Dynamic Systems and Control Theory

Lecture 6 : Stability
[Chapter 6 Nise, Chapter 5 Ogata]

Instructor: Dr. Ayse Tekes

Email: [email protected]
Learning Outcomes
— Make and interpret a basic Routh table to
determine the stability of a system
— Make and interpret a basic Routh table to
determine the stability of a system
— Make and interpret a Routh table whether the
first element of a row is zero or an entire row
is zero
— Use a Routh table to determine the stability of
a system represented in state space
 The concept of stability
Two types of response for LTI systems:
--Zero-state response: the response is due to input only; all
the initial conditions are zero;
Bounded-input-bounded-output(BIBO) stability: with zero
initial conditions, the system’s output y(t) is bounded to a
bounded input u(t).
--Zero-input response: the response is due to the initial
conditions only; all the inputs are zero;
Asymptotic stability: with zero input, for finite initial
conditions y(t0 ), y! (t0 ),", y ( n ) (t0 ) , a LTI system is asymptotic
stable there exist a positive number M which depends
on the initial conditions, such that 3

(1) y(t ) £ M < ¥ for all t > t0 ; and (2) lim y(t ) = 0.
t ®¥
Time-domain definition
The initial condition of the system is zero. When
system input is unit impulse function δ(t), the system
output is g(t).
If lim g (t ) = 0 , then the system is stable
t ®¥

n
Y ( s) n Ci
G ( s) = =å g (t ) = L-1[G ( s )] = å Ci e - pit
U ( s ) i =1 s + pi i =1

lim g (t ) = 0 Þe - pi t
decay with time
t ®¥
All poles should locate in the left side of s-plane
4
 Stability criterion in complex plane
A system is stable if and only if

all roots of the system characteristic equation have

negative real parts
or equivalently
all poles of closed-loop transfer functions must
locate in the left half of s-plane.

For LTI systems, both BIBO stability and asymptotic stability

have the same requirement on pole location. Thus if a system
is BIBO stable , it must also be asymptotic stable.

So we simply refer to the stability condition of a LTI system

as stable or unstable.
5
 LTI systems
Stability Conditions Location of poles

stable all poles in LHP

simple poles on the jw-axis

marginally stable
and no poles in RHF

at least one simple pole in

RHF or at least one multi-
unstable
order pole on the jw-axis
 Routh-Hurwitz stability criterion

All poles in left No poles in

s-plane > right s-plane

— The criterion tests whether any of the roots of the

characteristic equation lie in the right half of the s-
plane, without actually calculating the roots.
— Information about stability can be obtained directly
from the coefficients of the characteristic equation

7
 Routh-Hurwitz’s stability criterion

A necessary (but not sufficient) condition for

stability:

(1) All the coefficients of the characteristic

equation have the same sign.
(2) None of the coefficients vanishes.

Consider the characteristic equation of a LTI system

D( s ) = a0 s n + a1 s n -1 + ... + an -1 s + an = 0, a0 > 0

8
Example

• A sufficient condition for a system to be unstable is that all signs

of the coefficients of the denominator of the closed loop
transfer function are not the same.

• If powers of s are missing, the system is either unstable or, at

best, marginally stable.

• Unfortunately, if all coefficients of the denominator are positive,

and not missing, we do not have definite information about the
system’s pole locations. You better ask Matlab!!!
 Routh’s Criterion
A necessary and sufficient condition for stability:

all the elements of the first column of the Routh’

Tabulation are of the same sign.

The number of changes of signs in the elements

of the first column equals the number of roots in
the right-half s-plane.

10
Generating a Basic Routh Table
Example
s + 2 s + 3s + 4 s + 5 = 0
4 3 2

4
s 1 3 5
3
s 2 4
s2 1 5 Check
with
s1 -6 0 Matlab

0
s 5
Therefore, the system is unstable and has two roots
in the right-half s-plane.

13
Example, Nise

2 poles in the RHS

First-order: a0 s + a1 = 0
If a0 and a1 have the same sign, the system is stable.
Second-order: a0 s + a1 s + a2 = 0
2

If a0 , a1 and a2 have the same sign, the system is stable.

Third-order:

a0 s + a1 s + a2 s + a3 = 0
3 2

s3 a0 a2
s2 a1 a3 If a0,a1,a2,a3 are all
a0 a3 - a1a2 positive and
s 1
- a1a2>a3a0,the system is
a1
stable.
s0 a3
15
Special cases when applying Routh’s Tabulation
Case 1: only the first element in one of the rows of
Routh’s tabulation is zero
Solution: replace the zero with a small positive constant e
and proceed as before by taking the limit as e ® 0

s 4 + 3s 3 + 4s 2 + 12 s + 16 = 0 when e ® 0
s4 1 4 16 12e - 48 48
= 12 - <0
s3 3 12 e e
s2 0(e ) 16 The system is unstable and
12e - 48 has two roots not in the left-
s1 0 half s-plane.
e
s0 16 16
Example

There are two poles on the right half plane.

Case 2: an entire row of Routh’s tabulation is zero.

This indicates…
There are complex conjugate pairs of roots that are
mirror images of each other with respect to the
imaginary axis.
j j j

0 0 0

e.g. s1,2 = ±1 e.g. s1,2 = -1 ± j1 e.g. s1,2 = ±1 j

s3,4 = 1 ± j1 s3,4 = ±2 j
18
Example The characteristic equation of a system is
s 5 + 3s 4 + 3s 3 + 9s 2 - 4s - 12 = 0
Determine whether there are any roots on the
imaginary axis or in the RHP.
s
s 55
1
1
1 3 33 - 4 -- 44
s
s 44
3
3
3 9 99 - 12 -- 12
12
33 Introducing an auxiliary equation:
s
s 12
0
12 0118 8 0 00
A( s ) = 3s 4 + 9 s 2 - 12 = 0
s
s 22
2 -dA
9Its/ derivative: 12
(s)
= 12 s 3
+ 18s = 0
11 ds
s
s 50 0
s
s 00
-12
The sign in the first column changes once, so the system is
unstable and there is one root outside LHP. 19
Example [Nise]
Example [Nise]

-Find the number of poles in the right half plane.

 Application of Routh Tabulation
(1) Routh’s tabulation can not only be used to determine
the stability of a system, but also the spread of its
characteristic roots.
(2) For a control system with a regulator, Routh’s
tabulation can be used to select parameter values so that
the system is stable.
Determine the range of K so that the system is stable.
Example
R( s ) 10 Y ( s)
K
s ( s + 1)( s + 2)

2
23
 R( s ) 10 Y ( s)
K
s ( s + 1)( s + 2)

2
10k
C ( s) G forward ( s ) s( s + 1)( s + 2)
f ( s) = = =
R( s ) 1 + Gloop ( s ) 1 + 20k
s( s + 1)( s + 2)

Characteristic equation: 1 + Gl oop ( s ) = 0

20k
\1 + =0
s( s + 1)( s + 2)
24
Characteristic equation 20k
1+ =0
s( s + 1)( s + 2)

s + 3s + 2 s + 20k = 0
3 2

s3 1 2
s2 3 20k
3 ´ 2 - 20k
s 1
0 Þ 6 - 20k > 0 Þ k < 0.3
3
s 0
20k Þ k >0

So when 0 < k < 0.3 , the system is stable.

25
 Example, [Nise]

Find the range of gain K, for the system shown that will cause the system to
be stable, unstable and marginally stable.
Solution
If K<1386 --- > The system is stable with 3 poles in the Left Half Plane

If K>1386 --- > The system is unstable with 2 right half plane poles and 1
left half plane pole

If K=1386 (Not Fun) --- > Entire row is zero

If K=1386 --- > The system has two roots on the jw axis
Exercise
Consider that a 3rd-order system has the characteristic
equation 0.025s 3 + 0.325s 2 + s + k = 0
Determine the range of k so that the system is stable.
Solution. Reformulate the characteristic equation as

s + 13s + 40 s + 40k = 0
3 2

3
s 1 40
s2 13 40k
13 ´ 40 - 40k
s 1
Þ k < 13
13
s 0
40k Þ k >0 30
Dynamic Systems and Control Theory

Lecture 6 : Stability Example

Problems
[Chapter 6 Nise, Chapter 5 Ogata]

Instructor: Dr. Ayse Tekes

Email: [email protected]
Problem 1 --- > Bishop
Welding Control
Large welding robots are used in today’s auto
plants. The welding head is moved to different
positions on the auto body, and a rapid, accurate
response is required. A block diagram of a
welding head is shown below. Determine the
range of K and a for which the system is stable.
Solution
Problem 2 -- > Bishop, Skills Check
Problem 3
Problem 4 à Bishop
— A control system has the structure shown
below. Determine the gain at which the system
will become unstable.
 P(s) connection
=
parallel

+
P(s) =
-
1

t
TC =

= (2-5)k
s(s+ 1)(s + 4) + (2 -
5)k

↑ (s) = (2-5)k
s(s+ 1)(s + 4) + (2 -
5)k

Zhar
egei
, s(s + 1)(s + 4) + (2 -
s)k

her ege :
53 + 5s2 -
Ks + 4) + 2k

cher 5 + 55 + (4 17)s + 2k
-ege
-
Solution