WT 2 SECOND

WITH CONS
COEFFICIENTS
Structure
2.1 Introduction
Objectives

2.2 Some Terminology

2.3 Homogeneous Linear Equations With Constant Coefficients
2.4 Nonhomogeneous Linear Equations With Constant Coefficients
The Method of UndeterminedMultipliers

The Method of Variation of Parameters

2.5 Summary
2.6 Terminal Questions
2.7 Solutions and Answers

2.1 INTRODUCTION
In Unit 1 you have learnt to solve first order ordinary differential equations. These
equatioils provide a useful means of studying physical systems. For instance,
radioactive decay, free fall of a body, fluid flow, current growth in electrical
circuits, etc could be studied using these equations. But many systems require
solutions of ordinary differential equations of order higher than one. From Unit 1 of
PHE-02 course on Oscillations and Waves, you would recall that the equation of
motion of an undamped harmonic oscillator is a homogeheous second order differential
equation with constant coefficients. Similarly, to determine the depression in
horizontal beams,.we have to solve a second order differential equation with
constant coefficients. You will learn to solve such equations in Sec. 2.3. But can
you study time-variation of charge in maintained RLC circuits or the phenomenon of
resonance using a homogeneous second order differential equation? In such cases we
have to solve nonhomogeneous second order differential equations with constant
coefficients. But even this is not true in general. For instance, when we wish to
study field distribution around a charged sphere or a cylinder, we have to solve
second order differential equations with variable coefficients. Similar situations are also
'
encountered in heat, optics, electromagnetic theory, energy production in a nuclear
reactor, quantum mechanics, etc. In such cases we seek power series solutions or use.
Frobenills method. In the next unit you will learn these methods. But in this unit, we
have discussed the basic techniques of solving second order differential equations
with constant coeffidients. Some of their applications will be discussed in Unit.4.
d -

Objectives
After studying this unit you should be able to

* coinyute the Wronskian of a given ODE

obtain linearly independeit solutions of homogeneous second order ordinary differential

equations with constant coefficients .
useihe method of undetermined coefficients and variation of parametea to obtain
linearly independent solutions of non-homogeneoussecond order ordinary differential
equations with consthnt coefficients.
 Second Order O d i a r s ,
2.2 SOME TE DillFerential Equations witb
Constant CoeffEcients

While studying first order ordinary differential equations (ODES)in U N 1~ you have learnt
some basic terminology. You would come across some common terms, which you do not
know as yet, in the context of second order differential equations as well. It is important for
you to be familiar with them. This section is intended for this purpose.
You know that a second order linear ordinary differential equation can be written as

The function g(x) is termed as the forcingfunction, andpl(x) and po(x) are coefficient
functions. These are continuous over the interval where the solution exists.
Linearly Independent Solutions and the Wronskian
From Unit 1 (Sec. 1.4.3) we iecall that if yl and y2 are linearly independent solutions of the
homogeneous equation
Y" +P~(x)Y'+PO(~)Y
= 0 (2.2)

I then their linear combination

Y =C1~1+CzYz (2.3)
where C1 and C2 are arbilrary constants, is a general solution of Eq. (2.2). For example,
you know that y l = sin wt and yz = cos wt are linearly independent solutions of the ODE
2
for anxndamped harmonic oscillator: g+ 2 0 2 y = 0. S o the general solution of this
dt

I y(t) = C1sin ot + C2cos wt

You may ask: What do we mean by linearly independent solutions? How do we test linear
independence? Will a linear combination of linearly independent solutions necessarily lead
to a different solution? When does'a set of solutions constitute the general solution of a
linear differential equation? and so on. Let us now discover answers to these questions for
ODES of second order. We say that two solutions yl and yz are linearly independent on an
interval if the identity

is satisfied only when C,= C2= 0.For, if C1 and C2 were non-zero constants, Eq. (2.4)
would yield y2/yl = constant, is., yl aird yz would be proportional on sorns interval. Then,
by definition, yl and yz would be linearly depende~tfuiictions on thal interval. In other
words, linear independence of fl and yz means that the ratiodyl is not a constant. This implies
that the differential of this ratio

is not identically equal to zero. Therefore, we can write the co~iditio~i

of linear indepeadence
of two solutions yl and ~2 as

The determinant WCyl ,y,) is called the Wronski, determinant or the Wronskian of the
given differential equation. We may, therefore, conclude that

, , Two solutions y, and y, are linearly independent on an interval [a , b 1, if and only if,
. their Wronskian is non-zero for a s x s b.
I .
.$

For a harmonic oscillator, fbis means tbat

 1
Mathemstid Methods in
-
Physics IT sin wt cos o t
" -w
W(x) =
wcosot -osinot
showing that sin wt and cos o t are linearly independent. We also say that yl and y2are
linearly dependent solutions on an interval I, if and only if their Wronskian is zero for some-
x = xoinZ
-
Spend SAQ 1
5 min
The solutions of the equation

are given by y,' = sin 2x and y2 = cos 2x. Are these solutioi~slinearly independent?

Particular integral and complementary function

From Sec. 4.2 of the course Oscillations and Waves (PHE-02), you would recall that the
equation of motion of a forced damped harmonic oscillator is
my" + yyf + k y = FQ cos ot
which is usually rewritten as
y ff

+ 2by1+o8y = fOcosor
where 26 = y/m, wi = k/m andfo = ~ ~ / m .
Physically, the actual motion of this system is a sum of two oscillations: one of the
frequency of damped oscillations and the other of the frequency of the driving force.
Mathematically,we express it as

where yl is a solution of the homogeneous equation

yif + 2byi + m8yl a 0 (2.9)
On substituting Eq, (2.8) in Eq.(2.7) and using Eq.(2.9) in the resultant expression, you
will find that yz satisfies the equation
y; + 2byi + wi y2 - fo cos o t
In the language of mathematics, );I is called the complementary function and y2 is called
the particular integral. We can write the general solution of a second order non-
homogeneous linear differential equation with constant coefficients as the sum of a
complementary functionand the particular integral:

You know that the solution of a second order differential equation consists of only two atbitrary
constants. This implies that the particular integral will. not contain any arbitrary constant .
We hope that you are now dquipped with all the necessary basic terminology. Let us now
proceed to solve h6mogeneous linear ODESof second order with constant coefficients.

2.3 HOMOGENEOUS LINEAR EQUATIONS

WITH CONSTANT COEFFICIEBTS
A second order homogeneous linear ordinary differential equ.ation with constant coefficients
fmds wide applications in engineering, biological and physical systems. In particular, you
know its use in mechanical and electricaI vibrations. Many techniques of solving such .
equations have been developed. These include the reduction of order technique, which you
learnt iq Uiit 1, and the method of exponential functions, which we will discuss now.
A,homogeneous second order
'
ordinary differential equatiqrb@ith'constantcoefficients'can
be expressed in the form
 Second Order Odjaary,. '
ayl' + by1 + cy = 0 . (2.11) DifPerential Equations with-
Constant Coefflcients
where a, b and c are real constants.
From Unit 1, you would recall that the solution of the first order homogeneouslinear ordinary From Eq. (2.11), you wouldno~e
differential equation 01' + y = 0 ) is an exponential function of the form that y,y' and y" are linearly . ,

dependent. This demands that y

y = Aexp(-kx) be an exponential function. For,
the derivative of any order of an
Let us, therefore, seek a solution of Eq. (2.11) of the form exponential function is linearly
dependent with itself, i.e., some
y = A exp(mx) (2.12) multipliof the original
exponential.
where dimensions of m are inverse of those ofx. This ensures that the power of exponential
is dimensiollless,
Substituting this and its derivatives

and

in Eq. (2.11), you will obtain

(am 2 + bm + c)A exp(m x) = 0
SinceA exp( m x ) is finite, this equation will be satisfied only if
a m 2 + bm + c = 0 (2.13)

This quadratic equation is called the characteristic equation (or auxiliary equation). Its
roots are

-b +d c 4 ; - b - l K z c
ml = and m2 =
2a 2a
For example, the auxiliary equation for
y" + 5 y 1 - 7 y = O is m 2 + 5 m - 7 3 0
S o you will agree that
Y I(X) = A exp(m1 x) ' (2.14a)
and
Y2 (XI a= A exp(m2 x) (2.14b)
are solutions of Eq. (2.11). Using the principle of superposition, you can write its most
general solution as
y(x) = C1 exp(m1x) + C2exp(mzxf (2.14~)
for a suitable choice of constants ~1 and C2 determined by initial or boundary conditions.
The Wrol~skianof these solutions is
A exp(m1x ) A exp(m2x)
W(x) =
ml A exp(m1x ) m, A exp(m2x)
= (m2-ml)Bexp[(ml+ma)x] (2.15)
where B is a constant. This shows that for ml pc m2, the solutiol~swill be linearly independent:
You must have noticed that the process of solving a hoinogeneouslinear second order ordinary
differential equation with constait coefficients using an exponelltial fut~ctionas a solutioll
reduces to finding the mots of a quadratic equation. The roots of this equation can be
1. real and distinct for b2- 4ac > 0 or b 2 > 4ac
2. real a ~ equal
d when b 2 - 4ac = I-,0. or
-- ~- h 2= 4ac
. and
-
h h ~ m aM w dhdh .Let us now discover solutions corresponding to these roots.
-
P h y h II
Dbtlnct Real Roots
For distinct real roots, exp(ml x), and exp(m2x)are linearly independent and the general
solution is given by

Y = C, exp(m1 + (7, exp(m2x)

- up [-(g)] [c, e x ( a x ) + ~ . exp(-ad ] (2.16)

I f T = z C
where a =
2a

The constants C1and C2can be determined by using given initial and boundary conditions.
\ following example.
We now illustrate this method with the

Example 1 ,

Solve the equation

y" + 3y' + 2y - 0
subject to the initial conditions y( 0) = 1 and y'( 0) = 2.
Solution
In this case, thk auxiliary equation is

which has mots m = 1 and m = - - 2. Therefore, the general solution is

To determine C1 and Cz , we first use the condition that atx = 0, y = 1. This gives
1 - Cl +C2 (ii)
Further, since

we find that

You a n readily solve (ii) and (iii) for C1and Cz to obtain C1 = 4 and Cz = - 3. Hence, the
desired particular solution h

I
When two roots are equal (ml I mz), W k )= 0. This means that emlxand px are linearly
dependent. What docs this imply? It implies that (i) Eq. (2.14) does not hold and (ii) our,
starting assumpdon is false. You may now ask: How can we obtain two linearly independent ,
functions when auxiliary equation of a second order differential equation has two equal \

roots? In such a situation, we use the method of reduction of order to construct a second
linearly independent solution. This is illustrated helow.
Repeated Peal Roots
When a secdnd order differketial equation has two equal roots, we obtain the correct form
of the secord solution by assumiirg that' .

.where m id a root ofthe auxiliary equation (Eq. (2.13)). Differentsting Eq. (2.1'7) with .''-

respect?&, we get
 Secmd Order 0- I 1

Dillermtial EgumUoar with

Cwrinnt ~ o e ~ n b I
and

Substituting these in Eq. (2.11), we have 1jIii

(am2+bm+c)u(x)emr+(2ma+b)em"~'+ae*~" 0 - I 'I
I f!
The first term in this expression vanishes in view of Eq. (2.13). The coefficient of ut is zero
since m = - b/2a in this case. Hence, the above expression simplifies to
exp(mx) a u" - 0
Multiplying by exp( - m ) and integrating, you will get
u' = K
where K is an arbitrary constant of integration.
Integrating again, you will get

Hence, the desired solution is

y2 - x em = x e-hr/2a

where the arbitrary constants K and C have been dropped (since we are seeking only a
second linearly independent solution). Hence, the general solution of a second order
differential equation, when auxiliary equation has repeated real roots, is

To test that ewa and x e*& are linearly independent, you can compute their Wronskian

It implies that e-(bX/O)

and x e- are acceptable solutions. The arbitrary constants CI
and C2 occurring in Eq. (2.19) can be determined using specified initial or boundary
conditions.
You may, therefore, conclude as follows:
r
When the auxiliary equation for a second order ODE with constant coefficients has
repeated real roots (m, =' m2 = m), the general solution is given by
y = (C,+ C2x ) e x p ( 4
where CI and C2 are arbitrary constants.

You may now like to solve an SAQ to be sure that you have grasped this method.

SAQ 2
Solve the initial valWe problem
 Mathematical Methods in Complex Roots
-
Physics XI
Let us now consider the case for which the characteristic equation has complex roots of the
form m = a + i p, where a and p are real. From your school mathematics, you know that
complex roots of a real polynomial equation always occur in conjugate pairs. That is, if

is one of the roots, then

m2 = a - i @
is also a root.
As before, we can obtain the general solution as a linear combination of two linearly
independent solutions as
y = A exp(ml x ) + B exp(rn2x)

You will note that this solution is complex. Call you express it as a real solution? To do so,
we use Euler's forn~ula:

e'-
i
l+x+-+-+
X' -
.... e z i e = cos9*isin 9 (2.22)
2! 3!
-
Put x if3 in this expansion. The
result is
This gives
y = ear[A(cos~+isinpx)+B(cosplc-isil~@)]
oa e3 el
e"=l+ie---i-+-..+..
2! 3! 41 i-, = eaK[(A+B)cospx+(A-B)isinpx]
e2 e4
By letting C1. = A + B and C2 = (A - B ) i , you can write
y = ea"(C1 cos px + C2sin $x) (2.23a)
- wsetisine Putting C1 = C cos @ and C2 = C sin +, you can rewrite Eq. (2.23a) as

where C and (Iare arbitrary constants. These are related to C1 and C2 by

You will note that when the roots of characteiistic equation are complex, they generate
solutions of the form of a product of an exponential and a trigonometric function. Therefore,
you may conclude as follows:

If the characteristic equation of a second order ODE has complex roots of the form
m = a % i p, the general solution is of the form
' y =Cemcos(plc-@)
I

You know that the differential equationgoverning the motioi of an undamped spring-mass I
I
dl
system is -2+ wix = 0 where w i = k/m. You would readily note that the characteristic I
dt
equation for this case is . I
I
2 I
m +wz = O i
I
which has roots
I
ml = iwo and m2 = -ioo !
i
Hence, the general solution is 1
x(t) = C l w s w o t + C 2 s i n w o t I

38 I

:'
I a
 ~~~~~~~~~~
The oscillatory motion described by this exainple should continue indefinitely. Second Order Ordlnaty
In practice, however, several factors lead to loss of energy-of the system. As a ,
W i '
result, oscillations tend to decrease. In the followii~gexample, we illustrate the
ineihod discussed in this section to study the motion of a damped harmonic
oscillator.

Consider a spring-mass system which is damped by a viscous force (Fig.2.1) , which

oat1 be modelled so that it is linearly proportional to velocity. What differential equation
describes its motion and what are its acceptable solutions?
Solution
We know that viscous force opposes motion. Hence, the differential equation describing the
xvlotion of a damped spring-mass systein is

&x+ &+h., 0
dt2 dt

d 2x &
or -,+2b-+ oix = 0 (2.24)
- dt dt
where 26 = y/M and wi = k/M.
On comparing it with Eq. (2.1 1); you can write the characteristic equation as
m2+2bm+w$ = 0
which bas roots

ml - - b + m and rn2 - -6-4-

Fig. 21 I A damped
spring-masssystem.

These roots depend on damping and determine the inotioil of the oscillator. Depending on
the value of ( b2 - & )v2, we have three possibilities.
Case
Case 2: If b -
If b z wo ,
wo , b 2
is positive and we have two real distinct roots.
- & = 0 and we have real repeated roots.
Case 3: If b < wo , b2 - w$ is negative.and is iinaginary, i.e., we have a
co~nplexdnjugate pair of roots.
Let us now discuss these cases.
Case 1
When the roots are real and distinct, the system is said to be heavily damped and the general
solution of Eq.(2.24) is given by
x(t) = exp(- bt) [Cl exp(Pt) + C2 exp(- pt) I (2-25)

where fJ = m.
This represents non-oscillatory behaviour. The systein is said to be heavily damped and
such a motioil is called dead beat.

The actual displaceinentof any such system will be determined by the initial oo~~ditiom.
TO unnpute
this, we would like you to solve the following SAQ.
-.
SAQ 3 spend
A heavily damped oscillator in its equilibrium position is suddenly kicked so that at 5 min
h
t = 0 , x = 0 and
dt .
-
= vo. Compute the particular solutiol~and interpret the resulting
.,
expression fof displacement.
-.._
.On working out this SAQ, you will find that the displacement of a heavily damped
oscillator is determined by the interplay of an increasing byp-erbolic function and a
 -Mathematid M+ods in decaying exponential. As a result, the displacement increases initially, attains a maximum and
Physics I1- thereafter decays.

- -
B e hyperbolic €unctions are so
named becausex a cosh0, and

-
y a sinh 0 defme a rectangular
hyperbolaxz- y2 a'. Compare it

-
with tho parametric equation oftbe
circle x' + y2 a2 whicb is defined
byx=acos0andy-asin@. Assuming ~ ( 0>)0

t t

Fii2.2: Displacemmt-time graph for an Fig 23: Displncement-time graph for a critically damped
overdamped spring-mw system spring-mass system

Fig 2.2 shows two typical over-damped motions for drO

dt
= 0 and &m
Lit
Case 2
When we have repeated real roots, the general solution of Eq. (2.24) is given by
x(t) = (C, + C2t ) exp(- bt) (2.26)
Note that here C 1has dimensions of length and C2those of velocity. As before, these
constants can be determined by specifying initial conditions. You can easily verify that for
initial conditions given in SAQ 3, C1= 0 and C2 = voSO that the complete solution is
I
I x(t) = v, t exp(- bt) (2.27)
I--
Such a system is said to be critically damped. Typical graph of a critically damped system
I' dX ( 0 )
for -
dr ( O)= 0 and - > 0 is shown in Fig. 2.3.
dt dt
Case 3
When the roots are imaginary, let us write

I
where i = fiand wd = is a real positive quantity. Hence, the displacement is
given by
x(t) = exp(- bt) [C1exp( i t ) + C2exp(- i odt) ]
= C exp(- bt) cos (ad t + 4) (2.28)

You will note that Eq. (2.28) represents oscillatory motion whose amplitude decreases
expone~ltiallyat a rate governed by b. Such a system is said to be weakly damped. The
displacement of a weakly damped system is depicted in Fig. 2.4.
 Second Order Ordinary
Dflerential Equations with
Let us sum up what you have learnt so far: Constant Coef&ients

1. We can solve a second order homogeneous linear ordinary differential equa tibii with
conslant coefficients using exponential functions. The form o f the solution depends
on the roots of the charncteristicequation.
2. When we have distinct real roots, there exist two linearly independent functions of
the form ex&, x) and exp(m, x ) and the general solution is given by
y(x) C1 exp(ml x ) t C2 exp(m2 x)
3. When we have real equal roots, thetwo linearly independent functions in the general
solution are of the form exp ( m ) and x exp ( mx ), i.e.,
~ ( =4 ( e l + C2 exp(m)
4. When we have a complex coqjugate pair of roots, tbe two linearly independent
solutions are of the fonn exp(m) sin @ and exy(w) cos @, and the general solution
can be written as
y(x) = exp(cuc) (Cl sin @ t C2 cos @)

= Cexp(ox)cos($x-+)

So far we have considered homoge~leouslinear equations with constant coefficients. These

equations do not satisfactorily model forced mechanical and electrical systems. In fact, such
sjlstems can be fairly accurately represented by nonhomogeneous second order linear
equations. We now wish to obtain solutions of such equations. Can you use the method
discussed in the preceding section to solve non-homogeneous equations? NO, we have to
look for new methods. Let us lean1some of these now.

2.4 NONHOMOGENEOUS LINEAR EQUATIONS WITH

CONSTANT COEFFICIENTS ,

From Sec. 2.2, you would recall that the general solution of a secoild order nonhomogeneous
linear ordinary differential equatio~lwith co~istailtcoefficients Is composed of the particular .
integral (PI) and the complementary function (CF). You can easily verify this by substituting
y = y, t y, in the equation

You can obtain y, (x) by using the method of the pededing section. For instance, CF for the
equation

yc = C1 e-3z+ C2eYX

(You can check this by direct substitution.)

Y

This means that finding the particular integral is at the heart of the method of solving a
non-homogeneous equation. But how to get yp?One systematic approach to find y, is based
on the method of reduction of order, which you have learnt in Unit 1. The other commonly
used methods are the method of undetermined coemcients and the variation of
parameters. Let us pow learn these.

2.4.1 The Method of Undetermined Multipliers

The basic idea of this method is to first construct the general form of the particular integral
From the forcing function. The11we determine coefficients for yp that allow it to satisfy the
' given differential equation. In the following example, we have illustrated this concept. ,
Matbematicd Methods In
Physics. I1
Determine the particular integral ofy" + 3y' + 2y = exp(2r).
Solution
Since the driving function is exp(&), we assume the PI to be of the form yp
We substitute ye and its derivatives
- A exp(&).

y' = 2.4 exp(2x)

and
y'' = 4A exp(2x)
in,thegiven equation and solve for A. This gives
1
A=-
12'
Thus

is a PI of the given equation.

You may now ask: Will this method work always? No, we have to refine it when the forcing

-
function in the differential equation is a polynomial of order n. For example, consider the
differential equationy" + 3y' + 2y 5x2. Following the method outlined above, you may
choose

On substituting ye and its derivatives in the differential equation of interest, you will get

It is impossible to solve it for A. This means that the method illustrated in Example 3 fails to
give a particular solution. Then, how to find yp(x) ? We tabulate below the form ofyp(x)
depknding on the form of &).

Table 2.1

Form of forcing function Nature of the root of Form of particular integral

auxiliary equation

A e kx When k is not a root c e kx

k is single root cxekx
k is double root
cx2ekx
Polynomial k = 0 is not a root c0+ c l x + c 2 x 2 +...
A x n(n = 0 , 1 , ...) k = 0 is single root %(CotC l x + ...)
k = 0 is double root
xZ(c0+ C1 x + c;x2 + ..)

A coskx ik is not a root Ccoskx+Dsinkx

A sin kx ik is a single'root ~(CcoskxtDsinh)

From the table you will note that if g(x) is of the form given in column I, the corresponding
PI will be of the fonn given in colum 3.%c form of PI will also be determined by the
nature of the root of the auxiliary equation as given in c o l 2. ~Note also that if a term irl
g(x) is a solution of the homogeneous equation corresponding to the given non-homogeneous
. ODE, the fonn of ye is modified as follows: y, is multiplied by x or x 2 depending on whether
the root of the auxiliary equation is a single or a double root. This is termed the
modification rule.
Further, if g(x) is a sum of the functions in column 1, we choose yp as the sum of the
hnctions in the correspondi~lgrows.
 ?+

Second Order Ordinary

SAQ 4 DiRemtial Equations with
Find the PI of the equation i
- Coemcients
Constant

S end
x
yfl-y = x + -
2
14;min
C

Let us now consider an example involving a trigonometric forcing function.

Example 4
Find the general solution of the differential equation
y" + 4y - 2 sin.&
Solution
The solution of the homogeneous equation is
yc(x) = Clcos 2.x + C2sin 2x
Since the forcing function is itself a solution of the corresponding homogeneous equation,
using the modification rule, we assume a solution of the form
yp(x) = Ax cos 2x + BXsin 2.x
Substituting this into the original differential equation, we get

- 4Bx sin 2x + 4Axcos 2x + 4Rxsin 2x = 2 sin 2x

Equating coefficients of different trigonometric functions, we get

coefficient of sin h : -2 ~ 2~
- =2
coefficient of cos 2x : 2B+ 2B = 0
coefficient of x sin 2x : - 4B + 4B = 0
coefficient of x m s 2x : - 4A + 4A = 0.
These equations require that A = - - and B = 0 Thus,
1
2

A general solution is then ,

Y(X) YC(4 + yp(x)

1
= C1cos2x+C2si~12x--xcos&
2

We now hope that you have understood the method of solvillg second order non-homogeneous
ordinary differential equations with constant coefficients using the method of undetermined
multipliers. .
Let us now summarise:

1
-

In the method of undetermined multipliers, the particular integral is constructed

from the forcing function. The arbitrary constant(s) is (are) determined by solving
equations obtained on comparing coefficients of like feqns p p the two sides of the
given equation .

To ensure that your progress is satisfactory, we would like you to solve the following
SAQ.
 Mathematical Methods in
Physics .I1 SAQ 5

Spend Find the general solutions of the following differential equations:

20 min
(i) y U + y = xi
(ii) y" + 4y = 3 cos x

- -

From Unit 4 of PHE-02 course, you would recall that non-homogeneous linear'
differential equations find an immediate application to damped spring-mass systems
.
acted upon by external forces buppose that the applied force causes the weight in a
spring-mass system to move up and down in some prescribed manner. Denoting the
external applied force by F(t), the differential equation describing one-dimensional
motion of such a system is
m i + y i + k x = F(t)
Suppose the forcing function is given by F(t) = Focos of, where,,Fo is the constant
amplitude and o is the angular frequency. Then, you can write
m i + y i + k x = Focoswr (2.29)
Since the solution of this equation depends on the damping force, we have to consider
separately the cases y = 0 (undamped) and y > 0 (damped). Let us consider the first case
now.

Undamped Forced Vibrntions

If there is no damping force, the differential equation describing the motion of the'
spring-mass system becomes
m;+kx = Focosot (2,30)

Let us assume the weight to be initially at rest and that w * wo = 6.

By the method of
undetermined coefficiei~ts,you call show that the general solution of this equation is

x(t) = C, cos wo t + C, sin wo r + F0

cos wt
m(,a-w2)
Then, using i( 0) = x(0) r 0, we get

c; - Eb and C2 = 0
m(o8-w2)
The desired solution is then

x(t) =
Fo
(COS wt-COS 00t )
m(w0-w2)

If we use the identity cos A - cos B =

rewritten in the forin
- 2 sin(A- sin) ,I+( this expressio~ican be

x(t) = m(w8-w2)
2P0
sin ) (y
(y ) t sin t (2.31)

X
Since the two sine functions are of different frequencies, there will be occasions, especially
when w is close to wo, when their amplitudes will either magnify or cancel one another
o t (see Fig. 2.5). This magnification and cancellhtion occurs at regular intervals and is called '

beat. In acoustics, these fluctuations can be heard when two tuning forks of slightly
different frequencies are set into vibration simultaneously. The same phenomenon occurs in
'electronics, where it is called anlplitude nlodulation.
You must have noticed that the method of uilQtennhed coefficie~~ts is limited to those
Fig. 2.5 : Diflpl8cementhc equatioils whose driving fuilctions are ofvery special form. Let us tlow discuss the so-called
plot of undamped forced
vibrations variation of parameters method that is applicable to all linear differential equations with.
constant coefficients. - .
1 44 I
1
 The Method of Variation of Parameters Second Order Ordiaary
2.4.2 Differential Equstilom with
Let yl and y2 b e any linearly independent solutions of the homogeneous equation Constaat Coefficients
corresponding to the given non-homogkneous equation. To find the particular integral, let US
assume that
Yp = UYl + vY2 (2.32)

where u and v are unknown functions ofx. To determine these, we differentiate Eq. (2.32)
.
with respect to x This gives
yd = u'yl t u y i +v1y2t v y ; (2.33)

We seek a solution such that

' u'yI+v1y2=0 (2.34)

Using this condition in Eq. (2.33), you will get

y; = uyi +vy4
Differentiate this expression again. The result is
y: = u yil+ v y2" + U' yi + v1y4 (2.36)
'
On substituting y, , yi and yi' for , y' and y", respectively in the equation
a y l ' + b y ' + c y = g(x)
you will obtain ,
a ( u y i l +vy;' + u l y i t v 1 y ; ) + b ( ~ y ~ + v y ~ ) + c ( u y ~=+g(x)
vy~)

I This can be rearranged as

u (ayil+ b yi + cyl) + v (ayil+b yi + cy2)+ a (u'yi +.V' yi) = g ( ~ ) (2.37)
' Since yl and yz are solutiq& of the homogeneous equation
For two simultancous linear equations of
ay"+byl+cy 0
-
tile form a,, xi + a,, r, b,
5

the terms in both parentheses drop out. Hence, Eq.(2.37)reduces ta and a, x, + a, x, b,

Cramer'srulc tellti us that the solutions
for x, and r, are

This means that u' and v1satisfy the system of Eqs. (2.34) and (2.38) These can be solved
by. Cramer 's rule. Thus Xl
and
- D

Xa " U
'wherc

and
D w . l : : I:
is s non-zero determinat.

You will note that the denominator in thjs equation is the Wro~nskia~l
of two lincax bde-
pendent hnctibns yl and yz and is non-zero.
These equations can be integrated to obtain
MathematicalMethods in
-
Physics I1

'lhe particular solution is then

You would agree that the integration required to obtain u and v can make this method difficult
to use. So whenever possible, you should use the method of undetermined coefficients.
In SAQ 5, you obtained the PI of the equation
2
yl'+y = X

using the method.of i~ndetenninedmultipliers. Now let us obtain a particular solutioi~of this
equation using the method of variation of parameters.
Since the two independent solutions are
yl(x) = sin x and yz(x) = cos x
the Wronskian is given by

A particular s o i ~ o is
n then given by

S
y,,(x) = sinx x 2 cosx&- cosx
S x2sinx&

Integrating by parts, we get

This solution is, however, not unique. There can be other solutions as well.
To sum up this section, we say that

In the method of variation of parameters, we construct particular integral by a linear

combination of 'independentsolutions of the hombgeneous equation corresponding to the
given non-homogeneous equationsubject to the condition
u' y, + v' yz = 0
where u and v are given by

I aridi W is the Wro~lskianof yl and y, . ,

You may now ask: Can we use each of these methods for every second order ODE? The
applicability of a method essentially depends on the differential equation under
consideration. Yet the method of variation of parameters is the most general one and the
method of undetermined inultipliers is probably the most popular.
Let us now summarise this unit .
2.5 SUMMARY

If yl and y2 are solutions of a second order ODE, then they are linearly independent, if
and only if, their Wronskian, defined as
Yl . Y2
w
~i Y;
is non-zero.
 Second Order Ordinary
o A second order homogeneous ODE with constant coefficients can be solved using IWercntial Equations with
exponential functions. The form of the solution depends on the roots of the auxiliary . Constant CwCEcients
equation.
For distinct real roots, there exist two liiiearly independent functions of the form
exp(ml x) and exp(m2x), and the general solution is given by
Y(X) = C1 exp(m, x) + C2 exp(m2 x)
When rootsare equal (ml = m2 = m), the general solution is given by

For a complex coqjugate pair of roots, the two lincarl y independent solutions are of the
form exp(cuc) sin pic and exp(m) cos px, and the general solutioli can be written as
y(x) = e x p ( a ) (C1 sin pic + C2 cos fi)

o The solution of a non-homogeneous second order ODE with constant coefficients is a

sum of the particular integral and the complementary firnction.

The complementary function is the general solution of the homogeileous equation

corresponding to the given non-homogeneous equation.

0 In the method of undetermined muJtipliers,the particular integral is constructed from the

forcing function. Depending on the form of the forcing function, yp is given as follows:

Form of forcing functiol~ Nautre of the root of Form of particular integral

auxiliary equation

A e kx When k is not a root C ekx

k is single root cxekx
k is double root cx2ekX -

Polynomial ,k= 0 is not a root C, + C1x + c 2 x 2 + ...

A x n ( n = 0 ,lj ...) k = 0 is single root x(Co + C1x + ....)
k = 0 is double root
x2(c0 + c l x + c 2 x 2 +..)

Amsku i k is not a root C cos b + D sin lac

A sin kx i k is a single root x(C cos kx + D sill ku)

The arbitrary constants are determined by solving equatioils obtained on comparing

coefficients of like terms on the two sides of the given lion-homogeneousODE.

In the method of variation of parameters, the particular integral is given by

Y, = UYl+VY2 J,;

where y l and y2 are linearly independent solutions of the homoget~eouspart of the given
differential equation. The functions u and v can be computed using the relations

where W is the Wronskian of yl and y2.

2.6 TERMINAL QUESTIONS

1. The motion of a damped harmonic oscillator is described by the equation

Obtain the particular integral.

Mathematical Methodsin 2. Solve the differential equations
-
Physics I1
d2
i) --J'+y = secx
dx2

2.7 SOLUTIONS AND ANSWERS

1. The Wronskiap for these functions is '

sin 2x cos 2x
w (4=
2cos 2 x - 2sin 2x

= -2
Since W(x) * 0 for all x, the functions sin 2x and cos 2x are linearly independent.

2. The auxiliary equation corresponding to the given ODE is

m 2+ 6 m + 9 = 0

. -a double root m = -3. Hence, the general solution is

which h;ls
y(x) = (C1 + C2x) e-3X
The condition y( 0) = 2 gives
2 = C1 (ii)
Differentiate (i) with respect to x. This gives

-
dy
ak
P C2 em3'- 3 (C1 + C2x) e-3X

Us.hg the condition 1, we get

dx

Hence, the desired solution is

y(x) = (2 + 7x) e-3Y
3. x (1) = exp(-bt) [Cl exy(Pt) + C2 exp(-Pt)l
At t = 0 , x 0. This gives
0 = C1 + C2 I

or
C1 = -C2
Differentiate the given expression with respect to time. The result
-
ak
- b exp(-bt) [ Ci exp(Pt) + C2exp(-fit) ]
P

dt

Using the condition &@ vy ,we find that

dt
 Second Order Ordinary
Ditrerential Equstions wllb
Constant CoeMcients

Hence

This shows that the resultant motion of a heavily damped oscillator is determined by
the interplay of a decaying exponential and a hyperbolic function,

4. Assume that the PI is of the form

yp = C o + C 1 x + C 2 x 2
Substituting it and its second derivative .

in the given ODE, you would get

For (i) to be an acceptable solution, (ii) should be an identity. This gives

co= -1

and

Therefore
-
5. (i) The solutibn of the homogeneous equation

is found to b'e
y,(x) - C1cos x + C2sin x
A particular integral is assumed to have the form

This is substituted into the original differential equation to give

2~ + A . x 2 + ~ x + C= x2
Equating coefficients of the various gapers of x, we have

Coefficient of x1: B=O

Coefficient of x2 : A m 1
MathematicalMethods in These equations are solved simultaneously to give the particular solution
-
Physics 11

Finally, the general solution is

(ii) The solution of the corresponding homogeneous equation y" + 4y = 0 which has roots
* 2i, is
yc = C1sin 2x + C2cos 2\:
To find a particular solution of the given equation, we assume that the general form
ofyp is
yp = Asinx+Bcosx
This is the correct expression since neiker of these functions is in y,. Substituting yp
and yd' into. -"the given differential equation, we obtain

'Expanding and collectiiig like terms yields

3.4 and 3B = 3
-
r h ,

which has the solutionA = 0 and B = 1. Hence,yp m s x and the ~olutionis

(iii) Assume the particular solution to have the form

yp(x) = A ~ ~ + B X ~ + C X + D
Substitute this into the given diffe~entialequation, The result is
Aex+2B+AeX+2Bx+C%2Aex+
2Bx2+ 2Cx+2D = 4eX+2x2
Equating the various coefficients, you would obtain
/

Coefficient of d' : A +A + 24 'a 4

Coefficient of x0: 2B + C + 20 = 0

Coefficient of x1: 2Bt2C - 0

Coefficient of x2 : 2812
From the above equations, we find thatA = 1, B = 1, C = -1 and D = -1h. Thus,
yp(x) = e X + x 2 - x - 1h
Terminal Questions
1. The given differential equation is
mjl. + y i + iac - Focoswt (9
where dot over x denotes derivative with respect to time. As such, you have,learnt
to solve it in your PHE-02 course. But we repeat the solution for the sake of
completeness. The corresponding homogeneous equation is
mi+yi+kx = 0
which is the same as the equation describing damped vibration without a forcing
function. Its solutio~idepends upon the sign of 2- 4mk. Thus
If y2-4mk>0
x,(~) ~ ~ ~ - ( a - B ) t + ~ ~ ~ - ( 'a + B ) l
 Second Order Ordinnq'
DKPerentiml Eguntions with
coost~nt coemcients
1fy2- 4mk'< o
. xC(t)= e-at(C1 cos w't + C2sinoft)

where a. = y/2m , P E l,/=,wl

2m
= L~/=,c
2m
= Jf&z
, and tan 6 = C2/C1.
Since no constant multiple of the driving functionFo cos wt is a term ofxe (t), the
particular solution is of the form
xJt) = A cos wt + B sin wt
Differentiating twice with respect to time, we get
ip(t)= -oAsinwt+wBcos'wt

2p(t) = -w% cos wt - o2 B sin wt

Substituting these in (i) and collecting the coefficient of sine and cosine terms, we have

Equating the coefficients of the sine and cosine terms,on bo,th sides of this equality,
we get
- w y ~ + ( k - m o 2 ) B = .O

and ( k - m o 2 ) ~ + w y=~F o
Solving these forA and B, you will get

Recalling that 6 = wo, we can writeA and B as

We choose to write xp(t) in the form

xP(t) = C cos (wt - 6)

where C = h /d m2 ( wi - w2 l2+ ~ 0 and

4 ~tan 6 = wy/m (wi - w2).
For large values oft, the motion is essentially described by x,,(t). For this reason,
xp(t) is called the steady-state solution.
2.(i) Since two linearly independent solutioiis of the corresponding homogeneous equation
are cos x and sin x, the general solution of the give11equation is

where yp = u cos x + v sin x and u' and v' are, respectively given by

0 sinx COS X 0
,
secx cos x -sinx secx
- dv ,
and -
cosx sin x dx cos x sinx

-sin x cos x -sinx casx

ah = -tan x and dv
Therefore, -
dr
- = 1, from which it readily follows that u
ak
- In I cos x I
Mathe'matid Methods in and v The general solution is, therefore,
-
Physics 11
= X.

y = ~~cosx+~~sinx+ln)cosx)cosx+xsir~x

Note that the method of undetermined coefficients could not be used to obtain yp
because secx is not a solution of the homogeneous linear differential equation.
(ii) The corresponding homogeneous equation y" - y = 0 has the general solution
y = C1ex+ C 2 e-'. Because of the nature of the driving function, yp could not be

- -
found by the methad of undetermined coefficients. In the method of variation of
parameters, we put yl e x and y2 e-' in the expressions for u' ind v' to obtain

Therefore, u' = x / 2 and v' = from which it readily follows that

2
u = 2/4and v = - (xek/4 ) + ( ek/g ). The general solutioi~is then
y a e l e x + C 2 e - ~ +i x 2 e x - - e + -e
X x l x
4 4 8
1 1
Finally, we note that C l e X and - excan be combined as(C1 + - ) ex = C e X and the
8 8
general solution may be written as .

y = C l e X + C 2 e - X -1- x e X +1- x 2 e x
4 4
 UNIT 3 SECOND ORDER ORDI
DIFFE E
WITH E
Structure
3.1 Introduction
Objectives
3.2 Some Terminology
3.3 Power Series Method
3.4 The Frobenius' Method
3.5 Summary
3.6 Terminal Questions
3.7 Solutions and Answers

3.1 INTRODUCTION
In Unit 2, you have learnt how to solve second order ODEs with constant coefficients. The
solutions of these equations are simple exponentials, trigonometric or hyperbolic functioi~s
known from calculus. But in many physical and engineering problems, we have to solve
second order ordinary differential equations with variable coefficients. For example, we
have to solve such equations to study the field distribution around a charged sphere or a
cylindeqand energy production in a reactor. Similarly, when we wish to know how high
a vertical column of uniform cross-section can be extended upward until it buckles hnder its
own weight, we have to solve a second order ODE with variable coefficients. In such cases,
simple algebraic or transcendental solutions do not exist and methods discussed in Unit 2 do not
work. We, therefore; look for other methods.
One of the most elegant and efficient methods of solving such ODEs is the power series
method. This is so particularly because it facilitates numerical computations. Even so, it
has limited utility when coefficients of the given differential equation are not well
defined at some point. In such cases, we use an extension of the power series method,
called the Frobenius' method. You will learn these two methods in this unit. The
t properties of power series and certain other mathematical concepts are given in an
Appendix at tbe end of this unit. It would be better if you study the appendix before
studying this unit.
I
Objectives
After studying this unit, you should be able to

define ordinary and singular points

I locate and classify the type of siilgularity

use power series method to solve a second order ODE about an ordinary point

I use Frobenius' method to solve a second order ODE about a regular singular point

3.2 SOME TERMINOLOGY

While studying first and second eider ODEs with constant coefficients, you have learnt
some basic terminology. You would come across some more common terms, which you do
not know as yet, in reference to second order ODEs with variable coefficients. This section
is intended to familiarise you with these concepts,