Linearization, Transfer Function, Block

Diagram Representation, Transient Response

Automatic Control, Basic Course, Lecture 2

Gustav Nilsson
15 November 2016
Lund University, Department of Automatic Control
Content

1. Linearization

2. Transfer Function

3. Block Diagram Representation

4. Transient Response

2
Linearization
Linearization - Why?

Many systems are nonlinear. However, one can approximate them with
linear ones. This to get a system that are easier to analyze.
A few examples of nonlinear systems:

• Water tanks (Lab 2)

• Air resistance
• Action potentials in neurons
• Pendulum under the influence of gravity
• ...

4
Linearization - How?

Given a nonlinear system ẋ = f (x, u), y = g (x, u)

5
Linearization - How?

Given a nonlinear system ẋ = f (x, u), y = g (x, u)

1. Determine a stationary point (x0 , u0 ) to linearize around
ẋ0 = 0 , f (x0 , u0 ) = 0

5
Linearization - How?

Given a nonlinear system ẋ = f (x, u), y = g (x, u)

1. Determine a stationary point (x0 , u0 ) to linearize around
ẋ0 = 0 , f (x0 , u0 ) = 0

2. Make a first order Taylor series expansions of f and g around

(x0 , u0 ):
@ @
f (x, u) ⇡ f (x0 , u0 ) + f (x0 , u0 )(x x0 ) + f (x0 , u0 )(u u0 )
@x @u
@ @
g (x, u) ⇡ g (x0 , u0 ) + g (x0 , u0 )(x x0 ) + g (x0 , u0 )(u u0 )
@x @u
Notice that f (x0 , u0 ) = 0 and let y0 = g (x0 , u0 )

5
Linearization - How?

Given a nonlinear system ẋ = f (x, u), y = g (x, u)

1. Determine a stationary point (x0 , u0 ) to linearize around
ẋ0 = 0 , f (x0 , u0 ) = 0

2. Make a first order Taylor series expansions of f and g around

(x0 , u0 ):
@ @
f (x, u) ⇡ f (x0 , u0 ) + f (x0 , u0 )(x x0 ) + f (x0 , u0 )(u u0 )
@x @u
@ @
g (x, u) ⇡ g (x0 , u0 ) + g (x0 , u0 )(x x0 ) + g (x0 , u0 )(u u0 )
@x @u
Notice that f (x0 , u0 ) = 0 and let y0 = g (x0 , u0 )
3. Introduce x = x x0 , u = u u0 and y = y y0

5
Linearization - How?

Given a nonlinear system ẋ = f (x, u), y = g (x, u)

1. Determine a stationary point (x0 , u0 ) to linearize around
ẋ0 = 0 , f (x0 , u0 ) = 0

2. Make a first order Taylor series expansions of f and g around

(x0 , u0 ):
@ @
f (x, u) ⇡ f (x0 , u0 ) + f (x0 , u0 )(x x0 ) + f (x0 , u0 )(u u0 )
@x @u
@ @
g (x, u) ⇡ g (x0 , u0 ) + g (x0 , u0 )(x x0 ) + g (x0 , u0 )(u u0 )
@x @u
Notice that f (x0 , u0 ) = 0 and let y0 = g (x0 , u0 )
3. Introduce x = x x0 , u = u u0 and y = y y0
4. The state-space equations in the new variables are given by:
˙x = ẋ ẋ0 = f (x, u) ⇡ @ f (x0 , u0 ) x + @ f (x0 , u0 ) u = A x + B u
@x @u
@ @
y = g (x, u) y0 ⇡ g (x0 , u0 ) x + g (x0 , u0 ) u = C x + D u 5
@x @u
Example - Linearization

Example
The dynamics of a specific system is described by

ẋ1 = x2
x24 p
ẋ2 = 2 + x1 + u + 1
x1
y = x12 + u 2

a) Find all stationary points

b) Linearize the system around the stationary point corresponding to
u0 = 3

6
Transfer Function
Laplace Transformation

Let f (t) be a function of time t, the Laplace transformation L(f (t))(s) is

defined as Z 1
L(f (t))(s) = F (s) = e st f (t)dt
0
0 n 1
Assuming that f (0) = f (0) = · · · = f (0) = 0 (steady assumption
during this course) it has the property that
✓ n ◆
d f (t)
L (s) = s n F (s)
dt n

See Collection of Formulae for a table of Laplace transformations.

8
Example - Transfer Function

Example
A system’s dynamics is described by the di↵erential equation

ÿ + a1 ẏ + a2 y = b1 u̇ + b2 u.

After Laplace transformation we get

(s 2 + a1 s + a2 )Y (s) = (b1 s + b2 )U(s)

which can be written as

G (s)
z }| {
b1 s + b2
Y (s) = 2 U(s) = G (s)U(s)
s + a1 s + a2

G (s) is called the transfer function of the system.

9
Transfer Function

Relation between control signal U(s) and output Y (s):

Y (s) = G (s)U(s)

G (s) often fraction of polynomal, i.e.,

Q(s)
G (s) =
P(s)

Zeros of Q(s) are called zeros of the system, zeros of P(s) are called
poles of the system.
The poles play a very important role for the system’s behavior.

10
From State Space to Transfer Function

For a system on state space form

ẋ = Ax + Bu
y = Cx + Du

the transfer function is given by

1
G (s) = C (sI A) B +D

Observe: the denominator of G (s) is given by P(s) = det(sI A), so

eigenvalues of A are poles of the system.

11
From Transfer Function to State Space

Can be done in several ways, see Collection of Formulae.

Example
A system’s transfer function is
2s + 1
G (s) =
s 3 + 4s 8
Write the system on a state space form of your choice.

12
Three Ways to Describe a Dynamical System

Di↵erential equation
ÿ + a1 ẏ + a2 y = b1 u̇ + b2 u

x1 = y
⇣ ⌘
x2 = ẏ dn f (t)
L dt n
(s) = s n F (s)
...

Collection of Formulae

State space
Transfer function
ẋ = Ax + Bu
Y (s) = G (s)U(s) = Q(s)
P(s)
U(s)
y = Cx + Du

1
G (s) = C (sI A) B +D

13
Block Diagram Representation
Block Diagram - Transfer Function

When the blocks in a block diagram are replaced by transfer functions, it

is possible to describe the relations between signals in a easy way.

u y
GP

Y (s) = GP (s)U(s)

15
Block Diagram - Components

Most block diagrams consist of three components:

• Blocks - Transfer functions

• Arrows - Signals
• Summations

r e u y
+ GR GP

16
Determine Transfer Function From Block Diagram

r e u y
+ GR GP

Y = GP U, U = GR E , E =R Y
From the equations above the transfer function between r and y is
GP GR
Y = U
1 + GP GR

17
Example - Transfer Functions

Example
Two systems, G1 and G2 , are interconnected as in the figure below

G1

u y
+

+ G2

Compute the transfer function from u to y , Gyu .

18
Transient Response
Solution to State Space Equation

Given a system on state space form

ẋ = Ax + Bu
y = Cx + Du

The solution, y (t), is then given by

Z t
y (t) = Ce At x(0) + C e A(t ⌧)
Bu(⌧ )d⌧ + Du(t)
0

Initial state,
Weighted integral of Direct term, often
uninteresting except
the control signal, neglectable in
when the controller is
interesting part practical systems
initialized

20
Impulse Response

Shows how the system responds when the input is a short pulse, i.e., a
Dirac function
u(t) = (t)
The Laplace transformation is
Z 1
st
U(s) = e (t)dt = 1
0

Hence
Y (s) = G (s)U(s) = G (s)

Not so common in technological applications, can we think of other

applications?

21
Example - Impulse Response

Let the transfer function of the system be:

2
G (s) = 2
s + 3s + 2

0.6
0.4
y (t)

0.2
0
2 0 2 4 6 8 10
t
1
u(t)

0.5

0
2 0 2 4 6 8 10
t
22
Step Response

Shows how the system responds when the input is a step, i.e.,
(
1 t 0
u(t) =
0 t<0

The Laplace transformation is

Z 1 Z 1
st st 1⇥ ⇤
st 1 1
U(s) = e u(t)dt = e dt = e 0
=
0 0 s s

Very common in technological applications

23
Example - Step Response

Let the transfer function of the system be:

2
G (s) = 2
s + 3s + 2

1
y (t)

0.5

0
2 0 2 4 6 8 10
t
1
u(t)

0.5

0
2 0 2 4 6 8 10
t
24