Professional Documents
Culture Documents
SYSTEMS
Prof. Dr. Emin TACER
Phone: 0212 0212
4116100/22208
E-Mail:
emintacer@aydin.itu.tr
tacere@itu.edu.tr
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
2015-2016, Summer
Time and Location
Tuesday- 13.00-16.00
Course Outline
PART 1
Introduction to Control Engineering
Review of Complex Variables & Functions
Review of Laplace Transform
Review of Linear Algebra
PART 2
Linear Differential Equations
Obtaining Transfer Functions (TFs)
Block Diagrams
An Introduction to Stability for TFs
Concept of Feedback and Closed Loop
Basic Control Actions, P-I-D Effects
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
2/17
4
PART 3
Concept of Stability
Stability Analysis of the Closed Loop
System by Routh Criterion
State Space Representation and
Stability
Course Requirements
It is expected that the students do
the reading former lectures before
coming to the next class and be
prepared to participate in class
discussion and quizzes..
Homework assignments must be
turned in the following week.
A 70 % of attendance is a must to
take the final exam.
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
Grading
Evaluation criteria
contribution
Number
%
of final mark
Midterm Examinations
1
60 %
Quiz+Attendance 1 25 %
Homework 3 15%
TOTOL 100 %
Contrbution of in-term Assessment to
overall Grade 60 %
- Final Examination
40 %
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
Text Book
and
References
Academic Honesty
10
11
10/23/16
12
13
Best suited
for SISO, linear systems
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
14
15
16
17
18
20
21
23
25
26
Gc ( s ) isG 1 ( s )
If plant is known, the optimum controller
But, the plant is rarely known perfectly & disturbances always
exis
Example1: Most toaster are open-loops systems because
they are controlled by a timer. The time required to make
good toast must be estimated by the user,
Example2 : Washing machine *open loop system
27
28
29
30
31
- An automatic controller compares the actual value of the plant output with
the desired value,
determines the deviation. and produces a control signal that will reduce the
deviation to zero or a small value.
- The way in which the automatic controller produces the control signal is
called the
CONTROL ACTION
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
32
An automatic Controller
compares
deviation to
zero or a small value.
33
The actuator
plant
34
can be classified
control action
35
a two-position
control system
positions,
inexpensive
used extensively in both
industrial and household control systems.
36
37
Examples 1.
38
Solution
39
40
MATHEMATICAL MODELING
41
MATHEMATICAL MODELING
42
MATHEMATICAL MODELING
- A difference equation
an algebraic or transcendential
equality which involves more than one value of dependent
variable(s).
- The dependent varibles do not involve either differential or
derivatives.
- Differences equations are useful for relating the evulation of
variables from one discrete instant of time to another.
43
MATHEMATICAL MODELING
DIFFERENTIAL EQUATIONS, DIFFERENCE EQUATIONS and LINEAR SYTEMS
m(t ) &
x&
(t ) b(t ) x&(t ) k (t ) x(t ) 0
A time-invariant equation
an equation in which none of
the terms depend explicily on the independent variable time.
If the coefficients do not vary with time, the system is said to be a
time invariant system
&
mx&
(t ) bx&(t ) kx(t ) 0
44
d2
d
m 2 u (t ) b u (t ) ku (t ) 0
dt
dt
45
MATHEMATICAL MODELING
DIFFERENTIAL EQUATIONS, DIFFERENCE EQUATIONS and LINEAR SYTEMS
Examples
46
MATHEMATICAL MODELING
DIFFERENTIAL EQUATIONS, DIFFERENCE EQUATIONS and LINEAR SYTEMS
Examples
47
MATHEMATICAL MODELING
DIFFERENTIAL EQUATIONS, DIFFERENCE EQUATIONS and LINEAR SYTEMS
A linear term
and derivatives,
A linear equation
terms.
&
mx&
(t ) bx&(t ) kx(t ) 0
48
MATHEMATICAL MODELING
49
MATHEMATICAL MODELING
50
MATHEMATICAL MODELING
DIFFERENTIAL EQUATIONS, DIFFERENCE EQUATIONS and LINEAR SYTEMS
A linear term
and derivatives,
A linear equation
terms.
&
mx&
(t ) bx&(t ) kx(t ) 0
51
LTI
LTV
u1(t)
SYSTEM
y1(t)
u2(t)
SYSTEM
y2(t)
SYSTEM
y1(t)+y2(t)
u1(t)+ u2(t)
MATHEMATICAL MODELING
10/23/16
54
MATHEMATICAL MODELING
DIFFERENTIAL OPERATOR D and THE CHARACTERISTICS EQUATION
n
order linear constant-coefficient differential
equation, is the highest order of derivative .
Degree
the exponent of the higer power of the
highest order derivative has been cleared
D=d/dt
Dn=dn/dtn
55
56
57
58
59
60
61
62
63
64
65
66
67
MATHEMATICAL MODELING
10/23/16
68
MATHEMATICAL MODELING
Solving ODEs
Given a homogeneous 2nd order ODE of the
form
&
&
&
ax(t ) bx(t ) cx(t ) 0
10/23/16
69
MATHEMATICAL MODELING
10/23/16
70
MATHEMATICAL MODELING
10/23/16
71
MATHEMATICAL MODELING
Solving ODEs
10/23/16
72
MATHEMATICAL MODELING
or similar
Next, a particular solution with a form similar to the
forcing function and all its derivatives is assumed
x p (t ) Ad (t ) Bd&(t ) L
10/23/16
73
MATHEMATICAL MODELING
10/23/16
74
MATHEMATICAL MODELING
Example
75
76
Time
Input
Time
Input
Time
78
79
t<t0
V0 , t>t
may be expressed in terms of the unit step function as v(t) = V0u0(t t0)
If we let t0 = 0, then v(t) is simply the step voltage V0u0
80
(t) = du0(t)/dt
It is generally defined as
where
81
where t = 0 denotes the time just before t = 0 and t = 0+ is the time just after t
= 0.
For this reason, it is customary to write 1 (denoting unit area) beside the arrow that
is used to symbolize the unit impulse function, as in Fig.
82
Figure shows the impulse the impulse functions 5(t + 2), 10(t), and 4(t
3).
To illustrate how the impulse function affects other functions, let us evaluate
the integral
83
84
85
86
10/23/16
87
of Linear
Systems
Properties ofProperties
Linear
Systems
10/23/16
88
89
MATHEMATICAL MODELING
Y ( s)
G(s)
, u (0) u&(0) 0
U ( s)
G ( s ) G exp( jR G )
10/23/16
90
10/23/16
91
MATHEMATICAL MODELING
Example
Determine the dynamic behaviour of sepetatly excited
DC motor with connected mechanical load.
10/23/16
92
s j
s-domain algebraic equations
are easier to solve
Once a solution to the
algebraic equation is
computed in the s-domain,
an inverse Laplace
transform is used to revert
back to the time domain
10/23/16
93
A complex variable
L Laplace operator
F(s)
94
MATHEMATICAL
MODELING Theory
Review of Complex
Number
A complex variable
(karmak deiken)
consists of a real ()
and imaginary ()
part
A complex variable
can also be expressed
as a magnitude |s| &
phase s using the
s exp( jR s ),
phasor snotation
s 2 2 , R s tan 1
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
10/23/16
95
Eulers Theorem
j
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
96
Complex variable
97
98
For example
This is f (t)
This
limit approaches zero for >.
The
abscissa of convergence is therefore c
Laplace integral will converge only if s, the
The
real part of s, is greater than the abscissa of
convergence
99
Laplace transform of
Exponential Function
100
Lapla transform of
Step Functionce, 1(t)
101
Laplace transform of
Ramp Function
102
Laplace transform of
Sinusoidal Function
103
104
1. Linearity
3. Multiplication by eat
105
5. Real Differentiation
106
6. Real Integration
107
f (t)
A/t0
0
t0
108
f (t)
A/t0
0
t0
109
sF(s)
has no poles on the imaginary axis, this
110
111
112
Consider
where m<n.
If m=n, then find out the constant term and
separately write in the expansion, then invert.
If m>n, then find out the polynomial in s, and
write and invert it separately.
zis are zeros and pis are poles.
Poles and zeros may be complex numbers as
well
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
113
can be expanded as
where
114
Consider
where, deg A<q+m. This expression can be expanded as
with
k=0,1,,q1
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
115
116
An Example
Find the inverse Laplace transform of
Solution: Rewrite it as
117
118
119
120
121
>
>
>
>
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
>
>
122
Signals FG.
6
123
124
125
126
127
128
129
130
131
132
133
134
135
136
Evaluating the transfer function of a complex system from the block diagram
representation may become difficult (especially as the number of feedback
paths increase)
10/23/16
137
Branches
representing the
relationship
between the
output and input
to a sub-system
Nodes
representing the
variables or
signals
Paths which are
a continuous
sequence of
branches that
traverses from
one node the
another
Loops which are
closed paths that
originates and
terminates at the
same node
10/23/16
138
Aij
transmission function (Mathematical operatr mapping
Xj into Xi )
Xj and X i are
functions of s or z, Aij
functions , Aij (s) or Aij (s)
may
be a transfer
139
Multiplication rules
Paths
Forward Paths
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
140
Example
141
142
143
144
145
146
147
What is stability?
2
physical setup
u (t ) dt
10/23/16
148
149
150
151
152
153
Stability (cont.)
A system on the verge of instability is
termed a marginally stable system
Stability is a property of the
system and does not depend on the
inputs (similar to the transfer
function)
A plant may be designed to be
unstable for various reasons and a
stabilizing control can assure the
closed loop system is stable
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
10/23/16
154
155
Im
s-plane
Stable
x
Re(
Re
pi)<0
s-plane
t
h(t)
Im
x
Re(
Unstable
Re
pi)>0
156
Im
s-plane
Stable
Re
Re(
pi)<0
s-plane
h(t)
Im
x
Re(
pi)>0
Re
Unstable
t
157
s-plane
x
Re(
h(t)
Im
Re
pi)=0
s-plane
t
h(t)
Im
x
Re(
pi)=0
Re
158
159
10/23/16
160
161
162
163
164
165
The Hurwitz stability criterion, given next, gives conditions for all the
roots to have negative real parts in terms of the coefficients of the
polynomial. As stated in the discussions of Rouths stability criterion
for all the roots to have negative real parts, all the coefficients as must
be positive. This is a necessary condition but not a sufficient condition.
If this condition is not satisfied, it indicates that some of the roots have
positive real parts or are imaginary or zero. A sufficient condition for all
the roots to have negative real parts is given in the following Hurwitz
stability criterion: If all the coefficients of the polynomial are positive,
arrange these coefficients
in the following determinant:
166
167
168
169
170
171
172
( s ) an s n an 1s n 1 K a1s a0 0
If a0 = 0, the system will have a pole at the origin and
will become marginally stable
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
10/23/16
173
10/23/16
174
an
an 2
an 1 an 3
bn 1 bn 2
M
M
an 4 L
an 5 L
bn 3 L
10/23/16
175
an 1an 2 an an 3
an 1an 4 an an 5
bn 1
, bn 2
, K
an 1
an 1
and
10/23/16
176
10/23/16
177
10/23/16
178
Special Cases
1. First entry of a row in the RouthHurwitz array is zero, but the other
entries in the row are not zero (i.e.,
(s) is missing a term)
Substitute a small positive number ()
for the zero term
Continue to fill the array accordingly
Once array is finalized, let approach
zero
Check firstProf.Dr.M.Emin
column
for sign changes 10/23/16
Tacer - 2015-2016
Aydn niversitesi
179
10/23/16
180
10/23/16
181
10/23/16
182
10/23/16
183
10/23/16
184
10/23/16
185
10/23/16
186
Relative Stability
The border of absolute stability is the s = j line
The Routh-Hurwitz method determines the relative
location of the poles of the system with respect to
this s = j line
Even if the poles of a system are all stable, one
might be interested in determining the relative
stability of the poles with respect to each other
(e.g., which pole is most likely to become unstable
when perturbed)
A pole with a shorter settling time (ts) is more stable
Such a pole will be further away from the s = j axis
(i.e., will have a smaller negative real part)
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
10/23/16
187
Relative Stability
By modifying the Laplace
variable such that
s s 0
the j axis can be shifted
to the left by 0
The characteristic
equation can be
represented in terms of
the new variable
Routh-Hurwirtz based on
the modified polynomial
will provide the pole
locations relative to s =
0
Prof.Dr.M.Emin Tacer - 2015-2016
Aydn niversitesi
10/23/16
188
189
190
Determinant
191
Determinant
192
193
Characteristic Polynomial
Cayley-Hamilton Theorem
Every square matrix satisfies its characteristic
polynomial
194