Feedback Control Systems (FCS)

Lecture-28-29
Modern Control Theory

Dr. Imtiaz Hussain

email: imtiaz.hussain@faculty.muet.edu.pk
URL :http://imtiazhussainkalwar.weebly.com/
 Introduction
• The transition from simple approximate models, which are
easy to work with, to more realistic models produces two
effects.
– First, a large number of variables must be included in the models.

– Second, a more realistic model is more likely to contain

nonlinearities and time-varying parameters.

– Previously ignored aspects of the system, such as interactions with

feedback through the environment, are more likely to be included.
 Introduction
• Most classical control techniques were developed for linear constant
coefficient systems with one input and one output(perhaps a few
inputs and outputs).

• The language of classical techniques is the Laplace or Z-transform

and transfer functions.

• When nonlinearities and time variations are present, the very basis
for these classical techniques is removed.

• Some successful techniques such as phase-plane methods,

describing functions, and other methods, have been developed to
alleviate this shortcoming.
 Introduction
• The state variable approach of modern control theory provides a
uniform and powerful methods of representing systems of arbitrary
order, linear or nonlinear, with time-varying or constant coefficients.

• It provides an ideal formulation for computer implementation and is

responsible for much of the progress in optimization theory.

• The advantages of using matrices when dealing with simultaneous

equations of various kinds have long been appreciated in applied
mathematics.

• The field of linear algebra also contributes heavily to modern control

theory.
 Introduction
• Conventional control theory is based on the input–output
relationship, or transfer function approach.

• Modern control theory is based on the description of system

equations in terms of n first-order differential equations, which may
be combined into a first-order vector-matrix differential equation.

• The use of vector-matrix notation greatly simplifies the a

mathematical representation of systems of equations.

• The increase in the number of state variables, the number of inputs,

or the number of outputs does not increase the complexity of the
equations.
 State Space Representation
• State of a system: We define the state of a system at time
t0 as the amount of information that must be provided at
time t0, which, together with the input signal u(t) for t  t0,
uniquely determine the output of the system for all t  t0.

• This representation transforms an nth order difference

equation into a set of n 1st order difference equations.

• State Space representation is not unique.

• Provides complete information about all the internal

signals of a system.
6
 State Space Representation

• Suitable for both linear and non-linear systems.

• Software/hardware implementation is easy.

• A time domain approach.

• Suitable for systems with non-zero initial conditions.

• Transformation From Time domain to Frequency

domain and Vice Versa is possible.

7
 Definitions
• State Variable: The state variables of a dynamic system are
the smallest set of variables that determine the state of the
dynamic system.

• State Vector: If n variables are needed to completely

describe the behaviour of the dynamic system then n
variables can be considered as n components of a vector x,
such a vector is called state vector.

• State Space: The state space is defined as the n-

dimensional space in which the components of the state
vector represents its coordinate axes.
8
 Definitions
• Let x1 and x2 are two states variables that define the state
of the system completely .

dx
x2 dt
Velocity State (t=t1)
State (t=t1)

State
Vector
x1 Position
x

Two Dimensional State space State space of a Vehicle

9
 State Space Representation
• An electrical network is given in following figure,
find a state-space representation if the output is
the current through the resistor.
 State Space Representation
• Step-1: Select the state variables.

v c 
 
iL 

Step-2: Apply network theory, such as Kirchhoff's voltage and current

laws, to obtain ic and vL in terms of the state variables, vc and iL.
Applying KCL at Node-1 i L  i R  iC
iC  i R  i L
dvC vC
C   iL (1)
dt R
 State Space Representation
Step-2: Apply network theory, such as Kirchhoff's voltage and current
laws, to obtain ic and vL in terms of the state variables, vc and iL.

Applying KVL at input loop

diL
v(t )  L  vR
dt
diL
L  vC  v(t ) (2)
dt
Step-3: Write equation (1) & (2) in standard form.
dvC 1 1
 vC  i L
dt RC C
State Equations
diL 1 1
  vC  v(t )
dt L L
 State Space Representation
dvC 1 1 diL 1 1
 vC  i L   vC  v(t )
dt RC C dt L L

 1 1
d vc   RC   v  0
C c   1  v(t )
  1  
dt  i L   0   i L   L 
 L 

 1 1
v c   RC   v  0
C c   1  v(t )
    1  
iL   0   i L   L 
 L 
 State Space Representation
Step-4: The output is current through the resistor therefore, the
output equation is
1
i R  vC
R

1  v c 
iR   0  
R  iL 
 State Space Representation
 1 1
v c   RC   v  0
C c   1  v(t )
    1  
iL   0   i L   L 
 L  Where,
x(t) --------------- State Vector
x (t )  Ax(t )  Bu(t ) A (nxn) ---------------- System Matrix
B (nxp) ----------------- Input Matrix
u(t) --------------- Input Vector

1  v c 
iR   0  
R  iL  Where,
y(t) -------------- Output Vector
C (qxn) ---------------- Output Matrix
y(t )  Cx(t )  Du (t ) D ----------------- Feed forward Matrix
 Example-1
• Consider RLC Circuit Represent the system in Sate Space and find
(if L=1H, R=3Ω and C=0.5 F):
– State Vector
iL
– System Matrix
+ +
– Input Matrix & Input Vector Vc Vo
– Output Matrix & Output Vector - -

dvc diL
C  u ( t )  iL L   RiL  vc Vo  RiL
dt dt
• Choosing vc and iL as state variables

dvc 1 1 diL 1 R
  iL  u(t )  vc  iL
dt C C dt L L
 Example-1 (cont...)
dvc 1 1 diL 1 R
  iL  u(t )  vc  iL
dt C C dt L L

 1
v c   0   v   1 
C c
  C u(t )
 i    1   
R  iL   
State Equation
 L    0
L L

Vo  RiL

v c 
Vo  0 R   Output Equation
iL 
 Example-2
• Consider the following system
K
x(t)
M
f(t)
B
Differential equation of the system is:

d 2 x(t ) dx(t )
M B  Kx(t )  f (t )
dt 2 dt

18
 Example-2
• As we know
dx d 2x dv
v 
dt dt 2 dt
• Choosing x and v as state variables
d 2 x(t ) dx(t )
dx M B  Kx(t )  f (t )
v dt 2 dt
dt
dv B K 1
 v x f (t )
dt M M M

 x   0 1   x  0 
     K B      1  f (t )
  v   
 v   M M M 
 Example-2

 x   0 1   x  0 
     K B      1  f (t )
  v   
 v   M M M 
• If velocity v is the out of the system then output equation is given as

 x
y (t )  0 1 
v 
 Example-3
• Find the state equations of following mechanical translational
system.

• System equations are:

d 2 x1 dx1 d 2 x2
M1 2
D  Kx1  Kx2  0 f (t )  M 2 2
 Kx2  Kx1
dt dt dt
 Example-3
• Now dx1 d 2 x1 dv1
 v1 
dt dt 2 dt
dx2 d 2
x2 dv2
 v2 
dt dt 2 dt
• Choosing x1, v1, x2, v2 as state variables
dx1
 v1
dt
dv1
M1  Dv1  Kx1  Kx2  0
dt
dx2
 v2
dt
dv2
f (t )  M 2  Kx2  Kx1
dt
 Example-3
• In Standard form
dx1
 v1
dt
dv1 D K K
 v1  x1  x2
dt M1 M1 M1
dx2
 v2
dt
dv2 K K 1
 x2  x1  f (t )
dt M2 M2 M2
 Example-3
dx1 dv1 D K K dx2
 v1  v1  x1  x2  v2
dt dt M1 M1 M1 dt
dv2 K K 1
 x2  x1  f (t )
dt M2 M2 M2
• In Vector-Matrix form

 0 1 0 0  0 
 x 1   K D K  1 x 
     0    0 
 v1    M 1 M1 M1  v1   
   0  f (t )
 x 2   0 0 0 1  x2 
   K K    1 
 v 2   M 0  0  v 2   
 2 M2  M2 
 Example-3
 0 1 0 0  0 
 x 1   K D K  1 x 
     0    0 
 v1    M 1 M1 M1  v1   
   0  f (t )
 x 2   0 0 0 1  x2 
   K K    1 
 v 2   M 0  0  v 2   
 2 M2  M2 

• If x1 and v2 are the outputs of the system then

 x1 
1 0 0 0  v
 1

y (t )  
0 0 0 1
  x2 
 
 v2 
 Eigenvalues & Eigen Vectors
• The eigenvalues of an nxn matrix A are the roots of the
characteristic equation.

• Consider, for example, the following matrix A:

Eigen Values & Eigen Vectors
 Example#4
• Find the eigenvalues if
–K=2
– M=10
– B=3

 x   0 1   x  0 
     K B      1  f (t )
  v   
 v   M M M 
 Frequency Domain to time Domain
Conversion
• Transfer Function to State Space
K
x(t)
M
f(t)
B
Differential equation of the system is:
d 2 x(t ) dx(t )
M 2
 B  Kx (t )  f (t )
dt dt
Taking the Laplace Transform of both sides and ignoring
Initial conditions we get
 Ms 2 X ( s )  BSX ( s )  KX ( s )  F ( s )
The transfer function of the system is
X ( s) 1

F ( s ) Ms 2  Bs  K

State Space Representation:

X (s) 1
 2 M
F (s) s  B
M s K
M

X (s) 1
s 2 P ( s )
 2 M
 2
F (s) s  B
M s K
M s P( s)
2
X (s) 1
s P( s)
 M
F ( s) P( s)  B
M s P ( s )  MK s  2 P ( s )
1
30
 1 2
X (s)  s P( s) ……………………………. (1)
M
B 1 K 2
F ( s)  P( s)  s P( s)  s P( s) ……………………………. (2)
M M

From equation (2)

B 1 K 2
P( s)  F ( s)  s P( s)  s P( s) ……………………………. (3)
M M

Draw a simulation diagram of equation (1) and (3)

-K/M

-B/M
F(s) 1/s 1/s X(s)
P(s) 1/M
• Let us assume the two state variables are x1 and x2.
• These state variables are represented in phase variable form as
given below.
-K/M

-B/M
x 2 1/M
F(s) 1/s x  x 1/s X(s)
x1
P(s) 2 1

• State equations can be obtained from state diagram.

x1  x2
K B
x 2  F ( s )  x1  x2
M M
• The output equation of the system is
1
x(t )  x1
M
 K B
x1  x2 x 2  F ( s )  x1  x2
M M

 x 1   0 1   x  0 
 x    K B       f (t )
1

 2   M    x2  1
M

1
x(t )  x1
M

 1   x1 
x (t )   0  
M   x2 
 Example#5
• Obtain the state space representation of the following
Transfer function.
 State Transition
• Any point P in state space represents the state of the system
at a specific time t.
x2
P(x1, x2)

x1

• State transitions provide complete picture of the system

x2
t0
t1
t6 t2

t3
x1
t5 t4
Solution of State Equations & State Transition Matrix
• Consider the state space model

x (t )  Ax(t )  Bu(t )
• Solution of this state equation is given as
t
x(t )   (t ) ( 0 )    (t   )Bu( )d
0
• Where  (t ) is state transition matrix.

 (t )   [( SI  A) ]  e
1 1 At
 Example-1
• Consider RLC Circuit

iL
+ +
Vc Vo
- -

dvc diL
C  u ( t )  iL L   RiL  vc Vo  RiL
dt dt
• Choosing vc and iL as state variables

dvc 1 1 diL 1 R
  iL  u(t )  vc  iL
dt C C dt L L
 Example-1 (cont...)
 1
v c   0   v   1 
C   C u(t )
 i    1
c
  
R  iL   
 L    0
L L

R  3, L  1 and C  0. 5

v c  0  2 vc  2

 i      i    u(t )
 L  1  3  L  0 
 Example-1 (cont...)
v c  0  2 vc  2
 i      i    u(t )
 L  1  3  L  0 
• State transition matrix can be obtained as
  S 0  0  2  
1

 (t )  1[( SI  A)1 ]  1       

  0 S  1  3  

• Which is further simplified as

 S  3 2 
 
1 ( S  1)( S  2 )
 (t )    ( S  1)( S  2 ) 
1 S 
 
 ( S  1)( S  2) ( S  1)( S  2 ) 
 Example-1 (cont...)
 S  3 2 
 
1 ( S  1)( S  2 )
 (t )    ( S  1)( S  2 ) 
1 S 
 
 ( S  1)( S  2 ) ( S  1)( S  2 ) 
• Taking the inverse Laplace transform of each element

( 2e  e ) ( 2e  2e )
t 2 t t 2 t
 ( t )   t  2 t t  2t 
 ( e  e ) ( e  2e ) 
 State Space Trajectories
• The unforced response of a system released from any initial
point x(to) traces a curve or trajectory in state space, with time
t as an implicit function along the trajectory.

• Unforced system’s response depend upon initial conditions.

x (t )  Ax(t )
• Response due to initial conditions can be obtained as

x(t )   (t ) x( 0 )
 Example-2
• For the RLC circuit of example-1 draw the state space trajectory
with following initial conditions.

vc (0) 1

  
 iL ( 0 )   2 
• Solution

x(t )   (t ) x( 0)

 c  ( 2e  e ) ( 2e  2e )    3e  3e 
v  t 2 t t 2 t
 1  t 2 t

 i    t  2 t t  2t   
  t  2t 
 L   ( e  e ) ( e  2e )  2   e  3e 
 Example-2 (cont...)
• Following trajectory is obtained
State Space Trajectory of RLC Circuit
2

1.5

t-------->inf
1

0.5
iL

-0.5

-1
-1 -0.5 0 0.5 1 1.5 2
Vc
 Example-2 (cont...)
State Space Trajectories of RLC Circuit
2

1.5
0 
1
1  

0.5
1
0  1 0 
iL

0  
 
-0.5

0
-1 1
 

-1.5

-2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Vc
 Phase Plane
• When the state variables are represented by phase
variables (i.e. CCF), the state space is called Phase space
or Phase plane.

 x 1   0 1 0   x1  x1  x2
 x    0   
 2  0 1   x2  x 2  x3
 x 3   a3  a2  a3   x3  x 3  a3 x1  a2 x2  a3 x3

• And the trajectory plotted in phase plane is called phase

portrait.
 Equilibrium Point
• The equilibrium or stationary state of the system
is when
x (t )  0
State Space Trajectories of RLC Circuit
2

1.5

0.5

0
iL

-0.5

-1

-1.5

-2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Vc
To download this lecture visit
http://imtiazhussainkalwar.weebly.com/

END OF LECTURES-28-29