Professional Documents
Culture Documents
net/publication/308092864
CITATIONS READS
50 1,338
4 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Ali Mohebbi on 19 October 2018.
spherical capsule
1
Abstract
melting process of the PCM, its melting fraction was analyzed at different
times. The results indicated the existence of thermally stable layers on the top of
points along the vertical axis indicates the existence of some disturbances at the
bottom of the sphere due to the natural convection. After the validation of the
results, the effects of different parameters such as the surface temperature of the
capsule, the initial temperature and the size of the spherical capsules, on the
melting process were investigated. The initial temperature did not affect the
temperature of the capsule and also decreasing the diameter of the sphere. The
parameters and other operational conditions can have a greater influence on the
2
1. Introduction
Recently, latent heat thermal energy storage systems have been widely
applied in cool storage for central air-conditioning, hot storage for air heating,
solar energy, energy efficient buildings and waste heat recovery. Storage
systems allow for the efficient and rational utilization of the available resources
or renewable energies, by using the time lag between production, as well as the
availability of the energy and its consumption in the demanding systems. Phase
change materials (PCM) are mainly used to provide higher storage densities. As
transfer rate is very appealing. This can be done by packing a volume with a
great number of PCM capsules. The spherical geometry is one interesting case
for heat storage applications, since spheres are much employed in packed beds.
Due to the complexity of such systems, it is often more effective to first model
Roy and Sengupta [1] examined the melting process with the solid phase
gradient in the solid core, they modified the heat transfer equation. At each time
step, the unsteady conduction equation has been solved numerically using a
3
boundary to transform the infinite domain into a finite one. In another paper [2],
order to reduce the computational effort and the time, they made some
simplifications. They obtained the non-dimensional melting time and the heat
energy storage) and release for the phase change of paraffin wax in the spherical
shells. They showed that an increment in the Nusselt number of the sphere with
the PCM inside the sphere. The natural convection role is enhanced upon an
increase in the sphere diameter and the air temperature. On the other hand,
during solidification the wax layers are formed inward, nucleating on the sphere
walls. As solidification progresses, the melt volume becomes smaller and the
spherical containers. They reported that during the initial melting process, the
convection is strengthened due to the increase in the melting zone, the process
at top region of the sphere is much faster than at the bottom due to the
increment of the conduction mode of heat transfer. They found that buoyancy-
4
driven convection speeds up the melting process compared to the diffusion-
controlled melting.
Veerappan et al. [6] investigated the phase change behavior of 65 mol% capric
acid and 35 mol% lauric acid, calcium chloride hexa hydrate, n-octadecane, n-
heat storage material. They created analytical models for solidification and
heat generation and found a good agreement between the analytical predictions
and the experimental data. Both models were validated with the experimental
the melting process. The model results were in a good agreement with the
experimental data.
Tan [9] investigated the melting of the phase change material (PCM)
5
melting processes. For constrained melting, paraffin wax (n-octadecane) was
temperature data that were gathered along the vertical diameter of the sphere
during the melting process. Tan and Khodadadi [10] also experimentally and
was performed. The effects of different parameters like the capsule size, the
surface temperature, the initial temperature and the Stefan number, on the PCM
2. CFD simulation
CFD simulation was carried out using the commercial Fluent 6.3
software. The physical model and computational procedure are discussed below.
6
diagram of the physical model is shown in Fig. 1. At t (time) > 0, a constant
temperature (T0) greater than the melting temperature is imposed on the surface
of the sphere, i.e. T0>Tm. Melting starts at the surface, moving the solid-liquid
interface towards the middle of the field. Since the present work is focused on
the analysis of constrained melting, both the solid and liquid phases have the
same density.
7
Table 1. Thermophysical properties of n-octadecane.
coefficient
For the experimental data that we used [9], the initial temperature was
1°C below the melting temperature and the surface temperature was 40°C. The
Prandtl, Stefan and Rayleigh numbers for this problem were 59.6, 0.113, and
2.63×108, respectively.
melting process that take place inside the sphere. The flow was considered
constrained melting, it is supposed that both the liquid and the solid phases are
8
the phase-change region inside the PCM, enthalpy-porosity approach [11-12]
was used. In this technique, the melt interface is not tracked explicitly.
Alternatively, a quantity called the liquid fraction, indicating the fraction of the
cell volume that is in liquid form, is associated with each cell in the domain. In
this way, the governing equations for the PCM are as follows:
Continuity:
𝜕𝜌 (1)
+ ∇. (𝜌𝑉) = 0
𝜕𝑡
Momentum:
𝜕(𝜌𝑉) (2)
+ ∇. (𝜌𝑉) = −∇𝑃 + 𝜇∇2 𝑉 + 𝜌𝑔 + 𝑆
𝜕𝑡
Energy:
𝜕 (3)
(𝜌𝐻) + ∇. (𝜌𝑉𝐻) = ∇. (𝑘∇𝑇)
𝜕𝑡
where V is the fluid velocity vector, ρ is the density, μ is the dynamic viscosity,
and H is the specific enthalpy, which is defined as the sum of the sensible
𝐻 = ℎ + ∆𝐻 (4)
where
9
𝑇 (5)
ℎ = ℎ𝑟𝑒𝑓 + ∫ 𝐶𝑝 𝑑𝑇
𝑇𝑟𝑒𝑓
The content of the latent heat can vary between zero (for a solid) and L (for a
liquid):
∆𝐻 = 𝛽𝐿 (6)
at which the material is at solid state and minimum temperature at which the
solidus and liquidus temperatures are the same. In order to determine the liquid
fraction at mushy zone, according to the method was given by Voller and
(Tm) was considered and assumed that liquefaction occurs at a certain range of
10
Fig. 2. The relation between enthalpy and temperature of a pure substance.
The source term S in the momentum equation is the Darcy's law damping
term that is added to the momentum equation due to the phase change effect on
amplitude of the damping. The higher this value, the steeper is the transition of
the velocity of the material to zero as it solidifies. Values between 10 4 and 107
are the recommended values for most calculations [12] and values exceeding
this range may cause the solution to oscillate. In the present research, different
values of the constant C were investigated to examine the effect of the mushy
zone constant.
This scheme uses the discrete continuity balance for a “staggered” control
volume about the face to compute the “staggered” pressure [13]. For flows with
natural convection, rotating flows, and flows in strongly curved domain, using
energy) were discretized using power law scheme, which interpolates the value
energy, pressure correction and liquid fraction were 0.1, 0.3, 1 and 0.9,
respectively.
Different grid densities were tried to make sure that the solution is independent
from the adopted grid density (see Fig. 3). As a result, 7200 grids were found
sufficient for the present study. Adopting a fine grid distribution in the radial
direction allows for the use of longer time steps. The time required for
comparing the selected quantities obtained from simulations using 0.05, 0.1,
0. 5 and 1 s, the time step for integrating the time derivatives was set to 1 s.
12
1
0/9
0/8
0/7
Liquid fraction
0/6
0/5
0/4
0/3 12800 node
0
0 20 40 60 80 100 120 140
Time(min)
The effect of different values of the mushy zone constant on the melt
fraction with a Stefan number of 0.132 and a shell diameter of 101.66 mm, is
shown in Fig. 4. As one can see from this figure, the value of 107 is suitable for
the present simulation where the numerical result is in most accordance with the
experimental data. Fig. 4 also confirms the validation of the result of CFD work.
13
Fig. 4. Variations in the liquid fraction vs. time at different mushy zone
initial temperature of 27.2 °C are shown in Fig.5. The blue and red colors
represent the solid and melted parts, respectively. In the beginning of the
process, the outer surface of the solid PCM is in contact with the inner wall of
the sphere so that heat conduction dominates between the solid PCM and the
wall. This makes the formation of a thin layer of liquid between the solid PCM
and the wall of the sphere. As time progresses, the molten zone expands and the
liquid layer grows. The liquid film is then forced up on the interior surface of
14
the sphere, occupying the region above the solid PCM. By moving towards the
top part of the sphere, the warm liquid is replaced with the cold liquid.
Therefore, natural convection in combination with a hot rising curved wall jet
the upper zone compared to the side and bottom sections is observed. This
zone of the sphere. After 40, 60 and 80 minutes, a waviness at the bottom part
of the PCM is observed. This can be attributed to the existence of stable and
unstable thermal layers at the bottom part of the sphere. This result as well as
the observed shape of the solid, are in good agreement with the experimental
15
0 min 20 min 40 min
120 min
The temperature contours of the right side of the sphere as well as the
formation of recirculating vortexes between the solid PCM and the inner wall of
the sphere as a result of the replacement of the warm liquid with cold liquid at
the bottom part of the sphere is obvious. It can also be seen that more melting
16
occurs at the upper part of the sphere due to the contribution of natural
convection (see Fig. 5). The whole PCM is melted after 120 minutes.
20 min 40 min
17
60 min 80 min
simulation for three points A, B, and C along the vertical diameter of the sphere
18
42 40
40
38
38
36
Temperature (°c)
Temperature(°C)
36
34
34
32
32
30 30
28 28
present work present work
26 26
Experimental data[9] Experimental data[9]
24 24
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
Time (min) Time(min)
(a) (b)
40
38
36
Temperature(°C)
34
32
30
28
present work
26
Experimental data[9]
24
0 20 40 60 80 100 120 140
Time(min)
(c)
convection motion in the unstable liquid layer at the bottom of the sphere. The
computed temperatures at point C (located on the upper side of the sphere) are
19
in a good agreement with the experimental data. This is due to the stable nature
To prove the existence of liquid layer where the point A is located, the
variation of liquid fraction at the point A during melting time was calculated
(see Fig. 8). The results revealed that the PCM was at a transition state from
solid to liquid during 7.5 to 9.5 minutes and after 9.5 minutes it was completely
1
Liquid fraction at point A
0/8
0/6
0/4
0/2
0
7 7/5 8 8/5 9 9/5 10
Time (min)
20
process. The melting rates for initial temperatures of 18.2 °C and 8.2 °C are the
same but the initial temperature of 27.2 °C has a few faster melting rate.
0/8
0/6
Liquid fraction
0/4
initial temperature at 8.2 °C
temperatures.
Stefan numbers. The Stefan number (Ste) is the ratio of sensible to latent heat of
the PCM. The surface temperature (TS) directly affects the value of the Stefan
number:
𝐶𝑝𝑙 (𝑇𝑆 − 𝑇𝑚 )
𝑆𝑡𝑒 = (9)
𝐿
here Cpl is the specific heat of the liquid PCM and L and Tm are the latent heat
and melting temperature of the PCM, respectively. The present simulation was
21
performed at three different Stefan numbers, namely Ste= 0.065 (Ts= 35 °C),
Ste= 0.113 (Ts= 40 °C) and Ste= 0.161 (Ts=45°C). By increasing the surface
temperature, the Stefan number will be increased whereas the time required for
0/8
Liquid fraction
0/6
0/4
ste=0.065(Ts=35°C)
0/2 ste=0.113(Ts=40°C)
ste=0.161(Ts=45°C)
0
0 30 60 90 120 150 180 210 240
Time(min)
temperatures.
The effects of different shell diameters and Stefan numbers on the changes in
the PCM liquid fraction with time are shown in Fig.11.The slopes of these
graphs represent the melting rates. The highest melting rates belong to the
beginning of the melting process because of the heat conduction between the
inner wall of the shell and the solid PCM. As expected, for each shell diameter
the liquid fraction increases more rapidly when the Stefan number (or surface
temperature, Ts) is larger. Also, for each Stefan number, full melting is
accomplished more rapidly in a shell with a smaller diameter. However, the
effect of these parameters (i.e. diameter and surface temperature or Stefan
number) are not the same on the melting time and rate. With respect to the data
of Fig. 11, it can be extracted that, 28% increase in surface temperature (from
22
Ts = 35o C to Ts = 45o C) results in about 170% to 198% decrease in melting
time for each constant diameter. While 27% increase in diameter of container
(from D = 80 mm to D = 101.66 mm) results in about 80% increase in melting
time. Therefore, the effect of surface temperature on melting time as well as
melting rate is more than the effect of container’s diameter on these parameters.
23
1 1
0/8 0/8
Liquid fraction
Liquid fraction
0/6 0/6
0/4 0/4
Ste=0.065 Ste=0.065
0/2 Ste=0.113 0/2 Ste=0.113
Ste=0.161 Ste=0.161
0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 160 180 200
Time(min) Time(min)
0/8
Liquid fraction
0/6
0/4
Ste=0.065
0/2 Ste=0.113
Ste=0.161
0
0 30 60 90 120 150 180 210 240
Time(min)
(c) D=101.66 mm
Fig.11. Changes in the liquid fraction versus time at different Stefan numbers
and different shell diameters: (a) D=60 mm (b) D=80 mm (c) D=101. 66 mm.
Fig.12 indicates variations in the heat flux versus time at different shell
diameters and Stefan numbers. In all plots, a sharp drop in the heat flux is
observed at the beginning of the melting process but after a few minutes, it
decreases smoothly. When the whole PCM is melted, the heat flux almost
reaches zero. The results show that by increasing the Stefan number, the heat
24
7000 8000
6000 7000
5000 6000
Heat flux(Wm-2)
Heat flux(Wm-2)
4000 5000
4000
3000
3000
2000 ste=0.065 ste=0.065
ste=0.113 2000
ste=0.113
1000 ste=0.161 1000 ste=0.161
0
0
0 50 100 150
0 50 100 150 200
Time(min) Time(min)
(a) D=60 mm (b) D=80 mm
7000
6000
5000
Heat flux(Wm-2)
4000
3000
ste=0.065
2000
ste=0.113
1000 ste=0.161
0
0 50 100 150 200 250
Time(min)
(c) D=101.66 mm
Fig.12. Variations in the heat flux at different Stefan numbers and shell
diameters: (a) 60 mm (b) 80 mm (c) 101.66 mm.
transfer rate increases. Integration of the area under the heat flux graphs in
Fig.12 gives the total amount of heat collection. As expected, by increasing the
diameter of the sphere and due to the increase in the amount of the PCM, the
25
4. Conclusions
and various diameters of the container was successfully done. Detailed phase
and flow fields were provided inside the system by simulations and the results
were discussed and compared with the experimental data. At the beginning, the
heat conduction to the PCM is dominant but after formation of a thin layer of
melted liquid between the inner wall and the solid PCM, natural convection
becomes dominant at the top half of the sphere. The comparison between the
calculated and measured temperatures at the top half of the sphere clearly shows
the stable nature of the molten liquid layer in that zone. The unstable liquid
layer at the bottom of the sphere is responsible for the waviness of the PCM at
this zone. Based on the simulation results, change in surface temperature has the
most deterministic effect on melting rate, melting time and the rate of heat
26
Nomenclature
P Pressure, Pa
Ri Internal radius, m
t Time, s
Ti Initial temperature, K
Tm Melting temperature, K
Ts Surface temperature, K
V Velocity, ms-1
Greek symbols
Density, kgm-3
27
µ Dynamic viscosity, kgm-1s-1
References
[1] Roy SK, Sengupta S. Melting of a free solid in a spherical enclosure: effects
paraffin wax stored in spherical shells. ASME J Solar Energy Eng. 2005; 127:
357–65.
of melting in a spherical shell. Int J Heat Mass Transfer 2007; 50 (9-10): 1790–
804.
study of spherical PCMs in solar energy storage. Solar Energy 2009; 83: 1245–
52.
[7] Eames IW, Adref KT. Freezing and melting of water in spherical enclosures
of the type used in thermal (ice) storage systems. ApplTherm Eng. 2002; 22:
733–45.
28
[8] Regin AF, Solanki SC, Saini JS. Experimental and numerical analysis of
AIAA/ASME joint thermos physics and heat transfer conference (CD ROM).
[9] Tan FL. Constrained and unconstrained melting inside a sphere. Int
[10] Tan FL, Hosseinizadeh SF, Khodadadi JM, Fan L. Experimental and
29