Experimental Performance Analysis of Supercritical CO2

Thermodynamic Cycle Powered by Solar Energy

X. R. Zhanga*, H. Yamaguchia, K. Fujimab, M. Enomotoc, and N. Sawadad
a
Department of Mechanical Engineering, Doshisha University, Kyoto 630-0321, Japan
b
Mayekawa MFG. Co., Ltd., 2000 Tatsuzawa Moriya-city, Ibaraki-Pref., 302-0118, Japan
c
Showa Denko K. K., 1-480, Inuzuka, Oyama-city, Tochigi 323-8679, Japan
d
Showa Tansan Co., Ltd., 7-1, Ogimachi, Kawasaki-Ku, Kawasaki-city, Kanagawa, 210-0867, Japan

Abstract. The interests in using carbon dioxide as working fluid increase since the Montreal and Kyoto Protocols were
made. In this paper, a complete effort was made to study the performance of CO2 Rankine cycle powered by solar
energy experimentally. The system utilizes evacuated solar collectors to convert CO2 into high-temperature supercritical
state, used to produce electrical energy and thermal energy, which could be used for air conditioning and hot water
supply and so on. The system performances were tested not only in summer, but also in winter; not only in sunny day,
but also in cloudy day. The interest of the paper is the solar collector efficiency, because the absorbed heat quantity in
the collector can be utilized for power generation and heat supply and other useful outputs. The results show that
annually-averaged solar collector efficiency was measured at about 60.4%. The study shows the potential of the
application of the solar powered CO2 cycle as a distributed power/heat generation system.
Keywords: Supercritical CO2, Solar Energy, Rankine Cycle, Solar collector, Efficiency.
PACS: 42.79.Ek

INTRODUCTION HFC refrigerants is less restricted than in certain

European countries, carbon dioxide research has been
By looking back on the history of carbonic systems, vigorous and of high quality though small in volume
it can be easily seen that CO2 is an ‘old’ refrigerant [1]. for a large economy [3]. However, in the last four
As the CFC fluids were introduced in the 1930s and years the increase in the practical use of carbon
1940s, these ‘safety refrigerant’ eventually replaced dioxide has been small, relative to the size of the
the old working fluids in most applications. There is heating/cooling market worldwide.
no single reason why the use of CO2 declined, but a CO2 is the non-flammable and non-toxic fluid
number of factors probably contributed. These factors (friendly to environment). In addition, its vapor
included high-pressure problems, aggressive pressure is much higher and its volumetric
marketing of CFC products, low-cost tube assembly in refrigeration capacity is 3−10 times higher than CFC,
competing systems, and a failure of CO2 system HCFC, HFC and HC fluids. Thus, the system volume
manufactures to improve and modernize the design of would be smaller and the working fluid charge would
systems and machinery. be lower with CO2. In the range of lower (30 ℃) to
In 1992, Lorentzen and Pettersen [2] published the moderate (200 ℃) temperature heat source, a
first experimental results on a prototype CO2 system thermodynamic analysis was made among the working
for automobile air conditioning. Based on their results fluids for sake of comparison, such ammonia, water,
and other results, the interest in CO2 as a working fluid HFC-134a etc [4]. The analysis shows that there is an
increased considerably since the 1990s, in spite of obvious efficiency advantage of using carbon dioxide
resistance from the fluorocarbon industry and as working fluid in the temperature range shown above.
conservative parts of the automotive industry. At the But supercritical operation of CO2 needs high pressure
same time, the research activity has been provoked by equipments. In 2004, Yamaguchi and Zhang et al. [5]
the Montreal and Kyoto Protocols that the synthetic published a thermodynamic cycle ― solar energy
HFC refrigerants
CREDITmay LINEbe (BELOW)
made difficult
TOorBE impossible
INSERTED ONpowered RankinePAGE
THE FIRST cycleOFusing
EACHsupercritical
PAPER carbon
to obtain by legislation. Even inEXCEPT
the US, where use of
THE PAPER BEGINNING ON P. 64

CP832, Flow Dynamics, The Second International Conference on Flow Dynamics

edited by M. Tokuyama and S. Maruyama
© 2006 American Institute of Physics 0-7354-0324-4/06/$23.00
419
dioxide. Figure 1 shows a schematic diagram of the The solar collector is the heart of the
CO2-based solar Rankine cycle. thermodynamic cycle. Its characteristics play an
important role in the successful operation of such
systems. In the present study, all-glass evacuated solar
collectors with a U-tube heat removal system are used.
A sketch of the solar collectors is shown in Figure 3.
Efficient area of the solar collectors used in the
Rankine system is about 9.6 m2. A detailed description
of the used collectors can be seen in the Ref. [6].
To date there is no turbine available for
supercritical CO2. Therefore, in the experiment, a
pressure relief valve was used, instead of a turbine, in
order to complete the Rankine cycle. The pressure
relief valve can provide various extents of opening for
the cycle loop in order to simulate pressure drop
occurring in realistic turbine condition and
consequently a thermodynamic cycle can be achieved.
The two shell and tube heat exchangers are used
for achieving heat recovery. The higher-temperature
FIGURE 1. Schematic diagram of solar energy powered water and lower-temperature water are respectively
Rankine cycle using carbon dioxide. provided to the heat exchanger 1 and 2 to simulate the
Based on the previous study [5, 6], it can be known heat recovery processes. The total heat exchanger area
that the trans-critical cycle powered by solar energy utilized is about 0.76 m2. A mechanical-draft water
can be achieved using CO2 as working fluid. The cooling tower with a cooling capacity of 22 kW is used
objective of the present paper is to investigate the CO2 as a heat sink, dissipating heat recovered from the
Rankine cycle performance by experimental work and Rankine cycle to the ambient.
to have basic data on the cycle powered by solar A high-accuracy measurement and data acquisition
energy. system is used to achieve real-time data measurement,
data acquisition, and processing. Meteorological data,
EXPERIMENTAL SET-UP such as solar radiation, atmospheric temperature
values, can be acquired through the meteorological
Figure 2 shows a schematic diagram of the instruments installed in the experiment system, as
experimental set-up constructed. The experimental set- shown in Fig.2. Accuracies of sun radiation sensor and
up is mainly comprised of solar collector arrays, a air temperature gauge are ± 0.3 % and
pressure relief valve, liquid CO2 feed pump, heat ± 0.15 + 0.0002 t ℃, respectively. In addition, 5
exchangers and cooling tower. In addition, a thermal couples and 5 pressure transmitters are
measurement and data acquisition system is also mounted in the CO2 loop to measure CO2 temperatures
included. and pressures, respectively, with accuracy of ± 0.1 ℃
for temperature measurement and ± 0.2% for pressure
measurement. There are 5 platinum resistor
temperature sensors mounted to measure water
temperatures at the inlets and outlets of the heat
exchangers with accuracy of ± 0.15 + 0.0002 t ℃. CO2
mass flow meter is mounted at the downstream of
pump exit with accuracy of ± 0.1% and two water
flow meter mounted in the heat recovery systems with
accuracy of ± 0.5 %. Therefore, the accuracies of the
cycle parameters defined in the equations above are
calculated to be less than ± 1.0%. The measuring
points of all the sensors are shown in Fig. 2. Thermal
insulation coating is installed for carbon dioxide and
water loops to reduce heat losses from piping.
FIGURE 2. Schematic diagram of the experimental set-up.

420
 PERFORMANCE ANALYSIS loop, and the CO2 flow rate under the turbine condition
should be larger than that of the present condition.
No electricity power is outputted in the Therefore, the power generation and other useful
experimental test, but basic cycle performance for the outputs in the true turbine condition are larger than the
power/heat production can be known based on values estimated in the present analysis, which
thermodynamic estimations, although we are aware represent the minimal ones.
that the outlet temperature and thermo-physical
properties are somewhat different between the true RESULTS AND DISCUSSION
turbine condition and the present condition. The
following thermodynamic equation is used to calculate During the test time, the inlet water temperature and
the power generation from the cycle: flow rate for the heat exchanger 1 are respectively
controlled at 30.0 ℃ and 830.0 L/h (0.230 kg/s) and
W p = mcη p ( h − h ) (1)
ti to for the heat exchanger 2 at 9.0 ℃ and 200.0 L/h (0.056
Where mC is mass flow rate of CO2; η p generator kg/s). During the test hours, the pressure relief valve
was adjusted to a state of half close. The supercritical
efficiency, a value of 0.95 is used; hti , hto specific high-pressure side and subcritical low-pressure side of
enthalpy value of CO2 at the inlet and outlet of the the CO2–based Rankine cycle were kept at about 8.0
pressure relief valve. and 5.5 MPa respectively. All the experimental tests
The important point of estimating the performance are carried out in the Kyoto area of Japan.
of the CO2 Rankine cycle is how much heat energy can Solar radiation and air temperature measured with a
be obtained from the solar collector, because the heat function of time are shown in Figure 3. Summer and
energy can be utilized as useful energy for power winter weathers were selected, both of which are not
generation and heat recovery and so on. So in the so sunny, not so cloudy, and represent typical summer
present study, the solar collector efficiency is used as and winter weather condition. The experimental tests
the most important index to estimate the CO2 cycle were carried out respectively in the summer and winter.
performance. The solar collector efficiency is defined The time-averaged solar radiation and air temperature
as a ratio of describing the performance of collecting are about 0.58 kW/m2 and 36.3 ℃ in the summer, and
heat using supercritical CO2 in the collector, 0.30 kW/m2 and about 17.0 ℃ in the winter.
t
∫ m c ( h so − h si ) dt
ηs = 0 t (2) 1.0
[a]
45
Air temperature
∫ IAdt 40
Sol ar r adi at i on ( kW/m)

Solar radiation
2

Ai r t emper at ur e ( ℃ )
0.8
0 35
The following efficiencies are defined to describe 0.6
the performance of the CO2 thermodynamic cycle, 30

Wp 0.4 25
COPp = (3)
20
Qs
0.2
15
Qr
COPh = (4) 0.0 10
Qs 5:00 7:00 9:00 11:00 13:00 15:00 17:00 19:00
Heat quantity recovered from the cycle can be 1.0 35
[b]
calculated, 30
Sol ar r adi at i on ( kW/m)
2

Ai r t emper at ur e ( ℃ )

0.8
Qr = mc ( hto − h pi ) (5) Air temperature 25

where η s is collector efficiency, represents a ratio of 0.6 20

Solar radiation
the heat quantity absorbed into CO2 in the collector 0.4 15
during a time period of t to the total solar energy 10
striking the collector surface during the time; COPp 0.2
5
power generation efficiency; COPh heat recovery
efficiency; Qs solar energy striking the collector surface, 0.0 0
5:00 7:00 9:00 11:00 13:00 15:00 17:00 19:00
Qsolar=IA; I solar radiation; A efficient area of the solar
Ti me ( hh: mm)
collector.
It should be mentioned here that using the pressure FIGURE 3. Variations of solar radiation and air temperature
relief valve artificially reduces the CO2 flow rate in the measured with time, (a) summer; (b) winter.

421
Figure 4 shows the measured CO2 temperatures in the 200 1.2
Collector outlet temperature
Rankine cycle loop, in which the measured CO2 flow Relief valve outlet temperature
[a]

CO2 f l ow r at e ( kg/mi n)
rate is also included. It can be seen that the CO2 flow 1.0

CO2 t emper at ur e ( ℃ )
150
rate achieved in the cycle loop is about 0.7 kg/min and 0.8
0.5 kg/min, respectively in the summer and winter
condition, which is kept throughout most of the 100 0.6
CO2 flow rate
daylight time. It can be seen that the CO2 temperature Pump inlet temperature
at the solar collector outlet reaches up to about Pump outlet temperature 0.4
50
180.0 ℃ and 140.0℃ at noon, respectively in the
0.2
summer and winter day. The averaged temperature at
the collector outlet is about 160.0 ℃ for the summer 0 0.0
10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
day and 120.0 ℃ for the winter day. Such a 200 1.2
temperature achieved even in the winter day, makes it [b]
easy to collect thermal energy from the cycle and

CO2 f l ow r at e ( kg/mi n)
Collector outlet temperature 1.0

CO2 t emper at ur e ( ℃ )
Relief valve outlet temperature
achieve power generation in the turbine. Furthermore, 150
0.8
the time-averaged CO2 temperature at the outlet of the
pressure relief valve is respectively about 135.0 ℃ and 100 0.6
98.0 ℃, a relatively high temperature, possible to be
0.4
utilized for boiling water and air conditioning etc. 50
CO 2 flow rate
Pump inlet temperature
Figure 5 shows the calculated solar energy striking Pump outlet temperature 0.2
the collector and the heat quantity absorbed into CO2
in the collector per collector area and per hour based 0
9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00
0.0

on the measured data. Thermodynamic and transport Time (hh:mm)

properties of CO2 are calculated based on the
measured temperatures and pressures using a Program FIGURE 4. The measured CO2 temperatures and flow rate
Package for Thermophysical Properties of Fluids in the Rankine cycle loop, (a) summer; (b) winter.
database version 12.1 (PROPATH 12.1). During the
test time period, the averaged amount of solar
radiation is about 1.98 MJ/(m 2 ⋅ h) , the heat quantity 3.500

collected in the collector is about 1.40 MJ/(m 2 ⋅ h) for 3.000 Solar radiation per hour
Heat quantity collected per hour
the summer day and 1.17 MJ/(m 2 ⋅ h) and 0.71 2.500
[a]
MJ/(m 2 ⋅ h) for the winter day. Average collector
MJ/m2h

2.000

efficiency is estimated at 63.0% and 52.0%, 1.500

respectively. The result shows that even in the winter 1.000
day, the collector used is also effective in collecting
heat using supercritical CO2, which may explain to a 0.500

certain extent why a relatively high temperature of 0.000

10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
160.0 ℃ and 120.0 ℃ are achieved at the collector
2.000
outlet respectively for the summer and winter day. 1.800 Solar radiation per hour
Furthermore, the useful energy outputs and the 1.600 Heat quantity collected per hour
efficiencies are shown in Fig. 6. From Fig. 6, it can be 1.400
[b]
seen that Wp and Qr obtained are found to be relatively 1.200
MJ/m2h

1.000
stable throughout the test hours. No abrupt variations 0.800
of the useful outputs with time occur even in the 0.600
cloudy time, which is very different from the output of 0.400

solar cell. The time-averaged Wp is estimated at 0.29 0.200

0.000
kW and 0.18 kW for the summer and winter day, 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00
respectively. The time-averaged Qr is estimated at 4.16 Time (hh:mm)
kW and 1.23 kW respectively. The time-averaged
FIGURE 5. The amount of solar radiation and heat
COPp is found to be 6.25% and 5.30%, and COPh is
quantity collected in the solar collector per area and per hour,
48.5% and 35.3%. The total efficiency is estimated to (a) summer; (b) winter.
be about 54.8% and 40.6%, respectively. It is noted
that the power consumption of the liquid CO2 feed because the pump used is originally designed for water.
pump is not considered into the efficiencies above, If the power consumption is considered, COPp would

422
drop. Therefore a higher efficiency of CO2 feed pump 1.0 40
or solar energy powered pump will be considered to Air temperature (a)

Sol ar r adi at i on ( kW/m)

reduce the power consumption. It is also seen from Fig. Solar radiation

2
0.8 35

Ai r t emper at ur e ( ℃ )
6 that COPh is much higher than COPp, and therefore,
to a certain extent, the benefit obtained from the CO2- 0.6 30
based cycle also largely stems from the reasonable
utilizations of the recovered heat quantity. 0.4 25

6 1.0 0.2 20
Heat recovery Heat recovery efficiency (a)
5 0.0 15
0.8
Useful output (W)

5:00 7:00 9:00 11:00 13:00 15:00 17:00 19:00

4 1.0 40

Efficiency
0.6 (b)

Sol ar r adi at i on ( kW/m)

2
3 Air temperature

Ai r t emper at ur e ( ℃ )
0.8 Solar radiation
35
0.4
2
Power generation 0.6 30
Power generation efficiency 0.2
1

0.4 25
0 0.0
10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
3 0.8 0.2 20
(b)

H eat recovery efficiency 0.0 15

Useful output (kW)

0.6
5:00 7:00 9:00 11:00 13:00 15:00 17:00 19:00
2
H eat recovery Ti me [ hh: mm]
Efficiency

0.4
FIGURE 7. Variations of solar radiation and air
1 temperature measured with time, (a) sunny day; (b) cloudy
Pow er generation Pow er generation efficiency 0.2
day.

0 0.0
10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
3.500
Solar radiation per hour
T i me ( hh: mm)
Heat quantity collected per hour
3.000
(a)
FIGURE 6. Variations of the useful cycle outputs and 2.500
efficiencies, (a) summer; (b) winter.
MJ/(m2 h)

2.000

An experimental test was also carried out for the 1.500

cycle performance comparison between sunny day and 1.000
cloudy day. The tests were carried out in summer day.
0.500
Figure 7 shows solar radiation and air temperature
measured both in sunny day and cloudy day. Figure 8 0.000
10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
shows the solar energy and the heat quantity absorbed 3.500
Solar radiation per hour
into CO2 in the collector. Figure 9 shows the CO2 3.000
Heat quantity collected per hour

pressures measured. Average collector efficiency is 2.500

(b)

estimated at 62.0% and 52.3%, respectively. The result

MJ/(m2 h)

2.000
shows that even in the cloudy day, the collector used is
also effective in collecting heat using supercritical CO2 1.500

and the collector efficiency would not drop too much. 1.000
Furthermore, an annually experimental data was 0.500
presented in Fig. 10, which shows the monthly-
0.000
averaged solar energy and heat quantity absorbed into 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
CO2 in the collector. The experiment tests were carried Time [hh:mm]
out from July of 2004 to June of 2005. During the tests, FIGURE 8. The amount of solar radiation and heat quantity
the pressure relief valve was adjusted to a state of two- collected in the solar collector per area and per hour, (a)
third open. Because of the maintenances in 2004 sunny day; (b) cloudy day.
September and 2005 March, there is no data obtained

423
 10 increased greatly. The absorbed heat energy in the
[a] collector will also be used as useful output, power
9 generation, heat supply and so on.
CO2 pressure (MPa)

8 CONCLUDING REMARKS
Pump outlet pressure
Collector outlet pressure
7 Relief valve outlet pressure The basic performance of the CO2 Rankine cycle
Pump inlet pressure
powered by solar energy was investigated
6 experimentally. The interest of the paper focuses on
the solar collector efficiency, because the obtained
5 heat energy in the collector is considered as useful
10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
energy output from the thermodynamic cycle. The
10
[b] results show supercritical CO2 can effectively collect
heat in the evacuated solar collector. An annually-
9
averaged collector efficiency was measured at about
CO2 pressure (MPa)

60.4%. If in the real turbine condition, the value has a

8
potential of being further increased, because CO2
Pump outlet pressure
velocity is increased in the loop. Furthermore, the
7
Collector outlet pressure results show that the stable outputs (power generation
Relief valve outlet pressure
Pump inlet pressure and heat recovery) from the CO2 cycle can be achieved
6
not only in not only in summer time, but also in winter
day. The collector efficiency would not drop greatly in
5
cloudy day, compared to that in sunny day. A further
9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00
investigation is needed to understand the natures for
Time [hh:mm]
the flow and heat transfer of supercritical CO2 in the
FIGURE 9. Variations of CO2 pressures measured with time evacuated solar collector. And a turbine condition is
in the cycle loop, (a) sunny day; (b) cloudy day.
needed to further investigate the CO2-based Rankine
3 Solar radiation per hour cycle.
Heat quantity collected per hour
2.5
2 ACKNOWLEDGMENTS
MJ/m2h

1.5
This study was supported by the Academic Frontier
1
Research Project on “Next Generation Zero-Emission
0.5 Energy Conversion System” of Ministry of Education,
0 Culture, Sports, Science and Technology, Japan.
1 2 3 4 5 6 7 8 9 10 11 12
Month REFERENCES
FIGURE 10. The monthly-averaged solar radiation and heat
quantity collected in the solar collector per area and per hour. 1. M. H. Kim, J. Pettersen and C. W. Bullard, Progress in
Energy and Combustion Science 30, 119−174 (2004).
for the two months. It can be seen from the result that 2. G. Lorentzen and J. Pettersen, “New possibilities for
the monthly-average value of solar radiation increases non-CFC refrigeration” edited by J. Pettersen, IIR
with month until it reaches a maximum value in International Symposium on Refrigeration, Energy and
August, and then, the solar radiation begins to decrease. Environment, Trondheim, Norway, 1992, pp. 29-34.
3. J. Fleming, Carbon dioxide as the working fluid in
The annually-averaged solar energy striking the heating and/or cooling systems, Bulletin of the IIR-No
collector was measured at 1.82 MJ (m 2 ⋅ h) . The solar 2003-4.
collector efficiency varied from 43.0% to 70.0%. An 4. X. R. Zhang, H. Yamaguchi, D. Uneno, K. Fujima, M.
annually-averaged value of 60.4% was obtained for the Enomoto and N. Sawada, Renewable Energy, (in press).
collector efficiency. The value is encouraging, because 5. H. Yamaguchi, X. R. Zhang, K. Fujima, M. Enomoto
and N. Sawada, Applied Thermal Science, submitted for
the pressure relief valve artificially reduced the CO2
publication.
flow rate and in the real turbine condition, the CO2 6. X. R. Zhang, H. Yamaguchi, K. Fujima, M. Enomoto
velocity is increased by passing through a nozzle. So and N. Sawada, JSME International Journal, Series. B
the heat convective processes in the collector tubes can 48, 540-547 (2005).
be enhanced greatly in the turbine condition and then
the collector efficiency has a potential of being

424