Systems Analysis and Control

Matthew M. Peet
Illinois Institute of Technology
Lecture 6: Calculating the Transfer Function
Introduction
In this Lecture, you will learn: Transfer Functions
Transfer Function Representation of a System
State-Space to Transfer Function
Direct Calculation of Transfer Functions
Block Diagram Algebra
Modeling in the Frequency Domain
Reducing Block Diagrams
M. Peet Lecture 6: Control Systems 2 / 23
Previously:
The Laplace Transform of a Signal
Denition: We dened the Laplace transform of a Signal.
Input, u = u.
Output, y = y
Theorem 1.
For a bounded, linear, causal, time-invariant system, y = Gu, there exists a
Transfer Function,

G, so that the ratio of input to output is
y(s)
u(s)
=

G(s)
In this lecture, we will discuss several ways of nding the Transfer Function.
M. Peet Lecture 6: Control Systems 3 / 23
Transfer Functions
Example: Simple System
State-Space:
x(t) = x(t) + u(t)
y(t) = x(t) .5u(t) x(0) = 0
Apply the Laplace transform to the rst equation:

_
x(t) = x(t) + u(t)
_
which gives s x(s) + x(0) = x(s) + u(s).
Noting that x(0) = 0 and solving for x(s) gives
(s + 1) x(s) = u(s) and so x(s) =
1
s + 1
u(s).
Similarly, the second equation gives y(s):
y(s) = x(s) .5 u(s) =
1
s + 1
u(s) .5 u(s) =
1 .5(s + 1)
s + 1
u(s) =
1
2
s 1
s + 1
u(s)
Thus we have the Transfer Function:

G(s) =
1
2
s 1
s + 1
M. Peet Lecture 6: Control Systems 4 / 23
Transfer Functions
Example: Step Response
The Transfer Function provides a convenient way to nd the response to inputs.
Step Input Response: u(s) =
1
s
y(s) =

G(s) u(s) =
1
2
s 1
s + 1
1
s
=
1
2
s 1
s
2
+ s
=
1
2
_
2
s + 1

1
s
_
Consulting our table of Laplace
Transforms,
y(t) =
1
2

1
2
s + 1

1
2

1
1
s
= e
t

1
2
1(t)
0 1 2 3 4 5 6
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
0.5
Step Response
Time (sec)
A
m
p
l
i
t
u
d
e
M. Peet Lecture 6: Control Systems 5 / 23
Transfer Functions
Example: Sinusoid Response
Sine Function: u(s) =
1
s
2
+1
y(s) =

G(s) u(s) =
1
2
s 1
s + 1
1
s
2
+ 1
=
1
2
s 1
s
3
+ s
2
+ s + 1
=
1
2
_
s
s
2
+ 1

1
s + 1
_
Consulting our table of Laplace Transforms,
y(t) =
1
2
cos t
1
2
e
t
0 2 4 6 8 10
1
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
0.8
1
t
u
(
t
)
input
0 2 4 6 8 10
0.8
0.6
0.4
0.2
0
0.2
0.4
0.6
t
y
(
t
)
output
Note that this is the same answer we got by integration in Lecture 4.
M. Peet Lecture 6: Control Systems 6 / 23
Inverted Pendulum Example
Return to the pendulum.
Dynamics:

(t) =
Mgl
2J
(t) +
1
J
T(t)
y(t) = (t)
For the rst equation,
s
2

(s)

(0)s(0) =
Mgl
2J

(s)+
1
J

T(s)
Set x(0) = 0 and x(0) = 0 and solve for

(s):

(s) =
1
J
1
s
2

Mgl
2J

T(s)
Second Equation: y(s) =

(s)
Transfer Function:

G(s) =
1
J
1
s
2

Mgl
2J
M. Peet Lecture 6: Control Systems 7 / 23
Inverted Pendulum Example: Impulse Response
Impulse Input: u(s) = 1
y(s) =

G(s) u(s) =
1
J
1
s
2

Mgl
2J
=
1
J
1
(s
_
Mgl
2J
)(s +
_
Mgl
2J
)
=
1
J

2J
Mgl
_
_
1
s
_
Mgl
2J

1
s +
_
Mgl
2J
_
_
0 5 10 15
0
2
4
6
8
10
12
14
16
18
x 10
5
Impulse Response
Time (sec)
A
m
p
lit
u
d
e
Figure: Impulse Response with
g = l = J = 1, M = 2
In time-domain:
y(t) =
1
J

2J
Mgl
_
e

Mgl
2J
t
e

Mgl
2J
t
_
Pendulum Accelerates to innity!
M. Peet Lecture 6: Control Systems 8 / 23
Constructing the Transfer Function: Suspension System
x
1
x
2
m
c
m
w
u
Recall the dynamics:
z
1
(t) =
K
1
m
c
z
1
(t)
c
m
c
z
1
(t) +
K
1
m
c
z
2
(t) +
c
m
c
z
2
(t)
z
4
(t) =
K
1
m
w
z
1
(t) +
c
m
w
z
1
(t)
_
K
1
m
w
+
K
2
m
w
_
z
2
(t)
c
m
w
z
2
(t)
K
2
m
w
u(t)
y(t) =
_
z
2
(t)

M. Peet Lecture 6: Control Systems 9 / 23

Constructing the Transfer Function: Suspension System
x
1
x
2
m
c
m
w
u
Apply the Laplace Transform to the dynamics:
s
2
z
1
(s) =
K
1
m
c
z
1
(s)
c
m
c
s z
1
(s) +
K
1
m
c
z
2
(s) +
c
m
c
s z
2
(s)
s
2
z
2
(s) =
K
1
m
w
z
1
(s) +
c
m
w
s z
1
(s)
_
K
1
m
w
+
K
2
m
w
_
z
2
(s)
c
m
w
s z
2
(s)
K
2
m
w
u(s)
y(s) = z
2
(s)
M. Peet Lecture 6: Control Systems 10 / 23
Constructing the Transfer Function: Suspension System
We isolate the z
1
and z
2
terms:
_
s
2
+
c
m
c
s +
K
1
m
c
_
z
1
(s) =
_
K
1
m
c
+
c
m
c
s
_
z
2
(s)
_
s
2
+
c
m
w
s +
K
1
m
w
+
K
2
m
w
_
z
2
(s) =
_
K
1
m
w
+
c
m
w
s
_
z
1
(s)
K
2
m
w
u(s)
y(s) = z
2
(s)
Which yields
z
1
(s) =
_
K
1
m
c
+
c
m
c
s
_
_
s
2
+
c
m
c
s +
K
1
m
c
_
z
2
(s)
z
2
(s) =
K
1
m
w
+
c
m
w
s
s
2
+
c
m
w
s +
K
1
m
w
+
K
2
m
w
z
1
(s)
K
2
m
w
s
2
+
c
m
w
s +
K
1
m
w
+
K
2
m
w
u(s)
M. Peet Lecture 6: Control Systems 11 / 23
Constructing the Transfer Function: Suspension System
Now we can plug in for z
1
and solve for z
2
:
z
2
(s) =
K
2
(m
c
s
2
+ cs + K
1
)
m
c
m
w
s
4
+ c(m
w
+ m
c
)s
3
+ (K
1
m
c
+ K
1
m
w
+ K
2
m
c
)s
2
+ cK
2
s + K
1
K
2
u(s)
Compare to the State-Space Representation:
d
dt
_

_
z
1
z
2
z
3
z
4
_

_
(t) =
_

_
0 1 0 0

K
1
m
c

c
m
c
K
1
m
c
c
m
c
0 0 0 1
K
1
m
w
c
m
w

_
K
1
m
w
+
K
2
m
w
_

c
m
w
_

_
_

_
z
1
z
2
z
3
z
4
_

_
(t) +
_

_
0
0
0

K
2
m
w
_

_
u(t)
y(t) =
_
1 0 0 0
0 0 1 0
_
_

_
z
1
z
2
z
3
z
4
_

_
(t) +
_
0
0
_
u(t)
Note: We only used one output to nd the transfer function.
M. Peet Lecture 6: Control Systems 12 / 23
Block Diagrams
Series (Cascade) Interconnection
The interconnection of systems can be represent by block diagrams.
G H
u y y
1
Cascade of Systems: Suppose we have two systems: G and H.
Denition 2.
The Cascade or Series interconnection of two systems is
y
1
= Gu y = Hy
1
or
y = H(G(u))
M. Peet Lecture 6: Control Systems 13 / 23
Block Diagrams
Series Connection
G(s) H(s)
u(s) y(s) y
1
(s)
H(s)G(s)
u(s) y(s)
The Transfer function of a Series interconnection is Simple
The output of system 1 is the input to system 2.
Let

G(s) and

H(s) be the transfer functions for G and H.
Apply the Laplace transform to get
y
1
(s) =

G
1
(s) u(s) y(s) =

H(s) y
1
(s) =

H(s)

G(s) u(s)
The Transfer Function,

T(s) for the cascade of G and H is

T(s) =

H(s)

G(s)
Note: The order of the Transfer Functions!
M. Peet Lecture 6: Control Systems 14 / 23
Block Diagrams
Parallel Connection
The parallel Interconnection is even simpler.
G
H
u y
+
+
Parallel Interconnection: Suppose we have two systems: G and H.
Denition 3.
The Parallel interconnection of two systems is
y
1
= Gu y
2
= Hu y = y
1
+ y
2
or
y = H(u) + G(u)
M. Peet Lecture 6: Control Systems 15 / 23
Block Diagrams
Parallel Connection
G(s)
H(s)
u(s) y(s)
y
1
(s)
+
+
H(s)+G(s)
u(s) y(s)
The Transfer function of a Parallel interconnection is trivial
Apply the Laplace transform to get
y(s) = y
1
(s) + y
2
(s) =

G(s) u(s) +

H(s) u(s) =
_

H(s) +

G(s)
_
u(s)
The Transfer Function,

T(s) for the parallel interconnection of G and H is

T(s) =

H(s) +

G(s)
M. Peet Lecture 6: Control Systems 16 / 23
Block Diagrams
Lower Feedback Interconnection
G(s)
K(s)
+
-
y(s)
u(s)
Feedback:
Controller: z = K(u y) Plant: y = Gz
Applying the Laplace Transform gives
z(s) =

K(s) y(s) +

K(s) u(s) y(s) =

G(s) u
i
(s)
so
y(s) =

G(s) z(s) =

G(s)

K(s) y(s) +

G(s)

K(s) u(s)
Solving for y(s),
y(s) =

G(s)

K(s)
1 +

G(s)

K(s)
u(s)
M. Peet Lecture 6: Control Systems 17 / 23
Block Diagrams
Upper Feedback Interconnection
There is an alternative Feedback
interconnection
Let u be the external
input/disturbance
y is the output
y(s) =

G(s) z(s)
z(s) = u(s)

K(s) y(s)
G(s)
K(s)
+
-
y(s)
u(s)
Which yields
y(s) =

G(s)
_
u(s)

K(s) y(s)
_
=

G(s) u(s)

G(s)

K(s) y(s)
hence the Transfer Function is given by
y(s) =

G(s)
1 +

G(s)

K(s)
u(s).
M. Peet Lecture 6: Control Systems 18 / 23
The Eect of Feedback: Impulse Response
Inverted Pendulum Model
Transfer Function

G(s) =
1
Js
2

Mgl
2
Controller: Static Gain:

K(s) = K
Input: Impulse: u(s) = 1.
Closed Loop: Lower Feedback
y(s) =

G(s)

K(s)
1 +

G(s)

K(s)
u(s) =
K
Js
2

Mgl
2
1 +
K
Js
2

Mgl
2
=
K
Js
2

Mgl
2
+ K
First Case:
If K >
Mgl
2
, then K
Mgl
2
> 0, so
y(s) =
K/J
s
2
+
_
K/J
Mgl
2J
_
y(t) =
K
J
_
K/J
Mgl
2J
sin
_
_
K/J
Mgl
2J
t
_
0 2 4 6 8 10 12
2.5
2
1.5
1
0.5
0
0.5
1
1.5
2
2.5
Impulse Response
Time (sec)
A
m
p
l
i
t
u
d
e
M. Peet Lecture 6: Control Systems 19 / 23
The Eect of Feedback: Impulse Response
Inverted Pendulum Model
0 5 10 15 20 25
0
2
4
6
8
10
12
14
16
18
x 10
6
Impulse Response
Time (sec)
A
m
p
l
i
t
u
d
e
Second Case:
If K <
Mgl
2
, then K
Mgl
2
< 0, so
y(s) =
K
J
_
_
1
s
_
K/J
Mgl
2J
+
1
s +
_
K/J
Mgl
2J
_
_
y(t) =
K
J
_
e

K/J
Mgl
2J
t
+ e

K/J
Mgl
2J
t
_
Important: Value of K determines stability vs. instability
M. Peet Lecture 6: Control Systems 20 / 23
Block Diagrams
Reduction
Now lets look at how to reduce a more complicated interconnections
u(s)
y(s)
e(s)
-
+
-
+
1/s
1/s
K1
Label
The output from the inner loop z
The input to the inner loop u
First Close the Inner Loop using the Lower Feedback Interconnection.
z(s) =
K
1
s
K
1
s
+ 1
u(s) =
K
1
K
1
+ s
u(s)
M. Peet Lecture 6: Control Systems 21 / 23
Block Diagrams
Reduction
We now have a reduced Block Diagram
u(s)
y(s)
e(s)
-
+
1/s
K
1
/(K
1
+s)
Again, apply the Lower Feedback Interconnetion:
y(s) =
K
1
s(K
1
+s)
1 +
K
1
s(K
1
+s)
e(s) =
K
1
s(K
1
+ s) + K
1
e(s)
So the Transfer function is

T(s) =
K
1
s
2
+ K
1
s + K
1
M. Peet Lecture 6: Control Systems 22 / 23
Summary
What have we learned today?
Transfer Functions
Transfer Function Representation of a System
State-Space to Transfer Function
Direct Calculation of Transfer Functions
Block Diagram Algebra
Modeling in the Frequency Domain
Reducing Block Diagrams
Next Lecture: Partial Fraction Expansion
M. Peet Lecture 6: Control Systems 23 / 23