MESB374

System Modeling and Analysis

System Stability and Steady State
Response
Stability Concept
Describes the ability of a system to stay at its equilibrium position in the
absence of any inputs.
Stability
Ex: Pendulum
where the derivatives of all states are zeros
inverted
pendulum
simple
pendulum
hill
plateau
valley
A linear time invariant (LTI) system is stable if and only if (iff) its
free response converges to zero for all ICs.
Ball on curved surface
Examples (stable and unstable 1
st
order
systems)
Q: free response of a 1
st
order system.
( )
0
5 (0) y y u t y y + = =
1
5
( )
0
t
y t y e

=
Q: free response of a 1
st
order system.
1
5 1
G
s
=
+
TF:
Pole:
0.2 p =
t
y
t
y
( )
0
5 (0) y y u t y y + = =
1
5
( )
0
t
y t y e =
1
5 1
G
s
=
+
TF:
Pole:
0.2 p =
Stability Criterion for LTI Systems
Stability of LTI Systems
( ) ( 1) ( ) ( 1)
1 1 0 1 1 0
1
1 1 0
Characteristic Polynomial
Stable All poles lie in the left-half complex plane (LHP)
All roots of ( ) 0 l
n n m m
n m m
n n
n
y a y a y a y b u b u b u b u
D s s a s a s a

+ + + + = + + + +
:
: = + + + + = ie in the LHP
Complex (s-plane)
Re
Im
Marginally
stable/``unstable
Relative
Stability
(gain/phase margin)
Absolutely
Stable
Unstable
Comments on LTI Stability
Stability of an LTI system does not depend on the input (why?)

For 1st and 2nd order systems, stability is guaranteed if all the
coefficients of the characteristic polynomial are positive (of same
sign).


Effect of Poles and Zeros on Stability
Stability of a system depends on its poles only.
Zeros do not affect system stability.
Zeros affect the specific dynamic response of the system.
Stability of LTI Systems
0 0
2
1 0 1 2
( ) : Stable 0
( ) : Stable 0 and 0
D s s a a
D s s a s a a a
= + >
= + + > >
Passive systems are usually stable
Any initial energy in the system is usually dissipated in real-world
systems (poles in LHP);
If there is no dissipation mechanisms, then there will be poles on
the imaginary axis
If any coefficients of the denominator polynomial of the TF are
zero, there will be poles with zero RP
System Stability (some empirical guidelines)
Active systems can be unstable
Any initial energy in the system can be amplified by internal source
of energy (feedback)
If all the coefficients of the denominator polynomial are NOT the
same sign, system is unstable
Even if all the coefficients of the denominator polynomial are the
same sign, instability can occur (Rouths stability criterion for
continuous-time system)
In Class Exercises
(1) Obtain TF of the following system:

(2) Plot the poles and zeros of the system
on the complex plane.
(3) Determine the systems stability.

| |
L
2 5 y y y u u + + =
y y y u u u 6 3 4
( ) ( ) ( ) ( ) ( )
2
2 5 s Y s sY s Y s sU s U s + + =
( )
( )
( )
2
1
2 5
Y s s
G s
U s s s

= =
+ +
Poles:

Zero:

2
2 5 0 s s + + =
1,2
2 4 20
1 2
2
p j

= =
1 0 s =
1 z =
Real
Img.

1
1 2 p j = +
2
1 2 p j =
1 z =
Stable

(1) Obtain TF of the following system:

(2) Plot the poles and zeros of the system
on the complex plane.
(3) Determine the systems stability.

TF:

( )
( )
( )
2
3 2
3 4
6
Y s s s
G s
U s s s s
+
= =
+ +
Poles:

Zeros:

3 2
6 0 s s s + + = 1 2,3
1 23
0,
2
j
p p

= =
2
3 4 0 s s + =
1,2
3 7
2
j
z

=
Real
Img.

3
1 23
2 2
p j = Marginally
Stable
2
1 23
2 2
p j = +

1
0 p =
1
3 7
2 2
z j = +
2
3 7
2 2
z j =
Example
Inverted Pendulum
(1) Derive a mathematical model for
a pendulum.





(2) Find the equilibrium positions.
(3) Discuss the stability of the
equilibrium positions.
u
Bu
mg
l
EOM:

sin
o
I mgl B u u u =
is very small

Equilibrium
position:

0
0
0
u
u
u

=
=
`
=
)
Assumption:

u
Linearized EOM:

0
o
o
K
I mgl B
I B mgl
u u u
u u u
=
+ =
Characteristic
equation:

2
0
o
s I sB K + =
Poles:

2
1,2
4
2
o
o
B B KI
p
I
+
=
Real
Img.

Unstable

Example (Simple Pendulum)
u
Bu
mg
EOM:

sin
o
I mgl B u u u =
is very small

Equilibrium
position:

0
0
0
u
u
u

=
=
`
=
)
Assumption:

u
Linearized EOM:

0
o
o
K
I mgl B
I B mgl
u u u
u u u
=
+ + =
Characteristic
equation:

2
0
o
s I sB K + + =
Poles:

2
1,2
4
2
o
o
B B KI
p
I

=
Real
Img.

stable

2
4 0
o
B KI >
Real
Img.

stable

2
4 0
o
B KI =
Real
Img.

stable

2
4 0
o
B KI <
How do the positions of poles
change when K increases?
(root locus)

Transient and Steady State Response
Ex:
( )
5 u t t =
5 10 y y u + =
to a ramp input:
Lets find the total response of a stable first order system:
with I.C.: ( )
0 2 y =
- total response
( )
( )
( )
2
0
Transfer Function
Free Response
Forced Response
10 5 1
2
5 5
y
U s
Y s
s s s
= +
+ +
- PFE
( )
3 1 2
2
2
5 5
a a a
Y s
s s s s
= + + +
+ +
2
2 forced
0
( ) 10
s
a s Y s
=
( = =

( )
{ }
( )
0
2 1
2
1 forced 2 2 1
0
0
1 50 50
( ) 2
2 1 ! 5
5
s
s
s
d d
a s Y s
ds ds s
s
=

=
=

= = = =
`
+
) +
( )
3 forced
5
5 ( ) 2
s
a s Y s
=
= + = (

( )
5 5
3
Transient response
Steady state response
Transient response
free response
from Forced response
from Forced response
2 10 2
t t
y t t a e e

| |
|
= + + +
|
|
\ .
Transient and Steady State Response
to a input u(t) can be decomposed into two parts
In general, the total response of a STABLE LTI system
( ) ( ) ( )
Transient Response
Steady State Response
T SS
y t y t y t = +
( ) ( 1) ( ) ( 1)
1 1 0 1 1 0
n n m m
n m m
y a y a y a y b u b u b u b u


+ + + + = + + + +
where
Transient Response
contains the free response of the system plus a portion of forced response
will decay to zero at a rate that is determined by the characteristic roots (poles)
of the system
Steady State Response
will take the same (similar) form as the forcing input
Specifically, for a sinusoidal input, the steady response will be a sinusoidal
signal with the same frequency as the input but with different magnitude and
phase.
Transient and Steady State Response
Ex:
( ) ( )
5sin 3 u t t =
4 3 6 y y y u + + =
to a sinusoidal input:
Lets find the total response of a stable second order system:
with I.C.: ( ) ( )
0 0, 0 2 y y = =
- total response
( )
2 2 2 2
Forced Response Free Response
6 5 3 2 4 2
4 3 3 4 3
s
Y s
s s s s s
+
= +
+ + + + +
- PFE
( )
3 1 2 4 1 2
3 1 3 3 3 1
a a a a b b
Y s
s s s j s j s s
(
(
= + + + + +
(
(
+ + + + +


2
9
2
a =
1
5
2
a =
3
1
1
2
a j =
( ) { } ( ) ( )
( ) ( )
3 3
3 1 1 2 2
Steady state response Transient response
1 3
2Re
7 15
5sin 3 tan 2
2 2
jt t t
t t
y t a e a b e a b e
t e e


( = + + + +

(
= + + +
(

4
1
1
2
a j = +
2
3 b =
1
1 b =
Steady State Response
f f t sF s
t s
( ) lim ( ) lim ( ) = =
0
4 12 4 3 y y y u u + + = +
( )
2
4 3 5
4 12
s
Y s
s s s
+
=
+ +
Final Value Theorem (FVT)
Given a signals LT F(s), if all of the poles of sF(s) lie in the LHP, then f(t)
converges to a constant value as given in the following form
Ex.
(1). If a constant input u=5 is applied to the sysetm at time t=0, determine
whether the output y(t) will converge to a constant value?
(2). If the output converges, what will be its steady state value?
We did not consider the effects of IC since
it is a stable system
we are only interested in steady state response
A linear system is described by the following equation:
0
5
( ) lim ( ) lim ( )
4
t s
y y t sY s

= = =
Steady State Response
Given a general n-th order stable system
( ) ( 1) ( ) ( 1)
1 1 0 1 1 0
n n m m
n m m
y a y a y a y b u b u b u b u


+ + + + = + + + +
( )
1
1 1 0
1
1 1 0
m m
m m
n n
n
b s b s b s b
G s
s a s a s a

+ + + +
=
+ + + +
1
1 1 0
( )
( )
Free
n n
n
F s
Y s
s a s a s a

=
+ + + +
Free Response
Transfer Function
Steady State Value of Free Response (FVT)
1
0 0
1 1 0
0
0
( )
lim ( ) lim
0 (0)
lim 0
SS Free n n
s s
n
s
sF s
y sY s
s a s a s a
F
a

= =
+ + + +

= =
In SS value of a stable LTI system, there is NO contribution from ICs.