Effect of refrigerant charge variation on the energy and thermal performance of a domestic refrigera

Abstract

Keywords: Domestic Refrigerator; EER; Energy; Refrigerant Charge; Temperature

Introduction

Over the course of years, the domestic refrigerator has been maintained as one of the most widely used household appliances for food preservation, preventing the growth of microorganisms that damage food and affect human health. In 2020 in Mexico, approximately 31 million refrigerators were in use, which accounted for around 29% of the energy

consumption in the residential sector [1, 2]. This energy consumption is taken into consideration under normal operating conditions, but like any household appliance, the refrigerator can present problems in the functioning of its components, due to either electrical or mechanical aspects. In addition, problems can also occur due to microleaks or improper refrigerant gas charges, which could affect the

*Corresponding author. *E-mail address: jfbelman@ugto.mx This paper was recommended for publication in revised form by Editor-in-Chief Ahmet Selim Dalkılıç Published by Yıldız Technical University Press, İstanbul, Turkey Copyright 2021, Yıldız Technical University. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

thermal and energy performance of a refrigeration system [3, 4]. Thus, timely detection of improper operation in vapor compression refrigeration systems could help improve system performance as well as minimize unwanted environmental effects [5]. The evaluation of the amount of refrigerant in vapor compression systems is one of the main areas of study due to the effect it has on the energy performance of the system [6, 7], as well as the thermal performance according to the application of the refrigeration system. In this sense, in literature, there are experimental and theoretical works that allow for determining the causes for which refrigerant leaks occur. Li et al. [8] concluded, for instance, that refrigerant leaks are caused by prolonged operation of the refrigeration system, either by continuous vibrations in the pipeline or by corrosion inside the pipeline causing different types of cracks and holes. In most cases, the refrigerant leak is located at the joints of the valves and pipeline, at the elbows of the heat exchanger coils, and in the high-pressure liquid line [9]. On the other hand, a refrigerant overcharge can occur when the refrigeration system is charged with more refrigerant than those recommended by the manufacturer as a result of maintenance [10] or even due to a lack of calibration in the process and assembly line. Du et al. [11] investigated a refrigeration system working with inadequate refrigerant charges (overcharge and leak), and some of their results showed that refrigerant overcharging caused a decrease in the normalized coefficient of performance, COP. On the other hand, when there was a 20% refrigerant leak, it was observed that the temperature at the condenser outlet presented the greatest variation. Grace et al. [12] studied the performance of a vapor compression system when the R404A refrigerant charge varied from its design value (25% below and 25% above). The results showed that the COP did not present significant changes when the refrigerant charge was within a range of ± 25% of the design value. On the other hand, the COP decreased by 45% when the charge was below 25% and decreased by 13% for a refrigerant charge greater than 25%. In addition, the refrigerant overcharge caused the discharge pressure to increase significantly; otherwise, it happens with the suction pressure, where there were no significant changes for the different levels of refrigerant charge. Deymi-Dashtebayaz et al. [13] found that by optimizing the refrigerant charge, it was possible to save 785 GWh of energy consumption per year in the residential sector. The above when evaluating the effect of the charge of R22 from 540 g to 840 g in a split type of air conditioner. On the other hand, Kim and Kim [14] studied the cooling capacity of a refrigeration system working with different refrigerant charges. The authors concluded that the subcooling gradually decreased as the refrigerant overcharge condition increased due to the accumulation of refrigerant at the outlet of the condenser. Lee and Chang [15] proposed a method in which the amount of refrigerant and the detection of refrigerant leaks can be predicted based on the

starting characteristics of the air conditioner. Besides that, artificial intelligence has also been used to detect refrigerant leaks in real time. For example, Tassou and Grace [16] used neural networks for refrigerant leak detection; one neural network was used to predict the compressor discharge pressure and the other to predict the temperature of the refrigerant in the evaporator. With this, the system had the ability to detect small refrigerant leaks when the charge dropped below 33% of the nominal refrigerant charge. Yang et al. [17] used 35 fault identification rules for refrigerant leaks, considering a maximum leak of 30% of the rated charge. The results showed an accuracy of 91% in fault identification through these rules. Boeng and Melo [18] experimentally studied the thermodynamic behavior of a domestic refrigerator (factory refrigerant charge of 47 g) for refrigerant charges of 36.7 g and 64.7 g. Among the results, the authors concluded that as more refrigerant is added, it accumulates in the heat exchangers, increasing condensation and evaporation pressures, ultimately causing an increase in energy consumption of up to 30%. Gugulothu [19] evaluated the refrigeration effect by means of a design of experiments, varying three parameters, including refrigerant charge (86 g, 96 g, and 106 g). The results showed that when the refrigerator works with a refrigerant charge of 96 g, it shows a lower energy consumption and a better COP. As mentioned above, the variation of refrigerant charge in refrigeration systems based on vapor compression affects to some extent the energy and thermal performance of the system. Thus, the objective of this work is to evaluate the thermal and energy performance of a domestic refrigerator experimentally when it works with inadequate refrigerant charges, that is, charges below and above the optimum charge (86 g of R516A) [20], simulating leaks and excess refrigerant. As a novelty, the use of control graphs is proposed to better map and understand thermal behavior (refrigerant side and internal compartments) of the refrigerator under the operation of different refrigerant charges, the application of a statistical analysis is proposed. The foregoing with the purpose to evaluate the energy performance of the refrigerator according to those parameters that present greater thermal variation. Among the main results discussed are the EER, energy consumption, compressor power, and cost of energy consumption due to charge variation.

Experimental Methodology

The experimental system used to carry out this study consists of a domestic refrigerator, a data acquisition system, and an energy analyzer. The domestic refrigerator is from the Mabe brand with a volumetric capacity of 0.51 m3; it uses R516A as a low GWP alternative refrigerant and has a 280 W resistive element for automatic defrosting. It also has two fans, one attached to the condenser (forced convection heat dissipation) and another internal attached to the evaporator, which moves the flow of cold air inside the food

compartments. The compressor motor is 0.74 kW with a fixed speed, Embraco model EM3Y50HLP and a working voltage of 127 V at a 60 Hz frequency. Experimental System Instrumentation The refrigerator was instrumented to obtain the changes in thermal and energetic performance for the different charges of refrigerant R516A. Type K thermocouples were used for temperature measurement; these thermocouples were connected to a temperature module NI-9213 (operating temperature range of -40 °C to +70 °C). For the measurement of working pressures, suction pressure, and discharge pressure, two pressure transducers were used, which were connected to the NI-9207 module. Both the temperature and pressure sensors were connected to a NI cRIO-9030 chassis model and a computer with drivers and LabView SignalExpress. Power consumption was measured using a Fluke 438–II power quality analyzer. Table 1 shows the sensitivity and precision of the measurement made by the sensors used. Figure 1 shows schematically the detail of the instrumentation of the different components of the refrigerator (condenser, evaporator, and compressor) as well as the

internal compartments (freezer and fresh food compartment). The measurement points and the location of the temperature and pressure sensors are shown in blue, as shown in the following figure. For the condenser and evaporator, the refrigerant temperature was measured at points such as the inlet, in the middle position, and at the outlet with respect to the length of the heat exchanger. For the compressor, the temperature and pressure at the suction and discharge, as well as the casing, were measured. The air temperature inside the internal compartments of the refrigerator was also measured, with three measurement points determined within each compartment. Copper masses were used for the temperature measurements in the compartments, as specified by NOM-015-ENER-2018. The experimental tests were carried out under the following operating conditions: • Without food charge in the compartments and behind closed doors. • Testing started when the freezer and the fresh food compartment were at the same ambient temperature. • All the tests were carried out in a space where the ambient temperature was 32 °C ± 1.5 °C and the relative humidity of 40 ± 5%. • The thermostat level was set to position 3 (the middle position as shipped from the factory). • Data was recorded every 10 seconds for all temperature, pressure, and energy measurements. • 7 test hours were defined for each refrigerant charge in stable thermal conditions. • A replication was performed for each of the tests for data reliability purposes, thus obtaining an average for the parameters analyzed.

Figure 1. Schematic instrumentation of the experimental refrigerator.

Figure 2. Schematic process of refrigerant charge. Case Study Figure 2 shows schematically the vacuum and refrigerant charging process, which involves a digital manifold gauge, an analytical balance, and a vacuum pump. In this case, the refrigerator works with the azeotropic mixture R516A. A vacuum process was carried out for each refrigerant charge using a 0.37 kW vacuum pump (rotation speed of 2800 rpm, voltage of 110 v, and frequency of 60 Hz). A testo 550 brand digital manometer is used, which is connected to the compressor service line, to the vacuum pump, and to the refrigerant tank, which is located on a digital scale, to observe and record the refrigerant charge that enters the refrigerator. Initially, the refrigerator was evaluated with an optimum charge of 86 g for the refrigerant R516A, as established in the experimental work of Belman-Flores et al. [20], where the optimal charge was determined by the lowest energy consumption. Based on this charge, two refrigerant charges were proposed below (simulating a refrigerant leak) and two charges above (simulating an excess of refrigerant).

Table 2. Evaluated R516A refrigerant charges Conditions evaluated

Refrigerant overcharging may occur during maintenance, where the service technician may charge the system with more refrigerant than recommended by the manufacturer [10]. The different charges proposed in this work are presented in Table 2.

Results And Discussion

This section presents the thermal behavior and energy performance of the domestic refrigerator under different refrigerant charge conditions. First, all the temperature measurements are evaluated using the Tukey method to focus the statistical analysis on those temperature points that present greater variability for the different refrigerant charges. Thus, the measured points that represent significant variations with respect to the performance of the reference refrigerator (86 g optimum charge) are identified. According to this first study, a statistical analysis is proposed using control charts, which show the impact of the variation in the refrigerant charge on the operation of the refrigerator. The operation is also projected on a P-h diagram, and power consumption and energy efficiency performance (EER), are discussed. Finally, an energy cost analysis is shown for the different refrigerant charges. Data Analysis Using the Normal Distribution and the Tukey Method The measurements of the 15 temperature points (Fig. 1) are initially analyzed for the optimum charge since these data represent the normal behavior of the refrigerator; thus, they are used for checking if they follow a

Table 3. Statistics of the temperature data for the reference charge (86 g) Measuring point

SST is the sum of total squares, and SStrat is the sum of squares between treatments, and they are obtained by the equations (4) and (5). (4)

normal distribution. For this, the hypothesis method was used, which defines that if the P-value is less than the significance level (α=0.05), the data does not follow a normal distribution. Table 3 shows that the P-value for all the measured data is greater than the significance level; therefore, the data obtained for the optimum charge follow a normal distribution. All the temperature measurements for the different refrigerant charges were analyzed in order to evaluate the variability of the data and conclude which are most affected by the different refrigerant charges. Using the Tukey test, we can compare the difference between the mean and a critical value [21]. Equation (1) was used to find the calculated Tukey value (Tα). (1) Where, Qα (k, N - k) is the standardized range and its distribution obtained from the table [22], ni is the amount of data in each sample, N is the total number of elements comprising all the samples, and k is the number of groups of the samples being compared. MSE is the value of the mean square of the error obtained by the equation (2). (2) SSE is the sum of squares of the error, equation (3). (3)

Where, y is the experimentally measured value, − y is the average of the measured values. A significant difference occurs when the discrepancy between means is greater than the calculated Tukey value. In this case, the means of the distinct refrigerant charges were compared (μ70, μ80, μ100 and μ110) with regard to the mean of the optimum charge (μ86). Table 4 shows the differences in the averages of each refrigerant charge concerning the average of the reference charge. Additionally, the calculated Tukey value is presented, where the numbers shown in red are the data that represent a significant difference. For example, the value of 4.49 corresponds to the suction temperature; it exceeds the calculated Tukey value of 1.46, which indicates that the refrigerant temperature in the compressor suction is affected when the system works with a refrigerant overcharge (110 g). On the contrary, the discharge temperature did not present values that exceeded the calculated Tukey value; consequently, the discharge

Table 4. Tukey test for refrigerant charge variation and discrepancy analysis Measuring point

temperature in the compressor is discarded for a more detailed analysis. This behavior is similar to the findings by Li et al. [23], where the suction temperature presented greater sensitivity to excess refrigerant charge than the discharge temperature. Thus, the Tukey test initially allows for focusing the evaluation of the temperatures measured on the points that are really affected by the variation of refrigerant charge. Thermal Analysis for Refrigerant Charge Variation The effect of the variation of the refrigerant charge in a domestic refrigerator with respect to the optimal charge is presented in a visual and detailed way through individual control charts. According to the Tukey test, the temperatures that showed the greatest effect between the different refrigerant charges are in the compressor suction, compressor casing, middle position, and outlet of the condenser and evaporator, along with the three freezer measurements. The methodology and equations to estimate the control limits and the mean for the individual control charts are the same as those presented in the work of Pardo-Cely et al. [24]. Figure 3 shows the temperatures around the compressor, suction, and casing by means of control graphs for the

reference charge (optimal) and the other charges that simulate a refrigerant leak or overcharge. The figure also shows the upper and lower control limits (blue dotted lines), which were estimated to have three standard deviations. The red line is obtained from the mean of the reference data (optimum charge); each black point represents the average temperature measured every hour during 7 hours in a stable state, where for each test with its corresponding replicate an overall of 14 points were obtained. The orange points represent the temperature value that exceeds the control limits, indicating improper operation of the domestic refrigerator. Figure 3 shows that the temperature measured within the compressor casing tends to move away from the central value and approach the lower control limit when there is an excess refrigerant charge compared to the optimum charge. However, none of the points surpass the lower control limit, which indicates that an excess of refrigerant does not affect the temperature in the compressor casing. This is convenient since high temperatures around the compressor affect the stability of the lubricants and internal components of the compressor [25]. On the other hand, the temperatures measured in the suction of the compressor present lower points outside the control limits for a refrigerant charge of

100 g, moving away from the reference average by approximately 2.1 °C. For the 110 g charge, several points outside the range indicate inadequate compressor performance for the suction temperature, which deviated from the mean temperature by approximately 4.5 °C. This behavior reflects the fact that an overcharge of refrigerant in the refrigerator represents a reduction in the suction temperature. This is due to the fact that the refrigerant does not evaporate sufficiently in the evaporator causing it to flow through the suction line towards the compressor in the liquid phase. According to the Tukey test, the refrigerant temperature measured both in the condenser in the middle position and at the outlet presented a significant difference, therefore, a more detailed analysis was carried out using the control chart. The thermal behavior of these positions is presented in Figure 4, which shows that for the refrigerant charge of 110 g, there is a notable difference with respect to the reference temperature for the middle position and at the condenser outlet of approximately 4 °C. This increase in temperature is because the refrigerant accumulates in the condenser and therefore decreases the capacity of heat dissipation to the environment in the condenser; the same behavior is presented in the work of Bellanco et al. [10].

However, this behavior does not definitively represent a malfunction because the increase in temperatures is not outside the control limits. The excess refrigerant leads to its accumulation in the evaporator causing an increase in the average temperature of the refrigerant in the middle position and outlet of the evaporator as shown in Figure 5. The charge of 110 g causes an increase of 3 °C in the refrigerant temperature in these positions with respect to the 86 g charge. It can also be seen that for a low refrigerant charge (70 g), the temperature at the evaporator outlet moves away from the reference average; this refrigerant charge is insufficient for the correct operation of the evaporator. In conclusion, the analysis through the control charts does not indicate an inadequate operation in the evaporator as such since the data remains within the established control limits. Maintaining the proper temperatures in the internal compartments is ultimately the purpose of a domestic refrigerator. For example, according to the food safety guide, it is recommended that temperatures be less than

Acknowledgement

In this work, the impact on the operation of a domestic refrigerator when working with inadequate refrigerant charges was estimated through statistical analysis. Different refrigerant charges were evaluated experimentally; the charges of 70 g and 80 g represented a refrigerant leak, and the charges of 100 g and 110 g represented an excess of refrigerant. According to the results, the charge conditions (leakage or overcharge) near the optimal charge (86 g) did not present significant changes in the general functioning of the domestic refrigerator. On the other hand, the negative effects were presented for the charges of 70 g and 110 g. The 70 g charge generated an increase in the temperature in the freezer for the door position of 2.2 °C; the run time increased by 0.5%, causing an increase in energy consumption of 0.15 kWh/ day; and the EER decreased by 0.7 with respect to the optimal charge. The refrigerant overcharge of 110 g presented a greater effect on the suction temperature, moving away from the reference temperature by 4.5 °C, and the temperatures measured in the freezer moved away from the reference temperature by approximately 4 °C; the discharge pressure increased by 1.3 bar; and the run time increased by 21% regarding the run time of the reference charge. Consequently, these behaviors caused an increase in energy consumption of 0.56 kWh/day and a decrease in EER of

Data Availability Statement

The authors confirm that the data that supports the findings of this study are available within the article. Raw data that support the finding of this study are available from the corresponding author, upon reasonable request.

Ethics

There are no ethical issues with the publication of this manuscript.

References

1. Navarro-Esbrí J, Torrella E, Cabello R. A vapour [16] Tassou SA, Grace IN. Fault diagnosis and refrigerant compression chiller fault detection technique leak detection in vapour compression refrigeration based on adaptative algorithms. Application to systems. Int J Refrig 2005;28:680-688. [CrossRef] on-line refrigerant leakage detection. Int J Refrig [17] Yang J, Wu J, Yu X, Liang Y. Study on fault identifica- 2006;29:716-723. [CrossRef] tion rules for real refrigerant leakage in R290 room
2. Mehrabi M, Yuill D. Generalized effects of refrig- air conditioner based random forest algorithm. erant charge on normalized performance variables Expert Syst Appl 2024;238:122126. [CrossRef] of air conditioners and heat pumps. Int J Refrig [18] Boeng J, Melo C. Mapping the energy consumption 2017;76:367-384. [CrossRef] of household refrigerators by varying the refriger-
3. Chen G, Fang J, Li Z, Zhang S. Theoretical analysis ant charge and the expansion restriction. Int J Refrig of the impacts of refrigerant leakage on the perfor- 2014;41:37-44. [CrossRef] mance of a flash tank vapor injection heat pump. Int [19] Gugulothu SK. Enhancement of household refrig- J Refrig 2022;147:10-19. [CrossRef] erator energy efficiency by studying the effect of
4. Li Y, Yang J, Wu X, Liu Y, Zhuang Y, Zhou P, et al. refrigerant charge and capillary tube length. J Therm Leakage , diffusion and distribution characteristics Engineer 2021;7:1121-1129. [CrossRef] of refrigerant in a limited space: A comprehensive [20] Belman-Flores JM, Heredia-Aricapa Y, García- review. Therm Sci Engineer Prog 2023;40:101731. Pabón JJ, Pérez-García V, Pérez-Reguera CG. [CrossRef] Drop-in replacement of R134a in a household refrig-
5. Francis C, Maidment CG, Davies G. An investiga- erator with low-GWP refrigerants R513A, R516A, tion of refrigerant leakage in commercial refrigera- and R1234ze(E). Energies 2023;16:3422. [CrossRef] tion. Int J Refrig 2016;74:10-19. [CrossRef]
6. Bellanco I, Belío F, Vallés M, Gerber R, Salom J. 2nd ed. Colombia: Antioquia University; 2009. Common fault effects on a natural refrigerant, vari- [Spanish] able-speed heat pump. Int J Refrig 2022;133:259-
7. Du Z, Domanski AP, Payne WV. Effect of com- Tecnología Educativa; 2016. [Spanish] mon faults on the performance of different types of
8. Li K, Lan J, Zhou G, Tang Q, Cheng Q, Fang Y, et al. vapor compression systems. Appl Therm Engineer 2016;98:61-72. [CrossRef] Investigation on the influence of refrigerant charge
9. Grace IN, Datta D, Tassou SA. Sensitivity of refrig- amount on the cooling performance of air condi- eration system performance to charge levels and tioning heat pump system for electric vehicles. J parameters for on-line leak detection. Appl Therm Therm Sci 2018;28:294-305. [CrossRef] Engineer 2005;25:557-566. [CrossRef] [24] Pardo-Cely D, Belman-Flores JM, Heredia-Aricapa
10. Deymi-Dashtebayaz M, Farahnak M, Moraffa M, Y, Rodríguez-Valderrama DA, Morales-Fuentes Ghalami A, Mohammadi N. Experimental evalu- A, Gallegos-Muñoz A. Fault analysis in a domestic ation of refrigerant mass charge and ambient air refrigerator: Fan fault, condenser fouling, and area temperature effects on performance of air-condi- restriction. Int J Refrig 2022;154:290-299. [CrossRef] tioning systems. Heat Mass Transf Stoffuebertragung [25] Mohanraj M, Jayaraj S, Muraleedharan C, 2018;54:803-812. [CrossRef] Chandrasekar P. Experimental investigation of
11. Kim M, Kim MS. Performance investigation of a R290/R600a mixture as an alternative to R134a in a variable speed vapor compression system for fault domestic refrigerator. Int J Therm Sci 2009;48:1036- detection and diagnosis. Int J Refrig 2005;28:481- 1042. [CrossRef]
12. Lee C, Chang YS. Refrigerant charge prediction Inocuidad Alimentaria para Voluntarios. Available based on start-up characteristics of an air condi- at: https://www.fsis.usda.gov/sites/default/files/ tioner using a gray box model. J Mech Sci Technol media_file/2021-02/Cooking_for_Groups_SP.pdf. 2022;36:4859-4868. [CrossRef] Accessed Oct 8, 2024.

Share and Cite