Assignment Solutions of Partial Dierential Equations

Weijiu Liu
Department of Mathematics
University of Central Arkansas
201 Donaghey Avenue, Conway, AR 72035, USA
1 Assignment 1
1.2.3. Derive the heat equation for a rod assuming constant thermal properties with variable
cross-sectional area A(x) assuming no sources.
Denote by A the the cross-sectional area.
Physical quantities:
Thermal energy density e(x, t) = the amount of thermal energy per unit volume.
Heat ux (x, t) = the amount of thermal energy owing across boundaries per unit
surface area per unit time.
Heat sources Q(x, t) = 0.
Temperature u(x, t).
Specic heat c = the heat energy that must be supplied to a unit mass of a substance
to raise its temperature one unit.
Mass density (x) = mass per unit volume.
Fouriers Law: the heat ux is proportional to the temperature gradient
= K
0
u. (1)
Conservation of heat energy:
Rate of change of heat energy in time = Heat energy owing across boundaries per unit
time + Heat energy generated insider per unit time
heat energy = e(x, t)A(x)x.
Heat energy owing across boundaries per unit time = (x, t)A(x) (x+x, t)A(x+
x).
Then

t
[e(x, t)A(x)x] = (x, t)A(x) (x + x, t)A(x + x).
Dividing it by x and letting x go to zero give
A(x)
e
t
= A(x)

x
(x)
A
x
. (2)
Heat energy per unit mass = c(x)u(x, t)Ax. So
e(x, t)A(x)x = c(x)u(x, t)A(x)x,
and then
e(x, t) = c(x)u(x, t).
2
It then follows from Fouriers law that
cA(x)
u
t
= A(x)

x
_
K
0
u
x
_
+K
0
A
x
u
x
. (3)
and then the heat equation
A(x)
u
t
= k
_
A(x)

2
u
x
2
+
A
x
u
x
_
, (4)
where k =
K
0
c
is the thermal diusivity.
1.2.9. Consider a thin one-dimensional rod without source of thermal energy whose lateral
surface is not insulated. Let w(x, t) dente the heat energy owing out of the lateral sides
per unit surface area per unit time. Assume that w(x, t) is proportional to the temperature
dierence between the rod u(x, t) and a known outside temperature (x, t). Derive the
equation for the temperature.
Denote by A the the cross-sectional area, and P the lateral perimeter.
Physical quantities:
Thermal energy density e(x, t) = the amount of thermal energy per unit volume.
Heat ux (x, t) = the amount of thermal energy owing across boundaries per unit
surface area per unit time.
Temperature u(x, t).
Specic heat c = the heat energy that must be supplied to a unit mass of a substance
to raise its temperature one unit.
Mass density (x) = mass per unit volume.
Fouriers Law: the heat ux is proportional to the temperature gradient
= K
0
u. (5)
Conservation of heat energy:
Rate of change of heat energy in time = Heat energy owing across boundaries per unit
time + Heat energy generated insider per unit time
heat energy = e(x, t)Ax.
Heat energy owing across boundaries per unit time = (x, t)A (x + x, t)A.
Heat energy owing out of the lateral sides per unit time = w(x, t)Px = [u(x, t)
(x, t)]h(x)Px, where h(x) is a proportionality.
3
Then

t
[e(x, t)A(x)x] = (x, t)A(x) (x + x, t)A(x + x) [u(x, t) (x, t)]h(x)Px.
Dividing it by Ax and letting x go to zero give
e
t
=

x

P
A
[u(x, t) (x, t)]h(x). (6)
Heat energy per unit mass = c(x)u(x, t)Ax. So
e(x, t)Ax = c(x)u(x, t)Ax,
and then
e(x, t) = c(x)u(x, t).
It then follows from Fouriers law that
c
u
t
=

x
_
K
0
u
x
_

P
A
[u(x, t) (x, t)]h(x). (7)
4
2 Assignment 2
1.4.1. Determine the equilibrium temperature distribution for a one-dimensional rod with
constant thermal properties with the following sources and boundary conditions:
(a) Q = 0, u(0) = 0, u(L) = T.
(f) Q = K
0
x
2
, u(0) = T, u

(L) = 0.
Solution. (a) Equilibrium satises
u

(x) = 0,
whose general solution is
u = c
1
+c
2
x.
The boundary condition u(0) = 0 implies c
1
= 0 and u(L) = T implies c
2
= T/L so that
u = Tx/L.
(f) In equilibrium, u satises
u

(x) = Q/K
0
= x
2
,
whose general solution (by integrating twice) is
u = x
4
/12 +c
1
+c
2
x.
The boundary condition u(0) = T yields c
1
= T, while u

(L) = 0 yields c
2
= L
3
/3. Thus
u = x
4
/12 +L
3
x/3 +T.
1.4.11. Suppose
u
t
=

2
u
x
2
+x, u(x, 0) = f(x),
u
x
(0, t) = ,
u
x
(L, t) = 7.
(a) Calculate the total thermal energy in the one-dimensional rod (as a function of time).
(b) From part (a), determine a value of for which an equilibrium exists. For this value
of , determine lim
t
u(x, t).
Solution. (a) Integrating the equation, we obtain:
d
dt
_
L
0
u(x, t)dx =
_
L
0
_

2
u
x
2
+x
_
dx =
u
x

L
0
+
1
2
L
2
= 7 +
1
2
L
2
.
Integrating in t from 0 to t, we obtain the total thermal energy
_
L
0
u(x, t)dx =
_
L
0
f(x)dx +
_
7 +
1
2
L
2
_
t. (8)
(b) In order for an equilibrium to exist,
_
7 +
1
2
L
2
_
t must be 0. So
= 7 +
1
2
L
2
.
5
The equilibrium satises

(x) +x = 0.
Its general solution (after integrating twice) is
=
1
6
x
3
+c
1
+c
2
x.
The boundary condition yields
c
2
= 7 +
1
2
L
2
.
So
=
1
6
x
3
+c
1
+
_
7 +
1
2
L
2
_
x.
Since
lim
t
u(x, t) = (x),
using (8), we obtain
_
L
0
f(x)dx =
_
L
0
u(x, t)dx
= lim
t
_
L
0
u(x, t)dx
=
_
L
0
(x)dx
=
_
L
0
_

1
6
x
3
+c
1
+
_
7 +
1
2
L
2
_
x
_
dx
=
1
24
L
4
+c
1
L +
1
2
_
7 +
1
2
L
2
_
L
2
.
Solving it gives
c
1
=
_
L
0
f(x)dx
7
2
L
2

5
24
L
4
L
,
and then
=
1
6
x
3
+
_
7 +
1
2
L
2
_
x +
_
L
0
f(x)dx
7
2
L
2

5
24
L
4
L
.
1.5.2. For conduction of thermal energy, the heat ux vector is = K
0
u. If in
addition the molecules move at an average velocity V, a process called convection, then =
K
0
u+cuV. Derive the corresponding equation for heat ow, including both conduction
and convection of thermal energy (assuming constant thermal properties with no sources).
Solution. Physical quantities:
Thermal energy density e(x, t) = the amount of thermal energy per unit volume.
Heat ux (x, t) = the amount of thermal energy owing across boundaries per unit
surface area per unit time.
6
Heat sources Q(x, t) = heat energy per unit volume generated per unit time.
Temperature u(x, t).
Specic heat c = the heat energy that must be supplied to a unit mass of a substance
to raise its temperature one unit.
Mass density (x) = mass per unit volume.
Conservation of heat energy:
Rate of change of heat energy in time = Heat energy owing across boundaries per unit
time + Heat energy generated insider per unit time
heat energy =
_
R
e(x, t)dV .
Heat energy owing across boundaries per unit time =
_
ndS.
Heat energy generated insider per unit time =
_
R
Q(x, t)dV = 0.
Then

t
_
R
e(x, t)dV =
_
ndS.
The divergence theorem give

t
_
R
e(x, t)dV =
_
R
dV.
and then
e
t
= . (9)
Heat energy per unit volume = c(x)u(x, t). So
e(x, t) = c(x)u(x, t).
It then follows that
c
u
t
= (cuV) + (K
0
u). (10)
and then the convection-diusion equation
u
t
+ (uV) = k
2
u, (11)
where k =
K
0
c
is called the thermal diusivity.
7
3 Assignment 3
2.3.2. (d) Find the eigenvalues and the corresponding eigenfunctions of the eigenvalue prob-
lem

d
2

dx
2
= ,
(0) = 0,
d
dx
(L) = 0.
(i) If > 0, = c
1
cos(

x) + c
2
sin(

x). (0) = 0 implies c

1
= 0, while
d
dx
(L) = 0
implies c
2

cos(

L) = 0. Thus

L =

2
+ n (n = 1, 2, ). Then the eigenvalues
are
n
=
_

2
+n
_
2
/L
2
and the corresponding eigenfunctions are
n
= sin
_
(

2
+n
)
x
L
_
(n = 1, 2, ).
(ii) If = 0, = c
1
+c
2
x. (0) = 0 implies c
1
= 0, while
d
dx
(L) = 0 implies c
2
= 0. Thus
= 0 is not an eigenvalue.
(ii) If < 0, = c
1
exp(

x) + c
2
exp(

x). (0) = 0 implies c

1
+ c
2
= 0, while
d
dx
(L) = 0 implies c
1

exp(

L) c
2

exp(

L) = 0. Solving this system for

c
1
, c
2
gives c
1
= c
2
= 0. Thus < 0 is not an eigenvalue.
2.3.3. (c) Solve the initial boundary value problems:
u
t
= k

2
u
x
2
, (12)
u(0, t) = 0, u(L, t) = 0, (13)
u(x, 0) = 2 cos
_
3x
L
_
. (14)
Solution. The problem has innite series solution
u(x, t) =

n=1
c
n
sin
_
nx
L
_
e

kn
2

2
t
L
2
.
The initial condition yields
2 cos
_
3x
L
_
= u(x, 0) =

n=1
c
n
sin
_
nx
L
_
.
So
c
n
=
2
L
_
L
0
2 cos
_
3x
L
_
sin
_
nx
L
_
dx.
2.3.8. Solution. (a) The equation for the equilibrium is
k
d
2
u
dx
2
u = 0, (15)
u(0) = 0, u(L) = 0. (16)
8
The general solution is u = c
1
exp(
_

k
x) + c
2
exp(
_

k
x). u(0) = 0 implies c
1
+ c
2
= 0,
while u(L) = 0 implies c
1
exp(
_

k
L) + c
2
exp(
_

k
L) = 0. Solving this system for c
1
, c
2
gives c
1
= c
2
= 0. Thus u = 0.
(b) Separation of variable, u = (x)G(t), yields two ODEs:
dG
dt
= (k +)G
and

d
2

dx
2
= ,
(0) = 0, (L) = 0.
The G-equation has solution
G(t) = Ce
t
e
kt
.
The eigenvalue problem has the eigenvalues

n
=
n
2

2
L
2
, n = 1, 2, (17)
and the corresponding eigenfunctions

n
= sin
_
nx
L
_
, n = 1, 2, (18)
Thus by superposition,
u(x, t) = e
t

n=1
c
n
sin
_
nx
L
_
e

kn
2

2
t
L
2
.
The initial condition gives
f(x) = u(x, 0) =

n=1
c
n
sin
_
nx
L
_
,
which gives
c
n
=
2
L
_
L
0
f(x) sin
_
nx
L
_
dx.
Since
lim
t
e

kn
2

2
t
L
2
= 0, lim
t
e
t
= 0,
we have
lim
t
u(x, t) = 0.
The u(x, t) converges to the only equilibrium 0.
9
4 Assignment 4
2.4.1. (a). The solution is
u(x, t) = a
0
+

n=1
a
n
cos
_
nx
L
_
e

kn
2

2
t
L
2
.
where
a
0
=
1
L
_
L
L/2
dx =
1
2
,
a
n
=
2
L
_
L
L/2
cos
_
nx
L
_
dx =
2
L

L
n
sin
_
nx
L
_

L
L/2
=
2
n
sin
_
n
2
_
.
2.4.1. (b). The solution is
u(x, t) = a
0
+

n=1
a
n
cos
_
nx
L
_
e

kn
2

2
t
L
2
.
where
a
0
=
1
L
_
L
0
_
6 + 4 cos
_
3x
L
__
dx = 6,
a
3
=
2
L
_
L
0
_
6 + 4 cos
_
3x
L
__
cos
_
nx
L
_
dx = 4,
and others are 0.
2.4.2. Solution.
u
t
= k

2
u
x
2
, (19)
u
dx
(0, t) = 0, u(L, t) = 0, (20)
u(x, 0) = f(x). (21)
Look for a solution of the form of separation of variables:
u(x, t) = (x)G(t), (22)
Substitute the above expression into the equation (19), we obtain
(x)G

(t) = k

(x)G(t),
and then
G

(t)
kG(t)
=

(x)
(x)
= , (23)
10
where is constant to be determined. The boundary condition (20) yields that
u
dx
(0) = (L) = 0.
We then have an eigenvalue problem

d
2

dx
2
= ,

dx
(0) = 0, (L) = 0.
Auxiliary equations:
m
2
= .
Case 1: < 0. Distinct real roots m
1
=

and m
2
=

:
(x) = c
1
e

x
+c
2
e

x
.
The boundary conditions imply that c
1
= c
2
= 0. So no non-zero solutions exist and
then < 0 is not an eigenvalue.
Case 2: = 0. Repeated real roots m
1
= m
2
= 0:
= c
1
+c
2
x.
The boundary conditions imply that c
1
= c
2
= 0. So no non-zero solutions exist and
then < 0 is not an eigenvalue.
Case 3: > 0. Conjugate complex roots m
1
= i

and m
2
= i

:
= c
1
cos(

x) +c
2
sin(

x).

(0) = 0 implies that c

2
= 0. (L) = 0 gives
cos(

L) = 0.
So

L =

2
+n (n = 0, 1, 2, ) and then we obtain the eigenvalues

n
=
_

2
+n
_
2
L
2
, n = 0, 1, 2, (24)
and the corresponding eigenfunctions

n
= cos
_
_

2
+n
_
x
L
_
, n = 0, 1, 2, (25)
11
On the other hand, it follows from (23) that
dG
dt
= kG, (26)
which has solutions
G(t) = ce
kt
= ce

2
+n)
2
kt
L
2
.
We then derive the innite series solution:
u(x, t) =

n=1
a
n
cos
_
_

2
+n
_
x
L
_
e

2
+n)
2
kt
L
2
.
The initial condition gives
f(x) = u(x, 0) =

n=1
a
n
cos
_
_

2
+n
_
x
L
_
. (27)
To determine a
n
, we multiply (27) by cos
_
(

2
+n
)
x
L
_
and integrate from 0 to L. We then
nd
a
n
=
2
L
_
L
0
f(x) cos
_
_

2
+n
_
x
L
_
dx, n 1. (28)
2.5.1. (c) Solve Laplaces equation

2
u
x
2
+

2
u
y
2
= 0, (29)
u
x
(0, y) = 0, u(L, y) = g(Y ), u(x, 0) = 0, u(x, H) = 0, (30)
Look for a solution of the form of separation of variables:
u(x, y) = h(x)(y), (31)
Substitute the above expression into the equation (29), we obtain
(y)h

(x) +

(y)h(x) = 0,
and then
h

(x)
h(x)
=

(y)
(y)
= , (32)
where is constant to be determined. The boundary condition yields that
(0) = (H) = 0, h

(0) = 0.
We then have an eigenvalue problem
12

d
2

dx
2
= ,
(0) = 0, (H) = 0,
which has the eigenvalues

n
=
n
2

2
H
2
, n = 1, 2, (33)
and the corresponding eigenfunctions

n
= sin
_
ny
H
_
, n = 1, 2, (34)
On the other hand, it follows from (32) that
h

(x) =
n
2

2
H
2
h(x), h

(0) = 0. (35)
The general solutions are
h(x) = c
1
e
nx
H
+c
2
e

nx
H
.
The boundary condition h

(0) = 0 gives
c
1
n
H
c
2
n
H
= 0.
So
h(y) = c
1
_
e
nx
H
+e

nx
H
_
.
We then derive the innite series solution:
u(x, t) =

n=1
a
n
sin
_
ny
H
_
_
e
nx
H
+e

nx
H
_
.
The boundary condition u(L, y) = g(y) gives
g(y) = u(L, y) =

n=1
a
n
sin
_
ny
H
__
e
nL
H
+e

nL
H
_
. (36)
To determine a
n
, we multiply the above by sin
_
ny
H
_
and integrate from 0 to H. We then
nd
a
n
=
2
_
e
nL
H
+e

nL
H
_
H
_
H
0
g(y) sin
_
ny
H
_
dy, n 1. (37)
13
5 Assignment 5
See the solutions in the textbook.
14