ENGR 4220/5220:
Control Systems
Rick Hill, PhD
Dept of Mechanical Engineering
Lecture 1: Overview and
Intro
Introduction to the control system
design process
Control system example
ENGR 4220/5220 Lecture 1
open loop vs. closed loop
Introduction to modeling
Solving differential equations
Free response
Forced response
Control System Design
Process
Translate
types of models
physical vs.
empirical
mathematical
graphical
physical
system
Plant
Design (and
Constructio
n)
purpose of models
analysis
design
verification
diagrams
math
Model
behavior
Analyze
types of analysis
time domain
frequency domain
simulation
hardware in the
loop (HIL)
control
system
Controller
Design
types of control
supervisory
logic control
on/off control
P, PI, PD, PID
advanced
techniques
ENGR 4220/5220 Lecture 1
customer
input / govt
eng specs
regulations
Control System Example
desired
speed
Control
Engin
Algorith
e
throttle
m
force
angle
(voltage)
Car
ENGR 4220/5220 Lecture 1
Cruise Control Example
actual
speed
Open-loop Control
[feedforward]
desired
speed
advantages
simple to design
inexpensive
doesnt affect
stability
fast response
wind force,
gravity force
Control
Engin
Algorith
e
+
throttle
m
force
angle
(voltage)
Car
ENGR 4220/5220 Lecture 1
disadvantages
sensitive to errors
in model
sensitive to
disturbances
needs periodic
recalibration
actual
speed
Closed-loop Control
[feedback]
advantages
robust to errors
in model
robust to
disturbances
D wind force,
gravity force
CONTROLLER
ACTUATOR
PLANT
Control
R + E
-U
Engin
Algorith
Car
e
+
desired throttle
m
force
speed
angle
(voltage)
SENSOR
measured
speed
Speedomete
r
ENGR 4220/5220 Lecture 1
disadvantages
extra complexity
extra cost
can affect stability
can be slow to
respond
Y
actual
speed
Introduction to Modeling
A model is an abstraction of the
physical world
ENGR 4220/5220 Lecture 1
Used for analysis and design,
possibly before physical system
exists
Can be obtained from first principles
or experimentally
Purpose determines level of
abstraction, form
Complex enough, but no more
Model Derivation
ENGR 4220/5220 Lecture 1
From first principles
Use physical laws to derive
models
Provides understanding
Can use empirical data to
determine parameters, validate
model
8
Model Derivation
From empirical data
ENGR 4220/5220 Lecture 1
Feed a known input and observe output, fit model
to data
SYSTEM
Good for complicated systems (IC engine, battery)
Good for black-box systems (driver model)
Does not provide intuition, cant be widely applied
Complexity Depends on
Purpose
Simple enough to generate closed-form
solution
Less accurate, but provides intuition
di
K b& ea
dt
&
T b& J&
Ra ia La
ENGR 4220/5220 Lecture 1
Design/analysis model: simpler
10
Complexity Depends on
Purpose
Simulation model: more accurate
Static Systems
Dynamic Systems
Output is determined
only by the current
input, reacts
instantaneously
Relationship does not
change (it is static!)
Relationship is
represented by an
algebraic equation
Output takes time to
react
Relationship changes
with time, depends on
past inputs and initial
conditions (it is dynamic!)
Relationship is
represented by a
differential equation
input
SYSTEM
output
ENGR 4220/5220 Lecture 1
Static vs. Dynamic
Systems
12
Static vs. Dynamic
Systems
Motor from a Static
Motor from a Dynamic
Viewpoint
Viewpoint
500
450
400
Tstall
300
250
200
150
torque
100
50
0
ea1
0.5
1.5
Time
2.5
ea2
no-load
speed
Motor Torque
13
ENGR 4220/5220 Lecture 1
Motor Speed
350
0.5
1.5
Time
2.5
Solving Differential
Equations
Homogenous differential equations
Righthand side of equation equals 0
Represents free response of system
Solution consists of exponentials
1t
x(t ) a1e a2e
2t
where exponents are roots of the
characteristic eq.
ENGR 4220/5220 Lecture 1
& bx& kx 0
mx&
14
Solving Differential
Equations
Homogenous differential equations
& bx& kx 0
mx&
m b k 0
2
Roots can be found from the quadratic
formula
1,2
b
b 4km
2m
2m
2
ENGR 4220/5220 Lecture 1
For the above, the characteristic equation is
15
1t
2 t
displacement, x
Recalling thatx (t ) a1e a2 e
If the roots are completely real, then
the solution is exponential
If all negative, stable
If any positive, unstable
ENGR 4220/5220 Lecture 1
Solving Differential
Equations
16
time
Solving Differential
Equations
If the roots are complex, then can
rewrite in sines and cosines using
Eulers identity:
jt
Therefore,
cos t j sin t
a1e
( d j ) t
a1e
e
a2 e
t d jt
( d j ) t
a2 e
t d jt
( A cos d t B sin d t )
ENGR 4220/5220 Lecture 1
17
Solving Differential
Equations
( A cos d t B sin d t )
Above follows when have complex roots
of char. eq.
d j
real part = rate of decay (growth)
imag part = freq of oscillation
ENGR 4220/5220 Lecture 1
x(t ) e
18
Solving Differential
Equations
Forced differential
equations
Solution consists of two
xparts
(t ) xh (t ) x p (t )
xh is the homogenous solution
- same form as before, natural response of
system
xp is the particular solution
- generally same form as F(t), due to the
input
ENGR 4220/5220 Lecture 1
& bx& kx F (t )
mx&
19
xh determined from characteristic equation
Example
where 2 +8 +25=0 has roots 1,2 4 3 j
and x p has same form as F (t )
&
x& 8 x& 25 x 2t
has a solution of the form
x(t ) e
4 t
( A cos 3t B sin 3t ) at b
xh(t)
xp(t)
where
the homogenous portion dies out
(transient)
the particular portion remains (steady
state)
Example
Consider other types of forcing
functions:
&
x& 8 x& 25 x 5
x(t ) e 4t ( A cos 3t B sin 3t ) a
&
x& 8 x& 25 x 3e
x(t ) e
4 t
2 t
( A cos 3t B sin 3t ) ae
2 t
&
x& 8 x& 25 x 5sin 2t
x(t ) e
4 t
( A cos 3t B sin 3t ) a sin(2t )
C cos 2t D sin 2t
Example
Find the solution x(t) forx& 3 x 0, x(0) 5
Example
2 t
&
x
3
x
e
, x(0) 5
Find the solution x(t) for
Example (continued)