FINITE ELEMENT MODELING FOR INTEGRATED SOLID-SOLID PCM-BUILDING
MATERIAL WITH VARYING PHASE CHANGE TEMPERATURES
Dahai Zhang, Alan S. Fung and Omar Siddiqui
Dept. of Mechanical and Industrial Engineering, Ryerson University, Toronto, Canada
ABSTRACT
Phase change materials (PCMs) can be implemented
into building practice to enhance thermal storage
performance, reduce indoor temperature fluctuations
and improve the comfort of occupants. Solid-solid
PCMs (SSPCMs) have unique properties which make
them superior to the traditional solid-liquid PCMs
(SLPCMs). These include an enhanced structural
integrity and ease of incorporation with building
fabrics. In order to investigate new types of SSPCM, it
is desirable to have a finite element model (FEM) to be
validated by the experimental data. Due to the impurity
and thermal instability, practical PCMs usually do not
have the same phase change temperatures on heating
and cooling processes. This varying thermal property
will affect the validation of FEM, if it is not properly
represented. In current study, a mathematical model for
SSPCMs will be described and a corresponding FEM
will be created in COMSOL Multiphysics (COMSOL,
2007). The effective heat capacity method was used for
the simulation. The lumped model was found to show
good agreement with the discretized model. The
integrated PCM-building material was examined in the
lumped model based on the temperature and heat flux
profiles. The heat flux profile proves to be an effective
tool to understand the phase change process. The
simulation results indicate that the PCM with lower
phase change temperature can lengthen the energy
release at certain level, hence can improve the thermal
comfort in a building.
INTRODUCTION
Building energy consumption and its resulting
greenhouse gas emissions can be substantially reduced
if solar energy is adequately utilized. Storage of
thermal energy in buildings has gained prominence in
the past two decades due to a strong need to reduce the
total thermal energy requirement (both heating and
cooling) in buildings (Zalba et al., 2003). The
integration of phase change materials into building
fabrics can accumulate the gain from solar radiation
during the day and release the stored energy at night.
Therefore, the energy requirement at peak hours can be
reduced and pre-cooling or heating becomes possible.
Furthermore, the implementation of PCMs can reduce
the indoor temperature fluctuations and improve the
thermal comfort of occupants.
The most popular type of PCMs available currently is
of the solid-liquid type, which includes two major
groups, organic and inorganic compounds (Trp, 2005).
Organic compounds exhibit various advantages such as
being non corrosive, having low or no supercooling,
chemical and thermal stability, and compatibility with
conventional building materials. However, they
encounter several disadvantages like low phase change
enthalpy, low thermal conductivity and flammability.
Inorganic compounds present a greater phase change
enthalpy, usually twice the value of organic compounds,
and also possess high thermal conductivity. The main
disadvantages of inorganic compounds are corrosion,
supercooling and lack of thermal stability.
Despite their high energy storage density and other
advantages, most currently available PCMs suffer from
inefficient energy recovery and reduction in overall
thermal performance during the solid-liquid phase
transition (Farid et al., 2004). Considering the ever-
increasing application of PCMs in building practice,
PCMs in the liquid state affect the structural integrity
and cause undesirable “wetness” in the walls. The
aforementioned difficulties could be avoided if
SSPCMs are employed.
Practical PCMs also exhibit different phase change
temperatures on heating and cooling processes due to
their impurity and thermal instability. Paraffin wax, for
instance, melts and solidifies in the range of 30-40°C
and 50-60°C, respectively (Bentz and Turpin, 2007).
The varying phase change temperature on heating and
cooling will present different thermal performance in
buildings and will also affect the results of building
simulation.
In order to mathematically understand the heat transfer
phenomenon during solid-solid phase transition, a brief
review of the numerical methods of thermal
performance of the solid-liquid PCMs is helpful and
necessary. During the solid-liquid phase transition
process, heat transfer at the moving solid-liquid
interface becomes a transient, non-linear phenomenon,
which is often called the “moving boundary” problem
(Alexiades and Solomon, 1993). The existence of the
non-linear convective term in the liquid phase requires
the solution of partial differential equations, thus
making the moving boundary problems difficult to
solve (Myers, 1971).
The finite difference method or finite element method
are often used to solve phase change problems. The
most common numerical methods are the enthalpy
method and the effective heat capacity method. These
methods allow a wide range of phase transition
temperature of PCMs (Lamberg et al., 2004).
Lamberg et al. (2004) carried out finite element
analysis on the technical grade paraffin with FEMLAB,
and indicated that the effective heat capacity method,
when used with a narrow temperature range, gave the
closest prediction when compared with the
experimental data. However, the phase change
temperatures used for the simulation are the same for
heating and cooling processes, which was not
suggested from the differential scanning calorimetry
(DSC) curve. For the application of PCMs in building
practice, the characteristic of different phase change
temperatures on heating and cooling is especially
important since it determines at what degree the energy
is stored and released. The proper modeling on the
phase change temperature will help to understand the
effectiveness of PCMs on the energy performance in
buildings.
Zhang et al. (2007) studied the thermal performance of
a concrete compound integrated with a hypothetical
SSPCM of varying PCM concentration and thickness.
The finite element analysis on COMSOL indicated that
the overall thermal behavior of the compound is subject
to not only the thermal capacity of PCMs, but also the
thermal resistance of the compound. However, the
difference in phase change temperatures on the heating
and cooling processes were not taken into
consideration in the study.
OBJECTIVE
The objective of this study is to modify the previous
FEM (Zhang et al., 2007), and to analyze the impact of
different heating and cooling phase change
temperatures on the thermal performance. A script will
be written in the solver to achieve the automatic
alteration in thermal properties depending on the
heating and cooling mode. The new FEM will be used
for the proposed experimental validation.
NUMERICAL METHODS
The governing equations for the heat transfer process
in the solid-liquid PCMs are composed of the Navier-
Stokes (momentum) equations, the mass conservation
equation, and the energy conservation equation. From a
mathematical point of view, the momentum and mass
conservation equations can be neglected for the solid-
solid PCMs because there is no convective term in the
PCMs. This conclusion significantly simplifies the
numerical analysis. Therefore, the energy equation is
the only governing equation for the analysis of
SSPCMs, and can be expressed as follows,
⎛ ∂T ⎞
ρC p ⎜ ⎟ = ∇(k∇T ) (1)
⎝ ∂t ⎠
where ρ, Cp and k are the density, specific heat and
thermal conductivity of PCMs, respectively.
Upon heating or cooling, sensible and latent heat pass
through the PCM simultaneously and the phase
transition does not occur uniformly in the PCM.
Furthermore, the phase transition usually happens in a
non-isothermal temperature range, which means that
the transition range covers several degree instead of a
sudden change. In case of solid-liquid PCMs, it is
nearly impossible to track the solid-liquid interface,
thus making the heat transfer calculation difficult.
Therefore, it is desirable to solve the PCM problem in a
whole constrained domain (Lamberg et al., 2004).
There are two types of numerical methods often used to
solve the PCM problems, namely enthalpy method and
effective heat capacity method.
The enthalpy method enables the governing equations
to be applied over the entire domain of interest rather
than on the phase change interface. The total energy
required for the phase change, i.e. the overall sensible
heat and latent heat, can be determined by the enthalpy
function H(T). For the SSPCM, the enthalpy term for
the energy equation (Eq.1) can be expressed as follows
∂H (T )
ρ = ∇(k∇T ) (2)
∂t
An alternate way of solving the non-isothermal phase
transition in PCM is the effective heat capacity method.
The effective heat capacity of the material (Ceff) is a
linear function of the latent heat of fusion on both the
heating and cooling processes. It is inversely
proportional to the temperature difference between the
onset and the end of the phase transition (Peippo et al.,
1991). The effective heat capacity of the PCM during µV/Btu/Ft2Hr, and it comes with one K-type
the phase change is given as, thermocouple. The composite is insulated on the sides.
The bottom maintains constant temperature through a
L
Ceff = + Cp (3) constant temperature bath. During the heating mode, a
Te − To flexible heater will provide constant heat output into
the composite. The back of the heater is insulated to
where L is the latent heat of fusion, To is the onset minimize the heat loss. During the cooling mode, two
temperature of phase transition happens and Te is the constant temperature baths will be attached to the top
temperature when the phase transition completely and bottom of the composite. The heat flux across both
finishes. Thus, the governing equation (Eq.1) for the surfaces and the temperatures on the surfaces as well as
binary solid state PCM can be expressed as in the composite will be monitored during the whole
test. These collected data will be used to validate the
⎛ ∂T ⎞
ρC p (T )⎜ ⎟ = ∇(k∇T ) (4) finite element model.
⎝ ∂t ⎠
SIMULATION
⎧C p T < To The simulations were carried out in COMSOL
⎪ Multiphysics. It is assumed that the heat transfer occurs
⎪ L (5)
Cp = ⎨ + C p @ To To ≤ T ≤ Te only in the vertical direction, so that a 2D model is
⎪Te − To sufficient for the analysis. The first step of the work
⎪C p T > Te
⎩ was to examine that the model could represent the
phase change process, i.e., the temperature vs. time
EXPERIMENTAL SETUP profile should clearly show the distinctive slopes
within the phase change temperature range.
In this section, a brief description of experimental setup
will be introduced and the proposed experimental For the integrated PCM-building materials, the
procedure will be discussed for the study of integrated assumption of homogenous mixture is crucial for the
PCM-building materials. simulation of thermal performance. Because the PCM
and the building material have different thermal
properties, the interaction between the two might affect
the overall thermal behaviour. Thus, the ideal model is
to simulate the two discretized domains. However, the
discretization of different domains will increase the
Insulation

Insulation
Insulation

Insulation

calculation time significantly and hence is not a desired

solution for a long time simulation. An alternative
solution is to treat the mixture as a homogenous
lumped system, with the thermal properties as the
averaged values of the two domains. In order to
validate this assumption, the temperature and heat flux
Figure 1 Experimental setup across the composite were examined.
Table 1
The experimental setup for the heating and cooling
Thermo-physical properties of the PCM
modes is sketched in Figure 1. The composites are
composed of drywall, concrete or gypcrete with certain Property Value
weight ratios of PCM. The composite will be cast in a Density (ρ) kg·m-3 800
plastic mould with the interior dimension of 10×10×4
Thermal conductivity (k) W·m-1K-1 0.3
cm. Three T-type fine-gage thermocouples will be
embedded into the composite. Two pieces of aluminum Specific heat (Cp) J·g-1K-1 1.4
plates will be attached to the top and bottom of the Phase change temperature °C heating
22/23
composite. The aluminum plates are used to achieve Case 1 cooling
uniform temperature distribution across the surface. Phase change temperature °C heating 22/23
Two HFS-3 heat flux sensors with thermocouples made Case 2 cooling 19/20
from Omega will be attached between the plate and the
composite. The NI cDAQ-9172 data acquisition system Latent heat of fusion (L), J·kg-1, 100000
will be used to connect all the sensors (National
Instruments, 2007). The sensor has a resolution of 65
The thermo-physical properties of the hypothetical
PCM are presented in Table 1. The 2D discretized
model of the composite is 200 mm in length and 20
mm in thickness. The PCM is assumed to be circles
with a radius of 1mm (note: it would be cylindrical in
3D model). The array of PCM is added in the
composite with a distance of 4 mm between two
centers, as shown in Figure 2. The rest of the composite
is the building material. The lumped model is assumed
to have homogenous properties throughout the whole
domain. Table 2 lists the parameters used in the
simulation for the discretized model and the lumped
model, respectively. The values were calculated based
on the weight ratio of PCM in the composite.
Figure 2 Partial geometry of the discretized model
insulated. In order to investigate the effect of varying
heating and cooling phase change temperature on the
thermal performance, two scenarios were applied over
the domain. The first scenario assumes the phase
transition occurs in the same temperature range
between 22 and 23°C on heating and cooling. The
second scenario assumes a phase change temperature
between 22 and 23°C on heating, whereas between 19
and 20°C on cooling. The latent heat remains the same
as before, at 100 J/g on heating and cooling for both
cases.
RESULTS AND ANALYSIS
A spherical domain with 2mm in diameter was
modeled to exhibit the phase change process. The
effective heat capacity method was used for the
simulation. A constant heat flux with the value of 20
W/m2 was applied over the boundary. Figure 3 shows
the temperature variation in the heating mode in the
homogenous domain. The temperature profile
represents any point in the domain. As can be seen, the
curve departs distinctly at the phase change
temperatures between 22 and 23°C, respectively. The
smaller slope in between indicates the energy storage
during the phase transition.

26
Table 2 25

Parameters in the discretized and lumped models

24

23
Temperature, °C

Discretized Model 22

21
Parameter PCM BM 20

ρ (kg·m-3) 800 2400 19

k (W·m-1K-1) 0.3 0.4 18

17
Cp (J·g-1K-1) 1.4 1.4 16

vol% 19.6 80.4 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Time, sec
wt% 7.5 92.5
Lumped Model Figure 3 Temperature profile in the heating mode in
ρ (kg·m-3) 2084 the homogenous domain
k (W·m-1K-1) 0.39
L (J·kg-1) 7500 With the method mentioned above, the temperature and
the heat flux in the heating mode were compared
between the discretized model and the lumped model.
In the heating mode, a constant heat flux of 200 W/m2 Figure 4 shows the heat flux distribution at the 2000
is applied over the upper surface, with all the other second in the discretized model. PD_1 and PD_2 refer
boundaries insulated. The initial temperature and the to the positions in the center of PCM sphere and the
ambient temperature are assumed to be 19°C. middle point between two PCM spheres. Both points
Convective heat transfer on the upper surface is zero. are 2 mm in the depth of the composite. As can be seen,
In the cooling mode, the initial temperature was set to the heat flux is not uniformly distributed across the
be 26°C. The upper surface is subject to a constant horizontal surface. Heat flux has a lower value in the
temperature at 16°C, with all the other boundaries PCM compared to the surrounding building material.
This can be attributed to the high latent heat capacity of
the PCM as well as its low thermal conductivity. Thus, process than the discretized model, with the maximum
the thermal diffusivity is lower in the PCM (Kuznik et error of 12%.
al., 2008).
The temperature profiles in both models show very
good agreement, as shown in Figure 6. The maximum
error in temperature is 1%. It should be noted that the
PD_1 PD_2 temperature does not vary much between the two
domains in the discretized model, although the heat
flux value varies significantly. Figure 6 does not
exhibit distinct temperature transition as shown in
Figure 3, which can be attributed to the size of the
sample. Figure 3 indicates an intensive heat transfer
process (even with 20 W/m2) over a comparable small
domain. For a large domain (as in Figure 6) however,
the heat transfers slowly through the material even with
a much larger heat flux (200 W/m2), resulting in a
smooth temperature transition.

26 PD_1 PD_2 PL_1

Figure 4 Heat flux distribution at the 2000 second in 25
the discretized model
24
Temperature (°C)
23

In order to use the lumped model to replace the 22

discretized model, it is necessary to compare the 21

thermal performance of two models. Figure 5 shows 20

the heat flux distribution over 2000 seconds in both 19

models, where PD_ave refers to the averaged heat flux 18

0 200 400 600 800 1000 1200 1400 1600 1800 2000
on PD_1 and PD_2. PL_1 refers to the same position of Time (sec)
PD_2 in the lumped model, i.e. 2mm in depth into the
composite.
Figure 6 Temperature profiles over 2000 seconds in
the discretized model and lumped model
220 PD_1 PD_2 PD_ave PL_1
200

180
According to the above analysis, it is reasonable to
160
assume the integrated PCM-building material
Heat flux (W/m2)

140

120 composite as a lumped system, with the thermal

100 properties as the averaged values of different materials.
80
Other assumptions made in the simulation are given as
60
40
follows,
20

0
1. The density and the thermal conductivity of the
0 200 400 600 800 1000 1200 1400 1600 1800 2000
composite are constant.
Time (sec)
2. The heat transfer across the compound is
considered as one-dimensional.
Figure 5 Heat flux profiles over 2000 seconds in the
discretized model and lumped model
The lumped model was adopted to investigate the
As can be seen, heat flux in the building material thermal performance of the composite. The thermal-
shows a higher value than that in the PCM in the same physical properties in the following simulation are
depth. Both curves deviate from the heat flux profile in listed in Table 3.
the lumped model. However, the profile of averaged
heat flux shows reasonable agreement with the lumped
model, especially before and after the phase change.
The lumped model exhibits an earlier phase change
 Table 3 27 P1 P2 P3 P4 P5
26
Thermo-physical properties of the composite 25
24

wt% of PCM

Temperature (°C)
20 23
22
vol% of PCM 42.86 21

ρ (kg·m-3) 1714 20
19

k (W·m-1K-1) 0.38 18
17
L (J·kg-1) 20000 16
15
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (sec)
Figures 7 and 8 show the temperature profile and the
heat flux profile in the heating mode at the different
Figure 9 Temperature profile in cooling modes, case 1
depths of the composite. P1-P5 refer to 2 mm, 6 mm,
10 mm, 14 mm, 18 mm in depth, respectively. As can
be seen from the temperature profile, the phase change
range can be hardly identified, especially at the regions 1000
P1 P2 P3 P4 P5
that are close to the heat source. This is possibly due to 900

the fast thermal response to the comparably intensive 800

700
heat gain at these regions. At a lower region where the

Heat flux (W/m2)

600
heat slowly transfers to, e.g. at P3, a smaller slope 500
between the phase change temperatures (22 and 23°C) 400
can be identified. 300

200
29 P1 P2 P3 P4 P5 100
28 0
27 0 1000 2000 3000 4000 5000 6000 7000 8000
Time (sec)
26
Temperature (°C)

25

24 Figure 10 Heat flux profile in cooling modes, case 1

23

22 The heat flux profiles present a better indication of

21 phase change. As shown in Figure 8, the heat flux
20 shows a wave trend, which initially increases quickly,
19
0 500 1000 1500 2000 2500 3000 3500 4000
followed by the decrease in heat flux. At the valley of
Time (sec) the wave, the heat flux starts to increase again until it
maintains at a constant level. The wave shape can help
Figure 7 Temperature profile in the heating mode to understand the alternative dominate function of
sensible heat and latent heat. The increase in heat flux
indicates accumulation of sensible heat, which is
dominantly used to increase the body temperature. The
220 P1 P2 P3 P4 P5
200
first peak indicates that the sensible heat shifts to the
180 dominant effect of transferring heat to the lower
160 position. The temperature continues to increases,
Heat flux (W/m2)

140
however, the majority of the sensible heat passes the
120

100
point and is dedicated to increase the temperature at the
80 lower position. This process exhibits the decrease in
60 heat flux. The latent heat becomes dominant at the
40
valley of the heat flux wave. By comparing the time
20

0
when the phase change temperatures are reached, it is
0 500 1000 1500 2000 2500 3000 3500 4000 reasonable to believe that the valley is the start of phase
Time (sec)
change, i.e. the initial of dominance of the latent heat.
The heat flux increases as the energy is stored, until it
Figure 8 Heat flux profile in the heating mode starts to level off to remain constant, which can be
referred to the end of phase change in the temperature
profile. Thus, the alternative dominant function of
sensible heat and latent heat can be identified from the 800 P1 P2 P3 P4 P5
heat flux profile. The delay of occurrence of phase 700

change at lower positions also indicates that the great 600

latent heat serves to “block” the travel of heat in depth.

Heat flux (W/m2)

500

Figures 9 and 10 show the temperature profile and the 400

heat flux profile in the cooling mode for case 1. As can 300

be seen in Figure 9, the composite is completely cooled 200

down to the temperature at the upper surface after 8000 100

seconds. The position closer to the cooling source 0

shows a faster decrease in temperature, whereas the 0 1000 2000 3000 4000 5000 6000 7000 8000
Time (sec)
temperature decreases slowly at lower positions. The
phase change process can be clearly identified in the
lower region; however it is not obvious at upper region. Figure 12 Heat flux profile in cooling modes, case 2

According to the heat flux analysis explained above,
the phase change range can be identified easily from
the heat flux profile. The initial increase in heat flux
indicates the quick temperature drop due to the removal
of the sensible heat. The followed decrease in heat flux
can be attributed to the heat transfer from the lower
position. The onset of the valley is an indication of the
switch to the dominant function of latent heat, from
where the energy is released to compensate the
decrease in temperature. The second peak of the heat
flux wave indicates the end of phase change. The heat
flux shifts to decrease as the sensible heat decreases,
corresponding to the reduction in temperature at that
position.
27 P1 P2
26
25
24
Temperature (°C)
Figures 11 and 12 show the temperature profile and the
heat flux profile in the cooling mode for case 2. The
phase change temperatures correspond to 19 and 20°C.
In order to investigate the effect of varying phase
change temperatures on the thermal performance of the
composite, the temperature and heat flux profiles were
compared between two cases.
As shown in Figure 13, the rate of temperature change
in case 1 is more or less the same during the entire time.
The temperature decreases faster in case 2 than in case
1 before the phase change occurs. The rate of
temperature change in case 2 decreases after the phase
change happens. The temperate in case 2 becomes
greater than that in case 1 after 3700 seconds, which
indicates that the lower phase change temperature will
delay the temperature decrease. In terms of building
P3 P4 P5 comfort, the PCM with lower phase change
temperature could give a longer duration of thermal
comfort.

23
22
21
20 27 Case 1 Case 2
19 26
18 25
17 24
Temperature (°C)

16
23
15 22
0 1000 2000 3000 4000 5000 6000 7000 8000
21
Time (sec)
20
19

Figure 11 Temperature profile in cooling mode, case 2 18

17
16
15
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (sec)

Figure 13 Temperature profile at P3 in both cases

 1000 Case 1
900
800
700
Heat flux (W/m2)
600
500
400
300
200
100
0
0 1000 2000 3000 4000
Time (sec)
Figure 14 Heat flux profile at P1 in both cases
Figure 14 shows the heat flux profile at P1 in both
cases. As can be seen, the heat flux in case 1 is greater
than that in case 2 until 3700 second. The heat flux in
case 1 quickly decays afterwards, whereas the heat flux
in case 2 decays slowly at around 100 W/m2 until 7000
second. The heat flux comparison also indicates that
the lower phase change temperature can present a
longer release of latent heat, hence improve the thermal
comfort in a room.
The PCM may have different impact on the buildings
with various phase change temperatures. In the region
where cold weather dominates, such as in Canada,
PCM with lower cooling phase change temperature
might improve the thermal comfort more efficiently by
maintaining the room temperature at certain level for a
longer period of time. However, this hypothesis should
be examined by some building simulation tools.
CONCLUSIONS
A finite element model was developed for the
integrated SSPCM-building materials with varying
heating and cooling phase change temperatures. The
lumped model was found to show good agreement with
the discretized model. The FEM was examined based
on the temperature and heat flux profiles in the lumped
model. The heat flux profile clearly demonstrates the
alternative domination of sensible heat and latent heat
during the heating and cooling processes, which proves
to be an effective tool to understand the phase change
process. The simulation results indicate that the PCM
with lower phase change temperature can lengthen the
energy release at certain level, hence can improve the
thermal comfort in a building.
ACKNOWLEDGEMENTS
The authors would like to thank the NSERC Solar
Buildings Research Network for the financial support
of the project.
Case 2 REFERENCES
Alexiades, V. and Solomon, A.D. (1993),
Mathematical Modelling of Melting and Freezing
Processes, Hemisphere, Washington.
Bentz, D.P. and Turpin, R. (2007), ‘Potential
Applications of Phase Change Materials in Concrete
Technology’, Cement and Concrete Composites, 29
527-532.
5000 6000 7000 8000
COMSOL AB. COMSOL Multiphysics (2007),
Version 3.3, Stockholm. http://www.comsol.com
Farid, M.M., Khudhair, A.M., Razack, S.A.K and Al-
Hallaj, S. (2004), ‘A Review on Phase Change
Energy Storage: Materials and Applications’,
Energy Conversion and Management, 45 1597-
1615.
Kuznik, F., Virgone, J. and Noel, J. (2008),
‘Optimization of a Phase Change Material
Wallboard for Building Use’, Applied Thermal
Engineering, 28 1291-1298.
Lamberg, P., Lehtiniemi, R, and Henell, A.M. (2004),
‘Numerical and Experimental Investigation of
Melting and Freezing Processes in Phase Change
Material Storage’, International Journal of Thermal
Sciences, 43 277-287.
Myers, G.E. (1971), Analytical Methods in Conduction
Heat Transfer, McGraw-Hill, New York.
National Instruments (2007), National Instruments
Corporation, USA. http://www.ni.com
Peippo, K., Kauranen, P. and Lund, P.D. (1991), ‘A
Multicomponent PCM Wall Optimized for Passive
Solar Heating’, Energy and Buildings, 17 259-270.
Trp, A. (2005), ‘An Experimental and Numerical
Investigation of Heat Transfer during Technical
Grade Paraffin Melting and Solidification in a
Shell-and-Tube Latent Thermal Energy Storage
Unit’, Solar Energy, 79 648-660.
Zalba, B., Marín, J.M., Cabeza, L.F. and Mehling, H.
(2003), ‘Review on Thermal Energy Storage with
Phase Change: Materials, Heat Transfer Analysis
and Applications’, Applied Thermal Engineering,
23 251-283.
Zhang, D., Fung, A.S. and Siddiqui, O. (2007),
‘Numerical Studies of Integrated Concrete with a
Solid-Solid Phase Change Material’, 2nd Canadian
Solar Buildings Conference, Calgary, Canada.