Classical Control Theory

An Introduction to the Mathematics of System and Control Author : Efthimios Kappos

Introduction: Systems and Control

A system is like a black box with connections to the outside

world roughly divided into inputs and outputs. The role of a scientist or engineer is to study the behavior or a systems.

A system typically consists of a number of inter-connected subsystems.

A turbofan jet engine

T h e co m p l ex m e ch a n i l ch e m i l m a ch i e ca / ca n has a num ber of va ri b l s su ch a s fa n sp e e d , p re ssu re , te m p e ra tu re e tc th a t a re a e ca l e d sta te v a ria b le s . I i cru ci l th a t a l va ri b l s sta y w i i l t s a l a e th n re a so n a b l b o u n d s. e T h e o b j ct o f co n tro l th e o ry i to m a ke syste m s p e rfo rm sp e ci c ta sks e s fi b y u si g su i b l co n tro l a cti n s. Fo r exa m p l , w h e n th e p i o t p u sh e s n ta e o e l th e th ro ttl , th e a cti n s i tra n sl te d i to a co n tro l si n a l th a t m a ke s e o s a n g th e e n g i e i cre a se i p o w e r o u tp u t b y a sp e ci e d m o u n t. n n ts fi

A standard control system

The basic control arrangement involves measuring the outputs of a given system and comparing them with derivable responses to form the error signal. The errors are then fed into the controller to produce the control signal, which is finally applied to the system through actuator. an

The crucial feature is the closed loop that feeds the output back to the system. The sensory information telling you whether outputs are good or not. We call this job feedback.

Chapter 2: Linear Differential Equations and the Laplace Transform

* First order equations : W i x 0 is the starting value, the solution th is then Fo r k < 0 d e ca ys exp o n e n ti l y , th e syste m i ca l e d al s l sta b le Fo r k > 0 g ro w s exp o n e n ti l y , th e syste m i ca l e d al s l u n sta b le T h e l rg e r |k| i , th e fa ste r th e tra n si n t. a s e * Second order equations : 0> 0 the natural frequency the damping ratio

The solutions of equations 2.1 have form x(t) = x0est

E q u a ti n s 2 . 1 b e co m e s : o
2 x0 e st ( s 2 + 20 s + 0 ) = 0

dx = sx0 e st dt

d 2x = s 2 x0 e st dt 2

W e h a ve th e a u x ilia ry e q u a tio n :
2 ( s 2 + 20 s + 0 ) = 0

1 .N o dam pi g: 2 n

2 ( s + 0 ) = 0 s = i0 x(t ) = e0t ( A sin 0t + B cos 0t ) = ( A sin 0t + B cos 0t ) x0 = ( A sin 0 0 + B cos 0 0) B = x0 & x(t ) = 0 ( A cos 0t B sin 0t ) & x0 & & x0 = 0 ( A cos 0 0 B sin 0 0) x 0 = 0 A A = 0 T h e so l ti n i th i ca se u o n s & x0 i s x(t ) = ( x0 cos 0t + sin 0t ) 0

2.

U n d e rd a m p ed:

3 . C ri ca l ti dam pi g: n

4 . O ve rd a m p e d : th i ca se g i s tw o re a l p o l s a n d h e n ce a so l ti n s ve e u o th a t i a su m o f re a l exp o n e n ti l . s as

x(t ) = Ae s1t + Be s2t

The movement of poles as a function of the damping ratio

1< < 0

=0

> 1 c > 1
1< < 0

=0

Higher-Order Equations and Linear System Models

The general nth order homogeneous linear differential equation is

T h e h o m o g e n e o u s e q u a ti n d e scri e s th e u n fo rce d sy ste m . When o b w e h a ve n o n -zero right hand side , the system is a co n tro lle d sy ste m . Le t l o k a t a l n e a r syste m w h i w i l m e a n a syste m w i i p u t u and o i ch l th n o u tp u t y :

Tra n sfe r fu n ctio n s A n d s is called a co m p l x e

Laplace Transform Methods

Definition 2.1: If f(t), t 0 is a signal, the Laplace transform L(f) of f is defined by

W e ca n u se th e p o l r a fo rm : Pro p e rti s o f La p l ce e a Tra n sfo rm :

sh i : ft Pro o f:

L(e at ( x (t )) = x (t )e at e st dt = x (t )e (s + a )t dt = X ( s + a )
0 0

+ DifferentiationLaplace tra n sfo rm : Pro o f:

dX d de st st = x (t )e dt = x(t ) dt = ( t ) x (t )e st dt = L ( tx (t )) ds ds 0 ds 0 0 dnX d n e st = x (t ) dt = ( t ) n x(t )e st dt = L(( t )n x(t )) ds ds 0 0

Pro o f:

d de st dx ( xe st )dt = x dt + e st dt dt dt dt 0 0 0

dx dx st de st st L( ) = e dt = xe |0 x dt = x0 + xe st dt dt dt dt 0 0 0 dx = sX ( s ) x0 L( ) = sX ( s ) x0 dt
+ Positive Time Shift : W e h a ve : X ( s ) = L(h(t ) x(t )) Fo r a n y p o si ve ti e the transform of the function x shifted by ti m , ti e is m st

L(h(t ) x(t )) = e

X (s )

L(h(t ) x (t )) = x (t )e st dt
Le t ch a n g e th e va ri b l u a e = t -

L(h(t ) x (t )) = x (u )e s (u + ) du = e s X ( s )
0

Example 2.2:

&+ 2 y + 5 y = u (t ) & & y & y0 = 1, y0 = 1

Taking the Laplace transform of the both sides , we have :

& ( s 2Y ( s ) sy0 y0 ) + 2( sY ( s) y 0 ) + 2Y ( s) = U ( s) & U ( s ) + ( s + 2) y0 + y0 U ( s ) + ( s + 1) Y (s) = = 2 2 ( s + 2 s + 5) ( s + 2 s + 5) U (s) = 2 1 1 2s = 2 2 s s s

2 s + s 2 ( s + 1) s3 + s2 s + 2 Y (s) = 2 2 = 2 2 s ( s + 2 s + 5) s ( s + 2 s + 5) n( s ) Y (s) = d (s)

The Partial Fractions Expansion Theorem

P r o p o s i t i o n 2 . 1 : If the strictly proper rational function n ( s )/ d ( s ) of equation 2 . 29 has more than one distinct roots , then it can be written as a sum of rational functions of denominator degree less than deg d ( s ) , as follows :

a s+ p

where :

a=

n( p ) = lim( s p )G ( p ) q( p) s p

2 . To every multiple real root p of multiplicity k ,( k2) of the denominator d ( s ), so that d ( s )=( s - p ) k q ( s ), with q ( p ) 0 , corresponds exactly k terms in the partial fractions expansion ( PFE )

ck c1 c2 + + ... + 2 ( s p) ( s p) ( s p)k

where :

n( s ) h( s ) = q(s)

Example 2.3:

& 3 y + 2 y = u (t ) & & y u (t ) = (4t )h(t )

& y0 = 1, y0 = 1

Taking the Laplace transform , we have :

4 s2 & L( y ) = sY ( s ) y 0 = sY (s ) 1 L(4t ) =

& & & L( & = s ( L( y )) y 0 = s 2Y (s ) s + 1 y)

4 s Y ( s ) + 1 3( sY ( s ) 1) + 2Y ( s ) = 2 s 4 4 + s3 4 s 2 2 Y ( s )( s 3s + 2) = 2 + s 4 = s s2 s 3 4s 2 + 4 Y (s) = 2 s ( s 2)( s 1)
2

s 3 4s 2 + 4 Y (s ) = 2 s ( s 2)( s 1)

n( p ) a= = lim( s p )Y ( p ) q( p) s p

3 . To every simple pair of complex conjugate roots i of the denominator d ( s ), so that With q ( i ) 0, correspond exactly two terms of the form

4 . To every multiple pair of complex poles i of multiplicity k ,( k2), correspond exactly k terms of form

Example2.4:
Taking the Laplace transform , we have :

R0 =

iK / 0 K K |s =i0 = = 2 ( s + i0 )( s 2 + 2 ) 2i 0 ( 0 2 + 2 ) ( 0 2 )

K / 0 A1 = 2(R0 ) = 0, B1 = 2i( R0 ) = 2 ( 0 2 ) A2 = 0, B2 == K 2 (0 2 )

Y (s) =

K K 2 2 2 2 ( s 2 + 0 )( 2 0 ) ( s + 2 )( 2 0 )

0 K = 2 2 2 2 2 ( 0 ) 0 ( s + 0 ) ( s + 2 )
Taking the inversion of Laplace transform Y ( s ), we have :

With 0

Summary
The Laplace transform converts linear differential

equations into algebraic equations. The solution of these linear equations therefore leads to rational function expressions for the variables involved.
Simple real roots of the denominator correspond to

exponential time functions. A simple complex pair corresponds to a modulated trigonometric function, e at cos(t+)
Multiple roots imply the multiplication of the time

function corresponding to a simple root by a polynomial in t.

linear equations involving the sets of inputs and