Advanced Calculus and Numerical Methods 18MAT21

MODULE – 3
PARTIAL DIFFERENTIAL EQUATIONS

CONTENTS:

 Formulation of PDE by elimination of arbitrary constants

 Formulation of PDE by elimination of arbitrary functions,

 Solution of non- homogeneous PDE by direct integration,

 Solution of homogeneous PDE involving derivative with

respect to one independent variable only.

 Derivation of one dimensional heat and wave equations and

their solutions by variable separable method.

DEPT. OF MATHS/SJBIT Page 1

 Advanced Calculus and Numerical Methods 18MAT21

Introduction:
Many problems in vibration of strings, heat conduction, electrostatics involve two or more variables.
Analysis of these problems leads to partial derivatives and equations involving them. In this unit we first
discuss the formation of PDE analogous to that of formation of ODE. Later we discuss some methods of
solving PDE.

Definitions:

An equation involving one or more derivatives of a function of two or more variables is called a partial
differential equation.

The order of a PDE is the order of the highest derivative and the degree of the PDE is the degree of highest
order derivative after clearing the equation of fractional powers.

A PDE is said to be linear if it is of first degree in the dependent variable and its partial derivative.

In each term of the PDE contains either the dependent variable or one of its partial derivatives, the PDE is
said to be homogeneous. Otherwise it is said to be a nonhomogeneous PDE.

 Formation of pde by eliminating the arbitrary constants

 Formation of pde by eliminating the arbitrary functions

Solutions to first order first degree pde of the type

P p + Q q =R

Formation of pde by eliminating the arbitrary constants:

(1) Solve: x 2 y2
2z  2  2
a b
Sol: Differentiating (i) partially with respect to x and y,
z 2 x 1 1 z p
2  2 or 2  
x a a x x x

2z 2 y 1 1 z q
 2 or 2  
y b b y x y

Substituting these values of 1/a 2 and 1/b2 in (i), we get

(2) z = (x2 + a) (y2 + b)

DEPT. OF MATHS/SJBIT Page 2

Advanced Calculus and Numerical Methods 18MAT21

Sol: Differentiating the given relation partially

(x-a) 2 + (y-b) 2 + z2 = k2 …(i)

Differentiating (i) partially w. r. t. x and y,

z z
(x  a )  z  0, ( y  b)  z  0
x y

Substituting for (x- a) and (y- b) from these in (i), we get

  z  2  z  2 
z 1         k 2 This is the required partial differential equation.
2

  x   y  

(3) z = ax + by + cxy ...(i)

Sol: Differentiating (i) partially w.r.t. x y, we get

z
 a  cy..(ii )
x

z
 b  cx..(iii )
y

It is not possible to eliminate a,b,c from relations (i)-(iii).

Partially differentiating (ii),

 2z
 c Using this in (ii) and (iii)
xy

z  2z
a y
x xy

z  2z
b x
y xy

Substituting for a, b, c in (i), we get

DEPT. OF MATHS/SJBIT Page 3

Advanced Calculus and Numerical Methods 18MAT21

 z  2 z   z  2z   2z
z  x  y   y  x   xy
 x x  y   y x  y  xy

z z  2z
z  x  y  xy
x y xy

x 2 y2 z2
(5) 2  2  2  1
a b c
Sol: Differentiating partially w.r.t. x,

2x 2z z x z z
  0, or  
a 2 c 2 x a2 c 2 x
Differentiating this partially w.r.t. x, we get

1  z   2 z    2z 
2
 z 
2
1 c2 
  2    z 2  or      z 2
a 2
c  x  x  a 2
 x 
 x  

: Differentiating the given equation partially w.r.t. y twice we get

z z  z   2z
2 2
z z  z   2z
    z 2   z 2
y y  y  y x x  x  x

Is the required p. d. e..

Note:

As another required partial differential equation.

P.D.E. obtained by elimination of arbitrary constants need not be not unique

Formation of p d e by eliminating the arbitrary functions:

1) z = f(x2 + y2)

Sol: Differentiating z partially w.r.t. x and y,

DEPT. OF MATHS/SJBIT Page 4

Advanced Calculus and Numerical Methods 18MAT21

z z
p  f ' ( x 2  y 2 ).2 x, q   f ' ( x 2  y 2 ).2 y
x y

p /q = x / y or y p –x q=0 is the required pde

(2) z = f ( x +ct ) + g (x -ct)

Sol: Differentiating z partially with respect to x and t,

z 2 z
 f ' ( x  ct )  g ' ( x  ct ), 2  f " ( x  ct )  g" ( x  ct )
x x
Thus the pde is

2 z 2 z
 0
t 2 x 2
(3) x + y + z = f(x2 + y2 + z2 )

Sol:Differentiating partially w.r.t. x and y

z  z 
1  f ' ( x 2  y 2  z 2 ) 2 x  2 z 
x  x 

z  z 
1  f ' ( x 2  y 2  z 2 ) 2 y  2 z 
y  y 
1  (z / x) 1  (z / y )
2 f ' (x2  y2  z 2 )  
x  z (z / x) y  z (z / y )

z z
( y  z)  ( z  x)  x  y is the required pde
x y

(4) z = f ( xy / z ).

Sol: Differentiating partially w.r.t. x and y

z  xy  y xy z 
 f '    2 
x  z  z z x 

z  xy  x xy z 
 f '    2 
y  z  z z x 

DEPT. OF MATHS/SJBIT Page 5

Advanced Calculus and Numerical Methods 18MAT21

 xy  z / x z / y
f '   
 z  ( y / z ){1  ( x / z )(z / x} ( x / z ){1  ( y / z )(z / y}

z z
x y
x y

or xp = yq is the required pde.

(5) z = y2 + 2 f(1/x + logy)

z 1 
Sol :  2 y  2 f '(1/ x  log y)  
y  y

z  1
 2 f ' (1 / x  log y ) 2 
x  x 

z  z 
2 f ' (1 / x  log y )   x 2  y  2 y 
x  y 

z z
Hence x
2
 y  2y 2
x y

(6) Z = xΦ(y) + y (x)

z z
Sol :   ( y)  y '( x);  x '( y)  ( x)
x y

Substituting  ' ( y ) and  ' ( x)

2 z z z
xy  x  y  [ x ( y)  y ( x)]
xy x y

2 z z z is the required pde.

xy  x  y z
xy x y

DEPT. OF MATHS/SJBIT Page 6

Advanced Calculus and Numerical Methods 18MAT21

7) Form the partial differential equation by eliminating the arbitrary functions from
z = f(y-2x) + g(2y-x) (Dec 2011)
Sol: By data, z = f(y-2x) + g(2y-x)
z
p  2 f ( y  2 x)  g (2 y  x)
x
z
q  f ( y  2 x)  2 g (2 y  x)
y

2 z
r  4 f ( y  2 x)  g (2 y  x)...............(1)
x 2

2 z
s  2 f ( y  2 x)  2 g (2 y  x).........(2)
xy

2 z
t  2  f ( y  2 x)  4 g (2 y  x)................(3)
y

(1)  2  (2)  2r  s  6 f ( y  2 x)..............(4)

(2)  2  (3)  2 s  t  3 f ( y  2 x)............(5)

Nowdividing (4) by (5) we get
2r  s
 2 or 2r  5s  2t  0
2s  t

2 z 2 z 2 z
Thus 2  5  2  0 is the required PDE
x 2 xy y 2

LAGRANGE’S FIRST ORDER FIRST DEGREE PDE: Pp+Qq=R

(1) Solve: yzp + zxq = xy.

dx dy dz
Sol :  
yz zx xy

Subsidiary equations are

From the first two and the last two terms, we get, respectively

DEPT. OF MATHS/SJBIT Page 7

Advanced Calculus and Numerical Methods 18MAT21

dx dy dy dz
 or xdx  ydy  0 and  or ydy  zdz  0.
y x z y

Integrating we get x2 - y2 = a, y2 – z2 = b.

Hence, a general solution is

Φ(x2-y2, y2 –z2) = 0

(2) Solve: y2p - xyq = x(z-2y)

dx dy dz
Sol :  
y 2  xy x( z  2 y )

From the first two ratios we get

x2 + y2 = a from the last ratios two we get

dz z
 2
dy y

from the last ratios two we get

dz z
  2 ordinary linear differential equation hence
dy y

yz – y2= b

solution is Φ( x2 + y2, yz – y2) = 0

(3) Solve : z(xp – yq) = y2 –x2

dx dy dz
Sol :   2 2
zx  zy y  x

dx dy
 , or xdy  ydx  0 or d(xy)  0,
x y

on integration, yields xy = a

DEPT. OF MATHS/SJBIT Page 8

Advanced Calculus and Numerical Methods 18MAT21

xdx +ydy + zdz = 0 x2 + y2 + z2 = b

Hence, a general solution of the given equation

Φ(xy,x2+y2+z2)=0

yz zx x y
(4) Solve: p q
yz zx xy

yz zx xy
Sol : dx  dy  dz
yz zx x y

x dx + y dy + z dz = 0 …(i)

Integrating (i) we get

x2 + y2 + z2 = a

yz dx + zx dy + xy dz = 0 …(ii)

Dividing (ii) throughout by xyz and then integrating,

we get xyz = b

Φ( x2 + y2 + z2, xyz ) = 0

(5) (x+2z)p + (4zx – y)q = 2x2 + y

dx dy dz
Sol :   2 ..(i )
x  2 z 4 zx  y 2 x  y

Using multipliers 2x, -1, -1 we obtain 2x dx – dy – dz = 0

Using multipliers y, x, -2z in (i), we obtain

y dx + x dy – 2z dz = 0 which on integration yields

xy – z2 = b ….(iii)
5) Solve z xy  sin x sin y for which z y  2sin y when x  0 and z  0

when y is an odd multiple of .
2

DEPT. OF MATHS/SJBIT Page 9

Advanced Calculus and Numerical Methods 18MAT21

Sol: Here we first find z by integration and apply the given conditions to determine the arbitrary
functions occurring as constants of integration.

  z 
   sin x sin y
x  y 
The given PDF can be written as

Integrating w.r.t x treating y as constant,

z
 sin y  sin x dx  f ( y )   sin y cos x  f ( y )
y
Integrating w.r.t y treating x as constant
z   cos x  sin y dy   f ( y ) dy  g ( x)
z    cos x  ( cos y )  F ( y )  g ( x),
where F ( y )   f ( y ) dy.

Thus z  cos x cos y  F ( y )  g ( x)

z
Also by data,   2sin y when x  0. U sin g this in (1)
y
 2sin y  ( sin y).1  f ( y ) (cos 0 1)

Hence F ( y )   f ( y ) dy    sin y dy  cos y

With this, (2) becomes z  cos x cos y  cos y  g ( x)

U sin g the condition that z  0 if y  (2n  1) in (3) we have
2
 
0  cos x cos(2n  1)  cos x c(2n  1)  g ( x)
2 2

But cos (2n  1)  0. and hence 0  0  0  g ( x)
2
Thus the solution of the PDE is given by
z=cos x cosy + cosy

Method of Separation of Variables

x
1) Solve by the method of variables 3ux  2u y  0, giventhat u( x,0)  4e

u u
Sol: Given 3 2  0..............(1)
x x

DEPT. OF MATHS/SJBIT Page 10

Advanced Calculus and Numerical Methods 18MAT21

Assume solution of (1) as

U=XY where X=X(x); Y  Y ( y )

u u
3 ( xy )  2 ( xy )  0
x x
dX dY 3 dX 2 dY
 3Y  2X 0  
dx dy X dx Y dy
3 dX 3dX
Let K   kdx
X dx X
Kx
 3log X  kx  c1  log X   c1
3
kx
 c1
 X e 3

2 dY dY  Kdy
Let k  
Y dy Y 2
 ky
 Kdy  c2
 log Y   c2  Y  e 2
2
Substituting (2) & (3) in (1)
x y
K     c1  c2
U e 3 2

Also u ( x1o)  4e  x
 2x  kx
k 
x x
i.e., 4e  Ae  4e  Ae
 6  3

Comparing we get A  4 & K  3

x y
3  
U  4e 3 2
is required solution.

u u
2) Solve by the method of variables 4   3u, giventhat u (0, y)  2e5 y
dx y
u u
Sol: Given 4   3u
x y

DEPT. OF MATHS/SJBIT Page 11

Advanced Calculus and Numerical Methods 18MAT21

Assume solution of (1) as

u  XY where X  X ( x); Y  Y ( y )
 
4 ( XY )  ( XY )  3 XY
x y
dX dY 4 dX 1 dY
 4Y X  3 XY   3
dx dy X dx Y dy
4 dX 1 dY
Let  k , 3 k
X dx Y dy
Separating var iables and int egrating we get
kx
 log X   c1 , log Y   3  k  y  c2
4
kx
 c1
 X e 4
and Y  e 3 k  y  c2
kx kx
  3 k  y   3 k  y
 c1  c2 
Hence u  XY  e e 4
 Ae 4
where A  e c1  c2 
put x  0 and u  2e5 y
The general solution becomes
2e5 y  Ae 3 k  y  A  2 and k  2
 Particular solution is
x
5 y
u  2e 2

APPLICATION OF PARTIAL DIFFERENTIAL EQUATIONS:

ONE DIMENSIONAL HEAT EQUATION:

Consider a heat conducting homogeneous rod of length L placed along x-axis. One end of the
rod at x=0(Origin) and the other end of the rod at x=L.

DEPT. OF MATHS/SJBIT Page 12

Advanced Calculus and Numerical Methods 18MAT21

Assume that the rod as constant density  and uniform cross section A. Also assume that the rod
is insulated laterally and therefore heat flows only in the x direction. The rod is sufficiently thin
so that the temperature is same at all points of any cross sectional area of the rod.

Let u(x, t) be the temperature of the cross section at the point x at any time t.

The amount of heat crossing any section of the rod per second depends on the area A of the cross
u
section, the thermal conductivity k of the material of rod and the temperature gradient
x

i.e., the rate of change of temperature with respect to distance normal to the area.

Therfore q1 the quatity of heat flowing into the cross section at a distance x in uint
u 
time is q1  kA   per second
 x  x

Negative sign appears because heat flows in the direction of decreasing

temperature (as x increases u decreases )

q2 the quantity of heat flowing out of the cross section at a distance x   x

(i.e, the rate of heat flow at cross section x   x )

 u 
q2  kA   per second
 x  x  x

The rate of change of heat content in segment of the rod between x and x+ x must be equal to
net heat flow into this segment of the rod is
 u   u  
q1  q2  kA       per second.............(1)
 x  x  x  x  x 

But the rate of increase of heat in the rod

u
s  A x ...............(2)
t

Where S is the specific heat,  the density of material.

∴ From (1) & (2)

DEPT. OF MATHS/SJBIT Page 13

Advanced Calculus and Numerical Methods 18MAT21

u  u   u  
s  A x  kA     
t  x  x  x  x  x 
  u   u  
     
u  x  x  x  x  x 
or s  k
t  x 
 
 

Taking limit as  x  0 , we have

  2u u k  2u
s  k 2 or 
t x t s  x 2
u  2u k
or  c 2 2 ...........................(3) where c 2 
t x s

Is known as diffusivity constant.

Equation (3) is the one dimensional heat equation which is second order homogenous and
parabolic type.

Various possible solutions of standard p.d.es by the method of separation of

variables.

We need to obtain the solution of the ODEs by taking the constant k equal to

i) Zero ii) positive: k=+p2 iii) negative: k=-p2

Thus we obtain three possible solutions for the associated p.d.e

Various possible solutions of the one dimensional heat equation u t =c2uxx by the method of
separation of variables.

u  2u
Consider  c2 2
t x

Let u= XT where X=X(x),T=T(t) be the solution of the PDE

Hence the PDE becomes

  XT   2  XT  dT 2
2 d X
 c2 or X  c
t x 2 dt dx 2

DEPT. OF MATHS/SJBIT Page 14

Advanced Calculus and Numerical Methods 18MAT21

1 dT 1 d 2 X
Dividing by c2 XT we have 
c 2T dt X dx 2

Equating both sides to a common constant k we have

1 d2X 1 dT
=k and =k
X dx 2 c 2T dt

dT
d2X  c 2 kT  0
 kX  0 and dt
dx 2

D 2
 k  X  0 and  D  c 2 k  T  0

d2 d
Where D2 = 2
in the first equation and D = in the second equation
dx dt

Case (i) : let k=0

AEs are m=0 amd m2=0 amd m=0,0 are the roots

Solutions are given by

T = c1e0t  c1 and X   c2 x  c3  e0 x   c2 x  c3 

Hence the solution of the PDE is given by

U= XT= c1  c2 x  c3 

Or u(x,t) =Ax+B where c1 c2=A and c1 c3=B

Case (ii) let k be positive say k=+p2

AEs are m –c2p2=0 and m2 -p2=0

m= c2 p2 and m=+p

Solutions are given by

T  c'1ec and X  c '2e px  c '3e  px

2
p 2t

Hence the solution of the PDE is given by

u  XT  c'1ec .( c'2e px  c '3e px )

2
p 2t

DEPT. OF MATHS/SJBIT Page 15

Advanced Calculus and Numerical Methods 18MAT21

(A’ e px +B e  px ) where c1’c2’=A’ and c1’c3’=B’

2
p 2t
Or u(x,t) = c '1ec

Case (iii): let k be negative say k=-p2

AEs are m+ c2p2=0 and m2+p2=0

m=- c2 p2 and m=+ip

solutions are given by

T  c''1e c and X  c ''2 cos px  c ''3 sin px

2
p 2t

Hence the solution of the PDE is given by

u  XT  c''1e c .(c''2 cos px  c''3 sin px)

2
p 2t

u ( x, t )  e c ( A'' cos px  B '' sin px)

2
p 2t

u  2u
1. Solve the Heat equation  c 2 2 given that u(0,t)=0,u(l,0)=0 and u(x,0)= 100x/l
t x

n x n x
l l
2 100 x 200
Soln: bn   sin dx =  2  x sin dx
l 0 l l l 0 l

n x n x 
l

x.  cos  sin
200  l 1 l 
bn  2  2 
l  n / l  n / l  
 0

200  1 200  1

n n 1
200 1
bn  2 .  l cos n    .
l n n n

The required solution is obtained by substituting this value of bn

200  1 n  c t n 1
n x
2 2 2

Thus u ( x, t )   e 2 sin
n 1 n l l

DEPT. OF MATHS/SJBIT Page 16

Advanced Calculus and Numerical Methods 18MAT21

u  2u
2. Obtain the solution of the heat equation  c 2 2 given that u(0,t)=0,u(l,t)and
t x
 2Tx l 
 l in 0  x  2 
u(x,0) =f(x)where f ( x)   
 2T  l  x  in l  x  l 
 l 2 

n x
l
2
Soln: bn   f ( x) sin dx
l 0 l

 2l 
n x n x 
l
2  2Tx 2Tx
bn    sin dx   (l  x)sin dx 
l 0 l l l l
 l
2


 2l 
n x n x 
l
4T 
l  0
 x sin dx   (l  x)sin dx 
l l
 l
2


8T n
bn  sin
n
2 2
2

The required solution is obtained by substituting this value of b n

n  n  c t n x
2 2 2

8T 1
Thus u ( x, t )  2  2 sin e 2 sin
 n1 n 2 l l

u  2u
3. Solve the heat equation  with the boundary conditions u(0,t)=0,u(l,t)and
t x 2
u(x,0) =3sin  x

Soln: u ( x, t )  e p t ( A cos px  B sin px)............................(1)

2

Consider u(0,t)=0 now 1 becomes

0= e  p t (A) thus A=0

2

Consider u(1,t)=0 using A=0 (1) becomes

0= e  p t (Bsinp)
2

Since B≠0,sinp=0or p=n 

DEPT. OF MATHS/SJBIT Page 17

Advanced Calculus and Numerical Methods 18MAT21

 n2 2c 2t
u ( x, t )  e ( B sin n x)


In general u ( x, t )   bn e
 n2 2c 2t
sin n x
n 1

Consider u(x,0)= 3 sin n x and we have

3 sin n x  b1 sin  x  b2 sin 2 x  b3 sin 3 x

Comparing both sides we get b1  3, b2  0, b3  0

We substitute these values in the expanded form and then get

 2 t
u ( x, t )  3e (sin  x)

ONE DIMENSIONAL WAVE EQUATION:

Consider a tightly stretched elastic string of length l stretched between two points O and A and
displaced slightly from its equilibrium position OA. Taking O as origin and OA as x axis and a
perpendicular line through O as Y- axis. We shall find the displacement y a function of the
distance x and the time t.

We shall obtain the equation of motion of string under the following assumptions.

i) The string is perfectly flexible and offers no resistance to bending

DEPT. OF MATHS/SJBIT Page 18

Advanced Calculus and Numerical Methods 18MAT21

ii) Points on the string move only in the vertical direction, there is no motion in the
horizontal direction. The motion takes place entirely in the X Y plane .
iii) Gravitational forces on the string is neglected.

Let m be the mass per unit length of the string. Consider the motion of an element PQ
of length  s . Since the string does not offer resistance to bending, the tensions T 1

At P and T 2 at Q are tangential to the curve.

Since the is no motion in the horizontal direction, some of the forces in the horizontal direction
must be zero.

i.e., - T1cos + T2cosβ=0 or T1cos = T2cosβ=T=constant…..(1)

Since gravitational force on the string is neglected, the only two forces acting on the
string are the vertical components of tension - T1 sin at P and T2sinβ at Q with upward
direction takes as positive.
Mass of an element PQ is m  s . By Newton’s second law of motion, the
equation of motion in the vertical direction is
Resultant of forces = mass *acceleration
2 y
T2sinβ - T1sin =m  s ..............(2) .
t 2
2 T sin  T1 sin  m s  2 y
gives 2  
1 T2 cos  T1 cos  T t 2
m s  2 y
or tan   tan  
T t 2
.
2 y T
  tan   tan  
t 2
m s
2 y T  y   y  
     
t 2
m s  x  x  x  x  x 

(  s   xto a first approximation and tan  , tan  arethe slopes of thecurveof the string at x and x   x )

  y   y  
     
 y T  x  x  x  x  x
2
  
t 2 m  x 
 
 
Taking Limit as  x  0
2 y T 2 y 2 y 2 2 y T
 or c ...................(3) where c 2 
t 2 m x 2 t 2
x 2
m
DEPT. OF MATHS/SJBIT Page 19
Advanced Calculus and Numerical Methods 18MAT21

Which is the partial differential equation giving the transverse vibrations of the string .

Equation (3) is the one dimensional wave equation which is second order homogenous and
parabolic type.

Various possible solutions of the one dimensional wave equation u tt =c2uxx by the method of
separation of variables.

 2u 2  u
2
Consider 2  c
t x 2

Let u= XT where X=X(x),T=T(t) be the solution of the PDE

Hence the PDE becomes

 2  XT  2   XT 
2
d 2T 2
2 d X
 c or X  c
t 2 x 2 dt 2 dx 2

1 d 2T 1 d 2 X
2
Dividing by c XT we have 2 
c T dt 2 X dx 2

Equating both sides to a common constant k we have

1 d2X 1 d 2T
=k and =k
X dx 2 c 2T dt 2

d 2T
d2X  c 2 kT  0
 kX  0 and dt 2
dx 2

D 2
 k  X  0 and  D 2  c 2 k  T  0

d2 2
2 = d in the second equation
Where D2 = in the first equation and D
dx 2 dt 2

Case(i) : let k=0

AEs are m=0 amd m2=0 amd m=0,0 are the roots

Solutions are given by

T = c1e0t  c1 and X   c2 x  c3  e0 x   c2 x  c3 

Hence the solution of the PDE is given by

DEPT. OF MATHS/SJBIT Page 20

Advanced Calculus and Numerical Methods 18MAT21

U= XT= c1  c2 x  c3 

Or u(x,t) =Ax+B where c1 c2=A and c1 c3=B

Case (ii) let k be positive say k=+p2

AEs are m –c2p2=0 and m2 -p2=0

m= c2 p2 and m=+p

Solutions are given by

T  c'1ec andX  c '2e px  c '3e px

2
p 2t

Hence the solution of the PDE is given by

u  XT  c'1ec .( c'2e px  c '3e px )

2
p 2t

(A’ e px +B e  px ) where c1’c2’=A’ and c1’c3’=B’

2
p 2t
Or u(x,t) = c '1ec

Case (iii): let k be negative say k=-p2

AEs are m+ c2p2=0 and m2+p2=0

m=- c2 p2 and m=+ip

Solutions are given by

T  c''1e c and X  c ''2 cos px  c ''3 sin px

2
p 2t

Hence the solution of the PDE is given by

u  XT  c''1e c .(c''2 cos px  c''3 sin px)

2
p 2t

u ( x, t )  e c ( A'' cos px  B '' sin px)

2
p 2t

1. Solve the wave equation utt=c2uxx subject to the conditions u(t,0)=0 ,u(l,t)=0,
u
 x, 0   0 and u(x,0) =u0sin3(𝜋x/l)
t
 n x n ct
Soln: u  x, t    bn sin cos
n 1 l l

DEPT. OF MATHS/SJBIT Page 21

Advanced Calculus and Numerical Methods 18MAT21

Consider u(x,0) =u0sin3(𝜋x/l)

 n x
u  x, 0    bn sin
n 1 l

x  n x
u0 sin 3   bn sin
l n 1 l

3  x 1 3 x   n x
u0  sin 3  sin    bn sin
4 l 4 l  n 1 l

3u0  x u0 3 x x 2 x 3 x
sin  sin  b1 sin  b2 sin  b3 sin
4 l 4 l l l l
comparing both sides we get

3u0 u
b1  , b2  0 , b3  0 , b4  0 b5  0 ,
4 4

Thus by substituting these values in the expanded form we get

3u0 x  ct u0 3 x 3 ct
u ( x, t )  sin cos  sin cos
4 l l 4 l l

2. Solve the wave equation utt=c2 utt subject to the conditions u(t,0)=0 ,u(l,t)=0,
u
 x, 0   0 when t=0and u(x,0) =f(x)
t
 n x n ct
Soln: u  x, t    bn sin cos
n 1 l l

Consider u(x,0)=f(x) then we have

 n x
Consider u(x,0) =  bn sin
n 1 l
 n x
F(x) =  bn sin
n 1 l

The series in RHS is regarded as the sine half range Fourier series of f(x) in (0,l) and hence

DEPT. OF MATHS/SJBIT Page 22

Advanced Calculus and Numerical Methods 18MAT21

n x
l
2
bn  
l 0
f ( x) sin
l
dx

Thus we have the required solution in the form

 n x n ct
u  x, t    bn sin cos
n 1 l l

DEPT. OF MATHS/SJBIT Page 23