5/8/2020

Concept of Stability Stable System

• Stability of a Cone
• Natural (transient) response → 0 as t →∞

Unstable System Marginally Stable System

• Natural (transient) response → neither decays nor
• Natural (transient) response → ∞ as t →∞ grows

1
 5/8/2020

Stability Definition Stability Definition

(1) Asymptotic Stability (3) S-domain Stability
Stable system if the transient response decays to System Transfer Function : T(s)
zero y (n ) ( t )  a y (n 1) ( t )        a y( t )  0 Stable system if the poles of T(s) all lies in the left-
n1 0
half s-plane. j
I.C. y(0) , y (1) (0 ) ,       , y (n1) (0)
Unstable
lim y( t )  0 stable
t poles
poles
(2) BIBO Stability 
Stable system if the response is bounded for
bounded input signal y( t )  t g( t  ) u( ) d
 0

u( t )  M  y( t )  N
Marginally stable/unstable
The impulse response of a system is absolutely
The definitions of (1), (2), and (3) are equivalent
integrable.
for LTI system.

Location of Poles Stability Analysis

• Time Response vs. Pole Location  BIBO: Bounded Input Bounded Output
systems.
 For LTI systems this requires that all poles
of the closed-loop transfer function lie in
the left half of the complex plane.
 Determine if the transfer function has any
poles either on the imaginary axis or in the
right half of the s-plane.

2
 5/8/2020

Examples Routh-Hurwitz Criterion

20
M (s)  BIBO stable or a.s.
( s  1)( s  2)( s  3)  Any pole of system lies on the right hand side of
the origin of s plane, it makes system unstable.
20( s  1)
M (s)  unstable  On basis of this condition A. Hurwitz and E.J.
( s  1)( s 2  2 s  2) Routh started investigating necessary and
20( s  1)
sufficient conditions of stability of a system.
M (s)  marginally stable and NOT a.s.  Routh-Hurwitz criterion: determine if any roots of
( s  2)( s 2  4)
a polynomial lie outside left half of complex
10 plane.
M (s)  unstable
( s 2  4)2 ( s  10)
 Routh-Hurwitz criterion allow us to determine
M (s) 
10 stability without finding roots of characteristic
NOT a.s.
s ( s 3  30 s 2  s  10) equation.
 It does not find the exact locations of the roots.

Routh-Hurwitz Criterion Routh-Hurwitz Criterion

• Routh Tabulation (array)
• Characteristic Polynomial of Closed-loop System sn an a n 2 a n  4 
sn1 an1 an 3 an 5 
D( s)  ansn  an1sn1        a1s  a0 , an  0   1  a n an  2   1  an a n  4
s n2
b1 b2 b3  b1    , b 2    
 an1  an1 an 3  an1  an1 an 5
Hurwitz polynomial sn  3 c 1 c2 c3 
All roots of D(s) have negative real parts. stable system   1 a an  3   1 a a n 5
c 1    n1 , c 2    n1 
Hurwitz’s necessary conditions: All coefficients (ai) are to be positive.  b1  b1 b2  b1  b1 b3
1
s g1 0 
Define
s0 h1 0
D0 (s)  an sn  an 2sn 2   
• Routh-Hurwitz Stability Criterion
D1(s)  an 1sn 1  an3 sn 3    (1) The polynomial D(s) is a stable polynomial if  i are all positive, i.e.
an , an1, b1, c1,    h1 are all positive.
D (s) 1 a a b (2) The number of sign changes in an , an 1, b1, c 1,    h1 is equal to the
 0  1s  , 1  n ,  2  n1 ,  3  1 ,   
D1(s) 1 an 1 b1 c1 number of roots in the RH s-plane.
 2s 
(3) If the first element in a row is zero, it is replaced by a small ε, ε>0,
 3s  and the sign changes when   0 are counted after completing the

array.
Note: Any zero root has been removed in D(s). (4) If all elements in a row are zero, the system has poles in the RH plane
or on the imaginary axis.

3
 5/8/2020

Routh-Hurwitz Criterion Completed Routh table

T(s)
Equivalent
closed-loop
transfer function

Initial layout for

Routh table

Case - 1 Case - 2
No zero appear First element of a row is zero

Example: Q(s) = s3 + s2 +2s+ 8 Q(s) = s5 + 2s4 + 2s3 + 4s2 +11s+ 10

s3
1 2 s5 1 2 11
s2 1 8 s4 2 4 10
s1 – 6 s3 0/  6  is a small number
s0 8 s2 –12/ 10
2 sign change  2 roots lie in RHP s1 6
s1 10
2 sign change  2 roots lie in RHP

4
 5/8/2020

Case - 3 Case - 3
All elements of a row is zero Method 1
First factorized the polynomial and analyze each factor
Q(s)= s5 + s4 + s3 + s2 +2s +2 individually
s5 1 1 2 For the above polynomial we have
s4 1 1 2 s5 + s4 + s3 + s2 +2 s +2 = (s+1)(s4 +s2 +2)
s3 0 0 We then create Routh array for s+1 and for s4+s2+2
• A case 3 polynomial contains even polynomials, Method 2
called auxiliary polynomial, as a factor. Differentiate the aux polynomial
• Row above zero row indicates auxiliary polynomial Coefficient of the result replaces the zero row
• Auxiliary polynomial of s5 + s4 + s3 + s2 +2 s +2 is
Coefficient of 4s3 + 2s replaces the zero
Qa = s4 +s2 +2
s3 4 2
• Two method available to solve this problem.

Example Routh-Hurwitz Criterion

Ex: For a closed-loop system with characteristic polynomial
Q(s) = s4 +4 D(s)  s 5  2s 4  3s3  6s 2  5s  3
s4 1 0 4 Determine if the system is stable
s3 0 0 Sol: Routh Table Sign of first column for   0,   0
We have zero row here. Aux-polynomial is s5 1 3 5 s5 
s4 2 6 3 s4 
Qa(s) = s4 +4 s 3
0ε
7
s 3 
2
Qa’(s) = 4s3 j s2
6ε - 7
3 s 2 
two sign changes  two poles in RHP
ε
s1 
 unstable system
42ε - 49 - 6ε 2
The array becomes s 1

12ε -14 s0 

s4 1 0 4 s0 3

1 Ex: For D(s)  s3  2s2  4s  8, determine if the system is stable

s3 4 0
Sol: Routh Table 1 j
s2  4 D( s)  ( 2s2  8)( s  1)
2 2
s3 1 4
s1 -16/ 2
 (s2  4)(s  2)
s 2 8  Auxiliary eq. : D(s)  2s  8
2

s0 4  (s  2j)(s  2j)(s  2)  2
s1 poles : s  -2j, 2j, - 2
2 sign change  2 roots lie in RHP
0
 2
s0  stable system

5
 5/8/2020

Exercise Solution

Unstable System

Exercise Exercise
s3  3408.3s2 1204000s 1.5 107 K  0 K  273.57  A(s)  3408.3s2  4.1036 109  0
s3 1 1204000 s   j1097 roots when K  273.57
s2 3408.3 1.5 107 K
410.36 107 1.5 107 K
s1 0
3408.3
s 0
1.5107 K
410.36 107 1.5 107 K
a.s.  0 & 1.5 107 K  0
3408.3
For the system to be stable
0  K  273.57

6
 5/8/2020

Exercise Example
We have control system as follow
compensator plant
2
K
s  4 s 2  5s  2
3

Suppose that, for constant input, we want ess<2%.

The position error constant Kp is

 2 
K p  lim  K 3   K
s 0
 s  4s2  5s  2 

Steady state error is ess =1/(1+K) < 2%, thus K > 49.

Example Example
However we have to check the stability, and we will use To overcome the previous problem we have to
Routh-Hurwitz criterion
The CE is found to be
use PI controller
compensator plant
s3 + 4s2 +5s+ 2+2K = 0 2
KP+ KI /s
The array becomes s  4 s 2  5s  2
3

s3 1 5
s2 4 2+2K
s1 (18-2K)/4  K<9
• Using this controller ess= 0
s0 2 + 2K  K>-1
Here for stability requirement we must have • We will find values of KP and KI to assure the
-1<K<9 stability
We conclude that, using proportional compensator, our • Characteristic equation is found to be
requirement to have ess<2% cannot be satisfied. s4 +4s3 + 5s2 +(2+2KP)5s+ 2KI = 0

7
 5/8/2020

Example Root Locus Technique

The array becomes
s4 1 5 2KI Investigation of the paths of the roots of the
s3 4 2+2KP characteristic equation or simply the Root Locus
s2 (18-2KP)/4 2KI  KP<9
s1 c Root locus is a powerful tool for designing and
s0 2KI  KI > 0 analyzing feedback control systems
where Locus of roots in the s-plane can be determined by
4  ( 2  2 K p )(18  2 K p ) 
c 8K I  a graphical method
18  2 K p  4 

In general, root locus may be sketched by

4
18  2 K p

(1  K P )(9  K p )  8 K I  following some simple rules and properties
For plotting the root locus accurately the MATLAB
If we choose KP = 3<9 then c = (24-8KI)/3)>0
root locus tool in the Control System Toolbox
Thus 0< KI <30
can be used

Definition of Root Locus Definition of Root Locus

Roots of this equation are the closed-loop poles of Set of all points in the s-plane that satisfy
the feedback system the equation 1 + kG(s) = 0 for some value
Locus - set of all points whose location is of k
determined by stated conditions
Root locus is a graphical presentation of the
Stated conditions here are that 1 + kG(s) = 0 for
closed-loop poles as a system parameter
some value of k, and the “points” whose
locations matter to us are points in the s-plane is varied
Path taken by the root of the characteristic Root locus is a powerful method of analysis
equation when open loop gain k is varied from 0 and design for stability and transient
to ∞ are called root locus response

8
 5/8/2020

Root Locus Technique Root Locus Technique

Let's say we have a closed-loop transfer function In other words, the locations of the poles of a
for a particular system specific equation must satisfy the following
relationship

Where N is the numerator polynomial and D is the And from the above equation we can manipulate
denominator polynomial of the transfer functions, an equation such as
respectively.
Now, we know that to find the poles of the
equation, we must set the denominator to 0, and And finally by converting to polar coordinates, we
solve the characteristic equation get

Root Locus Technique Example

Now we have 2 equations that govern the Camera control system
locations of the poles of a system for all
gain values
Magnitude Equation
Angle Equation

How the dynamics of the camera changes as K is

varied ?

9
 5/8/2020

Effect of K on Pole Locations Pole Plots from the Table

From Root Locus Root Locus Technique

Technique which shows how changes in one
of a system’s parameter (usually the
controller gain, K) will modify the location
of the closed-loop poles in the s-domain

10
 5/8/2020

Root Locus Construction Rules Root Locus Construction Rules

Rule 1: Starting Point (K=0) Rule 4: Symmetry of the Root Loci
The root locus starts at open loop poles. Or The root loci are symmetrical with respect to the
there is one branch of the root-locus for every real axis and all complex roots are conjugate.
root of b(s). Rule 5: Angle of Asymptotes
Rule 2: Terminating Point (K=infinity) The root loci are asymptotic to straight lines at
The root locus terminates at open loop zeros large values and the angle of asymptotes is
which include those at infinity. given by
Rule 3: Number of Distinct Root Loci
There will be as many root loci as the highest
number of finite open loop poles or zeros.

Root Locus Construction Rules Root Locus Construction Rules

Rule 6: Asymptotic Intersection Rule 8: Locus Breakaway Point
The asymptotes intersects the real axis at the The points at which the root locus break away
point given by can be calculated by the following:

Rule 7: Root Locus Location on the Real Rule 9: Angle of Departure and Arrival
Axis Rule 10: Point of Intersection with the
The root loci may be found on portions of the Imaginary Axis
real axis to the left of an old number of open
loop poles and zeros.
Rule 11: Determination of K

11
 5/8/2020

Angle of Departure Example

Sketch the root locus of the following

system

Determine the value of K such that the

damping ratio ζ of a pair of dominant
complex conjugate closed-loop is 0.5
Angle of departure =180°- (ϴ1+ϴ2 )+Φ

Step #1 Step #2
Draw the n poles and m zeros Determine the loci on the
of G(s)H(s) using x and o real axis:
respectively
Choose a arbitrary test
1 point.
G s  
ss  1s  2  If the TOTAL number of
both real poles and zeros
3 poles:
is to the RIGHT of this
p1 = 0; p2 = -1; p3 = -2
point is ODD, then this
No zeros point is on the root locus

12
 5/8/2020

Step #3 Step #4
Determine the n - m
asymptotes: Find the breakpoints.
Locate s = α on the real Express K such as:
axis: 1
K   ss  1s  2 
q  0, 1, 2, 
G( s )H ( s )
K   s 3  3s 2  2 s
Compute and draw
angles: Set dK/ds = 0 and solve
  1800 2  0  1 for the
   600 2 poles.
 0 3 0  3s  6 s  2  0

   180 2  1  1  1800 s1  1.5774, s2  0.4226
0


1
3 0

Exercise
Step #5
Points on imaginary axis
satisfy:
s  j jω?
Points on root locus
satisfy:
1  KG s H s   0
- jω
Substitute s=jω into the
characteristic equation (-0.784,0)
and solve for ω.   0 or    2

13
 5/8/2020

Exercise Exercise

Exercise Solution

14
 5/8/2020

Exercise Exercise

15
5/8/2020

16