Application of Laplace and Fourier Transform
Application of Laplace and Fourier Transform
UNIT 15
APPLICATIONS OF
FOURIER AND LAPLACE
TRANSFORMS
Structure
15.1 INTRODUCTION
In Units 13 and 14, you have learnt the basic concepts related to Fourier and
Laplace transforms, their definitions, properties and a few applications in
physics. In this unit, we will discuss some more applications of these
transforms in solving initial and boundary value problems involving differential
equations, particularly in Physics.
Fourier and Laplace transforms are widely used in physics and engineering for
solving differential equations. For example, Fourier transforms are used in
solving partial differential equations such as Laplace’s equation, wave equation,
diffusion equation with given initial and boundary conditions. Laplace’s
transforms can be applied to study the motion of a particle under a dissipative
force, forced oscillations, radioactive disintegration, electrical circuits subjected to
short pulses, etc.
In order to study this unit effectively you require sound knowledge of integral
calculus and practice of evaluating definite integrals. You must also know how to
reduce an algebraic fraction to a partial fraction. If you have thoroughly studied
Units 13 and 14 and worked out the SAQs and TQs, then you already have
enough practice and you will find this to be an easy and interesting unit.
Expected Learning Outcomes
After studying this unit, you should be able to:
apply the method of Fourier transforms to solve initial and boundary value
problems in physics; and
apply the method of Laplace transforms to solve initial and boundary value
problems in physics. 49
Block 4 Fourier and Laplace Transforms
15.2 APPLICATIONS OF FOURIER TRANSFORMS
In Unit 13, you have already learnt of one application of Fourier transform in
physics, viz. the Dirac delta function. Let us now consider the applications of
Fourier transforms in solving differential equations. We first take up the example
of the wave equation, which you have studied in your UG physics courses.
Example 15.1
Solve the wave equation for the amplitude of vibrations of an infinitely long
vibrating string given by:
2 y 1 2 y
(15.1a)
x 2 v 2 t 2
Assume the initial condition
y ( x, 0) f ( x ) (15.1b)
2 y x, t i x 1 y x, t i x
2
x 2 e dx
v 2 t 2
e dx
Now, using the expression for the second derivative of the Fourier transform
from Unit 13 (given below for ready reference), we can write:
d 2 y ( x, t )
Fourier Transform ( i )2Y ( , t )
dx 2
So, we have:
1 2Y , t
i 2Y , t (15.2a)
v2 t 2
where we have used the fact that x and t are independent of each other along
with the following definition:
1
y x, t e
i x
Y (, t ) dx (15.2b)
2
Since no derivatives with respect to appear in Eq. (15.2), it is an ordinary
differential equation. In fact, it is the linear oscillator equation with the solution
Y (, t ) C e iv t (15.3)
1 i x
y ( x, t ) Y (, t ) e
2
dx
Since f(x) is the inverse Fourier transform of F , we obtain the solution of
Eq. (15.1a) as:
y ( x, t ) f ( x vt ) (15.6)
You have studied the heat conduction equation in your UG courses in Physics.
You know that this equation can also be used to model the diffusion of any
substance in a medium. Let us take up the example of diffusion of salt in
water.
Example 15.2
Consider the problem of diffusion of a certain quantity of salt in water.
Suppose at time t 0, salt of mass M is introduced at a point x 0 in water
flowing in a long and narrow pipe (Fig. 15.1). Obtain the concentration of salt
( x, t ) at a later time t that satisfies the following one-dimensional diffusion
equation:
2
D (15.7a)
t x 2
with the initial condition
M
( x,0) x x0 (15.7b)
A
The initial condition expresses the fact that at t 0, ( x,0) is zero
everywhere except at x x0 where salt of mass M has been introduced over a
cross-section A of the pipe. The boundary condition is:
The boundary condition reflects the fact that the total amount of salt is
conserved.
Salt introduced at t = 0
x0
We now use Eq. (15.7a) in Eq. (15.9) along with the expression for the second
derivative of the Fourier transform from Unit 13
d 2f ( x )
Fourier Transform ( ik )2 g ( k ) where g (k ) is the Fourier transform
dx 2
of f(x). Thus,
d 2
Fourier Transform ( ik )2 R( k, t )
dx 2
So, we get:
dR(k, t ) 1 2 x, t ikx
dt
2
D
x 2
e dx
Dk 2 R k , t
M 1 ik x x0 Dk 2t
A 2
e e dk
52
Unit 15 Applications of Fourier and Laplace Transforms
We can express the exponent as a square and write:
x x0
2
x x0 2
Dt k ik x x0 Dt k
2
2i Dt 4 Dt
Then changing the variable of integration to
x x0
k Dt k ,
2i Dt
we get:
M 1 1 x x 2
ek dk
2
( x, t ) exp 0
A 2 Dt 4 Dt
u 2 du
Using the result e , we obtain the expression for the
M 1 x x 2
( x, t ) exp 0
(15.11)
A 4Dt 4 Dt
The solution ( x, t ) is shown in Fig. 15.2. Notice that for small t the spread in
( x, t ) is less, and vice versa.
( x, t )
Small t
Medium t
Large t
x0 x
Before studying further, you might like to practice this technique. Solve SAQ 1.
SAQ 1
The one-dimensional Fermi age equation for the diffusion of neutrons slowing
down in some medium (such as graphite) is:
2q x, q x,
x 2
Here q is the number of neutrons that slow down (falling below some given
energy per second per unit volume) and , called the Fermi age is a measure
of the energy loss. If the source of neutrons is planar, emitting N neutrons at
x = 0, per unit area per second, i.e., qx,0 S x , obtain the solution q x,.
In the next example, we use the Fourier transform method to solve Laplace’s
equation for steady state temperature distribution. 53
Block 4 Fourier and Laplace Transforms
Example 15.3
Consider an infinite metal plate placed in the xy plane (Fig. 15.3). Its edge
along the y-axis is maintained at temperature 0°C and the temperature in its
edge along the x-axis is given by:
100 C 0 x 1
T ( x, 0)
0 C x 1
1 x
T 1000 C T 00 C
Fig. 15.3: Temperature distribution in an infinite metallic plate.
Solution : From Unit 1 of this course, you know that the general solution of
T x, y is:
When y 0, we have
T ( x,0) 0 B k sin kx dx (15.14)
This is a Fourier sine transform of B(k). Thus B(k) is the inverse Fourier sine
transform of T x,0.
2 2
B( k ) gs (k ) T ( x,0) sin kx dk
0
For T ( x,0) given in this problem, we get
2 1
0
B( k ) 100 sin kx dk
1
200 cos kx
k 0
200
1 cos k
54 k
Unit 15 Applications of Fourier and Laplace Transforms
Therefore,
T ( x, y )
200 1 x 1
tan1
x 1 1 tan1 x 1 (15.16)
tan
y 2 y 2 y
Example 15.4
when it is not acted upon by any external driving force. Here primes denote
differentiation with respect to time t. Let us solve this equation with the
particular values m 5, b 1, k 50 and the initial conditions:
The initial conditions of Eq. (15.17b) tell us that we start measuring time
when the displacement of the particle is 1 and at this position, its velocity is
3. If we take the positive y-axis to the right of the origin as per the normal
convention, then this means that at time t 0, the particle is to the right of the
origin (at 1) and is moving towards the origin (velocity, 3) (Fig. 15.4b).
55
Block 4 Fourier and Laplace Transforms
(a)
(b)
O P
Fig. 15.4: a) A spring mass system at its equilibrium position; b) the oscillating
system at the instant t 0.
What we really do in going from Eq. (15.17a) to Eq. (15.18) is that we multiply
every term of Eq. (15.17a) by e st and integrate over t from 0 to . Since m, b
and k are constants independent of t, they can be taken out of the operation L
by the linearity property of the Laplace transforms. Let us denote the Laplace
transform of y(t) by Y(x), that is,
L y (t ) Y ( x ) (15.19)
We now write Eq. (15.19) with particular given values of the parameters, using
the expressions for Laplace transforms of the first and second derivatives of
y (t ) [Eqs. (14.16a and b) of Unit 14]:
Here you must remember that y (0) and y (0) are known while Y ( x ) is the
unknown for which we must solve the above equation.
Substituting the initial values in Eq. (15.20) from Eq. (15.17b) and collecting
terms, we get the equation:
or (5 x 2 x 50)Y ( x ) 5 x 14 (15.21)
giving
5 x 14 x 2.8
Y (x) (15.22)
2 2
5 x x 50 x 0.2x 10
We must now obtain the inverse Laplace transform of Eq. (15.22). For this, we
rewrite the denominator of the function Y ( x ) as:
Y (x)
x 0.1 2.9
(15.23)
x 0.12 9.99
If we replace x 0.1 by x, we notice that the function on the right-hand side of
Eq. (15.23) becomes
x 2.9 x 0.918 3.161
F(x) (15.24)
x 3.161 x 2 3.161
2 2 2
x 9.99 2
But, from Eq. (15.19), the inverse Laplace transform of Y x is just the
unknown displacement y (t ), which we are trying to determine, that is,
L1 Y ( x ) y (t )
This is the solution of the differential equation (15.17a), with given values of m,
s, k, subject to the initial values given in Eq. (15.17b).
To complete the solution, we must also interpret Eq. (15.27).
You can see that the solution y t is a product of two factors, one of which is
e0.1t and the other is cos t 0.918 sin t , where we have used 3.161 .
Note that is the angular frequency of the oscillatory motion. The second
factor is a combination of the trigonometric function cos t and sin t, while
the first factor is an exponentially decreasing function of time.
Let us consider only the second factor first. We can write it as
cos t 0.918 sin t a sin(t c ),
or c 2.313 rad.
Thus, we have
f (t ) cos t 0.918sin t 1.357sin(t 2.313)
f(t)
1.0
e-0.1t
0.5
0
2 4 6 8 10 t
-0.5
-1.0
Fig. 15.5: The solid line shows the funct on f (t ) a sin(t c ) with the given
0.1t
values of a and c, dotted line shows the function e .
However, the complete solution y(t) is a product of this function and e 0.1t .
Thus, the motion of the mass attached to the spring would be like that shown
in Fig. 15.6, a sinusoidal function of decreasing amplitude and of period 1.988.
The decrease in the amplitude is a result of damping which may be due to
friction. This completes the solution of Example 15.4, with specific values of
parameters as given in the problem. You could have also worked out the
solution with the general parameters m, s, k without using any particular
values, and arrived at a general solution.
1.0
0.5
0
2 4 6 8 10 t
-0.5
-1.0
Differential equations like Eq. (15.17a) (along with initial conditions) arise in
several branches of physics – in solving problems in electrical networks, in
58 radioactivity where one element decays into another, which decays into a third
Unit 15 Applications of Fourier and Laplace Transforms
one and so on, in problems involving mechanical systems, etc. Another typical
problem is that of forced vibrations. To bring home the power of this method
we now consider a damped vibrating system subjected to a force that acts for
a short time. But before that, you should attempt an SAQ to practice the
method.
SAQ 2
y 4y 3y 0, y 0 3, y 0 1
Example 15.5
f(t)
y 5y 4y f t , y 0 0, y 0 0 (15.28)
with
1 , 0 t 1
f t
0 otherwise
0 1 t
The forcing function f(t) is shown in Fig. 15.7. This equation can model the
response of an LCR circuit to an applied square pulse or a damped vibrating Fig. 15.7: The forcing
system subjected to a single square pulse. function f(t).
Solution: We take the Laplace transform of Eq. (15.28) and set up the
subsidiary equation.
Let Y (s ) L y (t )
Then we have
s 2Y 5sY 4Y
1
s
1 e s
where we have used the result for SAQ 3 of Unit 14 with appropriate changes.
1 es
Y (s )
1 e s F (s )
2
s s 5s 4
1
where F (s )
s(s 1)(s 4)
1 1 1
4s 3 s 1 12 s 4
Using this result, we can apply the Laplace transformation method to impulsive
forces.
Example 15.6
Newton’s second law for impulsive force acting on a particle of mass m
moving along a straight line is
d 2x
m F t
dt 2
where F (t ) p (t ), with p constant, is an impulsive force acting for a very
short time. Obtain x(t ) given x(0) 0, x(0) 0.
Solution: Let Lx(t ) X (s ).
Applying Laplace transform to the equation of motion we get:
ms 2 X (s ) ms x(0) m x (0) p since L (t ) 1
60
so that
Unit 15 Applications of Fourier and Laplace Transforms
dx (t ) p
is a constant.
dt m
Thus the effect of the impulse pt is to transfer, instantaneously, p units of
linear momentum to the particle.
2 E x, t 1 E x, t
2
0
x 2 v2 t 2
E ( x, t )
E ( x,0) 0 and 0
t t 0
L E ( x, t ) e sx / v F (s )
15.4 SUMMARY
In this unit, we have discussed the following:
Applications of Fourier transforms and Laplace transforms in
solving
partial and ordinary differential equations;
PDEs and ODEs in physics.
or ( ik )2 g (k, ) g ( k , )
where we have used
g k, q x, e ikxdx
1
2
62
Unit 15 Applications of Fourier and Laplace Transforms
or k 2 g ( k , ) g ( k , )
This is an ODE and its solution is:
since
e
ikx
x dx e ikx x 0 1
Thus,
S
g ( k , ) exp( k 2 )
2
The inverse Fourier transform of g (k , ) yields the solution:
q( x, )
1
2
g ( k , ) e
ikx
dk
S
2
exp k 2 ikx dk
S x2
q ( x, ) exp
4 4
s 2Y s 3s 1 4sY s 3 3Y s 0
or Y (s )[s 2 4s 3] 3s 4 0
3s 4 5 1
or Y s
s 3 s 1 2 s 3 2 s 1
5 1 1 1 1 1
Thus, y (t ) L L s 1
2 s 3 2
5 3t 1 t
e e
2 2
1 t
2
e 5e 3t
Terminal Questions
1. We will follow the same procedure as in Example 15.3. However, in this
T
case we use the cosine solution as T 0 when x 0 but 0 when
x
x 0. Thus, the general solution will be of the form: 63
Block 4 Fourier and Laplace Transforms
ky
0 B(k ) e cos kx dk
When y 0, we have
T ( x, 0) 0 B(k )cos kx dk
This is a Fourier cosine transform of B(k). Thus
2
0
B( k ) T ( x,0)cos kx dx
2 2
200 100 x cos kx dx
0
B(k )
400 2 200 2
0
cos kx dx
0
x cos kx dx
200
1 cos 2k
k 2
Therefore,
1
r e
ip.r /
2. g ( p) dV
2 3/2
1 1 r / a0 ip.r /
g ( p) e e dV
2
32 12
a03
Let
I er / a0 eip.r / dV
2 r / a ipr cos /
0 0 0 0 e e r 2dr sin d d
ibr cos
2 dr r 2 e ar 0 e sin d
0
here a 1/ a0, b p /
64
Unit 15 Applications of Fourier and Laplace Transforms
Putting cos t, we have:
I1 2 dr r 2 e ar ibrt
1
0 1e dt
4 ar
b 0
re sin br dr
4
lm r e ar ibr dr
b 0
4 1 8a
Im
a ib 2 a2 b2
b 2
8 / a0 8a03 4
I
2 p2 a02
2 2
1 p2
a2 2
0
1 1 8a03 4
Therefore, g ( p)
2 3/2
1/2 2
a03 2
p2a02
23/2 a03 2 5 2
a02 p2 2
2
3. Taking the Laplace transform and using Eqs. (14.16a and b) of Unit 14, we
get
s 2Y s s 8 2sY s 2 8Y s 0
s 10 s 10
or Y (s )
2
s 2s 8 s 4 s 2
1 2
s4 s2
1 1 1
Therefore, y (t ) L1 2L s 2
s 4
e4t 2e2t
V 0 t 1
where v (t ) 0
0 t 1 65
Block 4 Fourier and Laplace Transforms
Taking the Laplace transform of Eq. (i), we get
I (s )
R I (s ) L v (t )
sC
We can express v(t) in terms of the unit step function as
v (t ) V0 1 H t 1
1 t 1
where H (t 1)
0 t 1
1 e s V0
Therefore, Lv (t ) V0
s
s
s
1 e s
and
I (s ) R
1 V0
sC
s
1 e s
or
I (s ) F (s ) 1 es F (s ) F (s )es
V0 / R
where F (s )
s 1/ RC
V0 t / RC
Now L1(F ) e
R
Then applying the second shifting theorem we get
V0 e t / RC e t 1 / RC H t 1
R
V0 t / RC
Thus i (t ) e 0 t 1
R
V0
R
1 e1/ RC e t / RC t 1
66