Stability Analysis

Prof. Enver Tatlicioglu

Department of Electrical & Electronics Engineering

Ege University

Chapter 5

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 1 / 72

BIBO stability

There are many classifications of stability in system analysis.

Figure: Stability, marginal stability and instability

For LTI systems, they all are basically same which is bounded input
bounded output (BIBO) stability.
If input is bounded (i.e., |u (t)| < K1 ), then output is also bounded
(i.e., |y (t)| < K2 ).

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 2 / 72

BIBO stability
In the time domain, the output has the form
+∞
Z
y (t) = h (τ ) u (t − τ ) dτ.
−∞

Bounding the output results in

+∞
Z
|y (t)| = h (τ ) u (t − τ ) dτ
−∞
+∞
Z
≤ |h (τ ) u (t − τ )| dτ
−∞
+∞
Z
= |h (τ )| |u (t − τ )| dτ.
−∞

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 3 / 72

BIBO stability

After noting that |u (t)| < K1 ∀t, we obtain

+∞
Z
|y (t)| ≤ K1 |h (τ )| dτ.
0

So if a system is BIBO stable, then

+∞
Z
|h (τ )| dτ < ∞.
0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 4 / 72

BIBO stability
In the Laplace domain, the transfer function has the form
m
Q
(s − zi )
K
Y (s)
H (s) = = ni=1
R (s) Q
(s − pi )
i=1

where we consider physical systems that satisfy m ≤ n.

We first consider distinct poles.
So, from inverse Laplace transform, we obtain
n
X
h (t) = ki exp (pi t)
i=1

thus
+∞
Z
|h (τ )| dτ < ∞ iff < {pi } < 0 ∀i.
0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 5 / 72

BIBO stability

If poles are not distinct, for an m–th order pole, h (t) will have terms
of the form
k
t (m−1) exp (pi t)
(m − 1)!
thus
+∞
Z
|h (τ )| dτ < ∞ iff < {pi } < 0 ∀i.
0

Notice that, this is the same condition as the distinct poles case.
So for Laplace domain condition for stability, all poles of the transfer
function must be in the open left half plane (i.e., none on jω axis).

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 6 / 72

BIBO stability
Example

2
Is T (s) = s 2 +3s+2
stable?
Note that
2
T (s) =
(s + 2) (s + 1)
where the poles are at s = −1 and s = −2.
Since all poles of the transfer function are in the open left half plane,
the system is stable.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 7 / 72

BIBO stability
Example

10s+24
Is T (s) = s 3 +2s 2 −11s−12
stable?
Note that
10 (s + 2.4)
T (s) =
(s + 1) (s − 3) (s + 4)
where the roots are at s = −1, s = 3 and s = −4.
Since not all poles of the transfer function are in the open left half
plane, the system is unstable.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 8 / 72

BIBO stability
Example
2s
Is T (s) = s 2 +1
stable?
Note that
2s
T (s) =
(s + j) (s − j)
where the roots are at s = ±j.
Since the poles of the transfer function are on jω axis, the system is
marginally stable.
Consider r (t) = sin (t), then the output is found as
2s 1 2s
Y (s) = T (s) R (s) = =
s2 2
+1s +1 (s 2 + 1)2
⇒ y (t) = t sin (t) 1 (t)

which obviously grows unbounded as time increases.

So marginally stable systems are unstable because some bounded
inputs may result in unbounded output.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 9 / 72
BIBO stability

Is there a relation between the coefficients of the denominator of the

transfer function and the type of stability?

Denominator of the transfer function Type of stability

s 2 + 3s + 2 stable
s 3 + 2s 2 − 11s − 12 unstable
s 2 + 0s + 1 marginally stable

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 10 / 72

Routh Hurwitz stability criterion

Factoring higher order polynomials to find their poles is tedious and

numerically not well conditioned.
We want other stability tests and we also want to find margins of
stability (i.e., how far is the system from instability?).
Routh Hurwitz stability criterion considers the denominator of the
transfer function without factoring.
– Routh test is a very mechanical test.
– It is not intuitive.
– Its proof is very difficult.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 11 / 72

Routh Hurwitz stability criterion

For a second order system, the denominator of the transfer function

has the form

a (s) = s 2 + a1 s + a0
= (s − p1 ) (s − p2 )
= s 2 − (p1 + p2 ) s + p1 p2 .

Note that, the poles are either real or they are complex conjugates.
Since, for stability, we need < {pi } < 0, it is clear that

− (p1 + p2 ) = a1 > 0
p1 p2 = a0 > 0

are required for stability.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 12 / 72

Routh Hurwitz stability criterion
For a third order system, the denominator of the transfer function has
the form

a (s) = s 3 + a2 s 2 + a1 s + a0
= (s − p1 ) (s − p2 ) (s − p3 )
= s 3 − (p1 + p2 + p3 ) s 2 + (p1 p2 + p1 p3 + p2 p3 ) s − p1 p2 p3 .

Note that, the poles are either real or two of them are complex
conjugates.
Since, for stability, we need < {pi } < 0, it is clear that

− (p1 + p2 + p3 ) = a2 > 0
(p1 p2 + p1 p3 + p2 p3 ) = a1 > 0
−p1 p2 p3 = a0 > 0

are required for stability.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 13 / 72
Routh Hurwitz stability criterion: Case 0
Trend: Stability is equivalent to none of the coefficients ai being less
than or equal to 0.
Does this trend continue?
Note that
an−1 = (−1) × sum of all poles
an−2 = sum of products of poles taken two at a time
an−3 = (−1) × sum of products of poles taken three at a time
..
.
a0 = (−1)n × product of all poles.
From this trend, we can conclude that:
? If any coefficient ai = 0, then not all poles in LHP.
? If any coefficient ai < 0, then at least one pole in RHP.
So our first test criteria is:
? If any coefficient ai ≤ 0, then the system is unstable.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 14 / 72
Routh Hurwitz stability criterion
Example

Check the stability of a (s) = s 2 + 1.

Since the coefficient of s 1 is equal to 0, we can conclude that not all
poles are in LHP.
The poles are at s = ±j, so the system is marginally stable.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 15 / 72

Routh Hurwitz stability criterion
Example

Check the stability of a (s) = s 3 + 2s 2 − 11s − 12.

Since the coefficient of s 1 and the constant term are negative, we can
conclude that there is at least one pole in RHP.
The poles are at s = −1, s = −4, and s = +3, so the system is
unstable.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 16 / 72

Routh Hurwitz stability criterion
Example

Check the stability of a (s) = s 3 + s 2 + 2s + 8.

We can not conclude if this system is stable or not using our first test
criteria.
√
1 15
The poles are at s = −2, s = 2 ±j 2 , so the system is unstable.

How do we find out this without factoring?

Once we have determined that ai ≥ 0, we can run the Routh test to
find the number of poles in the RHP.

Important: In this course, we consider the case where the

denominator of the transfer function is monic.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 17 / 72

Routh Hurwitz stability criterion: Case 1
We form the Routh array for a (s) = s n + an−1 s n−1 + · · · + a1 s + a0 as
sn 1 an−2 an−4 · · ·
s n−1 an−1 an−3 an−5 · · ·
s n−2 b1 b2 b3
s n−3 c1 c2 ···
..
.
s1 j1
s0 k1
where the entries of the third line are obtained as
−1 1 an−2 −1 1 an−4
b1 = , b2 = ,···
an−1 an−1 an−3 an−1 an−1 an−5
and similarly, the entries of the fourth line are obtained as
−1 an−1 an−3 −1 an−1 an−5
c1 = , c2 = ,···
b1 b1 b2 b1 b1 b3
number of unstable poles = number of sign changes in left column.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 18 / 72
Routh Hurwitz stability criterion
Example
Check the stability of a (s) = s 3 + s 2 + 2s + 8.
The Routh array is formed as
s3 1 2
s2 1 8
s 1 b1
s 0 c1
1 2 −1
where b1 = 1 = −6.
1 8
Updated Routh array is
s3 1 2
s2 1 8
s 1 −6
s 0 c1
−1 1 8
where c1 = −6 = 8.
−6 0
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 19 / 72
Routh Hurwitz stability criterion
Example

Final Routh array is

s3 1 2
s2 1 8
s 1 −6
s0 8
where in the left column, there are two sign changes

1 → −6 and − 6 → 8

thus there are two poles in RHP.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 20 / 72

Routh Hurwitz stability criterion
Example

Check the stability of a (s) = s 2 + a1 s + a0 .

The Routh array is
s 2 1 a0
s 1 a1
s 0 b1
1 a0
−1
where b1 = a1 = a0 .
a1 0
Final Routh array is
s 2 1 a0
s 1 a1
s 0 a0
where there will be no sign changes in the left column iff

a1 > 0 and a0 > 0.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 21 / 72

Routh Hurwitz stability criterion
Example

Check the stability of s 3 + a2 s 2 + a1 s + a0 .

The Routh array is
s 3 1 a1
s 2 a2 a0
s 1 b1
s 0 c1

−1 1 a1 −1 a0
where b1 = a2 = a2 (a0 − a1 a2 ) = a1 − a2 .
a2 a0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 22 / 72

Routh Hurwitz stability criterion
Example
Updated Routh array is
s3 1 a1
s 2 a2 a0
a0
s 1 a1 − a2
s0 c1

−1 a2 a0
where c1 = a a0 = a0 .
a 1 − a0
2
a1 − a2 0
Final Routh array is
s3 1 a1
s 2 a2 a0
s 1 a1 − aa02
s0 a0
where there will be no sign changes in the left column iff
a0
a2 > 0 , a1 > , a0 > 0.
a2
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 23 / 72
Routh Hurwitz stability criterion: Case 2

Sometimes, an element in the left column is equal to 0.

This may cause the first element of the next row be found as
constant
.
0
In this case, we replace this element with  as  → 0.
We count the number of sign changes in the left column for both
positive and negative  values.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 24 / 72

Routh Hurwitz stability criterion
Example

Check the stability of a (s) = s 5 + 2s 4 + 2s 3 + 4s 2 + 11s + 10.

The Routh array is
s5 1 2 11
s4 2 4 10
s3 b1 b2
s2 c1 c2
s1 d1
s0 e1

−1 1 2 −1 1 11
where b1 = 2 = 0 and b2 = 2 = 6.
2 4 2 10

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 25 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s 5 1 2 11
s 4 2 4 10
s3 0 6
s 2 c1 c2
s 1 d1
s 0 e1

Notice that, since we have 0 in the first element of the s 3 row, we

replace it with .

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 26 / 72

Routh Hurwitz stability criterion
Example

New Routh array is

s 5 1 2 11
s 4 2 4 10
s3  6
s 2 c1 c2
s 1 d1
s 0 e1

−1 2 4 −1 −12
where c1 =  =  (12 − 4) =  + 4 and
 6
−1 2 10
c2 =  = 10.
 0
Since for  → 0, | −12
 | >> 4, and the sign of c1 is not affected by the
presence of 4, for simplicity reasons, we can use c1 ≈ −12
 .

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 27 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s5 1 2 11
s4 2 4 10
s3  6
−12
s2  10
s1 d1
s0 e1

−1 6  
10 + 72

10 2
where d1 = −12 = 12
−12  = 12  + 6 and since as
 10 
10 2
 → 0, then 12  → 0, for simplicity reasons, we can use d1 ≈ 6.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 28 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s5 1 2 11
s4 2 4 10
s3  6
−12
s2  10
s1 6
s0 e1
−12
−1  10
where e1 = 6 = 10.
6 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 29 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

s 5 1 2 11 s5 1 2 11
s 4 2 4 10 s4 2 4 10
old s 3 0 6 new s 3  6
2 −12
s c1 c2 s2  10
1
s d1 s1 6
s 0 e1 s0 10

– if  > 0, there are two sign changes in the left column,

– if  < 0, there are also two sign changes in the left column.
So there are two poles in RHP.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 30 / 72

Routh Hurwitz stability criterion: Case 3
Sometimes an entire row in a Routh array is equal to 0.
This means that there are polynomial factors such that one factor has
conjugate mirror poles.

We need to complete Routh array by forming the polynomial ā (s)

from the last non–zero row in the array.
Note that, ā (s) is a factor of a (s).
Notice that, ā (s) has even or odd orders of s only.
Since it is missing some orders of s, it is unstable by case 0.
We want to see if it has roots in the RHP or only on the jω axis.
We replace the zero row with d ā(s)
ds and then continue.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 31 / 72
Routh Hurwitz stability criterion
Example

Check the stability of a (s) = s 4 + s 3 + 3s 2 + 2s + 2.

The Routh array is
s4 1 3 2
s3 1 2
s 2 b1 b2
s 1 c1
s 0 d1

−1 1 3 −1 1 2
where b1 = 1 = 1 and b2 = 1 = 2.
1 2 1 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 32 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s4 1 3 2
s3 1 2
s2 1 2
s 1 c1
s 0 d1

−1 1 2
where c1 = 1 = 0.
1 2

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 33 / 72

Routh Hurwitz stability criterion
Example
Updated Routh array is
s4 1 3 2
s3 1 2
s2 1 2
s1 0
s 0 d1

where the entire s 1 row is 0.

Thus, from the s 2 row, we form
ā (s) = s 2 + 2
and its derivative with respect to s is equal to
d ā (s)
= 2s
ds
from which the new s 1 row is obtained.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 34 / 72
Routh Hurwitz stability criterion
Example

New Routh array is

s4 1 3 2
s3 1 2
s2 1 2
s1 2
s 0 d1

−1 1 2
where d1 = 2 = 2.
2 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 35 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

s4 1 3 2 s4 1 3 2
s 3 1 2 s3 1 2
s2 1 2 s2 1 2
old s 1 0 new s 1 2
s 0 d1 s0 2

where there are no sign changes in the left column.

Thus there are no RHP poles.
√
The poles of ā (s) are at s = ±j 2, so the system is marginally stable.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 36 / 72

Routh Hurwitz stability criterion
Example
Check the stability of a (s) = s 4 + 4.
The Routh array is
s4 1 0 4
s3 0 0
s 2 b1 b2
s 1 c1
s 0 d1
where the entire s 3 row is 0.
Thus, from the s 4 row, we form
ā (s) = s 4 + 4
and its derivative with respect to s is equal to
d ā (s)
= 4s 3
ds
from which the new s 3 row is obtained.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 37 / 72
Routh Hurwitz stability criterion
Example
New Routh array is
s4 1 0 4
s3 4 0
s 2 b1 b2
s 1 c1
s 0 d1
1 0 −1 −1 1 4
where b1 = 4 = 0 and b2 = 4 = 4.
4 0 4 0
Updated Routh array is

s4 1 0 4
s3 4 0
s2 0 4
s 1 c1
s 0 d1

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 38 / 72

Routh Hurwitz stability criterion
Example

Note that, since we have 0 in the first element of the s 2 row, we

replace it with .
New Routh array is
s4 1 0 4
s3 4 0
s2  4
s 1 c1
s 0 d1

−1 4 0 −16
where c1 =  =  .
 4

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 39 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s4 1 0 4
s3 4 0
s2  4
−16
s1 
s0 d1

−1  4 − 64
where d1 = −16 −16 = −16

= 4.
  0 

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 40 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

s4 1 0 4 s4 1 0 4 s4 1 0 4
old s 3 0 0 new s 3 4 0 s3 4 0
2
s b1 b2 old s 2 0 4 new s 2  4
−16
s 1 c1 s 1 c1 s1 
0
s d1 0
s d1 s0 4

– if  > 0, there are two sign changes in the left column,

– if  < 0, there are also two sign changes in the left column.
So there are two poles in RHP.
Other two poles are mirrors in LHP.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 41 / 72

Routh Hurwitz stability criterion
Example
Find the values of K to make the following closed–loop system stable.
R (s) + s+1
Y (s)
+ K s(s−1)(s+6)
−

The transfer function can be found as

Y (s)
T (s) =
R (s)
s+1
K s(s−1)(s+6)
= s+1
1 + K s(s−1)(s+6)
K (s + 1)
= .
s (s − 1) (s + 6) + K (s + 1)
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 42 / 72
Routh Hurwitz stability criterion
Example

The denominator of the transfer function is found as

a (s) = s (s − 1) (s + 6) + K (s + 1)
= s 3 + 5s 2 + (K − 6) s + K .

We will check the stability of a (s).

The Routh array is
s3 1 K − 6
s2 5 K
1
s b1
s 0 c1

−1 1 K −6 −1 4K −30
where b1 = 5 = 5 (K − 5K + 30) = 5 .
5 K

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 43 / 72

Routh Hurwitz stability criterion
Example
Updated Routh array is
s3 1 K −6
s2 5 K
4K −30
s1 5
s0 c1

−1 5 K
where c1 = 4K −30 4K −30 = K.
5 5 0
Final Routh array is
s3 1 K −6
s2 5 K
4K −30
s1 5
s0 K
4K −30
where there will be no sign changes in the left column iff 5 >0
and K > 0.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 44 / 72
Routh Hurwitz stability criterion
Example

Figure: Unit step response for different values of K

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 45 / 72

Routh Hurwitz stability criterion
Example
Find the values of K to make the following closed–loop system stable.
R (s) +  
1
Y (s)
+ K 1 + T1I s (s+1)(s+2)
−

The transfer function can be found as

Y (s)
T (s) =
R (s)
KTI s+K
TI s(s+1)(s+2)
= KTI s+K
1 + TI s(s+1)(s+2)
KTI s + K
= .
TI s (s + 1) (s + 2) + KTI s + K
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 46 / 72
Routh Hurwitz stability criterion
Example
The denominator of the transfer function is found as
a (s) = TI s 3 + 3TI s 2 + TI (2 + K ) s + K
where its monic form is obtained as
K
a (s) = s 3 + 3s 2 + (2 + K ) s + .
TI
We will check the stability of the monic form of the denominator of
the transfer function.
The Routh array is
s3 1 2 + K
K
s2 3 TI
s 1 b1
s 0 c1
1 2+K  
I −K
where b1 = −1 3 K = −1 K
3 TI − 6 − 3K = 6TI +3KT
3TI .
3 TI
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 47 / 72
Routh Hurwitz stability criterion
Example

Updated Routh array is

s3 1 2+K
K
s2 3 TI
6TI +3KTI −K
s1 3TI
s0 c1

K
−1 3 TI K
where c1 = 6TI +3KTI −K 6TI +3KTI −K = TI .
3TI 3TI 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 48 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

s3 1 2+K
K
s2 3 TI
6TI +3KTI −K
s1 3TI
K
s0 TI

where there will be no sign changes in the left column iff

6TI +3KTI −K
3TI > 0 and TKI > 0.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 49 / 72

Routh Hurwitz stability criterion
Example

Figure: Unit step response for different values of K and TI

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 50 / 72

Routh Hurwitz stability criterion
Example

Find the values of K to make the following system stable

a (s) = s 5 + 5s 4 + 10s 3 + 10s 2 + 5s + K .

The Routh array is

s5 1 10 5
s4 5 10 K
s3 b1 b2
s2 c1 c2
s1 d1
s0 e1

−1 1 10 −1 1 5 25−K
where b1 = 5 = 8 and b2 = 5 = 5 .
5 10 5 K

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 51 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s5 1 10 5
s4 5 10 K
25−K
s3 8 5
s 2 c1 c2
s 1 d1
s 0 e1

−1 5 10 55+K −1 5 K
where c1 = 8 25−K = 8 and c2 = 8 = K.
8 5 8 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 52 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s5 1 10 5
s4 5 10 K
25−K
s3 8 5
55+K
s2 8 K
s1 d1
s0 e1
25−K
−1 8 5 −K 2 −350K +1375
where d1 = 55+K 55+K = 5(55+K ) .
8 8 K
Note that, e1 = K .

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 53 / 72

Routh Hurwitz stability criterion
Example
Final Routh array is

s5 1 10 5
s4 5 10 K
25−K
s3 8 5
55+K
s2 8 K
2
−K −350K +1375
s1 5(55+K )
s0 K
55+K
where there will be no sign changes in the left column iff 8 > 0,
−K 2 −350K +1375
5(55+K ) > 0 and K > 0.
55+K
From 8 > 0, we have K > −55.
= −(K +353.88)(K −3.88)
2
Note that, −K 5(55+K
−350K +1375
) 5(55+K ) > 0, thus
3.88 > K > −55 or −353.88 > K .
Combining these with K > 0 results in 3.88 > K > 0.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 54 / 72
Routh Hurwitz stability criterion
Example
Find the values of K to make the following system stable

a (s) = s 5 + 1.9s 4 + 5.1s 3 + 6.2s 2 + (2 + K ) s + 4K .

The Routh array is

s 5 1 5.1 2 + K
s 4 1.9 6.2 4K
s 3 b1 b2
s 2 c1 c2
s 1 d1
s 0 e1

−1 1 5.1
where b1 = 1.9 = 1.83 and
1.9 6.2
−1 1 2+K
b2 = 1.9 = 2 − 1.1K .
1.9 4K
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 55 / 72
Routh Hurwitz stability criterion
Example

Updated Routh array is

s5 1 5.1 2+K
s 4 1.9 6.2 4K
s 3 1.83 2 − 1.1K
s 2 c1 c2
s 1 d1
s 0 e1

−1 1.9 6.2
where c1 = 1.83 = 1.14K + 4.12 and
1.83 2 − 1.1K
−1 1.9 4K
c2 = 1.83 = 4K .
1.83 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 56 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s5 1 5.1 2+K
s 4 1.9 6.2 4K
s3 1.83 2 − 1.1K
s 2 1.14K + 4.12 4K
s 1 d1
s0 e1

where
−1 1.83 2 − 1.1K −(1.25K 2 +9.57K −8.24)
d1 = 1.14K +4.12 = 1.14K +4.12 .
1.14K + 4.12 4K
Note that, e1 = 4K .

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 57 / 72

Routh Hurwitz stability criterion
Example
Final Routh array is

s5 1 5.1 2+K
s4 1.9 6.2 4K
s3 1.83 2 − 1.1K
s2 1.14K + 4.12 4K
−(1.25K 2 +9.57K −8.24)
s1 1.14K +4.12
s0 4K

where there will be no sign changes in the left column iff

−(1.25K 2 +9.57K −8.24)
1.14K + 4.12 > 0, 1.14K +4.12 > 0 and 4K > 0.
From 1.14K + 4.12 > 0, we have K > −3.61.
−1.25(K 2 +7.65K −6.59)
Note that, 1.14K +4.12 = −1.25(K −0.78)(K +8.43)
1.14K +4.12 > 0, thus
0.78 > K > −3.61 or −8.43 > K .
Combining these with K > 0 results in 0.78 > K > 0.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 58 / 72
Routh Hurwitz stability criterion
Example

Consider the following system

R (s) + Y (s)
+ K s s+α
2 +ω 2
1
s(s+1)
0
−

Use Routh Hurwitz stability criteria to find the range of K (in terms
of ω0 and α) such that the closed–loop system remains stable.
Next, use ω0 = 1 and α = 0.25 to find the range of K numerically.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 59 / 72

Routh Hurwitz stability criterion
Example

The transfer function is found as

Y (s)
T (s) =
R (s)
K s s+α 1
2 +ω 2 s(s+1)
0
=
1 + K s s+α 1
2 +ω 2 s(s+1)
0

K (s + α)
= .
s 2 + ω02 s (s + 1) + K (s + α)

The denominator of the transfer function is

a (s) = s 2 + ω02 s (s + 1) + K (s + α)

= s 4 + s 3 + ω02 s 2 + ω02 + K s + αK .

We need to check the stability of a (s).

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 60 / 72
Routh Hurwitz stability criterion
Example

Routh array is
s4 1 ω02 αK
s 3 1 ω02 + K
s 2 b1 b2
1
s c1
s 0 d1

−1 1 ω02 −1 1 αK
where b1 = = −K and b2 = = αK .
1 1 ω02 + K 1 1 0

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 61 / 72

Routh Hurwitz stability criterion
Example

Updated Routh array is

s4 1 ω02 αK
s 3 1 ω02 + K
s 2 −K αK
s 1 c1
s 0 d1

where
−1 1 ω02 + K 1
αK + ω02 K + K 2 = α + ω02 + K .


c1 = −K = K
−K αK
Note that, d1 = αK .

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 62 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

s4 1 ω02 αK
s3 1 ω02 +K
s2 −K αK
s 1 α + ω02 + K
s0 αK

where there will be no sign changes in the left column iff K < 0,
α + ω02 + K > 0 and αK > 0.
Since we are given that α = 0.25, from αK > 0, we need K > 0, but
since K < 0, we can conclude that no values of K will result in
stability of the closed–loop system.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 63 / 72

Routh Hurwitz stability criterion
Example

Consider the following system

R (s) + kI 1
Y (s)
+ kP + s + kD s s 2 +K
−

Use Routh Hurwitz stability criteria to find the range of K (in terms
of kP , kD and kI ) such that the closed–loop system remains stable.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 64 / 72

Routh Hurwitz stability criterion
Example
The transfer function is found as
Y (s)
T (s) =
R (s)
 
kP + kD s + ksI s 2 +K
1
=  
1 + kP + kD s + ksI s 2 +K1

kP s + kD s 2 + kI
= .
s (s 2 + K ) + (kP s + kD s 2 + kI )

The denominator of the transfer function is

a (s) = s s 2 + K + kP s + kD s 2 + kI

= s 3 + kD s 2 + (kP + K ) s + kI .

We need to check the stability of a (s).

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 65 / 72
Routh Hurwitz stability criterion
Example

The Routh array is

s3 1 kP + K
s 2 kD kI
s 1 b1
s 0 c1
1−1 kP + K kD kP +kD K −kI
where b1 = kD = kD .
kD kI
Note that, c1 = kI .

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 66 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

s3 1 kP + K
s2 kD kI
kD kP +kD K −kI
s1 kD
s0 kI

where there will be no sign changes in the left column iff

kD kP + kD K − kI
kD > 0 , > 0 , kI > 0.
kD

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 67 / 72

Routh Hurwitz stability criterion
Example
Modify the Routh Hurwitz stability criterion so that it determines the
number of poles that are at the left of −2 for a (s) = s 2 + a1 s + a0 .
Let p = s + 2 thus s = p − 2.
Substituting s = p − 2 in a (s) results in
ā (p) = a (s = p − 2) = (p − 2)2 + a1 (p − 2) + a0
= p 2 + (a1 − 4) p + 4 − 2a1 + a0 .
Routh array is
p2 1 4 − 2a1 + a0
p 1 a1 − 4
p 0 4 − 2a1 + a0
where the poles of a (s) are to the left of −2 iff a1 − 4 > 0 and
4 − 2a1 + a0 > 0. From a1 − 4 > 0, we have a1 > 4, and from
4 − 2a1 + a0 > 0, we have 4 + a0 > 2a1 , so combining these results in
4 + a0 > 2a1 > 8 for the poles of a (s) to be at the left of −2.
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 68 / 72
Routh Hurwitz stability criterion
Example
Determine the number of poles of a (s) = s 2 + 2s + 1 that are at the
left of −2.
Note that, the roots of a (s) are at −1 (i.e., double roots).
Let p = s + 2 thus s = p − 2.
Substituting s = p − 2 in a (s) results in

a (p) = a (s = p − 2) = (p − 2)2 + 2 (p − 2) + 1
= p 2 − 2p + 1.

Routh array is
p2 1 1
p 1 −2
p0 1
where there are two roots on the modified right half plane (where the
modified right half plane is to the right of −2).
Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 69 / 72
Routh Hurwitz stability criterion
Example

Determine the number of poles of a (s) = s 2 + 4s + 3 that are at the

left of −2.
Note that the roots of a (s) are at −1 and −3.
Let p = s + 2 thus s = p − 2.
Substituting s = p − 2 in a (s) results in

a (p) = a (s = p − 2) = (p − 2)2 + 4 (p − 2) + 3
= p 2 − 1.

Routh array is
p 2 1 −1
p1 0
p 0 b1
where the entire p 1 row is 0.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 70 / 72

Routh Hurwitz stability criterion
Example

Thus, from the p 2 row, we form

ā (p) = p 2 − 1

and its derivative with respect to p is found as

d ā (p)
= 2p.
dp
Updated Routh array is

p 2 1 −1
p1 2
p 0 b1

Note that, b1 = −1.

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 71 / 72

Routh Hurwitz stability criterion
Example

Final Routh array is

p 2 1 −1 p 2 1 −1
old p1 0 new p1 2
p 0 b1 p 0 −1

where there is only one sign change (i.e., 2 → −1), so we have one
pole on the modified right half plane (where the modified right half
plane is to the right of −2).

Dr. Enver Tatlicioglu (EEE@EgeU) Control Systems 1 Chapter 5 72 / 72