This article was downloaded by: [Bibliotheek TU Delft]

On: 22 December 2014, At: 05:41

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

HVAC&R Research
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/uhvc20

Development of high-efficiency carbon dioxide

commercial heat pump water heater
a

Michael Petersen , Chad Bowers , Stefan Elbel

a

a b

& Pega Hrnjak

a b

Creative Thermal Solutions, Inc., 2209 North Willow Road, Urbana , IL 61802 , USA

Department of Mechanical Science and Engineering , University of Illinois , UrbanaChampaign , IL , USA

Accepted author version posted online: 19 Aug 2013.Published online: 25 Oct 2013.

To cite this article: Michael Petersen , Chad Bowers , Stefan Elbel & Pega Hrnjak (2013) Development of high-efficiency
carbon dioxide commercial heat pump water heater, HVAC&R Research, 19:7, 823-835
To link to this article: http://dx.doi.org/10.1080/10789669.2013.833543

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

HVAC&R Research (2013) 19, 823835

C 2013 ASHRAE.
Copyright 
ISSN: 1078-9669 print / 1938-5587 online
DOI: 10.1080/10789669.2013.833543

Development of high-efficiency carbon dioxide commercial

heat pump water heater
MICHAEL PETERSEN1,, CHAD BOWERS1, STEFAN ELBEL1,2, and PEGA HRNJAK1,2
1

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Creative Thermal Solutions, Inc., 2209 North Willow Road, Urbana, IL 61802, USA
Department of Mechanical Science and Engineering, University of Illinois, Urbana-Champaign, IL, USA

Although heat pump water heaters are widely accepted in both Japan and Europe, where energy costs are high and government
incentives for their use exist, acceptance of such products in the United States has been limited. This trend is slowly changing with
the introduction of heat pump water heaters into the residential market, but acceptance remains low in the commercial sector.
The objective of the presented work is the development of a high-efficiency R744 heat pump water heater of approximately 35kW (10-ton) heating capacity for commercial applications with effective utilization of the cooling capability for air conditioning
and/or refrigeration. This unit will be targeted at commercial use where some cooling load is typically needed year round, such
as restaurants, hotels, nursing homes, and hospitals. The improvement process concentrated on the heat exchangers of the system.
Further optimization potential was identified by investigating the gas cooler as well as the expansion device of the heat pump water
heater by using a two-phase ejector. In addition, a comparison to a commercially available baseline R134a unit of the same capacity
and footprint was made where significant package size reduction potential of the R744 heat pump water heater was discovered as
well as performance improvement, especially at high water temperature lifts.

Introduction
In todays world with increasing costs of energy, heat pump
water heaters (HPWHs) offer great potential to reduce energy consumption in water heating applications. HPWHs are
widely accepted in both Japan and Europe, where energy costs
are high and government incentives for their use exist. Acceptance of such a product in the United States has been slow,
with approximately 15,000 units sold in 2009 (D & R International 2010) compared to a few hundred thousand per year
in Japan with the help of generous government and utility
incentives (Heat Pump & Thermal Storage Technology Center of Japan [HPTCJ] 2012). Barriers to HPWH acceptance
have historically been performance, reliability, as well as initial
costs. The dominant styles of water heaters used today in the
United States are still electric resistance and gas, split roughly
50/50 in market share (D & R International 2010). The technology for these systems is quite mature, and all have primary
Received January 31, 2013; accepted August 6, 2013
Michael Petersen, MSc, is Research Engineer. Chad Bowers, PhD,
Associate Member ASHRAE, is Senior Research Engineer. Stefan Elbel, PhD, Member ASHRAE, is Adjunct Professor at Department of Mechanical Science and Engineering and Chief Engineer at Creative Thermal Solutions. Pega Hrnjak, PhD, Fellow
ASHRAE, is Research Professor in Department of Mechanical Science and Engineering and President of Creative Thermal
Solutions.

Corresponding author e-mail: michael.petersen@

creativethermalsolutions.com

energy efficiencies less than 1. Commercial water heaters are

rated by a combination of thermal efficiency and passive losses
to ambient, called stand-by losses. Heat pumps used for water
heating are essentially refrigeration machines and have been in
use for many years. Thermal efficiency, or heating coefficient
of performance (COP), of electrically driven heat pumps using
conventional refrigerants (R22, R134a, R410A) are in the 3
to 5 range compared to 0.8 to 0.95 seen in gas and electric
resistance water heaters, respectively.
The use of the natural refrigerant R744 as a refrigerant
for heat pumps is relatively new, starting first in the automotive industry but quickly moving to residential HPWHs in
the mid-1990s in Japan with sponsorship from the Japanese
government. Many Japanese manufacturers now have fully
commercialized residential R744 HPWHs, collectively called
Eco-Cute, which have become fully accepted as the most
promising technology for reducing Japans dependence on oil.
Many studies were published in the last two decades investigating the potential and improvement of R744 refrigeration cycles focusing mainly on the heat exchangers and the expansion
device. For heat pumps, one of the most attractive features of
using R744 as the working fluid is the utilization of the temperature glide in the gas cooler of the system, which reduces losses
in the heating process as described by Neksa et al. (1998). Stene
(2005) used a tripartite heat exchanger to reduce conduction
losses in the gas cooler and to utilize different temperature levels for hot water as well as space heating. Stene (2007) noted
that using R744 allowed for the production of hot water in the
temperature range from 60 C to 85 C (140 F to 185 F) while
obtaining the highest possible COP for an HPWH system.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

824
Kim et al. (2005) described the influence of an internal heat
exchanger (IHX) on an R744 HPWH system performance.
They saw COP improvement while capacity decreased due to
trade-offs between effectiveness and pressure drop in the IHX.
Sarkar (2006) found the main COP-influencing parameters of
an R744 heat pump system for simultaneous heating and cooling to be compressor speed and discharge pressure as well as
coolant temperatures for evaporator and gas coolers. The optimization process of an ejector expansion R744 heat pump
cycle was investigated by Sarkar (2008), where such ejector parameters as entrainment ratio and pressure lift and their effect
on COP were described. A theoretical comparison of R744
and R134a in a tap water heat pump was described by Cecchinato et al. (2005), where competitive results can be achieved
for R744 by using its beneficial properties.
This article presents the development and improvement
process of an R744 HPWH that was performed in several steps
investigating the heat exchangers and the expansion device of
the system. This HPWH was compared to a commercially
available R134a HPWH of the same footprint. The retrofit
of the working fluid R134a with R744 was done to improve
system performance characteristics and, at the same time, to
show the influence of the higher volumetric capacity of R744
compared to R134a.

Experimental facility
The experimental facility was designed according to ASHRAE
Standard 118.1 (ASHRAE 2003), which describes the testing

Fig. 1. Experimental facility.

HVAC&R Research
of Type IV HPWHs that can be operated without a water
storage tank. The heat pump system was installed in an environmental chamber in order to provide steady ambient conditions. The unit was instrumented in such a way as to achieve
energy balances on the cooling side and heating side of the cycle. The accuracy of the experimental results was within 5%
for the COP on the cooling and heating sides. On the cooling
side of the cycle, the balance was achieved on the air stream
and refrigerant stream. The cooling capacity on the air stream
was determined using a separate wind tunnel directly connected to the evaporator air discharge of the heat pump unit.
This wind tunnel was built and instrumented according to
ASHRAE Standard 37 (ASHRAE 2005). The heat pump unit
was equipped with a blower to provide airflow over the evaporator. The power consumption of this blower as a function of
air temperature was determined and added to the compressor
power to receive the total HPWH power consumption. After
determining the blower power consumption, it was removed
from the unit because it was not strong enough to generate the
necessary pressure head to overcome the pressure drop caused
by the flow nozzles used to determine airflow rate. Airflow was
provided by the wind tunnel blower that was able to provide
the pressure lift required to maintain a constant airflow rate
of 1800 l/s (3800 CFM) for all tests and in all system configurations. The second determination of the cooling capacity of
the system was made through measurements obtained on the
refrigerant flow stream of the evaporator of the system. As an
example of the installed instrumentation in the experimental
facility, a schematic of the R744 HPWH with IHX is shown
in Figure 1.

Volume 19, Number 7, October 2013

825

Table 1. Instrumentation.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Description
Temperature
Pressure

T-type thermocouple
Refrigerant: Sensotec TJE
Air: Setra 265

Mass flow
Mass flow
Humidity
Power

Water/glycol: MicroMotion DS100

Refrigerant: MicroMotion DS25
General Eastern: Model D2
Ohio Semitronics: GW5-008DY22

On the heating side of the system, three different energy

determination methods were employed. Using the instrumentation on the refrigerant cycle described above, the heating
capacity can be determined from the temperatures, pressures,
and mass flow values at the inlet and outlet of the condenser/gas cooler. The second heating capacity determination
was made using temperature and mass flow measurements on
the water stream. In order to reject the heat input into the
water stream by the heat pump, a glycol chiller was used. The
third determination of heating capacity was made on this glycol stream. For this, the glycol temperature was measured to
determine the specific heat necessary for the capacity determination. The glycol/pump facility was designed and constructed to transfer heat between the hot water stream and the
cold glycol stream through a brazed plate heat exchanger. In
addition to the brazed plate heat exchanger, the facility was
also equipped with mass flow meters for each fluid stream,
trim heaters, and a pump to control the water flow rate. The
instruments that were used and their accuracy are summarized
in Table 1.

Investigated systems
The investigation started with an evaluation of a commercially available R134a HPWH marketed toward indoor applications. This baseline unit was a packaged air-source HPWH
with a nominal heating capacity of 35 kW (10 ton) and COP
of 3.9. A scroll compressor was used with a condenser of
brazed plate design. The evaporator was a round tube plate
fin design, and the expansion device was a thermostatic expansion valve. This baseline system was then compared to an
R744 system of the same footprint. An R744 compressor that
would provide similar capacity at the rating condition for the
baseline R134a system was chosen. This compressor was of a
semi-hermetic reciprocating design. It has been demonstrated
several times before (Bullard 2004; Elbel and Hrnjak 2008)
that the performance of a transcritical R744 system can be
optimized using the high-side pressure. An electronic expansion device was used to vary the high-side pressure during
testing.
In addition to possible performance benefits of using R744
as the working fluid, the high volumetric capacity of R744
enables the construction of the same capacity unit with a substantially reduced system volume. Displaying this unit package

Range
200 C350 C (328 F662 F)
0138 bar (02000 PSIG)
0625 Pa
(02.5 W)
0455 kg/min (01000 lbs/min)
020 kg/min (045 lbs/min)
35 C25 C Td (1%100% RH)
060 kW (017ton)

Accuracy
0.5 K
0.1% full scale
1% full scale
0.05% (liquid) of rate
0.2 C
0.2% full scale

reduction potential was a secondary aim of this work beside

the COP improvement process. The volume reduction was
demonstrated in the reduced height of the unit, which was
primarily achieved through a more compact heat exchanger
size. The reduction of the evaporator volume was achieved by
using a round tube plate fin evaporator coil with a fin density
of 16 fins per inch and a tube diameter of 3/8 in., originally
designed for a 17-kW (4.8-ton) R134a system environmental
control unit where it was used successfully, as described by Elbel and Hrnjak (2010). This resulted in a 40% reduction in face
area, primarily in height, and a 55% reduction in evaporator
volume.
The first step in the system improvement process focused on
the modification of the heat exchangers of the system, namely
the IHX, the evaporator, and the gas cooler. The IHX consisted of microchannel tubes assembled in a sandwich configuration having the high-temperature stream in the center and
low-temperature streams on the outside. Two identical components in parallel set up in counter flow were used in order to get
sufficient capacity for the heat transfer from the high-pressure
refrigerant after the gas cooler to the low-pressure refrigerant
after the evaporator. An improved IHX of brazed plate design
was used in a later step of the improvement process. This component had similar capacity but lower pressure drop; it was
designed to provide an effectiveness of approximately 70%.
For the following improvement steps of the R744 HPWH
refrigeration system the full R134a housing volume was used.
The additional height allowed the use of a larger evaporator.
For this purpose, the baseline R744 evaporator volume was
enlarged by 50% by combining it with half of a second identical
component creating an evaporator of almost the same height
as the R134a evaporator. A comparison of the dimensions of
the three evaporators that were used is shown in Figure 2.
The R744 baseline gas cooler was a commercially available
model with a much narrower design compared to the R134a
condenser. The reduction of the gas cooler volume was approximately 50%. In the following system development steps,
a performance improvement potential was identified by reducing conduction losses, especially at high water outlet temperatures, as described by Kim et al. (2004). For this, a staged gas
cooling process using multiple brazed plate heat exchangers
was used to reduce conduction effects compared to one compact gas cooler. Four brazed plate heat exchangers were used
with 9 kW (2.5 ton) each. A comparison of the dimensions of
the three heat exchangers is shown in Figure 3.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

826

HVAC&R Research

Fig. 2. Evaporator dimensions.

The previously described R744 HPWH systems were

equipped with two electronic expansion valves (EEVs) that
were set up in parallel to provide sufficient refrigerant flow.
These valves allowed convenient control of the high-side pressure during operation. Part of the improvement process was
the use of an electronically actuated ejector that was used to reduce throttling losses during the expansion process. A system
schematic and the corresponding cycle plotted in the pressure
specific enthalpy diagram describe the main differences when
using an ejector compared to an expansion valve, as shown in
Figure 4.
The ejector system consists of two sides. A driven lowpressure side with the suction flow and a driving high-pressure
side with the motive flow. The high-pressure side contains
the gas cooler and high-pressure side of the IHX. After the
IHX, the motive flow enters the ejector. The ejector can be
compared to a pump in which the high-pressure motive flow
pumps the low-pressure suction flow. The motive flow enters
the ejector and is accelerated in the motive nozzle. The suction
flow is entrained by the high-velocity motive flow in the suction chamber and both streams mix in the mixing chamber.

In the diffuser, the kinetic energy of the refrigerant stream is

converted into pressure energy, creating an intermediate pressure that reduces the compressor power. After the ejector, the
refrigerant flow enters the phase separator that divides the liquid and the vapor phases of the flow. The liquid fraction is
expanded and enters the evaporator as a low-quality suction
flow. The vapor flow enters the low-pressure side of the IHX,
and the cycle starts over again.
For better control and more flexibility during testing, the
ejector was equipped with a stepper-motor-controlled needle
to vary the area of the motive nozzle. The ejector is of modular
design to allow the modification of the component dimensions
as described by Elbel et al. (2012). A schematic of the ejector
that was used with its main sections and the two refrigerant
streams is shown in Figure 5.
The investigated R744 HPWH system configurations are
summarized in Table 2.

Results and discussion

The strategy of the investigation was a comparison of the
R744 development stages to the R744 baseline system to
concentrate on the improvement potential of each step. Gas
cooler water inlet temperature and water flow rate variations were investigated for the first four R744 development
steps (BEVAP [baseline evaporator], BEVAP+MC IHX
Table 2. Investigated R744 HPWH system configurations.
MC BP
BEVAP EEVAP IHX IHX MGC EEV EJECTOR

Fig. 3. Gas cooler/condenser dimensions.

1
2
3
4
5
6a
6b

X
X

X
X
X
X
X
X

X
X

X
X
X

X
X
X
X
X
X
X

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Volume 19, Number 7, October 2013

827

Fig. 4. R744 HPWH ejector system schematic and cycle plotted in pressure specific enthalpy diagram (color figure available online).

[BEVAP+microchannel IHX], EEVAP [enhanced evaporator], and EEVAP+MC IHX). The investigation of the multiple gas coolers and the ejector was done at selected conditions
to show their performance improvement potential. Finally a
comparison of the R744 performance to the R134a system
allowed a comparison of the two fluids. The R744 HPWH development stages were investigated at the condition the R134a
unit as rated by the manufacturer. The rating temperature of
the water and air inlet of the HPWH is 26.7 C (80 F). The
water flow rate was adjusted to reach a water temperature lift
of 5 K, which was a water flow rate of approximately 1760 g/s
(3.9 lbs/s) for all systems. All systems were tested under these
conditions as well as under a broader range of water inlet
temperatures as shown in Table 3.
The combined COP was the main interest of the investigation. It was calculated according to Equation 1 as the ratio
of the useful output (heating and cooling capacity) divided
by the system power consumption caused by compressor and
system fan power consumption:
COPComb =

QEVAP + QGC
.
WHPWH

(1)

For the R744 systems, the high-side pressure at each condition was optimized to provide the highest heating COP when

operating in the transcritical mode. Even though this high

water flow rate was required by the standard to rate the system, this may not represent actual operation of such units,
especially if higher water temperature lift conditions are required. In order to understand these effects of different water temperature lifts on system performance and gain insight
into a better control strategy, the water flow rate was reduced
from the rating value of 1760 g/s (3.9 lbs/s) down to 1000
and 400 g/s (2.2 and 0.9 lbs/s). This was done for water inlet temperatures of 12 C, 26.7 C, and 50 C (54 F, 80 F, and
122 F) while maintaining the air inlet temperature at 26.7 C
(80 F). This ambient condition was chosen because this model
is marketed as an indoor unit, meaning the heat pump will be
pumping heat from the buildings indoor environment to the
water stream. Applications in which this would be most beneficial are situations where there is a constant need for simultaneous hot water and cooling. Restaurant kitchens and
laundry facilities are both excellent examples of such locations. The following paragraphs describe the HPWH improvement steps by using different IHX designs and an enhanced
evaporator. These improvements were investigated for different gas cooler water inlet temperatures and water flow rates.
In addition, the improvement potentials when using multiple gas coolers and an ejector were determined for selected
conditions.

Fig. 5. Ejector schematic with main sections and refrigerant flows (color figure available online).

828

HVAC&R Research

Table 3. System performance test conditions.

Air side

Temperature
Flow rate

26.7 C (80 F)
1800 l/s (3800 CFM)

Water side

Temperature

12 C, 26.7 C, 50 C (54 F,
80 F, 122 F)
400, 1000, and 1760 g/s (0.9,
2.2, and 3.9 lbs/s)

Flow rate

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

IHX and evaporator

An IHX allows heat transfer from the high-pressure refrigerant after the gas cooler to the low-pressure refrigerant leaving
the evaporator. The heating and cooling capacity increase by
further subcooling the refrigerant after the gas cooler, and
therefore, lower evaporator inlet quality is offset by the increasing compressor power consumption caused by higher
suction temperatures. Overall, net thermodynamic benefits
can be seen for R744 systems (Kim et al. 2004) because the
COP-optimizing high-side pressure is lower when an IHX is
used. The IHX that was used in the first system improvement
step was a microchannel heat exchanger.
The improvement on the evaporator side was accomplished
by increasing the component size by 50%. Consequently, the
component volume as well as face area were increased while
the air face velocity decreased. The combined COPs of the four
R744 development steps that were investigated at an ambient
temperature of 26.7 C (80 F) and different gas cooler water
inlet temperatures as a function of water temperature lifts are
shown in Figure 6.
The trends that can be seen in Figure 6 confirm the typical behavior of HPWH systems caused by varying gas cooler
water inlet temperature and water flow rate. The system per-

Fig. 6. Comparison of combined COP as a function of water

temperature lifts at different gas cooler water inlet temperatures
for R744 HPWH development stages.

formance decreases with increasing gas cooler water inlet temperatures. The refrigerant outlet temperature is determined by
the water inlet temperature because of no pinch temperature
during single-phase heat transfer in the gas cooler. Therefore,
a close approach between refrigerant outlet temperature and
water inlet temperature is very important for an optimum in
combined COP. Pettersen et al. (1998) described a 5% cooling
COP increase per degree of approach temperature reduction
in transcritical operation in the gas cooler. Higher water inlet
temperatures with high compressor discharge pressure levels showed lower combined COP values because more input
power was needed compared to lower water inlet temperatures.
The combined COP was more dependent on water temperature lifts at low water inlet temperatures. At 50 C (122 F) water inlet temperature, the results were almost steady, whereas
at 12 C (54 F), the COP dropped significantly when going
to larger water temperature lifts. This larger COP gradient at
low water inlet temperatures can be explained with the large
temperature differences in the gas cooler and the resulting potential losses. At high water inlet temperatures, the refrigerant
gas cooler outlet temperature varied less at increasing water
temperature lifts compared to lower water inlet temperatures.
Therefore, a much stronger decreasing effect on the COP was
seen at low water inlet temperatures compared to high water
inlet temperatures. In other words, the COP level at the 50 C
(122 F) condition was already relatively low compared to the
12 C (54 F) condition, which was much more sensitive to a
decreasing water flow rate that created larger temperature lifts.
Overall, the EEVAP+MC IHX system is considered the best
configuration for a wide range of gas cooler water inlet temperatures and water flow rates. Even though the EEVAP system
shows good performance for different conditions, its performance gets worse at increasing water inlet temperatures. The
IHX system suffers from the low-side pressure drop, especially
at low water inlet temperatures that cause a small compressor
pressure ratio. However, at high water inlet temperatures with
large pressure ratios, the relative pressure drop effect is less
severe, which helps to compensate the performance decrease
at high water inlet temperatures.
In order to better understand the influence of the heat
exchanger modifications (IHX, evaporator) on the system performance, the cycles were compared in an R744
pressure-specific enthalpy diagram (Figure 7).
The systems without an IHX (BEVAP, EEVAP) operated
at a higher high-side pressure than the systems with an IHX
(BEVAP+MC IHX, EEVAP+MC IHX). For the systems
without an IHX, a constant superheat at the evaporator outlet
of 5 K was used. This was done to prevent liquid refrigerant
from entering the compressor, which could lead to damage.
The IHX systems operated at a lower optimized high-side
pressure resulting from more subcooling and therefore lower
evaporator inlet qualities. The evaporator outlet condition was
kept constant at a quality of 0.95 by adjusting system charge
in order to receive moderate superheat on the compressor
suction side. The superheat of the systems with IHX was approximately 16 K. This larger amount of superheat caused a
lower suction density at the compressor inlet, which led to
smaller refrigerant mass flow rate.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Volume 19, Number 7, October 2013

Fig. 7. Comparison of R744 HPWH cycles in pressure versus specific enthalpy diagram at rating condition (color figure available
online).

The full potential of cycle improvements when using the

IHX was not realized due to a relatively high pressure drop
on the low-pressure side (Figure 8). The pressure drop on the
suction side was examined for all four systems in order to
compare the influence of the IHX. For the systems without
an IHX, the low-side pressure drop was measured across the
inlet and outlet of the evaporator. For the cycle with IHX, the
low-side pressure drop was measured between evaporator and
compressor inlet.
Two important effects of the IHX on the system can be
seen in Figure 8. As described earlier, the systems suffered
from a lower suction density, which decreased the refrigerant
mass flow rate. The second effect was a large pressure drop
prior to the suction of the compressor. This pressure drop had
to be compensated during the compression process, increasing
the compressor power consumption. Consequently, the overall
combined COP increases were offset. A minimization of these

829

Fig. 9. Pressure ratio versus water temperature lift at rating water

and air temperatures.

pressure losses would have further improved the COP values

significantly.
Another characterizing parameter used to analyze the influence of the IHX was the pressure ratio of the system, which
was calculated by the ratio of compressor discharge and suction pressure according to Equation 2:
=

pCp,Out
.
pCp,In

The pressure ratio at different water temperature lifts at an

air and water inlet temperature of 26.7 C (80 F) is shown in
Figure 9.
The pressure ratios of the systems with enhanced evaporator (EVAP and EVAP+IHX) were less dependent on the
water temperature lift. The increased low-side pressure for
these systems kept the pressure ratio on a lower level, which
was beneficial for the compressor power consumption. The
system without an IHX showed the best COP of the investigated systems at this condition due to the offsetting effect of
the low-side pressure drop in the IHX.
The evaporator approach temperature difference as defined
in Equation 3 describes the temperature difference between air
outlet and refrigerant inlet saturation temperature at evaporator inlet pressure. The saturation temperature was used to
get a good comparison for all four systems because different
evaporator outlet conditions were used as described earlier:
TEVAP,Approach = TEVAP,A,Out TEVAP,R,In,Sat .

Fig. 8. R744 HPWH suction pressure drop at different refrigerant

mass flow rates at rating water and air temperatures.

(2)

(3)

The smaller the approach temperature difference between

the two fluid streams was, the more effective was the heat exchanger. An approach temperature difference of 0 represents a
theoretical heat exchanger with infinite heat transfer area. The
results for the evaporator approach temperature difference are
shown in Figure 10. The dominant positive effect on the evaporator approach temperature difference is the increase of the

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

830

HVAC&R Research

Fig. 10. Comparison of evaporator approach temperature difference at different water temperature lifts for R744 HPWH at rating
water and air temperatures.

evaporator size from the BEVAP to the EEVAP systems. The

systems with an IHX reach even lower approach temperature
differences mainly due to the improved heat transfer characteristics caused by lower evaporator inlet qualities.
Another characteristic factor that had to be considered
for the modification of the evaporator was the overall heat
transfer coefficient U, which was calculated for the air side
by using the air-side heat transfer area and the logarithmic
mean temperature difference (LMTD) of the evaporator, as
described in Equation 4:
UEVAP,air =

QEVAP
,
AEVAP,A LMT D

(4)

Fig. 11. Evaporator overall heat transfer coefficient versus refrigerant mass flow rate for R744 HPWH at rating water and air
temperatures.

fore increased the evaporation pressure. At the same time, the

overall heat transfer coefficient was reduced.
The experimental test results were used to make predictions about HPWH performance at large water temperature
lifts, creating high gas cooler water outlet temperatures. For
this, the trends that were obtained from the test results of the
R134a system were extrapolated to make predictions about
the systems energy savings potential of R744 over R134a as a
function of water outlet temperature (Figure 12). The R134a
system cannot reach the high water temperatures required
in sanitary applications without additional supplement heating due to a limit of the compressor discharge temperature
of 107 C (225 F) specified by the manufacturer. That is why
an additional electric resistance heater was used to provide

where
LMTD
(TEVAP,A,Out TEVAP,R,Out,Sat ) (TEVAP,A,In TEVAP,R,In,Sat )
=
.
TEVAP,R,Out,Sat )
(T
ln EVAP,A,Out
(TEVAP,A,In TEVAP,R,In,Sat )
(5)
The airflow rate was kept constant during the tests. That is
why the results for the overall heat transfer coefficient U are
shown as a function of refrigerant mass flow rate in Figure 11.
The evaporator was modified to utilize the volume that was
provided by the R134a HPWH housing. For this, the evaporator volume was enlarged by 50%. This larger heat exchanger
volume led to an increase in heat transfer area, which helped in
decreasing the approach temperature difference between the
air and refrigerant side. The low-side pressure increased, which
reduced the necessary compressor work to provide the highside pressure. The trends for evaporator approach temperature
difference and overall heat transfer coefficient confirmed the
behavior that was expected. The increase of the heat transfer
area reduced the approach temperature difference and there-

Fig. 12. Energy savings of R744 HPWH systems compared to

R134a at various water outlet temperatures at rating water and
air inlet temperatures.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Volume 19, Number 7, October 2013

heating capacity for water outlet temperatures above 82 C
(180 F). This meant that solely HPWH operation was done
up to a water outlet temperature of 82 C (180 F). Above that
temperature, a hybrid operation of combined HPWH usage
and additional supplemental resistance heating with a COP of
1 was used to make up the remaining temperature difference.
For the R744 system with internal heat exchange, up to a
22% savings in energy consumption were achieved in comparison to the R134a system. The baseline R744 system achieved
an energy savings potential of 9%, which can be seen above
90 C (194 F). The assumption of a 100% efficient electric resistance heater does not represent the situation in real systems
and is therefore a conservative choice. For the situation with a
gas condensing water heater with an efficiency of 0.85 for the
additional heating capacity above 82 C (180 F), even larger
improvement potential can be seen. For the BEVAP system,
an improved performance of 14%, and for the BEVAP+MC
IHX system, 27% improvement at 95 C (203 F) water outlet temperature over R134a can be reached. The location of
the crossover points in performance between R134a and R744
were interesting for several reasons. The first and primary reason was that the temperature region in which R744 began to
excel over R134a was noticeably higher than what would be
needed in residential water heating applications but was exactly where the water needed to be heated to in commercial
applications. According to the U.S. Department of Health and
Human Services (Food Code 2005), fresh hot sanitizing water for ware-washing equipment should be no less than 74 C
(165 F) for stationary, single-temperature machines and no
less than 82 C (180 F) for all other machines. If a storage
device were to be implemented, water would likely need to be
produced at an outlet temperature at or above 85 C (185 F) to
account for losses encountered during storage. Currently, these
higher temperatures are achieved through secondary electric
resistance heating, starting somewhere between 43 C and 60 C
(109 F and 140 F). In the case of R744, no secondary heating
source would be needed; however, many installations would
likely still have this as a backup method of heating. In most
cases, even the primary heating source is electric or gas with
an energy factor of 1 or less. In these cases, the use of any heat
pump technology would result in significant energy savings,
especially if the free cooling is used to augment existing
building cooling.

831
heat transfer process differ from the ideal counter-flow configuration and, therefore, increases the approach temperature
difference between refrigerant and water side. In order to reduce these effects, the heat path must be interrupted. This can
be done for refrigerant to air gas coolers by preventing conduction through cutting fins, as described by Park and Hrnjak
(2007) or for water heating applications by a stepwise heat
transfer process using multiple heat exchangers. This stepwise
process not only reduces losses but also makes it possible to
use the different water temperature levels for domestic hot
water or hydronic heating, as described by Stene (2005).
The staged heating process was realized in the R744 HPWH
with four brazed plate heat exchangers that had the same combined capacity as the single component. These heat exchangers
had a smaller hydraulic diameter that created a larger pressure
drop compared to the single heat exchanger design. That is
why only lower water flow rates were investigated. However,
at the lower water flow rates, large temperature lifts were realized, and also the conduction effects that led to increasing
losses were most severe. Therefore, the investigation at these
conditions showed the greatest effects. One indicator for the
heat exchanger performance is a small approach temperature
difference between the refrigerant outlet and the water inlet, as
described in Equation 2; the importance of a close approach
between refrigerant and cooling fluid for an optimum in COP
was described many times before (Pettersen et al. 1998; Kim
et al. 2004):
TGC,Approach = TGC,R,Out TGC,W,In .

(6)

A comparison between the R744 HPWH systems with

a single gas cooler and multiple gas coolers is shown in
Figure 13.
The approach temperature difference as a function of water
temperature lift for the single and multiple gas cooler systems
showed the influence of conduction on the heat exchanger

Gas cooler
The transcritical operation on the high-pressure side of the
R744 refrigeration system is a gas cooling process with a gliding temperature instead of the constant temperature during
condensation of the R134a system. This gliding gas cooler
temperature profile is useful for heating applications where
the hot refrigerant rejects heat to the cooling fluid. However, there is also a larger potential for losses in the R744
gas cooler compared to the R134a condenser caused by the
very large temperature differences between hot gas coming
from the compressor and the cooling fluid (air, water). These
losses are often caused by conduction from the hot refrigerant
inlet to the cool refrigerant outlet. This heat bypass makes the

Fig. 13. Comparison of gas cooler approach temperature difference at ambient air and gas cooler water inlet temperature
of 26.7 C (80 F) and different water temperature lifts for R744
HPWH development stages.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

832

HVAC&R Research

Fig. 14. Comparison of combined COP at gas cooler water inlet

temperature of 26.7 C (80 F) and different water temperature
lifts for R744 HPWH development stages.

Fig. 15. Suction pressure drop of MC IHX and BP IHX at different refrigerant mass flow rates at rating temperatures and water
flow rate.

performance. The multiple gas cooler system showed much

lower approach temperature differences compared to the single gas cooler systems. These lower values represented better component effectiveness during the heat transfer process.
Since only small water mass flow rates were investigated that
created large water temperature lifts, a direct comparison was
only possible with the BEVAP and EEVAP systems. However,
the trends that were gained from the other systems at lower
water temperature lifts showed the larger approach temperature differences compared to the multiple gas cooler system.
The effect of the better heat transfer characteristics of the
multiple gas cooler system on the system COP can be seen in
Figure 14.
Performance improvement was seen for both combined
COP and heating capacity when using multiple gas coolers.
For large water temperature lifts, where conduction effects
are most severe, the multiple gas cooler system (EEVAP+MC
IHX+MGC [multiple gas cooler]) showed the best combined
COP values of the investigated R744 HPWH systems. A comparison between the EEVAP+MC IHX+MGC system and
the BEVAP system at a water temperature lift of 63 K shows
a 35.6% increase in combined COP for the multiple gas cooler
system. However, these good results were only achieved at
large water temperature lifts, where the benefits through an
improved heat transfer process outweigh the larger pressure
drop of the multiple gas cooler assembly. The situation would
change at smaller water temperature lifts, where the conduction effect would be less severe and the large pressure drop
would become more dominant, which would make the values
for combined COP less competitive compared to the single gas
cooler systems.

ment ratio is the ratio of ejector suction mass flow and the
motive mass flow:

Ejector
The performance of the ejector can be characterized by several
parameters, as described by Elbel (2007). The mass entrain-

m
Suction
.
m
Motive

(7)

The pressure lift describes the pressure difference between ejector diffuser outlet pressure and evaporator outlet pressure.
This pressure difference elevates the suction pressure compared to the evaporation pressure to an intermediate level and
therefore reduces the necessary compressor power:
pLift = pDiff ,Out pEVAP,Out .

(8)

Both mass entrainment ratio and pressure lift are coupled.

That means that optimizing these factors is a compromise between good results for either of them. For example, by opening
the liquid valve after the phase separator, more refrigerant suction mass flow is provided, which increases the entrainment
ratio. However, at the same time, the evaporation pressure
increases, which reduces the ejector pressure lift.
The pressure lift and its effect on the system performance
improvement show the importance of reducing the suction
pressure drop. That is why an enhanced IHX in brazed plate
design was chosen for the comparison between direct expansion (DX) and ejector system to realize the full potential of
the ejector. A comparison of microchannel IHX and brazed
plate IHX pressure drops as a function of refrigerant mass
flow rate showed the improved component performance (Figure 15). The brazed plate design (EEVAP+BP IHX [brazed
plate IHX]) created in average a 37% lower suction pressure
drop than the microchannel design.
The ejector efficiency is the main parameter to characterize
the performance of the component. It is defined as the ratio of
recovered work rate to the maximum possible expansion work

Volume 19, Number 7, October 2013

833

rate potential, as described by Elbel (2007):

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Ejector =

Rec
W
Rec,Max
W

(9)

In order to determine the effect of the ejector on the system

performance, the ejector test conditions were matched closely
with the DX system. The ejector condition was matched for
the heating capacity and the evaporator outlet temperature
and pressure simultaneously. In a second run, the heating capacity and the gas cooler outlet temperature and pressure were
matched. This approach of the investigation was done to cover
both influences of high- and low-pressure sides on the system
performance (Figure 16).
It must be pointed out that the ejector system performance
was not ideal due to low separation efficiencies in the phase
separator. This caused very low phase separator outlet qualities on the vapor port. Ideally, the vapor and the liquid port
are saturated, having 100% and 0% refrigerant qualities, respectively. However, the phase separator that was used only
provided lower separation efficiencies. That is why liquid refrigerant left the separator through the vapor port and vapor
refrigerant left the separator through the liquid port. A phase
separator with better separation efficiency would improve the
ejector and system performance. Lawrence and Elbel (2012)
described the influence of separation efficiency on the system
COP and pointed out that the COP improvement gained by
using an ejector can easily be canceled out by the effect of an
ineffective phase separator.
The test for the ejector and DX system were done at rating
water mass flow rate at an ambient temperature of 26.7 C
(80 F) and gas cooler water inlet temperatures of 37.8 C and
50 C (100 F and 122 F). The characteristic ejector parameters
that were recorded are shown in Figure 17.
The values for efficiency with 21% and 24%, respectively,
as well as the entrainment ratio with 59% and 67% confirmed
the good performance of the ejector. These were typical per-

Fig. 16. Ejector system suction and gas cooler exit condition
match with DX system (color figure available online).

Fig. 17. Characteristic ejector performance parameters.

formance values for ejector efficiency and entrainment ratio.

Comparable results were reported by Banasiak et al. (2012),
where a two-phase ejector in an R744 heat pump was used.
The pressure lift resulted from the optimized high-side pressure that was adjusted by the stepper-motor-controlled needle. The effect of the ejector on the system performance is
presented through the combined COPs in Figure 18.
The ejector system showed better combined COPs at the
two investigated water inlet temperatures; at 37.8 C (100 F),
an improvement of 5%, and at 50 C (122 F) water inlet temperature, an improved COP of 9% were seen.

Fig. 18. Ejector and DX system at matched gas cooler conditions

and heating capacities.

834

HVAC&R Research

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Conclusions
The development and improvement process of an R744
HPWH for commercial applications was described. The investigation aimed for a twofold objective. In the first step, the
system package reduction potential when using R744 compared to R134a was shown by realizing a more compact design
enabled by the beneficial fluid properties of R744. The component volume reduction was achieved mainly in the evaporator
of the system with a 40% smaller height and a 55% smaller
volume compared to R134a. In the second step, the IHX and
the evaporator of the R744 HPWH were improved, creating a total of four R744 HPWH systems. It was shown that
the combined COP increased during the improvement process
compared to the R744 baseline system. Trends were developed
based on the experimental data to predict the performance of
R744 compared to R134a. An energy savings potential of 22%
at a water outlet temperature of 95 C was discovered for the
R744 BEVAP+MC IHX system over R134a. This was mainly
caused by the necessary supplemental heating capacity that
the R134a system requires above water outlet temperatures
of 82 C due to compressor discharge temperature limitations.
This hybrid operation of HPWH and electric resistance heating is not necessary in the R744 HWPH that can provide large
temperature lifts in once through operation.
Further investigations were done to show the potential of
reducing conduction losses in the gas cooler by using multiple
heat exchangers. The multiple gas cooler system outperformed
the single gas cooler system with an enhanced evaporator as
well as the same system with an IHX for the combined COP by
35% and 22%, respectively. This improved performance was
achieved at large water temperature lifts when the conduction
effects have the strongest impact, and therefore, beneficial heat
transfer characteristics were more dominant than the larger
pressure drop in the component.
Finally, the effect of an ejector was investigated that was
used to reduce throttling losses during the expansion process.
The ejector test conditions were closely matched with a DX
system. At 37.8 C and 50 C (100 F and 122 F) gas cooler
water inlet temperatures, the ejector system showed 5% and
9% better combined COPs, respectively.

Nomenclature
BEVAP
BP
COP
DX
EEV
EEVAP
EJECTOR
HPWH
IHX
LMTD
MC
MGC
p
T

=
=
=
=
=
=
=
=
=
=
=
=
=
=

baseline evaporator
brazed plate heat exchanger
coefficient of performance
direct expansion
electronic expansion valve
enhanced evaporator (50% volume increase)
two-phase ejector
heat pump water heater
internal heat exchanger
logarithmic mean temperature difference, C
microchannel
multiple gas cooler
pressure, kPa
temperature, C

=
=
=
=

U
W

overall heat transfer coefficient, W/(m2K

power, kW
efficiency,
entrainment ratio,

Subscripts
A
Approach
BP
Comb
d
Diff
Ejector
EVAP
GC
In
Lift
Max
MC
Motive
Out
R
Ratio
Rec
Sat
Suction
W
X

= air
= approach
= brazed plate
= combined
= dew point
= diffuser
= ejector
= evaporator
= gas cooler
= inlet
= lift
= maximum
= microchannel
= motive flow
= outlet
= refrigerant
= ratio
= recovered
= saturation
= suction flow
= water
= expansion valve

Acknowledgment
The authors gratefully acknowledge the support given by the
U.S. Department of Energy. Project funding was made available through ARRA 2009 under contract DE-EE0003981.
The authors would also like to thank A.O. Smith Corporation
for their continued support in this project.

References
ASHRAE. 2003. ASHRAE Standard 118.1, Method of Testing for Rating
Commercial Gas, Electric, and Oil Service Water Heating Equipment.
Atlanta: ASHRAE.
ASHRAE. 2005. ASHRAE Standard 37, Methods of Testing for Rating
Electrically Driven Unitary Air-Conditioning and Heat Pump Equipment. Atlanta: ASHRAE.
Banasiak, K., A. Hafner, and T. Andresen. 2012. Experimental and numerical investigation of the influence of the two-phase ejector geometry on the performance of the R744 heat pump. International
Journal of Refrigeration 35:161725.
Bullard, C. 2004. Transcritical CO2 systemsrecent progress and new
challenges. Review article, Bulletin of the IIR. International Institute of Refrigeration, Paris, France.
Cecchinato, L., M. Corradi, E. Fornasieri, and L. Zamboni. 2005. Carbon dioxide as refrigerant for tap water heat pumps: A comparison
with the traditional solution. International Journal of Refrigeration
28:12508.
D & R International. 2010. Energy Star water heater market profile. U.S.
Department of Energy, Washington, DC.

Downloaded by [Bibliotheek TU Delft] at 05:41 22 December 2014

Volume 19, Number 7, October 2013

Elbel, S. 2007. Experimental and analytical investigation of a two-phase
ejector used for expansion work recovery in a transcritical R744 airconditioning system. Dissertation, University of Illinois, UrbanaChampaign, IL.
Elbel, S., and P. Hrnjak. 2008. Experimental validation of a prototype
ejector designed to reduce throttling losses encountered in transcritical R744 system operation. International Journal of Refrigeration
31(3):41122.
Elbel, S., and P. Hrnjak. 2010. Design, build-up, and performance investigation of a 35kW (10 ton) military environmental control unit using
transcritical R744 technology. Proceedings of International Refrigeration and Air Conditioning Conference, West Lafayette, IN, July
1215.
Elbel, S., C.D. Bowers, M. Reichle, J.M. Cristiani, and P. Hrnjak. 2012.
Vapor jet ejector used to generate free waste heat driven cooling in
military environmental cooling units. Proceedings of International
Refrigeration and Air Conditioning Conference, West Lafayette, IN,
July 1619.
Food Code. 2005. U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration, College Park,
MD.
Heat Pump & Thermal Storage Technology Center of Japan (HPTCJ).
2012. Data Book on Heat Pump and Thermal Storage System
20112012. Tokyo, Japan: HPTCJ.
Kim, M.H., J. Pettersen, and C.W. Bullard. 2004. Fundamental
process and system design issues in CO2 vapor compression
systems. Progress in Energy and Combustion Science 30:119
74.

835
Kim, S.G., Y.J. Kim, G. Lee, and M.S. Kim. 2005. The performance of a
transcritical CO2 cycle with an internal heat exchanger for hot water
heating. International Journal of Refrigeration 28:106472.
Lawrence, N., and S. Elbel. 2012. Experimental and analytical investigation of automotive ejector air-conditioning cycles using lowpressure refrigerants. Proceedings of International Refrigeration and
Air Conditioning Conference, West Lafayette, IN, July 1619.
Neksa, P., H. Rekstad, G.R. Zakeri, and P.A. Schiefloe. 1998. CO2 -heat
pump water heater: Characteristics, system design and experimental
results. International Journal of Refrigeration 30:38997.
Park, C.Y., and P. Hrnjak. 2007. Effect of heat conduction through the
fins of a microchannel serpentine gas cooler of transcritical CO2
system. International Journal of Refrigeration 21(3):1729.
Pettersen, J., A. Hafner, and G. Skaugen. 1998. Development of compact heat exchangers for CO2 air-conditioning systems. International
Journal of Refrigeration 21(3):18093.
Sarkar, J. 2006. Simulation of a transcritical CO2 heat pump cycle for simultaneous cooling and heating applications. International Journal
of Refrigeration 29:73543.
Sarkar, J. 2008. Optimization of ejector-expansion transcritical CO2 heat
pump cycle. Energy 33:1399406.
Stene, J. 2005. Residential CO2 heat pump system for combined space
heating and hot water heating. International Journal of Refrigeration
28:125965.
Stene, J. 2007. Integrated CO2 heat pump systems for space heating
and hot water heating in low-energy houses and passive houses.
International Energy Agency (IEA) Heat Pump ProgrammeAnnex
32, Kyoto, Japan.