We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 126
INTEGRAL TRANSFORMS
IN MATHEMATICAL
PHYSICS
by
C. J. TRANTER, M.A.
ASSOCIATE PROFESSOR OF MATHEMATICS,
MILITARY COLLEGE OF SCIENCE, SHRIVENHAM
WITH 7 DIAGRAMS
LONDON: METHUEN & CO. LTD.
NEW YORK: JOHN WILEY & SONS, INC.First published in 1951
CATALOGUE NO. 4050/U
PRINTED IN GREAT BRITAINCONTENTS
CHAPTER
INTRODUCTION
I INTEGRAL TRANSFORMS AND THEIR INVERSION
FORMULAE
Integral transforms. Fourier’s integral formula.
Inversion formulae. Summary. Examples.
Il THE LAPLACE TRANSFORM
Application of Laplace transform to ordinary
differential equations. A simple problem in heat
conduction. On the evaluation of the contour
integral of the inversion formula. An example in
the radial flow of heat. Special solutions for small
values of the time. Verification of solution.
Examples.
III FOURIER TRANSFORMS
Sine and cosine transforms. The heat conduc-
tion problem of § 2.3 solved by a sine transform.
An example of the repeated use of transforms.
The complex Fourier transform. The motion of
a very long string. Stresses in a long circular
cylinder. Examples.
IV HANKEL AND MELLIN TRANSFORMS
The problem of Boussinesq. The electrified
disc. An application of the Mellin transform.
Examples.
V THE NUMERICAL EVALUATION OF INTEGRALS IN
SOLUTIONS
Values at end points. Asymptotic series. Filon’s
method for trigonometrical integrals. A worked
example. Examples.
Ve
PAGE
vii
18
32
60vi TRANSFORMS IN MATHEMATICAL PHYSICS
CHAPTER
VI FINITE TRANSFORMS
Finite Fourier transforms. Example of the use
of the finite sine transform. Example of the
repeated use of finite transforms. Example of the
use of the finite cosine transform. Example in
heat conduction with radiation boundary condi-
tion. The finite Hankel transform. The problem
of §2.5 solved by a finite Hankel transform.
Heat flow in a cylinder with radiation at the
surface. Motion of viscous fluid between con-
centric cylinders. Extension to other kernels.
Legendre transforms. Electrified disc solved by
a Legendre transform. Examples.
VII THE COMBINED USE OF RELAXATION METHODS
AND INTEGRAL TRANSFORMS
Introduction. The finite cylinder. The semi-
infinite cylinder. Examples.
BIBLIOGRAPHY
INDEX
PAGE
77
104
Ii!
116INTRODUCTION
THE classical methods of solution of the boundary-value
problems of mathematical physics may be said to be
derived from Fourier’s pioneer work. Another technique,
that of integral transforms, which had its origin in Heavi-
side’s work, has been developed during the last few years
and has certain advantages over the classical method. It
is the purpose of this monograph to give an outline of
the use of integral transforms in obtaining solutions to
problems governed by partial differential equations with
assigned boundary and initial conditions.
Heaviside (about 1890) was originally interested in the
solution of the ordinary differential equations with con-
stant coefficients occurring in the theory of electric cir-
cuits. Later he extended his method to the partial
differential equations occurring in problems of electro-
magnetism and heat conduction. The power of his
method was such that he solved many hitherto intractable
problems and obtained solutions to problems already
solved in a form better adapted for numerical compu-
tation. Later investigations by Bromwich, Carson and
van der Pol placed the Heaviside calculus on a sound
foundation.
The theory developed by Heaviside, Bromwich and
Carson has been unified in recent work by Doetsch
(and others) on the Laplace transformation. The solu-
tion found from Heaviside’s calculus is obtained from
Laplace’s integral equation and the contour integral
appearing in Bromwich’s work is the integral in the
inversion theorem for the Laplace transform.
The use of an integral transform will often reduce a
partial differential equation in ” independent variables to
one in (”-1) variables, thus reducing the difficulty of the
problem under discussion. Successive operations of this
type can sometimes ultimately reduce the problem to the
vilvill TRANSFORMS IN MATHEMATICAL PHYSICS
solution of an ordinary differential equation, the theory
of which has been extensively developed. Successive
operations could, in fact, reduce the problem to the solu-
tion of an ordinary algebraic equation, but this is only
sometimes worth while.
Although the Laplace transform has been more widely
used, and is particularly suitable for problems governed
by ordinary differential equations and for problems in
heat conduction, other integral transforms can be very
useful in the solution of the boundary-value problems
of mathematical physics. Several different integral trans-
forms so far have been successfully used in the solution
of this type of problem, and there is no reason why the
method should not be extended by the use of other
kernels.
As already stated, the aim of the present monograph
is to outline the procedure to be followed in using an
integral transform in the solution of boundary-value
problems. It is proposed also to show that a similar
technique can be employed whatever the kernel or range
of integration of the transform. Once the technique has
been mastered, this method of solution is really more
direct and straightforward than the classical method,
which often demands great ingenuity in assuming at the
outset the correct form for the solution. The technique
can be reduced almost to a “ drill”.
The point of view adopted is one which may perhaps
be regarded as adequate for most investigations in mathe-
matical physics. In applying an integral transform in this
type of work, assumptions as to the commutability of
certain limiting operations have to be made and often the
derivation of the solution is not rigorous. Strictly speak-
ing, it should be verified that the solution obtained by a
purely formal procedure does in fact satisfy the differ-
ential equation and its boundary conditions. It is usually
possible to do this, although in some cases the verification
process is somewhat laborious.
I am greatly indebted to many friends who read theINTRODUCTICN ix
manuscript, to the Quarterly Journal of Mechanics and
Applied Mathematics and the Delegates of the Clarendon
Press for permission to include Figures 3-7.
Cole ae
MILITARY COLLEGE OF SCIENCE,
SHRIVENHAM.
October, 1949.CHAPTER I
INTEGRAL TRANSFORMS AND THEIR
INVERSION FORMULAE
1.1. We define the integral transform /(p) of a function
f(x) by the integral equation
He)={'f0K(e, ») ds, pees (ot)
K(p, «) being a known function of p and x, called the
Rernel of the transform. When the limits a, b are both
finite we shall speak of f(p) as the finite transform of f(x).
Such transforms and their applications are discussed in
Chapter VI. At present we shall consider only transforms
in which a is zero and b infinite except for the case of the
transform given in (1.4) below where both limits are
infinite.
In the application of integral transforms to the solution
of boundary-value problems, use has so far been made of
five different kernels. The general method could doubt-
less be extended by the use of other kernels. The five
transforms considered here are :—
Laplace transform.
T(r) )-|" f(xjer2 dx. . . (1.2)
Fourier sine and cosine transforms.
fief) 22 sas... (23)
Complex Fourier transform.
fie)-[" flaerede. . . . (1.4)
-—~o
I2 TRANSFORMS IN MATHEMATICAL PHYSICS
Hankel transform.
He)={" fodstulps) ds, . (25)
where J,,(px) is the Bessel function of the first kind of
order n.
Mellin transform.
Hey={"foresds. 5 (18)
As will be seen later, the effect of applying an integral
transform to a partial differential equation is to exclude
temporarily a chosen independent variable and to leave
for solution a partial differential equation in one less
variable. The solution of this equation will be a function
of p and the remaining variables. When this solution has
been obtained, it has to be “inverted” to recover the
“Jost” variable: thus if x is the variable eliminated and
F(p) is one of the transforms given above, we first obtain
auxiliary equations giving f in terms of p and the remain-
ing independent variables, solve for f and then invert to
obtain f(x).
The inversion process means, in effect, the solution of
one of the integral equations (1.2), .. . , (1.6), f(p) being
supposed known and f(x) to be found. Such solutions
are known and can be obtained formally from Fourier’s
integral theorem which is given in the next paragraph.
1.2. FOURIER’S INTEGRAL FORMULA
As a preliminary to obtaining the inversion formulae,
we shall require what is often known as Fourier’s integral
formula. A formal derivation is given below.
Suppose a function f(x), of period 27, is given by the
Fourier series
oO
Hla)=hagt lan cos (nx/A)+b, sin (nx/A)].
n=1TRANSFORMS AND INVERSION FORMULAE 3
The coefficients a, an, 6, are obtained by multiplying
successively by unity, cos (”x/A), sin (nx/A) and integrat-
ing with respect to x between —zA, 7A. Since the circular
functions are orthogonal, that is,
na ore
(ee
— na 008
7A sin nx Cos mx
cos A sin 4 4*-%
_ cos Asi
7A sin nx sin mx | =0, m+n,
> > dx
cos A cos 2 "=A, m=n,
o0 and to zero
for x0 and zero for xo0, the
integral taken round a semi-circle, centre the origin and
lying below the real axis in the p-plane, tends to zero as
the radius of the circle tends to infinity. The integral can
wo e—izp
-«TRANSFORMS AND INVERSION FORMULAE 9
therefore be replaced by —277 times the residue at the only
pole (at p=—7) within this contour, the negative sign being
affixed since the direction round the contour is clockwise.
Thus
f(x)=-2d e~ 9 /[(d/dp)(14+p*)]p= i
=e—*,
Similarly, for x0
and to zero for x0,
=0, ¥<0,
. (1.16)10 TRANSFORMS IN MATHEMATICAL PHYSICS
giving
o
[i felipe dunanf(s), #>0,
—0«
=0, Xx0,
—0
=27f(x), x
[Oxford Graduate Texts] Oliver Davis Johns - Analytical Mechanics for Relativity and Quantum Mechanics, Second Edition [2nd Ed] (Instructor Solution Manual, Solutions) (2011, Oxford University Press, OUP, O.U.P) - Libgen.