0% found this document useful (0 votes)
12 views5 pages

Exsheet 2

Uploaded by

pobisas812
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
12 views5 pages

Exsheet 2

Uploaded by

pobisas812
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Engineering Tripos Part IIB Module 4F3

Nonlinear and Predictive Control


Examples Paper 4F3/2

1. Consider the system

x˙1 = x2
x˙2 = −a1 x2 − a2 x1 − (b1 x2 + b2 x1 )2 x2

Using V (x) = a2 x21 + x22 , show that this system is globally asymptotically stable if
a1 > 0 and a2 > 0.

2. Suppose that ẋ = f (x) has an equilibrium point at x = 0. Define the matrix


 ∂f1 ∂f1 
∂x1
··· ∂xn
J(x) =  ··· .
∂fn ∂fn
∂x1
··· ∂xn

If there is a region ∆, which includes 0, and a matrix P = P T > 0, such that

−(J T (x)P + P J(x)) > 0

for all x in ∆, and 0 is the only equilibrium state in ∆, prove that x = 0 is


asymptotically stable. (Hint: Use V (x) = f T (x)P f (x).)

3. Recall that a matrix P = P T is said to be positive-definite if xT P x > 0 for all x 6= 0


and we write P > 0. If P = P T > 0 and Q = QT ≥ 0 show that

(i) P −1 > 0,
(ii) P + Q > 0,
(iii) P QP > 0 if Q > 0,
(iv) RT P R ≥ 0 for arbitrary R (of appropriate dimensions),
(v)
 
P 0
≥ 0.
0 Q

1
2

u R x2 R x1
− −

Figure 1: Linear system.

4. The linear system shown in Fig. 1 is to have feedback placed around it to speed up its
reponse and return to equilibrium from non-zero initial conditions. The input signal
u(t) is constrained in amplitude: |u(t)| ≤ U . Write the system’s state equation in
the form

ẋ(t) = Ax(t) + bu(t)

and find the symmetric matrix P such that

AT P + P A = −I.

Note that P > 0; why? Show that if the feedback law is chosen as

u(t) = −U sign(bT P x(t))

then the closed-loop system is globally asymptotically stable.

5. Find the describing function N (E) of a “relay with dead-zone”:



 −R if e < −δ
f (e) = 0 if |e| ≤ δ

+R if e>δ

and sketch its variation with E.

6. (a) If u = f1 (e) is single-valued, and

f1 (e) = f2 (e) + f3 (e),

show that

N1 (E) = N2 (E) + N3 (E)

where Ni (E) is the describing function of fi (e).


3

f (e)

δ
−3δ
2
−δ
2
e
δ 3δ
2 2

Figure 2: A-D converter.

(b) Figure 2 shows the charactersitic of an A-D converter. Using (a), find its
describing function. Assume that |f (e)| = N δ if |e| ≥ 2N2−1 δ.

7. Figure 3 shows a feedback system containing one nonlinear and one linear element.
The transfer function of the linear element is
20
g(s) = .
s(s + 1)(s + 2)

The nonlinear element is a ‘relay with dead-zone’, as defined in question 5, with


δ = 0.25 and R = 1. Use the describing function method to predict whether any
stable limit cycle exists and if so predict its frequency and amplitude.

r=0 e NONLINEAR u LINEAR y


ELEMENT ELEMENT

Figure 3: Feedback system.

8. Which of the following transfer functions are positive real or strictly positive real?
1 1
(a) 1+sT
(T > 0) (b) 1−sT
(T > 0)

1 100(s+1)
(c) s
(d) (s+10)2

sωn2
(e) s2 +ωn2 (ωn real)
4

9. Figure 4 shows the attitude control system of a satellite. The attitude sensor has a
nonlinear gain ψ(y), where

0.25 ≤ ψ(y) < β

and β can be chosen by the system designer. Use the Circle and Popov criteria
to predict the maximum value of β for which the system is globally asymptotically
stable. (Sketch the appropriate loci accurately for 1 ≤ ω ≤ 5).
satellite dynamics
controller z }| {

0 1 ẏ 1
y
5 s s
− −
rate gyro

1
attitude
filter sensor

10
1+s ψ(y)

Figure 4: Feedback system.

Would your answer change if you knew that the sensor characteristic ‘drifted’ with
time, although it remained within the limits 0.25 ≤ ψ(y, t) < β?

10. What form does the Circle Criterion take when

(i) α = 0?
(ii) α becomes negative?

11. Show that if the conditions of the Circle Criterion are satisfied, the nonlinear gain
does not vary with time, and is an odd function (ψ(−y) = −ψ(y)), then the de-
scribing function N (E) satisfies

α ≤ N (E) ≤ β

and hence that the describing function predicts that no limit cycle will exist.

I. Lestas, January 2009


(J.M. Maciejowski, January 2008)
5

Answers

4.
 5 1

4 4
1 1
4 4

5.
(
0 ifqE ≤ δ
N (E) = 4R
πE
1 − ( Eδ )2 if E > δ

6.


 0 if E ≤ δ/2

 4δ Pk n o1/2
(2i−1)δ 2
N (E) = πE i=1 1 − [ 2E
] if (2k−1)δ
2
<E≤ (2k+1)δ
2

 P n o1/2
 4δ N
 (2i−1)δ 2
i=1 1 − [ 2E ] if E > (2N2−1)δ .
πE


7. Frequency 2 rad/sec, Amplitude 4.237.

9.
Circle criterion: β ≤ 0.392 approx.
Popov criterion: β ≤ 0.6

Past Papers

The following past Tripos questions are suitable for further practice:
Part IIB Module 4F3. 2009 Q.2, 2008 Q.1, 2007 Q.1, 2006 Q.3, 2005 Q.2, 2004 Q.2 (a)–
(d). 2003 Q.1.
Part IIB Module I3. 2002 Q.1. 2001 Q.2(a), 3(a),(b). 2000 Q.1(c), 2. 1999 Q.2.
1998 Q.1, 2(a)(b), 4(b). 1997 Q.2. 1996 Q.1,2.
EIST Paper E3. 1994 Q.5(b),6. 1993 Q.4. 1992 Q.5(b),6.

You might also like