COURSE NAME: MODERN CONTROL SYSTEMS

COURSE CODE : EE-412

SESSION: 2021-22 (ODD-SEMESTER)
BRANCH: ELECTRICAL ENGINEERING (7TH SEM.)

Presentation By:
Dr. B B Sharma
Associate Professor
Department of Electrical Engineering
National Institute of Technology
Hamirpur HP-177005
Phone: +91-1972-254540
E-mail: [email protected]
CONTENTS

Subject: Modern Control Subject Code: EE- Session: 2021-22

Systems 412 (Odd Sem.)

Course Type Course Plan Pre-requisites

Lectures Tutorials Practicals Credits (i) Linear Control

Core Course Systems
(L) (T) (P) (C)
(ii) Differential Calculus
(iii) Laplace Transforms
3 1 0 4 (iv) Matrix Algebra
(v) MATLAB/SIMULINK
software
EVALUATION PATTERN
Class Timings, Evaluation Schedule and Grading:

Mid-Term 30% Note:

Examination *Schedule of the mid-term and end semester
Class Test/Quiz 10% examinations will be as per the academic
Assignments/Project/ 10% calendar for the year 2021-22.
*Grading will be as per the policy adopted by
Attendance
the institute.
End Semester 50%
Examination
Class Lectures: Monday [11:00-12:00 AM], Tuesday [09:00-10:00 AM],
Wednesday [11:00-12:00 AM], Thursday [11:00-12:00 AM]
Timings
COURSE OUTCOMES

Course Outcomes (CO’s):

On successful completion of the course, the students should be able to address the
following:
 CO1: Writing description of different linear and nonlinear systems in state space form

 CO2: Designing control mechanisms for linear dynamic systems

 CO3: Writing difference equation description of discrete systems and analysing the
stability

 CO4: Understanding the nonlinearities and analysing the stability of nonlinear

systems

 CO5: Analyzing stability of discrete time systems described by difference equations.

 CO6: Understanding the advanced methods of control design

COURSE PLAN
Course Schedule and Books:
COURSE PLAN
Course Schedule and Books:
BOOKS

Text/Reference Books:

(a) Text Books(TB)

1. Control System Engineering by I.J. Nagrath and M. Gopal, “New Age International”.
2. Discrete time Control Systems by K. Ogata, “Prentice Hall International”.
3. Control System Engineering by Norman N. Nise, 6th edition, John Wiley and Sons.
(b) Reference Books (RB):
1. Discrete time Control Systems by K. Ogata, “Prentice Hall International ”.
2. Digital control and state variable methods by M.Gopal, “Tata McGraw Hill”.
3. Digital Control Systems by B.C. Kuo, Oxford University Press
4. Applied Nonlinear Control by J.J.E. Slotine and W. Li, “Prentice Hall, Englewood Cliffs New Jersey”.
STATE VARIABLE APPROACH
 Introduction
 Concept of state, state variable and state model
 State space representation of systems
 Block diagram for state equation,
 Transfer function decomposition
 Direct, parallel and cascade decomposition
 Solution of state equations
 Concept of controllability and observability
 Controller design using pole placement by state feedback
 Controller design using state observer.
REVIEW OF CLASSICAL CONTROL
 Greece (BC) – Float regulator mechanism
 Holland (16th Century)– Temperature regulator
 18th Century James Watt’s (Scottish Engineer) centrifugal governor for the
speed control of a steam engine.
 1920s Minorsky (Russia) worked on automatic controllers for steering
ships.
 1930s H. Nyquist (USA) developed a method for analyzing the stability of
controlled systems
 1940s Frequency response methods made it possible to design linear
closed-loop control systems (H.W. Bode, USA)
 1950s Root-locus method due to W.R. Evans (USA) was fully developed
 1960s State space methods, optimal control, adaptive control
 1980s Learning controls were investigated and developed.
 Present and on-going research fields on application of modern control
theory includes such non-engineering systems such as biological,
biomedical, economic and socio-economic systems
REVIEW OF CLASSICAL CONTROL
Main Features and Limitations of Classical Control:

 The classical control theory and methods are based on a simple input-output
description of the plant, usually expressed as a transfer function.

 These methods do not use any knowledge of the interior structure of the
 plant, and limit us to single-input single-output (SISO) systems.

 Classical control allows only limited control of the closed-loop behavior when
feedback control is used.

 Classical control techniques are simple and requires small computations.

 Analysis of MIMO systems is quite cumbersome.

REVIEW OF CLASSICAL CONTROL
Main Features and Limitations of Classical Control (Continued):…….

o These approaches are not applicable for time variant systems

o Not amenable to adaptive and optimal control systems, which are nonlinear and
time-varying.

o Mostly transfer function based so consider zero initial conditions

o Mostly frequency domain approaches are used for analysis which are complex.

o Trial and error based procedures mostly

MODERN CONTROL (STATE SPACE CONTROL)
Advantages of Modern Control Theory (State Space Approach):

 New approach which can address complex systems.

 Developed in 1960s with advent of digital computers

 Applicable to LTI/LTV, Nonlinear, MIMO systems.

 It’s a time domain approach

 Enable design of optimal control and adaptive control

 Design can be done for a class of inputs rather then specific inputs

 This approach considers initial conditions also, hence leads to accurate response.
STATE SPACE APPROACH
 The state-space description provides the dynamics as a set of coupled first-order
differential equations in a set of internal variables known as state variables,
together with a set of algebraic equations that combine the state variables into
physical output variables.

Important Terminology:
 State: Behavior of a system at any time t , which is depicted by the values of
variables is called as state.

 State Variables: Smallest set of variables of a dynamical system such that

knowledge of these variables at t=t0 along with input for t>=t0, completely
determines the behaviour of a system for t>=t0.
STATE SPACE APPROACH
 State variables can be measurable (physical) variables, which is generally
preferred. Auxiliary variable selection as state variables can also be made to
develop model.
STATE SPACE APPROACH
State Space and Output Equations:
STATE SPACE APPROACH
State Space and Output Equations:
STATE SPACE APPROACH
State Space and Output Equations:
STATE SPACE APPROACH
State Space and Output Equations:
STATE SPACE APPROACH
State Space and Output Equations:
STATE SPACE APPROACH
State Space and Output Equations:
STATE SPACE APPROACH
Choice of State Variables:
 In electric circuits, the energy storage devices are the capacitors and
inductors
– They contain all of the state information or “memory” in the system
– State variables: Voltage across capacitors, Current through inductor

 In mechanical systems, energy is stored in springs and masses:

– State variables: Spring displacement, Mass position and velocity

 The underlying transformation z=Tx should be linear with T be non-

singular and invertible.
STATE SPACE APPROACH
Choice of State Variables:
 Example of electric circuits:

 Example of mechanical systems:

STATE SPACE APPROACH

State Space Models:

•The state space model of a continuous-time dynamic system can be derived
either from the system model given in the time domain by
•Differential equation
•or from its Transfer Function representation

•Four state space forms:

 Phase variable form or Controllable Canonical Form (CCF)
Observable Canonical Form (OCF)
Modal form/Diagonal Form
Jordan form

All the above forms are used in modern control theory and practice due to
their specific advantages.
STATE SPACE APPROACH
State Space Model &Differential Equations:
• Consider a general nth-order model of a dynamic system represented by a n-th
order differential equation

• In order to derive a systematic procedure that transforms a differential equation of

order ‘n’ to a state space form representing a system of n first-order differential
equations, we first start with a simplified version of above equation.

• Consider the case when no derivatives with respect to the input are present.
STATE SPACE APPROACH
State Space Model &Differential Equations:
• Introduce the following (easy to remember) change of variables:
STATE SPACE APPROACH
State Space Model &Differential Equations:
• The above selection in state space description (Matrix Form) is given as

• (I)

• The corresponding output equation is

• This state space form is known in the literature as the phase variable canonical
form (Bush Form) .
STATE SPACE APPROACH
• In order to extend this technique to the general case defined by initial
differential equation, which includes derivatives with respect to the input,
we form an auxiliary differential equation as

(A)

• Here, change of variables is selected as:

Then, applying the superposition principle to

differential equation with derivatives of inputs i.e.

(B)
STATE SPACE APPROACH
• Since is response of (A), then by superposition, response of (B) is given by

(C)

• Using change of variables as given earlier in equation (A):

(D)

• Using (D) and change of variables in (C), we get

(E)

• It is interesting to point out that for bn = 0, which is almost always the case, the
• output equation also has an easy-to-remember form given by

(F)
STATE SPACE APPROACH
General Phase Variable form:
STATE SPACE APPROACH
EXAMPLE:1
STATE SPACE APPROACH
EXAMPLE:1
STATE SPACE APPROACH
Block Diagram Realization
STATE SPACE APPROACH
Example of state space modeling:
 Derive the state space model for the following RLC circuit:
STATE SPACE APPROACH
(1) Example of state space modeling:
STATE SPACE APPROACH
(1)Example of state space modeling:
Block diagram representation:
STATE SPACE APPROACH
(2) Example of state space modeling:
STATE SPACE APPROACH

State Space Variable Form(s) from Transfer Functions:

 Two methods are used to derive the state space models:
(i) Direct programming techniques
(ii) Parallel programming technique

 For simplicity, we consider only single-input single-output (SISO) systems.

 The resulting state space models may or may not contain all the modes of the
original transfer function, where by transfer function modes we mean poles of the
original transfer function (before zero-pole cancellation, if any, takes place).
STATE SPACE APPROACH

 If some zeros and poles in the transfer function are cancelled, then the
resulting state space model will be of reduced order.

 The corresponding modes will not appear in the state space model.

 In the following, direct programming technique is used to derive the state

space forms known as the controller canonical form and the observer
canonical form

 Then, by the method of parallel programming, the state space forms

known as modal canonical form and Jordan canonical form are obtained.
STATE SPACE APPROACH
Direct Programming Technique & Controller Canonical Form:
This technique is convenient in the case when the plant transfer function is given
in a non-factorized polynomial form.

For this system an auxiliary variable V(s) is introduced such that the transfer
function is split as

(1)

(2)
STATE SPACE APPROACH
Controller Canonical Form Continued…..

Equation (1) has the same structure as Bush form, after the Laplace transformation is applied,
which directly produces the state space system equation as that of Bush Form.
To find output equation from second part, one can write:
STATE SPACE APPROACH
Controller Canonical Form Continued…..
The model obtained is same as that of PVF ( Replace V by ξ )
STATE SPACE APPROACH
Controller Canonical Form Continued…..

This form of the system model is called the controller canonical form. It is identical
to the one obtained earlier.
Controller canonical form (CCF) plays an important role in control theory since it
represents the so-called controllable system.
STATE SPACE APPROACH
Direct Programming Technique & Observer Canonical Form:
Let the transfer function is given as
STATE SPACE APPROACH
Observer Canonical Form Continued…
The above relationship can be implemented by using a simulation diagram composed of
integrators in a cascade, and letting the corresponding signals to pass through the
specified number of integrators.

For example, terms containing should pass through only one

integrator (1/s) and so on.

Term containing should pass through (n-1) integrators (1/s) and

so on.

1
+ −𝑎0 𝑌 𝑠 + 𝑏0 𝑈(𝑠)
𝑠
STATE SPACE APPROACH
Observer Canonical Form Continued…
1
+ −𝑎0 𝑌 𝑠 + 𝑏0 𝑈(𝑠)
𝑠
STATE SPACE APPROACH
Observer Canonical Form Continued…
STATE SPACE APPROACH
Observer Canonical Form Continued…
STATE SPACE APPROACH
Observer Canonical Form Continued:
STATE SPACE APPROACH
Direct Programming Technique & Observer Canonical Form:
The matrix form of observer canonical form is easily obtained from the earlier
differential equations as

• OCF form represents an observable system, which means that all state variables
have an impact on the system output, and vice versa.

• That from the system output and state equations, one is able to reconstruct the
state variables at any time instant.
STATE SPACE APPROACH
Direct Programming Technique Example:
STATE SPACE APPROACH
(2) Example of state space modeling:

Obtain Observable Canonical form for the above system.

STATE SPACE APPROACH
Parallel Programming Technique: For this technique; we distinguish two
cases: distinct real roots and multiple real roots of the system transfer function
denominator.

(a) Distinct Real Roots

•This state space form is very convenient for applications.
•Derivation of this type of the model starts with the transfer function in the partial
fraction expansion form.
•Let us assume, without loss of generality, that the polynomial in the numerator has
degree of m < n, then
STATE SPACE APPROACH
Parallel Programming Technique:
The simulation diagram of such a form (Modal Canonical Form) is as follows:

The state space model derived from this simulation diagram is given by:
STATE SPACE APPROACH
Parallel Programming Technique Example:

The state space form obtained by using the direct programming technique is
STATE SPACE APPROACH
Parallel Programming Technique Example Continued…

Using partial fraction expansion, model is written as:

STATE SPACE APPROACH
Parallel Programming Technique Example:
•Note that the parallel programming technique presented is valid only for the case of
real distinct roots.
•If complex conjugate roots appear, they should be combined in pairs corresponding to
the second-order transfer functions, which can be independently implemented as
demonstrated in the next example.
STATE SPACE APPROACH
Parallel Programming Technique:
•Then, distinct real poles are implemented like in the case of parallel programming.

•A second-order transfer function, corresponding to the pair of complex conjugate

poles, is implemented using direct programming, and added in parallel to the first
order transfer functions corresponding to the real poles.
STATE SPACE APPROACH
Parallel Programming Technique:
STATE SPACE APPROACH

Parallel Programming Technique:

Multiple Real Roots:
•When the transfer function has multiple real poles, it is not possible to represent the
system in uncoupled form.
•Assume that a real pole p_1 of the transfer function has multiplicity r and that the
other poles are real and distinct, that is
STATE SPACE APPROACH

Parallel Programming Technique:

Multiple Real Roots Case Continued…
STATE SPACE APPROACH
Parallel Programming Technique Example:
Multiple Real Roots:

This form of the system model is known as the Jordan Canonical Form.
STATE SPACE APPROACH
The System Characteristic Equation and Eigenvalues:
o As in algebraic equations, we solve set of equations Ax=y, where A maps given vector x to y.

o If there exists a x =v such that Ax is mapped along the direction of x vector itself i.e. Av= kv; the
v is called eigenvector and scalar k is called eigenvalue.
STATE SPACE APPROACH

The System Characteristic Equation and Eigenvalues:

•It follows that the last equation is the characteristic equation, and hence the eigenvalues are the
zeros of the characteristic equation.

•For the characteristic equation of order n, the number of eigenvalues is equal to n.

•Thus, the roots of the characteristic equation in the state space context are the eigenvalues of the
matrix A.

•These roots in the transfer function context are called the system poles, according to the
mathematical tools for analysis of these systems—the complex variable methods.
STATE SPACE APPROACH
Modal Transformation:

The above state transformation is known as the modal transformation.

Note that pure diagonal state space form can be obtained only in the cases where
system matrix has distinct eigenvalues i.e. no multiplicity of eigenvalues is there.
STATE SPACE APPROACH
Multiple Eigenvalues
• If the matrix A has multiple eigenvalues, it is possible to transform it into a block diagonal form by
using the transformation:

• Here the matrix V is composed of n linearly independent, so-called generalized eigenvectors and
J is known as the Jordan canonical form.
• This block diagonal form contains simple Jordan blocks on the diagonal.
• Simple Jordan blocks have the given eigenvalue on the main diagonal, ones above the
main diagonal with all other elements equal to zero.
• For example, a simple Jordan block of order four is given by:
STATE SPACE APPROACH
Multiple Eigenvalues:

Vander Monde Structure for distinct and multiple eigenvalues:

Vander Monde Matrix for distinct eigenvalues:

This matrix can be used to diagonalize the system model.

STATE SPACE APPROACH
Multiple Eigenvalues:
Vander Monde like Matrix for eigenvalues with multiplicity r can be represented as:

It can be used to convert controllable canonical state space model to Jordan Canonical Model.
STATE SPACE APPROACH
;
Transformation Matrix for Diagonal and Jordan Canonical Form:
(i) For the case with distinct eigenvalues:
The transformation matrix is constituted as Vander Mode Matrix which results in
diagonalized form:

(ii) For the case with eigenvalues with multiplicity:

The transformation matrix is constituted as following Matrix which results in
Jordan canonical form with Jordan blocks of size of multiplicity:

Matrix V leads to diagonal form and Matrix S leads to Jordan Canonical form.
STATE SPACE APPROACH
Similarity Transformation:

• Note: A system modelled by state space technique may have many state space
forms related to each other through linear transformation known as similarity
transformation.
STATE SPACE APPROACH
Similarity Transformation Continued:
•

 With P as invertible matrix, above transformation is known as similarity

transformation.
 Very important features of this transformation are that under similarity
transformation, both, the system eigenvalues and the system transfer function, are
invariant.
STATE SPACE APPROACH

Deriving Transfer Function From State Space Model:

STATE SPACE APPROACH

Invariance of Eigenvalues under linear Transformation:

 A new state space model obtained by the similarity transformation does not change
internal structure of the model.

 That is, the eigenvalues of the system remain the same.

STATE SPACE APPROACH
Invariance of Transfer Function under linear Transformation :

•Another important feature of the similarity transformation is that the transfer function
remains the same for both models, which can be shown as follows:
STATE SPACE APPROACH
The System Characteristic Equation and Eigenvalues:
•In the state space variable approach, we know:

•A third form of the characteristic equation is obtained in the context of the transfer
function approach. The transfer function of a single-input single-output system is

•The characteristic equation in this case is obtained by equating the denominator of

this expression to zero.
•No matter what form of the system model is considered, the characteristic equation is
always the same.