Arch Appl Mech (2009) 79: 679694

DOI 10.1007/s00419-008-0245-2
ORIGINAL
Amin Moosaie
Axisymmetric non-Fourier temperature eld in a hollow
sphere
Received: 18 September 2007 / Accepted: 9 May 2008 / Published online: 3 June 2008
Springer-Verlag 2008
Abstract The non-Fourier axisymmetric (2+1)-dimensional temperature eld within a hollow sphere is
analyticallyinvestigatedbythe solutionof the well-knownCattaneoVernotte hyperbolic heat conductionequa-
tion. The material is assumed to be homogeneous and isotropic with temperature-independent thermal proper-
ties. The method of solution is the standard separation of variables method. General linear time-independent
boundary conditions are considered. Ultimately, the presented solution is applied to a (1+1)as well as a
(2+1)dimensional problem, and their respective non-Fourier thermal behavior is studied. The present solu-
tion can be reduced to special cases of interest by choosing appropriate boundary conditions parameters.
Keywords Non-Fourier conduction Hyperbolic conduction Hollow sphere Analytical solution
1 Introduction
During the past few years there has been research concerned with departures from the classical Fourier heat
conduction law. The motivation for this research was to eliminate the paradox of an innite thermal wave
speed which is in contradiction with Einsteins theory of relativity and thus provide a theory to explain the
experimental data on second sound in liquid and solid helium at low temperatures [1, 2]. In addition to low
temperature applications, non-Fourier theories have attracted more attention in engineering sciences because
of their applications in high heat ux conduction, short time behavior as found, for example, in lasermaterial
interaction, etc.
The classical Fourier conduction law relates the heat ux vector q to the temperature gradient , by the
relation
q = , (1)
where the material constant is the thermal conductivity. Equation (1) along with the conservation of energy
gives the classical parabolic heat equation
=

t
, (2)
where = /c, , c and are thermal diffusivity, mass density, specic heat capacity and the Laplacian,
respectively. Equation (2) yields temperature solutions which imply an innite speed of heat propagation.
In order to eliminate this paradox, Cattaneo [3] and Vernotte [4] independently postulated a time-dependent
relaxation model for the heat ux in solids:
q +
0
q
t
= , (3)
Dedicated to Prof. Gholamali Ate, with appreciation and admiration on the occasion of his 65th birthday.
A. Moosaie
Department of Mechanical Engineering, Iran University of Science and Technology, Narmak, Tehran 16844, Iran
E-mail: aminmoosaie@gmail.com
680 A. Moosaie
where
0
is the so-called relaxation time (a non-negative constant). If the relaxation time
0
= 0, the heat ux
lawof the non-Fourier model dened by Eq. (3) reduces to the classical Fouriers model for heat conduction, i.e.,
Eq. (1). However, this model was rst proposed by Maxwell [5]. Inserting Eq. (3) into the energy conservation
equation, the hyperbolic heat transport equation, including source term, takes the form

t
2
+

t
= +

_
Q (x, t ) +
0
Q (x, t )
t
_
, (4)
where Q is the source term and

/
0
denotes the propagation speed of the temperature wave.
Various solution schemes for Eq. (4) with different initial and boundary conditions can be browsed for in
the literature. Baumeister and Hamill [6, 7] have given the solution for a semi-innite problem. Sadd and Cha
[8] have solved Eq. (4) for axisymmetric conduction in cylindrically bounded domains. Several solutions for
nite media have been given in [915]. Tang and Araki [16, 17] have solved the problem of non-Fourier heat
conduction in a nite medium under periodic surface disturbance. The periodic condition was a harmonic one.
Mikhailov and Cotta [18] solved the steady-periodic hyperbolic heat conduction in a nite slab with the use of
a computer algebra system Mathematica

. Abdel-Hamid [19] solved the non-Fourier heat conduction under

periodic surface disturbance using the nite integral transform. Tang and Araki [20, 21] investigated the non-
Fourier heat conduction in a nite mediumunder pulse surface heating. Wang [22] has investigated the solution
structure of the hyperbolic heat conduction equation. Glass et al. [23] presented an explicit analytical solution
for the hyperbolic heat conduction in a semi-innite solid subjected to a periodic onoff surface heat ux and
used a nite-difference scheme to solve the nonlinear cases with surface radiation. Yen and Wu [24] gave a
semi-analytical numerical solution of hyperbolic heat conduction in a nite medium with periodic thermal
disturbance and surface radiation. Their periodic condition was an onoff one. Lewandowska and Malinowski
[25] gave an analytical solution of the hyperbolic heat conduction for a nite medium symmetrically heated on
both sides. Moosaie has considered the non-Fourier heat conduction in a nite mediumunder arbitrary periodic
[26] and non-periodic [27] surface disturbances. Zhang and Liu [28] studied non-Fourier effects in a spherical
solid medium with either a sudden temperature change or a time-dependent pulsed heat ux prescribed on
the boundary surface. More recently, Jiang [29] investigated the hyperbolic heat conduction process in a
hollow sphere with its two boundary surfaces subjected to sudden temperature changes. He used the Laplace
transform method for an analytical solution of the problem. Trostel has investigated the axisymmetric Fourier
heat conduction in a spherical object, analytically [30].
This literature survey demonstrates that almost all of the research works done in the context of hyperbolic
heat conduction in spherical objects during past years are concerned with one-dimensional problems in the
radial direction under the inuence of special kinds of boundary and/or initial conditions or source terms. In this
paper however, the (2+1)-dimensional hyperbolic heat conduction equation in a hollow sphere is analytically
solved under the inuence of arbitrarily-chosen linear time-independent boundary conditions. The method of
solution is the well-known separation of variables method. Finally, the proposed solution is applied to a (1+1)
as well as a (2+1)-dimensional problem, and the behavior of the solution is studied for several choices of
boundary conditions and material properties.
2 Problem formulation
Let us consider a hollow sphere of inner and outer radii r
i
and r
o
, respectively, which is composed of an
isotropic heat conducting material with constant thermal properties. The boundary and initial conditions are
supposed to be such that axisymmetric two-dimensional heat conduction in the r and directions prevails.
In this case, the hyperbolic heat conduction equation (4) reduces to the following partial differential equation
in a spherical coordinate system. The spherical coordinate system used is depicted in Fig. 1.

t
2
+

t
=
_

r
2
+
2
r

r
+
1
r
2
_
cot

+

2

2
__
= . (5)
The following boundary and initial conditions are assumed:
a
1
(r
o
, , t ) +a
2

r
o
,,t
= f
o
() , (6a)
b
1
(r
i
, , t ) +b
2

r
i
,,t
= f
i
() , (6b)
Axisymmetric non-Fourier temperature eld in a hollow sphere 681
Fig. 1 Spherical coordinate system
(r, , 0) = 0, (7a)

r,,0
= 0. (7b)
where a
1
, a
2
, b
1
and b
2
are constants. Boundary conditions (6) are the most general case of linear boundary
conditions with time-independent non-homogeneities. Special cases can be achieved by different choices of
the parameters a
1
, a
2
, b
1
and b
2
.
The boundary conditions in the radial direction (6a) and (6b) are non-homogeneous. Therefore, it is not
possible to solve the problemdirectly. The solution structure for such problems is as follows [31]. The solution
is decomposed into two parts, a steady-state solution
0
(r, ) due to non-homogeneous boundary conditions
and a transient one
1
(r, , t ) with homogeneous boundary conditions but transformed initial conditions
[30, 31]. Thus, the complete solution is a superposition of these two steps which are presented in the sequel.
The present solution is of course an extension of the analytical solution done by Trostel [30] for classical
Fourier conduction in a hollow sphere. In the course of solution, use is frequently made of Bessel functions.
The interested reader is referred to Watsons excellent treatise [32].
3 Steady-state temperature eld
In order to fulll the non-homogeneous boundary conditions (6a) and (6b), we need to solve the following
steady-state boundary-value problem:

0
=

2

0
r
2
+
2
r

0
r
+
1
r
2
_
cot

+

2

2
_
= 0, (8)
a
1

0
(r
o
, ) +a
2

0
r

r
o
,
= f
o
() , (9a)
b
1

0
(r
i
, ) +b
2

0
r

r
i
,
= f
i
() . (9b)
Now, we solve this problem by the use of a separation ansatz as
0
(r, ) = R
0
(r)
0
(). Substituting this
ansatz into Eq. (8) yields
r
2
d
2
R
0n
dr
2
+2r
dR
0n
dr
n (n +1) R
0n
(r) = 0, (10)
d
2

0n
d
2
+cot
d
0n
d
+n (n +1)
0n
() = 0, (11)
682 A. Moosaie
where (10) is the well-known Eulers differential equation. Equation (11) can be rewritten as
1
sin
d
d
_
sin
d
0n
d
_
+n (n +1)
0n
() = 0.
Utilizing the transformation

0n
() =
0n
( ()) with () = cos , 1
2
= sin
2
and
d
d
= sin , (12a)
we obtain the Legendre differential equation in terms of .
d
d
_
_
1
2
_
d
0n
d
_
+n (n +1)
0n
() = 0. (12b)
Therefore, the steady-state temperature eld reads

0
(r, cos ) =
0
(r, ) =

n=0
R
0n
(r)
0n
() =

n=0
_
A
n
r
n
+
B
n
r
n+1
_
P
n
() , (13)
where P
n
() are the spherical functions, i.e. Legendre functions of the rst kind.
Now, we have to determine the coefcients A
n
and B
n
(n = 0, 1, 2, . . .) by enforcing the boundary
conditions (9). However, the non-homogeneities of boundary conditions are functions of while the tem-
perature eld is a function of through = cos . Therefore, writing the functions f
o
() and f
i
() as
f
o
= f
o
() = f
o
(cos ) and f
i
= f
i
() = f
i
(cos ), we have

A
n
_
a
1
r
n
o
+na
2
r
n1
o
_
+ B
n
_
a
1
r
(n+1)
o
(n +1) a
2
r
(n+2)
o
_
= c
(o)
n
,
A
n
_
b
1
r
n
i
+nb
2
r
n1
i
_
+ B
n
_
b
1
r
(n+1)
i
(n +1) b
2
r
(n+2)
i
_
= c
(i )
n
,
(14)
in which the functions f
o
() and f
i
() are expanded as
f
o
() =

n=0
c
(o)
n
P
n
() , (15a)
f
i
() =

n=0
c
(i )
n
P
n
() . (15b)
The coefcients c
(o)
n
and c
(i )
n
are determined fromthe orthogonality of spherical functions P
n
() on the interval
[1, +1], namely
+1
_
1
P
n
() P
m
() d =
_
0; m = n,
2
2n+1
; m = n,
(16)
which yields
c
(o)
n
=
2n +1
2
+1
_
1
f
o
() P
n
() d, (17a)
c
(i )
n
=
2n +1
2
+1
_
1
f
i
() P
n
() d. (17b)
Now, the constants A
n
and B
n
(n = 0, 1, 2, . . .) are obtained by solving the linear equation system (14).
A
n
=
(i )
n
c
(i )
n
+
(o)
n
c
(o)
n
, (18a)
B
n
=
(i )
n
c
(i )
n
+
(o)
n
c
(o)
n
, (18b)
Axisymmetric non-Fourier temperature eld in a hollow sphere 683
where

(o)
n
=
1

n
_
b
1
r
(n+1)
i
(n +1) b
2
r
(n+2)
i
_
, (19a)

(i )
n
=
1

n
_
a
1
r
(n+1)
o
(n +1) a
2
r
(n+2)
o
_
, (19b)

(o)
n
=
1

n
_
b
1
r
n
i
+nb
2
r
n1
i
_
, (19c)

(i )
n
=
1

n
_
a
1
r
n
o
+na
2
r
n1
o
_
, (19d)
in which

n
=
_
a
1
r
n
o
+na
2
r
n1
o
_
_
b
1
r
(n+1)
i
(n +1) b
2
r
(n+2)
i
_

_
a
1
r
(n+1)
o
(n +1) a
2
r
(n+2)
o
_ _
b
1
r
n
i
+nb
2
r
n1
i
_
.
(19e)
Thus, the steady-state temperature eld can be expressed as

0
(r, ) =

n=0

n
(r) P
n
() , (20)
in which

n
(r) =
_

(i )
n
c
(i )
n
+
(o)
n
c
(o)
n
_
r
n
+
_

(i )
n
c
(i )
n
+
(o)
n
c
(o)
n
_
r
(n+1)
= A
n
r
n
+
B
n
r
n+1
.
(21)
4 Transient temperature eld
So far, we were able to fulll our boundary conditions. Nevertheless, we have to satisfy the initial conditions
of our problem as well and, it cannot be done just with a steady-state solution. Therefore, we go further by
introducing a transient temperature eld
1
(r, , t ) which fullls the homogeneous boundary conditions and
also the initial conditions. Afterwards, the solution to the original problem will be constructed by summing
these two solutions up.
(r, , t ) =
0
(r, ) +
1
(r, , t ) , (22)
where
1
(r, , t ) is the solution of the following initial-boundary-value problem.

1
t
2
+

1
t
=
_

1
r
2
+
2
r

1
r
+
1
r
2
_
cot

+

2

2
__
=
1
, (23)
a
1

1
(r
o
, , t ) +a
2

1
r

r
o
,,t
= 0, (24a)
b
1

1
(r
i
, , t ) +b
2

1
r

r
i
,,t
= 0, (24b)

1
(r, , 0) =
0
(r, ) , (25a)

1
t

r,,0
= 0. (25b)
684 A. Moosaie
In order to solve this problem, use is made of a separation ansatz as
1
(r, , t ) = R (r) () T (t ). Replacing
this ansatz into (23) yields
1
R
_
d
2
R
dr
2
+
2
r
dR
dr
_
+
1
r
2
1

_
cot
d
d
+
d
2

d
2
_
=
1

0
T

T
+
T

T
_
=
2
. (26)
The conventional separation of variables routine gives

0
T

T
+
T

T
=
2
, (27a)
r
2
R
n
_
d
2
R
n
dr
2
+
2
r
dR
n
dr
+
2
R
n
_
=
1

n
_
cot
d
n
d
+
d
2

n
d
2
_
= n (n +1) . (27b)
Equation (27b) yields the following two ordinary differential equations for R
n
and
n
.
d
2
R
n
dr
2
+
2
r
dR
n
dr
+
_

n (n +1)
r
2
_
R
n
= 0, (28a)
d
2

n
d
2
+cot
d
n
d
+n (n +1)
n
= 0. (28b)
Equation (27a) is easy to solve. Equation (28b) is identical to Eq. (11) and thus its general solution is
known. Equation (28a) can be transformed into the Bessels differential equation for f
n
(r) by the use of the
Lommels transformation [33] which is
R
n
(r) =
1

r
f
n
(r) . (29)
Hence, the transient temperature eld
1
(r, , t ) can be written as

1
(r, , t ) = e

t
2
0
(E
1
cos t + E
2
sin t )
1

n=0
_
A
n
J
n+1/2
(r) + B
n
J
(n+1/2)
(r)
_
P
n
() , (30)
where = cos and
=
_
4
0

2
1
2
0
. (31)
Now, we have to enforce the homogeneous boundary conditions (24a) and (24b). They can be expressed
in terms of R
n
(r) as
a
1
R
n
(r
o
) +a
2
dR
n
dr

r
o
= 0, (32a)
b
1
R
n
(r
i
) +b
2
dR
n
dr

r
i
= 0, (32b)
which yield the following system of linear algebraic equations for A
n
and B
n
(n = 0, 1, 2, . . .).

A
n
_
a
1
J
n+1/2
(r
o
) +
a
2
2r
o
_
2r
o
J

n+1/2
(r
o
) J
n+1/2
(r
o
)
__
+B
n
_
a
1
J
(n+1/2)
(r
o
) +
a
2
2r
o
_
2r
o
J

(n+1/2)
(r
o
) J
(n+1/2)
(r
o
)
__
= 0,
A
n
_
b
1
J
n+1/2
(r
i
) +
b
2
2r
i
_
2r
i
J

n+1/2
(r
i
) J
n+1/2
(r
i
)
__
+B
n
_
b
1
J
(n+1/2)
(r
i
) +
b
2
2r
i
_
2r
i
J

(n+1/2)
(r
i
) J
(n+1/2)
(r
i
)
__
= 0.
(33)
Axisymmetric non-Fourier temperature eld in a hollow sphere 685
Equation system (33) is a homogeneous one and in order for this system to have non-trivial solutions, the
determinant of the coefcients matrix must vanish which gives the following transcendental equation:

n
() =
_
a
1
J
n+1/2
(r
o
) +
a
2
2r
o
_
2r
o
J

n+1/2
(r
o
) J
n+1/2
(r
o
)
_
_

_
b
1
J
(n+1/2)
(r
i
) +
b
2
2r
i
_
2r
i
J

(n+1/2)
(r
i
) J
(n+1/2)
(r
i
)
_
_

_
a
1
J
(n+1/2)
(r
o
) +
a
2
2r
o
_
2r
o
J

(n+1/2)
(r
o
) J
(n+1/2)
(r
o
)
_
_

_
b
1
J
n+1/2
(r
i
) +
b
2
2r
i
_
2r
i
J

n+1/2
(r
i
) J
n+1/2
(r
i
)
_
_
= 0. (34)
For each specic n (n = 0, 1, 2, . . .), Eq. (34) gives the eigenvalues
kn
(k = 0, 1, 2, . . .). Therefore, the
transient temperature eld can be written as

1
(r, , t ) =

n=0

k=0
e

t
2
0
_

A
kn
cos
kn
t +

B
kn
sin
kn
t
_

kn
(
kn
r) P
n
() , (35)
where

kn
=
_
4
0

2
kn
1
2
0
, (36a)

kn
(
kn
r) =
1

r
__
a
1
J
(n+1/2)
(
kn
r
o
) +
a
2
2r
o
_
2
kn
r
o
J

(n+1/2)
(
kn
r
o
) J
(n+1/2)
(
kn
r
o
)
_
_
J
n+1/2
(
kn
r)
_
a
1
J
n+1/2
(
kn
r
o
) +
a
2
2r
o
_
2
kn
r
o
J

n+1/2
(
kn
r
o
) J
n+1/2
(
kn
r
o
)
_
_
J
(n+1/2)
(
kn
r)
_
. (36b)
The coefcients

A
kn
and

B
kn
must be determined such that the temperature eld (35) fullls the initial
conditions (25a) and (25b). From (35), (25a) and (20) we have

1
(r, , 0) =

n=0
_

k=0

A
kn

kn
(
kn
r)
_
P
n
() =

n=0

n
(r) P
n
() , (37a)
which yields

k=0

A
kn

kn
(
kn
r) =
n
(r) . (37b)
Since the eigenfunctions
kn
(
kn
r) are orthogonal with respect to the weighting function r
2
, namely
r
o
_
r
i
r
2

kn
(
kn
r)
j n
_

j n
r
_
dr =
_
0; j = k,

kn
; j = k,
(37c)
where

kn
=
_
r
2
2
kn
__

2
kn
r
2
n (n +1)
_

2
kn
+
kn
r
kn

kn
+
2
kn
r
2

2
kn
_
_
r
o
r
i
, (37d)
686 A. Moosaie
in which derivatives are carried out with respect to
kn
r, we can assume that (37b) is the series representation
of
n
(r) in terms of
kn
(
kn
r). Thus we have

A
kn
=
1

kn
r
o
_
r
i
r
2

n
(r)
kn
(
kn
r) dr. (37e)
The coefcients

B
kn
are obtained from initial condition (25b).

1
t

r,,0
=

n=0

k=0
_

1
2
0

A
kn
+
kn

B
kn
_

kn
(
kn
r) P
n
() = 0, (38a)
which gives

B
kn
=
1
2
0

kn

A
kn
. (38b)
Using the coefcients

A
kn
and

B
kn
(k, n = 0, 1, 2, . . .) dened in (37e) and (38b), we can rewrite the transient
temperature eld (35) as

1
(r, , t ) =

n=0

k=0
1

kn
e

t
2
0
_
cos
kn
t +
1
2
0

kn
sin
kn
t
_

kn
(
kn
r) P
n
()

r
o
_
r
i
r
2

n
(r)
kn
(
kn
r) dr. (39)
5 Superposition, complete temperature eld
Recalling equation (22), one can obtain the complete non-Fourier temperature eld by superposing the steady-
state and transient solutions which will be done in this section.
(r, , t ) =
0
(r, ) +
1
(r, , t )
=

n=0

n
(r)

k=0

kn
e

t
2
0
_
cos
kn
t +
1
2
0

kn
sin
kn
t
_

kn
(
kn
r)

r
o
_
r
i
r
2

n
(r)
kn
(
kn
r) dr

P
n
() , (40)
which can be written as
(r, , t ) =

n=0

n
(r)

k=0
1

kn

kn
(t )
kn
(
kn
r)
r
o
_
r
i
r
2

n
(r)
kn
(
kn
r) dr

P
n
() , (41a)
where

kn
(t ) = e

t
2
0
_
cos
kn
t +
1
2
0

kn
sin
kn
t
_
. (41b)
Moreover, by dening the auxiliary functions

(i )
n
(r) =
(i )
n
r
n
+

(i )
n
r
n+1
, (42a)

(o)
n
(r) =
(o)
n
r
n
+

(o)
n
r
n+1
, (42b)
Axisymmetric non-Fourier temperature eld in a hollow sphere 687
one has
(r, , t ) =

n=0

(i )
n
(r)

k=0
1

kn

kn
(t )
kn
(
kn
r)
r
o
_
r
i
r
2

(i )
n
(r)
kn
(
kn
r) dr

c
(i )
n
P
n
()
+

n=0

(o)
n
(r)

k=0
1

kn

kn
(t )
kn
(
kn
r)
r
o
_
r
i
r
2

(o)
n
(r)
kn
(
kn
r) dr

c
(o)
n
P
n
()
=
(i )
(r, , t ) +
(o)
(r, , t ) . (43)
6 Demonstrations
In order to demonstrate the applicability of the presented solution procedure, we proceed to solve two problems
as special cases in this section. A C++ code was developed to evaluate the analytical solution numerically in
which the use was made of the special functions (Bessel and Legendre functions) provided by the GNU
Scientic Library (GSL).
6.1 Demonstration case 1
As the rst example, a radial one-dimensional non-Fourier heat conduction process in a hollow sphere with
an inner radius of r
i
and outer radius of r
o
is studied here. Constant thermal properties are considered. The
thermal disturbance is caused by a sudden change in heat ux on its outer surface while the inner surface is
kept at the initial temperature. No heat source is involved and heat convection and radiation are disregarded.
Boundary and initial conditions are as follows.
(r
i
, , t ) = 0,

r
o
,,t
= q
0
, (44a)
(r, , 0) = 0,

r,,0
= 0. (44b)
The presented solution reduces to the following temperature eld for this problem:
(r, t ) =

(o)
0
(r)

k=0
1

k0

k0
(t )
k0
(
k0
r)
r
o
_
r
i
r
2

(o)
0
(r)
k0
(
k0
r) dr

q
0

=
(o)
(r, t ) , (45)
where

(o)
0
(r) = r
2
o
_
1
r
i

1
r
_
, (46a)

k0
(
k0
r) =
1

r
__
1
2r
o
_
2
k0
r
o
J

1/2
(
k0
r
o
) J
1/2
(
k0
r
o
)
_
_
J
1/2
(
k0
r)

_
1
2r
o
_
2
k0
r
o
J

1/2
(
k0
r
o
) J
1/2
(
k0
r
o
)
_
_
J
1/2
(
k0
r)
_
, (46b)

k0
=
_
r
2
2
k0
_

2
k0
r
2

2
k0
+
k0
r
k0

k0
+
2
k0
r
2

2
k0
_
_
r
o
r
i
. (46c)
688 A. Moosaie
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
r

Fo = 0.1
Fo = 0.2
Fo = 1.0
Fo = 2.0
Fo = 0.5
o steadystate
Fig. 2 Temperature proles for q
0
= 1, r
i
= 0.6, r
o
= 1 and Ve
2
= 0.2 for various Fo
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
r

Fo = 0.1
Fo = 0.2
Fo = 2.0
Fo = 0.5
Fo = 10.0
Fo = 1.0
o steadystate
Fig. 3 Temperature prole for q
0
= 1, r
i
= 0.6, r
o
= 1 and Ve
2
= 1 for various Fo
The eigenvalues
k0
are the roots of the following transcendental equation:

0
() =
1
2r
o
_
2r
o
J

1/2
(r
o
) J
1/2
(r
o
)
_
J
1/2
(r
i
)

1
2r
o
_
2r
o
J

1/2
(r
o
) J
1/2
(r
o
)
_
J
1/2
(r
i
) = 0. (46d)
The results of this solution are depicted in Figs. 2, 3, 4, 5. The dimensionless parameters Fo, Ve and are
dened as
Fo =
t
r
2
o
, Ve
2
=

0
r
2
o
, =

q
0
r
o
/
. (47)
Figure 2 shows the temperature proles for q
0
= 1, r
i
= 0.6, r
o
= 1 and Ve
2
= 0.2. The temperature
gradually changes from the initial temperature to the steady-state temperature eld which is given by

ss
(r) = 1.

6
1
r
, (48)
for the above values. The temperature almost reaches the steady-state at Fo = 2. Figure 3 shows the proles
of the same problem but for Ve
2
= 1. As is expected, the process is slower and the temperature almost reaches
Axisymmetric non-Fourier temperature eld in a hollow sphere 689
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
r

Fo = 0.04
Fo = 0.1
Fo = 0.2
Fo = 0.5
o steadystate
Fig. 4 Temperature prole for q
0
= 1, r
i
= 0.6, r
o
= 1 and Ve
2
= 0 (Fourier conduction) for various Fo
the steady-state at Fo = 10. The extreme case of Fourier conduction, i.e. Ve
2
= 0, is depicted in Fig. 4 for the
same problem. Now, the nite propagation speed of non-Fourier model and the innite propagation speed of
the Fourier one is obvious by comparing Figs. 3, 4, 5. In Fourier conduction, the temperature has reached the
steady-state at Fo
2
= 0.5.
There are other interesting differences. In the Fourier case, the temperature proles are smooth while in the
non-Fourier case there are some kinks which show the hyperbolic and wave characteristic of the non-Fourier
model. Moreover, In the Fourier case, the temperature prole is always between the initial and the steady-state
proles. However, in the non-Fourier case, the temperature exceeds the steady-state prole and converges to
the steady-state in an oscillatory manner. The amplitude of oscillations is greater for greater values of Ve which
is better shown in Fig. 5.
Another phenomenon which is shown in Figs. 2, 3, 4, 5 is the lag in the temperature change at r = r
i
in
the non-Fourier case. The inner layers remain at their initial temperature for a while, and it takes some time
for the information to travel from the outer surface, where we applied the disturbance, to the inner layers. The
lag time is smaller for smaller values of Ve and it is zero for the Fourier case (Ve = 0).
Figure 5 shows the time history of temperature at r = 0.8 for the same values used in Figs. 2, 3, 4 for
Fo [0, 5]. The oscillatory convergence to the steady-state for the non-Fourier case, the fast and non-oscillatory
convergence of the Fourier one, and the kinks are quite obvious. Our solution reaches the steady-state value
of 0.41667 which is practically equal to the value of 0.41

6 obtained from the steady-state solution (48). The

small difference is due to the machine-dependent round-off error as well as the truncation error associated with
the computation of an innite series as a truncated nite series on a specic machine. Moreover, it can be seen
that the frequency of oscillations is smaller for the higher values of Ve. The amplitude trend is as mentioned
before.
6.2 Demonstration case 2
As the second example, a (2+1)-dimensional non-Fourier heat conduction process in a hollow sphere with
an inner radius of r
i
and outer radius of r
o
is studied here. Constant thermal properties are considered. The
thermal disturbance is caused by a sudden change in heat ux on its outer surface while the inner surface is
kept at the initial temperature. The applied heat ux in this case is a function of . No heat source is involved
and, heat convection and radiation are disregarded. Boundary and initial conditions are as follows:
(r
i
, , t ) = 0,

r
o
,,t
= q () , (49a)
(r, , 0) = 0,

r,,0
= 0. (49b)
690 A. Moosaie
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Fo

Fourier
Ve
2
= 0.2
Ve
2
= 1.0
Ve
2
= 2.0
Fig. 5 Temperature history for q
0
= 1, r
i
= 0.6, r
o
= 1 at r = 0.8 for various Ve
2
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.05
0.1
0.15
0.2
0.25
r

Fo = 0.5
Fo = 1.0
Fo = 2.0
Fo = 3.0
Fig. 6 Temperature proles for q = q
0
| cos |, r
i
= 0.6, r
o
= 1, Ve
2
= 0.2 at = /2 for various Fo
The presented solution reduces to the following temperature eld for this problem:
(r, , t ) =

n=0

(o)
n
(r)

k=0
1

kn

kn
(t )
kn
(
kn
r)
r
o
_
r
i
r
2

(o)
n
(r)
kn
(
kn
r) dr

c
(o)
n
P
n
()
=
(o)
(r, , t ) , (50a)
where = cos and

kn
(
kn
r) =
1

r
__
1
2r
o
_
2
kn
r
o
J

(n+1/2)
(
kn
r
o
) J
(n+1/2)
(
kn
r
o
)
_
_
J
n+1/2
(
kn
r)

_
1
2r
o
_
2
kn
r
o
J

n+1/2
(
kn
r
o
) J
n+1/2
(
kn
r
o
)
_
_
J
(n+1/2)
(
kn
r)
_
. (50b)

kn
is given by (37d).
(o)
n
(r) depends on the special choice of q (). The eigenvalues
kn
are the roots of the
transcendental equation (34) by setting a
1
= 0, a
2
= 1, b
1
= 1 and b
2
= 0. Here we consider the following
choices for q ().
q () = q
0
cos (m), q () = q
0
| cos |, m = 1, 2.
Axisymmetric non-Fourier temperature eld in a hollow sphere 691
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.05
0.1
0.15
0.2
0.25
0.3
r

Fo = 0.5
Fo = 1.0
Fo = 2.0
Fo = 3.0
Fo = 4.0
Fo = 5.0
Fig. 7 Temperature proles for q = q
0
| cos |, r
i
= 0.6, r
o
= 1, Ve
2
= 1.0 at = /2 for various Fo
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.05
0.1
0.15
0.2
0.25
0.3
r

Fo = 0.5
Fo = 1.0
Fo = 2.0
Fo = 3.0
Fig. 8 Temperature proles for q = q
0
cos , r
i
= 0.6, r
o
= 1, Ve
2
= 0.2 at = /3 for various Fo
The results of this solution are depicted in Figs. 3, 4, 5, 6. The denitions of Fo and Ve are given in (47) while
is dened here as
=

q
0
r
o
/
.
The results of this solution are depicted in Figs. 6, 7, 8, 9, 10, 11, 12. The geometrical parameters used are
r
i
= 0.6 and r
o
= 1. To check whether the (2+1)-dimensional code is working properly, it was used to compute
the case of q = q
0
cos (m) with m = 0, which is identical to the demonstration case 1. It reached to the
steady-state value of 0.41667 which is practically the same as the exact value of 0.41

6.
Figures 6 and 7 show the temperature proles for the case of q = q
0
| cos | at = /2 for Ve
2
= 0.2
and Ve
2
= 1. Again, the oscillatory behavior is obvious, i.e., the temperature exceeds the steady-state values.
Moreover, at Ve
2
= 0.2 the thermal wave propagation takes place much faster than Ve
2
= 1. The temperature
for Ve
2
= 0.2 almost reaches the steady-state at Fo = 3.0, while for Ve
2
= 1 it is not yet reached at Fo = 5.
Figures 8 and 9 show the temperature proles for the case of q = q
0
cos at = /3. The same
phenomena is observable as they are shown in Figs. 6 and 7. The temperature remains zero at = /2.
The reason is the geometrical symmetry and the anti-symmetry of the cosine function about = /2. The
occurrence of kinks is also obvious. Figure 10 shows the temperature proles for q = q
0
cos (2). Again,
the same trends are observable. The temperature is negative in this case which is due to the sign changes in
the applied heat ux q = q
0
cos (2). Figure 11 represents the temperature proles with xed Fourier and
692 A. Moosaie
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
r

Fo = 3.0
Fo = 4.0
Fo = 5.0
Fo = 0.5
Fo = 1.0
Fo = 2.0
Fig. 9 Temperature proles for q = q
0
cos , r
i
= 0.6, r
o
= 1, Ve
2
= 1.0 at = /3 for various Fo
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0.6
0.5
0.4
0.3
0.2
0.1
0
r

Fo = 0.5
Fo = 1.0
Fo = 2.0
Fo = 3.0
Fig. 10 Temperature proles for q = q
0
cos (2), r
i
= 0.6, r
o
= 1, Ve
2
= 0.2 at = /2 for various Fo
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
0.5
0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
0.5
r

= 0
= / 6
= / 4
= / 3
= / 2
Fo = 0.5
Fig. 11 Temperature proles for q = q
0
cos (2), r
i
= 0.6, r
o
= 1, Ve
2
= 1.0 at Fo = 0.5 for various
Axisymmetric non-Fourier temperature eld in a hollow sphere 693
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.05
0.1
0.15
0.2
0.25
Fo

Fourier
Ve
2
= 0.2
Ve
2
= 1.0
Fig. 12 Temperature history for q = q
0
| cos |, r
i
= 0.6, r
o
= 1 at r = 0.8 and = /2 for various Ve
2
Vernotte numbers, i.e. Fo = 0.5 and Ve
2
= 1, versus radius for various values of which gives a better idea
about the temperature distribution in 2D space. Due to the symmetry of thermal loading about = /2, plots
are just given for 0 /2.
Figure 12 shows the temperature history for Ve
2
= 0 (Fourier conduction), Ve
2
= 0.2 and Ve
2
= 1.
Again, the lag in inner layers temperature for non-Fourier conduction is observable. The oscillatory nature
of hyperbolic conduction and the non-oscillatory nature of the Fourier one is obvious. The propagation takes
place faster for smaller values of Ve
2
.
7 Concluding remarks
A brief literature review of analytical solutions to the non-Fourier (hyperbolic) heat conduction equation is
presented. Then the hyperbolic heat conduction equation is analytically solved for the problemof axisymmetric
heat transfer in a hollow sphere under the inuence of arbitrarily-chosen linear time-independent boundary
conditions. The principle of superposition decomposes the solution into two parts; a steady-state part which
fullls the non-homogeneous boundary conditions and a transient solution which satises the initial conditions.
Ultimately, as for demonstration, the proposed solution is applied to a (1+1)as well as a (2+1)dimensional
problems, and their thermal behavior is studied. This paper gives the solution structure of hyperbolic heat
conduction in an axisymmetric spherical coordinate system and the given general solution can be reduced to
special cases of interest by assigning appropriate values to boundary conditions parameters. The results can
be used for problems of technical interest, and as benchmarks to validate the numerical solutions.
Acknowledgments The anonymous referees whose comments improved the quality of the present work are acknowledged.
Many thanks also go to the providers of the GNU Scientic Library (GSL) which has been used for the numerical evaluation of
the present analytical solution.
References
1. Chester, M.: Second sound in solids. Phys. Rev. 131, 20132015 (1963)
2. Brown, J.B., Chung, D.Y., Matthews, P.W.: Heat pulses at low temperatures. Phys. Lett. 21, 241243 (1966)
3. Cattaneo, C.: Sur une forme de lequation de la chaleur elinant le paradoxe dune propagation instantance. C. R. Acad.
Sci. 247, 431432 (1958)
4. Vernotte, P.: Les paradoxes de la theorie continue de lequation de la chaleur. C. R. Acad. Sci. 246, 31543155 (1958)
5. Maxwell, J.C.: On the dynamic theory of gases. Philos. Trans. R. Soc. Lond. 157, 4988 (1867)
6. Baumeister, K.J., Hamill, T.D.: Hyperbolic heat conduction equationa solution for the semi-innite body problem. J. Heat
Transf. 91, 543548 (1969)
7. Baumeister, K.J., Hamill, T.D.: Hyperbolic heat conduction equationa solution for the semi-innite body problem. J. Heat
Transf. 93, 126127 (1971)
694 A. Moosaie
8. Sadd, M.H., Cha, C.Y.: Axisymmetric non-Fourier temperatures in cylindrically bounded domains. Int. J. Non-Linear
Mech. 17, 129136 (1982)
9. Taitel, V.: On the parabolic, hyperbolic and discrete formulation of the heat conduction equation. Int. J. Heat Mass
Transf. 15, 369371 (1972)
10. Carey, G.F., Tsai, M.: Hyperbolic heat transfer with reection. Numer. Heat Transf. 5, 309327 (1982)
11. zisik, M.N., Vick, B.: Propagation and reection of thermal waves in a nite medium. Int. J. Heat Mass Transf. 27, 1845
1854 (1984)
12. Frankel, J.I., Vick, B., zisik, M.N.: Flux formulation of hyperbolic heat conduction. J. Appl. Phys. 58, 33403345 (1985)
13. Glass, D.E., zisik, M.N., Vick, B.: Hyperbolic heat conduction with surface radiation. Int. J. Heat Mass Transf. 28, 1823
1830 (1985)
14. Glass, D.E., zisik, M.N., McRae, D.S., Vick, B.: Hyperbolic heat conduction with temperature-dependent thermal conduc-
tivity. J. Appl. Phys. 59, 18611865 (1986)
15. Gembarovic, J., Majernik, V.: Non-Fourier propagation of heat pulse in nite medium. Int. J. Heat Mass Transf. 31, 1073
1081 (1988)
16. Tang, D.W., Araki, N.: Non-Fourier heat conduction in a nite medium under periodic surface thermal disturbance. Int. J.
Heat Mass Transf. 39, 15851590 (1996)
17. Tang, D.W., Araki, N.: Non-Fourier heat conduction in a nite mediumunder periodic surface thermal disturbance-II. Another
form of solution. Int. J. Heat Mass Transf. 39, 33053308 (1996)
18. Mikhailov, M.D., Cotta, R.M.: Steady-periodic hyperbolic heat conduction in a nite slab. Int. Commun. Heat Mass
Transf. 24, 725731 (1997)
19. Abdel-Hamid, B.: Modelling non-Fourier heat conduction with periodic thermal oscillation using the nite integral trans-
form. Appl. Math. Model. 23, 899914 (1999)
20. Tang, D.W., Araki, N.: Analytical solution of non-Fourier temperature response in a nite medium under laser-pulse heat-
ing. Heat Mass Transf. 31, 359363 (1996)
21. Tang, D.W., Araki, N.: Non-Fourier heat condution behavior in nite mediums under pulse surface heating. Mat. Sci. Eng.
A 292, 173178 (2000)
22. Wang, L.: Solution structure of hyperbolic heat conduction equation. Int. J. Heat Mass Transf. 43, 365373 (2000)
23. Glass, D.E., zisik, M.N., Vick, B.: Non-Fourier effects on transient temperature resulting fromperiodic onoff heat ux. Int.
J. Heat Mass Transf. 30, 16231631 (1987)
24. Yen, C.-C., Wu, C.-Y.: Modelling hyperbolic heat conduction in a nite medium with periodic thermal disturbance and
surface radiation. Appl. Math. Model. 27, 397408 (2003)
25. Lewandowska, M., Malinowski, L.: An analytical solution of the hyperbolic heat conduction equation for the case of a nite
medium symmetrically heated on both sides. Int. Commun. Heat Mass Transf. 33, 6169 (2006)
26. Moosaie, A.: Non-Fourier heat conduction in a nite medium subjected to arbitrary periodic surface disturbance. Int.
Commun. Heat Mass Transf. 34, 9961002 (2007)
27. Moosaie, A.: Non-Fourier heat conduction in a nite medium subjected to arbitrary non-periodic surface disturbance. Int.
Commun. Heat Mass Transf. 35, 376383 (2008)
28. Zhang, Z., Liu, D.: Hyperbolic heat propagation in a spherical solid medium under extremely high heating rates. In: Armaly
B.F. et al. (eds.) AIAA/ASME joint thermophysics and heat transfer conference. New York, Vol 3, pp. 275283 (1998)
29. Jiang, F.: Solution and analysis of hyperbolic heat propagation in hollow spherical objects. Heat Mass Transf. 42, 1083
1091 (2006)
30. Trostel, R.: Instationre Wrmespannungen in einer Hohlkugel. Ingenieur-Archiv 24, 373391 (1956)
31. zisik, M.N.: Heat Conduction, 2nd edn. Wiley, New York (1993)
32. Watson, G.N.: A Treatise on the Theory of Bessel Functions, 2nd edn. Cambridge University Press, Cambridge (1966)
33. Rothe, R., Szab, I.: Hhere Mathematik, Vol 6. TeubnerVerlag, Stuttgart (1953)