- Email: [email protected]
How to achieve full liquid conditions at the capillary tube inlet of a household refrigerator
International Journal of Refrigeration 100 (2019) 265–273
Contents lists available at ScienceDirect
International Journal of Refrigeration journal homepage: www.elsevier.com/locate/ijrefrig
How to achieve full liquid conditions at the capillary tube inlet of a household refrigerator Laetitia Bardoulet a,∗, José M. Corberán a, Santiago Martínez-Ballester b a b
Institute for Energy Engineering, Universitat Politècnica de València, Camino de vera, s/n, Valencia 46022, Spain Ingersol Rand, Sant Feliu de Llobregat 08980, Spain
a r t i c l e
i n f o
Article history: Received 18 July 2018 Revised 2 January 2019 Accepted 4 February 2019 Available online 7 February 2019 Keywords: Household refrigerator Capillary tube inlet Refrigerant visualization Non-equilibrium Subcooled vapor COP
a b s t r a c t The capillary tube with a liquid-to-suction heat exchanger (CT-LSHX) is a component that is widely used in household refrigerators. Recent works have indicated that even when measuring subcooled conditions at the condenser outlet, the condition at the capillary tube inlet is a two-phase flow. The present work was dedicated to analyzing the actual refrigerant conditions at the capillary tube inlet and to investigating how full liquid conditions could be achieved. The research was performed using a typical household refrigerator with corresponding fresh food and freezer compartments, replacing the original refrigerant-to-air condenser with a refrigerant-to-water condenser. This allowed, first, the condensation conditions to be controlled and, second, the estimation of the refrigerant conditions at the condenser outlet from the heat exchanger balance. The obtained results indicated the presence of a non-equilibrium two-phase flow, composed of subcooled vapor and subcooled liquid, at the capillary tube inlet, with both liquid and vapor entering the capillary tube as a vortex with small, fast fluctuations of the liquid level. This non-equilibrium indicated that the subcooling, evaluated from the pressure and temperature of the refrigerant at the condenser outlet, was only apparent and did not allow the evaluation of the actual enthalpy. Finally, by using a smaller capillary tube diameter and increasing the compressor speed, full liquid conditions at the capillary tube inlet were achieved. Furthermore, a performance comparison between the original and the new design revealed that the COP was higher with full liquid conditions. © 2019 Elsevier Ltd and IIR. All rights reserved.
Comment atteindre des conditions de plein liquide à l’entrée du tube capillaire d’un réfrigérateur domestique Mots-clés: Réfrigérateur domestique; Ouïe du tube capillaire; Visualisation du frigorigène; Vapeur sous-refroidie; COP
1. Introduction In household refrigerators, a capillary tube with a liquidto-suction heat exchanger (CT-LSHX) is the most widespread expansion device used due to its simplicity and low cost. Liquid conditions at the capillary tube inlet are necessary to avoid negative consequences, such as noise (Hartmann and Melo, 2013) or hysteresis effects as well as the reduction of CT-LSHX effectiveness (Liu and Bullard, 1997). Another consequence of the absence of ∗
Corresponding author. E-mail address: [email protected] (L. Bardoulet).
https://doi.org/10.1016/j.ijrefrig.2019.02.006 0140-7007/© 2019 Elsevier Ltd and IIR. All rights reserved.
liquid at the inlet of the capillary tube is a lower cooling capacity due to a certain decrease in the refrigerant mass flow rate, which affects its global efficiency. Subcooled conditions at the capillary tube inlet are always reported in literature. Nevertheless, some recent works have concluded that the subcooled conditions at the capillary tube inlet may not be real, which, if true, would have consequences for efficiency and, more generally, energy consumption. Boeng and Melo (2014) performed an experimental study using a refrigerator-freezer equipped with R600a to identify the optimal combination of the refrigerant charge and the capillary tube by varying the expansion device’s capacity. These researchers then
266
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Nomenclature cp h m˙ P Q˙ SC T W˙
Specific heat capacity, [kJ.kg−1 .K−1 ] Specific enthalpy, [kJ.kg−1 ] Mass flow rate, [kg.s−1 ] Pressure, [bar] Rate of heat transfer, [W] Subcooling, [K] Temperature, [°C] Rate of work, power [W]
Greek symbols α Air flow ratio, [-] τ Time ratio, [-] Subscripts cap Capillary tube comp Compressor cool Cooling evap Evaporator p Pressure [bar] in Inlet out Outlet ref Refrigerant w Water Acronyms and abbreviations CT Capillary tube COP Coefficient of performance FF Fresh food compartment FZ Freezer compartment HX Heat exchanger LSHX Liquid-to-suction heat exchanger PFA Perfluoroalkoxy
complemented this work by applying a model (Hermes et al., 2008) to examine the correlations between the degree of the valve opening and the equivalent capillary diameter. This study revealed a good prediction of the mass flow rate, except for the tests with fully open valves or a subcooling of less than 5 K. Through visualization of the capillary tube inlet, the authors concluded that even with a subcooling of 5 K, a two-phase flow still occurred at the inlet. As a possible explanation, Boeng and Melo highlighted that at the capillary inlet, there was a non-equilibrium mixture of subcooled liquid and saturated vapor instead of a purely liquid phase. The findings also revealed that the vapor quality at the capillary tube inlet ranged from 2% to 12%, depending on the refrigerant charge. Additionally, Martinez-Ballester et al. (2017), through the visualization of the condenser outlet and capillary tube inlet of a household refrigerator filled with R600a, concluded that the capillary tube inlet possessed two-phase flow conditions although a 14 K subcooling was measured. Furthermore, Lee et al. (2016) visualized a two-phase refrigerant flow at the inlet of the capillary tube of a commercial refrigeration system as well, working with isobutane. These researchers set up a transparent tube at the condenser outlet and an oil separator at the compressor discharge; then, they measured both the vapor and liquid temperatures as the condensation pressure. This study revealed that the refrigerant was in a non-equilibrium state, where both the subcooled liquid and the subcooled vapor coexisted at the capillary tube inlet. The authors subsequently proposed a method to calculate the enthalpy for non-equilibrium refrigerant conditions. They were able to estimate the vapor quality for the
different condensing pressures, ranging from 2.44% to 4.29%, at the condenser outlet. Ko and Jeong (2017) complemented the previous work by studying the effects of a non-equilibrium subcooled two-phase flow on the performance of a vapor compression refrigeration system. The authors estimated that the COP may be overestimated by 19.4% if the thermodynamic properties table is used for a nonequilibrium refrigerant state. Thus, various accounts are provided to explain the presence of the two-phase flow at the capillary tube inlet. As stated by Lee et al. (2016), an explanation for the presence of the vapor, although a certain subcooling is measured, is the non-equilibrium state of the refrigerant, composed of the subcooled liquid and subcooled vapor. In comparison, Boeng and Melo (2014) proposed non-equilibrium between the saturated vapor and the subcooled liquid. In their study, Martinez-Ballester et al. (2017) set up a transparent tube of 1.5 m at the condenser outlet. The authors stated that the transparent section was relatively long and could be considered adiabatic compared to a normal tube. Furthermore, the refrigerant had enough length to achieve an equilibrium state or, at least, to reveal some evolution towards equilibrium. This study provided a hypothesis for the presence of vapor at the capillary tube inlet: the capillary tube could be oversized and, therefore, to match the compressor flow rate with the flow rate through the capillary tube, the inlet should be in two-phase flow conditions to produce a higher pressure drop. The main objectives of the present work were, first, to assess the actual conditions at the capillary tube inlet of a household refrigerator system and, second, to find a way to achieve full liquid conditions at the capillary tube inlet. In order to reach these goals, four different experimental campaigns were performed by means of an innovative test bench and the use of a household refrigerator. First, the operating conditions were determined with the original condenser and capillary tube for an optimal charge of the refrigerant. Second, these operating conditions were reproduced with a refrigerant-to-water condenser to enable the measurement of the heat produced during the condensation and, thus, to determine the vapor quality or the subcooling degree at the capillary tube inlet. Third, a smaller capillary tube was tested with different compressor speeds to fill the capillary tube inlet with liquid. Fourth and finally, a performance analysis comparison between both the original and the modified design was performed. 2. Experimental set-up 2.1. Test bench The original apparatus was a high-efficiency, two-door refrigerator-freezer from the A+++ energy class, with no frost, and operating with R600a. It contained a fin and tube evaporator, a tube and wire condenser, a variable-speed 7 cm3 hermetic reciprocating compressor, and a capillary tube with a liquid-tosuction heat exchanger as the expansion device. The original refrigerator was modified in order to test different configurations of condensers and capillary tubes (Fig. 1). A set of valves allowed operating with either the original refrigerant-to-air condenser or a refrigerant-to-water condenser, composed of twin copper tubes with a 3.16 mm diameter welded onto a 2.53 m length and carefully insulated. In order to visualize the flow, a PFA (Perfluoroalkoxy) transparent tube section was inserted at the outlet of each condenser configuration. In the case of the refrigerant-to-air condenser, the final 1.5 m of the original condenser was cut off and replaced with the same length of transparent tube. An extra transparent tube section of 0.5 m was added to the refrigerant-to-water condenser.
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
267
Fig. 1. Test bench.
Two different capillary tubes were set up in the same suction line: the original one, with a 0.6 mm diameter, and a smaller one, with a 0.55 mm diameter. They were both laterally configured, i.e., both capillary tubes were welded to the suction line. Each capillary tube was mounted with a vertical transparent filter of PFA, filled with zeolite balls (Fig. 1), which allowed the visualization of the entrance of the capillary tube. A set of valves allowed operating with one capillary tube or another. Fig. 1 presents a drawing of the test bench, depicting the installed instrumentation: thermocouples, RTDs, mass flow meters, and pressure transducers. The refrigerant loop was equipped with a Coriolis mass flow meter (MFM_ref), located at the compressor discharge, that measured the refrigerant mass flow rate. The temperature of the refrigerant at the inlet of the refrigerant-to-water condenser (TcondIN) and the refrigerant inlet temperature of each capillary tube were measured using RTDs (TcapIN). The water loop was powered by a diaphragm pump and a by-pass system that allowed the modification of the water mass flow rate, which was measured using a Coriolis mass flow meter (MFM_w). Furthermore, RTDs were placed at both the inlet and the outlet of the condenser’s waterside in order to measure the inlet (TwIN) and outlet (TwOUT) water temperatures. The rest of the instrumentation was shared by both condensers’ configurations. A thermowell, with a thermocouple inside, measured the temperature at the outlet of the condenser (TcondOUT), and a pressure sensor allowed the measurement of the condensation pressure (Pcond). The evaporation pressure was measured by a pressure transducer (Pev), and the compressor power input was determined with a Wattmeter. The actual uncertainties of the different measurements are listed in Table 1. These values were used for the assessment of the final uncertainties of the different results: COP, heating and cooling capacities and subcooling, by the uncertainty propagation method.
Table 1 Employed sensors and corresponding uncertainty. Temperatures
Uncertainty
Water inlet temperature: RTD Water outlet temperature: RTD Refrigerant inlet condenser temperature: RTD Refrigerant outlet condenser temperature: thermocouple Capillary tube inlet: RTD Pressures Condensation pressure: pressure transducer Evaporation pressure: pressure transducer Mass flow rates Water mass flow rate: Coriolis mass flow meter Refrigerant mass flow rate: Coriolis mass flow meter Power Compressor power: Wattmeter
±0.07 K ±0.07 K ±0.078 K ±0.2 K ±0.07 K ±0.04 bar ±0.078 bar ±0.1% of rate ±0.25% of rate ±1.2 W
2.2. Experimental test campaigns Table 2 describes the main parameters of the four experimental campaigns that were performed. The experiments were carried out in a climatic chamber maintained at 25 °C. As the best manner to analyze and understand a refrigerant cycle, steady tests were conducted. In order to assess the conditions at the capillary tube inlet, experimental campaign 1 dealt with the determination of the operating conditions of the apparatus, with both the original condenser and the capillary tube, for the optimal charge of the refrigerant, the one that maximized the COP. Experimental campaign 2 dealt with the use of the water condenser to enable the accurate evaluation of the enthalpy of the refrigerant at the outlet of the condenser, thereby, at the capillary tube inlet conditions. However, first of all, the refrigerant charge had to be determined in order to achieve the same operating
268
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Table 2 Performed experimental campaigns. Experimental campaign Aim
1 Determining the optimal refrigerant charge
2 Assessing the capillary tube inlet conditions
3 Getting full liquid conditions at the capillary tube inlet
Condenser Capillary tube diameter (mm) Compressor speed (rpm) Refrigerant charge (g)
air 0.6 1600 From 74.5 to 90
water 0.6 1600 64
water 0.55 180 0–250 0–350 0 64
conditions that were attained with the original condenser. Therefore, we first performed a series of tests, varying the charge and the water flow conditions in order to obtain exactly the same operating conditions for the refrigerant. Then, the conditions at the inlet of the capillary tube were determined and analyzed. Afterwards, the operating conditions were varied by adjusting the water flow rate in order to check the influence of the condensing conditions on the refrigerant conditions at the inlet of the capillary tube. Furthermore, this enabled the comparison of the differences between the apparent subcooling estimated from the pressure and temperature measurements and the subcooling calculated from the enthalpy of the refrigerant at the outlet of the condenser. The aim of experimental campaign 3 was to search for other possible balance points in the matching between the capillary tube design (diameter) and the compressor speed at which full liquid conditions could be permanently maintained at the inlet of the capillary tube. To do so, a capillary tube with a 0.55 mm diameter, instead of the original 0.6 mm tube, along with different compressor speeds were tested. It should be pointed out that full liquid conditions could not be obtained with the original capillary tube at the maximum compressor speed. To assess whether the performance obtained by the improved design determined in experimental campaign 3 (a higher compressor speed and a smaller capillary tube) was better than the original design, a fourth experimental campaign was carried out using two capillary tubes of different diameters (0.55 mm and 0.6 mm) at the same compressor speed. In order to obtain a fair comparison, an optimal refrigerant charge study was performed for each capillary tube size at the same compressor speed (2900 rpm). 3. Results and discussion 3.1. The optimal charge of the refrigerator with the original condenser In order to determine the optimal refrigerant charge, a charge variation study was performed. The charge was varied from 74.5 g to 90 g. Results at 90 g were not included in the data analysis because the suction became frozen, and, therefore, it was not possible to reach the required temperature inside the cabinets. The temperature inside the FF (Fresh Food) and the FZ (Freezer) were maintained constant by a PID that regulated the maximum power of a 200 W heater inside each cabinet. The average temperatures inside the FF and the FZ were equal to 4 °C and −18 °C, respectively. The optimal charge was that which maximized the COP. The cooling capacity, considering the presence of the CT-LSHX, was calculated as per the following equation.
Q˙ cool = m˙ re f (hsuct − hcondOUT )
(3.1)
The COP was calculated as per the following equation.
COP =
Q˙ cool W˙ comp
(3.2)
Fig. 2 presents the cooling capacity, calculated from the refrigerant side, the measured compressor power and corresponding COP,
4 Performance comparison: two phases vs. full liquid air 0.55/0.6 2900 From 56.8 to 78.7
Fig. 2. Cooling capacity, COP, and compressor power input for different refrigerant charges with the original condenser design.
and its uncertainty. The compressor power increased almost linearly with the refrigerant charge and in a proportion equal to 14%, from the lowest to the highest refrigerant charge. The cooling capacity increased when the charge increased, but it reached a maximum at around 82 g and, then, began to decrease. There, the COP had a moderate increase with the refrigerant charge, a maximum at approximately 80 g, and, then, a decrease with a significant negative slope. Similar results can be found in literature (Boeng and Melo, 2014; Choi and Kim, 2002). During the entire campaign, visualizations of the condenser outlet and capillary tube inlet confirmed the presence of the two-phase flow. 3.2. Assessment of the subcooling from the heat balance on the water side In experimental campaign 2, the original air condenser was replaced by the water condenser, first, to enable a larger potential for the variation of the condensation conditions and, second, to assess, from the heat balance on the water side of the condenser, the refrigerant enthalpy at the outlet, thereby, at the inlet of the capillary tube. The first part of the experimental campaign consisted of finding the combination of the refrigerant charge and the water flow rate that precisely replicated the operation of the system with the original design at its optimum charge, i.e., the evaporation and condensation temperatures, the refrigerant mass flow rate, the compressor power, and the average temperatures inside the refrigerator compartments. Using the refrigerant-to-water heat exchanger, the refrigerant charge of the system was increased slowly until the operating parameters were very close to those required. Then, a fine-tuning was conducted, varying the water mass flow rate in order to match the targeted operating conditions almost exactly. Table 3 presents the final charge selected and the operating conditions corresponding to two different water flow rates for the water condenser, 0.118 kg min−1 and 0.190 kg min−1 , respectively, which would provide operating conditions almost identical to the original setup with an optimum charge.
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
269
Table 3 Operating conditions with the original and the new water condenser.
Ambient temp. Average FF temp. Average FZ temp. Water mass flow rate Refrigerant charge Condensation temp. Evaporation temp. Ref. mass flow rate Compressor power
°C °C °C kg.min−1 g °C °C g.s−1 W
Water HX test conditions 1
Air condenser target test conditions
Water HX test conditions 2
25 4.3 −18.7 0.190 64 35.74 −20.73 0.246 31.4
25 4.1 −19.3 N.A. 80.9 36.83 −21.15 0.238 31.6
25 4.2 −19.5 0.118 64 37.21 −21.63 0.229 31.1
Fig. 4. Subcooling evaluated from the pressure and temperature measurements (SC_pT), and subcooling evaluated from the heat measured at the condenser water side (SC_HX), during experimental campaign 2. Fig. 3. Condensation capacity evaluated from the condenser water side (Q˙ w ), and condensation capacity (Q˙ cond ) evaluated from the refrigerant side, for different water mass flow rates.
With the new setup, the refrigerant enthalpy at the condenser outlet could be calculated from the heat balance between the water heat capacity (Eq. (3.3)) and the refrigerant condensing capacity (Eq. (3.4)).
Q˙ w = m˙ w c p w (TwOUT − TwIN )
(3.3)
Q˙ w = m˙ w c p w (TwOUT − TwIN ) = Q˙ cond = m˙ re f (hcondIN − hcondOUT _HX ) (3.4) From Eqs. (3.3) and 3.4, it was possible to calculate the enthalpy at the condenser outlet (Eq. (3.5)):
hcondOUT _HX = hcondIN −
m˙ w c p (TwOUT − TwIN ) m˙ re f w
(3.5)
Once the refrigerant charge was determined, a series of tests that varied the water flow rate at the condenser were performed to compare the estimation of the subcooling made from the pressure and temperature measurements at the outlet of the condenser and the subcooling that could be obtained from the refrigerant enthalpy hcondOU THX . Fig. 3 compares the condensation capacity (Q˙ w ), evaluated as the water heat capacity measured on the water side, and the condensation capacity (Q˙ cond ), evaluated from the refrigerant side by using the outlet enthalpy, calculated with the measured pressure and temperature while assuming equilibrium. The dots in the figure also include the associated uncertainties (which were hardly visible because of their low values). The results are represented
versus the water mass flow rate through the condenser, as this was the only parameter that was changed across the experiments. As can be observed, in general, the capacities evaluated from the heat balance from the water side were always lower than the ones evaluated from the pressure and temperature measurements; furthermore, the uncertainty of the measurements could not explain this difference in all cases. Both capacities remained constant for the water mass flow rates from 0.447 kg min−1 to 0.252 kg min−1 . Then, they continually decreased if the water flow rate decreased. When the water mass flow rate decreased, the capacity of the condenser to reject the heat to the water also decreased, and, therefore, the condensation temperature had to increase, further increasing the condensation pressure. At some point, the rise in the pressure ratio led to a decline in the refrigerant mass flow rate. This resulted in the observed decrease of both capacities (Fig. 3). A possible reason for the observed difference between Q˙ cond and Q˙ w could be the assumption of thermodynamic equilibrium at the refrigerant outlet for the evaluation of the condensing capacity (Q˙ cond ) when there is actually no equilibrium, as discovered by Lee et al. (2016) as well as Ko and Jeong (2017). Fig. 4 depicts the corresponding values for the subcooling. The values were evaluated first from the pressure and temperature measurements at the outlet of the condenser while assuming thermodynamic equilibrium for the refrigerant, namely, SCpT . Second, the subcooling was evaluated from the estimated outlet enthalpy, namely, SCHX . As can be seen in Fig. 4, SCpT was almost constant around 9 K, except for the lowest tested water flow rate, where it decreased to approximately 7 K. These subcooling values contradicted the visualization of the flow at the capillary tube inlet, which clearly exhibited two-phase flow conditions or, at least, alternating liquid/vapor inlet conditions during the entire campaign (Fig. 5a).
270
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Fig. 6. Balance point between the compressor mass flow rate and the mass flow rate through the capillary tube.
Fig. 5. Examples of visualization of filter inlet and condenser outlet during experimental campaign 2.
Martinez et al. (2017) already noted this contradiction between the apparent subcooled state and the actual visible conditions of the refrigerant at the capillary tube inlet, with the liquid level always appearing at the capillary inlet mouth. In contrast with the almost constant value of the SCpT subcooling, the SCHX subcooling was much lower and decreased alongside the decrease of the water flow rate until reaching a point where the subcooling ceased. For a water mass flow rate of 0.066 kg min−1 and 0.118 kg min−1 , the evaluated vapor quality was equal to 8.8% and 1.7%, respectively. Note that despite the attention paid to instrumentation selection and calibration, the analysis of the propagation error revealed that, unfortunately, the uncertainty in the SCHX subcooling was high. From these results, and others reported in literature, it is clear that it is difficult to conclude which is the actual state of the refrigerant at the capillary tube inlet. Visual observation continued, confirming the reported alternating vapor–liquid entrance for all performed tests (Fig. 5a). Video recording with a high-speed camera clearly showed the formation of a vortex with small but continuous variations in liquid levels and entering conditions. Despite the wide water mass flow rate variation and, thus, the very different condensing conditions, it was not possible to fill the inlet of the capillary tube with liquid. Additionally, variations in the condensing temperatures were visible at the condenser outlet, which strongly modified the flow patterns as the water mass flow rate changed, resulting in a higher presence of the vapor phase for
lower water mass flow rates and, thus, higher condensation pressure. Fig. 5b shows the flow pattern existing at the outlet of the condenser. The heat balance methodology indicated that, except for the two lowest flow rates, all others indicated a certain low subcooling (around 3 K). Pressure and temperature measurements and the assumption of equilibrium led to much higher subcooling values. Furthermore, the vapor temperature measured above the filter dryer, which acts as a vapor/liquid separator, was always below the saturation temperature and practically at the same temperature as the subcooled liquid. The present study did not find an explanation for the measured refrigerant state since the condensation process happened quite slowly, and, therefore, the vapor bubbles had enough time to condensate along the tube or inside the filter dryer’s small volume or, at least, to keep close to the corresponding saturation temperature. However, all measurements clearly indicated subcooled conditions for both the liquid and vapor, and the visualization of the inlet of the capillary tube revealed the entrance of vapor and liquid. Lee et al. (2016) as well as Ko and Jeong (2017) have observed this paradoxical state and attributed it to non-equilibrium conditions of the flow, where both subcooled gas and subcooled liquid coexist. From these considerations, it can be concluded that refrigerant conditions at the capillary tube inlet may have been in a non-equilibrium two-phase flow composed of subcooled liquid and subcooled vapor. Consequently, the enthalpy at the condenser outlet, thus, cannot be determined using the thermodynamic properties table, which is only valid for equilibrium conditions of the refrigerant. 3.3. Design modifications to achieve full liquid conditions at the capillary tube inlet A possible explanation for the permanent existence of vapor at the capillary tube entrance is that the capillary tube was oversized; therefore, the refrigerant conditions at the inlet of the capillary must be under two-phase flow conditions in order to augment the pressure drop across the capillary tube and to match the compressor flow rate. This potential situation is depicted in Fig. 6, where the compressor flow rate is characterized by a line with a negative slope since it decreases with the pressure ratio’s increase. A line with a positive slope represents the flow rate through the capillary tube, which increases with the pressure ratio. The curves present a qualitative and orientated situation and are not quantitatively representative of the actual operation. The balance point, the intersection of these two lines, represents the stable operation point, at which the refrigerant flow rate through the compressor
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
271
Fig. 7. Examples of visualization of filter inlet and condenser outlet during experimental campaign 3.
matches the one through the capillary tube. The flow rate through the capillary tube with the two-phase flow inlet is marked with a dashed line, whereas the zone inlet with the liquid refrigerant is indicated by a solid line. As depicted in Fig. 6, the balance requires that the flow through the capillary tube is in average in two-phase conditions. This could be the reason why a vortex is observed with a combination of liquid and vapor, or a certain alternating liquid– vapor inlet flow, at the entrance of the capillary tube. This argument would explain the presence of vapor at the capillary tube inlet and would lead to the conclusion that the capillary tube was oversized for this refrigerator. Therefore, a third experimental campaign was performed in which the capillary tube was changed to a smaller size (a nominal 0.55 mm diameter) to increase the restriction, and the compressor speed was varied from 1800 rpm to 3500 rpm to enable the identification of a new balance point at which the inlet of the capillary tube could be full of liquid. During this experimental campaign, the tests were run only with the freezer (i.e., the damper between the refrigerator and freezer was kept closed) because the purpose
was not to analyze a commercial refrigerator anymore but to focus on the possibility of filling the capillary tube with only liquid. To regulate the temperature in the freezer and to ensure the same temperature conditions for the entire test campaign, the previous PID and heater set-up was modified. The thermocouple used by the PID was placed at the evaporator air inlet. This type of control insured that the conditions of the secondary fluid at the evaporator inlet were always the same for each operating parameter tested despite the variation of the cooling capacity produced by the variation of the compressor speed. The air temperature at the evaporator inlet Tair_EvapIN of a household refrigerator is composed of a mix of air coming from the FF compartment (TFF ) and air coming from the FZ compartment (TFZ ). Eq. (4.1) enables the evaluation of the corresponding mixing temperature:
Tair_E vapIN = α TF F + (1 − α )TF Z ,
(4.1)
where α represents the air flow ratio supplied to the FF cabinet. This air flow ratio is usually between 0% and 30%. In this third
272
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Fig. 8. Subcooling evaluated from the pressure and temperature measurements (SC_pT), and subcooling evaluated from the heat measured at the condenser water side (SC_HX), during experimental campaign 3.
experimental campaign, Tair_E vapIN was fixed at −13.5 °C, which corresponded to an airflow ratio of 20%. As expected, tests with the smaller diameter capillary tube at 250 0 and 350 0 rpm, respectively, led to a filter full of liquid, as depicted in Fig. 7a. All transparent parts of the filter were full of liquid, and these conditions were maintained through the entirety of these tests. A test at 1800 rpm resulted in an alternation between two different liquid level situations at the capillary tube inlet: a vortex with a two-phase inlet (Fig. 7b) and a liquid level about 5 mm above the capillary tube inlet (Fig. 7c) without a vapor inlet. The visualization of the condenser outlet (Fig. 7d) allowed the observation of the void fraction at the outlet of the condenser for several compressor speeds. As it can be clearly observed in the picture, the higher the compressor speed, the lower the void fraction. Fig. 8 shows the results of the evaluation of the subcooling by the two employed methods for the 0.55 mm diameter capillary tube. As it can be observed, both methods indicated subcooling conditions at 1800 rpm when the visualization showed the twophase flow inlet, as revealed by experimental campaign 2. Interestingly, the subcooling obtained from the pressure-temperature measurements, SCpT , was still higher than the one obtained from the heat balance, SCHX , although the differences were smaller. Furthermore, when full liquid conditions were clearly visible (higher compressor speeds), the difference began to fall inside the uncertainty band. 3.4. Performance with the modified design Once it had been proven that it was possible to have full liquid conditions at the inlet of the capillary tube, the question arose as to whether that led to a better COP or not. The objective of experimental campaign 4 was to determine whether there was a performance difference between the original design of the capillary tube with a 0.6 mm diameter and the modified design with a slightly smaller diameter. For this comparison, we returned to the configuration of test campaign 1, i.e., the refrigerator using the original airto-refrigerant condenser. The comparison had to be performed at the same compressor speed to compare similar capacities. For this comparison, we selected a compressor speed of 2900 rpm, which was high enough to lead to full liquid inlet conditions at the inlet of the 0.55 mm diameter capillary tube. With both designs having differences (i.e., different capillary tubes), the COP values needed to be determined at their corresponding optimal charge in order to draw a fair comparison.
Fig. 9. COP and liquid level inside the inlet filter for different refrigerant charges for the two tested capillary tubes, of 0.55 mm and 0.6 mm diameter respectively.
Therefore, an optimization charge study was performed for each capillary tube. Fig. 9 shows the COP results obtained. As it can be observed, all COP values attained with the 0.55 mm diameter capillary tube were higher than those achieved with the 0.6 mm tube. The average COP difference between the two capillary tubes was approximately 4%, which is an important performance improvement in this kind of equipment. This indicates that the original capillary tube was slightly oversized and that better performance can be achieved with the adequate design of the capillary tube. Regarding the refrigerant visualizations at the capillary tube inlet, the different liquid levels for the different tested refrigerant charges are depicted in Fig. 9. As expected, for the 0.6 mm capillary tube, the inlet always had a two-phase flow, as represented in the bottom frame. For the 0.55 mm capillary tube, different liquid levels were obtained through the refrigerant charge test campaign, which are represented in the top frames. At 67.5 g, the filter was full of liquid, and the highest COP was found at exactly this refrigerant charge. For the remaining tested refrigerant charges, the inlet alternated between the two-phase flow and liquid, with a mainly liquid inlet. 4. Conclusions This work aimed, first, to assess the actual conditions present at the capillary tube inlet of a household refrigerator system and, second, to find out if a different design of the capillary tube could lead to full liquid conditions at the tube inlet as well as if this would lead to better performance. The main conclusions drawn from this study can be summarized as follows: •
•
•
With the original capillary tube, the inlet to the capillary tube always had a two-phase flow regardless of the refrigerant charge and the operating conditions. The liquid level was always very close to the capillary tube’s mouth, and a vortex was always present, indicating the entrance of vapor and liquid. At the same time, the subcooling, estimated from the measured pressure and temperature at the outlet of the condensers and the assumption of thermodynamic equilibrium, was significantly high, e.g., around 9 K.
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273 •
•
•
•
In contrast, the estimation of the subcooling from the evaluation of the refrigerant enthalpy at the outlet of the condenser led to lower values, even indicating a certain amount of vapor quality. These results agree with recent results published by other authors who postulate the lack of thermodynamic equilibrium at the outlet of the condenser. The present results seem to indicate the coexistence of subcooled liquid and subcooled vapor at the inlet of the capillary tube. It has been found that the presence of vapor at the inlet of the capillary tube could be forced by the matching of the flow rates between the compressor and the capillary tube, i.e., as the capillary tube is oversized with respect to the compressor, the inlet flow must contain some vapor in order to increase the pressure drop and become more restrictive. Considering this, an additional test campaign was carried out with a smaller capillary tube (0.55 mm instead of the original 0.6 mm) and by varying the compressor speed in order to check if it was then possible to fill the capillary tube inlet fully with liquid. It was found that with the smaller diameter, it was possible to have full liquid conditions for speeds above 2500 rpm. Finally, a refrigerant charge optimization was carried out with the original refrigerator, the two different capillary tubes (0.6 and 0.55 mm, respectively), and a higher compressor speed (2900 rpm) in order to be sure that full liquid conditions could be achieved at the inlet of the smaller capillary tube. This final experimental campaign showed that the COP with the smaller diameter was always approximately 4% higher and that the maximum COP was precisely achieved when full liquid conditions were clearly observable at the inlet of the tube.
273
Acknowledgements In this project, the work of Laetitia Bardoulet was partially supported by the Santiago Grisolía 2015 program, which is funded by the Generalitat Valenciana, with the reference number GRISOLIA/2015/021. References Boeng, J., Melo, C., 2014. Mapping the energy consumption of household refrigerators by varying the refrigerant charge and the expansion restriction. Int. J. Refrig. 41, 37–44. https://doi.org/10.1016/j.ijrefrig.2013.06.005. Choi, J.M., Kim, Y.C., 2002. The effects of improper refrigerant charge on the performance of a heat pump with an electronic expansion valve and capillary tube. Energy 27, 391–404. https://doi.org/10.1016/S0360-5442(01)0 0 093-7. Hartmann, D., Melo, C., 2013. Popping noise in household refrigerators: fundamentals and practical solutions. Appl. Therm. Eng. 51, 40–47. https://doi.org/10.1016/ j.applthermaleng.2012.08.054. Hermes, C.J.L., Melo, C., Gonçalves, J.M., 2008. Modeling of non-adiabatic capillary tube flows: a simplified approach and comprehensive experimental validation. Int. J. Refrig. 31, 1358–1367. https://doi.org/10.1016/j.ijrefrig.20 08.04.0 02. Ko, J., Jeong, J.H., 2017. Effects of a non-equilibrium two-phase refrigerant flow at subcooled temperatures on the performance of an R-600a refrigeration system. Int. J. Refrig.. https://doi.org/10.1016/j.ijrefrig.2017.10.029. Lee, W.-J., Seo, J.-Y., Ko, J., Jeong, J.H., 2016. Non-equilibrium two-phase refrigerant flow at subcooled temperatures in an R600a refrigeration system. Int. J. Refrig. 70, 148–156. https://doi.org/10.1016/j.ijrefrig.2016.07.005. Liu, Y., Bullard, C.W., 1997. An Experimental and Theoretical Analysis of Capillary Tube-Suction Line Heat Exchangers. University of Illinois Project reference: ACRC TR-109. Martínez-Ballester, S., Bardoulet, L., Pisano, A., Corberán, J.M., 2017. Visualization of refrigerant flow at the capillary tube inlet of a high-efficiency household refrigerator. Int. J. Refrig. 73, 200–208. https://doi.org/10.1016/j.ijrefrig.2016.09.019.
Contents lists available at ScienceDirect
International Journal of Refrigeration journal homepage: www.elsevier.com/locate/ijrefrig
How to achieve full liquid conditions at the capillary tube inlet of a household refrigerator Laetitia Bardoulet a,∗, José M. Corberán a, Santiago Martínez-Ballester b a b
Institute for Energy Engineering, Universitat Politècnica de València, Camino de vera, s/n, Valencia 46022, Spain Ingersol Rand, Sant Feliu de Llobregat 08980, Spain
a r t i c l e
i n f o
Article history: Received 18 July 2018 Revised 2 January 2019 Accepted 4 February 2019 Available online 7 February 2019 Keywords: Household refrigerator Capillary tube inlet Refrigerant visualization Non-equilibrium Subcooled vapor COP
a b s t r a c t The capillary tube with a liquid-to-suction heat exchanger (CT-LSHX) is a component that is widely used in household refrigerators. Recent works have indicated that even when measuring subcooled conditions at the condenser outlet, the condition at the capillary tube inlet is a two-phase flow. The present work was dedicated to analyzing the actual refrigerant conditions at the capillary tube inlet and to investigating how full liquid conditions could be achieved. The research was performed using a typical household refrigerator with corresponding fresh food and freezer compartments, replacing the original refrigerant-to-air condenser with a refrigerant-to-water condenser. This allowed, first, the condensation conditions to be controlled and, second, the estimation of the refrigerant conditions at the condenser outlet from the heat exchanger balance. The obtained results indicated the presence of a non-equilibrium two-phase flow, composed of subcooled vapor and subcooled liquid, at the capillary tube inlet, with both liquid and vapor entering the capillary tube as a vortex with small, fast fluctuations of the liquid level. This non-equilibrium indicated that the subcooling, evaluated from the pressure and temperature of the refrigerant at the condenser outlet, was only apparent and did not allow the evaluation of the actual enthalpy. Finally, by using a smaller capillary tube diameter and increasing the compressor speed, full liquid conditions at the capillary tube inlet were achieved. Furthermore, a performance comparison between the original and the new design revealed that the COP was higher with full liquid conditions. © 2019 Elsevier Ltd and IIR. All rights reserved.
Comment atteindre des conditions de plein liquide à l’entrée du tube capillaire d’un réfrigérateur domestique Mots-clés: Réfrigérateur domestique; Ouïe du tube capillaire; Visualisation du frigorigène; Vapeur sous-refroidie; COP
1. Introduction In household refrigerators, a capillary tube with a liquidto-suction heat exchanger (CT-LSHX) is the most widespread expansion device used due to its simplicity and low cost. Liquid conditions at the capillary tube inlet are necessary to avoid negative consequences, such as noise (Hartmann and Melo, 2013) or hysteresis effects as well as the reduction of CT-LSHX effectiveness (Liu and Bullard, 1997). Another consequence of the absence of ∗
Corresponding author. E-mail address: [email protected] (L. Bardoulet).
https://doi.org/10.1016/j.ijrefrig.2019.02.006 0140-7007/© 2019 Elsevier Ltd and IIR. All rights reserved.
liquid at the inlet of the capillary tube is a lower cooling capacity due to a certain decrease in the refrigerant mass flow rate, which affects its global efficiency. Subcooled conditions at the capillary tube inlet are always reported in literature. Nevertheless, some recent works have concluded that the subcooled conditions at the capillary tube inlet may not be real, which, if true, would have consequences for efficiency and, more generally, energy consumption. Boeng and Melo (2014) performed an experimental study using a refrigerator-freezer equipped with R600a to identify the optimal combination of the refrigerant charge and the capillary tube by varying the expansion device’s capacity. These researchers then
266
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Nomenclature cp h m˙ P Q˙ SC T W˙
Specific heat capacity, [kJ.kg−1 .K−1 ] Specific enthalpy, [kJ.kg−1 ] Mass flow rate, [kg.s−1 ] Pressure, [bar] Rate of heat transfer, [W] Subcooling, [K] Temperature, [°C] Rate of work, power [W]
Greek symbols α Air flow ratio, [-] τ Time ratio, [-] Subscripts cap Capillary tube comp Compressor cool Cooling evap Evaporator p Pressure [bar] in Inlet out Outlet ref Refrigerant w Water Acronyms and abbreviations CT Capillary tube COP Coefficient of performance FF Fresh food compartment FZ Freezer compartment HX Heat exchanger LSHX Liquid-to-suction heat exchanger PFA Perfluoroalkoxy
complemented this work by applying a model (Hermes et al., 2008) to examine the correlations between the degree of the valve opening and the equivalent capillary diameter. This study revealed a good prediction of the mass flow rate, except for the tests with fully open valves or a subcooling of less than 5 K. Through visualization of the capillary tube inlet, the authors concluded that even with a subcooling of 5 K, a two-phase flow still occurred at the inlet. As a possible explanation, Boeng and Melo highlighted that at the capillary inlet, there was a non-equilibrium mixture of subcooled liquid and saturated vapor instead of a purely liquid phase. The findings also revealed that the vapor quality at the capillary tube inlet ranged from 2% to 12%, depending on the refrigerant charge. Additionally, Martinez-Ballester et al. (2017), through the visualization of the condenser outlet and capillary tube inlet of a household refrigerator filled with R600a, concluded that the capillary tube inlet possessed two-phase flow conditions although a 14 K subcooling was measured. Furthermore, Lee et al. (2016) visualized a two-phase refrigerant flow at the inlet of the capillary tube of a commercial refrigeration system as well, working with isobutane. These researchers set up a transparent tube at the condenser outlet and an oil separator at the compressor discharge; then, they measured both the vapor and liquid temperatures as the condensation pressure. This study revealed that the refrigerant was in a non-equilibrium state, where both the subcooled liquid and the subcooled vapor coexisted at the capillary tube inlet. The authors subsequently proposed a method to calculate the enthalpy for non-equilibrium refrigerant conditions. They were able to estimate the vapor quality for the
different condensing pressures, ranging from 2.44% to 4.29%, at the condenser outlet. Ko and Jeong (2017) complemented the previous work by studying the effects of a non-equilibrium subcooled two-phase flow on the performance of a vapor compression refrigeration system. The authors estimated that the COP may be overestimated by 19.4% if the thermodynamic properties table is used for a nonequilibrium refrigerant state. Thus, various accounts are provided to explain the presence of the two-phase flow at the capillary tube inlet. As stated by Lee et al. (2016), an explanation for the presence of the vapor, although a certain subcooling is measured, is the non-equilibrium state of the refrigerant, composed of the subcooled liquid and subcooled vapor. In comparison, Boeng and Melo (2014) proposed non-equilibrium between the saturated vapor and the subcooled liquid. In their study, Martinez-Ballester et al. (2017) set up a transparent tube of 1.5 m at the condenser outlet. The authors stated that the transparent section was relatively long and could be considered adiabatic compared to a normal tube. Furthermore, the refrigerant had enough length to achieve an equilibrium state or, at least, to reveal some evolution towards equilibrium. This study provided a hypothesis for the presence of vapor at the capillary tube inlet: the capillary tube could be oversized and, therefore, to match the compressor flow rate with the flow rate through the capillary tube, the inlet should be in two-phase flow conditions to produce a higher pressure drop. The main objectives of the present work were, first, to assess the actual conditions at the capillary tube inlet of a household refrigerator system and, second, to find a way to achieve full liquid conditions at the capillary tube inlet. In order to reach these goals, four different experimental campaigns were performed by means of an innovative test bench and the use of a household refrigerator. First, the operating conditions were determined with the original condenser and capillary tube for an optimal charge of the refrigerant. Second, these operating conditions were reproduced with a refrigerant-to-water condenser to enable the measurement of the heat produced during the condensation and, thus, to determine the vapor quality or the subcooling degree at the capillary tube inlet. Third, a smaller capillary tube was tested with different compressor speeds to fill the capillary tube inlet with liquid. Fourth and finally, a performance analysis comparison between both the original and the modified design was performed. 2. Experimental set-up 2.1. Test bench The original apparatus was a high-efficiency, two-door refrigerator-freezer from the A+++ energy class, with no frost, and operating with R600a. It contained a fin and tube evaporator, a tube and wire condenser, a variable-speed 7 cm3 hermetic reciprocating compressor, and a capillary tube with a liquid-tosuction heat exchanger as the expansion device. The original refrigerator was modified in order to test different configurations of condensers and capillary tubes (Fig. 1). A set of valves allowed operating with either the original refrigerant-to-air condenser or a refrigerant-to-water condenser, composed of twin copper tubes with a 3.16 mm diameter welded onto a 2.53 m length and carefully insulated. In order to visualize the flow, a PFA (Perfluoroalkoxy) transparent tube section was inserted at the outlet of each condenser configuration. In the case of the refrigerant-to-air condenser, the final 1.5 m of the original condenser was cut off and replaced with the same length of transparent tube. An extra transparent tube section of 0.5 m was added to the refrigerant-to-water condenser.
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
267
Fig. 1. Test bench.
Two different capillary tubes were set up in the same suction line: the original one, with a 0.6 mm diameter, and a smaller one, with a 0.55 mm diameter. They were both laterally configured, i.e., both capillary tubes were welded to the suction line. Each capillary tube was mounted with a vertical transparent filter of PFA, filled with zeolite balls (Fig. 1), which allowed the visualization of the entrance of the capillary tube. A set of valves allowed operating with one capillary tube or another. Fig. 1 presents a drawing of the test bench, depicting the installed instrumentation: thermocouples, RTDs, mass flow meters, and pressure transducers. The refrigerant loop was equipped with a Coriolis mass flow meter (MFM_ref), located at the compressor discharge, that measured the refrigerant mass flow rate. The temperature of the refrigerant at the inlet of the refrigerant-to-water condenser (TcondIN) and the refrigerant inlet temperature of each capillary tube were measured using RTDs (TcapIN). The water loop was powered by a diaphragm pump and a by-pass system that allowed the modification of the water mass flow rate, which was measured using a Coriolis mass flow meter (MFM_w). Furthermore, RTDs were placed at both the inlet and the outlet of the condenser’s waterside in order to measure the inlet (TwIN) and outlet (TwOUT) water temperatures. The rest of the instrumentation was shared by both condensers’ configurations. A thermowell, with a thermocouple inside, measured the temperature at the outlet of the condenser (TcondOUT), and a pressure sensor allowed the measurement of the condensation pressure (Pcond). The evaporation pressure was measured by a pressure transducer (Pev), and the compressor power input was determined with a Wattmeter. The actual uncertainties of the different measurements are listed in Table 1. These values were used for the assessment of the final uncertainties of the different results: COP, heating and cooling capacities and subcooling, by the uncertainty propagation method.
Table 1 Employed sensors and corresponding uncertainty. Temperatures
Uncertainty
Water inlet temperature: RTD Water outlet temperature: RTD Refrigerant inlet condenser temperature: RTD Refrigerant outlet condenser temperature: thermocouple Capillary tube inlet: RTD Pressures Condensation pressure: pressure transducer Evaporation pressure: pressure transducer Mass flow rates Water mass flow rate: Coriolis mass flow meter Refrigerant mass flow rate: Coriolis mass flow meter Power Compressor power: Wattmeter
±0.07 K ±0.07 K ±0.078 K ±0.2 K ±0.07 K ±0.04 bar ±0.078 bar ±0.1% of rate ±0.25% of rate ±1.2 W
2.2. Experimental test campaigns Table 2 describes the main parameters of the four experimental campaigns that were performed. The experiments were carried out in a climatic chamber maintained at 25 °C. As the best manner to analyze and understand a refrigerant cycle, steady tests were conducted. In order to assess the conditions at the capillary tube inlet, experimental campaign 1 dealt with the determination of the operating conditions of the apparatus, with both the original condenser and the capillary tube, for the optimal charge of the refrigerant, the one that maximized the COP. Experimental campaign 2 dealt with the use of the water condenser to enable the accurate evaluation of the enthalpy of the refrigerant at the outlet of the condenser, thereby, at the capillary tube inlet conditions. However, first of all, the refrigerant charge had to be determined in order to achieve the same operating
268
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Table 2 Performed experimental campaigns. Experimental campaign Aim
1 Determining the optimal refrigerant charge
2 Assessing the capillary tube inlet conditions
3 Getting full liquid conditions at the capillary tube inlet
Condenser Capillary tube diameter (mm) Compressor speed (rpm) Refrigerant charge (g)
air 0.6 1600 From 74.5 to 90
water 0.6 1600 64
water 0.55 180 0–250 0–350 0 64
conditions that were attained with the original condenser. Therefore, we first performed a series of tests, varying the charge and the water flow conditions in order to obtain exactly the same operating conditions for the refrigerant. Then, the conditions at the inlet of the capillary tube were determined and analyzed. Afterwards, the operating conditions were varied by adjusting the water flow rate in order to check the influence of the condensing conditions on the refrigerant conditions at the inlet of the capillary tube. Furthermore, this enabled the comparison of the differences between the apparent subcooling estimated from the pressure and temperature measurements and the subcooling calculated from the enthalpy of the refrigerant at the outlet of the condenser. The aim of experimental campaign 3 was to search for other possible balance points in the matching between the capillary tube design (diameter) and the compressor speed at which full liquid conditions could be permanently maintained at the inlet of the capillary tube. To do so, a capillary tube with a 0.55 mm diameter, instead of the original 0.6 mm tube, along with different compressor speeds were tested. It should be pointed out that full liquid conditions could not be obtained with the original capillary tube at the maximum compressor speed. To assess whether the performance obtained by the improved design determined in experimental campaign 3 (a higher compressor speed and a smaller capillary tube) was better than the original design, a fourth experimental campaign was carried out using two capillary tubes of different diameters (0.55 mm and 0.6 mm) at the same compressor speed. In order to obtain a fair comparison, an optimal refrigerant charge study was performed for each capillary tube size at the same compressor speed (2900 rpm). 3. Results and discussion 3.1. The optimal charge of the refrigerator with the original condenser In order to determine the optimal refrigerant charge, a charge variation study was performed. The charge was varied from 74.5 g to 90 g. Results at 90 g were not included in the data analysis because the suction became frozen, and, therefore, it was not possible to reach the required temperature inside the cabinets. The temperature inside the FF (Fresh Food) and the FZ (Freezer) were maintained constant by a PID that regulated the maximum power of a 200 W heater inside each cabinet. The average temperatures inside the FF and the FZ were equal to 4 °C and −18 °C, respectively. The optimal charge was that which maximized the COP. The cooling capacity, considering the presence of the CT-LSHX, was calculated as per the following equation.
Q˙ cool = m˙ re f (hsuct − hcondOUT )
(3.1)
The COP was calculated as per the following equation.
COP =
Q˙ cool W˙ comp
(3.2)
Fig. 2 presents the cooling capacity, calculated from the refrigerant side, the measured compressor power and corresponding COP,
4 Performance comparison: two phases vs. full liquid air 0.55/0.6 2900 From 56.8 to 78.7
Fig. 2. Cooling capacity, COP, and compressor power input for different refrigerant charges with the original condenser design.
and its uncertainty. The compressor power increased almost linearly with the refrigerant charge and in a proportion equal to 14%, from the lowest to the highest refrigerant charge. The cooling capacity increased when the charge increased, but it reached a maximum at around 82 g and, then, began to decrease. There, the COP had a moderate increase with the refrigerant charge, a maximum at approximately 80 g, and, then, a decrease with a significant negative slope. Similar results can be found in literature (Boeng and Melo, 2014; Choi and Kim, 2002). During the entire campaign, visualizations of the condenser outlet and capillary tube inlet confirmed the presence of the two-phase flow. 3.2. Assessment of the subcooling from the heat balance on the water side In experimental campaign 2, the original air condenser was replaced by the water condenser, first, to enable a larger potential for the variation of the condensation conditions and, second, to assess, from the heat balance on the water side of the condenser, the refrigerant enthalpy at the outlet, thereby, at the inlet of the capillary tube. The first part of the experimental campaign consisted of finding the combination of the refrigerant charge and the water flow rate that precisely replicated the operation of the system with the original design at its optimum charge, i.e., the evaporation and condensation temperatures, the refrigerant mass flow rate, the compressor power, and the average temperatures inside the refrigerator compartments. Using the refrigerant-to-water heat exchanger, the refrigerant charge of the system was increased slowly until the operating parameters were very close to those required. Then, a fine-tuning was conducted, varying the water mass flow rate in order to match the targeted operating conditions almost exactly. Table 3 presents the final charge selected and the operating conditions corresponding to two different water flow rates for the water condenser, 0.118 kg min−1 and 0.190 kg min−1 , respectively, which would provide operating conditions almost identical to the original setup with an optimum charge.
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
269
Table 3 Operating conditions with the original and the new water condenser.
Ambient temp. Average FF temp. Average FZ temp. Water mass flow rate Refrigerant charge Condensation temp. Evaporation temp. Ref. mass flow rate Compressor power
°C °C °C kg.min−1 g °C °C g.s−1 W
Water HX test conditions 1
Air condenser target test conditions
Water HX test conditions 2
25 4.3 −18.7 0.190 64 35.74 −20.73 0.246 31.4
25 4.1 −19.3 N.A. 80.9 36.83 −21.15 0.238 31.6
25 4.2 −19.5 0.118 64 37.21 −21.63 0.229 31.1
Fig. 4. Subcooling evaluated from the pressure and temperature measurements (SC_pT), and subcooling evaluated from the heat measured at the condenser water side (SC_HX), during experimental campaign 2. Fig. 3. Condensation capacity evaluated from the condenser water side (Q˙ w ), and condensation capacity (Q˙ cond ) evaluated from the refrigerant side, for different water mass flow rates.
With the new setup, the refrigerant enthalpy at the condenser outlet could be calculated from the heat balance between the water heat capacity (Eq. (3.3)) and the refrigerant condensing capacity (Eq. (3.4)).
Q˙ w = m˙ w c p w (TwOUT − TwIN )
(3.3)
Q˙ w = m˙ w c p w (TwOUT − TwIN ) = Q˙ cond = m˙ re f (hcondIN − hcondOUT _HX ) (3.4) From Eqs. (3.3) and 3.4, it was possible to calculate the enthalpy at the condenser outlet (Eq. (3.5)):
hcondOUT _HX = hcondIN −
m˙ w c p (TwOUT − TwIN ) m˙ re f w
(3.5)
Once the refrigerant charge was determined, a series of tests that varied the water flow rate at the condenser were performed to compare the estimation of the subcooling made from the pressure and temperature measurements at the outlet of the condenser and the subcooling that could be obtained from the refrigerant enthalpy hcondOU THX . Fig. 3 compares the condensation capacity (Q˙ w ), evaluated as the water heat capacity measured on the water side, and the condensation capacity (Q˙ cond ), evaluated from the refrigerant side by using the outlet enthalpy, calculated with the measured pressure and temperature while assuming equilibrium. The dots in the figure also include the associated uncertainties (which were hardly visible because of their low values). The results are represented
versus the water mass flow rate through the condenser, as this was the only parameter that was changed across the experiments. As can be observed, in general, the capacities evaluated from the heat balance from the water side were always lower than the ones evaluated from the pressure and temperature measurements; furthermore, the uncertainty of the measurements could not explain this difference in all cases. Both capacities remained constant for the water mass flow rates from 0.447 kg min−1 to 0.252 kg min−1 . Then, they continually decreased if the water flow rate decreased. When the water mass flow rate decreased, the capacity of the condenser to reject the heat to the water also decreased, and, therefore, the condensation temperature had to increase, further increasing the condensation pressure. At some point, the rise in the pressure ratio led to a decline in the refrigerant mass flow rate. This resulted in the observed decrease of both capacities (Fig. 3). A possible reason for the observed difference between Q˙ cond and Q˙ w could be the assumption of thermodynamic equilibrium at the refrigerant outlet for the evaluation of the condensing capacity (Q˙ cond ) when there is actually no equilibrium, as discovered by Lee et al. (2016) as well as Ko and Jeong (2017). Fig. 4 depicts the corresponding values for the subcooling. The values were evaluated first from the pressure and temperature measurements at the outlet of the condenser while assuming thermodynamic equilibrium for the refrigerant, namely, SCpT . Second, the subcooling was evaluated from the estimated outlet enthalpy, namely, SCHX . As can be seen in Fig. 4, SCpT was almost constant around 9 K, except for the lowest tested water flow rate, where it decreased to approximately 7 K. These subcooling values contradicted the visualization of the flow at the capillary tube inlet, which clearly exhibited two-phase flow conditions or, at least, alternating liquid/vapor inlet conditions during the entire campaign (Fig. 5a).
270
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Fig. 6. Balance point between the compressor mass flow rate and the mass flow rate through the capillary tube.
Fig. 5. Examples of visualization of filter inlet and condenser outlet during experimental campaign 2.
Martinez et al. (2017) already noted this contradiction between the apparent subcooled state and the actual visible conditions of the refrigerant at the capillary tube inlet, with the liquid level always appearing at the capillary inlet mouth. In contrast with the almost constant value of the SCpT subcooling, the SCHX subcooling was much lower and decreased alongside the decrease of the water flow rate until reaching a point where the subcooling ceased. For a water mass flow rate of 0.066 kg min−1 and 0.118 kg min−1 , the evaluated vapor quality was equal to 8.8% and 1.7%, respectively. Note that despite the attention paid to instrumentation selection and calibration, the analysis of the propagation error revealed that, unfortunately, the uncertainty in the SCHX subcooling was high. From these results, and others reported in literature, it is clear that it is difficult to conclude which is the actual state of the refrigerant at the capillary tube inlet. Visual observation continued, confirming the reported alternating vapor–liquid entrance for all performed tests (Fig. 5a). Video recording with a high-speed camera clearly showed the formation of a vortex with small but continuous variations in liquid levels and entering conditions. Despite the wide water mass flow rate variation and, thus, the very different condensing conditions, it was not possible to fill the inlet of the capillary tube with liquid. Additionally, variations in the condensing temperatures were visible at the condenser outlet, which strongly modified the flow patterns as the water mass flow rate changed, resulting in a higher presence of the vapor phase for
lower water mass flow rates and, thus, higher condensation pressure. Fig. 5b shows the flow pattern existing at the outlet of the condenser. The heat balance methodology indicated that, except for the two lowest flow rates, all others indicated a certain low subcooling (around 3 K). Pressure and temperature measurements and the assumption of equilibrium led to much higher subcooling values. Furthermore, the vapor temperature measured above the filter dryer, which acts as a vapor/liquid separator, was always below the saturation temperature and practically at the same temperature as the subcooled liquid. The present study did not find an explanation for the measured refrigerant state since the condensation process happened quite slowly, and, therefore, the vapor bubbles had enough time to condensate along the tube or inside the filter dryer’s small volume or, at least, to keep close to the corresponding saturation temperature. However, all measurements clearly indicated subcooled conditions for both the liquid and vapor, and the visualization of the inlet of the capillary tube revealed the entrance of vapor and liquid. Lee et al. (2016) as well as Ko and Jeong (2017) have observed this paradoxical state and attributed it to non-equilibrium conditions of the flow, where both subcooled gas and subcooled liquid coexist. From these considerations, it can be concluded that refrigerant conditions at the capillary tube inlet may have been in a non-equilibrium two-phase flow composed of subcooled liquid and subcooled vapor. Consequently, the enthalpy at the condenser outlet, thus, cannot be determined using the thermodynamic properties table, which is only valid for equilibrium conditions of the refrigerant. 3.3. Design modifications to achieve full liquid conditions at the capillary tube inlet A possible explanation for the permanent existence of vapor at the capillary tube entrance is that the capillary tube was oversized; therefore, the refrigerant conditions at the inlet of the capillary must be under two-phase flow conditions in order to augment the pressure drop across the capillary tube and to match the compressor flow rate. This potential situation is depicted in Fig. 6, where the compressor flow rate is characterized by a line with a negative slope since it decreases with the pressure ratio’s increase. A line with a positive slope represents the flow rate through the capillary tube, which increases with the pressure ratio. The curves present a qualitative and orientated situation and are not quantitatively representative of the actual operation. The balance point, the intersection of these two lines, represents the stable operation point, at which the refrigerant flow rate through the compressor
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
271
Fig. 7. Examples of visualization of filter inlet and condenser outlet during experimental campaign 3.
matches the one through the capillary tube. The flow rate through the capillary tube with the two-phase flow inlet is marked with a dashed line, whereas the zone inlet with the liquid refrigerant is indicated by a solid line. As depicted in Fig. 6, the balance requires that the flow through the capillary tube is in average in two-phase conditions. This could be the reason why a vortex is observed with a combination of liquid and vapor, or a certain alternating liquid– vapor inlet flow, at the entrance of the capillary tube. This argument would explain the presence of vapor at the capillary tube inlet and would lead to the conclusion that the capillary tube was oversized for this refrigerator. Therefore, a third experimental campaign was performed in which the capillary tube was changed to a smaller size (a nominal 0.55 mm diameter) to increase the restriction, and the compressor speed was varied from 1800 rpm to 3500 rpm to enable the identification of a new balance point at which the inlet of the capillary tube could be full of liquid. During this experimental campaign, the tests were run only with the freezer (i.e., the damper between the refrigerator and freezer was kept closed) because the purpose
was not to analyze a commercial refrigerator anymore but to focus on the possibility of filling the capillary tube with only liquid. To regulate the temperature in the freezer and to ensure the same temperature conditions for the entire test campaign, the previous PID and heater set-up was modified. The thermocouple used by the PID was placed at the evaporator air inlet. This type of control insured that the conditions of the secondary fluid at the evaporator inlet were always the same for each operating parameter tested despite the variation of the cooling capacity produced by the variation of the compressor speed. The air temperature at the evaporator inlet Tair_EvapIN of a household refrigerator is composed of a mix of air coming from the FF compartment (TFF ) and air coming from the FZ compartment (TFZ ). Eq. (4.1) enables the evaluation of the corresponding mixing temperature:
Tair_E vapIN = α TF F + (1 − α )TF Z ,
(4.1)
where α represents the air flow ratio supplied to the FF cabinet. This air flow ratio is usually between 0% and 30%. In this third
272
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273
Fig. 8. Subcooling evaluated from the pressure and temperature measurements (SC_pT), and subcooling evaluated from the heat measured at the condenser water side (SC_HX), during experimental campaign 3.
experimental campaign, Tair_E vapIN was fixed at −13.5 °C, which corresponded to an airflow ratio of 20%. As expected, tests with the smaller diameter capillary tube at 250 0 and 350 0 rpm, respectively, led to a filter full of liquid, as depicted in Fig. 7a. All transparent parts of the filter were full of liquid, and these conditions were maintained through the entirety of these tests. A test at 1800 rpm resulted in an alternation between two different liquid level situations at the capillary tube inlet: a vortex with a two-phase inlet (Fig. 7b) and a liquid level about 5 mm above the capillary tube inlet (Fig. 7c) without a vapor inlet. The visualization of the condenser outlet (Fig. 7d) allowed the observation of the void fraction at the outlet of the condenser for several compressor speeds. As it can be clearly observed in the picture, the higher the compressor speed, the lower the void fraction. Fig. 8 shows the results of the evaluation of the subcooling by the two employed methods for the 0.55 mm diameter capillary tube. As it can be observed, both methods indicated subcooling conditions at 1800 rpm when the visualization showed the twophase flow inlet, as revealed by experimental campaign 2. Interestingly, the subcooling obtained from the pressure-temperature measurements, SCpT , was still higher than the one obtained from the heat balance, SCHX , although the differences were smaller. Furthermore, when full liquid conditions were clearly visible (higher compressor speeds), the difference began to fall inside the uncertainty band. 3.4. Performance with the modified design Once it had been proven that it was possible to have full liquid conditions at the inlet of the capillary tube, the question arose as to whether that led to a better COP or not. The objective of experimental campaign 4 was to determine whether there was a performance difference between the original design of the capillary tube with a 0.6 mm diameter and the modified design with a slightly smaller diameter. For this comparison, we returned to the configuration of test campaign 1, i.e., the refrigerator using the original airto-refrigerant condenser. The comparison had to be performed at the same compressor speed to compare similar capacities. For this comparison, we selected a compressor speed of 2900 rpm, which was high enough to lead to full liquid inlet conditions at the inlet of the 0.55 mm diameter capillary tube. With both designs having differences (i.e., different capillary tubes), the COP values needed to be determined at their corresponding optimal charge in order to draw a fair comparison.
Fig. 9. COP and liquid level inside the inlet filter for different refrigerant charges for the two tested capillary tubes, of 0.55 mm and 0.6 mm diameter respectively.
Therefore, an optimization charge study was performed for each capillary tube. Fig. 9 shows the COP results obtained. As it can be observed, all COP values attained with the 0.55 mm diameter capillary tube were higher than those achieved with the 0.6 mm tube. The average COP difference between the two capillary tubes was approximately 4%, which is an important performance improvement in this kind of equipment. This indicates that the original capillary tube was slightly oversized and that better performance can be achieved with the adequate design of the capillary tube. Regarding the refrigerant visualizations at the capillary tube inlet, the different liquid levels for the different tested refrigerant charges are depicted in Fig. 9. As expected, for the 0.6 mm capillary tube, the inlet always had a two-phase flow, as represented in the bottom frame. For the 0.55 mm capillary tube, different liquid levels were obtained through the refrigerant charge test campaign, which are represented in the top frames. At 67.5 g, the filter was full of liquid, and the highest COP was found at exactly this refrigerant charge. For the remaining tested refrigerant charges, the inlet alternated between the two-phase flow and liquid, with a mainly liquid inlet. 4. Conclusions This work aimed, first, to assess the actual conditions present at the capillary tube inlet of a household refrigerator system and, second, to find out if a different design of the capillary tube could lead to full liquid conditions at the tube inlet as well as if this would lead to better performance. The main conclusions drawn from this study can be summarized as follows: •
•
•
With the original capillary tube, the inlet to the capillary tube always had a two-phase flow regardless of the refrigerant charge and the operating conditions. The liquid level was always very close to the capillary tube’s mouth, and a vortex was always present, indicating the entrance of vapor and liquid. At the same time, the subcooling, estimated from the measured pressure and temperature at the outlet of the condensers and the assumption of thermodynamic equilibrium, was significantly high, e.g., around 9 K.
L. Bardoulet, J.M. Corberán and S. Martínez-Ballester / International Journal of Refrigeration 100 (2019) 265–273 •
•
•
•
In contrast, the estimation of the subcooling from the evaluation of the refrigerant enthalpy at the outlet of the condenser led to lower values, even indicating a certain amount of vapor quality. These results agree with recent results published by other authors who postulate the lack of thermodynamic equilibrium at the outlet of the condenser. The present results seem to indicate the coexistence of subcooled liquid and subcooled vapor at the inlet of the capillary tube. It has been found that the presence of vapor at the inlet of the capillary tube could be forced by the matching of the flow rates between the compressor and the capillary tube, i.e., as the capillary tube is oversized with respect to the compressor, the inlet flow must contain some vapor in order to increase the pressure drop and become more restrictive. Considering this, an additional test campaign was carried out with a smaller capillary tube (0.55 mm instead of the original 0.6 mm) and by varying the compressor speed in order to check if it was then possible to fill the capillary tube inlet fully with liquid. It was found that with the smaller diameter, it was possible to have full liquid conditions for speeds above 2500 rpm. Finally, a refrigerant charge optimization was carried out with the original refrigerator, the two different capillary tubes (0.6 and 0.55 mm, respectively), and a higher compressor speed (2900 rpm) in order to be sure that full liquid conditions could be achieved at the inlet of the smaller capillary tube. This final experimental campaign showed that the COP with the smaller diameter was always approximately 4% higher and that the maximum COP was precisely achieved when full liquid conditions were clearly observable at the inlet of the tube.
273
Acknowledgements In this project, the work of Laetitia Bardoulet was partially supported by the Santiago Grisolía 2015 program, which is funded by the Generalitat Valenciana, with the reference number GRISOLIA/2015/021. References Boeng, J., Melo, C., 2014. Mapping the energy consumption of household refrigerators by varying the refrigerant charge and the expansion restriction. Int. J. Refrig. 41, 37–44. https://doi.org/10.1016/j.ijrefrig.2013.06.005. Choi, J.M., Kim, Y.C., 2002. The effects of improper refrigerant charge on the performance of a heat pump with an electronic expansion valve and capillary tube. Energy 27, 391–404. https://doi.org/10.1016/S0360-5442(01)0 0 093-7. Hartmann, D., Melo, C., 2013. Popping noise in household refrigerators: fundamentals and practical solutions. Appl. Therm. Eng. 51, 40–47. https://doi.org/10.1016/ j.applthermaleng.2012.08.054. Hermes, C.J.L., Melo, C., Gonçalves, J.M., 2008. Modeling of non-adiabatic capillary tube flows: a simplified approach and comprehensive experimental validation. Int. J. Refrig. 31, 1358–1367. https://doi.org/10.1016/j.ijrefrig.20 08.04.0 02. Ko, J., Jeong, J.H., 2017. Effects of a non-equilibrium two-phase refrigerant flow at subcooled temperatures on the performance of an R-600a refrigeration system. Int. J. Refrig.. https://doi.org/10.1016/j.ijrefrig.2017.10.029. Lee, W.-J., Seo, J.-Y., Ko, J., Jeong, J.H., 2016. Non-equilibrium two-phase refrigerant flow at subcooled temperatures in an R600a refrigeration system. Int. J. Refrig. 70, 148–156. https://doi.org/10.1016/j.ijrefrig.2016.07.005. Liu, Y., Bullard, C.W., 1997. An Experimental and Theoretical Analysis of Capillary Tube-Suction Line Heat Exchangers. University of Illinois Project reference: ACRC TR-109. Martínez-Ballester, S., Bardoulet, L., Pisano, A., Corberán, J.M., 2017. Visualization of refrigerant flow at the capillary tube inlet of a high-efficiency household refrigerator. Int. J. Refrig. 73, 200–208. https://doi.org/10.1016/j.ijrefrig.2016.09.019.