Chapter 3

Steady Heat Conduction

※ Thermal Resistance (熱阻)

Recall from Ch. 1, the rate of conductive heat transfer through a plane wall of
thickness L can be obtained by

∆T T −T
Q cond =
−kA =
kA 1 2
∆x L

The above equation can be rewritten as

T −T V1 − V2
Q cond = 1 2 (analogous to the electric current I = )
Rwall Re

where

L
Rwall =
kA

is the thermal resistance of the wall against heat conduction or simply the conduction
resistance of the wall. Note that the thermal resistance of a medium depends on the
geometry and the thermal properties of the medium.

Similarly for convection heat transfer, the rate of convective heat transfer from a solid
surface of area A s can be obtained by

Q conv hAs (Ts − T∞ )

=

which can again be rewritten as

T −T 1
Q conv = s ∞ and Rconv =
Rconv hAs

R conv is the thermal resistance of the surface against heat convection, or simply the
convection resistance of the surface.

Again for radiative heat transfer, the rate of radiative heat transfer from a solid surface
of area A s can be obtained by

1
 Q rad εs As (Ts4 − Tsurr
= 4
)
which can again be rewritten as

T −T 1
Q rad = s surr and Rrad = , hrad =εs (Ts + Tsurr ) (Ts2 + Tsurr
2
)
Rrad hrad As

R rad is the thermal resistance of the surface against heat radiation, or simply the
radiation resistance of the surface.

When T surr ≈ T ∞ , the radiation effect can properly be accounted by combined heat
transfer coefficient

= hconv + hrad
hcombined

※ Thermal Resistance Network

Now consider steady one-dimensional heat flow through a plane wall

of thickness L, area A, and thermal conductivity k that is exposed to
T1
T2 convection on both sides to fluids at temperatures T ∞1 and T ∞2 (T ∞1
A A > T ∞2 ) with heat transfer coefficients h 1 and h 2 . Under steady
L conditions we can obtain rate of heat convection into the wall, rate of
Q heat conduction through the wall, and rate of heat convection from
the wall are the same:

T − T2
Q h1 A (T∞1 − T=
= 1) kA 1 = h2 A (T2 − T∞ 2 )
L

which can be rearranged as

T∞1 − T1 T1 − T2 T2 − T∞ 2
=Q = =
1/ h1 A L / kA 1/ h2 A

T∞1 − T1 T1 − T2 T2 − T∞ 2
or=Q = =
Rconv1 Rwall Rconv2

Adding the numerators and denominators yields

T −T
Q = ∞1 ∞ 2
Rtotal

where

2
 Rtotal = Rconv1 + Rwall + Rconv2

This indicates the rate of heat transfer through a wall separating two mediums is equal
to the temperature difference divided by the total thermal resistance between the
mediums (like electrical resistances connected in series).

※ Multilayer Plane Walls

T2 From the conclusion above which states the rate of heat transfer
T1 through a wall separating two mediums,
T3
A T −T
Q = ∞1 ∞ 2
L1 L2 Rtotal
Q Here the total thermal resistance Rtotal can be calculated by

1 L L 1
Rtotal = + 1 + 2 +
h1 A k1 A k2 A h2 A

This result for the two-layer case is analogous to the single-layer case, except that an
additional resistance is added for the additional layer.

※ Thermal Contact Resistance (接觸熱阻)

In the analysis of heat conduction through multilayer solids, we assumed “perfect

contact” at the interface of two layers, and thus no temperature drop at the interface.
However, a surface is microscopically rough no matter how smooth it appears to be.
That is, an interface will contain numerous air gaps of varying sizes that act as
insulation because of the low thermal conductivity of air. Thus, an interface offers
some resistance to heat transfer, and this resistance per unit interface area is called the
thermal contact resistance, R c . By applying Newton’s law of cooling

Q hc A∆Tinterface
=

which can be rewritten as

∆T Rc 1
Q = interface and Rinterface
= =
Rinterface A hc A

h c is the thermal contact conductance. Note that R c represents thermal contact

resistance per unit area. The thermal resistance for the entire interface is obtained by
dividing R c by the apparent interface area A.

3
 Note that, obviously the value of thermal contact resistance depends on the surface
roughness and the material properties as well as the temperature and pressure at the
interface and the type of fluid trapped at the interface (see Table 3-2).

※ Generalized Thermal Resistance Networks

The thermal resistance concept or the electrical analogy can also be used to solve
steady heat transfer problems that involve parallel layers or combined series-parallel
arrangements. Although such problems are often two- or even three-dimensional,
approximate solutions can be obtained by assuming one dimensional heat transfer and
using the thermal resistance network.

Consider the composite wall which consists of two parallel layers.

Q1
The total rate of heat transfer is the sum of the rate of heat
A1
transfers through each layer
T1 T2
L  1
T −T T −T 1 
A2 Q = Q1 + Q 2 = 1 2 + 1 2 = (T1 − T2 )  + 
Q 2 R1 R2  R1 R2 

Utilizing electrical analogy for resistors in parallel,

T −T 1 1 1 R1 R2
Q = 1 2 where = + or Rtotal =
Rtotal Rtotal R1 R2 R1 + R2

Similarly for a more complicated configuration with three layers, the

A3 total rate of heat transfer through this system can again be expressed
A1
as
T1 T∞
Q T −T
Q = 1 ∞
A2 Rtotal

where

R1 R2
Rtotal = R12 + R3 + Rconv = + R3 + Rconv
R1 + R2

L1 L2 L3 1
=
and R1 = , R2 = , R3 = , Rconv
k1 A1 k2 A2 k3 A3 hA3

Two assumptions commonly used in solving complex multidimensional heat transfer

problems by treating them as one-dimensional (x-direction) using the thermal
resistance network are (1) any plane wall normal to the x-axis is isothermal (i.e., to
assume the temperature to vary in the x-direction only) and (2) any plane parallel to

4
 the x-axis is adiabatic (i.e., to assume heat transfer to occur in the x-direction only).
These two assumptions result in different (but usually close) resistance networks, and
the actual result lies between these two cases.

※ Heat Conduction in Cylinders and Spheres

Consider a long cylindrical layer (such as a circular pipe) of inner radius r 1 , outer
radius r 2 , length L, and average thermal conductivity k. The two surfaces of the
cylindrical layer are maintained at constant temperatures T 1 and T 2 and there is no
heat generation in the layer. Through 1-D Fourier’s law

dT
Q cond, cyl = −kA
dr

where A = 2πrL is the heat transfer area at location r. integrating from r = r 1 , where
T(r 1 ) = T 1 , to r = r 2 , where T(r 2 ) = T 2 gives

T −T T1 − T2
Q cond, cyl 2=
= π Lk 1 2
ln ( r2 / r1 ) Rcyl

where

ln ( r2 / r1 )
Rcyl =
2π Lk

is the thermal resistance of the cylindrical layer against heat conduction, or simply the
conduction resistance of the cylinder layer.

Similarly for a spherical layer with A = 4πr2

T − T T1 − T2
Q cond, sph 4=
= p r1r2 k 1 2
r2 − r1 Rsph

where

r2 − r1
Rsph =
4p r1r2 k

is the thermal resistance of the spherical layer against heat conduction, or simply the
conduction resistance of the spherical layer.

Note that same electrical analogy for resistors in series can also be applied for
multilayer cylinders and spheres such as the case for a double layered cylinder

Rtotal = Rconv1 + Rcyl1 + Rcyl2 + Rconv2

5
※ Critical Radius of Insulation

Insulations are usually applied to prevent heat loss. However, the additional insulation
increases the conduction resistance of the insulation layer but decreases the
convection resistance of the surface because of the increase in the outer surface area
for convection. The heat transfer from the pipe may increase or decrease, depending
on which effect dominates.

Consider a cylindrical pipe of outer radius r 1 whose outer surface temperature T 1 is

maintained constant. The pipe is now insulated with a material whose thermal
conductivity is k and outer radius is r 2 . Heat is lost from the pipe to the surrounding
medium at temperature T ∞ , with a convection heat transfer coefficient h. The rate of
heat transfer from the insulated pipe to the surrounding air can be expressed as

T1 − T∞ T1 − T∞
=Q =
Rins + Rconv ln ( r2 / r1 ) 1
+
2π Lk h ( 2π r2 L )

The value of r 2 at which Q reaches a maximum is determined from the requirement

that dQ/dr 2 = 0. Solving for r 2 yields the critical radius of insulation for a cylindrical
body to be

k 2k
rcr,cylinder = (similarly, rcr,sphere = )
h h

The rate of heat transfer from the cylinder increases with the addition of insulation for
r 2 < r cr , reaches a maximum when r 2 = r cr , and starts to decrease for r 2 > r cr .

※ Heat Transfer from Finned Surfaces

When the temperatures T s and T ∞ are fixed by design considerations, as is often the
case, there are two ways to increase the rate of heat transfer: to increase the
convection heat transfer coefficient h (extra fan/pump) or to increase the surface area
A s (attaching fins made of highly conductive materials such as aluminum).

In the analysis of fins, we consider steady operation with no heat generation in the fin,
and we assume the thermal conductivity k of the material to remain constant. We also
assume the convection heat transfer coefficient h to be constant and uniform over the
entire surface of the fin for convenience in the analysis (in reality the value of h is
usually much lower at the fin base than it is at the fin tip). Consider a volume element
of a fin at location x having a length of ∆x, cross-sectional area of A c , and a perimeter
of p. Under steady conditions, the energy balance of the volume element is

6
  Rate of heat   Rate of heat   Rate of heat 
     
=
 conduction into   conduction from the  +  convection from 
 the element at x   element at x + ∆x   the element 
     

Q cond, x Q cond, x +∆x + Q conv

or = where Q conv =hp∆x (T − T∞ )

Q cond, x +∆x − Q cond, x

⇒ + hp (T − T∞ ) =
0
∆x

By taking ∆x → 0 ,

dQ cond
+ hp (T − T∞ ) =
0
dx

With Fourier’s Law of heat conduction,

d  dT 
 −kAc  + hp (T − T∞ ) =
0
dx  dx 

As constant conductivity is applied,

d 2T hp d 2θ
− (T − T∞ ) =
0 or − m 2θ =
0
dx 2 kAc dx 2

where θ= T − T∞ is the temperature excess and m 2 = hp / kAc . This is in fact a linear,

homogeneous, second-order differential equation with constant coefficients which the

general solution can be expressed as

( x ) c1emx + c2e− mx
θ=

where the unknown constants can be obtained by two boundary conditions.

※ Infinitely Long Fin (T fin tip = T ∞ )

For a sufficiently long fin of uniform cross section A c , the temperature of the fin at the
fin tip will approach the environment temperature T ∞ and thus θ will approach zero

θ ( L)= T ( L) − T∞= 0 as L→∞

By plugging in the boundary condition, the general solution becomes

7
 lim θ ( L ) = lim ( c1e mL + c2 e − mL )= lim ( c1e mL )= 0
L →∞ L →∞ L →∞

which gives c1 = 0 and leave the solution as θ ( x ) = c2 e − mx . Assume another

boundary condition is given by the base temperature θ ( 0=

) θ=b Tb − T∞ (which is
almost always the case in reality), the solution yields

T ( x ) − T∞ − mx −x
= e=
hp / kAc
e
Tb − T∞

dT
and Q long fin =
−kAc hpkAc (Tb − T∞ ) by Fourier’s law or the heat transfer
=
dx x =0
from the exposed surfaces of the fin is equal to the heat transfer to the fin at the base
under steady condition

Q fin= ∫ h T ( x ) − T∞  dAfin= ∫ hθ ( x ) dAfin

Afin Afin

※ Negligible Heat Loss from the Fin Tip (Adiabatic fin tip, Q fin tip = 0 )

Fins are not likely to be so long that their temperature approaches the surrounding
temperature at the tip. A more realistic situation is for heat transfer from the fin tip to
be negligible since the surface area of the fin tip is usually a negligible fraction of the
total fin area. Thus,

dθ dT
=0 (from Q fin =
−kAc =
0)
dx x=L dx x = L

Assume another boundary condition is again given by the base temperature

θ ( 0=
) θ=b Tb − T∞ ,

θ ( 0 ) = c1 + c2 =θ b  θb
 c1 =
  1 + e 2 mL
 dθ ⇒ 
 dx =mc1e mx − mc2 e − mx =0 c = θ b
 x=L  2 1 + e −2 mL


And thus

8
 T ( x ) − T∞ cosh m( L − x)
=
Tb − T∞ cosh mL

dT
and Q adiabatic tip =
−kAc hpkAc (Tb − T∞ ) tanh mL
=
dx x =0

※ Specified Temperature ( Tfin tip = TL )

In this case the temperature at the fin tip is fixed at a specified temperature, T L and the
boundary condition can be written as

θ ( L=
) θ=
L TL − T∞

Along with the base temperature boundary condition θ ( 0=

) θ=b Tb − T∞ and general

( x ) c1emx + c2e− mx ,
solution θ=

T ( x ) − T∞ (TL − T∞ ) / (Tb − T∞ )  sinh mx + sinh m( L − x)

=
Tb − T∞ sinh mL

dT cosh mL − (TL − T∞ ) / (Tb − T∞ ) 

and Q specified temp =
− kAc hpkAc (Tb − T∞ )
=
dx x =0 sinh mL

※ Convection from Fin Tip

The proper boundary condition for the fin tip is convection that also includes the

effects of radiation. From the energy balance at the fin tip Q cond = Q conv ,

dT
− kAc = hAc T ( L ) − T∞ 
dx x=L

Along with the base temperature boundary condition θ ( 0=

) θ=b Tb − T∞ and general

( x ) c1emx + c2e− mx ,
solution θ=

9
 T ( x ) − T∞ cosh m( L − x) + ( h / mk ) sinh m( L − x)
=
Tb − T∞ cosh mL + ( h / mk ) sinh mL

dT sinh mL + ( h / mk ) cosh mL
and Q convection =
−kAc hpkAc (Tb − T∞ )
=
dx x =0 cosh mL + ( h / mk ) sinh mL

A practical way of accounting for the heat loss from the fin tip is to replace the fin
length L in the relation for the insulated tip case by a corrected length defined as

Ac t D
Lc= L + ⇒ Lc , rectangular fin= L + & Lc , cylindrical fin= L +
p 2 4

where A c is the cross-sectional area, p is the perimeter of the fin at the tip, t is the
thickness of the rectangular fins, and D is the diameter of the cylindrical fins.
Multiplying the relation above by the perimeter gives A corrected = A fin (lateral) + A tip ,
which indicates that the fin area determined using the corrected length is equivalent to
the sum of the lateral fin area plus the fin tip area. The corrected length approximation
gives very good results when the variation of temperature near the fin tip is small
(which is the case when mL ≥ 1) and the heat transfer coefficient at the fin tip is about
the same as that at the lateral surface of the fin.

※ Fin Efficiency

In the limiting case of zero thermal resistance or infinite thermal conductivity (k →

∞), the temperature of the fin will be uniform at the base value of T b . The heat
transfer from the fin will be maximum in this case and can be expressed as

Q fin, max hAfin (Tb − T∞ )

=

where A fin is the total surface area of the fin. In reality, however, the temperature of
the fin will drop along the fin, and thus the heat transfer from the fin will be less
because of the decreasing temperature difference T(x) − T ∞ toward the fin tip. To
account for the effect of this decrease in temperature on heat transfer, we define a fin
efficiency as

Q fin Actual heat transfer rate from the fin

hfin
= =

Qfin, max Ideal heat transfer rate from the fin (fin were at base temperature)

10
or Q fin h=
= 
fin Qfin, max hfin hAfin (Tb − T∞ )

For the cases of constant cross section of very long fins and fins with insulated tips,
the fin efficiency can be expressed as

Q fin hpkAc (Tb − T∞ ) 1 kAc 1

h
= = = =
Q fin, max hAfin (Tb − T∞ )
long fin
L hp mL

Q fin hpkAc (Tb − T∞ ) tanh mL tanh mL

and hadiabatic
= = =
Q fin, max hAfin (Tb − T∞ )
fin
mL

※ Fin Effectiveness

The performance of the fins is judged on the basis of the enhancement in heat transfer
relative to the no-fin case. The performance of fins expressed in terms of the fin
effectiveness ε fin is defined as

Q fin Q fin Heat transfer rate from the fin of base area Ab
e fin
= = =
Q no fin hAb (Tb − T∞ ) Heat transfer rate from the surface of area Ab

Note that both the fin efficiency and fin effectiveness are related to the performance
of the fin, but they are different quantities. However, they are related to each other by

Q fin hfin hAfin (Tb − T∞ ) Afin

ε fin
= = = hfin
Q no fin hAb (Tb − T∞ ) Ab

The rate of heat transfer from a sufficiently long fin of uniform cross section under

=
steady conditions is given by Q long fin hpkAc (Tb − T∞ ) and the effectiveness of

such a long fin is determined to be

Q long fin hpkAc (Tb − T∞ ) kp

ε= = =
Q hAb (Tb − T∞ )
long fin
no fin hAc

With the rate of heat transfer for a surface containing n fins, we can also define an
overall effectiveness for a finned surface as the ratio of the total heat transfer from the
finned surface to the heat transfer from the same surface if there were no fins,

11
 Q total, fin h ( Aunfin + hfin Afin )(Tb − T∞ ) Aunfin + hfin Afin
e fin, overall
= = =
Qtotal, no fin hA (T − T )
no fin b ∞ A no fin

Conclusions:

• The thermal conductivity k of the fin material should be as high as possible.

• The ratio of the perimeter to the cross-sectional area of the fin p/A c should be as
high as possible.

• The use of fins is most effective in applications involving a low convection heat
transfer coefficient. Thus, the use of fins is more easily justified when the medium
is a gas instead of a liquid and the heat transfer is by natural convection instead of
by forced convection.

※ Proper Length of a Fin

Normally, one may be tempted to think that the longer the fin, the larger the surface
area and thus the higher the rate of heat transfer. Therefore, for maximum heat transfer,
the fin should be infinitely long. However, the temperature drops along the fin
exponentially and reaches the environment temperature at some length. The part of
the fin beyond this length does not contribute to heat transfer since it is at the
temperature of the environment.

We compare heat transfer from a fin of finite length to heat transfer from an infinitely
long fin under the same conditions. The ratio of these two heat transfers is

Q fin hpkAc (Tb − T∞ ) tanh mL

= = tanh mL
Q long fin hpkAc (Tb − T∞ )

The heat transfer from a fin increases with mL almost linearly at first, but the curve
reaches a plateau later and reaches a value for the infinitely long fin at about mL = 5.
Therefore, a fin whose length is L = 5/m can be considered to be an infinitely long fin.
Furthermore, reducing the fin length by half in that case (from mL = 5 to mL = 2.5)
causes a drop of just 1 percent in heat transfer which is of course acceptable. In
practice, a fin length that corresponds to about mL = 1 will transfer 76.2 percent of the
heat that can be transferred by an infinitely long fin, and thus it should offer a good
compromise between heat transfer performance and the fin size.

※ 1-D Heat Transfer Assumption for Fins

A common approximation used in the analysis of fins is to assume the fin temperature
12
varies in one direction only (along the fin length) and the temperature variation along
other directions is negligible (1-D assumption). Studies have shown that the error
involved in one-dimensional fin analysis is negligible (less than about 1 percent)
when

hδ
< 0.2
k

where δ is the characteristic thickness of the fin, which is taken to be the plate
thickness t for rectangular fins and the diameter D for cylindrical ones.

※ Heat Transfer in Common Configurations

Heat transfer problems involving two surfaces maintained at constant temperatures T 1

and T 2 are commonly seen. The steady rate of heat transfer between these two
surfaces is expressed as

=Q Sk (T1 − T2 )

where S is the conduction shape factor, which has the dimension of length, and k is
the thermal conductivity of the medium between the surfaces (see Table 3-7 for more
information). Note that conduction shape factor S is related to the thermal resistance R
by R = 1/kS or S = 1/kR.

Under steady conditions, the rate of heat transfer through any section of a building
wall or roof can be determined from

A (Ti − To )
Q= UA (Ti − To =
)
R

where T i and T o are the indoor and outdoor air temperatures, A is the heat transfer
area, U is the overall heat transfer coefficient (the U-factor), and R = 1/U is the overall
unit thermal resistance (the R-value) (see Table 3-8 and 3-11 for more information).

13