈ ⌉
𝐴 (𝑗 𝜔𝜋 )
 If 𝐺𝑚 is under the zero dB line the system is stable.

3. The phase margin is found by using the magnitude curve to get the
frequency 𝜔1 in the phase curve at that frequency.
4. The phase margin 𝜑𝑚 is the difference between the phase value and
−180° (𝜑𝑚 = 𝐴𝑟𝑔 (𝐴(𝑗𝜔1 )) + 𝜋)

Simplified Bode Diagram criterion (Reverse Bode Diagram)

A feedback control system with a unity feedback is stable if, for the
frequency𝑤𝜋 , the curve 20 log |𝐴(𝑗𝜔1) | possess below the zero dB (0dB) line. It
is unstable otherwise.
NYQUIST DIAGRAM

The Nyquist diagram is the representation of 𝐺(𝑗𝜔) = |𝐺(𝑗𝜔)|∠ in the

complex plane (𝑅𝑒 (𝐺 (𝑗𝜔)) 𝑣𝑠 𝐼𝑚(𝐺(𝑗𝜔)) )

𝐺(𝑗𝜔) = |𝐺(𝑗𝜔)|[cos 𝐴𝑟𝑔[𝐺(𝑗𝜔)] + 𝑗 sin 𝐴𝑟𝑔[𝐺(𝑗𝜔)]]

𝐺(𝑗𝜔) = |𝐺(𝑗𝜔)| cos 𝐴𝑟𝑔[𝐺(𝑗𝜔)] + 𝑗|𝐺(𝑗𝜔)| sin 𝐴𝑟𝑔[𝐺(𝑗𝜔)]

𝑅𝑒 (𝐺 (𝑗𝜔)) = |𝐺(𝑗𝜔)| cos 𝐴𝑟𝑔[𝐺(𝑗𝜔)]

𝐼𝑚(𝐺(𝑗𝜔)) = |𝐺(𝑗𝜔)| sin 𝐴𝑟𝑔[𝐺(𝑗𝜔)]

𝜔𝑐
𝜔 (rad/s) 0 2 𝜔𝑐 2𝜔𝑐 10𝜔𝑐 100𝜔𝑐

𝑅𝑒 (𝐺 (𝑗𝜔)) 2 1 .59 1 0.4 0.010 0

𝐼𝑚(𝐺(𝑗𝜔)) 0 -0.8 -1 -0.8 -019 0

BLACK (NICHOLS) DIAGRAM

The Black diagram is the representation of the Mag(dB) and phase

2. Frequency response stability criterion

• Nyquist criterion:

The Nyquist criterion relates the stability of a closed-loop system to the open-
loop frequency response and open loop pole location.

Review:complex analysis

The residue theorem:

Suppose that 𝑓(𝑧) is analytic on and inside a simple closed contour C except at
isolate singular points𝑧1 ...𝑧𝑛 lying inside C, then:
𝑚

∮ 𝑓(𝑧) ⋅ 𝑑𝑧 = 2𝜋𝑖 ∑ 𝑅𝑒𝑠𝑓(𝑧𝑖 )

𝐶 𝑖=1

The Argument principal:

Theorem:

Suppose 𝑓(𝑧) is analytic on and inside this sample closed contour C with the
exception of pole type singularities lying inside C.

Suppose further that 𝑓(𝑧) does not vanish on C but may have zeros inside C.

If P and Z demote the number of poles and zeros of flying inside C (each
counted according to is multiplicity) then:
1 𝑓 ′ (𝑧) 1
∮ ⋅ 𝑑𝑧 = ∆ arg(𝑓(𝑧)) = 𝑍 − 𝑃
2𝜋𝑖 𝐶 𝑓(𝑧) 2𝜋

When ∆ arg(𝑓(𝑧))denotes the change in the argument of 𝑓(𝑧) as Z trace the

curve C in the positive sense.

Meaning

If C enclosed Z zeros of 𝑓(𝑧) and no poles then Γ.

Encircle the origin in the w-plane exactly Z time.

Derivation of the Nyquist criterion

Consider the system

R(s) +
G(s) C(s)
-
G(s)
T(s) =
1 + G(s)

The Nyquist criterion can tell you how many closed-loop poles in RHP

Important concepts:

1- Relationship between the poles of 1 + G(s) and the poles of G(s).

2- Relationship between the zeros of 1 + G(s)and the poles of closed-loop
transfer function.
𝐹(𝑆) = 1 + G(s)

𝑛(𝑠)
Let 𝐺(𝑠) = (1)
𝑑(𝑠)

𝑛(𝑠) 𝑑(𝑠)+𝑛(𝑠)
1 + 𝐺(𝑠) = 1 + = (2)
𝑑(𝑠) 𝑑(𝑠)
 𝐺(𝑠) 𝑛(𝑠)𝑑(𝑠) 𝑛(𝑠)
𝑇(𝑠) = = 𝑑(𝑠)+𝑛(𝑠) = (3)
1+𝐺(𝑠) 𝑑(𝑠)+𝑛(𝑠)
𝑑(𝑠)

Conclusion

1-poles ( 1 + G(s) = 𝑝𝑜𝑙𝑒(G(s))

G(s)
2-poles (T(s) = = 𝑧𝑒𝑟𝑜𝑠(1 + G(s))
1+G(s)

Meaning:

If C enclosed Z zeros of 𝑓(𝑧) and no poles then Γ:

1- If we assume a clock wise direction for mapping the points on contour A,

then the contour B maps in clock wise direction 𝐹(𝑠) has zeros or poles
that are not encircled by the contour.
2- If we assume a clockwise direction for mapping the points on contour A,
then the contour B maps in a clockwise direction ,if 𝐹(𝑠) has just pole;
and in the clockwise direction if 𝐹(𝑠) has just zeros that are encircled by
the contour, also the mapping encircle the origin.

3.
Derivation of Nyquist stability

Assume 𝐹(𝑠) = 1 + G(s) has two zeros and 3 poles

𝑁 =𝑃−𝑍

Where:

• N equals the number of counterclockwise rotation of contour B about the

origin.
• P equals the number of poles of 1 + G(s) inside the contour A.
• Z equals the number of zeros of 1 + G(s) inside the contour A.

Since the pole of 1 + G(s) equal poles of G(s) there are known, thus pole of
G(s) = 𝑃.

Since the zero of 1 + G(s) equals pole of the closed-loop system thus poles of
closed-loop G(s) = Z.

Hence 𝑁 = 𝑃 − 𝑍 or Z= 𝑃 − 𝑁

Therefore if the plot 1 + G(s) encircles the origin, the plot of G(s) will encircles
-1 on the real axis.
Example:

Z= 𝑃 − 𝑁 = 0 − 0 ⇒ 𝑍 = 0

No closed-loop poles in the RHP (therefore the closed-loop system is stable)

Z= 𝑃 − 𝑁 = 0 − (−2) = 2 ⇒𝑍=2

Two poles of the closed-loop system in the RHP (therefore the system is stable)
1
𝐺(𝑠) =
(𝑠 + 1)(𝑠 + 2)

Let us consider a closed-loop system:

D(s)
+ +
R(s)
C(s) G(s)

- +
 𝑌(𝑠) 𝐶(𝑠)𝐺(𝑠) 𝐴(𝑠)
𝑇(𝑠) = = = ; 𝐴(𝑠) = 𝐶(𝑠)𝐺(𝑠)
𝑅(𝑠) 1+𝐶(𝑠)𝐺(𝑠) 1+𝐴(𝑠)

For simplicity we assume 𝐶(𝑠) = 1

Definition1:

A system is stable if the poles of the transfer function have their real parts
negative.

Definition 2:

A feedback closed-loop system, T(s) is stable if the pole of its characteristic

equation 1 + G(s) = 0doesn’t possess negative real parts.

1 + 𝐺(𝑠) = 0 ⇒ 𝐺 (𝑠) = −1 = 1∠ (2k+1)180°; k=0, ±1, ±2………

Alternatively |𝐺(𝑠)| = 1 and 𝐴𝑟𝑔(𝐺(𝑠)) = (2𝑘 + 1)180°

When the control system undergoes an oscillation (sinusoidal output signal, for a
constant input, or zero input, the system is in harmonic or critical mode. It is at
the stability, we call 𝑤𝑐 the frequency of oscillation, 1 + 𝐺(𝑗𝜔𝑐 ) = 0 gives the
stability limit conditions for 𝑘 = −1(clockwise)

|𝐺(𝑗𝜔𝑐 )| = 1 and 𝐴𝑟𝑔(𝐺(𝑗𝜔𝑐 )) = −𝜋.

Note: In the stability diagram representation of transfer function, the singular

point (1; −𝜋) infrequency domain correspond to the 𝑗𝜔 − 𝑎𝑥𝑖𝑠crossing

(Point on the imaginary axis).

Algebraic stability condition:

• Magnitude limit condition :|𝐺(𝑗𝜔𝑐 )| = 1

• Phase limit condition : 𝐴𝑟𝑔(𝐺(𝑗𝜔𝑐 )) = −𝜋

The criticalfrequency 𝑤𝑐 , is determined from the phase limit condition.

The closed-loop system is stable if for 𝜔𝑐 , |𝐺(𝑗𝜔𝑐 )|∠1 ( 𝐺𝑑𝑏 ∠0) and unstable
otherwise.

Example
R(s) + Y(S)
C(s)=k G(s)
-

Determine the range of stability.

Solution :
𝐾
𝐴(𝑠) = 𝐶(𝑠)𝐺(𝑠) =
(1 + 𝜏𝑠)3

Phase condition to determine,wc

𝐴𝑟𝑔(𝐴(𝑗𝑤𝑐 ) = −𝜋

K
𝐴𝑟𝑔 ( ) = −𝜋
(1 + 𝑗𝜏𝜔𝑐 )3
𝐴𝑟𝑔( 𝐾) − 𝐴𝑟𝑔((1 + 𝑗𝜔𝑐 )3 ) = −𝜋

−3𝐴𝑟𝑔( 1 + 𝑗𝜏𝜔𝑐 ) = −𝜋

−3𝐴𝑟𝑡𝑎𝑛(𝜏𝜔𝑐 ) = −𝜋
𝜋
𝐴𝑟𝑡𝑎𝑛(𝜏𝜔𝑐 ) =
3

𝜋 √3
𝜏𝜔𝑐 = tan ( ) = √3 ⇒ 𝜔𝑐 =
3 𝜏
𝐾 𝐾
|𝐶 (𝑠)𝐺(𝑠)| = =
√3 (√1 + 3)3
(√(1 + 𝜏 2 ( 𝜏 )2 )3

𝐾 𝐾
|𝐶(𝑠)𝐺(𝑠)| = =
23 8
𝐾
The stability limit condition is = 1 ⇒ 𝐾 = 8.
8

Range of stability:
𝐾
< 1 from the magnitude stability condition 𝐾 < 8.
8
Graphical criterion

The algebraic criterion can not be applied to complex systems.

It is therefore wise to use graphical criteria.

We will be interested with the three graphical criterion such as Nyquist; Bode
and Black-Nichols diagram, along with the simplified criteria.

Nyquist stability criterion (from last time)

Simplified Nyquist criterion

Theorem:

A feedback control system with unity feedback is stable if by traversing the

Nyquist diagram of the open-loop transfer function, in the increasing frequency,
we leave the critical point, coordinate (-1:0) at the left.It is unstable otherwise

Example𝐴(𝑠) = 𝐶(𝑠)𝐺(𝑠)

Stable Unstable

Note: the simplified Nyquist criterion is not applicable if 𝐴(𝑠) has poles and
zeros in the RHP.

Stability margins:

Two characteristics are used to measure stability: Gain margin,𝐺𝑚 and phase
margin,𝜑𝑚 .

Let 𝑤𝜋 the frequency when the Diagram crosses, the negative real axis.
Definition: Gain margin

The gain margin is the factor by which we raise 𝐴(𝑗𝜔𝜋 ) before it becomes
unstable.
1 1 1
|𝐴(𝑗𝜔𝜋 )| ∗ |𝐴(𝑗𝜔 =1, = 𝐴𝑚 =
𝜋 )| 𝐴𝜋 𝐴(𝑗𝜔𝜋 )

1 1
𝐺𝑚 = 20 log [ ] = 20 log ⌈ ⌉ = 20 log 𝐴𝑚
𝐴(𝑗𝜔𝜋 ) 𝐴𝜋
Phase margin

Let 𝜔1 the frequency where|𝐴(𝑗𝜔1 )| = 1, 𝜑1 = 𝐴𝑟𝑔(𝐴(𝑗𝜔1 ))

𝜑𝑚 = 𝜑1 − (−𝜋)

𝜑𝑚 = 𝜋 + 𝐴𝑟𝑔(𝐴(𝑗𝜔1 ))
The phase margin is the amount by which the phase of 𝐴(𝑗𝜔1 )exceed −𝜋.

The system is stable if 𝐺𝑚 > 0 (𝐴𝑚 > 1) and 𝜑𝑚 > 0

Typical values of margins

6𝑑𝐵 < 𝐺𝑚 < 12𝑑𝐵 (2 < 𝐴𝑚 < 4) and 40° < 𝜑𝑚 < 60°
1
= 𝐴𝑚 ⇒ 𝐺𝑚 = 20 log 𝐴𝑚 and 𝜑𝑚 = 𝜑𝜔1 − 𝜑𝜋
𝐴𝜋

Bode stability criterion

5. The gain margin is found by using the phase to get the frequency 𝜔𝜋
when the phase angle is −180°.
1
6. 𝐺𝑚 = 20 log ⌈ ⌉
𝐺(𝑗𝜔𝜋 )

If 𝐺𝑚 is under the zero dB line the system is stable.

7. The phase margin is found by using the magnitude curve to get the
frequency 𝜔1 in the phase curve at that frequency.
8. The phase margin 𝜑𝑚 is the difference between the phase value and
−180°
(𝜑𝑚 = 𝐴𝑟𝑔 (𝐺(𝑗𝜔1 )) + 𝜋)
Simplified Bode Diagram criterion (Reverse Bode Diagram)

A feedback control system with a unity feedback is stable if, for the
frequency𝑤𝜋 , the curve 20 log |𝐴(𝑗𝜔1) | possess below the zero dB (0dB) line. It
is unstable otherwise.
 Black Diagram

A feedback control system with a unity, feedback is stable if traversing the curve
representing the open-loop transfer function in the direction of increasing
frequency; we leave the critical point (0dB,-180°) at the right. It is unstable
otherwise

Critical w
point
(0°,-180°) 0 dB (0°,-180°) 0 dB

-180° -90° 0 -180° -90° 0

Stable Unstable

Illustrations des marges de gain et de phase dans le plan de Nyquist, de Black et de Bode