You are on page 1of 25

Chapter 1 EEE3001- 8013

Chapter #1

EEE3001 & EEE8013

Design of Linear Control Systems

ƒ Introduction
ƒ 1st order dynamics
ƒ 2nd order dynamics
ƒ Nonhomogeneous differential equations

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 1/25


Chapter 1 EEE3001- 8013

Introduction

System: is a set of objects/elements that are connected or related to each


other in such a way that they create and hence define a unity that performs a
certain objective.

Control: means regulate, guide or give a command.

Task: To study, analyse and ultimately to control the system to produce a


“satisfactory” performance.

Model: Ordinary Differential Equations (ODE):

dx d 2x
ΣF = ma ⇔ f − f friction = ma ⇔ f − Bω = ma ⇔ f − B = m 2
dt dt

Dynamics: Properties of the system, we have to solve/study the ODE.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 2/25


Chapter 1 EEE3001- 8013

dx
First order ODEs: = f ( x, t )
dt

Analytical solution: Explicit formula for x(t) (a solution which can be found
dx
using separate variables, integrating factor…) which satisfies = f ( x, t )
dt

Example:

dx
The ODE = a has as a solution x(t ) = at .
dt

dx
First order Initial Value Problem : = f ( x, t ), x(t 0 ) = x0
dt

dx
Analytical solution: Explicit formula for x(t) which satisfies = f ( x, t )
dt
and passes through x0 when t = t0

Example:

dx
= a ⇔ ∫ dx = ∫ adt ⇔ x(t ) = at + C ⇒ x(0 ) = C ⇒ x(t ) = at + x0
dt

Ö INFINITE curves (for all Initial Conditions (ICs)).

For that reason a different symbol is used for the above solution: φ (t , t 0 , x0 ) .

You must be clear about the difference to an ODE and the solution to an
IVP! From now on we will just study IVP unless otherwise explicitly
mentioned.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 3/25


Chapter 1 EEE3001- 8013

Numerical solution: Using Simulink it is possible to see how the solution


x(t ) behaves. In order to do that we place one integrator, set the IC (in the
integrator block) and then we create the signal f ( x, t ) at the input of the
integrator. Remember the set the properties of the “Simulation parameters”
(I would say set the maximum step size to 1e-3) and of the “Scope” (change
in the Data History the option “Limit data points to last:”

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 4/25


Chapter 1 EEE3001- 8013

Example: Numerical solution of x& = − x 2 cos(t ) , x(0 ) = 0.1

In order to improve the presentation of our results I set in the “Save data to
workspace” option the following:

That way, once we run the simulation, we have in the command window the
matrix s which has in tits first column the values of the time and in the
second the values of our solution:

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 5/25


Chapter 1 EEE3001- 8013

>> plot(s(:,1),s(:,2))

0.115

0.11

0.105

0.1

0.095

0.09
0 1 2 3 4 5 6 7 8 9 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 6/25


Chapter 1 EEE3001- 8013

First order linear equations - (linear in x and x’)

⎧a (t ) x'+b(t ) x = c(t ), Non autonomous


General form: ⎨
⎩ax'+bx = c, Autonomous

Example: x'+ kx = u

Numerical Solution:

ƒ k=5, u=0.5

Gain

1
0.5 s
Constant Integrator Scope

0.1

0.08

0.06

0.04

0.02

0
0 1 2 3 4

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 7/25


Chapter 1 EEE3001- 8013

ƒ k=0.5, u=sin(10t)

Gain

0.5

1
s
Sine Wave Integrator Scope

0.2

0.15

0.1

0.05

-0.05

-0.1
0 5 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 8/25


Chapter 1 EEE3001- 8013

Analytical Solutions:

Integrating factor (NOT ASSESSED MATERIAL):

kdt k = const kt
e∫ = e

( )
e kt ( x'+ kx ) = e kt u ⇒ e kt x ' = e kt u

( )
⇒ ∫ e kt x ' dt = ∫ e kt udt

⇒ e kt x = ∫ e kt udt + c ⇒ x = e − kt ∫ e kt udt + e − kt c

t
or: x = e −kt
x(0 ) + e −kt
∫e
kt1
udt1
0

The equation that we derived is:

t
x=e −kt
x(0 ) + e −kt
∫e
kt1
udt1
0

Assuming that k>0 the first part is called transient and the second is called
steady state solution.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 9/25


Chapter 1 EEE3001- 8013

Response to a sinusoidal input (Not assessed)

y1 '+ ky1 = k cos(ω t )

y2 '+ ky2 = k sin (ω t ) ⇒ jy 2 '+ kjy2 = kj sin (ω t )

jy2 '+ kjy2 + y1 '+ ky1 = kj sin (ω t ) + k cos(ω t )

~ y = ke jωt
y '+ k~

( )
y e kt ' = ke(k + jωt )
Integrating factor: e kt => ~

e (k + jωt ) ⇒
~ k
y e kt =
k + j ωt
k
~
y= e jωt ⇒
k + jωt

⎛ −1(ω ) ⎞
1 j ⎜ ωt − tan k

~
y= e ⎝ ⎠

1+ ω
2

k2

⎛ ⎛ −1(ω ) ⎞ ⎞
⎜ k ⎟⎟
−1(ω ) ⎞
j ⎜⎜ ωt − tan
1 ⎟ 1 ⎛
Re( y ) = Re⎜
~ e ⎝ ⎠⎟= cos⎜ ωt − tan k

⎜ ⎟ ⎝ ⎠
⎜ 1+ ω 2 1+ ω 2
2 2

⎝ k ⎠ k

1 ⎛ −1(ω ) ⎞
cos⎜ ωt − tan k ⎟
1+ ω
2 ⎝ ⎠
k2

= magnified/attenuated amplitude and phase shifted.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 10/25


Chapter 1 EEE3001- 8013

Example:

u = cos(2πt ) k=1, x(0)=0.1:

Gain2 X0

-1 eu 0.1
Clock
Math
Function

Gain3
Scope1
1 eu
Clock1 Product1
Math 1
Function1 s
Product Integrator1

Sine Wave1

0.25
Steady
state Overall Transient
0.2

0.15

0.1

0.05

-0.05

-0.1

-0.15

-0.2
0 2 4 6 8 10

Exercise: Find the response for k=0.5, u=1 and u=-1, initial conditions: 0, 1,
-1. Describe the system’s behaviour.

Autonomous 1st order ODEs => Liner Time Invariant (LTI) systems

dx
= f (x) (not t on RHS).
dt

Analytic solution: Can be solved as before: Transient and steady state part.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 11/25


Chapter 1 EEE3001- 8013

d 2x
Second order ODEs: 2 = f ( x' , x, t )
dt

Second order linear ODEs with constant coefficients: x' '+ Ax'+ Bx = u

u=0 => Homogeneous ODE; I need two “representative solutions”

x' '+ Ax'+ Bx = 0 , assume x = e rt => x' = re rt & x' ' = r 2 e rt =>

x' '+ Ax'+ Bx = 0 ⇔ r 2 e rt + Are rt + Be rt = 0 ⇔

r 2 + Ar + B = 0 ; Characteristic or Eigenvalue equation => Check its roots.

− A ± A2 − 4B
r= , these are the Characteristic values or Eigenvalues.
2

ƒ Roots are real and unequal: r1 and r2 ( A 2 > 4 B =>Overdamped system)

x1 = e r1t and x2 = e r2t are solutions of the ODE =>

x = C1 x1 + C 2 x2 = C1e r1t + C 2 e r2t . If r1 and r2<0 then x → 0.

Example:

x' '+4 x'+3 x = 0 ⇔ r 2 + 4r + 3 = 0 ⇔ (r + 3)(r + 1) = 0

x = C1e −3t + C 2 e −t . Assume that x(0 ) = 1 and x' (0) = 0 :

x(0 ) = C1 + C 2 = 1 and x' = −3C1e −3t − C 2 e −t ⇒ x' (0 ) = −3C1 − C 2 = 0 =>

C1=-0.5, C2=3/2 => x = −0.5e −3t + 3 e −t :


2

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 12/25


Chapter 1 EEE3001- 8013

Numerical Solution:

4
Check ICs
Gain5
1 1
s s
Integrator3 Integrator2 Scope

Gain1

0.8

0.6

0.4

0.2

0
0 5 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 13/25


Chapter 1 EEE3001- 8013

Analytical Solution:

Gain2 Gain3

-3 eu -0.5
t Scope2
Clock Math
Function

Scope1
Gain4 Gain6

-1 eu 3/2

Math
Function1 Scope3

1.5
Overall
1 Part 2

0.5

Part 1
-0.5
0 5 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 14/25


Chapter 1 EEE3001- 8013

ƒ Roots are real and equal: r1=r2 ( A 2 = 4 B Critically damped system)

x1 = e rt and x2 = te rt => x = C1 x1 + C2 x2 = C1e rt + C2te rt

Example:

A=2, B=1, x(0)=1, x’(0)=0 => c1=c2=1

Numerical Solution

2
Check ICs
Gain5
1 1
s s
Integrator3 Integrator2 Scope

Gain1

Analytical Solution

Gain2 Gain3

-1 eu 1
t
Clock Math
Function

Gain6
Scope1
1
t*exp()
Product

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 15/25


Chapter 1 EEE3001- 8013

0.9 e-t

0.8
te-t
0.7 overall

0.6

0.5

0.4

0.3

0.2

0.1

0
0 2 4 6 8 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 16/25


Chapter 1 EEE3001- 8013

ƒ Roots are complex: r=a+bj A 2 < 4 B


o Underdamped system A ≠ 0

So x = e rt = e (a +bj )t = e at +bjt = e at e jbt = e at (cos(bt ) + j sin(bt ) ) =Re+jIm.

Theorem: If x is a complex solution to a real ODE then Re(x) and Im(x) are
the real solutions of the ODE:

x1 = e at cos(bt ), x2 = e at sin(bt ) =>

x = c1 x1 + c2 x2 = c1e at cos(bt ) + c2 e at sin(bt )


= e at (c1 cos(bt ) + c2 sin(bt ) ) = e at G cos(bt − φ )

, & φ = tan −1 ⎛⎜ ⎞
c1 c2
where G =
⎝ c1 ⎟⎠
cos⎛⎜ tan −1 ⎛⎜ 2 ⎞⎟ ⎞⎟
c
⎝ ⎝ c1 ⎠⎠

A=1, B=1, x(0)=1, x’(0)=0 => c1=1, c2=1/sqrt(3)

Numerical Solution

1
Check ICs
Gain5
1 1
s s
Integrator3 Integrator2 Scope

Gain1

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 17/25


Chapter 1 EEE3001- 8013

Analytical Solution

Gain2
u
-0.5 e
t
Clock
Math
Function Product Scope1

Gain3
Gain4

sqrt(3)/2 cos 1
bt
Trigonometric
Function
Gain6

sin 1/sqrt(3)

Trigonometric
Function1

1.5
C 1 cos(bt)

C 1 cos(bt)+C 2 sin(bt)
0.5

-0.5
C 2 sin(bt)

-1

-1.5
0 2 4 6 8 10

1.5

1 eat

C 1cos(bt)+C 2sin(bt)
0.5

-0.5

Overall
-1

-1.5
0 2 4 6 8 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 18/25


Chapter 1 EEE3001- 8013

o Undamped system A=0

x' '+0 + Bx = 0 ⇔ r 2 e rt + 0 + Be rt = 0 ⇒ r 2 = − B => Imaginary roots (If B<0


then I would have two equal real roots).

So r = jb => x = c1 cos(bt ) + c2 sin(bt ) = G cos(bt − φ )

A=0, B=1, x(0)=1, x’(0)=0 =>c1=1, c2=0:

0
Check ICs
Gain5
1 1
s s
Integrator3 Integrator2 Scope

Gain1

Gain2
u
0 e
t
Clock
Math
Function Product Scope1

Gain3
Gain4

1 cos 1
bt
Trigonometric
Function
Gain6

sin 0

Trigonometric
Function1

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 19/25


Chapter 1 EEE3001- 8013

0.8

0.6

0.4 Overall
0.2

-0.2

-0.4

-0.6

-0.8

-1
0 2 4 6 8 10

In all previous cases if the real part is positive then the solution will diverge
to infinity and the ODE (and hence the system) is called unstable.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 20/25


Chapter 1 EEE3001- 8013

Root Space

jb
jb

a
a

Name Oscillations? Components of solution


Overdamped No Two exponentials:

e k1t , e k 2 t , k1 , k 2 < 0
Critically No Two exponentials:
damped
e kt , te kt , k < 0
Underdamped Yes One exponential and one
cosine e kt , cos(ωt ) , k < 0
Undamped Yes one cosine cos(ωt )

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 21/25


Chapter 1 EEE3001- 8013

Natural frequency, damping frequency, damping factor

(NOT ASSESSED MATERIAL)

2nd order systems very important with rich dynamic behaviour

So A = 2ζω n , B = ω n2 => x' '+2ζω n x'+ω n2 x = 0

ζ is the damping factor and ωn is the natural frequency of the system.

− A ± A 2 − 4 B − 2ζω n ± (2ζω n )2 − 4ω n2
r= =
2 2

r = −ζω n ± ζ 2ω n2 − ω n2

ζ >0
1. Real and unequal ζ 2ω n2 > ω n2 ⇔ ζ 2 > 1 ⇒ ζ > 1 => Overdamped system

implies that ζ > 1; r1, 2 = −ζω n ± ζ 2ω n2 − ω n2 => replace at

x = C1e r1t + C 2 e r2t .

2. Real and equal ζ 2ω n2 = ω n2 ⇔ ζ 2 = 1 ⇒ ζ = 1 => Critically damped

system implies that ζ = 1; r = −ω n => x = C1e −ωnt + C2te −ωnt .


ζ >0
3. Complex ζ 2
ω n2 < ω n2 ⇔ ζ < 1 ⇒ ζ < 1 => Underdamped systems
2

implies ζ < 1 ;

r1, 2 = −ζω n ± j ω n2 − ζ 2ω n2 = −ζω n ± jω n 1 − ζ 2


ωd =ωn 1−ζ 2
= − ζω n ± jω d => x = e −ζωnt G cos(ω d t − φ ) ; ω d is called damped
frequency or pseudo-frequency.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 22/25


Chapter 1 EEE3001- 8013

4. Imaginary roots ζ =0 and therefore the solution is r = ± jω n =>


x = G cos(ω n t − φ ) ; so when there is no damping the frequency of the
oscillations = natural frequency. x = G cos(ω n t − φ )

5. In all the previous cases if ζ > 0 then the transient part tends to zero. If
ζ <0 then the system will diverge to infinity with or without
oscillations.

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 23/25


Chapter 1 EEE3001- 8013

NonHomogeneous (NH) differential equations

x' '+ Ax'+ Bx = u

ƒ u=0 => Homogeneous => x1 & x2.


ƒ Assume a particular solution of the nonhomogeneous ODE: xp
R
o If u(t)=R=cosnt => x P =
B
ƒ Then all the solutions of the NHODE are x = x P + c1 x1 + c2 x2
ƒ So we have all the previous cases for under/over/un/critically damped
systems plus a constant R/B.
ƒ If complementary solution is stable then the particular solution is called
steady state.

Example:

x' '+ x'+ x = 2 ⇒ x P = 2

x = 2 + c1 x1 + c2 x2 = 2 + e at (c1 cos(bt ) + c2 sin(bt ) )

x(0)=1, x’(0)=0 => c1=-1, c2=-1/sqrt(3)

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 24/25


Chapter 1 EEE3001- 8013
Numerical Part

1 1
s s
Integrator1 Integrator Scope

Gain1
2

Constant
Gain2
Scope2
u
-0.5 e
t exp(-at)
Clock Scope1
M ath
Function
Gain4 Gain5
Product

sqrt(3)/2 sin -1/sqrt(3)


bt sin(bt)
T rigonom etric
Function
Gain3
Scope5
cos -1
cos(bt)
T rigonom etric Scope4
Function1

Scope3

2.5

1.5
Particular solution
Overall
1

Homogeneous solution
0.5

-0.5

-1
0 2 4 6 8 10

Module Leader: Dr Damian Giaouris - damian.giaouris@ncl.ac.uk 25/25