Journal of Energy Storage 8 (2016) 12–26

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Exergy analysis of cascaded encapsulated phase change material—

High-temperature thermal energy storage systems
Laura Solomon, Alparslan Oztekin*
Mechanical Engineering and Mechanics, P.C. Rossin College of Engineering and Applied Science, Lehigh University, Bethlehem, PA 18015, USA

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

Article history:
Received 8 July 2016 The second law analysis of an example thermal energy storage (TES) system was conducted to determine
Received in revised form 14 September 2016 the benefit of a system employing a multiple phase change materials. Six systems were considered: three
Accepted 15 September 2016 single PCM systems (NaNO3, NaNO2, and KNO3), a 2-PCM system a 3-PCM system, and a sensible heat
Available online xxx only system as a comparison. The latent heat-based systems were able to store more energy and exergy
with comparable efficiencies than a system that relies on sensible heat only. Furthermore, when the
Keywords: overall cyclic performance was examined and showed that systems with multiple PCMs outperform their
Encapsulated phase change materials corresponding single PCM-based systems. While for the operating conditions and PCMs chosen the 2-
Exergy analysis
PCM system (NaNO3/NaNO2) was superior, great care is required during the design of an EPCM-based
Second law
TES system as the difference between the melting point of the PCMs and the inlet temperatures during
Thermal energy storage
Void charging and discharging greatly affect the performance of the system.
Concentrated solar power ã 2016 Elsevier Ltd. All rights reserved.

1. Introduction While the energy efficiency of a TES system is an important

The limited supply of fossil fuels in the world coupled with
ongoing concerns over the environmental impact of their use as
spurred a revival in research into clean renewable alternatives. One
promising area is solar thermal power produced at concentrating
solar power (CSP) plants. Often the solar field at a CSP plant can
absorb more energy than is required for baseline electricity
generation, however, without a means of energy storage, it is
difficult for these plants to match changes in demand and produce
power when there is no sun. Therefore integrating thermal energy
storage (TES) into CSP plants allows for a reduction in the levelized
cost of electricity making solar power cost competitive with the
current fossil fuel-based power plants. Currently, the majority of
TES implemented at CSP plants around the world are molten salt
sensible heat storage systems [1,2]. These systems require a large
volume of storage material as well as insulated storage tanks to
minimize heat loss. By utilizing a latent heat or thermochemical
energy storage system the volume of storage material required to
store the same amount of energy is immensely reduced.
Abbreviations: CSP, concentrating solar power; EPCMs, encapsulated phase
change materials; HTF, heat transfer fluid; PCMs, phase change materials; TES,
thermal energy storage.
* Corresponding author.
E-mail address: alo2@lehigh.edu (A. Oztekin).
factor in its performance and the overall system cost, the exergy
efficiency of the system should also be considered as the purpose
of the system is to store useful work and not simply energy [3]. The
first law analysis of a system does not reflect the quality of the
energy that is stored thus the second law analysis is required [4].
Exergy is the maximum useful work that can be produced by a
system as it comes to equilibrium with the surroundings. Since the
exergy efficiency (the ratio of exergy output to exergy input) of a
system accounts for internal irreversibility, it is often lower when
compared to the energy efficiency of the system [5–12].
Numerous numerical and experimental studies have been
conducted into both researching novel materials to be used as
PCMs [1,2,13–19] and into numerically evaluating the heat transfer
that occurs within encapsulated phase change material (EPCM)
capsules [20–25]. Only a few studies have investigated the
exergetic efficiency of latent heat based TES systems [26,27] and
none have focused on an EPCM based system. El-Dessouky and Al-
Juwayhel [28] studied the effect of the inlet temperature of the HTF
on the exergetic efficiency and showed that a maximum occurs
when the smallest possible temperature difference is used
between the initial PCM temperature and the inlet of the system
during charging. Ramayya and Ramesh [6] studied the effect of
sensible heating and sub-cooling on the performance of a shell and
tube-based latent heat system and their results showed that
sensible heating improves the efficiency of the melting process.
Additionally, they found that the optimal melting temperature is

http://dx.doi.org/10.1016/j.est.2016.09.004
2352-152X/ã 2016 Elsevier Ltd. All rights reserved.
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26 13

HTF as it residence time decreases. It should be noted that their

Nomenclature
analysis neglected the pumping power required during the exergy
analysis.
G Gravitational acceleration [m/s2]
It has often been shown in the literature that using a cascaded
k Turbulent kinetic energy [J/kg]
multiple-PCM system improves the energetic efficiency of latent
H Total enthalpy [J/kg]
heat based systems [30–34], however, limited studies have
hs Sensible enthalpy [J/kg]
considered the exergetic efficiency of these systems [35–39]. As
hsref Reference enthalpy [J/kg]
the temperature of the HTF decreases in the streamwise direction
L Latent heat of fusion [kJ/kg]
in a single-PCM system, multi-PCM systems utilizing PCMs with
m_ Mass flow rate [kg/s]
decreasing melting temperatures have better performance. Wata-
p Pressure [N/m2]
nabe and Kanzawa [35] showed that the rapid charging and
r Radius [m]
discharging seen in multi-PCM systems leads to high charging,
t Time [s]
discharging, and overall exergy efficiencies. Domanski and Fellah
Tm Melting temperature [K]
[38] reported a 40% increase in overall efficiency for a two-PCM
Tref Reference temperature [K]
system over that of a single-PCM system. Gong and Mujumdar [39]
u Velocity component [m/s]
reported similar results that showed a 74% increase for a three-
x X coordinate [m]
PCM system due to the decrease in the charging and discharging
y Y coordinate [m]
time of the system. Li et al. [40] studied the use of a two-PCM
cp Specific heat [J/kg K]
system for TES using finite-time thermodynamics and showed that
Re Reynolds number [–]
an increase of 19–54% in efficiency is possible over that of a single-
Ste Stefan number [–]
PCM system. Shabgard et al. [41] examined a cascaded latent heat
Fo Fourier number [–]
storage system with gravity-assisted heat pipes and reported that
the cascaded system recovered 10% more exergy over a 24-h cycle
Greek symbols
compared to the best non-cascaded system considered. Recently,
a Volume fraction [–] Mosaffa et al. [42] reported on the energy and exergy performance
b Thermal expansion coefficient [1/K] of a multi-PCM system for free cooling applications. Their study
g Liquid fraction [–] showed that higher exergy efficiency was achieved when using
t ij Reynold’s stresses [N/m2] multiple PCMs and that decreasing the temperature difference
m Dynamic viscosity [Pa s] between the HTF and PCM yielded an increase in efficiency with
n Kinematic viscosity [m2/s] time.
r Density [kg/m3] While these initial studies are a starting point, there are no
v Specific dissipation rate [1/s] reported studies on the exergy performance of EPCM-based latent
k Thermal conductivity [W/m2 K] heat TES systems for high-temperature applications. Therefore a
mt Turbulent viscosity [Pa s] numerical investigation was conducted for an example EPCM-
based system to determine if the trend of an increase in
Subscripts
performance would be seen for these systems as well. Based on
i Component
the results of this investigation a greater understanding of the key
j Component
factors in the behavior of large-scale TES systems will be gained.
Present results can be used to further optimize high-temperature
TES modules to improve the performance of CSP plants.
higher for systems that include sensible heating and sub-cooling
and that increasing the inlet HTF temperature during discharging 2. Mathematical modeling and numerical solutions
increases the efficiency. Recently, Singh et al. [29] reported the
effect of varying inlet temperatures in the exergy performance of A 2-D numerical analysis was conducted to investigate the
NaCl impregnated graphite foam where they saw a similar trend increase in the energy and exergy efficiency of an EPCM-based
and concluded that the optimal inlet temperature was 880
C. In latent heat TES system by employing multiple PCMs. Six systems
addition to the inlet temperature, the flow rate of the HTF is an were considered: three single PCM systems (NaNO3, NaNO2, and
important factor in the performance of a latent heat-based TES KNO3), a 2-PCM system (NaNO3 and NaNO2), a 3-PCM system, and a
system. Although an increase in the flow rate increases the sensible heat only system as a comparison. The thermal properties
required pumping power and the resulting pressure drop across used are listed in Table 1. Each system consisted of 72 cylindrical
the system both leading to higher entropy generation and lower EPCM stainless steel capsules that had a diameter of 76.2 mm with
efficiency Erek and Dincer [11] showed the opposite to be true a shell thickness of 1.5875 mm. A staggered arrangement of the
possibly due to the smaller temperature drop that is seen in the capsules used to minimize the wake regions behind each capsules

Table 1
Properties of NaNO3, NaNO2, KNO3, air, and stainless steel used in exergy simulations.

NaNO3 NaNO2 KNO3 Air Stainless Steel

Melting Temperature (K) 581 [43] 555 [44] 610 [44] – 1672 [45]
Density (kg/m3) 1900 [43] 1812 [44] 1870 [44] 0.5214 [45] 7900[45]
Viscosity (Ns/m2) 0.00285 [43] 0.002666 [44] 0.002367 [44] 3.65  105 [45] –
Thermal Conductivity (W/mK) 0.550/0.680 [46] 0.665/0.765 [44] 0.481/0.878 [44] 0.0242 [45] 14.7[45]
Solid Heat Capacity(kJ/kg K) 1.588 [19] 1.733 [44] 1.240 [44] – 477 [45]
Liquid Heat Capacity(kJ/kg K) 1.650 [19] 2.553 [44] 1.341 [44] – –
Gas Heat Capacity(kJ/kg K) – – – 1.0064 [45] –
Latent Heat (kJ/kg) 162.5 [19] 180.12 [44] 99.73 [44] – –
14 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26

row typical of cylinders placed in a cross flow and therefore Turbulent flow of the HTF is modeled using the k-v SST model
enhance the overall heat transfer rate. The capsules are placed two [54]. The turbulent kinetic energy (k) and specific dissipation rate
diameters apart center to center within each row and the rows are (v) transport equations are listed in Eqs. (4) and (5), respectively,
one diameter apart. The inlet and outlet are 10 diameters above/ and the turbulent viscosity is defined in Eq. (6)
below the first/last row of EPCM capsules respectively. Lastly, the  
@ @ u @ @k 
wall is located one diameter from the center of the last column of ðrkÞ þ ðrkuj Þ ¼ t ij i þ ðm þ s k mt Þ  rb kv ð4Þ
capsules. A 20% void space was considered at the top of the EPCM
@t @xj xj @xj @xj
capsules based on the thermal expansion (b) values of the    
@ @ @ @v jr @u
materials considered [23–25]. The three chosen materials had ðrvÞ þ ðrvuj Þ ¼ ðm þ s v mt Þ þ t i
similar values of b approximately equal to 4.0  104 1/K. However, @t @xj @xj @xj mt ij @xj
rs v2 @k @v
convection within the capsules is neglected due to the required  brv2 þ 2ð1  F 1 Þ ð5Þ
computational resources for the system size considered in the v @xj @xj
present study. The computational domain is presented in Fig. 1. Air
ra1 k
was used as the HTF with an inlet mass flux of 0.2607 kg/m2s. The mt ¼ ð6Þ
maxða1 v; VF 2 Þ
initial temperature of the system was 551 K for the start of the
charging process. The capsules were then heated for 12 h with an where ui is the velocity vector, r is the density, m is the viscosity, V
inlet temperature of 611 K. At the end of the charging process, the is the vorticity magnitude (Eq. (13)) and s k, b*, s v, b, and a1 are
direction of the HTF was reversed and it now enters the system closure constants. Unlike in the standard k-v model, the constants
from the bottom with a temperature of 551 K. The capsules are then in the k-v SST model are a blend of the k-v and k-e model constants
cooled for 12 h. as defined in Eq. (7) for a constant f and additional functions are
The phase change within the capsules was modeled using the defined in Eqs. (8)–(12) and the model constants are listed in
enthalpy-porosity method within Ansys FLUENT. Numerous Table 2. Note that the flow properties are time averaged values
investigations have used the enthalpy-porosity method to study throughout. The above set of governing equations is solved
phase change [21–25,47–51]. First developed by Voller et al. [52] it iteratively until a convergence of 103 was reached for the
indirectly tracks the location of the solid-liquid interface using a continuity, momentum, and turbulence equations, and a value of
parameter called the liquid fraction, g . The computational domain 106 for the energy equation.
is broken down into three regions (solid, liquid, and mushy) based
on the value of the g within a given cell. The liquid fraction is a f ¼ F 1 f1 þ ð1  F 1 Þf2 ð7Þ
function of temperature and is defined as:
8 F 1 ¼ tanhðarg41 Þ ð8Þ
>
> 0 T < T solidus
< T T " ! #
solidus pffiffiffi
g¼ T  T solidus
T solidus < T < T liquidus ð1Þ k 500n 4rs v2 k
>
> arg1 ¼ min max ; ; ð9Þ
: liquidus b vd d2 v CDkv d2
1 T > T liquidus

where Tsolidus and Tliquidus are the solidus and liquidus temperatures  
1 @k @v
of the PCM, respectively. The size of the mushy zone is based on the CDkv ¼ max 2rs v2 ; 1020 ð10Þ
difference between the solidus and liquidus temperatures. For a
v@xj @xj
congruently melting material the solidus and liquidus temperature
are equal. A cell is solid when g = 0 and it is liquid when g = 1. The F 2 ¼ tanhðarg22 Þ ð11Þ
mushy zone (0 < g < 1) is modeled as a “pseudo” porous material !
pffiffiffi
using a source term in the momentum equation that dampens the k 500n
arg2 ¼ max 2 ; ð12Þ
velocity to zero as the liquid fraction goes to zero. The governing b vd d2 v
equations for heat transfer involving a phase change are given
below [53]: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Z T
V¼ 2W ij W ij ð13Þ
H ¼ hsref þ ðcp dTÞ þ g L ð2Þ  
T ref 1 @ui @uj
W ij ¼ þ ð14Þ
2 @xj @xi
  Table 2
@ @ @T
ðrHÞ ¼ k ð3Þ Model constants employed in the closure of the k-v SST Model.
@t @xj @xj
j1 ¼ bb1 spv1ffiffiffiffi
l
j2 ¼ bb2 spv2ffiffiffiffi
l
2 2

where j is summation index, H is the total enthalpy, k is the thermal

 
b b
s k1 ¼ 0:85
conductivity, T is the temperature, hsref is the sensible reference s v1 ¼ 0:5 b1 ¼ 0:075
enthalpy, Tref is the reference temperature, cp is the specific heat s k2 ¼ 1:0 s v2 ¼ 0:856 b2 ¼ 0:0828
capacity, and L is the latent heat of fusion. b ¼ 0:09 l ¼ 0:41 a1 ¼ 0:31

Fig. 1. The computational domain used for exergy evaluation for the charging process. Note that the inlet and outlet are reversed for the discharging process.
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26 15

To simplify the energy and exergy calculations, a uniform
temperature distribution in the capsules is assumed and the value
at the center of the capsule is used. The energy and exergy stored in
the PCM for each capsule in the system were individually
calculated using the following equations and the temperature at
the capsule center [4]:
where m _ HTF is the mass flow rate of the HTF, cp,HTF is the heat
capacity of the HTF, THTF,in is the inlet temperature of the HTF
during charging, and Dt is the elapsed time.
Again ideally the below set of equations would be solved for
each cell within the domain and then summed to calculate the total
energy and exergy released during the discharging process.

f
However, as with the charging process, a uniform temperature
mi cspi ðT  T i Þ T < T si
distribution within the capsules is assumed and the value at the
mi cspi ðT mi  T i Þ þ mi g Li T si < T < T li
Q stored ¼ temperature at the centers of the capsules was used for the energy
mi cspi ðT mi  T i Þ þ mi g Li þ mi clpi ðT  T mi Þ T > T li and exergy calculations during discharging. The energy and exergy
ð15Þ released from the PCM at any given instant are calculated

f
  
T
mi cspi ðT  T i Þ  T o ln T < T si
Ti
     
T To
mi cspi ðT mi  T i Þ  T o ln mi þ mi g Li 1  T si < T < T li
Ti T
Exstored ¼          ð16Þ
T To T
mi cspi ðT mi  T i Þ  T o ln mi þ mi g Li 1  þ mi clpi ðT  T mi Þ  T o ln T > T li
Ti T mi T mi

where the subscript i is used to denote the properties of one of the according to following equations:
three PCMs used, T is the temperature at the center of the capsule, 8
Tsi and Tli are the solidus and liquidus temperature of a given PCM, >
> mi clpi ðT char  T Þ T > T li
>
<
and To is the temperature of the surroundings taken as 300 K. Note Q relassed ¼ mi clpi ðT char  T mi Þ þ mi g Li T si < T < T li
that the materials considered here are congruently melting >
>
>
:
materials and therefore the solidus and liquidus temperatures mi clpi ðT char  T mi Þ þ mi g Li þ mi cspi ðT mi  T Þ T < T si
are equal. Ideally the above set of equations would be solved for ð19Þ
each cell within the capsules and then summed to determine the

f
  
T
mi clpi ðT char  TÞ  T o ln char T > T li
 T    
T To
mi cpi ðT char  T mi Þ  T o ln char
l
þ mi g Li 1  T si < T < T li
  T mi    T    
Exreleased ¼ T char To T ð20Þ
mi cpi ðT char  T mi Þ  T o ln
l
þ mi g Li 1  þ mi cspi ðT mi  TÞ  T o ln mi T < T si
T mi T mi T

total energy and exergy stored. However due to the large where Tchar is the final temperature of the EPCM capsule at the end
computational demand required, a uniform temperature distribu- of the charging process. F the discharging process, the efficiency is
tion within the capsules was assumed and the temperature at the defined by Eq. (24) as exergy gained by the HTF divided by the total
center of the capsules was used. Note that this method will under exergy stored by the PCM [4].
predict the total energy stored as the PCM near the capsule shell
will experience a higher degree of super heating than the center of
X
n   
T
the capsule. ExHTF ¼ _ HTF cp;HTF ðT HTF;out  T HTF Þ  T o ln HTF;out
m
The efficiency of the charging process is defined in Eq. (18) and j¼1
T HTF j

compares the exergy stored by the PCM to the maximum exergy  Dt j ð21Þ
supplied [4]. The exergy possessed by the HTF is given by Eq. (17)
and neglects the input power required to pump the HTF through
the system. ExHTF
hdis ¼ ð22Þ
Exstored
  
  T HTF;in where THTF,out is the outlet temperature of the HTF during
_ HTF cp;HTF
Exinput ¼ m T HTF;in  T o  T o ln  Dt ð17Þ
To discharging, THTF is the inlet temperature of the HTF during
discharging, Dtj is the time step, j is the index indicating time, and n
is the number of time steps to the current time. n assumes value 1
Exstored to N with 1 representing the initial time and N is representing final
hchar ¼ ð18Þ
Exinput time, tf = NDt. The overall energy and exergy efficiencies at any
16 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26
given instant are defined as:
Q released
hQ o verall ¼
Q stored
Exreleased
hExo verall ¼
Exstored
For the system considered here, the end of charging was taken as
the state of the system after a 12 h charging period. Alternatively,
the charging process can be considered complete when all of the
PCM within the systems has undergone a complete phase change.
3. Results and discussions
A numerical analysis of an EPCM-based TES system was
conducted to examine the first and the second law efficiency. As
the purpose of a TES system is to store useful work, the first law
analysis of the system is incomplete as it does not account for
irreversibility. Exergy is a measure of the quality of the energy that
was stored and determines the amount of useful work the system
can produce. The charging and discharging process are examined
individually in addition to the overall cyclic performance of the
system. A mesh independence analysis was conducted using a
quadrilateral dominant mesh with element sizes ranging from
5  104 to 5  105 using the pure NaNO3 system. All three mesh
sizes tested resulted in the same total energy and exergy stored in
the system and therefore an element size of 1 104 was used. A
detailed validation of the numerical methods used has been
previously reported in the literature [22,23].
3.1. Charging process results
The HTF as an inlet velocity of 0.5 m/s which results in a Re
number of 544 based on the diameter of the capsules. As the
velocity within the system is initially zero, there was an
approximately 10 s transient period as the flow within the system
is established. After this flow has reached steady state and remains
unchanged over the remainder of the charging process, Fig. 2. The
staggered arrangement of the capsules was used to enhance
momentum mixing in the vessel and promote a uniform heat
transfer coefficient around the capsules. A stagnation point forms
at the top of each capsule and large wakes are seen behind the last
row of EPCM capsules. The column of capsules that is next to the
wall of the system see a slight increase in velocity compared to the
rest of the system due to boundary layer effects. As expected of
flow around a cylinder, the highest velocity occurs 90
around the
Fig. 2. Velocity field after 5 s (top), 100 s (middle), and 40,400 s (bottom) at Re = 544.
capsules with a magnitude 2.6–2.8 times higher than the inlet
velocity, and flow separation occurs slightly downstream of the
ð23Þ high-velocity region.
The change in the outlet temperature during the charging
process for the six systems considered is shown in Fig. 3. In the
system that stores only sensible heat, the outlet temperature
ð24Þ increases asymptotically to the inlet temperature. A similar
behavior occurs for the KNO3 system, as the melting temperature
is extremely close to the inlet temperature. In the NaNO3, the outlet
temperature rapidly increases until the capsules reach the 581 K
melting point of NaNO3 (7000 s), then only a slight increase
occurs until the time at which all the capsules have completed
melting (32,000 s) when it begins to increase asymptotically
towards the inlet temperature. A similar trend was seen for the
NaNO2 system, however, it has a higher slope during the melting
period due to the larger temperature difference between the inlet
and melting temperatures. The 2-PCM system is a middle ground
between the NaNO3 and NaNO2 systems as it has a higher slope
than the NaNO3 system and a higher temperature than the NaNO2
system. The 3-PCM system has the most consistent energy storage
rate during the charging process with the presence of two
inflection points as the NaNO2 and NaNO3 capsules are completely
melted.
While the evolution of the outlet temperature lends insight into
the behavior of the system, a greater understanding can be gained
by looking at the temperature evolution of the entire system.
Initially, the sensible heat and KNO3 systems have the smallest
temperature drop in the HTF across the system with a row-wise
temperature distribution in the EPCM capsules. The NaNO3,
NaNO2, 2-PCM, and 3-PCM cases all have a similar temperature
drop within the HTF. The NaNO3 capsules in the NaNO3, 2-PCM, and
3-PCM systems are at the 581 K melting temperature. The first few
rows of NaNO2 capsules in the NaNO2, 2-PCM, and 3-PCM systems
have developed a radial thermal gradient as the PCM begins to
melt.
Nearly midway through the charging process, the temperature
drop in the sensible heat only system has greatly reduced, Fig. 4a.
An even larger reduction was seen in the temperature drop of the
HTF in the KNO3 system as well as in the row-wise gradient within
the EPCM capsules, Fig. 4d. A nearly identical temperature
distribution in the HTF is seen between the NaNO3 and 3-PCM
cases, Fig. 4b and f. The NaNO3 capsules in both cases are still at the
581 K melting temperature. Additionally, in the 3-PCM system, the
KNO3 capsules have a row-wise decreases temperature while the
radial gradient in the NaNO2 capsules has increased as the melt
fraction in the capsules has increased. Lastly, a similar temperature
drop of the HTF is seen in the NaNO2 and 2-PCM system, Fig. 4c and
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26 17

Fig. 3. Outlet temperatures during the charging process for the NaNO3 (solid), NaNO2 (square), KNO3 (circles), NaNO3/NaNO2 (dashed), KNO3/NaNO3/NaNO2 (dash-dot), and
sensible heat (dot) systems.

Fig. 4. Isotherms at 20,400 s for a) sensible heat only, b) NaNO3 (SteFo = 0.93), c) NaNO2 (SteFo = 1.85), d) KNO3 (SteFo = 0.066), e) NaNO3/NaNO2 (SteFo = 1.48), f) KNO3/NaNO3/
NaNO2 (SteFo = 1.27) systems.

e. The capsules in the top half of the NaNO2 system have finished capsules has increased as the capsules continue to melt. In the 2-
melting resulting in the formation of decreasing row-wise PCM system, the same radial temperature distribution is seen in
temperature distribution while the radial gradient in the bottom the NaNO2 system while the NaNO3 capsules remain at 581 K.
18 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26
After 30,400 s the KNO3 system was at a near uniform
temperature as the PCM finally beings to melt while a slightly
larger drop was seen in the sensible heat system. As the capsules in
the NaNO3 system are still melting they remain at 581 K and there
is a minimal change in the temperature of the HTF. The capsules in
the NaNO2 system have finished melting and now have a row-wise
temperature distribution. Furthermore, the temperature drop of
the HTF as decreased and is now closer to that seen in the NaNO3
system. Although only half of the NaNO3 capsules have melted, the
temperature distribution of the capsules and HTF in the 2-PCM
system resembles that of the NaNO2 system. The temperature drop
across the HTF for the 3-PCM system has reduced and is not smaller
than the 2-PCM system. Additionally, each block of PCM capsules is
at its own nominal temperature with a slight row-wise tempera-
ture gradient.
Near the end of the charging process, the systems are
approaching a uniform temperature distribution. Both the sensible
heat and the KNO3 system no longer exhibit a noticeable drop in
temperature across the system. The largest temperature drop that
remains in the HTF is in the NaNO3 system as the capsules have not
fully completed melting after 40,400 s. The row-wise temperature
gradient has decreased in the NaNO2 and the 2-PCM system again
lies between the NaNO2 and NaNO3 systems. For the 3-PCM
Fig. 5. Liquid fraction at 20,400 s for a) NaNO3 (SteFo = 0.93), b) NaNO2 (SteFo = 1.85), c) KNO3 (SteFo = 0.066), d) NaNO3/NaNO2 (SteFo = 1.48), e) KNO3/NaNO3/NaNO2
(SteFo = 1.27) systems.
system, there is a slight decrease in temperature across the HTF
and a row-wise distribution remains in the bottom of the system.
Although the staggered arrangement of the EPCM capsules
promotes a uniform heat transfer coefficient around the capsules,
the presence of the void reduces the heat transfer rate in the upper
portion of the capsules leading to a “U”-shaped solid-liquid
interface. While neglected here the effect of convection on the
evolution of the solid-liquid was previously examined by the
current authors [23–25]. With the exception of the KNO3 system,
partial melting over the length of the system was seen. The melt
fraction within the system at a given time varied based on the
PCMs within the system. The liquid fraction in all six systems is
shown in Fig. 5 after 20,400 s. Please note that the vertical line
present within the capsules is the void-PCM interface. Additionally,
red corresponds to a liquid phase whereas blue denotes a solid
region. All of the capsules in the NaNO3 systems have started
melting with the first 24 capsules being approximately halfway
melted, Fig. 5a. Whereas, the capsules in the top half of the NaNO2
system are completely melted and the bottom capsules are over
50% melted due to the higher heat transfer rate resulting from the
lower melting point of NaNO2 compared to NaNO3, Fig. 5b. In
contrast, the capsules in the KNO3 system have yet to start melting
due to their high melting point, Fig. 5c. Furthermore, when the
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26 19

Fig. 6. Temperature at the center (a) and the energy stored (b) for each EPCM capsule for the KNO3/NaNO3/NaNO2 system.

system does begin to melt it does so in a row-wise manner, unlike
the other systems. A similar melt fraction is seen in the NaNO3 and
NaNO2 capsules in the 2- and 3-PCM systems as seen in their single
PCM systems.
The NaNO2 system had the shortest melting time of 30,411 s.
The 2- and 3-PCM systems had melting times of 36,245 s and
35,050 s, respectively. The NaNO3 system takes 41,100 s to melt
completely while the KNO3 system barely completes melting by
the end of the charging process. Thus with the exception of the
NaNO2 systems, the multi-PCM systems improved the energy
storage rate over that of their base single PCM systems.
The temperature at the center of the EPCM capsules in the 3-
PCM system is shown in Fig. 6 along with the individual energy
stored in each capsule as a function of time. As the energy
calculations use the temperature at the center of the capsule, a
jump in stored energy occurs when the melting point is reached.
Following this, there is a period of time where the energy storage
remains constant while the melting front propagates inward. Once
the capsule has completely melted the temperature and thus
energy stored begin to increase.
The total energy stored by the system was determined by
summing the individual energy stored as a function of time. A
similar procedure was used to calculate the exergy content within
the system. Both the cumulative energy and exergy within the six
systems considered is presented in Fig. 7. For the 3-PCM system,
the NaNO2 capsules melt at 3750 s resulting in the first jump in
energy storage. After this, the increase in stored energy is from the
sensible heating of the NaNO3 and KNO3 capsules. At 10,000 s the
NaNO3 capsules melt resulting in the second jump in energy
storage. The energy storage rate then begins to level off as the only
increase in energy storage is from the KNO3 capsules until at
20,000 s the NaNO2 capsules sequentially complete melting and
20 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26

Fig. 7. Energy (a) and exergy (b) stored in the NaNO3 (solid), NaNO2 (square), KNO3 (circles), NaNO3/NaNO2 (dashed), KNO3/NaNO3/NaNO2 (dash-dot), and sensible heat (dot)
systems during the charging process.

begin super heating. Around 25,000 s the NaNO3 capsules are
completing the melting process at the same time the KNO3 have
begun melting. The row-wise melting of the KNO3 capsules leads to
the third increase in energy storage having a lower slope compared
to the first two. Similar trends are seen in the other five systems.
The NaNO2 system stored the most energy (142 MJ) while the KNO3
system stores the least (84.4 MJ). This resulted from the higher
energy storage density of the NaNO2 system compared to the KNO3
system and the lack of liquid superheating in the KNO3 system. As
the operating conditions of the systems are the same, the systems
store energy at nearly the same temperature, therefore the main
contributor to the exergy content within the system is the total
energy that has been stored. Thus, the NaNO2 system has the
highest exergy content at the end of the charging process.
The charging efficiency of the six systems considered is
presented in Fig. 8. A spike in the efficiency occurred when a
set of capsules melts. The large temperature difference between
the melting point of NaNO2 and the inlet coupled with the latent
heat of fusion result in the system storing a large amount of energy
in a short period of time leading to it having the highest peak
efficiency of approximately 38%. This also results in the 2-PCM
system having a higher efficiency over that of the 3-PCM system.
The KNO3 and sensible heat only systems have a similar behavior
where the efficiency increases during the initial stages while there
is still a large temperature difference and then decreases as the
system approaches equilibrium. In the KNO3 system, however,
there is an increase in efficiency when the capsules begin to melt,
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26 21

Fig. 8. Exergy efficiency of the charging process.

although as melting occurs over a longer period of time it is not as
steep as the ones seen in other systems.
3.2. Discharging process results
At the end of the 12 h charging process, the direction of the HTF
was reversed and it now enters the system from the bottom with a
temperature of 551 K. When the direction of the HTF is reversed
there is an initial transient period while the flow field reestablishes
into the mirror image of that seen during the melting process. The
evolution of the outlet temperature during the discharging process
is shown in Fig. 9. In a manner similar to that of the charging
process, the outlet temperature of the sensible heat system
asymptotically decreases to the inlet temperature and in the
NaNO3 system the temperature first decreases to the melting
temperature at which point it slowly decreases while the capsules
solidify and then asymptotically decreases towards the inlet
temperature after the capsules have completely solidified. The
biggest difference between the charging and discharging process is
in the behavior of the NaNO2 and KNO3 systems are reversed where
now the NaNO2 system behaves similarly to the sensible heat
system as its melting point is only four degrees above the 551 K
inlet temperature.
At the onset of the discharging process, the temperature within
the EPCM capsules rapidly decreases to their respective melting
points. A nearly identical temperature distribution occurs in the
sensible heat, NaNO3, NaNO2, and 2-PCM systems. The KNO3
capsules in the KNO3 and 3-PCM systems have a radial thermal

Fig. 9. Outlet temperatures during the discharging process for the NaNO3 (solid), NaNO2 (square), KNO3 (circles), NaNO3/NaNO2 (dashed), KNO3/NaNO3/NaNO2 (dash-dot),
and sensible heat (dot) systems.
22 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26
Fig. 10. Isotherms at 70,000 s for a) sensible heat only, b) NaNO3 (SteFo = 0.098), c) NaNO2 (0.15), d) KNO3 (2.80), e) NaNO3/NaNO2 (0.12), f) KNO3/NaNO3/NaNO2 (1.00)
systems (26,800 s into discharging).
gradient as the capsules solidify. This also leads to the systems
having a larger temperature increase within the HTF as latent
energy is being removed from the systems. As time progresses the
KNO3 capsules in the KNO3 and 3-PCM system complete
solidification leading to a row-wise temperature gradient of the
capsules. The NaNO3 capsules remain at the 581 K melting point.
There is also a row-wise temperature gradient present for the
NaNO2 capsules as they solidify row by row.
After 70,000 s (26,800 into the discharging process), while there
is a row-wise temperature distribution in the capsules in the
sensible heat system, there is no discernible temperature gradient
within the HTF in the flow direction, Fig. 10a. The NaNO3 system
has the largest increase in the temperature of the HTF as the
capsules continue to solidify, Fig. 10b. While the NaNO2 and KNO3
systems have similar temperature distributions, the capsules in the
NaNO2 system have just begun to solidify whereas the KNO3
system is completely solid, Fig. 10c and 10d respectively. In the 2-
PCM system, the NaNO3 capsules remain at 581 K while they
continue to solidify whereas there is a row-wise gradient in the
NaNO2 capsules as they begin to solidify, Fig. 10e. The 3-PCM
system has a slightly lower temperature increase within the HTF
than the 2-PCM system. Furthermore, as the KNO3 and NaNO3
capsules have solidified completely there is a row-wise thermal
gradient present in the capsules as the NaNO2 capsules solidify row
by row, Fig. 10f.
Towards the end of the discharging process, there is a near
uniform temperature distribution in the sensible heat system. The
NaNO3 system still has the largest increase in the temperature of
the HTF across the system as the capsules at the stop of the system
are still releasing energy. While there is not temperature change in
the HTF of the NaNO2 system, there is a row-wise temperature
distribution within the EPCM capsules as they continue to solidify
one row at a time. The temperature in the KNO3, 2-PCM and 3-PCM
systems is similar to that seen in the sensible heat only system as
these systems are either completely or nearly completely solidified
and therefore have released the majority of their stored energy.
Although convection within the molten PCM has been
neglected, previous studies have shown that the solidification
process is conduction dominated [22,23,55] and therefore the
result presented here will be unaffected. As the HTF enters the
system from the bottom during the discharging process, the PCM
begins to solidify along the bottom edge of the capsule first. The
solid-liquid interface is “U”-shaped throughout the solidification
process and it slowly propagates radially inward. The NaNO3 and
KNO3 capsules in the NaNO3, KNO3, 2-PCM, and 3-PCM systems
begin to solidify shortly after the discharging process is initiated.
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26
Fig. 11. Liquid fraction at 70,000 s for a) NaNO3 (SteFo = 0.098), c) NaNO2 (0.15), d) KNO3 (2.80), e) NaNO3/NaNO2 (0.12), f) KNO3/NaNO3/NaNO2 (1.00) systems (26,800 s into
discharging).
The NaNO2 capsules, however, take longer to reach the 555 K
melting point leading to a delay in the solidification process. As
discharging of the systems continues, the solid fraction within the
NaNO3 systems has increased where now the bottom third of the
systems is over halfway solidified. The capsules in the KNO3 system
have completely solidified while the capsules in the NaNO2 have
still yet to begin to solidify. This trend is seen in the multi-PCM
systems where the KNO3 capsules are completely solid, the solid
fraction within the NaNO3 capsules as increased, and the NaNO2
capsules have not begun to solidify.
After 26,800 s the solid fraction within the NaNO3 system has
continued to increase where now the last 24 capsules are
completely solid, the middle capsules are approximately 50%
solidified, and the top capsules are about 25% solidified, Fig. 11a.
The capsules in the NaNO2 system have finally begun to solidify
and as seen with the KNO3 system during melting, the capsules are
solidifying on diagonal rows, Fig. 11b. Again the capsules in the
KNO3 are completely solidified, Fig. 11c. In the 2-PCM system, the
NaNO3 capsules have almost completely solidified while the same
row-wise solidification is seen in the NaNO2 capsules, Fig. 11d. In
the 3-PCM system, the KNO3 and NaNO3 capsules have solidified
completely while the NaNO2 capsules have started to solidify in a
row-wise manner, Fig. 11e. The KNO3 system has the shortest
solidification time of 61,505 s while the NaNO3 system takes
82,900 s. The NaNO2 system does not solidify completely over the
23
course of the 12-h discharging process. Furthermore, with the
exception of the KNO3 system, the 2- and 3-PCM systems offer an
improvement over the single PCM systems and have total
solidification times of 81,695 s and 80,020 s, respectively.
As was done for the charging process, the energy released by
the individual EPCM capsules was summed to determine the total
energy storage within the system as a function of time, Fig. 12a.
Although energy is being released from the capsules, the shape of
the curves is similar to that seen during the charging process
where sudden drops occur when the capsules begin to solidify
followed by a period of constant energy storage as the solid-liquid
interface propagates radially inward before sub-cooling can occur.
The 2-PCM system releases the most energy (128 MJ) closely
followed by the 3-PCM system (125 MJ). The KNO3 systems only
released 83.8 MJ of energy but it also has the smallest maximum
energy storage capability due to the low latent heat of fusion. The
exergy extracted from the PCM follows a similar trend as that of the
energy released and is seen in Fig. 12b.
The efficiency of the systems considered during the discharging
process is shown in Fig. 13. The NaNO3 system has the highest
efficiency (74%) due to the prolonged solidification period whereas
the NaNO2 system has the lowest (26%). For all of the systems, the
efficiency increases during the beginning of the discharging
process due to the large temperature increase between the inlet
and outlet of the HTF as it extracts energy from the PCM. As the
24 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26

Fig. 12. Energy (a) and exergy (b) content in the NaNO3 (solid), NaNO2 (square), KNO3 (circles), NaNO3/NaNO2 (dashed), KNO3/NaNO3/NaNO2 (dash-dot), and sensible heat
(dot) systems during the discharging process.

temperature increase in the HTF decreases to the inlet temperature
the efficiency of the systems decreases. The 2-PCM system has a
higher discharging efficiency than the 3-PCM system due to the
larger amounts of latent energy released to the HTF over a greater
period of time during the solidification process.
3.3. Overall cycle efficiency
The performance of a TES system should not be judged on the
charging or discharging process alone but by looking at the overall
cycle efficiency. A prime example as to the importance of the cyclic
performance is the behavior of the NaNO2 system. While it has the
best performance during the charging process, it fails to release all
of its stored energy during discharging. This leads to it having the
slowest cycle efficiency of 88%. The converse is true for the KNO3
system where its strong performance during discharging makes up
for its poor performance during charging leading to the system
having the highest overall energy efficiency. The energy stored,
exergy stored, energy released, exergy released, charging and
discharging Ste numbers, overall charging and discharging
efficiencies, and the percent energy stored via latent heat are
listed in Table 3. The results show that the multi-PCM systems not
only store more energy than the KNO3 and NaNO3 single PCM
systems but they do so in a more efficient manner. Furthermore,
while it is possible to recover all of the energy stored in the sensible
heat system, it can only store 40% of the energy the NaNO3 system
stores. This highlights the improvement latent heat-based systems
 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26 25

Fig. 13. Exergy efficiency of the discharging process.

prolonged solidification time of the NaNO3 systems leads to the

Table 3
Results of the charging and discharging process for the NaNO3, NaNO2, KNO3, 2- system having the highest discharging efficiency. Furthermore, as
PCM, 3-PCM, and sensible heat systems. the purpose of TES systems is to produce work, the overall cyclic
performance of the systems must be analyzed. As the NaNO2
NaNO3 NaNO2 KNO3 2- 3- Sensible Heat
PCM PCM
system fails to release all of its stored energy it has the lowest cyclic
efficiency with only 88% of the stored energy being retrieved.
Energy Stored (MJ) 116 142 84.4 130 118 47.1
Exergy Stored (MJ) 55.9 66.4 42.0 61.4 56.5 22.8
Although complete solidification of the system occurs, the NaNO3
Stechar 0.30 0.79 0.013 0.55 0.37 – system only releases 92% of its stored energy. Additionally, a
Energy Released (MJ) 107 125 83.8 128 116 46.5 system should not be judged solely on efficiency but also on the
Exergy Released (MJ) 51.9 58.5 41.7 61.4 55.8 22.5 total energy and exergy stored. When this is considered, the 2-PCM
Stedis 0.029 0.038 0.73 0.034 0.27 –
not only stores the most energy but the HTF is also able to recover
hQ-overall (%) 92 88 99 98 98 98
hEx-overall (%) 93 88 99 100 99 99 this energy during discharging. This makes the 2-PCM system the
%LH 69 60 57 64 60 – best suited for the given operating conditions. The behavior of the
systems considered highlights the importance that the relation-
ship between the temperature of the HTF and the melting point of
offer over purely sensible heat TES systems and the added benefit the chosen PCMs have. Additionally, if a smaller operating
of employing a cascaded multi-PCM system. temperature range is employed the faction of energy stored via
latent heat would increase leading to an improvement in both the
4. Conclusions energetic and exergetic efficiencies of the system.
These results indicate that a cascaded multi-PCM has a better
The second law analysis of example EPCM-based TES systems energy and exergy performance over that of a single PCM system.
was conducted to investigate the benefit of using a cascaded multi- While these results show a 2-PCM system to be the optimal choice,
PCM system. Six systems were considered: three single PCM as the length of the system is increased the number of PCMs used
systems (NaNO3, NaNO2, and KNO3), a 2-PCM system (NaNO3 and will have a greater impact. Furthermore, care needs to be taken as
NaNO2), a 3-PCM system, and a sensible heat only system as a to the inlet and melting temperatures of the PCMs as they highly
comparison. In addition to their performance during the charging impact the performance of the system. Additionally, a constant
and discharging processes, the overall cyclic performance of the mass flow rate was assumed in this study, however this parameter
systems was also determined. As expected the latent heat-based has a large impact on the performance of the system. Thus a
systems outperform the sensible heat only systems and have a parametric investigation is required to determine the optimal
higher energy storage density. Furthermore, the thermal proper- operating conditions that would yield the highest overall system
ties of NaNO2 coupled with the chosen operating conditions lead to efficiency. While the results of this investigation lend insight into
the system having best performance during the charging process. key aspects to the performance of EPCM-based latent heat TES
With the exception of the NaNO2 system, the multi-PCM systems systems, additional research is required to determine the optimal
outperform the single PCM systems by storing more energy at a operating conditions.
higher rate and having a higher final exergy content with the 2-
PCM system slightly besting the 3-PCM system. References
The exceptional performance of the NaNO2 system during the
charging process is overshadowed but its dismal performance [1] A. Gil, M. Medrano, I. Martorell, A. Lazaro, P. Dolado, B. Zalba, L.F. Cabeza, State
during discharging. It is the only system that fails to solidify of the art on high temperature thermal energy storage for power generation.
Part 1—concepts, materials and modellization, Renew. Sustain. Energy Rev. 14
completely at the end of the 12-h discharging process. The (1) (2010) 31–55.
26 L. Solomon, A. Oztekin / Journal of Energy Storage 8 (2016) 12–26
[2] S. Kuravi, J. Trahan, D.Y. Goswami, M.M. Rahman, E.K. Stefanakos, Thermal
energy storage technologies and systems for concentrating solar power plants,
Prog. Energy Combust. Sci. 39 (4) (2013) 285–319.
[3] A. Bejan, Two thermodynamic optima in the design of sensible heat units for
energy storage, J. Heat Transfer 100 (4) (1978) 708–712.
[4] S. Jegadheeswaran, S. Pohekar, T. Kousksou, Exergy based performance
evaluation of latent heat thermal storage system: a review, Renew. Sustain.
Energy Rev. 14 (9) (2010) 2580–2595.
[5] M.A. Rosen, N. Pedinelli, I. Dincer, Energy and exergy analyses of cold thermal
storage systems, Int. J. Energy Res. 23 (12) (1999) 1029–1038.
[6] A.V. Ramayya, K. Ramesh, Exergy analysis of latent heat storage systems with
sensible heating and subcooling of PCM, Int. J. Energy Res. 22 (5) (1998) 411–
426.
[7] A. Sari, K. Kaygusuz, First and second laws analyses of a closed latent heat
thermal energy storage system, Chin. J. Chem. Eng. 12 (2) (2004) 290–293.
[8] H.H. Öztürk, Experimental evaluation of energy and exergy efficiency of a
seasonal latent heat storage system for greenhouse heating, Energy Convers.
Manage. 46 (9) (2005) 1523–1542.
[9] A. Koca, H.F. Oztop, T. Koyun, Y. Varol, Energy and exergy analysis of a latent
heat storage system with phase change material for a solar collector, Renew.
Energy 33 (4) (2008) 567–574.
[10] T. Kousksou, T.E. Rhafiki, A. Arid, E. Schall, Y. Zeraouli, Power, efficiency and
irreversibility of latent energy systems, J. Thermophys. Heat Transfer 22 (2)
(2008) 234–239.
[11] A. Erek, I. Dincer, A new approach to energy and exergy analyses of latent heat
storage unit, Heat Transfer Eng. 30 (6) (2009) 506–515.
[12] D. MacPhee, I. Dincer, Thermodynamic analysis of freezing and melting
processes in a bed of spherical PCM capsules, J. Solar Energy Eng. 131 (3) (2009)
031017.
[13] A. Sharma, V.V. Tyagi, C.R. Chen, D. Buddhi, Review on thermal energy storage
with phase change materials and applications, Renew. Sustain. Energy Rev. 13
(2) (2009) 318–345.
[14] M. Kenisarin, K. Mahkamov, Solar energy storage using phase change
materials, Renew. Sustain. Energy Rev. 11 (9) (2007) 1913–1965.
[15] M.M. Farid, A.M. Khudhair, S.A.K. Razack, S. Al-Hallaj, A review on phase
change energy storage: materials and applications, Energy Convers. Manage.
45 (9–10) (2004) 1597–1615.
[16] H. Paksoy, H. Mehling, L. Cabeza, Phase change materials and their basic
properties, Thermal Energy Storage for Sustainable Energy Consumption,
Springer, Netherlands, 2007 pp. 257–277.
[17] Y. Zheng, W. Zhao, J.C. Sabol, K. Tuzla, S. Neti, A. Oztekin, J.C. Chen,
Encapsulated phase change materials for energy storage—characterization by
calorimetry, Sol. Energy 87 (2013) 117–126.
[18] W. Zhao, Y. Zheng, J.C. Sabol, A. Oztekin, S. Neti, K. Tuzla, W.M. Misiolek, J.C.
Chen, Heat transfer analysis for thermal energy storage using NaNO3 as
encapsulated phase change material, ASME 2012 Heat Transfer Summer
Conference Collocated with the ASME 2012 Fluids Engineering Division
Summer Meeting and the ASME 2012 10th International Conference on
Nanochannels, Microchannels, and Minichannels (2012) 241–248.
[19] Y. Zheng, J.L. Barton, K. Tuzla, J.C. Chen, S. Neti, A. Oztekin, W.Z. Misiolek,
Experimental and computational study of thermal energy storage with
encapsulated NaNO3 for high temperature applications, Sol. Energy 115 (0)
(2015) 180–194.
[20] W. Zhao, A.F. Elmozughi, A. Oztekin, S. Neti, Heat transfer analysis of
encapsulated phase change material for thermal energy storage, Int. J. Heat
Mass Transfer 63 (0) (2013) 323–335.
[21] E. Assis, L. Katsman, G. Ziskind, R. Letan, Numerical and experimental study of
melting in a spherical shell, Int. J. Heat Mass Transfer 50 (9–10) (2007) 1790–
1804.
[22] A.R. Archibold, J. Gonzalez-Aguilar, M.M. Rahman, D. Yogi Goswami, M.
Romero, E.K. Stefanakos, The melting process of storage materials with
relatively high phase change temperatures in partially filled spherical shells,
Appl. Energy 116 (0) (2014) 243–252.
[23] A.F. Elmozughi, L. Solomon, A. Oztekin, S. Neti, Encapsulated phase change
material for high temperature thermal energy storage—heat transfer analysis,
Int. J. Heat Mass Transfer 78 (0) (2014) 1135–1144.
[24] L. Solomon, A.F. Elmozughi, S. Neti, A. Oztekin, High temperature thermal
energy storage using EPCM: the effect of void, Proc. ASME 2014 International
Mechanical Engineering Congress and Exposition, American Society of
Mechanical Engineers (2016).
[25] L. Solomon, A.F. Elmozughi, A. Oztekin, S. Neti, Effect of internal void
placement on the heat transfer performance–encapsulated phase change
material for energy storage, Renew. Energy 78 (2015) 438–447.
[26] M.H. Mahfuz, A. Kamyar, O. Afshar, M. Sarraf, M.R. Anisur, M.A. Kibria, R. Saidur,
I.H.S.C. Metselaar, Exergetic analysis of a solar thermal power system with
PCM storage, Energy Convers. Manage. 78 (2014) 486–492.
[27] M. Rezaei, M.R. Anisur, M.H. Mahfuz, M.A. Kibria, R. Saidur, I.H.S.C. Metselaar,
Performance and cost analysis of phase change materials with different
melting temperatures in heating systems, Energy 53 (2013) 173–178.
[28] H. El-Dessouky, F. Al-Juwayhel, Effectiveness of a thermal energy storage
system using phase-change materials, Energy Convers. Manage. 38 (6) (1997)
601–617.
[29] D. Singh, W. Zhao, W. Yu, D.M. France, T. Kim, Analysis of a graphite foam–NaCl
latent heat storage system for supercritical CO2 power cycles for concentrated
solar power, Sol. Energy 118 (2015) 232–242.
[30] H. Cui, X. Yuan, X. Hou, Thermal performance analysis for a heat receiver using
multiple phase change materials, Appl. Therm. Eng. 23 (18) (2003) 2353–2361.
[31] M. Fang, G. Chen, Effects of different multiple PCMs on the performance of a
latent thermal energy storage system, Appl. Therm. Eng. 27 (5) (2007) 994–
1000.
[32] R. Seeniraj, N.L. Narasimhan, Performance enhancement of a solar dynamic
LHTS module having both fins and multiple PCMs, Sol. Energy 82 (6) (2008)
535–542.
[33] H.A. Adine, H. El Qarnia, Numerical analysis of the thermal behaviour of a shell-
and-tube heat storage unit using phase change materials, Appl. Math. Modell.
33 (4) (2009) 2132–2144.
[34] G. Peiró, J. Gasia, L. Miró, L.F. Cabeza, Experimental evaluation at pilot plant
scale of multiple PCMs (cascaded) vs. single PCM configuration for thermal
energy storage, Renew. Energy 83 (2015) 729–736.
[35] T. Watanabe, A. Kanzawa, Second law optimization of a latent heat storage
system with PCMs having different melting points, Heat Recov. Syst. CHP 15 (7)
(1995) 641–653.
[36] Z.-X. Gong, A.S. Mujumdar, Thermodynamic optimization of the thermal
process in energy storage using multiple phase change materials, Appl. Therm.
Eng. 17 (11) (1997) 1067–1083.
[37] T. Kousksou, F. Strub, J.C. Lasvignottes, A. Jamil, J. Bedecarrats, Second law
analysis of latent thermal storage for solar system, Sol. Energy Mater. Sol. Cells
91 (14) (2007) 1275–1281.
[38] R. Doman  ski, G. Fellah, Exergy analysis for the evaluation of a thermal storage
system employing PCMs with different melting temperatures, Appl. Therm.
Eng. 16 (11) (1996) 907–919.
[39] Z.-X. Gong, A.S. Mujumdar, Finite element analysis of a multistage latent heat
thermal storage system, Numer. Heat Transfer Part A Appl. 30 (7) (1996) 669–684.
[40] Y.-Q. Li, Y.-L. He, Z.-F. Wang, C. Xu, W. Wang, Exergy analysis of two phase
change materials storage system for solar thermal power with finite-time
thermodynamics, Renew. Energy 39 (1) (2012) 447–454.
[41] H. Shabgard, C.W. Robak, T.L. Bergman, A. Faghri, Heat transfer and exergy
analysis of cascaded latent heat storage with gravity-assisted heat pipes for
concentrating solar power applications, Sol. Energy 86 (3) (2012) 816–830.
[42] A. Mosaffa, L.G. Farshi, C.I. Ferreira, M. Rosen, Energy and exergy evaluation of a
multiple-PCM thermal storage unit for free cooling applications, Renew.
Energy 68 (2014) 452–458.
[43] G.J. Janz, C.B. Allen, N.P. Bansal, R.M. Murphy, R.P.T. Tomkins, Physical
properties data compilations relevant to energy storage, 2. Molten salts: Data
on single and multi-component salt systems NASA STI/Recon Technical Report
N, 80 (1979) 10643.
[44] G.J. Janz, C.B. Allen, N. Bansal, R. Murphy, R. Tomkins, Physical Properties Data
Compilations Relevant to Energy Storage. II. Molten Salts: Data on Single and
Multi-component Salt Systems, Rensselaer Polytechnic Inst., Troy, NY (USA),
1979 (Cogswell Lab.).
[45] F.P. Incropera, D.P. DeWitt, Fundamentals of Heat and Mass Transfer, John
Wiley & Sons, Inc., New York, 2002.
[46] X. Zhang, M. Fujii, Simultaneous measurements of the thermal conductivity
and thermal diffusivity of molten salts with a transient short-hot-wire
method, Int. J. Thermophys. 21 (1) (2000) 71–84.
[47] A.D. Brent, V.R. Voller, K.J. Reid, Enthalpy-porosity technique for modeling
convection-diffusion phase change: application to the melting of a pure metal,
Numer. Heat Transfer 13 (3) (1988) 297–318.
[48] J.M. Khodadadi, Y. Zhang, Effects of buoyancy-driven convection on melting
within spherical containers, Int. J. Heat Mass Transfer 44 (8) (2001) 1605–1618.
[49] K.A.R. Ismail, Heat Transfer in Phase Change in Simple and
ComplexGeometries.Thermal Energy Storage Systems and Applications,
Wiley, Chichester, 2002.
[50] F.L. Tan, S.F. Hosseinizadeh, J.M. Khodadadi, L. Fan, Experimental and
computational study of constrained melting of phase change materials
(PCM) inside a spherical capsule, Int. J. Heat Mass Transfer 52 (15–16) (2009)
3464–3472.
[51] M. Pinelli, S. Piva, Solid/Liquid phase change in presence of natural convection:
a thermal energy storage case study, J. Energy Res. Technol. 125 (3) (2003) 190–
198.
[52] V.R. Voller, C. Prakash, A fixed grid numerical modeling methodology for
convection-diffusion mushy region phase-change problems, Int. J. Heat Mass
Transfer 30 (8) (1987) 1709–1718.
[53] 2010, ANSYS FLUENT Theory Guide Replease 13.0.
[54] F.R. Menter, Two-equation eddy-viscosity turbulence models for engineering
applications, AIAA J. 32 (8) (1994) 1598–1605.
[55] E. Assis, G. Ziskind, R. Letan, Numerical and experimental study of
solidification in a spherical shell, J. Heat Transfer 131 (2) (2008) 024502.