Energy 93 (2015) 1394e1403

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Experimental investigations on using phase change material for

performance improvement of storage-enhanced heat recovery room
air-conditioner
Jie Jia, W.L. Lee*
Department of Building Services Engineering, The Hong Kong Polytechnic University, Hung Hom, Hong Kong

a r t i c l e i n f o a b s t r a c t

Article history: The use of SEHRAC (storage-enhanced heat recovery room air-conditioner) for space cooling and free
Received 11 June 2015 water heating has been conrmed effective but its overall performance will unavoidably be affected by
Received in revised form the associated increase in water temperature in the heat recovery water tank. Measures to enhance the
19 September 2015
heat storage capacity of the water tank will thus be benecial to the overall performance of SEHRAC. But
Accepted 15 October 2015
virtually none has been done to date. For achieving this objective, the use of PCM (phase change ma-
Available online xxx
terial) in the water tank was proposed in this study. A prototype SEHRAC integrated with an expanded
graphite/parafn composite PCM was designed and setup in a test facility for laboratory experiments.
Keywords:
Phase change material
Two identical sets of experiments with and without the use of PCM (wPCM and woPCM scenarios) under
Heat recovery a range of outdoor temperature conditions were conducted. It was found that the wPCM scenario, as
Air-conditioner compared to the woPCM scenario, the overall coefcient of performance was 6.9%e9.8% higher and the
Space cooling heat retention time of the tank water was 21.1% longer. The results conrmed that the performance of
Water heating SEHRAC could be improved by the use of PCM.
Energy performance 2015 Elsevier Ltd. All rights reserved.

1. Introduction
Air-conditioners function to remove heat from indoor spaces
and subsequently discharge this heat to the outside. The discharged
heat, if recovered for water heating, can reduce energy use and
greenhouse gas emissions. Along this line, HRAC (heat recovery
room air-conditioner) was invented in the eighties for providing
simultaneous space cooling and water heating [1]. Its energy saving
potential and payback period were found attractive [2]. However,
due to the possible heat supply and demand mismatch, recent
research advocates the use of a water tank as an energy reservoir
for heat recovery, leading to the evolvement of SEHRAC (storage-
enhanced heat recovery room air-conditioner). When employed in
residential buildings, SEHRAC is often used to recover and store the
nighttime discharged heat for satisfying the daily water heating
demand of a household.
Many simulation and experimental investigations have been
done on SEHRAC. On simulation investigations, a study by the au-
thors conrmed the effective use of SEHRAC for households in a
* Corresponding author. Tel.: 852 27665852.
E-mail address: bewll@polyu.edu.hk (W.L. Lee).
typical residential area in Hong Kong [3]. A similar study was also
carried out by Shimoda et al. [4] in Japan. Techarungpaisan et al. [5],
based on fundamental principles of heat transfer, thermodynamics
and uid mechanics, developed a detailed mathematical model for
SEHRAC. The model was coded into a simulation program and was
used for system optimization. Yokoyama et al. [6] developed
another mathematical model to relate the energy performance of
SEHRAC with the outdoor air and tank water temperatures. Apart
from mathematical models, recent research focuses on incorpo-
rating SEHRAC with other control devices to enable its efcient
operation [7,8].
On experimental investigations, Ying et al. [9] examined the
overall performance of SEHRAC and found that it was more efcient
than electric water heating with an energy efciency ratio of 1.2.
Jiang et al. [10] found that SEHRAC could operate satisfactorily and
the associated savings in water heating energy use were evident.
Similar conclusions were also drawn by Monerasinghe et al. [11].
The energy performance, as compared to conventional air-
conditioners, was examined by the authors [12]. It was reported
that, for a typical household in Hong Kong, the use of SEHRAC could
reduce the annual water heating energy use by 49%. In another
study by the authors [13], the use of capillary tube and TEV

http://dx.doi.org/10.1016/j.energy.2015.10.053
0360-5442/ 2015 Elsevier Ltd. All rights reserved.
 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403 1395

Nomenclature 4 relative humidity

cp specic heat capacity (kJ/kg K) Subscript

C regression coefcient a air
COP coefcient of performance b atmosphere
h specic mass enthalpy (kJ/kg) cd compressor discharge
H heat storage capacity of the water tank (MJ) cl space cooling
m mass ow rate (kg/s) cs compressor suction
M mass (kg) ei evaporator inlet
P pressure (Pa or MPa) er evaporator return
Q space cooling/water heating capacity (kW) es evaporator supply

t temperature ( C) hi heat exchanger inlet
T absolute temperature (K) ho heat exchanger outlet
w moisture content (kg/kg dry air) oa overall
W power consumption (kW) p PCM
x refrigerant vapor content (kg/kg) pa parafn
r refrigerant
Greek symbol sl saturated liquid refrigerant
a time spent for the entire experiment sw saturated water vapor
l latent heat of fusion (kJ/kg) w water
s latent heat of vaporization (kJ/kg) wh water heating
t operating time

(thermostatic expansion valve) as expansion devices in SEHRAC
were side-by-side compared. It was identied that TEV was a more
suitable option for its better refrigerant ow control during the heat
recovery process.
The above studies have conrmed the effective use of SEHRAC
for space cooling and free water heating, but it has been
acknowledged that the associated increase in tank water temper-
ature will unavoidably affect the cooling and heating performances
of SEHRAC [11e13]. Therefore, measures to enhance the heat stor-
age capacity of the water tank will be benecial to the overall
performance of SEHRAC.
For achieving this objective, the use of PCM (phase change
material) in the water tank is proposed in this study. PCM is capable
of storing and releasing large amounts of energy when charging
and discharging. During charging, the material changes from solid
to liquid and vice versa for discharging. As reported, PCMs can
typically store 5e14 times more energy per unit volume than water
[14]. The phase change characteristics also allow heat storage or
release at an almost constant temperature. PCMs thus have been
widely investigated for applications in various heat recovery sys-
tems [15,16] and systems adopting natural cooling sources [17].
For SEHRAC, the benet of using PCM with a high energy storage
density is twofold. First, its overall performance can be improved
with the heat storage capacity of the water tank enhanced by PCM
charging. And when it is not operating, PCM discharging can
partially offset the heat loss from the water tank to better maintain
the hot water at desired temperature.
The temperature of hot water for direct residential use often
ranges between 35  C and 50  C [18]. There are a number of PCMs
with phase change temperatures within this range; namely salt
hydrates, fatty acids and parafn waxes [19]. Among them, parafn
waxes are considered the most suitable option for domestic hot
water applications owing to their non-toxicity, wide availability,
low cost and non-corrosive property [20]. But as parafn waxes are
of extremely low thermal conductivity, recent studies are investi-
gating the use of different heat transfer enhancement techniques to
shorten the charging and discharging processes [21]. The tech-
niques include insertion of a metal matrix into parafn, dispersion
of metallic particles, using parafn containers augmented with ns,
microencapsulation of parafn, and impregnating parafn into a
high thermal conductivity material with porous structure.
Despite the potential benets, there is virtually no study to date
that integrates the use of SEHRAC and PCM. To conrm and quantify
the performance improvement for their integration, a prototype
SEHRAC was designed and setup in a test facility for laboratory
experiments. Two identical sets of experiments with and without
the use of PCM (abbreviated as wPCM and woPCM scenarios) under
a range of outdoor temperature conditions were conducted. In the
experiments, the operating parameters on the refrigerant, air and
water sides of the system were closely monitored. Based on the
experimental results, the heat storage capacity of the water tank,
the refrigerant side operation characteristics as well as the space
cooling, water heating and overall energy performances were
examined and compared for the two scenarios.
Given the small equipment capacity due to residential use and
the highly volatile refrigerant properties, monitoring of the refrig-
erant side performance is difcult. Thus the results presented in
this study and the experimental protocol established will be useful
for further development of SEHRAC.
2. Experimental study
2.1. Test facility
The experiments were carried out in a test facility located at the
Hong Kong Polytechnic University. It comprised two environmental
chambers resembling indoor and outdoor conditions. Each cham-
ber measured 3.9 m in length, 3.8 m in width and 2.2 m in height,
which was constructed in conformity with ASHRAE Standard 16
[22] for rating of room air-conditioners. Given the same test facility
was used in earlier studies by the authors on performance evalu-
ation of SEHRAC [12,13], to avoid duplication, only brief de-
scriptions are given below.
The two chambers were completely insulated and separated.
The outdoor chamber was conditioned by a BIACS (built-in air-
conditioning system) and was provided with a set of sensible and
1396 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403
Table 1
Specications of the test facility.
BIACS
Cooling output 9.9 kW
Supply air ow rate 1700 m3/h
Supply air pressure head 1100 Pa
LGUs
Maximum sensible load generation rate 12 kW
Maximum latent load generation rate 4.8 kW
LGUs (latent load generation units). The indoor chamber was pro-
vided with another set of LGUs. The specications of the BIACS and
the LGUs are summarized in Table 1. A computer supervisory pro-
gram was adopted to regulate the outputs of the LGUs to maintain
the pre-set temperatures and humidities in the two chambers.
2.2. Prototype design
To accomplish the objectives of this study, a prototype SEHRAC
was designed and setup for laboratory experiments. A schematic
representation and a photograph of the prototype are shown in
Fig. 1(a) and (b), respectively. It mainly comprised a split-type air-
cooled air-conditioner and a heat recovery water tank. Inside the
water tank, there were a refrigerant-to-water heat exchanger and a
PCM container. Table 2 summarizes the specications of these key
components. The operating cycle of the prototype is shown in Fig. 2.
The split-type air-cooled air-conditioner had a nominal cooling
capacity of 2.5 kW. The indoor unit was basically a direct expansion
evaporator with copper tubes and aluminum ns. The outdoor unit
consisted of an air-cooled condenser and a rotary type compressor.
There was no capacity control provided for the compressor. A TEV
(Danfoss TX2) was used to replace the factory-provided capillary
tube for better refrigerant ow control, as recommended in an
earlier study [13]. The sensing bulb of the TEV was mounted on the
horizontal suction line near the evaporator outlet. Non-hydroscopic
insulation material was used to insulate the connecting pipings.
The amount of refrigerant charged into the system was determined
according to the actual piping length and the manufacturer's
recommendations.
The heat recovery water tank was cylindrical in shape and
contained 130 L of water (same as earlier studies [12,13]). It was
insulated with rock wool wrapped with a layer of aluminum foil
with a total thickness of 50 mm.
The refrigerant-to-water heat exchanger was a helical coil made
of copper. It was installed between the compressor and the air-
cooled condenser and was immersed vertically in the water tank
with top inlet to cater for the density change of vaporeliquid
refrigerant.
A PCM container of cylindrical shape was placed inside the he-
lical heat exchanger and was located centrally in the water tank
(Fig. 1(b)). It was made of aluminum with longitudinal ns added to
the external surface to enlarge the total heat transfer area for
minimizing the thermal resistance between the container and the
water. A cross-sectional view of the PCM container (with detailed
dimensions) is shown in Fig. 3. The dimensions were determined by
reference to a study by Castell et al. [23], which was also on do-
mestic hot water heating but was solar-powered.
Parafn wax RT44HC supplied by Rubitherm Technologies
GmbH was chosen as the latent heat storage material in this study.
Given the poor heat conduction performance of the parafn, EG
(expanded graphite) with a high thermal conductivity was mixed
into it in a typical compounding process to form an EG/parafn
composite PCM. The properties of the parafn and the EG used are
summarized in Table 3.
EG is highly porous and has a good compatibility with organic
materials. In an EG/parafn composite PCM, EG serves as a solid
matrix to enable the impregnation of parafn into its pores by
capillary and surface tension forces [24]. A conductive network is
then formed by EG to enhance the overall thermal conductivity. For
achieving a balance between thermal conductivity and energy
storage density, too high or too low an EG content is not desirable.
Previous literatures recommended 7%e15% by volume and 10% was
most commonly adopted [23,25,26]. Therefore, the composite PCM
(hereinafter referred to as the PCM) with an EG content of 10% by
volume was used in this study. Its thermal conductivity measured
by using transient hot-wire method was 1.2 W/m K, while the total
mass that could be encapsulated in the PCM container was about
12.6 kg.
2.3. Experimental measurements
The measuring instruments used in this study and their corre-
sponding accuracies are summarized in Table 4. The refrigerant
temperatures were measured by calibrated thermocouples (K
type). The refrigerant high and low side pressures were measured
by pressure transmitters of appropriate ranges for minimizing the
measurement error [27]. The refrigerant ow rate was measured by
a variable area ow meter installed between the air-cooled
condenser and the TEV. The pressure drop across the ow meter
was measured (2.1 kPa). It was much lower than the ASHRAE's
recommended limit of 82.7 kPa to conrm fullling the accuracy
requirement [28]. The evaporator return and supply air conditions
were measured by hygrometers. The supply air ow rate was
measured by thermal anemometers. The hygrometers and the
thermal anemometers were positioned according to the distribu-
tion method recommended by British Standard 5141 [29]. The
water temperatures at various radial and axial positions in the
water tank were measured by Pt100 RTDs. They were also used for
measuring the PCM temperatures along two longitudinal axes of
the container. Along each axis, temperature measurements were
taken at four equally spaced positions (T1-T4 and T5-T8 in Fig. 3).
The power consumption of the prototype was measured by a power
quality tester.
Based on the measurements, the space cooling and water
heating capacities as well as the overall coefcient of performance
of the prototype (detailed in Sections 3.3 and 3.4) could be ob-
tained. The associated uncertainties were estimated by using Kline
and McClintock's method [30,31]. The procedures were described
in detail in an earlier study by the authors [13], indicating that the
uncertainties were well within the ISO's recommendation of 10%
[32].
All the instruments used were connected to two data loggers
(Agilent 34970A), which allowed all measured data to be real-time
monitored, recorded and processed. In this experimental study, the
measurements were taken at a short time interval of 5-s to better
capture the dynamic variations in operating parameters of the
prototype.
2.4. Experimental procedures
The experiments carried out in this study were classied as heat
charging and discharging tests. Heat charging referred to water
heating in the water tank, while heat discharging referred to heat
loss from the water tank to the atmosphere. For each test, two
identical sets of experiments with and without the use of PCM
(wPCM and woPCM scenarios) were conducted.
A preliminary comparison was made on the dynamic operating
parameters of the prototype monitored in this and an earlier
studies [13]. It was revealed that they were comparable under
 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403 1397

Fig. 1. Schematic and photograph of the prototype SEHRAC. (T: Thermocouple; P: Pressure transmitter; H: Hygrometer; HX: Refrigerant-to-water heat exchanger; PCM: PCM
container; SG: Sight glass).
1398 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403

Table 2 adjusted to exactly offset the instantaneous cooling output of the

Specications of the prototype SEHRAC. prototype. For the wPCM scenario, the experiment was stopped on
Split-type air-cooled air-conditioner complete melting of the parafn in the PCM. The complete melting
Model Mitsubishi SRK09CMP was conrmed by checking if the PCM temperatures measured at
Nominal cooling capacity 2.5 kW positions T1 to T4 (Fig. 3) had exceeded the melting temperature
Nominal power consumption 0.925 kW
Compressor Hermetic and rotary type
range and reached a thermal equilibrium with the surrounding
Evaporator and condenser Copper tubes and aluminum ns water. The four measurement positions were chosen because they
a
Expansion device TEV (Danfoss TX2 with an orice number of 01) were furthest from the heat transfer boundary. The time spent for
Working uid R22 the entire experiment was recorded. An identical experiment, with
Refrigerant-to-water heat exchanger
the same time spent, was then carried out for the woPCM scenario
Material Copper
Tube diameter 10 mm to enable their direct comparisons. The above-described experi-
Total length 15.6 m ments, both for the wPCM and woPCM scenarios, were repeated for
Heat recovery water tank the three outdoor temperature conditions.
Material High-density polyethylene (HDPE)
Dimensions 400 mm (ID)  1300 mm (Height)
Insulation layer Rock wool and aluminum foil 2.4.2. Heat discharging experiments
PCM container The tank water and the PCM (in the case of the wPCM scenario)
Material Aluminum
were heated to an initial temperature of 50  C (maximum hot water
Dimensions 140 mm (ID)  1100 mm (Height)
Fin size 3 mm (Width)  60 mm (Length)  temperature for direct residential use [18]). The water was then left
1100 mm (Height) to cool down naturally to the minimum use temperature of 35  C
a
A TEV was used to replace the factory-provided capillary tube.
[18], during which the prototype was left idle and the outdoor
chamber (where the water tank was located) was maintained at
23  C. The temperature was chosen because it is Hong Kong's
similar experimental conditions to conrm the repeatability of the annual average outdoor temperature in the typical meteorological
experimental results and the reliability of the experimental pro- year [33]. The time elapsed for the entire experiment was recorded.
cedures detailed below. The procedures were repeated for the woPCM scenario to enable
their direct comparisons.

2.4.1. Heat charging experiments

3. Results and discussion
The indoor temperature was maintained at 22  C, while the
outdoor temperature was varied from 25  C to 35  C with an in-
This section discusses the performance of the prototype oper-
terval of 5  C to become three different experimental conditions
ated in the wPCM scenario as compared to that in the woPCM
(same as earlier studies [12,13]). The desired conditions in the
scenario. For heat charging, the performance is evaluated based on
chambers were achieved by the use of PID (proportional-integral-
the heat storage capacity of the water tank, the refrigerant side
derivative) controllers to provide real-time adjustments of the
performance, the space cooling and water heating capacities as well
LGUs outputs and were maintained unchanged throughout each
as the overall energy performance. For heat discharging, the per-
experiment.
formance is evaluated based on the heat retention time of the tank
The outdoor chamber was rst conditioned to the pre-set con-
water. Due to the large amount of experimental data collected,
dition. The tank water and the PCM (in the case of the wPCM sce-
performance evaluations for heat charging are presented for an
nario) were heated to an initial temperature of 26  C, which is Hong
outdoor temperature of 30  C. The results obtained for other
Kong's annual average tap water temperature [18]. Upon achieving
the desired conditions, the prototype, the LGUs in the indoor
chamber and all the measuring instruments were powered on to
maintain the indoor chamber at 22  C and 55% RH (relative hu-
midity). Under this condition, the indoor cooling load was actively

Fig. 3. Cross-sectional view of the PCM container. (T1eT4 and T5eT8: Temperature
Fig. 2. Operating cycle on pressure-enthalpy diagram. measurements numbered from top to bottom of the container).
 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403 1399

Table 3
Properties of the parafn and the EG.

Parafn
Melting temperature rangea 39.1  Ce42.5  C
Solidication temperature rangea 43.6  Ce42.3  C
Latent heat of fusiona 256.9 kJ/kg
Specic heat capacity 2 kJ/kg K (solid and liquid)
Thermal conductivity 0.2 W/m K (solid and liquid)
EG
Average particle size 180 mm
Bulk density 10 kg/m3
Carbon content 95%
Moisture content 5%
Thermal conductivity 23 W/m K
a
Determined from differential scanning calorimetry (DSC) analysis.

experimental conditions will also be analyzed and given where

necessary.

Fig. 4. Water and PCM temperatures during heat charging.

3.1. Heat storage capacity of the water tank

During heat charging, whether the use of PCM can enhance the
heat storage capacity of the water tank and can eventually lower
the tank water temperature are important to the performance
improvement of the prototype.
Fig. 4 compares the time varying average tank water tempera-
tures (tw) for the wPCM and woPCM scenarios. In the gure, tw is
taken as the arithmetic mean of the water temperatures measured
at 12 different positions in the water tank. It can be seen that upon
the prototype is switched on, owing to the condensation of the hot
vapor refrigerant from the compressor discharge, tw for both sce-
narios increase gradually to exceed 50  C. But the rate of the tem-
perature increase drops with operating time to indicate a reduced
amount of heat rejection in the water tank. It is further noted that
tw for the wPCM scenario is constantly lower during heat charging,
with a maximum temperature difference of 2  C. This is attributed
to the use of PCM in the water tank. As a result, part of the energy
stored in the water is absorbed by the PCM to lower the rate of
increase in tw.
To demonstrate a proper heat charging, Fig. 4 shows also the
time varying PCM temperatures for the wPCM scenario measured
at the four positions along the centered axis of the PCM container
(tp1 to tp4 for T1 to T4 in Fig. 3). By analyzing the parafn with DSC
(differential scanning calorimetry) technique, its melting temper-
ature range (39.1  Ce42.5  C) can be identied and is shown as two
horizontal lines in Fig. 4.
Based on the identied melting temperature range, the changes
in the PCM temperatures can be explained. It can be seen that the
proles for tp1 to tp4 are divided into three phases. They are solid
sensible heating, phase change and liquid sensible heating phases.
In the solid sensible heating phase, the energy absorbed by the PCM
is in the form of sensible heat, which raises its temperature linearly
from 26  C to 39.1  C. But the temperature increase, as compared to
that in the water, is at a relatively low rate. Upon reaching the
temperature of 39.1  C, the parafn in the PCM starts to change
phase and turn from solid to liquid, whereby the PCM temperature
remains nearly constant. As the temperature increases to exceed
42.5  C, the parafn in the pores of EG is melted. Energy is then
absorbed in the form of sensible heat. As the heat transfer in liquid
parafn can be enhanced by convection [34], the PCM temperature
therefore jumps immediately to reach a thermal equilibrium with
the water.
Besides phase change, the PCM also exhibits a vertical temper-
ature difference. A review of the proles for tp1 to tp4 shows that the
PCM in the upper part of the container spends much less time to
reach a common temperature with the water. This can be explained
by the top inlet design of the heat exchanger as described previ-
ously. With this conguration, the hot vapor refrigerant discharged
from the compressor ows from top to bottom of the water tank,
resulting in higher water temperature in the upper part and thus
higher heat transfer potential between the water and the PCM.
As for the heat storage capacity of the water tank (H), it can be
estimated by using Eq. (1) for the woPCM scenario. While for the
wPCM scenario, H can be quantied by using Eq. (2) when the
sensible heat stored in the EG is ignored.
H cp;w Mw Dtw (1)
H cp;w Mw Dtw cp;pa Mpa Dtpa lMpa (2)

where cp,w and cp,pa are the specic heat capacities of water and
Table 4 parafn (kJ/kg K), respectively; Mw and Mpa are the total masses of
Summary of the measuring instruments. water and parafn (kg), respectively; Dtw and Dtpa are the tem-
Instrument Model Measurement range Accuracy perature increments of water (31.9  C for woPCM; 29.9  C for


Hygrometer ROTRONIC HygroFlex 40  Ce85  C 0.1  C

wPCM) and parafn (29.9  C) for the entire experiment ( C),
0e100% RH 1% RH respectively; l is the latent heat of fusion of parafn (kJ/kg).
Thermal anemometer E E 70-VT62B5 0e5 m/s 2% F.S. In Eq. (2), the three product terms on the right-hand side
Pt100 RTD Omega 1/10 DIN 100  Ce400  C 0.05  C represent, respectively, the sensible heat stored in the water, the
Thermocouple Omega TT-K-24 0  Ce1250  C 0.1  C
sensible heat stored in the PCM and the latent heat stored in the
Pressure transmitter Danfoss AKS 32 and 33 0.1 MPae1.2 MPa 0.3% F.S.
0.1 MPae3.4 MPa PCM. According to Eq. (2), H for the wPCM scenario is estimated to
Variable area KROHNE H250/RR/M9 0 to 3.2 L/min 1.6% F.S. be 20.1 MJ, which is 15.7% higher than that for the woPCM scenario
ow meter (17.4 MJ). It can therefore be seen that, despite the lower tw, the
Power quality ISO-TECH IPM-3005 0.01 kWe9.999 kW 20 W wPCM scenario still has a higher H due to the higher energy storage
tester
density of the PCM.
1400 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403
Fig. 5. Comparisons of compressor suction and discharge pressures.
Based on the above, it is evident that the PCM can function as an
internal heat sink to lower the rate of water temperature increase
during heat charging and to enhance the heat storage capacity of
the water tank.
3.2. Refrigerant side performance
The refrigerant side performance of the prototype can be
revealed from the compressor suction and discharge pressures, the
refrigerant ow rate and the refrigerant vapor content at the
evaporator inlet.
To illustrate the refrigerant pressure uctuations during heat
charging, Fig. 5 compares the time varying compressor suction
(Pcs,r) and discharge (Pcd,r) pressures for the two scenarios. It can be
seen that Pcd,r for both scenarios display a continuous increase with
operating time. This is because as tw increases, the heat rejection
performance of the prototype drops to increase the condensing
pressure and thus Pcd,r. Between the two scenarios, the wPCM
scenario has a slightly lower Pcd,r due to the lower tw as revealed in
Fig. 4. While for Pcs,r, as explained in an earlier study [35], it is less
inuenced by the heat recovery process and thus a very small
change is observed for both scenarios.
Fig. 6 compares the time varying refrigerant ow rates (mr) for
the two scenarios, illustrating that mr for both scenarios are fairly
stable despite the pressure uctuations. This is due to the active
control characteristics provided by the TEV. The refrigerant ow in
a TEV system depends on the pressure difference across the TEV as
well as its opening extent [36]. At a particular opening position, the
rise in Pcd,r tends to increase mr, but that subsequently lowers the
degree of superheat at the evaporator outlet. To restore the pre-set
degree of superheat, the TEV will reduce its opening extent to
throttle the refrigerant ow. Therefore, mr does not vary much
during heat charging. For the same reason, it is noted that mr for the
two scenarios are almost identical despite the difference in Pcd,r
(Fig. 5).
The amount of vapor refrigerant at the evaporator inlet is a key
indicator of system performance because the presence of vapor can
reduce the amount of liquid refrigerant available for effective
cooling [36]. The vapor is called ash gas and is formed in the
expansion process due to a drop in saturated temperature with
reduced refrigerant pressure. The amount of ash gas at the
evaporator inlet is often represented by its fraction in two-phase
ow (vapor content, xei,r). xei,r is affected by the heat rejection
performance. A better heat rejection performance will result in a
Fig. 6. Refrigerant ow rate comparisons.
lower xei,r [37]. xei,r can be estimated by using the following
equation:
hei;r  hsl
xei;r (3)
s
where hei,r is the enthalpy of refrigerant at the evaporator inlet (kJ/
kg); hsl is the enthalpy of saturated liquid refrigerant (kJ/kg); s is the
latent heat of vaporization of refrigerant (kJ/kg).
The enthalpies in Eq. (3) can be evaluated by using a set of
curve-tted equations [38]. By assuming an isenthalpic expansion
process, hei,r is equal to the refrigerant enthalpy before the TEV,
which can subsequently be determined based on the correspond-
ing pressure and temperature.
Fig. 7 compares the time varying xei,r for the two scenarios,
illustrating that they both increase with operating time. This is
because as tw increases, the amount of heat rejection of the pro-
totype drops. Therefore, more ash gas has to be generated to cool
down the remaining liquid refrigerant to the required evaporating
temperature. Between the two scenarios, it can be seen that xei,r for
the wPCM scenario is constantly lower, especially after the early
stage. This can be explained by the lower tw (Fig. 4) and thus a
better heat rejection performance of the prototype for the wPCM
scenario.
In the above, the refrigerant pressure uctuations, ow rates
and vapor contents for the two scenarios are explained and
compared to conrm a better refrigerant side performance for the
wPCM scenario.
3.3. Space cooling and water heating capacities
The differences in refrigerant side operation characteristics
explained above will result in different space cooling and water
heating capacities between the two scenarios as discussed below.
In this study, the space cooling capacity (Qcl) of the prototype is
determined based on the enthalpy drop of air across the evaporator
(Eq. (4)), while the air enthalpy (ha) is quantied by using Eqs.
(5)e(7).


Qcl ma her;a  hes;a (4)
 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403 1401

Fig. 7. Refrigerant vapor content comparisons.

Fig. 8. Space cooling capacity comparisons.

ha 1:006ta w2501 1:86ta
4Psw
w 0:622
Pb  4Psw
ln Psw C1 =Ta C2 C3 Ta C4 Ta2 C5 Ta3 C6 ln Ta
where ma is the evaporator supply air ow rate (kg/s); her,a and hes,a
are the evaporator return and supply air enthalpies (kJ/kg),

respectively; ta is the air dry-bulb temperature ( C); w is the
moisture content (kg/kg dry air); 4 is the relative humidity; Psw is
the saturation vapor pressure over liquid water (Pa); Pb is the at-
mosphere pressure (Pa); C1 to C6 are the regression coefcients
indicated in Table 5 [37]; Ta is the absolute air temperature (K).
Similarly, the water heating capacity (Qwh) of the prototype is
deduced from the enthalpy drop of refrigerant in the heat
exchanger:


Qwh mr hhi;r  hho;r
(5) increase and tends to be stable, Qcl drops correspondingly and ap-
proaches a steady state.
Fig. 9 compares the time varying Qwh for the two scenarios at 20-
(6) min intervals. It is reasonable to see that Qwh for the wPCM scenario
is constantly larger. The average difference is 16.1%. This is attrib-
uted to the use of PCM, as identied previously, to lower the rate of
(7) increase in tw during heat charging, which can enhance the heat
transfer performance of the heat exchanger.
Regarding the heating capacity proles, it is noted that Qwh
drops signicantly with operating time for both scenarios. This can
be explained by the increase in tw to decrease the temperature
difference between the refrigerant and the water. The heat transfer
performance of the heat exchanger is thus adversely affected [39],
leading to reduced amount of vapor refrigerant condensed in the
heat exchanger. For the same reason, it can be seen that Qwh for
both scenarios are rather stable in the late stage due to the corre-
sponding slower increase in tw (Fig. 4).
It is evident from the above that both the space cooling and
water heating capacities of the prototype can benet from the
(8) enhanced heat storage capacity of the water tank.

where hhi,r and hho,r are the refrigerant enthalpies at the heat
exchanger inlet and outlet (kJ/kg), respectively.
Fig. 8 compares the time varying Qcl for the two scenarios at 20-
min intervals. It can be seen that owing to the slightly lower tw
(Fig. 4) and xei,r (Fig. 7) for the wPCM scenario, the associated Qcl is
larger than that for the woPCM scenario by 5.4% on average.
Regarding the cooling capacity proles, it is noted that despite
the rather stable mr for both scenarios (Fig. 6), Qcl drops substan-
tially in the early stage and becomes stable afterwards. The drop is
attributed to the increase in xei,r (Fig. 7), resulting in less amount of
liquid refrigerant available for vaporization. While xei,r continues to

Table 5
Summary of the regression coefcients in Eq. (7).

Coefcient Constant

C1 5.800E03
C2 1.391E00
C3 4.864E-02
C4 4.176E-05
C5 1.445E-08
C6 6.546E00
Fig. 9. Water heating capacity comparisons.
1402 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403

Fig. 11. Energy performance comparisons for different outdoor temperature

conditions.
Fig. 10. Overall energy performance comparisons.

3.4. Overall energy performance
In this study, the overall energy performance is quantied by the
use of overall coefcient of performance (COPoa), which takes into
account the space cooling and water heating capacities as well as
the power consumption of the prototype. COPoa is mathematically
presented as:
Qcl Qwh
COPoa
W
where W is the power consumption of the prototype (kW).
Fig. 10 compares the time varying COPoa for the two scenarios. It
is noted that COPoa drops gradually with operating time. According
to the previous discussions, this is attributed to the corresponding
decrease in Qcl and Qwh. Moreover, it is unveiled from Fig. 10 that
COPoa for the wPCM scenario is on average 8.9% higher than that for
the woPCM scenario.
To review if the above ndings are also valid for other experi-
mental conditions, the time-weighted averages (Y) of Qcl, Qwh, W
and COPoa for both scenarios are determined for different outdoor
temperature conditions by using the following equation:
Z SEHRAC could improve the system performance was investigated in
a
Xtdt
0
Y (10)
a
revealed from the heat discharging test described previously in
Section 2.4.2.
Fig. 12 shows the time varying average water (tw) and PCM (tp)
temperatures for the two scenarios. In the gure, tp is taken as the
arithmetic mean of the temperatures measured at positions T1 to
T8 in the PCM container (Fig. 3). It can be seen that the proles for
tw are similar in the early stage. But they start to deviate from each
other when tp drops below 43.6  C. This is because the PCM has
reached the solidication temperature range (43.6  Ce42.3  C)
(9) when the parafn in it starts to turn from liquid to solid. The latent
heat released therefore heats up the water to result in a lower rate
of drop in tw.
In the experiments, it was found that the time elapsed for the
water to cool down from 50  C to 35  C is 68.4 h for the wPCM
scenario, and is 56.5 h for the woPCM scenario. The results indicate
that the use of PCM as an internal heat source to partially offset the
water tank heat loss can extend the heat retention time by 21.1%.
4. Conclusions
Whether the use of PCM in the heat recovery water tank of

where X is the parameter being averaged; t is the operating time; a

is the time spent for the entire experiment.
Results are shown in Fig. 11. It is noted that a rise in outdoor
temperature decreases Qcl and simultaneously increases Qwh to
result in a cancellation effect in system output (sum of Qcl and Qwh).
Therefore, the system output displays only a slight decrease with an
increase in outdoor temperature from 25  C to 35  C. The decrease,
combined with a marginal increase in W, results in a drop in COPoa
with outdoor temperature for both scenarios. It is further noted in
Fig. 11 that COPoa for the wPCM scenario is constantly higher, with a
difference ranging from 6.9% to 9.8% for different experimental
conditions, primarily owing to the larger system output.

3.5. Heat retention time

The heat retention time, dened as the time elapsed for the tank
water to drop below the desired temperature (35  C), can be Fig. 12. Water and PCM temperatures during heat discharging.
 J. Jia, W.L. Lee / Energy 93 (2015) 1394e1403
this study. To facilitate the investigations, a prototype SEHRAC in-
tegrated with an expanded graphite/parafn composite PCM was
designed and setup in a test facility for laboratory experiments.
Both heat charging and discharging tests were carried out. For the
heat charging test, two identical sets of experiments with and
without the use of PCM (wPCM and woPCM scenarios) under a
range of outdoor temperature conditions were conducted. It was
conrmed that the PCM could function as an internal heat sink to
lower the rate of water temperature increase and to enhance the
heat storage capacity of the water tank. As a result, the space
cooling and water heating capacities for the wPCM scenario, as
compared to the woPCM scenario, were 5.4% and 16.1% larger,
respectively. The overall coefcient of performance was therefore
found 6.9%e9.8% higher. While for the heat discharging test, it was
revealed that the use of PCM could extend the heat retention time
of the tank water by 21.1%. The results of this study conrmed that
the performance of SEHRAC could be improved by the use of PCM.
The conclusion, together with the experimental protocol estab-
lished, will be useful for further development of SEHRAC.
Acknowledgment
This work is supported by Hong Kong Research Grants Council
General Research Fund No. 519412.
References
[1] D'Valentine ML, Goldschmidt VW. Mathematical model of a refrigerant
desuperheater for heat recovery. ASHRAE Trans 1989;95:215e23.
[2] Lee AHW, Jones JW. Thermal performance of a residential desuperheater/
water heater system. Energy Convers Manag 1996;37:389e97.
[3] Jia J, Lee WL. Applying storage-enhanced heat recovery room air-conditioner
(SEHRAC) for domestic water heating in residential buildings in Hong Kong.
Energy Build 2014;78:132e42.
[4] Shimoda Y, Okamura T, Yamaguchi Y, Yamaguchi Y, Taniguchi A, Morikawa T.
City-level energy and CO2 reduction effect by introducing new residential
water heaters. Energy 2010;35:4880e91.
[5] Techarungpaisan P, Theerakulpisut S, Priprem S. Modeling of a split type air
conditioner with integrated water heater. Energy Convers Manag 2007;48:
1222e37.
[6] Yokoyama R, Shimizu T, Ito K, Takemura K. Inuence of ambient temperatures
on performance of a CO2 heat pump water heating system. Energy 2007;32:
388e98.
[7] Fatouh M, Elgendy E. Experimental investigation of a vapor compression heat
pump used for cooling and heating applications. Energy 2011;36:2788e95.
[8] Nizetic S, Coko D, Marasovic I. Experimental study on a hybrid energy system
with small- and medium-scale applications for mild climates. Energy 2014;75:
379e89.
[9] Ying WM. Performance of room air conditioner user for cooling and hot water
heating. ASHRAE Trans 1989;95:441e4.
[10] Jiang ML, Wu JY, Xu YX, Wang RZ. Transient characteristics and performance
analysis of a vapor compression air conditioning system with condensing heat
recovery. Energy Build 2010;42:2251e7.
[11] Monerasinghe NJ, Ratnalingam R, Lee BS. Conserved energy from room air-
conditioners for water heating. Energy Convers Manag 1982;22:171e3.
[12] Jia J, Lee WL. Experimental study of the application of intermittently operated
SEHRAC (storage-enhanced heat recovery room air-conditioner) in residential
buildings in Hong Kong. Energy 2015;83:628e37.
[13] Jia J, Lee WL. Experimental investigations on the use of capillary tube and
thermostatic expansion valve in storage-enhanced heat recovery room air-
conditioner. Energy Build 2015;101:76e83.
1403
[14] Sharma A, Tyagi VV, Chen CR, Buddhi D. Review on thermal energy storage
with phase change materials and applications. Renew Sust Energ Rev
2009;13:318e45.
[15] Pandiyarajan V, Pandian MC, Malan E, Velraj R, Seeniraj RV. Experimental
investigation on heat recovery from diesel engine exhaust using nned shell
and tube heat exchanger and thermal storage system. Appl Energy 2011;88:
77e87.
[16] Maruoka N, Akiyama T. Exergy recovery from steelmaking off-gas by latent
heat storage for methanol production. Energy 2006;31:1632e42.
[17] Sun XQ, Zhang Q, Medina MA, Liu YJ, Liao SG. A study on the use of phase
change materials (PCMs) in combination with a natural cold source for space
cooling in telecommunications base stations (TBSs) in China. Appl Energy
2014;117:95e103.
[18] Chen H, Lee WL. Combined space cooling and water heating system for Hong
Kong residences. Energy Build 2010;42:243e50.
[19] Cabeza LF, Castell A, Barreneche C, de Gracia A, Ferna
ndez AI. Materials used
as PCM in thermal energy storage in buildings: a review. Renew Sust Energ
Rev 2011;15:1675e95.
[20] Zalba B, Marn JM, Cabeza LF, Mehling H. Review on thermal energy storage
with phase change: materials, heat transfer analysis and applications. Appl
Therm Eng 2003;23:251e83.
[21] Agyenim F, Hewitt N, Eames P, Smyth M. A review of materials, heat transfer
and phase change problem formulation for latent heat thermal energy storage
systems (LHTESS). Renew Sust Energ Rev 2010;14:615e28.
[22] ASHRAE. ASHRAE standard 16 e method of testing for rating room air con-
ditioners and packaged terminal air conditioners. Atlanta, USA: American
Society of Heating; 1983. Refrigerating and Air-Conditioning Engineers.

C, Medrano M, Roca J, Cabeza LF, Garca D. Natural convection
[23] Castell A, Sole
heat transfer coefcients in phase change material (PCM) modules with
external vertical ns. Appl Therm Eng 2008;28:1676e86.
[24] Xia L, Zhang P, Wang RZ. Preparation and thermal characterization of
expanded graphite/parafn composite phase change material. Carbon
2010;48:2538e48.
[25] Ottinger O, Bacher J. Material mixtures for heat storage systems and pro-
duction method. US Patent 7704405 B2; 2010.
[26] Cabeza LF, Ib
an~ ez M, Sole

C, Roca J, Nogue

s M. Experimentation with a water
tank including a PCM module. Sol Energy Mater Sol Cells 2006;90:1273e82.
[27] ASHRAE. ASHRAE standard 41.3 e standard method for pressure measure-
ment. Atlanta, USA: American Society of Heating; 1989. Refrigerating and Air-
Conditioning Engineers.
[28] ASHRAE. ASHRAE standard 116 e methods of testing for rating seasonal ef-
ciency of unitary air conditioners and heat pumps. Atlanta, USA: American
Society of Heating; 2010. Refrigerating and Air-Conditioning Engineers.
[29] BSI. British standard 5141 e specication for air heating and cooling coils.
London, UK:British Standards Institution; 1975.
[30] Doebelin EO. Measurement systems application and design. 3rd ed. New York,
USA: McGraw-Hill; 1983.
[31] Holman JP. Experimental methods for engineers. 5th ed. New York, USA:
McGraw-Hill; 1989.
[32] ISO. ISO standard 5151 e non-ducted air conditioners and heat pumps:
testing and rating for performance. Geneva, Switzerland: International Or-
ganization for Standardization; 2010.
[33] Chan ALS, Chow TT, Fong SKF, Lin JZ. Generation of a typical meteorological
year for Hong Kong. Energy Convers Manag 2006;47:87e96.
[34] Agyenim F, Hewitt N. The development of a nned phase change material
(PCM) storage system to take advantage of off-peak electricity tariff for
improvement in cost of heat pump operation. Energy Build 2010;42:1552e60.
[35] Bong TY, Hawlader MNA, Mahmood W. Inuence of expansion device on the
performance of an air-conditioner with a desuperheater. ASHRAE Trans
1988;94:661e72.
[36] Dossat RJ. Principles of refrigeration. 4th ed. New Jersey, USA: Prentice-Hall;
1997.
[37] ASHRAE. ASHRAE handbook e fundamentals. Atlanta, USA: American Society
of Heating; 2009. Refrigerating and Air-Conditioning Engineers.
[38] Ding GL, Wu ZG, Liu J, Inagaki T, Wang KJ, Fukaya M. An implicit curve-tting
method for fast calculation of thermal properties of pure and mixed re-
frigerants. Int J Refrig 2005;28:921e32.
[39] Bejan A. Heat transfer. New York, USA: John Wiley & Sons; 1993.