Scribd
Upload
61 views

C21 Nonlinear Systems: Introduction and Concepts of Stability

This document provides an introduction to nonlinear control systems. It discusses key concepts such as stability of equilibrium points, multiple equilibria, limit cycles, and chaotic behavior in nonlinear systems. Examples are given including mass-spring-damper systems, the Van der Pol and Lorenz attractors. Stability is defined for equilibrium points, and the region of attraction is introduced. Linear and nonlinear system properties are compared.

Uploaded by

Pra Surya
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
61 views

C21 Nonlinear Systems: Introduction and Concepts of Stability

This document provides an introduction to nonlinear control systems. It discusses key concepts such as stability of equilibrium points, multiple equilibria, limit cycles, and chaotic behavior in nonlinear systems. Examples are given including mass-spring-damper systems, the Van der Pol and Lorenz attractors. Stability is defined for equilibrium points, and the region of attraction is introduced. Linear and nonlinear system properties are compared.

Uploaded by

Pra Surya
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/ 12

C21 Nonlinear Systems

Mark Cannon
4 lectures webpage: www.eng.ox.ac.uk/conmrc/nlc

Michaelmas Term 2012

0-1

Lecture 1

Introduction and Concepts of Stability

1-2

Course outline

1. Types of stability 2. Linearization 3. Lyapunovs direct method 4. Regions of attraction 5. Linear systems and passive systems

1-3

Books
J.-J. Slotine & W. Li Applied Nonlinear Control, Prentice-Hall 1991.
Stability Interconnected systems and passive systems

H.K. Khalil Nonlinear Systems, Prentice-Hall 1996.


Stability Passive systems

M. Vidyasagar Nonlinear Systems Analysis, Prentice-Hall 1993.


Stability & passivity (more technical detail)

1-4

Why use nonlinear control?


Real systems are nonlinear
friction, non-ideal components actuator saturation sensor nonlinearity

Analysis via linearization


accuracy of approximation? conservative?

Account for nonlinearities in high performance applications


Robotics, Aerospace, Petrochemical industries, Process control, Power generation . . .

Account for nonlinearities if linear models inadequate


large operating region model properties change at linearization point

1-5

Linear vs nonlinear system properties

Free response Linear system x = Ax


Unique equilibrium point: Ax = 0 x = 0 Stability independent of initial conditions

Nonlinear system x = f (x)


Multiple equilibrium points f (x) = 0 Stability dependent on initial conditions

1-6

Linear vs nonlinear system properties


Forced response Linear system x = Ax + Bu
u nite x nite if open-loop stable Frequency response: u = U sin t x = X sin(t + ) Superposition: u = u1 + u2 x = x 1 + x 2

Nonlinear system x = f (x, u)


u nite x nite No frequency response u = U sin t x sinusoidal No linear superposition u = u1 + u2 x = x1 + x2

1-7

Example: step response


Mass-spring-damper system
k(x) m c(x)

u
x

Equation of motion: x + c(x) + k(x) = u c(x) = x k k(x) nonlinear:


0 x

Input u(t)
6 50 5 4 3 30 x(t) u 2 1 0 10 1 2 5 10 t 15 20 3 0 5

Response x(t)

40

20

0 0

10 t

15

20

apparent damping ratio depends on size of input step


1-8

Example: multiple equilibria


First order system: x = f (x)
f (x)

xc

xb

xa

x > xa xb < x < x a xc < x < x b x < xc

= = = =

f (x) > 0 f (x) < 0 f (x) > 0 f (x) < 0

= = = =

x(t) x(t) x(t) x(t)

increases decreases increases decreases

xa , xc are unstable equilibrium points xb is a stable equilibrium point


1-9

Example: limit cycle


Van der Pol oscillator: x + (x2 1)x + x = 0 Response x(t) tends to a limit cycle (= trajectory forming a closed curve) Amplitude independent of initial conditions

4 3 2

1 x, dx/dt dx/dt

1 0 1 2

3 4 3 2 1 0 x 1 2 3

3 0

0.5

1.5 t

2.5

Response with x(0) = 0.05, dx/dt = 0.05

State trajectories x(t), x(t)

1 - 10

Example: chaotic behaviour


Strange attractor

1 - 11

Example: chaotic behaviour


Lorenz attractor Simplied model of atmospheric convection: x = (y x) z = xy z y = x( z) y

State variables x(t): y(t): z(t): uid velocity dierence in temperature of acsending and descending uid characterizes distortion of vertical temperature prole

Parameters = 10, = 8/3, = variable

1 - 12

Example: chaotic behaviour


Lorenz attractor = 28 = strange attractor:

60

40 z 20 0 20 10 0 10
1 - 12

40 20 0 20 20 40 y

Example: chaotic behaviour


Lorenz attractor sensitivity to initial conditions

60

40 z 20 0 20 10 0 10
1 - 12

40 20 0 20 20 40 y

Example: chaotic behaviour


Lorenz attractor sensitivity to initial conditions blue: (x, y, z) = (0, 1, 1.05) red: (x, y, z) = (0, 1, 1.050001)

20 0 20 50 0 50 60 40 z 20 0
1 - 12

10

15

20

25

30

35

40

10

15

20

25

30

35

40

10

15

20 t

25

30

35

40

Example: chaotic behaviour


Lorenz attractor = 99.96 = limit cycle:

200 150 100 50 0 40 20 0 20 40


1 - 12

100 50 0 50 60 100 y

Example: chaotic behaviour


Lorenz attractor = 14 = convergence to a stable equilibrium:

30

20 z 10 0 15 10 5 0 y
1 - 12

15 10 5 5 0 x

State space equations


x = f (x, u, t) x : state u : input e.g. nth order dierential equation: dn y dn1 y = h y, . . . , n1 , u, t dtn dt

has state vector (one possible choice) x1 y x2 y x= . = . . . . . n1 n1 xn d y/dt

1 - 13

and state space dynamics: x1 x2 x2 x3 x= . = = f (x, u, t) . . . . . xn h x1 , x2 , . . . , xn , u, t

Equilibrium points
x is an equilibrium point of system x = f (x) i: x(0) = x i.e. f (x ) = 0 Examples: (a) y + y 2 + sin y = 0 (damped pendulum) y n x= , x = , n = 0, 1 y 0 implies x(t) = x t > 0

(b) y + (y 1)2 y + y sin(y/2) = 0 y 0 1 1 x= , x = , , y 0 0 0 Consider local stability of individual equilibrium points Convention: dene f so that x = 0 is equilibrium point of interest Autonomous system: x = f (x)
1 - 14

x = constant

Stability denition
An equilibrium point x = 0 is stable i: max x(t) can be made arbitrarily small
t

by making x(0) small enough

for any R > 0, there exists r > 0 so that x(0) < r = x(t) < R t > 0
R r x(0) 0 x(t)

Is x = 0 a stable equilibrium for the Van der Pol oscillator example? No guarantee of convergence to the equilibrium point
1 - 15

Asymptotic stability denition


An equilibrium point x = 0 is asymptotically stable i: (i). x = 0 is stable (ii). x(0) < r = x(t) 0 as t (ii) is equivalent to: for any R > 0, x(0) < r = x(t) < R
x r

t > T for some r, T

R 0 T t

1 - 16

Exponential stability denition


An equilibrium point x = 0 is exponentially stable i: x(0) < r = x(t) et t > 0

exponential stability is a special case of asymptotic stability

r et


0 t

1 - 17

Region of attraction
The region of attraction of x = 0 is the set of all initial conditions x(0) for which x(t) 0 as t

x(0) 0 x(t)

region of
attraction

Every asymptotically stable equilibrium point has a region of attraction r= = = entire state space is a region of attraction x = 0 is globally asymptotically stable

Are stable linear systems asymptotically stable?

1 - 18

Summary

Nonlinear state space equations: x = f (x, u) x = state vector, u = control input Equilibrium points: x is an equilibrium point of system x = f (x) if f (x ) = 0 Stable equilibrium point: x is stable if state trajectories starting close to x remain near x at all times Asymptotically stable equilibrium point: x must be stable and state trajectories starting near x must tend to x asympotically Region of attraction: initial conditions from which state trajectories converge to an asymptotically stable equilibrium x

1 - 19

You might also like