Session 8
Multi-Dimensional, Steady-State, Linear Heat Conduction: Analytical Methods

Ernesto Gutierrez-Miravete

Spring, 2002

1  Steady State Heat Conduction

If the temperature field inside a solid body does not change with time then steady state conditions prevail. Under steady state conditions the term H/t = 0 and the heat equation becomes
Ñ·(k ÑT) + g(r) = 0
where k is the thermal conductivity, T is temperature, g is the rate of internal heat generation and r and t are, respectively, position and time. The steady state heat equation must be solved subject to appropriate boundary conditions. The boundary conditions may consist of

at each boundary in the system.

The corresponding forms of the steady state heat conduction equation in Cartesian, cylindrical and spherical coordinates are, respectively


x
(k T
x
) +
y
(k T
y
) +
z
(k T
z
) + g(x,y,z) = 0
1
r

r
(k r T
r
) + 1
r2

f
(k T
f
) +
z
(k T
z
) + g(r,f,z) = 0
and
1
r2

r
(k r2 T
r
) + 1
r2 sinq

q
(k sinq T
q
) + 1
r2 sin2 q

f
(k T
f
) + g(r,f,q) = 0

2  One-dimensional Steady State Problems

The one dimensional forms of the steady state heat equation are particularly simple and readily solvable in many cases as they reduce to second order ordinary differential equations. The corresponding forms of the steady state heat conduction equation in Cartesian, cylindrical and spherical coordinates are, respectively

d
d x
(k d T
d x
) + g(x) = 0
1
r
d
d r
(k r d T
d r
) + g(r) = 0
and
1
r2
d
d r
(k r2 d T
d r
) + g(r) = 0

2.1  Steady State Conduction in a Slab

This section contains solutions to a couple of simple but important problems. Consider first the problem of steady state heat conduction in a slab of thickness L, without internal heat generation (g = 0) and with fixed temperatures at its boundaries T(0) = T1 and T(L) = T2. Furthermore, assume the thermal conductivity is constant. Integrating the heat equation twice and introducing the given boundary temperatures yields
T(x) = T1 + (T2 - T1) x
L
The heat flux through the slab is
q = - k dT
dx
= - k T2 - T1
L
and the rate of energy transfer (heat flow) through an area A of slab is
Q = q A = - k A T2 - T1
L
= - T2 - T1
Rslab
where Rslab = L/k A is the thermal resistance of the slab.

Exercise. Derive the above equation.

As a second problem consider the same slab as before but now let heat be generated internally at a constant rate g per unit volume. Furthermore, assume the slab exchanges heat by convection with an environment at temperature T¥. In this case the solution is

T(x) = T¥ + g L2
2 k
[1 - ( x
L
)2] + g L
h

Exercise. Derive the above equation.

The concept of thermal resistance introduced above is used to determine the steady state heat flow in more complex situations. For instance, consider the composite slab made up of two slabs 1 and 2, of respective thicknesses L1 and L2. The thermal contact between the slabs is perfect. Assume heat is exchanged by convection at x = 0 with an environment at T¥1 (heat transfer coefficient h1), and at x = L = L1 + L2 with an environment at T¥2 (heat transfer coefficient h2). Furthermore, let the convective thermal resistances be defined as

Rconv,1 = 1
h1 A
and
Rconv,2 = 1
h2 A
The steady state heat flow is simply given by
Q = T¥1 - T¥2
Rconv,1 + Rslab1 + Rslab2 + Rconv,2

2.2  Steady State Conduction in a Cylinder

Consider a hollow cylinder (inner radius a, outer radius b). Let the temperatures at A and b be fixed as T(a) = T1 and T(b) = T2. Assume g = 0 and constant k. The solution to this problem is
T(r) = T1 + T2 - T1
ln(b/a)
ln( r
a
)

The heat flux through the cylindrical wall is

q = - k dT
dr
= - k T2 - T1
r ln(b/a)
and the rate of energy transfer through an area A = 2 pr l of shell is
Q = q A = - k (2 pr l) T2 - T1
r ln(b/a)
= - T2 - T1
Rcyl
where l is the axial length and Rcyl = ln(b/a)/2 pl k is the thermal resistance of the cylindrical shell.

Exercise. Derive the above equation.

The concept of thermal resistance can also be applied to the determination of the steady state heat flow in composite cylindrical shells. Consider a hollow cylinder consisting of two concentric layers. let the inner radius be r1, the radius at the joint bewtween layers r2 and the outer radius r3. The heat flux is simply given by

Q = 2 pl (T¥1 - T¥2)
1/(h1 r1) + ln(r2/r1)/ k + ln(r3/r2)/ k + 1/(h2 r3)

The above expression can be used to determine the critical thickness of cylindrical insulation.

Exercise. Derive an expression for the critical thickness of cylindrical insulation.

Consider now a solid cylinder (radius b), inside which heat is generated at constant rate g per unit volume while its boundary exchanges heat with an environment at T¥. The solution to this problem is

T(r) = T¥ + g b2
4 k
[1 - ( r
b
)2 + 2 k
h b
]

Exercise. Derive the above equation.

Now, assume a cladding is placed around the cylinder but that no heat is generated in the cladding. Let the conductivities and radii of the inner cylinder and cladding be k1, r1 and k2, r2. The following expression allows the calculation of the temperature drop between the axis of the cylinder (Tc) and that of the surrounding fluid,

Tc - T¥ = g r12
4 k1
+ g r12
2 k2
ln( r2
r1
) + g r12
2 r2 h

In practice, the contact between the cladding and the inner cylinder is not perfect and and additional term accounting for this effect must be added to the right hand side of the above equation.

As a final example consider the problem of one-dimensional steady-state conduction in a hollow cylindrical shell subject to a boundary condition of the second kind on the inner surface (r = r1) and of the first kind on the outside surface (r = r2), i.e.

1
r
[
r
( r k T
r
)] = 0
subject to
-k ( T
r
)r1 = q1
and
T(r2) = T2
This problem has an analytical solution given by
T(r) = T2 - q1 r1
k
ln( r
r2
)

Exercise. Derive the above equation.

2.3  Steady State Conduction in a Sphere

Consider a hollow spherical shell (inner radius a, outer radius b). Let the temperatures at a and b be fixed as T(a) = T1 and T(b) = T2. Assume g = 0 and constant k. The solution to this problem is
T(r) = T1 + T2 - T1
1/b - 1/a
( 1
r
- 1
a
)

The heat flux through the spherical shell is

q = - k dT
dr
= k T2 - T1
r2 (1/b - 1/a)
and the rate of energy transfer through an area A = 4 pr2 of shell is
Q = q A = k (4 pr2) T2 - T1
r2 (1/b - 1/a)
= T2 - T1
Rsph
where Rsph = (b-a)/4 pab k is the thermal resistance of the cylindrical shell.

Exercise. Derive the above equation.

Consider now a solid sphere (radius b), inside which heat is generated at constant rate g while its boundary exchanges heat with en environment at T¥. The solution to this problem is

T(r) = T¥ + g b2
4 k
[1 - ( r
b
)2 + 2 k
h b
]

3   Steady State Conduction in Fins and Spines

Fins and spines are extended surface structural elements commonly used to enhanced the efficiency of heat exchange surfaces. Fins are thin and flat while spines are pointy. Consider a fin or spine protruding from a flat wall. At the base of the fin x = 0 and the fin stretches towards x ® ¥. The fin/spine has a nonuniform cross sectional area A(x) and its perimeter is P(x). As energy travels through the fin, part of it gets dissipated by convection though the exposed extended surface while the rest continues travelling along the fin. Let the heat transfer coefficient be h and the surrounding environment temperature T¥. A steady state heat balance on a small fin/spine element of size Dx gives
Q(x) = Qconv + Q(x+Dx)
but
Q(x+Dx) - Q(x) = d Q
dx
Dx = d
dx
(- k A(x) dT
dx
) Dx
and
Qconv = h P(x) (T - T¥) Dx
Combination of the above yields
d
dx
( A(x) dT
dx
) - h
k
P(x) (T - T¥) = 0

If the fin has constant cross section A and perimeter P and T¥ = 0, the above differential equation becomes

d2 T
dx2
- h P
k A
T = 0
which has the general solution
T(x) = C1 e- m x + C2 em x = C3 sinh(m x) + C4 cosh(m x)
where m = hP/kA. If the temperature at the base of the fin/spine is Tb and also the fin/spine is very long, T(x ® ¥) » Tinfty = 0 so that
T(x) = Tb e- [hP/kA] x

4  Steady State Temperature in a Rectangular Plate

At steady state the temperature satisfies Laplace's equation
Ñ2 T = 0
From Green's theorem, Laplace's equation requires that
ó
(ç)
õ

ó
(ç)
õ

T
n
d s = 0
which states that under steady state conditions the boundary heat flux cannot be chosen arbitrarily but must average zero.

Also, from Green's theorem, if T1 and T2 are two solutions of a steady state problem whose values coincide at the boundary

ó
õ
ó
õ
ó
õ


V 
[ Ñ(T2 - T1]2 d t = 0
so that T2 - T1 = constant. The constant is zero when the problem involves only prescribed temperatures at the boundary (Dirichlet problem) and can be nonzero when normal derivatives of T at the boundary are specified (Neumann problem).

Consider steady state heat conduction in a thin rectangular plate of width 1 and height 1. The edges x = 0, x = 1 and y = 0 are maintained at T = 0 while at the edge y = 1, T(x,1) = f(x). No heat flow along the z direction perpendicular to the plate. The required temperature T(x,y) satisfies

2 T
x2
+ 2 T
y2
= 0
To find T by separation of variables we assume the a particular solution can be represented as a product of two functions each depending on a single coordinate, i.e.
Tp(x,y) = X(x) Y(y)
substituting into Laplace's equation gives
- 1
X
d2 X
dx2
= 1
Y
d2 Y
dy2
= k2
where k2 is a constant and this is true since the LHS is a function of x alone while the RHS a function of y alone. The constant is selected as k2 in order to obtain a proper Sturn-Liouville problem for X (with real eigenvalues). With the above the original PDE problem has been transformed into a system of two ODE's, i.e.
X¢¢+ k2 X = 0
subject to X(0) = X(1) = 0 and
Y¢¢- k2 Y = 0
subject to Y(0) = 0.

The solution for X(x) is

X = Xn = An sin(n px)
with eigenvalues
kn = n p
for n = 1,2,3,....

The solution of Y(y) is

Yn = Bn sinh(n py)
so that the particular solution desired is
Tn = ¥
å
n = 1 
an sin(n px) sinh(n py)

The an's are determined by making the above satisfy the nonhomogeneous condition at y = 1, i.e.

f(x) = ¥
å
n = 1 
[ an sinh(n p) ] sin(n px)
which is a Fourier sine series representation of f(x) with coefficients
cn = an sinh(n p) = 2 ó
õ
1

0 
f(x) sin(n px) dx
so that the final solution is
T(x,y) = ¥
å
n = 1 
cn sin(n px) sinh(n py)
sinh(n p)

As a particular example consider the case when f(x) = 1. In this case

cn = 2 ó
õ
1

0 
f(x) sin(n px) dx = -2 [cos(n p) - 1]
n p
= ì
ï
ï
í
ï
ï
î
4
n p
n odd
0
n even

Therefore, as long as f(x) is representable in terms of Fourier series, the obtained solution coverges to the desired solution. Note also that the presence of homogeneous conditions at x = 0, x = l made feasible the determination of the required eigenvalues.

5  Steady State Temperature in a Circular Annulus

Consider now the problem of determining the steady state temperature in a thin annular plate where the temperatures are specified at the inner and outer radii r1 and r2 as
T(r1,q) = f1(q)
T(r2,q) = f2(q)
Laplace's equation in this case is
Ñ2 T = 2 T
r2
+ 1
r
T
r
+ 1
r2
2 T
q2
= 0
Assume now a particular solution is of the form
Tp(r,q) = R(r) Q(q)
Substituting leads to the following two ODEs
r2 R¢¢+ r R¢- k2 R = 0
Q¢¢+ k2 Q = 0
where the separation constant as been chosen as k2 in order to obtain periodic trigonometric functions as the solutions for Q.

The general solution for R is

R = ì
í
î
Ak rk + Bk r-k; k ¹ 0
A0 + B0 lnr; k = 0
whereas that for Q is
Q = ì
í
î
Ck cos(k q) + Dk sin(k q); k ¹ 0
C0 + D0 q; k = 0
The periodicity requirement is satisfied by taking k = n, with n = 1,2,3,....

The particular single valued solution Tp is then

Tp = (a0 + b0 lnr) + ¥
å
n = 1 
[(an rn + bn r-n) cos(n q)
+ (cn rn + dn r-n) sin(n q) ]
Finally, the desired solution must satisfy the stated boundary conditions. Substituting f1 and f2 leads to Fourier series representations and the relationships
a0 + b0 lnr1 = 1
2 p
ó
õ
2 p

0 
f1(q) d q
a0 + b0 lnr2 = 1
2 p
ó
õ
2 p

0 
f2(q) d q
an r1n + bn r1-n = 1
p
ó
õ
2 p

0 
f1(q) cos(n q) d q
an r2n + bn r2-n = 1
p
ó
õ
2 p

0 
f2(q) cos(n q) d q
cn r1n + dn r1-n = 1
p
ó
õ
2 p

0 
f1(q) sin(n q) d q
cn r2n + dn r2-n = 1
p
ó
õ
2 p

0 
f2(q) sin(n q) d q

A special case of interest is steady heat conduction in a disk (r1 = 0) of radius r2 = a subject to a temperature f(q) at its boundary. In this case, the solution is

T(r,q) = A0 + ¥
å
n = 1 
( r
a
)n [An cos(n q) + Cn sin(n q)]
with
A0 = 1
2 p
ó
õ
2 p

0 
f(q) d q
An = 1
p
ó
õ
2 p

0 
f(q) cos(n q) d q
Cn = 1
p
ó
õ
2 p

0 
f(q) sin(n q) d q
for n = 1,2,.... Note that the temperature at the center of the disk is simply the average value of the boundary temperature.

Poisson's integral formula allows direct determination of the temperature T(r,q) from the values of it at its boundary as follows

T(r,q) = 1
2 p
ó
õ
2 p

0 
a2 - rr
a2 - 2 a r cos(q- y) + r2
T(a,y) d y

Another important special case is that of computing the temperature field around a circular hole (r2 ® ¥) of radius r1 = a subject to a temperature distribution T(a,q) = f(q) at the hole boundary. In this case

T(r,q) = A0 + ¥
å
n = 1 
( a
r
)n [Bn cos(n q) + Dn sin(n q)]
with
A0 = 1
2 p
ó
õ
2 p

0 
f(q) d q
Bn = 1
p
ó
õ
2 p

0 
f(q) cos(n q) d q
Dn = 1
p
ó
õ
2 p

0 
f(q) sin(n q) d q
for n = 1,2,.... Note that the temperature at infinity is simply the average value of the boundary temperature.

6  Steady State Temperature in a Sphere

Consider now the problem of steady state heat conduction in a sphere (radius a) incorporating symmetry in the azimuthal direction (i.e. solution independent of q). Laplace's equation in this case is

r
(r2 T
r
) + 1
siny

y
(siny T
y
) = 0
The temperature at the surface of the sphere is prescribed as
T(a,y) = f(y)
As before, a particular solution of the form Tp = R Y is assumed leading to
r2 R¢¢+ 2 r R¢- k2 R = 0
1
siny
d
d y
(siny d Y
d y
) + k2 Y = 0

The general bounded solution for R is

R = Cn rn
while the general solution for Y, finite at y = 0, p in terms of Legendre polynomials is obtained by making k2 = n(n+1) giving
Y = An Pn(cosy)

Therefore the particular solution obtained is

T(r,y) = ¥
å
n = 0 
cn ( r
a
)n Pn(cosy)

The coefficients cn are determined from the imposed boundary condition as

cn = 2n+1
2
ó
õ
p

0 
f(y) Pn(cosy) sinyd y

7  Steady State Temperature in a Brick

Consider now the problem of steady state heat conduction in a rectangular parallelepiped (length l1, width l2, height d) subject to zero temperature on all five faces except that at z = d which is maintained at T(x,y,d) = f(x,y). Laplace's equation is
2 T
x2
+ 2 T
y2
+ 2 T
z2
= 0
Assume a particular solution is of the form
Tp = X(x) Y(y) Z(z)
Two separation constants are needed in this case giving
X¢¢+ k12 X = 0
Y¢¢+ k22 Y = 0
and
Z¢¢- (k12 + k22) Z = 0
Note that homogeneous conditions occur at both boundaries in both the x and y directions leading to
Xm = Am sin(k1 x) = Am sin( m px
l1
)
Yn = Bn sin(k2 y) = Bn sin( n py
l2
)
and
Zmn = Cmn sinh( kmn z)
where kmn2 = k12 + k22. The desired particular solution is then
Tp = ¥
å
m = 1 
¥
å
n = 1 
amn sin( m px
l1
) sin( n py
l2
) sinh(kmn z)
where the coefficients amn must be determined by incorporating the nonhomogeneous condition. This leads to a double Fourier series representation and the coefficients
amn = 1
sinh(kmn d)
4
l1 l2
ó
õ
l1

0 
ó
õ
l2

0 
f(x,y) sin( m px
l1
) sin( n py
l2
) dy dx


File translated from TEX by TTH, version 2.34.
On 24 Apr 2002, 11:58.


Last Updated: Wednesday, April 24 2002 11:58
Rensselaer at Hartford, 275 Windsor St, Hartford, CT 06120
For more information: 1-800-433-4723 or info@rh.edu
Please send questions, comments or suggestions to webmaster@rh.edu