Course roadmap

ME451: Control Systems Modeling Analysis Design

Laplace transform Time response

Design specs
• Transient
Lecture 9 Transfer function • Steady state
Root locus
Stability
Models for systems Frequency response
Frequency domain
• electrical • Bode plot
• mechanical
PID & Lead-
Lead-lag
• electromechanical Stability
Dr. Jongeun Choi Block diagrams • Routh-
Routh-Hurwitz
Design examples
Department of Mechanical Engineering Linearization • (Nyquist)
Nyquist)
Michigan State University

(Matlab simulations &) laboratories

1 2

Simple mechanical examples Stability

ƒ We want mass to stay at x=0, but wind gave some ƒ Utmost important specification in control design!
initial speed (F(t)=0). What will happen? ƒ Unstable systems have to be stabilized by
f(t)
f(t) K f(t)
f(t) feedback.
M M ƒ Unstable closed-loop systems are useless.
ƒ What happens if a system is unstable?
x(t)
x(t) x(t)
x(t)
K • may hit mechanical/electrical “stops”
stops” (saturation)
B
f(t)
f(t) f(t)
f(t) • may break down or burn out
M M
B
x(t)
x(t) x(t)
x(t)

ƒ How to characterize different behaviors with TF?

3 4
What happens if a system is unstable? Mathematical definitions of stability
Tacoma Narrows Bridge (July 1-Nov.7, 1940) ƒ BIBO (Bounded-Input-Bounded-Output) stability :
Any bounded input generates a bounded output.
ICs=0 y(t)
y(t)
u(t)
u(t)
BIBO stable
system

ƒ Asymptotic stability :
Wind-
Wind-induced vibration Collapsed!
Any ICs generates y(t)
y(t) converging to zero.
2008… ICs y(t)
y(t)
Asymp.
Asymp. stable
u(t)=0
u(t)=0 system

5 6

Stability condition in s-domain

Some terminologies (Proof omitted, and not required)
Ex. For a system represented by a transfer
function G(s),
ƒ Zero : roots of n(s) system is BIBO stable

ƒ Pole : roots of d(s) All the poles of G(s) are in the open left
half of the complex plane.
ƒ Characteristic polynomial : d(s)

system is asymptotically stable

ƒ Characteristic equation : d(s)=0
7 8
 Second order impulse response-
“Idea” of stability condition Underdamped and Undamped
Example

… Overdamped
… Critically damped
Asym.
Asym. Stability:
… Underdamped
(U(s)=0)
U(s)=0)
… Undamped
BIBO Stability:
(y(0)=0)

Re(α)>
Bounded if Re(α)>0
0
9 10

Second order impulse response – Second order impulse response –

Underdamped and Undamped Underdamped and Undamped
Changing / Fixed Changing / Fixed

Impulse Response Impulse Response

5 5

4 4

3 3

2 2
Amplitude

Amplitude
1 1

0 0

-1 -1

-2 -2

-3 -3

-4 -4
0 2 4 6 8 10 12 0 2 4 6 8 10 12
6 6
Time (sec) Time (sec)
4 4
2 2

0 0
-2 -2
-4 -4

-6 11 -6 12
-5 0 5 -5 0 5
 Second order impulse response –
Underdamped and Undamped
Remarks on stability
Changing / Fixed
ƒ For a general system (nonlinear etc.), BIBO
5
Impulse Response stability condition and asymptotic stability
4 condition are different.
ƒ For linear time-invariant (LTI) systems (to which
3

we can use Laplace transform and we can

Amplitude

-1
obtain a transfer function), the conditions
-2 happen to be the same.
-3

-4
0 2 4 6 8 10 12
ƒ In this course, we are interested in only LTI
6
Time (sec)
4 systems, we use simply “stable” to mean both
2

0
BIBO and asymptotic stability.
-2
-4

-6 13 14
-5 0 5

Remarks on stability (cont’d) Examples

ƒ Marginally stable if ƒ Repeated poles
ƒ G(s)
G(s) has no pole in the open RHP (Right Half Plane), &
ƒ G(s)
G(s) has at least one simple pole on -axis, &
ƒ G(s)
G(s) has no multiple poles on -axis.

ƒ Does marginal stability imply BIBO stability?

Marginally stable NOT marginally stable

ƒ TF:

ƒ Unstable if a system is neither stable nor

ƒ Pick
marginally stable.
ƒ Output

15 16
 Feedback Technique Positive Feedback

K will depends on the distance between the guitar and the amplifier.

17 18

Stability summary Mechanical examples: revisited

Let si be poles of G. K
f(t)
f(t) f(t)
f(t)
Then, G is … M M

ƒ (BIBO, asymptotically) stable if x(t)

x(t) x(t)
x(t)
Poles= Poles=
Re(si)<0 for all i. stable? stable?
ƒ marginally stable if
K
ƒ Re(si)<=0 for all i, and B
f(t)
f(t) f(t)
f(t)
ƒ simple root for Re(si)=0 M M
ƒ unstable if
B
it is neither stable nor x(t)
x(t) Poles= x(t)
x(t) Poles=
marginally stable. stable? stable?

19 20
Examples Summary and Exercises
Stable/marginally stable
/unstable ƒ Stability for LTI systems
ƒ (BIBO and asymptotically) stable, marginally stable,
? unstable
ƒ Stability for G(s)
G(s) is determined by poles of G.
?
ƒ Next
?
ƒ Routh-
Routh-Hurwitz stability criterion to determine stability
without explicitly computing the poles of a system.
?
ƒ Exercises
? ƒ Solve examples in the previous slide.

???

21 22