LTI System Analysis with

the Laplace Transform

 Laplace transforms
Linear
system Complex frequency domain
Time domain (t domain)

(s domain)
Differential Laplace Algebraic
equation transform L equation

Classical Algebraic
techniques techniques

Response Inverse Laplace Response

signal transform L-1 transform

The diagram commutes

Same answer whichever way you go
 Laplace Transform - definition

Function f(t) of time

Piecewise continuous and exponential order f (t ) < Kebt

st
F ( s ) f (t )e dt
=
0-

0- limit is used to capture transients and discontinuities at t=0

s is a complex variable (+j)
There is a need to worry about regions of convergence of
the integral
Units of s are sec-1=Hz
A frequency
 Laplace transform examples
Step function unit Heavyside Function 0, for t < 0
After Oliver Heavyside (1850-1925) u (t ) =
1, for t 0
" " ! st " !($ + j% )t "
! st ! st e e 1
F ( s ) = # u (t )e dt = # e dt = ! =! = if $ > 0
0! 0! s $ + j% s
0 0

Exponential function
After Oliver Exponential (1176 BC- 1066 BC)
" " #( s +$ )t "
#$t # st #( s +$ )t e 1
F (s) = ! e e dt = ! e dt = # = if % > $
0 0 s +$ s +$
0
Delta (impulse) function (t)
"
F ( s ) = ! $ (t )e # st dt = 1 for all s
0#
 Laplace Transform Pair Tables
Signal Waveform Transform
impulse ! (t ) 1
step u (t ) 1
s
ramp 1
tu (t )
s2
exponential 1
e "!t u (t )
s +!
1
damped ramp te
"!t
u (t )
( s +! ) 2
sine !
sin ( !t ) u (t )
s2 + ! 2
s
cosine cos( !t )u (t )
s2 + ! 2
damped sine e #"t sin ( !t )u (t )
!
( s +" ) 2 + ! 2
damped cosine #"t cos( !t )u (t ) s +"
e
( s +" ) 2 + ! 2
 Laplace Transform Properties

Linearity absolutely critical property

Follows from the integral definition

L{ Af1 (t) + Bf 2 (t)} = AL{ f 1 (t)} + BL{ f 2 (t)} = AF1 (s) + BF2 (s)
Example

$ A j"t # j"t ' A A

! L{ Acos("t)} = L % [e + e ]( = L{e } + L{e# j"t }
j"t

&2 ) 2 2
A 1 A 1
= +
2 s # j" 2 s + j"
As
= 2
s + "2

!
 Laplace Transform Properties
Integration property &t # F (s )
L % ' f (( )d( " =
$0 ! s

/t , # )t & ! st
Proof L . " f (0 ) d0 += " ' " f (0 ) d0 $e dt
-0 * 0 (0 %
"e " st t
Denote x= , and y = ! f (# ) d#
s 0
dx " st dy
so = e , and = f (t )
dt dt
"
)t & ) #e # st t & 1"
Integrate by parts L ' ! f (* ) d* $ = ' ! f (* ) d* $ + ! f (t )e # st dt
'(0 $% '( s 0 $% 0 s 0
 Laplace Transform Properties

Differentiation Property ' df (t ) $

L& # = sF ( s ) ! f (0!)
% dt "

Proof via integration by parts again

/ df (t ) , # df (t ) ! st ) df (t ) ! st & # #
! st dt
L. =
+ " e dt = ' e $ + s " f (t ) e
- dt * 0 ! dt ( dt %0 ! 0 !
= sF ( s ) ! f (0!)
Second derivative
(/ d 2 f (t ) %/ ( d . df (t ) + % ( df (t ) % df
L'
2 $ = L' , ) $ = sL ' $ ! (0 !)
/& dt /# & dt - dt * # & dt # dt

= s 2 F ( s ) ! sf (0!) ! f "(0!)
 Laplace Transform Properties
General derivative formula

)# d m f (t ) &# m m !1 m!2 " ( m)

L( m % = s F ( s ) ! s f ( 0 ! ) ! s f ( 0 ! ) ! L ! f (0 !)
#' dt #$

Translation properties
s-domain translation L{e "!t f (t )} = F ( s + ! )

t-domain translation

L{f (t ! a )u (t ! a )}= e ! as F ( s ) for a > 0

 Laplace Transform Properties

Initial Value Property lim f (t ) = lim sF ( s )

t "0 + s "!

lim f (t ) = lim sF ( s )
Final Value Property t !" s !0

Caveats:
Laplace transform pairs do not always handle
discontinuities properly
Often get the average value
Initial value property no good with impulses
Final value property no good with cos, sin etc
 Laplace Transform Properties
Multiplication-Convolution Property

L{ f (t) * g(t)} = F(s)G(s)

Critical property for notion of transfer function

!
More in a bit
 Rational Functions
We shall mostly be dealing with LTs which are
rational functions ratios of polynomials in s
bm s m + bm !1s m !1 + L + b1s + b0
F (s) =
an s n + an !1s n !1 + L + a1s + a0
( s ! z1 )( s ! z2 )L ( s ! zm )
=K
( s ! p1 )( s ! p2 )L ( s ! pn )
pi are the poles and zi are the zeros of the function
K is the scale factor or (sometimes) gain

A proper rational function has nm

A strictly proper rational function has n>m
An improper rational function has n<m
 How Laplace transforms are born
in MAE 143A

From linear ODEs:

y (n ) (t) + a1 y (n"1) (t) + ...+ an y(t) = b0 x (m ) (t) + b1 x (m"1) (t) + ...+ bm x(t)

Under zero initial conditions, Laplace tranforms produce

! (s n
+ a1sn"1 + ...+ an )Y (s) = (b0 sm + b1sm"1 + ...+ bm ) X(s)

For any given X(s) we can compute Y(s)!

b0 sm + b1sm"1 + ...+ bm
! Y(s) = H(s)X(s), H(s) := n
s + a1sn"1 + ...+ an

This is the same as convolution in the time domain:

$
! H(s) = L{h(t)}, Y (s) = H(s)X(s), y(t) = % #$
h(" )x(t # " )d"
 Example

Compute the impulse response of the RC circuit

described by the ODE:
1 1
y "(t) + y(t) = x(t)
RC RC
Apply Laplace tranform:

! "c 1
Y (s) = H(s)X(s), H(s) = , " c :=
s + "c RC
The impulse response is then
1
"1 1 " RC t
! h(t) = L {H(s)} = e u(t)
RC

!
 Inverting Laplace Transforms

We have a table of inverse LTs

Write F(s) as a partial fraction expansion

bm sm + bm"1sm"1 + L + b1s + b0
F(s) =
an sn + an"1sn"1 + L + a1s + a0
(s " z1 )(s " z2 )L(s " zm )
=K
(s " p1 )(s " p2 )L(s " pn )
#1 #2 # 31 # 32 # 33 #q
= + + + 2
+ 3
+ ...+
(s " p1 ) (s " p2 ) (s " p3 ) (s " p3 ) (s " p3 ) (s " pq )
Now appeal to linearity to invert via the table
! Surprise!
Nastiness: computing the partial fraction expansion is best
done by calculating the residues
 Inverting Laplace Transforms
Compute residues at the poles
1 d j"1
lim ( s ! a) F ( s) lim j"1 [(s " a) m F(s)]
s "a ( j "1)! s#a ds
Bundle complex conjugate pole pairs into second-order terms if
you want
[ ]
" # + j$ ) = s2 " 2#s + (# 2 + $ 2 )
(s " # " j$ )(s!

but you will need to be careful

Inverse
! Laplace Transform is a sum of complex exponentials

In Matlab, check out [ r,p,k] =residue(b,a), where

b = coefficients of numerator; a = coefficients of denominator
r = residues; p = poles; k = result of long division
 Strictly Proper Laplace Transforms
20( s + 3)
Find the inverse LT of F ( s ) =
( s + 1)( s 2 + 2s + 5)
k1 k2 k 2*
F(s) = + +
s +1 s +1 " 2 j s +1+ 2 j
20( s + 3)
k1 = lim ( s + 1) F ( s ) = = 10
s # "1 2
s + 2 s + 5 s = "1
5
20( s + 3) j !
k2 = lim (s + 1 " 2 j ) F (s) = = "5 " 5 j = 5 2e 4
s # "1 + 2 j ( s + 1)( s + 1 + 2 j ) s = "1 + 2 j

$ "t 5
("1+2 j )t + j # ("1"2 j )t " j # '
5
f (t) = &10e + 5 2e 4
+ 5 2e 4
) u(t)
% (
$ "t "t 5# '
= &10e +10 2e cos(2t + ) ) u(t)
% 4 (
 Laplace Transforms with Multiple Poles
Compute residues at the poles
j "1
1 d $ m F(s)'
lim (s " a)
( j "1)! s # a ds j "1 &% )(

2 s 2 + 5s 2( s + 1) 2 + ( s + 1) ! 3 2 1 3
Example 3
=
3
=
s + 1
+
2
!
3
! (s + 1 ) (s + 1 ) (s + 1) (s + 1)
( s + 1)3 (2 s 2 + 5s ) = d ( s + 1)3 (2 s 2 + 5 s ) =
lim 3 lim 1
s 1 ds 3
s 1 ( s + 1)3 ( s + 1)

1 d 2 ( s + 1)3 (2 s 2 + 5 s ) =
lim 2
2! s 1 ds 2 ( s + 1)3

1 2 s 2 + 5 s = t ( + 2)

L e 2 t 3t u (t )
3
( s + 1)

 Not Strictly Proper Laplace Transforms
3+ 2+
s 6 s 12 s + 8
Find the inverse LT of F ( s ) =
s 2 + 4s + 3
Convert to polynomial plus strictly proper rational function
Use polynomial division
s+2
F (s) = s + 2 + 2
s + 4s + 3
0.5 0.5
= s+2+ +
s +1 s + 3

Invert as normal

f (t) = [# "(t) + 2# (t) + 0.5e $t + 0.5e $3t ] u(t)

!
 System stability

Easy to determine using the transfer function H(s)

System is BIBO stable if

all the poles of the transfer function!lie in the
open left half of the s plane

The system is called marginally stable if there are

simple poles on the imaginary axis and no poles in
the right half-plane.

Marginally stable is a particular case of BIBO

unstable.
 System interconnections

H1 (s)
X(s) Y(s) X(s) Y(s)
H1 (s) H 2 (s) +
H 2 (s)
!
!
! !
! !!
Cascade interconnection Parallel interconnection
!

X(s) Y(s) X(s) Y (s)

H1 (s)H 2 (s) H1 (s) + H 2 (s)

! ! ! !
! !
 Feedback interconnection

X(s) + E(s) Y (s)

+ H1 (s)
"
! X(s) H1 (s) Y (s)
!
1+ H1 (s)H 2 (s)
!! !
!
H 2 (s)
! !
!
H1 (s) is the plant or original system we want to control
H 2 (s) is!the sensor/controller selected to make closed-loop
system have certain property (e.g., stability)

1 H1 (s) 1
Example: H1 (s) = , H 2 (s) = K, =
s 1+ H1 (s)H 2 (s) s + K

Much more about this in MAE143B

! !
!
 Block diagrams - revisited
Multiple ways of drawing a system block diagram for strictly
proper (nm) transfer functions
Y (s) bn sn + bn "1sn "1 +L + b1s + b0
H(s) = =
X(s) an sn + an "1sn "1 +L + a1s + a0

(if n>m, then some of the bi are zero)

H(s) as a product, H1 (s)H 2 (s)
! Y1 (s) 1
H1 (s) = ! =
X(s) an sn + an "1sn "1 +L + a1s + a0
! ! Y (s)
H 2 (s) = = bn sn + bn "1sn "1 +L + b1s + b0
Y1 (s)
! X(s) Y1 (s) Y(s)
H1 (s) H 2 (s)

!
! ! !
 Block diagrams - revisited
1
s Y1 ( s ) =
N

aN
{
X ( s ) ! "# aN !1s N !1 Y1 ( s ) +L + a1s Y1 ( s ) + a0 Y1 ( s ) $% }

H1 (s) H(s)
 Block diagrams - revisited
in cascade form
(s " z1 ) (s " zm ) 1 1
H(s) = A L L
(s " p1 ) (s " pm ) (s " pm +1 ) (s " pn )

in parallel form

K1 Kn
H(s) = +L +
(s " p1 ) (s " pn )