International Journal of Heat and Mass Transfer 52 (2009) 5371–5380

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

3D numerical simulation on shell-and-tube heat exchangers with

middle-overlapped helical baffles and continuous baffles – Part I: Numerical model
and results of whole heat exchanger with middle-overlapped helical baffles
Jian-Fei Zhang, Ya-Ling He, Wen-Quan Tao *
School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an 710049, China

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

Article history: In the present study, a 3D numerical simulation of a whole heat exchanger with middle-overlapped heli-
Received 17 March 2009 cal baffles is carried out by using commercial codes of GAMBIT 2.3 and FLEUNT 6.3. At first, the compu-
Received in revised form 9 July 2009 tational model and numerical method of the whole heat exchanger with middle-overlapped helical
Accepted 9 July 2009
baffles is presented in detail, and parallel computation mode is adopted for the simulation of a whole heat
Available online 21 August 2009
exchanger with six cycles of the middle-overlapped helical baffles of 40° helical angle on a grid system of
13.5-million cells; second, the validation of the computational model is performed by comparing the total
Keywords:
pressure drop and average Nusselt number of the whole heat exchanger with experimental data. Reason-
Heat exchangers
Helical baffles
ably good agreement is obtained, and the reasons causing to the discrepancy are analyzed. The shell-side
Numerical simulation fluid pressure and temperature fields of the whole area are then presented. Finally the cycle average Nus-
Turbulence selt number of different cycle in the heat exchanger are compared and it is found that within the accuracy
Pressure drop allowed in engineering computation, periodic model for one cycle can be used to investigate the heat
Heat transfer transfer and pressure drop characteristics for different heat exchanger to save computational source.
The companion paper will provide details of simulation results using the periodic model for different
helical angles and different structures of helical baffles.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction third the dramatic zigzag flow pattern also causes high risk of
vibration failure on tube bundle.
Shell-and-tube heat exchangers (STHXs) are widely used in To overcome the above-mentioned drawbacks of the conven-
many industrial areas, such as power plant, chemical engineering, tional segmental baffle, a number of improved structures were pro-
petroleum refining, food processing, etc. According to Master et al. posed for the purposes of higher heat transfer coefficient, low
[1] more than 35–40% of heat exchangers are of the shell-and-tube possibility of tube vibration, and reduced fouling factor with a mild
type due to their robust geometry construction, easy maintenance increase in pumping power [5–10]. However, the principal short-
and possible upgrades. Baffle is an important shell-side component comings of the conventional segmental baffle still remain in the
of STHXs. Besides supporting the tube bundles, the baffles form above-mentioned studies, even though the pressure drop across
flow passage for the shell-side fluid in conjunction with the shell. the heat exchangers has been reduced to some extent. A new type
The most commonly used baffle is the segmental baffle, which of baffle, called helical baffle, provides further improvement. This
forces the shell-side fluid going through in a zigzag manner, hence type of baffle was first proposed by Lutcha and Nemcansky [11],
improves the heat transfer with a large pressure drop penalty. This where they investigated the flow field patterns produced by such
type of heat exchanger has been well-developed [2–5] and proba- helical baffle geometry with different helix angles. They found that
bly is still the most commonly used type of the shell-and-tube heat these flow patterns were much close to plug flow condition, which
exchanger. The major drawbacks of the conventional shell-and- was expected to reduce shell-side pressure drop and to improve
tube heat exchangers with segmental baffles (STHXsSB) are heat transfer performance. Stehlik et al. [12] compared heat trans-
threefold: first it causes a large shell-side pressure drop; second fer and pressure drop correction factors for a heat exchanger with
it results in a dead zone in each compartment between two adja- an optimized segmental baffle based on the Bell–Delaware method
cent segmental baffles, leading to an increase of fouling resistance; [2–4] with those for a heat exchanger with helical baffles. Kral et al.
[13] discussed the performance of heat exchangers with helical
baffles (STHXsHB) based on test results of various baffles geome-
tries. A comparison between the test data of shell-side heat trans-
* Corresponding author.
E-mail address: [email protected] (W.-Q. Tao). fer coefficient versus shell-side pressure drop was provided for five

0017-9310/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2009.07.006
5372 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380

Nomenclature

Latin symbols S cross-flow area at the shell centerline, mm2

Ao heat exchange area based on the outer diameter of tube, Dtm logarithmic mean temperature difference, K
mm2 t temperature, K
B baffle spacing for segmental baffles or helical pitch for tp tube pitch, mm
helical baffles, mm u fluid velocity in the shell side, m s1
cp specific heat, J/(kg K)
Di inside diameter of shell, mm Greek symbols
Do outside diameter of shell, mm b helix angle
D1 tube bundle-circumscribed circle diameter, mm e turbulent energy dissipation
di tube inner diameter, mm U heat exchange quantity, W
do outer diameter of tube, mm k thermal conductivity, W (m K) 1
Eij mean rate-of-strain tensor l dynamic viscosity of fluid, Pa s
h heat transfer coefficient, W (m2 K)1 m kinematic viscosity of fluid, m2 s1
k turbulent kinetic energy q density of fluid, kg m3
l effective length of tube, mm
M mass flow rate, kg/s Subscripts
N tube number in inlet
Nt number of tube rows out outlet
Nu Nu number s shell side
Dp shell-side pressure drop, kPa t tube side
qs volume flow rate, m3 h1 w wall
Re Re number

helical baffles and one segmental baffle measured from a water-

water heat exchanger. The case of 40° helix angle behaved the best.
For the convenience of manufacturing, up to now all helical baffles
actually used in STHXs are noncontinuous approximate helicoids.
The noncontinuous helical baffles are usually made by four ellipti-
cal sector-shaped plates joined end to end. Elliptical sector-shaped
plates are arranged in a pseudohelical (noncontinuous) manner, as
shown in Fig. 1, where the helix angle, designated by b, and helical
pitch, B, are presented. As shown in Fig. 1 the helix angle is referred
to the angle between normal line of circular sector-shaped plates
and the axis of heat exchanger. Each baffle occupies one-quarter
of the cross section of the heat exchanger and is angled to the axis
of the heat exchanger. The two adjacent baffles may touch at the
perimeter of each sector, forming a continuous helix at the outer
periphery (Fig. 1); and this structure of connecting baffles together
is called continuous connection. Another connection between two
adjacent sectors is the middle-overlapped connection as shown in
Fig. 2. For heat exchangers with large shell diameters, such struc-
ture can reduce the helical pitch to shorten the length of heat ex-
changer, and can also reduce the cross-flow area to obtain a
higher shell-side velocity. Hence such connection is more popular
in engineering practice. In this paper helical baffle refers to such
noncontinuous baffle, otherwise special description will be given.
Above papers were mostly published in the last century. Typical
publications on this subject since the year of 2000 can be referred
to [14–18]. Experimental study on heat transfer enhancement of a
helically baffled heat exchanger combined with three-dimensional
finned tubes was conducted by Zhang et al. [15]. Peng et al. [16]
carried out an experimental study of shell-and-tube heat exchang-
ers with continuous helical baffles and they found that that the use
of continuous helical baffles resulted in nearly 10% increase in heat
transfer coefficient compared to that of conventional segmental
baffles based on the same shell-side pressure drop. Lei et al. [17]
presented some experimental results of pressure drop and heat
transfer of a heat exchanger with helical baffles. A comprehensive
comparison of heat transfer coefficient per unit pressure drop and
unit pumping power at the same shell-side flow rate was provided
for four helical baffle heat exchangers and one segmental baffle
heat exchanger by Zhang et al. [18]. Fig. 1. Schematics of parameters definition.
 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380 5373

Fig. 2. Four pieces middle-overlapped helical baffle arrangement.

Another important research method for the improvement of Recently, the rapid development of CFD commercial code and
heat exchangers is the numerical simulation. Although experi- computer hardware helps the direct 3D numerical simulation of
ment study can provide reliable test data for designers and complex flow phenomenon in STHXs and it is becoming more
researchers, it is very expensive and time consuming. Compared and more convenient and popular. Schröder and Gelbe [29] applied
to experiment, a validated CFD method can provide more infor- two- and three-dimensional simulation models for the computa-
mation in heat exchangers at much lower cost. In the following a tion of flow-induced vibration of tube bundles. Mohr and Gelbe
brief review on the numerical simulation study is presented. In [30] presented a method to produce equivalent velocity distribu-
1974, Patankar and Spalding [19] simplified a three-dimensional tions and corresponding cross-sectional areas in tube bundles of
STHX to a kind of porous media model and introduced the con- STHXs, which enables the designer to predict the vibration excita-
cept of distribution resistance to simulate the shell-side flow in tion more accurately than before. Philpott and Deans [31] investi-
STHX. In 1978, Butterworth [20] developed a three-dimensional gated the effects of the addition of ammonia on the enhancement
model for heat transfer in tube bundles. From 1980 to 1982, por- of steam condensation heat transfer in a horizontal shell-and-tube
ous media model and distribution resistance concept was im- condenser. Karlsson and Vamling [32] carried out 2D CFD calcula-
proved by Sha [21] and Sha et al. [22], and the concept of tions for vapor flow field and rate of condensation for a pure refrig-
surface permeabilities was introduced to account for the anisot- erant and a binary mixture in a shell-and-tube condenser. Lee and
ropy of tube bundles porosities. Prithiviraj and Andrews [23–25] Hur [33] investigated the heat transfer and flow in the shell side of
developed a three-dimensional CFD method (named as HEATX) a STHXSB, and the effects of locations and sizes of the sealing strips
based on the distributed resistance concept along with volumet- on heat transfer and flow were also investigated in their study.
ric porosities and surface permeabilities to simulate flow and Apart from the numerical study on STHXsSB mentioned above, sev-
heat transfer in STHXsSB. Their research results were compared eral numerical studies on STHXsHB were also conducted by some
with experiment data and good agreement between the simula- researchers. Shen et al. [34] established a mathematical model of
tion results and experimental data was obtained. In 2003, Deng the flow and heat transfer of the helical baffles heat exchanger to
[26] investigated the flow and heat transfer in shell side of simulate the influence of helical baffles on heat transfer and flow
STHXsSB and flow in shell side of STHXsHB based on the method characteristics of STHXsHB. And the numerical simulation results
suggested by Prithiviraj and Andrews [23–25], and the accuracy at the 35° helix inclination angle were compared with experimen-
of numerical model was validated by experimental data. In 2005, tal data. Lei et al. [35] investigated the effects of helix angle on the
Andrews and Master [27,28] employed HEATX to investigate the flow and heat transfer characteristics of STHXs with continuous
performance of a STHXHB. Their computed pressure drops com- baffles by using simplified periodical model. Lei et al. [17] also de-
pared reasonably well with ABB Lummus Heat Transfer Co. pres- signed a heat exchanger with two-layer helical baffles and com-
sure drop correlation results. The numerical model with porous pared its performance with heat exchanger with single-segment
medium concept mentioned above could reduce the requirement baffles and single-helical baffles by using simplified periodical
to computer capability greatly, but there are still some defaults numerical model. Jafari Nasr and Shafeghat [36] studied the veloc-
in it: (1) many additional parameters, such as volumetric poros- ity distribution in STHXsHB at different helix angles and developed
ities and surface permeabilities on shell side of heat exchanger a rapid algorithm for STHXsHB design.
must be calculated exactly. Because the shell-side configuration In most of the above referenced papers for the performance
of STHXs is very much complicated, so it is a big challenge for simulation of STHXsHB a periodic model is usually adopted for
researchers and designers to obtain such geometric parameters which the fluid flow and heat transfer are assumed to be fully
accurately; (2) the shell-side distributed resistances and heat developed. For the external fluid flow and heat transfer it is well
transfer coefficient must be provided from existing experimental accepted [37–39] that for the flow and heat transfer in a geometric
correlation formulas, and the accuracy of numerical results is af- periodic structure it needs about 4–6 cycles for the flow and heat
fected by such correlations to a great extent; (3) detailed and ex- transfer to be fully developed. For many STHXsHB in practical
act characteristics of flow and heat transfer on shell side cannot usage, the cycle numbers are not very far from above-mentioned
be obtained due to the simplified principle of porous media lower limitation. Thus only periodical model is not enough to
model. Among the three drawbacks, the second one is probably understand the full fluid flow and heat transfer process in the
the most fatal. STHXsHB. At least we have to reveal how many cycles (the space
5374 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380

within one helical pitch composes one cycle) are needed in a 2. Model for whole heat exchanger simulation
STHXHB for the shell-side flow and heat transfer to be fully devel-
oped. In addition, as indicated above the STHXsHB in practical 2.1. Computational model
application are all of noncontinuous type, and the experimental
comparative study between continuous and noncontinuous baffles The computational model of an experimental tested STHXHB
has not been reported in the literatures, probably because the with 40° helix angle [18] is shown in Fig. 3, and the geometry
expensive test cost. Thus numerical comparison will be very useful parameters are listed in Table 1. As can be seen from Fig. 3,
to proceed such an investigation. the simulated STHXHB has six cycles of baffles in the shell-side
In the present paper, a 3D simulation model on flow and heat direction with total tube number of 37. The whole computa-
transfer in the shell side of a whole STHXHB with middle-over- tional domain is bounded by the inner side of the shell and
lapped baffles at 40° helix angle will be established in detail by every thing in the shell is contained in the domain. The inlet
using the commercial software of FLUENT with grid systems being and outlet of the domain are connected with the corresponding
generated by GAMBIT. Numerical simulation for the whole heat ex- tubes.
changer with 40° helix angle will first be conducted at a series of To simplify numerical simulation while still keep the basic char-
shell-side flow rate for validating the numerical model. The simu- acteristics of the process, following assumptions are made: (1) the
lated results will be compared with some experiment data avail- shell-side fluid is of constant thermal properties; (2) the fluid flow
able to the present authors. Then numerical results of the whole and heat transfer processes are turbulent and in steady-state; (3)
heat exchanger simulation will be presented in details, including the leak flows between tube and baffle and that between baffle
the streamwise variation of the cycle average Nusselt number, and the shell are neglected; (4) the natural convection induced
the shell-side fluid pressure drop and temperature distributions. by the fluid density variation is neglected; (5) the tube wall tem-
It will be shown that after 5 or 6 cycles the flow and heat transfer peratures are kept constant in the whole shell side; (6) the heat ex-
in a STHXHB can be regarded as fully developed, validating the sig- changer is well-insulated hence the heat loss to the environment is
nificance of periodic model in the numerical simulation of totally neglected.
STHXsHB. In the companion paper the periodic model will be
adopted for the simulation. Investigation will be conducted for 2.2. Governing equations and boundary conditions
three periodic models with middle-overlapped baffles at different
helix angles to reveal the fully developed performance, and the ef- The renormalization group (RNG) k–e model [50–52] is adopted
fects of helix angle on the performance of STHXsHB are examined because it can provide improved predictions of near-wall flows and
and analyzed by the field synergy principle recently developed by flows with high streamline curvature [50,51]. The governing equa-
Guo et al. [40–42] and later enhanced in [43–49]. The performance tions for the mass, momentum, and energy conservations, and for k
of a periodic model with continuous helical baffle will also be and e can be expressed as follows:
established and simulations will be conducted. The numerical re- Mass:
sults will be analyzed from different aspects: the effects of the he-
@
lix angle, comparison between the results of continuous helical ðqui Þ ¼ 0 ð1Þ
@xi
baffle with noncontinuous middle-overlapped helical baffles for
the shell-side heat transfer and pressure drop at the same flow Momentum:
rate, and the comparison between the two types of helical baffles  
@ @ @uk @p
for the shell-side heat transfer coefficient based on the unit pres- ðqui uk Þ ¼ l  ð2Þ
@xi @xi @xi @xk
sure drop.

Fig. 3. Shaded partial scenograph of the computational model.

 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380 5375

Table 1 Table 2
Geometry parameters for the whole model. Thermophysical properties of oil.

Item Dimensions and description Parameter Value

Shell-side parameters cp (J/kg K) 2270.1
Do/Di (mm) 223/211 l (kg/ms) 0.0095
Material 0Cr18Ni9 q (kg/m3) 826.1
k (W/m K) 0.132
Tube parameters
do/di (mm) 19/15
Effective length (mm) 1703
Number 37
Layout pattern 45°
tured tetrahedral elements and the region adjacent to the tubes
Tube pitch (mm) 25 is meshed much finer to meet the requirement of wall function
Material 0Cr18Ni9 method. The meshes of the computational model are shown in
Baffle parameters Fig. 4. Grid independence tests are carried out to ensure that a
Baffle pitch (mm) 250 nearly grid independent solution can be obtained. In the test, three
Helix angle 40° different grid systems with 9840678, 13553984 and 17759679
Thickness (mm) 3
cells are adopted for calculation of the whole heat exchanger,
Number 24
and the difference in the overall pressure drop and the average
heat transfer coefficient between last two grid systems are around
2%. Thus, considering both the computational time cost and
Energy:
 
@ @ k @t
ðqui tÞ ¼ ð3Þ
@xi @xi C P @xi
Turbulent kinetic energy:
 
@ @ @ @k
ðqkÞ þ ðqkui Þ ¼ ak leff þ Gk þ qe ð4Þ
@t @xi @xj @xj
Turbulent energy dissipation:
 
@ @ @ @e e e2
ðqeÞ þ ðqeui Þ ¼ ae leff þ C 1e Gk  C 2e q ð5Þ
@t @xi @xj @xj k k
where
2
k gð1  g=go Þ
leff ¼ l þ lt ; lt ¼ qcl ; C 1e ¼ C 1e  ;
e 1 þ bg3
 
k 1 @ui @uj
g ¼ ð2Eij  Eij Þ1=2 ; Eij ¼ þ :
e 2 @xj @xi
The empirical constants for the RNG k–e model are assigned as
following [51]:
Cl = 0.0845, C1e = 1.42, C2e = 1.68, b = 0.012, go = 4.38, aj =
ae = 1.39.
Now boundary conditions are presented. Non-slip boundary
condition is applied on the inner wall of the shell and all solid sur-
faces within the computational domain. The standard wall func-
tion method is used to simulate the flow in the near-wall region.
The mass-flow-inlet and outflow boundary condition [51] are ap-
plied on the inlet and outlet sections, respectively. To the authors’
knowledge, such treatments of the inlet and out let conditions are
corresponding to the average velocity distribution at the inlet and
the fully developed condition [52] at the outlet. The temperature of
tube walls are set as constant and their values are taken from the
average wall temperature determined in the experiments [18]. The
shell wall of heat exchanger is set as adiabatic. Heat conduction of
baffles in heat exchanger is considered by using shell conduction in
thin-walls model [51] in FLUENT. The baffles are made from stain-
less steel and its thermal conductivity is taken as constant (15.2 W/
(m K)). The conductive-320 oil is taken as working fluid for shell
side of heat exchanger in simulation and thermophysical proper-
ties of the fluid are listed in Table 2 [18].

2.3. Grid generation and numerical method

The 3D grid system was established using the commercial code

GAMBIT based on the 3D geometry created in a commercial CAD
program. The computational domain is discretized with unstruc- Fig. 4. Meshes of computational model.
5376 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380

solution precisions the second grid system is taken for the whole 3. Computational model validation and result analysis
computation.
The commercial code FLUENT is adopted to simulate the flow 3.1. Model validation
and heat transfer in the computational model. The governing equa-
tions are discretized by the finite volume method [52,53]. The In order to validate above simulation model simulations are
QUICK scheme is used to discretize the convective terms. The SIM- first conducted for the STHXHB described in Table 1. Figs. 5 and
PLE algorithm is adopted to deal with the coupling between veloc- 6 provide the comparisons between experimental and numerical
ity and pressure. The convergence criterion is that the mass results for total shell-side pressure drop and average Nu number,
residual should be less than 106 for the flow field and the energy respectively. The experimental data are taken from [18]. It can be
residual less than 108 for the energy equation. A parallel compu- observed that for both fluid pressure drop and heat transfer their
tation is performed on four DELL workstations with two Quad-Core variation trends with mass flow are in good agreement with the
CPUs and 4 GB memory each by using FLUENT and every simula- test data. Quantitatively, the maximum differences between
tion case takes approximately 72 h to get converged solutions. numerical results and experimental data are around 25% for pres-
sure drop and 15% for Nu number. It is noticed that the pressure
2.4. Data reduction drop of the test data are lower than those of numerical predictions
and the Nusselt number of the test data are higher than those of
2.4.1. Determination of shell-side velocity and Re number numerical predictions. Apart from some un-avoidable measure-
The shell-side fluid mean velocity is defined by ment errors, such discrepancies between experimental data and
numerical results may be caused by the major simplification made
qs in the computational model: the effect of leakage flow was not ta-
u¼ ð6Þ
S
where S is the cross-flow area at the shell centerline [2–4,12,54].
For the noncontinuous helical baffles [12]:
 
D1  do
S ¼ 0:5B Di  D1 þ ðtp  do Þ ð7Þ
tp

where Di is the inside diameter of shell, D1 is the diameter of the

tube bundle-circumscribed circle, do is the tube outside diameter,
and tp is the tube pitch.
For the middle-overlapped helical baffle B is determined as
follows:
pffiffiffi
B¼ 2Di  tan b ð8Þ

With the mean velocity at hand, the Reynolds number of shell-side

fluid can be calculated:

udo
Res ¼ ð9Þ
ms
It should be emphasized that for the STHXsHB because the
shell-side flow pattern resulted from the helical-type structure is
Fig. 5. Comparison of overall pressure drop between experimental results and
close to helical flow, the cross section area is actually only half of simulation results in shell side.
the entire cross section at the shell centerline of the heat
exchanger.

2.4.2. Shell-side heat transfer coefficient and Nu number

Heat exchange rate of shell-side fluid:

Us ¼ Ms  cps  ðt s;in  t s;out Þ ð10Þ

The shell-side heat transfer coefficient hs is equal to [55]

Us
hs ¼ ð11Þ
A o  Dt m
Ao ¼ Nt  pdo l ð12Þ
Dt max  Dt min
Dt m ¼ ð13Þ
lnðDtmax =Dt min Þ
Dtmax ¼ t s;in  tw ð14Þ
Dtmin ¼ t s;out  t w ð15Þ
hs do
Nus ¼ ð16Þ
ks

where Ao is the heat exchange area based on the outer diameter of

tube; tw is the temperature of tube walls; the subscripts s and t refer Fig. 6. Comparison of Nu number between experimental results and simulation
to shell side and tube side, respectively. results in shell side.
 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380 5377

ken into account in the simulation. It can be easily understood that average Nusselt number presented later will further support this
this simplification will lead to a higher pressure drop, for the same observation.
flow rate the case without leakage and the case with leakage have
different actual flow areas. Obviously at the same nominal flow
area the one with flow leakage has larger actual flow area than 3.2. Pressure and temperature variations
the one without leakage. Thus the pressure drop of the former is
expected to be lower than the later. As far as the heat transfer is In order to observe of the pressure and temperature evolution
concerned, the case with leakage also has a larger heat transfer processes of the shell-side fluid some geometric specifications
area than the case without leakage for which the only heat transfer are made in Fig. 8 to identify each cycle. As can be seen there,
surface is the tube-banks. Thus the total heat transfer rate of the the four cycles located in the center part of the heat exchangers
case with leakage can be expected larger than that without flow are identified as 2–5, respectively. Around the four cycles there
leakage. are five cross sections which are designated as (a)–(e), respectively.
The flow path lines in the shell side of the heat exchanger are In Fig. 9 the shell-side pressure variations are presented for the
shown in Fig. 7. It can be clearly observed that except the inlet above specified geometric units. It should be noted that there are
end region the fluid passes though the tube bundles basically in several tube-like space units going through the five cross sections.
a helical pattern. By careful examining different flow path lines These are the shell-side fluid space in the central longitudinal sec-
shown by different colors, it can be observed that in the first four tion (x–z plane at y = 0 and y–z plane at x = 0). The hollow tube-like
cycles some path lines move forward in a more or less zigzag man- space units are actually the location of tubes. It can be seen that in
ner. Only in the fifth and sixth cycles most of the fluid path lines this visualized space the fluid pressure is the highest at the upper
have become quite smooth. This observation may be regarded as part of cross section (a) and the lowest at the lower part of cross
an indication of the development process of the periodically fully section (e). Totally speaking fluid pressure decreases gradually
developed flow of the shell-side, and is quite consistent with the from cross section (a) to cross section (e). Such variation trend is
previous measurement results [37–39]. The variation of the cycle very understandable.

Fig. 7. Path lines in shell side of heat exchanger (Ms = 3.91 kg/s).

Fig. 8. Specifications of specified surfaces and geometric cycle units.

5378 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380

The variation of shell-side fluid temperature is shown in Fig. 10 Table 3

with the same visualized method. The streamwise decrease of fluid Comparison of pressure drop and Nu number (Ms = 3.54 kg/s).

temperature can be clearly observed. In this computation the tube Item Pressure drop (Pa) Nu number
wall temperature is taken as 289.4 K, and the inlet oil temperature Cycle 2 in whole model 138.01 44.19
is 318.5 K. It can be seen that in the most part of the helical baffles Cycle 3 in whole model 130.97 43.71
the solid temperatures are basically dominated by the wall tem- Cycle 4 in whole model 130.95 43.51
perature via the conduction. While in the periphery region of each Cycle 5 in whole model 135.58 43.36

helical baffle, where the tube wall conduction effect is weakened

the solid temperatures become higher.
The same comparison results can be obtained for the fluid pressure
3.3. Variation of the cycle average Nusselt number with cycle drop. Therefore from fifth cycle the flow and heat transfer in STHXHB
can be regarded as periodically fully developed. As an engineering
A comparison of pressure drop and Nu number between cycles 2 computation, 1–2% discrepancy can be accepted. Thus from the pres-
and 5 in whole model (see Fig. 8) at a fixed mass flow rate are pre- ent simulation the periodic model for one cycle can be accepted for
sented in Table 3. Two basic features can be observed. First, as in the performance simulation of a STHXHB within the accuracy al-
the external flow passing through geometrically periodic structure lowed in engineering computation to greatly save the computer
[37–39], the cycle average Nusselt number is the largest in the inlet source.
cycle of the simulated STHXHB, and then its values gradually de-
ceases in the streamwise direction, even though the variation is
not as significant as that in [37–39]. Second as far as the absolute va- 4. Conclusion
lue is concerned, the Nusselt number of the fifth cycle differs from
that of the second cycle by less than 2.0 %. And the difference be- In this paper, a comprehensive simulation model for a whole
tween the fourth cycle and the fifth cycle is even less than 0.5%. STHXHB with middle-overlapped baffles is developed by using

Fig. 9. Pressure distribution in specified surfaces of shell side without inlet and outlet sections (faces (a)–(e) in Fig. 8, x–z longitudinal section at y = 0 and y–z longitudinal
section at x = 0, Ms = 3.91 kg/s).

Fig. 10. Temperature distribution in specific surfaces of shell side without inlet and outlet sections (faces (a)–(e) in Fig. 8, x–z longitudinal section at y = 0 and y–z longitudinal
section at x = 0, Ms = 3.91 kg/s).
 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380 5379

commercial code FLUENT and the grid generation program GAM- Schlunder (Eds.), Heat Exchanger Design and Theory Source Book, McGraw-
Hill, New York, 1974.
BIT. Validation for the helical angle of 40° helix angle is performed
[20] D. Butterworth, A model for heat transfer during three-dimensional flow in
and the numerical results show a reasonable agreement with the tube bundles, in: 6th International Heat Transfer Conference, Toronto, August
available experiment data. 1978.
Simulation of the shell-side fluid flow and heat transfer for the [21] W.T. Sha, An overview on rod-bundle thermal-hydraulic analysis, Nucl. Eng.
Des. 62 (1980) 1–24.
whole heat exchanger is conducted. Comparison between cycle [22] W.T. Sha, C.I. Yang, T.T. Kao, et al., Multi-dimensional numerical modeling of
average Nusselt number and the pressure drop shows that the heat exchangers, ASME J. Heat Transfer 104 (1982) 417–425.
absolute value of the cycle average Nusselt number decreases from [23] M. Prithiviraj, M.J. Andrews, Three-dimensional numerical simulation of shell-
and-tube heat exchanger, Part I: foundation and fluid mechanics, Numer. Heat
the inlet to outlet, so does the pressure drop. However, the relative Transfer A Appl. 33 (1998) 799–816.
differences are quite mild and for the case studied the difference [24] M. Prithiviraj, M.J. Andrews, Three-dimensional numerical simulation of shell-
between the 2nd cycle and the fifth cycles are both less than 2% and-tube heat exchanger, Part II: heat transfer, Numer. Heat Transfer A Appl.
33 (1998) 817–828.
for both pressure drop and heat transfer. Thus for the performance [25] M. Prithiviraj, M.J. Andrews, Comparison of a three-dimensional numerical
simulation of a STHXHB periodic model for one cycle can be used model with existing methods for prediction of flow in shell-and-tube heat
to investigate its performance without inducing large error. exchangers, Heat Transfer Eng. 20 (2) (1999) 15–19.
[26] Bin Deng, Experimental and numerical study of flow and heat transfer in the
In the companion paper simulations by using periodic models shell side of heat exchangers, Ph.D. dissertation, Xi’an Jiaotong University,
will be conducted and parameter effects will be investigated. Xi’an, China, 2003.
[27] M.J. Andrews, B.I. Master, 3-D modeling of the ABB Lummus Heat Transfer
Helixchanger using CFD, in: International Conference Compact Heat
Acknowledgments Exchangers, Banff, Canada, July 19–23, 1999.
[28] M.J. Andrews, B.I. Master, Three-dimensional modeling of a HelixchangerÒ
heat exchanger using CFD, Heat Transfer Eng. 26 (2005) 22–31.
This work is supported by the National Fundamental Research [29] K. Schröder, H. Gelbe, Two- and three-dimensional CFD-simulation of flow-
Program of China (973 Program) (2007CB206902), the Key Project induced vibration excitation in tube bundles, Chem. Eng. Process. 38 (1999)
621–629.
of Chinese Ministry of Education (No. 306014) and the Key Project
[30] U. Mohr, H. Gelbe, Velocity distribution and vibration excitation in tube bundle
of National Natural Science Foundation of China (No. 50636050). heat exchangers, Int. J. Thermal Sci. 39 (2000) 414–421.
[31] C. Philpott, J. Deans, The enhancement of steam condensation heat transfer in a
horizontal shell and tube condenser by addition of ammonia, Int. J. Heat Mass
References Transfer 47 (2004) 3683–3693.
[32] T. Karlsson, L. Vamling, Flow fields in shell-and-tube condensers: comparison
of a pure refrigerant and a binary mixture, Int. J. Refrigeration 28 (2005) 706–
[1] B.I. Master, K.S. Chunangad, V. Pushpanathan, Fouling mitigation using
713.
helixchanger heat exchangers, in: Proceedings of the ECI Conference on Heat
[33] S.H. Lee, N. Hur, Numerical analysis of the fluid flow and heat transfer in a shell
Exchanger Fouling and Cleaning: Fundamentals and Applications, Santa Fe,
and tube heat exchanger, in: Proceedings of 1st Asian Symposium on
NM, May 18–22 2003, pp. 317–322.
Computational Heat Transfer and Fluid Flow, Xi’an, China, October 18–21
[2] K.J. Bell, Delaware method for shell side design, in: S. Kakac, A.E. Bergles, F.
2007.
Mayinger (Eds.), Heat Exchangers-Thermal-Hydraulic Fundamentals and
[34] R.J. Shen, X. Feng, X.D. Gao, Mathematical model and numerical simulation
Design, Taylor & Francis, Washington, DC, 1981.
of helical baffles heat exchanger, J. Enhanced Heat Transfer 11 (2004) 461–
[3] K.J. Bell, Delaware method of shell side design, in: J.W. Pallen (Ed.), Heat
466.
Exchanger Sourcebook, Hemisphere, New York, 1986.
[35] Yong-Gang Lei, Ya-Ling He, Rui Li, Ya-Fu Gao, Effects of baffle inclination angle
[4] K.J. Bell, Delaware method of shell-side design, in: R.K. Shah, E.C. Sunnarao,
on flow and heat transfer of a heat exchanger with helical baffles, Chem. Eng.
R.A. Mashelkar (Eds.), Heat Transfer Equipment Design, Taylor & Francis,
Process. Process Intensification 47 (12) (2008) 2336–2345.
New York, 1988.
[36] M.R. Jafari Nasr, A. Shafeghat, Fluid flow analysis and extension of rapid design
[5] K.J. Bell, Heat exchanger design for the process industries, ASME J. Heat
algorithm for helical baffle heat exchangers, Appl. Thermal Eng. 28 (2008)
Transfer 126 (6) (2004) 877–885.
1324–1332.
[6] R. Mukherjee, Use double-segmental baffles in the shell-and-tube heat
[37] S.S. Lue, H.Z. Huang, W.Q. Tao, Experimental study on heat transfer and
exchangers, Chem. Eng. Prog. 88 (1992) 47–52.
pressure drop characteristics in the developing region for arrays of
[7] M. Saffar-Avval, E. Damangir, A general correlation for determining optimum
obliquely positioned plates of nonuniform length, Exp. Thermal Fluid Sci.
baffle spacing for all types of shell and tube exchangers, Int. J. Heat Mass
7 (1993) 30–38.
Transfer 38 (13) (1995) 2501–2506.
[38] H.Z. Huang, W.Q. Tao, An experimental study on heat mass transfer and
[8] H.D. Li, V. Kottke, Effect of baffle spacing on pressure drop and local heat
pressure drop characteristics for arrays of nonuniform plate length
transfer in shell-and-tube heat exchangers for staggered tube arrangement,
positioned obliquely to the flow direction, ASME J. Heat Transfer 115
Int. J. Heat Mass Transfer 41 (10) (1998) 1303–1311.
(1993) 568–575.
[9] P. Sthlik, V.V. Wadekar, Different strategies to improve industrial heat
[39] L.B. Wang, W.Q. Tao, Heat transfer and fluid flow characteristics of plate array
exchange, Heat Transfer Eng. 23 (6) (2002) 36–48.
aligned at angles to the flow direction, Int. J. Heat Mass Transfer 38 (1995)
[10] B.K. Soltan, M. Saffar-Avval, E. Damangir, Minimization of capital and
3053–3063.
operating costs of shell and tube condensers using optimum baffle spacing,
[40] Z.Y. Guo, D.Y. Li, B.X. Wang, A novel concept for convective heat transfer
Appl. Thermal Eng. 24 (17–18) (2004) 2801–2810.
enhancement, Int. J. Heat Mass Transfer 41 (1998) 2221–2225.
[11] J. Lutcha, J. Nemcansky, Performance improvement of tubular heat exchangers
[41] S. Wang, Z.X. Li, Z.Y. Guo, Novel concept and device of heat transfer
by helical baffles, Trans. Inst. Chem. Eng. 68 (1990) 263–270.
augmentation, in: Proceedings of 11th International Heat Transfer
[12] P. Stehlik, J. Nemcansky, D. Kral, Comparison of correction factors for shell-
Conference, vol. 5, 1998, pp. 405–408.
and-tube heat exchangers with segmental or helical baffles, Heat Transfer Eng.
[42] Z.Y. Guo, S. Wang, Novel concept and approaches of heat transfer
15 (1) (1994) 55–65.
enhancement, in: Cheng P. (Ed.), Proceedings of Symposium on Energy
[13] D. Kral, P. Stelik, H.J. Van Der Ploeg, Bashir I. Masster, Helical baffles in shell-
Engineering in the 21th Century, Begell House, vol. 1, New York, 2000, pp.
and-tube heat exchangers, Part One: experimental verification, Heat Transfer
118–126.
Eng. 17 (1) (1996) 93–101.
[43] B. Hua, P.S. Guo, S.S. Lu, et al. The study of field synergy theory in transfer
[14] S.L. Wang, Hydrodynamic studies on heat exchangers with helical baffles, Heat
process, in: Proceeding of 2001 IAMS International Seminar on Materials for
Transfer Eng. 23 (3) (2002) 43–49.
Use in Lithium Battery and Transport Phenomena in Material Processing,
[15] Z.G. Zhang, T. Xu, X.M. Fang, Experimental study on heat transfer enhancement
Japan, November 2001, pp. 26–27.
of a helically baffled heat exchanger combined with three-dimensional finned
[44] W.Q. Tao, Z.Y. Guo, B.X. Wang, Field synergy principle for enhancing
tubes, Appl. Thermal Eng. 24 (14–15) (2004) 2293–2300.
convective heat transfer - its extension and numerical verifications, Int. J.
[16] B. Peng, Q.W. Wang, C. Zhang, G.N. Xie, L.Q. Luo, Q.Y. Chen, M. Zeng, An
Heat Mass Transfer 45 (2002) 3849–3856.
experimental study of shell-and-tube heat exchangers with continuous helical
[45] W.Q. Tao, Y.L. He, Q.W. Wang, et al., A unified analysis on enhancing single
baffles, ASME J. Heat Transfer 129 (2007) 1425–1431.
phase convective heat transfer with field synergy principle, Int. J. Heat Mass
[17] Yong-Gang Lei, Ya-Ling He, Pan Chu, Rui Li, Design and optimization of heat
Transfer 45 (2002) 4871–4879.
exchangers with helical baffles, Chem. Eng. Sci. 63 (2008) 4386–4395.
[46] S. Shen, W. Liu, W.Q. Tao, Analysis on field synergy on natural convective heat
[18] Jian-Fei Zhang, Bin Li, Wen-Jiang Huang, Yong-Gang Lei, Ya-Ling He, Wen-
transfer in porous media, Int. Commun. Heat Mass Transfer 30 (8) (2003)
Quan Tao, Experimental performance comparison of shell side heat transfer for
1081–1090.
shell-and-tube heat exchangers with middle-overlapped helical baffles and
[47] Y.P. Cheng, Z.G. Qu, W.Q. Tao, et al., Numerical design of efficient slotted fin
segmental baffles, Chem. Eng. Sci. 64 (2009) 1643–1653.
surface based on the field synergy principle, Numer. Heat Transfer A Appl. 45
[19] S.V. Patankar, D.B. Spalding, A calculation procedure for the transient and
(6) (2004) 517–538.
steady state behavior of shell-and-tube heat exchanger, in: N.F. Afgan, E.U.
5380 J.-F. Zhang et al. / International Journal of Heat and Mass Transfer 52 (2009) 5371–5380

[48] Y.L. He, M. Wu, W.Q. Tao, et al., Improvement of the thermal performance of [52] W.Q. Tao, Numerical Heat Transfer, second ed., Xi’an Jiaotong University Press
pulse tube refrigerator by using a general principle for enhancing energy Xi’an, China, 2001.
transport and conversion processes, Appl. Thermal Eng. 24 (2004) 79–93. [53] S.V. Patakar, Numerical Heat Transfer and Fluid Flow, McGraw-Hill Company,
[49] Z.Y. Guo, W.Q. Tao, R.K. Shah, The field synergy (coordination) principle and its New York, 1980.
applications in enhancing single phase convective heat transfer, Int. J. Heat [54] Ernst U. Schlünder (Ed.), Heat Exchanger Design Handbook, vol. 3, Hemisphere
Mass Transfer 48 (2005) 1797–1807. Publishing Corporation, Washington, 1983. p. 3.3.5-16.
[50] V. Yakhot, S.A. Orszag, Renormalization group analysis of turbulence 1: basic [55] S.M. Yang, W.Q. Tao, Heat Transfer, third ed., High Education Press, Beijing,
theory, J. Sci. Comput. 1 (1996) 3–11. China, 1998. 163–164.
[51] FLUENT 6.3 user’s guide, FLUENT Inc., 2006, section 12.4.2, section 7.5, section
7.10, section 7.13.1.