Lecture Notes

Unit IV: Applications of Partial Differential

Equations
Dr. Madhukar Sharma

1 CLASSIFICATION OF PARTIAL DIFFERENTIAL

EQUATIONS
Consider the Equation

a
m
 
∂2u ∂2u ∂2u ∂u ∂u
ar
A 2 +B +C 2 +F x, y, u, , = 0. (1)
∂x ∂x∂y ∂y ∂x ∂y
Sh

Where A, B, C may be constants or functions of x and y. The equation (1) is

r
ka

1. Parabolic; if B 2 − 4AC = 0
hu

2. Elliptic; if B 2 − 4AC < 0

ad

3. Hyperbolic; if B 2 − 4AC > 0.

rM

Example 1: Classify the type of PDE:

D

∂2u ∂2u ∂2u

+ 2 + 5 = 0,
∂x2 ∂x∂y ∂y2
whether it is parabolic, elliptic or hyperbolic? [2023]
Solution: Comparing the given equation with the most general second order PDE,
 
∂2u ∂2u ∂2u ∂u ∂u
A 2 +B + C 2 + F x, y, u, , = 0.
∂x ∂x∂y ∂y ∂x ∂y

We have, A = 1, B = 2 and C = 5.
Now,

B 2 − 4AC = 22 − 4 × 5
= −16 < 0

Therefore, the given equation is Elliptic.

1
Example 2: Show that the equation

∂2u 2
2∂ u
= c ,
∂t2 ∂x2
is hyperbolic?
Solution: The given equation may be written as

∂2u 2
2∂ u
− c =0
∂t2 ∂x2
Comparing it with the most general second order PDE,
 
∂2u ∂2u ∂2u ∂u ∂u
A 2 +B + C 2 + F t, x, u, , = 0.
∂t ∂t∂x ∂x ∂t ∂x

We have, A = 1, B = 0 and C = −c2 .

Now,

a
B 2 − 4AC = 02 − 4 × (−c2 )
= m
4c2 > 0, Always positive
ar
Sh

Therefore, the given equation is Hyperbolic.

r

Example 3: Classify the equation:

ka

∂2u 2
2 ∂ u
2
2 ∂ u
hu

(1 + x2 ) + (5 + 2x ) + (4 + x ) = 0.
∂x2 ∂x∂t ∂t2
ad

Solution: Here, A = 1 + x2 , B = 5 + 2x2 and C = 4 + x2 .

rM

Now,
D

B 2 − 4AC = (5 + 2x2 )2 − 4(1 + x2 )(4 + x2 )

= 25 + 20x2 + 4x4 − 4(4 + x2 + 4x2 + x4 )
= 25 + 20x2 + 4x4 − 16 − 20x2 − 4x4
= 9>0

Therefore, the given equation is Hyperbolic.

2 METHOD OF SEPARATION OF VARIABLES

In this method, we assume that the required solution is the product of two functions
i.e.,

u(x, y) = X(x)Y (y). (1)

Then we substitute the value u(x, y) from (1) and its derivatives reduces the P.D.E.
to the form

2
 f1 (X, X  , . . .) = f2 (Y, Y  , . . .) (2)
which is separable in X and Y . Since f2 (Y, Y  , . . .) is a function of Y only and
f1 (X, X  , . . . is a function of X only, then equation (2) must equal to a common con-
stant say k. Thus equation (2) reduces to

f1 (X, X  , . . .) = f2 (Y, Y  , . . .) = k.
Example 1: Using the method of separation of variables solve

∂u ∂u
=2 + u,
∂x ∂t
where u(x, 0) = 6e−3x .
Solution: We have
∂u ∂u
=2 +u (1)
∂x ∂t
Let

a
u = X(x)T (t),
m
ar (2)
is the complete solution, where Xis the function of x only and T is the function
 of t
Sh
∂u ∂X dX ∂X dX
only. Then = T = T = As X is the function of x alone
∂x ∂x dx ∂x dx
r

∂u ∂T dX
ka

and =X =X .
∂t ∂t dt
hu

Putting these values in equation (1)

ad
rM

dX dT
T = 2X + XT
dx dt
1 dX 2 dT
D

⇒ = +1
X dx T dt
1 dX 2 dT
⇒ = +1=k
X dx T dt
Now
1 dX
=k
X dx
dX
⇒ = kdx (3)
X
On integrating, we get

loge X = kx + loge c1
X
⇒ loge = kx
c1
⇒X = c1 ekx .

3
 And again from (3), we get
2 dT
+1 = k
T dt
2 dT
⇒ = k−1
T dt
dT k−1
⇒ = t.
T 2
On integration, we obtain
k−1
log T = t + log c2
2
T k−1
⇒ log = t
c2 2
c e( 2 )t .
k−1
⇒T = 2

Putting the values of X and T in equation (2), we get

a
u = c1 ekx .c2 e( 2 )t m
k−1
ar
1
Sh

⇒ u = c1 c2 ekx+ 2 (k−1)t . (4)

r

On putting t = 0 in (4), we get

ka
hu

u(x, 0) = c1 c2 ekx
ad

⇒ 6e−3x = c1 c2 ekx
rM

⇒ c1 c 2 = 6 and k = −3.
D

Hence the required solution is

1
u = 6e−3x+ 2 (−3−1)t = 6e−3x−2t
⇒u = 6e−3x e−2t . Ans.

Example 2: Solve by the method of separation of variables

∂u ∂u
=2 + u,
∂x ∂y
where u(x, 0) = 3e−5x − 2e−3x.
Solution: We have
∂u ∂u
=2 +u (1)
∂x ∂y
Let
u = X(x)Y (y) (2)

4
is the complete solution of equation (1). Then
∂u ∂ ∂X dX
= (XY ) = Y = Y
∂x ∂x ∂x dx
 
∂X dX
= , as X is the function of x alone
∂x dx
Similarly,
∂u ∂ ∂Y dY
= (XY ) = X =X .
∂y ∂y ∂y dy
Putting these values in equation (1), we have
dX dY
Y = 2X + XY
dx dy
1 dX 2 dY
⇒ = +1
X dx Y dy
1 dX 2 dY
⇒ = +1= k (3)

a
X dx Y dy
Now
m
ar
1 dX
Sh

= k
X dx
r

dX
ka

⇒ = kdx
X
hu

On integrating, we get
ad

loge X = kx + loge c1
rM

X
⇒ loge = kx
c1
D

⇒X = c1 ekx .
And taking last two terms of equation (3), we get
2 dY
+1 = k
Y dy
2 dY
⇒ = k−1
Y dy
dY k−1
⇒ = y.
Y 2
On integrating,
k−1
log Y = y + log c2
2
Y k−1
⇒ log = y
c2 2
c e( 2 )y .
k−1
⇒Y = 2

5
 Putting the values of X and Y in equation (2), we get

c1 ekx.c2 e( 2 )y
k−1
u =
1
⇒u = c1 c2 ekx+ 2 (k−1)y
∞
1
⇒u = bn ekn x+ 2 (kn −1)y . (4)
n=1

(bn = c1 c2 ) and (k = kn )
Which is the most general solution of equation (1).
On putting y = 0 in (4), we get

∞

u(x, 0) = bn ekn x
n=1

⇒ 3e−5x − 2e−3x = b1 ek1 x + b2 ek2 x .

a
Comparing the terms on both sides, we get
m
ar
b1 = 3, b2 = −2, k1 = −5, k2 = −3.
Sh

Hence the required solution of given equation is

r
ka

u = 3e− 5x + 3y − 2e− 3x + 2y. Ans.

hu
ad

3 ONE DIMENSIONAL WAVE EQUATION

rM

3.1 Solution of One Dimensional Wave Equation

D

The wave equation is

∂2y ∂2y
= . (1)
∂t2 ∂x2
Let
y = X(x)T (t) (2)
where X is the function of x only and T is the function of t only.
Then

∂2y d2 T ∂2y d2 X
= X and = T .
∂t2 dt2 ∂x2 dx2
Putting these values in equation (1), we get
d2 T 2
2 d X
X = c T
dt2 dx2
1 d2 X 1 d2 T
= = k Let (3)
X dx2 c2 T dt2

6
 Now

1 d2 X
= k
X dx2
d2 X
⇒ − kX = 0
dx2
d
⇒ (D2 − k)X = 0; ≡D
dx
The A.E. is

m2 − k = 0
√
⇒m = ± k.

Hence
√ √
kx
X = c1 e + c2 e− kx

a
and, again from (3), we get

m
ar
1 d2 T
= k
Sh

c2 T dt2
d2 T
kc2 T
r

⇒ 2 =
ka

dt
2
d T
hu

⇒ 2 − kc2 T = 0
dt
ad

The A.E. is
rM

m2 − kc2 = 0
√
D

⇒m = ±c k.

Hence
√ √
T = c3 ec kt
+ c2 e−c kt
.
Thus, from equation (2), we get
√ √ √ √
kx
y = (c1 e + c2 e− kx
)(c3 ec kt
+ c2 e−c kt
).

There are arise following cases:

Case I: If k > 0, let k = p2 , then

y = (c1 epx + c2 e−px )(c3 ecpt + c2 e−cpt ). (4)

Case II: If k < 0, let k = −p2 , then

m2 = −p2 ⇒ m = ±pi

7
 therefore X = c1 cos px + c2 sin px
and
m2 = −p2 c2 ⇒ m = ±cpi

T = c3 cos cpt + c4 sin cpt

then
y = (c1 cos px + c2 sin px)(c3 cos cpt + c4 sin cpt). (5)
Case III: If k = 0, then

D2 X = 0 ⇒ m = 0, 0
therefore
X = (c1 + c2 x).
And
D2 T = 0 ⇒ m = 0, 0

a
m
therefore ar
T = (c3 + c4 t).
Sh

then
y = (c1 + c2 x)(c3 + c4 t). (6)
r
ka

Of these three solutions, we have choose the solution which is consistent with the
hu

physical nature of the problem.

Since the physical nature of one dimension wave equation is periodic so we con-
ad

sider the solution which has periodic nature.

rM

Here only the solution in equation (5) is periodic (as both sine and cosine are peri-
odic). thus the desired solution, for one dimensional wave equation is
D

y = (c1 cos px + c2 sin px)(c3 cos cpt + c4 sin cpt).

Example 3: A string is stretched and fastened  point’s l apart. Motion is started

to two
 πx
by displacing the string in the form y = a sin , from which it is released at a time
l
t = 0. Show that the displacement of any point at a distance x from one end at time t
is given by  
 πx  πct
y(x, t) = a sin cos .
l l
Solution: Let y(x, t) be the displacement at any point P (x, y) at any time t. Then by
the wave equation, we have
∂2y 2
2∂ y
= c (1)
∂t2 ∂x2
The solution of equation (1) is of the form
y(x, t) = (c1 cos px + c2 sin px)(c3 cos cpt + c4 sin cpt) (2)
Now using the boundary conditions

8
 (i) At x = 0, the displacement y = 0 ⇒ y(0, t) = 0 (3)
(ii) At x = l, the displacement y = 0 ⇒ y(l, t) = 0 (4)
Using boundary condition (i) in equation (2), we get

y(0, t) = (c1 cos 0 + c2 sin 0)(c3 cos cpt + c4 sin cpt)

⇒0 = c1 (c3 cos cpt + c4 sin cpt)
⇒ c1 = 0.

From (2), we get

y(x, t) = c2 sin px(c3 cos cpt + c4 sin cpt) (5)

Using (ii) boundary condition in equation (5), we get

y(x, l) = c2 sin pl(c3 cos cpt + c4 sin cpt)

a
⇒0 = c2 sin pl(c3 cos cpt + c4 sin cpt)
⇒ sin pl = 0
m
ar
⇒ sin pl = sin nπ
Sh

⇒ pl = nπ
r

nπ
⇒p =
ka

.
l
hu

Now the initial conditions

 
ad

∂y ∂y
(iii) At t = 0, the velocity =0⇒ =0 (6)
rM

∂t ∂t t=0
πx πx
D

(iv) At t = 0, the displacement y = a sin ⇒ y(x, 0) = a sin (7)

l l
From (5), we have

∂y
= c2 sin px [c3 (−cp) sin cpt + c4 (cp) cos cpt]
∂t
Using initial condition (6) in above equation, we get

0 = c2 c4 cp sin px
⇒0 = c2 c4 cp
⇒ c4 = 0, c2 = 0 otherwise there is trivial solution
nπ
Using p = and c4 = 0 in equation (5), we get
l
nπx nπct
y(x, t) = c2 c3 sin cos
l l
The general solution is

9
 ∞
 nπx nπct
y(x, t) = bn sin cos (8)
n=1
l l

Finally using initial condition (7) in equation (8), we get

∞
 nπx
y(x, 0) = bn sin
n=1
l
πx πx 2πx
⇒ a sin = b1 sin + b2 sin +···
l l l
⇒ b1 = 0, b2 = b3 = · · · = 0.

Hence the required solution of given problem is

πx πct
y(x, t) = a sin cos P roved.
l l

a
m
Example 4: A string is stretched and fastened to two point l apart. Motion is started
ar
by displacing the string into the form y = k(lx − x2 ) from which it is released at time
t = 0. Find the displacement of any point on the string at a distance of x from one end
Sh

at a time t.
r

Example 5: Find the displacement of a finite string of length l that is fixed at both
ka

ends and is released from rest with both ends and is released from rest with an initial
hu

displacement f(x). Solution: Let y(x, t) be the displacement at any point P (x, y) at
ad

any time t. Then by the wave equation, we have

rM

∂2y 2
2∂ y
= c (1)
∂t2 ∂x2
D

The solution of equation (1) is of the form

y(x, t) = (c1 cos px + c2 sin px)(c3 cos cpt + c4 sin cpt) (2)

Now using the boundary conditions

(i) At x = 0, the displacement y = 0 ⇒ y(0, t) = 0 (3)

(ii) At x = l, the displacement y = 0 ⇒ y(l, t) = 0 (4)

Putting x = 0 in equation (2), we get

y(0, t) = (c1 cos 0 + c2 sin 0)(c3 cos cpt + c4 sin cpt)

⇒0 = c1 (c3 cos cpt + c4 sin cpt)
⇒ c1 = 0.

From (2), we get

y(x, t) = c2 sin px(c3 cos cpt + c4 sin cpt) (5)

10
 Putting x = l in equation (5), we get

y(x, l) = c2 sin pl(c3 cos cpt + c4 sin cpt)

⇒0 = c2 sin pl(c3 cos cpt + c4 sin cpt)
⇒ sin pl = 0
⇒ sin pl = sin nπ
⇒ pl = nπ
nπ
⇒p = .
l
The intial conditions are:
 
∂y ∂y
(iii) At t = 0, the velocity =0⇒ =0 (6)
∂t ∂t t=0

(iv) At t = 0, the displacement y = f(x) ⇒ y(x, 0) = f(x) (7)

From (5), we have

a
m
∂y
= c2 sin px [c3 (−cp) sin cpt + c4 (cp) cos cpt]
ar
∂t
Sh

Putting t = 0, we have
 
r

∂y
ka

= c2 sin px [c3 (−cp) sin 0 + c4 (cp) cos 0]

∂t t=0
hu
ad

Using initial condition (6), we get

rM

0 = c2 c4 cp sin px
⇒0 = c2 c4 cp
D

⇒ c4 = 0, c2 = 0 otherwise there is trivial solution

nπ
Putting p = and c4 = 0 in equation (5), we get
l
 
nπx nπct
y(x, t) = c2 sin c3 cos +0
l l
nπx nπct
= c2 c3 sin cos
l l
The general solution is

∞
 nπx nπct
y(x, t) = bn sin cos (8)
n=1
l l

On putting t = 0 in equation (8), we get

11
 ∞
 nπx
y(x, 0) = bn sin cos 0
n=1
l
∞
nπx
⇒ f(x) = bn sin , Using (7)
n=1
l

Which represents the half range Fourier Series Sine Series, so

l
2 nπx
bn = f(x) sin dx
l 0 l
On putting the value of bn in equation (8), we get the required solution.

4 ONE DIMENSIONAL HEAT FLOW

a
4.1 Solution of One Dimensional Heat Equation
m
ar
We know that the heat equation
∂u ∂2u
Sh

= c2 2 (1)
∂t ∂x
Let u(x, t) = X(x)T (t) (2)
r
ka

is complete solution of equation (1).

hu

Where X = X(x) and T = T (t).

∂u dX ∂2u d2 X
ad

Then =T or = T
∂x dx ∂x2 dx2
rM

∂u dX
and =X .
∂t dt
Putting these values in equation (1), we get
D

dT d2 X
X = c2 T
dt dx2
2
1 d X 1 dT
⇒ =
X dx2 c2 T dt
1 d2 X 1 dT
⇒ = = k, Let (3)
X dx2 2
c T dt
1 d2 X 1 dT
Now = k and =k
X dx2 c2 T dt

d2 X
− kX = 0
dx2
dT
and − kc2 T = 0. (4)
dt

12
There arises following cases:
Case I: If k > 0, let k = p2 , from equation (4), we have

d2 X
− p2 X = 0
dx2
dT
and − c2 p 2 T = 0.
dt
On solving, we have

X = c1 epx + c2 e−px
2 2
and T = c3 ec p t .

From equation (2), we have

2 2
u = c1 epx + c2 e−px .c3 ec p t
(5)

Case II: If k < 0, let k = −p2 , from equation (4), we have

a
d2 X
+ p2 X = m
0
ar
dx 2
Sh

dT
and + c2 p 2 T = 0.
dt
r
ka

On solving, we have
hu

X = c1 cos px + c2 sin px
ad

2 2
and T = c3 e−c p t
.
rM

From equation (2), we have

D

2 2
u = (c1 cos px + c2 sin px) .c3 e−c p t
(6)

Case III: If k = 0, then from equation (4), we have

d2 X
= 0
dx2
dT
and = 0.
dt
On solving, we have

X = c 1 x + c2
and T = c3 .

From equation (2), we have

u = (c1 x + c2 ) .c3 (7)

13
 Here only the solution is given by Case II, the temperature reduces with time and
u → 0 as t → ∞, which is the physical nature of heat flow.
Hence the required solution is given by
2 2
u(x, t) = (c1 cos px + c2 sin px) .c3 e−c p t
.

∂u ∂2u
Example 6: Determine the solution of one-dimensional heat equation = c2 2
∂t ∂x
subject to the boundary conditions u(0, t) = 0, u(l, t) = 0, (t > 0) and the initial con-
dition u(x, 0) = x, l being the length of the bar.
Solution: We have,
2
∂u 2∂ u
=c (8)
∂t ∂x2
We know that the solution of equation (1) is given by
2 2
u(x, t) = (c1 cos px + c2 sin px) .c3 e−c p t
. (9)

a
Boundary conditions are:
m
ar
(i) u(0, t) = 0
Sh

(ii) u(l, t) = 0
r
ka
hu

Initial condition is:

ad

(iii) u(x, 0) = x
rM

Putting x = 0 in equation (2), we get

D

2 2
u(0, t) = (c1 cos 0 + c2 sin 0) .c3 e−c p t
2 2
0 = c1 c3 e−c p t

c1 = 0

Putting c1 = 0 in equation (2), we have

2 2
u(x, t) = c2 c3 sin pxe−c p t
(6)

Putting x = l in equation (6), we have

2 2
u(l, t) = c2 c3 sin ple−c p t
2 2
0 = c2 c3 sin ple−c p t

sin pl = 0
pl = nπ
nπ
p =
l

14
 On putting in equation (6), we have
nπx − n2 π22 c2 t
u(x, t) = c2 c3 sin ·e l
l
nπx − n2 π22 c2 t
= bn sin ·e l , let c2 c3 = bn
l
Hence the general solution of equation (1), is
∞
 nπx − n2 π22 c2 t
u(x, t) = bn sin ·e l (7)
n=1
l

Putting t = 0, in equation (7), we have

∞
 nπx 0
u(x, 0) = bn sin ·e
n=1
l
∞
nπx
x= bn sin , Using equation (5)

a
l

m
n=1 ar
Which represents half range Fourier Sine Series, so
Sh

l
2 nπx
bn = x sin
r

dx
ka

l 0 l
l
1  nπx 
hu

2 1 nπx
= x · nπ − cos −1· − · sin
l l nπ 2 l
ad

l l 0
    2  l
rM

2 lx nπx l nπx
= − cos + 2 2
· sin
l nπ l n π l 0
  2  2  
D

2 l l
= − cos nπ + sin nπ − 0
l nπ n2 π 2
  2 
2 l n
= − (−1)
l nπ
2l
= (−1)n+1
nπ
Putting the value of bn in equation (7), we get the required solution
∞
 2l  nπx  n2 π 2 c2 t
u(x, t) = (−1) n+1
sin · e− l2
n=1
nπ l

Example 7: Find the temperature in a bar of length 2 whose end are kept at zero and
lateral surface is insulated if the initial temperature is
 πx   
5πx
sin + 3 sin .
2 2

15
 Solution: We have, One Dimensional Heat Equation
2
∂u 2∂ u
=c (1)
∂t ∂x2
We know that the solution of equation (1) is given by
2 2
u(x, t) = (c1 cos px + c2 sin px) .c3 e−c p t
. (2)

Boundary conditions are:

(i) At x = 0, u = 0, i.e.,
u(0, t) = 0 (3)

(ii) At x = 2, u = 0, i.e.,
u(2, t) = 0 (4)

The given Initial condition is:

πx 5πx
(iii) At t = 0, u = sin + 3 sin , i.e.,

a
2 2

m
 πx   
5πx
ar
u(x, 0) = sin + 3 sin (5)
2 2
Sh

Putting x = 0 in equation (2), we get

r
ka

2 2
u(0, t) = (c1 cos 0 + c2 sin 0) .c3 e−c p t
hu

2 2
0 = c1 c3 e−c p t
ad

, Using Equation (3)

c1 = 0
rM

Putting c1 = 0 in equation (2), we have

D

2 2
u(x, t) = c2 c3 sin px · e−c p t
(6)

Putting x = 2 in equation (6), we have

2 2
u(2, t) = c2 c3 sin 2p · e−c p t
2 2
0 = c2 c3 sin 2p · e−c p t
, Using equation (4)
sin 2p = 0
2p = nπ
nπ
p =
2
Putting the value of p in equation (6), we have
 nπx  n2 π 2 c2 t
u(x, t) = c2 c3 sin · e− 4
2
 nπx  n2 π 2 c2 t
= bn sin · e− 4 , let c2 c3 = bn
2

16
Hence the general solution of equation (1), is
∞
  nπx  n2 π 2 c2 t
u(x, t) = bn sin · e− 4 (7)
n=1
2

Putting t = 0, in equation (7), we have

∞
  nπx 
u(x, 0) = bn sin · e0
n=1
2
 πx    ∞  nπx 
5πx
sin + 3 sin = bn sin , Using equation (5)
2 2 n=1
2
 πx     πx     
5πx 2πx 3πx
sin + 3 sin = b1 sin + b2 sin + b3 sin +···
2 2 2 2 2

On comparing both sides, we have b1 = 0, b5 = 3 and b2 = b3 = b4 = b6 = · · · =

0. Putting these values in equation (7), we have

a
 πx   
m
2 2 2
− 1 π4 c t 5πx 52 π 2 c2 t
u(x, t) = b1 sin ·e + · · · + b5 sin · e− 4 + · · ·
ar
2 2
 
Sh

 πx  π 2 c2 t 5πx 52 π 2 c2 t
= sin · e− 4 + 3 sin · e− 4
2 2
r
ka
hu

5 TWO DIMENSIONAL HEAT FLOW

ad

5.1 Two Dimensional Heat Equation

rM

The Two Dimensional Heat Equation equation is

D

 2 
∂u 2 ∂ u ∂2u
=c + 2 (1)
∂t ∂x2 ∂y
∂u
In steady state, u always independent of time, t, so that ∂t = 0. Hence from equation
(1), we get
∂2u ∂2u
+ 2 = 0. (2)
∂x2 ∂y
Equation (2) is known as the Laplace Equation.

5.2 General Solution Solution of Laplace Equation in Two Dimen-

sions
The Laplace Equation in 2D is

∂2u ∂2u
+ 2 = 0. (1)
∂x2 ∂y

17
Let
u = XY (2)
is complete solution of equation (1), where X and Y are functions of only x and y
respectively, i.e., X = X(x) and Y = Y (y). From equation (2), we have
∂u dX
=Y
∂x dx
∂u dY
=X
∂y dy
∂2u d2 X
=Y
∂x2 dx2
∂2u d2 Y
=X 2
∂y2 dy
Putting these values in equation (1), we get

d2 X d2 Y
Y +X 2 =0

a
dx2 dy
2
1 d X 1 d2 Y m
ar
+ =0 (Dividing both sides by XY )
X dx2 Y dy2
Sh

1 d2 X 1 d2 Y
=− = k (Let) (3)
r

X dx2 Y dy2
ka
hu

From equation (3), we have

ad

1 d2 X 1 d2 Y
= k and − =k
rM

X dx2 Y dy2
d2 X d2 Y
− kX = 0 and + kY = 0 (4)
D

dx2 dy2

Case I: If k > 0, let k = p2 , from equation (4), we have

d2 X 2 d2 Y
2
− p X = 0 and 2
+ p2 Y = 0
dx dy
On solving, we have

X = c1 epx + c2 e−px and Y = c3 cos py + c4 sin py

Putting these values in equation (2), we have

u = c1 epx + c2 e−px (c3 cos py + c4 sin py) (5)

Case II: If k < 0, let k = −p2 , from equation (4), we have

d2 X 2 d2 Y
2
+ p X = 0 and 2
− p2 Y = 0
dx dy

18
On solving, we have
X = c1 cos px + c2 sin px and Y = c3 epy + c4 e−py
Putting these values in equation (2), we have
u = (c1 cos px + c2 sin px) c3 epy + c4 e−py (6)
Case III: If k = 0, from equation (4), we have
d2 X d2 Y
= 0 and =0
dx2 dy2
On solving, we have
X = c1 x + c2 and Y = c3 y + c4
Putting these values in equation (2), we have
u = (c1 x + c2 ) (c3 y + c4 ) (7)

a
Of these three solutions, we have to choose that solution which is consistent with the

m
physical nature of the problem and the given boundary conditions. Solution given by
ar
equation (6) is required solution.
Sh

u(x, t) = (c1 cos px + c2 sin px) (c3 epy + c4 e−py )

r
ka

Example 1: Solve the following Laplace equation

hu

∂2u ∂2u
+ 2 = 0.
ad

∂x2 ∂y
rM

in rectangle with u(0, y) = 0, u(a, y) = 0, u(x, b) = 0 and u(x, 0) = f(x) along

x-axis.
D

Solution: Given Laplace equation is

∂2u ∂2u
+ 2 =0 (1)
∂x2 ∂y
Let
u = XY (2)
is complete solution of equation (1), where X and Y are functions of only x and y
respectively, i.e., X = X(x) and Y = Y (y). From equation (2), we have
∂u dX
=Y
∂x dx
∂u dY
=X
∂y dy
∂2u d2 X
=Y
∂x2 dx2
∂2u d2 Y
=X 2
∂y2 dy

19
Putting these values in equation (1), we get
d2 X d2 Y
Y + X =0
dx2 dy2
1 d2 X 1 d2 Y
+ =0 (Dividing both sides by XY )
X dx2 Y dy2
1 d2 X 1 d2 Y
= − = −p2 (Let) (3)
X dx2 Y dy2
From equation (3), we have
1 d2 X 2 1 d2 Y
= −p and − = −p2
X dx2 Y dy2
d2 X 2 d2 Y
2
+ p X = 0 and 2
− p2 Y = 0
dx dy
On solving, we have
X = c1 cos px + c2 sin px and Y = c3 epy + c4 e−py

a
Putting these values in equation (2), we have
m
ar
u(x, y) = (c1 cos px + c2 sin px) c3 epy + c4 e−py
Sh

(4)
r

Given boundary conditions are:

ka

(i) u(a, y) = 0 (5)

hu

(ii) u(a, y) = 0 (6)

ad

(iii) u(x, b) = 0 (7)

rM

(iv) u(x, 0) = f(x) (8)

D

Putting x = 0 in equation (4), we have

u(0, y) = (c1 cos 0 + c2 sin 0) c3 epy + c4 e−py
0 = c1 c3 epy + c4 e−py , Using equation (5)
c1 = 0
Putting the value of c1 in equation (4), we have
u(x, y) = c2 sin px c3 epy + c4 e−py (9)
Putting x = a in equation (9), we get
u(a, y) = c2 sin pa c3 epy + c4 e−py
0 = c2 sin pa c3 epy + c4 e−py , Using equation (6)
sin pa = 0
pa = nπ
nπ
p=
a

20
Putting the value of p in equation (9), we have
nπx  nπy − nπy

u(x, y) = c2 sin c3 e a + c4 e a (10)
a
Putting y = b in equation (10), we get
nπx  nπb − nπb

u(x, b) = c2 sin c3 e a + c4 e a
a  
nπx nπb
− nπb
0 = c2 sin c3 e a + c4 e a , Using (7)
a
nπb nπb
c3 e a + c4 e− a = 0
nπb nπb
c4 e− a = −c3 e a

2nπb
c4 = −c3 e a

Putting the value in equation (10), we have

nπx  nπy 2nπb
− nπy

u(x, y) = c2 sin c3 e − c3 e e

a
a a a
a
m

nπx nπy 
− nπ (y−2b)
u(x, y) = c2 c3 sin −e
ar
e a a
a
Sh

The general solution is given by

r
ka

nπx  nπy 
∞

− nπ (y−2b)
u(x, y) = bn sin e a −e a , Let c2 c3 = bn (11)
hu

n=1
a
ad

Putting y = 0 in equation (11),

rM

nπx  
∞
 2nπb
u(x, 0) = bn sin 1−e
D

n=1
a

That represents the half range Fourier Sine series, therefore

 2nπb
 2 a
nπx
bn 1 − e a = f(x) sin dx
a 0 a
a
2 nπx
bn =  2nπb
 f(x) sin dx (12)
a 1−e a 0 a

On putting the value of bn in equation (11), we get the required solution.

Example 2: Solve the Laplace equation

∂2u ∂2u
+ 2 = 0,
∂x2 ∂y

in a rectangle in the xy-planne with u(x, 0) = 0, u(x, b) = 0, u(0, y) = 0 and

u(a, y) = f(y) parallel to y-axis.

21
 Solution: Given Laplace equation is
∂2u ∂2u
+ 2 =0 (1)
∂x2 ∂y
Let
u = XY (2)
is complete solution of equation (1), where X and Y are functions of only x and y
respectively, i.e., X = X(x) and Y = Y (y). From equation (2), we have
∂u dX
=Y
∂x dx
∂u dY
=X
∂y dy
∂2u d2 X
=Y
∂x2 dx2
∂2u d2 Y
=X 2
∂y2 dy

a
Putting these values in equation (1), we get
m
ar
d2 X d2 Y
Sh

Y +X 2 =0
dx2 dy
r

2
1 d X 1 d2 Y
ka

+ =0 (Dividing both sides by XY )

X dx2 Y dy2
hu

1 d2 Y 1 d2 X
=− = −p2 (Let) (3)
ad

Y dy 2 X dx 2
rM

From equation (3), we have

1 d2 Y 1 d2 X
D

2
2
= −p and − 2
= −p2
Y dy X dx
2 2
d Y 2 d X
2
+ p Y = 0 and 2
− p2 X = 0
dy dx
On solving, we have
Y = c1 cos py + c2 sin py and X = c3 epx + c4 e−px
Putting these values in equation (2), we have
u(x, y) = (c1 cos py + c2 sin py) c3 epx + c4 e−px (4)
Given boundary conditions are:
(i) u(x, 0) = 0 (5)
(ii) u(x, b) = 0 (6)
(iii) u(0, y) = 0 (7)
(iv) u(a, y) = f(y) (8)

22
Putting y = 0 in equation (4), we have
u(x, 0) = (c1 cos 0 + c2 sin 0) c3 epx + c4 e−px
0 = c1 c3 epx + c4 e−px , Using equation (5)
c1 = 0
Putting the value of c1 in equation (4), we have
u(x, y) = c2 sin py c3 epx + c4 e−px (9)
Putting y = b in equation (9), we get
u(x, b) = c2 sin pb c3 epx + c4 e−px
0 = c2 sin pb c3 epx + c4 e−px , Using equation (6)
sin pb = 0
pb = nπ
nπ
p=
b

a
Putting the value of p in equation (9), we have
u(x, y) = c2 sin
nπy nπx

m
c3 e b + c4 e− b
nπx
ar
(10)
b
Sh

Putting x = 0 in equation (10), we get

nπy
r

u(0, y) = c2 sin c3 e0 + c4 e0
ka

a
nπy
hu

0 = c2 sin (c3 + c4 ) , Using (7)

a
ad

c 3 + c4 = 0
rM

c4 = −c3
Putting the value in equation (10), we have
D

nπy nπx nπx

u(x, y) = c2 sin c3 e b − c3 e− b
b
nπy nπx nπx
u(x, y) = c2 c3 sin e b − e− b
b
nπy nπx
u(x, y) = bn sin sinh , Let c22c3 = bn
b b
The general solution is given by
∞
 nπy nπx
u(x, y) = bn sin sinh (11)
n=1
b b
Putting x = a in equation (11), we have
∞
 nπy nπa
u(a, y) = bn sin sinh
n=1
b b
∞
nπa nπy
f(y) = bn sinh sin
n=1
b b

23
 That represents the half range Fourier Sine series, therefore
b
nπa 2 nπy
bn sinh = f(y) sin dy
b b 0 b
b
2 nπy
bn = f(y) sin dy. (12)
a sinh nπa
b 0 b

On putting the value of bn in equation (11), we get the required solution.

a
m
ar
r Sh
ka
hu
ad
rM
D

24