Nonlinear Systems
Dr. P. Kavitha
Associate Professor
SELECT, VIT University
kavitha.p@vit.ac.in
Lecture Outline
• Introduction
2
Introduction
3
Introduction
• One of the most important characteristics of nonlinear system is
the dependence of the system response on the magnitude and type
of the input.
F=kx
0.3 3
F=kx-119x
0.25
0.2
Forces (N)
0.15
0.1
0.05
0
0 2 4 6 8 10
x (cm)
4
Introduction
• For example, a nonlinear system may behave completely differently in
response to step inputs of different magnitudes.
5
Frequency Amplitude Dependence
• Consider the free oscillation of the mechanical systems shown in
figure.
• The differential equation of the system is Nonlinear
Spring
𝑚𝑥 + 𝑏𝑥 + 𝑘𝑥 + 𝑘 ′ 𝑥 3 = 0
M
𝑤ℎ𝑒𝑟𝑒, 𝑘𝑥 + 𝑘′𝑥3 is nonlinear spring force x
𝑥 is displacement
𝑚 is mass B
𝑏 is coefficient of damping
7
Frequency Amplitude Dependence
• Frequency amplitude dependence is one of the most fundamental
characteristics of the nonlinear systems.
Amplitude
Frequency 8
Multivalued Responses and Jump Resonances
• The differential equation of spring mass damper system with
sinusoidal forcing function exhibits multivalued response and
jump resonance.
𝑚𝑥 + 𝑏𝑥 + 𝑘𝑥 + 𝑘 ′ 𝑥 3 = 𝑃 cos(𝜔𝑡)
9
Sub-Harmonic & Super-Harmonic Oscillations
• The differential equation of spring mass damper system also
exhibits periodic motions such as sub-harmonic and super-
harmonic oscillations.
𝑚𝑥 + 𝑏𝑥 + 𝑘𝑥 + 𝑘 ′ 𝑥 3 = 𝑃 cos(𝜔𝑡)
10
Self-excited Oscillations or limit cycles
• Another phenomenon that is observed in certain nonlinear systems is a self-
excited oscillation or limit cycle.
𝑚𝑥 − 𝑏(1 − 𝑥 2 )𝑥 + 𝑘𝑥 = 0
• It is nonlinear in the damping term, that is, for small values of 𝑥 the damping
will be negative and will actually put energy into the system, while for large
values of 𝑥 it is positive and removes energy from the system.
• Thus it can be expected that such a system may exhibit a sustained oscillation.
• Since it is not a forced system, this oscillation is called self excited oscillation
or a limit cycle.
11
Frequency Entrainment
• If a periodic force of frequency 𝜔 is applied to a system capable of
exhibiting a limit cycle of frequency 𝜔𝑜 , the well known phenomenon of
beat is observed.
Suppose two tuning forks having frequencies 256 and 257 per
second respectively, are sounded together. If at the beginning
of a given second, they vibrate in the same phase so that the 𝜔
compressions (or rarefactions) of the corresponding waves
reach the ear together, the sound will be reinforced
(strengthened). Half a second later, when one makes 128 and
the other 128/2 vibrations, they are in opposite phase, i.e., the
compression of one wave combines with the rarefaction of the
other and tends to produce silence. At the end of one second, 𝜔𝑜
they are again be in the same phase and the sound is
reinforced. By this time, one fork is ahead of the other by one
vibration.
Thus, in the resultant sound, the observer hears maximum
sound at the interval of one second. Similarly, a minimum
loudness is heard at an interval of one second. As we may
consider a single beat to occupy the interval between two
consecutive maxima or minima, the beat produced in one
second in this case, is one in each second. If the two tuning
forks had frequencies 256 and 258, a similar analysis would 𝜔 − 𝜔𝑜
show that the number of beats will be two per second. Thus,
in general, the number of beats heard per second will be equal
to the difference in the frequencies of the two sound waves. 12
Frequency Entrainment
• In a self excited nonlinear system the frequency 𝜔𝑜 of the limit cycles
falls in synchronistically with, or entrained by, the forcing frequency 𝜔,
within a certain band of frequencies.
∆𝜔
𝜔
𝑜
13
Asynchronous Quenching
14
COMMON NONLINEARITIES IN CONTROL SYSTEMS
• Consider the typical block diagram shown in Figure. It is composed of four
parts: a plant to be controlled, sensors for measurement, actuators for control
action, and a control law, usually implemented on a computer.
r(t) + y(t)
Controller Actuators Plant
-
Sensors
• Nonlinearities may occur in any part of the system, thus make it a nonlinear
control system.
• Many different types of nonlinearities may be found in practical control
systems, and they may be divided into
Inherent Nonlinearities: The non-linear behavior that is already present
in the system. eg. Saturation
Intentional Nonlinearities: The non-linear elements that are added into a
system. eg. Relay
15
Inherent Nonlinearities
16
Intentional Nonlinearities
17
Approaches to the analysis and Design of
nonlinear Control Systems
• There is no general method for dealing with all nonlinear systems
18
Describing Function Method
19
Describing Function (DF)
• The describing function (DF) method is an approximate procedure for
analyzing certain nonlinear control problems.
• Transfer function of a true LTI system cannot depend on the amplitude of the
input function.
• Suppose that the output is periodic with the same period as the
input.
21
Describing Function Method
22
Describing Function Method
• The describing function or sinusoidal describing function
of a nonlinear systems is defined as the “complex ratio of
the fundamental harmonic component of the output to the
input”
• That is
𝑌1
𝑁 = ∠∅1
𝑋
• Where,
𝑁 = describing function
𝑋 = amplitude of input sinusoid
𝑌1 = amplitude of the fundamental harmonic components of output
∅1 = phase shift of the fundamental harmonic component of output.
23
Describing Function Method
𝑌1
𝑁 = ∠∅1
𝑋
24
Describing Function Method
• In calculating the describing function for a given nonlinear
system, we need to find the fundamental harmonic component of
the output.
𝑥 𝑡 = 𝑋 sin(𝜔𝑡)
25
Describing Function Method
∞
2𝜋 2𝜋
1 1
𝐴𝑛 = 𝑦 𝑡 cos 𝑛𝜔𝑡 𝑑(𝜔𝑡) 𝐵𝑛 = 𝑦 𝑡 sin 𝑛𝜔𝑡 𝑑(𝜔𝑡)
𝜋 𝜋
0 0
𝑦 𝑡 = 𝐴𝑜 + 𝑌𝑛 sin 𝑛𝜔𝑡 + ∅𝑛
𝑛=1
𝐵𝑛
𝑌𝑛 = 𝐴2𝑛 + 𝐵𝑛2 ∅𝑛 = tan−1
𝐴𝑛
26
Describing Function Method
• If the nonlinearity is skew symmetric, then 𝐴𝑜 = 0 . The
fundamental harmonic component of the output is
𝑦1 (𝑡) = 𝑌1 sin 𝜔𝑡 + ∅1
𝑌1 𝐴12 + 𝐵12 −1
𝐵1
𝑁 = ∠∅1 = ∠ tan
𝑋 𝑋 𝐴1
27
Relay Nonlinearity (On-Off Nonlinearity)
• The relay-type (on-off) nonlinearity is shown in following figure.
y(t)
on
M t0 M
y (t )
M t0
0 t
-M
off
𝑦 𝑡 = 𝐵𝑛 sin 𝑛𝜔𝑡
28
𝑛=1
Relay Nonlinearity (On-Off Nonlinearity)
𝑦1 𝑡 = 𝐵1 sin 𝜔𝑡
• Where,
2
y (t ) sin t d t
1
B1
0
M sin t d t
2
B1
0
sin t d t
2M
B1
0
4M
B1
• Therefore,
4𝑀
𝑦1 𝑡 = sin 𝜔𝑡
𝜋 29
Relay Nonlinearity (On-Off Nonlinearity)
• Where,
−1
𝐵1
𝑌1 = 𝐴12 + 𝐵12 ∅1 = tan
𝐴1
Y1 4M
N 0
X X
30
Saturation
When one increases the input to a physical device, the following phenomenon is often
observed: when the input is small, its increase leads to a corresponding (often
proportional) increase of output: but when the input reaches a certain level, its further
increase does produce little or no increase of the output. The output simply stays around
its maximum value. The device is said to be saturation when this happen. A typical
saturation nonlinearity is represented in following figure, where the thick line is the real
nonlinearity and the thin line is an idealized saturation nonlinearity.
Most actuators display saturation characteristics. For example, the output torque of a
servo motor cannot increase infinitely and tends to saturate, due to the properties of
magnetic material.
31
The input-output relationship for a saturation nonlinearity is plotted in following
figure, with a and k denoting the range and slope of the linearity. Since this
nonlinearity is single-valued, we expect the describing function to be a real
function of the input amplitude.
32
Consider the input x(t)=Asin(ωt). If A≤ a, then the input remains in the linear range, and
therefore, the output is w(t)=kAsin(ωt). Hence, the describing function is simply a constant
k.
Consider the case A>a. The output is seen to be symmetric over the four quarters of a
period. In the first quarter, it can be expressed as
-δ
δ x
dead-zone
Figure 6. A dead-zone nonlinearity.
Dead-zones can have a number of possible effects on control systems. Their most
common effect is to decrease static output accuracy. They may also lead to limit cycles
or system instability because of the lack of response in the dead-zone. The response
corresponding to a sinusoidal input x(t)=Asin(ωt) into a dead-zone of width 2δ and
slope k, with A≥δ, is plotted in Figure 7. Since the characteristics is an odd function,
a1=0. The response is also seen to be symmetric over the four quarters of a period. In
one quarter of a period, i.e., when 0≤ωt≤p/2, one has
34
x(t)
35
0 0 t
w (t)
k A sin t t / 2
where sin 1 / A
2k 2
NA sin 1 1 2
2 A A A
Backlash
Backlash often occurs in transmission systems. It is caused by the small gaps which
exist in transmission mechanism. In gear trains, there always exists small gaps
between a pair of mating gears as shown in Figure 9.
output angle
C B
slope 1
O A
-b b input angle
D
E
w ( t ) A b k t
2
3
w ( t ) A sin t b k t
2
3
w ( t ) A b k t 2
2
5
w ( t ) A sin t b k 2 t
2
where sin 1 1 2b / A
38
Figure 10. Backlash nonlinearity.
Unlike the other nonlinearities, the function w(t) here is neither odd nor even. Therefore, a1
and b1 are both nonzero.
4kb b
a1 1
A
2
sin 1
Ak 2 b 2 b 2 b
b1 1 1 1 1
2 A A A
Therefore, the describing function of the backlash is given by
1
N ( A) a12 b12
A
b1
angle( N ( A)) tan 1
a1
The amplitude of the describing function for backlash is plotted in Figure 11.
The characteristics of the nonlinear systems
(distinguishing features with linear system)
3 Have two kind of work states: Have stable, unstable and self-
stable and unstable. oscillation.
4 The form of the output is the The form of the output is different
same as input. from the input.
40
NONLINEAR SYSTEM ANALYSIS
41
DESCRIBING FUNCTION
42
Describing function method
43
Describing function
r ( t ) 0 x (t )
KN
y(t )
G(s)
c (t )
Typical structure of the nonlinear
systems
H(s)
nonlinear linear
44
Describing function
2
y(t )d (t )
1
A0
2
2 0
For x( t ) X sint ,
y( t ) A1 cos t B1 sint Y1e j1
N(X )
Y1 sin(t 1 ) X
46
2
Y1 A B
2 2 A1
1 1
0
B1
1 tan 1 2
y( t ) sintd (t )
1
A1 B1
0
47
How to get the describing function?
(2) Solve y(t) and obtain the fundamental component of
y(t).
(3) Calculate the describing function N(X) according to
following formula:
2
1
A1 y( t ) cos td (t ) Y A2 B 2
1
1 1
0
A y(t ) Y1 sin(t 1 )
2 1 1
1 tan
1
B1 y( t ) sintd (t ) B1
0
Y1e j1
N(X )
X
48
Determine the describing function of the saturation
non linearity
y(t ) y
b
a k t
x(t ) 2
0 a 0 1
b
Solution
0 x(t )
a
1 1 sin 1
X
x(t ) X sint
t 49
How to get the describing function?
y
kXsint 0 t 1
2 t y( t )
0 1 kX 1 t 2
2
1
A1 0 B1 y(t ) sin td (t )
0
2kX 1 a a a 2
sin 1 ( )
X X X
B1 2k 1 a a a 2
KN (X ) sin 1 ( ) X a
X X X X
50
Determine the describing function of the dead zone
non linearity
x(t ) y
1
k
a t t
0 a 0 1
k
solution
x(t )
0
1 1 sin 1
X
x(t ) X sint
t 51
y
0 0 t 1 1
y(t) t
k ( Xsint - ) 1 t 2 0 1
A1 0
2
B1 2k 1 2
K N ( X ) sin 1 ( ) ( X )
X 2 X X X
52
The describing function of some typical nonlinearity
M
4M M
KN (X ) KN (X )
4M h
1 ( )2
-M X h
-M X X
Relay
k 2k 1 a a a 2
KN (X ) sin 1 ( )
a X X X
saturation
k 2k 1 2
N(X ) sin 1 ( )
dead zone
2 X X X
53
The describing function of some typical nonlinearity
M 4M h 2 4M
KN (X ) 1 ( ) j
h
X X X 2
1 2h
sin (1 )
k 2 X
k
KN (X )
h
2h h h
2(1 X ) X (1 X )
backlash
hysteresis 4kh h
j ( 1)
X X
54
How to get the describing function?
characters
For the “single value” nonlinearity the describing function must be a “real number”.
such as the dead zone, saturation and the ideal relay nonlinearity etc.
For example:
k2
y(t ) N1 ( X ) N2( X ) N3( X )
k1
a x(t )
0 a = + +
N ( X ) N1 ( X ) N 2 ( X ) N 3 ( X )
55
Stability analysis of the nonlinear system by describing function
γc
1 G ( j ) 0 stable
unstable Critical
G ( j ) 1 j 0
Stability
56
Stability analysis of the nonlinear system by
describing function
Compare the nonlinear system with the linear system
Nonlinear Linear
r (t ) c(t ) r (t ) x(t ) y(t ) c(t )
G(s) KN(x)
G(s)
1
: Negative inverse describing function.
KN (X )
Graphical explanation is shown as following:
Im X increasing Im G ( j ) Im
1 direction
KN (X )
Re Re Re
G ( j )
1
G ( j )
1
KN (X ) KN (X )
G ( j ) G ( j )
do not circle -1/KN(X) G ( j ) circle -1/KN(X) Intersect with -
(stable) (unstable) 1/KN(X)
(self-oscillation)
58
Stability analysis of the nonlinear system by describing function
Self-oscillation of the nonlinear system
Im
A special motion of the nonlinear system:
A : stable point
System will be at a continuous Re
oscillation, which has a constant
stable unstable
amplitude and frequency, when
zone zone
the system come under a light
1
disturbance. B:
KN (X )
unstable point G( j )
Self-oscillation
Using the describing function to analyze the nonlinear system, the Linear parts of
the system must be provided with a good characteristic of the low-pass filter→so
that the harmonics produced by the nonlinear element can be neglected.
Generally the describing function method can only be used for analyzing the
stability and self-oscillation of the nonlinear systems, not the steady-state error and
transient specifications.
60
Phase Plane Analysis
61
Introduction
• Introduced in the end of 19th century by Henry Poincare
62
Introduction …
• Phase Plane Analysis gives intuitive insights to nonlinear
systems; without having to solve the nonlinear equations
analytically
64
State Space Trajectories
• The unforced response of a system released from any initial
point x(to) traces a curve or trajectory in state space, with time t
as an implicit function along the trajectory.
66
Phase Plane Analysis
1. Analytical method
67
Phase Plane Analysis of Linear Systems
68
- Consider the second-order linear system
x ax bx 0
is the general solution
1t 2t
x(t ) k1e k2e
• Where the eigenvalues 1 and 2 are the solutions of the
characteristic equation
s 2 as b ( s 1 )( s 2 ) 0
(a a 2 4b ) (a a 2 4b )
1 , 2
2 2
69
There is only one singular point (assuming b 0 ), namely the
origin.
1. λ1 and λ2 are both real and have the same sign (positive or
negative)
2. λ1 and λ2 are both real and have opposite signs (saddle point)
𝑟= 𝑥12 + 𝑥22
• And the angle θ is
𝑥2
0≤𝜃= tan−1 < 2𝜋
𝑥1
Equilibrium States
𝑟 = 𝑟(1 − 𝑟)
𝑟= 𝑥12 + 𝑥22
𝜃 −1
𝑥2
𝜃= sin2( ) 0 ≤ 𝜃 = tan < 2𝜋
2 𝑥1
Where,
L: the pendulum’s length
M: its mass
b: the friction coefficient at the hinge
g: the gravity constant.
Example: Simple Pendulum
• Letting
𝑥1 = 𝜃 and 𝑥2 = 𝜃
x1 x2
b g
x2 x sin x1
2 2
MR R
b g
x1 x2 x2 x sin x1
2 2
MR R
x2 = 0, sin x1 = 0
(0, 0) and (π, 0)
Answer
(1, 1) and (−1, −1).
82
Example
𝑦 = 𝑦 + 2𝑥𝑦
Answer
(0, 0) and (−1/2, 1).
83
Local Behavior of Nonlinear Systems
84
Local Behavior of Nonlinear Systems
𝑥1 = 𝑎𝑥1 + 𝑏𝑥2
𝑥2 = 𝑐𝑥1 + 𝑑𝑥2
85
Limit Cycle
86
Limit Cycle
• Depending on the motion patterns of the trajectories in the vicinity of the
limit cycle, we can distinguish three kinds of limit cycles.
– Stable Limit Cycles: all trajectories in the vicinity of the limit cycle
converge to it as t →∞
– Unstable Limit Cycles: all trajectories in the vicinity of the limit cycle
diverge to it as t →∞
87
Limit Cycle
88
Example
𝑥 = 𝑦 − 𝑥(𝑥 2 + 𝑦 2 − 1)
𝑦 = −𝑥 − 𝑦(𝑥 2 + 𝑦 2 − 1)
89
Example
• Find the equilibrium point(s) of each nonlinear system given
below. Then determine the type and stability of each equilibrium
point.
90
Exercise
• Find the equilibrium point(s) of each nonlinear system given
below. Then determine the type and stability of each equilibrium
point. If system exhibits limit cycles then determine the nature
of limit cycle as well.
1. 𝑥 = 𝑥𝑦 − 3𝑦 4. 𝑥 = 𝑥 2 𝑦 + 3𝑥𝑦 − 10𝑦
𝑦 = 𝑥𝑦 − 3𝑥 𝑦 = 𝑥𝑦 − 4𝑥
2. 𝑥 = 𝑥 2 − 3𝑥𝑦 + 2𝑥 5. 𝑥 = 2 − 𝑥2 − 𝑦2
𝑦 =𝑥+𝑦−1 𝑦 = 𝑥2 − 𝑦2
3. 𝑥 = 𝑥 2 + 𝑦 2 − 13
𝑦 = 𝑥𝑦 − 2𝑥 − 2𝑦 + 4
91
Solution
1. 𝑥 = 𝑥𝑦 − 3𝑦
𝑦 = 𝑥𝑦 − 3𝑥
92
Solution
2. 𝑥 = 𝑥 2 − 3𝑥𝑦 + 2𝑥
𝑦 =𝑥+𝑦−1
93
Solution
3. 𝑥 = 𝑥 2 + 𝑦 2 − 13
𝑦 = 𝑥𝑦 − 2𝑥 − 2𝑦 + 4
• Equilibrium points are (2, 3), (2, −3), (3, 2) and (−3, 2).
94
Solution
4. 𝑥 = 𝑥 2 𝑦 + 3𝑥𝑦 − 10𝑦
𝑦 = 𝑥𝑦 − 4𝑥
• Equilibrium points are (1, 1), (1, −1), (−1, 1), and (−1, −1).
95
Solution
5. 𝑥 = 2 − 𝑥2 − 𝑦2
𝑦 = 𝑥2 − 𝑦2
• Equilibrium points are (0, 0), (2, 4), and (−5, 4).
96
Phase Plane Analysis Method 1
Analytical Method
97
1. Analytic method
Example 1
Spring - mass motion system : mx Kx 0 k=1
2
Because x(0) x 0 , x(0) 0 x x x 0
2 2 2
s2
-
Plot the phase loci of the system:
Solution:
1
Because : Y ( s) 2 C ( s) Y ( s) s 2C ( s) y c
s
d 2c M r c 0
So we have : y M Msign(r c)
2 r c 0
dt
99
make : c x1 , c x2
then : x1 x2 dx2 Msign(r x1 )
x 2 c
Msign(r x1 ) dx1 x2
x2 x1
that is x dx
x2 ( 0 )
2 2 Msign(r x )dx
x1 ( 0 )
1 1
x2
when x1 r :
x1
we have x 2Mx1 2Mx1 (0) x (0)
2
2
2
2
r
when x1 r :
we have x22 2Mx1 2Mx1 (0) x22 (0)
stable
focus
unstable
focus
stable
nodes
unstable
nodes
102
centers
saddle
points
103
Analysis of the phase plane
2. limit cycle
A kind of phase locus with the closed loop form
→Corresponding to the self-oscillation.
Types of the limit cycle :
(1) stable limit cycle
(2) unstable limit cycle
(3) semi - stable limit cycle
104
Phase Plane Analysis Method 2
Isocline Method
105
Graphic method--- isoclinal method
dx
For the systems : x f ( x, x ) 0 x f ( x, x ) 0
dx
dx2
make : x1 x, x2 x then : x2 f ( x1 , x2 )
dx1
dx2 f ( x1 , x2 )
dx1 x2
dx2 f ( x1 , x2 )
make : isocline equation
dx1 x2
: the slope of the phase loci
Example 3
Spring - mass motion system : x x 0
plot the phase loci by the isoclinal method.
106
Solution
dx x
In terms of : x x 0 we have :
dx x
dx 1
make then : x x
dx
The isocline is the beelines passing the coordinate origin and
1
with slope .
1
The values are shown in following table for different
2 1 0.5 0 -0.5 -1 -2
1
0 -0.5 -1 -2 2 1 0.5 0
x
(x 0 , x 0 )
108
Attentions:
1) x - axis and x should have the same scale.
2) The direction of the phase loci always are clockwise :
For x 0 : from left to right with x increasing ;
For x 0 : from right to left with x decreasing .
3) The slope of the phase loci through x - axis is ,
so the phase loci intersect x - axis uprightly.
4) apply the symmetry of the phase locus to reduce work .
For the symmetry about x axis : f(x, x ) f( x, x )
For the symmetry about x axis : f(x, x ) f(x, x )
For the symmetry about origin : f(x, x ) f( x, x )
109
Analysis of the phase plane
How to analyze the performance of the nonlinear systems from
the phase locus
(1) We can analyze the stability directly from the phase locus:
the phase locus is convergent or divergent.
r e k1 y K c k1 0.0625 ; k2 1;
e0 s(Ts 1)
T 1; K 4; e0 0.2
• Also the phase plane method is used to analyze the stability of some
intelligent control systems, such as the Fuzzy control systems.
112
Phase Plane Analysis Method 3
Delta Method
113
End of Slides
114