- Email: [email protected]
Comparison of heat pump performance using fin-and-tube and microchannel heat exchangers under frost conditions
Applied Energy 87 (2010) 1187–1197
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Comparison of heat pump performance using fin-and-tube and microchannel heat exchangers under frost conditions Liang-Liang Shao a, Liang Yang a,b, Chun-Lu Zhang b,* a b
Institute of Refrigeration and Cryogenics, Shanghai Jiaotong University, Shanghai 200240, China China R&D Center, Carrier Corporation, 3239 Shen Jiang Road, Pudong, Shanghai 201206, China
a r t i c l e
i n f o
Article history: Received 31 March 2009 Received in revised form 12 July 2009 Accepted 14 August 2009 Available online 11 September 2009 Keywords: Heat pump Heat exchanger Microchannel Frost Model Experiment
a b s t r a c t Vapor compression heat pumps are drawing more attention in energy saving applications. Microchannel heat exchangers can provide higher performance via less core volume and reduce system refrigerant charge, but little is known about their performance in heat pump systems under frosting conditions. In this study, the system performance of a commercial heat pump using microchannel heat exchangers as evaporator is compared with that using conventional finned-tube heat exchangers numerically and experimentally. The microchannel and finned-tube heat pump system models used for comparison of the microchannel and finned-tube evaporator performance under frosting conditions were developed, considering the effect of maldistribution on both refrigerant and air sides. The quasi-steady-state modeling results are in reasonable agreement with the test data under frost conditions. The refrigerant-side maldistribution is found remarkable impact on the microchannel heat pump system performance under the frost conditions. Parametric study on the fan speed and the fin density under frost conditions are conducted as well to figure out the best trade-off in the design of frost tolerant evaporators. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Air-source heat pumps are becoming popular for energy saving. Frost is one of the challenges for air-source heat pumps. When evaporator surface temperature is below the freezing point and the ambient air humidity is higher than the saturation humidity at the evaporator surface temperature, frost forms on the coil surfaces. Frost increases the heat resistance of the evaporators and the air-side pressure drop through the coil, which decreases heat pump system performance. Microchannel heat exchangers, which are commonly used in automotive air-conditioning systems, are compact and offer higher performance per unit weight than conventional finnedtube heat exchangers. Microchannel heat exchangers consist of aluminum flat tubes and louvered fins, which makes less material cost than finned-tube heat exchangers. Refrigerant inventory of microchannel heat exchanger systems is also less than that of finned-tube heat exchangers. Therefore, the replacement of conventional finned-tube heat exchangers with microchannel heat exchangers is drawing more attention. The characteristics of conventional finned-tube heat exchangers in heat pump and refrigerator systems have been investigated by many researchers [1–5]. Some researchers have * Corresponding author. Tel.: +86 21 3860 3010; fax: +86 21 3860 3156. E-mail address: [email protected] (C.-L. Zhang). 0306-2619/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2009.08.021
experimentally and numerically investigated the frost formation and the performance of finned-tube heat exchangers under frosting conditions. Hayashi et al. [6] reported there were three stages of the frost formation: crystal growth period, frost layer growth period, and frost layer full growth period. Yonko and Sepey [7] found the thermal conductivity of frost layer varies with the frost density. Yang and Lee [8,9] investigated frost properties on a cold plate experimentally, and frost density and thermal conductivity correlations were given. Xia and Jacobi [10] gave an exact solution of the frosted fin efficiency with the onedimensional fin and two-dimensional frost assumption. Seker et al. [11] developed a quasi-steady state finned-tube heat exchanger model to investigate the frost formation on a finnedtube heat exchanger. The frost growth on the finned-tube heat exchangers in Seker’s model was divided into two parts: the frost thickness growth and the frost density growth. The model results were also compared with experimental data [12]. Chen et al. [13] developed a finned-tube heat exchanger frost model, and investigated the frosted heat exchanger performance combined with a fan. Yao et al. [14] developed a frosted finned-tube heat exchanger model to investigate the frost formation and its impact on performance in air-source heat pump water heater/ chiller units. Some researchers have developed various models of the microchannel heat exchangers. Yin et al. [15] developed a CO2 microchannel gas cooler model. In their model, each pass was separated
1188
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
Nomenclature A Bo C COP cp D DS Fh Fp f G H h h hd k Le M m P p Q R Re T Ts t v
vm V Vc W Wel x
area (m2) 1 boiling number ðqG1 hfg Þ constant coefficient of performance = capacity/power consumption specific heat (kJ kg1 K1) diameter (m) diffusivity of water vapor in air (m2 s1) fin height (m) fin pitch (m) friction factor mass flux (kg s1 m2) tube height (m) enthalpy (kJ kg1) heat transfer coefficient (W m2 K1) mass transfer coefficient (kg s1 m2) thermal conductivity (W m1 K1) Lewis number refrigerant charge (kg) frosting rate (kg s1), mass flow rate (kg s1) power (W) pressure (Pa) heat transfer (W) universal gas constant (kJ kmol1 K1), heat resistance (m2 K W1) Reynolds number temperature (°C) saturation temperature (°C) time (s) specific volume (m3 kg1) average specific volume (m3 kg1) volumetric flow rate (m3 s1) air velocity through minimum free-flow area (m s1) humidity ratio (kg kga1) Weber number based on liquid, ðG2 D=ql rÞ coordinate
into 10 equal-length element. Asinari et al. [16] developed a microchannel gas cooler model considering the thermal conduction in heat exchanger. They found the accuracy of capacity prediction could be improved if thermal conduction was included in the gas cooler model. Shao et al. [17] developed a port-by-port serpentine microchannel condenser model. In their model, each port was divided into several control volumes and a set of conservation equations were solved to calculate the air and refrigerant states. The thermal conductivity between ports was also considered in the model. Kim and Bullard [18] developed a microchannel evaporator model for CO2 systems. Each slab of the microchannel evaporator was divided into several segments. Yun et al. [19] developed a microchannel evaporator model with finite volume method. The air and refrigerant leaving properties of each element were calculated with the energy and mass conservation equations. So far, the studies on the microchannel evaporator modeling under frosting conditions haven’t yet been found in open literature. The investigation on the comparison of microchannel and finned-tube heat exchangers is rare, too. Kim and Groll [20] experimentally compared four different microchannel evaporators with a baseline finned-tube evaporator in a unitary split heat pump system. The defrost frequency of the microchannel heat exchangers systems were about three times of that of the baseline finned-tube heat exchanger system. The microchannel heat exchangers system showed low average heating capacity and system performance lower than those of the finned-tube heat exchanger system.
Greek symbols d thickness (m) g efficiency h louver angle (°) k eigenvalue l viscosity (Pa s) q density (kg m3) r contraction ratio of the fin array surface tension (N m1) Subscripts a air cal calculation d dry air dis discharge f fluid fin fin fr frost g gas i ice in inlet lat latent heat m metal out outlet r refrigerant s surface sb sublimation sc subcooling sens sensible heat sh superheat suc suction v vapor wat water d thickness q density
Modeling study of microchannel evaporators is not sufficient compared to the investigation on modeling of finned-tube evaporators. In this study, the comparison of heat pump performance using microchannel and finned-tube evaporators under frost conditions are conducted numerically and experimentally. First, tube-by-tube models for microchannel and finned-tube evaporators under dry, wet, and frost conditions and a quasi-steady-state heat pump system model are developed for the numerical investigation. Then, the microchannel heat pump and the finned-tube heat pump are tested under frosting conditions. The test results are compared with the simulation results using the numerical models developed in this study. The effects of the refrigerant-side maldistribution, the fan speed, and the fin density on the system performance under frost conditions are numerically investigated as well.
2. Heat pump system model The air-source heat pump system model is a quasi-steady state model. In each time interval, the system model is solved with the steady state assumption. The growth of the frost layers is calculated at the end of the time interval. The calculated frost layer growth is used in the next time interval. Including the microchannel evaporator model and the finned-tube evaporator model in the previous sections, the heat pump system model also contains a compressor model, a parallel-flow shell-and-tube condenser mod-
1189
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
el, an expansion valve model, and a fan model. Refprop 7.1 [21] is used to calculate the refrigerant properties in the heat pump system model. 2.1. Evaporator models
2.1.1. Frost model The frost model calculates the frost layer growth of an element, which can be integrated into both the microchannel evaporator model and the finned-tube evaporator model for the frosting conditions analysis. Following assumptions are made for the frosting process: the frosting process is quasi-steady state; the frost layer is homogeneous over the surface; the frost layer properties are average properties; the surface temperature is same for one element; frost grows layer-by-layer on the surface; one-dimensional fin and two-dimensional frost layer; the water vapor in frost layers are treated as ideal gas.
There are two ways for the sublimation of water vapor on frost layer [11,14,22]: part of the water vapor frosting on the surface on the frost layer, which increasing the frost layer thickness; part of the water vapor diffusing in the frost layer and sublimating inside frost layer, which increasing the frost layer density. Fig. 1 shows the schematic of a frost layer on a fin. The frosting rate can be divided into two parts: the frosting rate increasing frost layer thickness, md, and the frosting rate increasing the frost layer density, mq. The total frost rate can be expressed as:
mfr ¼ md þ mq
ð1Þ
The water vapor diffusing rate through a porous frost layer contributes the increasing of frost layer density. Kondepudi and O’Neal [22] calculated the water vapor diffusing as follows:
" mq ¼ As DS
1 ðqfr =qi Þ
#
1 þ ðqfr =qi Þ0:5
dqv dx
ð2Þ
dT s þ mq hsb dx Qa
hsb þ DS
mq ¼ As DS
1 ðqfr =qi Þ
1 þ ðqfr =qi Þ
0:5
#
Δδfr Frost Fin Fig. 1. Schematic of a frost layer.
δfr
pv ðhsb RT s Þ
mq Dt As dfr md Ddfr ¼ Dt As qfr
ð6Þ ð7Þ
The Hayashi’s correlation [1] is used to calculate the density of the increasing thickness part frost layer. The Yonko and Sepey’s correlation [7] is used to calculate the frost layer conductivity. The Xia and Jacobi’s one term approximation solution [10] of the frost fin efficiency is used to calculate the fin efficiency of frost fins, in which the one-dimensional fin and two-dimensional frost layer assumptions are used. 2.1.2. Microchannel evaporator model A port-by-port incremental microchannel evaporator model is developed for the air-source heat pump air-conditioning system model to compare the performance of the microchannel evaporator and the finned-tube evaporator under frosting conditions. Frosted, wet, and dry fins are considered in the model to deal with different air entering conditions. The heat transfer and pressure drop of each port in one flat tube are calculated by solving the convention equations. Each port is divided into several elements along the refrigerant flow direction. Each element can be treated as a cross-flow heat exchanger, as shown in Fig. 2. Each element contains a single port and the accompanying fins. Following governing equations are solved to calculate the heat transfer and pressure drop in one element. Refrigerant mass conservation equation:
mr;in ¼ mr;out
ð8Þ
Refrigerant momentum conservation equation:
mr,out pr,out hr,out
ð3Þ
The total heat transferred from the air to the fin includes the heat conducting through the frost layer and the latent heat of the water vapor which increases the density of the frost layer, as shown in Eq. (4):
1ðqfr =qi Þ 0:5 1þðqfr =qi Þ
Dqfr ¼
pv hsb dT s 1 dx RT 2s RT s
ð5Þ
kfr R2 T 3s
With the quasi-steady state assumption, the frost layer density and thickness don’t change in one time interval in the quasi-steady state frost model. After the states of one time interval are solved, the frost layer density and thickness are calculated for next time interval in Eqs. (6) and (7), respectively.
where the molecular diffusivity correlation of Ecker and Drake [23] is used to calculate the frost density increasing rate. Substitute the ideal gas state equation and the Clausius–Clapeyron equation into Eq. (2), yields:
"
ð4Þ
Combine Eqs. (3) and (4), yields:
mq ¼
In air-source heat pump systems, performance of evaporators affect system heating performance a lot, especially under frosting conditions. A microchannel evaporator model and a finned-tube evaporator model are developed for frosting conditions to investigate and compare the performance of microchannel evaporators and finned-tube evaporators and their system performance.
(1) (2) (3) (4) (5) (6) (7)
Q a ¼ As kfr
ma,in pa,in ha,in
ma,out pa,out ha,out
mr,in pr,in hr,in
Fig. 2. Schematic of a microchannel heat exchanger element.
1190
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
pr;in pr;out ¼ G2
v r;out v r;in
þf
G2 v m 2D
mr,out pr,out hr,out
ð9Þ
Refrigerant energy conservation equation:
Q r ¼ mr;in hr;out hr;in
ð10Þ
Heat transfer equation:
Q a ¼ Q a;sens þ Q a;lat
ð11Þ
where
Q a;sens ¼
As ðT a T r Þ 1=ðha gs Þ þ ðRm þ Rfr Þ þ ðAs =Ar Þ=ha
Q a;lat ¼ ma;d W a;in W a;out hlat
ð12Þ
ma,out pa,out ha,out mr,in pr,in hr,in
ð13Þ
The air-side humidity balance is
ma dW a ¼ hd ðW a W w Þ dAa
ð14Þ
where
hd ¼
ha Le cp;a
ma,in pa,in ha,in
Fig. 3. Schematic of a finned-tube heat exchanger element.
ð15Þ
Following correlations are used to calculate the heat transfer coefficients and friction factors in the governing equations of the microchannel evaporator model. The selection of the heat transfer and the pressure drop correlations has little influence on the model algorithm. In other words, the correlations used in the microchannel evaporator model in this study could be easily replaced with some more proper correlations if any. Yun et al. correlation [24] is used to calculate two-phase heat transfer coefficient. Yan et al. correlation [25] is used to calculate two-phase friction factor. Jin et al. correlation [26] is used to calculate the refrigerant distribution in the microchannel evaporator. For dry surfaces, Kim and Bullard’s correlations [27] were used to calculate air-side heat transfer. For wet surfaces, Kim and Bullard’s correlations [28] were used to calculate air-side heat transfer. The algorithm of the microchannel evaporator model is as follow: First, the conservation equations of each element in one port were solved along the refrigerant flow direction. After all the ports in one tube were calculated, the refrigerant pressure drop of each port was balanced by adjusting refrigerant mass flow rate of each port in the tube. Second, the refrigerant pressure drop of each tube is balance by adjusting refrigerant mass flow rate of each tube in the evaporator, after all the flat tubes are calculated. 2.1.3. Finned-tube evaporator model A tube-by-tube incremental finned-tube evaporator model was developed for the air-source heat pump air-conditioning system model to compare the performance of the microchannel evaporator and the finned-tube evaporator under frosting conditions. Frosted, wet, and dry fins are considered in the model to deal with different air entering conditions. The heat transfer and pressure drop of each tube in finned-tube evaporators are calculated by solving the convention equations. Each tube is divided into several elements along the refrigerant flow direction. Each element can be treated as a cross-flow heat exchanger, as shown in Fig. 3. The governing equations of the elements in the finned-tube evaporator model are the same as those of the elements in the microchannel evaporator model: the refrigerant mass conservation equation, Eq. (8), the refrigerant momentum conservation equation, Eq. (9), the refrigerant energy conservation equation, Eq. (10), the heat transfer equations, Eqs. (11)–(13), the air water vapor mass conservation equation, Eq. (14). But the calculations of the finned-tube heat exchanger geometries in each element were
different from those of the microchannel heat exchanger geometries, such as the heat transfer area and the surface ratio. Following correlations are used to calculate the heat transfer coefficients and friction factors in the governing equations of the finned-tube evaporator model. Gungor and Winterton correlation [29] is used to calculate the refrigerant-side two-phase heat transfer coefficient. Pierre correlation [30] is used to calculate the refrigerant-side two-phase pressure drop. Wang and Du correlation [31] is used to calculate the air-side heat transfer and pressure drop. The algorithm of the finned-tube evaporator model is as follow: First, the conservation equations of each element in one tube are solved along the refrigerant flow direction. After all the tubes in one circuit are calculated along the refrigerant flow direction, the calculation move onto the next circuit until all the circuits in the finned-tube evaporator are all solved. Second, the refrigerant pressure drop of each circuit is balance by adjusting refrigerant mass flow rate of each circuit in the finned-tube evaporator until the refrigerant pressure drop of each circuit is the same. 2.2. Compressor model An ARI polynomial equation compressor model [32] is used in the heat pump system model. The compressor mass flow rate is:
X ¼ C1 þ C2 Tse þ C3 Tsc þ C4 Ts2e þ C5 Tse Tsc þ C6 Ts2e þ C7 Ts3e þ C8 Tsc Ts2e þ C9 Tse Ts2c þ C10 Ts3c
ð16Þ
where X can represent the compressor power input or the refrigerant mass flow rate. The coefficients in Eq. (16) are provided by the compressor manufacturer. 2.3. Condenser model A parallel-flow shell-and-tube refrigerant-to-water condenser is used in the tested heat pump system. Refrigerant is inside the tube and water is on the shell side. The condenser model is divided into three regions: desuperheated region, two-phase region, and subcooled region. The conservation equations of each region were solved to calculate the heat transfer and the leaving states of refrigerant and water. Refrigerant mass conservation equation:
mr;cond;in ¼ mr;cond;out
ð17Þ
Refrigerant energy conservation equation:
Q cond ¼ mr;cond;in hr;cond;in hr;cond;out
ð18Þ
1191
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
Heat transfer from the refrigerant to the water:
Q cond ¼
Water energy conservation equation:
Awat ðT r T wat Þ 1=hwat þ Rm þ ðAwat =Ar Þ=hr
Q cond ¼ mwat;in cp; wat T wat;out T wat;in
ð19Þ
Start
Input system geometries
Input air entering state: T db,in and Twb,in Input water entering state: Twat,in and V wat,in Initialize V a, Ts c, Ts e, and ΔT sc
Calculate compressor model with Tsc, Ts e, and ΔTsh to find m and Tdis
Adjust Ts c
Calculate condenser model with Ts c, T dis , and ΔT sh to find m and ΔT sc,cal
N
ΔTsc,cal = ΔT sc ? Increase time
Y Calculate valve model with Tsc and ΔTsc to find h val,out
Adjust Tse
Calculate evaporator model with Va, Ts e, h val,out , and m to find ΔT sh,cal and Δp a
Adjust ΔTsc
Adjust V a
Calculate fan model with Va to find Δp a,cal
N
Δpa,cal = Δp a ? Y N
ΔT sh,cal = ΔT sh ? Y Sum charge Mcal
N
Mcal = M ? Y Frost layer growth
End ?
N
End Fig. 4. Flowchart of heat pump system model.
ð20Þ
1192
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
Cavallini et al. correlation [33] is used to calculate the refrigerant two-phase heat transfer coefficient. Gnielinski correlation [34] is used to calculate the water heat transfer coefficient. 2.4. Expansion valve model The expansion process is considered as an isenthalpic process.
hin ¼ hout
ð21Þ
Using an expansion valve, the refrigerant superheat of at evaporator outlet is assumed constant. 2.5. Fan model A polynomial fan curve was integrated in the heat pump system model to evaluate the effect of the frost growth on the heat pump performance. With the growth of the frost layer, the air-side pressure drop increasing. At each time step, the air flow rate was balanced with the air-side pressure drop of the evaporator and the air pressure rise of the fan, Eq. (22):
dprise ¼ C 0 þ C 1 V þ C 2 V 2
channel heat exchangers circuit. When the finned-tube circuit was under testing, the microchannel circuit was off and vice versa. The microchannel heat exchangers in the heat pump system were designed capacity-equivalent to the conventional finnedtube heat exchanger in the same heat pump system. In the cooling mode test, the microchannel heat exchanger system obtained the same level performance. The geometries of the microchannel heat exchanger and the finned-tube heat exchanger are shown in Tables 1 and 2, respectively. As the air-side pressure drop of the microchannel heat exchanger is lower than that of the finned-tube heat exchanger, the fan power of microchannel heat exchanger system is less than that of the finned-tube heat exchanger system. Hence, the COP (defined as system capacity over power) of the microchannel heat exchanger system is higher than that of the finned-tube system, shown in Table 3. Fig. 6 shows the test facility
ð22Þ
where the coefficients C0, C1, and C2 are reduced with the fan test data. 2.6. System algorithm The heat pump system model is developed for system performance (heating mode) prediction. The inputs of the heat pump system model are divided into two parts: (1) the heat pump system configurations, including the refrigerant type and charge, the compressor type, the condenser geometries, the evaporator geometries, the fan type, the suction superheat set point and the fan speed; (2) the operating conditions, including the air dry-bulb and wet-bulb temperature, the water flow rate, and water entering temperature. The component models of the heat pump system model are solved sequentially. The heat pump system algorithm is shown in Fig. 4. The air flow rate (Va), saturation condensing temperature (Tsc), saturation evaporating temperature (Tse), and condenser leaving subcooling (DTsc) are initially guessed. Va, Tsc, Tse, and DTsc are iterated to find the fan pressure rise, refrigerant mass flow rate, suction superheat, and refrigerant charge in the system model at one time step. After the system is solved at one time step, the frost growth during the time interval is calculated for the next time interval. 3. Experimental results and model validation The microchannel heat exchangers in this study were designed capacity-equivalent to the conventional finned-tube heat exchangers in the cooling model, so that both the microchannel heat pump and the finned-tube heat pump could provide almost same cooling capacity. The heating performance of the microchannel heat pump and the finned-tube heat pump were compared experimentally and numerically under rating condition and frosting condition.
Fig. 5. Tested heat pump.
Table 1 Characteristics of the microchannel heat exchanger. Characteristics of the microchannel heat exchanger Face area (m2) Fin density (fins/in.) Tube depth (mm) Tube length (mm) Pass number Port inner diameter (mm) Port number Fin height (mm) Fin depth (mm) Louver angle (°) Header inner diameter (mm)
2.12 20 25 1030 1 0.64 32 7.6 0.1 25 27
Table 2 Characteristics of the finned-tube heat exchanger. Characteristics of the finned-tube heat exchanger
3.1. Heat pump system and test apparatus The tested heat pump in this study, shown in Fig. 5, has two separated refrigerant circuits. All the components in the two refrigerant circuits are the same, except the coils. One refrigerant circuit contains the finned-tube heat exchangers, while the other contains the microchannel heat exchangers. In the test, the finned-tube heat exchangers circuit is compared with the micro-
Face area (m2) Fin density (fins/in.) Tube length (mm) Row number Tube number Tube outer diameter (mm) Circuit number Row pitch (mm) Tube pitch (mm)
2.18 17 1950 3 44 10 22 19 25.4
1193
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197 Table 3 Performance in cooling mode.
Tested COP
Finned-tube system
Microchannel system
99.8 2.77
99.2 2.86
Water
Test chamber
M
T
Simulated COP
2.6
Heating COP
Capacity (kW) COP
2.8
2.4
2.2
T
2.0 P,T
P,T
Condenser
1.8 0
30
60
90
Time (min)
Compressor EXV
Fig. 8. Heating COP of the finned-tube system (2/1 °C).
Air
Tdb , Twb
P,T
P,T
Reversing valve
Evaporator
Fig. 6. Schematic of the tested system and testing facility.
Table 4 Performance in heating mode.
Capacity (kW) COP
Finned-tube system
Microchannel system
103.8 2.69
114.1 2.73
for the heating mode tests of both microchannel and finned-tube systems. The ambient dry-bulb and wet-bulb temperature in the test chamber, the water flow rate, water entering and leaving temperature, compressor discharge pressure and temperature, condenser leaving pressure and temperature, evaporator leaving pressure and temperature, compressor suction pressure temperature, and system power are measured. The air entering the test chamber is controlled to keep the dry-bulb and wet-bulb temperatures in the test chamber for the tested heat pump systems. The heat pump system entering water temperature is also controlled to keep the entering water temperature during the test. The uncertainties for pressure, temperature and flow rate measurements were about ±8 kPa, ±0.5 °C and ±3 105 m3/s, respectively. 3.2. Experimental results and model results The microchannel heat pump system and the finned-tube heat pump system were tested under the rating condition
120
110
Tested capacity
100
100
Heating capacity (kW)
Heating capacity (kW)
Simulated capacity
90
80
80
60
Tested capacity Uniform distribution model Maldistribution model
40
70
20
0
30
60
90
Time (min) Fig. 7. System heating capacity of the finned-tube system (2/1 °C).
0
10
20
30
Time (min) Fig. 9. System heating capacity of the microchannel system (2/1 °C).
1194
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
(dry-bulb temperature 7 °C and wet-bulb temperature 6 °C) and the frost condition (dry-bulb temperature 2 °C and wet-bulb temperature 1 °C), respectively. Table 4 shows the performance comparison of the microchannel system and the finned-tube system under the rating condition. The microchannel heat pump system could provide more heating capacity than the finnedtube heat pump system. The microchannel system model and the finned-tube system model were calibrated with the tested results under the rating condition, respectively. The calibrated microchannel system model and the finned-tube system model were used to predict the system performance under frosting conditions, respectively. The finned-tube heat pump system was tested as the baseline in the comparison with the microchannel evaporator heat pump system under frosting condition (2/1 °C). Figs. 7 and 8 show the comparison of the test and the predicted results of the system capacity and COP of the finned-tube system, respectively. The heating time
Heating COP
3.0
Tested COP Uniform distribution model Maldistribution model
2.5
2.0
of the finned-tube heat pump system was about 80 min. In the beginning several minutes, the system capacity and COP swung, as the heat pump finished the defrost cycle and the 4-way reversing valve turned the heat pump into heating mode. The 4-way reversing valve caused the refrigerant pressure changing sharply. The simulated heating time is slightly longer than the tested heating time. In Fig. 7, the finned-tube system frost model showed good accuracy on the system capacity. The simulated heating time was little longer than the tested heating time. In Fig. 8, the simulated finned-tube system COP is a little lower than the tested COP. The microchannel heat pump system was tested under frosting condition (2/1 °C). Figs. 9 and 10 showed the tested system capacity and COP of the microchannel heat pump system, respectively. The heating time of the microchannel heat pump system was much shorter than that of the finned-tube heat pump system under frosting condition. In Figs. 9 and 10, the results of the refrigerant-side uniform distribution model and the maldistribution model were compared with the test data. Fig. 11 shows the frost distribution profile on the microchannel evaporator of the tested heat pump. The frost on the microchannel evaporator in the test was not uniform because of the inertia effect in the microchannel evaporator header. The microchannel evaporator manifolds distribution model developed by Jin et al. [26] was used to simulate the microchannel heat pump system. The maldistribution model showed good agreement with the test results in Figs. 9 and 10. The simulated heating time of the uniform distribution system was about 100% longer than that of the maldistribution case. The uniform distribution model simulation results showed the performance and heating time of the microchannel heat pump could be both improved if the refrigerant distribution of the microchannel evaporators was uniform. The frosting condition test results of the microchannel system and the finned-tube system showed that the heating time of the microchannel system was less than one-fifth of that of the finned-tube system. The microchannel heat pump system needed much more defrosting times than the finned-tube heat pump system for a same frosting duration.
1.5 0
10
20
30 4. Parametric study
Time (min) Fig. 10. Heating COP of the microchannel system (2/1 °C).
The heating time of the microchannel heat pump system was much less than that of the finned-tube heat pump system. The
Fig. 11. Frost on the microchannel evaporator (2/1 °C).
1195
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
parametric study of the microchannel heat pump system was used to investigate the factors on the system performance under frosting conditions to improve the system heating time. The frosting condition in the parametric study was 2 °C dry-bulb temperature and 1 °C wet-bulb temperature. In the simulation, the inputs of the heat pump model are as follow: the air dry-bulb temperature is 2 °C, the wet-bulb temperature is 1 °C, the water entering temperature is 40 °C, and the entering water flow rate is 20 m3/h. 4.1. Effects of the fan speed The fan speed effect of the microchannel heat pump system on frost performance was investigated with the microchannel system model. Figs. 12 and 13 show the microchannel system capacity and COP of the fan speed changing ±10%. System capacity and heating time were improved by increasing the fan speed, as the air volu-
110
Designed fan speed +10% fan speed -10% fan speed
100
Designed fan speed +10% fan speed -10% fan speed
100
Heating capacity (kW)
Heating capacity (kW)
110
metric flow rate increases with the fan speed. But the high fan speed COP is lower than the low fan speed COP at initial part of the frosting process, as high fan speed needs more power. But the low fan COP drops faster than the high fan speed COP, as the frost on the microchannel evaporator increases faster for the low fan speed system. The fan speed effect of the finned-tube heat pump system on frost performance was compared with that of the microchannel system. The simulated finned-tube system capacity and COP of the fan speed changing ±10% were shown in Figs. 14 and 15. Same trends of the system capacity and COP of the microchannel system were found. For both the microchannel and finned-tube heat pumps, the fan speed affected the system performance and the heating time with similar trends. High fan speed could increase the heating time. But for the microchannel heat pump with the refrigerant maldistribu-
90
80
90
80
70
60
70 0
5
10
15
0
20
20
Time (min)
Designed fan speed +10% fan speed -10% fan speed
2.6
80
100
Fig. 14. Fan speed effect on the finned-tube system capacity.
2.8
2.8
60
Time (min)
Fig. 12. Fan speed effect on the microchannel system capacity.
2.4 2.2
Designed fan speed +10% fan speed -10% fan speed
2.6
Heating COP
Heating COP
40
2.4
2.2
2.0
2.0
1.8
1.8
1.6 0
5
10
15
Time (min) Fig. 13. Fan speed effect on the microchannel system COP.
20
0
20
40
60
80
Time (min) Fig. 15. Fan speed effect on the finned-tube system COP.
100
1196
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
tion, the heating time after increasing the fan speed was still much less than that of the finned-tube heat pump. The heating time of the microchannel heat pump with refrigerant maldistribution, shown in Figs. 14 and 15, is still much less than that of the heat pump with uniform distribution, shown in Figs. 9 and 10, even if the fan speed has been increased by 10%. The refrigerant maldistribution in the microchannel heat exchangers affected the system performance and heating time much than the fan speed.
4.2. Effects of the fin density The fin density effect of the heat pump system on system capacity and COP are shown in Figs. 16 and 17, respectively. When the refrigerant maldistribution happened in the microchannel evaporator manifold, the fin density changing did not significantly affect the system performance. From the simulation results in Figs. 16
and 17, the lower and higher fin density microchannel systems provide almost the same system capacity and COP as the designed fin density microchannel system. The fin density effect of the finned-tube heat pump system on frost performance was compared with that of the microchannel system. Figs. 18 and 19 show the simulated finned-tube system capacity and COP of the fan density changing ±10% and ±20%. Lower fin density design can improve the heating time of the finnedtube heat pump with less negative effect on the finned-tube system capacity and COP. The fin density effect on the microchannel system was much different from that on the finned-tube system. Lower fin density of the finned-tube heat exchangers could improve the finned-tube system heating time. When refrigerant maldistribution occurred in the microchannel heat exchangers, changing the fin density of the microchannel heat exchangers couldn’t affect the system performance and the heating time.
110
100
Heating capacity (kW)
Heating capacity (kW)
100
90
Designed fin density +20% fin density +10% fin density -10% fin density -20% fin density
80
70
60
0
5
10
90
80
Designed fin density +10% fin density -10% fin density
70
60
15
0
Time (min)
30
60
90
Time (min)
Fig. 16. Fin density effect of the microchannel evaporator on system capacity.
2.6
2.6
2.4
2.4 Heating COP
Heating COP
Fig. 18. Fin density effect of the finned-tube evaporator on system capacity.
2.2
Designed fin density +20% fin density +10% fin density -10% fin density -20% fin density
2.0
1.8
2.2
2.0
Designed fin density +10% fin density -10% fin density
1.8
1.6
1.6
0
5
10
15
Time (min) Fig. 17. Fin density effect of the microchannel evaporator on system COP.
0
30
60
90
Time (min) Fig. 19. Fin density effect of the finned-tube evaporator on system COP.
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
5. Conclusions The performance of microchannel evaporator and finned-tube evaporator in the heat pump system under frosting conditions has been compared in this study. Quasi-steady-state microchannel and finned-tube heat pump system models for frost conditions have been developed, respectively. The refrigerant maldistribution effect has been considered in the microchannel heat pump model. Both test results and simulation results show that the frost period of the tested finned-tube heat pump system was much longer than that of the microchannel heat pump system. The parametric study of the microchannel heat pump system has been conducted to investigate the factors for the performance of the microchannel evaporator under frosting conditions. It reveals that the refrigerant distribution in the microchannel heat exchangers is the key factor to the microchannel heat pump performance during the frost period.
Acknowledgements The authors would like to thank Mr. Jing Li and Dr. Xinghua Huang from Carrier China R&D Center for their efforts in the experiments.
References [1] Comakli K, Simsek F, Comakli O, Sahin B. Determination of optimum working conditions R22 and R404A refrigerant mixtures in heat-pumps using Taguchi method. Appl Energy 2009;86(11):2451–8. [2] Huang D, Li QX, Yuan XL. Comparison between hot-gas bypass defrosting and reverse-cycle defrosting methods on an air-to-water heat pump. Appl Energy 2009;86(9):1697–703. [3] Wang ZY, Wang XM, Dong ZM. Defrost improvement by heat pump refrigerant charge compensating. Appl Energy 2008;85(11):1050–9. [4] Barbosa Jr JR, Melo C, Hermes CJL, Waltrich PJ. A study of the air-side heat transfer and pressure drop characteristics of tube-fin ‘no-frost’ evaporators. Appl Energy 2009;86(9):1484–91. [5] Gholap AK, Khan JA. Design and multi-objective optimization of heat exchangers for refrigerators. Appl Energy 2007;84(12):1226–39. [6] Hayashi Y, Aoki A, Adachi S, Hori K. Study of frost formation based on a theoretical model of the frost layer. Heat Transfer Jpn Res 1977;6(3):79–94. [7] Yonko JD, Sepey CF. An investigation of the thermal conductivity of frost while forming on a flat horizontal plate. ASHRAE Trans 1967;73:I.1.1–I.1.10. [8] Yang DK, Lee KS. Dimensionless correlations of frost properties on a cold plate. Int J Refrig 2004;27:89–96. [9] Yang DK, Lee KS. Modeling of frosting behavior on a cold plate. Int J Refrig 2005;28:396–402.
1197
[10] Xia Y, Jacobi AM. An exact solution to steady heat conduction in a twodimensional slab on a one-dimensional fin: application to frost heat exchangers. Int J Heat Mass Transfer 2004;47:3317–26. [11] Seker D, Karatas H, Egrican N. Frost formation on fin-and-tube heat exchangers. Part I – Modeling of frost formation on fin-and-tube heat exchanger. Int J Refrig 2004;27:367–74. [12] Seker D, Karatas H, Egrican N. Frost formation on fin-and-tube heat exchangers. Part II–Experimental investigation of frost formation on fin-andtube heat exchangers. Int J Refrig 2004;27:375–7. [13] Chen H, Thomas L, Besant RW. Fan supplied heat exchanger fin performance under frosting conditions. Int J Refrig 2003;26:140–9. [14] Yao Y, Jiang YQ, Deng SM, Ma ZL. A study on the performance of the airside heat exchanger under frosting in an air source heat pump water heater/chiller unit. Int J Heat Mass Transfer 2004;47:3745–56. [15] Yin JM, Bullard CW, Hrnjak PS. R-744 gas cooler model development and validation. Int J Refrig 2001;24:692–701. [16] Asinari P, Cecchinato L, Fornasieri E. Effect of thermal conduction in microchannel gas coolers for carbon dioxide. Int J Refrig 2004;27:577–86. [17] Shao LL, Yang L, Zhang CL, GU B. Numerical modeling of serpentine microchannel condensers. Int J Refrig 2009;32:1162–72. [18] Kim MH, Bullard CW. Development of a microchannel evaporator model for CO2 air-conditioning system. Energy 2001;26:931–48. [19] Yun R, Kim YC, Park C. Numerical analysis on a microchannel evaporator designed for CO2 air-conditioning systems. Appl Therm Eng. 2007;27:1320–6. [20] Kim JH, Groll EA. Performance comparisons of a unitary split system using microchannel and fin-tube outdoor coils. ASHRAE Trans 2003;109(2):219–29. [21] Lemmon EW, Huber ML, McLinden MO. NIST reference fluid thermodynamic and transport properties – REFPROP. NIST standard reference database 23version 7.1; 2006. [22] Kondepudi SN, O’Neal DL. Performance of finned-tube het exchangers under frosting conditions: Part I – Simulation model. Int J Refrig 1993;16(3):175–80. [23] Eckert ER, Drake RM. Analysis of heat and mass transfer. New York: McGrawHill; 1972. [24] Yun R, Heo JY, Kim YC. Evaporative heat transfer and pressure drop of R410A in microchannels. Int J Refrig 2006;29:92–100. [25] Yan YY, Lin TF. Evaporation heat transfer and pressure drop of refrigerant R134a in a small pipe. Int J Heat Mass Transfer 1998;41:4183–94. [26] Jin DH, Hwang Y, Radermacher R. Refrigerant distribution in evaporator manifolds. ASHRAE 1260-TRP; 2006. [27] Kim MH, Bullard C. Air-side thermal hydraulic performance of multi-louvered fin aluminum heat exchangers. Int J Refrig 2002;25:390–400. [28] Kim MH, Bullard C. Air-side thermal performance of micro-channel heat exchangers under dehumidifying conditions. In: Tree DR, (Ed.), Proceedings of the 2000 International refrigeration conference, West Lafayette (IN): Purdue University; 2000. p. 119–26. [29] Gungor KE, Winterton RHS. A general correlation for boiling in tubes and annuli. Int J Heat Mass Transfer 1986;19:351–8. [30] Pierre B. Flow resistance with boiling refrigerants. Part I. ASHRAE J 1964:58–65. [31] Wang CC, Du JY. A heat transfer and friction correlation for wavy fin and tube heat exchangers. Int J Heat Mass Transfer 1999;42:1919–24. [32] ARI Standard 540-99, Performance rating of positive displacement compressors and compressor units; 1999. [33] Cavallini A, Censi G, Del Col D, Doretti L, Longo GA, Rossetto L. In-tube condensation of halogenated refrigerants. ASHRAE Trans 2002;108(1):146–61. [34] Gnielinski V. New equations for heat and mass transfer in turbulent pipe and channel flows. Int Chem Eng 1976;16:359–68.
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Comparison of heat pump performance using fin-and-tube and microchannel heat exchangers under frost conditions Liang-Liang Shao a, Liang Yang a,b, Chun-Lu Zhang b,* a b
Institute of Refrigeration and Cryogenics, Shanghai Jiaotong University, Shanghai 200240, China China R&D Center, Carrier Corporation, 3239 Shen Jiang Road, Pudong, Shanghai 201206, China
a r t i c l e
i n f o
Article history: Received 31 March 2009 Received in revised form 12 July 2009 Accepted 14 August 2009 Available online 11 September 2009 Keywords: Heat pump Heat exchanger Microchannel Frost Model Experiment
a b s t r a c t Vapor compression heat pumps are drawing more attention in energy saving applications. Microchannel heat exchangers can provide higher performance via less core volume and reduce system refrigerant charge, but little is known about their performance in heat pump systems under frosting conditions. In this study, the system performance of a commercial heat pump using microchannel heat exchangers as evaporator is compared with that using conventional finned-tube heat exchangers numerically and experimentally. The microchannel and finned-tube heat pump system models used for comparison of the microchannel and finned-tube evaporator performance under frosting conditions were developed, considering the effect of maldistribution on both refrigerant and air sides. The quasi-steady-state modeling results are in reasonable agreement with the test data under frost conditions. The refrigerant-side maldistribution is found remarkable impact on the microchannel heat pump system performance under the frost conditions. Parametric study on the fan speed and the fin density under frost conditions are conducted as well to figure out the best trade-off in the design of frost tolerant evaporators. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Air-source heat pumps are becoming popular for energy saving. Frost is one of the challenges for air-source heat pumps. When evaporator surface temperature is below the freezing point and the ambient air humidity is higher than the saturation humidity at the evaporator surface temperature, frost forms on the coil surfaces. Frost increases the heat resistance of the evaporators and the air-side pressure drop through the coil, which decreases heat pump system performance. Microchannel heat exchangers, which are commonly used in automotive air-conditioning systems, are compact and offer higher performance per unit weight than conventional finnedtube heat exchangers. Microchannel heat exchangers consist of aluminum flat tubes and louvered fins, which makes less material cost than finned-tube heat exchangers. Refrigerant inventory of microchannel heat exchanger systems is also less than that of finned-tube heat exchangers. Therefore, the replacement of conventional finned-tube heat exchangers with microchannel heat exchangers is drawing more attention. The characteristics of conventional finned-tube heat exchangers in heat pump and refrigerator systems have been investigated by many researchers [1–5]. Some researchers have * Corresponding author. Tel.: +86 21 3860 3010; fax: +86 21 3860 3156. E-mail address: [email protected] (C.-L. Zhang). 0306-2619/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2009.08.021
experimentally and numerically investigated the frost formation and the performance of finned-tube heat exchangers under frosting conditions. Hayashi et al. [6] reported there were three stages of the frost formation: crystal growth period, frost layer growth period, and frost layer full growth period. Yonko and Sepey [7] found the thermal conductivity of frost layer varies with the frost density. Yang and Lee [8,9] investigated frost properties on a cold plate experimentally, and frost density and thermal conductivity correlations were given. Xia and Jacobi [10] gave an exact solution of the frosted fin efficiency with the onedimensional fin and two-dimensional frost assumption. Seker et al. [11] developed a quasi-steady state finned-tube heat exchanger model to investigate the frost formation on a finnedtube heat exchanger. The frost growth on the finned-tube heat exchangers in Seker’s model was divided into two parts: the frost thickness growth and the frost density growth. The model results were also compared with experimental data [12]. Chen et al. [13] developed a finned-tube heat exchanger frost model, and investigated the frosted heat exchanger performance combined with a fan. Yao et al. [14] developed a frosted finned-tube heat exchanger model to investigate the frost formation and its impact on performance in air-source heat pump water heater/ chiller units. Some researchers have developed various models of the microchannel heat exchangers. Yin et al. [15] developed a CO2 microchannel gas cooler model. In their model, each pass was separated
1188
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
Nomenclature A Bo C COP cp D DS Fh Fp f G H h h hd k Le M m P p Q R Re T Ts t v
vm V Vc W Wel x
area (m2) 1 boiling number ðqG1 hfg Þ constant coefficient of performance = capacity/power consumption specific heat (kJ kg1 K1) diameter (m) diffusivity of water vapor in air (m2 s1) fin height (m) fin pitch (m) friction factor mass flux (kg s1 m2) tube height (m) enthalpy (kJ kg1) heat transfer coefficient (W m2 K1) mass transfer coefficient (kg s1 m2) thermal conductivity (W m1 K1) Lewis number refrigerant charge (kg) frosting rate (kg s1), mass flow rate (kg s1) power (W) pressure (Pa) heat transfer (W) universal gas constant (kJ kmol1 K1), heat resistance (m2 K W1) Reynolds number temperature (°C) saturation temperature (°C) time (s) specific volume (m3 kg1) average specific volume (m3 kg1) volumetric flow rate (m3 s1) air velocity through minimum free-flow area (m s1) humidity ratio (kg kga1) Weber number based on liquid, ðG2 D=ql rÞ coordinate
into 10 equal-length element. Asinari et al. [16] developed a microchannel gas cooler model considering the thermal conduction in heat exchanger. They found the accuracy of capacity prediction could be improved if thermal conduction was included in the gas cooler model. Shao et al. [17] developed a port-by-port serpentine microchannel condenser model. In their model, each port was divided into several control volumes and a set of conservation equations were solved to calculate the air and refrigerant states. The thermal conductivity between ports was also considered in the model. Kim and Bullard [18] developed a microchannel evaporator model for CO2 systems. Each slab of the microchannel evaporator was divided into several segments. Yun et al. [19] developed a microchannel evaporator model with finite volume method. The air and refrigerant leaving properties of each element were calculated with the energy and mass conservation equations. So far, the studies on the microchannel evaporator modeling under frosting conditions haven’t yet been found in open literature. The investigation on the comparison of microchannel and finned-tube heat exchangers is rare, too. Kim and Groll [20] experimentally compared four different microchannel evaporators with a baseline finned-tube evaporator in a unitary split heat pump system. The defrost frequency of the microchannel heat exchangers systems were about three times of that of the baseline finned-tube heat exchanger system. The microchannel heat exchangers system showed low average heating capacity and system performance lower than those of the finned-tube heat exchanger system.
Greek symbols d thickness (m) g efficiency h louver angle (°) k eigenvalue l viscosity (Pa s) q density (kg m3) r contraction ratio of the fin array surface tension (N m1) Subscripts a air cal calculation d dry air dis discharge f fluid fin fin fr frost g gas i ice in inlet lat latent heat m metal out outlet r refrigerant s surface sb sublimation sc subcooling sens sensible heat sh superheat suc suction v vapor wat water d thickness q density
Modeling study of microchannel evaporators is not sufficient compared to the investigation on modeling of finned-tube evaporators. In this study, the comparison of heat pump performance using microchannel and finned-tube evaporators under frost conditions are conducted numerically and experimentally. First, tube-by-tube models for microchannel and finned-tube evaporators under dry, wet, and frost conditions and a quasi-steady-state heat pump system model are developed for the numerical investigation. Then, the microchannel heat pump and the finned-tube heat pump are tested under frosting conditions. The test results are compared with the simulation results using the numerical models developed in this study. The effects of the refrigerant-side maldistribution, the fan speed, and the fin density on the system performance under frost conditions are numerically investigated as well.
2. Heat pump system model The air-source heat pump system model is a quasi-steady state model. In each time interval, the system model is solved with the steady state assumption. The growth of the frost layers is calculated at the end of the time interval. The calculated frost layer growth is used in the next time interval. Including the microchannel evaporator model and the finned-tube evaporator model in the previous sections, the heat pump system model also contains a compressor model, a parallel-flow shell-and-tube condenser mod-
1189
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
el, an expansion valve model, and a fan model. Refprop 7.1 [21] is used to calculate the refrigerant properties in the heat pump system model. 2.1. Evaporator models
2.1.1. Frost model The frost model calculates the frost layer growth of an element, which can be integrated into both the microchannel evaporator model and the finned-tube evaporator model for the frosting conditions analysis. Following assumptions are made for the frosting process: the frosting process is quasi-steady state; the frost layer is homogeneous over the surface; the frost layer properties are average properties; the surface temperature is same for one element; frost grows layer-by-layer on the surface; one-dimensional fin and two-dimensional frost layer; the water vapor in frost layers are treated as ideal gas.
There are two ways for the sublimation of water vapor on frost layer [11,14,22]: part of the water vapor frosting on the surface on the frost layer, which increasing the frost layer thickness; part of the water vapor diffusing in the frost layer and sublimating inside frost layer, which increasing the frost layer density. Fig. 1 shows the schematic of a frost layer on a fin. The frosting rate can be divided into two parts: the frosting rate increasing frost layer thickness, md, and the frosting rate increasing the frost layer density, mq. The total frost rate can be expressed as:
mfr ¼ md þ mq
ð1Þ
The water vapor diffusing rate through a porous frost layer contributes the increasing of frost layer density. Kondepudi and O’Neal [22] calculated the water vapor diffusing as follows:
" mq ¼ As DS
1 ðqfr =qi Þ
#
1 þ ðqfr =qi Þ0:5
dqv dx
ð2Þ
dT s þ mq hsb dx Qa
hsb þ DS
mq ¼ As DS
1 ðqfr =qi Þ
1 þ ðqfr =qi Þ
0:5
#
Δδfr Frost Fin Fig. 1. Schematic of a frost layer.
δfr
pv ðhsb RT s Þ
mq Dt As dfr md Ddfr ¼ Dt As qfr
ð6Þ ð7Þ
The Hayashi’s correlation [1] is used to calculate the density of the increasing thickness part frost layer. The Yonko and Sepey’s correlation [7] is used to calculate the frost layer conductivity. The Xia and Jacobi’s one term approximation solution [10] of the frost fin efficiency is used to calculate the fin efficiency of frost fins, in which the one-dimensional fin and two-dimensional frost layer assumptions are used. 2.1.2. Microchannel evaporator model A port-by-port incremental microchannel evaporator model is developed for the air-source heat pump air-conditioning system model to compare the performance of the microchannel evaporator and the finned-tube evaporator under frosting conditions. Frosted, wet, and dry fins are considered in the model to deal with different air entering conditions. The heat transfer and pressure drop of each port in one flat tube are calculated by solving the convention equations. Each port is divided into several elements along the refrigerant flow direction. Each element can be treated as a cross-flow heat exchanger, as shown in Fig. 2. Each element contains a single port and the accompanying fins. Following governing equations are solved to calculate the heat transfer and pressure drop in one element. Refrigerant mass conservation equation:
mr;in ¼ mr;out
ð8Þ
Refrigerant momentum conservation equation:
mr,out pr,out hr,out
ð3Þ
The total heat transferred from the air to the fin includes the heat conducting through the frost layer and the latent heat of the water vapor which increases the density of the frost layer, as shown in Eq. (4):
1ðqfr =qi Þ 0:5 1þðqfr =qi Þ
Dqfr ¼
pv hsb dT s 1 dx RT 2s RT s
ð5Þ
kfr R2 T 3s
With the quasi-steady state assumption, the frost layer density and thickness don’t change in one time interval in the quasi-steady state frost model. After the states of one time interval are solved, the frost layer density and thickness are calculated for next time interval in Eqs. (6) and (7), respectively.
where the molecular diffusivity correlation of Ecker and Drake [23] is used to calculate the frost density increasing rate. Substitute the ideal gas state equation and the Clausius–Clapeyron equation into Eq. (2), yields:
"
ð4Þ
Combine Eqs. (3) and (4), yields:
mq ¼
In air-source heat pump systems, performance of evaporators affect system heating performance a lot, especially under frosting conditions. A microchannel evaporator model and a finned-tube evaporator model are developed for frosting conditions to investigate and compare the performance of microchannel evaporators and finned-tube evaporators and their system performance.
(1) (2) (3) (4) (5) (6) (7)
Q a ¼ As kfr
ma,in pa,in ha,in
ma,out pa,out ha,out
mr,in pr,in hr,in
Fig. 2. Schematic of a microchannel heat exchanger element.
1190
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
pr;in pr;out ¼ G2
v r;out v r;in
þf
G2 v m 2D
mr,out pr,out hr,out
ð9Þ
Refrigerant energy conservation equation:
Q r ¼ mr;in hr;out hr;in
ð10Þ
Heat transfer equation:
Q a ¼ Q a;sens þ Q a;lat
ð11Þ
where
Q a;sens ¼
As ðT a T r Þ 1=ðha gs Þ þ ðRm þ Rfr Þ þ ðAs =Ar Þ=ha
Q a;lat ¼ ma;d W a;in W a;out hlat
ð12Þ
ma,out pa,out ha,out mr,in pr,in hr,in
ð13Þ
The air-side humidity balance is
ma dW a ¼ hd ðW a W w Þ dAa
ð14Þ
where
hd ¼
ha Le cp;a
ma,in pa,in ha,in
Fig. 3. Schematic of a finned-tube heat exchanger element.
ð15Þ
Following correlations are used to calculate the heat transfer coefficients and friction factors in the governing equations of the microchannel evaporator model. The selection of the heat transfer and the pressure drop correlations has little influence on the model algorithm. In other words, the correlations used in the microchannel evaporator model in this study could be easily replaced with some more proper correlations if any. Yun et al. correlation [24] is used to calculate two-phase heat transfer coefficient. Yan et al. correlation [25] is used to calculate two-phase friction factor. Jin et al. correlation [26] is used to calculate the refrigerant distribution in the microchannel evaporator. For dry surfaces, Kim and Bullard’s correlations [27] were used to calculate air-side heat transfer. For wet surfaces, Kim and Bullard’s correlations [28] were used to calculate air-side heat transfer. The algorithm of the microchannel evaporator model is as follow: First, the conservation equations of each element in one port were solved along the refrigerant flow direction. After all the ports in one tube were calculated, the refrigerant pressure drop of each port was balanced by adjusting refrigerant mass flow rate of each port in the tube. Second, the refrigerant pressure drop of each tube is balance by adjusting refrigerant mass flow rate of each tube in the evaporator, after all the flat tubes are calculated. 2.1.3. Finned-tube evaporator model A tube-by-tube incremental finned-tube evaporator model was developed for the air-source heat pump air-conditioning system model to compare the performance of the microchannel evaporator and the finned-tube evaporator under frosting conditions. Frosted, wet, and dry fins are considered in the model to deal with different air entering conditions. The heat transfer and pressure drop of each tube in finned-tube evaporators are calculated by solving the convention equations. Each tube is divided into several elements along the refrigerant flow direction. Each element can be treated as a cross-flow heat exchanger, as shown in Fig. 3. The governing equations of the elements in the finned-tube evaporator model are the same as those of the elements in the microchannel evaporator model: the refrigerant mass conservation equation, Eq. (8), the refrigerant momentum conservation equation, Eq. (9), the refrigerant energy conservation equation, Eq. (10), the heat transfer equations, Eqs. (11)–(13), the air water vapor mass conservation equation, Eq. (14). But the calculations of the finned-tube heat exchanger geometries in each element were
different from those of the microchannel heat exchanger geometries, such as the heat transfer area and the surface ratio. Following correlations are used to calculate the heat transfer coefficients and friction factors in the governing equations of the finned-tube evaporator model. Gungor and Winterton correlation [29] is used to calculate the refrigerant-side two-phase heat transfer coefficient. Pierre correlation [30] is used to calculate the refrigerant-side two-phase pressure drop. Wang and Du correlation [31] is used to calculate the air-side heat transfer and pressure drop. The algorithm of the finned-tube evaporator model is as follow: First, the conservation equations of each element in one tube are solved along the refrigerant flow direction. After all the tubes in one circuit are calculated along the refrigerant flow direction, the calculation move onto the next circuit until all the circuits in the finned-tube evaporator are all solved. Second, the refrigerant pressure drop of each circuit is balance by adjusting refrigerant mass flow rate of each circuit in the finned-tube evaporator until the refrigerant pressure drop of each circuit is the same. 2.2. Compressor model An ARI polynomial equation compressor model [32] is used in the heat pump system model. The compressor mass flow rate is:
X ¼ C1 þ C2 Tse þ C3 Tsc þ C4 Ts2e þ C5 Tse Tsc þ C6 Ts2e þ C7 Ts3e þ C8 Tsc Ts2e þ C9 Tse Ts2c þ C10 Ts3c
ð16Þ
where X can represent the compressor power input or the refrigerant mass flow rate. The coefficients in Eq. (16) are provided by the compressor manufacturer. 2.3. Condenser model A parallel-flow shell-and-tube refrigerant-to-water condenser is used in the tested heat pump system. Refrigerant is inside the tube and water is on the shell side. The condenser model is divided into three regions: desuperheated region, two-phase region, and subcooled region. The conservation equations of each region were solved to calculate the heat transfer and the leaving states of refrigerant and water. Refrigerant mass conservation equation:
mr;cond;in ¼ mr;cond;out
ð17Þ
Refrigerant energy conservation equation:
Q cond ¼ mr;cond;in hr;cond;in hr;cond;out
ð18Þ
1191
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
Heat transfer from the refrigerant to the water:
Q cond ¼
Water energy conservation equation:
Awat ðT r T wat Þ 1=hwat þ Rm þ ðAwat =Ar Þ=hr
Q cond ¼ mwat;in cp; wat T wat;out T wat;in
ð19Þ
Start
Input system geometries
Input air entering state: T db,in and Twb,in Input water entering state: Twat,in and V wat,in Initialize V a, Ts c, Ts e, and ΔT sc
Calculate compressor model with Tsc, Ts e, and ΔTsh to find m and Tdis
Adjust Ts c
Calculate condenser model with Ts c, T dis , and ΔT sh to find m and ΔT sc,cal
N
ΔTsc,cal = ΔT sc ? Increase time
Y Calculate valve model with Tsc and ΔTsc to find h val,out
Adjust Tse
Calculate evaporator model with Va, Ts e, h val,out , and m to find ΔT sh,cal and Δp a
Adjust ΔTsc
Adjust V a
Calculate fan model with Va to find Δp a,cal
N
Δpa,cal = Δp a ? Y N
ΔT sh,cal = ΔT sh ? Y Sum charge Mcal
N
Mcal = M ? Y Frost layer growth
End ?
N
End Fig. 4. Flowchart of heat pump system model.
ð20Þ
1192
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
Cavallini et al. correlation [33] is used to calculate the refrigerant two-phase heat transfer coefficient. Gnielinski correlation [34] is used to calculate the water heat transfer coefficient. 2.4. Expansion valve model The expansion process is considered as an isenthalpic process.
hin ¼ hout
ð21Þ
Using an expansion valve, the refrigerant superheat of at evaporator outlet is assumed constant. 2.5. Fan model A polynomial fan curve was integrated in the heat pump system model to evaluate the effect of the frost growth on the heat pump performance. With the growth of the frost layer, the air-side pressure drop increasing. At each time step, the air flow rate was balanced with the air-side pressure drop of the evaporator and the air pressure rise of the fan, Eq. (22):
dprise ¼ C 0 þ C 1 V þ C 2 V 2
channel heat exchangers circuit. When the finned-tube circuit was under testing, the microchannel circuit was off and vice versa. The microchannel heat exchangers in the heat pump system were designed capacity-equivalent to the conventional finnedtube heat exchanger in the same heat pump system. In the cooling mode test, the microchannel heat exchanger system obtained the same level performance. The geometries of the microchannel heat exchanger and the finned-tube heat exchanger are shown in Tables 1 and 2, respectively. As the air-side pressure drop of the microchannel heat exchanger is lower than that of the finned-tube heat exchanger, the fan power of microchannel heat exchanger system is less than that of the finned-tube heat exchanger system. Hence, the COP (defined as system capacity over power) of the microchannel heat exchanger system is higher than that of the finned-tube system, shown in Table 3. Fig. 6 shows the test facility
ð22Þ
where the coefficients C0, C1, and C2 are reduced with the fan test data. 2.6. System algorithm The heat pump system model is developed for system performance (heating mode) prediction. The inputs of the heat pump system model are divided into two parts: (1) the heat pump system configurations, including the refrigerant type and charge, the compressor type, the condenser geometries, the evaporator geometries, the fan type, the suction superheat set point and the fan speed; (2) the operating conditions, including the air dry-bulb and wet-bulb temperature, the water flow rate, and water entering temperature. The component models of the heat pump system model are solved sequentially. The heat pump system algorithm is shown in Fig. 4. The air flow rate (Va), saturation condensing temperature (Tsc), saturation evaporating temperature (Tse), and condenser leaving subcooling (DTsc) are initially guessed. Va, Tsc, Tse, and DTsc are iterated to find the fan pressure rise, refrigerant mass flow rate, suction superheat, and refrigerant charge in the system model at one time step. After the system is solved at one time step, the frost growth during the time interval is calculated for the next time interval. 3. Experimental results and model validation The microchannel heat exchangers in this study were designed capacity-equivalent to the conventional finned-tube heat exchangers in the cooling model, so that both the microchannel heat pump and the finned-tube heat pump could provide almost same cooling capacity. The heating performance of the microchannel heat pump and the finned-tube heat pump were compared experimentally and numerically under rating condition and frosting condition.
Fig. 5. Tested heat pump.
Table 1 Characteristics of the microchannel heat exchanger. Characteristics of the microchannel heat exchanger Face area (m2) Fin density (fins/in.) Tube depth (mm) Tube length (mm) Pass number Port inner diameter (mm) Port number Fin height (mm) Fin depth (mm) Louver angle (°) Header inner diameter (mm)
2.12 20 25 1030 1 0.64 32 7.6 0.1 25 27
Table 2 Characteristics of the finned-tube heat exchanger. Characteristics of the finned-tube heat exchanger
3.1. Heat pump system and test apparatus The tested heat pump in this study, shown in Fig. 5, has two separated refrigerant circuits. All the components in the two refrigerant circuits are the same, except the coils. One refrigerant circuit contains the finned-tube heat exchangers, while the other contains the microchannel heat exchangers. In the test, the finned-tube heat exchangers circuit is compared with the micro-
Face area (m2) Fin density (fins/in.) Tube length (mm) Row number Tube number Tube outer diameter (mm) Circuit number Row pitch (mm) Tube pitch (mm)
2.18 17 1950 3 44 10 22 19 25.4
1193
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197 Table 3 Performance in cooling mode.
Tested COP
Finned-tube system
Microchannel system
99.8 2.77
99.2 2.86
Water
Test chamber
M
T
Simulated COP
2.6
Heating COP
Capacity (kW) COP
2.8
2.4
2.2
T
2.0 P,T
P,T
Condenser
1.8 0
30
60
90
Time (min)
Compressor EXV
Fig. 8. Heating COP of the finned-tube system (2/1 °C).
Air
Tdb , Twb
P,T
P,T
Reversing valve
Evaporator
Fig. 6. Schematic of the tested system and testing facility.
Table 4 Performance in heating mode.
Capacity (kW) COP
Finned-tube system
Microchannel system
103.8 2.69
114.1 2.73
for the heating mode tests of both microchannel and finned-tube systems. The ambient dry-bulb and wet-bulb temperature in the test chamber, the water flow rate, water entering and leaving temperature, compressor discharge pressure and temperature, condenser leaving pressure and temperature, evaporator leaving pressure and temperature, compressor suction pressure temperature, and system power are measured. The air entering the test chamber is controlled to keep the dry-bulb and wet-bulb temperatures in the test chamber for the tested heat pump systems. The heat pump system entering water temperature is also controlled to keep the entering water temperature during the test. The uncertainties for pressure, temperature and flow rate measurements were about ±8 kPa, ±0.5 °C and ±3 105 m3/s, respectively. 3.2. Experimental results and model results The microchannel heat pump system and the finned-tube heat pump system were tested under the rating condition
120
110
Tested capacity
100
100
Heating capacity (kW)
Heating capacity (kW)
Simulated capacity
90
80
80
60
Tested capacity Uniform distribution model Maldistribution model
40
70
20
0
30
60
90
Time (min) Fig. 7. System heating capacity of the finned-tube system (2/1 °C).
0
10
20
30
Time (min) Fig. 9. System heating capacity of the microchannel system (2/1 °C).
1194
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
(dry-bulb temperature 7 °C and wet-bulb temperature 6 °C) and the frost condition (dry-bulb temperature 2 °C and wet-bulb temperature 1 °C), respectively. Table 4 shows the performance comparison of the microchannel system and the finned-tube system under the rating condition. The microchannel heat pump system could provide more heating capacity than the finnedtube heat pump system. The microchannel system model and the finned-tube system model were calibrated with the tested results under the rating condition, respectively. The calibrated microchannel system model and the finned-tube system model were used to predict the system performance under frosting conditions, respectively. The finned-tube heat pump system was tested as the baseline in the comparison with the microchannel evaporator heat pump system under frosting condition (2/1 °C). Figs. 7 and 8 show the comparison of the test and the predicted results of the system capacity and COP of the finned-tube system, respectively. The heating time
Heating COP
3.0
Tested COP Uniform distribution model Maldistribution model
2.5
2.0
of the finned-tube heat pump system was about 80 min. In the beginning several minutes, the system capacity and COP swung, as the heat pump finished the defrost cycle and the 4-way reversing valve turned the heat pump into heating mode. The 4-way reversing valve caused the refrigerant pressure changing sharply. The simulated heating time is slightly longer than the tested heating time. In Fig. 7, the finned-tube system frost model showed good accuracy on the system capacity. The simulated heating time was little longer than the tested heating time. In Fig. 8, the simulated finned-tube system COP is a little lower than the tested COP. The microchannel heat pump system was tested under frosting condition (2/1 °C). Figs. 9 and 10 showed the tested system capacity and COP of the microchannel heat pump system, respectively. The heating time of the microchannel heat pump system was much shorter than that of the finned-tube heat pump system under frosting condition. In Figs. 9 and 10, the results of the refrigerant-side uniform distribution model and the maldistribution model were compared with the test data. Fig. 11 shows the frost distribution profile on the microchannel evaporator of the tested heat pump. The frost on the microchannel evaporator in the test was not uniform because of the inertia effect in the microchannel evaporator header. The microchannel evaporator manifolds distribution model developed by Jin et al. [26] was used to simulate the microchannel heat pump system. The maldistribution model showed good agreement with the test results in Figs. 9 and 10. The simulated heating time of the uniform distribution system was about 100% longer than that of the maldistribution case. The uniform distribution model simulation results showed the performance and heating time of the microchannel heat pump could be both improved if the refrigerant distribution of the microchannel evaporators was uniform. The frosting condition test results of the microchannel system and the finned-tube system showed that the heating time of the microchannel system was less than one-fifth of that of the finned-tube system. The microchannel heat pump system needed much more defrosting times than the finned-tube heat pump system for a same frosting duration.
1.5 0
10
20
30 4. Parametric study
Time (min) Fig. 10. Heating COP of the microchannel system (2/1 °C).
The heating time of the microchannel heat pump system was much less than that of the finned-tube heat pump system. The
Fig. 11. Frost on the microchannel evaporator (2/1 °C).
1195
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
parametric study of the microchannel heat pump system was used to investigate the factors on the system performance under frosting conditions to improve the system heating time. The frosting condition in the parametric study was 2 °C dry-bulb temperature and 1 °C wet-bulb temperature. In the simulation, the inputs of the heat pump model are as follow: the air dry-bulb temperature is 2 °C, the wet-bulb temperature is 1 °C, the water entering temperature is 40 °C, and the entering water flow rate is 20 m3/h. 4.1. Effects of the fan speed The fan speed effect of the microchannel heat pump system on frost performance was investigated with the microchannel system model. Figs. 12 and 13 show the microchannel system capacity and COP of the fan speed changing ±10%. System capacity and heating time were improved by increasing the fan speed, as the air volu-
110
Designed fan speed +10% fan speed -10% fan speed
100
Designed fan speed +10% fan speed -10% fan speed
100
Heating capacity (kW)
Heating capacity (kW)
110
metric flow rate increases with the fan speed. But the high fan speed COP is lower than the low fan speed COP at initial part of the frosting process, as high fan speed needs more power. But the low fan COP drops faster than the high fan speed COP, as the frost on the microchannel evaporator increases faster for the low fan speed system. The fan speed effect of the finned-tube heat pump system on frost performance was compared with that of the microchannel system. The simulated finned-tube system capacity and COP of the fan speed changing ±10% were shown in Figs. 14 and 15. Same trends of the system capacity and COP of the microchannel system were found. For both the microchannel and finned-tube heat pumps, the fan speed affected the system performance and the heating time with similar trends. High fan speed could increase the heating time. But for the microchannel heat pump with the refrigerant maldistribu-
90
80
90
80
70
60
70 0
5
10
15
0
20
20
Time (min)
Designed fan speed +10% fan speed -10% fan speed
2.6
80
100
Fig. 14. Fan speed effect on the finned-tube system capacity.
2.8
2.8
60
Time (min)
Fig. 12. Fan speed effect on the microchannel system capacity.
2.4 2.2
Designed fan speed +10% fan speed -10% fan speed
2.6
Heating COP
Heating COP
40
2.4
2.2
2.0
2.0
1.8
1.8
1.6 0
5
10
15
Time (min) Fig. 13. Fan speed effect on the microchannel system COP.
20
0
20
40
60
80
Time (min) Fig. 15. Fan speed effect on the finned-tube system COP.
100
1196
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
tion, the heating time after increasing the fan speed was still much less than that of the finned-tube heat pump. The heating time of the microchannel heat pump with refrigerant maldistribution, shown in Figs. 14 and 15, is still much less than that of the heat pump with uniform distribution, shown in Figs. 9 and 10, even if the fan speed has been increased by 10%. The refrigerant maldistribution in the microchannel heat exchangers affected the system performance and heating time much than the fan speed.
4.2. Effects of the fin density The fin density effect of the heat pump system on system capacity and COP are shown in Figs. 16 and 17, respectively. When the refrigerant maldistribution happened in the microchannel evaporator manifold, the fin density changing did not significantly affect the system performance. From the simulation results in Figs. 16
and 17, the lower and higher fin density microchannel systems provide almost the same system capacity and COP as the designed fin density microchannel system. The fin density effect of the finned-tube heat pump system on frost performance was compared with that of the microchannel system. Figs. 18 and 19 show the simulated finned-tube system capacity and COP of the fan density changing ±10% and ±20%. Lower fin density design can improve the heating time of the finnedtube heat pump with less negative effect on the finned-tube system capacity and COP. The fin density effect on the microchannel system was much different from that on the finned-tube system. Lower fin density of the finned-tube heat exchangers could improve the finned-tube system heating time. When refrigerant maldistribution occurred in the microchannel heat exchangers, changing the fin density of the microchannel heat exchangers couldn’t affect the system performance and the heating time.
110
100
Heating capacity (kW)
Heating capacity (kW)
100
90
Designed fin density +20% fin density +10% fin density -10% fin density -20% fin density
80
70
60
0
5
10
90
80
Designed fin density +10% fin density -10% fin density
70
60
15
0
Time (min)
30
60
90
Time (min)
Fig. 16. Fin density effect of the microchannel evaporator on system capacity.
2.6
2.6
2.4
2.4 Heating COP
Heating COP
Fig. 18. Fin density effect of the finned-tube evaporator on system capacity.
2.2
Designed fin density +20% fin density +10% fin density -10% fin density -20% fin density
2.0
1.8
2.2
2.0
Designed fin density +10% fin density -10% fin density
1.8
1.6
1.6
0
5
10
15
Time (min) Fig. 17. Fin density effect of the microchannel evaporator on system COP.
0
30
60
90
Time (min) Fig. 19. Fin density effect of the finned-tube evaporator on system COP.
L.-L. Shao et al. / Applied Energy 87 (2010) 1187–1197
5. Conclusions The performance of microchannel evaporator and finned-tube evaporator in the heat pump system under frosting conditions has been compared in this study. Quasi-steady-state microchannel and finned-tube heat pump system models for frost conditions have been developed, respectively. The refrigerant maldistribution effect has been considered in the microchannel heat pump model. Both test results and simulation results show that the frost period of the tested finned-tube heat pump system was much longer than that of the microchannel heat pump system. The parametric study of the microchannel heat pump system has been conducted to investigate the factors for the performance of the microchannel evaporator under frosting conditions. It reveals that the refrigerant distribution in the microchannel heat exchangers is the key factor to the microchannel heat pump performance during the frost period.
Acknowledgements The authors would like to thank Mr. Jing Li and Dr. Xinghua Huang from Carrier China R&D Center for their efforts in the experiments.
References [1] Comakli K, Simsek F, Comakli O, Sahin B. Determination of optimum working conditions R22 and R404A refrigerant mixtures in heat-pumps using Taguchi method. Appl Energy 2009;86(11):2451–8. [2] Huang D, Li QX, Yuan XL. Comparison between hot-gas bypass defrosting and reverse-cycle defrosting methods on an air-to-water heat pump. Appl Energy 2009;86(9):1697–703. [3] Wang ZY, Wang XM, Dong ZM. Defrost improvement by heat pump refrigerant charge compensating. Appl Energy 2008;85(11):1050–9. [4] Barbosa Jr JR, Melo C, Hermes CJL, Waltrich PJ. A study of the air-side heat transfer and pressure drop characteristics of tube-fin ‘no-frost’ evaporators. Appl Energy 2009;86(9):1484–91. [5] Gholap AK, Khan JA. Design and multi-objective optimization of heat exchangers for refrigerators. Appl Energy 2007;84(12):1226–39. [6] Hayashi Y, Aoki A, Adachi S, Hori K. Study of frost formation based on a theoretical model of the frost layer. Heat Transfer Jpn Res 1977;6(3):79–94. [7] Yonko JD, Sepey CF. An investigation of the thermal conductivity of frost while forming on a flat horizontal plate. ASHRAE Trans 1967;73:I.1.1–I.1.10. [8] Yang DK, Lee KS. Dimensionless correlations of frost properties on a cold plate. Int J Refrig 2004;27:89–96. [9] Yang DK, Lee KS. Modeling of frosting behavior on a cold plate. Int J Refrig 2005;28:396–402.
1197
[10] Xia Y, Jacobi AM. An exact solution to steady heat conduction in a twodimensional slab on a one-dimensional fin: application to frost heat exchangers. Int J Heat Mass Transfer 2004;47:3317–26. [11] Seker D, Karatas H, Egrican N. Frost formation on fin-and-tube heat exchangers. Part I – Modeling of frost formation on fin-and-tube heat exchanger. Int J Refrig 2004;27:367–74. [12] Seker D, Karatas H, Egrican N. Frost formation on fin-and-tube heat exchangers. Part II–Experimental investigation of frost formation on fin-andtube heat exchangers. Int J Refrig 2004;27:375–7. [13] Chen H, Thomas L, Besant RW. Fan supplied heat exchanger fin performance under frosting conditions. Int J Refrig 2003;26:140–9. [14] Yao Y, Jiang YQ, Deng SM, Ma ZL. A study on the performance of the airside heat exchanger under frosting in an air source heat pump water heater/chiller unit. Int J Heat Mass Transfer 2004;47:3745–56. [15] Yin JM, Bullard CW, Hrnjak PS. R-744 gas cooler model development and validation. Int J Refrig 2001;24:692–701. [16] Asinari P, Cecchinato L, Fornasieri E. Effect of thermal conduction in microchannel gas coolers for carbon dioxide. Int J Refrig 2004;27:577–86. [17] Shao LL, Yang L, Zhang CL, GU B. Numerical modeling of serpentine microchannel condensers. Int J Refrig 2009;32:1162–72. [18] Kim MH, Bullard CW. Development of a microchannel evaporator model for CO2 air-conditioning system. Energy 2001;26:931–48. [19] Yun R, Kim YC, Park C. Numerical analysis on a microchannel evaporator designed for CO2 air-conditioning systems. Appl Therm Eng. 2007;27:1320–6. [20] Kim JH, Groll EA. Performance comparisons of a unitary split system using microchannel and fin-tube outdoor coils. ASHRAE Trans 2003;109(2):219–29. [21] Lemmon EW, Huber ML, McLinden MO. NIST reference fluid thermodynamic and transport properties – REFPROP. NIST standard reference database 23version 7.1; 2006. [22] Kondepudi SN, O’Neal DL. Performance of finned-tube het exchangers under frosting conditions: Part I – Simulation model. Int J Refrig 1993;16(3):175–80. [23] Eckert ER, Drake RM. Analysis of heat and mass transfer. New York: McGrawHill; 1972. [24] Yun R, Heo JY, Kim YC. Evaporative heat transfer and pressure drop of R410A in microchannels. Int J Refrig 2006;29:92–100. [25] Yan YY, Lin TF. Evaporation heat transfer and pressure drop of refrigerant R134a in a small pipe. Int J Heat Mass Transfer 1998;41:4183–94. [26] Jin DH, Hwang Y, Radermacher R. Refrigerant distribution in evaporator manifolds. ASHRAE 1260-TRP; 2006. [27] Kim MH, Bullard C. Air-side thermal hydraulic performance of multi-louvered fin aluminum heat exchangers. Int J Refrig 2002;25:390–400. [28] Kim MH, Bullard C. Air-side thermal performance of micro-channel heat exchangers under dehumidifying conditions. In: Tree DR, (Ed.), Proceedings of the 2000 International refrigeration conference, West Lafayette (IN): Purdue University; 2000. p. 119–26. [29] Gungor KE, Winterton RHS. A general correlation for boiling in tubes and annuli. Int J Heat Mass Transfer 1986;19:351–8. [30] Pierre B. Flow resistance with boiling refrigerants. Part I. ASHRAE J 1964:58–65. [31] Wang CC, Du JY. A heat transfer and friction correlation for wavy fin and tube heat exchangers. Int J Heat Mass Transfer 1999;42:1919–24. [32] ARI Standard 540-99, Performance rating of positive displacement compressors and compressor units; 1999. [33] Cavallini A, Censi G, Del Col D, Doretti L, Longo GA, Rossetto L. In-tube condensation of halogenated refrigerants. ASHRAE Trans 2002;108(1):146–61. [34] Gnielinski V. New equations for heat and mass transfer in turbulent pipe and channel flows. Int Chem Eng 1976;16:359–68.