Lecture Notes

Introduction to Control Systems

Instructor: Dr. Huynh Thai Hoang

Department of Automatic Control
Faculty of Electrical & Electronics Engineering
Ho Chi Minh City University of Technology
Email: hthoang@hcmut.edu.vn
huynhthaihoang@yahoo.com
H
Homepage: www4.hcmut.edu.vn/~hthoang/
4h t d / hth /

16 November 2011 H. T. Hoang - www4.hcmut.edu.vn/~hthoang/ 1

 Chapter 8

ANALYSIS OF
DISCRETE CONTROL SYSTEMS

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 2

 Content
Stability conditions for discrete systems
Extension of Routh-Hurwitz criteria
Jury
J criterion
it i
Root locus
Steady
S d state error
Performance of discrete systems

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 3

 Stability conditions for discrete systems

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 4

 Stability conditions for discrete systems
A system is defined to be BIBO stable if every bounded
input to the system results in a bounded output.
I s
Im I z
Im

Stable
Re s Stable Re z
1
Re{s} < 0 | z |< 1
z = eTs

The region of stability for a The region of stability for a

continuous
ti system
t i the
is th di
discrete
t system
t is i the
th interior
i t i
left-half s-plane of the unit circle
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 5
 Characteristic equation of discrete systems
Discrete systems described by block diagram:
R(s) Y(s)
+ GC(z) ZOH G(s)
T

()
H(s)

Characteristic equation: 1 + GC ( z )GH ( z ) = 0

Discrete systems described by the state equation

x( k + 1) = Ad x( k ) + Bd r ( k )

y ( k ) = Cd x( k )
Characteristic equation: det( zI Ad ) = 0
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 6
 Methods for analysis the stability of discrete systems
Algebraic stability criteria
The extension of the Routh-Hurwitz criteria
Jurys
J stability
t bilit criterion
it i
The root locus method

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 7

 The extension of the Routh-
Routh-Hurwitz criteria

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 8

 The extension of the Routh
Routh--Hurwitz criteria
Characteristic equation of discrete systems:
a0 z n + a1 z n 1 + L + an = 0
Im z I w
Im

Region of Region of
stability Re z stability
Re w
1 1+ w
z=
1 w

The extension of the Routh-Hurwitz criteria: transform

zw, and then apply the Routh Hurwitz criteria to the
characteristic equation of the variable w.
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 9
 The extension of the Routh-
Routh-Hurwitz criteria Example

A l
Analyze th stability
the t bilit off the
th following
f ll i system:
t
R(s) + Y(s)
ZOH G(s)
T = 0.5

H(s)
3e s 1
Gi
Given t G( s) =
that:
th H ( s) =
s+3 s +1
Solution:
S l ti
The characteristic equation of the system:
1 + GH ( z ) = 0
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 10
 The extension of the Routh-
Routh-Hurwitz criteria Example (cont)

s
G ( s ) H ( s ) 3 e
GH ( z ) = (1 z 1 )Z G ( s) =
s ( s + 3)
s
1 3e
3 e 1
= (1 z )Z H (s) =
s ( s + 3)( s + 1) ( s + 1)
1 2 z ( Az + B)
= 3(1 z ) z
( z 1)( z e 30.5 )( z e 10.5 )
(1 e 30.5 ) 3(1 e 0.5 )
A= = 0.0673
3(1 3) 1 z ( Az + B)
Z =
s3(s0+.5 a)( s + b) ( z 1)( z e aT )( z e bT )
3e 30.5 (1 e 0.5 ) e 0.5 (1 e ) aT
B= b(1 e = 0).0346
a(1 e bT )
3(1 3) A=
ab(b a)

0 .202 z + 0. 104
aeaT (1 e bT ) be bT (1 e aT )
GH ( z ) = 2
z ( z 0.223)( zB = 0.607) ab(b a)
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 11
 The extension of the Routh-
Routh-Hurwitz criteria Example (cont)

The characteristic equation:

1 + GH ( z ) = 0
0.202 z + 0.104
1+ 2 =0
z ( z 0.223)( z 0.607)
z 0.83z + 0.135z + 0.202 z + 0.104 = 0
4 3 2

1+ w
Perform the transformation: z =
1 w
1+ w 1+ w 1+ w 1+ w
4 3 2

0.83 + 0.135 + 0.202 + 0.104 = 0

1 w 1 w 1 w 1 z w
0.202 + 0.104
GH ( z ) = 2
G
z ( z 0.223)( z 0.607)
1.867 w + 5.648w + 6.354 w + 1.52 w + 0.611 = 0
4 3 2

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 12

 The extension of the Routh-
Routh-Hurwitz criteria Example (cont)

The Routh table

Conclusion: The system is stable because all the terms

in the first column of the Routh table are positive.
1.867 w4 + 5.648w3 + 6.354 w 2 + 1.52 w + 0.611 = 0
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 13
 Jury stability criterion

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 14

 Jury stability criterion
Analyze the stability of the discrete system which has the
characteristic equation: a0 z n + a1 z n 1 + L + an 1 z + an = 0

Jury table: consist of (2n+1) rows.

The first row consists of the coefficients of the
characteristic polynomial in the increasing index order.
order
The even row (any) consists of the coefficients of the
previous row in the reverse order.
The odd row i = 2k+1 (k1) consists (nk+1) terms, the
term at the row i column j defined by:
1 ci 2,1 ci 2,n j k +3
cij =
ci 2,1 ci 1,1 ci 1,n j k +3

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 15

 Jury stability criterion (cont)

Jury criterion statement: The necessary and sufficient

condition for the discrete system to be stable is that all the
fi t terms
first t off the
th odd
dd rows off the
th Jury
J t bl are positive.
table iti

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 16

 Jury stability criterion Example
Analyze the stability of the system which has the
characteristic equation: 5 z 3 + 2 z 2 + 3z + 1 = 0
Solution: Jury table
Row 1
Row 2
Row 3

Row 4
Row 5

Row 6
Row 7

Since all the first terms of the odd rows are positive, the system is stable.
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 17
 The root locus of discrete systems

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 18

 The root locus (RL) method

RL is
i a sett off all
ll the
th roots
t off the
th characteristic
h t i ti equation
ti off
a system when a real parameter changing from 0 +.
Consider a discrete system which has the characteristic
equation:
N ((zz )
1+ K =0
D( z )

N ( z)
Denote: G0 ( z ) = K
D( z )
Assume that G0(z) has n poles and m zeros.
The rules for construction of the RL of continuous system
can be applied to discrete systems, except the step 8.
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 19
 Rules for construction of the RL of discrete systems

Rule
R l 1:
1 The
Th number b off branches
b h off a RL = the
th order
d off the
th
characteristic equation = number of poles of G0(z) = n.
Rule 2:
For K = 0: the RL begin at the poles of G0(z).
As K goes to + : m branches of the RL end at m zeros of
G0(z), the nm remaining branches goes to approaching
the asymptote defined by the rule 5 and rule 6.
Rule 3: The RL is symmetric
y with respect
p to the real axis.

Rule 4: A point on the real axis belongs to the RL if the

totall number l andd zeros off G0(z)
b off poles ( ) to its
i right
i h is
i odd.
dd
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 20
 Rules for construction of the RL of discrete system (cont)

Rule 5: The angles between the asymptotes and the real

axis are given by: ( 2l + 1)
= (l = 0,1,2,K)
nm
Rule 6: The intersection between the asymptotes and the
real axis is a point A defined by:
n m

l zeros
poles pi zi (pi and zi are
poles and zeros
OA = = i =1 i =1
nm nm of G0(z) )

Rule 7: : Breakaway / break-in points (or

dK
break p
points for short),
), if any,
y, are located in =0
d
dz
the real axis and are satisfied the equation:
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 21
 Rules for construction of the RL of discrete system (cont)
Rule 8: The intersections of the RL with the unit circle can
be determined by using the extension of the Routh-Hurwitz
g z=a+jb
criteria or byy substituting j ((a2+b2 =1)) into the
characteristic equation.
Rule 9: The departure angle of the RL from a pole pj (of
multiplicity 1) is given by:
m n
j = 1800 + arg(( p j zi ) arg(( p j pi )
i =1 i =1,i j

The geometric
Th t i form
f off the
th above
b formula
f l isi
j = 1800 + (angle from zi (i=1..m) to pj )
(angle
( l pi (i=1..m,
( ij)) to pj )
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 22
 The root locus of discrete systems Example
Consider a discrete system described by a block diagram:
R(s) Y(s)
+ ZOH G(s)

T = 0.1

5K
G( s) =
s( s + 5)
Sketch the RL of the system when K=0+. Determine
the critical gain Kcr

Solution: The characteristic equation of the system:

1 + G( z) = 0
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 23
 The root locus of discrete systems Example (cont)

1 G(s) 5K
5 K
G ( z ) = (1 z )Z G ( s) =
s s ( s + 5)
1 5K
= (1 z )Z 2
s ( s + 5)
0. 5
1 z[(0.5 1 + e ) z + (1 e 0.5 0.5e 0.5 )]
= K (1 z ) 0 .5

5( z 1) ( z e )
2

0.021z + 0.018
G( z) = K
( z 1)( z 0.607)
0.021z + 0.018
The
Th characteristic
h ti : 1 + K
t i ti equation = 0 (*)
( z 1)( z 0.607)
Poles: p1 = 1 p2 =Z 0.607a = z[(aT 1 + e ) z + (1 e aTe ]
aT aT aT
)
aT
s 2
( s + a ) a ( z 1) 2
( z e )
z
Zeros: 1 = 0 .857
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 24
 The root locus of discrete systems Example (cont)
The asymptotes:
(2l + 1) (2l + 1)
= = =
nm 2 1
OA =
poles zeros [1 + 0.607] ( 0.857)
= OA = 2.464
nm 2 1
The breakaway/break-in points:
( z 1)( z 0.607) z 2 1.607 z + 0.607
((*)) K = =
0.021z + 0.018 0.021z + 0.018
dK 0.021z 2 + 0.036 z 0.042
=
dz (0.021z + 0.018) 2

dK z1 = 2.506
Then =0
dz z2 = 0.792
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 25
 The root locus of discrete systems Example (cont)
The intersection of the root locus
loc s with
ith the unit
nit circle:
(*) ( z 1)( z 0.607) + K (0.021z + 0.018) = 0

z + (0.021K 1.607) z + (0.018 K + 0.607) = 0

2
(**)

Method 1: Apply the extension of Routh Hurwitz criteria:

w +1
Perform the transformation z = , (**) becomes:
w 1
w +1 w +1
2

+ (0.021K 1.607) + (0.018K + 0.607) = 0

w 1 w 1

0.039 Kw2 + (0.786 0.036 K ) w + (3.214 0.003K ) = 0

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 26

 The root locus of discrete systems Example (cont)

According to the corollary of the Hurwitz criterion,

criterion the
stability conditions are:
K > 0 K > 0

0.786 0.036 K > 0 K < 21.83 K cr = 21.83
3.214 0.003K > 0 K < 1071

Substitute Kcr = 21.83 into ((**),

), we have:
z 2 1.1485 z + 1 = 0 z = 0.5742 j 0.8187

Then the intersection of the RL with the unit circle are:

z = 0.5742 j 0.8187

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 27

 The root locus of discrete systems Example (cont)

M th d 2:
Method S b tit t z = a + jb into
2 Substitute i t (**) :
( a + jb)2 + (0.021K 1.607)( a + jb) + (0.018K + 0.607) = 0

a 2 + j 2ab b 2 + (0.021K 1.607)a + j (0.021K 1.607)b +

(0.018K + 0.607) = 0

a 2 b2 + (0.021K 1.607)a + (0.018K + 0.607) = 0

j 2ab + j (0.021K 1.607)b = 0

z 2 + (0.021K 1.607) z + (0.018 K + 0.607) = 0

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 28

 The root locus of discrete systems Example (cont)
Combine with a2 + b2 =1,
=1 we have the set of equations:
a 2 b2 + (0.021K 1.607)a + (0.018K + 0.607) = 0

j 2ab + j (0.021K 1.607)b = 0
a 2 + b 2 = 1

Solve the above set of equation, we obtain 4 intersection:

z =1 when
h K =0
z = 1 when K = 1071
z = 0.5742 j 0.8187 when K = 21.83

K cr = 21.83

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 29

 The root locus of discrete systems Example (cont)
Im z

0 5742+j0 8187
0.5742+j0.8187
+j

22.506
506 0.792
Re z
3 2 1 0.857 0 0.607 +1

j
0.5742j0.8187

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 30

 Frequency response of discrete systems

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 31

 Frequency response of discrete systems
jT
Exact
E act frequency
freq enc response: substitute
s bstit te z = e into the
transfer function G(z) G ( e jT )
10
Ex: Transfer function: G ( z ) =
z ( z 0.6)
10
Frequency response: G ( e ) = jT jT
jT

e ( e 0.6)
Draw the exact Bode diagram of discrete systems:
Difficult
Cannot apply the addition property of the Bode plot

fs
Note: Sampling theorem: f
2 T
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 32
 Frequency response of discrete systems Example
Consider the discrete system:
R(s) Y(s)
+ ZOH G(s)

T = 0.1
12
G( s) =
s( s + 3)
Plot the frequency response of the open-loop system
Solution:The transfer function of the open-loop system:
1 G( s) 0.0544 z + 0.0493
G ( z ) = (1 z )Z G( z) = 2
s z 1.741z + 0.741
jT
jT 0 .0544 e + 0.0493
Frequency response: G ( e ) =
( e jT ) 2 1.741e jT + 0.741
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 33
 Frequency response of discrete systems Example (cont)

Exact Bode diagram (Matlab)

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 34

 The bilinear transformation
1 + Tw / 2 2 z 1
Bilinear transformation: z = w= z + 1
1 Tw / 2 T
j jw

w
1

Z-plane W-plane

Frequency response of discrete system after bilinear

transformation:
G ( z ) z =e j G ( w) w= j
w

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 35

 Relationship between frequency in w-
w-plane and continuous frequency

I the
In th imaginary
i i axis
i off the
th w-plane:
l w = jw
In the unit circle of the z-plane:
jT
2 z 1 2 e 1
2 T
z + 1 jT = jT = j tan
T z =e
T e + 1 T 2
2 z 1
Due to the bilinear transformation: w =
T z + 1
j = j 2 tan T
w
T 2
T T
At low frequency: T / 2 0 tan , then:
2 2
jw j
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 36
 Procedure for drawing approximate Bode diagram

Step 1: Perform the bilinear transformation:

1 + Tw / 2
z=
1 Tw / 2
p 2: Substitute w = jw , then apply
Step pp y the procedure
p for
drawing the Bode diagram presented in chapter 4.

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 37

 Stability analysis of discrete systems using frequency response

When determine
Wh d i the
h gaini crossover frequency,
f phase
h
crossover frequency, remember the relationship:
2 T
jw = j tan
T 2

Gain margin and phase margin are determined in a similar

way as continuous
i systems.

The stability of discrete systems can be analyzed by using

Bode diagrams as continuous systems.

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 38

 Performance of discrete systems

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 41

 Time response of discrete systems
Time response of a discrete system can be calculated by
using one of the two methods below:
Method 1: if the discrete system described by a transfer
function, first we calculate Y(z), and then apply the
inverse z-transform to find y(k).
y( )
Method 2: if the discrete system described by state
equations,
q first we find the solution x(k)( ) to the state
equations, then calculate y(k).

Dominant
D i t poles l off a discrete
di t system
t are the
th poles
l lying
l i
closest to the unit circle.

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 42

 Transient performances
Method 1: Analyzing the transient performance based on
the time response y(k) of discrete systems.
ymax yss
Percentage of overshoot: POT = 100%
yss
ymax and yss are the maximum and steady-state
steady state values of y(k)

Settling time: ts = ksT

where ks satisfying the condition:
. yss
y ( k ) yss , k k s
100

1 yss y ( k ) 1 + yss , k k s
100 100
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 43
 Transient performances
Method 2:2 Analyzing the transient performances based
on the dominant poles.
ln r
=
The dominant poles: z1, 2 = re
* j

(ln r ) 2
+ 2

= 1 (ln r ) 2 + 2
n T

Percentage of overshoot: POT = exp 100%
2
1
3
Settling time: ts = (according to 5% criterion)
n
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 44
 Steady state error

R(s) E(z) Y(s)

+ GC(z) ZOH G(s)
T

H(s)

R ((zz )
E
Error expression:
i E ( z) =
1 + GC ( z )GH ( z )

Steady state error: ess = lim e( k ) = lim(1 z 1 ) E ( z )

k z 1

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 45

 Performances of discrete system Example 1

R( )
R(s) Y( )
Y(s)
+ ZOH G(s)

T = 0.1 10
G(s) =
( s + 2)( s + 3)
1. Calculate the closed-loopp transfer function of the system.
y
2. Calculate the time response of the system to step input.
3. Evaluate the performance of the system: POT, settling
time, steady-state error.

Solution:
G( z)
1. The closed-loopp TF of the system:
y G cl ( z ) =
1 + G( z)
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 46
 Performance of discrete system Example 1 (cont)

G( s)
1 G ( s) =
10
G ( z ) = (1 z )Z ( s + 2)( s + 3)
s
1 10
= (1 z )Z
s( s + 2)( s + 3)
1 z ( Az + B )
= 10(1 z )
( z 1)( z e 20.1 )( z e 30.1 )
1 z ( Az + B)
Z =
0.042 z + 0.036 s s + a s + b z z e aT
z e bT
G( z) = ( )( ) ( 1)( )( )
( z 0.819)( z 0.741
b(1) e aT ) a(1 e bT )
A=
ab(b a)
aeaT (1 e bT ) be bT (1 e aT )
B=
ab(b a)
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 47
 Performance of discrete system Example 1 (cont)
0.042 z + 0.036
G( z) G( z) =
Gcl ( z ) = ( z 0.819)( z 0.741)
1 + G( z)

0.042 z + 0.036
( z 0.819)( z 0.741)
=
0.042 z + 0.036
1+
( z 0.819)( z 0.741)

0.042 z + 0.036
Gcl ( z ) = 2
z 1.518 z + 0.643

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 48

 Performance of discrete system Example 1 (cont)

2. The
Th time
ti response off the
th system
t tot step
t input.
i t
Y ( z ) = Gk ( z ) R( z ) 0.042 z + 0.036
Gk ( z ) =
0.042 z + 0.036 z 2 1.518 z + 0.643
= 2 R( z )
z 1.518 z + 0.643

0.042 z 1 + 0.036 z 2
= 1 2
R( z )
1 1.518 z + 0.643z
(1 1.518 z 1 + 0.643 z 2 )Y ( z ) = (0.042 z 1 + 0.036 z 2 ) R ( z )

y (k ) 1.518 y (k 1) + 0.643 y (k 2) = 0.042r (k 1) + 0.036r (k 2)

y (k ) = 1.518 y (k 1) 0.643 y (k 2) + 0.042r (k 1) + 0.036r (k 2)

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 49

 Performance of discrete system Example 1 (cont)

U it step
Unit t r ( k ) = 1,
t input:
i kk 0
Initial condition: y ( 1) = y ( 2) = 0
Substitute the initial condition to the recursive equation
off y(k),
(k) we have:
h
y ( k ) = {0; 0.0420; 0.1418; 0.2662; 0.3909; 0.5003;...
0.5860; 0.6459; 0.6817;0.6975; 0.6985; 0.6898;...
0.6760; 0.6606; 0.6461; 0.6341; 0.6251; 0.6191;...}

c(k ) = 1.518c(k 1) 0.643c(k 2) + 0.042r (k 1) + 0.036r (k 2)

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 50

 Performance of discrete system Example 1 (cont)
Step Response
0.7

0.6

0.5

0.4
Amplitude

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5 3
Time (sec)

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 51

 Performance of discrete system Example 1 (cont)
3 Transient performances:
3. 0.042 z + 0.036
Gk ( z ) = 2
The steady state response: z 1.518 z + 0.643
li (1 z 1 )Y ( z )
yss = lim 1
R( z ) =
z 1
1 z 1
= lim(1 z 1 )Gk ( z ) R( z )
z 1
0.042 z + 0.036 1
1
= lim(1 z ) 2 1
z 1 z 1.518 z + 0.643 1 z
y ss = 0.624

The maximum value: ymax = 0.6985

Percentage of overshoot:
ymax yss 0.6985 0.624
POT = 100% = 100% = 11.94%
yss 0.624
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 52
 Performance of discrete system Example 1 (cont)
Settling time (5% criterion): cxl = 0.624
First, we need to find ks satisfying: = 5% = 0.05
(1 ) yss y (k ) (1 + ) yss , k ks
0.593 y (k ) 0.655, k ks
From the time response calculated before ks = 14
ts = ksT = 14 0.1
ts = 1.4 sec

kt) t= {error:
St d c(state
Steady 0; 0.0420; 0.1418; 0.2662; 0.3909; 0.5003;...
Since the system0.5860is;unity
0.6459negative
; 0.6817;feedback,
0.6975; 0.6985we ;have:
0.6898;...
ess = rss 0.6760
yss =; 10.
6606 ; 0.6461;
0.624 0.6341e; 0=.6251
0. ; 0.6191;...}
376
ss

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 53

 Performance of discrete system Example 1 (cont)
Note: It is possible to calculate POT and ts based on the dominant poles
The poles of the closed-loop system are the roots of the equation:
z 2 1.518 z + 0.643 = 0
z1*, 2 = 0.7590 j 0.2587 = 0.80190.3285
ln r ln 0.8019
= = = 0.5579
(ln r ) +
2 2
(ln 0.8019) + 0.3285
2 2

1 1
n = (ln r ) + =
2 2
(ln 0.8019) 2 + 0.3285 2 = 0.3958
T 0 .1


POT = exp
.100% = exp 0.5579 3.14 .100% = 12.11%
2
1 2
1 0. 5579
3 3
t q = = = 1.36 sec
n 0.5579 0.3958
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 54
 Performance of discrete system Example 2

r(t) e(t) e(kT) eR(t) y(t)

+ ZOH G (s )

T

2( s + 5)
with T = 0.1 G( s) =
( s + 2)( s + 3)
1. Formulate
1 F l t the
th state
t t equations
ti d
describing
ibi the
th system
t
2. Calculate the response of the system to unit step input
(assuming the initial conditions are zeros) using the state
equation formulated above.
3 Calculate POT,
3. POT settling time,
time steady state error

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 55

 Performance of discrete system Example 2 (cont)

Solution:
S l ti
1. Formulate the state equation:
Y ( s) 2( s + 5) 2 s + 10
G( s) = = = 2
E R ( s ) ( s + 2)( s + 3) s + 5s + 6
The state equation of the continuous plant:
x&1 (t ) 0 1 x1 (t ) 0
x& (t ) = 6 5 x (t ) + 1 eR (t )
2 1424 3 2 {

A B
x1 (t )
y (t ) = [10 2]
123 x2 (t )
C
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 56
 Performance of discrete system Example 2 (cont)
The transient matrix:
1 1
1 0 0 1 s 1
( s ) = (sI A) = s =
-1

0 1 6 5 6 s + 5
s+5 1
1 s + 5 1 ( s + 2)( s + 3) ( s + 2)( s + 3)
= =
s ( s + 5) 6 6 s 6 s
( s + 2)( s + 3) ( s + 2)( s + 3)
1 3 2 1 1 1
L L
1 s + 2 s + 3 s + 2 s + 3
(t ) = L [( s)] ==
L 1 6 + 6 L 1 2 + 3
s + 2 s + 3 s + 2 s + 3
(3e 2t 2e 3t ) (e 2t e 3t )
(t ) = 2t 3t 2t 3t
(6e + 6e ) (2e + 3e )
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 57
 Performance of discrete system Example 2 (cont)
The state equation of the x[(k + 1)T ] = Ad x(kT ) + Bd eR (kT )

discrete open-loop system: y (kT ) = Cd x(kT )

(3e 2T 2e 3T ) (e 2T e 3T ) 0.9746 0.0779

Ad = (T ) = 2T 3T 2T 3T
= 0.4675 0.5850
( 6 e + 6 e ) ( 2 e + 3e ) T = 0. 1

T
T 2 3
(3e 2e ) (e 2 e 3 ) 0
Bd = ( ) Bd = 2 3 2 3
d
0 0(6e + 6e ) (2e + 3e ) 1
2 3 0.1
T (e 2 e 3 ) ( e +
e
0.0042
= d = ) =
2 3 2 3
0 ( 2e + 3e )
(e e )
2 3 0 .0779
0

Cd = C = [10 2]

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 58

 Performance of discrete system Example 2 (cont)
The state equation of the discrete closed-loop system:
x[(k + 1)T ] = [ Ad Bd Cd ]x(kT ) + Bd r (kT )

y (kT ) = Cd x(kT )
with
0.9746 0.0779 0.0042 0.9326 0.0695
[Ad Bd Cd ] = [10 2] =
0.4675 0.5850 0.0779 1.2465 0.4292

x1 (k + 1) 0.9326 0.0695 x1 (k ) 0.0042

x (k + 1) = 1.2465 0.4292 x (k ) + 0.0779 r (kT )
2 2

x1 (k )
y (k ) = [10 2].
x
2 ( k )

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 59

 Performance of discrete system Example 2 (cont)
2 Time response of the system:
2.
From the closed-loop state equations, we have:
x1 (k + 1) = 0.9326 x1 (k ) + 0.0695 x2 (k ) + 0.0042r ( k )

x2 (k + 1) = 1.2465 x1 (k ) + 0.4292 x2 (k ) + 0.0779r (t )
With initial condition x1(1)=x2(1)=0, unit step input, we can
calculate the solution to the state equation:
x1 (k ) = 10 3 {0; 4.2; 13.5; 24.2; 34.2; 42.6; 49.1; 54.0; 57.4; 59.7;...
61.2; 62.0; 62.5; 62.7; 62.8; 62.8; 62.7; 62.7; 62.6; 62.6 ...}
x2 (k ) = 10 3 {0; 77.9; 106.1; 106.6; 93.5; 75.4; 57.2; 41.2; 28.3; 18.5; ...
11 4; 6.5;
11.4; 6 5; 3.4;
3 4; 1.4; 0 3; -0.3;
1 4; 0.3; 0 3; -0.5;
0 5; -0.5;
0 5; -0.5; 0 4 ...}
0 5; -0.4
The closed-loop system response: y (k ) = 10 x1 (k ) + 2 x2 (k )
y (k ) = {0; 0.198;; 0.348;; 0.455;; 0.529;; 0.577;; 0.606;; 0.622;; 0.631;; 0.634;...
;
0.635; 0.634; 0.632; 0.630; 0.629; 0.627; 0.627; 0.626; 0.625; 0.625...}
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 60
 Performance of discrete system Example 2 (cont)
Step Response
espo se
0.7

0.6

0.5

0.4
Amplitude

0.3

0.2

0.1

0
0 0.5 1 1.5 2 2.5
Time (sec)

16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 61

 Performance of discrete system Example 2 (cont)
3 Performances of the system:
3.
Percentage of overshoot:

ymax = 0.635
63 ymax yss
POT = 100% = 1.6%
yss
yss = 0.625
The settling time:
(1 0.05) y y ( k ) (1 + 0.05) y , k ks
According to the response of the system:
0.594 y (k ) 0.656, k 6
ks = 6 ts = ksT = 0.6 sec
c(k ) = {0; 0.198; 0.348; 0.455; 0.529; 0.577; 0.606; 0.622; 0.631; 0.634;...
S d state error: ess = rss yss = 1 0.625 = 0.375
Steady
0.635; 0.634; 0.632; 0.630; 0.629; 0.627; 0.627; 0.626; 0.625; 0.625...}
16 November 2011 H. T. Hong - www4.hcmut.edu.vn/~hthoang/ 62