Digital Control Systems

(EP-5511)

Lecture-4
Z-plane Analysis of DTCS

Belaynesh Belachew (capt.)

Defence Engineering College

1
 Obtaining z-Transform By The Convolution
Integral Method
Consider the impulse sampler output
 
x (t )   x(t ) (t  kT )  x(t )  (t  kT )
k 0 k 0

L  (t  kT )  ekTs
Since  

=X (s)  L[ x (t )]  L= x(t ) (t  kT ) 
* *

 k 0 
The Laplace transform of the product of two Laplace-
transformable functions f(t) and g(t) can be given by
 1 c  j
= L[ f (t ) g (t )]   f (t ) g (t )e dt 
2 j 
 st
F ( p)G(s  p)dp
0 c  j
2
 Continued
Let us substitute x(t) and for f(t) and g(t),
respectively. Then
 

=X ( s)  L= x(t )  (t  kT ) 
*

 k 0  Convolution
1 c  j 1 Integral

2 j c  j 
X ( p)
1 e T ( s  p )
dp

Since
  1
= L= (t  kT )   Ts
 k 0  1  e
where the integration is along the line from c - j∞ to c +
j∞ and this line is parallel to the imaginary axis in
the p plane and separates the poles.
3
 Continued
An integral can be evaluated in terms of residues by
forming a closed contour consisting of the line from
c - j∞ to c + j∞ and a semicircle of infinite radius in
the left or right half-plane, provided that the integral
along the added semicircle is a constant (zero or a
nonzero constant).
1 c  j 1
2 j 
=X ( s) 
*
X ( p) T ( s  p )
dp
c  j 1 e
1 X ( p) 1 X ( p)
 o dp   dp
2 j 1  e T ( s  p )
2 j  1  e  T ( s  p )

where Γ is a semicircle of infinite radius in the left or

right half p plane
4
 Continued
There are two ways to evaluate this integral: one using
an infinite semicircle in the left-half plane and the
other an infinite semicircle in the right-half plane.
Evaluation of the convolution integral in the left-half
plane: * 1 X ( p) 1 X ( p)
=X (s)  o dp   dp
2 j 1  e T ( s  p )
2 j 1  e
 T ( s  p )

, Nothing that degree of P(s) in s is greater

than q(s), the integral along ΓL vanishes. Hence
1 X ( p)
=X ( s) 
*
o dp
2 j 1  e T ( s  p )

 X ( p) 
   residue of T ( s  p )
at pole of X ( p ) 
 1  e 5
 Continued
Closed contour in the left-half of the p plane.
By substituting z for
e-Ts and changing the
complex variable
notation from p to s.
 X ( p) z 
 residue of
=X ( z )   z  eTp 
 
 at pole of X ( p) 

 X ( s) z 
=X ( z )   residue of at pole of X ( s ) 
 z  e Ts

6
 Continued
Assume that X(s) has poles s1, s2, …, sm. If a pole at s=sj
is a simple pole, the corresponding residue Kj is
 X ( s) z 
K j  lim  s  s j Ts 
s s j
 z  e 
If a pole at s=si is a multiple pole of order ni, the
1 d ni 1  ni X ( s) z 
Ki  lim ni 1  s  si 
(ni  1)! ssi ds  z  eTs 
Example 1: Given

obtain X(z) by use of the convolution integral in the left

half-plane.
7
 Continued
Solution: Two poles at s=0 and one at s=-1
 X ( s) z 
=X ( z )   residue of at pole of X ( s ) 
 1  eTs

1 d  2 1 z   1 z 
 lim  s 2   lim  s  1 2
(2  1)! s0 ds  s ( s  1) z  e  s1 
Ts
s (s  1) z  eTs 

 z  z  eTs  ( s  1)(T )eTs  1 z

 lim 
s 0 ( s  1)2 ( z  eTs )2 (1)2 z  eT
z  z 1 T  z z 2 (T  1  eT )  z (1  eT  TeT )
  
( z  1)2  z  eT 
T
( z  1) 2
z e
(T  1  eT ) z 1  (1  eT  TeT ) z 2

(1  z 1 )2 1  eT z 1  8
 Continued
Evaluation of the convolution integral in the right-
half plane: The closed contour consisting of the line
from c - j∞ to c + j∞ and ΓR.
The closed contour does not enclose any poles of X(p).
1 X ( p) 1 X ( p)
=X (s)  o dp  
*
dp
2 j 1  e T ( s  p )
2 j R 1  e T ( s  p )

There are two cases to evaluate

Case 1: X(s) has p(s) two or more degrees higher in s
than q(s). For this case, since X(s) possesses at least
two or more poles than zeros.

1 X ( p)
=  dp  0
2 j R 1  e T ( s  p )
9
 Continued
Closed contour in the right-half of the p plane.

Then
1 
=X ( s)   X (s  js k )
*

T k 

Thus

1 
=X ( z )   X ( s  js k )
T k  s  (1/T)lnz

10
 Continued
Case 2: X(s) has p(s) one degree higher in s than q(s).
For this case, and the integral
along ΓR is not zero.
1 X ( p) 1
=  dp   x (0 
)
2 j R 1  e
 T ( s  p )
2
Therefore:

1 1
=X ( s)   X ( s  js k )  x(0 )
*

T k  2
Example 2: show that X*(s) is periodic with period 2π/ωs
Referring to equation

1 1
=X ( s)   X ( s  js h)  x(0 )
*

T h 2
11
 Continued
Hence,

1 1
=X (s  js k )   X (s  js k  js h)  x(0 )
*

T h 2
Let k+h=m, then
1  1
=X ( s  js k )   X (s  js m)  x(0 )  =X * (s)
*

T m 2
Therefore,
 X * (s  js k ), k  0,1, 2,...
X * (s) =
Thus, X*(s) is periodic, with period 2π/ωs. This means
that, if a function X(s) has a pole at s=s1 in the s
plane, then X*(s) has poles at s=s1 ± jωsk
(k=0,1,2,...).

12
 Continued
Obtaining z Transforms of functions involving the
term (1-e-Ts)/s: Suppose
1  eTs
X ( s)  G( s)
s
Ts G ( s )
X ( s)  1  e  1  eTsG1 ( s)
s
Consider the function (product of two LT functions)
x1 (t )   g0 (t  ) g1 ( )d
t
Ts
x1 ( s)  e G1 ( s) 0

Where g0(t)= £-1[e-Ts] =δ(t-T) and g1(t)= £-1[G1(s)]

Thus,
x1 (t )    (t  T  ) g1 ( )d  g1 (t  T )
t

0
13
 Continued
Ȥ[g1(t)]= G1(z) and Ȥ[x1(t)]= Ȥ[g1(t-T)]= Ȥ-1G1(z)
X ( z )  [G1 ( s)  eTs G1 ( s)]
 [ g1 (t )]  [ x1 (t )]
 G1 ( z )  z 1G1 ( z )
 (1  z 1 )G1 ( z )
Or
 G( s) 
X ( z )  [ X ( s)]  1  z   
1

 s 
Example 3: Obtain the z transform of
1  eTs 1
X ( s) 
s s 1
14
 Continued
Solution:
1  eTs 1 
X ( z)    
 s s  1
1
 1 
 (1  z )   
 s  s  1 
1 1 1 
 (1  z )    
 s s  1 
1  1 1 
 (1  z )  1
 T 1 
 1  z 1  e z 
(1  eT ) z 1

1  eT z 1 15
 Pulse Transfer Function
The transfer function for the continuous-time system
relates the Laplace transform of the continuous-time
output to that of the continuous-time input, while the
pulse transfer function relates the z transform of the
output at the sampling instants to that of the sampled
input.
Convolution Summation:


  y(t )  Y ( z )   y(kT ) z  k
k 0
16
 Continued
 For the continuous time-system
y(t )   g (t  ) x( )d   x(t  ) g ( )d
t t

0 0

 For the discrete-time system

 
x (t )   x(t ) (t  kT )   x(kT ) (t  kT )


k 0 k 0
• For a physical

system a response cannot precede the input
y(t )   g (t  hT ) x(ht ), 0  t  kT
h 0
The value of the output y(t) at the sampling instants
t=kT are given by Convolution summation
 
y(kT )   g (kT  hT ) x(hT )   x(kT  hT ) g (hT )
h 0 h 0
y(kT )  x(kT )  g (kT ) 17
 Continued
 It is noted that if G(s) is a ratio of polynomials in s and if
the degree of the denominator polynomial exceeds that
of the numerator polynomial only by 1 the output y(t) is
discontinuous.
 In analyzing discrete-time control
systems it is important to
remember that the system
response to the impulse-sampled
signal may not portray the correct
time-response behavior of the
actual system unless the transfer
function G(s) of the continuous-
time part of the system has at least
two more poles than zeros, so that 18
 Continued
Pulse Transfer Function

y(kT )   g (kT  hT ) x(hT ), k  1, 2,...
h 0

 The z transform of y(kT)

  
Y ( z )   y(kT ) z  k   g (kT  hT ) x(hT ) z  k
k 0 k 0 h 0
   
  g (mT ) x(hT ) z ( m h )  g (mT ) z  m  x(hT ) z  h
m 0 h 0 m 0 h 0

 G( z ) X ( z )
 Pulse transfer function
Y ( z)
G( z ) 
X ( z) 19
 Continued
General procedures for obtaining pulse transfer
functions:

Y(s)=G(s)X*(s) and Y*(s)=G*(s)X*(s)

Y*(s)=[G(s)X(s)]*=[GX(s)]*
Y(z)=Z[Y(s)]=Z[G(s)X(s)]=Z[GX(s)]=GX(z)≠G(z)X(z)
20
 Continued
Pulse Transfer Function of Cascaded Elements

U(s)=G(s)X*(s), Y(s)=H(s)U*(s),
U*(s)=G*(s)X*(s), Y*(s)=H*(s)U*(s)
Y*(s)=H*(s)G*(s)X*(s) Y (z)=H(z)G(z)X(z)

Y(s)=G(s)H(s)X*(s)=GH(s)X*(s)
Y*(s)=[GH(s)]*X*(s) Y(z)=GH(z)X(z)
Y ( z)
 GH ( z )  Z[GH (s)]
X ( z)
Note that G(z)X(z)≠GX(z)=Z[GX(s)] 21
 Continued
Pulse Transfer Function of Closed-Loop Systems

E (s)  R(s)  H (s)C (s)

C (s)  G(s) E  (s)
E ( s)  R( s)  H ( s)G ( s ) E  (s )
E  ( s)  R ( s)  GH  ( s) E  ( s)

R ( s)
E  ( s) 
1  GH  ( s)
 
G ( s ) R (s) G ( z ) R( z ) C( z) G( z )
C  (s)   C ( z )   
1  GH  ( s) 1  GH ( z ) R( z ) 1  GH ( z ) 22
 Continued
Five typical configurations for closed-loop discrete-time control
systems

23
 Continued
Pulse Transfer Function of a Digital Controller:
 The input to the digital controller is e(k) and the
output is m(k)
m(k )  a1m(k  1)  a2m(k  2)   amm(k  m)
 b0e(k )  b1e(k  1)   bne(k  n)
 The z transform of the equation
M ( z )  a1 z 1M ( z )  a2 z 2 M ( z )   am z  m M ( z )
 b0 E ( z )  b1 z 1E ( z )   bn z  n E ( z )
 1
1  a z 1
 a2 z 2
  am z  m  M ( z )
  b0  b1 z 1   bn z  n  E ( z )
M ( z) b0  b1 z 1   bn z  n
GD ( z )  
E ( z ) 1  a1 z 1  a2 z 2   am z  m 24
 Continued
Closed-loop Pulse Transfer Function of a Digital
Control System:

1  eTs
G p ( s)  G( s), C ( s)  G( s)GD ( s) E  (s), C  ( s)  G  ( s)GD ( s) E  (s)
s
C ( z )  G( z )GD ( z ) E ( z ), E ( z )  R( z )  C ( z )
C( z) G( z )GD ( z )
C ( z )  G( z )GD ( z )  R( z )  C ( z )   
R( z ) 1  G( z )GD ( z ) 25
 Continued
Pulse Transfer Function of a Digital PID Controller:
 The PID control action in analog controllers

 Discretization of the equation to obtain the pulse

transfer function
 T  e(0)  e(T ) e(T )  e(2T ) e(k  1)  e(kT )  
 e ( kT )      
Ti  2 2 2
m(kT )  K  
  T e(kT )  e((k  1)T ) 
 d
T 
 T k e((h  1)T )  e(hT ) Td 
m(kT )  K e(kT )    e(kT )  e((k  1)T ) 
 Ti h 1 2 T 
26
 Continued
Define, e((h  1)T )  e(hT )
 f (hT ), f (0)  0
2
k
e((h  1)T )  e(hT ) k

h 1 2
  f (hT )
h 1

 k e((h  1)T )  e(hT )  k  1 1

Z    Z   f ( hT )   1  F ( z )  f (0)   1
F ( z)
 h1 2   h1  1 z 1 z
1  z 1 k
e((h  1)T )  e(hT ) 1  z 1
F ( z )  Z  f (hT )   E( z)    1
E( z)
2 h 1 2 2(1  z )
 T k e((h  1)T )  e(hT ) Td 
m(kT )  K e(kT )    e(kT )  e((k  1)T ) 
 Ti h 1 2 T 
 T 1  z 1 Td 1   1 
M ( z )  K 1  1
 1  z   E ( z )  K 1 
T T 1
 1

Td
1  z   E( z)
 2Ti 1  z T   2Ti Ti 1  z T 
 1 
D 
Ki
 K p  1
 K 1  z  E( z)
 1 z 
27
GD ( z )
 Continued
Obtaining response between consecutive sampling
instants
 Laplace transform method
 Modified z transform method
 State-space method
Laplace Transform Method

R( s )
C ( s )  G ( s ) E  ( s)  G( s)
1  GH  (s)
 R( s ) 
c(t )  L1 C (s)  L1 G(s)
 1  GH  (s) 

28
29