Scribd
Upload
46 views18 pages

Automatic Control III (Reglerteknik III) Fall 2012: (Chapter 12)

This document discusses stability analysis techniques for nonlinear systems, including: 1) Linearizing nonlinear systems around an equilibrium point and analyzing the stability of the linearized system. 2) Using Lyapunov functions to prove global asymptotic stability by showing the Lyapunov function decreases along system trajectories. 3) Applying the circle/Nyquist criteria to analyze stability of systems with nonlinear feedback by reformulating them in a way that satisfies the assumptions of the small gain theorem.

Uploaded by

mareladra
Copyright
© Attribution Non-Commercial (BY-NC)
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
46 views18 pages

Automatic Control III (Reglerteknik III) Fall 2012: (Chapter 12)

This document discusses stability analysis techniques for nonlinear systems, including: 1) Linearizing nonlinear systems around an equilibrium point and analyzing the stability of the linearized system. 2) Using Lyapunov functions to prove global asymptotic stability by showing the Lyapunov function decreases along system trajectories. 3) Applying the circle/Nyquist criteria to analyze stability of systems with nonlinear feedback by reformulating them in a way that satisfies the assumptions of the small gain theorem.

Uploaded by

mareladra
Copyright
© Attribution Non-Commercial (BY-NC)
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
You are on page 1/ 18

Automatic Control III (Reglerteknik III) fall 2012

4. Nonlinear systems, Part 2


(Chapter 12)

Hans Norlander Systems and Control Department of Information Technology Uppsala University

REPETITION: EQUILIBRIA AND STABILITY 0 = (0 0 ) (0 0 ) equilibrium point

Nonlinear system: = ( ), Linearized model: = +

where

 (0 0 )  = 0 = 0
=

 (0 0 ) 

Let (0 ) be close to 0 and ()

stable if () 0 is small for all

0 , then (0 0 ) is 0 0 as

(globally) asymptotically stable if ()

(0 ))

unstable if it is not stable (Denitions 12.12) asymptotically stable/unstable if is asymptotically stable/unstable (Theorems 12.12)
Automatic Control III:4 part 2

Fall 2012

Page 1


Case 1:

STABILITY, EXAMPLE Case 2:


2 3 1 3 2

1 + 3 2

2 + 3 1

For both cases = 0 is a stationary point with


Fall 2012

1 0

eigenvalues

Theorems 12.12 does not apply what can be said about stability?

Automatic Control III:4 part 2

Page 2

STABILITY, EXAMPLE, contd

The equilibrium point can be a stable or unstable focus, or a center. The distance to the origin is

= Examine the squared distance: Set () = Case 1:


2

2 + 2 1 2

= 2 + 2 1 2

() = 21 1 + 22 2

() = 21 (2 3 ) + 22 (1 3 ) = 2(4 + 4 ) 1 2 1 2 () = 21 (2 + 3 ) + 22 (1 + 3 ) = 2(4 + 4 ) 1 2 1 2

Case 2: 0

The distance decreases for case 1, and increases for case 2.


Automatic Control III:4 part 2

Fall 2012

Page 3

STABILITY AND LYAPUNOV FUNCTIONS

Assume there is a function (), satisfying

( ) = 0

( )

0 for

=
 

() ()

 () =  1

Interpret () as a generalized distance from to . Along the solutions to the system = ()

(()) = (())() = (()) (())

Such a function is called a Lyapunov function.

Fall 2012

Automatic Control III:4 part 2

Page 4

STABILITY AND LYAPUNOV FUNCTIONS, contd

Theorem 12.3: If a Lyapunov function satisfying

(()) () = 0 for =

and ()

as

can be found, then the equilibrium point is globally asymptotically stable. Theorem 12.4: The rst condition can be replaced by: No solution to = () lies completely in the area where () () = 0. The tricky part is to nd the Lyapunov function!

Fall 2012

Automatic Control III:4 part 2

Page 5

STABILITY AND LYAPUNOV FUNCTIONS, contd

Let the properties of apply in a set


N

Md xo

: = : () Theorem 12.5: If (0) , then 0 as long as () . But the solution cannot leave . Hence, all solutions starting in
Introduce the subset will converge to the stationary point.
Automatic Control III:4 part 2

Fall 2012

Page 6

STABILITY AND LYAPUNOV FUNCTIONS, contd (

For the linear system

=
try the Lyapunov function

0 = 0 equilibrium)

() =
where is a positive denite matrix. Along the solutions (()) = () () + () () = ()( + )() = 0

If is a solution to the Lyapunov equation

then (()) = ()()

0 = 0 asymptotically stable.

Fall 2012

Automatic Control III:4 part 2

Page 7


STABILITY AND LYAPUNOV FUNCTIONS, contd Lemma 12.1: 1. If has all eigenvalues strictly in the left half plane, then for every matrix = 0 (or 0), there is a matrix = 0 (or 0) that satises the Lyapunov equation.

Fall 2012

2. If there are matrices = 0 and = 0 satisfying the Lyapunov equation, and ( ) is detectable, then has all eigenvalues strictly inside the left half plane.

Automatic Control III:4 part 2

Page 8

STABILITY AND LYAPUNOV FUNCTIONS, contd

Consider the nonlinear system

= + ()

() = ( )

Let have all eigenvalues in the left half plane, and let and be positive denite matrices fullling

Then () = is a Lyapunov function in an area around = also for the nonlinear system. (Theorem 12.6)

Fall 2012

Automatic Control III:4 part 2

Page 9

REPETITION: SMALL GAIN THEOREM (Theorem 1.1)

1 2

Assume 1 and stable if

2 stable. The closed loop system is input-output 1 2


1

Fall 2012

Automatic Control III:4 part 2

Page 10

THE CIRCLE CRITERION

Consider a linear system, with a nonlinear static feedback:

()

2
Assume (0) = 0 0

()
1

2
for = 0

( )

Direct application of small gain theorem guarantees stability if


2

sup

( )

which is too restrictive to be practical.


Automatic Control III:4 part 2

Fall 2012

Page 11


CIRCLE CRITERION, contd Set = and = 1+
1

+ 2
2

( ) = ( )

Fall 2012

Then

( )

Automatic Control III:4 part 2

Page 12

CIRCLE CRITERION, contd () + P +P

Reformulation of the closed loop system: +P

() ()

() ()

() Cancel each other

P +

Apply the small gain theorem for () = () 1+ () and ( ) = ( )

Fall 2012

Automatic Control III:4 part 2

Page 13


that is

CIRCLE CRITERION, contd

Applying the small gain theorem to the modied system description gives a sucient stability condition:

sup ( )
1 )

1 = ( )

( ( )

1 + ( )

Fall 2012

Interpretation: The curve 1 ( ) must not intersect the circle in the complex plane with center in and radius . More convenient and familiar if this can be reformulated to conditions on the Nyquist curve!
Automatic Control III:4 part 2

Page 14

THE NYQUIST CRITERION

Let be a semicircle with radius Im

() maps onto .
Im

()

Re Re

The Nyquist criterion: If encircles -1 times counter clockwise, the closed loop system has more poles in the RHP than () has. If encircles -1 times clockwise, the closed loop system has fewer poles in the RHP than () has.
Automatic Control III:4 part 2

Fall 2012

Page 15

CIRCLE CRITERION, contd

Theorem 12.7: Assume that () has no poles in the right half plane, ( ) and that (0) = 0, 1 2 for = 0. Then the closed loop system is input output stable if the Nyquist curve ( ), 0 does not enter, nor encircle the circle which intersects the negative real axis (perpendicularly) in 1 1 and 1 2 .
Im

1/k1 1/k

Re

Fall 2012

Automatic Control III:4 part 2

Page 16

CIRCLE CRITERION, contd () as long as () is

The circle criterion also applies for unstable stable. 5 Example: () = 5+2+ , 1 = 0 5, 2 = 2 2
2 1.5 1 0.5 0 0.5 1 1.5 2 3 2 1 0 1/k = 0
2

stable closed loop.

Im

1/k = 2
1

Re

Fall 2012

Automatic Control III:4 part 2

Page 17

You might also like