Applied Thermal Engineering 143 (2018) 561–569

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Research Paper

Numerical investigation on double tube-pass shell-and-tube heat exchangers T

with different baffle configurations
⁎
Li Hea,b, , Peng Lib
a
State Key Laboratory of Hydraulic Engineering Simulation and Safety, Tianjin University, Tianjin 300072, China
b
School of Renewable Energy, North China Electric Power University, Beijing 102206, China

H I GH L IG H T S

• RNG k-ε model and standard wall treatment with boundary layers are adopted.
• STP-STHE and DTP-STHE are compared in the point of exergy theory.
• Pressure drop and heat transfer coefficient of DTP-STHEs with different baffles are compared.
• A new and more effective method is introduced to evaluate the economy of STHE.

A R T I C LE I N FO A B S T R A C T

Keywords: Thermo-hydraulic performances of shell-and-tube heat exchangers with different baffle types have been widely
Thermo-hydraulic performances evaluated using computational fluid dynamics (CFD) methods. In this paper, double-tube-pass shell-and-tube
Segmental baffle heat exchanger (DTP-STHE) is proposed to improve the recovered heat quality from the point of exergy analysis.
Heat exchangers Results reveal that the recovered heat quality in DTP-STHE is almost twice as that in single-tube-pass shell-and-
Exergy analysis
tube heat exchanger (STP-STHE). Furthermore, the heat transfer rates in both heat exchangers are nearly the
same. In all heat exchangers, increasing the mass flow rate of water in shell side can increase the exergy utilizing
efficiency and heat transfer rate, but the mass-weighted average temperature of water decreases. In order to
investigate thermo-hydraulic performances of different DTP-STHEs, three kinds of DTP-STHEs with segmental,
helical and flower baffles respectively are compared under the same conditions. Results show that flower baffles
produce the lowest pressure drop and lowest heat transfer coefficient. For better evaluating the economic per-
formance of these three kinds of heat exchangers, heat transfer rate per effective pumping power (i.e., QH/P0) is
introduced. The simulation results show that QH/P0 of flower baffle is the highest and QH/P0 of segmental baffle
the lowest.

1. Introduction However, the heat exchanger with segmental baffles normally has
some inherent defects resulting from structure limitations [9–12]: (1)
For last several decades, shell-and-tube heat exchangers (STHEs) in high pressure drop on the shell side because of the sudden contraction
their various forms are widely used in process industries because of and expansion of the flow at baffle cut section; (2) fouling formed in the
their easy manufacture, maintenance and upgrading, such as power stagnation zone near the rear of the baffles; (3) low shell-side mass
plants, air-conditioning, chemical plants and petroleum [1–3]. It is es- velocity across the tubes due to significant bypass streams and leakage
timated that about 30% of the heat exchangers in current use are of the streams caused by manufacturing tolerance and installation, and (4)
shell-and-tube type [4–6]. Conventional single baffles play a significant short operational lifetime as a result of flow induced tube vibration. In
role in STHEs. The heat transfer coefficient in the heat exchanger is order to minimize these shortcomings, it is necessary to investigate
enhanced by introduction of baffles. This is because the baffles guide other baffles to lower the shell side pressure drop and hence reducing
the shell side fluid to flow in a zigzag pattern between the tube bundle, the operating cost of heat exchanger. One successful attempt is STHE
which enhances the turbulence intensity and the local mixing, thus with helical baffles (HB-STHE) which was first invented in Czechoslo-
increasing the heat transfer coefficient [7,8]. vakia [13] in late 1980th. Compared with traditional segmental baffles,

⁎
Corresponding author at: State Key Laboratory of Hydraulic Engineering Simulation and Safety, Tianjin University, Tianjin 300072, China.
E-mail address: helix111@tju.edu.cn (L. He).

https://doi.org/10.1016/j.applthermaleng.2018.07.098
Received 20 August 2017; Received in revised form 4 July 2018; Accepted 17 July 2018
Available online 19 July 2018
1359-4311/ © 2018 Elsevier Ltd. All rights reserved.
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

velocity distribution in helical baffle system is more uniform and Maakoul et al. [22] studied thermo-hydraulic performances of STHEs
homogenous and the applications of the helical baffles greatly decrease with three different baffle types: segmental baffles, continuous helical
the proportions of bypass streams and leakage streams. Heat transfer baffles and trefoil-hole baffles. The results indicated that HB-STHEs
coefficient and pressure drop for shell side equipped with helical baffle could provide a good balance between heat transfer and pressure drop
system were also introduced in Ref. [14]. By creating different ar- characteristics. While the STHE with trefoil baffles (TB-STHE) could
rangements of the helical baffles, the comparison between these types enhance the heat transfer at the expense of a large pressure drop.
of baffles and the segmental one has been performed. Some equations Ambekar et al. [23] used a numerical method to investigate STHEs with
for both turbulent and laminar regimes were also proposed based on the different baffle segmental configurations. In this study, single, double,
derived pressure drop relationship and the rapid design algorithm. triple and flower segmental baffles were adopted to study the heat
Additionally, other effects have been undertaken for identifying the transfer and pressure drop in the shell side. Results showed that single
technical and economic performances of heat exchangers based on si- segmental baffles could provide good overall heat transfer coefficient at
mulation techniques. For example, Prithiviraj and Andrews [15] first cost of large pressure drop. Double, triple segmental baffles, flower
developed a 3D numerical code based on distributed resistance method baffles and helical baffles could reduce pressure drop while heat
along with volumetric porosities and surface permeabilities to model transfer coefficient reduced as well. However, the standard for identi-
the tubes in a baffled shell and tube heat exchanger. In this method, fying the most economical one was not presented in this study, resulting
fluid flow and temperature field were simulated and the agreements in its incapability of enhancing and verifying the simulation solutions.
between simulation results and experiment values were within ± 7.4% According to the above literature review, it can be generally con-
and 15%, respectively. Andrews and Master [16] performed detailed cluded that different heat exchangers, involving STHE with single,
three-dimensional CFD simulations to explore the performance of a double, triple segmental baffles, helical baffles with different helix
helical baffle heat exchanger. They discussed detailed mathematical angles, flower baffles, trefoil-hole baffles and multiple shell-pass, have
and numerical modeling of STHEs with helical baffles, in which three been extensively applied for improving flow and heat transfer perfor-
cases with helix angles of 10°, 25°, and 40° were considered. Inspection mance and reducing pressure drop. However, limited attention has
of the computed flows revealed distinct inner and outer regions, with been paid on the pressure drop per unit between single-tube-pass STHE
the outer region showing a very desirable plug flow characteristic. Xiao (STP-STHE) and double-tube-pass STHE (DTP-STHE) under the similar
et al. [17] used numerical method to simulate helical baffles heat ex- input conditions. Moreover, the previous studies merely considered the
changer with different Prandtl number fluids. Results found that the overall heat transfer coefficient on the shell side of the heat exchanger
helical baffles were used to direct the shell side fluid flow spirally across which could hardly judge the quality of heated water. Furthermore, the
the tubes, and the guide function would weaken when large Prandtl flow and heat transfer performance of multiple tube-pass STHE with
number fluids involved. Results also reveal that heat exchanger con- segmental baffles, flower baffles and non-continuous baffles were not
tained water as the shell side fluid achieved the best heat transfer compared. In this study, pressure drop per meter and recycled heat per
performance when helical tilt angle is at 40°. Gao et al. [18] focused on kilogram of water in the tube side will be compared between STP-STHE
the research on flow resistance and heat transfer of several shell-and- and DTP-STHE. After that, heat transfer coefficient, pressure drop and
tube heat exchangers with discontinuous helical baffles at different heat transfer rate per effective pumping power on DTP-STHE with
helical angles of 8°, 12°, 20°, 30°and 40°. Results disclosed that the different baffle configurations will be compared.
shell-side pressure drop and heat transfer coefficient of the heat ex-
changer with smaller helix angle were higher than those with larger 2. Modeling details
helix angle at a given shell-side volume flow rate. However, in the
condition of the same shell-side Reynolds number, the flow resistance 2.1. Physics model
with larger helix angle is lower and the heat transfer performance is
better. Zhang et al. [19] employed water as working fluid to simulate This study mainly focuses on the STHE with segmental, helical,
shell-and-tube heat exchangers with small-angle helical baffles. The trefoil-hole and flower baffles. Configuration of DTP-STHE with helical
paper focused on non-continuous helical baffles with a helix angle of baffles is illustrated in Fig. 1, and the detailed geometry parameters
10–30°, and their shapes are 1/4 ellipse, 1/4 sector and 1/3 sector. This DTP-STHE are listed in Table 1. It should be specially mentioned that all
study assumed that heat exchanger tubes were kept at a constant parameters, except for baffle types of these three heat exchangers,
temperature. Result indicated that the helical baffles increased the heat
transfer rate and decreased the pressure drop compared with conven-
tional segmental baffles heat exchangers.
In terms of STHE with helical baffles, other attempts were also make
to achieve better flow and heat transfer performance. For example, You
et al. [20] developed a numerical model based on the concepts of
porosity and permeability for the shell-side flow and heat transfer of
STHXs, where the distributed resistance and heat source, and the dis-
tributed turbulence kinetic energy and its dissipation rate were in-
troduced to account for the impacts of tubes on fluid. This study com-
pared contours between STHE with segmental baffles (SG-STHE) and
STHE with flower baffles (FB-STHE), and results pointed out that the
FB-STHE has a better overall thermal hydraulic performance than the
SG-STHX. However, this study didn’t investigate the heat transfer per-
formance between these two heat exchangers. Wang et al. [21] com-
pared the combined multiple shell-pass shell-and-tube heat exchanger
(CMSP-STHE) with SG-STHE by means of computational fluid dynamics
(CFD) method. They found that under the same mass flow rate and
overall heat transfer rate, the average overall pressure drop of the
CMSP-STHE was lower than that of SG-STHE by 13%, and that under
the same pressure drop in the shell side, the overall heat transfer rate
and mass flow rate of CMSP-STHE were higher than those of SG-STHE. Fig. 1. Configurations of DTP-STHE with helical baffles.

562
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

Table 1
Geometry parameters of MTP-STHEs.
Item Dimensions and description

Baffle shape Segmental Helical Flowers

Shell inside diameter, Di/mm 87 87 87

Length of heat exchanger/mm 500 500 500
External tube diameter, d0/mm 19 19 19
Tube thickness/mm 1.5 1.5 1.5
Effective tube length, L/mm 365 365 365
Tube pitch, pt/mm 24 24 24
Tube number, NT 6 6 6
Baffle spacing, B/mm 57 57 57
Baffle thickness/mm 3 3 3
Inlet and outlet spacing/mm 61 61 61
Baffle cut, Di% 25 – –

Table 2 equations are shown as follows:

Thermo physical properties of shell-side hot water and tube-side cold water. Continuity Equation:
Parameter Unit Value ∂ (ρui )
=0
∂x i (1)
Hot water Cold water
Momentum Transport Equation:
Density, ρ kg·m−3 974 998
Conductivity, λ W·m−1·K−1 0.66 0.60 ∂ (ρui uk ) ∂p ∂ ⎛ ∂u
Specific heat, Cp J·kg−1·K−1 4194 4182 =− + ueff k ⎞ ⎜ ⎟

Dynamic viscosity, μ kg·m−1·s−1 0.00037 0.001

∂x i ∂xk ∂x i ⎝ ∂x i ⎠ (2)
Energy equation:

remain constant for better comparisons among the three different baf- ∂ (ρui T ) ∂ ⎡⎛ μ μ ∂T ⎤
= + t⎞
fles [24]. The working fluid in both the shell and tube side of the heat
⎜ ⎟

∂x i ∂x i ⎢
⎣⎝ Pr σ ⎥
T ⎠ ∂x i ⎦ (3)
exchanger is water. All properties of water are seen as constants be-
cause of small temperature change both in shell side and in tube side Turbulent kinetic energy:
whose values are shown in Table 2. Also, all parts of heat exchanger are
∂ (ρkui ) ∂ ⎡ ∂k ⎤
made from carbon steel with a thermal conductivity λ = 45 W/m2·K. = αk μeff + Gk + ρε
∂x i ∂x j ⎢
⎣ ∂xj ⎥
⎦ (4)
2.2. Governing equations Turbulent dissipation energy:

The commercial code FLUENT is adopted to simulate the flow and ∂ (ρεui ) ∂ ⎡ ∂ε ⎤ ε ε2
= αε μeff + C1∗ε Gk−C2ε ρ
heat transfer in the computational model. The numerical simulation is ∂x i ∂x j ⎢
⎣ ∂x ⎥
j⎦ k k (5)
performed with a three-dimensional steady-state turbulent flow system.
where
In the standard k-ε model, for each component of Reynolds Stress,
assuming that the viscous coefficient μt is isotropic scalar. However, in μeff = μ + μt
⎧
the case of curved streamline, turbulence is anisotropic, μt thus should ⎪ k2
⎪ μt = ρCμ
be an anisotropic scalar [25]. Additionally, RNG k-ε and Realizable k-ε ε
⎪C = 0.0845, αk = αε = 1.39
models are proposed to remedy the defects of k-ε model. The RNG k-ε ⎪ μ
put forward by Yakhot and Orissa [26] is similar in form to the standard ⎪ η (1 − η η0 )
C1∗ε = C1ε−
⎪ 1 + βη3
k-ε model, but has the following refinements: (1) the turbulent viscosity ⎪
C1ε = 1.42, C2ε = 1.68
has been modified, so the effect of swirl on turbulence is included in the ⎨
η = (2Eij Eji )1 2k
RNG model, enhancing accuracy for swirling flows; and (2) the RNG ⎪ ε
model has an additional term in its ε equation that improves the ac- ⎪
curacy for rapidly strained flows. Therefore, compared with standard k-
⎪
⎪
⎪
Eij =
1
2 ( ∂ui
∂xj
+
∂uj
∂xi )
ε model, the RNG k-ε model can better handle the flows with high strain η0 = 4.377, β = 0.012
⎪
rate and large curved streamline [25,26]. ⎪ σT = (0.9 ∼ 1.0)
⎩ (6)
In terms of the Realizable k-ε model, it is usually applied to the
circumstance when the time-averaged strain rate is significantly large.
RNG k-ε model is thus recognized as the most appropriate model among 2.3. Boundary conditions and near wall treatment
these three k-ε models when simulating flows in the shell side of STHEs.
All the governing equations are modified according to the conditions of The momentum boundary condition of no slip and no penetration is
the simulation setup. As the problem is assumed to be steady, the time set for all solid walls. The thermal boundary condition of zero heat flux
dependent parameters are dropped from the equations. The resulting is set for the shell wall and inlet and outlet nozzle walls, heat transfer

563
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

between water and baffles is neglected. While the walls of tubes and 2.6. Data reduction
tube bundle, which also represent the solid-fluid interfaces between the
two fluid domains, have the thermal boundary condition of coupling 2.6.1. Heat flux and shell-side heat transfer coefficient
heat transfer. The inlets for the shell and tube sides are set as boundary Heat flux is an important parameter in heat exchanger analysis. QĊ is
conditions of mass flow inlet, the outlets are set as pressure-outlet. The defined as transferred heat to cold water, which can be calculated as
outlets are assumed to have a pressure of constant so the pressure dif- follow:
ference between inlet and outlet is equal to the pressure drop on both
QĊ = GC × cp, C × (tout , c−tin, c ) (7)
shell and tube sides.
In order to simplify the model and save computation time, standard where Gc is the mass flow rate of cold fluid and cp,c is the specific heat of
wall treatment is utilized, typically y+ of the first layer grid is 30 ≤ y+ cold water at certain temperature. Qḣ is the heat released by hot water
≤ 300. which can also be calculated as follow:
Q̇H = GH × cp, H × (tin, H −tout , H ) (8)
2.4. Numerical method where GH is the mass flow rate of hot water and cp,H is the specific heat
of hot water. In this study, the heat transfer between STHE and en-
The CFD code FLUENT is used to calculate the fluid flow and heat vironment is neglected, QĊ thus equals to Q̇H .
transfer in the computational domain. The governing equations are The shell side heat transfer coefficient of heat exchanger αs is an-
solved by the finite-volume method with pressure-velocity coupling other important parameter, which is defined by the following equation:
scheme and the SIMPLE algorithm. The second-order upwind scheme is
Q̇H
adopted for the momentum, energy, turbulence and its dissipation rate. αs =
A0 Δtm (9)
The pressure term is treated with the standard scheme. In addition to
under relaxation factor of pressure (0.5), default under relaxation fac- where A0 is the heat exchanger area based on the outer diameter of tube
tors of the solver are used, which are considered as 0.7, 0.8, and 0.8 for d 0 and effective tube length l. Δtm can be calculated by the Eq. (11).
the momentum, turbulent kinetic energy, and turbulent energy dis-
A0 = nπd 0 l (10)
sipation, respectively. The convergence criterions for residual mon-
itoring were assumed to be 10−3 for flow equations and 10−7 for the Δt − Δt
⎧ Δtm = ln(Δmax min
tmax Δtmin )
energy equation. Some physically meaningful variables, such as tem- ⎪
perature and velocity, were also monitored. ⎨ Δtmax = max(Th, in−Tc, out , Th, out−Tc, in )
⎪ Δtmin = min(Th, in−Tc, out , Th, out−Tc, in) (11)
⎩

2.5. Mesh generation and independence verification

2.6.2. Pressure drop and effective pumping power
Pressure drop is another important factor to consider when de-
Although the structure of heat exchanger is complex, the compu-
signing the STHE, because pumping power is highly related to pressure
tation domain is meshed with unstructured grids by commercial code
drop, and therefore lower pressure drop results in lower operating
ANSYS ICEM CFD. The regions adjacent to the tubes and shell are
costs. As for STHE in this paper, if potential energy is neglected, total
meshed finer with boundary layer to better fit in enhanced wall treat-
pressure drop can be expressed as,
ment in FLUENT and thus to improve the calculation accuracy, as
shown in Fig. 2. The quality of the mesh for each STHX (elements ΔPt = ΔPs + ΔPd, (12)
quality, skewness, angle and so on) is evaluated using the built-in Mesh
where ΔPs is static pressure drop and ΔPd is dynamic pressure drop.
Metrics in ICEM (Table 3).
Effective pumping power is largely determined by pressure drop,
Before the numerical simulation, mesh independence test is carried
given by:
out with adopting three different mesh size groups for the STP-STHE.
The total elements in the geometry are 3684238, 4598626, and Po = ΔPt ·V , (13)
5806253, respectively (Table 4). It can be seen from Tables 5 and 6 that where V denotes volume of STHE in the shell side.
the differences in pressure drop and heat transfer coefficient between
the last two mesh systems are less than 1%. Thus, the second mesh 2.6.3. Determination of shell side vs and Res number
system is adopted for the simulation considering the calculating time vs and Res number on the shell side has significant effect on shell
and solution precision. side flow and heat transfer performance. For the same Re number,

Fig. 2. Mesh: (a) cold water inlet; (b) helical baffles.

564
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

Table 3
Mesh details and metrics.
STHE baffles type Total nodes Total elements Min quality Min angle Min skew Min determinant

Segmental 5,569,794 1,334,821 0.2747 16.15 0.2410 0.3574

Helical 6,168,013 1,381,053 0.2801 12.87 0.2686 0.3355
flower 6,226,362 1,442,593 0.2837 15.89 0.2677 0.3000

regardless of the specific structure of the STHE, the flow of the fluid on Table 5
the shell side should have similarity. Comparison of heat transfer coefficient between the experimental results and
As for SG-STHE and HB-STHE, the mean shell side velocity vs is simulated results in the shell side.
defined by: Reynolds number Experimental data Simulation results
(W·m−2·K−1) (W·m−2·K−1)
GH
Vs = ,
ρs Across (14) 46,941 2326 2245
55,476 2404 2305
where GH is the mass flow rate of hot water on the shell side and Across 64,011 2477 2353
represents the cross-flow area at the shell centerline, ρs is the density of 72,546 2546 2393
hot water. For the segmental baffles in this case: 81,081 2587 2428

Across = (Ds −Nc d 0) B (15)

where Ds is the internal diameter of the shell, d0 the external diameter Table 6
Comparison of pressure drop between the experimental results and simulated
of the tubes, B the baffle spacing and Nc the number of tubes in the
results in the shell side.
central row.
For helical baffles: Reynolds number Experimental data (kPa) Simulation results (kPa)

Across = 0.5(Ds −Nc d 0) B (16) 46,941 8.99 7.98

55,476 11.70 11.19
where D1 is the tube bundle-circumscribed circle, pt the tube pitch. 64,011 15.70 14.94
With the vs the Re number of the shell side fluid in SG-STHE and HB- 72,546 20.01 19.24
STHE can be calculated by as follow: 81,081 25.22 24.10

Vs·de ·ρs
Res =
μs (17) outlet spacing are both 61 mm, and the baffle cut is 25% of the shell
inside diameter. Comparison of the simulation results with experiment
where μs is the dynamic viscosity of hot water, de the equivalent
results is shown in Fig. 3(a) and (b), Tables 4 and 5. In the figure, Gs
diameter. For the square tube layout, de can be gained from the fol-
donates mass flow rate of water in the shell side. It can be seen from
lowing equation.
Fig. 3, Tables 4 and 5 that the difference between simulation results and
Ds2−Nt d 02 experimental results is between 3.5% and 6.1% for heat transfer coef-
de =
Ds + Nt d 0 (18) ficient and between 3.8% and 11.3% for pressure drop, so the numer-
ical method in this study is reliable and valid.
For flower baffles, Res is defined through Eq. (19):
ρs Ve de
Res = 4. Results and discussion
μs (19)
where Ve refers to the axial velocity at the breaches of flower baffles 4.1. Comparison between STP-STHE and DTP-STHE
on the shell side.
Exergy theory is a powerful tool to analyze the heat transfer effi-
3. Model validations ciency because this can consider both heat quality and heat quantity. In
this study, STP-STHE and DTP-STHE are compared from point of heat
The model is validated by using a frequently-used STHE design transfer performance (Table 7). In this table, Gs represents mass flow
software HTRI. The inputs of geometrical parameters and boundary rate of hot water in shell side, Δt represents increasing temperature of
conditions in the validation are the same as in simulations. A conven- cold water in tube side, Ex is the total absorbed energy of cold water
tional STHX with segmental baffles is used in the experiment, where the and Ex/m denotes the absorbed energy per kilogram. It should be no-
warm water flows in the shell side, while the cool water flows in the ticed that in order to better identify heat transfer efficiency, mass flow
tube side in a counter current configuration. The shell has an internal rate of cold water per tube, inlet temperature of cold water and inlet
diameter of 87 mm and a length of 500 mm. Six tubes with an external temperature of hot water both in STP-STHE and DTP-STHE remain the
diameter of 19 mm are installed inside the shell with a square ar- same (i.e., 0.2 kg/s, 293.15 K and 353.15 K). Results show that Ex in
rangement. The number of baffles is 5 with a baffle spacing of 57 mm these two kind of heat exchangers is almost the same, which is due to
and the baffles thickness is 3 mm. Furthermore, the inlet spacing and that heat exchange area in both exchangers are the same. However, Δt

Table 4
Comparison of three different mesh systems.
STHE baffles type Total nodes Total elements Min quality Min angle Min determinant

System1 3,684,238 1,053,987 0.2697 16.38 0.3253

System2 4,598,626 1,334,821 0.2732 16.72 0.3125
System3 5,806,253 1,586,956 0.2837 17.03 0.3343

565
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

26.00 main cross-flow streams, the fluid velocities near the shell wall surface
Experimental results
24.00 CFD results
and at the corner of baffles are relatively small, resulting in the stag-
22.00 nation flow generation. The turbulence and local mixing result from the
20.00
segmental baffles are positive to heat transfer but rather negative to
huge pressure loss formation as pressure changes suddenly at each
18.00
P/kPa

baffle section. Therefore, most of fluid energy is spent in areas where

16.00
there are few tubes, leading to an inefficient conversion of pressure
14.00 drop in heat transfer.
12.00 As for heat exchanger with helical baffles, the function of helical
10.00 baffles is used to form a spiral fluid flow passage forcing the fluid to
8.00 flow through the channels between baffles as depicted in the Fig. 4b.
6.00 With the help of helical baffles, the flow pattern in the shell side is
45000 50000 55000 60000 65000 70000 75000 80000 rotational, and the shell fluid passes though the tube bundles close to an
Res ideal helical pattern. Compared to heat exchangers with helical baffles,
(a) no obvious changes could be found for the direction and the magnitude
of velocity at the near-baffle edge areas, and the uniformity of the ve-
2600 Experimental Results locity distribution with helical baffles are much better than that with
2550 CFD results
segmental baffles. Due to the effect of baffle inclination angle, the fluid
2500 flow in the axial direction is enhanced and turned into a plug flow.
W·m ·K )

2450 Fluid flow with helical baffles is continuous and the dead zones do not
-1

2400 occur near the helical baffle. Since fluid in the shell side in the helical
-2

2350 baffles heat exchanger keeps moving forward spirally as the cross plug
2300 flow with little back flow or vortex, both the helical baffles and seg-
2250 mental baffles have a similar heat transfer effect. The spiral motion
2200
brings about better mixing, and the heat transfer in this region is sig-
nificantly enhanced while the pressure drop is reduced.
2150
45000 50000 55000 60000 65000 70000 75000 80000 From Fig. 4c, it is clearly seen that the fluid velocity magnitude on
Res the shell side changes periodically in the heat exchanger with flower
(b) baffles. When the fluid passes through a baffle, it is accelerated rapidly
and flows across the breaches with large velocity magnitude. After
Fig. 3. Comparison of the experimental results and simulated results in the shell
side: (a) pressure drop ΔP; (b) heat transfer coefficient α. rushing out of the breaches, the fluid is expanded suddenly and the
velocity magnitude is decreased gradually. This periodic flow pattern is
caused by the periodic changes of the arrangement of flower baffles.
Table 7 Moreover, it is noticed that two recirculation flow regions are generated
Heat transfer performance comparison between STP-STHE and DTP-STHE.
in the downstream just behind a baffle, where the velocity magnitude is
Type Gs (kg/s) 1.1 1.3 1.5 1.7 1.9 very small. Lastly, like heat exchanger with helical baffles, vortex forms
just behind the baffles, which is positive to heat transfer and heat
STP-STHE Δt (K) 3.29 3.41 3.52 3.62 3.68
Ex (J) 16,496 17,096 17,645 18,165 18,484
mixing.
Ex/m (J·kg−1) 13,747 14,247 14,704 15,138 15,403 Fig. 5 represents the vectors at the shell side plane as is showed in
DTP-STHE Δt (K) 6.6 6.9 7.2 7.3 7.4 the left side for each of the computed STHE configurations. Segmental
Ex (J) 16,436 17,408 17,998 18,266 18,689 baffles have the highest shell side maximum velocities, located at the
Ex/m (J·kg−1) 27,393 19,013 29,997 30,443 31,148
gaps between baffles and shell side. The high velocity jets can be seen
clearly in Fig. 5b. By comparison to segmental baffles, helical baffles
produce lower shell side maximum velocities, located at the baffle
and Ex/m in DTP-STHE are obviously higher than those in STP-STHE.
window and near the central tube respectively Fig. 5a and b. As can be
This is because the heat exchange time of water in DTP-STHE is almost
seen in Fig. 5a and c, flower baffles produce no dead zones with re-
the twice of that in STP-STHE.
markable fluid recirculation zones compared with segmental baffles.
According to exergy analysis, the heat exergy of the heat stream can
Flower baffles and helical baffles produce relatively lower maximal
T
( )
be expressed as E = Q 1− T0 , where Q is the total heat of cold water
velocities, better flow distributions, without dead zones and fluid re-
after heat transfer, T0 is the environmental temperature and T is the
circulation areas. Higher maximal velocities, dead zones, and fluid re-
temperature of cold water after heat transfer. Δt in the tube side of DTP-
circulation areas always produce an increase in frictional pressure
STHE is almost twice of that in the tube side of STP-STHE. So T in DTP-
drops. Thus, it is expected that the pressure drop will go from highest to
STHE is higher than that in STP-STHE. Therefore, DTP-STHE has a
lowest in the following order: segmental, helical and flower baffles.
higher recovered heat quality without reducing heat quantity.
4.3. Pressure drop and heat transfer performance
4.2. Flow field characteristics on the shell side of DTP-STHE
The pressure drop is of great importance in the design of shell-and-
The flow behavior in the shell side for the STHX is different de- tube heat exchangers, pumping cost are highly related to pressure drop,
pending on the baffle type, Fig. 4 shows the streamlines in the shell side and therefore lower pressure drop results in lower operating costs.
for the three kinds of heat exchangers. For the conventional segmental Fig. 6 depicts the variation of the pressure drop ΔP versus the shell side
baffles, as shown in Fig. 4a, the baffles in the shell side work as ob- mass flow rate for the three heat exchangers. The pressure drop in-
stacles which divide the original axial flow into cross flow, fluid re- creases proportional to the mass flow rate for the three heat exchangers.
circulation and some vortex. The zigzag flow pattern also causes large As expected previously, segmental baffles produce the highest pressure
dead zones at the back of the baffles. Both of magnitude and direction drop, higher on average by 135% and 218% than STHEs with helical
of velocity changes dramatically as the fluid has to climb over baffles in and flower baffles respectively. This high pressure drop is caused by the
the zones near the edge of baffles. Compared to the velocities of the high velocity jets at the baffles opening and large fluid recirculation

566
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

Fig. 4. Shell path lines of three different heat exchangers: (a) segmental baffle heat exchanger; (b) helical baffle heat exchanger; (c) flower baffle heat exchanger.

zones. The reason is that the flow distribution with segmental baffles on 4.4. Comprehensive performance analysis
the shell side is zigzag, flow separation at the edge of baffles causes
abrupt momentum change and severe pressure drop. Whereas the pri- As mentioned above, the design of a shell-and-tube heat exchanger
mary flow direction of helical baffles and flower baffles does not change is a balance between higher heat transfer rate and lower effective
dramatically and flow velocity magnitude of both helical baffles and pumping power in the fluids (these two parameters is closely related to
flower baffles does not change severely. heat transfer coefficient and pressure drop respectively, but can directly
Fig. 7 represents the comparisons of the shell side heat transfer evaluate the economy of STHE), since the two parameters depend
coefficient for the three heat exchangers versus Reynolds number of highly on each other. So it is necessary to increase the fluid velocities
shell side. A key point should be stressed that the dots in same position when improving the heat transfer rate. Such measure always produces
at each line are in same mass flow rate. It can be observed that the shell an increase in frictional pressure drops, thus increasing the effective
side heat transfer coefficient α increases with the increase of mass flow pumping power. In order to better evaluate the economy of STHEs with
rate Gs or Reynolds number in the shell side. It can be seen clearly from different kinds of baffles, heat transfer rate per unit effective pumping
Fig. 7 that under same mass flow rate of water in the shell side Reynolds power at the shell side, QH/Po is introduced to evaluate the economy of
number of STHE with helical baffles is obviously higher than those of different heat exchangers. In practical application the pressure drops of
STHE with segmental baffles and flower baffles. So α for the STHE with the heat exchangers are usually limited, and the goal is to find the
helical baffles is a little higher than that for STHE with segmental and design parameter combination that results in the highest heat transfer
helical baffles. However, the difference of α among these three heat rate within the effective pumping power limitations. Thus, the ratio QH/
exchangers is within 6%. Furthermore, it can be judged that under Po is a useful parameter.
Reynolds number, flower baffles has the highest heat transfer coeffi- Fig. 8 shows the comparisons of the ratio of the heat transfer rate to
cient compared to other two baffles. In order to evaluate the economy effective pumping power for the three heat exchangers with different
of STHEs with different baffles, a new method needs to seek to study the baffles. From Fig. 8, it can be seen that the performance parameter QH/
heat exchanger. Po decreases with increases in the shell side flow rate. It is clearly seen
that the STHE with flower baffles has the best performance ratio QH/Po
among the three STHXs. On average, the performance ratio QH/Po for

567
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

Fig. 5. Velocity vectors of three kinds of heat exchangers in longitudinal section: (a) segmental baffle (b) helical baffle (c) flower baffle.

Fig. 6. Shell side pressure drop of different kind heat exchanger versus mass Fig. 7. Shell side heat transfer coefficient versus Reynolds number of shell side.
flow rate.
be designed to replace an existing one, if the two heat exchangers have
STHX with flower baffles is 223% and 49% higher on average than equal effective pumping power, the new STHE must have a larger heat
STHXs with segmental and helical baffles respectively. Based on the transfer rate, and if the two heat exchanger have equal heat transfer
results discussed above, it is clear that STHX with flower baffles per- rate, the new STHE must have a lower effective pumping power.
forms better for the flow rates interval studied. Furthermore, the shell
side fluid mass flow rate in this section range from 1.1 kg/s to 1.9 kg/s, 5. Conclusion
QH/Po rate decreases with the shell side fluid mass rate increasing.
However, the recovered heat quality decreases at the same time. From This study compares the performances of STP-STHE and DTP-STHE
this point of view, the conclusion can also draw that if a new STHE is to with segmental baffles. Results reveal that DTP-STHE can effectively

568
L. He, P. Li Applied Thermal Engineering 143 (2018) 561–569

4000
Heat Transfer 126 (6) (2004) 877–885.
Segmental [5] B.B. Gulyani, Estimating number of shells in shell and tube heat exchangers: a new
3500 Helical approach based on temperature cross, Trans. ASME J. Heat Transfer 122 (3) (2000)
Flower 566–571.
3000 [6] B.I. Master, K.S. Chunangad, V. Pushpanathan, Fouling mitigation using he-
lixchangers, in: Engineering Conference International, 2003.
2500
[7] E. Pal, I. Kumar, J.B. Joshi, N.K. Maheshwari, CFD simulation of shell-side flow in a
shell-and-tube type heat exchanger with and without baffles, Chem. Eng. Sci. 143
QH/Po

(2016) 314–340.
2000 [8] L. Huadong, V. Kottke, Effect of baffle spacing on pressure drop and local heat
transfer in shell-and-tube heat exchangers for staggered tube arrangement, Int. J.
1500 Heat Mass Transfer 41 (1998) 1303–1311.
[9] P. Stehlik, J. Nemcansky, D. Kral, L.W. Swanson, Comparison of correction factors
1000 for shell-and-tube heat exchangers with segmental or helical baffles, Heat Transfer
Eng. 15 (1994) 55–65 (Compendex).
[10] H.D. Li, V. Kottke, Visualization and determianation for local heat transfer coeffi-
500
cients in shell-and-tube heat exchangers for staggered tube arrangement by mass
1.0 1.2 1.4 1.6 1.8 2.0 transfer measurements, Exp. Therm. Fluid Sci. 17 (3) (1998) 210–216.
[11] D. Kral, P. Stehlik, H.J. Van der Ploeg, B.I. Master, Helical baffles in shell-andtube
-1
Gs(kg·s ) heat exchangers, Part I: Experimental verification, Heat Transfer Eng. 17 (1) (1996)
93–101.
Fig. 8. Heat transfer rate per effective pumping power versus mass flow rate. [12] Q.W. Wang, Current status and development of shell-side heat transfer enhance-
ment of shell-and-tube heat exchangers with helical baffles, J. Xi’an Jiaotong Univ.
38 (9) (2004) 881–886 (in Chinese).
improve recovered heat quality in the shell side without decreasing [13] J. Lutcha, J. Nemcansky, performance improvement of tubular heat exchangers by
helical baffles, Chem. Eng. Res. Des. 68 (1990) 263–270 (Compendex).
heat transfer rate. Then, a numerical model is used to simulate and [14] M.R. Jafari Nasr, A. Shafeghat, Fluid flow analysis and extension of rapid design
compare the thermo-hydraulic performances of DTP-STHEs with dif- algorithm for helical baffle heat exchangers, Appl. Therm. Eng. 28 (2008)
ferent baffle types: segmental, helical and flower baffles. Flow analysis 1324–1332.
[15] M. Prithiviraj, M. Andrews, Three-dimensional numerical simulation of shell-and-
in shell side shows that velocity distribution in helical baffles is more tube heat exchangers. Part 1: heat transfer, Int. J. Comput. Methodol. 33 (8) (1998)
uniform and homogenous as compared to segmental and flower baffles. 799–816.
This leads to less dead zones and less fluid recirculation areas inside the [16] M.J. Andrews, B.I. Master, Three-dimensional modeling of a helixchanger heat
exchanger using CFD, Heat Transfer Eng. 26 (6) (2005) 22–31.
shell as compared to segmental baffles. Moreover, results indicate that [17] X.M. Xiao, L.H. Zhang, X.G. Li, B. Jiang, X.L. Yang, Y.M. Xia, Numeral investigation
the use of flower baffles can provide the best balance between heat of helical baffles heat exchanger with different Prandtl number fluids, Int. J. Heat
transfer rate and effective pumping power when compared to the Mass Transfer 63 (2013) 434–444.
[18] B. Gao, Q.C. Bi, Z.S. Nie, J.B. Wu, Experimental study of effects of baffle helix angle
conventional segmental baffles and helical baffles.
on shell side performance of shell-and-tube heat exchangers with discontinuous
However, simulation in this model still has some drawbacks, further helical baffles, Exp. Thermal Fluid Sci. 68 (2015) 48–57.
points should be investigated: (1) the comparison between baffle types [19] M.H. Zhang, F. Meng, F.Z. Geng, CFD simulation on shell-and-tube heat exchangers
on induced fouling resistances, flow vibrations and leakages, (2) the with small-angle helical baffles, Front. Chem. Sci. Eng. 9 (2) (2015) 183–193.
[20] Y.H. You, A.W. Fan, S.Y. Huang, W. Liu, Numerical modeling and experimental
study of the effect of varying other design parameters that have an validation of heat transfer and flow resistance on the shell side of a shell-and-tube
effect on the shell-side thermo-hydraulic performances, and (3) un- heat exchanger with flower baffles, Int. J. Heat Mass Transfer 55 (2012)
certainties associated with modeling inputs and sensitive analysis of 7561–7569.
[21] Q.W. Wang, Q.Y. Chen, G.D. Chen, M. Zeng, Numerical investigation on combined
simulation results [27–29]. multiple shell-pass shell-and-tube heat exchanger with continuous helical baffles,
Int. J. Heat Mass Transfer 52 (2009) 1214–1222.
Acknowledgements [22] A.E. Maakoul, A. Laknizi, S. Saadeddine, M.E. Metoui, A. Zaite, M. Meziane,
A.B. Abdellah, Numerical comparison of shell-side performance for shell and tube
heat exchangers with trefoil-hole, helical and segmental baffles, Appl. Therm. Eng.
The authors thank the editor and the anonymous reviewers for their 109 (2016) 175–185.
helpful comments and suggestions. This research was supported by the [23] A.S. Ambekar, R. Sivakumar, N. Anantharaman, M. Vivekenandan, CFD simulation
study of shell and tube heat exchangers with different baffle segment configura-
Strategic Priority Research Program of Chinese Academy of Sciences tions, Appl. Therm. Eng. 108 (2016) 999–1008.
(XDA20040302) and Fundamental Research Funds for the Central [24] J.F. Yang, M. Zeng, Q.W. Wang, Numerical investigation on combined single shell-
Universities. pass shell-and-tube heat exchanger with two-layer continuous helical baffles, Int. J.
Heat Mass Transfer 84 (2015) 103–113.
[25] Fluent Inc., FLUENT User’s Guide. Fluent Inc., 2017.
References [26] V. Yakhot, S.A. Orissa, Renormalization-group analysis of turbulence: Basic theory,
J. Sci. Comput. 1 (1) (1986) 3–11.
[1] B. Sunden, Computational fluid dynamics in research and design of heat ex- [27] Y.Z. Chen, L. He, Y.L. Guan, H.W. Lu, J. Li, Life cycle assessment of greenhouse gas
changers, Heat Transfer Eng. 28 (2007) 898–910. emissions and water-energy optimization for shale gas supply chain planning based
[2] Z.G. Zhang, C.S. Wu, X.M. Fang, X. Gao, Z. Wang, Experimental study of shell-side on multi-level approach: case study in Barnett, Marcellus, Fayetteville, and
heat transfer coefficient and pressure drop for an integrally helical baffled heat Haynesville shales, Energy Convers. Manage. 134 (2017) 382–398.
exchanger combined with different enhanced tubes, Ind Eng. Chem. Res. 48 (8) [28] Y.Z. Chen, H.W. Lu, J. Li, G.H. Huang, L. He, Regional planning of new-energy
(2009) 4040–4044. systems within multi-period and multi-option contexts: a case study of Fengtai,
[3] B.I. Master, K.S. Chunangad, A.J. Boxma, D. Kral, P. Stehlík, Most frequently used Beijing, China, Renew. Sust. Energy Rev. 65 (2016) 356–372.
heat exchangers from pioneering research to worldwide applications, Heat Transfer [29] X. Cheng, L. He, H.W. Lu, Y.Z. Chen, L.X. Ren, Optimal water resources manage-
Eng. 27 (6) (2006) 4–11. ment and system benefit for the Marcellus shale-gas reservoir in Pennsylvania and
[4] K.J. Bell, Heat exchanger design for the process industries, ASME Trans. ASME J. West Virginia, J. Hydrol. 540 (2016) 412–422.

569