Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

MODERN CONTROL

SYSTEMS ENGINEERING
COURSE #: CS421
INSTRUCTOR:
DR. RICHARD H. MGAYA

Date: October 25th, 2013

Analysis of Steady-State Error


Final Value Theorem
Examination of the asymptotic behavior of a discrete system
without the need for the evaluation of the z-transforms
associated with transfer functions
Recap: Sequences that have a limiting value for large k must be
stable functions
All poles must lie inside the unit circle

Function should be expandable into the form


n
Bi z
Az
F ( z)

z 1 z 1 z pi

pi 1

... i

Inverse transformation
n

f (k ) A Bi pik
i 1

Dr. Richard H. Mgaya

Analysis of Steady-State Error


Final Value Theorem Cont
As k gets large the terms contributed by the summation sign
will approach zero.
Evaluation of A by partial fraction expansion:
From eqn. i
z 1
Multiply by
and let z approaches unity
z

z 1
A lim z 1
F ( z)
z

Formal mathematical theorem:


z 1
limk f (k ) lim z 1
F ( z)
z
Dr. Richard H. Mgaya

Analysis of Steady-State Error


Effect of Sampling on the Steady-State Error
No general conclusion about the steady-state error
Error depend on the placement of the sampler

Continuous Systems:
Steady-state error is based on the open-loop transfer function
Static error constants
estep ()

1
1 lims 0 G ( s )

Kp

, eramp ()

1
lims 0 sG( s )

Kv

, and e parabola ()

1
lims 0 s 2G ( s)

Ka

Discrete Systems:
Placement of the sampler changes the open-loop transfer
function
Dr. Richard H. Mgaya

Analysis of Steady-State Error


Effect of Sampling on the Steady-State Error Cont
Consider the following digital system:
Digital computer is represented as a sampler and a zero-order hold

Block reduction techniques


Dr. Richard H. Mgaya

Analysis of Steady-State Error


Effect of Sampling on the Steady-State Error Cont
From the figure: E( z) R( z) E( z)G( z)
E( z)

R( z )
1 G( z)

Final value for discrete signal


z 1
e () lim z 1
E( z)
z
*

e* () limz 1 1 z 1 E ( z )

e*() is the final sampled value of e(t) or e(kT)


e*() for unity feedback system

e* () limz 1 1 z 1

1 RG( z()z)

... ii
Dr. Richard H. Mgaya

Analysis of Steady-State Error


Unit Step
Unit step input:R(s) 1s
R( z )

z
z 1

Substitution to eqn. ii
e* ( )

1
1 lim z1 G ( z )

K p limz 1 G( z )

Thus,
e* ( )

1
1 K p
Dr. Richard H. Mgaya

Analysis of Steady-State Error


Unit Ramp
Unit ramp input: R(s) s1

R( z )

Tz
z 12

From unit step procedure eqn. ii


1

e* ( )

1
limz 1 ( z 1)G( z )
T
1
K v limz 1 ( z 1)G( z )
T

Thus,
e* ( )

1
Kv
Dr. Richard H. Mgaya

Analysis of Steady-State Error


Unit Parabolic
Unit parabolic input: R(s) s1

T 2 z ( z 1)
R( z )
3
2z 1

Similarly
e ( )

Ka

1
2
lim
(
z

1
)
G( z )
z 1
2
T

1
2
lim
(
z

1
)
G( z)
z 1
2
T

Thus,
e* ( )

1
Ka
Dr. Richard H. Mgaya

Analysis of Steady-State Error


Conclusion:
The error constants for digital systems are similar to those of
analog systems
Multiple pole placement at the origin of the s-plane reduces the error to
zero in case of analog systems
Multiple placement of pole at z = 1 of the z plane reduces the steadystate error to zero for the case of discrete system of the type discussed

Thus,
s 0 maps to z 1 under z eTs
Stability
Continuous systems stability can be determined by the Routh
Hurwitz criterion
Number of unstable poles
The number of sign changes indicates the number of unstable poles and
not their location
Dr. Richard H. Mgaya

Stability of Digital Control System


Jury test is employed to asses the stability of discrete systems
Consider the characteristic equation s a sampled-data system:
Q( z ) an z n an 1 z n 1 a1 z a0 0
z0
a0
an
b0
bn 1

l0
l3
m0

z1
a1
an 1
b1
bn 2

l1
l2
m1

z2
a2
an 2
b2
bn 3

l2
l1
m2

m2

m1

m0

z n 1
an 1
a1
bn 1
b0

l3
l0

zn
an
a0

a0
bk
an

an k
ak

b0
ck
bn 1

bn 1 k
bk

c0
dk
cn 2

cn 2 k
ck

Dr. Richard H. Mgaya

Stability of Digital Control System


Necessary and sufficient conditions for stability
No roots outside or on the unit circle
Condition 1: Q(1) > 0
Condition 2: (-1)nQ(-1) > 0
Condition 3: a0 an
b0 bn 1
c0 cn 2

Condition n: m0 m2
Dr. Richard H. Mgaya

Stability of Digital Control System


Example
Given the characteristic equation below find the value of K to
make the system just unstable
1

K (0.092 z 0.066)
0
2
z 1.368 z 0.368

Solution:
z 2 (0.092 K 1.368) z (0.368 0.066 K ) 0

Condition 1: Q(1) > 0


Q(1) 1 (0.092 K 1.368) (0.368 0.066 K ) 0
Q(1) 0 if K 0

Condition 2: (-1)nQ(-1) > 0


(1) 2 Q(1) {1 (0.092 K 1.368) (0.368 0.066 K ) 0
2.736 0.026 K 0
or
K 105.23
Dr. Richard H. Mgaya

Stability of Digital Control System


Example Cont
Condition 3: a0 an
0.368 0.066 K 1
0.368 0.066 K 1

1 0.368
K
9.58
0.066

The system is marginal stable when K = 9.58 and K = 105.23

Dr. Richard H. Mgaya

Stability of Digital Control System


Digital System Stability via s-Plane
Mapping of j axis points on the s-plane to unit circle points
on the z-plane, i.e., bilinear transformation
Transformation maps right-half plane points on the s-plane to point
outside the unit circle on the z-plane. It also maps the left-half points on
the s-plane to the point inside the unit circle on the z-plane.

Transformation of the denominator of the pulse transfer


function D(z) to the denominator of a continuous transfer
function G(s)
Routh-Hurwitz stability criterion can then be applied to the
transformed system
Dr. Richard H. Mgaya

Stability of Digital Control System


Digital System Stability via s-Plane
Bilinear transformation:
Provides tools for applying s-plane design and analysis to
digital system
If
z eTs
and the inverse

1
ln z
T

Expansion of the logarithmic function


z 1 1 z 1 3

ln z 2


z 1 3 z 1

Bilinear transforms

2 z 1
s
T z 1

Dr. Richard H. Mgaya

Stability of Digital Control System


Digital System Stability via s-Plane
Inverse
z

Substitute s = +j
z
z

s 1
s 1

( 1) j
( 1) j
( 1) 2 2
( 1) 2 2

Thus,
z 1 when 0
z 1 when 0
z 1 when 0
Dr. Richard H. Mgaya

Stability of Digital Control System


Digital System Stability via s-Plane
Example: Given T(z) = N(z)/D(z), where
D( z ) z 3 z 2 0.2 z 0.1

Use Routh-Hurwitz criterion to find the number of z-pane poles


of T(z) inside, outside, on the unit circle. Is the system stable?
Substitute: Inverse of bilinear transform into D(z) = 0
s 3 19s 2 45s 17 0

Routh-Hurwitz table:
s3

- 45

s2

- 19

- 17

s1

- 45.89

s0

- 17

0
Dr. Richard H. Mgaya

Stability of Digital Control System


Digital System Stability via s-Plane
Example: Cont
Routh-Hurwitz table shows one root on the right half-plane
and two roots on the left half-plane.
Thus, T(z) has one pole outside the unit circle, two poles in the
unit circle and no pole on the unit circle
Conclusion: System is unstable poles outside the unit circle

Dr. Richard H. Mgaya

Root locus Analysis in z-plane


Root Locus Construction Rules
1.

Starting points (K = 0): The root loci starts at open-loop poles

2.

Terminating points (K = ): The root loci terminate at open-loop zeros


when they exists, otherwise at infinity.

3.

Number of distinct root loci: This is equal to the order of the characteristic
polynomial

4.

Symmetry of root loci: The root loci are symmetrical about the real axis

5.

Root locus location on real axis: A point on the real axis is part of the loci
if the sum of the open-loop poles and zeros to the right of the point is zero

6.

Break away points: The points where a locus breaks away from real axis
are the root of the equation

d
GH ( z ) 0
dz

Dr. Richard H. Mgaya

Root locus Analysis in z-plane


Root Locus Construction Rules
7. Unit circle crossover: Can be obtained by determining the
value of K for marginal stability using the Jury test
Example: For a given transfer function G(s), determine the
breakaway point, the value of K for marginal stability and the unit
circle cross-over for T = 0.5 seconds
e Ts
G ( s ) K 1
s

s ( s 2)

z-Transform:
0.092 z 0.066
G( z) K 2

1
.
368
z

0
.
368

Dr. Richard H. Mgaya

Root locus Analysis in z-plane


Root Locus Construction Rules
Characteristic equation:
1

K (0.092z 0.066)
0
2
z 1.368z 0.368

z 2 (0.092 K 1.368) z (0.368 0.066 K ) 0

Breakaway points:
d
{GH ( z )} 0
dz
( z 2 1.368 z 0.368) K (0.092 z 0.066)(2 z 1.368) 0

0.092 z 2 0.132 z 0.1239 0


z1 0.647
z 2 2.084
Dr. Richard H. Mgaya

Root locus Analysis in z-plane


Root Locus Construction Rules
K for marginal stability: Jury test
K1 9.58
K 2 105.23

Unit circle crossover: K into the characteristic equation


z 2 0.487 z 1 0

Roots
z1 0.244 j 0.97

Dr. Richard H. Mgaya

Root locus Analysis in z-plane


Root Locus Construction Rules
Root locus analysis of a discrete system is the plot of the roots
of the characteristic equation 1+G(z)=0 in z-plane as a function
of K.

Dr. Richard H. Mgaya

You might also like