Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

Chapter 2Q

Fourier Integral, Fourier


Transforms and Integral
TransforJDs

INTRODUCTION ~ i d e r the Fourier series expansion of


f(x) many mterval [- L, L] given by
Just as the Fourier series decomposes a periodic func-
uon into a discrete set of contributions of various .✓ . 00

frequencies (all multip les of one fundamental fre-


' f(x ) = ao + ~ ( an cos -nrrx + bn sin -nrrx) (2)
~
2 n= I L L
quency), the Fourier transform provides a continu-
ou~ frequency resolution of a (possibly nonperiodic)
!unction. Fourier transform is useful in the study of
where an = -l fL
L
- L
- t ) dt
f (t) cos (nrr
L
(3)

= 2_ f CJ!L t) dt
frequency response of a filter, solution of PDE, dis-
bn L f (t) sin (4)
crete Fourier transform and Fast Fourier transform L -L
msignal analysis.
Two more integral transforms Hankel transforms* By substituting the coefficients an and bn from (3)
useful in problems involving Bessel functions and and (4), the Fourier series (1) takes the form
Hilben transform are considered.
20.1 FOURIER INTEGRAL THEOREM
f(x) = -1 JL f(t)dt +
-- 2L -L
A periodic function f (x) defined in a finite interval
(-L. L) can be expressed in Fourier series. By ex- +L L [f
1
00
L
-L J(t)cos Lt ) dt ] cos nrrx
(nrr L
n=l
tending this concept, non-periodic functions defined
+L L [f
00
m-x < x < 00 (for all x) can be expressed as a 1
-L
L
J(t)sin L dt ] sin (nrrx)
(nrrt) L
Founer integral, since in practice periodic functions n=l
are fairly rare.
Theorem: A function f(x), which is piecewise = 2~ f_LLJ(t)dt + L?;
1 1L-Lf(t) [cos (nrrr)
00

T
: 0 ntinuous in every fini te interval and is absolutely

integrable on the x-axis can be represented by a n:rrx) (n:rrt) . (mrx)]d t


Fou ·
rier integral
' xcos ( L +sin L -sm L
f(x) == 1')0 [A(a) cos ax + B (a) sin ax] da (1)
v.:hich i . o . .
At s valid at all points of contznu, ty. .
int a Point of discontinuity xo, the Fourier
ag?;al === ![
j(xo - 0) + f (xo + O)] i.e., aver-
0
th e left and right hand limits.
20.1
>---
1\va,]ab]c
on our Web Sile http://www.mhhe.co
m/rarnanahelD
.
PW

20.2 - v...--------..--~~-
HIGHER ENGINEERING MAlHIMAncs--;;
II Jr
Put a 11 = - nnd 6.a11 = a11 + 1 - ct 11 Then.
the
.
. d
.. l ( l) rcuccstoh
Fourier itlte g 1ct
L cos tne tnlcgral l C Fourier
j[ II Jr Jf
= (11 + 1) -L - -
L
= -
L
(6) 00

f (x) =f J\(a) cos(ax)cla


Jr0 (l3)
Then
Fourier Sine Integral
J(.r) = -;,- !LI
f (r)dr + When .f (x) is odd function , then A(a) = 0 and
_L . - L

f = ~ fo
00

+ !_ L dr] 6.a
N [ L B(a) .f(l) sin(ett)clt (l 4J
f(r) cos (a11U - .r)) 11 (7)
;r 11 = l - L
Then the Fourier integral ( 1) reduces to the Fourier
As L ➔ oo. t
➔ 0 and 6.a 11 = ➔ 0, the infinite I sine integral
00
series in (7) becomes an integral from O to oo. Thus f(x) = fo B(a ) sin axda (ISJ

.f(x) = -1
Jf
lN[Joo
O -00
.f(r) • cos (a(t - x)) dt da ] (8) Suppose f(x) is defined in the interval (0, oo). Then
for x > 0, f (x) can be represented by Fourier cosine
since the first tern.1 in the 1ight hand side of (7) be- integral (13) by redefining f(x) in (-oo, 0) such that
comes zero because f (x) is absolutely integrable. f(x) is even function in (-oo, oo).
Thus as L ➔ oo the Fourier series becomes aFoUiier Similarly, by redefining f (x) in (-oo, 0) such that
Integral. f (x) is an odd function in (-oo, oo ), the giv~n fu~c-
E;.panding cos(a(t - x)), (8) is rew1itten as tion f (x) can be represented by the Founer srne
integral (15) valid for x > 0.
f(x ) = 2- [
rr lo
00
(j 00

- oo
J(t)cosat dt) cosaxda
Fourier Integral in Complex Form

+~ L:o (J_: J(t) sin at dt) sinaxda (9) Since cos(a (t - x)) is an even function of a ' then (8)
can be written as gr
The Fourier integral expansion of f (x) is the right
side (expression) of (8) or (9).
f(x) =~.~Joo
2 rr -00
r1L 00

-oo
.f(t)cos ( a(t - x)) dt1 da
( t6)
Introducing C

A (a ) = !_ 1 00
J(t) cos(at)dt
(10) Since sin(a(t - x)) is an odd function of a. then
rr -oo
= ~ . ~ Joo [f" xl) dt] da.
= 2_ 1
f(t) sill ( a(t -
00
(11) 0
B (a) J(t) sin(at)dt 2 rr -oo - 00 l l 7)
rr -oo
(9) can be rewritten as
Multiplying (17) by -drr and adding it to l 16). we
get
f(x) = la?? [A(a) cos(ax) + B(a) sin(a.x)] da (1)
.
f(x)= -
l Joo [f cv ·/'(t). co~ ( LYV - x))
. 2Jr - N - N
ote: Fourier integral is very useful_ in solving dif-
ential equations and int~gr~I equations ..
- i sin ( cx(t .d) tit] da
Particular cases of Founer rntegral (10). 0v [ I 1·~1./(. t). e iat dr] eiaxda (18)

r cosine Integral
·

.
f'(x) =
f N

I
2Jt N

1•N[~ 1•Nf(t)e-i"'dt]i"-'dOC·
f . h
(x) is an even functlon, t en
B( ) - 0 and
a - or .f (x) = ./frr _ J2ir 00
- oo
- . h are very often
A(a) = -
fc 2
00
J(t)cos(at)dt 0 2) Note.
. A list of standard results wh1c
~··
.
rr o used are given below:

2024103.2n ~3:~s
FOUIIIB' ~ '°'1RIER ~ & INTEGRAL TRANSFORMS - 2CL""JI
dard Results
,tan . e'1' . Fouri er transfonn method is the process of obtain-
• ,;_r. sID bxdx= a -,. . .b,- (a sin bx b
l' ./ e . b
- cos bx ) ing F (a) fo r a given function f (x).
.-,; e-•L\
/
. sin bxdx = ~
; , .,0
a-,-b-
at Fourier Cosine Transform of f(x)
1
. a.r.
}e
= a {_,_b1
cos bxdx ·
(a cos b t + b · b
- SUl x) Th e Fou1ier integral of an even function f (x ) reduces
·· . .__ - or cos bxdx = ~ to Fourier cosin e integral given by (1 3)
~,- .o / e a--1,-

=~1
00
,
:..L.: --=
2
e-.r dx = fi f(x) [1'XI f(t ) cos(rx t )dt] cos(a x )da
d r = .!2 tan
0
·--= I" -c- ~
6• •I() C-:r - f - -
'
f • 2

-x ~ ~e-• d t = .!. sec ~


Put Fc_(a ) ~-- fooo J ~x ) ~os a x ~x _, (21 )
~. f, e-•-e_-, . : 2
2 00.
b d t - I b Then f( x ) = - {
s. J~-.;•; ~
--al
sm x x = an ; . c >

o. b > 0 rr lo
Fc(ct) cos(a x )dct (22)

9.. -.:
1
~dx
t
= f if a > 0. - Fc(a) given by (21) is called the Fourier cosine trans-
form of f(x) in the interval O < x < oo and f (x)
20.2 FOURIER TRANSFORM given by (22) as the inverse Fourier cosine transform
of Fc(a).
fiom the Fourier integral representations (18),
Fourier Sine Transform of f (x)
13) and (15). we get Fourier transform, (which is
complex). Fourier cosine transform ( which is real) The Fourier integral of an odd function f (x ) reduces
md Fourier sine transform (which is real) of f(x) to the Fourier sine integral given by (15)
ES follows:
~ fo
00 00

f(x) = [fo f(t) sin(at)dt] sin(ctx )da


Fourier Transform of f (x)
= fo
00
Take Fs(a) f(x) sin(ax)dx (23)
The Fourier integral of f (x) in the complex form
~ren by (1 8) 00
Then f(x) = -21
f(x) = 1'.X) [-I
- x 2rr - co
l oo
f(t)e-iardt] eiaxda (18)
rr o
Fs(a) sin(ax)da

The function Fs(a) defined by (23) is known as the


(2-l)

can be written as Fourier sine transform of f(x) in O < x < oo. The

✓ f(x) = 1_: F(a)eiaxda (19)


function f(x) given by (24) is known as the inverse
Fourier sine transform of Fs(a).

'J,here F (a ) = _I_
2rr
loo
_ 00
f(x)e-iaxdx (20)
Linearity Property
Fourier transform, Fourier cosine tran~fonn .and
Fourier sine transform are all linear operauons_ (smce
Fru.Jdefined by (20) is known as the Fourier trans-
form of f(x) . the integral operation is linear). For example for any
two functions f(x) and g(x) and for any_tw~) con-
Fof (x)
· defined by (19) is known as the inverse stants a an d b , the Fourier transform of a·f tx) +
unertransform of F(a). F(a) and f (x) are known bg(x) is given by
as Fourier transform pair which differ in form only
iIJ the s"

!Vote• Th
ign of the exponent.
. al
F(af(x) + bg(x)) = 2~ 1-: [ df(x) + hg(x)] f - ,a\dx
(19 :
th/
e factor -1... can multiply the f(x) 111tegr
instead of the 'p(a) integral (20). Alternative~y
==a. _I loo f(x~ - ia.tdx
2,r - oo
(l g factor ~ can multiply each of the integrals m
J anct (20). +-b loo ( \.-ia.tdx
2n -oo
gx~
Pou· ·
cont· tier transform breaks up the function rnto a
inuou =aF ( f(x>) + bF ( g(x>)
s spectrum of frequencies a•
20.4 - HIGHER INGl- - ~~--..--.
In a similar way
Fe(af(x) + bg(x))
~ ffliiS-V
-
b. Similarly,
--=
-------~-
~

= fo
00

(af + bg)cos a x da
Fd f' 1:::: r )O f ,sin. (ax)c/a -
Jn f .
o - s1n ~x \()(,
00 0
= a fooo f cos ax da + b fooo g cos a x da -a
1
O
f cos ax dx

:::: 0 - a Fc{ J) .
= aFeCJ) + bFc(g).
Corollary I:
Fourier Transform of Derivatives
Fe{/" ) == a l~r{f' ) _ f' (O)
Fourier transform of a derivative of a function f( x )
correspo~1ds_ to multiplication of the Fourier trans- = a ( - a Fe If)) - ! ' (0)
fo rm by 1a 1.e., Fe{J'' ) = - a 2 Fc{f) - J'(O)
(26) 11ew

F{ J'(x) } =F{:~} = (i a)F{ J(x)}


Corollary 2:

Fs U " } = - aFc{f' ) = -a [ aF5 {f) - f (0)1


Proof: By definition

F { J'(x) } = -2rr 1
f 00
_ 00
J'(x)e-iaxdx
FsU " } = -a 2 FsU) + af(O)
Application to Initial Boundary Value
mi

Integrating by parts Problem (IBVP)


No
=-1 [ f · e - iax loo -(-ia)f oo f-e-ia xdx] The solution of a IBVP cons1stmg of a partial ass
2rr - oo _00 differential equation together with boundary and
A ssuming that f( x )-+ 0 as lxl -+ oo, we have initial conditions can be solved by the Fourier
transform method. If the boundary conditions are
F { J '(x) } = (ia)F { f(x)} of the Dirichlet type where the function value is \
prescribed on the boundary, then the Fourier sine
In general F { ~:(,} = (ial F {f}. transform is used. If the boundary conditions are of
In particular the Neumann type where the derivative of function
is prescribed on boundary, then Fourier cosine
E~
sh
F { ~: ;} = -h{f). (25)
transform is applied. In either case, the P.D.E.
reduces to an O.D.E. in Fourier transform which
is solved. Then the inverse Fourier sine (or cosine)
Fourier Cosine and Sine Transforms
transform will give the solution to the problem.
of Derivatives
Prove that 20.3 CONVOLUTION s
a. Fe { J'(x) } = aFs { f(x)} - f(O) Convolution of two functions f (x) and g(x) denoted
Fs { f '(x )} = -aFc { f(x)} by f * g is defined as

f:
By definition and applying integration by parts
h(x) = (f * g)(x) = 1-: f(s)g(x - s)ds

= 1-: f(x - s)g(s)ds


Fcf f' }= Jo{ oo J' cos(ax)dx = f · cos ax looo
00 Theorem•• The Fourier transform of th e conv_oe,·
+w fo f · sin axdx h · Fouri
lution off and g is the product oft eir

Fe {/' } = - f( O) + wFsU}
transforms.

~~
~ B dllQlfi· · .. ER ~Nbftaa.. !
f· y e muon - -~&INTEGRALTRAN
pro<' ' SFORMS - - j]
1
00
1
f(
f*g) :::= 2:rr f_oo (J * g)e-iaxd.x Consider I =
00

f (t)e - lrxt dt
-oo
L: [1-oo
00

:::= 2~ f(s)g(.x - s)d s ] e - i a.\. dx

bange the order of integra..;


!Jl(erc . uon e-l ( l + irx) loo _ 1 - ia
=
~b C,[i: f(s)g(x - s),-iaxdx ds 1 So f( x )
- (l + ia)

= -2n1 J oo leirxxd o:
1=0
- l + ia = 1 + ,,,2
...

NoW put x - _s = q' ~o x = s + q .th


-oo
wvariable of mtegratlon instead of w1 q as the 1 J oo (1 - io:) .
~ x. = 2n -oo 1 + o:2 e'cxx do:

~ J oo Joo oo 1
::
2Jr -00 -00
f (s )g(q )e-ia(s+q) dq ds
= J-oo 2n(l + o:2) (1 - io:) X

:: _I Joo ( )
f s e - ictsd s ] [2n
- Joo g(q)e-iaq dq ]
oo
x (cos o:x + i sin o:x )do:

[ 2rr -oo

::2rr •F{/} · F{g} .


2n _ 00
J 1 [ .
= _002n(l+o: 2) (coso:x+o:sinax)+
+i (sin o:x - o: cos ax)] do:

Note: Convolution is commutative f * g = g * f, The second integral on the right side is zero because
the integrand is an odd functio~.
associative f * (g * h) = (f * g) * h.
00
2
f(x)= [ (cosax+asinax)da
2rr(l + a 2) Jo
WORKED OUT EXAMPLES For x > 0, f(x) = e-x so
e
-x
=;
11
o
00
_ax+
cos a sin ax d a
_--,,---
1 +a2
(1) -
Fourier integral
Example 1: Using Fourier integral representation
For x < 0, f(x) = 0 so
show that 1 loo cos ax+ a sin cu da (2)

!:
0 == - 1 +a2
Jr 0
if X < 0
00
At x == 0, f(x) bas a discontinuity. So
- - -+
cosxa ifx=O
-a-sinxa
-da= - 1 l
1 + a2 ~ ~ ( /(x +O) + /(x -0)} = 2[1 +OJ= 2
[
o 2 f(x)
;re-x if X > 0.

For x == 0 • -
Solution: Consider the function defined by 1 1 100 cos ax+ a 2smax da. (3)

- == - 1 +a
0 ifx<O 2 ,r 0

1 3) the result follows.


f(x) = - if X =0 Frotn (1), <2>, ( . · gra1and
2 . Fourier cosme mte
e-x if X > 0. ple2· F111dthe(a). resentation) of
Now ftnd · f f(X) ESIIID(b)Fouri~~mem~gral(rep
~ th the Fourier integral representation
e exnr. -•
° (x) == slllX
.
-
ifO<X~,r
t'vnential form: By defirution f ifx>"
r

~
f (x),,,
I
I; ~L: ~
f (t) _e -iat dt] ,;ax d<t
::0
·
20.6 - HIGHIRRliiNiicCIINIIIIIN8~iiiiiiriMini i l4•111!Jl(~f l~-~ ,,
~-~==~~===~~
So lution: a. The Fourier cosine integral of For a = I , B( t) :::: I. ~
J(x) = fo
00
A(a) cos a x da where Thus .f (,,·) -- [oo
.o -
2 sin
-- ·s arr .
rr(l - a2) ,1nco:da

00
2 [ sin 0'.7r
A(a) = -2
J[
lc °"
0
f( t ) cos al dt = ; lo (\ - a2) sinaxda.
J(,t

A(a) =; rUorr sin


1
1· cosatdt 1
+ :rc O1j
00
Fourier transform

Example 3: Represent J(x) as an exponential

= ~~ fo:r
Fourier transfonn when
Lsin(l +a)t +sin(l-a)t]dr
I
sin x, 0 < x < rr
J(x) = 0, otherwise
= - ~ f cos(1 + a)I + cos(l - a)t]rr
rr l l +a l - a r=O show that the result can be written as

l Ll - cos(l + a)rr 1 - cos(l - a)rr] 1 laoo cos ax+ cosa(x - rr)d


= +- - - - - - - + - - - - - - f(x) = -rr o 1-a
2 a
rr l+a 1-a
Solution: Fourier transform in the exponential form
A(a ) =~I l + cos arr
rr
2(1
L l+a
+ cos arr)
rr(l - a.2)
+ 1 + cosarr1
1-a is given by
F { f(x)} = F(a) = -2rr1 _ 00
.
Joo f(x) · e-,axdx 1
For a = 1, A(l) = 0.
Substituting A(a), we get the Fourier cosine integral
= _!_
2rr
[!-oo + lo
0
0 rrr sin X • e-iax dx +1rr ol'1
00

of f(x) as
00 2(1 + cos arr)
= -1 1:rc (eix e-ix)
- -_ - .
- e-,axdx
f(x)=
1 00o rr(l - a 2 )
-cosaxda
. _
2rr o

(eix _ e-ix)
2i

= -21 1 + cos arr cos ax da smce sm x = -----


2i
rr o 1-a2
= -. .
I l:rc [ez(l-a)x _ e-i(a+l)x1 dx
b. Fourier sine integral of J(x): 4rrz 0
00
f(x)= fo B(a)sinaxda, where = ~ [~i(l-a)x _ e-i(l+a)x ]rr
4rrz i(l - a) -i(l + a)
2100 f l
0
B(a) = -
1l' 0
(t) · sin atdt
- 1 - ei(l-a)rr e-i(l+a)rr 1 1 1
- 4rri i(l - a)+ i(l + a) - i(l - a) - i(l +al
= ~ [Ion- sin t · sin atdt + loo o]
1 e-irra +1 = -1
= 2rr since e±irr
=; 2 lorn- [ cos(!
2 1
- a)t - cos(l + a)t] dt l- a2

= ~ [sin(1
TC
- a)I _ sin(l + a)t ]rr
I-a l+a t=O
Then f(x) = L: F(a)eiaxdx

1
=-
' 7! 1- a 1
- - _ 2 s.marr
- - +sinarc]
[sinarr
+ a - rc(l - a2)
= Joo
-oo
_1 (e-i:rca
2rr
+
1 - a2
1) eiaxdx

1,
equExpanding
15t11e0'.011- .
lf
. edFOuRIER
exponential
.
~
Fout·i·e . l'OliRIER
the 1ntegra ct1 transform repre-
& NINIIAL 1RANSFORMS - 20,7

se
nto _ j,r<X)e'cu
. . 11 The inverse Fourier cosine tran c; form
(1-t- e . . 2
I(x) = f rx, f,~(a)co s u.xd.x
:;: :; [I + cos arr - 1 sina rr ][coso, 11 + ; si n O'x l :n- lo
-cos ax+
00 coso-(rr - x) + il +:, in ax. sinO' :n- l
- -2 1a cosax +cosa(.n- - x)
- (, ,l - -a)2
-,;,- - - - d x
_~ 1,x, ~
- rr +
co~ a.xdu
o a2 u2
27f O
1+
(( X):;:::;

• 1
00
sin arr · sin ax Since J(x)
ten as
= e - ax' the above integral can be rewnt-
+-2JT - oo 1 - a2 dx .
00

1 C.OS a.x if e
-ax
since u1e Znd integral is zero (odd function), result - -da = - -
0 a2 + a2 2a
follo,1·s. b. The Fourier sine tramform of f( x) is
ie 4: Find the Fourier transform of
£~11)11 P
Fs \ f (.x) \ = Fs(a) = {:,o f(x) sin axdx
1
f(x) = 2a' if Ix! :Sa
00 CJ_

= 0, if lxl > a
= fo e- ax sin axdx = a2 + u2
Now the inverse Fourier sine transform
Solution: By defi m. r·I on ' the Fourier transform of
f(x) is . f(x) = -2100 Fs (a) sin axdx
TC 0
. 1 Joo J(x) · e -1cu dx 00
f ( f(x) } = 2rr - oo
= -21 TC 0
_a_ _ sinaxd x
a2 +a.2

= 2~ [L: + L: + f] 00
with f(x)
.
= e-ax ' this above mtegr al takes the form
-ax

= TCe 2
= _l
2n
[! -a
- 00
0 +
fa 1 -iaxdx
-a 2a e
+
11
a
0 l oo a sin ax da

o a
2 +a.2
'
roo a smi°x da = Joroo ~da=,
a -.
a Note: For a= O, sform
e -iax
Jo a .

Find the inverse Fo urier sme tran


= 4na -ia x =-a Example6:
1
of -e
s
-as
·
-1-
- [ e-iaa - e +iaa]
Solution: -as)
= 4naai
- ~
1 [eiaa - e
2i
. ]
-1aa
=
sin(aa) .
2rraa
)
f(x)=Fs
-1

d.
l ~
s

- 2naa ·ne and (b 2100


f- -
~ sinsx s
s
-(IS

. cos1 - JC 0
) Founer d a > 0.
Example 5: Find the (a -ax for x ": O an iegrals"
sine transform of f (x) = e "Laplace 10 · g W.r.t., X
ntiatin
known
Deduce the integra1s ;n ax dNas Differe 2 oo e-a.f ~(sinsJ)ds
r00 ~
rx.
lo ~ de, and Jo a1+a2
I.A,,

sine trans
- df - -
- - n Jof -;- ,l;i;
a +a rier co dx ·os sxtls
. • n Fou
Solution: a. By dejimtw 2100 e- us . S. C

lorn, off (x) is dx - -


- 7f O -;- ti
2 .----;;
oo - as cossx ds == -7f x2 +w
_!.
F,\ f(x )j = F,(a) -- Jo{oo J(x)cosax a
----1
df
-
dx - 1C l 0
e

Fc{f) = Jo{oo e-ax cos axdx === a2 + a

I ~
I ~• 1111 . R bounded·1.e i RIii ----~•INTIGRALTRAN
~ •
.,~51(irst term ts .I-LS. becomcs.,z , .~ 0 as .r ➔ oo . SFORMS - 20.91
~if e10 . So/111,ow 1
U(f. s) = Fe { 11 (x, r ) } :::::: { C\:) th
. o solve e Laplace's equ ation
~11 d lo 11 (.r, 1) . cos sxdx iJ 2u iJ 211
(I)
fhCll
- U
di
=- s:2c2U
.
v2
X
+ iJ y 2 = 0
with the boundar·y con d"1t1.ons
•11e 1J1al t11e res_ult of Fouiier trai . . = 0, for O < y ,, b
11 (0, y)
~l 1i111inate denvatives of r f. 1sfor1111ng on x is (2)
ll1 e . • I om th h u(x, b) = o for O < x <
ill t11us Jeavmg an ordinary diff e_ eat equa- 00 (3)
ot . '(. erenLinl equation u(x, O) = e - ax . h
Wil a > 0. (4)
,1 r.t., . b
·1111egra1JJ1g y separation of variables Region is: 0 < x < oo O b
~ ? A I· ' < Y< ·
U(s, r) == A e-c·s·r PP ymg Fourier sine transform of (I) on both sides

r== 0.U(s,0)=A
00
1o
00
a2 u
-ax2 sin sx d x + 1oo 0
2
iJiJyu sin sx dx
2 =O (5)

]lJUS
A== U(s. 0) = ]f u(x , 0) · cos (sx)dx The
tion first
b integral is_· sunp
· l"fi
1 ed by successive integra-
0
y parts and gives
from the initial condition u(x , O) = f (x ), roo a2u . au loo
00 J~0 a1 sm sxdx = - . sin sx
(hen A== fo f(x) cos (sx)dx
X ax 0
00

1
au
- - · s · cos sxdx
O ax
Taking inverse Fomier cosine transform
Assuming that :~ ➔ O as x ➔ oo
u(x, t) = F; l {U} = F e-I { Ae-c2s21}

21
00

00 =-s[ u-cossx[- fo u-s·(-sinsx)dxj


=- Ae-c2 s2 t • cos (sx)ds
rr o using the boundary condition (2), u(O, y) = 0 and
Special Case: when f(x) = e-ax then assuming u is bounded i.e., u ➔ 0 as x ➔ oo
00

A=
oo
e-ax cos (sx)dx =
a
b2
= -s2 fo u sin sxdx
lo o a
2
+
Thus the solution in this case is
Put U(s, y) = Fs { u(x, y)} = J;° u(x, y)sin sxdx
Thus
u(x, r) = F-1 - -a- e - c2 s2 t } oo a2 u sin sxdx = -s 2 (6)

u(x, t)
c

=_21
{ a2 + b2
00
a . e-c 2s2 ' . cos (sx)ds
L 0
-
ax2
U.

Substituting (6) in (5) and rewriting


TC o a2 + b2
00

Exa~ple 10: Solve the Laplace's equation in the _ 2u+~[f u(x,y)sinsxdJ = o


s dy2 Jo J
senu infinite strip shown in Fig. 20 .1.
y d2U
-s2U + dy2 == O
. ( l) has been reduced to an
Thus Laplace Equation . .in )' (withs as a pa-
. differential equauon .
ordinar}' . f (7) is
rameter). Sotuuon o . (8)
sh sy + B smh sy
f(x)=e-ax U(s, y) == A cO

Fig. 20.1
FOURIER INTIG~ ...... _---=--...,__
--- 2 "VURIER ~
f _ { l - -'" • if Ix I ::: I
1·r I
~ I&
~ lli
ltm~'n
o;R-;;
A-,L ;::
TR~A;:-N- S':"'F-0-~
R-M_S_ 2
o. 11
9. • (·r ) - O, .r l ::::,. 1 · "L,
Qcnce evalu ate

I
- - - - - - - ~~ -
(ccv , co~ :.-,in_, CO S ½c!.r
.o 18. p·ind P.C.T.of/(r t. 01-
- .;-2
y - ,
)== 2 r,
f { f (.r)} = - 4(s cos s - sin s); 3
A"··~ ·
Integral = 76
- JT S ,
Ans. ~OH-co~2, - JJ
0. X > 2.
,i

,o. f( X) = l
~ (as cos sa - sin sa )
:r .
0.
for lx l .:S a
for lx l > a I 9. Find F.S.T. and F.C.T. of 2e s, + Se -2•
Ans. F.S.T.: A +
s +25
Ss Fe T ·
:;I:i:i ' · · ..
JO
,1-l.4
10
+ ,1.i.25
.411s. s- 20. Find F.C.T. of / (x) == e- a, cos ax.
-,-,
I J. ! (·'·) = e-r , -OO < X < 00
?
r
&e-z Inverse Fourier transform

AflS, \ { Q, - 00 < X < a 21. Find f(x) if itsF.C.T. is


1
L andF.S.T. is
2
~,~.
1
f (x) = x. a =sx :S b Ans. f (x) == e- x, f (x) == e- •
1?
-· 0, X > b 22. Find f (x) if its F.C.T. is t; (a - ½)if s < 2a
f(x ) ::::: and zero if s ::'.:. 2a .
. - b eisb) + l..(ei sb _ eisa)
Ans. sI ( ae'sa 2
s Ans. (2 sin2 ax)/rr 2 x2
ds
IJ, . !(X ) -- xe-x ' O -< x < oo 23. Find the inverse F.S.T. of sn e-a5
Ans. 2·n! sin[(n+!!±l
l)x]
f (x) == 1 ( 1+1.s )?- rr •(a2+x2) 2-
Ans. 2.,7 • (1+s 2)2
24. Find the inverse Fourier transfonn of e- •
sin sxds COS X , if Q < X < 1 y
Ans. [rr (y2+x2)]'
b2+s2 ) . 14. J(x) = { O, otherwise
Integral equations
, show Ans. I [ sin(s+ l) + sin(s -1 ) +
4rr s+l s -1 Solve the integral equations:
t
. { 1- cos(s+ l )
i s+ I
+ l -co~{-l)
s
}] ·
O~a ~l
< 1
25 _ J;•J(x) -sinaxdx=\~.- a. a-,. l
. transforms
Fourier cosine and srne 2(x- sin .\ )
Ans. f(x) =~ b
15. Find F.C.T. and F.S.T. of oo f (u)du .. = ~, 0 < ll <
k, if O < x < a 26 . J-oo (x-u )z+a•(1,- u)a ,\-+
f(x) = o if x > a.
! , k(I-cosas
)
Ans. J(x) = lnr(x2+(1>- al11 0 ..... l~ ..... l
- 5- , Fs {f(x)} ==
Ans. Fe{f (x)} = k · sinas ·s roo f(x)cosaxdx == \~)~ tY. lY ...... l.
how that 27. Jo
. .T. 0 f e-lx l. Hences
16. FlildFS
•N ~ tlr
. ,re-m O. Hence evaluate .lll ,-

1 00

o
xsm mx d - - , m >
x-
1 + x2
2
An s, !( . \
·)::::
~, - cos .0
2 ~!\ II

Ans. F.S.T. == 7 re,.) J'(x) cos a."· h· == t'


l+s
28, JO
---1-,r, .
l?. Find F.S.T. of e~ax • Ans. iro+r>
Ans. tan-I s
;;+c
20.12 - HIGHER ENGi .. Zlllliii.......
u e problems 32 · Fiuct th e temperature u(x, t) in ~
Solution to boundary val_
(B.V.P) : Laplace's equation
bar with il,'
a.~
=
. /k when x = 0 andsel'n.i-infi1\°
w· . ite
te mperature 0. Assume that i!!!. 1th 1nitia.1
. d the steady-state temperature disu·ibution · ilx ~ {.) .
29 . F 111 . I 00 an d 1s bounded . as x . . _,
. ") in an infinit e metal plate covenng t 1c
[/ (.\ ' I
first quadrant with the edge along the y-ax is

Ans. u(x, t) = lrr Jor
00
l'--(1 _ ~-
s2' e
2
v c21)
cos sxds
lJla,5~
(,d iJ1
held at 0° and 1J1e edge along the x-axis he ld of f
33. Solve a,, -
il t -
.2 a2 11
c ~ ,x >
o, l > 0 With ~ \
100, 0 < X < 1
= x, 0 < x < l •J~ ,
X > 1

l
at 11 (.r, 0) 0, 'J
[ = Oandu(x,0) = - -
0, X > 1 .
Hint: Use F.S.T. Lo solve Laplace's equation.

Ans. u(x, y) = 2: J; l - ~_o ss e - sy sin sxds


Ans. u(x, t) = l7r Jt [sin s + -h, (cos s - 1)1 c I cl,
S S
2
I

x cos sxds
or u (x, y) = 2: [ arc tan ( f) - ½arc tan ( x; 1
)
34 . SO1ve au =
a, C
2 au X
a;r,
2
> O, t > 0 With
and u(O, t) = u0 f0r gesult
-½arc tan (x;t)] u(x , 0)
t > 0.
= 0, x > 0

30. Solve 1-1 .u + u yy = 0


in the upper half-
plane when the temperature along the x-axis Ans. u(x, t) = u0 [ 1 - 2 roo e-s2c2/
-; Jo - 5
- •
.
sm sxds.
1 fs

u(x, 0) = f(x). Deduce (a) when u(x, 0) = 1


Vibrating string
2uo, x < -1
if \x\ < 1 (b) u(x, 0) = uo, -1 < x < 1 35. An infinite string is initially at rest and has By de
{ 0, X > 1 an initial transverse displacement y(x, 0) :::
f (x), -oo < x < oo. Show that the dis-
A,~s
' .
u(x
. '
)') = lrr j'-oo
00
yf(s)
(x -s)2+y2
ds known as placement y(x, t) of the string is y(x, t) ==
Poisson integral formula for the half-plane ½[f(x +ct)+ f (x - ct)].
y > 0 or Schwarz integral formula.
af
a. u(x, y) = ¼[tan- 1 ( 1
/) + tan- 1 ( x; 1 )] 20.4 FINITE FOURIER SINE AND COSINE = ax
TRANSFORMS
b. u(x, y) = ~ [rr - tan- (
1 1
:x)
Let f (x) be a function defined in a finite interval
+ tan- 1 (1;x) J. 0 < x < L i.e., when the range of one of the variables
say x is finite. Suppose f (x) is neither periodic nor
even nor odd. Now by redefining f(x) as an odd
One-dimensional heat equation
function in - L < x < L, the half range Fourier sine
31. Find the temperature distribution u (x, t) in a series expansion of f (x) can be obtained as
semi-infinite metal bar with the end x = 0 at 00

zero temperature and initial temperature dis- f(x) = Lb 11


nrr x)
sin ( L
tribution f(x). Deduce when f (x) = e-x. 11=1
s· .
Ans. u(x , t) = ¾ft U(s,
2
0)e -c s 21 sin sxds where b, 1 = ~fol J(x)sinC~x)dx. lea
where U(s , 0) = ft u(x, 0) • sin sxdx = Then the finite Fourier sine transfonn of J(x) in
J; f (x) sin sxdx 0 < x < L is defined as

when f(x) = e-x : U(s, 0) = ~-


l+s~'
Fs(n) = fol f(x) sin ( n~x) dx
2 roo -c2 ,i11
u(x , t) = "ir Jo se l+s 2 sin sxds which is a function of n, an integer.
fhe inverse finite Fourier .
. ,en by s111e tr
jS g11 2 Q1tsfon11 of F.r (11)
By definition
J(x) = -L t"' F

rnas1nu· ·1arw
•) in O < x
L.., s(n). sin ("Jrx
ay, thefiniteF0 .
L • uner .
r:-) F.1~1
c 2 : : : 1l -
iJ
x
82/
o 8x2 . cos
(/IJTX
l ) dx ,
of r J
( . < IS defined COSine trnnsf
. Fc(n) fl = as onn integrating by parts

fhe inverse finite Fo .


]
0
f (x) cos ( ~ )
L dx
-- -8f . cos -mrx IL
ax L
0
+ -mr
L
1 8/
0
L
. -mr x dx
-aX ·Sin L
. b uner co . ::::: [ - fx(O,t)+( - l tfx(L , t )]
fr(11) is given y sme transform of
l 2 oo +~[f
L . -mr xlL _
·Stn
f(x) = zFc(O) + -LL..,
t"' F ( ) nnx
c n cos-
L o
n==I L .

uo
With
for Result 1:
-n; lo L f cos n~xdx]
f) = l f)
2

I a2 2 2
-n rr Fe -a - [( l)" 2 2
Fs ax2 £2 Fs(/) - iJx2 - - fx(L, t) - f x(O, t)] - n ;~ Fe(/).
Xds].
-nn Note 3:
--Z:- [(-It f(L, T)- f(O, t)]
1. Apply finite F_ourier sine transform when the
and has Bydefinition . bou ndary conditon f prescribed.
(x,0)==
2. Apply finite Fourier cosine transform when the
e dis-
boundary condition f x prescribed.
(x, t) ==
integrating by parts

- aJ .
--·M-
mrx IL -mr
-
loL -aJc o omrx- b WORKED OUT EXAMPLES

OSINE ax L o L o ax L

interval
=0- n; {[t(x,t)-cosC~x)]: Example 1: Find the finite Fourier sine and cosine
transform of f(x) = x(rr - x) in O < x <re.

ariables mr [L
+L (nrrx) dx }
Jo J(x, t) · sin L
Solution: Finite Fourier sine transform of f(x) is
die nor
(L
I f}
nrcx
an odd Fs(n)= Jo f(x)sinLdx
·er sine
Fs a2 = - nJr [(-It f(L, t) - f(O, t)]
ax 2 L = Lrr x(rc - x) . sin nxdx

-cosnx \II'
=x(n-x)· [ --;;-- 0 - l(

8itnilar1
Y, - sin,u)\71' +(-l) · cos;.'C l'
-(rr - 2x) · ( --;;r O n 1o
Result 2:

I JI
f (x) ill
+ !..[1- cosnrrl
2
F,(n) == 0 + 0
Fc a = - [fxCO, t) - (-It fx(L, t)] -
a;I ,a3

11
== ~{1-(-1) 1
n2;r2
- ---,;r- Fe(/). "
'IO!~ ~ u ~
-=11-~·-··-_-_-,_~~:,&-t-----..._,_~
fi1io.:;.1ico~-;.;-~Hiii1oiiHil1fi11■iNN01Niiiiiiii1ii1iR1iNOio7~iinii....l•IIUAiiJ■&J.__:UV=i
whrn ,. = /,, U(s. /,) I ~, / 11( 1 . /,)) - F, (Ol O {
]. f(x) -
~-----
~
I, \x \ .:::: I H
~
· (J). 11( 1. h)
si nL'l' frn111 (). l •·rnlll (8). at y h. 9· j
0' Ix I > I . cncc cva\ U,\tc
, 1 l'l'~ h s h I H si nil sh (9) j
:!E..
. ·oo sin>.. co~>.. r 2' if \x\ " 1
Whc11 y = 0. U (s. ()) = . ' '-'
/ll 11( 1. 0) Sill .,· 1 d 1 /
,o ~ d>- = n
ii ' if \x\ ::::: 1
Fro rn (-1-). 11 ( ., • 0) - <' " ' so 0,
\ if \x\ > 1
l ' l•' · 0) = /''-' <' ·" si n .\'.\d1 = .,·
( I 0)
Ans. /(x)
'
= 1n J()r 00
sin> co, >,x
),
d'A.
.n a -1 -1s 2 10•
4. Find the Fourier cosine and sine .
Pulling y = 0 in (8) ·
f (x) = e- kx , for x > 0, k > o.
tntegra\
~ rif
s Ans. Fourier cosine integral (F.C.l.)
, ,=
a-+ s-
U(s. 0) = A+ B . 0

or A=
a2 + s2 (LI)
_ - kx _ 2k
j '( x ) -e - -
n o
1 00

-
cos sx
k2 + s2
ds

Substituting ( l I) in (9) Fourier sine integral (F.S.l.) = f(x) == C '


= lJT Jor
00
ssin sx ds
k2+s2 ·
12
s
B= 2 coth sb (12)
a +s 2 5. Find the Fourier cosine integral of f(x ):::.
e-x cosx.
Thu s the solution (8) becomes
Ans. f(x) = e-x COSX = ¾ft (s2+2;4c~~ sxds
U(s. r)
-
=- -s - cosh sy
a 2 + s2
6. Find the Fourier sine integral of f(x )-==
s e-ax _ · e-bx.
An
coth sb sinh sy (13)
a 2 +s 2 2 2
A f(x) = e-ax _ e - bx = l r00 (b -a )s sin sxds
ns. 7T Jo (a2 +s2)(b2-'-s2 l .
Taking inverse Fourier sine transform of (12) a> 0, b > 0
u(x. y) = F -l { U(s , y)} 7. Using Fourier integral representation, show

l!'
5
that
00
=~ f
n Jo a2
s
+ s2
(cosh sy-cothsb · sinh sy)sin sxds.

In
0
00
sins . cos x s
----ds -
s
-
4,
~if Ox - = x <
I
1

0, If X > l.
EXERCISE
Hint: Find Fourier integral of
Fourier integrals 1, if Q :'.:: X < 1
f(x) = { 0, if X ~ 1.
Find the Fourier integral representation of f(x):
Fourier transforms
lxl X, < I
1. /(x) ={ Find the Fourier transform of f(.r):
0, lxl > 1
A
ns. f(x) = J -oooo sina_ll[a
- aco~aeiaxda 1, for lxl < a
2
=
cosx, - JT < X < ~
8. f(x)
1 O,
for lxl
.
> a
2. f(x) = { 0,
2 2 Hence evaluate f0
00
811
'.t·dx
!xi >~
Ans. F {f (x)} = Zsi:'. rn, for s -=/= 0, For s == O,
Ans. f (x) = .!. J oo co~(aJT /2) eiax da
7r -oo l- a 2 F(s) = 2, Integral = f
Integrating f(x) = ~1 ;r
a
x2
dx
+ a2 = ;;:
2 [ ['
lo I• sin t x dt + J,
2
2sin IXdt + r ·l
From (1) at x
= -;r2 tan -
= 0 , J(O) = 0. Using this in (2)
l X
-
a
+A (2)
= ~ [ - C:S IX \ I
t= O
_ 2 COS I X \
X
2
1
l== I 111eP
2
0 = f(O) =0 + A .·.A= 0 J(x) = -:rrx [1 + cos x - 2cos2x).
...r0 te t
oat 1
i~ eliJlllp2
Thus f(x) = -rr2 tan
- IX
-
a Example 9: Find the temperature distributio . tO tbtlS
semi-infinite bar with its end point and lateral su~ in tiOJl• 't'.
is the required inverse Fourier sine transform. wr.t.,
insulated and with initial temperature distributio:: · 1J).tegf3
Note: When a= 0, J(x) = F s- l { ¼} the bar is prescribed by f (x). Deduce the solution U(s
when J(x) = e-ax.
= l;r tan - 1 oo = 1
At
Example 7: Find f (x) whose Fourier cosine trans- Solution: This problem is represented by the one-
form is sina s . dimensional heat equation 'fbUS
s
au 2 a2 u froIIl
Solution : It is given Fclf(x)} = sin/°. -=c-
2
at ax

J(x) = F- 1
sinsa1
--
2 Joo sinsa
= - --cossxds for 0 < x < oo, t > 0 with boundary condition men
C S Jr Q S
l
=0 (insulated) Taking
= ~ ~ [ 00 sin(s(a + x)) + sin(s(a - x)) ds Ux(O, t)
:rr 2 } 0 s
and with initial condition
1 J oo sin s(a + x) 1 looo sin s(a - x)ds
=- ----ds+- u(x, 0) = f(x) (given) for O < X < 00
rr O s :rr o s

=~(I+I) ifa-x>Oi.e.,x<a Since the boundary condition is of the Neumann


(derivative) type, apply Fourier cosine transforma- Speci
- ~ (~ - ~) if x - a > 0 i.e., x > a tion to the equation
-:rr 2 2
since J;° sin/x dx = I when a > 0. (See note of WE .
oo au cos sxdx = c2 [ 00 -a2 u cos sxdx
-
5 on page 20.7) la
O at O ax
2
Thus
Thus

f(x) = g if X < a
if X > a.
Integrating the R.H.S. by parts

a laO
-;-
ut
00
u cos sxdx =c 2
ax [
o
~
a coi sx 1
00

Example 8: Solve for f(x) the integral equation


1, O~t<l + f s sin SX • au dx]
00
E
00 ax
fo ~t
)0
f(x) sin xtdx = 2, 1 < 2 Se
➔ 0 as x ➔ oo and using the bound·
{
0, t ::".: 2 Assuming ~:
Solution: By definition ary condition ~'.:. \x=O = 0, the first tenn in R.H,S,

rXJ {1.
Fs(t)=Fs { f(x)} = Jo f(x)sinxtdx = 2,
o~t<l
1~ t 2 <
becomes zero. Integrating by parts again

:t fo
00

0, t ::".: 2 u · cossxdx

= c2s [• · sinsxl: - J.
00
f(x) = -2100 Fs(t) sin txdt
rr o us -cossxd•]
20.14 - HIGHER ENGINitliilillJJL •Nliif'.~;;-r--:--:-~~
nt . .-----~
Finite Fourier cosine transform of f( x ) is Solution: This probl . t
. . em is o so\ve
dimen t· au
. sional heat eq ua 10n a, = c2 3 II f
2 the I)\\t ,
{L nrrx
Fc(n ) = Jo f(x)cosL d x L , with boundary
d . .
conct·t·
l ions U x
(0 a
) x
ihi l)r
, t ::::: Lt
a" X '
an wi th mitial condition u(x 0) - x\l , 1ho aJ1
= fo :rr x(rr - x) · cos nx d x th e boundary conditions are of ~he f (x) . Since N
(derivatives) apply finite Fourier co . eumann ty\le
. sine transf
sinn x both sides of the equation. l)tTnto
= [ x(rr - x) · ---
11
{L au . cos (nn x ) dx = c2 [L 32u cos (nnx \
_ ( - cos 11 x )
-(.rr - 2x) + (-2)
- sin nx ] rr lo ar L }0 ax 2 L )dx
2 3
11 n 0 (Ii

Fc(n ) = 0 -

Example 2:
(ncosnn +rr)
,.,
11-

Find the inverse finite Fourier cosine


= -n2
- n
[1 + (-1)'1].
a2u
Fe ax2
{
l Put

= [(-1)
U =
n
Fe {u(x, t)} and use

Ux (L , t) - Ux(O, t )]-
n27r2
v Fc(U)
(
2)

. sin( IIJT )
transf01mf (x )1f Fc(n) = 2; forn = 1, 2, 3, ... the equation (1) reduces to
and = ~ when n. = 0 in O < x < rr
-d
dt
lL
O
u -cos - - dx
L
(nnx)
Solution:
n2rr2 \
f(x) = Inverse finite Fourier cosine transform ofFe(n) = c2 [(-lluxCL, t) -
{
Ux(O , t)] - o Fc(U)\

L Fe(n) · cos (nnx)


00

= -l
rr
F e(O) + -n2
11=1
- .
L or dU
- -C
dt
_ 2 (-n rr 2

L2
2
) U (l,

= 1
rr .
n + n2 L sin (nrr) • ( 1n )
00
(nnx)
· cos -;- since the first term in the R.H.S. is zero because of the
4 n=l
2 2
zero boundary conditions (ux(L, t) = ux(O , t) = 01
Here U = U(n, t).
f ~ ! · sin ( nn ) ·cos nx.
(x) = ~4 + ~rrL....,n Solving (3), we get
2
n2rr2 c2 t)
Example 3: Find the inverse finite Fourier sine U(n , t) = Ae- ( ~ (~l
2,r(- l)n-l _
transform J(x) if fs(n) = 112 , n - 1, 2, • • •
To determine the arbitrary constant A in (4) use th!
when O < x < rr.
initial condition.
Solution: Put t = 0 in (4) this yields
00
= Fe {u(x, 0)} = Fe {f(x)}
J(x) = J; 1
{F5 (n)} = 2
L L Fs(n) · sin (nrrx)
L A= U(n, 0)

(mrx)
l.
(5)
= lo[L
n=l
f(x) • cos L dx
2rr ( - on -1
= -2 ~
00 {
- - - - • sm nx 2
Jr L.., n2 L (mr x) ] -n21t2c21/L
n= I

oo (-l)n-1 .
:. U(n, t) = [la 0
f(x) • cos L dx e
~
f (x ) = 4 L n2
smnx .
n=l
Taking inverse finite Fourier cosine transfol'IIl
Example 4: Find the temperature distribution in a u(x, t) = F;- 1 {U(n, t)}
bar oflength L , with its both ends and lateral surface
1 2 oo (n1rx}
insulated when the initial temperature in the bar is = -Fc(O) +- LFc(n)COS L
f (x). Deduce when f (x) = x 2 and--L- = 10. L L n=l

~~
. ~ L - --.;.:;: ER TRi r....s -
·c Fc(O) = { f(x)dx .. . . .JNil:.:G
::=-::RA
~ L-T-IW,SF
1v11c1 lo .
'and
Fc(n) = Fe {u (x,
I 1L t)} = U(n , t) g .
iven by (6)
Put U :::: F ( ( )} 1l
(nn.x)
s u .x,y === o u(.x,y)sin - l d.x
fJttlS
u(x, I) = -L
O
f(x)dx + :::: U(n , y )
00
Then c12u

<l,t
+L L
2
n= l
[

1
O
L
f(x)cos ( ~
L
)
dx
]
x
-dy2 + () - L
/77[ [
u(.x , y) • cos lllTC.X IL
0
Tl lrX 2
11 TC { L
Cl) x cos L e - < 11 2
1r c 2 r;L 2) +7: lo u(.x,y) -sin C:.x )c1.x] === o
2 (7) 2
d U nn:
(2) ,,,,n1e11 J (r)=x
· , andL=lO ,
dy2 - L [u (L, y)( - l l - u(O, y) ]
c-(tt) r 10 x3 10 n 2TC 2
Fc(O)= Jo x2 dx =3 =~ - v u === O
0 3
~s~~g the boundary condition u(L , y ) = u(O , y)
U(11. O) = F, {u(x, 0)) = folO x' . cos ("~x) dx
d2U
dy2 - ), = ~
2.(-10 0
)..2U === where
2 2 (1)
=x . (~).sin(~)_
n.n 10
) L
x n 2rr2
The s?lut_ion of this 2nd order ordinary differential
equat10n 1s
X COS c~x) + 2 ( ~,1::) sin n~x 1:0 U (n, y) === A cosh ·Ay + B sinb >-.y (2 )
(3)
= 2000 (-1)'1. Using the boundary condition u(x , 0) = O
n2n2
the 0 = U(n, 0) =A· 1 + B · 0 .-. A= 0 (3)
00
,. 0) 100 2 2000 I )ll
Thusu (x, t ) =
3 + 10
· ---;;z- L~ (-
Using u(x, L) = f(x)
n=I U(n, L) = Fs {u(x , L)} = Fs (f(x)}
X e-(n2n2c2t / IOO ) . COS (n;Ox). = foL f(x) sin c~x) dx (4)
(4) . ,Pu
Example 5: Solve the Laplace's equat10n ax2 + Substituting (3), (4) in (2)
211
the ~
oy2 = O insquare piate of length L with the fol-
U(n, L) = B · sinhH
lowing conditions u(O , y) = u(L, y) = 0, u(x O) .
0, u(x, L) = f(x). Deduce when J(x) = x 2 and
U(n, L) (5)
B = sinh>-..L
L-==n.
Thus (2) reduces to
Solution: ·
Since the boundary condi ttons are
.
(5) n·IIlchlet
· type (u prescribed), take fi mte· FOurier sme
U(n, y) = B. sinh>-..y

s
transform of the Laplace equation on both ides d). given by (5) and (1). -
where Ban . are fl ·t Fourier sine transfom1 of
2 Talcing the mverse ru e
{La
lno -a u. (nnx)
2 sm - dx + a2u .
- 2 sm L
loL (~)dx==O (6), we get .
(6) x L ay O 00 ,UT.\

Integrating by parts - F - l{U} = -2 ·l...J U,,,1..v) . sin - L°""'


u(x, y)- s L n= l
~-.
ax (nnx)jn
-L
nn loL aau -cos(~)dx
- -L
SlQ L 2 00 • (~)-sin(~) (?)
0 0 X
) - - "°'B•smh L L
u(x,Y - L /.J
+~ (L u(x
ay2 J0 •
y) sin(~)
L
== O
n=l
I

20. 16 - HIGHER ENGINE_IR __MA:.


_I_NG ~_ ~;..;;::ncl
_ IMA:
:rH .:: _______ __ _~ ---....:
::..;.;;;:;;:;___ V f0 1

\\' hL'H ( (.1) 1' .llld / - ;r. Crom (I). " l' )!L' t c1 • n ctcnni nc the i11vc1sc F.F.S T

1 " l. 2, J . ... and O < '< -- I>\ 8. !.¼:Jr


l/ (11 . l ) l'
. ()
(( 1),111(
11
; )i11 \
I 1/ S. I
.
• (.\)
,
-= -~ ""
L,
"\,l
l h( I )"
I
I
' Ill nrn
~ -
- R
~

11 - \ II

7 · Determine the inverse F.F.C.T. of ~ - -

-+ 2 ( ➔ II
\) • cos 11 1 I,()
for 11 = l, 2. 3 .... and equals to 1 r~r" ~
in0 < X < 4
1

rr 11::::;
00
- ;r ·\ - 1) 11 ') 2
(8) Ans. ·f'(x) = .!.4 . lrr + ~4 "L.., 6(sm(,m/2)- cos1t:t)
(2n + l)rr
1 .
L ~II .
l• ) = - -II- - _I
7
II
-
~ ( - I )II -;-
11 ·' 11 = 1
C.05(•,
\'j'

Substirnting (5) . (8) in (7) . we get 8. Determine F.F.S.T. of (1 - cos n:r) ..


11: ~-
where O < x < n
Il l,\. \')
·
-
- [ (-
. -l 11,;r
= ;r"'~ smh "
-..J
1) II (?
-.c. - rr~?
11 3
) ?]
- --
11 3 Ans. f (x) = }3 f (l-~~inrr) sin nx
II I 11=1
>-- sinh 11y • sin 11.,
Solve the one-dimensional heat equation ~~ == 2~
w~ ~
Qb:;er-inQ
EXERCISE 9. u(O, t) = u(4, t) = 0, u(x, 0) = 2x ~-fa
? oo 3?(-l)n+l 2 2 ' . ,
Ans. u(x ' t) = =- "
4L..,
-
/Irr
e-11 rr rr / 16. sm~
!
Find the fini te Fourier sine transfonn (F.F.S.T.) and 11=1
finite Fourier cosine transform (F.F.C.T.) off (x ): 10. u..(0, t) = u..(6, t) = 0, u(x, 0) = 2x
Thus L
1. f (x) = 2x in O < x < 4 Ans. u(x, t) = 6 + 2rr-~ °'L..,"' - ,-
oo ( 1)11
11-
I , ' .,
e-,nr-rr /
36
x
Coroll
11=1 , I
32 (- 1)11+1. F.F.C.T.: 32 ((-l)11 _ l)
Ans. F.F.S.T.: ~~ : -. Tu,
srr X COS~

!:~
6
0 < X < 7f /2
+ ~ = Oin
l=
l
11. Solve the Laplace's equation
2. f (x) - 1'. 7f / 2 <X<TC
square metal plate of siden with u(x .·rr) = ~-
Ans. F.F .S.T. (-2 cos(srr /2) + 1 + cos srr /2)/ s u(O, y) = 0, u(rr, y) = 0. 11(x, 0) = 0. Ass~
F.F.C.T. = 2 sin(sn / 2)/2, s = 1, 2, 3, ... u (x, y) is bounded
3. J(x) = x2 rn O< x < L Ans. u(x, y) = 4110rr "
00
sinlr((211+l).'·)sin(:!11+n.t
L.., (211+l) -sinlr~C::11+l\~rr)
Ans. F.F.S.T.: - L co~ nrr + ~ (cos nrr - l ),
11rr II rr
3
n = 11 = 0

1, 2, 3, .. . 12. Find the displacement y(x. t) of a s~ 3-,·


of i.e.,
F.F.C.T.: 2L 3(cos nn - I)/ n 2rr
2 length rr governed by the wave equation a,! ~
4. Find F.F.S.T. off (x) = ea). in (0, L) c2 a2 v
one end fixed y(rr. t) = O, ·uu'tiallY
axf with

Ans. srrL[(-1)s+I •eaL+ 1]/(a 2L2 +s 2rr 2] at rest y(x, 0) = 0, with initial displaceillClll
11(0, t) = a sin wt.
5. Find F.F.C.T. of f(x) = "2Jr12 - Ir.
6 in (0 -rr)
l JL
Ans. y(x, t) = a sin wt . sin w (1r~x) cosec (!f) +
Ans. (-l)"/n for n 2
= 1, 2, 3 and zero for n = 0 00
(2a;,c) Z:(w2 _ n2c2)-1 sinnx sin net.
rr=I
FOURIER INTEGRAL' POURIER TRANSFO>IRtaM~St•-;;;;;;-;;:~ ::--=:'""--:- ~ ~ - ~ - - - -
pARSEVAL'S* IDENTITY F & INTEGRAL TRANSFORMS - 20.17
t0·5 FOURIER TRANSFORMS OR

• f(O') and G(a) be respectively the F . Parscval's identity for Fouri er sine transform are
It 1 ,f ((X) and g (x). Then ouner trans-
¾ft F~(a)G s(a )cla = ft f(x)g(x)dx
. (Ill~ l
[ll

.( ~
·

..._ F(O!)G(a)da = -2JT1 f ex-,


_ 00
f( _
x)g(x)dx (I)
(c)

(d) ¾fa°° \Fs(a)\ 2 cla = fa"" \f (x)\2 dx


0 baf implies the complex conJ·
,d1crc ' u gate . These results can be proved on similar lines as above.
f· Using the inversion form 1 f
proo · . --:-) · tl R u a or Fourier
rr:uisfoon for g(.1 m 1e .H.S. of (1),

Cwig(Xldx - i: f (x) ( l: G(a)e;., da l dx


WORKED OUT EXAMPLES

Example 1: r00 <x cos xx6-~m d cl X -_


Prove that JO 2.
15.
Interchanging the order of integrain in R.H.S .

== 1-: {i: G (a ) f(x)eiaxdx} da


Solution: Consider the function j (x) defined as

f(x) =\
l a.2 - x2 if \x \ ::: a

Observing that inner most integral is the Fourier 0 if \x\ > a


{i1ll1Sfonn of f (x) , we have Fourier transform of f(x ) = F{f (x)} = F (e1. ) =
:x;
= 2rr Joo l Joo
27[ ·
-00 f(x)e-l~Xdx = 2~ + f~a + u~: J; 1-
f
::: -:x: G(a ) {2JT F (a)} da - oo F(a)G(a)da Smce f(x) = 0 m (-oo, -a) and (a, oo), we have

= -1 fa (a2 - . l
Thus i f~oo J (x)g(x)dx = J~ 00 F(a)G(a)dt(, F(a)
2Jr -a
x 2 )e-,axdx = -l/1
2rr
- h1

Corollary 1: Put g(x) = f(x) and note that zz = where


: 2• Then (1 ) reduces to

in
o,
L: F(a )F(a)da = 1-: \F(a)\ da
2

e = -1
2rr
J 00

-oo
J(x)f(x)dx
and integrating by parts.

= -1 Joo \J(x)\ dx 2

2rr -oo

of
te..
I f x
-')Q
\F:a) \2 da = -1
2rr
f 00
-00
\f(x)\ dx
2 (2)

a2 [ ja(I - iall1 + ~ rt'Hhl t t'-hl~1 +


y ==-e
ia
-e
a
iL
Results (2) and its generalization (1) are known
ot
~ Parseval's identity for Fourier transforms 2 ( - 11111 - i'"1
rintegrah;). +:-:-
,a·'
e
C~~
c . ary 2: The Parseval's identity for Founer
. So l
..!.. r~
2 .
os1ne stn
• l (2cosa,1)+ 32•SlUaa 1
e transform are F(a) • f,r [Ii -'21•2,r l al a·
(a) l roo
nJo Fc(a)Gc(a)da = J;° J(x)g(x)dX
)== +n'oJ
2 [ainC.Wa - aa cosaal
>--
Marc An . . ..
toine Parseval (1755-1836) French mathemaucian,
F(a

· ~~
20.18 - HIGHER INGINl~f@IM'IHIMAMl-!..:..
V- - -:e~~.;._'----.~ -~
Apply Parsevar s identity for f(x) and F(a). Thus ~
C
N
IF( )l2da =
100 - 4 (SillaO - O<XCOSCW)
~
2
da
r,o
Jo (a 2
x
2

+ x 2)(1J2 + x2) d x == ~1t (


1 -N a - oo :,r 2 a( ·
= b = ['.ro 2
a --1-b)

= l oo I
_l
'2rr -oo .
f(x)/'1.dx
For Q 00
(
?:+I
x )
clx = l-

= -I
2rr
la-a
(/a 2 -
2
x /)2dx
EXERCISE

= -l
2rr
la +
-a
(a 4 x 4 - 2a 2 x 2 )dx
Solve the ~allowing problems using Parseval's.
tity for Fourier tansforms. tden.

oo 5111
· 2
=
l [
2:,r a4x
'"5 x3 ] a-a
+ j - 21!123
1. Show that !c 0 ~
ax dx = n:a2
Hint: Consider f (x) = 1, \x\ <; n•~'fF{ODU
0
and f(x) = 0 for lxl > a. Then
~5 - ~)
1 2
= -2:,r · 2a ( 1 + F{f(x)} = F(a) = 2si:aa' apply P.I. then oifferenc
3
a (l)2dx = __l_ Joo 4sin2 aa da
16a 5
15
1
· 2rr =
8 as
15 rr
f -a
2. Evaluate (a)
2n: -oo a2

ft (a 2+x 2~;b 2+x 2) and hence find ~:::~le


rather th .
Thus (b) ft cx/;1)2 · rial equa

?
-
1~ 0
00

Jr
2
(ax cos ax - sin ax )
3
2
dx
= 8 as
15 rr
Hint: Fc{e-ax}
use P.I.
= a2
+a 2 ,
Ci
Fc{e-bx} = ----b/
,
ta- '
tion relati
. ... AstrJ
X
equally SJ
00
(ax cos ax - sinax)'-
']
rras Ans. (a) 2ab~+b) (b) For a =b= 1, ¼- Recurren
or
1 o
-------dx
x6
=
15 3. Prove that ft (s~x)4 dx = t by powe1
enceeqi
. I

For a = 1, we get the required result. Hint: Take f(x) = 1 - Ix\, for \xi < 1,0,
to differ]
otherwise, F {f (x)} = [f (l-~~s a), use P.l.
Example 2: Evaluate Jo
roo x2 d
x and z-tram
r
(a 2+x2)(b2+x2)
4. Evaluate (a) f000 (1- ~osx) 2 dx (b) ft s~~xdx
hence find ft C2:1 dx. Hint:._ Take f (x) = 1, for O ::; x < 1, 0 oth·
Solutiot
enceeqi
erwise, then Fe{/ (x)} = 1-:osa, Fs{f(x)) ==
Solution: We know that if f(x)=e·-ax then Fourier analysi:
sina a , use PI
..
sine transform of f(x)=F5 {f(x)}=Fs(a) = --r2· commu
a +a
Similarly for g(x)=e-bx then Gs{g(x)} = b2 :a 2•
Ans. (a) I (b) I
Recall Parseval 's identity for sine transform 5. Evaluate r 00 sinax dx
21.1
Jo x(a2+x2)

-2100 Fs(a)Gs(a)da = 100 f(x)g(x)dx Hint: Fc{e-ax} = :b, Fc{g(x)} == ~


a +a Differ,
7r O 0 where g(x) = 1, 0 < x < a, zero for x > a. define
use (a) in cor. 2.
functi1
Ans. ~2 (
2
1-e-" )
a2

-
2 -(a+b)
x=O

You might also like