CONTROL THEORY

COURSE

Chapter 3: Stability of Control

System

Lecturer: Vũ Văn Phong

Content
 Definition of stablity of control system
 Routh-Hurwitz Criterion
 Root locus
 Frequency Criterion
Definition of Stability
Definition of Control System
 Definition: A stable system is a dynamic
system with a bounded response to a
bounded input.
Step Singal

system

Stable Neutral Unstable

Definition of Control System
Poles position
Definition of Control System
 Poles and Zeros
 Consider transfer function of system as below:

 Denominator:

Numerator:
Definition of Control System
 Poles: Poles are defined as the roots of
Denominator of transfer function. It means
that they are roots of equation: A(s)=0. Poles
donate pi with i=1,2 …,n
 Zeros: are defined as the roots of Numerator
of transfer function. It means that they are
roots of equation: B(s)=0. Poles donate Zi
with i=1,2 …,m
Definition of Control System
 Root locus Poles and Zeros
 Depict the positions of Poles and Zeros on
complex plane.

Poles
Zeros
Definition of Control System
Definition of Control System
 Stability of control system bases on position
of poles on complex plane:
 System is stable if all poles locate on left hand
side
 System is neutral if there are the poles locate
on imaginary (truc ao) axis
 System is unstable if there exist at least one
pole on the right hand side.
Definition of Control System
 Characteristic Equation: A(s)=0

Feedback control system State equation

Characteristic Equation Characteristic Equation

Routh-Hawirt Criterion
Routh-Hawirt Criterion

 Consider a characteristic equation of closed-loop

system:

 Necessary condition: All coefficients ( ai ) of

characteristic equation must be the same sign
and is not equal to zero
 Example:

Unstable

Unstable

Not sure
Routh-Hawirt Criterion
 Routh- Array
 Consist of n+1 rows
 First row is even coefficients of characteristic
equation
 Second row is odd coefficients of
characteristic equation
 The element at row i and columm j is
calculated by formula:
Routh-Hawirt Criterion
 Routh-Hawirt Array
Routh-Hawirt Criterion
 Routh Stability Criterion
 (1) All elements in the first row is positive. The
number of sign changes in first row is equal to the
number of roots in the right hand-side s-plane.

 (2) First row are positive and all of the

coefficients in the first column be positive.
Routh-Hawirt Criterion
• Routh Stability Criterion
(2) If the first element in a row is zero, it is replaced by a small ,   0,
and the sign changes when   0 are counted after completing the

array.
Routh-Hawirt Criterion
 Example:
Characteristic Equation
Routh-Hawirt Criterion
 Routh-Array
Routh-Hawirt Criterion
Routh-Hawirt Criterion
 Routh-Array
Routh-Hawirt Criterion
 Special Case: in Routh Array, if there is any
element of column 1 is equal to zero and
others are not, let’s change 0 into the small
positive constant  and then continue
carrying out the above steps.
 Example:
Routh-Hawirt Criterion

The system is unstable

 Special case 2: If all elements of any row
are zeros. We have to build a auxiliary
polynomial A0 ( s) from elements of previous
row. Then we change all zeros element by
the coefficients of polynomial dA0 ( s)
ds
 Example:
Routh-Hawirt Criterion
Routh-Hawirt Criterion
 Auxiliary polynomial A0 ( s )

 Note: roots of auxiliary polynomial A0 ( s ) are

also roots of characteristic equation.


 The system is neutral

Routh-Hawirt Criterion
 Hurwitz Matrix:

 Hurwitz Matrix is a nxn matrix

 Diagonal matrix is coefficients from a1 to an
Odd rows are the odd coefficients
Even rows are the even coefficients
 For example:

 Hurwitz criterion: The system is stable if

determinant of all diagonal matrix of Hurwitz
matrix is positive.
 Example:

 Determinants:
Root locus
What is Root Locus??

A graphical representation of
closed loop poles as a system
parameter varied
Why do we need Root Locus?

 Based on Root-Locus graph we can

choose the parameter for stability and
the desired transient response
What is the Root-Locus look like?
Pole 1

Pole 2
How to draw a Root-Locus graph?
 Transfer function:

 Characteristic equations:
 Rule 1: Find the zeros and poles.

 Define n is the number of poles and m is the

number of zeros of

Amplifier condition

Angle condition
 Rule 2: The number of branches of the root
locus equals the number of closed loop poles
 Rule 3:
 When K=0, then the root-locus start from the
poles.
 When K approach infinity, m branches will
approach to m zeros and n-m branches
approach infinity.
 Rule 4: Since the poles appear as complex
conjugate pairs, root locus is symmetric about
real axis
 Rule 5: If the total number of open loop poles
and zeros on the right of a point is odd then
that point is part of root-locus on the real axis.
 Rule 6: The angles between asymptotes and
real axis:
 Rule 7: The point A where the asymptotes
merge on the real axis
 Rule 8: Determine the break-away and
break-in point.
 Break-away point: The point where root-locus
leaves the real axis.
 Break-in point: The point where root locus
enters the real axis.
 Break-away and break-in points are the root
of the equation:
 Rule 9: Imaging Axis Crossing points.
 Use Routh-Hurwitz .The Routh-table has
zeros at a row
 Substituting into the characteristic
equation.
 Rule 10: Angles of Departure from the pole pj
Frequency Criterion
What is the Frequency Response?
 By the term of Frequency-Response, we mean the
steady-state response of a system to a sinusoidal
input. In Frequency-Response methods, we vary the
frequency of the input signal over a certain range and
study the resulting response

The Steady-State output of a transfer function system

can be obtained directly from the sinusoidal transfer
function, that is, the transfer function in which s is
replaced by j , where  is frequency.
Bode Diagram
 Transfer function in Frequency Domain:

C( jω)
G( jω)  G( jω)  P(ω)  jQ(ω)  A(ω)e jφ( ω)
R( jω)

 Magnitude:
A(ω)  G( jω)  P(ω)   Q(ω) 
 Phase:
 Q(ω) 
φ(ω)  Arg (G( jω))  arctg  
 P(ω) 
Bode Diagram
 Bode Diagram includes two parts:
 Magnitude:
 Plot of the logarithm
 of the magnitude of a
 transfer function
 when  change: 20logG() [dB].
 Phase:
 Phase angle of transfer
 Function when  changes
Bode diagram

Basic factors of G ( j ) H ( j )

The basic factors that very frequently occur in any arbitrary

transfer function are:

1.Gain K
2.Integral and derivative factor  j  1
3.First  order factors 1  jT 
1


4.Quadractic factors 1  2 ( j / n )  ( j / n ) 2  1
1. The Gain K
R2
Electrical circuit R1
-

+
R
R0

Transfer function G(s) = K dB

20lgK
Bode diagrams
lg
L( ) dB  20 lg G ( j )  20 lg K 


 2a. Integral
1
Transfer Function G (s ) 
s
   lg | G( jω) |  lg ω
 G( jω)  
jω φ(ω)  G( jω)   

20lg|G(j)| dB
() độ
Slope -20dB/dec

lg 
-90 0
2b. Derivative

Transfer function G (s )  s

  lg | G( jω) |  lg ω
 G( jω)  jω  
φ(ω)  G( jω)   

20lg|G(j)| dB () độ
Slope 20dB/dec
900
lg
lg
 ±1 1K
3. First - order factors ( 1+ jw T) G(s) =
Ts  1 (Som Pha)

The log of magnitude of the first-order factor 1 / 1  jT 

is: 1
20 lg  20 lg 1   2T 2 dB
1  jT
1
*    20 lg 1   2T 2  20 lg 1  0dB
T
1
*    20 lg 1   2T 2  20 lg T dB
T
The phase angle of the first-order factor 1 / 1  jT  is:
*  0    0
   arctan(T ) *   1    450
T
*       900
 Slope -20dB/Dec

The maximum error occurs at the corner frequency and

approximately equal to -3dB since:  20 lg 1  1  10 lg 2  3,03dB
 ±1
3. First - order factors ( 1+ jwT) G(s) = (Ts + 1) (Quan Tinh)

The log of magnitude of the first-order factor 1  jT 

is:
20 lg 1  jT  20 lg 1   T dB
2 2

1
*    20 lg 1   2T 2  20 lg 1  0dB
T
1
*    20 lg 1   2T 2  20 lg T dB
T
The phase angle of the first-order factor 1  jT  is:
*  0    0
  arctan(T ) *   1    450
T
*       900
 ±1
3. First - order factors ( 1+ jwT)

Slope 20dB/Dec

Log-Magnitude curve, together with the

asymptotes, and phase angle curve for 1  jT 
 4.Quadratic factors
Control systems open possess quadratic factors of the form:

ω n 
G( s )  

s  ξω n s  ω n T  s   ξTs  

1
n  Natural Frequency  ξ  Damping coefficient
T

khi    n Slope of asymptotes, -40dB/dec

 4.Quadratic factors

The asymptotic frequency response curve may be obtained as follows:

1
20 lg G  j   20 lg 
1  2 ( j / n )  ( j / n ) 2

 20 lg [1  ( / n )2 ]2  ( 2 / n )2
 2
4. Quadratic factors [1+ 2z jw / w n + jw / w n ]±1
( ) ( )
if   n the log magnitude become :  20 lg(1)  0 dB

if   n the log magnitude become :


 20 lg( / n )  40 lg
2
dB
And the phase-angle curves: n

if 0    1 : G ( j ) is the product of two complex  conjugate

factors

1  2 ( / n ) 
   arctan 2
1  2jT( j / n )  ( j / n ) 2
 1  ( / n ) 
 2
4. Quadratic factors [1+ 2z jw / w n + jw / w n ]±1
( ) ( )

At   0    00
At   n    900
At       1800
- Delay
Hàm truyền G (s )  e  Ts
 jTω   lg | G( jω) |  lg   
 G( jω)  e  cos Tω - jsinTω  
φ(ω)  G( jω)   Tω

|G(j)|dB lg

lg
Summary of Individual Factors
1 1
0 dB  
 
1
1  j  90

 90
1  j 0 dB

0 dB
 90
j

1 0 dB
 90
of
Department j EECS University of California, Berkeley
Bode Plots
Find magnitude and phase of each term
and sum them up!!!

K ( s  z1 )(s  z 2 )
G ( s) 
s m ( s  p1 )(s  p 2 ) mag(num)-mag(den)
K ( s  z1 ) ( s  z 2 ) phase(num)-phase(den)
G ( s) 
s m ( s  p1 ) ( s  p 2 )

20 log G ( s )  20 log K  20 log ( s  z1 )  20 log ( s  z 2 )

 20 log s m  20 log ( s  p1 )  20 log ( s  p 2 )

G ( s )  K  ( s  z1 )  ( s  z 2 )  s m  ( s  p1 )  ( s  p2 )
3. Procedure for plotting Bode diagram

Step 1: Transform the transfer function of system into the standard form
of the basic functions such as gain K, integral, derivative, first-order
factors, quadratic factors.

Step 2: Find all break frequencies and arrange them from small to big
 -(+)20*n

Step 5: If the bode diagram meet the break frequency of first-order

factor ( quan tinh) then the slope of Bode magnitude will increase to 20
db/dec.
If the bode diagram meet the break frequency of first-order factor ( som
pha) then the slope of Bode magnitude will reduce 20 db/dec.
If the bode diagram meet the break frequency of quadratic factor then
the slope of Bode magnitude will be deducted to 40 db/dec.
 Step 6: Plot the bode magnitude diagram
form left to right.
 Step 7: Plot the bode phase diagram
m

 s  c  m

n

 
e
G s  K e 1  arctg  arctg
n e 1
c e i 1 di
 s  d 
i 1
i
 Example 1: plot Bode diagram for the 100
system with following transfer function G (s ) 
s(s  1)(s  10)
Break frequency  (rad/s) at: 1, 10.
Because G(s) has zero poles therefore, |G(j0)|dB with 0 is selected
such that it is too small with respect to the smallest break frequency.
Choose 0 = 0,1 rad/s 100
| G ( j 0 ) | dB  20 lg  40dB
0,1.1.10
|G(j)| dB  
A   90  arctg  arctg
40 1 10
  0    90 0
0

      270
20 -2 -lg
- 1 lg
-900
1 -
20 -3
-2700
 10 5 (s  100)
Example 3: Plot Bode diagram : G (s ) 
(s  1)(s  10)(s  1000)
Break Frequency  (rad/s): 1, 10, 100, 1000.
Chú ý G(s) không có cực ở gốc.

10 7
Giản đồ Bode biên độ: | G ( j0) | dB  20 lg 4
 60dB
10

40dB
Bode stability Criterion

• Gain margin is the magnitude value when

the phase is -1800.

• Phase margin is the phase difference from

-1800 when the magnitude is zero.
 Gain m arg in

Phase m arg in
Bode stability Criterion

The system is stable if the Gain margin >0

and the Phase margin >0
 Plot Bode diagrams using Matlab
bode(num,den)
grid on
title(‘ title…’)

Example:
Stability via Bode plots

Using Matlab

G = tf(num,den)
Bode(G)
Margin(G)
Nyquist Criterion
Nyquist diagrams

The Nyquist diagram is a polar plot of the frequency

response of the system

On the Nyquist diagram the output, for a multi-amplitude sinusoidal input

at particular angular frequency, is specified by drawing a line of length
G ( j )
equal to the magnitude at the phase angle with the real axis.

Imaginary
R esponse at 
for G ( j )
G ( j )

 Input
0 Real
Nyquist diagrams

In plotting Nyquist diagram there are four key points to

be represented on the diagrams:

1. The start of the plot where 0

2. The end of the plot where 
3. Where the plot crosses the real axis, i.e.   0 or  180
0 0

4. Where it crosses the imaginary axes, i.e.    90

0
 Nyquist diagrams
First-order systems:

1
G s  where  is const ant
Transfer functions is: s  1
1
 G  j  
1  j

Magnitude: 1
G  j  
1    
2

Phase: j =0 - arctan(w t)
 Nyquist diagrams

1
( )
G jw =
2 j =0 - arctan(w t)
1+ wt
( )
Nyquist Diagram

Where :   0  G( j)  1,   00 0.1

Where :     G( j)  0,   900 0

-0.1

1 1
Imaginary Axis
Where :    G( j)  ,   450 -0.2

 2 -0.3

-0.4

-0.5

-0.2 0 0.2 0.4 0.6 0.8 1 1.2

Real Axis
Nyquist diagrams
Second-order systems:
n2
Transfer functions is: G  s   2
s  2n s  n2
where : n is freely oscillate angular frequency
 is the damping ratio

Magnitude:
1
G  j  
1   /     2  /   
n
2 2
n
2

Phase:  2   / n  
   arctan  2
 1    / n  
Nyquist diagrams

1
G  j  
1   /     2  /   
n
2 2
n
2

 2   / n  
Nyquist Diagram
1.5

   arctan  2
1

 1    / n   0.5

-0.5

Imaginary Axis
-1

-1.5

-2

-2.5

-3

-3.5
-1 0 1 2 3 4 5
Real Axis
Given a system.

K
R(s) C(s)
s( s  2) 2

where K = 5
Plot Nyquist diagram
Nyquist diagrams

Nyquist diagrams using Matlab

nyquist(num,den)
grid on
title(‘ title…’)
Example 1
Example 2
Example 3

Stable
Unstable
Unstable
Stable Unstable
THANKS YOU