CHAPTER 6

Laplace Transform Methods

6.1 INTRODUCTION
Laplace transform is essentially a mathematical tool which can be used to solve several
problems in science and engineering. This transform was first introduced by Laplace, a
French mathematician, in the year 1790 in his work on probability theorem. This technique
became popular when Heaviside applied to the solution of an ordinary differential equation
referred hereafter as ODE, representing a problem in electrical engineering. To the basic
question as to why one should learn Laplace transform technique when other techinques are
available, the answer is very simple. Transforms are used to accomplish the solution of certain
problems with less effort and in a simple routine way. To illustrate, consider the problem of
finding the value of x from the equation
x1.85  3 (6.1)
It is an extremely tedious task to solve this problem algebraically. However, taking logarithms
on both sides, we have the transformed equation as
1.85 ln x  ln 3 (6.2)
In this transformed equation, the algebraic operation and exponentiation have been changed
to multiplication which immediately gives
ln 3
ln x 
1.85
To get the required result, it is enough if we take the antilogarithm on both sides of the above
equation, which yields

⎛ ln 3 ⎞
x = ln −1 ⎜ ⎟
⎝ 1.85 ⎠
With the help of any ordinary calculator, we can now compute x. Following this simple
example, the Laplace transform method reduces the solution of an ODE to the solution of an
316
 LAPLACE TRANSFORM METHODS 317

algebraic equation. In fact, this method has a particular advantage in finding the solution of
an ODE with appropriate ICs, without first finding the general solution and then using ICs
for evaluating the arbitrary constants. Also, when the Laplace transform technique is applied
to a PDE, it reduces the number of independent variables by one.

Definition 6.1 Suppose f (t ) is a piecewise continuous function and if it has an additional

property that there exists a real number H 0 and a finite positive number M such that
Lt | f (t ) | eH t c M for H  H 0
t nf

and the limit does not exist when γ < γ 0 , then such a function is said to be of exponential
order H 0 , also written as

| f (t ) |  0(eH 0t )
Variables such as velocity and current are always finite; which means that f (t ) is bounded.
Thus for any bounded function f (t ), | f (t ) | eH t n
0 for all H  0. The order of such a function
is zero. However, variables such as electrical charge and mechanical displacement may increase
without limit but of course proportional to t. Such functions are also of exponential order.
For illustration, let us consider the following examples:
(i) Lt teH t  0
t nf

The fact that t n is of exponential order zero can be seen as follows:

¦ tn µ ¦ nt n 1 µ
Lt t n eH t  Lt § ¶  Lt § ¶ (using L’Hospital’s rule)
t nf t n f § eH t ¶ t nf § Ht ¶
¨ · ¨ He ·

Applying the L’Hospital’s rule repeatedly, we get

¦ n! µ
Lt t n eH t  Lt ¶0
t nf t n §
f § H n eH t
¨
¶
·

(ii) In an unstable system a function may increase as eat and we can see that

Lt eat eH t  0 if H  a
t nf

Thus the function eat is of exponential order a.

(iii) exp (t n ) (n  1) is not of exponential order, since

Lt exp (t n ) eH t  Lt exp [t (t n 1  H )]  f

t nf t nf

for any finite value of H .

318 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Definition 6.2 Let f (t ) be a continuous and single-valued function of the real variable t
defined for all t , 0  t  f, and is of exponential order. Then the Laplace transform of f (t )
is defined as a function F ( s ) denoted by the integral
f
L [ f (t ); s ]  F ( s )  ± e st f (t ) dt (6.3)
0

over that range of values of s for which the integral exists. Here, s is a parameter, real or
complex. Obviously, L[ f (t ); s ] is a function of s. Thus,

L[ f (t ); s ] = F ( s)

f (t ) = L−1[ F ( s ); t ]

where L is the operator which transforms f (t ) into F ( s ) , called Laplace transform operator,
and L1 is the inverse Laplace transform operator.
The Laplace transform belongs to the family of “integral transforms’’. An integral transform
F ( s ) of the function f (t) is defined by an integral of the form
b
∫ a k (s, t ) f (t ) dt = F (s) (6.4)

where k ( s, t ), a function of two variables s and t, is called the kernel of the integral transform.
The kernels and limits of integration for various integral transforms are given in Table 6.1
(which is not exhaustive).

Table 6.1 Kernels and Limits for Various Integral Transforms

Name of the transform k ( s, t ) a b

Laplace transform e st 0 e
Fourier transform eist / 2π e e
2
Fourier sine transform
Q
sin st 0 e
2
Fourier cosine transform
Q
cos st 0 e
Hankel transform tJ n ( st ) 0 e
Mellin transform t s 1 0 e
The integral transforms defined above are applicable, either for semi-infinite or infinite
domains. Similarly, finite integral transforms can be defined on finite domains.
Now, we are in a position to verify the following important result.
 LAPLACE TRANSFORM METHODS 319

Theorem 6.1 If f (t ) is piecewise continuous in the range t s 0 and is of exponential order

H, then the Laplace transform F ( s ) of f (t ) exists for all s  H .
Proof From the definition of Laplace transform,
f T f
L[ f (t ); s]  ± e st f (t ) dt  ± e st f (t ) dt  ± e st f (t ) dt  I1  I 2
0 0 T

Since f (t ) is piecewise continuous on every finite interval 0  t  T , I1 exists, whereas

f
| I 2 | c ± | e st f (t ) | dt
T

But f (t ) is a function of exponential order; therefore,

| f (t ) |  MeH t for H real
Hence,
| e  st f (t ) |  Me( s H ) t
Thus,
f  ( s H ) t Me ( s H )T
| I2 |  ± e M dt  , s H
T s H

In other words, I 2 can be made as small as we like provided T is large enough and, therefore,
I 2 exists. Hence, L[ f (t ); s ] exists for s  H .

6.2 TRANSFORM OF SOME ELEMENTARY FUNCTIONS

Following the definition of Laplace transform by the integral (6.3), we shall compute the
Laplace transform of some elementary functions.

EXAMPLE 6.1 Find the Laplace transform of

(i) 1, (ii) 0, (iii) t, (iv) eat, (v) e–at.
Solution Using the definition of Laplace transform, we have
f
f  st ¦ e st µ 1
(i) L [1; s ]  ± e (1) dt § ¶  if s  0
0 § s ¶ s
¨ ·0
f
(ii) L [0; s ]  ± e st (0) dt  0
0
f
f  st f ¦ e  st µ ¦ e st µ f e  st 1
(iii) L [t; s ]  ± e —t dt  ± td § ¶  §t ¶ ± dt 
§ s ¶ § s ¶
0 0
¨ · ¨ ·0
0 s s2
320 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

f
at f  st at f  ( s  a )t © e  ( s  a )t ¸ 1
(iv) L [e ; s ]  ± e e dt  ± e dt ª ¹  , sa
0 0 ª« ( s  a) ¹º 0 sa
f f 1
(v) L [e at ; s]  ± e st e at dt  ± e( s  a )t dt 
0 0 sa

EXAMPLE 6.2 Find the Laplace transform of

(i) cos at, (ii) sin at.
Solution Following the definition of Laplace transform, we have
f f
(i) L [cos at; s]  ± e st cos at dt  Re ± e st eiat dt  Re L[eiat ; s]
0 0

1 s + ia s
Re = Re 2 = 2
s − ia s +a 2
s + a2
s + ia a
(ii) L [sin at; s] = Im L [eiat ; s ] = Im 2 2
=
s +a s + a2
2

EXAMPLE 6.3 Find the Laplace transform of

(i) cosh at, (ii) sinh at.
Solution Using the results established in Example 6.1, we have
⎡ eat + e− at ⎤ 1
(i) L [cosh at; s ] = L ⎢ ; s ⎥ = {L[eat ; s ] + L[e− at ; s ]}
⎢⎣ 2 ⎥⎦ 2

1⎛ 1 1 ⎞ s
= ⎜ + ⎟=
2 ⎝ s − a s + a ⎠ s2 − a2

⎡ eat − e− at ⎤ 1
(ii) L [sinh at ; s ] = L ⎢ ; s ⎥ = {L[eat ; s ] − L[e− at ; s ]}
⎢⎣ 2 ⎥⎦ 2

1⎛ 1 1 ⎞ a
= ⎜ − ⎟=
2 ⎝ s − a s + a ⎠ s2 − a2

EXAMPLE 6.4 Find the Laplace transform of t n , where n is a positive integer.

Solution Using the definition of Laplace transform, we have
f
f  st n f n ¦ e st µ ¦ n e  st µ n f n 1  st n f
n
L [t ; s ]  ± e t dt 
± 0 t d §§  s ¶  §t ¶  ± t e dt  ± t n 1e st dt
0 ¶ § s ¶ s 0 s 0
¨ · ¨ ·0
 LAPLACE TRANSFORM METHODS 321

Hence,
n n 1
L[t n ; s ]  L[t ; s ]
s
Similarly, we can prove the following:
n  1 n2
L[t n 1; s ]  L[t ; s ]
s
n − 2 n −3
L[t n − 2 ; s ] = L[t ; s ]
s

2
L[t 2 ; s ] = L[t ; s ]
s
1
L[t ; s ] =
s2
Therefore,
n n 1 n  2 2 1 n!
— L[t n ; s ] 
— ... — 2  n1
s s s s s s
which can be expressed in Gamma function as
n 1
L[t n ; s ] 
s n 1

6.3 PROPERTIES OF LAPLACE TRANSFORM

We present a few important properties of the Laplace transform in the following theorems
which will enable us to find the Laplace transform of a combination of functions whose
transforms are known.
Theorem 6.2 (Linearity property). If c1 and c2 are any two constants and if F1 ( s ) and F2 ( s )
are the Laplace transforms, respectively of f1 (t ) and f 2 (t ), then

L[{c1 f1 (t )  c2 f 2 (t )}; s ]  c1 L[ f1 (t ); s ]  c2 L[ f 2 (t ); s ]  c1F1 ( s )  c2 F2 ( s )

Proof Following the definition of Laplace transform, we have

f  st
L[{c1 f1 (t )  c2 f 2 (t )}; s ]  ± e {c1 f1 (t )  c2 f 2 (t )} dt
0

f f
 c1 ± e st f1 (t ) dt  c2 ± e st f 2 (t ) dt
0 0

 c1L[ f1 (t ); s ]  c2 L[ f 2 (t ); s ]

 c1F1 ( s )  c2 F2 ( s )
322 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Theorem 6.3 (Shifting property). If a function is multiplied by eat, the transform of the
resultant is obtained by replacing s by ( s  a) in the transform of the original function. That
is, if
L[ f (t ); s ]  F ( s )
then

L[eat f (t ); s ]  F ( s  a)

Proof From the definition of Laplace transform,

f f
L [eat f (t ); s ]  ± e st eat f (t ) dt  ± e( s a ) t f (t ) dt  F ( s  a)
0 0

Similarly,

L[e  at f (t ); s ]  F ( s  a ) (6.5)

Theorem 6.4 (Multiplication by power of t). If

L[ f (t ); s ]  F ( s )
then
dn
L[t n f (t ); s ]  (1)n n
F ( s )  (1)n F ( n) ( s )
ds
where n = 1, 2, 3, …

Proof From the definition of Laplace transform

f
F ( s)  L [ f (t ); s ]  ± e st f (t ) dt
0

Hence,
d d ⎡ f − st ⎤
ds
[ F ( s )] = ⎢
ds ⎣ ∫0 e f (t ) dt ⎥
⎦
Interchanging the operations of differentiation and integration for which we assume that the
necessary conditions are satisfied, and since there are two variables s and t, we use the notation
of partial differentiation and obtain
d f w − st f − st

ds
[ F ( s )] = ∫0 ws
{e f (t )} dt = − ∫0 e tf (t ) dt = − L[tf (t ); s ]

Therefore,
d
L[tf (t ); s ]   F (s)
ds
 LAPLACE TRANSFORM METHODS 323

By repeated application of the above result, it can be shown that

dn
L[t n f (t ); s] = (−1)n F ( s ) = (−1)n F ( n) ( s ) (6.6)
ds n

Theorem 6.5 (Differentiation property). If

L [ f (t ); s ] = F ( s )
then

L [ f ( n ) (t ); s ] = s n F ( s ) − s n −1 f (0) − s n − 2 f ′ (0) − L − sf n − 2 (0) − f n −1 (0)

Proof From the definition of Laplace transform, we have

∞ ∞
∫0 e ∫0 e
− st
L [ f ′ (t ); s ] = f ′ (t ) dt = [e− st f (t )]0∞ + s − st
f (t ) dt

= − f (0) + sL [ f (t ); s ] = sF ( s ) − f (0)

Similarly, it can be shown that

L[ f ′′ (t ); s ] = sL[ f ′ (t ); s ] − f ′ (0) = s{sF ( s ) − f (0)} − f ′ (0) = s 2 F ( s ) − sf (0) − f ′ (0)

L[ f ′′′ (t ); s ] = s 3 F ( s ) − s 2 f (0) − sf ′ (0) − f ′′ (0)

Thus, in general,

L[ f ( n ) (t ); s ] = s n F ( s ) - s n -1 f (0) - s n - 2 f ¢(0) - s n -3 f ¢¢(0) - L - f ( n -1) (0) (6.7)

This property is very useful for solving differential equations.

EXAMPLE 6.5 Find the Laplace transform of

(i) eat cos bt , (ii) e at sin bt , (iii) eat cosh bt , (iv) eat t n , and (v) cos at cosh bt.

Solution Using the shifting property

s s−a
(i) L[eat cos bt ; s ] = 2 2
=
s +b s→( s − a) ( s − a)2 + b2

b b
(ii) L[e at sin bt ; s ] = 2 2
=
s +b s→( s− a) ( s − a)2 + b2

s s−a
(iii) L[eat cosh bt ; s ] = 2 2
=
s −b s→( s − a) ( s − a)2 − b 2
324 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

n! n!
(iv) L[e at t n ; s ] n 1
s so( s  a ) ( s  a ) n 1

ª § ebt  ebt · º
(v) L[cos at cosh bt ; s ] L « cos at ¨ ¸¹ ; s »
¬ © 2 ¼
1
{L[ebt cos at ; s ]  L[ebt cos at ; s ]}
2

1­ sb sb ½

2 ®¯ ( s  b)2  a 2 ( s  b)2  a 2 ¾¿

EXAMPLE 6.6 Fing the Laplace transform of the following:

(i) t 2 eat , (ii) t sin at , (iii) t 2 cos at , (iv) t n e  at .

Solution Using the result established in Theorem 6.4, we have

d2 d2 § 1 · d § 1 · 2
(i) L[t 2 eat ; s ] (1)2 L[eat ; s ] ¨ ¸
ds 2
ds 2 © s  a ¹ ds ¨© ( s  a)2 ¹¸ ( s  a )3
Alternatively,
2! 2
L[eat t 2 ; s ] 3
(using the shifting property)
s so( s  a) ( s  a )3

d d § a · 2as
(ii) L[t sin at; s ] (1)1 L[sin at; s ]  ¨ 2
ds ds © s  a 2 ¸¹ ( s  a 2 )2
2

d2 d2 § s ·
(iii) L[t 2 cos at ; s ] (1)2 L[cos at ; s ] ¨ ¸
ds 2 ds 2 © s 2  a 2 ¹

d ª a2  s 2 º 2s3  6sa 2
« »
ds «¬ ( s 2  a 2 )2 »¼ ( s 2  a 2 )3

n! n!
(iv) L[e at t n ; s ] n 1
(using the shifting property)
s so( s  a) ( s  a )n 1

EXAMPLE 6.7 Find the Laplace transform of

(i) te 4t sin 3t , (ii) sin 2t sin 3t , (iii) sin 3 2t.
 LAPLACE TRANSFORM METHODS 325

Solution We may note that

(i) L [te 4t sin 3t ; s ] L [e4t (t sin 3t ); s ]

Using the result of Theorem 6.4, we get
d § 3 · 6s
L[t sin 3t; s] 
ds ¨© s 2  9 ¸¹ ( s  9)2
2

Now, using the shifting property, we obtain

6s 6( s  4) 6( s  4)
L [e4t (t sin 3t ); s ] 2 2 2 2
( s  9) s o ( s  4)
{( s  4)  9} ( s  8s  25) 2
2

1
(ii) Since sin 2t sin 3t (cos t  cos 5t ),
2
1
L[sin 2t sin 3t ; s ] {L[cos t ; s ]  L[cos 5t ; s ]}
2
1§ s s · 12 s
 2
2 © s  1 s  25 ¸¹
¨ 2
( s  1) ( s 2  25)
2

3 3 1
(iii) Since sin 6t sin 3 (2t ) 3 sin 2t  4 sin 3 2t , we have sin 2t sin 2t  sin 6t.
4 4
Thus,
3 1 3§ 2 · 1§ 6 ·
L[sin 3 2t; s ] L[sin 2t; s ]  L[sin 6t ; s ] 
4 4 4 ¨© s 2  4 ¸¹ 4 ¨© s 2  36 ¸¹
48
( s  4)( s 2  36)
2

EXAMPLE 6.8 Find the Laplace transform of f (t ) defined as

­sin t , 0 t π
f (t ) ®
¯0, t !π

Solution Using the definition of Laplace transform, we have

π  st f π  st it π (i  s ) t
³0 e ³π e ³0 e ³0 e
 st
L[ f (t ); s ] sin t dt  (0) dt Im e dt Im dt

π
ª e(i  s )t º ª e(i  s )π  1º is
Im « » Im « » Im [1  e sπ (cos π  i sin π )]
¬ i  s ¼0 ¬ is ¼ 2
s 1

is 1  e sπ
Im [1  e sπ ]
s2  1 s2  1
326 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Theorem 6.6 (Initial value theorem). If f (t ) and f c(t ) are Laplace transformable and
F ( s ) is the Laplace transform of f (t ) , then the behaviour of f (t ) in the neighbourhood of
t 0 corresponds to the bebaviour of sF ( s) in the neighbourhood of s f. Mathematically,

Lt f (t ) Lt sF ( s )
t o0 s of

Proof From the property of derivative, we have

L[ f c (t ); s ] sF ( s )  f (0) (6.8)
Taking the limit as s o f on both sides, we get
f
Lt ³
s of 0
e st f c (t ) dt Lt sF ( s )  Lt f (0)
s of s of
(6.9)

since s is independent of t, we can take the limit before integrating the left-hand side of Eq.
(6.9), thus getting
f f
ª Lt e st f c (t ) º dt
s of ³0 ³0
 st
Lt e f c (t ) dt 0
«¬ s of »¼
and Eq. (6.9) becomes
Lt sF ( s) f (0) Lt f (t )
s of to 0

Hence the result. For example, let f (t ) be a polynomial of degree n of the form

f (t ) a0  a1t  a2t 2    an t n
Its Laplace transform is
a0 a1 2a2 n !a
F (s)  2  3    n 1k
s s s s
Now, taking the limit on both sides as s o f, we obtain
Lt sF ( s ) a0 f (0)
s of

EXAMPLE 6.9 Verify the initial value theorem for the function

f (t ) 1  e t (sin t  cos t )

Solution Given f (t ) 1  e t (sin t  cos t ) , we have

 LAPLACE TRANSFORM METHODS 327

F (s) L[ f (t ); s ] L[1; s ]  L[(et sin t  et cos t ); s ]

1 § 1 s ·
¨ 2  2 ¸
s © s  1 s  1 ¹ s o( s 1)
1 § 2s ·

s ¨© s 2  2 s  2 ¸¹
Hence,

2s  s 2
sF ( s ) 1 
s 2  2s  2
Therefore,
2/ s  1
Lt sF ( s ) Lt 1  11 2
s of s of 1  2/ s  2 / s 2

But f (0) 1  1 2. Thus,

Lt sF ( s ) f (0)
s of
Hence the result.

Theorem 6.7 (Final value theorem). If f (t ) and f c (t ) are Laplace transformable and F ( s )
is the Laplace transform of f (t ), then the behaviour of f (t ) in the neighbourhood of t f
corresponds to the behaviour of sF ( s ) in the neighbourhood of s 0. Mathematically,
Lt f (t ) Lt sF ( s )
t of so0

Proof From the property of derivative, we have

L[ f c (t ); s ] sL[ f (t ); s ]  f (0)
i.e.
f
³0 e
 st
f c (t ) dt sF ( s )  f (0)

Taking the limit as s o 0 on both sides of the above equation, we have

f
Lt
s o0 0 ³ e st f c(t ) dt Lt sF ( s)  Lt f (0)
so 0 so 0

But,
f f
³0 so 0
Lt e st f c(t ) dt ³0 f c(t ) dt [ f (t )]0f Lt f (t )  f (0)
t of
328 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Using this result in the above equation, we get

Lt sF ( s ) Lt f (t )
so 0 t of

Theorem 6.8 (Division by t). If

L[ f (t ); s ] F ( s)
then

ª f (t ) º f
L«
¬ t
; s»
¼ ³s F ( s ) ds (6.10)

Proof From the definition of Laplace transform

f
³0 e
 st
L[ f (t ); s] F (s) f (t ) dt

Integrating the above equation with respect to s between the limits s to f, we get

f fª f º f § f ·
³s ³s ³0 est f (t) dt » ds ³0 ³s e
st
F(s) ds «¬ f (t) ¨ ds ¸ dt (by changing the order of integration)
¼ © ¹

f
f § est · f f (t) st ª f (t ) º
³0 f (t ) ¨
© t ¸¹s
dt ³0 t
e dt L «
¬ t
; s»
¼

Hence the result.

Note: In applying this rule, one should be careful. Since f (t)/t may have an infinite discontinuity
at t 0, it may not be integrable. If f (t)/t is not integrable, then its Laplace transform does
not exist. For example, at t 0, the function sin t/t does not have an infinite discontinuity,
while the function cos t/t has an infinite discontinuity.

EXAMPLE 6.10 Find the Laplace transform of

1  cos t cos 2t  cos 3t
(i) , (ii) .
t t

Solution Using the result of Theorem 6.8, we have

ª1  cos t º f
(i) L«
¬ t
; s»
¼ ³s F ( s ) ds

where,
1 s
F (s) L[(1  cos t ); s] L[1; s]  L[cos t; s]  2
s s 1
 LAPLACE TRANSFORM METHODS 329

Therefore,

ª1  cos t º f §1 s ·
L«
¬ t
; s»
¼ ³s ¨© s  2 ¸ ds
s  1¹
f f
ª 1 2 º ª s º
«¬ln s  2 ln ( s  1) »¼ «ln 2 1/2 »
s ¬ ( s  1) ¼ s
f
ª 1 º s
«ln 2 1/2 »
0  ln
¬ (1  1/s ) ¼ s ( s  1)1/2
2

( s 2  1)1/2
ln
s

ª § cos 2t  cos 3t · º f
(ii) L Ǭ
© t
¸¹ ; s »
¼ ³s F ( s ) ds
where
s s
F (s) L[(cos 2t  cos 3t ); s ] 2
 2
s 4 s 9
Therefore,

ª cos 2t  cos 3t º f§ s s ·
L«
¬ t
; s»
¼ ³s ¨© 2  2 ¸ ds
s  4 s  9¹
f f
1 ª § s2  4 ·º 1 § 1  4/ s 2 ·
«ln » ln
2 ¬ ¨© s 2  9 ¸¹ ¼ s 2 ¨© 1  9/ s 2 ¸¹
s

1 § s2  9 ·
ln
2 ¨© s 2  4 ¸¹

6.4 TRANSFORM OF A PERIODIC FUNCTION

A function f (t) is called periodic with period T, if f (t  T ) f (t ) for all values of t and T > 0.
For example, the trigonometric functions sin t and cos t are periodic functions of period
2 Q . Periodic functions occur very often in a variety of engineering problems.

Theorem 6.9 If f (t ) is a periodic function with period T, then

T
³0 e
 st
L [ f (t ); s ] f (t ) dt/(1  e st ) (6.11)
330 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Proof From the definition of Laplace transform, we have

f T f
³0 e ³0 e ³T e
 st  st  st
L [ f (t ); s ] f (t ) dt f (t ) dt  f (t ) dt

If we substitute t u T in the second integral on the right-hand side and

write dt du, we obtain
T f
³0 e ³0 e
 st  s (u T )
L [ f (t ); s ] f (t ) dt  f (u  T ) du

T f
³0 e ³0 e
 st
f (t ) dt  e sT  su
f (u ) du

T
³0 e
 st
f (t ) dt  e sT L[ f (t ); s ]

Rearranging, we get
T
(1  e  sT ) L[ f (t ); s ] ³0 e
 st
f (t ) dt

Thus,
T
³0 e
 st
L[ f (t ); s ] f (t ) dt/(1  e sT )

Hence the result.

EXAMPLE 6.11 Obtain the Laplace transform of the periodic saw-tooth wave function
given by
t
of period T , 0  t  T
f (t )
T
Solution The graph of the periodic saw-tooth function is described in Fig. 6.1. Since
f (t )
is periodic with period T, we have

1 T t 1 T 1 T § e st ·
L [ f (t ); s ]
1  e sT ³0 e st
T
dt
T (1  e sT ) ³0 e st tdt
T (1  e sT ) ³0 td ¨
©  s ¸¹

ª §  st ·T T º
1 « te ¸  1
³0 e »
 st
 sT « ¨
dt
T (1  e ) ¬ ©  s ¹ 0 s ¼»

1 ª Te sT 1 º
 sT «  s
 2 (e sT  1)»
T (1  e ) ¬ s ¼
 LAPLACE TRANSFORM METHODS 331

f (t)

t
O T 2T 3T
Fig. 6.1 Saw-tooth function with period T.

Therefore,

1 e sT
L[ f (t ); s ] 
s 2T s (1  e  sT )

EXAMPLE 6.12 Find the Laplace transform of the following full wave rectifier function:

­ E sin ω t , 0  t  λ /ω
f (t ) ®
¯0, λ /ω  t  2λ /ω
Given that

§ 2M ·
f ¨t  ¸ f (t )
© X ¹

Solution Since the given function f (t) is periodic with period 2λ /ω , we have

1 2λ /ω
L [ f (t ); s ]
1 e 2λ s /ω ³0 e st f (t ) dt

1 ª λ /ω  st 2λ /ω º
1 e 2λ s /ω «¬ ³0 e E sin ω t dt  ³λ /ω e st (0) dt »
¼
λ /ω
E ª e st º
« 2 ( s sin ω t  ω cos ω t ) »
1  e2λ s /ω ¬s ω 2
¼0

Therefore,

E ª e  sλ /ω ω º»
L[ f (t ); s ] « ( s sin λ  ω cos λ ) 
1  e2λ s/ω ¬« s 2  ω 2 s 2  ω 2 »¼
332 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

EXAMPLE 6.13 Find the Laplace transform of

⎧ 1, 0≤t<2
f (t ) = ⎨
⎩−1, 2≤t<4

f (t + 4) = f (t )

Solution In this problem, f (t ) is a periodic function of period 4; we, therefore, have

1 4

1 − e−4 s ∫0
− st
L [ f (t ); s ] = e f (t ) dt

1 ⎛ 2 4 ⎞
∫0 e ∫2 e
− st − st
= ⎜⎝ (1) dt + (−1) dt ⎟
1− e −4 s ⎠

1 ⎛ −2e−2 s e−4 s 1 ⎞
= ⎜⎝ + + ⎟
1 − e −4 s s s s⎠

6.5 TRANSFORM OF ERROR FUNCTION

The error function denoted by erf (t) is defined as
2 t 2
erf (t ) =
π ∫0 e−u du (6.12)

This function occurs in many branches of science and engineering; for example, in probability
theory, the theory of heat conduction, and so on. In terms of the power series, we have
∞
2 t (−1)n 2 n
erf (t ) =
π ∫0 n∑=0 n!
u du (6.13)

Alternatively, it can be written as

∞
2 ( −1)n t 2 n +1
erf (t ) =
π ∑ n ! (2n + 1)
(6.14)
n=0

We can easily verify that these series converge everywhere and, therefore, erf (t) is an entire
function. From the definition (6.12), it can be verified at once that
erf (0) = 0 (6.15)

2 ∞ 1/2
−u 2
π ∫0
erf (∞ ) = e du = =1 (6.16)
π
 LAPLACE TRANSFORM METHODS 333

The graph of error function is shown in Fig. 6.2.

erf(t)

t
O

Fig. 6.2 Error function.

In solving heat conduction equation, it has been found useful to introduce the complementary
error function defined as
2 f u 2 2 § f u 2 t
u 2 ·
erfc (t )
Q t ³
e du ¨© e du  e du ¹¸
Q 0 0 ³ (6.17) ³
Therefore,
erfc (t ) 1  erf (t ) (6.18)
Now we shall find the Laplace transform of erf (t): From the definition of Laplace
transform,
f 2 t
u 2
³0 e Q ³0
 st
L[erf (t ); s ] e du dt

Changing the order of integration, we obtain

2 f f
u 2
³0 e ³0 e
 st
L[erf (t ); s ] dt du
π
2 f
 (u 2  su )
s π ³0 e du

2 s 2/4 f
 (u  s /2)2
s π
e ³0 e du

Setting x u  s /2, we get

2 s 2/4 f
 x2
L[erf (t ); s ]
s π
e ³s /2 e dx

1 s 2/4 (6.19)
e erfc ( s /2)
s
334 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

EXAMPLE 6.14 Find the Laplace transform of erf (t1/2).

Solution From the definition of Laplace transform, we have
‡  st 2 t1/2  u 2
L [erf (t1/2 ); s ] ³0 e
π ³0 e du dt

Changing the order of integration, we get

2 f f
u 2
L [erf (t1/2 ); s ]
π ³0 e du ³u 2
e st dt

2 f
 u 2  su 2
s π ³0 e du

2 f
 (1 s )u 2
s π ³0 e du

Setting (1  s )u 2 x 2 or 1  su x, we have

du dx / 1  s
Then
2 f 2 1 § 2 f 2 ·
L [erf (t1/2 ); s]
s π 1 s ³0 e x dx ¨
s 1 s © π ³0 e x dx ¸
¹
or
1
L [erf (t1/2 ); s] (6.20)
s 1 s

EXAMPLE 6.15 Find the Laplace transform of

cos t
t

Solution Let f (t ) sin t ; then

cos t
f c (t )
2 t

Using the property of the Laplace transform of the derivative of a function, we have
L [ f c (t ); s ] sL [ f (t )]  f (0)
 LAPLACE TRANSFORM METHODS 335

Therefore,
ª cos t º
L« ; s» sL [sin t ; s ] (6.21)
¬ 2 t ¼
But,

ª§ ( t )3 ( t )5 · º
L [sin t ; s ] L «¨ t    ¸ ; s »
¬ © 3! 5! ¹ ¼

ª§ t 3/2 t 5/2 · º
L « ¨ t1/2    ¸ ; s »
¬© 3! 5! ¹ ¼

ª 3/2 1 5/2 1 7/2 º

« 3/2  5/2
 7/2
 »
¬s 3! s 5! s ¼

ª1/2 π 1 3/2 ˜ 1/2 π 1 5/2 ˜ 3/2 ˜ 1/2 π º

« 3/2  5/2
 7/2
 »
¬ s 6 s 120 s ¼

π ª §1· 1§1· º
2 3
1§1·
«1  ¨ ¸  ¨ ¸  ¨ ¸  »
2 s3/2 ¬ © 4 s ¹ 2! © 4 s ¹ 3! © 4s ¹ ¼

π
3/2
e1/4 s (6.22)
2s

Now substituting Eq. (6.22) into Eq. (6.21), we get

ª cos t º π 1/4 s
L« ; s» e
¬ 2 t ¼ 2s 1/2

Therefore,

ª cos t º π 1/4 s
L« ; s» e
¬ t ¼ s

Hence the result.

6.6 TRANSFORM OF BESSEL’S FUNCTION

Bessel functions arise in several problems involving circular or cylindrical geometry. It is
therefore useful to find the Laplace transform of Bessel functions of the first kind.
336 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

EXAMPLE 6.16 Find the Laplace transform of

(i) J 0 (t ), (ii) tJ 0 (t ), (iii) e  at J 0 (t ).

Solution (i) From the definition of the Bessel function, we have

f n  2r
(1)r § t ·
J n (t ) ¦ ¨ ¸
r! n  r 1 © 2 ¹
r 0

For n 0 , we have

f 2r f 2r
(1)r § t · (1)r § t ·
J 0 (t ) ¦ ¨ ¸
r! r 1 © 2 ¹ ¦ ¨ ¸
(r !)2 © 2 ¹
r 0 r 0

t2 t4 t6
1   
22 2 2 u 42 2 2 u 42 u 6 2
Thus,

1 1
L [ J 0 (t ); s ] L [1; s ]  2
L[t 2 ; s ]  2 2
L[t 4 ; s ]  
2 2 u4

1 1 2! 1 4! 1 6!
 2 3 2 2 5
 2 
s 2 s 2 u4 s 2 u 4 u 6 s7
2 2

1 ª 1 § 1 · 1u 3 § 1 · º
2 3
1u 3 u 5 § 1 ·
«1  ¨ 2 ¸  ¨ ¸  ¨ ¸  »
s ¬ 2 © s ¹ 2 u 4 © s2 ¹ 2 u 4 u 6 © s2 ¹ ¼
1/2
1§ 1· 1
1
s ¨© s 2 ¸¹ 1  s2

Hence,
1
L [ J 0 (t ); s ]
1  s2
(ii) From the properties of Laplace transform, we have

dn
L [t n f (t ); s] (1)n F (s)
ds n
 LAPLACE TRANSFORM METHODS 337

Therefore,

d d § 1 ·
L[tJ 0 (t ); s ] (1) [ L[ J 0 (t ); s ]]  ¨
ds ds © 1  s 2 ¸¹

Thus,
s
L [tJ 0 (t ); s ]
(1  s 2 )3/2
(iii) From the shifting property of the Laplace transform, we have

L [e at f (t ); s ] F (s  a)
Therefore,

1 1
L[e  at J 0 (t ); s ]
1  s2 so( s  a) 1  ( s  a)2

6.7 TRANSFORM OF DIRAC DELTA FUNCTION

The concept of impulse function or Dirac delta function has been introduced in Chapter 3
itself. In certain applications involving a sudden excitation of a system or a large voltage over
a short interval of time, the Laplace transform of Dirac Delta function is useful. From the
property of Dirac Delta function, we have
f
³0 E (t  a) f (t ) dt f (a)

In particular, if f (t ) e st , then
f
³0 e
 st
L[E (t  a); s ] E (t  a ) dt e as , a ! 0 (6.23)

6.8 INVERSE TRANSFORM

So far we have discussed various properties of the Laplace transform and studied the Laplace
transform of some simple functions. However, if the Laplace transform technique is to be
useful in applications, we have to consider the reverse problem too, i.e., we have to find the
original function f (t ) when we know its Laplace transform F ( s ). Thus, if

L [ f (t ); s ] F (s)
then

f (t ) L1[ F ( s ); t ] (6.24)
338 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

In other words, the inverse Laplace transform of a given function F (s) is that function f (t)
whose Laplace transform is F ( s ). It can be established that f (t ) is unique. Here, L–1 is known
as inverse Laplace transform operator. From the elementary definition (6.24) and from the
results obtained thus far in finding the Laplace transform of some elementary functions, we
can immediately generate the following table of transforms:

Table 6.2 Table of Laplace Transform

f (t ) L[ f (t ); s]  F ( s) F (s) L1[ F ( s ); t ]  f (t )

0 0 0 0
1 1/s 1/s 1

eat 1/( s − a ) 1/( s − a ) eat

eat 1/( s + a ) 1/( s + a ) eat

t 1/s 2 1/s 2 t

tn n !/s n +1 1/s n +1 t n/ n !

a a
sin at sin at
s2  a2 s2  a2

s s
cos at cos at
s2  a2 s2  a2

a a
sinh at sinh at
s2  a2 s2  a2

s s
cosh at cosh at
s2  a2 s2  a2

2as 2as
t sin at t sin at
( s 2  a 2 )2 ( s 2  a 2 )2

s2  a2 s2  a2
t cos at t cos at
(s  a )
2 2 2
( s 2  a 2 )2

In most of the problems we have considered earlier, L[ f (t ); s ] is a simple rational function.

The linearity property holds true even in the case of inverse transform. That is, if F1 ( s ) and F2 ( s )
are the Laplace transform of f1 (t ) and t2 (t ), and if c1 and c2 are any two constants, then

L1[{c1F1 ( s ) q c2 F2 ( s )}; t ]  c1L1[ F1 ( s ); t ] q c2 L1[ F2 ( s ); t ]

By expressing L[ f (t ); s ] as partial fractions, we should be able to recognise them as the

transform of some known functions, with the help of which we can write down the inverse
transform. Similarly, shifting property is also useful in constructing the inverse transform of
some functions, which is stated in the following theorem:
 LAPLACE TRANSFORM METHODS 339

Theorem 6.10 If L[ f (t ); s ]  F ( s ), then

L1[ F ( s  a ); t ]  e  at L1[ F ( s ); t ] (6.25)

Proof Since L[ f (t ); s ]  F ( s ), we have

L1[ F ( s ); t ]  f (t )

Recalling the shifting property of Laplace transform, we find that

L[e at f (t ); s]  F ( s  a)

L1[ F ( s  a); t ]  e at f (t )

Thus,

L−1[ F (s + a); t ] = e−at L−1[ F (s); t ]

EXAMPLE 6.17 Obtain the inverse Laplace transform of

4 s 2  3s  5
( s  1)( s 2  2 s  2)

Solution Using partial fraction expansion, we can write

4 s 2  3s  5 A Bs  C
 
( s  1)( s  2 s  2)
2 s 1 s  2s  2
2

Therefore,

4s 2  3s  5  A ( s 2  2 s  2)  ( Bs  C ) ( s  1)

Let s = –1; then A = 12/5. Equating the coefficient of s on both sides, we have
B + C = 9/5

Equating the coefficient of constant on both sides, we get 2 A  C  5 which gives C = 1/5,
and hence B = 8/5. The given expression can now be written as

4 s 2  3s  5 12 1 8 s 1 1
  
( s  1)( s  2 s  2)
2 5 s  1 5 ( s  1) 2  12 5 ( s  1)2  12
340 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Thus, we find that

⎡ 4 s 2 − 3s + 5 ⎤ 12 ⎡ 1 ⎤ 8 ⎡ ( s − 1) + 1 ⎤ 1 −1 ⎡ 1 ⎤
L−1 ⎢ ; t ⎥ = L−1 ⎢ ; t ⎥ + L−1 ⎢ ; t⎥ + L ⎢ ; t⎥
2
⎢⎣ ( s + 1) ( s − 2 s + 2) ⎥⎦ 5 ⎣s +1 ⎦ 5 2 2
⎣ ( s − 1) + 1 ⎦ 5
2 2
⎣ ( s − 1) + 1 ⎦

12 −t −1 ⎡ 1 ⎤ 8 t −1 ⎡ s + 1 ⎤
= e L ⎢ ; t⎥ + e L ⎢ 2 ; t⎥
5 ⎣s ⎦ 5 ⎣s +1 ⎦

1 ⎡ 1 ⎤
+ et L−1 ⎢ 2 ; t ⎥ (by using the shifting property)
5 ⎣ s +1 ⎦

12 t 8 t 1
 e  e (cos t  sin t )  et sin t
5 5 5
12 t 8 t 9
 e  e cos t  et sin t
5 5 5

Theorem 6.11 If f (t ) is a piecewise continuous function and satisfies the condition of

exponential order c0 such that Lt f (t )/t exists, then for s  c0 ,
t→ 0

⎡ F (s) ⎤ t
L−1 ⎢
⎣ s
; t⎥ =
⎦ ∫ 0 f ( x) dx
t
Proof Let G (t ) = ∫ 0 f ( x) dx. Then G (0)  0 and G b(t )  f (t ); Also,

L[G b(t ); s]  sL[G (t ); s ]  G (0)  sL[G (t ); s ]

i.e.,
F ( s )  sL[G (t ); s ]
Therefore,
F ( s)
L[G (t ); s ] 
s
Hence,

⎡ F (s) ⎤ t
L−1 ⎢
⎣ s
; t ⎥ = G (t ) =
⎦ ∫ 0 f ( x) dx (6.26)

This result can be generalized to show that

⎡ F (s) ⎤ t t t t
L−1 ⎢ n ; t ⎥ = ∫ 0 ∫ 0 ∫ 0 ∫ 0 f (t ) dt
n
 (6.27)
⎣ s ⎦
 LAPLACE TRANSFORM METHODS 341

Theorem 6.12 (Change of scale property). If

L−1[ F ( s ), t ] = f (t )
then
1 ¦ t µ
L1[ F (B s ); t ]  f §
B ¨ B ¶·
Proof From the definition of Laplace transform, we have
f
F ( s)  ± e st f (t ) dt
0

Therefore,
f
F (B s )  ± eB st f (t ) dt
0

Let B t  x, so that dt = dx/α . Then we get

1 f  sx ¦ xµ 1 © ¦ xµ ¸ 1 © ¦ t µ ¸
F (B s)  ± 0 e f §¨ B ¶· dx  B L ª« f §¨ B ¶· ; s ¹º  B L ª« f §¨ B ¶· ; s ¹º
B
Thus,
1 ¦ t µ
L1[ F (B s ); t ]  f §
B ¨ B ¶· (6.28)

EXAMPLE 6.18 Find the inverse Laplace transform of

1 s2 2s − 3 s3
(i) , (ii) , (iii) , (iv) .
(s  a) n
s 2  4s  13 s 2 − s − 3/4 ( s 2  a 2 )2

Solution (i) Using the shifting property, we at once have

© 1 ¸ © 1 ¸ e at t n 1
L1 ª ; t  e at L1 ª
n ¹
;t 
n ¹
« (s  a) º «s º (n  1)!

s2 ( s  2)  4
(ii) 
s  4 s  13
2
( s  2)2  32
Therefore,
© s2 ¸ © ( s  2) ¸ © 1 ¸
L1 ª ; t ¹  L1 ª ; t  4 L1 ª
2 ¹
;t
2 ¹
« s  4 s  13 º
2
« ( s  2)  3 º
2
« ( s  2)  3 º
2

© s ¸ © 1 ¸ (by using the

 e2t L1 ª 2 2 ; t ¹  4e2t L1 ª 2 2 ; t ¹
«s  3 º « s  3 º shifting property)
342 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Thus,

⎡ s+2 ⎤
L−1 ⎢ 2 ; s ⎥ = e 2t cos 3t + (4/3)e 2t sin 3t
⎣ s − 4 s + 13 ⎦

2s − 3 2s − 3 2( s − 1/2) − 2
(iii)
2
= 2
=
s − s − 3/4 ( s − 1/2) − 1 ( s − 1/2)2 − 1
Thus,

⎡ 2s − 3 ⎤ ⎡ ( s − 1/2) ⎤ ⎡ 1 ⎤
L−1 ⎢ 2 ; t ⎥ = 2 L−1 ⎢ 2
; t ⎥ − 2 L−1 ⎢ 2
; t⎥
⎣ s − s − 3/4 ⎦ ⎣ ( s − 1/2) − 1 ⎦ ⎣ ( s − 1/2) − 1 ⎦

⎡ s ⎤ ⎡ 1 ⎤ (by using the

= 2et /2 L−1 ⎢ 2 2 ; t ⎥ − 2et /2 L−1 ⎢ 2 2 ; t ⎥
⎣ s −1 ⎦ ⎣ s − 1 ⎦ shifting property)

Therefore,

⎡ 2s − 3 ⎤
L−1 ⎢ 2 ; t ⎥ = 2et /2 cosh t − 2et /2 sinh t
⎣ s − s − 3/4 ⎦

s3 s(s 2  a 2  a 2 ) s a2 s
(iv)   
( s 2  a 2 )2 ( s 2  a 2 )2 s2  a2 ( s 2  a 2 )2
Hence,

⎡ s3 ⎤ ⎡ s ⎤ ⎡ s ⎤
L−1 ⎢ 2 ; t = L−1 ⎢ 2
2 2 ⎥
; t − a 2 L−1 ⎢ 2
2 ⎥
;t
2 2 ⎥
⎣⎢ ( s + a ) ⎦⎥ ⎣s + a ⎦ ⎣ (s + a ) ⎦
Thereore,

© s3 ¸ a
L1 ª 2
;
2 2 ¹
t  cos at  t sin at
ª (s  a )
« º¹ 2

EXAMPLE 6.19 Find the inverse Laplace transform of

s2 + 1 , ⎛s⎞
(i) ln (ii) cot −1 ⎜ ⎟ .
s ( s + 1) ⎝k⎠

Solution (i) From Theorem 6.4, we have

L[t n f (t ); s ]  (1)n F ( n ) ( s )
 LAPLACE TRANSFORM METHODS 343

In particular, n  1 gives
L[tf (t ); s ]   F b( s )
i.e.
d
F ( s )   L[tf (t ); s ] (6.29)
ds
Let

s2 + 1
L [ f (t ); s ] = ln = F (s)
s ( s + 1)
Then,

d d
F ( s )  [ln ( s 2  1)  ln s  ln ( s  1)]
ds ds
2s 1 1
     L [tf (t ); s]
s 1 s
2 s 1

Now, using Eq. (6.29), we get

1 1 2s
L [ tf (t ); s]    2
s s 1 s 1
Hence,

⎛1 ⎞ ⎛ 1 ⎞ ⎛ s ⎞
tf (t ) = L−1 ⎜ ; t ⎟ + L−1 ⎜ ; t ⎟ − 2 L−1 ⎜ 2 ;t⎟
⎝s ⎠ ⎝ s +1 ⎠ ⎝ s +1 ⎠

= 1 + e−t − 2 cos t

Therefore,

© s 2  1 ¸ 1  et  2 cos t
f (t )  L1 ªln ; t¹ 
« s ( s  1) º t

¦sµ
(ii) Let L[ f (t ); s ]  cot 1 §¶  F (s)
¨k·

Then,
d d ¦ sµ k
F ( s )  § cot 1 ¶   2
ds ds ¨ k· k  s2
344 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Now using Eq. (6.29), we obtain

d k
F (s) = 2 = L [tf (t ); s]
ds k + s2
Therefore,

© k ¸
tf (t )  L1 ª 2 2
; t ¹  sin kt
«k s º

Hence,

⎡ ⎛ s ⎞ ⎤ sin kt
f (t ) = L−1 ⎢cot −1 ⎜ ⎟ ; t ⎥ =
⎣ ⎝k⎠ ⎦ t

6.9 CONVOLUTION THEOREM (FALTUNG THEOREM)

We often come across functions which are not the transform of some known function, but
then, they can possibly be expressed as a product of two functions, each of which is the
transform of a known function. Thus we may be able to write the given function as F ( s ) G ( s ),
where F ( s ) and G ( s ) are known to be transforms of the functions f (t ) and g (t ), respectively.

Theorem 6.13 If F ( s ) and G ( s ) are the Laplace transforms of f (t ) and g (t ) respectively,

then F ( s ) G ( s ) is the Laplace transform of
t
∫ 0 f (t − u) g (u) du
i.e.
t
L−1[ F ( s) G ( s)] = ∫ 0 f (t − u) g (u) du (6.30)

This integral is called the convolution of f and g and is denoted by the symbol f * g .

Proof From the definition of Laplace transform, we have

© f  sv ¸© f  su ¸
F ( s )G ( s )  ª ± e f (v) dv ¹ ª ± e g (u ) du ¹
« 0 º« 0 º

f f
 s (v u )
 f (v) g (u ) dv du
±0 ±0 e

f ¬ f
(v u ) »

± 0 g (u ) ­ ± e
® 0
f (v) dv ¼ du
½
 LAPLACE TRANSFORM METHODS 345

Let u  v  t in the inner integral. Then,

¬ ff »
g (u ) ­ ± e st f (t  u ) dt ¼ du
F ( s) G ( s)  ±
0 ® u ½
Change the order of integration as shown in Fig. 6.3.
u
t
=
u

t=u t=‡

t
O
Fig. 6.3 Convolution integral.

Then, we get

f ¬ t  st »
F (s) G(s)  ± ­ ± e f (t  u ) g (u ) du ¼ dt
0 ® 0 ½
f  st ¬ t »
± e ­ ± f (t  u ) g (u ) du ¼ dt
0 ® 0 ½

© t ¸
 L ª ± f (t  u ) g (u ) du; t ¹
« 0 º

Hence the result.

Definition 6.3 We define the Laplace convolution of f (t ) and g (t ) by the integral

t
f *g = ∫ 0 f (t − u) g (u) du (6.31)

It can be verified that f and g can be interchanged in the convolution, i.e., f and g are
commutative. Let t  u  v in Eq. (6.31) so that  du  dv. Then,
0 t
f *g = − ∫t f (v) g (t − v) dv = ∫ 0 g (t − v) f (v) dv
Therefore,
f *g = g * f (6.32)
346 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

EXAMPLE 6.20 Apply the convolution theorem to evaluate

© s ¸ ⎡ s ⎤ © 1 ¸
(i) L1 ª 2 2 2
; t¹ (ii) L−1 ⎢ 2
; t⎥ (iii) L1 ª 2
; t¹.
« (s a ) º ⎣ ( s + a ) ( s + 1) ⎦ « s ( s  1) º

Solution Consider
s 1© s a ¸
(i) 
(s 2
a
2 2
) a « s  a s  a 2 ¹º
ª 2 2 2

Now choosing
s a
F (s) = 2 2
, G ( s) = ,
s +a s + a2
2

we can write the inverse transform for F ( s ) and G ( s ) as

¦ s µ
L1 § 2 2
; t ¶  cos at  f (t )
¨ s a ·

¦ a µ
L1 § 2 2
; t ¶  sin at  g (t )
¨s a ·

Hence, using the convolution theorem, we obtain

⎡ s ⎤ 1 t
L−1 ⎢ 2 ;t =
2 2 ⎥ a
⎣ (s + a ) ⎦
∫ 0 cos a(t − u) sin au du
1 t
=
a ∫ 0 (cos at cos au + sin at sin au) sin au du
1 t sin at t
=
2a
cos at ∫ 0 sin 2au du + 2a ∫ 0 (1 − cos 2au) du
t t
cos at ⎛ cos 2au ⎞ sin at ⎛ sin 2au ⎞
= ⎜− ⎟ + ⎜u − ⎟
2a ⎝ 2 a ⎠0 2a ⎝ 2a ⎠0
1 cos at t sin at
=− 2
(cos at cos 2at + sin at sin 2at ) + 2
+
4a 4a 2a
t sin at
=
2a
Therefore,
© s ¸ t sin at
L1 ª 2 2 2
; t¹ 
« (s  a ) º 2a
 LAPLACE TRANSFORM METHODS 347

(ii) Consider
s 1 s

( s  a) ( s  1) 2 s  a s2  1
Now choosing
1 , s
F (s) = G( s) = 2
s+a s +1
the inverse transforms for this pair are obtained as

⎡ 1 ⎤
L−1 ⎢ ; t = e− at = f (t )
⎣ s + a ⎥⎦

⎡ s ⎤
L−1 ⎢ 2 ; t ⎥ = cos at = g (t )
⎣s +1 ⎦
Hence, using the convolution theorem, we obtain

⎡ 1 ⎤ t − a ( t −u )
L−1 ⎢
⎣ ( s + a ) ( s 2
+ 1)
; t⎥ =
⎦
∫0e cos u du

t au
= e− at ∫0e cos u du (a standard integral)

t
− at ⎡ eau ⎤
=e ⎢ 2 (sin u + a cos u ) ⎥
⎣a +1 ⎦0

Therefore,

© s ¸ 1
L1 ª 2
; t¹  2
(a cos t  sin t  ae at )
« ( s  a ) ( s  1) º a 1

(iii) We shall write

1 1 1

s ( s  1)
2 2
s ( s  1)2
2

For this pair, we can write inverse transforms as

⎡1 ⎤
L−1 ⎢ 2 ; t ⎥ = t = f (t )
⎣s ⎦

⎡ 1 ⎤ ⎡1 ⎤
L−1 ⎢ ; t = e−t L−1 ⎢ 2 ; t ⎥ = e−t t = g (t )
2 ⎥
⎣ ( s + 1) ⎦ ⎣s ⎦
348 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Hence, using the convolution theorem, we obtain

⎡ 1 ⎤ t
L−1 ⎢ 2 ∫ 0 (t − u) ue
−u
;t =
2 ⎥
du
⎣ s ( s + 1) ⎦
t −u t 2 −u
=t ∫0e u du − ∫0u e du

⎡ t −u ⎤ t
= −t ⎢(ue−u ) t0 − ∫0e du ⎥ + [u 2 e−u ] t0 − 2 ∫ 0 ue
−u
du
⎣ ⎦

= (t + 2) e−t + t − 2

Therefore

© 1 ¸
L1 ª 2 2
; t ¹  (t  2) et t 2
«s ( s  1) º

EXAMPLE 6.21 Prove that

t
∫ 0 J 0 (t ) J0 (t − u) du = sin t
Solution From Table 6.2, we note that

© 1 ¸
L1 ª 2
; t ¹  sin t
« s 1 º

We shall also write

1 1 1

s 12
s2  1 s2  1
Now taking
1 1
F (s)  , G (s) 
s 1
2
s 1
2

their inverse transforms give

© 1 ¸
L1 ª ; t ¹  J 0 (t )  f (t )  g (t )
2
s 1 º
«
Hence, using the convolution theorem, we have
⎡ 1 ⎤ t
L−1 ⎢ 2 ; t⎥ =
⎣s +1 ⎦ ∫ 0 J 0 (t ) J 0 (t − u ) du = sin t
 LAPLACE TRANSFORM METHODS 349

6.10 TRANSFORM OF UNIT STEP FUNCTION

Definition 6.4 The unit step function or Heaviside unit function is defined as
¬0 for t  a
H (t  a)  ­
®1 for t s a, a s 0
Graphically, it can be depicted as in Fig. 6.4.
H(t – a)

t
O a
Fig. 6.4 Illustration of Heaviside unit function.

EXAMPLE 6.22 Find the Laplace transform of unit step function.

Solution From the definition of Laplace transform, we have
f − st
L [ H (t − a); s ] = ∫0 e H (t − a ) dt

a − st f − st
= ∫0 e ⋅ 0 ⋅ dt + ∫a e 1 dt

e− as
= , s>0
s

Theorem 6.14 (Second shifting property). If L[ f (t ); s ]  F ( s ), then

L[ f (t  a ) H (t  a ); s ]  e as F ( s)

L1[e as F ( s ); t ]  f (t  a ) H (t  a )

Proof From the definition of Laplace transform, we have

f  st
L [ f (t  a ) H (t  a ); s ]  ± e f (t  a ) H (t  a ) dt
0

a f
 ± 0 — dt  ± e st f (t  a) dt
0 a
350 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Let t – a = v, and hence dt = dv, the right-hand side of the above equation becomes
f f
³0 e s ( a  v ) f (v) dv e  as ³0 e  sv f (v) dv

e  as F ( s )
Therefore,

L[ f (t  a ) H (t  a ); s ]  e as F ( s) (6.33)

L1[e as F ( s ); t ]  f (t  a ) H (t  a ) (6.34)

EXAMPLE 6.23 Find the inverse Laplace transform of

e as
, a0
s2  1

Solution From the second shifting property, we have

L1[e as F ( s ); t ]  f (t  a ) H (t  a )
In the given problem,
1
F (s) 
s 1
2

Hence,

© 1 ¸
f (t )  L1[ F ( s ); t ]  L1 ª 2
; t ¹  sin t
« s 1 º

Therefore,

© 1 ¸
L1 ªe as 2 ¹
 sin (t  a ) H (t  a )
« s  1º

¬0, ta
­
®sin (t  a ), tsa

Theorem 6.15 (Heaviside expansion theorem). Let F ( s ) and G ( s ) be two polynomials in s

where the degree of F ( s ) is lower than that of G ( s ) and if G(s) has n distinct
roots α i (i = 1,…, n), then
 LAPLACE TRANSFORM METHODS 351

n
⎡ F (s) ⎤ F (αi ) αit
L−1 ⎢ ; t⎥ =
⎣ G(s) ⎦
∑ G ′(α i )
e (6.35)
i =1

Proof Since G ( s ) is a polynomial in s having n distinct roots and F(s) is a polynomial

whose degree is less than that of G ( s ) , we have

F (s) F (s) A1 A2 An
= = + ++ (using partial fractions)
G(s) n s − α1 s − α 2 s − αn
Π ( s − αi )
i =1

Multiplying both sides by ( s  B i ) and taking the limit as s n B , we obtain the coefficients
i

F ( s) ( s  Bi ) s  Bi
Ai  Lt  F (B i ) Lt
s nBi G(s) s nB i G(s)

which take indeterminate form and, therefore, using L’Hospital’s rule, we get
1 F (B i )
Ai  F (B i ) Lt 
s nBi G b( s ) G b(B i )

Hence,
F ( s ) F (α1 ) 1 F (α 2 ) 1 F (α n ) 1
= + + +
′ ′
G ( s ) G (α1 ) ( s − α1 ) G (α 2 ) ( s − α 2 ) ′
G (α n ) ( s − α n )
Thus,

⎡ F ( s) ⎤ F (α1 ) −1 ⎡ 1 ⎤ F (α 2 ) −1 ⎡ 1 ⎤
L−1 ⎢ ; t⎥ = L ⎢ ; t⎥ + L ⎢ ; t⎥ +
⎣ G ( s ) ⎦ G ′(α1 ) ⎣ s − α1 ⎦ G ′(α 2 ) ⎣ s − α 2 ⎦

n
F (α n ) −1 ⎡ 1 ⎤ F (α i ) αit
+
′
G (α n )
L ⎢ ; t⎥ =
⎣ s − αn ⎦
∑ G ′(α i )
e (6.36)
i =1
Hence the result.

EXAMPLE 6.24 Using the method of Heaviside expansion theorem, find the Laplace inverse
of
s2 + 1
s3 + 3s 2 + 2s

Solution We shall take F ( s )  s 2  1, and

G ( s )  s3  3s 2  2 s  s ( s  1) ( s  2)
352 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Here, G ( s ) has three distinct roots 0,  1,  2 and the degree of F ( s ) is lower than that of
G ( s). Hence, using the Heaviside expansion theorem, we have

© s2  1 ¸ F (0) 0 r F (1) t F ( 2) 2t

L1 ª 3 2
; t¹  e  e  e
« s  3s  2 s º G b(0) G b(1) G b( 2)

But, G ′( s ) = 3s 2 + 6 s + 2. Therefore,

© s2  1 ¸ 1 5 2 t
L1 ª 3 2
; t¹   2e
t
 e
« s  3s  2 s º 2 2

6.11 COMPLEX INVERSION FORMULA (MELLIN-FOURIER INTEGRAL)

In solving some complicated problems using the Laplace transform method, the direct approach
so far followed may not be helpful in finding the inverse Laplace transform. Methods based
on complex variable theory may come in handy for finding the inverse transform. Also, it can
be noted that the Laplace transform of f (t ) is expressed as an integral. Similarly, the inverse
Laplace transform of F ( s ) can be expressed as in integral which is known as inverse integral.
This integral can be evaluated by using contour integration methods. The complex inversion
formula is stated in the following theorem.

Theorem 6.16 Let f (t ) and f b(t ) be continuous functions on t s 0 and f (t ) = 0 for t < 0. In
addition, if f (t ) is 0(eγ 0t ) and
F ( s ) = L [ f (t ); s ]
Then
1 H if st
L1[ F ( s ); t ]  f (t )  e F ( s ) ds
2Q i ± H if

t > 0 and γ is a positive constant.

f
Proof Let g (t ) and g b(t ) be continuous functions and if ± f g (t ) dt converges absolutely

and uniformly then g (t ) may be represented by the Fourier integral formula

1 f ⎡ f ⎤
g (t ) =
2π ∫ −f g (ν ) ⎢⎣ ∫ −f cos ω (t − ν ) dω ⎥⎦ dν
1 f ⎡ f ⎤
=
2π ∫ −f ⎢⎣ ∫ −f g (ν ) cos ω (t − ν ) dν ⎥⎦ dω (6.37)
 LAPLACE TRANSFORM METHODS 353

Since sin ω (t − ν ) is an odd function of X, we have

1 f © f ¸

± ª ± f
g (O ) sin X (t  O ) dO ¹ dX 0
2Q  f « º

Combining this expression with Eq. (6.37), we can write

1 f © f
g (O ) eiX (t O ) dO ¹ dX
¸
g (t ) 
2Q ±  f ª ± f
« º

1 f iX t © f
e ª ± g (O ) eiXO dO ¹ dX
¸
 (6.38)
2Q ±  f «  f º

In addition, we assume that g (t ) is of exponential order  0(eH 0t ).

Now we consider the function

⎧⎪e−γ t f (t ), t≥0
g (t ) = ⎨
⎪⎩0, t<0

where H is a real number greater than H 0 . Thus, g (t ) satisfies all the conditions required by
the Fourier integral theorem and, therefore, we have from Eq. (6.38), for t s 0 the relation

1 f iX t © f HO ¸
eH t f (t )  e ª ± e f (O ) eiXO dO ¹ dX
2Q ±  f «  f º

1 f iX t © f (H iX )O ¸
 e ª± e f (O ) dO ¹ dX
2Q ±  f «  f º

1 f iX t
 e F (H  iX ) dX [From the definition of Laplace transform]
2Q ±  f

Let H  iX  s, so that dω = ds/i. It follows that

1 H if t ( s H )
eH t f (t )  e F ( s) ds
2Q i ± H if
Therefore,

1 H if st
f (t )  e F ( s ) ds, ts0 (6.39)
2Q i ± H if

Hence the proof.

354 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

EXAMPLE 6.25 Find the inverse Laplace transform of

1
(s + 1) (s − 2)2

Solution From complex inversion formula, we get

⎡ 1 ⎤ 1 γ +i f e st ds
L−1 ⎢
⎣ ( s + 1) ( s − 2)
;t =
2 ⎥ 2π i
⎦
∫ γ −i f (s + 1) (s − 2)2
e st
= sum of the residues of
( s + 1) ( s − 2)2

at the simple pole s  1 and double pole s  2. Therefore,

⎡ 1 ⎤ ( s + 1) e st d ⎛ e st ⎞
L−1 ⎢ ; t =
2 ⎥ s →−1
Lt + Lt ⎜ ⎟
⎣ ( s + 1) ( s − 2) ⎦ ( s + 1) ( s − 2)2 s →2 ds ⎝ s + 1 ⎠

e −t ⎡ e st {( s + 1) t − 1} ⎤
= + Lt ⎢ ⎥
9 s →2 ⎢⎣ ( s + 1)2 ⎥⎦

e−t t 2t 1 2t
= + e − e
9 3 9

EXAMPLE 6.26 Find the inverse Laplace transform of

sinh ( x s )
, 0 xl
sinh (l s )

Solution Let
sinh ( x s )
F (s) 
sinh (l s )
Then,
1 H if st sinh ( x s ) 1 sinh ( x s )
L1[ F ( s ); t ]  ± e ds  ± e st ds
2Q i H if sinh (l s ) 2Q i C sinh (l s )

= sum of the residues of the integrand at infinitely many simple

poles given by the roots of transcendental equation sinh (l s )  0
 LAPLACE TRANSFORM METHODS 355

i.e.,

el s
 e l s
 0, implying 2l s  1  e2 nQ i .

Hence, sl 2  n2Q 2 , i.e.,

n2Q 2
s , n  0, 1, 2, ...
l2
Thus, we have to compute the residues at the poles
 n2π 2 ,
s 0, s sn n 1, 2, }
l2
Now, the residue at s  0 is
sinh ( x s )
Lt se st 0
s n0 sinh (l s )
The residues at s  sn are obtained as

sinh ( x s ) 2 s
Lt ( s − sn ) e st = e st sinh ( x s ) s = s
s → sn sinh (l s ) l cosh (l s ) n

−n 2π 2
2 ⎛ −n 2π 2 ⎞
l2⎛ −n 2π 2t ⎞
= exp ⎜ ⎟ sinh ⎜⎜ x ⎟⎟
⎛ −n 2π 2 ⎞ ⎝ l2 ⎠ ⎝ l2 ⎠
l cosh ⎜⎜ l ⎟
⎟
⎝ l2 ⎠

2inπ ⎛ −n 2π 2 t ⎞ sinh [i (nπ /l ) x]

= exp ⎜ ⎟
l2 ⎝ l 2 ⎠ cosh (inπ )

2nπ (−1) ⎛ −n 2π 2 t ⎞ sin (n π x/l )

= exp ⎜ ⎟
l2 ⎝ l 2 ⎠ cos nπ

(−1)n +1 ⋅ 2nπ ⎛ nπ ⎞ ⎛ −n 2π 2t ⎞
= sin ⎜ x ⎟ exp ⎜ ⎟ ; n = 1, 2, ...
l2 ⎝ l ⎠ ⎝ l2 ⎠

Therefore,

¥
f
© sinh ( x s ) ¸ 2Q ¦ nQ µ ¦  n 2Q 2 t µ
L1 ª ;t¹  2 (1) n 1 n sin § x ¶ exp § ¶
« sinh (l s ) º l n 1 ¨ l · ¨ l2 ·
356 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

6.12 SOLUTION OF ORDINARY DIFFERENTIAL EQUATIONS

The Laplace transform technique is one of the powerful tools for solving physical problems
involving ordinary differential equations (ODE), particularly initial value problems. It reduces
the solution of ODE to the solution of an algebraic equation. This method has a particular
advantage in finding the solution of an initial value problem, without first finding the general
solution and then using the given ICs for evaluating the arbitrary constants. We shall first
apply the Laplace transform technique to find the solution of a typical initial value problem
and demonstrate the steps involved. Consider a second order linear differential equation

d2y dy
2
B  C y  f (t ) (6.40)
dt dt
subject to the ICs
y (0)  A, y b(0)  B (6.41)

where B and C are constants. Taking the Laplace transform of (6.40) on both sides, we obtain
L[{ y bb  B y b  C y}; s ]  L[ f (t ); s ]
Using the property of the Laplace transform of the derivatives, we get

{s 2 L[ y (t ); s ]  sy (0)  y b(0)}  B{sL[ y (t ); s ]  y (0)}  C L[ y (t ); s]  L[ f (t ); s]

Using the ICs, we have, after regrouping the terms,

( s 2  B s  C ) L[ y (t ); s ]  L[ f (t ); s ]  ( s  B ) A  B

Alternatively,

(s 2  B s  C ) Y (s)  F (s)  (s  B ) A  B
i.e.

(s  B ) A  B F (s)
Y (s)   (6.42)
s  Bs  C
2
s  Bs  C
2

Taking the inverse Laplace transform on both sides, we get the solution in the form

© ( s  B ) A  B ¸ 1 © F ( s ) ¸
y (t )  L1 ª ; t¹  L ª 2 ; t¹ (6.43)
« s Bs  C º «s Bs  C º
2

EXAMPLE 6.27 Solve the equation using the Laplace transform method
y bb  4 y b  8 y  cos 2t

Given that y  2 and y b  1 when t  0.

 LAPLACE TRANSFORM METHODS 357

Solution Taking the Laplace transform of the given ODE, we obtain

s
{s 2Y ( s )  sy (0)  y b(0)}  4{sY ( s)  y (0)}  8Y ( s)  2
s  4
Using the initial conditions, we get
s
s 2Y ( s)  2s  1  4sY ( s)  8  8Y ( s) 
s 4
2

s
( s 2  4s  8) Y ( s )   2s  9
s 4
2

Therefore,

( s /20 + 1/5) ( s /20 + 2/5) 2s 9

Y (s) = 2
− 2
+ 2
+ 2
s +4 s + 4s + 8 s + 4s + 8 s + 4s + 8

1 s 1 1 1 ( s + 2) − 2 2
= × 2 + 2 − × 2 2
−
20 s + 4 5 s + 4 20 ( s + 2) + 2 5

1 2( s + 2) − 4 9
× 2 2
+ 2 2
+
( s + 2) + 2 ( s + 2) + 2 ( s + 2)2 + 22

Taking the inverse Laplace transform, we obtain

1 1 1 1 −2 t
y (t ) = cos 2t + sin 2t − e −2t cos 2t + e sin 2t
20 10 20 20
2 −2 t 9
− e sin 2t + 2e −2t cos 2t − 2e−2t sin 2t + e −2t sin 2t
10 2
On simplification, we get
e −2 t 1
y (t ) = (39 cos 2t + 47 sin 2t ) + (cos 2t + 2 sin 2t )
20 20

EXAMPLE 6.28 Using the Laplace transform technique, solve

x  3 x  2 x  te t

Given
x (0)  1 and x (0)  0

Solution Taking the Laplace transform of the given ODE, we get

1
{s 2 X ( s )  sx(0)  xb(0)}  3{sX ( s )  x(0)}  2 X ( s ) 
( s  1)2
358 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Applying the ICs, we have

1
( s 2  3s  2) X ( s )  s3
( s  1)2
Therefore,

1 s+3
X (s) = +
3
( s + 1) ( s + 2) ( s + 1) ( s + 2)

1 1 1 1 1 1
= × − × + ×
2 s + 1 2 ( s + 1) 2 2 ( s + 1)3
1 1 2 1
− × + − (using partial fractions)
2 s + 2 s +1 s + 2
The inverse Laplace transform yields

©¬ 5 1 1 1 1 1 3 1 » ¸
x(t )  L1 ª ­ t  t  t  t ;t
«®
2 s 1 2 ( s  1)2 2 ( s  1)3 2 s  2 ¼½ ¹º

5 t 1 t 1 t t 2 3 2t e t ¦ t 2 µ 3 2t
 e  e t e  e  §5  t  ¶ e
2 2 2 2 2 2 ¨ 2· 2

EXAMPLE 6.29 Using the Laplace transform technique, solve the following initial value
problem:
ty ′′ + y ′ + ty = 0, y (0) = 1, y ′(0) = 0

Solution This is an example of ODE with variable coefficients. Taking the Laplace
transform on both sides of the given ODE and using Eq. (6.7), we get
d d
 L[ y bb; s ]  L[ y b; s ]  L[ y; s ]  0
ds ds

d 2 d
 {s Y ( s )  sy (0)  y b(0)}  {sY ( s )  y (0)}  {Y ( s )}  0
ds ds
Applying the ICs, we obtain
d 2 d
{s Y ( s )  s}  {sY ( s )  1}  Y ( s )  0
ds ds
or
dY
( s 2  1)  sY  0
ds
 LAPLACE TRANSFORM METHODS 359

which is a first order ODE. On rewriting, we get

dY s ds
+ 2 =0
Y s +1
Integrating, we see that

1
ln Y + ln ( s 2 + 1) = ln c
2

c
Y=
2
s +1

Now, taking the inverse Laplace transform, we find y (t ) = cJ 0 (t ), where J 0 (t ) is a Bessel

function of order zero. Since y (0) = 1 = cJ 0 (0) = c, the required solution is

y (t ) = J 0 (t )

EXAMPLE 6.30 Solve the simultaneous equations

dx dy
− y = et , + x = sin t
dt dt
using the Laplace transform method which satisfies the conditions x(0) = 1, y (0) = 0.

Solution Taking the Laplace transform of both the equations and using the notation
Y = L[ y; s ], we have

1
sX − x(0) − Y = L[et ; s ] =
s −1
1
sY − y (0) + X = L [sin t ; s ] = 2
s +1
Using the given ICs, the above equations reduce to
s
sX − Y = (6.44)
s −1

1
X + sY = 2
(6.45)
s +1
Solving Eqs. (6.44) and (6.45), we get

s2 1 (6.46)
X= 2
+
( s − 1) ( s + 1) ( s + 1)2
2
360 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

 s3  s 2  2s
Y (6.47)
( s  1) ( s 2  1) 2

s s
= 2 2
− (6.48)
( s + 1) ( s − 1) ( s 2 + 1)
Using partial fractions, we get
s2 A Bs  C 1 ¦ 1 s 1 µ
    
( s  1) ( s  1)
2 ( s  1) s 2  1 2 ¨§ s  1 s 2  1 s 2  1 ·¶
Hence Eq. (6.46) gives
1¦ 1 s 1 µ 1
X  §   ¶ 2
2 ¨ s  1 s  1 s  1 · ( s  1)2
2 2

Taking inverse Laplace transform, we obtain

1
x  [et  cos t  sin t  (sin t  t cos t )] (6.49)
2
Also, from Eq. (6.48), we have
s 1¦ 1 s 1 µ
Y    
(s 2
 1)
2 2 ¨ s  1 s  1 s  1 ·¶
§ 2 2

Again, taking inverse Laplace transform, we get

1
y (t sin t  et  cos t  sin t ) (6.50)
2
Equations (6.49) and (6.50) constitute the solution of the given system.

6.13 SOLUTION OF PARTIAL DIFFERENTIAL EQUATIONS

A large number of problems in science and engineering involve the solution of linear partial
differential equations. A function of two or more variables may also have a Laplace transform.
Suppose x and t are two independent variables; consider t as the principal variable and x as
the secondary variable. When the Laplace transform is applied with t as a variable, the PDE
is reduced to an ordinary differential equation of the t-transform U ( x, s ), where x is the
independent variable. The general solution U ( x, s) of the ODE is then fitted to the BCs of
the original problem. Finally, the solution u ( x, t ) is obtained by using the complex inversion
formula. Thus, the Laplace transform is specially suited to solving initial boundary value
problems (IBVP), when conditions are prescribed at t  0. We have already noted in Chapters 3
and 4 that such situations naturally arise in the case of heat conduction equation and wave
 LAPLACE TRANSFORM METHODS 361

equation. Before we consider the Laplace transform method of solution of IBVP, we consider
the following example to prove some of the useful elementary results.

EXAMPLE 6.31 If u ( x, t ) is a function of two variables x and t, prove that

wu
(i) L ª« ; s º» sU ( x, s )  u ( x, 0)
¬wt ¼

ª w 2u º
(ii) L « 2 ; s » s 2U ( x, s )  su ( x, 0)  ut ( x, 0)
¬wt ¼
wu
(iii) L ⎡⎢ ; s ⎤⎥ =
dU ( x, s)
⎣w x ⎦ dx

ª w 2u º d2
(iv) L « 2 ; s » U ( x, s )
¬w x ¼ dx 2

ª w 2u º d d
(v) L « ; s» s U ( x, s )  u ( x, 0) .
¬w x w t ¼ dx dx

where U ( x, s )  L[u ( x, t ); s ].

Proof Following the definition of the Laplace transform, we have

ªw u º f wu p wu
(i) L « ; s»
¬wt ¼ ³0 e  st
wt
dt Lt
p of 0 ³ e st
wt
dt

ª p º
Lt {e st u ( x, t )}0p  s ³0 e
 st
u ( x, t ) dt »
p of «
¬ ¼
f
³0 e
 st
u ( x, 0)  s u ( x, t ) dt

Therefore,
ªw u º
L « ; s» sU ( x, s )  u ( x, 0)
¬wt ¼

ª w 2u º ªwV º wu
(ii) L « 2 ; s » L« ;s , V
¬wt ¼ ¬ w t »¼ wt
sL [V ; s ]  V ( x, 0)

s {sU ( x, s )  u ( x, 0)}  ut ( x, 0)
362 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Thus,
⎡ ∂ 2u ⎤
L ⎢ 2 ; s ⎥ = s 2U ( x, s ) − su ( x, 0) − ut ( x, 0)
⎣∂t ⎦

⎡∂ u ⎤ ∞ ∂u d ∞ dU
∫0 e ∫0 e
− st − st
(iii) L ⎢ ; s ⎥ = dt = u ( x, t )dt = ( x, s )
⎣∂ x ⎦ ∂x dx dx

⎡ ∂ 2u ⎤ ⎡ ∂ u ⎤ d ⎛ dU ⎞ d U
2
∂u
(iv) L ⎢ 2 ; s ⎥ = L ⎢ ; s ⎥ = ⎜⎝ ⎟⎠ = ( x, s ), u =
⎣∂ x ⎦ ⎣ ∂ x ⎦ dx dx dx 2 ∂x

⎡ ∂ 2u ⎤ d dU du
(v) L ⎢ ; s ⎥ = [ sU ( x, s ) − u ( x, 0)] = s ( x, s ) − ( x, 0)
⎣ ∂ x ∂ t ⎦ dx dx dx
Thus, we notice from the above results that the partial derivatives are transformed into ordinary
derivatives.

6.13.1 Solution of Diffusion Equation

EXAMPLE 6.32 Solve the following IBVP using the Laplace transform technique:
PDE: ut = u xx , 0 < x < 1, t>0
BCs: u (0, t ) = 1, u (1, t ) = 1, t>0
IC: u ( x, 0) = 1 + sin π x, 0 < x <1

Solution Taking the Laplace transform of both sides of the given PDE, we have

d 2U
sU ( x, s ) − u ( x, 0) =
dx 2
Thus, the solution of the second order PDE reduces to the solution of second order ODE given
by

d 2U
− sU ( x, s ) = −1(1 + sin π x) (after using IC) (6.51)
dx 2
The general solution of Eq. (6.51) is found to be
1 sin π x
U ( x, s ) = Ae sx
+ Be− sx
+ + 2 (6.52)
s π +s

But, u (0, t ) = 1, u (1, t ) = 1, and their Laplace transforms are

1 1
U (0, s ) = , U (1, s ) =
s s
 LAPLACE TRANSFORM METHODS 363

From Eq. (6.52), we have

1 1
A B 
s s
Hence, A  B  0, and

1 1
Ae s
 Be s
 
s s
Therefore,

Ae s
 Be s
0
This is a homogeneous system; the determinant of the coefficient matrix is

1 1
s  s
 e s
e s
y0
e e

Thus, the only possible solution is the trivial solution and, therefore,
AB0
From Eq. (6.52), we now have
1 sin Q x
 U ( x, s )  (6.53)
s Q2  s
Taking the inverse Laplace transform of Eq. (6.53), we get

¦1 µ © sin Q x ¸
u ( x, t )  L1 § ; t ¶  L1 ª 2 ; t¹
¨s · «Q  s º

Thus,
2
u ( x, t ) = 1 + sin π x e−π t (6.54)
is the required solution.

EXAMPLE 6.33 Using Laplace transform, solve the following initial boundary value problem:
PDE: kut = u xx , 0  x  1, 0tf

BCs: u (0, t )  0, u (l , t )  g (t ), 0tf

IC: u ( x, 0) = 0, 0< x<l

Solution Taking the Laplace transform of both sides of the PDE, we get

d 2U
K [ sU ( x, s ) − u ( x, 0)] =
dx 2
364 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Thus the solution of PDE reduces to the solution of ODE which, after using the IC, can be
rewritten as

d 2U
− KsU ( x, s) = 0 (6.55)
dx 2
Its general solution is found to be

U = cosh ( ks x) + B sinh ( ks x) (6.56)

Taking the Laplace transform of the BCs, we have

U (0, s )  0, U (l , s )  G ( s ) (6.57)
Applying Eq. (6.57) in Eq. (6.56), we get 0 = A. Therefore,

G(s) B sinh ( ks l )

which implies

G ( s)
B (6.58)
sinh ( ks l )

Taking the inverse Laplace transform, we obtain

1 H  if sinh ( ks x)
u ( x, t )  ± G ( s ) e st ds (6.59)
2Q i H if sinh( ks l )

To evaluate the integral on the right-hand side of this equation, we use the method of residues
for which we note that the integrand has poles given by

sinh ( ks l ) = 0 = e ks l
− e− ks l

e2 ks l
= 1 = e2nπ i

implying the simple poles at

n2π 2 ,
sn = − n = 0, ± 1, ± 2, …. (6.60)
Kl 2
Now,

1 H f sinh ( ks x)
± G ( s ) e st ds  sum of the residues of the integrand at the poles.
2Q i H  f sinh ( ks l )
 LAPLACE TRANSFORM METHODS 365

But the residue at the pole s  0 is zero. The residues at s  sn are

sinh ( ks x)
Lt G ( s ) e st
s nsn d
[cosh ( ks l )]
ds
2 sG ( s )
 Lt e st sinh ( ks x)
s nsn l k cosh ( ks l )

 n 2Q 2 ¦  n 2Q 2 µ
G§
¨ kl 2 ¶·
2 ¦  n 2Q 2 µ
kl 2 ¦  n 2Q 2 µ
 exp § § x¶ , n  1, 2, 3, ...
¨ kl 2 ¶·
sinh
¦  n 2Q 2 µ ¨ l2 ·
l k cosh § l ¶
¨ l2 ·

¦  n 2Q 2 µ
2inQ ¨ kl 2 ¶·
G§
¦ inQ µ ¦  n 2Q 2 t µ
 2 sinh § x¶ exp § n  1, 2, ...
¨ kl 2 ¶·
,
l k cosh (inQ ) ¨ l ·

Using the fact that cosh (inQ )  cos n Q , and

¦ inQ µ ¦ nQ µ
sinh § x ¶  i sin § x¶
¨ l · ¨ l ·
the above expression becomes

2nQ (1) n ¦ nQ µ ¦  n 2Q 2 µ ¦  n 2Q 2 µ
u ( x, t )  sin § x ¶G § ¶ exp § t ¶, n  1, 2, ...
l 2k ¨ l · ¨ kl
2
· ¨ kl
2
·
Therefore, the required solution is
f ⎛ n 2π 2 ⎞
2π ⎛ nπ ⎞
u ( x, t ) =
l 2k
∑ (−1)n nG ⎜ 2 ⎟ sin ⎜
⎝ kl ⎠ ⎝ l
x ⎟ exp[(−n 2π 2/ kl 2 ) t ]
⎠
(6.61)
n =1

EXAMPLE 6.34 Find the solution of the BVP given by

w 2u 1 wu , 0 c x c a, t  0
PDE:
wx 2 k wt

BCs: u (0, t ) = f (t ), u ( a, t )  0

IC: u ( x, 0) = 0
using the Laplace transform method.
366 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Solution Taking the Laplace transform of the given PDE, we have

d 2U1
= [ sU ( x, s) − u ( x, 0)]
2 (6.62)
dx k

Using the initial condition u ( x, 0)  0, we get

d 2Us
− U ( x, s ) = 0
2
(6.63)
dx k
Its general solution is found to be

U ( x, s) = Ae ( s/k ) x
+ Be− ( s/k ) x (6.64)
Now taking the Laplace transform of the BCs, we have
U (0, s )  F ( s) (6.65)

U ( a, s )  0 (6.66)
Using Eqs. (6.64) and (6.65), we obtain
F ( s)  A  B (6.67)
Substituting Eq. (6.66) into Eq. (6.64), we get
U (a, s ) = A exp [ ( s / k ) a ] + B exp [− ( s / k ) a ] (6.68)
Combining Eqs. (6.67) and (6.68), we get
F ( s) = A (1 − e2 a s/ k
)
Therefore,
A = F ( s)/(1 − e2a s/k
)
and

F (s) F ( s )e 2 a s /k
B = F (s) − −
(1 − e2 a s /k
) 1 − e2a s /k

or

A = F ( s) e− a s/k
/(e− a s/k
− ea s/k
)
and

B = F ( s ) ea s/k
/(ea s/k
− e− a s/k
)
Therefore,

U ( x, s ) =
(e a s/k
F (s)
−e −a s / k
)
{e s / k (a − x)
− e− s / k (a− x)
}
 LAPLACE TRANSFORM METHODS 367

© s ¸
sinh ª (a  x ) ¹
« k º (6.69)
 F (s)
¦ s µ
sinh § a¶
¨ k ·

Its inverse Laplace transform yields

© © s ¸ ¸
ª F ( s ) sinh ª (a  x) ¹ ¹
1 « k º ¹
u ( x, t ) L ª ;t
ª ¦ s µ ¹ (6.70)
ª sinh § a¶ ¹
« ¨ k · º

© s ¸
e st F ( s ) sinh ª (a  x) ¹
1 H  if k
« º
 ds (6.71)
2Q i ± H if ¦ s µ
sinh § a¶
¨ k ·

6.13.2 Solution of Wave Equation

EXAMPLE 6.35 Using the Laplace transform method, solve the IBVP described as
1
PDE: u xx = utt − cos ω t , 0 c x  f, 0ctf
c2
BCs: u (0, t ) = 0, u is bounded as x tends to f
ICs: ut ( x, 0) = u ( x, 0) = 0

Solution Taking the Laplace transform of PDE, we obtain

d 2U 1 s
 [ s 2U ( x, s)  su ( x, 0)  ut ( x, 0)] 
dx 2
c 2
s X2
2

Using the ICs, we get

d 2U s2 s
 U ( x, s )   (6.72)
dx 2
c 2
s X2
2

Its general solution is found to be

c2
U ( x, s ) = Ae( s / c ) x + Be − ( s / c ) x +
s(s 2 + ω 2 )
368 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

As x n f, the transform should also be bounded which is possible if A  0; thus,

c2
U ( x, s ) = Be− ( s / c ) x + (6.73)
s(s 2 + ω 2 )
Taking the Laplace transform of the BC, we get
U (0, s)  0
Using this result in Eq. (6.73), we have
c2
B
s(s 2  X 2 )
Hence,

c2
U ( x, s ) = [1 − e− ( s / c ) x ] (6.74)
s(s 2 + ω 2 )
Now, taking its inverse Laplace transform, we get
−( s / c ) x
⎡ 1 ⎤ 2 −1 ⎡ e ⎤
u ( x, t ) = c 2 L−1 ⎢ 2 ; t ⎥ − c L ⎢ ; t⎥ (6.75)
⎣ s(s + ω ) ⎦ ⎣⎢ s ( s + ω ) ⎦⎥
2 2 2

But,
© 1 ¸ 1 ¬ 1 © 1 ¸ 1 © s ¸» 1
L1 ª 2 ; t¹  2 ­L ª ; t ¹  L ª 2 ; t ¹ ¼  2 (1  cos X t ),
« s( s  X ) º X «s º «s X º½ X
2 2
®

⎡ e − ( x /c ) s ⎤ 1 ⎧ ⎛ x ⎞⎫ ⎛ x⎞
L−1 ⎢ 2 ; t ⎥ = 2 ⎨1 − cos ω ⎜ t − ⎟ ⎬ H ⎜ t − ⎟ ,
⎣⎢ s ( s + ω ) ⎦⎥ ω ⎩
2 ⎝ c ⎠⎭ ⎝ c ⎠

where H is the Heaviside unit function. Substituting these results in Eq. (6.75), we arrive at
c2 c2 ⎡⎧ ⎛ x ⎞⎫ ⎛ x ⎞⎤
u ( x, t ) = (1 − cos ω t ) − 2 ⎢⎨
1 − cos ω ⎜ t − ⎟ ⎬ H ⎜ t − ⎟ ⎥
ω 2
ω ⎣⎩ ⎝ c ⎠⎭ ⎝ c ⎠⎦

EXAMPLE 6.36 Solve the IBVP described by

PDE: utt = u xx , 0 < x < 1, t>0
BCs: u (0, t ) = u (1, t ) = 0, t>0
ICs: u ( x, 0) = sin π x, ut ( x, 0) = − sin π x, 0 < x <1
Solution Taking the Laplace transform of the given PDE, we get
d 2U
2
 s 2U ( x, s)  su ( x, 0)  ut ( x, 0)
dx
 LAPLACE TRANSFORM METHODS 369

Using the initial conditions, this equation becomes

d 2U
− s 2U = (1 − s) sin π x (6.76)
dx 2
Its general solution is found to be
( s  1) sin Q x
U ( x, s)  Ae sx  Be sx  (6.77)
Q 2  s2
The Laplace transform of the BCs gives
U (0, s )  0, U (1, s )  0 (6.78)
Using Eq. (6.78) into Eq. (6.77), we find A  B  0. Hence, we obtain
( s  1) sin Q x
U ( x, s )  (6.79)
Q 2  s2
Taking the inverse Laplace transform, we get

¬ © s 1 ¸» ¬ © s ¸ © 1 ¸»
u ( x, t )  sin Q x ­ L1 ª 2 ; t  sin Q x ­ L1 ª 2
2 ¹¼
; t  L1 ª 2
2 ¹ 2 ¹¼
;t
® «s Q º½ ® «s Q º «s Q º½

¦ sin Q t µ
 sin Q x § cos Q t  ¶
¨ Q ·
Hence the required solution of the given IBVP is
¦ sin Q t µ
u ( x, t )  sin Q x § cos Q t  ¶
¨ Q ·

EXAMPLE 6.37 If the function u ( x, t ) satisfies the following:

1
PDE: xxx = utt + k , 0 ≤ x ≤ l, t >0
c2
BCs: u (0, t ) = u x (l , t ) = 0, t  0,
ICs: u ( x, 0) = ut ( x, 0), 0c xcl
then, show that
⎡ kc 2 ⎛ cosh {s (l − x)/c} ⎞ ⎤
u ( x, t ) = L−1 ⎢ 3 ⎜ − 1⎟ ; t ⎥
⎣⎢ s ⎝ cosh {sl /c} ⎠ ⎦⎥
and find the solution.
370 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Solution Taking the Laplace transform of the given PDE, we get

d 2U 1 k
2
 2
[ s 2U ( x, s )  su ( x, 0)  ut ( x, 0)] 
dx c s
Using the initial conditions, we obtain

d 2U s2 k
2
− 2
U ( x, s ) = (6.80)
dx c s
Its general solution is found to be

kc 2
U ( x, s ) = Ae( s/c ) x + Be−( s/c ) x − 3
(6.81)
s
Taking the Laplace transform of the BCs, we get

dU
U (0, s ) = (l , s ) = 0 (6.82)
dx
Using these boundary conditions, Eq. (6.81) gives

A + B = kc 2/ s3

As ( s / c )l s
e − B e − ( s / c )l = 0
c c
Eliminating B, we get
s kc
A [e ( s / c ) l + e − ( s / c )l ] = 2 e − ( s / c ) l
c s
which gives
kc 2 ⎛s ⎞
A= e−( s / c)l /cosh ⎜ l ⎟
2s 3 ⎝c ⎠

kc 2 ⎛s ⎞
B= e( s / c )l/cosh ⎜ l ⎟
2s 3 ⎝c ⎠
Hence, from Eq. (6.81), we have
©s ¸
cosh ª (l  x) ¹
kc 2 «c º kc 2
U ( x, s )  
s3 ¦s µ
cosh § l ¶ s3 (6.83)
¨c ·
 LAPLACE TRANSFORM METHODS 371

Applying the complex inversion formula, we obtain

⎡ ⎡s ⎤ ⎤
⎢ cosh ⎢ (l − x) ⎥ ⎥
2 −1
u ( x, t ) = kc L ⎢ ⎣ c ⎦ ; t − kc 2 L−1 ⎡ 1 ; t ⎤ (6.84)
⎥ ⎢ 3 ⎥
⎢ s3 cosh ⎛ l ⎞
s ⎥ ⎣s ⎦
⎜ ⎟
⎣⎢ ⎝c ⎠ ⎦⎥

©s ¸
e st cosh ª (l  x) ¹
kc 2 H if «c º kc 2 t 2
 ds  (6.85)
2Q i ± H if ¦s µ 2
s3 cosh § l ¶
¨c ·

But

⎡s ⎤
γ +if
e st cosh ⎢ (l − x) ⎥
1 ⎣c ⎦ ds =
2π i ∫ γ −if 3 ⎛ s ⎞
sum of the residues at the poles s = 0 of order
s cosh ⎜ l ⎟ three and at the poles of cosh ( s/c)l , i.e., at
⎝c ⎠
e(2 s/c )l = −1 = ei (2nπ +π )

or at
(2n  1)cQ i
s , n  0, q 1, q 2, ...
2l
Now the residue at the pole s = 0 of order 3 is

© st ¬s »¸ © ¬s »
2 ªe cosh ­ (l  x) ¼ ¹ ª cosh ­ (l  x) ¼
1 d
ª
®c ½
¹
1 d
ªte st
®c ½ xt l  x
Lt Lt e
s n0 2 ds 2 ª ¦s µ
cosh § l ¶ ¹ 2 s n 0 ds ª ¦ s
cosh § l ¶
µ c
ª« ¨c · ¹
º ª
« ¨c ·

¬s » ¬s » ¦ s µ¸
sinh ­ (l  x) ¼ cosh ­ (l  x) ¼ sinh § l ¶ ¹
® c ½ l
st ¦ µ ® c ½ ¨c ·
t e § ¶ ¹
¦s µ ¨c· 2¦s µ ¹
cosh § l ¶ cosh § l ¶ ¹
¨c · ¨c · º

Again differentiating and taking the limit, we get

1© 2 (l  x ) 2 l2 ¸ t2 x2 2lx
ªt   ¹  
2
2« c c2 º 2 2c 2
2c 2
372 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Also, the residue at the poles

i (2n + 1) π c
sn = , n = 0, 1, 2, …
2l
1 1 ⎛l − x ⎞
= exp ( sn t ) cosh ⎜ sn ⎟ , n = 0, 1, 2, ...
sn3 l ⎛l ⎞ ⎝ c ⎠
sinh ⎜ sn ⎟
c ⎝c ⎠

⎧⎛ 2n + 1 ⎞ ⎫ ⎧⎛ 2n + 1 ⎞ ⎛ l − x ⎞⎫
c exp ⎨⎜ ⎟ π ict ⎬ cosh ⎨⎜ ⎟ π ic ⎜ ⎟⎬
⎩ ⎝ 2l ⎠ ⎭ ⎩⎝ 2l ⎠ ⎝ c ⎠⎭
=
(2n + 1)3π 3i 3c3
3 3
2 l
l sinh
(2n + 1)
2l {
π ic
l
c }
¬¦ 2n  1 µ » ¬¦ 2n  1 µ »
8l 2 c exp ­§ c ¶ Q ict ¼ cosh ­§ ¶ Q i (l  x) ¼
 ®¨ 2l · ½ ®¨ 2l · ½
(2 n  1)
(2n  1)3 Q 3 (i )c3 sinh \ Q i^
2

¬ ¦ 2n  1 µ » ¬¦ 2n  1 µ »
8l 2 exp ­i § ¶ Q ct ¼ cos ­§ ¶ Q (l  x) ¼
 ® ¨ 2l · ½ ®¨ 2l · ½
(since sinh iR  i sin R
Q cosh iR  cos R )
(2n  1)3 Q 3c 2 sin (2n  1)
2

Its real part is

⎛ 2n + 1 ⎞ ⎧⎛ 2n + 1 ⎞ ⎫
8l 2 cos ⎜ ⎟ π ct cos ⎨⎜ ⎟ π (l − x) ⎬
⎝ 2l ⎠ ⎩⎝ 2l ⎠ ⎭, n = 0, 1, 2, ...
(−1) (2n + 1) π c
n 3 3 2

Therefore,
© ¬s »¸
ª cosh ­ (l  x) ¼ ¹ t 2 x2 lx
®c
L1 ª ½
¹ 
2

ª
s 3
cosh
¦ s
l
µ ¹ 2 2c c2
ª« § ¶
¨c · ¹ º

© ¦ 2n  1 µ ¸ ©¦ 2n  1 µ ¸
(1)n 8l 2 ªcos § ¶ Q ct ¹ cos ª§ ¶ Q (l 
¥
f x) ¹
« ¨ 2l · º «¨ 2l · º

n 0 (2n  1)3 Q 3c 2
 LAPLACE TRANSFORM METHODS 373

Thus, the required solution from Eq. (6.84) is

¥
f
kx 8kl 2 (1)n © ¦ 2 n  1 µ ¸ ©¦ 2 n  1 µ ¸
u ( x, t )  ( x  2l )  3 cos ª§ ¶ Q ct ¹ cos ª§ ¶ Q (l  x) ¹
2 Q n  0 (2n  1) 3
«¨ 2l · º «¨ 2l · º

EXAMPLE 6.38 A string is stretched and fixed between two points (0, 0) and (l, 0). Motion
is initiated by displacing the string in the form u  M sin (Q x/l ) and released from rest at time
t  0. Find the displacement of any point on the string at any time t.

Solution The displacement u ( x, t ) of the string is governed by

PDE: utt = c 2u xx , 0 < x < l, t0

BCs: u (0, t ) = u (l , t ) = 0
ICs: u ( x, 0) = λ sin (π x / l ), ut ( x, 0)  0
Taking the Laplace transform of the given PDE, we have

d 2U
s 2U ( x, s)  su ( x, 0)  ut ( x, 0)  c 2
dx 2
Using the ICs, we get

d 2U s 2U Ms Qx
2
 2
 2
sin (6.86)
dx c c l
Its general solution is found to be
λ s sin (π / l ) x
U ( x, s) = Ae( s / c ) x + Be−( s / c ) x + (6.87)
s 2 + π 2 c 2/l 2
The Laplace transform of the BCs is given by
U (0, s )  0, U (l , s )  0
Applying these conditions in Eq. (6.87), we obtain
A+ B = 0
Ae sl/c + Be − sl /c = 0
On solving the above set of equations, we get only the trivial solution, viz.
AB0
Thus,
λ s sin (π / l ) x
U ( x, s ) =
s 2 + π 2 c 2/ l 2
374 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Its inverse Laplace transform yields

⎡ s ⎤ πx ⎛πc ⎞ πx
u ( x, t ) = λ L−1 ⎢ 2 ; t sin
2 2 2 ⎥
= λ cos ⎜ t ⎟ sin
⎣ s + π c /l ⎦ l ⎝ l ⎠ l

which gives the displacement of the string for a given time.

EXAMPLE 6.39 An infinitely long string having one end at x  0 is initially at rest on the
x-axis. The end x  0 undergoes a periodic transverse displacement described by A0 sin ω t , t > 0.
Find the displacement of any point on the string at any time t.

Solution Let u ( x, t ) be the transverse displacement of the string at any point x and at
any time t. Then the transverse displacement of the string is described by the
PDE: utt = c 2u xx , x > 0, t >0
BCs: u (0, t ) = A0 sin ω t , t > 0 ( A0 = constant)
ICs: u ( x, 0) = 0, ut ( x, 0) = 0, x≥0
The Laplace transform of the PDE gives

d 2U
s 2U  su ( x, 0)  ut ( x, 0)  c 2 (6.88)
dx 2
After using the ICs, Eq. (6.88) becomes

d 2U s2
 U ( x, s )  0 (6.89)
dx 2 c2
Its general solution is found to be

U ( x, s ) = Ae( s / c ) x + Be −( s / c ) x

Since the displacement U ( x, s) is bounded as x n f, we get A = 0 and, therefore,

U ( x, s ) = Be−( s / c ) x (6.90)
Now taking the Laplace transform of the BC, we obtain
A0X
U (0, s ) 
s X2
2

Using this expression, from Eq. (6.90), we have

A0X
B
s X2
2
 LAPLACE TRANSFORM METHODS 375

Hence the solution of Eq. (6.89) is found to be

A0ω
U ( x, s ) = e−( s / c ) x (6.91)
s +ω2
2

Finally, taking the inverse Laplace transform of Eq. (6.91), we get

⎡ Aω ⎤
u ( x, t ) = L−1 ⎢ 2 0 2 e− ( s / c ) x ; t ⎥
⎣s +ω ⎦
⎧ ⎛ x⎞ x
⎪⎪ A0 sin ω ⎜⎝ t − c ⎟⎠ if t >
c
=⎨
⎪0 x
if t <
⎪⎩ c

6.14 MISCELLANEOUS EXAMPLES

¦ 1 µ
EXAMPLE 6.40 Find the Laplace transform of erf § ¶.
¨ t·
Solution Following the definition of Laplace transform, we have

f  st 2 1/ t u 2
L[erf (1/ t ); s ]  ± e e du dt
0 Q ±0
2 f 1/ t  st u 2
 e du dt
Q ±0 ±0

Now, changing the order of integration (see Fig. 6.5), we get

2 f u 2 1/u 2  st
L [erf (1/ t ); s]  e du ± e dt
Q ±0 0

2f u 2  s / u 2 2
 ± (e  eu ) du (6.92)
s Q 0

To evaluate this integral, consider an integral of the form

f 2 2
u b / u 2
I ( a, b)  ± e  a du (6.93)
0

We can easily verify that

f 2 2 Q
I (a, 0)  ± e  a u
du  (6.94)
0 2a
376 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

u = 1/–
t

t
O

Fig. 6.5 Curve of 1/ t.

since
2 f u 2
e du  1
Q ±0
Differentiating Eq. (6.93) partially with respect to b, we get
wI f 2 2 2 2
 2b ± u 2 ea u b /u du (6.95)
wb 0

Let au = x; then du = dx /a, and we can easily verify that

1 f 1
I ( a, b)  ± exp ( x 2  a 2b 2/ x 2 ) dx  I (1, ab) (6.96)
a 0 a

Also, let b /u = x; then we find from Eq. (6.95) that

w I ( a , b) 0 x2 §  a 2b2 · b
wb
2b ³f b 2
exp ¨
© x 2
 x 2 ¸ 2 dx
¹x

f §  x 2  a 2b 2 ·
2 ³0 exp ¨
© x2
¸¹ dx

2 I (1, ab) (6.97)

From Eqs. (6.96) and (6.97), we can eliminate I (1, ab) and arrive at

wI
2aI
wb
 LAPLACE TRANSFORM METHODS 377

On integration, we get
ln I  2ab  ln c, I (a, b)  ce2 ab
Thus,

I (a, b)  I (a, 0) e2 ab

f  a 2u 2 b 2/u 2 Q
±0 e du  e2 ab (6.98)
2a
By making use of Eq. (6.98), the Laplace transform of the given function is obtained from
Eq. (6.92) as
2 ⎛ π −2 s π⎞ 1 −2 s
L[erf (1/ t ); s ] = − ⎜ e − ⎟ = (1 − e )
s π ⎝ 2 2 ⎠ s

−1/ s
EXAMPLE 6.41 Find the inverse Laplace transform of e .
s

Solution The given function can be rewritten as

¥ ¥
f f
1 1/ s 1 (1)n (1)n
e  
s s n  0 n ! s n n  0 n ! s n 1 2

Therefore,

© e 1/s
¥ (1)n t n 1/2
¥
f f
¸ (1)n 1 © 1 ¸
L1 ª ; t¹  L ª n 1/2 ; t ¹ 
« s º n 0
n! «s º n 0 n ! n  1/2

But,
1 (2n)!
n+ = π
2 22 n n !
Hence,
© e 1/ s ¸ f (1) n 22 n t n 1/2
L1 ª ; t¹  4
« s º n 0 (2n)! Q

t 1/2 4t1/2 24 t 3/2

   
Q 2! Q 4! Q

1 © (2 t )2 (2 t )4 ¸
 ª1    ¹
Qt « 2! 4! º

cos 2 t

Qt
378 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

EXAMPLE 6.42 Solve

x  ky  a sin kt , y  kx  a cos kt
using Laplace transform technique, given that
x(0)  0, y (0)  b

Solution Taking the Laplace transform of the given equations, we get

ak
sX − x(0) + kY = (6.99)
s + k2
2

as
sY − y (0) − kX = (6.100)
s + k2
2

Utilizing the given ICs, we obtain

ak
sX + kY = 2 2
(6.101)
s +k
as
sY  kX 
b (6.102)
s  k2 2

Solving Eqs. (6.101) and (6.102) and using Cramer’s rule, we obtain

ak
k
s  k2
2 s k bk
X  2 2
as k s s k
b s
s2  k 2
Similarly, we can show that
a bs
Y= 2 2
+
s +k s + k2
2

Taking the inverse Laplace transform of the above two equations, we get
a
x  b sin kt , y sin kt  b cos kt
k
which is the required solution.

EXAMPLE 6.43 A beam which is coincident with the x-axis is simply supported at its end
x  0 and is clamped at the other end x  l. Let a vertical load W act transversely on the
beam at x = l /4. The differential equation for deflection at any point is given by

d4y
W
δ ( x − l /4)
4
=
dx EI
where EI is the flexural rigidity of the beam. Find the deflection at any point.
 LAPLACE TRANSFORM METHODS 379

Solution Taking the Laplace transform on both sides of the governing differential
equation and noting the transform of Dirac delta function from Eq. (6.23), we obtain
W
s 4Y − s 3 y (0) − s 2 y ′(0) − sy ′′(0) − y ′′′(0) = exp [−(l /4) s ] (6.103)
EI
Since the beam is simply supported at x  0, we have
y (0)  y bb(0)  0 (6.104)
Also, it is given that the beam is clamped at x  l , which means that
y (l )  y b(l )  0 (6.105)

Let y b(0)  A and y bbb(0)  B. Then using Eq. (6.104) into Eq. (6.103), we get

W −(l /4) s
s 4Y − As 2 − B = e
EI

A B W e−(l /4) s
Y= + +
s2 s4 EI s 4
Taking the inverse Laplace transform, we obtain

Bx3 W −1 ⎡ −(l /4) s 1 ⎤

y = Ax + + L ⎢{e } ⋅ 4 ; x⎥ (6.106)
3! EI ⎣ s ⎦
But, from the second shifting property (Theorem 6.14), we have

⎡ 1 ⎤
L−1 ⎢{e− (l /4) s } ⋅ 4 ; x ⎥ = f ( x − l /4) H ( x − l /4)
⎣ s ⎦
where

© 1 ¸ x3
f ( x)  L1 ª 4
; x¹ 
«s º 3!
Hence Eq. (6.106) becomes

B 3 W ( x  l/4)3
y  Ax  x  H ( x  l/4)
3! EI 3!
Thus, the deflection is given by

B 3
y = Ax + x for 0 < x ≤ l /4
6
B 3 W ( x − l /4)3
= Ax + x + for l /4 < x < l
6 EI 6
380 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

The unknowns A and B can now be determined by using the conditions (6.105). Thus, we
have

B 3 9 Wl 3
0 = Al + l +
6 128 EI

B 2 9 Wl 2
0 = A+ l +
2 32 EI
whose solution gives

9 Wl 2 81 W
A= , B=−
256 EI 128 EI
Hence the resulting deflection y is given as

⎧ 9 Wl 3 27 Wl 3
⎪ − , 0 < x ≤ l /4
⎪ 256 EI 256 EI
y=⎨
⎪ 18 Wl 3 W
⎪− + ( x − l /4)3 , l /4 < x < l
⎩ 256 EI 6 EI

EXAMPLE 6.44 Find the solution of the

PDE: ut  ku xx , 0  x  f, t0
subject to

BCs: u ( x, t ) n 0 as x n f
u (0, t )  g (t )
IC: u ( x, 0)  0

assuming

© ¦ s µ ¸ x ¦ x2 µ
L1 ªexp §  x ¶; t ¹  exp §  ¶
« ¨ k · º 2 kQ t 3 ¨ 4kt ·

Solution Taking the Laplace transform of the PDE, we have

d 2U
sU ( x, s )  u ( x, 0)  k (6.107)
dx 2
 LAPLACE TRANSFORM METHODS 381

Using IC: u ( x, 0) = 0, we obtain

d 2U s
2
 U ( x, s )  0 (6.108)
dx k
whose general solution is given by

¦ s µ ¦ s µ
U ( x, s )  A exp § x ¶  B exp §  x ¶
¨ k · ¨ k ·
The Laplace transform of the first BC gives:
U n 0 as x n f
Using this in the solution, we get A  0 and

¦ s µ
U ( x, s )  B exp §  x¶ (6.109)
¨ k ·
The Laplace transform of the second BC gives
U (0, s) = G ( s)
Using this condition, Eq. (6.109) becomes
¦ s µ
U ( x, s )  G ( s ) exp §  x¶ (6.110)
¨ k ·
Taking inverse Laplace transform, we get

Ë È s Ø Û Ë Ë x ÛÛ
u ( x, t ) L1 ÌG ( s ) exp É  xÙ ; s Ü L1 Ì L [ g (t ); s] L Ì exp(  x 2/4kt ); s Ü Ü
Í Ê k Ú Ý ÌÍ ÌÍ 2 kS t 3 ÜÝ ÜÝ

Finally, through the use of the convolution theorem, we arrive at the result
t x exp[ − x 2 /4k (t − u )]
u ( x, t ) = ∫0 2 π k (t − u )3/2
g (u ) du

EXAMPLE 6.45 Find the solution of IBVP described by

PDE: ut ( x, t ) = u xx ( x, t ) − hu ( x, t ),
h = constant, 0 < x < π , t > 0
BCs: u (0, t ) = 0, t>0
u (π , t ) = 1, t>0
IC: u ( x, 0) = 0, t=0
382 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

Solution Taking the Laplace transform of the PDE with respect to the variable t, we
have

d 2U
sU ( x, s )  u ( x, 0)   hU ( x, s) (6.111)
dx 2
Using the IC: u ( x, 0) = 0, we get

d 2U
 (h  s)U ( x, s)  0 (6.112)
dx 2
Its general solution is found to be
U ( x, s) = A cosh ( h + s x) + B sinh ( h + s x) (6.113)

Taking the Laplace transform of BCs, we obtain

U (0, s)  0, U (Q , s )  1/s (6.114)

Using these BCs into Eq. (6.113), we have

U ( x, s ) = B sinh ( h + s x)

1
= B sinh ( h + s π )
s
implying thereby
1 1
B=
s sinh ( h + s π )
Hence, the required solution of Eq. (6.112) is

1 sinh ( h + s x)
U ( x, s ) =
s sinh ( h + s π )
By means of complex inversion formula, we get

1 H  if e st sinh ( h  s x )
u ( x, t ) 
2Q i ± H if s sinh ( h  s Q )

 sum of the contributions from all the poles of the integrand

The poles are given by s = 0, and sinh ( h + s π ) = 0 = sin (i h + s π ). Therefore,

i h + s π = nπ , implying h + s = −n 2
 LAPLACE TRANSFORM METHODS 383

Thus, the integrand has poles at s  0, and

s = − n 2 − h, n = 1, 2, 3, …
The residue of the expression

e st sinh ( h + s x)
at s = 0
s sinh ( h + s π )
is
sinh ( h x )
sinh ( h π )

The residue of the integrand at s   n 2  h is

2
e( − n − h )t
sinh ( −n2 x)
, n = 1, 2, 3, ...
( − n 2 − h) π
sinh −n π + 2
cosh ( −n π ) 2
2 −n 2

Using the relations sinh x  i sin ix, and cosh x  cos ix, the above residues become

exp [(n 2  h)t ]i sin nx 2n exp [(n 2  h)t ]sin nx


(n 2  h)Q (n 2  h)Q cos n Q
cos nQ
2in

2n exp [(n 2  h)t ]sin nx


(n2  h)Q (1)n

Hence, the required solution is

f
sinh ( h x) 2e − ht n( −1) n exp ( − n 2t ) sin nx
u ( x, t ) =
sinh ( h π )
+
π ∑ n2 + h
n =1

EXERCISES
1. Find the Laplace transform of the following:
(i) t cos at , (ii) te t sin t ,

(iii) tet sin 2t , (iv) (1 + t )2 e − at .

384 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

2. Find the Laplace transform of f (t ) defined as

⎧t /τ , 0 < t <τ
(i) f (t ) = ⎨
⎩1, t >τ

⎧t , 0 < t <1
⎪
(ii) f (t ) = ⎨(1 − t ), 1< t < 2
⎪0, t>2
⎩
3. Find the Laplace transform of

(ii) cos 2t  sin 3t .

1 − et
(i) ,
t t
4. Obtain the Laplace transform of
¬t , 0 c t 1
f (t )  ­
®0, 1c t  2

given f (t  2)  f (t ) for t  0.
5. Find the Laplace transform of
¬t , 0ct 6
f (t )  ­
®12  t , 6 c t  12

given f (t  12)  f (t ) for t  0.

6. Find the Laplace transform of
¬1, 0ct b
f (t )  ­
®1, b c t  2b

given f (t  2b)  f (t ).
7. Find the Laplace transform of erfc (1/ t ).
8. Find the Laplace transform of
(i) J1 (t ), (ii) tJ1 (t ).
9. Find the inverse Laplace transform of
1 , s
(i) (ii)
s ( s + 1)( s + 2) ( s  2)3

(iii) 1 (iv) 4s + 5 ,
,
s ( s  1)
2
( s − 1)2 ( s + 2)

(v) s+3 , (vi) s2 + 1 .

( s 2 + 6s + 13)2 s3 + 3s 2 + 2s
 LAPLACE TRANSFORM METHODS 385

10. Find the inverse Laplace transform of

s+a, ⎛s⎞
(i) ln (ii) tan −1 ⎜ ⎟ ,
s+b ⎝k⎠

s2 + 1 , s−4 .
(iii) ln (iv) ln
( s − 1) 2 4 + s2
11. State and prove the convolution theorem for Laplace transforms.
12. Using the convolution theorem, find the inverse Laplace transform of
1 1 s .
(i) , (ii) , (iii)
2 2
s (s + a ) 2 s ( s − a) ( s + 4)3
2

13. Find the inverse Laplace transform of

e −π s /( s 2 + 1)
14. Using the Heaviside expansion theorem, find the inverse Laplace transform of
1 3s + 1 .
(i) , (ii)
3
s +1 ( s − 1)( s 2 + 1)
15. Find the inverse Laplace transform of
cosh ( x s ) ,
F (s) = 0 < x <1
s cosh s
using the complex inversion formula.
16. Solve the following ODE using the Laplace transform
y bb  3 y b  2 y  4e 2t
given that y (0) = −3, y ′(0) = 5.
17. Solve the ODE by the Laplace transform method

d4y
 k4 y  0
dt 4
with the initial conditions y (0)  1, y b(0)  y bb(0)  0, y bbb(0)  0.
18. Solve the initial value problem using the method of Laplace transform
y bb  ty b  y  0, y (0)  1, y b(0)  0
19. Using the Laplace transform method, solve the system of equations
dx dy
  x  y 1
dt dt
dy
 2x  y
dt
given that x  0 and y  1 when t  0.
386 INTRODUCTION TO PARTIAL DIFFERENTIAL EQUATIONS

20. Applying the Laplace transform technique, solve the system:

dx dy
  x  e t  0
dt dt
dx dy
 2  2x  2 y  0
dt dt
given that x  1 and y  1 when t  0.
21. Using the Laplace transform, find the solution of the system of ODEs

d2x d2x
 y  1, x0
dt 2 dt 2
satisfying the ICs x(0) = y (0) = x′(0) = y ′(0) = 0 .
22. Show that, by means of the Laplace transform, the solution R ( x, t ) of one-dimensional
diffusion equation
w 2θ 1 wθ
, 0 d x dπ, t!0
wx 2 k wt
satisfying the boundary conditions

θ (0, t ) = 1 − e−t , t >0

θ (π , t ) = 0, t >0
θ ( x, 0) = 0, t = 0, 0 ≤ x ≤ π
is given by the formula

et sin {(Q  x) / k } 2

¥
f
Q x sin nx exp (n 2 kt )
R ( x, t )   
Q sin(Q / k ) Q n 1 n(n 2 k  1)
23. Using the Laplace transform method, solve
PDE: ut = 3u xx
⎛π ⎞
BCs: u ⎜ , t ⎟ = 0, u x (0, t ) = 0
⎝2 ⎠
IC: u ( x, 0) = 30 cos 5 x
24. Using the Laplace transform, solve the following problem in wave propagation:

PDE: c 2u xx = utt , 0 < x < l, t >0

BCs: u = 0 at x = 0
Eut = P at x = l , t > 0 ( E , P are constants)
ICs: u = ut ( x, 0) = 0, 0< x<l
 LAPLACE TRANSFORM METHODS 387

25. Solve the BVP using the Laplace transform method

PDE: utt  c 2u xx , 0 c x  f, 0ct f

BC: u (0, t )  0

ICs: u ( x, 0)  0,

ut ( x, 0)  1

26. If G is the potential, i the current, l the inductance per unit length of cable, c the
capacitance to ground per unit length, then G satisfies the wave equation
G xx  lcGtt
Initially, the line is considered to be dead, i.e.
φ ( x, 0) = φt ( x, 0) = 0
The other boundary conditions are
φ (0, t ) = f (t ), t>0

φ x (l , t ) = φ (l , t ) = 0, t>0

Find the potential at any point on the cable at any time t, assuming l  f.