0% found this document useful (0 votes)
3 views30 pages

Lecturenote6 (Stability)

This lecture covers various aspects of stability in dynamic systems, including Internal Stability, BIBO Stability, and Lyapunov Stability. It discusses the conditions for stability in linear time-invariant (LTI) and linear time-varying (LTV) systems, along with the application of Lyapunov's methods for stability analysis. Key theorems and definitions related to equilibrium states and stability criteria are also presented.

Uploaded by

hieule123pro
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
3 views30 pages

Lecturenote6 (Stability)

This lecture covers various aspects of stability in dynamic systems, including Internal Stability, BIBO Stability, and Lyapunov Stability. It discusses the conditions for stability in linear time-invariant (LTI) and linear time-varying (LTV) systems, along with the application of Lyapunov's methods for stability analysis. Key theorems and definitions related to equilibrium states and stability criteria are also presented.

Uploaded by

hieule123pro
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

LECTURE 6

Stability

This lecture includes:


- Internal Stability
- BIBO Stability
- Lyapunov Stability

1
Stability Analysis for Linear Dynamic Systems

x0

u Dynamic System y

(1) Internal Stability (I-S)


 Assume : External Input u = 0 (or known fixed)
 Equilibrium xe  
x  f (x)  xe  f (xe )  0
at xe

 Behavior of x(t) near xe is of concern.


 “Stable” if x(t)  nhd(xe),  t  t0
(2) BIBO Stability
 If any BI guarantees BO, the I/O Description is called BIBO-S.

2
Internal Stability

For LTI systems,


(1) If the transient remains bounded in the nhd of “O” (or dies out), the system is
stable.
(2) Poles  LHP
x0

x
u y

For LTV systems,


a. “Internal Stability”

x(t )  A(t ) x(t ), x(t0 )  x0
Boundedness w.r.t. x0 & t0 .
How do you know the stability ?

b. Lyapunov Theory

x (t )  f ( x (t ), t )
  
 x  A (t ) x (t )  B (t )u (t ) 
 
3

Given x(t )  A(t ) x(t ) : Unforced System

The response due to x(t0 )  x0


x (t ) x ( t ; t 0 , x 0 )   (t , t 0 ) x 0

Definition 1
State xe is an equilibrium state at t0 iff
x(t )  x(t; t0 , xe )  t  t0
 xe
Definition 2
An equilibrium state xe is stable in the sense of Lyapunov
at t0 iff for every   0     (t0 ,  )  0
( depending on t0 and   0 ) s.t. if x0  xe   ,

( x(t; t 0 , x0 )  xe )   ,  t  t0

4
Definition 3
xe is uniformly stable i.s.L. iff for   0,    ( )  0 s.t.
x0  xe    x(t ; t0 , x0 )  xe   ,  t  t0

Remarks

(1) 0 is always an equilibrium for x  A(t ) x

(2) Let xe  0.

Then, for any   0,    ( ) s.t. x0    x(t , t0 , x0 )   t


if and only if  r  0 s.t.
x(t , t0 , x0 )  r x0

 Hence, for any  , if we choose    / r , then x0   implies that 


 
 x(t )   ,  t  t0 . 
  5
Theorem

Every equilibrium state of x  A(t ) x is stable in the sense of Lyapunov at t0
iff there exists some constant r which depends on t0 such that
 (t , t0 )  r  ,  t  t0 .
If r is independent of t0 , it is uniformly stable i.s.L.

Pf)
(Sufficiency)(  )

Let xe be an equilibrium state of x  A(t ) x; that is
xe   (t , t0 ) xe ,  t  t0 .
Then we have
x(t )  xe   (t , t0 ) x0  xe   (t , t0 )( x0  xe ),  t  t0 .
 x(t )  xe   (t , t0 ) x0  xe  r x0  xe  t  t0 .
Hence, for any  , if we choose    / r , then x0  xe   implies that
x(t )  xe   ,  t  t0 .
(Necessity)(  ) We can prove this by contradiction. (Details are omitted.)
6
Theorem

The zero state of x  A(t ) x is asymptotically stable at t0
iff (t , t0 )  r (t0 )   & (t , t0 )  0 as t  .
The zero state is asymptotically stable over [0, )
iff there exist positive numbers k1 and k2 such that for any t0  0
 (t , t0 )  k1e  k2 (t t0 ) ,  t  t0

Pf ) Omitted.

Note)

If x  A(t ) x is asymptotically stable, the zero state is the only equilibrium state.
( If there were another equilibrium state different from the zero state, then by
choosing that equilibrium state as an initial state, the response would not
approach the zero state.)

7
A. Stable i.s.L.

x2
x1

t0
t

x
r x0

x0
t0 t

8
Lyapunov Stability Criteria

Lyapunov Methods for Stability


- 1st Method, 2nd Method
- Lyapunov divides nonlinear systems into 2 groups

1st Class : Systems whose solutions are available.

The solution is examined and stability or

instability is noted.
2nd Class : Systems about which stability can be

determined directly by V-function*

( a Lyapunov function ),

9
- “The total energy of an unforced, dissipative mechanical

systems decreases as the state of the system evolves

in time”

Generalized concept  V function


x V


x x
x

10
(1) Consider

x(t )  A(t ) x(t ), x(t0 )  x0
Under what conditions do the solutions x(t) behave such
that x(t ) 2  0 (monotonically decreases) as t   ?

(observation)
2
Let x(t )  xT (t ) x(t ) .
d  
2 T
Then x(t )  x (t ) x(t )  xT (t ) x(t )
dt


 xT (t ) AT (t )  A(t ) x(t )
If AT (t )  A(t )  0 at each t , d x(t ) 2  0 and
dt
2
hence x(t ) decreases for t  t0

(not increasing)

11

x(t )  A(t ) x(t ), x(t0 )  x0 (1)

Theorem LS-II ( Uniform Exponential Stability )

Eq. (1) is uniformly exponentially stable


if  an n  n Q(t) such that

I  Q(t )  I (†)

T
A (t )Q(t )  Q(t ) A(t )  Q(t )  I (††)
where  ,  ,  are some finite positive numbers.

Brief Pf )
Let V (t )  xT (t )Q(t ) x(t ), where Q(t )  0
then V (t )  0 and

 

V (t )  xT (t )  AT (t )Q(t )  Q(t ) A(t )  Q(t )  x(t ).
 
 
So if A (t )Q(t )  Q(t ) A(t )  Q(t )  0, then V (t )  0.
T

It means that the system is asymptotically stable.


12
(  uniformly exponentially stable.)
Lyapunov Theory for Constant System

Q is Hermitian iff Q = Q* (complex conjugate transpose).

(1) Let Q be Hermitian. Then


2 2
min x  x*Qx  max x

(2) Hermitian Q is called positive (semi) definite iff

x*Qx  0  x0
( x*Qx  0  x  0)

13
(3) Hermitian Q is positive (semi) definite

 1. All eigenvalues are positive


(nonnegative)

2.  nonsingular N s.t.

Q = N* N.

(4) Lyapunov Theorem.



x  Ax is asymptotically stable iff , for any M > 0,
the equation
A*Q  QA   M (+)

has a unique solution Q > 0 with Q being Hermitian.

Note : (+) is called the Lyapunov equation.


14
( Proof )

() Let M  I , where   0.



Existence of Q s.t., with Q  0 ,

AT Q  QA   M

implies that all the hypothesis of Theorem LS-II are satisfied.


Therefore x  Ax, x(0)  x0 is exponentially stable.

(or asymptotically stable).

() Necessity can be proved by the following:


15
Theorem

If A has negative real part eigenvalues,


then, for each symmetric M,  ! solution of

AT Q  QA   M
of the form
 AT t
Q   e 0
Me At dt (++)
Further, if M > 0, then Q > 0.

16
( Proof ) If all eigenvalues of A have negative real parts,
Q in (++) is well defined. Now,
 At  At
A Q  QA  0 A e Me dt  0 e Me Adt
T T AT t AT t

 d
 0
dt
e A t Me At dt T

T
 e Me
A t At
0
 M

To prove uniqueness, suppose Qa is also a solution. Then

AT Qa  Q   Qa  Q  A  0

 e A t AT Qa  Q  e At  e A t Qa  Q  Ae At  0
T T

d  AT t
 e (Qa  Q)e At   0
dt  
T 
0  e A t (Qa  Q)e At 0
 (Qa  Q)

 Qa  Q
17
Finally, suppose that M > 0.

Q   e Me At dt
AT t
0

 Q is symmetric.

Now, for a x  0 , e At x  0 , and as



T
x Qx 
T
x T e A t Me At xdt  0
0

since ( e At x ) T M ( e At x )  0  t

 For x  0, xT Qx  0  Q  0.

18
Theorem

Every equilibrium state of x  Ax is stable i.s.L.
if all the eigenvalues of A have nonpositive real parts and
those with zero real parts are distinct roots of the minimal
polynomial of A.  
Note)
 
 minimal polynomial of A : the monic polynomial  ( A) 
 of least degree s.t.  ( A)  0 
Theorem  

The zero state is asymptotically stable iff all the


eigenvalues of A have negative real parts.
j
  m
t j 1 k t
Proof. 0
e At
dt  
0

k 1 j 1
Wk j
( j  1)!
e dt

 t j 1
  Wk j  e k t dt
0 ( j  1)! 
e Rei t
 finite 19
BIBO - Stability

x  
Input Output
System
u( t ) y (t )

Definition ( BIBO Stability )

A relaxed system (i.e., a system with zero initial

conditions) is BIBO - stable iff, for a bounded input,

the output is bounded.

20
(i.e u (t )  k1  , t  (, ), k1  ,

 k2   y (t )  k 2  , t  (, ))

x0  0
u (0) y (0)
system
u (t ) y (t )

k1 k2

t t

21
Theorem Ⅰ

A relaxed single-variable system


t
y (t )  

g ( t ,  ) u ( ) d 

is BIBO stable iff  k  s.t.

t

g ( t ,  ) d   k   ,  t  (  ,  ).

22
( Proof )
Sufficiency (←)

t
y (t )  

g ( t ,  ) u ( ) d
t
 g ( t ,  )  u ( ) d

t
 k   g ( t ,  ) d   k  k t .


Necessity (→)
t1
Suppose  g (t1 ,  ) d   for some t


Define u ( )  sgn  g (t1 ,  )

Then u ( )  1 but

y ( t1 )   ;  ( Contradiction )
23
Theorem Ⅱ
A relaxed MIMO System
t
y (t )   
G ( t ,  ) u ( ) d 

is BIBO stable iff


t
 
g ij ( t ,  ) d   k

for  i , j , t  (  ,  )

( Note )
G ( t ,  )   g i j ( t ,  )  i  1 , , q
j  1 , , p

Corollary to Theorem I
 t

 cf ) y (t )   u (t   ) g ( )d 
Time-invariant SISO case : 

0



   u (t   ) g ( ) d 
 
 
o

 
g (t) dt  k   


  g ( ) d  M


 0 
  k M   
 
 BIBO stable. 24
Theorem III ( BIBO stability of LTI System in terms of the system poles )

A relaxed SISO system with the transfer function g ( s)

given as a proper rational function is BIBO stable if

and only if all the poles of g ( s) have negative real parts.

( Proof ) : omitted

25
Internal Stability vs BIBO Stability

Theorem IV 
Consider x  A(t ) x  B(t )u
y  C (t ) x

If B ( t ) & C ( t ) are bounded on (, ) , then the


asymptotic stability of the zero-state ( equilibrium )
implies the BIBO stability of the zero-state response.
( The converse is not true. )

( Proof )

t

t0
C ( t )  ( t ,  ) B ( ) d   
t

t0
g (t , ) d    
  C ( t ) )   ( t ,  )  B ( ) d 

 k1  k 2   (t , ) d 


 k1  k 2 k  e d
 k3 (t  )
 
26
Ex 1)      
x x   u
    

y   x 
Note that
1
gˆ ( s ) 
s 1
det(I  A)  (  1)(  1)

The system is BIBO stable but the zero-state is not (internally) stable,
because  a positive eigenvalue.

Definition ( Total-Stability )
A system is T-stable (totally stable ) if for any initial state and for any

bounded input, the output and the state variable are bounded.

27
Theorem V

A system is T-stable if B (  ) and C (  ) are bounded



on (   ,  ) and the equilibrium (O) of x  A x

is asymptotically stable.

( Proof ) Exercise.

28
Time-Invariant Case

x  Ax  Bu
y  Cx
Recall :
Transfer Function G ( s)  C ( sI  A )  B
The zero-state response is BIBO stable iff all the poles
have negative real parts.

Theorem VI

Every equilibrium state of x  Ax is stable iff all
the eigenvalue of A have nonpositive real parts and
those with zero real parts are simple zeroes of the
minimal polynomial of A.

 Note) 
 
 minimal polynomial of A : the monic polynomial  ( A) 
 of least degree s.t.  ( A)  0  29
 
Theorem VII


The zero state of x  Ax is asymptotically stable iff
all the eigenvalues of A have negative real parts.

Theorem VIII

If an LTI dynamic equation FE is controllable and


observable, then the following are equivalent.
(1) Total stability

(2) BIBO stability

(3) Asymptotic stability

(4) All the poles have negative real parts

(5) All the eigenvalues of A have negative real parts.


30

You might also like