Experimental and numerical investigation of thermal characteristics of a novel concentric type tube heat exchanger ...
[ Pobierz całość w formacie PDF ]
INTERNATIONAL JOURNAL OF ENERGY RESEARCH
Int. J. Energy Res. (2012)
Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/er.2919
Experimental and numerical investigation of thermal
characteristics of a novel concentric type tube heat
exchanger with turbulators
Hacı Mehmet Şahin
1,
*
,
â€
,EÅŸref Baysal
2
and Ali Rıza Dal
3
1
Gazi University, Faculty of Technology, Energy Systems Engineering, Teknikokullar, Ankara 06503, Turkey
2
Batman University, Vocational School, Department of Mechanical and Metal Technologies Machine Program, Batman, Turkey
3
Undersecretariat for Maritime Affairs, Tando
ÄŸ
an, Ankara, Turkey
SUMMARY
In this study, the heat transfer performance and friction characteristics of a novel concentric tube heat exchanger with
different pitches of helical turbulators were investigated experimentally and numerically for a Reynolds number range from
3000 to 14 000. An experimental system was established to obtain experimental data. The numerical simulations were
performed by using a three dimensional numerical computation technique, a commercial CFD computer code. Then, the
heat transfer performance and friction characteristics of several helical turbulators were compared. The experimental,
numerical and empirical correlation results were in a good agreement with each others. As a result, the heat transfer
enhancements using turbulators were 2.91, 2.41, 2.18 and 1.99 times better than the smooth tube for pitch distances of
p = 20, 40, 60 and 80 mm, respectively. Copyright © 2012 John Wiley & Sons, Ltd.
KEY WORDS
concentric tube heat exchangers; heat transfer enhancement; friction factor; numerical simulation; CFD
Correspondence
*Hac
ahin, Gazi University, Faculty of Technology, Energy Systems Engineering, Teknikokullar, Ankara 06503, Turkey.
â€
E-mail: mesahin@gazi.edu.tr
ı
Mehmet
Åž
Received 16 November 2011; Revised 27 February 2012; Accepted 3 March 2012
1. INTRODUCTION
turbulent or swirl
ow devices, coiled tubes and additives
for liquids and gases [5]. In this respect, several types of
turbulators, replacing a nned surface have been used in
gas to liquid heat exchangers in boiler applications [6].
The turbulators improve heat transfer efciency; however,
they cause a pressure drop, and, therefore, the system may
sometimes need fan power.
Swirl ow devices based on passive method have been
developed and are most widely applied to concentric pipe
heat exchangers as heat transfer enhancement technologies
[7
–
10]. In this case, there has been effective improvement
of heat exchangers based on extensive research via exper-
imental and numerical investigation on heat transfer
performance and pressure drop characteristics of swirl ow
applications [11
–
15]. There has been more research based
on generating swirl ow devices as found in the literature
[16
–
21]. Some series of studies have been performed by
Eiamsa-ard et al. [11,22
–
26] that their experimental inves-
tigations of heat transfer and friction characteristics in a
circular tube with turbulators insert as swirl generators
which are conical ring, conical nozzle, V nozzle, twisted
tape and screw tape.
Heat exchangers have been widely used in the industry
because of supply heat transfer between two uids that
are at different temperatures and separated by a solid wall.
Particularly, they have been used in the thermal engineering
applications, such as power stations, chemical plants, food
industries, heating-cooling systems, aircraft, automotive
sector, solar water heaters, combustion chambers, heat
recovery processes, etc. The heat transfer coefcient and
pressure drop are the most important parameters to deter-
mine in reducing the size and cost of a heat exchanger
which, in term, reduces the energy usage [1
–
3]. Heat
transfer enhancement techniques for the design of more
compact heat exchangers can be divided to two groups:
active and passive methods [4]. Active methods require
extra external power sources, such as mechanical aids,
surface and uid vibration, electrostatic elds, injection
or suction of uid and jet impingement. Passive methods,
on the other hand, perform without additional external
power, by mean of surface coating, rough surfaces,
extended surfaces, displaced enhancement devices,
Copyright © 2012 John Wiley & Sons, Ltd.
H. M.
Åž
ahin, E. Baysal and A. R. Dal
Enhancement of heat exchangers
Experimental study of heat exchangers has proved very
useful but is very expensive because of the high cost of the
tools involved. Nevertheless, some studies have been done
experimentally to determine heat transfer and friction char-
acteristics of the heat exchangers using turbulators. The
effects of geometrical parameters on heat transfer and pres-
sure drop for various types of the turbulators have been
investigated during recent decades [27
–
30].
Some studies in the literature have been made only
numerically. Chen and Dung have been studied numeri-
cally the heat transfer characteristics of the ows in parallel
and counter ow concentric tube heat exchangers with
alternating horizontal or vertical oval cross section pipes
as inner tubes [31]. Lei et al. reported a study on numerical
simulations of the impacts of various bafe inclination
angle on uid ow and heat transfer of heat exchangers
with helical bafes [32].
Compared to experimental studies, a suitable numerical
method and/or computational uid dynamics (CFD) code
requires much lower cost and also provides ow eld
information. Because of the fast development of computer
hardware, numerical simulations of heat processes as well
heat exchangers have been performed.
Although some studies related to this area have been
investigated only experimentally, some have been simu-
lated only numerically [33]. The most useful investigations
are to compare both experimental and numerical studies.
Hence, the best numerical model for suitable experimental
study can be selected. Then, more advanced studies for the
real work can be done with this selected numerical model.
Therefore, new results have been presented experimentally
and numerically in this study.
In reviewing the literature, some studies have been done
either experimentally or numerically to determine heat
transfer and friction characteristics of the concentric tube
heat exchanger. However, the effectiveness of the types
of the turbulators on heat transfer and the pressure drop
across a heat exchanger should be analyzed both numeri-
cally and experimentally. Therefore, a suitable numerical
method and/or CFD code is frequently used to solve the
governing equations of uid ow. The CFD code predicts
some characteristics of the ow speeds, pressures, resi-
dence times, ow patterns, heat transfer, etc. Thus, numer-
ical methods have been performed by several researchers
for the analysis of heat exchangers included in several
geometries and various boundary conditions [31
–
33]. The
heat transfer and ow friction behaviors in a circular tube
with the helical turbulators have rarely been reported and
studied with using the hot and cold water as the test uids
by Eiamsa-ard and Promvonge [25].
The objective of this paper is to analyze experimentally
and numerically for Re numbers ranging from 3000 to 14
000 and for a novel concentric tube heat exchanger with
different pitches (p = 20, 40, 60 and 80 mm) of helical
turbulators. Numerical analyses have been done using a
three dimensional (3-D) numerical computation technique,
and the FLUENT, a CFD computer code [34]. The heat
transfer performance and friction characteristics of helical
turbulators have been compared with experimental and
numerical results. The obtained results have been pre-
sented in dimensionless form such as the average Nusselt
N
ðÞ
number and Darcy fraction factor (f) as function of
Reynold (Re) number. In addition, new engineering corre-
lations of Nunumber and f have been presented.
2. MATERIALS AND METHODS
2.1. Experimental study
2.1.1. Experimental setup and procedure
The schematic diagram of the experimental setup
connected to the data logger and analysis system, positions
of the thermocouples, other measurement instruments, a
concentric tube heat exchanger and used tools are shown
in Figure 1. The experiments were operated at steady-state
conditions and fully developed ow. The inlet bulk air at
25
C from the laboratory room was directed through a
600-W blower adjusted ow rates by changeable motor
speed via the inverter (ABB ACS350 model) located at
the system seen in Figure 1. The KR 25/48 model blower
is manufactured in OSTIM [35] and its working pressure
range 0
–
4903.325 Pa. The air velocity in the circular pipe
was measured by a PROVA AVM-05 model calibrated
digital anemometer and its measurement accuracy
4%.
The air was heated by an electrical heater (manufactured in
OSTIM) adjusted via process controller namely digital
temperature controller unit (
0.5%measurement accuracy).
In the test section, the pressure drop of the hot air across
the test section was measured by U-type manometer with
water (
2% measurement accuracy). The inlet hot air
temperature was heated to 100
C and the cold water was
at 23
C during the experiments. The cold water in the tank
was circulated through the annulus by a pump. The ow
rate of the water was controlled by an adjusting valve
and measured by a calibrated rotameter (
3% measure-
ment accuracy) having measurement range between 25
and 250 liter/h. The thermocouples were calibrated within
0.2
C measurement accuracy and measured at steady-state
conditions. The local wall temperature of the pipe at
11 points, inlet and outlet of air and water temperatures,
the ambient air temperature and the water in the tank were
measured by K type (Ni-NiCr alloy) mineral insulated
thermocouples, connected to the Agilent Model 34970A
data logger sets and recorded by a computer to save the
temperature measurements.
Figure 2 shows a schematic diagram of the test section,
which is the concentric tube heat exchanger with helical
turbulator. The concentric tube heat exchanger obtains
two different uids and counterow. The hot air and cold
water enter in the opposite ends of the heat exchanger
and ow in opposite directions. The hot air ows in the
inner tube, while the cold water ows in the outer tube,
shown in Figure 2. The diameters of the inner tube (D)
and outer tube are 40 mm and 69 mm, respectively. The
length of tube (L) is 1 000 mm (for fully developed ow),
Int. J. Energy Res.
(2012) © 2012 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Enhancement of heat exchangers
H. M.
Åž
ahin, E. Baysal and A. R. Dal
Control
Panel
Process
Controller
Invert
er
Thermocouples
Turbulator
Insulation
Data
Logger
Electrical
Heater
Manometer
Rotameter
Computer
Anemometer
Water
Tank
Valve
Pump
Blower
Figure 1. Schematic diagram of the experimental setup.
t
p
Water out
Steel tube
Copper tube
Turbulator
Air out
Air in
Water in
L
Figure 2. Schematic diagram of the concentric tube heat exchanger with helical turbulator.
the inner tube is made of copper and its thickness is 1 mm.
The outer tube is made of steel, thickness is 3.5 mm and
insulated with berglass (k = 0.04 W/mK) to minimize
heat loss. The helical turbulator was inserted into the in-
ner tube for turbulent ow. The helical turbulator is made
of stainless steel. Four different pitches of the helical
turbulator (p = 20, 40, 60 and 80 mm) were used in the
experiments. The geometric dimensions of the helical
turbulator are W=37mm,d =15mm,t = 1 mm, as shown
in Figure 2.
Int. J. Energy Res.
(2012) © 2012 John Wiley & Sons, Ltd.
DOI: 10.1002/er
H. M.
Åž
ahin, E. Baysal and A. R. Dal
Enhancement of heat exchangers
T
w
¼
X
11
The thermo-physical properties of the uids and materi-
als used in our experiments were taken from the literature
[1]. All experiments were done for Re number for hot air
ranging from 3000 to 14 000.
T
Wi
=
11
(6)
i
¼
1
The average heat transfer coefciency
ðÞ
canbecalcu-
lated from the combination of Equations (1) to (3), as follows:
2.1.2. Experimental uncertainty
The uncertainties of the experimental results are
achieved by the method of Kline and McClintock [36]
using data obtained from the measurement accuracy. This
method is expressed by following equation:
Q
con
AT
m
T
w
h
¼
(7)
ð
Þ
Reynolds (Re) number based on hydraulic diameter
D
h
¼
4A
c
S
of the tube is as follow:
"
#
1
=
2
@
x
1
w
1
2
@
x
2
w
2
2
@
x
n
w
n
2
@
R
@
R
@
R
w
R
¼
þ
þ ...þ
uD
h
v
Re
¼
(8)
(1)
Nu number can be expressed as
The average Nusselt
where, R is a given function of the independent variablesx
1
,
x
2
,
...
, x
n
. It can be expressed as: R= R(x
1
, x
2
,
...
, x
n
). w
1
, w
2
,
, w
n
are the uncertainties in the independent variables and w
R
is the uncertainty in the result.
The obtained uncertainty ranges using measurement
accuracy for m
;
Q
;
h
;
Nu
;
Re
; Δ
P and f were estimated to be
4.25%,
4.32%,
4.44%,
4.7%,
4.69%,
5.85%
and
6.7%, respectively. In the same conditions, the
experimental results can be reproduced for these uncertainty
ranges in this study.
follow:
Nu
¼
hD
h
v
(9)
In forced convection, Nu number is function of Re
number and Prandtl (Pr) number [1
–
3]:
Nu
¼
C
1
Re
n
1
Pr
n
2
(10)
Pr number of air was considered as constant at a value
of 0.7 for temperature range during the experiments [1]
and C
2
=C
1
.Pr
0.3
, Equation (10) can be written as follow:
2.1.3. Data reduction
A concentric tube heat exchanger was selected for
experiment in the present work. Hot air in the inner side
was used as the working uid; water in the outer side
was used to absorb heat from the hot air. In order to inves-
tigate the effect of using helical turbulator for heat transfer,
the average heat transfer coefcient (h) and average Darcy
friction factor (f) of the inner side were determined. There-
fore, as in the literature [12,37
–
40] as well our study, the
steady state of the heat transfer rate by convection was
thought to be equal to the heat loss rate from test section
in the hot air side which can be written as follow:
Nu
¼
C
2
Re
n
1
(11)
The average Darcy friction factor (f) can be expressed
as follow:
Δ
P
D
h
r
u
2
2
f
¼
(12)
L
The friction factor can be a written function of Re
number [1
–
3]:
Q
a
¼
Q
con
(2)
where
f
¼
C
3
Re
n
3
(13)
Q
a
¼
mc
p
;
a
T
o
T
i
ð
Þ
(3)
In Equations (10), (11) and (13), C1, C2 and C3 are
constant coef
and
cients; n1, n2 and n3 are power indices.
Q
con
¼
hA T
m
T
w
ð
Þ
(4)
2.2. Numerical model and method
The mean temperature of air can be determined as follow:
2.2.1. Mathematical modeling
The governing equations for continuity, momentum
and energy in the computational procedure can be written
as follows.
Continuity equation:
T
m
¼
T
o
þ
T
i
ð
Þ=
2
(5)
The average wall temperature T
ðÞ
was calculated from
local wall temperatures T
ðÞ
at 11 points, located linearly
at the outer wall surface of the inner tube between the inlet
and the outlet of the tube. It can be described as follow:
@
@
x
i
r
u
i
¼
0
ð
Þ
(14)
Int. J. Energy Res.
(2012) © 2012 John Wiley & Sons, Ltd.
DOI: 10.1002/er
 Enhancement of heat exchangers
H. M.
Åž
ahin, E. Baysal and A. R. Dal
Momentum equation:
velocity coupling algorithm and a second-order upwind
numerical scheme were utilized to discretize the governing
equations. The continuity, momentum and energy equations
for the heat transfer and pressure drop were iteratively solved
by the nite volume method for the thermal and uid dynam-
ics analyses, a commercial CFD code, ANSYS FLUENT
12.0 [34] in the test section of the experimental setup. Fluent
code provides mesh exibility by structured and unstruc-
tured meshes. Fluent has many turbulence models: two types
of k-
o
turbulence model (standard k-
o
and shear-stress
transport SST k-
o
) have been used for transition region
and three types of k-
e
models (standard k-
e
, renormaliza-
tion-group k-
e
and realizable k-
e
) have been used for fully
developed turbulence ow. The realizable k-
e
model is a rel-
atively recent development and differs from others k-
e
mod-
els. It contains a new formulation for the turbulent viscosity
and a new transport equation for the dissipation rate (
e
)[41].
Moreover, the realizable k-
e
model provides the best perfor-
mance of all the k-
e
@
p
@
x
j
¼
@
@
@
x
i
r
u
i
u
j
@
x
i
m
@
u
j
(15)
@
x
i
Energy equation:
(16)
@
@
x
i
r
u
i
T
Þ¼
@
@
x
i
c
p
@
T
k
ð
@
x
i
Where,
r
is density, u velocity,
m
dynamic viscosity,
p pressure, k thermal conductivity, T temperature and
c
p
specic heat.
2.2.2. Boundary conditions
In this study, air and water were used as the working
uids. Below are boundary conditions applied to solve
Equations (14
–
16) in the test section, shown in Figure 2:
model version for several validations of
Inlet boundary conditions: the temperature and velocity
of air and water at the inlet of the test section were taken
from experiments.
Outlet boundary conditions: air and water leaving
from the test section to atmospheric conditions.
separated
ow fea-
tures [34]. In this study, full turbulence ow was assumed in
case of Re
>
3000 and using turbulators [1]. Therefore, a re-
alizable based k-
e
turbulence model was used to predict the
heat transfer and uid ow characteristics. As a result, the
numerical simulations were performed with 3-D geometry,
steady-state, and turbulent ow system by FLUENT code.
To assure the accuracy of the results, a grid indepen-
dence study was done using ve different grid sizes for
Re number 8000 in all of the calculation, shown in Figure 4.
This gure shows that the heat transfer rate did not change
after a residual value of 10
3
. In the computational domains
in Figure 3, ranges of cell numbers of 850 000
900 000
ows and
ows with complex secondary
The material of turbulators and outer tube were stainless
steel, and inner tube was copper.
In addition, some conditions were assumed as follow:
The ow was steady state and turbulent.
The working uids were incompressible.
The outer tube wall was assumed to be adiabatic.
The thermo-physical properties of air and water were
considered constant.
The thermo-physical properties of the materials were
taken as constant.
2.2.3. Computational domain and calculation
procedure
Figure 3 shows the computational domain as 3-D geom-
etry for the studied model of the concentric tube heat
exchanger with helical turbulator. The computational
domain was meshed via unstructured Tet/Hybrid grids
with GAMBIT code [34]. The model drawing was created
by using SOLIDWORS due to its excellent merit of man-
aging very complex 3-D geometries and then transferred
to GAMBIT software to mesh. A SIMPLE pressure
–
Figure 3. Schematic of the computational domain.
Figure 4. The heat transfer rate versus cell number.
Int. J. Energy Res.
(2012) © 2012 John Wiley & Sons, Ltd.
DOI: 10.1002/er
[ Pobierz całość w formacie PDF ]