Prüfung Regelungstechnik I (Control Systems I)

Prof. Dr. Lino Guzzella

29. 8. 2011

Übersetzungshilfe / Translation aid (English)

To be returned at the end of the exam!

Do not mark up this translation aid - The German exam is the only valid version!

All answers must be written on the regular exam sheets (which are in German).

1
Question 1 (Modeling and Linearization) 8 Points

The figure below illustrates a capacitor with variable distance between the plates. The lower
plate of the capacitor is fixed. The upper plate of mass m is movable and connected to two
springs and a damper. The springs are equal and their string stiffness is k. At the position
y = y0 , the springs are in their zero force position. The damper force is proportional to the
velocity of the plate. The damping coefficient is b. The upper plate is attracted by the lower
plate with the force

c · U2
FEF = ,
4 · y2
where U is the voltage across the capacitor. The position y is measured by an appropriate
sensor.

m
E-field
U y

a) (3 points) Determine the differential equations which describe the vertical motion of the
upper plate. Select appropriate input, output and state variables. Give the system equa-
tions in the standard form, i.e., as a system of nonlinear first-order differential equations
of the form

ż(t) = f (z(t), v(t)), w(t) = h(z(t), v(t)), z(t) ∈ R2 , v(t), w(t) ∈ R.

b) (2 points) Determine the voltage Ue which keeps the upper plate at equilibrium at the
position ye = y0 /2. Note: y0 > m · g/k.

c) (3 points) Linearize the system equations at the position ye = y0 /2 (a normalization is

omitted). Give the system equations in the standard form, i.e., in state-space representa-
tion {A, b, c, d}.

1 / 12
Question 2 (Frequency domain, time domain) 8 Points

The open-loop transfer functions (loop gain) (L1 (s) , L2 (s) , L3 (s) , L4 (s)) of 4 control systems
are given. Further the Nyquist plots (diagrams A, B, C und D) of these transfer functions, and
the resulting step responses (step responses 1 bis 4) of the corresponding closed loop systems
are given.
Assign the correct nyquist plot and the correct step response to each of the open loop transfer
functions. Use the table provided below for your solution. You do not need to justify your
answers.
Credits:
Per correct assignment: +1 credit
Per incorrect assignment: −1 credit
Minimum amount of credits for the whole question: 0 credits

Table for solution

L1 (s) = L2 (s) = L3 (s) = L4 (s) =
Transfer functions
1 1 −0.5s+1 2s+1
s2 +s+1 (3s)·(s2 +s+1) s2 +s+1 s2 +s+1

Nyquist plot
(open loop)
Step response
(closed loop)

Nyquist plot A Nyquist plot B

1 1.5

1
0.5
0.5
Im

Im

0
0

-0.5 -0.5

-1
-1
-1.5
-1 -0.5 0 0.5 1 -1 0 1 2
Re Re

2 / 12
 Nyquist plot C Nyquist plot D

1 1

0.5 0.5
Im

Im
0 0

-0.5 -0.5

-1 -1

-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1

Re Re

step response 1 step response 2

1 1

0.8 0.8
amplitude [-]
amplitude [-]

0.6 0.6

0.4 0.4

0.2 0.2

0 0

0 5 10 15 0 5 10 15
time [s] time [s]

step response 3 step response 4

1 1

0.8 0.8
amplitude [-]
amplitude [-]

0.6 0.6

0.4 0.4

0.2 0.2

0 0

0 5 10 15 0 5 10 15
time [s] time [s]

3 / 12
Question 3 (Controller Design) 10 Points

The following plant is given

1
P (s) =
s (s + 1) (s + 2)

Your supervisor asks you and your colleague to design a controller for the given plant. You
suggest to use a PID controller and to design its parameters with the Ziegler/Nichols method.

a) (4 points) Determine the critical gain kp∗ and the critical frequency ω ∗ .

b) (1 point) Now calculate according to the Ziegler/Nichols method the parameters kp , Ti

and Td of the following PID controller

 
1
CP ID (s) = kp · 1 + + s · Td
s · Ti

Afterwards, your colleague computes with MATLAB the phase margin of the control system of
your controller. This poor margin turns out to be 24.8 ◦ . Since your colleague wants to have
more phase margin, he suggests to use instead of CP ID (s) the following controller

C2 (s) = 2 · s · (2 − s)

c) (5 points) Since your supervisor wants to be sure that the phase margin is larger with
C2 (s), she asks you to compute the phase margin of the control system C2 (s) · P (s).

4 / 12
Question 4 (Laplace-Transformation) 9 Points

The subtasks a) and b) can be solved independently.

a) (4 points) Determine the time-domain output signal y(t) of the system P (s)
s+2
P (s) =
(s2 + 2 · s + 1) · (s + 3)

for the following input signal u(t)

u(t) = h(t).

b) A linear time-invariant SISO system is excited by

u(t) = h(t).

Its step response in the time-domain is

 
−t 1
y(t) = 1 − e · (cos(2 · t) + · sin(2 · t)) · h(t).
2

i) (3 points) Calculate the transfer function Σ(s) of the SISO system.

ii) (1 point) Determine the poles of the system.
iii) (1 point) Approximate the rise time t90 as well as the overshoot ǫ̂ of the step response
by respecting only the relevant system dynamics.

5 / 12
Question 5 (Constraints) 10 Points

Figure 1 shows a sketch of the system to be analyzed. A ferromagnetic solid sphere is posi-
tioned in the magnetic field of an electromagnet. The magnitude of this field can be adjusted
very rapidly by an amplifier. Therefore, the force acting on the solid sphere is assumed to be
proportional to the input u(t).

u(t)

y(t)

Figure 1: Floating ferromagnetic solid sphere in a magnetic field.

A controller C(s) (whose design is not a part of this question) has to keep the ferromagnetic
sphere levitating in the magnetic field. The position y(t) of the sphere is measured by a photo
cell. The measurement y(t) is corrupted by the electromagnetic radiation that is produced by
the amplifier. The lowest frequency of the noise signal is 50 Hz ≈ 300 rad/s. Figure 2 shows the
set-up of the control system.

r(t) u(t) y(t)

C(s) P (s)
−
n

Figure 2: Control system with input and output signals.

The dynamic behavior of the plant can be represented perfectly by the following state space
model:
0 1 0
" # " #
d
x(t) = · x(t) + · u(t) (1)
dt 900 0 9000

where x1 (t) = y(t) is the position and x2 (t) is the velocity of the ferromagnetic solid sphere.

a) (1 point) Determine the transfer function P (s) of the plant with the input signal u(t) and
the output signal y(t).

b) (4 points) Sketch in the Bode-diagram provided on next page the magnitude |P (j ω)| and
phase plot ∠P (j ω) of the plant transfer function. Tip: Determine the asymptotes for very
low and very high frequencies.

c) (1 point) Indicate in the Bode-diagram the frequency range where the noise signal n(t) is
present.

d) (4 points) Before starting with the design of the controller C(s), answer first the question,
whether a controller exists that can full fill the requirements. What are your considera-
tions? What is your answer? Justify your answer mathematically.

6 / 12
|P (j ω)|dB
20

-20

-40

1 2
10 10

ω rad/s

∠P (j ω)◦
-20

-60

-100

-140

-180

1 2
10 10

ω rad/s

Figure 3: Bode-Diagram for representing the solution.

7 / 12
Question 6 (Nyquist-plot, Nyquist-theorem) 8 Points

For a plant P (s) the following Bode plot was measured.

Bode Diagram

40
Magnitude [dB]

-40

-80

-120
10−2 10−1 100 101 102

-90
Phase [deg]

-180

-270
10−2 10−1 100 101 102
Frequency [rad/s]

Figure 4: Bode plot of the system

a) (2 points) Draw the Nyquist plot of the plant P (s) qualitatively in the provided figure.
Thereby, use the information that

lim P (jω) = −4 − j · ∞.
ω→0

b) (2 points) Identify the transfer function of the plant P (s) using the Bode plot.

c) (1 point) You want to control the plant using a P-controller C(s) = kp mit kp = 1. What
are the gain and the phase margins of the resulting closed loop system.

d) (1 point) Use the Nyquist theorem in order to find the values of kp , for which the system
is stable.

e) (2 points) You want to improve the step response of the closed loop system. Therefore,
you use a PD-controller with the transfer function

C(s) = 0.5 + 0.5s.

What is the gain margin of the resulting closed loop system?

8 / 12
 Nyquist Plot
4

1
Imaginary Axis

-1

-2

-3

-4
-4 -3 -2 -1 0 1 2 3 4
Real Axis
Figure 5: Nyquist plot of the plant P (s)

9 / 12
Question 7 (System Analyses, Block Diagram) 8 Points

A state-space representation is given

ẋ(t) = A · x(t) + b · u(t)

y(t) = c · x(t) + d · u(t)

with
   
0 0 1 0  
A =  0 −1 0  , b =  1 , c= 1 −1 0 , d=0
−1 0 2 0

a) (2 points) In the sense of Lyapunov, is the system stable, asymptotically stable, or unsta-
ble? Justify your answer mathematically.

b) (1 point) Is the control system completely controllable? Justify your answer mathemati-
cally.

c) (1 point) Is the control system completely observable? Justify your answer mathematically.

d) (2 points) The detailed flow chart of a different control system is given with the input
signal u(t) and the output signal y(t) (Fig. 6). Derive the state-space matrices {A, b, c, d}
for the given system!
y(t)
-3

u(t) R x1 (t) R x2 (t) R x3 (t)

Figure 6: Signal flow chart, u(t) = input signal, y(t) = output signal

e) (2 points) The transfer function Σ(s) is given, describing the input/output behavior of
another linear dynamic system

s2 + 3s + 2
Σ(s) =
s3 + 4s2 + s + 5
Determine the system matrices {A, b, c, d} of an accordant state-space representation.

10 / 12
Question 8 (Multiple-Choice) 8 Points

Decide whether the following statements are true or false and check the corresponding check box
with an X (
×).
You are not required to justify your answers. All questions are equally weighted (1 point).
There will be a reduction of one point for a wrong answer 1 . Unanswered questions will get 0
points. The minimum sum for all questions is 0 points.

a) The differential equation δẋ = −19·δx+12·δu represents the linearization of the non-linear
system ẋ = −7 · x2 − 5 · x + 3 · u2 around the equilibrium point {xe = 1, ue = 2}.

 True  False

b) A constant signal u(t) = 1 at the input of a system with the transfer function Σ(s) =
s−4
s2 +3s+2 produces for t → ∞ a constant output signal of −2.

 True  False

c) The following state space model {A, b, c, d} represents a realization for a system with the
s+3
transfer function Σ(s) = s2 −5s−7 :
   
0 1 0    
A= , b= , c= 3 1 , D= 0
−7 −5 1

 True  False

2
d) An unstable system with the transfer function P (s) = s(s−3) can be stabilized in a closed
loop control system by a PD-controller C(s) = kp + kd · s with the parameters kp = 2 and
kd = 1.5.

 True  False

1
e) A plant with the transfer function P (s) = s(s+5) (an integrator 1s and a first-order element
1
(LP-1) s+5 in series) is controlled by a P-controller (kp > 0). There is a disturbance signal
w at the input of the plant (see figure below).

+ +
r kp P s! y
–

The integrator of the plant will guarantee that a constant disturbance w(t) = h(t) will be
completely rejected for r(t) = 0 and t → ∞, i.e. lim y(t) = 0.
t→∞

 True  False

1
Be aware of this fact!

11 / 12
f) A PI-controller C(s) = kP (1 + T1i s ) is used to control an asymptotic stable system. For the
adjusted controller parameters kp (gain of the proportional part) and Ti (time constant
of the integral part), the output signal of the control system shows a harmonic oscillation
(control system is critically stable). By increasing the time constant Ti of the integral part,
the control system will become asymptotically stable .

 True  False

g) An asymptotically stable control system has at least a guaranteed gain margin of k < 2 if
its sensitivity function fulfills the condition max |S(jω)| < 2.
ω

 True  False

h) The open loop gain of a control system L(s) = C(s) · P (s) has two unstable poles. The
Nyquist-plot L(jω) is shown below for ω von −∞ bis +∞. It can be cocluded that he
closed loop control system is asymptotically stabel.

L(jω)
2

1
Im

0
−1

−1

−2

−3
−2 −1 0 1 2 3
Re

 True  False

12 / 12