Energy Conversion and Management 52 (2011) 1612–1621

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

The inaccuracy of heat transfer characteristics for non-insulated and insulated

spherical containers neglecting the influence of heat radiation
King-Leung Wong a,⇑, José Luis León Salazar a, Leo Prasad b, Wen-Lih Chen a
a
Department of Mechanical Engineering, Kun-Shan University of Technology, 949, Da-Wan Road, Yung-Kang City, Tainan County 710, Taiwan, ROC
b
Faculty of Built Environment, University of New South Wales, Sydney, NSW 2052, Australia

a r t i c l e i n f o a b s t r a c t

Article history: In this investigation, the differences of heat transfer characteristics for insulated and non-insulated
Received 27 July 2009 spherical containers between considering and neglecting the influence of heat radiation are studied by
Received in revised form 31 May 2010 the simulations in some practical situations. It is found that the heat radiation effect cannot be ignored
Accepted 6 October 2010
in conditions of low ambient convection heat coefficients (such ambient air) and high surface emissivi-
Available online 13 November 2010
ties, especially for the non-insulated and thin insulated cases. In most practical situations when ambient
temperature is different from surroundings temperature and the emissivity of insulation surface is differ-
Keywords:
ent from that of metal wall surface, neglecting heat radiation will result in inaccurate insulation effect
Insulation
Non-insulation
and heat transfer errors even with very thick insulation. However, the insulation effect considering heat
Sphere radiation will only increase a very small amount after some dimensionless insulated thickness (such
Heat radiation insulation thickness/radius =0.2 in this study), thus such dimensionless insulated thickness can be used
Emissivity as the optimum thickness in practical applications. Meanwhile, wrapping a material with low surface
Inaccuracy emissivity (such as aluminum foil) around the oxidized metal wall or insulation layer (always with high
surface emissivity) can achieve very good insulated effect for the non-insulated or thin insulated
containers.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction
Non-insulated and insulated cold/hot containers and ducts are
widely used in the various industries and domestic devices. Be-
cause heat radiation equation contains the 4th order exponent of
temperature, it is very difficult to obtain precise theoretical or
numerical heat radiation solutions before the time when modern
computing tools were not available. In the past, many heat transfer
experts and scholars avoid this problem and argue, from their own
experiences, that the heat radiation effect can be ignored when the
temperature difference between non-insulated or insulated
cold/hot containers and surrounding is small. As a result, heat radia-
tion has been commonly neglected in many practical heat transfer
applications even in cases involving low heat convection coefficients.
For the non-insulated situations, log mean temperature differ-
ence (LMTD) method, which neglects the influence of heat radia-
tion, is introduced in most heat transfer [such as 1,2], air
conditioning, and refrigeration text books [such 3,4], and is
⇑ Corresponding author. Tel.: +886 62057121; fax: +886 62050509.
E-mail address: klwong@mail.ksu.edu.tw (K.-L. Wong).
conventionally used to calculate the total heat transfer rate of heat
exchangers. Recently, Hsien et al. [5] studied the complete
heat transfer characteristics of non-insulated circular ducts consid-
ering heat radiation effect. They reported that in some practical sit-
uations, the heat radiation effect cannot be ignored when a non-
insulated duct has high surface emissivities and the ambient air’s
convection heat coefficient is low. Even in situations when the
temperature difference between the fluid inside a non-insulated
circular duct and the ambient air is very low, the errors generated
by neglecting heat radiation are still very large and hence cannot
be ignored. Unfortunately, it can be seen from Table 1 that most
surface emissivities of oxidized metal are greater than 0.64; and
Table 2 shows that the heat convection coefficients of ambient
air with medium wind speed are with quite small. Consequently,
the errors associated with the heat transfer characteristics ob-
tained with conventional LMTD method applying to condensers
and evaporators with heat exchanging in ambient air can be very
large. Thus, LMTD method should not be applied to evaporators
and condensers. To solve this problem, Wong et al. [6] developed
a log mean heat transfer rate (LMHTR) method which considers
the influence of heat radiation to calculate the exact heat transfer
rate of heat exchangers. Their study showed that the results
obtained by LMHTR method are the same as those obtained from

0196-8904/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enconman.2010.10.016
 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621 1613

Nomenclature

e emissivity of insulation surface qr radiation heat transfer rate of insulated container con-
e0 emissivity of container surface sidering heat radiation
A1 inner surface area of a spherical container qr0 radiation heat transfer rate of non-insulated container
A2 external surface area of a spherical container considering heat radiation
A3 external surface area of a insulated spherical container QR error of heat transfer rate generated by neglecting heat
EF insulation effect radiation
EFa insulation effect of an insulated spherical container con- r1 inner radius of spherical container
sidering heat radiation r2 external radius of spherical container
EFn insulation effect of an insulated spherical container r3 external radius of insulated spherical container
neglecting heat radiation SR error of surface temperature generated by neglecting
hi inner heat convection coefficient heat radiation
ho external heat convection coefficient TD the surface temperature difference generated by
hr external radiation heat convection coefficient neglecting heat radiation
HR convective heat coefficient ratio t1 thickness of spherical container wall
KA conductivity of duct material t thickness of insulated layer
Ks conductivity of insulated material t/r2 dimensionless insulated thickness
q total heat transfer rate of insulated container neglecting T2 bare wall surface temperature generated by neglecting
heat radiation heat radiation
q0 total heat transfer rate of non-insulated container T2a bare wall surface temperature generated by considering
neglecting heat radiation heat radiation
qa total heat transfer rate of insulated container consider- T3 insulated surface temperature generated by neglecting
ing heat radiation heat radiation
qa0 total heat transfer rate of non-insulated container con- T3a insulated surface temperature generated by considering
sidering heat radiation heat radiation
qc convection heat transfer rate of insulated container con- Ti temperature of the fluid inside the duct
sidering heat radiation To temperature of the fluid outside the spherical container
qc0 convection heat transfer rate of non-insulated container Tsur temperature of the outside surrounding
considering heat radiation

LMTD method while surface emissivities equal to zero; and they especially in the situations involving low ambient air convec-
indicated that it is risky to neglect the influence of heat radiation tion coefficients and heat exchanger (such as condensers and
evaporators) surfaces with high surface emissivity. Thus, in order
Table 1 to obtain more accurate results, in stead of the conventional LMTD
The emissivities e of various substances from the manual of infrared temperature method, the LMHTR method should be employed. However, inves-
demonstrator [24]. tigations related to non-insulated heat transfer to ambient air
Human skin 0.98 neglected the heat radiation are still quite common. For example,
Gold 0.02 Elsayed et al. [7] studied free convection of air around a constant
Silver 0.02
Aluminum Weathered = 0.83; Foil (bright)=0.04 Disk, rough = 0.96
Copper Polished = 0.05
Table 2
Oxidized = 0.78
Referred approximate values of convection heat transfer [13].
Iron Cast(ox) = 0.64
Sheet, rusted = 0.69 Approximate values of convection heat transfer, h (W m2 K1)
Stainless steel Polished = 0.16 Mode h
Oxidized = 0.85 (W m2 K1)
Steel Polished = 0.07
Free convection
Oxidized = 0.79
Temp. diff. = 30 °C horizontal plate 0.3 in high in air 4.5
Nickel Electro pole = 0.05
Temp. diff. = 30 °C vertical plate 0.3 in high in air 6.5
Brick 0.81
Horizontal cylinder, 2 cm diameter, in water 890
Carbon 0.95
Heat transfer across 1.5 cm vertical air gap with temp. 2.64
Concrete 0.95
diff. = 60 °C
Glass 0.84–0.97
Paint oil 0.94 Forced convection
Paper, white 0.70 Air flow at 2 m/s over 0.2-m square plate 12
Paper 0.89 Air flow at 35 m/s over 0.75-m square plate 75
Plaster 0.86 Air at 2 atm flowing in 2.5 cm diameter tube at 10 m/s 65
Rubber, black 0.95 (=36 km/h)
Wood, oak 0.90 Water at 0.5 kg/s flow in 2.5 cm diameter tube 3500
White ceramic 0.91 Air flow across 5 cm diameter cylinder with velocity of 180
Black painting 0.96 50 m/s(=180 km/h)
Oil, lubricant Film 0.03 mm = 0.27
Boiling water
Film 0.13 mm = 0.72
In a pool or container 2500–35,000
Thick = 0.82
Flowing in a tube 5000–
Soil Dry = 0.92
100,000
Saturated water = 0.95
Water Distilled = 0.96 Condensation of water vapor, 1 atm
Frost = 0.98 Vertical surfaces 4000–11,300
Snow = 0.85 Outside horizontal tubes 9500–25,000
1614 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621

heat flux elliptic tube without considering heat radiation; Lazaro 2.1. Cases with the influence of the heat radiation being neglected
et al. [8] investigated the PCM-air heat exchangers for free cooling
applications in buildings neglecting the heat radiation effect; Here, we assume that heat transfer is steady state without heat
Karadag and Teke [9] developed a new approach relevant to floor generation inside solid materials. All thermal properties are uni-
Nusselt number in floor heating system without considering heat form and independent of temperature, and hi and ho are constants.
radiation; Linhui et al. [10] reported an experimental study on If the influence of external surface heat radiation is not considered,
natural convective heat transfer from a vertical plate with discrete it can be seen from Fig. 1(a) that the total thermal resistance can be
heat sources mounted on the back neglecting heat radiation; Chen derived from heat transfer text book [such as 15] is:
et al. [11] solved the inverse problem in determining convection    
1
heat transfer coefficient of an annular fin without considering heat X 1 r1
 r12 1
r2 1  r13
Rth ¼ þ þ þ ð1Þ
radiation; Chang et al. [12] investigated the heat pipe for cooling of hi 4pr 21 4pK A 4p K S h0 4pr 23
electronic equipment neglecting heat radiation; and the list goes
on. and the relative total heat transfer rate neglecting heat radiation is:
For the insulated situations, it is commonly shown in many heat Ti  T0 T3  T0
transfer text books [such as 13–15] that heat radiation is neglected q¼ P ¼ 1
ð2Þ
Rth h 4pr 2 0
in insulated duct/container even in situations with ambient air at 3

low heat convection coefficients. Meanwhile, the critical heat Eq. (2) implies that heat transfer rate q and insulated surface
transfer does occur with ambient air at a very small heat convec- temperature T3 of situations without considering the influence of
tion coefficient. From Table 1, it can be seen that the surface emis- external surface heat radiation can be obtained. From Fig. 2(a),
sivities of insulated material are greater than 0.8. Thus, the heat the heat transfer rate of non-insulated spherical container without
radiation should be considered in the analysis of critical heat trans- considering the influence of outside heat radiation can be written
fer. But the critical heat transfer characteristics introduced in most as:
heat transfer text books [such as 13–15] neglect the heat radiation.
In addition, many investigations associated with insulation in open
literature also neglected the influence of heat radiation, for exam-
ple: Chen and Yang [16] used an iterative regularization method in
estimating the transient heat transfer rate on the surface of the
insulation layer of a double circular pipe; Ahmed et al. [17] inves-
tigated the heat transfer characteristics across the insulated walls
of refrigerated truck trailers by the application of phase change
materials; Karadag and Teke [18] investigated the floor Nusselt
number in floor heating system for insulated ceiling conditions;
Atayilmaz and Teke [19] obtained the experimental and numerical
study of the natural convection from a heated horizontal cylinder
wrapped with a layer of textile material; Chen et al. [20] found
out the reliable one-dimensional approximate solution of insulated
oval duct, Lee et al. [21] investigated the complete heat transfer
solutions of an insulated regular polyhedron by using a RPSWT
model, Hsien et al. [22] found out the reliable one-dimensional
approximate solutions for insulated oblate spheroid containers,
etc.
Recently, Hsien et al. [5] conducted a comparative study on the
heat transfer characteristics of an insulated circular duct consider-
ing and neglecting the influence of heat radiation. They found that
the heat radiation effect cannot be ignored in conditions of thin
insulation with low ambient convective heat coefficient and high
surface emissivity. In order to highlight the importance of heat
radiation for non-insulated and insulated ducts/containers in situ-
ation where heat convection coefficients of ambient air is small,
this paper, a follow-up study to Hsien et al. [5], examines the dif-
ferences in heat transfer characteristics of a non-insulated and
insulated spherical container resulted by considering and neglect-
ing the influence of heat radiation. Then the inaccuracy of heat
transfer characteristics for a non-insulated and insulated spherical
container neglecting the influence of heat radiation can be
demonstrated.

2. Problem formulation

Insulated and non-insulated spherical containers are shown in

Figs. 1 and 2, respectively. Both spherical containers are exposed
to a surrounding temperature of Tsur, and have a wall thickness
of t1, conductivity KA, bare wall surface emissivity e0. The insulation
layer has thickness t, conductivity KS and surface emissivity e. The
internal and external fluids have convection heat transfer coeffi- Fig. 1. An insulated spherical container and relative parameters: (a) neglecting heat
cients hi and ho, temperatures Ti and To, respectively. radiation and (b) considering heat radiation.
 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621 1615

where T3a is the actual surface temperature considering heat radia-

tion. Then from Fig. 1(b), the surface convective heat transfer
becomes:

qc ¼ h0 4pr 23 ðT 3a  T 0 Þ ð6Þ
and the surface radiation heat transfer rate is:
 
qr ¼ re4pr 23 T 43a  T 4sur ð7Þ

Comparing Eqs. (2) and (6), under the conditions of e – 0,

T3a – T3, qc – q, the energy conservation among qa, qc and qr can
be written as:
qa ¼ qc þ qr ð8Þ
Therefore, the qa, qc and qr and T3a can be readily deduced from Eqs.
(5)–(8).
With both q and qa being solved, the error of heat transfer rate
generated by neglecting heat radiation effect can be defined as:
 
q
QR ¼ 1   100% ð9Þ
qa
The equivalent heat convection coefficient of heat radiation [15] is
defined from Eq. (7) as:
 
hr ¼ re T 23a þ T 2sur ðT 3a þ T sur Þ ð10Þ

Here, hr is normally used to compare with the heat convection

coefficient ho which in turn shows how significant the effect of
radiation is. The ratio between equivalent heat convection coeffi-
cient of heat radiation and original heat convection coefficient is
defined as:
hr hr 4pr23 ðT 3a  T sur Þ q
HR ¼  100%   100% ¼ r  100% ð11Þ
h0 h0 4pr23 ðT 3a  T 0 Þ qc
where Tsur is always close to To. Similar to Eq. (9), the error of surface
temperature generated by neglecting heat radiation effect, based on
T3 and T3a using Celsius temperature scale, is defined as:
 
T3
SR ¼ 1  100% ð12Þ
T 3a
In the case of the insulated cold spherical container, condensed
water may be formed if the insulated surface temperature is less
than the dew point of ambient air. Therefore, predicting the insu-
lated surface temperature is very important. Dimensionless SR
Fig. 2. A non-insulated spherical container and relative parameters: (a) neglecting tends to reduce the magnitude of error; hence SR is not suitable
heat radiation and (b) considering heat radiation. to demonstrate the heat characteristic of cold spherical container.
Instead, the surface temperature difference is used and is written
Ti  T0 as:
q0 ¼   ð3Þ
1
1 1
r 1 þr 2 TD ¼ T 3a  T 3 ð13Þ
hi 4pr 21
þ 4pK A
þ h 41pr2
0 2
While the influence of external surface heat radiation is consid-
Eqs. (2) and (3) lead to an insulated effect EFn of an insulated ered, the complete heat transfer rate from the surface of a non-
spherical container comparing with non-insulated situation insulated spherical container shown in Fig. 2(b) can be written as:
neglecting heat radiation being:
T i  T 2a
  qa0 ¼   ð14Þ
q 1 1
EF n ¼ 1   100% ð4Þ 1
r 1 r 2
q0 hi 4pr 21
þ 4pK A

2.2. Case with the influence of heat radiation being considered where T2a is the actual surface temperature of a non-insulated
spherical container shown in Fig. 2(b) with the influence of heat
From Fig. 1, when the influence of external surface heat radia- radiation taken into account. Additionally, its surface convective
tion is considered, the complete heat transfer rate from the surface heat transfer rate is:
of an insulated spherical container can be expressed as: qc0 ¼ ho 4pr 23 ðT 2a  T o Þ ð15Þ
T T
qa ¼  i 3a   ð5Þ and the surface radiation heat transfer rate is:
1 1
r 1 r 2
1 1
r2 r 3  
1
hi 4pr21
þ 4pK A
þ 4pK S qr0 ¼ re4pr23 T 42a  T 4sur ð16Þ
1616 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621

Table 3 The heat balance can be expressed as:

Referred approximate values of thermal conductivities [13].
qa0 ¼ qc0 þ qr0 ð17Þ
Thermal conductivity of various materials at 20 °C
Metals Similarly, qco, qro, qao and T2a of non-insulated spherical container
Material K (W m1 K1) considering heat radiation can be readily obtained from Eqs. (14)–
Copper (pure) 386
Aluminum (pure) 204
(17). And finally, the insulated effect EFa of an insulated spherical
Carbon steel, l% C 73–77 container comparing with non-insulated situation when heat radi-
Carbon steel (18%Cr, 8%Ni) 43 ation is considered can be obtained by:
Cast iron 16  
q
Nonmetallic solids EF a ¼ 1  a  100% ð18Þ
Glass, window 0.78 qa0
Plaster, gypsum 0.48
Metal lath 0.4
Woof lath 0.28 3. Calculation of heat transfer results
Teflon 0.35
Asphalt 0.7
The one-dimensional heat transfer solutions of the insulated
Wood fiber sheet 0.047
Wool 0.038 and non-insulated spherical containers are exact. These solutions
Glass fiber 0.035 are calculated by one dimensional LabVIEW [23] programming,
Building brick common 0.69 which provide an easy and friendly interface to input all needed
Building brick face 1.32 parameters and also offer live graph, table data, and automatic
Concrete, cinder 0.76
Stone, 1–2–4 mix 1.37
connection with excel. With this handy tool, it is possible to obtain
accurate values with one-dimensional spherical heat equation. The
Graphite, pyrolytic
Perpendicular to layers 5.6
major parameter in this program is the insulation thickness, and
Polyethylene 0.33 according to the input insulation thickness, Eqs. (5) and (8) and
Polypropylene 0.16 equations from (14) to (17) are solver iteratively until the differ-
Polyvinylchloride 0.09 ence of heat transfer rates obtained between two consecutive iter-
Rubber, hard 0.1
ations is less than 104. The iteration process converges quickly

ºC ºC

ºC ºC ºC ºC
Fig. 3. The relations between QR and t/r2 in the situation of Ti = 100 °C, Fig. 4. The relations between HR and t/r2 in the situation of Ti = 100 °C,
KA = 77W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm and KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm and
ho = 10 W m2 K1 with e = e0: (a) To = Tsur = 30 °C and (b) To = 30 °C, Tsur = 32 °C. ho = 10 W m2 K1 with e = e0: (a) To = Tsur = 30 °C and (b) To = 30 °C, Tsur = 32 °C.
 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621
and only takes a few seconds to complete, therefore, with each
change of parameters, accurate solutions can be returned in a very
short period of time. According to the emissivities shown in Table
1, e = 0.8 and 0.9 and e = 0.1 and 0.2 are adopted to represent the
cases of high and low surface emissivity, respectively. Also Table
2 shows that the natural convection coefficients of air are below
10 W m2 K1; and even in the cases with a very high air speed,
the forced convection coefficients of air are less than
100 W m2 K1. Therefore, hi = 30 W m2 K1 can be used to repre-
sent the situation of low/medium convection coefficients inside
the spherical container, and ho = 8.3 and 10 W m2 K1 are chosen
to represent the natural or low forced convection coefficients of
ambient air. Meanwhile, the natural convection coefficients of
water are more than 890 W m2 K1, and the values of convection
coefficients are between 2500 and 35,000 W m2 K1 for boiling
water in a pool or container. Nevertheless, in insulated cases, con-
vection coefficients may become smaller, thus hi = 800 W m2 K1
is chosen to represent the cases of high convection coefficients of
hot or cold liquid (may be other than water) inside the insulated
spherical container. In practical applications, carbon steel is the
common material applied to construct containers. According to
the Table 3, KA = 77 W m1 K1 is used as the carbon steel spherical
container’s conductivity; and the insulated material conductivity
with the value KS = 0.035 W m1 K1shown in Table 3 is adopted
in this study.
In order to check if the computer results are reliable, the follow-
ing measures are adopted:
ºC
ºC ºC
Fig. 5. The relations between SR and t/r2 in the situation of Ti = 100 °C,
KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm and
ho = 10 W m2 K1 with e = e0: (a) To = Tsur = 30 °C and (b) To = 30 °C, Tsur = 32 °C.
1617
(a) Let surface emissivity e = 0, i.e., heat radiation does not exist
and HR = hr/ho = 0; check if the resulted QR being close to
zero.
(b) Let surface emissivity e = 1 and external convection coeffi-
cient ho = 50,000 W/m2 K, i.e., heat radiation effect becomes
very small by comparing the very big heat convection. And
HR = hr/ho becomes very small; check if the resulted and
QR being close to zero.
4. Results and discussions
For the results to be generalized, they are shown in dimension-
less parameters except for TD, the surface temperature difference
between considering and neglecting heat radiation effect for insu-
lated cold spherical container. This is because that the value of dif-
ference in dimensionless temperatures is very small, and it might
result in a misunderstanding of physical meaning among readers.
The first investigation is on the case of insulated hot spherical
containers with the following thermal and geometrical parame-
ters: Ti = 100 °C, KA = 77 W m1 K1, Ks = 0.035 W m1 K1,
r1 = 190 mm, r2 = 200 mm and ho = 10 W m2 K1 at different
ambient-air conditions: (a) To = Tsur = 30 °C and e = e0 and (b)
To = 30 °C, Tsur = 32 °C, e = e0 or e – e0, respectively. Figs. 3–6 shows
the heat transfer rate error (QR), convective coefficients ratio (HR),
error of surface temperature (SR), and insulated effect (EF) are af-
fected by dimensionless insulated thickness (t/r2), internal convec-
tion coefficient (hi), insulated surface emissivity (e), and metal wall
surface emissivity (e0).
ºC
ºC ºC
Fig. 6. The relations between EF and t/r2 in the situation of Ti = 100 °C,
KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm and
ho = 10 W m2 K1: (a) To = Tsur = 30 °C with e = e0 and (b) To = 30 °C, Tsur = 32 °C
with e – e0.
1618 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621
ºC
ºC ºC
Fig. 7. The relations between QR and t/r2 in the situation of Ti = 20 °C,
KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm
ho = 8.3 W m2 K1 with e = e0: (a) To = Tsur = 30 °C and (b) To = 30 °C, Tsur = 32 °C.
Figs. 3, 4 and 6 show that QR and HR decrease, and EF increases
as t/r2 increases. The higher the e and hi are, the larger the QR, HR
and EF will be. The plots in Fig. 3(a) indicate that in the condition
of To = Tsur = 30 °C and e = e0, QR gets close to 0 as t/r2 approaches
to 1. This phenomenon, however, does not happen at the condition
of To = 30 °C, Tsur = 32 °C, and e = e0 as shown by the plots given in
Fig. 3(b). Here QR gets close to 3% as t/r2 approaches to 1; thus
in the condition of To – Tsur, neglecting the heat radiation produces
error in QR even with very thick insulation. Fig. 4 suggests that HR
value becomes larger at t/r2 = 0. It is noticeable from Fig. 4(a) that
even at t/r2 = 1, HR is about 50% for e = 0.8 and HR ; 12% for
e = 0.2 under the condition of To = Tsur = 30 °C. Fig. 4(b) also shows
that even at t/r2 = 1, HR reaches 58% for e = 0.8 and HR is 7.5% for
e = 0.2 in the condition of To = 30 °C and Tsur = 32 °C. HR represents
the ratio between heat radiation and heat convection as shown in
Eq. (11), thus heat radiation effect cannot be neglected if HR > 5%.
Fig. 5 shows that SR exists, and its absolute value reaches maximum
around t/r2 = 0.025; the absolute value of SR is generally propor-
tional to e. If the value of e is fixed, smaller hi tends to result in lar-
ger absolute value of SR before t/r2 ; 0.025 then the trend reverses
after that value. Fig. 6(a) shows that for same values of hi and e = e0,
EFa values at the condition of e = e0 – 0 (considering heat radiation)
are greater than that at e = e0 = 0 (neglecting heat radiation), mean-
ing neglecting heat radiation will underestimate insulation effect.
The slope of EFa (considering heat radiation) becomes much smaller
after t/r2 = 0.2, and the value of EFa is within 3% of between
t/r2 = 0.2 and t/r2 = 1; thus t/r2 = 0.2 can be used as the optimum
dimensionless insulation thickness in practical applications.
The most meaningful practical characteristics of non-insulated
and insulated spherical container are shown in Fig. 6(b). Fig. 6(b)
ºC
ºC ºC
and
Fig. 8. The relations between HR and t/r2 in the situation of Ti = 20 °C,
KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm and
ho = 8.3 W m2 K1 with e = e0: (a) To = Tsur = 30 °C and (b) To = 30 °C, Tsur = 32 °C.
shows that at t/r2 = 0, the value of EFa (;36%) at e = 0.1, e0 = 0.79,
and hi = 800 W m2 K1 is much greater than that (EFa ; 3%) at
e = 0.9, and with the same values of e0 and hi; meanwhile EFa
(;24%) at e = 0.1, e0 = 0.79, and hi = 30 W m2 K1 is much greater
than that (EFa ; 3%) at e = 0.9, and with the same values of e0
and hi. Hence, it can be concluded that for non-insulated and thin
insulated cases, wrapping an aluminum foil (e = 0.04 as shown in
Table 1) around the oxidized metal wall or insulation layer (e at
high values as shown in Table 1) can result in excellent insulated
effect. This practice works if t/r2 5 0.15. The slope of EFa becomes
much smaller after t/r2 = 0.2, and the value of EFa is within 3% be-
tween t/r2 = 0.2 and t/r2 = 1. Therefore, t/r2 = 0.2 can be treated as
the optimum dimensionless insulation thickness in practical appli-
cations. When t/r2 = 0.0025 and for same values of hi, EFa value at
e – e0 > 0 (considering heat radiation) is greater than that at
e = e0 = 0 (neglecting heat radiation). Thus neglecting heat radia-
tion will result in inaccurate insulation effect in most situations.
The second simulation is to examine insulated cold spherical
containers with the following thermal and geometrical parameters:
Ti = 20 °C, KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm,
r2 = 200 mm and ho = 8.3 W m2 K1 at different ambient-air condi-
tions: (a) To = Tsur = 30 °C with e = e0 and (b) To = 30 °C, Tsur = 32 °C
with e = e0 and e – e0, respectively. Figs. 7–10 give the heat transfer
rate error (QR), convective coefficients ratio (HR), the difference of
surface temperature (TD), and insulated effect (EF) are affected by
the dimensionless insulated thickness (t/r2), internal convection
coefficient (hi), insulated surface emissivity (e), and metal wall sur-
face emissivity (e0). The characteristics observed in Figs. 7–10 are
very similar to those in Figs. 3–6. It can be seen in Figs. 7, 8 and
10 that QR decreases, but EF and HR increase as t/r2 increases. The
 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621
ºC
ºC ºC
Fig. 9. The relations between TD and t/r2 in the situation of Ti = 20 °C,
KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm
ho = 8.3 W m2 K1 with e = e0: (a) To = Tsur = 30 °C and (b) To = 30 °C, Tsur = 32 °C.
higher the e and hi are, the larger the QR, HR and EFa will be. Fig. 7(a)
shows under the condition of To = Tsur = 30 °C and e = e0, QR ap-
proaches zero value as t/r2 approaches 1. But this is not the case
for the condition of To = 30 °C, Tsur = 32 °C, and e = e0 as shown in
Fig. 7(b). Fig. 7(b) shows that in the above condition, QR gets close
to 5% as t/r2 approaches 1. This suggests that in practical situation
of To – Tsur, neglecting the heat radiation produces errors in QR
even with very thick insulation. Fig. 8 shows that HR value remains
stable at t/r2 = 0 but begins to increase at t/r2 = 1. It can be seen in
Fig. 8(a), at t/r2 = 1, HR is about 60% for e = 0.8 and about 15% for
e = 0.2 in the condition of To = Tsur = 30 °C. In Fig. 8(b), even at t/
r2 = 1, HR is about 70% for e = 0.8 and about 8% for e = 0.2 in situa-
tion of To = 30 °C and Tsur = 32 °C. HR represents the ratio between
heat radiation and heat convection as shown in Eq. (11), and heat
radiation effect cannot be neglected if HR > 5%. The results indicate
that heat radiation cannot be neglected within the whole range of
insulation thickness. In Fig. 9, TD reaches its maximum value near
t/r2 ; 0.025, and the absolute value of TD becomes higher with lar-
ger e; but if the value of e is fixed, smaller hi tends to result in larger
absolute value of TD at some point before the maximum TD then
the trend reverses after that point. Fig. 10(a) shows that for the
same values of hi and e = e0, EFa values of e = e0 – 0 (considering
heat radiation) are larger than those of e = e0 = 0 (neglecting heat
radiation). This means that neglecting heat radiation will result in
inaccurate the insulation effect. The slope of EFa becomes much
smaller after t/r2 = 0.2, and the value of EFa between t/r2 = 0.2 and
t/r2 = 1 is within 4%, thus t/r2 = 0.2 can be used as the optimum
dimensionless insulation thickness in practical application.
In the second set of cases, the most significant characteristics of
non-insulated and insulated spherical container are shown in
1619
ºC
ºC ºC
Fig. 10. The relations between EF and t/r2 in the situation of Ti = 20 °C,
KA = 77 W m1 K1, Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm and
ho = 8.3 W m2 K1: (a) To = Tsur = 30 °Cwith e = e0 and (b) To = 30 °C, Tsur = 32°Cwith
and e – e0.
Fig. 10(b). It can be seen that at t/r2 = 0, the value of EFa (;28%)
at e = 0.1, e0 = 0.79, and hi = 800 W m2 K1 is much greater than
that (EFa ; 2%) at e = 0.9 and with same values of e0 and hi. An-
other example is that the value of EFa (;21%) at e = 0.1, e0 = 0.79,
and hi = 30 W m2 K1 is much greater than that (EFa ; 2%) at
e = 0.9 and with same values of e0 and hi. The results indicate that
for non-insulated and thin insulated cases, wrapping an aluminum
foil (e = 0.04 as shown in Table 1) around the oxidized metal wall
or insulation layer (e at high values as shown in Table 1) can
achieve very good insulated effect; this advantage exists until
t/r2 5 0.15. The slope of EFa becomes much smaller after
t/r2 = 0.2, the value of EFa is within 4% between t/r2 = 0.2 and
t/r2 = 1, thus t/r2 = 0.2 can be used as the optimum dimensionless
insulation thickness in the practical applications. While t/r2 =
0.0025 and for same values of hi, the values of EFa at e – e0 > 0
(considering heat radiation) are greater than those at e = e0 = 0
(neglecting heat radiation), suggesting that neglecting heat radia-
tion will underestimate the insulation effect in most situations.
The data shown in Tables 4 and 5 serve to demonstrate the inac-
curacy of heat transfer characteristics for insulated and non-insu-
lated spherical containers neglecting the influence of heat
radiation. It can be seen that the values of considering heat radia-
tion are very different from those of neglecting heat radiation. For
the non-insulated case, data at t/r2 = 0 with e0 = 0.9 listed in Table 4
show that qa is 367.0 W m2, which is larger than q(=256.7
W m2), and qc is 215.8 W m2, which is quite different from q.
Thus, q is obtained by neglecting heat radiation but its value does
not stand for heat convection. Due to the effect of heat radiation,
T2a is 72.93 °C, a smaller value than T2 (=81.07 °C). This is because
that qc ¼ ho 4pr 22 (T2a  To) is not equal to q ¼ ho 4pr 22 (T2-To). Since q
1620 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621

Table 4
Ti = 100 °C; To = Tsur = 30 °C, KA = 77 W m1 K1; Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm, t1 = 10 mm, ho = 10 W m2 K1, e = e0 = 0.9.

t (mm) t/R2 qa (W m2) qr (W m2) qc (W m2) q (W m2) QR (%) HR (%) T3a (°C) T3 (°C) SR (%)
0 0 367.0 151.2 215.8 256.7 30.06 70.09 72.93 81.07 11.16
0.5 0.003 318.6 129.2 189.3 233.3 26.77 68.27 67.48 76.18 12.90
1.0 0.005 281.9 113.0 168.9 213.9 24.12 66.89 63.27 72.13 14.00
1.5 0.008 253.1 100.5 152.6 197.5 21.95 65.81 59.92 68.72 14.69
2.5 0.013 210.7 82.39 128.3 171.5 18.59 64.23 54.90 63.28 15.28
3.5 0.018 180.8 69.97 110.9 151.7 16.11 63.11 51.30 59.15 15.29
5.0 0.025 149.6 57.22 92.37 129.5 13.40 61.94 47.49 54.53 14.82
7.5 0.038 116.7 44.10 72.63 104.6 10.43 60.72 43.42 49.33 13.59
10 0.05 96.17 36.05 60.12 87.99 8.50 59.96 40.85 45.88 12.31
30 0.15 42.38 15.55 26.83 41.01 3.23 57.98 34.04 36.17 6.27
40 0.2 34.05 12.45 21.59 33.23 2.40 57.68 32.98 34.59 4.87
50 0.25 28.83 10.52 18.30 28.29 1.88 57.49 32.33 33.60 3.93
60 0.3 25.25 9.21 16.05 24.87 1.52 57.37 31.89 32.93 3.26
80 0.4 20.67 7.52 13.15 20.45 1.08 57.21 31.33 32.08 2.36
100 0.5 17.86 6.49 11.37 17.71 0.81 57.12 31.01 31.57 1.81
200 1 12.09 4.39 7.70 12.05 0.31 56.94 30.38 30.60 0.71

Table 5
Ti = 20 °C; To = Tsur = 30 °C, KA = 77 W m1 K1; Ks = 0.035 W m1 K1, r1 = 190 mm, r2 = 200 mm, t1 = 10 mm, ho = 8.3 W m2 K1, e = e0 = 0.9.

t (mm) t/R2 qa (W m2) qr (W m2) qc (W m2) q (W m2) QR (%) HR (%) T3a (°C) T3 (°C) TD (°C)
0 0 221.7 81.28 140.4 159.5 28.04 57.89 3.65 8.24 4.58
0.5 0.003 198.4 73.64 124.7 146.8 25.99 59.04 0.25 5.01 5.26
1.0 0.005 179.4 67.27 112.2 136.0 24.20 59.97 3.38 2.28 5.66
1.5 0.008 163.8 61.89 101.9 126.7 22.62 60.74 5.94 0.07 5.87
2.5 0.013 139.5 53.35 86.13 111.6 19.97 61.95 9.86 3.90 5.96
3.5 0.018 121.5 46.90 74.63 99.8 17.84 62.84 12.72 6.88 5.84
5.0 0.025 102.0 39.72 62.25 86.33 15.34 63.81 15.80 10.30 5.49
7.5 0.038 80.68 31.75 48.94 70.69 12.39 64.87 19.10 14.26 4.84
10 0.05 66.99 26.52 40.46 60.06 10.34 65.55 21.20 16.94 4.26
30 0.15 30.03 12.09 17.94 28.77 4.20 67.38 26.75 24.79 1.96
40 0.2 24.18 9.76 14.42 23.42 3.15 67.66 27.60 26.10 1.50
50 0.25 20.50 8.29 12.21 19.99 2.48 67.84 28.13 26.93 1.19
60 0.3 17.98 7.27 10.70 17.61 2.02 67.96 28.48 27.50 0.98
80 0.4 14.73 5.97 8.76 14.52 1.44 68.11 28.93 28.22 0.70
100 0.5 12.73 5.16 7.57 12.60 1.09 68.20 29.19 28.66 0.54
200 1 8.63 3.50 5.12 8.59 0.42 68.37 29.69 29.49 0.21

is obtained by neglecting heat radiation, its value does not stand
for heat convection, and qr(=151.2 W m2) cannot be neglected
while comparing with the value of qc (=215.8 W m2); the resulted
HR ; qr/qc is 70.09%. If it is neglected, a very big QR (=30.06%) is
produced for non-insulated case. For the insulated case in Table 4,
data at t/r2 = 1 with e0 = 0.9 show that qa is 12.09 W m2, which is
very close to q(=12.05 W m2). However, qc is 7.07 W m2, whose
value is very different from q. Due to the effect of heat radiation,
T3a is 30.38 °C, a smaller value than T3(=30.60 °C). This is because
that qc ¼ ho 4pr 23 (T3a-To) is not equal to q ¼ ho 4pr23 (T3-To). Thus, q
is obtained by neglecting heat radiation but its value does not
stand for heat convection. Also qr(=4.39 W m2) cannot be ne-
glected while comparing with the value of qc (=7.07 W m2) be-
cause the resulted HR ; qr/qc would be 56.94%. It is interesting
that a very small QR (=0.31%) is produced for thick insulated case
(t/r2 = 1) even when heat radiation plays a very important role. It
can be explained that both T3a(=30.38 °C) and T3(=30.60 °C) are
very close to To and Tsur (both temperatures are assumed to be
30 °C) in the very thick insulated case (t/r2 = 1). Thus the difference
between qa(=12.09 W m2) and q(=12.05 W m2) is quite small.
The above explanations can be applied to all data shown in Tables
4 and 5.
5. Conclusion
This paper shows that the influence of the heat radiation effect
is very important in situation where the convection coefficient of
ambient air is of low value. Neglecting heat radiation will result
in large errors for non-insulated and thin insulated spherical con-
tainers with a high surface emissivity. Meanwhile, using computer
programming (like LabVIEW [23] in this study) which takes the
heat radiation effect into account in a common PC can easily obtain
highly accurate results. Since taking heat radiation effect into ac-
count by modern computers only requires some additional trivial
effort, the heat radiation should not be neglected in any theoretical
analysis of the heat transfer characteristics of non-insulated and
insulated ducts and containers.
Acknowledgment
The authors would like to thank the National Science Council of
Taiwan, ROC for the financial support of this study, which was
completed under the project numbered NSC-97-2221-E168-044-
MY2.
References
[1] Bejan A. Heat transfer. John Wiley & Sons Inc.; 1993. p. 42–5.
[2] Kreith F, Bohn MS. Principles of heat transfer. 5th ed. Harper International
Edition; 1993. p. 21–41.
[3] Dossat RJ. Principles of refrigeration. 3rd ed. Prentice Hall; 1991. p. 167–78.
[4] Chatenever R. Air conditioning and refrigeration for the professional. John
Wiley & Sons Inc.; 1988. p. 445–8.
[5] Hsien TL, Wong KL, Yu SJ. The inaccuracy of heat transfer characteristics of
insulated and non-insulated circular duct while neglecting the influence of
heat radiation. Energy Convers Manage 2009;50(11):2856–63.
 K.-L. Wong et al. / Energy Conversion and Management 52 (2011) 1612–1621
[6] Wong KL, Ke MT, Ku SS. The log mean heat transfer rate method of heat
exchanger considering the influence of heat radiation. Energy Convers Manage
2009;50(11):2693–8.
[7] Elsayed AO, Ibrahim EZ, Elsayed SA. Free convection from a constant heat flux
elliptic tube. Energy Convers Manage 2003;44(15):2445–53.
[8] Lazaro A, Dolado P, Marin JM, Zalba B. PCM-air heat exchangers for free-cooling
applications in buildings: experimental results of two real-scale prototypes.
Energy Conver Manage 2009;50(3):439–43.
[9] Karadag R, Teke I. New approach relevant to floor Nusselt number in floor
heating system. Energy Convers Manage 2008;49(5):1134–40.
[10] Linhui C, Huaizhang T, Yanzhong L, Dongbin Z. Experimental study on natural
convective heat transfer from a vertical plate with discrete heat sources
mounted on the back. Energy Convers Manage 2006;47(18):3447–55.
[11] Chen WL, Yang YC, Lee HL. Inverse problem in determining convection heat
transfer coefficient of an annular fin. Energy Convers Manage 2007;48(4):
1081–8.
[12] Chang YW, Cheng CH, Wang JC, Chen SL. Heat pipe for cooling of electronic
equipment. Energy Convers Manage 2008;49(11):3398–404.
[13] Holman JP. Heat transfer. ninth edition of international edition. McGraw-Hill
Inc.; 2002. p. 26–37.
[14] Incropera FP, DeWitt DP, Theodore LB, Adrienne SL. Fundamentals of heat and
mass transfer. 6th ed. John Wiley & Sons Inc.; 2007. p. 116–25.
[15] Yunus AC. Heat transfer-a practical approach. second edition in SI
units. McGraw-Hill Inc.; 2004. p. 151–5.
1621
[16] Chen WL, Yang YC. An iterative regularization method in estimating the
transient heat-transfer rate on the surface of the insulation layer of a double
circular pipe. Energy Convers Manage 2009;50(12):3096–103.
[17] Ahmed M, Meade O, Medina MA. Reducing heat transfer across the insulated
walls of refrigerated truck trailers by the application of phase change
materials. Energy Convers Manage 2010;51(3):383–92.
[18] Karadag R, Teke I. Investigation of floor Nusselt number in floor heating system
for insulated ceiling conditions. Energy Convers Manage 2007;48(3):967–76.
[19] Atayilmaz SO, Teke I. Experimental and numerical study of the natural
convection from a heated horizontal cylinder wrapped with a layer of textile
material. Int Commun Heat Mass Transfer 2010;37(1):58–67.
[20] Chen W, Wong KL, Hsien TL, Huang CT. Reliable one-dimensional
approximate solution of insulated oval duct. Energy Convers Manage
2008;49(8):2214–24.
[21] Lee JF, Wong KL, Chen WL, Ku SS. Complete heat transfer solutions of an
insulated regular polyhedron by using a RPSWT model. Energy Convers
Manage 2005;46:2232–57.
[22] Hsien TL, Wong KL, Chen WL, Siao CC. Reliable one-dimensional approximate
solutions for insulated oblate spheroid containers. Energy Convers Manage
2009;50(8):1871–8.
[23] LabVIEW 7.1 Edition. National Instruments Corporation Product.
[24] Manual of IR Thermography Device: ThermaCAM™ E45, FLIR Systems, Sweden
Product.