Stability - System Analysis Techniques - Lecture Notes | ECSE 6400, Study notes of Electrical and Electronics Engineering

Material Type: Notes; Class: SYSTEMS ANALYSIS TECHNIQUES; Subject: Electrical & Comp. Sys. Engr.; University: Rensselaer Polytechnic Institute; Term: Fall 2004;

Typology: Study notes

Pre 2010

Uploaded on 08/09/2009

koofers-user-v96-1
koofers-user-v96-1 🇺🇸

10 documents

1 / 18

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
21. Stability
Topics
Be able to
.find equilibrium points.
define and give an example of
asymptotic stability.
define and give an example of Lyapunov
stability. .
.
.
.define and give an example of global
stability,
define and determine if a function is
positive definite, negative definite,
positive semi-definite, or negative semi-
definite. '
determine stability using Lyapunov's 2nd
Method.
.
.
.derive and use Lyapunov's Equation.
Ref: Sections 9.1-9.3, 9.6. Fall2004
pf3
pf4
pf5
pf8
pf9
pfa
pfd
pfe
pff
pf12

Partial preview of the text

Download Stability - System Analysis Techniques - Lecture Notes | ECSE 6400 and more Study notes Electrical and Electronics Engineering in PDF only on Docsity!

21. Stability

Topics

Be able to

. (^) find equilibrium points.

define and give an example of asymptotic stability.

define and give an example of Lyapunov

stability..

.

.

. (^) define and give an example of global

stability,

define and determine if a function is

positive definite, negative definite,

positive semi-definite, or negative semi-

definite. '

determine stability using Lyapunov's 2nd Method.

.

.

. (^) derive and use Lyapunov's Equation.

Ref: Sections 9.1-9.3, 9.6. Fall 2004

Equilibrium Points ~ f

A ball can be made to rest at A, E, F, G and anywhere between Band D:

Classify the Equilibrium Points

A andF are

E and G are

Cis

Time varying systems:

Asymptotic Stability

Def. An equilibrium point, xe, is asymptotically

stable if 1

. 1m II. el - t-> 00 II <1>(t,to' x(to), 0) - x 11 - 0

for any to and for all t > to'

If a trajectory reaches -xe,and no input is

applied, the system will stay at xe. Therefore,

ie(t) = (^0) for all t > to.

Example

Find the equilibrium points for a pendulum with damping. (^) ... Set) + Set) + ksinS(t) = 0 Solution

Stability in the Sense of

Lyapunov

Def. An e~uilibrium point, xe,is stable in the

sense oJ Lyapunov (i.s.L.) iff for any E> 0 , ~ 8(E) 3

IIxO - xell < 8(E) ~ II (t;to' xO, 0) - xell < E

for any to and for all t > to.

Loosely speaking, if an initial state, xo, is near xe then the response to Xowill also stay near xe.

Lyapunov's 2ndMethod

The system i(t) = f(x(t)) is asymptotically stable if:3 V(x(t)) 3.

Note: V(x(t)) satisfies the definition of a positive definite function.

V(x (1)) satisfies the definition of a negative definite function.

Example

If x(t) is 2 xl, explain asymptotic stability when

2 2 V(X (t» = Xl (t) + X2(t)

Explanation

Example

Investigate the stability of the following

.. nonlinear system,.

Xl (t) = X2 (t) - axl (t) [X12 (t) + X22 (t) ]

X2 (t) = -Xl (t) - ax2 (t) [X12 (t) + X22 (t) ]

Solution

Example

Classify 2 2 a.) Xl +X2 + 1

b.) Xl +X

c.) sin2XI +sin2X 2 2 d.) -Xl -X 2 2 2 e.) Xl +2X2 +X3 +2xlx

x is 2x 1 x is 2x 1

x is 2x 1

x is 3x 1

x is 3xl

Answers

a.)

b.)

c.)

d.)

e.)

When Is xTNx

  • nIl < Negative Definite?
    • nIl nl
    • nl2 n
    • nIl n12 nl
    • nl2 n22 n23 I <
    • nl3 n23 n

Table 9.8-1, page 478

IAI Rank andIndex CharacteristicValues A Form^ PrincipalLeading Minors Positive Definite Real quadraticform ~1'~2'... , ~fI

Hermitian form IAI>O r=p=fr^ Ai^ >^0

A =C7'C ICI>O A = C*7'C Iq> 0

Q = Yl1 + y,: +... + Yfl

Q (^) .-- z *z 1 1 + z *zIt. + ... +^ z *z" fI all positive

Positive. Semideflnite Real quadraticform. IAI = () Hermitian fortn

n - r roots r = p < n equalothers zero; positive

A = C7'C Iq = 0 A = CU'C ICI=O

Q = Yll+ YII + ... + yrl

A1'~2".. .Ar

all positive;

. Ar+1>'" , AfI-l non-negative Q - -11^ z *z^ +^ z *zII + ... +^ z *zrr

Negative Definite Real.quadratic form Hermitian form (-1)"^ IAI^ > () J'

r=1I p=O Ai<^0

A = -C7'C ICI>O A = -C*7'C ICI>O

Q = -Yll - Ytl-... - Ynl -A1'~I"'" (-:-l)"AfI Q - -^ -z *z 1 1 -^ z *%II -'"^ -^ z *zn n all positive

Negative Semidefinite Real quadraticform

Hermitian form IAI^ =^ ()

r<n p=O

n ,. r roots othersequal zero; negative

A = -CTC ICI=O A = -C*TC ICI=O

Q = -Yll- YII- ... - Yrl -AhAI".. , (-l)rAr all positive; Q = -zl*ZI - ZIzl - ... - zrzr (-l)r~r+1"'" (-l)"A"- non positive

Example

Investigate the stability of.

i(t) + O.li(t) + x(t) = 0 or

0 1 i(t) = I Ix(t) -1 -0.

Lyapunov' s Equation

. Generalize these ideas to. iCt) = Ax (t).

Let V(x(t)) = X(t)T Qx(t) where Q> 0, then,

V(x(t)) = d V(x(t)) - dt

Theorem: i(t) = Ax(t) is asymptotically

stable iff for any positive definite, symmetric

matrix N :3Q 3