Solar Energy Vol. 65, No. 3, pp.

189–196, 1999
 1999 Elsevier Science Ltd
Pergamon PII: S 0 0 3 8 – 0 9 2 X ( 9 8 ) 0 0 1 2 4 – 8 All rights reserved. Printed in Great Britain
0038-092X / 99 / $ - see front matter

PERFORMANCE OF A HEAT PUMP USING DIRECT EXPANSION SOLAR

COLLECTORS†
S. ITO‡, N. MIURA and K. WANG
Department of Mechanical Systems Engineering, Kanagawa Institute of Technology, Atsugi 243-0292,
Japan

Received 25 September 1997; revised version accepted 9 September 1998

Communicated by DOUG HITTLE

Abstract—Theoretical and experimental studies were made on the thermal performance of a heat pump that
used a bare flat-plate collector as the evaporator. The analysis used empirical equations to express the electric
power consumption of the compressor and coefficient of performance (COP), as functions of temperature of
evaporation at the evaporator and that of the heat transfer medium (water) at the inlet of the condenser. The
experimental heat pump had a compressor with a rated capacity of 350 W and collectors with the total area of
3.24 m 2 . Around noon in winter the evaporator temperature was found to be about 178C higher than the
ambient air temperature of 88C, and a COP of about 5.3 was obtained when the water temperature at the
condenser inlet was 408C. These measured evaporation temperatures and COPs were in good agreement with
those predicted by the analysis. According to the analysis, the total area of the collectors in the experiment was
appropriate for the heat pump system. Also, the 1-mm thickness of the collector’s copper plate used in the
experiment could be 0.5 mm with little reduction of COP. The pitch of the tube soldered to the copper plate for
the refrigerant flow was 100 mm in the experiment, but the COP would only be reduced by about 4% if the
pitch were changed to 190 mm.  1999 Elsevier Science Ltd. All rights reserved.

1. INTRODUCTION equation for obtaining the amount of the refrig-

erant delivered by the compressor, simultaneous-
Solar energy can be used to heat the refrigerant in
ly. Also solved simultaneously were the equation
the evaporator of a heat pump, by using a solar
for finding the specific volume of the superheated
collector as the evaporator. The evaporation tem-
vapor in the collector, which was assumed to be
perature of the refrigerant is thereby increased
proportional to the absolute temperature, and the
(due to effective absorption of solar thermal
equation for finding the superheated region in the
energy by the evaporator) resulting in a higher
collector, which were derived from the common
heat pump performance. Charter and Taylor
collector theory. The resulting coefficient of per-
(1976) reported on such a system using a flat-
formance of the heat pump was compared with
plate collector without a cover as the evaporator.
the experimental results. The analysis was rather
Following this, various types of collectors have
complex and the details of the experiment were
been considered and the usefulness of the collec-
not given.
tors in heat pump systems has been investigated.
O’Dell et al. (1984) obtained the heat gain at
Studies of heat pumps using direct expansion
the condenser and the COP as functions of the
solar collectors, were reviewed by Shinobu and
evaporation temperature, from data available for a
Matsuki (1990) and Ito (1992). Very few ana-
heat pump operated with the temperature of the
lytical investigations on the performance of heat
heat source air at the condenser held constant.
pumps were reported, and the analytical results
Using these equations and the equation for the
were only rarely confirmed by experiments.
heat collected by the solar collector, the condenser
The analysis of Chaturvedi and Abazeri (1984)
heat gain and COP were predicted under various
assumed that the refrigerant at any tube cross-
weather conditions.
section in the collector was in a single phase.
This paper presents a simplified analysis of a
They solved the continuity and momentum equa-
heat pump that uses bare flat-plate collectors as
tions, the equation of heat balance, and the
the evaporator. The analysis can predict heat
†Paper presented at the ISES Solar World Congress, Taejon, pump performance under various weather con-
South Korea, 24–29 August 1997. ditions. Examined are the effects of solar radia-
‡Author to whom correspondence should be addressed. tion, sky radiation, ambient air temperature, wind

189
190 S. Ito et al.

speed, collector design, and the hot water tem-

perature at the condenser. The analysis was
validated using experimental results obtained in
the earlier study by Ito and Miura (1991). The
model having been established, the influence of
the collector area, the pitch of the tube for the
refrigerant flow, and the thickness and the collec-
tor plate material were also investigated analyti-
cally.
It should be noted that refrigerant R-12 was
used in those experiments, even thought it has
been banned for several years. On the other hand,
the analysis of the paper can be applied in the
same way to heat pumps using other kinds of
refrigerants. (Refrigerant R-22 has been more
recently tested in the study by Nisikawa and Ito, Fig. 2. The bare flat-plate collectors (evaporator).
1995.)
pitch between tubes of 100 mm. A 12-mm thick
wooden sheet was applied to the back of the
2. EXPERIMENT
collector plate, as an insulator. The collectors
Shown in Fig. 1, the apparatus used in this were installed on the roof of a cabin, facing south
study is the same as the one used by Ito and at an angle of inclination of 50 degrees. The
Miura (1991). There were two types of collectors temperatures at various points in the system, the
in that study, but only the data from the radiative ambient air temperature, the solar radiance, the
type evaporator (flat-plate collector) was used in wind speed and the compressor electric power
the present study. The compressor was of the consumption were recorded automatically by a
electrical rotary type, with a rated capacity of 350 data logger, every 5 min.
W, and R-12 was used as the refrigerant. The The heat gain at the condenser Q c was obtained
temperature of the circulating water at the inlet of by the following equation;
the condenser was kept constant during operation.
The collectors are shown in detail in Fig. 2. Q c 5 mc p (T w,2 2 T w,1 ) (1)
The total area of the two sets of collectors was where m is the mass flow rate of the water, c p is
2
3.24 m . The collector plate was a copper plate, the specific heat, and T w,1 and T w,2 are the
1-mm thick, with its surface painted black. A temperatures of the water at the inlet and the exit,
copper tube, having inside and outside diameters respectively. The flow rate m was kept at 0.0778
of 8.0 mm and 9.52 mm respectively, was sol- kg s 21 (280 kg h 21 ). The coefficient of perform-
dered to the backside of the copper plates, with a ance, defined as the ratio of the heat gain Q c to
the compressor electric power Wcp was calculated
from the measurements.

3. ANALYSIS

3.1. Performance of heat pump with collectors

Fig. 3 shows the heat pump cycle on a pres-
sure-specific enthalpy diagram. The water with
average temperature of T w,av at the condenser
receives heat Q c from the refrigerant at condens-
ing temperature T c . The temperature of the re-
frigerant evaporating at the collector (T e ) is higher
than the ambient air temperature T a , because of
the solar energy input.
The coefficient of performance COP and the
Fig. 1. Schematic diagram of the heat pump system for electric power consumption of the compressor Wcp
experiments. can be determined if the temperature of the
 Performance of a heat pump using direct expansion solar collectors 191

Wcp 5 f2 (T e ,T w,av ) (3)

or
COP 5 f3 (T e ,T w,1 ) (4)

Wcp 5 f4 (T e ,T w,1 ) (5)

The above equations were obtained considering a

heat pump with a non-solar evaporator. It is
assumed that the same equations can be also
obtained for the heat pump that uses a solar
collector instead of the evaporator in the system.
This analysis can be adapted to the heat pump for
which Eqs. (2) and (3) or Eqs. (4) and (5) are
Fig. 3. A heat pump cycle in the pressure-specific enthalpy available. The present analysis is able to predict
diagram. the performance of the system with solar collec-
tors for variable weather condition and for vari-
condensation of the refrigerant T c and the tem- able temperature of the heat source.
perature of the evaporation T e are known. When The following assumptions are made. (1) The
the performance of the heat pump in a particular refrigerant in the collectors exits as saturated
operation is to be predicted, the condensation vapor. (2) There is no pressure loss along the tube
temperature is not generally known a priori. of the collector, so that the evaporation tempera-
Instead, the temperature of the heat source at the ture, which is the saturation temperature, is the
condenser would be given. In this analysis, the same at any location. (3) Steady state conditions
flow rate of the heat transfer medium is assumed apply. (4) The thermal resistance of the bond
to be constant, so the heat transfer coefficient in between the tube and the collector plate can be
the side of the heat transfer medium is constant. neglected.
Now COP is defined by the ratio of the heat Considering that the sum of the energy con-
gained by the heat transfer medium Q c to the sumed by the compressor and the heat collected at
electric power consumption of the compressor the evaporator is equal to the thermal energy
Wcp . Since the heat transfer coefficient between gained by the heat transfer medium at the con-
the refrigerant and the wall of the tube in the denser, the following equation is obtained.
condenser is much larger than that between the
heat transfer medium and the wall of the tube, it Q c 5 Q e 1 Wcp (6)
may be assumed that the overall heat transfer
coefficient between the refrigerant and the con- The heat collected by the collector per unit area
denser heat transfer medium varies only with the and per unit time qe is given as follows
flow rate of the heat transfer medium. As a result,
the thermal resistance between the refrigerant and qe 5 Ia 2 qL (7)
the heat transfer medium can be assumed to be where I is the solar irradiance on the collector
constant when the flow rate of the heat transfer plate, a is the solar absorptivity of the collector
medium does not vary. Therefore, the heat gain plate, and qL is the heat loss from the collector,
Q c is considered to be a function of T c and the which is given as follows (Ito and Miura, 1989)
average temperature of the heat transfer medium
T w,av or the inlet temperature to the condenser qL 5 ´q0 1 U(T p 2 T a ) (8)
T w,1 . On the other hand, Q c is equal to the product
of COP and Wcp , and COP and Wcp for a given where ´ is the emissivity, T p is the temperature of
heat pump are considered to be functions of T e the collector plate at a local point, and T a is the
and T c . Consequently, Q c becomes a function of ambient air temperature. Also, q0 is the difference
T e and T c . Thus, T c becomes a function of T w,av between the emissive power per unit area from a
and T e or a function of T w,1 and T e . As a result, black body at the ambient air temperature and the
the COP and the power consumption of the emissive power from the sky q` , as given
compressor are given by q0 5 s T a4 2 q` (9)
COP 5 f1 (T e ,T w,av ) (2) The symbol U is the overall heat loss coefficient
192 S. Ito et al.

from the collector to the ambient air that is given

by
U 5 h w 1 4´s T a3 (10)
where s is the Stefan–Boltzmann constant and h w
is the wind heat transfer coefficient, given by

h w 5 5.7 1 3.8V [W m 22 ] (11)

where V is the wind speed expressed by the unit
of m s 21 .
From Eqs. (7) and (8), qe is given as follows
qe 5 S 2 U(T p 2 T a ) (12)
where S is the difference between the solar
Fig. 4. The relation between the electric power consumption of
radiation absorbed by the collector per unit area compressor and the evaporation temperature of the refrigerant
and the total radiation heat loss from the collector in the evaporator.
surface, and is given by
S 5 Ia 2 ´q0 (13) Wcp 5 4.3(T w,1 2 273.15) 1 151 [W] (18)

Following analysis given by Duffie and Beck- The relations between the evaporation tempera-
man (1980), the heat rate obtained by the collec- ture and COP are shown under various conditions
tor Q e is expressed by in Fig. 5. The results were also plotted in a graph
Q e 5 AF9qe (14) of COP versus (T w,av 2T e ) as shown in Fig. 6.
The supposed relation for COP as a function of
where F9 is the collector efficiency factor. As- (T w,av 2T e ), COP was fitted by
suming that the thermal resistance of the bond
between the collector plate and tube can be COP 5 a i 1 b i (T w,av 2 T e ) (i 5 1 to 4) (19)
neglected, F9 is given by
where
F9 5 F 1 (1 2 F )(D/W ) (15) a 1 5 8.0, b 1 5 2 0.17 for 0 , T w,av 2 T e # 20 K
where F is the fin efficiency, D is the outside
diameter of the tube and W is the pitch of the a 2 5 6.6, b 2 5 2 0.1 for 20 K , T w,av 2 T e # 30 K
tube. For the collector used in the experiment, F
and F’ were about 0.97. The COP is given by a 3 5 5.7, b 3 5 2 0.07 for 30 K , T w,av 2 T e
COP 5 Q c /Wcp (16) # 50 K

from Eqs. (6), (12), (14), (16), the following a 4 5 3.7, b 4 5 2 0.03 for 50 K , T w,av 2 T e
equation can be derived:
# 60 K
AF9hS 2 U(T e 2 T a )j 2 Wcp (COP 2 1) 5 0 (17)

Introducing Eqs. (2) and (3) or Eqs. (4) and (5)

into Eq. (17), T e can be found from Eq. (17),
once T w,i is specified.

3.2. Analysis for the heat pump used in the

experiment
The heat pump was operated under constant
water temperatures of 308C, 408C and 508C at the
inlet of the condenser. From the results of the
experiments, the relation between the power
consumption of the compressor and the evapora-
tion temperature, as shown in Fig. 4, are given by Fig. 5. The relation between COP and the evaporation
the equation temperature.
 Performance of a heat pump using direct expansion solar collectors 193

T e # T sat 2 DT (22)

When the evaporation temperature obtained with-

out considering the condition given by Eq. (21)
does not satisfy Eq. (22), the evaporation tem-
perature becomes T sat 2DT.

4. RESULTS AND DISCUSSION

4.1. Comparison of experimental and analytical

results
Fig. 7 shows the experimental results obtained
during a summer day when the ambient air
temperature was in the range between 308C and
Fig. 6. The dependence of the average water temperature at
368C. The temperature of the water at the inlet of
condenser and the evaporation temperature on COP.
the condenser was kept constant at 308C. The
experimental data plotted were instantaneous val-
The above equation is plotted in Fig. 6. ues except the solar irradiance, for which the
From Eqs. (17) and (18), the following equa- 30-min average was plotted. The variations of the
tion can be derived: solar irradiance and wind speed with time were
large. The evaporation temperatures and COP
AF9(S 1 UT a ) 2 Wcp (a i 1 b i T w,av 2 1) were mainly affected by the solar irradiance.
T e 5 ]]]]]]]]]]]
AF9U 2 b iWcp When the solar irradiance was about 800 W m 22
(20) at 11:00 a.m., the evaporation temperature was
308C and COP reached a value of 7.55. In all
cases, Eq. (22) was satisfied for DT5108C.
To obtain COP and the heat gained by the
Fig. 8 shows experimental results obtained
water at a temperature T w,i at the inlet of the
during a winter day when the ambient air tem-
condenser for a given weather condition, Wcp
perature was in the range between 68C and 118C.
should be obtained from Eq. (18) first. Then,
The temperature of the water at the inlet of the
assuming T w,av is equal to T w,i , an approximate
condenser was kept constant at 408C. The degree
value of T e can be found from Eq. (20). Using the
of superheat at the exit of the collector was less
values of T e and T w,av in Eq. (19), an approximate
than 108C.
COP can be obtained. Using these COP and T e ,
The variations of the evaporation temperature,
approximate Q c can be found from Eq. (16).
COP and the heat gain Q c with time are shown in
From this Q c and the flow rate of the heat transfer
Figs. 9–11, respectively. The evaporation tem-
medium m, the outlet temperature of the heat
transfer medium T w,2 can be obtained by using
Eq. (1). From T w,2 and T w,1 , the average tempera-
ture of the water T w,av can be found. This
procedure is then repeated until converged values
of COP,T e and Q c can be found. If the collector is
exposed to the solar radiation without collecting
heat, the temperature of the collector T sat is
obtained from Eq. (12) by adopting Q c 50 as
follows.
T sat 5 T a 1 S /U (21)

The maximum temperature of the refrigerant at

the exit of the collector as well as the temperature
of the collector is T sat . Therefore, if the degree of
superheat at the exit of the collector is set to a
certain value DT, the following condition should Fig. 7. The experimental results obtained on July 17,
apply: t w,1 5308C.
194 S. Ito et al.

Fig. 8. The experimental results obtained on Jan. 30,

t w,1 5408C.
Fig. 10. Comparison between the experimental and analytical
results of COP.
peratures became about 358C around noon on July
23rd when the inlet water temperature T w,1 was perimental and analytical results, they generally
608C. Referring to the previous study by Ito and agreed with each other quite well.
Miura (1989), q0 was chosen to be 40 W m 22 in
the summer and 80 W m 22 in the winter. The 4.2. Sensitivity of analytical results to the
evaporation temperatures in the analysis were collectors and other parameters
about 38C higher than the experimental ones, and Fig. 12 shows the variation of COP with total
the analytical COP was 0.4 higher than COP of collector area. The other dimensions of the collec-
about 4 in the experiment. tor are the same as those used in the experiment.
There was a larger difference between the The wind speed and the inlet water temperature
experimental and the analytical COP results at 2 are chosen to be 2 m s 21 and 408C, respectively.
p.m. on July 17th. This is probably due to the The ambient air temperature is chosen to be 108C.
steady state assumption used in the analysis. The The results show the COP increases with increas-
amount of the heat released at the condenser ing area. Theoretically, the COP will go to 1.0 as
would not be influenced spontaneously by an area A goes to zero. The area of the collector used
abrupt change of the ambient air temperature, in the experiment is 3.24 m 2 . The further increase
solar radiation and the wind speed. of COP with an increase of the collector area is
Despite some disagreement between the ex- rather small. Therefore, the size of the collector

Fig. 9. Comparison between the experimental and analytical Fig. 11. Comparison between the experimental and analytical
results of the evaporation temperatures. results of the heat gain at the condenser.
 Performance of a heat pump using direct expansion solar collectors 195

Fig. 12. Variation of COP with the collector area.

Fig. 14. Variation of COP with the pitch of the tube.

used in the experiment is considered to be appro-

priate.
The effect of the collector plate thickness d on if T e .T a . Therefore, relative magnitudes of heat
COP is shown in Fig. 13. For copper plate gain by convection and radiation can be estimated
thicknesses greater than 0.5 mm, the effect of the if the evaporation temperature is obtained. COP
thickness on COP is small. For a plate of iron, goes to 1.0 as the collector area becomes zero.
whose thermal conductivity is one seventh that of The heat collected by evaporator is only by
copper, the effect of the thickness is only small convection at I50.
when d .2 mm. Fig. 14 shows the variation of COP with the
Heat is absorbed from the ambient air if the pitch W of the tube. When the pitch was increased
evaporation temperature T e is lower than the to 190 mm from 50 mm, the reduction of COP
ambient air temperature, but heat is lost to the air was only 4% for the copper plate, about 7% for
the aluminum plate, and about 20% for the iron
plate.

5. CONCLUSIONS

Using experimental data, the power consump-

tion of the compressor and the coefficient of
performance of a heat pump can be represented
by empirical functions of the evaporation tem-
perature and the temperature of the water at the
condenser. Using these relations and the theory of
a bare flat-plate collector, the evaporation tem-
perature, the coefficient of performance and the
heat gain at the condenser for the heat pump with
the flat-plate collector can be closely predicted,
for various weather conditions. Good agreements
between the experimental and analytical results
were obtained in the present studies. According to
the simulations, the effect on COP of enlarging
the area of the collector over that used in the
experiment would be small. Therefore, the total
area of the collectors used in the experiment was
appropriate. The thickness of the (copper) collec-
tor plate could be made 0.5 mm and the pitch of
Fig. 13. Variation of COP with the thickness of the collector the tube on the collector could be 190 mm instead
plate (fin). of 100 mm, without significant reduction of COP.
196 S. Ito et al.

NOMENCLATURE condensation or condenser

e evaporation or evaporator
A collector area, m 2 p collector plate
ai constant w water (heat transfer medium)
21
bi constant, K 1 inlet
COP coefficient of performance 2 exit
cp specific heat, J kg 21 K 21
D outside diameter of tube, m
F fin efficiency
F9 collector efficiency factor
hw wind heat transfer coefficient, W m 22 K 21 REFERENCES
I solar irradiance on collector, W m 22
m mass flow rate, kg s 21 Charter W. W. S. and Taylor L. E. (1976) Some performance
Qc heat gained by heat transfer fluid (water) at con- characteristics of a solar boosted heat pump. In Proceedings
of I.I.R. Conference, pp. 641–648, Melbourne, Australia.
denser, W
Chaturvedi S. K. and Abazeri M. (1984) Thermal performance
Qe heat collected at evaporator, W
of a direct expansion solar-assisted heat pump. Solar Energy
qe heat collected at local place of collector, W m 22 33, 155.
qo cooling power of blackbody at ambient air tempera- Duffie J. A. and Beckman W. A. (1980) Solar Engineering of
ture, W m 22 Thermal Processes, pp. 197–249. John Wiley and Sons,
qL heat loss from collector, W m 22 New York.
22
q` sky radiation, W m Ito S. and Miura N. (1989) Studies of radiative cooling
22
S 5Ia 2 ´qo , W m systems for storing thermal energy. J. Solar Energy, En-
T absolute temperature, K gineering 111, 251–256.
t Celsius temperature, Ito S. and Miura N. (1991) A comparison of heat pump
systems using different types of direct expansion solar
U heat loss coefficient, W m 22 K 21
collectors. Proceedings of Biennial Congress of the Interna-
V wind speed, m s 21 tional Solar Energy Society, pp. 1655–1659, Denver, USA.
W pitch of tube, m Ito S. (1992) Heat pumps with direct expansion solar collec-
Wcp electric power consumption of compressor, W tors. Trans. Japanese Association of Refrigeration ( in
Japanese) 9, 101–115.
Greek letters Nisikawa M. and Ito S. (1995) Heat pump using photovoltaic
a absorptivity of collector and thermal hybrid panels. Proceedings of 1995 ASME /
d thickness of collector plate, m JSME / JSES International Solar Energy Conference Maui,
´ emissivity Hawaii. Solar Engineering 1, 17.
s Stefan–Bolzmann constant, W m 22 K 24 O’Dell M. P., Michell J. W. and Beckman W. A. (1984) Design
method and performance of heat pumps with refrigerant-
sat without collection of heat
filled solar collectors. J. Solar Energy Engineering 106,
159–164.
Subscripts Shinobu Y. and Matsuki K. (1990) Recent technical trends in
a ambient air direct expansion solar heat pump systems. Proceedings of
av average between inlet and outlet at condenser the 3 rd International Energy Agency Heat Pump Confer-
c ence, pp. 487–497, Tokyo, Japan.