You can access complete document on following URL.

Contact me if site not loaded

https://unihelp.xyz/
sm
Chapter 1
tb9
A PRELUDE TO
CONTROL THEORY
8@
1.5 Exercises and projects
The exercises in this chapter are meant to stimulate student thinking about
interconnected systems. Students should be encouraged to use their background
knowledge in physics, signals and systems, electronics, logic design, and other
systems-level courses to piece together the concepts of time constants, transport
delays, logical elements, time-frequency relationships, and such to be able to for-
gm

mulate a high-level system description. Given this breadth for problem solving,
students are expected to develop the hierarchical structure of any system.
The fact is that there is a diversity of answers, and the advice is to let
students brainstorm their ideas. These exercises are meant to expand student
thinking about sensors and possible feedback architectures that will improve
overall controllability and observability of systems.
ail

Exercise 1 Hero’s control system is a classic system for beginning engineers

to formulate what an open-loop control structure looks like. While there is no
problem in terms of stability, the exercise demonstrates that in this, and in a
multitude of other open-loop systems, humans do indeed become the feedback
provider. In Hero’s system, the fire on the altar is the control input. The human
is the one who through the observations can control the altar fire either to open
.co

or close the temple doors.

To obtain a processing structure one needs to consider each element of the
process as a processing block. The input and output of each block needs to
be specified. If there are time delays, then such delay elements must be iden-
tified and marked as such. Where there are pulleys and spindles there must
be appropriate conversion of forces that translate linear to rotational motion,
and vice versa. Appropriate coupling equations must be shown that couple the
m

dynamics of water discharged from the kettle to increase in mass of the bucket.

Contact me in order to access the whole complete document - Email: [email protected]

WhatsApp: https://wa.me/message/2H3BV2L5TTSUF1 - Telegram: https://t.me/solutionmanual
2 CHAPTER 1. A PRELUDE TO CONTROL THEORY

The notion of time constant in the input-output behavior of subsystems must

be exploited.
A diagram showing the sequential processing of various elements of the con-
trol system results in a block diagram.

Exercise 2 In this exercise the student is expected to recognize the shortfall

in the response of open-loop systems for various open-loop examples of the text.
Once the shortfalls are recognized, the question asked is how one converts such
systems into more reliable systems that have feedback. This question is meant
to force the student towards critical thinking about how to obtain a desired
response.
Obtaining a desired response requires knowing what the plant is producing
and comparing the response with some desired quantity. The resulting error is
naturally a difference between desired and actual response. Say in the case of a
toaster oven the desired level of “toasting” can be specified by how dark or light
the toast should be, then an appropriate sensor, say a camera that can take the
picture of the toast and compare with a desired toast level can be used to yield
a more “controlled” response.
Chapter 2

MATHEMATICAL
MODELS IN CONTROL

2.10 Exercises and projects

Exercise 1 To show that a system is or is not linear we must prove that the
principle of superposition does or does not apply.
a. Suppose v (t) = v1 (t) + v2 (t). The given system is linear if and only if
d d
y (t) = v1 (t) (v1 (t)) + v2 (t) (v2 (t))
dt dt
However, given v (t) = v1 (t) + v2 (t),
d
y (t) = (v1 (t) + v2 (t)) (v1 (t) + v2 (t))
dt
d d
= v1 (t) (v1 (t) + v2 (t)) + v2 (t) (v1 (t) + v2 (t))
dt dt
d d
= v1 (t) (v1 (t)) + v1 (t) (v2 (t))
dt dt
d d
+v2 (t) (v1 (t)) + v2 (t) (v2 (t))
dt dt
Hence the system is not linear.
b. For the cascade connected system, the time-domain output q (t) = N2 ∗
N1 ∗ (v1 (t) + v2 (t)), where “∗” represents the convolution operator. The
output is
y (t) = N2 ∗ q (t) = N2 ∗ N1 ∗ v (t)
Suppose the input is v (t) = v1 (t) + v2 (t). Then
y (t) = N2 ∗ N1 ∗ (v1 (t) + v2 (t))
= N2 ∗ N1 ∗ v1 (t) + N2 ∗ N1 ∗ v2 (t)
Hence the system is linear.

3
4 CHAPTER 2. MATHEMATICAL MODELS IN CONTROL

Exercise 2

a. Consider a complex sinusoidal input signal v(t) = Aej( t+φ) in which

the term Aejφ represents the phasor value of v(t), and Aej( t) is the
sinusoidal carrier of v(t). Expressing the complex sinusoid in terms of real
and imaginary quantities yields

v(t) = Re A cos( t + φ) + Im A sin( t + φ)

For a linear system the input v(t) produces an output y(t). Let y(t) =
Bej( t+θ) . Expressing the output in terms of real and imaginary quanti-
ties yields
y(t) = Re B cos( t + θ) + Im B sin( t + θ)
We can conclude therefore that an input Re v(t) = Re A cos( t + φ) pro-
duces an output Re y(t) = Re B cos( t + θ).

b. The response to an input v(t) is y(t). For a linear system this can be
expressed as y(t) = Av(t), where A is a linear operator. Therefore

dy(t) dv(t)
=A
dt dt

Exercise 3

(i) For the system

∙ ¸ ∙ ¸∙ ¸ ∙ ¸ ∙ ¸
ẋ1 1 0 x1 1 £ ¤ x1
= + u y= 2 1
ẋ2 2 2 x2 0 x2

we have the matrices

∙ ¸ ∙ ¸
1 0 1 £ ¤
A= B= C= 2 1
2 2 0

(a) For this system the controllability matrix is

∙ ¸
1 1
C0 = [B AB] =
0 2

A system is fully controllable if and only if this matrix is of full rank,

in this case rank 2 since the system is of second-order. A matrix has
full rank if and only if its determinant is not zero. We see that
∙ ¸
1 1
det C0 = det = 2 6= 0
0 2

and hence the system is fully controllable.

 5

(b) For this second-order system, the observability matrix is

∙ ¸
£ ¤ 2 4
O0 = [C CA]T = C T AT C T =
1 2

A system is fully observable if and only if the observability matrix is

of full rank, in this case 2 since the system is of second-order. This
matrix has full rank if and only if the determinant is non-zero. We
can either compute the determinant or rank, or both. Observing that
∙ ¸
2 4
det O0 = det =0
1 2

or that Rank [O0 ] = 1 < 2, proves the system is not fully observable.
Note: Finding the rank of the matrix does not by itself explain why
the system is not observable. To understand the reason why, we need
to examine the transfer function as requested in (c).
(c) The transfer function of a linear time-invariant system is given by

Y (s) = [C] [sI − A]−1 B + D

where A, B, C, and D are the matrices obtained from the state and
output equations, and I is the identity matrix. In this system, as
well as in the others, D = 0.
½∙ ¸ ¸¾−1 ∙ ¸
∙
£ ¤ s 0 1 0 1
Y (s) = 2 1 − +0
0 s 2 2 0
½∙ ¸¾−1 ∙ ¸
£ ¤ s−1 0 1
= 2 1
−2 s−2 0
⎧⎡ ⎤⎫
⎪ (s − 2) ⎪
⎪
⎨ 0 ⎪∙ ¸
£ ¤ ⎢ (s − 1)(s − 2) ⎥⎬ 1
= 2 1 ⎢ ⎥
⎪ ⎣ 2 (s − 1) ⎦⎪ 0
⎪
⎩ ⎪
⎭
(s − 1)(s − 2) (s − 1)(s − 2)
⎧⎡ ⎤⎫
⎪ (s − 2) ⎪
⎪ ⎪
£ ¤ ⎨⎢ (s − 1)(s − 2) ⎥⎬
= 2 1 ⎢ ⎥
⎪ ⎣ 2 ⎦⎪
⎪
⎩ ⎪
⎭
(s − 1)(s − 2)
2(s − 2) 2 2(s − 2) + 2
= + =
(s − 1)(s − 2) (s − 1)(s − 2) (s − 1)(s − 2)
2(s − 1) 2
= =
(s − 1)(s − 2) (s − 2)

The pole-zero cancellation makes the system clearly unobservable for

the pole located at s = 1.
6 CHAPTER 2. MATHEMATICAL MODELS IN CONTROL

(d) To determine whether or not the system is stable, we need to examine

the roots of the characteristic polynomial. If all the roots lie in the
left-half s-plane, the system is stable. If not, the system is unstable.
It is important to note that if even one root of the characteristic
polynomial lies in the right-half s-plane, the system is unstable.
To find the characteristic polynomial, we look at the equation

det [λI − A] = 0

Now
½∙ ¸ ∙ ¸¾
λ 0 1 0
det [λI − A] = det −
0 λ 2 2
½∙ ¸¾
λ−1 0
= det = (λ − 1) (λ − 2)
−2 λ−2

which gives us the equation

(λ − 1) (λ − 2) = 0

Since λ = 1, and λ = 2, both roots lie in the right-half s-plane

showing that the system is unstable.

(ii) For the system

⎡ ⎤ ⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ẋ1 0 1 0 x1 0 £ ¤ x1
⎣ẋ2 ⎦ = ⎣ 0 0 1 ⎦ ⎣x2 ⎦ + ⎣1⎦ u y = 1 0 1 ⎣x2 ⎦
ẋ3 0 −2 −3 x3 1 x3

we have the matrices

⎡ ⎤ ⎤ ⎡
0 1 0 0 £ ¤
A=⎣ 0 0 1 ⎦ B=⎣ 1 ⎦ C= 1 0 1 D=0
0 −2 −3 1

For third- and higher-order systems it is convenient to use a computer

algebra system to perform the computations. The following Matlab code
can be used to obtain the solutions.
% Define the A, B, C, and D matrices
A=[0 1 0;0 0 1;0 -2 -3];
B=[0;1;1];
C=[1 0 1];
D=[0];
% Set up the controllability matrix
CO=[B A*B (A^2)*B];
% Test for Full Controllability
% You can either use the determinant or directly compute the rank
tCO1=det(CO);
 7

tCO2=rank(CO);
% Set up the observability matrix
OB=[C’ A’*C’ (A’^2)*C’];
% Test for full observability
% You can either use the determinant or directly compute the rank
tOB1=det(OB);
tOB2=rank(OB);
% Obtain the numerator and denominator coefficients of the transfer function
[num,den]=ss2tf(A,B,C,D,1);
% To test the stability compute the eigenvalues.
% Eigenvalues are the roots of the characteristic equation
e=eig(A);

For this system the answers are as follows:

⎡ ⎤
0 1 1
(a) Controllability matrix C0 = ⎣ 1 1 −5 ⎦, det C0 = −24 (6= 0),
1 −5 13
rank C0 = 3. Therefore the system is fully controllable.
⎡ ⎤
1 0 1
(b) Observability matrix O0 = ⎣ 0 −1 6 ⎦; det O0 = 10 (6= 0);
1 −3 8
rank)o = 3. Therefore the system is fully observable.
s2 − s + 4 s2 − s + 4
(c) Transfer function Y (s) = = .
s3 + 3s2 + 2s s (s + 1) (s + 2)
(d) The eigenvalues are λ = 0, λ = −1, λ = −2, and therefore the system
is stable.

(iii ) For the linear time-invariant system

⎡ ⎤ ⎡ ⎤⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ẋ1 1 0 0 x1 1 £ ¤ x1
⎣ẋ2 ⎦ = ⎣ 0 1 1 ⎦ ⎣x2 ⎦ + ⎣0⎦ u y = 0 1 1 ⎣x2 ⎦
ẋ3 0 −2 −1 x3 1 x3

we have the matrices

⎡ ⎤ ⎡⎤
1 0 0 1 £ ¤
A=⎣ 0 1 1 ⎦ B=⎣ 0 ⎦ C= 0 1 1 D=0
0 −2 −1 1
⎡ ⎤
1 1 1
(a) The controllability matrix C0 = ⎣ 0 1 0 ⎦ has determinant
1 −1 −1
−2 (6= 0) and rank 3; therefore the system is fully controllable.