You are on page 1of 4

Examination

Control Systems Theory and Design


06.09.2011

Problem 1 (25 Marks)

z
R i

u m
b

Figure 1: Electrohydraulic servovalve

A simplified model of an electrohydraulic servovalve is shown in Figure 1; the dynamic


equations are given as

d2 d
m 2
z(t) + b z(t) = ci(t)
dt dt ,
d u(t)
τ i(t) + i(t) =
dt R
where m, b, c, τ and R are the physical parameters of the system.

a) Determine a state space representation of the system with input voltage u(t) as
input and position z(t) as output.
In the following parts of this problem, suppose m = b = c = τ = R = 1.

b) Is the system stable? Give reasons.

c) Is the system controllable? Give reasons.

d) What is the dimension of the observable subspace if only position z(t) can be mea-
sured?

e) What is the dimension of the observable subspace if only current i(t) can be mea-
sured?

f) Suppose that the system matrices A,b,c,d are available in the workspace of MAT-
LAB. Give MATLAB code for checking the observability and controllability of the
system.

1
Problem 2 (25 Marks)

The following state space realization of a system is given as

ẋ = Ax + Bu
y = Cx + Du

where
   
−3 −2 0 2 0    
1 2 0 0 3
A= 1 0 0 , B =  0 0 , C= , D= .
0 0 1 0 0
0 0 −3 0 1

a) Obtain the transfer function matrix representing this system.


 
s+3 2 0
Hint: Note that the inverse of a matrix  −1 s 0  is
0 0 s+3

 
s(s + 3) −2(s + 3) 0
1  (s + 3) (s + 3)2 0 .
(s + 1)(s + 2)(s + 3)
0 0 (s + 1)(s + 2)

b) Is the given state space model minimal? Give reasons.

c) Does the system have a transmission zero? Give reasons.

d) Suppose that the system matrices A,B,C,D are available in the workspace of MAT-
LAB.

i) Give MATLAB code to generate weighting matrices with appropriate dimen-


sions to be used for linear quadratic optimal state feedback and state estimator
design.
ii) Give MATLAB code to calculate linear quadratic optimal state feedback and
state estimator gains.

e) If the measurement noise is too high, which weighting matrix should be modified
and how?
How does this affect the speed of estimation?

2
Problem 3 (25 Marks)

a) Consider a discrete-time system G(z) represented by the transfer function

z+3
G(z) =
z2 + 3z + 2

and the ramp input signal u(k) = kσ(k).

i) Is the system G(z) stable?


ii) Compute the response of the system G(z) to the input signal u(k) for
k = 0, 1, ... , 4.

b) Assume the system in (a) was generated by sampling the input and output signal
of a continuous-time system. What will happen with the poles of the discrete-time
system if the sampling time approaches zero?

c) Sketch the lines of constant real parts of the poles of a second order system in the
s-plane and map them into the z-plane.

d) i) Sketch a block diagram of the system G(z) which can be used to derive the
observer canonical form assuming that u(k) is the input and y(k) is the output
of G(z).
ii) Derive a state space model in terms of the matrices Φ, Γ, C and D in observer
canonical form for G(z).

e) Give MATLAB code to plot the response of the system G(z) to the input u(k) as
given in problem (a.ii).

3
Problem 4 (25 Marks)

Given is the discrete-time system G(z) represented by the following state-space model

xk+1 = Φxk + Γuk


yk = cxk ,

where    
0.5 1 1
Φ= , Γ= , c = [ 1 1] .
0 0.2 0

a) Is the system G(z) controllable and observable?

b) What is the rank of the Hankel matrix

g(1) g(2) g(3) · · · g(k)


 
g(2) g(3) ··· g(k + 1) 
 .
..

Hk = g(3) g(k + 2)  ,
 
 . ..
 ..

. 
g(k) g(k + 1) · · · g(2k − 1)

where g(k) represents the impulse response of the system G(z)?

c) i) Let k = 6. Rewrite the Hankel matrix H6 in terms of the matrices Φ, Γ and c


of the state space model of the system G(z).
ii) Which steps are required to determine the extended observability and control-
lability matrices if the Hankel matrix H6 is given and the measurements are
assumed to be ideal?
iii) How can the system matrix be computed using the extended controllability
and observability matrices?

d) Assume that the Hankel matrix Hk given in part b) is in the workspace of MATLAB
as Hk. Give MATLAB code to calculate the input and output matrices.

Hint: Use your knowledge of the rank of Hk obtained in part b).

You might also like