Reading guide (first lecture)

 Read the introduction and introductory examples of feedback control from

slides.
 Learn Matlab and Simulink, especially how to write small m-code
programs in Matlab and doing simple simulations in Simulink.
These tools are needed in homeworks. See Tutorials in the
MyCourses page of the course. The first homework is given next
week.
 Familiarize yourself to the basic dynamic and static models of systems
(first principles models of physics). (Dorf: Chapters 1 and 2.)

 Note again that the textbook (Dorf and Bishop) contains it all
and much more than covered in this course. Reading every
chapter takes much time. The purpose is to use the textbook
when needed (and personally desired) for clarifications and
additional information.
Dynamic models, differential equations,
Laplace transformations and block diagram
algebra
 Example. Dynamical/Static Models
 The system is dynamical when its state is a F(t)
function of an earlier state (the system has
x(t)
memory and time-dependency).
m
 E.g. The effect of external power F on the
position of the mass position x-derived from
the force balance (M is mass, K spring B
k
constant and B attenuation coefficient)
mx(t )  Bx (t )  kx (t )  F (t )
 Static system does not depend on previous p(t) T(t)
state (memoryless and inertial system).
 E.g. Temperature T effect pressure p in n
closed, insulated container-derived from ideal
gas (n is the number of substances, V volume
V
and R gas constant at standard conditions)
p(t )V  nRT (t )
 Example. Dynamic/Static Models
 Simulations are carried out to change the
temperature in the external power and gas
container in the mechanical system.
 In a dynamic system, the response is
changing long after the impulse has entered
in input.
 In the static system, the input and response
are changing at the same time, and the
response can be determined directly by the
value of the input at the same instant.
Example of First-principles modelling (modelling by laws of
physics): Basic components of electrical circuits
i(t)
R
 Resistor (resistance)
v(t )  Ri (t )
v(t)
i(t)
Coil (inductance) L
di (t )
v(t )  L
dt
v(t)

C i(t)
 Capacitor (capacitance) i (t )  C
dv (t )
dt v(t)
 Example. Electrical circuit
Making a model for the electrical i1(t) R2 i2(t)

R1
circuit
 The input is v0 (t) and the output
vR1(t) vR2(t)
quantities are the voltages v1 (t)
v0(t) v1(t) C1 v2(t) C2
and v2 (t).
i3(t)
 Electric currents and resistors can
be modelled as

dv1 (t ) dv (t ) vR1 (t )  R1i1 (t ), vR 2 (t )  R2i2 (t )

C1  i3 (t ), C2 2  i2 (t )
dt dt
 Kirchoff's First Law
 Second Kirchoff law i1 (t )  i2 (t )  i3 (t )
 v0 (t )  v1 (t )
 i1 (t ) 
 0
v ( t )  v R1 (t )  v1 ( t )  0
v ( t )  R i
11 (t )  v1 (t )  R1
    
1v (t )  v R2 ( t )  v 2 (t ) 1v (t )  R i
2 2 ( t )  v 2 (t ) i (t )  v1 (t )  v2 (t )
 2 R2
 Example. Electrical circuit
 This model has voltages as states (memory elements), so it is advisable
to eliminate the electrical currents as unnecessary variables from the
developed equations
 dv1 (t ) 1 1 v0 (t )  v1 (t ) v1 (t )  v2 (t )
 dt  i3 (t )  1
i ( t )  i2 (t )   
 C1 C1 R1C1 R2C1

 dv2 (t )  1 i (t )  v1 (t )  v2 (t )
 dt C2
2
R2C2

 dv1 (t )  1 1   1   1 
     v1 (t )    v2 (t )    v0 (t )
 dt  R1C1 R2C1   R2C1   R1C1 
 
 dv2 (t )   1  v (t )   1  v (t )
 dt   1   2
 R C
 2 2 R
 2 2C
 Solutions of Differential Equations

 To understand and map the system under consideration, it is

essential to know how the output y(t) behaves as a function
of time in different conditions and with different control input
u(t).
 Since the models are differential equations, the output
function can be determined through the input response and
initial values given to solve the differential equation.
 Analytical solution cannot be guaranteed for a nonlinear or
distributed parameter differential equation group.
 An analytical solution can always be determined for a linear,
centralized, precisely defined, differential equation group.
 Linear Differential Equation solution

Linear Differential Equation/Equation Group can be solved

algebraically using Laplace transform

 Take Laplace transform of the system describing differential equations and

system input quantities
 Solve the algebraic equation obtained for output to input relation
 Take Laplace inverse transform and get the output function in time domain.
In control engineering applications, all time functions are
assumed zero before the start of the time period:

f (t )  0, for t  0
 Example
Time domain equation Time domain solution

 y (t )  2 y (t )  e  t
 y (t )  e  t
 y (0)  1

Laplace domain equation Laplace domain solution

  1  1
sY ( s )  y (0)  2Y ( s )    Y (s) 
  s  1 s 1
 y (0)  1

 Laplace transformation
 Definition: f(t) is the time domain function and F(s) is the corresponding
Laplace domain function
 b  j
1
f (t )  L1  F ( s ) 
2 j b j
F ( s)  L  f (t )   f (t )e st dt F ( s )e st ds
0

 If the limit values exist, then

 The final value theorem lim f (t )  lim sF ( s)
t  s 0

 The initial value theorem lim f (t )  lim sF ( s )

t 0 s 

 Laplace tables are shown in different sources with slightly different

expressions (usually either so that the time functions are easy to
convert to Laplace domain or so that the Laplace domain form can be
reversed easily.
Laplace Transforms (common examples)

Laplace-muunnos
Laplace functions Ajanfunctions
Time funktio
F (s) f (t ) T1
C1 F1 ( s )  C2 F2 ( s ) C1 f1 (t )  C2 f 2 (t ) T2
F (s  a) e at f (t ) T3
 as 0, ta
e F (s)  T4
 f (t  a ), t  a
1 s
F  f (at ) T5
a a
d
 F (s) f (t )t T6
ds
 Laplace Transforms (common examples)
Laplace-muunnos
Laplace functions Time Ajan funktio
functions

1
 F ( )d
s
f (t )
t
T7

t
F1 ( s ) F2 ( s )  f ( ) f (t   )d
0
1 2 T8

sF ( s )  f (0) f (t ) T9

s 2 F ( s )  sf (0)  f (0)  f(t ) T10


s n F ( s )  s n 1 f (0)  s n  2 f (0)   f ( n 1) (0)  f ( n ) (t ) T11

1 
t t
1
F ( s )    f ( )d   f ( )d T12
s s0  0
t 0
 Laplace Transforms (common examples)
Laplace-muunnos
Laplace functions Ajan funktio
Time functions Laplace-muunnos
Laplace functions Ajan funktio
Time functions
1  (t ) M1 1 1
1 ( s  a )( s  b) a b
 e bt  e  at  M9
1 M2
s 1 1 1
1 s ( s  a)( s  b)

ab ab(b  a )
 ae  bt  be  at  M10
t M3
s2 a
sin( at ) M11
1 tn s2  a2
M4
s n 1 n! s
cos(at ) M12
1  at s2  a2
e M5
sa a
e  bt sin(at ) M13
1 ( s  b) 2  a 2
te  at M6
( s  a)2 sb
n  at e bt cos(at ) M14
1 t e ( s  b) 2  a 2
M7
( s  a ) n 1 n! sa
 (t )  (a  b)e bt M15
1 1 sb
s( s  a) a
1  e  at  M8
Impulse Dirac delta

Step

Staircase
 Determinis

 The system input u(t) is commonly modeled as the following

functions:
 Unit Impulse function (Dirac delta function) u(t)
R z

; t  0
u (t )   (t )  S  ( t ) dt  1

T0; muulloin
otherwise 
U  ( s)   ( s)  1
 Unit step function
0 t
u(t)
R
u (t )  S
0; t 0
1
1
s
T1; t 0 U s ( s) 
s
 Staircase function 0 t
u(t)
R
u (t )  S
0; t0
r
Tt; t0 1
U r ( s)  2
s 0 t
 Example: Mass block
F(t)
An earlier example showed the mass equation:
mx(t )  Bx (t )  kx(t )  F (t ) x(t)
m
Solve for the response, when
R|k  5 x(0)  1 k B
S|B  2 x (0)  1
F (t )  2 (t ) (impulse)
Tm  1
Mass force balance in time domain: x(t )  2 x (t )  5 x(t )  2 (t )
and in Laplace domain:
( s 2 X ( s)  sx(0)  x (0))  2  ( sX ( s )  x(0))  5  X ( s )  2 1

cs X (s)  s  1h  2bsX (s)  1g  5 X (s)  2

2
 cs 2
h
 2 s  5 X ( s)  s  3
 Example: Mass block
Solving the expression for position of the mass block in Lapce domain
X(s): s3 s  1 2 s 1 2
X ( s)    
s  2s  5
2
( s  1)  2
2 2
( s  1)  2
2 2
( s  1) 2  2 2

Taking reverse Laplace transform to get position equation in time domain:

l q b
x ( t )  L X ( s)  e  t cos( 2t )  e  t sin(2t )  e  t sin(2t )  cos( 2t ) g
 Free and Forced responses

 The response can be divided into initial y0(t) (Free response)

and external control response yu(t) (Forced response). The
overall response of the linear system is the sum of the two
responses.
y ( t )  yu ( t )  y 0 ( t )

 Free response y0(t) is the response when external controls

do not affect the system ui(t) = 0.
 Forced response yu(t) is the response when all initial values
of the system y(n)(0) and ui(n)(0) are zero.
 Often, the term "response" refers to a forced response or
response to a given input, ignoring initial values.
 Example 2: Mass block

 Separating the last example into free and forced responses:

( s 2 X ( s )  sx(0)  x (0))  2  ( sX ( s )  x(0))  5  X ( s )  2   ( s )

( s 2  2s  5) X ( s )  ( sx (0)  x (0)  2 x (0))  2( s )

( sx(0)  x (0)  2 x (0)) 2 ( s )

X (s)    X 0 (s )  X u (s)
s  2s  5
2
s  2s  5
2

s 1 2
X ( s)  X 0 ( s)  X u ( s)  
( s  1)  2
2 2
( s  1) 2  2 2

l q l q
x (t )  L X ( s)  L X 0 ( s)  X u ( s)  e  t cos(2t )  e  t sin(2t )  x0 (t )  xu (t )
 Example 2: Mass block

 The free response starts from the initial state due to the
system's initial values (the position of the mass track at the
start is 1 and its initial speed-1)
 The forced response moves away from rest as a result of
external control or force (impulse-like nudge at the start)

R|x (t )  e cos(2t )
0
t

S|x (t )  e sin(2t )
u
t

Tx(t )  e bcos(2t )  sin(2t )g

t
 Transfer function
 Control techniques normally examine how external inputs
and disturbances affect the response; The effect of the initial
values is ignored and the focus is on the forced response.
 When initial values are zero, the expression of the response
is given in the form of Laplace output Y(s) which is a product
of the impulse response in Laplace domain G(s) and the
Laplace input function U(s).
 The Laplace representation of the model G(s) is called the
transfer function.
 U(s) G(s) Y(s)
Y ( s)  G ( s)  U ( s ) heräte Control
siirtofunktio vaste
input function response
 Block Diagram conversions: Signals
Y1 (s)
 In block diagrams, a single signal can be
exported to more than one block (signal
branching).
U(s) Y2 (s)
 A block diagram is an information chart and it
can branch information, but it does not
reduce that information. Each branch has Y3 (s)
the same information.
Y1 ( s)  Y2 ( s)  Y3 ( s)  U ( s) U 1 (s)
 The different signals can be combined using
a summation block. The combination can +
U 2 (s) + Y(s)
be either an addition or subtraction of
individual signals _
 Signs on the summation block indicate the
U 3 (s)
signs of the individual signals in the total.
Y ( s)  U 1 ( s)  U 2 ( s)  U 3 ( s)
 Passing a signal through a block
 As stated in previous lectures, in Laplace domain the output signal is
obtained by multiplying the input signal with a transfer function
U(s) Y(s) = G(s)U(s)
G(s)

 This basic formula can be used to derive a transformation from the serial
association of the equation blocks. Introduce the auxiliary variable e (s),
which is subsequently eliminated

U(s) e (s) Y(s)
G1(s) G2(s)

RSY (s)  G (s)e (s)

2
 Y ( s)  G1 ( s)G2 ( s)U ( s)  GTOT ( s)U ( s)
Te (s)  G (s)U (s)
1
U(s) Y(s)
 GTOT ( s)  G1 ( s)G2 ( s)  G1(s)G2(s)
 Passing a signal through a block
 The derivation for parallel blocks:
U(s) e (s)
G1 (s)
U(s) + Y(s)
U(s) e (s)+
G2 (s)

Y ( s )  e1 ( s)  e 2 ( s )

e1 ( s )  G1 ( s )U ( s )  Y ( s )  G1 ( s )U ( s )  G2 ( s)U ( s )
e ( s )  G ( s)U ( s )
 2 2

Y ( s )   G1 ( s)  G2 ( s )  U ( s)  GTOT ( s )U ( s)  GTOT ( s )  G1 ( s )  G2 ( s )

U(s) Y(s)
G1(s) + G2(s)
 Passing a signal through a block
 Loop relation (feedback) to the conversion formula is calculated as:
U(s) + e (s) Y(s)
G1 (s)
_
e (s) Y(s)
G2 (s)
R|Y (s)  G ( s)e (s)
1 1

S|e ( s)  U ( s)  e ( s)
1 2 b
 Y ( s)  G1 ( s) U ( s)  G2 ( s)Y ( s) g
Te ( s)  G ( s)Y ( s)
2 2

 Y ( s)  G1 ( s)U ( s)  G1 ( s) G2 ( s)Y ( s)  b1  G (s)G (s)gY (s)  G (s)U (s)

1 2 1

G1 ( s) G1 ( s)
 Y ( s)  U ( s)  GTOT ( s)U ( s)  GTOT ( s) 
1  G1 ( s)G2 ( s) 1  G1 ( s) G2 ( s)
Thus, the overall transfer function is the tf
U(s) G1 (s) Y(s)
of the forward path divided by the term (1 +
1 + G1 (s)G2(s) open loop transfer function) This will be
essential in future.