You are on page 1of 6

2.

4 Inverse Laplace Transform


Definition

f (t ) = L−1 [F ( s )]

−1
Where: L is called the Inverse Laplace Transform Operator.

How to do it?

• F(S) is simple use table of Laplace pairs


• F(s) is complicated Apply PFE first, then use table

PFE?
Partial Fraction Expansion

F ( s ) = F1 ( s ) + F2 ( s ) + LL + Fn ( s )

ZG2004 Where Fi(s) is in simple form and can be found in table 1

Steps of PFE
N (s)
F (s) =
D( s )
Steps:

1. Make the order of N(s) less than that of D(s)


2. Factor D(s) into linear factors and/or quadratic factors

N (s) N ( s)
F (s) = =
D( s) ( s + p1 )( s + p2 ) L (as 2 + bs + c) L

3 Expand out F(s) as

N (s) A1 A2 D1s + D2
F (s) = = + +L+ +L
D( s) ( s + p1 ) ( s + p2 ) (as 2 + bs + c)

4 Apply either “cover-up” or “undetermined coefficients” method to find A1,A2, C1,D1…

ZG2004 2

1
Examples of PFE

See examples on blackboard/slide first,


Then see following slides for “suggestions”

ZG2004 3

Example of “Cover-Up’
Example 1:
1
F ( s) = Find inverse Laplace
s ( s + 5)

Solution:

ZG2004 4

2
Example of “Undetermined Coefficients”
Example 2: ω2
F (s) = Find inverse Laplace
s( s + ω 2 )
2

Solution:

ZG2004 5

Suggestions for PFE (1)


Suggestions:

1 For powers, if using “cover-up”, you must have

1 A1 A2 A
= + +L+ n
(s + a )n ( s + a1 ) n ( s + a2 ) n −1 s+a

1 d11s + d12 d 21s + d 22 d n1s + d n 2


= + +L+
(as 2
+ bs + c )n
(as 2 + bs + c) n (as 2 + bs + c) n −1 (as 2 + bs + c)

2 Comsider general situation where N(s) is lower in order than D(s)

N (s) N ( s)
Case 1 F ( s) = = F(s) has distinctive poles
D( s ) ( s + p1 )( s + p2 ) L ( s + pn )
A1 A2 An
= + +L+
( s + p1 ) ( s + p2 ) ( s + pn )

Use “cover-up” to find A1-An

ZG2004 6

3
Suggestions for PFE (2)
N (s) N ( s)
Case 2 F ( s) = = F(s) has multiple poles
D( s ) ( s + p1 )( s + p2 ) n
A B1 B2 Bn
either = + + L+ Good for “cover-up”
( s + p1 ) ( s + p2 ) n ( s + p2 ) n −1 ( s + p2 )
n −1 n−2
A Bs + B2 s + L + Bn
or = + 1 Good for “undetermined coeffcients”
( s + p1 ) ( s + p2 ) n

Case 3 N (s) N ( s)
F ( s) = =
D( s ) as 2 + bs + c
N (s)
= If two real roots
( s + r1 )( s + r2 )
N ( s)
=
(s + r )2 If two identical real roots

N (s)
=
(s + σ )2 + ω 2 If two complex conjugate roots

ZG2004 7

Suggestions for PFE (3)


Case 3 (continued)
N (s) N (s)
F (s) = = If two complex conjugate roots
D( s ) as 2 + bs + c
N (s)
How to find F ( s) =
(s + σ )2 + ω 2

N (s) N (s)
F (s) = =
as 2 + bs + c ⎛ b c⎞
a⎜ s 2 + s + ⎟
⎝ a a⎠
1 N ( s) 1 N (s)
= =
a⎡ 2 b ⎛ b ⎞ c ⎛ b ⎞ ⎤ a⎛
2
⎞ ⎛ 4ac − b ⎞
2 2 2 2
⎢s + s + ⎜ ⎟ + − ⎜ ⎟ ⎥ b
⎜ s + ⎟ + ⎜⎜ ⎟

⎣⎢ a ⎝ 2a ⎠ a ⎝ 2a ⎠ ⎦⎥ ⎝ 2a ⎠ ⎝ 2a ⎠

We have assumed 4ac-b^2 >0

3 If “distinctive poles”, use “cover-up”, otherwise, “undetermined coefficients” is better

ZG2004 8

4
MATLAB-Symbolic Solution-PFE
N (s) 1
Problem: F ( s) = = Find PFE
D( s) s 2 + 2s

>> num = [0 1]
num =
0 1
N ( s) r r r
>> den = [1 2 0] F (s) = = 1 + 2 +L+ n + k
den = D( s) s − p1 s − p2 s − pn
1 2 0

>> [r, p, k] = residue (num,den)


− 0.5 0.5 1 ⎛ 1 1 ⎞
r= F (s) = + = ⎜ − ⎟
-0.50000000000000
s − (−2) s − 0 2 ⎝ s s + 2 ⎠
0.50000000000000

p= This part will be discussed in detail in CH4.3


-2
0

k=
[]
ZG2004 9

MATLAB-Symbolic Solution-Inverse Laplace


>> syms a s t w x
>> ilaplace (s/(s*(s+2)))

ans =

exp(-2*t)

>> ilaplace (a/(s*(s+2)))

ans =

a*(1/2-1/2*exp(-2*t))

>> ilaplace (a/(s^2*(s+2)^3))

ans =

a*(1/8*t+1/4*t*exp(-2*t)+1/8*t^2*exp(-2*t)-3/16+3/16*exp(-2*t))

>> ilaplace(a/(s^2+2*s))

ans =

1/4*a*4^(1/2)*(exp((-1+1/2*4^(1/2))*t)-exp((-1-1/2*4^(1/2))*t))
ZG2004 10

5
MATLAB-Symbolic Solution-Inverse Laplace
>> syms a s t w x
>> ilaplace (s/(s*(s+2)))

ans =

exp(-2*t)

>> ilaplace (a/(s*(s+2)))

ans =

a*(1/2-1/2*exp(-2*t))

>> ilaplace (a/(s^2*(s+2)^3))

ans =

a*(1/8*t+1/4*t*exp(-2*t)+1/8*t^2*exp(-2*t)-3/16+3/16*exp(-2*t))

>> ilaplace(a/(s^2+2*s))

ans =

1/4*a*4^(1/2)*(exp((-1+1/2*4^(1/2))*t)-exp((-1-1/2*4^(1/2))*t))
ZG2004 11

You might also like