International Journal of Heat and Mass Transfer 148 (2020) 119095

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Thermal performance investigation in a novel corrugated plate heat

exchanger
Salman Al zahrani a,b, Mohammad S. Islam a, Feng Xu c, Suvash C. Saha a,∗
a
School of Mechanical and Mechatronic Engineering, University of Technology Sydney, Ultimo, NSW 2007, Australia
b
Department of Mechanical Engineering, Al baha University, Al baha, Saudi Arabia
c
School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, China

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

Article history: Compact heat exchangers have become an essential necessity for power production and multi other pur-
Received 2 April 2019 poses on a daily basis. The corrugated plate heat exchangers (CPHEs) are well-known for their high ther-
Revised 15 November 2019
mal performance. This study proposes a unique CPHE with a simple modification that can boost its ther-
Accepted 21 November 2019
mal performance significantly. The overall tests have been conducted on four CPHEs for two symmetric
chevron angles (β ) of 30°/30° and 60°/60° Two CPHEs belong to the newly CPHEs, and the other two
Keywords: belong to the well-known basic CPHE. Data are obtained for steady-state, single-phase (water-water),
Corrugated plate heat exchanger counter-current arrangements, and for Reynolds number (Re) ranges from 500 to 2500. Sophisticated
Thermal performance mesh techniques have been adopted to develop the mesh for the plates and the fluids between the plates.
Numerical modelling
An appropriate grid refinement test has been carried out for the accuracy of the numerical results. The
Nusselt number
results have been validated with benchmark experimental and numerical data. A realizable k − ε turbu-
lence model with scalable wall treatment found to provide the most consistent and accurate prediction
of the thermal performance of CPHE. The numerical results showed that the Nusselt number (Nu) and
the effectiveness ( ) of the newly developed CPHEs are much higher than that of the basic one, which
can be very useful when a heavy heat duty is required. The enhancement for Nu is up to 75% and for  is
up to 42%, and generally both exhibit a direct proportional relationship with Re. Based on the numerical
result, a new correlation to predict Nu has been developed.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction been employed in food, paper/pulp, power, pharmaceutical indus-

tries, HVAC and many other applications [8].
The use of CPHEs has become an essential necessity on daily In CPHEs, the turbulence flow can be achieved at low Re,
basis. CPHE is the most efficient among other conventional type Re > 400 [9,10]. In addition, CPHEs are lighter and require less
of heat exchangers (HEs). Many researchers have studied differ- space.
ent approaches to enhance the flow mixing and the heat trans- Troup et al. [11] performed one of the earliest studies on CPHE
fer i.e. introducing passive techniques to control the energy dis- performance, who considered washboard plate type HE. However,
sipation rate, and using vortex generators to reduce the wake re- later the chevron plate has become the most popular due to its
gion and enhance the turbulence intensity [1–4]. Other research high thermal performance among other plate types (e.g. wavy, and
conducted to optimize the performance for different types of the zig-zag plate). The impact of different β on CPHE’s thermal per-
HEs [5,6]. CPHEs are compact in nature, have high thermal effec- formance was investigated by Okada et al. [12], where the β was
tiveness, and hence close approach temperatures (2 °C tempera- considered with respect to the horizontal centreline. However, it
ture difference) can be reached [7]. Therefore, CPHEs are impor- is more common that β to be considered with respect to the lon-
tant particularly for heat recovery and regeneration applications. In gitudinal centreline as shown in Fig. 1. Later Muley and Manglik
addition, CPHEs have modular nature that eases the cleaning pro- [13] studied heat transfer characteristics (HTC) in CPHE for β =
cess as well as making the system flexible, by adding or remov- 30°/30°, 60°/60° and 30°/60° The correlations for the whole study
ing plates to meet the energy requirements quickly. CPHEs have were incorporated in one formula. Nevertheless, each correlation
should be separately reported, in order to be able to test the agree-
∗
ment between each β correlation and the general formula. One-
Corresponding author.
E-mail address: suvash.saha@uts.edu.au (S.C. Saha).
one pass, water-water fluids for the same β ’s have been tested

https://doi.org/10.1016/j.ijheatmasstransfer.2019.119095
0017-9310/© 2019 Elsevier Ltd. All rights reserved.
2 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095

Nomenclature

A effective heat transfer area, m2

A◦ channel flow heat transfer area, m2
b corrugation depth, m
Cp specific heat, J/kg.k
de equivalent diameter, de = 2b, m
f fanning friction factor
G core mass velocity, kg/m2 .s
h heat transfer coefficient, W/m2 .k
k thermal conductivity, W/m.k
Lh horizontal length from port to port, m
Lp plate’s channel effective length, m
Lv vertical length from port to port, m
Lw flow channel width, m
m ˙ mass flow rate, kg/s
N number of channels
Nu Nusselt number Fig. 1. Geometrical parameter illustration for chevron plate type.
Q heat transfer rate, W
P pressure drop, Pa
Pc corrugation pitch, m [19] carried out both experimental and numerical studies about
Re Reynolds number thermo-hydraulic characteristics in CPHE with one corrugated plate
t plate thickness, m (β = 60◦ ). The study used one corrugated plate, while the other
plate was flat. The port effect and the pressure drop were ignored.
Greek
In fact, to be able to study heat transfer inside the CPHE, at least
β Chevron angle, °
three plates should be considered (two channels) to allow heat
μ dynamic viscosity, Pa.s
transfer process to take place. One year later, Kanaris et al. [20] ex-
ρ fluid density, kg/m3
ecuted another numerical investigation using CFD, a three plates
 effectiveness
were generated with β = 60◦ . The shape of the corrugation they
Subscripts considered was trapezoidal. The real corrugation however is sinu-
avg average soidal shape. Also, the port effect was neglected again.
b bulk fluid temperature A numerical study for CPHE with two channels was performed
c cold stream by Tsai et al. [21]. The flow maldistribution was investigated by
h hot stream applying Bassiouny and Martin formula [22]. The heat transfer was
HE heat exchanger ignored, and the range of Reynolds number was small (Re ≤ 1700).
HTC heat transfer characteristics A two symmetric β = 30°/30° and 60°/60° CPHEs were numerically
i inlet condition studied using CFD by Asif et al. [23]. The thermo-hydraulic charac-
min minimum teristics was investigated in the form of Nusselt number. An essen-
max maximum tial details were not provided for supporting the appropriateness
m measured of adopting the turbulence model, and no information of the num-
o outlet condition ber of channels or plates. The Wilson plot technique was applied
CPHE1 basic plate heat exchanger in their numerical study. However, a comprehensive understand-
CPHE2 the modified plate heat exchanger ing of CFD would help to avoid going through this long iterative
w wall process.
One good aspect of considering CFD study is that, it enables
the user to find the temperature at any spot on the model, and
by Khan et al. [14]. An equal heat transfer coefficient (h) and Nu the average temperature for any side of the plate. Consequently h
were considered for the both sides as the Reynolds number was can directly be calculated from numerical data. This would be very
the same at the cold and at the hot side. The same Re on the both hard to get from an experimental model. In all previous numerical
sides of the CPHE does not imply h will be the same, as h depends studies reviewed in this study, the mesh dependency tests were
on many other factors i.e. fluid viscosity, fluid density, fluid veloc- not provided. In addition, the mesh statistics was also insufficient.
ity and many other parameters. For the cold side, h is likely greater Al zahrani et al. [15] has numerically investigated the effect of the
than that of hot side, because thermal boundary layer resistance is Prandtl number (Pr) on heat transfer and friction factor (f), by con-
lower in case of cold fluids [15]. In addition, heat capacity rate (C) ducting two sets of tests at hot fluid side, while the cold water
of cold water is higher than that of hot water as the viscosity of kept at the cold side for all cases. First set was for hot air, and the
cold water is higher than that of hot water. Hence the velocity of second set was for hot water. The result concluded that, both Nu
the cold water would be higher in order to meet the same Re, and and f increase as Pr increases.
consequently the mass flow rate of cold water will be higher than An essential objective of the present study is to introduce a
that of hot water. Therefore, considering Nu as the same on both new passive technique that could enhance the thermal perfor-
sides is not an accurate assumption. Similar studies performed on mance of the current CPHE, and consequently reduce its size and
CPHE to investigate the impact of different chevron angles [16,17]. make the system more compact. Most research either studied the
The experimental and numerical studies have concluded that impact of different β on HTC or studied the fluid flow pattern
the flow inside the CPHE is non-uniform and tends to flow to- inside CPHE’s channels. However, in order to be able to reduce
ward the lateral edges of the plate [18]. However, the study did its size and make the system more compact, there is a large
not consider the thermal performance of CPHEs. Kanaris et al. number of ongoing efforts to find new techniques to enhance
 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095 3

the performance of the heat exchanger. In general, heat transfer

enhancement methods are active, passive, and compound. Active
technique requires an external power for the enhancement such
as induced pulsation, and surface vibration. Passive technique
involves a geometrical modification to the fluid flow passage, or
using inserts in the flow passage, or both. An example for passive
technique is inserting twisted tape to promote turbulence flow
regime, adding fins, and extending surface. In compound method,
both active and passive techniques are used. In general, passive
technique is the preferred one [24], because the potential of active
technique is limited due to the design complexity [25].
The present study introduces a new modification in CPHE flow
mechanism as described in the following section, which could en-
hance convective heat transfer significantly between the cold and
the hot fluids, and consequently the fuel consumption can be re-
duced [26]. The numerical thermo-hydraulic performance tests car-
ried out on counter-current flow arrangement, and for two sym-
metric β = 30◦ /30◦ , and 60◦ /60◦ . Nu is employed as an indicator
for heat transfer improvement, and CPHE effectiveness ( ) is em-
ployed to compare the thermal performance between the basic and
the new CPHE design. The CPHE comprised of four channels (five
plates), two of them are pertaining to the cold side, which repre- Fig. 2. The current gasket design.

sent the utility fluid, and the other two pertaining to the hot side,
which represent the product fluid. Therefore, the present study is
3. Numerical method
performed on the hot side of CPHE. The port effect, and the cor-
rugations shape have been considered for all cases in order to get
3.1. Governing equations
as closer as possible to simulate thermal-hydraulic performance in
real CPHE.
Fluids inside CPHE are subjected to turbulent flow due to the
change in velocity in the corrugations for the Reynolds number
considered here. In addition, heat transfer takes place between
2. The new CPHE design approach
cold and hot sides, and consequently a transition of physical prop-
erties occurs such as temperature, pressure, viscosity, density, and
The material technology of CPHE has been constantly improved,
velocity. Therefore, Navier-Stockes Eq. (NS) (1) is used to estimate
which allows the usage of CPHE for further applications such as
changes on these properties during the thermal and dynamic in-
chemical processes. In addition, the brazed CPHE has been intro-
teraction as shown below:
duced in order to resist higher pressures and temperatures. How-  
ever, the basic design of CPHE has not been changed a lot since it ∂   ∂p ∂ ∂ ui
ρ ui u j = − + (μ + μt ) (1)
was invented in 1920s. All studies of the CPHEs were to test the ∂xj ∂ xi ∂ x j ∂xj
performance, flow patterns inside the CPHE, or to study the effect
NS can be considered as Newton’s second law applied to the
of a specific parameter. In the basic design, CPHE’s thermal per-
fluid motion. Inertia, pressure, and viscous forces are estimated.
formance is higher in comparison with other types of HEs such
ANSYS FLUENT 19.0 solves these equations along with continuity
as shell and tube HE [7, 27]. However, the current study intro-
Eq. (2). Whereas NS refers to conservation of momentum, and con-
duces a new modification in the basic design that improves the
tinuity equation refers to conservation of mass.
CPHE’s thermal performance significantly. Additionally, the fluid is
distributed on the plate’s surface randomly in the basic design. The ∂
( ρ ui ) = 0 (2)
modification implies more degree of control of the fluid flow on ∂ xi
the plate’s surface.
In addition, the energy equation is included to resolve heat
In the basic and new CPHEs, the gasket is used to regulate the
transfer among the cold fluid, the hot fluid, and the plates as
fluid directions through the CPHE. The design of the new gasket
shown in Eq. (3):
shown in Fig. 2, where the separator has been installed at the
middle of the plate for two reasons. Firstly, it guides the fluid to ∂   ∂ 2 T   ∂ ui
ρC p u T = ke f f + Ti j (3)
the desired direction. Secondly, it replaces the contact points at ∂xj j ∂ x2j ef f ∂x
j
the middle between every two consecutive plates, where the heat
transfer magnitude is negligible at this area [28]. In addition, it SIMPLE algorithm scheme is enabled to resolve the pressure-
also assures equal fluid distribution on each plate side. velocity coupling. The finite volume based technique is used to dis-
The flow mechanism in the new CPHE is as following; the hot cretize the governing Eqs. along with second order upwind scheme
fluid is flowing from bottom to top and then from top to bottom. to descretize the convection term. Additionally, first order upwind
At the same time on the adjacent plate, the cold fluid is flowing is enabled to discretize the turbulent kinetic energy and dissipa-
in the opposite direction with respect to the hot fluid as shown in tion rate.
Fig. 3 (b). In all well-known CPHEs configurations, the fluid enters
from one side and exits from the opposite side either vertically (as 3.2. Turbulence model
shown in Fig. 3(a)) or diagonally. However, in the new design, the
fluid enters and leaves from the port on the same side, in order Up today, there is no single or favorite turbulence model that
to maximize the amount of heat recovery (reducing temperature can be applied for any turbulence flow modelling. ANSYS FLU-
gradient) between the cold and the hot fluids, and consequently ENT provides a large selection of turbulence models, however, care
enhancing CPHE’s thermal performance. must be taken during choosing an appropriate model. One of the
4 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095

Fig. 3. Illustrative schematics of, (a) flow arrangement for 1–1 pass in the basic CPHE, and (b) The current flow mechanism.

most common turbulence model that is considered as an industry spreading rate of planar and round jet can accurately be predicted.
standard model is k − ε model [29]. The turbulence kinetic energy, In addition, a realizable k − ε model can provide a superior per-
k represents the diversity of fluid fluctuations. The turbulence eddy formance for simulating complex flows, that includes boundary
dissipation, ɛ represents the dissipation rate of the velocity fluctu- layer reattachment, circulation, rotation, and strong adverse pres-
ations. On the other hand, the shear stress transport (SST) k − ω sure gradient. Both k − ε and k − ω are y+ independent. However,
turbulence model is also used to simulate some heat transfer pro- adopting the most appropriate near wall function depends on the
cesses [30,31]. Whereas ω represents the dissipation rate between degree of mesh refinement near the wall. Near wall treatment
ɛ to k. The two equations turbulence models, k − ε and k − ω dif- methods are very useful when the prism boundary layers are not
fers from each other. The SST k − ω model is the combination of sufficient to resolve those layers. Standard wall treatment denotes
k − ω and standard k − ε . Also, the SST k − ω model can resolve the that, the whole boundary layer mesh is located within the log-law
boundary layers near the wall, whereas, the standard k − ε model region. However, for engineering applications, this is difficult to be
can resolve the boundary layer away from the wall. The realizable fulfilled. That is because of existing of different geometrical scales
k − ε model is relatively new approach that differs from standard as well as arbitrary refinement, especially for geometries that con-
one. It shows an outstanding capabilities to capture the mean flow tain narrow curves and passages. Enhanced and non-equilibrium
for a very complicated structures [29]. In realizable k − ε model, a wall functions can be adopted for ɛ based. However, the mesh res-
new formulation approach for the calculation of eddy viscosity μt olution should be high, which is computationally expensive. Al-
(4) (also called turbulent viscosity) has been developed, where Cμ together, both enhanced and non-equilibrium approaches are not
is no longer constant. recommended if viscous sub-layer region is the area of interest.
k2 Instead SST k − ω can perform better in this region [29].
μt = ρ Cμ (4) On the other hand, scalable wall function introduces an elegant
ε
solution for the issue of arbitrary refinement. Particularly in com-
The transport Eqs. (5) and (6) are used to obtain the turbulence
plicated geometries. The mesh is virtually shifted to the log-law
kinetic energy k, and the dissipation rate ɛ values, respectively.
region (y+ ≈ 11.225 ). Hence the invalid modelling of the laminar
∂ ∂   sub-layer and buffer region is avoided.
(ρ k ) + ρ ku j
∂t ∂xj Generally, all turbulence models have been tested for this prob-
 
μ  ∂k
lem. The realizable k − ε model with scalable wall function has
∂
= μ+ t + Gk + Gb − ρε − YM + Sk (5) been adopted. Because it showed the most accurate result that is
∂xj σk ∂ x j close to the experimental one as well as its consistency with dif-
ferent Re.
∂ ∂  
(ρε ) + ρε u j
∂t ∂xj
  3.3. Data formulation
∂  μt  ∂ε ε
= μ+ + ρC1 Sε + C1ε C3ε Gb
∂xj σε ∂ x j k In the current study, Re set as the same for both cold and hot
fluids inside CPHE’s channels. However, Nu was not considered to
ε2
−C2 ρ √ + Sε (6) be the same.
k+ εν
m˙ de
where Gk and Gb characterize the generation of k due to the mean Re = (7)
velocity gradient and buoyancy, respectively. YM depicts the addi-
μ A◦ N
tion of fluctuating enlargement in compressible turbulence to the Re, μ, A◦ , N, and de are known, m˙ is calculated to meet the re-
total dissipation. C1ɛ , C2 , and C3ɛ are constants. σk and σε charac- quired Re. de is twice the plate’s corrugation depth.
terize the turbulence Prandtl numbers for turbulence kinetic en- The essential measurements are the outlet temperatures of cold
ergy k, and its dissipation ɛ, respectively. User-define source terms and hot fluids, and the hot walls temperature. The fluids’ inlet
Sk and Sɛ can be implemented when needed. temperatures and velocities have been set in the initial bound-
A realizable k − ε model satisfies a specific mathematical con- ary conditions. Two main non-dimensional parameters have been
straint as well as complied with turbulent flow physics. The employed to express heat transfer data. Nu is used to a scale
 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095 5

heat transfer improvement, and  is calculated to compare the en-

hancement in thermal performance between the new and the basic
CPHEs for two symmetric β = 30°/30°, and 60°/60° The hot side of
CPHE is considered as the product fluid, while the cold side is con-
sidered as the utility fluid. Therefore, Nu, and  are considered for
the hot side. Nu is given by:

hh de
Nu = (8)
k
The heat transfer coefficient, hh is calculated as follows:
 
Qh = m˙ h c p,h Th,i − Th,o (9)

cp, h has been extracted from the tables of thermodynamics at hot

fluid bulk mean temperature as follows:

(Th,i + Th,o )
Th,b = (10)
2
Note, Qc is calculated from Eq. (11) below. cp, c is also extracted
at the bulk mean cold fluid temperature as shown in Eq. (12) be-
low. The difference between Qh and Qc should always be zero to
Fig. 4. (a) A side view for one corrugation mesh (The whole mesh is not visible
fulfil the energy balance. However, the difference of about 95% of due to the high-density elements), and (b) Close front view for mesh at the cold
the simulations is less than ± 2%, and ±4–6% for the rest of the inlet port.
simulations. Therefore, Qavg is taken as the average value of the
hot and cold heat load and considered for the current calculations. Table 1
Mesh independent test for basic (CPHE1 ) and new (CPHE2 ) CPHEs.

Qc = m˙ c c p,c (Tc,o − Tc,i ) (11) Mesh elements Outlet cold average Outlet hot average
(million) temperature (K) temperature (K)

CPHE1 β =30°/30° 44 295.44 305.32

(Tc,i + Tc,o ) 53.1 295.75 305.82
Tc,b = . (12)
2 62 295.79 305.84
CPHE2 β =30°/30° 56.6 296.57 303.98
Now Qavg is known, and hh is given by: 73 296.78 304.02
80 296.79 304.01
Qavg CPHE1 β =60°/60° 8 293.62 308.45
hh =   (13) 14.8 293.95 308.85
A Th,b − Tw,h 32 293.89 308.87
CPHE2 β =60°/60° 9.5 296.59 303.92
Qmax is the maximum possible amount of heat that could be 16 296.75 303.82
exchanged between hot and cold fluids and is given by: 47 296.73 303.82
 
Qmax = Cmin Th,i − Tc,i (14)

Cmin is the minimum heat capacity. For all cases in the current 4.2. Mesh optimization
study Ch < Cc :
The CPHE contains a large number of curved and tilted narrow
Ch = m˙ h c p,h
(15) passages. Therefore, in order to ensure sufficient mesh element
Cc = m˙ c c p,c
in these narrow passages, an unstructured tetrahedron mesh ele-
Then,  is determined as follows: ments are adopted. An advance technique has been employed to
get a good quality mesh. Patch conforming and patch independent
Qavg
= (16) algorithms have been adopted simultaneously for the same geom-
Qmax etry. The patch conforming is a Delaunay [29] mesher, where mesh
refinement is carried out by using an advancing front point inser-
4. Model setup tion technique. The meshing process uses the bottom up approach,
meshing edges, faces, and volume in sequence. In patch conform-
4.1. CAD geometry creation ing algorithm, excluding the de-featuring tolerance, all faces, and
their boundaries are conformed.
Four CPHEs are drawn by using Solidworks CAD 2016. Two sym- On the other hand, patch independent algorithm is based on
metric CPHEs with β = 30◦ /30◦ , and 60◦ /60◦ are developed ac- spatial subdivision. The mesh refinement is carried out where nec-
cording to CPHE’s basic design. The other two symmetric CPHEs essary, particularly in holes, curves, and narrow passages. Large
with β = 30◦ /30◦ , and 60◦ /60◦ are drawn according to the new mesh elements are developed where possible such as on flat sur-
design criterion. Each CPHE consists of five plates, which gener- faces. Hence, one should look into covering the important compli-
ate four channels. Two channels belong to the cold side and the cated areas on the geometry and allowing for faster computation
other two belong to the hot side. All CPHE’s geometric parameters at the same time. The meshing process uses top down approach.
have been developed carefully. The corrugations have a sinusoidal The volume mesh is carried out first, and then the surface mesh is
shape similar to that one in real CPHE. The port effect is consid- created by projecting the volume mesh on to faces and edges.
ered. The hot port is created and merged with the hot side, and Samples of the mesh is shown in Fig. 4(a) and (b). To ensure so-
same is developed for the cold port side for both basic and new lution stability, the mesh dependency tests have been carried out
design CPHEs. for each CPHE as shown in Table 1. Mesh elements of 53.1 and
6 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095

Fig. 5. Comparison of present Nu data with other experimental and numerical studies, (a) for β = 60◦ /60◦ , and (b) for β = 30◦ /30◦

73 million for β = 30◦ /30◦ , for basic and new design CPHEs have glik [13] are +6% and +2% respectively. The maximum and the
been adopted, respectively. Mesh elements of 14.8 and 16.1 mil- minimum deviation are found -9% and -4% respectively with the
lion for β = 60◦ /60◦ , for basic and new design CPHEs have been findings of Okada et al. [12]. The maximum and the minimum de-
adopted, respectively. viation compared with Lee and Lee [34] are -4.7% and +1.3%, re-
spectively. The numerical correlations of the present study for both
4.3. Boundary conditions and material properties specification β = 30◦ /30◦ and 60◦ /60◦ are in the range of all other published
correlations and shows an increasing trend.
Both hot and cold fluids inlet boundary conditions are set as Several reasons could contribute in the deviation between the
velocity inlet. The mass flow rate (m˙ ) is calculated from Eq. (7) to results. Some studies calculated average Nu between the cold and
meet the required Re value. Thus, velocity at the inlet can be calcu- the hot sides of the CPHE’s [12]. Muley and Manglik [13] calculated
lated from m˙ . The working fluid is (water-water). The hot and the Nu for the hot side, hence the current results are very close to this
cold inlet water temperatures are set to 40 °C and 18 °C respec- one.
tively. The fluid’s thermodynamics properties (ρ , cp , kf , and μ) have Janusz et al. [35] have investigated a number of published
been set for each fluid according to its temperature. Zero gauge Nusselt number correlations to test their accuracy and reported
pressure has been set at both cold and hot ports outlets. Accord- that, Muley and Manglik [13] correlation can predict Nu values
ing to Ansys FLUENT user manual [29], the optimum value for tur- reasonably well. Furthermore, the differences in the geometrical
bulence intensity for the current flow pattern is 5%. The conjugate dimensions such as corrugation depth, aspect ratio (Ar ), and even a
heat transfer is enabled, where the plate thickness is set to 0.5 mm. small difference in the corrugation roundness may result in change
Stationary and no slip boundary conditions are set for all walls. in Nu values up to 18% [36]. However, the maximum deviation is
Since almost all studies performed on CPHEs have used stain- always ≤ ±10% except one case the maximum deviation is 11%.
less steel, and to be able to validate the numerical study, stainless The calculated Nu of the present study shows good agreement
steel is defined as the plate’s material. The plate’s ρ , cp , and k are with the published literature, which sufficiently indicate that
8030 kg/m3 , 502.48 J/kg.k, and 16.27 W/m.k, respectively. All sim- present CFD model is accurate to predict the thermal performance
ulations are carried out on a high performance computing cluster of the corrugated CPHE’s.
using nodes with 3.3 GHz, 28 processor, and with 128 GB of RAM.
For the mesh that contains 14.8 and 16.1 million element, the sim- 5. Results and discussion
ulation time is approximately10-13 hours. For the mesh that con-
tains 53.1 and 73 million elements, the time for each simulation is In the present study, four symmetric CPHEs have been em-
approximately 48-55 hours. ployed. Two of them are belonging to the well-known basic design
with β = 30◦ /30◦ , and 60◦ /60◦ The other two are based on the
4.4. Model validation new design criterion with the same chevron angle of β = 30◦ /30◦ ,

The present study has been comprehensively validated with

available experimental and CFD studies. The same working fluids
and CPHE’s material from the available literatures are used in this
study. The CFD study compared the Nu empirical correlation data
with the published Nu empirical results for different β . Fig. 5(a)
shows the empirical Nu deviation for β = 60◦ /60◦ The maximum
and the minimum deviations are found + 9% and +6% respectively
with the findings of Okada et al. [12]. The findings of the CFD study
also compared with the measurement obtained by Gherasim et al.
[32] and Thonon [33], and the numerical results show good agree-
ment with the published literatures.
The findings of the present study have also been compared
with the results of Muley and Manglik [13] and Lee and Lee
[34], and found a negligible deviation with the published data.
In case of β = 30◦ /30◦ , as shown in Fig. 5(b), the maximum
and the minimum deviation with the study of Muley and Man- Fig. 6. Nu values versus Re for novel and basic CPHEs.
 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095 7

Fig. 7. ZY plane (side view) illustrates fluids velocity contour inside new CPHE’s channels with β 30◦ /30◦ at Re = 500.

Fig. 8. ZX plane (Top view) for velocity vectors for basic CPHE with β 60◦ /60◦ at Re = 500.

Table 2 The non-uniformity of the CPHE’s surfaces causes velocity fluc-

Comparison between Nu of the basic (CPHE1 ) and the new (CPHE2 )
tuation. The fluid has the maximum velocity at the corrugation’s
CPHEs.
ridge and the minimum velocity at the corrugation’s furrows as
Re Nu Nu I% Nu Nu I% shown in Fig. 7. This fluctuation of velocity causes disruption,
CPHE1 CPHE2 CPHE1 CPHE2
boundary layer re-attachment, and secondary flow development.
60°/60° 60°/60° 30°/30° 30°/30°
Consequently, provide heat transfer enhancement.
500 20.6 26.7 29.8 17.1 20.9 22 The longitudinal vortices in the corrugation’s furrows are ob-
1000 31.6 45.7 44.6 26 35.4 35.9
served by Blomerius and Mitra [37]. In addition, Won and Ligrani
1500 40.6 65.1 60.2 33.7 50.8 50.9
2000 49.4 81.5 65.1 40.3 62.6 55.3 [38] has carried out an instantaneous visualization of fluid flow in-
2500 58.3 102.2 75.3 46.2 72 56 side the channel. The result showed a strong spanwise secondary
flow, that moved in an opposite direction in the bottom and the
top halves. The boundary layer re-attachment, and its rule in heat
transfer augmentation was also confirmed by Focke and Knibbe
and 60◦ /60◦ The numerical simulations have been carried out for [39].
the single phase (water-water). All four CPHEs have an identi- In the case of new CPHE, Fig. 9. shows that, the Re varies signif-
cal geometrical dimension ( b, de , β , and Ar ) and physical conditions icantly over the plate as the fluid is forced to flow along the speci-
( Th, i Tc, i , and Re). Reynolds number ranges from 500 to 2500. fied half of the plate. The fluid velocity inside the modified channel
Nusselt number has been calculated for all CPHEs as shown in is about 3 times higher than that of the basic channel. Thus, more
Fig. 6. The results show that, Nu is directly proportional to the turbulent flow is developed, which would contribute in heat trans-
plate’s chevron angle as well as to the Re. In addition, the new de- fer augmentation. Fig. 8. shows the velocity vectors inside the basic
sign exhibits significant heat transfer enhancement for all cases. At CPHE. For the same mass flow rate, the fluid velocity is lower than
the same Reynolds number, Nusselt number for the new CPHE in- that of the new one. Because in basic CPHE the fluid flows and
creases up to 75% compare to the basic CPHE as shown in Table 2. distributes randomly over the plate.
8 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095

Fig. 9. ZX plane (Top view) for velocity vectors for new CPHE with β 60◦ /60◦ at Re = 500.

Table 3
Comparison between  of the basic (CPHE1 ) and the new (CPHE2 ) CPHEs.

Re  CPHE1  CPHE2 I%  CPHE1  CPHE2 I%

60°/60° 60°/60° 30°/30° 30°/30°

500 34.1 42 23 35.6 41.5 16.6

1000 26.4 34.3 30 27.7 34.2 23.5
1500 22.5 31.4 39.6 23.5 30.7 30.6
2000 20 28.6 42.5 20.8 28.5 36.8
2500 18 26.1 42.2 18.7 25.8 38

Thermal performance of basic and new CPHEs has been calcu-

lated in the form of CPHE’s effectiveness ( ). Fig. 10 shows the
CPHE’s  versus Reynold number for all cases. In addition, quan-
titative data of the  along with the enhancement percentage for
the new CPHE’s  comparing with the  of the basic one is given
in Table 3.
Fig. 10. Comparison between the effectiveness ( ) of new and basic CPHEs versus
Re.
Figs. 11(a,b) and 12(a,b) show temperature contour and profile
on the hot channel that locates at the middle (as also shown in

Fig. 11. For β = 60◦ /60◦ , and Re = 500 (a) temperature contour for new CPHE, and (b) temperature profile for new CPHE.
 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095 9

Fig. 12. For β = 60◦ /60◦ , and Re = 500 (a) temperature contour for basic CPHE, and (b) temperature profile for basic CPHE.

Fig. 13.) of the new and the basic CPHE, respectively. In the case
of new CPHE, the hot fluid’s temperature decreasing rate is higher
than that of the basic one. In addition, the hot fluid’s temperature
is more homogeneous. While in the case of basic CPHE, the hot
fluid’s temperature shows less homogeneity. In the basic CPHE, the
temperature of the hot fluid on the left side is lower than that on
the right side as shown in Fig. 12a. That is because the cold fluid is
entering from the left side in the fore and the next adjacent cold
channel as shown in Fig. 13.
The temperature trend throughout the port of hot channel out-
let as shown in Fig. 14 is presented in Fig. 15 (a) and (b). For all
cases, the temperature of the last hot channel is higher than that
of the first hot channel. Whereas, the last hot channel transfers
heat with cold fluid from one side only (the side that locates just
before the last channel).
The first hot channel transfers heat with cold channels from
the front and the back sides. Therefore, its temperature is always
less than the hot channel that locates at the end. All 4 trends at
the beginning show descending in temperature as hot fluid moves
through z axis. Because, the hot fluid of the last channel mixes
with the hot fluid of the first channel, which has lower temper-
ature, the temperature increases up to a specific limit, allowing for
the temperature balance to take place.
In the newly modified CPHE, more degrees of contact be-
tween the thermal plates are presented from the middle area of
each plate where gaskets are installed which would enhance the
mechanical strength and integrity of the HE. Furthermore, the con-
vective heat transfer and the effectiveness of the modified CPHE
are considerably higher than those of the well-known CPHE, hence Fig. 13. Illustrative schematic for fluid flow arrangement in basic CPHE.

its size could be further reduced for the same heat duty allowing
for more compact heat exchanger. Therefore, the newly CPHE could represent an obstacle that needs to be solved [40], hence compact
be implemented in applications that require superior performance HEs are important to overcome such difficulties. Similarly, the
such as the solar receiver powered systems. The design of the new CPHE could be implemented in applications where area and
new generation of the solar receivers in the Concentrated Solar weight are limited such as in rockets to cool their nozzles and in
Power (CSP) systems is required to be very efficient. Whether the ships to integrate with power generation system and to control
pressurized air or the carbon dioxide is the working fluid in the the lube oil temperature. In addition, it could be implemented for
solar receiver, both would require an efficient thermo-hydraulic the heat recovery applications where closer temperature approach
performance of the HE to guarantee the overall efficiency of the could be achieved in comparison with the well-known CPHE as
system. Additionally, the geometrical constraints in these systems presented earlier in Figs. 11, 12, and 15.
10 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095

Fig. 14. Illustrative sketch for hot channel outlet port.

Fig. 15. Temperature profile through the hot channel’s outlet axis at Re = 500 for basic and new CPHEs, (a) β = 60◦ /60◦ , and (b) for β = 30◦ /30◦ .

5.1. Heat transfer correlations And Nu for the basic and the new CPHEs for β = 30◦ /30◦ is as
follows, respectively.
The heat transfer correlations are essential to help the design-  μ 0.14
ers to predict the HE’s thermal performance and to estimate the Nu = 0.2332 Re0.6175 P r 1/3 (20)
required heat transfer area. They are also important for the re- μw
searcher to compare the harmony level of their data with those in
 μ 0.14
the literature. For the present study, classical Sieder [41] empirical
Nu = 0.1134 Re0.7721 P r 1/3 (21)
correlation (17) was adopted to predicts the Nu correlations. All μw
thermodynamics properties are estimated at the bulk temperature
as shown earlier in Eq. (10).
6. Conclusions
 μ 0.14
p n
Nu = CRe P r (17) Thermal performance of new CPHE for symmetric = 30◦ /30◦
μw
and 60◦ /60◦ has been carried out. A comprehensive validation has
In order to find the constant values C and p for each CPHE, lin- been performed with a wide range of published experimental and
ear regression technique has been applied to the numerical data in numerical investigations. A realizable k − ε turbulence model with
Fig. 5 (a,b). The Pr exponent (n) value ranges from 0.333 up to 0.5 scalable wall treatment found to provide the most consistent and
as reported in the literature [12,13,42–44]. However, different val- accurate prediction of the thermal performance of CPHE. The nu-
ues for n has already been tested in Al zahrani et al. [15], and 0.333 merical simulations have been conducted for steady-state single
is found to provide the best approximation. Hence, Nu for the ba- phase (water-water), and counter-current flow arrangement.
sic and the new CPHEs for β = 60◦ /60◦ is as follows, respectively. The CFD results showed that, the thermal performance of
 μ 0.14 the new CPHE is significantly higher than that of the well-
Nu = 0.2354 Re0.6415 P r 1/3 (18) known CPHE. For the newly developed CPHE, the calculated Nu
μw is up to 75% higher than that available for CPHE. The effec-
tiveness for the new CPHEE is significantly higher than that of
 μ 0.14 the well-known CPHE, and generally exhibits a direct propor-
Nu = 0.096 Re0.8273 P r 1/3 (19)
μw tional relationship with Re. An empirical based correlation for
 S. Al zahrani, M.S. Islam and F. Xu et al. / International Journal of Heat and Mass Transfer 148 (2020) 119095 11

each CPHE is developed to estimate Nu values. The newly de- [14] T. Khan, M. Khan, M.-C. Chyu, Z. Ayub, Experimental investigation of single
veloped CPHE would increase the understanding of the thermal phase convective heat transfer coefficient in a corrugated plate heat exchanger
for multiple plate configurations, Appl. Therm. Eng. 30 (8-9) (2010) 1058–1065.
behaviour of the CPHE. The findings of the present study could [15] S. Al zahrani, M.S. Islam, S.C. Saha, A thermo-hydraulic characteristics investi-
be very useful for many applications, especially for heat recov- gation in corrugated plate heat exchanger, Energy Proc. 160 (2019) 597–605.
ery applications, where high heat transfer coefficients are re- [16] W. Focke, J. Zachariades, I. Olivier, The effect of the corrugation inclination an-
gle on the thermohydraulic performance of plate heat exchangers, Int. J. Heat
quired, and for applications where area and weight are limited Mass Transf. 28 (8) (1985) 1469–1479.
and require large amount of heat to be removed such as in ships [17] M.B. Kim, C.Y. Park, An experimental study on single phase convection heat
and airplanes. transfer and pressure drop in two brazed plate heat exchangers with different
chevron shapes and hydraulic diameters, J. Mech. Sci. Technol. 31 (5) (2017)
Since the performance of the newly CPHE is much higher than
2559–2571.
the well-known CPHE, its size can be further reduced for the [18] A. Lozano, F. Barreras, N. Fueyo, S. Santodomingo, The flow in an oil/water
same heat duty allowing for more compact heat exchanger. Fur- plate heat exchanger for the automotive industry, Appl. Therm. Eng. 28 (10)
(2008) 1109–1117.
ther extension of numerical simulations could be performed on the
[19] A.G. Kanaris, A.A. Mouza, S.V. Paras, Flow and heat transfer prediction in a
present work to study the difference in size between the new and corrugated plate heat exchanger using a CFD code, Chem. Eng. Technol. Ind.
the well-known CPHE for the same thermal performance. Chem. Plant Equip. Process Eng. Biotechnol. 29 (8) (2006) 923–930.
[20] A. Kanaris, A. Mouza, S. Paras, Flow and heat transfer in narrow channels with
corrugated walls: a CFD code application, Chem. Eng. Res. Des. 83 (5) (2005)
Declaration of Competing Interest 460–468.
[21] Y.-C. Tsai, F.-B. Liu, P.-T. Shen, Investigations of the pressure drop and flow
distribution in a chevron-type plate heat exchanger, Int. Commun. Heat Mass
The authors declare no conflict of interest. Transf. 36 (6) (2009) 574–578.
[22] M. Bassiouny, H. Martin, Flow distribution and pressure drop in plate heat ex-
changers—I u-type arrangement, Chem. Eng. Sci. 39 (4) (1984) 693–700.
CRediT authorship contribution statement [23] M. Asif, H. Aftab, H. Syed, M. Ali, P. Muizz, Simulation of corrugated plate
heat exchanger for heat and flow analysis, Int. J. Heat Technol. 35 (1) (2017)
Salman Al zahrani: Writing - original draft. Mohammad S. Is- 205–210.
[24] E. Elshafei, M. Awad, E. El-Negiry, A. Ali, Heat transfer and pressure drop in
lam: Writing - review & editing. Feng Xu: Writing - review & edit- corrugated channels, Energy 35 (1) (2010) 101–110.
ing. Suvash C. Saha: Supervision, Visualization, Writing - review & [25] A. Dewan, P. Mahanta, K.S. Raju, P.S. Kumar, Review of passive heat transfer
editing. augmentation techniques, Proc. Inst. Mech. Eng. Part A J. Power Energy 218 (7)
(2004) 509–527.
[26] R. Gugulothu, K.V.K. Reddy, N.S. Somanchi, E.L. Adithya, A review on enhance-
Acknowledgment ment of heat transfer techniques, Mater. Today Proc. 4 (2) (2017) 1051–1056.
[27] D. Walraven, B. Laenen, W. D’haeseleer, Comparison of shell-and-tube with
plate heat exchangers for the use in low-temperature organic Rankine cycles,
The authors would like to thank the high-performance comput- Energy Convers. Manag. 87 (2014) 227–237.
ing Unit ARCLab, University of Technology Sydney, Australia. [28] S. Jain, A. Joshi, P. Bansal, A new approach to numerical simulation of small
sized plate heat exchangers with chevron plates, J. Heat Transf. 129 (3) (2007)
291–297.
Supplementary materials [29] Ansys, "Ansys fluent 12.0 user’s guide." (2009).
[30] F. Menter, T. Esch, Elements of industrial heat transfer predictions, in: Proceed-
ings of the Sixteenth Brazilian Congress of Mechanical Engineering (COBEM),
Supplementary material associated with this article can be 2001 sn..
found, in the online version, at doi:10.1016/j.ijheatmasstransfer. [31] D.C. Wilcox, Formulation of the kw turbulence model revisited, AIAA J. 46 (11)
2019.119095. (2008) 2823–2838.
[32] I. Gherasim, M. Taws, N. Galanis, C.T. Nguyen, Heat transfer and fluid flow in
a plate heat exchanger part I. Experimental investigation, Int. J. Therm. Sci. 50
References (8) (2011) 1492–1498.
[33] B.. Thonon, Design method for plate evaporators and condensers, in: Proceed-
[1] G. Biswas, K. Torii, D. Fujii, K. Nishino, Numerical and experimental determi- ings of the BHR Group Conference Series Publication, Mechanical Engineering
nation of flow structure and heat transfer effects of longitudinal vortices in a Publications Limited, 1995.
channel flow, Int. J. Heat Mass Transf. 39 (16) (1996) 3441–3451. [34] J. Lee, K.-S. Lee, Flow characteristics and thermal performance in chevron type
[2] P. Saha, G. Biswas, S. Sarkar, Comparison of winglet-type vortex generators pe- plate heat exchangers, Int. J. Heat Mass Transf. 78 (2014) 699–706.
riodically deployed in a plate-fin heat exchanger–a synergy based analysis, Int. [35] J.T. Cieśliński, A. Fiuk, K. Typiński, B. Siemieńczuk, Heat transfer in plate heat
J. Heat Mass Transf. 74 (2014) 292–305. exchanger channels: experimental validation of selected correlation equations,
[3] A. Sinha, K.A. Raman, H. Chattopadhyay, G. Biswas, Effects of different orienta- Arch. Thermodyn. 37 (3) (2016) 19–29.
tions of winglet arrays on the performance of plate-fin heat exchangers, Int. J. [36] E. Sparrow, L. Hossfeld, Effect of rounding of protruding edges on heat transfer
Heat Mass Transf. 57 (1) (2013) 202–214. and pressure drop in a duct, Int. J. Heat Mass Transf. 27 (10) (1984) 1715–1723.
[4] L. Sirovich, S. Karlsson, Turbulent drag reduction by passive mechanisms, Na- [37] H. Blomerius, N. Mitra, Numerical investigation of convective heat transfer and
ture 388 (6644) (1997) 753. pressure drop in wavy ducts, Numer. Heat Transf. Part A Appl. 37 (1) (20 0 0)
[5] T. Rush, T. Newell, A. Jacobi, An experimental study of flow and heat transfer 37–54.
in sinusoidal wavy passages, Int. J. Heat Mass Transf. 42 (9) (1999) 1541–1553. [38] S. Won, P. Ligrani, Comparisons of flow structure and local nusselt numbers in
[6] C.-C. Wang, W.-S. Lee, W.-J. Sheu, A comparative study of compact en- channels with parallel-and crossed-rib turbulators, Int. J. Heat Mass Transf. 47
hanced fin-and-tube heat exchangers, Int. J. Heat Mass Transf. 44 (18) (2001) (8-9) (2004) 1573–1586.
3565–3573. [39] W. Focke, P. Knibbe, Flow visualization in parallel-plate ducts with corrugated
[7] Z.H. Ayub, Plate heat exchanger literature survey and new heat transfer and walls, J. Fluid Mech. 165 (1986) 73–77.
pressure drop correlations for refrigerant evaporators, Heat Transf. Eng. 24 (5) [40] Q. Li, G. Flamant, X. Yuan, P. Neveu, L. Luo, Compact heat exchangers: A re-
(2003) 3–16. view and future applications for a new generation of high temperature solar
[8] M. Young, Plate heat exchangers as liquid cooling evaporators in ammonia re- receivers, Renew. Sustain. Energy Rev. 15 (9) (2011) 4855–4875.
frigeration system, Proceeding of the IIAR Sixteenth Annual Meeting, 1994. [41] E.N. Sieder, G.E. Tate, Heat transfer and pressure drop of liquids in tubes, Ind.
[9] B.P. Rao, S.K. Das, An experimental study on the influence of flow maldistribu- Eng. Chem. 28 (12) (1936) 1429–1435.
tion on the pressure drop across a plate heat exchanger, J. Fluids Eng. 126 (4) [42] G. Longo, A. Gasparella, Refrigerant r134a vaporisation heat transfer and pres-
(2004) 680–691. sure drop inside a small brazed plate heat exchanger, Int. J. Refrig. 30 (5)
[10] R.K. Shah, D.P. Sekulic, Fundamentals of Heat Exchanger Design, John Wiley & (2007) 821–830.
Sons, 2003. [43] A. Talik, L. Fletcher, N. Anand, L. Swanson, Heat Transfer and Pressure Drop
[11] R. Troupe, J. Morgan, J. Prifiti, The plate heater versatile chemical engineering Characteristics of a Plate Heat Exchanger Using a Propylene-Glycol/Water Mix-
tool, Chem. Eng. Prog. 56 (1) (1960) 124–128. ture As the Working Fluid, American Society of Mechanical Engineers, New
[12] K. Okada, M. Ono, T. Tomimura, T. Okuma, H. Konno, S. Ohtani, Design and York, NY (United States), 1995.
heat transfer characteristics of new plate heat exchanger, Heat Transf. Jpn. Res. [44] A. Wanniarachchi, U. Ratnam, B. Tilton, K. Dutta-Roy, Approximate Correlations
1 (1) (1972) 90–95. For Chevron-Type Plate Heat Exchangers, American Society of Mechanical En-
[13] A. Muley, R. Manglik, Experimental study of turbulent flow heat transfer and gineers, New York, NY (United States), 1995.
pressure drop in a plate heat exchanger with chevron plates, J. Heat Transf.
121 (1) (1999) 110–117.