Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
14 views

digital control homework2

Uploaded by

talha manzoor
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
14 views

digital control homework2

Uploaded by

talha manzoor
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

EE 561: Digital Control Systems

Problem Set # 2
Spring 2020

Due date: May 21, 2020

You are required to collect the MATLAB code for all plots in a single m file
and upload it along with your solution on LMS.

Question # 1 (10 marks)


Consider the following discrete-time system

        
x1 (k + 1) 0 1 x1 (k) 0   x1 (k)
= + u(k); y(k) = 1 0
x2 (k + 1) −0.16 −1 x2 (k) 1 x2 (k)

Assuming full state feedback, determine the matrix K such that if u(k) = −Kx(k), then
the poles of the closed loop system are located at exactly z = 0.5 ± j0.5.

Question # 2 (15 marks)


Consider the following continuous-time open-loop system

Y (s) 1
G(s) = = 2
U (s) s (s + 2)

a) Find the equivalent state space representation in the discrete time by assuming a zero
order hold and a sample time of 0.1 sec.

b) Find the gain matrix Lp for a predictor estimator for the given system if the settling
time for the error of the estimator is required to be less than 0.5 seconds.

c) Verify that the settling time for the estimator error is indeed less than 0.5 seconds by
plotting the error response for any initial condition of your choosing.

Question # 3 (15 marks)


Consider the following system

        
x1 (k + 1) 1 T x1 (k) T 2 /2   x1 (k)
= + u(k); y(k) = 1 0
x2 (k + 1) 0 10 x2 (k) T x2 (k)

a) Design a full-state prediction observer of the system where it is desired that all observer
poles be zero.
b) Write down the dynamics for the estimator error. Then assume that the initial state
x(0) = [a1 b1 ]T and the initial estimate x̄(0) = [a2 b2 ]T . Show that the estimation error
goes to zero in at most two sampling periods. (hint: solve the dynamics of the error
recursively, to compute x̃(2)).

c) A response such as the one obtained above that attains a value of exactly zero in finite
time is called a dead-beat response. The concept of a dead-beat response is unique to
discrete-time systems and has no counterpart in continuous time. This can only happen if
the response dynamics are given by a nilpotent matrix.
Show that a matrix with every eigenvalue zero (such as the one designed above), is nilpo-
tent.

Question # 4 (20 marks)


1
Consider the following state space description of the discretized double integrator 2 pre-
s
ceded by a Z.O.H and a sampling time of T = 0.2 secs
        
x1 (k + 1) 1 0.2 x1 (k) 0.02   x1 (k)
= + u(k); y(k) = 1 0
x2 (k + 1) 0 1 x2 (k) 0.2 x2 (k)

a) Determine whether the system is controllable and observable.

b) Assuming full state feedback, determine the matrix K such that if u(k) = −Kx(k),
then the poles of the closed loop system exhibit a damping ratio ζ = 0.486 and a natural
frequency ωn = 3.364.

c) Notice that the state x1 (k) is being measured by the sensors directly. Keeping this is
mind, determine the gain matrix Lr for a reduced order observer of the system so that all
poles of the estimator error are placed at the origin.

d) Determine the pulse transfer function of the compensator consisting of the combined
control law and estimator designed above.

Question # 5 (20 marks)


Consider the system
        
x1 (k + 1) 0 1 x1 (k) 0   x1 (k)
= + u(k); y(k) = 1 0
x2 (k + 1) −0.16 −1 x2 (k) 1 x2 (k)
Assume that the system is designed to track a reference input and the control scheme is
given as u(k) = N r(k) − Kx(k), where N and K are design parameters (assume r(k) is a
scalar).
a) Determine the gain matrix K, if the desired characteristic polynomial of the closed-loop
poles is αc (z) = z 2 − z + 0.5. You may assume that r(k) = 0 in this step.

b) Now, after incorporating the control law u(k) = N r(k)−Kx(k) with the gains K chosen
as above, determine the transfer function Y (z)/R(z).

c) Using the result of part b, and the final value theorem in discrete-time, choose the design
parameter N so that the steady-state value of the step-response is unity (the steady state
error is zero).

Question # 6 (20 marks)


Consider the following pulse transfer function for a discrete time plant

Y (z) z −2 + 0.5z −3
=
U (z) 1 − z −1 + 0.01z −2 + 0.12z −3

a) Obtain the state space description for the system in the control canonical form.

b) Show that the system exhibits a non-zero steady state error to a step input.

c) In order to track inputs described by step functions, we implement an integral control


via state augmentation. Determine the integral gain constant KI and the state feedback
gain matrix K such that all poles of the closed loop system are placed at the origin.

The following questions have been taken from the end of chapter problems of
your textbook

Question # 7 (25 marks)


Problem 8.13

Question # 8 (25 marks)


Problem 8.14 parts (a)-(d)

Question # 9 (30 marks)


Problem 8.24 parts (a)-(d)

Question # 10 (20 marks)


Problem 8.25

You might also like