EE-379

LINEAR CONTROL SYSTEMS

Lecture No 14, 15
“CONTROL SYSTEMS ENGINEERING”
Text Book: Chapter 6 (nn 6th Ed)
6.1 Introduction Page 302
6.2 Routh-Hurwitz Criterion Page 305
6.3 Routh-Hurwitz Criterion:
(Special Cases) Page 308
Instructor: Dr. Farid Gul
Class: BEE 4 A/B
Electrical Engineering Department

1
 Introduction

Control systems analysis and design

focus on three specifications:
(1) transient response
(2) stability
(3) steady-state errors

2
We can control the output of a system
if
the steady-state response consists of The system is stable
only the forced response

The total response of a system is sum

of the forced and natural responses

c(t) = cforced(t) + cnatural(t)

Transient and steady state responses of a system

can be controlled only if the system is stable
3
 Control Systems (Example) From Lecture 01

Consider an elevator. When the fourth-floor button is pressed,

the elevator rises to the fourth floor with a speed and floor-
leveling accuracy designed for passenger comfort.

Two measures of
performance are :
(1) the transient
response
(2) the steady-state
error.

Passenger comfort and passenger patience are dependent

upon the transient response

4
 Definitions
(from natural response perspective )

Stability : A linear, time-invariant system is stable if the

natural response approaches zero as time approaches
infinity.

Instability : A linear, time-invariant system is unstable if the

natural response grows without bound as time approaches
infinity.

Marginal Stability : A linear, time-invariant system is

marginally stable if the natural response neither decays nor
grows but remains constant or oscillates as time approaches
infinity.

5
 Definitions
(from total response perspective)

1. A system is stable if every bounded input yields a bounded

output.
(bounded-input, bounded-output (BIBO) definition of stability)

2. A system is unstable if any bounded input yields an

unbounded output.

6
7
Cause of problems in finding
closed-loop poles:

Poles for the forward TF

are known

Closed-loop system

We need to factorize the

characteristic polynomial
to find roots of the
closed-loop TF
equivalent system

8
Location of the poles of the closed-loop system are not known unless the
characteristic polynomial is factored or otherwise solved for the roots.

Some conclusions can be drawn about the stability of the system.

If the transfer function has poles only in the left-half-plane, then the factors of the
denominator consist of products of terms:
s + ai
ai is real and positive, or complex with a positive real part

a. The product of terms (s + ai ) is a polynomial with all positive coefficients

b. No term of the polynomial can be missing
(no cancellation between positive and negative coefficients in the factors)

9
• A sufficient condition for a system to be unstable is that all
signs of the coefficients of the denominator of the transfer
function are not the same.
• If powers of s are missing, the system is either unstable or
marginally stable
• If all coefficients of the denominator are positive and not
missing, we are uncertain about the system's pole locations

How can we find about stability of the system if all

coefficient are greater than zero ????

10
 Routh-Hurwitz Criterion
The Routh-Hurwitz criterion for stability is a method that
yields stability information without the need to solve for the
closed-loop system poles.

The method requires two steps:

(1) Generate a data table called a Routh table
(2) Interpret the Routh table to tell how many system
poles are in the left half-plane, in the right half-plane,
and on the jω-axis

11
Generating Routh Table:

Q(s)  ansn  an1sn1  an2sn2  an3sn3  an4sn4   a1s  a0

sn an an-2 an-4 an-6 …… 1 an an2

b1  
sn-1 an-1 an-3 an-5 an-7 …… an1 an1 a n 3

sn-2 b1 b2 b3 b4 …… 1 an an  4
b2  
sn-3 c1 c2 c3 c4 ……
an 1 an 1 an 5
. .
1 an 1 an 3
. . c1  
. . b1 b1 b2
s2 k1 k2
1 an1 an5
s1 l1 c2  
b1 b1 b3
s0 m1
 Routh-Hurwitz Criterion

Closed-loop transfer
function

Initial layout for

Routh table

14
Completed Routh table

15
Example 6.1 Make the Routh table for the system shown in the Figure.

Feedback system

SOLUTION Find the equivalent closed-loop system because we want to

test the denominator of the closed-loop transfer function.

How the transfer function is found ???

Equivalent closed-loop system

16
Make a Routh Table and label the rows with powers of s from
s3 down to s0 in a vertical column

Form the first and

second row using the
coefficients of the C P

Elements of subsequent
rows are computed
from the entries of the
preceding rows
Completed Routh table
17
s3 1 31 0

s2 1 103 0

s1 -72 0 0

s0 103 0 0
18
 Case 1
(No Zero element in first column)

19
Example 1: (Case: 1) (No zero element in first column)

–Number of sign changes in the first column = number of

unstable poles

20
Example 2: (Case: 1) (No zero element in first column)
P(s) = 2s4 + 3s3 + 5s2 + 2s + 6

s4 2 5 6 Two sign
changes
s3 3 2 +

s2 +11/3 6 RHP=2 -
LHP=2
IA = 0
+
s1 -32/11
Two
0
unstable
s +6 poles
–Number of sign changes in the first column = number of
unstable poles 21
 Interpreting the Basic Routh Table
The basic Routh table applies to systems with poles in
the left and right half-planes

The Routh-Hurwitz criterion declares that the

number of roots of the polynomial that are in the
right half-plane is equal to the number of sign
changes in the first column of Routh Table.

s3 1 31 0

How many sign s2 1 103 0

changes in this s1 -72 0 0
Routh table ???
s0 103 0 0

22
 Case 2
(Zero element in first column)
Un-stable system

23
Example1: (Case: 2) (Zero element in first column) Un-stable
system
Case 2 (Steps)
1. First Element of a row is Zero (0)
2. Replace “0” by small number . (=0.00000000..01)
3. Continue the array

Q(s) = s5 + 2s4 + 2s3 + 4s2 + 11s + 10

11 2 1
s5 1 2 11 b1     (1x4  2 x 2)  0
22 4 2
s4 2 4 10
1 an1 an3
then c1   
b1  0 b1 b2
s3 b1 0  6
let b1   (very very small value )
12 4 1 12
s2 c1 -12/ 10 then c1    (12  4 )  
s1 6   6  
s0 10 Two sign changes whether ‘’ is
assumed +ve or –ve. 2-roots in RHS 24
Example2: (Case 2)
Case 2 (Steps)
1. First Element of a row is Zero (0)
2. Replace “0” by small number . (=0.00000000..01)
3. Continue the array
Q(s) = 3s4 + 6s3 + 2s2 + 4s +5 (2.16)
s4 3 2 5
s3 6 4
RHP=2

s2 0 2 LHP=2
IA = 0
s1 -12/
s0 2
• Two sign changes whether ‘’ was
assumed +ve or –ve. 25
Example3: (Case 2)

Q(s) = s5 + s4 + 2s3 + 3s2 + s + 4

s5 1 2 1
s4 1 3 4
s3 -1 -3

s2 0 4
RHP=2
LHP=3
s1 4/
IA = 0

s0 4
• Two sign changes whether ‘’ was
assumed +ve or –ve. 26
 Home Assignment
PROBLEM:
Make a Routh table and tell how many roots of the
following polynomial are in the right half-plane and in the
left half-plane.

Don’t submit

27
Examples:
Stable,
Unstable,
and
Marginally Stable Systems

28
Example:1

C(s) 2 2 n2
T(s)  G(s)    
R(s) (s  1)( s  2) s2  3s  2 s 2  2 n  n2

Stable or Unstable?
S-plane
Characteristic equation (s+1)(s+2)=0
x x
-2 -1
Hence s+1=0 or s=-1 and s+2=0 or s=-2

The natural-response terms for the system are k1e-t and k2e-2t

29
 10 s  24 10s  24
Example 2:
T (s)  3 
s  2s  11s  12 ( s  1)( s  3) s  4)
2

Hence roots are s= -1, s= 3, & s= -4

Stable or Unstable? Unstable due to s=3

The natural-response terms for the system are k1e-t , k2e3t, and k3e-4t
Unstable due to term k2e3t

MATLAB Program: x x x
-4 -1 2.4 3
>> p = [1 2 -11 -12];
>> r = roots(p)

Note: Numerator 10s-24 has NO role in the stability of the system

30
 s s
Example:3 T(s)  
s2  1 (s  j)( s  j)

Hence roots are s= -j, s= j

xj
Stable or Unstable? Marginally stable
x -j
The natural-response term for the system is k sin(t+)

Marginally stable because no exponential term

MATLAB Program:
>> p = [1 0 1];
>> r = roots(p)

31
 s s
Example:4 T(s)  
s2  1 (s  j)( s  j)

Hence roots are s= -j, s=j Marginally stable

1
Let the input r(t) = sint  R(s)  x j

s2  1
s 1 x -j
C(s)  T(s)R(s) 
(s 2  1) (s 2  1)

The response of the system c(t) = t sint

Stable or Unstable? Unstable due to t

sint is marginally stable but the multiplication of time (t) makes it unstable

NOTE: The natural-response of the system are ksin(t+)

bounded (marginally stable) but produce unbounded
output (unstable) for certain bounded input 32
 Case 3
(All Element of a row are Zero
(Premature Termination )

33
Example1: (Case: 3)
Case 3: (Steps)
1. All Element of a row are Zero. (Premature Termination )
2. Auxiliary Polynomial
3. Aux Poly is differentiated with respect to ‘s’
4. Coefficients of the Polynomial replaces the zero row

Ex. Q(s) = s2 + 1

s2 1 1 Aux Pol = s2 + 1
s1 0
2 d/ds(s2 + 2)
s0 1
2s
No sign change in the first column
 implies that system is STABLE.
 Aux poly indicates roots on jw axis, which implies
MARGINALLY STABLE S2=-1; S=j 34
What is Auxiliary Polynomial ??
Q(s) = s3 + s2 + 2s + 2
= (s+1) (s2 + 2) s2 1 2
s1 0
Q(s) contains an even polynomial s0
as factor. An even polynomial is one
in which the exponents of s are
Aux Pol = s2 + 2
even integers or zero
s3 1 2
This even polynomial is called
“Auxiliary Polynomial” s2 1 2
s1 0
In Routh array, coefficients of “Aux s0
Poly” are those directly above the
zero row.
Aux Pol = s2 + 2
35
• Even polynomials only have roots that
are symmetrical about the origin.
• This symmetry can occur under three
conditions of root position:
(1) The roots are symmetrical and real
(2) The roots are symmetrical and
imaginary
(3) The roots are quadrantal

36
Example2: (Case: 3)
Case 3: (Steps)
1. All Element of a row are Zero. (Premature Termination )
2. Auxiliary Polynomial
3. Aux Poly is differentiated with respect to ‘s’
4. Coefficients of the Polynomial replaces the zero row

Ex. Q(s) = (s+1) (s2 + 2) = s3 + s2 + 2s + 2

s3 1 2
Aux Pol = s2 + 2

s2 1 2 d/ds(s2 + 2) = 2s
No sign change in the first column:
MARGINALLY STABLE
s1 02 RHP=0
LHP=1
s0 2 IA = 2 37
Example3: (Case: 3)

Ex. Q(s) = s4 + s3 + 3s2 + 2s + 2

s4 1 3 2
Aux Pol = s2 + 2

s3 1 2 d/ds(s2 + 2) = 2s

s2 1 2 No sign change in the first column:

MARGINALLY STABLE

s1 0
2
RHP=0
LHP=2
s0 2 IA = 2
38
Example4: (Case: 3)

Q(s) = s5 +2s4 + 8s3 + 11s2 + 16s + 12

s5 1 8 16
Aux Pol = 3s2 + 12
s4 2 11 12
d/ds(3s2 + 12)
s3 2.5 10 6s

s2 3 12
No sign change in the first column
MARGINALLY STABLE.
s1 6
0
 Aux poly indicates roots on jw axis
s0 12 3s2 = -12 s2 = -4 s1,2=j2
39
Example 5: (Case: 3)

s6 + s5 + 5s4 + s3 + 2s2 - 2s – 8
s6 1 5 2 -8

s5 1 1 -2
Aux Pol = 4s4 +4s2 - 8

s4 4 4 -8 d/ds(4s4 +4s2 - 8) = 16s3 +8s

One sign change in the first column:

s3 0
16 08 UNSTABLE

RHP=1
s2 2 -8 LHP=3
s1 72 IA =2
s0 -8 40
Unknown Parameter
in the
Characteristic Polynomial

41
Example 1:

P(s) = S4 + 2s3 + 4s2 + 2s + K

S4 1 4 K
S3 2 2

S2 3 K

S1 (6-2K)/3 → (6-2K)/3 > 0 Stable

S0 K → K> 0 Stable
Stability range 0<K<3
For what value of K the system will be marginally stable?
42
Example 2: In the figure below, determine the range of K for the system to
be stable

Solution: The characteristic equation is:

s 3  3s 2  2s  K  0
Construct the Routh array
For the system to be stable
we must have the following
two conditions satisfied
simultaneously:

For K = 0, the characteristic equation becomes: s3 + 3s2 + 2s + 0 = 0

43
For K = 6
the characteristic equation becomes
s3 + 3s2 + 2s + 6 = 0
or s2 (s + 3) + 2(s + 3) = 0

The frequency of oscillation is 2 rad/sec

44
Example 3

P(s) = S4 + 2s3 + 4s2 + Ks + 6

S4 1 4 6
S3 2 K

S2 4-K/2 6 → K < 8 Stable

12
S1 K → >0 Stable
4 K/2
S0 6
K2  4  16  24
  4K  12  0 K  roots complex
2 2( 0.5)
Stability range can not be satisfied for any value of K
(complex roots). Polynomial has RHP roots for all K 45
46