Scribd
Upload
25 views

Lecture 11

The document discusses stability concepts in control systems, including linear and nonlinear systems. It introduces key stability concepts like Lyapunov stability and asymptotic stability. For linear systems, stability is determined by the eigenvalues of the system matrix A, as given by the characteristic equation. Nonlinear system stability depends on the stability of its local linear approximation. Historical examples like Maxwell's governor equation and the Furuta pendulum are analyzed. Numerical tools like MATLAB can evaluate stability without algebraic criteria like Routh-Hurwitz.

Uploaded by

Edutam
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
25 views

Lecture 11

The document discusses stability concepts in control systems, including linear and nonlinear systems. It introduces key stability concepts like Lyapunov stability and asymptotic stability. For linear systems, stability is determined by the eigenvalues of the system matrix A, as given by the characteristic equation. Nonlinear system stability depends on the stability of its local linear approximation. Historical examples like Maxwell's governor equation and the Furuta pendulum are analyzed. Numerical tools like MATLAB can evaluate stability without algebraic criteria like Routh-Hurwitz.

Uploaded by

Edutam
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Lecture 11 Stability

Introduction
Risk for instability is the main drawback of feedback

K. J. strm
1. Introduction

Instabilities were frequently encountered in early use of


feedback

2. Stability concepts

Created a pressing need for theory

3. Linear Systems

Understand mechanisms that create instability


Criteria for stability - beginning of control theory
Ways to avoid instabilities

4. Nonlinear Systems
5. Summary
Theme: Nonlinear systems have much richer behavior. Stability of nonlinear systems is determined by their local linear
approximation. The main reason why linera control theory
works.

Maxwells Observations 1868

Linear systems are very special


Stability problems appear in many different contexts,
buckling, critical speed, oscillations in combustion, etc.

Maxwells Qualitative Observation

It will be seen that the motion of a machine with its governor


consist in general of a uniform motion, combined with a disturbance
which may be expressed as the sum of several component motions.
These components may be of four different kind:

Stable

Unstable

0.5

1. The disturbance may continually increase.

0.5

0
0

2. It may continually diminish.

10

20

30

40

50

Stable
1

3. It may be an oscillation of continually increasing amplitude.

0
0

10

20

30

40

50

30

40

50

Unstable
1

4. It may be an oscillation of continually decreasing amplitude.


0

The first and third cases are evidently inconsistent with the stability
of the motion: and the second and fourth alone are admissible in
a good governor. Stability is mathematically equivalent to the
condition that all roots of an algebraic equation (the characteristic
equation) are in the left half plane.

1
0

1
10

20

30

40

50

10

20

How to formalize the ideas?

c K. J. strm August, 2001


&

Stability Concept

Lyapunov Stability

Consider a solution x(t, a) to the differential equation


dx
= f ( x)
dt

with initial conditions x(0, a) = a. Investigate what happens to a


solution x(t, b) with initial condition x(0, b) = b where b is close
to a.

The solution x(t, a) is called stable if h x(t, a) x(t, b)h < for all
b such that ha bh < .
The solution is called asymptotically stable if it is stable
and if in addition h x(t, a) x(t, b)h goes to zero as t increases
towards .
Notice that we can only talk about stability of a particular
solution. One solution may be stable and another unstable.
Example: the pendulum.
In control we will require asymptotic stability.
It is convenient to normalize so that the interesting solution is
x(t, a) = 0.

Stability of Nonlinear Systems

Linear Systems
Consider the solutions x(t, a) and x(t, b) to the equations

Consider the nonlinear system


dx
= f ( x)
dt

dx
= Ax
dt

Assume that it has an equilibrium x = a, i.e. f (a) = 0. The


equilibrium is stable if the linearized equation
dx
= Ax
dt

where A = f P (a) is stable

with initial conditions x(0, a) = a and x(0, b) = b. We have


x(t, a) = eAt a,

x(t, b) = eAt b

Hence x(t, a) x(t, b) = eAt (a b)


The solution x(t, a) is stable if the matrix A has all eigenvalues
in the left half plane or on the imaginary axis, and if the
eigenvalues on the imaginary axis are simple. The solution is
asymptotically stable if all eigenvalues of A are in the proper
left half plane.
c K. J. strm August, 2001
&

Linear Time Invariant Systems are Very Special


In general we can only talk about the stability of a specific
solution
This means that some solutions may be stable and other
unstable
Linear time invariant system are very special because if
one solution is stable all other solutions are also stable

The Characteristic Equation


The system

dx
= Ax has the characteristic equation
dt
det(sI A) = 0

The system

dn y
dtn

has the characteristic equation

It is thus possible to talk about the stability of a solution


This is a very unusual property for linear systems

n1

+ a1 ddtn1y + . . . + an1 dy
+ an y = 0
dt

s n + a1 s n 1 + . . . + a n 1 s + a n = 0

so does the system


Y ( s)
b1 sn1 + . . . + b1 s + bn
= n
U ( s)
s + a1 s n 1 + . . . + a n 1 s + a n

Example Inverted Pendulum

Eigenvectors and Eigenvalues

Linearize around x1 = 0.

An eigenvector v of a matrix A has the property


Av = v

where is the eigenvalue. This means that the equation


( A I )v = 0

d
dt

x1

x2



x1

x1 ( t)

x2 ( t)

x2

 


1
1
t
= c1 e
+ c2 e
1
1
t

det ( sI A) = s 1
2

Characteristic equation has roots s = 1, solution is unstable!

has a non-trivial solution, hence det( I A) = 0

Linearize around x1 = .

Now consider the differential equation

d
dt

dx
= Ax
dt

x1

x2

1 0

d
dt

x1
x2

d
dt

x1
x2


= c1

sin t
cos t


+ c2

cos t

sin t

det ( sI A) = s + 1
2

The function x(t) = e t v is a solution with the initial condition


x(t) = v

Characteristic equation has roots s = i, the solution is stable


but not asymptotically stable

c K. J. strm August, 2001


&

Using Matlab

Algebraic Criteria Routh-Hurwitz


All zeros of polynomial a0 s + a1 are in left half plane if all
coefficients are positive

Does the equation


s3 + s2 + s + k = 0

have roots in the right half plane for k = 1 or k = 10?


>> roots([1 3 2 1])
>> roots([1 3 2 10])
ans =
ans =
-2.3247
-3.3089
-0.3376 + 0.5623i
0.1545 + 1.7316i
-0.3376 - 0.5623i
0.1545 - 1.7316i
Algebraic stability conditions (Routh-Hurwitz) were been
important historically, but are now less important because of
computational tools like Matlab. The commands roots and
eigen give numerical solutions. What do we mean by solution
to a problem?

All zeros of polynomial a0 s2 + a1 s + a2 are in left half plane if all


coefficients are positive
All zeros of polynomial a0 s3 + a1 s2 + a2 s + a3 are in left half
plane if all coefficients are positive and if
a1 a2 a0 a3 > 0

Example: The polynomial s3 + 3s2 + 2s + k has all zeros in LHP


if k < 6 because
a1 a2 a0 a2 = 3  2 1  k = 6 k

The Furuta Pendulum


Model
J p( 2 sin cos ) mgQ sin = 0

tilt angle of pendulum


rate of rotation of arm
Stationary solutions for = constant.

mgQ 
C D sin = 0
J p 2 cos +
J p 2

The Furuta Pendulum


Stationary solutions = const.

mgQ 
C D sin = 0
J p 2 cos +
J p 2
p
Two solutions if < mgQ/ J p
Four solutions if >

= 0, =
p
mgQ/ J p  g/Q  7rad/s

= 0, = , = 0 , = 0
where 0 = arccos (mgQ/ J p 2). Physical interpretation!

c K. J. strm August, 2001


&

Linearization

Stability of Stationary Solutions


Solution x1 = = 0 has

Model

J p( 2 sin cos ) mgQ sin = 0


Introduce x1 = och x2 =

A=

dx1
= x2
dt
dx2
mgQ
1
=
sin x1 + 2 sin 2x1
dt
Jp
2

Linearize around x = x0
V f
A=
=

V x x= x0

mgQ
cos x1 + 2 cos 2x1
Jp

J p

!


0

A=

x= x0

The matrix A has eigenvalues on the imaginary axis if mgQ >


J p 2 the solution is then stable. If mgQ < J p 2 the matrix A
has one eigenvalue in the RHP and the solution is unstable.

Summary

Stability concepts

The matrix A has eigenvalues on the imaginary axis if

the solution is then stable. Physical interpretation!

Buckling, critical speeds, combustion instability, acoustics

J p 2 > mgQ

0
mgQ

+ 2
Jp

Stability important in control and in many other fields

Solution x1 = =

Solutions = 0 = arccos (mgQ/ J p 2),


!
A=

The matrix A has one eigenvalue in the RHP, unstable

Stability of Stationary Solutions ...

0
 mgQ 2

0
mgQ
+ 2
Jp

Stability of solutions and stability of systems


Linear systems: Characteristic equation det (sI A) tells
all
Nonlinear systems
Only stability of particular solutions
A solution is stable if linearized equation stable
The main reason why linear control theory is so useful

c K. J. strm August, 2001


&

You might also like