Journal of Energy Storage 35 (2021) 102312

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Numerical simulation and experimental verification of constrained melting

of phase change material in cylindrical enclosure subjected to a constant
heat flux
Ahmed Saad Soliman a, b, Shuping Zhu a, Li Xu a, Junguo Dong a, Ping Cheng a, *
a
School of Environmental and Chemical Engineering, Shanghai University, Shanghai 200444, China
b
Mechanical Power Engineering Department, Mansoura University, El-Mansoura, 35516, Egypt

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

Keywords: The transient melting process of the phase change material (PCM) is controlled by heat conduction and natural
Natural Convection convection. Several experimental investigations are essential to understand the heat transfer mechanisms and
Heat flux performance for different designs of PCM thermal storage systems. Therefore, the present study experimentally
Melting
investigated paraffin wax’s melting process in a vertical cylindrical enclosure. The PCM, initially at 30 ◦ C, was
PCM
Cylindrical Enclosure
heated using an electric heater located at the center of the enclosure. The heat flux density varied to 1300, 1000,
and 700 W/m2. Local temperatures of the PCM were measured, and the solid-liquid interface was tracked. A CFD
(computational fluid dynamic) model was developed to numerically investigate the melting process with and
without considering the convection effect. The CFD model was validated with the current experiments and with
data from the literature. A very good agreement was obtained. The results showed that; the melting was initially
dominated by conduction at an early stage, and this period increased with a decrease in heat flux. Natural
convection was promoted, at a later time, leading to a curved shape of the solid-liquid interface. Numerical
results indicated a robust thermal stratification of the molten liquid in the upper half of the storage unit. It was
observed that increasing the input power from 700 W/m2 to 1000 W/m2 and 1300 W/m2 decreased the total
melting time by 24.82% and 43.58%, respectively. It is recommended for the future modeling of the PCM melting
process to consider the convection effect.

heat, and cycle stability [5,6]. During the heat transfer investigations,
the PCM thermophysical properties, including the melting point, heat of
1. Introduction fusion, thermal conductivity, density, and specific heat, should be
detected in the solid and liquid state [7]. Research has been performed
Thermal energy can be stored in sensible, latent, and thermochem­ on the types of PCM, and the results suggested that paraffin wax is the
ical heat storage. The latent heat storage in phase change materials most valuable type due to some desirable thermal properties, including
(PCMs) has more desirable properties than other heat storage systems high latent heat capacity, the low vapor pressure in the melt, and good
because PCM has various useful features. PCM can be used as a heat thermal and chemical stability. However, the main problem in using
source at a constant temperature with a minimum temperature drops paraffin wax in applications is its low thermal conductivity (0.21- 0.24
during the heat recovery. It has a high energy density with a small W/m.K) [8].
storage volume, chemical stability, non-corrosion, and low vapor pres­ Several ideas have been proposed to enhance the thermal conduc­
sure at working temperature [1]. Therefore, PCM is one of the most tivity of PCM, especially during the melting process, which could lower
effective methods used in thermal energy storage applications [2,3]. heat transfer rates. The improved methods can be divided into two
Many engineering applications, such as electronic cooling technology, types: first, using the extended surface to increase the surface area, such
waste heat recovery, ventilation, and air conditioning (HVAC), are also as fins [9,10], and porous material such as metal foams [11-13]. Second,
applicable to PCM [4]. adding fine particles [14,15] to nanoparticles [16,17].
The main parameters for selecting a PCM material for any applica­ Using nanoparticles has a significant concern in various
tion are thermal conductivity, phase change temperature, high latent

* Corresponding author.
E-mail address: Pingcheng@shu.edu.cn (P. Cheng).

https://doi.org/10.1016/j.est.2021.102312
Received 28 November 2020; Received in revised form 2 January 2021; Accepted 6 January 2021
Available online 22 January 2021
2352-152X/© 2021 Elsevier Ltd. All rights reserved.
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Nomenclature l Tube length [m]

E Stored energy per kg of PCM [J/kg]
Qt Total input heat [W] t Time of charging [sec]
Qus Useful heat [W] Cp Specific heat at constant pressure [kJ/ (kg. K)]
Qloss Heat loss [W] L Latent heat [kJ/kg]
V Output voltage [V]
R Electrical resistance [Ohm] Greek symbols
Kins Insulation thermal conductivity [W/(m. K)] ⍴ Density [kg/m3]
r Radial distances [m] β Thermal expansion coefficient [1/K]
m Mass of paraffin wax [kg] λ Liquid fraction
∆Tins Temperature difference [K] Subscripts
q Heat flux [W/m2] s Solid-phase
As Surface area [m2] l Liquid-phase
Do Outer diameter [m]

investigations to enhance the thermal performance of fluids [18] and Mehta et al. [37] compared the vertical and horizontal storage units
PCM [19-22]. By adding nanoparticles to PCM, the specific heat capacity and concluded that the convective heat transfer is operational in the
and thermal conductivity of the nanocomposite were increased vertical unit during the melting process, resulting in a melting rate is
compared to the pure one. The temperature was a dominant parameter almost constant compared to the horizontal units. In the horizontal
that affects the thermal conductivity. configuration, the liquid PCM forms a recirculation in the normal di­
The PCM melting process is a transient process controlled by two rection and around the heated tube. The melted PCM thus fills the upper
heat transfer modes: heat conduction and natural convection [23]. part of the device easily, and once it melts completely, the temperature
Solving the phase change heat transfer problem is more complex and has becomes uniform, reducing the convection intensity. In the vertical unit,
attracted the attention of some researchers. Due to a small change of the melted PCM fills the upper part gradually, and hence, there is a
PCM density, the natural convection can be ignored. However, the in­ molecular movement through the process.
fluence of natural convection exists and increases gradually with time in Khillarkar et al. [38] numerically studied the melting mechanism of a
the actual process. The melting method of PCM determines which heat PCM in tube geometries of two different configurations. The first was a
transfer mechanisms is the dominant model. The solid PCM position in a square external tube with a circular tube inside, and the second was a
container divides the melting process in PCM into unfixed and fixed circular external tube with a square tube inside. Their findings showed
modes. that thermal stratification is obtained in the upper part of the cavity
The solid PCM descends to the bottom of the container during the because of natural convection.
unfixed melting, and the heat transfer to the PCM occurs through both Sari and Kaygusuz [39] conducted an experimental analysis on
heat conduction and natural convection. Heat transfer by conduction is melting and solidification in a vertical double pipe energy storage de­
the dominant mode of heat transfer in the solid PCM, starting from a vice. They found that the average heat transfer coefficient and heat flow
completely solid-state and passing through the first layer of liquid PCM rate were further affected during the melting process due to the natural
during the melting process. Natural convection can only be initiated convection effect.
when sufficient PCM is first melted by conduction [24-27]. Natural Anica Trp [40] performed numerical and experimental heat transfer
convection transfers heat to the upper part of the enclosure as the PCM studies during the melting and solidification process of the latent heat
melting progresses. Therefore, relative to the bottom part, more PCM thermal storage system (HTES) unit without considering the effect of
melts at the upper part of the enclosure [28-30]. This has been observed natural convection in liquid PCM. The experimental study has shown
in many geometries, such as rectangular [31,32] and annular spaces for that the PCM charging process occurs non-isothermally, while the dis­
both horizontal and vertical directions [27]. charging process is isothermal and is in good agreement with the nu­
During the fixed melting, heat conduction plays a significant role merical study.
only during the initial stage of melting, and the solid PCM is restrained The transient output of LHTES was studied by Bagheri et al. [41]
from settling down to the bottom of the container. Simultaneously, as using the enthalpy method to disregard the influence of the shell wall
the gap between the solid PCM and container wall increases, natural thermal resistance of the storage device on the PCM charging time.
convection becomes the dominant heat transfer mode [33,34]. When the thermal resistance of the shell wall was used in the model, a
In liquid PCM, the strength of natural convection flows depends on delay in the charge time of the storage unit was recorded by the authors.
the container’s geometry and a group of non-dimensional numbers, Various studies investigated the melting process of PCM in a rect­
including Rayleigh, Stefan, and Fourier [32]. Experimental in­ angular cavity that was subjected to constant heat flux. During the initial
vestigations are required to understand heat transfer mechanisms and stages of melting, the natural convection plays a dominant role, and the
the performance differences resulting from different PCM thermal strength of natural convection reduces as melting is completed. The
storage systems designs. important parameter controlling heat transfer and melting velocity was
Experimental and numerical investigations have been performed in the applied heat flux [42,43].
recent years to improve the understanding of heat transfer within the In previous studies, the melting process of the PCM exposed to a
PCM, particularly at the solid-liquid moving interface. This will allow constant heat flux has been used in many thermal applications. How­
optimal storage device designs to be built for specific applications, such ever, highlighting the impact of considering the conduction or natural
as rectangular enclosures, spherical capsules, tubes or cylinders, and convection in the melting of PCM contained in a cylindrical energy
annular cavities [35]. storage unit is still not investigated in the literature. Therefore, In the
Natural convection was confirmed by tracer particles showing the present work, the PCM melting process was experimentally and
fluid motion during melting in the liquid phase by Menon et al. [36]. The numerically evaluated in a vertical cylindrical tank subjected to a con­
flow near the hot tube wall is shown to be upward due to differences in stant heat flux at the center. The present study aims to understand the
density, and the flow near the cold solid-liquid interface is downward. effects of natural convection on PCM thermal performance under

2
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

various heat flux values. Understanding the melting process in a cylin­

drical tank is widely useful for many engineering applications of thermal
energy storage.

2. Experimental Setup and Procedures

An experimental apparatus was designed and built to measure the

temperature distribution during the PCM melting process accurately.
The PCM storage system was composed of two concentric cylinders. The
outer cylinder with an inner diameter of 92 mm was made of PVC
(polyvinyl chloride). The inner cylinder with an outer diameter of 30
mm; was made of aluminum with a thickness of 9 mm. An electric heater
was inserted inside a glass tube. The glass tube with an electric heater
was inserted inside a hollow aluminum rod to obtain a constant and
uniform heat flux on the outer surface of the aluminum rod. The length
of the whole system was 270 mm. The vertical cylinder was insulated by
1.5 cm thick insulator sheets (K= 0.042 W/m.K) for insulation,
The storage system with a temperature distribution inside the PCM
was experimentally studied. This setup consisted of a data acquisition
system, recorder, electric heater, test section, variable transformer, and
digital voltmeter. Uniform heat flux condition was achieved by an
electric heater (nickel-chrome wire) wrapped on the inner wall of the
aluminum tube. The power supplied to the heating coil was adjusted
using a variable output power supply (transformer). Therefore, the heat
flux used to heat the PCM was adjustable. Heat transfer characteristics at
various heat flux values were also studied. Throughout all experiments,
the resistance of the electric heater was measured in each experiment,
Fig. 2. The thermocouples arrangement on the test section.
which is reasonably dependent on the temperature, and the volt sup­
plied by the transformer was measured by using a voltammeter. The
schematic layout of the experimental setup is shown in Figure 1. Three Table 1
sets of experiments were conducted using three different heat fluxes of Thermophysical properties of the PCM and aluminum tube.
1300, 1000, and 700 W/m2 to study the effect of convection heat
Thermophysical properties Paraffin wax [44] Aluminum tube (Al)
transfer on PCM melting behavior. All experiments were initiated using
solid PCM at a uniform temperature of 30◦ C and repeated three times to K s (W/m.K) 0.15 211
K l (W/m.K) 0.15 -
obtain accurate results. ⍴ s (kg/m3) 890 2675
Two longitudinal slots were created in the heater tube wall with a ⍴ l (kg/m3) 712 -
depth of 2.0 mm to fix the thermocouples to ensure that the thermo­ CP s (kJ/kg. K) 2.384 0.903
couples reading is more sensitive to the temperature of the outer wall of CP l (kJ/kg. K) 2.44 -
Kinematic viscosity (m2/s) 8.31 £ 10¡5 -
the tube. The heat flux was adjusted to study the effect of heat flux
Melting point (◦ C) 54.22/56.22 -
density on heat transfer characteristics inside the PCM, and the results Heat of fusion (kJ/kg) 278.8 -
were validated with the numerical results. The geometrical configura­ Thermal exp. Coeff (l/K) 0.000714 -
tion of the test section, heater, and thermocouples distribution is shown
in Figure 2. The annular space between the outer cylinder and an
aluminum rod was filled with 1.6 kg of paraffin wax as a phase change 0.21 mm were placed at the axial and radial position of the storage unit.
material. The thermophysical properties of the PCM and aluminum tube In the PCM domain, 16 thermocouples were fixed, and the average
are represented in Table 1. temperature was measured. First, four thermocouples were vertically
Twenty-four K-type thermocouples with a small wire diameter of distributed at a radius of 23 mm. The vertical distances were 33, 101,

Fig. 1. Schematic representation of the experimental setup.

3
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

168, and 236 mm for thermocouples number 15, 16, 17, and 18, Heat flux (q) can be calculated from the following equations:
respectively. Second, twelve thermocouples were set at three heights of
As = π × Do × l (4)
50, 135, and 202 mm from the tube base with the radii of 15, 25, 35, and
40 mm, in the radial direction. In the electric heater aluminum pipe, six Qus
thermocouples were fixed into the longitudinal slot of the aluminum q = (5)
As
tube wall at a depth of 2 mm, to secure the thermocouples and ensure
that thermocouples accurately sense the outer wall temperature of the The total stored energy per kg of PCM can be calculated as:
tube. Figure 3 shows the geometrical configuration of the aluminum (q × As × time)
tube and the arrangement of thermocouples. E = (6)
m

3. Experimental Measurements The energy storage can also be estimated using the experimental
obtained liquid fraction values. The experimental liquid fraction can be
To calculate the total input heat, the voltage applied to the electric experimentally obtained if the PCM container is transparent. In this case,
heater was measured using a voltmeter with a minimum readable value capturing and processing real instantaneous images for the melting
of ± 0.1 V. The resistance of the electric heater was measured with a history are used to estimate the liquid fraction as used in [45]. This
minimum readable value of ± 0.1 Ohm. The temperatures were cannot be obtained in the present experiment because of the opaque
measured every 5 seconds and displayed on a computer monitor by a outside cylindrical shell. Therefore, using a transparent outside cylinder
data logger with an accuracy of ± 0.1 ◦ C. containing the PCM could be a promising option to capture melting
The total input heat to the electric heater (Qt), wrapped in an history experimentally and estimate the energy storage as a function of
aluminum tube and transferred to the PCM, was divided into useful heat liquid fraction in future research work.
(Qus), that flows to the PCM through the heater wall and part of the heat
transferred to the surroundings as heat loss (Qloss). The total input heat 4. Uncertainty Analysis
can be determined as follows:
Errors generated by calculations or measurements were presented
Qt =
V2
(1) due to personal and instrumental errors. The individual error may be
R overcome or minimized by repeating measurements several times. As
Ignoring the axial heat transfer (l >> d), the heat lost from the outer shown in Table 2, the instrumental errors are due to the accuracy of the
surface of the PVC tube to the surrounding (Qloss) can be obtained by measuring devices. The uncertainty of the calculation results was esti­
fixing the two thermocouples inside the insulation. One thermocouple mated based on the uncertainty of the primary measurements.
was fixed at a radial distance of 53 mm and the other at a radial distance Error propagation is calculated using the root sum square method.
of 62 mm. The amount of heat lost from the tube wall was calculated Given the specific result Z, we have:
using Eq 2: Z = z (x1, x2, . . ., xn) (7)
ΔTins Where xi is independently measured quantity, the uncertainty ωi is
Qloss = ( ) (2)
1
ln rr21 the uncertainty of independent variables, and the uncertainty ωz in the
2π Kins l
result, Z is given by:
The useful heat transferred to the PCM can be calculated as follow: [( )2 ( )2 ( )2 ( )2 ]0.5
∂Z ∂Z ∂Z ∂Z
ωz = .ω 1 + .ω 2 +…+ .ωn− + .ωn
Qus = Qt − Qloss (3) ∂x1 ∂x2 ∂xn− 1
1
∂xn
(8)

5. Numerical Modeling

Brent’s [46] numerically investigated research that used the

enthalpy porosity technique to model the combined
convection-diffusion phase transition. By using this method, the inter­
face in solid-liquid is not followed directly. The solid-liquid mushy zone
is regarded as a porous zone with a quantity marked as a liquid fraction
as its "porosity" with suitable momentum sink terms increased to mo­
mentum equations, which is the cause of the pressure drop as solid
material results existence.
The liquid PCM density difference in the buoyancy term is repre­
sented by the Boussinesq approximation, which includes thermal
buoyancy. Therefore, the 3D transient laminar flow containing buoy­
ancy driven-conventional governing equations can be expressed as
follows:
The continuity equation is as follows:

Table 2
Errors in the measurement devices.
Measured Value Error in Measured Value

Temperature ± 0.1 oC
Diameter ± (.5) mm
Voltage ± 0.1 Volt
Resistance ± 0.1 Ohm
Fig. 3. The thermocouples arrangement on the aluminum tube.

4
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

∂ρ 1 ∂(rρVr ) 1 ∂(ρVθ ) ∂(ρVz ) liquid fraction λ can be articulated as follows:

+ + + =0 (9)
∂t r ∂r r ∂θ ∂z ⎧
⎪ ΔH
⎪ = 0 T < Tm
The momentum equations are as follows: ⎪
⎪
⎪ L
⎪
⎪
⎪
⎪
. (Momentum r:) ⎨ ΔH T − Tsolidus
λ= = Tm < T < Tm + Δ Tm (17)
( ) ⎪
⎪ L Tliqudus − Tsolidus
∂(Vr ) ∂(Vr ) Vθ ∂(Vr ) Vθ2 ∂(Vr ) ⎪
⎪
⎪
ρ + Vr + − + Vz ⎪
⎪
∂t ∂r r ∂θ r ∂z ⎪
⎪
⎩ ΔH
= 1 T > Tm + Δ Tm
[ ( ) ] L
∂P 1 ∂ ∂(Vr ) Vr 1 ∂2 (Vr ) 2 ∂(Vθ ) ∂2 (Vr )
=− +μ r − 2+ 2 − + +S
∂r r ∂r ∂r r r ∂θ2 r2 ∂θ ∂z2 where Tm is the melting temperature of the PCM, and ∆Tm is the range
(10) between phase transition, which is defined as the variance between
Momentum θ: liquidus and solidus temperatures, as shown in Eq. (18).
(
∂(Vθ ) ∂(Vθ ) Vθ ∂(Vθ ) Vr Vθ ∂(Vθ )
) Δ Tm = Tliquidus − Tsolidus (18)
ρ + Vr + + + Vz
∂t ∂r r ∂θ r ∂z The density and thermal conductivity of PCM depends on the phase
1 ∂P
[
1 ∂
(
∂(Vθ )
)
Vθ 1 ∂2 (Vθ ) 2 ∂(Vr ) ∂2 (Vθ )
] change process as follows:
=− +μ r − 2+ 2 2
− 2 + 2
+S ⎧
r ∂θ r ∂r ∂r r r ∂θ r ∂θ ∂z ρs T < Tm
⎪
⎪
(11) ⎪
⎪
⎨
ρs + ρl
Momentum Z: ρpcm = Tm < T < Tm + ΔTm (19)
⎪
⎪ 2
( ) ⎪
⎪
⎩
∂(Vz ) ∂(Vz ) Vθ ∂(Vz ) Vr Vθ ∂(Vz ) ρl T > Tm + ΔTm
ρ + Vr + + + Vz
∂t ∂r r ∂θ r ∂z
[ ( ) ] ⎧
∂P 1 ∂ ∂(Vz ) 2 2
1 ∂ (Vz ) ∂ (Vz ) ⎪
⎪ ks T < Tm
=− +μ r + 2 + + ρgβ(T − Tm ) + S ⎪
⎪
∂z r ∂r ∂r r ∂θ2 ∂z2 ⎪
⎨
ks + kl
(12) kpcm = Tm < T < Tm + Δ Tm (20)
⎪
⎪ 2
⎪
⎪
The parameter S is Darcy’s law damping term added to the mo­ ⎪
⎩
k l T > Tm + Δ Tm
mentum equation as a convective phase change effect. This term de­
pends on both the existing constant Amush and a liquid fraction (λ), that
can be expressed as follows: 5.1. Boundary and initial conditions

Initially, the whole system was at the ambient temperature of 30◦ C,

2
→ (1 − λ) →
S = 3 Amush. V
(λ + γ) and the storage unit was filled with solid paraffin wax, which was
maintained at a temperature less than the melting point. The boundary
Where: γ is a minor number, naturally about 10− 3, and is recognized
conditions were as follows: the inner side was subjected to constant heat
to avoid division by zero. The mushy zone constant Amush explains how
flux, and the outer walls of the storage unit were insulated.
steeply its velocity is condensed to zero when the material solidifies. Its
value depends on the morphology of the medium. This value is
5.2. Numerical procedure and validation
frequently high and ranges between104 and 108. The same can also be
said when the local liquid friction turns to be high, and the velocity is
The numerical simulation of the model was developed using ANSYS
condensed to zero. In this study, Amush is supposed to be constant
19.2 software. The pressure-velocity coupling equations in the PCM
throughout the whole simulation and is set to be 106, where it attains the
were determined using the Semi-Implicit Method for the Pressure-
best conformity amongst the predicted, experimental and numerical
Linked Equations (SIMPLE) scheme. The governing equations in
data as presented in the following model validations. Many researchers
conjunction with boundary and initial conditions are described by a
recommend considering the Amush value of 106. This value provides the
control volume approach. Central difference and QUICK schemes are
best overall relationship between their experimental and numerical data
used to describe the diffusion and convection terms, respectively. The
[47].
SIMPLER algorithm is applied for the velocity-pressure coupling. By
The energy equation is as follows:
( ) solving the governing equations at each time step, the liquid fraction can
(→ )
∂h ∂H k be updated using Eq. (17). The grid size and time steps were chosen after
+ + ∇⋅ V h = ∇⋅ ∇h (13)
∂t ∂t ρcp careful examination of the independence of the results for high accu­
racy. The iterative calculations continue until a relative convergence
where: H is the enthalpy of the material and is calculated as the total criterion of 10− 7 is satisfied by all the field variables of the problem.
value of the sensible enthalpy, h1 , and the latent heat ΔH: To validate the computational modeling of melting in our finite
H = h1 + ΔH (14) volume CFD code, an initial run was performed and compared to the
experimental data of Kamkari and Amlashi [45] for a vertical enclosure.
∫T Numerical model and experimental study were conducted at the same
h1 = href + Cp dT (15) operating conditions. Figure 4.shows the relationship between the liquid
Tref
fraction and time in the experimental study of [45] with the present
numerical model. It can be observed from the figure that a good
Where href is the enthalpy reference at temperature reference Tref and agreement between the present predicted study and experimental
latent heat term can be expressed in the form of the latent heat of the results.
PCMs (L) as follows:
ΔH = λ L (16) 5.3. Mesh dependency study

Where ΔH can vary from zero (solid) to L (liquid). Therefore, the ANSYS-Mesh 19.2 software was used to generate the mesh

5
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Fig. 4. Liquid fraction vs. time in the experimental model of [45] with the Fig. 6. History of average temperatures on the surface of the heater pipe.
present numerical study.

aluminum pipe with various heat fluxes with the numerical results. It is
independence test. The melting process was numerically predicted in the observed that the experimental and numerical results are very matching
storage unit using the ANSYS 19.2 program. The average PCM temper­ with each other.
ature and liquid fraction were compared for a different number of mesh The instantaneous contours of the temperature fields in the two
grid sizes to examine the influence of the grid size on the solution. The models (with and without convection) along the length of a cylindrical
PCM temperature and liquid fraction were plotted as a function of the container after 0.5, 1, 2, 3, 4, and 5.5 hr are given in Table 3. It is
melting time for various mesh cell sizes (6600, 10200, and 26112) in observed that in the case of without convection, the solid-liquid inter­
Figure 5-a and b, respectively, at a heat flux value of 1300 W/m2. It can face is in a flat wall shape through the melting process. The PCM tem­
be seen that increasing the element size from 10200 to 26112 does not perature is uniformly along the axial side with a high value between the
significantly affect the temperature and liquid fraction. Consequently, a heated wall and the interface layer, but after the interface, it is still at a
mesh distribution with an element size of 10200 was selected for the low value near to the solidification value.
present calculations. The finite volume method was followed in FLUENT In the case of convection, the PCM temperature is not uniformly and
to solve the continuity, momentum, and energy equations that describe changes through the whole of the domain, and referring to the natural
the heat transfer and fluid flow for the given boundary conditions. The convection effect, there is a recirculation region create in the narrow
residuals of the continuity and momentum equation were estimated to melting area, and the solid-liquid interface is in a curved shape.
be 10− 4 and 10− 6 in the energy equation. Due to the small melting area, a small recirculation region is pro­
duced at the beginning of the process. Then, due to the increase in the
6. Results and Discussion buoyancy effect, the molten region grows by advancing in time. By
raising the melting area, many recirculation regions merge to gather and
Experiments were started by changing the heat flux values from 1300 create a main recirculation region, and the dominant heat transfer
to 700 W/m2 through the melting process. Figure 6 illustrates the time mechanism would be the convection.
history of average heater temperatures recorded on the surface of the Figure 7 shows the variation of the PCM temperature versus the

Fig. 5. Parametric study on the element size of PCM (a) temperature and (b) liquid fraction.

6
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Table 3
Melting temperature contours for the cases of with and without convection.

(continued on next page)

7
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Table 3 (continued )

8
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

thermocouples (TC5, TC6, TC7, TC8) as shown in Figures 8 and 9,

respectively. The axial direction was chosen at a radius equal to 23 mm
(mid-radius), while the radial direction was selected at the height of 135
mm (mid-height).
It is observed from the Figures that the local temperature increases
with heat flux density. The simulated numerical results closely follow
the trends of the experimental data. The calculated temperature values
consistently overestimate the experimental data. The deviation between
the two results may be due to the consideration of the outer surface of a
well-insulated theoretical model. At the same time, since inadequate
insulation was used in the experimental work, it caused heat transfer
through the outer surface. Also, the theoretical model has many as­
sumptions not found in the experimental model.
The effect of heat flux on the time variation of PCM average nu­
merical and the experimental temperature is illustrated in Figure 10.
Experimental and numerical results tend to be the same, with increasing
heat flux shortening the melting time. The charging times are 5.50, 7.33,
and 9.75 hr for the heat fluxes of 1300, 1000, and 700 W/m2,
respectively.
Table 4 shows the changes in the instantaneous contours of the liquid
fraction in the two models (with and without convection). The liquid
Fig. 7. Numerical and experimental PCM temperature with time at 1300 fraction is uniform along the axial side in the absence of convection.
W/m2. However, it is not uniform in the convection case, and that refers to the
natural convection effect. Due to the convection effect, in the early
melting time at a heat flux of 1300 W/m2 in both experimental and stages of melting (t = .5 hr), the buoyant force cannot overcome the
numerical results for the two cases of with and without convection. The molecular viscous force; therefore, conduction is dominant for PCM.
numerical results in the presence of the convection effect are very near The layer of the molten liquid rises as melting continues, and
to the experimental value throughout the melting process. Through the buoyancy-driven convection is improved. The melted hot liquid moves
first 30 minutes, the numerical results in the case of without convection toward the top portion of the test cell adjacent to the heated wall. As
effect match the experimental results; after that, there is a big difference compared to the lower portion, the melting rate is more pronounced in
between them due to the convection effect. the upper part of the container, creating a noticeable curvature of the
From the experimental work and numerical results for the case of melting interface.
with convection effect, the temperature data versus time was recorded Natural convection obviously improves the melt fraction and the
through the PCM at heat flux densities of 700, 1000, and 1300 W/m2 for melting rate compared to the case of without convection. In the first
axial thermocouples (TC15, TC16, TC17, TC18) and radial stage, both models have the same value of average liquid reaction until

Fig. 8. Transient numerical and experimental temperature variations in the axial position.

9
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Fig. 9. Transient numerical and experimental temperature variations in the radial position.

Figure 12 shows a comparison of experimental and numerical results

of total PCM stored energy as a function of time at different heat flux
densities of 700, 1000, and 1300 W/m2. It can be observed from the
figure that a good agreement between the theoretical and experimental
results of total PCM stored energy. The total stored energy is approxi­
mately 470 kJ/kg in the storage unit after a complete melting of the
PCM.

7. Conclusions

In this study, the melting process of a phase change material (PCM)

was studied experimentally and numerically in a cylindrical tank heated
by three constant heat flux. Numerical investigations were performed to
evaluate the effect of convection on the thermal performance of melting
PCM.
Based on the current experimental/computational results of melting
PCM inside a cylindrical tank, the following conclusions can be drawn:

1 Heat transfer to the phase change material is greatly affected by

natural convection in the melting layer section, and the melting time
of PCM can be reduced due to the natural convection effect compared
to that without convection.
Fig. 10. PCM average numerical and experimental temperature at different
2 Melting is initially dominated by conduction; after that, natural
heat fluxes.
convection is promoted, leading to curved shapes of the solid-liquid
interface.
the convection mechanism started, and this period increased with the 3 Numerical results indicate a robust thermal stratification of the
decreasing of the heat flux, which begins roughly from .5, 1, and 1.5 hr molten liquid in the upper half of the storage unit.
at 1300, 1000, and 700 W/m2. 4 Several experiments have been performed to investigate the influ­
Figure 11 shows the effect of heat flux on the time variation of the ence of the intensity of the heat flux on the total melting time of PCM.
PCM liquid fraction. There is a big difference between the liquid fraction The applied heat flux controls the melting process, and by increasing
of the numerical results for the two cases of with and without convec­ the heat flux from 700 to 1000 and 1300 W/m2, the melting time was
tion. The liquid fraction in the convection case is greater than the value decreased by 24.82% to 43.58%, respectively.
in the case of without convection. Melting time decreases as a result of
increased enthalpy flow due to increased heat flux. It can be seen that
increasing the heat flux from 700 to 1000 and 1300 W/m2 reduces the
total melting time by 24.82 % to 43.58 %, respectively.

10
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Table 4
Liquid fraction contours for the cases of with and without convection.

(continued on next page)

11
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Table 4 (continued )

(continued on next page)

12
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

Table 4 (continued )

Fig. 11. Liquid fraction vs. time with and without convection.

Author statement presentation of the published work. Li Xu: presentation of the published
work, Supervision. Ping Cheng: presentation of the published work,
Credit author statement Ping Cheng: idea, Methodology, Writing - Supervision, Funding acquisition. Junguo Dong: Writing - original draft,
original draft, writing the initial draft. Ahmed Saad Soliman: Writing - Writing - review & editing, writing the initial draft, preparing literature
original draft, idea, computer program, writing the initial draft, Vali­ survey.
dation, Writing - review & editing, data writing, writing, review, and
editing. Shuping Zhu: Writing - original draft, writing initial draft,

13
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

[14] W. Wang, X. Yang, Y. Fang, J. Ding, J. Yan, Enhanced thermal conductivity and
thermal performance of form-stable composite phase change materials by using
β-Aluminum nitride, Appl. Energy. 86 (2009) 1196–1200.
[15] R.V. Seeniraj, R. Velraj, N.L. Narasimhan, Heat transfer enhancement study of a
LHTS unit containing dispersed high conductivity particles, J. Sol. Energy Eng.
Trans. ASME. 124 (2002) 243–249.
[16] A.A. Ranjbar, S. Kashani, S.F. Hosseinizadeh, M. Ghanbarpour, Numerical heat
transfer studies of a latent heat storage system containing nano-enhanced phase
change material, Therm. Sci. 15 (2011) 169–181.
[17] I. Zarma, M. Ahmed, S. Ookawara, Enhancing the performance of concentrator
photovoltaic systems using Nanoparticle-phase change material heat sinks, Energy
Convers. Manag. 179 (2019) 229–242.
[18] A. Hussanan, S. Aman, Z. Ismail, M.Z. Salleh, B. Widodo, Unsteady natural
convection of sodium alginate viscoplastic casson based based nanofluid flow over
a vertical plate with leading edge accretion/ablation, J. Adv. Res. Fluid Mech.
Therm. Sci. 45 (2018) 92–98.
[19] N. Aslfattahi, R. Saidur, N.A.C. Sidik, M.F.M. Sabri, M.H. Zahir, Experimental
assessment of a novel eutectic binary molten salt-based hexagonal boron nitride
nanocomposite as a promising PCM with enhanced specific heat capacity, J. Adv.
Res. Fluid Mech. Therm. Sci. 68 (2020) 73–85.
[20] N. Aslfattahi, R. Saidur, A. Arifutzzaman, R. Sadri, N. Bimbo, M.F.M. Sabri, P.
A. Maughan, L. Bouscarrat, R.J. Dawson, S.M. Said, B.T. Goh, N.A.C. Sidik,
Experimental investigation of energy storage properties and thermal conductivity
of a novel organic phase change material/MXene as A new class of
nanocomposites, J. Energy Storage. 27 (2020), 101115.
[21] N. Aslfattahi, A. Zendehboudi, S. Rahman, M.F. Mohd Sabri, S.M. Said,
A. Arifutzzaman, N.A. Che Sidik, Optimization of Thermal Conductivity of
NanoPCM-Based Graphene by Response Surface Methodology, J. Adv. Res. Fluid
Fig. 12. Stored energy vs. time for experimental and numerical results at Mech. Therm. Sci. 75 (2020) 108–125.
different heat fluxes. [22] N. Aslfattahi, R. Saidur, M.F.M. Sabri, A. Arifutzzaman, Experimental investigation
of thermal stability and enthalpy of eutectic alkali metal solar salt dispersed with
MgO nanoparticles 10 (2019) 1112–1119.
Declaration of Competing interest [23] N.I. Ibrahim, F.A. Al-Sulaiman, S. Rahman, B.S. Yilbas, A.Z. Sahin, Heat transfer
enhancement of phase change materials for thermal energy storage applications: A
critical review, Renew. Sustain. Energy Rev. 74 (2017) 26–50.
The authors declare that they have no known competing financial
[24] M.Z.M. Rizan, F.L. Tan, C.P. Tso, An experimental study of n-octadecane melting
interests or personal relationships that could have appeared to influence inside a sphere subjected to constant heat rate at surface, Int. Commun. Heat Mass
the work reported in this paper. Transf. 39 (2012) 1624–1630.
[25] Y. Kozak, T. Rozenfeld, G. Ziskind, Close-contact melting in vertical annular
enclosures with a non-isothermal base: Theoretical modeling and application to
Acknowledgment thermal storage, Int. J. Heat Mass Transf. 72 (2014) 114–127.
[26] A. Rozenfeld, Y. Kozak, T. Rozenfeld, G. Ziskind, Experimental demonstration,
modeling and analysis of a novel latent-heat thermal energy storage unit with a
The author would like to gratefully acknowledge the support of helical fin, Int. J. Heat Mass Transf. 110 (2017) 692–709.
Shanghai University, China, on the development of this research by [27] H. Shmueli, G. Ziskind, R. Letan, Melting in a vertical cylindrical tube: Numerical
providing access to their facility. investigation and comparison with experiments, Int. J. Heat Mass Transf. 53 (2010)
4082–4091.
[28] C. Liu, D. Groulx, Experimental study of the phase change heat transfer inside a
References horizontal cylindrical latent heat energy storage system, Int. J. Therm. Sci. 82
(2014) 100–110.
[29] R.E. Murray, D. Groulx, Experimental study of the phase change and energy
[1] M.J. Hosseini, A.A. Ranjbar, K. Sedighi, M. Rahimi, A combined experimental and
characteristics inside a cylindrical latent heat energy storage system: Part 1
computational study on the melting behavior of a medium temperature phase
consecutive charging and discharging, Renew. Energy. 62 (2014) 571–581.
change storage material inside shell and tube heat exchanger, Int. Commun. Heat
[30] M. Avci, M.Y. Yazici, Experimental study of thermal energy storage characteristics
Mass Transf. 39 (2012) 1416–1424.
of a paraffin in a horizontal tube-in-shell storage unit, Energy Convers. Manag. 73
[2] D. Guerraiche, C. Bougriou, K. Guerraiche, L. Valenzuela, Z. Driss, Experimental
(2013) 271–277.
and numerical study of a solar collector using phase change material as heat
[31] N.S. Dhaidan, J.M. Khodadadi, T.A. Al-hattab, S.M. Al-mashat, Experimental and
storage, J. Energy Storage. 27 (2020) 101–133.
numerical investigation of melting of phase change material / nanoparticle
[3] Alireza Haghighi, Aziz Babapoor, M. Azizi, Zahra Javanshir, Hooman Ghasemzade,
suspensions in a square container subjected to a constant heat flux, Int. J. Heat
Optimization of the thermal performance of PCM nanocomposites, J. Energy
Mass Transf. 66 (2013) 672–683.
Manag. Technol. 4 (2019) 14–19.
[32] B. Kamkari, H. Shokouhmand, F. Bruno, Experimental investigation of the effect of
[4] N.S. Dhaidan, J.M. Khodadadi, Melting and convection of phase change materials
inclination angle on convection-driven melting of phase change material in a
in different shape containers: A review, Renew. Sustain. Energy Rev. 43 (2015)
rectangular enclosure, Int. J. Heat Mass Transf. 72 (2014) 186–200.
449–477.
[33] L.W. Fan, Z.Q. Zhu, M.J. Liu, C.L. Xu, Y. Zeng, H. Lu, Z.T. Yu, Heat Transfer during
[5] B. Zalba, J.M Mar ın, L.F Cabeza, H. Mehling, Review on thermal energy storage
constrained melting of nano-enhanced phase change materials in a spherical
with phase change: materials, heat transfer analysis and applications, 2003.
capsule: An experimental study, J. Heat Transfer. 138 (2016) 1–9.
[6] H. Mehling, L.F. Cabeza, Phase change materials and their basic properties, Therm.
[34] N.S. Dhaidan, J.M. Khodadadi, T.A. Al-Hattab, S.M. Al-Mashat, Experimental and
Energy Storage Sustain. Energy Consum. NATO Sci. 234 (2007) 257–277.
numerical study of constrained melting of n-octadecane with CuO nanoparticle
[7] M.M. Kenisarin, K. Mahkamov, S.C. Costa, I. Makhkamova, Melting and
dispersions in a horizontal cylindrical capsule subjected to a constant heat flux, Int.
solidification of PCMs inside a spherical capsule: A critical review, J. Energy
J. Heat Mass Transf. 67 (2013) 523–534.
Storage. 27 (2020), 101082.
[35] M.H. Joneidi, M.J. Hosseini, A.A. Ranjbar, R. Bahrampoury, Experimental
[8] S. Jesumathy, M. Udayakumar, S. Suresh, Experimental study of enhanced heat
investigation of phase change in a cavity for varying heat flux and inclination
transfer by addition of CuO nanoparticle, Heat Mass Transf 48 (2012) 965–978.
angles, Exp. Therm. Fluid Sci. 88 (2017) 594–607.
[9] S.K. Saha, P. Dutta, Heat transfer correlations for PCM-based heat sinks with plate
[36] A.S. Menon, M.E. Weber, A.S. Mujumdar, The dynamics of energy storage for
fins, Appl. Therm. Eng. 30 (2010) 2485–2491.
paraffin wax in cylindrical containers, Can. J. Chem. Eng. 61 (1983) 647–653.
[10] V. Shatikian, G. Ziskind, R. Letan, Numerical investigation of a PCM-based heat
[37] D.S. Mehta, K. Solanki, M.K. Rathod, J. Banerjee, Thermal performance of shell and
sink with internal fins: Constant heat flux, Int. J. Heat Mass Transf. 51 (2008)
tube latent heat storage unit: Comparative assessment of horizontal and vertical
1488–1493.
orientation, J. Energy Storage. 23 (2019) 344–362.
[11] N. Japili, H. Rosali, N. Bachok, Suction effect on stagnation point flow and heat
[38] D.B. Khillarkar, Z.X. Gong, A.S. Mujumdar, Melting of a phase change material in
transfer over an exponentially shrinking sheet in a porous medium, J. Adv. Res.
concentric horizontal annuli of arbitrary cross-section, Appl. Therm. Eng. 20
Fluid Mech. Therm. Sci. 73 (2020) 163–174.
(2000) 893–912.
[12] B. Wu, Y.M. Xing, Numerical study of graphite foams saturated with phase change
[39] A. Sarı, K. Kaygusuz, Thermal and heat transfer characteristics in a latent heat
materials in power system thermal management, Asia-Pacific Power Energy Eng.
storage system using lauric acid, Energy Convers. Manag. 43 (2002) 2493–2507.
Conf. APPEEC. (2010).
[40] A. Trp, An experimental and numerical investigation of heat transfer during
[13] D. Gao, Z. Chen, M. Shi, Z. Wu, Study on the melting process of phase change
technical grade paraffin melting and solidification in a shell-and-tube latent
materials in metal foams using lattice Boltzmann method, Sci. China Technol. Sci.
thermal energy storage unit, Sol. Energy. 79 (2005) 648–660.
53 (2010) 3079–3087.

14
A.S. Soliman et al. Journal of Energy Storage 35 (2021) 102312

[41] G.H. Bagheri, M.A. Mehrabian, K. Hooman, Numerical study of the transient [45] B. Kamkari, H.J. Amlashi, Numerical simulation and experimental verification of
behaviour of a thermal storage module containing phase-change material, Proc. constrained melting of phase change material in inclined rectangular enclosures,
Inst. Mech. Eng. Part A J. Power Energy. 224 (2010) 505–516. Int. Commun. Heat Mass Transf. 88 (2017) 211–219.
[42] M. Mbaye, E. Bilgen, Phase change process by natural convection-diffusion in [46] A.D. Brent, V.R. Voller, K.J. Reid, Enthalpy-porosity technique for modeling
rectangular enclosures, Heat Mass Transf. Und Stoffuebertragung. 37 (2001) convection-diffusion phase change : application to the melting of a pure metal,
35–42. Numer. Heat Transf. 5720 (1988) 297–318.
[43] D. Pal, Y.K. Joshi, Melting in a side heated tall enclosure by a uniformly dissipating [47] M. Emam, M. Ahmed, Cooling concentrator photovoltaic systems using various
heat source, Int. J. Heat Mass Transf. 44 (2001) 375–387. configurations of phase-change material heat sinks, Energy Convers. Manag. 158
[44] Marimuthu.P Thirugnanam.C, Experimental analysis of latent heat thermal energy (2018) 298–314.
storage using paraffin wax as phase change material, Int. J. Eng. Innov. Technol. 3
(2013) 372–376.

15