Sie sind auf Seite 1von 195

EDIT|ON2009

Differential
Equations
Module

Normah Maan
Zaiton Mat Isa
Halijah Osman
Sharidan Shafie
Khairil Anuar Arshad
Department
of Mathematics
Facultyof Science,Universiti
Teknologi
Malaysia

DifferentialEquations
Module

by
Normah Maan
Halijah Osman
Zatton Mat Isa
SharidanShafie
Khairil Anuar Arshad

Contents
First Order Differential Equation

1.1 Basic Definitions and Terminologies

1.1.1 Solution of Difierential Equation


t.t.z Initial and boundary value problems

1.1.3

Forming a diferential equation

L.2 TVpesof First Order ODEs and Methods of Solution .


1.2.t SeparableEquation

5
8
8

L.2.2 HomogenousEquation

L2

1.2.3 Exact Differential Equations

18

1.2.4 Linear First Order Differential Equations


1.2.5 The Bernoulli Equation

26

1.3 Applications of First ODEs

31
34

1,3.L Newtcn'scoolinglaw .

34

t.3.2

36

Circuits Model

1.3.3 Motion along a straight line . .

38

Second Order Differential Equations

4L

2.1 SecondOrder Linear Differential Equations

42

2.2 Solution of HomogenousEquation

44

2.3 Solution of NonhomogenousEquations

49

2.3.L Method of UndeterminedCoefficients

49

2.3.2 Method of Variation of Pararneters

58

2.4 Applications. .
2.4,t

Mechanicalvibrations

2.4.2 Electric Circuit

63
63
67

Contents
3 The Method of Laplace Tbansforms
3.1 The Laplace tansform
3.1.1 Linearity of the transform
3.L.2 Tlanslation in s .
3.1.3 Derivative of the Laplace transform
3.1.4 Laplace transform of derivatives
3.2 The InverseLaplace Tlansform
3.2.L Ttanslation property for inverseLaplacetransform
3.2.2 Completing the square
3.2.3 The method of partial fractions
3.2.4 Convolution theorem
3.3 Initial Value Problems
3.4 Boundary Value Problems

3 . 5 SpecialFunctions

7L
72
76
78
80
82
83
85
87
89
92
95
97
100

3.5.1 Unit Step (Heaviside)Function


3.5.2 Dirac Delta Function

100
10g

3.5.3 PeriodicFunction

LL2

Fourier Series
4.t Periodic Function
4.2 Even and Odd F\rnctions

115

4.2,L Even and odd functions


4.2.2 Basic preperties of the even and odd functions

4.3 Fourier Series


4.4 Half-range Series
4.5 The Sum of the Series
5 Partial Differential Equations
5.1 Basic Def,.nitionsand Terminologies
5.2 The Metbod of Separationof Va^riables
5.3 The Heat Equation
5.3.1 Rod with zero temperatureat the endpoints.
5.3.2 Rod with insulated endpoints
5.3.3 Rod with mixed bounda^ryconditions
5.4 The Wave Equation

116
116
7L7
118
119
L25
130
135
136
139
L42
L42
150
157
161

Contents
161
5.4.L String with fixed endpoints
168
5.4.2 String with free endpoints
L7L
5.4.3 String with mixed boundary conditions
t72
5.5 The Laplace Equation
5.5.1 Prescribedtemperature at the boundaries- type A . t72
5.5.2 Prescribedtemperature at the boundaries- type B . 178

Preface
This moduleis written for undergraduatedifferential equationscourse(SSH 1743/
SSII 2713/ SSE 1793) taken by students majoring in mathematics, mathematics
educationor engineeringprograrnmes.The prerequisiteis the introductory calculus (SSH 1713/ SSE 1792).
The main goal of this module is to free students from notes taking during the
lecture sessions,so that they can concentratemore on understandingthe theories.
We want the module to be easily read and understoodby the students. Therefore
our presentationsof the theories are as simple as possibleand emphasisis more
on how to usethem. Besidesthat the students a,realso encouragedto do on their
own all the quizzesprovided in each sub-sections.
This module consistsof five chapters. Chapter t providesthe introduction to Differential Equations which emphasizeson the First Order Differential Equations.
SecondOrder Differential Equationsa,rediscussedin Chapter 2 and we confineour
attention io linear differential equa'r,ions.Chapter 3 gives the Laplace Tlansforms
method while Fourier Seriesare discussedin Chapter 4. Finally, introductory partial differential equationsare presentedin Chapter 5 which specificallyfocuseson
solving the Heat, Wave aud Laplace equations using the method of separation of
va,riables.
On completingthe course,students should be able to:
o Find the solution of first order differential equations.
o Find the solution of secondorder linear diferential equationswith constant
coeffrcientsusing the method of undeterminedcoefficientsand the method
of variation of pa,rameters.
o Obtain the Laplace transforms and im'ersesof the unit step function, the
Dirac delta functions and periodic functions.
r Solveinitial and boundary value problemsusing Laplacetransforms.
r Determine Fcurier seriesof given functions.
r Solvewave, heat and Laplace equationsusing the method of separation of
variables.
To the students,if you want to sail through this coursewith flying colours, heed
our advice: "the only way to leam mathemat'icsis to do mathematics"' Good
Luck!

Chapter 1

First Order Differential


Equation
Objectives
On completingthis chapter,the students should
be able to:
1:

classify the d.ifferential equation according to the order, degree


and the linearity of the given difierential equation,

2:

identify the different form of the first order difierential equation


given,

3:

find the solution ofthe first order equation accordingly.

In introductory calculusthe studentshave learnedthat given a function


y: f (s), then ihey have to find its derivative that is

9:
l'@).
dx '\
But the problem that we are concernedin this courseis not: given a function
A: f @), find its derivative. Rather, our problem is: if we are given an
find.a function a : l@\ which satisfiesthe
equationsuch as
H:2ty,
equation. In simplt-words, we wish to solue adifierential equation.

First Order Differential

1.1

Equation

Basic Definitions and Terminologies

of one
Definition L An equationcontainingthe ileriuatiuesor differentials
aarior tnore ilepenilent aariables,with respectto one or rnoreindepenilent
ables,is sai'ilto be a difierential equation(DE)'
propDifferential equations are classified according to the following three
erties:
1. Classiffcation bY tYPe:
An equationthat contains
ordinary Differential Equation(oDE)
only ordiuary derivatives of one or more dependent variables,
with respectto a singie independentvariable'
Example t

OO

#-s':t
Example 2

^(#)':tff+u
ExamPleB
fu-sg+6s:o
-dx '
dn2

An equation involving the


partial Differeni;ial Equation(FDE)
partial derivatives of one or more dependent variables of two or
more indePendentvariables.
ExamPle 4

ou

ou

ag: a*

B;arnple 5

,0u

0u

' ' a ' + 0 6 :u

1.1 Basic Deffnitions and Terminologies


Example 6
02u 02u n}u
-0s2:M-"A t
2. Classification by order/ degree
order of Differential Equations
ative

Determinedby the highestderiv-

Degree Exponent of the highest derivative


Example 7 g" + 3ty3 : L is secondorder first degreeordinary
difi erentiaI equation.
Example a (t/')3*z5siny - a4 is first oriler third degreeordinary differential equation.
3. Classification as linear or nonlinear
A d.ifferentialequation !s said to be linea.rif it has the form

: g(r)
o*@)#* o^-r@)#+ ...+ "r@)#* as(n)s
In words, a differential equation is classifiedas a linear differential
equation if the dependentva^riable,y and all its derivatives are of
the first degreeand eachcoefficientdepenrisonly on the independent
',,a,riabler. Otherwisethe equation is said to be nonlinear'
Example I The equations
xd'g* gdt :0

a" -2a'* a:0

+r!u *5s : s a
,'4dr3 -,'*
'dr2'"dr
are I'i,near first, seconil anil thi,rd order ODE respectiuely'

First Order Differential

Equation

Example LO The equations

&a ^du
:
Yd i ; - ' #
'
and

4
* y 2 r :o
drz'
are nonlinear secondorder ODE respectiaely.
As mentioned before, our objective in this courseis to solve,or find
solutionsof a differential equation.
1.1.1

Solution of Differentiat

Equation

Definition 2 Any function y : f (x) defined on sonle interual I which


satisfiesthe giuen differential equationis said to be a soluti,ono! that differential eqitationon the interual.
Example lL Thefunction A : e-a/2 is a solution of the fi,rst ord,erlinear
difrerential equation
ZY t+Y :9.
Since

Idn2: -L"-'1'
we seethat

z# * o: zFI"-,/2'1
s e-,/2: 0.
There a,retwo types of solutions which are:
General solution solution with arbitrary constantsdependingon the order of the equation.
Particular Solution solution that satisfiesgiven bounda,ryor initial conditions.
A solution of a difierential equation that is identically zero on an interval
.I is often referred to as trivial solution.

1.1 Basic Definitions and Terminologies


1.1.2

Initial

and boundary

value problems

Initial conditions(IC): Condition specifiedat a given point. The number of


initial conditions dependon the order of the DE.
Boundary conditions(BC) : Condition specifiedat someor difierent points.
The number of bounda,ryconditions again dependingon the order of the
DE.
Initial value problem(IvP) : A DE with initial conditions.
Boundary value problem(BVP): A DE with boundary conditions.
Example 12 Initiol aalue problem
a" +2a'*Y :coss
IC: Y(0): 0, Y'(0): 1
Example L3 Boundary aalue Problem
a" +2Y ' *Y :cosz
BC : Y( 0 ): t, a ( L):2
Example L4 Initiol aalue problem
a"'+2a'*g:coss
I C : y( 0 ) : 1 , g '( 0): 0, A "(0) :2
Example t5 Boundary aalue Problem
g"'+2a'*a:cosx,
B C : Y( 0) : 1 , Y( 1):0, A (2): -L

First Order Differential

1.L.3

Equation

Forming a differential equation

Mathematically,
a differentialequationcanbeformedfroma givenfunction.
Example 16 Form a suitableDE from thelollowinglunction
A
A: f i- - .

(1 . 1 )

Solution
Sincethe given function has only one constant A, we difierentiate the function once with respect to x. Thus we have
dy
A
:, -'nz'
*
d,r

G'2)

The resulting equation is a DE and the next task is to eliminate .A, or


otherwise express,4.in terms of r and gl .
Divide (1.1)by o:

a:L -4.
n
n2'
Eliminating .4 by finding the sum of (1.2) and (1.8), rhat is:
dy,A

fi * L:L

(1 ' 3 )

(1.4)

Note that the order of DE dependson the the numberof arbitrary constants
appeared in the given function. Thus if the given function two arbitrary
constants,we would have secondorder oDE, etc. Also, note that there
should not be any constant appearedin the DE.

1.1 Basic Definitions and Terminologies

Quiz 1.1
L. In Problems (a)-(b) state whether the given difierential equationsare
linear or nonlinear. Give the order and the degreeof each equation.

(') (r- d# - 4"#* 5s:6esg

@ , #- r (#f*s:o
(c)

&r k
d7

2. Verify that the indicated function is a solution of the given difierential


equation.
(^) Y' * 2ag :2 * n2 + Y2i

U : x.1 tanr

First Order Differential

L.2

Equation

Types of First Order ODEs and Methods of


Solution

In this course,we will introduce five difierent forms of first order oDEs.
we begin our study of solving a first-order equation with the simplest of
all differential equations.

1.2.1 Separable Equation


Deffnition 3 A generalforrn of a first orderODE is giuenby
dY

Fr
fr: f@,il.

(1.5)

Supposef (*,d : g@)/h(y), whereh(d * 0, then we can write (I.i) as


da _ g@)
dr
h(y)
and i,ssaiil to be separableor to haaeseparabieuariables.
Method of Solution
Observethat a separableequation can be written as

nti#: s@)'

(1 . 6 )

Now if y: i@) denotesa solution of (1.6), we must have


h(f (n))tt@): g(*),
an<itherefore

ff

JWAnt'@)da:
Jo@)tu+c.

( 1.7)

But dg = f'(r)d,x so (1.7) is the sameas

Iorrron:lo@)d,,+c.

(i. 8 )

Equation (1.8) indicates the procedure for solving separabledifferential


equations.

1.2 Types of First order oDEs and Methods of solution


Example t7 Solaeydn - r2d,Y- g
Solution
Multiolvine"-bv
a t"

j; 1 we can write

Ioo:#o*,
from which it follows

: \a"
|t'
lnlsf *c1 : -l*.,

rolsl : -l*,
w h e r eC:c2- ct.
Notes: In future, we will write only one constant'

ExampleL8 Solue#: #
Solution
The equation can be written as

\
do:(j
dn \t +7 )a"
from whictr it follows

au:

12

\, *A)

d"

Integrate both sities,


12 \

J ;,tu:J l * a) n
f L

f /

-i:.-lfin
-1 -z-tan-1

r*C

10

First Order Differential Equation

ExampleLg Solue
u"Pr: r,
Solution
The equation can be written as

fr
dA: | 2re-, d,x,
I
JJ
from which it follows

u : -2 n e -'-2 e -0 +C

11

1.2 Types of First Order ODEs and Methods of Solution

Qrriz L.2
In this quiz the students must be able to rearrange the differential
equation given in the separable form first.
In Problems 1-4 solve the given differential equation by the method of
separationof variables.
I

1. :s _ e30+2y
afr

2. 2 y ( n+l) d ,g:s6s
3 . secndY : tcotY du
4.

*2a2
da
d , a 1 +r

Ans:{ :

+C

A n s:Y2:n- Inlo*Ll+C
Ans: -lncos A :

tsins * cosn * C

A ns :-| :+ - c * ln 1 1 + 0 1

+C

First Order Differential Equation


1.2.2

Homogenous Equation

Definition A If fQ,n,)g) : \"f(s,y), tor some real nurnbern, then


l@,y) i,ssaid to be homogenous
function of degreen.
Example 2O f (r,g) : 2rsy - 4*ty' is homogenous
function of ilegreeI
since
f Q,r,\y)

z((,\r)3(,\y)- +(>,a)2)(xy)2

2\an3y -4\au2y2

\4(2a3y- Ar'a')

f(\a,\Y ):

\af(r,s ),

w he r en :4.
Example 2t f (a,A) : A2- s is not htomogenous
lunction since
f (\r,\y)

Definition

(\y)2 - ),r

\2y2 _ ),r

\"f (s,y).

5 An equationin the differential farm


dy

rt_.\

M (r, y )

#: I @' v ) : ffi

(1 .e )

is a homogenous
equationif M(x,y) ond,N(r,y) are homogenous
functions
ol the same degree.
we can also define the homogenousequation by the following definition.

Definition 6 The firstord,erdifierenti,al


equation
#*:
as thehomogenous
:
equation
if tQ,x,)y) f @,d.--

f (r,y) i,sd,enoted,

1.2 Types of First Order ODEs and Methods of Solution

13

Example 22 Determi,newhetherthe followi,ngequationi,shomogenousor


not.
d,y_zt/nY - Y
d s fi

Solution
Comparethe given equation with equation (1.9). We can write

dy ztFA-a : M(a,A)
N 1* * 1
,
d* :and N(r,g) are homogenousfunc-

we need to determinewhether M(r,il


tions of the samedegree.
'Iake
M(r,y)

21pg-g

M(\r,\y) :

(1 ' 1 0 )

ZtN^s-XY

: zJ@-xa
:

),(2.{ry - y)

).M(n,g),

and

N(.b,.\s) :

\r

N(*,9)

,\N(r,g).

ThereforeM(x,y\ and N(u, gf)axehomogenousfunctionsof degree1. Hence


the equation is homogenousequation.
We also could usedefinition 6 to showthat the given equation is a homoge-

L4

First Order Differential

Equation

nous equation. That is

f @, il: ' +" F i- Y


f(M,\y)

l.@ _ \ g
=---i"

: zr/w - ^y
.\r
:

:
Therefore

- y)
^(Zr/ny
),r

f (x,il'

d'v-2tF0-a

d rfi

is a homogeneousequation.
Notes: If f (r,g,) is a homogeneousfunction of degreen, notice that we
can write

(1.11)

t,!)
!(s ,v ): x "I(
' ' ,I,

Exanrple 23 We seef (r,U):3s2 *2xg*y2


Thus

is homogenous
at degree2.

t@ ,y ) :* ( t* r # ). (1)')
: y6,f,).
"r
Example 24

dy _ s2+a2
dn
3ny
can be wri,tten os

x2 (t + (y/")')
_M_

dy

d"
:

l$,ylx).

(1 . 1 2 )

1.2 Types of First Order ODEs and Methods of Solution

15

The important point in the subsequentdiscussionis the fact that a


homogenousdifferential equation can always be reduced to a separable
equation through a sirple algebraicsubstitution'
Method of Solution
An equation of the form
d rt

M (r,A)

7*:f(r,il:-ffi'
where M and,N have ihe same degree of homogeneity can be reduced to
uc, where tr is a new dependent
separable variables by the substitutionY:

variable.'thenffi : u+affi.Hence the difierentialequation(1.9)becomes


da

u * x 7r:

M(r,at)

W ,nr),

where this substitution will lead to a separableequation'


equationof Emmple 2l'
Example 25 Soluethe homogeneous
Solution
By inspectionEquation (1.12)'
rht

fr2 + A2

7 r: f (*'Y):6

:
is homogeneousequation of degree2 and use substitution 3r ur and
du
du
*'7*, Equation(1'12)will become
#:
"
d,u
:
u *:D-

i + (yl x)2

dT

u * fi -

du

!+a2

afr

3u
!*u 2 -3 u z
3u
l -2 a 2
3u'

du
d,fi

du
dfi

First Order Differential Equation

16

which can be written as separable equation of


3u1
du : - ct':D
t -W
To solve this equation we need to integrate both side of the equation, ie

3u

f \ar.

I m*:

J;

For the integral on the left hand side, use integration by substitution
to get

method, that is

f 3u '
J t-fr"o
f3

-A^

=, [!a".
J r
:

ln lcl+ C

-fr"1 "1: lnlc l+ c


Tbke exponent for both sidesof the equation,we obtain

ollnrlO

"-3/4lnu
:
"lnu-3/4

u-3/+ :

(L - 2ru2'1-3:

whereD :

"lnn"c

,"c
(1 . 1 3 )

g4D,

ec. Then we needto substituteback

|El

equation (1.13). Then the solui;ionof the differential equation is


s2 : D(x2 -zg'),
where D is an arbitra"ry constant.

into the

1.2 Types of First Order ODEs

"t

d M"thod"

of Solttti

t7

Quiz 1.3
L. In Problems(a)-(e), determinewhether the given function is homoge'
nous or not. If so, state the degreeof homogeneity'
n}o,

(u)f@,il:ffi- ''2A2
(b)/(",il: -n#*3 L

Ans: NH

rrS

Ans: H

.1 2

Ans: NH

(c) /(t, U) : cos


. +g
o,2 t42_gy

(d)/(", il: Y--h:

Ans: H
Ans: NH

(e) /(c, ,1: r/Gi@r +3a)

2. In Problem (a)-(b), solvethe given homogenousdifierential equations.


,\

d' y

Ans,$z :lna * C
Ans: gl: -rlns * Cr

Y.fi

\ a ) d n - - ;a
(b) (" - ildn * nd'Y: g

3. Solve the given difi'erential equation subject to the indicated initial


condition.

'*' # : Y a Y

* u2

; Y ( t): - 2

Ans: gt2:

4n(r * y)2

18
L.2,3

First Order Differential Equation


Exact Differential

Equations

The students need to understand the following extendedchain rule before


proceedingto definition of Exact Difierential Equations.
The Extended Chain Rule
Supposed@,y) is a function of two indepen<ieni;
va.riabless and y, where
6@,g) has continuouspartial derivatives with respect to c and y. If the
secondindependent va^riablegris replaced with a difierentiable function of
r, call it g(n),.we obtain a d(s,y(r)) which is now a function of o only.
d6^

ir

Nowwnarrs
fr!.
The appropriatechainrule is

f too,u@ ) ):ry.ryW

(1 . 1 4 )

Equaticn (1.14)is an extensionof lhe chain rule for the functions of a single
v-ariable.
Theorem 1 Consider the differential equation
)lq

M(r,y)* N(c,
ilffi = o.

(1 . 1 5 )

Let the functions M, N,0M/0y,


und 0Nl0r he enntinuousin an open
rectangleR of the x{-plane. Then the differentiat equationis eaact in R
if and only i,f
A M _A N
0y
0x'

(1 . 1 6 )

for all points (a,y) in R.


Example 26 which of the lollowing differential equationsis esact?
. du

r.y+n +a 7 ;:0
2 . 2a yd r * ( r 2 -L)dy:g

3.sins+ (2y*tcos
ilff : O

1.2 Types of First Order ODEs and Methods of Solution

19

Solution
1. Comparingwith Equation (1.15),we have
M ( r,A ):y+lt

N (r, y ): 6

and

Calculating the partial derivative, we find

W:t

an d

aN t

_:

0r

Therefore,by Theorem 1, the differential equation is exact.


2. Exact equation.
3. Exaci equation.
Theorem L provides an easy test to determine whether or not a given
differential equation is exact. The theorem however does not tell us how
to construct the solution of the equation. The basic idea underlying the
solving of exact difierential equationsis to reversethe exiended chain rule.
To that end, considerthe difierential equation of the form (1.15).
Notice the similarity between the form of Equation (1.15) and that
of the chain rule (1.14). Supposethere exist somefunction, $(n,y), that
satisfiesthe following two conditions:

H:

M@'Y) and

#:

N@'v)'

(1 . 1 7 )

Becauseof (1.17),w.ecan rewrite difierentialequation(1.15)a^s

P
*P 9: 0.
0x' }gdt

(1 .1 s )

By the chain rule (1.14),equation(1.18)is the sarneas

:0.
4o@,e)
AT
Thereforewe obtain the solution of equation (1.18)'

d@,a):c.

(1.1e)

If there is a function 4(r,g) satisfyingthe condition in (1.17)' then differential equation (1.15) is called an exact equation. If we can identify the
function 4@,a), the solutionof the exact equationis given by (1.19).

First Order Differential Equation


Method of Solution
Rewrite the equation(1.15) as
M(r,y)dr * /V(r, a)da:0.
The students have first to shour
AM

(1.20)

AN

W : a*'

Then assumethe solution of the equation (1.20)is g(n,d : C that satisfies

y:

Y : M@,a)
and
-'- \*t
0n

ug

"''

N@,0.

So we can find 6@,0 by either integrating M(a,y) with respectto s while


holding g constant or integrating N(o, g) with respectto y while holding r
constant. Let here, we consider the firsi; case.We write

6@,d : I u6,y) dr + h(y),


J

(1.21)

where the arbitrary function h(y) is the "constant" of integration. Now


difierentiate(1.21)with respectto y,
A d,
Af

+
o y:

*oluUJ ( r , s ) d , r + h ' ( y)

N (s,a ).

This gives

h'(v\: N(r,a) -

6 J

M(x,v)dx.

It is importantto observe
that theexpression
i,I(x,g)-(|1Ail
is independentof o since

(1.22)
f

M(a,y) dr

!^^l*a,il- * [ *@,ila,1: q{ - ! ( ! [ *6,y)a")


dxL.v'
0 a" ,
ar
ay\a"J'
J
/

: g{:s
.
ot
ou

Finally, integrate (L.22)with respectto y and substitute the result in (1.21).


The solutionof the equationis g(r,il : C.
Example 27 Consi,ilerthe initial ualueproblem

: 1.
L+y2
+2 @+t)yfl:s, s ( o)
Verifu that the ilifferenti,alequationis eract and solaethe problem.

1.2 T\rpes of First Order ODEs and Methods of Solution

21

Solution
By Theorem 1,
M(n,U) - I + A2

and

N(r,Y\ :2(r

+ l)U'

Then
AM
:'o
Uo

and

AN
: za'
6

sin c e
ry:o
!:za,
0n
0g
the differential equation is exact.
Now we would like to frnd,$(r,y) that satisfies

4:
0n

and

u@
---\--',y
e' )

*OA
:N (" , v ) .

we can fiI'd Q(n,gr)by integratitg M(n,g) with respectto u while holding


y constant. Let here, we considerthe first case. We write
f

4 @ ,i l :

l U (r,ddn+h( y ) ,
J
f

! L +yz d r,

a+ryz

+h(g),

(1 . 2 3 )

where the arbitra,ry function h(g) is the "constant" of integration. Novr


difierentiate (1.23) with respectto y,

a6

: Zny+h'(y)

fr

N(*,g).

This gives
h'(il

2(t+L)Y -2tY '


2y.

Finally, integrate (L.24) with respectto y, gives

h(v \:a2+c

(1.24)

22

First Order Differential Equation

a^ndsubstitute this result in (1.23). Hence the solution of the equation is


Q@,y):fr*svz +A2 +c:C
or we couldwrite the solutionas
r+rg2

+a2+K : 0 , .

that K = c- C is the arbitrary constant.


Imposing the initial condition, y(0) : 1, we obtain
K :-L.
Therefore,the solution of IVp is
r+ry2

*u2:1,

Example 28 show that thetollowi,ngdifierentiat equati,ons


is exact.Hence
find the solution.
(2n2 + 2ay * a2)dn+ (r2 + 2xy)d,y: e

Solution
Step 1: Show that the equation is exact
By Theorem L,
M(n,il : 2r2 + 2ny * y2

and

N(s,il : n2 + 2ry.

Then
AiuI

W
Since

:2u * 2Y

and

A/V

#:2n

aM atr
w: ar:2x*2Y'

the differential equation is exact.

*2y.

1.2 Types of First Order ODEs and Methods of Solution


Step 2: Find 6@,il that satisfies

9:
0f

up,y'1 and yO U : N@,a).


- ' - \ - ' d'

Now take

M@,g):2n2 * 2rv-r s2

#:
Integrate Lt(rrg)

with respect to r, we obtain,

d@,y) :

I zr, *2rs + y2dr

2r3

! + a r z + y 2 r + h ( s)

Step 3: Find h(A) bv difierentiating / with respect to g, we obtain

y : x2+zry + h'(a)
dy

Compare the above equation with.l[(r,y),

we get

h'(a):o
Thereforeh(y) :

"

Step 4: Rewrite $@,il,


t-3

d@ , i l : + + a r z + y 2 t l c : C
or the solution of the exact equation is
,-3

d@,y):?+s*2ry2r-K,
wh e r eK: C- c.

23

24

First Order Differential

Elquation

Example 29 Show whether or not that the following ilifferential equations


is exact. If so, fi,nd the solution.
( "" -ycosxy) d,r+(2se2Y-uc o s a g * 2 9 ) d , y : e

Solution
Step 1: Show whether the equation is exact or not.
By Theorem 1,
M(s,y):

e2a- ycosfiy

and

N(r,y):2xe2u - ocosxy *2,g.

aM
: yxsinty - cossg
Ay

and

aN
o-.
:2e"s + gasinxy - costgf.
i

Then

Since

aM raN
0y ' 0n'

the difierential equation is not exact.

1.2 Types of First Order ODEs and Methods of Solution

26

Quiz 1.4
1. In Problems (a)-(b) determine whether the given difierential equations are exact. Ifexact find the solutions.

(u)4 a, -*
y a"

da : o
Ans:Exacttt:C

(b) (1*In t+!)

a": (1- rnr)d,y


Ans: Exact;-g + glnc * rlnr:

2. Find the valueof /cso that the following differential equation is exact.
(a) (6xys * cosg) da + (kr2g2- xsiny) dy :0

Ans:/c:9

(b) (2ry2 * Ae') d,x* (2a2y{ keo - L) dy : e

Ans: /c : 1

First Order Differential Equation


1.2.4

Linear First Order Differential

Equations

In Section1.1 we defined the generalform of a linear differential equation


of order n to be
i:,r
, ,ilu
, ff-ru
:
""(")ffi + ant@)ffi + ...+ "t(")# I as(x)yg(n).
Now when n : L, we obtain the first order linear difierential equation

" t @ ) # * * a s ( n ) y:s( n ) .
Dividing by o1(c), gives
du

#+p(* )y :
oo(")
,,,L^-^
^/-\ :wnerep\a)

;Ta

q(n)

(1.25)

9@)

^r^t :ancr
--, g(0)

ffi

We seekthe solution of (1.25) on an interval f for which p(r) and g(c) are
continuous.
An integrating t-actor
Let us supposeequation (1.25) is vrritten in difierential form
dy + fu(t)y - q(c)l d,r :0.

(1.26)

Linear equa,tionshave a property that a function p,(x) canalwaysbe found


such that the multipie of (1.26)
' p(n)dy p(s)lp@ )a- q@)ld r: 0
+

(1 . 2 7 )

is an exact differential equation. By Theorem 1, the left side of equation


(L.27) will be an exact differential if

aa

*u@)

*u@)lp(")v

- q(r))

(1.28)

or
d,p,

fi: un@).
This is a separableequation from which we can determine p(r). We have
d1t'

l" lpl :

P(n)d'x
f

I p(r) dx
J

(1.29)

1.2 Tlpes of First Order ODEs and Methods of Solution

2T

so that
n@)d'a
tt(r) : sI

(1.30)

The function p,(u) definedin (1.30)is calledan integrating ractor for


the linear equation.
Method of Solution
To solve a linear first order difierential equation first write the equation
into generalform, that is,
dy.,\

#+p(")y:

q(n).

Then multiply the entire equation by the integrating factor p(r). The left
side of :

: r.t(r)q(u),
p@)H + p@)1-t(n)u
is the derivative of the product of the integrating factor and the dependent
variable:
Write the equation in the form
)

q@)p(")
frlu@)u{")l:
Integrate both sidesof equation (1.31)

a@)p(r):

Iq@)pr(r)dr

a@): h(l
Example 3O Solue
"#

dn+c)
r@)q(n)

- 2g : LLnl

Solution
This equationcanbe written in the linearequationform:
d v2

- -U:1012
dfrr
-

(1.31)

First Order Differential


,
where p (x) : -a and I (r) : 10a2.
Find the integrating factor p,(u)z

p(u):

enp{f-?-ar]

: ;ztl4'
:

:#

"-2lna

"lna-z

Write the equation in the form


d

:
frtu@)u@\ q(n)p(r)
: Lorz\:1s
#(,",)
integrate both sidesof the equation, we obtain

4
:L}r+C
frz
or

u :L 0 r3 *C rz

Example 3L Find the geneml solution of

g '+s)!**ry-0.
Solution
Step 1: Rewrite the equation in the followine form,

ar* (

,,

d r ' \ x z + s ) a :s'
I
t
\
p(r):
where

ili

q(r) : g.
a^nd

Equation

1.2 lypes of First Order ODEs and Methods of Solution


Step 2: Find the integrating factor, that is,
et fr'*

: (s2+ g )L / 2
"Lhlo2+el
Step 3: Rewrite the equation in form of
p( r ) :

q(u)t'(,),
frlu@)a@)l:
from which it follows

:0.
#U" +e)L/zy(x)J
Step 4: Integrate both sideswith respectto r, we get,
(r2 +911/2, - Q
:

c(rz+9)- t / z

Example 32 Find the generalsolutiond ,#+

(1 + n)y : er.

Solution
Step 1: Rewrite the equation

+ I+" - y :{ ,
+
df inf r
where p(x) :

| * t
T

and,q(Q : t.

Step 2: Find the integrating factor, that is,


P(x) :

d'r -

ulna*x :

"!

eln'e' : frev

Step 3: Rewrite the equation in form of


n

q@)pr,
fr[r@)a{r)]:

from which it follows

ftlr",u{dl:"rh

Step 4: Integrate both sideswith respectto t, we get,

(r"')y

(ru')A :

l ez'd*

J
oZs

U-+Cd6

Y :z * 'r

L ,* 9"-'
*

29

First Order Differential

Equation

Quiz 1.5
1. Find the generalsolution of
dy *et.
z_A
A n s : Y : e ' (t * C)
2. Find the general solution of the following equation
d,u
cosnfi*ysino= 1 .
A n s : g : c o s s (t a n r+ C)
3. Solve the following equation

@ +t)fl*u:tnlsl
subject to initial condition g(1) : 10
An s : g (1 * r):

s ln t -n * 2 L

1.2 Types of First Order ODEs and Methods of Solution


t.2.5

The Bernoulli

31

Equation

The differential equation


d,u

: y"q(r),
fr + r@)a

(1.32)

where n is any real number, is called Bernoulli's equation. For zr I 0 and


n * L, the substitution z : yr-n lead to the linear equation
iq

(1- n )q( a) .

# *tt-n )p (n )z:

Example BB Solue
H.

( 1.33)

- naz.

*,

Solution
Comparingwith equation(1.32),we foundthat p(r) : lis, q(r) : n,
and n: 2. Thusye takethe substitutionI z : + lwhere
l ,l
dz

d,-r,

-)

*\a

# : - rL ,f"#=:-r '#
grves
1
-A:

,dz
-U'
**

o
uA"

a:- L( 1) : - " ,
-'
d n n \a l
d z1
d. i":

where p(n) : -l and q(n) : -u.


T
equation is
I t, dc "-

-*'

The integrating factor fou this linear

u-Inlal -

ulnlal-r :

Hence
d r - ti

--W
dn"

-Z i :

- fi

-:
fi

-I'

fr-L.

32

First Order Differential Equation

Integrating this latter form gives


z :

u-r :
y :

-r2+Cs

-a2+ c r
I

Wgn

Example 34 Find the general solutionof thefoilowing Bemoulli equation.


dy

fr-a-"*s2
Solution
Step !.: Rewrite the equation in the generalform, that is:
dy

fi-u- "'u2

(1.s4)

step 2: Identify n: In this problem n : 2. Then use substitution z :


yL-n : g-1, which then it follows,
d,z _ -L dy
y2 d,n'
dr
Step 3: Substitutethe equations(from Step 2) into (1.14):

-A "d
-az r-Y
d z1
dfra
--+:
dz

*,

= e'aDA'
:

- e'

-e'

(1.8b)

Note: After the substitution the Bernoulli equation will be reduced to a


linear equation.
step 4: solve the first order linea,rdifferentialequation (1.3b) by the linear
method as explained in the previous section. In this problem the general
solutionof (1.35)is

e'z:-!*".

2
step 5: Find the solution of the original Bernoulli's equationby substitute
back z = a-t, we get
e,
e2,

T:- , *"'

1.2 Types of First Order ODEs and Methods of Solution

33

Quiz 1.6
1. Find the generalsolution of the equation
dat

:'(t" - t)
A n s:$ :- t+ t+ C" - "

2. Solve the following difierential equation subject to indicated initiat


condition

*' # - 2 n y - k y a ,

s ( 1) :1
Ans:$: -$c-t + fc-o

First Order Differential Equation

1.3 Applications of First ODEs


First or<ier difierential equations are usually used in describing physical
phenomenasuch as cooling of bodies,current in a seriescircuit, velocity of
a falling body, radioactive decay population growth and so on. Here we
will look at a few examples.
1.3.1

Newtonts cooling law

Newton's law of cooling states that the rate at which the temperature ?(t)
changesin a cooling body is proportional to the difierencebetweenthe temperature in the body and the constant temperature 2] of the surrounding
medium. That is

# : -k(r-rs)

(1 . 3 6 )

where k is a constant of proportionality. we also assumeat time t : 0 the


temperaiure of the object is ? : ?s, that is, the initial condition of this
problem is given by

:2i.

(r.37)

"(0)

Equation (1.3s) and initial condition (1.37) will form an initial value problem which is known as Newton's cooling model.
Equation (1.36) can be written in the separableequationform:
dT

(T - T')

: -kdt

(1 . 3 8 )

Integrate both sidesofequation (1.38), gives

tnf?- T,l : -k t + C

(1.3e)

Take exponent for the both sidesof (1.39), gives

T -?|,

"-ktD,
Tr*De-kt.

where D:

ec

(1.40)

Determinethe integrationconstantD by considering


the initial condition
: ?b.
"(0)
Substitutet : 0 and T : To in equation(1.40),gives

1.3 Applications of First ODEs

T o :Tr * De - o :Tr+D

35

D : To - Ts.

(1.41)

Thereforethe solutionof ((1.36)and initial condition (1.92)is givenby


T :Ts + ("0 - Tr) u-ot.

(r.42)

The constant /c can be determinedif we know one value of ? at a given


f as shown by the following examples.
Example 35 A cup of coffeeaf temperature80oC is placeil in a room ol
temperatureLI,C. In 20 minutes, the temperatureof the coffeeis d,ecreas,ing
ta 50oC. How long wi,Ilit take to cool off from 80oC to.JToC. Then, how
long wi,il it take to cool off from 50oC to 30oC.
Solution
Use the abc.,'eNewton's cooiing law where the temperaturesare given by
(L.42). Flom the iuformation given, we know that ?i - 10, ?o : 80.
Therefore.
?:

10+ (S0- 10)s-rt : 10* T\e-ht

(1.48)

We also know that if t : 20, the temperature of the object is ? : 50.


Substitute these values in equa'uion(1.43) to get the value of k. Fbom
(1.43),we obtain
50 : 10+70e-k'20

::)

-20k:tnf,

ore -20k
ork

t
-1

l;;ur l ;,

/4\-

'r t./

T : L}-r.7gr-fit"(t)t

(1,.44)

Now, we will useequation (L44) to determinethe time t if the temperature


? is given. In this example,given that T :30,, t :?
Flom (1.44):

3o: Lo*7g.-+ot"(I)t

36

First Order Differential Equation

* i:"

-*r" (i )t

or

-fr'"
(+),:'(;)
? f : -201fi
Ar 44.??l::^

The time neededfor the temperature to


cool ofi from gOoc to BO"c in a
room of temperature 10"c is approximately
4b minutes. Then, the time
neededfor the temperature to decrease
E0oc to 30oc, is 4b - 20 : 2b min.
Think about the results of this example
and whether they are consistent
with your own experiencein your home
or other places.What would be the
efiectsof changing the value of k?. Next
we look at different apprications
of first order differential equations,which
is Circuit Model.
1.3.2

Circuits Model

Circuit - RC
In a seriescircuits (seeFigure 1.r.) containing

only a resistor and an induc_

Figure 1..1:An R-C SeriesCircuit


tor, Kirchoff's secondlaw states that
the sum of the voltage drop across

1.3 Applications of First ODEs

37

the inductor L(dl/dt) and the voltage drop acrossthe resistor IR is the
sarneas the impressedvoltage E(t\ on the circuit.
Thus the differential equation of an electric circuits for the current .I(t)
is given by

r'* +R I:E (t),


dt

where Z and fi are co-qstantsknown as the inductance and the resistance


respectively
*dq-dI
d t'
The solution of the electric circuit's problem can be divided into 2 cases:
o Case (i) :

E(t) :.Es where .86 is a constant


iT

Thus the DE is L# + RI :.8s, where L, R, Es axeall constants.


For this casethe Off is a separableequation,
dI
r,---rL:;: dt
L

Integrate both side,

hl$ - 4rl
I

_TR
or

_!_La
-v

'l?-f4:-f,*o

Then,

Eo R;-;r-ke-tt
Lastly, we get the solution;

. -& +

k:eD.

Ir:f+e"-E'
r:E ;+B e-* t
Case(ii):

E:

Eosinuf or.E : E ncoswt

Consider
E:

Eosinut.

s8

First Order Differential Equation


The DE equation is
dr
"i*RI:E osin t rt
Therefore
dI , R,
E o -.
:
*
dt E t
Tsnut
This is a linear equationwith p(t) = !r, ilt) =
!sinwt
The integrating factor, p(t) is

:*o
p(t)

to4:
"r'

Therefore

uE'I: + t ettsinwtdt
LJ1.3.3

Motion

along a straight line

when we appl.yNewton's secondlaw of motion, that is, the rate of changeof


momentumis proportional to force acted on the body to an object moving
in a straight line, we obtain a differential equation of the form
d,\
: F,
4(mu)

(1.4b)

where F is external force , rn is the massof the body, u is the velocity of


the body with the samedirection F and t is the time.
Newton's second law (1.4b) can also be written in the following forms:

,:*#

: *#;

because
u:f,

: *nd,a d,s
at

: nO +
dr
one of the simplest examplesof linear motion is an object falling under
the influenceof gravity (seeFigure 1.2).

39

1.3 Applications of First ODEs

Let y(t) represent the height of the object above the surface of the
earth at time t, and let g denote the constant accelerationdue to gravity,
tsewton'ssecondlaw reducesto

fiwo- -ms.

T
v(t

(1.46)

-ms

Figure 1.2: Falling Body


Here, the mlnus sign on the right-hand side is presentbecausewe measure y(t) and u(t) as positive upward while the force of gravity acts downward, where we considerthe air resistanceis neglected.

du

or 4s : -gdt
-g
-dt:
Rearrangethe equation,we obi;ain
a--gt*C

(r.47)

If the initial condition is given, that is when t : 0, u : uo. Then substitute


thesevaluesinto equation (i.47), obtaining

+
C:uo
Therefore,
u - -st * ao
The position of the body at time t can be calculatedfrom (1,.+8):
dr
- - -gt*us

(1.48)

40

First Order Differential Equation

or
: (-gt * us)dt

(1.49)

Integrateboth sidesofequation (l.ag):


12

a (t):-g T +u st*D
W h e n f :0 ,s:0 ,9 D :0 .

r ( t ) : 4 + r ot.
2

Now, assumethat velocity o(t) is positive in the upward direction and


the drag force is proportional to velocity. If /c is the positive constant of
proportionality Newton,s secondlaw of motion leadsto
d,
\
*lmu)--rng-m k u

+
du
- - -9 - Ica

(i.b0)

Rewrite (1.50) in the form of separableequation:


du
4'

ka:

dt

(1 ' 5 1 )

Integraie both sidesof equation (1.b1), u,eobtain


-1

nl nl g* k ul : t + C
where C is the arbitrary constant.
up to this point we have related our discussionof difierential equations to mathematical models of cooling of a body, circuit and a falling
body. The derivation of these rnodels are not shown in this chapter, but
we concentrateon identifiiing which models should be used for a certain
application and how to get the sclution of the problem.

1.3 Applications of First ODEs

4L

Quiz L.7
L. The temperature of a dead body when it was found at 3 o'clock in
the morning is 85oF. The surroundingtemperature at that time was
68".F. After two hours, the temperature of the dead body decreased
bo 74"F. Determine the time of murdered.
Ans: L.52am
2. Supposethat the differential equation of an electric circuits for the
current /(t) ofa batter y is given by

t* +R I:E (t),
dt

where -L and R are cons'uantlknown as the inductance and the resistance respectivelyand g :
Let tire resistanceanC the inductance
A.
be 12O and 4H respectively.If the battery gives a constant voltage
of 60I/ and the switch of the circuit is closedwhen t : 0, that is the
current starts with /(0) :0, Find
(a) I(t)
(b) the current after Ls
(c) the limiting value of the current.

A n s : . I : t --lu -3 i
Ans: 0.99
Ans: 1

42

First Order Differential Equation

Chapter 2

Second Order Differential


Equations

Objectives
On completingthis chapter,the studentsshould
be able to:
determinethe characteristicequation of the homogenousequation
and solveaccordingly
solve homogenousequation with initial or boundary conditions
identify the appropriate generalform of the forcing term for the
nonhomogenousequations
solve nonhomogenousequationsusing the method of undetermined
coefficients
solvenonhomogenousequationsusing the method of variation of
pararneters
apply appropriate techniquesfor problems in rr.echanicalan<i
electrical circuits systems

Tt'e begin the discussionof secondorder differential equations as we did


with first order equations. However,we shall confineour attention to linear
differential equations.

44

2.L

Second Order Differential Equations

Second Order Linear Differential Equations

A linea.rsecondorder difierential equationswith constantcoefficientstakes


the form of:

"ffi*rH*cu:f @).

( 2.1)

o, b, and c are constants. Such an equation (2.1) is known as homogenous


equation if /(r) : 0 for all s. Otherwise, it is called a nonhomogenous
equation. /(o) is refered to the forcing term.
The next two concepts which are linearly independenceand superposition
principle are basic to the study of finding a solution to a linear differential
equations.

Definition 7 A set of tunction y1(n) end y2@) is said to be linearly dependent on an i,nieraal I if there en'i,stconstant iq and k2(not all zero),
such that
ks{x) * k2y2(r):9.

Therefore,if we assumeh * 0, then


k2
/\
_frur@).
yt(n):
That is, if two functions are linearly dependentthen one is simply a ccnstant multiple of the other.

Definition 8 A set of tuncti,on y1(x) and,g2@) is saiil to be linearly indepeir,denton an i,nteraalI if it is not linearly depenCenton the interual.
We can concludethat two functions a,relinearly independentwhen neither
is a constant multiple of the other on an interval. Or, in other word, U{r)
and,y2(r) is said to be linea,rlyindependentif kr: kz - 0.

l-!. Second Order Linear Differential Equations

45

rrrnple 36 Determinewhetherthe functions cosr and 2cosr are li,nw*g dependentor li,nearlyindependenton arbitrary interval.
Sotution
Im'eBeedto checkif there exist constant &1 and ,k2such that
rklcosr * k2(2cosr): 6.

(2.2)

!f,sw. from (2.2), we can write the equation in the form


cosz: -2^ kz
ktcosr.
Theefore, the given functions are linea,rlydependenton any interval since
mecan choosekr: -2 and kz : 1.

F:rarnple 37 Determine whether lhe functions 4r3 arui -2n3 are linearly
r,ryndent or linearly i,nd,epend,ent
on arbitrary interual.
Solution
fu(4r3) + k2(-2x3) : g or ,t : ?.!' *t.'
4lq"
Therefore,the given functions are linearly dependenton any interval since
we ca,nchoosekt: L and ,t2 : 2r so that it satisfy the equation.
Ilere, we supposethat

Example 38 Show that the functions etand e2t are linearly inilepend,ent
cln any interual.
Solution
kt@t)* k2(e2t): g or

: -9"t'.
"'
K1
This equation only has trivial solution which is k1 - kz : 0.
Therefore,the given functions are linea^rlyindependenton any interval.
flere, we supposethat

46

Second Order Differential

Equations

Theorem 2 (superposition principre) Letgl(r)


andg2(x) besolutions
of the homogeneous
l,inearsecondoriler difierential equati,onon an interaal
I. Then the linear combination
A :tcgt(a)+kzyz@)
uhere k1 and k2 a,rearbitrary anstants, is alsoa solution
on the interaal.

Example 39 The functions gt(r) : es, and y2(x) :


solutions of the homogeneous
secondord,ereguation

s-Bn bothare

at' - gu :0.
By the superpositionprinciple, the linear combination
y:

creS'*c2e-3'

is also a solution of the equati,on.

2.2

Solution of Homogenous Equation

We would like to find the solution for

"#,*u#*q:0.

(2.3)

In order to solve the equation, it is well known that


the best way is by
assumingthe solution in the form of
?): e^"

(2.4)

wherern is an arbitrary constant. since (2.4) is a solution,


it must satisfy
(2.3). So, we difierentiate (2.4) tc get yt andy,, giving,
an1,2e*' * bmem, * cema : o

or
e^t (amz + Arn + c) : 0.

47

ar Sotrution of Homogenous Equation


for y(o) to be a solution of the equation, we need
T.trnnus

(2.5)

a m2+A m+ c:0

or auxiliary
d'r' I 0. Equation (2.5) is called the characteristic
cqmation. Its significance lies in the fact that if rn is a root of the polynomunl {2-5), then gr: e*'isa solutionof the difierential equation (2.3).
$tu

Since (2.5) is a quadratic equation with real coefficients, the roots can
hne'{rqsffied into 3 cases which are:
:) two distinct real roots if.bz > Aac,
ri} repeated real roots if * :4ac,
nil} conplex conjugate roots if b2 < 4ac.
Seref,ore, we can conclude that the solution of (2.3) basically depends on
afoeroots of the characteristic equation. Using the concept of linearly indeewdetrt and superposition principle, we can now write the general solution
1q1i
the linear second order differential equations with constant coefficients
w rn Thble 2.1.
Table 2.1: Roots of characteristic equation and its general solution

Generalsolution

Roots

Case
1

two real roots: mt * mz

U :Ae mtx* B emzt

real double toot: m1:62

A:

complex conjugate roots:

A :eas(A c o s p r* B s in P x )

m:at0i

(A*

Bx)emn

Second Order Differential

Equations

Example 4O Find the generalsolution of A" + 5A' * 69 : 6.


Solution
The characteristicequation for this diferential equation is
m2+5**6:0.
Solving this quadratic equation for m will give m : -2 and rn : -3. Be'
causeof the two distinct real roots; we concludethat the generalsolution
is gr: Ae-z' * Be-g'.

Example 4L Find the general solutionof A" +9y - g.


Solution
Characteristicequation: m2 +9 : 0
Therefore,the generalsolution is

m : !3i.

a:A cosS x1.B sin 3 r.

Example 42 Find the general solutiono! y" - 2y'* g : 0.


Solution
Characteristicequation: m2 -2m* 1 : 0
Flom Table 2.1, the generalsolution is
y : (A* Bx)e'

m:

L.

of Homogenous Equation
43 Find the solution of the initial aalueproblem y" -4y'+4y : o
: 3 an il U '( 0):1.

m2-4m*4:0.
m:2r2,
y: (A+ Ba)ez'

solution:
the value for A and B:
Em the first initial condition

9(0 ):3 '

A :3 .

now gr: (3 + Ba\e2'.


we can substitute the second initial cond.ition, we must find the
which is
BeZ' +2(3 + Bn)e2a
-

{(r):
the secondcondition,
the answer is

y'(0) = L

A :(g-5a)e2'.

=)

.8+6 : l.

B : -5.

Second Order Differential Equations


Quiz 2.L
L. Find the generalsolution for
(a) Zy,,+ hyt + lig : g.
Ans: y(o) : rts-|r * Be-a.
( o) a " + 4yt+5y:o.
Ans: y(r) : e-2'(Acosz * Bsinr).
(") g" - 6y :0.
Ans: y(o) : as{6a + Be-{6'.
(d) 9y', - B}y' * 25y : g.
Ans: y(r) : (A+ Bx)e|'.
(e) LGyt'- n2g :0.
Ans: g(r) : Aei' + Be-T'.
2. Solve the following equai;ions:

(u) y" * g :0,

y(0): g, V(i): -g.


Ans:y( r ) :3cosr - Ssinc

(b) g" * g' - r2y - 0,

y(2) :2,

y'(2): g.

Ans:s(c) :
(") 4y" - 4y' - 3y : 0,

|"*p3('-4

*f,exp-a(o-z;

ael) : e, Ut(-2) : -;.


Ans:g( o) :s- l' .

2.3 Solution of Nonhomogenous Equations

2.3

5L

Solution of Nonhomogenous Equations


.flr,

;qr

Thesolutionot affi * b# * cu: /(c) is givenby


a@)-- sn@)+ao@)
where y{x) is the solution of the correspondinghomogenousequation and
yo@) is the particular integral or complimentaryfunction. In this chapter,
we will learn two methods of finding the parrticular integral yo(o). They
are
(1) The method of undeterminedcoefficients
(2) The method of va,riationof parameters
Note:

i;n@), homogenous solution, yp(s): particular integral

2.3.1

Method

of Undetermined

Coefficieni;s

This method is used to solvelinear secondorder differential equationswith


constant coefficientswhere /(s) is in one of the following forms:
(u) /(") is a polynomial with degreen
(b) /(") - Deo',, D and o are constanis
(c) /(r):

CcosBt or l@):CsinBx,

C and p are constants

(d) Sumsor products of the abovefunctions (a -

")
yo@)
integral,
is found basedon
particular
of
the
form
the
method,
this
In
the form of the forcing term. Pleaserefer Table 2.2.
Rules for the method of Undetermined

Ccefficients:

(1) Modification Rule If a term in the particular integral appearsin


the homogeneoussolution, it is necessaryto multiply by a sufficiently large power of r in order to make two solutions linearly
independent.
(2) Sum Rule If the forcing term, /(r) is a sum of severalfunctions,
then chooseyp(x) as the sum of the correspondingfunctions.

52

Second Order Differential

Equations

Table 2.2: The correspondingtrial particular integral,


yo(z)

t@)

(")

: anfin * an-1xn-L + ... a(D


+
* ao

bnsn *bo-Lfin-L +... + ha *bo


Ged'

Ccospxor CsinBa
Pn(a)eaa

P"(s) cospr or Pn(x) sinBa


Deaa cospn ot Deo, sinpa
Po(r)eaa cosBx or pn(c)ea, sinpr

pcospr * qsinpa
(bnxtu* bo-Lan-L+ ... + Ur, + boh*
(anao *an-1on-t +... + otz*aa)cosBr

*(b"r" t bo-Lon-r+ ... + hr * fo) sinBo


eor(pcospn * qsinpa)
(anrn * ao-1xn-L+ ... + otz * q)e*

tpg

Example 44 Determine the form of the particurar


i,ntegrurto the difier_
ential equationy,, - 4At* 4y - 6ea.
Solution
The corresponding homogenousequation is
y"-4yt*4y:g
Therefore,the characteristicequation : rn2 - 4m 4:0
*
+
Here, the homogenoussolution is

m:2,2

yn(r): (A+ Bn\e2,.


To determinethe particular integral,
ap@),we first rook at the forcing term
given on the right hand sicieof the differential
equation. We can see that
:6e3.
Flom
Tb,ble
!(s)
2.2, thepa,rticurarintegral that we must chooseis
VP(x)= Pu' '
Becauseof no terms in 96(o) appeaxsin yo(x), we
don,t have to multiply
it by c. In conclusion,the form of the particular integral
to the differential
equation is
Vp(x) : P"'.

cos0i

* bn-1rn-t + ...+ ht * h,)"* sinBr

53

2.3 Solution of Nonhomogenous Equations

Example 46 Determi,nethe forrn of particular i,ntegralto the ilifferential


equati,on
a" + A' - l\a :2e3*.
Solution
g
Homogenousequation: Y" + a' - t2Y :
+
equation:m2 +m- L2:0
Characteristic
Homogenoussolution: yn@): Ae\'+ Be-4'

m:3,-4

Particular integral:
Given l@):2e3'Ftom
chooseis

Tbble 2.2, the particular integral that we must


uP@): Pe3''

need to
Becauseof there is a term in 96(r) that appears in 3rr(c), we
multiply the solution by c. Now, there are no more similar terms. Hence,in
equation
conclusion,the form of the particula,r integral to the differential
is
gP(r): tPes''

ili'fferenti'al
Exarnple 46 Determi,nethe lorrn o! parti'cularzntegralto the
equationa" * g : x2 sinSx * es
Soluiion
Homogenousequation: g" + A :0
m:
+
Characteristicequation: m2 + I :0
Homogenoussolution: yn@) : 'Acosc * B sinr

Ii

Particular integral:
Using the
Given f (r) : s2sinSc * e' which is the sum of two functions.
sum rule,
gp(x):ae{s)*A ,pz @).

54

Second Order Differential Equations

For yr1(r), considerthe first forcing term which


is s2sin3o. Fyom Tb,ble
2.2,
ye{s) - (D*' * Ex * .F cosBa + (Gu2 Hn*
)
*
.I) sin Sc
Becauseof no terms in y1r(r) appearsin yo{x), we
don,t have to multiply
it by c.
Fot yo2(c), considerthe secondforcing term which
is eo. Here,
Ypz(a): Jea'
Becauseof no terms in g{x) appeirs in yo2(x),
we don,t have to multiply
it by r.
In couclusion, the form of the particula.r integral
to the difierential equation
is
ap@): (D*2 + Eu * F) cosJn * (Grz * Hr*

-t)sinBr * Je,.

Example 47 Determine the


form oJparti,cular integral to the dffirential
equationy" +U - ,2u5r * Bcosa
Solution
Hom.cgenous
solution:
Flom the above exampre, we know that the homogenous
solution is
an(x): A coss*B sin r
Particular integral:
Given i@) : r2e5' + Bcoso which is the sum of two
functions. using the
sum rule,
Up(x):ye{r)+yez(a ).
For yr1(r), considerthe first forcing term which
is x2e1r. Flom Table 2.2,
ye{r):(Drz*E x*F)"5,

2.3 Solution of Nonhomogenous Equations

OD

Becauseof no terms in y6(r) appearsin ypr(n), we don't have to multiply


it by o.
For ypz(r), considerthe secondforcing term which is cosr. Here,
Uez@ ):Gcosr*I/sin z .
Becauseof there is a term it y6@) that appears in gez(s), we need to
multiply tbieypz(n) by r. Now,
Yez@): t(G cosn* f/sinr),
where there are no more simila,r terms.
In conclusion, the form of the particula,r integral to the differential equation
is
yp($ : (Drz + Er * F)uu'* r(Gcosr * f/sinr).

The steps in fiirding the solutionz g(r) : An(x) + Ap@)


1. Find the homogeneoussolution, yh@).
2. Determine the correct generaiform of particula,r integral go(r) hom
Tabte 2.2.
3. gp(r) in step 2 must satisfiesthe given ODE. Therefore, substitute
tbe go@\ and its derivativesinto the given ODE in order to determine
the constant in yo@).
4. The generalsolution is y(z) : Ah(n)+ Up@).

Example 48 Firul the general soluti,onfo, A" i A' - 2A : 5e3'.


Solution
Homogeneoussolution:
m2+m-2:o
(m+z)(m - 1):0 +
A e'* B e-2"
"'Y n@\:

m: 1 , -2

56

Second Order Differential Equations

Particular integral:
Given I (r) : 5e3',

:. ap@) : pssa (no terms in y6 appears in yo)


YL@):3De3'
a'i@):9De3"

substituting yr(a) and its derivative into the difierential equation gives
gDes' + 3De3' - 2De3' : 5e3,.
Equatethe coefficients:

L}D :5

D:

:.ap@): i"t,
In conclusion,the generalsolution is
Y (r):

A e'+ B e-2'*t

"u'.

Example 49 Find the general soluti,onfo, y,, + 4yt + 4: !ss-2x


Solution
Homogeneoussolution:
m2+4m*4:0
(m+z)(m*2):ge
m : -2 , -2
... yn(") : (,4 + Bu)s-za
Particular int:gral:
Givenf(r):3re-2',
We choose Ap@): xz(Dr+E1e-z'

Next step is find its derivatives.


Ae@): P*3u-2a1B*2"-2a

(multiply by r,,so that no terms


in y6 appearsin gr)

2.3 Solution of Nonhomogenous Equations

67

:. {o@) : 3Ds2e-2' - 2Ds3e-za I Zune-2a - 2gr2


"-2x
g,i@):6Dre-2' - !2Ds2e-2',+ 4Dr3e-2',+ZEe-z',
_ggse-z, q 46r2"-2a
Substituting ye@) ard its derivative into the difierential equation and
simplify it will gives
6Dre-2' + 2Ee-2') : yxe-b .
Equatethe coefficients: -2D:3

=+

O:

w hi l e.E:0.

:'aP@)- L""-"
In conclusion,the generalsolution is

a@): (A-rBa)e-z'+!;"-zu.

-t

Second Order Differential Equations


Quiz 2.2
1. Choosean appropriate form for a particular integral of
(u) a" * 4y - ,2u3n
Ans:yo(r) - (Arz * Bs + C)esa
(b) u" * 4y - 0e2'coss
Ans: yo(c) : (An + B)ezacosc * (Dr + E)e2, sinr
(") y" * 4y - 2s2 +

+ ess
'sin2x
Ans:gru(r): Ar2 * Br + C +c(Dcos 2x * E sin1u)+ FeB,

(d) y" *4a - x2 cos2s


Ans:yo(o): r(As2 * Bx * C)(Dcos 2r * Esin2r)
2. Find the generalsolution for:
( u ) y" * a ' -2a:2x2

- 4a
Ans:Y(c) : 1s-2x * Bex - n2 +, -,

(b) y" * y' - 2y : eos2a


Ans: y(a) : /s-2a * Be' -

*r 2a +

*sin2c

&) a"-a'- 2 u : e - ' + a


A n sy: ( r ): Ae 2 ' *Be - ' - !,*u - ' - ;- i
( d ) y" * 2 y' -8y:

re-,
L=r"-*
Ans: g(c) : Ae2' + Be-4' 9 -"

2.3 Solution of Nonhomogenous Equations

3. Solvethe followingequations:
( ^ ) y " *y ' :2 *2 s*a 2 ,

y(o ):8 , g' ( o) : - 1


Ans:y(r): }e-a+ 5 + 2r +lrs

( b ) y "*4 a - "':::',,r,(0 ;o l

:
Ans:y(r)

y'( o) :g2
^2r + 1o - 1B_
cos
tUsin2c fr
)r2 , * i*

60

Second Order Differential

2.3.2

Method

of Variation

Equations

of Parameters

In this section, we describe another method of finding a particular integral, ye(r). The main advantage of variation of pararnetersis that the
method is more general and it can be applied to any equation with variable
coefficients.Besidesthat, the forcing term can be other than the elementary continousfunctions discussedin previous section. However,our main
concern is solving second order linea^rdifferential equation with constant
coefficients.
Consider

a g t'+b y'+" U :t@ ) .

(2.6)

As a startingpoint, we assumethat the generalsoluiionis


v : u(r)ut@)+ u(u)s2(r)
where 4(r)

(2.7)

and y2(a) are the corresponding homogeneous equation

ay "+W ' +

"u:0 .

(2.8)

Then we try to determine u(r) and u(r) so that the expressionin (2.7)
is a solution of the nonhomogeneousequation(2.6). Thus, we differentiate
(2.7) obtaining
y' : u'(x)yr(s) + u,(n)yl(x)* a'(r)y2(r) + a(r)yL@).

(2.s)

We now set the terms involving u'(o) and u'(c) in (2.9) equal to zero, that
is we require that
(2.10)
u'(n\y1(a)+ ut(a)y2(x): 0.
Then, from (2.9), we have
y' : u(x)g't(") + u(r)y'2@).
Further, by difierentiating again, we obtain
y't: u'(t)yi(") + u(x)yi@) + u'(x)yt2@)
+ u(n)yi@).
Now, we substitute for y, y' and y" in (2.6). After rearranging the terms
in the resulting eguation, we find that

: f @) (2.11)
u(ay'l+ W| + cgi + a("y'l+ Wi + caz)+ au'yl+ au'gt2

2.3 Solution of Nonhomogenous Equations

61

Each of the expressionin the bracketsin (2.11) is zero becauseboth y1 and


equation (2.g).Therefore,(2.11)reduces
a2 aresolutionsof the homogeneous
to
au'g'1+aa'yt2: f @)
or

lY)
u'a'r*u'rL:

(2.12)

Equations (2.10) and (2.12)form a systemof two linea,ralgebraicequations


for the derivatives ut and,o, of the unknown functions.
By solving the system(2.10) and (2.12)usingcramer,s RuIe, we obtain

-,, azt@)
,:--Mt

u, :

aW '

(2.13)

_ afl,z
_ ^,
yy : -lart ) sr@)l
_
with
wrUurz
^,, I,Z is calledas Wronskian.
",.^,,
a2a,L.
|u,,@) oLi,l:

By integrating (2.13), we find the desired functions u(r) and,u(r),


namely

u(x):-l#*
and

ub): Iu!@)
'^
1 ofr-o''
If the integral can be evaluatedin terms of elementaryfunctions, then we
substitute the results in (2.7), thereby obtaining the general solution of
(2.6).
Basedon the discussion,we can now summarizedthe stepsta,kenin the
method of variation of parameters.

62

Second Order Differential Equations

Theseare the stepstaken when using the method of variation of para,meters:


1. Find the homogeneorur
equation and write it in the form
yn(a):A u@ )+Bs 2 (n ).
2. Calculate 17, called the Wronskia^n,which is giv311fy

* :l':l't ':1:ll: ytaL


- aza',
lul@ aL@,r
3. Calculate

rtr:- [Y'lt"'l '


1 ffi-o*
and

)
"- IIa i@
':
ffio'' ^
4. The particular integral is

ye@) : uAt * uyz.

5. Write the generalsolution

y(x):u n @)+yp( x) .

Example 5O Find the gerieralsolution


fo, y,, - 2at + y :

#.

Solution
Cha^racteristic
equation: m2 - Zm* I : 0 =+
Homogenoussolution: ynb): Aet + Bre,
Particular integral:
Let Ar@) - ea

:1":
Then,
' ,
l"'

rn : l,I

and Uz(x): rer.

: s2a
,,1: e,(e,* ne.)- xea(e,)

-"u
ea+xett

1r'=- f xe', e0
J &GAda:

-rttt +z.2)+D

2.3 Solution of Nonhomogenous Equations

6S

e'

fe',

': Jb ;(fu)d ' u: t an- l

n+ E

For simplicity, we take D and .E to be zero.


:.yp(r) : (-i tn(l + n2))e' * (tarr-t r)reo
Generalsolution:

v (r):an@ )+s o@)


.'.y(s) :
:

Ae' * Bnea+(-| r"1r+ x2))e'* (tan-lr)re'


V-tLn(l+ *\le, + [B + tan-lnfne,

Example 5L Find the generalsolutionfo, U" | 4y :3 csc2o


Solution
m: *2i
Characteristicequation: m2 + 4: 0 +
Homogenoussolution: Un(r) : Acos2x * B sin2a
Particular integral:
Let Y{t) : cos2s and gz(t) : sin 2c.
Then.' w :l

cos2x
l-zsin2n

sin2t
|: z
2cos2rl

1t:- /1$"z "l fe c s c 2 r')


fd t : - r t * D
L
u:

/ (cos2r)(3csc2x)

J tT':

3, ,
ln ( : i- Z t ) + E

Finally, we will found that the generalsolution is

u@) :

AcosLr* B sin2r+

sin2z ln(sin2c) -

f,* "o"zr.

Second Order Differential

Equations

Quiz 2.3
Find the generalsolution for
(u ) y " - 2 y ' + z y:

glno

Ans: g(c) : (-o * D)e' coso * (ln lsinol * E\er sina


( b ) 4 u "- 4 y'+g :etlnzs

Ans:
s(a): (A+Bn)"t
+.-t*rurzr+fia\"t*(Irbzx-lgret
(" ) 4 g "*y : Z s e c(|)

An s:y(c):l .co sf +B si nf + csinf +ln( ccsil*"t


(d ')y" - 2 y ' + g : e tl n t
Ans:y(r) : (t -*^,

* A)et+ (rlnr - t + B)tut

2.4 Applications

2.4

65

Applications

one of the rbasonwhy secondorder linear equations with constant coefficients are worth studying is that they serve as mathematical models of
some important physical processes.Two importa,nt areas of application
a,rein the fields of mechanicaland electrical oscillations. For example,the
motion of a masson a vibrating spring and the flow of electric current in a
simple seriescircuit.

2.4.1

Mechanical vibrations

Mechanicalvibrationmodelis givenby

* # * r40 ,+kr:F( t )

(2.14)

where rn denotes ma,ss,b is a damping constant, /c is a spring consta,nt


and F(t) is an external force (damping force).The system is said to be
free vibration when F'(t) : 0. Oiherwise is said to have forced vibration.
Basically, there are four casesthat most discusseddepending on the value
of b and F(t).
(a) Damped free vibrations
In this $ystem,F(t) : 0 (freevibrations). From (z.L4),the differential
equation governingthe moiion of the mass is

ka:o
^# *uLor*

(b) Undamped free vibrations


In this system,F(t) :0 (free vibrations) and b : 0 (undamped).
Ftom (2.14), the systemfor undampedfree vibrations can be written
as:
&n
m*p*kn:0

66

Second Order Differential

Equations

(c) Undampedforced vibrations


In this system,b: 0 (unda-ped). Therefore,from (2.14),
d2r
rnTp+ko:F (t )

(d) Damped forced vibrations


This is given by (2.14),whichis
dzr

-dx + kr: F(t)


*W + bi

Foreverycases,
ifar: rE tnuo the fr,eQuenc]:u
y m,
1;.
Example 52 A uibratingstring with dampingc&nbe moilelled,
bg the ini,tial
valueproblemgiaen by

&u

. 133r* -^
0,
'
';t
-oda
dt2

g(0) : 1 , y ' (0 ) : 1

(a) find the equa+"ion


of motion of thi,silampeil uibrating string.
(b) fi,nd the frquency of the oscillation
(c) describethe motion as t *) oo .
Solution
(a) This is secondorder homogenousequationwith initial condition. Based
on our discussionbefore,below are the steps in finding the solution.
Characteristicequation:
Generalsolution:

rn2 * 6m * 13 : 0. =+ m: _J *.2i,.
y(t) : e-\t(Acos}t + Bsin2t)

Determine the value for .,4.and B:


Flom the first initial condition,
y( 0 ): 1 ,

L=

* B s in 0 )
"o(A cosO

=+

A:

l.

2.4 Applications

67

Thereforenow 9(t) : e-3t(cos2t+ B sin2t).


y'(t) : e-3t1-2sin 2t + zB cos2t)- Be-3t(cos2t + B sin2t).
Flom the secondcondition,
y,(0): t ... l" : e0(-2sin0* 28cos0)_ Beo(cos0*Bsin0)
:+ 2B -3:I
+ B :2.
y(t) : e-3t(cos2t * 2sin2t)

Finally,
(b)Flequen"y:

,.L
;o: ;

(c)As f -+ oo,

y + 0.

The amplitu,ie of oscillation dirninishes.

Example 53 Find the solution of an undampedsystemgoaernedby


& y ..
u.
*Tfr*ky:5sin5t

' , :
where

l-k
:,
|

and the i,nitial conditions-3i(0): 0, yr(0) : g.

Solution
This is secondorder nonhomogenousequation with initial condition.
Flom the given forcing term, we know that we can use the method of
undetermined coefficients. Based on our discussionbefore, below a^rethe
steps in finding the solution.
Instead ofusing m for the roots ofthe characteristicequation, here we use
the symbol ).
Characteristicequation: m\2* k : 0.
Homogeneous
solution:
Particular integral:

Given
f (t): ssinft,

=+

X: +tlTf, : *wi.
v rn

yh$) : Acoswt* Esina/

Second Order Differential

ae$) = D cos + n ri"f,t


it

Equations

(no termsin 916


appearsin gro)

: -|o sinlt+ n
aL$)
f, "oslt

: -+"
yi4)
- -+r siolt
"o"f,t
Substitutin9Ap@) a^ndits derivative into the differential equation gives

*e+o

-{-" rt
:
it\ +k(Dcos
"o"f,tf,t+ n "i,,|| tsinf,t.

Consider the coefficient for the term sin?t z

- **r*kE :5

Considerthe coefficientfor the term


(.2

-r n + D * k D : o
4

E::+
t

"o"|t

D:o

: zfi"irit
...,i/pe)
In conclusion,the generalsolution is

: Acosut*Bsina.rr
+zfi"i"|t.
sp(t)
Determine the value for.4 and B:
Ftom the first condition.

u(o):0,

A:0 .

Therefore
nowy(t): Bsinc..rf
+ frflsinft.
a'$): Bwcosut*
# "o"|t.

69

2.4 Applications
From the secondcondition,

g'(0) :0

= + n:#

.'.

0: Bw + #

Finallv.
u,
10
. . 20
.".,
A( t) : - * s ina/+ gk sintf

2.4.2

Electric

Circuit

Consider first the series.R.tC network. A voltage source .E(t) having the
polarity is connected in series with circuit elements having resistance R,
inductance.L, and capacitanceC. A current 1(t), assumedpositive flows in
the loop. In essence,Kirchhofr's voltagelaw assertsthat the voltageat each
point in the network is a well-defined single-valuedquantrty. Therefore, as
we make an orcursion around the loop, the sum of the voltage rises must
equal the sum of the voltage drops. If we proceed around the loop in a
clockwise mannel, the volta.gerise is the source voltage .E(t), while the
voltage drops a,rethe drops aclossthe three circuit elements.The voltage
drop acrossthe resistor is /(t)R, the drop acrossthe inductor is L(dIldt),
and the <irop acrossthe capacitor is (l/c)Q(t), where 8(t) representsthe
electric cha,rgeon the capacitor. An application of Kirchoff's voltage law
therefore leads to the equation

E(t):Rr+Llldt+LA|j).

(2.15)

To obtain a differential eo;ration for a singie dependent variable, we use


the fact that electric current is the rate of change of electric charge with
respect to time,
t/,\
t\t) :

dQ

E'

One approach is to rewrite equation (2.15) as a secondorder differential


equation for the electric charge,obtaining
.*Q.ndQ,1
Lffi
* R; * eQ: E(t)
This equation, supplementedby initial conditions specifying the charge
t6, can be solvedfor
Q(to) and current Q'$d = /(t6) at some initial time

70

Second Order Differential Equations

the charge Q(t). Ditrerentiating this solutions yields the desired current,

r(t).
Elxample 54 The differential equationdescribingan RLC seri,escircuit is
giuen by

Lq"+ nd +

E(t)

In the aboae,L is the inductance,in"l:


henries, R is the resistancein ohrns,
c is the capacitancein famds and E(t) i,s the electromoti,ue
lorce in aolts.
Note that q(t) is the charge in coulombson the capacitor at time t, and
i(t) : ffi l" tn" cwrent in amperesat time t.
(a) Supposethat L : 0.5 henries, R : 8 ohms, C : 0.02i farads, and
no eluiromoti,ueforce is opplid. Finil the chargefunction q(t) rf the
initial' cur-rentis 2 amperesanil the chargeon the capacitoris i,nitialty
zeno.
(b) Then use the results obtained to show that the cument i(t\ can be
written as i(t) :2,,1$,"-at cos(4t- a) wherea : tan-r 2.
Hint: usethe id,entity
Pcos(z -y):
Pcosscosy*psinrsiny

2.4 Applications
Quiz 2.4
1. A vibrating string with damping can be modelledby the equation

)^.
*#&* * u# * kY- o,

: a,s'(o)
: B
Y(o)

(u) If rn:l0kg, b:60 kg/s, k: 250kg1"2, a: 0.2 arLdP : -0,2,fiud


the solution of thc abovesystem.
Ans: y(t) : e-st (0.2cos4t) * 0.1sin 4t
(b) Given that when b: 0, the motion is oscillatory with frequency
5^.
Discusswhat efiects damping has on the frequency of oscilfr.
lation and the solution as t -+ oo
2. Find the solution of an unda,mpedsystem governed by

&u

rn# * kY:5sinc.rf
where, :

tk
| *,

*U the initial conditionsg(0) : 0, y'(0) : g.

3. Consider the vibration of a mass-springsystem given by the initial


value problem
rln

&n
nxffi * b; * kr : 0,

c(0): 0,r'(0): 1

where m,b,k,are nonnegativeconstantand b2 14mk.


solution to the problem is given bY

Showthat a

n(t):#"-*'"r'ff,
4. Sol.'.ethe initial-value problem governingthe chargeq(t) in a,nelectrical circuit consistingof inductance.L and capacitor C,
q1-* fiu: Essinlt
-'&orz
whereEorLrC,l arc constantsandf

q(0): 0,g'(0): g,
I fi

:,4cos
q(r)
Ans:
hr*Bsin hr+

ffi-rr"t't,

72

Second Order Differential Equations

Chapter 3

The Method of Laplace


Tlansforms

Objectives
On completing this chapter,the studentsshould
be able to:
L: determine the Laplacetransform of standard functions and
special functions (step, delta, and periodic functions).
2: use the properties of Laplacetransform and inverseLaplace
3:
4:

transform appropriatelY.
determine the inverseLaplacetransform by way of Laplacetable,
partial fractions, completing of squares,and convolution theorem.
use the method of Laplace transform.qto solve initial and boundar;'
value problems for linear, first and secondorder ODEs with constant
coefficients.

The ultimate aim of this chapter is to demonstrate the techniqueof the


Laplace transforms to solvelinear, constantcoefficientsODEs with emphasison the second-orderequations.The techniqueis particularly usefulwhen
the forcing term in the ODE involves other than continuous elementary
functions discussedin chapter 2. In other words, the previoustwo methods

74

of Laplace f1'ansforms

The Method

(the method of undetermined coefficientsand va.riiationof pa,rameters)fail


when the forcing term involves discontinuous functions or having impulse
force.
We will describesomeimportant tools beforeberngable to usethe technique
successfully. Fbr ease and convenience,we will apply the technique for
problems having continuous forcing terms where the two previous methods
can still be utilized. Then we introduce three special functions which will
eventualiy be included in our forcing functions describing the ODEs.

3.L

The Laplace Transform

The Laplace transform is an integral transform which has the property of


translating certain complicated operations into simple algebraic operations
in the image (Laplace) space. It is a solution techniquewhich transforms
difie.rential equations in the time domain into algebraic equations in the
s-domain.
Definition I The Laploce transform
f (t) i,silefined by

function

t-(s) of the tunction

f g)e-'td,t
Io*

F(s): L{f (t)} :

for all numberss for whi,chthe improperintegralconaerges.


Example 65 Fi,rulF(s) i,! f (t) : t.
Solution

r(") :

L {l (t)}: L{ L}

at
fo*"-*
:

lim IN

N-+oo./g

"-"'dt

lim f{.|"-r,

N-+o l-sJo

s) 0 .

3.1 The Laplace TYansform

75

Example 56 Find F(s) if /(t) : eotfor a ) 0.


Solution

F(")
: Jo'i{,?),:,::,:,
[*
Jo "b_4'41
r
rt
1N
lim |
lim I
N-+oo./g "@-s)t41: N+oolo-s

1N

,(a-s)t I
Jo

*"!x"["r"-"liv-t1
f

i f s<a ,

f-

,-"1-t

i f s > o.

We see that the improper integral for F(s) does not convergeif s ( a.
Therefore,
F(") : L{f(r)} :
for s > a.
*

Note that for the Laplacetransform of cosineand sine functions, it is much


easier to obtain the formula using result from example 56. We postpone
this to the next section.
Example 57 Find F(s) if l(t) : t.
Solution

F'(s) :

f6

L{ty : I
J

1N

,"-" dr: _lim I


N_+oo

;16

1
:
--,,Jo
"-ttl
-"
"r'
',TLL
,.

lt"-t'

Example 58 Determine the Laplacetransform of

( z , o < r< b
/ ( r ): 1 0 , 5 < r< 1 0
[ "nt, 10< r.

te-'tdt

The Method of Laplace Tlansforms


Solution

F(s) :

yp'1.-*a,

[*

JO

75

r10

l " r" -" rd t+ ['" o e- ,tdt* [*


Jo
Js
Jro "4t"- st41
:

2 fu
JO

"-"rdt

+ tim [*
ly-+oo /1g

"-a-c,rry,

,ftlu*
' u*f ";"-'1il'
L -" l o N _+m
L_( " _ a ) J r o

-2e-5t *?
8a

f
u-r(s-a) I
*liml-|
"-{"-a)r +iv-+ool s-4
r-4
j
-

Ze-i"
"

2
*;+A "-L0@-a) f o r s > 4 '

3.1 The Laplace Tlansform

77

Quiz 3.1
.- 2

. 1..Find L{t,-z}.

Ans: -

-.

2. Find the t aplacetransformof

Io ir

/(,):tl
il

t<1
L<t< 2
t>2.
e-s * e-2t
,
Ans: -.
s

3. Find L{lt - 4l}.

. -4 p . ----_ -.

2e-4" *4s -I

The Method of Laplace Transforms

78

So far, we have learned how to find the Laplace transform function using
definition. It is quite cumbersometo compute this every time. Therefore,
the Laplace transform of some elementary functions are given in a table
called the Laplace transform table. You can refer to this table unlessotherwise noted. Flom now on, we will refer to this table. We will state useful
theorems of which some are given without proof.
3.1.1

Linearity

of the transform

Theorem 3 Giuentunctions f anil g and constantsa and b, then

L{af + bgl,: aL{f} + bL{s}.

Here we illustrate the use of this theorem to determinethe formula for the
Laplace transfbrm of cosine and sine functions.
Example 69 DetermineL{cosbtl and f,{sinbt}:
Solution Recall that.
eibt: cosbt + isinbt.
If we set a : ib with i : GTrthen

from example56,

L{edny.:
*:ffi+ny
:

s*ib

frF:7ryFi+xlrT6t

Therefore,
L{d*}

.C{cosbt * zsin bt}

.C{cosbt],+ i0{sinttt}
sb

+xlz

+62'

Equating the real and imaginary parts respectively gives

b4:
,C{cos

=
and f,{sinbt}

#*.

3.1 The Laplace Tlansform


Exarnple 6O DetermineL{\eat - 6 + Scos2t.}
Solution
L{gtt - 6 + cos2t}

L{te4t} + 4{-6} * L{5 cos2t}

l${eatlr - 4{6} + \L{cos2t}


365s
s-4--=' , J a a '

Exarnple 6L.DetennineL{coszl * Ssinh2t}


1\

A. nsI :l 5( ;sr y a+ ; ) * T4

80

The Method

of Laplace Transforms

3.1.2 Tlanslation in s
Here is one important property, also referredto as the First shifting

theorem

Theorem 4 If L{f (t)} : f(s) exi,sts,then


L{""' f(t)I : F(s - a),

a ) 0.

Proof

L{e"tye)l :
:
:

lo* "-,r"oty1tyat
f*

J,

^F(s- a).

Example 62 DetermineEte"t sinbt),


Solution

e-("-")t
I (t)dt

a,b > 0.

We know
l{sinbt}:F(s):-3--.
s2+ b 2 '

Thus, by the translation property of F(s),


L{"o' sin bt} : F(s - a) : ,
?- , =.
'
(s - a)z .162'

3.1 The Laplace Transform

81

Quiz 3.2
L. Find L{et sin2t}.
2. Find L{t2 e3t}.
3. CalculateE{e-lt(t2 - 3t + 5)}.

a*'1":il*a'
,'r 2
ns:1"r 3) a'
.2 35
Ans:
GJ-f I".BP- ft'

The Method of Laplace Transforms

3.1.3

Derivative of the Laplace transform

Theorem 6 Let L{f (t)I: F(s). Then

E{t"f (t)}:(-\"#.
Proof

F(s): Eg@\:

lo*

e-'t1g)dt.

Then

#: * l,* e-'tvg)dt'
By the Leibniz's rule,
dF
ds

ft"-"'y1t1at
*
" -,,tfg) dt

r/(r)).

Thus

L{t lQ)} : (-L)


dsL!.
The general result followq by induction on n.
Example 63 Determine L{tsin2t}.
Solution

F(s): L{sin2t}:

Differentiating .F'(s)we obtain


dF

-4s

6:Wry'
Thus

im

4s

L{tsin2t}:- d"=:
Gfu.
Example 64 Detennine L{tcosSt}.
.

ens:

s2 -g

1;z",-gp

3.1 The Laplace Transform


Quiz 3.3 Determine
L. L{testsin2f}.
2. L{tsin22t}.

a(s-3)
^
Ans:
[G-przp'
1fl
s 2 -1 6 l
.
Ans:;Lp-lF+16)zJ

The Method of Laplace Tlansforms


3.1.4

Laplace transform

of derivatives

Theorem 6 Deriaatiae theorem Giuen o function f (t) with Laplace


transforrnL{l(t)} : F(s\, then
(i) L{f'(t)}:
(iil L{1" (t)} :

sF(s) - /(0)
"2tr(")

- s/(o) - /'(o)

(iiil L{f(n) (t)} : snF(s) - s"-1.f(0) - sn-2f'(0)


where/(0) dsthe initial aalueof f (t) at f :0.
Example 65 Find the Laplacetranslorm ol f (t):
Solution

/"-1(0)

cos2t.

Note: We can apply trigonometric identity for cos2t and solveaccordingly.


However,we give another alternative to illustrate the use of this theorem.
f' (t) : 2 cost(- sint) : - sin2t.
Using the Laplace transform of sin, we have
C{f'(t)}:

L{-sin2t} : -* V , .

By the derivative theoreno,


L{f'(t)}:

sF(s) - /(0) : sL{f (t)} - f (0),

so

sL{f(t)}- /(o): -h,.


And given that /(0) : cos20 : 1, we have

s}ifg)\-1--h
s2+2
2
sL {f(t)}:l-;;T4:-fr.
Thus
\r-)
\.tl\"t
-t{fi
Example

t'.*2,.
s(sZ+ 4)

66 Find the Laplace transform of l(t) =3

usi,ngthis theorem.

3.2 The Inverse Laplace Transform

85

Example 67 Finil F(s) if /(0):"J., and l,(0):2.

1. f"(t) - 2f'(t) : L.

Ans:F(s)

2. 2f't(t) +3f'(t) + l(t):

s r(s _ 2 ),

+ -L.

s_2.

egt.
I

2s*7
A
ns:F(d:,
,'
--'---L - \-/ (s- B)(2s
+ 1)(s+ r) - 12rl g1s1 r).

3.2

The fnverse Laplace Transform

Definition LO It L{l(t)}:
of f is giaen by

F(s), then the inaerse Laplace transform

c-l{r1s;y : f (t).
Moreover, for any F(s) and G(s) whoseinverseLaplace transform exist.
we have
t-L {ar g) + bc(s)} : aL-L{F (s)} + ac-l{c(")}.
Example 68 use the Laplacetable to find L-r{F(s)}

fn ,\(s ):

s 3-r4l B

for the function

Solution

+ + - = L- - s
t@: L - {lF-A+%-741

3 '(
,-' * sL-,1#*)
- 2-r{"=}
tpj

= 2 ' '(2 )
:

Zt2+lsinbt - cosh2t.
o

The Method of Laplace Transforms


Quiz 3.4 use the Laplace table to find ^c-1{.F(s)} for the folowing functions.

1 .r(s):h .
2.F(s):W
3..F( s) :* -

Ans:/(f) :
Ans:/(t): o

+rsint-,zt.

["ortX*r"*;]

eor,f; -f,sin2t.

3.2 The fnverse Laplace Ilansform


Next, we will look at important techniques when the inverse Laplace transform function can not be obtained directly from the Laplace table. We will
illustrate by way of examples.

3.2.1

Translation

property

for inverse Laplace transform

: f (t) and a is a constant,then

Theorem 7 Il L-r{F(")}

- o)} : e"tL-L{F(s\} : u"tl(t).

ft{r@

(
'
^-1
Example 69 Deter
-nu'ne

'

Solution

t6-ET )

2
L^ --,[(G : E E JI :

f t t- L { 3 }
[s" J

:
^-1
Example 7o Detet-rn,ne,

Solution
3s

GTT

3s

eatt2.
)

t1"J';u.1

:
:

3 (s+1 )-3

I;TTZ_
33

6aTs

Is+l)4-:

Then

3 \_.-tI
3 I
-r
: c-'tGTtrj
L-t
r("+Fi
tcw]

L -t{=3 t

e-,1-'{*} - e-,'-r{i}

_,!,"_,t_,{$}
: !ru_,r_,ti)
3o- t + 2

2""
:

- lo- t t l

2"

- r3).
Ue-t1tt2

The Method of Laplace Tlansforms


Quiz 3.5 Determine

L L-L{t"_,irz}

2L L{ff i }
3L-L{I;5,i

4 Lrt# )

Ans: e2t[cos2t* sin2t].

Ans:
e-4tPr-ZUl
Ans:
et
V*r'.rtt]
Ans:e-a' - hu].
[*u

3.2 The fnverse Laplace Ilansform

3.2.2

Completing the square

Examprerr Detennine
L-t
Solution

ih=r}

{"

,-'{F#m} =,-'{#iTd}
:

"_2tp_t{'_%}
: su_zt2_r{;uh}
Example
12 Determine
L-t
{#'si13t
Solution

,-'{#*h} =,-,{#.T6}

:,-'19fi*)
=
:

"_s,'_L{f*}
- rt"tt) .
"-3t (cosat |

89

The Method of Laplace Transforms


Quiz 3.6 Find the inverseLaplace transform of the given function.

L.L-Li"r*Trr)
2L-'{#}
s.L-L{F#"-*}

.e.*,
Ans:e2'
[.*tt*

|u-tsinBt.
]

r"tt] .

Ans:e-ztlL - ztl

3.2 The Inverse Laplace Transform


3.2.3

91

The method of partial fractions


'l

Example 73 Deterrnine L-L { }-

l.s(s+ 1)J'

Solution
I
B
_A *
's*
1
s(s*1)
s
Then
l.:.4(s+1)+B s.
Set s :0 and s : -1. Hence
s:0:
A:L
=+ B :-t.
s : - 1 : 1 :0- B
Now we have

rL!
'

1 I-'-r11-

i ;F + tl :'

Use the linearity property,

1 \

t;-"+11'

.-'{#} :,-'{i}-p,{#}
:

L-e-t.

Exampre14 DetermineL-L{#-r}
Solution

B
Cs * D
r
PGITII:; P+ffi

Then
r :A s(s2 +1 )+ B (sz+1 )+ ( cs *D) s2
reducesto
t:

(A * C )s3+ (B + D )s2+ As * B.

Equatingcoefficientsgives
s3: A+C:0
s 2: B + D : 0
s l : A: 0
s o : B: L .

g2

The Method of Laplace Transforms


Thus C : 0 and D : -l,so that
1

-1

;tlF+lt:p+s-t+1

and

: E-{+} - L'{#}
"-'/-l-=.I
t"r(rr+L)l
: ,_rrl:l

3.2 The fnverse Laplace Tlansform

93

Quiz 3.7 Find the .C-l{f(s)} for

L L'{G=fuo}
2 L-L{#rr}
3 L-L{#-r}

Ans:
!["r, - "-ur1.
a*'

cos2r].
| [1-

a*' -| * tb" +
"-r'1.

The Method

3.2.4

of Laplace Transforms

Convolution theorem

Theorem I If L-L {f(")} : f (t) and,L-r {G(")} : s(t), thenthe conaoIution.! tQ) ands(t\ is

:
- u)du.
I * s : t-L{r@yc(s)}
lo' ilu)nt,

Example75 Deterrnine
L-r{1;#r=rr}
Soiution: We can obtain the solution by the method of
(a) completing ihe squaxe
(b) partial fractions
(c) convolution theorem.
We want to use convolution theorem. Write
111

s+ 1' s - 2

Gm_ zI:
and let

I,I
:;=
r(s)-tr,
G(s)

Then

: L_L
r(t) : r_l1r1sy1
t#i
s( t ) :

: e_t

c - 1 { c 1 s ;:1L - L{ *} :O'

so that
f(")

g(t - u) :

e-u

"2(t-u)

- u2tu-2u.

95

3.2 The fnverse Laplace Transform


Hence we obtain

-' (
c-'tG
.1 -z)
=,) : Jo
[' 'l@)ne-u)d,u
+ r)(,
)
:

ft
Jo "-uuzt"-zupu

::k,
ft e-uudu
.,
"n Jo

- ,-r).

Erxample76 DetermineL-r{#}
Solution: Write

331

;(F+lt:;';41"
and let

:
F(s)
h,

c(")::.

Then

: L-t
f (t) = z-11r1s;y
t#)
s(t) :

: sinr

L-r{G(s)}: L-t {1} : t


IsJ

so that
f (") :

sinu

and

9(t - u) : 3.

Hence we obtain

L-r{_

r"1Fht]

ft

Jor@)rft-u)du

/t ,irrr.a d,u: -|cos u l[


Jo

3(1- cost).

The Method of Laplace Ttansforms


Quiz 3.8 Find the inverseLaplace transform of the given function.
t'1
- . s2 ( s2+ 1 ).

Ans: t - sinf

o3 "
- ' ( "2+ l) ( sz +4)'

Ans: cost - cos2t.

.1
o.

.
e 2 t-L -2 t
Ans: -7-'

ffi_ z).

, G(")
^'s2+L'

7t

Jo

n$ - u)sinudu.

3.3 Initial Value Problems

3.3

97

Initial Value Problems

Now we are ready to solve initial value problems using the method of
Laplace transforms.
Example 77 Solue
a" - 6a' * 9Y: t'"t',

Y(0\ :2,

g'(0) : 6.

Solution
L {y" - 6y' + gy} : L {tzest}
By the linearity property,
L{v"} -6L{y'}

+sL{yl:

E { t 2 " 3 r} .

s2v1r7-sy(o)
-a'itri)-6[sr(s)-y(0)]
+er(s) :

2.,=.

(s - 3 ; e '

Using the initial conditions

s ) :- ,
[s2 r( s ;-2s -6] -6[s r(s )-2]+
-t ' es.r (r"r

2
1"I$

and simplifuinggive

(s2- 6s+ 9)f(s) : 2s- 6.

Ffu

: 2(.s
(.*- 3)2v(s)
- 3).

ufo

22
Y (s):-*(;=j E
Thus

v(t) :

t-r 1v1s1i

:,1-{*}*1'-'{#}

2e't+
iru"t .

Example 78 Usethe Laplacetranslormsto solae


a ' +3 Y : 1 0si n f, Y (0):0.
Ans: e-3t- cost* 3sint.

The Method of Laplace Transforms


Quiz 3.9 Use the Laplace transforms to solve

r. a'+2a: sin4t, s(0): 1.


2. U" - 2 y'+2 y:9,
3. A" - 2 g t + 2y:

g(0):1,
e-t,

g(0) :0,

eor'

sinat.
$"-u |cos4t+fr

y'(0): 0 .

A n s : e t [ c o s t -s in t ] .

g'(0 ): t .
An s :

11

i " -' -

ie ' [ c o s t

- T s in t ] .

4. Showthat
L- ' {, ='=t'

*10,"'=}

t(" 2-2s +5)(s+L)J

B e t c o2s t * 4 e ts in 2 t- e -t .

Using this result, solve the fVP


A" - 2Y'* 5Y : -8"-',

A(0) : 2,

Y'(0) : tZ.

Ans: Setcos2t * 4et sin2t - e-t.

3.4 Boundary Value Problems

3.4

Boundary Value Problems

The method of Laplacetransformsapplicabledirectly to conditions at the


origin (t : 0) only. For the caseof one boundary condition not at the
origin, we have to let that condition to be a constant, say a. Later, we can
determine its value.
Example 79 Soluethe BVP

a" + a - cos2tt Y'(0): 0, a'(n 12): -1'


Solution
t {a " } + L {v}: L {cos2t}

- sy(o)
+ r(s) -- #
- s'(o)]
[s2r(s)
Letting g(0) : a and applying boundary condition give

's2f(s;- ds -01+ Y(s):

h.

Simplifying,

:
(s2+l)Y(s)

"t+fi

(r"+ l)Y (s): W


y(r) :

alternative?
Whatis another

ffi.

The partial fraction decompositionfor Y(s) is

os3+ (4o+ 1)s =


( s 2+ 1 )(s2
+4 )

D
As_*B
*C2^*
sz*4
sz+L

(o+| )s * 0 , - | s * o
: __A+L
r_fr4.

100

The Method of Laplace Tfansforms

Hence we obtain

v(t) :

^c*1{r(8)}

pL-,[(a.+ *)s * -i" \


I;E-V+
|

: 1o+
_
|)c_,{"*,} Ir-{"_}
:

1'*|lcosf-|coszr.

To find a, usecondition y,(n/2'): -1.

a'U) :

-(o+

1r.

i)sint+ isin2t

-1 : -(a+|l.t* l. o
Thus '
o*|:r

o:!.

Substituting a into y(t) give


v(t) :cosr -

coszt.

3.4 Boundary Value Problems


Quiz 3.10 Use the Laplacetransformsto solve
L. U" + 4A : \et ,

y(0) : 1, A'kr 12)- er/2

2 . g" - 2 yt+ 2y:

e-t,

g(0):0,

3. A'+9y : sin3f, Akl2\: O.

Ans: y(t) = et.

y'(1) : 0

Ans:y(t): -W.

LO2

The Method of Laplace Transforms

3.5 Special tr\rnctions


In this sectionwe will learn how to use the method of Laplace transforms
when the forcing term in the IVps or BVps involves unit step functions,
delta function, or periodic functions. Specificallg the emphasisis
on the
first two functions.
3.5.1

Unit Step (Heaviside) Function

Definition Lt The unit stepfunction H(t - a) is definedby

H (t-o):F

0'
I t,

0(t(a
t]a.

o In engineering we frequently encounter functions that


can be either
*onn of uof.'

o For o(ample, an external force acting on a mechanical


system or a
voltage impressed on a circuit can be turned offafter a period
of time.
o we can produce all sorts of effects by multiplyi*g
H(t - a).
Example

functions /(t) with

8O Write the function


(

fft):J

s i ntr
I t,

o( t < z r
t) r .

in terms of unit stepfunctions.


Solution
f( t)

(
sintr
{

I t,

olt1r

s i ot+J

t)r.
0'

I t-sinf,

s i n r* ( r- s i n'rr) 0,
I t,

sint * (t - sint)I/(t -

0Stczr
t lr.

01t 1r
t2 r .
").

103

3.5 Special Ftrnctions

In general we can usethe formula (for discrete functions having 3 intervals),


( gr,

s (t ) :

0(t(o

au-tu-b

lnr,

t>b
I ge'
h * @2- giil(t

- a) + (gs- sz)H(t - b).

Note that this formula can be extended to n indervals.


3.5.1.1

The Lapiace transform

Theorem I The Loplacetransform of H(t - a) with o 2 0 fs

L{H(t- o)}: "1".


3
Proof

L{H(t- a)} :
:

Io*"-"r(t
rim [*

N-rm /o

- o)a,:

"-,rdr:

I"*

e-"tdt

lim dl"

JV-+oo -S

:"-o" .
lo

Example 8L Detennine the Laplocetranstorm ot

( g.
/(r):{-t,

I z,

c <t< z
z St< s

t-- b.

Solution
Write /(t) in terms of unit step functions.

f g.

/(r) : { -t,
[ 7,

o<t<z

zZt<s
t>5

gt* bz - giH(t -2) + (gs- gz)H(t- b)


3 + (-1 - 3)rr(f- 2) +[7 - (-1)]II(' - 5)

= 3 -4 H (t-2 )+8 .H (t-5) .

LO4

The Method of Laplace Transforms

Hence,

L{l(t\}

E{s} - 4L{H(t- 2)} + 8L{H(t - 5)l

I
.e-2t
--4-r3-

^e-\t

88a

3.5.1.2

Translation in f

The following theorem is also known as the second shifting theorem


Theorem LO If .F'(s): L{l(t)}

ond a} 0, then

L{f (t - a)H(t - a)} : e-"'L{f (t)} : e-"'p'(s).


Conuersely,
L-L {e-as F'(")} : f (t - a)H (t - a).

Example 82 EualuateL{tH(t
B)}.
Solution Hereo:3.

L{tH(t - 3)} :
:

':

t{[(t -3) +3]]/(r - 3)]


e-3'L{t+ g} : e-3' 1t1ty+ c{g}l

gl.

' [1*
"-'"
["''sl'

Example 83 EualuateL{t2H(t - g)}.


Soluticn

L{*Hft -3)} :
- :
:
:

t{t(t -B) +sl2fl(r- s)}


L{[(t -3)2 + 6(t- 3)+ e]r/(t* 3)]
u-tc21sz+ 6t + g)
9l
*;] '
"-t'[3*
l s o6 P

105

3.5 Special Functions


Example 84 EualuateL{costF(t - r)}.
S o l uti on He r e f:n.
L{costH(t- ?r)} :,C{cos[(t - ?r)+ nlg(t- ?r)]
:

e-"'L{cos(t + zr)}

e-"sL{cost cos?r- sin t sin n}

-e-otL{cost} ::-s-"t

p-h.

Alternatively we can also do the following.


cosf :

cos[(t-r)+rl

cos(t - zr)cosn - sin(t - zr)sinzr

-cos(t-zr).

Thus
L{costfl(t - ?r)} :

-L{[cos(t - r) H(t - 1ri-t


-e-"" E{cost} :

-e-o"

;{il

Example 85 DetermineL-L{+)
L8'J

Solution

L-'

{#}

: L-L{e-2'r1s;},a : 2

F(s ):i

f( t ) : t .

From the translation propertS

L-{* \

t"'J

,f(t-z ) H( t - z )

(t - 2)H(t - 2).

106

The Method

of Laplace Transforms

Exampre86 Determt""L-r{#}.
Solution

c-'{d#}:

t-{e-4,r1s;}
, a:4

F("):

Fb'
f(t): ,-'{#tr}
:

"-2tx-L{*}

:}

t'

Flom the translation property

' - ' {#}

:
:

rQ-4)H(t-4)
o-2(t-4)

?(t_q2 H( t _ 4 ) .

3.5 Special Functions

LO7

Quiz 3.11 Find the inverseLaplacetransform for the following functions.


1 . G( s) :

,":=

Ans:9(f) : (t - t)e-Q-s)g(r - 3).

2l . H ( s ) : # s . An s:h (t)_ " *2 (t_ r)si n (t_tr ) H( t_r r ) .

108

The Method of Laplace Transforms

3.5.1.3

Initial Value Problems

Example 87 Solae
r(0) : g,

r" (t) + ar(tl = s(t),

o'(0; : g

where

s (t) :

t,
{ -1,
( o,

o < r < i.
t< t < 2
t> 2 .

Solution
Write g(t) in terrns of unit step functions aiid find its Laplace transform,

s (t) :

t-z H (t-1) + H( t - 2 )

L{s(t)r:
:_T.+.
Then take the Laplace transform oi both sides of the ODE,

t, {n"(t)} + aL{a(t)} : L {s(t)l

: -'+.+
- sr(o)
- 0'(o)]
+ax(s)
[s2x(s)
:
1

2e-8

(s^ z +4)x (s ):;-;+-.

e-2t

Thus

x(s) : _=1-J:-"
,=+ "1" ,.
- s(s2
s(sz+4)
+4) ' s (s 2+ 4 )
:

n(t) :
where

F(") - 2e-sF(s) + e-2'F(s)

c-l 1r1s;1- zL-t {e-'r(s)} + L-r {e-2'r1s;}

r() ' '\


W
r ,(:;(;tTZ):A\;/-4\
s) :+ :lr l\

giving
r

l(t): i- ie.os2t.

3.5 Special F'rrnctions

109

Tb^kingthe inversetransform of r(t) gives

r(t) : l(t) - zf (t - r)H(t- 1)+ f (t - z)H(t- 2).


Hence, via the translation property,

n(t) : l-1-lcoszl
L4 4*"'" J
l-r i

- . - 1)] rr(t - 1)
,cosz(t
|.;
fr 1 _.
.l
*
[a ncosz(t-2)JH(t-2).
-

110

of Laplace Transforrns

The Method

Quiz 3.12 Solvethe following initial value problems.


L . A" + y:

g(0): 1, y'(0) :6

H(t-3o),

Ans: g : cost + H(t2. y" +A:

g(t),

3n) - cos(t- zr)H(t - 3n).

g(0) :0, g'(0) : 1 where

s(t) :

( t

{ ;.,'
L

o< t < rl2" ' '


;t2.-r/=2:;
.

Ans: gr- 1 - cosf * sint * (sint - L)H(t - T 12)


3. x"(t) + 16r(t) : g(t),

r(0) :0, c'(0) : 1 where

,r\
s(t) :

Ans:r(t) :

( cosht,, 0 3 t < n

1\,

o.

t) r .

-r)H(t-r).
sinat+ t rio+t- -zr)sin4(t
l,t
f

111

3.5 Special Functions


3.5.2

Dirac Delta Function

Definition

L2 The Dirac Deito function 6(t' - a) is definedby


6(t-a)

( ,o.
{ ;'U,
I

xl t : a
otherwise

and

6(t-a)dt:L,
f6' r fi

J,

, f ( r ) d ( t - a ) d t:f( a ) .

. li&" o \
f ;fl"- + jlf

\J .l

o Problems involving short impulses (a very large yalue over a very


short interval) can be describedby the delta function.
o Exa,rnples:a vibrating airplane wing siruck by a bolt of lightning, an
airplane makes a ha"rdlanding, a terrnis ball is hit, a ship is irit by a
single wave,etc.
3.5.2.1

Laplace TYansform

Theorem LL For a ) 0,
L{6(t- o)} : e-o"and, l-t 1"-osy: d(t - a). ^ '*"fu
,[
:
(a)e-"".
a)}
(t)d(t
Theorem L2 For a) 0, L{f
f

4&t'" t:=
: !

Example 88

- 9 r3s.
f, {t2 t1t - 3)} : /(3)
"-3s

Example 89

4 {costd(t - r)} : f (rr) e-rs : cosr e-os : -e-r$ '

3,6.2.2

},1a"1

Initial Value Problems

Example 9O Solae
a" + 3a' *2Y - t(t - 1), Y (0): 0 ,

g ' (o ): 0 '

LLz

The Method of Laplace lbansforms

Solution Taking the Laplace transform, gives

L { v " } + 3 L { a ' }+ z L { v } : r { d ( t- 1 ) }
["'v(") - ry(o)- /(0)] + 3[sr(s)- s(0)]+zY(s): e-'.
Usingthe initial conditionsand solvingfor lf(s) yields
( s 2 +3 s+2 )r(s) :

e -s

1
y(s) : e-,Wfu:

e-eF(s)

where

r(s) :
:

11

;,fu:

(s+ l)G + 2),

(usepartialfractions!)

11

,+1--s +Z'

Hence,

/(t) : c-l 1r1s11: e-t - e-Zt


and
v(t) :
:

f-1 {e-'.F(s)}
f(t - l)H(t - 1) : [e-(t-t) -

"-z(t-t)1p(t

- 1).

Example 9L Solue
2y" +A ' +2A : d(t - 5), g(0): 0 ,

g ' (0 ): 0 .

:
Ans:
sr(t)
rya(t-s).
h"-u-s)/asin

3.5 Special Fnnctions

113

Quiz 3.13 Solve


L. y' - 5y : (t + \26{s,- 2), g(0) : 3.
Ans: g(t) :3e5t 19"5(t-z)g$-2).
2. u' - 2y : et d(t - 1), y(2) : e4.
Ans: y(t) : (, -1)
H(t - 1)
' ee2(t-r)
e J ""
\
3. g" * a : 6(t - ?rr) cost, g(o) : o, yt(o) : 1.
Ans: 9(t) : sinf * sin(t - 2r)H(t - hr).
4. g" +4y : 6(t-r\

- 6(t -hr\,

An s: s(t) :|sinz(t

y(0) : 0, y'(0) : $.

- r)H(t - r) -|s in z (t

-z n )H(t -2 o ).

The Method of LaPlace Transforms

LL4

3.5.3

PeriodicFunction
t3 A lunction f(t) is sai,ilto bepetiodic of peri,odT if

Definition

t(t +r) : l(t).


Example 92 Sketchthe gmph of the periodic function

ilt):

( t-r
{ ' ^ "0' ,
[

/(r):

f(t+2).

o<r<t
I< t< z

Left as on enerc'i,se.
3.5.3.1

Laplace Transform

Theorem Lg If i(t) has periodT, then


'l

fT

E{f(t)}: T#

l,

e-'t761at

Example 93 Find the Laplace transfonn for

f(t) :

( t.

o<t<1

I ;:

L<t<z

( "t

f (t) :

f (t +2).

Solution

L{f(t)}: #
:

fo'"-"t{\a,
L fl

I e-ttdt
-L _ e-zt Jo
1
s(l + e-s ; '

3.5 Special Functions

115

Quiz 3.14 Find the Laplace transform of

l(t) :

( r
{ 0,
[ ^'

f(t):

f(t+2).

o<r<1
L<t < 2

L-e-s
e-s
r
r(s):AG\-;G_;=t

An s :.,-r

116

The Method of Laplace


Transforms

Chapter 4

Fourier Series

Objectives
On completing this chapter, the students should :
be able to:
1: determine the Fourier seriesrepresentationfor continuousand
piecewisecontinuousperiodic functions on l-LrL].
2: determine the convergenceof the Fourier series.
3; compute the Fourier seriesrepresentation for even and odd function.
4: obtain the Fourier seriesrepresentationfor functions definedon [0,Z].

Fourier seriesa,reseriesof cosineand sine terms and arise in the important practical task ofrepresentinggeneralperiodic functions. They constitute a very important tool in solving problcms that invoive ordinary anC
partial differential equations. Areas of application include electrical engineering, vibration analysis,acoustics,optics, signal and image processing,
and data comprassion. The Fourier series is named after the FYenchscientist and mathematicianJosephFourier, who used them in his infiuential
work on heat conduction,Thorie Analytique de la Chaleur (The Analytical
Theory of Heat), publishedin 1822.

118

4.L

Fourier Series

Periodic Function

A function /(r) definedon [-.t,.[] is said to be periodic with period T : 2L


i f f ( x + ?) : f ( n),T lo.
Elxample 94 Show that f (") : cos2nx is a peri,odicfunction and d,eterrnine its peri,od.
Solution

C h e ck
f ( n ) :f @ +r)
Since /(c) : cos2na

+ f (a+T):cos2r(n+ f ).
Find the value of 7 so that
cos2r (r +T) : cos?nx.
We know that from trigonometric identity

cos?tr(x*T)

cos2rx cos2rT - sin2rssin2rT


cos2nr.

If cos2rT : 1 and sin?rT : 0, this is true for ? : n, where n :


L,2,3,.... The smallestvalue for T is T = 1. The value ? : L is known as
fundamentalperiod. Therefore / (r) : cos2trr is a periodic function with
period 1 (the smallest period).

4.2

Even and Odd F\rnctions

In mathematics,even functions and odd functions are functions which satisff particular symmetry relations, with respect i,otaking additive inverses.
They a^reimportant in many areasof mathematical analysis,especiallythe
theory of power seriesand Fourier series.

119

4.2 Dven and Odd Ftrnctions


4.2.L

Even and odd functions

Definition L4 Let l@\ be a real-aalueilfunction of a rea! uari,able.Then


f (s) i,seaenfunction if
(4 . 1 )
f (-a\: f (s).

Geometrically,an evenfunction is symmetric with respectto the 3ra:ris,


meaning that its graph remains unchangedafter reflection about the g-a:ris.
Examplesof evenfunctions arelxlrn2rt4rcosfi 6nd cosbc.
Definition L5 Again,Iet f (r) be o real-ualuedtunction of a real uariable.
Then f (x) is oilil function if
f (-"):

(4 . 2 )

-/ (") .

Geometrically,an odd function is symmetric with respect tc the origin,


meaning that its graph remains unchangedafter rotation of 180 degrees
about the origin. Examplesof odd functions ate 0,rr3.sinc and sinhc.
Example 95 Showthat !(r):

sitrrc,is an odd'function'

Solution
We can checkeither from the graph or from the definition.
From the graph( left as an exercise)'/ (-") : -f (n).
.'. f ("): sinr is an odd function.
From definition, / (-") : sin(-r) : - sins
.'. f (") : sinr is an odd function'
Elxample 96 Showthat t@):

coss is an euentunction'

Solution
Flom the graph (left as an exercise),I (;r):
.'. /(o) : coso is an even function'
Flom definition, / (-") : cos(-r) : coso.
.'. f (r) : cosr is an even function'

/ (r) .

Fourier Series

4.2.2

Basic properties of the even and odd functions

and the productof two evenfunctionsis an even


1. The sum(difference)
function.
Even * Even -' Even
Even x Even -+ Even
2. The sum (difference)of two odd functions is an odd function. The
product of two odd functions is an even function.
Odd + Cdd -r Odd
Odd x Odd -+ Even
3. The product of an odd and even functions is an odd function. The
sum (difierence) of odd and even function is neither even nor odd
function.
Odd x Even -r Odd
Even * Odd -+ Neither Even nor Odd
P fL,

If /(r) is anevenfunctionthen,

l_"f

'1L

If f (s) is an odd function then,

l_"f

(r)d* :,

Jo

t @)dr.

(a)d,x:0.

Therefore,if /(s) is an even function then


7L

J_L

t @sinnrdn :0

pL

and

J-L

1L

f e: cosnrdx:2 I t (x) cosnrda


JO

and if /(o) is an odd function then

1L

J_L

pL

:0
t @ cosnxd,r

and

J-L

1L

t @sinnxdr :2 I / (r) sinnndr.


JO

Lzt

4.3 Fourier Series

4.3 Fourier Series


Fourier series is a series representing some real periodic function as an
expansion in a seriesof sines and cosines. Fourier series of a periodic
function J(r) on l-L,Ll is given bY

f (s):?.i
-

n=L

lo""o,
ff+b,sinW]

(4.3)

where

, =

ao: i I:"r @)d,x


an: LI : r / ( r ) co s T o ,

(4.4)

bn: LI:,/(c)sinTo,

(4.6)

(4.5)

and ? is the period of the periodic function.

If /(r) is an even function brz: 0.


.'. Fourier seriesis given by :

f ( x ) : ?. E on " o ,ff
where

: ?|,,

! (r)d,

: ?1,'

/ (r) cosTon

g .7 )
(4.8)

(4.e)

It f (r) is an odd function d0 :0 and or, : g.


.'. Fourier seriesis given bY :

f (r):i r"sinff

(4.10)

n=l

where

u^--Z,I, / (r)sinffar.

(4.11)

Fourier Series
Example 97 Find the Fouri,er seriesfor
( Zo, -n1c (0

f@):l r, s:o
| 2r, 0<r < r
l @ ):f(n+2r).
Solution

Figure 4.1: Graph of f (x) for -zr < n < r.


Flom the graph the function f (s) is neither evennor odd function. we
needto find o6,an a"ndbn

ao :

* [" zra,l
:i
f 2trdr
T LJ-"
Jo
J

:
:
:3r

an :

)lr*,11"+,rl;l
|lrr";r)l+r2

[0"2xcosn'a'l
+lf:"2trcosnrd'a*

: +ft'+']:".1'ry.'#f,l
s i n cesi n( - n r ) - -sinnzr:0, n:!r2,8,...
and cos(-nzr)= cosno : (-7), n = 1,213,...
we obtain,

4.3 Fourier Series

An

bn

L23

fiter r - tt
lf f 2nsinnad,x+
[" zasinnrd'f
---l
r Ll -"
Jo
i | | -2rcnsnslo, | -2rcosnx. 2sinn"l" I

;L | . "] - , *L

_?

" *;r.J,J

.'.f (s) :

3u *r$

[ (-r)" - r ] cos??,
_ 2i
T-"k-W

:k;

1 rior".

Example 98 Find the Fourierseriestor


f (x\:4 -

r', -2 < s <2

l@):f@+Q.
Solution

Figure 4.2: Graph of /(o) fot -Z I r < 2.

Flom the graph the function is even,so b,, - 0. Fourier seriesis given

L24

Fourier Series

by @.7)and os and o, are given by (4.8) and (a.9) respectivelyas


2 l f2.
ao : ,l,Jr
4-nzdol

:
:

f.

s312

[ n"- r lo
16
3

an: ?l Ir'*- o2)


cos
(T) *l
:

8s
l n r n \ -a-*'(ry)
-#'t(ry) ].
.,,i'h(,
/
L.
I u- r') 3

t@):;.#i#*'(ry)
Example 99 Find the Fourier Seri,esfor
./ \

f l0l: (

| -1 +a,
[ 1*c,

-t1 o < 0
01 r1 r

l@) : f@ +ztr).
Solution
Flom the graph the fu;rction is odd, so o2 : o0 : 0.

4.3 Fourier Series

L25
t(x)
1+1E

.n

fr

0
-1

-1-fr

Figure 4.3: Graph of /(r) for -zr 2 * < o.

Fourier seriesis given by (a.10) and b,, is giverj by (4.11).

bn: ?llr"U+x)sinnndrl
:

2l

(1+r)

7(L

I --

COSnJf -l

I
-n

SlIlI}O

l"
I
J0

: L-rrr- (1+ rr)(-1)"


l
since
( -?.
,,'
|

b": I

a+z r.

\NT

n eaen

n odd

...f (r): -2i fisinznx.ry,pr+

sin(2n
- t)r

Fourier Series
Quiz 4.t
1. Find the fourier series of
(a)

r + n , - n 1 a .- o ;

fbr:I

I o -" , 0< o< zr ;


l @):f(x+2 zr).

(b)
(

-3<c(0;
"'
I r, 0 < o < 3 ;
l @ ):/(o+6).

l @ ):l

2. Showthat the Fourier seriesof


(a)
f7l

| 0 , -r 1r <- ;;

f(s ):

Irtf'

n'
1
:i"<,i
0,
[
i,., .o,

t@): f (n +2r).
is given by

f (s):2 +"9
7f

[.*r

- ]
cosbs
- 1*, ,. *.f
"orss+ *

(b)

ro:

,.1,0t
{'.:,;:;

l@):f@+a\
is given by

:| - :,,:ts)1 "o"ff-1,io
f(s)
ryl

4.4 Half-range Series

4.4

Half-range Series

Half-range series is a periodic function with period 2-L defined on [0, Ll. Tt
the Fourier seriesis given bY

f (s):?.:

on*,ff,

then /(r) must be expanded to an even function. If the Fourier series is


grven by

r@,):iu"sinff,
n=I

then /(r) must be expanded to a.nodd function.

Example LOO Firul the half-mnge Fourier series of


f(x'):7t-fit

0<t1r.

Solution
1. To obtain cosine Fourier series,we must expaod the above function
to an even function.

Figure 4.4: Graph of /(c) fot -r < n < lt.

Fourier seriesof an even extension is given by (4.7) where

L28

Fourier Series

o ,o: ? lr" r-a)dx

: b,-*1",

: 71

n,a,l
i[ 1,",- c)cos

an:

2 l (tr-x).
1
1o
srn t?$- Acos"r lo
;L
"

...t@):

ftft-tr*'+rl

;.+,:1'],

cos(2n_L),.

2. To obtain sine Fourier series, we must expand the


above function to
an odd function. Fourier seriesof an odd extension given
is
by (a.10)

Figure4.5: Graph of f (x) for -r 1a 1r.

where

L29

4.4 Half-range Series

I
2l f".
;LJ r(n-r)s in n a d n f
1
21fu-r)
l"
cosnn
Asnnn Jo
; ),-;

.
bn:
:

:? n
...f(n) :

,=irlsinrao.

Elxample LOL Fi,ndthe cosi,neFourier seri,esof


0<a < r'

f(r\:2a,

Solution
CosineFourier seriesis for an evenfunction. Therefore,b,, :0 and the
cosine Fourier series is given by (4.7) where

an:
-

? ["rra*
rJo

:hr
,. 1 f1(

I
'l
| Znco s n n d , n l
nLJo
J

an:
"

_ tl

;pIFt)"

sin ce

o":1

(0,
I

.. .(*)
/ , ':

o-

nea e n i

-9.

ss

rn2'

n odd;

; *L_FL1

' - - 1)*
cos(2n

Fourier Series

130
Example LOZ Find the sine Fourier series ot
( t
"'o
{ -J,
[

f(*\:

o<n z - ;
:
+<r<8
;

Solution
Sine Fourier seriesis odd function. For an odd functiorl, 40 : an:0
and sine Fourier seriesis given by (a.10) where

bn: ?l Ir^rsinffdn
* In'-Bsin
ry*l
6r.nr'l

(-1)"- zcos
7J
;L, +

.'.f(*): *i#
" n :l

\' "/

-2cosT]'.
[r+1-r1"
ry
-

131

4.4 Half-range Series

Quiz 4.2
1..Sketchthe graphof
o <t<r
f('):-s
:
l @) f@ +%r)
and then find the SinusFourier seriescoefficient,brr.
2. Find the cosineFourier seriesof
(2

o< c < f ;

l;''
/("):

{
It

I i( s - " ) , 8 < ' < s ;

3. Show that the sine Fourier seriesof

l@):

( r.

0< r< 1 ;

I J'_r, L1 r . - 2;

is given by

k Lw

" " \ z / *; J

s m\2 /'

Fourier Series

4.5

The Sum of the Series

We can find the sum of a seriesby the respectiveFourier seriesof /(z). For
this, we needto understand the Fourier ConvergenceTheorem.
Theorem t4 Let t@l arul tt(r\ be periodic function haaing perioil 2L,
wheref (n) and f'(x) are pi,ecewise
continuousonlL,-Ll. Let the Fouri,er
cofficients be defined bV (4.5) and (J.6). Then the Fourier series (j.3)
oonaerge
at eo,ch
u i,nl-L,L).
Series (1.3) conaergesto the aalue
l@);

if a is a point of conti,nuityin [L,-L].

(4.L2)

Howeuer seri,es(J.3) conuergesto the ualue


1

if n is a poiroiof disrcntinui,ty
in lL,-Ll.
+
,tf @n) /(cot)l;

Example 103 Il a seriesof ihefollowingfuncti,on


-2 <s< 2

f(s\:4 -*2 ,

f(a):f@+a)
is giuenby
4 . 8 g (-1 )"

,, \

/(c) :-g * p

_!_T

nr r

cos
2

show that by taki,ngan appropriate uolue of x the series

- , * # - b#*
,o
conaerges

-2

i.

That is,
11112

-r*F-g2+

+ r -...:T

(4.13)

133

4.5 The Sum of the Series


Solution
Tbkings:0,

: i.3i+,1
/(o)
ura

e T'- t 1 \
4 - 4
s+7 (-t* fr-F+ . . . )

+ : -r*#-#*...Example tO4 Show that the tourier seri,esof


f

t*r'
|. 0 ,

r\fo l' :l

(o
-t.u
O<s< n

l @):f(n +ztr)
is giaenby

- i sinnz
I *?i "",{'" =.tJ"
f (,):-4'rf;o(2n-t)2
n
?-

Then, by chaosingan o.ppropriateaalueof u, finil the sum of the infinite


series
111
1 * t+sr+F+
...

Solution

ao :

L f0.

; J__@tr)dn
L rr2
1o
;lz * n'l -o
7l

an :

2
L fo .
; l__@*r)cosnndn
o
r,
, sinnc cosrzc]

l@+ n)-

t (L -(-1 )" \
:;\-r1

+ -Tl_^

Fourier Series

b n:

L fo o *tr)si n n rd ,r
Tt J-n

| ,
\cosnr.sin n rlo
-.lrl-."
f -(s*r)-*. n

:-1

...r@): zlilz *:iri-=r\ (t -st)")


- i sinns
rto
n
"o,,,"
/
7=,
r,2 $ co s(2 n -L )x

$sinna

A- ;kw- k

We chooses : 0. But r : no :0

"
is a point of discontinuity since

f bt): o and!(ai) : o.
:
.'.f (x) converses
,o
0) :
*"obtain
|tlt";l + /(coF)l *b+
i.
jr

tr2
.'. g :

--!-

2g\

1
-

4 'rQ(Z n- t1z

Ln=,

I+,+N +U+ . . . .

135

4.5 The Sum of the Series


Quiz 4.3
L. Let

., \ [ o, -3<r<o
Iw): r", o<r<B
l
f (*): /(s + 6).

Showthat the Fourier seriesof /(r) is

- rl*r(T1= a1_r1"sin(ff)
3 . s 6[(-1)"
t-k- @Then, by choosingan appropriate value of c, estimate the sum of
111
1*F
+ g+ a; ry + ...

2. rf
f (*):

T2 - n2, -'tt < n <

'ti

l @ \:f(r+2r).
*2
(_r)"*t
Show
'---' l:L2 $
---" that
n2
3
3'L e t

,,\
{z n, -trls ( o
I\n):7 z r. 0 ( s ( z r
\,
l @ ):f(n+2r)

(a) Sketchthe graph of /(r) for -3zr 1 s 13r.


(b) Show that the Fourier seriesof /(r) is

3z r,

(-r)"-r
;*2 ^S
r\_Acosnn(c) Henceby taking r :0, showthat
1
I
r,1
g : t + s r + F *F + - ..

3n2

S s io n s

"

Fourier Series

136
4. Let
/('):

( n.
"'
{
[ ",

-r1n10
01a1n

l@),: f (r +zn)
(a) Sketchthe graph of J(c) for -3zr I s 13r.
(b) Show that the Fourier seriesoi /(r) is

3 " , S (-1 )" -1

_- _- _ gsinnc

n * kfficosnn - k

(c) Henceby taking s:0,

"

showthat

7P 111

8:1* gr* p+ F + . . A function / is given as follows


fb\:[",0(a(3

"-'- I o, -3<r<o'

l @ ):/(o+o)
(a) Find the value of /(20).
(b) Sketchthe graph of i (c) in the interval of -G < s < 6.
(c) Expand / in a"Fourier seriesan<ishow that

(d) Use the result in (c) to esiimate the sum of


111
''92'b2'(2n-l)2 " "

'lJ---L-J---l-

Chapter 5

Partial DifferentialEquations

Objectives
On completingthis chapter, the students should
be able to:
1.: use the method of separation of variables to solve simple
linear partiai difierential equations
2: solve initial boundary value problems for one-dimensional
linear heat and wave equations in Cartesian coordinate system
using the method of separation of variables.
3: solve boundary value problems for two dimensional Laplace
equation in Cartesiancoordinatessystem using the method of
separationof variables.

what is a partial difierential equation (PDE)? What types of PDEs are


there? What do they model? How do you solve them? These are someof
the questionsthat will be answeredin this chapter. First, let us look at
somebasic definitions and terminologies.

Partial Differential Equations

138

5.1

Basic Definitions and Terminologies

A simple definition of partial difierential equation is as follows:


Definition LS A portiol differentiol equation (PDE) is an equati,on
containing at least one parti,al deri,aatiue.
We will sometimesdenote partial derivatives of functions by means of subscripts. Thus, for u: u(xrt)
ur:

0u
6,

uan:

02u
W,

ut:

02u
0u
u,nv:
At'
@:

0 /0u\
a, \e, ) '

etc'

Most of the classificai;ionschemefor ODEs appliesto PDEs as well. Thus,


a PDE can be of first or higher order, it may be a linea,ror nonlinear, and
may be homogeneousor nonhomogeneorr.

Definition

L7 The order of the PDE is the order of the highestderi,uatiue.

Definition LB The PDE is mlled,linear il it is of the first ilegreei,n the


ilepenilentaoriable and also in its partial deriuatiaes.

Example !05 Generaltorm of a first order lineot PDE in two aari,ables

: f (x,a)
b(n,
AH * c(r,y)u
"Oru\# *

(5.1)

Example LO6 Generalform ol o second order linear PDE in two uariables

--f (,,y)
y\ffi *c(r,
op,
dff+"6,ilft +o{r,s)u
il ffi +af*,
fiffi+b@,
(5.2)

5.1 Basic Definitions and Terminologies

139

Definition L9 The linear PDE in equati,on(5.1) is said to haueconstant


coefficients if arb, and c ane constants. Othertaise,it is saiil to haue
aariable coefficients,
Definition 2O The linear PDE in equation(5.1) i,s cailed.hornogeneous
it is callednonhomogeneous.
if I :0. Othenui,se,
W.pcan generalize definition 19 and 20 to higher order equations as well.
What can you say about equation (5.2)?
Deffnition 2L A function u is calleil a solution of o PDE if all its ileriaatiaes con be e,alculated,and the PDE is satisfiedupon substitution of all
the deriaatiuesappeareilin the PDE.
Example LOT Veri,fythat the tunction
u(t, Y) : coss sinh3r
i,s a soluti,onto the equation
02u

02u

w+@ :u .
Solution
ula: - sin r sinh y,

'ltrns: - cost sinh 3r

ug : Cost coshgt, uyy : cosc sinh g


Therefore,
uao* uyu: - cososinhY* cososinhg: g.
Thus, u(r,y) : cosr sinhy is a solutionto the PDE.
In this chapter, we a,reconcernedwith linear secondorder PDEs. They fall
into three types. The classificationdependson the sign of b2- 4ac as given
in the following definition.
Definition

22 The equation

o@,d#
* b(u,il#* c(r,rr#:r (",n,",H,H)

140

Partial Differential

Equations

i,scalled
pa lb o lic
hgperbolic
elliptic

,f b2- 4ac:0,
xl b2- 4ac > 0,
Lt b2- 4ac < 0.

These classesrepresent completely diferent physical processes.For example, parabolic equationsrepresent diffusion. The simplest exampleis
Example lO8 The one-dimensi,onalheat equation

ou :
o-W
'o2u
at
whereu@,t) is the temperatureond a2 i,s the d,ifiusion cofficient, calleil
thermal diffusiaity which goaernsthe rate of iliffusion.
Hyperbolic equationson the other hand representpropogation An example
is
Exampie tO9 The one-dimensionalwaueequation
02u

,0 2 u
: o-ffi
at2

where u(.r,t) nxay represent the displacementoJ a stretcheil stri,ngfrom


equlibrium, and a2 is a propagation constant that describesthe uelocity of
ihe wave.
Finally, elliptic equationsrepresentsteadgsfate(equilibrium). An exa,mple
is
Example tLO The tuo-dimensional Laplaceequation
02u

02u

a;t+@:u
whereu(x,g) may describetemperature,electric, gmuitational, or fluiil potentials in equilibri,um.
Note that the equations in Example 108 - 1L0 are linear. Why? In the
next section we will lea,rn how to solve those three important equations
analytically using the method of separationof va,riables.For that, we need
one important theorem. In simple words,

5.2 The Method of Separation of Variables

L4L

Theorem LE (principre of superposition): The comprete


sorutionof a
Iinear PDE is the sum of all possiblesolutions.
The following section explains a simple technique on how
to solve a liaear
PDE analytically.

5.2

The Method of Separation of Variables

There are various techniques (analyticar and numerical)


for solving pDEs.
separation of variables is the brdest and most efficient
sorution technique
for a certain class of pDE problens. we will apply the
method to initial
boundary value problems (IBVp) for the one-cimensional
heat and wave
equations,and to bounda.ryrnlue problems (BVp) for
the two-dimensional
Laplace equation. Before ihat, lets look at simple examples
on how the
method works.
The basic idea is to assumethat the original function
of two variables can
be written as a product of two functions, each of which
is only dependent
upon a single independentvariable. Based on the assumed
solution form,
the PDE is convertedinto two oDEs that can be solved
separatery.Then
we combine the two solutions to give the solution for the
corresponding
PDE. We will demonstratethe method through examples.

Example ttL (Ise the methoilof separationo! aoriablesto


find the geneml
solution of the equation

0u
0u
a":Y ay '
Solution
(a) since the dependentvariable z dependson o md y,
we seeka solution
of the form
u(r,a): X (x)Y (y),
wherex and Y are independentof eachother. Becauseu(x,g)is
a solution,
it must satisfy the given pDE. Differentiating u with respect
to u and g

Partial Differential Eqpations

L42
respectively,gives
ou

' fltt' : *(*)Y'(d'


: x' (a)Y(s)and
d
fi

Putting theseinto the PDE yields


X'(r\Y(s) : y X(x\Y'(y).
Dividing through by X(r)Y(y),

leads to

X'(s)

Y'(y)

W:Y

Y@)'

Since the left hand side of the equation is independent of 3rand the right
hand side is independentof r, the ratio must equal a constant,say /c.Thus,
we ca,nsplit up the equation into two ODEs

W:i
ffi:r and

To solve the ODEs, we have to use appropriate techniques that you have
lea,rnedbefore. For this czule,we jusi need (separation of variable - Chapter
1) to integrate with respect to appropriate variable to obtain

fx,(n),

== [ua,
J

JW""

lnX (r)

kx*c

X(")

Aek'

and
Iffi^r= Ii*
lnY(sr) -- klng * c
lny&+ lnB
Y(d

BY''

Finally, combinethe solution to give


u(x,u\: X(r)Y(y) : Aek'Bgk : C"k'Yk
whereC is a constant.,t can be any number.It couldbe zero,positiveor
negative.Couldyou statethe solutionfor the 3 possibilities?As a guide,
you couldwrite k = -12 for negativevalueof k, etc.

5.2 The Method of Separation of Variables

L43

Quiz 5.1 Find the generalsolution of the equations

,.H * H:0.
2 . au r * yuo - 9 .

Ans: u(r, y) : g"k(o-r'1

Ans: u(r, A):C(;)-

Partial Differential

L44

Equatio4s

Now, lets move to a more challenging and realistic problem. Remember!


Pay attention to the stepstaken in the processin obtaining the final solution
to the a.ssociatedproblem. The procedure is similar for all S classesof PDEs
involving the linear heat equation, wave equation, and Laplace equation.

5.3 The Heat Equation


Consider the one'dimensional heat conduction equation
7trt:Q2rJss,

0<f

<L,

t>0

(5.3)

describingthe evoluiioa 6f fsmperature inside a thin uniform rod of length


.6. The unknown function u(art) describesthe temperature at the point o
and i. r denotesspatial coordinate while t denotes time. o2 is the thermal
diffusivity (a constant). The dyna,rmicof the temperature dependson the
bounda,ryand initial coaditions.
Sinceequation (5.3) is of first order in time, we need only one condition
that prescribesinitial temperature distribution in the rod, na,rrely
u(o,0):l(r),

0<r < L .

Likewise. equation (5.3) is of secondorder in the spacevariable. So, we


needtwo bounda,ry(endpoint) conditions. That is, one condition at r : 0
and another condition at a:.6. The way the condition is prescribedgives
rise to difierent physical significance as illustrated next.
5.3.1

Rod with zero temperature

at the endpoints

For this problem we prescribe the temperai;ureat both endpoints specifu


cally with zero temperature. The correspondingIBVP model for this problem is
0u

o02u

t>0

(5.4)

u(0,t) : g, u(L,t) : g, , > 0

(5.5)

U :o'f,* r,

0<x < L '

with boundary conditions

5.3 The Heat Equation

L45

and initial conditions


u(r,O):f(s),

0<n< L .

(b . 6 )

Now, we want to find the lsmperature distribution in the rod over time.
Basically, solutions of the heat equation are characterized by a gradual
smoothing of the initial temperature distribution by the flow of heat from
wa,rmer to colder areasof an object.
Solution
We seek a solution of the form
u@,t): X (a)T(t).
Because
0u

at:

X(x)r,(t) and

fr2o,

#:

X,'(r)T(t)

(5.4) becomes

x@)rt$): azx"(r)T(t).
Dividing both sidesby a2X@)f(r), yields
L T, (t)

X,,(r)

eN:w:o

where,bis the separationconstant.The aboveequationyieldstwo oDEs


L T'(t)

,, , X"(r)
--n
oz7111 W:*

which can be written as


T' - a2kT :0

and x't - kX : 0.

At this point ,k is still arbitrary whosesign is yet to be cetermined. Next,


we will look at the boundary conditions.

Partial Differential

L46

Equations

Boundary conditions
Remember, we have two oDEs that need to be solved separately. so, we
needto separatethe boundary condition: (5.5) so that we have at least one
completeset for one of the ODEs. Ftom the boundary conditions (5.5)' we
seethat
X (0 ) : 6

u(O,t):X (o)"(t):0
u(L,t) : X(L)T(I) :0

=+ X(I) : g.

: 0 then u(n,t) : 0r a trivial solution which is of no


Note that if
interest. Thus"(t)
10. Hencewe have
"(t)
Tt - azkT:o

x" - kx :o, x(o): o, x(r) : g.


We will find solutionsfor X first. For that, we will look at 3 casesdepending on the valueof k.
Case1: ,t :0
=+ X(r) - Ar * B

Xtt:0
X (0 ):0

=+ B :0

X (L I:O

X:Ar

0 :A L

A:0

Thus X(r) : gCase2: /c:.\2 > 0


: o

x" ^2x
X (0 ):g
X(L ):0

x(a) : Ae\' + Be-\'

=+ A --A +B

+ 0 :A e L '-A e -L'

B:- A
A:0:B

Thus X(r) : g.
Case3: lc: -.\2 < 0
xt' +.\2x : o

X(x) :Acos.\r * Bsin.\r

5.3 The Heat Equation

X ( 0 ):0

L47

X (L ) :0

=+ X:Bsin,\o

0 :A
+

0 : B si n ).L, B *0.

Hence
sin.\.t:0

\L:ntr

n:L,2,3,...

^*:T,

These .\r, are referred to as the eigenvaluesfor the correspondingeigenfunctions Xrr(r),

X,(r): Bnsinry ,

n-: L,2,3,.
.

Tio solve ?(t), consider the 3 casesagain. For this particula,r problem, only
case3 gives nonzerosolution of u(r,t). Thus, for

k: -\2- -^7:- (ry\'


\L /
ri, + a2x?rTn: o

Tn(i) : Cne-a2\?*t

Hence i;hefunctions
un(x,t)

X"(n)y"(t)

Br,sin
ff

b rsinS

C""-a2)'f;t
"-a2\7t,

bn: B n x Cn , n :

I r2 rlr. . .

are solutions of the heat equation satisfying the given boundary conditions.
Since we have one more condition (5.6) to be satisfied, we will find the
entire solution tc the whole probiem using series as discussednext.
Series solution
By the principle of superposition (the complete solution is the sum of aii
possiblesolutions),

u(n,t) : i

n=r

u,r*,4: i

6,,sinW
IJ

"-o'x'^t

"=

: iu" sioffu-#'
n=l

(r.z)

148

Partial Differential

Equations

Initial condition
Applying the initial condition (5.6) to (5.7), gives

: !(n)::
u(a,o)

b^sinff.

The problem is thus reduced to determining a,nexpansionfor /(o) of the


form

oo

l @ ):Ie"s i n$,
a Fourier sine series. Thus

b,: ?

0<a<L
lJ

"=-t

..
fr"talsinff aa,n= 1,2,.
Final solution

Finally, substitute bn back-into the u(r,t) in (5.7) to give the solution for
the wholeproblem (5.4)-(5.6).
Note: In future, we can ignore case I (ft : 0) and case2 (k > 0) if the
heat conduction problem have zero iemperature at both boundaries.

Example tl?

Use the aboaesolutiontormula to solve the IBVP


Au

02u

at:w '

0<t<r,

t>0

.r(0,; :

u(n,t)-c , t > o

u(a,0) :

0,

0<s< r.

Ans:u(r,t):2ikl]"fl.
n=r

e-n'tsinn*

5.3 The Heat Equation

L49

Figure 5.1: Temperaturedistribution of exa,mple112.

Example LL3 Solaethe IBVP


0u

02u

i:ffi,
u(0,t) :
u(x,0) :

o < c < 1 , t> o

u(1,t) : 6, t > 0
sin2nr- 3sin6zrc, 0 ( n I L.

Ans: u(2, t\ : s-4nzt sin1trr - g"-36r2t sin6urr

Figure 5.2: Temperaturedistribution of example 113.

150

Partial Differential

Equafions

Elxample tI4 Solaethe IBVp


0u

02t

d:#,

0 < n < 1, t> o

u (0 ,t) :

u (!,t):9 ,

u (t,O) :

l @), 0 <s<.1

l@):l ''
I o,

Ans:
u(r,4:

t>0

o< '<L
i<r<r.

E &sinff-

a"o,

T) "--'-","iono,

o.6
o.45
o.4
o.s
o-9

$o.aa
o-e
o.15
o.1
o,o5
o

Figure b.B:.Temperature distribution of


example 114.
Example 7t6 show that the sorutionto
theinitial-boundary aarueproblem
ou

ofu

o <o <1 ,

d t-d 'd ;r,


u (O,t) :

u (l ,t):9 ,

u (n ,O) :

x(l -x),

t> o
t>0

0 <c<1

is giaenbg

u(x,t):

i@.t

t +U
# n=l

"-n2trzazt
"inn'rn.

5.3 The lleat Equation

151

o.?5

o.2

o.tt

o.1

o.o6

Figure 5.4: Temperaturedistribution of example 115.


Quiz 5.2
L. Solvethe IBVP
0u

02u

A r:W'
u (O,t) :
u (o '0 ) :

n (c<1,
0

t>0

u (l ,t):9,
2 x*t,

t> 0

0< o< 1.

Ans: z(c,D : ?i
n u -L

t - (at)"
n

sinnnr.
"-n2tr2t

2. Find the solution of the heat conduction problem

*ot: r o o * ,ofi'
u (},t) :
u (r,O) :

o ( a 1 L , t> o

u (l ,t):9 ,

t> 0

si n ? rs-si n 5zr r , 0( r ( 1.

Ans: u(c, t) -

s-40or""in2o,

- ezlo}r2tsinbzrr.

Partial Differential Equations

t52
5.3.2

Rod with insulated endpoints

Now the rod is insulated at both endpoints. This meansthe heat flux there
is zero. Mathematically, we presentthis by letting the derivative of u(a,t)
The IBVP model for this
with respect to r to be zero' nunelyu' :0'
problem is of the form
0u

o02u

( 5 .8 )

# : o t # , 0 < a < L , t> o


with boundary conditions
ur(O,t):9,

ur(L,t ): 0 ,

t>0

(5 ' 9 )

and initial conditions


(5 . 1 0 )

0<n < L .

u(n,0):f(x),
Solution
Again, we seeka solution of the form
u(r,t):

X (s)T (t )'

We proceedas in section (5.3.1) to get the two ODEs


T'"- azkT :0

and X " - k X : a

where ,t is still arbitrary consta,nt. Then, we separatethe boundary conditions.


BoundarSr conditions
Ftu,o ihe boundary conditlons (5.9)' we seethat
u,(O,t): X'(O)f(t) : O
u,(L,t):

X '(t'\T(t):0

=+ X'(O; : g
+

Hencewe have two sets of ODEs

T' - a2k r :o

X ' (' D)- O '

153

5.3 The l{eat Equation

xt, - kx :0,, X'(o;: 6, x'(L) :0.


Find solutions for X. Again, we will look at the S.casesdependingon the
value of /c.

C a se 1: &:0
X't:0

X (r):A n*B

X '(01:g

=+ -A :0

X t (n ' 1 : 4 .

X'(u7:g

Xt(L'1 : g does not affect B


Thus X(x):

B.

Case 2z lc: )2 > 0

x"-^zx:o

= + .\.4 -,\B :0

X '(0 1: 6

X'(*):.\Aer'-IBe-\'

Ae\'+Be-\'

x(u):

A: B

= + A:o:B

+ A \e L '-A \e -L ':o

x ' ( L ) :o

X' ( ,n) : \Ae\' - ) .Ae- \'

Thus X(r) : g.
Case 3: ,t : -,\2 < 0
x"+^2x : 0
X'(0;:g

X(r) :,4,cos,\s*Bsin.\r

B ):0
X'(L1 : g

=+ B :0
+

=+ X'(r) : -4.\sin.\r*B)cos.\c
X' ( *\:- .A.\sin.\z

-4.\sin).0 : 0.

Hence
si n .tr.t:0 +

),L:ntr

rt,:L ,2,3,,..
I
^r:T
X"(x),
yieki eigenfunctions
Theseeigenvalues
X ^ ( x ) : A zrco s\n r : A n co sry ,,

n : L,2,3r ...

To find ?(t), we consideronly cases1 and 3. Why?

Partial Differential Equations

L64

Ca s e1 : k : 0
=> T(t) : C.

T' :0
The solutionfrom case1 is

u(x,t) : X(s)T(t) : B x C : D.
Case3: /r: -tr2 < 0
Since

k : -\2- -^7: - (T)' ,

the associatedfunction is
Tn $ ):

- Cnu-A ',
Cns-az\?ot

n:

L r2 r3 r. . . .

Comhine X" and ?lo to give


un(r,t) :
:

Xo@)Tn$):,4n cos
T"^e-#t
ancos"T

an : A n X Cn

"-4#',

n : ! r2 r3 r. . . .

Series solution
Sofar, the solution consistsof solutionsfrom casesL and 3. By the principle
of superposition,
u(x,t)

3)
u1(case1) * u3(ca'se

: D*i n= L un(n,t)

: D*i

oo

n=L

oncos
T"-49'

For convenience(for obvious rea,sonlater), lets put D : ao/Z. Then u(r, t)


becomes
oo

u(a,t): ?. i @n
cos
ry"-ryt
fL=

'

Initial condition

1r.tr1

5.3 The Heat Elquation

155

Applying the initial condition (b.10), gives

: f(r) :
u(r,o)

\ +in :l r" "o"T.


u

The problem is thus reduced to determining an expansionfor


/(r) of the
form

oo

f(r):3z* I

o n"o"ff, o< x < L


n*-=t

a Fourier cosine series. Thus


2fL

"o: i Jo f @)dn

o*:?

d ,* ,
lo " ro l"o "n n* n: r , 2, . . .
Final solution

Finally, substitute as and o' back into the u(x,t\ in (b.11) to give the solution for the wholeproblem (b.8)-(b.10).
Note: In firture, for the caseof iusulated endpoints, we just needto consider case 1 and case3 only.

Example 1L6 use the aboae solutionformula to obtain the solution


for
the IBVP
ou
,o2u
0<s11,
t>0
d:o'W;
ur(O,t) :

u'(1,f):g,

u(n,0) :

fr, 0<s<1.

Ans:z(r,t):;.

t>0

3 2#"-,"2n2o21cosn*

156

Partial Differential Equations

I
o.9
o.g
o.7
o.c

Fo.o
o.1
o.3
o.2
o.l
o

Figure 5.5: Temperature distribution of example 1L6.


Example tt7

Find the solution u(n,t\ for the IVBP


0u
,02u
E :a"*-z;
ur(O,t)

u(r,O) :

0<c<1,

ur(L,t) :0,
x(l-n),

o.5

t>0
,> 0

0 < r< 1 .

0.

o.7

0.4

o.9

Figure 5.6: Temperature distribution of ocample LL7.


Example LLB Solaethe IBVP

#= #,

oc o<1 ,t> o

5.3 The Heat Equatiorr

L57

u r(O,t) :

ur(l,t):S ,

u (arO) :

3-2cos4na,

t> 0
0<s<1.

Ans: u(c, t) : 3 - 2u-r,tt2t cosgtrs.

Figure 5.7: Temperaturedistribution of sxample 118.

Partial Differential Equations


Quiz 5.3
1. Solvethe IBVP
0u

02u

t>0

0<o < 1 '

A --t'
ur(0,t) :

u'(1,1): 6 ,

u(c,O) :

( o, o < r < o . b
I ;
0.5<o<1.

t>0

lofis

e'osn
- *'T]) "-n2r2t
*
Q+"sinff #ift-tl"

u(o,a: i.E
Ans:
2. Solvethe IBVP
0u

02u

o<o<1,t>o

fr:# ,
ur(Drt'1 :

zr(l,t): 0 ,

u(r.r}) :

fi2, 0 < r< 1 .

,>0

Q"-n2n""ornor.
A ns :u(o,r):1* 4 i
3' 7 P 3 n 2
3. Find the solutioirof the heat conductionproblem
0u

L02u

fr:;# ,
ur(0,t) :

u(r,o):

0<r < 40, t > 0

us(L,t): 9 ,

#,

t>0

o(u <40

and determine the steady-statetemperature (i.e, as t + oo).

. 200 g -lp1]1l3t?srln}
Ans:u(u,t):
-"- 40
"o"#
"-n2tr2t1*4oo
9 *_a_Te
Steady-statetemperature: 200/9.

5.3 The Heat Equation


5.3.3

Rod with mixed boundary

conditions

Try to derive the solution formula for the following IBVP using similar
approachas in previous sections. This time the right endpoint of the rod
is insulated while the right endpoint is fixed at zero temperature.
0u

,02u

fr:a'/r2,0<r<L,

r>0

( 5 .1 2 )

t > 0

(5.13)

Boundary conditions:
u(0,t) : $,

ur(L,t) :9,

Initial conditions:

u (t,O):f(r),

A <x <L.

(5.14)

Example Ll9 Referring to quations 5.12 - 5.11,


(a) show that the funilamental solutions sati'sfyingthe boundaryconditi,ons
are giuen by

un(o,g):

o2(2n_

gn"---,#&sin

l2efr)e.

(b) f,nd a tornral seriesexpansionfor the temperatureu(a,t) that satisfies


the initial condition.
( c ) s upp o se tha ta2:L, L:30,
the temperatureu(u, t).

f o r0 < x 1 -3 0 .F in d

and I@):30 -n

Solution
Answer to part (c),
r'o
Ans-Qn-r)zr2tl36oo
u(x,t): i
n=L
where
An:

(2n - 1)zro
60

l2}f2cosntr*(2n-1)"]
(2n - l)znz

160

Partial Differential Equations

Elxample L2O Show that the solution lor the following IBV?
u1(art) :

urr(rrt),

0<x<1,t>0

u(O,t) :

0,

us(L,t) : Q,

u(a,0) :

1,

0<o<1

,>0

i,sgiaen by

u(a,t):t

n=l

(2n - l)rs

_(zn-r\2t2t

Wsitr--e-----7-

I
o.c
o.
o.7
o.6
go.E
o.4
o.3
o.2
o.1
o

o.1

0.2

o .7

0 .9

Figure 5.8: Temperature distribution of exa^rnple


1.20.
The following problem is an exa,mpleof a left errdpoint being insulate.l
while the right endpoiut is fixed at zero temperature. what behaviour can
we observefrom the picture in examples(120) and (121) respectively?
Example t2L Solae the IBVp

0u 02u
A t:A r2 ;0 (c<1 ,

t>0

ur(Drt)

u (l ,t):9 ,

t>0

u(x,0)

a (L -x),

0< o< 1.

161

5.3 The Heat Equation

o.26

o.?

o.t 6

o.1

o.

o.4

0.E

o.z

o.

distributionof example121.
Figure5.9: Tlemperature
Quiz 5.4
L. Solvethe followingIBVP
u 1 @rt) :

0 <o < 1' ' > o


t>0
z(l ,t) : g,

u " (a ,t),

u*(0 ,ti -'0 ,


u (x'D ) : 4 ' 0 (s<L
(-t)_+
z" _ i

.' 16S
Ans: u,',\n ,t1 :;k

cos(2n -l)nt .
u-(2n-L)2#r/4

2. Solvethe IBVP
7o-t, Eu

t> o

u:a*r'o<c<1'
u (C ,t) :
u (r'0 ) :

u " (l ,t)=9,
-3 ,

t> 0

0 (c<1.

sin(2n-l)rt .

Ans:u(r,D: -+
">^#"-(2n-r)ztr2t/a
3. Show that the solution for the following IBVP
u1(t't) =

u"(x't\'

u'(0,t) :

0'

u(L'f):g'

u(r'O) =

fr'

0<r<1

0<o<L't>0
t>0

L62

Partial Differential
is given by

u(x,t):i l,n"-@4&*"Y
n:l

^
--"

A(-l'Y+t
(2n - l'1tr

8
(2n - l\znz'

Equations

163

5.4 The'Wave Equation

5.4

The Wave Equation

Consider the one-dimensional wave equation


Q2 l tra a r

Ittt:

0 1t

< L,

t ) 0,

(5.15)

describingthe motion of an elastic string with length .L where u(r,t) is the


vertical displacement(deflection) of the string. u(r,t) describesthe displacement at the point o and time t. Horizontal displacement is assumed
negligible.
We will have two initial conditions as the wave equation is of secondorder
in time. Usually one coudition is for initiat position of the points in the
string written as
0<n<L ,
u(o,O):!(r),
while the other is for initial'relocity given by
u1(n,0):9(x),

0<n< L .

Similarly, we will have two boundary conditions. As in the case of heat


equation, difierent conditions prescribed at endpoints gives difierent physical interpretation. We poseseveralcases.However,we will work thoroughly
only for one ca^sewhere both endpoints are fixed at zero displacement. For
the rest, you can take the challengeto solve them.

5.4.1

String with fixed endPoints

Here,both endpointsof the string arefixed. They are not allowedto move.
The modelgoverningthis problemis
02u

,02u

d:"'#,

o1r<L'

t>o

(5.16)

>0

(5.17)

0 1x < L.

(5.18)

with boundaryconditicns
u(0't) : g' u(L't) : g'

'

and initial conditions


) (t),
u ( r , 0 ): f (n ), u 1 (r,0:9

t64

Partial Differential

Equations

where /(c) and g(c) are given functions. We will solvethis model to find
out how the string vibrates.
Solution
Again, we begin with
u(n,t) : X(a\T(t).
Then

..

02rt.

fl%t

: x(0)7"(t\,
X(dT"(t):

1.T"(t)

: x't (n)T(t)

a2X"(x)T(t)

X"(*)

aconsta,nt

A-6:ffi:k,
Therefore

I Ttt(t\

X" (r)

orm:o,

S :o '

We now have two ODEs


T" -a2kT:0

and X" -k X : 0 .

Boundar;r conditions
Sepa^rate
the boundary conditions to give

Notethat if

u(0,t): X(0)"(r) : 0 +

x(0) : g

u(L,t): X(I)?(t) : 0 +

X(L) :0.

:0) u@3): 0. Thus


wehave
I 0. Hence
"(t)
"(t)
T" -az k r:o

x,, - kx :0, x(o): g, x(r) : g.


Case1: /c:0
X " :0
X (0 )-g

+
+

X (x):Ar *B
B :0

X:As

5.4 The'Wave Equation


X (L) :0

=+ 0: A L

=+

A:0

Thus X(c) : g.
C a s e 2z k:,\2 > 0

x" -

^2x:o
X( 0):0 +

x(a): Ae\'+ Be-\'


:+ B:-A

0:A +B

X( L ) :0 =+0 : A eL'- A e-! ' * A : 0 :

Thus X(o) : g.
Case 3: k:

-12 1
=+ X(a):

x" + \2X : o

X(I) : g

0:A

X( 0):0

Ccos.\c* Bsin,\s
X : B s in . \ o

=+ 0: B sin) . L ,

B +O'

Hence
sin)Z:0

\L:

nr

n : 7 , 2 ,3 ,...
^r:7,
Therefore

Xn@\:Bnsinry ,

n: !,2,3','"

Tiosolve?(t), considerthe 3 casesagain. For this particular problem, only


case3 is applicable.

k:-\2--^?*:-(T)'
Now we have

: o
: o =) T" + az\?nT
/' + a2\?^T
so that
T"(t) :

Cncosc\nt + Dnsino)r' t

crr,rt
mrt
^ rn.
Uncos
+rns
L
L.

166

Partial Differential

Equations

Combining X,, and ?r, giving


u"(a,t)

Xn(a)To(t)

: B , , s in
i n]
r y [ " " " o Y! * D o sry

: siory[""""'
ry*Qosin
ry]

with

Pn=BoxCn

and

Q n=BnxD'.

Series solution

By the principle of superposition,

u(a,t) =

u,(a,t)

n=l

cnnt
,
: f$ siniL""o"ff*e4sinryl
^,_nra | ^
fnitia! conditions
Applying the first initial condition
yields

u(x,o)=_iP,,sin
T= f @l
n=1
+
a Fourier sine series. Thus

P*= 1
ffirt

f (s) =LS
n=l ""

,iono,
tJ

It talsinffax,n= 1,2,...

theotherinitiarcondition,difierentiat
e u(x,r) with respectto f,

u1(a,t)
: I
*ry].
"' T [-"ry "ioy!+A,ry
@

L67

5.4 The'Wave Equation

u1(r,o):
i "i'ff0*ff: s@)
L

'L=

+ g( x ) : i, ioT O " T
n=l

sin
a-ry : ? [^"s(c)
To'
tJ J0

tJ

.t

Qn:

*Jr

rL

e(c)sinTo,

Final solution"
Substitute P,, and 8" back into the seriessolution u(n,t),
Note: In future, we can ignorecase1 (k : 0) and case2 (k > 0) whenever
we have fixed endpoints.

Example t22 Soiuethe IBVP


7trtt:Itrsa, 0<c<L,
u(0,t) : g, u(l,t) -- g,
u (r,0) : 1, u1(r,0):$,

t>0
t> 0
0< s < 1 .

sin(2n- 1)zro'cos(2n- l)a't.


Ans:
_
_ _ _z(o,
_, ,\_il:_ ,_lf/ L rr
2n_L
Example L23 Solaethe IBVP

r >0
t> 0
z(0,t) : g, u(L,t) : g,
u(4,0) : lQ, u1(tr0) : L, 0< s< L.
I t r tt:' ttra a t

0 (r(1 '

400

q:3D
Ans:
u(o,
T'n=L

sin(2n- l)rx. cos(2n- L)rt

(2n- r)z

Partial Differential

168
Example

Equations

L24 Solue the IBVP


utrtt:'ltraat 0<O<1''

t>0

t>0
u (O,t):6 , u (L 't):0 ,
: x, 0 ( s< 1' .
u (r,0 ): l , u s(u ,Q)

Ans:u(o,t\:2f

rio",or(#

e'snnt.#sio",ot)

5.4 The'Wave Equation

169

Quiz 5.5
l. Show that the half range sine seriesfor
(n

I ro'
I\n): 1 zo_,
,O '

o(r(10
10( r ( 20.

is representedby
,/\
I\n):

aS r
n r. n f ir
p )_ -5-srnl-sln 20 .
"
n= l' "

2. Consider the waveequation


02u

02u

0<r<2 0 ,

ffi :p,

t>0

wiih bounda^ryconditions
u(0, t) : g and u(20,t) : Q
and initial conditions
u(r,0)

u1(r,O) :

f (")

(as in pa,rt (1))

0'

Use the method of separationof variablesto shcw that the solution


to the given problem is
u(c,;"
c):

gS t

tr2k;,

flT-' --n n x '

n rt

sm sm
kos;o-'
lld-

Partial Differential

t70

5.4.2

Equations

String with free endpoints

The physical phenomenaseemawkward. However, mathematically we still


can find the solution. The boundary conditions for this problem a,reanalogor$ to that of the insulated endpointsfor the heat equation. Hence,the
model is governedby
o2u

oo2u

o<r1L, ,>o

d:o'ffi,

(5.19)

with boundary conditions


u'(O,t):Q,

(5 ' 2 0 )

ur(L,t):0 ,
'>0

and initial conditions


u( r ,O) : l@ ),

u1(a,0):9(r ),

0 < r< L .

(5 . 2 1 )

If you like to challengeyourself, why not try to work on that problem. You
should be able to find the solution formula as given below:

a1t*i*, ryI""*ff*Q,,sin
u(x,t):ao*
ryl
L

n=l

where
LfL

"o:iJo f@)dn
1fL

dr
"t: i Jo s@)
:1,2,...
," : ?rf:/(r) cos
T 0,, n
a,: * lo " sAl"o"ff d, rn:
, 1, 2, . . .
Feeling good? Yes. You can do it! Why not continue working on the
following IBVP. A simple one.

5.4 The'Wave Equation

17t

Example t26 Soluethe IBVP


02u 02u
ffi:6rz'

0(r(1'

t>0

with bounilary cond,itions


ut(O,t):0,

u'(L,t) :6,

t > 0

and initial unditions


u(arD): coszlrs, u1(rr0) : -2rcosrsr

0 < c < 1.

Ans: u(r, f) : -2 cosrx sinnt * cos2tra sin2rt.

Elxample t26 Solaethe IBVP


02u 02u
6;;z:6rz'

0<o<1,

t>0

with boundaryconilitions
ur(O,t):0,

ur(L,t) : g,

t > 0

and initial mnditions

u(a,0): -}cos2rc, u1(xr}):2s - l.

Ans: u(r,t) : -l cosbtrrcos2rt*i

n=L

[(-1)' - L]#

cosnro sinnrt

L72

Partial Differential

Equations

Quiz 5.6 Solvethe following IBVP.

o< o < 1 't> o

#: #'
with boundary conditions
u'(O,t) :0,

ut(l,t) :0,

t> 0

and initial conditions


u(o,0) : $,

( -t,
g(c):

I o,

u1(u,A): g(r\

I.r<1
otherwise.

sinnrt.
Ans:
u(o,t'):+
fi"i'ff"o"ff*"nn*
E

L73

5.4 The Wave Equation

5.4.3

String with mixed boundary conditions

We could haveoneendpointfixed (no displacement)while the other is set


free. Here is one exa,mple.
02u
,02u
@:u'#,

t>0

(5 . 2 2 )

t >0

(5.23)

0101L ,

with boundary conditions


u(0,f) : 9, ur(L,t) :9,
and initial conditions
u ( r ,0) : f(n),

u1(r,0):9(n),

0<r

<L.

(5 . 2 4 )

Example t27 Referring to equations5.22 - 5.24,


(a) show that the funilamental solutionssatisfuing the boundaryconditions
are gi,uenby

)r,a,
un(nrY):/r, sin)rrocos
^,

ry'

(b) fi,nd a tormal series expansionfor the temperatureu(r,t) that satisfies


the ini,ti,alconditions.
(c) supposethe string is set i,n motion with no initial uelocity,4(r,0) : Q,
lrom the initial positionu(x,O) : l@) where

/( ") :

( U

* -t.nc l +1, ( L > 2 )

{
[ 0,

othertai,se.

u(r,t).
Find the displacement
Solution Left as an exercise.
Answer to part (c),

sin
u(r,t):iA,,
ry
n=I

where

.
A":

"o"w

_,- (2n- 1),rsmlz-:.


_,. (2n- 1)r

ffism-a:

Partial Differential

L74

5.5

Equations

The Laplace Equation

In the previous two sections,we have dealt with transient problems. The
solutions are normally died away, leaving steady state (equilibrium). The
PDE describingthe two-dimensionalheat equation in steady state is the
Laplace equation
ltrxiu^g :0,

O 1a 1 a ,

0<Y<b'

Problems of this type are called boundary value probtems (BVPs) since
there is no time variable involved. And so, there are no initial conditions.
Only the boundary conditions specify the solution.
Like the heat and wave equations, there axe many possibilities for the
boundary conditions. However, in this section we will be dealing with
ouly one type of bounda^ryconditions where the temperature is prescribed
at the boundary. Since there are 4 boundaries,we will heve 4 boundary
conditions. We will fix zero temperature at 3 of the boundarieswhile the
fourth having temperature defined by a function of I or a function of 3t
alone. Exa,mplesare given next.
5.5.1

Prescribed temperature

at the boundaries - type A

By this we mean the nonzero temperature is a function of o alone. One


exarnple is as urodelled by the following BVP.

u'u **:0 ,
W
oy'
u(o,0):-f(s),
u (o ,g ):9 ,

o <r< a,

o<y< b

u(t,b) : Q,
u (a rY ):0,

0< x < a
0<Ycb'

Solutiun
We seeka solution of the form
u (s,v): X (n )Y( y) .
Then

x@)Y"(y).
#: x@)"Y(y),
#:

L75

5.5 The Laplace Equation


Substituting these into the PDE, gives
Xt'(a)Y(y) + x(n)Y't(u) :0
and dividing through by X(a)Y(y), we obtain

: u, a constant.

: -"#

(b.2b)

We nonr have two ODEs


Xtt - kX : 0 and Yt' + lcY :0.

(5.26)

Note: The choiceof k here is crucial. We omit the caseof Ic :0 because


it gives trivial solution. We can take only one possible value of k, either
negativeor positive. Why? Sincethe nonzerotemperature is a function of
fi, we want that tc be representedby a Fourier series, For this reiuton,we
take k to be negativeby letting k : -\2.
As a result, the ODEs in (5.26) becomes
x" + \2 X :0

:0.

and Y" -

(5.27)

^2Y
What is another alternative to equation (5.25) if you want the sameeq'rations as in (5.27)?
Next, we tackle the three zero boundary conditions.
Zero boundary conditions
Sepa^ratethe three boundary conditions to give
u(u,b): X(O)Y(b): 0

r(b) : 6

u(0,s): X (O)Y (g):0

=+ X (0 ) : 0

u(a,A): X(a)Y(Y) :0

=+ X(o) : g.

Thus, we obtain two sets of ODEs

x,, +,\2x :0,

x(o) - g, x(o) : g

ytt - 72Y :0,

Y (b ) : 6 .

Partial Differential Equations

t76

We will solve for X first. The solution is


X (o):A cos)c*B sin, \ r.
Then apply the two boundary conditions.
X( 0) :9

=+

X( o) : g

A :0

=+

X (x ):

B s in )r'

B +0

B sin.la:0,

wh Y ?

Hence
sin.\a:0
\n-nn

\a:nn

n:1r2r3 r. . .

The corresponding eigenfunctions are

X^(n):B' sinry ,

n: L,2,3,
"'

Now,lets solve
y, - s2Y = 0, Y(b):9.
For convenience,we don't replace the .\r, yet.
Y :0.
Recall: ConsiderYt' ^2 is 12 - \2 r :
The characteristic equd,tion
r : *). The solution can be written as:
Y(il

0, which upon solving grves

a1"^Y+a2e-^v

fu cosh\y +bzsinh,\gr

csinh)(g+d ).

It is convenient to take the third form for the solution of our Y. Thus,

Y(s) -- C sinhXs+ d)
where { is a constant to be determined. Applying boundary condition
f(b) :0 gives
C sin h .\( b * 6) :0

sinh.\(b+{):0

(wh v Cl0 ? )

I (b + 4 1 -g

L77

5.5 The Laplace Equation


Thus { - -5. Updating Y gives

Y(il:
n:!,2,3,

Z:::W,

Then we combine X" and Y' to give the eigenfunctions


un(s,y) :
:

XnYn: Bnsin
D,nsinW

na-(y-'!)'
T

"rrio6

si*M,

n:!,2,3,....

where Dn - Bn x Cn.
Series solution
By the principle of superposition,

u(r,t\ : i

u,(n,t)

n=L

," si,.W.io6

tt(Y {)-.

n=L

Nonzero boundary condition


To satisfythe remainingboundaryconditionu(r,0) : t@),0 < s < a, we
should have

u(c,o): I@) :

r*in9ar66

i'(-b)

n=l

6-L

D -o*"ioyW "inY2: t bnsinry,


n=l
n=I

a Fourier sine series.


By the Fourier seriestechnique,
b n:

- D n"iohnnb :? f/(r)sin
alo'
a

giving

Dn: -

,.

fo

\ a as n : L,2,...

: 1 ; . ; r l - t t * l " i o naT ^d',, n :L ,2 ,....


osinhff Jo

L78

Partial Differential Equatiois


Final solution

Finally, substitute Dn back into the seriessolution to obtain


the final solution.

Example !28 Show that the solution


for the BVp
u a a*usa:0,

0<a<1,

u(rrD) : a, u(nr2) : g,
u(0,y) :9,
u(l,y\ -0,

0<y<2
0< o< L
0 <y <2.

is

u(a'a): : i
7t
-

nsr
#

sinntrs
sinhnn(v
- 2).

5.5 The Laplace Equation

L79

Quiz 5.7 Obtain the solution formula for the BVP


u aa*uyy:0,

01.x<a,

u(a,O): Q, u(r,b) : l@),


u( 0,9):6, u(a,g):0,

0<y<b
0< n< a
0<y<b.

and usethe result to solve


IJrn*uuy:0,
u(c,O):Q,
u(0,y) :9,

0<r(1,

0<A<2

u(a,2):s,
u(!.,y): g,

0<c<1
0 < y <2.

t;t]"=.t sinntrasinhnnu.
Arrr, ? i
-----'-"v'
n ?-r nsinh2nn

Partial Differential Equations

180
5.5.2

Prescribed

temperature

at the boundaries - type B

By this, we mean the nonzero temperature is a function of y alone. One


exa,mpleis as modelled by the following BVP.
u ru**:0,
M
oy-

o<y<b

o1n<a,

u(c,0) : Q, u(t,b) : Q,
u(0,9r): g, u(a,U): 9(A),

0 1a < a
O < A <b.

Solution
The solution is obtained in a similar manner as in the previoussection. We
seeka solution of the form
u(a,t\ : X(a)T(t).
Then
02t,

#:

A2,,t

x(x)Y"(il.

x(n)"v(y),
#:

Substituting these into the PDE, gives

x" (r)Y (y) * X (r)Y't(e): o


and dividing through by X(x)Y(g), we obtain

ry:W
x(*)

Y(y)

- 7r, a constant.

(b.2g)

We now have two ODEs


X" +'lcX :0
Again, we choosek:

and Y" - lsY :0.

(5.29)

-\2. Why? As a result, the ODEs in (5.29)becomes

x" -.\2 x : o and Y" + \2Y :0.


Zero boundary conditions
Separatethe three boundary conditions to give
u(r,0) : X (s)Y (0) :0

Y (0 ) : g

(5.30)

5.5 The Laplace Equation

181

u ( x,b): X (r)Y (b) :0

r(b ) : g

u ( O,Y ):X (O)Y (Y ):0

X (0 ): s .

Thus, we obtain two sets of ODEs


tt-s2x:0,
y" + \2 )i:

x(o) : g
y(o) : g,

0,

y(b) : 9.

We will solve for Y first. The solution is


Y(g) :.Acos)g * Bsin)y.
Then appiy the two boundary conditions.
f( 0) =0

Y( b ) :0

=+

A :0

Y (y ): B s in )g r.

B sin)b:0,

B*A

wh y ?

Hence
sin.\b:0
,
:
,\"

n7l

\b:n t r

n: L,2 , 3 r. .

i: )

The corresponding eigenfunctions axe

: Br,sin ,
Yn(y)
ry

n:1,2,3,.-.

Now, lets solvc


X t'-^2X =0,

X (0): 0 .

For convenience,we don't replacethe .\r, yet.


Thus,
X (a):Csinh)(c*/).
Applyrng boundary condition X(0) :0 gives
C s i n h,\/:Q

:+

sinh.\{:g

(whyC+07 )

}d:0

Updating X gives
X(t)

C sinh'\o

C"nb s i nhff,

n: 1 , 2 , 3 , . . . ,

6:0.

Partial Differential Equations

L82

and Yn to give the eigenfunctions

Then we combine &

un(x'
v)
: :w"_'^fq "'*n*:,2,
3,.. .
whereDn:

Bn x Cn.
Series solution

By the principle of superposition,

u(r,t) :

u^(a,t)

n=I
oo

: D o,sinff sinnff.
n=I
Nonzero boundary condition
To satisfy the remaining boundary condition u(a,U) : g(U), 0 < y ( b, we
must have

u(a,,a):
sfu): i ,, sinffsinhff
rl:

: ir"sinhff sinT:i

bnsinff,

"Z'ooa Fourier sine series.


By the Fourier series technique,
bo : Dn"it'

nrto

"hT

giving

Dn:
'- #
b

:; 2

fb ' \

l"s(y)sinf f at

n: 1, 2, . . .

rb

n:1,2,....
b
smtrjs/ol- s@)sinffay
Final solution

Finallg substitute D, back into the seriessolution to obtain the final so.
lution.

183

5.5 The Laplace Equation


Example L29 Usethe aboueresult to solue
u ao*uug:0,

0<n<2,

u(c,0) : Q, u(ar2) - 0,
u(0,Y) : 0, u(2''d : 9(2 - Y)''

o<y< 2
0<u< 2
o <y <2.

Partial Differential

184

Quiz 5.8 Derive a solution formula for the BVP


Uaa*uyA :0,

0<n<a,

a(o,0):9,
u(u,b):9,
u(O,d : g(U\, u(a,y) : 0,

O<Y<b
0 < r< a
0 < y < b.

and use the result to solve the problem


uan*uvs:0,

01r<2,

u( o,0):9,
u( 0,il :U(2-A \,

u(t,2):9,
u(2,y) :9,

0<Y<2
0 < r< 2
0 <y <2.

Equations

Bibliography
[1] Alan Jefirey Ailaanceil Engi'neeringMathematics, Academic Press;
2002.
[2] Boyce W.E.and DiPrima R.C.Elementary Differcntial Equaticns and
Boundary ValueProblents,8th Edition, John Wiley and Sons,2004.
[3] Glynn Jamcs et al. AduancedModem Engineeri'ngMathematics,Addison Wesley,1993.
[4] Kohler W. and Johnson L. Elementary differential Equations with
Bounilarg ValueProblenus,SecondEdition, 2005.
[5] Nagle, Saff and Snider, h.rndamentalsOt Differential Equations fth
Edition Addison WesleyLongman, 2000.
[6] Stroud K.A.AilaanceilEngineeringMathematics MacMillan Ltd., London; 1996.

LISTS OF FORMULAE

Trigonometric
cos2o+sinzc:1

sl nno:

1+tan2 0:seczt

,
=
cosno

cotzg*1=cogec2r
sin2r = 2sinrcosc

tan2n= T=ffiG

sin(r * g) = sin r cos3l* cosr sin Y


cos(ot g) : cosrcos gia sincsinY
ianr * ta'nY
'
tan(t
2sinr cos! = sin(r * g) + sin(c - Y)
2sincsin gr: - cos(o+ Y) + cos(o- Y)
2 cosncos! : cos(c* Y) * cos(o- Y)

Logarithm
Ox "a ln o

lo96r
.
logo o =
I;E;;

n_

cosh2Jt-sinh2c:1
1-tanh2n=sechzn

l:rls2r=cos2r-sinzn
:2cos2 n - L
=1-2sin2r
2tans

e' - e-'
2
er* e-,

'

cot}2r-1=cosech2r
sinh2c = 2sinhrcoshs
cosh2c=cosh2c+sinh2c
=2cosh?o-1
- 1* 2sinh2c

tanh2r= '2t42 0
I + tanhsinh(c * g) = sinh c coshy t coshr sinh y
cosh(o* y) = coshccosh3t* sinhc sinhy
tanhc * tanhY

tanh(cLY): G t*ffiffit

Inverse Hiperbolic

-oo ( 0 < oo
sinh-l c = ln(s + tffi1,
x > I'
cosh-lc : ln(s I 1Pa),

ra nh- r , = t "( i#) , - 1< r < 1

Source:Department of Mathematics, UTM'

Integrations

Differentiations

=
constant
ftWl o, lc
*rO,

= nrn-r

=|
Su"rn

I r " *= # + c ,n l -t
I *:hlcl+c

fiV*a: -sinr
ftbi"a: coso
n
sec2

S'rt*rrl:

I u*:kx*c

!*[,,*tl='-cosec2o

sinndn=-cosc* C

I cnsndr: sina * C
J
| """'

nd,a = tann * C

co"""',d,,= - cotx* C
+C
s""rtuord,r= seca

ftb..rl= secotano

fibo"u"rl= -cosecrcotc

;["".".t.ot

=""
frb"l

1" " *=e'+C

St"o.rrtl
$f

*'f

lffr*o'f

= sinhc
: cosh'

=sech2s

I
I
I

sinhnd,n= coshr* C
d,r= sinho* C
cosnn
tanho* C
sech2xd,r=

ilcothcl = -cosedr'c
d,t'

$ttu*tt

fr["or""t

= -sechctanhc

| "o""Af
/

xds = -cosecc * C

,ar= - cothc* C

sectrrtanhx.Tr = -sechz * C

= -cosechc
* C
t "or""nr"othad,s

tl = -cosechccothc

Source: Department of Mathematics, UTM.

Differentiations of
Inverse Ftrnctions

#t""-'
"t:f,#.ff,W.r.
fr!"o"-'"t:
f,# .ft,d.,
d ,.

-l- u'1 : f f i .L8 .

du

Integrations
in fnverse
dfr

.-1

Resulting
Functions
,

J f fi :s i n -r(I)* " .
fda
,-J azqs2-

j**-'G)*
".

f iFa n
d ,.-r

fiLcot

'u1:

du

-1

fifi.8.

flb""-'ut:
WJ#=.ft,urt

d=ffi_ r.fl, wr r.
fr\"o"""-'
d,.,-r

ftsmn

-uJ= 1
JFV.

du

St," " o-'4 = fi .fi ,u.t.


-',r]= -1
f [ru"t
u{L -

-1

f["os".t-lrJ=

wrt.
du

uz

(;) . c, a>0.

I &:cosh-l

(;) . c, x)0.

fd ,

tm

r.
f t"*o-'
"t=#r.fi, i,utr

fl["oto-,q=
*.ft,

I #:sinh-1

'6,

:!

(1)*", u>o,
:tanh-l

t j.orn-'(1)*", t t>o.

Iffi--!su"h-, (:)::,.,

0<u<1.

'#'uro'

I#--1"o,""h-'L;1::,,

Source: Department of Mathematics, UTM.

r(s) :

Laplace Transforms

/-

7(t)e-"tat

r(s)
k

t n, n= t r 2r 3r . ..

nl

eot

8-&
5

kosot

s2+a2
&

sin oJ

wF
I

koshat

s 2 -a 2
&

sinhaJ

F=7

e"tf(t)

F(s - a)

t" f (t), n:1,2,3,.. .

/\ - 1\-/n*

rr'-

ltLl

#,

n=1r2r3,.,,

intr
dsn

s'F(s) - s'-l.f (0) - s"-2 f '(0)


e-o"

H(t - a)

f (t - a\H(t - a)

e-o"F(s)

0\t - a)

e-o"

f(t)6(t - a)

e-"'f(a)

7t

I f (")sQ- u)auJO
fL

I f(u)du
JO

r(s)G(s)
F(")
s
trT

l$ ) : l (t+r)

'-, _e

| e-" y 1 t 1 a t

-" Jo

ConvolurionTheorem L-L {F(s)G(")} = [" f @)n$ - u) d,u


JO

Source: Department of Mathematics, UTM.

J('?-1)(0)

Das könnte Ihnen auch gefallen