Professional Documents
Culture Documents
Shivamoggi
Integral transform methods provide effective ways to solve a variety of problems arising in the
engineering, optical, and physical sciences. This concise, easy-to-follow reference text
introduces the use of integral transforms, with a detailed discussion of the widely applicable
Laplace and Fourier transforms.
The text is suitable as a self-study for practicing engineers and applied mathematicians, as well
as a textbook for students in graduate-level courses in optics, engineering sciences, physics,
and mathematics. The math is straightforward. In most sections, applications relevant to
engineers and applied scientists are used in place of formal proofs. Numerous examples,
exercise sets, illustrations, and tables of transforms enhance the books usefulness as a
teaching tool and reference.
Integral Transforms
for Engineers
Contents: Special functions. Fourier integrals and Fourier transforms. Applications involving
Fourier transforms. The Laplace transformation. Applications involving Laplace transforms.
The Mellin transform. The Hankel transform. Finite transforms. Discrete transforms.
Bibliography. Appendix A: Review of complex variables. Appendix B: Table of Fourier
transforms. Appendix C: Table of Laplace transforms. Index.
P.O. Box 10
Bellingham, WA 98227-0010
ISBN-10: 0819432326
ISBN-13: 9780819432322
SPIE Vol. No.: PM178
Larry C. Andrews
Bhimsen K. Shivamoggi
Larry C. Andrews and Bhimsen K. Shivamoggi are professors of mathematics at the University
of Central Florida. Andrews is also a member of the Department of Electrical and Computer
Engineering and associate member of the Center for Research and Education in Optics and
Lasers (CREOL) and the Florida Space Institute. Shivamoggi is also a member of the
Department of Physics at U.C.F.
Larry C. Andrews
Bhimsen K. Shivamoggi
Integral Transforms
for Engineers
Larry C. Andrews
Bhimsen K. Shivamoggi
University of Central Florida
=-==
SPIE
===
=~
PTICAL ENGINEERING
PRESS
Published by
SPIE-The Intemational Society for Optical Engineering
P.O. Box 10
Bellingham, Washington 98227-0010
Phone: 360/676-3290
Fax: 360/647-1445
E-mail: spie@spie.org
WWW: http://www.spie.org/
Contents
Preface to the 1999 Printing
Preface
Introduction
vii
lX
SPECIAL FUNCTIONS
1.1
1.2
1.3
1.4
1.5
Introduction
The Gamma Function
The Error Function and Related Functions
Bessel Functions
U seful Engineering Functions
6
7
16
21
29
37
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
Introduction
Fourier Integral Representations
Proof of the Fourier Integral Theorem
Fourier Transform Pairs
Properties of the Fourier Transform
Transforms of More Complicated Functions
The Convolution Integrais of Fourier
Transforms Involving Generalized Functions
Hilbert Transforms
Additional Topics
37
38
47
49
58
65
78
85
91
97
102
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
Introduction
Boundary Value Problems
Heat Conduction in Solids
Mechanical Vibrations
Potential Theory
Hydrodynamics
Elasticity in Two Dimensions
Probability and Statistics
102
103
113
125
131
141
151
156
162
4.1
4.2
Introduction
The Transforms of Some Typical Functions
162
164
v
vi
Contents
4.3
4.4
4.5
4.6
4.7
170
182
190
200
210
5
5.1
5.2
5.3
5.4
5.5
218
218
218
221
229
238
245
6.1
6.2
6.3
6.4
6.5
Introduction
Evaluation of Mellin Transforms
Complex Variable Methods
Applications
Table of Mellin Transforms
245
246
254
262
273
7
7.1
7.2
7.3
7.4
274
Introduction
Evaluation of Hankel Transforms
Applications
Table of Hankel Transforms
274
276
285
290
FINITE TRANSFORMS
291
8.1
8.2
8.3
8.4
Introduction
Finite Fourier Transforms
Sturm-Liouville Transforms
Finite Hankel Transform
291
291
298
303
9
9.1
9.2
9.3
9.4
9.5
DISCRETE TRANSFORMS
310
310
311
321
330
333
Introduction
Discrete Fourier Transform
The Z Transform
Difference Equations
Table of Z Transforms
BffiLIOGRAPHY
335
337
340
344
INDEX
349
Preface
Preface
Introduction
Introduction
J~= K(s,t)f(t) dt =
F(s)
(0.1)
K(s,t)
0,
<o
= { e-r, 1 ;;::: 0
(0.2)
K(s,t)
= -=em
(0.4)
y27T
J""
y27T -=
eist f(t)
dt
F(s)
(0.5)
which, when tis restricted to the positive realline, leads to the Fourier
sine and Fourier cosine transforms
principal 1'{1/ue
(0.6)
of the integral.
=!i~ rRj(x)dx.
t Other definitions of K(s,t) for Fourier transforms involve the choices e'", e-'-'',
(I/2'1T)e'", among others.
lntroduction
and
(0.7)
The Laplace and Fourier transforms are by far the most prominent
in applications. Many other transforms have been developed, but most
have limited applicability. ln addition to the Laplace and Fourier transforms,
the next most useful transforms are perhaps the Hankel transform of
order v
(0.8)
where lv(x) is the Bessel function of the first kind (see Sec. 1.4), and
the Mellin transform
(0.9)
The Hankel transform arises naturally in solving boundary value problems
formulated in cylindrical coordinates while the Mellin transform is useful
in the solution of certain potential problems formulated in wedge-shaped
regions.
The integral transforms mentioned thus far are applicable to problems
involving either semnfinite or infinite domains. However, in applying
the method of integral transforms to problems fonnulated on finite domains
it is necessary to introduce finite intervals on the transform integral.
Transfonns of this nature are called finite integral transforms.
A basic problem in the use of integral transforms is to determine the
function f(t) when its transform F(s) is known. We refer to this as the
inverse problem. In many cases the solution of the inverse problem is
another integral transform relation of the type
fon(s,t)F(s) ds
= f(t)
(0.10)
V127T
Joo e-'s'F(s)
. ds = f(t)
(0.11)
-oo
which is very much like the transform itself in Eq. (0.5). This means
that the problems of evaluating transfonns or inverse transforms are
lntroduction
essentially the sarne for Fourier transforms. This is not necessarily the
case for other transforms like the Laplace transform, however, where
the inversion formula is quite distinct from that of the transform integral.
Also, in the case of finite transforms, the inverse transform is in the
form of an infinite series.
The basic aim of the transform method is to transform a given problem
into one that is easier to solve. ln the case of an ordinary differential
equation with constant coefficients, the transformed problem is algebraic.
The effect of applying an integral transform to a partial differential equation
is to reduce it to a partial differential equation in one less variable. The
solution of the transformed problem in either case will be a function of
the transformed variable and any remaining independent variables. lnversion of this solution produces the solution of the original problem.
The exponential Fourier transform does not incorporate any boundary
conditions in transforming the derivatives. Thus, it is best suited for
solving differential equations on infinite domains where the boundary
conditions usually only require bounded solutions. On the other hand,
the Fourier cosine and sine transforms are well suited for solving certain
problems on semiinfinite domains where the governing differential equation
involves only even-order derivatives. We will see that the Fourier transform
lends itself nicely to solving boundary-value problems associated with
the following partial differential equations:
(a) the heat equation:
V2u
= a- 2u, -
q(x,y,z,t)
(0.12)
= c- 2u,, - q(x,y,z,t)
(0.13)
= u(x)
J~=k(x, t)u(t) dt
(0.15)
lntroduction
f~ooeitx p(x) dx
(0.16)
J: U(T)k(t -
T) dT == f(t),
>o
(0.17)
1
Special Functions
1.1
lntroduction
* For a more thorough treatment of special functions, see L. C. Andrews, Special Functions of
Mathematicsfor Engineers (SPIE Press, Bellingham, Wash.; Oxford University Press, Oxford, 1998)
6
1.2
One of the simplest but very important special functions is the gamma
function. Although it has less direct application than some of the other
special functions, knowledge of the properties of this function is a prerequisite for the study of Bessel functions and others which do have
direct .pplication.
Historically, the gamma function was discovered by L. Euler (17071783) in 1729 who was concerned with the problem of interpolating
between the numbers
n!
L"" e- tn dt,
1
n == O, 1,2, ...
with nonintegral values of n. His studies eventually led him to the gamma
function relation
Re(z) >O*
(1.1)
J:
e- f dt
= -e- 1 f
~~ + z L"" e-
tz-l
dt
from which we deduce the very simple but important recurrence formula
f(z
The value
to the result
z=
+ 1) = zf(z)
(1.2)
f(l)
= L"' e-
dt
=1
x + iy. Similarly,
I)
= n!,
n = 0,1,2,...
(1.3)
Thus, we see that the gamma function is Euler' s extension of the factorial
function to nonintegral values of n. ln fact, the gamma function is an
extension of the factorial function to all complex numbers with a positive
real part. ln the next section we will extend this domain even further.
1.2.1
= f(z + 1)/z,
z:fO
= f(z +
2)/(z + 1),
(1.4)
z :f O, and
z by z + 1
z :f -1
f(z) = f(z
f(z + k)
z(z + l)(z + 2)(z + k - 1)'
z :f
+ 1 (1.6)
= 0,1,2, ...
(1.7)
f(x)
\)
II
I
I
I
I
I
I
I
I
I
I
-41
I
-3
I
I
-21
I
I
I
I
I
I
f(z)
= log(l I u)
= Jori
in ( 1.1) yields *
gamma function.
Solution:
(1.8)
x3 , we have
(1.9)
10
f(x)f(y) = 2
4
J:
e-u2U2x-l
du
u = r cosO,
r sinO
leads to
which reduces to
f(x)f(y) = 2f(x + y)
f"/2
Jo
cos2x-o sin2y-o dO
(1.10)
f"/2
Jo
cos
f"/2
Jo
x>O,y>O
(1.11)
Jo
But, f(3)
2 and
f(3)f(5/2)
2f(1 1/ 2)
rC 1) = r(1 + ~) = ~ ~ ~rG)
2
11
Jo
2 2 2
= 9 7 5 = 3I5
sin 6 cos 6 d6
=2
~/2
tan2' - 16 dO
(1.12)
f(z)f(l - z)
(1.13)
fW =I+~
which has a simple pole at ~ = - I and a branch point at ~ = O, around
the contour shown in Fig. 1.2 and then let p ~ O and R ~ ao. If we
write ~ = u along the upper boundary of the cut along the positive real
axis, then we must write ~ = ue 2~i along the lower boundary of this cut.
lm(~)
ReW
~-plane
12
tfWd'
f fWd'
JcR
=O
uz-1
fo (ue21Tiy-l
- - du +
du
ol+u
oo1+u
oo
2m"Res{ -1}
or
z-1
_u_ du = 2TriRes{ -1}
o 1+u
oo
(1 _ e2,.;z)
(1.15)
However, ( -l)z-t
e"i<z-tJ
oo
uz-1
7T
--du=--
o 1
+U
SD7TZ '
(1.16)
z. By setting z
(1.17)
= 1/2, we
= 7T
y;.
(1.18)
Jo
cot
13
112
x dx.
f(l/4)f(3/4)
2f(l)
1
1T
2 sin(Tr/4)
and hence
There are many other identities involving the gamma function which
are too numerous to mention here. For reference purposes, a short list
of basic identities follows.
L=e-rtz-J dt,
(Gl): f(z)
(G2): f(z)
Re(z) >O
12
Re(z) >O
= 0,1,2, ...
f(z + k)
(G6): f(z)
= (
(G8): f(n
z z + 1)(z + 2) (z + k
(G7): y;. f(2z) = 22z-J f(z)f(z + !)
(2n)!
1T
(G9): f(z)f(l - z) = - . - ,
- 1)
k = 1,2,3, ...
n = 0,1,2, ...
z nonintegral
SID 1TZ
(GlO): f(n
14
EXERCISES 1.2
ln Probs. 1-6, give numerical values for the expressions (use the result
f(l/2) =
where necessary).
y;
1. f(6)
2. f(3/2)
3. f(7 /2)
4. f( -1/2)
5. f( -9/2)
6. f(8/3)/f(2/3)
7. f(a + n)
a(a
8. f(n - a)/f(- a)
1,2,3, ...
9. f(a)/f(a - n)
+ l)(a + 2) .. (a + n - l)f(a),
= ( -1ta(a
(a - l)(a - 2) .. (a - n),
1) ..
) =
k!(n
n~ k)!'
(c)
!/
(a) =
(d)
(~a)= (-1i(a
(b) (-
2
)
= 0,1,2, ... ,
= ( ~~(~~~)!,
f(a
1)
Z-
1,2,3, ...
k = 1,2,3, ...
= 0,1,2, ... ,n
n = 0,1,2, ...
k!f(a- k + 1)'
1,2,3, ...
0)
=I=
1)(a - 2) .. (a - n + 1),
Show that
= 0,1,2, ...
k=
1
),
0,1,2, ...
11. f(x)
12. f(x) =
13. f(x)
p"
L= e-pttx-I dt,
J~=exp(xt - e~dt,
Hint: Let u
> O, p > O
x >O
x >O, b > 1
t log b.
ln Probs. 14 and 15, use properties of the gamma function to obtain the
result.
14.
oo
2
2
a e ax-x
dx
15
1.,.;-z
= 2 v 7Tea
Hint: 2ax - x2
-(x - a)
{.,.;z
16.
Jo
1f/2
18.
1
0
{.,.;z
sin x dx
5
Jo
17.
~dx
19
sin xcos x dx
100 ~
o 1
+ x4
Hint: Let
i1t/2
cos 2n+i(Jd(J
22n( ')2
n.
(2n
I)!
x2 = tan
8.
= 0,1,2, ...
= Li tx-i(l
X> O, Y >O
- t)Y-i dt,
Show that
(a) B(x,y)
f(x)f(y)/f(x
+ y)
L= (1 :x~;x+y du
Hint: Let t
u/(1 + u).
= zi-2x l.,.;z
o
. 2x-l,~,. d..l.
sm
=!
'+'
'+'
sin 2x.
y; f(2x) =
z2x-i
f(x)f(x
+ 1/2)
16
I/2)
(2n)!
= 22nn.I y'fl',
n = O,I,2, ...
:f -I,-2,-3, ... )
f(2n + 2v + 1)
f(n +v+ I)
1.3
y;
1/2)
= y:;;: Jo e-
12
(l.I9)
dt
2 Jor ~o ----;;!t
<-Ir
y;
00
erf(z) =
2
n
dt
00
= -= L
(-1tin+l
1)
(1.20)
= -
erf(z)
(1.21)
= erf(z)
(1.22)
(1.23)
17
1r
2_
v7T
f'"' eJo
12
dt =
f(l~)
= 1
(1.24)
v7T
(1.25)
2 f'"' _,z d
erfc(z) = v; Jo e t -
t,
=1-
(1.26)
erf(z).
Hence, ali properties of erfc(z) are easily derived from those of erf(z).
1.3.2 Fresnel Integrais
Closely associated with the error function are the Fresnel integrais
C(x)
2
COS(7Tt /2) dt
(1.27)
erf(x)
-2
18
and
S(x)
J: sin(1rf /2) dt
(1.28)
(1.29)
2
COS(1TX /2),
S'(x)
= sin(1Tx2/2)
(1.30)
and thus we deduce that both C(x) and S(x) are oscillatory. Namely,
C(x) has extrema at the points where
= 2n + I (n = O,I,2, ... ), and
S(x) has extrema where x2 = 2n (n = I ,2,3, ... ). The largest maxima
occur first andare found to be C(l) = 0.77989 ... and S('\1'2) = 0.71397 ....
For x ~ oo, we can use the integral formulas (see Prob. I2 in Exer. 1.3)
f'
~ J~
(1.31)
The graphs of C(x) and S(x) for positive real x are shown in Fig. I.4.
C(x)
0.5
19
EXERCISES 1.3
1. Show that
(a)
ra e- ~
(b)
Ja
1 dt =
rb
-(2
dt
y; erf(a)
v; [erf(b) -
erf(a)] =
v; [erfc(a) -
erfc(b)]
2. Show that
- erf(z) = -=e-z
dz
v1T
J erf(z)dz = z erf(z) + y; e
+C
e-px
1
erf(x) dx = - eP 214 erfc(p/2)
p
L=
y;
= (x/2 +
p) 2
e-px-x2f4
Hint: Write x2 /4 + px
variable u = x/2 + p.
dx
e- 1 dt
e -z2
=-
2z
ep2
erfc(p)
i"" -
e -12
dt
2 z t2
- -
20
erfc(z)-
-z
2
[
v'e-
7r
2:= (- l t 1 X
I +
X X
(2z
n=l
I)]
(2n-
( ) =-=-e
px
-(x-mP/2rr2
y2-rrCT
where m is the mean value of X and CT2 the variance. The probability
that X $: y is defined by
P(X $: y)
f=
p(x) dx
$;
y)
H1
$;
err(Y; 2:) J
y)
in the limit
y -
oo?
2:
O, b >O
7r
2b2
=b
ea erfc(ab).
-C(x)
= ~ ( -lt{-rr/2)2n x4n+l
n=O
(2n)!(4n
1)
(b) S(x)
=
2:
n=O
=
(_
(2n
I )n( /2)2n + I
x4n+3
7r
1)!(4n
3)
1.4/Bessel Functions
21
oo
- a2t2
y;
t=-
2a
Jor= cos t
dt
= f"" sin r
Jo
dt
= ! J~
2
13. Using the definition of the error function (1.19), show that
(a) erf(Vfx) = (l + i) [C(xV2/7T) - iS(xV2/1T)]
(b) erf(V -ix) = (l - i) [C(xy'2/";) + iS(xV2/1T)]
(c) erf(Vfx) +erf(y' -ix) = 2 [C(xV2/7T) + S(xy'2/";)]
fv(Z)
-l)k (z/2ik+v
1)
= k~O k! f(k + v +
(1.32)
where the parameter v denotes the order of the given Bessel function.
When v = n (n = 0,1,2, ... ), Eq. (1.32) defines the Bessel function of
integer order
"" (- 1)k(z/2ik+n
Jn(Z)
= t:o
k! (k
+ n)! '
= 0,1,2, ...
(1.33)
J.o(Z)
f:o
(-lt(z/2)2k
(k!)2
(1.34)
The graphs of Jn(x), n = 0,1,2, are shown in Fig. 1.5, where xis real.
The parameter v in (1.32) may also take on negative values. For
example, when v = - n (n = O, l ,2, ... ), we get
22
l_iz)
k=O
-l)k(z/ 2)zk-n
k' (k _ )'
oo
= k~n
-1i(z/2)2k-n
k!(k- n)!
where we have used the fact that 1/(k - n)! = O (k = 0,1, ... , n - 1)
by virtue of Eq. (1.7). Finally, the change of index k = m + n yields
oo
j_n(Z)
=L
-l)m+n(z/ 2)zm+n
)I
m.I(m + n.
m=O
= (- 1Y Jn(Z),
n = 0,1,2, ...
(1.35)
Jv(Z)
oo
-1)k(z/ 2)zk
1)
= ( 2 ~o k!f(k + V +
it can be shown that the series on the right converges in the whole zplane. Therefore, the function (2/ztlv(Z) is ao entire function of z. However,
this does not necessarily imply that lv(z) is entire. If v< Oand nonintegral,
then clearly lv(z) has ao infinite discontinuity at z = O, and hence, cannot
represent an entire function. But, if v = n, n = 0,1,2, ... , then it can
be shown that Jv(x) is entire- a result that depends upon the relation
(1.35).
The Bessel functions are named in honor of F. W. Bessel (17841846) who in 1824 carried out the first systematic study of the properties
of these functions and derived their goveming differential equation (see
Prob. 3 in Exer. 1.4). Nonetheless, Bessel functions were discovered
1.4/Bessel Functions
23
years earlier by Euler and others who were concerned with various
problems in mechanics, and the infinite series (1.34) was obtained by D.
Bernoulli in 1703- more than 120 years before Bessel's famous study
- in connection with bis investigation of the oscillatory behavior of a
hanging chain.
1.4.1
Basic Properties
~ [z"J.(z)] = z"lv-l(z)
(1.36)
and
d
dz [z-vlv(Z))
= - Z-vlv+ 1(z)
(1.37)
both of which follow from termwise differentiation of the series for z"J"(z)
and z-"lv(Z) (see Prob. 1 in Exer. 1.4). If we carry out the differentiation
in (1.36) and (1.37) and simplify the results, it follows that
(1.38)
and
J~(z)- ~lv(Z) =
The substitution of v
-lv+I(Z)
(1.39)
-JI(Z)
(1.40)
J z"lv-l(z) dz = z"lv(Z) + C
(1.41)
and
(1.42)
where C denotes a constant of integration. As a general rule, any integral
of the forro
m
+ n >O
where m and n are integers, can be evaluated with the use of (1.41) and
(1.42), coupled with standard integration techniques such as integration
24
J J (z)dz when
0
m + n is even.
Examp/e 1.4: Reduce
Jlliz) dz
= -
z21t(Z) + 3
Jzlt(Z) dz
3zJo(z)
Jlo(z) dz
,. ( _
(z/Z)2k +"
t:o k! f(k0 +
(Jl): J.(z)
(J2): 1 0(0)
= 1; J.(O) = O,
11
11
(J4):
~ [z"J.(z)] = z"l.-t(Z)
(J5):
~ [z-"J.(z)] =
-z-"lv+t(Z)
-lv+t(Z)
= 2J~(z)
(J7):J~(z)-
(J9): lv-t(Z)
-Jv(Z)
>O
n = O, 1,2, ...
1)
2v
+ J,+ t(Z) = -
fv(Z)
1.4/Bessel Functions
(JlO):
(Jll):
25
= - Z-vJ.,(z) + C
Z-vJ.,+ J(Z) dz
(J12): Jo(Z)
1 f21r
= 21T Jo eizcos8d(J
(z/2)"
(Jl3):J.,(z)-f(v + l)'
(114): J.,(z)-
J~z
cos[ z - (v +
1/2)~],
lzl-
ao,
larg(z)l < 1T
J.,(ty)
(yj2)2k+v
= l k~O k!f(k +
1)
(1.43)
Except for the multiplicative factor i", the right-hand side of (1.43) defines
a real function, which is called the modified Bessel function of the first
kind and denoted by the symbol lv(Y ). Thus,
/.,(y)
= ;-"J.,(iy)
(1.44)
= ~o
(z/2)2k+v
k!f(k + v + 1)
(1.45)
Comparing this series representation with Eq. (1.-32) for Jv(z), it would
appear that /.,(z) and J.,(z) have many properties in common. lndeed, the
modified Bessel functions satisfy relations analogous to ali those for the
standard Bessel functions. ln particular, the modified Bessel functions
satisfy properties similar to those for J.,(z) given by (1.36)-(1.42) (see
26
3
Figure 1.7 Graph of Kn(x), n = 0,1,2
1.4/Bessel Functions
27
= !!_I -v(Z!
- /,(z)
sm v7r
(1.46)
This is called the modified Bessel function of the second kind and its
graph for certain integer values of v is given in Fig. 1. 7. Some properties
of this function are taken up in the exercises.
EXERCISES 1.4
1. Using the series representation (1.32), show that
(a)
d
(b) dz [z-vlv(z)]
= -
z-vlv+ ,(z)
= J,_,(z)
(a) 2J~(z)
- J,+t(z)
2v
+ J,+I(Z)
zJ;(z)
(z
v2)1v(Z)
7T'Z
n= -oo
show that
(a) cos(x sin O)
= J0(x) + 2
2: 1 n(x)cos(2n0)
2
n=l
L1 n_ (x)sin[(2n 2
n=l
(c) cos x
= 10(x) + 2
2: (-lt1 ix)
2
n=l
(d) sin x
=2
2: (-1tlzn-l(x)
n=l
1)0]
=O
7T'Z
cos z
28
= ~ { cos xt dt
7TJoyl1=7
1
Jo(x)
8.
9.
10.
I
I
I
x J 0(x) dx
= x 2J 1(x) + xJ0(x)
x 3J 0(x) dx
= (x3
x- J 2(x) dx
J 0(x) dx
4x)J1(x) + 2x 2J 0 (x) + C
1
= - 2x2 J,(x) - 3J,(x)
+ 1x Jo(x) + 3li Jo(x) dx + C
3
12. By expressing J0 (bx) in its series representation, use termwise integration to show that
+ b2 ,
a> O, b >O
(a)
Jo(""cos(cx)J0(bx) dx = { 1/ybO, -
(b)
b>c
b <c.
bb:
~.
14. Integrate both sides of Prob. 13(a) with respect to c to deduce that
{"" sin x
Jo
--;-Jo(bx)dx=
{ 71'/2,
sin-'(1/b),
O< b < 1
b>l
~ [zvlv(Z)]
= zvlv-!(z)
d
(b) dz [z-vlv(Z)]
= Z-vlv+!(Z)
= Iiz),
29
= 0,1,2, ...
(b) /~(z)
(c) l~(z)
= lv+t(Z) + -lv(Z)
z
=i
[lv-t(Z)
+ lv+ t(Z)]
2
=-
vz
lv(Z)
1TZ
(z)
112
1TZ
cosh z
d
(b) dz [z-"K.(z)]
1.5
= -z-"K.+ 1(z)
30
h(t - a)
helpful to introduce the unit step function, also widely known as the
Heaviside unit function in honor of its discoverer. * We denote this function
by the symbol h(t - a) and define it by (see Fig. 1.8)
h(t - a)
= {~:
:~ :
(1.47)
is clearly zero for t < 1 and assumes the graph of the cosine function
for t > 1 as shown in Fig. 1.9.
A related function is the rectang/e function defined by
f(t)
= {o1',
a<t<b
otherwise
(1.48)
h(t - 1) cos21Tt
31
= h(t
(1.49)
- a) - h(t - b).
= { f2(t),
J;(t),
t< a
a<t<b
t>b
(1.50)
+ [j3(t) -
f2(t)]h(t - b)
(1.51)
J
oo
I =
-= da{t) dt =
fa+e
a-e da{t) dt
(e> 0)
(1.52)
32
da(t)
2s
a-s
this constant value in such a way that the total impulse given by (1.52)
is unity. Hence, we write
da(t)
= (1/2s)[h(t
- a + s) - h(t - a - s)]
(1.53)
Now let us idealize the function da(t) by requiring it to act over shorter
and shorter intervals of time by allowing s - O. Although the interval
about t = a is shrinking to zero, we still want I = 1; i. e.,
lim I
e~o
= lim J=
e---~>0
da(t) dt
(1.54)
-oo
J~oo (t -
a) dt
= O,
=f
(1.55)
=1
f~oo (t -
a)f(t) dt
= ~~ J~oo da(t)f(t) dt
1
e"'"'""+O 2B
= lim -
fa+e
f(t) dt
a-e
Recalling that
J: f(t) dt = f()(b -
a),
a<<b
(1.56)
33
which is the mean value theorem of the integral calculus, we find that
J-= ll(t -
a)f(t) dt
= ~~ 2e J(g) 2e
J~= ll(t -
a)f(t) dt
= f(a).
(1.57)
Obviously the "function" 8(t - a), also known as the Dirac delta
function, * is not a function in the usual sense of the word. It has significance
only as part of an integrand. ln dealing with this function, therefore, it
is best to avoid the idea of assigning "functional values" and instead
refer to its integral property (1.57), even though it has no meaning as
an ordinary integral. Following more rigorous lines, the impulse function
can be defined as a limit of an infinite sequence of well-behaved functions
(see Probs. 11 and 12 in Exer. 1.5).
There are certain operational properties of the impulse function that
prove useful in practice. Our derivations of such properties, however,
will be based strictly upon formal manipulations of the symbols, i.e.,
they will not be rigorous. To begin we make the observation
J~= ll(t -
a)f(t) dt
= J~= ll(t)f(t + a) dt
(1.58)
I~= [f(t)ll(t
- a)]g(t) dt ==
I~= 8(t -
a)[f(t)g(t)] dt
== f(a)g(a)
Since
g(a)
= J~= ll(t
- a)g(t) dt
we see that
J~"" [f(t)(t -
a)]g(t) dt
= I~= [f(a)(t
- a)]g(t) dt
= f(a)(t
- a)
(1.59)
* Named after Paul A. M. Dirac (1902-1984), who was awarded the Nobel prize (with
E. Schrdinger) in 1933 for his work in quantum mechanics.
34
J~oo h'(t -
= h(t
a)f(t) dt
- a)f(t)
= f(=)
a)f'(t) dt
L"' f'(t) dt
= f(a)
where we are assuming that f(t) is both continuous and bounded. By
comparison of this result with Eq. (1.57), we deduce that
J~oo h'(t -
a)f(t) dt
J~oo i3(t -
a)f(t) dt
(1.60)
h'(t - a) = 8(t - a)
EXERCISES 1.5
1. Show that
h(t)
=i
[1
+ sgn(t)]
t<O
= { - 1,1
>o
2. Show that
rb (t -
Ja
3. Show that
a< t0 < b
otherwise
to)f(t) dt = {f(to).
O,
oo
-=
1
i3(at)f(t) dt
= jaj /(0)
35
= 28(t)
J~oo 8'(t)f(t) dt =
(b)
-f'(O)
= 1,2,3, ...
= 8(t)f'(t) + '(t)f(t)
9. Show that
l>'(t)f(t)
= f(0)8'(t)
- f'(0)8(t)
g(t)
= f(t)
ak h(t - tk)
k=f
where f(t) is a piecewise continuous function having jump discontinuities of magnitude a], a2' ... ' an at the points t], t2' ... ' tn.
Assuming that f'(t) is defined everywhere except at these
discontinuities,
(a) show that g(t) is everywhere continuous and that g'(t) = f'(t)
except for a finite number of points.
(b) Deduce that the generalized derivative of the piecewise differentiable functionf(t) with finite jumps is the ordinary derivative,
where it exists, plus the sum of impulse functions at the discontinuities multiplied by the magnitude of the jumps.
11. Consider the sequence of rectangle functions defined by (n
= 1,2,3, ... )
36
ltl <
ltl >
1/n
1/n
(a) Show that for each n the area enclosed by the rectangle is unity,
and deduce that
!~ f~= t/Jit) dt =
O and
2
Fourier Integrais and
Fourier Transforms
2.1
Introduction
The concept of an infinite series dates back as far as the ancient Greeks
such as Archimedes (287-212 B.c.), who summed a geometric series in
order to compute the area under a parabolic are. ln the eighteenth century,
power series expansions for functions like eX, sin x, and arctan x were
first published by the Scottish mathematician C. Maclaurin (1698-1746),
and British mathematician B. Taylor (1685-1731) generalized this work
by providing power series expansions about some point other than
X= 0.
By the middle of the eighteenth century it became important to study
the possibility of representing a given function by infinite series other
than power series. D. Bernoulli (1700-1783) showed that the mathematical
conditions imposed by physical considerations in solving the vibratingstring problem were formally satisfied by functions represented as infinite
series involving sinusoidal functions. ln the early 1800s, the French
physicist J. Fourier* carne across similar representations and announced
*Jean Baptiste Joseph Fourier (1768-1830) is known mainly for his work on the
representation of functions by trigonometric series in his studies on the theory of heat
conduction. His basic papers, presented to the Academy of Sciences in Paris in 1807 and
1811, were criticized by the referees for a Jack of rigor and consequently were not published
then. However, when publishing the classic Thorie analytique de la Chaleur in 1822, he
also incorporated his earlier work almost without change.
37
38
2.2
wbere
n = 0,1,2, ...
Power series sucb as tbis are useful for numerical calculations in addition
to various otber uses. lf tbe function f is periodic witb period 2p, it may
bave a Fourier series representation*
f(x)
1 + LJ
~ ( ancos -n1rx + bnsm. n1rx)
= 2a
0
P
n=l
(2.2)
wbere
1 JP
n1rt
an = f(t)cos- dt,
p -p
p
= 0,1,2, ...
(2.3)
= 1,2,3, ...
(2.4)
and
1 JP
. n1rt
bn = f(t)sm- dt,
p
-p
39
p -p
f(t)dt
n=!
[.!.p
r
-p
or
f(x)
= 2p
JP
-pf(t)dt
x)
dt
(2.5)
f~ if(t)ldt <
00
(2.6)
so that
1
lim 2p
p-=
Jp
f(t)dt
=O
(2.7)
-p
= rr/ p
40
f(x)
1 f"'/!!>s
lim f(t)
lls-->0 1T
-w/lls
2:
00
cos[nas(t - x)]as dt
(2.8)
n= I
L"" cos[s(t -
x)]ds
in the limit as as ~ O. While this does not mean that the limit of the
series in (2.8) is defined to be the above, we may take, under appropriate
conditions on f 1 that (2.8) tends to the integral form
1
f(x) = ;
(2.9)
x)]ds dt
(2.10)
- x)]dt ds
The purely formal procedure we just went through (since the passage
to the limit cannot be rigorously justified) has led us to an important
result known as Fourier's integral theorem.* We will state the theorem
here but not present its rather lengthy proof until Sec. 2.3.
Theorem 2.1 (Fourier Integral Theorem). If f and f' are piecewise continuous functions on every finite interval, and if
00
then
f(x)
= ;1 Joroo I""_""f(t)cos[s(t
- x)]dt ds
* For a rigorous discussion and a precise statement of the conditions under which
(2.10) holds, see E. C. Titchmarsh, Theory of Fourier Integrais, Oxford: Clarendon Press,
1937.
t Right-hand and left-hand limits are defined, respectively, by f(x+) = lim f(x + e)
e-O+
and f(x-)
e-O+
= f(x+) = f(x).
41
The conditions listed in Theor. 2.1 are only sufficient conditions, not
necessary conditions. That is, there exist functions f that have valid
integral representations but which do not satisfy the conditions of this
theorem. Moreover, the conditions stated in Theor. 2.1 are not the most
general set of sufficient conditions that have been established over the
years. Nonetheless, these conditions are broad enough to cover most of
the functions commonly occurring in practice.
To emphasize the analogy between Fourier series and the Fourier
integral theorem, we rewrite (2.10) in the form
f(x)
= -7T'
-oo
or equivalently,
f(x)
(2.11)
where
A(s)
(2.12)
B(s)
(2.13)
and
Example 2.1:
f(x)
-1
Figure 2.1 Graph of f(x) = h(l - lxi)
42
1
A(s) = 7T
J=
-=
1
h(l - lxi)cos sx dx = 7T
and
Jl
= -1
B(s)
7T
-1
J
1
2 sin
s
cos sx dx = 7TS
-1
sin sx dx = O
= ;2 Jor= (sin
- 8
s) cos sx ds
= 1=~
7T
r= sins s ds
Jo
r= sin s ds =!
Jo
(2.14)
; L(2
sin s
8
-)
1/2,
cos sx ds =
{1 2
O,
X=
-1
-l<x<l
X = 1
otherwise
(2.15)
S~(x) =
;J:
(2.16)
(si;s) cossxds
oo.
= sin(A + B) + sin(A -
B)
we have
* This, of course, is a standard integral result that can also be derived by the use of
complex variable theory. It is an important result that we will refer to on several occasions.
43
J"'
1"'( +x)
1"'(1-x)
(2.17)
- x)]}
7T
iz o
sin t dt
t
(2.18)
2.2.1
= 4,16,128.
A(s)
= -7T1 J""
f(x)cos sx dx
-oo
= -7T21""'
f(x)cos sx dx
O
(2.19)
and
J""
B(s) = ;1 -=f(x)sin sx dx = O
(2.20)
f(x)
= L"" A(s)cos sx ds
(2.21)
44
f"
B(s)sin sx ds
(2.22)
O and
B(s)
= -7T21""
f(x)sin sx dx
o
(2.23)
x<
Example 2.2: Find a Fourier cosine and Fourier sine integral representation of the function (see Fig. 2.3)
f(x)
= {cos x,
X> 7T/2
O,
=-
1r0
1r
2
1 -s
and, therefore,
f(x)
=;
= -2 i1T
12
cos x sm
sx dx
1r0
- 2sin_ 1rsj2)
= -2 (s---;;_..;.__
1r
s-1
45
= ; Jo
f(x)
EXERCISES 2.2
1. By using the result of Eq. (2.14), show directly that
(a) ("" sin s cos s ds = :!!.
Jo
s
4
a>0
2. If
x<O
x>O
= { 0e :..x ,
f(x)
f(x)
1r
Jo
i +1
{1O, -
xZ,
lxl <
lxi >
L""
C-
;osxy dx
Jo x2
=~
= .:!!.
4
ln Probs. 5-10, obtain the Fourier integral representation of the given
function.
-l<x<O
5. f(x) = e-lxl, -ao <X< ao
6. f(x) = { O<x<l
O,
otherwise
~:
46
lxl < 7T
lxl > 7T
1 f(x) = {sin x,
O,
9 f(x)
8. f(x)
O<x<7T
otherwise
= {cos x,
O,
= { si~'x,
O<x<7T
otherwise
e-kx
7T
Jo
X>
s +
(b) e-kx
X> 0
=; Loo (:: :
!)cos sx ds,
-x
2 ioo s sin sx
(b) e cos x = ds,
4
7T o s + 4
x >O
x>O
-2x
- e
roo
= ; Jo
s sin sx
(i
+ l)(i + 4) ds,
>O
ln Probs. 14-17, obtain the Fourier cosine and Fourier sine integral
representations of the given function.
o.'K
14. f(x)
={
15. f(x)
= { ~'- x,
16. f(x)
= {~:n x,
17. f(x)
= {o,'
>p
~~~< 1
~~:
< 7T
O<x<1
x> 1
(c) f(x)
= e-lxl
lxl < 1
lxl > 1
{!ir,
(b) J(x)
(d) f(x)
= sin x
X
<e>J<x>
= (si:xr
lxl < 1
lxl > 1
2.3
47
!~ J~=f(t)cos t dt = !~ J~=f(t)sin t dt =O
or equivalently,
Proof: We will present the proof only for the case when f is continuous
and has a bounded derivative f' on the real axis. A slight modification
of the proof is required for the case whenfhas some finite discontinuities.
Using integration by parts over the finite interval - p s t s p, we
get
P f(t)eit dt
J -p
= f(t~eit
l
lP -p
J- JP-p f'(t)eit dt
l
Clearly, f(t)eit is bounded on ali finite intervals for any . Thus, the
first term on the right-hand side vanishes in the limitas tends to infinity.
Also, because we assume that f' is bounded, it follows that
-p
48
. 1 Joo
sin t
hm f(x + t) - - dt
->OO 1T
-oo
f
= f(x)
. 1 JP sin-t
d Im
I" -1 JP sin tt
d
t=
Itm1T -p
t
->OO 1T
-p
t
-+00
_ 1 Joo sin-t dt
1T -oo
[
--
=1
this Iast resuit following from (2.14). Hence, to prove the Iemma we wish
to show that
. 1 JP [f(x + t) - f(x)] sm
. t dt
Itm-+OO
1T
-p
=0
+ t)
- f(x)
= f'(x)
2.3 .1
We are now prepared to prove our maio result, which is Theor. 2.1. We
will present the proof only for points of continuity of the function f,
leaving the proof for points of discontinuity to the exercises.
By changing the order of integration in the following iterated integral,
we obtain
:;;1rfoo
Jo -oo f(t)cos[s(t
- x)]dt ds
1Joo
If we now a11ow
result
-oo
x)]ds dt
f(t) sin (t - x) dt
-oo
= -1 Joo
1T
Jor cos[s(t f - X
f(x
sin t
/
t) - - dt
1 r=
; Jo
foo
-=f(t)cos[s(t - x)] dt ds
= f(x)
49
(2.24)
EXERCISES 2.3
1. Prove that
JP sin t
(a) hm
-+00
(b) lim
Jo
-+00
7T
- - dt = t
2
-p
sin t dt
t
= '!!.
2
2. Based on Lemma 2.1, show that if f and f' are piecewise continuous
on (O,p) and (- p,O), then
. 2 JP
sin t
(a) hmf(x + t) - - dt = f(x+)
-+00 7T
o
t
. 2
sin t
(b) hmf(x + t) - - dt = f(x-)
-+00 7T
-p
t
3. Prove Theor. 2.1 for points of finite discontinuity of f, i.e., prove that
Jo
; L=
2.4
f~=f(t)cos[s(t -
x)]dt ds
= ~[f(x+) + f(x-)]
(2.25)
Through the use of Euler' s formula, cos x = !( eix + e- ix), we can express
(2.25) in terms of complex exponential functions. That is,
f(x)
Joo-= f(t)cos[s(t
= ;1 1""
0
- x)]dt ds
= _1
= -1
27T
or
50
f(x) = - 1
27T
Joo
e -isx
Joo
-oo
eist f(t) dt ds
(2.26)
-oo
} Joo
~ ;;c_
V 27T
eist f(t) dt
(2.27)
e-ist F(s) ds
(2.28)
-oo
and
f(t)
= v'1 Joo
27T
-oo
= @i{f(t);s}
(2.29)
and f(t) as the inverse Fourier transform of F(s), which may be written
as
f(t) = g;- 1{F(s);t}
(2.30)
The location of the constant 1/27T in the definition of the transform pairs
is arbitrarily selected as long as (2.26) is satisfied. For reasons of symmetry
we have split the constant between the transform pairs, but in the literature
no universal agreement exists on the location of these constants. ln some
texts, the constant 1/27T is positioned in front of one of the transform
pairs with no constant in front of the other. There is also some variation
as to which integral represents the transform and which one represents
the inverse transform. ln practice, of course, these differences are of
little consequence but the user should be aware of them when consulting
different reference sources.
As an immediate consequence of (2.27), we observe that
} Joo
!F(s)! :s ~ r-;--
y27T
lf(t)l dt
(2.31)
-oo
(2.32)
11--=
* Unless stated otherwise, we will generally assume that both t and sare real variables.
51
ln Sec. 2.2.1 we found that when the function f is even, the Fourier
integral representation of f(x) reduces to
= E"' A(s)cos sx ds
f(x)
=-
'1T'
(2.33)
J~L'"'f(t)cosstdt = Fc(s),
s>O
(2.34)
.<f'c 1{Fc(s);t} =
>o
(2.35)
These results are interesting in that they imply the equivalence of the
operators .<1'c and .<f'c 1 ln other words, the cosine transform and its
inverse are exactly the sarne in functional form.
Similarly, whenfis an odd function its Fourier integral representation
becomes
= -2
f(x)
'1T'
(2.36)
s>O
(2.37)
>o
(2.38)
52
Hence, we see that the Fourier sine transform and its inverse are also
exactly the sarne in functional form.
lf the function f is neither even nor odd, but defined only for t > O,
then it may have both a cosine transform and a sine transform. Moreover,
the even and odd extensions of f will then have exponential Fourier
transforms. To see the relations between these various transforms, let
us construct the even extension of f by setting
-oo<t<oo
fe(t) = f(ltl),
(2.39)
1 Joo
= --=
fe(t)eist dt
-oo
O
= .. 11 Joo fe(t)cos st dt + i .. 11 Joo~
fe(t) st dt
y21f' -oo
y21f' -oo
=
Loo f(t)cos st dt
Y21f'
J!;
~c{f(t);s},
-oo<s<oo
(2.40)
Based on (2.40), it is clear that the Fourier transform and cosine transform
of an even function give identical results. ln particular, their transforms
are even functions of s (see Prob. 20 in Exer. 2.4). The odd extension
of f is constructed by setting
fo(t)
= f(ltl)sgn(t),
-oo<t<oo
(2.41)
-1
:
1
t<O
t >o
(2.42)
\(i;
-oo
1!-oo fo(t)
~
os st dt + i .. }Joo fo(t)sin st dt
y21f' -oo
00
= .. ;;:c_
y21f'
J!; Loo
f(t)sin st dt
53
= igf5{f(t);jsj}sgn(s),
gf{fo(t);s}
-oo<s<oo
(2.43)
r= e-arcos st dt == s
Jo
+ a2 '
a>O
(2.44)
a>O
(2.45)
and
J =
ioo
o
-ar .
sm st dt =
s2 +a
2,
1
ioo e-arsin st dt == -1 - -J
S
I=--~
a
a o
and
J
= (s/a)I
Solving these last two equations simultaneously for I and J yields the
results given by (2.44) and (2.45).
Example 2.3:
Solution:
= gfc{e-ar;s}
=
AL=
e -arcos st dt
54
2
-
1rs
a>O
2'
~{e-altl;s} =
1
y21r
{ s2
+ a2 '
1 J""
= __
vz:;;.
-oo
e-ist ds
i + a2
+a
J;
a> O (2.46)
= _1 _ e-altl
a
'
a>O
(2.47)
'
a..Vz
'
a>O
(2.48)
Example 2.4: Find the Fourier sine and cosine transforms of te- 01 ,
a> O.
Solution: Formal differentiation of both sides of (2.45), first with
respect to a and then with respect tos, gives us, respectively,
- 2as/(i + a2f
55
and
L= te-a
COS
st dt
.r s te
at ;s} =
and
-at.,s} -3- c te
01:: {
= (a2
s 2)/(s 2
J2
) ,
-7T (s2 2as
+a2 2
-2 ( a2 - s2)2,
7T s +a
+ a 2) 2
a >O
a>o
l=
- e -atsin st dt
o t
= f=
s
da
s 2 + a2
7T
Ia
= - - tan- -
= tan -1 -s
[!li{!
s t
If we allow a
e-ats} =
'
a> O
= J~
(2.49)
This result is only a formal result since neither 1/t nor V7T/2 satisfy the
conditions of the Fourier integral theorem. Nonetheless, it can be useful
to treat (2.49) as a limiting case of the transform relation given in Exam.
2.5. Using (2.43), we obtain the similar relation*
f!Ji{l/t;s}
= iV7T/2 sgn(s)
(2.50)
* Formal results like (2.49) and (2.50) are discussed in more detail in Sec. 2.8.
56
Example 2.6:
Solution:
21
\
a > O.
.:r
= . . 11 Joo eist-aZtZ dt
v21T'
-oo
By writing
we find
y:x;
e-(at-is/za)Z dt
-=
1
- sZf4aZ
=--=e
av'27r
Joo e - xZ dx
-oo
By setting a =
interesting relation
1/0
a>O
3P{e_,zl2;s}
= e-sz/2
(2.52)
which says that the function e- 1212 is self-reciprocal, i.e., it is its own
transform.
EXERCISES 2.4
1. Given the following functions, develop the even and odd extensions,
fe(t) and fo(t), respectively:
(a) f(t)
(c) f(t)
=
=
e-at
(1
t)e-a'
57
= J~ (2e-bs
b >O
- 1),
= {~~ar,
s.
= { 1..
f(t)
~~~.a >O
ltl < b
otherwise
7. f(t)
1
= --=2 -
9. f<t)
= <r + 4)2
4. f(t)
= {~:
~t~~s~
6 f(t)
= {osi'n t,
o< t < 1T
otherwise
s. J<t) = r+
5t + I
m>O
(a)
1T.
f~~ (si:xy dx = 1T
= 1/Yt in the sine and cosine forms of Fourier's integral
theorem given by (2.33) and (2.36), respectively, show that
14. Lettingf(t)
(a)
r~ cos t dt = r~ sin_t dt = J~
Jo Vi
Jo Vt
(1 - i)/2 to derive
58
(b) @i{sin(r/2);s} =
[cos(ii) - sin(i/2)]
= J~ [C(s/y;) + S(sjy';)]*
(b) ~c{e
-ar .
sm at;s}
2
3
-2 as4 + 2a4 ,
n s + 4a
JJ
2 2a - as
n s4 + 4a4
a>O
a>O
t4 ~
e;t}
k >o
g;S{r4: e;s}
k 4) 2
;s}
<b>
g;s{l: e;s}
(d)
g;s{ (14 :
k4)2
;s}
20. Prove the following properties of the Fourier transform for real functions f(t):
(a) If f(t) is even, then F(s) is real and even.
(b) If f(t) is odd, then F(s) is imaginary and odd.
(c) If f(t) is neither even nor odd, then F(s) has ao even real part
and ao odd imaginary part.
2.5
* C(x) and
59
= ~ ;;:;--:_
1
y2rr
J"'
eisr f(t) dt
(2.53)
-oo
Thus, there is only one transform function F(s) associated with each
functionf(t). However, if/(t) and g(t) are two functions that are identical
everywhere except at certain isolated points, then both f(t) and g(t) will
have the sarne transform, say F(s). This means that the inverse transform
of F(s) can be either f(t) or g(t). Of course, the distinction between
functions that differ only at isolated points is mostly of academic interest
and has little effect in practical applications. If we agree to define a
function at a point of finite discontinuity as the average of its left-hand
and right-hand limits, then f(t) is uniquely related to F(s) by the inverse
transform relation
-1 [/(t+)
2
1
+ J(r)] = ~ ;;:;--:_
(2.54)
y2rr -"'
Proof:
@i{Ctf(t)
Czg(t);s}
= ~ ;;:;--:_
1 J""
V 2rr
eist [Ctf(t)
+ C2g(t)]
eist f(t) dt
c J""
yz;
::= _ I _
-oo
= CtF(s)
dt
-oo
c J"'
yz;
_2_
eist g(t) dt
-oo
+ C2G(s)
60
8Ji{f(at);s}
= ~ ;;c
1 J""
y27T
eit f(at) dt
-oo
avz; Joo
= _1-
eiu(s/a) f(u) du
-oo
where we have set u = at. Thus, if F(s) is the Fourier transform off(t),
we have just shown that
8Ji{f(at);s}
= (1/a)F(s/a),
a>O
(2.55a)
-(1/a)F(s/a),
a<O
(2.55b)
f: O
(2.56)
= (1/iai)F(s/a),
ya + r
2
iaiVI + (t/af
8Ji{
2.5.1
sin bya +
.. !2
..2
va +
r. } = {yr;;ji lo (yb2a2 -
,s
rii),
O,
_{yr;;jio, lo(iajyb
-
s2),
!as!< bla!
!as!> b!a!
!si< b
Is!> b
Shift Properties
= ia F(s)
61
Y27T
-oo
= F(s
+a)
= 1
J=
{/a'F(s);t}
Y27T
-oo
1 JQO
.
e-<t-a>s F(s) ds
Y27T -oo
=-=
= f(t
- a)
@'{f(t - a);s}
= eias F(s)
Example 2.8: Find the Fourier inverse transform of 1/(i + ias + b),
b >o.
Solution: By completing the square, we have
i +
1
1
ias + b - [s + (ia/2)f + [(az/4) + b]
r} = e-at/2@'-t {
s 2 + ias + b'
=Jaz
t}
+ a);t}
e;3'- 1{F(s);t}.
62
V27T -oo
= _1_f(tkst
vz:;
I-oo
00
is
I""
V27T -oo
eist f(t) dt
=O
we then obtain
.?F{f'(t);s} = - isF(s)
(2.57)
/t/->oo
/t/->=
then
.?F{J<n>(t);s}
= ( -ist F(s),
= 1,2,3, ...
a,
= sF5 (s)
y2/7T /(0)
(2.58)
63
= -sFc(s)
(2.59)
(see Prob. 9 in Exer. 2.5). For second derivatives, we are led to the
relations
~c{f"(t);s}
(2.60)
and
(2.61)
F(s)
1 Joo
y27T
~ ;;c_
eist f(t)
(2.62)
dt
-00
and formally differentiate both sides with respect to s. This action yields
F'(s)
~ ;;c_
= -iF'(s)
(2.63)
Of course, the validity of (2.63) requires that the transform of tf(t) exist.
Continued differentiation of (2.62) with respect to s leads to
p<m>(s)
= _1_ Joo
\12;
eist
[(it)mf(t)] dt,
m = 1,2,3, ...
(2.64)
-00
= (- i)m p<m>(s),
= 1,2,3, ...
64
Jm
m,n
= 1,2,3,...
(2.65)
= e-s2Jz
EXERCISES 2.5
1. lf f(t) is a real function with transform F(s), show that the complex
conjugate of F(s) satisfies
F(s)
= S'{f(t);- s}
4. Show that
(a) Sic{f(at);s} = (1/a)Fc(s/a),
(b) $ 5{f(at);s} = (1/a)Fs(s/a),
a> O
a>O
a>O
5. Show that
(a) S'c{f(t)cos at;s}
(b) S'5{f(t)cos at;s}
= i[Fc(s + a) + Fc(s
= ![F5(s + a) + F5(s
6. Show that
(a) S'c{f(t)sin at;s}
(b) $ 5{f(t)sin at;s}
= UF5 (s +
= UFc(s -
- a)]
- a)]
a) - F5 (s - a)]
a) - Fc(s + a)]
(b) :'Jic{e
at
cos at;s}
-at .
smat;s}
J= J
=
65
-2 as4 + 2a4
'1f' s + 4a
3
a>O
2 2a - as
4
4
'1f' s + 4a
a>O
= -sFc(s)
~s{f'(t);s}
11. f(t)
13. f(t)
15. f(t)
12. f(t)
1' 1
14. f(t)
= ebr-P
= e _,212 cos 2t
e-2is
1
2
s + 4s + 7
17. F(s) =
s 2 + 4s + 7
= iFc(s) +
2.6
= s4F 5 (s)
66
= J~
= J~ ~ -1)* ik
1r k=O
(2k)!
Jo
where we have replaced the cosine function with its power series
representation. The substitution t = a sin (} in the above integral
leads to
J: (az -
y:;f(2z)
= 22z- 1 f(z)f(z + !)
we find that
k!f(k + p + 1)
This last power series is recognized as the Bessel function JP(as) (see
Sec. 1.4), and thus we have our result
67
2p-l/ 2f(p
+ !)(a/sr Jp(as),
~c{t P JP(at);s}
(az - szy-1/Zh(a - s)
P_ 112aPr(p + !)
,
1
2
-
sz'
O,
2.6.1
lsl <a
lsl >a
(2.67)
(2.68a)
. (7Ta / 2) ,
= J2f(a)
-7T - sa sm
(2.68b)
ar; {a-I }
:1'
st
;s
Let us define the complex function f(z) = za- 1 e- sz and integrate it around
the closed contour shown in Fig. 2.4. From Cauchy's integral theorem
(Theor. A. I in Appendix A), it follows that
(2.69)
fcf(z)dz =O
or
Cp
f(z)dz
(2.70)
iy, so that
= ;a-lya-le-isy
~ oo,
68
p
X
lim
p->0
J f(z)dz = O,
Cp
lim
R->=
J f(z)dz = O
(2.71)
CR
e-hralz,
= s-"f(a)
from which we deduce
= y a-1 e -isydy=--e
f(a) -iwa/2
o
s"
(2.73)
(2.74a)
1/VS
(2.74b)
and
@'s{l/yt;s}
69
~I r j(x)esxdX +
k=l
JLk
k=l
r f(z)szdz + r
Jck
JcR
f(z)eiszdz
n
27Ti
L Res{f(z)eiz;ak}
k=l
where L 1 , L 2 , ,L;..+, are the straight line segments along the x axis
and C~o C2 , ,Cm are small semicircles with centers at the simple poles
b 1, b2 , , bm. ln the limitas R~ oo and the radii ofthe small semicircles
tend to zero, we obtain
f~= f(x)eisxdx -
7T ktl
(2.75)
JcR
70
= iy27T [
+!
Res{f(z)i;ak}
~ Res{f(z)eiz;bk}]
2 k=l
k=l
(2.76)
for s >O.
lf f is either even or odd we can extend the result (2.76) to include
s < O by utilizing the relations (2.40) and (2.43). For more general f we
can set s = -a-< Oand integrate the functionf(z)e-;"z around a contour
similar to that in Fig. 2.5, but in the lower half-plane. Equivalently, we
can replace z by - z and integrate the function f(- z)e-iz around a
contour in the upper half-plane. The result of this latter approach is the
transform relation (see Prob. 9 in Exer. 2.6)
@i{j(t);s}
= iVbr
[~ Res{f( -z)e-i;ak}
"f;
.
]
1 M
+2
Res{f( -z)e_,.z;f3k}
(2.77)
z=
e)
and
Res{f(z)eiz;ik}
= lim (z
z--+ik
g;{ 1(1z
71
e-ks)
( I
2JC - 2/C
+ JC) ;s = iV2TT
1T i 1
= "2/C(
-e -ks) ,
s> O
1
t(t2
s} =
+e)'
J~2 _i_(l
e
- e-kJI)sgn(s)
Solu1ion: By definition,
1
g;- 1{F(s);t} =---=
V2TT
= ~ 11
J""
e-lls F(s) ds
-=
J" 00
y27T
eits F(-s) ds
-=
= zz - zaz
. +b
= z 1 and z = z2 ,
where
Zt
Z2
= 2(a
Ya
+ 4b)
Clearly, z1 lies in the upper half-plane while z2 is in the lower halfplane. Calculating the residue at z = z1 leads to*
Res{F(- z)eitz;zt}
eitz
.
2Z - la
I
z=zl
* Recall that if z =
Res{f(z);a} = lim (z
z-tJ
P(z)/Q(z), then
-Q~);(z) = P(a)/Q'(a).
Z
72
f!F-1{ i+ ias+
1
r}
b'
= Ja 2 ~ 4b exp[ -i(a +
ya + 4b)t].
t >O
z3 =
Z4
1
iaz + b
= z2 +
=
z3 and z
i
-Z(a
z4 ,
where
+ ya 2 + 4b)
..
I 2
-Z(a- ya
+ 4b)
Therefore,
;JF-1{ s
t}
+ ias + b'
=Jaz
~ 4bexp[ -~(a- ya
+ 4b)t].
t<O
+ 4bltl>]
f(z)
= cosh z
around the rectangular contour shown in Fig. 2.6. The function f(z)
has simple poles at z = (n + l/2)7Ti, n = O, 1, 2, ... , but the
* Recall that we previously found this inverse transform relation througb use of the
shifting property (see Exam. 2.8).
73
-R + i1r
u i7T/2
-R
fcf(z)dz
z = R + iy,
fc2 f(z)dz
"'
I Ii "'
i
and thus
eaR-sy ei(sR+ayl
. dy
cosh(R + ty)
e-sy
:5 eaR
o jcosh(R
+ iy)j
dy
lt can be shown that the last integral vanishes in the limit as R oo (see Prob. 20 in Exer. 2.6). The sarne in true of the integral along
the line segment C4 We have z = x + i7T along the segment C3 ,
which leads to
-R
c,
f(z)dz
e<a+is)(x+iTr)
cosh(x
= e(ia-s)Tr
Therefore, as R (1
or
oo
. dx
l7T)
(a+is)x
_e_ _
-R
cosh x
dx
+ e(ia-s)Tr)
ao
(a+is)x
_e__ dx
-ao cosh x
= 27Te(ia-s)Tr/2
74
J
oo
-oo
COS h X
2Tre(ia-s)Tr/2
dx -
1+e
(ia-s)Tr
2Tr
1T
= ------::-------=
cosh[ (ia
~ s)Tr]
By splitting this last expression into real and imaginary parts, we get
oo
-oo
cosh ax
cos sx dx
COS h X
foo
sinh ax .
h sm sx dx
-oo
COS
1T
a1r
cos- cos
2
h s1r
. a1r . h s1r
2- l. smsm 2
2
a1r
s1r
. a1r . h s1r
T
cosh
1r sm 2 sm 2
2
=------------------------+i----------------1r cos
S1T
at. } _
:JP s
h ,s cos t
oz: {sinh
~v;;;-::
2~1T
lal < 1
a1r . hs1r
810
cos a1r
810
lal <
EXERCISES 2.6
l. Find the Fourier cosine transform of h(a - t), a > O and compare
this with Exam. 2.10 to deduce that
Jl/2(x)
1TX
sin x
:JP{f(t);s}
-IYn! (2)n+l/2
V2
ln+ l/2(s)
;n
ltl);s} = Vs In+tds),
= 0,1,2, ...
that
75
= 0,1,2, ...
4. Show that
~5{t(a 2
a> O,
1/2)Jp(as),
> 1/2
6. Show that
CJi::
(a) 3-{ltl
-a
;s}
JZ
= :;;
(b) ~{ltl-"sgn(t);s}
7. Given that
ya
f(l - a) . 1TQ'
=
2
sm 2,
lslt-a
z f(llslt-a
-
a)
:;;
+i and tan
O<a<1
1Ta
cos 2'
O<a<1
show that
s/a,
a> O,
~s{e
- - J2
attp
;s}
= -
f(p) sm
. p(J,
a> O, p >O
-p
1T r
p >O
9. f(t)
=f +9
l
11. j(t)
= t4 +
13. f(t)
(12
+ 4)2
10. f(t)
2it
= -2-1
t +
12. f(t) =
14. j(t)
-1
-6
P-t
= t(t2 + I)
ln Probs. 15-17, use residue theory to find the inverse Fourier transform
of the given rational function.
15. F(s)
= -2 - l
s + 1
16. F(s)
2is + 1
~
s~ + 1
76
1-
17. F(s)
= (i + 4)2
t2k
y--
+1
-=[
d2k
e,.., dt = (- 1i d 2k
s
f=
eist
- 2- -
-=t
+1
dt
19. Following the suggestion in Prob. 18, derive the Fourier sine transform
of t(r + 1)- 1Jo(at).
20. Show that
(a) icosh(R + iy)i ;::: eR /4,
R ~ oo.
(b) Use (a) to show that the integral along C2 in Exam. 2.12 satisfies
the inequality
ILJ<z)dz
R~ oo
r f(z)dz
Jc2
o.
.
.
.
sinh(z/2) . .
cosh(z/2) ..
21. By mtegratmg the funct10ns
. h e''z and
. h
e''" around the
sm z
sm z
contour shown in the accompanying figure, deduce that
cn: {sinh(t/2). } _
3'c
sinh t ,s cn:
3's
{cosh(t/2).
} _
h
,s SIO t
-R+
27Ti~-----.._
-R
-p
77
g;- {F(s);t} = O,
t < O
1 L= e-fr
= --=
-= d~
'
v'1r
o v'~
'
>o
f:F c{(l
- t)h(l - t);s} ==
J~'TTS~(1
- cos s)
cn;; { sinh at . } _ _
1_
sin a
24 3-c . h ,s - ~ ;;::;-::_
h
'
sm 1rt
y 21r cos s + cos a
lal < Tr
cn;; {cosh at }
1
sinh s
3-s sinh 1rt ;s == y'21r cosh s + cos a'
ial < Tr
25
78
26.
e-altl }
{ --;s
Vi
sin;ts }
27. ~s { t
(y's2 + a2 + a)l/2
,
2?T(i + a 2)
1
= ~ ~log
y2?T
a>O
11-+-sI
1- s
28.
~c{r 1 (e-b'
- e-a');s}
29.
~s{r 1 (e-b'
- e-a');s}
j~ tan-{ ~~ ~ :~].
30.
~ei~~a~; 1\s} =
2.7
1z
a> b
::).
lo(ay'b2 - s2)h(l
a> b
ls/bl)
(fo g)(t)
1 Joo
=~~
y2?T
-oo
f(u)g(t - u) du
(2.78)
= -1 Joo
2?T -oo
1 Joo
2?T -oo
-oo
- u) du dt
(2.79)
= -1 Joo foo
eis(u+x>.J(u)g(x) dx du
-oo -oo
1 Joo eisuf(u)du ~ ~
1 Joo eisxg(x)dx
=~~
y2?T -oo
y2?T -oo
2?T
g)(t);s} = F(s)G(s)
(2.80)
where F(s) and G(s) are the Fourier transforms, respectively, off(t) and
79
g;- 1{F(s)G(s);t}
= (fo g)(t)
(2.81)
(2.82)
u) du
= O, which
= f(u),
(2.83)
then*
* ln
= (Cf) a g = C(fa g)
most cases of interest to us the function f(t) is real and thus f(t) = f(t).
(2.85)
80
+ k) =Jog +Jok
(2.86)
By definition,
(g a f)(t)
1 f=
= ;;;-:_
v 27T
g(u)f(t - u) du
-=
= -./
y27T
y27T
t - u, we have
f=
=. 11-
f(v)g(t- v) dv
-=
Jog=goJ
Finally, the convolution integral (2. 78) also satisfies the associa tive law
Jo(gok)
= (fog)ok
(2.88)
but we leave the proof of this result to the exercises (see Prob. 8 in
Exer. 2.7).
Example 2.14: Use the convolution theorem to evaluate the inverse
Fourier transform
Qj;"-l{
Yft
Solution:
sin s . ;t }
s(l -
IS)
Let us define
sin s
F(s) = -
1
and G(s) = - - .
I - IS
J7T2 h(l -
Iti)
;t} = V27Te_,
h(t)
g;- 1{ -sin- s ;t } =
5
and
1
g;- 1{ --.
I -
IS
81
-ao
J~2
J~2
or
e- 1 fr eu du,
e-r fi eu du,
~-~{
sins t } =
s(l - is)'
-1
_ e -<r+ ll),
J~ (e-
e- )e-
ltl < 1
1
,
t> 1
t < -1
O,
Example 2.15:
I-f=
-
Solution:
>1
t < -1
O,
J~ (I
;'f
ltl < 1
-1
- =
(xz
dx
az)(xz
bz)
I = f~= F(x)G(x) dx
where F(x) = 1/(x2 + a 2) and G(x) = 1/(x2 + b2). Thus, by the use
of Eq. (2.83), we immediately can relate this integral to
J
f~oo J(t)g(- f) d/
where f(t) and g(t) are the inverse Fourier transforms of F(x) and
G(x), respectively. Recalling Exam. 2.3, we see that
f(t)
and using the fact that these are even functions, we deduce that
I
= 2 Loo j(f)g(f) df
=
.!!._
ab
r= e -(a+b)t dt
Jo
82
or
2. 7.I
Other convolution integrais involving the cosine and sine transforms can
be derived similarly, but the resulting integrais are more complicated
than (2.82). For example, ifjand gare functions defined for t >O which
have cosine transforms Fc(s) and Gc(s), respectively, then from previous
results we know that
~{f(ltl);s}
= Fc(s)
~{g(ltl);s}
= Gc(s)
and
The substitution of these expressions into (2.82) yields
(2.89)
Since both Fc(s) and Gc(s) are necessarily even functions, we can write
(2.89) as
2
ui) du + J:oof(iui)g(it
- ui) du +
ui) du
roo
l roo
Jo cos(st)Fc(s)Gc(s) ds = 2Jo f(u) [g(it
- ui) + g(t +
u)] du
(2.90)
roo
I roo
Jo cos(st)Fs(s)Gs(s) ds = 2Jo f(u) [g(u + t) + g(u - t)] du
roo
I roo
Jo sin(st)F5 (s)Gc(s) ds ; 2Jo f(u) [g(iu - ti) - g(u + t)] du
roo
l roo
Jo sin(st)Fc(s)Gs(s) ds = 2Jo f(u) [g(u + t) - g(u - t)] du
(2.91)
(2.92)
(2.93)
83
EXERCISES 2.7
1. Verify the convolution integral (2.82) for
(a) f(t) = g(t) = h(l - ltl>
(b) f(t) = g(t) = e-r2f2
1
F(s) = (1 - ls)2 = 1 --is. 1 -is
u) du
and use this result to evaluate the Fourier transform of e-ltl sin t.
t
oo
-oo
(x 2
dx -
_'Tr_
a+b
+ a2)(x 2 + b2)
@'{(a 2
J~J0(as),
>O
to show that
(""
Jo
J 0(ax)J0(bx) dx
O<a<b
= 1rbK(ajb),
f"/2
Jo
(1 -
m2sin28) -l/2 dO
oo
b dx
= e-a
21
7ra
smaxsm x -2 = x
2
a> O
84
8. Show that the Fourier convolution (2. 78) satisfies the associative
property
Jo(gok)
= (fog)ok
9. For real g(t) = f(t), show that (2.90) leads to the Parseval relation
o< a < 1
= r'",
=
(1 - t
Y- 112h(l
- t),
v> -1/2
a;)ro -
~ sin(
(b) Using properties of the beta function (see Prob. 21 in Exer. 1.2),
evaluate the integral on the right in (a) and deduce that
r=
2a-v-lr( /2)
Joxa-v-llv(x)dx=f(v+ ::a/ ),
O<a<l,v> -1/2
3- c
0(at) . }
2
1 ,s
t+
J~ e
1T
-s ia " cosh
u d
I 2
2 u,
oya-u
>O
o;;; {
lo(at) . } _
:-!/'c 12
+ 1 ,s
J~2 e -sIo(a ),
85
a>O
2.8
ln Sec. 2.4 we stated that the Fourier transform F(s) is a bounded function
provided the inverse transformf(t) is absolutely integrable, i. e., provided
00
(2.94)
Impulse Function
86
J~= (t -
a)f(t) dt = f(a)
(2.95)
1
--=
v27r
(2.96)
.:1'- 1{1;1}
J""
= .. 11
v2Tr -=
e-m (1) ds
- (s)
= V2Tr
(2.97)
1
v27r
(2.98)
where iF(w)i is the spectrum amplitude function and cp(w) describes the
spectrum phase.
The sinusoidal functions f(t) = cos w0t, which represents one of the
simplest waveforms possible, does not satisfy the condition (2.94), and
hence, does not have a Fourier transform in the strict sense. Yet, from
a purely physical point of view we know this waveform has a single
frequency component, or fine spectrum, at w = w0 This apparent contradiction can be explained by recognizing that the function cos w0t
cannot exist for all time < t <
as we assume in the formal
definition of the Fourier transform. Like all waveforms, the cosine
waveform exists for only some finite interval of time, and as such will
87
= ~ 11-
g;{f(t);w}
2'7T
JT
eiwt
cos Wol dt
-T
dt
or
F(w)
(2.99)
The graph of jF(w)j is shown in Fig. 2.7 for a fixed value of T. By allowing
T to become unbounded, it can be shown that the graph becomes more
and more narrow and peaked around w = w0 and w = - w0 Hence, in
a formal sense we may consider the Fourier transform of cos w0t for ali
time as the limit
~{
'
.
= 1lm
r-.oo
-=
(2.100)
1Joo
00
1
f(x) = :;
= _!. Joo
1T
-oo
(2.101)
f -
Based on the sifting property (2.95), it becomes clear that the limit in
(2.101) leads to the formal definition
jF(w)j
-wo
Figure 2. 7 Spectrum of sinusoidal pulse
Wo
88
1 . sin (t - x)
"'(
)
-tm
=ut-x
1
7T -->co
t - X
(2.102)
With this interpretation of the limit, we see that (2.1 00) gives the expected
result*
$'{cos w0t;w} = YTT/2 [(w + w0) + (w - w0 )]
(2.103)
= iJ~ sgn(s)
= J21
-:7T l(
-co
J Jco
-=
Y27T
eist
sgn(t) dt =
-J~~
7T
IS
-co
-!..
7TS
Example 2.17: Find the Fourier transform ofthe Heaviside unit function
h(t).
= 2[1 + sgn(t)]
Hence,
CJJof
89
[ :s]
5(s) +
-is~{h(t);s}
= - -.zsyl2;
= G(s),
but
= G(s) + k8(s)
EXERCISES 2.8
l. Show formally that
(a)
~{5(t
- a);s}
(b) ~{eiat;s}
= yl2; e.a
= yl2; 5(s + a)
= (- istF(s),
1,2,3, ...
(a) ~{'(t);s}
(b) ~{(n)(t);s}
= (-ist,
1,2,3, ...
*Observe that Bl{l;s} = Bl- 1{l;s} dueto the even property of f(t) = l.
90
= 1,2,3, ...
= 1,2,3, ...
.!. = (- on-1
t"
(!)
dn-i
(n -l)!dtn-i
f"
(is)n-i
_ 1)! sgn(s),
= 'y2 (n
= 1,2,3, ...
7. If
J~oc f(t) dt =
F(O)
f O
J~oc h(t -
theorem.
8. Starting with the identity t8(t) = O,
(a) show formally that
'(t)
= - ! S(t)
t
s(n)(t)
= 1,2,3, ...
= (-on-m(n!/m!) a(m)(t)
2.9/Hilbert Transforms
91
(n) (
k=O
(b) f(t)S(n)(t)
-l)k f(n-k)(O)S(k)(t)
k
ln Probs. 11-15, verify the given Fourier transform relation.
k=O
J~ (
13 .9i'{C"h(t);s}
n!
. )n+l,
7T -ls
J~ ;n [s<n>(s)
n = O,I ,2,...
(~!:n~!J.
= r(~a)isi"- 1 i"(l-a)sgn(s>,
n == 0,1,2, ...
a > 1 (a nonintegral)
27T
14 .9i'{ltl-";s}
J~ f(l
> 1 (a non-
integral)
15. .9i'{ltl-"sgn(t);s}
> 1
(a nonintegral)
2.9
Hilbert Transforms
= R(w) +
iX(w)
(2.104)
from only knowledge of the real functions R(w) or X(w). If the function
f(t), which is the inverse Fourier transform of F(w), is a causal function
[i.e., f(t) = O for t < 0], then this mathematical problem can be solved
by means of Hilbert transforms, which fundamentally provide integral
relations between the real functions R(w) and X(w).
A related problem involving Hilbert transforms is to find a complex
representation of a real signal. For instance, if x(t) is a real signal and
x(t) is its Hilbert transform, then the analytic signal
u(t)
= x(t) +
ix(t)
(2.105)
92
Bf{/(t);w} =
(2.106)
F(w)
= -1
g;- 1{F(w);t}
27T
(2.I07)
= f~""f(t)cos wt dt
(2.I08)
and
=-
X(w)
J~= f(t)sin wt dt
(2.I09)
where R(w) and X(w) are the real and imaginary parts of F(w) as given
in Eq. (2.104). We can immediately deduce from this that R(w) is an
even function and that X(w) is an odd function.
ln the special case when f(t) is an even function, we see that
R(w)
= 2 L= f(t)cos wt dt
(2.110)
and X(w) = O, and through the inverse cosine transform relation, it can
be shown that
f(t)
= -I
7T
i""o R(w)cos wt dw
=-
L= f(t)sin wt dt
(2.111)
= O and
(2.112)
Thus,
f(t)
= - -I
7T
i=
o
X(w)sin wt dw
(2.Il3)
2.9/Hilbert Transjorms
93
Uf(t) -
J<- t)J
we have
= fe(t) + fo(t)
(2.114)
fe(t)
(2.115)
fo(t)
= ![f(t)
(2.116)
f(t)
where
is an odd function. From our discussion above, it now follows that R(w)
is the Fourier transform of fe(t) while X(w) is the Fourier transform of
fo(t). Hence, we have the transform pairs
R(w)
= 2 L"" fe(t)cos wt dt
(2.117)
1""o R(w)cos wt dw
(2.118)
X(w)
= -2 L"" / (t)sin wt dt
0
(2.119)
fo(t)
= -7T1 1""
X(w)sin wt dw
o
(2.120)
fe(t) = -I
7T
and
ln the case when f(t) is causal, it happens that f(- t) == Ofor t > O, and
therefore we deduce from (2.115) and (2.116) that
f(t)
= 2/e(t)
== 2/0 (1),
t>O
(2.121)
21""o R(w)cos wt dw
= -7T
(2.122)
and
f(t)
21""o X(w)sin wt dw
= -7T
(2.123)
94
(1)
= fo(t)sgn(t)
= J.,(t)sgn(t)
(2.124)
(2.125)
Hence, using (2.124) and properties of even and odd functions, we can
rewrite (2.117) as
R(w)
= J~ e-;,, fe(t) dt
= J~oo e-;,, /
(t)sgn(t) dt
(2.126)
(2.127)
.cf{sgn(t);w} = 2/iw
(2.128)
and
1
2
-oo
R(w)
=-
= - 1T Joo
iX(y) [.(
l
)] dy
(IJ -
or
1 Joo X(y)
dy
-oo w-y
(2.129)
1T
Similarly, since
iX(w)
dt
-oo (IJ -
dy
(2.130)
Equations (2.129) and (2.130) form what we call a Hilbert transform pair.
They are also called Kramers-Kronig relations in electromagnetic theory
2.9/Hilbert Transforms
95
= - .!_ J""
7T
X(y) dy
-co (J)
+Y
= R(w)
(2.131)
=~roo
Jo
7T
y;(y) dy
(2.132)
w - y2
=-
2w
7T
roo R(y) dy
Jo w2 - /
(2.133)
EXERCISES 2.9
ln Probs. 1-4, verify that the given functions satisfy the Hilbert transform
relations (2.129) and (2.130).
a
1. R(w)
=-
2_ R(w)
=1-
4 R( w)
w +a2
cos w ~ w sin w,
-w
= acosw- asinw
2
w +a
'
a>O
96
J"'
~-o
-a>
w- y
f(t).
8. Show that
(a) X(w)
= .!.. J"'
1T
l"'
->
R(w) - R(y) dy
w - y
R~y) dy
(b) X(w)
= 2w
(c) X(w)
=;Jof~ R'(~lo~~~
R(w)2-
+ yl
w - y
1T
dy
9. Let F(w) be a complex function of w that is analytic in the upperhalf w-plane and tends to zero as w ~ oo. [The condition that F(w)
has no poles in the upper-half w-plane is equivalent to the causality
condition that the inverse transform of F(w) vanishes for t < 0.]
Then show that (2.129) and (2.130) follow by applying Cauchy's
integral theorem to the integral
/(wl)
CW-
F(w)
e~
o+
w1
ie
Re(w)
2.10/Additional Topics
97
=!
7T
Joo
-oo (
x(r) dr
-
J~= x(t)x(t) dt = o
11. Let x(t) be a real waveform which has no constant component. Show
that x(t) and its Hilbert transform .X(t) satisfy the Parseval relation
F(s)eirG<s> ds
(2.134)
98
(2.135)
Similarly, we approximate F(s) by F(s0 ) , the first term in its Taylor
expansion. Under the assumption that the major contribution to the
integral occurs in the neighborhood of S0 , these approximations lead to
(2.136)
By formally extending the limits of integration over (- oo,oo) and using
the result (see Prob. 11 in Exer. 2.10)
yl2;"r + i1r/4
e-+ ia2u2 du =--e-
oo
(2.137)
lal
-=
we deduce that
l(t)- F(s 0 ) y2?T/tiG"(s0 )1 exp[itG(s0 ) i?T/4],
-H>O
(2.138)
where the plus sign goes with G"(s 0 ) > O and the minus sign with G"(s 0 )
<o.
cos[t(s2
oo of
s)] ds
we find s0
Re{ f
2
eit(s -s)
= 1 and G(s)
= 2s
ds}
= s2
- 1, G"(s)
s. Thus, calculating
=2
1T)
'
00
2.10/Additional Topics
00
99
(2.139)
= -1
21T
(2.140)
and
f(x,y)
= _1 Joo foo
21T
-oo -oo
e-i(X+Tjy)
F(,Tj)dd'Yj
(2.141)
We can interpret (2.140) as an iterated Fourier transform applied sequentially to the variable x and then to y. That is, we define
f*(,y)
= "'}
V 27T
f"" x f(x,y) dx
-oo
followed by
F(,Tj)
1 J""
= y;i;
-= ei.,y f*(,y) dy
(2.146)
aY
8i'(2){iJxa/x,y);x~,
Y~'YI } = -'Y/F(,'Y/)
(2.147)
8i'<2>{
100
with a similar expression for the inverse transform. For reasons of theoretical tractability, one normally does not use Fourier transforms in
dimensions greater than two. Actually, even though physical systems ali
have three-dimensions, we are often able to ignore one or two dimensions
in our analysis by using some reasonable simplifying assumption, such
as symmetry, and so on. The classical example where simplification to
one or two dimensions is not possible, however, is the problem of diffraction
of X-rays by crystals, which must be analyzed in three dimensions.
EXERCISES 2.10
ln Probs. 1-3, verify the given asymptotic behavior of each integral.
1.
3.
t~=
- s) ds - 1 J1T ,
21
(b) As t ~ oo, show that the result on the right-hand side in (a)
approaches that given in Exam. 2.18.
S. lf G'(s 0 ) = G"(s0 ) = ... = G<m-t)(s 0 )
Eq. (2.138) is generalized to
b
a
J F(s)e
itG(s)
ds -
f(1/ m )
m
F(s 0 )
= O,
G(m)(S 0 )
=f O, show that
.
I ]1/mexp [ztG(s
m.
tjG<m>(so)l
l1T
]
0)
2m
2.10/Additional Topics
101
Jix)
7T
- ns) ds,
= 0,1,2, ...
8.
9.
cos(ts4)tan s ds
(I
eil(s-sin s)
10.
ds
(a
) -1-
y:I;r
J""
J2 [l"" yv
-00
7T
i"" -Vv- vd ]
COS V d
---vz
sin V
-1-
Vf;
J""
1 iw/4
e ia2u2 du=-e
lal
-oo
ei<a~+ln,JF({,'r/)
wx;x-, Y-"1}
= ![F(+w,'r/) + F(-w,'r/)]
= (f* g)(x,y)
where
<f* g)(x,y)
J"" I'"'
;::: -211'
1 _.., _.., f(u,v)g(x-u, y-v) du dv
3
Applications Involving
Fourier Transforms
3.1
Introduction
103
(3.1)
(3.2)
oo
oo
= Y(s)
= - s2 Y(s)
and
[1i{h(l - lxl);s}
=- 1
vfi;
-1
1T
* For
Andrews, Elementary Partia/ Differential Equations with Boundary Value Problems, Orlando:
Academic Press, 1986.
104
J2
s
-sin
7T
with solution
Y(s) =
v;g
sin s
s(s2
1)
Finally, the solution y(x) that we seek is simply the inverse Fourier
transform of Y(s), which we can obtain through use ofthe convolution
theorem. That is, we have
fffi- 1 {
and
= fffi- 1{Y(s);x}
}Joo e-iui h(l
=2
- lx- ul)du
-oo
1lx+l e-iui du
=-
x-1
-oo<x< -1
sinh (l)e\
1 - e- 1 coshx, -1~x~1
{
sinh (l)e-x,
1 <x<oo
105
used to solve the problem. The decision as to which of these two transforms
to use will depend upon the type of boundary condition specified at the
finite boundary x = O. That is, from Eqs. (2.60) and (2.61) in Sec. 2.5,
we recall
J~ y'(O)
(3.3)
+ J~ sy(O)
(3.4)
,q;c{y"(x);s}
-s2 Yc(s) -
g;s{y"(x);s}
-s Y5 (s)
and so if y(O) is specified we use the sine transform, whereas the cosine
transform is called for when y'(O) is specified. Problems involving these
transforms are taken up in the exercises.
Finally, we remark that the technique illustrated in Ex. 3.1 can readily
be generalized to boundary v:alue problems of the more general form
y"+ay'-by=f(x),
-oo<x<oo
y(x) --+ O, y'(x) --+ Oas lxl --+ ao
(3.5)
=-
F(s)
i+ ias+ b
(3.6)
The inversion of (3.6) by use of the convolution theorem [see Eq. (2.81)
in Sec. 2.7] gives us the solution formula
y(x)
= - f~""f(u)g(x
- u) du
(3.7)
(3.8)
where we define
Oj;-1
g(X ) -- _1_
~ ~ ~
y21T
s + zas +
b .,X }
i<ax + \la
+ 4bjxj)J
(3.9)
106
-oo
< x < oo
y(x) =
y<4>(x) =
f(x)
L= (Els
(Eii + k)A(s)
= 2F0 sin s
1T
(Eis
+ k)B(s)
107
y
Figure 3.1 Infinite beam on elastic foundation
= 2F0
1r
sin s
s(Els4 + k)
2Fo f=
sin s
y(x) = -:;;:- Jo s(Els4 + k) cos sx ds
The final evaluation of this integral by using residue theory is left to
the exercises (see Prob. 8 in Exer. 3.2).
3.2.1
(3.10)
k(x,t)u(t) dt
where f(x) and k(x,t) are known functions and the function u(x) is to be
determined. Such equations are called linear integral equations of the
first kind. A familiar example of this kind of equation is the integral
transform
F(s)
= ~ ;;:;:
1 J""
y2Tr
eist f(t)
dt
(3.11)
-oo
108
propriate conditions)
f(t)
1 I""
= ~ r::>=
V 2'7T
. F(s) ds
e-rts
(3.12)
-oo
J: k(x,t)u(t) dt
(3.13)
u(x) -
J: k(x,t)u(t) dt
(3.14)
J: J: jk(x,tW dt dx <
oo
1
r::>=
y2'7T
foo
-oo
u(t)k(x - t) dt
= f(x),
=< x < =
(3.15)
109
in which the functions f(x) and k(x) are prescribed for all real values of
x andare assumed to possess Fourier transforms F(s) and K(s), respectively.
Taking the Fourier transform of each side of (3.15), we have formally
= F(s)
U(s)K(s)
(3.16)
= ~-
{L(s);x}
(3.17)
1
y27T
= r,;-:
foo
f(t)f(x - t) dt
(3.18)
-oo
ln some cases it may tum out that L(s) = 1/K(s) does not have the
inverse transform given by (3 .17), but the related inverse transform
m(x)
= ~- 1 {(-is:nK(s);x} = ~-
{M(s);x}
(3.19)
= [(- is)nF(s)]M(s)
(3.20)
1 I""
= r,;-:
y27T
J<n>(t)m(x - t) dt
(3.21)
-oo
Solve
f(x)
!--: J=
y27T
-oo
Solution:
F(s)
and thus,
U(s)
110
U(s)
= lxl- 112sgn(x)
and
8fr- 1{( -is)F(s);x}
= f'(x)
=-
1
-
1
lx - gl- 11'i'(g)dg - - -
IX
V2rr -
oo
112
f'(g)dg
11
u(t) dt
= f(x),
-oo<x<oo
-oo
U(s) - 22U(s)
s +4
= F(s)
where U(s) and F(s) are the Fourier transforms, respectively, of u(x)
and f(x). This algebraic equation has solution
U(s)
= F(s) + 2F(s)/(i + 2)
1 Joo
= f(x) +----=
y2
e-v21x- 11 /(t) dt
-oo
EXERCISES 3.2
1. Given the integral
x+l
I=
x-1
e-lul du
111
show that
r+
(a) I
L-1
(b) I
l
l
(c) I=
x-1
x+l
x-1
eu du, - = <X< -1
eu du +
ix+l
e-u du,
-1sxs1
2. y" - y
-=<X<=,
e-lxl,
O< x < =
4. y" - y = e-x,
= O,
O, y'(x)
y'(O)
6. y" - y
y(x)
= xe-x,
= O,
O as x
O< x < =
=
Hint: $ 5{fe-';s} = V2f; 2s(3 - s )/(i +
y(O)
y(x)
O, y'(x)
O as x
1)3
-=<x<=
1 J""
= ~ r-c
y2'7T
j(g)g(x -
g) dg
-00
where
g(x)
lvT
= v;
K e-Kix! 2 [cos(Kx/\1'2) + sin(KJxJ/y2)]
2 3
8. Use residue theory to show that the integral solution in Exam. 3.2
can be reduced to
_ Fo [ -b(l+x)/vT b(l + x)
Y(x) - 2k e
sm v'
2
where b4
= k/El.
-b(l-x)/vT
b(l-x)]
sm y'2
112
ln Probs. 9-14, solve the given integral equation of the first kind.
9.
oo
.
u(t)erxr dt
-oo
oo
Sn X
= -X-
10.
2av-:;i
oo
13.
J-oo
14.
J~oo u()u(x -
-=
u()
1:)2
X-~
+a
d =
u(t)cos xt dt =
7T e -x
2
e -x2f4a20 +r>
y'1 + t
b2 , O < a < b
) dg = e-x
e-lxl
0<<1
x<O
x2:::0
1 Joo
-oo [1
u(x) = :;
tu(t)dt
(x - t) 2](1
t 2)
-loo -,-sinxt
y27T
dt
o e - 1
2xoo
7T L
=- +7TX
1
2
n=IX
+n
1)
2x
3.3
113
= a- 2u,
- q(x,y,z,t)
(3.22)
114
the temperature u along each finite surface of the solid (Dirichlet condition);
the second condition is to prescribe the flux of heat across the surface,
which is accomplished by specifying the normal derivative of u at the
surface (Neumann condition); and the third boundary condition is to
prescribe the rate at which heat is lost from the solid due to surface
radiation into the surrounding medium (Newton's law of cooling). This
last boundary condition is sometimes called Robin's condition.
When the unknown function u depends only upon the spatial variable
x and on time t, then (3.22) becomes
Uxx
= a- 2U
1 -
q(x,t)
(3.24)
=a -2U 1
(3.25)
Both (3.24) and (3.25) are called one-dimensional heat equations. Among
other areas of application, these one-dimensional heat equations govern
the temperature distribution in a long rod or wire whose lateral surface
is impervious to heat (i. e., insulated). For modeling purposes we assume
in this case that the rod coincides with a portion of the x axis, is made
of homogeneous material, and has uniform cross section. Further, we
will assume that the temperature u(x,t) is the sarne at any pont in a
particular cross section of the rod, but may change from cross section
to cross section.
3.3.1
Let us first consider the flow of heat in the infinite medi um - oo < x <
when the initial temperature distribution f(x) is known and the region
is free of any heat sources. Physically, this problem might represent the
linear flow of heat in a very long slender rod whose lateral surface is
insulated. ln such cases, the solution will represent the temperature
distribution in the middle portion of the infinitely long rod, prior to the
time when such temperatures are greatly influenced by the actual boundary
conditions of the rod. The problem is mathematically characterized by
ao
-oo<x<oo,t>O
B.C.:
I. C.:
u(x,O)
= f(x),
oo
lxl ~ oo
<x<
(3.26)
oo
115
1 J<X>
= ~ r:c_
y27T
eixu(x,t) dx
U(s,t)
(3.27)
-<X>
=-
isU(s,t)
(3.28)
~{u.u(x,t);x~s}
= -
iU(s,t)
(3.29)
J=
eix
and
~{U1(X,t);x~s} = . . }
v 27T
U 1(x,t)
dx
-00
1 afoo .
== ... 1 e"x u(x ,t) dx
v 27T iJt
-<X>
or
(3.30)
Using the above results together with the relation F(s) = @i{f(x);s},
we find that the Fourier transform applied to the problem described by
(3.26) leads to the transformed problem
UI
I. C.
+ a2s2 U
U(s,O)
=O,
= F(s),
>o
(3.31)
-oo<s<oo
The solution of the original problem is now found by taking the inverse
Fourier transform of (3.32). The product form of (3.32) suggests use of
the convolution integral, which yields
u(x,t)
I J<X>
= ~ r:c_
y27T
-00
f()g(x-,t) d
(3.33)
116
where
1
(3.34)
= ~ J""
f()e-(x-t;J2f4a2r d
2ay'TTt -=
(3.35)
The integral (3.35) expresses the solution u(x,t) as a sum of the effects
of the initial temperature distribution f() at various points along the
infinite rod. The heat kemel
- -1- e -(x-t;)2f4a2t
2av:;i
1 J""
= y;
-=f(x + 2azy't)e-z
dz
(3.36)
= . . To;- Joo
v 1T
e z dz
= T0
(3.37)
-00
Oas
lxl ---+ oo
* Recall our discussion in Sec. 3.2. Also, since solutions given by (3.35) and (3.36)
can be derived independent of the Fourier transform method, their validity may extend
beyond that of the latter.
117
Example 3.5:
-oo<x<oo
Using the transform relation
~{ -x2J4a2 }
~ ;-2
-a2s2
:!/'e
;s = v a e
Solution:
= Vae-azszo
U(s,t)
+ I)
Rather than use the convolution theorem, we can idvert this solution
directly to obtain
U(X,f)
Vl+t
e -x2f4a2(1
+ I)
a- 2u1
B.C.:
I.C.:
u(x,O)
= O,
t>0
q(x,t),
>O
(3.38)
$i{q(x,t);x--+s}
= Q(s,t)
U(s,t)
t>O
I. C.:
(3.39)
-=<s<=
U(s,O) =O,
= a2
e-azszu- T)Q(s,T) dT
(3.40)
= ~ ~ 2 foo
V 27T
-oo
ii
g(x-, t-T)q(,T) dT d
(3.41)
118
=.
Uxx
=a -2U 1 ,
B.C.:
ux(O, t)
I. C.:
u(x,O)
0 <X<<=, t
>0
(3.42)
The fact that the interval is semiinfinite, together with the prescribed
boundary condition at x = O, suggests that the Fourier cosine transform
be used in this case. Hence, if we define*
.?fc{u(x,t);x~s}
U(s,t)
(3.43)
= -
=-
s2 U(s,t)
(3.44)
U,
+ a2s2U = O,
U(s,O)
= F(s),
t >O
(3.45)
O<s<=
= F(s)e-a s
2 21
(3.46)
and by applying the inverse cosine transform, we get the formal solution
u(x,t)
(3.47)
= yffj;
i"" f()cos
s d
* To avoid confusion with partia! derivatives, we have chosen not to subscript the
transfonned functions by C to denote a cosine transform.
119
which yields
u(x,t)
= -
00
7T o o
1
=e-azszt f() {cos[s(x-)] + cos[s(x+)]} d ds
7T o o
1"" 1""
(3.48)
Finally, interchanging the order of integration and using the result (see
Prob. 7 in Exer. 3.3)
f"
e-bxzcos ex dx
= iJ~ e-cz/4b,
b>O
(3.49)
u(x,t)
1""
ef]
{ [
..1c
f() exp - (x - 2
4a t
2a y1Tt o
+ exp[
(x + 2 )
4a t
]}
(3.50)
u(O, t)
I. C.:
u(x,O)
= f(t),
= O,
u(x,t)
~O,
O< x <
oo,
t>0
ux(x,t)
Oas x
~ oo
(3.51)
oo
This time we will use the Fourier sine transform, which leads to the
transforrned problem
I. C.:
Ut + a2s2 U
U(s ,0) = O,
= yl2f; a2sf(t),
O< s <
t>O
(3.52)
oo
where U(s,t) denotes the Fourier sine transform of u(x,t). The solution
of this initial-value problem is
U(s,t)
(3.53)
* Equation (3.50) could also be derived directly from the convolution integral relation
given by Eq. (2.90) in Sec. 2.7.1.
120
and by using the transform relation [see Prob. 7(b) in Exer. 3.3)
f'
_ _ x_
f(T)
r= (
) e -x2f4a2(t-T) dT
2ay?T o t - T 312
u(x,t) -
(3 .55)
= T1 (constant).
Solution: The solution for any f(t) is that given by Eq. (3.55). By
making the change of variable
z = x/2ay't-T
we find that (3.55) becomes
T1 reduces to
u(x,t)
= T1 V2r?T f""
e -z2 dz
x/2a\t
2 {""
= y; Jx e-z
dz
u(x,t)
= T1 erfc(x/2ayt)
The physical interpretation of the solution suggests that for any fixed
value of x, the temperature in the rod at that point will eventually
approach T1 ifwe wait long enough (t-oo). However, at any particular
instant of time t, the temperature u(x,t)- O as x - =, in agreement
with our prescribed boundary condition. Finally, we recognize that
the temperature u(x,t) is constant along any member of the family of
parabolas in the xt plane defined by
x/2aVt = constant
121
+ Uyy = a -2U
I.C.: u(x,y,O)
= f(x,y),
-oo<x<oo, -oo<y<oo,t>O
1,
-oo
-oo
< y < oo
(3.56)
We also assume that u and its normal derivatives vanish near the infinite
edges of the plate.
By introducing the double Fourier transforms (see Sec. 2.9.2)
@i(2){u(x,y,t);x~e. Y~'11}
@i<z>{f(x,y);x~e. Y~'11}
= U(,'7/,t)
= F(,11)
(3.57)
(3.58)
U, + a 2(2 + 11 2)U
U(,'71,0) = F(,'1/)
I.C.:
= O,
t>O
(3.59)
U(,'7/,t) = F(,'1/)e-aZ(z
71z) 1
(3.60)
u(x,y,t)
= _1 I""
2'7T
-oo
J""-= F(,'1/)e-a2(z+TJzlt
e-i(fx+TJYl
d d'11
(3.61)
This solution can also be expressed in the equivalent form (see Prob.
23 in Exer. 3.3)
u(x,y,t)
= 21'7T J""
-oo
(3.62)
where
(3.63)
EXERCISES 3.3
ln Probs. 1-5, solve the heat conduction problem described by (3.26)
when the initial temperature distribution f(x) is prescribed as given.
1./(x) =
{~~ ~~~~~
2.f(x)
={~o. ~~~
122
3. f(x) = {To,
O,
S. f(x)
0
x<O
x>O
e '
Hint: See Prob. 6
O<x<e
otherwise
= e-lxl,
<x<
-oo
4. f(x)
oo
= { '-x
6. Show that
7. Show that
(a)
ex
b >O
Jo
J!
e-c2f2b2
'
>o
u(x,t)
= T0 +
2av;;i
2j4a2t
= u,
- S(x- vt)q(t)h(t),
I.C.: u(x,O)
= O,
-oo
-oo<x<oo,t>O
lxl ~ oo
< x < oo
= ~ rt q(T) e-(x-VT)2j4(t-T) dT
2y7TJo~
= a- 2u,
u(x,O;T)
= f(x,T),
-oo
<x<
oo
123
show that
v(x,t)
= J: u(x,t- T;T) dT
is a solution of
2
a vxx
=V
1 -
f(x,t),
v(x,O)
= O,
= a -2Un
u(O,t) = O,
u(x,O) = f(x),
U:u
B.C.:
I.C.:
O< x <ao
(a) is given by
u(x,t)
= J~ L"" e-a s
2 21
F(s)sin sx ds
2
2; ) ]
exp[-
(x :
2; ) ] }
{0,T xO>c
<x<c
0,
{0,T O
<x<c
x >c
0,
<ao.
15. Given the boundary-value problem
u(O,t)
= O,
u(x,t)
Oas x ~ ao
O<x<ao
= e-x,
<
124
show that
u(x,t) =
1 1/y't, where 1 1 is
= {11 ,
O< t < b
t 2: b
o,
show that
u(x,t)
T1erfc(x/2ay't),
O<t<b
T1[erf(x/2ay't=b) - erf(x/2ay't)],
t2::b
=a -2 u,,
ux(O, t) = - f(t),
B.C.:
I. C.:
u(x,O) =O,
O<x<oo
show that
U (X,f )
u(x,t) ~O as x ~ oo
lt
= v-7 r 0~ yc-e
f-T
a
f(-r)
-x2j4u2(t-T)
d
T
19. Show that the solution in Prob. 18 can also be expressed n the form
u(x,t)
= v'x_ r~-
/(t-x~/4lh~)z-~ e-:O dz
7T J.r/1a\/
J~ e-x
214 2
a r
x erfcCavr)]
3.4/Mechanical Vibrations
u(O,t)
u(x,O)
B.C.:
I. C.:
= O,
u(x,t)
= O,
O<x<oo
~O
as x
125
~ oo
ux(O,t)
B.C.:
I. C.:
= O,
u(x,O) =O,
u(x,t)
g;(;lt
Oas x
t>0
~ oo
O<x<oo
oo,
{e-a2(2+...,2)r;~x, TJ~Y}
= (1/2a2t) e-<x2+yZ)f4a2t
(b) Use the result of (a) to deduce the solution formula (3.62) of the
problem described by (3.56).
126
problem is characterized by
Uxx
= C-2 u,,,
-oo<x<=,t>O
lxl ~ ao
B.C.:
u(x,t) ~ Oas
I. C.:
u(x,O) = f(x),
u,(x,O)
(3.65)
The infinite extent on x once again suggests the use of the Fourier
exponential transform. Hence, by introducing
.'Y'{u(x,t);x~s}
= V(s,t)
(3.66)
t>O
(3.67)
+ c 2s2V = O,
V(s,O) = F(s),
Uu
I. C.:
= G(s)
V,(s,O)
G(s) .
sm cst
+-
(3.68)
cs
=~
~ f=
y27T
+ G(s)sin cst]
ds
(3.69)
CS
-oo
2\(i;r
I=
[e-i<x-ct>
e-is<x+ct>] F(s) ds
-oo
+ _1_ f=
2y'z;
[e-is(x-ct) - e-is(x+ct)]
~(s) ds
lCS
-oo
(3.70)
+ f(x +
ct)]
=v'1271' foo
-oo
e-,.zG(s)ds
(3.71)
3.4/Mechanical Vibrations
127
r+ct g(z)dz
Jx-ct
y'21r
I""
-oo
u(x,t)
1[
= 2 f(x
1 r+ct
2
+ c Jx-cr g(z)dz
(3.72)
3.4.2
Consider a uniform elastic beam lying along the x axis in its equilibrium
position as shown in Fig. 3.2. If the beam is initially displaced from this
equilibrium position in the vertical plane it will cause the beam to vibrate
freely in the transverse direction. From the elementary theory of small
beam deftections it is known that the bending moment M at any cross
section of the beam satisfies the relation*
M = EI/R
= Eluxx
(3.73)
Mxx
yA
= - -uu
g
(3.74)
where 'Y is the specific weight of the beam, A is the area of the cross
section, and g is the acceleration of gravity. The term on the right-hand
side in (3.74) representing an inertia force actually behaves like a load
intensity along the entire length of the beam. Combining (3.73) and (3.74),
* See S. Timoshenko, Strength of Materiais, 3rd ed., New York: Van Nostrand,
1955.
128
- ---
....
....
X
\ :it:l displacement
Figure 3.2 Freely vibrating beam in transverse direction
+ a -2 u,, = o
(3.75)
where a = ElgjyA.
For illustrative purposes, let us consider the case involving the free
vibrations of a uniform beam of infinite length where the motion is
produced by distorting the beam initially in the shape f(x) with zero
velocity. Assuming units such that a2 = 1, the problem is formulated
by
Uxxxx
+ u,
B.C.:
u(x,t)
I. C.:
u(x,O)
= O,
~O,
-oo<x<oo,t>O
~
ux(x,t)
= f(x),
O as lxl ~
u,(x,O) = O,
-oo
oo
(3.76)
< x < oo
I. C.:
+ s4 U =O,
U(s,O))
t>O
= F(s),
(3.77)
U,(s,O) =O
with solution
U(s,t) = F(s)cos ts 2
(3.78)
u(x,t)
= -1- J""
yl2;
e-zsx
F(s)cos ti ds
(3.79)
-oc
g;;- 1 {cos(ti);s~} = ~1 r
2yt
[cos(e/4t) + sin(e/4t)]
(3.80)
3.4/Mechanical Vibrations
129
} Joo
= ~ ;;:;-::-.
2 V 211'1
+ sin(e/4t)] d
f(x-) [cos(e/4t)
(3.81)
-oo
~ lc
2 V 11'1
Joo
. [(x-71)
f(TJ) sm
4(
-oo
l
+ -11'
d71
4
(3.82)
EXERCISES 3.4
ln Probs. l-4, solve the vibrating string problem described by (3.65)
when the initial conditions are prescribed by the given functions.
1. f(x)
2. f(x)
= e-lxl,
= O,
g(x)
g(x)
= O,
e-lxl,
<x <
-oo
-oo
<x <
oo
oo
V0 constants)
B.C.:
I. C.:
u(x,t) ~O as x ~ oo
u(O, t) == O,
u(x,O) == f(x),
u,(x,O) = O, O < x <
oo
= Ufo(X +
= O,
B.C.:
ux(O,t)
I.C.:
u(x,O) =
e-x,
u(x,t)
oo,
t>0
~O
as x
~ oo
oo
130
( ' t)
U X
e-' cosh x,
cosh t'
x<t
x>t
= { e- r
8. Solve
Uxx =C
-2
u,,,
B.C.:
u(O,t) = f(t),
I.C.:
u(x,O) = O,
9. Solve
Uxx
= Uu,
B.C.:
ux{O,t) = -1,
I.C.:
u(x,O) = O,
u(x,t) ~O as x ~ oo
10. Solve
=U
Uxx
11 ,
u(x,t) ~O as x ~ oo
B.C.:
I.C.:
12. If f(x)
O< x < oo
-oo<x<O
= {F.o,o'
=
lxl < 1
lxl >I
,
I [e--/4(1+ir>
u(x,t) = -
e-x-/4(1-it>]
+ ---:==-
2 y'I + it
(b) By setting I + it = Reitt>, where
R =
yf1+7,
y'I - it
tan cf>
=t
u(x,t)
= R-l/2e-r2(costt>t/4Rcos(x2~;cf>- ~)
U 11
= O,
B.C.:
u(O,t)=I,
I.C.:
u(x,O) = O,
U.u(O,t)=O,
u(x,t)~Oasx~oo
u,(x,O) = O, O< x
< oo
3.5/Potentilll Theory
131
= 2J~
(1 - :os
tsz)
= 1-
c(~)- s(~)
where C(x) and S(x) are the Fresnel integrais (see Sec. 1.3.2).
+ Uu = 0,
B.C.: u(O,t)
= f(t),
I.C.: u(x,O) = O,
0 <X <
uxx<O,t) =O,
oo,
t>0
u(x,t) ~O as x ~
oo
oo
show that
(a) the solution of the transformed problem is
U(s,t) =
3.5
1 100 f(t
= ~c
V 1T x/V'ft
- x 2/2v 2) [cos(!v2)
+ sin(!v2)] dv
Potential Theory
Uyy
+ Uyy = 0
+ Uzz = 0
(3.83a)
(3.83b)
132
V2u =O in R
(3.84)
u =fone
V2u =O in R
iJu/iJn = f on C
(3.85)
* For example, see O. D. Kellog, Foundations of Potential Theory, New York: Dover,
1953.
= Vu
3.5/Potential Theory
133
= U(s,y)
= F(s)
(3.87)
(3.88)
Uyy - s2 U = O, y >O
U(s,O) = F(s), U(s,y) ~O as y
(3.89)
134
+ B(s)eY
where A(s) and B(s) are arbitrary functions of s. A valid solution for all
values of s is readily found to be
-ao< s <ao
U(s,y) == F(s)e-IIY,
(3.90)
g;-l{e-IIY;s~x} == f!:_
V;.x
(3.91)
+l
and applying the convolution theorem, we are led to the solution formula
Y
u(x,y) == -
1T'
Joo ( J(}1:) d+
-oo X Y
2
2,
(3.92)
Y >O
C,
<b
>b
0Jb
yT
u(x,y) == - 1T'
-b
(t
To [ tan== -:;;:
dt
-
)2
+ y2
(X-Y+ b)
(X -
b)]
- tan- 1 -Y-
-I(
2by b )
2
x2 + y
- 2
3.5/Potential Theory
135
x 2 + l- Cy
= b2
uy(x,O)
= f(x),
-oo
<x<
(3.93)
oo
B.C.:
+ Vyy
v(x,O)
= -(uxx
ay
+ Uyy)
= O
(3.94)
= uy(x,O) = f(x)
v(x,y) = 7T
J""
-oo
f() d
X -
<:)2
+ Y2
y>O
T0
(3.95)
136
Chap.
The solution function u(x ,y) that we are seeking can now be obtained
by a simple indefinite integration of (3.95), leading to
u(x,y)
= Jv(x,y) dy
f""
l
:::: :;; -ooj(g)
J(x - yeidy + l
dg
1 J""
= Z1r
-oc/(g) log[(x
g)2 + /] dg
(3.96)
-=<x<=,O<y<a
u(x,O) = f(x), u(x,a) = g(x), -ao< x < oo
(3.97)
(3.98)
u==f
Figure 3.5 Infinite slab
3.5/Potential Theory
137
B.C.:
O<y<a
U(s,a) = G(s)
= F(s),
(3.99)
where F(s) and G(s) are Fourier transforms, respectively, of f(x) and
g(x). The solution of this boundary-value problem is easily shown to be
(see Prob. 13 in Exer. 3.5)*
U(s,y)
- y) + G(s) s~nh sy
smh sa
smh sa
= F(s) sinh.s(a
(3.100)
k(x y) =
'
sin(7Ty/a)
2 cosh(7Tx/a) + cos(7Ty/a)
(3 .10l)
the details ofwhich we leave to the reader. t Hence, using the convolution
theorem we can express the inverse transform of (3.100) in the form
u(x,y)
.Jz; [J~=f()k(x
- , a - y)d +
f~= g()k(x -
, y)d]
(3.102)
B.C.:
Uyy
= O,
u(x,O) =!(x),
{ ux(O,y)- O,
O< x <
oo,
O< y < a
oo
(3.103)
138
3'duxix,y);x~s}
= U(s,y)
(3.104)
=-
iU(s,y) - V2/Trux{O,y)
=-
s 2 U(s,y)
(3.105)
and
3'df(x);s}
= F(s)
(3.106)
(3.107)
= F(s) cosh(a
(3.108)
with solution
U(s, )
y
- y)s
cosh as
{cosh(a - y)s.
}
h
,s~x
cos as
sin(TTy/2a) cosh(Trx/2a)
= cosh(Trx/a) - cos(Try/a)
_ ar.- I
g ( x,y ) -ore
(3.109)
1
= ;;c
1= f()[g(x + ,
y21T o
y) + g(lx- l, y)] d
(3.ll0)
3.5/Potential Theory
139
EXERCISES 3.5
ln Prob. 1-4, use the prescribed function
Dirichlet problem (3.86).
1. f(x)
= T0 , -= <
3. f(x) =
<=
2. f(x) = {
g: ~ ~ ~
~: ~ ~ ~
= L>O
Y(s,y) [A(s)cos sx
+ B(s)sin sx] ds
B(s)
= ;1 Joo-oo f(x)sin sx dx
ln Probs. 6 and 7, use the result of Prob. 5 and the prescribed function
= {~.-
~~~ ~!
xz,
7. f(x)
{~~X,
~~~
B.C.:
Uyy
= O,
u(O,y) =O,
{ u(x,y)- O
u(x,O)
= f(x),
l) -
as (x 2 +
oo
= ; Jo
x)2
+ y2
- (t
+ x)2 + y2
dt
+ Uyy = 0,
140
B.C.:
oo
and show that the solution has a form similar to that in Prob. 8.
11. Solve Prob. lO when
f(x)
{1,0, O<x<c
C< X<
oo
uxx + Uyy
B.C.:
u(x,O)
= O,
-xe
-x2
-oo<x<oo,y>O
oo,
(a) show that the transformed problem via the Fourier transform has
the solution
U(s,y)
= (1
e-ls!y)~e-s2f4
2V2s
u(x,y )
= -21Tl
l""o (1 - e
-sy)
sin-xse -s2j4 ds
s
13. Verify that (3 .1 00) is the solution of the transformed problem (3 .99).
14. Find a solution similar to (3 .1 02) for the potential problem
Uxx
B.C.:
+ Uyy = 0,
u(x,O) = f(x),
-oo<x<oo,y>O
uy(x,a)
=O
u(O,y) =O
{ u(x,O) = f(x), u(x,a) = g(x)
B.C.:
Uxx
B.C.:
-oo<x<oo, -oo<y<oo,z>O
{ u(x,y,z) """"O as (x 2 + l +
r)
112
""""
oo
u(x,y,z)
= 2}
1T Joo
-= J""
-=f(-,TJ)g(x-
,y - TJ,Z) ds d71
3.6/Hydrodynamics
141
3.6 Hydrodynamics
A ftuid ftow in three-dimensional space is called two-dimensional if the
velocity vector V is always parallel to a fixed plane (xy plane), and if
the velocity components parallel to this plane along with the pressure p
and ftuid density p are all constant along any normal to the plane. This
permits us to confine our attention to just a single plane which we
interpret as a cross section of the three-dimensional region under consideration. Our discussion here will be limited to two-dimensional ftow
problems.
An ideal jluid is one in which the stress on an element of area is
wholly normal and independent of the orientation of the area. * ln contrast,
the stress on a small area is no longer normal to that area for a viscous
fluid in motion. lf the density p is constant, we say the ftow is incompressible. Of course, the notions of an ideal ftuid or incompressible ftuid
are only idealizations that are valid when certain effects can be safely
neglected in the analysis of a real ftuid.
The velocity, pressure, and ftuid density are all interrelated through
a set of differential equations consisting of a continuity equation, equation
of motion, and an equation of state (such as the density equal to a
constant, etc.). The continuity equation is an expression ofthe conservation
of mass, which states that the ftow of mass into a region equals the ftow
of mass out of it, and assumes the form
~+V (pV)
=O
(3.111)
(3.112)
which implies there are no sources nor sinks within the region of interest
(i.e., points at which fluid appears or disappears). The equation ofmotion
is a consequence of Newton's second law of motion. By equating the
rate of change of momentum of an element of ftuid to the total force
(i. e., the resultant F of the extemal forces and the net stresses acting
on the element), we obtaint
av
-+(V V)V
ilt
= F- -Vp
p
1
+ vV2V +- vV(V V)
3
(3.113)
* An ideal ftuid
142
F = - VA,
(3.114)
av-v
X
at
n +
v(A + !vz
+ !!..)
2
p
+ vVxn
=o
(3.115)
aw
aw
aw
at
iJx
Jy
- + u- + v-
vV w
(3.116)
= (u,v,O)
(3.117)
where
fi
and V2
= (O,O,w),
= - t/ly,
t/Jx
(3.119)
~ vzt/1 + (Jt/1 ~ at
Jy iJx
(3.121)
3.6/Hydrodyno.mics
143
+ 2t/Jxxyy + t/lyyyy
= O
(3.123)
(3.124)
This relation combined with (3 .119) leads to Laplace' s equation
(3.125)
Clearly, solutions of Laplace's equation V t/J = O are also solutions of
the equation of motion given by (3.121).
The irrotational ftow of an ideal ftuid can also be described in terms
of a velocity potential function cf>. That is, the condition V x V = O
implies the existence of a potential function cf> such that V = - Vcf>, or
2
= -cJ>x,
= -c/>y
(3.126)
By combining (3.126) with (3.118), we find that the velocity potential cf>
is likewise a solution of Laplace' s equation
(3.127)
Thus, for irrotational ftows we have the choice of solving (3.125) for the
stream function t/J or solving (3.127) for the potential function cf>.
Example 3.8: Consider the irrotational ftow of ao ideal ftuid filling the
half-space y ~ O through the strip lxl :::: a (see Fig. 3.7). If the ftuid
is introduced normal to the region with prescribed velocity v = f(x),
find the resulting velocity components u(x,y) and v(x,y) within the
half-space.
Solution: Let us formulate and solve the problem in terms of the
velocity potential. ln this case the problem is characterized by
B.C.:
+ c/>yy =O,
-oo<x<oo,y>O
cf>y(x,O) = -f(x)h(a - lxl),
-oo < x < oo
cf>xx
144
</Jx, </Jy
B.C.:
~O
as y
oo
= - 217T fa-af()log[(x-)2 + /] d
u(x,y)
=-
<!Jx(x,y)
= -7T
v(x,y)
= -
cf>y(x,y)
= -7T
fa ((X - 1:.))f()
+y
-a X ~
and
Y
Ia
-a X -
J()
1:.)2
+ y2 d
The solution of Exam. 3.8 in terms of the stream function 1/J is left
to the exercises (see Prob. 4 in Exers. 3.6).
3.6.2 Surface Waves
We consider here the development of two-dimensional gravity waves on
a semiinfinite body of ftuid y :s Ofrom the action of an impulsive pressure
on the free surface y = O of the ftuid. We regard the ftuid as ideal and
initially at rest. lt then follows from the laws of hydrodynamics that the
motion will be irrotational throughout ali time, and hence we can describe
such motion in terms of a velocity potential q, satisfying Laplace's equation
cPxx + cPyy = O
(3.128)
3.6/llydrodynalnics
145
(3.129)
= -p(<f>,
gy
+ !V2)
(3.130)
= TJ (x,t)
1
(3.131)
which expresses the fact that a ftuid particle initially on the free surface
continues to remain on the free surface during the wave motion. The
dynamical boundary condition, corresponding to the requirement p = O
at the free surface, is
(3.132)
<f>,(x,O, t) - gTj(x,t) = 0
Combining (3.131) and (3.132), we obtain the single boundary condition
B.C.:
(3.133)
(3.134)
= J p(x,O, t) dt
(3.135)
146
l/>1(x,O,O) = O
I.C.:
(3.136)
<l>yy - i<l> = O,
-ao
B.C.:
LC.:
<l>(s,O,O)
(3.137)
= (1/p)Q(s),
<l>(s,y,t)
= (1/p)Q(s)cos(Vglslt)ell.v
(3.138)
tf>(x,y,t)
= - - J""
pyfi;
(3.139)
-oo
lf>(x,y,t)
= .. ~
py27T
[r=
e.v Q(s)cos(sx - vgst) ds
Jo
+
(3.140)
.. r-L
(3.141)
where the plus sigo goes with G"(s 0 ) > O and the minus sigo with
G"(s 0) <O.
Now, in (3.140) there are two symmetrical groups of waves moving
in the two directions from the origin. If we consider only the rightrunning waves, then we may drop the second of the two integrais. Writing
the first integral as
3.6/Hydrodynamics
1
p
100 e YQ(s)cos(sx
5
v'21T
147
- ygst) ds
(3.142)
v'g-r - 2xv's =
2n(s
(3.143)
J7
cp(x,y,t) - 1
3.6.3
(gt
(gt
ao
(3.144)
(3.145)
(3.146)
* A moving
path.
== 'l'(s,y)
(3.147)
148
and the transformed problem takes the form of a fourth-order boundaryvalue problem
'l'yyyy - 2s 2'1'yy
B.C.:
'l'y(s,O)
= O,
+ s4'1'
= O, y
-is'l'(s,O)
>O
= F +(s)
(3.148)
where
(3.149)
The characteristic polynomial of this differential equation is*
m4
with m
2im2
+ s4
= O
= [A(s) + B(s)y]e-iIY
(3.150)
where A(s) and B(s) are arbitrary functions. Application of the boundary
conditions in (3 .148) yields the relations
+ B(s) = O
is'l'(s,O) = - isA(s) = F +(s)
which we deduce A(s) = iF +(s)/ s and B(s) = isiA(s).
'l'(s,y) = is- 1(1 + isiy)F +(s)e-I!Y
'l'y(s,O)
from
=-
isiA(s)
Hence,
(3.151)
(3.152)
and then determine
8f- 1 {G(s,y);s~x}
= 8f-
{is- 1 e-iIY;s~x}
iy81'- 1 {e-11Ysgn(s);s~x}
y >O
(3.153)
3.6/Hydrodynamics
149
and similarly,
~- 1 {ie- 1 1Ysgn(s);s~x}
= ~s {e-Y;s~x}
= V2{; x/(x2 + l),
1
(3.154)
y>O
Thus,
~- 1 {G(s,y);s~x}
(3.155)
=;
g) + (x ~ ~/~ l] d
(3.156)
The velocity components u(x,y) and v(x,y) can now be determined from
Eq. (3.119) (see Prob. 9 in Exers. 3.6).
EXERCISES 3.6
1. Show that the elimination of u and v between (3.119) and (3.126)
leads to the Cauchy-Riemann equations
1/Jx
-<{>y
and use these relations to calculate the stream function 1/J from the
expression for<{> given in Exam. 3.8.
2. Solve Exam. 3.8 for the special case where f(x)
3. For the special case of Exam. 3.8 where f(x)
(a) show that
""( ) = ! J""
o/
x,y
-oo
lo(a)
ll
V0 , constant.
= (a 2
x 2)- 112 ,
;~x-j~jy d/:.
(b) From (a), determine integral representations for the velocity components and verify that u(x,O) = O when lxl < a.
4. ln terms of the stream function, the problem discussed in Exam. 3.8
is characterized by
+ l/lyy =O,
-=<X<=, Y >O
1/Jx{x,O) = f(x)h(a - lxi),
-= < x < =.
1/Jxx
B.C.:
Solve this problem for 1/J and show that it leads to the sarne expressions
for the velocity components u(x,y) and v(x,y) as derived in Exam.
3.8 from the potential function.
5. Verify that (3 .138) is the solution of the transformed problem given
by (3.137).
150
(a) cp(x,O,t)
- ygst)
+ cos(sx + ~t)] ds
{ = \,lxjg (~ + gt/2x),
= yfgt2/4x
show that
cp(x,O,t)
where
J =
-zv'i!X
dJ
-= -Po
Trpg dt
sin(w2
2
{ )
d{
where C(u) and S(u) denote the Fresnel integrais (see Sec. 1.3.2)
and where u2 = gf /2Trx.
=.
+ c{lyy = O,
- = <X < =, -h < Y < O
{cflu(x,~t) + !!_cfly(x,O,t) = O
cflxx
B.C.:
cpy(x, h, t) - O
I.C.:
cp,(x,O,O)
= f(x)
{1/ly(x,O)
= g(x)h(a
1/Jx,l/Jy ~o as y
- lxl),
<Xl
-=<x<oo,y>O
1/lx(x,O)
=O
151
3.7
Many of the great mathematicians since the time of Euler have attacked
the problem concerning the state of stress in various elastic bodies under
the action of given forces. ln this section we will briefty discuss some
two-dimensional problems for which the solution may be obtained by
Fourier transform methods.
The effect of body or surface forces on a two-dimensional body will
be to produce internal forces between various parts of the body. The
magnitudes of these internal forces are defined by the ratio of the force
to the area over which it acts, called the average stress. ln the limit as
the area shrinks to zero, we obtain the components of stress at a point
in the elastic medium. This stress is composed of two normal components,
CTxx and CTyy, and two shearing components, CTxy and CTyx for which CTxy =
CTyx * We adopt the convention that stresses are positive when a tension
is produced and negative when a compression occurs.
The DEs satisfied by the components of stress in an elastic medium
under the action of a force per unit mass having components <Fx,Fy)
may be obtained by applying Newton's second law of motion to a small
rectangular element of the medium. Writing the displacement vector as
u = (u,v), these equations of motion aret
2
iJCTXX
O(Txy
F
au
--+-+p x=pdX
ay
af
CT
o2V
CT
ax + --1!
ay + pFY = Paf-
(3.157a)
(3.157b)
where p is the mass density ofthe elastic body. For equilibrium problems
the time derivatives on the right-hand sides can be set to zero. Also, in
the absence of body forces we have Fx = Fy = O, and in these cases
the equations of equilibrium take the simpler form
dCTxx
ax
+ dCTxy =O
ay
(3.158a)
(3.158b)
152
v(uxx
+ --:2 [uyy
Uyy)]
- v(uxx
ou
+ CTyy)] = 2 ~ (3.159)
k~
a2x
(1'
xy
= - -xy
(3.160)
~
X
(u - a2x)
(3.161a)
axx)
(3.161b)
a ( Uyy
y 2
XX
+ 2_il + 4X =O
arai
y4
(3.163)
153
a4x
a4y
-+2--+-=0
iJx4
iJx2iJy2
al
'
B.C.: o-xx(O,y)
=-
O<x<oo, -oo<y<oo
(3.164)
p(y),
(3.165)
0 as X - oo
a2x
iJy
= -
2 (O,y)
a2x
p(y), -(O,y)
iJxiJy
(3.166)
= G(x,s)
(3.167)
d 4G
dx4
B.C.:
d 2G
4
2s dx 2 + s G = O,
2
x>O
(3.168)
. dG
[A(s) + B(s)x]e-llx
(3.169)
(3.170)
vz;
J""
-oo
P(s)
- - 2 (1
S
+ isix)e- 11
SX
_.
rsy
ds
(3.171)
The stress components that we desire are now calculated using (3.162),
and are given by
154
<Txx(x,y)
= - ~ ~ J""
y27T
P(s)(l + isix)e-lslx-isy ds
(3.172)
~;
P(s)(l - isix)e-lslx-isy ds
(3.173)
-oo
<Tyy(x,y) = -
<Txy(x,y) = -
V 27T
J""
-oo
~ J""
\[i;
sP(s)e-llx-isy ds
(3.174)
-oo
For the special cases where P(s) is either an even function or an odd
function, these equations take on a simpler form. Specifically, if P(s) is
an even function, then
J:
<Txx<x,y)
= -V2/7T
P(s)(l
+ sx)e -sxcos sy ds
(3.175)
<Tyy(x,y)
= -V2/7T L= P(s)(l
- sx)e-xcos sy ds
(3.176)
<Txy(x,y) =
-V2/7T X L= sP(s)e-xsin sy ds
(3.177)
<Txx(x,y) =
<Tyy(x,y)
= -
<Txy(x,y) =
(3.178)
(3.179)
(3.180)
EXERCISES 3.7
1. Use the convolution theorem to show that Eqs. (3.172)-(3.174) can
be expressed in the altemate form
2x3
<Txx
= ---:;;
<Tyy
= --;;
<Txy
= ---:;;
Joo
p(y _ 7])
-=(x2 + 112fd1]
(r
z.x2 [
-=
7]p(y - 7])
(xz + 7]2)2 d1]
155
2. Find a result similar to that in Prob. 1 for the special case when p(y)
is an even function.
3. Use the result of Prob. 1 to show that, when the prescribed pressure
is p(y) = polz(y), p 0 constant,
(a) the stress components ofthe problem described by (3.164) become
1 . 2() )
+ -rr() + -sm
2rr
- Po ( -1
2
1 . 2() )
+ -() - -sm
rr
2rr
Po 2
= -cos
8
rr
a-xy
where x
Po ( -1
2
=-
a-
r cos () and y
r sin ()
= !(a-xx
- a-yi
+ u;S 12
IYI),
p 0h(a -
Po constant.
= Po (a 2 rr
IYI), p0 constant
Show that
(a) a-xx
+ O"yy = -\{i]; Po
(c) At y
+ 2ia-xy =
Jo(as)e -sxcos sy ds
j ;2 PoX Jo
r~
sJ0(as)e-s<x+zyl ds
a2)-t/2
= V2/rr PoXz(~ +
az)-3/2
6. Find the stress components inside the semiinfinite medium y :::::: O such
that
-oo<x<oo,y>O
B.C.:
a-yy(x,O)
= O,
O"xy(x,O)
= q(x)
156
2. lim P(x)
=1
J:= p(u) du
(3.181)
The function p(x) is called the probability density function (PDF) of the
random variable X. Once p(x) has been determined, various properties
of the random variable X can be calculated, such as the statistical moments
of X.
ln statistics, the moments m 1, m2 , , of the random variable X are
defined in terms of the expectation operator E. For example,
m 1 = E[X]
= J~= xp(x)dx
(3.182a)
(3.182b)
whereas in general,
mk = E[x*] =
I~"" x*p(x)dx,
= 1,2,3, ...
(3.183)
The first moment gives the average value of a random variable, and the
higher-order moments give additional information about the spread of
the distribution defining the random variable. The variance of the distribution is defined by
rl
f~"" (x -
m1i p(x) dx
= m2
mi
(3.184)
Characteristic Functions
157
(3.185)
= Y2Tr @'{p(x);t}
(3.186)
= I~"" p(x) dx =
(3.186a)
C'(O)
= i I~"" xp(x) dx = im 1
C"(O)
=-
I~"" x p(x) dx = 2
(3.186b)
m2
(3.186c)
I""
'7T
-ao
e -ltx C(t) dt
(3.188)
Example 3.9: Find the characteristic function associated with the normal
or Gaussian density function
p(x)
1
= --e-<x-m)2f2rr2
-yi2;(T
158
= (1/<T).cf{e-(x-m)2/2u2;t}
= eilm.cf{e-u !2;u~<Tt}
2
= px(x)py(y)
1
-x2/2
=--=e
-1- e -y2j2
y27T
y'2;
= _.!._e -(x2+y2)/2
27T
which leads to the double integral
Cz(t) = _1_
27T
I=
-oo
e-y2f2
I""
eitxy e-x2!z dx dy
-oo
-t
* When working with more than one PDF, we find it convenient to use the more
distinguishing notation Px(x) instead of just p(x).
159
= _1_ J""
Cz(t)
vz:;
y1
e-o + ,zlyz/z dy
-ao
t2
J""
1 I""
-= -1
e-itz Cz(t) dt
27T
-oo
27T
-oo
= _!_ (""'
7Tj0
e-itz
y't+7 dt
cos zt dt
y't+7
where we have used the fact that C2 (t) is an even function. This last
integral is not an elementary integral nor does it lend itself to evaluation
by conventional means using basic complex variable theory. Nonetheless, it is a well-known integral which leads to the final result*
pz(z)
= - Ko(izi),
7T
EXERCISES 3.8
1. A certain random variable X is known to have the characteristic
function given by C(t) = (1 - 2it) -t.
(a) Determine the first moment m 1 and variance
(b) Find the PDF of X.
a-2 of X.
{1/Za-,
o,
ix-mi s aix-mi> a-
160
1 ""
mk- k.
j~
- 2j
cr mk-lj
(2j
+ l)!(k- 2j)!
where
[k/2]
= {~z_,l)/2,
keven
kodd
1
= -e-lx-ml/u
2o-
= eimt/(1 + a-2t2)
(b) Use (3.188) to invert the characteristic function in (a) and show
that it leads to the Laplace PDF.
1 = 2~
e x/lUZ h(x)
= X + Y,
show that
161
z = X1Y1
(a) when N = 2.
(b) when N = 4.
(c) Compute the variance of Z for arbitrary N.
8. Objects illuminated by laser light exhibit what is called a speckle
pattem. The normalized intensity E of the speckle pattem is usually
govemed by the negative exponential PDF, PE(x) = e-xh(x). The
addition of N independent speckle pattems on an intensity basis leads
to a total intensity pattem described by
I
= E) + E2 + ... + EN
= !e-W+Y>I2J0(AVY)h(y)
SQUARE
LAW
LOWPASS
FILTER
Z=LlJ
j=l
INTEGRATOR
4
The Laplace Transform
4.1
Introduction
162
4.1/Introduction
163
= e -cr f(t)h(t)
(4.1)
does have this property, where c is a positive real constant and h(t) is
the Heaviside unit function. It follows therefore that f must satisfy the
condition
oo
= _!_I""
27T
e-ist
-oo
I""
g(x)eisxdx ds
-oo
or equivalently,
f(t)h(t)
= ect I""
27T
e-ist
-=
r= f(x)e-(c-is)xdx ds
Jo
(4.2)
= -1.
ic+i=. i=
21T l c-=
eP1
e-Py(x)dx dp
(4.3)
when his powerful new mathematical methods perplexed them, and he also alienated his
fellow electrical engineers who didn't know what to make ofhim. Because ofpoor relations
with his fellow scientists, much of his work did not receive the credit he deserved. Even
today his name is often not mentioned in connection with some of his most important
contributions to science, such as his pioneering work on discontinuous functions using
operational methods and his mathematical solution for the distortionless transmission line.
164
f'
e-pr f(t) dt
= .X{f(t);p}
(4.4)
=- .
27Tl c-=
(4.5)
4.2
joo
Re(p)
C -
joo
165
a real variable. For now we will proceed with such integrations and later
examine the conditions under which this formalism is valid.
Example 4.1: Find the Laplace transform of ear.
Solution: By definition,
L= e-pt eatdt
= L= e-(p-a)tdt
2{eat;p} =
t=
e-(p-a)t
-(p - a) o
Thus, the integral diverges for Re(p) :::; a, while for Re(p) > a, we
get
Re(p) >a
2{ea1 ;p} = 1/(p - a),
The restriction Re(p) > a that is required in Exam. 4.1 is typical in
the evaluation of Laplace transforms. That is, the transform integral (4.4)
will be meaningful only for complex values of p in some half-plane and
not for others, if indeed the integral converges at ali. Such restrictions
are important in developing the general theory of Laplace transforms
but are of little consequence in many of the applications of the transform.
That is, in practice we are generally satisfied if we know that the transform
exists for some values of p without knowing specifically for which values.
Notce that by allowing a --7 o+ in the result of Exam. 4.1, we get
the limiting case
2{1;p}
= 1/ p,
(4.6)
Re(p) >O
Example 4.2: Find the Laplace transforms of cos at and sin at.
Solution: From definition we have
2{cos at;p}
= L= e-pr cos at dt =
2{sin at;p} =
p/(p2 + a 2),
+ a 2),
Re(p) >O
Re(p) >O
which follow directly from the results of Eqs. (2.44) and (2.45) in
Chap. 2. Also, we note the relations
2{cos at;p} =
2{sin at;p}
166
where f!fc and f!f s denote lhe Fourier cosine and Fourier sine transforms,
respectively.
Example 4.3:
Solution:
Existence Theorem
Our evaluation of Laplace tra'.sforms thus far has been purely formal,
using elementary integration techniques of real variables. We have not
addressed the question as to which class of functions actually have
Laplace transforms. That is, like the Fourier transform, not ali functions
(even continuous functions) have a Laplace transform.
Suppose that f is a piecewise continuous function with the further
property that there exists a real number c0 such that
lim lf(t)le-c' = { 0 . .
no hm1t,
~->=
c> Co
c < c0
(4.7)
= o,
c>a
/-->00
167
e1\ since
==
Remark: Functions that are identically zero for t :::: t 0 > O are said
to be of exponential order -=.
To establish that a given functionf(t) has a Laplace transform F(p),
we must show that the Laplace transform integral
F(p)
f'
e-P1 f(t)dt
:5
Jo
Jo
where the first integral with finite limits exists because f is piecewise
continuous. Furthermore, the second integral satisfies
Jro
Jro
Jro
But this last integral exists for c > c., and thus we have established
conditions under which the Laplace transform integral converges absolutely
in the half-plane Re(p) > c0 lt can also be shown that the Laplace
transform integral converges uniformly for Re(p) :::: c2 > c0 , where c2 is
any real number satisfying c0 < c2 :::; c.
ln summary, we have the following existence theorem.
Theorem 4.1 (Existence theorem). lf f is piecewise continuous on t :::: O
and is O(ec01), then f(t) has a Laplace transform F(p) in the half-plane
168
L= e-pt r dt
'
=r=
e-ptrll2dt = VTT/p
Jo
Also, the Laplace transform may exist in certain instances whenfis not
of exponential order, although we will not pro vide any general discussion
of this case. t
4.2.2 Analytic Continuation
Theorem 4.2.
ln some cases the function F(p) may be analytic to the left of the
line Re(p) = c0 , although for our purposes it suffices to know that there
exists a half-plane where F(p) is indeed analytic.
Up to this point we have produced Laplace transforms
F(p)
* Since p is complex, one really needs also to specify the branch of the multivalued
or in general.
function p' 1Z, which may be either
t For a discussion of the convergence of the Laplace transform integral in the general
case, see W. R. LePage, Complex Variables and the Laplace Transform for Engineers,
New York: Dover, 1980.
:1: For a proofofTheor. 4.2, see R. V. Churchill, Operational Mathematics, New York:
McGraw-Hill, 1972.
yp
VP
169
EXERCISES 4.2
ln Probs. 1-10, evaluate the Laplace transform of each function directly
from the defining integral.
=f
2. f(t)
3. f(t) = cosh at
4. f(t)
1. f(t)
5. f(t)
7. f(t)
= t sin kt
9. f(t)
h(t - a),
a>O
6. f(t)
8. f(t)
10. f(t)
sin at sin bt
= e-ar - e-bt
= sinh at
= te 21
= cos 2kt
=: ea
cosh kt
=r
14. f(t)
= sin
16. f(t)
= log 1
17. f(t)
= 1100
= l
= (sin t)/t
= r 312log t
18. f(l)
= sin(e12)
19. f(t)
= Vltan ti
20. f(t)
11. f(t)
13. f(t)
15. f(t)
112
2
e' log t
2~ J~
J~
;p} = y; t!' erfc(p)
22 .P{t5/2;p} = 81;3
23 .P{e- 1214
2
Va2y'1rp
eap
a >O
erfc(y'ap),
a>O
170
26 .P{log t;p} =
~ [ f'(l) -
log p]
ln Probs. 27-30, the given functions are various types of pulses of unit
height and duration T. Determine their Laplace transforms.
27. f(t)
= { o:
1
O<t<T
t>T
28. f(t)
=
a<t<a
otherwise
1,
{ O,
30. f(t) =
29. f(t) =
2t/T,
2- 2t/T,
O,
+ T, a > O
sin at,
{ o,
O<t<1r/a
t > 1r/a
4.3
+ C2g(t);p}
= C,F(p)
+ C2G(p)
171
= -1
a o
the last step of which is the result of making the variable change u =
at. Thus, if F(p) denotes the Laplace transform off(t), the above relation
suggests that
1
a>O
.T{f(at);p} = - F(p/a),
(4.8)
Example 4.4:
= y; eP
Solution:
.T{e- 1 ;p}
4.3.1
y;
= 2 in the scaling
= T eP erfc(p/2)
2j4
.T{ea'l'(t);p}
Proof:
= F(p
- a)
From definition,
.:t{ealj(t);p}
= F(p-
a)
172
Example 4.5:
Solution:
= pf(p2 + 9)
5t{e
cos 3t;p}
p +2
= (p + 2)2 + 9 =
p2
p +2
+ 4p + 13
- a)h(t - a) dt
- a) dt
'
Jo
du
(4.9)
2{h(t - a);p}
1
= -e
-ap,
173
(4.10)
a>O
~:
~~~ < 2
4.3.2
=2
-
p3
4 -2)
+ e- 2 p (2-p
p2
p3
f"
e-pt f'(t) dt
= e-pt f(t)
[ + p
L= e-pt f(t) dt
~O
= pF(p)
as t
~ao,
- f(O)
and consequently,
(4.11)
174
= p.:l'{f'(t);p}
- f'(O)
which simplifies to
.:l'{f"(t);p}
= p 2F(p)
(4.12)
- pf(O) - f'(O)
= pnF(p)
and therefore
,p} -- n.lj p n+l , n -- 1,2 ,3 , ... *
(J){ln.
..L
1/p.
lf
5t {J:f(u)du;p}
175
f is piecewise continuous on
= F(p)jp
Let us define
g(t)
J: f(u)du
or
5t {J:f(u)du;p}
= F(p)jp
(4.13)
176
= - F'(p)
(4.14)
= 1,2,3, ...
= (-It p<n>(p),
n = 1,2,3, ...
= pf(p2 +
I)
d ( p )
p - 1
.P{t cos t;p} = - dp p2 + 1 = (pz + 1)2
Finally, we apply the shifting property (Theor. 4.4) to obtain
-21
.P{te
2
(p + 2) 2 - 1
p + 4p + 3
cos t;p} = [(p + 2)2 + tf = (pz + 4p + 5i
L"" (I:
L""
[f(t)/tl dt
e-pt
*Recai! from Theor. 4.1 that the integral in (4.13) is uniformly convergent in a halfplane. Thus, differentiation under the integral sign is permitted as longas each new integral
produced in this fashion also converges uniformly.
177
Y:{f(t)/t;p}
2:::
= L'>O F(u) du
Solution:
= {' sin u du
Jo
= 1/(p2 +
Y:{sin t;p}
1)
y;
{si~ t ;p} =
f' d:
u2
= tan -I
fi {Si(t);p} = Y: {
fi
}
1 j""
Jo(' f(u)
-;;-du;p = p
F(s) ds
P
{f
oo
f(u)
}
---;;du;p
p1 JofP F(s) ds
(4.15)
(4.16)
and
Y: {
("" f( )
Jo :
du;p
f""
= pJo
F(s) ds
(4.17)
the last of which is a simple consequence of the first two. The verification
of these properties is left to the exercises.
4.3.4
Periodic Functions
178
= [l/(1
- e-pT))
e-pt f(t) dt
e-pt f(t) dt +
J;
e-pt f(t) dt
r
=r
r
.P{f(t);p} =
e-pt f(t) dt +
= [1/(1 -
17'
Since T
Solution:
179
21r, we compute
z,.
e-P1 f(t) dt
o
i"'
o
1 + e-,.p
= --,:--2
e-p1 sin t dt
+1
and therefore
EXERCISES 4.3
1. Prove the linearity property (Theor. 4.3)
:J!{C1/(t)
Czg(t);p}
= C 1F(p) + CzG(p)
= aF(ap + b),
a> O
3. Given that :J!{(sin t)/t;p} = tan- (1/p), find the Laplace transform
of (sin at)/t where a > O.
1
4. Use Probs. 2 and 3 to find the Laplace transform of e31 (sin 5t)/t.
7. f(t)
= 3t4e51
= e41 cosh 5t
9. f(t)
= {;:
5. f(t)
ll. f(t)
13. f(t)
O,
O<t<1r
t > 1T
!,
O<t<l
l,
3 - t,
l<t<2
2<t<3
t>3
O,
e-"214 du
15. f(t)
17. f(t)
= t2 sin kt
= (sinh t)/t
19. f(t)
~:: < 1
= {sin t,
=
21. If .,<t'{f(t);p}
~: ~ < 3
10. f(t)
= {: !_1,
12. f(t)
{C?S
t,
sm t,
16. f(t)
18. f(t)
5te 31 sin2t
20. f(t)
= (e-a1
4
"
0<t<1T
t > 1T
cosh 5u du
- e-b1)/t
180
24
=!1
,p
.
25. 2{tSI(t);p}
tan- 1p
p
+ a2
+ bz
= - -2 -
26. 2 { e;;p} =
p
2 og 2
p
erfc
p p + 1)
J~ ea 1
24
C~),
a> O
27. 2 { e -aVr.;p} = -1 - -a
p
2p
J'T-e
r
p
a2f4 P
erfc ( -a-) ,
2yp
a>O
29.
=p P
=; J:
30. 2:{Jo
: du;p
foo F(s) ds
F(s) ds
1(
00
= pJo F(s) ds
oo J(u)
ioo F(s) ds
-du =
io
fi cosud
00
t>O
U,
E (t) =
1
>o.
181
Show that
;t'{Ln(t};p}
34. If f(t)
= (1/t)g'(t),
= ~ (p
1) n
show that
= Loo sG(s) ds
F(p)
b ~O, p >O
=o
= ! j~ e-b2f4p
2
ti
37. f(t)
38. J'"'(t)
= { - 11 ',
lsin
40. f(t)
= {631,'
f(t
o< t <c
c< t < 2c
o< t < 2
2<t<4
f(t
+ 2c) = Jr.(t)
f(2t
2c) = f(t)
+ 4) = f(t)
182
4.4
Example 4.11: Find the Laplace transform of erf(t), which is the error
function discussed in Sec. 1.3.
Solution: From the defining integral of the Laplace transform and
that of the error function, we have
.'l'{erf(t);p}
= L= e-pi erf(t) dt
2 r
= Jor= e-pi y;
Jo e-xz dx dt
.'l'{erf(t);p}
2 r=
r=
= y;
Jo e-xz Jx e-pi dt dx
= _2_
r= e-(xZ+px) dx
pyi;Jo
= _2_ eP2j4
pyi;
r= e
-(x+p/2)2
dx
Jo
t
Figure 4.3 Region of integration
183
+ px = (x + p/2) 2
.2C ePz/4
py7T
lI 4
=x +
f""
Jp/2
p/2 leads to
e -uz du
Re(p) >O
y; -2ab
e -aZxZ-bZfxZ dx=-e
2a
'
a> O, b ~O,
(4.18)
2:
O, Re(p) >O
(4.19)
184
2 loo e-x2
.P{erfc(l/yt);p} = ~r
v 1T
i""
e-pt
dt dx
l/x2
p~ Loo e-x2-pfx2 dx
Re(p) >O
.P{erfc(a/yt);p}
= (ljp)e- 2aYii,
Re(p)>O
(4.20)
X=
Vt
185
Solution:
= (- 1t(yt)2n+t
Slfi
Vt = n~O
= (- 1 1n+I/2
(2n + 1)!
= n~O
(2n + 1)!
-1r
(L){ r.. } - ~ (
(L){ n+l/2. }
sm yt,p - n~o(2n + l)!oL t
,p
oL
"" (-1r
= ~o (2n +
f(n + 3/2)
1)!
pn+3/2
= 22n+2(n +
f(n + 3/2)
1)!
v'rr =
(2n + 1)!
22n+ln!
v'rr
J1i- L --1t, ( -1 )n
=
n.
n=O
4p
4.4.1
Re(p) >O
(2y't)
L _..:.(-____.1)n:.___
[n+v/2
lv/2
n=O
n!f(n +
i (-tr tn+v
n=O
n!f(n +
+ 1)
+ 1)
186
(f){
aL
v/2
r.) } _
lv 2 V t ,p -
~
( -l)n
:=-o
n!f(n + v +
1)
5f{ n+v. }
t
,p
"" <-W
= ~O n!pn+P+ I
1 (-l)"(l)n
p
oo
= pv+t~0 ~
from which we deduce
Re(p) >O
The scaling property, Eq. (4.8), allows us to generalize the result of
Exam. 4.15 to
5f{tv12Jv(2yat);p}
= O then yields
5f{Jo(2'0);p} = (1/ p)e-afp,
a>O,Re(p)>O
(4.21)
Example 4.16:
Solution:
(4.22)
we have
and thus
5f{tvJ (1)
v
;p
~
00
n~O n!f(n
(-
CO{
( - l)n
= :=-o n!f(n +
1)22n+v
aL
2n + 2v
;p
1r f(2n + 2v + 1)
+
1)22n+vp2n+Zv+l
f(2n + 2v + 1)
f(n + v + 1)
v;1 2
2n+2v
f(n
+ 1/2)
2v
y; p2v+l n=O
p2
187
+ x)-a
(1
n=O
(-a)
=~
x"
n=O
n.f(a)
lxl < 1
= (-1Yf(n
+a)
n!f(a)
or
v
;t'{t Jv(t);p}
When v
2v f(v + 1/2)
C 2
v1r(p
+ 1Y
+ 1/2'
= 1/VJT+l,
Re(p) > 1
(4.23)
Also, by using Eq. (4.11) together with the relation J(t) == -J1(t), we
have
2{J(t);p} == P 2{Jo(t);p} - Jo(O)
or
= <PIVP2 + o -
-5t{J,(t);p}
= (Vp 2 +
1 - p)/VJT+l,
Re(p) > 1
(4.24)
EXERCISES 4.4
ln Probs. 1-8, verify the Laplace transform relation involving error
functions.
= (lj p)e
2{erf(I/yt);p} = (1/ p)(l
02 2
P
1. 5t{erf(t/2a);p}
2.
3. 2{erf(yt);p}
erfc(ap),
a >O
- e- 2v'P)
= 1/ PYP+l
4. 5t{ea'erf(Vaf);p}
= J~
(1
-),
P p- a
a> O
188
5 .X{ea'erfc(yrat);p}
ypp(V P1 + Vaa) ,
a> O
r.
3p + 8
6 .X{t erf(2 v t);p} = p 2 (p + )312
4
4
21
=V
8. .x{eb<bt+alerfc(byt + _a_);p}
2yt
p+a
a> O
= --==-e--a=v_p_
VP<VP + b)
ln Probs. 9-15, use infinite series to derive the given Laplace transform
relation.
9. .X{(sin2t)/t;p}
= (1/4)log(l + 4/ p 2)
Jp
= tan- 1 a/p,
12. .X{erfc(l/yt);p}
a> O
e-afp,
a>O
a>O
v> -1/2
ln Probs. 16-22, verify the Laplace transform relation involving Bessel
functions.
16. !t{e-a'J0 (bt);p} =
1/Vp 2 +
2ap
= pj(p 2 + a 2) 312,
18 . .X{te-'J0(t);p} = (p - l)/(p2 19 .;l{tJI(t);p} = 1/(p2 + 1)3/2
17 .X{tJ0(at);p}
+ a 2 + b2 ,
a> O
2p + 2) 3/ 2
(2n)!
__
20 .X{fJn(at);p} = -2n'(p2 + a2) n 1/2,
n.
b >O
a> O, n
= 1,2,3, ...
22. 2{Ju(t);p} =
<Vll+l -
189
p)" ;\(il+t
sinG tan -I
~)
V"PVl+4
(b) .5t{J0(t) cos t;p}
1
_ _ cosG tan -I ;
VPVP + 4
= .P{f(t);p},
show that
= F(log p)
p
f"
27.
28.
Jo(2Vxf)cos x dx
= sin t
29. Use the result of Prob. 31 in Exer. 4.3 to deduce the value of
~ Ln~t) = e Jo(2y/t)
n=O
n.
where C(x) and S(x) are the Fresnel integrais (see Sec. 1.3.2).
32. Using the sifting property of the impulse function (see Sec. 1.5.2),
show that
(a) .P{cS(t);p} = 1
(b) 2{8(t - a);p}
= e-ap,
a> O
190
4.5
Thus far we have concerned ourselves only with the problem of finding
the transform function F(p), given the function f(t). However, the use
of Laplace transforms is effective in applications only if we can also
solve the inverse problem of finding f(t), given the function F(p). ln
symbols, we write the inversion formula
f(t)
= ,;e- 1{F(p);t}
(4.25)
Theorem 4.11.
lim F(p) =O
IPI->oo
The proof of Theor. 4.11 follows that of Theor. 4.1 and is left to the
exericises (see Prob. 37 in Exer. 4.5). The real significance of Theor.
4.11 is that if F(p) is any function for which lim F(p) =/= O, then it does
IPI-+oo
J: n(u) du
O for all t.
191
directly from existing tables of transforms (see Appendix C). For instance,
in Exams. 4.1 and 4.2 we derived the transform relations
= pj(p2 + a 2)
;e-t
{-1-;t} =
p-a
eat
and
;e-t {
pz
+ az
;t} = cos
at
(4.26)
and
::e- 1{F(p
Example 4.17:
Solution:
- a);t}
(4.27)
+ 3) - 8
+ 3)2 + 4
:;e-1 {
p2
p - 5
+ 6p + 13 ;t
}
=
:;e-1 { (p + 3) - 8 }
(p + 3)2 + 4 ;t
= e-Jr;e-1 {
t}
p - 8
p2 + 4'
= e- 3t [ 5f- I { pz :
;t} -
4 ;t'- I
;e-t {
Pp + 6p
2
+ 13
pz :
;t}]
192
4.5 .1
Partia[ Fractions
= R(p)/Q(p)
F(p)
Solution:
Example 4.18:
1) (p
1); t}
2
=~+Bp+C
(p + 1)(p2 + 1)
p + 1
p2 + 1
and clearing fractions yields
2
= A(p2 +
1) + (Bp + C)(p + 1)
Setting p = - 1, we find A
and p 0 gives the equations
O= A+ B,
2 =A+ C,
from which we deduce B
If:-1 {
. (p
-1 and C
1. Thus, we find
2 + r}- {-1-r}
+
l)(p2
1) ,
It-1
= e -t
Let us write
1,
- 5f- I
Solution:
p2 :
cos t
;t} + If-
+ sin t
I {
p2
~ 1 ;t}
193
+ 1 = Ap(p + 2) 3 + B(p + 2) 3
+ Cp 2(p +
2)
2 + Dp 2(p +
= 1/16,
-A
2)
+ Ep 2
= 1/8,
x-
{p2rp ++ 2)3;r}
= -/6 x1
+ 16 X
-1 {
= _ _!_ + ! 1 + I.e-21
16
41 X
! 12 e-21
+ 2 ;t -
16
-1 {
(p
+ 2)3 ;t
x- 1{log (1
Find
+ 1/p2); t}
=-
2 = -2(!- _P_)
+
p(p2 + 1)
p2
= - F'(p)
x- 1{F'(p);t} =
tf(t)
x-
{F'(p);t} =
-tf(t)
194
= 2(1
f(t)
- cos t)/t
F(p) = p
a1
a2
a3
+ p2 + p3 + p4 +
(4.28)
12
f(t)
= a0 + a1t + a2 21 + a3 3! + ...
(4.29)
Pe
-1
IP
00
= P~o
<-1t
n!
Pn
00
<-1r
= ~o n! Pn+ I
::t-1
_1 e-lfp;t }
p
1t ::t-1 { --;t
1 }
2: __
(-
n!
n=O
00
= n=O
2:
pn+l
-1t ln
(n!)
4.5.3
{~ e-l/p;t} = lo(2Vt)
Convolution Theorem
195
and G(p) are known. This sarne situation arises when one of them represents
an arbitrary function - typically an input to some physical system. The
inversion of the product F(p )G(p) can then be obtained by application
of what we call the convolution theorem.
ln order to derive the convolution theorem, let us begin by writing
the product of the transforms of f(t) and g(t) as the iterated integral
F(p)G(p)
t - u then leads to
- u)g(u) dt du
(4.30)
u)g(u) du dt
u)g(u) du dt
We now define the Laplace convolution integral of f(t) and g(t), i.e.,
(f * g)(t)
(4.31)
f(t - u)g(u) du
t
Figure 4.5 Region of integration
196
(4.32)
Lt(t-
u)g(u)du
(4.33)
<f* g)(t)
Once again we point out that the conditions stated in Theor. 4.12 are
more stringent than required for the validity of (4.33). That is, there exist
functions satisfying (4.33) that are not piecewise continuous on t ~ O.
Some examples illustrating this point are included in the exercises (see
also Exam. 4.23).
Example 4.22:
Find
.x- 1 {1/p2(p 2 +
e);t}.
and
.x-l
<f* g)(t)
(t - u)(l/k) sin ku du
197
=-
f(v)g(t - v)dv
= J: g(t
- v)f(v)dv
= (g * f)(t)
simply f * g = g * f. From definition,
(f* g)(t)
* g = CU* g)
(4.34)
it is clear
(4.35)
(4.36)
= ~ {I
cos tx dx
x2
yl -
7T Jo
= ~ r:::r-71
+
yp
1 . ;t }
I
yp-z
yp+z
=.!.. t
7T Jo
1
=-
i/
u-1/2
e-iu
ei(t -
(t- u)-1/2
2u)
7T 0 yu(t - u)
du
ei(l- u)
du
198
By letting u
ii
7T
eit(l-2v)
y'v(l - v)
= 1 - 2v, we have
II itx
=- I y'l""=7
1
Jo(t) = 7T
7T
-I
-t
dv
~dx
V 1 -X
cos tx dx+-i
7T
-I
sin tx dx
y't - x2
Jo(t) = 3_
7T
cos tx dx
Jo v'1
-r
EXERCISES 4.5
ln Probs. 1-10, determine the inverse Laplace transform using the table
in Appendix C and various operational properties.
1
2. F(p) =
1. F(p) = 2p + 3
3. F(p)
2p
+5
1
6 p + 10
+ 13
6. F(p)
= 3p 2 + 4 p + 8
2p + 3
4
5
p + p +
8. F(p)
= (p + 2)4
p
6
e-sp
9. F(p)
= p2 -
- 3)s
p -
7. F(p) = 4
4. F(p)
= (p
5. F(p) =
3p + 7
= (p
- 3t
5p- 2
p2
+ 1)
11. F(p)
13. F(p)
= (p + Z)3
p(p
p2
12. F(p)
= (p
- l)(p
3p- 2
14. F(p) = p3(p2 + 4)
+ 2)(p + 4)
15. F(p)
p+l
- 4p)(p + 5)z
= (pz
16. F(p)
17. F(p)
19
= (pz + az)(pz +
18. F(p)
bz)
199
= - 4- -
p - 1
4p 2 - 16
p3(p + 2)z
i- 3
3p 2 - 6p + 7
20. F(p) = (pz - 2p + 5)z
21. Given that ::r 1{F(p);t} = f(t), show for constants, a, b, and k that
(a) 5/!- 1 {F(kp);t} = (1/k)f(t/k),
k >O
(b) 5f- 1{F(ap + b);t} = (l/a)e-bt/a f(t/a),
a> O
22. Given that f(t) = t sin t is the inverse Laplace transform of F(p)
= 2p/(p2 + 1)2 , use Theor. 4.7 to evaluate
2 -1 {l/(pz + l)z;t}
ln Probs. 23-28, use infinite series to find the inverse Laplace transform
of the given function.
23. F(p)
25. F(p)
21. F(p)
log(l + 1/ p 2)
24. F(p)
= log--
26. F(p)
= 1og--b
= 1/pyp + 4
log(l
1/p)
p+1
p- 1
p-a
p-
ln Probs. 29-36, use the convolution theorem to find the inverse Laplace
transform of the given function.
29. 1/
PY p + 4
32. 1/ p 2(p
1)
33. 1/yp(p - 1)
2
36. 1/(yp - I)
35. 1/(p + 1) (p + 4)
37. Prove Theor. 4.1 1.
38. Show that
2
(a) I * 1 * 1 == t /2
5
(b) t * t * t = 1 /5!
(c)t
m-1
(m -
*I=
l)!n!tm+"
)I
m + n.
;m,n=1,2,3, ...
Lux- (t 1
> O,
F(p)
= f(x)f(y)/ px+\
200
J
1
lf- 1 (I - u)Y- 1 du
= f(x)f(y)/f(x + y),
>O, y >O
J:
(a)
J 0 (t - u)J0 (u) du
= sin t
(d)
I:
= J (t)
0
-'1T 1,v
I
COS
ezau
u(t - u)
du
= eat lo(at)
!t-t
4.6
fc+ico
27TZ
c-ico
= -.
ept
F(p) dp
(4.38)
where f(t) and F(p) are Laplace transform pairs. ln order to provide a
201
more rigorous derivation of this formula, let us assume that both f(t)
andf'(t) are continuous functions on t ~O and thatf(t) is O(ec1). Then,
based on Theors. 4.1 and 4.2, we know that the Laplace integral
F(p)
(4.39)
f:i~ eP
F(p) dp
(4.40)
Our goal at this point is to show that by allowing ~ oo, we can derive
Eq. (4.38) from Eq. (4.40). Because ofthe requirements we have imposed
on f(t), we can interpret (4.40) as an iterated integral and interchange
the order of integration. Doing so yields
c+iA ept F(p) dp
f c-iA
= loo f(u)
fc+iA eP<t-u) dp du
c-iA
u) du
t-u
-t
27TZ
e~'1
= t + x.
c- iA
= f(t),
>o
(4.41)
1 fc+ioo
= 27Ti
c-ioo
ePI F(p) dp
202
Laplace transform is F(p) for Re(p) > c. Also, the function f(t) is
O(ec"') and is continuous everywhere, and f(t) = O when t :5 O.
The conditions stated in Theor. 4.13 are quite severe in that they
exclude, for example, the simple function 1/p, which is O(p-~ where
k = 1. Also, these conditions are not satisfied by transforms of functions
that are discontinuous or for which f(O) =/= O. Nonetheless, by using a
Fourier integral theorem and stating conditions on the functionf(t) rather
than on F(p), these conditions can be relaxed so that the inversion
integral formula is valid in nearly all practical cases of interest to us. *
We do note that the conditions stated in Theor. 4.13 ensure the existence
of the inverse Laplace transform of F(p), and moreover, ensure that the
inverse Laplace transform f(t) is that function for which 5t'{f(t);p} =
F(p).
~ fi=
2m
'''2 dp
-i=
= _1 I""
21T
eiry e- y2 dy
-=
-.
fi=
eP 1'
dp
-i=
= -1- Bf{e-Y2;t}
y-2;
1
-t /4
=--=e
2
(4.42)
2y'7T
.;e{ 2~ e-'
214
;p}
= ~ '
erfc(p)
=/=
F(p)
(4.43)
What we are illustrating here is that the inversion integral of F(p) may
exist without representing the function f(t) whose Laplace transform is
F(p). The reason it happens in this particular case is because the function
2
F(p) = ' is not O(p-~ for any positive k. Thus we have violated a
condition of Theor. 4.13.
* See, for instance, R. V. Churchill, Operational Mathematics, New York: McGrawHill, 1972, Chap. 6.
203
The complex inversion integral for the inverse Laplace transform can
often be evaluated quite readily through use of the theory of residues.
Let us suppose that F(p) is an analytic function in the complex p-plane,
except for a finite number of isolated singularities a~> a 2 , , aN. By
integrating the function
eP' F(p)
around the closed contour shown in Fig. 4.6, where c and R are selected
such that no singularity ak lies to the right of Re(p) = c and the radius
R of the circular are CR is large enough to enclose all singularities of
F(p), we have the result
J~ eP'F(p )dp
N
= 27Ti 2:
Res{eP'F(p);ak}
(4.44)
k=l
cR
eP'F(p )dp
={
JnJ
+J
JKL
fu
eP' F(p)dp
lm(p)
J
Re(p)
Re(p) = c
L
Figure 4.6 Contour of integration
(4.45)
204
lim
(4.46)
R-oo JJKL
However, sinee the ares BJ and LA do not lie in the seeond and/ or third
quadrants, the vanishing of the remaining two integrais in (4.45) in the
limit must be treated separately. Under the assumption of Theor. 4.13,
F(p) is of order O(p- k) and therefore it follows that there exists a eonstant
M sueh that
IF(p)l
<M/R\
M >O, k > 1 on CR
We now eonsider
II1I = I
::s
JBJ
epr F(p)dp
iepriiF(p )iidPi
JBJ iepriidPi
::s (M/Rk)
Along the are BJ we set p = Re;8 ,
()0
idpi = RdO
and therefore
/ I
1I
::s __
Rk-1
= Tr/2
- () and c/Jo
eRr cos
e dO
Bo
::5
where c/J
l.,.;z
Rk-1
Jo
'I'
= sin- 1 (e/R).
c/R
we deduee that
M
rtf>o
As R
ect
dcfJ
Rk-1
R>>c
to find
lim
R.....~J>oo
liii ::5
lim (Me/R")
R--+oo
ect
=o
(4.47)
205
Hence, the integral along the are BJ vanishes in the limit R~ oo, Similar
arguments can be used to show that the integral in (4.45) along LA also
vanishes as R ~ oo. Based on the above results, we deduce that
lim { epr F(p) dp = O
R->oo
(4.48)
JcR
f(t)
= ;e- {F(p);t} =
1
L Res{eP F(p);ak}
1
(4.49)
k=l
lf F(p) = R(p)/Q(p), e.g., when both R(p) and Q(p) are polynomials,
then Eq. (4.49) can be expressed in more explicit forms. For instance,
if Q(p) has simple zeros at p = a~> a 2 , ... , aN, then (4.49) becomes
f(t)
= ;e-I {R(p)
t} = k=li Q'(ak)
R(ak) eakr
(4.50)
Q(p)'
;t' _ 1 {
}
p
} = Res { peP ;w
+ Res
2 ;t
2
2
2
p+a
p+a
}
peP ; - za
2
2
p+a
+ e -iat)
!(eiat
cos at
We can formally extend the result (4.49) to the case in which F(p)
has infinitely many isolated singularities by allowing N to become infinite.
ln this case, we obtain
f(t)
;e-I {F(p);t}
L Res{eP F(p);ak}
1
k=l
coshxyp
.. I '
p cosh v p
o< X < 1
(4.51)
206
O is a branch point of
F(p) because of the presence of yp. That this is not the case can
be seen by writing
+ (xyp)4/4! + ...
P cosh yp
p[l + (ypi/2! + (y p) 4/4! + ... ]
1 + x 2p/2! + x4p 2/4! + .. .
= p(l + p/2! + p 2/4! + ... )
from which it is clear that p = O is not a branch point but a simple
F(p) =
coshxyp
1 + (xyp) 2/2!
.= _
__:_~~-__;_=-=-__;_--
pole of F(p). ln addition, there are infinitely many other simple poles
given by the roots of the transcendental equation
cosh
YP =O
YP = (n -
1/2)7ri, n
= 1,2,3, ...
or
p
= an = -
(n - 1/2)277"2 , n
= O and
= 1,2,3, ...
p = an yields
cosh YP p=o
=1
and
Res{eP' F(p);an}
=
=
2YP smh YP
1
I
p=a.
xyp ;t}
!!_ 1 { cosh
pcosh yp
=1+-
L (2n-l)n
cos[(n - 1)
oo
1/2)7rx]e-<n-t/2)2.r2t
1T' n=l
4.6.1
Multivalued Functions
207
Fig. 4.6 to exclude the branch point(s) and then take the limit as the
radius of the small circle around the branch point(s) tends to zero. Let
us illustrate the procedure with an example.
Example 4.26: Find the inverse Laplace transform of F(p) == e-aYP,
a> O.
Solution: The function F(p) has a branch point at the origin p
O. Hence we integrate the function
epr
F(p)
+ r ept-avP dp _
J~
JR e-xt+iaVx dx +
p
r ept-avP dp == 0
J~
lt can be shown that the integrais along C~> CP, and C4 all tend
to zero as p ~O and R ~ oo. Hence, it follows that
lm(p)
208
Ii""
=-
7T
e -xt sin(ay'x) dx
= u 2 further leads to
.P- 1{e-aYP;t}
= -2
7T
3P{ue-uzt;a} =
-i! [
(2t)-l/2 e-a2/4t]
ia(2t)-3/2 e-a2f4t
However, for a > O the Fourier and Fourier sine transforms are
related by
3'{ue-u2t;a} = i 3's{ue-uzr;a}
and so we finally deduce the intended result*
a> O
EXERCISES 4.6
ln Probs. 1-12, use the complex inversion formula to evaluate the inverse
Laplace transform of the given function having isolated singularities.
1
p
1. F(p) = (p + l)(p _ 3)2
2. F(p) = (p + 1)3(p _ ti
p2
p2
3. F(p)
= (p2 + 4i
S. F(p)
= p 2 cosh p
7. F(p) =
sinhxy'p
.
r:,
smh vP
O<x<l
4
4. F(p)
=~
6. F(p)
= co~h xp '
+4
psmhp
O< x < 1
sinhxy'p
= r:
y' ,
v p cosh p
O<x<I
8. F(p)
* See aJso Prob. 22(a) for another derivation of the Fourier sine transform
.'9is{ue-\a}.
9. F(p)
11. F(p)
= p(' +
1
= 2p -
1)
1
psm p
12. F(p)
. h
1
p cosh
209
O<x< 1
vp
t} = ~ ~
2:-1 {/o(ryp)
Jo(byp)'
knlo(knr)
b n= 1 J1(knb)
e-k~r
O, n
1,2,3, ...
2:-1 {
J]/-avP;t} = J1T/-a
a>
/4t,
17. F(p)
=c
v p(p-
2
a)
18. F(p)
=c
20. F(p)
= log(l + 1I p 2)
VP +a
= L~ e-bu
cos tu du,
t::::: O, b >O
= ! J~ e-r2f4b
2
Jo
= ~JTT e- 02121
t
2t
'
t>O
ZlO
L"" (1ju)e-
1 2
sin au du
= (Tr/2) etf(a/2Vt)
= O, show that
a> O
a>O
24. From the defining integral of the Laplace transform, show that
(a)
::e{--v:;i +
1
-
e'etf(Vt);p}
YP
p-1
::t{ 1 yt
+ ;p } = Vr;p 1
TrePerfc(y p)
::e L~ 1 ;p} =
lP -::e{yr(t
+ 1)
;p}
4. 7 Additional Topics
ln earlier sections we have found that the transform function F(p) has
certain useful properties, most of which can be established by relying
on the Laplace transform integral for the definition of F(p). For example,
under appropriate conditions on the inverse transformf(t), we have found
that F(p) is analytic (Theor. 4.2) and that F(p) ~O as IPI ~ = (Theor.
4.11). ln addition, certain operational properties of this function were
developed in Sec. 4.3. Here we will discuss briefty further asymptotic
properties of F(p) and extend the definition of the Laplace transform to
4.7/Additional Topics
211
functions defined on the entire realline rather than only on the positive
realline.
4.7.1.
= 1,2,3, ...
a0
F(p) = p
a1
2a2
k!ak
+ 2p + -p3 + + ~
p
!PI-+=
= ao = /{0)
(4.52)
Not only is this particular property true of all polynomials, but applies
to a relatively large class of functions.
Theorem 4.14 (lnitial Value Theorem). Iff(t) is piecewise continuous on
t ~ O and is O(ec01 ), then its Laplace transform F(p) satisfies
lim pF(p) = limf(t) = /(0)
,_.o
IPI-+=
Proof: We will present the proof only for the stronger case where
f(t) is continuous on t ~ O.
By Eq. (4.11) in Sec. 4.3,
.P{f'(t);p}
pF(p) - /(0)
f"" e -p'f'(t) dt = O
Jo
IPI-+oo
=O
or
lim pF(p)
IPI-+oo
= limf(t) = /(0)
t->0
212
L"" f'(t) dt
exists, then the transform F(p) satisfies
lim pF(p)
p-+0
t-+a:J
Proof: As in the proof of Theor. 4.14, we will present the proof here
only for the case when f(t) is continuous on t ;;::= O.
Again we start with the relation
.T{f'(t);p}
- /(0)
Noting that
lim
p .....
r=
oJo
e-pt
f'(t) dt
r= f'(t) dt
Jo
= lim
t-+oo
Jor' f'(u) du
= lim f(t)
- /(0)
/-+00
we deduce that
limf(t) - /(0)
= lim pF(p)
- /(0)
p-+0
(-+00
= f(oo)
p-+0
Example 4.27:
defined by
Ci(t)
t >O
00
= cos t
4.7/Additional Topics
213
= P2 +
dp [pF(p) - f(O)]
p2: 1 dp
or
pF(p)
= (1/2) log(l +
p 2)
+C
= limf(t) = O
p-+0
t-+oo
Theorem 4.16 (Watson's lemma). If f(t) is O(ec01 ) and if, in some neighborhood of t = O, the function f(t) has the Maclaurin series expansion
00
f(t)
L
n=O
a~ tn,
n.
ltl <R
F(p)-
L
n=O
~:I'
Example 4.28:
*For a rigorous proof of Watson's lemma, see I. N. Sneddon, The Use of Integral
Transforms, New York: McGraw-Hill, 1972, pp. 188-190.
214
2
erf(t) = -
v;
(-1tt2n+l
2:oo _:__;.____
n=O
n!(2n
+ 1)
~ ~ ( -1t(2n)!
. . 1- LJ
y'Tr
n=O
1
n.p
2n+2
__
1 ~ ( -1t f(n + 1/2)
1rp n=O
(p/2)2n+i
where we have used the duplication formula of the gamma function.
Finally, setting x = p /2, we get the desired asymptotic expansion
erfc(x)-
4.7.2
e-x2
1r
n=O
-l)n
r~n+~
1/2)'
lxl--'
00
lf the function f(t) is defined over the entire real axis, we may consider
the integral
2 +{f(t);p}
(4.53)
J~ e-ist f(t) dt
(4.54)
4.7/Additional Topics
215
(4.55)
The second transform in (4.55) exists in the plane Re(p) > c~> where c 1
is the abscissa of convergence. Similarly, the first transform will exist
in the plane Re(- p) > - c2 , or Re(p) < c2 , where - c2 is the abscissa
of convergence in this case. The two-sided Laplace transform of f(t)
exists only if these two half-planes overlap. That is, it will exist only in
the strip c 1 < Re(p) < c2 lf c2 < c~> the two-sided Laplace transform
does not exist, and if c2 = c 1 , the strip contracts to the vertical line
Re(p) = c 1 Finally, if c2 = c 1 = O, the two-sided Laplace transform
is then actually a Fourier transform [see Eq. (4.54)].
Example 4.29: Find the two-sided Laplace transform of e-
12
oL
+ { e -t2 ;p}
= I""
-oo
e -pt e -t2 dt
The inversion theorem for the two-sided Laplace transform has the
sarne form as for the ordinary Laplace transform, except that for the
two-sided transform the interval of convergence must also be established
in order to uniquely establish the inverse transform. For example, consider
the two functions
f(t) = {
0
:_21
-t
e
-e ,
t<O
t>O
and
g(t) =
{e=:;
e '
t <o
t >o
216
Re(p) > -1
and
f(t)
= - 7Tl.
fc+ioo
.
eP
F +(p) dp,
c> -1
c-100
and
1
g(t)
= - 7Tl.
2
fc+ioo
.
1
eP G+(P)
dp,
-2<c< -1
c-100
EXERCISES 4. 7
1. Verify the initial value theorem for the following functions:
(a) f(t) = 5 + 4 cos 2t
(b) f(t) = (3t - 2)3
(c) f(t) = erf(t)
2. Verify the final value theorem for the following functions:
(a) f(t) = 3 + e- 21(cos t + sin t)
(b) f(t) = 1 + e_,z
(c) f(t) = erfc(l/y't)
ln Probs. 3 and 4, use the technique of Exam. 4.27 to find the Laplace
transform of the given function.
3. Si(t)
r sinu u du,
Jo
t >O
4. Ei(t)
J,
t >O
O, f"(t) is piecewise
4.7/Additional Topics
217
= f'(O)
= (""
e-px
dx
Jo 1 + x2
(a) show that its inverse Laplace transform leads to the Maclaurin
expansion
f(t)
~ ( -1r<2n)!
f:o p2n+l '
F( ) -
IPI ~ oo
ln Probs. 8-10, use the technique of Prob. 7 to derive the given asymptotic
formula.
oo
8.
e-px
- -dx-L.J
o 1+ X
n=O
-lt n!
p
n+l
00
(""
9. j<
e-px
COS X
dx-
-lt
L 2n+i
(
n=O
1(""
where (a)n
= f(a + n)/f(a),
= 0,1,2, ....
ln Probs. 11-14, find the two-sided Laplace transform of the given function
and state the strip of convergence.
e"1
11. .f(/) = { a/
e .
t<O
t>O
12 .ftt>
= e -ltl
14. f(t)
=e-a't'
a<h
sin t
13. .f(t) = t-
15. By integrating the function F + (p) = eP 214a along the imaginary axis
of the complex p-plane, find the inverse transform of this two-sided
Laplace transform.
5
Applications Involving
Laplace Transforms
5.1 lntroduction
Like the Fourier transform, the Laplace transform is used in a variety
ofapplications. Perhaps the most common usage ofthe Laplace transform
is in the solution of initial value problems. However, there are other
situations for which the properties of the Laplace transform are also
very useful, such as in the evaluation of certain integrais and in the
solution of certain integral equations. ln this chapter we will briefly
discuss applications of the Laplace transform in ali of the above named
areas.
5.2/Evaluating Integrais
219
r= sin t dt
Jo
sin t d
t
e 1-- t
1=
w{sin t ;p } =
sin-tdt = tan _ 1 -1
e -pr t
o
t
p
.z, - -
By setting p
= O and
I=
Ii{ si~
==~
;p = 1} = tan -I 1 =
r= cos tx dx
t>
Jo x2 + 1
Solution: Let us define the integral by f(t) and take the Laplace
transform with respect to t. This action leads to
F(p)
= L"" (xz +
l)~xz + pz) dx
r=(xz + 1 -
= pz - 1 Jo
=
p2
xz + pz dx
~ 1(~ -2:)
7T/2
=p+
1'
= (7T/2)e-
t >O
* We might also recognize the integralf(t) as a multiple ofthe Fourier cosine transform
of the function l/(x 2 + 1).
220
EXERCISES 5.2
ln Probs. 1-10, use known Laplace transforms or transform properties
to evaluate the given integral.
1. L=te- 21cos t dt
3. Joo sinh t dt
-ao
S.
J:
4.
oo
-31
dt
-61
- e
6. LootJo(t) dt
J0 (t) dt
8. L=e- 1erf(y/t) dt
10. L=xe-x2erfc(x) dx
t>
15.
t
-
>o
12.
rjx2 )dx,
i=Xsin txdx
2
1 '
O X +
14 ioosin txdX,
O
LOOS0Xdx,
t>O
16.
t >o
J~ooCOS tx
t>O
dx,
t >o
J""
sin x dx =
18.
r=
Jo
x 3 dx
-oo
X COS
19 r=sin X dx
. Jo xa
20 r=cos X dx
Jo xa
J:!
2
1T
3 3f(l/3)
1T
2f(a)sin 0!7T/2'
7T
2f(a)cos 0!7T /2'
O< a< 1
O<a<1
5.3/Solution of ODEs
221
So/ution of ODEs
5.3
Solution:
= 2, y'(O) = 6
p 2 Y(p) - 2p - 6
and
2{t2e3';p}
= 2!/(p
- 3) 3
we find that the given initial value problem is transformed into the
algebraic equation
[p 2 Y(p) - 2p - 6] - 6[pY(p) - 2]
+ 9Y(p) = 2/(p - 3) 3
or
(p 2
6p + 9)Y(p)
= 2(p
- 3) + 2/(p -
222
=P
Y(p)
- 3
(p - 3)5
= .~r'{Y(p);t} =
:e-'{p ~
;t} + ..ce-'Lp
~ 3>5;1}
= 2e 31 + (1/12)t4e31
y(t)
+ 4y = f(t),
where
f(t)
y(O)
{41,4,
= 1, y'(O) = O
:S;t:S;l
t> 1
= 4t[l
= 4t -
4
4
= -p2
- -p2
e-P
+ 4Y(p) = 2 p
2 e-p
p
with solution
4
Y(p) =
l + 4 + pz(pz + 4)
4
-p
- pz(pz + 4) e
Using a partial fraction expansion on the last two terms on the righthand side of the above expression, we obtain
Y(p)
_P_
p2
+4
_1_)
e-p
= cos 2t + t
5.3/Solution of ODEs
223
The above examples illustrate the basic procedure used in the method
of Laplace transforms. And although these problems can be solved by
other techniques, the Laplace transform method offers the advantage of
solving the problem directly without first producing the general solution
of the DE. Moreover, in the case of Exam. 5.4 we found the solution
without splitting the problem into two problems, one over each interval
where f(t) is defined, as required by more conventional methods. Forcing
functions of this nature, as well as discontinuous or impulsive ones, are
commonplace in circuit analysis problems and in certain problems involving
mechanical vibrations.
While in general the Laplace transform method works best on constantcoefficient equations, there are some variable-coefficient equations which
also lend themselves to the transform method. Consider the following
examples.
Example 5.5:
Solution:
we find
+ y' +
ty
o,
y(O)
1, y'(O)
=O
- dp[p Y(p) - p]
+ [pY(p)
- 1] - dp Y(p) =O
which reduces to
(p2
dY
1) dp
+ py
+1
= AJ0 (t)
where J0 (t) is the Bessel function of arder zero. The initial conditions
require the choice A = 1.
Example 5.6:
= o,
y(O) = 1, y'(O)
=O
224
[p Y(p) - p] - dp[pY(p) - 1]
+ Y(p) =O
or
dY
- - pY = -1
dp
= ep2/2(A
= eP212 [A
- y27T erf(pjy'2)]
e-p2f2
dp)
= f(t),
y(O)
= ko, y'(O) = kt
(5.1)
where a and b are known constants. The coefficient of y" has been set
to unity for mathematical convenience. If we introduce the Laplace
transforms .P{y(t);p} = Y(p) and .P{f(t);p} = F(p), then (5.1) reduces
to the algebraic equation
[p 2 Y(p) - pko - kt1 + a[pY(p) - ko] + bY(p) == F(p)
5.3/Solution of ODEs
225
or
(p2 + ap
+ b)Y(p) =
(p
+ a)k0 + k1 + F(p)
(5.2)
(p + a)k0 + k 1
F(p)
+ -=--;:__2
p2 + ap + b
p + ap + b
(5.3)
= 2-lr~2++ a~~o: :
;t} + 2-l{
p2 :;:; + b ;t }
~
YH(t)
(5.4)
~
yp(t)
+ by = O,
y(O) = k0 , y'(O) = k 1
(5.5)
= O, y'(O) = O
(5.6)
+ ay' + by = f(t),
y(O)
= 2- 1{
F(p)
p 2 + ap + b
;t}
= J:g(t - u)f(u) du
(5.7)
=2
-1{ p
+ ap +
b ;t }
(5.8)
226
= ..P-l{p2
which yields
g(t)
= !e
sin 2t
- u)
= J:g(t -
u)8(u) du
= g(t)
(5.9)
Hence, we see that the response function g(t) is actually the response
of the system to a unit impulse. For this reason, it is often called the
impulse response function of the system. According to (5.7), all other
solutions for general forcing functions f(t) are simply superpositions of
f(u) with the "fundamental solution" g(t - u).
Example 5.8: Use the impulse response function to finda general solution
of
y" - 2y'
Solution:
+ 5y = f(t),
y(O)
= Yo, y'(O) = Vo
tion is
g(t)
=! e
sin 2t
5.3/Solution of ODEs
y(t)
- 1)
(p -
+ (vo - Yo)
1l + 4
227
u)f(u) du
r} + ! Jo
'
y(t)
= e1[y0cos 2t + !(vo
- Yo)sin 2t]
+ !J:e -"sin2(t1
u)f(u)du
EXERCISES 5.3
ln Probs. 1-15, use the Laplace transform to solve the given initial value
problem.
= e cos
+ 2y' + y ==
1
1. y" - y
2. y"
3. y" - 4y'
t,
y(O)
3te-
y(O)
+ 4y = t,
2y'
5y
= e-r
= f(t)
10. y" + 4y
= sin t
== {
y(O)
sin t,
= t2e
4, y'(O) == 2
y(O)
,
1
O,
= 1,
y(O)
=5
y'(O) =
-3, y'(O)
y(O)
= e-
9. y" + 4y
y(O) = 1, y'(O) = O
O, y'(O)
y'(O)
~~s 41
~ ~ ~ :s
'IT
y(O)
O, y'(O)
y(O)
O, y"(O)
y(O)
=1
O, y'(O)
1
y(O) == O, y'(O)
O, y"(O) == -2
O, y'(O) = O, y"(O) = 1
y(O) == O, y'(O)
t,
3e- , y(O) = O
= 1
O, y'(O)
228
14. ty"
+ (t - l)y' + y = O,
y(O)
y(O)
=2
16. M =(D-a?
18. M
= D2 + 5
20. M = 4D2
8D
= D2 + 4D + 7
M = D2 - D - 2
19. M
+5
21.
22. The small motions y(t) of an undamped spring-mass system are govemed by the initial value problem
my" + ky
= f(t),
(1/w0}sin w0t,
23. Using the impulse response function given in Prob. 22, find the
response of the spring-mass system given that
(a} f(t) = P (constant)
(b} f(t) = P cos wt,
w i= w0
(c) f(t) = P cos w0t
24. When resstve forces are taken into account for the spring-mass
system in Prob. 22, the motions are called damped. ln such cases
the goveming DE is modified to
my"
+ cy' + ky
f(t)
5.4/Solutions of PDEs
229
(5.10)
Heat Conduction
O< X<
B.C.:
I. C.:
u(O, t)
= f(t),
u(x,O) =O,
ao,
t >O
u(x,t) - O as x O<x<oo
oo
(5.11)
B.C.:
where U(x,p)
U{O,p)
= F(p),
O<
X<
oo
U(x,p)- O as x -
oo
(5.12)
* The problem described by (5.11) can also be solved by applying the Fourier sine
transform to the variable x, but the Laplace transform is an easier tool to use in this case.
230
U(x,p)
(5.13)
= F(p)e-xvp/a
(5.14)
e-x2f4a2t
2a\ITTt 312
'
(5.15)
u(x,t)
1t
j(T)
= 2avTT
.. r o (t-T)3/2
u(x,t)
= T0 erfc(x/2a\lt)
T0
(5.18)
B.C.:
I. C.:
= a -2U1,
u(O,t)
u(x,O)
= O,
= O,
O<x<1,t>O
u(l,t)
= To
(5.19)
O<x<1
Uxx- (p/a 2) U = O,
O< x < 1
U(O,p) =O,
= T0 /p
U(l,p)
(5.20)
5.4/Solutions of PDEs
231
+ B(p)sinh(xy'pla)
(5.21)
sinh(xy'p I a)
p sinh(Y pia)
(5 22)
eP sinh(xyPia)
1
YP
.
;p == O
p smh( pia)
~
{eP sinh(xy'pla) ;p ==
+ LJ Res
p sinh(y'PI a)
n= 1
2 2 2} ]
-n1ra
_n2,-2a21] t
(5.23)
B.C.:
I. C.:
u(O, t) == f(t),
u(x,O) ==O,
oo,
t>0
as x ~ oo
ulx,O) == O, O < x < oo
u(x,t)
~O
(5.24)
* As a general rule, we use hyperbolic functions in the general solution when the
domain is finite and exponential functions [see (5.13)] when the domain is infinite.
t The standard method of solving this problem is by use of separation of variables,
e.g., see L. C. Andrews, Elementary Partia/ Differential Equations with Boundary V alue
Problems, Orlando: Academic Press, 1986.
232
Uxx - (p/c) 2 U
B.C.:
= O,
= F(p),
U(O, p)
O< X<
(5.25)
00
U(x,p) ~O as x-? oo
= F(p)e-xp/c
(5.26)
= f(t
- x/c)h(t - x/c)
= {J(t
t -.5 xf c
t >x/c
_ x/c),
(5.27)
= U11 + / 03(!
B.C.:
u(O,t) = O,
I. C.:
u(x,O)
- x/v),
O <X
u(x,t) ~ Oas x ~ oo
= O,
= O, O < x <
u,(x,O)
(5.28)
oo
Uxx - (p/c) 2 U
= (fo/c2) e-xpfv,
U(O,p)=O,
U(x,p)~Oasx-oo
O< X<
oo
(5.29)
fov2
U(x,p)
(e -xpjv - e -xpfc)
2)2
c - v P
{ _fox e-xpfc
2cp
..J.
T
=c
(5.30)
'
=I=
=c
(5.31)
5.4/Solutions of PDEs
233
+ a- 2U = 0,
11
{"((O,t)) = b ,
uxx(O,t) = O
u x,t ~ 0 asx~ oo
u(x,O) = O,
u1(x,O) = O, O< x <
B.C.:
I.C.:
(5.32)
oo
O< X<
B.C.:
oo
=O
(5.33)
= e-xvPti [A(p)cos(xyp/2a)
+ B(p)sin(x~)]
(5.34)
= (b/p)e-xvPti cos(xyp/2a)
= !(e;z
=
=
(5.35)
(5.36)
Then, using the inverse transform relation (see Prob. 14 in Exer. 4.6)
2- 1{(1/p)e-avP;t} = erfc(a/2Yt)
(5.37)
we deduce that
u(x,t) =
(x J-i)
(x Jl)J
1 erf 2 at
1 erf 2 -;;! - 2
=b [1- 2
(5.38)
234
= b [1
u(x,t)
- C(xjy2mrt) - S (x/y2mrt)]
(5.39)
EXERCISES 5.4
1. lf the boundary condition in (5.11) is
u(O,t) =f(t) =
{~:
O<t+b
t?:. b
O< t < b
T1erfc(x/2ayt),
u(x,t) = { T1[erf(x/2ayt - b) - erf(x/2ayt)], t?:. b
(b) Show that u(x,t) given in (a) is continuous at t
= b.
f(t)
Tt/Vt.
3. Given the heat conduction problem
Uxx
B.C.:
I. C.:
= a -2U
ux(O, t)
0 <X<
1,
= - f(t),
u(x,O) =O,
oo,
u(x,t)
t>0
~O
as x
~ oo
O<x<oo
ac_
= ~y7T
l'
f(r) exp [
~ r:--:
Oyt-r
-r Jdr
(
4a 2 t - r
= ..V~-J=
vf(t
7T x/2a
5. For the special case f(t) = K (constant), show that the solution of
Prob. 3 is
u(x,t)
= K [2ayt/7Te-x
2
/
4 2
ar -
x erfc(x/2ayt)]
5.4/Solutions of PDEs
235
peratures of the face x = O and the medium x < O. Given that the
mathematical formulation of the problem is
0 <X<
B.C.:
I. C.:
ux(O, t)
= ku(O, t),
u(x,O)
= To,
oo,
I> 0
u(x,t)-+ Oas x -
oo
O<x<oo
show that the temperatures inside the conducting solid are given by
u(x,t)
= 2k10
1T
z2 +
ln Probs. 7-10, use the Laplace transform to solve the given heat conduction
boundary-value problem.
7.
B.C.:
I. C.:
8.
B.C.:
I. C.:
9.
Uxx = a -2U1 ,
O<x<l,t>O
u(O,t) = T0 ,
Ux(l,t) = O
u(x,O) =O
Uxx = a -2U1,
O<x< l,t>O
u(O,t) = O,
u(l,t) = O
u(x,O) = T0
0.25 Uxx
B.C.:
I. C.:
10.
=U
1 -
ux(O,t) =O,
u(x,O)
u(lO,t)
= 20
= 50
Uxx = U1
B.C.:
I.C.:
1,
O<x<l,t>O
2x,
u(l,t) = O
u(O,t) =O,
u(x,O)
= x(l - x)
B.C.:
I.C.:
Uxx = Uro
O<x<I,t>O
u(O,t) = O,
u(l,t) = T0
u(x,O) = T0 ,
O<x<l
(a) show that the solution of the transformed problem can be expressed
in the form
-Z(l-x)Vp]}
[ 1 - e
T0 {
U(x,p) = - 1 - e-vp
V
p
1 - e- 2 P
236
=; [1-
e-xv'P
+ .. ]
- erfc[(2+x)/2yt] + }
12. A heat source of strength q(t)h(t), where h is the Heaviside unit
function, appears at the origin of a long rod at time t = Oand moves
along the positive x axis with constant speed v. The problem is
characterized by
Uxx
B.C.:
I.C.:
= u,
-oo<x<oo,t>O
- (x - vt)q(t)h(t),
u(x,t)- Oas
lxl -
u(x,O) = O,
oo
<x<
-oo
oo
= -1- ft q(T)(t
zv; o
1JT)ZJ4(t- T)
dT
B.C.:
I.C.:
= a - 2 u1 -
u(x,t) -O as
(x)(t),
lxl -
oo
<x<
oo,
t>O
oo
u(x,O) = T0
14. Using the Laplace transform, finda bounded solution of the exterior
temperature distribution problem for a sphere described by
(2/r)u,
u(l,t)
= T1,
= T0
u,,
B.C.:
I.C.:
u(r,O)
= u,,
1 <r<
u(r,t) -
oo,
O as r -
t >O
oo
u, + (1/r)u, = u,,
B.C.:
I.C.:
u(l,t)
= T1
u(r,O) = O
5.4/Solutions of PDEs
237
u_.(O,t)
I.C.:
u(x,O)
= f(t),
= O,
oo, t
>0
u(x,t) ~O as x ~ oo
u,(x,O)
=O
=-
g(z)
Lf(T) dT,
z > O,
and
z 2: O
17. Consider the motions of a string fastened at the origin but whose
far end is looped around a frictionless peg that exerts no vertical
force on the loop. The string is initially supported at rest along the
x axis and is released at time t = O, moving downward under the
action of gravity. Determine the subsequent displacements given that
the problem is characterized by
u(O,t) =O,
I. C.:
u(x,O)
u(x,t) ~O as x ~ oo
= O,
u,(x,O)
=O
u..(O,t) =O,
I. C.:
u(x,O)
oo, t
>0
u(x,t) ~O as x ~ oo
= e-x,
ur(x,O) =O
e- 1 coshx,
{ e-"cosht,
x<t
x>t.
ln Probs. 19-25, use the Laplace transform to solve the given boundaryvalue problem.
19.
Uxx
B.C.:
u(O,t)
I. C.:
u(x,O)
20.
Uxx
B.C.:
I. C.:
-2
U 11 ,
= O,
= O,
C
-2
Um
0 <X<
oo,
ux(x,t)
~O
u,(x,O)
0 <X<
>0
as x
~ oo
Vo
oo,
>0
u(O,t) =O,
u(x,t) ~ O as x ~
u(x,O) =A,
u,(x,O) =O
oo
O<x<l,t>O
B.C.:
u(O,t) =O,
u(l,t) =O
I. C.:
u(x,O) =O,
u,(x,O) =O
21.
238
22.
Uxx =
B.C.:
I. C.:
23.
24.
25.
u(l,t) = 1
u,(x,O) = O
-2
O<x< 1,t>O
Um
Ux(O,t) =O,
u(l,t) = 1
u(x,O) = O,
ub,O) =O
Uxx
B.C.:
I. C.:
O<x<1,t>O
Um
u(O,t) = O,
u(x,O) =O,
Uxx =
B.C.:
I. C.:
-2
=C
-2
O<x<l,t>O
Um
u(O,t) = O,
ux(l,t)
u(x,O) =O,
u,(x,O) = O
Uxx
=C
-2
=Ar
O<x<1,t>O
Um
B.C.:
u(O,t) = O,
ux(l,t) = O
I. C.:
u(x,O) =O,
u,(x,O) = x
t>
J:u(r)k(t - r) dr = f(t),
(5.40)
are known as V o/terra equations of convolution type (see also the discussion
in Sec. 3.2.1). The Laplace transform provides a useful technique for
the solution of such equations in which f(t) and k(t - r) are known
functions and u(t) is to be determined.
Using the result of Eq. (4.32) in Sec. 4.5.3 to take the Laplace transform
of (5.40), we obtain
U(p )K(p) = F(p)
(5.41)
from which it follows that
U(p)
= F(p)/K(p)
(5.42)
= oT- 1 {:~~;t}
(5.43)
2-
1
{
K:p)
;t} =
g(t)
(5.44)
239
then we can use the convolution theorem (Theorem 4.12) to express the
solution (5.43) as
u(t)
f(T)g(t - T) dT,
>o
(5.45)
ln some cases it may happen that the inverse transform (5.44) does
not exist, but 1/pK(p) has an inverse transform. Thus, (5.42) becomes
U(p)
= pF(p) L(p)
(5.46)
.:f- 1{pF(p);t}
= f'(t)
(5.47)
f'(T)e(t - T) dT,
>o
(5.48)
J: U(T)J (t -
T) dT
Sn t,
>0
Recalling that
.:t'{J0(t);p} = 1/Vp2 + 1
+ 1 = 1/(p2 +
1)
= 1/y'p2 +
= J 0(t)
lo(T)J0(t - T) dT
= sin t
240
Example 5.10:
J: e
u(t) -
-T
u('r) dr
= f(t),
>O
Solution:
U(p)
U(p)- - p- 1
= F(p)
which leads to
U(p)
= (p
- l)F(p)
= F(p) +
p-2
F(p)
p-2
Now taking the inverse transform of this last expression, we get the
formal solution
u(t) = f(t)
5.5.1
e2<t--T> /(r) dr
P(x,y)
241
(5.49)
1 J.Q
t =-
\(ii
ds
(5.50)
yy:::-;,
-u(TJ)dTJ
(5.51)
where the negative sigo reftects the fact that 11 is a decreasing variable
from P to Q. Substituting this last expression into (5.50) yields
and therefore the total time of descent T from P to the origin is given
by the Volterra integral equation
T
= 1 (Y
y2g Jo
U(TJ)
yy:::-;,
dTJ
(5.52)
or
U(p)
= Ty2g/TTp
(5.53)
(5.54)
Rewriting Eq. (5.51) in terms of x and y, and substituting (5.54) into the
square of the resulting expression leads to
242
1 + (dx) == 2gT ==
TTzY
dy
where a
2gT2/TT2
a=y
-y-dy
(5.55)
= (a/2)(0 + sinO),
y == (a/2)(1 - cos O)
(5.56)
f
t
O ( t - T)
dT
= J(t),
>o
(5.57)
where f(t) is given and a is a constant such that O < a < 1. If we formally
apply the Laplace transform to (5.57), we get
f(1 - a)pa- 1U(p)
= F(p)
or
U(p)
= f(l
F(p)
- a)pa
pF(p)
a)pa
(5.58)
ro -
= O,
then
= f'(t)
(5.59)
~ a)p";t} == -r(-1-~-a-~-)f-(a-)
1
2 -l{ro
sin
TTa
==--(a- 1
1T
( ) = sin TTa
U(
1T
1
1
f'(T)
(
)l-a
O t-T
d
T,
>o
(5.60)
(5.61)
243
EXERCISES 5.5
ln Probs. 1-10, we finda continuous solution ofthe given integral equation.
1.
u(r)
Jo y!l--;
dr
,fi
2.
r vu(r)
dr = 1 + t + 3t
t - r
Jo
3. J:u(r)u(t - r) dr = 4 sin 2t
4. u(t)
= t + J:(t - r)u(r) dr
S. u(t) = 4t
6. u(t)
+ J:e-r u(t - r) dr = 1
- 3 J:u(r)sin(t - r) dr
,i
= t + l J: (t - du(r) dr
u(r)
(
) 113 dr
O t- r
= t(l + t)
+ J: u(r)u(t
- r) dr
u(O)
L(t-
u(O)
= u'(t) - u(t),
=1
u(O)
=3
=O
u(O) = O, u'(O) = O
O<a<1
= f(t)
0 2 = w(l + w)
244
=L l0(2~)u(T)
f(O) =O
dT,
= f'(0)/0(2yt) +
J: 10(2~)/(T) dT
U(T)
;-:---_
l vtt
dT
= f(t),
u(O)
= /(0) = O
Lr(t -
T)g(T) dT
(
o t2
u(s)
20. f(t)
21. f(t)
= 2 J,
su(s)
~ds
22. f(t)
=t
~
ds
s- t
(""
),
- s2
)"ds,
O< a < 1
6
The Mellin Transform
6.1
Introduction
. Joo g(t)e't'. dt de
= -1 Joo e-tu
211"
-~
-~
(6.1)
g(logy)y-c
= ~i
y- L""g(logx)x-cx'- 1 dxds
(6.2)
c-oo
If we now define
f(x)
= g(log x )x -c
(6.3)
245
246
F(s)
= L""xs-l f(x) dx
(6.4)
and
c+ioo
f(x)
1
= 27Ti
Jx -s F(s) ds
(6.5)
c-ioo
We define (6.4) as the Meflin transform of f(x), and (6.5) is the related
inversion formula. We also use the notations
F(s)
= .M{f(x);s}
(6.6)
and
(6.7)
both converge for real a and b such that a< b, then the Mellin transform
of f(x) converges uniformly to F(s) in any finite region interior to the
infinite vertical strip a < u < b, where u = Re (s). ln such cases the
transform F(s) is analytic in this vertical strip. Also, if xk -y(x) is absolutely
integrable on the positive real axis for some k > O, and if F(s) is defined
by (6.4), then the inversion formula (6.5) is valid for c > k.
= L""x-le-x dx
247
At{cos x;s} =
f"
x'- 1cos x dx
= J~f'x-l cos x dx
f:Fc{x'- 1 ;}
'
= f(s)cos(!7Ts ),
At{cos x;s}
Similarly, the result
follows directly from the sine transform given by Eq. (2.68b) in Sec.
2.6.1.
Example 6.3
1
}
l~-x'-1
At { - - ;s =
-dx
1+x
o1+x
By imposing the restriction O < Re(s) < 1, the integral is exactly
that given by Eq. (1.13) in Sec. 1.2.2. Hence, it follows that
AtL
~ x ;s} = f(s)f(l
- s),
248
= -Tr-
Operational Properties
(6.9)
where F(s) and G(s) are the Mellin transforms, respectively, off(x) and
g(x). This property is a simple consequence of the linearity property of
integrais.
By making the simple change of variable x = t/a in the following
integral
At{f(ax);s}
Looxs-J f(ax) dx
=~[r-'
J(t) dt
a o
we deduce the scaling property
a>O
At{f(ax);s} = (1/as)F(s),
(6.10)
Al{xaf(x);s}
= Loor+a- f(x) dx
Al{xaf(x);s} = F(s + a)
(6.11)
At{f(xa);s}
= (1/a)F(s/a),
a>O
(6.12)
and
At{(l/x)f(l/x);s}
= F(l
- s)
(6.13)
the proofs of which are left to the exercises (see Probs. 16 and 17 in
Exer. 6.2).
lff(x) is continuous on x 2:: O and has a Mellin transform F(s), then
..ttl{f'(x);s}
249
= L~xs-l f'(x) dx
= x- 1f(x)~~-
(s- 1)
L~x.- 2 f(x)dx
= O,
<T 1
(6.14a)
= O,
<Tt
(6.14b)
x-+0
limr- 1/(x)
<Tz
<Tt
(6.15)
-(s- l)..ttl{f'(x);s- 1}
or
..tU{f"(x);s} = (s - 1)(s - 2)F(s - 2)
(6.16)
= ( -1r
f(s)
r (s
- n)
F(s - n),
= 1,2,3, ...
(6.17)
provided
limx-k-tf<k>(x)
= O,
x-+0
= 1,2, , n-1
(6.18)
=
d
f(x/u)g(u) __!!_
o
u
(6.19)
= L~(l ~-~)m dx
250
but recalling the definition of the beta function (see Prob. 21 in Exer.
1.2)
B(x,y)
Lao(I
:x~;x+y dt,
x>O,y >O
= B(s, m
- s)
f(s)f(m - s)
f(m)
Al { (1
+ axr;s =
f(s)f(m - s)
aSf(m)
'
Examp/e 6.5: Find the Mellin transform of x-vlv(ax), a> O, v> -1/2.
Solution: ln this case the Mellin transform
Al{x-vlv(ax);s} = Laoxs-v-Ilv(ax) dx
and
~c{xs-I
;}
= y2/7T f(s)cos(!'TTs)g-s
[see Eqs. (2.66) and (2.68a) in Sec. 2.6]. Then, using the cosine
transform relation
Laof(x)g(x) dx
= LaoFc()Gc() d
Jo X
ra-s(a2-
Jo
gzy-1/2
d = !a2v-s f r<s+l)/2(1
2
Jo
t)v-1/2
=-a
dt
251
-1
v-sf(s)f(!- ls)cos(i7Ts)
xs v J(ax)dx =a
o
v
v1Trf(v -is+ 1)
oo
= f(7Tf(S)
= ~C
v 1T 2s-t f(lzS )
2 + zS
1
EXERCISES 6.2
ln Probs. 1-15, evaluate the Mellin transform of the given function.
When possible, use known integral results from previous chapters.
1. f(x)
3. f(x)
= h(a - x),
= xae-bX,
a>O
2. f(x)
= e-hx,
>O
4. f(x)
a,b
7. f(x)
= x"/(1 +
>O
9 f( ) = h(t - x)
x
e-b'x\
xt,
a,b >O
b>O
a>O
(x- 1)",
10. f(x)
[y?+t- x]"
.. 12"":-:
,
vx2 + 1
a>O
Hint: Let x
I
15. f(x)
= Jix2),
lv(x),
> - l/2
= b/W+J..
14. f(x)
= Jv(V-;;),
v> -l/2
v> -l/2
Hint: ln Probs. 14 and 15, use the result of Prob. 13 and Eq. (6.12).
16. Al{f(x");s}
a> O
(l/a) F(s/a),
17. Al{(l/x)f(l/x);s}
F(l - s)
252
= (-stF(s),
= ( -1tf(~(:) n) F(s),
1,2,3, ...
1,2,3, ...
+ xf'(x);s} = s2F(s)
+ 1)
23 .,u{J:f(u) du;s}
= (1/s) F(s +
25 .,u-'{F(s)G(s);x}
26. Al{f(x)g(x);s}
-(1/s) F(s
1)
Loof(x/u)g(u) du/u
1 Jc+ioo
27
28
+m -
s)G(s)
+ m +
s)G(s)
= ~{f(t);x}
= y:;;j2 ~c{f(t);x}
s);x} = v:;;;2 ~slf(t);x}
g(s)
an
= "'
LJ--:;
n=l
= Looxs-l f(x) dx
g(s )f(s)
where
n=l
21'fl
c-100
At{ex ~
;s}
= '(s )f(s)
2: -..
n
n=l
At{ ex ~
(b) At{
ex ~
1 ;s} -
1 ;s}
At{ ~ ~
1 ;s}
zl-s '(s)f(s)
(1 - 2(-sms)
At { -1 -+1 x's }
1'(
=-
sin 1rs
s} =
1
At{-- 2
1+x '
--7T...,.-2 sin(!7Ts)
253
254
6.3
Me/lin Transforms
O"z
O"z
(6.20)
z-->0
limt'f(z)
lzl--->=
= O,
Jc
k=t
Res{z-t_r(z);ak}
or
rRXs-t_r(x)dx
JP
JcR
JR
Jcp
= 2?TiL Res{z-t_{(z);ak}
k=l
(6.21)
Because of the limit relations (6.20), it can be shown that the integral
around the large circle of radius R vanishes in the limitas R --+ =, as
does the integral around the small circle of radius p as p --+ O. Hence,
in the limit as R --+ = and p --+ O, we obtain
(6.22)
255
Figure 6.1
Contour of integration
However,
J<
= - ~e-i"'s ~ Res{z-t_{(z);ak}
sm 'Tf'S
(6.23)
k=l
Eq. (6.23) can be easily modified to include the case wheref(z) has poles
on the positive real axis (see Prob. 8 in Exer. 6.3).
Example 6.6:
Solution:
s-1
}
_z__
.i
l+l'
s-1
_z-
2i
= - - ei"'s/2
and
256
oo
e3i'ITs/2)
= __;!!.--. !(e-i'lrs/2 +
ei'ITs/2)
sm 7TS 2
7T cos (7Ts/2)
sin 7TS
O<Re(s)<2
Observe also that we could obtain this result from Exam. 6.3 and
the operational property given by Eq. (6.12).
(6.24)
To formally recover the function e-x from the transform function f(s),
we consider the inversion formula
c+ioo
(6.25)
c-ioo
= -n
(n = 0,1,2, ... ). Thus, since f(s) is analytic in the right-halfplane, we
can evaluate (6.25) by considering the sarne kind of contour used in
evaluating inverse Laplace transforms [see Fig. 6.2(a)]. Dueto the asymptotic behavior of f(s) for large Is I, it can be shown that the integral along
the large circular are tends to zero for ali x as the radius of the circular
are tends to infinity. Based on the calculus of residues, we then conclude
that
(6.26)
n=O
7T
.
f(l - s )sm 7TS
(6.27)
257
lm(s)
lm(s)
c + i
Re(s)
Re(s)
----+------+----~--
c - i
c - i
(a)
(b)
Figure 6.2 Contours of integration
we find
Res{x-f(s); -n}
= f(l
x-
)
SCOS1TS
s=-n
( n'.
-l)nx"
n!
(6.28)
Example 6.7:
F( )
= f(s)f(m
- s)
ar(m)
'
a>O,O<Re(s)<m
(n
258
R { -Fi( )
}
es x
s ;- n
- s)
= ax-r(m
r(m )'f(l - S ) COS 7TS
I
s= -n
_ ( -ltf(m + n)( )n
n!f(m)
ax
However, using the identity (see Prob. lO in Exer. 1.2)
(-1rr(m +n)
n!f(m)
= (- 1r(m + nn
1) = (-m)
n
_,u- {F(s);x}
It is known that this series converges for Jaxl < I,* and hence we
have
_,u- 1{F(s);x} = 1/(1 + ax)m,
O<x< I/a
(ax) f(m)
f(s )f(m - s + n + l)
s->m+n(axY(m- s)(m- s + 1) (m- s
= _ Iim
+ n-
1)
*These are the sarne values of x for which the integral along the circular are in Fig.
6.2(a) vanishes in the limit as the radius tends to infinity.
259
(-ltf(m+n)
n=O
af(m)
'
(1
+ ax)m'
O<x<=
= L=r-1 f(reie) dr
= e -is8L=s-1 f() d
= e-is8 F(s)
*W. J. Harrington, "A property ofMellin transforms," SIAM Review, 9, No. 3, 542547, July (1967).
260
which leads to
(6.29a)
(6.29b)
.,.u-
I{
1T
}
sin S1T ;x
= 1 +1 x
1 ]
1 + rei8
1 + cos (}
1 + 2r cos (} + r 2
{1Tsins(}
---s
sin S1T '
11 _ 1
.;n
~r
rsin(}
= ---------=
2
1 + 2r cos (} + r
EXERCISES 6.3
ln Probs. 1-6, use residue calculus to evaluate the Mellin transform of
the given function.
1
a,b >O
l. f(x) = (x + a)(x +h)'
1
1
3. f(x) = - - 3
2. f(x) = (1 + x) 3
1 +x
4. f(x)
=-
6. f(x)
S. f(x)
1 +x
(1
+ rf
JU{f(x);s}
261
.M{f(x);s}
= __;!!--
LRes{(e-i,.zY-1(z);ak}
sm rrs k=t
ln Probs. 9 and 10, use the result of Prob. 8 to find the Mellin transform
of the given function.
1
1
9. f(x) = - 10. f(x) = 4 -x2
1- X
11. Show that
(a) .M{erfc(x);s}
r(s ;
);v;
r(s
.M
- 1{
1)
-2-
y;
1TS
}
;x
=1-
2 "" (-1tx2n+l
r= L
'(2n
V1Tn=On.
) = erfc(x)
+1
f(s)/s
E (x)
1
{""e-t dt
Jx
(b) Using residue methods, find an ascending series for the inverse
Mellin transform of F(s) = f(s)/s.
=-
f(s) cos(rrs/2)
(b) Using residue methods, find an ascending series for the inverse
Mellin transform of F(s) = - f(s)cos(rrs/2)/s.
14. Use residues to show that
(a) .M-1{
1T
x} = ~ 2s COS(rrs/2) '
2
~ ( -l)nx2n+l
2n + 1 '
n=O
O<x<1
262
(b) .At
7r
2s cos(Trs/2) ;x
= ~o
(2n
-IY
x>I
l)x 2n+l'
.At- 1{
7r
2s COS(Trs/2)
x>O
15. F(s)
= r~:.
17 ' F(s)
= 2 sin(Trs/2)
19. F(s)
21. F(s)
7r
16. F(s)
= -.-
18. F(s)
= f(s)cos(7Ts/2)
= f(s)sin(Trs/2)
20. F(s)
2- 1r(s/2)
f(l - s/2)
= c~s scp,
22. F(s)
= -.-, - - <
7r
smsa
7r
sm 1TS
sin scp
7r
smsa
<P
7T
2
< -
7r
= 7r c_os sO
24. F(s) = 7T
~in sO
s sm S7T
s sm sTr
6.2.
25. F(s)
= cos s(}
COS
S7r
26. F(s)
= sin s()
COS
S1T
6.4 Applications
Although the Mellin transform is not nearly as versatile in applications
as are the Fourier and Laplace transforms, there are some areas of
application where it can be a useful tool. ln particular, it is useful in the
summation of certain series, in finding the distribution function for products
of random variables, and in solving for the potential function in a wedgeshaped region. Our discussion of these applications, however, will be
intentionally brief since any deeper treatment would require mathematical
knowledge beyond the stated prerequisites.
6.4/Applications
6.4.1
263
Summation of Series
f(x)
J -
1
7Ti
x s F(s) ds
2 c-ioo
=
=
c+iC/C)
f(nx)
2 Jn-sx-sF(s) ds,
= ~i
n = 1,2,3, ...
(6.30)
c-ioo
Now let us sum both sides of (6.30) over ali positive integer values of
n, interchanging the order of summation and integration on the righthand side. This action leads to
c+ioo
00
l;((nx)
where
~(s)
1
7Ti . x-s~(s)F(s) ds
2 c-too
(6.31)
00
(6.32)
n=ln
(6.33)
n=l
The importance of Eq. (6.33) is that we can replace the series on the
left, which in some cases may converge very slowly, by the inverse
Mellin transform on the right. This inverse transform leads to another
infinite series, but in certain cases it may be summed exactly. ln other
cases this new series may converge faster than the original series, which
is important for computational purposes. Proficiency in this technique
of summing series requires knowledge of some properties of the zeta
function, most of which were developed in 1859 by G. Riemann (1826-
264
1866). Some of the relations involving the zeta function that prove useful
in our work here include the following: *
,(0) = -1/2
,(1)
00
,(2)
2
7T
=
'( -2n) = O,
4
7T
,(4)
(6.34)
(6.35)
(6.36)
/6
/90,
(6.37)
= 1,2,3, ...
(6.38)
(6.39)
7T''(l - s) == i-r(s),(s)cos('ITs/2)
Example 6.9:
2: (cos an) / n
(6.40)
n~t
Al{f(x);s}
= L"'x.- 3 cos ax dx
= Al{cos ax; s- 2}
= - (1/a'- 2)f(s - 2)COS(7TS/2)
ln deriving this transform we have used the result of Exam. 6.2 and
the scaling property (6.10). Substituting this result into Eq. (6.33)
yields
~ cos an ==
LJ-2
n~l
_{
-j{ 1
1_ ,(s)f(s- 2)COS(7Ts/2);x= 1}
2
U sing property (6.40) of the zeta function, we can express this inverse
Mellin transform in the more convenient form
~ cos an = _ a
n=t
= _ a .M-t{(27T)'
'(1 - s)
(s - l)(s -
x==
2) '
t}
t}
*For a detailed discussion ofthe zeta function, see E. T. Whittaker and G. N. Watson,
A Course of Modern Analysis, Cambridge: Cambridge University Press, 1965.
6.4/Applications
265
O, 1, 2, the residues of
= 1r/a, Res{2} =
-7r
/3a 2
The details of evaluating these residues are left to the exercises (see
Probs. 1 and 2 in Exer. 6.4). Combining results, we finally obtain
~ cos an = _ a
n= 1
n2
2
(
_! + ~ _
2
7r )
2
3a
=a
2
_
o+
6.4.2
2
7r
L-=~
n= I n2
7ra
(6.41)
=~
JLpx.y(x,y) dy dx
(6.42)
where Dz is the region of the xy plane such that xy :s z. (ln this section
the variable z is considered to be real.) Since X and Y are independent
random variables, we can express their joint density function as a product
of their respective PDFs, and thus (6.42) becomes
r=
d
rz/x
Pz(Z) == dz Jo Jo Px(x)py(y) dy dx
266
which simplifies to
Pz(z)
i"
o
dx
Px(x)py(z/x)X
(6.43)
z = nxj
(6.45)
j=l
= .l- 1ll]Fis);z}
(6.46)
where Fj(s) is the Mellin transform of the PDF for Xi, j = 1,2, ... ,N.
Further generalizations to products involving both positive and negative
random variables have been established, as well as generalizations to
certain quotients of randorn variables.
Find the PDF of the product Z = XY, where X and Y
are independent Cauchy random variables, each having density function
Example 6.10:
p(x)
Solution:
= 2/7T(l + x2),
x >O
4 (""
Pz(z)
= 'TTz J0
(xz
l)(xz
+ l)
.M{f(x);s}
2(z2
l)sin(TTs/2)
6.4/Applications
267
2(zs-2 - 1)
= hm 7T(z2 -
Pz(z)
.~2
. ('Tf'S / 2)
1)sm
Pz(z)
4log z
7T2(z2 _ 1)'
z>O
We could equally well use Eq. (6.44) to find the PDF by first
noting that
Al{px(x);s}
= Al{py(y);s} = 1/sin(7Ts/2)
However, the details of this approach are left to the exercises (see
Prob. 12 in Exer. 6.4).
Finally, observe that, because Px(x) and py(y) are even functions,
we could extend our result to products involving Cauchy random
variables that extend over -ao< x <ao, -ao< y < oo. The PDF
for Z = XY in this case assumes the form
Pz(Z)
log z2
-ao<z<oo
7Tz(l _ l),
r 2u,,
B.C.:
u(r,-a) = f(r),
{ u(r,O) ~O as r~
O<r<oo, -a<O<a
u(r,a)
ao,
= g(r)
(6.47)
101 <a
268
Among other areas of application, this problem describes the steadystate temperature distribution in an infinite wedge, given the temperature
distribution along the boundaries of the wedge.
Using the differentiation property of the Mellin transform (see Prob.
22 in Exer. 6.2)
(6.48)
we find that under the Mellin transform the potential problem (6.47) is
transformed to
uflfl
B.C.:
+ s2U = o,
U(s,-a)
-a<()< a
= F(s),
U(s,a)
(6.49)
= G(s)
(6.50)
,.m
[ {""
= 2a cos(m(J) Jo g2m
(""
+ Jo
gm-1 J(g) dg
+ 2rngmsin(m()) + ,zm
gm-1 g(g) d
em - 2r"'msin(m8) + ,zm
(6.51)
where we are using the inverse Mellin transform relation (see Prob. 22
in Exer. 6.3)
s~r} = _!_.
2a
rmsin(mc/J)
1 + 2rmcos(mc/J) + r 2m
where m = 7T /2a.
EXERCISES 6.4
1. Given the function
F(s) = (2/as) '(s)f(s - 2)cos(7Ts/2)
(6 _52 )
6.4/Applications
269
,(1 -
s)
(s - l)(s - 2)
show that
27T
7T
a
a
47T2
7T2
(b)Res{F(s);2}=2 '(-1)= - -2
a
3a
(a) Res{F(s);I}
=-
-'(0) =-
3. Show that
00
"" sm an _ 7T - a
L.J
n=l
n 2 ,
4. Show that
~ _,2
L.Je
= --+
2
n=l
.. ~-~
y7TL.Je
-1TZk2
k=O
5. Show that
(a)
~ (-I~"- I = (I
n
n=l
(b)
- 21 -ms)
~o (2n ~ IY =
(I - 2 -.),(s)
(a)
L (-1)"-Y(nx) = AC 1{(1
- 2'-ms)F(s);x}
n=l
n=O
ln Probs. 7-10, use Eq. (6.33) or the results of Prob. 5 to sum the given
series.
oo ( -l)n-1
7.
n=l
00
cos an
n=l
L
n=l
oo (- l)n-1
9.
S.
sin an
sm an
2n - 1
~ J 1[(2n
10. L.J
n=O
+ 1)a]
2n + I
7T(Zs-2 - I)
1
}
{ (x2 + l)(x2 + z2) 's = ---:=---'----'--2(l - 1)sin(7Ts/2)
270
1
}
= Ai -1{ sin2(7Ts/2)
;z
(b) U sing the method of residues, show that the inverse Mellin transform in (a) leads to the result of Exam. 6.10.
Pz(Z) =
+ 1)2
2
= (a +
Z h(l - z)log ~
0
l)xah(l - x),
z> 0
(a + l)N-1
(
1)N-1
(N _ 1)! zah(l - z) log ~
,
z>O
= Al{py(y);-s+2}
[(a
{ [(a
+ 1)/2] Z
+ 1)/2] Z-a-2,
0
'
O:::;z:::;1
z > 1.
15. Recall from Exam. 3.10 in Sec. 3.8 that the product of two independent,
zero-mean, Gaussian random variables with equal variances has the
PDF
Pz(Z)
= (1/7T)Ko(lzl)
16. Use the result of Prob. 15 to deduce that the PDF of Z = XY, where
X and Y have the sarne PDF p(x) = e-xh(x), is given by
pz(z)
= (2/7T)Ko(Vz /2)h(z)
6.4/Applications
271
17. Use the result of Prob. 14 to find the PDF of Z = X/Y, where X
and Y are positive independent random variables, each with PDF
p(x) == V2/TT e-x 212 h(x)
18. For the special case of (6.47) where the boundary conditions are
u(r, -a) == u(r,a) == f(r)
At _ 1 { cos s() .s ~
cos sa'
r} =
r"'
12
where m == TT /a.
(c) Using the result in (b), find a formal solution of the potential
problem by using the convolution theorem.
19. Finda formal solution of the problem described by (6.47) when the
{~:
O<r<a
r>a
20. Find the steady-state temperature distribution inside the infinite wedge-
shaped region O < r< oo, O < () < a, if the boundary () = O is held
at temperature zero while the other boundary is maintained at
u(r,a)
T.
O< r< a
= { 0,0 '
r>a
d
Jor= u()g(x/) T
1
2m
u(x) == - .
fc+i= X-sF(s) ds
c-i= 1 - G(s)
x) in Prob. 21,
'TT
272
(b) lf 7T
Jtl
-I
SID a7T
SID S7T -
SID a7T
t an a7T ( X -a
,X-
1-
7T
XI
2
+a)
+ x (tan7Ta7T)
u(x) =f(x)
= L= g(x~)u(~) d~
1 fc+i= F(l - S) -s d
=-
27Ti
c-i= G(l - S)
24. Use the result of Prob. 23 to solve the Laplace integral equation
x>O
= f(x) +
L= g(x)u() d
=-
-s
r=
= f(x) + y Jo
e-xl;
u(~) d~,
x >O
where
log(l + x)
f(x)
~ { lo~/+ x) _ !,
X
O<x<1
x>1
6.5
273
No.
Mellin Transforms
I ic+i=
f(x) =---:
x- F(s) ds
F(s) =
27Tt '-i=
x'- 1 f(x) dx
7T
I +X
2
3
I - X
I
(I + ax)m'
sin 1rs
7T cot 1rs
a >O.m >O
a >O,m >O
f(s)f(m- s)
a'f(m)
f(s/a)f(m - s/a)
a f(m)
f(s)/a'
f(s/2)/2a"
cos ax,
a> O
f(s)
- . COS(7Ts/2)
sin ax,
a > O
f(s) .
-sm(7rs/2)
x-"lv(ax),
a'
a'
a"-'2-- 1f(s/2)
f(11 + 1 - s/2)
JJ)
s +
2'-lf ( -2-
lO
J"(ax),
s
a'r -v - -+
I)
(
7
The Hankel Transform
7.1
Introduction
_!_f= J=
2'7T
-oo
ei<x~+.v">f(x,y)dxdy
(7. 1)
(7.2)
-oo
and
f(x,y)
= _!_ J=
2'7T
-oo
J=
-oo
7.1/lntroduction
x =r cosO,
= p cos ~.
y = r sinO
Ti= p sin ~
275
(7.3)
"')
,p,"'
= _1 f""
2'11"
where we have recognized that x + Y'Y/ = pr cos (O property (J12) in Sec. 1.4.1, we have
-1
12" e'Pr...
2rr o
cos(8-t/J)
(7.4)
~).
Recalling
dO = lo(pr)
(7.5)
where J 0(x) is the Besselfunction of order zero. This result suggests that
is a function of p alone, i.e., F(p,~) = F(p), and in this case,
(7 .4) leads to
F(p,~)
(7.6)
called the Hankel transform of order zero. Clearly, the substitutions (7.3)
into Eq. (7.2) will produce the similar result
f(r)
= L"" pF(p)J0(pr) dp
(7.7)
= L"' rf(r)lv(pr) dr =
F(p),
v> -1/2*
(7.8)
Yt'; 1 {F(p);r} =
v> -1/2
(7.9)
*The restriction v > -1/2 may be extended to a larger interval in certain transform
results.
tSee I. N. Sneddon, The Use of Integral Transforms, New York: McGraw-Hill, 1972,
pp. 79-83.
276
The basic requirement for the existence of the Hankel transform (7 .8)
is that the function Vrf(r) be piecewise continuous and absolutely integrable
on the positive realline. The proof of the Hankel inversion formula (7.9)
is similar to, but more complicated than, the corresponding proof of the
Fourier inversion theorem. This is due primarily to the fact that the
Hankel inversion formula relies on a good understanding of the properties
of Bessel functions, which are more complicated than those of the corresponding kemels of the Fourier transform. The interested reader can
consult Sneddon for the proof. *
J: r v+ 1v(pr) dr
1
= v\2lap Xv+IJv(X) dx
p
the last step resulting from a change of variable. Now using Eq. (1.42)
in Sec. 1.4.1, we deduce that
a>O,v> -1/2
Example 7.2: Find the Hankel transform of rv-te-a', a >O.
Solution: Here we see that the Hankel transform can be expressed
as a Laplace transform, i.e.,
*1. N. Sneddon, The Use of Integral Transforms, New York: McGraw-Hill, 1972,
pp. 301-309.
277
= O in
1 -ar ;p } = . . 1
-e
1 2
vP +a2
~o{;;p} = ~'
a> O
(7.10)
p>O
(7.11)
or
a>O
(7.12)
2
Example 7.3: Find the Hankel transform of order zero of e-a r2.
Solution:
'Je0{e -aZrZ;p}
<11- o e
-a2r2, }
,p
= ~ ( -l)n(p/2)2n f""
LJ
n=O
( ') 2
n.
,Zn+l -a2r2
278
the last step of which follows a change of variables. This last integral
is simply n!, and thus
'Je { -az~. }
Oe
,p
= _1 ~
( -1t(p /4a
2a2 n~O
n!
or
ou; {
dLo
-a2~.
} -
_1_
,p - 2a2
-p2/4a2
'
a >o
Operational Properties
Because the Hankel transform and inverse Hankel transform are exactly
the sarne in functional form, it follows that each operational property of
the transform is likewise a property of the inverse transform. For example,
as a consequence of the linearity property of integrais, we deduce that
'Je~{Ctf(r)
1
'Je; {C1F(p)
(7.14a)
(7.14b)
where C 1 and C2 are arbitrary constants and F(p) and G(p) are the Hankel
transforms, respectively, of f(r) and g(r).
If we make the simple change of variable x = ar in the defining
integral
a>O
we obtain
'Je~{f(ar);p}
= 2a1 100
xf(x)Jv(pxfa) dx
o
= (l/a2)F(p/a),
a>O
(7.15)
279
m=-cx:>
f'
pF(p)G(p) dp
Solution:
L"" xF(x)G(x) dx
F(x)
= (1/x)Jv+l(ax)
G(x)
= (1/x)Jv+l(bx)
where
and
(7.16)
280
We recognize from Ex. 7.1 that F(x) and G(x) are the Hankel transforms,
respectively, of
and
= V/bv+t)h(b-
g(r)
r),
b >O
fa
1I=r2v+l dr
(abt+ 1 o
which leads to
r=
1
1
Jo ~lv+ 1 (ax)Jv+l(bx)dx = 2(v + 1)
(a)v+l
b
(7.17)
=;v L=
or
Yt'v{;f(r);p}
=;v [
Yt'v+t{f(r);p}
= - pYt'o{f(r};p}
(7.19)
and
Yt'o{; :)rf(r)];p}
= pYt't{f(r);p}
(7.20)
= L= rf'(r)J1(pr) dr
= rf(r)J1(pr}
r=
d
Jo f(r) dr [rJ 1(pr)] dr
(7.21)
281
Recalling the asymptotic relations [see (J13) and (J14) in Sec. 1.4.1]
(7 .22a)
z -7
00
(7.22b)
we see that the first term on the right in (7 .21) vanishes provided
lim y;f(r) = O
lim rf(r) =
o.
The first limit condition is satisfied by any function whose Hankel transform
exists, and the second condition imposes on f(r) a certain behavior near
r = O. ln addition, we note that [see Eq. (1.36) in Sec. 1.4.1]
d
dr[rJ 1(pr)]
= prJo(pr)
=-p
f"
rf(r)Jo(pr) dr
(7 .23)
or
;tevV- 1 :,[r1 -vf(r)];p}
-p;tt'v-I{f(r);p}
(7.24)
= p,rt'v+I{f(r);p}
(7.25)
The verification of these results is also left to the exercises (see Probs.
26-28 in Exer. 7.2).
Hankel transforms provide significant simplifications in solving partial
differential equations that lead to Bessel's equation
ry"(r) + ry'(r) + (r 2
v2 )y(r)
=o
(7.26)
d
= y"(r) + -r1 y'(r) = -1r -d
[ry'(r)]
r
(7.27)
282
and apply the Hankel transform of order zero. This action leads to
:1eo{go(r);p}
= p:1et{y'(r);p}
= -p2:1eo{y(r);p}
(7 .28)
?: ![,zv+t :r(~y(r)) J
(7.29)
and apply the Hankel transform of order v, we obtain the similar result
(7 .30)
by use of (7.25) and (7.24), respectively. Hence, we see that the Hankel
transforms of g0(r) or gv(r), which involve derivatives of y(r), are related
directly to Hankel transforms of y(r).
EXERCISES 7.2
ln Probs. 1-15, verify the given Hankel transform relation. When possible,
use known integral transforms results from previous chapters.
ou; { v -ar. }
dL v
re
,p
a>O
2.
ou; {
1e
dL v -
-ttr ;p} = . . I
v p +a
P
... I 2
2
a+ vP +a
)v
a >O
'
6. :1evV(a2
a> O,
r 2)~'-"- 1 h(a -
IL >v~ O
= 21-L-If(~-t)(~rJ,..(ap),
r);p}
= 2~'-"-
f<~-t
a> o
v)a~'p"-~'J,..(ap),
7.
283
a> O
p),
v-1'
. _ pv(a2 _ p2)"'-v-l
Jiar),p} - 2~'- v-lf(JL - v)a~'-h(a - p),
a> O, JL >v;:::: O
8. ;r{v{r
9. ;reo{ Va 2
a> O
10. ;rev{
rvh(a - r)
v'a2 -
r2 ;p
JTT2/
v+ 1/2
lv+ 1/z(ap),
>o
= ~h(a
1 2
1 .
}
12. ;re 1{ - sm ar;p
r
13.
va -
p)
p
a> O
ah(p - a)
p p2 -a2
14. ;rt' 1{
~ cos (br2);p} = ~ [ 1 -
15. ;rt'1{
>O
a >O
cos(:;)].
b >O
b >O
16. Integrate both sides of the Hankel transform relation in Prob. 7 with
respect to a from O to b and deduce that
Hint:
b>O
J 1(ar) da
= - [1 - J0(br)]
r
.!l{lv(pr);r~(s
+ 1)}
(b) Evaluate the Mellin transform in (a) and thus show that
_1
;rt'v{r'
;p}
= ps+lf(v/2
-1-v<s<1+v
284
a2 ,
O<a<b
p> 1
a>O
Yfo{e-arl lo(br);p} =
~exp(p ~ b )/o(~:).
a> O, b >O
v'rz
+ a2
1_
r= e-(rZ
VTr Jo
a2)x
X-1/2
dX
to deduce that
e-ap,
a> o
22. Use the result of Eq. (7 .10) and other appropriate properties to show
that
2az - pz
a>O
(a) .rt'o{re-a';p} = ( 2
z)s;z
p +a
(b) .rt'o{e-a';p}
= (pz
+a a2)3/2
a>O
-ar
;p}
a + vy'pz + az (
(pz
+ az)3/2
P
)"
a+ ypz + az '
a> O
7.3/Applications
7.3
285
Applications
Potential Problems
u(r,O) = f(r),
{ u(r,z) ~ O as
O < r < oo
+ z2 ~ oo,
y?-
z >o
(7.31)
Uzz - p U
B.C.:
U(p,O)
{ U(p,z)
= O,
z >O
= F(p)
~O
as z ~
(7.32)
oo
where
:1t'0{u(r,z);r
p}
Jeo{flr);p}
= U(p,z)
= F(p)
(7.33)
(7.34)
and we are using the result of Eq. (7.28). Clearly, the solution of (7.32)
is
U(p,z)
= F(p)e-pz
(7.35)
or
u(r,z)
Example 7.5:
(7.36)
where
u(r,O) = f(r) =
Yr I+a
2
a>O
286
Solution:
Hence, substituting this result into (7 .36) leads to the Laplace transform
integral
u(r,z)
f'
= 2{J0(pr);p ~ (z + a)}
from which we conclude
u(r,z)
= ---r:;;====;;
2
2
v'r + (z + a)
u, + (1/r)u, +
B.C.:
u(r,O)
= u0 ,
uz(r,O)
= O,
Uzz
{ u(r,z) ~O as
= O,
O< r<
O< r< 1
1 <r< oo
2
+r~ oo,
Vr
oo, z
>O
(7.37)
z>O
z >O,
(7.38)
= A(p}e-pz
(7.39)
(7.40)
for the determination of the function A(p). Now imposing on this formal
7.3 I Applications
287
L"" pA(p)Jo(pr) dp = u
O< r< 1
(7.41a)
1 < r< oo
(7.41b)
0,
Equations of this variety are known as dual integral equations, but the
general theory concerning them goes beyond the intended scope of this
text. * However, to solve this particular pair of equations we simply start
with the observations (see Probs. 13 and 14 in Exer. 1.4)
(""
Jo
sin p
7f'
= 2'
O<r<1
(7.42a)
1 <r<oo
(7.42b)
J 0(pr)pdp
0
2u0
smp
7rp 2
=-
(7.43)
= -2u0
'
u(r,z)
7f'
7.3.2
(7.44)
Vibration Problems
+ (1/r)u, = c- 2u11
r~
(1/T)p(r,t),
(7.45)
oo
u1(r,O)
oo
288
I. C.:
c
+ c2p2U = pTP(p,
t),
U(p,t)
= F(p),
>O
(7.46)
= G(p)
U,(p,t)
Using standard solution techniques, we find that (7.46) has the solution
U(p,t)
Jo
- T)]dT
(7.47)
The subsequent inversion of (7 .47) then gives us the formal result for
the displacements
u(r,t)
(7.48)
EXERCISES 7.3
1. Finda formal solution of the heat-conduction problem
u,,
+ -1 u, = a -2u,,
B.C.:
I.C.:
u(r,O)
= f(r)
2. Solve the problem described by (7.31) for the special case where
u(r,O) = f(r) = 1/(r2
+ a 2 ) 312 ,
a> O
B.C.:
(1/r) u,
+ Uzz =
u(r,O), =O
{ u(r,z) ---+O as
O< r<=,
y'r + l---+ =,
-f(r),
roo F(p)
= Jo
(1
- pe
z>O
-pz) J (pr) dp
0
= a/(r2 + a 2) 312 ,
a >O
z >O
7.3/Applications
( ) 1[1
1- 1og
w(r,O)-
u r z =- og
'
2
w(r,O)
where
+1
J+
289
1]
w(r,z)w(r,z) + 1
w(r ,z)
(z
a)
+1
(l/r)
- Kuz(r,O)
B.C.:
U,
= (Q/ a )h(a
2
{ u(r,z) ~O as
yr + z
2
O< r<
O,
Uzz =
oo,
- r),
a> O
~ oo,
z >o
z >O
= Qjy,Z + i +
u(r,z)
B.C.:
(1/r)u,
u(r,a)
Uzz
=O,
= -Q/~
v (r ,z) -
l""
cosh zp -ap 1 ( ) d
Q
_ Q
h
e
o pr p
1_2 + z2
vr
o cos ap
= h(c
- r)
a2y
t:r
ay
af = g~:r ax (x a) + p(x,t),
0 <X< oo, t
>0
290
of the lower end and the x axis is taken along the equilibrium position
of the chain, pointing upward. The function p(x,t) is the intensity of
the extemal transverse force. If the initial conditions are given by
I. C.:
y(x,O) =O,
iJy
-(xO)=O
iJt
'
= ~o{~
~ 2j~}
where
7.4
f(r)
h(a - r),
2
,_, ,
pF(p)Jo(pr) dp
F(p)
a<O
lsl
2'f[(s
>I
+ 1)/2]
1/r
(1/r)e-"',
e -o' ,
(1/r)sin ar,
-u2,2
1/p
a<O
l/yp2 + a2
a>O
,
a>O
a>O
a>O
h(a -
p)/ya2
p2
rf(r)l0(pr) dr
8
Finite Transforms
8.1
Introduction
8.2
292
O :::; x :::;
1T
f(x) =
2:
bn sin nx,
O<x<1r
(8.1)
n=J
where
bn
n = 1,2,3, ...
(8.2)
where
2 f1T
an = - f(x)cos nx dx,
1T
8.2.1
= 0,1,2, ...
(8.4)
S.,.{f(x);n}
= Fs(n) =
/(x)sin nx dx,
n = 1,2,3, ...
(8.5)
n = 1,2,3, ...
(8.6)
s; I {F.(n);x} = f(x)
F.(n)sin nx,
(8.7)
O<x<1r
n=J
ln this case Eq. (8.7) represents the inversion formula for the inverse
finite sine transform, which is generally attributed to Doetsch. *
The finite sine transform is appropriate for solving differential equations
containing only even-order derivatives. (Recall the discussion in Sec.
(3.1.) Fortunately, many of the DEs of interest fali into this category.
=
1T
293
f"(x)sin nx dx
= f'(x)sin nx~~-
L" f'(x)cos nx dx
The first term on the right is identically zero and a further integration
by parts leads to
S1T{f"(x);n}
-nf(x)cos nx
I~
-r
n
f(x)sin nx dx
= - n 2Fs(n) + n[f(O)
- ( -l)".f(7r)]
(8.8)
Hence, we see that the finite sine transform of f"(x) depends upon the
transform of f(x) and upon the values of f(x) at the boundary points x =
O and x = 'TT'.
Although we have defined the finite sine transform for the particular
intervalO:::; x:::; 'TT', it is easy to generalize to other intervallengths. For
instance, the generalization of (8.5) and (8.7) to an interval of length p,
rather than 'TT', merely involves a scale transformation with x replaced
by 'TT'X/p. This leads to the more general sine transform
SP{f(x);n}
= F.(n) =
n'TT'X
f(x)sin- dx,
o
p
= I ,2,3,...
(8.9)
o< X < p
(8.1 O)
sp I {F.(n);x}
= f(x)
2 ~
. n'TT'x
=-L.. Fs(n)sm - ,
Pn=l
nz'TT'z
n'TT'
(8.11)
Still other generalizations are possible, but we will not discuss them. *
*See Chap. II in R. V. Churchill, Operational Mathematics, 3rd ed., New York:
McGraw-Hill, 1972.
294
C11'{f(x);n}
= Fc(n) =
J:
f(x)cos nx dx,
= 0,1,2,...
(8.12)
O<x<1r
(8.13)
= f(x)
=-
n2Fc(n)
+ ( -1tf'(7T) - f'(O)
(8.14)
Cp{f(x);n} = Fc(n) =
c;t {Fin);x}
P
n1rx
f(x)cos- dx,
p
= 0,1,2,...
(8.15)
O<x<p
(8.16)
- f'(O)
(8.17)
f(x)
= -1 Fc(O) + -2
P
LJ
P n=i
n1rx
Fc(n)cos - ,
P
and
2
CP{f"(x);n}
= - n ~ Fc (n) + ( -l)Y'(p)
p
8.2.3 Applications
The steady-state temperature distribution u(x,y) in a long square bar with
one face held at constant temperature T0 and the other faces held at zero
temperature is governed by the boundary-value problem
Uxx
B.C.:
+ Uyy
0,
u(O,y) =O,
{ u(x,O) =O,
O<x<1r,O<y<1T
U(1T,y) =
u(X,1T) = To
(8.18)
If we. apply the finite sine transform with respect to the variable x, we
arrive at
Uyy - n2 U
B.C.:
= O,
O< y <
where*
U(n,y) =
295
1T
0, n even
U(n,O) = O,
u(x,y)sin nx dx
(8.19)
(8.20)
= 2T0 sinh ny
'
. h
n sm n'TT
n = 1,3,5, ...
(8.21)
4T0 ~
u(x,y) = - L.
1T
sinh ny .
. h
sm nx
n sm n'TT
n=l
(8.22)
(odd)
B.C.:
=a -2 U 1,
O<x<1,t>O
ux(O,t) =O,
I. C.:
ux(1,t)
u(x,O) = f(x),
= O,t>O
(8.23)
O<x<1
U,
I. C.:
(8.24)
where
L
1
U(n,t) =
u(x,t)cos n'TTX dx
(8.25)
*For simplicity of notation, we will drop the s subscript on the transform function.
296
and
F(n)
f(x)cos mrx dx
(8.26)
= c~(n) e -u2n2x,2t
(8.27)
n,t)
= F(O) +
2 2
(8.28)
n=l
However, using (8.26) we can express this solution in the more convenient
forrn
u(x,t) =
2 2 21
EXERCISES 8.2
ln Probs. 1-10, evaluate the finite sine transforrn and finite cosine transforrn
of the given function.
1. f(x) = 1,
0 <X < 1T'
2. f(x) = X,
0 <X < 1T'
3. f(x)
= (x/2)(1T'
- x),
4. f(x)
= eX,
O< x < 4
O<x<1r
O<x<1
7. f(x)
= sin
kx,
0 <X<
8. f(x)
= cos
kx,
0 <X<
1T',
1T',
=f 1,2,3, .. .
k =f 1,2,3, .. .
10. f(x) = (x/6)(x2 -
O< x < p
3x
O<x<1
O<x<l
= UF.(n +
+ F.(n - k)]
k) + Fc(n + k)]
k) - Fc(n + k)]
k) - F.(n - k)]
k)
= ( -l)n-lps(n)
= ( -ltFc(n)
+ 2),
= - n 2Fc(n) + ( -1tf'(7r)
297
- f'(O)
2 2
(b) CP {f"(x);n}
=-n~
p
Fc(n)
+ ( -1tf'(p)
- f'(O)
= - n C"{f(x);n}
C" {f'(x);n} = n S"{f(x);n} -
/(0)
+ ( -l)j(1r)
ln Probs. 16-20, use the finite sine transform or finite cosine transform
to solve the given boundary-value problem
16.
Uxx
a- 2u1 ,
O<x< 1,1>0
= O,
u(l,t) = O
u(x,O) = 3 sin 1TX - 5 sin 47Tx
B.C.:
I. C.:
u(O,t)
17.
B.C.:
u(O,t)
= 10,
I. C.:
u(x,O)
=O
u(lO,t)
= 30
0 <X< p, l > 0
18.
B.C.:
I. C.:
19.
u(x,O)
U 11
c 2Uxx-
u(x,O)
20.
Uxx
Uyy
ux(O,y)
{
u(x,O)
ux(p,t) = O
T0 sin2(7Tx/p)
0 <X< 1, t > 0 (0 < k < 7TC)
2ku,,
u(O,t)
B.C.:
I. C.:
B.C.:
= O,
u..(O, t)
= O, u(l,t) = O
= sin 1TX, u,(x,O) = O
= O,
= O,
O<
ux(1r,y)
= T0 cos x,
u(x,l)
= T0 cos2x
B.C.:
I. C.:
u(O,t) = O,
u(x,O)
u(1r,t) = O
= f(x)
= C- 2U 11
298
8.3
~ 1 . mrb . mrx
mrct
b) Li 2 sm -sm -cos - n=l n
p
p
p
2ep2
= 7T2b( p
Sturm-Liouville Transforms
= O,
> O in a <
= O,
h 2 y(b)
a<x<b
(8.30)
+ y'(b) = O
J:
r(x)<f>n(x)f(x) dx,
n = 1,2,3, ...
(8.31)
where the function r(x) is called a weighting function. We call F(n) the
Sturm-Liouville (S-L) transform of the function f(x), and K(x,n) =
r(x)<f>n(x) is the kernel of the transform. t
For example, the finite Fourier sine transform (Sec. 8.2) is a S-L
transform associated with the Sturm-Liouville problem
y"
=O,
O<x<1r
y(O)
= O,
y(7T)
=o
(8.32)
8.3/Sturm-LiouviUe Transforms
299
= sin nx,
= 1,2,3, ...
(8.33)
=O,
O<x<1r
y'(O)
= O,
y'(1T)
(8.34)
=o
= cos nx,
= 0,1,2, ...
(8.35)
J: r(x)cf>m(x)cf>n(X) dx = O,
f=
(8.36)
L Cnc/>n(X),
a<x<b
(8.37)
n=l
= Cm
I: r(x)[cf>m(.~)f
dx,
= 1,2,3, ...
Solving now for Cm (and changing the dummy index m to n), we find
that
n
= 1,2,3, ...
(8.38)
300
where
= 1,2,3, ...
(8.39)
2:
F(n)
ll<f>ix)ll- 2 <f>ix),
a<x<b
(8.40)
n=l
lt can be shown that when f, f', and f" are all continuous functions over
the closed interval a ::s x ::s b, the series (8.40) converges to f(x) over
the open interval a < x < b (and possibly at the endpoints).
8.3.1
ln addition to the finite Fourier transforms, the next most useful transforms
are those in the category of generalized finite Fourier transforms. These
transforms are associated with Sturm-Liouville problems usually having
the sarne differential equation as in (8.32) and (8.34), but with different
boundary conditions. For example, let us consider the Sturm-Liouville
problem
y" + y = 0,
0<X <b
(8.41)
h>O
y(O) = O,
hy(b) + y'(b) = O,
k2.
n'
= 1,2,3, ...
(8.43)
= F(n) =
= 1,2,3,...
(8.44)
8.3/Sturm-Liouville Transforms
301
2
but since sin knb
111/>ix)ll2
kn
1( 1 b) ,
= 2 b + h cos2kn
= 1,2,3, ...
(8.45)
{F(n);x}
= f(x)
~(
F(n)
) .
= 2n~t b + (1/h)cos2knb sm knx,
0 <X< b
(8.46)
=-
k~ F(n)
(8.47)
B.C.:
u(O,t) =O,
I. C.:
u(x,O)
u(l,t)
= u0 ,
+ ux(I,t)
= O,t>O
(8.48)
O<x<l
it follows that
(8.49)
302
= - k~
T{uxx(x,t);x ~ n}
T{u,(x,t);x ~ n}
U(n,t)
(8.51)
U,(n,t)
and
T{uo;x
~ n} = u
(8.50)
~:(I
fsin knx dx =
cos kn)
(8.52)
U, + a 2k~ U = O, t > O
U(n,O) = (uo/ kn)(l - cos kn)
I. C.:
(8.53)
with solution
U(n,t)
= (uo/kn)(l
COS
kn)e-a k~t
2
(8.54)
U(X,t) -
~ (1 2Uo L.J
n= 1
COS
ki1
kn)sin knX
COS
kn)
-a2k2t
(8.55)
"
EXERCISES 8.3
ln Probs. 1-4, determine the generalized finite Fourier transform defined
by (8.44) of the given function.
x
2. f(x)
1 - x
= h(a - x),
O<a<b
4. f(x)
1. f(x)
3. f(x)
5. Show that
T{f"(x);n}
-k~ F(n)
[hf(b)
+ f'(b)]sin knb +
knf(O)
y =O,
y'(O) =O,
O<x<b
hy(b)
y'(b)
=o
+ y
=O,
hy(O) - y'(O) = O,
O<x<1
y'(l)
=o
303
d
dx (eV) + exy == O,
O<x<l
y(l)
y(O) == O,
=o
9. Use the generalized finite Fourier transform (8.44) to solve the heat
conduction problem (u 1 and u2 are known constants)
O<x<l,t>O
B.C.:
u(O,t)
U~o
I.C.:
u(x,O) =
U1
u(l,t)
ux{l,t)
= u2
8.4
B.C.:
ux(O,t) =O,
I.C.:
u(x,O)
2u(l,t)
+ ux{l,t) = u0
=O
,?
dx (xy')- ;-Y
xy
= 0,
o< X<
b, Jy(O)I
< =, y(b)
==
(8.56)
n = 1,2,3, ...
(8.57)
n = 1,2,3, ...
(8.58)
= F(n) =
J: xf(x)Jv(knx) dx,
= 1,2,3,...
(8.59)
304
J: x[J.,(k,x)f dx
=l
b2 [J.,+,(k,b)f
(8.60)
it follows that the inversion formula for the finite Hankel transform
assumes the form
_, {
2 ~ F(n) J., (k,x)
H v F(n);x} = f(x) = b2 ;!:, [J.,+,(k,b)f ,
0<
< b (8.61)
J: xf'(x)J.,(k,x) dx
= xf(x)J.,(knx)
rb d
Jo f(x) dx [xJ.,(knx)] dx
(8.62)
The first term on the right in (8.62) for v ?: O is zero at both endpoints
[recall Eq. (8.58)], and by using the identity (see Prob. 8 in Exer. 8.4)
[xJ,(kx)]
1(kx)
(v :V 1)k,H,_ {f(x);n}
- (v;:, 1)k,Hv+l{f(x);n},
1
v> o (8.63)
*See Chap. 6 in L. C. Andrews, Special Functions of Mathematics for Engineers (SPIE Press, Bellingham, Wash.; Oxford University Press, Oxford, 1998)
H 0{f'(x);n}
= - Ho{;f(x);n} + knHl{f(x);n}
305
(8.64)
{f(x);n}],
> O (8.65)
and
Hv {xv-y'(x);n}
V> O
-knHv-l {xv-lf(x);n},
(8.66)
(8.67)
rb
= Jo
d
dx [xf'(x)]J0(knx) dx
= xf'(x)J0(knx)
+
1:-
knxf(x)J(knx)
xf(x) [k~J(knx) +
1:
k
..!!.
J(knx)] dx
(8.68)
The first term on the right above vanishes at both endpoints provided
that f'(x) is bounded at x = O. Also, using the identity JMx) = -J1(x)
and the fact that (see Prob. 13 in Exer. 8.4)
(8.69)
k~ J: xf(x)Jo(knx) dx
= knbf(b)J1(knb)
- k~F(n)
(8.70)
+ (1/x)f'(x)
- (v2/r)/(x);n}
= knbf(b)Jv+l(knb) -
k~F(n)
(8.71)
Other versions of the finite Hankel transform are also possible, but
306
we will not discuss them. The interested reader may consult Chap. 8 in
I. N. Sneddon, The Use of Integral Transforms, New York: McGrawHill, 1972.
8.4.2 Applications
Consider the heating of a long circular cylinder of radius b, whose initial
temperature throughout is zero. If the temperature on the lateral surface
is described by the function f(t), the subsequent temperature u(r,t)
throughout the cylinder is govemed by
O< r< b, t >O
u(b,t) = f(t),
u(r,O) =O,
B.C.:
I. C.:
t>O
O<r<b
(8.72)
U, + a2k;U
U(n,O) = O
I.C.:
= a2knbf(t)Jt(knb),
t >O
(8.73)
where
U(n,t) =
J: ru(r,t)J (knr) dr
(8.74)
U(n,t) = a2knbJt(knb)
f('r)e-a
kMt - T)
dT
(8.75)
u(r,t) = :
~1 It(~b)
[L f(T)e-a k~(t2
T)
dT ]J0 (knr)
(8.76)
EXERCISES 8.4
ln Probs. 1-5, find the finite Hankel transform of order zero of the given
function.
1. f(x)
3. f(x)
= 1 - r, 0 <X< 1
= log x, O<x<1
S. f(x)
1,
<X<
2. f(x)
{~:
O<x<1
1<x<2
O<x<1
307
= (bv+l/kn)Jv+l(knb)
d [Xlv(kx)]
dx
v>O
ln Probs. 9-12, verify the given property ofthe finite Hankel transform.
9. Ho{f'(x);n} + H0{(1/x)f(x);n}
10. Hv{(l/x)f(x);n}
knH1{f(x);n}
= J0(x)
v > O
v >O
= knbf(b)lv+l(knb)
- k~F(n)
+ (1/kr)J(kr) + J 0(kr) = O
B.C.:
u(l,t)
=O
I.C.:
u(r,O)
1,0<r<1/2
{ O, 112 <r< 1 u,(r,O)
u(r,t)
where J 0(kn)
O, n
1,2,3, ...
u(l,t) =O
u(r,O)
O, u,(r,O)
O<r<l,t>O
308
show that
u(r,t)
2 ~ sin(k,ct)
LJ k 2 J (k) 10(k,r)
=-
C n=l
= O,
where J0(k,)
1,2,3, ....
B.C.:
I.C.:
u(b,t)
u(r,O)
= a- 2u,,
=O
= f(r)
B.C.:
u(b,t)
I.C.:
u(r,O)
= (1/1))U1,
= u0
=O
+ 71k~U =
U(n,O) = O
U,
I.C.:
-71u0bk,J0(k,b),
t >O
(b) Solve the problem in (a) and show that its inversion leads to
u(r,t)
= - 2uo
b
nsl
J,(k,r) (1 k,Jo(k,b)
e-'IJki~, J,(k,b) = o
O<r<b
a2y
a ay
ar
ax
ax
O<x< 1,t>O
309
(a) If the chain is initially at rest under gravity, show that the horizontal
deflection at any subsequent time is given by
y(x,t)
~ lo(kn yx)
L.J
CTVgn=l
~r
where
q(t)
1r
=2
Jo
p(x,t)J0(kn yx) dx
a2y
cr t2
(a 2y
= 4g cr az2
x = z\
+
z1 ay)
az + p(z ,t)
2
(b) If the chain is released from rest in the position y = e(l - x),
O < x < 1, and swings freely, show that the horizontal deflections
at subsequent times are described by
y(x,t)
~ lo(kn yx)
= Se f:t
,A r
k!Jt(kn) cos(!~~. v g t)
9
Discrete Transforms
9.1
lntroduction
311
9.2
Suppose we are able to sample the continuous function x(t) at the discrete
times t = kT, k = 0,1,2, ... (see Fig. 9.1).* The sampled function
x*(t) then consists solely of the sample values {x(kT)}, obtained through
our sampling procedure. If we idealize the situation by assuming the
sampling is done instantaneously, it is convenient to represent the sampled
function by
x*(t) =
2:
k=
x(t)S(t - kT)
= L
k=
-oo
x(kT)S(t - kT)
(9.1)
-oo
where S is the impulse function (see Sec. 1.5.2). The sampled function
x*(t) is really a traio of impulses in this sense, but is treated as if it were
a continuous function of t through use of the properties of the impulse
function.
ln reality, we cannot obtain an infinite number of samples as suggested
by Eq. (9.1). Hence, in practice we must settle for N samples over a
total time duration NT, and in this case, Eq. (9.1) is replaced by the
finite sum
N-1
x*(t)
x(kT)S(t - kT)
(9.2)
k=O
9.2.1
= f~"" x(t)e-iwt dt
(9.3)
and
x(t)
= -1
27T
f""
X(w)eiwt dw
(9.4)
-00
*We are using the term "continuous" here merely to distinguish the function x(t) from
its sampled version x*(t). Our only real assumption regarding x(t) is that it has a Fourier
transform.
312
-31
-21
-1
21
Figure 9.1 Sampled function
31
41
= w/2TT,
X(f)
= J~"" x(t)e-Z1rift dt
(9.5)
and
(9.6)
Equations (9.5) and (9.6) as transform pairs have a certain appeal because
the constant 1/2Tr has been absorbed. ln our discussion to follow, we
will henceforth convert to these definitions of Fourier transforms in order
to be consistent with the majority of the literature on discrete transforms.
Let us begin by taking the Fourier transform of the sampled function
x*(t) given by (9.2); thus, we obtain
X*(f)
= ~~ f~"" x(k1)(t
- k1)e- -rrift dt
N-1
2:
x(k1)e- 2-rrifkT
(9.7)
k=O
313
= x(kT),
x(k)
X(j)
= X*(j / NT)
2:
X(j)
x(k)e-Z7Tiik/N,
(9.8)
k=O
2:
N-1
e-27Ti(k-m)j/N
= { O,
N,
j=O
:f m
=m
(9.9)
the verification of which is left to the reader (see Prob. 6 in Exer. 9.2).
Now multiplying (9.8) by e27Timi/N and summing from j = O to N -1, we
find
N-1
X(j)e27Timi/N
N-1
N-1
x(k)
k=O
j=O
e-27Ti(k-m>i/N
j=O
= Nx(m)
(9.10)
..!_
N
X(j)e 27Tiik/N,
(9.11)
j=O
Equations (9.8) and (9.11) constitute what are called DFT pairs. ln
some areas of the literature, however, the factor 1/N that appears in
(9.11) is found in (9.8) instead. Thus, once again, we caution the reader
to carefully check the definitions when using properties of these transforms
found in other reference sources.
Example 9.1: Find the DFf of the four-point sequence {x(k)}
and then find the inverse DFT of the result.
Solution: The DFT in this case is given by
3
X(j)
2: x(k)e-i7Tjk/Z,
k=O
= 0,1,2,3,
= {1,1,0,0},
314
which leads to
3
X(O)
L x(k) = 2
k=O
3
X(l)
L x(k)e-i"'k 12 = 1 -
k=O
3
X(2)
L x(k)e-i"'k = O
k=O
3
X(3)
L x(k)e-3iTrk/2 = 1 + i
k=O
x(k)
=!L X(j)eiTrjk/2,
4 k=O
= 0,1,2,3
x(l)
1,
x(2) = O,
x(3) = O
+ N) = x(k),
ali k
(9.12)
X(j
+ N) = X(j),
allj
(9.13)
and
the verification of which we leave to the exercises (see Prob. 8 in Exer.
9.2). To geometrically display this periodicity property, the sample values
are often represented as equally spaced marks around a circle as depicted
in Fig. 9.2.
Other than the periodicity property, the operational properties of the
DFT correlate very closely with the corresponding operational properties
315
of the continuous Fourier transform. For example, the DFT has the shift
properties
N-1
2:
[x(k)e-21Tiik/N] e-21rimk!N
= X(j + m)
(9..14)
k=O
and
N-1
2:
x(k
m)e -21Tijk/N
= X(j)e21Tiim/N
(9.15)
k=O
-1 2:
N
..
N-1
X(j)Y(j)e 21T'Jk/N
2:
= (x * y)(k) =
j=O
x(m)y(k - m)
(9.16)
m=O
~i~
N-1
lXU)iZ =
~o lx(k)!Z
(9.17)
The proofs of these properties, and some additional ones, are left to the
exercises.
(x
* y)(k) =
2:
x(m)y(k - m)
m=O
k=
k=1
Figure 9.3 Convolution
316
* y)(l) = (1)(6)
= 68
* y)(2) = (1)(7)
and
(x * y)(3)
9.2.2
= (1)(8) +
= 60
X(J)
x(K)W1K,
= 0,1, ... ,N -1
(9.18)
K=O
and
N-l
x(K) =
_!_
N
2: X(J)W-
(9.19)
J=O
where
(9.20)
*J. W. Cooley and J. W. Tukey, "An algorithm for the machine calculation ofcomplex
Fourier series," Math. Comp., 19, pp. 297-301, April 1965.
317
in this century to the field of numerical computations. Whereas N 2 operations are required for computing a DFT, the FFT only requires N
log 2N operations, when N is a power of 2. Today there are several
variations of this algorithm, but they are ali based on the sarne principie.
ln order to derive the original form of the algorithm let us assume
that N = 2m, where m is a positive integer. Using base 2, the ndices
J and K can be represented by
J
+lo= (Jm-IJilo)
(9.21)
and
(9.22)
where each li and Ki is either zero or one. For example, the number
121 in base 2 is represented by
121 = 26 1
+ 25 1 + 24 1 + 231 + 22 0 + 20 + 1 = (1111001)
and
X(J)
= 2: ... 2:
Ko
x(Km-IKo)WJ(2m-lKm-l+...+Ko)
(9.23)
Km-l
Certain simplifications can now take place by making use of the following
theorem and its obvious generalizations.
e-ZTTi/N
Proof: Consider
WJ2m-lKm-l
=1
exp(- 1Tizm-zlm-zKm-1) = 1
318
WJo2m-IKm-l
= 2 x(Km-IKo)WJo2m-IKm-l
Km-1
.,. J 1K m-3 K o) -A 2( Jo
Km-p
.,. J ) _ ""A
(.,. J
K )W(2m-1Jm-I+ ... +Jo)Ko
A m (Jo
m-1 - L.J m-1 Jo m-2 O
Ko
(9.25)
The last result Am(J0 Jm_ 1) is our desired output, but in bit-reversed
order; that is,
(9.26)
This bit-reversal is an inherent property of the algorithm. To obtain the
output in proper order, the input could first be scrambled. Either way,
the scrambling of input or output data is simply part of the entire FFT
process.
Example 9.3: Use a four-point FFT to compute the Fourier transform
of
{x(K)}
= {1,2,3,4}
AI(JoK0)
x(K 1K 0 )W 21oK1
K1=0
I
A2(Joll)
AI(JoKo)W2J1Ko+JoKo
Ko=O
X(Jilo)
where W
= e -i'IT/2 =
i.
= A2(JoJ1)
319
Step 1:
I
A1(00)
x(K10)
= x(OO) + x(lO) = 1 + 3 = 4
x(K11)
= x(01) + x(11) = 2 + 4 = 6
K1=0
I
A1(0l)
L
K1=0
I
A 1(10)
= x(OO)-
x(lO)
= x(Ol)-
x(ll)
= 2-
1- 3
-2
-2
K1=0
I
A1(11)
L x(K11)W
2
K'
K1=0
Step 2:
I
Az(OO)
A1(0K0)
= A1(00) + A1(01) = 4 + 6 = 10
Ko=O
I
Az(Ol) =
Ko=O
I
Az(IO)
Az(ll)
= A1(10)
A10Ko)W 3Ko
- A1(11)i
-2- 2i
Ko=O
Step 3:
X(O)
X(l)
= X(OO) = A2(00) =
= X(01) = A2(10) =
10
-2 + 2i
= X(lO) = A 2(01) = -2
X(3) = X(ll) = A2(11) = -2- 2i
X(2)
= {10,
EXERCISES 9.2
ln Probs. 1-4, find the DFf of the given four-point sequence and then
find the inverse DFf of the result.
1. {1,2,3,4}
2. {0,1,0,-1}
3. {1,2/3,1/3,0}
4. {1 ,2, 1,2}
320
= { O,
k =f m
k =m
N,
j=O
+ y(k)]e-Z1rijk/N = X(j) +
Y(j),
k=O
N-1
(b)- ~ [X(j)
Y(j)]e 21Tijk/N
= x(k) + y(k),
= 0,1, ... ,N -1
j=O
all k
allj
(a)
[x(k)e-21TijkfN]e-21Timk/N
= X(j + m)
k=O
N-1
(b) ~ x(k
+ m)e-21Tijk/N = X(j)e21Tijm/N
k=O
(a)
N-1
= ~ x(m)y(k
- m)
N j=O
m=O
(b) From (a), deduceParseval's relation
N-1
~ j~ IXU)I
N-1
2
~o lx<kW
11. Verify Parseval's relation for the sequences given in Exam. 9.1.
12. A sequence {x(k)} is said to be even if x(N - k) = x(- k) = x(k).
We say the sequence is odd if x(N - k) = x(- k) = - x(k). Prove
the following properties concerning even and odd sequences involving
DFT pairs:
(a) {X(j)} is even if and only if {x(k)} is even.
9.3/The Z Transform
321
15. {1,2/3,1/3,0}
9.3
The Z Transform
L x(nT)8(t -
nT)
(9.27)
n=O
where T denotes the time interval between samples. ln developing properties of the Z transform, we normally assume that the sampled function
consists of an infinite number of samples as indicated in Eq. (9.27). Of
course, in practice there may be only a finite number of these sample
322
values that are actually available for processing. The Laplace transform
of the above sampled function formally leads to
;t'{x*(t);p}
= ~ L~ e-pt x(nT)8(t
- nT) dt
n=O
L x(nT)e-pnT
(9.28)
n=O
z = epT
(9.29)
= ;t'{x*(t);p}
of the Laplace
z plane we have the
CT
= X*(p)
2: x(nT)z-n,
(9.30)
n=O
We say that X(z) is the Z transform of x(nT), but which we will also
call the Z transform of the continuous function x(t). Using the principie
of analytic continuation, we can deduce that X(z) is an analytic function
of z outside the circle lzl = errr. The series on the right in (9.30) is the
Laurent series expansion of X(z) about the origin. Because the spacing
T between samples has no effect on developing properties and on the
use ofthe Z transform, it is conventional to set T = 1. Also, corresponding
to the notation ;t'{f(t);p} = F(p) that is used in the Laplace transform,
we introduce the similar notation
Z{x(t);z}
= X(z) =
2: x(n)z-n
(9.31)
n=O
1.
9.3/The Z Transform
9.3.1
323
Evaluating Z Trans/orms
+ z- 1 + 5z- 3
X(z) = 2
1, find Z{l;z}.
= 0,1,2, ...
Hence,
00
Z{l;z}
L z-n = 1
lzl > 1
-1'
n=O
Z{1 ;z}
- 1),
lzl > 1
= 0,1,2, ... ,
so that
= z/(z
=I=
O.
an, n
Z{a 1 ;Z} =
anz-n
n=O
(a/zt
n=O
or
Z{a 1 ;z}
= z/(z -
lzl >a
a),
Z{t;z}
nz-n
n=O
=z
2:
nz-(n+ I)
n=O
00
""
-n
-z-L.JZ
dz n=O
324
= z/(z -
1)
lzl >
+ l);z}
1/f(n
1)
oo
Z{1/f(t
+ l);z} =
-n
2: ~
= e 11z,
n.
all z
n=O
9.3.2
+ C2 y(t);z}
= CtX(z)
+ C2Y(z)
(9.32)
+ l);z} =
2: x(n + l)z-n
n=O
= z 2:
x(n + l)z-<n+ll
n=O
=z
2: x(n)z-n
n=t
(9.33)
9.3/The Z Transform
325
= z-a X(z)
(9.34)
L nx(n)z-n
n=O
L
00
- zx(n)z-n
dz n=O
= - z X'(z)
(9.35)
= X(eaz)
(9.36)
X(z)
L CnZ-n
(9.37)
n=O
n = 0,1,2, ...
(9.38)
z+3
z- 2
*Because the uniqueness of the Z transform and its inverse extends only to the sample
values {x(n)} and not to x(t), we will consider {x(n)} as the inverse Z transform of X(z)
rather than x( t).
326
Solution: By writing
z
3
= -- + --
X(z)
z-2
z-2
it follows from Exam. 9.5 and the second shift property (9.34) that
x(n)
= z- 1{X(z);n}
= r + 3. 2n-lh(n
- 1)
n=O
1,
== { 5r- 1,
n = 1,2,3, ...
= z
4z3
9l + 5z sl + 5z -
Solution: Since the numerator is of the sarne degree as the denominator, we first divide and then apply a partial fraction expansion to
obtain
2
!+
X(z) =
Therefore, with the aid of entries 13, 15, and 5 in the table and the
second shift property, we deduce that
9(1)n-l
2 h(n-1)
~h(n) + [ 03 :
5
n)Gr-
~]h(n-1).
n=0,1,2, ...
z3
4z3
9l + 5z 8z2 + 5z -
1
1 7 _1
1 = 4- 4z
41 _2
- 16z
9.9 is to
+ ...
from which we conclude x(O) = 1/4, x(l) = -7/4, x(2) = -41/16, ....
This method, of course, provides no general formula for x(n) but it does
eliminate the need to factor the polynomial in the denominator.
9.3/The Z Transform
327
z- {X(z)Y(z);n} = (X*Y)(n)
1
(9.39)
where
(X*Y)(n)
2: x(k)y(n -
(9.40)
k)
k=O
Example 9.10:
Use the convolution theorem to find the inverse Z transform of the product
(z - 2)(z - 3)
Solution:
By setting
X(z)
= - z- ,
Y(z)
z- 2
= -z-
z-
z- {X(z);n} = x(n) = 2n
z- {Y(z);n} = y(n) = 3n
1
k=O
= 3n
(~)k
k=O
Co
CJZ-
+ c 2z- 2 + +
CnZ-n
+ ...
CnZ-1
+ ...
CoZn-1
CJZn-2
+ ... +
328
lzl = R such that X(z) is analytic on and outside the closed contour. On
the right ali terms will vanish except the one involving z-I, and we are
left with the inversion formula
Cn
f
2~; izi=R
n = 0,1,2,...
X(z)zn-J dz,
= ako
(9.41)
x(n)
= cn =
L Res{X(z)zn- 1;ak},
= 0,1,2, ...
(9.42)
k=l
Example
=z+
z-
X(z)
Solution:
3
2
= (z + 3)zn-J
z-
= Res{O} + Res{2} =
x(n) = Res{2} = 5
zn-
-23 + 25 = 1
= 1,2,3, ...
EXERCISES 9.3
ln Probs. 1-10, evaluate the Z transform of the given function.
2.
3. cos bt
4. sin bt
S. h(t - 1)
6. h(t) - h(t - 1)
7. h(t - 1) - h(t - 2)
s. oI t)h(t
9. h(t - k),
1,2,3, ...
1)
10. a 1- 1 h(t - 1)
z-ax(z),
a >O
9.3/The Z Transform
329
= X(eaz)
Z{e-a1X(t);z}
l+1
15. -2--1
16. -2--1
z -
z +
17.
(z
+ l)(z +
18.
3)
(z
l +1
19. (z
- 1)3
20. (zz
z
+ l)(z +
3)
+ l)(z -
2)
z- 1{X(z)Y(z);n} = L x(k)y(n
- k)
k=O
where X(z) and Y(z) are, respectively, the Z transforms of x(n) and
y(n).
= x(O)
z-->oo
23
_z_
"l+l
25.----(z + l)(z + 3)
l+l
24. -2--1
z -
26.
(z
28
30
(i +
i -
l)(z
3)
1)(z - 2)
zsinb
2z cos b + 1
330
9.4
Difference Equations
= 1T->0
1m
y(t
+ T) T
y(t)
= y(t +
(9.43)
1) - y(t)
+ 1) - y(t)]
+ 2) - y(t + 1)]
[y(t
= [y(t
- [y(t
1) - y(t)]
or
2
y(t)
= y(t + 2)
- 2y(t
1)
y(t)
(9.44)
L YnZ-n
(9.45)
n=O
= Y(z), then
z[Y(z) - Yol
= z2[Y(z) - Yol - ZY1
lf Z{yn;z}
(a) Z{Yn+I;z}
(b) Z{Yn+z;z}
9.4/Difference Equations
331
n=O
n---i>n-1
n=l
= Z [~
YnZ-n -
= z[ Y(z)
- Yo]
Yo]
n=O
Similarly,
n=2
(9.47)
Replacing y' in (9.46) with the difference (9.43), we get the corresponding
difference equation
Ay- Yn =O,
=3
(9.48)
Yo = 3
(9.49)
Yo
or
Yn+l - 2yn =O,
332
= 3z/(z
Y(z)
- 2)
= 3. zn,
Yn
(9.50)
One of the things that this simple example is illustrating is that the
"natural base" for exponential functions in difference calculus is 2 rather
than e as in ordinary calculus. There are several other correspondences
of this nature that become evident by pursuing how the di:fference operator
ll works on various types offunctions. However, here we will not develop
such correspondences since our primary interest is in solving difference
equations. t
Example 9.12: Use the Z transform to solve the difference equation
Yn+2
Yo = 1,
Yt = 2
or
(z
3z
+ 3z[Y(z)
- 1]
+ 2Y(z) = O
+ 2)Y(z) = z2 +
5z
l + 5z
= (z + l)(z +
4z
z+
2)
3z
z+2
= 4(-lt-
3(-2t
= <- 1r[4 -
3 zn1.
= 0,1,2, ...
EXERCISES 9.4
ln Probs. 1-10, use the Z transform to solve the given di:fference equation.
= O;
= O;
Yo = O,
Yo
= O,
Yt =
Yt
on the other hand, the Z transform applied directly to (9.47) yields Y(z) = 3z/(z - e).
tFor more discussion of finite differences, see K. S. Miller, An lntroduction to the
Calculus of Finite Differences and Difference Equations, New York: Dover, 1960.
9.4/Difference Equations
333
= O; Yo = 1, Y1 = O
+ 4yn = O; Yo = 1, Y1 = 4
+ 2yn = O;
Yo = O, Y1 = 1
+ 6yn = 4n;
Yo = O, Y1 = 1
+ lOyn = 16n;
Yo = 6, Y1 = 2
+ 6yn = 2n + 1; Yo = O, Y1 = 1
- 5yn = 24n - 8;
Yo = 3, Y1 = -5
Yo = O, Y1 = O
4. Yn+z - 4Yn+l
S. Yn+2 - 2Yn+l
6. Yn+2 - 5yn+l
1. Yn+2 - 1Yn+l
8. Yn+2 - 5yn+l
9. Yn+2 + 4Yn+l
= y(t +
3) - 3y(t
Z{Yn+3 ;z}
= l[Y(z)
- Yo] - Z Y1 - ZY2
2)
+ 3y(t +
1) - y(t)
ln Probs. 14 and 15, use the result ofProb. 13 to solve the given difference
equation.
2yn
= O;
Yo
= n + r;
2
O, Y1
Yo
= O,
1, Y2
Y1
1, Y2
9.5
Table of Z Transforms
334
= 2: x(n)z-n = X(z)
n=O
No.
x(t),
~o
n = 0,1, ...
x(n),
z
z- I
h(n)
z(z + I)
(z - 1)3
a',
ta',
cos bt
8
9
a> O
z
z- a
a"
a > O
na"
cos bn
sin bn
sin bt
cosh bt
cosh bn
sinh bt
sinh bn
e'"'cos bt
11
e''sin bt
12
h(t - 1)
h(n -
13
h(t) - h(t - I)
h(n) - h(n -
14
h(t - k),
h(n -
z2
z2
z2
z sinb
2zcosb +I
z(z - cosh b)
2zcoshb + 1
z sinh b
2z cosh b
ze'sin b
k = 1,2, ...
1)
k)
z-
I)
(z- I)
15
17
z(z - cos b)
2z cos b + I
z2
z- a
1
z(z - e'cos b)
10
16
(z-
X(z)
-h(t- 1)
t
-h(n-1)
f(t + 1)
n!
log-z- 1
Bibliography
336
Bibliography
Appendix A
Review of
Complex Variables
=X
+ iy
(A.l)
where x and y are real variables and i = y'-=1. The variable xis called
the real part of z, denoted by x = Re(z), and y is called the imaginary
part of z, also written as y = lm(z).
If f is a function depending on the complex variable z, we say that
f is a complex function. Such functions can always be represented in
the form
f(z)
= u(x, y) + iv(x, y)
(A.2)
au
av
au
ax
ay'
ay
-=-
av
ax
(A.3)
337
338
Appendix A
at each of its isolated singular points, the value of which rnay be zero.
ln general, the value of the residue is the value of the integral
1
- . i, f(z) dz
27Tl Yc
around any closed contour containing the isolated singularity. Multiplevalued functions have singularities that are called branch points. For
example, the function f(z) = z 112 has a branch point at z = O.
One kind of isolated singular point is called a pole. If f(z) is not
finite at sorne point z = a, but the product (z - a)mf(z) is analytic for
sorne integer m, we say that f(z) has a pole of order m at z = a. A
pole of order one is also called a simple pole. The residue for a function
having a pole of order m at z = a is given by
1
dm-1
Res{f(z); a} = ( - 1)' lirn d m-I [(z - a)mf(z)]
(A.4)
m
. z-->a z
If f(z)
Res{f(z); a}
a, then also
p(a)
q'(a)
(A.5)
Appendix A
339
= 27Ti
i-,
Res{f(z); ak}
JcR
z=
=o
f(z)dz
Iim
f(z)dz
p--.o Jc.
= o.
Jc"
f(z) dz
= ai Res{f(z); a}
JcR
eim'f(z) dz
=o
(m >O)
JcR
em'f(z) dz
=O
(m
> O)
AppendixB
Table of
Fourier Transforms
No.
f(t)
I
= yl2;
21T
r.
-=
e-'"F(s)ds
I
= -yl2;
F(s)
r.
-=
e'''f(t )dt
y21r (s)
-I
tz + az
i sgn(s)
+ az)z '
f- a 2
t(f + a 2 )'
iul
e-11,
te-11,
340
a>O
J;i
a>O
is e-11
- J'IT2 2a
a>O
I -11
--e
2a
a>O
a>O
~;a
sz + az
(s2
2ais
+ az)z
l)sgn(s)
Appendix B
No.
f(t) =
ltle-lrl,
.r .
y'2;
21!'
e-"'F(s)ds
t!"f(t)dt
-00
ji
2
2
a -s
1T (s2 + a2)2
a>O
-u2t2
y'2;
21!'
F(s) =
-oo
a>O
-s2/4a2
10
II
cos(t2/2)
~ [cos(s2/2) + sin(s2/2)]
12
sin(t2/2)
~ [cos(s2/2)
13
e-ltllv'i[cos(at/VZ>
'
Via e
+ sin(alti/VZ)],
a>O
14
_ , sin t
11 e
t
15
sgn(t)
- sin(s 2/2)]
2a 3
y; s4 + a4
I
y'2; arctan(2/s 2)
211'
ji!
'ITS
16
t sgn(t)
-/,!
17
h(l - ltl)
ji sins
18
(I - ltl)h(l - ltl)
I ( sin s/2r
\/211'
s/2
19
h(t)
20
8(t - a)
--e'"'
21
Jtj-a,
22
ltl-a sgn(t ),
23
P.(t )h(l -
1T s2
1T
[a(s)
~J
y'2;
O<a<l
O<a<l
Iti)
/!;.
)2
:;;
f(l - a) . 'l!'a
isl'-a
f(l - a)
StnT
'll'a
JsJ' a COST
i"
v~Jn+I/2(S)
341
342
No.
Appendix B
f(t)
=A r
Fc(s)cos st ds
J;r
Fc<s> =
f(t)cosstdt
Y2Tr 8(s)
2
Vi
Vs
P + a2'
e -at ,
te-a',
e- cos at,
J"!!
e-
2a
a>O
A
A
a a2
s2 +
a>O
-a212
a2-s2
(s2 + a2)2
a>O
a> O
e- sin at,
cos(f/2)
lO
sin(f/2)
ll
t-1'
12
No.
(a 2
as
+ 4a 4
[cos(i/2)
/2f(p)
; 7COS(7rp/2)
=A r
2- 112f(p + l/2)(a/sYJ.(as),
p > -l/2
Fs(s)sin st ds
Fs(s)
= J!;.
J~ sgn(s)
l
Vi
Vs
f + a2'
+ sin(s2/2)]
O<p<l
0
0
l
2
+ 4a
A2a
s4
a>O
a>O
/(t)
j"'J. as2 + 2a
Tr
_,2,..2
aVi
a>O
'
a>O
J~ e-
/(t)sin st dt
Appendir B
Table 8.3.
==A r
(continued)
Ar
Fs(s)sin st ds
Fs(s) =
+ a2)2,
a>O
1 -Se -as
-yZ; a
t(t2 + a 2)'
a>O
-(1 - e-as)
fz1
a2
e -ai ,
te-a',
No.
f(t)
(t2
JI
JI
s
:;;: s2 + a2
a>O
a>O
7T
te -a2t2'
1 -ar
-e
t
f(t)sin st dt
2as
(s2 + a 2f
_s_ e-s2;<a2
2yZa 3
a>O
J!
JI
JI
fz 7
arctan(s I a)
lO
a>O
ll
a>O
12
tp-1/2 ,
13
:;;: s + 4a4
2
2a s
4
:;;:s +4a 4
4
f(p) .
S10(7Tp/2)
O<p<1
a>O
2p-3/2aPsl-pf(p - l/2)Jp(as ),
p > 1/2
343
Appendix C
Table of
Laplace Transforms
No.
F(p)
e -p'f(t)dt
f(t)
1 f+i>O
=~
. eP'F(p)dp
7Tl c-=
-1
p
pi
-p"1
(n
= 1,2,3, ... )
r-1
(n-
v'P
\(;i
2\/t/rr
PJ/2
,.,-I
x>O
p;'
1
-p-a
e"'
1
(p- a)2
te"'
344
1)!
f(x)
Appendix C
No.
9
10
=r
F(p)
(n
(p - a)"
f(t)
e-P'f(t)dt
=1, 2, 3, ...)
12
13
14
15
16
17
18
19
20
21
22
23
24
25
(p - a )(p - b) '
p
pz
a~ b
a~ b
1n-!eat
ebt
a-b
ae"' - beh'
a-b
1
- sin at
+ az
cos at
+ az
Pz _ az
. h at
-1 sm
p
pz _ az
cosh at
(p - a)z
!b e"'sin bt
+ bz
p-a
(p - af
eP'F(p)dp
f(x)
p
Pz
1 c-i~
(n - 1)!
eat -
11
1x-leat
x>O
(p- aY'
f+'~
=-.
27T
e"'cos bt
+ b2
p(pz
+ az)
20 - cos at)
pz(pz
+ az)
~(at
sin at)
1 .
)
a (sm at - at cos at
2 3
(pz
+ a2)z
(pz
+ az)2
a sin at
2
p2
(pz
+ az)z
1a (sm
. at
2
(pz
+ az)(pz + bz)
P4
+ 4a4
az ~ bz
+ at cos at)
1
;;---z<cos at - cos bt)
-a
4~ 3 (sin at cosh at -
345
346
No.
Appendix C
F(p)
/(t)
Lc+i~
= -'1TI.
. eP'F(p)dp
c-IQO
. h
26
a SID at SID at
2 2
27
2~3 (sinh at -
28
29
10(at)
30
31
a> O, v> O
yP(p- a)
32
33
34
sin at)
8(1 -
l
-e-op
p
,
a>O
35
a)
h(t - a)
(t - a)h(t - a)
a>O
36
r= cos(2y'at)
y7Tt
37
38
_I_
-a/p
PVPe
.!_
p "e
-a/p
'
,
a>O
. c= sin(2y'at)
yTTa
a> O, v> O
f)(v-ll/2
(~
_a_
39
J,_,(2y'at)
e-a2f4t
2v:;f
40
41
1
- e-v'p
p
,
42
arctan(a/p)
a~
erfc(a/2yi()
sin at
Appendix C
Table C.l. (continued)
No.
F(p)
==r e-'f(t)dt
- arctan(a/p)
p
44
e"
45
- e" P erfc(as ),
p
46
eP erfc( v'ap),
47
48
erf(a/yp)
49
VP1 e"1Perfc(v'aJP),
2 2
erfc(as ),
S(at)
_I_ e-2t4a2
a
v;,
a>O
2 2
27TI c-=
43
f+l=
1
f(t) = - .
a2=0
erf(t/2a)
Va
a>O
1TVt(t + a)
a2=0
y:;,(t +a)
_!_ sin(2aVt)
1TI
a2=0
I e-2,,i;;,
v'1Tt
e'"F(p)dp
347
Index
349
350
Index
applications, 330-333
fast Fourier, 316-319
Fourier, 3ll-319
inverse, 313, 328
properties of, 315-316, 324-325
sampling continuous functions, 3ll
table of, 334
z, 321-334
Doetsch, G., 292
Double Fourier transform, 98-101, 121,
274-275
Dual integral equations, 287
Duhamel's principie, 122-123
Duplication formula, 13, 15, 66
Eigenfunctions, 108, 298-299
Eigenvalues, 108, 298-299
Elasticity, 151-155
Airy stress function, 152
compatibility condition, 152
equations of motion, 151
equilibrium, 152-154
stress components. 151, 153-154
Elliptic integral, 83
Energy spectrum, 79
Entire function, 16
Erdelyi, A., 98
Error function, 16-21, 124, 182-184, 187188, 200, 210, 214, 216, 220, 224,
233-234, 236
complementary. See complementary error function
graph of, 17
properties of, 16-17
Euler formulas, 49, 126, 233
Euler, L., 7, 23, 106, 151
Exponential integral, 180, 216, 261
Exponential order, 166
Fast Fourier transforms, 316-319
Fibonacci sequence, 333
Final value theorem,
Laplace transform, 212
Z transform, 329
Finite transforms, 291-309
applications, 294-296, 306
cosine, 294
Fourier, 291-298
generalized Fourier, 300-301
inverse, 292, 294, 300, 302, 304
Hankel, 303-309
sine, 292-293
Sturm-Liouville, 298-309
Fluid ftow. See Hydrodynamics
Fourier integral representations, 38-49,
106-107
Index
cosine, 43-46
sine, 43-46
Fourier integral theorem, 40
exponential form of, 50
proof of, 47-49
Riemann-Lebesgue lemma, 47
Fourier, J., 37, 113
Fourier series, 38, 299
cosine, 292
generalized, 299
sine, 292
Fourier's Iaw, 113
Fourier transform, 2, 49-161, 208
applications, 102-161
convolution integral, 78-85, 99-100,
115, 117, 129, 137-138, 149
cosine, 51-53
definition of, 2, 50
discrete, 311-334
fast, 316-319
inverse, 50, 59
method of stationary phase, 97-98
multi pie, 98-101, 121, 274-275
properties of, 58-65
residue methods, 67-74
sine, 5!-53
table of, 340-343
Fredholm integral equation, 108
Fresnel integrais, 17-21, 58, 131, 189, 234
graph of, 18
series representation of, 20
Gamma function, 7-16, 66, 166, 170, 247,
250-251, 256-259, 264
argument negative, 8
graph of, 9
poJes of, 8
properties of, 13
Gaussian random variable, 20, 157
Generalized finite Fourier transform, 300301
Generalized functions, 85-91
See also Impulse function
Goodier, J. N., 152
Gradshteyn, I. S., 159
Green's function, 226
Hankel transform, 3, 274-290
applications, 285-290
definition of, 3, 275
finite, 303-309
inverse, 275
properties of, 278-282
table of, 290
Harmonic function, 132
Harrington, W. J., 259
Heat equation, 4, I 13-125, 229-231
351
axisymmetric, 288
finite interval, 230-231, 295-296, 301302
infinite line, 114-117
infinite rectangle, 121
semiinfinite tine, 118-120, 229-230
Heaviside expansion theorem, 205
Heaviside, 0., 162
Heaviside unit function, 29-31, 41, 66,
88-89, 104, 172-173, 222, 232, 236,
276, 328
Helstrom, C. W., 157
Hilbert-Schmidt kernel, 108
Hilbert transform, 91-97
Hydrodynamics, 141-151
continuity equation, 141
equation of motion, 141
ideal fluid, 141
incompressible flow, 141
irrotational flow, 143-144
sources and sinks, 141
steady flow, 142, 147-149
stream function, 142
surface waves, 144-147
two-dimensional flow, 141
viscous fluid, 141
vorticity, 141
Impulse function, 31-36, 85-91, 122, 150,
189, 226, 232, 236, 311, 321-322
definition of, 32
sifting property of, 33
Impulse response function, 224-227
Influence function, 196
Initial value problems, 221-228
Initial value theorem,
Laplace transform, 211
Z transform, 329
Integral equations, 4-5, 107-112, 238-244
convolution type, 108, 238-244
dual, 287
Fredholm, 108
Hilbert-Schmidt kernel, 108
of the first kind, 107, 238-239
of the second kind, 108, 240
Volterra type, 108, 238-244
Inversion formula, 3
cosine transform, 51
discrete Fourier transform, 313
double Fourier transform, 98
finite cosine transform, 294
finite Hankel transform, 304
finite sine transform, 292
Fourier transform, 50, 59
generalized finite Fourier transform, 301
Hankel transform, 275
Laplace transform, 164-201
352
Index
Index
Mellin transform, 254-262
Z transform, 327-328
Residue theorem, 338-339
Riemann, G., 263
Riemann-Lebesgue lemma. 47
Riemann zeta function. See Zeta function
Robin's condition, 114
Ryzhik, I. M., 159
Sampled function, 311-312, 321
Scaling property of
Fourier transform. 60
Hankel transform, 278
Laplace transform, 171
Mellin transform, 248
Schwartz, L., 85
Self-reciprocal function, 56, 278, 284
Shift properties of
discrete Fourier transform, 315
double Fourier transform, 99
Fourier transform, 60-61
Laplace transform, 171-173, 191
Mellin transform, 248
Z transform, 324-325
Shivamoggi, B. K., 141
Signum function, 34, 52. 88, 90-91, 110,
148-149
Sine integral, 43, 177, 216
Sine integral representation, 43-46
Sine transform, 51-53, 63, 119-120, 208
inverse, 51
Singularity,
branch point, 338
isolated, 338
poJe, 338
Sneddon, I. N., 213, 275, 276, 287, 306
Special functions, 6-36
beta function, 15
Bessel functions, 21-29
cosine integral, 180, 261
elliptic integral, 83
error functions, 16-21
Fresnel integrais, 17-21, 58, 131, 189,
234
gamma function, 7-16
Heaviside unit function, 29-31
impulse function, 31-36
Laguerre polynomials, 181
Legendre polynomials, 175
rectangle function, 30
353
signum function, 34
sine integral, 43, 177, 216
Spectrum, 86
Springer, M. D., 265
Stationary phase method, 97-101, 146-147
Statistical moments, 156
Stream function, 142
Sturm-Liouville problem, 298, 300, 303
Sturm-Liouville transform, 298-309
applications, 301-302, 306
finite Hankel, 303-309
generalized finite Fourier, 300-301
inverse, 300, 304
Tautochrone problem, 240-242
Taylor, B., 37
Thompson, W. E., 265
Timoshenko, S., 127, 152
Titchmarsh, E. C., 40
Translation property of
Laplace transform, 172, 232
Mellin transform, 248
Tukey, J. W., 316
Two-sided Laplace transform, 214-216
Unit step function. See Heaviside unit
function
Vibrating
beam, 127-129, 233-234
membrane, 287-288
string, 125-127, 231-232
Volterra integral equation, 108, 238, 241
Watson, G. N., 264
Watson's lemma, 213-214
Wave equation, 4, 125-131, 231-234
infinite line, 125-127
semiinfinite line, 231-232
Whittaker, E. T., 264
Widder, D. V., 216
Zemanian, A. H., 85
Zeta function, 253, 263-264, 268-269
properties of, 264
Z transform, 321-334
applications, 330-334
inverse, 325-328
properties of, 324-325, 329
table of, 334
Integral Transforms
for Engineers
Contents: Special functions. Fourier integrals and Fourier transforms. Applications involving
Fourier transforms. The Laplace transformation. Applications involving Laplace transforms.
The Mellin transform. The Hankel transform. Finite transforms. Discrete transforms.
Bibliography. Appendix A: Review of complex variables. Appendix B: Table of Fourier
transforms. Appendix C: Table of Laplace transforms. Index.
P.O. Box 10
Bellingham, WA 98227-0010
ISBN-10: 0819432326
ISBN-13: 9780819432322
SPIE Vol. No.: PM178
Larry C. Andrews
Bhimsen K. Shivamoggi
Larry C. Andrews and Bhimsen K. Shivamoggi are professors of mathematics at the University
of Central Florida. Andrews is also a member of the Department of Electrical and Computer
Engineering and associate member of the Center for Research and Education in Optics and
Lasers (CREOL) and the Florida Space Institute. Shivamoggi is also a member of the
Department of Physics at U.C.F.
Larry C. Andrews
Bhimsen K. Shivamoggi