International Communications in Heat and Mass Transfer 39 (2012) 582586

Contents lists available at SciVerse ScienceDirect

International Communications in Heat and Mass Transfer

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

A simple method for predicting bulk temperature from tube wall temperature with
uniform outside wall heat ux
M. Lin a, Q.W. Wang a,, Z.X. Guo b
a
b

MOE Key Laboratory of Thermo-Fluid Science and Engineering, Xi'an Jiaotong University, Xi'an, 710049, China
Department of Mechanical and Aerospace Engineering, Rutgers, the State University of New Jersey, NJ 08854, USA

a r t i c l e

i n f o

Available online 5 April 2012

Keywords:
Bulk temperature
Wall temperature
Axial wall conduction
Laminar ow
Uniform heat ux

a b s t r a c t
An empirical approach is proposed to estimate the bulk temperature in practical laminar tube ow. To
examine the correlation, heat transfer in different types of tubes with wall conduction and uniform constant
heat ux at tube outer wall surface is numerically investigated. The predictions from the proposed correlation
match well with the numerical results in all the cases studied for air ow in the Pe range from 105 to 1032
and for water ow in the Pe range from 70 to 700. The method is further testied via comparison with
experimental data and numerical results of mini (micro)-channel water ow available in literature.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
Extensive studies on conjugate heat transfer in a tube with
internal laminar ow have been performed using analytical and
numerical methods in the past three decades. Mori et al. [1]
investigated the wall conduction effects in a circular tube with either
uniform constant heat ux or constant temperature at the outer
surface of the tube. Faghri and Sparrow [2] numerically studied the
coupled effect between solid wall and uid axial conductions. It was
found that the effect of the wall axial conduction can readily
overwhelm the effect of the uid axial conduction. The generalized
conjugate Graetz problem with axial conduction was studied in Refs.
[36]. All of the above-mentioned literature obtained temperature
proles assuming a known and temperature-independent velocity
prole. As a matter of fact, the real velocity prole is not known and
both the velocity and temperature elds are simultaneously developing at the entrance regime in most cases.
Experimentally the tube outer wall temperature can be easily
measured whereas it is impossible to measure the local uid bulk
temperature. It would be very useful if the bulk temperatures can be
acquired via the measured wall temperatures in order to calculate
heat transfer coefcients. To this end, a simple method oriented from
the experimental point of view is presented in this paper which
avoids the complexity of the afore-mentioned analytical and
numerical approaches. Air is considered as the working uid, with
Peclet number varying from 105 to 1032. In addition, the proposed
method for predicting water bulk temperature is veried through
comparison with published experimental data and numerical results.
Communicated by P. Cheng and W.Q. Tao
Corresponding author.
E-mail address: wangqw@mail.xjtu.edu.cn (Q.W. Wang).
0735-1933/$ see front matter 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.icheatmasstransfer.2012.03.017

Meanwhile, two common tube materials (copper and stainless steel)

are also considered to extend the exibility of the method.
2. Physical model and proposed method
The physical model is a circular tube with nite wall thickness
shown in Fig. 1, whose inner diameter is Di and the ratio of length to
inner diameter (L/Di) is set to 100 in order to reach fully-developed
state. The conjugate heat transfer problem is analyzed in laminar ow
condition. When air is considered as the working uid, natural
convection is negligible. The laminar ow is characterized by steadystate, constant physical property and three-dimensional situations
with no viscous dissipation. The general governing equations are
omitted here due to space limitation.
A mass-ow-inlet is set as the inlet boundary condition and a
pressure-outlet is set as the outlet boundary condition. The ambient
uid temperature (Te) is 283.15 K for air (Pr = 0.7). A uniform
constant heat ux boundary condition is imposed on the outer
surface of the tube wall (qwo = const) and the adiabatic boundary
condition is imposed on both ends of the tube wall. The numerical
models are solved by commercially available computer CFD software
(FLUENT 6.3).
It is well known that, under the boundary condition of uniform
constant heat ux at the circular tube without consideration of wall
thickness, the uid temperature distribution is linearly increased
along the streamwise direction [7]. At the fully-developed region the
temperature difference between the wall temperature and the uid
bulk temperature (Tw Tf) is a constant.
In fact, any circular tube comes with a nite wall thickness. In
experiments, the outer-wall temperature (Two) of a tube can be
measured by thermocouples; and the inner-wall temperature (Twi) is
obtained via one-dimensional Fourier heat conduction analysis. The

M. Lin et al. / International Communications in Heat and Mass Transfer 39 (2012) 582586

Nomenclature
c
d
Dh
Di
L
q
Pe
Pr
Re
T
x*
x
x, y, z

coefcient in Eq. (2a)

coefcient in Eq. (2a)
hydraulic diameter (m)
inner diameter of outer-tube (m)
tube length (m)
wall heat ux (W/m 2)
Peclet number (=RePr)
Prandtl number
Reynolds number
temperature (K)
dimensionless axial position (4x/DhPe)
axial position (m)
Cartesian coordinates

Greek symbols

thermal conductivity (W/m K)

thickness (m)

error indicator

dimensionless temperature

auxiliary function, Eq. (2a)

Subscripts
in
inlet
e
ambient uid temperature
f
uid
L
tube end
m
mean
max
maximum
out
outlet
w
wall
wi
inside wall
wo
outside wall
x
along x direction, local
0
tube entrance
1
results based on Eq.(5)
2
results based on Eq.(6)
3
results based on numerical results
*
dimensionless

583

laminar tube ow without wall heat conduction or in turbulent ow

[8,9]; but not appropriate for lower Reynolds number laminar
ow because of the existence of axial wall heat conduction. If the
inlet uid temperature is to be measured, there exist ve known
parameters, namely, Te, Tf(0) (=Tin), Tf(L) (=Tout), Twi(0) and Twi(L);
then a linear correlation for the dimensionless temperature, , is
obtainable in laminar tube ow. The dimensionless temperature,(x),
is dened as
x T wi xT f x=T wi xT e

where Twi is the inner-wall temperature of the tube wall, Te is the

ambient uid temperature, and Tf is the bulk temperature.
For an ideal circular tube with negligible wall thickness, the
reciprocal of can be expressed as a linear function of the axial
coordinate in the absence of axial wall heat conduction and (0) = 1
because of Tf(0) = Tin = Te. Fig. 2a shows the numerical results of the
variation of (1/) versus x* = 4x/Dh/Pe. As shown clearly in Fig. 2a, the
distributions of (1/) vs. x* are nearly linear for air (Pe = 105 and
1032).
For an actual circular tube with nite wall thickness, however, the
uid will be preheated (Tin > Te) before entering into the heated tube
because of the axial wall conduction. So the dimensionless temperature (0) at x = 0 is less than 1 ((0) b 1), and the difference of (1
(0)) could be congenially regarded to represent the axial
heat transfer rate. The numerical results of the variation of 1/((x)+1
(0)) are illustrated in Fig. 2b and c, from which a quasi-linear equation
can be obtained. It is seen that the linearity of 1/((x)+1(0)) for a
circular tube with nite wall thickness is quite good at low thermal
conductivity and high Peclet number but poor at high thermal
conductivity and low Peclet number. However, the error for predicting
uid bulk temperature is satisfactory in terms of engineering application as described in the next section because (x) is a decimal between
0 and 1.
Now the quasi-linear equation for a circular tube with axial wall
heat conduction is given as
x

1
c dx
x 10

2a

at
x 0; x 0 : 0 T wi 0T in =T wi 0T e ;0 1

2b

at
inner-wall temperature (Twi) may be assumed of the outer-wall
temperature (Two) for tubes of high thermal conductivity or thin-wall
tubes. The ambient uid temperature (Te) could be measured by
Mercury thermometer or other measuring instruments. Usually, the
uid outlet temperature (Tout) of a test tube is measured. The inlet
temperature of the uid is approximately equal to the ambient uid
temperature (Tin = Tf(0) = Te). Such consideration is correct in

4L=Dh
:
x L; x
Pe

1
L T wi LT out =T wi LT e ;L
L 10

For a circular tube, the hydraulic diameter is Dh = Di.

Fig. 1. Sketch of a circular tube with specied thermal boundary conditions.

2c

584

M. Lin et al. / International Communications in Heat and Mass Transfer 39 (2012) 582586

(a) ideal tube with wall neglected

and the dimensionless temperature, (x), is

20

x 011=x:

1/

15

Air
Circular tube (w/Dh=0)
Pe=105
Pe=1032

Then, the uid bulk temperature in the circular tube with wall
conduction is obtained as follows
T f1 x T wi x011=xT wi xT e

10

0
0

T f2 x T e T out T e x=L:

x*

Air, Copper, w/Dh=0.5

Pe=105
Pe=1032

1/((x)+1-(0))

25

In order to compare the results based on Eqs. (5) and (6),

respectively, the corresponding error indicators () are dened as

(b) real tube with copper wall

30

i T fi T f3 m =T out T e ; i 1;2

15

where i represents the dimensionless mean deviation of the bulk

temperature and Tf3 is directly obtained from the numerical
simulation and treated as the correct value.

10

3. Results and discussion

20

3.1. Circular tube with wall conduction

0
0

x*

(c) real tube with stainless steel wall

Air, Stainless steel, w/Dh=0.5
Pe=105
Pe=1032

20

1/((x)+1-(0))

Eq. (5) is the proposed correlation in this study. Traditionally, the

bulk temperature of uid in a circular tube without wall heat
conduction is assumed to be linearly increased along the uid ow
direction under constant heat ux boundary condition [7]:

15

10

0
0

x*
Fig. 2. Variation of dimensionless temperature in different circular tubes.

The two coefcients in Eq. (2a) are solved as follows

c1

3a



4L=Dh
:
d L1=
Pe

3b

So the quasi-linear equation for circular tubes with wall conduction is

x 1

4b

L1 
x
4L=D
Pe

4a

To check the grid-independence the present computations were

primarily performed with three grid densities: 180,044, 787,336, and
1,365,048 mesh points. It is found that the relative deviation of Nu
was only 2.5% between the coarse and ne meshes and 0.8% between
the intermediate and ne meshes for air in the case of Re = 100 and
x/L = 0.6. Therefore, the intermediate mesh system was adopted in
the numerical modeling calculations thereafter.
The air temperature distributions of circular tubes with thick
wall (w/Dh = 1.5) are presented in Fig. 3ab. Both low and high
thermal conductivity materials for the tube are considered: copper
( = 398 W.m 1.K 1) and stainless steel ( = 16.6 W.m 1.K 1).
There are three dened local bulk temperatures: the black solid
squares in the gures represent Tf1 estimated from Eq. (5), the hollow
circles indicate Tf2 predicted by Eq. (6), and the black solid triangles
represent Tf3 calculated from numerical modeling. The hollow
squares in the gures indicate the wall temperature (Twi) resulted
from the numerical modeling. From Fig. 3, it is found that the linear
Tf2 proles obtained assuming negligible axial wall heat conduction
deviate obviously from the Tf3 proles obtained from the numerical
modeling; whereas the Tf1 proles from the currently proposed
Eq. (5) match closer to the numerical results (Tf3). For the tube made
of copper, the maximum temperature difference between Tf1 and Tf3
is 4.4 K and 5.1 K, respectively for Pe = 105 and 1032. However,
the maximum temperature difference between Tf2 and Tf3 is 42.1 K
and 20.6 K, correspondingly. For the tube made of stainless steel,
the maximum temperature difference between Tf1 and Tf3 is 3.4 K
and 4.8 K for Pe = 105 and 1032, respectively; while the maximum
temperature difference between Tf2 and Tf3 is 27.6 K and 7.2 K for
Pe = 105 and 1032, respectively. The errors of mean temperature
deviation are listed and compared in Table 1. It is seen that for the
tubes made of high thermal conductivity copper, 1 is 4.7% whereas
2 is 51.3% at Pe = 105. When Pe increases to 1032, 1 is only 0.5%
whereas 2 is 19.8%. The same tendency appears in Table 1 for
the tubes made of low thermal conductivity stainless steel. At
Pe = 105, 1 is 3.3% whereas 2 is 33.9%. It can be concluded
that the proposed Eq. (5) has a more accurate estimation of the air
bulk temperature than that of Eq. (6) for thick-wall tubes.

M. Lin et al. / International Communications in Heat and Mass Transfer 39 (2012) 582586

400

(a) Pe=105
360

Circular tube ( w/Dh=1.5)

Tf1 (Te=292.05K)
Tf1 (Te=291.75K)
Tf3 (Fig.14 in Ref.[11])
Twi

Pe=105

340
360

-4.4K

Tf1(Eq.5)
Tf2(linear)
Tf3(simulation)
Twi(simulation)

T,K

300
Copper

280
320

T,K

320

585

320

Stainless steel

300
-3.4K

280

0.000

260

0.002

0.004

0.006

0.008

0.010

x, m
0

x*

Fig. 4. Comparison of temperature distributions with results in Ref. [11].

(b) Pe=1032
390

3.2. Micro-channel with wall conduction

Circular tube (w/Dh=1.5) Pe=1032

360
330
300

T,K

Tf1 (Eq.5)
Tf2 (linear)
Tf3 (simulation)
Twi (simulation)

-5.1K
Copper

270
360

Stainless steel

330

-4.8K

300
0.0

0.1

0.2

0.3

0.4

x*
Fig. 3. Comparison of predicted temperature distributions in different circular tubes
(w/Dh = 1.5).

The case of circular tube with thin-wall (w/Dh = 0.5) is also

studied. The errors of mean temperature deviation are shown in
Table 1. It is seen that, even for thin-wall tubes, the estimation from
Eq. (5) is more accurate than that from the linear Eq. (6), in particular
at low Peclet number.
The applicability of Eq. (5) is extended to water ow in circular
tubes (w/Dh = 0.5) made of either copper or stainless steel. Two
different Peclet numbers (Pe = 70 and 700) are selected. It was found
that the water bulk temperature proles predicted by the numerical
modeling for the water tube made of high thermal conductivity
copper was not linear and the proles predicted by Eq. (5) matched
closely to the numerical results. The errors of mean temperature
deviation were that 1 was 4.1% and 0.7% at Pe = 70 and 700,
respectively; while 2 was 18.5% and 2.9%, correspondingly. For
the tube made of stainless steel, both the 1 and 2 were every small
and within acceptable accuracy.

w/Dh = 1.5

Copper
Stainless Steel

w/Dh = 0.5

Copper
Stainless Steel

4. Conclusions
In this study, a simple correlation with acceptable engineering
accuracy is proposed to estimate the bulk temperature from the
measured wall temperature in laminar tube ow. In this correlation,
ve measurable parameters should be prior known, namely Te, Tin,
Tout, Twi(0), and Twi(L). The correlation is testied in circular air tubes
with Pe between 105 and 1032 as well as in micro-channels of water
ow (Pe = 70 700). A good agreement between the correlative
results and the corresponding numerical and experimental results
is found. The proposed method has a wide scope of engineering
applications and can be applied to either macro-channels or mini
(micro)-channels.
Acknowledgments

Table 1
Relative errors of mean uid temperature deviation (%).
Cases

The proposed correlative Eq. (5) is further testied with the

experimental data in Ref. [10] and the numerical results in Ref. [11],
in which water-ow heat transfer in a 10 mm long micro-channel,
57 m wide and 180 m deep, was considered. The inside wall
temperature can be thought to equal to the measured outside wall
temperature because of the high thermal conductivity of silicon and
thin wall ( = 0.45 mm). The wall temperatures were Twi(0) = 313.15 K
and Twi(L) = 382.15 K. The uid temperatures were Tin = 293.15 K and
Tout = 378.15 K. The ambient uid temperature was not given, and thus,
is assumed as Te = 291.75 K or 292.05 K. Fig. 4 shows the predicted bulk
temperature proles based on Eq. (5) with the wall temperatures
extracted from Fig. 14 in Ref. [11] for the case Re = 42 (P = 6kPa). The
bulk temperatures (Tf3) in the gure were also extracted from Fig. 14 in
Ref. [11] and are regarded as the correct values. It is observed that the
two Tf1 proles with different Te values (292.05 K or 291.75 K) are
nearly identical and the mean temperature difference between these
two bulk temperatures is only 0.4 K. Moreover, the Tf1 proles are very
close to the Tf3 prole. The relative error of the mean temperature
deviation (1) is 1.3% and 0.8% at Te = 292.05 K and 291.75 K,
respectively. The results calculated from Eq. (5) are quite accurate.

Error

Pe = 105

Pe = 1032

1
2
1
2
1
2
1
2

4.7
51.3
3.3
33.9
3.6
45.3
1.1
12.3

0.5
19.8
3.2
7.7
2.5
15.0
2.4
2.5

This work was supported by the National Natural Science

Foundation of China (grant no. 51025623) and Foundation of Key
Laboratory of Low-grade Energy Utilization Technologies and Systems of MOE (grant no. LEUTS 201003).
References
[1] S. Mori, M. Sakakibara, A. Tanimoto, Steady heat transfer to laminar ow in a
circular tube with conduction in tube wall, Heat Transfer Japanese Research 3
(1974) 3746.

586

M. Lin et al. / International Communications in Heat and Mass Transfer 39 (2012) 582586

[2] M. Faghri, E.M. Sparrow, Simultaneous wall and uid axial conduction in laminar
pipe-ow heat transfer, Transactions of the ASME Journal of Heat Transfer 102
(1980) 5863.
[3] B. Vick, M.N. zisik, An exact analysis of low Peclet number heat transfer in
laminar ow with axial conduction, Letters in Heat and Mass Transfer 8 (1)
(1981) 110.
[4] H.M. Soliman, Analysis of low Peclet heat transfer during slug-ow in tubes with
axial wall conduction, ASME Journal of Heat Transfer 106 (1984) 782788.
[5] G.S. Barozzi, G. Pagliarini, A method to solve conjugate heat transfer problems:
the case of fully developed laminar ow in a pipe, Transactions of the ASME
Journal of Heat Transfer 107 (1985) 7783.
[6] S. Piva, Axial wall conduction preheating effects in high Peclet number laminar
forced convection, International Journal of Heat and Mass Transfer 39 (16)
(1996) 35113517.

[7] W.M. Kays, H.C. Perkins, Handbook of Heat Transfer, Chapter 7, McGraw-Hill,
New York, 1972.
[8] B. Yu, J.H. Nie, Q.W. Wang, W.Q. Tao, Experimental study on the pressure drop
and heat transfer characteristics of tubes with internal wave-like longitudinal
ns, Heat Mass Transfer 35 (1) (1999) 6573.
[9] M. Huq, A.M.A. Huq, Experimental measurements of the heat transfer in an
internally nned tube, International Communications in Heat and Mass Transfer
25 (1998) 619630.
[10] C.P. Tso, S.P. Mahulikar, Experimental verication of the role of Brinkman number
in microchannels using local parameters, International Journal of Heat and Mass
Transfer 43 (2000) 18371849.
[11] J. Li, G.P. Peterson, P. Cheng, Three-dimensional analysis of heat transfer in a
micro-heat sink with single phase ow, International Journal of Heat and Mass
Transfer 47 (2004) 42154231.