Anda di halaman 1dari 38

LAPLACE TRANSFORMS RECAP

Laplace transforms replace an ODE with an


algebraic equation
Sometimes we can solve the algebraic equation easily
Then we invert the transform to get the response

From a signals & systems perspective we shall
characterize systems by their;
Transfer function instead of their impulse response
Frequency response instead of impulse response

State variable techniques will conform as well
1

MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM DEFINITION
Function of time: f(t)

Lower limit to capture !-functions etc
s is a complex variable s="+j#
The units of s are sec
-1
=Hz; a frequency
If f(t) is in pascals then F(s) is in pascal-seconds

There are technical requirements on the convergence
of the integral and on the piecewise continuity of f
Thanks to Oliver Heaviside, these shall not concern us
2

F(s) =
Z

0

f(t)e
st
dt
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM EXAMPLES
Exponential function


Note the convergence requirement Re(s)>!

Unit step/Heaviside function


Same as exponential with $=0
3

F(s) =
Z

0

e
t
e
st
dt =
Z

0

e
(s)t
dt =
e
(s)t
s

=
1
s
F(s) =
Z

0

1(t)e
st
dt =
e
st
s

=
1
s
MAE143A Signals & Systems Winter 2014
BACK OF THE BOOK LAPLACE TABLE
4

Signal Waveform Transform
impulse
step
ramp
exponential
damped ramp
sine
cosine
damped sine
damped cosine
) (t !
2 2
) ( ! "
"
+ +
+
s
s
2 2
) ( ! "
!
+ + s
2 2
!
!
+ s
2 2
! + s
s
1
s
1
2
1
s
! + s
1
2
) (
1
! + s
1(t)
t1(t)
e
!!t
1(t)
te
!!t
1(t)
sin !t
( )
1(t)
cos !t
( )
1(t)
e
!!t
sin "t
( )
1(t)
e
!!t
cos "t
( )
1(t)
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM PROPERTIES
Linearity critically important


Example

5

L{Af
1
(t) + Bf
2
(t)} = AL{f
1
(t)} + BL{f
2
(t)} = AF
1
(s) + BF
2
(s)
L[Acos t] = L

A
2

e
jt
+ e
jt

=
A
2
L(e
jt
) +
A
2
L(e
jt
)
=
A
2
1
s j
+
A
2
1
s + j
=
As
s
2
+
2
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM PROPERTIES
Integration property
Proof
6

L

Z
t
0
f() d

=
F(s)
s
L
Z
t
0
f() d

=
Z

0
Z
t
0
f() d

e
st
dt
=
Z

0
Z
t
0
f() d

d
dt

e
st
s

dt
=
e
st
s
Z
t
0
f() d

0
+
1
s
Z

0
f(t)e
st
dt
=
F(s)
s
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM PROPERTIES
Differentiation property

Proof





Second and higher derivatives
7

L

df(t)
dt

= sF(s) f(0)
L

d
2
f(t)
dt
2

= L

d
dt

df(t)
dt

= sL

df(t)
dt

df
dt
(0

)
= s
2
F(s) sf(0

) f
0
(0

)
L

df(t)
dt

=
Z

0

df(t)
dt
e
st
dt
= f(t)e
st


Z

0

f(t)

se
st

dt
= f(0

) + s
Z

0

f(t)e
st
dt
= f(0

) + sF(s)
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM PROPERTIES
General derivative formula


Time translation
An important formula to make your life easy


Remember about delayed signals and writing pulses
as successive delayed steps
8

L

d
m
f(t)
dt
m

= s
m
F(s) s
m1
f(0

) s
m2
f
0
(0

) f
(m1)
(0

)
L{f(t a)1(t a)} = e
as
F(s) for a > 0
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM PROPERTIES
Final value property
Used very frequently, especially in control systems
One needs to be careful to ensure that the final value
really exists often get the average value
No good with cos and sin, unstable, etc


Initial value property
Does not work with impulses etc
9

lim
t
f(t) = lim
s0
sF(s)
lim
t0
+
f(t) = lim
s
sF(s)
MAE143A Signals & Systems Winter 2014
RATIONAL FUNCTIONS
Most of our problems will yield Laplace transforms
which are rational functions



p
i
are the poles (which might be repeated)
z
i
are the zeros (which might be repeated)
K is the gain or scale factor
A proper rational function has n"m
A strictly proper rational function has n>m
Otherwise it is improper
1
0

F(s) =
b
m
s
m
+ b
m1
s
m1
+ + b
1
s + b
0
a
n
s
n
+ a
n1
s
n1
+ + a
1
s + a
0
= K
(s z
1
)(s z
2
) . . . (s z
m
)
(s p
1
)(s p
2
) . . . (s p
n
)
MAE143A Signals & Systems Winter 2014
PARTIAL FRACTION EXPANSION
We can write the rational function as




This is the partial fraction expansion
The !
ij
are residues at the poles p
i
of various order
The residues are given by a (possibly nasty) formula
1
1

F(s) =
b
m
s
m
+ b
m1
s
m1
+ + b
1
s + b
0
a
n
s
n
+ a
n1
s
n1
+ + a
1
s + a
0
= K
(s z
1
)(s z
2
) . . . (s z
m
)
(s p
1
)(s p
2
) . . . (s p
n
)
=

1
s p
1
+

2
s p
2
+

31
s p
3
+

32
(s p
3
)
2
+

33
(s p
3
)
3
+ +

q
s p
q

1
= lim
sp
1
{(s p
1
)F(s)}

3j
=
1
(3 j)!
d
3j
ds
3j

(s p
3
)
3
F(s)

sp
3
MAE143A Signals & Systems Winter 2014
PARTIAL FRACTION EXAMPLE
Find the inverse Laplace transform of

Find the poles and residues


1
2

F(s) =
20(s + 3)
(s + 1)(s
2
+ 2s + 5)
F(s) =

1
s + 1
+

2
s + 1 j2
+

2
s + 1 + j2

1
= lim
s1
(s + 1)F(s) =
20(s + 3)
s
2
+ 2s + 5

s=1
= 10

2
= lim
s1+j2
(s + 1 2j)F(s) =
20(s + 3)
(s + 1)(s + 1 + j2)

s=1+j2
= 5 5j = 5

2e
j5/4
MAE143A Signals & Systems Winter 2014
PARTIAL FRACTION EXAMPLE CONTD
Determine the pole these give the exponential modes
Determine the residues relative presence in solution
Real rational functions have real inverse transforms
Multiple poles yield terms like
Matlab has a residue function
1
3

F(s) =
20(s + 3)
(s + 1)(s
2
+ 2s + 5)
=
10
s + 1
+
5

2e
j5/4
s + 1 j2
+
5

2e
j5/4
s + 1 + j2
f(t) =
h
10e
t
+ 5

2e
(1+j2)t+j5/4
+ 5

2e
(1j2)tj5/4
i
1(t)
=
h
10e
t
+ 10

2e
t
cos (2t + 5/4)
i
1(t)
t
k
e
at
1(t)
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORMS AND
CONVOLUTION
Yeah! No more convolution!
But wait
H(s) must be something special the transfer function
The Laplace transform of the impulse response
1
4

L
Z

h()u(t ) d

=
Z

0

h()u(t ) de
st
dt
=
Z

0

h()e
s
u(t )e
s(t)
d dt
=
Z

h()e
s
Z

u()e
s
d d
=
Z

h()e
s
Z

0
u()e
s
d d
= H(s)U(s)
= t , d = dt
t = , =
t = 0, =
MAE143A Signals & Systems Winter 2014
OUR FAVORITE SYSTEM EXAMPLE
Take Laplace transforms


Solve the algebraic equation for Y(s)



Expand as a partial fraction

1
5

d
2
y
dt
2
+ 5
dy
dt
+ 6y = 3 cos(5t), y(0) = 1, y
0
(0) = 0
s
2
Y (s) sy(0) y
0
(0) + 5sY (s) 5y(0) + 6Y (s) =
3s
s
2
+ 25
(s
2
+ 5s + 6)Y (s) s 5 =
3s
(s + j5)(s j5)
Y (s) =
s + 5
s
2
+ 5s + 6
+
3s
(s
2
+ 5s + 6)(s + j5)(s j5)
=
s + 5
(s + 2)(s + 3)
+
3s
(s + 2)(s + 3)(s + j5)(s j5)
Y (s) =
a
s + 2
+
b
s + 3
+
c
s + j5
+
c

s j5
MAE143A Signals & Systems Winter 2014
LAPLACE TRANSFORM EXAMPLE
Residue formula
1
6

Y (s) =
s + 5
(s + 2)(s + 3)
+
3s
(s + 2)(s + 3)(s + j5)(s j5)
=
a
s + 2
+
b
s + 3
+
c
s + j5
+
c

s j5
a = lim
s2
(s + 2)Y (s) =
3
1
+
6
29
=
81
29
b = lim
s3
(s + 3)Y (s) =
2
1
+
9
1 34
=
59
34
c = lim
sj5
(s + j5)Y (s) =
j15
(2 j5)(3 j5) j10
=
3/2
19 j25
MAE143A Signals & Systems Winter 2014
SHOWING OFF YES!
The matlab solution

Are these the same?
1
7

Y (s) =
81
29
1
s + 2

59
34
1
s + 3
+
3/2
19 j25
1
s + j5
+
3/2
19 + j25
1
s j5
y(t) =

81
29
e
2t

59
34
e
3t
+
3/2
19 j25
e
j5t
+
3/2
19 + j25
e
j5t

1(t)
y(t) =

81
29
e
2t

59
34
e
3t

51
986
cos 5t +
75
986
sin 5t

1(t)
19 j25 =
p
19
2
+ 25
2
exp [j atan2(25, 19)]
=

986 exp

+ tan
1
25
19

19 + j25 =

986 exp

+ tan
1
25
19

MAE143A Signals & Systems Winter 2014


SHOWING OFF .. OH NO!
A little trig
1
8

25
19
%(986)
&
3/2
19 j25
e
j5t
+
3/2
19 + j25
e
j5t
=
3

986
1
2
h
e
j(5t++tan
1 25
19
)
i
+
3

986
1
2
h
e
j(5t++tan
1 25
19
)
i
=
3

986
cos

5t + + tan
1
25
19

=
3

986
cos

5t + tan
1
25
19

=
3

986
cos(5t) cos

tan
1
25
19

+
3

986
sin(5t) sin

tan
1
25
19

= tan
1

25
19

cos

tan
1

25
19

=
19

986
sin

tan
1

25
19

=
25

986
3/2
19 j25
e
j5t
+
3/2
19 + j25
e
j5t
=
51
986
cos(5t) +
75
986
sin(5t)
MAE143A Signals & Systems Winter 2014
INVERSE LAPLACE TRANSFORMS
The Bromwich integral

We need help from Cauchy to do this integral
Tables inside the back cover of the book
In undergraduate studies, we shall see only rational
functions poles and zeros
Use the residue formulas
Main idea simple pole at s=$ yields mode e
$t
Pole of multiplicity n at s=$ yields mode t
n-1
e
$t
Stability is tied to poles only in left half of s-plane

1
9

f(t) =
1
2j
Z
c+j
cj
X(s)e
st
ds
MAE143A Signals & Systems Winter 2014
SYSTEMS IDEAS
First-order system RC


Apply Laplace transforms
2
0

R
u(t)=i(t) C v
C
(t)=y(t)
i
R
(t)
i
C
(t)
+
-
dy(t)
dt
+
1
RC
y(t) =
1
C
u(t), y(0)
Y (s) = L[y(t)], U(s) = L[u(t)]
sY (s) y(0

) +
1
RC
Y (s) =
1
C
U(s)

s +
1
RC

Y (s) = y(0

) +
1
C
U(s)
Y (s) =
1
s +
1
RC
y(0

) +
1
C
1
s +
1
RC
U(s)
y(t) = e
t/RC
y(0) +
1
C
Z
t
0
e
(t)/RC
u() d
MAE143A Signals & Systems Winter 2014
SYSTEMS IDEAS CONTINUED
Laplace transform domain




System transfer function

System impulse response
For zero state
2
1

Y (s) =
1
s +
1
RC
y(0

) +
1
C
1
s +
1
RC
U(s)
Zero-input response Zero-state response
y(t) = e
t/RC
y(0) +
1
C
Z
t
0
e
(t)/RC
u() d
H(s) =
1
C
1
s +
1
RC
Y (s) = H(s)U(s), y(t) = h(t) u(t)
h(t) =
1
C
e

t
RC
1(t)
MAE143A Signals & Systems Winter 2014
SYSTEMS IDEAS CONTINUED
Characterizing a systems response to an arbitrary
input u(t) can be performed using the transfer
function or the impulse response
The two pieces of system information are
interchangeable
The initial condition response has the same
exponential modes as the impulse response
The initial condition transform has the same poles as
the transfer function
Stability is tied to poles/modes
2
2

Y (s) = H(s)U(s), y(t) = h(t) u(t), H(s) = L[h(t)]
MAE143A Signals & Systems Winter 2014
A SECOND-ORDER EXAMPLE (6.29)
Consider the input-output ODE

Laplace transforms



Transfer function

Lets look at the step response
2
3

d
2
y(t)
dt
2
+ 6
dy(t)
dt
+ 8y(t) = 2u(t)
s
2
Y (s) sy(0

) y
0
(0

) + 6sY (s) 6y(0

) + 8Y (s) = 2U(s)
(s
2
+ 6s + 8)Y (s) = sy(0

) + y
0
(0

) + 6y(0

) + 2U(s)
Y (s) =
sy(0

) + y
0
(0

) + 6y(0

)
s
2
+ 6s + 8
+
2
s
2
+ 6s + 8
U(s)
H(s) =
2
s
2
+ 6s + 8
MAE143A Signals & Systems Winter 2014
EXAMPLE 6.29
Transfer function
What is the step response?
Step response; u(t)=1(t) and zero i.c.s



Partial fraction expansion
2
4

H(s) =
2
s
2
+ 6s + 8
u(t) = 1(t) U(s) =
1
s
Y (s) = H(s)U(s) =
2
s
2
+ 6s + 8

1
s
=
2
s(s + 4)(s + 2)
Y (s) =
2
s(s + 4)(s + 2)
=
a
s
+
b
s + 2
+
c
s + 4
=
0.25
s
+
2.5
s + 2

1.75
s + 4
y(t) =

0.25 + 2.5e
2t
1.75e
4t

1(t)
MAE143A Signals & Systems Winter 2014
EXAMPLE 6.30
System with transfer function
Input with zero i.c.s
Output transform
2
5

H(s) =
s
2
+ 2s + 16
s
3
+ 4s
2
+ 8s
u(t) = e
2t
1(t),

U(s) =
1
s + 2

Y (s) =
s
2
+ 2s + 16
(s
3
+ 4s
2
+ 8s)(s + 2)
=
s
2
+ 2s + 16
[(s + 2)
2
+ 4]s(s + 2)
=
s + 1
(s + 2)
2
+ 4
+
1
s
+
2
s + 2
=
s + 2
(s + 2)
2
+ 4

1
(s + 2)
2
+ 4
+
1
s
+
2
s + 2
y(t) =

e
2t
cos 2t
1
2
e
2t
sin 2t + 1 2e
2t

1(t)
MAE143A Signals & Systems Winter 2014
OUR FAVORITE EXAMPLE AGAIN
Laplace transforms



Transfer function
Stable poles in left half s-plane
Sinusoidal steady-state solution
What part of the output signal remains as
A sinusoid at frequency 5 rad/s
2
6

d
2
y
dt
2
+ 5
dy
dt
+ 6y = 3 cos(5t), y(0) = 1, y
0
(0) = 0
Y (s) =
s + 5
(s + 2)(s + 3)
+
3s
(s + 2)(s + 3)(s + j5)(s j5)
=
a
s + 2
+
b
s + 3
+
c
s + j5
+
c

s j5
H(s) =
1
s
2
+ 5s + 6
t
MAE143A Signals & Systems Winter 2014
STILL OUR FAVORITE EXAMPLE
For sinusoidal steady-state we need to know c

2
7

Y (s) =
s + 5
(s + 2)(s + 3)
+
3s
(s + 2)(s + 3)(s + j5)(s j5)
=
a
s + 2
+
b
s + 3
+
c
s + j5
+
c

s j5
c = lim
sj5
(s + j5)Y (s) = lim
sj5
(s + j5)H(s)U(s)
= lim
sj5
(s + j5)H(s)
3s
s
2
+ 25
= lim
sj5
(s + j5)H(s)
3s
(s + j5)(s j5)
= H(j5)
15j
10j
= H(j5)
3
2
MAE143A Signals & Systems Winter 2014
EVEN MORE OF OUR FAVORITE E.G.
Sinusoidal steady-state solution
2
8

Y (s) =
s + 5
(s + 2)(s + 3)
+
3s
(s + 2)(s + 3)(s + j5)(s j5)
=
a
s + 2
+
b
s + 3
+
c
s + j5
+
c

s j5
c = H(j5)
3
2
, c

= H(j5)
3
2
y
ss
(t) = ce
j5t
+ c

e
j5t
=
3
2

H(j5)e
j5t
+ H(j5)

e
j5t

= |H(j5)|
3
2
h
e
jH(j5)
e
j5t
+ e
jH(j5)
e
j5t
i
= |H(j5)|
3
2
h
e
jH(j5)+j5t
+ e
jH(j5)j5t
i
= |H(j5)|3 cos [5t +H(j5)]
MAE143A Signals & Systems Winter 2014
SINUSOIDAL STEADY-STATE SOLUTION
System has stable transfer function
All the poles in the left half s-plane
Input is a sinusoid at frequency !
Sinusoidal steady-state response is

A magnitude change by
And a phase change by
The quantity
is called the frequency response of the system
2
9

H(s)
A cos(t + )
y
ss
(t) = A|H(j)| cos(t + +H(j))
H(j)
|H(j)|
H(j) = H(s)|
s=j
MAE143A Signals & Systems Winter 2014
FREQUENCY RESPONSE
Knowing the transfer function of a system is the
same as knowing the frequency response is the
same as knowing the impulse response
The frequency response only makes sense for
stable systems
The zero-input response tends to zero
The zero-state response tends to the sinusoidal
steady state
Funny thing: even for unstable system we can define
the frequency response as H(j#)
3
0

H(j) = H(s)|
s=j
MAE143A Signals & Systems Winter 2014
FREQUENCY RESPONSE CONTINUED
Applies to stable systems with transfer function H(s)
Steady-state sinusoidal response to Acos(!t+") is

We have to wait for the transients to die away
Two pieces; magnitude and phase


The joint plots log
10
! versus log
10
and
is called the Bode Diagram after Henrik Bode
3
1

y
ss
(t) = A|H(j)| cos(t + +H(j))
|H(j)| H(j)
H(j) = |H(j)|e
jH(j)
H(j) |H(j)|
MAE143A Signals & Systems Winter 2014
BODE DIAGRAM
RC circuit
3
2

h(t) =
1
C
e
t/RC
1(t)
H(s) =
1/C
s + 1/RC
H(j) =
1/C
j + 1/RC
|H(j)| =
1/C
p

2
+ 1/(RC)
2
H(j) = tan
1
(/RC)
Impulse response
Transfer function
Frequency response
Magnitude (frequency) response
Phase (frequency) response
MAE143A Signals & Systems Winter 2014
BODE DIAGRAM FOR RC CIRCUIT
>> R=100;C=1e-6;
>> H=tf(1/C,[1 1/R/C])
Transfer function:
1e06
---------
s + 10000
>> bode(H)
>> print -dpdf bodeRC.pdf
3
3

MAE143A Signals & Systems Winter 2014
BODE DIAGRAM FOR RC CIRCUIT
3
4

0
5
10
15
20
25
30
35
40
M
a
g
n
i
t
u
d
e

(
d
B
)
10
2
10
3
10
4
10
5
10
6
!90
!45
0
P
h
a
s
e

(
d
e
g
)
Bode Diagram
Frequency (rad/s)
MAE143A Signals & Systems Winter 2014
BODE DIAGRAM FOR RC CIRCUIT
A low-pass system
Low frequency inputs
get through
High frequency inputs
are attenuated

The magnitude is
measured in decibels
3
5

0
5
10
15
20
25
30
35
40
M
a
g
n
i
t
u
d
e

(
d
B
)
10
2
10
3
10
4
10
5
10
6
!90
!45
0
P
h
a
s
e

(
d
e
g
)
Bode Diagram
Frequency (rad/s)
|H(j)|
dB
= 20 log
10
|H(j)|
MAE143A Signals & Systems Winter 2014
STATE-SPACE AND LAPLACE
Remember calculus and matrices
Element-by-element differentiation or integration
The Laplace transform of a vector is the vector of
Laplace transforms
It is linear
3
6

x(t) = Ax(t) +Bu(t)
y(t) = Cx(t) +Du(t)

sX(s) x(0

) = AX(s) +BU(s)
Y (s) = CX(s) +DU(s)
[sI A]X(s) = x(0

) +BU(s)
X(s) = [sI A]
1
x(0

) + [sI A]
1
BU(s)
Y (s) = C[sI A]
1
x(0

) +

D +C[sI A]
1
B

U(s)
Zero-input response Zero-state response
MAE143A Signals & Systems Winter 2014
STATE-SPACE AND LAPLACE
>> H=tf(1,[1 5 6])
Transfer function:
1
-------------
s^2 + 5 s + 6
>> [A,B,C,D]=ssdata(H)
A =
-5 -3
2 0
B =
0.5000
0
C =
0 1
D =
0
>> syms s, t
>> D+C*inv(s*eye(2)-A)*B
ans =
1/(s^2 + 5*s + 6)
>> expm(A*t)
ans =
[ 3/exp(3*t) - 2/exp(2*t), 3/exp(3*t) - 3/exp(2*t)]
[ 2/exp(2*t) - 2/exp(3*t), 3/exp(2*t) - 2/exp(3*t)]
>> inv(s*eye(2)-A)
ans =
[ s/(s^2 + 5*s + 6), -3/(s^2 + 5*s + 6)]
[ 2/(s^2 + 5*s + 6), (s + 5)/(s^2 + 5*s + 6)]
>> ilaplace(inv(s*eye(2)-A))
ans =
[ 3/exp(3*t) - 2/exp(2*t), 3/exp(3*t) - 3/exp(2*t)]
[ 2/exp(2*t) - 2/exp(3*t), 3/exp(2*t) - 2/exp(3*t)]
3
7

MAE143A Signals & Systems Winter 2014
END OF CONTINUOUS-TIME SIGNALS
& SYSTEMS
Physical, chemical, electrical models ODEs and ics
State-space and n-vector first-order ODEs and ics
Zero-input response + zero-state response
Impulse response, step response, convolution
A systems zero-state response is captured entirely by its
impulse response
Laplace transforms: zero-input + zero-state responses
Transfer function and zero-state response
Frequency response and sinusoidal steady-state
State variables and computational tools - matlab

3
8

MAE143A Signals & Systems Winter 2014

Anda mungkin juga menyukai