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:

marquesc@lsbu.ac.uk(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 Fig. 2: Cooling capacity for

different compressor displacements 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 cm3. 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.

Qfridge=UAΔT (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

1 4.0 22% 92.2 2.262 5.7 14% 87.2 2.393 7.48 11% 87.2 2.394 8.05 10% 74.2 2.815 8.76 9% 76.1 2.746 11.2 7% 78.7 2.65

COPCompressor

modelRun Time (%)

Energy Consumption (kWh/year)

Displacement (cm3)

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

Ambient Temperature (°C)

Qfridge

(W)Qevaporator (W)

Tevaporator = -10°CQfridge

(W)Qevaporator (W) Tevaporator = 0°C

20 18.7 262 18.7 025 23.8 246 23.8 030 28.9 230 28.9 0

FREEZING PROCESS MELTING PROCESS

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 Fig. 5: Effect of thickness on the PCM

melting time 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 < 109. 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 c) Horizontal and Vertical

Temperature (°C)

Velocity vectors (m.s-1)

Fig. 6: Simulation comparison between PCM orientationsAs 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

Average Maximum Average MaximumIce at 0°C 11.2 14.9 0.022 0.23

Eutectic at -6°C 7.5 12.2 0.023 0.25Ice at 0°C 8.8 12.6 0.049 0.34

Eutectic at -6°C 5.0 9.8 0.054 0.36Ice at 0°C 6.1 9.8 0.042 0.23

Eutectic at -2°C 4.5 8.9 0.043 0.28

Temperature (°C) Velocity (m.s-1)

Vertical PCM

Horizontal PCM

Horizontal and Vertical PCM

Melting Point

(a)

(b)

(c)

PCM orientation

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 under-counter 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).

PumpThermostatic bath

Tin

Tout

FlowmeterPCM

Heat Exchanger

Ethylene glycol / water solution

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 -10C) 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 Fig. 10: PCM temperature during the freezing

and melting cycle at 20°C ambient 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

Predicted Measured Predicted Measured20 191 21.3 57 57 294 290 0.3 -1.525 220 27.1 54 55 237 243 1.2 2.3

Ambient Temperature

(°C)

Measured Cooling Duty

(W)

Measured Heat Gain

(W)

Freezing error (%)

Melting error (%)

Freezing Time (min) Melting Time (min)

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 Fig. 12: Experimental temperatures (°C) in the vertical symmetry plane at z = 0.23 m vertical symmetry plane at z = 0.23 m

The predicted centre temperatures shown for the vertical symmetry plane were marginally lower (by approximately 1.6C) 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 < 0.2C). 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.

Fig. 13: CFD predicted temperatures (°C) in the Fig.14: Experimental temperatures (°C) in the horizontal symmetry plane at y = 0.38 m 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 < 0.4C).

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

Position – x axis (m) Position – x axis (m)

Ambient Temperature

Measured Predicted Measured Predicted Measured PredictedAverage 8.05 7.45 9.95 8.70 11.69 10.40Maximum 10.09 10.31 12.97 12.50 15.41 14.65

20°C 25°C 30°C

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 Surface area (m2) Greek symbolsk Thermal conductivity (W.m-1.K-1) α Thermal diffusivity (m2.s-1)Qevaporator Cooling capacity (W) λ Latent heat of fusion (J.kg-1)Qfridge Refrigerator heat gain (W) ρ Density (kg.m-3)s Solid-liquid interface

t Time (s)T Temperature (K)U Global heat transfer coefficient (W.m-2.K-1)

References

[1] DEFRA. Act on CO2 Calculator: Data, Methodology and Assumptions Paper. 2008 www.defra.gov.uk/environment/climatechange/uk/individual/actonco2 [Accessed 16 May 2010].

[2] 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.

[3] 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] 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.

[5] 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.

[6] 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.

[7] Cheng W-L, Mei B-J, Liu Y-N, Huang Y-H, Yuan, X-D. A novel household refrigerator with shape-stabilized PCM (Phase Change Material) heat storage condensers: An experimental investigation. Energy 2011; 36:5797-804.

[8] 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.

[9] 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.

[10] James S J, Evans J, James C. A review of the performance of household refrigerators. Journal of Food Engineering 2008;87:2-10.

[11] 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].

[12] P. Domanski, Theoretical evaluation of the vapour compression cycle with a liquid-line/suction-line heat exchanger, economizer and ejector, NISTIR 5606 (1995)

[13] Coolpack Software (1998) [Online]. Available: http://www.et.web.mek.dtu.dk/Coolpack/UK/download.html: Denmark Technical University, Department of Mechanical Engineering.

[14] 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.

[15] Laguerre O, Flick D. Heat transfer by natural convection in household refrigerators. Journal of Food Engineering 2004;62:79-88.

[16] 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.

[17] 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.

[18] 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.