Você está na página 1de 28

Tutorials 1

Linearization
Laplace Transform

Linearization

Problem 3.7 Consider a system, with input u(t) and output y(t), having a
model


d
y(t) + 2 + 0.1y(t)2 y(t) = 2u(t)
dt
Assume that we associate with this system a linear nominal model
d
y(t) + 2y(t) = 2u(t)
dt
Simulate the two systems and plot the model error for
u(t) = A cos(0.5t),

A = 0.1, 1, 10

Solution
The Simulink model of the nonlinear and linear nominal model is the file
Problem3 7.mdl. It will be sent and you can explore it.

Linearization
6

10

12

14

16

18

20

Figure 1: The
between responses of nonlinear
 error 

 and linear system for u(t) =
0.1 cos 0.5t 1000 , for u(t) = 1 cos 0.5t 10 and u(t) = 10 cos 0.5t
3

Linearization

Problem 3.8 Consider the following nonlinear state space model


x 1 (t) = 2x1 (t) + 0.1x1 (t)x2 (t) + u(t)

2
x 2 (t) = x1 (t) 2x2 (t) x1 (t)
y(t)

= x1 (t) + (1 + x2 (t))

Build a linearized model around the operating point given by uQ = 1.


Solution


For any operating point x1Q , x2Q , uQ the linearized model is


d
4x1 (t) = 24x1 (t) + 0.14x1 (t)x2Q + 0.1x1Q 4x2 (t) + 4u(t)
dt
d
4x2 (t) = 4x1 (t) 24x2 (t)x21Q 2x2Q 2x1Q 4x1 (t)
dt
4y(t) = 4x1 (t) + 2 (1 + x2Q ) 4x2 (t)

Linearization

Let us find first the operating point defined by uQ = 1, to do this we have to


solve the equations
0 = 2x1Q + 0.1x1Q x2Q + uQ

(1)


0 = x1Q 2x2Q x21Q = x1Q 1 + 2x2Q x1Q

(2)

The second equation (2) has two solutions


x1Q = 0,

1 = 2x2Q x1Q

The first solution does not satisfy to (1) with uQ = 1, then the operating
point is defined by




1
1
1
x1Q = [0.1x1Q x2Q + uQ ] =
0.1
+ 1 = 0.475
2
2
2
and
x2Q

1
= 1.05263157
=
2x1Q
5

Laplace Transform

Problem 4.2 A system transfer function is given by


1s
H(s) =
(s + 1)2
Compute the time instant tu , at which the step response exhibits maximum
undershoot
Solution The Laplace transform of the system response to the step input is
Y (s) =
=

1s 1
k1
k2
k3
=
+
+
(s + 1)2 s
s
s + 1 (s + 1)2
1
2
1

s s + 1 (s + 1)2

(3)

I have found that


k1 = 1,

k2 = 1,

k3 = 2

by calling for Matlab command


residue([-1 1], [1 2 1 0])
6

Laplace Transform

Taking the inverse Laplace transform, see Table 4.1, we have


h
i
1
Y (s) = 1 et 2t et
y(t) = L
To find time for maximum undershoot let us find time moments where
y(t)
= 0, that is


t
t
t
t
2t 1
0 = y(t)
= e 2e + 2t e = e
The time derivatives becomes 0 at
t = 0.5,

t = +

while the values of y(t) in these times moments are


y(0.5) = 1 e0.5 e0.5 0.213,

y(+) = 1

Therefore, the maximum undershoot occurs at


tu = 0.5
7

Laplace Transform

Problem 4.3 The unit step response of a system with zero initial conditions
is given by
y(t) = 3 2e2t e3t , t 0
4.3.1 Compute the system transfer function
4.3.2 Compute the system response to a unit impulse
Solution We know that the Laplace transform Y (s) of the step response of
the system is
1
Y (s) = G(s)
s
Making the Laplace transform of y(t) and bring the fractions to common
denominator, we get
h
i 3
2
1
7s + 18
1

=
L y(t) =
s s+2 s+3
(s + 2)(s + 3) s
Therefore, the transfer function of the system is
7s + 18
G(s) =
(s + 2)(s + 3)
8

Laplace Transform

The Laplace transform Yim (s) of the impulse response h(t) of the system
satisfies to
Yim (s) = G(s) 1 = G(s)
Therefore to find the impulse response, we should apply the inverse Laplace
transform to Yim ,


h
i
7s + 18
h(t) = L1 Yim (s) = L1
(s + 2)(s + 3)


4
3
= L1
+
s+2 s+3
= 4e2t + 3e3t
The same result we get if differentiate y(t). Indeed,
i
d
d h
2t
3t
= 4e2t + 3e3t
y(t) =
3 2e
e
dt
dt

Laplace Transform

Problem 4.4 A nonlinear system has an input-output model given by




d
y(t) + y(t) 1 0.2y(t)2 = 2u(t)
dt
4.4.1 Determine the transfer function for the linearized model, as a function
of operating point
4.4.2 Find an operating point, where the above linear system is unstable
Solution Given an operating point (yQ , uQ ), the linearized model around this
point is then
i
h
d
2
4y(t) + 1 0.2yQ 4y(t) = 24u(t)
dt
To be unstable the coefficient
i
h
2
1 0.2yQ
should be negative or zero. For example, if yQ = 5, then
i
h
2
= 4
1 0.2yQ
10

Laplace Transform

So the linearized model becomes


d
4y(t) 44y(t) = 24u(t)
dt
Its solution is (via the Cauchy formula)
4y(t) = e4t 4y(0) + 2

e4(ts) 4u(s)ds

To determine a operating point (yQ , uQ ), one has to solve equation




2
yQ 1 0.2yQ
= 2uQ
which is also a nontrivial task here.

11

Laplace Transform

Problem 4.5 A system transfer function has its poles at


2,

2,

1 j

and its zeros at


1,

Determine the dominant poles. If the system D.C. gain is equal to 5, build the
system transfer function
Solution By definition, the dominant poles of the stable system are those
which are closer than others to the imaginary axis. In our case the poles
1 j
satisfy this condition. Therefore they are dominant.

12

Laplace Transform

The transfer function of the system is

G(s) = K

(s 1)(s + 3)
(s + 2)2 (s2 + 2s + 2)

The parameter K is not given, by the DC gain, that is the value G(0), of the
system is known
(1) 3
G(0) = K
=5
42
Then
K=

40
3

13

Laplace Transform

Problem 4.6 A system transfer function is given by


H(s) =

s+1+
(s + 1)(s + 2)

4.6.1 Compute the unit step response


Solution The Laplace transform Y (s) of the unit step response is
Y (s) =
=

s+1+ 1
k1
k2
k3
=
+
+
(s + 1)(s + 2) s
s
s+1 s+2
1+1

1 1

+
2 s s+1
2 s+2

Taking advantage of Table 4.1, pp. 6869, we get


1 2t
1+
t
e +
e
y(t) =
2
2

14

Laplace Transform

Problem 4.7 The input-output model for a system is given by


d2
d
y(t) + 12y(t) = 3u(t)
y(t)
+
7
2
dt
dt
4.7.1 Determine the system transfer function
4.7.2 Compute the unit step response with zero initial conditions
4.7.3 Repeat with initial conditions y(0) = 1, y 0 (0) = 2
Solution Applying the Laplace transform to the both sides, we get
s2 Y (s) sy(0) y 0 (0) + 7sY (s) 7y(0) + 12Y (s) = 3U (s)
or the same
(s + 7)y(0) + y 0 (0)
3
U (s) +
Y (s) = 2
s + 7s + 12
s2 + 7s + 12
Then the system transfer function is
G(s) =

3
s2 + 7s + 12
15

Laplace Transform

To calculate the step response of the system with zero initial conditions, we
should apply the inverse Laplace transform for
1
3
1
3
1
G(s) = 2
=
s
s + 7s + 12 s
(s + 3)(s + 4) s
which is

L1

3
1
s2 + 7s + 12 s

= L1

k2
k3
k1
+
+
s
s+3 s+4

1 1
1
3 1

+
4 s s+3 4s+4
3 4t
1
3t
e
+ e
4
4

= L1
=

16

Laplace Transform

Now consider the case of given nonzero initial conditions


y(0) = 1,

y 0 (0) = 2

Then the Laplace transform of the step response is


3
(s + 7)y(0) + y 0 (0)
Y (s) =
U (s) +
2
s + 7s + 12
s2 + 7s + 12
1
2 (s + 7)
3
+
=
s2 + 7s + 12 s s2 + 7s + 12
We have already considered the first term, consider now the second




s+5
k5
k4
L1 2
+
= L1
s + 7s + 12
s+3 s+4


1
2
+
= L1
s+3 s+4
= 2e3t + e4t
Summing up we get
3 4t
1
7 4t
1
3t
3t
4t
3t
+ e
+ 2e
+e
= 3e
+ e
y(t) = e
4
4
4
4
17

Laplace Transform

Problem 4.18 Find the impulse and step responses of the following linear
transfer functions
1
2
3
4

1
s2 + 2s + 1
10 s
: G(s) = 2
s + 2s + 1
0.1 s
: G(s) = 2
s + 2s + 1
1
: G(s) = 2
s + 0.2s + 1

: G(s) =

Comment on the differences observed.

18

Laplace Transform

Solution for (1) The step response of the system (1) is

y(t)

1
k2
k3
1
1 k1
=
L
+
+
s2 + 2s + 1 s
s
s + 1 (s + 1)2


1
1
1
t
t
= L1

=
1

te
s s + 1 (s + 1)2
= L1

The impulse response h(t) is related to y(t) as


h(t) =

d
y(t)
dt

So
h(t) =

i
d h
1 et tet = et et + tet = tet
dt

19

Laplace Transform
Step Response
1

0.9

0.8

0.7

Amplitude

0.6

0.5

0.4

0.3

0.2

0.1

10

15

Time (sec)

Figure 2: Step response for the system (1)


20

Laplace Transform

Solution for (2)


The step response of the system (2) is

y(t)




1
k
k
k
10

s
2
3
1
+
+
= L1
= L1 2
s + 2s + 1 s
s
s + 1 (s + 1)2


10
10
11
t
t
= L1

=
10

10e

11te
s
s + 1 (s + 1)2


The impulse response h(t) is then

h(t)

i
d h
t
t
= 10et 11et + 11tet
10 10e 11te
dt

= 11tet 10et

21

Laplace Transform
Step Response
10

Amplitude

10

15

Time (sec)

Figure 3: Step response for the system (2)


22

Laplace Transform

Solution for (3)


The step response of the system (3) is

y(t)




1
k
k
k
0.1

s
2
3
1
+
+
= L1
= L1 2
s + 2s + 1 s
s
s + 1 (s + 1)2


0.1
0.1
1.1
t
t
= L1

=
0.1

0.1e

1.1
te
s
s + 1 (s + 1)2


The impulse response h(t) is then

h(t)

i
d h
t
t
= 0.1et 1.1et + 1.1 tet
0.1 0.1e 1.1 te
dt

= 1.1 tet et

23

Laplace Transform
Step Response
0.15

0.1

0.05

Amplitude

0.05

0.1

0.15

0.2

0.25

0.3

0.35

10

15

Time (sec)

Figure 4: Step response for the system (3)


24

Laplace Transform

Solution for (4)


The step response of the system (4) is

y(t)

1
1
n
1
1
=
L
s2 + 0.2s + 1 s
s2 + 2n s + n2 s

 p
en t
= 1 p
sin n 1 2 t + ,
1 2
= L1

where = cos and the parameters , n are defined by


2n = 0.2,

n2 = 1,

n = 1,

= 0.1

25

Laplace Transform
Step Response
1.8

1.6

1.4

Amplitude

1.2

0.8

0.6

0.4

0.2

10

20

30

40

50

60

Time (sec)

Figure 5: Step response for the system (4)


26

Laplace Transform

Problem 4.22 A parallel connection of two systems is


1
s+1
Y

SYSTEM 1
3
s+4
SYSTEM 2

Figure 6: System Diagram


4.22.1
4.22.2
4.22.3
4.22.4

What is the transfer function from U to Y ?


What is the system poles?
What is the system zeroes (if any)?
Calculate the system step response

27

Laplace Transform

Solution The transfer function from U to Y can be obtained from diagram


3
1
2s 1
Y (s) =
U (s)
U (s) =
U (s)
s+4
s+1
(s + 4)(s + 1)
The poles of the system are located at
1 = 1,

2 = 4

The zero of the system is located at


= 0.5
The step response of the system is




2s 1
1
k2
k3
1
1 k1
+
+
y(t) = L
=L
(s + 1)(s + 4) s
s
s+1 s+4


1
1
2
1
2 4t
1
1
t
1
+

= +e e
= L

4s s+1 3s+4
4
3
28

Você também pode gostar