18MAB201T-U3-Lecture Notes

18MAB201T-Transforms and Boundary Value

Problems

S
N
A
H
IT
H
AT
Prepared by
F

Dr. S. ATHITHAN
O

Assistant Professor
S
TE

Department of of Mathematics
O

Faculty of Engineering and Technology

N

SRM INSTITUTE OF SCIENCE AND TECHNOLOGY

E

Kattankulathur-603203, Kancheepuram District.

R
TU
C
LE

SRM INSTITUTE OF SCIENCE AND TECHNOLOGY

Kattankulathur-603203, Kancheepuram District.
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Unit-3
ONE DIMENSIONAL WAVE & HEAT EQUATION
T OPICS :
? Classification of partial differential equations. Method of separation of variables. One
dimensional Wave Equation and its possible solutions
? Initial and Boundary value Problems with zero velocity - related problems
? Initial and Boundary value Problems with Nonzero velocity related problems
? One dimensional heat equation and its possible solutions

S
? Steady state conditions and zero boundary conditions- related problems

N
? Steady state conditions and Non-zero boundary conditions related problems

A
H
IT
H
AT
F
O
S
TE
O
N
E
R
TU
C
LE

Page 1 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

D EAR A LL , H ERE I HAVE SOLVED FEW PROBLEMS ONLY AND SOME TOPICS MAY BE
MISSED . P LEASE FOLLOW THE CLASSWORK TO HAVE ALL THE TOPICS FOR PREPARA -
TION . TAKE E XRECISE PROBLEMS GIVEN AT THE END FOR YOUR PRACTICE . A PART
FROM E XERCISE , YOU CAN FOLLOW ANY REFERENCE BOOK FOR YOUR PRACTICE .

S OME OF THE SECTIONS / TOPICS IN THIS UNIT ARE PRELIMINARY IDEAS WHICH ARE
BASICS NEEDED TO DO OUR REGULAR COURSE EXAMPLES AND EXERCISES .

1 What is a Partial Differential Equation?

S
You’ve probably all seen an ordinary differential equation (ODE); for example the pendulum
equation,

N
d2 Θ g

A
2
+ sin Θ = 0, (1.1)
dt L

H
IT
describes the angle, Θ, a pendulum makes with the vertical as a function of time, t. Here g and
L are constants (the acceleration due to gravity and length of the pendulum respectively), t is

H
the independent variable and Θ is the dependent variable. This is an ODE because there is
AT
only one independent variable, here t which represents time.
A partial differential equation (PDE) relates the partial derivatives of a function of two or more
F

independent variables together. For example, Laplace’s equation for Φ(x, y),
O

∂ 2Φ ∂ 2Φ
S

+ =0 (1.2)
∂x2 ∂y 2
TE

arises in many places in mathematics and physics. For simplicity, we will use subscript notation
O

for partial derivatives, so this equation can also be written Φxx + Φyy = 0.
N

We say a function is a solution to a PDE if it satisfy the equation and any side conditions given.
E

Mathematicians are often interested in if a solution exists and when it is unique.

R

Exercise: 1. Show that Φ1 = x and Φ2 = x2 − y 2 are solutions to Laplace’s equation (1.2).

TU

How can you combine them to create a new solution?

C
LE

Exercise: 2. Show that  

sin(y)
u(x, y) = ln
sin(x)
is a solution to the minimal surface equation,

(1 + u2y )uxx − 2ux uy uxy + (1 + u2x )uyy = 0, (1.3)

in the region 0 < x < π, 0 < y < π. What happens on the boundary of this region?
Suppose we consider a constant multiple of u(x, y) – is it still a solution of the PDE?

Page 2 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

2 Classifying PDE’s: Order, Linear vs. Nonlinear

When studying ODEs we classify them in an attempt to group similar equations which might
share certain properties, such as methods of solution. We classify PDE’s in a similar way. The
order of the differential equation is the highest partial derivative that appears in the equation.
So, for example Laplace’s Equation (1.2) is second-order.
Some other examples are the convection equation for u(x, t),

ut + Cux = 0, (2.1)

which is first-order. Here C is the wave speed. The minimal surface equation,

S
N
(1 + u2y )uxx − 2ux uy uxy + (1 + u2x )uyy = 0, (2.2)

A
describes an area minimizing surface, u(x, y), and is second-order. Finally, the Korteweg-

H
deVries equation (sometimes called KdV),

IT
ut + 6uux = uxxx (2.3)

H
AT
is a model of the amplitude of a wave, u(x, t), on the surface of a fluid and is third-order.
We also define linear PDE’s as equations for which the dependent variable (and its derivatives)
F

appear in terms with degree at most one. Anything else is called nonlinear. So, for example,
O

the most general first-order linear PDE for u(x, t) would be

S

a(x, t)ut + b(x, t)ux + c(x, t)u = d(x, t), (2.4)

TE

where a, b, c and d are known functions (called coefficients).

O
N

Exercise: 1. Which of Laplace’s equation (1.2), the convection equation (2.1), the minimal
surface equation (2.2) and the Korteweg-deVries equation (2.3) are linear?
E
R
TU

Exercise: 2. Write down the most general constant coefficient linear second-order equation for
C

Φ(x, y).
LE

Page 3 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

3 Homogeneous PDE’s and Superposition

Linear equations can further be classified as homogeneous for which the dependent variable
(and it derivatives) appear in terms with degree exactly one, and non-homogeneous which may
contain terms which only depend on the independent variable. So, the convection equation

ut + cux = 0

is homogeneous, but its cousin, the general first-order linear PDE for u(x, t), is non-homogeneous

a(x, t)ut + b(x, t)ux + c(x, t)u = d(x, t),

S
unless d(x, t) = 0.

N
Because partial differentiation is distributive, you can quickly convince yourself that if two

A
solutions, say u1 and u2 , satisfy a linear homogeneous PDE, that any linear combination of

H
them

IT
u = c1 u1 + c2 u2 (3.1)

H
is also a solution. So, for example, since AT
Φ1 = x2 − y 2 Φ2 = x
F

both satisfy Laplace’s equation, Φxx + Φyy = 0, so does any linear combination of them
O

Φ = c1 Φ1 + c2 Φ2 = c1 (x2 − y 2 ) + c2 x.
S
TE

This property is extremely useful for constructing solutions which satisfy certain initial condi-
tions and boundary conditions.
O
N
E
R
TU
C
LE

Page 4 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

4 The Transport Equation

One of the driving motivations for studying PDE’s is to describe the physical world around us.
We can use a flux argument to derive equations describing the evolution of a density, which is
just a fancy word describing the concentration of something (mass in a region, heat in a metal
bar, traffic on a highway) per unit volume.
Consider a one-dimensional freeway and let ρ(x, t) be the density of cars per unit length on
the freeway.

S
N
A
H
IT
H
AT
F
O
S

Figure 1.1: Flux argument for cars on a freeway.

TE

(draw your own figure).

O

Then the mass of cars in the region a < x < b is given by

N

Z b
M = ρ(x, t) dx . (4.1)
E

a
R

Now suppose we are measuring the flux, Q, of cars into this region measured in mass/unit time.
TU

It can written in terms of the number of cars crossing into the region at x = a, called q(a),
minus the number of cars that flow out of the region at x = b, called q(b),
C

Q = q(a) − q(b). (4.2)

LE

Now, by conservation of mass, the rate of change of the mass between a and b is given by the
flux into the region,
dM
= Q. (4.3)
dt
We can rewrite the flux by a clever application of the fundamental theorem of calculus:
Z b
x=b
Q = q(a, t) − q(b, t) = − q(x, t)|x=a = − qx dx . (4.4)
a

We can now rewrite the conservation of mass equation as

d b
Z b Z b
dM
Z
= ρ dx = ρt dx = Q = − qx dx, (4.5)
dt dt a a a

Page 5 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

or, rearranging Z b
ρt + qx dx = 0. (4.6)
a
Since this is true for every interval a < x < b, the integrand must vanish identically. So

ρt + qx = 0. (4.7)

Equations of this form are called transport equations or conservation laws – they are a very
active area of study in PDE’s.
We can propose a simple model for the flux function q(x, t) – suppose we assume the cars are
all moving at a constant speed C. Then we can argue that the flux is just equal to the product

S
of the number of cars time the speed they are moving at,

N
A
q(x, t) = Cρ(x, t). (4.8)

H
IT
Substituting into the transport equation yields

H
ρt + Cρx = 0, AT (4.9)

which is just the convection equation. If we specify the initial distribution of cars,
F

ρ(x, 0) = F (x), (4.10)

O

we can show fairly easily that the solution to the convection equation with this initial condition
S

is just
TE

ρ(x, t) = F (x − Ct), (4.11)

O

corresponding to cars moving uniformly to the right.

N
E
R
TU
C
LE

Page 6 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Physically, we just see the distribution of cars translating to the right with a speed of C.

S
N
A
H
IT
Figure 1.2: Solution to the convection equation.
(draw your own figure).

H
AT
To verify this solution let ξ = x − Ct, and look for a solution F (ξ). Then, by the chain rule

Ft = Fξ ξt = −CFξ Fx = Fξ ξx = Fξ (4.12)
F
O

Substituting ρ(x, t) = F (ξ) into the convection equation (4.9), we find

S

ρt + Cρx = Ft + CFx = −CFξ + CFξ = 0. (4.13)

TE

Moreover, when t = 0, we find ξ = x so that the initial condition ρ(x, 0) = F (x) is satisfied
O

also.
N
E
R
TU
C
LE

Page 7 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 1
Classify the follow differential equations as ODE’s or PDE’s, linear or nonlinear, and
determine their order. For the linear equations, determine whether or not they are homo-
geneous.
1. The diffusion equation for u(x, t):

ut = c2 uxx

2. The wave equation for w(x, t):

wtt = c2 wxx

S
3. The thin film equation for u(x, t):

N
A
ut = −(uuxxx )x

H
4. The forced harmonic oscillator for y(t):

IT
ytt + ω 2 y = F cos(Ωt)

H
AT
5. The Poisson Equation for the electric potential Φ(x, y, z):

Φxx + Φyy + Φzz = 4πρ(x, y, z)

F
O

where ρ(x, y, z) is a known charge density.

S

6. Burger’s equation for u(x, t):

TE

ut + uux = νuxx
O
N

Example : 2
E

Suppose when deriving the convection equation, we assumed the speed of the cars was
R

given by βx for x > 0.

TU

1. Explain why the flux function now is given by q(x, t) = βxρ and the associated
transport equation is given by
C
LE

ρt + (βxρ)x = 0.

2. Explain why
ρ(0, t) = 0, ρ(x, 0) = xe−x
correspond to a boundary condition of no flux of cars in from the origin and an
initial condition specifying the distribution of cars at t = 0.
3. Verify that
−βt
ρ(x, t) = xe−(2βt+xe )
is a solution to both the transport equation given in (a) and the initial and boundary
conditions given in (b).

Page 8 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 3
Show that the helicoid
u(x, y) = tan−1 (y/x)
satisfies the minimal surface equation,

(1 + u2y )uxx − 2ux uy uxy + (1 + u2x )uyy

MAPLE may be helpful with the algebra.

S
N
A
Example : 4

H
IT
Show that the soliton

H
u(x, t) = 2α2 sech α(x − 4α2 t)


AT
satisfies the the Korteweg-deVries equation,
F

ut + 6uux = uxxx
O

MAPLE may be helpful with the algebra, in particular if you don’t remember your hy-
S

perbolic trigonometric identities.

TE
O
N
E
R
TU
C
LE

Figure 7.1: The Vibrating Membrane

(draw your own figure).

Page 9 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

5 Examples of Wave Equations in Various Settings

As we have seen before the ”classical” one-dimensional wave equation has the form:

utt = c2 uxx , (5.1)

where u = u(x, t) can be thought of as the vertical displacement of the vibration of a string.
The string can be fixed at both ends, or just at one end, or we can think of an ”infinite” string,
that is not bound at any end. Each will yield different boundary conditions for the well-posed
wave equation. We can also consider the case where the string is ”pushed” with an external

S
force h(x, t), or where we take under consideration the friction coefficient from the air that the

N
string displaces. These two equations will be called ”forced” and ”damped” respectively. In the

A
”forced” case, the wave equation is:

H
utt = c2 uxx + h(x, t),

IT
where an example of the acting force is the gravitational force. In the ”damped” case the

H
equation will look like: AT
utt + kut = c2 uxx ,
where k can be the friction coefficient.
F
O

If we have more than one spatial dimension (a membrane for example), the wave equation will
look a bit different. In the case of the vibrating membrane we have two spatial variables and the
S

wave equation will look like:

TE

utt = c2 (uxx + uyy ).

O

For n−dimensions (whatever THAT means...) the n-wave equation will be:
N

utt = c2 (ux1 x1 + ux2 x2 + · · · + uxn xn ).

E
R

The initial conditions for the one-dimensional wave equation will be:
TU

u(x, 0) = f (x), ut (x, 0) = g(x).

C

For the finite string the boundary conditions will be:

LE

u(0, t) = A(t), u(L, t) = B(t).

Page 10 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

6 Dirichlet Problem and Separation of Variables

If we tie the string at both ends we can have the following boundary conditions:

u(0, t) = A(t), u(L, t) = B(t),

where A, B are C 1 piecewise functions. For example, we can have a sinusoidal function at one
end and a Heaviside function at the other.
When the boundary values A and B are 0 we obtain the Dirichlet Problem for the wave

S
equation:

N
utt = c2 uxx , 0 < x < L, t > 0 DE

A
u(0, t) = 0, u(L, t) = 0, t>0 BC

H
u(x, 0) = f (x), ut (x, 0) = g(x) 0<x<L IC.

IT
H
As you have seen in Lecture 5 for the diffusion equation, the method of separating the variables
is a very convenient way to obtain solutions for PDEs. In the case of the Dirichlet Problem we
AT
will quickly review the method.
Theorem 6.1. A solution of the problem:
F
O

utt = c2 uxx , 0 < x < L, t > 0 DE

u(0, t) = 0, u(L, t) = 0, t>0 BC
S
TE

u(x, 0) = f (x), ut (x, 0) = g(x) 0<x<L IC.

O

is given by:
N

N  
X L nπct nπx nπct
u(x, t) = Ān sin( ) + Bn cos( ) sin( ),
E

n=1
nπc L L L
R
TU

N N
X nπx X nπx
where: f (x) = Bn sin( ), and g(x) = Ān sin( ). The coefficients Ān
L L
C

n=1 n=1
Z L L
2 nπx 2 nπx
Z
LE

and Bn are given by: Bn = f (x) sin( ), Ān = g(x) sin( ).

L 0 L L 0 L

Proof. We will use the method of separation of variables, namely think of the solution u(x, t)
as a product of a function that depends only on the variable t and of a function that depends
only on the variable x.
Let u(x, t) = X(x)T (t) and substitute in the equation utt = c2 uxx , to obtain:

X(x)T̈ (t) = c2 Ẍ(x)T (t),

T̈ (t) Ẍ(x)
or = , thus the equality is one of functions of different variables, so both
c2 T (t) X(x)
quotients have to be constant.

Page 11 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

T̈ (t) Ẍ(x)
Say = = ±λ2 , then we can solve each ordinary differential equation sepa-
c2 T (t) X(x)
rately. We have the following three cases: −λ2 , λ2 , and λ = 0.
Ẍ(x)
Case 1 When the constant is −λ2 , then the solutions for = −λ2 , are: X(x) =
X(x)
T̈ (t)
c1 sin(λx)+c2 cos(λx), and the solutions for = −λ2 , are: T (t) = d1 sin(λct)+
c2 T (t)
d2 cos(λct). Then u(x, t) = (d1 sin(λct) + d2 cos(λct))(c1 sin(λx) + c2 cos(λx)).
Ẍ(x)
Case 2 When the constant is λ2 , then the solutions for = λ2 , are: X(x) = c1 eλx +

S
X(x)
T̈ (t)

N
c2 e−λx , and the solutions for = λ2 , are ±cλ, are T (t) = d1 eλct + d2 e−λct . Then
c2 T (t)

A
u(x, t) = (d1 eλct + d2 e−λct )(c1 eλx + c2 e−λx ).

H
IT
Case 3 When the constant is 0, then the equations become Ẍ(x) = T̈ (t) = 0, and X(x) =
c1 x + c2 , and T (t) = d1 t + d2 . Then u(x, t) = (d1 t + d2 )(c1 x + c2 ).

H
AT
Let’s take a look at the boundary conditions: u(0, t) = 0, u(L, t) = 0. The only solution
for u(x, t) that can satisfy them is u(x, t) = (d1 sin(λct) + d2 cos(λct))(c1 sin(λx) +
c2 cos(λx)), and the boundary conditions translate into:
F
O

(d1 sin(λct) + d2 cos(λct))(c1 sin(0) + c2 cos(0)) = 0

(d1 sin(λct) + d2 cos(λct))(c1 sin(λL) + c2 cos(λL)) = 0, ∀t > 0,
S
TE

namely:
O

c2 = 0
N

c1 sin(λL) = 0.
E

πn
R

From the last condition we obtain λ = , and

L
TU

∞  
X πn πn πn
u(x, t) = d1n sin( ct) + d2n cos( ct) cn sin( x).
C

n=1
L L L
LE

The only conditions left to check are the initial conditions:

N
X nπx
u(x, 0) = f (x) = Bn sin( ),
n=1
L
N
X nπx
ut (x, 0) = g(x) = Ān sin( ).
n=1
L

N  
X L nπct nπx nπct
Then u(x, t) = Ān sin( ) + Bn cos( ) sin( ),
n=1
nπc L L L

Page 12 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Remark 6.0.1. In the more general case for the Dirichlet Problem, when initial conditions (IC)
change to more general homogeneous conditions: k1 u(x, 0)+k2 ux (x, 0) = 0, we can solve
the problem in the same manner, using separation of variables.
Exercise: 1. Check that Case 1 is the only one that verifies the boundary conditions in the proof
above.

Exercise: 2. Check that the solution found above verifies the initial conditions.

Exercise: 3. A string of length π is held fixed at both endpoints. Its initial position is f (x) =
sin(x) and its initial velocity is g(x) = cos x. Assuming that c = 1, find the position of
the string u(x, t) for every x ∈ [0, π] and for every t > 0. Find an approximate value for

S
u(x, t) by adding several terms of the series. Animate the approximation and draw a 3D plot.
Note. Use Maple carefully, there might be some tricky answers. Finally, use Maple to check

N
that your (approximate) solution satisfies the PDE, the boundary conditions, and the initial

A
conditions (at least approximately).

H
IT
Exercise: 4. Solve the following problem for the string equation:

H
PDE utt = uxx ,
AT
BC u(0, t) = 0, ux (π, t) = 0 for every t > 0;
F

IC u(x, 0) = sin(x), ut (x, 0) = 0, for every x ∈ [0, π] .

O

Notice the change in the boundary conditions. This will lead to different eigenvalues and eigen-
S

functions.
TE

Use Maple to animate the solution you found, to draw a 3D plot, and to check that the solution
satisfies the conditions of the problem.
O
N
E
R
TU
C
LE

Page 13 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

7 Galerkin Method for the ”Damped” Wave Equation

The ”damped” wave equation looks like:
utt + νut = c2 uxx , 0 < x < L, t > 0 DE
u(0, t) = 0, u(L, t) = 0, t>0 BC
u(x, 0) = f (x), ut (x, 0) = g(x) 0 < x < L. IC.

From our previous discussion, we know that the solution will have the form:
N
nπx

S
X
u(x, t) = fn (t) sin( ).
n=1
L

N
Plugging into the equation, we obtain the following equations for fn :

A
H
f¨n + ν f˙n + (αn )2 fn = 0,

IT
cnπ
where αn = with eigenvalues:

H
L AT
r
ν ν
λn12 = ± ( )2 − (αn )2 ,
2 2
F

r
ν ν
O

i.e.λn12 = ± iωn , where: ωn = ((αn )2 − )2 .

2 2
S

So, if the friction coefficient ν is small enough, then we will have:

TE

ν
fn (t) = e− 2 t (cn cos(ωn t) + dn sin(ωn t)),
O

and
N

∞
X ν nπx
u(x, t) = e− 2 t (cn cos(ωn t) + dn sin(ωn t)) sin( ),
E

n=1
L
R

We input the initial conditions:

TU

∞ ∞
X nπx X ν nπx
u(x, 0) = f (x) = cn sin( ), ut (x, 0) = g(x) = (ωn dn − cn ) sin( )
C

n=1
L n=1
2 L
LE

ν
Then cn is the n−th Fourier coefficient for f, and ωn dn − cn is the n−th Fourier coefficient
2
for g, so EUREKA! The problem has been SOLVED:

L
2 nπx
Z
cn = f (x) sin( )dx
l 0 L
ν 2 L nπx
Z
ωn dn − cn = g(x) sin( )dx
2 l 0 L
L 
2 ν nπx
Z
dn = f (x) + g(x) sin( )dx.
ωl 0 2 L

Page 14 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

8 The Example of the Plucked String

The plucked string refers to the initial condition for the Dirichlet problem , where f (x) looks
like a ”plucked string”, namely:

utt = c2 uxx , 0 < x < L, t > 0 DE

u(0, t) = 0, u(L, t) = 0, t>0 BC
u(x, 0) = f (x), ut (x, 0) = 0 0<x<L IC,

S
where

N
A
 u0

H
 x,
 0 ≤ x ≤ x0
x0

IT
f (x) = x−L
u0
 , x0 ≤ x ≤ L

H
x0 − L AT
F
O
S
TE
O
N
E
R
TU
C

Figure 7.: The initial condition for the ”Plucked string” wave.
LE

(draw your own figure).

The behaviour of the solutions is exemplified in the Maple animations.
Let’s find a formal solution to the ”plucked string” equation. In order to do this we need to find
Bn , the Fourier sine coefficients of f (x).
L
nπx
Z
Bn = f (x)sin( )dx
0 L
L
2 −L nπx 2 L nπx
Z
= f (x)( ) cos( )|L
0 + f 0 (x) cos( )dx.
L nπ L L nπ 0 L

Page 15 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

We have that:
u
0
 ,
 0 ≤ x ≤ x0
f˙(x) = x0u0

 , x0 ≤ x ≤ L,
x0 − L
and f (0) = f (L) = 0, thus:
Z x0 Z L
2 L u0 nπx 2 L u0 nπx
Bn = ( ) cos( )dx + ( ) cos( )dx
L nπ 0 x0 L L nπ x0 x0 − L L
2 L 2 u0 nπx0 u0 nπx0
= ( ) ( sin( )− sin( ))

S
L nπ x0 L x0 − L L
2L2 u0

N
1 nπx0
= sin( ).

A
π 2 x0 (L − x0 ) n2 L

H
Now we can write the formal solution to the plucked string equation:

IT
∞
2L2 u0 1 nπx0 nπct nπx

H
X
u(x, t) = sin( ) cos( ) sin( ).
π 2 x0 (L − x0 ) n=1 n2 L L L
AT
F

8.1 Musical instruments

O

Many instruments produce sound by making strings vibrate; such are the harp, the piano, the
S

harpsichord, the guitar, the violin, and others. Strings are kept fixed at the endpoints, but they
TE

way the instruments are played create different initial conditions. In instruments like the guitar,
the string is plucked; this produces an initial perturbation with no initial velocity. In the piano,
O

on the other hand, the string is hit, which creates an initial velocity but no initial perturbation
N

from the initial position.

E

The oscillations of the string are described by

R

∞
TU

X
u(x, t) = (An cos ωn t + Bn sin ωn t) sin λn t,
n=1
C
LE

where π π
λn = n and ωn = cλn = cn .
L L
The sound we hear is thus a combination of the main harmonic sounds (eigenfunctions)

un (x, t) = (An cos ωn t + Bn sin ωn t) sin λn t.

The contribution of each particular harmonic is measured by its energy, which turns out to be
equal to:
ω2 M
En = n (A2n + Bn2 ),
4
where M = DL is the total mass of the string (recall that D was the density).

Page 16 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

For the plucked string (Section 7.5), the energy is given by:

M u20 L2 c2 πnx0
En = sin2 .
n2 π 2 x20 (L − x0 )2 L

The energy decreases as n−2 , so only the main tone u1 and a few other harmonics are audible.
On the other hand, if we hit the string with a flat hammer of length 2δ with center at x0 and
producing an initial velocity v0 , the energy of the nth harmonic is

4M V02 πnx0 πnδ

π 2 sin2 sin2

S
En = ,
n2 L L

N
and the energy again decreases as n−2 . However, if the hammer is sufficiently narrow, letting δ

A
tend to zero (the blade of a knife), we get the model of a string getting an impulse concentrated

H
at a point x0 . The corresponding energy is:

IT
v02 πnx0

H
En = sin2 .
L L
AT
Thus, for a very narrow hammer the energies of all harmonics are of the same order and the
F

generated sound will be saturated with harmonics. This can be checked experimentally, by
O

hitting a string with the blade of a knife. The sound will have a metallic quality.
S

Not all harmonics are desirable. The first ones, u2 up to u6 , sound well together with the main
TE

harmonic u1 . However, the 7th and the first harmonics sounding together produce a sense of
dissonance.
O

There are several ways to try to “kill” those harmonics by percussion (as in the piano).
N

πnx0
a). The position of the hammer. The presence of the factor sin shows that by choosing
E

L
the center x0 of the hammer at the node of the undesired harmonic we may make it disappear
R

(make the corresponding An and Bn be equal to zero). In modern pianos the position of the
TU

hammer is chosen near the nodes of the 7th and the 8th harmonics, to “kill” them.
C

b). The shape of the hammer. In modern pianos the hammers are not flat, but rather round. One
LE

can model this situation by choosing the initial velocity to be, say, a parabola on the interval
[x0 − δ, x0 + δ], instead of a horizontal line. Older pianos, which had flatter and narrower
hammers, produced a more piercing, shrilled sound.
c). The rigidity of the hammer. If instead of being rigid the hammer is softer. In this case
the motion is not described by its initial position and velocity, but rather by a short-time acting
force, which varies in time.
Exercise: 1. Find the solution u(x, t) of the string equation if l = π, c = 1, when both
endpoints are fixed, the initial velocity is zero, and u(x, t) is the following piecewise linear
function:

Page 17 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

9 Uniqueness of the solution to the wave equation

Theorem 9.1. Let u1 (x, t) and u2 (x, t) be two solutions of the problem:

utt = c2 uxx , 0 < x < L, t > 0 DE

u(0, t) = A(t), u(L, t) = B(t), t>0 BC
u(x, 0) = f (x), ut (x, 0) = g(x) 0 < x < L. IC,

where A, B, f, g are C 1 piecewise continuous. Then u1 (x, t) = u2 (x, t) for all points in
the domain.

S
Proof. Let v(x, t) = u1 (x, t) − u2 (x, t), then v satisfies the wave equation with initial

N
conditions: u(x, 0) = ut (x, 0) = 0, and boundary conditions u(0, t) = u(L, t) = 0. Our

A
goal is to prove that v(x, t) = 0 ∀x, t.

H
In order to accomplish this, define:

IT
Z L  2
c vx (x, t)2 + vt (x, t)2 dx.

H

H(t) =
0
AT
We will prove that H(t) = 0 first. Differentiating with respect to t we obtain:
F

Z L
 2
O


Ḣ(t) = c 2vx vxt + 2vt vtt dx
0
S

Z L
2
= 2c [vx vxt + vt vxx ] dx
TE

0
Z L
δ
O

= 2c 2
(vx vt )dx = 2c2 [vx (x, t)vt (x, t)]L
0
0 δx
N

= 2c2 (vx (L, t)vt (L, t) − vx (0, t)vt (0, t)) = 0.

E
R

Since Ḣ(t) = 0, H(t) is constant, and as H(0) = 0, we conclude that H(t) = 0.

TU

Z L
Then vt (x, t) = 0, and v(x, t) = v(x, t) − v(x, 0) = vt (x, t)dt = 0.
0
C

Remark 9.0.1. The energy integral of the string at time t is:

LE

Z L
T0 ux (x, t)2 + Dut (x, t)2 dx,
 
E(t) =
0

where D is the mass per unit length and T0 is the constant tension when the string is straight.
We can see that the energy is proportional to H if we construct H using u instead of v. So the
uniqueness proof comes from the conservation of energy for the unforced string.

Page 18 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

10 Cauchy Problem for the infinite string, D’Alembert’s So-

lution
When we have an infinite string, with no boundary, then we have the following Cauchy Prob-
lem:
utt = c2 uxx , 0 < x < L, t > 0 DE
u(x, 0) = f (x), ut (x, 0) = g(x) 0 < x < L. IC
Theorem 10.1. The solution of the wave equation for the infinite string is:
Z x+ct

S
1 1
u(x, t) = [f (x − ct) + f (x + ct)] + g(s)ds,
2 2c x−ct

N
A
when f is C 2 and g is C 1 .

H
Proof. The wave equation can be written in the alternate form:

IT
δ2 δ2

H
( − c2 )u(x, t) = 0,
δt2 δx2
AT
and we can factor the ”differential operator”
δ2 δ2 δ δ δ δ
F

( − c2 )=( −c )( ) +c
O

δt2 δx 2
δt δx δt δx
Now we will use this factorization to find the general solution to the wave equation. Denote:
S

δ δ δ δ
TE

v(x, t) = ( + c )u(x, t), then the wave equation becomes: ( − c )v(x, t) = 0.

δt δx δt δx
We know that the last equation has solutions of the form: v(x, t) = F (x − ct), so the wave
O

equation becomes:
N

δ δ
( + c )u(x, t) = F (x − ct).
δt δx
E
R

To solve it, we make the change of variable: w = x − ct, z = x + ct, and let U (w, z) =
TU

u(x, t), and we replace in the original equation to get:

C
LE

ut + cux = Uw wt + Uz zt + c(Uw wx + Uz zx )
= −cUw + cUz + cUw + cUz
= 2cUz = F (z).
1
Z
Thus, U (w, z) = F (z)dz + G(w) = H(z) + G(w), or:
2c
u(x, t) = H(x + ct) + G(x − ct).
We have obtain so far that the general solution is a superposition of waves traveling in opposite
direction with speed c. We will use now the initial conditions:
f (x) = u(x, 0) = H(x) + G(x)
g(x) = ut (x, 0) = Ḣ(x)c − Ġ(x)c.

Page 19 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

We obtain:

H(x) + G(x) = f (x)

1 x
Z
H(x) − G(x) = g(s)ds + C,
c 0
and
 x 
1 1
Z
H(x) = f (x) + g(s)ds + C
2 c 0
1 x
 
1
Z
G(x) = f (x) − g(s)ds − C

S
2 c 0
1 0
 
1
Z

N
= f (x) + g(s)ds − C .
2 c

A
x

H
Then:

IT
u(x, t) = H(x + ct) + G(x − ct)

H
1 x+ct 1 0
   
1 1
Z Z
f (x − ct) + g(s)ds − C ,
=
2
f (x + ct) +
c 0
g(s)ds + C +
2
AT
c x−ct
F

which completes the proof.

O

Exercise: 1. Show that odd/even initial conditions yield odd/even solutions.

S
TE
O
N
E
R
TU
C
LE

Exercise: 2. Show that periodic initial conditions yield periodic solutions.

Exercise: 3. Show that we can deduct the solution to the finite string problem from the solution
of the infinite string. You may need help on this one!

Page 20 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 5
A damped string of length 1 has equation

utt = c2 uxx − γut ,

where γ is a small damping coefficient. Find the solution u(x, t) assuming that both
endpoints are fixed, the initial condition is x(1 − x) and the initial velocity is zero.
1 1
Plot and animate the solution for the case when c = and γ = .
4 5

Example : 6

S
Find the solution u(x, y, t) of a square membrane with side 1 fixed on the boundary, if

N
the initial position is u(x, y, 0) = (x − x2 )(y − y 2 ) and the initial velocity is zero.

A
Animate several eigenfunctions un,m (x, y, t), say, u1,1 , u1,2 , u3,5 , assuming that c =
1.

H
IT
Example : 7

H
Solve the string equation utt = c2 uxx for L = 1, with the boundary conditions
AT
u(0, t) = 0 and u(1, t) = 1, with zero initial velocity, assuming that the initial
position is
(a) u(x, 0) = sin x (easier),
F

(b) u(x, 0) = x2 (harder).

O

Hint: You cannot use the superposition principle, since the boundary condition at x = 1
S

is not homogeneous. Try a change of coordinates first, v(x, t) = u(x, t) + h(x),

TE

where h(x) is a suitable (easy) function that would guarantee that v also satisfies the
string equation, now with homogeneous boundary conditions.
O
N

Example : 8
E

Solve the equation utt = c2 uxx + sin x For 0 ≤ x ≤ π and t > 0, with the boundary
R

conditions u(0, t) = 1, ut (π, t) = 2, and the initial conditions u(x, 0) = f (x),

TU

ut (x, 0) = g(x).
Hint: Make the change of coordinates u(x, t) = y(x) + v(x, t), where y(x) satisfies
C

c2 y 00 + sin x = 0,
LE

y(0) = 1, and y(π) = 2. Find y.

What will be the PDE for v?
What are the (boundary and initial) conditions for v?

Page 21 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 9
Solve the string equation utt = c2 uxx for L = 1, with the boundary conditions
u(0, t) = 0 and ux (1, t) + u(1, t) = 0. This corresponds to the case when the
left end is fixed and the right end is attached to an elastic hinge. The initial conditions
2
are u(x, 0) = x − x2 and ut (x, 0) = x.
3
Note. This exercise is hard! The eigenvalues λn will be solutions of a transcendental
equation.

Example : 10
In the ”damped” case, what happens if the friction coefficient is large, say that you im-

S
merse your string in a high viscosity liquid?

N
A
H
IT
H
AT
F
O
S
TE
O
N
E
R
TU
C
LE

Page 22 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

11 An Introduction to Heat Flow

A classical example of the application of ordinary differential equations is Newton’s Law of
Cooling which, basically, answers the question “How does a cup of coffee cool?” Newton
hypothesized that the rate at which the temperature, U (t), changes is proportional to the differ-
ence with the ambient temperature, which we call Ū ,
dU
= −κ(U − Ū ). (11.1)
dt
Here κ is a positive rate constant (with units of inverse time) that measures how fast heat is lost
from the coffee cup to the ambient environment. If we specify the initial temperature,

S
U (0) = U0 , (11.2)

N
we can solve for the evolution of the temperature,

A
U (t) = Ū + (U0 − Ū )e−κt .

H
(11.3)

IT
H
AT
F
O
S
TE
O
N
E
R

Figure 2.1: (a) A coffee cup (b) Its temperature as a function of time.
TU

(draw your own figure).

C

If we graph the temperature as a function of time, we see that it decays exponentially to the
LE

ambient temperature, Ū , at a rate governed by κ.

When we derived Newton’s Law of cooling we made several assumptions – most importantly
that the temperature in the coffee cup did not vary with location. If we account for the variation
of temperature with location, we can derive a PDE called the heat equation or, more generally,
the diffusion equation. If the temperature, U (x, t) is a function of a single spatial variable, x,
we will show that it satisfies the diffusion equation,
Ut = c2 Uxx , (11.4)
where D is a constant known as the thermal diffusivity. In higher dimensions, the equation can
be written
Ut = c2 ∇2 U, (11.5)
where ∇2 is the Laplacian.

Page 23 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

12 Derivation of the Diffusion Equation

The diffusion equation will be our second example of a conservation law; we can derive the
equation by accounting for the flow of thermal energy. Suppose we consider a metal bar, with a
uniform cross-sectional area, A, whose temperature, U (x, t), is a function of time, t, and the
position, x, along the bar (that is we assume the temperature is uniform in every cross-section).

S
N
A
H
IT
H
AT
F

Figure 2.2: Conservation of heat in a metal bar

O

of cross-sectional area A.
S

(draw your own figure).

TE

Let the thermal energy in the region a < x < b is given by

O

Z b
E = ρ0 cv A U (x, t) dx (12.1)
N

a
E

The important term in the integral is the temperature, U (x, t), measured in degrees. The
R

remaining constants, A, the cross-sectional area (with units of [(length)2 ]), ρ0 , the density
TU

[mass/(length)3 ] and cv , the heat capacity [energy/(degree · mass)] are physical properties of
the material – think of them as being obligatory for making the units work out.
C

We wish to equate the change in thermal energy to the heat flux out of the bar through the planes
LE

at x = a and x = b. To do this we use Fourier’s heat law which states that the flux density
of thermal energy, q(x, t) is proportional to the temperature gradient,

q(x, t) = −kUx , (12.2)

where the negative sign reflects the fact that heat flows from hot to cold, just as in New-
ton’s law of cooling, with a constant of proportionality, k, called the thermal conductivity
[(energy·length)/(degrees·time)].

Page 24 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

S
N
A
H
IT
Figure 2.3: Heat flux is from hot to cold!!
(draw your own figure).

H
AT
Now, the total flux of thermal energy into the into the region a < x < b is given by
Q = A[q(a, t) − q(b, t)], (12.3)
F
O

where we multiply by the area A to get the total flux through the cross-section.
By conservation of energy, the rate of change of the energy between a and b is given by the
S
TE

flux into the region,

dE
= Q. (12.4)
O

dt
N

Once again we can rewrite the flux by a clever application of the fundamental theorem of cal-
culus,
E

Q = A[q(a, t) − q(b, t)] = − Aq(x, t)|x=b (12.5)

R

x=a
Z b
TU

= −A qx dx. (12.6)
a
C

We now rewrite the conservation of energy equation as

LE

 Z b  Z b Z b
dE d
= ρcv A U dx = ρcv AUt dx = Q = −A qx dx, (12.7)
dt dt a a a

or, rearranging Z b
ρcv A Ut + Aqx dx = 0. (12.8)
a
Since this is true for every interval a < x < b, the integrand must vanish identically. So
ρcv AUt + Aqx = 0. (12.9)
Substituting for the flux function q(x, t) = −kUx yields
ρcv AUt − kA(Ux )x = 0. (12.10)

Page 25 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Rearranging the equation yields the diffusion equation,

Ut = c2 Uxx , (12.11)

where the diffusivity, D = k/(ρcv ), is a constant which is determined by the geometry and
physical properties of the metal bar.
To complete the description of the problem, we need to supplement the diffusion equation
with boundary conditions and initial conditions. Suppose we consider a bar of finite length
L, occupying the region 0 < x < L. At the boundaries of the metal bar we can specify a fixed
temperature,
U (0, t) = U0 U (L, t) = U1 , (12.12)

S
which are usually referred to as Dirichlet boundary conditions. Alternatively, we could specify

N
a heat flux,

A
q0 = q(0, t) = −kUx (0, t) q1 = q(L, t) = −kUx (L, t).

H
(12.13)

IT
Specifying the gradient across the boundary is referred to as Neumann boundary conditions.

H
Finally, we also need to specify the initial temperature distribution,
AT
U (x, 0) = f (x) 0 < x < L. (12.14)
F

We will demonstrate below that the solution to this problem (if it exists) is unique; later in this
O

course we will solve this problem using the method of separation variables.
S

For completeness, we also comment here that the problem can be posed on the infinite line,
TE

−∞ < x < ∞ sometime called the Cauchy problem – in this case one usually replaces the
boundary condition with the specification that the temperature remains bounded as we approach
O

infinity,
N

lim |U (x, t)| < C, (12.15)

x→±∞
E

for some constant C. This condition may seem superfluous at first glance, but actually is neces-
R

sary to stop heat from leaking in from infinity (speaking very, very informally an infinite source
TU

of heat infinitely far away can have a finite effect in a short amount of time). If you are interested
in details, look for the examples of Tychonov in a PDE’s text1 .
C
LE

13 Examples of Solution to the Diffusion Equation

We can summarize the last section by restating a well-posed problem for the diffusion equation
on the interval 0 < x < L with Dirichlet boundary conditions,
T HE D IRICHLET P ROBLEM FOR THE D IFFUSION E QUATION
(N ON - HOMOGENEOUS B OUNDARY C ONDITIONS )
1
See, for example, T. W. Körner, “Fourier Analysis,” Cambridge University Press, p. 338.

Page 26 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Ut = c2 Uxx 0 < x < L, t > 0 DE

U (0, t) = U0 U (L, t) = U1 t>0 BC
U (x, 0) = f (x) 0 < x < L. IC

Solving the general problem will have to wait, but we can find some specific solutions to the
problem using the ideas of Separation of Variables. For the moment, we will restrict ourselves
to homogeneous boundary conditions,
T HE D IRICHLET P ROBLEM FOR THE D IFFUSION E QUATION
(H OMOGENEOUS B OUNDARY C ONDITIONS )

S
Ut = c2 Uxx

N
0 < x < L, t > 0 DE

A
U (0, t) = 0 U (L, t) = 0 t>0 BC

H
U (x, 0) = f (x) 0 < x < L, IC

IT
H
If you want, you can skip the derivation for the moment and jump ahead to Exercise 1, if you
AT
don’t mind the solution appearing deus ex machina ( a fancy term for “out of thin air”).
F

13.1 A Solution to the Homogeneous Dirichlet Problem

O
S

Let us look for solutions to the homogeneous Dirichlet problem of the form
TE

U (x, t) = X(x)T (t) (13.1)

O
N

we find from the differential equation (DE) that

XTt = c2 Xxx T
E

(13.2)
R

and dividing by XT we find

TU

Tt Xxx
= = −λ. (13.3)
c2 T X
C

where λ is to be determined. Now because Tt /DT is only a function of t and Xxx /X is only
LE

a function of x we know that λ must be independent of x and t respectively, and therefore must
be a constant – consequently it is known as the separation constant. We can now solve the
resulting ODE for T (t)

Tt = −λDT ⇒ T (t) = e−λt , (13.4)

or some constant multiple of it.

We now look for a solution for the X(x) equation that also satisfies the homogeneous boundary
conditions. From the boundary conditions (BC), we know that

U (0, t) = X(0)T (t) = 0 ⇒ X(0) = 0 (13.5)

U (0, t) = X(L)T (t) = 0 ⇒ X(L) = 0 (13.6)

Page 27 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

So finally we conclude that we are looking for solutions to the Boundary Value Problem for
X(x),
Xxx + λX = 0, X(0) = 0 X(L) = 0. (13.7)
Solving the DE, we find that
√ √
X(x) = B cos( λx) + C sin( λx) (13.8)

and applying the boundary conditions we see that X(0) = 0 implies that B = 0, and that
√
C sin( λL) = 0. (13.9)

S
Consequently, a non-trivial solution (that is a solution for which X(x) 6= 0) for X(x) can be
found if and only if

N
A
 2
nπ
λ = λn ≡ for n = 1, 2, 3 . . . (13.10)

H
L

IT
for which we find

H
 
nπx
X(x) = Xn (x) ≡ sin
AT
for n = 1, 2, 3 . . . , (13.11)
L
F

or some constant multiple of it. These special values of λ are called eigenvalues and the asso-
O

ciated functions, Xn (x), are known as eigenfunctions.

S

Multiplying the solution for Xn (x) and T (t) together finally yields a solution for Un (x, t),
TE

 
nπx nπ 2 2
U (x, t) = Un (x, t) ≡ sin e−( L ) c t for n = 1, 2, 3 . . . . (13.12)
O

L
N

The method of separation of variables is very powerful – it will be one of our primary tools for
finding solutions to PDE’s in the coming lectures.
E
R
TU

Exercise: 1. Verify that

C

 
nπx 2
LE

nπ
Un (x, t) ≡ sin e−( L ) t
for n = 1, 2, 3 . . . , (13.13)
L

satisfies the diffusion equation Ut = c2 Uxx and the homogeneous boundary conditions
U (0, t) = U (L, t) = 0. Explain why any linear combination of Un ,
∞ ∞  
nπx nπ 2 2
e−( )
X X
c t
U (x, t) = an Un (x, t) = an sin L (13.14)
n=1 n=1
L

also satisfies the diffusion equation and the homogeneous boundary condition. Does it worry
you that this is an infinite sum? What initial condition, U (x, 0), does this correspond to?

Page 28 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

13.2 A Solution to the Cauchy Problem

We can also consider a solution to the Cauchy problem for the diffusion equation, which you
hopefully remember is the problem posed on the entire real line,
T HE C AUCHY P ROBLEM FOR THE D IFFUSION E QUATION

Ut = c2 Uxx −∞ < x < ∞, t > 0 DE

lim |U (x, t)| < C t > 0, BC
x→±∞
U (x, 0) = f (x) −∞ < x < ∞. IC

S
N
While there are many clever derivation for the solution to this problem, for the moment I will

A
simply give you the most important solution, usually called the fundamental solution or the

H
diffusion kernel,

IT
1 x2
U (x, t) = G(x, t + τ ) ≡ p e− 4D(t+τ ) . (13.15)

H
4πD(t + τ )
AT
where τ is a constant (which we will assume is positive). This solution can be used to construct
a general solution of the diffusion equation for an arbitrary initial condition, f (x).
F
O
S
TE
O
N
E
R

Exercise: 2. Verify that

TU

1 − x2
G(x, t + τ ) ≡ p e 4c2 (t+τ ) . (13.16)
C

4πc2 (t + τ )
LE

satisfies the diffusion equation and the boundary conditions for the Cauchy problem when τ >
0 . Show that this solution corresponds to a Gaussian with time varying width and height. How
does the Gaussian’s width, height and area vary in time?

Page 29 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

14 The Maximum Principle

Looking at solutions to the heat equation, we note that they tend to average out maximums
and minimums. We can develop some intuition for this by considering what the equation says.
Basically, Ut = c2 Uxx means: The temperature is decreasing when the profile is convex down
and the temperature is increasing when the profile is convex up.

S
N
A
H
IT
H
AT
F

Figure 2.4: The heat equation interperted graphically

O

(draw your own figure).

S

From which we conclude that interior maximums in temperature are decreasing and interior
TE

minimums of temperature are increasing. This reasoning is not quite airtight (how to make it
tighter is a good question to ponder). We can give a rigorous statement (without proof) of the
O

maximum principle:
N

Theorem 14.1 (Maximum Principle for the Diffusion Equation). If

u(x, t) satisfies the Dirichlet problem for the diffusion equation in the semi-infinite strip 0 <
E

x < L, 0 < t, then it assumes its maximum value (as a function of x and t) either initially
R

(when t = 0) or on the lateral boundaries (where x = 0 or x = l).

TU

The same is also true of the minimum of u(x, t). A proof can be found in most advanced PDE
C

texts.
LE

Exercise: 1. Interpret the solutions we have found for the diffusion equation in terms of the
maximum principle. Show examples where the maximum value of u(x, t) occur in the initial
condition and on the lateral boundaries.

15 Energy Dissipation and Uniqueness

By looking at what is normally known as energy for the diffusion equation, we can show that
the solution for the Dirichlet problem is unique. Note this energy is a mathematical construct,
not to be confused with the thermal energy discussed in the derivation of the diffusion equation.

Page 30 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

First, suppose that U (x, t) is a solution to the homogeneous Dirichlet problem,

Ut = c2 Uxx 0 < x < L, t > 0 DE

U (0, t) = 0 U (L, t) = 0 t>0 BC
U (x, 0) = f (x) 0 < x < L, IC

Let’s define, the energy,

L
1
Z
W = U 2 dx, (15.1)
2 0

S
which is a function of t dependent on the particular solution U (x, t) (technically it is a function

N
of t and a functional of U (x, t)). Note that W ≥ 0 with W = 0 only for the trivial solution
U (x, t) = 0.

A
H
If we differentiate the energy with respect to time, we find

IT
Z L
dW
= U Ut dx,

H
dt 0 AT
Z L
= D U Uxx dx,
0
F

Z L
= − (Ux )2 dx + U Uxx |x=L
x=0 ,
O

0
S

where we have substituted the DE and used integration by parts. Now, applying the BC’s, we
TE

fine that the boundary terms from the integration by parts vanish, so,
O

Z L
dW
=− (Ux )2 dx ≤ 0
N

dt 0
E

Now, we can conclude that W is decreasing (that is energy is dissipated!!)unless Ux = 0, that

R

is to say that U is constant. As the only constant solution satisfying the boundary conditions is
TU

U = 0, we might be tempted to conclude that the solution always decays to this trivial state.
This turns out to be true, although one must invest some analysis to show it rigorously.
C
LE

Page 31 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

S
N
A
H
IT
Figure 2.4: (a) A solution to the homogeneous Dirichlet problem for the heat equation. (b) The
corresponding energy, W, which is decreasing to zero.

H
(draw your own figure).
AT
A second conclusion one can reach is that if f (x) = 0, that U (x, t) = 0 for all t > 0. This
follows quickly because W = 0 at t = 0, it is non-increasing and non-negative. While this
F

seems like a trivial result, it has a very powerful consequence. Suppose we had two solutions to
O

the non-homogeneous Dirichlet problem, call them V1 and V2 . You should be able to convince
S

yourself that there difference U = V1 − V2 satisfies the homogeneous Dirichlet problem with
TE

f (x) = 0. Consequently, we know that U (x, t) = 0 for all t > 0, which implies V1 = V2 .
From this we conclude that The solution to the non-homogeneous Dirichlet problem is unique,
O

a powerful result indeed.

N
E
R
TU
C
LE

Exercise: 1. Convince yourself the energy argument for uniqueness of solutions in the previous
paragraph is correct. Show that a similar argument can be made for the Neumann problem.

Page 32 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 11
Consider the diffusion equation with homogeneous Neumann boundary conditions.

Ut = c2 Uxx 0 < x < L, t > 0, DE

Ux (0, t) = 0 Ux (L, t) = 0 t > 0, BC
U (x, 0) = f (x) 0 < x < L. IC

1. Explain physically why this corresponds to the diffusion of heat in a metal bar with insu-
lated ends. Make sure you understand what each of the equations corresponds to.

S
2. Show that

N
(i) U0 (x, t) = 1

A
 
nπx

H
nπ 2 2
(ii) Un (x, t) = cos e−( L ) c t
n = 1, 2, 3, . . .
L

IT
H
satisfy both the diffusion equation (DE) and the homogeneous Neumann boundary con-
ditions (BC).
AT
3. Write down a general solution as a linear combination of the solutions you found in part
(b). What does this say about f (x) if we assume that this solution also satisfies the initial
F

condition (IC)?
O
S
TE

Example : 12
O

In this problem, we will argue that for the homogeneous Neumann problem discussed in
N

Problem 1, that the solution approaches a constant temperature, given by the average of
the initial temperature.
E

1. Suppose we define the total heat energy in the bar as

R

Z L
TU

Q(t) = U (x, t) dx. (15.2)

0
C

Show that Q is conserved, that is that it is independent of time (Hint: compute

LE

dQ
).
dt
2. Use the initial condition to compute Q in terms of f (x).
3. Modify the energy argument in the previous section show that the energy is de-
creasing unless U (x, t) is constant. Use this to argue that U (x, t) approaches a
constant solution as t → ∞.
4. Finally, use parts (a) and (b) of the problem to show that there is only one possible
constant solution for U that is consistent with the conservation of Q. Show that
solution corresponds to the bar approaching the average temperature of the initial
condition.

Page 33 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 13
Previously we showed that the diffusion kernel,
1 x2
U (x, t) = G(x, t + τ ) ≡ p e− 4D(t+τ ) , (15.3)
4πD(t + τ )

satisfies the diffusion equation with an initial condition

U (x, 0) = G(x, τ ). (15.4)

1. Show that the total heat ,

S
Z ∞
Q(t) = U (x, t) dx, (15.5)

N
−∞

A
is conserved, and in fact Q(t) = 1, for the heat kernel.

H
2. Show that as τ → 0, that G(x, τ ) → 0 for x 6= 0 and that G(0, τ ) → ∞.

IT
3. Explain why the solution G(x, t) (i.e. with τ = 0) corresponds to introducing
a unit amount of heat concentrated at the origin when t = 0. This is called a

H
δ-function initial condition.
AT
F
O
S

Example : 14
TE

A curious property of the diffusion equation is that both derivatives and integrals of solu-
tion satisfy the diffusion equation also. This can be used to generate new solutions from
O

existing ones.
N

1. Show that if U (x, t) satisfies the diffusion equation then ψ(x, t) = Ux satisfies
the diffusion equation also (Hint: differentiate both sides of the diffusion equation
E

with respect to x).

R

2. Generate a new solution for the heat equation by differentiating the heat kernel
TU

with respect to x. Graph this solution (MAPLE may be useful here) – what does it
look like?
C

3. ZShow that if ψ(x, t) satisfies the heat equation, then so does U (x, t) =
LE

x
ψ(ξ, t) dξ.
x0
4. Show that √
x/ 4c2 t
11
Z
2
U (x, t) = + √ e−z dz (15.6)
2 π 0

satisfies the heat equation, by showing its derivative is the heat kernel. Graph the
solution at various times – what physical problem does this solution correspond to
?

Page 34 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

16 The heat equation with a radiation boundary condition

In this lecture, we consider the initial boundary value problem (IBVP) with nonhomogeneous
boundary data,
ut = Kuxx : 0 < x < L, t>0

u(0, t) = T1   : t > 0,
ux (L, t) = −h u(L, t) − T2
u(x, 0) = T3 : 0 < x < L,
where h, T1 , T2 , T3 , and K are (strictly) positive constants. We have seen all of these ex-

S
pressions except for the boundary condition at x = L: this is called a radiation or Robin
condition. It describes how heat radiates radiates from the end into the surrounding medium.

N
The general form of a homogeneous Robin condition at x = a is

A
H
−kux (a, t) + βu(a, t) = 0 t > 0

IT
where k is the thermal conductivity of the bar (introduced in Lecture 2 (Bernoff)), a positive

H
constant. If β < 0 (and, of course, u(a, t) > 0), then heat flows into the bar, an absorption
AT
condition. If, as in our IBVP, β > 0, then heat flows out of the bar, a radiation condition.
Rewriting our IBVP so that the radiation condition is more readable, we see
F

ut = Kuxx : 0 < x < L, t > 0, (16.1)

O


u(0, t) = T1
: t > 0,
ux (L, t) + hu(L, t) = hT2
S

u(x, 0) = T3 : 0 < x < L.

TE
O

A straightforward application of the separation of variables technique that worked so well for the
N

heat equation with homogeneous Dirichlet or Neumann boundary data leads to a hard problem
(see Problem 1), so we will try to transform the problem into one we actually can handle.
E

Motivated by the examples shown in Lectures 2, 5 (Wittwer), and 8 (Vajiac) we will begin by
R

seeking a steady-state solution.

TU

16.1 The steady state solution

C
LE

Finding a steady state solution means that we seek a solution that is independent of time. That
is, find a function us (x) that satisfies ut = Kuxx on 0 < x < L, together with the boundary
data us (0) = T1 and u0s (L) + hus (L) = hT2 . Because {d/dt}(us (x)) = 0, we have
u00s (x) = 0.
Integrate twice to find us (x) = ax + b. Requiring that this line satisfies the boundary data
allows us to compute the integration constants a and b. After this work, we find the steady-state
solution:
h(T2 − T1 )
us (x) = x + T1 . (16.2)
1 + hL
Exercise: 1. Do the algebra to prove that this is the correct form of the steady solution.

Page 35 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

16.2 Homogeneous boundary conditions

Now we transform our problem by setting v(x, t) = u(x, t) − us (x). A straightforward

calculation (Problem 2) demonstrates that v must satisfy the new IBVP

vt = Kvxx : 0 < x < L, t>0


v(0, t) = 0
: t > 0,
vx (L, t) + hv(L, t) = 0
v(x, 0) = T3 − us (x) : 0 < x < L.

S
How is this an improvement? Notice that both boundary conditions are now homogeneous and
it is only the initial condition that varies in x. Guided by the previous lectures, we expect that

N
this non-constant v(x, 0) will not pose any difficulties.

A
The goal now is to find a solution v by separation of variables, and then to find a solution

H
u(x, t) = v(x, t) + us (x) of (16.1).

IT
H
16.3 Separation of variables
AT
Applying the standard separation of variables argument leads us to two ordinary differential
F

equations. For v(x, t) = X(x)T (t), we find, as in Lecture 5,

O

T 0 (t) X 00 (x)
S

= (16.3)
TE

KT (t) X(x)

which we set equal to −λ to find

O
N

T 0 (t) = −λKT (t)

X 00 (x) = −λX(x).
E
R

Again, the general solution for the time dependence has the form
TU

T (t) = Ce−λKt .
C
LE

16.4 A Sturm-Liouville Equation

From this point forth, we seek only nontrivial (nonzero) solutions v(x, t).
From the separation of variables argument T (t) = Ce−λKt . Note that T is nonzero for all t.
We now must find nontrivial solutions of the equation

− X 00 = λX. (16.4)

Recall the boundary conditions, for all t > 0

v(0, t) = 0 and vx (L, t) + hv(L, t) = 0.

Page 36 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Substituting v(x, t) = X(x)T (t) in each of these conditions provides

X(0)T (t) = 0 and X 0 (L)T (t) + hX(L)T (t) = 0
for all t > 0. Since T (t) is nonzero, divide these equations by T (t) to capture
X(0) = 0 and X 0 (L) + hX(L) = 0.
Putting all of this together leads to a Sturm-Liouville problem,
− X 00 = λX, X(0) = 0, X 0 (L) + hX(L) = 0. (16.5)
If a nontrivial solution of equation (16.5) exists, then the constant λ is called an eigenvalue and
the solution X is called its associated eigenfunction.

S
Postponing until Lecture 9 a discussion of the possibility of complex eigenvalues, we consider

N
three distinct cases:

A
H
i. λ < 0,

IT
ii. λ = 0, or

H
iii. λ > 0. AT
Let’s first consider the possibility that λ is negative.
F

16.4.1 Case i: λ < 0

O
S

A purely analytical approach is described in Appendix 19, but here we introduce a useful and
TE

more general approach called an energy argument. First, multiply both sides of −X 00 = λX
by X and integrate both sides with respect to x on the interval [0, L]:
O

Z L Z L
N

00
− X(x)X (x) dx = λ X 2 (x) dx.
0 0
E

Then integrate by parts to change the form of the left-hand side of this equation:
R
TU

L Z L  2 Z L 2
0 0

− X(x)X (x) +  X (x) dx = λ X(x) dx. (16.6)
0 0 0
C

However,
LE

L
−X(x)X (x) = −X(L)X 0 (L) + X(0)X 0 (0).
0

0
The boundary conditions require X(0) = 0 and X 0 (L) = −hX(L). Thus,
L
−X(x)X (x) = −X(L) − hX(L) + 0 · X 0 (0) = hX(L)2 ≥ 0.
0
  
0

Now equation (16.6) becomes

RL
hX 2 (L) + 0 (X 0 (x))2 dx
RL =λ (16.7)
X 2 (x)dx
0

from which it follows that λ ≥ 0.

Page 37 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

16.4.2 Case ii: λ = 0

If λ = 0, then −X 00 = λX becomes X 00 = 0, which has general solution

X(x) = A + Bx.

The boundary condition X(0) = 0 implies that A = 0, so the solution is now X(x) = Bx,
with derivative X 0 (x) = B. The second boundary condition provides

0 = X 0 (L) + hX(L) = B + h(BL) = B(1 + hL).

Because both h > 0 and L > 0, it follows that B = 0. Hence, X(x) = 0 for all x. Since

S
λ = 0 has only the trivial solution, we conclude that λ = 0 is not an eigenvalue.

N
A
16.4.3 Case iii: λ > 0

H
IT
Since λ > 0, we now know

H
√
√

X(x) = A cos x λ + B sin x λ .
AT
Applying the first boundary condition gives
F

0 = X(0) = A cos 0 + B sin 0 = A.

O

√

S

Thus, X(x) = B sin x λ . Notice that we must have B 6= 0 to obtain a nontrivial

TE

solution. The second boundary condition requires that

√ √
√

O

0 = X 0 (L) + hX(L) = B λ cos L λ + hB sin L λ .

N

√

Dividing through by B cos L λ shows
E

√ √

R

0= λ + h tan L λ .
TU

To obtain nontrivial solutions X(x), we must hope that this trigonometric equation has at least
C

one solution for λ. This is a perfect opportunity to use a computer application such as Maple.
LE

In fact, we will see that there are infinitely many solutions! Write these eigenvalues of the BVP
as λj such that p
p
λj
− tan L λj = , (16.8)
h
with corresponding eigenfunctions
p

Xj (x) = sin x λj .

Page 38 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

16.5 Orthogonality

The eigenvalues of the Sturm-Liouville problem (16.5) are the positive solutions of
√
√
λ
− tan L λ = (16.9)
h
and have associated eigenfunctions
p

Xj (x) = sin x λj . (16.10)

In the Fourier theory developed in Lecture 6, the eigenfunctions had the remarkable property of

S
orthogonality. Because we can find only approximate (numerical) solutions of (16.9), it will be
difficult to show that (for j 6= k)

N
A
Z L p
p

sin x λj sin x λk dx = 0,

H
0

IT
for our eigenfunctions (16.10) . Returning to the differential equation itself may provide more

H
insight into the nature of these eigenfunctions.
AT
Suppose that λ1 6= λ2 are two distinct eigenvalues with associated eigenfunctions X1 and X2
that both satisfy the Sturm-Liouville problem
F

−X 00 = λX, X(0) = 0, X 0 (L) + hX(L) = 0.

O

Thus,
S
TE

−X100 = λ1 X1
−X200 = λ2 X2 .
O
N

Multiply the first equation by X2 and the second by X1 and subtract:

E

−X2 X100 + X1 X200 = (λ1 − λ2 )X1 X2 .

R
TU

Integrate this expression with respect to x on the interval [0, L] to find

Z L Z L
C

00 00
− X2 X1 + X1 X2 ) dx = (λ1 − λ2 ) X1 X2 dx.
LE

0 0

Note that the integrand on the left is exact, so

Z L   Z L
d 0 0
X1 X2 − X2 X1 dx = (λ1 − λ2 ) X1 X2 dx.
0 dx 0

Thus, L Z L
X1 X20 X2 X10 

− = (λ1 − λ2 ) X1 X2 dx. (16.11)
0 0

Because X1 (0) = 0 and X2 (0) = 0,

L
X1 X2 − X2 X1  = X1 (L)X20 (L) − X2 (L)X10 (L).
0 0

(16.12)
0

Page 39 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

The next step is to argue that the right-hand side of equation (16.12) also equals zero.
Both X1 and X2 must satisfy the second boundary condition of the Sturm-Liouville equation.
Thus,

X10 (L) + hX1 (L) = 0

X20 (L) + hX2 (L) = 0.

This system can be written in matrix form:

 0    
X1 (L) X1 (L) 1 0
0 = .
X2 (L) X2 (L) h 0

S
The fact that this matrix equation has a nontrivial solution demands that the determinant of the

N
coefficient matrix equal zero. Thus,

A
 0 
X (L) X1 (L)

H
0= 0 1  = X 0 (L)X2 (L) − X 0 (L)X1 (L).
X2 (L) X2 (L) 1 2

IT
This last result, along with equations (16.11) and (16.12), provide

H
Z L AT
0 = (λ1 − λ2 ) X1 X2 dx.
0
F

Finally, because λ1 and λ2 are distinct eigenvalues,

O

Z L
S

0= X1 X2 dx,
TE

and X1 and X2 are orthogonal.

O
N

16.6 A formal solution and orthogonality

E
R

We may construct a formal solution for v by superposition:

TU

∞
X
Cj e−λj Kt sin x
p

v(x, t) = λj (16.13)
C

j=1
LE

We want this formal solution to satisfy the initial condition. This requires that
∞
X p
T3 − us (x) = v(x, 0) = Cj sin(x λj ). (16.14)
j=1

How do we know that such Cj exist?

We have shown that the eigenfunctions
 p
∞
sin x λj j=1
(16.15)

are pairwise orthogonal. (Refer to Problem 3 in which you will demonstrate via Maple that
these functions are orthogonal.) Further, this set forms a basis for L2 in which we may write

Page 40 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

our linear function v(x, 0) as a linear combination of elements of (16.15). (This idea will be
explored further in Lecture 9.)
We may obtain the “Fourier” coefficients Cj just as the Fourier
p coefficients were constructed in
Lecture 6 (Aarao). Multiply both sides of (16.14) by sin(x λk ) and integrate over the interval
[0, L]. This gives
Z L p
Z ∞
LX p
p

[T3 − us (x)] sin x λk dx = Cj sin x λk sin x λj dx.
0 0 j=1

Assuming uniform convergence, we may interchange the order of integration and summation.

S
By orthogonality, the right-hand side is non-zero only when j = k. Hence,

N
Z L p
Z L p

[T3 − us (x)] sin x λk dx = Ck sin2 x λk dx. (16.16)

A
0 0

H
We may now write

IT
RL √

0
[T3 − us (x)] sin x λk dx
Ck = √
(16.17)

H
RL
0
sin2 x λk dx AT
2
for integers k ≥ 1.
Now we have completely determined the series solution for v, and hence for u(x, t) =
F
O

v(x, t) + us (x):
S

∞  p  h(T − T )
2 1
X
−λj Kt
TE

u(x, t) = Cj e sin x λj + x + T1 (16.18)

j=1
1 + hL
O

with (λj , Cj ) solutions of equations (16.8) and (16.17), respectively.

N
E

16.7 Convergence
R
TU

How do we know that the formal series solution (16.18) converges uniformly on [0, L]? Further,
the series formulations of ut , ux , uxx obtained through term-by-term differentiation also ought
C

to converge uniformly. After all, in order to claim that a certain function u satisfies the IBVP, it
LE

had better be true that u is actually a function, differentiable once in t and twice in x!
Lemma 16.1. For all j ≥ 1, the coefficients Cj given by (16.17) satisfy |Cj | ≤ M for some
constant M .
Exercise: 2. Prove the lemma.
This claim allows us to prove that all of the series formulations for u, ut , ux , and uxx are
uniformly convergent on [0, L].
2
Using the neat inner product and norm notations introduced in Lecture 6, we may write equation (16.17) above
as √


v(x, 0), sin x λk
Ck = √
.
k sin x λk k2

Page 41 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

16.8 Long-term behavior (Asymptotics)

Finally, what happens to the series solution for u(x, t) as t → ∞? Since λj K > 0 we can
use the Lemma to show that
 
  p 
Cj e−λj Kt sin x λj  ≤ |Cj |e−λj Kt ≤ M e−λj Kt → 0 (16.19)
 
 

as t → ∞. By uniform convergence, then,

∞
X  p 
Cj e−λj Kt sin x λj −→ 0 as t → ∞. (16.20)

S
j=1

N
Thus, u(x, t) → us (x), the steady state solution!

A
H
IT
17 Summary: The Ten-Step Program

H
AT
At this point we review the procedure discussed in the previous pages.
1. Find a steady-state solution us .
F

2. Change variables by v = u − us , transforming our IBVP with inhomogeneous boundary

O

conditions to an IBVP with homogeneous boundary conditions.

S
TE

3. Apply separation of variables, using v(x, t) = X(x)T (t) to obtain two ODEs, in x
and in t, from
X 00 (x) T 0 (t)
O

= = −λ.
N

X(x) kT (t)

4. Solve for the t-dependence: T (t) = Ce−λKt .

E
R

5. Use an energy argument to determine sign of λ.

TU

6. Solve the x-ODE to determine pairs λj , Xj .
C

7. Create a formal solution v(x, t) by superposition and use orthogonality to determine the
LE

series coefficients.
8. Determine a formal solution for u = v + us .
9. Prove that the series expressions for u, ut , ux , and uxx converge uniformly. Thus, u in
series form is genuinely a solution.
10. Study long-term behavior. Show that

lim u(x, t) = us (x).

t→∞

Page 42 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 15
Use separation of variables to try to solve the IBVP

ut = Kuxx : 0 < x < L, t>0


u(0, t) = T1
: t > 0,
ux (L, t) + hu(L, t) = hT2
u(x, 0) = T3 : 0 < x < L.

Example : 16
Show that v(x, t) = u(x, t) − us (x) satisfies the IBVP

S
N
vt = Kvxx : 0 < x < L, t>0

A

v(0, t) = 0
: t > 0,

H
vx (L, t) + hv(L, t) = 0

IT
v(x, 0) = T3 − us (x) : 0 < x < L.

H
when u(x, t) satisifies AT
ut = Kuxx : 0 < x < L, t>0

u(0, t) = T1
F

: t > 0,
ux (L, t) + hu(L, t) = hT2
O

u(x, 0) = T3 : 0 < x < L.

S
TE

and us (x) is its steady state solution.

O

Example : 17
N

p

Use Maple to demonstrate that functions sin x λj are orthogonal when λj are the
E

solutions of the transcendental equation (16.8).

R
TU
C
LE

Page 43 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 18
(Logan, p. 131) Consider a large, circular, tubular ring of circumference 2L that contains
a chemical of concentration c(x, t) dissolved in water. Let x be the arc-length parameter
with 0 < x < 2L. If the concentration of the chemical is initially given by c0 (x), then
c(x, t) satisfies the IBVP

ct = Dcxx : 0 < x < 2L, t>0


c(0, t) = c(2L, t)
: t > 0,
cx (0, t) = cx (2L, t)
c(x, 0) = f (x) : 0 < x < 2L.

S
These boundary conditions are called periodic boundary conditions, and D is the dif-

N
fusion constant. Apply the separation of variables method and show that the associated

A
Sturm-Liouville problem has eigenvalues λn = (nπ/L)2 for n = 0, 1, 2, . . . and

H
eigenfunctions Xn (x) = An cos(nπx/L) + Bn sin(nπx/L) for n = 1, 2, . . . .
Show that the concentration is given by

IT
∞

H
A0 X 2 π 2 Dt/L2
c(x, t) = + (An cos(nπx/L) + Bn sin(nπx/L)) e−n
AT
2 n=0

and find the formulae for the An and Bn .

F
O

Example : 19
S

Write one dimensional wave equation and its possible solutions.

TE
O

Hints/Solution:
∂ 2u ∂ 2u
N

2
One dimensional wave equation is given by =c (or) utt = c2 uxx
∂t2 ∂x2
Its possible solutions are given by
E

(i) u(x, t) = (c1 epx + c2 e−px )(c3 epct + c4 e−pct ),

R

(ii) u(x, t) = (c5 cos px + c6 sin px)(c7 cos pct + c8 sin pct) and
TU

(iii) u(x, t) = (c9 x + c10 )(c11 t + c12 ).

C
LE

Example : 20
Write one dimensional heat equation and its possible solutions.

Hints/Solution:
∂u ∂ 2u
One dimensional heat equation is given by = c2 (or) ut = c2 uxx
∂t ∂x2
Its possible solutions are given by
2 2
(i) u(x, t) = (c1 epx + c2 e−px )ec p t ,
2 2
(ii) u(x, t) = (c3 cospx + c4 sin px)e−c p t and
(iii) u(x, t) = (c5 x + c6 ).

Page 44 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 21
∂ 2u ∂ 2u
In one dimensional wave equation = c2 (or) utt = c2 uxx what does c2
∂t2 ∂x2
stands for?.

Hints/Solution:
T
In one dimensional wave equation the value of c2 is stands for c2 = , where T is the tension
m

S
and m is mass per unit length.

N
A
H
Example : 22

IT
∂u 2
∂ 2u
(or) ut = c2 uxx what does c2 stands

H
In one dimensional heat equation =c
∂t ∂x2 AT
for?.

Hints/Solution:
F

k
In one dimensional heat equation the value of c2 is stands for c2 =
O

, where k is thermal
ρh
conductivity, ρ is the density of the material and h is the specific heat of the material.
S
TE
O
N

Example : 23
The ends of a rod of length 20 cm are maintained at the temperature 10◦ C and 20◦ C
E

respectively until steady state prevails. Determine the steady state temperature of the rod.
R
TU

Hints/Solution:
1-D heat equation is ut = c2 uxx .
C

In steady state, uxx = 0 =⇒ u(x) = ax + b.

LE

Applying given conditions, we get

Tl − T0 u(l) − u(0)
b = u(0) = T0 and a = =
l l
Tl − T0
=⇒ u(x) = x + T0 − − − − − (1)
l
Here, T0 = 10◦ C, Tl = 20◦ C and l = 20.
1
∴ (1) becomes u(x) = x + 10.
2

Page 45 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Example : 24
When the ends of a rod of length 30 cm are maintained at the temperature 20◦ C and
80◦ C respectively until steady state prevails. Determine the steady state temperature of
the rod.

Hints/Solution:
1-D heat equation is ut = c2 uxx .
In steady state, uxx = 0 =⇒ u(x) = ax + b.
Applying given conditions, we get
Tl − T0 u(l) − u(0)
b = u(0) = T0 and a = =

S
l l
Tl − T0

N
=⇒ u(x) = x + T0 − − − − − (1)
l

A
Here, T0 = 20◦ C, Tl = 80◦ C and l = 30.

H
∴ (1) becomes u(x) = 2x + 20.

IT
H
AT
Example : 25
A tightly stretched string with fixed end points x = 0 and x = l is initially at rest in its
F

equilibrium position. If it is set vibrating giving each point a velocity 3x(l − x), find
O

the displacement.
S
TE

Hints/Solution:
O

∂ 2u ∂ 2u
N

2
One dimensional wave equation is given by =c (or) utt = c2 uxx − − − − − (1)
∂t2 ∂x2
E

The boundary conditions becomes,

R

(i) u(0, t) = 0, t ≥ 0,
(ii) u(l, t) = 0 t ≥ 0,
TU

(iii) u(x, 0) = 0, 0 ≤ x ≤ l
∂u
C

(iv) (x, 0) = 3x(l − x), 0 ≤ x ≤ l

∂t
LE

The suitable solution is given by

u(x, t) = (c1 cos px + c2 sin px)(c3 cos pct + c4 sin pct) − − − − − (2).
Applying boundary
   (BC’s) (i) and (ii) in (2), we get
conditions
nπx nπct nπct
u(x, t) = c2 sin c3 cos + c4 sin − − − − − (3).
l l l

Applying boundary
   (BC) (iii) in
conditions  (3), we get
nπx nπct nπx nπct
u(x, t) = c3 sin c4 sin = cn sin sin − − − − − (4).
l l l l

Page 46 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Since cn is an arbitrary constant, and n is any integer and since (1) is homogeneous and linear,
the most general solution of (1) becomes,
∞
X nπx nπct
u(x, t) = cn sin sin − − − − − (5).
n=1
l l

Using BC (iv), we have

∞
∂u X nπa nπx
(x, 0) = cn sin = 3x(l − x), 0 ≤ x ≤ l
∂t n=1
l l

Expanding 3x(l − x) as Fourier series in (0, l), we have

∞
nπx

S
X
3x(l − x) = bn sin , then
l

N
n=1 
Zl 2 0 if n = even

A
nπc 2 nπx 12l
b n = cn = 3x(l−x) sin dx = 3 3 [1−(−1)n ] = 24l2

H
l l l n π  if n = odd
0 n3 π 3

IT

0 if n = even

H
=⇒ cn = 24l3
 if n = odd AT
cn4 π 4
∞
X 24l3 nπx nπct
u(x, t) = sin sin
F

4
cn π 4 l l
n=1,3,5,...
O

∞
X 24l3 (2r − 1)πx (2r − 1)πct
i.e. u(x, t) = sin sin
S

r=1
c(2r − 1)4 π 4 l l
TE
O
N

Example : 26
E

A rod of length l through which heat flows is insulated at its sides. The ends are kept at
R

zero temperature. If the initial temperature at the interior points of the bar is given by
TU

 
3 πx
k sin , find the temperature distribution in the bar after time t.
l
C
LE

Hints/Solution:
∂u ∂ 2u
One dimensional heat equation is given by = c2 (or) ut = c2 uxx − − − − − (1)
∂t ∂x2
The boundary conditions becomes,
(i) u(0, t) = 0, t ≥ 0,
(ii) u(l, t) = 0 t ≥ 0, 
3 πx
(iii) u(x, 0) = k sin , 0≤x≤l
l

The suitable solution is given by

2 2
u(x, t) = (c1 cos px + c2 sin px)e−c p t − − − − − (2).
Applying boundary conditions (BC’s) (i) and (ii) in (2), we get

Page 47 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

    
nπx 2 2
−c2 n 2π t nπx −c2 n22π2 t
u(x, t) = c2 sin e l = cn sin e l − − − − − (3).
l l

Since cn is an arbitrary constant, and n is any integer and since (1) is homogeneous and linear,
the most general solution of (1) becomes,
∞  
X nπx −c2 n22π2 t
u(x, t) = cn sin e l − − − − − (4).
n=1
l

Using BC (iii), we have

  ∞
3 πx X nπx
k sin = bn sin , then

S
l n=1
l

N
     ∞
k πx 3πx X nπx
sin − sin = bn sin ,

A
4 l l n=1
l

H
IT
Equating like co-efficients and solving for bn we get
3k πx −c2 π22 t k 3πx −c2 322π2 t

H
u(x, t) = sin e l − sin e l
4 l 4 l AT
F
O

Example : 27
The ends A & B of a rod of length ‘`’ units long have their temperature kept at
S

0◦ C & 120◦ C until steady state condition prevails. The temperature of the end B is
TE

suddenly reduced to 0◦ C and that of A is maintained to 0◦ C. Find the temperature

distribution in the rod after time t.
O
N
E

Hints/Solution:
∂u ∂ 2u
R

One dimensional heat equation is given by = c2 (or) ut = c2 uxx − − − (1)

TU

∂t ∂x2
In steady state, uxx = 0 =⇒ u(x) = ax + b. Applying given conditions, we get
Tl − T0 u(l) − u(0) Tl − T0
C

b = u(0) = T0 and a = = =⇒ u(x) = x + T0 −

l l l
LE

− − (1)
120
Here, T0 = 0◦ C, Tl = 120◦ C and length = l. ∴ (1) becomes u(x) = x.
l
The boundary conditions becomes,
(i) u(0, t) = 0, t ≥ 0,
t ≥ 0, 
(ii) u(l, t) = 0 
120x
(iii) u(x, 0) = , 0≤x≤l
l

The suitable solution is given by

2 2
u(x, t) = (c1 cos px + c2 sin px)e−c p t − − − (2).
Applying boundary conditions (BC’s) (i) and (ii) in (2), we get

Page 48 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

    
nπx 2 2
−c2 n 2π t nπx −c2 n22π2 t
u(x, t) = c2 sin e l = cn sin e l − − − (3).
l l
120x
Using BC (iii) and expanding as Fourier series in (0, l), we have
l
∞ 2 2 2
X 240 n+1 nπx − c n 2π t
`
Ans.: u(x, t) = (−1) sin e
n=1
nπ `

S
N
A
H
IT
H
AT
F
O
S
TE
O
N
E
R
TU
C
LE

Page 49 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

18 Exercise/Practice/Assignment Problems
1. Classify the following PDE’s
(a) 3uxx − 4uxy − 3uy = 0.
Ans.: B 2 − 4AC = 16 > 0. ∴ The PDE is hyperbolic ∀ x, y.
(b) y 2 uxx − 2xyuxy + x2 uyy − 3ux = 0.
Ans.: B 2 − 4AC = 0 ∀ x & y. ∴ Parabolic in this case.
(c) x2 uxx + 2xyuxy + (1 + y 2 )uyy − 2ux = 0.
Ans.: B 2 − 4AC = −4x2 < 0 for x < 0 or x > 0. ∴ Elliptic in this case.

S
If x = 0, then B 2 − 4AC = 0. ∴ Parabolic in this case.

N
(d) x2 uxx + (1 − y 2 )uyy = 0.
Ans.: B 2 − 4AC = 4x2 (1 − y 2 ) > 0 ∀x, x 6= 0 and y > 1, y < −1. ∴ The

A
PDE is hyperbolic in this case.

H
B 2 − 4AC = 4x2 (1 − y 2 ) < 0 ∀x, x 6= 0 and −1 < y < 1. ∴ The PDE is

IT
elliptic.

H
If x = 0 ∀y or ∀x, y = ±1, then B 2 − 4AC = 0. ∴ The PDE is parabolic in
AT
this case.
2. Derive the solution of one-dimensional wave equation.
F

3. Write down all the possible solutions of one-dimensional wave equation.

O

4. Derive the solution of one-dimensional heat-flow equation.

S
TE

5. Write down all the possible solutions of one-dimensional heat equation.

6. The points of trisection of a tightly stretched string of length ‘l’ units with fixed ends
O

pulled aside through a distance of ‘d’ units on opposite sides of the position of equilibrium
N

and the string is released from rest. Find the expression for the displacement of the string
at any subsequent time. Show also that the midpoint of the string remains always at rest.
E
R

Hint: Use Figure. 18.1 for the boundary conditions.

TU

∞
X 36d nπ nπx nπct
C

Ans.: u(x, t) = sin sin cos

n2 π 2 3 l l
LE

n=2,4,6,...
∞
X 9d 2rπ 2rπx 2rπct
i.e. u(x, t) = sin sin cos
r=1
r2 π2 3 l l
7. A tightly stretched string with fixed end points x = 0 and x = l is initially in a position
3πx 2πx
given by y(x, 0) = k sin cos . If it is released from rest from this position,
l l
determine the displacement y(x, t).
k πx πct k 5πx 5πct
Ans.: sin cos + sin cos
2 l l 2 l l
8. A tightly stretched string with fixed end points x = 0 and x = l is initially in a position
πx
given by y(x, 0) = k sin3 . If it is released from rest from this position, determine
l

Page 50 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

y
A(l/3, d)

O(0, 0) (2l/3, 0) C(l, 0)

x
(l/3, 0) M

B(2l/3, −d)

S
N
A
Figure 18.1: Trisection Graph.

H
IT
the displacement y(x, t).
3k πx πct k 3πx 3πct

H
Ans.: y(x, t) = sin cos − sin cos
4 l l 4 l l
AT
9. An elastic string is stretched and fastened to two points with distance ‘π’ apart. In its
initial position the string is in the shape of the curve y(x, 0) = k(sin x − sin3 x).
F

Obtain y(x, t).

O

k k
Ans.: y(x, t) = sin x cos ct − sin 3x cos 3ct
4 4
S
TE

10. A tightly stretched string with fixed end points x = 0 and x = l is initially at rest in
its equilibrium position. The string is in the form kx(l − x) and then released, find the
O

∞
X 8kl2 nπx nπct
displacement. Ans.: u(x, t) = sin cos
N

3
n π 3 l l
n=1,3,5,...
E

11. A tightly stretched string with its end points x = 0 and x = l are fastened. The
R

mid point of the string is displaced transversally through a small distance b and the
TU

string is released from rest from that position. Find an expression for the transverse
displacement of the string at any time during the subsequent motion. Ans.: u(x, t) =
∞
C

X 8b nπ nπx nπct
sin sin cos
LE

2
n π 2 2 l l
n=1

12. A tightly stretched string with its end points x = 0 and x = l are fastened. The
point x = l/3 of the string is displaced transversally through a small distance h and
the string is released from rest from that position. Find an expression for the transverse
displacement of the string at any time during the subsequent motion. Ans.: u(x, t) =
∞
X 9h nπ nπx nπct
sin sin cos
n=1
n2 π 2 3 l l
13. A tightly stretched string with its end points x = 0 and x = 2l are fastened. The
mid point of the string is displaced transversally through a small distance b and the
string is released from rest from that position. Find an expression for the transverse
displacement of the string at any time during the subsequent motion. Ans.: u(x, t) =

Page 51 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

∞
X 8b (n−1)
2
nπx nπct
(−1) sin cos
n=1,3,5,...
n2 π 2 l l

14. A tightly stretched string with its end points x = 0 and x = l is initially at rest in its
equilibrium position. If it is set vibrating giving each point a velocity kx(l − x). Find
∞
X 8kl3 nπx nπct
the displacement y(x, t). Ans.: u(x, t) = 4 4
sin sin
n=1,3,5,...
cn π l l

15. If a string of length ‘l’ is initially at rest in its equilibrium position and each of its points
l

ax,
 when 0 < x <
are given a velocity v = 2 . Find the displacement at
l

S
ax(l − x), when < x < l

2

N
∞
X 4al2 nπx (n−1)
2
nπct
any time t. Ans.: u(x, t) = sin (−1) sin

A
n=1,3,5,...
cn3 π 3 l l

H
16. If a string of length ‘l’ is initially at rest in its equilibrium position and each of its points

IT
(
ax, when 0 < x < l
are given a velocity v = . Find the displacement at

H
ax(l − x), when l < x < 2l AT
∞
X 16al nπx (n−1)
2
nπct
any time t. Ans.: u(x, t) = sin (−1) sin
n=1,3,5,...
cn3 π 3 2l 2l
F
O

17. A uniform string of length ‘l’ is struck in such a way that an initial velocity of v0 is
l 3l
imparted to the portion of the string between and while the string is in its equi-
S

4 4
TE

librium position. Find the displacement of the string at any time. Ans.: u(x, t) =
∞
X 4v0 l nπ nπ nπx nπct
O

sin sin sin sin

cn2 π 2 2 4 l 2l
N

n=1

18. A uniform string of length ‘l’ is struck in such a way that an initial motion is started by
E

3πx
displacing the string in the form 10 sin and also by imparting the string with initial
R

l
9πx
TU

velocity 15 sin . Find the displacement of the string at any time. Ans.: u(x, t) =
l
3πct 3πx 15l 9πct 9πx
C

10 cos sin + sin sin

l l 9πc l 2l
LE

19. A string is stretched and fastened to two points with distance ‘l’ apart. Motion is started
by displacing the string into the form of the curve u = x(l − x) and also by imparting a
constant velocity k to every point of the string in this position at time t = 0. Determine
the displacement of the string.
20. A uniform string of length ‘l’ is struck in such a way that an initial motion is started by
l
x,
 when 0 < x <
displacing the string in the form u = 2 and also by im-
l
(l − x), when < x < l

2
parting the string with initial velocity x(l − x). Find the displacement of the string at
any time.

Page 52 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

∞
8l3
 
X nπx 4l nπct nπ nπct
Ans.: u(x, t) = sin cos sin + sin
n=1,3,5,...
l n2 π 2 l 2 n4 π 4 c2 l

21. The ends A & B of a rod of length 10cm long have their temperature kept at 20◦ C & 70◦ C.
Find the steady state temperature distribution on the rod.
Ans.: u(x) = 5x + 20
22. When the ends of a rod of length 20 cm are maintained at the temperature 30◦ C and
70◦ C respectively until steady state prevails. Determine the steady state temperature of
the rod. Ans.: u(x) = 2x + 30
23. A rod of length ‘`’ units is heated so that its ends A and B are at zero temperature. If

S
kx(` − x)
initially the temperature is given by u = . Find the temperature distribution
`2

N
of the rod at any time t at a distance x.

A
∞ 2 2 2
X 8k nπx − c n`2π t
Ans.: u(x, t) = sin e

H
n=1,3,5,...
n3 π 3 `

IT
24. A rod of length ‘`’ units is heated so that its ends A and B are at zero temperature. If

H
initially the temperature is given by u = x. Find the temperature distribution of the rod
AT
at any time t at a distance x.
∞ 2 2 2
X 2` nπx − c n`2π t
Ans.: u(x, t) = sin e
nπ `
F

n=1
O

25. A rod of length ‘`’ units is heated so thatits ends A and B are at zero temperature. If
`
S

x,
 if 0 < x <
TE

initially the temperature is given by u = 2 . Find the tempera-

`
` − x, if

<x<`
O

2
ture distribution of the rod at any time t at a distance x.
N

∞ 2 2 2
X 4` nπ nπx − c n`2π t
Ans.: u(x, t) = sin sin e
n2 π 2 2 `
E

n=1
R

26. A homogeneous rod of conducting material of length ‘`’ units has its ends A and B are
TU

at zero temperature. The temperature at the center is T and falls uniformly to zero at the
two ends. Find the temperature distribution of the rod at any time t at a distance x.
C

∞ 2 2 2
X 8T nπ nπx − c n`2π t
Ans.: u(x, t) = sin sin e
LE

n=1
π2 2 `
27. The ends A & B of a rod of length ‘`’ units long have their temperature kept at 0◦ C & 100◦ C
until steady state condition prevails. The temperature of the end B is suddenly reduced to
0◦ C and that of A is maintained. Find the temperature distribution in the rod after time
t. ∞ 2 2 2
X 200 n+1 nπx − c n`2π t
Ans.: u(x, t) = (−1) sin e
n=1
nπ `
28. The ends A & B of a rod of length ‘`’ units long have their temperature kept at 0◦ C & 160◦ C
until steady state condition prevails. The temperature of the end B is suddenly reduced
to 0◦ C and that of A is maintained to 0◦ C. Find the temperature distribution in the rod

Page 53 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

after time t. ∞ 2 2 2
X 320 n+1 nπx − c n 2π t
`
Ans.: u(x, t) = (−1) sin e
n=1
nπ `
29. The ends A & B of a rod of length ‘30’ units long have their temperature kept at 20◦ C & 80◦ C
until steady state condition prevails. The temperature of the ends A and B are suddenly
reduced to 0◦ C. Find the temperature distribution in the rod after time t.
∞
X 40  n nπx − c2 n900
2 π2 t

Ans.: u(x, t) = 1 − 4(−1) sin e

n=1
nπ 30
30. The ends A & B of a rod of length ‘20’ units long have their temperature kept at 30◦ C & 90◦ C
until steady state condition prevails. The temperature of the ends A and B are suddenly

S
reduced to 0◦ C. Find the temperature distribution in the rod after time t.

N
∞
X 60  n nπx − c2 n400
2 π2 t

Ans.: u(x, t) = 1 − 3(−1) sin e

A
n=1
nπ 20

H
31. The ends A & B of a rod of length ‘10’ units long have their temperature kept at 50◦ C & 100◦ C

IT
until steady state condition prevails. The temperature of the end B is suddenly reduced to

H
60◦ C and that of A is increased to 90◦ C and the temperatures are maintained thereafter.
AT
Find the temperature distribution in the rod after time t.
∞
X 80 nπx − c2 n25
2 π2 t

Ans.: u(x, t) = −3x + 90 − sin e

nπ 5
F

n=1
O

32. The ends A & B of a rod of length ‘10’ units long have their temperature kept at 20◦ C & 40◦ C
until steady state condition prevails. The temperature of the end B is suddenly reduced to
S

10◦ C and that of A is increased to 50◦ C and the temperatures are maintained thereafter.
TE

Find the temperature distribution in the rod after time t.

∞
O

X 120 nπx − c2 n100

2 π2 t

Ans.: u(x, t) = −4x + 50 − sin e

N

n=2,4,...
nπ 10

33. The ends A & B of a rod of length ‘30’ units long have their temperature kept at 20◦ C & 80◦ C
E
R

until steady state condition prevails. The temperature of the end B is suddenly reduced to
60◦ C and that of A is increased to 40◦ C and the temperatures are maintained thereafter.
TU

Find the temperature distribution in the rod after time t.

∞
2 40 nπx − c2 n225
2 π2 t
C

X
Ans.: u(x, t) = x + 40 − sin e
LE

3 n=1
nπ 15
34. The ends A & B of a rod of length ‘l’ units long have their temperature kept at 40◦ C & 90◦ C
until steady state condition prevails. The temperature of the end B is suddenly reduced to
40◦ C and that of A is increased to 90◦ C and the temperatures are maintained thereafter.
Find the temperature distribution in the rod after time t.
∞ 2 2 2
50 X 200 nπx − c nl2π t
Ans.: u(x, t) = − x + 90 − sin e
l n=2,4,...
nπ l

35. The ends A & B of a rod of length ‘l’ units long have their temperature kept at a◦ C & b◦ C
until steady state condition prevails. The temperature of the end B is suddenly reduced
to d◦ C and that of A is increased to c◦ C and the temperatures are maintained thereafter.
Find the temperature distribution in the rod after time t.

Page 54 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

∞ 2 2 2
d−c
  X
2 n
nπx − c nl2π t
Ans.: u(x, t) = x+c + [(a − c) − (b − d)(−1) ] sin e
l n=1
nπ l
36. The temperature at one end of a bar ‘l’ units long and with insulated sides is kept at 0◦ C
and that at the other end is kept at 60◦ C until steady state condition prevails. The two
ends are then suddenly insulated, so that the temperature gradient is zero at each end
thereafter. Find the temperature distribution of the bar.
37. The ends A & B of a rod of length ‘l’ cm long have their temperature kept at 30◦ C & 80◦ C
until steady state condition prevails. The temperature of the end B is suddenly reduced
to 60◦ C and that of A is increased to 40◦ C. Find the temperature distribution in the rod
after time t.

S
N
A
19 Appendix

H
IT
Here is the analytical argument that the Sturm-Liouville problem

H
− X 00 = λX, X(0) = 0, X 0 (L) + hX(L) = 0
AT (19.1)

has only nonnegative eigenvalues. Suppose that λ < 0 and, so, write λ = −ω 2 , where
ω > 0. Then equation (19.1) becomes
F
O

X 00 − ω 2 X = 0.
S

It is easily seen that the general solution is3

TE

X = A cosh ωx + B sinh ωx.

O
N

The boundary condition X(0) = 0 forces

E

0 = X(0) = A cosh 0 + B sinh 0 = A,

R
TU

and X(x) = B sinh ωx, which has derivative X 0 (x) = Bω cosh ωx. The second bound-
ary condition now provides
C

0 = X 0 (L) + hX(L) = Bω cosh ωL + hB sinh ωL.

LE

Dividing by B,
0 = ω cosh ωL + h sinh ωL,
or equivalently,
ω
tanh ωL = − .
h
It is not difficult to show that the only solution of this equation is ω = 0, contradicting our
assumption that ω > 0.
Therefore, it cannot be the case that λ is negative. That is, λ must be nonnegative (λ ≥ 0).
3
The hyperbolic sine and cosine are defined as follows: cosh x = (ex + e−x )/2 and sinh x = (ex −
−x
e )/2.

Page 55 of 56 https://sites.google.com/site/lecturenotesofathithans/home
 18MAB201T-Transforms and Boundary Value Problems S. ATHITHAN

Acknowledgement:
Some of the portions of this material are taken from the sources available from various sources.
I thank the authors for those who prepared the calculus books and related materials.
Contact: (+91) 979 111 666 3 (or) athithan.s@ktr.srmuniv.ac.in
Visit: https://sites.google.com/site/lecturenotesofathithans/home

S
N
A
H
IT
H
AT
F
O
S
TE
O
N
E
R
TU
C
LE

Page 56 of 56 https://sites.google.com/site/lecturenotesofathithans/home