I.

C Laplace Transform
Powerful method in analyzing/solving LTI differential
equations

I.C.1 Definition Function of

complex variable s
L
f (t ) 
 F ( s)
Function of t

F(s)  L [ f (t )]   0
e st f (t )dt
• F(s) complex variable function of complex variable s
• 0  0
• f (t ) is such that f (t )  0 for t  0
• Example of functions considered in this course:
f(t)=t f(t)=5

t t

f(t)=sin(t)

• Actually, any function f(t) u(t)

considered in this course can 1
be written as f(t)= f(t) u(t), with
u(t) being the step function: t
• Laplace Transform example:
f(t)  u ( t )  Step function at t  0
  st u(t)
F (s)  L [ f ( t )]  
0
u ( t )e dt
1
 st  st
e e
 t  t0
s s t
1
 (This term is 0 with real s > 0 )
s
Q: Laplace transform F(s) exists s

only on complex plane with

real s>0?
 Q: Laplace transform F(s) exists only for real s>0?
A: Analytic Extension Theorem extends existence of F(s) to
whole s-plane except at some singular point(s), if any.
Location(s) on
complex plane where
F (s ) becomes 
• Hence, Laplace Tansforms F(s) existing on part of s-
plane implies existence in whole s-plane except of some
singular point(s)! s

• For unit step function u(t) above,

F(s) =1/s over whole s-plane
except for singular point at s=0
Laplace Transform of
1
unit step function u(t): F(s) 
singular point at s=0 s
I.C.2 Properties of Laplace Transform

(a) Superposition of functions f1(t) and f2(t)

L [ f 1 ( t )   f 2 ( t )]   F1 ( s )   F2 ( s )

(b) Time Delay (Function shift to right by duration   0 )

L [ f ( t   )]  e  s  F ( s )
Given f (t )
L F (s)
t
Then
f (t   )
L  s
F (s)
e
 t
 (c) Time Scaling (Expansion/contraction of time axis)

L [ f ( at )]  1
a F ( as ) , a  0

Given
f (t )
L F (s)
to t
Then
f (at )
L
Correspond to
1
a F ( as )
to to t
contraction or
extension of t-axis a a
a >1 a<1
Contraction Extension
of time axis of time axis
(d) Modulation by Exponential factor

L [ e  at f ( t )]  F ( s  a )

Given  at
e
f (t )
L F (s)
Then t

 at
e f (t )
L F (s  a)
t
 (e) Differentiation of a function
L [ dtd f ( t )]  sF ( s )  f ( 0  )
n 1
Value of dtd n  1 f ( t )
at t  0 
Value of f ( t ) Value of dtd f ( t )
- Generalization at t  0  at t  0 

( n 1 )

L[ dn
f ( t )]  s F ( s )  s
n n 1 
f (0 )  s n 2
f ( 0 )    f ( 0  )

dt n

- For function and derivatives all starting at zero

L[ dn
dt n
f (t )]  s n F ( s )

(f) Integration of a function

t F (s)
L[ f ( ) d  ] 
0 s
(g) Convolution Theorem
L [ f 1 ( t )  f 2 ( t )]  F 1 ( s ) F 2 ( s )

where “convolution” of function f1(t) and f2(t) defined as:

t
f1 (t )  f 2 (t )   0
f1 (t   ) f 2 ( )d
t
  0
f1 ( ) f 2 (t   )d

• Laplace Transform of selected function in Table 2.3

• Laplace Transform of complicated functions maybe

obtained from Laplace Transforms of simpler functions
using Transform properties
I.C.3 Laplace Transform Theorems

• Poles and zeros of F(s): factorize F(s) to obtain

C ( s  z1 )...( s  z m )
F (s) 
( s  p1 )( s  p2 )( s  p3 )...( s  pn )

Then z1, z2, …., zm are the m zeros of the F(s), and
p1, p2, …., pn are the n poles of the F(s)

• If some of the pi are the same  Repeated pole case

• Repeated zero case not of interest
• Final Value Theorem: If all poles of sF(s) inside left
half plane (LHP) of s-plane, then f (  ) exists and
lim f ( t )  f   lim sF ( s ) Condition
t  s 0
f ( )

implies
existence
Example: 3( s  3) of f (  )
F ( s)  2
 f ( )  ? (Later!)
( s  2 s  10)
s
Y ( s)  2  y( )  ?
( s  10)

• Initial Value Theorem: Given any F(s),

f ( 0  )  lim sF ( s )
s 

s 
Example: Y ( s)  2
 y ( 0 )?
( s  10)
I.C.4 Inverse Laplace Transform
L
f (t ) L-1
F (s)
• Partial Fraction Expansion and term-by-term Inversion
The image part with relationship ID rId8 was not found in the file.

Partial Fraction
1 1 1 1 Expansion
 
2 s  2  2 s  4 
Term-by-term
L-1 L-1 Laplace Inversion

1  2t 1  4t
g (t )  e  e
2 2
• Partial Fraction Expansion: Distinct Pole Case
k ( s  z1 )...( s  z m )
- F ( s) 
( s  p1 )( s  p2 )...( s  pn )
p1  p2    pn
(All poles different)
- Write
C1 C2 Cn
F (s)    ... 
( s  p1 ) ( s  p2 ) ( s  pn )
- Determine C1,…,Cn as

Ci  ( s  pi ) F ( s ) s  pi i = 1,…,n
- Formulation applicable to distinct complex poles as well
• Partial Fraction Expansion: Repeated Pole Case

k ( s  z1 )...( s  z m )
- F (s) 
( s  p1 ) r ( s  pr 1 )...( s  pn )
p1  p1  ...  p1  pr 1  ...  pn
(first r poles equal)

- Write

 C1 C2 Cr  Cr1 Cn
F(s)      r
  
(s  p1) (s  p1) (s  p1)  (s  pr1) (s  pn )
2

Cr+1,…,Cn of non-repeated
poles determined by formula
of distinct pole case
 - Determine C1,…,Cr as
1 di
C r i  [( s  p1 ) r
F ( s )] s  p1 , i=0,…,r-1
i! ds i

* Special case: r =2
k ( s  z1 )...( s  zm )
F ( s) 
( s  p1 ) 2 ( s  pr 1 )...( s  pn )
 C C2  Cr1 Cn
F(s)   1  2
  ...
 (s  p1) (s  p1  (s  pr1)
) (s  pn )

1 di C 2  [( s  p1 ) 2 F ( s )] s  p
C 2 i  [( s  p1 ) 2
F ( s )] s  p1  1
i! ds i
 d
i=0,1  1
C  [( s  p 1 ) 2
F ( s )] s  p1
r-1  ds
 * Special case: r =3
k ( s  z1 )...( s  zm )
F ( s) 
( s  p1 )3 ( s  pr 1 )...( s  pn )
 C1 C2 C3  Cr1 Cn
F(s)     3
  ...
 (s  p1) (s  p1)2
(s  p1  (s  pr1)
) (s  pn )


 3
C  [( s  p 1 ) 3
F ( s )] s  p1
1 di 
 [(  ) 3
F ( s )] s  p1  d
C 3 i
i! ds i
s p1
 2
C  [( s  p 1 ) 3
F ( s )] s  p1
 ds
i=0,1,2  1 d2
r-1  1
C  [( s  p 1 ) 3
F ( s )] s  p1
 2! ds 2
• Term-by-Term Laplace Transform Inversion

- Distinct Real Pole: C L 1


  Ce pt
s p
- Multiple Real Pole: C L 1 1 r  1 pt
 
  C t e
(s  p) r
( r  1)!

- Distinct Complex Pole:

The image part with relationship ID rId8 was not found in the file.

where C  | C | e
j C
, p    j

- Multiple Complex Pole: Complicated!

• Example with complex pole:
- Partial Fraction Expansion:
2( s  5) C1 C2 C3
Y (s)    
s ( s  2 s  5)
2
s ( s  p ) ( s  p *)
( s  p )( s  p *), p   1  j 2
C 1  sY ( s ) s0
 2,
2( p  5) 2(4  2 j)
C 2  ( s  p )Y ( s )     1,
s p
p ( p  p *) (  1  2 j )( 4 j ) Im
C 3  C *2   1
-1
- Inverse Laplace Transform: C2  1
Re
 180 o

   1,   2 ,
| C 2 | 1, y ( t )  2  2 C 2 e  t cos(  t   C 2 )
by formula
 C 2   180 o
 2  2 (1) e  t cos( 2 t  180 o )
 2  2 e  t cos( 2 t )
 - Another method to solve (without PFE):
2 ( s  5)
Y (s) 
s ( s 2  2 s  5)
A Bs  C A ( s 2  2 s  5 )  s ( Bs  C )
  2 
s ( s  2 s  5) s ( s 2  2 s  5)
Equating coefficients A  B  0, A  2,
A ( s 2  2 s  5 )  s ( Bs  C ) 2 A  C  2, B   2,
 2 ( s  5) A  10 C  2
2 2 ( s  1) 2 2 ( s  1)
Hence, Y (s)   2  
s ( s  2 s  5) s 
( s  1) 2  2 2 
L-1 L-1
y ( t )  2  2 e  t cos( 2 t )
• Example with complex pole (modified from previous):
2 (3 s  5) C1 C2 C3
Y (s)    
s ( s  2 s  5)
2
s ( s  p ) ( s  p *)
( s  p )( s  p *), p   1  j 2

C 1  sY ( s ) s0
 2,
2 (3 p  5 ) 2(2  6 j)
C 2  ( s  p )Y ( s )    (  1  j ),
s p
p ( p  p *) (  1  2 j )( 4 j )
C 3  C *2 Im

Inverse Laplace Transform: -1 Re

 135 o
C2  (1  j ) -j
   1,   2 ,
| C 2 | 2,
y ( t )  2  2 2 e  t cos( 2 t  135 o )
 C 2   135 o

 2  2 e  t sin( 2 t )  cos( 2 t ) 
 - Another method to solve (without PFE):
2 (3 s  5)
Y (s) 
s ( s 2  2 s  5)
A Bs  C A ( s 2  2 s  5 )  s ( Bs  C )
  2 
s ( s  2 s  5) s ( s 2  2 s  5)
Equating coefficients A  B  0, A  2,
A ( s 2  2 s  5 )  s ( Bs  C ) 2 A  C  6, B   2,
 2 (3 s  5) A  10 C 2

Hence, 2  2s  2 2  2 ( s  1)  4
Y (s)   2  
s ( s  2 s  5) s  
( s  1) 2  2 2
2 2 ( s  1) 4
  
s 
( s  1) 2  2 2  
( s  1) 2  2 2 
L-1 L-1 L-1
y ( t )  2  2 e  t cos( 2 t )  2 e  t sin( 2 t )
• Example with complex pole (more challenging!):
- Partial Fraction Expansion: p  1  j4
1 C1 C2 C3
Y (s)    
s ( s  2 s  17 )
2
s ( s  p ) ( s  p *)
1 1 1
C1  , C2   e  j , C 3  C *2
17 (  32  8 j ) 1088

-32

- Inverse Laplace Transform: (  32  8 j ) -8j

1 1
   1,   4 , y (t )  2 e  t cos( 4 t   )
17 1088
1 1 2
| C 2 | ,   e  t cos( 4 t   )
1088 17 1088
 C 2   1 1 t  1 
  e  cos( 4 t )  sin( 4 t ) 
17 17  4 
Example 2.2 Application of Laplace Transform to solving
LTI differential equation
y(t )  4 y (t )  3 y(t )  r (t )
Initial conditions : y ( 0 )   , y ( 0 )  
Excitation on RHS: r(t)= 2 (meaning 2u(t) )

Apply Laplace transform : Step function

2
( s Y ( s )  sy(0)  y (0))  4( sY ( s )  y(0))  3Y ( s ) 
2
s
( s  4)   2
Y ( s)  
( s  1)( s  3) s( s  1)( s  3)
Taking Inverse Laplace Transform

3   t (   ) 3t 2 t 1 3t
y(t )  ( e  e )( e  e )
2 2 3 3

Response due to Initial Response due to

Condition excitation r(t)
(zero if =0, =0) (zero if r(t)=0)
 Second Reading Assignment:

PP. 87-95, PP. 101-106, PP. 128-131 of Textbook

I.C.5 Pole Locations and Time Response
• Information about y(t) from Y(s)?
-- From FVT and IVT:
 information on final and initial values of y(t)
-- From pole location pi of Y(s):
 more information on time behavior of y(t)

• Partial Fraction Expansion yields

C1 C C*
Y (s)   ...    ...
s  p1 s p s p *

Ci is real for real pole; and complex for complex pole

• Inverse Laplace Transform
 Each pole contributing a term to y(t)
• Contribution from a single real pole:
1
in Y ( s )  e   t in y (t )
s  
( pole at s    )
• Single real pole location and its contribution to y(t):

e t
e t e t e t e t
• Contribution from a pair of complex poles: jd
(a) With p in Cartesian Representation
-- p     j  d -
-- Inverse Laplace Transform -jd
C C* L1
 *
   2 C e  t
cos(d t  C )
s p s p
-- C  C e jC from Partial Fraction Expansion

-- Contribution to y(t): sinusoidal component with

Frequency:  d
Decay/growth Envelop: e   t
-- Different C values yield different phase shift  C
and amplitude scaling 2|C| of the sinusoidal signal
-- Graphic illustration: Location (of complex pole Growth Envelop
pair) and time contribution e t
Decay Envelop Frequency  d
y(t) e t y(t) y(t)

t t
Increasing
Frequency

j d

Faster Faster
decay growth
envelop envelop
 

 j d

Increasing
Frequency
(b) ( n ,  ) -Representation for p in LHP
 n : Distance between pole and origin
 : Cosine of pole’s angular position   cos 
relative to –ve real axis or 
  cos 1


p    n  j  n 1   2

-- Conversion to/from Cartesian

Representation
  n   2   d2 , 
    n ,   
  or   
  d   n 1   2
    
   d
2 2

-- Correspondingly, y(t) becomes sinusoidal signal with
Frequency:  d   n 1   2
Decay/growth Envelop: e   t  e   n t
-- Pole location and time response
in (n, ) representation
y(t)

Fixed  n , varying  : 0  1 t

 t
Response: 2 C e n cos(n 1   t  C )
2

Complex pole pair on j-axis

  90 o j 
y(t)   0 n

Complex pole pair

  0 . 707
t
  45 o
0    90 o


y(t)
 1
  0o
Double real pole at
t s   n
 j n
  0
-- Pole location and  value   0

p , p *    n  j  n 1   2
j n
0    1

0    90 o

 1


 1  1

 j n

-- System of systems according to its poles

- Underdamped: 0    1 , pole p,p* complex pole pair
- Critically damped:   1 , pole p,p* real and overlapping
- Overdamped:   1 , pole p,p* distinct real roots
 • Example: Spring-Mass-Damper system
d2 d
M 2 y(t )  b y(t )  ky(t )  F (t )
dt dt Wall friction, b k
No external
excitation!
M y
- With F(t)=0, y(0)  yo , y (0)  0
F
 s  b 

Y ( s)  2 M y
s b sk  o
 M M
- Two independent design parameters b/M and k/M to
yield  and n
b 2 k  s  2 n 
2 n  ,n  Y ( s )   2 y
2  o
M M  s  2 n s  n 
b
 
2 kM
- Underdamped case: 0    1 , i.e., 0  b  2 kM
Y(s) has two complex poles:
s  2 n C C*
Y (s)  yo  
( s  p)(s  p*) ( s  p) ( s  p*)

p    n  j  n 1   2
( n ,  ) representation!

  j 1 2 yo
C yo  C  ; C    90
j2 1   2
2 1  2

where   cos  , and hence

yo
y (t )  e nt sin( 1   2 nt   )
1  2
- Critically damped case:   1 or b  2 kM
Y(s) has two overlapping real poles:

Y (s) 
 s  2n 
y 
A

B
( s   n )2 ( s  n ) ( s   n )2
o

A  yo
y (t )  yo (1  nt )e nt
B   n yo

Response reaches its final value at the shortest time

without oscillation, and b is the minimum damping
that results in non-periodic motion (i.e. simple
decay).
- Overdamped case:   1 or b  2 kM
Y(s) has two distinct (non-overlapping) real poles:

Y ( s) 
s  2n  yo 
A

B
( s  p1 )( s  p2 ) ( s  p1 ) ( s  p2 )

p1 , p 2    n   n  2
1

   2 1     2 1
A yo , B yo
2  1 2
2  12

    2 1 p t    2 1 p t 
y(t )   e1  e 2  yo
 2  2 1 2  2
 1 
 
- Example of responses: Recall F(t)=0, y (0)  yo , y (0)  0

yo  1

b  2 kM

b  2 kM

b  2 kM

- For the case F(t)=0, y(0)  yo , y (0)  v0 , see

http://www.efunda.com/formulae/vibrations/sdof_free_damped.cfm
- Question: is overdamped/underdamped response better?
Design Concept #1:
To properly place the poles of Y(s) on complex s-plane
so that the resulting y(t) yields satisfactory performance

Location for poles of output Y(s)? s

? ?
? ?

 j n
• Example 2.15: Design of an accelerometer
(illustrating Design Concept #1: pole placement effect on y(t))

- Idea: constant Force Fo on sled mass Ms

Want y as measurement of acceleration: y  a  Fo
Ms

x = position

Fo
- Absolute movement of mass inside case: x+y
d2 d
M 2
( y  x )   ky  b y
dt dt
- Force acting on sled mass:
d2
M s 2 x  Fo
dt
- Combining
b k Fo
y  y  y
M M Ms
- Apply Laplace Transform:
 b 
sy ( 0 )   y ( 0 )  y(0)  Fo
Y (s)   M  
Ms
s2 
b
s
k  b k 
s s 2  s 
M M  M M 
- Conduct PFE, with s1 and s 2 roots of  s 2  s   ,
b k
 M M 
 b 
sy ( 0 )   y ( 0 )  y (0)  Fo
Y (s)   M  Ms
( s  s1 )( s  s 2 ) s ( s  s1 )( s  s 2 )

A B E C D
    
( s  s1 ) ( s  s 2 ) s ( s  s1 ) ( s  s 2 )

Due to initial value Due to excitation

- Constants A, B, C, D and E can be determined as in PFE
Specially,  M  Fo
E   
 k Ms
- Inverse Laplace transform,
 M  Fo
y ( t )  Ae  Be  
s1t s2t
  Ce s1t  De s 2 t
 k Ms
- Acceleration measurement:
* With b, k, M +ve, roots s1 and s 2 inside LHP
* As t  large, e s1t , e s 2 t  0
Fo/Ms =a is acceleration!
and y ( t )  y ss    M  Fo  y ss  acceleration!
 k Ms
Can also obtain from FVT M 
y (t ) y ss    a
 k 
t
k
* Or, plotting  y (t ) vs. t
M
yield a as steady state value! 
k
y (t )
M

Fo
Steady state value is acceleration a 
Ms
- Specific numerical design in Example 2.15
b k F
y ( 0 )   1, y ( 0 )   2 ,  3,  2 , o  a  3  Fig. 2.46
M M Ms

k
 y (t )
M

2.46

* What are the roots s1 and s 2 ?

• How long to reach a “reasonable” acceleration
measurement?
• Can we redesign to get as a faster measurement?
- Design issues:
* How fast y(t) becomes yss depends on how fast
e s1t , e s 2 t  0
* Design b/M and k/M such that s1 and s 2 are farther
in the LHP so that e s1t , e s 2 t  0 and hence
y(t)  yss fast enough
* However, faster decay  expensive components
 need to trade-off

- Question: what if the applied force F , hence a, is

time varying?