Chapter 3: Solution of Differential Equations of Dynamic Systems

Spring, 2025

Ohmin Kwon
School of Electrical Engineering
Chungbuk National University
Republic of KOREA

Homepage: http://cel.cbnu.ac.kr
E-mail: madwind@chungbuk.ac.kr
©2025 All rights reserved.

3-1 INTRODUCTION TO DIFFERENTIAL EQUATIONS

- A wide range of systems in engineering are modeled mathematically by differential
equations.

<EX> A series electric RLC network

1
3-1-1 Linear Ordinary Differential Equations
- nth-order differential equation
d n y (t ) d n 1 y (t ) dy (t )
n
 a n 1 n 1
   a1  a0 y (t )
dt dt dt
(3-5)
d m u (t ) d m 1u (t ) du (t )
 bm m
 bm 1 m 1
  b1  b0u (t )
dt dt dt
u(t): the input of the system
y(t): the output of the system
- A first-order linear ordinary differential equation
dy (t )
 a0 y (t )  f (t ) (3-6)
dt
- A second-order linear ordinary differential equation
d 2 y (t ) dy (t )
2
 a1  a0 y (t )  f (t ) (3-7)
dt dt

3-1-2 Nonlinear Differential Equations

- The motion of a pendulum of mass m and length l can be described
d 2 (t ) g
m  mg sin  (t )  0 (   sin   0 ) (3-8)
dt 2 l

3-2 LAPLACE TRANSFORM

The Laplace transform is one of the mathematical tools used to solve Linear Ordinary
Differential Eq (LODE).
♣ Two features:
 The homogeneous and particular solutions of the eq. are obtained simultaneously in
one operation.
 The Laplace transformation converts Diff Eq into an algebraic equation n s.

ℒ {nth diff eqn}  nth polynomial eqn

3-2-1 Definition of Laplace Transform(LT)


F (s)   f (t )e st dt  L[ f (t )]
0


where  0
f (t )e t dt   .

2
♣ Example: See example 3-2-1 and 3-2-2 on page 86.
<Ex 3-2-1>
0, t  0
f (t )  us (t )   (3-13)
1, t  0

<Ex 3-2-2>
f (t )  e  st t  0 (3-16)

3-2-2 Important Theorems of the Laplace Transform

TABLE 3-1 Theorems of Laplace Transforms
Multiplication by a constant L[kf (t )]  kF ( s )

Sum and difference L[ f1 (t )  f 2 (t )]  F1 ( s)  F2 ( s)

Differentiation  df (t ) 
L  sF ( s )  f (0)
 dt 

 d n f (t )  n 1 n 2 (1)
  s F ( s )  s f (0)  s f (0)
n
L n
 dt 
   sf ( n2) (0)  f ( n1) (0)

d k f (t )
where f ( k ) (0) 
dt k t 0

2th Differentiation
 d 2 f (t )  2
 2   s F ( s)  sf (0)  f (0)
 dt 
3rd Differentiation

 d 3 f (t )  3
 3   s F ( s )  s f (0)  sf (0)  f (0)
2

 dt 

Integration F (s)
L   f ( ) d  
t

 0  s

3
 F (s)
L  f (t ) d dt1dt 2  dt n1   n
t1 t2 tn
 
 0 0 0  s
Shift in time
L[ f (t  T )u s (t  T )]  e Ts F ( s )

Initial-value theorem
lim f (t )  limsF ( s )
t 0 s 

Final-value theorem
lim f (t )  limsF ( s ) if sF (s ) does not have poles on or
t  s 0

to the right of the imaginary axis in the s-plane

Complex shifting L[e t f (t )]  F ( s   )

Real convolution
F1 ( s ) F2 ( s )  L   f1 ( ) f 2 (t   ) d 
t

 0 

 L   f 2 ( ) f1 (t   ) d   L[ f1 (t ) * f 2 (t )]
t

 0 

Complex convolution L f1 (t ) f 2 (t )  F1 ( s ) * F2 ( s )

<Ex 3-2-3> Final-value theorem

5
F ( s)  (3-37)
s ( s  s  2)
2

5
lim f (t )  (3-38)
t  2
<Ex 3-2-4>
w
F ( s)  (3-39)
s  w2
2

lim f (t )  ?
t 

3-2-3 Transfer Function

The transfer function G (s ) is related to the LT of the input and the output through the
following relation:
Y (s)
G (s) 
U (s)
Consider the system

4
 d n y (t ) d n 1 y (t ) dy (t )
n
 a n 1 n 1
   a1  a 0 y (t )
dt dt dt
(3-18)
d m u (t ) d m1u (t ) du (t )
 bm m
 bm1 m 1
   b1  b0 u (t ).
dt dt dt
Then the LT of the system with zero initial conditions is

( s n  a n1 s n 1    a1 s  a 0 )Y ( s )  (bm s m  bm 1 s m 1    b1 s  b0 )U ( s ). (3-20)

The transfer function is

Y ( s ) bm s m  bm 1 s m 1    b1 s  b0
G (s)   n . (3-21)
U (s) s  a n 1 s n 1    a1 s  a0
♣ The Properties of the transfer function (TF)
 TF is defined only for LTIS (Linear Time Invariant System).
 TF is LT of impulse response.
 All initial conditions are set to zero.
 TF is independent of input.
 TF is s-function.

♣ Proper Transfer Functions

Types of Transfer Function:
 Strictly proper: n  m
 Proper: n  m
 Improper: n  m
<4-1 page> 2th Differentiation’ Transfer function
d 2 y (t ) dy (t )
 a1  a0 y (t )  b0u (t )  (1)
dt dt
Transforming equation (1) Laplace,

5
<Cf>
unit-step response

 0, t  0 1
u (t )   , (u (t ))  U ( s ) 
1, t  0 s
1  1
y (t )   G ( s)   : unit  step response.
 s

Unit-impulse response

1, t  0
u (t )   , (u (t ))  U ( s )  1
 0, else
y (t )  1 (G ( s ) 1)  g (t ): unit  impulse response.

3-2-4 Characteristic Equation

Definition: The equation obtained by setting the denominator polynomial of the transfer
function to zero.

s n  an 1s n 1    a1s  a0  0 (3-22)

Meaning: From the roots of characteristic equation, the response of output can be predicted.

3-2-5 Analytic Function

3-2-6 Poles of a Function
3-2-7 Zeros of the function
3-2-8 Complex Conjugate Poles of Zeros
3-2-9 Final-Value Theorem

lim f (t )  limsF ( s ) if sF (s ) does not have poles on or to the right of the imaginary axis in
t  s 0

the s-plane

6
 3-3 INVERSE LAPLACE TRANSFORM BY PARTIAL-
FRACTION EXPANSION
The final-value theorem is very useful for the analysis and design of control systems because it
gives the final value of a time function by knowing the behavior of its Laplace transform at s=0.
The theorem states: If the Laplace transform f (t ) is F ( s ) , and if sF ( s ) is analytic (see
Sec. 3-2-5 on the definition of an analytic function) on the imaginary axis and in the right
half of the s-plane, then
1 c  j

2 j 
L-1 [ F ( s )]  f (t )  F ( s )e st ds. (3-41)
c  j

The final-value theorem is not valid if sF ( s ) contains any pole whose real part is zero or
positive, which is equivalent to the analytic requirement of sF ( s ) in the right-half s-
plane, as stated in the theorem. The following examples illustrate the care that must be taken
in applying the theorem.

3-3-1 Partial Fraction Expansion

Consider the rational function
Q( s)
G( s)  (3-42)
P( s)
where

P ( s )  s n  an1s n1    a1s  a0 (3-43)

♣ G (s ) has simple poles:

Q( s) Q(s)
G (s)   (3-44)
P ( s ) ( s  s1 )( s  s2 ) ( s  sn )
K1 K2 Kn
G (s)     (3-45)
s  s1 s  s2 s  sn

Q(s )
K1   ( s  s1 ) (3-47)
( s  s1 )( s  s2 ) ( s  sn ) s  s 1

K2 Kn
K1  K1  ( s  s1 )  (   )
s  s2 s  sn s  s
1

 Q( s) 
where K i   ( s  si ) .
 P( s )  s  si

7
<Cf>
 a   bt  bt
L   ae u (t ) or ae , t  0
 s  b 
<Ex 3-3-1>
5s  3 5s  3
G (s)   (3-49)
s  6 s  11s  6 ( s  1)( s  2)( s  3)
3 2

♣ G (s ) has multiple-order poles:

Q( s) Q( s)
G (s)   (i  1, 2,..., n  r ) (3-56)
P ( s ) ( s  s1 )( s  s2 ) ( s  sn  r )( s  si )r
K1 K2 K nr
G (s)   
s  s1 s  s2 s  sn  r
(3-57)
A A2 Ar
 1  
s  si ( s  si ) 2
( s  si ) r
where

Ar  ( s  si )r G ( s )  (3-58)
s  si

d
Ar 1  ( s  si ) r G ( s )  (3-59)
ds s  si

1 d2
Ar  2  ( s  si ) r G ( s ) 
2 
(3-60)
2! ds s  si


1 d r 1 (3-61)
A1  ( s  si ) r G ( s ) 
r 1 
(r  1)! ds s  si

8
<Ex 3-3-2> Consider the function
1 1
G (s)   5 (3-62)
s ( s  1) ( s  2) s  5s  9 s3  7 s 2  2 s
3 4

By using the format of Eq. (3-57), G(s) is written

K 0 K 2 A A2 A3
G (s)    1   (3-63)
s s  2 s  1 ( s  1) ( s  1)3
2

The coefficients corresponding to the simple poles are

1
K 0   sG ( s)  s 0  (3-64)
2
1
K 2   ( s  2)G ( s ) s 2  (3-65)
2
and those of the third-order pole are

<Cf>

9
<Ex 3-3-2> (sol2)
1 1
G (s)   5 (3-62)
s ( s  2)( s  1) 3
s  5s  9 s 3  7 s 2  2 s
4

k 0 k 2 A A2 A3
G (s)    1   (3-63)
s s  2 s  1 ( s  1) ( s  1)3
2

1  k k  
G ( s )  ( s  1)3    0  2  ( s  1)3  A1 ( s  1)2  A2 ( s  1)   A3
s ( s  2) s 1  s s  2  
  
s 1
0

1
 A3   1 (3-66)
(1)  (1  2)
d d  1   d  k k   
 G ( s )  ( s  1)3        0  2  ( s  1)3   A1  2( s  1)   A2
ds ds  s ( s  2)  s 1  ds  s s  2   
 
s 1
0

2 s  2
 A2  0 (3-67)
( s 2  2s) 2 s 1

d2 d2  1   d 2  k0 k2  3 
2  
G ( s )  ( s  1) 3
 2     2    ( s  1)   2  A1
ds ds  s ( s  2)  s 1  ds  s s  2   s 1

0

1 d2  1 
 A1   
2 ds 2  s ( s  2)  s 1
1 d  (2s  2) 
  
2 ds  ( s 2  2 s ) 2  s 1 (3-68)
d  s 1  ( s 2  2s )2  ( s  1)  2  (2 s  2)
  2 2 

ds  ( s  2s )  s 1 ( s 2  2s)4 s 1

(1  2) 2  0 1
    1
(1  2) 4 1
1 1 1 1
G (s)     (3-69)
2s 2( s  2) s  1 ( s  1)3
1  t e 2 t t 2 e  t
g (t )  e   (t  0) (3-70)
2 2 2
♣ G (s ) has simple complex-conjugate poles:
Q( s) Q( s )
G( s)  
P ( s ) ( s  s1 )( s  s1 )( s  s2 )( s  s2 ) 
K1 K1 K2 K2
    
s  s1 s  s1 s  s2 s  s2

10
where

K1  ( s    j )G ( s ) s  s   j ,
1

K1  ( s    j )G ( s ) s s   j .
1

<Ex 3-3-3> Consider the second-order prototype function

wn 2
G (s)  2 (3-73)
s  2 wn s  wn 2

11
3-4 APPLICATIN OF THE LAPLACE TRANSFORM TO THE
SOLUTION OF LINEAR ORDINARY DIFFERENTAL EQUATIONS
♣ Procedure:
 Transform DE to s-function by LT
 Manipulate the transformed Eq. and solve for the output variable
 Perform partial-fraction expansion
 Obtain the ILT from LT table

3-4-1 First-Order Prototype System

12
3-4-2 Second-Order Prototype System
d 2 y (t ) dy (t )
2
 2 wn  wn 2 y (t )  wn 2u (t ) (3-94)
dt dt
0, t  0,
u (t )  us (t )   (3-95)
1, t  0
dy (0) 1
If y (0)   0 , (us (t ))  U ( s )  , and ( y (t ))  Y ( s) , the output relation in the s-
dt s
domain is
1 wn 2
Y (s)  (3-96)
s s 2  2 wn s  wn 2
Where the transfer function of the system is
Y (s) wn 2
G (s)   (3-97)
U ( s ) s 2  2 wn s  wn 2
The characteristic equation of the prototype second-order system is obtained by setting the
denominator of Eq. (3-97) to zero:

 ( s )  s 2  2 wn s  wn 2  0 (3-98)

The two poles of the system are the roots of the characteristic equation, expressed as

s1 , s2   wn  wn  2  1 (3-99)

(1) System Is Critically Damped (   1 )

1 wn 2 1 wn 2
Y (s)  2  (3-100)
s s  2 wn s  wn 2 s ( s  wn ) 2

13
Further, the transfer function in Eq. (3-98) becomes
wn 2
G (s)  (3-101)
( s  wn ) 2
Where G(s) has two repeated poles at s   wn , as shown in Fig 3-7. In order to find the
solution of the differential equation, in this case, we obtain the partial fraction representation Eq.
(3-100) following the process defined is Example 3-3-2. Hence, by using the format of Eq. (3-
57), Y(s) is written as
K0 A1 A2
Y (s)    (3-102)
s ( s  wn ) ( s  wn ) 2
where

 1 wn 2 
K 0  ( s) 2
1 (3-103)
 s ( s  wn )  s  0

 1 wn 2 
A2  ( s  wn ) 2   1 (3-104)
 s ( s  wn )2  s  w
n

d  2 1 wn 2 
A1   ( s  wn )   1 (3-105)
ds  s ( s  wn ) 2  s  w
n

The completed partial-fraction expansion is

1 1 1
Y (s)    (3-106)
s ( s  wn ) ( s  wn ) 2
y (t )  1  e wnt  te  wnt , t  0 (3-107)

(2) System Is Overdamped (   1 )

1 wn 2
Y (s)  2 (3-108)
s s  2 wn s  wn 2

14
Further, the transfer function in Eq. (3-108) becomes
wn 2
Y (s)  (3-109)
s 2  2 wn s  wn 2
where G(s) has two poles at

s1 , s2   wn  wn  2  1 (3-110)

Let’s define
   wn (3-111)

w  wn  2  1 (3-112)

Is, for the purpose of reference, loosely called the conditional (or damped) frequency of the
system- note the system will not exhibit oscillations in the overdamped case, so usage of the
term frequency is not an accurate term. We use the following numerical example for easier
understanding of the approach.

<Ex 3-4-1>

1 wn 2 3 2
Consider Eq. (3-108) Y ( s )  with   , and wn  2 rad / s :
s s  2 wn s  wn
2 2
4

1 2 1 2
Y (s)   (3-113)
s s  3s  2 s ( s  1)( s  2)
2

K0 K K 2
Y (s)   1  (3-114)
s ( s  1) ( s  2)
where

 1 2 
K0  s  1 (3-115)
 s ( s  1)( s  2)  s  0

 1 2 
K 1  ( s  1)  2 (3-116)
 s ( s  1)( s  2)  s 1

 1 2 
K 2   ( s  2) 1 (3-117)
 s ( s  1)( s  2)  s 2

The completed partial-fraction expansion is

15
 1 2 1
Y (s)    (3-118)
s ( s  1) ( s  2)
Taking the inverse Laplace transform of Eq. (3-118),
y (t )  1  2e  t  e 2t t0 (3-119)

d 2 y (t ) dy (t ) dy (t )
<Ex 3-4-2> 2
3  2 y (t )  5us (t ) y (0)  1 and y (0)   2.
dt dt dt t  0

16
 (3) System Is underdamped (   1 )
1 wn 2
Y (s)  (3-128)
s s 2  2 wn s  wn 2
Further, the transform function in Eq. (3-128) becomes
wn 2
G (s)  2 (3-129)
s  2 wn s  wn 2
where G(s) has two complex-conjugate poles at

s1 , s2   wn  jwn 1   2 (3-130)

Let’s define
   wn (3-131)
as the damping factor, and

w  wn 1   2 (3-132)

As the conditional (or damped) frequency of the system. Figure 3-9 illustrates the relationships
among the location of the characteristic equation roots and  ,  , wn , and w. For the complex-
conjugate roots shown,
 wn is the radial distance from the roots to the origin of the s-plane, or wn 

( wn ) 2  wn 2 (1   2 ).

  is the real part of the roots.

 w is the imaginary part of the roots.
  is the cosine of the angle between the radial line to the roots and the negative axis
when the roots are in the left-half s-plane, or   cos  .

<Figure 3-9> Relationships among the characteristic-equation roots of the prototype second-
order system and  ,  , wn , and w.

17
<Ex 3-4-3> y(t )  34.5 y(t )  1000 y (t )  1000us (t ), y (0)  0, y(0)  0 (3-140)

18
3-4-3 Second-Order Prototype System - Final Observations

19
3-5 IMPULSE RESPONSE AND TRANSFER FUNCTIONS OF
LINEAR SYSTEMS

3-5-1 Impulse Response

Unit-impulse function

<Figure 3-14> Graphical representation an impulse function.

♣ The Properties of unit-impulse function

0 , t  t0
1)  (t  t0 )  
 , t  t0

2)   (t  t0 )dt  1

 t0 

3)  f (t ) (t  t0 )dt   f (t ) (t  t0 )dt
 
t0

t0

  f (t0 ) (t  t0 )dt  f (t0 )

t0 

4)   (t )    (t )e  st dt  1
0

5)   f (t )   (t )  F ( s )

Impulse response, g (t ) , is defined as the output when the input is a unit-impulse function
 (t ) in linear time-invariant system.
3-5-2 Time Response Using the Impulse Response
Once the impulse response of a linear system is known, the output of the system for any input
can be found by using the transfer function.

20
 L( y (t )) Y ( s )
G (s)   (3-154)
L(u (t )) U ( s )

3-5-3 Transfer Function (Single-Input, Single-Output Systems)

Transfer function of a linear time-invariant system, G (s ) , is defined as the Laplace transform
of the impulse response, with all initial conditions set to zero.
Transfer function G(s) is defined as
Y ( s)
G ( s )  L( g (t ))  (3-160)
U ( s)
with a linear time-invariant system, all the initial conditions set to zero, and Y(s) and U(s) are
the Laplace transforms of y(t) and u(t), respectively.

<Ex>

3-6 SYSTEMS OF FIRST-ORDER DIFFERENTIAL EQUATIONS:

STATE EQUATIONS
di (t ) 1
- In Eq. Ri (t )  L   i (t )dt  e(t ) if we let x1 (t )   i (t )dt and
dt C
dx1 (t ) di (t ) 1
x2 (t )   i (t ) , then Ri (t )  L   i (t )dt  e(t ) is decomposed into
dt dt C
the following two first-order differential equations:
dx1 (t )
 x2 (t ) (3-164)
dt
dx2 (t ) 1 R 1
 x1 (t )  x2 (t )  e(t ) (3-165)
dt LC L L
- In a similar manner, for Eq.
d n y (t ) d n 1 y (t ) dy (t )
n
 an 1 n 1
   a1  a0 y (t )  f (t )
dt dt dt
let us define

21
 x1 (t )  y (t )
dy (t )
x2 (t ) 
dt
(3-171)

d n 1 y (t )
xn (t ) 
dt n 1
then, the nth-order differential equation is decomposed into n first-order differential
equations:

dx1 (t )
 x2 (t )
dt
dx2 (t )
 x3 (t )
dt (3-172)

dxn (t )
  a0 x1 (t )  a1 x2 (t )    an  2 xn 1 (t )  an 1 xn (t )  f (t )
dt
 In control systems theory, the set of first-order differential equations in Eq. (2-
106) is called the state equations, and x1 (t ),..., xn (t ) are called the state
variables.
<Ex 3-6-1>

22
3-6-1 Definition of State Variables
The state variables must satisfy the following conditions:
 At any initial time t  t 0 , the state variables x1 (t 0 ), x2 (t 0 ), xn (t 0 ) define the
initial states of the system.
 Once the inputs of the system for t  t 0 and the initial states just defined are specified,
the state variables should completely define the future behavior of the system.
The state variables of a system are defined as a minimal set of variables,
x1 (t ), x2 (t ), xn (t ) , such that knowledge of these variables at any time t 0 and information on
the applied input at time t 0 are sufficient to determine the state of the system at any time t  t 0 .
Hence, the space state form for n state variables is
x (t )  Ax(t )  Bu(t ) (3-191)
where x(t) is the state vector having n rows,

 x1 (t ) 
 x (t ) 
x(t)   2  (3-192)
  
 
 xn (t )

and u(t) is the input vector with p rows,

 u1 (t ) 
 u (t ) 
u(t)   
2
(3-193)
  
 
u p (t )

The coefficient matrices A and B are defined as:

 a11 a12  a1n 

a a22  a2 n 
A   21 ( n  n) (3-194)
     
 
an1 an 2  ann 

23
 b11 b12  b1 p 
b b22  b2 p 
B
21
(n  p ) (3-195)
    
 
bn1 bn 2  bnp 

<Ex 3-6-2>
x (t )  Ax(t )  Bu (t )

 x 1 (t ) 
 x (t ) 
x(t )   2  (3-196)
 x3 (t ) 
 
 x4 (t ) 
u (t )  f (t ) (3-197)
 0 1 0 0
 (K  K ) K2 
 1 2
0 0
 M1 M1 
A  (4  4) (3-198)
 0 0 0 1
 K2 ( K  K3 ) 
 0  2 0
 M2 M2 
 0 
 0 
 
B   0  (4  1) (3-199)
 
 1 
 M 2 

3-6-2 The Output Equation

- An output of a system is a variable that can be measured, but a state variable does not
always satisfy this requirement.
- In general, an output variable can be expressed as an algebraic combination of the state
variables.

24
  y1 (t ) 
 y (t ) 
y (t )   2   Cx(t )  Du(t ) (3-200)
  
 
 yn (t ) 

 c11 c12  c1n 

c c22  c2 n 
C
21
(3-201)
    
 
cq1 cq 2  cqn 

 d11 d12  d1 p 
d d 22  d 2 p 
D
21
(3-202)
     
 
 d q1 dq2  d qp 

<Ex 3-6-3>
y(t )  3 y(t )  2 y (t )  2u (t ) (3-203)
If we let
x1 (t )  y (t ) (3-204)
and
x2 (t )  y(t )  x1 (t ) (3-205)
then second-order differential equation is decomposed into the following two first-order
differential equations:
x1 (t )  x2 (t ) (3-206)

x2 (t )  2 x1 (t )  3x2 (t )  2u (t ) (3-207)

As a results,

 x1(t )   0 1   x1 (t )   0 
       u (t )
 x2 (t )   2 3  x2 (t )   2 
 x (t ) 
y (t )  1 0  1 
 x2 (t ) 
<Ex 3-6-6>
y(t )  5 y(t )  y(t )  2 y (t )  u (t ) (3-225)

25
 If we let
x1 (t )  y (t ), x2 (t )  y(t )  x1 (t ), x3 (t )  y(t ) (3-227)
Then second-order differential equation is decomposed into the following two first-
order differential equations:
x1 (t )  x2 (t )
x2 (t )  x3 (t )
x3 (t )  2 x1 (t )  x2 (t )  5 x3 (t )  u (t )
As a result,
 x1(t )   0 1 0   x1 (t )  0 
      
 x2 (t )    0 0 1   x2 (t )   0  u (t )
 x (t )   2 1 5   x3 (t )  1 
 3    
A B

 x1 (t ) 
y (t )  1 0 0  x2 (t ) 
 
C
 x3 (t ) 

3-7 SOLUTION OF THE LINEAR HOMOGENEOUS STATET

EQUATION
x (t )  Ax(t )  Bu(t ) (3-231)

 sX( s)  x(0)   AX( s)  BU(s)

(3-232)
 sX( s)  AX(s)   x(0)  BU(s)   sI  A  X( s)
X( s )  ( sI  A) 1 x(0)  ( sI  A) 1 BU ( s ) (3-233)

x(t )  L1 ( sI  A)1 x(0)  L1 ( sI  A) 1 BU ( s ) (3-234)

y (t )  Cx(t )  Du (t ) (3-235)
<Ex 3-7-1>
 x1(t )   0 1   x1 (t )  0 
   u (t ) (3-236)
 x2 (t )  2 3  x2 (t )   2 
  
A B

26
3-7-1 Transfer Function (Multivariable Systems)
X ( s )  ( sI  A) 1 x(0)  ( sI  A) 1 BU ( s ) (3-245)
Y( s )  CX( s )  DU( s ) (-246)

Y ( s )  C  ( sI  A) 1 x(0)  ( sI  A) 1 BU ( s )   DU( s ) (3-247)

Y ( s )  C( sI  A) 1 x(0)   C( sI  A) 1 B  D  U( s ) (3-248)
 
G(s)

Y( s )
 C( sI  A)1 B  D  G ( s ) (3-249)
U( s ) x (0) 0

Y( s)  G ( s)U( s) (3-250)
In general, if a linear system has p inputs and q outputs, the transfer function between the
j th input and the i th output is defined as

Yi ( s )
Gij ( s )  (3-251)
U j (s)
with U k ( s )  0, k  1, 2, , p, k  j.

♣ i th output Output:

Yi ( s )  Gi1 ( s )U1 ( s )  Gi 2 ( s )U 2 ( s)    Gip ( s)U p ( s ). (3-252)

♣ Matrix-Vector Form:
Y( s)  G ( s)U( s)
where

27
  Y1 ( s )   U1 ( s )   G11 ( s ) G12 ( s )  G1 p ( s ) 
Y ( s )  U ( s )  G ( s ) G ( s )  G2 p ( s ) 
Y( s)   2  , U( s)   2  , G (s)  
21 22
.
           
     
Yq ( s )  U p ( s )  Gq1 ( s ) Gq 2 ( s )  Gqp ( s ) 

 Y1 ( s )   G11 ( s ) G12 (s )  G1 p ( s )   U1 ( s ) 
Y ( s )  G ( s ) G (s )  G2 p ( s )  U 2 ( s ) 
 2    21 22
.
          
    
Yq ( s )  Gq1 ( s ) Gq 2 (s )  Gqp ( s )  U p ( s ) 

<Cf> Two-Input Two-Ouput (TITO) system

 Y1 ( s )   G11 ( s ) G12 ( s )  U1 ( s ) 

Y ( s )   G ( s ) G ( s )  U ( s ) 
 2   21 22  2 

<Ex 3-7-2>


y1  4 y1  3 y2  u1 (3-254)

y1  y 2  y1  2 y2  u2 (3-255)

28
<Ex 3-7-3>

0 1 0 
x    x    u Y (s) 2
 2 3 2    C ( sI  A) 1 B  2
U ( s) s  3s  2
y  1 0 x 

2
G (s)  (3-268)
s  3s  2
2

3-7-2 Characteristic Equation from State Equations

From the state-variable approach, we can write Eq. G ( s )  C( sI  A ) 1 B  D as

29
 adj( sI  A)
G u ( s)  C BD
sI  A
(3-269)
C[adj( sI  A)]B  sI  A D
 .
sI  A

Thus, by setting the denominator of transfer function to zero, we have

sI  A  0, (3-270)

which is an alternative form of characteristic equation.

Note that if the coefficients of A are real, then the coefficient of sI  A are also real.

The roots of the characteristic equation are also referred to as the eigenvalues of the matrix A

<Ex 3-7-4>

<Ex 3-7-5>

<Ex 3-7-6>

30
3-9 LINEARIZATION REVISITED – THE STATE-SPACE
APPROACH

Let us consider the following vector-matrix state equations:

dx(t )
 f [x(t ), r (t )] (3-329)
dt
Let the nominal operating trajectory be denoted by x 0 (t ) , which corresponds to the nominal
input r0 (t ) . Then,

n
f i (x, r ) p
f i (x, r )
xi (t )  f i (x 0 , r0 )   ( x j  x0 j )   (rj  r0 j ) (3-332)
j 1 x j j 1 rj
x0 , r0 x0 , r0

Where i  1,..., n. Let

xi  xi  x0 i (3-333)
and
ri  ri  r0i (3-334)
Then
xi  xi  x0 i (3-335)
Since
x0i  f i (x0 , r0 ) (3-336)
Equation (3-332) is written as
n
f i (x, r ) p
fi (x, r )
xi   x j   rj (3-337)
j 1 x j j 1 rj
x0 , r0 x0 , r0


x  A x  B*r
*
(3-338)
 f1 f1 f1 
 x 
x2 xn 
 1 
 f 2 f 2 f n 

A   x1
*
x2 xn  (3-339)
     
 
 f n f n

f n 
 x1 x2 xn 

31
  f1 f1 f1 
 r 
r2 rn 
 1 
 f 2 f 2 f n 

B   r1
*
r2 rn  (3-340)
     
 
 f n f n

f n 
 r1 r2 rn 

<Cf>
x1  f1 ( x1 , , xn , u1 , , um )  f1 ( x(t ), u (t ))
x2  f 2 ( x1 , , xn , u1 , , um )  f 2 ( x(t ), u (t ))

xn  f n ( x1 , , xn , u1 , , um )  f n ( x(t ), u (t ))
 f1 ( x(t ), u (t )) 
 (t )  
X    F( x(t ), u (t ))
 
 f n ( x(t ), u (t )) 

<Ex 3-9-1>
x1  x2
g
( (t )  sin( (t ))  0)
g l (3-341)
x   sin( x1 (t ))
l

32
<Cf>
2
fi ( x1 , x2 , 0)
xi (t )  xi (t )   xi (t ), (i  1, 2)
i 1 xi
 f1 ( x1 , x2 , 0) f ( x , x , 0) 
 x1  1 1 2 x2 
 x1 (t )   x1 x2

 x (t )    f ( x , x , 0) f ( x , x , 0) 
 2  2 1 2
x1  2 1 2 x2 
 x1 x2
 
 f1 f1 
 x x   x 
 1 2
 1
 f 2 f 2   x2 
 x x 
 1 2 

33