Department of Mechanical Engineering

EMT2101/ AEN2104: ENGINEERING MATHEMATICS III

Course objectives
o To introduce students to integral transformations and their application to the solution
of Ordinary Differential Equations
o Introduce the power series solution technique to Ordinary Differential Equations
o To develop problem solving skills and proof skills by working on specific problems
o To focus on analytical techniques for the classical linear PDE of physics and
engineering (heat, wave and Laplace equations), and their frequent occurrence in
applications.
Course outline
1. Fourier and Laplace transforms
2. Power series solution to ODEs
3. Gamma and Beta functions
4. Bessel functions
5. Legendre functions
6. Applied Statistics
7. Applied Probability
Reference text books
1. Advanced Engineering mathematics by Stroud
2. Calculus and analytical geometry by Edwards
3. Advanced Engineering mathematics by Wylie
4. Advanced engineering mathematics by Kreyszig
5. Advanced Engineering mathematics by Bajpai
Course assessment
Three course work assignments with an assignment at the end of every pair of chapters
and two tests with a test at the end of three chapters.

o Course work: 15%

o Continuous assessment tests: 25%
o Final examination: 60%

1 EM
 THE LAPLACE TRANSFORM

2 EM
 Department of Mechanical Engineering

INTRODUCTION

The Laplace Transformation is a powerful method for solving linear differential

equations with constant coefficients to investigate either the transient or the steady-state
behaviour of physical systems some arising in engineering mathematics, especially if a
particular solution satisfying given initial conditions is required. It consists essentially of
three steps;

1. The given differential equation is transformed into an algebraic equation

2. Then the latter is solved by purely algebraic manipulations
3. Finally, the solution of the algebraic equation is transformed back so that it
becomes a required solution of the original differential equation.

In this way the laplace transformation reduces the problem of solving a differential
equation to an algebraic problem. Relatively simple operations on the transform
correspond to more complicated operations on the function being transformed.

DEFINITIONS AND BASIC THEORY

The Laplace transformation is an operation, denoted by the symbol L, which associates

with each function f (t ) satisfying suitable conditions for t  0, a unique function F (s )
called the Laplace transform of f (t ) , according to the rule

L  f (t ) = F ( s) =  f (t )e − st dt
0

if the integral exists.

Furthermore the original function f (t ) is called the inverse transform or inverse of

F (s ) and will be denoted by L -1 (F ) ; that is

f (t ) = L -1 (F )

Example 1. Find f (t ) = k when t>0

 
k
L (f) = L (k) =  ke dt = − e − st
− st

0
s 0

Hence when s>0,

L (1) =
k
s
Example 2. What is L (cos at ) ?

3 EM
  
e − st (− s cos at + a sin at)
L (cos at ) =  cos ate dt = − st

0
s2 + a2 0
s
= 2 s0
s + a2

Conversely, we can also write

 s 
cos at = L−1  2 2
s + a 

Example 3. Let f(t) = e at when t>0, where a is a constant. Then

 
1 −( s − a )t
L ( e )=  e e dt =
at at − st
e ;
0
a−s 0

Hence, when s-a >0,

1
L ( e at )=
s−a

INVERSE LAPLACE TRANSFORMS

We have noted that given a function f (t ) , we can find the Laplace transform F ( s ). Now
we would like to find f (t ) , given F ( s ).

For instance, knowing that

s
L cos at =
,
s + a2 2

it follows from the uniqueness property of Laplace transform that the function f (t ) of t ,
s
whose Laplace transform is 2 must be cos at. In other words,
s + a2
 s 
L−1  2 2 
= cos at.
s + a 

The symbol L−1 denotes the Inverse Laplace Transform Operator. Below we tabulate a
few transforms with the corresponding inverses.

4 EM
F (s) f (t ) F (s) f (t )
1 1 a a
s s
1 t 1 e − at
s2 s+a
1 e at n! tn
s−a s n +1
1 sin at 1 t n −1
s + a2
2
a sn (n − 1)!
s cos at s cosh at
s + a2
2
s − a2
2

Theorem 1 (Linearity property) The Laplace transformation is a linear operation, that

is, for any functions f(t) and g(t) whose Laplace transforms exist and any constants a and
b we have

Laf (t ) + bg(t ) = aL ( f ) + bL (g )

Proof. By definition,

Laf (t ) + bg(t )=  e − st af (t ) + bg (t )dt

0
 
= a e − st
f (t )dt + b  e − st g (t )dt = aL ( f ) + bL (g )
0 0

Example 4. Since f(t) = cosh at =

(e at
+ e− at ) . Then, using Theorem 1 and the result of
2
Example 3, we find

1 1 1 
L (cosh at ) =
1 1
L(e at ) + L(e − at ) =  + 
2 2 2s−a s+a
That is, when s>a,

L (cosh at ) =
s
s − a2
2

5 EM
Example 5. Prove that L (t n ) =
n!
s n +1
We begin by briefly introducing a special function which will enable us do the proof.
This is the gamma function which is defined as


(x ) =  t x −1e −t dt ,
0
Using several values of x:

(1) =  e −t dt = −e −t

=1
0
0

 
(2) =  te −t dt = −t e −t

+  e −t dt =1
0
0 0

And more generally,

 
(n + 1) =  t n e −t dt =− t n e −t +  (n )t n −1e −t dt


0
0 0


= 0 − nt n −1e −t + n(n − 1)  t n − 2 e −t dt
0
0

Since n  0 , the last term decreases to n=0 with progressive integrations yielding

(n + 1) = n(n − 1)(n − 2)...1 e −t dt = n(n − 1)(n − 2)...1 = n!

0

Thus,
(n + 1) = n! (1)

We can therefore proceed with the Laplace transformation as;

L ( t ) =  t n e − st dt
n

Let p = st , dp = sdt and the limits remain the same such that

 n
 p dp
L ( t )=    e − p
n

0
s s

1
p e
n −p
= dp
n +1
(2)
s 0
The improper integral above can be related to the Gamma function. One has to put into
consideration that p is only a dummy variable and that

6 EM
 

 p e dp can be written as  t e dt without error later on allow close comparison from

n −p n −t

0 0

which its quickly visible that x = n+1. Hence from (2)

p
n
e − p dp = (n + 1)
0

And,

(n + 1) n!
L ( t n )= n +1
= n +1 , using (1)
s s

Theorem 2 (First shifting theorem)

If L−1 F ( s) = f (t ) , then L−1 F (s − a) = eat f (t ) .
 1  1 −1  1  e 2t
Thus, for instance, since L−1  2  = sin 3t , we have L  = sin 3t .
s + 9 3  ( s − 2) + 9  3
2

Theorem 3(change of scale)

1 t
If L−1 F ( s) = f (t ) , then L−1  F (cs ) = f   .
c c
 1 
Thus, for instance, since L−1  2  = sin t , there follows that
 s + 1
 1  −1  1  1 t
L−1  2 =L   = sin .
 4s + 1  (2s) + 1  2
2
2

Theorem 4 ( Inverse Laplace transform of derivatives)

 dn 
If L−1 F ( s) = f (t ) , then L−1  n F ( s )  = (−1) n t n f ( t ) .
 ds 
1
Thus, for instance, since L−1  2  = t , there follows that
s 
 d  1   −2  2  2! 
L−1   2   = L−1  3  = −L−1  3  = −L−1  2+1  = (−1)t  t = −t 2 .
 ds  s   s  s  s 

It is easy to find the inverse Laplace transforms of standard tabulated transforms.

However, in many instances are not either in their standard forms or easily recognised.
 s−4 
For instance, it is not immediately obvious as to what L−1   is. In such
 ( s − 2)( s + 3) 
situations one needs to transform the given transform into simpler and easily recognisable
partial fractions.

7 EM
PARTIAL FRACTIONS

Rules of Partial Fractions

1. The numerator must be of lower degree than the denominator. If it is not then
we must first divide it out and then manipulate the remainder.
2. Factorise the denominator into its prime factors. These factors will determine
the shapes or natures of the partial fractions.
A
3. A linear factor ( s + a ) gives a partial fraction , where A is a constant to
s+a
be determined.
A B
4. A repeated factor ( s + a ) 2 gives + .
s + a (s + a)2
A B C
5. Similarly, ( s + a )3 gives + + .
s + a ( s + a) 2
( s + a )3
Ps + Q
6. A quadratic factor ( s 2 + ps + q) gives a partial fraction 2 .
s + ps + q
7. Similarly, repeated quadratic factor ( s 2 + ps + q)2 gives a partial fraction of
Ps + Q Rs + T
the form 2 + 2 .
s + ps + q ( s + ps + q ) 2

Examples:
 5s 2 + 14s + 2 
1. Find L−1  2 
;
 ( s − 2)( s + 3) 
We note that
5s 2 + 14s + 2 A B C
= + + .
( s − 2)( s + 3) 2
( s − 2) ( s + 3) ( s + 3) 2
On putting the right hand on the same denominator, and equating the numerators,
we find that
5s 2 + 14s + 2 = A( s + 3)2 + B( s − 2)( s + 3) + C ( s − 2) ,
where upon A = 2, B = 3 and C = −1. Therefore
 5s 2 + 14 s + 2  −1  2  −1  3  −1  1 
L−1   = L   + L  - L  2 
 ( s − 2)( s + 3) 2
  ( s − 2)   ( s + 3)   ( s + 3) 
= 2e 2t + 3e −3t − te −3t .

 s 2 + 3s − 7 
2. Find L−1  .
 ( s − 1)( s + 2) 
2

We note that

8 EM
 s 2 + 3s − 7 A Bs + C
= + 2 .
( s − 1)( s + 2) ( s − 1) ( s + 2)
2

On putting the right hand on the same denominator, equating the numerators, we
find that
s 2 + 2s − 7 = A( s 2 + 2) + Bs( s − 1) + C (s − 1) ,
whereupon A = −1, B = 2 and C = 5. Therefore
 s 2 + 3s − 7  −1  −1  −1  2s  −1  5 
L−1  =L  +L  2 +L  2 
 ( s − 1)( s + 2)   ( s − 1)   ( s + 2)   ( s + 2) 
2

5
= −et + 2 cos 2t + sin 2t.
2

Exercises: Express the following transforms in partial fractions and find the
corresponding inverse Laplace transforms:
7s − 5 2s 2 + 1 s 2 + 5s + 3 s2
i) ; ii) ; iii) ; iv)
( s − 1)( s − 2)( s − 3) s ( s + 1) 2 s 2 + 7 s + 10 ( s + 2)3
s +1
v) 2 .
( s + 7 s + 2) 2

Other Specialised Expressions:

 6s − 4 
3. Find L−1  2 
 s − 4 s + 20 
Note that we can first express the denominator as squares and thus recast the
transform as
6s − 4 6s − 4 6s − 4
= = .
s − 4 s + 20 ( s − 2) + 16 ( s − 2) 2 + 4 2
2 2

Moreover, the numerator can be re-expressed as 6( s − 2) + 8. Therefore,

 6s − 4  - −1  6( s − 2)   8 
L- −1  2 =L  2 
+ L- −1  2 
 s − 4 s + 20   ( s − 2) + 4   ( s − 2) + 4 
2 2

 ( s − 2)   4 
= 6L- −1  2 
+ 2L- −1  2 
 ( s − 2) + 4   ( s − 2) + 4 
2 2

= 6e 2t cos 4t + 2e 2t sin 4t.

 2s − 8 
4. Find L−1  2 .
 ( s − 8s + 15) 
Note that this is a straightforward case, since the degree of the numerator is lower
than that of the denominator! Hence

9 EM
 2s − 8 A Bs + C
= + .
s − 8s + 15 ( s − 3) ( s − 5)
2

Again on putting the right hand on the same denominator, equating the
numerators, we find that
2s − 8 = A( s − 5) + B( s − 3) ,
whereupon A = 1 and B = 1. Therefore
 2s − 8  −1  1  −1  1 
L−1  2 =L  +L  =e +e .
3t 5t

 s + −8s + 15   ( s − 3)   (s − 5 

 5s − 8 
5. Find L−1  .
 s ( s − 4) 
Note that this is also a straightforward case, since the degree of the numerator is
lower than that of the denominator! Hence
5s − 8 A B
= + .
s( s − 4) s ( s − 4)
Once again, putting the right hand on the same denominator, equating the
numerators, we find that
5s − 8 = A( s − 4) + Bs ,
where upon we find that A = 2 and B = 3. Therefore
 5s − 8  −1  2  −1  3 
L−1   = L  +L   = 2 + 3e .
4t

 s ( s − 4)  s s − 4

 3s + 7 
6. Find L−1  2 .
 s − 2s − 3 
Note that we can give a shift of 1 to s and thus recast the transform as
3s + 7 3( s − 1) + 10 3( s − 1) + 10 ( s − 1) 2
= = =3 +5 .
s − 2 s − 3 ( s − 1) − 4 ( s − 1) − 2
2 2 2 2
( s − 1) − 2
2 2
( s − 1) 2 − 2 2
Hence
 3s + 7  −1  ( s − 1)   2 
L−1  2  = 3L  2 
+ 5L−1  2 
 s − 2s − 3   ( s − 1) − 2   ( s − 1) − 2  .
2 2

= 3et cosh 2t + 5et sinh 2t

Special Techniques for Evaluating Coefficients in Partial Fractions

So far we have used classical techniques in evaluating the constants in the partial
fractions. There are, however, other shorts to the evaluation of these coefficients at the
poles, that is, at those values of s for which the denominators of the partial fractions
varnish. From example, consider the transform
5s − 8 A B
= +
s( s − 4) s s − 4

10 EM
Multiplying through by s and taking the limits as s → 0 , we have
5s − 8 5s − 8 As Bs
lim s  = lim = 2 = lim + lim = A.
s →0 s( s − 4) s →0 ( s − 4) s →0 s s →0 s − 4

Likewise, multiplying by s − 4 and evaluating the limits as s → 4, we have

5s − 8 5s − 8 A( s − 4)
lim( s − 4)  = lim = 3 = lim + B = B.
s →4 s( s − 4) s →4 s s →0 s
This is called the “Cover-Up Rule”.

Hence
5s − 8 2 3
= +
s( s − 4) s s − 4
and
 5s − 8  −1  2  −1  3 
L−1   = L  +L   = 2 + 3e .
4t

 s ( s − 4)   
s  s − 4 
Using the same “cover-up” rule, we can express transforms in forms that we can
recognise as standard ones and thus easily generate the corresponding inverse transforms.

−1  2s 2 − 4 
Example 1: Find L  
 ( s − 3)( s − 2)( s + 1) 
Expressing the transform in partial fractions as
2s 2 − 4 A B C
= + +
( s − 3)( s − 2)( s + 1) s − 3 s − 2 s + 1
Using the “cover-up rule”, we have
2s 2 − 4 2s 2 − 4 14 7
A = lim( s − 3)  = lim = = ,
s →3 ( s − 3)( s − 2)( s + 1) s →3 ( s − 2)( s + 1) 1  4 2
2s 2 − 4 2s 2 − 4 4 4
B = lim( s − 2)  = lim = =− ,
s →2 ( s − 3)( s − 2)( s + 1) s →2 (s − 3)(s + 1) (−1)  3 3
2s − 4
2
2s − 4
2
−2 1
C = lim( s + 1)  = lim = =− .
s →−1 ( s − 3)( s − 2)( s + 1) s →3 ( s − 2)( s − 3) (−3)  (−4) 6
Therefore,
2s 2 − 4 7 1 4 1 1 1
= − − ,
( s − 3)( s − 2)( s + 1) 2 s − 3 3 s − 2 6 s + 1
and
−1  2s 2 − 4  7 −1  1  4 −1  1  1 −1  1 
L  = L  − L  − L  
 ( s − 3)( s − 2)( s + 1)  2  s − 3 3 s − 2 6  s + 1 ,
7 4 1
= e 3t − e 2 t − e − t .
2 3 6

11 EM
  2 − 11s 
Example 2: Find L−1  
 ( s − 2)( s + 2s + 2) 
2

Expressing the transform in partial fractions as

2 − 11s A Bs + C
= + 2
( s − 2)( s + 2 s + 2) s − 2 s + 2 s + 2
2

Using the “cover-up rule” for the simple pole at s = 2 , we have

2 − 11s 2 − 11s −20
A = lim( s − 2)  = lim 2 = = −2.
s →2 ( s − 2)( s + 2 s + 2) s →2 ( s + 2 s + 2) 10
2

With this result, the transform can then be recast as

2 − 11s −2 Bs + C
= + 2
( s − 2)( s + 2 s + 2) s − 2 s + 2 s + 2
2

Putting the right-hand terms over the same denominator, we have

2 − 11s −2( s 2 + 2 s + 2) + Bs( s − 2) + C ( s − 2)
= .
( s − 2)( s 2 + 2s + 2) ( s − 2)( s 2 + 2s + 2)
Equating the numerators, we have
2 − 11s = −2( s 2 + 2s + 2) + Bs( s − 2) + C ( s − 2),
or
2 − 11s = −2s 2 − 4s − 4 + Bs 2 − 2 Bs + Cs − 2C = ( B − 2)s 2 + (C − 4 − 2B)s − (4 + 2C ).
Equating coefficients of the various powers of s , we have
B − 2 = 0,
C − 4 − 2 B = −11,
4 + 2C = −2.
Solving these expressions, we have
B = 2, C = −3.
Therefore,
2 − 11s −2 2s − 3 −2 2( s + 1) − 5
= + 2 = + ,
( s − 2)( s + 2 s + 2) s − 2 s + 2 s + 2 s − 2 ( s + 1) 2 + 1
2

and
 2 − 11s  −1  1  −1  ( s + 1)  −1  1 
L−1   = −2 L   + 2L   − 5L  
 ( s − 2)( s + 2 s + 2)  s − 2  ( s + 1) + 1   ( s + 1) + 1  ,
2 2 2

= −2e 2t + 2e − t cos t − 5e − t sin t.

In case of repeated simple poles we can use the “cover-up rule” also. For example, for the
s+3
transform F ( s ) = , we may first of all recast it as
( s − 2)( s + 1) 2
s+3 1 s+3 1  A B 
= = +
( s − 2)( s + 1) 2 
s + 1 ( s − 2)( s + 1) s + 1  ( s + 1) ( s − 2) 
where
s+3 s+3 2
A = lim( s + 1)  = lim =−
s →−1 ( s − 2)( s + 1) s →−1 ( s − 2) 3

12 EM
and
s+3 s+3 5
B = lim( s − 2)  = lim =
s →2 ( s − 2)( s + 1) s → 2 ( s + 1) 3
With these coefficients, the transform then becomes
s+3 5 1 2 1
= −
( s − 2)( s + 1) 2
3 ( s + 1)( s − 2) 3 ( s + 1) 2
Further still, the first partial fraction on the right can be expressed as
1 C D
= +
( s + 1)( s − 2) s + 1 s − 2
where upon we have
1 1 1
C = lim( s + 1)  = lim =− ,
s →−1 ( s + 1)( s − 2) s →−1 s − 2 3
1 1 1
D = lim( s − 2)  = lim = ,
s →2 ( s + 1)( s − 2) s →2 s + 1 3
With the latter results, the whole transform them becomes
s+3 5 1 5 1 2 1
= − −
( s − 2)( s + 1) 2
9 ( s − 2) 9 ( s + 1) 3 ( s + 1) 2
Hence
 s+3  5 −1  1  5 −1  1  2 −1  1 
L−1  2 
= L  − L  − L  2 
 ( s − 2)( s + 1)  9 s − 2 9  s + 1 3  ( s + 1) 
5 5 2
= e2t − e−t − te −t
9 9 3

 1 
Example 3: Find L−1  2 
 ( s + 1)( s + 4) 
2

Here by regarding s 2 as a simple pole, we first express the transform as

1 A B
= 2 + 2
( s + 1)( s + 4) s + 1 s + 4
2 2

where
1 1 1
A = lim ( s 2 + 1)  2 = lim = ,
s →−1
2
( s + 1)( s + 4) s →−1 ( s + 4) 3
2 2 2

1 1 1
B = lim ( s 2 + 4)  2 = lim =− .
s →−4
2
( s + 1)( s + 4) s →−4 ( s + 1)
2 2 2
3
With these values of the coefficients, the transform becomes
1 1 1 1 1
= − 2 .
( s + 1)( s + 4) 3 s + 1 3 s + 4
2 2 2

Therefore,
 1  1 −1  1  1 −1  2  1 1
L−1  2  = L  2 − L  2  = sin t − sin 2t.
 ( s + 1)( s + 4)  3 s +1  6 s + 4 3
2
6

13 EM
  s 
Example 4: Find L−1  2 
 ( s + 1)( s + 4) 
2

Regarding s 2 as again a simple pole and using the preceding results, we may recast the
transform as
s 1 s 1 s
= − 2 .
( s + 1)( s + 4) 3 s + 1 3 s + 4
2 2 2

Therefore,
 s  1 −1  s  1 −1  s  1 1
L−1  2 = L  2 − L  2  = cos t − cos 2t.
 ( s + 1)( s + 4)  3 s +1  3 s + 4 3
2
3

 1 
Example 5: Find L−1  5 
 ( s + 4) 
 1  t
n
−1
Invoking the Shift Theorem, and remembering that L  n +1  = w, e get
 s  n!
 1  e t −4 t 4 −4 t 4
e t
L−1  5 
= = .
 ( s + 4)  4! 24
 1 
Example 6: Find L−1  2 
 s + 4s + 9 
Giving a shift to s , the transform becomes
1 1 1 1 5
= = = .
s + 4s + 9 ( s + 2) + 5 ( s + 2)2 + 5
( ) ( 5)
2 2 2 2
5 ( s + 2) 2 +
Hence
 

−1 1  1 −1  5  1 −2t
L  2 = L  2 
= e sin 5t.
 s + 4s + 9  5  ( s + 2) 2 +
 ( )5 

5

Exercises: Find the inverses of the following Laplace transforms:

s−2 3s − 4 5
i) F (s) = ; ii) F (s) = 2 ; iii) F ( s) = .
( s − 1) ( s − 3)
2
s +9 s+3
2s + 1 s
iv) F (s) = ; v) F ( s) = 2 ;
( s − 3) 2
s − 2s + 2
3s + 2 1
vi) F ( s) = 2 ; vii) F (s) = 4 ;
s − 4s − 21 s − 2s3

14 EM
 3s − 8 1 s +1
viii) F (s) = ; ix) F (s) = ;
4 s 2 + 25 s2 s2 + 1
s+2 1
x) F (s) = 2 ; xi) F ( s) = ;
s ( s + 3) s ( s + 4 s + 13)
2

s+3 3s + 7
xii) F ( s) = 2 ; xiii) F ( s) = 2 .
( s + 6 s + 25) 2 s − 2s − 3

Solutions:
 2 s + 1  3t
iv) L−1  2 
= e (2t + 7);
 ( s − 3) 
 s 
L−1  2  = e ( cos t + sin t ) ;
t
v)
 s − 2 s + 2 
 3s + 2  1
L−1  2  = ( 23e + 7e ) ;
7t −3 t
vi)
 s − 4 s − 21  10
 1  1 2t
vii) L−1  4 3 
= ( e − 1 − 2t − 2t 2 ) ;
 s − 2 s  8
 3s − 8  3 5t 4 5t
viii) L−1  2  = cos − sin ;
 4 s + 25  4 2 5 2
 1 s +1 
ix) L−1  2 2  = 1 + t − cos t − sin t ;
 s s + 1
 s+2  2 1 1 −3t
x) L−1  2 = t+ − e ;
 s ( s + 3)  3 9 9
 s+3  1
xi) L−1  2 2 
= ( 3 − 2e−2t sin 3t − 3e−2t cos 3t ) ;
 ( s + 6s + 25)  39
 3 s + 7  1 −3 t
xii) L−1  2  = te sin 4t.
 s − 2s − 3  8

15 EM
LAPLACE TRANSFORMS OF DERIVATIVES

Theorem 6: If L  f (t ) = F ( s) , then L  f (t ) = sF (s) − f (0) .

Proof: Let f (t ) denote the first derivative of f (t ). Then, by definition,


L  f (t ) =  e− st f (t )dt.
0

dv du
Setting u = e − st and = f (t ), so that = − se − st and v = f (t ), and invoking
dt dt
the formula for integration by parts to perform the integrand on the right-hand
side, we have


L  f (t ) = e− st f (t )  + s  e− st f (t )dt.
0
0

But since lim  e − st

f (t )  = 0 , because s  0 and large enough to ensure that the
t →

decay of the exponential function e − st overshadows any growth characteristic of

f (t ), and at the same time lim  e − st f (t )  = f (0) , there follows that the integrated
t →0

part yields only − f (0). Therefore, we are left with the expression

L  f (t ) = − f (0) + s  e− st f (t )dt.
0

If then we recognise that the remaining integral is in fact the Laplace transform of
f (t ) , there immediately follows that
L  f (t ) = sF (s) − f (0),
which confirms our assertion.

Using the Laplace transform notation then, we have

L  f (t ) = sL  f (t ) − f (0).
If we need to find the Laplace transform of the second derivative we simply replace
f (t ) by f (t ) and f (t ) by f (t ) .
L  f (t ) = sL  f (t ) − f (0).
Inserting the already established expression L  f (t ) = sF (s) − f (0), we then find that
L  f (t ) = s  sL  f (t ) − f (0). = s 2 L  f (t ) − sf (0) − f (0).
By similar procedures, we find the
L  f (t ) = s3 L  f (t ) − s 2 f (0) − sf (0) − f (0).
L  f (iv ) (t ) = s 4 L  f (t ) − s3 f (0) − s 2 f (0) − sf (0) − f (0).
L  f ( v ) (t ) = s5 L  f (t ) − s 4 f (0) − s3 f (0) − s 2 f (0) − sf (0) − f (iv ) (0).
and so on. In general, we find that

16 EM
L  f ( n ) (t ) = s n L  f (t ) −  s n − k f ( k −1) (0)
n

k =1

where f (0)
(0) = f (0).

The above results or formulae are very fundamental in the solution of differential
equations. Thus if we let
dx d2x dnx
x t =0 = x0 , = x1 , = x2 , …, = xn
dt t =0 dt 2 t =0 dt n t =0
with L  f (t ) = X , then we have
L  f (t ) = sX − x0 .
L  f (t ) = s 2 X − sx0 − x1 ,
L  f (t ) = s3 X − s 2 x0 − sx1 − x2 ,
L  f (iv ) (t ) = s 4 X − s3 x0 − s 2 x1 − sx2 − x3 ,
L  f ( v ) (t ) = s5 X − s 4 x0 − s3 x1 − s 2 x2 − sx3 − x4 ,
and so on. In general, therefore,
L  f ( n ) (t ) = s n X −  s n − k xk −1 .
n

k =1

Exercises:
a) Find L sin(bt −  ) ;
b) Find L cos(at +  )
c) Find L e−2t ( cos t + 6sin t )
1 1 
d) Find L  sin at − cos bt 
a b 

17 EM
SOLUTIONS OF DIFFERENTIAL EQUATIONS BY USE OF LAPLACE
TRANSFORMS

In solving differential equations by use of Laplace transforms, four distinct steps are
followed:
1. The differential equation in question is written in terms of its Laplace transform,
2. The initial conditions are inserted or applied,
3. The transform equation is re-arranged to give the transform of the solution,
4. The inverse of the transform of the solution is determined.

Given a function f (t ) , we recall that the Laplace transforms of the derivatives of the
various orders of this solution are given by:
L  f (t ) = sF (s) − f (0)
L  f (t ) = s 2 F (s) − sf (0) − f (0),
L  f (t ) = s3 F (s) − s 2 f (0) − sf (0) − f (0),
L  f (iv ) (t ) = s 4 F (s) − s3 f (0) − s 2 f (0) − sf (0) − f (0),
and so on, such that in general we have
L  f ( n ) (t ) = s n F ( s ) −  s n − k f ( k −1) (0), n = 1, 2,3,
n

k =1

We need to develop a convenient notation for solving differential equations using

Laplace transforms. For the function f (t ) , let
x = f (t ) ,
L  f (t ) = F ( s) = X ,
and for the initial conditions at t = 0, let
dx d2x d3x
xt = 0 = x0, = x1 , = x2 , = x3 ,
dt t =0 dt 2 t =0 dt 3 t =0
and so on, such that
dnx
= xn , n = 0,1, 2,
dt n t =0
Accordingly, therefore, let us set
f (0) = x0 , f (0) = x1 , f (0) = x2 , f (0) = x3 ,
such that, in general,
f ( n) (0) = xn , n = 0,1, 2,

With these notations, we then have

L  f (t ) = sX − x0 ,
L  f (t ) = s 2 X − sx0 − x1 ,
L  f (t ) = s3 X − s 2 x0 − sx1 − x2 ,
and so on, such that in general we have

18 EM
L  f ( n ) (t ) = s n X −  s n − k xk −1 , n = 1, 2,3,
n

k =1

Note the following features when we are writing the n − th derivative:

1. the first term is s n X ,
2. the powers of s then decrease with succeeding terms,
3. all terms after the first are negative,
4. the subscripts increase in order of x0 , x1 , x2 , x3 ,

dx
Example 1: Solve − 4 x = 8 , given that x = 2 at t = 0.
dt
A convenient notation would of course have been x − 4x = 8 , x = 2 at t = 0.

Step 1: Re-write the entire equation in terms of Laplace transforms:

 dx 
L   − 4L  x = L 8
 dt 
or
8
sX − x0 − 4 X =
s
Step 2: Insert the initial condition x0 = 2 :
8
sX − 2 − 4 X =
s
so that we have
8
( s − 4) X = + 2
s
Step 3: Re-arrange to make X the subject of the expression:
8 2
X = +
s( s − 4) ( s − 4)
Step 4: Take the inverse transform of X :
8
First break X = into partial fractions as
s( s − 4)
8 A B
= +
s( s − 4) s s − 4
where using “Cover-up Rule”, we have
8 8
A = lim s  = lim = −2,
s →0 s( s − 4) s →0 ( s − 4)
and
8 8
B = lim( s − 4)  = lim = 2,
s →4 s( s − 4) s →4 s
Thus
4 2
X = − .
s−4 s

19 EM
 Taking the inverse transform, we thus find that
x = 4e − 2.
4t

dx
Example 2: Solve 2 + 3 x = e 4t , given that x = 5 at t = 0.
dt
Let us convenient write the equation as
2 x + 3 x = e 4t , with x = 5 at t = 0.

Step 1: Express the equation in terms of Laplace transforms:

 dx 
2L   + 3L  x = L e 4t 
 dt 
or
1
2sX − 2 x0 + 3 X = .
s−4
Step 2: Insert the initial condition x0 = 5 :
1
2 sX − 10 + 3 X = .
s−4
so that we have
1
(2 s + 3) X = + 10.
s−4
Step 3: Re-arrange to make X the subject of the expression:
1 10
X = + .
( s − 4)(2s + 3) 2s + 3
Note that the second term on the right can be expressed as
10 5
= .
2s + 3 s + (3 / 2)
The first term on the right can be split into partial fractions as
1 A B
= + ,
( s − 4)(2s + 3) s − 4 2s + 3
whence, by the cover-up rule, we have
1 1 1
A = lim( s − 4)  = lim = ,
s →4 (2s + 3)( s − 4) s → 4 (2s + 3) 11
and
1 1 2
B = lim (2s + 3)  = lim =− .
s →− (3 / 2) (2s + 3)( s − 4) s →− (3 / 2) ( s − 4) 11
Hence,
1 1 1 1 5
X = − + .
11 s − 4 11 s + (3 / 2) s + (3 / 2)
Step 4: Take the inverse transform of X :
x = e4t − e − (3 / 2) t + 5e − (3 / 2) t = ( 54e − (3 / 2) t + e4t ) = f (t ).
1 1 1
11 11 11

20 EM
 dx
Example 3: Solve 2 + 5 x = e −2t , given that x = 1 at t = 0.
dt
Let us convenient write the equation as
2 x + 5 x = e −2t , with x = 3 at t = 0.

Step 1: Express the equation in terms of Laplace transforms:

 dx 
2L   + 5L  x = L e −2t 
 dt 
or
1
2sX − 2 x0 + 5 X = .
s+2
Step 2: Insert the initial condition x0 = 1:
1
2 sX − 2 + 5 X = .
s−4
so that we have
1
(2 s + 5) X = + 2.
s+2
Step 3: Re-arrange to make X the subject of the expression:
1 2
X = + .
( s + 2)(2s + 5) 2 s + 5
Note that the second term on the right can be expressed as
2 1
= .
2s + 5 s + (5 / 2)
The first term on the right can be split into partial fractions as
1 A B
= + ,
( s + 2)(2 s + 5) s + 2 2 s + 5
whence, by the cover-up rule, we have
1 1
A = lim( s + 2)  = lim = 1,
s →−2 (2s + 5)( s + 2) s →−2 (2s + 5)
and
1 1
B = lim (2s + 5)  = lim = −2.
s →− (5 / 2) (2s + 5)( s + 2) s →− (5 / 2) ( s + 2)
Hence,
1 1 1 1
X = − + =
s + 2 s + (5 / 2) s + (5 / 2) s + 2
Step 4: Take the inverse transform of X :
x = f (t ) = e−2t .

dx
Example 4: Solve 3 − 4 x = sin 2t , given that x = 1/ 3 at t = 0.
dt
Let us conveniently write the equation as

21 EM
3x − 4x = sin 2t , with x = 1/ 3 at t = 0.

Step 1: Express the equation in terms of Laplace transforms:

 dx 
3L   − 4L  x = L sin 2t
 dt 
or
2
3sX − 3 x0 − 4 X = 2 .
s +4
Step 2: Insert the initial condition x0 = 1/ 3 :
2
3sX − 1 − 4 X = 2 .
s +4
so that we have
2
(3s − 4) X = 2 + 1.
s +4
Step 3: Re-arrange to make X the subject of the expression:
2 1
X = 2 + .
( s + 4)(3s − 4) 3s − 4
Note that the second term on the right can be expressed as
1 1/ 3
= .
3s − 4 s − (4 / 3)
The first term on the right can be split into partial fractions as
2 A Bs + C
= + 2 .
( s + 4)(3s − 4) 3s − 4 s + 4
2

Using the cover-up rule, we have

2 2 18 9
A = lim (3s − 4)  = lim 2 = = ,
s →4 / 3 (3s − 4)( s + 4) s →4 / 3 ( s + 4) 52 26
2

Noting now that

2 9 (Bs + C )
= + 2 ,
( s + 4)(3s − 4) 26(3s − 4) ( s + 4)
2

or
2 (9 / 26)( s 2 + 4) + ( Bs + C )(3s − 4)
= .
( s 2 + 4)(3s − 4) ( s 2 + 4)(3s − 4)
Multiplying out the numerator on the right-hand side and collecting coefficients
of powers of s , we have
2 [(9 / 26) + 3B) s 2 + (3C − 4 B) s + [18 /13 − 4C )
= .
( s 2 + 4)(3s − 4) ( s 2 + 4)(3s − 4)
Equating coefficients of powers of s on both sides, we get the system of
equations:
(9 / 26) + 3B = 0,
18 /13 − 4C = 2,
3C − 4B = 0.

22 EM
 3
From the first equation, we have B = − . From the second equation we
26
2
have C = − .
13
With these values of the coefficients, the transform equation becomes:
2 9 (3s + 4)
= − ,
( s + 4)(3s − 4) 26(3s − 4) 26( s 2 + 4)
2

or
2 3 1 3 s 4
= − − ,
( s + 4)(3s − 4) 26 s − (4 / 3) 26 s + 4 26( s 2 + 4)
2 2

Therefore,
1 2
X = + 2
3s − 4 ( s + 4)(3s − 4)
1 1 3 1 3 s 4
= + − −
3 s − (4 / 3) 26 s − (4 / 3) 26 s + 4 26( s 2 + 4)
2

35 1 3 s 4
= − − .
78 s − (4 / 3) 26 s + 4 26( s 2 + 4)
2

Hence, taking the inverse Laplace transform, we get

35 (4 / 3) t 1
x(t ) = e − ( 3cos 2t − 2sin 2t ) .
78 26

d2x dx
Example 5: Solve 2
− 4 x = cos 2t , , given that x = 2, = 3 at t = 0.
dt dt
dx
A convenient notation would of course have been x − 4x = cos 2t , x = 2, = 3 at t = 0.
dt

Step 1: Re-write the entire equation in terms of Laplace transforms:

d2x 
L  2  − 4L  x = L cos 2t
 dt 
or
s
s 2 X − sx0 − x1 − 4 X = 2
s +4
Step 2: Insert the initial condition x0 = 2 and x1 = 3
s
s 2 X − 2s − 3 − 4 X = 2
s +4
so that we have
s
( s 2 − 4) X = 2 + 2s + 3
s +4
Step 3: Re-arrange to make X the subject of the expression:

23 EM
 s s 3
X = +2 2 + 2
( s + 4)( s − 4)
2 2
s −4 s −4
Step 4: Take the inverse transform of X :
s
First break 2 into partial fractions as
( s + 4)( s 2 − 4)
s As + B C D
= 2 + + .
( s + 4)( s − 4) s + 4 s + 2 s − 2
2 2

On gathering the right-hand side over the same denominator, we have

s ( A + C + D) s 3 + ( B − 2C + 2 D) + 4(C − A + D ) s + 2( D − C − 2 B )
= .
( s 2 + 4)( s 2 − 4) ( s 2 + 4)( s 2 − 4)
Upon equating the coefficients of the various powers of s in numerators, we get
A + C + D = 0,
B − 2C + 2 D = 0,
4(C − A + D ) = 1,
2( D − C + 2 B) = 0.
Solving this system of equations, we find that
1 1 1
A = − , B = 0, C = , D = .
8 16 16
Hence
s 1 s 1 1
=− 2 + + .
( s + 4)( s − 4)
2 2
8 s + 4 16( s + 2) 16( s − 2)
and
3 2s 1 s 1 1
X = 2 + 2 − 2 + + .
s − 4 s − 4 8 s + 4 16( s + 2) 16( s − 2)
Therefore
3 17 1
x(t ) = sinh 2t + cosh 2t − cos 2t.
2 8 8

d2x dx dx
Example 6: Solve 2 − 3 + 2 x = 0, given that x = 4, = 3 at t = 0.
dt dt dt

Step 1: Re-write the entire equation in terms of Laplace transforms:

d2x   dx 
L  2  − 3L   + 2L  x = 0.
 dt   dt 
or
s 2 X − sx0 − x1 − 3sX + 3x0 + 2 X = 0
Step 2: Insert the initial condition x0 = 4 and x1 = 3
s 2 X − 4s − 3 − 3sX + 12 + 2 X = 0
so that we have

24 EM
( s 2 − 3s + 2) X − 4s + 9 = 0
Step 3: Re-arrange to make X the subject of the expression:
4s − 9 4s − 9
X = 2 =
s − 3s + 2 ( s − 2)( s − 1)
Step 4: Take the inverse transform of X :
4s − 9
First break into partial fractions as
( s − 2)( s − 1)
4s − 9 A B
= + .
( s − 2)( s − 1) s − 2 s − 1
On gathering the right-hand side over the same denominator, we have
4s − 9 ( A + B) s − ( A + 2 B)
= .
( s − 2)( s − 1) ( s − 2)( s − 1)
Upon equating the coefficients of the various powers of s in numerators, we get
A + B = 4,
A + 2 B = 9.
Solving this system of equations, we find that
A = −5, B = 9.
Hence
5 9
X =− + .
s − 2 s −1
and therefore
x(t ) = 9et − 5e2t .

d2x dx dx
Example 7: Solve 2 − 7 + 12 x = 2, given that x = 1, = 5 at t = 0.
dt dt dt

Step 1: Re-write the entire equation in terms of Laplace transforms:

d2x   dx 
L  2  − 7L   + 12L  x = 0.
 dt   dt 
or
2
s 2 X − sx0 − x1 − 7 sX + 7 x0 + 12 X =
s
Step 2: Insert the initial condition x0 = 1 and x1 = 5
s 2 X − s − 5 − 7 sX + 7 + 12 X = 0
so that we have
2 s 2 − 2s + 2
( s 2 − 7 s + 12) X = + s − 2 =
s s
Step 3: Re-arrange to make X the subject of the expression:
s 2 − 2s + 2 s 2 − 2s + 2
X = =
s ( s 2 − 7 s + 12) s ( s − 4)( s − 3)

25 EM
Step 4: Take the inverse transform of X :
s 2 − 2s + 2
First break into partial fractions as
s( s − 4)( s − 3)
s 2 − 2s + 2 A B C
= + +
s( s − 4)( s − 3) s s − 4 s − 3
Using the cover-up rule, we have
s 2 − 2s + 2 s 2 − 2s + 2 1
A = lim s  = lim = ,
s →0 s( s − 4)( s − 3) s → 0 ( s − 4)(s − 3) 6
s 2 − 2s + 2 s 2 − 2s + 2 10 5
B = lim( s − 4)  = lim = = ,
s →4 s( s − 4)( s − 3) s →4 s( s − 3) 4 2
s 2 − 2s + 2 s 2 − 2s + 2 5
C = lim( s − 3)  = lim = ,
s →3 s( s − 4)( s − 3) s →3 ( s − 4) s 3
Hence
11 5 1 5 1
X = + − .
6 s 3 s −4 2 s −3
and therefore
1 5 5
x(t ) = + e 4t − e3t .
6 2 3

d2x dx dx
Example 8: Solve 2 − 6 + 8 x = e3t , given that x = 0, = 2, at t = 0.
dt dt dt

Step 1: Re-write the entire equation in terms of Laplace transforms:

d 2x   dx 
L  2  − 6L   + 8L  x = L e3t  .
 dt   dt 
or
1
s 2 X − sx0 − x1 − 6sX + 6 x0 + 8 X =
s −3
Step 2: Insert the initial condition x0 = 0 and x1 = 2
1
s 2 X − 2 − 6sX + 8 X = ,
s −3
so that we have
1
( s 2 − 6 s + 8) X = + 2.
s−3
Step 3: Re-arrange to make X the subject of the expression:
s 2 − 2s + 2 2s − 5
X = 2 =
( s − 6 s + 8) ( s − 4)( s − 3)( s − 2)
Step 4: Take the inverse transform of X :
4s + 9
First break − into partial fractions as
( s − 4)( s − 2)

26 EM
 13 − 4s A B C
= + +
( s − 4)( s − 2) s − 4 s − 3 s − 2
Using the cover-up rule, we have
2s − 5 2s − 5 3
A = lim( s − 4)  = lim = ,
s →4 ( s − 4)( s − 3)( s − 2) s →4 ( s − 3)( s − 2) 2
2s − 5 2s − 5
B = lim( s − 3)  = lim = −1,
s →3 ( s − 4)( s − 3)( s − 2) s →3 ( s − 4)( s − 2)
2s − 5 2s − 5 1
C = lim( s − 2)  = lim =− .
s →2 ( s − 4)( s − 3)( s − 2) s → 4 ( s − 3)( s − 4) 2

Hence
3 1 1 1 1
X = − −
2 s −4 s −3 2 s −2

and therefore
3 1
x(t ) = e 4t − e3t − e 2t .
2 2

Example 9: Use Laplace transform techniques to determine the expression for the current
i (t ) flowing in the circuit at any time t . Find the steady-state value.
i (t ) E0

L R

An RL-Circuit
Recall that the differential equation governing the system is given by the expression
di
L + Ri = E0
dt
and we assume that i = 0 at t = 0. Taking the Laplace transform of each term, we have
 di 
LL   + R L i = L  E0  .
 dt 
On invoking the formula for the Laplace transform of derivatives, we get
E
sLI − Li0 + RI = 0 ,
s
Under the assumption that initially there is no current flowing in the circuit, we get

27 EM
 E0
sLI + RI =,
s
The Laplace transform of the current we wish to determine is therefore given by
E0 E 1 E 1
I = = 0 − 0 ,
s ( sL + R ) R s R s + R / L
Taking the inverse Laplace transform, we get the desired expression for the current as
As t →  , the current in the circuit attain a steady-state value given by
E
lim i(t ) = i = 0 .
t → R

i (t )
Graph of i (t )

E0
R

Example 10

Determine the current in the RLC electrical circuit below, assuming that initially no
current is flowing in the circuit and the applied voltage is a constant E0 .

R
e
i

L C

The differential equation governing the current flow in the circuit is given by
di q
L + Ri + = e (0.1)
dt C
Using Laplace transformation to re-write this equation, we have
1
sLI − Li (0) + RI + L q = L e . (0.2)
C
Given that

28 EM
 dq
i=
dt
there follows that
 dq 
L i = I = L   = sQ − q (0).
 dt 
Consequently,
I q(0)
+ = Q = L q .
s s
The transformed equation thus becomes
1
sLI − Li (0) + RI +  I + q (0)  = L e . (0.3)
sC 
Considering the initial conditions i(0) = 0, e(0) = E0 , q(0) = 0; this expression becomes
I E
sLI + RI + = 0 .
sC s
since Solving for the transform of the current, we have
1 E0 E0
I = E0 = = (0.4)
1  1  R   L ( s +  ) + n 2 
2 2 2
s 2 L + sR + R 
L  s +  + −    
C
 2L  LC  2 L  
where
2
R 1  R 
= , n2 = −  . (0.5)
2L LC  2 L 

There are three distinct cases to consider here, namely,

2
1  R 
(i) n2 = −    0,
LC  2 L 
2
1  R 
(ii) n = 2
−   = 0,
LC  2 L 
2
1  R 
(iii) n = 2
−  = −k , with k  0.
2 2

LC  2 L 

Taking the inverse of the transforms of the current for the three cases, we have
E
I = 0 e− t sin nt (0.6)
nL
for case (i),
E
I = 0 te− t (0.7)
nL
for case (ii), and
E
I = 0 e− t sinh kt (0.8)
kL
for case (iii).

29 EM
Example:
Now determine the current in the RLC electrical circuit below, assuming that there is no
applied voltage but that the initially is initially charged to a voltage E0 = q(0) / C and no
current initially flows within the circuit.

R
e
i

L C

The differential equation governing the current flow in the circuit is given by
di q
L + Ri + = 0. (0.9)
dt C
Using Laplace transformation to re-write this equation, and imposing the initial
conditions, we have
I CE
sLI + RI + = − 0 . (0.10)
s s
Replacing I by Q
I q(0)
L q = Q = + ,
s s
Consequently,
I q(0)
+ = Q = L q .
s s
The transformed equation thus becomes
1
sLI − Li (0) + RI +  I + q (0)  = 0 .
sC 
q (0)
Considering the initial conditions i (0) = 0, = E0 , this expression becomes
C
I E
sLI + RI + =− 0 .
sC s
Since I = sQ − q(0) , we may re-write this expression in terms of Q as

Solving for the transform of the current, we have

30 EM
add some thing here

Example 11
d 2x dx dx
Solve the equation 2 − 2 + 5 x = e2t , given that x = 0 and = 0 when t = 0.
dt dt dt

In terms of of Laplace tyransform, this equation can be re-written as

1
s 2 X − sx0 − x1 − 2sX + 2 x0 + 5 X = .
s−2
On inserting the initial conditions, we have
1
( s 2 − 2 s + 5) X = + 1.
s−2
On solving for X , we get
1 1
X = 2 +
s − 2 s + 5 ( s − 2)( s − 2 s + 5)
2

Expanding the second term as partial fractions, we have

1 A Bs + C
= + 2
( s − 2)( s − 2 s + 5) s − 2 s − 2 s + 5
2

Using the cover-up rule, we have

1 1
A = lim( s − 2)  = .
s →2 ( s − 2)( s − 2 s + 5) 5
2

With this value of the coefficient, we hav

1 (1/ 5) Bs + C [(1/ 5) + B]s 2 − [(2 / 5) + 2 B − C ]s + 1 − 2C
= + =
( s − 2)( s 2 − 2 s + 5) s − 2 s 2 − 2 s + 5 ( s − 2)( s 2 − 2 s + 5)
Comparing the coefficients in the numerators, we get
1 2
+ B = 0, + 2 B − C = 0, 1 − 2C = 1.
5 5
Solving this system of equations, we get
1
B = − , C = 0.
5
Hence
1 1 1 1 s
X = + 2 −
5 s − 2 s − 2 s + 5 5 ( s − 2 s + 5)
2

By noting that s 2 − 2s + 5 = ( s − 1)2 + 4, we get

1 1 1 1 s −1 1 1
X = + − −
5 s − 2 ( s − 1) + 4 5 ( s − 1) + 4 5 ( s − 1) 2 + 4
2 2

1 1 1 4 1 s −1
= + −
5 s − 2 5 ( s − 1) + 4 5 ( s − 1) 2 + 4
2

Taking the inverse Laplace transform, we have

31 EM
 1 2 1
x(t ) == e 2t + et sin 2t − et cos 2t.
5 5 5

32 EM
SOLVING SIMULTANEOUS EQUATIONS USING LAPLACE TRANSFORMS

Consider solving the system of equations

dx 
+ 4 x + 5 y = 7 cos 2t 
dt 
 , x = y = 0, when t = 0.
dy
− 4 y − 5 x = 8cos 2t 
dt 
Writing the equations in terms of Laplace transforms, we have
s
sX + 4 X + 5Y = 7 2 ,
s +4
s
sY − 4Y − 5 X = 8 2 ,
s +4
or
s
( s + 4) X + 5Y = 7 2 ,
s +4
s
( s − 4)Y − 5 X = 8 2 .
s +4
Multiplying the first equation by s − 4 and the second equation by −5 , we have
s ( s − 4)
( s + 4)( s − 4) X + 5( s − 4)Y = 7 2 ,
s +4
s
−5( s − 4)Y + 25 X = −40 2 .
s +4
Adding the two expressions, we get
7 s 2 − 68s
[( s + 4)( s − 4) + 25] X = .
s2 + 4
Solving for X , we get
7 s 2 − 68s
X = .
( s 2 + 4)( s 2 + 9)
Noting that s 2 + 4 = ( s + 2 j )( s − 2 j ) and s 2 + 9 = ( s + 3 j )( s − 3 j ) , we can expression the
right-hand side as partial fractions and thus have
7 s 2 − 68s A B C D
= + + + .
( s + 4)( s + 9) s + 2 j s − 2 j s + 3 j s − 3 j
2 2

Invoking the cover-up rule, we have

 7 s 2 − 68s  1
A = lim   = − (34 + 7 j ),
s →−2 j ( s − 2 j )( s + 9)
2
  5
 7 s 2 − 68s  1
B = lim   = − 5 (34 − 7 j ),
s → 2 j ( s + 2 j )( s 2 + 9)
 
 7 s 2 − 68s  1
C = lim   = 10 (68 + j 21),
s →−3 j ( s − 3 j )( s 2 + 4)
 

33 EM
  7 s 2 − 68s  1
D = lim   = 10 (68 − j 21),
s →3 j ( s + 3 j )( s 2 + 4)
 

Hence
1 (34 + 7 j ) 1 (34 − 7 j ) 1 (68 + 21 j ) 1 (68 − 21 j )
X =− − + +
5 s+2j 5 s−2j 10 s + 3 j 10 s − 3 j
68 s 28 1 68 s 63 1
=− − + +
5 s + 4 5 s + 4 5 s + 9 5 s2 + 9
2 2 2

Taking the inverse transform, we have

68 14 68 21
X = − cos 2t − sin 2t + cos 3t + sin 3t.
5 5 5 5

On the other hand, had we multiplied the first equation by −5 and the second equation by
s + 4 , we would have
s
−5( s + 4) X − 25Y = −35 2 ,
s +4
s ( s + 4)
( s − 4)( s + 4)Y − 5( s + 4) X = 8 2 .
s +4
On subtracting the first expression from the second and solving for Y , we then find that
8s 2 + 67 s
Y = 2 .
( s + 4)( s 2 + 9)
Next we express the left-hand side as partial fractions and thereby have
8s 2 + 67 s A1 B1 C1 D1
= + + + .
( s + 4)( s + 9) s + 2 j s − 2 j s + 3 j s − 3 j
2 2

Using the cover-up rule, we have

 8s 2 + 67 s  −32 − j134 1
A1 = lim 
s →−2 j ( s − 2 j )( s 2 + 9)
 = (−4 j )(5) = 10 ( 67 − j16 ) ,
 
 8s 2 + 67 s  −32 + j134 1
B1 = lim 
s → 2 j ( s + 2 j )( s 2 + 9)
 = (4 j )(5) = 10 ( 67 + j16 ) , .
 
 8s + 67 s  −72 − j 201
2
1
C1 = lim 
s →−3 j ( s − 3 j )( s 2 + 4)
 = (−6 j )(−5) = − 10 ( 67 − j 24 ) ,
 
 8s 2 + 67 s  −72 + j 201 1
D1 = lim   = = − ( 67 + j 24 ) .
s →3 j ( s + 3 j )( s + 4)

2
 (6 j )(−5) 10
With these coefficients, we then have

34 EM
 8s 2 + 67 s
Y =
( s 2 + 4)( s 2 + 9)
1 67 − j16 1 (67 + j16) 1 67 − j 24 1 67 + j 24
= + − −
10 s + 2 j 10 s − 2 j 10 s + 3 j 10 s − 3 j
67 s 32 s 67 s 72 s
= − − +
5 s + 4 5 s + 4 5 s + 9 5 s2 + 9
2 2 2

Taking the inverse, we have

67 16 67 24
y (t ) = cos 2t − sin 2t − cos 3t + sin 3t.
5 5 5 5

Alternatively, having obtained the transforms

7 s 2 − 68s
X = 2
( s + 4)( s 2 + 9)
and
8s 2 + 67 s
Y = 2 ,
( s + 4)( s 2 + 9)
we could instead express them in partial fractions as
7 s 2 − 68s As + B Cs + D
X = 2 = 2 + 2
( s + 4)( s + 9) s + 4
2
s +9
and
8s 2 + 67 s A s + B1 C1 s + D1
Y = 2 = 12 + 2 ,
( s + 4)( s + 9)
2
s +4 s +9
Putting the right-hand terms under over the same denominators, we would then have
7 s 2 − 68s ( A + C ) s 3 + ( B + D) s 2 + (9 A + 4C ) s + (9 B + 4 D)
X = 2 =
( s + 4)( s 2 + 9) ( s 2 + 4)( s 2 + 9)
and
8s 2 + 67 s ( A1 + C1 ) s 3 + ( B1 + D1 ) s 2 + (9 A1 + 4C1 ) s + (9 B1 + 4 D1 )
Y = 2 = ,
( s + 4)( s 2 + 9) ( s 2 + 4)( s 2 + 9)
Equating coefficients of the various powers of s , we have
A + C = 0, B + D = 7, 9 A + 4C = −68 , 9B + 4D = 0.
A1 + C1 = 0, B1 + D1 = 8, 9 A1 + 4C1 = 67 , 9 B1 + 4 D1 = 0.
Solving this system of equations, we have
68 28 68 63
A=− , B=− , C = , D= ,
5 5 5 5
67 32 67 72
A1 = − , B1 = − , C1 = , D1 = ,
5 5 5 5
In this way, we would then find that

35 EM
 (−68 / 5) As + (−28 / 5) (68 / 5) s + (63 / 5)
X = +
s2 + 4 s2 + 9
68 s 28 1 68 s 63 1
=− − + +
5 s + 4 5 s + 4 5 s + 9 5 s2 + 9
2 2 2

and
(−67 / 5) s + ( −32 / 5) (67 / 5) s + (72 / 5)
Y = +
s2 + 4 s2 + 9
,
67 s 32 1 67 s 72 1
=− − + +
5 s2 + 4 5 s2 + 4 5 s2 + 9 5 s2 + 9
Looking –up the inverses then, we have
68 14 68 21
x(t ) = − cos 2t − sin 2t + cos 3t + cos 3t
5 5 5 5
and
67 16 67 24
Y = − cos 2t − sin 2t + cos 3t + sin 3t. ,
5 5 5 5

36 EM
SPECIAL CASES OF LAPLACE TRANSFORMS

a) Laplace Transform of Integrals

t
F ( s)
Theorem: If L  f (t ) = F ( s) , then  f (u)du = .
0
s

t
Proof: Let g (t ) =  f (u )du . Then g (t ) = f (t ) and g (0) = 0. Taking the Laplace
0

transform of g (t ) , we know that

L  g (t ) = sG( s) − g (0) = sG( s) = L  f (t ) = F ( s). Hence sG ( s ) = F ( s ), or
t
F ( s)
G( s) = = L  g (t ) =  f (u ) ( du ), as asserted.
s 0

a t  1 a
Example: Since L sin at = 2 , then L   sin audu  = .
s +a  s (s + a )
2 2 2
0
 i (t )  I
Likewise, since the earlier example we had L   = , there follows that
C  C
1 t  I
L   i (u )du  = .
C 0  sC

b) Multiplication by Integral Powers of t

d n F (s)
Theorem: If L  f (t ) = F ( s) , then L t n f (t ) = (−1)n n
= (−1)n F ( n ) ( s).
ds
Example: Since we know that
a
L sin at = 2 ,
s + a2
using this Theorem we find that
d  a  2as
L t sin at = −  2 2 
= 2 ,
ds  s + a  ( s + a 2 )2
Likewise,
d2  s  d2  s  2s (3s − 2a )
2 2
L t 2 cos at = (−1) 2 2  2 = = .
ds  s + a 2  ds 2  s 2 + a 2  ( s 2 + a 2 )3

Tasks: Evaluate the following expressions:

i) L t sinh 2t.

37 EM
 2 d  2  4s
We note that L sinh 2t = . Therefore L t sinh 2t = −  2  = 2 .
s −4
2
ds  s − 4  ( s − 4) 2
ii) L t 2 sinh 2t.
2
We again note that L sinh 2t = . Therefore
s −4 2

d 2  2  d  4s  4(3s 2 + 4)
L t 2 sinh 2t = (−1)2 2  2 =  = .
ds  s − 4  ds  ( s 2 − 4) 2  ( s 2 − 4)3

b) Division by t

 f (t )  f (t )
Theorem: If L  f (t ) = F ( s) , then L   =  F (u )du, provided lim exists.
 t  s t →0 t
f (t )
Proof: Let g (t ) = . Then clearly f (t ) = tg (t ) . Taking the Laplace transform of both
t
sides, we have
d
L  f (t ) = L tg (t ) = − G ( s) = F ( s)
ds
In other words,
s  
f (t ) − st
G( s) = −  F (u )du =  F (u )du =  e dt ,
 s 0
t
as asserted.

Alternatively, know that


L  f (t ) =  f (t )e− st dt = F ( s).
0

Hence, integrating from s to  , we obtain

  
 
s s  0 f (t )e dt  ds
− st
F ( s ) ds =

f (t )
Under the assumption that lim exists, and that the function f (t ) is itself regular and
t →0 t
of exponential order on the interval [0, ), it follows that the order of integration on the
right-hand side can be interchanged, so that we have

 
  − st  
 e − st  
 e − st 
s F (s)ds = 0 f (t )  s e ds  dt = 0 f (t )  (−t )  dt =0 f (t ) 0 + t  dt
s

f (t ) − st  f (t ) 
= e dt = L  .
0
t  t 
For example, knowing that

38 EM
 a
L sin at = ,
s + a2 2

we use the Corollary to find that

 
 sin at  a  s  s s
L = 2 ds =  tan −1  = − tan −1 = cot −1 .
 t  s s +a  a s
2
2 a a

Note that by taking inverses in the assertion of the foregoing theorem and then solving
for f (t ) , we obtain the following useful result as a Corollary:
 
If L  f (t ) = F ( s) , then f (t ) = L F ( s) = t L   F ( s)ds  .
−1 −1

s 
Restated as a tool for finding the inverses, this Corollary states that “The inverse of a
transform is equal to t times the inverse of the integral of the transform from s to . ” This
Corollary is often invoked to find inverses when then the integral is simpler to work with
than the transform itself.

Thus, for example, if Likewise,

s
L  y (t ) = 2 ,.
( s − 1) 2
by the Corollary, we have
 s  
−1   −1  

−1  −1 
y (t ) = t L   2
−1
ds  = t L   2   = tL  2 
 s ( s − 1)   2( s − 1)  s   2( s − 1) 
2

1  1 1  t t t et − e − t t sinh t
= t L−1    4 ( ) 2 2 = 2
−t
− = e − e =
 4  ( s − 1) ( s + 1)  

Homework Assignments:

a) Find the Laplace transform of the following functions:

1 − cos 2t
i) f (t ) = cosh 2t ; ii) f (t ) = t 2 sin 2t; iii) f (t ) = ;
t
sin t
iv) f (t ) = e−3t cosh 3t; v) f (t ) = ; vi) f (t ) = e −2t sin 3t ;
t
vii) f (t ) = 2e−4t − 3e 4t ; viii) f (t ) = te−2t sin t; ix) f (t ) = sin kt cos kt ;
x) f (t ) = (sin t − cos t )2 ; xi) f (t ) = e−t sin 2 t.


a) Evaluate  te−2t sin tdt.
0

 cos at − cos bt  1  s + b 
2 2
b) Show that L   = ln  2 .
  2 s +a
2
t 

39 EM
 
3
c) Show that  te−3t sin tdt = .
0
50

cos 6t − cos 4t 3
d) Show that 
0
t
dt = ln .
2

40 EM
THE HEAVISIDE STEP FUNCTION

Hitherto we have been involved in finding the Laplace transforms of continuous

functions of independent variables, usually time. However, in practical applications there
are functions which change abruptly with time, i.e., “switch-on” and “switch-off”, at
certain specific values of time.

f (t ) 1, t  a,
f (t ) = 
0, t  a.
1

0 a t

The above sketch is the Heaviside Unit Step Function, and is denoted by
f (t ) = H (t − a ).
Thus, the Heaviside Unit Step Function for a = 4 is denoted by
f (t ) = H (t − 4).
and has the graph depicted below.

f (t ) 1, t  4,
f (t ) = 
0, t  4.
1

0 4 t

Thus H (t − a ) is a two-valued function.

If the Heaviside Unit Step Function occurs at the origin, then we have
f (t ) = H (t )
and the graph is depicted below.

41 EM
 f (t ) 1, t  0,
f (t ) = 
0, t  0.
1

0 t
We may express composite functions through the use of Heaviside Unit Step Functions,
as depicted in the following examples.

y1 (t ) 1, t  a,
y1 (t ) = 
0, t  a.
1

0 a t

y2 (t )
e −t , t  0,
1 y2 (t ) = 
 0, t  0.

0 a t

f (t )
f (t ) = y1 (t ) y2 (t ) = H (t − a)e−t
1

0 a t

42 EM
 f (t )
f (t ) = H (t −  / 2) sin t
1

 /2  3 / 2 2 t

f (t )
f (t ) = H (t − a) sin(t − a)
1

2
a  t

Laplace Transform of Heaviside Function: H (t − a )

From definition

L  H (t − a ) =  e − st H (t − a )dt
0
 
e − as
a
=  e H (t − a )dt +  e H (t − a )dt = 0 +  e − st dt =
− st − st

0 a a
s
Hence
e− as
L H (t − a) =
s
and, consequently,
1
L  H (t ) =
s

43 EM
Laplace Transform of Composite Functions: H (t − a ) f (t − a ).
From definition,

L  H (t − a ) f (t − a ) =  e − st H (t − a ) f (t − a )dt
0
a 
=  e − st H (t − a ) f (t − a )dt +  e − st H (t − a ) f (t − a )dt
0 a
 
= 0 +  e − st f (t − a )dt =  e − st f (t − a )dt.
a a

Setting t − a =  , so that dt = d and t = a +  , we have

 
=  e− st f (t − a)dt = e− sa  e− s f ( )d = e− sa F (s) = e− sa L  f (t ).
a 0

Hence
L H (t − a) f (t − a) = e− sa F (s) = e− sa L  f (t ).

Examples:
L  H (t − 3)(t − 3) 2  =
2! −3s 2 −3 s
e = 3e .
s3 s
e−5 s
L H (t − 5)et −5  = e−5 s L et  = .
s −1
s se−2 s
L  H (t − 2) cos(t − 2) = e −2 s
= .
s2 + 1 s2 + 1
4!e− s
L H (t − 1)(t − 1) 4  = 5 e − s = 5 .
4!
s s
2 2e− ( / 3) s
L H (t −  / 3)sin 2(t −  / 3) = 2 e− ( / 3) s = 2 .
s +4 s +4

Now we know that

L H (t − c) f (t − c) = e−cs F (s) = e −cs L  f (t ).
The reverse is true. That is, if F (s) = L  f (t ) , then
L e−cs F (s) = H (t − c) f (t − c),
where c is real and positive.

Examples:
 1 
L−1 e −2 s t −5
 = H (t − 2)e ),
 s − 5 

44 EM
  3 
L−1 e − s 2  = H (t − 1) sin 3(t − 1),
 s + 9
 s   1
L−1 e −3 s 2  = H (t − 3) cos 4(t − 3), L−1 e −4 s 2  = H (t − 4)(t − 4),
 s + 16   s 
 3  2
L−1 e −5 s  = 3H (t − 5), L−1 e −3 s 3  = H (t − 3)(t − 3) 2 .
 s  s 

Summary:
1, t  c,
1. H (t − c) =  .
0, t  c,
e − cs
2. L  H (t − c) = ,
s
1
3. L  H (t ) = ,
s
4. L H (t − c) f (t − c) = e− cs F (s) = e−cs L  f (t ) ,
5. If F ( s) = L  f (t ) , then e−cs F (s) = L H (t − c) f (t − c)
The above statements are all that we need for Laplace transforms of and inverse Laplace
transforms of functions involving the Heaviside Unit Step Function.

Some Further Examples

Demonstrate that
 s 
1. L−1 e − s 2  = H (t − 1) cos 3(t − 1),
 s + 9
 8 
2. L−1 e −4 s 2  = 4 H (t − 4) sin 2(t − 4),
 s + 4
 3 
3. L−1 e −2 s 2  = 3H (t − 2) sinh(t − 2),
 s − 1

Grasp the Following Important Results:

 1, t  a,
1. H (t − a ) = 
0, 0  t  a.
e− sa  e− sa 
2. L H (t − a) =  L−1   = H (t − a );
s  s 
1 1 
3. L  H (t ) =  L−1   = H (t );
s s
e− sa
4. L H (t − a) f (t − a) = F ( s ), where F ( s) = L  f (t ) ;
s
e− sa
5. If F ( s) = L  f (t ) , then F ( s) = L H (t − a) f (t − a).
s

45 EM
Exercises:
Determine the functions f (t ) whose Laplace transforms are given here below:
3 4e− s 5e−2 s
1. L  f (t ) = − 2 + 2 ;
s s s
−s
2 5e 8e−3s
2. L  f (t ) = − + ;
s s s
1 e− s e−2 s e−3s
3. L  f (t ) = 2 − 2 + 2 − 2 .
s s s s
Answers:
1. f (t ) = 3H (t ) − 4(t − 1) H (t − 1) + 5(t − 2) H (t − 2);
2. f (t ) = 2 H (t ) − 5H (t − 1) + 8H (t − 2);
3. f (t ) = tH (t ) − (t − 1) H (t − 1) + (t − 2) H (t − 2) − (t − 3) H (t − 3).

Further Exercises:
1. A function is defined by the statement
3, 0  t  1,
f (t ) = 
 t, 1  t.
Sketch the graph of the function and find its Laplace transform

f (t )
3, 0  t  1,
f (t ) = 
 t, t  1.
3

0 1 2 3 t

Note the following:

i) Between t = 0 and t = 1, f (t ) = 3, a constant.
ii) From t = 1, onwards f (t ) = t.

46 EM
Hence
f (t ) = 3H (t ) − 3H (t − 1) + tH (t − 1). (1.1)
In other words, f (t ) = 3 is switched off at t = 1 , while f (t ) = t is switched on at t = 1.
To find the Laplace transform of (1.1), we first note that
t = (t − 1) + 1 (1.2)
Using this equivalent expression in equation (1.1), we get
f (t ) = 3H (t ) − 3H (t − 1) + (t − 1) H (t − 1) + H (t − 1)
(1.3)
= 3H (t ) − 2 H (t − 1) + (t − 1) H (t − 1).
Taking Laplace transforms of the various terms on the right-hand side, we have
3 2e− s e− s
L  f (t ) = 3L H (t ) − 2L H (t − 1) + L (t − 1) H (t − 1) = − + 2 ..
s s s

2. A function is defined by the statement

t 2 , 0  t  2,

f (t ) =  4, 2  t  5,
 0, t  52.

Sketch the graph of the function and find its Laplace transform

t 2 , 0  t  2,
f (t ) 
f (t ) =  4, 2  t  5,
 0, t  5.

4

0 t
1 2 3 4 5

Note the following:

iii) Between t = 0 and t = 2, f (t ) = t 2 , a quadratic.
iv) Between t = 2 and t = 5, f (t ) = 4, a constant,
v) From t = 5 onwards, f (t ) = 0, a constant.

47 EM
Hence
f (t ) = t 2 H (t ) − t 2 H (t − 2) + 4 H (t − 2) − 4 H (t − 5). (1.4)
To find the Laplace transform, we first note that
t 2 = (t − 2)2 + 4(t − 2) + 4 (1.5)
Using this equivalent expression in equation (1.1), we get
f (t ) = t 2 H (t ) − (t − 2) 2 H (t − 2) − 4(t − 2) H (t − 2) − 4H (t − 5). (1.6)
Taking Laplace transforms of the various terms on the right-hand side, we have
2 2e−2 s 4e−2 s 4e−5 s
L  f (t ) = 3 − 3 − 2 − .
s s s s

Special Consideration of H (t − t0 )

f (t )

(t − t0 ) H (t − t0 )

t
t0
For any function f (t ) , we have
 f (t ), t  t0 ,
f (t ) H (t − t0 ) = 
 0, t  t0 ,

Periodic Functions
Periodic Functions may be represented by a series of terms involving Heaviside Unit Step
Functions. Fpr example,


  H (t − 2n) − H (t − 2n − 1) =  H (t ) − H (t − 1) +  H (t − 2) − H (t − 3) +

n =0

Recall that
e− st0
L  H (t − t0 ) =
s

48 EM
 f (t )

Theorem: If f (t + nT ) = f (t ) for all integers n , then f (t ) has the period T and

T
1
L  f (t ) = − sT 
e− st f (t )dt.
1− e 0
Proof:

f (t )

f (t ) f (t1 )

t
t T t1 2T 3T

Because of periodicity f (t ) = f (t1 ) , where t = t1 − T .

By definition

L  f ( t ) ) =  e− st f (t )dt , s  0. (1.7)
0

Since when t = 0 , t1 = T , and when t → , t1 → , in terms of t1 we have

 
L  f ( t ) ) =  e − s (t1 −T ) f (t1 )dt1 = e sT  e − st1 f (t1 )dt1
T T
In other words,

e − sT
L  f ( t ) ) =  e − st1 f (t1 )dt1 (1.8)
T
Subtracting (1.8) from (1.7), we have

49 EM
   T
1 − e − sT
 L  f ( t ) ) =  e − st
f (t )dt −  e − st1
f (t1 )dt1 =  e− st f (t )dt (1.9)
0 T 0

Thus,
T
L  f ( t ) ) =
1
− sT 
e− st f (t )dt (1.10)
1− e 0
for a periodic function of period T , as asserted above.

Example: Consider the square wave function below, defined by the expression
 1, 2n  t  2n + 1,
f (t ) =  n = 0,1, 2,
−1, 2n + 1  t  2n + 2,
with a period of T = 2.

f (t )

0 t
1 2 3 4

−1

Using the above theorem

1  − st 
2 1 2
L  f ( t ) ) =
1
1 − e −2 s 0 1 − e −2 s  0 1
− st − st
e f (t ) dt =  e f (t ) dt − e f (t ) dt 

1 1 − e − s e − s − e −2 s  (e −2 s − 2e − s + 1)
=  −  = −
1 − e −2 s  s s  s (1 − e − s )(1 + e − s ) (1.11)
(e − s − 1) 2 (1 − e − s ) (e s / 2 − e − s / 2 )
=− = =
s (1 − e − s )(1 + e − s ) s (1 + e − s ) s (e s / 2 + e − s / 2 )
1 s
= tanh .
s 2

50 EM
Example: Consider the sawtooth wave function below, defined by the expression
f (t ) = t / 2, 0  t  3; f (t ) = f (t + 3).
with a period of T = 3.

f (t )

0 t
3 6 9 12

Using the Theorem established above again, we have

3
 3 − st 
L  f ( t ) ) =
1 1
1 − e−3s 0 2(1 − e−3s )  0
− st
e f (t ) dt = −  te dt  .

Hence
3 
2(1 − e −3 s )L  f ( t ) ) =   te − st dt  .
0 
Integrating the right-hand side by parts, we get
2(1 − e −3s )L  f ( t ) ) = 2 1 − 3se3s − e −3s  .
1
s
Therefore
L  f ( t ) ) = 2
1
1 − 3se3 s − e −3 s  .
−3 s 
2 s (1 − e )
However, on dividing by (1 − e−3s ) , we get
 3s 
L  f ( t ) ) =
1
2s 2 1 − e3 s − 1  .

51 EM
THE IMPULSE FUNCTION

We have met the Heaviside Unit Step Function, which provides us with the means of
expressing the switching on and off items in the physical problem.

We now consider the impulse function, representing an extremely large force acting for
acting for a minutely small interval of time, or a “shock” voltage applied to an electrical
circuit. It is equivalent to a sudden blow, rather than a continuous force.
f (t )

1
b

0 a t
b
1
Figure: A Rectangular Pulse of Width b and Height
b
1 b
The are of the pulse is b  = 1. Now, if the width of the pulse is reduced to , and the
b 2
2
value of the area maintained constant (unity), then the height becomes . Evidently, as
b
1
the with b → 0, the height → . In such a case the function exists for a minutely small
b
interval, but during that interval, the magnitude of the function is extremely large. The
function represented by the limiting stage of this process is called the Dirac Delta
Function or the Unit Impulse Function, and is denoted by  (t − a) .

Graphically, it would require a rectangular pulse of zero width and infinite height, and for
convenience, is represented by a single vertical line with an arrow at its head.

 (t − a)

0 a t
Figure: The Dirac Delta Function

52 EM
Note that in all cases and at all times the area of the pulse is unity.
Mathematically, the Dirac Delta Function is defined as below:
 , t = a ,
 (t − a) = 
 0, t  a,
For a Dirac Delta Function centred at the origin, we have
, t = 0,
 (t ) = 
 0, t  0,
Schematically, it is depicted in the Figure below.

 (t )

0 t
Figure: The Dirac Delta Function Centred
at the Origin

Properties of Dirac Delta Function

q a q

1.   (t − a)dt =   (t − a)dt +   (t − a)dt + (Area at t = a) = 0 + 0 + (Area at t = a) = 1.

p p a
q

In other words,   (t − a )dt = 1, provided p  a  q.

p

q a q

2.  f (t ) (t − a)dt =  f (t ) (t − a)dt +  f (t ) (t − a)dt + (Area at t = a) = f (a).

p p a
q

In other words  f (t ) (t − a)dt =

p
f (a ), provided p  a  q. This is sometimes

referred to as the Filtering Property of the Dirac Delta Function. In obtaining the
result we invoke the fact that the integrand f (t ) (t − a) is non-zero only at t = a.

53 EM
 3. Laplace Transform of The Dirac Delta Function

If p = 0 and q = , we have  f (t ) (t − a)dt = f (a),
0
a  0, 0  a  . If then


e  (t − a)dt = e− sa , a  0, 0  a  .
− st
we set f (t ) = e − st , there follows that
0

Hence,
L  (t − a) = e− sa , a  0, 0  a  .
Consequently,
L  (t ) = 1.


4. L  f (t ) (t − a) =  e− st f (t ) (t − a)dt = f (a)e− sa .
0

To demonstrate this we note that e− st f (t ) (t − a) = 0, except at t = a, where

e st = e− sa and f (t ) = f (a ). Considering these latter values as constants at
t = a, we may thus pull them out and have instead
 

 e f (t ) (t − a)dt = f (a)e   (t − a)dt = f (a)e ,

− st − sa − sa

0 0

as asserted. In particular, if a = 0, then we have



e f (t ) (t )dt = f (0).
− st

Example 1: Impulses of magnitudes 1, 2 and 4 units occur at t = a, t = 2a, and t = 3a,

respectively, as depicted in the sketch below.
4   (t − 3a )

1   (t − a )

0 t

(−2)   (t − 2a)

54 EM
The function f (t ) we have here is given by
f (t ) =  (t − a) − 2 (t − 2a) + 4 )t − 3a).
Taking the Laplace transform, we then have
L  f (t ) = e− sa − 2e−2 sa + 4e−3sa .

5. Relationship Between the Dirac Delta and the Heaviside Unit Step Functions.

Recall that
t
0, t  a,
  ( − a)d = 1,
0
t  a.
But the right-hand side is precisely the Heaviside Unit Step Function! Therefore,
t

  ( − a)d = H (t − a).
0

Differentiating both sides, we find that

d
 (t − a) = H (t − a ).
dt
In particular,
d
 (t ) = H (t ).
dt

Examples:
1. L 5 (t − 3) = 5e−3s ,
2. L t 2 (t − 2) = 4e−2 s ,
3. L cos 2t (t −  ) = cos 2 e− s = e− s ,
4. L sinh t (t ) = sinh 0  e−0 = 0.

Summary:
q

1.   (t − a)dt = 1,
p
p  a  q.

2.  f (t ) (t − a)dt = F (a),
p
provided p  a  q.

3. L  (t − a) = f (a).
4. L  (t ) = 1.
5. L  f (t ) (t − a) = f (a)e− sa .

55 EM
Some Illustrative Examples
Example 1: An impulsive voltage E (t ) is applied to a circuit containing an
inductance L and capacitance C in series. If current and charge are
initially zero, find the expression for the current at time t .
di
VL = L
dt

E (t ) q
VC =
C

Let q be the charge on the capacitor at time t , whereby the current through the capacitor
is given by
dq
i= .
dt
The voltage across the inductance is of course given by
di
VL = L .
dt
As the total drop in voltage in the circuit must equal the applied voltage, the differential
equation governing the voltage in the circuit is given by
di q
VL + VC = L + = E (t ).
dt C
Expressing this in Laplace transform, we have
Q
sLI − Li (0) + = E.
C
dq I q(0)
As i = , there follows that I = sQ − q(0) or + = Q. Accordingly, he Laplace
dt s s
transform of the circuit voltage equation becomes
I q(0)
sLI − Li(0) + + = E.
sC sC
With i (0) = 0 and q (0) = 0 , this equation degenerates into
 1 
 sL +  I = E.
 sC 
The transform of the current is thus given by
s E s
I = E0 = 0 .
1 L 2 1
s L+
2
s +
C LC
The desired current is thus given by
E 1
i (t ) = 0 cos t.
L LC

56 EM
CONVOLUTION
Consider a linear system in which the effect at the present time t of a stimulus f ( ) at
any past time  is proportional to the stimulus. On physical grounds we assume that the
proportionality constant depends on the elapsed time t −  and hence has the
form g (t −  ) .

Input (Stimulus) Output (Response)

System

The effect at the present time t is therefore

f ( ) g (t −  ) . (1.1)
Since the system is linear the response to the whole past history can be obtained by
adding these separate effects, and we are led to the integral
t
h(t ) =  f ( ) g (t −  )d . (1.2)
0

The lower limit of this integral is 0 because we assume that the process started at time
t = 0; in other words, f ( ) = 0 for   0.

Expression (1.2) is called the convolution of f and g . It gives the response at the present
time as the weighted superposition over the inputs at the times   t. The weighting
factor g (t −  ) characterises the system, and f ( ) characterises the past history of the
input. Because of this physical interpretation, the convolution is sometimes called the
superposition integral.

The importance of convolution can hardly be overestimated. It plays a prominent role in

the study of heat conduction, wave motion, plastic flow and creep, and other areas of
mathematical physics modelled by partial differential equations. The convolution is also
encountered in several branches of mathematics, for example, a large part of the theory of
distributions can be based on expression (1.2) rather than on the concept of linear
functional.

The function h in expression (1.2) is often denoted by f  g , so that f  g (t ) = h(t ) is the

value of f  g at the argument t . A similar notation applies in other cases; for example,
t
(a  b)(t ) =  a( )b(t −  )d (1.3)
0

Note that the product  behaves very much like the normal multiplication, in that
a  (b + c) = a  b + a  c, a  b = b  a, (a  b)  c) = a  (b  c).

57 EM
We now show that convolution in the time domain actually corresponds to multiplication
in the frequency domain.

Theorem: If f  g is defined, then

L  f  g = L  f  L  g = F (s)G(s).

Proof: Let F ( s) = L  f  and G( s) = L  g . From the general definition of Laplace

transform, we have
 
F ( s) =  e− st f (t )dt and G(s) =  e− s g ( )d  .
0 0

We note that in terms of Heaviside function we can re-write these definitions as

 

e e
− s − s
F (s) = H ( ) f ( )d and G ( s ) = H ( ) g ( )d  ,
− −

wherein we see that the lower limits of integration have now shifted from 0 to  . With
these expressions, we thus find that the product of the Laplace transforms of f (t ) and
g (t ) may be expressed as

  − s    − s 
F ( s)G ( s) =   e H ( ) f ( )d    e H ( ) g ( )d   .
−  − 
The separated integrals on the right may of course be expressed in the form of a double
integral as
 

 e
− s ( + )
F ( s )G ( s ) = H ( ) H ( ) f ( ) g ( )d  d .
− −

Re-organising the sequence of integration in this double integral, we may recast it as


 
F ( s)G ( s) = 
−
f ( ) H ( )   e − s ( + ) H ( ) g ( ) d  d .
 − 
In the integral in the brackets suppose we now change the variable of integration from 
to t through setting  +  = t , so that  = t −  . Then the product of the two Laplace
transforms becomes

  − st 
F ( s)G ( s) = 
−
f ( ) H ( )   e H (t −  ) g (t −  )dt d .
 − 
On changing the order of integration, we may write

 
−  −     
− st
F ( s)G ( s) = e  f ( ) H ( ) H (t − ) g (t − ) d dt .

Now we know that because of the definition of the Heaviside function, we know that

58 EM
 1,   0,
H ( ) = 
 0,   0,
1, t −   0, or   t ,
H (t −  ) = 
0, t −   0, or   t.
as can be deduced from the schematic representation.

H ( )

t 

H (t −  )

t 

H ( ) H (t −  )

t 

The integrand of the integral in the brackets is thus zero everywhere, except over the
interval 0    t , since

 f ( ) g ( ), 0    t ,
f ( ) H ( ) H (t −  ) g (t −  ) = 
0, elsewhere.
Hence we may express the non-vanishing portion as

59 EM
 
t 
−  0 f ( ) H ( ) H (t −  ) g (t −  )d dt .
− st
F ( s)G ( s) = e

However, considering further that by virtue of the definition of the Heaviside function,
H (t ) H ( − t ) g ( − t ) = 1, 0    t ,
the product of the two Laplace transforms may actually be written as

t 
−  0 f ( ) g (t −  )d dt .
− st
F ( s)G ( s) = e

Recognising the integral in the brackets to the function h(t ), derived in expression (1.2),
we thus have

F ( s )G ( s ) = e
− st
h(t )dt = L h(t ) = L  f  g .
−

Note that the interchanges of order of integrals are justified because we assume that the
conditions for the existence of the Laplace transforms of the functions f and g are in
place. It thus follows that
t
L−1 F ( s)G( s) = f  g =  f ( ) g (t −  )d .
0

In other words, if a given Laplace transform is expressible as the product of two

components and the inverse Laplace transforms of the components are known, then the
inverse Laplace transform of the product is obtainable as the convolution integral of the
inverse transforms of the components.

The convolution is often used in solving integral equations where the unknown function
to be determined is under the integral sign, as illustrated in the following example.
t
Example: Solve the integral equation y (t ) = 3t +  y ( )sin(t −  )d .
0

Solution: We first note that the integral can be re-written as

y (t ) = 3t + y (t )  sin t.
Taking the Laplace transforms, we have
3 Y ( s)
Y (s) = 2 + 2 .
s s +1
Solving for Y ( s ), we get
3( s 2 + 1) 3 3
4
Y ( s) =
= 2 + 4
s s s
Then taking the inverse Laplace transform, we have
1
y (t ) = 3t + t 3 .
2

60 EM
The convolution theorem is often useful in obtaining solutions to differential equations in
which there are functions whose Laplace transforms are difficult or even impossible to
find, as illustrated by the following example.

Solve the initial value problem y  + y = et , y(0) = 0, y (0) = 0.

2
Example:

Setting f (t ) = e−t in the differential equation and taking the Laplace

2
Solution:
transform, we have
s 2Y ( s) − sy (0) − y (0) + Y ( s) = F ( s).
Applying the initial conditions, this expression reduces to
( s 2 + 1)Y ( s) = F ( s),
so that
1
Y ( s) = 2 F ( s).
s +1
Viewing the right-hand side as the product of two Laplace transforms, and taking the
inverse transform of both side we have
t
y (t ) =  e− sin(t −  )d
2

Although this integral cannot be evaluated exactly, it can be handled numerically. Hence
application of the convolution theorem has led to an alternative route of solving the initial
value problem.

It is worthwhile to note that if the right-hand side of the differential equation in the above
initial value problem had been, say, et instead of e − t , the Laplace transform would not
2 2

exist, but the solution without application of Laplace transform is in fact

t
y (t ) =  e sin(t −  )d .
2

Thus, as in the above example, formally the method of Laplace transform can be used to
arrive at possible solutions which can then be checked.

The fact that Laplace transform techniques can lead to correct results even in cases where
functions do not have Laplace transforms seems to indicate that there is something more
basic than the Laplace transform, but still closely related to it, which can be used in its
place. From the above example it would seem that the convolution itself is the desired
concept. This is indicated further by the fact that the convolution obeys many of the usual
laws of algebra, such as the Commutative, Associative and Distributive Laws. This has
thus lead some authors to avoid Laplace transforms altogether and deal only with
convolutions. By use of this procedure it is possible to make rigorous impulse functions
such as the Dirac Delta Function.

61 EM
INITIAL AND FINAL VALUE THEOREMS
The Initial Value Theorem:

This theorem states that

lim f (t ) = lim sF ( s)
t →0 s →

Recall that

L  f (t ) =  f (t )e− st dt = sF ( s) − f (0)
0

However, if f (t ) is continuous piecewise and of exponential order, meaning there are
real constants M and  , such that f (t )  Me− t , then

lim  f (t )e− st dt = 0,
s →
0

because lim e − st = 0. It then follows that lim sF (s) − f (0) = 0 or

s → s →

lim sF (s) = f (0) = lim f (t ).

s → t →0

The Final Value Theorem:

This theorem states that

lim f (t ) = lim sF ( s)
t → s →0

Again recall that



 f (t )e
− st
dt = L  f (t ) = sF ( s) − f (0)
0

Considering the left-hand side, we have

 
lim  f (t )e dt =  f (t )dt ,
− st
s →0
0 0
− st
because lim e = 1. But, we also note that
s →0
 p


0
f (t )dt = lim  f (t )dt.
p →
0

Upon integrating the right-nad side we have



 f (t )dt = lim  f ( p) − f (0) = lim f (t ) − f (0) .

0
p → t →

Hence,
lim  sF ( s) − f (0)  = lim  f (t ) − f (0) .
s →0 t →

and so we have
lim sF (s) = lim f (t ).
s →0 t →

62 EM
THE SINE, COSINE AND EXPONENTIAL INTEGRALS

The Sine Integral of t , written as Si(t ) , is given by the expression

t
sin u
Si(t ) =  du .
0
u
Its Laplace transform is thus given by
 t sin u  1 −1 1
L Si(t ) = L   du  = tan ..
0 u  s s
t
sin u
Proof: Let f (t ) =  du. Then quite clearly
0
u
sin t
f (0) = 0 , f (t ) = and tf (t ) = sin t.
t
Invoking the formula for the Laplace transform of a derivative, we get
d
L tf (t ) = −  sF ( s ) − f (0) 
ds
Therefore, with
L tf (t ) = L sin t ,
there follows that
d d 1
 sF ( s) − f (0) =  F ( s) = − 2
ds ds s +1
Hence
ds
sF ( s ) = − 
= − tan −1 s + C
s +1 2

where C is an arbitrary constant of integration.

Invoking the Initial Value Theorem, we have


lim sF ( s) = lim f (t ) = f (0) = − lim tan −1 s + C = − + C = 0.
s → t →0 s → 2

Therefore, C = . Hence
2
 1
sF ( s ) = − tan −1 s = tan −1 .
2 s
Accordingly,
1 1
F ( s ) = tan −1 = L Si(t )
s s
as asserted.

The Cosine Integral of t , written as Ci(t ) , is given by the expression


cos u
Ci(t ) =  du .
t
u

63 EM
Its Laplace transform is thus given by
 cos u  ln( s 2 + 1)
L Ci(t ) = L   du  = ..
t u  2 s

cos u
Proof: Let f (t ) =  du. Then quite clearly
t
u
cos t
f () = 0 , f (t ) = − and tf (t ) = − cos t.
t
Invoking the formula for the Laplace transform of a derivative, we get
d
L tf (t ) = −  sF ( s ) − f (0) 
ds
Therefore, with
L tf (t ) = −L cos t ,
there follows that
d d s
 sF ( s) − f (0) =  F ( s) = 2
ds ds s +1
Hence
sds 1
sF ( s ) = 
= ln( s 2 + 1) + C
s +1 2
2

where C is an arbitrary constant of integration.

Invoking the Final Value Theorem, we have

1
lim sF ( s) = lim f (t ) = f () = ln1 + C = 0.
s →0 t → 2
Therefore, C = 0. Hence
1
sF ( s ) = ln( s 2 + 1)
2
Accordingly,
ln( s 2 + 1)
F (s) = = L Ci(t )
2s
as asserted.

The Exponential Integral of t , written as Ei(t ) , is given by the expression


e−u
Ei(t ) =  du .
t
u
Its Laplace transform is thus given by
  e−u  ln( s + 1)
L Ei(t ) = L   du  = ..
t u  s

64 EM
 
e−u
Proof: Let f (t ) =  du. Then quite clearly
t
u
e−t
f () = 0 , f (t ) = − and tf (t ) = −e−t .
t
Invoking the formula for the Laplace transform of a derivative, we get
d
L tf (t ) = −  sF ( s ) − f (0) 
ds
Therefore, with
L tf (t ) = L e−t .
there follows that
d d 1
 sF (s) − f (0) =  F (s) =
ds ds s +1
Hence
ds
sF ( s ) = 
= ln( s + 1) + C
s +1
where C is an arbitrary constant of integration.

Invoking the Final Value Theorem, we have

lim sF ( s) = lim f (t ) = f () = ln1 + C = 0.
s →0 t →

Therefore, C = 0. Hence
sF ( s ) = ln( s + 1)
Accordingly,
ln( s + 1)
F ( s) = = L Ei(t )
s
as asserted.

65 EM