Applied Thermal Engineering 26 (2006) 1593–1603

www.elsevier.com/locate/apthermeng

Experimental evaluation of a domestic refrigerator

working with LPG
M. Fatouh *, M. El Kafafy
Department of Mechanical Power Engineering, Faculty of Engineering, Ibrahiem Abd El-Razik, Mattaria,
Masaken El-Helmia P.O., Cairo 11718, Egypt

Received 6 April 2005; accepted 21 November 2005

Available online 18 January 2006

Abstract

Liquefied petroleum gas (LPG) of 60% propane and 40% commercial butane has been tested as a drop-in substitute for R134a in a
single evaporator domestic refrigerator with a total volume of 10 ft3 (0.283 m3). Continuous running and cycling tests were performed on
that refrigerator under tropical conditions using different capillary tube lengths and various charges of R134a and LPG.
Continuous running and cycling results showed that R134a with a capillary tube length of 4 m and charge of 100 g or LPG with cap-
illary tube lengths from 4.0 to 6.0 and charge of 50 g or more satisfy the required freezer air temperature of 12 C. The lowest electric
energy consumption was achieved using LPG with combination of capillary tube length of 5 m and charge of 60 g. This combination
achieved higher volumetric cooling capacity and lower freezer air temperature compared to R134a.
Experimental results of the refrigerator using LPG of 60 g and capillary tube length of 5 m were compared with those using R134a of
100 g and capillary tube length of 4 m. Pull-down time, pressure ratio and power consumption of LPG refrigerator were lower than those
of R134a refrigerator by about 7.6%, 5.5% and 4.3%, respectively. Also, actual COP of LPG refrigerator was higher than that of R134a
by about 7.6%. Lower on-time ratio and energy consumption of LPG refrigerator by nearly 14.3% and 10.8%, respectively, compared to
those of R134a refrigerator were achieved. In conclusion, the proposed LPG seems to be an appropriate long-term candidate to replace
R134a in the existing refrigerator, except capillary tube length and initial charge.
 2005 Elsevier Ltd. All rights reserved.

Keywords: LPG; Continuous running test; Cycling test; Capillary tube

1. Introduction of performance characteristics while others are related to

sizing of capillary tube for suitable drop-in replacements.
Domestic refrigerators working with R12 consume Several researchers have examined R134a, propane
about 12% of the total residential energy budget annually (R290), iso-butane (R600a), cyclopropane (RC270) and
[1]. Unfortunately, R12 was considered as a harmful fluid dimethyl ether (DME) as working fluids to replace R12
to the environment [2]. As a result, considerable efforts in domestic refrigerators. Maclaine-cross and Leonardi
have been done to identify R12 replacements among chlo- [3] confirmed that hydrocarbon refrigerants have environ-
rine-free pure refrigerants, mixture of environmentally mental advantages and are safe if they used in small quan-
friendly refrigerants, natural refrigerants and hydrocarbon tities, thereby R290 can replace R22 and hydrocarbon
refrigerants for energy saving and environmental protec- mixtures can replace both R12 and R134a in applications
tion. Some of these investigations dealt with evaluation using positive displacement compressors. Rogstam [4]
showed that RC270 is more energy efficient than R12
and R600a in household refrigerators. Operating parame-
ters for R12 and DME were compared by Kustrin and Por-
*
Corresponding author. Tel.: +20 249 376 60; fax: +20 263 323 98. edos [5] who concluded that DME can be a very suitable
E-mail address: [email protected] (M. Fatouh). replacement for R12.

1359-4311/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.applthermaleng.2005.11.026
1594 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603

Nomenclature

COP coefficient of performance (–) ISO International Standard Organization

h specific enthalpy (kJ/kg) LPG liquefied petroleum gas
L capillary tube length (m) LPS low pressure side, Fig. 1
m mass flow rate (kg/s) R134a 1,1,1,2-tetrafluoroethane
N number of revolution per second TP thermostat position
p pressure (MPa)
P power (kW) Subscripts
PR pressure ratio (–) 1–5 refers to Fig. 1
q cooling effect (kJ/kg) act actual
Q heat transfer rate (kW) amb ambient
Qvol volumetric cooling capacity (kJ/m3) c cooling
RH relative humidity (%) com compressor
T temperature (C) con condenser
V volume (m3) elec electrical
w compressor specific work (kJ/kg) eva evaporator
r refrigerant
Abbreviations st stroke
F flow meter, Fig. 1 th theoretical
HPS high pressure side, Fig. 1

Experimental performance characteristics of domestic or 25% propane, 57.5% butane and 17.5% iso-butane were
refrigerators operating with propane/iso-butane mixture used as possible alternative replacements to R12. The
were evaluated by many investigators. Agarwal et al. [6] authors showed that the hydrocarbon mixture with 50%
and Richardson and Butterworth [7] reported that the pro- propane, 38.3% butane and 11.7% iso-butane is the most
pane/iso-butane mixture at 50% mass fraction of propane suitable alternative refrigerant. Akash and Said [13]
yields higher COP, lower charge and lower compressor reported that LPG compares very well to R12. Their results
shell temperature as compared to R12. Liu et al. [8] con- showed that LPG mass charge of 80 g had the best results
firmed that energy savings of up to 6% were achieved with when used in the same refrigerator.
a blend of 70% propane and 30% n-butane compared to In addition, sizing of capillary tubes commonly used as
R12. Maclaine-cross [9] stated that R134a leakage and ser- expansion devices in household refrigerators and freezers
vice emissions yield 15% increase in Total Equivalent working with R12 alternatives was experimentally carried
Warming Impact (TEWI). He concluded that R290/ out by few investigators. Melo et al. [14] and Driessen
R600a mixture with 55% propane matches the performance et al. [15] conducted several tests using R12, R134a and
of R134a. Experimental results of Jung et al. [10] indicated R600a at different condensing pressures and degrees of sub-
that the mixture of propane and iso-butane with 60% mass cooling. Liu et al. [16] tested the hydrocarbon mixtures as
fraction of propane has higher COP, faster cooling rate, drop-in substitutes for R12 in domestic refrigerators. They
shorter compressor on-time and lower compressor dome reported that a propane/n-butane blend of 70% propane
temperatures than R12. Such results confirmed the success- with charge of 70 g and an additional capillary tube length
ful use of the propane/butane mixture as an alternative to of 5 ft (1.52 m) in a domestic refrigerator/freezer yields
R12 in domestic refrigerators. lower energy consumptions compared to R12. Fatouh
Liquefied petroleum gas (LPG), which consists of pro- and El-Kafafy [17] reported that a propane/iso-butane/n-
pane (R290), iso-butane (R600a) and n-butane (R600), butane mixture with 60% propane mass fraction is an ade-
was also examined as a working fluid in domestic refriger- quate drop-in refrigerant for R134a.
ators. Alsaad and Hammad [11] examined LPG of 24.4% Literature survey revealed that many investigators have
propane, 56.4% n-butane and 17.2% iso-butane as a examined R134a, R600a, cyclopropane, dimethyl ether,
drop-in refrigerant for R12 in domestic refrigerators. They propane/iso-butane mixture and a blend of R290/R600a/
reported that the evaporator temperature of 15 C and R600 known as liquefied petroleum gas (LPG) as working
coefficient of performance of 3.4 were achieved at a con- fluids to replace R12 in unmodified domestic refrigerators.
denser temperature of 27 C when LPG was used. Ham- Among the examined refrigerants, R134a has been
mad and Alsaad [12] investigated the performance accepted as a working fluid in domestic refrigerators but
parameters of an unmodified domestic refrigerator when it has high global warming potential (GWP). In order to
100% propane; 75% propane, 19.1% butane, 5.9% iso- avoid its green house effect, propane/iso-butane mixtures
butane; 50% propane, 38.3% butane, 11.7% iso-butane; or LPG as drop-in replacement in existing domestic
 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603 1595

refrigerator were tested by few researchers under normal or inner shells to minimize heat gain. Specifications of the
subtropical conditions. However, LPG of 60% propane evaporator, compressor, condenser and capillary tube are
and 40% commercial butane fraction seems to be an appro- given in Table 1 [18]. Filter-drier is installed before the cap-
priate drop-in replacement for R134a. illary tube to absorb the moisture which may exist in the
The objective of the present investigation is to examine refrigerant circuit. As the refrigerant is condensed in the
LPG of propane/commercial butane by 60%/40% as a condenser, it flows through the high-side filter-drier into
drop-in candidate for R134a in the existing domestic refrig- a capillary tube attached to a section of the suction line.
erator under tropical operating conditions. This requires This provides a heat exchange between the capillary tube
performance characteristics evaluation for LPG refrigera- and the suction line.
tor. Thus, a single door, manual defrost and tropical class This domestic refrigerator was instrumented with eleven
refrigerator with total volume of 0.283 m3, which was orig- Type-T thermocouples, two pressure gauges, a refrigerant
inally manufactured to work with R134a, was equipped flow meter, a watt meter and watt-hour meter. Five of these
with necessary instrumentations. Then, influences of capil- thermocouples were used to measure the refrigerant tem-
lary tube length and LPG charge on the performance char- peratures as shown in Fig. 1. Other five thermocouples,
acteristics of the domestic refrigerator were investigated which were attached to five copper cylinders of 25.4 mm
through continuous running and cycling tests. Finally, per- diameter and 25.4 mm height [19], were installed at differ-
formance characteristics and capillary tube size of both ent levels in the freezer (two) and cabinet (three) according
LPG and R134a were compared. to ISO specifications [19] to monitor refrigerated space air
temperatures. Another thermocouple was used to measure
2. Experimental test rig and procedures the dry bulb air temperature inside the climatic chamber.
All thermocouples were well calibrated using a constant
Fig. 1 shows a schematic diagram of a single door, man- temperature bath. The pressure gauges with resolution of
ual defrost and tropical class refrigerator with total and 0.01 bar were used to indicate the pressure at suction and
freezer volumes of 10 ft3 and 2.1 ft3, respectively, which discharge of the compressor in the refrigerant circuit. These
was originally manufactured to work with R134a. It con- pressure gauges were calibrated using a dead weight tester.
sists of a cabinet, an evaporator, a compressor, a condenser A digital watt meter which has 0.01 W resolution and a
and a capillary tube. The cabinet was made of pressed steel watt-hour meter were used to record the instantaneous
with smooth and water proof outside shell. Expanded poly-
styrene panels were installed between the outer and the Table 1
Specifications of main components [18]
Evaporator
Type Roll bond-type
Tamb = 43 °C Material Aluminum
T5 RH = 50% Length of evaporator panel 1184 ± 1 mm
Thickness of evaporator panel 1.2 ± 1 mm
Evaporator
Passage way surface of panel 0.35 m2
Volume capacity of panel 0.024 m3
T4 Fins Outlet tube diameter 6.5 mm
Maximum working pressure 20 bar
Condenser
Insulation Type Wire-on-tube
Tube material Steel
Condenser Tube diameter (inner/outer) 4.76/6.18 mm
Outer surface area 0.724 m2
Maximum working pressure 40 bar
Compressor
Type Hermetic reciprocating
Refrigerant R134a
T3 Number of cylinders One
Stroke volume 6.64 cm3
Capillary tube F Filter Power supply 1 phase, 220–240 V
T2 , p2 Current 6 A (starting), 1.2 (running)
T1, p1 Frequency 50 Hz
HPS Speed 2900 rpm
LPS Capillary tube
Hermetically sealed compressor Material Copper
Inner diameter 0.78 ± 0.02 mm
Outer diameter 2.0 ± 0.05 mm
Climatic chamber
Length 4.0 ± 0.03 m
Initial charge 100 g
Fig. 1. Schematic diagram of the experimental set-up.
1596 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603

power input to the refrigerator and the integrated energy Theoretical coefficients of performance (COPth) and
consumption, respectively. A precalibrated refrigerant flow actual coefficients of performance (COPact) can be
meter with resolution of 0.01 kg/h was used to indicate expressed as
refrigerant mass flow rate. A programmable charging Qeva
meter of 1 g resolution was used to determine the amount COPth ¼ ð5aÞ
P com
of refrigerant charge input to the system during the charg-
ing process. A digital data logger along with a personal Q
COPact ¼ eva ð5bÞ
computer was used to collect the measured data. P elec
During the entire test period, the refrigerator with closed Pelec is the measured electric power supplied to the com-
door was placed in an environmental chamber in which dry pressor and Pcom is the theoretical power consumed by
bulb air temperature of 43 ± 1 C and relative humidity of the compressor. Pcom can be written as follows:
50% were controlled to obtain reliable data. Moreover,
P com ¼ mr wcom ð6Þ
some random tests were repeated to ensure reproducibility
of the data. The system was allowed to run and readings of wcom is the specific work, which equals enthalpy difference
temperatures, pressure gauges, watt meter, refrigerant flow through the compressor (h2  h1). Specific enthalpy values
meter and energy meter were collected every ten seconds at the inlet and outlet of each component were computed
from starting time of each experiment till steady state of using SUPERTRAPP [20] based on the measured pressures
the operating conditions was reached. Ranges and condi- and temperatures of R134a and LPG. Then, refrigerant
tions of experiments conducted in the present work are cooling capacity (Qeva), and power consumed by the com-
listed in Table 2. pressor (Pcom) were computed using Eqs. (3) and (6),
Measured temperature, pressure and mass flow rate of respectively.
R134a and LPG in addition to electric power and energy Experimental error analysis was conducted according to
were used to evaluate the performance characteristics of the differential method described by Moffat [21]. This anal-
the domestic refrigerator. Parameters such as pressure ysis indicates that the experimental errors in measuring
ratio, volumetric cooling capacity and coefficient of perfor- refrigerant pressure, temperature and mass flow rate are
mance are considered. 1.3%, 0.5% and 0.68%, respectively, while those in measur-
Pressure ratio (PR) is defined as follows: ing electric power and energy are around 1.5% and 2.0%,
p respectively.
PR ¼ con ð1Þ
peva
3. Results and discussion
where pcon and peva are the absolute condensation and
evaporation pressures, respectively. Volumetric cooling
Experimental results were obtained for continuous and
capacity of the refrigerator (Qvol) can be calculated by
cycling operating modes at ambient air temperature of
the following equation:
43 ± 1 C, which simulates tropical operating temperature
Q according to ISO-8187 specification [19]. It can be stated
Qvol ¼ eva ð2Þ
V st N that freezer air temperature of 12 C or less should be
Vst is the stroke volume, N is the number of revolutions per achieved to accept the domestic refrigerator according to
second and Qeva is the cooling capacity which can be writ- ISO standard for this refrigerator class.
ten as
Qeva ¼ mr qeva ð3Þ 3.1. Continuous operating results

where mr is the measured refrigerant mass flow rate and Firstly, baseline continuous operating tests were carried
qeva is the cooling effect of the evaporator, i.e. (h5  h4). out using a capillary tube length of 4 m with various R134a
Hence Eq. (2) can be rewritten as charges. Then, various capillary tube lengths with different
mr qeva initial LPG charges were used to evaluate the performance
Qvol ¼ ð4Þ
V st N of the same refrigerator. In addition, continuous operating
results of LPG were compared against those of R134a
Table 2 baseline.
Ranges and conditions for continuous and cycling experiments
Parameter Range 3.1.1. Operating temperatures
Refrigerant type R134a LPG As stated earlier, freezer and cabinet air temperatures
Capillary tube length (m) 4 4.0, 4.5, 5.0, 5.5 and 6.0 were measured at two and three locations, respectively.
Initial charge (g) 60, 70, 80 30, 40, 50, 60 and 70 Based on these measured temperatures, arithmetic averages
and 100 of instantaneous freezer and cabinet air temperatures were
Out-door conditions calculated for all R134a and LPG experiments. Figs. 2 and
Dry bulb temperature (C) 43 ± 1 43 ± 1 3 show the average freezer and cabinet air temperatures
Relative humidity (%) 50 50 versus operating time for R134a and LPG, respectively.
 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603 1597

Clearly, longer operating time yields lower freezer and cab- 45

inet air temperatures. Since continuous running tests were LPG charge
30g
carried out with closed door, the refrigerator can be consid-
35 40g
ered as a constant volume system. Thus, as operating time
increases, heat absorbed from the internal air volume in 60g
both freezer and cabinet increases causing the air tempera- 25 80g

Freezer air temperature (°C)

tures to decrease as shown in Figs. 2 and 3 for R134a and 100g
LPG, respectively. Similar trends were observed for other
15
LPG charges and capillary tube lengths.
Figs. 2 and 3 reveal that the freezer and cabinet air tem-
peratures for LPG refrigerator have similar trends to those 5
of the R134a refrigerator. It can be seen that a capillary
-5
45
R134a charge
-15
60g
35
70g
80g -25
25 0 1 2 3 4
Freezer air temperature (°C)

100g
(a) Time (h)

15 45
LPG charge
30g
5
40g
35
60g
Cabinet air temperature (°C)

-5 80g
100g
25
-15

-25
15
0 1 2 3 4
(a) Time (h)

45
R134a charge 5
60g
70g
35
80g -5
0 1 2 3 4
Cabinet air temperature (°C)

100g
(b) Time (h)
25
Fig. 3. Effect of LPG charge on the measured air temperatures. (a)
Freezer air temperature and (b) cabinet air temperature.

15 tube length of 4 m with R134a charge of 100 g yields steady

state freezer and cabinet air temperatures of about
16.15 C and 2.5 C, respectively. Fig. 3 shows that free-
5 zer air temperature of 12 C can be achieved using LPG
charge of 40 g or larger. Clearly, instantaneous and steady
state freezer and cabinet air temperatures decrease as initial
charge increases.
-5
Variations of steady state freezer and cabinet air temper-
0 1 2 3 4 atures with LPG charge for different capillary tube lengths
(b) Time (h)
under continuous operating conditions are shown in
Fig. 2. Effect of R134a charge on the measured air temperatures. (a) Fig. 4a and b, respectively. It is seen that steady state free-
Freezer air temperature and (b) cabinet air temperature. zer and cabinet air temperatures decrease when either LPG
1598 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603

charge increases or capillary tube length decreases. It is evi- Thus, these arrangements of charges with their correspond-
dent that capillary tube length of 4 m leads to the lowest ing capillary tube lengths are not suitable to be used. How-
steady state air temperatures for both freezer and cabinet ever, the rest of LPG charge and capillary tube length
over the entire range of the considered LPG charges. For combinations achieve steady state freezer and cabinet air
a given LPG charge, longer capillary tube leads to increase temperatures equal or lower than those of R134a baseline
in flow resistance causing refrigerant mass flow rate to test.
decrease (see Fig. 5b). As a result, heat absorbed from
the refrigerated space decreases and the freezer and cabinet 3.1.2. Pull-down time
air temperatures increase. Clearly, LPG charge of 30 g with Pull-down time is defined as the duration needed to
all capillary tube lengths, LPG charge of 40 g with capillary change the freezer air temperature from tropical ambient
tube longer than 4 m or LPG charge of 50 g with capillary condition (43 C) to the desired final temperature
tube length of 6 m yield higher steady state freezer or cab- (12 C) according to ISO standard for the considered
inet air temperatures than those of R134a (Fig. 4a and b). refrigerator class. With reference to Figs. 2 and 3, one can
Steady state freezer air temperature (°C)

4
L=4.0m
0 L=4.5m

-4 L=5.0m
L=5.5m
-8
L=6.0m
-12

-16

-20
20 30 40 50 60 70 80
(a) LPG charge (g)
Steady state cabinet air temperature (°C)

25
L=4.0m
20 L=4.5m
15 L=5.0m
L=5.5m
10
L=6.0m
5

-5
20 30 40 50 60 70 80
(b) LPG charge (g)

1.6

1.5
Pull-down time (h)

1.4 L=4.0m
L=4.5m
L=5.0m
1.3
L=5.5m
L=6.0m
1.2
30 40 50 60 70 80
(c) LPG charge (g)

Fig. 4. Effect of capillary tube length on air temperatures and pull-down time of the domestic refrigerator during continuous tests. (a) Freezer air
temperature, (b) cabinet air temperature and (c) pull-down time.
 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603 1599

115 500

450
110

Volumetric cooling capacity (kJ/m3)

Discharge temperature (°C) 107.6°C 400
383kJ/m3

105
350

300
100

L=4.0m L=4.0m
250
L=4.5m L=4.5m
95 L=5.0m
L=5.0m
200
L=5.5m L=5.5m
L=6.0m L=6.0m
90 150
30 40 50 60 70 80 30 40 50 60 70 80
(a) LPG charge (g) (c) LPG charge (g)

2.5 230

210
Electric power consumption (kW)

2 190 186 kW
Mass flow rate (kg/h)

170

1.5 150

130
L=4.0m L=4.0m
1 L=4.5m 110 L=4.5m
L=5.0m L=5.0m
L=5.5m 90 L=5.5m
L=6.0m L=6.0m
0.5 70
30 40 50 60 70 80 30 40 50 60 70 80
(b) LPG charge (g) (d) LPG charge (g)

Fig. 5. Characteristics of the domestic refrigerator with LPG charge during continuous test. (a) Discharge temperature, (b) mass flow rate, (c) volumetric
cooling capacity and (d) electric power consumption.

draw a horizontal line at average temperature of 12 C to the pull-down time line of reference test (87 min) is drawn.
intersect the curves at which pull-down time can be deter- It can be seen that pull-down time decreases as either cap-
mined. Hence, pull-down time of about 87 min is achieved illary tube length decreases or LPG charge increases. The
for capillary tube length of 4 m and R134a charge of reason behind this trend is the reduction in refrigerant mass
100 g (baseline test) as indicated in Fig. 2. Freezer air tem- flow rate when either capillary tube length increases or
perature of 12 C can be obtained using capillary tube LPG charge decreases. Clearly, pull-down times for LPG
length of 4 m with LPG charges larger than 30 g as shown charge of 40 g with capillary tube length of 4.5 or longer
in Fig. 3, which indicates that the pull-down times for and LPG of 50 g with capillary tube length of 6 m are
LPG charges of 40, 50, 60, 70, 80 and 100 g are approxi- higher than that of R134a. All capillary tube lengths with
mately 87, 75, 73, 70, 66 and 61 min, respectively. LPG charge greater than 50 g attain pull-down time lower
The effect of capillary tube length and LPG charge on than that of R134a reference test over the entire range of
the pull-down time can be predicted from Fig. 4c on which operating conditions.
1600 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603

3.1.3. Performance characteristics capillary tube length. Fig. 5d presents measured electric
The influence of LPG charge and capillary tube length power consumption versus LPG charge for various capil-
on the performance characteristics of the domestic refriger- lary tube lengths. It is clear that the electric power increases
ator is shown in Fig. 5. as LPG charge increases. This is mainly due to increase in
Discharge temperature is a very important parameter mass flow rate as shown in Fig. 5b. Capillary tube lengths
because of its influence on compressor components and of 6 m and 4 m require minimum and maximum electric
stability of lubricants. Measured discharge temperatures power consumptions, respectively, over the entire range
as a function of LPG charge are presented in Fig. 5a. It of LPG charges. Capillary tube lengths of 4 and 4.5 m with
is seen that as LPG charge increases discharge temperature LPG charge of 60 g or more and capillary tube length of
increases. This can be attributed to the increase in conden- 5 m with charge of 70 g LPG are not suitable because of
sation pressure when LPG charge increases. Clearly, the their higher electric power consumption compared to that
longest capillary tube yields the highest discharge tempera- of R134a. However, other combinations of capillary tube
ture while the shortest capillary tube produces the lowest length and LPG charge can be used because of their low
discharge temperature. This is because longer tubes yield power consumption.
lower mass flow than that of the shorter tubes for a given In order to accept a refrigerant as a drop-in replace-
LPG charge as shown in Fig. 5b. This leads to decrease ment, similar or better volumetric cooling capacity and
in available condenser heat transfer surface area, thereby power consumption should be achieved. Analyzing the
the condensing temperature (pressure) increases causing continuous test results revealed that LPG charge of greater
the discharge temperature to increase with capillary tube than 50 g with all capillary tube lengths achieves pull-down
length for the given LPG charge. Capillary tube length of time lower than that of R134a. Capillary tube length of 5 m
6.0 m with charge of 60 g or 70 g yields higher discharge with LPG charge of 50 g or capillary tube length of 5.5 m
temperature than that of R134a (107.6 C) while other with 60 g produce similar volumetric cooling capacity
combinations of capillary tubes and LPG charges yield and lower power consumption compared to R134a. How-
lower discharge temperature. Thus, using LPG guarantees ever, capillary tube lengths of 4.0, 5.0 and 5.5 m with
lower impact on compressor components and stability of LPG charge of 40, 60 and 70 g, respectively give slightly
lubricants. This means that longer compressor lifetime higher volumetric cooling capacity and slightly lower
can be expected when LPG is used as a drop-in replace- power consumption compared to R134a. These combina-
ment for R134a in the domestic refrigerator. tions of LPG charge and capillary tube length were sub-
Measured mass flow rates of LPG versus LPG charge jected to cycling tests to determine the appropriate
for different capillary tube lengths are shown in Fig. 5b, combination of capillary tube lengths and LPG charges
which shows that the actual mass flow rate increases when to be used as a drop-in refrigerant for R134a.
LPG charge increases or capillary tube lengths decrease.
This is because flow resistance increases as the capillary 3.2. Cycling (on/off) operating results
tube length increases. It should be noted that mass flow
rate for R134a was about 4.0 kg/h, which is more than In order to simulate the actual operating conditions, the
double that of LPG. considered domestic refrigerator was subjected to cycling
Actual volumetric cooling capacity as a function of LPG tests. Percentage of running time (ratio of on-time to total
charge for different capillary tube lengths is presented in time) and energy consumption per day were determined
Fig. 5c. Actual volumetric cooling capacity increases as using the measured data during the cycling (on/off) tests
LPG charge rises. This is mainly due to the increase in mass at five thermostat positions. Based on energy consumption
flow rate of LPG. This figure indicates that the highest and per day, the appropriate combination of LPG charge and
lowest actual volumetric cooling capacity can be obtained corresponding capillary tube length can be determined.
using capillary tube lengths of 4 m and 6 m, respectively. Cycling results of LPG (shown in solid lines) were
Clearly, capillary tube length of 6 m with all LPG charges compared with those of R134a (shown in dash lines) in
gives lower volumetric cooling capacity compared to Fig. 6.
R134a; thereby this length is not suitable. However, volu-
metric cooling capacities of other tube lengths with LPG 3.2.1. Operating temperatures during cycling tests
charge of 60 g or more are higher than that of R134a. Using measured values of instantaneous freezer and
The same volumetric cooling capacity of R134a (383 kJ/ cabinet air temperatures, their average values were calcu-
m3) can be achieved with a capillary tube length of 5 m lated during on/off cycles. Fig. 6a and b presents average
and LPG charge of 50 g or with 5.5 m and 60 g. Using freezer and cabinet air temperatures, respectively, as a
LPG charge of 50 g or higher, the suitable capillary tube function of thermostat position during on/off cycles. It is
lengths are from 4 m to 5.5 m based on same or higher vol- clear that average freezer and cabinet air temperatures
umetric cooling capacity. Thus, the same compressor size are dependent on capillary tube length, LPG charge and
can be used for particular operating conditions. thermostat position. It is seen that LPG charge of 70 g
Electric power consumption is the main criterion to and capillary tube length of 5.5 m yields the lowest air tem-
determine the suitable combination of LPG charge and peratures whereas LPG charge of 40 g and capillary tube
 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603 1601

0 70
L=5.0m, charge=70g LPG
L=5.5m, charge=70g R134a
L=5.0m, charge=60g 60

Average freezer air temperature (°C)

-4 L=5.5m, charge=60g
L=5.0m, charge=50g
50
L=4.0m, charge=40g

On-time ratio (%)

-8 40

30 L=5.0m, charge=70g
L=5.5m, charge=70g
-12
L=5.0m, charge=60g
20
L=5.5m, charge=60g
LPG
L=5.0m, charge=50g
R134a
-16 10
0 1 2 3 4 5 6 0 1 2 3 4 5 6
(a) Thermostat position (c) Thermostat position

20 5
L=5.0m, charge=70g LPG
L=5.5m, charge=70g R134a
L=5.0m, charge=60g
16 4
Average cabient air temperature (°C)

L=5.5m, charge=60g
Energy consumption (kW.h/day)

L=5.0m, charge=50g
L=4.0m, charge=40g
12 3

8 2

L=5.0m, charge=70g
L=5.5m, charge=70g
4 1 L=5.0m, charge=60g
LPG L=5.5m, charge=60g
R134a L=5.0m, charge=50g
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
(b) Thermostat position (d) Thermostat position

Fig. 6. Comparison of some characteristics of the domestic refrigerator working with R134a and LPG during cycling tests. (a) Freezer air temperature, (b)
cabinet air temperature, (c) on-time ratio and (d) energy consumption.

length of 4 m achieves the highest freezer and cabinet air foods for longer period than R134a. Clearly, TP-5 yields
temperatures. Clearly, LPG charge of 40 g with capillary satisfactory freezer and cabinet air temperatures for
tube length of 4 m yields unsatisfactory freezer and cabinet R134a charge of 100 g with capillary tube length of 4 m
air temperatures; thereby it will not be subjected to further and for all considered combinations of LPG charges and
analysis. However, freezer and cabinet air temperatures of capillary tube lengths, except LPG charge of 40 g. Also,
LPG refrigerator are higher than 12 C at both TP-1 and TP-4 can be used for LPG charge of 60g with capillary tube
TP-2. Desired freezer air temperature (12 C) or lower length of 5.0 m whereas TP-4 and TP-3 can be used for
can be obtained at TP-5 for R134a and all considered com- LPG charge of 70 g with capillary tube lengths of 5.5 m.
binations of capillary tube lengths and LPG charges, In conclusion, the desired freezer air temperature of
except 4 m and 40 g. 12 C or better can be obtained at different thermostat
In general, LPG refrigerator yields lower freezer and positions (except both TP-1 and TP-2) using various com-
cabinet air temperatures than those of R134a. This means binations of LPG charges and their corresponding capil-
that LPG (as a drop-in replacement) is capable to preserve lary tube lengths.
1602 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603

3.2.2. On-time ratio during cycling tests Table 3

On-time ratio of refrigerator is defined as the ratio of Comparison between R134a and LPG results
operating time (during complete cycles) to total time of Test Refrigerant
these cycles [19]. Fig. 6c shows the on-time ratio as a func- R134a LPG
tion of thermostat position. The on-time ratio increases Continuous run test
with thermostat position. This is mainly because the tem- Capillary tube length (m) 4.00 5.00
perature inside the freezer is changed to reach lower values Suitable refrigerant charge (g) 100.00 60.00
as thermostat position increases. Thus, the refrigerator Pull-down time (min) 87.00 80.40
Pressure ratio (–) 13.23 12.50
needs more cooling time to satisfy the required lower tem- Discharge temperature (C) 107.60 103.50
perature. It can be seen that on-time ratio increases when Refrigerant mass flow rate (kg/h) 4.01 1.97
either capillary tube length increases or LPG charge Cooling capacity (W) 123.00 127.00
decreases while thermostat position was kept unchanged. Volumetric cooling capacity (kJ/m3) 383.00 395.00
It is evident from Fig. 6c that on-time ratio of all LPG Electric power consumption (W) 186.00 178.00
Actual coefficient of performance (–) 0.66 0.71
combinations is lower than that of R134a. It should be sta-
ted that for given charge the shortest capillary tube length Cycling (on/off) tests at thermostat position (5)
Average freezer air temperature (C) 12.10 13.10
achieves the smallest on-time ratio.
Average cabinet air temperature (C) 5.00 4.00
On-time (min) 70.00 60.00
3.2.3. Energy consumption Off-time (min) 30.00 40.00
Energy consumption is defined as the product of electric On-time ratio (%) 70.00 60.00
power and running time. Energy consumption as a func- Energy consumption (kW h/day) 3.61 3.22
tion of thermostat position for each suitable LPG charge
with its corresponding capillary tube length is shown in
Fig. 6d. Clearly, energy consumption increases with the at higher refrigeration capacity because of its higher refrig-
thermostat position for all cases. The reason behind this erating effect. Pressure ratio, discharge temperature, mass
trend is mainly due to increase of on-time ratio (percentage flow rate and electric power consumption of LPG are lower
of running time) when the thermostat position changes than those of R134a by about 5.5%, 3.8%, 51.0% and 4.3%,
from 1 to 5 as shown in Fig. 6c to satisfy lower freezer respectively. Also, LPG as a drop-in substitute to R134a
and cabinet air temperatures. Fig. 6d reveals that energy yields higher coefficient of performance by 7.6%. Freezer
consumption decreases when either capillary tube length and cabinet air temperatures of LPG are 1 C lower than
decreases or LPG charge increases. This can be attributed those of R134a at the same thermostat position (5). Using
to the effect of both refrigerant mass flow rate and on-time LPG as a working fluid in the domestic refrigerator
ratio as mentioned earlier. Fig. 6d reveals that energy con- achieves energy saving of 10.8% and shorter compressor
sumption using capillary tube length of 5.5 m with LPG on-time ratio by 14.3% compared to R134a. The R134a
charge of 70 g is the highest one whereas energy consump- refrigerator has 3.61 kW h/day energy consumption with
tion using capillary tube length of 5 m and LPG charge of run-time ratio of 70% while the LPG refrigerator has
60 g is the lowest one; it is also lower than that of R134a. 3.22 kW h/day energy consumption with an on-time ratio
In conclusion, LPG charge of 60 g with capillary tube of 60%.
length of 5 m achieves similar volumetric cooling capacity
(Fig. 5c), lower energy consumption (Fig. 6d) and lower 4. Conclusions
freezer air temperature (Figs. 4a and 6a) than those of
R134a. Hence, LPG is a suitable drop-in replacement for Continuous running and cycling tests were conducted to
R134a in domestic refrigerators with minor modifications examine LPG of 60% propane and 40% commercial butane
which include initial mass of LPG charge and its capillary as a drop-in substitute for R134a in a single evaporator
tube length. domestic refrigerator with a total volume of 10 ft3
(0.283 m3) under tropical conditions. The considered
3.3. Comparison between R134a and LPG parameters include filling charges of R134a and LPG and
capillary tube lengths for LPG. The test results revealed
Table 3 summarizes the experimental results for both that
R134a baseline test and the best arrangement of LPG test
(capillary tube of 5 m and 60 g charge). Continuous run- • Freezer air temperature of 12 C can be achieved using
ning and on/off results confirm that the suitable mass of LPG charge of 50 g or more for capillary tube length
LPG filling charge is 0.6 and its capillary tube length is ranges from 4 m to 6 m.
1.25 of the original R134a values. The time required to pull • As LPG charge increases, freezer and cabinet air
down the temperature of the freezer space from 43 C to temperature and pull-down time decrease whereas dis-
12 C under no load condition, decreases from 87 min charge temperature, refrigerant mass flow rate, electric
for R134a to 80.4 min for LPG, showing a reduction in power consumption and volumetric cooling capacity
pull-down time of 7.6%. This means that LPG is working increase.
 M. Fatouh, M.E. Kafafy / Applied Thermal Engineering 26 (2006) 1593–1603 1603

• As capillary tube length increases, the freezer and cabi- Proceedings of IIF-IIR Conference, Oslo, Norway, 1998, pp. 636–
net air temperatures, pull-down time and discharge tem- 646.
[5] I. Kustrin, A. Poredos, Use of dimethyl ether in domestic refrigerator,
perature increase while the refrigerant mass flow rate, in: Proceedings of IIF-IIR Conference, Oslo, Norway, 1998, pp. 618–
electric power consumption and volumetric cooling 625.
capacity decrease. [6] S.A. Agarwal, M. Ramaswamy, A. Kant, Evaluation of hydrocarbon
• The same volumetric cooling capacity of R134a (383 kJ/ refrigerants in single evaporator domestic refrigerator freezer, in:
m3) can be obtained using LPG with a capillary tube Proceedings of International CFC and Halon Alternative, Washing-
ton, USA, 1995, pp. 248–257.
length of 5 m and charge of 50 g or with a capillary tube [7] R.N. Richardson, J.S. Butterworth, The performance of propane/
length of 5.5 m and charge of 60 g. isobutane mixtures in a vapor compression refrigeration system, Int.
• For a constant LPG charge, the on-time ratio decreases J. Refrig. 18 (1995) 58–66.
as the capillary tube length increase. [8] Z. Liu, I.H. Haider, R. Radermacher, Simulation and test results of
• Minimum electric energy is achieved with LPG charge hydrocarbon mixtures in a modified Lorenz Metzener cycle, ASH-
RAE Trans. 101 (1995) 1318–1328.
of 60 g and capillary tube length of 5 m combination. [9] I. Maclaine-cross, Hydrocarbon refrigerants for car air conditioners,
in: Proceeding of Seminar on ODS Phase-out Solution for the
Comparison of experimental results of both R134a (with Refrigeration Sector, Kuta, Bali, Indonesia, 1999, pp. 11–17.
its custom capillary tube length of 4 m and charge of 100 g) [10] D. Jung, C. Kim, K. Song, B. Park, Testing of propane/isobutane
and the best combination of LPG (capillary tube length of mixture in domestic refrigerators, Int. J. Refrig. 23 (2000) 517–
527.
5 m and charge of 60 g) revealed that [11] M.A. Alsaad, M.A. Hammad, The application of propane/butane
mixture for domestic refrigerators, Appl. Therm. Eng. 18 (1998) 911–
• Pull-down time, discharge temperature, pressure ratio, 918.
mass flow rate and electric power were reduced by about [12] M.A. Hammad, M.A. Alsaad, The use of hydrocarbon mixtures as
7.6%, 3.8%, 5.5%, 51% and 4.3%, respectively, when refrigerants in domestic refrigerators, Appl. Therm. Eng. 19 (1999)
1181–1189.
LPG is used as a drop-in replacement for R134a. [13] B.A. Akash, S.A. Said, Assessment of LPG as a possible alternative
• Coefficient of performance of the domestic refrigerator to R12 in domestic refrigerators, Energy Convers. Manage. 44 (2003)
working with LPG is improved by 7.6% compared to 381–388.
the one achieved when operating with R134a. [14] C. Melo, R.T.S. Ferreira, C.B. Neto, J.M. Goncalves, M.M.
• Freezer and cabinet air temperatures of LPG are 1 C Mezavila, An experimental analysis of adiabatic capillary tubes,
Appl. Therm. Eng. 19 (1999) 669–684.
lower than those of R134a. [15] J.L. Driessen, R.H. Pereira, C. Melo, R.T.S. Ferreira, Sizing capillary
• Energy consumption and on-time ratio using LPG are tubes for household refrigeration systems, International Appli-
lower than those of R134a by 10.8% and 14.3%, ance Manufacturing, Ashokan International Ltd., UK, 1998,
respectively. pp. 55–59.
[16] B.Y. Liu, M. Tomasek, R. Radermacher, Experimental results with
hydrocarbon mixtures in domestic refrigerator/freezers, ASHRAE
In conclusion, the reported results confirm that LPG is Trans. 101 (1995) 1415–1421.
an attractive drop-in substitute for R134a in domestic [17] M. Fatouh, M. El Kafafy, Assessment of propane/commercial butane
refrigerators. mixtures as possible alternatives to R134a in domestic refrigerators,
Energy Convers. Manage., in press.
[18] DIC, Delta Industrial Company, Specification of domestic refriger-
References ators, 2004, Processes and Specifications Department, Delta Indus-
trial Co., Imbaba, Giza, Egypt.
[1] R. Radermacher, K. Kim, Domestic refrigerator: recent develop- [19] ISO, International Standard Organization, International Standard-
ments, Int. J. Refrig. 19 (1996) 61–69. 8187, Household refrigerating appliances (refrigerators/freezers)
[2] B.E. Ben Adamson, Dimethyl ether as an R-12 replacement, in: characteristics and test methods, 1991, International Organization
Proceedings of IIF-IIR Conference, Oslo, Norway, 1998, pp. 610– 617. for Standardization, Switzerland.
[3] I.L. Maclaine-cross, E. Leonardi, Performance and safety of LPG [20] SUPERTRAPP, NIST Thermo-physical properties of hydrocarbon
refrigerants, in: Proceeding of the ‘Fuel for Change’ Conference of mixture database (Version 3.1), 2003, National Institute of Standards
Australian Liquefied Petroleum Gas Association Ltd., Australia, and Technology, Gaithersburg, Maryland, USA.
1995, pp. 149–168. [21] R. Moffat, Describing the uncertainties in experimental resultsExper-
[4] J. Rogstam, Experimental investigation of the chemical stability and imental Thermal and Fluid Science, Elsevier Science, New York,
the system performance with cyclopropane as refrigerant, in: 1988.