Case Studies in Thermal Engineering 32 (2022) 101873

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

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

Numerical investigation of the flow dynamics and heat transfer in

a rectangular shell-and-tube heat exchanger
Marwa Ben Slimene a, *, Sébastien Poncet b, Jamel Bessrour a, Ftouh Kallel c
a
Mechanical Engineering Department, National Engineering School of Tunis, BP37, El Belvedere, 1002, Tunis, Tunisia
b
Mechanical Engineering Department, Université de Sherbrooke, Sherbrooke, QC, J1K 2R1, Canada
c
Energetic Engineering Department, Private University of Tunis, 30 Av. Kheireddine Pacha 1002, Tunis, Tunisia

A R T I C L E I N F O A B S T R A C T

Keywords: Numerical simulations have been carried out to investigate the turbulent flows and coupled
RANS modeling conductive/convective heat transfers in a glycol to water shell-and-tube heat exchanger (one
Shell-and-tube heat exchanger shell, one tube doing six passes). The refrigerant (glycol) flows in the tube, and a secondary fluid
Rectangular geometric shell (water) flows in the shell. The k-ω SST model is used to close the system of equations in the
Concentric annular pipe turbulent regime. The present model is first favorably compared and validated against experi­
Baffles
mental and numerical researches for a concentric annular heated pipe for two radius ratios
(R* = 0.1 and 0.5) at a bulk Reynolds number ReDh = 8900 and a Prandtl number Pr = 0.71. It is
then extended to consider the shell-and-tube configuration with a rectangular shell. Its thermo-
hydraulic performances in the tube side are quantified for different Reynolds numbers at the
cold inlet ranging from 1.03 × 103 to 1.47 × 105. The performance of the heat exchanger is then
enhanced by introducing baffles in the rectangular shell. The results are finally discussed in terms
of three global performance metrics.

1. Introduction
Heat exchangers found many engineering applications such as for heating/cooling, refrigeration or air conditioning systems,
thermal energy storage systems, chemical industries and power generation in large thermal plants, and so on. The heat exchanger
design depends on the targeted applications. The most common ones are the heat pipes exchangers [39], straight tube and helical coil
heat exchangers [38], plate heat exchangers [1] to name just a few. Recently, Edreis and Petrov proposed a detailed review of the pros
and cons of different types of heat exchangers, while [32] focused on the applications of porous materials, foams and wire meshes as
heat transfer enhancement techniques applied to heat exchangers.
The most currently employed technology of heat exchangers in process or food industries is the shell and tube heat exchanger
(STHE) thanks to its versatility, robustness, easy maintenance and possible geometrical improvements, as stated by Ref. [4]. The
literature review demonstrates that most works focused on shell and tube heat exchangers having a cylindrical cross-section on the
shell side. Huge research efforts have been made during the past decades to improve their thermo-hydraulic performances by intro­
ducing baffles (segmental, helical, inclined), fins of different geometries along the tubes or the shell or by changing the tube
configuration and their arrangement. Though they drastically increase the effectiveness of STHE, baffles introduce problems
commonly encountered in experimental prototypes: increased pressure drop, decreased lifetime due to vibrations and enhanced

* Corresponding author.
E-mail address: marwa.benslimene@enit.utm.tn (M. Ben Slimene).

https://doi.org/10.1016/j.csite.2022.101873
Received 13 August 2021; Received in revised form 28 December 2021; Accepted 12 February 2022
Available online 15 February 2022
2214-157X/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

fouling in recirculation and dead regions. Such STHEs have some disadvantages due mainly to the cylindrical shell geometry. [22]
indicated that having more than two shell passes is unpractical in a standard shell and tube heat exchanger, which was taken up later in
their recommendations by TEMA [40]. Patel et al. [30] listed all the main limitations of cylindrical shaped shell, pointing out that it
contains stagnant/dead zones within the shell, which can lead to corrosion issue. Recently [13], pointed out, in their exhaustive re­
view, that for larger space and combined cycle, a standard shell and tube heat exchanger may become too large and space consuming.
Due to the large number of degrees of freedom (geometrical parameters and operating conditions), many researches recently
focused on the design enhancement of shell and tube heat exchanger using multiobjective algorithms, artificial neural networks, or
metamodels with improved effectiveness and reduced capital cost as objective functions. The so-called MINLP model was developed by
Ref. [34] for the design of STHEs using the Bell-Delaware method and the TEMA standards [40] for the shell side. Their model included
fouling effects and pressure drop and was used to minimize either the heat transfer area or the capital cost. [18] used a genetic al­
gorithm with multiple design parameters to optimize a SHTE with segmental baffles based on the maximization of the field synergy
number. [42] proposed a generalized disjunctive programming model for the optimization of STHE design. It included 12 combina­
tions (four tube side and three shell side geometries). They compared two solvers, namely BARON and DICOPT, to decrease the capital
cost. They showed that their optimization strategy led to a better configuration of STHE compared to conventional design procedures.
[9] combined a CFD analysis using the standard k-ε model and the Taguchi method to investigate various shapes of baffles for STHE
and propose an optimal design. Three-zonal baffles exhibited the lowest pressure drop while keeping high thermal performances. [43]
proposed a very detailed and interesting review on about 100 different definitions of the Performance Evaluation Criterion (PEC) based
either on the first or second laws of thermodynamics. Though many of them are related to each other, PEC numbers remain useful
indicators to quantify the thermo-hydraulic performance of STHE for optimization purpose or even for experiments and CFD
calculations.
Computational Fluid Dynamics (CFD) simulations remain a valuable tool to investigate in details the flow dynamics and associated
heat transfer in STHE before improving their design. It enables indeed to well capture the heat transfer, which occurs in the near-wall
regions, but also possible 3D coherent structures embedded in the mean flow, which may greatly affect the heat transfer. Moreover, it
gives access to dead zones and local heat transfer coefficients, which are usually not captured experimentally, and the fluid flow
maldistribution can also be considered numerically. However, most of the time, authors prefer to increase the design complexity by
increasing the number of tubes or introducing innovative baffle geometries, to the detriment of a well-captured near-wall modeling
and computational efforts and of a global validation. Sometimes, no validation is provided as for [16]; who modeled a fractal
branch-like STHE without besides mentioning the applied turbulence closure. Most of the time, high Reynolds number approaches
based on a k-ε model with standard wall function are preferred. As an example [23], investigated two STHEs with 6 louver baffles and
61 tubes using a RNG k-ε model with standard wall function coupled to a 2.35 million cell mesh grid. They validated their model
against experimental data in terms of the pressure drop distribution on the shell side as a function of the fluid velocity. [4] optimized a
SHTE with segmental baffles using the SolidWorks Flow Simulation. They considered a standard k-ε model with standard wall function
without precising the mesh grid. They validated their approach in terms of pressure drop as well as shell-side heat transfer coefficient
for three operating conditions, getting a maximum discrepancy of 24%. [36] modeled a STHE with a 25% baffle cut and three types of
tubes using within COMSOL. A standard k-ε model with standard wall function associated with a coarse 3D mesh was considered and
the model validated in terms of the pressure drop as a function of the mass flow on the shell side. [24] carried out an experimental and
numerical study of a STHE with a tube bundle (19 tubes doing one pass each) and different innovative baffle arrangements. They used a
realizable k-ε model with standard wall function and a mesh grid composed of 2.32 million cells. For such a complex geometry, this
mesh grid is not sufficient to capture the fluid flow and heat transfer up to the viscous sublayer, which may explain the 17%
discrepancy on the Nusselt number. [41] performed the influence of longitudinal fins on the performance of a shell and helically coiled
tube heat exchanger using a RNG k-ε model. The benefit from using fins is of the same order as the discrepancy between their CFD and
experimental results. Liu et al. [25] compared three turbulence models, namely the RNG and realizable k-ε and k-ω SST models for a
STHE with baffles. Surprisingly, the latter model performs worst but it has to be noticed that the mesh grids and wall resolutions are not
provided. The same mesh grid has been probably used but then, no definitive conclusion can be done. A last example is the numerical
work of [14]; who modeled a STHE with curved segmental baffles with a standard k-ε closure. The wall resolution was not indicated
and though their predictions agree well with experimental data (average deviation <9% in terms of heat transfer coefficient and
pressure drop), their grid convergence analysis was not conclusive and the mesh grid used did not lead to grid-independent solutions.
[2] were one of the few authors to perform large eddy simulations (LES) for a shell and coiled tube heat exchanger with fins on the
external wall of the tube. LES were compared to the realizable k-ε and k-ω models for three test cases at a fixed spatial resolution.
However, as the wall resolution is not mentioned, it was surprising to obtain indistinguishable results between the three models. A
detailed review on the CFD modeling of various heat exchangers has been proposed by Ref. [8]; which has been completed by
Ref. [29].
To the best of the authors’ knowledge, although, the flow dynamics and heat transfer in a square or rectangular shape are discussed
in Ref. [30]; the literature on STHE with a rectangular cross-section on the shell side is scarce, especially by advanced numerical
simulations. These authors addressed the weakness of conventional STHE with a cylindrical shell. They designed a new experimental
set-up with a rectangular shell side and tested three configurations. Besides the facts that it provides the flexibility to increase the
number of shell pass and a complete counter flow may be performed, this novel shell configuration provided, for the same heat transfer
rates, an enhanced Effective Mean Temperature Difference and less heat surface area compared to cylindrical STHE. In industrial
applications, such as, cooling systems, wastewater heat exchangers, air preheaters and combined cycle installations heat exchanger
can be specifically designed to suit customer’s requirements like high heat transfer with less flow area, little maintenance, low pressure
drop and less corrosion problems. All these reasons highlight the increasingly need to investigate the enhancement of the flow

2
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

dynamics and heat transfer in the rectangular shell and tube heat exchanger.
The objectives and novelties of the present study are twofold: (1) modeling the fluid flow and conjugated heat transfer in an in­
dustrial shell and tube heat exchanger with or without baffles using low-Reynolds number closure. Compared to most published works,
this is deemed necessary to well capture the heat transfer in the near-wall regions; (2) demonstrating the benefit from a shell with a
rectangular cross-section compared to a cylindrical one commonly used in the literature.
The paper is organized as follows: the numerical method is first described in Section 2 together with the numerical parameters, the
mesh grid and boundary conditions. Two configurations, namely a concentric heat pipe and a double tube pass shell and tube heat
exchanger (DTP-STHE) with baffles, are then considered as validation test cases in Section 3. The results for a one-rectangular shell –
one-tube (six passes) heat exchanger with and without baffles are discussed in Section 4. It includes the analysis of the fluid flow and
heat transfer in addition to the performance of the heat exchanger quantified using three global metrics for a wide range of the tube
Reynolds number. The paper ends with some concluding remarks and future views in Section 5.

2. Numerical modeling
2.1. Geometrical modeling
Fig. 1 presents a shell-and-tube heat exchanger composed of one rectangular shell in which the fluid does one pass and one tube
doing six passes. The hot fluid is water and flows within the tank through three tubes at an averaged velocity Vh = 1 m/s and inlet
temperature Th = 14◦ C. The Reynolds number based on the tube diameter Dh is Reh = 60 545. On the tube side, glycol flows through
carbon steel pipe of diameter Dc = 48.2 mm at an inlet averaged velocity Vc, which varies between 1 and 10 m/s and an inlet tem­
perature Tc = − 8◦ C. So the Reynolds number for the glycol flow Rec varies between 14 703 and 147 034. Five rectangular baffles
alternatively placed along the top and bottom walls of the shell may be added to increase the performance of the heat exchanger. The
required properties of fluids and solid are given in Table 1, while Table 2 lists the detailed geometrical parameters. In this study, a
rectangular shell and tube heat exchanger with carbon steel tubes is used. The melting point of this steel ranges between 1450 and
1500 ◦ C, the module of elasticity is equal to 200 GPa and the Poisson ratio is 0.29. Note that deformations of the solid walls such as the
ones reported by Refs. [5,7] for the buckling and bending, respectively, of laminated plates are not accounted for in the present
simulations. These are considered as rigid and undeformable.

2.2. Numerical method

The numerical computations are carried out by solving continuity, momentum and energy equations in cylindrical coordinates (r,θ,
z). The fluids are assumed to be Newtonian and incompressible.
The CFX 17.1 solver available in ANSYS Workbench 17 [3] based on the finite volume method is used. Second-order upwind
schemes are employed for the spatial discretization together with second-order implicit schemes for the temporal discretization. The
SIMPLE algorithm enables to overcome the velocity-pressure coupling. All gradients are evaluated using the least squares cell-based
method without skewness correction.
As it will shown in the following section, the temperature gradients are relatively weak such that the Boussinesq approximation
remains valid in both cases. In the turbulent regime, the Shear Stress Transport (SST) k-ω model proposed by Ref. [27] is chosen.
Blending functions are introduced in the transport equation of k and ω to switch from the k-ω in the near wall region to the k-ε out of the
boundary layer. [37] demonstrated its superiority over other two-equation closures and also over Reynolds Stress Models for a tur­
bulent heated pipe flow, which confirmed the former results of [29] for a STHE with tube bundle.
Convergence of the solution is achieved when all the residuals are less than 10− 5 and the mass flow rate is conserved at ± 0.1%.

2.3. Boundary conditions and mesh grid

Prescribed velocity and temperature are imposed at all inlets. The mean inlet velocity and temperature imposed at the shell side are
fixed to Vh = 1 m/s and Th = 14◦ C, respectively, in all calculations. At the pipe inlet, the averaged velocity Vc is varied between 1 and
10 m/s, while temperature remains constant (Tc = − 8◦ C). A turbulent intensity of 5% is imposed at all inlets but results revealed that
the flow distribution and heat transfer are not sensitive to that parameter. A pressure boundary condition is prescribed at both outlets.

Fig. 1. Schematic view of the shell-and-tube heat exchanger with baffles.

3
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

Table 1
Thermophysical properties of materials (Physical Data of Thermal Fluids). Note that the fluid properties are evaluated at their corresponding inlet temperatures.
Sources: [31], and [20].

Water Glycol (MEG30%) Steel

Density ρ (kg.m− 3) 999 1057 7854

Specific heat capacity Cp (J.kg− 1.K− 1) 4198 3611 434
Thermal conductivity λ (W.m− 1.K− 1) 0.592 0.464 60.5
3 3
Dynamic viscosity μ (Pa.s) 1.11 × 10− 3.465 × 10− –

Table 2
Geometrical characteristics of the heat exchanger.

Parameters Values

Width of the shell [mm] 1200

Length of the shell [mm] 5000
Inlet and outlet diameters of the shell tube [mm] 80
Outer diameter of the tube [mm] 48.2
Thickness of the tube wall [mm] 5
Effective length of the tube [mm] 4200
Pitch of the tube [mm] 750
Number of tubes [-] 1
Length of the baffles [mm] 900
Width of the baffles [mm] 5
Spacing between two baffles [mm] 900
Number of baffles 5

No slip wall boundary conditions are adopted on all walls. The shell and its inlet and outlet pipes are assumed as adiabatic. Note
also that conductive heat transfer within the pipe wall is accounted for.
For the grid independence study, the simulation was run at Rec = 3.67 × 104 and Reh = 6 × 104, keeping in mind that the mesh
nodes need to be small enough to resolve the wall boundary layers and meet the requirements of the k-ω SST model. Five different grid
sizes were tested to find out its influence on the average outlet hot temperature. It decreases from 286.7 K to 286.66 K by increasing the
mesh size. It has been found that there is no significant change in average outlet hot temperature beyond the grid size with 13 769 418
elements. This mesh has been selected in the following for all simulations.
A mesh grid with tetrahedral elements is used for the rectangular shell-and-tube heat exchanger. The details about the mesh grid
are given in Table 3 and a zoom on the mesh grid at the junction between the tubes and the shell is displayed on Fig. 2.
For all calculations, the associated computational time step is fixed to 10− 3 s leading to a maximum value of the CFL number lower
than 1 for any set of operating conditions.

3. Validation test cases

3.1. Annular turbulent flows in a pipe
The first validation concerns the isothermal turbulent flow between concentric cylinders investigated numerically using direct
numerical simulations (DNS) by Ref. [10] and experimentally by Ref. [28]. It has been later extended to the non-isothermal case using
DNS by Refs. [11,21].
The flow is confined between two stationary cylinders of length Lz and radius R1 and R2, respectively. The Prandtl number Pr is
equal to 0.71 (air). The Reynolds number based on the mean inlet axial velocity Um = 0.66 m/s and the hydraulic diameter Dh = 4δ is
ReDh = Um Dh /ν = 8900, where ν is the fluid kinematic viscosity. Two radius ratios R* = R1 /R2 are considered in the following: R* =
0.1 and R* = 0.5. In the isothermal case, the computational length in the axial direction is Lz = 15δ for R* = 0.1 and Lz = 18δ for R* =

Table 3
Details of the mesh grid for the shell and tube heat exchanger with and without baffles.

Hot fluid Cold fluid Solid

Without With Without With Without With

baffles baffles baffles

Smallest cell height (m) 1 × 10− 2 4 × 10− 2 4 × 10− 2

Thickness of the first layer (m) 10–4 10–4 10–4
Total number of elements 11 849 667 11 475 434 1 473 066 1472 575 446 685
Maximum value of y+ 6.38 5.2 1.68 1.34 –
Skewness Avg 0.2 0.19 –
Max 0.85 0.86 0.82 0.81
Orthogonal quality Avg 0.87 0.85 0.9 –
Min 0.19 0.1 0.21 0.23

4
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

Fig. 2. Zoom on the mesh grid at the junction between the tubes and the shell.

0.5 and the walls are supposed to be adiabatic. In the non-isothermal case, the computational length in the axial direction is Lz = 25δ
and Lz = 30δ for R* = 0.1 and R* = 0.5 respectively. The inner and outer walls are then subject to uniform heat fluxes denoted qi * and
qo * respectively, such that qo * = qi * = 1520W/m2 .
In both cases, no-slip boundary conditions are also imposed at the inner and outer walls for the velocity components. Periodic
boundary conditions for the velocity components and the temperature are also imposed in the streamwise direction z. To properly
apply periodic conditions, the following transformation proposed by Ref. [26] is done for the temperature field:
⃒ ⃒
⃒∂Tw ⃒
T = Tp − ⃒⃒ ⃒ (1)
∂z ⃒

where Tp is the periodic temperature field and:

⃒ ⃒
⃒∂Tw ⃒ qw
⃒ ⃒
⃒ ∂z ⃒ = ρC U δ (2)
p m

An unstructured mesh grid composed of tetrahedral elements was built with ICEM. 20 prismatic layers are imposed on both walls to
guarantee the appropriate modeling up to the viscous sublayer. The ratio for the mesh refinement was set to 1.1. The main charac­
teristics of the mesh grid used for the annular pipe cases are summarized in Table 4. The chosen mesh grids lead all to a maximum wall
coordinate y+ around 0.2, which is well suited for low-Reynolds number model.
The computational time step used is 1.2 × 10− 3 s for both radius ratios R* = 0.1 and R* = 0.5.
The velocity distributions normalized by the bulk inlet axial velocity Um are presented in Fig. 3 for the two radius ratios R*. The
present model compares quite well with the DNS data of [10] for R* = 0.1 and R* = 0.5. The profiles are asymmetrical with higher
axial velocities along the outer wall in both cases. This asymmetry is besides more pronounced for the lowest radius ratio. The present
model is also capable of well reproducing the thin boundary layers developed along both walls.
To investigate in more details the near-wall regions, Fig. 4 shows the mean wall velocity profiles for R* = 0.1 and 0.5 obtained by
the present model and compared with the experimental results of [28]. The dimensionless axial velocity Vz + = Vz /Uτ is plotted against
√̅̅̅̅̅̅̅̅̅̅
the wall coordinate y+ = Δy.Uτ /ν, where Δy is the size of the cell in the radial direction, Uτ = τw /ρ is the friction velocity and τw =
⃒
⃒
μ∂∂Uy ⃒⃒ the wall shear stress. The present model perfectly predicts the linear profile Vz + = y+ in the viscous sublayer, for y+ ≤ 5 (resp.
y=0
6) at R* = 0.1 (resp. 0.5). On the outer wall, the velocity profile recovers the law of the wall Vz + = 1κ ln y+ + C+ in the log-layer region,
for y+ ≥ 35 at R* = 0.1. One recovers the von Karman constant κ ~ 0.4 and the smooth wall constant C+ = 5.5, for fully developed
turbulent flows over smooth walls [44]. At the same time, a weak velocity deficit is obtained on the inner wall, which may be attributed
to the curving effect, generated by the decrease of the radius of the inner cylinder [10]. For the highest radius ratio R* = 0.5, the
present model agrees fairly well with the experiments of [28] and can model the viscous sublayer and the log-law region without any
velocity deficit along the inner wall.
Fig. 5 displays the distributions of the dimensionless temperature θ+ = (Tw − T)/Tτ in the near-wall regions, with Tτ = qw / (ρCp Uτ )

Table 4
Mesh grid characteristics for the annular pipe case.

R* Lz Smallest cell height (m) Thickness of the first layer (m) Total number of elements Max (y+)
− 4 − 5
0.1 15 δ 1.04 × 10 5 × 10 5 236 081 0.191
4
25 δ 1.57 × 10− 8 666 341 0.194
4 5
0.5 18 δ 1.24 × 10− 5 × 10− 7 850 188 0.204
4
30 δ 1.87 × 10− 13 106 073 0.203

5
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

Fig. 3. Mean axial velocity profiles for (a) R* = 0.1 and R* = 0.5. Comparisons with the DNS of [10].

Fig. 4. Mean wall velocity profiles for R* = 0.1 and 0.5. Comparisons with the experimental results of [28].

Fig. 5. Mean wall temperature profiles for (a) R* = 0.1 and (b) R* = 0.5. Comparisons with the DNS results of [21].

6
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

the friction temperature. The law of wall for the temperature profile is recovered here on both walls with a linear profile θ+ = Pr.y+ (Pr
= 0.7) in the viscous sublayer and a logarithmic distribution θ+ = 2.62 ln y+ + 2.91 in the log-law layer. On the outer wall, the model
shows an excellent agreement with the DNS results of [21]. On the contrary, along the inner wall, a temperature deficit is observed
starting from the buffer region (y+ > 10) for R* = 0.1. This is not observed for the highest value of the radius ratio for R* = 0.5.
Table 5 summarizes the mean flow and thermal parameters, namely the friction Reynolds number Reτ = Uτ δ/ν, the skin friction
( )
coefficient Cf = τw / 12 ρUm 2 and the Nusselt number based on the hydraulic diameter NuDh = h Dh /λ as a function of the radius ratio.

R* and the bulk Reynolds number based on δ, Reδ = Um δ/ν. One recalls that the Reynolds number based on the hydraulic diameter
Dh remains constant in all cases: ReDh = 8900 as well as the normalized heat flux: q* = 1. The present model agrees fairly well on both
walls against the DNS results of [11]. The deviations remain relatively small for the hydrodynamic field and increase up to 11.7% for
the Nusselt number along the outer wall for the highest radius ratio.
To fully validate the present model, one considers also the turbulent heat flux Vr θ , which can be approximated by Vr θ = − αt ∂∂θr ,
′ ′ ′ ′

where αt = νt /Prt is the eddy diffusivity. The distributions of this turbulent heat flux are displayed on Fig. 6 for both radius ratios as a
function of the global coordinate y/δ and the wall coordinate y+. In all cases, the present model compares well with the DNS of [11]
with only a small shift between the curves. The locations and values of the peaks are particularly well predicted as well as the shapes of
the profiles in the near-wall regions.

3.2. Double tube pass shell and tube heat exchanger (DTP-STHE) with baffles
The present solver is also validated regarding the counter current DTP-STHE with segmental baffles considered experimentally and
numerically by Ref. [19]. The main geometrical parameters are recalled in Table 6. Note that the shell has a rectangular cross-section.
The inlet mass flow rate varies between 1.1 and 1.9 kg/s at the inlet of the shell side with a temperature fixed to 353.15 K. The mass
flow in the tubes is set to 0.2 kg/s with a temperature of 293.15 K. Pressure is imposed at the shell and tube outlets. All walls are
non-slip and adiabatic. The working fluid is water for both shell and tubes. The computational domain is meshed with 5.5 millions of
unstructured tetrahedral elements including refinement in the inlet and outlet regions and the near-wall tubes.
Fig. 7a displays the variation of the heat transfer coefficient on the shell side as a function of the Reynolds number based on the
mass flow rate imposed at the shell inlet. The present CFD results are compared to the experimental data of [19]and their numerical
predictions using a RNG k-ε model with standard wall functions. As expected, the overall heat transfer coefficient on the shell side
increases by increasing the shell Reynolds number. More interestingly, the present k-ω SST model predicts quite well the profile of the
heat transfer coefficient for this range of Reynolds number. Despite some weak discrepancies, which may be in the uncertainty range of
the experiments (not precised), the present model greatly improves the predictions of their RNG k-ε model. First, their mesh is slightly
coarser with 4.6 elements, but it can be mainly attributed to the poor performance of the standard wall function, whereas the thin
boundary layers are accounted for here with a k-ω closure. The present CFD can be then used confidently to model the turbulent flow
and associated conjugate heat transfer in a rectangular shell and tube heat exchanger.
The design of a cylindrical-shape shell leads indeed to some limitations such as, corrosion problems due to dead zones, long heat
exchanger circuits for optimum conditions, high pressure drops, limited number of shell passes, working volume and space require­
ment, etc (TEMA). A rectangular shell provides the flexibility to increase the number of shell passes and a complete counter flow can be
achieved due to the rectangular form of the shell. While a maximum of two shell passes are possible for a cylindrical STHE, the number
of shell passes for a rectangular STHE may be equal to the number of tube passes. The objective of the present study is then to
characterize the performance of a rectangular shell and tube heat exchanger.
Fig. 7b displays the evolution of the heat transfer rate Qh per effective pumping power P0 as a function of the mass flow rate Gs on
the shell side. The present results obtained using the k-ω SST model for a rectangular STHE are compared to the experimental data of
[19] for a double-tube-pass shell and tube heat exchanger equipped with flower baffles. Both shells have the same length (500 mm) and
the width of the rectangular one is equal to the diameter (87 mm) of the cylindrical one to keep the same overall volume required to
install the heat exchanger. All operating parameters are set to the same values between the two configurations. [19] demonstrated that
such baffles lead to better overall performance compared to helical and segmental baffles. It can be seen here that these two shell
geometries lead to similar values of the Qh/P0 ratio for this range of Gs values. The present results confirm then that a rectangular shape
shell leads to improved overall performance compared to their cylindrical counterpart. A rectangular shell and tube heat exchanger

Table 5
Mean flow and thermal parameters for the non-isothermal turbulent flow in the concentric annular case. Comparisons with the DNS results of [11].

Present results (% deviation) [11] Present results (% deviation) [11]

R* 0.5 0.5 0.1 0.1

Reδ 3355 3355 3487 3487
Reτ,i 156 (2.6%) 152 176 (1.7%) 179
Reτ,o 148 (2.8%) 144 130.65 (8.6%) 143
Cf,i 0.00988 (1.8%) 0.01006 0.0125 (3.7%) 0.0130
Cf,o 0.00883 (3.3%) 0.00913 0.00863 (4.6%) 0.00825
NuDh ,i 31.06 (11.3%) 35.03 64 (2.3%) 62.55
NuDh ,o 27.4 (3.9%) 28.51 24.39 (11.7%) 27.63

7
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

Fig. 6. Turbulent wall-normal heat flux for R* = 0.1 and 0.5: (top) global coordinates; (bottom) zoomed view in wall coordinates. Comparison with the DNS of [11].

Table 6
Geometrical parameters for the DTP-STHE of [19].

Parameters Values

Inner diameter of the shell [mm] 87

Length of the heat exchanger [mm] 500
Outer diameter of the tube [mm] 19
Thickness of the tube [mm] 1.5
Effective length of the tube [mm] 365
Pitch of the tube [mm] 24
Number of tubes 6
Number of baffles 5
Spacing of the baffle [mm] 57
Thickness of the baffle [mm] 3

will be then considered in the following.

4. Results and discussion for the rectangular shell and tube heat exchanger
The results are discussed hereafter in terms of the mean flow and temperature fields and performance metrics by varying the pipe
Reynolds numbers between 1.03 × 103 and 1.47 × 105 and for a rectangular shell equipped or not with baffles.

4.1. Mean flow and thermal fields

The results are mainly discussed here for the tube side, which corresponds to the cold flow of ethylene glycol. Fig. 8a presents the
mean axial velocity profile close to the wall for three Reynolds numbers. All the profiles have been extracted in the second tube pass at
a position z* = z/L = 0.07, where L is the tube length. As for the concentric pipe flows discussed in Section 3.1, once again, the
predictions of the k-ω SST follow quite well the law of the wall, from the viscous sublayer to the log region for the three values of Rec. It

8
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

Fig. 7. Comparison with [19] for a shell and tube heat exchanger: (a) Variations of the shell side heat transfer coefficient versus the Reynolds number; (b) Variations of
the heat transfer rate Qh per effective pumping power P0 versus the mass flow rate Gs on the shell side (flower baffles).

Fig. 8. Mean wall (a) axial velocity and (b) temperature profiles within the second pass of the tube at z* = 0.07. Results obtained without baffles for three Reynolds
numbers: Rec = 14703 (− ), 73517 (− ) and 147034 (− ), and comparison with (—) the law of the wall.

confirms in particular that the mesh grid within the tube is refined enough to capture such profiles. The corresponding dimensionless
mean temperature profiles are displayed in Fig. 8b. All profiles respect the law of the wall in the thermal sublayer for the three
Reynolds numbers. In the logarithmic region, there is a clear deficit of the temperature profile, which increases when increasing the
Reynolds number. It may be partly explained by the constant value of turbulent Prandtl number (Prt = 0.85) imposed in the present
model, whereas different authors as [35] showed that it may slightly vary in the log region. It besides decreases by increasing the

Table 7
Mean parameters on the tube side for the shell and tube heat exchanger.

Without baffles Vc Rec Reδ Reτ Cf Nuc

0.7 m/s 10 292.4 1617.37 95.2 0.028 69.82
1 m/s 14 703.43 3197.98 175.15 0.024 75.8
2.5 m/s 36 758.58 6469.5 354.66 0.022 96.3
5 m/s 73 517.17 14 335.81 660.81 0.017 111
7 m/s 102 924 19 886.3 861 0.015 115.14
10 m/s 147 034 28 818.6 1201.87 0.014 118.73
With baffles 0.7 m/s 10292.4 1617.37 106.96 0.035 162.94
1 m/s 14 703.43 2426 179.83 0.03 192.5
2.5 m/s 36 758.58 6469.5 354.3 0.024 305.57
5 m/s 73 517.17 14 005 700.16 0.02 265.51
7 m/s 102 924 19 886.3 969.11 0.019 313.24
10 m/s 147 034 28 818.6 1328.18 0.017 324.93

9
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

Prandtl number to reach approximately Prt = 1 for Pr = 1. Unfortunately, there is a lack of data for fluids like ethylene glycol at Pr = 27.
The second explanation comes from the lower resolution when the Reynolds number increases (at constant mesh grid) as explained by
Ref. [6].
Table 7 summarizes the main mean parameters obtained on the tube side for the whole range of Reynolds number considered here
with and without baffles. One recalls that 5 rectangular baffles are added, and their geometrical characteristics are given in Table 2. On
the tube side, the bulk Reynolds number Rec is varied between 10292 and 1.47 × 105, which corresponds to frictional Reynolds
numbers Reτ between 95 and 1202 and between 107 and 1328 for the cases without and with baffles, respectively.
The variations of the mean friction factor for the refrigerant fluid on the tube side as a function of the pipe Reynolds number are
assessed by comparing the predictions of both shell and tube heat exchanger with and without baffles. The average friction factor is
expressed by Equation (3) and the present numerical values are compared to the Blasius correlation for a turbulent pipe flow given by
Equation (4):
8τm
Cf ,c = (3)
ρVc 2

Cf ,c = 0.316 × Re−c 0.25 (4)

The profiles of the friction factor for both configurations with and without baffles as a function of the Reynolds number in the tube
are displayed in Fig. 9a. A reasonable agreement between the numerical results and theoretical values calculated with the Blasius
equation is observed. The friction factor decreases with increasing Rec in the two configurations with and without baffles. The wall
shear stress increases then slower than the inlet velocity. However, the configuration with baffles predicts higher values of the friction
factor compared to the configuration without baffles, but it gets a good approaching to the Blasius equation.
Fig. 9b shows the distributions of the Nusselt number on the tube side Nuc as a function of the inlet Reynolds number Rec with and
without baffles. As expected, the Nusselt number increases with increasing Reynolds number. It can be noticed that it increases very
slowly for the case without baffles reaching almost an asymptotical value close to Nuc = 120 at 1.47 × 105. With baffles, Nuc increases
much faster. Baffles play their role by enhancing the mixing in the shell and then increasing the heat transfer rate with the pipe. The
enhancement ratio ER (the ratio between the Nuc values with and without baffles) varies from 2.33 to 2.75 for this range of Rec values,
with an average value of ER = 2.64.
The preponderant influence of the five rectangular baffles is confirmed by looking at the temperature contours in the medium plane
of the shell, as displayed in Fig. 10. One recalls that the inlet velocity imposed in the three inlet tubes of the shell for the hot water is
fixed to Vh = 1 m/s. At Rec = 1.47 × 104, temperature is almost homogeneously distributed within the shell with a temperature
difference for the shell side of only ΔTh = 0.23 K for the case without baffle. By increasing the pipe Reynolds number by a factor of 10,
this temperature difference increases slightly up to 0.7 K. By adding baffles in the shell, ΔTh increases from 2.2 K at Rec = 1.47 × 104 to
6.6 K at 1.47 × 105. The Nusselt number Nuh on the shell side is then proportional to (Rec)0.48.

4.2. Performance of the rectangular shell and tube heat exchanger with and without baffles
The results are discussed here in terms of three global performance criteria: Performance Evaluation Criterion (PEC), the field
synergy number and the effectiveness of the heat exchanger.
The PEC number on the tube side is defined as the ratio of heat flow rate transferred to the required pumping power [37]:

Fig. 9. Variations of the (a) friction faction and the (b) Nusselt number as a function of the pipe Reynolds number. Comparison between the cases with and
without baffles.

10
 M. Ben Slimene et al.
11

Case Studies in Thermal Engineering 32 (2022) 101873

Fig. 10. Temperature contours in the medium plane of the shell for three values of the pipe Reynolds number: (a,b) Rec = 1.47 × 104, (c,d) Rec = 7.35 × 104 and (e,f) Rec = 1.47 × 105. Results obtained without (a,c,e) and with
(b,d,f) baffles.
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

ṁc Cp,c ΔTc

PEC = (5)
V̇c Δpc

Where ṁc and V̇c are the mass and volumetric flow rates within the pipe, respectively, ΔTc and Δpc are the temperature and pressure
differences between the pipe outlet and inlet, respectively.
The field synergy number Fc on the tube side may be employed for quantifying the dimensionless heat source strength over the
entire convection domain. The larger the Fc parameter, the better the synergy between the velocity and temperature fields is. It writes
[12,17]:
Nuc
Fc = (6)
Rec Prc
The effectiveness of the heat exchanger ε is defined as:
{( )/( ) }
Q T − Th,o)/( Th,i − Tc,i) if min(ṁCp) = (ṁCp)h
ε= = ( h,i (7)
Qmax Tc,o − Tc,i T h,i − Tc,i if min( ṁCp) = (ṁCp)c

Fig. 11a displays the PEC number as a function of the pipe Reynolds number for the two configurations with and without baffles.
The presence of baffles increases the PEC number significantly, especially at lower Reynolds numbers. In the two cases, PEC decreases
with increasing Rec. It confirms the former results of [15] obtained for the convective heat transfer of SiO2/water nanofluids in a
heated pipe. They showed indeed that PEC decreases exponentially with the bulk Reynolds number for both pure water and nanofluid
flows. On the contrary [4], reported numerically that PEC remains almost constant for five combinations of baffle and tube types when
the mass flowrate on the shell side increases.

Fig. 11. Variations of the three-performance metrics as a function of the pipe Reynolds number: (a) PEC, (b) field synergy number Fc and (c) effectiveness of the heat
exchanger ε. Comparison between the cases with and without baffles.

12
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

The field synergy number Fc corresponds to the I factor introduced by Ref. [17]. Fig. 11b shows the variations of Fc as a function of
the pipe Reynolds number. For both configurations, it decreases when Rec increases, which proves the former results of [17] for
laminar flows in various configurations including a stagnation point, a boundary layer plate flow and a duct flow. For these last ge­
ometries, Fc scales with Re− 1/2. [18] used this parameter based on the shell side characteristics to optimize a shell-and-tube heat
exchanger with baffles. For one of their set of operating conditions, they reported an initial value of 0.0021 (ε = 0.576) and an optimal
one of 0.003 (ε = 0.6044). Though one considers here the Fc number for the tube side, these values are quite comparable to the
maximum value obtained at the lowest Reynolds number for the case with baffles (Fig. 11b).
More interestingly is the variation of the heat exchanger effectiveness as a function of the pipe Reynolds number displayed in
Fig. 11c. As expected, the effectiveness decreases by increasing the pipe Reynolds number, its maximum value being reached at Rec =
1.03 × 103. This is a direct consequence of its definition. At a given Reynolds number Rec = 1.03 × 103, the presence of baffles in­
creases ε from 49% to about 76%. Though the operating conditions are different, two interesting points must be noticed: (1) the
maximum effectiveness is achieved when the other metrics are maximum too, confirming the results of [18]; (2) though the conditions
are somehow different, the present heat exchanger is more efficient than the optimal one (cylindrical shell) proposed by Ref. [18];
while the synergy field number remains always lower than in their case. It may lead to two conclusions, the first one being that Fc may
be not the most reliable metrics to characterize the heat exchanger performance and so to optimize it. The second conclusion is that this
last result confirms that rectangular shells are usually more efficient than their cylindrical counterpart as already discussed in Section
3.2.

5. Conclusions
Turbulence modelling using a k-ω SST closure has been carried to study the turbulent flow and coupled heat transfer in a rect­
angular shell and tube heat exchanger. The numerical solver has been first validated for fully developed turbulent flows in a concentric
annular pipe with two radius ratios R* = 0.1 and 0.5 at a Reynolds number ReDh = 8900 and Prandtl number Pr = 0.71. The transverse
curvature effects are emphasized, and the near-wall thermal structures close to the inner and outer walls are identified by applying a
heat flux ratio q* = 1. It compared quite well with former DNS and experimental results, proving the ability of the model to solve the
turbulent flow and heat transfer within the pipes. The second validation concerned a double tube pass shell and tube heat exchanger
with baffles studied experimentally by Ref. [19]. Once again, a very close agreement was obtained, the present k-ω SST closure
improving besides the predictions of the RNG k-ε model. The most interesting feature was that the rectangular cross-section of the shell
enabled to get the same performance as the optimal heat exchanger proposed by Ref. [19] and equipped with flower baffles.
The calculations have been then extended to the case of an industrial shell and tube heat exchanger (one rectangular shell, one tube
with six passes). With its rectangular shell form, it represents a realistic design for the dairy industry, which was not really addressed in
the literature. The glycol Reynolds number in the tube side has been varied from 1.03 × 103 to 1.47 × 105. The maximum effectiveness
ε = 0.76 was obtained at the lowest Reynolds number for a shell equipped with baffles. Baffles enable to increase the residence time of
the hot fluid on the shell side while inducing large recirculations, which improve mixing and consequently homogenize the fluid
temperature. Though the PEC number led to similar conclusions, the field synergy number appeared as a less reliable metrics to
optimize heat exchanger. The overall high effectiveness of the present design may be attributed here to the rectangular cross-section
shell, which, by comparisons with literature data, outperforms its cylindrical counterpart.
Future work includes the geometrical optimization the rectangular shell and tube heat exchanger equipped with bio-inspired
baffles. The optimization will be achieved using Dakota, which offers design of experiments and metamodels and based on the
maximization of the heat exchanger efficiency and the PEC number. Entropy generation analysis as proposed recently by Ref. [33] for a
corrugated channel will be also performed to better highlight the main locations of exergy losses within the STHE.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgements
S.P. would like to thanks the NSERC chair on industrial energy efficiency established at Université de Sherbrooke in 2019 with the
financial support of Hydro-Québec, Natural Resources Canada and Emerson. The calculations have been performed using the su­
percomputer Mammouth Parallèle 2 from Compute Canada, which is also gratefully acknowledged.

Nomenclature

Cp Specific heat capacity at constant pressure (J. K− 1 .kg− 1 )

Cf Skin friction coefficient (− )
Dh Hydraulic diameter (m)
h Heat transfer coefficient (W.m− 1 .K− 1 )
LZ Length in the axial direction (m)
NuDh Nusselt number (− )

13
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

P Pressure (Pa)
ΔP Pressure drop (Pa)
Pr Prandtl number (− )
Prt Turbulent Prandtl number (− )
q* Wall heat flux ratio (− )
qi * , qo * Wall heat flux at the inner and the outer wall, respectively (W.m− 2 )
qw Wall heat flux (W.m− 2 )
r, θ, z Spatial coordinates in the r, θ, z directions, respectively
R* Radius ratio (− )
R1 , R2 Radii of the inner and outer walls, respectively (m)
ReDh , Reδ , Reτ Reynolds numbers based on the bulk mean velocity, laminar maximum velocity and friction velocity, respectively
(− )
T Temperature (K)
Tw , Tp , Tτ Wall, periodic and friction temperatures, respectively (K)
Tc , Th Cold inlet and hot inlet temperatures, respectively (K)
y Distance from the wall (m)
vr , vθ , vz Fluctuating velocities in the r, θ, z directions, respectively (m.s− 1 )
Vr , Vθ , Vz Mean velocities in the r, θ, z directions, respectively (m.s− 1 )
Vh , Vc Hot inlet and cold inlet velocities, respectively (m.s− 1 )
Um , Uc , Uτ Bulk mean, laminar maximum and friction velocities, respectively (m.s− 1 )

Greek Symbols
μ Dynamic viscosity (Pa.s)
γ Kinematic viscosity (m2 .s− 1 )
γt Eddy viscosity (m2 .s− 1 )
αt Eddy diffusivity (m2 .s− 1 )
ρ Density (kg.m− 3 )
λ Thermal conductivity (W.m− 1 .K− 1 )
τw Average wall shear stress (Pa)
δ Distance between the inner and outer walls (m)
κ Von Karman constant (− )

Superscripts
Fluctuating component
′
()
( )+ Dimensionless component

Acronyms
CFD Computational Fluid Dynamics
DNS Direct Numerical Simulation
LES Large Eddy Simulation
RANS Reynolds-Averaged Navier-Stokes

References
[1] M. Abou Elmaaty, A.E. Kabeel, M. Mahgoub, Corrugated plate heat exchanger review, Renew. Sustain. Energy Rev. 70 (2017) 852–860.
[2] A. Alimoradi, M. Olfati, M. Maghareh, Numerical investigation of heat transfer intensification in shell and helically coiled finned tube heat exchangers and
design optimization, Chem. Eng. Process: Process Intensification 121 (2017) 125–143.
[3] ANSYS Inc, ANSYS CFX-Pre User’s Guide, 2013.
[4] A.A.A. Arani, R. Moradi, Shell and tube heat exchanger optimisation using new baffle and tube configuration, Appl. Therm. Eng. 157 (2019), 113736.
[5] T. Becheri, K. Amara, M. Bouazza, N. Benseddiq, Buckling of symmetrically laminated plates using nth-order shear deformation theory with curvature effects,
Steel Compos. Struct. 21 (6) (2016) 1347–1368.
[6] R. Bergant, I. Tiselj, Simulations of the Near-Wall Heat Transfer at Medium Prandtl Numbers, International Conference on Nuclear Energy for New Europe,
Portoroz, 2003.
[7] M. Bouazza, T. Becheri, A. Boucheta, N. Benseddiq, Bending behavior of laminated composite plates using the refined four-variable theory and the finite element
method, Earthq. Struct. 17 (3) (2019) 257–270.
[8] M.M.A. Bhutta, N. Hayat, M.H. Bashir, A.R. Khan, K.N. Ahmad, S. Khan, CFD applications in various heat exchangers design: a review, Appl. Therm. Eng. 32
(2012) 1–12.
[9] N. Bicer, T. Engin, H. Yasar, E. Buyukkaya, A. Aydin, A. Topuz, Design optimization of a shell-and-tube heat exchanger with novel three-zonal baffle by using
CFD and taguchi method, Int. J. Therm. Sci. 155 (2020), 106417.
[10] S.Y. Chung, G.H. Rhee, H.J. Sung, Direct numerical simulation of turbulent concentric annular pipe flow Part 1: flow field, Int. J. Heat Fluid Flow 23 (2002)
426–440.

14
M. Ben Slimene et al. Case Studies in Thermal Engineering 32 (2022) 101873

[11] S.Y. Chung, H.J. Sung, Direct numerical simulation of turbulent concentric annular pipe flow Part 2: heat transfer, Int. J. Heat Fluid Flow 24 (2003) 399–411.
[12] Y. Cui, Y. Zhang, W. Wang, B. Li, B. Sunden, Unified formula for the field synergy principle, Numer. Heat Tran. B: Fundamentals 77 (4) (2020) 287–298.
[13] E. Edreis, A. Petrov, Types of heat exchangers in industry, their advantages and disadvantages, and the study of their parameters, IOP Conf. Ser. Mater. Sci. Eng.
963 (2020), 012027 article.
[14] E.M.S. El-Said, A.H. Elsheikh, H.R. El-Tahan, Effect of curved segmental baffle on a shell and tube heat exchanger thermohydraulic performance: numerical
investigation, Int. J. Therm. Sci. 165 (2021), 106922.
[15] S. Ferrouillat, A. Bontemps, J.P. Ribeiro, J.A. Gruss, O. Soriano, Hydraulic and heat transfer study of SiO2/water nanofluids in horizontal tubes with imposed
wall temperature boundary conditions, Int. J. Heat Fluid Flow 32 (2) (2011) 424–439.
[16] N. Foster, D. Sebastia-Saez, H. Arellano-Garcia, Fractal branch-like fractal shell-and-tube heat exchangers: a CFD study of the shell side performance, IFAC-
PapersOnLine 52–1 (2019) 100–105.
[17] Z.Y. Guo, D.Y. Li, B.X. Wang, A novel concept for convective heat transfer enhancement, Int. J. Heat Mass Tran. 41 (14) (1998) 2221–2225.
[18] J. Guo, M. Xu, L. Cheng, The application of field synergy number in shell-and-tube heat exchanger optimization design, Appl. Energy 86 (2009) 2079–2087.
[19] L. He, P. Li, Numerical investigation on double tube-pass shell-and-tube heat exchangers with different baffle configurations, Appl. Therm. Eng. 143 (2018)
561–569.
[20] F. Incropera, D. DeWitt, T. Bergman, A. Lavine, Fundamentals of Heat and Mass Transfer, sixth ed., John Wiley & Sons, 2007.
[21] N. Kasagi, Y. Tomita, A. Kuroda, Direct numerical simulation of passive scalar field in a turbulent channel flow, Trans. ASME: J. Heat Tran. 114 (3) (1992)
598–606.
[22] D.Q. Kern, Process Heat Transfer, McGraw-Hill Book Company Japon, Tokyo, 1983.
[23] Y. Lei, Y. Li, S. Jing, C. Song, Y. Lyu, F. Wang, Design and performance analysis of the novel shell-and-tube heat exchangers with louver baffles, Appl. Therm.
Eng. 125 (2017) 870–879.
[24] N. Li, J. Chen, T. Cheng, J.J. Klemes, P.S. Varbanov, Q. Wang, W. Yang, X. Liu, M. Zeng, Analysing thermal-hydraulic performance and energy efficiency of shell-
and-tube heat exchangers with longitudinal flow based on experiment and numerical simulation, Energy 202 (2020), 117757.
[25] Y. Liu, J. Wen, S. Wang, J. Tu, Numerical investigation on the shell and tube heat exchanger with baffle leakage zones blocked, Int. J. Therm. Sci. 165 (2021),
106959.
[26] D.M. Lu, G. Hetsroni, Direct numerical simulation of a turbulent open channel flow with passive heat transfer, Int. J. Heat Mass Tran. 38 (17) (1995)
3241–3251.
[27] F.R. Menter, Two-equation eddy-viscosity turbulence models for engineering applications, AIAA J. 32 (8) (1994) 1598–1605.
[28] J.M. Nouri, H. Umur, J.H. Whitelaw, Flow of Newtonian and non-Newtonian fluids in concentric and eccentric annuli, J. Fluid Mech. 253 (1993) 617–641.
[29] E. Pal, I. Kumar, J.B. Joshi, N.K. Maheshwari, CFD simulations of shell-side flow in a shell-and-tube type heat exchanger with and without baffles, Chem. Eng.
Sci. 143 (2016) 314–340.
[30] V. Patel, R. Patel, V. Savsani, Novel heat exchanger design with rectangular shell geometry, in: Proceedings of the ASME 2014 International Mechanical
Engineering Congress and Exposition, 2014. Montreal, Canada.
[31] Physical data of thermal fluids, in: www.celsius-process.com.
[32] S. Rashidi, M. Kashefi, C.K. Kyung, S.A. Omid, Potentials of porous materials for energy management in heat exchangers – a comprehensive review, Appl. Energy
243 (2019) 206–232.
[33] S. Rashidi, M. Akbarzadeh, R. Masoodi, E.M. Languri, Thermal-hydraulic and entropy generation analysis for turbulent flow inside a corrugated channel, Int. J.
Heat Mass Tran. 109 (2017) 812–823.
[34] M.A.S.S. Ravagnani, J.A. Caballero, A MINLP model for the rigorous design of shell and tube heat exchangers using the TEMA standards, Trans. IChemE, Part A
85 (A10) (2007) 1423–1435.
[35] L. Redjem-Saad, M. Ould-Rouiss, G. Lauriat, Direct numerical simulation of turbulent heat transfer in pipe flows: effect of Prandtl number, Int. J. Heat Fluid
Flow 28 (2007) 847–861.
[36] M.R. Safarian, F. Fazelpour, M. Sham, Numerical study of shell and tube heat exchanger with different cross-section tubes and combined tubes, Int. J. Energy
Environ. Eng. 10 (2019) 33–46.
[37] G. Sekrani, S. Poncet, P. Proulx, Modeling of convective turbulent heat transfer of water-based Al2O3 nanofluids in an uniformly heated pipe, Chem. Eng. Sci.
176 (2018) 205–219.
[38] N.D. Shirgire, P.V. Kumar, Review on comparative study between helical coil and straight tube heat exchanger, J. Mech. Civ. Eng. 8 (2) (2013) 55–59.
[39] W. Srimuang, P. Amatachaya, A review of the application of heat pipe heat exchangers for heat recovery, Renew. Sustain. Energy Rev. 16 (2012) 4303–4315.
[40] Tubular Exchanger Manufacturer Association (TEMA). http://www.wermac.org/equipment/heatexchanger_part5.html.
[41] A.D. Tuncer, A. Sozen, A. Khanlari, E.Y. Gürbüz, H.I. Variyenli, Upgrading the performance of a new shell and helically coiled heat exchanger by using
longitudinal fins, Appl. Therm. Eng. 191 (2021), 116876.
[42] Z. Yang, Y. Ma, N. Zhang, R. Smith, Design optimization of shell and tube heat exchangers sizing with heat transfer enhancement, Comput. Chem. Eng. 137
(2020), 106821.
[43] M. Yilmaz, O. Comakli, S. Yapici, O. Nuri Sara, Performance evaluation criteria for heat exchangers based on first law analysis, J. Enhanc. Heat Transf. 12 (2)
(2005) 121–157.
[44] E.S. Zanoun, F. Durst, H. Nagib, Evaluating the law of the wall in two-dimensional fully developed turbulent channel flows, Phys. Fluids 15 (2003) 3079.

15