274

HEAT TRANSFER

g· 2 A∆ x which is identical to the difference equation (Eq. 5–11) obtained earlier.

Again, this equation is applicable to each of the M 2 1 interior nodes, and its
T1 – T2 T2 – T3 application gives M 2 1 equations for the determination of temperatures at
k A ——— k A ———
∆x ∆x M 1 1 nodes. The two additional equations needed to solve for the M 1 1 un-
known nodal temperatures are obtained by applying the energy balance on the
1 2 3 two elements at the boundaries (unless, of course, the boundary temperatures
Volume are specified).
element
of node 2
You are probably thinking that if heat is conducted into the element from
both sides, as assumed in the formulation, the temperature of the medium will
have to rise and thus heat conduction cannot be steady. Perhaps a more realis-
T1 – T2 T2 – T3 · tic approach would be to assume the heat conduction to be into the element on
k A ——— – k A ——— + g2 A∆ x = 0
∆x ∆x the left side and out of the element on the right side. If you repeat the formu-
or lation using this assumption, you will again obtain the same result since the
T1 – 2T2 + T3 + g· 2 A∆ x 2 / k = 0 heat conduction term on the right side in this case will involve Tm 2 Tm 1 1 in-
stead of Tm 1 1 2 Tm, which is subtracted instead of being added. Therefore,
(a) Assuming heat transfer to be out of the the assumed direction of heat conduction at the surfaces of the volume ele-
volume element at the right surface.
ments has no effect on the formulation, as shown in Figure 5–12. (Besides, the
actual direction of heat transfer is usually not known.) However, it is conve-
g· 2 A∆ x nient to assume heat conduction to be into the element at all surfaces and not
worry about the sign of the conduction terms. Then all temperature differences
T1 – T2 T3 – T2 in conduction relations are expressed as the temperature of the neighboring
k A ——— k A ———
∆x ∆x node minus the temperature of the node under consideration, and all conduc-
tion terms are added.
1 2 3
Volume
element Boundary Conditions
of node 2 Above we have developed a general relation for obtaining the finite difference
equation for each interior node of a plane wall. This relation is not applicable
T1 – T2 T3 – T2 ·
to the nodes on the boundaries, however, since it requires the presence of
k A ———
∆x
+ k A ———
∆x
+ g2 A∆ x = 0 nodes on both sides of the node under consideration, and a boundary node
or
does not have a neighboring node on at least one side. Therefore, we need to
T1 – 2T2 + T3 + g· 2 A∆ x 2 / k = 0
obtain the finite difference equations of boundary nodes separately. This is
best done by applying an energy balance on the volume elements of boundary
(b) Assuming heat transfer to be into the
nodes.
volume element at all surfaces. Boundary conditions most commonly encountered in practice are the spec-
FIGURE 5–12 ified temperature, specified heat flux, convection, and radiation boundary
The assumed direction of heat transfer conditions, and here we develop the finite difference formulations for them
at surfaces of a volume element has for the case of steady one-dimensional heat conduction in a plane wall of
no effect on the finite difference thickness L as an example. The node number at the left surface at x 5 0 is 0,
formulation. and at the right surface at x 5 L it is M. Note that the width of the volume el-
ement for either boundary node is Dx/2.
The specified temperature boundary condition is the simplest boundary
condition to deal with. For one-dimensional heat transfer through a plane wall
of thickness L, the specified temperature boundary conditions on both the left
and right surfaces can be expressed as (Fig. 5–13)
T(0) 5 T0 5 Specified value
T(L) 5 TM 5 Specified value (5-19)

where T0 and Tm are the specified temperatures at surfaces at x 5 0 and x 5 L,

respectively. Therefore, the specified temperature boundary conditions are
 275
CHAPTER 5

incorporated by simply assigning the given surface temperatures to the bound-

ary nodes. We do not need to write an energy balance in this case unless we Plane wall
decide to determine the rate of heat transfer into or out of the medium after the
temperatures at the interior nodes are determined. 35°C 82°C
When other boundary conditions such as the specified heat flux, convection,
radiation, or combined convection and radiation conditions are specified at a
boundary, the finite difference equation for the node at that boundary is ob-
tained by writing an energy balance on the volume element at that boundary.
L
The energy balance is again expressed as 0
…
0 1 2 M
·
o
·
Q 1 Gelement 5 0 (5-20)
all sides
T0 = 35°C
TM = 82°C
for heat transfer under steady conditions. Again we assume all heat transfer to
FIGURE 5–13
be into the volume element from all surfaces for convenience in formulation,
Finite difference formulation of
except for specified heat flux since its direction is already specified. Specified
specified temperature boundary
heat flux is taken to be a positive quantity if into the medium and a negative
conditions on both surfaces
quantity if out of the medium. Then the finite difference formulation at the of a plane wall.
node m 5 0 (at the left boundary where x 5 0) of a plane wall of thickness L
during steady one-dimensional heat conduction can be expressed as (Fig.
5–14) ∆x
—–
2
· T1 2 T0 Volume element
Q left surface 1 kA 1 g·0(ADx/2) 5 0 (5-21) of node 0
Dx g· 0

T1 – T0
where ADx/2 is the volume of the volume element (note that the boundary ele- ·
Qleft surface k A ———
∆x
ment has half thickness), g·0 is the rate of heat generation per unit volume (in
W/m3) at x 5 0, and A is the heat transfer area, which is constant for a plane L
wall. Note that we have Dx in the denominator of the second term instead of 0
0 1 2 … x
Dx/2. This is because the ratio in that term involves the temperature difference ∆x ∆x
between nodes 0 and 1, and thus we must use the distance between those two
nodes, which is Dx. · T1 – T0 · ∆ x
The finite difference form of various boundary conditions can be obtained Qleft surface + k A ——— + g0 A —– = 0
∆x 2
·
from Eq. 5–21 by replacing Q left surface by a suitable expression. Next this is FIGURE 5–14
done for various boundary conditions at the left boundary. Schematic for the finite difference
formulation of the left boundary
1. Specified Heat Flux Boundary Condition node of a plane wall.
T1 2 T0
q·0A 1 kA 1 g·0(ADx/2) 5 0 (5-22)
Dx

Special case: Insulated Boundary (q·0 5 0)

T1 2 T0
kA 1 g·0(ADx/2) 5 0 (5-23)
Dx

2. Convection Boundary Condition

T1 2 T0
hA(T` 2 T0) 1 kA 1 g·0(ADx/2) 5 0 (5-24)
Dx
 276
HEAT TRANSFER

3. Radiation Boundary Condition

Tsurr
4
T1 2 T0
«sA(Tsurr 2 T04) 1 kA 1 g·0(ADx/2) 5 0 (5-25)
ε ∆x Dx
—–
2
4. Combined Convection and Radiation Boundary Condition
g· 0
εσ A(T 4surr – T 40 ) (Fig. 5–15)
T1 – T0 4
T1 2 T0
k A ——— hA(T` 2 T0) 1 «sA(Tsurr 2 T04) 1 kA 1 g·0(ADx/2) 5 0 (5-26)
∆x Dx
hA(T` – T0)
L
0 or
0 1 2 … x
A ∆x ∆x T1 2 T0
hcombined A(T` 2 T0) 1 kA 1 g·0(ADx/2) 5 0 (5-27)
Dx
hA(T` – T0) + εσ A(T 4surr – T 40 )
5. Combined Convection, Radiation, and Heat Flux Boundary
T1 – T0 · ∆x
+ kA ——— + g0 A —– = 0 Condition
∆x 2
FIGURE 5–15 4
T1 2 T0
Schematic for the finite difference q·0A 1 hA(T` 2 T0) 1 «sA(Tsurr 2 T04) 1 kA 1 g·0(ADx/2) 5 0 (5-28)
Dx
formulation of combined convection
and radiation on the left boundary 6. Interface Boundary Condition Two different solid media A and B
of a plane wall. are assumed to be in perfect contact, and thus at the same temperature
at the interface at node m (Fig. 5–16). Subscripts A and B indicate
Interface properties of media A and B, respectively.
Medium A Medium B Tm21 2 Tm Tm11 2 Tm
g·A,m g·B,m kAA 1 kBA 1 g·A, m(ADx/2) 1 g·B, m(ADx/2) 5 0 (5-29)
kA kB Dx Dx

Tm – 1 – Tm
kA A ————–
Tm + 1 – Tm
kB A ————–
In these relations, q·0 is the specified heat flux in W/m2, h is the convection
∆x ∆x coefficient, hcombined is the combined convection and radiation coefficient, T` is
the temperature of the surrounding medium, Tsurr is the temperature of the
m–1 m m+1 x surrounding surfaces, « is the emissivity of the surface, and s is the Stefan–
∆x ∆x Boltzman constant. The relations above can also be used for node M on the
A A
right boundary by replacing the subscript “0” by “M” and the subscript “1” by
∆x ∆x
—– —–
2 2
“M 2 1”.
Note that absolute temperatures must be used in radiation heat transfer
Tm – 1 – Tm Tm + 1 – Tm calculations, and all temperatures should be expressed in K or R when a
kA A ————– + kB A ————–
∆x ∆x
boundary condition involves radiation to avoid mistakes. We usually try to
· ∆x · ∆x
+ gA, m A —– + gB,m A —– = 0
2 2
avoid the radiation boundary condition even in numerical solutions since it
causes the finite difference equations to be nonlinear, which are more difficult
FIGURE 5–16
to solve.
Schematic for the finite difference
formulation of the interface boundary
condition for two mediums A and B
that are in perfect thermal contact. Treating Insulated Boundary Nodes as Interior Nodes:
The Mirror Image Concept
One way of obtaining the finite difference formulation of a node on an insu-
lated boundary is to treat insulation as “zero” heat flux and to write an energy
balance, as done in Eq. 5–23. Another and more practical way is to treat the
node on an insulated boundary as an interior node. Conceptually this is done
 277
CHAPTER 5

by replacing the insulation on the boundary by a mirror and considering the

reflection of the medium as its extension (Fig. 5–17). This way the node next Insulated
Insulation
to the boundary node appears on both sides of the boundary node because of boundary
symmetry, converting it into an interior node. Then using the general formula node
(Eq. 5–18) for an interior node, which involves the sum of the temperatures of 0 1 2 x
the adjoining nodes minus twice the node temperature, the finite difference
formulation of a node m 5 0 on an insulated boundary of a plane wall can be Mirror
expressed as
Tm11 2 2Tm 1 Tm21 g·m T1 2 2T0 1 T1 g·0 Mirror Equivalent
1 50 → 1 50 (5-30)
image interior
Dx2 k Dx2 k
node

which is equivalent to Eq. 5–23 obtained by the energy balance approach. x 2 1 0 1 2 x

The mirror image approach can also be used for problems that possess ther-
mal symmetry by replacing the plane of symmetry by a mirror. Alternately, we FIGURE 5–17
can replace the plane of symmetry by insulation and consider only half of the A node on an insulated boundary
medium in the solution. The solution in the other half of the medium is sim- can be treated as an interior node by
ply the mirror image of the solution obtained. replacing the insulation by a mirror.

EXAMPLE 5–1 Steady Heat Conduction in a Large Uranium Plate

Uranium
Consider a large uranium plate of thickness L 5 4 cm and thermal conductivity plate
k 5 28 W/m · °C in which heat is generated uniformly at a constant rate of 0°C h
g· 5 5 3 106 W/m3. One side of the plate is maintained at 0°C by iced water k = 28 W/m·°C
T`
while the other side is subjected to convection to an environment at T` 5 30°C g· = 5 × 106 W/m3
with a heat transfer coefficient of h 5 45 W/m2 · °C, as shown in Figure 5–18. L
Considering a total of three equally spaced nodes in the medium, two at the 0
0 1 2 x
boundaries and one at the middle, estimate the exposed surface temperature of
the plate under steady conditions using the finite difference approach.
FIGURE 5–18
Schematic for Example 5–1.
SOLUTION A uranium plate is subjected to specified temperature on one side
and convection on the other. The unknown surface temperature of the plate is
to be determined numerically using three equally spaced nodes.
Assumptions 1 Heat transfer through the wall is steady since there is no in-
dication of any change with time. 2 Heat transfer is one-dimensional since
the plate is large relative to its thickness. 3 Thermal conductivity is constant.
4 Radiation heat transfer is negligible.
Properties The thermal conductivity is given to be k 5 28 W/m · °C.
Analysis The number of nodes is specified to be M 5 3, and they are chosen
to be at the two surfaces of the plate and the midpoint, as shown in the figure.
Then the nodal spacing Dx becomes

L 0.04 m
Dx 5 5 5 0.02 m
M21 321

We number the nodes 0, 1, and 2. The temperature at node 0 is given to be

T0 5 0°C, and the temperatures at nodes 1 and 2 are to be determined. This
problem involves only two unknown nodal temperatures, and thus we need to
have only two equations to determine them uniquely. These equations are ob-
tained by applying the finite difference method to nodes 1 and 2.