Chapter 2

Mathematical Models of Control

Systems
 2.1. Mathematical model
Mathematical model: A set of equations representing the
dynamics of the systems accurately or at least fairly well

How to get it?

 Theoretical approaches
 Experimental approaches

Types of mathematical model

 Differential equations
 Transfer function
 State variables

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 2/27

 2.2. Differential Equations
Consider a SISO linear time invariant (LTI) system

Differential equation:
d n y(t ) d n1 y (t ) dy(t ) d m r (t ) d m1r (t ) dr (t )
a0  a   a n 1  a y (t )  b  b    bm 1  bm r (t )
dt n1 dt m1
1 n 0 1
dt n dt dt m dt

 For proper systems, m ≤ n

 ai, bj are constants and depend on the system’s parameters
Differential equations can be obtained by utilizing physical laws
governing a particular systems, e.g., Newton’s law for mechanical
systems, Kirchhoff’s laws for electrical systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 3/27

 2.2. Differential Equations
Mechanical systems – linear motion

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 4/27

 2.2. Differential Equations
Mechanical systems – rotational motion

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 5/27

 2.2. Differential Equations
Mechanical systems – Example

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 6/27

 2.2. Differential Equations
Electrical systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 7/27

 2.2. Differential Equations
Electrical systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 8/27

 2.2. Differential Equations
Electrical systems - Example

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 9/27

 2.2. Differential Equations
Thermal systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 10/27

 2.2. Differential Equations
Thermal systems-Example

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 11/27

 2.2. Differential Equations
Electrical equivalent of thermal systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 12/27

 2.2. Differential Equations
Hydraulic systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 13/27

 2.2. Differential Equations
Hydraulic systems - Example

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 14/27

 2.3. Transfer function
Laplace transform - Review

Definition: L : f (t )  F ( s), F ( s)  L f (t )   f (t )e  st dt
0
where:
 s: complex variable (Laplace variable)
 L: Laplace operator
 F(s): Laplace transform of f(t)
Properties:
 Linearity: L . f1 (t )   . f 2 (t )   .F1 ( s )   .F2 ( s )
 Time shifting: L f (t  T )  e  sT .F ( s)
 Final value theorem
 df (t )  
lim f (t )  lim s.F ( s )
 Differentiation: L    sF ( s )  f ( 0 )
t  s 0
 dt 
t  F ( s)
 Integration: L f ( )d  
0  s
Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 15/27
 2.3. Transfer function
Laplace transform - Review
Laplace transform of some basic functions
Unit-step, us(t) Unit-impulse, δ(t)
1 t  0 dus  t  0
 s
u (t )    Lu (t ) 
1
  (t )    L (t )  1
 dt  0 t  0
s
0 t 0 s
 Useful in evaluating the system’s  Impulse response = L-1 of system’s
quickness in responding the inputs transfer function
with abrupt changes Rt 2
Prabolic-function input r (t )  u s (t )
2
Step-function input r (t )  Ru s (t ) u
(t )
Rt

s

Ramp-function input r(t )  Rus (t ) t)
R r(
 r (t ) = Rtus (t ) δ(t)
 Has the ability to test how the us (t)
system would response to a signal 1
Rt 2
r (t )  u s (t )
that changes linearly with time 2

0 t 0 t

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 16/27

 2.3. Transfer function
Transfer function
Consider a SISO linear time invariant (LTI) system

Define: R ( s )  Lr (t ); Y ( s )  Ly (t )

Transfer function:
Y ( s)
G(s) 
R( s)
 G(s) is defined as the ratio of the Laplace transform of the output to
the input
 All initial conditions are zero (initial relaxed systems)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 17/27

 2.3. Transfer function
Transfer function
Consider the differential equation of a TLI system
d n y(t ) d n1 y (t ) dy(t ) d m r (t ) d m1r (t ) dr (t )
a0  a   a n 1  a y (t )  b  b    bm 1  bm r (t )
dt n1 dt m1
1 n 0 1
dt n dt dt m dt

Taking the Laplace transform both sides (with all initial conditions are
set to zero)
Y ( s) b0 s m b1s m1  bm1s  bm
G( s)   n 1
, mn
R( s) a0 s  a1s  an1s  an
n

 G(s) is independent of the input and output of the system

 G(s) is a function of the complex variable s m
Y ( s) b0 s m b1s m1  bm1s  bm  ( s  zi )
i 1
G( s)    K
R( s) a0 s n  a1s n 1  an 1s  an n
 (s  p j )
j 1
 zi i = 1,2,..,m are zero of the system
 pj, j =1,2,..,n are poles of the system
Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 18/27
 2.3. Transfer function
Example: Obtain the transfer function of the following systems
a) y(t )  3 y (t )  y(t )  r(t )  r (t )

Take the Laplace transform both sides

s Y (s)  sy(0 )  y (0 ) 3sY (s)  y(0 ) Y (s)  sR(s)  r (0 ) R(s)
2    

Since all initial conditions are assumed to be zero

Y ( s) s 1
G(s)   2
R( s) s  3s  1
MATLAB
>> num=[1 1];
>> den=[1 3 1];
>> G=tf(num,den)
G=
s+1
-------------
s^2 + 3 s + 1
Continuous-time transfer function
Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 19/27
 2.3. Transfer function
Example: Obtain the transfer function of the following systems
b) ks(t )

v(t )  s(t )

ODE  
ms(t )  f (t )  ks(t )  s(t )  ms 2  s  k S ( s)  F ( s)

Transfer function S ( s) 1
G1 ( s)   2
F ( s) ms  s  k

V (s) s
G2 ( s)   2 Proof?
F ( s) ms  s  k

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 20/27

 2.3. Transfer function
Example: Obtain the transfer function of the following systems
c) Modeling of DC motor

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 21/27

 2.3. Transfer function
Example: Obtain the transfer function of the following systems
c) Modeling of DC motor
 Electrical part:
dia (t )
ua (t )  Ra ia (t )  La  e(t ) where e(t )  k a  (t )
dt
 Mechanical part: assume that no load is connected, i.e., Mt = 0
d (t )
J  M (t )  B (t ) where M (t )  k m ia (t )
dt
Take the Laplace transform
U a ( s )  Ra I a ( s )  La sI a ( s )  k a ( s )
Js  B
Js( s )  k m I a ( s )  B( s )  I a ( s) 
km
( s) 1
G( s)  
U a ( s) La J s 2  Ra J  La B s  Ra B  k 
km km km
a

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 22/27

 2.3. Transfer function
Example: Obtain the transfer function of the following systems
c) Modeling of DC motor
If there is no friction, i.e., B = 0, and assume ka = km = k
1
( s ) 1 k
G(s)   
U a ( s) La J s 2  Ra J s  k La J 2 Ra J
s  s 1
k k k 2
k  2

L
Define: Ta  a electromagnetic time constant
Ra
R J
Tc  a 2 mechanical time constant
k 
Since Ta<<Tc
( s) 1
 G( s)   ( s) 1
U a ( s) TaTc s 2  Tc s  1  G( s)  
U a ( s) Tc s  1

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 23/27

 2.3. Transfer function
Example: Obtain the transfer function of the following systems
d) Analog PID controller
V0 ( s)
Find G( s) 
Vi ( s)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 24/27

 2.3. Transfer function
Block diagram algebra
 Series system

u(t) w(t) y(t) u(t) y(t)

G1(s) G2(s) G1(s)G2(s)

 Parallel system

u(t) y1(t) y(t) u(t) y(t)

G1(s) G1(s)±G2(s)
±

G2(s)
y2(t)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 25/27

 2.3. Transfer function
Block diagram algebra
 Feedback system
u(t) e(t) y(t) u(t) G1 ( s ) y(t)
G1(s)
± 1 G1 ( s )G2 ( s )
w(t)
G2(s)

 Moving a summing point behind a block

u1(t) y(t) u1(t) y(t)

G(s) G(s)
± ±
u2(t)
u2(t)
G(s)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 26/27

 2.3. Transfer function
Block diagram algebra
 Moving a summing point ahead a block
u(t) y1(t) y(t) u(t) y(t)
G(s) G(s)
± ±
y2(t)
1 y2(t)
G (s)
 Moving a pickoff point behind a block

u(t) y1(t) u(t) y1(t)

G(s) G(s)
y2(t)
1 y2(t)
G (s)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 27/27

 2.3. Transfer function
Block diagram algebra
 Moving a pickoff point ahead a block

u(t) y1(t) u(t) y1(t)

G(s) G(s)
y2(t)
y2(t)
G(s)

 Interchange the positions of two consecutive summing points

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 28/27

 2.3. Transfer function
Block diagram algebra
Example: Find the equivalent transfer function of the following systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 29/27

 2.3. Transfer function
Block diagram algebra
Example: Find the equivalent transfer function of the following systems

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 30/27

 2.4. State Space Equation
y1 (t )
 State space equations r1 (t )
x1 (t )
r2 (t ) y 2 (t )
x (t )  Ax(t )  Br(t ) 
x2 (t )

 
y(t )  Cx(t )  Dr(t ) rm (t )
x n (t )
y p (t )

x(t )  x1 (t ), x2 (t ), , xn (t )
T
State variable vector
r (t )  r1 (t ), r2 (t ), , rm (t )
T
Input vector

y (t )  y1 (t ), y2 (t ), , y p (t ) 
T
Output vector

A  R nn , B  R nm , C  R pn , D  R pm

State matrix, input matrix, output matrix, and direct transmission matrix,
respectively
 Block diagram

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 31/27

 2.4. State Space Equation
 For SISO systems
x (t )  Ax(t )  br (t )

 y (t )  cx(t )  dr (t )
 Correlation between the state space equations and transfer function
Taking the Laplace transform of the state space equations (for SISO
systems)
 X( s)  sI  A  bR( s)
0 1
sX( s)  x(0 )  AX( s)  bR(s)

 Y ( s)  cX( s)  dR(s)
 
Y ( s)  csI  A  b  d R( s)
1

Transfer function
adjsI  A 
 csI  A  b  d  c
Y ( s) 1
G(s)  bd
R( s ) det( sI  A)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 32/27

 2.4. State Space Equation
 Choice of state variables

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 33/27

 2.4. State Space Equation
Example of mechanical system

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 34/27

 2.4. State Space Equation
Example of mechanical system

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 35/27

 2.4. State Space Equation
Example of mechanical system

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 36/27

 2.4. State Space Equation
Example of electrical system

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 37/27

 2.4. State Space Equation
Example of thermal system

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 38/27

 2.4. State Space Equation
Example of DC motor

dia (t )
ua (t )  Ra ia (t )  La  k a  (t )
dt
d (t )
J  km ia (t )  B (t )
dt
Derive the state space equations?

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 39/27

 2.4. State Space Equation
Example of DC motor workspace

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 40/27

 2.4. State Space Equation
Method to derive the state space equations from the differential equation

d n y (t ) d n 1 y (t ) dy(t )
 Case #1 n
 a1 n 1
 an 1  an y(t )  br (t )
dt dt dt
Define the state variables as
x1 (t )  y (t ); x2 (t )  x1 (t ); x3 (t )  x 2 (t ); ; xn (t )  x n 1 (t )

x (t )  Ax(t )  br (t )
State space equation 
 y (t )  cx(t )

 x1 (t )   0 1 0  0  0 
 x (t )   0 0 1  0  0 
 2    
x(t )     A         b     c  1 0  0 0
     
x (
 n 1 t )  0 0 0  1  0 
 xn (t )   an  an 1  an  2   a1  b 

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 41/27

 2.4. State Space Equation
Method to derive the state space equations from the differential equation
d n y (t ) d n 1 y (t ) dy (t )
 a   a n 1  an y (t )
 Case #2 dt n 1
dt n 1
dt
d n 1r (t ) d n  2 r (t ) dr (t )
 b1 n 1
 b2 n2
 bn 1  bn r (t )
dt dt dt
The equivalent transfer function
Y (s) b1s n 1  b2 s n  2  bn 1s  bn b( s )
G(s)   n n 1 n2

R( s ) s  a1s  a2 s   an 1s  an a( s )

R( s)
Y ( s )  b( s )  b( s ) F ( s )
a( s)
R( s)
R( s)  a( s)  a( s) F ( s)
a( s)

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 42/27

 2.4. State Space Equation
Method to derive the state space equations from the differential equation
 Case #2
R( s )  a( s ) F ( s )  s n F ( s )  a1s n 1 F ( s )  a2 s n  2 F ( s )    an 1sF ( s )  an F ( s )
Taking the inverse Laplace transform
d n f (t ) d n 1 f (t ) d n  2 f (t ) df (t )
r (t )  n
 a1 n 1
 a 2 n2
   a n 1  an f (t )
dt dt dt dt
Define d n f (t )
x1   a1 x1  a2 x2   an 1 xn 1  an xn  r
d n 1 f (t ) dt n
x1 (t ) 
dt n 1 d n 1 f (t )
x 2   x1
d n  2 f (t ) dt n 1
x2 (t ) 
dt n  2 

d 2 f (t )
df (t ) x n 1  2
 xn  2
xn 1 (t )  dt
dt df (t )
xn (t )  f (t ) xn (t )   xn 1
dt
Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 43/27
 2.4. State Space Equation
Method to derive the state space equations from the differential equation
 Case #2
Y ( s )  b( s ) F ( s )  b1s n 1 F ( s )  b2 s n  2 F ( s )    bn 1sF ( s )  bn F ( s )
Taking the inverse Laplace transform
d n 1 f (t ) d n  2 f (t ) df (t )
y (t )  b1 n 1
 b2 n2
   bn 1  bn f (t )
dt dt dt
 b1 x1  b2 x2  bn 1 xn 1  bn xn  x1 (t ) 
 x (t ) 
State space equations x (t )  Ax(t )  br (t )  2 
 x(t )    
 y (t )  cx(t )  
x (
 n 1 t )
 a1  a2   an 1  an  1  xn (t ) 
 1 0  0 0  0 
  
A       b   c  b1 b2  bn 1 bn 
   
 0 0  0 0  0 
 0 Controllable canonical form
0  1 0  0

Nguyen Le Hoa, Ph.D., Danang University of Science and Technology 44/27