Control Engineering

Dr. Viswanath Talasila

Department of Telecommunication Engineering
Ramaiah Institute of Technology, Bengaluru
 Module 12
Lecture 1
Introduction to State Space Systems
 The transfer function modeling approach

The good stuff ! The drawbacks

• Transfer functions are defined in the • Can be used only for linear
Laplace domain and are a
systems
generalization of Fourier transforms
• Can be extended for (some)
• System: input-output view. nonlinear systems, but tricky !
• Input can be signals of various
frequencies • Limited to single input single
• Bandwidth of the control system output systems
• Can handle MIMO systems, with
• Algebraic properties of transfer some jugglery !
functions
• Combine different systems • The internal behaviour of the
system is not known – black box !
• Very dominant in the industry !

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 What is a state space system?

• A state space representation of a system is a mathematical model where the inputs,

outputs and states (internal variables) are related by first order differential equations
• Chemical process or Audio Video signal amplifiers (multiple inputs, states and outputs)

• Can isolate individual system “components” (states) into controllable/uncontrollable or

observable/unobservable types

• Really useful techniques from linear algebra for analysis and design

Standard differential equation analysis is now supplemented with (really easy) linear algebraic properties !

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 General procedure to obtain a state space
system
• State space model: a representation of the dynamics of a system in the form of a set
of N first order differential equations
• Convert the Nth order differential equation that governs the dynamics into N first-order
differential equations

• The number of states is (usually) equal to the order of the differential equation
which models your system
• example: for an electrical system it is usually equal to the number of energy storing elements

• Sounds easy? It is !!

Standard differential equation analysis is now supplemented with (really easy) linear algebraic properties !

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 A simple example
A mass-spring damper system

A mass-spring system would

Ideally behave like a harmonic
oscillator. With the damper in
place we would (hope to) see
b damped oscillatory behaviour.

Let p denote the position of the mass in the above mass spring damper (MSD)
system. From the force balance equation we know that the dynamics of the MSD is
given by a second order differential equation

where m is the mass (kg), b is the damping coefficient (Ns/m) and k is the spring constant(N/m).
F is the applied force(N).

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 A simple example

b The final state space model will be as below

If the measured output of the system is the position, then we have that

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Simulation of the MSD
MATLAB code
m = 1000 kg
k = 2000 Nm-1
b = 1000 N / ms-1

We have a stable damped

step response, with a small
overshoot.

Can we guess the pole

locations from this?

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Simulation of the MSD
We observed a damped stable
m = 1000 kg response in the previous slide.
k = 2000 Nm-1 Can we guess the pole locations
b = 1000 N / ms-1
from this? Lets see…

A is a symmetric matrix and a natural question is Interestingly the poles

“what are the eigen values of A”? We will obtain the computed from the transfer
following eigenvalues function are also the same !

We shall see in a couple of slides that the eigenvalues of the A matrix are exactly
equal to the poles of the corresponding transfer function

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 What is a state?
x(t) is called the state of the system of time t because

• The future output of the system depends only on the current state and
input

• The future output depends on the past input only through the state

• The state summarizes the effects of all past inputs on future outputs…
• In other words the state acts like the memory of the system

Example: Rechargeable mobile phone – the state is the current state of charge of the
battery. If you know that state, then you do not need to know how that level of charge was
achieved (assuming a ideal battery) to predict the future working of the phone
• But to consider all nonlinear effects, you might also need to know how many cycles the
battery has gone through

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 General Form of a linear time invariant state space model

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Obtaining state space equations from differential equations

consider an nth order linear plant model described by the differential

equation

We can define a useful set of state variables as

Finally:

How does this look in matrix form?

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Obtaining state space equations from differential equations

If we are measuring only one state, then we have

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 General Form of a linear time invariant state space model

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Simple RLC example

+
Vout
-

The final state space model will be as below

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example

The state space model will be

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example – coupled mass spring damper

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example – coupled mass spring damper

Define new variables:

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example – coupled mass spring damper

The state
space eqns
in standard
form are
then:

States are position and velocity of each mass

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila
 State Space to Transfer Function
Take Laplace transform of the basic state equation:

Combine this with the output equation

The following relationship holds

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example
Obtain the transfer function for the system described in the state space model.

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example
Obtain the transfer function for the system described in the state space model.

Now we only need

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example
Obtain the transfer function for the system described in the above state space model.

We finally obtain

Don’t forget to check if the poles match with the eigen values !!
Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila
 Eigenvalues of A and relation to poles of a transfer
function
Recall that the roots of the characteristic equation of a transfer function determine stability and the
performance of a system.

From the slide on relationship

between TFs and SS systems…
we had…

Rewriting this we have…

Setting the denominator of the transfer-

function matrix Y(s) to zero, we get the
characteristic equation (in the state space
setting) as
The solutions to this are nothing but the eigenvalues of A !!

Thus the roots of the characteristic equation of a TF are the same as the
eigenvalues of the A matrix of the equivalent SS model

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example

1) Obtain the state space model for the system described by the following differential
equation
2) Show that the roots of the characteristic equation of the transfer function are
equal to the eigenvalues of the SS model

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example
1) Obtain the state space model for the system described by the following differential equation
2) Show that the roots of the characteristic equation of the TF are equal to the eigenvalues of
the SS model

We can easily show…

Define the state variables

o/p eqn is our choice

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example
1) Obtain the state space model for the system described by the following differential equation
2) Show that the roots of the characteristic equation of the TF are equal to the eigenvalues of
the SS model

C = éë 1 0 0 ùû

From the differential equation, we can calculate the transfer function and we obtain:

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Example
1) Obtain the state space model for the system described by the following differential equation
2) Show that the roots of the characteristic equation of the TF are equal to the eigenvalues of
the SS model

Compare the roots of the TF with the eigenvalues of the A matrix…

Thus the roots of the characteristic equation of a TF are the same as

the eigenvalues of the A matrix of the equivalent SS model

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila

 Overview

Summary: Lecture 1
Contents: Lecture 2
 Introduction  Difference between
 Simple example (mass-spring- linear and nonlinear
damper)  Taylor series and the
 Definition of a state space system Jacobian
 Three examples for state space  2 detailed examples of
modeling – mechanical and linearization (predator-
electrical prey and Van der Pol
 Relation between state space and oscillator
transfer function

Control Engineering Module 12 – Lecture 1 Dr. Viswanath Talasila