PROBLEM 2.1.

FAM

GIVEN:
Consider a steady-state, two-dimensional heat flux vector field given by

q = 3x2 sx + 2xy sy .

The control volume is centered at x = a and y = b, with sides 2∆x and 2∆y (Figure Pr.2.1).

SKETCH:
Figure Pr.2.1 shows a control volume centered at x = a and y = b with side widths of 2∆x and 2∆y.

y, sy
sn

sn
2Dy b
sn

sn
x, sx
a
2Dx
Figure Pr.2.1 A finite control volume in a two-dimensional heat transfer medium.

The depth (along z direction) is w.

OBJECTIVE:
(a) Using the above expression for q show that
R
A
q · sn dA
lim = ∇ · q,
∆V →0 ∆V
where the divergence of the heat flux vector is to be evaluated at x = a and y = b.
Use a length along z of w (this will not appear in the final answers). (Hint: Show that you can obtain the same
final answer starting from both sides.)
(b) If the divergence of the heat flux vector is nonzero, what is the physical cause?
(c) In the energy equation (2.1), for this net heat flow (described by this heat flux vector field), is the sum of the
volumetric terms on the right, causing the nonzero divergence of q, a heat source or a heat sink? Also is this a
uniform or nonuniform volumetric source or sink? Discuss the behavior of the heat flux field for both positive
and negative values of x and y.

SOLUTION:
(a) To prove the validity of q above for the region shown in Figure Pr.2.1, we calculate separately the left-hand
side and the right-hand side. For the right-hand side we have
µ ¶
∂ ∂
· 3x2 sx + 2xy sy .
¡ ¢
∇ · q = sx + sy
∂x ∂y

Performing the dot product we have

∂ 3x2
¡ ¢
∂ (2xy)
∇·q= + ,
∂x ∂y

which results in

∇ · q = 6x + 2x = 8x.

10
Applying the coordinates of the center of the control volume, we have finally
∇ · q|(x=a, y=b) = 8a.
The left-hand side can be divided into four integrals, one for each of the control surfaces:
(i) Control Surface at x = a − ∆x :
The heat flux vector across this control surface and the normal vector are
q1 = 3(a − ∆x)2 sx + 2(a − ∆x)y sy
sn1 = −sx .
The dot product between q and sn is
q1 · sn1 = 3(a − ∆x)2 (sx · −sx ) + 2(a − ∆x)y(sy · 0) = −3(a − ∆x)2 .
The net heat flow over this control surface is
Z Z b+∆y
Q|A1 = (q1 · sn1 ) dA = −3(a − ∆x)2 dyw = −6(a − ∆x)2 ∆yw.
A1 b−∆y

(ii) Control Surface at x = a + ∆x:

The heat flux vector across this control surface and the normal vector are
q2 = 3(a + ∆x)2 sx + 2(a + ∆x)y sy
sn2 = sx .
The net heat flow over this control surface is
Z Z b+∆y
Q|A2 = (q2 · sn2 ) dA = 3(a + ∆x)2 dyw = 6(a + ∆x)2 ∆yw.
A2 b−∆y

(iii) Control Surface at y = b − ∆y:

The heat flux vector across this control surface and the normal vector are
q3 = 3x2 sx + 2x(b − ∆y) sy
sn3 = −sy .
The net heat flow over this control surface is
Z Z a+∆x
Q|A3 = (q3 · sn3 ) dA = −2x(b − ∆y)dxw
A3 a−∆x
2 2
(a + ∆x) − (a − ∆x)
= −2(b − ∆y) w = −4a∆x (b − ∆y) w.
2
(iv) Control Surface at y = b + ∆y:
The heat flux vector across this control surface and the normal vector are
q4 = 3x2 sx + 2x(b + ∆y) sy
sn4 = sy .
The net heat flow over this control surface is
Z Z a+∆x
Q|A4 = (q4 · sn4 ) dA = 2x(b + ∆y)dxw
A4 a−∆x
2 2
(a + ∆x) − (a − ∆x)
= 2(b + ∆y) w = 4a∆x (b + ∆y) w.
2
Adding up the heat flow across all the surfaces, we have
Q|A = Q|A1 + Q|A2 + Q|A3 + Q|A4
Z Z Z Z
= q1 · sn1 dA + q2 · sn2 dA + q3 · sn3 dA + q4 · sn4 dA
A1 A2 A3 A4
= [−6(a − ∆x)2 ∆y + 6(a + ∆x)2 ∆y − 4a∆x (b − ∆y) + 4a∆x (b + ∆y)]w
= (24a∆x∆y + 8a∆x∆y)w = 32a∆x∆yw.

11
Now, applying the limit
R
A
q · sn dA 32a∆x∆yw
lim = lim = lim 8a = 8a,
∆V →0 ∆V ∆x,∆y→0 (2∆x)(2∆y)w ∆x,∆y→0

which is identical to the result found before. These are the two methods of determining the divergence of the
heat flux vector for a given location in the heat transfer medium.

(b) Since this is a steady-state heat flux vector field (i.e., q is not a function of time t), the only reason not to have
a divergence-free field would be the presence of a heat generation or sink. In this case, the differential-volume
energy equation is
X
∇·q= ṡi .
i

The heat generation or sink is caused by the conversion of work or other forms of energy to thermal energy. In
the energy equation, these energy conversions are called source terms. The source terms ṡi could be due to
(i) conversion from physical or chemical bond to thermal energy
(ii) conversion from electromagnetic to thermal energy
(iii) conversion from mechanical to thermal energy

(c) The divergence of the heat flux vector q given above is 8x. For x > 0, this is a positive source term indicating
a heat generation. For x < 0, the source term becomes negative indicating a heat sink. Also, since the source
term is a function of x, it is a nonuniform source term in the x direction and a uniform source term in the y
direction.

COMMENT:
The application of the divergence operator on the heat flux vector (as in the differential-volume energy equa-
tion) results in an expression valid for any position x and y while the application of the area-integral (as in the
integral-volume energy equation) results on a number which is valid only for that specific point in space x = a and
y = b. The integral form of the energy equation gives an integral or overall energy balance over a specified closed
region within the medium, while the differential form is pointwise valid, i.e., is satisfied for any point within the
medium.
For the control surfaces parallel to the x axis, the dot product between the heat flux vector and the surface
normal was a function of x (variable). That required the integration along x. The integration is simplified in the
case of a constant heat flux vector normal to the control surface, as obtained for the control surfaces parallel to
the y axis. Although the first case is more general, here we will mainly deal with situations in which the heat
flux normal to the control surface is constant along the control surface. This will allow the use of the thermal
resistance concept and the construction of thermal resistance network models, as it will be discussed starting in
Chapter 3.

12
PROBLEM 2.4.FUN

GIVEN:
A nitrogen meat freezer uses nitrogen gas from a pressurized liquid nitrogen tank to freeze meat patties as
they move carried by a conveyor belt. The nitrogen flows inside a chamber in direct contact with the meat patties,
which move in the opposite direction. The heat transfer mechanism between the nitrogen gas and the meat patties
is surface convection. Meat patties are to be cooled down from their processing (initial) temperature of Ti = 10◦C
to the storage (final) temperature of To = −15◦C. Each meat patty has a mass M = 80 g, diameter D = 10
cm, and thickness l = 1 cm. Assume for the meat the thermophysical properties of water, i.e., specific heat in
the solid state cp,s = 1,930 J/kg-K , specific heat in the liquid state cp,l = 4,200 J/kg-K, heat of solidification
∆hls = −3.34 × 105 J/kg, and freezing temperature Tls = 0◦C. The average surface-convection heat transfer
between the nitrogen and the meat patties is estimated as qku = 4,000 W/m2 and the conveyor belt moves with
a speed of uc = 0.01 m/s.

OBJECTIVE:
(a) Sketch the temperature variation of a meat patty as it move along the freezing chamber.
(b) Neglecting the heat transfer between the conveyor belt and the meat patties, find the length of the freezing
chamber. Use the simplifying assumption that the temperature is uniform within the meat patties. This allows
the use of a zeroth-order analysis (lumped-capacitance analysis).

SOLUTION:
(a) The temperature variation of the meat patties as they move along the freezing chamber is given in Figure
Pr.2.4.

Cooling Solid
Cooling Liquid Regime
Regime
T , oC
Phase-Change
Ti = 10 oC Regime

Tls = 0 oC

T0 = - 15 oC

ti = 0 t1 t2 t0 t,s
Figure Pr.2.4 Variation of meat patty temperature with respect to time.

(b) To calculate the necessary length for the freezing chamber, the cooling process is divided into three regimes
(shown in Figure Pr.2.4).
(i) Regime 1: Cooling of Liquid
During this period of time, the meat patties are cooled from their initial temperature down to the solidification
temperature. Application of the integral-volume energy equation for a control volume enclosing the meat gives
Z µ ¶
d
Z
qku · sn dA = − ρcp T dV.
Aku V dt

Assuming that qku is constant and normal to the surface and that the meat temperature and properties are
constant throughout the meat patty (lumped-capacitance analysis), the energy equation becomes

dT
qku Aku = −ρcp V .
dt

13
Integrating the equation above from t = ti = 0 to t = t1 , for a constant qku , gives
Z t1 Z Tls
qku Aku dt = − ρcp V dT
ti =0 Ti

ρcp V (Ti − Tls ) M cp (Ti − Tls )

t1 = =
qku Aku qku Aku
πD 2
From the data given Aku = 4 + πDl = 0.011 m2 and

0.08(kg) 4,200(J/kg-K) × [10(◦C) − 0(◦C)]

t1 = 2 = 76.36 s = 1.273 min.
4,000(W/m ) 0.011(m2 )

(ii) Regime 2: Solidification

During this regime the meat patties change phase from liquid to solid. Application of the integral-volume energy
equation gives
Z Z
qku · sn dA = ṡls dV.
As V

Again, assuming that qku is uniform and normal to the surface and that the meat properties are constant
throughout the meat patty (lumped-capacitance analysis), the energy equation becomes

qku Aku = ṡls V,

where the volumetric heat consumption due to phase change ṡls is obtained from Table 2.1,

ṡls = −ṅls ∆hℓs .

The volumetric solidification rate ṅls (kg/m3 -s) is given by

m
ṅls = .
V (t2 − t1 )

Using the relations above, the energy equation becomes

m∆hls
qku Aku = − ,
(t2 − t1 )

and solving for t2 − t1 ,

m∆hls
t2 − t1 = − .
qku Aku

From the values given,

0.08(kg) (−3.34 × 105 )(J/kg)

t2 − t1 = − 2 = 607.3 s = 10.12 min.
4,000(W/m ) 0.011(m2 )

(iii) Regime 3: Cooling of Solid

During this period of time, the meat patties are cooled from the melting temperature down to the final tem-
perature. Application of the lumped-capacitance analysis for a control volume enclosing the meat results in an
equation similar to t1 = ρcp V (Ti − Tls )/qku Aku = M cp (Ti − Tls )/qku Aku , i.e.,

ρcp V (Tls − To ) mcp (Tls − To )

to − t2 = = .
qku Aku qku Aku

14
From the data given,

0.08(kg) × 1,930(J/kg-K) × [0(◦C) − (−15)(◦C)]

to − t2 = 2 = 52.64 s = 0.8773 min.
4,000(W/m ) 0.011(m2 )

The total time for cooling of the meat patties is therefore

to = t1 + (t2 − t1 ) + (to − t2 ) = 1.273 + 10.12 + 0.8773 = 12.27 min.

For the velocity of the conveyor belt uc = 0.01 m/s, the total length necessary is

L = uc to = 0.01(m/s) × 12.27(min) × 60(s/min) = 7.362 m.

COMMENT:
Phase change at constant pressure for a pure substance occurs at constant temperature.
The temperature evolution for regimes 1 and 3 are linear because qku has been assumed constant (note that all
the properties are treated as constants). In practice, the heat loss by surface convection depends on the surface
temperature and therefore is not constant with time when this surface temperature is changing. This will be
discussed in Chapter 6.
The freezing regime accounts for more than 80 s of the total time, while the cooling of solid accounts for only
7 s of the total time. This is a result of the high heat of solidification exhibited by water and the relatively smaller
specific heat capacity of ice compared to liquid water. Liquid water has one of the largest specific heat capacities
among the pure substances. The specific heat capacity of substances will be discussed in Chapter 3.

15
PROBLEM 2.5.FUN

GIVEN:
While the integral-volume energy equation (2.9) assumes a uniform temperature and is applicable to many
heat transfer media in which the assumption of negligible internal resistance to heat flow is reasonably justifiable,
the differential-volume energy equation (2.1) requires no such assumption and justification. However, (2.1) is
a differential equation in space and time and requires an analytical solution. The finite-small volume energy
equation (2.13) allows for a middle ground between these two limits and divides the medium into small volumes
within each of which a uniform temperature is assumed. For a single such volume (2.9) is recovered and for a
very large number of such volumes the results of (2.1) are recovered.
Consider friction heating of a disk-brake rotor, as shown in Figure Pr.2.5. The energy conversion rate is Ṡm,F .
The brake friction pad is in contact, while braking, with only a fraction of the rotor surface (marked by R).
During quick brakes (i.e., over less than t = 5 s), the heat losses from the rotor can be neglected.
Note that Ṡm,F remains the same, while ∆V changes.

Ṡm,F = 3×104 W, Ro = 18 cm, Ri = 13 cm, l = 1.5 cm, R = 15 cm, ρcp = 3.5×106 J/m3 -K, T (t = 0) = 200 C,
t = 45.
SKETCH:
Figure Pr.2.5 shows the rotor and the area under the pad undergoing friction heating.

Pad Contact Area

Sm,F
l
dT
-rcpDV
dt

Ro
QA= 0 R
Ri

T(t = 0)
Initial Temperature

Figure Pr.2.5 A disc-brake rotor heated by friction heating. The region under the brake pad contact is also shown.

OBJECTIVE:
Apply (2.13), with (i) the volume marked as the pad contact region, and (ii) the entire volume in Figure
Pr.2.5, and determine the temperature T after t = 4 s for cases (i) and (ii) and the conditions given above.
Note that the resulting energy equation, which is an ordinary differential equation, can be readily integrated.

SOLUTION:
Starting from (2.13), we have
d
Q|A = − (ρcp T )∆V ∆V + Ṡm,F
dt
dT
0 = −ρcp ∆V + Ṡm,F
dt
or by separating the variables, using T |t=0 = T (t = 0) and integrating from 0 to t, we have
Z T Z t
Ṡm,F
dT = dt.
T (t=0) t=0 ρcp ∆V

Ṡm,F
T (t) − T (t = 0) = (t − 0)
ρcp ∆V
or
Ṡm,F
T (t) = T (t = 0) + t.
ρcp ∆V

16
Then using the numerical values we have

3 × 104 (W)
T (t = 4 s) = 20(◦C) + × 4(s)
3.5 × 106 (J/m3 -K) × ∆V (m3 )
3.429 × 10−2 ◦
= 20(◦C) + ( C).
∆V
(i) The smaller volume gives

∆V = π(Ro2 − R2 )l
= π(0.182 − 0.152 )(m2 ) × 0.015(m) = 4.665 × 10−4 m3

T (t = 4 s) = 20(◦C) + 73.50(◦C) = 93.50◦C.

(ii) The larger volume gives

∆V = π(Ro2 − Ri2 )l
= π(0.182 − 0.132 )(m2 ) × 0.015(m) = 7.305 × 10−4 m3

T (t = 4 s) = 20(◦C) + 46.94(◦C) = 66.94◦C.

COMMENT:
For more accurate results, the radial length as well as the length along l are divided into small-finite volumes
and then heat transfer is allowed between them. This is discussed in Section 3.7.

17
PROBLEM 2.8.FUN

GIVEN:
In some transient heat transfer (i.e., temperature and heat flux vector changing with time) applications, that
portion of the heat transfer medium experiencing such a transient behavior is only a small portion of the medium.
An example is the seasonal changes of the air temperature near the earth’s surface, which only penetrates a very
short distance, compared to the earth’s radius. Then the medium may be approximated as having an infinite
extent in the direction perpendicular to the surface and is referred to a semi-infinite medium. Figure Pr.2.8 shows
such a medium for the special case of a sudden change of the surface temperature from the initial (and uniform
throughout the semi-infinite medium) temperature of T (t = 0) to a temperature Ts . Under these conditions, the
solution for the heat flux is given by

k[Ts − T (t = 0)] 2
/4αt
qk,x (x, t) = 1/2
e−x ,
(παt)

where α = k/ρcp is called the thermal diffusivity.

k = 0.25 W/m-K (for nylon), α = 1.29 × 10−5 m2 /s (for nylon), Ts = 105◦C, T (t = 0) = 15◦C, xo = 1.5 cm,
to = 30 s.
This conduction heat flux changes with time and in space.

SKETCH:
Figure Pr.2.8 shows the semi-infinite slab, the conduction heat flux, and the local energy storage/release.

Storage
-rcp ¶T xo
¶t Dx 0
x
sn x 1
Semi-Infinite Slab
Initially at Uniform
Temperature
T = T(t = 0) Far-Field
Constant Surface
Temperature, Ts -qk,x(x,t)
T(t,x 1
Temperature
) = T(t = 0)

Conduction

Figure Pr.2.8 A semi-infinite slab with an initial temperature T (t = 0) has

its surface temperature suddenly changed to Ts .

OBJECTIVE:
(a) Using (2.1), with no energy conversion and conduction as the only heat transfer mechanism, determine the
time rate of change of local temperature ∂T /∂t at location xo and elapsed time to .
(b) Determine the location of largest time rate of change (rise) in the temperature and evaluate this for the
elapsed time to .

SOLUTION:
(a) Starting from (2.1) and for no energy conversion and a one-dimensional (in the x direction) conduction only,
we have
µ ¶
∂ ∂T
∇·q = sx · (qk,x sx ) = −ρcp
∂x ∂t
∂ ∂T
qk,x = −ρcp
∂x ∂t
or

∂T 1 ∂
− = qk,x .
∂t ρcp ∂x

18
Using the given expression for qk,x (x, t), we have

µ ¶ µ ¶ −x2
∂T k Ts − T (t = 0) −2x
= − e 4αt
∂t ρcp (παt)1/2 4αt
−x2
Ts − T (t = 0)
= xe 4αt .
2π 1/2 α1/2 t3/2
Evaluating this at xo and to , we have

−x2o
∂T Ts − T (t = 0)
= xo e 4αto .
∂t 2π 1/2 α1/2 t3/2
o

Using the numerical values, we have

−
(1.5 × 10−2 )2 (m)2
∂T (105 − 15)( C) ◦ −5 2
= × 1.510−2 (m)e 4 × 1.29 × 10 (m /s) × 30(s)
∂t 2π 1/2 (1.29 × 10−5 )1/2 (m2 /s)1/2 (30)3/2 (s)3/2
= 0.6453(◦C/s) × e−0.1453 = 0.5580◦C/s.

(b) We now differentiate the above expression for ∂T /∂t, with respect to x, at which we find the location of the
largest ∂T /∂t occurs. Then by differentiating and using t = t0 , we have

∂ ∂T ∂ 2 qk,x
=
∂x ∂t ∂x2
x2 x2
 
Ts − T (t = 0)  - 4αto 2x2 - 4αto 
= e − e = 0.
2π 1/2 α1/2 t3/2
o
4αto

Then
2x2
1− =0
4αto
or

x = (2αto )1/2 .

Now using the numerical values, we have

x = [2 × 1.29 × 10−5 (m2 /s) × 30(s)]1/2

= 0.02782 m = 2.782 cm.

From part (a), we have

∂T
= 1.1968(◦C/s) × e−0.5 = 0.7259◦C/s.
∂t
COMMENT:
Note that as expected, ∂T /∂t = 0 at x = 0 (because Ts is assumed constant). In Section 3.5.1, we will discuss
this transient problem and define the penetration front as the location beyond which the effect of the surface
temperature change has not yet penetrated and this distance is given as x ≡ δα = 3.6(αt)1/2 , as compared to
x = 1.414(αt)1/2 for the location of maximum ∂T /∂t.

19