The use of phase change materials in domestic
refrigerator applications
by
C. Marques* a,b,, G. Davies a, G. Maidment a, J. Evans a, I. Wood b
(a) Department
of Urban Engineering, London South Bank University, 103
Borough Road, London, SE1 0AA, UK; *email: [email protected]
(b) Adande Refrigeration, 45 Pinbush Road, South Lowestoft
Industrial Estate, Lowestoft, Suffolk, NR33 7NL, UK
1. Introduction
A growing global environmental awareness and the rising costs of energy are driving the demand for the
development of sustainable cooling technologies. In the UK, cold appliances are responsible for 17% of
average domestic electricity use [1]. Worldwide it has been estimated that there are approximately 1 billion
domestic refrigerators in use [2], and although their direct GHG emissions have been greatly reduced by
the introduction of hydrocarbon refrigerants, their indirect emissions remain very high due to the
electricity input needed for operation of the refrigerator. Most governments have implemented minimum
energy performance standards and energy labelling programs for household refrigerators aiming to regulate
the market and further improve the efficiency of these appliances.
The most common approaches to reduce the energy consumption of household refrigerators to date
include the use of optimised insulation and energy efficient compressors. Vacuum insulated panels (VIPs)
offer twice the level of insulation of polyurethane foam; nevertheless their reliability over lifespan and high
manufacturing and disposal costs has prevented their widespread use [3]. The compressor is the single
largest energy user in a refrigerator, responsible for more than 80% of the total energy consumed by the
appliance [4]. Compressor manufacturers have developed variable speed compressors that adjust the
refrigeration capacity in relation to the load by controlling the motor speed resulting in energy savings of up
to 40% [5]. However this technology is still very expensive limiting its use in a market that is particularly
sensitive to price.
Phase change materials (PCMs) are substances with high latent heat content that freeze and melt at a nearly
constant temperature, accumulating or releasing large amounts of energy during the process. The
application of PCMs in domestic refrigerators is a novel solution with the potential to improve the
appliance thermal stability and efficiency. The cooling energy stored in the PCM can be used to cool the
compartment, increasing the refrigerator energetic autonomy, when the power supply is switched off. This
approach was taken by Azzouz et al. [6], who tested a domestic refrigerator with 5 × 10-3 and 10 × 10-3 m
(i.e. 5 and 10 mm) ice slabs in contact with the evaporator surface. Their results showed that the time for
which the refrigerator could be operated without power supply increased by up to 5 and 9 hours
respectively, depending on the thermal load. It was observed however, that only 60% of the 10 mm slab
was frozen when the compressor switched off during the tests, which could be due to the low thermal
conductivity of the PCM, and/or the low cooling capacity of the 5 × 10-6 m3 (i.e. 5 cm3) swept volume
compressor employed. Another experimental study carried out by Cheng et al. [7] analysed the
performance of a fridge-freezer with a PCM fitted around the condenser pipes, which lowered the
condensing temperature and produced energy savings of 12% compared to the same fridge-freezer without
thermal storage. Gin et al. [8] covered 26% of the internal walls of a frost free upright freezer with PCM
panels reducing peak air and product temperature by 3°C and 1°C respectively during an electric defrost.
Numerous studies have shown that conventional household refrigerators frequently run at a higher
temperature than recommended by health and food safety legislation during real usage operation. In fact an
analysis of various surveys reported in the last 30 years has shown that 61.2% of refrigerators throughout
the world run at temperatures above 5°C [9, 10]. There are currently two types of household refrigerators
on the market: static and frost free; these distinctions indicate differences in the airflow and heat transfer
mechanism within the refrigerated compartment. The temperature stability is generally better in forced
convection (i.e. frost free) refrigerators than in static natural convection models, however, their energy
consumption is also considerably higher.
This paper presents a study on the application of PCMs to domestic refrigerators. It is the result of a 4 year
Industrial Case Award PhD supported by the EPSRC and Adande Refrigeration. The research focused on
the design and operation of the thermal storage refrigerator aiming to improve the appliance efficiency,
temperature stability and energetic autonomy. To achieve the project goals first the performance of a
conventional refrigerator was analysed, in particular the size and efficiency of the compressor and its impact
on the overall system performance. The operation of the thermal storage refrigerator was investigated by
estimating the PCM melting and freezing time. A computational fluid dynamics (CFD) model was used to
predict the airflow and temperature distribution within the thermal storage refrigerator. Several design
options were simulated to identify the most effective PCM configuration (horizontal or vertical) and phase
change temperature. A test rig was designed and constructed to validate experimentally the theoretical
models which predicted the performance of the refrigerator with a horizontal PCM. The experimental
results obtained for a prototype refrigerator will be presented and discussed.
2. Theoretical model of a conventional refrigerator
This section describes an analysis of compressor performance for a range of swept volume models. The
influence of compressor displacement on the refrigerator efficiency is demonstrated.
2.1 Compressor performance
Compressor manufacturer datasheets provide information on compressor performance under ASHRAE
conditions, i.e. condensing temperature of 54.4°C, and ambient, liquid and suction gas temperatures of
32.2°C. However, compressor performance under these conditions may not reflect how well compressors
will perform under more realistic conditions. In practice, both ambient and suction gas temperatures are
likely to be well below 32°C during normal refrigerator operation.
Compressor performance can be evaluated by its isentropic efficiency, which compares the actual
compressor energy consumption with the energy necessary for an ideal (i.e. reversible and adiabatic)
compression process. The isentropic efficiency varies between 0 and 1. An analysis of the compressor
performance for different displacements (i.e. compressor sizes) was carried out, to compare isentropic
efficiencies under typical refrigerator operating conditions. The RS+3 compressor unit selection tool
program developed by Danfoss [11] was used to predict performance data for each compressor
displacement at typical domestic refrigerator operating conditions. These conditions were a condensing
temperature of 35°C, ambient temperature of 25°C, evaporator superheat of 1K and compressor suction
temperature of 15°C, assuming a suction line heat exchanger (SLHE) efficiency of 65% [12].The CoolPack
software, version 1.46 [13] was used to calculate the compressor isentropic efficiency for each compressor
size at a range of evaporating temperatures (from -35°C to -5°C). The input data used in the program
(cooling capacity, power consumption and COP) were obtained from the RS+3 program. The isentropic
efficiency and cooling capacity for a range of compressor sizes are shown in Fig. 1 and 2 respectively.
Fig. 1: Isentropic efficiency for different
compressor displacements
Fig. 2: Cooling capacity for different
compressor displacements
As can be seen in Fig. 1, in general larger compressors are more efficient, with the isentropic efficiency
increasing from 0.4 to 0.6 as the displacement increased from 4 to 8 cm 3. The 8 cm3 compressor was the
most efficient size amongst the compressors considered, for typical domestic refrigerator storage
conditions (lower pressure ratio). Fig. 2 shows that the cooling capacity also increases with compressor
displacement.
2.2 Conventional refrigerator model
A theoretical model for a static conventional refrigerator was developed to establish the impact of
compressor selection on the system performance. The refrigerator efficiency was evaluated in terms of
total energy consumption and running time for the appliance. The running time was estimated from the
heat load into the refrigerator compartment and the cooling capacity of the compressor.
The heat load into the refrigerator of internal dimensions 0.75×0.45×0.46 m was calculated by considering a
steady state heat gain and then by adding an additional gain of 10% to account for real usage operation i.e.
periods of door openings and introduction of warm food into the compartment.
The ambient temperature was assumed to be 25°C, which corresponds to the test temperature of the
European Standard EN 153 [14] for subtropical refrigerators. It was considered that the back wall of the
refrigerator would be warmer than the side walls and door due to the condenser being mounted on this
wall. The condenser temperature was assumed to be 40°C. Additional heat from the compressor
compartment was assumed to provide an extra 7.5°C to the bottom surface of the refrigerator. The
compartment design temperature was 5°C and as a baseline, an evaporating temperature of -10°C was
considered. The three mechanisms of heat transfer, conduction, convection and radiation were considered
in the heat load to the refrigerated compartment. These were combined in the form of a global heat
transfer coefficient (U) within the model. The heat load into the refrigerator was then calculated using
equation1.
Q fridge  U A T
(1)
Table 1 presents the system performance for the different compressor models evaluated in section 2.1.
Table 1: Static refrigerator performance with different compressor models
Compressor Displacement
3
model
1
(cm )
4.0
2
3
4
5
6
Run Time (%)
Energy Consumption
COP
22%
(kWh/year)
92.2
5.7
7.48
8.05
14%
11%
10%
87.2
87.2
74.2
2.39
2.39
2.81
8.76
11.2
9%
7%
76.1
78.7
2.74
2.65
2.26
As can be seen in Table 1, the 8 cm3 compressor was optimal for the refrigerator studied under typical
operating conditions. An energy reduction of 19.6% was obtained by replacing the smaller 4 cm3
compressor by an 8 cm3 model and the run time decreased from 22% to 10%. Since the run time of the
appliance decreases as the compressor displacement increases the time required to cool the compartment
becomes very short but the compressor needs to be used frequently.
In order to fully exploit the higher performance of large compressors it is necessary to apply them in a
novel way so that their excess cooling capacity is used effectively in cooling the refrigerator compartment.
This can be achieved by storing cooling energy in a PCM. This results in the number of start/stop
operations of the compressor being reduced and the overall energetic autonomy of the refrigerator being
increased. The accumulated thermal energy also improves the compartment temperature stability when the
heat load suddenly increases e.g. during intense periods of door openings or product loading.
3. Modelling of a thermal storage refrigerator
This section describes the numerical model used to calculate the PCM melting and freezing time. The design
and operation of the thermal storage refrigerator were investigated by determining the influence of PCM
thickness and ambient temperature on the duration of the refrigerator off-cycle.
Water was selected as the PCM for the tests due to its high latent heat content and sharp freezing/melting
point. The encapsulated PCM was placed in thermal contact with the evaporator surface and both
components were placed in the refrigerator roof. The side walls of the container were insulated, so the
heat transfer was considered to be one dimensional. The upper wall of the PCM was cooled by the
evaporator and the bottom wall of the PCM enclosure was subjected to natural convection from the
refrigerator compartment.
In order to simplify the model is was assumed that the thermophysical properties of the PCM were
constant and different for each phase, density remained constant in the liquid and solid phases, the
convection in the melted regions was neglected and subcooling effects were not considered.
Since conduction was considered to be the only heat transfer mechanism inside the PCM, the conservation
of energy equations applied in the solid and liquid phases are described in Fig. 3.
Fig. 3: Phase change process – mathematical formulation
The boundary condition at the bottom surface (Qfridge) corresponds to the heat load into the refrigerator
compartment (determined in section 2.2). At the top boundary a cooling capacity (Qevaporator) was applied
during the freezing process. This is the cooling capacity estimated for the 8 cm3 compressor evaporating at
-10°C. The 8 cm3 compressor was identified in section 2.1 as the most efficient model for normal domestic
refrigerated storage conditions. It was assumed that the PCM was initially completely solid at a temperature
of 0°C and remained at this temperature during the melting process. The PCM was assumed to be in
thermal equilibrium with the refrigerator air at the beginning of the freezing cycle, therefore the PCM was
initially all liquid at 5°C, but its temperature was reduced to 0°C before the phase change process started.
The differential equations shown in Fig. 3 were discretized on a fixed grid using the finite difference method
[17]. The boundary conditions used to determine the PCM melting and freezing time are shown in Table 2.
Table 2: Boundary conditions used during the freezing and melting processes
FREEZING PROCESS
MELTING PROCESS
Ambient
Qfridge
Qevaporator (W)
Qfridge
Qevaporator (W)
Temperature (°C)
(W)
Tevaporator = -10°C
(W)
Tevaporator = 0°C
20
18.7
262
18.7
0
25
23.8
246
23.8
0
30
28.9
230
28.9
0
3.1 Effect of PCM thickness on the PCM melting and freezing time
Four PCM thicknesses were considered by the model. The PCM total storage capacity varied between 138
kJ for a 2 mm slab and 345 kJ for a 5 mm slab. The heat load and cooling capacity used to predict the PCM
melting and freezing times with different thicknesses correspond to an ambient temperature of 25°C and
the model results are shown in in Figs. 4 and 5 respectively.
Fig. 4: Effect of thickness on the PCM
melting time
Fig. 5: Effect of thickness on the PCM
freezing time
From Fig. 4 it can be seen that the melting time increased from 101 minutes (i.e. 6060 s) to 252 minutes
(i.e. 15120 s) as the PCM thickness increased from 2 mm to 5 mm. The freezing time was 13 minutes for
the 2 mm slab and 34 minutes for the 5 mm slab.
3.2 Effect of ambient temperature on the PCM melting and freezing time
The effect of ambient temperature on the PCM melting and freezing time was evaluated for a 5 mm PCM
slab. The fridge heat load and evaporator cooling capacity used in the model are listed in Table 2. The
simulation indicated that the melting time decreased from 320 minutes to 208 minutes when the ambient
temperature was increased from 20°C to 30°C, which corresponded to a reduction of 65% in refrigerator
autonomy i.e. without power supply.
The freezing time increased slightly with ambient temperature, with an additional 7 minutes required to
freeze the PCM at 30°C ambient, as compared to the time required at 20°C ambient.
4. Computational Fluid Dynamics simulation
The Computational Fluid Dynamics (CFD) software ANSYS 13.0 was used to simulate the airflow and
temperature profile within the natural convection thermal storage refrigerator. The influence of PCM
orientation and temperature were predicted by the CFD model.
4.1 Assumptions and boundary conditions
Only half of the refrigerator compartment was modelled due to the symmetry of its parallelepiped
geometry. The hexahedral mesh consisted of 455,400 cells, which were uniformly distributed across most
of the compartment, but with a finer grid i.e. more cells, close to the walls to more effectively simulate the
thermal boundary layer. In order to simplify the model, the melting/freezing process for the PCM was not
included; instead the PCM was simulated as a boundary condition of constant temperature (0°C) located at
either the top or back wall of the refrigerator compartment. The other compartment walls were assumed
to be subjected to a constant external temperature of 25°C. The thermal resistance of the walls in the
normal direction was calculated using the material thermal conductivity and thickness.
4.2 Natural convection and radiation models
The Rayleigh (Ra) number was calculated for both PCM orientations (horizontal/vertical), based on the
characteristic length of the PCM and the temperature difference between the PCM surface and the average
internal air (assumed to be 5°C). The calculated Ra number was 1×106 for the horizontal PCM and 7×107
for the vertical PCM, therefore a laminar flow model was assumed for both cases since Ra < 10 9. All of the
physical properties of the air were assumed to be constant except for the air density for which the
Boussinesq approximation was used. The Boussinesq approach assumes that the air density is uniform
except for the buoyancy term in the momentum equation.
Laguerre and Flick [15] have shown that the radiation heat transfer coefficient in refrigerators is of the
same order of magnitude as the natural convection heat transfer coefficient; hence radiation cannot be
ignored in the household refrigerator. In the present model, the radiation between the refrigerator surfaces
was taken into account using the discrete ordinates radiation model [16].
The equations were solved in steady state using the finite volume method with a sequential solver. The
solver employed the SIMPLE pressure-velocity coupling method. Solution convergence was considered to
be reached when the velocity residuals (conservation of momentum) and continuity residuals (conservation
of mass) decreased to 10-3 and the energy and discrete ordinates radiation (conservation of energy)
residuals decreased to 10-6.
4.3 CFD simulations for various PCM orientations
Fig. 6 presents a comparison of the simulations carried out for the various PCM orientations, i.e. (a)
vertical; (b) horizontal; and (c) combined horizontal and vertical, PCM. The contours of predicted
temperatures and the velocity vectors in the symmetry plane are presented for each simulation.
a) Vertical
b) Horizontal
Temperature (°C)
c) Horizontal and Vertical
Velocity vectors (m.s-1)
Fig. 6: Simulation comparison between PCM orientations
As can be observed in Fig. 6 orienting the PCM vertically (a) resulted in a stratified temperature profile with
a cold zone at the bottom and a warm zone at the top. The average temperature in the compartment was
11.2°C. The velocity vectors indicated a circular airflow pattern, with air flow along the walls and a region
of air recirculation in the bottom right corner. The air in the centre of the compartment was effectively
stagnant. This simulation is in agreement with the results of the experiments carried out by Amara et al.
[17] on a household refrigerator prototype fitted with a vertical heat exchanger.
The temperature contours for the horizontally oriented PCM (b) showed a slightly more homogeneous
temperature distribution than that found with the PCM positioned vertically and the average temperature
in the compartment was 2.4°C lower. The surface area of the PCM was the same for both orientations
simulated; therefore the results were directly comparable. The velocity vectors show that the cold dense
air flowed along the PCM and then downwards towards the centre of the cavity producing recirculation in
the bottom corners as the flow was directed upwards along the (warmer) walls of the enclosure. The
horizontal PCM simulation was validated by comparison with experimentally measured data obtained during
the current study. The results are presented in section 6.
The CFD models indicated that both the horizontal and vertical PCM orientations result in an average air
temperature above 5°C, which is the required compartment temperature specified in EN 153 [14]. One
option for reducing the temperature in the refrigerator would be to increase the heat transfer surface area
of the PCM, for example, by using a combination of horizontal and vertical PCMs as shown in Fig. 6 (c). This
scenario resulted in a more homogeneous temperature distribution in the refrigerator and an average air
temperature of 6.1°C was predicted. The velocity profile indicated that the airflow was a combination of
the two previous simulations. A more detailed description of the CFD analysis is provided in Marques et al.
[18].
4.4 Eutectic PCM
The effect of employing a eutectic PCM on the thermal storage refrigerator temperature was also
modelled. Eutectic mixtures are solutions of salts in water that have a phase transition temperature below
0°C. The proportion of salts in the solution defines the freezing point temperature at which all the
constituents crystallise in a eutectic reaction. Similar simulations to those undertaken for the ice PCMs
were carried out using eutectic PCMs at two different melting temperatures i.e. -2°C and -6°C, the results
are summarized in Table 3.
Table 3: Air temperatures and velocities in refrigerator compartment for all PCM configurations
PCM orientation
Temperature (°C)
Average
-1
Velocity (m.s )
Maximum Average
Maximum
Ice at 0°C
11.2
14.9
0.022
0.23
Eutectic at -6°C
Ice at 0°C
7.5
8.8
12.2
12.6
0.023
0.049
0.25
0.34
Eutectic at -6°C
5.0
9.8
0.054
0.36
Horizontal and
Ice at 0°C
6.1
9.8
0.042
0.23
Vertical PCM
Eutectic at -2°C
4.5
8.9
0.043
0.28
(a)
Vertical PCM
(b)
Horizontal PCM
(c)
Melting Point
The data presented in Table 3 shows that using a lower phase change temperature resulted in higher
velocities and lower temperatures in the thermal storage refrigerator. It also indicates that the required
refrigerator temperature can be achieved by using eutectic PCMs, e.g. for orientations (b) and (c).
5. Experimental setup and methods
An experimental rig was designed to validate the numerical models i.e. both those for the PCM melting and
freezing time, and for the CFD simulation with a horizontal PCM. The rig consisted of a prototype undercounter refrigerator cooled by an external coolant system as illustrated in Fig. 7. The coolant system was
used instead of a traditional vapour compression system, in order to prevent temperature cycling, thereby
maintaining steady state conditions within the heat exchanger throughout the experiments.
The PCM employed was 1 kg of pure (distilled) water. The PCM container (0.45×0.46×0.005 m) was an
open topped copper box, enclosed at the top with a heat exchanger supported by eight plastic spacers,
which ensured that the PCM thickness was 5 mm across the whole surface area. The heat exchanger
consisted of a copper plate with an 8.1 m copper tube soldered onto the plate surface. The tube was
arranged in a spiral shape to ensure heat transfer to the PCM was spatially uniform (Fig. 8).
Tout
Tin
Flowmeter
Heat Exchanger
PCM
Ethylene glycol /
water solution
Thermostatic
bath
Pump
Fig. 7: Schematic of the experimental test rig
Fig. 8: Heat exchanger
The ethylene glycol/water solution was cooled in the thermostatically controlled bath (to -10C) and
pumped through the horizontal heat exchanger, freezing the PCM beneath, which then cooled the
refrigerator compartment. After exiting the heat exchanger the coolant was returned to the thermostatic
bath completing the circuit. A ball valve was used to control the inlet flow to the heat exchanger. It was
opened at the beginning of each PCM charge cycle and closed once the PCM temperature readings
indicated that the PCM was frozen.
5.1 Temperature measurement
The temperatures within the PCM, at the inlet and outlet of the heat exchanger, within the refrigerator
compartment and the ambient temperature were measured with calibrated T-type thermocouples. The
average uncertainty of the temperature measurements was 0.3°C. A total of 14 thermocouples were
distributed uniformly (in 2D) inside the PCM. The air temperature distribution inside the empty refrigerator
compartment was measured using 37 hair-thin thermocouples (8 × 10-5 m diameter) to minimize the impact
on the compartment airflow. The thermocouples were distributed in five horizontal layers and three
vertical planes in half the refrigerator compartment. The ambient air temperature in the test room was
controlled to 20°C, 25°C or 30°C ± 0.5°C during the experiments.
5.2 Cooling duty and heat gain by the experimental refrigerator prototype
The cooling duty for the prototype was calculated from the measured coolant mass flow rate and the
temperature difference between the glycol inlet and outlet for the heat exchanger. The glycol coolant mass
flow rate was measured to an accuracy of +/- 0.015 kg s-1using a GEM turbine type volume flow meter and
a tachometer.
The overall heat transfer coefficient (U-value) for the refrigerator prototype was determined using
measured experimental data to determine the actual cabinet heat gain from the ambient environment using
the method reported by Azzouz et al. [6]. The U-value was determined to be 0.61 W.m-2.K-1.
6. Comparison of Numerical and Experimental results
6.1 Validation of numerical predictions for PCM melting and freezing time
The PCM initial temperature; the estimated heat gain into the refrigerator prototype; and the cooling duty
measured during the experiments, were introduced into the finite difference model described in section 3
and used to calculate the PCM melting and freezing time. The numerical model predictions and the
experimental results were then compared. Fig. 9 and 10 show the predicted and experimentally measured
PCM temperatures during a freezing/melting cycle at ambient temperatures of 20°C and 25°C respectively.
Fig. 9: PCM temperature during the freezing
and melting cycle at 20°C ambient
Fig. 10: PCM temperature during the freezing
and melting cycle at 25°C ambient
The PCM temperatures predicted by the numerical model closely matched the temperature profile
measured during the experiments as demonstrated in Figs. 9 and 10. A similar profile was obtained at 30°C
ambient. PCM melting times of 175 minutes (2.9 hours) and 290 minutes (4.8 hours) were predicted for
ambient temperatures of 30°C and 20°C respectively. There was a slight discrepancy in the PCM
temperature at the beginning of freezing cycle as the sensible heat was predicted to be removed from the
PCM more slowly by the numerical model than that measured in the experimental rig. It is believed that the
small difference between the model predicted results and the experimental data is due to the fact that the
model neglected convection currents within the PCM (while in liquid state), which would have enhanced
the heat transfer rate and therefore resulted in faster freezing during the first part of the cycle. Table 4
shows the average difference between the predicted and experimental results for total freezing and melting
times for a 5 mm slab.
Table 4: Comparison between predicted and experimental freezing and melting times at different ambient
temperatures for a 5 mm slab
Ambient
Measured
Measured
Temperature Cooling Duty Heat Gain
Freezing Time (min)
Melting Time (min)
Freezing
Melting
Measured error (%) error (%)
(°C)
20
(W)
191
(W)
21.3
Predicted
Measured
Predicted
57
57
294
290
0.3
-1.5
25
220
27.1
54
55
237
243
1.2
2.3
As can be seen in Table 4 there was a good agreement between the numerical model prediction and
experimentally measured PCM freezing and melting times. Overall, the experimental results indicated that
the integration of a 5 mm PCM slab into the refrigerator would allow 3 to 5 hours of continuous operation
without power supply depending on the ambient conditions.
6.2 Validation of the CFD model with a horizontal PCM
The horizontal PCM configuration was validated against experimental data collected from the test rig. The
temperatures measured by the thermocouples were compared with the CFD predictions at exactly the
same locations. Figs. 11 and 12 illustrate the compartment temperature in the vertical symmetry plane at z
= 0.23 m, for: the CFD prediction (Fig. 11) and for the experimentally measured data (Fig. 12), at an
ambient temperature of 25°C.
Fig. 11: CFD predicted temperatures (°C) in the
vertical symmetry plane at z = 0.23 m
Fig. 12: Experimental temperatures (°C) in the
vertical symmetry plane at z = 0.23 m
The predicted centre temperatures shown for the vertical symmetry plane were marginally lower (by
approximately 1.6C) than the measured temperatures recorded for the refrigerator prototype at the same
location. However, the temperatures measured close to the cabinet walls were in close agreement with the
CFD prediction (difference < 0.2C). The compartment temperature in the horizontal symmetry plane at y
= 0.38 m for both the CFD prediction and the experimentally measured data are presented in Figs. 13 and
14 respectively.
Position – x axis (m)
Position – x axis (m)
Fig. 13: CFD predicted temperatures (°C) in the
horizontal symmetry plane at y = 0.38 m
Fig.14: Experimental temperatures (°C) in the
horizontal symmetry plane at y = 0.38 m
From Figs. 13 and 14, it is seen that the results predicted and measured for the y = 0.38 m horizontal plane
showed close agreement (difference < 0.4C).
Table 5 presents the average and maximum air temperatures predicted by the CFD model and measured in
the refrigerator compartment at steady state for a range of ambient conditions.
Table 5: Predicted vs. measured air temperatures in the thermal storage refrigerator
Ambient
20°C
Temperature
25°C
30°C
Measured
Predicted
Measured
Predicted
Measured
Predicted
Average
8.05
7.45
9.95
8.70
11.69
10.40
Maximum
10.09
10.31
12.97
12.50
15.41
14.65
The average predicted compartment temperature was 1°C lower than the average measured temperature.
Overall, the predicted CFD temperatures showed close agreement with the experimental temperatures
measured in the refrigerated compartment at all ambient temperatures. The error in the average predicted
compartment temperature was approximately 10%. This could be due to a radiation effect on the centre
thermocouples leading to higher temperatures being measured. However, the maximum temperatures
predicted by the CFD simulation were very close to those measured in the compartment, with an average
error of only 2%.
Conclusions
The conventional refrigerator analysis demonstrated that for current single speed compressors, efficiency
increases with compressor displacement. The method proposed to exploit the superior performance of
large compressors is to accumulate their high cooling capacity in a PCM increasing the refrigerator
autonomy i.e. off-cycle period, without power supply, from a few minutes to several hours. Employing thin
PCM slabs (≤ 5 mm) ensures that the net volume of the compartment is not substantially reduced, while
moderate length compressor run times (i.e. on-cycle times) are obtained. Both the numerical simulation
and the experimental results demonstrated that the integration of a 5 mm PCM slab into the refrigerator
allowed between 3 and 5 hours of continuous autonomous operation depending on the thermal load. The
combination of a large displacement compressor and a thin PCM is a novel design approach with the
potential to significantly enhance refrigerator efficiency and temperature stability.
The CFD model predicted the airflow and temperature distribution within the thermal storage refrigerator.
Several design options were simulated to identify the most effective PCM configuration (horizontal or
vertical) and the optimum phase change temperature. This analysis enabled identification of the best design
options to optimise the performance of the thermal storage refrigerator. A horizontal PCM configuration
was found to be more efficient than a vertical PCM. For the horizontal PCM, the CFD predicted
temperatures were compared with experimentally measured values and found to be in close agreement.
Both the simulation and the experiments results suggested that a eutectic with a phase change temperature
below 0°C would need to be employed to maintain the compartment temperature within the required
range for domestic refrigerators.
Acknowledgements
The financial support of the Engineering and Physical Sciences Research Council (EPSRC) and Adande
Refrigeration is gratefully acknowledged.
Nomenclature
A
k
Qevaporator
Qfridge
s
t
T
U
Surface area (m2)
Thermal conductivity (W.m-1.K-1)
Cooling capacity (W)
Refrigerator heat gain (W)
Solid-liquid interface
Time (s)
Temperature (K)
Global heat transfer coefficient (W.m-2.K-1)
Greek symbols
α
Thermal diffusivity (m2.s-1)
λ
Latent heat of fusion (J.kg-1)
ρ
Density (kg.m-3)
References
[1]
[2]
[3]
DEFRA. Act on CO2 Calculator: Data, Methodology and Assumptions Paper. 2008
www.defra.gov.uk/environment/climatechange/uk/individual/actonco2 [Accessed 16 May 2010].
D. Coulomb, Refrigeration and cold chain serving the global food industry and creating a better
future: two key IIR challenges for improved health and environment. Trends in Food Science and
Technology, 2008;19: 413-417.
Market Transformation Programme. Innovation Briefing Note IBNC16: Vacuum insulated panels for
refrigerated appliances. 2006. http://efficient-products.defra.gov.uk/spm/download/document/id/667
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
Jähnig D I, Reindl D T, Klein S A. A semi-empirical method for representing household
refrigerator/freezer compressor calorimeter test data. ASHRAE Transactions 2000;106.
Bansal P K. Developing new test procedures for household refrigerators: harmonisation issues and
future R&D needs - a review. International Journal of Refrigeration 2003;26:735-48.
Azzouz K, Leducq D, Gobin D. Enhancing the performance of household refrigerators with latent
heat storage: An experimental investigation. International Journal of Refrigeration 2009;32:1634-44.
Cheng W-L, Mei B-J, Liu Y-N, Huang Y-H, Yuan, X-D. A novel household refrigerator with shapestabilized PCM (Phase Change Material) heat storage condensers: An experimental investigation.
Energy 2011; 36:5797-804.
Gin B, Farid M M, Bansal P K. Effect of door opening and defrost cycle on a freezer with phase
change panels. Energy Conversion and Management 2010;51:2698-706.
Laguerre O, Derens E, Palagos B. Study of household refrigerator temperature and analysis of factors
affecting temperature: A French survey. International Journal of Refrigeration 2002;25:653-9.
James S J, Evans J, James C. A review of the performance of household refrigerators. Journal of Food
Engineering 2008;87:2-10.
Danfoss RS+3 Compressor and Condensing unit selection tool. 3.6 ed. (1998) [Online].
http://www.danfoss.com/BusinessAreas/RefrigerationAndAirConditioning/Product+Selection+Tools+
Details/RSplus3.htm: Danfoss. [Accessed 10 May 2011].
P. Domanski, Theoretical evaluation of the vapour compression cycle with a liquid-line/suction-line
heat exchanger, economizer and ejector, NISTIR 5606 (1995)
Coolpack Software (1998) [Online]. Available:
http://www.et.web.mek.dtu.dk/Coolpack/UK/download.html: Denmark Technical University,
Department of Mechanical Engineering.
EN 153:2006 Methods of measuring the energy consumption of electric mains operated household
refrigerators, frozen food storage cabinets, food freezers and their combinations, together with
associated characteristics. European Committee for Standardization. 2006.
Laguerre O, Flick D. Heat transfer by natural convection in household refrigerators. Journal of Food
Engineering 2004;62:79-88.
Colomer G, Costa M, Cònsul R, Oliva, A. Three-dimensional numerical simulation of convection and
radiation in a differentially heated cavity using the discrete ordinates method. International Journal of
Heat and Mass Transfer 2004;47:257-69.
Amara B S, Laguerre O, Charrier-Mojtabi M-C, Lartigue B, Flick D. Numerical and Experimental
study of convective airflow in a household refrigerator model. Eurotherm Seminar 77 - Heat and
Mass transfer in Food processing, 2005, Parma, Italy.
Marques A C, Davies G F, Evans J A, Maidment G G, Wood I D. Theoretical modelling and
experimental investigation of a thermal energy storage refrigerator. Energy 2013;55:457-65.