136 Chapter 3 j One-Dimensional, Steady-State Conduction

3.3 Radial Systems

Cylindrical and spherical systems often experience temperature gradients in the radial direc-
tion only and may therefore be treated as one-dimensional. Moreover, under steady-state
conditions with no heat generation, such systems may be analyzed by using the standard
method, which begins with the appropriate form of the heat equation, or the alternative
method, which begins with the appropriate form of Fourier’s law. In this section, the cylin-
drical system is analyzed by means of the standard method and the spherical system by
means of the alternative method.

3.3.1 The Cylinder

A common example is the hollow cylinder whose inner and outer surfaces are exposed to
fluids at different temperatures (Figure 3.7). For steady-state conditions with no heat gener-
ation, the appropriate form of the heat equation, Equation 2.26, is

r dr 1
1 d kr dT 5 0
dr 2 (3.28)

where, for the moment, k is treated as a variable. The physical significance of this result
becomes evident if we also consider the appropriate form of Fourier’s law. The rate at which
energy is conducted across any cylindrical surface in the solid may be expressed as

qr 5 2kA dT 5 2k(2prL) dT (3.29)

dr dr
where A 5 2prL is the area normal to the direction of heat transfer. Since Equation 3.28
dictates that the quantity kr(dT/dr) is independent of r, it follows from Equation 3.29
that the conduction heat transfer rate qr (not the heat flux q0r ) is a constant in the radial
direction.

Hot fluid
T∞,1, h1
Cold fluid Ts,1
T∞,2, h2
r

Ts,2

Ts,1 r1 r2
r

r1
L

r2
T∞,1 Ts,1 Ts,2 T∞,2
qr
Ts,2 ________
1 In(r2/r1)
________ ________
1
h12 π r1L 2 π kL h22 π r2L

FIGURE 3.7 Hollow cylinder with convective surface conditions.

3.3 j Radial Systems 137

We may determine the temperature distribution in the cylinder by solving Equation

3.28 and applying appropriate boundary conditions. Assuming the value of k to be constant,
Equation 3.28 may be integrated twice to obtain the general solution
T(r) 5 C1 ln r 1 C2 (3.30)

To obtain the constants of integration C1 and C2, we introduce the following boundary
conditions:
T(r1) 5 Ts,1 and T(r2) 5 Ts,2
Applying these conditions to the general solution, we then obtain

Ts,1 5 C1 ln r1 1 C2 and Ts,2 5 C1 ln r2 1 C2

Solving for C1 and C2 and substituting into the general solution, we then obtain

12
Ts,1 2 Ts,2
T(r) 5 ln rr 1 Ts,2 (3.31)
ln (r1 /r2) 2

Note that the temperature distribution associated with radial conduction through a cylindri-
cal wall is logarithmic, not linear, as it is for the plane wall under the same conditions. The
logarithmic distribution is sketched in the inset of Figure 3.7.
If the temperature distribution, Equation 3.31, is now used with Fourier’s law, Equation
3.29, we obtain the following expression for the heat transfer rate:

2pLk(Ts,1 2 Ts,2)
qr 5 (3.32)
ln (r2 /r1)

From this result it is evident that, for radial conduction in a cylindrical wall, the thermal
resistance is of the form

ln (r2 /r1)
Rt,cond 5 (3.33)
2pLk

This resistance is shown in the series circuit of Figure 3.7. Note that since the value of qr is
independent of r, the foregoing result could have been obtained by using the alternative
method, that is, by integrating Equation 3.29.
Consider now the composite system of Figure 3.8. Recalling how we treated the com-
posite plane wall and neglecting the interfacial contact resistances, the heat transfer rate
may be expressed as

Ty,1 2 Ty,4
qr 5 (3.34)
1 ln (r2 /r1) ln (r3 /r2) ln (r4 /r3) 1
1 1 1 1
2pr1Lh1 2pkAL 2pkBL 2pkCL 2pr4Lh4
The foregoing result may also be expressed in terms of an overall heat transfer coefficient.
That is,
Ty,1 2 Ty,4
qr 5 5 UA(Ty,1 2 Ty,4) (3.35)
Rtot
138 Chapter 3 j One-Dimensional, Steady-State Conduction

Ts,4 T∞,4, h4
T∞,1, h1
T3
T2

Ts,1

r1 r2
r3
r4 L
T∞,1, h1 A B C

T∞,4, h4
T∞,1
Ts,1

T2
T3

Ts,4
T∞,4

T∞,1 Ts,1 T2 T3 Ts,4 T∞,4

qr

1
__________ In(r2/r1)
_________ In(r3/r2)
_________ In(r4/r3)
_________ 1
__________
h12 π r1L 2π kAL 2π kBL 2π kCL h42 π r4L

FIGURE 3.8 Temperature distribution for a composite cylindrical wall.

If U is defined in terms of the inside area, A1 5 2pr1L, Equations 3.34 and 3.35 may be
equated to yield

U1 5 1 (3.36)
1 1 r1 ln r2 1 r1 ln r3 1 r1 ln r4 1 r1 1
h1 kA r1 kB r2 kC r3 r4 h4
This definition is arbitrary, and the overall coefficient may also be defined in terms of A4 or
any of the intermediate areas. Note that

U1A1 5 U2A2 5 U3A3 5 U4A4 5 (SRt)21 (3.37)

and the specific forms of U2, U3, and U4 may be inferred from Equations 3.34 and 3.35.

EXAMPLE 3.6

The possible existence of an optimum insulation thickness for radial systems is suggested
by the presence of competing effects associated with an increase in this thickness. In partic-
ular, although the conduction resistance increases with the addition of insulation, the con-
vection resistance decreases due to increasing outer surface area. Hence there may exist an
insulation thickness that minimizes heat loss by maximizing the total resistance to heat
transfer. Resolve this issue by considering the following system.
3.3 j Radial Systems 139

1. A thin-walled copper tube of radius ri is used to transport a low-temperature refrigerant

and is at a temperature Ti that is less than that of the ambient air at Ty around the tube.
Is there an optimum thickness associated with application of insulation to the tube?
2. Confirm the above result by computing the total thermal resistance per unit length of
tube for a 10-mm-diameter tube having the following insulation thicknesses: 0, 2, 5,
10, 20, and 40 mm. The insulation is composed of cellular glass, and the outer surface
convection coefficient is 5 W/m2 z K.

SOLUTION

Known: Radius ri and temperature Ti of a thin-walled copper tube to be insulated from

the ambient air.

Find:
1. Whether there exists an optimum insulation thickness that minimizes the heat transfer
rate.
2. Thermal resistance associated with using cellular glass insulation of varying thickness.

Schematic:
T∞
h = 5 W/m2•K
r
ri
Ti Air

Insulation, k

Assumptions:
1. Steady-state conditions.
2. One-dimensional heat transfer in the radial (cylindrical) direction.
3. Negligible tube wall thermal resistance.
4. Constant properties for insulation.
5. Negligible radiation exchange between insulation outer surface and surroundings.

Properties: Table A.3, cellular glass (285 K, assumed): k 5 0.055 W/m z K.

Analysis:
1. The resistance to heat transfer between the refrigerant and the air is dominated by con-
duction in the insulation and convection in the air. The thermal circuit is therefore
Ti T∞
q'
In(r/ri)
________ 1
_______
2π k 2 π rh

where the conduction and convection resistances per unit length follow from Equations
3.33 and 3.9, respectively. The total thermal resistance per unit length of tube is then
ln (r/ri )
R9tot 5 1 1
2pk 2prh
140 Chapter 3 j One-Dimensional, Steady-State Conduction

where the rate of heat transfer per unit length of tube is

Ty 2 Ti
q9 5
R9tot
An optimum insulation thickness would be associated with the value of r that minimized
q9 or maximized R9tot. Such a value could be obtained from the requirement that
dR9tot
50
dr
Hence

1 2 1 50
2pkr 2pr 2h
or

r5k
h
To determine whether the foregoing result maximizes or minimizes the total resis-
tance, the second derivative must be evaluated. Hence
d 2R9tot
5 2 1 2 1 13
dr 2 2pkr pr h
or, at r 5 k/h,

d 2R9tot
dr 2 1
5 1 2 12 1 5
p(k /h) k 2k
1
2
2pk 3/h2
.0

Since this result is always positive, it follows that r 5 k/h is the insulation radius for
which the total resistance is a minimum, not a maximum. Hence an optimum insulation
thickness does not exist.
From the above result it makes more sense to think in terms of a critical insulation
radius
rcr ; k
h
which maximizes heat transfer, that is, below which q9 increases with increasing r and
above which q9 decreases with increasing r.
2. With h 5 5 W/m2 z K and k 5 0.055 W/m z K, the critical radius is

rcr 5 0.055 W/m z K 5 0.011 m

5 W/m2 z K
Hence rcr . ri and heat transfer will increase with the addition of insulation up to a
thickness of

rcr 2 ri 5 (0.011 2 0.005) m 5 0.006 m

3.3 j Radial Systems 141

The thermal resistances corresponding to the prescribed insulation thicknesses may be

calculated and are plotted as follows:
8

R'tot
6

R't (m•K/W)
R'cond
4

R'conv
2

0
0 6 10 20 30 40 50
r – ri (mm)
Comments:
1. The effect of the critical radius is revealed by the fact that, even for 20 mm of insula-
tion, the total resistance is not as large as the value for no insulation.
2. If ri , rcr, as it is in this case, the total resistance decreases and the heat rate therefore
increases with the addition of insulation. This trend continues until the outer radius of
the insulation corresponds to the critical radius. The trend is desirable for electrical
current flow through a wire, since the addition of electrical insulation would aid in
transferring heat dissipated in the wire to the surroundings. Conversely, if ri . rcr, any
addition of insulation would increase the total resistance and therefore decrease the
heat loss. This behavior would be desirable for steam flow through a pipe, where insu-
lation is added to reduce heat loss to the surroundings.
3. For radial systems, the problem of reducing the total resistance through the application of
insulation exists only for small diameter wires or tubes and for small convection coeffi-
cients, such that rcr . ri. For a typical insulation (k < 0.03 W/m z K) and free convection
in air (h < 10 W/m2 z K), rcr 5 (k/h) < 0.003 m. Such a small value tells us that, normally,
ri . rcr and we need not be concerned with the effects of a critical radius.
4. The existence of a critical radius requires that the heat transfer area change in the direction
of transfer, as for radial conduction in a cylinder (or a sphere). In a plane wall the area per-
pendicular to the direction of heat flow is constant and there is no critical insulation thick-
ness (the total resistance always increases with increasing insulation thickness).

3.3.2 The Sphere

Now consider applying the alternative method to analyzing conduction in the hollow sphere
of Figure 3.9. For the differential control volume of the figure, energy conservation requires
that qr 5 qr1dr for steady-state, one-dimensional conditions with no heat generation. The
appropriate form of Fourier’s law is

qr 5 2kA dT 5 2k(4pr 2) dT (3.38)

dr dr
2
where A 5 4pr is the area normal to the direction of heat transfer.